{"full_text": "Hydrogen sulfide (H2S) is an acidic, corrosive, toxic, and harmful waste gas pollutant with a rotten egg smell [1\u20133]. It is commonly found as an impurity in raw natural gas and as a byproduct gas pollutant in crude oil processing, the coal industry, iron, and steel smelting, etc., with the presence of sulfur-containing. The combustion of H2S or the gas\u2212containing streams will produce highly toxic and corrosive byproducts (such as SO2, CS2, COS, H2SO4, etc.), which will again be harmful to cause a threat to the health of human beings and the ecological balance of the environment if they are freely discharged [4,5]. In addition, it is easy to cause the deactivation upon many metal catalysts such as Ni, Fe, and Pt due to the high toxicity of H2S [6,7]. Therefore, it is urgent to selectively remove H2S in industrial processes, which has derived great focus on both academic and practical perspectives [8]. More importantly, there has been an increase in interest in how to use the highly purified H2S separated from absorption separation.Ionic liquids (ILs), as a novel type of green solvent, have attracted great attention because of their unique properties, such as negligible vapor pressure, structural designability, high thermal stability, good affinity to gas, etc [9\u201314]. In terms of the good affinity for acidic gases by ILs, it has also been well explored in H2S capture with good selectivity and high absorption capacity [15\u201317]. Jou and Mather first investigated H2S absorption in 1-N-butyl-3-methylimidazolium hexafluorophosphate ([bmim][PF6]) and observed that H2S was only physical dissolution in the IL [18]. Pomelli and coworkers extended this investigation with a variable ion pair to demonstrate ILs with extremely high H2S absorption capacities [19]. Wu et\u00a0al. reported that cost-effective protic ionic liquids (PILs) derived from alkanolamines MDEA and formic acid or acetic acid were contributing to the absorption separation of H2S from CO2 [20]. As for the functionalized ILs, our group first developed 1-alkyl-3-methylimidazolium carboxylates for the highly efficient absorption of H2S [21]. Whereafter, tertiary amine and carboxylate group as two functionalized sites were incorporated simultaneously into triethylbutylammonium-based ILs toward selectively separating of H2S from CO2 [22,23]. We also synthesized a class of hydrophobic tertiary amine functionalized half protonated PILs paired with Tf2N for highly selective absorption of H2S from CO2 [24]. To achieve both high H2S uptake capacity and superior H2S/CH4 and/or H2S/CO2 selectivity, we prepared strongly basic azole-based PILs to separate H2S with the highest capture capacity up to 6.81\u00a0mol\u00a0kg\u22121 [25,26].The aim of H2S separation is to make it profitable. In the aspect of H2S conversion into an inorganic sulfur chemical, our group developed a highly efficient ILs-mediated Claus reaction under mild conditions [27,28]. Yu et\u00a0al. reviewed metal-based ILs for the efficient H2S absorption and oxidative conversion into Sulphur [29]. Wu et\u00a0al. disclosed a liquid-phase Claus reaction mediated in lactate-based ILs and ethylene glycol (EG) or H2O mixture [30]. It should be noted that the deep processing utilization of H2S discussed above is low-value-added sulfur. The conversion of H2S into value-added organic chemicals is another option that makes more sense.Recently, our group has made progress on the conversion of H2S into a value-added organosulfur compound mediated in ILs. For example, a series of task-specific hydrophobic PILs catalyzed H2S conversion with unsaturated acids into value-added sulfydryl acids was pioneered [31]. To prepare sulfydryl alcohols, two different pathway-the reaction between H2S and epoxides or unsaturated alcohols catalyzed by hydrophilic PILs-was developed [32,33]. In addition, Sen's group used the trihexyltetradecylphosphonium chloride (THTDPC) as the phase transfer catalyst for converting H2S into bis(2-phenylethyl) sulfide (PES) with a high product selectivity [34]. Although the above-mentioned ways have the advantages of solvent-free, high conversion, good recoverability of catalyst, and excellent substrate universality, the separation of reactant and product requires water addition to split them into two phases. The removal of water is highly energy-intensive because of the large latent heat required for the subsequent recovery of catalyst or purification of products [35]. Generally, the addition of a third component for the extraction operation is typically necessary to separate the catalyst from the product in a homogenous reaction. Can a direct phase separation procedure or methodology be designed once the reaction is finished for the resource utilization of H2S without involving the third component?Herein, we develop a novel H2S resource way through the reaction with \u03b1,\u03b2\u2212unsaturated carboxylate esters to prepare thiols mediated by functionalized carboxylate-based ILs because of their recognition for H2S. Importantly, the self-separation was realized to facilitate the progress of reactions and make the convenient separation of catalysts and products possible. The product can be controlled by regulating the molar ratio of H2S to \u03b1,\u03b2\u2212unsaturated carboxylate esters. The reaction was systematically investigated by means of condition optimization, product selectivity control, substrate expansion, etc. The gaussian calculation was performed to explore the reaction mechanism. It is anticipated that this self-separation method mediated by task specific ILs provides a potential strategy for utilizing H2S.The specifications and sources of the chemicals used in this work were summarized in Table S1. All reagents were used without further purification. In addition to the ILs which were directly purchased, the three ILs were prepared by one-step neutralization reaction, including tetraethylammonium acetate ([N2222][Ac]), tetraethylammonium 2\u2212hydroxypropanoate ([N2222][LAc]), and tetraethylammonium formate ([N2222][For]). In brief, taking the synthesis of [N2222][Ac] as an example, the equimolar acetic acid (HAc) was added into tetraethylammonium hydroxide 25\u00a0wt.% aqueous solution ([N2222][OH]), followed by stirring for 12.0\u00a0h at ambient temperature to form a transparent liquid. Then, the final product was purified on a Schlenk line at 60\u00a0\u00b0C and 0.15\u00a0mbar for 4.0\u00a0h until no bubbles evolved to ensure the [N2222][Ac] free of water.Nuclear magnetic resonance (NMR) spectra were obtained at ambient temperature on a Bruker DPX spectrometer with CDCl3 as the internal standard solvent at 400\u00a0MHz for 1H and 101\u00a0MHz for 13C NMR, respectively. Conversion and selectivity were determined by GC analysis (Shimadzu, GC-2014C). FT\u2212IR spectra (Tensor II, Bruker) were carried out with the spectral resolution and the number of scans of 4\u00a0cm\u22121 and 32, respectively. ESI-MS (Thermo Fisher) was used to characterize the molecular weight of the product at m/z. For the separation of the mixture after the reaction, the pure product was obtained by thin\u2212layer chromatography (TLC) with petroleum ether (PE) and ethyl acetate (EA) as eluent (3:1). The structures of thiols and thioethers were verified by 1H and 13C NMR spectra with CDCl3 as solvent.For typical experiments on the reaction of \u03b1,\u03b2-unsaturated carboxylate esters with H2S, butyl acrylate (1.0\u00a0g, 8\u00a0mmol) was added into a reaction vessel. The amount of catalyst is 10\u00a0mol% of butyl acrylate, and the reaction condition was 30\u00a0\u00b0C and 2.0\u00a0h. In order to ensure the equimolar reaction between H2S and substrate, the molar amount of H2S was calculated from its density (mol\u00b7L\u22121) according to our previous work [25,26,36,37]. Generally, the required H2S molar is ensured by adjusting the pressure drop (\u0394P) of the large tank. Since the substrate (butyl acrylate) can be accurately weighed, the required H2S can be determined by its mass (Eq. S1), so as to obtain the partial pressure of H2S at the corresponding temperature through the NIST database (DOI: https://doi.org/10.18434/T4D303, Eq. S2). The detailed description can be found in ESI.The characterizations of the catalyst ([Emim][Ac]) are presented in the ESI (Figs. S1\u2013S2). Taking butyl acrylate (1a) as a probe substrate of \u03b1,\u03b2-unsaturated carboxylate esters, we explored its reaction effect under different catalyst conditions (Table 1\n). The substrate butyl acrylate has no reaction activity with H2S without the addition of ILs catalyst in a blank experiment, indicating that the autocatalysis is negligible (Entry 1, Figs. S3\u2013S4). We also performed the investigations on the reaction of H2S and butyl acrylate with conventional [Emim][BF4] and [Emim][PF6] as catalysts (Entries 2\u20133). It is found that the conversion of the substrate is significantly restricted with a catalyst loading of 10\u00a0mol% under reaction conditions. This is because the alkalinity of [Emim][BF4] and [Emim][PF6] is too weak to activate H2S [33,38]. [Hmim][Cl] was also fed into the system to explore this reaction process (Entry 4). It is found that the conversion of butyl acrylate is only 6%, indicating that this catalyst has a poor activation effect on H2S. The main reason may be the fact that [Hmim][Cl] brings a relative acidic environment to the entire reaction system so that H2S can hardly be activated [19].When the [Emim][Ac] with weak alkalinity was fed to the system, a quantitative conversion of substrate was realized, indicating that the presence of carboxylate-based ILs is favorable for the conversion of butyl acrylate. As the reaction proceeds, the system comes to turbid and the [Emim][Ac] is split out in the lower phase because [Emim][Ac] is almost insoluble in the product mixture (Fig.\u00a01\na, see Fig.\u00a0S5 for an enlarged view). No signal of ILs was detected from the upper phase except for the mixed components of the two products, which can be assigned to butyl 3-sulfanylpropanoate (1b) and dibutyl 3,3\u2032-thiodipropionate (1c), respectively (Fig.\u00a01b, see Fig.\u00a0S6 for the corresponding 1H NMR). To investigate the self-separation process, the FT-IR spectra of pure butyl acrylate, butyl 3-sulfanylpropanoate, dibutyl 3,3\u2032-thiodipropionate, and upper phase have been carried out. As demonstrated in Fig.\u00a0S7, it is found that the FT-IR of upper phase are almost consistent with the dibutyl 3,3\u2032-thiodipropionate and butyl 3-sulfanylpropanoate, demonstrating the formation of products. In addition, the signal of [Emim][Ac] cannot be detected from the upper phase after reaction, which further verifies the reliability of self-separation of catalyst and products. The conversion of butyl acrylate (1a) was as high as 99%, and the selectivity of 1b and 1c were 7% and 93%, respectively (Entry 5, Table 1) at 30\u00a0\u00b0C and 2.0\u00a0h. The two products can be separated easily by TLC (PE/EA\u00a0=\u00a03:1). The structure characterization and molecular weight of 1\u00a0b and 1c were verified by NMR spectra and ESI\u2212MS, respectively (Fig.\u00a0S8\u2212S13).To evaluate the effect of ILs with different cations on the conversion of butyl acrylate, we carried out the reaction with other carboxylated-based ILs by increasing the chain length and changing the configuration of cations, including [Bmim][Ac], [Hmim][Ac], [Omim][Ac], [N2222][Ac], [Pyr12][Ac], and [Epy][Ac]. Clearly, the conversion and selectivity keep almost unchanged, indicating the effective catalytic role of the carboxylate (Entries 6\u201310), which means the cations have almost no effect on the conversion of butyl acrylate. To evaluate the influence of various carboxylate-based anion, [N2222][Ac], [N2222][Lac], and [N2222][For] were employed to be as catalysts for the conversion of butyl acrylate (Entries 11\u221213). It is found that the conversion of butyl acrylate is up to 99% with different catalysts, and the selectivity of 1\u00a0b and 1c is around 10% and 90%, respectively, further indicating that the carboxylate has a good activation to H2S.It is well known that almost all industrial gases are accompanied by humidity. In order to study the effect of water on the conversion of butyl acrylate, [Emim][Ac] containing 3\u00a0wt% H2O was used as the reaction medium (Entry 14). It is shown that the conversion of butyl acrylate was as high as 99% within 2.0\u00a0h, which was almost the same as those under anhydrous conditions. Therefore, the effect of H2O on the reaction of H2S and butyl acrylate can be negligible. We also prepared the potassium acetate (KAc) into a 4\u00a0mol\u00a0L\u22121 aqueous solution as the representative inorganic salt to catalyze the reaction at 30\u00a0\u00b0C for 2.0\u00a0h (Entry 15). In contrast, no conversion of butyl acrylate is discovered. This may be due to the obvious interface between KAc aqueous solution and butyl acrylate, which separates the two substances from each other, so the chemical mass transfer between them is greatly limited. Besides, the pH of KAc aq. (4\u00a0mol\u00a0L\u22121) was measured to be 8.3, which means its alkalinity is too weak to activate H2S. The carboxylate\u2212activated H2S cannot be transferred to the butyl acrylate side, resulting in an ineffective reaction process. To confirm the above assumption, the solubility of H2S in KAc aq. (4\u00a0mol\u00a0L\u22121)\u00a0has also been measured at the temperature of 30\u00a0\u00b0C and pressures up to 1.0\u00a0bar. As is shown in Fig.\u00a0S14, the absorption isotherm of H2S in KAc (4\u00a0mol\u00a0L\u22121 aq.) demonstrates an ideal linear type with increasing pressure, indicating the entire absorption process follows a physical behavior. Moreover, the solubility of H2S in KAc (4\u00a0mol\u00a0L\u22121 aq.) is only 0.75\u00a0mol\u00a0kg\u22121, significantly lower than those of our prepared ILs [26], implying that the KAc (4\u00a0mol\u00a0L\u22121 aq.) cannot activate H2S.It is well known that the selective preparation of intermediate products in cascade reactions is not easy to realize. Cascade reaction, described as two-step reaction (A \u2192 B \u2192 C) [39], one of which accelerated from A to B and the other accelerated the reaction from B to C (Fig.\u00a02\n). As a kind of value-added product, butyl 3-sulfanylpropanoate has been widely applied in pharmaceutical science and chemical engineering [40\u201344]. To control the selectivity of butyl 3-sulfanylpropanoate, the experiments of different molar ratio of H2S to butyl acrylate were carried out with a 10\u00a0mol% loading of [Emim][Ac] at 30\u00a0\u00b0C and 2.0\u00a0h. As demonstrated in Fig.\u00a03\n, the selectivity of butyl 3-sulfanylpropanoate increases up to 96% with increasing molar ratio of H2S to butyl butyrate from 0.25 to 28. When the molar ratios of H2S to butyl butyrate are 0.25 and 0.5, the corresponding conversion rates of butyl acrylate are 48% and 96%, respectively. In these points case, the selectivity of dibutyl 3,3\u2032-thiodipropionate was achieved as high as 100%. The maximum selectivity of butyl 3-sulfanylpropanoate is as high as 96% when the molar ratio of H2S to butyl acrylate is 28, indicating that the intermediate product can be effectively prepared through optimizing the molar ratio of H2S to butyl acrylate. It is believed that the selectivity of butyl 3-sulfanylpropanoate is affected by the concentration of H2S, which would reach up to 99% if we continue to improve the molar ratio of H2S to butyl acrylate.To investigate whether the generation of butyl 3-sulfanylpropanoate is related to the feeding mode, an experimental process of intermittent feeding was carried out herein. It is known that every kilogram (kg) of [Emim][Ac] can absorb 2.8 and 5.0\u00a0mol of H2S at 1.0 and 3.0\u00a0bar, respectively [21]. In the process of intermittent feeding operation, H2S is first captured by excessive ILs to enrich the concentration of SH\u2212, and then the butyl acrylate is directly added to the H2S-saturated [Emim][Ac] to participate in the reaction. The schematic diagram of the reaction device is shown in Fig.\u00a0S15 and the detailed operation steps can be found in Supporting Information. It is found that the selectivity of butyl 3-sulfanylpropanoate and dibutyl 3,3\u2032-thiodipropionate were up to 77% and 23% (Table 1, Entry 16), respectively, suggesting that the H2S concentration dissolved in the liquid played a decisive role in the control of product selectivity during the reaction.The cyclic catalytical performance of ILs catalyst is of great practical significance for the conversion of H2S. The reutilization of catalyst in this system is quite convenient to realize. For example, butyl acrylate (1.3\u00a0g, 10\u00a0mmol) together with 60\u00a0mol% [Emim][Ac] loading (1.0\u00a0g) was added for reaction in the H2S atmosphere at 30\u00a0\u00b0C for 2.0\u00a0h. After the reaction was completed, the self-phase separation mixture composed of the product and [Emim][Ac] was exhibited, which was consistent with that shown in Fig.\u00a01a. Then, the upper phase was removed and the lower phase was retained for the next reaction with the addition of fresh butyl acrylate. It should be noted that the reused [Emim][Ac] catalyst, without any activation or regeneration steps, was investigated in the next addition reaction of H2S and butyl acrylate. As demonstrated in entries 17\u221218 (Table 1), after the fifth and tenth catalyst recycling, the conversion of butyl acrylate and the selectivity of butyl 3,3\u2032-thiodipropionate show no significant decrease, suggesting that [Emim][Ac] has good durability, robustness, and recoverability after ten catalytic cycles.To better understand the process of the consecutive H2S conversion, Fig.\u00a04\n shows an example of the pressure\u2212time kinetic curve of butyl acrylate conversion in [Emim][Ac] within 6500\u00a0s [27]. During the entire reaction, three stages can be distinguished: preparation (keep vacuum), reaction, and equilibrium stage. It is impressive that after the introduction of H2S in a relatively short period of time, the pressure in the reaction tank decayed rapidly to a quite low level, indicating that H2S was quickly trapped in the liquid phase and completely reacted with butyl acrylate. The intersection point of the tangents of the next two parts was taken as a reference, and the whole reaction was almost completed within 460\u00a0s. According to the data in Table 1 (Entry 5), it can be considered that it has been completely converted in the equilibrium stage, suggesting that the reaction kinetics of the H2S conversion are at a rapid rate, showing a good industrial prospect.With the data aforementioned, a plausible mechanism for the H2S conversion by butyl acrylate mediated in [Emim][Ac] was proposed in Fig.\u00a05\na. Firstly, the carboxylate group on [Emim][Ac] interacts with free H2S, which will promote the release of nucleophilic SH\u2212 (1). In the effect of induction effect by the electron-withdrawing ester group, the secondary carbon in the CC double bond shows electropositivity (\u03b4+) [45], indicating it is easier to be attacked by the nucleophilic SH\u2212 group [46]. At the same time, the hydrogen protons on H2S activated by carboxylate of [Emim][Ac] are transferred to the tertiary carbon on the other side of the CC double bond to generate the target product 3-sulfanylpropionate because (2) and (3) can reach equilibrium at a lower barrier [47]. Once there is adequate butyl acrylate, it will continue to react with the 3-sulfanylpropionate in the presence of basic [Emim][Ac] to generate dibutyl 3,3\u2032-thiodipropionate. Similar to the previous part, the carboxylate group on [Emim][Ac] interacts with the sulfhydryl group on 3-sulfanylpropionate to release nucleophilic SR\u2212 to attack the secondary carbon on butyl acrylate, and the reaction formula was demonstrated in (4). The hydrogen proton attracted by the carboxylate can be transferred to the tertiary carbon. The ultimate dibutyl 3,3\u2032-thiodipropionate can be obtained since a chemical equilibrium of the processes (5) and (6) is realized at a lower barrier [48]. The catalyst [Emim][Ac] can be easily separated from the product mixture for the catalytic circulation because it is almost insoluble in the binary mixture system of 3,3\u2032-thiodipropionate and 3-sulfanylpropionate. As shown in Fig.\u00a05b\u2013d, the red and blue colors on the molecular surface indicate regions with more negative and positive electrostatic potentials (ESPs), respectively. It is obvious that the side of the sulfydryl group chain in [Emim][Ac]-H2S adducts and butyl 3-sulfanylpropylene exhibits increased electronegativity, which visually supports the idea that the sulfydryl group will attack the CC double bond on butyl acrylate.Immediately, the catalytic mechanism between [Emim][Ac] and butyl acrylate (1a) was discussed through DFT calculations. The bond length and transition state of the structures were further examined to further determine the reaction mechanism of the formation of thiols between H2S and butyl acrylate by using DFT-based theoretical calculations at the B3LYP/6-311G (d, p) level (more details can be seen in Supporting Information). All calculations were performed using the Gaussian 09 program, and the results are shown in Fig.\u00a06\n. First, the binding between H2S and the carboxylate anion of the catalyst is slightly exothermic (\u22126.7\u00a0kcal\u00a0mol\u22121), it is found that the bond length of H\u2013S of free H2S molecule was obviously elongated due to its affinity with carboxylate anion (Int1, 1.35 and 1.35\u00a0\u00c5 vs. 1.35 and 2.09\u00a0\u00c5). And then, Int2 is formed after the addition of butyl acrylate with a corresponding energy change of 3.1\u00a0kcal\u00a0mol\u22121. The activated H2S can easily attack the CC double bond with a low barrier of 12.8\u00a0kcal\u00a0mol\u22121 via transition state 1 (TS1). The intermediate product, 3-sulfanylpropionate, is produced by the nucleophilic attack of the SH group in the TS1 (2.25 vs. 1.85\u00a0\u00c5, Pro1b). As the reaction we investigated is a cascade reaction, the reactant butyl acrylate will continue to react with the reaction product (1b) in the first step. The Int3 is formed with the addition of 1a. At this time, two different pathways of transition state were discovered. In terms of configuration, the interaction of carboxylate anions on [Emim][Ac] to H on the sulfhydryl group of 3-sulfanylpropionate can be almost ignored during the formation of transition state 0 (TS0, bond length of H\u2212OIL\u2212anion: 2.53\u00a0\u00c5). Different from TS0, the configuration of TS2 is more moderate. The carboxylate anion has an intense interaction with H on the sulfhydryl group of 3-sulfanylpropionate (bond length of H\u2212OIL\u2212anion: 1.02\u00a0\u00c5), which can promote the release of nucleophilic SR\u2212 to attack the secondary carbon on butyl acrylate, which matches the proposed reaction mechanism in Fig.\u00a05. The energy barrier of TS2 is only 10.8\u00a0kcal\u00a0mol\u22121, which is roughly a 4/5 reduction compared to TS0, considerably decreasing the reaction energy and making this catalytic reaction more plausible, which is reasonable for the reaction to proceed readily under such mild conditions [49]. The bond length TS2 was tightened (2.14\u00a0\u00c5 vs. 1.83\u00a0\u00c5) to obtain the Pro1c, making the structure more stable. According to calculation results, the energy barriers of TS1 and TS2 are 12.8 and 10.8\u00a0kcal\u00a0mol\u22121, respectively, indicating that it is easier to approach the formation of thioether-based compounds, which is in good agreement with the experimental results under such mild conditions.The optimization procedure enabled us to efficiently examine substrate scope and limitations. Herein, we continue to study the range of substrates for such \u03b1,\u03b2-unsaturated carboxylate esters catalyzed by [Emim][Ac] (10\u00a0mol%) at 30\u00a0\u00b0C and 1.0\u00a0MPa\u00a0H2S for 2.0\u00a0h (Fig.\u00a07\n). It was found that seven substrates were investigated, and good to excellent conversion was achieved in all cases. The selectivity of products is affected by the chain length of the substrates (1a\u22124a) because of their different miscibility of corresponding products with [Emim][Ac]. Surprisingly, it is found that the structures of the products are affected by the steric hindrance of the groups connected on the CC bond (5a\u22127a), indicating that the increase of steric hindrance on the CC double bond inhibited the formation of thioether-based products. The conversion of 7a is as high as 99% at 70\u00a0\u00b0C and 6.0\u00a0h because of its high steric hindrance (>CC\u2212) [31,33], while the other six substrates achieve quantitative conversions with the Anti\u2212Markovnikov addition at 2.0\u00a0h. For the whole substrate expansion, TLC was utilized to separate the products with EA/PE as eluent. NMR and ESI\u2212MS spectra of all products are presented in ESI (Figs. S16\u2013S42).In summary, a self-separation carboxylate-based ILs catalyst was developed for the effective conversion of H2S by \u03b1,\u03b2\u2212unsaturated carboxylate esters. These ILs are immiscible with ether products, making the recovery and catalytic reuse of catalyst very convenient. It is found that the IL catalyst [Emim][Ac] can be reused for ten cycles without activity loss, revealing the highly efficient catalytic performance and excellent catalyst durability. This cascade reaction may effectively produce the sulfydryl based intermediate, butyl 3-sulfanylpropionate, whose selectivity is as high as 96% when the molar ratio of H2S to butyl acrylate is 28. The pressure\u2013time kinetic behavior shows that the conversion of butyl acrylate can be achieved almost quantitatively within 460\u00a0s with the catalyst loading of 10\u00a0mol% at 30\u00a0\u00b0C, suggesting extremely fast reaction kinetics. Besides, several \u03b1,\u03b2-unsaturated carboxylate esters were successfully converted into the corresponding products with high quantitative conversion. Furthermore, a new reaction mechanism for the H2S conversion by task-specific ILs [Emim][Ac] was proposed, in which the formation of thioether-based products between \u03b1,\u03b2\u2212unsaturated carboxylate esters and H2S takes place with a low energy barrier of 10.8\u00a0kcal\u00a0mol\u22121. It is believed that this green, efficient, and simple strategy would present great potential for industrial application.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was sponsored by the National Natural Science Foundation of China (Nos. 22208140 and 22078145).The following is the Supplementary data to this article.\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.gee.2023.03.001.", "descript": "\n                  The deep-processing utility of pure hydrogen sulfide (H2S) is a significant direction in natural gas chemical industry. Herein, a brand-new strategy of H2S conversion by \u03b1,\u03b2-unsaturated carboxylate esters into thiols or thioethers using task-specific carboxylate ionic liquids (ILs) as catalyst has been developed, firstly accomplishing the phase separation of product and catalyst without introducing the third component. It can be considered as a cascade reaction in which the product selectivity can be controlled by adjusting the molar ratio of H2S to \u03b1,\u03b2-unsaturated carboxylate esters. Also, the effects of ILs with different anions and cations, intermittent feeding operations, as well as pressure\u2212time kinetic behaviors on cascade reaction were investigated. Furthermore, the proposed interaction mechanism of H2S conversion using butyl acrylate catalyzed by [Emim][Ac] was revealed by DFT-based theoretical calculation. The approach enables the self\u2212phase separation promotion of catalyst and product and achieves 99% quantitative conversion under mild conditions in the absence of solvent, making the entire process ecologically benign. High-efficiency reaction activity can still be maintained after ten cycles of the catalyst. Therefore, the good results, combined with its simplicity of operation and the high recyclability of the catalyst, make this green method environmentally friendly and cost\u2212effective. It is anticipated that this self-separation method mediated by task-specific ILs will provide a feasible strategy for H2S utilization, which will guide its application on an industrial scale.\n               "}
{"full_text": "Data will be made available on request.Lower olefins (C2\u2013C4) are crucial chemical building blocks in the chemical industry as they are used to produce the necessary basic chemicals, such as propylene oxide, as well as a wide variety of polymers [1,2]. Considering the increasing demand for polyethylene and polypropylene and the transition towards fossil-free chemical building blocks, new ways to produce light olefins are needed [1]. The Methanol-to-Hydrocarbons (MTH) process offers an alternative route to produce lower olefins, such as ethylene and propylene, in which methanol and/or dimethyl ether is converted to olefins and gasoline-range products [3\u20136].The MTH reaction is an acid-catalyzed reaction in which various zeolites and zeotypes have been studied as catalyst materials during the past decades [3]. ZSM-5 and SAPO-34 zeolites have proven to be the most promising solid catalysts. Even though both of these catalysts are great for the MTH process, they have striking differences in their catalytic performance [3,5,7\u20139]. Zeolite framework structures with small-sized pores, e.g., SAPO-34 (with CHA framework structure), exhibit high selectivity towards light olefins, while they deactivate relatively fast. In contrast, zeolite framework structures with medium-sized pores, such as ZSM-5 (with MFI framework structure), are more stable and less prone to deactivation, while also having a much larger product distribution (i.e., the formation of light olefins and aromatics) [3,10,11].There are many strategies to enhance the selectivity towards ethylene and propylene and both acidity and porosity play a crucial role. Firstly, a facile way to increase olefin yields is to regulate the reaction conditions, such as low pressures and high reaction temperatures. Altering the strength and the density of acid sites by modifying the Si/Al ratio has been another approach to increase the production of lower olefins [12,13]. Post-synthetic modifications, such as dealumination and/or desilication, can influence selectivity and the lifetime of the catalyst [7,9,14\u201323]. Metal doping is proven to help enhance the efficiency of the MTH process. Alkaline earth metal (e.g., Mg, Ca, Sr, and Ba), rare earth metals (e.g., La), transition metals (e.g., Co, Ni, Fe, and Zr), and other chemical elements, such B and P, are, among others, used to achieve higher selectivity to light olefins [20,24\u201337].Chen et al. studied the effect of Mg modification on ZSM-5 zeolites for the methanol conversion. Mg addition reduced the number of strong acid sites, while new medium strength acid sites were created. Regarding the catalytic performance, Mg-modified ZSM-5 zeolites exhibited enhanced propylene selectivity and possessed increased lifetime [38]. Similar results were found in a study of Bakare et al., in which the effect of alkaline earth metals, such as Ca, Mg, and Ba, was investigated. Their research also revealed the formation of extra Lewis Acid Sites (LAS) [25]. Goetze et al. have investigated and compared the coke formation of H-ZSM-5 and Mg-ZSM-5 zeolites using operando UV\u2013Vis spectroscopy. It was found that Mg-ZSM-5 showed a prolonged lifetime due to the slower progression of the coke front along the catalyst bed. The latter can be attributed to decreased Br\u00f8nsted acidity, which hinders the formation of secondary coke [24]. Yarulina et al. confirmed these findings as it was revealed that propylene selectivity is firmly related to the isolation of Br\u00f8nsted Acid Sites (BAS), while the formation of LAS upon addition of Mg or Ca inhibits reactions, involving aromatic moieties and, thus, prevents coke formation [30].It is important to mention here that the majority of the studies in the literature are based on zeolite powder samples, and only a few studies shed light on shaped multi-component catalyst bodies, which are often the real catalyst material used in an industrial reactor. P\u00e9rez-Ram\u00edrez et al. investigated the impact of various binders on the performance of zeolite ZSM-5-based technical catalyst materials in the MTH reaction. Attapulgite (which is a magnesium aluminum phyllosilicate clay) has been shown to prolong the lifetime of the catalyst material and increased the selectivity towards light olefins. The authors attributed these observations to mobile Mg species migrating from the attapulgite binder to the zeolite [39]. De Jong's group has underlined the importance of nanoscale intimacy in a bifunctional zeolite-binder catalyst for the conversion of hydrocarbons. They showed that the control over Pt location on the zeolite or binder significantly influenced the catalytic performance [40\u201343].Nonetheless, the influence of Mg in zeolite-based catalyst materials on the MTH reaction has not been investigated thoroughly. To the best of our knowledge, the effect of Mg location in zeolite-based extrudates on the methanol conversion has not yet been studied. Here, by tuning the Mg location in zeolite alumina-bound shaped catalyst bodies, and, thus, the interaction between Mg and zeolite and/or binder has been investigated. To accomplish this goal, three different approaches for Mg addition have been followed; namely before, during, and after the extrusion process. Altering the step in which Mg was added (i.e., pre-, during, and post-extrusion) was crucial for the physicochemical properties of the technical catalyst bodies and their performance in the MTH reaction. The pre-extrusion modification resulted in a prolonged lifetime and large increase in the yield of light olefins due to a better Mg-zeolite interaction. A clear and distinct link was established between acidity, molecular transport, and deactivation during the MTH process.The following chemicals, gases, and materials were used: methanol (Acros, HPLC grade, 99.99% pure), N2 (Linde, 99.998%), He (Linde, >99%), Ar (Linde, 99.998%), methylcellulose (Sigma Aldrich, 4000 CP), acetic acid (Sigma Aldrich, 99.5%), boehmite (CATAPAL D, SASOL), Zeolite HZSM-5 (Zeolyst, 3024E), magnesium nitrate (Mg(NO3)2 6H2O, Sigma Aldrich, 99%), and magnesium oxide (MgO, Sigma Aldrich, 99%).Unmodified or modified powder H-ZSM-5 and boehmite were mixed in a 50-50\u00a0wt% zeolite-to-binder ratio. Methylcellulose was added and further mixed to acquire a homogeneous solid mixture. Acetic acid diluted in ultra-pure water was then added to form a paste. A Mini-Screw Extruder (Caleva) equipped with a 2-mm-diameter die plate with a cylindrical shape was used to extrude the paste. The formed catalyst extrudates were dried in air overnight, followed by a calcination step in a tubular oven at 600\u00a0\u00b0C for 6\u00a0h (with a ramp of 1\u00a0\u00b0C/min) in a flow of air. In the case of Mg addition during the extrusion process, the proper amount of MgO powder, to achieve a loading of 0.5 and 1\u00a0wt%, was added at the same moment together with the unmodified powder H-ZSM-5 and boehmite.Mg modification of ZSM-5 and boehmite powder, as well as zeolite-alumina catalyst extrudates, was done by wet impregnation. In brief, the proper amount of magnesium nitrate hexahydrate (aiming for 0.5 and 1\u00a0wt% of Mg in the final samples) was dissolved in 15\u00a0ml of ultra-pure water and added with the support in a round-bottom flask. The mixture was left to mix for 10\u00a0min to ensure homogeneity. The round-bottom flask was attached to a rotary evaporator, and vacuum was applied until the water was evaporated and the support was dry. The samples were calcined at 550\u00a0\u00b0C for 4\u00a0h (with a ramp of 1\u00a0\u00b0C/min). This synthesis method was used to prepare the individual Mg-modified components, such as ZSM-5 and boehmite (further denoted as the pre-extrusion modification). Furthermore, zeolite-alumina catalyst extrudates were impregnated in the same way (further termed as the post-extrusion modification).N2 physisorption was performed using a Micromeritics TriStar 3000 instrument operating at liquid N2 temperature. Before the measurements, a drying step was applied at 300\u00a0\u00b0C under an N2 flow for 15\u00a0h. Ammonia Temperature-Programmed Desorption (NH3-TPD) was measured on a Micromeritics AutoChemII 2920 instrument. X-ray diffraction (XRD) was measured on a Bruker D2 X-ray powder diffractometer with a Co K\u03b1 X-ray tube (\u03bb\u00a0=\u00a01.7902\u00a0\u00c5) as the source. The imaging of the spent catalyst samples after the MTH reaction was done using Confocal Fluorescence Microscopy (CFM), and the procedure followed is described in detail elsewhere [44].The performance of the various catalyst samples under study for the MTH reaction was tested in an operando UV\u2013Vis Diffuse Reflectance Spectroscopy (DRS) set-up.\u00a0\u223c\u00a069\u00a0mg of catalyst was placed in a fixed-bed reactor operating at a Weight Hourly Space Velocity (WHSV) of 6\u00a0h\u22121\u00a0at 400\u00a0\u00b0C. Methanol and the products formed during the reaction were detected using an Interscience Compact Gas Chromatograph (GC). More details of the set-up can be found elsewhere [11,24].In this study, we investigated the effect of the location of Mg in zeolite-based catalyst extrudates. Three different synthesis approaches were chosen to modify the zeolite-alumina shaped catalyst bodies with Mg. These methods are summarized in Scheme 1\n and Table S1. In these methods, Mg was introduced as both Mg2+ and MgO and is likely present as a mix of both Mg2+ ions and MgO in the final catalysts. When we refer to Mg we are referring to all potential Mg species present and not suggesting that Mg is present in the metallic form.Firstly, Mg was added to the individual components, such as zeolite and/or binder, before extrusion, in the so-called pre-extrusion-addition. In this case, Mg was impregnated in the ZSM-5 zeolite (further denoted as MgZ) and/or the alumina precursor binder (further denoted as MgA). Then, the modified samples were used to make the shaped catalyst bodies. The samples that belong in this category are: a sample which consists of Mg impregnated zeolite (MgZ) and unmodified alumina binder (A) (further denoted as Ext. MgZ/A), a sample with unmodified zeolite (Z), and Mg impregnated alumina (MgA) (further denoted as Ext. Z/MgA), and, finally, a sample with Mg impregnated zeolite (MgZ) and Mg impregnated alumina (MgA) (further denoted as Ext. MgZ/MgA). Secondly, the addition of Mg was done during the extrusion process, in the so-called during-extrusion-addition. Mg was added in the form of MgO, and it was mixed with the unmodified zeolite material and binder powders during the extrusion process. The samples are denoted as Ext. Z/A/0.5\u00a0Mg and Ext. Z/A/1\u00a0Mg where 0.5 and 1 represent the wt% of Mg in the sample. Thirdly, Mg modification was performed after the extrusion process, in the so-called post-extrusion-addition. In this case, unmodified zeolite and binder powder samples were used to make technical catalyst bodies, and then Mg was impregnated in the extrudates. The samples are further denoted as x/yMg/Ext. Z/A, where x and y represent the amount of Mg in the final sample.We used Scanning Electron Microscopy with Energy-dispersive X-ray spectroscopy (SEM-EDX) to validate the synthesis method and the location of Mg on selected samples. The results are shown in Figs. S1\u20133. Regarding the pre-extrusion modification, point spectra in zeolite and/or alumina binder\u2013rich areas proved the presence of Mg only in the zeolite for the Ext. MgZ/A sample and only in the binder for the Ext. Z/MgA. SEM-EDX for the Ext. Z/A/1\u00a0Mg sample shows larger agglomerates of Mg and no presence of Mg in zeolite and/or the alumina binder. Lastly, Mg appears in the zeolite and/or the alumina binder for the 1Mg/Ext. Z/A sample.\nFig. 1\n provides an overview of the textural properties of the shaped catalyst bodies before and after the addition of Mg as a function of the three ways of preparation. The reference catalyst extrudate sample has a bimodal pore size distribution. A first small peak is observed at 1\u00a0nm which appears to be the tail of a peak, suggesting the presence of pores <1\u00a0nm. This is attributed to the presence of micropores, associated with the zeolite material. The second peak centered at \u223c8\u00a0nm is attributed to the presence of mesopores, most probably from the alumina binder and interparticle domains formed upon extrusion, such as zeolite-binder and/or binder-binder interactions. Regarding the pre-extrusion-addition, Fig. 1a and Fig. S4 indicate that when only the zeolite was impregnated prior to extrusion (Ext. MgZ/A, pink color), a lower amount of micropores of \u223c 1\u20132\u00a0nm is observed, resulting in a subsequent slight decrease in surface area (Fig. 1d). When Mg is added before extrusion with the alumina binder, the second peak is shifted to pores of smaller sizes. Regarding the Ext. Z/MgA and the Ext. MgZ/MgA materials, the boehmite, used as alumina precursor, was impregnated with Mg and then calcined before using it for extrusion. Thus, it can be expected that strong interaction between Mg-boehmite can be achieved and Mg will fill the porous structure of Al2O3 and decrease its porosity. Similar results have been found in literature when Mg\u2013Al2O3 has been evaluated as catalytic support in a hydrodesulphurization reaction [45]. The authors claim that when Mg is introduced, a mixed Mg\u2013Al oxide is formed which fills the surface of the alumina support material. An alternative explanation could be that Mg addition before extrusion prevented the formation of interparticle pore space between zeolite-binder or binder-binder materials. Moving to the addition of Mg during the extrusion process, the Ext. Z/A/0.5\u00a0Mg sample shows an extra third peak centered at \u223c10.5\u00a0nm. An increase of the Mg amount to 1\u00a0wt% shows a decrease of the extra third peak and further decrease in the pore size as the peak at \u223c8\u00a0nm shifts to smaller pore size (\u223c6\u00a0nm).However, as shown in Fig. 1d, the addition of Mg during the extrusion process showed no significant effect on the surface area. All the findings mentioned above imply that Mg mainly interacts with the alumina binder in this approach. Lastly, Mg impregnated in the extrudate samples (i.e., the post-extrusion-addition samples) shows a minor decrease in the amount of micropores (Fig. 1c) as well as in the total surface area (Fig. 1d), suggesting blockage of the zeolite pores during the impregnation process.All the extrudate samples, i.e., both the unmodified and Mg-modified ones, show a crystal phase characteristic of an MFI topology, as shown in Fig. 2\n, as assessed by X-ray Diffraction (XRD) [38]. No significant framework change on the samples can be noticed upon adding Mg. Further inspection of the XRD patterns shows a lack of additional peaks related to the presence of Mg species, which can be justified by the low content of Mg (0.5\u20131\u00a0wt%) and/or the high level of Mg dispersion. Regarding the samples in which Mg was added before extrusion (Fig. 2a and Fig. S5), the Ext. MgZ/A and Ext. MgZ/MgA samples showed a decrease in the overall XRD intensity. The impregnation of metals in zeolites ZSM-5 is known to decrease the overall crystallinity due to the formation of defects and dealumination [46].However, when only the binder is impregnated with Mg (Ext. Z/MgA, orange color), no decrease in XRD intensity is observed compared to the reference sample. The observations mentioned above are implicit of the interaction between Mg and zeolite materials. On the other hand, when Mg is added during the extrusion process (Fig. 2b and Fig. S5), no major differences can be found compared to the reference sample, which implies no or poor interaction between Mg and the zeolite. Furthermore, as illustrated in Fig. 2c and Fig. S5, post-extrusion modification of extrudates with Mg also shows a decrease in the XRD peak intensities correlated to an increasing amount of the Mg loading.The Ammonia Temperature-Programmed Desorption (NH3-TPD) experiments, of which the results are summarized in Fig. 3\n, were used to measure the total amount of acid sites as well as their strength. Furthermore, the NH3-TPD curves were deconvoluted and used to quantify the amount of weak, intermediate and strong acid sites, which is illustrated in Fig. S6. The reference sample shows two major TPD peaks, centered at \u223c 220\u00a0\u00b0C and \u223c410\u00a0\u00b0C. According to literature, the first peak is attributed to the presence of weak acid sites, while the second is due to the presence of strong acid sites [47\u201350]. In previous reports, zeolite modification using Mg affected the overall acidity of the samples [24,30,38]. Firstly, regarding the pre-extrusion modification (Fig. 3a), the sample when only the zeolite is impregnated with Mg (Ext. MgZ/A, pink color) shows a significant decrease in the number of strong acid sites. This is likely due to Br\u00f8nsted acid sites being exchanged for Mg2+ cations. The addition of Mg resulted in an increase in the number of weak acid sites as well as in the formation of extra intermediate strong acid sites. Similar results were noticed for the sample where both components were impregnated with Mg before extrusion (Ext. MgZ/MgA, red color). Unexpectedly, impregnating only the binder material resulted in an increased strong acidity, but also in the formation of weak acid sites. Even though the increase of the number of weak acid sites could be expected, as Mg species acts as LAS [24,30,38], the formation of strong acid sites are clearly not. However, the latter observation could be explained by the migration of Al species from the binder material to the zeolite framework, which is also reported in the literature [51,52]. Mg modification of the samples prior to the extrusion process causes an increase in the total amount of acid sites, as shown in Fig. 3d. At the same time, the relative ratio between weak, medium, and strong acid sites vastly differs among the samples under study. To further assess the changes in acidity resulting from the addition of Mg during the extrusion process (Fig. 3b), our NH3-TPD measurements show that there is an increase in weak and medium acidity, while there is a minor decrease in the number of strong acid sites. The latter implies no significant Mg-zeolite interaction. It can only be assumed that the increase of weak and medium acidity is due to the presence of Mg species.On the other hand, the decrease in the number of strong acid sites is due to the interaction of Mg species with the acid site of the external zeolite surface or the migration of Mg species in the zeolite pores. The post-extrusion modification showed a decrease in the overall acidity (Fig. 3d), especially of the weak and the strong acid sites, and the formation of intermediate acid sites (Fig. 3c). This could arise from a combination of Mg exchanging with BAS in the zeolite, as well as blocking pores and lowering the overall accessibility to acid sites. Based on the results as mentioned earlier, the pre-extrusion modification with Mg shows the highest level of interaction, as shown from the XRD and NH3-TPD measurements performed. This can be concluded from the decreased overall relative intensity in the XRD pattern when Mg is impregnated in the zeolite. At the same time, NH3-TPD experiments showed a significant loss of the number of strong acid sites and the formation of new medium-strong acid sites. Weak interaction between Mg and the zeolite material was achieved in the post-extrusion impregnation. Zeolite-based extrudate catalyst materials are known to suffer from pore blockage phenomena, which makes diffusion of Mg precursor more difficult and causes further pore narrowing [51,53]. On the other hand, Mg-ions added to the technical catalyst bodies during the extrusion lead to a minimum interaction between Mg and the zeolite.The samples were tested on the activity and selectivity in the MTH reaction in order to investigate the effect of Mg modification in the zeolite-based alumina-bound shaped catalyst bodies on their catalytic performance. Fig. 4\n illustrates the methanol conversion and the yield of the ethylene and propylene formed during the MTH reaction operating at a Weight Hourly Space Velocity (WHSV) of 6 h\u22121 and 400\u00a0\u00b0C for 35\u00a0h Time-on-Stream (TOS). The reference sample (Ext. Z/A, black color), as Fig. 4a illustrates, exhibits a high level of conversion at \u223c94%. However, the methanol conversion decreased fast and reached a value of \u223c84% within the first 20\u00a0h TOS. The same trend is observed for the sample in which Mg is impregnated only in the alumina binder, which implies that Mg species present in the binder and the slight increase in acidity have no effect on the methanol conversion. Ext. MgZ/A (pink color) and Ext. MgZ/MgA (red color) samples can achieve a high level of conversion\u00a0\u223c\u00a092 and \u223c89%, respectively. Although this is slightly lower than the reference sample and Ext. Z/MgA sample, the decrease of strong acid sites can explain this due to Mg addition. Impregnating the zeolite with Mg before extrusion shows a prolonged lifetime as both samples are less prone to deactivation as they preserve a higher level of conversion than the reference sample. This is in line with findings in literature that show Mg addition reduces strong acidity, increasing the catalyst's stability [24,30,38].Regarding the samples in which Mg is added during extrusion (Ext. Z/A/0.5\u00a0Mg, dark green color, and Ext. Z/A/1\u00a0Mg, light green color), it can be seen that they achieve a lower level of conversion compared to the reference material. This can be explained by the decrease in surface area, as well as the minor decrease in the strong acidity, as shown in Figs. 1b and 3b, respectively. Further decrease in the conversion was observed for the samples in which Mg was impregnated after the extrusion process, with the sample containing 1\u00a0wt% exhibiting \u223c70% of methanol conversion.Nonetheless, the reference sample showed similar trends in ethylene and propylene yield. On the contrary, the Ext. Z/MgA (Mg impregnated only in the binder) sample exhibited the same yield of propylene at the same time as decreased ethylene yield underlying the major effect of the properties of the binder. As in this case, Ext. Z/MgA (orange color) sample showed pores of smaller size, while the Mg addition foresees the basicity of the binder. These changes promote the olefin cycle over the aromatic cycle. Mg modification of only the zeolite before extrusion (Ext. MgZ/A, pink color) resulted in a \u223c133% increase in propylene yield and a simultaneous decrease in ethylene yield. Impregnating both components prior to extrusion (Ext. MgZ/MgA, red color) caused an increase in propylene yield (\u223c166%). These findings also point out that Mg present in the binder holds a crucial role in the selectivity towards light olefins. Even though Mg and zeolite have little interaction when Mg is added during the extrusion process, based on the physicochemical characterization and the results mentioned above, it can be seen that Ext. Z/A/0.5\u00a0Mg and Ext. Z/A/1\u00a0Mg samples favored the propylene formation. Post-extrusion modification produced lower yields of ethylene and propylene in total due to the lower conversion levels recorded in these sets of samples. However, it can be seen that propylene formation is favored over ethylene due to the presence of intermediate strong acid sites. Last but not least, Fig. S7 illustrates the selectivities towards ethylene (black), propylene (red), C4 olefins (blue), C5 olefins (pink), and paraffins (sum of methane, ethane, propane, C4, and C5) (green) of all samples under study versus TOS for the MTH reaction. The latter was done in order to rule out any possibility to mislead due to the fact that the yield could be affected by the different conversion levels. Similar trends were noted for ethylene and propylene selectivities as those described above for the yields. Moreover, the paraffins selectivity draws great interest. We note that the reference sample (Ext. Z/A) exhibits the higher selectivity towards paraffins which can be attributed to the high content of strong acid sites. Relatively high selectivity towards paraffins can also be noted for sample Ext. MgZ/A in which Mg is added only in the zeolite before extrusion. In parallel, it can be seen that in all the other samples in which Mg is added either in the binder prior to extrusion, during, and/or after extrusion, the selectivity towards paraffins is considerably lower compared to Ext. Z/A and Ext. MgZ/A samples. A representative example is Ext. Z/MgA (in which only the binder is modified with Mg) and it can be noted that it exhibits\u00a0\u223c\u00a010% less selectivity towards paraffins compared to the reference sample (Ext. Z/A). Hydrogen transfer reactions could take place in acid sites located in the alumina binder that could lead to the formation of aromatics and alkanes. The above mentioned observation could imply that modification of the binder with Mg can inhibit these type of successive reactions leading to lower selectivity towards paraffins. This consideration underlines the importance of the presence of the binder as well as its properties in the physicochemical properties and the catalytic performance in the MTH reaction. The zeolite powder material and Mg-modified zeolite powder was also tested and compared for the MTH reaction, as shown in Fig. S8. The latter experiments confirmed the beneficial effect of Mg in the catalytic activity and the increase towards propylene.To further understand the effect of Mg, Thermogravimetric Analysis (TGA) measurements were performed on a sample modified by each approach (pre-extrusion, during extrusion, and post-extrusion) with 0.5\u00a0wt% of Mg and the reference sample after 35\u00a0h time on stream. As illustrated in Fig. 5\na, the TGA curves for every sample showed a low-temperature weight loss at below 150\u00a0\u00b0C and a high-temperature weight loss at above 150\u00a0\u00b0C, which are attributed to removal of water and coke, respectively. The reference sample, the sample in which Mg was impregnated prior to extrusion, and the one in which Mg was added during extrusion showed a similar amount of weight loss. On the contrary, the 0.5Mg/Ext. Z/A shows lower weight loss.To further evaluate the differences in the deactivation of these samples, the individual percentages of the two types of weight losses (low- and high-temperature) are presented in Fig. 5b. Focusing on the coke content, it is clear that the Mg addition reduces the formation of coke deposits. Post-extrusion modification showed less formation of coke, which can be ascribed to the lower methanol conversion levels. Comparing Ext. MgZ/A and Ext. Z/A/0.5\u00a0Mg samples, it is obvious that higher interaction reduces further the coke formation. The CO2 fragment signal is plotted against the temperature and shown in Fig. 5c to gain more insights into the type of coke species formed during the reaction. The reference sample shows different coke species that can be separated into two categories, the \u201csoft\u201d and \u201chard\u201d coke. \u201cSoft\u201d coke can consist of smaller aromatic compounds, such as alkylated benzenes and naphthalene, which need a lower temperature to burn off, while \u201chard\u201d coke can consist of larger polyaromatic compounds and even graphite-like coke which need higher temperatures to be removed. Comparing the reference sample with the one pre-extrusion modified, it is clear that Mg-zeolite interaction strongly reduces the formation of \u201chard\u201d coke. The reduction of strong acidity and the formation of moderate strong acid sites play a significant role in the coke species formed. However, the samples 0.5Mg/Ext. Z/A (dark blue) and Ext. Z/A/0.5\u00a0Mg (dark green) both show the presence of \u201csoft\u201d and \u201chard\u201d coke with \u201chard\u201d coke dominating.Confocal Fluorescence Microscopy (CFM) was used to visualize the nature of the coke species and their spatiotemporal distribution throughout the catalyst extrudate in the spent unmodified and modified samples after 15 and 75\u00a0min TOS in the MTH reaction. Various reaction products can be formed during the MTH process. These products can be separated into \u201cless conjugated\u201d species, such as alkylated benzenes and naphthalene, and \u201cmore conjugated\u201d, such as alkylated phenanthracenes (PH), pyrenes (PY), and (LPAs). Two lasers (i.e., 488 and 642\u00a0nm) were used to detect and separate the two types of reaction products. Green fluorescence can be emitted from the \u201cless conjugated\u201d species, while red fluorescence originates from the \u201cmore conjugated\u201d species.The top-view 3D CFM images of the reference sample, as shown in Fig. 6\na, show the presence of two different color areas, which translates into different types of coke deposits. After 15\u00a0min TOS, a green/yellow fluorescent near-edge region exists in the catalyst extrudates, while the core of the catalyst extrudates shows an orange fluorescence. The green/yellow color in the near-edge region indicates the presence of \u201cless conjugated\u201d hydrocarbon species, while the orange core region indicates the presence of \u201cmore conjugated\u201d hydrocarbon species. According to the literature, there is a molecular transport boundary towards the core of the catalyst extrudate [44,54]. Thus, \u201cless conjugated\u201d species produced in the core of the catalyst extrudate in their way to diffuse out can fall into secondary oligomerization reactions to form larger and more conjugated species, which are trapped in the core of the catalyst extrudate, explaining the orange core of the extrudate. The existence of less conjugated species in the near-edge region (bearing a green/yellow fluorescent color) could be explained by the cracking reaction of larger species to form smaller aromatic species. At 75\u00a0min TOS, it is evident that the two coke regions are still present, and their colors are more red due to the larger conjugated species formed.Upon adding Mg during the extrusion process, as illustrated in Fig. 6c, no major change in the nature and the distribution of coke species is observed in the 15\u00a0min spent samples. After 75\u00a0min TOS, the near-edge and the core regions of the catalyst extrudate show less red color, implying the formation of smaller aromatic species upon addition of Mg. Regarding the Ext. Z/A/1\u00a0Mg sample, it is clear that the near-edge region is thicker. N2 physisorption results and pore volume distribution showed a decrease in mesoporosity, which could explain the thickening of the near-edge region. The latter is in line with the literature as Whiting et al. reported that a decrease in porosity and accessibility resulted in the trapping of larger molecules in the core of the zeolite-containing catalyst extrudates [44]. Similar results regarding the thickness of the near-edge region were observed for the post-extrusion modified samples. After 75\u00a0min TOS, both samples, 0.5Mg/Ext. Z/A and 1Mg/Ext. Z/A, appear to have an even darker red core region compared to the reference sample. Even though Mg impregnation on the zeolite-based catalyst extrudates decreases both weak and strong acidity, we observe a higher formation of larger hydrocarbon species in the core region of these catalyst extrudates. This could be explained by the pore narrowing and decrease in porosity, as explained previously in Fig. 1c and d.Interestingly, pre-extrusion modification of the samples with Mg drastically changed the nature and molecular distribution of coke deposits throughout the catalyst extrudate, as demonstrated in Fig. 6b. Regarding the samples in which Mg was added only in the binder material, it still shows two areas, namely near-edge and core of the catalyst extrudates. However, the Ext. Z/MgA sample is characterized by a thicker yellow near-edge region and a red core region. The thickness of the yellow near-edge region could be explained by the smaller pore size after Mg modification, as shown in Fig. 1. The darker and/or more red hue of the two regions of interest could be attributed initially to the pore narrowing as entrapment of small molecules would be more profound. Furthermore, NH3-TPD analysis showed an increased acidity (for both the weak and strong acid sites) upon impregnating Mg in the alumina binder, which could increase secondary reactions of aromatic moieties and, thus, the genesis of the red fluorescence. Regarding Mg addition before extrusion in the zeolite (Ext. MgZ/A and Ext. MgZ/MgA), CFM images on the 15\u00a0min spent samples show a uniform coke formation of \u201cless conjugated\u201d hydrocarbon species. After 75\u00a0min TOS, conjugation into larger species is noticed, as shown from the red fluorescence emitted. The coke formation in these modified samples is vastly different from the reference sample, where two distinct areas of small and large aromatic species are formed. This is attributed to the different acidic properties of the samples. According to the literature, the formation of LAS induced by Ca or Mg modification prevents cyclic hydrocarbon pool intermediate species from participating in reactions involving aromatic moieties [30]. This could justify the existence of mainly \u201cless conjugated\u201d species, as the Ext. IZ/A and Ext. IZ/IA samples mainly contain weak and intermediates acid sites.About the post-extrusion modification, as shown in Fig. 6d, a similar deactivation pattern was observed compared to the reference sample and the samples in which Mg was added during extrusion. It seems also that in this case the near-edge region of the catalyst extrudates is enlarged for both 15 and 75\u00a0min TOS. Both samples show a more red color in the core region of the catalyst extrudates, implying the dominant presence of \u201clarge conjugated\u201d hydrocarbon species. The latter is in line with the results from the TGA measurements, which are shown in Fig. 5c. This observation can be explained by the pore narrowing which would explain the entrapment of aromatic compounds and their further oligomerization to larger polyaromatic hydrocarbon.Whiting et al. proved the existence of a molecular transport boundary has been proven [44]. They showed that tailoring the level of accessibility and porosity is strictly related to molecular transport. Even though the effect of the pore architecture on the molecular transport and the deactivation was established, there was no correlation on the effect of acidity. Here, our findings underlying the importance of the type of acid sites and the location of Mg in zeolite-alumina catalyst extrudates in their performance during the MTH reaction. Acidity and pore architecture are both important contributing factors in catalyst deactivation. However, acidity appears to be the most determining factor as weak and intermediate acid sites, as formed by the addition of Mg, inhibit the formation of larger polyaromatic moieties.As shown in Fig. 7\na\u2013b, the ethylene and propylene selectivities were correlated to the changes in acidity upon Mg modification (pre-, during, and post-extrusion). The NH3-TPD analysis were used to calculate the amount of weak, intermediate and strong acid sites, as previously explained. Then, the concentration of each type of acid sites were calculated based on the amount of acid sites divided by the SSA of the catalyst shaped bodies, to normalize for the changes in textural properties upon Mg modification. The above mentioned approach clearly shows a linear correlation between concentration of the strong acid sites (SAS) and ethylene selectivity as well as the concentration of the weak and intermediate acid sites (WAS) and propylene selectivity. Regarding ethylene selectivity, we show that at high concentration of strong acid sites, a slight increase in the ethylene selectivity can be observed. The latter observation can be explained by the presence of strong acid sites in close proximity and to the consecutive reactions which could enhance the aromatic cycle. Simultaneously, high concentration of weak acid sites resulted in high propylene selectivity. Incorporation of Mg induced lower strong acid sites while increased the concentration of weak and intermediate acid sites, as shown in Fig. 3. The latter could lead to inhibit aromatization and coke formation, promoting the alkene cycle, and thus, propylene selectivity. Similar approach has been followed from Yarulina et al. to describe the structure-performance descriptors and to underline the role of LAS for the MTO reaction [30]. The latter research study confirms our findings.To further understand the effect of Mg in the deactivation of zeolite-based shaped catalyst bodies, the concentration of strong acid sites was correlated to the deactivation rate calculated for each sample by the conversion over TOS. As illustrated in Fig. 7c, it is clear that there is a linear correlation between the strong acidity and the deactivation rate. This phenomenon can be attributed to the higher formation rate of methylated aromatic species, and thus, faster coke formation due to the high amount of strong acid sites. Last but not least, the relationship between the coke content and the ratio of the sum of the concentration of weak and intermediate acid sites is shown in Fig. 7d. The reference sample, Ext. Z/A (black), with a relative low ratio of weak and intermediate to strong acid sites exhibits a high coke formation. Moving to the sample in which Mg was added during extrusion, it is obvious that coke content is reduced. Furthermore, the status is completely reversed when maximum Mg-zeolite intimacy succeeded by addition of Mg before extrusion (Ext. MgZ/A, pink). In the case of Ext. MgZ/A (pink) we see a high ratio of weak and intermediate to strong acid sites resulted relatively low coke content. We conclude that the ratio of acid sites is crucial for the coke formation and catalyst deactivation.In this study, we have investigated the influence of the location of magnesium in zeolite-based shaped catalyst bodies on their physicochemical properties and catalytic performance in the Methanol-to-Hydrocarbons (MTH) reaction. Magnesium has been introduced in different steps of the extrusion process, namely before, during, and after the extrusion process. Physicochemical characterization of the different samples prepared proved that the pre-extrusion modification of the samples resulted in a decrease in strong acid sites, while new weak and intermediate acid sites were formed due to strong magnesium-zeolite interaction. At the same time, magnesium added prior to the extrusion process showed a significant increase in the propylene yields and the lifetime of the catalyst material prepared. A clear correlation between magnesium location, molecular transport, and catalyst deactivation during the MTH reaction of zeolite-based catalyst extrudates was made. We anticipate that this approach may contribute to designing a better catalyst material for the MTH process as well as applying this knowledge to other zeolite-based catalytic systems and acid-catalyzed chemical reactions.\nNikolaos Nikolopoulos: Writing \u2013 review & editing, Writing \u2013 original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Luke A. Parker: Writing \u2013 review & editing, Writing \u2013 original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Maurits W. Vuijk: Formal analysis, Data curation. Bert M. Weckhuysen: Writing \u2013 review & editing, Writing \u2013 original draft, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research has received funding from the European Union's EU Framework Program for Research and Innovation Horizon 2020 under Grant Agreement No. 721385 (MSCA-ETN SOCRATES - https://etn-socrates.eu/) and from the US Army Research Office (ARO, with reference number W911NF-18-1-0284). Joren Dorresteijn (Utrecht University, UU), Sebastian Haben (UU), and Silvia Zanoni (UU) are acknowledged for performing the N2 physisorption measurements. We would like to thank Dennie Wezendonk (UU) for the TGA measurements. The authors would like to thank Christia R. Jabbour (UU) for her contribution and help during the revision process.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.micromeso.2023.112553.", "descript": "\n                  One of the main challenges for the chemical industries is finding new ways to produce lower olefins, such as propylene and ethylene, to satisfy the increase in demand for e.g., polymers, namely polypropylene and polyethylene. The Methanol-to-Hydrocarbons (MTH) process is an alternative manufacturing process that can help to address this increasing demand for these important chemical building blocks. It has been proposed that the addition of magnesium to zeolites, in the form of powdered catalyst materials, enhances the selectivity towards light olefins. In this work, the impact of the location of magnesium (present as Mg2+ and MgO) in zeolite-based shaped catalyst bodies on their physicochemical properties and catalytic performance in the MTH reaction has been studied. By adjusting one of the preparation steps of the overall extrusion process in which magnesium is added tuning the location of magnesium, higher interaction between magnesium and the zeolite material could be achieved. Pre-extrusion modification showed the most favorable results in terms of physicochemical properties and catalytic activity. We found that the magnesium location could be crucial for altering molecular transport, coke formation, and catalyst deactivation during the MTH reaction due to its pronounced effects on the acidity as well as porosity of the shaped catalyst bodies. These new insights can be applied to other zeolite-based extrudate materials and other acid-catalyzed reactions as it can be crucial for the design of better and more efficient catalyst materials in their industrially shaped form.\n               "}
{"full_text": "Data will be made available on request.The design of efficient electrocatalytic materials for the development of green electrochemical energy storage and conversion devices, such as unitized regenerative fuel cells or rechargeable metal-air batteries, has become a promising way to solve the global energy demand and the environmental-related problems. However, the commercial applications are severely hampered by the high cost and poor stability of the bifunctional catalysts for oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) [1\u20134].Carbon-based materials have been studied in the last decades for their use as electrocatalysts for oxygen reduction and evolution reactions due to their good electrochemical properties such as high conductivity and catalytic activity. However, their main shortcoming is related to the weak resistance to corrosion and oxidation of carbon materials during OER, which restricts their use as bifunctional oxygen catalysts. To solve this issue, advanced carbon nanostructures, such as graphene, carbon nanofibers or carbon nanotubes, have shown promising performance and stability as bifunctional catalysts [5\u20137]. Among them, one dimensional carbon materials like carbon nanofibers (CNFs) have shown great potential because of their graphite-like structure together with more exposed active sites [8]. Therefore, CNFs contribute to increase the electrical conductivity, improve the resistance to corrosion processes and help to the diffusion of reagents and products [9\u201314].One of the most relevant properties of carbon materials in increasing activity is the presence of heteroatoms at the surface, such as N, S and P. The insertion of heteroatoms into sp2-hybridised carbon structures has been proven to be an effective way of modifying their electrical properties, chemical activity and stability [15\u201320]. Among heteroatoms, nitrogen has been extensively investigated for fuel cells and bifunctional catalysts [21,22], including those based on carbon nanofibers [23\u201326], but sulfur is gaining increasing attention due to the interesting electrochemical properties generated on sulfur-doped carbons. Higgins et al. developed a sulfur-doped graphene supported Pt-based catalyst with excellent activity for the ORR [27]. They reported that the presence of sulfur led to stronger adsorptive and cohesive binding energies with Pt nanoparticles, providing both beneficial catalytic activity and stability enhancements. Pt/S-Graphene provided 139\u00a0mA mg\u22121Pt at 0.9\u00a0V vs. RHE, better than commercial Pt/C (121\u00a0mA mg\u22121Pt) and Pt/Graphene (101\u00a0mA\u00b7mg\u22121Pt). Hoque et al. designed Pt nanowires/sulfur doped graphene as ORR catalysts in acidic electrolyte. The amount of sulfur significantly affected the ORR kinetics of the Pt nanowires. They got 182\u00a0mA\u00b7mg\u22121Pt at 0.9\u00a0V vs. RHE at a content of 1.4\u00a0wt% sulfur [28,29]. Li et al. enhanced the electrocatalytic stability of Pt supported on sulfur doped pristine carbon for the ORR in 0.1\u00a0M KOH using in-situ solution plasma with a sulfur content of 4\u00a0wt% [30]. However, bifunctional electrocatalysts based on sulfur-doped carbons have been less investigated for both oxygen evolution and reduction reactions [6,31,32]. Gao et al. synthesized manganese oxide/sulfur-doped graphitized carbon as bifunctional catalyst for ORR/OER. The current density reached \u22123\u00a0mA\u00a0cm\u22122 at 0.81\u00a0V vs. RHE (for ORR) and 10\u00a0mA\u00a0cm\u22122 at 1.62\u00a0V vs RHE (for OER) in 0.1\u00a0M KOH [32]. El-Sawy et al. incorporated heterocyclic sulfur into the carbon nanotube-graphene structure by a bidoping strategy, which not only enhanced OER activity with an overpotential of 0.35\u00a0V at a current density of 10\u00a0mA\u00a0cm\u22122, but also retained 100% of stability after 75\u00a0h. Furthermore, the sulfur-doped carbon showed high catalytic activity for the ORR [6].In previous works, we investigated a microemulsion procedure to create tantalum based catalysts [33], and carbon nanofibers as support for such tantalum-based nanoparticles, which demonstrated to be active and durable for OER [34]. Oxides from metals of groups 4 and 5 of the periodic table are known as highly durable catalysts, which combined with CNFs, represent a good opportunity towards durable and active bifunctional oxygen catalysts. In this work, we seek an enhancement of electroactivity by doping the CNF support with sulfur. We report the preparation of tantalum oxides supported on different sulfur-doped carbon nanofibers obtained by varying the temperature and duration time of the doping process. These materials were investigated as bifunctional catalysts for ORR and OER in alkaline medium. Physical and chemical properties of the supports as well as the corresponding bifunctional catalysts were characterized using several techniques. The electrocatalytic performance for ORR/OER was assessed using a rotating ring-disk electrode and the stability through time of these new materials was evaluated.The catalyst for CNF synthesis, composed by Ni\u2013Cu\u2013Al2O3 (Ni:Cu:Al molar ratio of 78:6:16), was prepared by coprecipitation of metal nitrates (Ni(NO3)2\u00b76H2O, Cu(NO3)2\u00b73H2O and Al(NO3)3\u00b79H2O, Sigma-Aldrich, > 99%), followed by calcination (air) at 450\u00a0\u00b0C for 8\u00a0h and subsequent reduction in hydrogen (Messer, > 99.5%) at 550\u00a0\u00b0C for 1\u00a0h. The catalyst composition and preparation correspond to previous investigation showing high methane conversion and stability [35]. For the growth of CNFs, 300\u00a0mg of the Ni-based catalyst was placed into a vertical fixed bed reactor under nitrogen flow (Messer, > 99.8%) and heated up to 700\u00a0\u00b0C. Then, methane (Air Products, 99.995%) was fed to the catalyst sample for 620\u00a0min at ambient pressure. The reactor was then cooled to room temperature under inert atmosphere (N2). Finally, CNFs were washed with 0.1\u00a0M HClO4 aqueous solution at 60\u00a0\u00b0C for 15\u00a0min in order to eliminate the nickel used for the growing process, followed by thorough washing with deionized water and drying at 60\u00a0\u00b0C overnight.CNF was mixed with elemental sulfur powder (Alfa Aesar), with a mass ratio of 95:5 (carbon:sulfur), in an agate mortar and deposited in a ceramic boat. The mixture was introduced in a horizontal reactor and thermally treated under inert atmosphere (N2). Two procedures were used to dope the CNF. One of them consisted of treating the sample at 250\u00a0\u00b0C for 6\u00a0h and the other one at 400\u00a0\u00b0C for 3\u00a0h, in both cases the heating rate was 5\u00a0\u00b0C min\u22121. As-prepared samples were washed with carbon disulfide (99.5%, Panreac) to eliminate the non-doping sulfur in the material, rinsed with ethanol, then with water, filtered, and finally dried in an oven at 60\u00a0\u00b0C. The sulfur-doped CNFs were labeled as CNF-SX where X stands for the CNF doping temperature (250 or 400\u00a0\u00b0C).Tantalum-based oxides (general formula TaOx) were deposited on the sulfur doped CNF by a microemulsion procedure [34]. The microemulsion (ME) consisted of mixing 0.25\u00a0mL of 75\u00a0mM NaOH aqueous solution (Alfa Aesar, 99.99%) with an oil phase composed of 2.3\u00a0g of surfactant (Igepal CO-520, Aldrich), 20\u00a0mL of n-heptane (Honeywell) and 0.75\u00a0mL of ethanol (Labkem, 99.5%). Then 0.05\u00a0mL (0.3\u00a0mmol) of tantalum (V) ethoxide (Aldrich, 99.98%) was added to the ME under continuous stirring at room temperature. According to previous works, the mixture reacts producing tantalum oxide nanoparticles within 5\u00a0min [36]. Afterwards, 312\u00a0mg of the S-doped CNF was added to the suspension and stirred overnight. The two doped CNF described in the previous section were used as supports, as well as undoped CNF for comparison purposes. In the next step, the materials were washed with ethanol and then with water, followed by drying at 60\u00a0\u00b0C overnight. The final step consisted of a heat treatment in inert atmosphere (N2) for 90\u00a0min at 900\u00a0\u00b0C. Finally, the material was washed with deionized water and dried at 60\u00a0\u00b0C overnight. The catalysts were labeled as TaOx/CNF\u2013S250 and TaOx/CNF\u2013S400.The concentration of nickel and tantalum was determined by inductively coupled plasma atomic emission spectroscopy (ICP-OES) using a Xpectroblue-EOP-TI FMT26 (Spectro). The weight percentages of sulfur and carbon were obtained by elemental analysis in a Thermo Flash 1112 equipment.Nitrogen physisorption experiments were carried out in a Quantachrome equipment and analyzed with Quadrawin software. Adsorption-desorption isotherms were obtained at \u2212196\u00a0\u00b0C. Brunauer-Emmet-Teller equation was used to calculate the BET specific surface area and the Barret-Joyner-Halenda model, applied to the desorption branch of the isotherms, was considered to determine the pore size distribution.X-ray photoelectron spectroscopy (XPS) was used in order to determine the concentration of species and the oxidation state of the doping sulfur and tantalum. The analyses were carried out in an ESCA+ (Omicron) and analyzed using CasaXPS software.The X-ray diffraction (XRD) analyses were obtained in a Brucker D8 Advance diffractometer with CuK\u03b1 radiation of 1600\u00a0W. The diffractograms were analyzed using TOPAS and EVA software and compared to the patterns of the different phases from the International Center for Diffraction Data (ICDD).Transmission electron microscopy (TEM) pictures were taken in a Tecnai F30 (FEI) microscope. For this purpose, the electrocatalysts were dispersed in ethanol and dropped on a carbon film coated Cu grid. For each catalyst, the particle size distribution of the deposited tantalum oxides was calculated using ImageJ software on TEM images and then analyzed with the statistic tools of OriginLab software.The ratio between oxygen and tantalum was determined with a Scanning Electron Microscopy-Energy Dispersive X-ray spectroscopy (SEM-EDX) SEM-EDX Hitachi S-3400\u00a0N with EDX R\u00f6ntec XFlash of Si(Li).Electrochemical analyses were carried out in a three-electrode cell at room temperature using a 0.1\u00a0M NaOH aqueous solution as electrolyte (NaOH 99.99%, Alfa Aesar), prepared with ultrapure water (Milli-Q, 18.2\u00a0M\u03a9\u00a0cm). For ORR experiments, the NaOH solution was saturated with O2 (Messer, 99.5%). The reference electrode was a reversible hydrogen electrode (RHE) and the counter electrode a glassy carbon rod. The catalysts were placed on a rotating ring-disk electrode (RRDE) composed of a glassy carbon disk (5\u00a0mm diameter) and a Pt ring. The catalytic layer was made using 1\u00a0mg\u00a0mL\u22121 ink obtained by sonicating the catalyst in isopropanol/water (50:50) and Nafion\u00ae (30\u00a0wt% of the catalytic layer), which acted as a binder. The required drops of ink were located onto the glassy carbon disk to get the needed mass loading. The catalyst loading on the disk was estimated to be 500\u00a0\u03bcg\u00a0cm\u22122. The study of the ORR was performed by linear sweep voltammetry (LSV) from 1.0 to 0.3\u00a0V vs. RHE and at a scan rate of 5\u00a0mV\u00a0s\u22121 with varying rotating speeds from 400 to 1600\u00a0rpm. The study of the OER was performed between 1.0 and 1.9\u00a0V vs. RHE and at a scan rate of 5\u00a0mV\u00a0s\u22121 with a rotating speed of 1600\u00a0rpm to ease the diffusion of evolved oxygen.The percentage of H2O2 formation during the ORR experiments was calculated according to \nEquation (1)\n, with j\n\nr\n\u00a0=\u00a0current at the ring (at constant potential of 1.2\u00a0V vs. RHE), j\n\nd\n = current at the disk and N\u00a0=\u00a00.249 (collection efficiency).\n\n(1)\n\n\n%\n\nH\n2\n\n\nO\n2\n\n=\n100\n\u00b7\n\n\n2\n\u00b7\n\nj\nr\n\n\n\nN\n\u00b7\n\nj\nd\n\n+\n\nj\nr\n\n\n\n\n\n\n\nThe kinetic current density in the ORR, j\n\nk\n, was calculated by \nEquation (2)\n considering that the measured current density, j, can be expressed as the separate contribution from j\n\nk\n and the diffusion limiting current density, j\n\nlim\n, as follows:\n\n(2)\n\n\n\n1\nj\n\n=\n\n1\n\nj\nk\n\n\n+\n\n1\n\nj\n\nl\ni\nm\n\n\n\n\n\n\n\nThe oxygen efficiency in the OER, \u03b5, this is the percentage of current actually transformed in oxygen, was calculated according to \nEquation (3)\n, where N, j\n\nr\n and j\n\nd\n stand again for the collection efficiency, current at the ring (at 0.6\u00a0V vs. RHE) and current at the disk, respectively, whereas n\n\nORR\n is the number of electrons for the reduction at the ring of the oxygen evolved from the disk (for Pt ring, n\n\nORR\n\u00a0=\u00a04).\n\n(3)\n\n\n\u03b5\n\n\n(\n%\n)\n\n=\n100\n\u00b7\n\n\n\n4\n\nn\n\nO\nR\nR\n\n\n\n\u00b7\n\nj\nr\n\n\n\nN\n\u00b7\n\nj\nd\n\n\n\n\n\n\n\nTafel slopes (b) for both ORR and OER were determined upon ohmic resistance correction (iR correction) of potential values. For this, the ohmic resistance was estimated from Newman's equation to be 8\u00a0\u03a9\u00a0cm2. \nEquation (4)\n describes the correlation of current (j\n\nk\n for ORR and j for OER) and iR-corrected potential (E\n\niR-free\n), with Tafel slope (b), exchange current density (j\n\n0\n) and reversible potential (E\n\n0\n\u00a0=\u00a01.23\u00a0V vs. RHE):\n\n(4)\n\n\n\u03b7\n=\n\n|\n\n\nE\n\ni\nR\n\u2212\nf\nr\ne\ne\n\n\n\u2212\n\nE\n0\n\n\n|\n\n=\nb\n\u00b7\nl\no\ng\n\n(\n\nj\n\nj\n0\n\n\n)\n\n\n\n\n\nEndurance tests were done to determine the variation of the behavior of the catalysts over time. For this purpose, chronopotentiometric tests were carried out, consisting of consecutive square cycles alternating either a positive or negative current density of 1\u00a0mA\u00a0cm\u22122 maintained for 180\u00a0s. Cut-off potentials of 1.9\u00a0V and 0.2\u00a0V vs. RHE were established. These tests were also done on a three-electrode cell with the same characteristics as those previously described, but now using a rotating disk electrode (RDE), with a glassy carbon tip of 5\u00a0mm and working at a rotating speed of 400\u00a0rpm.The chemical composition of the electrocatalysts was studied by elemental analysis, ICP-OES, SEM-EDX and XPS (see Tables 1 and 2\n\n). The weight percentage of sulfur for the two doped supports is similar with slightly higher concentration for the sample treated at 400\u00a0\u00b0C, as observed from elemental analysis and EDX results. The content of sulfur decreases slightly after introducing the tantalum-based catalytic particles (Table 1), most probably because of the increase of the relative content of tantalum species, which is consistent with the decrease of carbon concentration. With regard to the metallic phase, ICP analysis showed less content of Ta for S-doped CNF catalysts than for the undoped one following the same synthesis procedure (Table 2) [34]. The porous structure of supports was evaluated by nitrogen physisorption. The adsorption/desorption isotherms as well as the pore size distribution are included in the supplementary information (Fig. S1). The incorporation of sulfur species did not significantly alter the porous structure of the filaments, with similar values of both BET and external surface area (in between 55 and 70\u00a0m2\u00a0g\u22121, Table S1), a pore volume between 0.22 and 0.32\u00a0cm3\u00a0g\u22121, and a negligible content of micropores, as expected for this kind of materials [10]. The different content of Ta on supports with similar porosity indicates that the anchorage of tantalum nanoparticles is less efficient when there are sulfur species on the surface of carbon nanofibers. By comparison of the two S-doped TaOx catalysts, ICP analyses evidence a larger concentration of Ta on the support doped at 250\u00a0\u00b0C whereas XPS indicates a larger amount of Ta on the one doped at 400\u00a0\u00b0C. From this point of view, the CNF doped at 400\u00a0\u00b0C (CNF\u2013S400) clearly favors the presence of tantalum particles on the surface compared to CNF\u2013S250. The difference could be associated to the chemical speciation of sulfur, as discussed next. ICP results also evidenced the presence of nickel (1.8\u20132.6\u00a0wt%) even after the acid leaching of the samples. Interestingly, nickel was not detected in XPS analyses, most probably because it is encapsulated by carbon inside the CNF.A relevant parameter is the oxygen/tantalum atomic ratio, which is summarized in Table 2 from EDX and XPS analyses. The determination of the composition of the tantalum oxides by SEM-EDX was done by selecting areas with metal oxide nanoparticles, with the idea of minimizing the carbon and oxygen peaks from the supporting materials. The oxygen vacancies/substoichiometry has been correlated to favor the electrochemical activity for oxygen related reactions in oxides of metals from groups 4 and 5 [37]. In our case, Both EDX and XPS results do not indicate a ratio below the stoichiometric one for the most abundant phase (O/Ta\u00a0=\u00a02.5 for Ta2O5), but it must be considered that the support itself provides around 3\u00a0at.% oxygen (Table 1). Only in the case of TaOx/CNF\u2013S400 the XPS analysis indicates O/Ta below 2.5, which points to a larger substoichiometry in this catalyst compared to the other formulations.XPS analyses also show the different speciation of sulfur in the most external surface of the doped CNFs, as represented in Fig. 1\n. Regarding the supports (Fig. 1a), the S2p3/2 signal point out the presence of two main peaks at binding energies close to 164\u00a0eV and 169\u00a0eV, corresponding to two different oxidation states of sulfur: C\u2013S\u2013C and sulfoxide (S VI), respectively. The proportion of each functional group for each support is show in Table 3\n, indicating a larger content of C\u2013S\u2013C in the CNF doped at higher temperature (400\u00a0\u00b0C). This difference may partially explain the higher concentration of tantalum at the surface for TaOx/CNF\u2013S400 discussed before. Sulfur doped carbon materials are known for their capacity to adsorb metals, e.g. for decontamination purposes [38]. Since the main difference between our two S-doped supports is sulfur speciation, the enrichment of tantalum at the surface for TaOx/CNF\u2013S400 could be related to more sulfur bonded to carbon. Interestingly, after tantalum is incorporated on the support, sulfur appears mainly as C\u2013S\u2013C, with 83\u201385% as shown in Table 3 and Fig. 1b. In the catalysts, there is also a small contribution of sulfur bonded to metal appearing at lower binding energy of 162\u00a0eV (S-M), most probably nickel sulphide as identified in XRD analyses.XPS analyses revealed a different Ta concentration at the most external surface of each catalyst. The Ta 4f signal for the catalysts is showed in Fig. 2\n. The signal with a binding energy close to 27\u00a0eV, which corresponds to Ta 4f7/2, indicates the presence of Ta (V) oxidation state [39].XRD diffractograms for the S-doped and undoped catalysts are found in Fig. 3\n. Different phases were detected in the different materials: carbon, nickel and nickel sulphides (Ni3S2, NiS2, NiS) from the support (see supplementary information, Fig. S2, Table S2 and Table S3), and up to three phases containing Ta oxides (Ta2O5, TaO and NaTaO3). Regarding the support-related species, the S-doped supports without Ta oxides contain both NiS2 (major) and NiS. Interestingly, the catalysts reveal that nickel sulphides have been completely converted to the nonstoichiometric Ni3S2 upon thermal treatment.With regard to the tantalum related phases, the major contribution to XRD reflections comes from Ta2O5, with a minor contribution of TaO (in particular for TaOx/CNF\u2013S250) and even lower signal associated to NaTaO3, only for the undoped catalyst. It must be said that the presence of a mix of tantalum oxide species has been recently reported to tailor surface behavior by creating a charge transfer accumulation at their interface, caused by significant changes in the work function of the tantalum species, which results in enhanced electrocatalytic behavior [40].To delve into the crystallinity of the catalysts, XRD patterns were analyzed with Topas software (Lebail method) and the lattice parameters calculated from XRD for the CNF-supported TaOx electrocatalysts are summarized in Table 4\n, whilst the crystallite sizes are gathered in Table 5\n. It is interesting to mention that the lattice parameters for Ta2O5 are slightly lower compared to the reference pattern (JCPDS#89\u20132843) with stoichiometric formula. This is more evident in peaks (0 0 1) and (0 0 2) of Ta2O5, which appear about 0.2\u20130.3\u00b0 shifted to higher Bragg angles (2\u03b8) compared to the ICDD reference (JCPDS#89\u20132843). The contraction of the unit cell, together with chemical composition discussed before, could be associated to oxygen deficiency, as stated in previous works [33,34,41].TEM and STEM images were evaluated in order to study the morphology of S-doped CNF-supported TaOx catalysts. Some images are collected in Fig. 4\n. Both electrocatalysts are composed of carbon nanofibers with tantalum-based particles on their surface. The metal oxide nanoparticles are distinguished from the filaments by darker contrast in TEM images (left side of figure) and by lighter contrast in STEM images (right side of figure).The particle size distributions of the tantalum particles are depicted in Fig. 5\n for both catalytic materials from TEM images. The main difference is that TaOx/CNF\u2013S250 shows a broader distribution and particles with bigger size than TaOx/CNF\u2013S400. The average particle size is 24.1\u00a0\u00b1\u00a06\u00a0nm for TaOx/CNF\u2013S250, and 13.6\u00a0\u00b1\u00a03\u00a0nm for TaOx/CNF\u2013S400, in line with crystallite sizes reported in Table 5 for Ta2O5 from XRD results (23.5 and 17.8\u00a0nm, respectively).The electrochemical ORR activity in alkaline electrolyte for the different catalysts is depicted in Fig. 6\n. The disk current density (Fig. 6a) presents a sigmoidal wave form, as typically occurs for the oxygen electroreduction with oxygen saturated in the electrolyte, reaching a limiting current density at high overpotential attributed to oxygen diffusion limitation. First, by comparing the ORR activity of CNF\u2013S250 and CNF\u2013S400 with the catalysts TaOx/CNF\u2013S250 and TaOx/CNF\u2013S400, it is clear that the addition of tantalum oxides has a positive effect on the activity, with a potential shift of about 40\u00a0mV for TaOx/CNF\u2013S250 and 60\u00a0mV for TaOx/CNF\u2013S400, in terms of half-wave potential. This comparative study is useful to discard the eventual effect of nickel on ORR activity over the effect of tantalum oxides. There is a clear positive effect of tantalum oxide phases on the electroactivity regardless the presence of nickel traces.On the other hand, by comparison with the undoped support (TaOx/CNF), the introduction of sulfur has also a significantly positive effect on the ORR activity of doped catalysts, with more than 50\u00a0mV enhancement. The S-doped TaOx catalysts present a small amount of Ni3S2, as discussed from XRD characterization. A certain contribution of this phase to the activity cannot be discarded, even if nickel sulfide is at the level of traces.\nFig. 6b depicts the ring current at 1.2\u00a0V vs. RHE for the three TaOx catalysts, which is attributed to the oxidation of the hydrogen peroxide evolved at the disk. In the inset of Fig. 6b the evolution of H2O2 percentage with potential is shown, with about 50% average production. This indicates a number of electrons close to 3. There are no significant differences among the evaluated catalysts, which indicate a similar reaction mechanism in terms of the number of electrons. Most probably, the tantalum-based catalysts present a balanced mix of active sites towards both 2e\u2212 and 4e\u2212 (or 2x2e\u2212) pathways, which does not change with sulfur doping of the support.The most relevant electrochemical parameters related to ORR are collected in Table 6\n. The equivalent data for the experiments with only the supports are summarized in Table S4. The kinetic current density at 0.6\u00a0V vs. RHE is significantly higher (2.3 and 3.2\u00a0mA\u00a0cm\u22122) for the two sulfur-doped CNF-supported catalysts than for the undoped TaOx/CNF (1.4\u00a0mA\u00a0cm\u22122). In line with the previous description of results, the overpotential (\u03b7iR-free) at \u22121\u00a0mA\u00a0cm\u22122 decreases around 50\u00a0mV upon sulfur doping of the CNF support. Also the onset potential is 70\u00a0mV more positive for the doped materials. All these indicators point to the positive effect of sulfur doping in this class of catalysts. Still, despite presenting proper stability, the tantalum-based catalysts presented herein are not as active as other published catalysts in terms of activity. For example, when combining sodium tantalate (Na2Ta8O21) with tantalum oxide (Ta2O5) and tantalum nitride (Ta3N5), the ORR activity is much higher (Eonset\u00a0=\u00a00.9\u00a0V vs. RHE in 0.1\u00a0M KOH) [40]. Compared to bifunctional catalysts, titanium oxide (another group 4 metal) combined with N-doped graphene [7], or some other combinations of metal oxides (Fe, Co, Ni) with carbon nanofilaments (nanofibers, nanotubes) [23,26,42], present much better ORR activity (Eonset above 0.9\u00a0V vs. RHE in alkaline conditions), which is closer to benchmark commercial Pt/C catalyst (Eonset\u00a0=\u00a01.01\u00a0V vs. RHE [7]). In any case, the results of this work offer a new perspective of ORR improvement by means of sulfur doping of carbon supports.Tafel plots are shown in Fig. 7\n. At low overpotential, TaOx/CNF\u2013S250 showed a Tafel slope of 70\u00a0mV dec\u22121 whilst TaOx/CNF\u2013S400 and TaOx/CNF a Tafel slope of 76\u00a0mV dec\u22121. Considering an associative mechanism for ORR in alkaline medium and the theoretical simulation of Shinagawa et al. [43], a Tafel slope close to 60\u00a0mV dec\u22121 indicates that the hydrolysis of adsorbed oxygen is the rate determining step, whilst values closer to 120\u00a0mV dec\u22121 are related to the first electron transfer to adsorbed oxygen contributing to the overall reaction rate. The latter takes place at higher overpotential values, as indicated in Fig. 7. Although S-doping affects slightly to the Tafel slope, indicating that the overall reaction mechanism is not very much influenced. However, the exchange current density (j\n\n0\n) is positively influenced by S-doping (about two orders of magnitude higher), which explains the better behavior compared to undoped catalyst.The OER electrocatalytic activity of the investigated electrocatalysts (polarization curves) together with the ring current is reported in Fig. 8\n. The rotating speed for the electrode was maintained at 1600\u00a0rpm to favor the removal of O2 bubbles from the electrode surface. There is not iR-correction applied to the OER data. By comparison of supports themselves and TaOx catalysts, the incorporation of tantalum oxides clearly contributes to the enhancement of OER activity, confirming the main conclusions derived from our previous work on undoped CNF as supports [34].On the other hand, the doping of the support with sulfur does not appear to have a positive influence in OER as it does for the ORR since a slightly lower current density is observed. The ring current in Fig. 8b accounts for the oxygen evolved at the disk, which is reduced at the Pt ring. The trend is similar to that obtained of the disk current, with oxygen efficiencies slightly lower for the S-doped catalysts. The main electrochemical parameters obtained for the OER are collected in Table 7\n, while those for the supports without Ta-phases are found in Table S5. The two S-doped TaOx/CNF catalysts have less OER activity than the TaOx/CNF in terms of current at a fixed potential (j\n\n1.65\u00a0V vs RHE\n) or overpotential (\u03b7), with exception of TaOx/CNF\u2013S250 at low current density. By comparing the two doped samples, the one treated at 250\u00a0\u00b0C presents a better OER activity as reflected by its lower overpotential and higher current density, although the oxygen efficiency is higher for the catalyst doped at 400\u00a0\u00b0C. Compared to a benchmark commercial IrO2 catalyst (\u03b7\u00a0=\u00a0370\u00a0mV at 10\u00a0mA\u00a0cm\u22122 [7]), the TaOx catalysts present between only 35 and 120\u00a0mV higher overpotential.Tafel plots were used to determine the OER rate determining step (rds) for every catalyst, as presented in Fig. 9\n. Sulfur-doped catalysts exhibit a Tafel slope slightly over 100\u00a0mV dec\u22121, whilst TaOx/CNF has a lower Tafel slope of 72\u00a0mV dec\u22121. This indicates a clear change of reaction mechanism related to the doping of the support. A Tafel slope of 120\u00a0mV dec\u22121 appears when the surface species formed in the step just before the rate-determing step is predominant. In other cases, the Tafel slope is lower than 120\u00a0mV dec\u22121 for the overpotential values evaluated in Fig. 9. An intermediate value of 100\u00a0mV dec\u22121 is thus the result from a mixture of active species, with some of them favoring the pathway related to a Tafel slope of 60\u00a0mV dec\u22121 (the preferential adsorption of reaction intermediates is the rds) and some other acting with a Tafel slope of 120\u00a0mV dec\u22121 (the rds is the formation of hydroxide) [43].The different catalysts were investigated under a chronopotentiometric test in order to determine the electrode stability with time under ORR and OER conditions. Fig. 10\na and b show the variation of the potential with the number of cycles for OER and ORR, respectively. The test consisted of consecutive cycles of OER/ORR, implementing currents of +1\u00a0mA\u00a0cm\u22122 and -1\u00a0mA\u00a0cm\u22122. A duration of 3\u00a0min for sequence was programmed with cut-off potential values of 0.2\u00a0V and 1.9\u00a0V vs. RHE. In OER, the potential is quite stable with time, indicating a good stability of the set of catalysts for this reaction regardless the support used. Whereas, in ORR there is a decrease of potential in the first 10\u201315 cycles of approximately 50\u201360\u00a0mV with a much slower loss in the next ones. TaOx/CNF shows a sharp decrease of potential in ORR which is recovered after 10 cycles, indicating a reversible loss at the beginning of the experiment. Based on these results, it appears that the TaOx catalysts suffer some deactivation in ORR but good stability in OER. Upon cycling, the active sites responsible for ORR loss partial activity which does not affect OER behavior. It occurs for the three catalysts, independently of the support used. This phenomenon points to the presence of different active sites for both reactions, which is in line with other bifunctional catalysts with multiactive centers.To sum up, Fig. 10c includes the values of overpotential for both ORR and OER for the different formulations, considering both polarization curves and chronopotentiometric experiments. The sum of ORR and OER overpotentials accounts for the reversibility of catalysts defined as \u0394E\u00a0=\u00a0EOER \u2013 EORR\u00a0=\u00a0\u03b7OER\u00a0+\u00a0\u03b7ORR. The sulfur doping of the CNF leads to a significant improvement of \u0394E from 970\u00a0mV to 905\u00a0mV in the polarization curves, and from 1100\u00a0mV to 1020\u00a0mV in chronopotentiometric experiments. This is an enhancement of up to 80\u00a0mV, i.e. considering both ORR and OER, mostly coming from the better behavior for the oxygen reduction. This enhancement is maintained upon endurance tests, with particular better results for the catalyst doped at 400\u00a0\u00b0C. This particular catalyst presents a lower particle size for Ta oxides and lower O/Ta ratio than its S-doped counterpart treated at 250\u00a0\u00b0C as main differences, which supports the hypothesis of substoichiometry and low particle size conditioning the electrocatalytic behavior. The reversibility of published bifunctional catalysts in terms of \u0394E is variable and its determination depends on the electrolyte used and the criteria for calculation. For example, Luque-Centeno et al. reported \u0394E of 846\u00a0mV for a catalyst based on titanium supported on N-doped graphene, at 5\u00a0mA\u00a0cm\u22122 in 0.1\u00a0M NaOH [7]. Other transition metal-based bifuncional catalysts from the state of the art exhibit \u0394E values above 800\u00a0mV (spinels or metals combined with N-doped carbons) or above 900\u00a0mV (perovskites), with nanocomposites presenting the best bifunctional behavior with \u0394E in the range 700\u2013800\u00a0mV, similar to the best noble metal combinations [44\u201346].To sum up, the work described herein presents an alternative strategy to improve the activity and durability of oxygen bifunctional catalysts by sulfur doping of highly resistant carbon nanofibers and using a durable active phase like tantalum oxide. More work is needed to further increase the activity towards the current state of the art noble metals.In summary, we have studied new sulfur-doped CNF-supported tantalum based catalysts with bifunctional characteristics for the ORR/OER in alkaline media. The combination of high temperature and short time (400\u00a0\u00b0C for 3\u00a0h) and lower temperature for longer time (250\u00a0\u00b0C for 6\u00a0h) have been evaluated over CNF grown at 700\u00a0\u00b0C. Sulfur doping has been successful on both supports. The combination of tantalum oxide and S-doped CNF has been tested to determine the bifunctional activity of these new materials. The support doping improved the activity of the catalysts, in particular on the oxygen reduction reaction. However, the difference in activity between the doped materials is not enough remarkable to determine the effect of each doping procedure on the overall activity of the catalyst. These new materials have interesting bifunctional activity both for the ORR and OER, being slightly better for TaOx/CNF\u2013S250. TaOx/CNF\u2013S new catalysts also show good stability through time, especially for the OER. Future studies will be focused on improving the OER activity preserving the interesting bifunctionality of TaOx/CNF\u2013S electrocatalysts.\nJuan Carlos Ruiz-Cornejo: Methodology, Investigation, Data curation, Formal analysis, Writing \u2013 original draft. David Sebasti\u00e1n: Conceptualization, Methodology, Validation, Resources, Writing \u2013 review & editing, Visualization. Juan Ignacio Pardo: Validation, Writing \u2013 review & editing. Mar\u00eda Victoria Mart\u00ednez-Huerta: Supervision, Funding acquisition, Writing \u2013 review & editing. Mar\u00eda Jes\u00fas L\u00e1zaro: Supervision, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors wish to acknowledge the grants PID2020-115848RB-C21 and PID2020-115848RB-C22 funded by MCIN/AEI/10.13039/501100011033, and to the \nGobierno de Arag\u00f3n (DGA) for the funding to \nGrupo de Conversi\u00f3n de Combustibles\n (T06_17R). J.C. Ruiz-Cornejo acknowledges also DGA for his PhD grant.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.jpowsour.2022.231988.", "descript": "\n                  Highly efficient, low-cost and stable bifunctional catalysts for oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) are highly desirable for the development of green electrochemical energy storage and conversion devices. Here, we report the synthesis of sulfur-doped carbon nanofibers (CNF) as support for tantalum oxide nanoparticles. Carbon nanofibers and tantalum oxides represent a durable choice for oxygen electrocatalysis, but an improvement of catalytic activity is required. Upon doping, sulfur is found to be mainly bonded to carbon as C\u2013S\u2013C species (more than 80%). The results show that the effective incorporation of sulfur in the support has a clear positive effect on the electroactivity of the tantalum oxide catalysts. It causes a decrease of OER/ORR overpotential of 80\u00a0mV with respect to the undoped counterpart, with special improvement in the ORR. The new catalysts have shown an interesting bifunctional behavior for the OER and ORR, as well as a good stability through time.\n               "}
{"full_text": "Biorefinery using lignocellulosic biomass is a sustainable technology that contributes to climate change mitigation by sustainably converting valuable organic resources originating from natural polymers, such as lignin, cellulose, and hemicellulose, into value-added products [1,2]. During a typical biorefinery process, covalently bonded holocellulose (cellulose and hemicellulose) and lignin in the lignocellulosic biomass are separated by chemical and biological depolymerization processes [3\u20136]. Most of the fractionation focuses mainly on providing high-quality cellulose for its subsequent conversion into biofuels and chemicals [7\u201311] rather than the valorization of lignin or hemicellulose to fuels, chemicals, and materials [12]. According to recent techno-economic research, aromatic monomers derived from lignin as well as polyols from hemicellulose can be utilized as chemical feedstocks or as additives for polymers [10,12\u201321]. For example, the production of xylan and xylose, which are the main derivatives of hemicellulose, is profitable as the dehydration of hemicellulose-derived pentoses (C5 sugars) can produce furfural and its derivatives, which are important renewable platform chemicals [22]. The conversion of lignin to its corresponding aromatic monomers and polymers is highly desirable, however, the process is limited by the irreversible condensation that occurs during the lignin depolymerization; thus, minimizing this condensation is a key factor in successful lignin valorization [6,23].Reductive catalytic fractionation (RCF) of the lignocellulosic biomass enables the extraction and conversion of the majority of lignin into soluble monomers, dimers, and oligomeric alkyl phenols, while retaining most of the holocellulose in the pulp; hence, this process is categorized as \u201clignin-first\u201d in biorefinery [6,12,13,24]. Indeed, RCF yields an uncondensed low molecular weight lignin oil technically up to the theoretical yields via lignin depolymerization and stabilization [1,25\u201327].The major three factors affecting the RCF process are: (i) feedstock; (ii) solvent; and (iii) catalyst [1]. First, the efficiency of RCF is affected by the feedstock mass composition and structural features of lignin [28]. Lignin comprises three basic structural units: p-hydroxyphenyl (H, phenolic ring without any methoxy group), guaiacyl (G, phenolic ring with a methoxy group), and syringyl (S, phenolic ring with two methoxy groups). Understanding the structure of lignin is required for adjusting its depolymerization and the possible repolymerization of prepared phenolic monomers. For example, hardwood, which is a common feedstock used for RCF [13,23,29\u201332], contains approximately 18\u201325\u00a0wt% of lignin mainly comprising G and S units. Because S units lack a free ortho-position, the hardwood lignin cannot form 5\u20135 and \u03b2-5 interunit bonds by radical coupling during delignification [1,33], resulting in an abundance of easily cleaved \u03b2-O-4 moieties [1,34].Second, polar solvents can depolymerize biomass into oligomers and small amounts of phenolic monomers and dimers [1,6,18,24,35,36], even in the absence of a catalyst [12]. Particularly, methanol is highly efficient in the delignification and formation of solid fiber pulp [12,13,23,25,37]. Therefore, the RCF process can be initiated via solvolytic extraction and further proceed to partial fragmentation mainly through ether bond (\u03b2-O-4 linkages) cleavage by a catalyst [1]. Unsaturated fragments, such as G and S units, which are easily repolymerized, are produced in the initial solvolytic extraction [1,25]. The addition of organic chemicals, particularly sugar derivatives, also significantly improved the reductive depolymerization of lignin, achieving a yield of phenolic monomers as high as 83.0 % [38].Finally, heterogeneous transition-metal catalysts (e.g., Ni, Ru, Rh, Pt, Pd, and Cu) can catalyze the depolymerization of lignin oligomers to generate stable monomers while hindering unwanted repolymerization [34,36,39\u201341]. The combination of acid (from solvent) and metal catalyst improved RCF, contributing to hydrogenation and hydrolysis from their metal and acid components, respectively [41]. From this point of view, tungsten-based catalysts can be promising in the RCF process because the catalyst can allow hydrocracking, dehydrogenation, and alcohol dehydration at the Br\u00f8nsted acid sites or oxygen vacancies on tungsten oxide [36,42,43].Despite numerous efforts to develop the economically feasible RCF process or catalyst, the efficiency of the RCF process is insufficient for industrial applications [20]. Based on these observations, the objectives of this study are (i) to adjust the major products of lignocellulose (Mongolian oak, MO) using bifunctional catalysts comprising metal and acid components for the RCF process; (ii) to establish the reaction pathway based on the roles of metal and acid components; and (iii) to optimize the reaction conditions for the efficient fractionation of lignocellulose. For these objectives, the efficient depolymerization of lignin and holocellulose to their corresponding monomers was attempted while retaining cellulose as a solid pulp. Based on their lignin depolymerization capability [23,25,34,37,44] and hydrogenation or hydrodeoxygenation activity [6,34,45,46], Ru, Pd, Ni, and Co metal catalysts were prepared as bifunctional catalysts with a WZr support for fractioning the biomass.All chemicals were used without further purification, unless mentioned otherwise. Mongolian oak (MO) provided by the Seoul National University Gwanak Arboretum (Seoul, Korea) was milled and sieved into particles (< 0.5\u00a0mm). Compositional analysis of biomass was performed in accordance with the NREL\u2019s analytical procedure [47]: 74.9 %\u00a0\u00b1\u00a00.2 % holocellulose, 26.1 %\u00a0\u00b1\u00a00.1 % lignin, and 0.42 %\u00a0\u00b1\u00a00.1 % ash were measured; these values were used to calculate the yields of products in this study. Tungstate-zirconia (WZr) powder was purchased from Luxfer MEL Technologies (Manchester, UK). WZr support was calcined at 900\u00a0\u00b0C for 6\u00a0h at a heating rate of 10\u00a0\u00b0C/min. Ruthenium(III) chloride hydrate (RuCl3\u2219xH2O), palladium(II) nitrate hydrate (Pd(NO3)2\u00b7xH2O), nickel(II) chloride hydrate (NiCl2\u00b7xH2O), cobalt(II) chloride hexahydrate (CoCl20.6H2O), pyridine (anhydrous, 99.8 %), and acetic anhydride (99 %) were purchased from Sigma-Aldrich (St. Louis, Missouri, USA). 1-Butanol (99 %) was procured from Junsei chemical (Tokyo, Japan). Sulfuric acid (95.0\u201398.0 %) was procured from J.T. Baker (Phillipsburg, New Jersey, USA). Tetrahydrofuran (> 99.9 %) was procured from Honeywell B&J (Morris Plains, New Jersey, USA). Hydrogen gas (H2, 99.999 %), nitrogen gas (N2, 99.9 %), 0.5 % O2/N2 (v/v), and 5 % H2/Ar (v/v) were procured from Shinyang Medicine (Ansung, Korea). Deionized (DI) water was prepared using an aquaMAX-Ultra 370 series water purification system (YL Instruments, Anyang, Korea).The catalysts were synthesized using the wet impregnation method. To prepare 5\u00a0wt% metal/WZr (where metal refers to Ru, Pd, Ni, and Co) catalyst, the metal precursor (i.e., ruthenium(III) chloride hydrate, palladium(II) nitrate hydrate, nickel(II) chloride hydrate, and cobalt(II) chloride hexahydrate) was dissolved in 150\u00a0mL of DI water and 15\u00a0g of calcined WZr was subsequently added to the solution. The mixture was stirred at ambient conditions for 24\u00a0h, rotary evaporated, and dried for 16\u00a0h at 105\u00a0\u00b0C. The catalyst was thermally reduced in a tube furnace at 400\u00a0\u00b0C for 2\u00a0h under 200\u00a0mL/min of 5 % H2/Ar at a heating rate of 5\u00a0\u00b0C/min. Finally, passivation was performed at room temperature with 200\u00a0mL/min of O2/N2 for 30\u00a0min. The catalyst was then ready for use.RCF was performed in a 200\u00a0mL stainless-steel batch reactor equipped with a magnetic drive stirrer. A mixture of 2.5\u00a0g of MO, 1.0\u00a0g of WZr-supported metal catalysts, and 50\u00a0mL of 65 % methanol aqueous solution (v/v) was placed in the reactor (hereafter referred to as 65 % MeOH/H2O (v/v)). After sealing the reactor, a leak test was performed with N2. Thereafter, the reactor was flushed with H2 three times and pressurized with H2 to 30\u00a0bar at room temperature. The reactants were stirred at 500\u00a0rpm and heated to 100\u2013250\u00a0\u00b0C with a heating rate of 10\u00a0\u00b0C/min. The pressure of ~200\u00a0bar was reached at a temperature of 250\u00a0\u00b0C or above. When the reaction temperature was reached, it was maintained for 2\u00a0h before the mixture was cooled rapidly to room temperature. The reaction mixture was further depressurized to atmospheric pressure, and the products (liquid and solid phases) were collected. The residual solids (mixture of remaining biomass and the catalyst) were separated from the liquid by centrifugation (1200\u00a0rpm for 2.5\u00a0min at 4\u00a0\u00b0C), dried for 16\u00a0h at 60\u00a0\u00b0C, and exposed to ambient air. Solid residue was not observed after centrifuging, confirming the complete removal of solid from the liquid product (Fig. S1). The prepared products were classified into five fractions: (A) solid, (B) dichloromethane-extracted (DCM-extracted) monomers, (C) silylated DCM-extracted dimers, (D) DCM-extracted polymers, and (E) aqueous phase sugars (\nScheme 1). The recovered spent Ru/WZr catalyst was washed with water and methanol prior to its reuse for the catalysis reaction.The liquid products were identified using GC-MS and quantified using GC-FID. 1-Butanol (1\u2009mL/L in liquid product) was used as an internal standard. The HP-5MS column (60\u2009m \u00d7 250\u2009\u00b5m \u00d7 0.25\u2009\u00b5m) was used for both GC-MS (Agilent 7890/5795\u2009C, Agilent, Santa Clara, California, USA) and GC-FID (YL6500GC, Young In Chromass, Anyang, Korea). For the GC measurements, 1\u2009\u00b5L of liquid product was injected with a split ratio of 50:1 at an inlet temperature of 300\u2009\u00b0C. The oven temperature was increased from 50\u00a0\u00b0C to 150\u00a0\u00b0C at a ramping rate of 10\u2009\u00b0C/min, maintained at 150\u2009\u00b0C for 2\u2009min, increased to 250\u2009\u00b0C at a ramping rate of 8\u2009\u00b0C/min, maintained at 250\u2009\u00b0C for 5\u2009min, increased to 300\u2009\u00b0C at a ramping rate of 10\u2009\u00b0C/min, and maintained at 300\u2009\u00b0C for 2\u2009min. In addition to the monomer compounds, the dimers (mostly phenolic compounds from lignin) were silylated and observed using GC. DCM-extracted dry oil was silylated using 1 % trimethylchlorosilane dissolved in N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA). A mixture of DCM-extracted oil, pyridine, and BSTFA solution was heated at 105\u2009\u00b0C for 2\u2009h. The product was injected to the GC-MS (Agilent 7890/5975\u2009C) or GC-FID (YL6500GC) with 4-phenoxy phenol as the internal standard. The HP-5MS column (60\u2009m \u00d7 250\u2009\u00b5m \u00d7 0.25\u2009\u00b5m) was used for both GC-MS and GC-FID measurements. For the GC measurements, 1\u2009\u00b5L of liquid product was injected with a split ratio of 50:1 at an inlet temperature of 280\u2009\u00b0C. The oven temperature was increased from 50\u00b0 to 150\u00b0C with a ramping rate of 10\u2009\u00b0C/min, maintained at 150\u2009\u00b0C for 2\u2009min, then further increased to 300\u2009\u00b0C with a ramping rate of 5\u2009\u00b0C/min, and maintained at 300\u2009\u00b0C for 18\u2009min. A solvent delay of 8\u2009min was used to prevent overloading of the detector with the solvent. The mass fragments and spectra were clarified through comparison with dimer structures reported in the literature [48]. Gel-permeation chromatography (GPC) was performed using an Agilent 1200 HPLC device (Santa Clara, California, USA). Two Shodex LF-804 columns (Showa Denko, Tokyo, Japan) were used for the separation of the polymeric compounds, and a UV detector (\u03bb = 270\u2009nm) was used to measure the molecular weight distributions of the reactants and products dissolved in the eluent flow of tetrahydrofuran (THF, 1.0\u2009mL/min). The acetylation of lignocellulose was performed to improve the dissolution of polymeric compounds in the THF eluent [38,49,50]. The dried lignocellulose reactant (0.5\u2009g) or the solid residue (<0.5\u2009g, from the depolymerization product) was mixed with acetic anhydride (5\u2009mL) and pyridine (5\u2009mL). The mixture was stirred for 24\u2009h under ambient conditions prior to mixing with ethanol (20\u2009mL) and then further stirred for 30\u2009min under ambient conditions. The solution was dried at 60\u2009\u00b0C using a rotary evaporator and further dried at 55\u2009\u00b0C in vacuum for 16\u2009h to obtain the acetylated polymer. The acetylated polymer was dissolved in THF (1\u2009g/L), filtered using a Whatman syringe filter (0.45\u2009\u00b5m), and analyzed using GPC. The observed GPC results were converted to molecular weight distributions of the polymeric compounds and used to measure the number-average molecular weight (Mn), weight-average molecular weight (Mw), and polydispersity index (PDI = Mw/Mn). The GPC system was calibrated using a simulated polystyrene standard ReadyCal set (250\u20132,000,000\u2009g/mol, Sigma-Aldrich, Milwaukee, Wisconsin, USA). The quantity of sugar monomers was measured via HPLC analysis (Agilent 1260 Infinity equipped with an RI detector, Agilent, Santa Clara, California, USA) of the aqueous phase after extraction with DCM. The yields of aromatic and holocellulose-derived monomers or dimers were determined using the following equation:(Yield of monomers or dimers, wt%) =\u2009(weight of monomers or dimers)/(weight of reactant) \u00d7\u2009100.The standard deviation of monomer yields was less than 0.45 % of the average yield for Ru/WZr, confirming the reliability of the observed reaction results (Table S1).After centrifuging and drying, the mass of solid residue was measured, and the yield was calculated from the following equation:(Yield of solid products, wt%) =\u2009[(weight of recovered solid residue after the reaction) - (weight of catalyst)]/(initial weight of holocellulose fraction in the biomass reactant) \u00d7\u2009100.The X-ray diffraction (XRD) spectra of solid products were observed using D8 ADVANCE (Bruker, Billerica, Massachusetts, USA) to determine their crystal structure. The crystallinity index (%) was measured by subtracting baseline (as an amorphous fraction) at 2\u03b8\u2009=\u200910\u201327\u00b0 [51]. The 2\u03b8 range for measuring the crystallinity index was selected to contain the major diffraction peaks of cellulose and avoid the diffraction peaks of remaining catalysts. Scanning electron microscopy (SEM) images were obtained using an Inspect F50 field emission-scanning electron microscope (Thermo Fischer Scientific, Waltham, Massachusetts, USA). For SEM, the samples were coated with platinum using a Hitachi-1045\u2009ion sputter (Hitachi, Tokyo, Japan) at 15\u2009mA for 60\u2009sN2 physisorption was performed using an ASAP2020 system (Micromeritics, Norcross, Georgia, USA) to measure the BET surface area (SBET) of catalysts. NH3 temperature-programmed desorption (NH3 TPD) and H2 temperature-programmed reduction (H2 TPR) analyses were performed using a BELCAT-B catalyst analyzer (MicrotracBEL Corp., Osaka, Japan) equipped with a thermal conductivity detector. The surface metal atoms were quantified by CO chemisorption, which was performed using a BELCAT-1 automated analyzer (MicrotracBEL Corp., Osaka, Japan). The acidity of WZr was measured by NaOH titration using an automatic potentiometric titrator (AT-700). WZr powder (0.1\u2009g) was dispersed in DI water (100\u2009mL) and stirred for 12\u2009h. The mixture was titrated with aqueous NaOH (0.01\u2009M) to determine the number of acid sites and the pKa was measured using the titration curve. Powder XRD (Dmax2500-PC, Rigaku, Tokyo, Japan) was performed with Cu K\u03b11 radiation (\u03bb\u2009=\u20091.54056\u2009\u00c5, 40\u2009kV, and 200\u2009mA). X-ray photoelectron spectroscopy (XPS) was performed using a Theta Probe AR-XPS system (Thermo Fisher Scientific) with monochromated K\u03b1 excitation (h\u03bd = 1486.6\u2009eV) operated at 15\u2009kV and 150\u2009W at the Korea Basic Science Institute (Busan, Korea).The fractionation of MO using WZr-supported metal catalysts was performed to produce liquid products along with solid residue (precipitated in the mixture), as depicted in Scheme 1. The observed monomer products (fraction B in Scheme 1) dissolved in the DCM phase are listed in \nFig. 1 and Tables S2-S9; the yields of GC-measurable monomers were 1.15\u201318.6\u2009wt%. The negligible yield of 0.15\u2009wt% was achieved without the catalyst, which indicates that the RCF reaction requires a catalyst (Table S8). The highest yield of 18.6\u2009wt%, containing lignin- and holocellulose-derived monomers, was achieved at 250\u2009\u00b0C using Ru/WZr. Ni/WZr also exhibited a high yield of 16.8\u2009wt% at 250\u2009\u00b0C. Only a slight deactivation was observed when the spent Ru/WZr was used (Table S9), confirming that it is viable to reuse the Ru/WZr catalyst. Based on the 26.1 % lignin fraction in MO, the maximum yield (61.9\u2009wt% from the lignin fraction) of lignin-derived aromatic monomers was obtained for Ru/WZr at 250\u2009\u00b0C, which is one of the highest among those reported in recent literature (Table S8) [17,23,34,38,52\u201357]. A recent life-cycle analysis also confirmed a 16.2\u2009wt% aromatic monomer yield (61.9\u2009wt% lignin-based aromatic monomer yield) at a lower reaction pressure and shorter reaction time is considered a feasible strategy for reducing capital expenses [20]. The major components in fraction B were propyl- or allyl-branched syringols and guaiacols derived from lignin. Detailed product distributions were altered by the process conditions, including reaction temperature and type of catalyst.For the production of monomers (fraction B) dissolved in the liquid phase, the major lignin-derived monomers included G and S units. The formation of H units was negligible because of the native composition of hardwood MO [58]. Dimeric phenolic products were also formed, which is discussed in Section 3.6. Besides lignin-derived aromatic compounds, holocellulose-derived monomers were also formed, including ethylene glycol and furfural, which indicated significant cracking (ethylene glycol from polyol-like holocellulose) and acid-catalyzed dehydration (furfural from xylose), respectively.The metal used for catalysis affected depolymerization activity (Fig. 1). A WZr support without metal deposition, which is a solid acid catalyst, exhibited poor depolymerization activity (Table S3). The yield of G unit derivatives was 0.24\u2009wt% at 200\u2013250\u2009\u00b0C and negligible at 100\u2013150\u2009\u00b0C, indicating a significantly lower production of G units compared with S units in the absence of metal catalysts. The depolymerization activity was significantly increased when metal particles were deposited on the WZr, indicating the beneficial roles of metals on the depolymerization [50]. Among the catalysts tested in this study, Ru/WZr exhibited the highest lignin monomer yield of 18.6\u2009wt% at 250\u2009\u00b0C. When using Ru/WZr, the major components of 4-propyl guaiacol and 4-propyl syringol were obtained with high selectivity (54.8\u201385.7 % of monomers) at all reaction temperatures. The selectivities for propyl-branched guaiacol and syringol also increased with an increase in reaction temperature [23]. Pd/WZr produced propanol-branched guaiacol and syringols as the major products with the highest yield of total monomers, 14.1\u2009wt% at 200\u2009\u00b0C, which increased slightly to 15.2\u2009wt% at 250\u2009\u00b0C (Table S4). Because a larger increase in the yield of total monomers was observed for Ru/WZr (8.24\u201318.6\u2009wt%), Co/WZr (3.39\u20139.34\u2009wt%), and Ni/WZr (11.5\u201318.3\u2009wt%) when the temperature was increased from 200\u00a0\u00b0C to 250\u00a0\u00b0C, the smaller increase in the yield produced with Pd/WZr indicated that cracking or degradation of the aromatic monomer may have occurred at 250\u2009\u00b0C. Ru/WZr, Co/WZr, and Ni/WZr produced propyl-branched guaiacols and syringols as the major products, which may be less reactive than the propanol-branched products produced with Pd/WZr. Allyl-branched products were obtained when Ni/WZr and Co/WZr were used, which indicated their poor hydrogenation activity. In addition, the selectivity to allyl-substituted guaiacol and syringol decreased in the order of Ni (49.7 %) >> Co (48.9 %) > Pd (~0.0 %) \u2248 Ru (~0.0 %), indicating the greater hydrogenation activity of Ru/WZr and Pd/WZr [59]. Based on these observations, Ru/WZr was a selective hydrogenation catalyst that did not crack alkyl branches during RCF, but instead formed saturated alkyl branches.Focusing on the holocellulose-derived compounds, the highest yield of holocellulose-derived small molecules was observed with the acidic WZr support without metal deposition (0.39\u2009wt%). The acid sites of WZr catalyzed hydrolysis, hydrogenolysis, or de/rehydration of holocellulose [43,60,61] to produce sugar-degraded ethylene glycol, propylene glycol, and furfural (Tables S2-S6). Compared to a WZr support without metal deposition, the deposition of Pd, Co, and Ni onto WZr decomposed holocellulose to smaller sugar derivatives at 200\u2013250\u2009\u00b0C (Tables S2-S6). The deposition of metals, however, did not increase the yields of holocellulose-derived small molecules, achieving yields of 0.10\u20130.29\u2009wt%.Among the observed DCM-dissolved monomers (fraction B) in the liquid products, the yields of lignin-derived monomers (mostly, the derivatives of S and G units) increased at higher reaction temperatures (Fig. 1 and Tables S2-S6). For G units, propyl guaiacol and dihydroconiferyl alcohol comprising saturated propyl branches formed preferentially, indicating less cracking of alkyl branches under mild reaction conditions (30\u2009bar\u2009H2 and 100\u2013250\u2009\u00b0C), although cracking to ethyl- and methyl-branched phenolic compounds has been reported to occur in the depolymerization performed under harsher reaction conditions [49].The appreciable formation of GC-detectable (distillable) holocellulose-derived monomers (fraction B) was observed at a temperature of 200\u2009\u00b0C or higher, and the highest yield was achieved at 250\u2009\u00b0C, indicating that the degradation of holocellulose, particularly cellulose, required a higher temperature than that required by lignin (Fig. 1). The degradation products identified in the GC results included acetic acid, diols, furans, alcohols, ketones, and esters (Tables S2-S6). The observed products are mainly obtained by the degradation of hemicellulose [62]. The production of dehydrated alcohols and aldehydes, including ethylene glycol, 1,2-propanediol, xylitol, and furfural, observed at 200\u2013250\u2009\u00b0C indicated the further degradation of sugars at higher reaction temperatures. Although Ru/WZr, Pd/WZr, and Co/WZr exhibited larger productions of these degradation products compared with that exhibited by a WZr support without metal deposition, no distinct increase in the production of sugars was observed at 200\u2013250\u2009\u00b0C. The increased formation of degradation products may indicate a greater depolymerization of holocellulose at higher reaction temperatures, while the produced sugars may undergo further conversion to the degradation products. Because the sugars could not be observed using GC, the liquid products (fraction E) dissolved in the aqueous phase were observed using HPLC after DCM extraction, and negligible amounts of glucose and xylose were observed (Table S9), indicating that the majority of sugars were converted to furans and other degradation products.The dimeric products, particularly lignin-derived varieties, were observed using GC analysis of silylated products [12,48]; the dimers composed of S and G units coupled via \u03b2-1, \u03b2-\u03b2, and \u03b2-5 linkages were identified (\nFig. 2 and S2-S9) as reported in the literature [48]. Among the catalysts, Ru/WZr produced greater quantities of dimers, with a yield of 3.87\u2009wt% (based on the total weight of MO feed) at 200\u2009\u00b0C, containing more \u03b2-\u03b2 bonds and fewer \u03b2-1 bonds (Fig. 2). In comparison, Pd/WZr and Co/WZr produced dimers containing more \u03b2-1 bonds and fewer \u03b2-\u03b2 bonds. The formation of \u03b2-5 dimers was observed only for Pd/WZr. The reaction using WZr without metal deposition exhibited the negligible formation of monomers and dimers along with the formation of low molecular weight oligomers. The further depolymerization of lignin to monomers and dimers was not observed. These observations indicate that the deposited metals are required to more efficiently depolymerize lignin polymers. For all catalysts, more significant formation of dimers composed of S-G or S-S\u00a0units was observed. The formation of G-G dimers was observed for Ru/WZr, but not clearly observed for the other catalysts. \u03b2-\u03b2 dimers were observed for all catalysts, exhibiting 2\u2009wt% or higher yields. The formation of uncondensed \u03b2-\u03b2 dimers with a \u03b2-\u03b2 resinol structure (\u03b2-\u03b2 and \u03b1-2 in Fig. 2) can be attributed to the reductive catalytic cleavage of ether bonds of the resinol structure [48]. For the S units, which are the major fractions of the hardwood used in this study, the production of \u03b2-\u03b2 or \u03b2-O-4 bonds is kinetically preferred [48,63]; thus, formation of syringyl monomers occurred predominantly (Fig. 1), and led to the preferred formation of \u03b2-\u03b2 dimers compared with \u03b2-5 or \u03b2-1 dimers (Fig. 2). These C-C linkages were connected through unsubstituted or -CH2OH substituted alkyl bridges [12]. These bridges were partially removed and converted to ethylene glycol at high temperatures (> 200\u2009\u00b0C) using catalysts via C\u03b2-C\u03b3 cleavage of the linked propanol side-chains during the RCF process [12]. The effects of reaction temperature on the catalysis were studied for Ru/WZr (Fig. 2(b)). The yields of uncondensed \u03b2-\u03b2 resinol (\u03b2-\u03b2, \u03b1-2 of S-G) and the \u03b2-\u03b2 bonds of G-G dimers increased with temperature. An increase in lignin monomer yield was observed with temperature (Fig. 1), indicating that the depolymerization activity of Ru/WZr increased with temperature.In addition to the small molecules prepared from holocellulose and lignin, the formation of cellulose-rich solid residue was observed in RCF. The yield of solid residue decreased from 82.1\u201388.2\u2009wt% to 8.4\u201327.7\u2009wt% with an increase in reaction temperature from 100\u00a0\u00b0C to 250\u00a0\u00b0C (\nFig. 3), indicating the facile decomposition of holocellulose at higher temperatures. This trend is consistent with the increased yields of holocellulose-derived compounds at higher reaction temperatures (Fig. 1). The pulp was not completely recovered, even at the lowest reaction temperature of 100\u2009\u00b0C, indicating the formation of lower molecular weight holocellulose oligomers by the degradation and the loss of oligomers through their dissolution in the solvent. Although the yield of monomers was affected by the catalyst used, as depicted in Figs. 1 and 2, the yield of pulp was not significantly dependent on the deposited metal (Fig. 3). These observations indicated that MO pulping could be achieved with acidic WZr, regardless of the deposited metals.The crystal structures of cellulose-rich solid residue were analyzed using powder X-ray diffraction (\nFig. 4). Both MO feed and its recovered solid residue (fraction A) exhibited similar peaks at 2\u03b8\u2009=\u200914\u201317\u00b0 and 23\u00b0, which can be assigned as the diffractions of cellulose I\u03b1 or I\u03b2 [51,64,65]. Because of the similarity in the powder diffraction patterns of cellulose I\u03b1 and I\u03b2, it was difficult to determine if the phase transitions between I\u03b1 and I\u03b2 occurred [65]. Notably, the absence of the (1\n\n\n1\n\n\u0305\n\n0) peak at 2\u03b8\u2009=\u200912\u00b0 confirmed that cellulose II, which has been reported to form through the recrystallization of dissolved cellulose [66\u201370], was not formed, confirming that the recovered pulp was prepared by delignification but not by extraction and recrystallization. Among the catalysts, the WZr support without metal deposition exhibited weak diffraction intensities of cellulose I\u03b1 or I\u03b2, indicating the poor crystallinity of pulp recovered after the reaction at 200\u2009\u00b0C (\nTable 1). Because of the lower yields of small molecule compounds (Table S3 and Fig. 1) compared to those of other metal-deposited catalysts, the formation of less crystalline solid residue on WZr indicates that the depolymerization of lignin and holocellulose was not achieved without metal-catalyzed hydrocracking or hydrogenolysis. It also indicates that cellulose was significantly decrystallized by WZr without depolymerization, exhibiting the removal of intramolecular hydrogen bonding in cellulose. The decrystallization was, however, not significant when the metal components were present, indicating that excess hydrogen adsorbed on the metal surface selectively catalyzed lignocellulose, while preserving the crystal structures of cellulose. Among the metal-deposited catalysts, Ru/WZr at the reaction temperature up to 200\u2009\u00b0C did not cause significant modification of cellulose I\u03b1 or I\u03b2, but almost complete degradation of cellulose crystals was observed at 250\u2009\u00b0C, which is consistent with the lowest yield (8.40\u2009wt%) of solid residue (Table 1).The morphology of solid residue was further investigated using SEM (\nFig. 5). For RCF using WZr and Co/WZr, the morphology of solid residue did not significantly change except for the formation of small cracks (Fig. 5(a, c, f)). In contrast, the formation of fibrous structures was observed in RCF using Ni/WZr, Pd/WZr, and Ru/WZr (Fig. 5(b, d, e)). Reaction temperature also significantly adjusted the morphology of solid residue (Fig. 5(g, h, i, j)). For RCF using Ru/WZr, at 100\u2009\u00b0C, minor cracks were observed (Fig. 5(g)), which became significant with increasing temperature, forming fibrous structures at 200\u2009\u00b0C (Fig. 5(h, i)). As reaction temperature increased to 250\u2009\u00b0C, the solid residue decomposed to form particles (Fig. 5(j)). The XRD results of the solid residue at 250\u2009\u00b0C also confirmed the decomposition of these fibrous materials from the weak diffractions of cellulose (Fig. 4(b)).Based on the observed formation of monomers, dimers, and cellulose-rich solid residue produced by RCF, a plausible reaction pathway is proposed in \nFig. 6. Although the depolymerization of holocellulose was achieved with the acid (WZr), an increase in the formation of lignin-derived aromatic monomers and dimers was observed after the addition of metals. As illustrated in the monomer and dimer analysis results, the product compositions differed because of the reaction mechanism, which, in turn, depended on the catalyst type. The major dimeric compounds prepared using Ru/WZr and Ni/WZr contained \u03b2-\u03b2 and \u03b1\u22122 (S-G) bonds, while those obtained using Pd/WZr contained \u03b2-1 and \u03b3-OH (S-G). This suggested the presence of a different reaction mechanism depending on the catalyst type (\nScheme 2). The cleavage of the \u03b2-O-4 bond produced unstable radical products, including phenolic radicals (1 and 2) and the \u03b3-OH radical (3), which could be stabilized via the repolymerization, demethoxylation, dehydration, and combination of these reactions [71]. The \u03b3-OH radicals were stabilized via the dehydration of \u03b1-OH (4) or coupling with another \u03b3-OH to form the \u03b2-\u03b2 dimer (8). The presence of the lone pair electron of phenolic radicals can lead to equilibrium between phenolic radicals and quinone methide intermediates (5 and 6). The intermediate can be further demethoxylated (7) or cross-linked with other intermediates to form \u03b2-5 linkage (9) or \u03b2-1 (10). Based on these pathways, the \u03b3-OH radicals can be stabilized by dehydration and further hydrogenated on Pd/WZr and Co/WZr to form major products of dihydroconiferyl alcohol and dihydrosinapyl alcohol. Ru/WZr and Ni/WZr catalysts produced kinetically preferred \u03b2-\u03b2 dimers [48], and the catalysts formed alkylated phenols (4-propyl guaiacol and 4-propyl syringol), indicating the further dehydration and hydrogenation of the terminal hydroxyl group. These observations indicate that Ru/WZr and Ni/WZr catalysts not only stabilized the radicals via dehydration but also catalyzed the demethoxylation or cleavage of the C-C bonds of \u03b2-1 or \u03b2-5 dimers. Hence, it can be understood that supported metals stabilize and catalyze C-C cracking to produce the aromatic monomers. WZr support provides acid sites to mainly depolymerize the lignin-holocellulose complex. However, WZr support alone cannot stabilize reactive intermediates, resulting in severe repolymerization, as indicated by the GPC results and the calculated molecular weights (Fig. S10, Tables S12 and S13). On the contrary, it was found that the addition of metals to the WZr support can successfully suppress further repolymerization via the stabilization of radicals and improved C-C cracking.The surface active sites of metals and the acidity of the WZr support were measured to understand their contributions to catalytic activity. CO chemisorption exhibited the largest metal dispersion for Ru/WZr, which was 3.9 times larger than the smallest observed for Co/WZr (\nTable 2), indicating that Ru/WZr demonstrates better catalytic reduction activity. NH3 TPD results exhibited the quantity of acid sites modified by the deposited metals (\nFig. 7(a)). While the NH3 desorption peak temperature of the WZr support without metal deposition was in the broad range of 100\u2013450\u2009\u00b0C, those of Pd/WZr and Co/WZr slightly increased and decreased, respectively, and new peaks emerged for Ni/WZr and Ru/WZr. These observations indicate that the acidity was initially from the tungsten oxide. However, the acidic nature of tungsten oxide was affected by the deposited metals (when reduced) or metal oxides (when not reduced). NaOH titration results also confirmed that the quantity of acid sites of the catalysts varied depending on the deposited metals (Table 2). The quantity of surface active sites was in the order of Ru/WZr > Ni/WZr > Pd/WZr > WZr > Co/WZr. This catalyst-dependent acidity ultimately influenced the product distributions of the aromatic compounds (Fig. 2) and, especially, the yields of the aromatic monomers. Furthermore, for Ru/WZr and Ni/WZr, whose surface acidities are higher, the observed major dimer products were the \u03b2-\u03b2 dimers, whereas the catalysts with lower acidity (Pd/WZr and Ni/WZr) exhibited \u03b2-1 or \u03b2-5 dimers. Therefore, it can be concluded that a higher surface acidity, particularly Br\u00f8nsted acidity [43], improves the facile C-C bond cleavage. In addition, W 4f XPS results confirmed the adjusted electronic structures of tungsten oxides. The binding energy peak of W 4f7/2 shifted to a lower energy when the metals were deposited (Fig. S11 and Table S14), indicating the electron transfer from W to metals. Interestingly, the W 4f7/2 binding energy of WZr-supported metal catalysts decreased considerably for the catalysts containing a higher number of surface Br\u00f8nsted acid sites, confirming that these are the major active sites for RCF, and that the deposited metal oxides manipulate the acidity of tungsten oxide to enable the facile production of lignocellulose-derived small molecules (\nFig. 8).The crystal structures of WZr, as described by the peak intensity ratio of tetragonal ZrO2 (PDF#80-0965, at 2\u03b8\u2009=\u200930.2\u00b0) to monoclinic ZrO2 (PDF#37\u20131484, at 2\u03b8\u2009=\u200928.2\u00b0) (hereafter, for convenience, (tetragonal)/(monoclinic) ratio will be referred to as T/M ratio), were adjusted by the deposition of metals (Table 2 and Fig. 7(b)). While the WZr support without metal deposition exhibited a T/M ratio of 6.51, it increased (9.20 for Co/WZr) or decreased (5.36 for Ni/WZr, 2.79 for Pd/WZr, and 6.43 for Ru/WZr) depending on the metal catalyst. Although the monoclinic ZrO2 has been reported to provide more (Lewis) acidic sites because of the higher Zr4+ density [72,73], the T/M ratio did not exhibit this expected linear correlation with the measured quantity of surface acid sites. These observations indicate that the quantity of acid sites was more dependent on the presence of a metal and the interaction between metal and WZr than the crystal structure of ZrO2.To illustrate the interaction between metal and acid sites, the aromatic monomer yields were plotted as a function of the ratio of the quantities of metal and acid sites (M/A ratio) (Table 2, Fig. 7(c) and S12). While the addition of metals to WZr significantly increased the monomer yields, higher yields of aromatic monomers were obtained for Ni, Ru, and Pd rather than Co. The larger Co sites did not significantly increase the yield of aromatic monomers, although the metals can enhance reactions involving hydrogen. These observations indicate that the synergy between metal and acid sites can achieve the optimal RCF performance. Catalysts composed of only acid sites (zero M/A ratio) can easily crack the C\u2013C bonds; however, they cannot stabilize the active radical intermediates. In contrast, catalysts containing many metal sites (higher M/A ratio) cannot effectively crack C\u2013C bonds.The activity of the metal component was further investigated using H2 TPR (Fig. 7(d)). While the metal-free WZr did not exhibit distinct reduction at temperatures lower than 400\u2009\u00b0C, Ni, Co, and Ru were reduced at 200\u2013400\u2009\u00b0C. Because of the low reduction temperature (< 100\u2009\u00b0C) of Pd [74,75], it did not exhibit the similar reduction peak at 100\u2013400\u2009\u00b0C. These observations confirm that Ru and Pd can excellently facilitate adsorbed-hydrogen-involved reactions, including hydrogenation, hydrocracking, and hydrogenolysis [76]. These results were in accordance with the product distribution of Ni/WZr and Co/WZr, which produced allyl aromatic monomers (Tables S5 and S6).The yield of lignin-derived monomers under different reaction conditions was investigated. The yield of phenolic monomers increased with increasing H2 pressure (measured at room temperature) (\nFig. 9(a)). A negligible yield of phenolic monomers was observed at 1-bar H2, confirming that the efficiency of RCF was highly dependent on the metal-adsorbed hydrogen-involving reactions, including hydrogenolysis and hydrocracking. Because the results at 30\u2009bar\u2009H2 and 50\u2009bar\u2009H2 were not significantly different, the former may be sufficient to generate the maximum yields.The weight ratio of the Ru/WZr catalyst to the MO reactant was varied from 0.2 through 0.4\u20130.8 w/w, and all monomer yields from substrates lignin and hollocellulose were measured. The monomer yield was the lowest (6.1\u2009wt%) at a catalyst/MO ratio of 0.2 w/w. Increasing the amount of catalyst to catalyst/MO =\u20090.4 w/w increased the monomer yield to 8.2\u2009wt%; however, a further increase in the catalyst/MO ratio to 0.8 w/w did not significantly improve the monomer yield (8.7\u2009wt%). Therefore, among the three ratios, the catalyst/MO ratio of 0.4 w/w was sufficient to achieve the optimum monomer yield (Fig. 9(b)).The reaction temperature was also varied to examine the efficient recovery of the holocellulose pulp. As depicted in the XRD results and SEM images (Figs. 4(b) and 5(j)), RCF at 250\u2009\u00b0C led to the decomposition of the holocellulose pulp. Based on these observations, although Ru/WZr exhibited the highest aromatic monomer yield, Ni/WZr demonstrated a more efficient recovery of the hollocellulose pulp; hence, Ni/WZr can be used to obtain less degraded hollocellulose.The RCF of MO (hardwood) was performed to obtain: (1) phenolic monomers derived from lignin fragments and holocellulose-derived monomers for platform chemicals; and (2) cellulose-rich solid products (pulp) for further commercial industrial biorefinery. High yields of low molecular weight lignin fragments (monomeric, dimeric, and short oligomeric compounds) were produced, along with a cellulose-rich solid fraction. The maximum yield of aromatic small molecules (23.6\u2009wt% of aromatic monomer and dimer yield based on the weight of total lignocellulose) was obtained when Ru/WZr catalyst was used at 250\u2009\u00b0C, which is one of the highest yields among those reported in recent literature, as well as a techno-economically viable yield. With an increase in temperature, the components began to decompose (150\u2009\u00b0C), with further decomposition occurring at an even higher temperature (250\u2009\u00b0C). Under reductive conditions (30\u2009bar of H2 measured at room temperature) in the presence of metal catalysts, the monomeric phenols were stabilized and 4-propyl-substituted compounds were mainly generated. In the case of lignin, most of the labile ether bonds (e.g., \u03b2-O-4) were cleaved into monomeric compounds, whereas the cleavage of C-C bonds (e.g., \u03b2-\u03b2, \u03b2-1, and \u03b2-5) differed depending on the acidity of catalysts. A solid holocellulose-rich pulp was also obtained, which exhibited its cellulose crystallinity up to 200\u2009\u00b0C. Reaction conditions were optimized to achieve an efficient RCF process, and the effects of H2 pressure and catalyst/biomass ratio were investigated. The roles of catalyst components on fractionation were discussed, concluding the occurrence of depolymerization on solid acids and stabilization on supported metals. The findings of this study provide further insight into the lignin and hollocellulose components during the RCF process and are beneficial in the future development of a feasible process for fractionating lignocellulose to obtain valuable phenolic compounds and pulp.\nShinyoung Oh: Investigation, Writing \u2013 original draft. Sangseo Gu: Investigation, Writing \u2013 original draft. Jae-Wook Choi: Methodology, Investigation. Dong Jin Suh: Conceptualization. Hyunjoo Lee: Methodology, Investigation. Changsoo Kim: Methodology, Investigation. Kwang Ho Kim: Methodology, Investigation. Chun-Jae Yoo: Methodology, Investigation. Jungkyu Choi: Writing \u2013 review & editing, Supervision. Jeong-Myeong Ha: Conceptualization, Writing \u2013 review & editing, Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the Technology Development Program to Solve Climate Change of the National Research Foundation (NRF) funded by the Ministry of Science and ICT of Republic of Korea (NRF-2020M1A2A2079798). This work was also supported by the Technology Innovation Program (KEIT-20015401; NTIS-1415180841) funded by the Ministry of Trade, Industry & Energy (MOTIE, Republic of Korea).Shinyoung Oh: Investigation, Writing \u2013 original draft. Sangseo Gu: Investigation, Writing \u2013 original draft. Jae-Wook Choi: Methodology, Validation. Dong Jin Suh: Methodology, Resources. Hyunjoo Lee: Validation, Methodology. Changsoo Kim: Resources, Methodology. Kwang Ho Kim: Resources, Methodology. Chun-Jae Yoo: Validation, Conceptualization. Jungkyu Choi: Writing \u2013 review & editing, Data Curation. Jeong-Myeong Ha: Writing \u2013 review & editing, Funding acquisition..Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jece.2022.108085.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  Mongolian oak (MO), a lignocellulosic biomass feedstock comprising lignin, hemicellulose, and cellulose, was fractionated via reductive catalytic fractionation (RCF) into holocellulose-rich solid residue and lignin-derived phenol-rich liquid oil. To achieve an economically feasible RCF process, tungstate-zirconia (WZr)-supported metal catalysts, exhibiting bifunctionalities of hydrogen-adsorbing metal and acidic WZr, were used for depolymerizing and valorizing lignocellulose, and their catalytic activity was found to be highly dependent on the deposited metal. Ru/WZr exhibited excellent fractioning ability, achieving a maximum yield of 23.6\u00a0wt% of monomeric and dimeric compounds from MO and exhibiting the potential to be techno-economically viable. The superior activity of Ru/WZr can be attributed to the synergistic effects of metal and acid, which were studied by investigating the product distributions of aromatic small molecules depending on the properties of WZr-supported metal catalysts. The stabilization of reactive radical intermediates depending on the surface Br\u00f8nsted acidity of acid catalysts and hydrogen-adsorbing ability of metals were also investigated. RCF reaction conditions were optimized for the maximum yield of monomeric compounds, which can be beneficial for the further development of industrial processes.\n               "}
{"full_text": "Metal Organic FrameworksCovalent Organic FrameworksOrdered Mesoporous SilicaOxygen Reduction ReactionHydrogen Evolution ReactionOxygen Evolution ReactionCatalytic Fast PyrolysisCatalytic PyrolysisSilica to Alumina RatioGlycerol to AromaticsBenzene, Toluene, XylenePolyethylenePolypropylenePolystyreneLow Density PolyethyleneHigh Density PolyethyleneCetyltrimethylammoniumCetyltrimethylammonium BromideIndexed Hierarchy FactorWaste Cooking OilPolyethylene TerephthalateTurn Over FrequencyDensity Functional TheoryFatty AcidDimethyl etherParaxyleneBicyclic Aromatic HydrocarbonsResearch Octane NumberWeight Hourly Space VelocityGlycerol Steam ReformingPalm oil fuel ash wasteCO2 reforming of methaneDry reforming of methanetungstophosphoric acidSpace-time yieldSecondary Building UnitHydrodeoxygenationAnodized Aluminium OxideDirect Methanol Fuel CellPoly (vinylpyrrolidone)Methanol Oxidation ReactionEthanol Oxidation ReactionDirect Alcohol Fuel CellsFormic Acid OxidationDirect Formic Acid Fuel CellsPoly (ethylene oxide)-b-poly (methyl methacrylate)Many efforts have been devoted to establishing alternative resources and technologies for fuel production [1\u20134]. Not only because the fossil-based resources have been continuously depleted, but also the fossil fuel is not environmentally friendly due to the uncontrolled carbon emission. It should be noted that the alternative resources must be renewable, environmentally benign, and easy to be converted into high calorific fuels. Biomass, carbon dioxide, and water are among the resources that fulfill those requirements [5\u20138]. Consequently, various chemical or electrochemical processes are needed to convert these resources into renewable fuels. Therefore, catalysts and electrocatalysts are needed to control the course of reactions kinetically, i.e., accelerate the reaction rate and shift the selectivity towards the desired products.Several heterogeneous catalysts with various functionalities such as acidic, oxides, sulfides, metallic, metal complexes, and conductive sites have been introduced to produce renewable fuels [9\u201316]. In general, heterogeneous catalysts rely on the number of catalytic sites per surface area, proportional to the catalytic performance. It could be achieved by down-sizing the catalyst particle into the nanoscale (less than 100\u00a0nm) and/or creating a foam-like particle with nanopore cages or nanopore channels as the interior part. The latter is known as nanoporous materials, which are classified into three types based on the nanopore size: microporous (pore size less than 2\u00a0nm), mesoporous (pore size from 2 to 50\u00a0nm), and macroporous (pore size more than 50\u00a0nm) [17]. Interestingly, the nanoporous materials do not only enhance the catalytic activity but also possess the capability to direct the selectivity based on the shape or size of the nanopores [18,19].The production of renewable fuels has benefited from nanoporous materials since they significantly improve the feed conversion and yield of desired products. Several nanoporous materials have shown the up-and-coming catalytic performance for generating renewable fuels (Fig.\u00a01\n). For example, zeolites and ordered mesoporous silica (OMS) have been prominent in the conversion of biomass, waste, and carbon dioxide to renewable fuels [20\u201327]. The features of the zeolite framework provide a selective entrance for a guest molecule. Furthermore, combining the spatial confinement effect with the catalytic active sites results in a high conversion to the desired products [28]. For instance, the high selectivity of isoparaffins in gasoline-range hydrocarbon products from CO2 conversion could be boosted up to \u223c70% after coupling the Fischer\u2013Tropsch (FTs) catalyst with zeolite [29,30]. This selectivity is much higher than that obtained from classical FTs catalysts, which is limited by the Anderson-Schulz-Flory (ASF) law distribution.For biomass conversion, zeolite provides the main active sites, either Br\u00f8nsted, Lewis, or Basic site for producing fuels or intermediate platform from renewable feedstock through various processes such as pyrolysis, hydrolysis, condensation, esterification, and cracking [31]. Closely related to the zeolite materials, the ordered mesoporous silica (OMS) also provides high pore volume and large surface area with an ordered mesoporous network. Furthermore, it contains a silanol-rich surface, which is favorable for functionalization to generate the catalytic active sites [32]. These properties are beneficial for catalysis, especially for renewable related technologies, e.g., the use of SBA-15 for producing biodiesel from soybean oil\u00a0results in \u223c83% conversion [33], and the production of hydrogen from steam pyrolysis-gasification of biomass using\u00a0MCM-41 impregnated with Ni [34]. Moreover, OMS was\u00a0reported as effective support for semiconductor nanostructures for hydrogen production from solar energy via water splitting [35].Metal\u2013organic frameworks (MOFs) and covalent organic frameworks (COFs) offer diverse nanopore sizes and shapes thanks to their reticular properties [36,37]. The ability to modify their structural properties without negatively impacting their structural integrity is their key distinctive features compared to zeolites [38,39]. MOFs and COFs have been successfully applied as heterogeneous catalysts in several reactions for producing renewable fuels. For instance, they have been investigated as the efficient photocatalysts for hydrogen production due to the suitable band gap for visible light and electron\u2013hole recombination [40\u201343]. As a new kind of COFs, the covalent triazine frameworks (CTFs) are currently the cutting edge photocatalyst for hydrogen production. It can produce hydrogen 20\u201350 times higher compared to the graphitic photocatalyst [42,44]. Furthermore, they also have shown decent performance in CO2 reduction and hydrodeoxygenation of biomass for generating renewable fuels. Ultimately, nanoporous metals combine the interesting characteristics of nanomaterials with the mechanical strength of dense materials. They emerge as a new class of nanoporous materials showing catalytic activity in several reactions to produce renewable fuels. Yang et\u00a0al. [45] reported the porous cobalt-based thin film as a bifunctional catalyst to generate hydrogen and oxygen from water electrolysis. Other nanoporous metals, such as Ni\u2013Fe hydroxyl phosphate (NiFe-OH-PO4) supported on Ni foams' surface have also been reported for the same reaction in alkaline solution [46].Based on the above mentioned, it indicates the interest in the field of nanoporous materials and renewable fuels has grown massively in the past decade. Furthermore, It was also indicated by the exponential increase in the number of publications (Fig.\u00a02\n). Numerous review articles have pointed out the applications of nanoporous materials for the production of\u00a0renewable fuels. Nevertheless, they focused on either specific nanoporous materials or certain catalytic reactions [6,20,24,47\u201354]. Herein, we comprehensively review various nanoporous materials limited to the zeolite, OMS, MOFs, COFs, and nanoporous metals and their applications as heterogeneous catalysts and electrocatalysts in diverse reactions for generating renewable fuels. It should be noted that other materials, such as nanoporous carbon, are also a significant catalyst that has been widely used as supports or carriers in renewable energy-related technology, e.g., as electrocatalyst for Oxygen Reduction Reaction (ORR) and support for photocatalyst [55]. However, this material is not covered in our review. The discussion is classified by the different types of materials. Each section covers a detailed discussion of the physicochemical properties of particular nanoporous material, followed by a description of their applications in several catalytic reactions. This review is concluded with the critical remarks, opportunities, and prospects of nanoporous materials for the generation of renewable fuels in the future.Zeolites are defined as microporous (<2\u00a0nm) crystalline aluminosilicate materials composed of TO4 (T\u00a0=\u00a0Si, Al) tetrahedra as the primary building units [56\u201359]. Their frameworks bear the net negative charge because Si atom with zero formal charges undergoes an isomorphous substitution with Al atom, whose formal charge is \u22121 [60\u201363]. Therefore, zeolites also possess counter cations, commonly from the alkaline or alkaline-earth metal groups, that are present to balance the negative charge of the frameworks. Zeolites can be naturally formed due to volcanic activity or synthesized via the solvothermal (mostly hydrothermal) method [56,57,59,64,65].To date, more than 250 zeolite framework types have been reported in the database of zeolite structure provided by the International Zeolite Association (IZA) [66]. Each framework type is indicated by a three-capital-letter code, which is ordinarily derived from the names of the type materials. For example, the code LTA and MFI stand for Linde Type-A and Mordenite Framework Inverted. The zeolite porosity is originated from the uniform micropore channels and cages. The accessible pore opening is arranged by the rings comprising n TO4 units. Based on the size of pore openings, zeolites can be classified into small-pore, medium-pore, and large-pore zeolites. The small-pore zeolites have an 8-membered ring micropore with a diameter of around 3 to 4\u00a0\u00c5. They include various framework types, including LTA, CHA, AEI, and AFX. The medium-pore zeolites (e.g., MFI, MWW, MEL, and FER) comprise pore opening arranged by 10-membered rings with a diameter of around 5 to 6\u00a0\u00c5, whereas the large-pore zeolites possess pore openings whose diameter of about 7\u20139\u00a0\u00c5, which are formed by 12 or more membered rings. FAU, MOR, \u2217BEA, and EMT are included. Several structures of common zeolites are depicted in Fig.\u00a03\n.The catalytic activity of zeolites originates from their acid-active sites. In general, the counter cations of zeolites can be exchanged with ammonium ions (NH4\n+), followed by calcination to release ammonia molecules (NH3), and produce proton (H+) as the Br\u00f8nsted acid sites. In addition, the calcination also induces dehydration resulting in the three-coordinated Al species that act as Lewis acid sites. Despite the acidic properties, zeolite could also have the basicity, although less well-defined. Zeolite as a base catalyst usually could be represented by alkali-exchanged low-silica zeolite [67,68]. Principally, the basic site of zeolite was generated from the negative charges of the framework oxygen atom close to the tetrahedral Al atoms. The basicity of the framework oxygen atoms increases with decreasing the cation electronegativity. In this case, the order base strength of cation-exchanged-zeolite follows Li-\u00a0<\u00a0Na-\u00a0<\u00a0K-\u00a0<\u00a0Rb-\u00a0<\u00a0Cs-exchanged zeolites [69]. In addition to the activity, zeolites also display highly selective capability due to the uniform distribution of micropores. It should be noted that the size and shape restrictions play a major role in determining the accessibility for certain molecules [70\u201372]. The size and/or shape selectivity of zeolites results from the interaction between molecules with the well-defined pore architecture [73,74].However, there are three variants of selectivity that may overlap, i.e., selectivity based on reactant, product, and transition state molecules. The selectivity of the reactant means that only raw materials of a certain size and shape can access the zeolite pores and react at the active sites. The term \u201cmolecular sieve\u201d is, thus, coined. On the other hand, product selectivity occurs when specific molecules with a size and shape can diffuse out of the pores. The third variant of selectivity is based on the formation of intermediate molecules in chemical reactions. The pore system merely allows certain intermediates, which are suitable within the pore system. This selectivity is favored when mono- and bimolecular rearrangements are probable. Nevertheless, it is very easy to differentiate the intermediate selectivity from the product selectivity since they may display similar effects. In practice, they may proceed simultaneously [70]. The schematic illustration of the size and shape selectivity of zeolites is depicted in Fig.\u00a04\n.Despite the traditional applications, e.g., adsorption [75], separation, and catalysis related to the petrochemical industry [76], the application of zeolite has been extended to attain sustainable goals as a primary, hybrid, or support for other materials [31,77,78]. In this part, the role of zeolite catalysts in the production of renewable fuels is discussed and summarized in Table 1\n. It could be classified into the following routes: catalytic pyrolysis reaction biomass and plastic wastes and CO2 conversion reactions.Catalytic fast pyrolysis (CFP) reaction plays a major role in the production of high-quality biofuel and biomass. The catalyst also play a major role in deoxygenating bio-oil and improving its fuel properties. Several chemical reactions, i.e., deoxygenation, cracking, hydrocarbon pool mechanism, aromatization, and ketonization-aldol condensation, may coincide during fast pyrolysis catalytic reaction. Several reports have shown that zeolite provided promising results in the application for CFP reactions [1,31,79]. A highly deoxygenated, hydrocarbon-rich compound and stable pyrolysis oil product could be obtained due to the excellent capability of zeolite in promoting cracking reaction during pyrolysis. Furthermore, the tunable properties of zeolite, such as the acidity, morphology, and pore structure, make us possible to engineer the desired product of the reaction [31,79].Stefanidis et\u00a0al. [80] reported that ZSM-5 catalyst presented superior catalyst performance compared to magnesium oxide (MgO) and alumina (Al2O3), nickel monoxide (NiO), zirconia (ZrO)/titania (TiO2), tetragonal zirconia, titania, and silica-alumina (SiO2/Al2O3). It displayed the highest surface area with moderate selectivity towards hydrocarbons, reducing unwanted products and yielding organic liquid products at acceptable amounts (Fig.\u00a05\na). AbuBakar and Titiloye [81] reported the application of ZSM-5 for catalytic pyrolysis of Brunei rice husk in the fixed bed reactors. They showed that ZSM-5 increased the production of aromatic hydrocarbons and light phenols. Also, ZSM-5 increased the calorific value and water content in the bio-oil. At the same time, it decreased the viscosity, density, and acid number of the bio-oil. Thangalazhy-Gopakumar et\u00a0al. [82] introduced ZSM-5 for the catalytic pyrolysis of pinewood chips under helium and hydrogen environments. The reaction was run under two methods of catalyst-bed method and the catalytic mixing method. The best condition was achieved at 1:9 biomass catalyst mixture under helium environment employing catalytic mixing method. It yielded aromatic carbon of 41.5%, which is about 51% of the theoretical yield.In addition to their superior performance for CFP, several factors affect the performance of zeolite during the reaction. Jae et\u00a0al. [83] reported the influence of zeolite pore size and shape selectivity on the conversion of glucose to aromatic products. They utilized three different categories of zeolites as grouped for their different pore size and shape, i.e., small pore ZK-5, SAPO-34, medium pore Ferrierite, ZSM-23, MCM-22, SSZ-20, ZSM-11, ZSM-5, IM-5, TNU-9, and large pore SSZ-55, beta zeolite, Y zeolite. The result exhibited that the yield of aromatic products is dependent on the function of pore size of zeolite catalysts. Small pore zeolites did not yield any aromatic products whilst producing large quantities of coke. The medium pore size zeolite had the highest aromatic product and least amount of coke. The large pore zeolite results in the highest coke yield, low aromatic yield, and low oxygenated products yields. Yu et\u00a0al. [84] also reported the influence of pore size and shape selectivity of zeolites in the catalytic fast pyrolysis of lignin. They utilized ZSM-5, mordenite, beta, and Y zeolites, which have various static pore sizes between 5.6 and 7.6\u00a0\u00c5. ZSM-5 produced the highest aromatic yield among the four zeolites, followed by beta, mordenite, and Y zeolites, respectively (Fig.\u00a05b).Moreover, Beta and Y zeolites were the most effective catalysts for the deoxygenation reaction of lignin-derived oxygenates. This result indicated that ZSM-5 is the optimal catalyst for CFP of softwood due to its ability to achieve satisfactory deoxygenation and aromatic production simultaneously. For hardwood feedstock, the beta zeolite may be a prominent candidate. In addition to the influence of pore size and shape selectivity of zeolites, the function of the framework silica-to-alumina ratio (SAR) of ZSM-5 might also define the CFP product's yield. Foster et\u00a0al. [85] reported that ZSM-5 with SAR of 30 exhibited a higher aromatic yield concentration than SAR of 23, 50, and 80 (Fig.\u00a05c). They suggested that tuning the SAR might influence the acid concentration within the zeolite framework and maximize the aromatic production of CFP reaction. Ben and Ragauskas [88] carried out softwood (SW) kraft lignin pyrolysis using various H-ZSM-5 zeolites with different SiO2/Al2O3 mole ratios from 23 to 280. The result demonstrated that H-ZSM-5 zeolite with a relatively higher SiO2/Al2O3 mole ratio was more effective at the elimination of methoxyl groups, ether bonds, aliphatic C\u2013C bonds, and dehydration of aliphatic hydroxyl groups. However, the H-ZSM-5 zeolite with a very large SiO2/Al2O3 mole ratio, such as 280, has only limited effects on the properties of upgraded pyrolysis oil. After using zeolite, the pyrolysis oils contain some polyaromatic hydrocarbons, the content of which decreased with an increasing SiO2/Al2O3 mole ratio of zeolite.Moreover, several factors, i.e., porosity and the acidic properties of zeolite introduced to the catalytic reaction, should also be considered. Li et\u00a0al. [89] investigated the effect of mesoporosity on the performance of ZSM-5 for CFP of lignocellulosic biomass. The presence of mesopores improved the diffusion property of the ZSM-5 and their catalytic activity for cracking the bulky oxygenates (e.g., syringols derived from the lignin) as schematically shown in Fig.\u00a06\n. Therefore, it produces a higher yield of aromatic hydrocarbons (26.2\u201330.2%) and less coke formation (39.9\u201341.2%). Wang et\u00a0al. [90] Reported the catalytic performances of H-ZSM-5 catalyst with various porosity and acidic property in glycerol to aromatics (GTA). Both the GTA reaction and coking process were varied with the different mesoporosity of HZSM-5. Among all catalysts, the HZSM-5 catalyst with the highest mesoporosity of 0.385 (cm3\u00a0g\u22121) exhibited the highest BTX aromatics yield, lowest coking rate, and most extended catalyst lifetime.Furthermore, the different feedstocks of biomass also influenced the bio-oil yield produced via catalytic conversion. Huang et\u00a0al. [91] reported that the catalytic conversion of several biomass feedstocks into olefins using HZSM-5 with the addition of 6\u00a0wt.% La was decreased in the order: cellulose\u00a0>\u00a0hemicellulose\u00a0>\u00a0sugarcane bagasse\u00a0>\u00a0rice husk\u00a0>\u00a0sawdust\u00a0>\u00a0lignin. Biomass comprising a larger amount of cellulose or hemicellulose produced higher olefins yield than feedstocks with higher lignin content. While the HZSM-5 zeolite was catalytically active, incorporating La at 2.9 and 6.0\u00a0wt.% increased the production of olefins from rice husk by 15.6% and 26.5%, respectively.In order to increase the CFP activity of zeolites, the modifications of zeolite via the impregnation of transition metals have been reported. The introduction of metals, i.e., Pb, Ni, Zn, Fe, Mo, Ga, and Co, into the zeolite framework has also improved product selectivity. Liang et\u00a0al. [92] utilized ZSM-5 modified Co, Ni, and Zn for the catalytic pyrolysis of rice straw. The result exhibited that further introduction of transition metal into zeolite catalyst improved bio-oil compound selectivity. The product exhibited major contents of aldehydes/ketones and phenols with a composition of more than 50% on average of the bio-oil. Vichaphund et\u00a0al. [86] reported the application of HZSM-5 promoted Co and Ni metals via liquid ion exchange for catalytic upgrading pyrolysis vapors of Jatropha (Fig.\u00a05d). The catalytic waste was investigated using biomass to catalyst ratios of 1:0, 1:1, 1:5, and 1:10 for both types of metals (Co and Ni). The result demonstrated that both biomass to catalyst ratios and type of metals determined the aromatic hydrocarbons yields nd the oxygenated and N-containing compounds. Also, the introduction of metals, especially Ni, might inhibit the formation of coke. Therefore, it increased the catalyst lifetime.Moreover, Sun et\u00a0al. [93] reported the improved performance of ZSM-5 in CFP of biomass to aromatic after introducing Fe. It was found that the Fe/ZSM-5 catalyst exhibited higher catalytic activity by increasing the yields of monocyclic aromatic hydrocarbons and hindered its further polymerization. The introduction of Cu-metals into beta zeolite was reported by Widayatno et\u00a0al. [94]. The small amount loading (5%) of Cu on beta zeolite has improved the selectivity of hydrocarbons and the coking resistance. The introduction of metal on zeolite promoted the synergetic effect between the doped metal sites and the protonic sites on the zeolite structure, which may play an important role in improving catalyst performance. However, further introduction of Cu on zeolite has resulted in the formation of Cu aggregates, which blockage the zeolite pore and decrease the surface area. It also caused an increase in coke formation and decreased activity and selectivity. Recently, the modification of CFP catalysts using metal oxides exhibited a higher yield of the organic compounds in the bio-oil and lower content of undesired polyaromatic hydrocarbons and coke. The incorporation of MgO and ZnO reported by Fermoso et\u00a0al. [87] increased the gas yield as high formation CO and CO2. The bio-oil products also demonstrated higher H/C and O/C ratios and larger heating values (Fig.\u00a05e). It might be related to partial blockage of zeolite pore and decrease of the Br\u00f8nsted acid site, and the increase of Lewis acid site, which was created after the deposition of both metal oxides.In addition to their application for CFP of biomass wastes, zeolites are also widely applied as catalysts for non-biomass wastes such as plastics and rubber wastes. L\u00f3pez et\u00a0al. [95] introduced ZSM-5 compared to red mud for catalytic pyrolysis of plastic wastes. Both catalysts have been examined in pyrolysis of a mixture of plastics which resembles municipal plastic wastes, at 440 and 500\u00a0\u00b0C in a 3.5\u00a0dm3 semi-batch reactor. The result exhibited better performance as its higher porosity, and strong acidity contributed to producing a greater proportion of gases and liquids with a higher aromatics content than the condition without catalyst. Santos et\u00a0al. [96] reported the better performance of USY compared to ZSM-5 for the CFP of polyethylene (PE) and polypropylene (PP) wastes. USY catalyst exhibited regenerable properties, as reported by Kassargy et\u00a0al. [97].Additionally, Boxiong et\u00a0al. [98] explored the catalytic performance of USY and HZSM-5 in the CFP of waste tyres, and it was concluded that USY zeolites exhibited better conversion capability than HZSM-5 in the production of aromatic hydrocarbons. Recently, Wang et\u00a0al. [99] reported using USY zeolites for CFP of rubber wastes. It is obtained that alkenes and aromatic hydrocarbons were the main products obtained from the CFP of rubber wastes. They showed that the USY zeolite with a low 5.3 was more effective for producing aromatic hydrocarbons, while the higher SiO2/Al2O3 mole ratio (11.5) led to greater alkenes formation. In this reaction, the pyrolysis temperature also played a vital role, in which the formation of the highest concentration of aromatics compound was achieved at 750\u00a0\u00b0C.Kassargy et\u00a0al. [100] reported the catalytic degradation of PP, PE, and their mixtures to produce gasoline and diesel-like fuels using USY zeolites. The catalytic pyrolysis reaction resulted in a liquid fraction dominated by a (C5\u2013C7) hydrocarbons fraction and the gaseous products, which are major constituents of C3 and C4. Although the synergistic effect of the plastic mixtures is still elusive, their proportion influenced the liquid fractions and the yields of the products. Both natural and synthetic zeolite (ZSM-5) for catalytic pyrolysis of polystyrene (PS), PP, PE, and their mixtures exhibited remarkable performance [101]. Interestingly, the catalytic reaction of PS plastic employed natural zeolites could result in the highest liquid oil yield of 54%. In contrast, the mixing of PS with other plastic wastes might decrease the liquid oil yield. Nevertheless, the mixture PP and PE feedstocks demonstrated a higher liquid oil yield than their individual using both catalysts.Despite the choice of catalyst and feedstock used for catalytic reaction, it is also important to choose both suitable catalyst bed temperature and the SAR of the catalysts to obtain the optimum condition of the reaction. Onwudili et\u00a0al. [102] reported that the increase of catalyst bed temperature led to increased gas production, particularly C2\u2013C4 hydrocarbons, during the catalytic reaction of simulated mix plastic feedstocks using several zeolite-based catalysts with different SAR. The catalyst with lower SAR exhibited better performance than the catalyst with higher SAR (16.4-80). It produced the highest aromatic compounds yield at both pyrolysis temperatures of 500 and 600\u00a0\u00b0C. The catalyst also generated higher hydrogen gas and higher benzene and toluene composition of 90% in the aromatic fractions. Similar behavior was also found in the hydroprocessing of thermal cracked-low density polyethylene (LDPE) plastics oil using ZSM-5 as a catalyst. The low SAR ZSM-5 also exhibited an extensive increase in the production of gaseous hydrocarbons. However, in this type of reaction (upgrading LDPE oil to fulfill the properties of fuel oil), a higher SAR zeolite demonstrated more suitable properties as its ability to attenuate the cracking activity while keeping the contribution of hydroisomerization and olefin hydrogenation. Therefore, the higher SAR catalyst exhibits a good combination of hydroisomerization and aromatization reactions along with a limited extension of end-chain cracking reactions so that it achieves high selectivity towards liquid fuels (over 95%) [103].Susastriawan et\u00a0al. [104] studied the application of zeolite for the low-temperature pyrolysis of LDPE plastic waste. They also studied the influence of catalyst particle size on their performance. The results demonstrated that smaller zeolite sizes led to an increase in heat transfer rate, pyrolysis temperature, reaction rate, and oil yield. It also found that the elevated pyrolysis temperature gives rise to a higher oil yield. However, the author reported that the oil yield percentage is still relatively low compared to gas yield and remaining char. From 1000\u00a0g of LDPE plastic, the obtained oil yields of 138, 134, 126\u00a0mL from 1, 2, and 3\u00a0mm in diameter of the zeolite, respectively. Furthermore, Kadja et\u00a0al. [22] examined the effect of hierarchical porosity of ZSM-5 for their catalytic activity over LDPE pyrolysis. Hierarchical ZSM-5 was developed via sequential mechanochemical treatment and recrystallization in the presence of cetyltrimethylammonium (CTA+) molecules (Fig.\u00a07\na). Here, ZSM-5 was mechanically treated under a ball mill and sequentially recrystallized with the addition of CTA+ molecules in the autoclave at 180\u00a0\u00b0C for 10\u00a0h. In addition to its inevitable role in preventing excessive coalescence and crystal growth, CTA+ molecules also help assemble the nanosized zeolites and promote the formation of the hierarchical porous structure of the zeolite. The result of the temperature-programmed LDPE pyrolysis test exhibited that T\n50, the temperature at of 0.5 or 50%\u00a0LDPE conversion, shifts gradually to decrease values relative to the blank test without catalyst (476\u00a0\u00b0C) as increase amount of CTAB added (Fig.\u00a07b). The decline\u00a0of T50 was described as following order, M\u2212the zeolite after mechanochemically treated (461\u00a0\u00b0C)\u00a0>\u00a0initial ZSM-5 (423\u00a0\u00b0C)\u00a0>\u00a0MR\u2212mechanochemically treated and recrystallized without CTAB (419\u00a0\u00b0C)\u00a0>\u00a0MRCTAB0.004\u2212mechanochemically treated and recrystallized with CTAB, Si/CTAB of 0.012 (397\u00a0\u00b0C)\u00a0>\u00a0MRCTAB0.012\u2212recrystallized with CTAB, Si/CTAB of 0.012 (395\u00a0\u00b0C). Also, the observed activation energy (Eobs) calculated from Coats-Redferns plot [105] exhibited a similar trend (Fig.\u00a07c). The E\nobs of LDPE without catalyst is 450 kJmol-1, while The E\nobs of LDPE with addition of catalysts exhibited lower values as the following order, M (341\u00a0kJ\u00a0mol\u2212 1)\u00a0<\u00a0initial ZSM-5 (224\u00a0kJ\u00a0mol\u2212 1)\u00a0<\u00a0MR (216\u00a0kJ\u00a0mol\u2212 1)\u00a0<\u00a0MRCTAB0.004 (146\u00a0kJ\u00a0mol\u2212 1)\u00a0<\u00a0MRCTAB0.012 (132\u00a0kJ\u00a0mol\u2212 1). The introduction of hierarchical zeolite for LDPE pyrolysis was also reported by Wardani et\u00a0al. [21]. They introduced post alkaline treated SSZ-13 zeolite for temperature-programmed LDPE pyrolysis reaction. The relative activity of the catalyst is measured as T50.\u00a0Post alkaline treated hierarchical SSZ-13 exhibited the lowest T50 (C250-AT\u00a0=\u00a0460\u00a0\u00b0C) compared to blank test (T50\u00a0=\u00a0476\u00a0\u00b0C), catalysts calcined at 550 without post alkaline treatment (C550\u00a0=\u00a0468\u00a0\u00b0C), and as-synthesized catalyst with post alkaline treatment (AS-AT\u00a0=\u00a0463\u00a0\u00b0C) (Fig.\u00a07d). Also, the C250-AT catalyst\u00a0exhibited the highest IHF (Indexed Hierarchy Factor) value, |\u0394Eobs|/nAl, and |\u0394Eobs|/\u2211\nacid sites (Fig.\u00a07e). The higher activity of hierarchical zeolite in LPDE cracking reaction was related to the presence of the additional meso- and/or macropore, which provide such a molecular highway to alleviate the diffusion constraints of bulky LPDE molecules [106].Furthermore, the zeolite-based catalyst also exhibited remarkable performance for catalytic reaction of high-density polyethylene (HDPE) waste. As reported by Hassan et\u00a0al. [107], the utilization of faujasite-type zeolite for the co-pyrolysis of sugarcane baggase and HDPE increased the calorific value of product oil. It enhanced the liquid yield with the maximum bio-oil yield of 68.56\u00a0wt% and hydrocarbon yield (74.55%) and a minimum yield of oxygenated compounds (acid\u00a0=\u00a00.57% and ester\u00a0=\u00a00.67%). The suitability of ZSM-5 over catalytic cracking of HDPE was also reported by Elordi et\u00a0al. [108]. HZSM-5 zeolites with SAR of 30 and 80 were employed to catalyze the polyethylene feedstock under feed flow of 1g.min\u22121 in 10\u00a0h to a 30\u00a0g catalyst bed. The catalyst demonstrated a very low deactivation nature and a moderate acidity that is useful to modify the product distribution. The catalyst also exhibited common behavior, where the increase of SAR ratio led to a higher yield of C2 \u2013C4 olefins and that of the non-aromatic C5\u2013C11 fraction, and a decrease in the yields of aromatic components and C1\u2013C4 paraffin. Also, the rate of coke production is suppressed as the SAR is increased.The utilization of zeolite for the conversion of waste oil to fuel exhibited remarkable results. The application of zeolite for the conversion of waste cooking oil (WCO) was reported by Li et\u00a0al. [109]. They have effectively carried out the catalytic conversion of WCO to liquid hydrocarbon fuels by utilizing USY zeolites as the catalyst. The catalyst exhibited higher performance in comparison to traditional base catalysts, e.g., Na2CO3 and K2CO3. Interestingly, the reaction could generate liquid hydrocarbon fuels containing C8\u2013C9 alkanes or olefins, which is likely similar to the chemical composition of gas oil-based fuels with the high yield of liquid products was over 75% and low coke formation of 24.7%. Khowatimy et\u00a0al. [110] reported the study of hydrocracking of waste lubricant into gasoline and diesel fraction using the combination of Y-Zeolite and ZnO (Y-Zeolite/ZnO). The catalytic reaction achieved a higher total conversion of 99.49\u00a0wt.% compared to the reaction without catalyst (thermal hydrocracking) with a conversion of 98.99\u00a0wt.%. The catalyst also results in the highest liquid product of 24.75\u00a0wt.% and gasoline and diesel selectivity of 25.92 and 74.08%, respectively.Moreover, AbuKhadra et\u00a0al. [111] introduced alkali modified clinoptilolite for transesterification of commercial waste cooking oil into biodiesel with technical properties of EN 14214 [112] and ASTM D-6751 [113] standards. All the catalysts, i.e., K/clinoptilolite (K/Clino), Na/clinoptilolite (Na/Clino), Ca/clinoptilolite (Ca/Clino), and Mg/clinoptilolite (Mg/Clino) showed promising catalytic activities by achieving biodiesel yields of 93.6%, 95.2%, 96.4%, and 98.7%, respectively. Recently, Fan et\u00a0al. [114] reported the fast catalytic co-pyrolysis of lignin and waste cooking oil for aromatics production using ZSM-5 as a catalyst. The catalyst to feedstock achieved the highest yield of aromatics at the ratio of 3:1. The further increased ratios of catalyst to feedstock enhanced the alkylation and demethoxylation of phenols.Moreover, the suitable ratio of WCO: lignin should also take into account. The optimum ratio of WCO to lignin was achieved at 1:1, which results in the highest mono-aromatic selectivity of 82.6% and a synergistic extent of 52.1%. Ding et\u00a0al. [115] investigated the catalytic characteristics of modified HZSM-5 with separate NaOH/steaming treatment or integrated NaOH-steaming process in the catalytic fast pyrolysis of wasting cooking oil, and it was indicated that the integrated modified HZSM-5 exhibited higher yields of desired BTX aromatics and long-term stability due to the established micro-mesopore hierarchical system and improved acidic properties in HZSM5 zeolite. Nevertheless, most of these studies ignored the effects of the sequential order of desilication and dealumination treatment on the physicochemical property and catalytic performance of HZSM-5.The introduction of zeolites has shown remarkable results for syngas production from the gasification of plastic waste (PE, PP, and terephthalate of polyethylene (PET). Al-asadi et\u00a0al. [116] reported the application of Ca, Ce, La, Mg, and Mn to promote the Ni/ZSM- 5 catalyst for syngas production. The modified catalysts enhanced the reaction rate of the pyrolysis process, resulting in high syngas in the product yields. The maximum syngas production was obtained with a La catalyst. Catalysts can also accelerate the methanization reactions and isomerize the main carbon frame. Increasing temperature and oxygen in the atmosphere led to a higher n-paraffin/n-olefin ratio and more multi-ring aromatic hydrocarbons in pyrolysis oils. The concentration of hydrocarbons containing oxygen and branched compounds was also significantly affected by catalysts.The chemical transformation of CO2 to value-added chemicals and fuels is currently a research challenge, especially in developing more efficient catalysts. Zeolite comprises unique porosity that had been widely reported for direct conversion of CO2 into fuel products. In this process, zeolite is usually coupled with metal or metal oxide catalyst, in which metals promote the CO2 activation process, and zeolites provide the selectivity role by subsequent catalysis process (Fig. 8). Several valuable products, either C1 or C2+, have been extensively produced through various processes [29,30]. Delmelle et\u00a0al. [117] studied the sorption enhanced methanation of CO2 by loading Ni on zeolite 5A and 13X via wet impregnation. The result Ni/zeolite catalysts exhibited high activity and selectivity over CO2 methanation reaction. Ru is more active than Ni due to its high CH4 selectivity and low coke forming properties [118]. Ahmed [119] reported Ru/Y zeolite catalysts with loadings between 1 wt.%-5.4\u00a0wt.% Ru prepared toward ion-exchange method. The optimum loading was achieved at 2.2\u00a0wt% highest selectivities of 72% and 100% gas conversion at 170\u00a0\u00b0C. The utilization of ZSM-5 and silica MFI (S-1) as support for Rh nanoparticles catalyst also exhibited higher methane selectivity. Compared to conventional metal oxide supports, it exhibited high activity for CO2 conversion and high selectivity for CO under equivalent conditions [19]. The application of zeolite as catalyst support for metal loading catalyst was also reported for Pt. S\u00e1pi et\u00a0al. [120] deposited size-controlled Pt nanoparticles on ZSM-5 supports with SAR of 30, 80, and 280. Size-controlled Pt nanoparticles significantly improved the catalytic activity of the conventional H-ZSM-5 resulted in \u223c16 times higher CO2 consumption rate. Moreover, the catalytic activity increased \u223c4 times higher, and CH4 selectivity at 873\u00a0K was \u223c12 times higher.Bahari et\u00a0al. [121] studied the effectiveness of various bimetallic on Fe-Zeolite over CO2 hydrogenation to formic acid. Several metals-loaded catalysts, i.e., Co, Cu, Pd, and Ni, had been examined for their function as the promoter for CO2 hydrogenation. The catalyst of Pd: Fe: Zeolite (0.1:1.25:2) demonstrated the highest formic acid yield of 275.91\u00a0ppm compared to other catalysts. More advanced utilization of bimetallic clusters within zeolite catalyst was reported by Sun et\u00a0al. [122]. The Pd\u2013Mn clusters encaged within S-1 zeolites prepared via ligan-protected method was employed for catalyzing CO2 hydrogenation into formate (Fig.\u00a09\na). The obtained catalysts exhibited extraordinary catalytic activity and durability due to the formation of ultrasmall metal clusters and the synergic effect of bi-metallic components. The highest performance catalyst, PdMn0.6@S-1, exhibited the formate generation rate of 2151 molformate\u2009molPd\n\u22121\u00a0h\u22121\u00a0at 353\u00a0K and an initial TOF of 6860\u00a0mol H2 \u2009mol Pd\u22121\u00a0h\u22121 for CO-free fatty acid (FA) decomposition at 333\u00a0K without any additive (Fig.\u00a09b). Moreover, the DFT calculations were employed to explain the excellent catalytic performance of Pd\u2013Mn clusters over FA decomposition (Fig.\u00a09c). The results demonstrated that alloying Pd with Mn led to the formation of a more compact structure. Also, the addition of Mn slightly passivated Pd\u00a0active sites, preventing overly strong binding with intermediates in FA decomposition. These are key factors affecting the high performance of the catalyst.Furthermore, the metal order on bimetal-modified zeolite support has also significantly influenced the properties of the catalyst. Bacariza et\u00a0al. [123] studied the effect of bimetallic order of Ni-Ci on the USY zeolite support to its performance for hydrogenation of CO2 to methane. Here, three different orders impregnation i.e., Ni before Ce (Ce/Ni), Ce before Ni (Ni/Ce), and co-impregnation (Ni\u2013Ce) are examined to obtain the best way for obtaining the highest catalytic activity. Ce/Ni and Ni\u2013Ce catalyst exhibited stronger Ni\u2013Ce interaction and smaller Ni average size ot 2.5\u00a0nm, at the same time, it enhances the reducibility of the Ni species. However, these catalysts exhibited lower CO2 adsorption capacity than Ce/Ni catalyst. Among all the catalyst tested, Ni\u2013Ce catalyst was the best ordering method as it exhibited the highest catalytic performance.In addition to the effect of particle size of metal loaded and impregnation order of the metals, the preparation, and pre-reduction of the metal loaded catalyst also played an important role in affecting the catalyst's performance. Bacariza et\u00a0al. [124] reported that the drying method, the calcination temperature, and the pre-reduction temperature had influenced the catalytic performance of Ni-based zeolite catalysts for CO2 methanation. A suitable drying method and calcination temperature may result from more reducibility of Ni species and\u00a0at the same time produce good structural and textural properties of support and more homogenous Ni particle size. Drying method under microwave radiation, the calcination temperature at 300\u00a0\u00b0C and pre-reduction temperature at 500\u00a0\u00b0C exhibited maximum catalytic performance of Ni-based zeolite catalyst.The effect of zeolite topology on the direct conversion of CO2 into hydrocarbon products was also studied by Ramirez et\u00a0al. [125]. Both MOR and ZSM-5 are introduced separately with Fe2O3/KO2 for the direct conversion of CO2 to light olefins and aromatics. MOR and ZSM-5 exhibited different catalytic behavior as MOR directly converted CO2 into light olefins (Fig.\u00a010\na) and ZSM-5 into aromatics (Fig.\u00a010b). In addition, both MOR and zeolite boosted the total selectivity to desired hydrocarbon product and enabled the further conversion of undesired CO. Moreover, the remarkable difference in selectivity between the two zeolites is further rationalized by first-principles simulations, which show a difference in reactivity for crucial carbenium ion intermediates in MOR and ZSM-5 (Fig.\u00a010c). Further study using DFT simulations exhibited the future potential of ZSM-5 to activate long alkenes via carbenium ions which may lead to higher selectivity toward the formation of aromatic products (Fig.\u00a010d).Moreover, Bonura et\u00a0al. [126] also introduced two different zeolites of FER and MOR for the catalyst support of CuZnZr. The catalyst was utilized for CO2 conversion to DME in a fixed bed reactor. The hybrid CuZnZr-FER catalyst exhibited better activity-selectivity and an interesting DME productivity, with no coke formation under the experimental condition (Fig.\u00a011\na). The better dispersion of CuZnZr catalyst on the 2D FER framework than MOR zeolite leads to a more efficient mass transfer of MeOH from CuZnZr sites to the zeolite surface; therefore, favoring the formation of DME with higher yields. Furthermore, Bonura et\u00a0al. [127] also studied the effect of catalyst acidity of CuZnZr-FER for the direct CO2 to DME reaction. The catalyst exhibited a different behavior in terms of stability which was influenced by different acidity (Fig.\u00a011b). The acidity of the catalyst was controlled by varying the alumina amount to obtain different Si/Al molar ratios of 8, 30, and 60. Catalysts with higher SAR experienced higher metal loss during the reaction. However, all three catalysts demonstrated only a slight difference in conversion, selectivity, and product yield (Fig.\u00a011c).Ayodele et\u00a0al. [128] reported the introduction of a tandem zeolite-based catalyst for CO2 hydrogenation to methanol (MeOH). The prepared catalyst Cu/ZnO/ZSM-5 denoted CZZSM (Fig.\u00a012\na\u2013c), exhibited impressive CO2 conversion of 20.25% and higher selectivity of MeOH (77.7%) in comparison with other catalysts supported on Al2O3, SiO2, ZrO2 (Fig.\u00a012d). Other tandem zeolite-based catalysts were reported by Li et\u00a0al. [129]. They utilized Pd/ZnO/ZSM-5 for the hydrogenation reaction of CO2 to dimethyl ether (DME). The spatial configuration of Pd/ZnO and ZSM-5, which was adjusted by tuning the size of Pd/ZnO, defined the DME yield. The close proximity of both catalysts was unfavorable for DME production, resulting in lower selectivity due to the displacement of Br\u00f8nsted acids in ZSM-5 by low-valent Zn cation. The synthesis method of the catalyst via powder mixing demonstrated the best configuration that promoted the highest CO2 conversion with the largest selectivity of DME. Compared to conventional ZSM-5 catalysts, the introduction of Pd/ZnO on ZSM-5 results in excellent long-term stability in 60\u00a0h with less coke. Another application of tandem zeolite-based catalyst was carried out by Dai et\u00a0al. [130] for direct hydrogenation of CO2 to aromatics. It comprises an iron-potassium bimetal-modified alkaline Al2O3 catalyst and a phosphorus-modified ZSM-5 zeolite denoted as Fe\u2013K/Al2O3\u2013P/ZSM-5. The catalyst was prepared by several methods, i.e., powder-mixing, granule-mixing, dual-bed, and multi-bed techniques. Catalysts prepared by granule-mixing produce the highest aromatic products and lowest CO formation. During the reaction, P/ZSM zeolite provides acid sites which may transform lower olefin intermediates into aromatics. On the other hand, the Fe\u2013K/a-Al2O3 served as the metal active center to hydrogenate CO2 to lower olefin intermediates. Dai et\u00a0al. [130] also underlined the proximity of two components as its effect on the selectivity of the reaction.Furthermore, the appropriate loading of phosphorus on ZSM-5 (0.8\u00a0wt%) increased the amount of medium-strength acid site, resulting in higher production of aromatics and higher conversion of CO2. The high efficient conversion of CO2 to aromatic products was also obtained by utilizing a tandem Cr2O3/H-ZSM-5 catalyst [131]. The synergistic effect between two components enables aromatics selectivity to reach \u223c76%, CO2 conversion of 34.5%, and there was no catalyst deactivation after 100\u00a0h, which indicates the catalyst's long-term stability. Nevertheless, the catalytic performance of the catalyst could be further optimized by adjusting the acid strength of zeolites and the mass ratio of Cr2O3/H-ZSM-5. The proximity between two catalysts was also an important factor. It played an emerging role during the direct conversion of CO2 to aromatics.Furthermore, the utilization of a zeolite-based catalyst for CO2 hydrogenation was also in the form of a bifunctional zeolite-based catalyst. This catalyst comprises two different zeolite materials which bond together. Cr2O3/H-ZSM-5@S-1 comprising a core\u2013shell structured H-ZSM-5@S-1 zeolite capsule catalyst was found to be a promising bifunctional catalyst for CO2 hydrogenation aromatics [131]. The passivation effect of S-1 has suppressed the undesired reaction on the external site of zeolite and increased the selectivity of aromatic products. The selectivity of BTX and PX increases from 13.2% to 7.6%\u201343.6% and 25.3%, respectively. Noreen et\u00a0al. [132] also reported utilizing bifunctional zeolite-based catalyst over CO2 hydrogenation to multibranched isoparaffins. The catalysts were configured of Na/Fe3O4 (NaFe), Zeolites SAPO-11, and ZSM-5. The zeolites have improved the long-chain-hydrocarbon selectivity by performing oligomerization on the shorter chains. The catalyst test of dual-bed reactions with SAPO-11 and ZSM-5 coupled with NaFe separately exhibited that SAPO-11 plays a pivotal role in enhancing the isomerization activities on the short-chain hydrocarbons. On the other hand, ZSM-5 boosted the isomerization activity in all aspects, especially on the long-chain hydrocracking and\u00a0promoting aromatization activity. Combining triple-bed SAPO-11, ZSM-5, and SAPO-11 results in enhanced isomerization and suppressed aromatization activities. Therefore, isoparaffins' selectivity and multibranched isomers yield increased at the same time the aromatic products decreased.Ordered mesoporous silica (OMS) materials have attracted growing interest to be considered important in heterogeneous catalysis. Their large surface area highly ordered porous architecture, and the ability for metal atoms to load within the mesopores lead them to be emerging support materials for designing various catalysts. M41S was the first OMS material reported. Discovered by the scientist of Mobil corporation in 1992 [133], this catalyst exhibits well-defined hexagonal cylindrical mesopores with a relatively narrow pore size distribution. This solid possesses periodic arrangements of mesoscale porosity, but the framework pore walls are built of amorphous silica. Since the discovery of the MCM family in the early 90s, considerable progress has been achieved regarding the application of ordered mesoporous silica in various fields. The tremendous applications in the various fields have boosted the development of many other OMS materials such as the folded sheet mesoporous material-16 (FSM-16) [134], families of Santa Barbara Amorphous (SBA-n) [135,136], Fudan University Material (FDU) [137], Korea Advanced Institute of Science and Technology (KIT) [138] and anionic-surfactant-templated mesoporous silica (AMS) [139].OMS materials are generally synthesized by using a surfactant forming regularly aligned assemblies that are used as a template for the metal oxide, followed by template removal (Fig.\u00a013\na) [140]. The flexibility of the templating methods permits the synthesis of materials with a controlled pore size and structure, controlled wall compositions, and highly interconnected surface areas, all of which allow the optimization of the material for the specific application required. Fig.\u00a013b shows the structure of several OMS materials. SBA-15 and MCM-41 were put as examples for OMS materials that comprise two-dimensional hexagonal phases with the P6mm symmetry, composed of close-packed hexagonal arrays of cylindrical surfactant micelles [141,142]. Fm \n\n\n3\n\u00af\n\n\n m and Im \n\n\n3\n\u00af\n\n\n m phases are included in the cubic mesophases. Both symmetries are built by spherical micelles in a cubic cage structure. The OMS materials possess face-centered cubic (Fm \n\n\n3\n\u00af\n\n\n m) pore structure, such as FDU-12, is considered as close packing of spherical mesopores with every mesopore connected to 12 nearest neighboring mesopores while for body-centered cubic (Im \n\n\n3\n\u00af\n\n\n m) pore structure such as SBA-16, each mesopore is connected to 8 neighboring mesopores [143]. Other OMS materials, such as MCM-48 and KIT-16, are characterized by their bicontinuous cubic gyroid phase with the Ia \n\n\n3\n\u00af\n\n\n d symmetry. This cubic mesophase possesses networked and interconnected pores, regarded as two interwoven cylindrical channels that exhibited similar adsorption properties to 2-dimensional hexagonal materials without a pore-blocking effect. On the other hand, the mesophases with lamellar symmetry, such as MCM-50, consist of silica layers, not ordered mesoporous as others mentioned [144].OMS materials are characterized for their water-soluble, chemically, and thermally stable with mechanical strength and are toxicologically safe. These remarkable properties made OMS currently applied in countless applications. In the catalysis field, they are widely used as both catalyst and catalyst supports. They are also applied for liquid\u2013solid or gas\u2013solid adsorption [145\u2013147], pollutant removal and remediation [148] as well as sensors [149] and drug delivery systems [150]. Furthermore, in energy-related devices, OMS materials have attracted much attention due to their unique properties compared to bulk materials, i.e., large surface area, tuneable pore size, and high metal loading and functionalization ability. So that, these materials are widely applied during renewable energy production, whether as electrocatalyst or photocatalysis in the renewable energy generator, i.e., hydrogen production [139,141\u2013156] and CO2 conversion to fuels [157\u2013171].Several works reported the utilization of OMS as a catalyst for hydrogen production. The OMS materials are introduced in several routes of hydrogen production, i.e., hydrocarbon steam and dry reforming, photo (electro)catalytic water splitting, and biomass gasification. The utilization of OMS on these three processes will be discussed in the following subsection. In addition, their catalytic performance on the renewable energy production was summarized in Table 2\n.Hydrocarbon steam reforming is the production of syngas from the hydrocarbon (usually employed naphtha as feedstock) and steam. This is different to dry reforming, in which syngas was generated from the reaction between methane (CH4) and carbon dioxide (CO2). Generally, these reactions employed several noble metals (Pt, Pd, Rh, and Ru) or low-cost transition metals (Ni, Fe, Co) as the active component. During reactions, the deposition of the carbonaceous species on the active site could lead to the deactivation of the catalyst. In that sense, embedding the active sites onto suitable support, such as ordered mesoporous silica could enhance the coking and sintering resistance of the catalyst [152]. The utilization of MCM-41 and Zr incorporated MCM-41 (25Zr-MCM-41) supported Ni, or Cu catalysts was reported by Cakiryilmaz et\u00a0al. [153] for steam reforming of acetic acid reaction, at 750\u00a0\u00b0C. Physicochemical characterization reveals the presence of some deformation in the ordered pore structure of MCM-41 after Zr incorporation. Furthermore, catalysts supported zirconia incorporated MCM-41 exhibited higher catalytic stability compared to the MCM-41 supported materials. The impregnation of Ni was more suitable for enhancing the catalytic performance. On the other hand, the impregnation of copper leads to the decarboxylation reaction of acetic acid. This reaction yields large quantities of methane. Recently, Liu et\u00a0al. [154] loaded different compositions of LaNiO3 on MCM-41 supported as the catalyst for steam reforming of biomass in-situ tar reaction in the double fixed-bed reactor. The pure LaNiO3 was used as control catalysts. The in-situ tar was obtained from the pyrolysis of rice husk at 450\u00a0\u00b0C in the first reactor bed. Simultaneously, at second bed the LaNiO3 and XLaNiO3/MCM-41 (X\u00a0=\u00a00.025, 0.05, 0.075 and 0.1) catalysts were performed for hydrogen-rich syngas production at different reforming temperature of 500\u00b0C\u2013900\u00a0\u00b0C and steam to carbon ratio (S/C)\u00a0=\u00a00.6-1 (w.t). The catalytic test results exhibited that 0.1LaNiO3/MCM-41 produced a higher hydrogen yield of 61.9Nm3/kg at 800\u00a0\u00b0C and S/C of 0.8. Moreover, the catalyst also experienced good stability as its ability to produce hydrogen gas composition of around 50% after five-time cycles.Furthermore, Calles et\u00a0al. [155] reported other types of OMS catalysts, SBA-15, as support for Co-based catalysts in steam reforming of acetic acid. A series of Co\u2013Cr/SBA-15 extrudates prepared by varying the binder (bentonite) content and particle size were introduced into the acetic acid steam reforming tests at 600\u00a0\u00b0C and WHSV of 30.1 h\u22121. The catalytic performance test of the catalysts demonstrated that the extruded particles with an effective diameter of 1.5\u00a0mm and 30\u00a0wt% of bentonite get similar conversion and hydrogen selectivity than the powder sample. Al-salihi et\u00a0al. [156] introduced a series of SBA-15 supported catalysts, i.e., Co-SBA-15, Ni-SBA-15, Co\u2013MgO-SBA-15, Ni\u2013MgO-SBA-15, and Co\u2013Ni-SBA-15 prepared using a one-pot hydrothermal and impregnation methods for glycerol steam reforming (GSR) reaction at reforming temperature range of 450\u00b0C\u2013700\u00a0\u00b0C. Results from the GSR catalytic studies showed that both Co\u2013Ni-SBA-15 prepared via impregnation and one-pot hydrothermal method were resistant to deactivation, and both yielded 100% glycerol conversion for a continuous of 40\u00a0h. Moreover, the catalyst comprises 10wt.% Co and 5wt.% of Ni, which achieved H2 selectivity of (70\u201378) % and (60\u201378) %, respectively. The effect of MgO addition was also studied, the catalyst incorporated MgO exhibited higher activity and stability.The introduction of SBA-15 prepared from renewable sources as catalyst support was reported by Abdullah et\u00a0al. [157]. SBA-15 support was synthesized using extracted silica from palm oil fuel ash waste (POFA). Different impregnation techniques were used to prepare Ni/SBA-15 catalysts, i.e., the ordinary impregnation technique (Ni/SBA-15(IM)), rotary evaporator-assisted impregnation, (Ni/SBA-15 (RE)), shaker-assisted impregnation (Ni/SBA-15(SH)) and ultrasonic-assisted impregnation (Ni/SBA-15(US)). Ni/SBA-15 catalysts' performance and stability were tested in the stainless steel fixed-bed reactor setup at 800\u00a0\u00b0C for up to 24\u00a0h. The highest catalytic performance was achieved by Ni/SBA-15(US) owing to the cavitation effect of ultrasonic irradiation, resulting in better dispersion and smaller Ni particles inside the SBA-15 micelles. Both mentioned properties lead to stronger Ni\u2013O\u2013Si interaction, higher catalyst basicity, and suppress graphite carbon formation on Ni/SBA-15(US).Moreover, Abdullah et\u00a0al. [158] also studied the effect of different amounts of Ni loaded on SBA-15(POFA) for CO2 reforming of CH4 (CRM) in a stainless steel fixed-bed reactor\u00a0under a temperature of 800\u00a0\u00b0C and an atmospheric pressure of 1:1 CO2:CH4 (v:v). As observed through N2 adsorption\u2013desorption and XRD analyses, the increase of Ni loading on SBA-15(POFA) from 1 to 5\u00a0wt.% decreased the BET surface area and crystallinity of catalyst. Moreover, the increment of Ni loading from 1 to 3% enhanced the catalytic performance of CRM. However, the catalytic performance then decreased at 5% Ni loading. This result was due to Ni's even distribution and good Ni\u2013O\u2013Si interaction of 3\u00a0wt.% Ni/SBA-15(POFA) as proved by TEM, FTIR, and XPS. Lowest H2/CO ratio and catalyst activity and stability of 1\u00a0wt.% Ni/SBA-15(POFA) was due to the weaker Ni\u2013O\u2013Si interaction and a small amount of basic sites that favor the reverse water gas shift (RWGS) reaction and carbon formation.Furthermore, Jiang et\u00a0al. [159] reported the deployment of nanocomposite of mesoporous silica-supported Ni nanocrystals modified by ceria clusters (Ni\u2013CeO2/SiO2) for CO2 reduction by CH4 to produce CO and H2 (CRM) under focused UV\u2013visible\u2013infrared illumination in the absence of additional heater. High production rates of CO and H2 (41.53 and 33.42\u00a0mmol\u00a0min\u22121 g\u22121) and high light-to-fuel efficiency of 27.4% are obtained. The catalyst's marked CRM activity was due to highly effective light-driven thermocatalytic CRM due to good photothermal behavior of entire solar spectra of the nanocomposite. The presence of oxygen atom in ceria cluster participates in CRM reaction on Ni nanocrystals. It helps to significantly decrease the activation energies of C\u2217 and CH\u2217 the oxidation, which is CRM dominant steps. Also, the focused illumination for the experiment considerably decreases the\u00a0activation energy of CRM, thus also contributing to the enhancement of the photothermocatalytic activity of the catalyst.The Ni\u2013CeO2/SiO2 catalyst system also displayed good stability over long-term photothermocatalytic CRM reactions. The result of the photothermocatalytic durability test (Fig.\u00a014\na and b) demonstrated that the gradual decline of catalytic performance happened during the initial 27\u00a0h of illumination time. After that, the catalyst could maintain its remarkable performance even until 100\u00a0h, as evidenced by the production reaction rates of CO and H2O of 24.61 and 27.64\u00a0mmol\u00a0min\u22121 g\u22121, respectively. Moreover, in comparison with Ni/SiO2 (0.53\u00a0mmol\u00a0min\u22121g\u22121), Ni\u2013CeO2/SiO2 possesses a low carbon deposition rate of 0.034\u00a0mmol\u00a0min\u22121g\u22121 (Fig.\u00a014c). Here, it is indicated that the presence of ceria clusters on the surface of Ni nanocrystals markedly inhibits the carbon deposition rate. The carbon deposition inhibiting the ability of the ceria cluster was confirmed by HRTEM observation. HR-TEM observation demonstrated {111} facets (0.204\u00a0nm) of Ni within Ni\u2013CeO2/SiO2 catalyst (Fig.\u00a04d). In contrast, the surface of Ni nanocrystals in the used Ni/SiO2 catalyst is covered by graphite layers (Fig.\u00a04e), thus resulting in its deactivation.Generally, the photocatalytic process begins with the absorption of light by photocatalyst to create electron (e\u2212)\u2013hole (h+) pairs, then followed by the charge separation or migration process to the surface of photocatalyst, and simultaneous hydrogen evolution reaction (HER) and oxygen evolution reaction (OER) to overall water splitting (Fig.\u00a015\n). Several studies showed that incorporating the photocatalyst with OMS could enhance the activity due to the benefits of the well-connected pore channel structure. For instance, the multiple light reflection in the pore channels and the enrichment of the isolated photocatalyst site in the transparent mesoporous silica matrix allow full photon exposure and utilization [160]. Therefore, the use of OMS materials in photocatalysis water-splitting has been extensively studied. Shen and Guo [161] utilized Cr incorporated, and Cr and Ti co-incorporated MCM-41 photocatalysts synthesized by hydrothermal method for photocatalytic water splitting under visible light irradiation (l\u00a0>\u00a0430\u00a0nm). The catalytic examination demonstrated that the photocatalytic activities of Cr-MCM-41 and Cr\u2013Ti-MCM-41 decreased with an increase in the amount of Cr incorporated. Compared with the Cr-MCM-41, which had the same amount of incorporated Cr, the Cr\u2013Ti-MCM-41 exhibited much higher hydrogen evolution activities. Liu et\u00a0al. [162] reported the application of CdS/M(x)-MCM-41 (M\u00a0=\u00a0Zr, Ti, x\u00a0=\u00a0molar ratio of M/Si) photocatalysts prepared via three different methods, i.e., hydrothermal, ion-exchange, and sulfidation process for photocatalytic hydrogen production. The CdS/Zr (0.005)-MCM-41 and CdS/Ti (0.02)-MCM-41 was the highest hydrogen evolution activity in triethanolamine aqueous solution under visible light (l\u00a0>\u00a0430\u00a0nm) irradiation owing to the diffusion velocity of the reactants and\u00a0resultants and the protection which MCM-41 provided for\u00a0CdS.Furthermore, Al-doped MCM-41 (Al-MCM-41 zeolite) was introduced as catalyst support of NiCr2O4 [163]. The catalyst was synthesized via the facile sol\u2013gel method and followed by calcination at 700\u00a0\u00b0C. According to the results of photocatalytic hydrogen evolution examination, the NiCr2O4/Al-MCM-41 hybrid demonstrated higher hydrogen generation compared with individual NiCr2O4 or Al-MCM-41. NiCr2O4/Al-MCM-41 (50%wt.) exhibited the highest hydrogen production of 8.92\u00a0mmol/g, up to 2.4 times of NiCr2O4 (3.75\u00a0mmol/g) without mesoporous silica support (Fig.\u00a016\na and b). The excellent synergistic effect of NiCr2O4 and Al-MCM-41 zeolite is thought to be responsible for the remarkable enhancement, where the Al-MCM-41 mesoporosity increased the dispersion of NiCr2O4, thus improving the number of active sites. Also, Al-MCM-41 could enhance the charge transfer and inhibit the fast recombination of photo-generated electrons\u2013holes during a reaction. Here, the Lewis acids and bases sites of Al-MCM- 41 zeolites act as electron acceptors and donors, respectively. Thus, in the case of NiCr2O4/Al-MCM-41, the photogenerated electrons are transferred from NiCr2O4 to Al-MCM-41 zeolite support. This electron transfer mechanism leads to the inhibition of electrons and holes recombination (Fig.\u00a016c).In addition to MCM-41, MCM-48 was also reported as a catalyst for hydrogen production via a photocatalytic water splitting system. Zhao et\u00a0al. [164] also reported the application of CdS incorporated MCM-48 prepared via the ion-exchange method for photocatalytic hydrogen evolution under visible irradiation. The SEM observation revealed that the cubic phase of MCM-48 was destroyed during the ion-exchange reaction and sulfidation process. However, the CdS/MCM-48 exhibited high photocatalytic activity for hydrogen evolution. Zhao et\u00a0al. [165] also reported using Ti-modified MCM-48 (Ti-MCM-48) for photocatalytic hydrogen evolution from the methanol\u2013water mixture under UV irradiation. The catalyst achieved high activity without platinum cocatalyst, which is usually essential to TiO2 photocatalyst for hydrogen evolution. Moreover, Peng et\u00a0al. [166] introduced Ti-MCM-48 as support for CdS catalyst. The hybrid catalyst was examined for visible light-illuminated photocatalytic hydrogen production via water splitting without Pt as a co-catalyst. The highest hydrogen production rate and the most apparent quantum yield were as high as 2.726\u00a0mmol/h/gcatalyst and 36.3%, respectively.Moreover, Sanches et\u00a0al. [167] introduced Ti modified SBA-15 photocatalyst synthesized using two titanium sources (Ti tetra-isopropoxide and TiO2\u2013P25) with two Si/Ti molar ratios (20 and 40) for hydrogen production via water splitting. SBA-15 was introduced to enhance TiO2 efficiency. TiO2 prepared from isopropoxide source exhibited highly dispersed Ti-nanoclusters and Ti-coordinated species (penta-, hexa-, or octahedral) in the SBA-15 hexagonal mesoporosity. On the other hand, TiO2\u2013P25 introduction led to Ti dispersion into the mesostructure and octahedrally coordinated into the SBA-15. In addition, the increment of Si/Ti molar ratios induced more concentrated Ti coordinated isolated species and enhanced the photoactivity of the catalyst prepared from Ti tetra-isopropoxide as Ti source. The Al-modified SBA-15 was also utilized as catalyst support toward the water-splitting process for hydrogen generation/storage under visible light irradiation [168]. Iron oxide NPs (5 %wt.) were loaded onto Al (25 %wt.)-modified SBA-15 supports via microwave-assisted and ultrasonic-assisted routes.Interestingly, the different applied loading routes result in different catalytic properties, where the ultrasonic-assisted loading technique exhibited more prevailing behavior through the photocatalytic water-splitting reaction. In addition to SBA-15, Mac\u00edas-S\u00e1nchez et\u00a0al. [169] introduced mesoporous silica SBA-16 as catalyst support of Cd(1-x)ZnxS solid solutions (x\u00a0=\u00a00.05\u20130.3) for hydrogen production from water splitting under visible light. Cd(1-x)ZnxS nanoparticles are unevenly distributed on both external surfaces and within the pore network. The increment of Zn loading up to x\u00a0=\u00a00.2 leads to an enhancement of the\u00a0bandgap energy, which inhibits the enhancement of the bandgap energy, which inhibits the enhancement of the bandgap energy, which inhibits fast photorecombination. Subsequently, H2 evolution was also improved.Biomass gasification is also a promising route for renewable hydrogen production. In that sense, several reports have demonstrated the successful utilization of ordered mesoporous silica for this reaction. It acts as efficient support for various metal nanoparticles, acidic and basic sites, that provide remarkable catalytic performance [32]. Moreover, it plays a vital role in increasing the catalyst efficiency by allowing the high dispersion of metal catalyst, providing large surface areas and affinity for the formation of strong metal-support interaction, and high hydrothermal stability [170]. Wu et\u00a0al. [34] reported the utilization of Ni(x)/MCM-41 (x\u00a0=\u00a05, 10, 20, and 40\u00a0wt.%) for the steam pyrolysis-gasification of wood sawdust for hydrogen production in a two-stage fixed bed reaction system. At the first stage, the wood sawdust was pyrolyzed, and at the second stage, the derived products were gasified to produce hydrogen. The 5\u201320\u00a0wt.% Ni/MCM-41 catalysts exhibited homogeneously dispersed NiO particles inside the pores; however, the 40\u00a0wt%. Ni/MCM-41 experienced more bulky NiO particles (up to 200\u00a0nm particle size) detected outside the pores. Moreover, the increment of Ni loading from 0.5 to 4.0\u00a0wt.% increases gas production and hydrogen production from 40.7 to 62.8\u00a0wt.% and 30.1 to 50.6 vol.%, respectively. This catalyst also exhibited low coke deposition at the range of 0.5\u20134.0\u00a0wt.%.Furthermore, Wu et\u00a0al. [171] detailed their research to describe the effect of different location of Ni particles (inside; I-series and outside; O-series) at the pores of the supports. Interestingly, the I-series catalyst comprises 20%wt. Ni loading generated more gas and hydrogen and a lower oil fraction (21.26\u00a0mmol H2/gwood) than the O-series catalyst with 20 %wt. Ni loading (16.46\u00a0mmol H2/gwood). The better performance of the I-series catalysts is thought to be higher interaction between reactants and active Ni sites inside the pore of the supports.The modern world needs to find an alternative fuel that can replace nonrenewable fossil fuels. Being a versatile, sustainable, efficient, and clean energy carrier, hydrogen has the potential to play that role. CO2 reforming of CH4 (CRM) is a progressing technology for hydrogen production. CO2 reforming of methane is among promising renewable routes for the production of syngas. Kaydough et\u00a0al. [172] employed SBA-15 contained two series of Ni (2.5\u20137.5\u00a0wt%) and Ce (6\u00a0wt%) for dry reforming of methane. In the Ni enriched samples, the NiO particle was formed and entrapped in the porous channels to preserve the porosities. Also, the Ce-modified samples demonstrated highly dispersed CeO2 nanoparticles. The synthesized catalysts show high activity and selectivity towards H2 and CO at atmospheric pressure with full CH4 conversion below 650\u00a0\u00b0C. Moreover, the catalyst also exhibited high stability as it demonstrated low carbon amount formation and limited sintering of the Ni nanoparticles after prolonged tests performed at 500\u00a0\u00b0C for 12\u00a0h.Furthermore, Ibrahim et\u00a0al. [173] examined the effect of different catalyst promoters on the catalytic performance of 5%Ni\u00a0+\u00a01%x/MCM-41 (x\u00a0=\u00a0Ga, Gd, Sc, Ce, or Cs) catalysts for the production of synthetic gas. The chemical analyses demonstrated that the promoters could increase the metal-support interactions, thus improving the catalytic performance. Moreover, the introduction of Gd, Sc, Cs, or Ce-promoted catalysts yielded the lowest amounts of coke formation. However, the catalyst exhibited different CH4 and CO2 conversions. The introduction of 1% of Ga, Gd, or Ce catalyst promoters increased CH4 and CO2 conversions by 38%.On the other hand, 1% Sc or Cs decreased CH4 and CO2 conversion by 18% or 93% and 16% or 92%, respectively. Additionally, the effect of organic promoters for Ni-MCM-41 was also interesting to be discussed. Xu et\u00a0al. [174] reported the utilization of alcohol-promoters Ni-MCM-41 catalyst for CRM. The alcohol-during impregnation-promotes Ni2+ species into the channels of MCM-41, thereby strengthening the metal-support interaction. Also, the introduction of alcohol decreases the particle size of Ni. It increases the surface adsorbed oxygen species over the surface of the support, thus promoting the coke resistance of the catalysts.The advantages of the structural properties of OMS have also had a positive effect on CO2 hydrogenation. It has been investigated as efficient support owing to the high surface area, which provides a high catalyst dispersion and the ordered structure that confined the metals nanoparticle catalyst, inducing a favorable reaction pathway for high selectivity to the desired product [175]. Kiatphuengporn et\u00a0al. [176] carried out CO2 hydrogenation to alcohols over Cu-based MCM-41 catalyst. The series of Cu (10\u00a0wt.%) based catalysts with different pore characteristics of MCM-41 supports including unimodal (SS) and bimodal (T) pore structures, loading amount of Fe (0, 0.5, and 3\u00a0wt.%), and reaction temperature on the catalytic performance were introduced. The result showed that the activities of CO2 hydrogenation over bimodal support catalysts were higher than unimodal support catalysts. The bimodal 3\u00a0wt.% Fe\u201310Cu/MCM-41 catalyst exhibited the highest CO2 conversion of 20.8% (at 350\u00a0\u00b0C), alcohol selectivity of 80\u201399% (at 160\u2013200\u00a0\u00b0C), and highest TOF of alcohols and CO of 1.08 \n\n\u00d7\n\n 10\u221225 and 5.47\u00a0\u00d7\u00a010\u221225\u00a0mol surface metal atom\u22121 min\u22121, respectively. These outstanding catalytic activities could not be indispensable to the pore characteristics of the supports, where the presence of larger mesopore promote the formation of metals with larger sizes, resulting in weaker metal\u2013support interaction of which more favorable in this reaction.Moreover, Kiatphuengporn [177] interestingly examined the effects of magnetic field orientation and magnetic flux density on activity and selectivity of ferro/ferrimagnetic xFe/MCM-41 for CO2 hydrogenation. The result demonstrated that the magnetic field promotes the reactant adsorption and surface reaction over the magnetized Fe catalysts, resulting in the lower apparent activation energy and the increase of selectivities to hydrocarbons and CH3OH. Fig.\u00a017\na illustrates the proposed mechanism of CO2 hydrogenation without and with the magnetic field. Without the magnetic field, CO2 is converted into CH4 and CO products. On the other hand, CO2 is selectively converted to C2\u2013C3 hydrocarbons and methanol under magnetic field-induced reaction. Moreover, the presence of an external magnetic field, especially in the north-to-south (N\u2013S) direction, could significantly improve CO2 conversions by 1.5\u20131.8 times (Fig.\u00a017b); nevertheless, the activation energy was 1.12\u20131.15 times lower than those without a magnetic field. The increase of CO2 conversion in the magnetized catalyst might be attributed to the advancement of CO2 adsorption of the catalyst, then the apparent activation energy of the reaction is reduced. In addition to the advancement of higher CO2 conversion, applying external magnetic field also result in higher selectivity to C2\u2013C3 hydrocarbons and methanol (Fig.\u00a017c). These findings indicated that the magnetic field might facilitate the chain growth probability of the reaction product to form C2\u2013C3 hydrocarbons and methanol. Here, the optimum CO2 conversion and selectivity to CH3OH and C2\u2013C3 hydrocarbons were achieved at the condition of the presence of 27.7\u00a0mT of an external magnetic field in N\u2013S direction.The utilization of MCM-41 based catalyst was also reported by Taherian et\u00a0al. [178]. Here, different amounts of Yttria (Y2O3) were doped with Ni over MgO-modified MCM-41 support. The study found that the addition of NiO and promoters over MCM-41 support doesn't alter its mesoporous structure. However, it may decrease the BET surface area, and pore volume as the presence of both NiO and MgO could cause partial blockage of the pores. The presence of Yttria promotes higher Ni dispersion and that of improving CO2 adsorption sites on basicity sites. The result exhibited that all catalysts which contained Yttria showed improved catalytic activity compared to those of Ni/MgO-MCM-41. Also, the addition of yttria could lessen the temperature of maximum conversion and selectivity. Here, the highest catalytic performance was owned to catalyst with 2\u00a0wt% of yttria. The catalyst possesses CO2 conversion and methane selectivity as high as 65.55% and 84.44% at 673\u00a0K, respectively, and high stability after 30\u00a0h.Recently, the utilization of MCM-41 catalyst support for CO2 hydrogenation was reported by Seker et\u00a0al. [179]. The hybrid MCM-41-supported tungstophosphoric acid (TPA) catalyst with a commercial CuO\u2013ZnO\u2013Al2O3 catalyst was performed in the dimethyl ether (DME) synthesis by CO2 hydrogenation. Different amounts of TPA loading of 30, 40, 60, and 80\u00a0wt% lead to various TPA cluster structural distortions and the change of acid properties. A suitable amount of TPA loading could result in a high density of acid sites, thus improving catalytic performance. Here, the optimum condition was obtained at TPA loading of 60\u00a0wt% with CuO\u2013ZnO\u2013Al2O3:TPA/MCM-41\u00a0=\u00a04:1\u00a0at 40,000\u00a0mL CO2 gcat\n\u22121h\u22121 and H2:CO2\u00a0=\u00a03:1\u00a0at 250\u00a0\u00b0C and 45\u00a0bar. At these conditions, the DME production rate was 1551.5 gDME kgcat\n\u22121h\u22121. Moreover, Wang et\u00a0al. [180] utilized Ni/xMg@MCM-41 (x\u00a0=\u00a00, 0.05, 0.1) catalysts synthesized via a novel in-situ one-pot method for CO2 methanation. The effect of Mg concentration was examined. Interestingly, the introduction of Mg into the support framework can significantly improve the basicity of the catalyst, which induces the adsorption and activation of CO2. So, all the synthesized catalysts demonstrated good thermal stability and catalytic activity. However, the best catalytic performance was owned to Ni/0.05\u00a0Mg@MCM-41 catalyst as it demonstrated the highest low-temperature reaction activity.Besides MCM-41, SBA-15 also demonstrated good catalytic performance for CO2 hydrogenation reactions. Lin et\u00a0al. [181] reported the utilization of SBA-15 as support for three kinds hybrid catalysts of CuO\u2013ZnO/SBA-15 (CZ/SBA-15), CuO\u2013ZnO\u2013MnO2/SBA-15 (CZM/SBA-15) and CuO\u2013ZnO\u2013ZrO2/SBA-15 (CZZ/SBA-15). The catalysts were performed for catalytic hydrogenation of CO2 to methanol on a fixed bed reactor. The results show that the introduction of metal oxide in the catalyst changes the pore size and specific surface area of the SBA-15 molecular sieve support. The utilization of CuO\u2013ZnO\u2013ZrO2/SBA-15 achieved the optimum methanol selectivity of 25.02%. It is 28% and 136.9% higher than CuO\u2013ZnO/SBA-15 and CuO\u2013ZnO\u2013MnO2/SBA-15, respectively. Moreover, Mureddu et\u00a0al. [182] utilized Cu/ZnO@SBA-15 and Cu/ZnO/ZrO2@SBA-15 nanocomposites synthesized by innovative impregnation-sol\u2013gel auto combustion combined strategy for carbon dioxide hydrogenation to methanol. The composites comprise different active phase loading (20 and 35\u00a0wt.%) and Cu/Zn molar ratio (1.0\u20132.5\u00a0mol\u00a0mol\u22121). According to the result of characterization techniques, the active phase of the catalyst was highly dispersed into/over the well-ordered mesoporous channels, especially at low loading and low Cu/Zn molar ratio. Thus, this result corroborated the catalytic results, where the catalyst with the lowest Cu/Zn molar ratio (1.0\u00a0mol\u00a0mol\u22121) exhibits the best catalytic performance with a STY of methanol of 376 mgCH3OH\u00b7h\u22121\u00b7gcat\n\u22121. This obtained STY value was much higher than the unsupported catalyst (10 mgCH3OH\u00b7h\u22121\u00b7gcat\n\u22121).Furthermore, Li et\u00a0al. [183] reported utilizing monometallic Pd or Ni/SBA-15 and bimetallic Ni\u2013Pd/SBA-15 alloy catalysts with different ratios of Ni/Pd as a catalyst for CO2 methanation. The catalytic examination demonstrated that bimetallic Ni\u2013Pd/SBA-15 catalysts owned higher catalytic activity than monometallic Pd- or Ni/SBA-15. The bimetallic catalyst exhibited the highest catalytic performance with Ni:Pd atom ratio of 3:1, which yielded 0.93\u00a0mol CH4 per mol CO2 at 430\u00a0\u00b0C. This outstanding performance was due to Ni and Pd's synergistic effect with the high dispersion of active sites. Another Pd-based SBA-15 supported catalyst was reported by Jiang et\u00a0al. [184]. In their work, a series of Pd/In2O3/SBA-15 catalysts prepared by the citric acid method was performed toward CO2 methanation. The presence of Pd species helps to facilitate H2 dissociation. On the other hand, the introduction of In2O3 induced CO2 activation, resulting in the promotion of high-efficiency conversion of CO2 to methanol with maximum methanol selectivity of 83.9%, CO2 conversion of 12.6% corresponding to a space-time-yield (STY) of 1.1\u00a0\u00d7\u00a010\u22122\u00a0mol\u00a0h\u22121\u00b7gcat\n\u22121 under reaction conditions of 260\u00a0\u00b0C, 5 Mpa and 15,000\u00a0cm3\u00a0h\u22121\u00b7gcat\n\u22121. In addition to an ordinary rod-like one, the fibrous type of SBA-15 was also reported as catalyst support for CO2 methanation [185]. The fibrous type of SBA-15 (F-SBA-15) was obtained by transforming the rod-like SBA-15. The obtained Ni/F-SBA-15 catalyst exhibited higher superior catalytic performance with CO2 conversion of 99.7%, and CH4 yield of 98.2% compared to those of rod-like Ni/SBA-15 with CO2 conversion of 91.1%, and CH4 yield of 87.5%. Also, Ni/F-SBA-15 exhibited higher catalytic stability and coke resistance than Ni/SBA-15. The distinctive SBA-15 morphology led to a higher homogeneity of finer Ni, which reinforced the Ni\u2013F-SBA-15 interaction and increased the amount of moderate basic sites.Photoreduction of CO2 to hydrocarbons is one of sustainable energy technology which not only mitigates emissions but also provides alternative fuels. However, one of the largest challenges is to increase the overall CO2 photo-conversion efficiency when water is used as the reducing reagent. The use of ordered mesoporous silica to construct the isolated photoactive centers in porous matrices or to provide the high surface area for nanoparticles photocatalyst loading is an effective strategy to develop a highly active photocatalyst [186]. It is clear that the activity and selectivity of photocatalysts for different products strongly depend on the chemical nature of the supports. In brief, highly dispersed titanium dioxide in mesoporous silica materials (KIT-6, FSM-16, SBA-15, and TUD-1) leads to relatively high yields of CH4 or/and CH3OH, which makes the OMS-supported photocatalysts a promising candidate for CO2 photoreduction.A number of works have been reported the utilization OMS as support for photocatalyst for CO2 reduction. Li et\u00a0al. [187] utilized mesoporous silica-supported Cu/TiO2 nanocomposites obtained through a one-pot sol\u2013gel method for the photoreduction reaction of CO2 (Fig.\u00a018\na and b). The catalyst test was conducted in a continuous-flow reactor using CO2 and water vapor as the reactants under the irradiation of a Xe lamp. The presence of high surface area OMS support (>300\u00a0m2/g) greatly improved CO2 photoreduction by improving TiO2 dispersion and increasing adsorption of CO2 and H2O on the catalyst (Fig.\u00a018c). The addition of Cu species significantly improved the overall CO2 conversion efficiency and the selectivity to CH4 compared to TiO2\u2013SiO2 catalysts without Cu whose CO is the primary product of CO2 reduction. The optimum production rates of CO and CH4 achieved by using 0.5%Cu/TiO2\u2013SiO2 as their produced CO and CH4 with production rate of 60 and 10\u00a0mol gcat\u22121\u00a0h\u22121, respectively; the peak quantum yield was calculated to be 1.41% (Fig.\u00a018d). Moreover, Sasirekha et\u00a0al. [188] and Yang et\u00a0al. [189] employed the metal-doped TiO2 with mesoporous silica for CO2 photoreduction. Ordered mesoporous silica leads to the enhancement of the reaction rate because of the highly dispersed photocatalyst over the support and the improvement CO2 and H2O adsorption on the support surface.The improved surface area and better dispersion of the active sites were also demonstrated for the introduction of SBA-15 support to Ce\u2013TiO2 photocatalyst for CO2 photoreduction reactions [190]. It shows that Ce\u2013TiO2 dispersion on the silica matrix (Ti:Si\u00a0=\u00a01:4) is responsible for the enhanced textural properties of the catalyst compared to pristine TiO2. In addition, the unique mesoporous structure was one of the contributing factors for highly localized CO2 concentration near the TiO2 surface, thus leading to higher photocatalytic activity. Moreover, Tasbihi et\u00a0al. [186] reported the utilization of Pt/TiO2\u2013COK-12 photocatalysts prepared by a deposition\u2013precipitation method for the photocatalytic reduction of CO2 under UV light in a continuous flow gas-phase photoreactor. Pt/TiO2 was used as a comparison. Carbon monoxide is the major product obtained over TiO2 photocatalyst regardless of the presence of COK-12 as mesoporous support, while the addition of an appropriate amount of Pt to the catalysts leads to CO2 reduction towards CH4 formation, with selectivity as high as 100% in optimum loading (Fig.\u00a019\na). Fig.\u00a019b exhibited the mechanism of CO2 photoreduction reaction over TiO2 and Pt/TiO2. Here, the total reduction of CO2 into CH4 was promoted by the strong chemisorption of CO over Pt active site. Also, the phenomena could prevent CO from being the main product as it is the case for bare TiO2, either unsupported or supported onto mesoporous silica. Moreover, introducing COK-12 support on Pt/TiO2 catalysts maintains the selectivity of the reaction towards CH4 and further improves the overall activity of the Pt/TiO2 materials, as observed in the photocatalytic tests.More recently, Fu et\u00a0al. [191] developed new spherical hybrid materials cobalt oxide-coated spherical mesoporous silica for visible-light-driven photocatalytic reduction of CO2 to CO (Fig.\u00a020\na-o). Among all the catalysts tested, CoO/s-SBA-15 exhibited the best performance with an average generation rate of CO of up to 25,626\u00a0mol\u00a0h\u22121\u00a0g\u22121 with a selectivity of 83.0% after 2.5\u00a0h illumination time under the white LED lamp (Fig.\u00a020n). It achieved maximum production of CO as high as 71.69\u00a0\u03bcmol and selectivity of 84.8% when the 3% wt. In comparison with Co3O4 catalysts, the intrinsic properties of CoO catalyst towards CO2 reduction to CO were also examined using DFT calculation (Fig.\u00a020o). The result exhibited that the conversion of OCOH\u2217, a chemical species generated after the transfer of hydrogen atom to O atom of elongated CO bond in CO2, to CO\u2217 and OH\u2217 using CoO catalyst exhibited a lower energy barrier of 1.51 compared to 3.76\u00a0eV for Co3O4. Moreover, the total energy change for Co3O4 was 1.52\u00a0eV larger than that of 0.24\u00a0eV for CoO. This result corroborated the result of the catalytic test, in which the reaction pathway is favored in the case of using CoO as the catalyst. In this work, the role of the support is also examined. This study shows that the presence of interaction between the Si\u2013OH functional groups of the support and reaction substrate molecules around the catalytically active center leads to the enhancement of CO2 reduction to CO.Metal\u2013organic frameworks (MOFs) are a class of nanoporous materials consisting of metal ions or clusters coordinated to organic ligands to form two- or three-dimensional crystal structures [192,193]. Analog to that of zeolite, the porosity of MOF is dictated by its secondary building units (SBUs) which are the basic structure of the metal ions or clusters. The SBUs are linked in an infinite lattice by the organic ligands, often referred to as linkers. Fig.\u00a021\na shows a schematic representation of the formation of MOF from its building blocks. By designing the SBUs and the linkers, the structure of MOF can be tuned easily. In fact, the designability of MOF has been emerging a new branch of chemistry, i.e., reticular chemistry [194]. Numerous researchers have been pursuing MOF development in the last decades, as shown by steadily increased publications on the topic. Furthermore, approximately 70,000 crystal structures of MOF have been stored in the Cambridge Structural Database [195]. Fig.\u00a021b shows the crystal structure of the commercially available MOF\u2014 HKUST-1.The wide variety of structures with tuneable physicochemical properties leads MOFs to be applicable in various fields. With a pre-designable porous size and shape, MOF is an impeccable choice of material for separations and selective catalysts. The surface area of MOF could be tailored up to >14,000\u00a0m2\u00a0g\u22121 theoretically [196], which is much higher than those of zeolites and active carbons. This feature may tackle the problems among gas and energy storage materials. The synergy between the metal ions and organic ligands is also a great interest in sensor development. In the field of bio-related materials, the potential of MOF as drug delivery [197] and a carrier of genetic material [198] also has been investigated with quite promising results. However, some may be concerning the toxicity of the metal moiety [199]. Furthermore, the current progress shows the development of MOF with anisotropic crystal structures, enabling it to possess sequences of unique chemical properties. A comprehensive review by Xu et\u00a0al. on anisotropic reticular chemistry is available elsewhere [200].In catalysis, MOF has gone through a continued enhancement since firstly studied in the 1990s [201]. Thermal and chemical stability were the main issues in developing MOF as a catalyst [202]. MOFs suffered a structural collapse at a high temperature and in the presence of certain functional groups, organic solvents, or moisture. Nevertheless, the development of MOF catalysts was continuously pursued due to their irreplaceable unique properties. The combination of designable pores and intrinsic Lewis acidity out of the metal ions is a promising feature for catalysts.Moreover, additional catalytic active sites could be generated as well by introducing metal nanoparticles (NPs) into the pores. The insights on the MOF/NPs catalysts have been reviewed by D. Astruc's group recently [43]. Nowadays, most MOF structures can endure severe conditions\u2014to certain extents\u2014during the catalytic process. In the following sub-sections, we will highlight the remarkable success of MOFs as catalysts, particularly for the generation of renewable fuels.In addition to MOF, the rapidly growing research on reticular chemistry also brought up a new type of material, namely covalent organic frameworks (COFs). COF is often considered the \u201ccousin\u201d of MOF due to its similarity in designability, porosity, and high surface area. Unlike MOFs, however, COFs are built up out of entirely organic molecules as the building blocks. Fig.\u00a021c and d shows the building up of the COF structure from its components and a COF structure with hexagonal pores, respectively. The topology of COF is governed by the origin of the building blocks and their arrangement [203]. For instance, a combination of building blocks with C3 symmetries resulting in a hexagonal skeleton, a combination of C2 and C4 symmetries giving in a tetragonal skeleton, a combination of C2 and C6 symmetries giving in a trigonal skeleton, and so on. Various structures and pores can be built up extensively by varying the symmetries, sequences, orientations, and length of the building block. Commonly, the skeletons of COFs are constructed in a two-dimensional lattice due to planar organic molecules. The 2D skeletons are then assembled in 3D construction material via \u03c0\u2013\u03c0 stacking [204].COFs are potentially applied in many areas such as energy storage, electronic devices, biomaterials, and catalysis [204]. In the field of electronic devices, COF showed a promising result and unique characteristics as a light-emitting diode (LED) [205] and organic semiconductor [206], thanks to their highly crystalline and porous structure. Note that conventional OLEDs and organic semiconductors are typically fabricated using semi-crystalline polymers. As a biomaterial, COFs are gaining more attention than MOFs due to the absence of metal moieties. Although toxicity is not always related to the presence of metals and some organic molecules that can be highly toxic due to the ability to penetrate cell membranes, COFs are considered as a potential candidate for drug delivery [207] and other bio-related materials [208]. In catalysis, COFs are commonly combined with metal nanoparticles (NPs) [209]. Pores in COFs are selective to some substrates because of the size and the nature of functional groups in the building blocks. In typical reactions requiring both acid and base catalytic sites, metal nanoparticles provide the Lewis acid sites, and the COF framework acts as the base sites. Therefore COF/NPs can be a powerful catalyst for the generation of renewable fuels.Even though there are many similarities between MOFs and COFs, the chemistry is quite different. MOFs are based on metal coordination chemistry which the type of metal orbitals and the type donor ligands are essential [193]. On the other hand, COFs are based on polymer chemistry, in which C\u2013C bonds and \u03c0\u2013\u03c0 stacking play a crucial role [204]. The differences certainly affect the techniques of synthesis and characterization of each material. Reviews on the synthesis techniques of MOFs [210] and COFs [211] have been available and are beyond the scope of this review. In the following sub-sections, the use of MOFs and COFs as catalysts in generating renewable fuels is elaborated alternately.The role of MOF and COF catalysts in the production of renewable fuels could be classified into the following routes: hydrogenation of CO2 or CO, hydrodeoxygenation of biomass, hydrogen evolution reaction (HER), and photocatalysis. The latter might also include hydrogenation and HER, but we would separate the discussion for photocatalysis since the mechanism is quite different. In summary, we have tabulated the role of MOF and COF on the production of renewable energy in Table 3\n.Production of hydrocarbons, alcohols, or carboxylic acids through hydrogenation of CO2 using MOF catalysts has been extensively studied. The high affinity of MOFs in capturing CO2 has been demonstrated in numerous studies [212\u2013214], thanks to the remarkably high surface area and porosity. In the successful hydrogenation of CO2, precious metals and/or other transition metals are typically assembled into the MOF catalysts. Xu et\u00a0al. [215] reported a selective conversion of CO2 to CH4 using Ru/Zr-MOF catalyst in a plasma-assisted system. Under the optimized condition and molar ratio of CO2:H2\u00a0=\u00a04:1, a yield of 39.1% with 94.6% selectivity to methane could be achieved. The MOF catalyst also showed good stability under a plasma environment, as indicated by preserved crystallinity and morphology after catalytic reactions. Zhao et\u00a0al. [216] developed another Zr-based MOF (UiO-66) embedded with Ni nanoparticles. The Ni@UiO-66 catalyst showed excellent activity with 57.6% conversion of CO2 and 100% selectivity toward methane. The absence of precious metals in this catalyst is certainly favored in the economic viewpoints. The catalyst was stable up until 100\u00a0h under a reaction temperature of 300\u00a0\u00b0C. Introducing Ni nanoparticles into another MOF, MIL-101, showed a remarkable performance in selective methanation of CO2 [217]. Zhen et\u00a0al. optimized the reaction conditions, studied the thermodynamic, and proposed a reaction mechanism in this work. The 20Ni@MIL-101 catalyst could achieve 100% CO2 methanation under a temperature of 300\u2013340\u00a0\u00b0C with 100% selectivity. According to the DFT calculation, the key to the outstanding performance is that the MIL-101 framework enables Ni particles to be exposed on the crystal plane of 111, which has a lower potential energy barrier (10\u00a0kcal/mol) for CO2 dissociation into COads and Oads compared to those of other planes. The potential of COF catalysts for selective CO2 reduction has also been studied in recent years. Wang et\u00a0al. [218] investigated the catalytic performance of triazine-based COF for selective methanation of CO2. Using an electroreduction system, a perfluorinated covalent triazine framework (FN-CTF-400) catalyst converted CO2 to CH4 with a faraday efficiency of 99.3%. The DFT study suggests that fluorine doping is essential to improve electrocatalytic efficiency since it regulates the activity of N and enhances the conductivity of CH4 production.To generate liquid fuels, CO2 can be directly converted into methanol. Yin et\u00a0al. [219] developed a PdZn alloy on ZnO (PZ8-400) catalyst out of direct pyrolysis of palladium at zeolitic imidazole framework-8 (Pd@ZIF-8) for methanol synthesis. The ZIF-8 structure enables the growth of Pd particles at the sub-nano level, which then facilitates the formation of small-sized PdZn particles well-dispersed on the surface of oxygen defect ZnO. The best conversion of CO2 to methanol over the PZ8-400 catalyst was 15.1%, with a selectivity of 54.2%. A fundamental study on the kinetic and mechanistic of CO2 to methanol conversion over Pt nanoparticles encapsulated in Zr-based MOF, Pt@UiO-67, was investigated by Gutter\u00f8d et\u00a0al. [220]. Based on the spectroscopy analysis and DFT modeling, it was revealed that methanol formation is taken place at the interface between Pt nanoparticles and defect Zr nodes through the formation of formate species on Zr nodes. Fig.\u00a022\na shows the 3D representation of the reaction pathways. Under total pressure of 8\u00a0bar and temperature of 170\u00a0\u00b0C, approximately 1% conversion of CO2 led to the formation of methanol (\u224820%) and methane (\u224880%). Conversion of CO2 to methanol over another Zr-based MOF, UiO-66, was studied by Stawowy et\u00a0al. [221]. After an exchange of less than 50% Zr content in the UiO-66 framework with Ce, the selectivity toward methanol production was enhanced from 3.5% to 28.7%. Further improvement of selectivity up to 59% was achieved by introducing Cu nanoparticles onto the UiO-66 (Ce/Zr) framework. Zeng et\u00a0al. [222] reported an interesting technique to activate the Cu center at Ru-UiO MOF for excellent and controllable catalytic performance. Using light as a trigger, Cu0@Ru-UiO\u2014which is selective (99%) to methanol production\u2014can be converted to CuI with 99% selectivity to ethanol production. Fig.\u00a022b shows the controllable catalytic selectivity of Cu@MOF catalysts. An alternative product of CO2 conversion into liquid fuels other than methanol is formic acid (HCOOH). Notwithstanding, in recent years, most of the study on MOF catalysts was still focusing on DFT calculation and simulation [223,224].Another route to produce renewable fuels is through Fischer\u2013Tropsch synthesis (conversion of syngas\u2014, i.e., CO and H2\u2014to hydrocarbons). Typically, MOFs were employed as the precursor material to develop highly dispersed fine metal nanoparticles such as Co, Fe, and Ni to catalyze Fischer\u2013Tropsch synthesis. Sun et\u00a0al. [225] developed a highly loaded Co on silica (Co@SiO2) using ZIF-67 as the hard template. The Co@SiO2 catalyst prepared using MOF precursor could achieve CO conversion of >15% with >90% selectivity toward C5+ products. Qiu et\u00a0al. [226] further investigate the use of MOF as a precursor by comparing nitrogen-rich (ZIF-67) and nitrogen-free MOF (Co-MOF-74) to prepare Co-based catalysts. The catalyst prepared using nitrogen-rich MOF (Co@NC) exhibited 10% CO conversion with 31% selectivity to C5+ products. On the other hand, the nitrogen-free MOF (Co@C) exhibited 30% CO conversion with 65% selectivity to C5+ products. Ping et\u00a0al. used Ni/UiO-66 to prepare the fine structure of Ni/ZrO catalyst through impregnation followed by calcination [227]. The Ni/ZrO catalyst was able to carry out selective production of CH4 from CO reactant less than 10\u00a0ppm, which was considered as a non-poisonous concentration of CO, with a selectivity of >50%. Recently, Panda et\u00a0al. [228] incorporated Pt to Co nanoparticles catalyst using MOF as the phase controller. The phase of Co nanoparticles (FCC or HCP) on the Pt@Co/C catalyst derived from Co2(bdc)2 (dabco) could be controlled by adjusting synthesis conditions. CO conversion through the catalyst was at around 35% with high selectivity (\u224870%) toward C1 product.One of the most effective ways to convert biomass into fuels is through a hydrodeoxygenation (HDO) reaction. Typical biomass such as lignin is a big organic polymer containing a high amount of oxygen. Thus, depolymerization through transfer hydrogenation and deoxygenation is required to decrease the oxygen content so that the fuel quality can be improved. HDO of biomass can generate hydrocarbons with comparable quality to that of fossil fuels. In recent years, MOF catalysts were employed to realize this reaction with high efficiency. Zang et\u00a0al. [229] developed palladium nanoparticles well dispersed onto amine-functionalized UiO-66 MOF catalyst (Pd@NH-UiO-66) for selective conversion of vanillin\u2014a typical model of lignin\u2014into 2-methoxy-4-methyl phenol. A 100% conversion of vanillin was achieved using the catalyst under mild conditions. The catalyst was stable, at least for the 6-cycle of stability test. The authors pointed out that the remarkable performance is attributed to the synergic of well-dispersed Pd nanoparticles and the presence of the amino group. Bakuru et\u00a0al. studied the same reaction\u2014selective conversion of vanillin into 2-methoxy-4-methyl phenol\u2014using a similar Pd@UiO-66(Hf) catalyst without amine functionalization (Fig.\u00a023\nai) [230]. Note that the UiO-66(Hf) contains \u03bc3-OH groups. The catalyst also showed excellent performance with vanillin conversion of >99% and >99% selectivity (Fig.\u00a023aii). The synergic between Pd nanoparticles and the Br\u00f8nsted acid sites are responsible for the remarkably good performance (Fig.\u00a023aiii). Phan et\u00a0al. [231] demonstrated a similar effect on the synergetic of metal nanoparticles and Br\u00f8nsted acid sites for converting fatty acid into heptadecane (Fig.\u00a023bi). Using a phosphoric acid-enhanced Pt-encapsulated MOF catalyst (Pt/P@MIL101-Cr), 95% of fatty acid conversion could be achieved with 75.5% selectivity (Fig.\u00a023bii). The potential of COF in HDO of biomass is, unfortunately, not widely explored yet. In COF, the Br\u00f8nsted acid could be designed as the intrinsic structure of the lattice instead of through post-functionalization. Thus, COF catalysts may show outstanding performance as well in HDO of biomass.As an alternative to synthetic hydrocarbons, hydrogen energy has gained much attention since it promises a cleaner combustion system. MOF-derived materials are recently extensively studied for preparing electrodes in an electrocatalytic reactor for hydrogen production through HER. MOF-derived materials can substitute noble metals as electrodes and generate unique porous structures that increase the efficiency of hydrogen adsorption during hydrogen evolution. Nivetha et\u00a0al. [232] prepared a Meso-Cu-BTC MOF catalyst to\u00a0perform hydrogen evolution in 1M NaOH solution. Compared to those of conventional Pt electrode (onset potential\u00a0=\u00a00.002\u00a0V, overpotential\u00a0=\u00a079.0\u00a0mV, Tafel slope\u00a0=\u00a033.41\u00a0mV dec\u22121), the electrocatalyst developed in this study showed a faster kinetic in producing hydrogen (onset potential\u00a0=\u00a00.025\u00a0V, overpotential\u00a0=\u00a089.32\u00a0mV, Tafel slope\u00a0=\u00a029.0\u00a0mV dec\u22121). Xu et\u00a0al. [233] developed another method by mixing Ni\u2013Co MOF, ammonium molybdate, and thiourea to prepare a Ni0.15Co0.85S2@MoS2 electrocatalyst via hydrothermal synthesis. Using 1M KOH as the reactant, ultrahigh HER was performed with an overpotential and Tafel slope of 79\u00a0mV and 52\u00a0mV dec\u22121, respectively. Incorporating Ni and Co from the MOF precursor successfully enhanced the electrocatalytic performance of MoS2\u2014a potential non-precious metal electrocatalyst that often-suffered low conductivity. Another recent study on MOF-modified molybdenum sulfide electrocatalyst for HER was reported by Do et\u00a0al. [234]. Using a facial solvothermal method, MoSx was anchored on the surface of rod-like Co-MOF-74 particles, resulting in MoSx/Co-MOF-74 composite as an electrocatalyst. The catalyst could carry out a high HER performance at the optimized conditions with a low onset potential of \u2212147 mV and a Tafel slope of 68\u00a0mV dec\u22121. An XPS analysis showed the formation of CoMoS species, which may correspond to the decreasing electron transfer resistance of Co-MOF-74.The potential of COFs and their derived materials as an electrocatalyst for HER also has been investigated in recent years. Siebels et\u00a0al. [235] studied the electrocatalytic performance of Rh nanoparticles supported on covalent triazine framework-1 (Rh@CTF-1) and compared it to commercial Pt/C (Fig.\u00a024\na). The Rh@CTF-1 electrocatalyst showed overpotential and onset potential of \u221258 mV and \u221231 mV, respectively, while the commercial Pt/C exhibited an overpotential of \u221277 mV with an onset potential of \u221238 mV. The results demonstrated the great potential of precious metals/COF electrocatalyst to be applied to HER. As an alternative to using precious metals, Qiao et\u00a0al. [236] developed CTF@MoS2 electrocatalysts to proceed HER. The catalyst demonstrated excellent performance with an overpotential of 93\u00a0mV and a Tafel slope of 43 mVdec\u22121. The study also suggests that the inherent \u03c0-conjugated crystal channels in the CTF support mass diffusion and electron transmission during the HER process. The further interesting potential of COFs was demonstrated by Zheng et\u00a0al. [237] as the CTF-based material could be electrocatalytic active without any metal content. A nitrogen-doped hollow carbon nanoflowers (N-HCNFs) were prepared using CTF and melamine-cyanuric acid (MCA) as the precursors (Fig.\u00a024b). In carrying out HER in acidic media, the N\u2013HCNF electrocatalyst exhibited an overpotential of 243\u00a0mV with a Tafel slope of 111 mVdec\u22121. Metal-free catalysts are certainly a great interest in the future.Besides solar cell technology, photocatalysis is the key to harvesting unlimited solar energy and turning it into renewable fuels. MOFs and COFs provide unique properties that potentially tackle the main issues in conventional photocatalysts, e.g., the bandgap is barely suitable for visible light and electron\u2013hole recombination problems. In MOF photocatalysts, the organic linker acts as the semiconductor absorbing photons and generating electrons (light harvester) [43]. Subsequently, the electrons can be transferred into the adjacent metal center through ligand to metal charge transfer (LMCT) and eventually shift to the doped metal NPs to carry out the desired catalysis reactions. Likewise, COFs act as organic semiconductors, and with the addition of metal nanoparticles, can be an efficient photocatalyst [204]. Such a combination of organic semiconductor and metal mimics the natural system of chlorophyll which arguably has the highest quantum efficiency under sunlight exposure.In the effort to generate renewable fuels, photocatalysts are mainly employed to carry out the production of H2 through water splitting or HER and the production of hydrocarbons through CO2 reduction. Bai et\u00a0al. [238] fabricated noble-metal-free g\u2013C3N4\u2013MIL-53(Fe) composite by a simple grinding method as a photocatalyst for H2 production from water splitting. Under simulated sunlight, hydrogen evolution was performed with a rate of 0.9054\u00a0mmol\u00a0g\u22121\u00a0h\u22121. The synergy between graphitic carbon nitride (g-C3N4) and MOF intimates interfacial contact and increases active sites on the photocatalyst, which can substitute the role of noble metals. Another noble-metal-free photocatalyst was developed by Tian et\u00a0al. [239] by introducing annealed Ti3C2Tx MXenes to Zr-MOFs (UiO-66-NH2) precursors via hydrothermal process to form Ti3C2/TiO2/UiO-66-NH2. The hydrogen production rate under simulated sunlight was 1980\u00a0\u03bcmol\u00a0h\u22121\u00a0g\u22121, significantly higher than its precursors. The synergistic effects of Schttoky junctions for Ti3C2/TiO2/UiO-66-NH2, Ti3C2/TiO2, and Ti3C2/UiO-66-NH2 interfaces was the key to the enhanced photocatalytic performance. A schematic illustration of the photocatalytic mechanism involving the charge transfer is shown in Fig.\u00a025\na. Wang et\u00a0al. [240] developed a different approach using a porphyrin-based MOF with the Ti-oxo cluster as the metal center. With the small addition of Pt (3\u00a0wt%) as the co-catalyst, the rate of H2 production could reach as high as 8.52\u00a0mmol\u00a0g\u22121\u00a0h\u22121 under visible light irradiation up to 700\u00a0nm. The outstanding photocatalytic performance was attributed to the feasible LMCT from the porphyrin (photon harvester) to the Ti-oxo, then to the Pt NPs. A schematic illustration of the proposed photocatalytic mechanism is shown in Fig.\u00a025b.The synergy of Pt co-catalyst, TiO2, and COF was investigated by Chen et\u00a0al. [241] using 2,2\u2032-bipyridine-5,5\u2032-diamine (Bp-COP). The light-harvesting properties of the COF led the catalyst suitable under visible light up to more than 600\u00a0nm. The presence of TiO2 as a charge transfer media from the Bp-COP to Pt NPs significantly enhanced the H2 production rate to 1333\u00a0\u03bcmol\u00a0h\u22121\u00a0g\u22121. Biswan et\u00a0al. [242] developed a noble-metal-free photocatalytic system by combining thiazolo [5,4-d]thiazole-linked COF (TpDTz) as the photon-harvester, Ni-thiolate cluster as the co-catalyst, and triethanolamine (TEoA) as the sacrificing agent. The H2 production rate over this photocatalyst reached 931\u00a0\u03bcmol\u00a0h\u22121\u00a0g\u22121 for more than 70\u00a0h. A kinetics study over the photocatalyst reveals that an outer-sphere electron transfer from the photo absorber to the catalyst is the rate-limiting step (Fig.\u00a025c). The further potential of COF photocatalyst was demonstrated by Li et\u00a0al. [243] in the production of H2 directly from seawater, which has been a challenging problem because it contains various salts. Using a thioether-functionalized covalent organic framework (TTR-COF), selective adsorption of Au over other metal ions was demonstrated; thus, Au@TTR-COF photocatalyst could endure decomposition under seawater. The H2 production of over the photocatalyst using pure water was 501\u00a0\u03bcmol\u00a0h\u22121\u00a0g\u22121 and only slightly decreased in the presence of dissolved salts. DFT calculation shows that the chelation of TTR-COF and the metal ion in seawater such as Mg2+ is energetically not favored (Fig.\u00a025d).Photocatalytic reduction of CO2 into fuels over MOF and COF catalysts was also continuously studied in recent years. Ma et\u00a0al. [244] developed a hierarchically porous TiO2/UiO-66 photocatalyst via a simple solvothermal and assembly method. The deposition of ultrafine TiO2 NPs on the surface of UiO-66 provides interlaced spacing owing to electrostatic repulsions. Conversion of CO2 to CH4 over the photocatalyst composite gave a production rate of 17.9\u00a0\u03bcmol\u00a0g\u22121\u00a0h\u22121 with a selectivity of 90.4%, even when the CO2 concentration was decreased to \u22642%. Wang et\u00a0al. performed conversion of CO2 to methanol over porous Cu\u2013Zn oxide derived from Cu/Zn-bimetal MOF [245]. Under simulated sunlight, fast production of methanol (3.71\u00a0mmol\u22121\u00a0g\u22121\u00a0h\u22121) over the photocatalyst could be achieved. The synergy between CuO and ZnO on a large specific surface area with the unique mesoporous structure of the catalyst dictated the photocatalytic activity. Chowdhury et\u00a0al. [246] designed a sheet-like nanoporous covalent organic framework (TFP-DM COF) catalyst to convert CO2 into HCOOH and HCHO. The rapid production rate for both HCOOH (19.2\u00a0mol\u00a0g\u22121\u00a0h\u22121) and HCHO (0.54\u00a0mol\u00a0g\u22121\u00a0h\u22121) was obtained over the photocatalyst under a white light LED irradiation. The photocatalyst showed good stability over a 5-cycle reusability test.During the past several years, nanoporous metals have also emerged as one of the most studied catalytic materials due to their many promising opportunities. This class of materials was reminiscent of the great success of Raney Nikel by Muray Raney in the early of 20th century. It was described as the spongy and porous state of highly active unsupported Ni materials, which could be easily produced by leaching the Si or Al from Ni\u2013Si or Ni\u2013Al alloys in aqueous NaOH. It has been widely used especially in the hydrogenation process. Also, it has been utilized as powerful catalyst for conversion of lignocellulosic biomass feedstock in recent years. Its emergence was significant for the development of nanoporous materials, in which people finally realized that active nanoporous metal catalysts could be prepared through a simple alloy corrosion method [247\u2013250].In general, the vast majority of the applications were emphasized where the exploitation of the nanoporous metal's unique porous structures and their large specific surface area is much needed [251]. In most cases, full control in structure and pore features would allow one to significantly enhance the performance of such materials in various applications, such as in catalysis [252\u2013254], actuation [255\u2013257], sensing devices [258\u2013260], energy storage (batteries) [261\u2013263], and many more. It is primarily due to the porous structures' unique ability to amplify the effectiveness of several local processes at the material's surface or its interface with bulk, such as mass transport, electric and thermal conductivity, light scattering, etc. Nevertheless, the conflicting requirement of material's porous structures for certain application types has been a significant challenge in nanoporous metals. For instance, in the application of nanoporous metals as sensing devices, fabrication of nanoporous metals with high surface area and thus having small pores are required to provide many surface-active areas. Simultaneously, large pores are also desired to facilitate efficient and fast mass and ion transport. Thereby, controlling the materials' structural hierarchy has been one of the most potential approaches for reconciling these conflict requirements between small and large pores.Recently, tremendous efforts have been made to develop strategies for efficiently fabricating nanoporous metals with the desired structural hierarchy. In general, three main synthetic methods (Fig.\u00a026\n) can be used to manufacture nanoporous metals with hierarchical and multimodal structures, i.e., (i) dealloying-based, (ii) templated-based, and (iii) assembly method [251,264]. In the dealloying method, porous structures of metals are generally obtained via corrosion processes. This synthetic method has been proven reliable for the fabrication of various types of nanoporous metals, such as single and binary or multi-component noble nanoporous metal alloys and 3D nanoporous structures. Depending upon its corrosion process, the dealloying method can be categorized into chemical and electrochemical dealloying. Typically, the fabrication of nanoporous metals via chemical dealloying can be carried out using either acid or base solutions, depending on the type of metals and the desired porous structures. For example, a well-defined crystal structure of nanoporous PdPt alloy with a typical pore size of about 5\u00a0nm was successfully prepared from the chemical corrosion of ternary alloy of PdPtAl in 0.5\u00a0M of NaOH [265]. Meanwhile, Xu and co-workers [266] have also reported that a very similar nanoporous PdPt alloy could also be obtained from chemical corrosion of Pd16Pt4Al80 in acid conditions using HCl or H2SO4 solution. Based on the result, it was found that chemical corrosion in an acid condition of the precursor was able to form a similar uniform sponge-like nanostructure with a bicontinuous 3D network structure. Fig.\u00a027\na shows the structure comparison of nanoporous PdPt alloy prepared from chemical corrosion using acid and base solutions.In literature, other noble mesoporous metals and/or alloys with unique 3D structures, such as PtAu, AuAg, and PdAg have also been successfully prepared using this chemical dealloying method using both acid and base solutions [267\u2013271]. However, recent efforts have been shifted towards the fabrication of nanoporous non-noble metals and alloys driven by the need for economic and natural abundance consideration. It is reported that low-cost and naturally abundant metals, such as Cu, Ni, or Ti, could efficiently be used to reduce the high loading of nanoporous Pt metal catalysts and significantly enhance their catalytic activities. For instance, Qiu and co-workers have successfully prepared various kinds of nanoporous PtCu alloys with different structures, including wire-like structure, core\u2013shell porous structure, and aligned 3D structure, using chemical dealloying method from PtCuAl alloy precursors [272\u2013274]. Furthermore, chemical dealloying at basic condition using NaOH have also been reported to be able to successfully prepare various kinds of low cost and earth-abundant nanoporous binary metal alloys, such as PdNi, PdFe, PdCu, PdCr, PdCe, and PdZr, from the corresponding PdNiAl, PdFeAl, PdCuAl, PdCrAl, PdCeAl, and PdZrAl alloy precursors, respectively [257,259,275\u2013278]. Recently, the fabrication of multicomponent quaternary metal alloys with 3D nanoporous structures has received plenty of attention due to their unique physicochemical properties, and they are proven to exhibit excellent performance in various applications. For example, nanoporous PdAuCu metals with unique well-aligned 3D bicontinuous ligaments and pores have successfully synthesized and showed excellent electrocatalytic activity towards oxygen oxidation reaction (ORR) [279].In the dealloying-based method, nanoporous metals can also be prepared via electrochemical dealloying. In several cases, the electrochemical method is preferred due to the ability to control the chemical composition of the resulting nanoporous metals, and it is considered cheaper and easier to perform [264]. According to literature, both structural and physicochemical properties of the resulting nanoporous metals prepared using electrochemical dealloying are highly dependent on two key factors, i.e. (i) the parting limit, which is the concentration of the alloy precursors; and (ii) critical potential, which refers to the applied potential threshold during the electrochemical desolation [280]. For instance, electrochemical dealloying has been successfully employed to synthesize nanoporous PdAu alloys from PdAuNi ternary precursors in a 0.5M H2SO4 solution [281]. Here, the structural morphology and Pd/Au ratio of the as-prepared nanoporous metals could easily be controlled by tuning the appropriate critical potentials. Fig.\u00a027b shows the micrographic images of the nanoporous PdAu alloys prepared at various Pd/Au ratios. A similar approach was also used to fabricate nanoporous PtAu alloys with open 3D nanoporous network structures [282]. Based on the report, the nanoporous was fabricated using electrochemical dealloying by selectively etching Cu from the Pt10Au10Cu80 ternary alloy precursor. In another report, Zhang and co-workers [283] have also successfully fabricated ferromagnetic nanoporous PtFe by dealloying an amorphous FePtB alloy precursor. The result revealed that the nanoporous PtFe was composed of a single face-centered cubic phase and exhibited excellent performance as an electrocatalyst for methanol oxidation.The second synthetic method that can be used to fabricate nanoporous metal is the templated-based approach. In this approach, the structure of nanoporous is typically obtained within several steps of synthetic pathways, such as (i) the preparation of the original template; (ii) impregnation of the metal precursor into template's void spaces; (iii) metallization via reduction of precursors by either chemical or electrochemical methods; (iv) crystallization at desired temperature; and (v) removal of the template. Depending on the template, the synthesis of nanoporous metals via template-based method can generally be divided into two types, i.e., hard template-based and soft template-based synthetic methods [284,285]. In general, rigid natural and artificial minerals or biological molecules are commonly used in the hard template-based method. For example, Qiu et\u00a0al. [286] have successfully used SiO2 nanospheres as the hard template to prepare hollow porous PdPt alloy nanospheres, proving to be efficient for methanol electro-oxidation (MOR). According to the result, it was revealed that the SiO2 template played a critical role in controlling PdPt alloys' thickness due to the presence of its strong electrostatic interaction with charged metal precursors. In another report, 3D nanoporous of PdNi alloys were also successfully prepared using nanoporous alumina as the hard template [287]. Meanwhile, Nguyen and co-workers [288] were also able to efficiently synthesize and control the porous structure of PdCo thin films using anodized aluminum oxide (AAO) as the template. Recently, Fang et\u00a0al. successfully used several types of mesoporous silica, such as KIT-6, SBA-15, and EP-FDU-12, as hard templates for synthesizing mesoporous noble metal networks using chemical reduction process [289]. Fig.\u00a028\na shows the micrographic images of several mesoporous noble nanoporous metal structures prepared using silica templates.On the other hand, the soft template can be in the form of both biological and artificial self-assembled structures such as surfactant micelles or reverse micelles, microemulsions, polymers, or gas bubbles [264]. For instance, a study reported by Kang and co-workers [290] revealed that mesoporous PtCu nanostructures were successfully synthesized using self-assembled block copolymer micelle as a soft template and ascorbic acid as a reducing agent. Based on the result, it is reported that the composition of the as-prepared mesoporous alloys could easily be controlled by tuning the ratio of metal precursors. Fig.\u00a028b presents the schematic illustration and micrographic images of mesoporous PtCu alloys prepared using diblock copolymer, i.e., poly (ethylene oxide)-b-poly (methyl methacrylate) (PEO-b-PMMA) as the soft template. Furthermore, a honeycomb-like AuPt nanoporous alloy structure was also successfully prepared via electroreduction process in the presence of in-situ hydrogen gas bubbles as a soft template [291]. According to the result, the as-prepared AuPt alloys exhibited excellent sensitivity and selectivity for non-enzymatic glucose sensing applications due to their large surface area and the homogenous spread of Au and Pt throughout the surface.Finally, nanoporous metals could also be prepared using the assembly method. However, this method is less common than the previously discussed dealloying- and templated-based method. It is primarily because the metal precursor's assembly process is considerably hard to control and susceptible to various external factors, such as temperature, solvent, and the type and concentration of metal precursors. In this method, the porous structure is typically growing randomly during the assembly process, leading to less-ordered and low-intensity nanostructure formation. Nevertheless, several types of nanoporous metal alloys have been successfully prepared using this method. In literature, solvothermal and/or hydrothermal processes are considered one of the most common techniques in facilitating the precursors' self-assembly to form nanomaterials with nanoporous features. For instance, Li and co-workers [292] have successfully developed a simple one-pot solvothermal method to prepare porous PdCu alloys with a unique nano frame structure. Their report revealed that such a unique nano frame structure was achieved by simultaneous co-reduction of Pd and Cu precursors in the presence of oleylamine and NH3. In another report, a series of nanoporous Pd-M alloys (PdCd, PdPb, PdIr, and PdPt) have been successfully prepared using a robust and straightforward hydrothermal method [293]. According to the report, the as-prepared nanoporous alloys showed excellent performance as electrocatalysts for formic acid oxidation due to their small dendritic and random array structures.Nanoporous metals have been widely considered one of the most efficient catalysts for generating renewable fuels, especially in designing high-performance fuel cells. This is not only due to their exceptional catalytic activity as the consequences of the large surface area resulted from their unique porous structures but also their high electric conductivity, thermal and chemical stability, and excellent optical properties. One of the most common applications of nanoporous metals-based catalysts is in the non-spontaneous water-splitting reaction. In general, nanoporous metals are introduced to suppress the high overpotential required to make the reaction occur. Additionally, nanoporous metals could also be used to overcome one of the major fundamental challenges in water-splitting reaction: the low electrical conductivity of conventional electrocatalyst due to the resistance overpotential. For example, Lei and co-workers [46] have successfully prepared, and utilized nanoporous Ni\u2013Fe hydroxyl phosphate (NiFe-OH-PO4) supported on Ni foams' surface as a bifunctional electrocatalyst for whole-cell water electrolysis in alkaline solution. Based on the result, it was found that the system was able to generate a current density of 20 and 800\u00a0mA/cm2 at oxygen evolution overpotential of 240 and 326\u00a0mV, respectively, and a current density of 20 and 300\u00a0mA/cm2 at hydrogen evolution overpotential of 135 and 208\u00a0mV, respectively. Interestingly, the as-prepared electrocatalyst was also found to exhibit exceptional prolonged stability under continuous and intermittent electrolysis reactions.In another report, Detsi and co-workers [294] have also prepared ultrafine nanoporous NiFeMn alloys using the dealloying method and used them as electrocatalysts in water splitting reaction. The report revealed that such material could facilitate efficient water oxidation with a current density of 500\u00a0mA/cm2 at 360\u00a0mV overpotential in 1\u00a0M KOH solution. It is believed that such excellent catalytic activity was primarily due to the small size of the ligaments and pores of the nanoporous, which was proven by the high BET surface area of 43\u00a0m2/g. Additionally, the as-prepared nanoporous NiFeMn alloys' high electrical conductivity was also responsible for effective current flow. Another strategy was also developed by Dong et\u00a0al. [295] by selective dealloying of Al97\u00acNixFe3-x to form high-performance porous NiFe nanowire network alloy as water splitting electrocatalyst. Based on the result, it was found that the as-prepared electrocatalyst exhibited excellent performance to oxidize water with only \u223c244\u00a0mV overpotential in 1\u00a0M of KOH solution. Further suppression in water oxidation overpotential was also achieved by partial dealloying removal of Cu from Cu-rich NiFeCu ternary alloys [296]. According to the report, the as-prepared porous electrocatalyst required only about \u223c180\u00a0mA/cm2 for water oxidation in an alkaline solution.Aside from water oxidation reaction, nanoporous metal's performance as a catalyst has also been widely investigated for the generation of renewable fuels from other types of sources, such as methanol, ethanol, and formic acid. For example, Zhang and co-workers [297] reported that nanoporous PdPt alloys were able to exhibit exceptional performance in facilitating methanol oxidation reaction (MOR) in a direct methanol fuel cell (DMFC) system. According to the report, the high stability and catalytic activity of the as-prepared nanoporous PdPt alloys were primarily caused by the high density of twinned and ultrathin ligaments, which creates large curvatures between concave and convex region as well as low as forming low-coordination surface atomic steps and kinks. Consequently, these features render the appearance of many active low-coordination atoms sites essential for catalytic activity. A similar phenomenon was reported elsewhere when dendrite-like nanoporous PtAu was used as the catalyst [298]. Here, the catalyst was prepared using a soft-template method where l-histidine and PVP were used as the template and as structure-directing and dispersing agents. Based on the result, it was reported that the as-prepared dendrite-like nanoporous PtAu alloys showed a superior mass activity (MA) and specific activity (SA) in MOR than that of Pt black. Fig.\u00a029\n presents the micrographic images and catalytic performance of the as-prepared dendrite-like nanoporous PtAu alloys in MOR.Furthermore, various types of palladium-based of ternary nanoporous alloys have also been developed and used in MOR. It is reported that introducing a different metal into binary alloy would allow more abilities to tailor the overall catalytic properties of the material. For example, Li and co-workers have successfully prepared tri-metallic mesoporous PdPtAu alloys, demonstrating excellent MOR performance [299]. According to the report, the electrocatalytic system efficiently facilitated MOR with mass activity (MA) of 1010\u00a0mA/mg. It was believed that such excellent catalytic activity was primarily caused by the synergistic effect between the mesoporous structure and the ternary metal alloy. Recently, tremendous efforts have been made to develop non-noble metal-based nanoporous alloys due to their low-cost, non-toxicity, and high abundance properties with comparable catalytic activity with that of noble metal-based alloys. In literature, various noble/non-noble nanoporous metal alloys, such as PtCu, PdCu, PdNiO, PdZr, PtCo, PdCo, PtFe, etc., have been successfully prepared and could potentially be used in MOR. Table 4\n summarizes the performances of several types of nanoporous metals-based electrocatalyst in MOR.Nanoporous metals and alloys have also been widely investigated for ethanol oxidation reaction (EOR) in direct alcohol fuel cells (DACFs). In general, ethanol is preferred in renewable energy generation over methanol due to its higher energy density, lower toxicity, less volatility, and easier storage and transport [308]. Moreover, ethanol can also be produced from renewable sources, such as biomass like cellulose or starch. Nevertheless, most conventional electrocatalysts, such as bulk and nanosized mono-noble metal catalysts (Pd, Pt, and Au), are easily poisoned by the adsorbed CO during the electrocatalysis, leading to the significant reduction in electric conductivity and catalytic activity [264]. Therefore, the unique porous structure of nanoporous metals and alloys has been investigated and considered as the potential solution for such issues. For example, Chen et\u00a0al. [281] have successfully prepared a typical nanoporous bimetallic PdAu alloy with a tunable metal ratio for EOR. Here, the nanoporous PdAu at various PdAu ratios were synthesized by electrochemically dealloying ternary metal alloy precursor of PdAuNi, which led to geometrically formation controllable nanoporous structure. As a result, the resulting catalyst exhibited superior catalytic performance in EOR compared to the conventional Pt/C, Pd nanoparticles, and non-porous PdAu nanoparticle alloy.In another report, mesoporous PdPt alloys have also been efficiently utilized as the electrocatalyst in EOR [306]. The report revealed that the formation of unique porous and hollow structures of the materials was obtained due to the presence of halide ions (Br- and I- ions) during the facile one-pot hydrothermal process. Furthermore, the result also demonstrated that the as-prepared mesoporous PdPt alloys showed much higher specific and mass activities than the commercial Pt black and Pt/C electrode. A similar excellent catalytic performance in EOR was also observed when nanoporous bimetallic PdAg alloy was used as the electrocatalyst [269]. This exceptional catalytic activity was believed to be originated from the synergistic effect of bimetallic alloy and the unique porous structure of the material, which was obtained by dealloying melt-spun AlPdAg ribbon in 10wt% of H3PO4 solution containing polyvinylpyrrolidone (PVP). Table 5\n lists the comparison of different types of nanoporous metal and alloy-based electrocatalysts in facilitating EOR.Another potential application of nanoporous metal-based catalysts in renewable fuels is the formic acid oxidation (FAO) reaction in direct formic acid fuel cells (DFAFC). Recently, renewable fuel generation via the DFAFC process has gained much attention due to its high-power output and proton exchange's low membrane permeation rate [311,312]. In DFAFC, the FAO process can typically proceed via either a direct or indirect pathway. In the earlier pathway, CO2 is formed directly by the hydrogenation reaction of formic acid. Meanwhile, the latter pathway suggests that the CO2 product is obtained via the formation of intermediate CO molecules due to the dehydration reaction of formic acid. Therefore, there is a chance for such intermediate molecules to poison and inhibit the active site of electrocatalyst, which was also often observed in methanol and ethanol oxidation. During the past several years, various collections of nanoporous metal-based catalysts have been employed as anode materials to efficiently facilitate the FAO process due to their ability to prevent CO poisoning during electro-oxidation. For example, Xu and co-workers [266] reported that nanoporous PdPt alloy with uniform pore size prepared by chemical dealloying could efficiently facilitate FAO reaction. Based on the result, it was found that the as-prepared nanoporous PdPt demonstrated superior catalytic activity than the conventional Pd/C electrode and the corresponding mono-metal analogs to facilitate not only the oxidation of formic acid but also methanol and ethanol. Fig.\u00a030\na and b presents the as-prepared nanoporous PdPt alloy's micrographic images and its electrocatalytic performance in FAO reaction (Fig.\u00a030c). Assaud and co-workers also observed a similar result when the 3D-nanoarchitecture PdNi alloy catalyst was employed as the anode material [287]. According to the result obtained from cyclic voltammetry, it was found that the oxidation of formic acid proceeded via a direct pathway instead of an indirect route. Recently, other types of nanoporous metal alloys, such as PtGa, PdNi, PdCu, PdAg, AuPt, and PdM (M: Pb, Cd, and Ir), have also been proven to exhibit excellent catalytic activity in FAO reaction [270,293,313\u2013316].Nanoporous metals and alloys have also been investigated for their application to generate renewable fuel via carbon dioxide (CO2) reduction. During the past several years, this research direction has been attracting much attention due to the ability to simultaneously generate clean energy and reduce the effect of global warming due to the large accumulation of anthropogenic CO2 emissions [53]. In general, nanoporous metals and alloys were employed to facilitate the catalytic conversion of CO2 into several types of reusable carbon fuels such as CO, CH3OH, C2H5OH, C2H5, or CH4. For example, Selective and efficient conversion of CO2 to CO has been reported by Lu and co-workers [317] using nanoporous gold prepared by electrochemical dealloying. Based on the result, it is reported that the porous structures of the metal were responsible for the formation of many high-density steps or kink sites, exposing the essential high-index facets inside the curve of the metal's structure (Fig.\u00a031\na). As a result, the as-prepared nanoporous gold exhibited excellent catalytic performance in CO2 to CO conversion with Faradaic efficiency (FE) of 98% at an overpotential of 390\u00a0mV (Fig.\u00a031b). Furthermore, the result of the long-term stability test exhibited that the reduction of surface step/kink sites and the deposition of metal impurities are responsible for catalysis decay. However, by applying potential cycling, the catalytic performance of the deactivated electrode can be recovered. The reactivation of catalysts was caused due to the rejuvenation of the reduced step/kink sites and the removal of surface metal impurities (Fig.\u00a031c).Meanwhile, Hong et\u00a0al. [318] have also reported that nanoporous copper could be efficiently used to convert CO2 to various types of renewable fuels. Here, the unique porous structure of Cu film was obtained by highly-controlled electrodeposition using 3,5-diamino-1,2,4-triazole (DAT), which was responsible for directing the crystal's growth with high exposure to catalytic sites. According to the report, the as-prepared electrocatalyst exhibited an auspicious catalytic performance for electroreduction of CO2 to C2H5 and C2H5OH with FE of 40% and 20% at \u22120.5\u00a0V vs. RHE, respectively. Furthermore, it was also found that the overall mass activity for the CO2 reduction was \u223c700 A/g at \u22120.7\u00a0V vs. RHE.Furthermore, selective electroreduction of CO2 to C2H4 was also carried out using nanoporous copper film electrodes [319]. In this report, Peng and co-workers [319] used the chemical dealloying technique to fabricate the nanoporous Cu surface structure from the Cu\u2013Zn surface alloy obtained from electrodeposition and thermal treatment processes (Fig.\u00a032\na). Based on the result, the as-prepared electrocatalysts were found to suppress the FE of CO2 conversion to methane down to 1% while keeping selective reduction of CO2 to C2H4 with FE of 35% in an aqueous solution of 0.1\u00a0M KHCO3 at \u22121.3\u00a0V vs. RHE (Fig.\u00a032b). The high selectivity of C2H4 products could be attributed to several synergic factors, including the exposed (100) facets, along the possible existence of step and edge atoms (Fig.\u00a032c).In another report, the influence of ligament size of nanoporous Ag network was also evaluated for the application in CO2 reduction [320]. Here, two nanoporous Ag networks with the average ligament sizes of 21\u00a0nm and 87\u00a0nm were fabricated by dealloying binary Mg80A20 alloy ribbon in the presence of citric acid and phosphoric acid (Fig.\u00a033\na). According to the report, it was found that the ligament size of the as-prepared Ag network was significantly influencing the catalytic performance of the catalyst in facilitating CO2 conversion to CO (Fig.\u00a033b). Results demonstrated that Ag network with smaller ligament exhibited a superior catalytic performance (FE of 85% at \u22120.8\u00a0V vs. RHE) than that of larger ligament size (FE of 41.2% at \u22120.8\u00a0V vs. RHE) (Fig.\u00a033c). Recently, Lu et\u00a0al. [321] have also successfully prepared and used nanoporous AuSn alloy to enhance and selective electroreduction of CO2 to CO. Compared to the analogous bulk nanoporous Au, the as-prepared nanoporous AuSn alloy exhibited a significantly higher catalytic activity for CO2 conversion with FE of 92% at \u22120.85\u00a0V vs. RHE. It is believed that such exceptional activity was primarily due to the presence of a trace amount of Sn solute in Au lattice, which results in a more pronounced tensile strain on the surface of 3D nanoporous. As a result, this would allow the metal surface's d-band center to be shifted and ultimately lead to the enhancement for the adsorption of key intermediate species of \u2217COOH during the electroreduction process.Renewable fuels are urgently needed as alternative energy sources to the increasingly depleted fossil-based resources. These fuels are expected to tackle the depletion issue and realize an environmentally benign energy cycle. In this sense, a catalyst is inevitable to assist the conversion of renewable sources into fuels. Nanoporous materials have served as efficient heterogeneous catalysts that possess high catalytic activity and control product selectivity. This review summarizes recent advances in nanoporous materials, i.e., zeolites, ordered mesoporous silica (OMS), metal- and covalent organic frameworks (MOFs and COFs), and nanoporous metals, and their use in diverse chemical processes as a catalyst for producing renewable fuels.Zeolites have exhibited superior catalysis performance in producing renewable fuels through fast catalytic pyrolysis (CFP) and CO2 conversion. In the former case, zeolites simultaneously reduce undesirable products and yield organic liquid products at an acceptable amount. Zeolites also increase aromatic hydrocarbons and light phenols while decreasing the bio-oils viscosity, density, and acid number. On the other hand, zeolites are up-and-coming candidates in CO2 conversion into renewable fuels, which exhibited remarkable performance when combined with CO2 conversion into renewable fuels, which were remarkably successful when combined with a metallic catalyst. In both applications, several parameters need to be considered. The first is the type of zeolite frameworks. Note that each framework possesses a unique pore size and shape, which could govern the product selectivity. The second is the framework silica-to-alumina ratio (SAR) of zeolites, an essential factor since it determines the number of acid sites within the zeolite frameworks. The third is the incorporation of metallic sites. In the catalytic fast pyrolysis, introducing transition metals such as Pb, Ni, Zn, Fe, Mo, Ga, and Co into zeolite framework has improved the bio-oil yield and lowered the content of undesired polyaromatic hydrocarbons and cokes. Furthermore, metallic (e.g., Pd, Pt, and Ni) sites are indispensable as hydrogenation sites for converting CO2 into various renewable fuels. In addition, the use of bimetallic or trimetallic sites is intriguing owing to the synergistic effect, which might increase the overall catalytic performance.The performance of zeolite catalysts can be enhanced by introducing larger porosity (mesopores and/or macropores), so-called hierarchical porosity. Nevertheless, it should be emphasized that the larger pores should be interconnected with the micropores. The inlet flow of reactants must first go through the larger pores prior to the micropores. These requirements should be met to maximize the utilization of acid sites and overcome the diffusion issue within purely microporous zeolites. Moreover, metal size, shape, and facet engineering should be pursued since many molecules exhibit preferences over particular crystal facets. The ability to control the physicochemical properties of zeolites and metallic sites will enable the highly active, selective, and efficient catalyst system to produce desired products.OMS has demonstrated unmistakable performance in the hydrogen production reaction and the conversion of carbon dioxide to fuels and chemicals, respectively. The well-defined mesoporosity of this material has played a major role in improving the catalyst's performance in both types of reactions, where the pores in the OMS play a pivotal role in providing an active site for the reactants during the reaction. In addition, the pores in OMS can increase the dispersion of metal loading, thereby reducing the presence of catalyst particle agglomeration, which can cause pore blocking. In addition, a larger mesopore promotes the formation of metals with larger sizes, resulting in weaker interaction of metal\u2013support which is more favorable for both reactions.Several types of OMSs reported as catalysts in the hydrogen production and carbon dioxide conversion reactions include MCM-41, MCM-48, SBA-15, SBA-16. MCM-41 is the most widely used type. The use of the MCM-41 as support is owing to its interpenetrating 3D pores and high stability. MCM-48 is claimed to be a better candidate than MCM-41. This OMS is characterized by the interwoven and continuous three-dimension pore system. Mesoporous silica SBA-15 has a hexagonal structure (p6mm) similar to the MCM-41, but SBA-15 presents greater hydrothermal stability due to the thicker silica walls and has larger pores when compared to the MCM-41 mesostructure. In addition to SBA-15, which has a 2D hexagonal porosity, SBA-16 possesses cage-like mesopores organized in a three-dimensional cubic body-centered Im3m symmetry. The structure of SBA-16 can be described by a triply periodic minimal surface of I-WP (body-centered, wrapped package). The mesophase might also be a triply periodic minimal surface. In comparison with SBA-15, the synthesis of the cubic SBA-16 material is more challenging. This factor makes the use of SBA-16 support relatively rare.OMS materials are generally used to support several types of metal or metal oxide catalysts. The addition of these catalysts can improve the performance of CSO support. Several types of catalysts supported on the OMS surface include metal, both monometallic and bimetallic, such as Ni, Fe, Mg, Cu, Pd, Pt, Ni\u2013Pd; metal oxides such as metal oxides as CeO2, CdS, CuO, ZnO, ZrO2, NiCr2O4, Fe2O3, Al2O3. The results show that the introduction of metal oxide in the catalyst changes the pore size and specific surface area of the support. Among several types of metal and metal oxide catalysts added to the catalyst's surface, based on the literature that has been described, the addition of Ni is the most widely used because of the significant impact on improving the catalyst's performance. The incorporation of Ni could enhance catalyst coke resistance. However, adding Ni with high concentrations can also cause a decrease in the catalyst's performance. Therefore, attention to proper concentration is necessary.Furthermore, several things should be noted and worthy of being developed in the search for a reaction catalyst for hydrogen production and CO2 conversion, as follows. (i) The addition of a catalyst to the promoter. The addition of catalyst promoters could increase the metal-support interactions, decrease metal and/or metal oxide particles, thus improving the catalytic performance. In addition, the addition of a promoter can also reduce coke formation. The type of promoter used can be derived from metal particles such as Ga, Gd, Ce; organic compounds, such as alcohol; or metal oxide compounds such as Yttria (Y2O3). (ii) The incorporation of the acid or basic sites. The addition of the acid sites can increase the incorporation of metal catalysts on the support. Meanwhile, the addition of the base, such as magnesium atoms, could induce the adsorption and activation of CO2 in the CO2 conversion reaction. (iii) The morphology of the catalyst used. The different types of morphology in OMS support can determine the performance of the OMS catalyst. As a comparison of the performance of rod-like SBA-15 and fibrous type-SBA-15. The fibrous-type SBA-15 has higher catalyst performance than a rod-like one. Also, fibrous-type SBA-15 exhibited higher catalytic stability and coke resistance. In addition, the fibrous-type morphology led to higher homogeneity of a finer metal catalyst, which subsequently reinforced the metal-SBA-15 interaction and increased the amount of moderate basic sites.Despite the excellent catalytic activities for renewable energy production, some might still be concerning the stability of MOF and COF catalysts since the main frameworks are organic compounds. Compared to the other nanoporous materials, e.g., zeolites and porous metals, MOFs' thermal and mechanical stability are generally somewhat lower. Likewise, COFs with their fully organic contents need significant enhancement in thermal stability. Nevertheless, the potential of COFs in photocatalytic reactions is remarkably good. Manipulating the band gab in COFs (and MOFs) with their designability feature is certainly more feasible than inorganic crystals. The generation of renewable energy, either hydrogen or synthetic hydrocarbons, using photocatalysts suitable for visible light, would be a breakthrough to achieve a clean and renewable fuels supply.From the perspective of material science, we should note that the number of MOF and COF materials is extremely big, increasing in the coming years. To deal with such types of big data, incorporating machine learning can be useful. Quantitatively, machine learning with appropriate algorithms can reveal the structure\u2013properties relation of MOFs and COFs. However, the number of studies incorporating a machine learning approach for reticular chemistry is still limited. Moghadam et\u00a0al., for instance, reported the structure\u2013mechanical stability relationships for 3385 MOF materials with 41 distinct topologies using a combination of molecular mechanics calculations, machine learning, and molecular dynamics simulations [322]. Once computational chemistry is combined with a machine learning algorithm, the synthesis of new MOF and COF structures can be completely guided by the insight of numerous previous data concerning the designated applications.Another potential issue with MOF and COF catalysts is, in general, their reliance on the use of precious metals as active catalytic sites. Although there has been a continuous effort in recovering precious metals out of the end-of-life products [323,324], the use of common metals or even metal-free catalysts would be favored in the economic and environmental viewpoints. The concern of metal scarcity should be taken into account in mining sectors and in fine chemicals industries, including catalysts. In recent years, the trend shows the concern for developing the precious metal-free catalyst [325]. Continuing this path would be of great interest in the realization of sustainable development goals.Moreover, it is no secret that the current trend on utilizing nanoporous metals holds great potential for generating renewable fuels. Additionally, the ability to fabricate earth-abundant non-noble nanoporous metals and metal alloys-based catalysts has also enabled researchers to expand the utilization of such materials for large-scale applications. This would significantly reduce production and waste handling costs, making such technology one of the most attractive options for replacing fossil fuel-based energy generation. However, there are still plenty of challenges that need to be addressed regarding applying nanoporous metals and alloys in the generation of renewable fuels. For instance, albeit the ultra-small porous structure is essential for high surface area in the water oxidation reaction, the resulting O2 gas is often trapped inside the nanoporous bulk structure and could not easily escape from the electrode.Consequently, this O2 gas build-up would cause an increase in the system's internal resistance, which ultimately leads to the rise in the reaction overpotential. A similar phenomenon is often observed in other renewable energy generation reactions, such as methanol/ethanol/formic acid oxidations and CO2 reduction. One of the most promising solutions for such an issue is fabricating the catalysts with bimodal hierarchical porosity where macroporous and meso-/nano-porous features coexist. In such hierarchical-based material, it is believed that the meso-/nano-porous feature would still provide the large surface area needed for the reaction. At the same time, macropores would allow the escape of the reaction products and prevent the increased internal resistance.Furthermore, large-scale industrial application of nanoporous metal for renewable energy is also hindered by the fabrication method of the material itself. As previously discussed, most nanoporous metal-based catalysts that have been proven to exhibit exceptional ability in facilitating the generation of renewable fuels such as water/methanol/ethanol/formic acid oxidation and CO2 reduction are mostly prepared via the dealloying method. It is reported that the utilization of certain types of expensive binary and ternary metal alloys as a precursor is often considered to be one of the significant drawbacks of the dealloying technique. Besides, it is also known that the most convenient way to employ nanoporous metals as a catalyst in renewable fuels generation is to have them supported on conductive substrates such as carbon-based support.\nTable 6\n summarizes all the advantages, drawbacks, and future potential developments of the nanocatalysts discussed in this article. It is, thus, clear that although possessing several superiorities owing to physicochemical properties, however, based on the information on the table above, significant effort must still be devoted in the upcoming years to rationally design a better candidate of nanoporous catalyst for renewable fuel production. Ultimately, a further extensive investigation should be carried out to realize the nanoporous catalyst with high activity, low cost, and applicable for the fabrication on a large scale. Here, the role of computer-assisted catalytic experimentations should also be considered. In addition to the well-established methods, such as DFT and molecular dynamics, which has been discussed above in several catalytic reactions [122,125,191,220,223,224,243], the application of underdeveloping methods such as machine learning (ML) might also be put on the table in order to accelerate new heterogeneous catalyst discovery for renewable energy [326,327]. Several works have been devoted in order to employ machine learning (ML) for enabling catalyst discovery even also predicting its catalytic behavior [328\u2013330]. Thus, it does not only accelerate the discovery of novel catalysts but also provides a powerful tool to explore a deeper understanding of relationships between the properties of catalyst materials and their catalytic performance, i.e., activities, selectivities, and stabilities. Therefore by this knowledge, we could better design the catalysts and enhance their efficiencies [331]. Furthermore, apart from the catalyst perspective, a more comprehend the fundamental reaction mechanisms need to be also developed as a strong foundation to improve the catalytic performance. Also, understanding surface chemistry at the molecular level is crucial for designing a catalyst system with the desired nanoarchitecture that is expected to exhibit high activity and selectivity.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work is supported by Hibah P2MI (Research, community services, and innovation) \u2013 Institut Teknologi Bandung 2021, and the Ministry of Research and Technology (Kemenristek)/National Agency for Research and Innovation (BRIN) of the Republic of Indonesia, and Indonesia Endowment Fund for Education (LPDP) under the Program of Prioritas Riset Nasional No. 78/E1/PRN/2020.", "descript": "\n                  The rapid and continuous depletion of fossil-based resources has boosted extensive research on alternative energy from renewable resources. In this sense, heterogeneous catalysts play an inevitable role in converting renewable resources into fuels. The performance of heterogeneous catalysts strictly depends on their structures and physicochemical properties. As a rule of thumb, heterogeneous catalysts with large specific surface areas possess more catalytic sites to enhance the overall catalytic performance. Nanoporous materials have emerged as highly active heterogeneous catalysts due to their large internal surface area, enabling a high density of active catalytic sites. The presence of nanopores also allows the selectivity towards the desired products. Herein, we provide a comprehensive review of recent advances of several typical nanoporous catalysts, i.e., zeolites, ordered mesoporous silica (OMS), metal- and covalent organic frameworks (MOFs and COFs), and nanoporous metals. Each nanoporous catalyst's characteristics and synthesis strategies are elaborated in detail, followed by discussions on their applications in various chemical processes to produce renewable fuels. Finally, challenges and opportunities for future improvement are provided.\n               "}
{"full_text": "Solid propellant is widely used in aerospace and military fields, one of the most effective methods to improve the combustion performance of solid propellant is to advance the thermal decomposition process of energetic components (such as ammonium perchlorate (AP) and 1,3,5-trinitroperhydro-1,3,5-triazine (RDX)) by adding a few combustion catalysts (3%\u20135%) [1\u20134]. Nowadays, several combustion catalysts, such as metals [5\u20138], metal oxides [9\u201313], bimetals [14,15], and GO-based composition [3,16,17], have been proved to present compelling advantage in reducing the pressure exponent and forming platform combustion effect.The copper ferrite (CuFe2O4) nanoparticle, as a typical magnetic semiconductor with narrow band gap, has been extensively applied as gas-sensing materials [18], supercapacitor electrodes [19], and efficient catalysts [20\u201323] owing to its magnetic properties, excellent electronic conductivity and high thermal stability. Therefore, its application in solid propellent has been concerned as well. Zhang et al. prepared hollow CuFe2O4 nanosphere by hydrothermal method and confirmed its higher catalytic activity on RDX and FOX-7 than that of CuO and Fe2O3 [24]; Liu et al. further introduced graphene oxide (GO) to synthesize CuFe2O4/GO composite by self-assembly method and demonstrated that it could reduce the apparent activation energy of RDX by 98.71\u202fkJ/mol [25]. However, CuFe2O4 particles are inevitably to aggregate into clusters because of its nanometers size effect, which is detrimental for combustion of solid propellant [26,27]. The most common strategy for inhibiting aggregation and improving the dispersion of CuFe2O4 nanoparticles is to introduce appropriate nanoparticles carriers. Besides, pursuing the carrier that could display certain catalytic synergy properties with CuFe2O4, improve the combustion performance and anti-electrostatic discharge sensitivity of energetic components are more preferably.Many researchers have devoted themselves in seeking for layer carbon materials with two-dimension (such as GO) as right catalyst carrier, whereas the high-price, complex processing technology and waste acid contamination limit its engineering application [2,3]. But for silicon-based composites, holding abundant resources, low prices and facile preparation process, have been widely studied in solid propellent [28\u201330]. Kline et al. investigated the influences of thermally conducting (graphite) and insulating (SiO2) particles on film propagation rate of solid propellent, and demonstrated that the latter has a higher burning rate with low mass percentage additives [31]; Wang et al. demonstrated that Al/AP/SiO2 (7\u202fwt%) composite particles was more reactive and that is approximately 7 times higher than Al/AP particles [32]; Chen et al. successfully prepared AP/RDX/SiO2 nanocomposite via the sol-gel method and confirmed that SiO2 could accelerate the decomposition of AP and reduce the sensitivity of energetic materials [33]. Therefore, using SiO2 as combustion catalyst carrier will produce many positive effects in thermal decomposition, combustion and safety properties.For CuFe2O4/SiO2 composite combining silica and copper ferrite through in situ growth or sol-gel method, has been widely applied in fields of peroxidase and visual biosensing [34], magnetic materials [35], chemical looping gasification [36] and coupled with Li2O as anode material [37], yet there isn\u2019t any report of CuFe2O4/SiO2 applied in solid propellent so far. On the basis of above analysis, we speculated that CuFe2O4/SiO2 composite would be an excellent candidate material as combustion catalysts for AP and RDX.In this work, CuFe2O4/SiO2 binary nanocomposite was successfully prepared via the solvothermal method. This material was further used as combustion catalyst and its effect on thermal decomposition of AP and RDX was studied. Moreover, the combustion and safety properties of RDX over CuFe2O4/SiO2 nanocomposite were investigated. All results indicated that CuFe2O4/SiO2 nanocomposite could produce excellent catalytic effect on energetic components of solid propellent.All chemical reagents are of analytical grade and were used without further purification. FeCl3\u00b76H2O and CuCl2\u00b72H2O were purchased from Sinopharm Chemical Reagent Co., Ltd. Anhydrous CH3COONa and polyvinylpyrrolidone (PVP) were sourced from Damao Chemical Reagent Co., Ltd. Ethylene glycol was purchased from Kelong Chemical Reagent Co., Ltd. SiO2 (micrometer grade) was obtained from Macklin Biochemical technology Co., Ltd. The energetic materials (AP and RDX) were provided by Xi\u2019an Modern Chemistry Research Institute.CuFe2O4/SiO2 composite was synthesized through a versatile solvothermal method [3], and the preparation process is shown in Fig. 1\n. Firstly, 10\u202fmmol FeCl3\u00b76H2O and 5\u202fmmol CuCl2\u00b72H2O were separately dissolved in ethylene glycol (35\u202fmL) accompanied by magnetic stirring. Secondly, different proportions SiO2 (1\u202fwt%, 3\u202fwt%, 5\u202fwt%, 7\u202fwt%) were dissolved in FeCl3\u00b76H2O solutions, respectively. Meanwhile, 1\u202fg PVP was dissolved in CuCl2\u00b72H2O solution as a dispersant [38]. Those solutions were dispersed by ultrasonication under ultrasonic powder of 600\u202fW for 1\u202fh. Thirdly, the Fe3+ solution was mixed with the Cu2+ solution under continuous magnetic stirring until the complete mixture was achieved. Then, 2.46\u202fg CH3COONa, used as a sedimentation agent to promote the formation of the final product, was added into the mixture accompanying with vigorous stirring. Subsequently, the obtained homogenous reaction solution was sealed in a 100\u202fmL Teflon-lined stainless-steel autoclave and maintained at 180\u202f\u00b0C for 12\u202fh. The final product was rinsed by ethanol and deionized water for three times, respectively, the suspension was centrifuged and dried at 60\u202f\u00b0C for 12\u202fh. In contrast, pure CuFe2O4 nanoparticles were prepared under the same condition without adding SiO2 carrier.The phase structures of as-obtained samples were investigated by X-ray diffraction (Rigaku, Smart LAB SE); The surface morphology, size and element component were recorded by Scanning electron microscopy (Zeiss SIGMA) coupled with energy dispersive spectrometry (Oxford 51-XMX); Functional groups on the surface of samples were investigated by Fourier transform infrared (Shimadzu, IRAffinity-1S); Element composition and chemical valence state were determined via X-ray photoelectron spectra (ThermoFisher, NEXSA); Zeta potential analysis (Anton Paar, Litesizer\u2122 500) of SiO2 carrier was determined by Dynamic Light Scattering method; The specific surface areas and pore size distributions were measured by Nitrogen adsorption/desorption (Micromeritics, ASAP 2460) through the Brunauer-Emmett-Teller method.The catalytic decomposition behavior of energetic materials was studied by Differential scanning calorimetry (Netzsch, DSC-200F3), the N2 flow rate is 100\u202fmL/min, the heating rate is 10\u202f\u00b0C/min and temperature ranges are 30\u2013460\u202f\u00b0C and 30\u2013310\u202f\u00b0C for AP and RDX, respectively; The ignition characteristics were recorded by a CO2 laser ignition system (SLC 110) at air atmosphere. The stacking density of sample in crucible is around 0.76\u202fg/cm3; The electrostatic discharge sensitivity (EDS) was analyzed by a JGY-50(III) electrostatic test apparatus [39\u201341]. The test energy was determined by the energy calculation equation: E\n50\u202f=\u202f1/2CV2, in which C is the capacitance of the capacitor, V is the charge voltage in volts [42]. The charge capacitance was set as 10,000\u202fpF, and the electrode gap length was set as 0.12\u202fmm. In the above three types of characterizations, the catalyst was physically mixed with energetic materials at a mass ratio of 1:4.SEM was used to observe the loading of CuFe2O4 on the surface of SiO2 from apparent morphology, as shown in Fig. 2\n(a)\u20132(d). Fig. 2(a) shows the SiO2 substrate with a micrometer sheet structure, while the CuFe2O4 nanoparticles in Fig. 2(b) are spherical with an average particle size of 160\u2013170\u202fnm. Through further observation of CuFe2O4/SiO2-3% composite in Fig. 2(d), CuFe2O4 nanoparticles are uniformly distributed on the surface of SiO2 substrate. In comparison with CuFe2O4/SiO2-1% composite in Fig. 2(c), the agglomeration phenomenon of CuFe2O4 nanoparticles is effectively suppressed. Besides, the elemental mapping images in Fig. 2(d-1)\u20132(d-4) reveal that the elements of Fe, Cu, Si, and O are distributed on the surface of samples, which further demonstrates the uniform loading of CuFe2O4 particles on the surface of SiO2 substrate.XRD patterns and FT-IR spectra are often applied to study the composition and chemical bonding information of materials. The crystal structures of CuFe2O4, SiO2 and CuFe2O4/SiO2 composite were verified via XRD, as shown in Fig. 2(e). Feature peaks at 18.50\u00b0, 30.17\u00b0, 35.64\u00b0, 43.04\u00b0, 57.05\u00b0 and 62.77\u00b0 can be connected well with (111), (220), (311), (400), (511) and (440) panels of CuFe2O4, matching well with the standard ICDD card NO.025-0283 [26,27,43,44]. This proves that high purity CuFe2O4 was successful prepared. Meanwhile, feature peaks at 20.88\u00b0, 26.58\u00b0, 50.08\u00b0, 60.02\u00b0 and 67.86\u00b0 are correspond to (100), (101), (112), (211) and (212) panels of SiO2, which are well compatible with the standard ICDD card NO. 001\u20130649 [43]. It is essential that the intensity of feature peak at 26.58\u00b0 presents a gradually increasing tendency with larger SiO2 content in CuFe2O4/SiO2 composite. These typical feature peaks of CuFe2O4 and SiO2 co-exist in CuFe2O4/SiO2 binary composite, indicating that CuFe2O4 nanoparticles are successfully coupled with SiO2 substrate. Furthermore, the FT-IR spectrums of as-obtained samples are shown in Fig. 2(f), the bands around 1662\u202fcm\u22121 and 3400\u202fcm\u22121 belong to stretching vibrations of absorbed water molecules and hydroxyl group, respectively. The strong absorption bands at 468\u202fcm\u22121, 798\u202fcm\u22121 and 1060\u202fcm\u22121 are assigned to oscillatory vibration, symmetric stretching vibration and asymmetric stretching vibration of Si\u2013O\u2013Si bond, respectively [20,37,45,46]. The Si\u2013O\u2013Si bond at 451\u202fcm\u22121 exists in the tetrahedral position is overlapped with metal oxides vibration [37]. The absorb peak at 543\u202fcm\u22121 is attributed to metal-oxygen stretching [26,27,37]. XRD and FT-IR results demonstrate the successful synthesis of CuFe2O4/SiO2 composite.In the survey spectrum of CuFe2O4/SiO2-3% composite in Fig. 3\n(a), the Fe, Cu, Si and O elements can be seen clearly, there is no extra element in comparison with the survey spectrum of CuFe2O4 (Fig. 3(d)). The broad and asymmetric peak of O1s (Fig. 3(b)) spectrum implies that there can be more than one chemical state for O element, the two peaks centered at 530.16\u202feV and 532.91\u202feV are assigned to lattice oxygen in CuFe2O4 and nonmetal oxides (Si\u2013O), respectively [27]. Whereas, the peak centered at 531.55\u202feV belongs to the surface hydroxyl group (-OH) [43]. As shown in Fig. 3(c), the peaks centered at 103.38\u202feV and 100.30\u202feV indicate that Si element is presented in the form of SiO2 [30]. The Cu2p spectrum (Fig. 3(e)) exhibits four peaks located at 933.92\u202feV for Cu2p3/2, 953.40\u202feV for Cu2p1/2, 941.58\u202feV for the satellite feature of Cu2p3/2 and 962.56\u202feV for the satellite feature of Cu2p1/2 [43]. Meanwhile, the Fe2p spectrum (Fig. 3(f)) can be fitted into three contributions (FeO, Fe2O3 and Fe3O4), the peaks at 711.08 and 724.13\u202feV are assigned to the binding energies of 2p3/2 and 2p1/2 of Fe3+, 714.73 and 726.93\u202feV are attributed to the binding energies of 2p3/2 and 2p1/2 of Fe2+. In addition, the peak locating at 718.63\u202feV indicates that Fe3+ and Fe2+ coexist in the CuFe2O4/SiO2-3% composite [27,35]. XPS results further confirm the successful preparation of CuFe2O4/SiO2 composite.To illustrate the combination state between CuFe2O4 nanoparticles and SiO2 substrate, Zeta potential measurement of SiO2 particles was carried out at room temperature (25\u202f\u00b0C) and laser wavelength of 660\u202fnm. SiO2 particles were firstly suspended in ethylene glycol solution and then sonicated for 2\u202fh to form a homogeneous dispersion. The measurement was calculated for three times and results are shown in Fig. 4\n (a) and 4(b). The large zeta potential value (\u221254.0\u202fmV) of SiO2 particles indicates its good dispersion stability in ethylene glycol solution, and suggests that particles repel each other and do not undergo flocculation [47]. The negatively charged surface of SiO2 supplies active sites to absorb cations of Fe3+ and Cu2+ through electrostatic attraction. Results demonstrate that SiO2 can be an effective carrier for CuFe2O4 nanoparticles.The specific surface areas of as-obtained catalysts were investigated by a nitrogen adsorption/desorption instrument and calculated by the multipoint Brunauer-Emmett-Teller (BET) method, nitrogen adsorption-desorption isotherms curves and pore size distributions are shown in Fig. 4(c) and (d), respectively. Pure CuFe2O4 has a relatively small specific surface area (\u223c42.6125\u202fm2/g), but the specific surface areas of CuFe2O4/SiO2 composites progressively increases to 49.0999\u202fm2/g, 70.7566\u202fm2/g, 56.7653\u202fm2/g and 53.6471\u202fm2/g with SiO2 content increasing from 1 to 7\u202fwt%. Moreover, the maximum value (70.7566\u202fm2/g) achieved at 3\u202fwt% content of addition suggests that 3\u202fwt% content of SiO2 carrier possess enough capacity for dispersing CuFe2O4 nanoparticles. The results further suggest that the SiO2 carrier can inhibit the aggregation of CuFe2O4 nanoparticles, which is consistent with SEM image in Fig. 2(d). Additionally, the adsorption-desorption isotherms curve in Fig. 4(c) present typically type IV hysteresis loops when P/P\n0 range is 0.4\u20131.0 [34,44]. The maximum pore size can reach 148\u202fnm (Fig. 4(d)), indicating that the introduction of SiO2 carrier increases the pore content, pore size and specific surface area of catalysts, thus greatly improving the catalytic activity. Large specific surface area of CuFe2O4/SiO2 could offer more adsorption and reaction sites, which consequently result in better catalytic activity on energy components [43].The catalytic abilities of CuFe2O4/SiO2 composites on AP and RDX were investigated and shown in Fig. 5\n(a) and (b). DSC curves in Fig. 5(a) show a distinct endothermic peak of AP at 245\u202f\u00b0C assigning to transformation from orthorhombic phase to cubic phase [26]. The pure SiO2 advances the high temperature decomposition (HTD) and the low temperature decomposition (LTD) of AP from 403.8\u00b0C to 309.2\u202f\u00b0C\u2013389.8\u202f\u00b0C and 300.6\u202f\u00b0C, respectively, presenting a weaker catalytic decomposition on AP. Comparatively, the two exothermic peaks merge into one broad peak under the action of CuFe2O4 nanoparticles, and the peak temperature is advanced to 334.9\u202f\u00b0C. It can be seen clearly in Fig. 5(a) that there exists a remarkably synergistic catalytic effect after CuFe2O4 and SiO2 combining, the exothermic peak temperatures of AP decrease to 315.8\u202f\u00b0C, 310.7\u202f\u00b0C, 313.3\u202f\u00b0C and 318.6\u202f\u00b0C with SiO2 content increasing from 1 to 7\u202fwt%. In comparison with pure CuFe2O4 (334.9\u202f\u00b0C) and SiO2 (389.8\u202f\u00b0C), the 3\u202fwt% content of SiO2 in CuFe2O4/SiO2 composites can achieve the optimum catalytic effect and the exothermic peak temperatures of AP is advanced by 24.2 and 79.1\u202f\u00b0C, respectively.Taking good catalytic effect of CuFe2O4/SiO2 composites on AP thermal decomposing into consideration, its catalytic activity on RDX was further investigated, as shown in Fig. 5(b). The peak temperature of RDX exothermic decomposing is advanced to 241.1\u202f\u00b0C and 239.8\u202f\u00b0C under the catalytic effects of CuFe2O4 and SiO2, respectively, and the exothermic peak temperature of RDX decreases to 238.3\u202f\u00b0C, 234.9\u202f\u00b0C, 236.4\u202f\u00b0C and 237.4\u202f\u00b0C, with SiO2 content increasing from 1 to 7\u202fwt%. Interestingly, the variation tendency of catalytic effect of CuFe2O4/SiO2 composites on RDX is consistent well with that of AP, in which the peak temperatures of them all decrease firstly and then increase accompanied with the increasing of SiO2 content in CuFe2O4/SiO2 composites.The synergistic effect of CuFe2O4 and SiO2 can be explained as: (1) The uniform dispersion of CuFe2O4 nanoparticles on the surface of SiO2 by electrostatic interaction and the effective inhibition of aggregation of nanoparticles proved by SEM analysis in subsection 3.1, both RDX (234.9\u202f\u00b0C) and AP (310.7\u202f\u00b0C) realized the lowest thermal decomposition peak temperature under the catalytic effect of CuFe2O4/SiO2 (3\u202fwt%) composites; (2) The specific surface area of CuFe2O4/SiO2 composite achieves the maximum at the carrier content of 3\u202fwt%, as described in subsection 3.1, which generates more active sites and higher reaction activity, resulting in better catalytic ability; (3) CuFe2O4 nanoparticles show more obvious catalytic effect than pure SiO2 carrier for thermal decomposition of AP and RDX, hence higher content of SiO2 (5\u202fwt% and 7\u202fwt%) would weaken the synergistic catalytic effect.In addition, the apparent activation energy (E\u03b1) is an essential indicator reflecting the difficulty of decomposition process of energetic materials, which is of great significance for the study of catalytic decomposition performance [48,49]. In order to contrast the catalytic effect of CuFe2O4/SiO2 on AP and RDX, the E\u03b1 values of AP\u00a0+\u00a0CuFe2O4/SiO2 (3\u00a0wt%) and RDX\u00a0+\u00a0CuFe2O4/SiO2 (3\u202fwt%) were calculated from DSC data recorded at different heating rates (\u03b2 = 5.0, 10.0, 15.0 and 20.0\u202f\u00b0C/min) using Kissinger method [50] and Flynn-Wall-Ozawa method [51]. As shown in Fig. 5(c) and (d), the peak shape remains unchanged while the decomposition peaks advance to high temperature successively with heating rate increasing. With the addition of CuFe2O4/SiO2 catalyst, the E\u03b1 values of AP and RDX decrease from 161.0\u202fkJ/mol and 239.6\u202fkJ/mol to 139.3\u202fkJ/mol and 113.6\u202fkJ/mol for Kissinger method, 141.7\u202fkJ/mol and 116.1\u202fkJ/mol for Flynn-Wall-Ozawa method, respectively. Comparing with previous research [26], CuFe2O4/SiO2 (3\u202fwt%) decreases the E\u03b1 value of AP and RDX by 21.7\u202fkJ/mol and 126\u202fkJ/mol, respectively. It can be seen in Table 1\n that fewer catalyst carrier (3\u202fwt%) but higher catalytic activity is achieved, which illustrates the prominent catalyst effect of CuFe2O4/SiO2 composites on AP and RDX, and further confirms that SiO2 can be used as an effective carrier for CuFe2O4 nanoparticles.Multiple catalysts have been used to improve the thermal decomposition properties of AP and RDX, results all confirmed the catalytic effects on advancing the high exothermal peak temperature, reducing the apparent activation energy, and even transforming the slow two-stage decomposition process into a rapid one-step process. For comparison, the catalytic capacities of several catalysts or the combination of them are listed in Table 2\n, which further verifies the excellent synergistic catalytic effect of the CuFe2O4/SiO2 composites on both AP and RDX.Both ignition delay time and flame propagation velocity are important parameters to investigate the ignition and combustion performance of energetic materials [56], in which the former is defined as the time interval between the triggering of laser and the appearance of igniting flame [39,40], and the latter is related to the amount and rate of gas production.As shown in Fig. 6\n(a), the ignition delay time of all samples decreases sharply with the increase of laser power density. CuFe2O4 nanoparticles can shorten ignition delay time of RDX in lower power density of 109.3\u202fW/cm2 and improve the ignition properties of RDX, revealing that CuFe2O4 nanoparticles tend to be more sensitive to laser radiation. However, the ignition delay time of RDX increases after introducing SiO2 and higher SiO2 content is corresponding to longer ignition delay time. Moreover, the ignition delay time values of all composites tend to be close to each other after the power density exceeding 155.3\u202fW/cm2. In Fig. 6(b), the flame propagation velocity of RDX\u00a0+\u00a0CuFe2O4 is investigated, which exhibits a lower flame propagation velocity than that of pure RDX but the longest combustion time (272\u202fms). Besides, the flame propagation velocity of CuFe2O4/SiO2(3\u202fwt%) is 2.73 times faster than pure RDX. Considering the thermally insulating property of SiO2 carrier, the abnormal phenomenon may be explained as (1) The poor thermal conductivity of SiO2 makes it a barrier for thermal conduction, therefore the temperature at the vicinity of SiO2 particles increases rapidly and results in multiple ignition points for combustion, thereby promoting the combustion progress [57,58]; (2) Once ignited, particulate product is ejected along with high-pressure gas to increase the size of the flame, facilitating laser feedback and heat transfer [31,32]; (3) It is worthy to note that higher content (5\u202fwt%) of SiO2 may decrease the energy density and reactivity because of heat-sink effect [32].To illustrate the combustion performance of CuFe2O4/SiO2 composites, the burning snapshotting of RDX, RDX\u00a0+\u00a0CuFe2O4 and RDX\u00a0+\u00a0CuFe2O4/SiO2 (1\u202fwt%, 3\u202fwt%, 5\u202fwt%) composites were recorded under the same condition for comparison. As shown in Fig. 7\n(a), the combustion flame of pure RDX is dark red and weakly propagated, while the flame of RDX\u00a0+\u00a0CuFe2O4 is more luminous and lasts for 272\u202fms in Fig. 7(b), proving that CuFe2O4 could promote burning process of RDX. This might be attributed to the good thermal diffusivity of CuFe2O4 nanoparticle which facilitates heat transfer and accelerates combustion progress. Yet the combustion time decreases to 213\u202fms (Fig. 7(c)) after the introduction of SiO2 carrier (1\u202fwt%). As shown in the 5\u202fms picture (Fig. 7(d)) of RDX\u00a0+\u00a0CuFe2O4/SiO2 (3\u202fwt%), the mixture immediately forms an accelerated flame propagation fronter at the highest velocity and the combustion process completes in the shortest time (103\u202fms). The reaction intensity in Fig. 7(e) becomes weaker and flame propagation velocity declines sharply, which might be due to the high content of SiO2 in composite. All results reveal that the influence of reaction front area is greater than that of thermal diffusion at low mass percentage (3\u202fwt%) additives in solid propellent [32]. Therefore, it is necessary to control the appropriate content of SiO2 carrier to achieve the expected effect in practical application.The production, transportation and storage process of energetic material has high requirements due to its high energy density, therefore the sensitivity regulation is particularly important [59]. The electrostatic discharge sensitivities (EDS) of samples were investigated by apparatus JGY-50(III) introduced in subsection 2.4. As shown in Fig. 8\n, the introduction of CuFe2O4 nanoparticles produces a lower EDS value of RDX, which may easily produce security problems. However, E\n50 values of RDX\u00a0+\u00a0CuFe2O4/SiO2 composites increase to 7.32\u202f\u00b1\u202f0.05\u202fmJ as the SiO2 content rises to 7\u00a0wt%, indicating that the EDS value decreases by 189% and 230% when compared it with pure RDX (3.87\u00a0\u00b1\u00a00.05\u00a0mJ) and RDX\u00a0+\u00a0CuFe2O4 (3.20\u00a0\u00b1\u00a00.05\u00a0mJ), respectively. Additionally, the value of electrostatic discharge sensitivity of RDX\u00a0+\u00a0SiO2 (100\u202fwt%) decreases by 253% compared with that of pure RDX, which might be attributed to the micron-size structure of SiO2 which facilitates the dissipation of energy when electrostatic forces act on the composites [33,41], and the charge accumulation caused by aggregation of CuFe2O4 nanoparticles is eliminated to a great extent after introducing the SiO2. Therefore, in carrier content range of 3\u20135\u202fwt%, the CuFe2O4/SiO2 catalyst not only possess excellent catalytic and combustion properties, but also present stable safety characteristics.In this work, CuFe2O4/SiO2 composite was successfully synthesized through solvothermal method, proving by phase structure, morphology and chemical bond. The effect of SiO2 content on the catalytic property of CuFe2O4/SiO2 composite was explored, and it is found that CuFe2O4/SiO2 (3\u202fwt%) reduced the exothermic decomposition temperature of AP and RDX by 93.1\u202f\u00b0C and 7.4\u202f\u00b0C, respectively. In addition, the promotion effect of CuFe2O4/SiO2 composite on the combustion and safety performance of RDX were investigated and the results indicate that SiO2 also plays a significant role in accelerating the flame propagate and enhancing the anti-electrostatic ability of RDX, in which the flame propagate velocity increases from 0.705\u202f\u00b1\u202f0.005 to 1.923\u202f\u00b1\u202f0.025\u202fm/s after introduction of SiO2 carrier (3\u202fwt%) and the E\n\n50\n value rises from 3.87\u202f\u00b1\u202f0.05 to 7.32\u202f\u00b1\u202f0.05\u202fmJ after introduction of SiO2 carrier (7\u202fwt%). Therefore, SiO2 is confirmed to be an excellent carrier that can synergistic with CuFe2O4 to form combustion catalysts and be used in solid propellants. Furthermore, the range of SiO2 carrier content in which the catalyst exhibits remarkably catalytic performance, certain combustion promotion effect, and controllable electrostatic discharge sensitivity is found to be 3\u202fwt% to 5\u202fwt%.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This investigation received financial assistance from the National Nature Science Foundation of China (Grant Nos. 21673178, 22105160), the Natural Science Foundation of Shaanxi Province (Grant No. 2023-JC-ZD-07), the Foundation of Key Laboratory of Defense Science and technology (Grant No. 6142603032213) and the Key Science and Technology Innovation Team of Shaanxi Province (Grant No. 2022TD-33).", "descript": "\n                  To enhance the catalytic activity of copper ferrite (CuFe2O4) nanoparticle and promote its application as combustion catalyst, a low-cost silicon dioxide (SiO2) carriers was employed to construct a novel CuFe2O4/SiO2 binary composites via solvothermal method. The phase structure, morphology and catalytic activity of CuFe2O4/SiO2 composites were studied firstly, and thermal decomposition, combustion and safety performance of ammonium perchlorate (AP) and 1,3,5-trinitroperhydro-1,3,5-triazine (RDX) with it affecting were then systematically analyzed. The results show that CuFe2O4/SiO2 composite can remarkably either advance the decomposition peak temperature of AP and RDX, or reduce the apparent activation energy at their main decomposition zone. Moreover, the flame propagation rate of RDX was promoted by about 2.73 times with SiO2 content of 3\u202fwt%, and safety property of energetic component was also improved greatly, in which depressing the electrostatic discharge sensitivity of pure RDX by about 1.89 times. In addition, the effective range of SiO2 carrier content in the binary catalyst is found to be 3 to 5\u202fwt%. Therefore, SiO2 opens a new insight on the design of combustion catalyst carrier and will promote the application of CuFe2O4 catalyst in solid propellant.\n               "}
{"full_text": "Hydroformylation, also known as oxo-synthesis reaction, is one of the most important homogeneously catalyzed industrial processes for the production of aldehydes from alkenes and syngas with 100% atom economy, which was found by Otto Roelen as early as 1938.\n1\n\n,\n\n2\n\n,\n\n3\n\n,\n\n4\n\n,\n\n5\n Up to now, the production of chemical produces through this hydroformylation has exceeded 12 million tons every year, which is one of the most significant commercial uses of soluble homogeneous metal catalyst in the chemical industry.\n6\n\n,\n\n7\n The obtained aldehyde products for hydroformylation can be converted to the valuable and stable products (alcohols, ketones, acetals, amines products, etc.) through oxidation, hydrogenation, or reductive amination, which are extensively utilized in the synthesis of fine compounds such as insecticides, spices, food additives, and plasticizers.\n8\n\n,\n\n9\n\nThe typical catalysts for hydroformylation of olefins are homogeneous complexes of the type [HM(CO)xLy], where L can be further CO or an organic ligand. A generally accepted series of the activities of the unmodified metal is as follows:\n10\n Rh >> Co\u00a0>\u00a0Ir\u00a0>\u00a0Ru\u00a0>\u00a0Pd\u00a0>\u00a0Mn\u00a0>\u00a0Fe\u00a0>\u00a0Ni >> Re. To date, only cobalt and rhodium catalysts can be used in industrial application; other metals only stay in the stage of academic research. Rhodium-phosphine complex catalyst has the advantages of mild reaction conditions, well catalytic performance, and low energy consumption. It has gradually become the mainstream in the industrial hydroformylation instead of cobalt carbonyl catalyst.\n11\n The uniform distribution of active sites, excellent catalytic activity, and superior chem/regioselectivity of homogeneous catalysts are only a few of their many benefits. However, the issue of catalyst separation leads to the loss of active metal and phosphine ligand, which is not conducive to large-scale application in industrial production. In contrast, heterogeneous catalysts can overcome catalyst separate deficiencies. Due to the surface properties of the support, the interaction between metal and support, and the microenvironment of the catalytic sites, heterogeneous catalysts can demonstrate excellent performance. Therefore, the development of heterogeneous catalysts with high activity and high stability for hydroformylation has important theoretical and practical significance.\n12\n The term \u201cheterogeneous catalyst\u201d describes a catalyst that immobilizes the active metal or metal complex on a solid support. Molecular sieve,\n13\n\n,\n\n14\n\n,\n\n15\n carbon materials,\n5\n\n,\n\n16\n\n,\n\n17\n inorganic oxides,\n18\n\n,\n\n19\n\n,\n\n20\n magnetic nanoparticles,\n21\n and organic polymers\n22\n\n,\n\n23\n\n,\n\n24\n are the examples of supports.In contrast to traditional heterogeneous catalysts, single-atomic catalysts (SACs) are a recently emerging class of catalytic material featured with unique single-atom dispersion and maximum atomic utilization of active metal.\n25\n\n,\n\n26\n The atomically dispersed metal anchored on support brings similar catalytic behavior to homogeneous catalyst. In addition, the heterogenous property of SACs makes them easy to be separated from the liquid-phase reaction mixture and achieve convenient recovery as well as recycling. Combining the advantages of homogeneous catalysts and heterogeneous catalysts, SACs exhibit high catalytic activity and selectivity in hydroformylation.\n27\n\n,\n\n28\n\n,\n\n29\n\n,\n\n30\n\n,\n\n31\n\n,\n\n32\n\n,\n\n33\n\nHowever, up to now, very few reviews of SACs in hydroformylation have been reported. In this paper, we\u00a0summarize recent advances of SACs for hydroformylation. The effects of microstructure of SACs on the reactivity and chem/regioselectivity of hydroformylation are discussed. The support effect, ligand effect, and electron effect on the performance of SACs in hydroformylation are proposed. The mechanism of SACs in hydroformylation is elaborated. Finally, we summarize the current problems and challenges in this field, and propose the design and research ideas of SACs for hydroformylation (Figure\u00a01\n).The application of SACs in hydroformylation is still in the early stage. According to the current results, SACs have great potential to achieve high activity and selectivity of hydroformylation (Figure\u00a02\n) since they have extremely high metal dispersion, low coordination environment in the metal center, and the strong interaction between metal atoms and support. Herein, we focus on the recent development of SACs in the field of hydroformylation.Homogenous phosphines-modified Rh catalysts have shown remarkable performance in the hydroformylation process before the application of SACs. In the 1950s, Union Carbide applied RhCl(PPh3)3 to the industry. The \u201clow-pressure oxo-progress\u201d has much higher stability and milder conditions than Co-based catalysts.\n47\n Later, Rhone-Poulenc Company and Ruhrchemie Company jointly developed RCH/RP process to achieve a new two-phase (organic/water) catalytic system. In this process, the water-soluble Rh-P complex was dissolved in the water phase; the products were dissolved in the oil phase.\n48\n The effective separation of the products and catalyst can be achieved by simple static layering and decanting operation. Compared with Co-based catalysts, Rh-based system possessed higher catalytic activity, selectivity and stability, and the milder conditions.\n49\n\n,\n\n50\n Therefore, Rh SACs are the potential supported catalysts for hydroformylation under mild conditions.In 2016, Zhang et\u00a0al.\n42\n synthesized Rh SACs by the impregnation method to adsorb Rh3+ onto ZnO nanowires (Rh1/ZnO-nw) for the hydroformylation process (Figure\u00a03\nA). Compared to the typical Wilkinson\u2019s catalyst RhCl(PPh3)3 (Turnover number (TON)\u00a0= 19000), the Rh1/ZnO-nw showed excellent activity (TON\u00a0= 40000), and can be recycled and reused for four times without significant loss of reactivity and selectivity. However, the ratio of linear to branched aldehyde (L/B) was only 1.0. The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images clearly distinguished that isolated Rh atoms were distinguishable from the ZnO nanowires (Figure\u00a03B). In the characterization of in situ diffuse reflectance infrared Fourier transform spectroscopy study of CO adsorption (CO-DRIFTS) (Figure\u00a03C), the absence of a Rh0-CO bridge adsorption peak implied that Rh was dispersed on ZnO support in the form of single atoms, which was compatible with the characterization results of HADDF-STEM.\n54\n\n,\n\n55\n\n,\n\n56\n\nIn the same year, Wang et\u00a0al.\n51\n reported CoO-supported Rh SACs (Rh/CoO) for the hydroformylation of propylene. The selectivity of butyraldehyde was as high as 94.4% and the turnover frequency (TOF) can reach 2065 h-1. After five cycles, the activity and selectivity of the catalyst remained at high levels (Figure\u00a03D). Extended X-ray absorption fine structure (EXAFS) revealed that Rh atoms were atomically dispersed for the sample of 0.2% Rh/CoO. As shown in Figure\u00a03E, the peak at ca. 2.0\u00a0\u00c5 was attributed to Rh-O shell, and no other peaks for Rh-Rh contribution were observed. Additional mechanistic studies revealed that the structural reconstruction of the Rh SACs took place during the catalytic process, which facilitated in the adsorption and activation of the reactants.Gao et\u00a0al.\n52\n synthesized a phosphorus coordinated Rh SACs (Rh1/PNP-ND) through the metal-ligand coordination approach (Figures\u00a03F and 3G), and applied it to the hydroformylation of styrene to achieve high conversion (>99%) and high selectivity (>90%) under mild conditions. The abundant carboxyl groups loaded on the surface of nanodiamond (ND) selectively react with the amino groups of pincer ligands (PNP) to obtain PNP-ND, which provided a large number of sites for the highly dispersed anchoring of Rh. The 31P solid-state NMR spectra showed that the chemical shift of P atoms migrated to the low-field region after the Rh species were anchored by PNP-ND. This provided direct evidences for the successful anchoring of Rh species in PNP-ND. In order to further demonstrate the adaptability to different substrates, the Rh1/PNP-ND was applied to the hydroformylation of a series of styrene derivatives, which demonstrated exceptional selectivity and activity, comparable to homogeneous catalysts.In 2020, Li et\u00a0al.\n45\n synthesized 0.5% Rh/CeO2 SACs by electrostatic adsorption method innovatively, and coupled hydroformylation with low-temperature water-gas shift reactions. Without using any ligand, the catalytic system not only avoided the hydrogenation of styrene and phenylpropyl aldehyde occurred as the side reaction but also achieved high selectivity to obtain linear aldehyde in the hydroformylation of styrene and its derivatives (L/B\u00a0= 3). Due to the high Ce vacancy density in CeO2 support, high loading of active Rh sited can be achieved. HAADF-STEM images and CO-DRIFTS proved that Rh existed in the form of single atom in 0.5% Rh/CeO2 catalyst. To further comprehend the new reaction route, a number of comparative tests were conducted. In contrast to the conventional reaction with styrene, CO, and H2 as substrates, the authors found that the higher linear/branched aldehyde ratio obtained was related to the reactants of CO and H2O. When secondary alcohol dehydrogenation was coupled with hydroformylation reaction, the selectivity of linear products was still higher than that of branched products, although the activity was lower. Based on the above facts, the author proposes a six-membered transition formed by the combination of C=C unsaturated double bond and formic acid. According to Marcovnikov's rule, the active H addition the end of C=C bond. This intermediate not only facilitates the insertion of carbonyl groups into the terminal C=C bond to form the linear aldehydes but also prevents the formation of phenyl Rh species, which ultimately inhibits the formation of linear aldehydes.\n57\n\nZhao et\u00a0al.\n46\n successfully encapsulated Rh within porous monophosphine polymers (POPs) by one-pot method to prepare Rh@POP-PTBA-HA-50.\n58\n\n,\n\n59\n\n,\n\n60\n According to the characterization of HAADF-STEM and EXAFS (Figure\u00a03H), it is proved that Rh species were encapsulated as a single-atom in the POPs skeleton. Fourier transform infrared spectroscopy (FT-IR) spectrum of Rh@POP-PTBA-HA-50 showed that a strong C=N stretch at 1623\u00a0cm-1, and the peaks at 1700 and 3345\u00a0cm-1 attributed to aldehyde group were obviously weakened compared to 4,4\u2032,4\u2019\u2019-phosphanetriyltribenzaldehyde (PTBA) and N2H4H2O (HA). In addition, the 13C magic angle spinning NMR peak of Rh@POP-PTBA-HA-50 at 162 ppm matched to the carbon atom of the C=N bond. Both of them indicated the formation of imine bonds. Compared to Rh(CO)2(acac)-PTBA, Rh@POP-PTBA-HA-50 showed a significant improvement in regioselectivity (linear aldehydes) from 62% to 92% in hydroformylation of 1-octene (Figure\u00a03I). Due to the robust coordination of dispersed phosphine ligands with metal active species, the catalyst demonstrated remarkable catalytic activity (TON\u00a0= 60000) and good thermal stability.The obtained aldehyde products from hydroformylation can be further converted into high-valuable chemicals like amines, carboxylic acids, and alcohols through additional oxidation, reduction, and hydrogenation.\n61\n\n,\n\n62\n\n,\n\n63\n\n,\n\n64\n\n,\n\n65\n Hydroformylation followed by other reactions through one-pot method has been\u00a0extensively explored, such as \u201chydroformylation-hydrogenation\u201d, \u201chydroformylation-acetalization\u201d, \u201chydroformylation-aldol condensation\u201d, \u201chydroformylation and reductive amination\u201d, and so on. Following the atom economy and low energy consumption in green chemistry, combining SACs with tandem hydroformylation have become a powerful and promising synthetic method. Li et\u00a0al.\n53\n successfully prepared hydroxyapatite (HAP)-supported single-atom Rh catalyst (Rh1/HAP) for the tandem hydroaminomethylation of olefins. (Figure\u00a03J). HAADF-STEM and CO-DRIFTS results revealed that Rh atom was atomically dispersed on the HAP support. 1-hexene was almost entirely converted over 0.5Rh1/HAP under moderate reaction conditions, and the selectivity was 93.2%. The hydroformylation, condensation, and hydrogenation reactions are all parts of the overall hydrocarbamoylation reaction. According to the mechanistic study, the hydrocarbamoylation process is a speed-regulating step. Through separate evaluation of hydroformylation reaction, 0.5Rh1/HAP guaranteed high activity of hydroformylation reaction, thus ensuring the excellent catalytic activity of the tandem reaction.Co is another metal catalyst applied for hydroformylation industry. The catalytic activity of Rh is 103\u2013104 times than that of Co.\n66\n\n,\n\n67\n However, the shortage and the high price of precious Rh limit its development and application in hydroformylation. Co continues to have a long-term role, since its effective antitoxic performance and the weak requirement for olefin\u2019s purity. As early as 1952, the carbonyl cobalt catalysts with HCo(CO)4 as the active ingredient were first applied in the oxo-synthesis of propylene in 1952. Later, cobalt carbonyl modified by phosphines could decrease pressure to 5\u201310 MPa in the 1950s and the CO was replaced by PR3, P(OR), etc. Compared to CO, the phosphines possessed stronger \u03c3-electron-donating ability and weaker \u03c0-receptor-accepting ability. The selectivity of linear aldehydes is significantly increased in phosphines modified Co system.\n68\n\n,\n\n69\n However, the hydrogenation of olefins to alkanes occurred, which reduced the activity relatively. In recent years, supported cobalt-based catalysts have attracted a lot of attention. Basic researches have been done in the laboratory, but there is still a significant gap between these efforts and industrial manufacturing.Recently, Cong et\u00a0al.\n34\n developed the ultrasound-assisted impregnation method to design Co SACs supported by zirconium phosphate (CoZrP-2.0). The tight coordination of Co atom with phosphate group of ZrP prevented the leaching of Co, and enhanced the activity and stability of catalyst (Figures\u00a04A and 4B). In CoZrP-2.0-catalyzed hydroformylation, the conversion of 1-octence was about 100%, and the selectivity of C9 aldehyde was 91.3%. After six cycles, the activity and selectivity of the catalyst remained at high levels. Based on the pyridine adsorption FT-IR spectrum and inductively coupled plasma-atomic emission spectrometry (ICP-AES) analysis of CoZrP-X catalyst (Figure\u00a04C), it was clearly observed that the leaching of Co was closely related to the loss of Br\u00f8nsted acid. On the CoZrP-2.0, the higher the loss of B-acid sites (CoZrP-2.0: 174.7\u00a0mmolg-1), the lower leaching of Co species. With the increase of P/Zr ratio, more Co atoms combined with phosphate groups by replacing the protons of the B-acid sites, which promoted the formation of ionic Co atoms. The coordination structure and chemical surroundings Co cites were examined in depth using EXAFS structural characterization (Figures\u00a04D and 4E). Similar peaks at 1.5\u00a0\u00c5 and 2.6\u00a0\u00c5 corresponded to the initial coordination shells of Co-O and Co-Co, respectively, were presented in samples of CoZrP-0.5, CoZrP-1.0, and CoZrP-1.5. Only the Co-O bond (1.5\u00a0\u00c5) was observed on CoZrP-2.0, revealing the single atom characteristic. With the increase of Br\u00f8nsted acid site and BET surface area with the increase of P/Zr (from 0.5 to 2.0), Co atoms are more evenly dispersed on the support. According to the EXAFS characterization results, when P/Zr\u00a0= 2.0, the Co atoms are dispersed as a single atom on the support. By comparing the catalytic performance of different P/Zr catalysts for 1-octene hydroformylation, it can be found that with the increase of P/Zr from 0.5 to 2.0, the conversion rate of the catalyst decreased slightly, but the selectivity for aldehydes increased from 59.4% to 89.6%, and the leaching rate of Co decreased sharply from 29.1% to 0.5%. Co SACs mainly take aldehydes as the main product, which has better activity, selectivity, and stability.Current industrial production of hydroformylation mainly employs Rh-based catalyst. The expensive price of rhodium has promoted the research of other alternative transition metal catalysts in hydroformylation. A pioneering work on ruthenium catalysts for homogeneous hydroformylation in 1965 by Wilkinson et\u00a0al\n70\n is worth noting that Ru can significantly promote the catalytic activity of cobalt catalyst in hydroformylation reaction. Masanobu Hidai et\u00a0al.\n71\n studied the synergistic effect of bimetallic catalysts in the hydroformylation of olefin. The initial reaction rate of hydroformylation of cyclohexene catalyzed by Co2(CO)8/Ru3(CO)12 was 19 times higher than that of Co2(CO)8, and this bimetallic catalyst has a wide range of substrate applicability. The addition of 1 wt% Ru to 10 wt% Co/AC increased the conversion of 1-hexene by nearly 60%, inhibited the side reaction of alkene isomerism, and improved the selectivity of aldehydes. Zhang et\u00a0al.\n16\n believed that the addition of Ru can greatly improve the reducibility, provided more cobalt metal centers for the reaction.Escobar-Bedia et\u00a0al.\n35\n developed a novel Ru-based catalyst (Ru@NC) containing isolated single atoms and disordered clusters in nitrogen-doped carbon matrix, that applied to hydroformylation of 1-hexene with good activity and selectivity (Figures\u00a05A and 5B). The strong interaction between Ru and N atoms can improve the dispersion of metal and change the electronic properties of Ru atoms on the surface, thus affecting the stability and activity of catalyst. According to scanning electron microscope energy dispersive spectrometer findings, N and Ru atoms embedded in the carbon substrate were distributed uniformly on the support and their signals overlapped. Moreover, the potential connection between surface N and Ru atoms was confirmed by a prominent Ru-N bond at 460\u00a0cm-1 on the Raman spectra. According to the X-ray absorption near-edge region (XANES, left) spectral and EXAFS spectral (right) analysis (Figure\u00a05C), Ru mainly existed in the form of single atoms in Ru@NC. However, Ru-Ru scattering intensity increased with the increase of the metal loading. Ru-Ru is detected to exist in a highly disordered state in the highly loaded catalyst, indicating the formation of small, dispersed, and highly disordered Ru clusters. In order to further investigate the impact of N atoms on the performance of catalyst, 0.2Ru@NaC and 0.2Ru@NC were conducted as the comparison experiments under the same circumstances (Figure\u00a05D). The results showed that the rate of 0.2Ru@NaC decreased significantly with the prolongation of reaction time; the leaching of Ru in solution increased significantly. Moreover, the regioselectivity of 0.2Ru@NaC in hydroformylation was lower than that of 0.2Ru@NC. The interaction of surface N atom with Ru atom can stabilize and change the electronic properties of Ru atom.Various transition metals including Rh, Co, Ir, Ru, and Fe have been proven to be efficient catalysts for hydroformylation. Although Au is conventionally considered inactive for hydroformylation, numerous studies have shown that Au exhibits high olefin activity,\n72\n H2 dissociation,\n73\n\n,\n\n74\n and CO bonding capabilities.\n75\n At\u00a0same time, Au was applied for CO oxidation,\n75\n water gas conversion reaction,\n76\n and methanol synthesis.\n77\n\n,\n\n78\n\nWei et\u00a0al.\n40\n encapsulated dispersed Au into purely siliceous zeolite to prepare Au(0.2%)@S-1 (Figure\u00a06\nA). The Au SACs showed high activity noticeable stability after 5 cycles in the hydroformylation of propylene, which was one order of magnitude greater than Au nanoparticle catalysts (Au(0.8%)@S-1 and Au(0.2%)/S-1) (Figures\u00a06B and 6C). In addition, transmission electron microscopy, X-ray diffraction,\u00a0and X-ray photoelectron spectroscopy (XPS) provided evidence that Au was coated in the molecular\u00a0sieve and the morphology of the molecular sieve was unaffected by Au species. The utilization of Cs-corrected HAADF-STEM (Figure\u00a06D), XANES spectra (Figure\u00a06E), and EXAFS spectra (Figure\u00a06F) confirmed that the oxygen bridge bond of molecular sieve enclosed the atomically scattered Au to create the Au-O-SiOx structure, which maximizes the active site\u2019s density and structural stability.The supports of heterogeneous catalysts often employed inorganic oxides, POPs, metal-organic frameworks (MOFs), carbon materials, etc. The chem/regioselectivity of reactions increased through the modification of organic phosphine ligands or inorganic materials, and MOF domain limitation. However, the deactivation of catalysts and the loss of active components are still needed to be explored. Table 1\n summarizes the catalytic performance of hydroformylation on SACs.POPs are a new material composed of C, N, O, and H atoms with high specific surface area, low skeleton density, controllable pore structure, and excellent stability.\n79\n\n,\n\n80\n\n,\n\n81\n\n,\n\n82\n\n,\n\n83\n It provides a new class of polymer support for the preparation of SACs that possess the advantages of both homogeneous and heterogeneous catalysts. The utilization of POPs was conducive to the diversification of ligand modification, due to the immobilization of phosphines in polymer chains by covalent bonds. Further coordination of Rh atoms with P atoms can realize high loading of active metal. The high concentration of ligand can stabilize the metal atoms and prolong the life of the catalyst. In addition, POPs are typically insoluble in the majority of solvents, which prevents catalyst loss via dissolution.\n84\n However, the poor mechanical strength, poor thermal conductivity, complicated synthesis steps, and strict preparation conditions limit the large-scale production of POPs. For the first time, POL-PPh3 was synthesized via solvothermal polymerization in 2014 by the Xiao team and Ding team.\n58\n N2 adsorption/desorption curves showed high specific surface area of POL-PPh3 (1086 m2/g), and the concentrated pore size distribution at 0.7, 1.5, and 3\u201370\u00a0nm. The synthesized POL-PPh3 with graded porosity is conducive to the uniform dispersion of active centers. Jiang et\u00a0al.\n85\n used POL-PPh3 as a support to synthesize Rh SACs (Rh/POL-PPh3), which exhibited outstanding activity in the hydroformylation of ethylene in a fixed-bed reactor. After long-term stability test for more than 1000 h, the conversion of ethylene maintained at a stable level, and the loss of Rh catalyst was only 0.0046%. HAADF-STEM image clearly shows that the isolated Rh atoms are uniformly dispersed on the POL-PPh3 with porous structure. In addition, no sintering or aggregation was seen in the Rh species after long-term stability tests. Only Rh-P and Rh-C bonds were found according to EXAFS spectroscopy, which demonstrated that the strong coordination of Rh atoms with the exposed P atoms in the POL-PPh3 framework prevented the loss of active metal during the reaction.Metal oxides (ZnO, CoO, CeO2, and Al2O3) are frequently employed as support due to their outstanding chemical, thermal, and mechanical stability. The interactions between metal and support, the hydrogen overflow effects, and synergistic effects influence the catalytic activity of oxide-supported metal catalysts.\n86\n\n,\n\n87\n However, the interaction between the active metal and the oxide support may lead to the migration of surface particles, and finally the inert oxide will coat the active metal particles, thus deactivating the catalyst.\n88\n Amsler et\u00a0al.\n66\n investigated the activity and stability of Rh SACs loaded with different oxides (MgO, CeO2, and ZnO) in the hydroformylation of olefin by combining theoretical calculation and experimental study. Through calculating the free energy of supported catalysts in comparison to the complex HRh(CO)4, the atomically dispersed Rh/MgO was determined to be the most stable. HRh(CO)4 on flat oxide surfaces (CeO2 (111)) has catalytic activity comparable to that of molecular complexes. However, for the step edge on the MgO (301) surface, the calculation shows that the catalytic activity was significantly reduced. EXAFS characterization showed Rh atoms on MgO in higher coordination environments and higher degrees of confinement. The strong contact between Rh and the support, which interfered with the recovery of active species and the product\u2019s desorption, resulted in the low activity.Molecular sieves have been applied in numerous catalytic fields because of their special shape selectivity, adjustable acidity, and high water/thermal stability. In addition, molecular sieves can also be used as carriers to support, coat, disperse, and stabilize metal-active species (nanoparticles, clusters, single atoms, and isolated ions), achieving excellent activity and stability in heterogeneous process. The main problem of using molecular sieve as support is that the pore size of the support itself will affect the activity and selectivity of the catalyst. Moreover, there are abundant proton acid sites on the surface of molecular sieve, which may promote side reactions such as aldol condensation.\n89\n Shang et\u00a0al.\n39\n prepared Rh SACs (Rh@Y) utilizing an in situ guiding agent. In hydroformylation of olefins, Rh@Y showed significant catalytic activity, cycle stability, and substrate suitability under relatively mild conditions. Under the same experiment conditions, compared to other kind of catalysts (Rh/S-1, Rh/USY, Rh/ZSM-5, Rh/Beta, Rh/Mor, Rh/Y, and Rh-Y), Rh@Y showed higher catalytic efficiency. The characterization of MAS NMR, XPS, XAS, and HAADF-STEM images revealed that the introduced Rh species had no effect on the structural stability of zeolite. The isolated Rh&+ (&\u00a0= 2.5) was successfully confined to the molecular sieve structure. The researchers explored the reaction mechanism of hydroformylation on the catalyst through density functional theory (DFT) calculation. The results show that the combination of \u03b1-C atom on 1-hexene and C atom on CO is the rate-determining step (RDS) of the reaction. The energy barrier of RDS for straight chain aldehydes is obviously lower than that of RDS for branched chain aldehydes, which is consistent with the experimental results. In addition, the calculation results further revealed the dynamic change information of the active Rh site in the reaction, which confirmed the space limitation of the molecular sieve and the stabilizing effect of the skeleton oxygen atom on the active Rh site in the catalytic system.Carbon materials have adjustable physical and chemical characteristics, adequate pore size distribution, and appropriate pH, which can improve the high dispersion of active components and accelerate the diffusion of reactants and products. Therefore, carbon materials, of which activated carbon is the most common, offer enormous potential in catalytic reactions.\n90\n Ligands play a major decisive role in the reactivity of Rh metal, especially in the selectivity. Due to the limitations of carbon material itself, it is difficult to build a stable bond structure between carbon material and organic phosphine ligand, and it is easy to lose phosphine ligand in the reaction, which will lead to the reduction of selectivity and the loss of Rh atoms, and ultimately affect the life of the catalyst. Feng et\u00a0al.\n91\n constructed Rh SACs (Rh1/AC) on activated carbon for the carbonylation of methanol. The activity of Rh1/AC was three times than that of homogeneous catalyst ([Rh(CO)2I2]-) and 30 times than that of carbon-supported Rh nanoparticle catalyst. The HAADF-STEM diagram of Rh1/AC showed that isolated Rh atoms were dispersed on the surface of support before or after the reaction. DFT calculation and differential charge density calculation showed that the carbonyl group with electron donor properties were the optimal anchoring sites of Rh atoms, and the coordination bonds enhanced the electron density of the central Rh atoms and reduced the energy barrier of the speed control step.MOFs are crystals with adjustable pore structure, large specific surface, high porosity, and stable multi-dimensional network structure generated by the coordination and hybridization of multi-dentate organic ligands with transition metal ions. Selecting an appropriate metal precursor and confining it to the MOF pore through the nanoconfinement effect is the efficient method to achieve atomic-level dispersion of metal. The main problem of using MOFs as support is that the pore size of the support itself will affect the activity and selectivity of the catalyst, just like molecular sieves.\n92\n\n,\n\n93\n\n,\n\n94\n So far, Rh-based nanocluster catalysts combined with MOFs have been applied in the hydroformylation of alkenes, such as Rh@MIL-101,\n95\n Rh@IRMOF-3,\n96\n Rh@ZIF-8,\n97\n and Rh/MnMOF.\n98\n This paved the way for the continued development of atomically distributed catalysts based on MOFs.Phosphines-modified transition metals are widely employed in the hydroformylation of olefin, due to the high activity and selectivity under mild conditions. When coordinating with transition metals, P atoms can provide lone pair electrons to the empty tracks of transition metal to form \u03c3-bond; at the same time,\u00a0they also accept the filled d-orbital feedback electronic of metal atoms to form feedback \u03c0-bond. The \u03c3-donor and \u03c0-acceptor properties of phosphines can regulate the performance of transition metal\u00a0catalysts in organic reaction.\n99\n Compared to ligands composed of other elements of the same main group, phosphine ligands have the highest catalytic activity in hydroformylation (activity sequence: PPh3\u00a0>\u00a0NPh3\u00a0>\u00a0NP3\u00a0>\u00a0AsPh3, SbPh3\u00a0>\u00a0BiPh3).There are two or more P atoms in bi-/multi-dentate phosphines to coordinate and chelate with the transition metal to form a more stable catalyst active intermediate and avoid the deactivation. A bi-dentate phosphine ligand vinyl-biphephos functionalized by vinyl was copolymerized with vinyl monomer through solvothermal synthesis method reported by Li et\u00a0al.\n43\n The prepared corresponding Rh catalyst Rh/CPOL-1bp&10P showed higher regioselectivity (L/B\u00a0>\u00a024) and activity (TOF\u00a0>\u00a01200 h-1) in propylene hydroformylation of fixed-bed reactor in comparison to Rh/POL-PPh3, Rh/POL-dppe, Rh/CPOL-bp&DVB, Rh/CPOL-bp&dppe, and Rh-biphephos/SiO2 (Figures\u00a07A and 7B). The metal Rh atomically dispersing on the porous polymer support achieved high reactivity of catalysts according to results of EXAFS and HAADF-STEM. Moreover, the same group synthesized Xantphos-doped Rh/POPs-PPh3 via the copolymerization of vinyl-Xantphos and vinyl-PPh3 (Figure\u00a07C).\n44\n Although the lower conversion of 1-octene was obtained in Xantphos-doped Rh/POPs-PPh3 system, the selectivity and regioselectivity of the target product were significantly superior to that of Rh/POPs-PPh3.In addition to phosphine ligands, some other ligands can also be used to regulate the hydroformylation of olefins in SACs. Yuan et\u00a0al.\n100\n synthesized the hydrophilic catalyst (Rh1/PIPs) through alkalization, polymerization, impregnation, and other steps. When CO feed contained 1000 ppm H2S, the hydrocarboxylation of olefin was facilitated unexpectedly. The characterization of HAADF-STEM (Figures\u00a07D and 7E) and EXAFS demonstrated that Rh existed as single atom in the Rh1/PIPs. Ex situ EXAFS and in situ DRIFTS revealed a ternary cycle mechanism of olefin hydrocarboxylation reactions (Figure\u00a07F). The authors used CO and CO-H2S (1000 ppm H2S) as probe molecules, and added mixed liquids (cyclohexene, water, iodomethane, and other reactants and auxiliaries) in the form of bubbles to perform in situ DRIFTS (Figure\u00a07G). The findings demonstrated that, in contrast to Rh-H bonds in CO systems, Rh-H bonds in CO- H2S systems exhibit a red-shift, which is attributable to the coordination of the strong electron ligand S species with Rh atoms. DFT calculation confirmed that the energy barrier of each step can be reduced with the addition of H2S, including the speed control step. This work offered a sulfur-resistant strategy for the carbonylation reactions, and advanced the theory of SACs in heterogeneous catalysis.In the catalytic process, the formation and breakage of chemical bonds of substrates, intermediates, or products on the catalyst surface are the results of the interactions between reactant molecules and metal atomic orbitals. Besides the properties of metal elements, the electronic structure of active metals is also influenced by metal size, supports, and the coordination environment of surface atoms. Wei et\u00a0al.\n41\n developed a high-performance Rh SACs (Rh-Co-Pi/ZnO) by adding heteroatoms to regulate the microenvironment of active metals (Figure\u00a08\nA). In addition, the HAADF-STEM image (Figure\u00a08B) of Rh-Co-Pi/ZnO clearly showed that the introduction of Pi greatly significantly increased the dispersion degree of Rh atoms. According to the characterization of CO-DRIFTs (Figure\u00a08C) and XPS, the presence of Co atoms reduced the electron density around Rh and impaired the interaction of Rh-CO. Compared to Rh/ZnO system, the selectivity for linear aldehydes increased from 32.1% to 54.9%, the L/B ratio increased from 0.7 to 2.1, in Rh-Co-Pi/ZnO-catalyzed hydroformylation of 1-decene (Figure\u00a08D). Inductively coupled plasma optical emission spectrometer showed that 94.1% of Rh, 96.4% of Co, and 95.9% of P remained in the recycled Rh-Co-Pi/ZnO. The well thermally stability and recyclability of Rh-Co-Pi/ZnO was reused for five cycles without noticeably decline of catalytic activity.In previous studies, ionic liquids (ILs) have been shown to be effective in protecting and stabilizing nano- and homogeneous catalysts. Ding et\u00a0al.\n101\n found that ILs can increase the activation energy of single atoms aggregation and adjust the oxidation valence state of metal atoms, and first proposed a simple and universal strategy of stabilizing SACs. In 2021, the team extended this strategy to Rh SACs to investigate the effect of ILs on the stability of styrene hydroformylation on Rh1/TiO2 (Figure\u00a08E). After five cycles of reaction, the TOF of unmodified Rh1/TiO2 decreased from 1250 h-1 to only 10 h-1, and the loading of Rh decreased from 0.1% to 0.05%. The initial TOF of the IL-stabilized catalyst was lower than that of Rh1/TiO2, but its stability was significantly increased. In particular, the TOF value of 1-(2-hydroxyethyl)-3-methylimidazolium bis (trifluoromethanesulfonyl) imide ([OHEmim][Tf2N])-stabilized Rh1/TiO2 only decreased from 878 to 800 h-1, and the loading of remained at about 0.1%. DFT calculation showed that ILs can increase the binding energy from 0.69 to 1.19 eV by acting as the linker between Rh atoms and TiO2, thus improving its anti-leaching performance.There are few studies on the mechanism of hydroformylation catalyzed by SACs; many models and inferences need to be further explored for verification. Lee et\u00a0al.\n36\n studied the influence of ReOx promoter and the mechanism of hydroformylation of ethylene catalyzed by atomically dispersed Rh-ReOx/\u03b3-Al2O3 (Figure\u00a09\nA). The synergistic effect of atomically dispersed Rh-ReOx with \u03b3-Al2O3 was revealed using DFT calculations and microkinetic modeling. In contrast to the typical Wilkinson\u2019s catalyst, the stable Rh(CO)2 precursor exhibited a 16-electron square planar structure by coordinating with two oxygen atoms on the surface. The RDS of hydroformylation depended on the local surroundings of Rh atoms. In the absence of ReOx, the rate was controlled by the CO insertion step; in the presence of ReOx, the Rh-CO coordination bond weakened, and CO coordination was the rate-controlling step. Meanwhile, ReOx improved the selectivity of propanal by blocking the main hydrogenation pathway through steric effects.Rh/POL-PPh3 synthesized by Jiang et\u00a0al.\n85\n showed excellent performance and ultra-high stability in fixed bed of olefins hydroformylation, which was due to the similar catalytic function to that of homogeneous catalyst HRh(CO)(PPh3)3 (Figure\u00a09B). Ma et\u00a0al.\n102\n revealed the dual role of the polymer 3V-PPh3 monomer as both support and ligand in Rh/3V-PPh3 catalyzing hydroformylation of ethylene through quantum chemistry method. Compared with PPh3 as support, the adsorption energy of Rh atoms on 3V-PPh3 increased, indicating that the introduction of vinyl increased the Rh-P bond interaction. Secondly, the high density of P atoms (Rh: P\u00a0= 1:3) exposed on supports helped to improve the dispersion degree of Rh atoms, to increase the energy barrier formed by Rh-Rh bond, finally to improve the stability of catalyst. In this paper, the dual action mechanism of 3V-PPh3 as support and ligand in Rh/3V-PPh3 is studied, and the reason why Rh atoms are not easy to lose is explained from a microscopic perspective, which is helpful to understand the relationship between microstructure and electronic effect, and provides theoretical guidance for the development and design of efficient heterogeneous catalyst.Gao et\u00a0al.\n52\n used DFT calculation to explore the regioselective mechanism of Rh1/PNP-ND-catalyzed hydroformylation of styrene (Figure\u00a09C). The Rh1/PNP-ND model was established based on the coordination of one Rh atom with two P atoms on the similar single-layer graphene. After optimization, a square planar structure was obtained under reaction conditions, with the coordination of one Rh, two P, one H, and one CO. It is generally recognized that the step of olefin insertion is crucial for determining the regioselectivity. The reaction barrier of the formation of branched aldehydes (0.69 eV) was significantly lower than that of linear aldehydes (0.74 eV). After calculation, the ratio of the relative rates of branched aldehydes and linear aldehydes was 5.75. The ratio of two products was predicted to be 85:15, which was consistent with the experimental data. Further thermodynamic analysis revealed that, starting from the same alkene coordination state, the \u0394G of the branched alkyl complex and linear alkyl complex is -3.36 and -1.21\u00a0kcal/mol, respectively. In conclusion, the coordination environment of Rh atoms in Rh1PNP-ND was favorable for the formation of branched chain products in both thermodynamics and dynamics.Wang et\u00a0al.\n51\n thoroughly investigated the significant performance of 0.2% Rh/CoO in hydroformylation of propylene combining DFT calculation with characterization data (Figure\u00a09D). The DRIFT spectrum of 0.2% Rh/CoO demonstrated that the adsorption of propylene was greatly enhanced when exposed to H2 and CO atmosphere. The binding energy of Rh 3d in the mixed atmosphere (H2, CO, and propylene) was 1.3 eV lower than that without any gas treatment, which was obviously larger than the bias generated by the catalyst exposed to H2 or CO atmosphere, according to the XPS spectrum of 0.2% Rh/CoO. It can be inferred that Rh atoms in Rh/CoO undergo structural reconstruction during the catalytic process, which promoted the adsorption and activation of reactants. Based on the calculation of DFT, four stable co-adsorption configurations were proposed, designated as configurations \u2160, \u2161, \u2162, and IV. The relative positions of the H atoms and adjacent unsaturated C atoms helped to conclude that configurations \u2160 and \u2162 tend\u00a0to form linear products, whereas configurations \u2161 and IV tend to form branched products. Further investigation of the paths of configurations \u2160 and \u2161 showed that the last step of product formation was the rate-limiting step with the highest energy barrier. The energy barrier of the rate-limiting step in configuration I is 0.063 eV lower than that in configuration II, indicating the favorable formation of linear products.Wei et\u00a0al.\n37\n created an effective and stable Co SACs (Co/\u03b2-Mo2C) by utilizing the potent electronic metal\u2013support interaction (EMSI) effects between Co atoms and \u03b2-Mo2C. This catalyst outperformed all previously reported heterogenetic Co-based catalyst in hydroformylation of propylene (TOF up to 749 h-1). After repeated use for five times, the activity of catalyst did not decline noticeably. Compared to the bulk Co particles, Co/\u03b2-Mo2C with single Co atom has better activity for hydroformylation of olefin, and the TOF value was increased by 8.7 times. Based on the characterization of XPS, Baber charge, charge density, and Co density of states, it can be seen that the EMSI effect between single Co atoms and the support tailored the electronic properties of metal to be positively charged Co2+ and reduced the electronic density of Co. According to DFT calculations, the electron deficient property of the Co atoms contributed to CO insertion, thereby increasing the activity of hydroformylation of olefins (Figure\u00a09E). For Co-based hydroformylation, the significant EMSI interaction between single-atom Co and the support in Co/\u03b2-Mo2C was crucial in optimizing the charge density, lowering the reaction potential energy, and stabilizing the active site.Insoo Ro et\u00a0al.\n38\n reported heterogeneous Rh-WO\nx\n pair site catalysts for ethylene hydroformylation. Two active sites (Rh atoms and WO\nx\n species deposited on Al2O3) worked together to catalyze various stages of the reaction. The structure of catalyst can be altered by varying the loading of WO\nx\n on the support, which regulated the catalytic activity of ethylene hydroformylation in turn (Figure\u00a09F). According to HAADF-STEM (Figure\u00a09G) and CO-Fourier transform infrared spectrometer characterization, Rh and WO\nx\n were located in Rh/0.7W as independent sites and were also forming Rh-W pair sites. Due to the synergistic interaction between the active sites on Rh and WO\nx\n, Rh/0.7W possessed the highest activity (0.1 gpropanal cm-3h-1) and selectivity (\u00a0>\u00a095%). A bifunctional mechanism was proposed based on the experimental kinetics and First-principles microdynamics simulations (Figure\u00a09H). Rh assisted WO\nx\n reduction, which binded ethylene molecules; ethylene was transferred from WO\nx\n to Rh; H2 dissociated at the Rh-WO\nx\n interface to form two\u00a0hydrogen atoms, one of which binded to the Rh-WO\nx\n interface. The bi-functional catalyst also depended on the geometry of the Rh-WO\nx\n interface, the energetics of reconfiguring the coordination of the pair site during the reaction, and the capacity to transfer molecules between the active centers of the pair site.In SACs, active metal is loaded on the surface of the support in the form of a single atom, and the requirements for characterization methods of SACs have also reached unprecedented atomic-level accuracy. In recent years, the rapid development of electron microscopy and spectroscopy technology has provided support for analyzing the spatial distribution, electronic structure, and coordination environment of metal centers, and provided reliable evidence for exploring the catalytic performance, structure-activity relationship, and catalytic mechanism. The characterization techniques applicable to SACs include HAADF-STEM, XAS, and CO-DRIFT.HAADF-STEM improves the measurement accuracy to atomic level, and can clearly observe isolated metal atoms and their spatial distribution on the support, becoming the most direct means to characterize SACs.\n103\n It provides strong evidence for further understanding the mechanism of atomic catalytic reaction and identifying the coordination structure of metal center on the support. The brightness of the atoms in the image is proportional to the square of the atomic number, so as to distinguish between heavy atoms (such as Pd, Pt, Ru, Rh, Co, etc.) and light atoms (such as N, O, C, etc.).\n104\n\n,\n\n105\n In the prepared Rh1/HAP, Li et\u00a0al.,\n53\n using HAADF-STEM, can clearly observe that the isolated Rh atoms are evenly distributed on the HAP (Figure\u00a010\nA). Shang et\u00a0al.\n39\n observed atomically dispersed Rh species distributed in molecular sieve using Cs-HADDF-STEM. Due to the difference in atomic contrast (Si\u00a0= 14, O\u00a0= 8, Al\u00a0= 13, and Rh\u00a0= 45), the brightest spot in the image can be identified as the Rh atoms (Figure\u00a010B).X-ray absorption spectroscopy (XAS) is used to measure the structure of X-ray absorption coefficient varying with energy. The sample excites its core electrons to transition to the empty orbit by absorbing X-ray (XANES) or transition to continuous state to form wave dry radiation with surrounding atoms (EXAFS). The chemical valence state and electronic structure of elements can be obtained from XANES, and the two-dimensional local structure information of adjacent atoms can be obtained from EXAFS. Therefore, XAS is widely used to study the structural model of active sites and explore the mechanism of monatomic catalysis.\n27\n Ding et\u00a0al.\n101\n used XAS to characterize the synthesized atomically dispersed Rh1/TiO2. The K-edge XANES of Rh was studied with Rh foil and Rh2O3 as reference materials. Compared with the standard samples, the energy absorption curve of Rh1/TiO2 was very close to that of Rh2O3, indicating that the average oxidation state of Rh was close to 3+ (Figure\u00a010C). According to EXAFS, Rh displayed a dominant peak at around 1.6\u00a0\u00c5, which was assigned to the Rh-O first shell. There was no obvious peak at 2.3\u00a0\u00c5, which was attributable to Rh-Rh scattering (Figure\u00a010D).The probe molecule infrared spectroscopy shows different vibration frequencies for metal atoms in different chemical environments, which is an effective mean to characterize the dispersion state and electronic state of metal particles in supported catalysts. In the CO-DRIFT spectrum, the adsorption form of CO on metal can be judged according to the position of CO adsorption peak, and then the dispersion state of metal particles can be determined.\n45\n\n,\n\n106\n\n,\n\n107\n\n,\n\n108\n Li et\u00a0al.\n45\n used CO-DRIFT technology to identify the existence of Rh loaded on CeO2. For Rh1/CeO2, the positions of infrared absorption peaks of CO were 2010 and 2052\u00a0cm-1, which are attributed to the symmetric and asymmetric vibration of gem-dicarbonyl doublet CO on positively charged Rh atoms (Figure\u00a010E). A peak centered at 2052\u00a0cm-1 was also observed which corresponds to the linear CO adsorption on Rh atoms. For NP-Rh/CeO2 and 5Rh/CeO2, the positions of the infrared adsorption peaks of CO were 1860 or 1800\u00a0cm-1, and 2046 and 1960\u00a0cm-1, which respectively correspond to the bridge adsorption between two Pt atoms, the linear adsorption of CO molecules on the surface of Rh atoms and the adsorption at the interface (Figure\u00a010F).DFT calculation is often combined with relevant experiments to further explore the reaction mechanism by\u00a0studying the properties of catalytic materials (such as bond length, adsorption energy, etc.). DFT calculation in catalyst research mainly starts from the following four aspects: structural stability judgment, reaction free energy calculation, electronic structure analysis, and molecular diffusion/adsorption dynamics simulation. It is helpful to predict catalyst structure and stability, evaluate catalyst performance, innovate catalyst design strategies, and finally achieve SACs with high activity, high selectivity, and strong\u00a0stability.\n109\n\n,\n\n110\n\n,\n\n111\n Wei et\u00a0al.\n40\n clarified the reaction mechanism of propylene hydroformylation on the catalyst through DFT calculation. Firstly, the most reasonable model of Au(0.2%)@S-1 is determined, that is, a single Au atom replaces the Si atom at the T8 site on the crystal S-1 zeolite. Secondly, the adsorption energy of H, CO, and propylene is calculated on the Au(0.2%)@S-1 model. Among them, the adsorption capacity of H is the strongest (2.19 eV), followed by propylene (1.18 eV) and CO (1.27 eV), indicating that the adsorption capacity of propylene on the Au(0.2%)@S-1 model is moderate, which is conducive to the adsorption and desorption of reactants on the active center. It is also calculated that the adsorption energies of CH3CH2CH2\n\u2217 and CH3CH2CH2CO\u2217 are much more negative than those of propylene, indicating that the olefin insertion and CO insertion reactions are thermodynamically favorable. Based on the above analysis, a possible mechanism of propylene hydroformylation on Au(0.2%)@S-1 is proposed (Figure\u00a010G).It is difficult to separate and detect the free radicals and intermediates in the chemical reaction process,\u00a0which makes it difficult to speculate the reaction mechanism. In situ Raman, in situ XPS, isotope\u00a0labeling, and other technologies can monitor the dynamic evolution of catalysts and reaction intermediates in real time under experimental conditions, which helps to accurately understand the structure of catalysts and build theoretical models, making outstanding contributions to the design of various effective catalysts.In this paper, the catalytic application and reaction mechanism of SACs in hydroformylation of olefins are\u00a0summarized. The effects of microstructure regulation on catalytic activity, chemical/regioselectivity, and stability are discussed. The strategies of support effect, ligand effect, and electronic effect are proposed to adjust the performance of SACs. Advanced characterization techniques HADDF-STEM,\u00a0XAFs, and DFT calculations are used to further study the mechanism. Although the application of SACs in hydroformylation is still in its infancy, SACs have already shown excellent performance. Existing\u00a0research demonstrates that the SACs, notably the Rh SACs, have distinctive electronic/coordination structure, high atom utilization, unsaturated active center coordination, and tunable central metal electronic structure; the above characteristics make its catalytic activity equal to or even better than that of homogeneous catalyst. More importantly, the strong coordination between active metals and supports can effectively avoid the loss of Rh, which provides a new direction for the development of heterogeneous hydroformylation.Despite the significant development, there are still a dearth of pertinent studies and numerous pressing issues that need to be resolved.\n\n(1)\nIn contrast to nanocatalyst and cluster catalyst, the active metal in SACs is atomic dispersion on the support, and the metal surface energy in SACs increases sharply. In the process of preparation and reaction, metal atoms are easy to migrate and agglomerate, which lead to the instability and deactivation of the catalyst. Therefore, the key to the synthesis of catalyst is to select an appropriate support. The coordination between the defect sites on the support surface and the single metal\u00a0atom to prevent the agglomeration phenomenon can not only stabilize the single metal atom but also expose the active sites of the metal. The atomic dispersion of metal precursors can be achieved by means of space limitation, defect capture, and ligand anchoring, which can effectively limit the migration and aggregation of monodisperse metal atoms on the support. In addition, SACs cannot provide multiple adjacent metal sites, and its metallicity is often regulated\u00a0by the support. Therefore, when multiple metal active sites are required to be activated cooperatively and active metals are required to have strong metallicity for catalytic reactions, SACs\u00a0are difficult to achieve efficient catalytic activity. Fully exposed cluster catalysts (FECCs) can\u00a0not only provide adjacent metal active sites but also partially maintain its metallicity on the basis\u00a0of 100% metal dispersion. Metals in FECCs are mainly composed of very small clusters, and all atoms in the clusters are in the state of coordination unsaturated. FECCs have been widely used in alkane dehydrogenation, toluene hydrogenation, CO2 reduction, LGWS, and other reactions, and become an important field in heterogeneous catalysis. FECCs, as an extension of the concept of SACs, can well solve the problem of single active site in SACs, which makes it possible to efficiently carry out multi-step and complex catalytic reaction systems. As a conceptual extension of SACs, FECCs can solve the problem of single active sites in SACs, and provide a new way to design efficient catalysts.\n\n\n(2)\nIn order to fill the defect that SACs have a single metal center and low loading, a second metal is introduced to synthesize dual-atom-site catalysts (DASCs) and nano-single-atom-site catalysts (NSASCs). As a further extension of the concept of SACs, DACs/NSASACs achieve low-cost, high selectivity, high stability, and antitoxicity catalysts. They retain the advantages of SACs, and introduce a variety of interactions, such as synergistic effect, geometric effect, and electronic effect. With the diversity of metal atoms in DACs/NSASACs, it is of great significance to save precious metal resources, reduce production costs, and realize industrial applications. So far, the reported DACs/NSASACs have been successfully applied to hydrogen evolution reaction, O2 reduction/evolution reaction, N2 reduction reaction, CO oxidation reaction, and other catalytic fields. However, how to control the structure of diatomic sites, improve the density of catalytic sites, and reveal the synergistic effect between atomic sites and the structure-activity relationship of catalysts through accurate structural characterization or theoretical calculation is still a major challenge.\n\n\n(3)\nMost of the reported catalyst supports with remarkable performance are limited to metal oxides and porous polymers. The synthesis of SACs using metal oxide as the support has the advantages of simple synthesis process and straightforward catalyst model, which is conducive to exploring the reaction mechanism of olefin hydroformylation. In addition, the absence of phosphine ligand is extremely valuable for environmental preservation. However, the surface modification of this catalyst is limited, and the poor regioselectivity is difficult to reach the level of homogeneous catalyst. Due to the diversity of synthesis methods of porous polymers, mono/multidentate phosphine ligands can be modified into porous polymer materials. The SACs supported by this method have excellent catalytic activity and selectivity for hydroformylation. The high density of phosphine ligand can avoid the loss of Rh and improve the stability of catalyst. However, the complex synthesis of porous polymer materials, the expensive ligand, and the poor mechanical strength seriously limit the mass production of catalysts.\n\n\n(4)\nAs the products of hydroformylation, aldehydes are high value-added intermediates that can be converted into amines, alcohols, or acetals by further reactions. The one-pot tandem hydroformylation-hydrogenation reaction, hydroformylation-adol condensation reaction, hydroformylation-acetalization reaction, and hydroformylation-reductive amination reaction are economical methods to obtain above productions. At present, SACs or even supported catalysts are rarely reported in this area, which calls for more investigation.\n\n\n(5)\nHydroformylation with synthesis gas as raw material has become the mainstream of modern chemical industry for its mature process, low cost, and suitability. However, due to the high toxicity and explosiveness of syngas, researchers are committed to studying green and efficient alternatives, such as HCHO, CO2, HCOOH, aldehydes, etc. Among them, the hydroformylation of HCHO as raw material has achieved good progress in reactivity and regioselectivity, and HCHO is cheap and easy to obtain, convenient for storage, transportation, and atmospheric pressure application. Therefore, HCHO is a promising substitute for synthesis of gas. CO2 is a clean, low-cost, and abundant raw material. However, the inert carbon-oxygen bond in CO2 makes it difficult to add metal-activated species, resulting in the poor selectivity of target products. The utilization of CO2 in hydroformylation is still in the laboratory research and development stage. The hydroformylation of olefins using these syngas substitutes often requires the modification of precious metal catalysts with complex phosphine ligands, which leads to high production costs, so it is still a long way for the industrial application.\n\n\n(6)\nThe prepared SACs are still in the early stages of fundamental research, with the defects such as poor thermal stability, high metal surface energy, and low active metal loading. Therefore, there are still great challenges in industrial production. The development of high stability and applicability of SACs is crucial for meeting the demands of industrial applications. The macroscopic preparation of SACs is the long-term pursuit and the most challenging ultimate goal in hydroformylation.\n\n\n(7)\nFurther research is still required to fully understand the catalytic mechanism of SACs in hydroformylation, including the catalytic active species, reaction mechanism, and inactivation process. The synthesis and regulation of particular catalysts from the atomic scale can be realized by the means of in situ electron microscopy, in situ synchrotron radiation, and other contemporary characterization technologies. Combined with DFT calculations, the catalyst structure and reaction pathway can be simulated. Above methods provide a crucial scientific foundation for explaining the structure-activity relationship of SACs.\n\n\nIn contrast to nanocatalyst and cluster catalyst, the active metal in SACs is atomic dispersion on the support, and the metal surface energy in SACs increases sharply. In the process of preparation and reaction, metal atoms are easy to migrate and agglomerate, which lead to the instability and deactivation of the catalyst. Therefore, the key to the synthesis of catalyst is to select an appropriate support. The coordination between the defect sites on the support surface and the single metal\u00a0atom to prevent the agglomeration phenomenon can not only stabilize the single metal atom but also expose the active sites of the metal. The atomic dispersion of metal precursors can be achieved by means of space limitation, defect capture, and ligand anchoring, which can effectively limit the migration and aggregation of monodisperse metal atoms on the support. In addition, SACs cannot provide multiple adjacent metal sites, and its metallicity is often regulated\u00a0by the support. Therefore, when multiple metal active sites are required to be activated cooperatively and active metals are required to have strong metallicity for catalytic reactions, SACs\u00a0are difficult to achieve efficient catalytic activity. Fully exposed cluster catalysts (FECCs) can\u00a0not only provide adjacent metal active sites but also partially maintain its metallicity on the basis\u00a0of 100% metal dispersion. Metals in FECCs are mainly composed of very small clusters, and all atoms in the clusters are in the state of coordination unsaturated. FECCs have been widely used in alkane dehydrogenation, toluene hydrogenation, CO2 reduction, LGWS, and other reactions, and become an important field in heterogeneous catalysis. FECCs, as an extension of the concept of SACs, can well solve the problem of single active site in SACs, which makes it possible to efficiently carry out multi-step and complex catalytic reaction systems. As a conceptual extension of SACs, FECCs can solve the problem of single active sites in SACs, and provide a new way to design efficient catalysts.In order to fill the defect that SACs have a single metal center and low loading, a second metal is introduced to synthesize dual-atom-site catalysts (DASCs) and nano-single-atom-site catalysts (NSASCs). As a further extension of the concept of SACs, DACs/NSASACs achieve low-cost, high selectivity, high stability, and antitoxicity catalysts. They retain the advantages of SACs, and introduce a variety of interactions, such as synergistic effect, geometric effect, and electronic effect. With the diversity of metal atoms in DACs/NSASACs, it is of great significance to save precious metal resources, reduce production costs, and realize industrial applications. So far, the reported DACs/NSASACs have been successfully applied to hydrogen evolution reaction, O2 reduction/evolution reaction, N2 reduction reaction, CO oxidation reaction, and other catalytic fields. However, how to control the structure of diatomic sites, improve the density of catalytic sites, and reveal the synergistic effect between atomic sites and the structure-activity relationship of catalysts through accurate structural characterization or theoretical calculation is still a major challenge.Most of the reported catalyst supports with remarkable performance are limited to metal oxides and porous polymers. The synthesis of SACs using metal oxide as the support has the advantages of simple synthesis process and straightforward catalyst model, which is conducive to exploring the reaction mechanism of olefin hydroformylation. In addition, the absence of phosphine ligand is extremely valuable for environmental preservation. However, the surface modification of this catalyst is limited, and the poor regioselectivity is difficult to reach the level of homogeneous catalyst. Due to the diversity of synthesis methods of porous polymers, mono/multidentate phosphine ligands can be modified into porous polymer materials. The SACs supported by this method have excellent catalytic activity and selectivity for hydroformylation. The high density of phosphine ligand can avoid the loss of Rh and improve the stability of catalyst. However, the complex synthesis of porous polymer materials, the expensive ligand, and the poor mechanical strength seriously limit the mass production of catalysts.As the products of hydroformylation, aldehydes are high value-added intermediates that can be converted into amines, alcohols, or acetals by further reactions. The one-pot tandem hydroformylation-hydrogenation reaction, hydroformylation-adol condensation reaction, hydroformylation-acetalization reaction, and hydroformylation-reductive amination reaction are economical methods to obtain above productions. At present, SACs or even supported catalysts are rarely reported in this area, which calls for more investigation.Hydroformylation with synthesis gas as raw material has become the mainstream of modern chemical industry for its mature process, low cost, and suitability. However, due to the high toxicity and explosiveness of syngas, researchers are committed to studying green and efficient alternatives, such as HCHO, CO2, HCOOH, aldehydes, etc. Among them, the hydroformylation of HCHO as raw material has achieved good progress in reactivity and regioselectivity, and HCHO is cheap and easy to obtain, convenient for storage, transportation, and atmospheric pressure application. Therefore, HCHO is a promising substitute for synthesis of gas. CO2 is a clean, low-cost, and abundant raw material. However, the inert carbon-oxygen bond in CO2 makes it difficult to add metal-activated species, resulting in the poor selectivity of target products. The utilization of CO2 in hydroformylation is still in the laboratory research and development stage. The hydroformylation of olefins using these syngas substitutes often requires the modification of precious metal catalysts with complex phosphine ligands, which leads to high production costs, so it is still a long way for the industrial application.The prepared SACs are still in the early stages of fundamental research, with the defects such as poor thermal stability, high metal surface energy, and low active metal loading. Therefore, there are still great challenges in industrial production. The development of high stability and applicability of SACs is crucial for meeting the demands of industrial applications. The macroscopic preparation of SACs is the long-term pursuit and the most challenging ultimate goal in hydroformylation.Further research is still required to fully understand the catalytic mechanism of SACs in hydroformylation, including the catalytic active species, reaction mechanism, and inactivation process. The synthesis and regulation of particular catalysts from the atomic scale can be realized by the means of in situ electron microscopy, in situ synchrotron radiation, and other contemporary characterization technologies. Combined with DFT calculations, the catalyst structure and reaction pathway can be simulated. Above methods provide a crucial scientific foundation for explaining the structure-activity relationship of SACs.This work was supported by National Natural Science Foundation of China (Nos. 22108306, 22102214), Taishan Scholars Program of Shandong Province (No. tsqn201909065), Shandong Provincial Natural Science Foundation (Nos. ZR2021YQ15, ZR2020QB174), and the Fundamental Research Funds for the Central Universities (No. 22CX07009A).Writing - Original Draft, Conceptualization, S.T. and D.Y.; Review\u00a0& Editing, M.W., G.X., W.W., and Y.Z.; Writing - Review\u00a0& Editing, Supervision, Funding acquisition, Y.P.The authors declare no competing interests.", "descript": "\n                  Hydroformylation is one of the most significant homogeneous reactions. Compared with homogeneous catalysts, heterogeneous catalysts are easy to be separated from the system. However, heterogeneous catalysis faces the problems of low activity and poor chemical/regional selectivity. Therefore, there are theoretical and practical significance to develop efficient heterogeneous catalysts. SACs can be widely applied in hydroformylation in the future, due to the high atom utilization efficiency, stable active sites, easy separation, and recovery. In this review, the recent advances of SACs for hydroformylation are summarized. The regulation of microstructure affected on the reactivity, stability of SACs, and chem/regioselectivity of SACs for hydroformylation are discussed. The support effect, ligand effect, and electron effect on the performance of SACs are proposed, and the catalytic mechanism of SACs is elaborated. Finally, we summarize the current challenges in this field, and propose the design and research ideas of SACs for hydroformylation of olefins.\n               "}
{"full_text": "Polyurethanes are one of the most versatile polymer materials in the world and can be used for a wide range of end-user applications, such as furniture, coatings, adhesives, building materials, fibers, cushions, paints, elastomers and synthetic leathers [1,2]. Methylene diphenyl diisocyanate (MDI) is one of the main raw materials to produce polyurethane. MDI is prepared by phosgenation of methylene diphenyl diamine (MDA). MDA is synthesized by the reaction of aniline and formaldehyde in the presence of hydrochloric acid catalyst. The purity of aniline determines the quality of MDI and affect polyurethane production [3]. Therefore, aniline synthesis is an inevitable part of the ecosystem of polyurethane industry. Nitrobenzene reduction is a widely used process for aniline production. Mostly, various oxides (silicates, alumina) and carbon-supported (carbon nanotubes, graphene, carbon black) catalysts are used with palladium, platinum, or rhodium metals in nitrobenzene reduction [4\u201310]. Other special carbon containing systems including different carbide composites (e.g. MXene, Ti3C2(OHxF1-x)2), defect-rich nitrogen-doped reduced graphene oxide (RGO) and its chitosan combined hydrogel form are also promising catalyst supports for the hydrogenation of nitro-compounds, such as 2-nitrophenol, 4-nitrophenol [11\u201314].Recently, magnetic nanocatalysts have received a lot of attention and are widely used in catalytic processes due to their many advantages (high specific surface area, good dispersibility, efficient separation due to their magnetic properties, reusability) [15\u201318]. Several research groups have synthesized ferrite nanoparticles as catalyst supports and have effectively used them for the reduction of nitroaromatic compounds [19\u201321]. Ferrites are spinel transition metal oxides with the general formula of MFe2O4 (M is a transition metal such as Fe, Ni, Mn, and Zn). Several methods have been successfully applied in the synthesis of ferrite nanoparticles such as co-precipitation [22], sol\u2013gel [23], microemulsion [24], hydrothermal [25], thermolysis [26], and mechanical alloying [27]. A relatively new technique for the synthesis of ferrite nanoparticles is based on sonochemical treatment, which includes the exposure of the reaction mixture to intense ultrasonic irradiation [28\u201330]. Sound waves entering the liquid medium create high as well as low pressure cycles, depending on the frequency. During the low\u2013pressure cycle, high-intensity ultrasonic waves create small vacuum bubbles, or rather cavities in the liquid. When the bubble reaches a volume that is no longer able to absorb energy, it collapses during the high\u2013pressure cycle, consequently \u201cbreaking\u201d the solid particles in the liquid. This phenomenon is called acoustic cavitation, during which the released energy can cover the needs of certain chemical reactions [31].In the current work, ferrite supported Pd catalysts were prepared by combining sonochemical and combustion methods. The prepared catalysts are in active form after the final step and do not require post-treatment, and thus, a simplified catalyst preparation is achieved. Due to their magnetic properties, the produced spinel ferrite catalysts can be easily recovered from the liquid phase by magnetic separation using a magnetic field.To synthesize the spinel catalyst supports, zinc(II) nitrate\u00a0hexahydrate (Zn(NO3)2 \u2219 6H2O, ThermoFisher GmbH, 76870 Kandel, Germany), nickel(II) nitrate hexahydrate (Ni(NO3)2 \u2219 6H2O, ThermoFisher GmbH, 76870 Kandel, Germany), and iron(III) nitrate nonahydrate (Fe(NO3)3 \u2219 9H2O, VWR Int. LtD., B-3001 Leuven, Belgium) were applied. Polyethylene glycol (PEG 400, Mw: \u223c400\u00a0g\u00a0mol\u22121, Molar Chemicals Ltd., Budapest, Hungary) was used as reducing agent and dispersion media of the metal precursors. During the catalyst preparation step, palladium(II) nitrate dehydrate (Pd(NO3)2 \u2219\u00a02H2O, Merck Ltd., Darmstadt, Germany) was applied as precursor of the catalytic active metal, and patosolv, a mixture of aliphatic alcohols (90 vol% ethanol and 10 vol% isopropanol, Molar Cemicals Ltd., Budapest, Hungary) was used as reducing agent to carry out the conversion of Pd(II) ions to elemental Pd.To prepare the ferrite samples and decompose the palladium nanoparticles on the surface of the ferrite supports, Hielscher UIP100 Hdt. tip homogenizer (1000W, 20\u00a0kHz) was applied. Bs4d22 ultrasonic block sonotrode (D\u00a0=\u00a022\u00a0mm) was used to initiate the formation of metal hydroxides in polyethylene glycol dispersion. The spinel nanoparticles were examined by high-resolution transmission electron microscopy (HRTEM, FEI Technai G2 electron microscope, 200\u00a0kV) and their morphology has been characterized. The samples were prepared by dropping aqueous suspension of the nanoparticles on 300 mesh copper grids (Ted Pella Inc.). X-ray diffraction (XRD) measurements were used by Rietveld analysis to identify and quantitatively characterize the different oxide phases in the samples. Bruker D8 Advance diffractometer (Cu-K\u03b1 source, 40\u00a0kV and 40\u00a0mA) in parallel beam geometry (G\u00f6bel mirror) with Vantec detector was applied. Average crystallite size of the domains was calculated by the mean column length calibrated method by using of full width at half maximum (FWHM) and the width of the Lorentzian component of the fitted profiles. The quantity of the deposited palladium in the catalysts have been analyzed by Varian 720\u00a0ES inductively coupled optical emission spectrometer (ICP-OES). For the ICP-OES measurements, the samples have been solved in aqua regia. The specific surface area (SSA) measurements of the samples were carried out by nitrogen adsorption\u2013desorption method at 77\u00a0K. For this, the Micromeritics ASAP 2020 equipment was used, and the evaluation was carried based on the Brauner-Emmett-Teller (BET) method. The ferrite samples were examined by applying a Vario Macro CHNS element analyzer to quantify the carbon content. Certified phenanthrene (C: 93.538%, H: 5.629%, N: 0.179%, S: 0.453%; from Carlo Erba Inc.) was used as standard. The carrier gas was helium (99.9990%) while oxygen (99.995%) was used for oxidation of the carbon content. The quantitative analysis of the samples after hydrogenation tests was carried out by Agilent 7890A gas chromatograph coupled with Agilent 5975C Mass Selective detector with RTX-624 column (60\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm x 1.4\u00a0\u03bcm). The injected sample volume was 1\u00a0\u03bcL\u00a0at 200:1 split ratio, while the inlet temperature was set to 473\u00a0K. Helium was the carrier gas with constant flow (2.28\u00a0mL/min), and the oven temperature was set to 323\u00a0K for 3\u00a0min and after it was heated up to 523K with a heating rate of 10\u00a0K/min and kept there for another 3\u00a0min. The analytical standards of the main product, the by-products, and intermediates purchased from Sigma Aldrich and Dr. Ehrenstorfer Ltd.Zinc and nickel containing ferrite spinel nanopowders were synthesized by using a two-step process including ultrasonic cavitation and combustion and thus, nickel ferrite, zinc ferrite, and zinc doped nickel ferrite nanoparticles were achieved. In the first step, iron(III) nitrate nonahydrate and one of the precursors (Table 1\n) were dissolved in 20\u00a0g polyethylene glycol (PEG 400). The solutions were sonicated by using a Hielscher UIP1000 Hdt tip homogenizer for 5\u00a0min (130\u00a0W, 23\u00a0kHz). Due to the high energy caused by the ultrasonic cavitation, in the presence of polyol, a brownish red, high viscosity colloid system was formed which contained hydroxide and oxide nanoparticles. Thereafter, the total polyol content of the system was burned, and the oxides and hydroxides of the transition metals converted to ferrite nanoparticles with water outlet.The prepared nickel ferrite, zinc ferrite, and zinc doped nickel ferrite (4.00\u00a0g) supports were dispersed in ethanolic solution of palladium nitrate dihydrate (0.50\u00a0g). The dispersions were sonicated by using the high energy ultrasound homogenizer (130\u00a0W) for 4\u00a0min. During this process, the Pd(II) ions were reduced to elemental palladium nanoparticles due to the released energy caused by the cavitation, and Pd nanoparticles were deposited onto the surface of the magnetic spinel supports which led to the formation of the final catalysts (Pd/NiFe2O4, Pd/ZnFe2O4, and Pd/NiZnFeO4). Then, the catalyst samples were removed from the dispersion with a Nd magnet, washed with patosolv, and dried at 105\u00a0\u00b0C overnight. The final palladium content of the magnetic catalysts was determined by ICP-OES measurements.The catalysts (0.10\u00a0g) were tested in nitrobenzene hydrogenation. To carry out the test, a methanolic solution of nitrobenzene (c\u00a0=\u00a00.25\u00a0mol\u00a0dm\u22123) was used in a B\u00fcchi Uster Picoclave reactor (200\u00a0ml volume) (Fig.\u00a01\n). The pressure of hydrogenation was constant (20\u00a0bar) during the tests, and the reactions were carried out at four different reaction temperatures (283\u00a0K, 293\u00a0K, 303\u00a0K, and 323\u00a0K). Rotational speed of agitation was 1000\u00a0rpm. Sampling took place after the beginning of the reaction at 0, 5, 10, 15, 20, 30, 40, 60, 80, 120, 180, and 240\u00a0min.The catalytic activity of the palladium decorated ferrite catalysts was determined by calculating nitrobenzene conversion (X%) based on the following equation (Eq.(1)):\n\n(1)\n\n\nX\n\n%\n=\n\n\nc\no\nn\ns\nu\nm\ne\nd\n\n\nn\n\nn\ni\nt\nr\no\nb\ne\nn\nz\ne\nn\ne\n\n\n\n\ni\nn\ni\nt\ni\na\nl\n\n\nn\n\nn\ni\nt\nr\no\nb\ne\nn\nz\ne\nn\ne\n\n\n\n\n\n\u00d7\n\n100\n\n\n\n\n\nFurther, the yield (Y%) of aniline was calculated as follows (Eq. (2)):\n\n(2)\n\n\nY\n\n%\n=\n\n\n\nn\n\na\nn\ni\nl\ni\nn\ne\n\n\n\nn\n\nn\ni\nt\nr\no\nb\ne\nn\nz\ne\nn\ne\n\n\n\n\n\n\u00d7\n\n100\n\n\n\nwhere \nn\n\n\naniline\n and \nn\n\n\nnitrobenzene\n are the corresponding chemical amounts of the compounds.To determine the catalytic efficiency selectivity (S%) is very important parameter, which was calculated as follows (Eq. (3)):\n\n(3)\n\n\nS\n%\n=\n\n\nn\n\na\nn\ni\nl\ni\nn\ne\n\n\n\n\u03a3\n\nn\n\np\nr\no\nd\nu\nc\nt\ns\n\n\n\n\n\u00d7\n100\n\n\n\n\nThe phase composition of the synthetized ferrites was examined with XRD measurements by using the Rietveld analysis (Fig.\u00a02\n). Next to the ferrite phases, oxide forms of the transition metals were identified. On the diffractogram of the nickel ferrite sample, there are peaks at 30.3\u00b0, 35.7\u00b0, 43.4\u00b0, 53.8\u00b0, 57.4\u00b0, and 62.9\u00b0 two Theta degrees, and these can be associated with the (220), (311), (400), (422), (511), and (440) reflections of the NiFe2O4 phase (PDF 10\u20130325) which was the target product (Fig.\u00a02 A). In the nanopowder, NiO was also found next to the NiFe2O4, indicated by the peaks at 37.3\u00b0 (111), 43.4\u00b0 (200), two theta degrees (PDF 47\u20131049). Moreover, awaruite (FeNi3) was also identified in the nickel ferrite sample as peaks at 44.1\u00b0 (111), and 51.3\u00b0 (200) 2\u0398 degrees (PDF 38\u2013419) were located.In case of the zinc ferrite support, reflections which belongs to the ZnFe2O4 phase were identified at 18.2\u00b0 (111), 29.9\u00b0 (220), 35.3\u00b0 (311), 36.9\u00b0 (222), 42.8\u00b0 (400), 53.1\u00b0 (422), 56.6\u00b0 (511), and 62.2\u00b0 (440) two Theta degrees (Fig.\u00a02 B), (PDF 22\u20131012). Furthermore, a zinc(II) oxide phase was also identified at 31.8\u00b0 (100), 34.4\u00b0 (002), and 36.3\u00b0 (101) 2\u0398 degrees (PDF 36\u20131451).The zinc doped nickel ferrite support was also characterized and reflections of the NiZnFeO4 phase were located at 30.0\u00b0 (220), 35.5\u00b0 (311), 43.1\u00b0 (400) 53.4\u00b0 (422), 56.9\u00b0 (511), and 62.5\u00b0 (400) two Theta degrees (Fig.\u00a02C). During the formation of the zinc doped nickel ferrite spinel, by-products, ZnO, NiO, and FeNi3, were also formed, which were identified on the diffractograms of the other two support, the nickel ferrite and zinc ferrite. The reflections of these by-products are visible on the deconvoluted diffractogram of NiZnFeO4.Quantitative analysis of the different phases in the three ferrite samples was carried (Table 1). In case of the zinc ferrite support, the content of the spinel phases was the highest 94.7\u00a0wt% compared to the other two ferrite samples. In this case, ZnO was also formed in relatively low quantities (5.3\u00a0wt%). Presence of the non-magnetic phases, as zinc oxide, nickel oxide may cause a problem in the magnetic separation of the catalyst unless it adsorbs well on the magnetic particles or forms a stable aggregate with it. On other hand, these oxides can also play a role in the catalytic behavior of the spinel samples. Since the production of the samples includes a combustion step, it was assumed that the burning of polyethylene glycol was not complete, and thus, carbon forms can also remain in the samples. To verify this, CHNS element analysis was carried out, which confirmed that the samples also contain carbon in low quantities (<1\u00a0wt%) (Table 2\n). It has to be noted, that minimizing carbon content is also important, because the remaining carbon can cover the spinel particles, and thus, the supports\u2019 promoter effect cannot fully prevail during the catalytic hydrogenation processes.The size of the particles in the support samples was also determined by applying the Scherrer equation and using the full width at half maximum (FWHM) intensity of the reflexion peaks (Table 2). The mean particle sizes of the different nanoparticles are between 10 and 25\u00a0nm.The small size of the particles can be explained by the applied PEG400 dispersant, which prevented aggregation of the particles during the sonochemical step. To get a more detailed picture about the formation of the nanoparticles, additional samples were prepared for which the synthetic procedure of the supports was not carried out completely. After the ultrasonic treatment the processes were stopped, and the formed nanoparticles were washed with distilled water and were separated by centrifugation. The drying of the solid phases was carried by lyophilization (at 213\u00a0K and 1.0\u00a0mbar vacuum). XRD measurements were carried out on the separated nanoparticles to examine the solid phases after the sonication step. Metal oxide and oxyhydroxide structures such as nickel oxide hydroxide (Ni3O2(OH)4, PDF 06\u20130144), franklinite (ZnFe2O4, PDF 22\u20131012), zincite (ZnO, PDF 36\u20131451), maghemite (Fe2O3, PDF 39\u20131346), and trevorite (NiFe2O4, PDF 10\u20130325) were identified which formed from the nitrate salts of the precursors (SI Fig.\u00a01). Bunsenite (NiO) and FeNi3 was not identified after in samples and thus, these phases are formed only after the combustion step (Fig.\u00a02 A and C). Furthermore, the results shows that the nickel oxide hydroxide phase was eliminated by the combustion step, possible by water elimination due to the high temperature, which reacted with the zincite or maghemite and formed the spinel phases.The spinel particles were examined by transmission electron microscopy (Fig.\u00a03\n). On the HRTEM images carbon layers are visible, which cover the surface of the metal oxide nanoparticles. To further analyze the carbon layers and identify the surface functional groups, Fourier transform infrared (FTIR) spectroscopy measurements were carried.On the FTIR spectrum of the nickel ferrite sample two main broad bands at 424\u00a0cm\u22121 and 601\u00a0cm\u22121 wavenumbers were identified which can be associated with the \u03bdFe-O stretching vibration of the tetrahedral metal\u2013oxygen bond and the metal\u2013oxygen vibrations in the octahedral sites (Fig.\u00a04\n). A weak vibration band was located around 1062\u00a0cm\u22121 and can be assigned to the \u03bdC\u2212O, which originates from the carbon content remained after the combustion. Although the carbon content only 0.6\u00a0wt% (Table 2) in this case, it is also detectable by the symmetric and asymmetric stretching vibration modes of the \u2013CH2 bonds, which absorbs at 2852\u00a0cm\u22121 and 2924\u00a0cm\u22121. The presence of hydroxyl groups was also verified as a band was identified at 1386\u00a0cm\u22121. The hydroxyl functional groups may belong to the carbon or to the surface of the metal oxide nanoparticles. The presence of hydroxyl functional groups on the surface of the developed catalytic ferrites improves their polar feature and thus, their wettability and dispersibility in polar solvents, such as methanol (SI Fig.\u00a02). Two bands of the hydroxyl groups are located at 3438\u00a0cm\u22121 and 1640\u00a0cm\u22121, which can be attributed to the stretching modes of the H\u2013O\u2013H stretching and bending vibrations of free or absorbed water. There are only slight differences in the spectra of the other two support samples, ZnFe2O4 and NiZnFeO4, compared to NiFe2O4.Specific surface areas (SSA) of the ferrite supported palladium catalysts have been measured and it was found that the SSA of the Pd/NiFe2O4 is almost three times higher (64.0\u00a0m2\u00a0g\u22121) than in case of the Pd/ZnFe2O4 and Pd/NiZNFeO4 catalysts where it is only 22.6\u00a0m2\u00a0g\u22121 and 22.2\u00a0m2\u00a0g\u22121, respectively.On the XRD pattern of the palladium decorated spinel catalysts, reflections at 40.0\u00b0, 46.6\u00b0, and 68.2\u00b0 two Theta degrees were located which can be associated with Pd(111), Pd (200), and Pd (220) (ICDD card number 046\u20131043), respectively (Fig.\u00a05\n A, B and C). This indicates that the palladium is in the elemental state in the catalytic systems. The efficient ultrasonic treatment of the palladium(II) nitrate precursor which was carried out in alcoholic phase led to formation of elemental palladium particles. On the Rietveld refined diffractograms palladium oxide is not detectable, which shows that the reduction step was successful, and the total amount of palladium ions converted to elemental Pd. The palladium content of the ferrite supported catalysts were measured by ICP-OES analysis. Based on the spectroscopic results, the palladium content of the catalysts was the highest, 4.20\u00a0wt% in case of Pd/ZnFe2O4, which was followed by Pd/NiFe2O4 (4.10\u00a0wt%), and NiZnFe2O4 (3.82\u00a0wt%).The particle size of palladium was calculated based on the XRD results. The mean particle size of the Pd nanoparticles in the prepared catalysts were small, 5\u00b12\u00a0nm for Pd/NiFe2O4, and 5\u00b11\u00a0nm for Pd/ZnFe2O4, while a slight increase is experienced (6\u00a0\u00b1\u00a02\u00a0nm) in case of Pd/NiZnFeO4. The particle size of the phases presents in the catalysts (e.g. supports) were not changed significantly after the deposition of palladium nanoparticles (Table 3\n and SI Table 1).The individual particles cannot be distinguished from each other on the HRTEM images of the palladium decorated ferrite catalysts. The palladium nanoparticles are not identifiable next to the other metal oxide particles (SI Fig.\u00a03). On the TEM images, significant change in the particle morphology cannot be detected, and aggregates or new structures cannot be identified in the samples.The magnetic catalyst supports were tested also in nitrobenzene hydrogenation. After 3\u00a0h only 23%, 29.2%, and 22.8% nitrobenzene conversions were achieved by using the NiFe2O4, ZnFe2O4, and NiZnFeO4 supports, respectively (SI Fig.\u00a04). The aniline yields were also low (18.8 n/n%, 21.2 n/n% and 21.6 n/n%) after 3\u00a0h. Thus, the activity of the ferrite nanoparticles is not adequate for aniline synthesis. The catalysts were also tested, and their activity was compared. Pd/NiFe2O4 showed the lowest catalytic activity, after 3\u00a0h 92.8 n/n% nitrobenzene conversion was achieved at 323\u00a0K (Fig.\u00a06\nA). In case of the Pd/ZnFe2O4 and Pd/NiZnFeO4, 99.8 n/n% conversion was reached after only 2\u00a0h, and 99.0 n/n% after only 80\u00a0min hydrogenation at 323\u00a0K. Thus, by the presence of Zn in the support, the activity of the catalysts increased.Aniline yield (Y%) and selectivity (S%) were calculated in case of each catalyst applied at different temperatures (Fig.\u00a07\n A and B). After 3\u00a0h of hydrogenation, the highest yield (99.2 n/n%) was achieved by the Pd/ZnFe2O4 catalyst at 323\u00a0K, while the corresponding selectivity was also high, 98.8 n/n%. In general, all three catalysts showed high selectivity for aniline formation at all temperatures (Fig.\u00a07 B), but the yield was lower when the Pd/NiFe2O4 catalyst was applied.During the catalytic tests, several intermediates such as nitrosobenzene (NOB), azoxybenzene (AOB), and azobenzene (AZB) have been identified in each case (Fig.\u00a08\n). However, these species converted to aniline at end of the reaction, so the yield was not affected by them. Furthermore, by-products were also formed, but only in very small quantities, and thus, the selectivity of aniline was not significantly impaired.By applying Pd/NiFe2O4 and Pd/NiZnFeO4, N-methylaniline (NMA) formed as a by-product, but only in concentration <2\u00a0mmol\u00a0dm\u22123. NMA can be formed directly from aniline with methylation in the presence of methanol (solvent). By using Pd/ZnFe2O4 as catalyst, two more by-products (<1\u00a0mmol\u00a0dm\u22123) were observed, dicyclohexylamine (DCHA) and N-methyl-1-phenylethanimine (NMPE). It is interesting to note that, all catalyst contained the same catalytically active metal (palladium), nonetheless by using the Pd/ZnFe2O4 sample additional by-products were formed. This phenomenon cannot be explained by only the effect of palladium. Moreover, the catalysts with Zn in their support, also contained a ZnO phase (5\u00a0wt%), but DCHA and NMPE were formed only when Pd/ZnFe2O4 was applied. Thus, the presence of zinc oxide does not justify the emergence of new molecules either. In this sense, it can be assumed, that the zinc ferrite opened a new reaction pathway towards the formation of the imine and DCHA.The highest aniline yield and selectivity was achieved by using the Pd/ZnFe2O4 catalyst. Thus, this catalyst was tested in reuse tests and at 323\u00a0K 4 cycles of nitrobenzene hydrogenation was carried out. After 3\u00a0h reaction time the aniline yield and selectivity values were calculated in each cycle. Significant decrease in the yield was not experienced during the first three cycles. However, in the 4th cycle, the aniline yield dropped to 67% (SI Fig.\u00a05 B). In contrast, the selectivity remained similar during the reuse cycles and were between 98 and 99 n/n%. The palladium content of the Pd/ZnFe2O4 catalyst was determined after the 4th cycle, and it was found that the initial palladium content decreased from 4.20\u00a0wt% to 1.8\u00a0wt%. This palladium loss led to a decrease in the catalytic activity during the reuse tests. Based in this, it will be necessary to further improve the stability of the catalyst, to create an stronger interaction between the noble metal and the support.All in all, three highly selective, easy to use, magnetic catalysts (Pd/NiFe2O4, Pd/ZnFe2O4, and Pd/NiZnFeO4) were successfully prepared, and their applicability was tested in nitrobenzene hydrogenation. Based on the results, Pd/ZnFe2O4 catalyst was found to be the most efficient for aniline synthesis. First, the nickel ferrite, zinc ferrite and zinc doped nickel ferrite spinel nanoparticles were synthesized. These nanoparticles show high dispersibility, due to their surface functional groups (-OH) and their small mean particle sizes (21\u00a0\u00b1\u00a05\u00a0nm, 17\u00a0\u00b1\u00a04\u00a0nm, and 20\u00a0\u00b1\u00a05\u00a0nm, for NiFe2O4, ZnFe2O4, and NiZnFeO4, respectively). Therefore, they are excellent support materials and were applied in catalyst preparation. Palladium nanoparticles were deposited onto the surface of the magnetic supports, by using ultrasonic cavitation which assisted the adsorption and reduction of the Pd. After the process, the catalyst is ready to use, and further activation step is not necessary. In this sense, the applied catalyst preparation method is fast and simple. By using the Pd/ZnFe2O4 and Pd/NiZnFeO4 catalyst, high aniline selectivity (99.8 and 99.9 n/n%) and yield (99.2 and 92.8 n/n%) were achieved, while with the nickel ferrite supported catalyst, lower selectivity and yield was reached. Reuse tests were further verified the applicability of the prepared catalyst in the hydrogenation of nitro compounds, but the improvement its stability is necessary.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was supported by the European Union and the Hungarian State, co-financed by the European Regional Development Fund in the framework of the GINOP-2.3.4-15-2016-00004 project, aimed to promote the cooperation between the higher education and the industry. Prepared with the professional support of the Doctoral Student Scholarship Program of the Co-operative Doctoral Program of the Ministry of Innovation and Technology financed from the National Research, Development and Innovation Fund. Further financial support has been provided by the National Research, Development and Innovation Fund (Hungary) within the TKP2021-NVA-14 project.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.jmrt.2022.06.113.", "descript": "\n                  Aniline is one of the most important chemical in the polyurethane industry and it is produced by the catalytic hydrogenation of nitrobenzene. The development of novel, multifunctional catalysts, which are easily recoverable from the reaction mixture is therefore of paramount importance. Transition metal-containing ferrites decorated with palladium were prepared by using a combination of sonochemical and combustion steps. First of all, magnetic ferrites were produced to be used in catalyst preparation as supports for palladium nanoparticles. On the surface of the ferrite particles palladium nanoparticles were deposited by applying ultrahigh sonication in an alcoholic phase. All in all, three magnetic catalysts, Pd/NiFe2O4, Pd/ZnFe2O4, and Pd/NiZnFeO4 have been created. The catalysts have been tested and their activity have been compared in nitrobenzene hydrogenation to synthesize aniline at four different temperatures, and 20\u00a0bar pressure. The most active catalysts were the Pd/ZnFe2O4 and Pd/NiZnFeO4 systems with which aniline yields of 99.2 and 92.8 n/n% were achieved after 3\u00a0h of hydrogenation, respectively. In contrast, by applying the Pd/NiFe2O4 catalyst, a significantly lower aniline yield was achieved. It was proved that, due to their magnetic properties, the prepared catalysts are easily removable from the reaction medium by using a magnetic field. Thus, catalysts with excellent properties have been successfully developed and tested in nitrobenzene hydrogenation.\n               "}
{"full_text": "Data will be made available on request.According to the low reactivity and high corrosion resistance of platinum, the global demand for platinum is increasing. Platinum has catalytic properties and hydrogen adsorption and is used in the manufacture of catalysts for the refining, petrochemical and automotive industries. Industrial wastes including platinum are considered secondary sources of this metal. (Diederen, 2009; Morcali, 2020 [1,2]; Yakoumis et al., 2020). Despite the use of platinum in catalysts, especially automobile catalysts, high amounts of soil and water contamination with platinum group metals have been observed in the tests of soil and running water along the roads. Also, the recovery of platinum reduces hazardous wastes and proper disposal of residues [3\u20135]. Based on the use of platinum in high technologies, this metal is in the near-critical category [6,7]; Henckens et al., 2014; Sverdrup et al., 2017). South Africa and Russia are the world's largest mining suppliers of platinum. According to the report of the platinum group metals market by Johnson-Matthew; the supply of platinum from recycling is about one-fifth of primary sources. The supply and demand of platinum are shown (Fig. 1\n).The main uses of platinum are in the chemical, electrical, glass industry, petroleum refining, medical/biomedical and other industries. Platinum is mostly used in the production of chemicals such as industrial catalysts. Platinum catalysts have different applications in industries including increasing the octane number in reforming units, producing hydrocarbons from synthesis gas, reducing toxic gases in the automotive industry, oxidizing ammonia and converting it to nitrogen oxide or nitric acid, catalytic reforming, etc. [8]. Consumption rates in these industries over the past years are compared (Fig. 2\n).South Africa, as the major mining supplier of platinum, has always insisted on producing this metal from the mine and primary sources. But in consuming countries, the recovery of this metal from secondary sources is important. Therefore companies like Johnson Matthey (UK and USA), BASF (USA), and Umicore (Belgium), tend to recover platinum from secondary sources (Saguru et al., 2018). Concentrations of platinum in the secondary sources are higher than natural ores. Platinum from natural ores requires 18, 860\u2013254, 860 MJ per kg of metal and 100,000\u20131,200,000 m3 of water per ton of metal extracted, while recycled platinum needs 1400\u20133400 MJ and 3000\u20136000 m3, respectively. So platinum recovery is cost-effective (Comisi\u00f3n Europea, 2018). One of the largest sources of platinum extraction from secondary sources is industrial spent catalysts. Based on the value of platinum and its amount in the waste of spent catalysts, its extraction and recovery have been considered. [9\u201311]; Zanjani & Baghalha, 2009). To separate and purify platinum from spent catalysts, the catalysts need to be crushed and prepared. Pyrometallurgical and hydrometallurgical methods are used to separate platinum. The important issue in the latest research is attention to the economic aspect (such as the amount of energy consumption), environmental constraints and standards. The innovation of this research is that proposes the optimal parameters for the pre-treatment and leaching stages in the hydrometallurgical method for platinum recovery from spent catalyst.A catalyst increases the rate of reaction but remains unchanged at the end of the reaction. Catalysts make processes more economical and environmentally friendly [12]; Trimm, 2008). The selectivity and activity of catalysts decrease after a period of cycle life and are so-called deactivated. Deactivated catalysts are regenerated and returned to the process. If it is not possible to regenerate the catalyst, it will be unusable and become hazardous industrial waste. Spent catalysts in the refining, petrochemical and automotive industries contain precious metals: platinum, aluminum, iron, nickel and several other (Table 1\n).One of the most important goals in selecting catalysts is their cycle life. Catalyst consumption and spent catalyst production are reduced with longer cycle life. Catalysts are deactivated for a variety of reasons which is categorized into six intrinsic mechanisms in the following cases: 1. Poisoning of active sites, 2. Fouling, coking, and carbon deposition, 3. Thermal degradation and sintering of the catalyst, 4. Vapor compound formation and/or leaching accompanied by transport from the catalyst surface or particle, 5. Vapor\u2013solid and/or solid-solid reactions, and 6. Attrition/crushing [12,17\u201319]; Marafi et al., 2017; Zhou et al., 2020). Various catalysts are used in petrochemical industries, the main metals in these catalysts are cobalt, molybdenum, nickel, platinum and tungsten. The surface of this group of catalysts is deactivated for three main reasons, which are: coking, sintering and poisoning [18]. In Table 2\n, the amount of contamination of each catalyst is given according to the deactivation factors. In Platinum spent catalysts, the coking deposit has the highest rate of catalyst deactivation factor. Sintering is reduced due to the presence of chloride in the catalyst. But the moisture in the reactant stream can wash away the chloride in the catalyst and increase sintering. For this reason, the inlet stream must be completely free of moisture [12].According to the importance of supplying platinum from secondary sources, the recovery and extraction of this metal from spent catalysts in oil, petrochemical, automotive, and pharmaceutical are considered [20]; Naghavi et al., 2016; Shams & Goodarzi, 2006; Yoo, 1998). The following steps are required for metal recovery are: recyclability of material composition, availability of related compounds, affordable, safe waste collection system, complete the recovery chain steps, use optimal set-up for recovery chain and build enough capacity to the end of the chain [21]. Pyrometallurgical and hydrometallurgical methods are mainly used to recover precious metals from primary/secondary sources. The pretreatment steps such as heating and crushing are necessary to remove excess compounds from the industrial spent catalyst. Finally, the platinum is separated and purified. The choice of method for processing of metal recovery depends on the purity and final value of the product [10,11,22\u201324]; Yakoumis et al., 2021). State-of-the-art, recovery methods focus not only on maximum recovery but also on economic and environmental priorities (Yakoumis et al., 2021). The use of renewables and safer solvents, less waste generation, lower energy, hazardous chemicals, space, time consumption need to be considered to achieve economic and environmental goals. The recovery method is selected based on the number and type of metals in the spent catalyst, the concentration of metals, the basic nature of the catalyst, the conditions and materials and equipment required. Due to high energy consumption, production of polluting gases and low purity products in the pyrometallurgical method, this study will explain the hydrometallurgical method for the recovery of platinum. Pre-treatment is required before the hydrometallurgical method that increases the leaching and recovery efficiency. Therefore, the pre-treatment will also be explained. The general schematic of the aforementioned methods for platinum recovery is shown in Fig. 3\n.In pyrometallurgy, there are problems such as the production of polluting gases and high energy consumption. According to new environmental regulations and restrictions, the use of new and alternative methods of pyrometallurgy has been considered. Hydrometallurgy techniques will fill the gap in the industry [8]. The pyrometallurgical method requires industrial equipment to supply high temperatures, which may not be suitable for some small-scale recovery units. BASF (New Jersey, USA), Johnson Matthey (2 refineries in England, 1 in USA and 1 in China), Multi Metco (Alabama, USA), Umicore (Hoboken, Belgium), Stillwater (Colorado, USA), and NonX21 (Quebec, Canada) are some pyrometallurgical refineries in the world (Saguru et al., 2018). The pyrometallurgical process for the recovery of platinum from the spent catalyst involves three methods: smelting, vaporization, and sintering (Peng et al., 2017). Crushed spent catalysts are mixed with various collectors such as lead, copper, iron, matte, or PCB (printed circuit board) for helping platinum recovery. Then, this mixture melted at a furnace (plasma furnace, electric arc furnace, or inductive furnace) at a temperature above 1000 \u2e30C to become low-viscosity liquids (Morcali, 2020; Fern\u00e1ndez et al., 2021; Peng et al., 2017; Yakoumis et al., 2021). After the smelting process, the metal vaporizes and purifies. Platinum chloro-complexes are formed with the chlorination process. Separation of chloro-complexes is performed by volatilization or adsorption on an activated carbon bed [25]. Sintering is based on the chemical-physical properties of materials. In this method, with the help of plasma, the amount of oxide of platinum compounds is reduced at high temperatures (>1200 \u2e30C). Sintering and volatilization are more limited methods than smelting (Peng et al., 2017). The most important factor for recovering platinum from spent catalysts is temperature and the retention time on process temperature [23]. A summary of the three methods is given in Table 3\n.In the overview of the pyrometallurgical process, spent catalysts are mixed with fluxes and collectors. This mixture enters the furnace according to the aforementioned methods. The outlet of the furnace is purified to extract platinum. An overview of the pyrometallurgical process is shown in Fig. 4\n.Pre-treatment is used before the hydrometallurgical process to remove surface contaminants and increase leaching efficiency by improving chemical attack (Saguru et al., 2018; Yakoumis et al., 2021). Crushing and particle size smoothing of the spent catalyst is also required to improve the leaching process (Yakoumis et al., 2020). Preheating or thermal pretreatment is one of the pretreatment methods (Yuliusman et al., 2020). This method uses heating in an environmentally friendly atmosphere such as air, oxygen, nitrogen, and hydrogen [26]. Of course, it depends on the type of contamination on the spent catalysts [8]. According to the high oxidation potential of platinum (Marinho et al., 2010), its dissolution in the acidic leaching process is difficult. The use of various agents and the formation of chloro-complexes reduce the oxidation potential. Calcination and crushing of spent catalysts into micron-sized particles are also pre-treatment methods [16,27]; Zanjani & Baghalha, 2009). Fig. 5\n shows an overview of the pretreatment process.In most studies, calcination and crushing and homogenization of spent catalysts have been considered as a pretreatment method to the metallurgical process. However, in some cases, heating and calcination have not been performed (Yakoumis et al., 2020). The spent catalysts are oxidized in a furnace at 500\u00a0\u00b0C for 5h (Marinho et al., 2010), crushing into 10\u2013100 \u03bcm particles and calcination at 800\u00a0\u00b0C [16], drying at 900\u00a0\u00b0C for 3 hour and broking down into 300 \u03bcm (Yuliusman et al., 2020), calcination for 30 min at 600\u00a0\u00b0C and crushing into of 500 \u03bcm [27], for increasing the efficiency of the platinum leaching. Pretreatment at high temperature for increasing the efficiency of the platinum leaching is given in Table 4\n.Reduce energy consumption with drying in an oven at 110\u00a0\u00b0C and crushed into 106 \u03bcm (Zanjani & Baghalha, 2009), crushing to 200 mesh and drying at 120\u00a0\u00b0C for 24 hour [28], crushing into 100-\u03bcm and drying in an oven [29], crushing into 0.3 mm and drying by 8% H2, at 100\u00a0\u00b0C for 20 hour [26], crushing to 200 mesh [30], crushing to 0.16 mm (Lanaridi et al., 2021), crushing to 2 \u03bcm (Atia et al., 2021), just crushing (Hasani et al., 2015), using formic acid to improvement of leaching (Equation (1)) (Upadhyay et al., 2013), that has more applications in industries. Pretreatment at minimum energy consumption for increasing the efficiency of leaching is given in Table 5\n.\n\n(1)\n\n\nP\nt\n\nO\n\n(\nS\n)\n\n\n+\nH\nC\nO\nO\nH\n\u2192\nP\nt\n+\n\nH\n2\n\nO\n+\n\n\nC\nO\n\n2\n\n\n\n\n\nThe hydrometallurgical process has practical and economic advantages compared to the pyrometallurgical process. These advantages include lower energy consumption, achievement of higher purity materials, lower process temperature, easier process control(Saguru et al., 2018), applicable on large and small scales, and less toxic gas wastes [29]; Lanaridi et al., 2021). The steps of the hydrometallurgical process are explained in Fig. 6\n. In the process of hydrometallurgy, first the metal is separated and leached in different ways and purified based on the required purity of the product [30]. After the leaching step with acid, amine family solvents were used to extract selectively and purify platinum group metals. Iron is always present in solutions obtained from leaching of the platinum group metals, but iron and aluminum do not interfere in the extraction of platinum with solutions of amines [24]; Sun & Lee, 2011). According to the dissolution of aluminum in the leaching process, the least dissolution of aluminum is always the optimal method for leaching (M\u00e9ndez et al., 2021). On the laboratory scale, precious metals are separated from leaching solutions by precipitation, adsorption of activated carbon, ion exchange, and solvent extraction. But solvent extraction has also been used for industrial scales. So far, some ionic liquids have been used in solvent extraction for laboratory scales (Xing & Lee, 2019). In the present work, leaching and purification steps are considered as a subset of the hydrometallurgical process. Leaching, microwave leaching, and bio-leaching are mentioned in detail. The purification steps also include precipitation, solvent extraction, ion exchange, and ionic liquid extraction, which are described at the same time as the leaching steps. Precipitation and solvent extraction methods have been the most widely used in recent research. The use of ionic liquids or any other solvent on industrial scale requires economic estimation (Fig. 6).Spent catalysts leached after pretreatment. At this stage, platinum is extracted from the waste catalyst and then purified. The recovery rate of Platinum has increased by increasing acid concentration and temperature. The level of acidity in the acidic leaching process is always high, which causes environmental hazards and the consumption of large amounts of acidic agents. In the latest studies, non-volatile chloride salts are used instead of HCl in acid leaching that increases the dissolution of platinum and decreases the dissolution of catalyst base. Aluminum chloride is a good choice for oxidizing agents due to its 3 chloride ions per molecule. Increasing H2O2 accelerates the reaction by reducing the potential required for the reaction. The effect of different operating parameters such as temperature, solid to liquid ratio, time was investigated and optimized. Homogeneous spent catalyst particles have been used in direct and single-step leaching with 3\u00a0M HCl, 4.5 M NaCl, 1% v/v H2O2, at 70\u00a0\u00b0C, S/L of 0.7, for 2 hr. 100% platinum is recovered in this leaching with low acidity (Yakoumis et al., 2020). At a low concentration of HCL of about 0.66 M, 75% of platinum is recovered in I2 30 g/l, L/S of 20 g/g, at 75 \u2e30C and for 60 min (Zanjani & Baghalha, 2009). For reducing the hazardous effect of HCl, aluminum chloride with a low concentration of nitric acid was used as the oxidizing agent. Aluminum chloride is a good choice for oxidizing agents due to its 3 chloride ions per molecule. Platinum is recovered at 96.8\u201398.8% at 103\u00a0\u00b0C, for 1 hr, 2 M aluminum chloride, and 1 M nitric acid [28]. Aqua regia (HCl\u00a0+\u00a0HNO3, 3:1 v/v) was used for leaching of oxidized spent catalyst at 75\u00a0\u00b0C, for 20\u201325 min. 99\u00a0wt% of platinum is purified with 15\u00a0vol% Aliquat 336 in kerosene at 25\u00a0\u00b0C (Marinho et al., 2010). 98.1% of platinum is recovered by using the response surface methodology when leaching experiments were performed with HCl of 1.45 M, NaCl of 4.55 M, 10% H2O2/spent catalysts of 0.66 mL/g, and L/S of 4.85:1 at 90\u00a0\u00b0C for 2 h [16]. More than 80% of platinum is recovered by 11.6 M HCl leaching solution with 1% vol of H2O2, at 60 \u2e30C and 1 hr. Using 8 M HCl/2 M CaCl2 combination instead of HCl under the same conditions has similar responses (M\u00e9ndez et al., 2021). The leaching of catalyst particles is done with 15 mL HCl 37% and 5 mL HNO3 65%, L/S of 8, and heated at 109\u00a0\u00b0C for 3h. In this process, platinum is completely recovered [29]. In leaching with 1 M oxalic acid, 5.58% of platinum is extracted at 60\u00a0\u00b0C for 10 hr. 19.72% of Platinum is recovered under similar conditions to the UOP method 896\u2013930 and leaching with aqua regia (HCl: HNO3 3: 1) at S/L of 20, at 300\u00a0\u00b0C and time depends on the rate of evaporation and boiling of solution (Yuliusman et al., 2020). The use of HCl/H2O2 in the leaching process is environmentally friendly. 96% of platinum is separated with 0.8\u00a0vol% H2O2, and 9.0 M HCl at 60\u00a0\u00b0C for 2.5 hour [26]. In leaching with sulfuric acid, the base of the catalyst also dissolves. Dissolution in sulfuric acid for the Catalyst based on \u03b3-Al2O3 (such as R-134) is higher than the Catalyst based on a mixture of \u03b3-Al2O3 and \u03b1-Al2O3 (such as AR-405). According to the dissolution of the base of the catalyst, the separation of the target metal becomes more difficult. Catalyst particles are dissolved in 6 M\u00a0H2SO4, at 100 \u2e30C for 4 hours to leach alumina. 90% of aluminum sulfate crystals are formed after adding distilled water to the leaching solution. 52% of Platinum in AR-405 and 83% of Platinum in R-134 are dissolved in solution [27]. Leaching of spent catalysts in 60% H2SO4, 0.1 M NaCl, S/L of 1/100, at 135\u00a0\u00b0C after 2 hour has been investigated to avoid the use of aqua regia that platinum recovery was 95% [30]. The reaction of hydrochloric acid and nitric acid with platinum is given in equations 2\u2013(4.\n\n(2)\n\n\nP\nt\n+\n\n\nC\nl\n\n\n3\n\n\n(\n\na\nq\n\n)\n\n\n\u2212\n\n+\n\n\nC\nl\n\n\n\n\n(\n\na\nq\n\n)\n\n\n\u2212\n\n\u2192\nP\nt\n\n\nC\nl\n\n\n4\n\n\n(\n\na\nq\n\n)\n\n\n\n2\n\u2212\n\n\n\n\n\n\n\n\n(3)\n\n\nP\nt\n\n\nC\nl\n\n\n4\n\n\n(\n\na\nq\n\n)\n\n\n\n2\n\u2212\n\n\n+\n\n\nC\nl\n\n\n2\n\n\n(\n\na\nq\n\n)\n\n\n\n\u2192\nP\nt\n\n\nC\nl\n\n\n6\n\n\n(\n\na\nq\n\n)\n\n\n\n2\n\u2212\n\n\n\n\n\n\n\n\n(4)\n\n\n3\nP\nt\n+\n4\nH\nN\n\nO\n3\n\n+\n18\nH\nC\nl\n\u2192\n3\n\nH\n2\n\nP\nt\n\n\nC\nl\n\n6\n\n+\n4\nN\nO\n+\n8\n\nH\n2\n\nO\n\n\n\n\nTo avoid energy consumption and large volumes of liquid waste production in hydrometallurgical and pyrometallurgical processes, we need processes with high efficiency and environmentally friendly achievements. Deep eutectic solvent ionic liquids are always available, inexpensive, and chemically stable and can be used in the extraction and recovery processes of precious metals [22,31,32]; Olga Lanaridi et al., 2022). Aliquat-336 has been used for selective extraction of precious metals from a mixture under optimal conditions. The results show that the use of this solvent is suitable for the separation of high purity metal from the mixtures [24]; Wei et al., 2016). Polymerized ionic liquid has high efficiency for separating platinum from spent materials. The complete recovery of platinum occurs by leaching with 1% H2O2 in 8 M HCl, poly SILPs 20%, in S/L of 1:5, at 65\u00a0\u00b0C for 3 hour (Lanaridi et al., 2021). Aluminum and iron are separated by adding sodium phosphate to the acid leaching solution. The solution of 0.01 M Aliquat 336 in kerosene is used to extract platinum from the residues. After loading the organic phase, 0.5 M HCl and thiourea were used and 99.91% of platinum was obtained from stripping (Raju et al., 2012). After achieving the best conditions in the leaching stage, platinum is separated from the leach liquor by resins. These resins are formed from Merrifield beads (M) with triethylenetetramine (TETA), ethane-1,2-dithiol (SS) and bis((1H-benzimidazol-2-yl)methyl)sulfide (NSN) to form M-TETA, M\u00a0\u2212\u00a0SS and M-NSN, respectively [33]. Platinum recovery efficiency in acidic leaching is high. It is recommended to replace the acidic agent with salt or metal oxide to reduce the effects of acid on equipment and the environment.According to the results in Table 6\n, the efficiency of acidic leaching based on temperature and energy consumption is shown in Fig. 7\n. In the process of leaching with hydrochloric acid and hydrogen peroxide or iodine, platinum can be completely recovered at an optimum temperature of 65\u201370\u00a0\u00b0C.Using the microwave saves the time and energy required for the pretreatment of spent catalysts. The use of microwaves in the absence of oxidants has also led to an increase in platinum recovery (Trinh et al., 2020). In the leaching method with aqua regia at L/S of 5, at 109 \u2e30C and for 2.5 hr, 96.5% of platinum with a purity of 94.2% was recovered. In the second leaching method with aqua regia at L/S of 2, at 400 \u2e30C, and for 5 min, microwave radiation was used for heating and 98.3% of platinum with a purity of 98.9% was obtained [15]. The microwave is used for leaching in two stages, 500 W and 900 W. Two leaching samples were performed with a concentration of 6 M HCl in the presence and without 10% vol. H2O2 at 150\u00a0\u00b0C for 25 min. As H2O2 increases, the recovery efficiency of platinum slightly increases. With the addition of 10% vol. H2O2, the recovery efficiency of platinum has been increased from 90% to 96% (Atia et al., 2021). The combination of the spent catalyst, nickel matte and sodium salts with microwaves, reduces the viscosity and melting temperature, which results in 98.59% of the platinum being recovered at temperature of 1250\u00a0\u00b0C, time of 2 h, and N2 atmosphere. High efficiency in a short time is the result of using microwaves (Huimin Tang et al., 2021).In leaching and metal extracting, acidic and chemical substances are widely used. This consumption is not only costly but also poses risks to the environment. Therefore, replacing bio and nano solvents can be an alternative method. Bioleaching is done in two groups of direct and indirect processes. The direct method occurs in the presence of microorganisms, but in the indirect method, biometabolite is first produced by microorganisms and the leaching process is performed with this solution in the absence of microorganisms. The second method is more applicable on an industrial scale [20,34]. According to the biological methods and with the help of desferrioxamine B, about 30% of platinum has been obtained from the spent catalyst (H [13]. The refinery reforming catalyst was tested to recover platinum by the bioleaching method. The percentage of platinum recovery by oxalic acid in the bioleaching process increases from 13% to 30%. Comparing the bioleaching process with chemical leaching, it is concluded that oxalic acid has an effective role in both processes. The recovery of platinum in chemical leaching is slightly higher than in the bioleaching process [14]. About 93% platinum in the spent car exhaust catalyst was transferred to the solution phase of hydrochloric acid 12\u00a0M and nitric acid 9.5 M. Magnetite nanoparticles with silicate coating have been used to recover platinum from acidic solution (Hasani et al., 2015). The Gemini process (a liquid-solid resin ion exchange system) is used to recover platinum from a spent platinum-rhenium catalyst based on the alumina of the reforming unit. This process involves the steps of catalyst base dissolution, platinum separation, and purification. 91.71% of platinum is recovered in leaching with sodium borohydride at 60\u00a0\u00b0C for 15 min (Soltan mohammadzadeh et al., 2003). Selective recovery of platinum from automotive spent catalysts by anionic platinum chloride and cationic biocarbon adsorbents has been compared. The adsorbent Tu\u2013N\u2013SCG\u2013C\u2013A for selective recovery of precious metals is a priority not only economically but also environmentally [35].The reduction of time with the high efficiency of platinum recovery in the microwave leaching method is significant. But how this method is used on an industrial scale needs further investigation. The conditions for using this method on a laboratory scale are summarized in Table 7\n. The bioleaching process is both environmentally and economically beneficial but needs to be further explored for use on an industrial scale due to its low efficiency and longtime requirement. The factors influencing each research are presented in Table 7.The efficiency of platinum recovery in microwave and bioleaching leaching processes increases with increasing temperature. But in the same conditions of microwave leaching process, with increasing temperature from 109 to 400\u00a0\u00b0C, the efficiency increases from 96.5 to 98.9. It is not economical to use energy for increasing 2% of efficiency, so the suitable temperature is about 110\u2013100\u00a0\u00b0C for acquiring an efficiency of about 96% is very good. Also in the bioleaching process, increasing the temperature from 60 to 90\u00a0\u00b0C only increases the 1% in efficiency. Therefore, a temperature of 60\u00a0\u00b0C seems appropriate (Fig. 8\n).The importance of supplying platinum for various industries and removing environmental pollution from industrial wastes have led to the recovery of this metal from secondary sources. The general methods of extracting and separating platinum from spent catalysts include pyrometallurgy and hydrometallurgy. In this article, the hydrometallurgy method is described as an optimal method and its steps are explained in detail. The spent catalyst as feed is prepared. In the pre-treatment stage, heating removes pollutants from the surface of the spent catalyst, but it is suggested to use a method similar to washing and avoid energy consumption. Investigating, studying and comparing factors such as the use of economic and environmentally friendly solvents, lower temperature, lower solid-to-liquid ratio, shorter required time, reducing the chemical potential of the reaction and accelerating the separation of the metal from the base of spent catalyst, will improve the leaching process. It is suggested that experiment design software and artificial neural network; be used to predict the optimal values of leaching parameters and recovery percentage.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.", "descript": "\n                  According to the technological importance of platinum in the modern industries, demand for platinum has been increasing. On the other hand, the supply of platinum from mineral resources had environmental and economic challenges. Therefore, recovery of platinum from secondary sources such as spent catalysts is inevitable and important as a valuable solution for waste management and supply of required platinum. The general methods of extracting and separating platinum from spent catalysts include pyrometallurgy and hydrometallurgy. Choosing and optimizing the method of platinum recovery includes advantages such as: lower energy consumption, production of more platinum and less gas pollutants, less production of industrial waste, more safety and less investment. In this article, the hydrometallurgical method was chosen by examining various references, which includes pretreatment and leaching steps. Also, the effect of parameters such as: amount of crushing and temperature in the pre-treatment stage, type of solvents, temperature, solid-to-liquid ratio and time in the leaching stage has been investigated on the efficiency of platinum recycling. Consequently, crushing to 100 \u03bcm and calcination at about 500\u00a0\u00b0C or oven drying at about 100\u00a0\u00b0C are suggested for pretreatment. In the leaching step, the replacement of metal salts with acid, microwave heating speed and the use of bio/nano solutions help the environment. Washing efficiency increases with increasing temperature up to 100\u00a0\u00b0C, ratio of acidic agent to solid catalyst and washing time. Attention to economic and environmental issues is suggested for future works.\n               "}
{"full_text": "Nowadays, one of the main concerns of global society is public health and environmental safety. Industrial and agricultural activities have released various organic toxic compounds, which can contaminate surface and groundwaters, thus threatening access to fresh drinking water [1]. One of the various contaminants coming from anthropogenic activities is p-nitrophenol (PNP), widely known as a toxic and carcinogenic nitroaromatic chemical compound. The primary source of industrial activities facilitating the PNP contamination is the production of rubbers, pesticides, textile, dyes, and pharmaceuticals [1\u20133]. It can easily intoxicate living animals and humans and lead to serious health problems such as confusion, skin and eye irritation, loss of consciousness, and even potential carcinoma [2]. Moreover, exposure to the PNP might cause a negative effect on blood cells and damage the central nervous system and other human organs [4].Various treatment technologies, including advanced oxidation processes, thermal degradation, photodegradation, electro-coagulation, biological treatment, adsorption, and others have been used to efficiently remove PNP [1,5,6]. Although biological treatment can efficiently degrade PNP, it has several disadvantages, such as a slow start-up time and decreased efficiency at low temperatures and high PNP concentrations [7]. At the same time, purification techniques, such as electro-coagulation, photodegradation, and adsorption, also have several drawbacks, including high cost, long operation time, and reduced efficiency [6]. Moreover, there are conventional methods available for the reduction of PNP to p-aminophenol (PAP), such as the use of hazardous Sn/HCl, or Fe/HCl, catalytic transfer hydrogenation (CTH), or molecular hydrogen (H2) [8]. However, those methods have several disadvantages since they demand complex experimental design, high pressure, and temperature [8].Recently, metal nanoparticles such as Au, Ag, Ni, Pt, Co, and Pd have attracted attention for the catalytic reduction of PNP due to their good initial activity [9\u201314]. However, the agglomeration of the nanoparticles resulting in decreased removal efficiency has been reported as a fatal defect [15]. Various immobilization techniques have been developed to prevent the agglomeration of nanoparticles using support materials such as graphene hydrogel, polystyrene beads, graphene oxide, magnetite, etc. [10,15\u201317]. However, no significant study has been conducted to use nanoscale zerovalent iron (NZVI) as a support material for the immobilization of metal nanoparticles to degrade PNP efficiently. In recent years, the well-known advantages of NZVI, such as high reductive capacity and economical synthesis method, made it one of the most widely studied and used environmental materials for the treatment of various surface and groundwater pollutants found in the industrial and agricultural sectors [2,7,18]. For example, NZVI has been proven to effectively remove diverse halogenated organic compounds, oxy-onions, and heavy metals [19]. However, NZVI has several disadvantages, such as rapid oxidation of its surface to Fe oxides, which decreases the activity of NZVI acting as a reductant. Another serious drawback of NZVI is its tendency to agglomerate due to magnetic forces, which also decreases the reactive surface and reduces the reductive efficiency of the material [1]. On the other hand, the magnetic property of NZVI allows the easy collection of NZVI-supported catalysts from the suspension system after catalytic reaction [7,18]. Hence, NZVI could be a promising support material with a great potential for the enhanced reduction of PNP.Previous studies showed that a variety of metallic catalysts with the promoter and noble metals on the surface of NZVI were successfully applied to remove nitrate, trichloroethylene, and tetrabromobisphenol [18,20\u201324]. The type of the promoter metal highly affects the degradation kinetics of the nitrate removal [25]. Hence, promoter metals including Cu, Sn, In, and Zn on the surface of various supports were extensively tested and evaluated for the degradation of nitrates in combination with noble metals such as Pd, Pt, and Au, where Pd was the most widely and successfully used noble metal [7,25\u201328]. However, the combination of metal catalytic components deposited on the NZVI surface have not been used and investigated for the reductive degradation of PNP.The present work aimed to investigate the reduction of PNP by NZVI-supported metal catalysts. First, different types of promoter metals have been tested for the enhancement of the rate of reduction of PNP. Then, more suitable promoter metals (Cu, In, Ni, Zn, Sn) have been tested along with a noble metal (Pd). Finally, the effect of significant factors such as catalyst loading, nature of chemical promoter, and noble metal loading were investigated. Based on the present results, the reaction mechanism was suggested.NZVI was synthesized by a well-established method [18]. 50\u00a0mL of NaBH4 solution (0.9\u00a0M) was first prepared using deaerated and deionized water (DDIW). An exact concentration of FeCl3\n.6H2O (0.11\u00a0M) was prepared in ethanol and DDIW (1:8\u00a0v/v) and the NaBH4 solution was added dropwise into FeCl3\n.6H2O under constant mixing for >15\u00a0min to remove the remaining H2 gas. The suspension was sonicated for 2\u00a0min and washed with DDIW three times. The resulted suspension was used for the synthesis of bimetallic catalysts. Precursors for the promoter and noble metals were prepared by dissolving an appropriate amount of the relevant metal salt in DDIW, respectively. The solution was then added dropwise into the NZVI suspension under vigorous stirring. After addition of precursors, the suspension was stirred for 3\u00a0min to ensure reduction of metals by NZVI, and then washed with DDIW three times. The resultant slurry was used for the batch catalytic experiments.A morphological analysis of the catalyst was conducted using Scanning Electron Microscopy (SEM) with Energy-dispersive X-ray spectroscopy (EDX, Hitachi S-4700). Dried catalysts were placed onto metal sample holders and covered with a gold film. Catalytic activity experiments were conducted in a batch reactor (20\u00a0mL amber vial), and details are provided in the ESI.SEM/EDX analysis was conducted to investigate the morphological characteristics of NZVI and the dispersion of Pd particles on its surface. Fig. S1a-b (Electronic Supporting Information, ESI) illustrates the SEM images of 1.5%Pd/NZVI particle surface with magnifications of 20\u00a0k and 100\u00a0k, respectively. Fig. S1a shows that plate-shaped NZVI particles were synthesized. During the synthesis of NZVI, an ultrasonication process was applied [29], and thereby, round-shaped NZVI particles (~50\u00a0nm) can also be seen in Fig. S1b. The results indicate a successful synthesis of nano-sized iron particles. In addition, EDX mapping of surface elements of the catalyst was carried out to investigate Pd distribution on the NZVI surface. Fig. S1c-d shows e EDX mapping images of Pd and Fe, respectively, indicating that the chemical elements were well-mixed. It is also shown that Pd particles were uniformly dispersed on the surface of NZVI support. These results suggest that the applied synthesis method of Pd/NZVI provides proper dispersion and distribution of metal catalysts on the surface of NZVI support.Kinetic experiments in a batch reactor mode were conducted to evaluate the catalytic reduction of PNP by the bimetallic 4%Zn-1.5%Pd/NZVI, and the monometallic 1.5%Pd/NZVI and 4%ZnNZVI catalysts (Fig. 1\n). The reduction kinetics of PNP by bare-NZVI is also shown in Fig. 1 and compared to that obtained by the other catalysts. The control test (absence of catalyst) showed no removal of PNP throughout the experiment, indicating that no adsorption of PNP on the reactor's wall and no reduction by photolysis in the amber vial (reactor) occurred during the reaction. The reduction of PNP by bare NZVI reached 93.7% in 5\u00a0min, while a monometallic catalyst (4%Zn/NZVI) can completely degrade PNP in 3\u00a0min. The presence of promoter metal (Zn) could facilitate an electron transfer from the NZVI surface compared to the relatively slow direct electron transfer from the bare-NZVI surface, resulting in the accelerated catalytic reduction kinetics of PNP [7]. The complete reduction of PNP by 4%Zn-1.5%Pd/NZVI occurred in 1\u00a0min, and its pseudo-first-order kinetic rate constant k1 (0.0954\u00a0s\u22121, R2\u00a0=\u00a00.979) was found to be 3.6 and 11.8 times higher than that of 4%Zn/NZVI and bare-NZVI, respectively. Much faster reduction kinetics of PNP by the 4%Zn-1.5%Pd/NZVI solid could be originated from the additional formation of activated hydrogen on the surface of noble metal (Pd) during the facilitated electron transfer at the Zn/NZVI interface. This can rapidly and strongly degrade PNP on the Pd surface inducing much higher catalytic activity for the enhanced PNP reduction [4,30,31]. Hence, the addition of promoter and noble metal to the bare-NZVI can increase the catalytic reduction rate of PNP by facilitating electron transfer and subsequent hydrogenation [32]. In contrast, 1.5%Pd/NZVI showed the fastest reduction kinetics of PNP (k1\u00a0=\u00a00.248\u00a0s\u22121, R2\u00a0=\u00a01), of which the kinetic rate constant k1 is 4.1 times higher than that of 4%Zn-1.5%Pd/NZVI. It indicates how the hydrogenation occurred on the Pd surface could overwhelmingly contribute to the enhanced reductive catalysis of PNP with the fastest reduction kinetics. We show here the superiority of NZI-supported mono noble metal (Pd) catalyst over the bimetallic one for the enhanced reduction of PNP.The batch kinetic experimental results for the removal of PNP by the 1.5%Pd/NZVI catalytic system were compared to those obtained by other catalysts recently reported. Table S1 summarizes the kinetic rate constant for the removal of PNP by each of the catalysts under diverse experimental conditions. It can be seen that the 1.5%Pd/NZVI has the highest catalytic activity for the PNP reduction among the catalysts reported to date. Most of the previously reported catalysts for the PNP removal used passive support materials that cannot donate electrons and facilitate the electron transfer from the support, while NZVI-supported catalysts can actively donate electrons to the promoter metal or directly to the contaminant [1,15,33\u201335]. For instance, Chen et al. [35] investigated the performance of Au/Pd bimetallic catalyst deposited on the surface of graphene nanosheets, which did not possess any reductive capacity, and they were simply used to prevent the agglomeration of the nanoparticles. Here, NZVI-supported mono- and bimetallic catalysts showed high activity for the enhanced PNP removal. It can be concluded that the synthesized Pd/NZVI catalyst appears as one of the most promising nanocatalysts for the enhanced PNP removal.NZVI-supported monometallic catalysts with different promoter metals including Cu, Sn, Zn, Ni, and In were tested for the reduction of PNP. 4%Zn/NZVI showed the fastest reduction kinetics; hence, it was selected for further experimental studies. An increase in the catalyst loading resulted in the saturation point of its catalytic reactivity at a concentration of 500\u00a0mg/L. Monometallic Pd/NZVI catalyst showed a faster reduction kinetics than bimetallic Zn-Pd/NZVI since Zn particles could block available reactive surface sites of Pd. However, an increase in Pd loading of Pd/NZVI catalyst led to a decreased reduction kinetics of the catalyst. The details of the section are provided in the ESI.\nFig. 2\n shows the variation of UV\u2013Vis spectra during the reduction of PNP by Pd/NZVI. Once PNP was added to a weak basic aqueous solution (pH ~7.5), it could be easily deprotonated to form p-nitrophenolate. The catalytic reduction is initiated by the addition of Pd/NZVI. As the catalytic reaction proceeds, the peak at 400\u00a0nm corresponding to p-nitrophenolate is decreased, while a peak at 300\u00a0nm corresponding to p-aminophenol (PAP) is increased [2]. It indicates that the main reduction product of the catalytic reduction of PNP by Pd/NZVI is PAP.The catalytic reduction of PNP to PAP on the surface of Pd/NZVI can be explained via two main reduction pathways: (i) direct reduction of PNP to PAP via electron transfer from the reactive NZVI support in the form of Fe(II) and Fe(0) (Eqs. (1), (2)) and (ii) indirect reduction (hydrogenation) via reactive Hads generated on the Pd surface (Eqs. (2)\u2013(4)) [4,7,31].\n\n(1)\n\n\nFe\n\n2\n+\n\n\n\u2192\n\nFe\n\n3\n+\n\n\n+\n\ne\n\u2212\n\n\n\n\n\n\n(2)\n\n\nFe\n0\n\n\u2192\n\nFe\n\n2\n+\n\n\n+\n2\n\ne\n\u2212\n\n\n\n\n\n\n(3)\n\n2\n\nH\n+\n\n+\n2\n\ne\n\u2212\n\n\u2192\n2\n\nH\nads\n\n\u2192\n\nH\n2\n\n\n\n\n\n\n(4)\n\n\nPd\n0\n\n+\n\nH\n2\n\n\u2192\nPd\n\u2212\n2\n\nH\nads\n\n\n\n\nThe surface of NZVI could be oxidized to Fe(II) oxides with the generation of electrons until its surface reached complete passivation by Fe(III) oxides. They could further react with aqueous H+ forming the reactive Hads adsorbed species on the Pd surface that is the main overwhelming driving force to vigorously reduce PNP to PAP in the monometallic system (Eq. (3)). Pd particles were able to continuously activate H2 to the reactive Hads species on their surface (Eq. (4)), leading to the enhanced PNP reduction kinetics by the continuous catalytic reduction system of PNP. Lai et al. [31] demonstrated that the generation of Hads was the main reducing power in the reductive degradation of PNP by Fe/Cu catalyst. It could not completely reduce the PNP under high pH conditions since low H+ concentration limited the generation of Hads [31]. Moreover, since the addition of promoter metal and its loading increase have deteriorated the catalytic reduction kinetics of PNP, we can conclude that the indirect reduction of PNP via hydrogenation pathway with the reactive Hads species played the main role in the reaction mechanism of the catalytic PNP reduction.The study provided insights on the proper synthesis of NZVI-supported metal catalysts for the enhanced catalytic reduction of PNP. The effect of significant factors such as catalyst loading, promoter type and loading, and noble metal loading on the performance of catalytic PNP reduction were evaluated for the optimal operation of the batch catalytic system. Monometallic catalyst with a noble metal (1.5%Pd/NZVI) showed the fastest PNP reduction kinetics (k1\u00a0=\u00a00.248\u00a0s\u22121, R2\u00a0=\u00a01) among the catalysts reported to date, while bimetallic catalyst (4%Zn-1.5%Pd/NZVI) has shown much faster PNP reduction kinetics (k1\u00a0=\u00a00.095\u00a0s\u22121, R2\u00a0=\u00a00.979) than the Pd monometallic catalysts with different promoters. The optimal catalyst loading was observed at 500\u00a0mg/L for the enhanced catalytic reduction of PNP. Indirect reductive transformation of PNP to PAP via hydrogenation with reactive Hads on Pd surface was suggested as the main reduction pathway since 1.5%Pd/NZVI has shown the highest rate for the catalytic reduction of PNP to PAP. The type and content of noble metal influencing the catalytic activity for an application to practical water treatment systems need to be carefully selected and evaluated by considering its role and behavior in the catalytic reduction of PNP. The limitations of this study are the absence of activity tests under different pHs of the suspensions and the absence of stability test of the catalyst during repeated cycles, which will be both our near-future research tasks.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was supported by the Research Grants of Nazarbayev University (091019CRP2106 and 021220FD1051) and the Ministry of Education and Science of the Republic of Kazakhstan (APO9260229). The authors would like to thank Prof. Sungjun Bae of Konkuk University for basic environmental monitoring training. The authors also would like to extend their gratitude to the anonymous reviewers that helped significantly to improve the quality of the paper.\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2021.106337.", "descript": "\n                  Nano-sized zerovalent iron (NZVI) - supported metal catalysts were synthesized to characterize their reactivity for the reductive degradation of p-nitrophenol (PNP). Among the tested monometallic catalysts using metal promoters, Zn/NZVI showed the highest reactivity with complete reduction of PNP in 5\u00a0min (k\u00a0=\u00a00.0263\u00a0s\u22121). The addition of Pd accelerated the degradation kinetics of PNP with complete reduction in 1\u00a0min (k\u00a0=\u00a00.095\u00a0s\u22121) but promoter's presence on bimetallic catalyst surfaces simply decreased their reactivity. A proper Pd amount (1.5\u00a0wt% Pd/NZVI) showed the highest degradation rate (k\u00a0=\u00a00.248\u00a0s\u22121), while after its content increased to 10\u00a0wt% the rate was reduced by 5.8 times.\n               "}
{"full_text": "Most of non-renewable energy sources, such as coal, oil, and natural gas contribute to the increase in the CO2 emission [1,2]. Turning CO2 into high-value fuels or chemicals becomes an alternative strategy to replace fossil fuels with renewable energy and reduce the emissions of CO2 in the atmosphere [3,4]. Converting of CO2 to single carbon (C1) products, such as formic acid, carbon monoxide, methane, and methanol through direct hydrogen reduction or hydrothermal-chemical reduction in water have gained many interests of researchers. Methane (CH4) produced from hydrogenation of CO2 or CO2 methanation becomes one of the renewable energy carriers. Extensive research works in CO2 methanation to CH4 have been reported [2,5\u20137]. CO2 methanation is considered to be an effective strategy to reduce the CO2 emission by generating CH4 in places where H2 is produced from renewable energy sources and afterwards to use it everywhere [2,5\u20137]. CO2 methanation, also called the Sabatier reaction, is exothermic reaction and limited by the equilibrium at high temperatures. It usually conducts at temperatures of 250-500\u00a0\u00b0C and pressures of 1\u201380\u00a0bar in the presence of hydrogen and heterogeneous catalysts [8\u201310]. Since this equilibrium limitations, a superior catalyst with high activity and selectivity towards CH4 at moderate temperatures and pressures condition is needed to be developed [5].Nickel-based catalysts are known as the most studied for CO2 methanation due to their low cost and high natural abundance [5,10,11]. The catalytic performance of the Ni-based catalysts depends on various parameters such as the Ni content, type of promoters and supports, preparation method, and reaction conditions [12]. Many studies have been reported in order to improve their catalytic performance at low temperature and stability at higher temperature [10,13,14].Modification of physicochemical properties by adding the promoters, such as transition metals [14], alkali and alkaline earth metals [15], and rare-earth [16] was reported as an alternative way to improve the catalytic performance of Ni-based catalysts. The transition metal additives, such as V, Cr, Mn, Ce, Fe, Co, Y and Zr was investigated as efficient promoters in CO2 methanation [17\u201321]. For example, Y and Zr are the most metal used as dopants to modify the properties of support. Zr is used to stabilize the CeO2 structure and enhance its oxygen vacancies population [17], while Y can produce oxygen vacancies in ZrO2-based supports [18,19]. Xu, et al. reported the incorporation of rare earth (La, Ce, Sm, and Pr) increased the surface basicity and electron properties of the Ni-based catalysts which contributed to the activation of the CO2. The Ni species were dispersed well and formed the strong metal framework interaction. The thermal sintering of the metallic Ni nanoparticles during CO2 methanation conditions was minimized and there was no obvious deactivation was observed after 50\u00a0h stability test. Thus, Ni-promoted catalyst could be considered as promising catalyst candidates in low-temperature CO2 methanation [22].The introduction of a second metal, such as Fe, Co or noble metals which are Ru, Rh, Pt, Pd and Re as a dopant was reported to influence the properties of Ni, increase the dispersion, stability, and reducibility. Thus, it enhanced the CO2 methanation activity of nickel-based catalyst by creating bimetallic and multi-metallic catalysts system [14,23,24]. For example, NiFe alloys were reported as an active and stable catalysts for dry reforming of methane due to the addition of Fe promoted carbon gasification and minimized coking [25]. Ru and Ni mostly form monometallic heterostructures that rely on the synergistic effect between the two separate metallic phases, while Pt and Pd mostly lead to the creation of NiPt and NiPd alloys. It has been shown that small addition of noble metal (e.g., 0.5% or 1%) enhanced the reducibility and low-temperature activity of Ni-based catalysts [26].The catalyst support also plays an important role on the morphology of the active phase, adsorption ability and catalytic properties [27]. The use of metal oxides as supports for nickel, such as \u03b3-Al2O3, SiO2, CeO2, ZrO2, and different combinations of mixed oxides CeO2-Al2O3, CeO2-ZrO2, Y2O3-ZrO2 was reported [5,9,11,13,21,28]. Among the supports, alumina is commonly used due to its high specific surface area and strong interaction with active metal [29]. The modification of the catalyst support becomes an alternative solution to increase the catalytic activity and reducibility. Different support modifiers, such as ZrO2, SiO2, MgO, La2O3, CeO2, and TiO2 showed better conversion, higher redox property, higher thermal stability and resistance against sintering due to their excellent properties [30].Yttria-stabilized zirconia(YSZ) as known as SOEC material has been utilized for catalytic applications due to their properties [31]. YSZ as an oxygen ion conductor is a ceramic material combines different functionalities such as good thermal stability, selective bulk oxygen mobility and high surface oxygen vacancy concentration. YSZ has been considered as a promising support for metallic nanoparticles or as a catalyst itself [19,32]. Nickel and yttria-stabilized zirconia (Ni\u2013YSZ) cermet has been commonly employed as cathode material of SOEC due to its high catalytic activity, excellent electronics and conductivity, low cost, sufficient thermal expansion coefficient, and mechanical-chemical compatibility with other components [33]. Ni plays the main role of catalyst for oxidation of the fuel, and YSZ acts as a support to hold the porous structure and prevent Ni coarsening [13]. However, few researches have been done regarding to methanation over YSZ as catalyst support so far. Kosaka, et al. prepared Ni-yttria-stabilized zirconia (Ni-YSZ) tubular catalysts with different NiO contents ranging from 25 to 100\u00a0wt% and investigated the effect of Ni content on the CO2 methanation performance. The results showed that catalysts with Ni contents >75\u00a0wt% produced CH4 yields >91% and high CH4 selectivities (>99%) [5]. Watanabe, et al. developed an yttria-stabilized zirconia catalyst-supported nickel (Ni/YSZ) with high tolerance to coke deposition during methane steam reforming (MSR). Ni/YSZ prepared by an electroless plating method exhibited better stability during the MSR than prepared by a conventional impregnation method [34]. Fakeeha, et al. studied the use of yttria stabilized zirconia support with different loadings (5, 10, 15 and 20\u00a0wt%) of yttria. The results showed that Y2O3 stabilized ZrO2 supported catalysts produced higher conversions of CH4 and CO2 and higher stability compared with unstabilized ZrO2 supported catalysts at 700\u00a0\u00b0C [35].In this research, 60wt%NiO-40wt%YSZ and 60wt%NiO\u00a0+\u00a040wt%YSZ catalyst were investigated their catalytic activity for CO2 methanation. The ratio of NiO and YSZ which are 60\u00a0wt% and 40\u00a0wt%, respectively, was the optimum ratio for NiO-YSZ as a SOEC cell stack. The effect of feed ratio, catalyst reduction temperature, reaction temperature, and support to the activity and selectivity of the catalyst were investigated. The common methanation catalyst, 10wt%Ni/Al2O3 was compared as a benchmark. Furthermore, the stability test was also investigated through long-term tests.Catalyst precursors used are NiO (Sumitomo), YSZ (Tosoh), Ni(NO3)2.6H2O (Wako) and KHO-12 (spherical alumina (1\u20132\u00a0mm) Sumitomo Chemicals). Reactant gases used are hydrogen, argon, carbon dioxide, and carbon monoxide. Each of their purity are 99.999%, 99.995%, 99.95%, and 99.95%, respectively.Three type of catalysts used in this research are 60wt%NiO-40wt%YSZ, 60wt%NiO\u00a0+\u00a040wt%YSZ, and 10wt%Ni/Al2O3 catalyst. 60wt%NiO-40wt%YSZ was prepared in Chubu Centre of the National Institute of Advanced Industrial, Science and Technology (AIST). The powders of NiO and YSZ were calcined at 1400\u00a0\u00b0C, and reduced at 700\u00a0\u00b0C at 2\u00a0h by hydrogen flowing. Then, the bulks were crushed to small powder. While, 60wt%NiO\u00a0+\u00a040wt%YSZ was prepared in Department of Applied Chemistry and Biochemical Engineering, Shizuoka University. As of 60\u00a0wt% NiO and 40wt%YSZ were mixed physically by using mortar until homogeneous. After that, the mixture was reduced in situ with 10\u00a0ml\u00b7min\u22121 H2 flow at various temperature in the range of 300\u00a0\u00b0C until 500\u00a0\u00b0C. 10wt%Ni/Al2O3 catalyst was prepared by impregnation method from Ni(NO3)2.6H2O and KHO-12 (spherical alumina (1\u20132\u00a0mm)). The Ni/Al2O3 was calcined at 500\u00a0\u00b0C for 5\u00a0h, and reduced in a reactor with 10\u00a0ml\u00b7min\u22121 H2 flow at 500\u00a0\u00b0C for 5\u00a0h before reaction.Powder X-ray diffraction patterns were collected using a Rigaku RINT 2000 equipped with a Cu K\u03b1 (\u03bb\u00a0=\u00a01.5418\u00a0\u00c5) source and the Brag-Brentano \u03b8\u2013\u03b8 configuration in the 10\u201390\u00b0 2\u03b8 range, with 0.05\u00b0 step size and 2\u00a0s acquisition time. The crystallite size (d) of the catalyst samples was determined using Scherrer formula [30]:\n\n\nd\n=\n\n\n0.9\n\u03bb\n\n\u03b2cos\u03f4\n\n\n\n\nWhere \u03bb is the x-ray wavelength, \u03b2 is the full-width for the half maximum intensity peak, and \u03b8 is the diffraction angle.The surface area of the catalysts was measured by N2 adsorption desorption isotherms at 77\u00a0K using a Micromeritics ASAP 2010 apparatus. Before measurement, the catalysts were degassed at 300\u00a0\u00b0C in N2 for 5\u00a0h. The surface area was calculated by the Brunauer-Emmet-Teller (BET) method in the equilibrium pressure range 0.05\u00a0<\u00a0P/P\u00b0\u00a0<\u00a00.3.H2-temperature programmed reduction (H2-TPR) of the samples were performed using BEL MULTI-TASK-TPD. Analysis was carried out on 300\u00a0mg of sample, heating from 50 to 500\u00a0\u00b0C at a heating rate of 5\u00a0\u00b0C\u00b7min\u22121 with 10%H2 flow rate of 5\u00a0ml\u00b7min\u22121 and holding at the final temperature for 5\u00a0h. The H2 consumption was measured by a mass spectrometer (MS).CO2-temperature programmed desorption (CO2-TPD) of the samples were performed using BEL MULTI-TASK-TPD. Analysis was carried out on 1000\u00a0mg of sample, heating from room temperature to 500\u00a0\u00b0C at a heating rate of 5\u00a0\u00b0C\u00b7min\u22121 with helium gas flow rate of 50\u00a0ml\u00b7min\u22121 for 30\u00a0min. After the cleaning with He gas, the samples were cooled to 50\u00a0\u00b0C, switched in CO2 and saturated down at 50\u00a0\u00b0C. Then, the samples were purged again with He. The CO2 consumption of it was measured by an MS detector with heating from 50\u00a0\u00b0C to 500\u00a0\u00b0C.The performance of the catalysts was evaluated for CO2 methanation. The reactions were performed in a fixed bed reactor operating at atmospheric pressure by means of gaseous mixtures of H2/ CO2/Ar with different volumetric ratios in a temperature range 160\u2013440\u00a0\u00b0C and GHSV 18,750\u00a0h\u22121. Prior to the activity tests, 1000\u00a0mg of catalyst was placed inside a stainless steel fixed bed reactor (inner diameter: 9.0\u00a0mm), with quartz wool at both ends, and reduced in situ with 10\u00a0ml/min H2 flow, increasing the temperature from room temperature up to 700\u00a0\u00b0C and isothermally kept at this temperature for 60\u00a0min. Different reduction temperatures were employed to determine the effect of reduction temperature. Afterwards, the feed mixture was flowed through the reactor. The products were analyzed by two online gas chromatographs (Shimadzu GC-14B) consisted of two detectors, thermal conductivity detector (TCD) and flame ionization detector (FID) in series. Each of GCs have a Molecular Sieve 5A and a Porapak T columns, for analyses of gases and liquids including CO2, respectively.The specific surface area of the catalyst samples was calculated from their respective N2 adsorption isotherms. YSZ showed higher specific surface area compared to 60wt%NiO-40wt%YSZ and 60%NiO\u00a0+\u00a040%YSZ. It indicated that the deposition of Ni0 changed the textural properties of the YSZ support as shown by BET surface areas (6.93\u20137.70\u00a0m2. g\u22121) that are lower than YSZ support (13.7\u00a0m2. g\u22121). There was no significant difference observed between the specific surface area of the 60%NiO\u00a0+\u00a040%YSZ reduced at range temperature of 300-500\u00a0\u00b0C. The specific surface area of 60%NiO\u00a0+\u00a040%YSZ catalysts after reduced at range temperature of 300-500\u00a0\u00b0C tend to remain stable in the range of 6.93\u20137.70\u00a0m2. g\u22121. 60wt%NiO-40wt%YSZ catalyst reduced at 700\u00a0\u00b0C also exhibited the similar value of specific surface area. It indicated that YSZ as support contributed to maintain the specific surface area and minimize the sintering effect of NiO particles in 60wt%NiO-40wt%YSZ at higher reduction temperature.The XRD analysis was carried out to find out the presence of the metallic states of active metals. Fig. 1\n shows the XRD patterns of the catalysts before (fresh) and after reaction (used) at various reduction temperatures. Both of fresh and used 60wt%NiO-40wt%YSZ and 60wt%NiO\u00a0+\u00a040wt%YSZ catalysts showed the existence of cubic fluorite structure of YSZ support at 2\u03b8\u00a0=\u00a030o, 34o, 59o, and 73o (JCPDS 81\u20131550). There were significant differences observed between fresh and used catalysts. The 60wt%NiO-40wt%YSZ fresh showed the characteristic peaks of cubic nickel oxide (NiO) at 2\u03b8\u00a0=\u00a037.3o, 43.3o and 62.9o (JCPDS 73\u20131519). Whereas, either 60wt%NiO-40wt%YSZ or 60wt%NiO\u00a0+\u00a040wt%YSZ used catalysts, all of NiO phase peaks were disappeared and metallic fcc-Ni0 phase existed in all the reduced catalysts at 2\u03b8\u00a0=\u00a044.5o, 51.8o, and 76.5o (JCPDS 04\u2013850). It suggested that catalyst reduction was succeeded to reduce the Ni2+ into Ni0. As can be seen from the XRD patterns that the characteristic peak of YSZ as support shifted to the lower 2\u03b8. It indicated that addition of NiO into YSZ decreased the crystallinity and crystallite size of YSZ. The crystallite sizes of YSZ, d (nm), calculated by Scherrer's equation decreased from 20.45\u00a0nm to 17.92, 19.04, and 18.79\u00a0nm for 60wt%NiO\u00a0+\u00a040wt%YSZ 350\u00a0\u00b0C-red-5\u00a0h, 60wt%NiO\u00a0+\u00a040wt%YSZ 400\u00a0\u00b0C-red-5\u00a0h, and 60wt%NiO\u00a0+\u00a040wt%YSZ 500\u00a0\u00b0C-red-5\u00a0h catalysts, respectively.The reducibility of the NiO, YSZ, 60wt%NiO-40wt%YSZ and 60wt%NiO\u00a0+\u00a040wt%YSZ catalysts were studied by H2-TPR as shown in Fig. 2\n. NiO, 60wt%NiO-40wt%YSZ and 60wt%NiO\u00a0+\u00a040wt%YSZ reduced samples showed that there were two main peaks observed for hydrogen consumption and the reduction started at temperature about 200\u00a0\u00b0C. The first peak at lower temperature (200-300\u00a0\u00b0C) is associated with larger NiO particles that are of similar nature to pure bulk NiO or weakly interact with the YSZ support. These particles can be reduced at low temperatures. Whereas, the second peak at higher temperature (342-485\u00a0\u00b0C) is attributed to the greater dispersion of metallic oxide that strongly interact with the YSZ support. The absence of peaks for YSZ indicates that YSZ was not a highly reducible support material.The CO2-TPD patterns were obtained as shown in Fig. 3\n. All the catalysts had only weak low-temperature TPD peak implying that there were only weak basic sites for CO2 chemisorption. 60wt%NiO-40wt%YSZ and 60wt%NiO\u00a0+\u00a040wt%YSZ exhibited two peaks which first peak was at 55-144\u00a0\u00b0C and second peak was at 152-335\u00a0\u00b0C. The first peak of 60wt%NiO-40wt%YSZ is higher than 60wt%NiO\u00a0+\u00a040wt%YSZ indicated the higher CO2\u00a0methanation activity of 60wt%NiO-40wt%YSZ at lower reaction temperature.To study the effect of feed ratios, several test runs were performed at reaction temperatures of 160 to 440\u00a0\u00b0C with a total feed flow rate of 100\u00a0ml/min by varying only the volume percentages of the reactant gases CO2 and H2 while the volume percentage of inert gas Ar was kept constantly. CO2 conversion, CH4 selectivity and CH4 production rate results are displayed in Fig. S1. The excess amount of H2 enhanced CO2 conversion and CH4 selectivity. However, an increase in the CO2 reactant amount favored CO formation, since a lower CH4 selectivity was produced in the case of 1/1 (H2/CO2) ratio. These results suggested that the probability of CO2 reacting with the other reactant H2 and converting into CH4 was higher in the presence of an excess amount of hydrogen. As shown in Fig. S1, the highest CH4 production rate of those tested was obtained at a H2/CO2 ratio\u00a0=\u00a04/1, this ratio was used for the next catalytic activity tests.Different reduction temperatures were chosen to determine the effect on the performance of catalysts for CO2 hydrogenation reaction at the optimum of feed gas ratio H2: CO2: Ar\u00a0=\u00a072:18:10\u00a0ml\u00b7min\u22121. The higher reduction temperature of 60wt%NiO\u00a0+\u00a040wt%YSZ catalysts improved the performance of catalysts on methanation reaction. 60wt%NiO\u00a0+\u00a040wt%YSZ catalysts reduced at 300\u2013500\u00a0\u00b0C produced CO2 conversion more than 60% and CH4 selectivity is almost near 100% as shown in Fig. 4\n. This result was better than previous research using Ni/YSZ catalysts that performed lower CO2 methanation activity. CO2 conversion and CH4 selectivity reported were 60% and 75%, respectively [19].\nFig. 5\n shows CH4 production rates of nickel supported on YSZ at various reduction temperatures. For 60wt%NiO\u00a0+\u00a040wt%YSZ catalyst, the CH4 production rates tend to increase by the increase of reduction temperature in the following order: 350\u00a0\u00b0C\u00a0<\u00a0400\u00a0\u00b0C\u00a0<\u00a0500\u00a0\u00b0C. The 60wt%NiO\u00a0+\u00a040wt%YSZ catalyst reduced at 500\u00a0\u00b0C for 5\u00a0h produced the highest CH4 production rates of 31.81\u00a0mmol\u00b7gcat\n\u22121\u00b7h\u22121 at reaction temperature of 400\u00a0\u00b0C. It was little bit higher compared to 60wt%NiO-40wt%YSZ catalyst reduced at 700\u00a0\u00b0C for 1\u00a0h which produced CH4 production rates of 30.79\u00a0mmol\u00b7gcat\n\u22121\u00b7h\u22121 at 400\u00a0\u00b0C.Methanation activity of nickel supported on YSZ was better than those on YSZ. 60wt%NiO-40wt%YSZ reduced at 700\u00a0\u00b0C for 1\u00a0h performed higher CH4 production rate per catalyst weight compared with YSZ catalyst in all reaction temperatures carried on. YSZ as the support performed poor methanation activity at all reaction temperatures conducted. It suggested that the existence of YSZ minimized the sintering of nickel metal particle at higher reaction temperature and contributed to the catalytic activity of 60wt%NiO-40wt%YSZ catalyst for methanation at higher reaction temperature.Ni/Al2O3 catalysts are well-known as common catalysts for methanation due to the high specific surface area of Al2O3. The specific surface area of 10wt%Ni/Al2O3 catalyst prepared was 210 m2/g. However, one big drawback of Ni/Al2O3 catalyst is these catalysts are easily deactivated due to sintering of Ni particles and coke deposition during the exothermic methanation [6,9]. Fig. 6\n shows the turnover frequency (TOF) for methane (methane produced per nickel site per second) production by CO2 methanation on nickel supported on YSZ compared with nickel supported on Al2O3, nickel unsupported, and YSZ. The TOF for methane production by CO2 methanation on 60wt%NiO-40wt%YSZ was higher than 10wt%Ni/Al2O3 catalyst due to the existence of YSZ support as described in Fig. 6. At lower temperature, the TOF of 60wt%NiO-40wt%YSZ was lower and increased continuously by the increase of reaction temperature. It remained stable about 15\u00a0\u00d7\u00a0103-16\u00a0\u00d7\u00a0103 h\u22121 after 360\u00a0\u00b0C.The activation energy produced by 60wt%NiO-40wt%YSZ catalyst is 3.46\u00a0kJ/mol and it is lower than other Ni-YSZ published with activation energy of 93.6\u00a0kJ/mol [5]. This lower value of activation energy is apparent activation energy by reaction data. However, 60wt%NiO-40wt%YSZ catalyst produced activity and selectivity as high as published.As shown in Fig. 7\n, the 60wt%NiO-40wt%YSZ catalyst was stable during 220\u00a0h of CO2 methanation. There was no distinguishable difference observed in the activity of the catalyst. A long-term reaction at 400\u00a0\u00b0C revealed the remarkably stable performance of the 60wt%NiO-40wt%YSZ catalyst under CO2 methanation conditions with 29.56\u00a0mmol\u00b7gcat\n\u22121\u00b7h\u22121 of CH4 production rate and 1.52\u00a0mmol\u00b7gcat\n\u22121\u00b7h\u22121 of CO production rate. This result is similar as previous reported that Ni-based catalysts could maintain a very good methanation activity over a reaction time of nearly 100\u00a0h with high CH4 selectivity (about 100%) [5].The CO2 methanation activity was studied over nickel supported on YSZ catalyst. 60wt%NiO-40wt%YSZ catalyst exhibited high CO2 conversion close to the equilibrium one from 360\u00a0\u00b0C and above with CH4 selectivity of 100% under a GHSV of 18,750\u00a0h\u22121 at H2/CO2\u00a0=\u00a04. It was confirmed from XRD and TPR results that the catalytic pretreatment in H2 completely reduced the NiO to Ni0. CO2-TPD demonstrated the amount of CO2 adsorbed on 60wt%NiO-40wt%YSZ was much larger than that on 60wt%NiO\u00a0+\u00a040wt%YSZ. 60wt%NiO-40wt%YSZ catalyst produced higher TOF for methane production than 10wt%Ni/Al2O3 catalyst due to the existence of YSZ support and stable during more than 220\u00a0h.\nAnis Kristiani: Experimental work, Data analysis, Writing- Original draft preparation, Writing- Review and Editing. Kaoru Takeishi: Conceptualization, Methodology, Supervision, Funding acquisition. All authors have read and agreed to the published version of the manuscript.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to acknowledge funding from Japan Science and Technology Agency \u2013 Core Research for Evolutional Science and Technology (JST-CREST) through a project entitled \u201cDevelopment of Innovative Technology for Energy-Carrier Synthesis using Novel Solid Oxide Electrolysis Cell (JPMJCR1343)\u201d.\n\n\n\nSupplementary mateial\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106435.", "descript": "\n                  The CO2 methanation becomes promising solution to mitigate global warming and energy issues. Yttria-stabilized zirconia (YSZ) has been utilized for solid oxide fuel cell (SOFC) and solid oxide electrolysis cell (SOEC) applications. However, few researches about the use of YSZ as nickel-based catalyst support for CO2 methanation. Herein, we investigated the catalytic performance of 60wt%NiO-40wt%YSZ. The results showed that 60wt%NiO-40wt%YSZ produced methane with almost 100% selectivity and stable during more than 220\u00a0h. It confirms the excellent performance of 60wt%NiO-40wt%YSZ catalyst.\n               "}
{"full_text": "Data will be made available on request.The high energy demand due to the rapid growth of modernisation has accelerated the predicted depletion of conventional fossil-based resources such as coal, petroleum and natural gas. Fossil fuels are not eco-friendly as they are related to poisonous gas exhausts from automobile and industrial internal combustion engines, which eventually lead to global warming, acid rain and ozone depletion [1\u20133]. Our carbon demand for chemicals and energies along with the vast amount of greenhouse gas (GHG) emission attracts the development of renewable, non-polluting fuels and cleaner energy production methods [4,5]. In the EU, 80\u201395% GHG reduction is aimed to achieve in 2050 which is portrayed to be achievable with the use of renewable energy resources [6]. To promote the use of renewable energy resources as circular economy approach, political enactments such as the renewable portfolio standard (RPS), renewable fuel standard (RFS) and feed-in tariffs (FiTs) have been made in both developed and emerging economies countries [7\u20139].The use of renewable biofuel is the most promising alternative source of energy considering its biodegradability and low emission of carbon dioxide, free sulphur and non-toxic nature and possesses all the physicochemical properties of traditional fossil fuels such as liquid, solid and natural gas in terms of improved cetane number and high flash point [10,11]. Biofuels are biobased products, in solid, liquid or gaseous form which are produced from organic material such as plants, and their residues, agricultural crops, and by-products that can be an adequate substitute for petroleum-derived fuel [12\u201314]. The types of biofuels can be classified based on the feedstock used for production [13]. Waste cooking oil (WCO) is one of the most prudent sources of renewable energy feedstock because of its restrictions on reusability as food, freely available and its suitable properties for fuel production. Their valorisation can concurrently solve the WCO mismanagement issue and substantially reduce the cost of biofuel as feedstock constitutes 75\u201390% [15] of the overall cost which impeded the transition to the greener fuel option. WCO which originated from a plant (palm, soybean, rapeseed, etc.), consist of triglycerides with a long hydrocarbon chain ranging from C16 and C18, which has a similar molecular network to hydrocarbon fossil fuel [16]. During the repeated high-temperature exposure during frying, three main chemical reactions took place (thermolytic, oxidative and hydrolytic reaction) [17], which increases the oxidation variations (acidic compound and polymerised materials), and water content [18,19]. The difference between the fatty acid composition in fresh cooking oil and waste cooking oil has been reported in depth previously by Thawatchai et al. [20].Two routes that are mostly studied for WCO conversion to biofuel are the transesterification process and the pyrolysis process [21]. Biodiesel consists of alkyl esters produced by transesterification and contributes to most of the biofuel production [21,22]. Feedstock pre-treatment which requires an extra step in the production process is unavoidable for the sample with high moisture content, high impurities and FFA values such as WCO [5,23\u201325]. Moreover, a significant amount of glycerol is produced as a by-product of the conventional transesterification method [26]. Green diesel on the other hand contains deoxygenated hydrocarbon which may be obtained from the catalytic deoxygenation reaction [27]. Pyrolysis reaction is a rapid thermochemical conversion method that breaks the chemical bonds of the feedstock in an oxygen-free environment at a temperature range of 300\u2013500\u00a0\u00b0C into valuable products for biofuel or as chemical feedstock [16,28,29]. The product obtained from this pyrolysis reaction includes the organic liquid product, gaseous product, coke and water. In terms of the conversion process, the catalytic deoxygenation process has many advantages especially its great flexibility in the choice of raw materials, rapid conversion gradient, straightforward process, and industrial upscaling possibility [30,31]. The deoxygenated product (green diesel) possesses a higher heating value, higher energy density, higher cetane number and lower NOx emissions as compared to FAME biodiesel [27].Pyrolysis, however, is deemed to be cost-ineffective due to the high-temperature requirement compared to the product yield. Catalytic cracking with selective heterogeneous catalysts is a potential candidate for renewable-process-based industrialization to increase the pyrolysis yield and hence decrease the cost of liquid fuels [21]. These catalysts can be easily regenerated, recycled and environmentally friendly [31\u201333]. The development of heterogeneous catalysts is aimed to overcome the technical shortcomings associated with conventional non-recyclable homogeneous catalysts such as KOH and H2SO4 for biodiesel production. Up to date, a great deal of research on the development of a solid base heterogeneous catalyst for catalytic pyrolysis has been conducted such as biochar derived from chicken manure [34], Mg/Activated carbon [35], and K2O/Ba-MCM-41 [30]. On the other hand, only several solid acid heterogeneous catalysts have been proposed as catalytic material for biodiesel production such as composite zeolites [36], ZrO2 [37], Bentonite [38] and etc.With the conventional optimization method, it is time-consuming and requires a large number of experiments to determine optimum levels, which may be unreliable as the yield of organic liquid product (OLP), gas, water and coke are influenced by various factors such as pyrolytic temperature, residence time, heating rate, nitrogen flow rate [35,37,39]. By using the Taguchi method, the limitations of the conventional optimization method can be eliminated by optimising all the process parameters collectively using statistical experimental design. The most important feature of the Taguchi method is the use of an orthogonal array that can stipulate the way of conducting the minimal number of experiments which could give the full information of all the factors that affect the performance parameter [40,41]. The use of an adequate experimental design such as the Taguchi method for WCO pyrolysis is particularly important. To the best of our knowledge, no other work has been reported on the utilization of this heterogeneous acid, sulphated-ferric (II) oxide/alumina oxide catalyst via catalytic deoxygenation of WCO. Herein, in this study, the Taguchi method was used to investigate the optimal conditions for catalytic deoxygenation of WCO using sulphated-ferric (II) oxide/alumina oxide as a deoxygenation catalyst. This paper demonstrates the usefulness of using Taguchi coupled with product pyrolysis with GC-MS to predict pyrolysis yields with a great reduction in the number of experiments. Table 1\n represents the nomenclature of this paper.Alumina-supported-sulphated-ferric oxide (II) (SO4\n2--Fe2O3/Al2O3) catalyst (patented commercial catalyst Patent File No.: PI 2017702072 [42]) with purity >90% was provided by Catarim Sdn Bhd, Malaysia. The physicochemical properties of these heterogeneous catalysts are summarised in Table 2\n. The specific surface area and pore distribution of the synthesized catalysts were evaluated using a Brunauer-Emmett-Teller (BET) method using Thermo-Finnigan Sorpmatic 1990 series with an N2 adsorption/desorption analyzer. The particle size distribution of the synthesized catalyst was measured by laser diffraction (Malvern Mastersizer Hydro 2000S, Malvern Instruments Ltd., UK). While the acidity of the synthesized SO4\n2--Fe2O3/Al2O3 catalysts were investigated using temperature-programmed desorption with NH3 as probe molecules. The NH3-TPD analysis was carried out by using Thermo Finnigan TPD/R/O 1100 instrument equipped with a thermal conductivity detector (TCD. Based on Table 2, Brunauer-Emmett-Teller (BET) analysis showed that SO4\n2--Fe2O3/Al2O3 catalyst is classified as a macroporous structured catalyst with an average pore diameter >50\u00a0nm (67.77\u00a0nm). This heterogeneous catalyst has a small pore volume of around 0.19\u00a0cm3/g which will limit the rapid access of reactant into the porosity of the catalyst and the small amount of coke will be expected to form after the reaction. Besides that, through its relatively high surface area of 109.87\u00a0mg2/g, this catalyst will provide sufficient contact area between the WCO substrate and active sites for the reaction to take place. The particle size of the SO4\n2--Fe2O3/Al2O3 catalyst is smaller than 28\u201329\u00a0nm. The smaller the catalyst particle size, the larger the surface area for a given mass of particles as shown in Table 2. Based on a previous study conducted by Ng et al. [43], SO4\n2--Fe2O3/Al2O3 catalyst has composed of mild weak acid sites and a majority of strong acid sites with 1.03\u00a0\u00d7\u00a01021 atom/g. As reported by Nur Azreena et al. [44], weak or medium acid sites played an important role in the removal of oxygenating species, while strong acidic sites possessed the ability to catalyse the cracking reaction and dehydrogenation. This justified that SO4\n2--Fe2O3/Al2O3 catalyst has the potential to be used as a deoxygenation heterogeneous acid catalyst in green diesel production via deoxygenation of WCONitrogen gas (N2) with 99% purity was supplied by Smart Biogas Sdn Bhd. GC-MS analysis using n-hexane with purity >98% from Merck was utilized. The feedstock i.e. palm oil waste cooking oil (WCO) was collected from a local restaurant in Serdang, Selangor. The WCO was filtered, centrifuged at 3000\u00a0rpm for 30\u00a0min and heated to 100\u00a0\u00b0C prior to experimental work. To eliminate the differences due to feedstock differences, all the pyrolysis tests used the WCO from the same batch. The major composition of WCO consists of high fatty acids such as oleic acid (43.2%), linoleic acid (30.1%) and palmitic acid (19.4%) as summarised in Table 3\n. As reported by Hafriz et al. [45], these fatty acid structures were still maintained even after deep frying due to high boiling points and these fatty acid contents are the major indicators of the properties of biofuels produced via catalytic deoxygenation.Catalytic deoxygenation processes were conducted at temperatures ranging from 350\u00a0\u00b1\u00a05 to 450\u00a0\u00b1\u00a05\u00a0\u00b0C in a three-neck round bottom flask along with varied reaction parameters. Fig. 1\n [16] displays the apparatus setup of the experiment. Both the WCO feedstock (150\u00a0g) and SO4\n2--Fe2O3/Al2O3 catalyst (according to the experimental design) were placed in the triple neck glass reactor, stirred at 400\u00a0rpm continuously and heated at the rate of 20\u00a0\u00b0C/min with a flow of 10\u201320\u00a0cm3/min of N2. The thermocouple was calibrated to control the temperature of the reactor and the tip was inserted into the WCO feedstock. In the catalytic deoxygenation process, the WCO cracked and vaporized when the reaction temperature was reached. The vapour left the reactor through the graham coil rectification column and condensed in the second graham coil condenser unit. The condensed liquid products were collected in the receiving flask (round bottom flask) and the residue (catalyst and coke) was left in the reaction flask. The experiment was allowed to settle and cooled for approximately 30\u00a0min for each experiment, and the liquid product from the condensation of the oil was analysed using the GC-MS analysis. All the products and glassware were weighed before and after the experiment for mass balance calculation.In the present work, the Taguchi method (Minitab 16 software) was used to design optimized WCO cracking experiments with the temperature (\u00b0C), catalyst loading (wt.%), residence time (min) and N2 flow rate (cm3/min) as the independent variables, while the selected output response was the yield (%) of the biodiesel fractions. The Taguchi method suggested a set of 9 experiments for optimization which is a significant reduction from the original 81 experiments without the use of the Taguchi method. Table 4\n represents the suggested L9 orthogonal array for the waste cooking oil cracking experiments.Gas Chromatography-Mass Spectrometry (GC-MS) was used to analyse the composition of cracked liquid oil. The samples of pyrolysis oil (PO) were diluted with GC-grade n-hexane to 100\u00a0ppm. The obtained oil was analysed qualitatively and quantitatively in a non-polar ZB-5MS model column (30\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm I. D x 0.25\u00a0\u03bcm film thickness) in a split mode. The oven temperature was programmed to hold at 40\u00a0\u00b0C for 3\u00a0min, ramp at 7\u00a0\u00b0C/min to 300\u00a0\u00b0C and hold at 300\u00a0\u00b0C for 5\u00a0min. The injector temperature was set at 250\u00a0\u00b0C and the flow rate of the He carrier gas was 3.0\u00a0cm3/min. A different class of compound present in the liquid oil were identified using the National Institute of Standards and Testing (NIST) library. As reported by Hafriz et al. [46,47], the identification of the feedstock and the major products of pyrolysis oil using GC-MS analysis was based on the probability match between 95% and 100%. The WCO, product yield of pyrolysis oil and hydrocarbon selectivity can be determined by comparing the peak areas of the chromatogram as it is proportional to the relative content of the products as shown in Eq. (1) and Eq. (2), respectively [48,49].\n\n(1)\n\n\nY\ni\ne\nl\nd\n\no\nf\n\nh\ny\nd\nr\no\nc\na\nr\nb\no\nn\n\n\n(\n%\n)\n\n=\n\n\nT\no\nt\na\nl\n\na\nr\ne\na\n\no\nf\n\n\nC\n8\n\n\u2212\n\nC\n24\n\n\n\n\u2211\na\nr\ne\na\n\no\nf\n\nt\no\nt\na\nl\n\np\nr\no\nd\nu\nc\nt\n\n\n\n\n\n\n\n\n(2)\n\n\nH\ny\nd\nr\no\nc\na\nr\nb\no\nn\n\ns\ne\nl\ne\nc\nt\ni\nv\ni\nt\ny\n\n\n(\n%\n)\n\n=\n\n\nD\ne\ns\ni\nr\ne\nd\n\nh\ny\nd\nr\no\nc\na\nr\nb\no\nn\n\nf\nr\na\nc\nt\ni\no\nn\n\n\n\u03a3\n\na\nr\ne\na\n\no\nf\n\nh\ny\nd\nr\no\nc\na\nr\nb\no\nn\n\n\n\nx\n\n100\n\n%\n\n\n\n\nThe reaction conditions and the activity of synthesized sulphated-ferric (II) oxide/alumina oxide catalyst (patented solid acid catalyst) in the catalytic deoxygenation of waste cooking oil (WCO) were investigated based on the design given by the Taguchi method. In the receiving flask, two phases of liquid were detected in the flask, which was the pyrolysis oil (PO) on the top and the acid phase at the bottom layer. The liquids were separated by using the decantation technique. From Table 5\n, the PO yield ranged from 37.88% to 75.86%, varied based on the parameter, mostly affected by the temperature. In addition, the presence of SO4\n2--Fe2O3/Al2O3 catalyst has improved the yield of PO due to the strong acidic sites of the catalyst which possessed the ability to catalyse the cracking reaction and dehydrogenation.No liquid product was obtained at 350\u00a0\u00b0C (Runs 2, 4 and 9) as the vapour was unable to vaporise through the column of the condenser. Hence, from this point onwards, all the runs from 350\u00a0\u00b0C (Run 2, 4 and 9) had been opt-out from the figures. It is speculated that below 400\u00a0\u00b0C, the conversion of acetic acid was reduced due to catalyst deactivation, and the deoxygenation process was incomplete. At low temperatures or atmospheric pressure, deactivation of the catalyst might occur and faster carbon deposition occurred on the catalyst surface which affected the selectivity and the quality of the gasoline phase [50]. The same observation was reported by Lam et al. [51] in which no liquid product was obtained after 1\u00a0h of catalytic deoxygenation reaction at 350\u00a0\u00b0C. However, other works obtained pyrolysis oil at 300\u00a0\u00b0C with Co3O4\u2013La2O3/ACnano catalyst [52] and at 350\u00a0\u00b0C with NiO\u20135CaO/SiO2\u2013Al2O3 catalyst [33], in which the type of catalyst used might be played a better role for the deoxygenation reaction or the feedstock load was higher than in this present work.It was observed that all reactions at a higher temperature (450\u00a0\u00b0C) gave the highest PO yield (75.86% in Run 8), which shows that temperature was the most influencing parameter for obtaining a higher amount of PO. The high PO yield was obtained accompanied by the low amount of coke and slightly higher amount of gaseous product compared to other temperatures. The PO yield amount was only slightly lower than that obtained by Wako et al. [37] (83% PO yield) with zirconium oxide as another type of acid catalyst at 475\u00a0\u00b0C used in the catalytic deoxygenation reaction. However, the PO obtained at 450\u00a0\u00b0C appeared to be darker than the product at a lower temperature as displayed in Fig. 2\n. The colour changes were observed when the temperature rose from 430 to 450\u00a0\u00b0C during the experiment. As reported by Makcharoen et al. [53], the colour of the liquid product was changed with increasing reaction temperature and the colour of pyrolysis oil did not significantly change when the reaction temperature was below 400\u00a0\u00b0C. However, the samples appeared distinctively darker at 420\u00a0\u00b0C possibly due to the contribution of the thermal cracking reaction that occurred in the deoxygenation of crude palm kernel oil (CKPO). According to Shurong et al. [50], the dark colouration was due to the high amount of oxygenates in the product, which is undesirable in biofuel production. Wako et al. [37] also reported the unsuitability of using higher temperatures for biofuel production as it leads to secondary cracking that shall favour gaseous products.\nFig. 2 clearly showed that the catalytic reaction at a temperature of 400\u00a0\u00b0C produced a lighter yellow colour of PO despite the variations in the reaction parameters. The highest PO yield observed at 400\u00a0\u00b0C was 42.66% at Run 6. Although results at 400\u00a0\u00b0C were the most promising in terms of PO obtained, a high amount of coke was also obtained at 400\u00a0\u00b0C.In the present work, 1\u00a0wt% of catalyst loading gave the highest yield of product across different temperatures and other parameter variations. Increasing the catalyst loading in some cases increases the yield of biodiesel but overloading beyond the optimum amount may lead to non-proper mixing and over-saturate the catalyst's active sites. Besides the parameter investigated in this study, Shurong et al. [50] showed that a pressurised system could also promote oil phase yield, 32.2% of PO at 3\u00a0MPa compared to 10.8% of PO at 0\u00a0MPa.The presence of high oxygenates in the brownish-coloured oil obtained at 450\u00a0\u00b0C was confirmed by the GC-MS analysis, presented in Fig. 3\na. From Fig. 3a, the oxygenates were high in carboxylic acid and alcohol content, while a minor quantity of ketone was detected in most of the runs except in Run 6. The highest amount of carboxylic acid (mainly C12 lauric acid) was found in Run 7, while traces of other compounds including aldehydes and esters found in three runs with the sequence of Run 7\u00a0>\u00a0Run 8\u00a0>\u00a0Run 1. A high carboxylic acid value indicates a high acidity value of the pyrolysis oil. Hence it can be deduced that Run 7 has the highest acidity value and vice versa for Run 6. The high amount of carboxylic acid obtained in the present work might be affected by the heterogeneous acid catalyst used [16], but as shown in Fig. 3a, it can be minimised by parameter control.The hydrocarbon composition is dominated by aliphatic hydrocarbons which were alkene and alkane, followed by cyclic hydrocarbons; cycloalkane and cycloalkene as depicted in Fig. 3b Catalytic deoxygenation reaction may take place through decarbonylation (Eq. (2)) or via decarboxylation (Eq. (3)), which is usually influenced by the catalyst type. The high amount of alkane in most of the results shows domination by a decarboxylation reaction. In their work, Hafriz et al. [16] and Ali et al. [54] revealed that alkene is the main product from the WCO deoxygenation reaction with Ni-dolomite catalyst which they concluded that deoxygenation via decarbonylation dominates the conversion reaction.\n\n(3)\nDecarbonylation: R\u2013COOH \u2192 Cn-H2n(Alkenes)\u00a0+\u00a0CO\u00a0+\u00a0H2O (acid phase)\n\n\n\n\n(4)\nDecarboxylation: R\u2013COOH \u2192Cn-H2n+2 (Alkanes)\u00a0+\u00a0CO2\n\n\n\nNo traces of aromatic compounds were detected in the runs except for Run 1 (0.73%) and Run 7 (0.71%) as depicted in Fig. 3b. Aromatic compounds detected in the pyrolysis oil were from benzene, naphthalene, xylene and their derivatives (butylbenzene, pentylbenzene, ethylbenzene). Aromatics are essential in gasoline or petrol fuel as they gave them high octane rating and have higher knocking resistance. A small amount of aromatic compound in a form of xylene and toluene in aviation fuel will prevent fuel from freezing at elevated temperatures. In biomass-source diesel, low aromatic compounds are desirable as aromatics are carcinogenic and posed health-threat to humans [55].The result in Fig. 3b shows high hydrocarbon content in the PO, but also accompanied by a high amount of coke as given in Table 4. In the deoxygenation of WCO, coke may be produced by two solid phase reactions: aromatic hydrocarbon polymerization (Eq. (5)) or condensation of WCO (Eq. (6)) [16].\n\n(5)\nPolymerization: Cn-Hn (Aromatics) \u2192 Carbon(s)\n\n\n\n\n\n(6)\nCondensation: WCO \u2192 Carbon(s)\n\n\n\nIt was suggested that the acidity of the catalyst might enhance both polymerization and cyclization [33]. The two runs with the highest coke amount which were Run 5 and Run 6 also had a higher yield of cycloalkane as in Fig. 3b, which verifies the claim [33]. According to Asikin-Mijan et al. [33], acidic and transition metal catalysts have high deoxygenation activity but at the same time, are more susceptible to coke formation compared to basic catalysts.\nFig. 4\n shows the selectivity of the hydrocarbon chain from C8 to C24 of the PO and the corresponding hydrocarbon liquid fraction to the biofuel range (gasoline, kerosene and diesel) based on the GC-MS analysis. Typical palm-oil-based WCO contains carboxylic acids in the range of C16 and C18 which reflects a high amount of palmitic acid (C16) and oleic acid (C18) [16,33]. From the decarboxylation/decarbonylation reaction (Eq. (1) and Eq. (2)), the product should contain liquid hydrocarbon in the range of C15\u2013C17. From Fig. 4, the highest amount of liquid hydrocarbon detected in the range of C11\u2013C17, mixtures of the long and shorter hydrocarbon chain, indicates the presence of deoxygenation and catalytic cracking reaction. The experiment runs at 450\u00a0\u00b0C (Run 1, 3 and 7) had a shorter hydrocarbon chain (<C11), indicating catalytic cracking was favoured at higher temperatures as opposed to deoxygenation reaction. Shimada et al. [56] suggested that the after deoxygenation reaction where carbon was lost in a form of CO (decarbonylation) or CO2 (decarboxylation), intermediate hydrocarbon was formed and catalytic cracking of these intermediate products produced light hydrocarbons.The liquid hydrocarbon from the GC-MS analysis could be further classified into a range of alkenes C10\u2013C19, alkanes C10\u2013C22, cycloalkenes C10\u2013C13, and cycloalkanes C11\u2013C15. Liquid hydrocarbon obtained was in accordance with the petroleum fractions [57]; where a range of C5\u2013C10 indicates the presence of gasoline, while C11\u2013C12 is in the range of kerosene-like fractions, C13\u2013C17 is in the range of light diesel-like fractions, and C18\u2013C25 is in range of heavy diesel-like fractions.\nFig. 4 strikingly shows that the diesel fractions were the highest in the various experimental runs conducted, followed by kerosene and gasoline. Run 6\u00a0at the reaction temperature of 400\u00a0\u00b0C, 90\u00a0min of reaction time, 20\u00a0cm3/min N2 flow and 1\u00a0wt% catalyst loading had the highest yield of 49.66% of green diesel fractions. This was followed by Run 5 and Run 3 with 33.93% and 39.98% respectively and the remainder of runs (1,7 and 8) falls within the ranges of 20%\u201326%.ANOVA was used in the Taguchi method to assess the result from catalytic deoxygenation and to determine the influence of every factor on the product yield [58]. ANOVA is just like the regression analysis, used to access the relationship among the response variables [59].\nFig. 5\n\n shows an analysis of variance (ANOVA) results on the parameter contributions to produce biodiesel, tabulated from Table 3. The temperature was the most significant attribute with a per cent contribution of 86.62%, while the catalyst contributed the least, which influenced only 1.53%. As the experimental study did not compare with a blank experiment (i.e. without a catalyst), resulted in a less apparent influence of the catalyst on the diesel yield. Due to this, Taguchi's analysis of the influence of time and nitrogen flow rate had a greater impact than the catalyst tested.\nTable 6\n illustrates the ANOVA analysis obtained from the Taguchi method, using Minitab 16 software. The effect of temperature on the diesel fraction yield was significant, the model showed that the F-value of 17.61 implies the model was significant and Prob\u00a0>\u00a0F less than 0.0500 indicates the model terms were significant. If Prob\u00a0>\u00a0F values are greater than 0.100, the system indicated the model as insignificant for the regression model.From Table 7\n\n, the fit of the model can be confirmed by R-squared values. The Predicted R-Squared of 0.7279 was in reasonable agreement with the Adjusted R-Squared of 0.8925. This verifies that this model was statistically significant. Additionally, adequate precision estimates the signal-to-noise ratio, this helps in determining the validity of the model. It is desirable when the ratio is greater than 4. The ratio of 11.265 indicates an adequate signal. Therefore, it is convinced that the model was significant. Therefore, the above explanation of the ANOVA results deduced that this model could be used for deoxygenation and catalytic cracking reaction in this work, hence can be used to operate in the design space in terms of green diesel production [58].\nTable 9\n below shows the suggested parameter by Taguchi which predicts the diesel fraction yield of 44.78%. Experimental work according to the suggested conditions was carried out to validate the optimized suggested parameter. The targeted result obtained 38.96% of the diesel fraction. This showed a reasonable agreement between the experimental and predicted results under the conditions given to be optimized. Fig. 7\n shows the selectivity comparison between the targeted result from Taguchi and Run 6 which shows the similar height of C14 peaks, the lower peak of C11 and C17, but a broader range of C8\u2013C10 fraction. It is speculated that the gap between the experimental and predicted result may result from a more catalytic deoxygenation reaction favoured with the slightly altered parameter, which in this case, with lower N2 flowrate, from 20\u00a0cm3/min in Run 6\u201310\u00a0cm3/min of N2.As a comparison study, the effect of individual parameters is investigated while maintaining other process parameters constant at unspecified levels. Wako et al. [37] reported the optimum condition achieved with 83\u00a0wt% of pyrolysis oil obtained with 4\u00a0wt% ZrO2 as a catalyst, reaction temperature of 475\u00a0\u00b0C, 20\u00a0min of resident time, and 10\u00a0\u00b0C/min heating rate [37]. In the work by Wako et al. [37], the effect of N2 flow was not studied. It was proved in this study that a high temperature of reaction gave the highest yield of pyrolysis oil (>70\u00a0wt%). On the other hand, with a continuous feed system, Jungjaroenpanit and Vitidsant [35] reported optimum conditions at 430\u00a0\u00b0C, 154.20\u00a0mL/h flow rate of raw material (WCO), 102.73\u00a0cm3/min N2 flow rate, and 60\u00a0wt% catalyst loading resulted in the production of 57.07% diesel fraction. Only Chen et al. [39] and Ahmad et al. [59] reported optimization work for pyrolysis by the Taguchi method. With castor meal as the feedstock, Chen et al. [39] reported that the effective order of pyrolytic parameters in their work was nitrogen flow rate\u00a0>\u00a0heating rate\u00a0>\u00a0pyrolytic temperature\u00a0>\u00a0residence time. Whereas, Ahmad et al. [59] stated that all three parameters investigated gave a significant impact on the bio-gasoline yield in the following order; temperature\u00a0>\u00a0time\u00a0>\u00a0catalyst loading. As compared to this study, reaction temperatures gave a significant impact on the deoxygenation of WCO using SO4\n2--Fe2O3/Al2O3 catalyst followed by reaction time\u00a0>\u00a0nitrogen flow\u00a0>\u00a0catalyst loading.Waste cooking oil could be catalytically deoxygenated and cracked using a solid acid catalyst (sulphated-ferric (II) oxide/alumina oxide catalyst) into bio-gasoline, bio-kerosene and green diesel fractions at different reaction conditions and optimized with L9 orthogonal array by Taguchi method. The temperature of 400\u00a0\u00b0C gave a better-quality product with a PO yield of 42.66% whereby the liquid hydrocarbon yield was as high as 78.45% and less amount of oxygenated compound of 21.55%. The high liquid hydrocarbon gave rise to high biodiesel fractions of 49.66% and gasoline 28.79% whereas a temperature of 450\u00a0\u00b0C awarded the highest yield of 75.86% with a high oxygenated compound of 63.89% and less liquid hydrocarbon yield of 36.11%. As no result was obtained at 350\u00a0\u00b0C, it is safe to deduce that the temperature was too low for the reaction to occur with the catalyst used. Optimization of the catalytic deoxygenation was determined using the Taguchi method with the optimum conditions, temperature 400\u00a0\u00b0C, catalyst loading 1\u00a0wt%, time 90\u00a0min and the N2 flow rate 20\u00a0cm3/min. It can be concluded that the catalytic deoxygenation process of WCO was achieved and the DoE and optimization of operating conditions are realistic with fewer numbers an experiment. Finally, the predicted optimized values and the actual yield of pyrolysis oil data obtained were in close agreement.\nShafihi U: Conceptualization, Investigation, Methodology, Software, Formal analysis, Validation, Writing - original draft. R.S.R.M. Hafriz: Conceptualization, Investigation, Data curation, Methodology, Visualization, Writing \u2013 review & editing. N.A. Arifin: Conceptualization, Investigation, Validation, Writing \u2013 review & editing. Nor Shafizah I: Conceptualization, Investigation, Validation, Writing \u2013 review & editing. Idris A: Project administration, Resource. A. Salmiaton: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Writing \u2013 review & editing. N.M Razali: Project administration, Resource.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors acknowledge the financial support from the Ministry of Higher Education of Malaysia for the Fundamental Research Grant Scheme (FRGS/11/TK/UPM/02), the AAIBE Chair of Renewable Energy Grant No. 201801KETTHA and 202102 KETTHA for funding this research publication.", "descript": "\n                  This work investigates the optimization of reaction parameters with the Taguchi method for catalytic deoxygenation of waste cooking oil (WCO) as an alternative renewable fuel process. Commercial sulphated-ferric (II) oxide/alumina oxide catalyst has the potential as a deoxygenation catalyst due to its good physicochemical properties which enhance the removal of oxygenated species. The obtained pyrolysis oil analysed by GC-MS revealed the selectivity of the pyrolysis oil mostly in the range of light diesel and kerosene fraction. From an analysis of variance (ANOVA), temperature awarded the most significant impact (86.62%) in this catalytic deoxygenation as compared to three other parameters followed by reaction time\u00a0>\u00a0N2 flow\u00a0>\u00a0catalyst loading. From the GC-MS analysis, the maximum renewable diesel fraction of 49.66% was obtained at 400\u00a0\u00b0C, 1\u00a0wt% of catalyst, 90\u00a0min of reaction time and 20\u00a0cm3/min N2 flow. The predicted model by Taguchi in the present study validated by the experimental work shows a promising application in optimising the catalytic pyrolysis process for future use.\n               "}
{"full_text": "Energy utilization has become a necessity for the normal functioning of human life, growth, and survival. The primary energy consumption across the globe is growing day by day due to the increasing energy usage in industries, buildings, and transportation sectors (Balajii\u00a0and Niju,\u00a02019). Presently, most of the energy comes from fossil fuel that is finite. The combustion of fossil fuels emits various greenhouse gasses like CO2, CO, SO2, NOx, etc. which have adverse effects on the environment and are the primary factors for global warming in the 21st century. To protect nature and human beings, sustainable and suitable alternative energy sources to fossil fuels are in utmost need. Utilization of bioethanol and biodiesel as alternative to fossil fuels has attracted global attention due to their renewability, carbon-neutral character, and good combustion efficiency (Dehkhoda\u00a0et\u00a0al., 2010; Ezebor\u00a0et\u00a0al., 2014). Chemically, biodiesel is known as the fatty acid methyl esters (FAME) commonly produced via catalytic esterification (Scheme\u00a01) and transesterification (Scheme\u00a02) reactions from biological sources like animal fats, non-edible oils, edible oils, and fatty acids (Brahma\u00a0et\u00a0al., 2022; Basumatary\u00a0et\u00a0al., 2022).Biodiesel is widely endorsed as an alternative to fossil fuels because of its better performance and eco-friendly nature. However, the main obstacle for biodiesel to be used as an alternative source of fossil fuels is due to its higher price characteristics compared to diesel (Knothe,\u00a02002). Therefore, recent research aimed at reducing the production cost of biodiesel which is mainly caused by the high price of feedstocks (such as soybean oil, sunflower oil. etc.) used for biodiesel production and catalyst used for its production (Konwar\u00a0et\u00a0al., 2014a). In search of viable feedstocks of biodiesel, several oil sources including edible and non-edible oils for biodiesel synthesis have been reported by several researchers. Since the non-edible feedstocks are low valued, locally available throughout the year, and do not create the food vs. fuel impingement, they are consistently targeted as a viable feedstocks for the preparation of biodiesel.Catalyst can increase the rate of a reaction and plays a very important role in biodiesel synthesis in terms of efficiency and production cost. The catalysts which are applied in the synthesis of biodiesel are classified into three classes viz. homogeneous, enzyme and heterogeneous catalysts. Homogeneous catalysts are of two types, they are homogeneous acid and homogeneous base. Homogeneous acid catalysts are not commonly employed in the reaction for the synthesis of biodiesel due to the requirement of the high ratio of alcohol to oil, high pressure, and high reaction temperature, and also due to the problems linked to the reusability of catalyst (Canacki and Van Gerpen, 1999). Though the transesterification reaction time is shorter for homogeneous base-catalyzed reaction, it possesses difficulty due to soap formation during conversion producing a low yield of biodiesel and the catalyst cannot be separated easily (Basumatary\u00a0et\u00a0al., 2018). Because of higher costs and time-consuming reactions, enzyme catalysts are not feasible for broad application in biodiesel production (Sandoval\u00a0et\u00a0al., 2017; Kalita\u00a0et\u00a0al., 2022).Heterogeneous catalysts are also divided into two categories viz. solid acid and base catalysts. Due to the simpler and fast reaction process even at low temperatures, easy separation process and reusability for several cycles of reactions, heterogeneous base catalyst can reduce the overall production cost of biodiesel, and hence, it has been gaining importance in recent years (Kondamudi\u00a0et\u00a0al., 2011; Konwar\u00a0et\u00a0al., 2014a). Nowadays, researchers prefer heterogeneous catalysts prepared from renewable natural sources because of their renewability, catalytic efficiency, and eco-friendliness. Few researchers reported the biodiesel production using various biomass-based heterogeneous catalysts such as palm kernel fronds (Ameen\u00a0et\u00a0al., 2013), Musa paradisiacal\u00a0peel (Betiku\u00a0and Ajala,\u00a02014),\u00a0Musa balbisiana peels and underground stem (Gohain\u00a0et\u00a0al., 2017; Sarma\u00a0et\u00a0al., 2014; Aslam\u00a0et\u00a0al., 2014), banana peels, and cocoa pod husk (Odude\u00a0et\u00a0al., 2019), Musa acuminata peel (Pathak\u00a0et\u00a0al., 2018), coconut waste (Sulaiman\u00a0and Ruslan,\u00a02017), wood ash (Sharma\u00a0et\u00a0al., 2012), Lemna perpusilla (Chouhan\u00a0and Sarma,\u00a02013), rubber seed shell (Onoji\u00a0et\u00a0al., 2017), camphor tree (Li\u00a0et\u00a0al., 2018), rice husk ash (Chen\u00a0et\u00a0al., 2013; Zeng\u00a0et\u00a0al., 2014),\u00a0Musa paradisiaca\u00a0 (Basumatary\u00a0et\u00a0al., 2021a), sugarcane bagasse (Basumatary et\u00a0al., 2021b),\u00a0Heteropanax fragrans\u00a0 (Basumatary et\u00a0al., 2021c),\u00a0Sesamum indicum\u00a0plant (Nath\u00a0et\u00a0al., 2020), Brassica nigra\u00a0plant (Nath\u00a0et\u00a0al., 2019),\u00a0etc.Several researchers also reported various heterogeneous base catalysts which include oxides of metals such as CaO (Refaat,\u00a02011; Verma\u00a0and Sharma,\u00a02016), MgO (Refaat,\u00a02011; Vyas\u00a0et\u00a0al., 2010), SrO (Refaat,\u00a02011), and transition metal oxides and derivatives such as ZrO2 (Zabeti\u00a0et\u00a0al., 2009; Mahdavi\u00a0et\u00a0al., 2015), TiO2, ZnO2, zeolite (Lee\u00a0et\u00a0al., 2009) and basic hydrotalcite (Silva\u00a0et\u00a0al., 2010). However, heterogeneous base catalysts also have some insufficiencies like leaching of catalyst and inappropriateness of the feedstocks due to high free fatty acids (FFAs) contents (Tangy\u00a0et\u00a0al., 2016; Wilson\u00a0and Lee,\u00a02012). Heterogeneous acid catalysts are developed for biodiesel synthesis from the feedstocks with FFAs contamination, and also for the simultaneous reactions of esterification and transesterification (Patel\u00a0and Narkhede,\u00a02012). Nonetheless, the heterogeneous acid catalyst has also some disadvantages like it requires higher reaction time and temperature in comparison to heterogeneous base catalysts. To resolve these problems and to develop a more efficient catalyst for biodiesel synthesis, researchers are now investing their time in the preparation of nanocatalysts due to their high surface areas, higher catalytic activity and longer stability. A nanocatalyst is a nanoparticle with size ranging between 1 and 100\u00a0nm. The nanocatalyst can be synthesized by different methods such as hydrothermal method, wet impregnation method, co-precipitation method, precipitation method, sol-gel method, combustion method, and self-polymerization based grafting technique. Due to the small size of the particle, the surface-to-volume ratio increases which leads to the accumulation of a higher number of atoms on the surface of the catalyst (Zambre\u00a0et\u00a0al., 2012). The present work is aimed to review the recent applications of the different types of nanocatalysts employed in biodiesel synthesis which include metal oxide nanocatalysts (CaO, ZnO, MgO, TiO2, CuO, and ZrO2), magnetic nanocatalyst, bifunctional nanocatalyst, nano-zeolite, nano-hydrotalcite, etc. The catalytic performances of these nanocatalysts are highlighted and discussed herein.Biodiesel is a renewable diesel fuel that contains short chain esters (methyl or ethyl) made by transesterification process in the presence of a suitable catalyst. Heterogeneous catalysts are widely used due to easy recovery of the catalysts by simple gravity filtration which has a great contribution in reducing the overall biodiesel production cost. The metal oxide catalyst gives a high yield of biodiesel and requires less time for the transesterification reaction (Vasi\u0107 et\u00a0al., 2020). The higher catalytic activity of metal oxide is attributed to its high surface property (Gawande\u00a0et\u00a0al., 2012). Besides alkaline earth, the metal oxides, transition metal oxides such as zinc oxide, iron oxide, tin oxide, and zirconium oxide are widely used as a catalyst.The use of CaO nanocatalyst deals with the effective, economic, and eco-friendly conversion of vegetable oil and animal fat into biodiesel. The catalyst is subjected to different calcination temperatures and their effects on complete transesterification of different oils at different concentrations are discussed. Many researchers found naturally available CaO from waste materials such as dolomite, snail shell, waste eggshell, waste mussel shell, chicken bone (Nakatani\u00a0et\u00a0al., 2009; Diaz\u00a0and Borges,\u00a02012; Niu\u00a0et\u00a0al., 2014; Moradi\u00a0and Mohammadi,\u00a02014). The performances of CaO nanocatalysts derived from several renewable sources in biodiesel synthesis from different feedstocks are represented in Table\u00a01\n. In this table, the catalyst preparation methods and conditions along with the particle size (nm) and surface area (m2 g\u22121) of the prepared nanocatalysts were mentioned. It was found that various parameters such as reaction time, temperature, and catalyst loading/amount for various feedstocks and various catalysts influenced the FAME yield. Reddy\u00a0et\u00a0al.\u00a0(2006) investigated the catalytic activity of nanocrystalline CaO, a powder having a crystallite size of 20\u00a0nm with the specific surface area of 90 m2 g\u22121, in the reaction of poultry fat and soybean oil at room temperature.The nanocrystalline material CaO showed much superior activity than the laboratory-grade CaO and it was due to higher surface area of the nanocrystalline material. It was observed that when pure CaO is exposed to the atmosphere its surface basic sites get poisoned because it absorbed CO2 and H2O which were converted to CaCO3 and Ca(OH)2. To improve basic strength and to remove poisoned species, CaO was calcined using ammonium carbonate solution at higher temperatures (>750 \u00b0C) (Kesic\u00a0et\u00a0al., 2012), and above 850 \u00b0C, calcium carbonate decomposes into calcium oxide and carbon dioxide. Zik\u00a0et\u00a0al.\u00a0(2020) investigated the coconut residue and chicken bone-derived catalyst of nano-crystal cellulose (NCC) and CaO for biodiesel synthesis from waste cooking oil (WCO). They prepared the catalyst by calcining the dried, crushed, and blended bone at 700 \u00b0C, 800 \u00b0C, and 900 \u00b0C for 4, 5, and 6\u00a0h, respectively. They carried out the transesterification reaction in a packed bed reactor in the presence of CaO/NCC/PVA (PVA\u2013polyvinyl alcohol). They found the highest biodiesel yield of 98.4% under the optimum reaction conditions (ORCs) of 65 \u00b0C of temperature, 6:1 of methanol to oil molar ratio (MTOMR), and 0.5\u00a0wt.% of catalyst loading. This study also reported that the catalyst was reusable up to the 4th cycle with a product yield of above 90%.\nKrishnamurthy\u00a0et\u00a0al.\u00a0(2020) demonstrated the synthesis of CaO nanocatalyst from a snail shell for biodiesel production from dairy scum and\u00a0Hydnocarpus wightiana\u00a0oil. They prepared the catalyst by hydrothermal method, in which the clean and dry snail shells were finely crushed using a blender, washed with nitric acid for 3\u20134\u00a0min, the snail shell was meshed and rinsed with distilled water and then dried and calcined at 900 \u00b0C for 4\u00a0h. The average crystallite size of nano CaO catalyst was found to be 40\u00a0nm. In the BET analysis, the surface area was found to be 9.37 m2 g\u22121, and the average pore diameter and pore volume were 2.29\u00a0nm and 0.0538 cm3 g\u22121, respectively. They found that the maximum biodiesel yield for scum oil and\u00a0Hydnocarpus\u00a0wightiana oil was 96.92% and 98.93% at the ORCs of MTOMR of 12.7:1 and 12.4:1, catalyst dosage of 0.866 wt% and 0.892 wt%, reaction temperatures of 58.56 \u00b0C and 61.6 \u00b0C, and the reaction time of 2\u00a0h and 2.42\u00a0h, respectively. They also found that the catalyst can be reused up to 5th catalytic cycle though the product yield decreased slightly and after the 5th cycle, the yield declined to a greater extent due to leaching which results in a decrease in surface area, pore volume, and total basicity of the catalyst.\nKaur\u00a0and Ali\u00a0(2011) investigated lithium-ion impregnated calcium oxide as a nanocatalyst for biodiesel production from karanja and\u00a0jatropha\u00a0oils. They prepared the Li-CaO nano-catalyst by wet impregnation method in which CaO was suspended in deionized water, and an aqueous solution of LiNO3 was added, then stirred for 2\u00a0h and evaporated to dryness and heated at 120 \u00b0C for 24\u00a0h. The CaO impregnated with 1.75 wt% of lithium was used as a solid catalyst for the transesterification of karanja and\u00a0jatropha\u00a0oils which contains 3.4 and 8.3 wt% of free fatty acids, respectively. They found that the surface area, pore-volume, and pore size of the catalyst were 1.7 m2 g\u22121, 0.004 cm3 g\u22121, and 95.02\u00a0\u00c5, respectively. The ORCs for the transesterification of karanja and\u00a0jatropha\u00a0oils were achieved in 1 and 2\u00a0h, respectively, MTOMR of 12:1, 5 wt% of catalyst, and a reaction temperature of 65 \u00b0C resulting in >99% conversion of oils to FAME. Zhao\u00a0et\u00a0al.\u00a0(2013) reported the transesterification of canola oil catalyzed by nanopowder CaO. They reported that the nano-CaO catalyst with the surface area of 89.25 m2 g\u22121 displayed a faster chemical reaction and adsorption due to a larger surface area and stronger basicity. They found that the ORCs for production of 99.85% of biodiesel yield was obtained at 2\u00a0h when 3\u00a0wt% of the catalyst was used with 9:1 MTOMR at 65\u00a0\u00b0C. They also investigated the reusability and lifetime of the nano-CaO catalyst under ORCs for 15 catalytic cycles although a slight decrease in yield was observed after the 10th cycle. After the 15th cycle, the yield dropped by around 70%. This decreased in biodiesel production was due to the loss of catalyst during the recovery process. Borah\u00a0et\u00a0al.\u00a0(2019a) synthesized the Zn substituted waste eggshell-derived CaO nanocatalyst for biodiesel production from WCO. They found that the maximum FAME conversion of 96.74% was obtained under the reaction conditions of 20:1 MTOMR, 5\u00a0wt% catalyst loading, 65\u00a0\u00b0C of reaction temperature, and 4\u00a0h of reaction time. In this experiment, they found that the catalyst was reusable for 5 consecutive cycles under ORCs. Seffati\u00a0et\u00a0al.\u00a0(2019) reported biodiesel production from chicken fat using CaO/CuFe2O4 nanocatalyst. The results of their study showed that a maximum biodiesel yield of 94.52% was obtained at the MTOMR of 15:1, a reaction time of 4\u00a0h, a reaction temperature of 70\u00a0\u00b0C, and a catalyst amount of 3%.\nBharti\u00a0et\u00a0al.\u00a0(2019) reported biodiesel production from soybean oil by the use of CaO nano-catalyst. From the BET analysis, they found that the average surface area and the pore diameter of the CaO nanocatalyst were 67.781 m2 g\u22121 and 3.302\u00a0nm, respectively. The average particle size of the catalyst found from the TEM image was ranging from 5.68 to 8.33\u00a0nm. They reported that the maximum biodiesel yield of 97.61% was found at 3.675 wt% of catalyst loading, 11:1 MTOMR at 60\u00a0\u00b0C within 2\u00a0h of reaction time. Ahmad\u00a0et\u00a0al.\u00a0(2020) demonstrated the synthesis of nano-CaO catalyst for biodiesel production from algal biomass (Chlorella pyrenoidosa). The catalyst was prepared by the use of calcination-hydration-dehydration method in which finely crushed powder of waste eggshells was calcined at 900 \u00b0C for 3\u00a0h in a muffle furnace. Characterization of catalyst showed that the average particle size was 23.65\u00a0nm, the surface area was 64.51 m2 g\u22121, and the average pore size was 9.28\u00a0nm. The catalyst at the operating conditions of 2.06 wt% of catalyst amount, 30:1 MTOMR, 60\u00a0\u00b0C of reaction temperature, and 180\u00a0min reaction time provided maximum biodiesel production of 93.44%. They also reported that the catalyst could be reused up to the 6th time and after that, the biodiesel production decreases rapidly due to the decrease in the number of active sites which was blocked by the byproduct. Erchamo\u00a0et\u00a0al.\u00a0(2021) investigated biodiesel production from WCO using eggshell-derived CaO nanocatalyst. Since the use of ethanol in the transesterification reaction arises with various problems such as emulsification and difficulty in the separation process, they used a mixture of methanol-ethanol. Ethanol has better solvability than methanol and hence, it can easily mix the oil, alcohol, and catalyst, while methanol can minimize the emulsification effect of ethanol. After considering various optimization parameters like catalyst amount, mixed methanol-ethanol (8:4) to oil ratio, reaction time, and reaction temperature as 2.5 wt%, 12:1, 120\u00a0min, and 60 \u00b0C, respectively, a biodiesel yield of 92% was obtained.\nGupta\u00a0and Agarwal\u00a0(2016) investigated biodiesel production from soybean oil using calcium nitrate (CaO/CaN) and snail shell (CaO/SS) derived CaO nanocatalyst. In the study, it was that CaO/SS catalyst was more basic than CaO/CaN that enhanced the catalytic activity. They reported that the maximum respective biodiesel yields of 96% and 93% were found at 8 wt% of catalyst loading, 12:1 MTOMR at 65\u00a0\u00b0C within 6\u00a0h of reaction time. They also reported that these catalysts were reusable up to the 5th transesterification cycle. Degirmenbasi\u00a0et\u00a0al.\u00a0(2015) reported biodiesel production from canola oil using CaO nanocatalyst. In the study, CaO nanocatalyst was prepared using the incipient-wetness impregnation process in which CaO was taken in a flask, and a vacuum was applied. Then K2CO3 solution was added to the CaO nanoparticles and dried at 393\u00a0K. Finally, the impregnated CaO particles were calcined at 773\u00a0K for 3\u00a0h. From the BET analysis, they found that the surface area of the catalyst was in the range of 10.24\u201314.65 m2\ng\u00a0\u2212\u00a01. They reported that the pore size and particle size of the catalyst were between 2 and 300\u00a0nm and 20\u2013160\u00a0nm, respectively. They found that the maximum biodiesel of 97.67% was obtained at the ORCs of 9:1 MTOMR, 3 wt% of catalyst amount, a reaction time of 8\u00a0h, and a reaction temperature of 65 \u00b0C. They also reported that the catalyst could be reused up to five successive times.The catalytic performances of MgO nanocatalysts in biodiesel synthesis from different feedstocks are shown in Table\u00a02\n. In this table, the methods of catalyst preparation along with the particle size (nm) and surface area (m2 g\u22121) of the prepared nanocatalysts were also mentioned. Amirthavalli\u00a0and Warrier\u00a0(2019) reported the production of biodiesel from WCO using MgO nanocatalyst. They prepared the nanocatalyst using the sol-gel method in which magnesium acetate tetrahydrate was dissolved in absolute ethanol under constant stirring. Oxalic acid was added to maintain the pH of the solution. The mixture was continuously stirred until they form a thick white gel. Then it was dried in an oven at 100 \u00b0C for 15\u00a0h. It was powdered using the mortar and calcined at 600 \u00b0C for 2\u00a0h. They carried out the transesterification in the presence of MgO nanocatalyst and found the highest yield of 80% biodiesel was produced at the ORCs of 60 \u00b0C, 10:1 MTOMR, and 2 wt% of catalyst loading. Vahid\u00a0and Haghighi\u00a0(2017) investigated biodiesel production from sunflower oil over MgO/MgAl2O4 nanocatalyst. They prepared the MgO/MgAl2O4 catalyst by the combustion method. The ORCs for the transesterification were achieved in 3\u00a0h, MTOMR of 12:1, and a reaction temperature of 110 \u00b0C, which resulted in 95.7% conversion of oils to biodiesel. They also found that the catalyst was reusable up to six successive rounds of reaction though the yield decreased slightly. After the 6th cycle, the yield declined to a greater extent due to poisoning of catalyst active sites by adsorption which also resulted in the decrease of surface area, pore volume, and basicity of the catalyst.\nAshok\u00a0et\u00a0al.\u00a0(2018) prepared a nanostructured MgO catalyst following the co-precipitation method and produced biodiesel from WCO. They found that the maximum biodiesel yield of 93.3% was achieved using 2 wt% of nanocatalyst, MTOMR of 24:1, reaction temperature of 65\u00a0\u00b0C, and reaction time 1\u00a0h. They also found that the catalyst was reusable at least 5 times with a decrease in activity which may be due to the change in the structural form of the MgO catalyst transforming to Mg(OH)2 (Boro\u00a0et\u00a0al., 2011). Feyzi\u00a0et\u00a0al.\u00a0(2017) prepared MgO-La2O3 nanocatalyst and used it in the production of sunflower oil biodiesel. The maximum biodiesel yield of 97.7% was found using 3\u00a0wt.% of catalyst loading, 18:1 MTOMR at 65 \u00b0C within 5\u00a0h of reaction time. They also reported that the catalyst could be reused up to the 4th cycle without significant loss in catalytic activity. Rasouli\u00a0and Esmaeili\u00a0(2019) demonstrated the production of biodiesel from goat fat by the use of MgO nanocatalyst. They found that the specific surface area, total pore volume, average diameter, and volume of pores were 40.44 m2 g\u22121, 9.29 cm3 g\u22121, 36.69\u00a0nm, and 0.371 cm3 g\u22121, respectively. They reported that the catalyst was mesoporous because the average pore diameter was less than 50\u00a0nm. They found that the average particle size of the catalyst was 5.5\u00a0nm. Under the ORCs of 1 wt% of catalyst amount, 12:1 of MTOMR, 3\u00a0h of reaction time, and reaction temperature of 70 \u00b0C, the yield of biodiesel was 93.12%.\nEsmaeili\u00a0et\u00a0al.\u00a0(2019) studied biodiesel production from Moringa oleifera oil by the use of MgO nanocatalyst. The different physicochemical properties of the catalyst were determined by using SEM, TEM, EDX, and BET techniques. The specific surface area and volume of pores of the catalyst were 14.19 m2 g\u22121 and 0.045 cm3 g\u22121, respectively. They reported that the maximum biodiesel yield of 93.69% was found at 1\u00a0wt.% of catalyst loading and 12:1 MTOMR at 45 \u00b0C within 4\u00a0h of reaction time. Rafati\u00a0et\u00a0al.\u00a0(2019) demonstrated the synthesis of MgONaOH nanocatalyst for the production of biodiesel from WCO by electrolysis method. They prepared the catalyst by adding NaOH to the solution of magnesium nitrate hexahydrate and ammonia solution with constant stirring. The material was dried and calcined at 400 \u00b0C. They reported that the catalyst was spherical in shape and the average size of the particle ranged from 40 to 80\u00a0nm, but the actual particle size was determined by XRD analysis and was found to be 66.77\u00a0nm. They found that the ORCs for transesterification reaction were catalyst amount of 3\u00a0wt.%, MTOMR of 6:1, a reaction time of 6\u00a0h, reaction temperature of 50 \u00b0C, and the yield of biodiesel reached above 98%.Nanocatalyst plays a fateful role and resolves the problems associated with the transesterification reaction which leads to the reduction in the biodiesel production cost. Nanocatalysts have high selectivity and catalytic activity due to their non-dimensional pores present on the surface of the catalyst (Baskar\u00a0and Aiswarya,\u00a02016). The ZnO nanocatalysts are well known for their non-toxic and biodegradable property. Due to its hexagonal wurtzite structure, ZnO has high transparency and oxygen vacancy and possesses a higher affinity for the polar substrate (Dantas\u00a0et\u00a0al., 2020). Nowadays, researchers are concentrating on the doping of transition metals like Mn, Co, Cu, Ni, and Fe which have a variety of applications in the field of semiconductor devices, drug carrier molecules, and biodiesel production. The performances of ZnO nanocatalysts derived from several sources in biodiesel synthesis from different feedstocks are summarized in Table\u00a03\n. Baskar\u00a0et\u00a0al.\u00a0(2018) studied biodiesel production from castor oil by the use of heterogeneous Ni-doped ZnO nanocatalyst. They prepared the catalyst with the help of the co-precipitation method in which Ni acetate solution was mixed with Zn acetate solution and stirred constantly and then ammonia solution was added to the mixture. Then NaOH was added to the mixture dropwise and then the mixture was filtered and dried at 80 \u00b0C for 3\u00a0h. and then calcined for 3\u00a0h. Ni-doped ZnO nano-composite showed an average particle size of 35.1\u00a0nm. They found that a maximum biodiesel yield of 95.20% was obtained under the reaction conditions of 55\u00a0\u00b0C of reaction temperature, 60\u00a0min of reaction time, 8:1 MTOMR, and 11 wt% of catalyst amount. They also found that the catalytic activity was maintained up to three cycles and the yield of biodiesel decreased slowly from 95.2% to 91.5% in the fourth cycle, and in the fifth cycle, it was found to be 85%. The decrease in biodiesel production was mainly due to the accumulation of organic matter on the surface of the catalyst.\nRaj\u00a0et\u00a0al.\u00a0(2019) investigated biodiesel production from microalgae (Nannochloropsis oculata) oil using heterogeneous polyethylene glycol (PEG) encapsulated ZnOMn2+ nanocatalyst. They reported that the particle size of the catalyst ranged from 20 to 42\u00a0nm. They also reported that the maximum biodiesel yield of 87.5% was achieved using 3.5 wt% of catalyst loading, and 15:1 MTOMR at 60\u00a0\u00b0C within 4\u00a0h of reaction time. They found that the catalytic activity was maintained up to four cycles and reported that the yield of biodiesel decreased gradually to 85.8% in the fifth cycle and in the sixth cycle, it decreased to 73.5%. This decrease was mainly due to the loss of PEG capping on ZnO on the surface of the catalyst. Baskar\u00a0and Aiswarya\u00a0(2015) demonstrated the synthesis of Cu doped ZnO nanocomposite and used it as a heterogeneous catalyst for biodiesel production from WCO. They found that the average size of the nanocatalyst was 80\u00a0nm. After considering various optimization parameters, they reported 97.71% of the yield of biodiesel at catalyst amount of 12 wt%, MTOMR of 8:1, and reaction time of 50\u00a0min at 55 \u00b0C. They also found that the catalytic activity was maintained up to five cycles and after the 5th cycle, the biodiesel yield decreased by 10%.\nBaskar\u00a0et\u00a0al.\u00a0(2016) reported the production of biodiesel production from mahua oil using Mn-doped ZnO nanocatalyst containing an average particle size of 24.18\u00a0nm. The catalyst was prepared by co-precipitation method followed by calcination at 600 \u00b0C for 2\u00a0h. They found a 97% yield of biodiesel under the reaction conditions of 50\u00a0\u00b0C of reaction temperature, 50 of min reaction time, 7:1 MTOMR, and 8 wt% of catalyst loading. They concluded that the catalyst could be reused up to the 5th cycle and after that, the biodiesel yield decreased sharply due to the deactivation of the active sites of the catalyst. Baskar\u00a0and Soumiya\u00a0(2016) studied the production of biodiesel from castor oil using Fe (II) doped ZnO nanocatalyst. They obtained a maximum yield of 91% biodiesel in 50\u00a0min at 55 \u00b0C with 14 wt% catalyst loading and 12:1 MTOMR ratio. They also reported that the catalyst was reusable up to the 4th reaction cycle and the major drop of the biodiesel yield was observed from the 4th cycle (87%) which was due to the deactivation of active sites. Thangaraj\u00a0and Piraman\u00a0(2016) demonstrated the production of biodiesel from Madhuca indica oil with the use of a heteropoly acid-coated ZnO nanocatalyst. The particle size of the catalyst was within the range of 5\u201329\u00a0nm. They reported that the ORCs for the maximum biodiesel yield of 95% was found at 0.6\u00a0wt.% of catalyst loading, 6:1 MTOMR at 55 \u00b0C within 5\u00a0h of reaction time. Gurunathan\u00a0and Ravi\u00a0(2015) investigated Cu doped ZnO as a catalyst in the production of neem oil biodiesel. Under the ORCs of catalyst amount of 10 wt%, MTOMR of 10:1, and a reaction time of 60\u00a0min at 55 \u00b0C, the yield of biodiesel could reach above 97.18%. They also reported that the catalyst was reusable up to six consecutive cycles beyond which yield decreased sharply due to the deposition of organic materials on the surface of the catalyst.\nNagaraju\u00a0et\u00a0al.\u00a0(2017) utilized Ag-doped ZnO material as a nanocatalyst in biodiesel synthesis from simarouba oil. This catalyst could produce a maximum yield of 84.5% biodiesel under the ORCs of 64 \u00b0C of temperature, 9:1 of MTOMR, and 2\u00a0h of reaction time with 1.5 wt% catalyst loading. A magnetic ZnO/BiFeO3 nanocatalyst was used in the synthesis of canola oil biodiesel by Salimi\u00a0and Hosseini\u00a0(2019). They reported that the average crystallite size and the particle size of the catalyst were 31.27\u00a0nm and 20\u201360\u00a0nm, respectively. The investigation yielded 95.43% of biodiesel at the best ORCs of 65 \u00b0C, 6\u00a0h of reaction time, 4 wt% catalyst loading, and 15:1 MTOMR. They also reported that the catalyst after the 5th cycle could yield 92.08% of biodiesel and this decrease might be due to a decrease in the number of basic sites on the catalyst. Borah et\u00a0al., 2019b studied the synthesis of Co-doped ZnO and used it as a nanocatalyst for the reaction of Mesua ferrea oil. They reported that a maximum yield of 98.03% biodiesel was obtained under the ORCs of 2.5 wt% of catalyst loading, 3\u00a0h of reaction time, 9:1 MTOMR at 60 \u00b0C. They also reported that at the end of the 4th cycle, the biodiesel yield decreased to 43.13%.It has earlier been mentioned that a heterogeneous catalyst is preferred over homogeneous catalyst for biodiesel synthesis. Heterogeneous solid base catalysts are widely applicable for biodiesel production due to their high catalytic activity, low-temperature requirement, easy separation, and reusability, which could potentially reduce the biodiesel production cost. However, it has some disadvantages too such as catalyst leaching (Wilson\u00a0and Lee,\u00a02012), unsuitability for feedstocks containing high FFAs (Borges\u00a0and D\u00edaz,\u00a02012), and dissolution of catalyst in the reaction medium (Mbaraka\u00a0et\u00a0al., 2006), and some catalysts have low activity and low porosity with low surface area (Taufiq-Yapa\u00a0et\u00a0al., 2011). Furthermore, solid base catalysts are widely reported for the conversion of edible oils to biodiesel, which tends to create a competition between food and fuel (Lin\u00a0et\u00a0al., 2011; Qiu\u00a0et\u00a0al., 2011). Due to strong acid sites, high activity and sensitivity, and low-cost properties, solid acid catalysts were also studied for biodiesel production (Sharma\u00a0and Singh,\u00a02011). Many researchers reported the studies of biodiesel production using solid acid catalysts such as heteropolyacid impregnated on different supports (silica, alumina, zirconia, and activated carbon), WO3\u2013ZrO2, SO4\u2013ZrO2 (Kulkarni\u00a0et\u00a0al., 2006; Laosiripojana\u00a0et\u00a0al., 2010), etc. The synthesis of biodiesel from different feedstocks using ZrO2 nanocatalysts is mentioned in Table\u00a04\n.\nTakase\u00a0et\u00a0al.\u00a0(2014) studied biodiesel synthesis from Silybum marianum oil using ZrO2 modified with KOH as a nanocatalyst. The surface area and pore volume of pure ZrO2 were found as 7.02 m2 g\u22121 and 0.01 cm3 g\u22121, whereas the nanocatalyst showed 3.05 m2 g\u22121 and 0.01 cm3 g\u22121, respectively. A biodiesel yield of 90.8% was obtained at ORCs of 6 wt% of catalyst amount, 15:1 of MTOMR, and 2\u00a0h reaction time at 60 \u00b0C. They reported that the catalyst could be reused up to five times after washing with methanol and re-calcination at 530 \u00b0C and after the 5th cycle, the biodiesel yield decreased to 82.4%. Qiu\u00a0et\u00a0al.\u00a0(2011) investigated the reaction of soybean oil using ZrO2 coupled C4H4O6HK heterogeneous solid base nanocatalyst. The catalyst preparation was done by the incipient wetness impregnation method. They reported that when the reaction was carried out with MTOMR of 16:1, a reaction temperature of 60 \u00b0C, a reaction time of 2\u00a0h, and a catalyst amount of 6 wt%, the highest biodiesel yield reached 98.03%. They also reported the catalyst was reused up to the 5th cycle of reaction and in the 5th cycle, the biodiesel yield decreased from 98.03% to 89.65% which was basically due to the leaching of metal.\nMahdavi\u00a0et\u00a0al.\u00a0(2015) utilized oleic acid as feedstock for biodiesel production using ZrO2/Al2O3 as the catalyst. They found the particle size and surface area of the catalyst in the range of 20.59\u201329.86\u00a0nm and 253\u2013283 m2 g\u22121, respectively. The nanocatalyst at the ORCs of 1 wt% of catalyst amount, 8:1 of MTOMR, 67\u00a0\u00b0C of reaction temperature, and 2\u00a0h reaction time could provide 90.47% of biodiesel. They also investigated the reusability of the catalyst and found it reusable up to the 4th cycle. Booramurthy\u00a0et\u00a0al.\u00a0(2021) studied the transesterification of animal fat using ferric-manganese doped sulfated zirconia (Fe-Mn-SO4/ZrO2) as the catalyst. It was reported that using an optimized catalyst amount of 6 wt% and alcohol to oil molar ratio of 12:1, biodiesel yield was found to be 96.6% at the reaction temperature of 65\u00a0\u00b0C in 5\u00a0h. They recycled and reused the catalyst several times and easily activated the catalyst by washing it with hexane and methanol followed by being dried at 120 \u00b0C for 8\u00a0h. They reported that the catalyst could be reused up to the 5th cycle and after that, the biodiesel yield decreased because of the loss of acid sites from the catalyst surface due to the weak bonding strength and aggregation of the catalyst particles. Saravanan\u00a0et\u00a0al.\u00a0(2016) demonstrated the application of sulfated zirconia as the nanocatalyst in the reaction of palmitic acid. They reported that the ORCs for the production of 90% of biodiesel from palmitic acid were found as 6\u00a0wt.% of catalyst loading, 20:1 of MTOMR, 60 \u00b0C of reaction temperature, and 5\u20137\u00a0h of reaction time. They also reported that the catalyst was reusable up to five times and at the end of the 5th cycle, the yield decreased to 59%. Faria\u00a0et\u00a0al.\u00a0(2009) studied the synthesis of SiO2/ZrO2 nanocatalyst and utilized it in biodiesel synthesis from soybean oil. They found that the surface area of the catalyst was 135 m2 g\u22121 and particle diameter was 200\u00a0nm. They achieved 96.2% of biodiesel under the ORCs of 0.5\u00a0g of catalyst amount, and 10:1.5 of MTOMR at 50 \u00b0C in 3\u00a0h. The catalyst was reused up to six times.\nHelmiyati\u00a0et\u00a0al.\u00a0(2021) reported the synthesis of biodiesel from lauric acid using cellulose@hematite-zirconia as the catalyst. For the preparation of nanocatalyst, they isolated the cellulose from rice straw and converted it into nanocellulose with the help of H2SO4 and then filtered out the precipitate followed by washing it with water and dried. The nano-\u03b1-Fe2O3 was prepared by mixing the solutions of FeCl2\u22c54H2O and FeCl3\u22c56H2O and then NH4OH was added to the above mixture with constant stirring. The precipitate was filtered and washed with water and ethanol, and calcined at 600 \u00b0C for 1\u00a0h. For the preparation of nano-ZrO2, NaOH solution was added to ZrOCl2\u22c58H2O solution with constant stirring. Then the white precipitate was filtered out and wash with water and acetone, and calcined at 700 \u00b0C for 1\u00a0h. \u03b1-Fe2O3-ZrO2 composites were prepared by adding \u03b1-Fe2O3 into a solution containing H2O, ethanol, and ammonia. Then ZrOCl2\u22c58H2O was added to the above mixture with constant stirring. The precipitate was separated followed by washing with ethanol and water and dried at 60 \u00b0C for 12\u00a0h. Then nanocellulose was mixed with an aqueous solution of NaOH and urea and then mixed with \u03b1-Fe2O3-ZrO2 in aqueous NaOH with constant stirring. The product was then separated followed by washing with ethanol and water and dried. The nanocatalyst contained a surface area of 852 m2 g\u22121, pore volume of 0.85 cm3 g\u22121, the pore size of 13\u00a0nm, and average particle size of 42.5\u00a0nm. At the ORCs of 2 wt% catalyst amount, and 12:1 MTOMR at 60 \u00b0C in 3\u00a0h, the biodiesel yield of 92.50% was obtained. They found that the catalyst was reusable and at the end of the fifth cycle, biodiesel yield decreased to 80% which was mainly due to the deactivation of catalyst by absorbing unreacted lauric acid and by-product species.Metal oxide-based solid catalysts are being conventionally used in biodiesel production. Due to the large surface area, acid-base properties, strong metal-support interactions, and chemical stability, titanium dioxide (TiO2) nanoparticles are widely used in the transesterification reaction for biodiesel production (Carlucci\u00a0et\u00a0al., 2019; Li\u00a0and Wang,\u00a02012). Nowadays, binary metal oxides, for example, TiO2\u2013ZnO nano mixed metal oxide, etc. are receiving interest due to their high surface acidity for biodiesel production (Li and Wang, 2012; Gurusamya\u00a0et\u00a0al., 2019). The performances of TiO2 nanocatalysts in biodiesel synthesis from different feedstocks are summarized in Table\u00a05\n. Gurusamya et\u00a0al. (2019) reported the synthesis of biodiesel from Ulva lactuca seaweed using TiO2-ZnO nanocomposite catalyst. They found 82.8% yield of biodiesel at the ORCs of MTOMR of 6:1, catalyst dosage of 4 wt%, reaction temperature of 60\u00a0\u00b0C, and reaction time of 4\u00a0h. They also reported that the catalyst was reusable up to the 5th cycle. Madhuvilakku\u00a0and Piraman\u00a0(2013) prepared TiO2\u2013ZnO mixed oxide nanocatalyst for synthesis of palm oil biodiesel. They also compared the TiO2\u2013ZnO nanocatalyst with the ZnO nanocatalyst. They reported that TiO2\u2013ZnO mixed oxide catalyst showed 92.2% yield in 5\u00a0h, whereas ZnO nanocatalyst showed only 83.2% of biodiesel yield in 5\u00a0h under the ORCs of 6:1 MTOMR at 60 \u00b0C with 200\u00a0mg of nanocatalyst.\nZulfiqar\u00a0et\u00a0al.\u00a0(2021) prepared lipase-PDA-TiO2 (PDA\u2013polydopamine) nanoparticles using the hydrothermal method and self-polymerization-based grafting technique and utilized as the nanocatalyst for the synthesis of jatropha oil biodiesel. The ORCs for the transesterification to obtain the maximum yield of 92% biodiesel were 10\u00a0wt.% of catalyst loading, 6:1 MTOMR at 37 \u00b0C in 30\u00a0h of reaction time. They also reported that the catalyst could be reused up to the 4th reaction cycle with the decrease in the catalytic activity. Chen\u00a0et\u00a0al.\u00a0(2018) demonstrated the synthesis of biodiesel from Jatropha curcas oil with the help of a nano-sized SO4\n2\u2212/TiO2 catalyst. 85.3% yield of biodiesel was reported at the ORCs of 4 wt% of catalyst amount, 9:1 of MTOMR, 24\u00a0h of reaction time, and 140 \u00b0C of reaction temperature. They also reported that the catalyst was reusable up to the 3rd cycle and at the 3rd cycle, the biodiesel yield decreased to 25.3%. This decrease was due to the aggregation of cokes on the surface of the catalyst which leads to the decrease in the catalytic activity. A nanocatalyst, Ti(SO4)O, was utilized by Gardy\u00a0et\u00a0al.\u00a0(2016) for the preparation of biodiesel from WCO. They reported that the surface area, mean pore size, and total pore volume of the nanocatalyst were 44.4563 m2 g\u22121, 22.7347\u00a0nm, and 0.312459 cm3 g\u22121, respectively, and the average particle size of the catalyst was 45\u00a0nm. They reported that the ORCs for production of 97.1% of biodiesel were 1.5 wt% of catalyst amount, 9:1 of MTOMR, 3\u00a0h of reaction time, and 75 \u00b0C of reaction temperature. The nanocatalyst was reused up to the 8th cycle and the biodiesel yield decreased to 85.91%. Biodiesel was synthesized from WCO by Gardy\u00a0et\u00a0al.\u00a0(2017) using TiO2/PrSO3H as the catalyst. TEM analysis showed that the average particle size of the catalyst was 23.1\u00a0nm. The BET analysis revealed a surface area of 38.59 m2 g\u22121, pore volume of 0.192 cm3 g\u22121, and mean pore size of 24.55\u00a0nm. The ORCs of MTOMR of 15:1, 4.5% of catalyst amount, 60 \u00b0C of reaction temperature, and reaction time of 9\u00a0h resulted in 98.3% of biodiesel yield. The reusability of nanocatalyst was investigated and at the 4th cycle, the biodiesel yield was 94.16% and after the 6th cycle, it was 20.64%. The decrease in yield was due to the blockage of the active site of the catalyst by the organic or carbonaceous material. Mihankhah\u00a0et\u00a0al.\u00a0(2018) also studied biodiesel synthesis from waste olive oil with the help of a TiO2 nanocatalyst. They reported that the surface area and average particle size of the catalyst were 238 m2 g\u22121 and \u223c30\u00a0nm, respectively. They also reported that conversion of 91.2% was obtained at an ORCs of 30:1 of MTOMR, 200\u00a0mg of catalyst amount, 120 \u00b0C of reaction temperature, and 4\u00a0h of reaction time. The catalyst at the 3rd cycle of reaction yielded 88% of biodiesel.The utilizations of different CuO nanocatalysts in the synthesis of biodiesel and their results reviewed are shown in Table\u00a06\n. Santha\u00a0et\u00a0al.\u00a0(2021) investigated biodiesel synthesis from WCO using CuO nanoparticles as the heterogeneous catalyst. They found that the CuO nanocatalyst at the operating conditions of 2 wt% catalyst amount, 4:1 MTOMR at 60\u00a0\u00b0C reaction temperature, and 2.5\u00a0h reaction time provided a maximum yield of 88.64% biodiesel. Varghese\u00a0and Prabu\u00a0(2017) reported the synthesis of a needle-shaped CuO nanomaterial for biodiesel production. They reported that the biodiesel yield of 86.56% was found at relatively low catalyst loading (0.75 wt%), 3.5:1 MTOMR within 2\u00a0h of reaction time. Varghese\u00a0et\u00a0al.\u00a0(2017) demonstrated the synthesis of Mg-CuO heterogeneous nanocatalyst for the synthesis of sunflower oil biodiesel. They reported that a biodiesel yield of only 71.78% was obtained at the reaction conditions of 0.25\u00a0wt% catalyst amount, 6:1 MTOMR, and 30\u00a0min of reaction time at 60 \u00b0C. Suresh\u00a0et\u00a0al.\u00a0(2021) studied the preparation of biodiesel from pig tallow using CuO nanocatalyst. The CuO nanocatalyst was synthesized using C. tamala leaves. They reported that the average particle size of the nanocatalyst was 19.01\u00a0nm. The CuO catalyzed biodiesel preparation yielded 97.82% of the product under the ORCs of 60 \u00b0C of temperature, 29.87:1 MTOMR, and 2.07 wt% of catalyst loading.Commercialization and application of biodiesel to replace fossil fuels are hindered by the outrageous cost of production which is mainly due to the cost of raw materials. The cost can be reduced up to 77% of the total cost by the use of non-edible vegetable oil feedstocks (Skarlis\u00a0et\u00a0al., 2012). In biodiesel production, the use of acid or alkaline homogeneous catalysts is linked to some kind of problem which hinders commercial production (Seffati\u00a0et\u00a0al., 2019). Nowadays, nanocatalysts are receiving more attention for the synthesis of biodiesel due to their high recovery factor, large surface area, high energy consumption recovery, and requirement of low reaction temperature (Shahid\u00a0and Jamal,\u00a02011). Magnetic nanoparticles are the most popular materials due to their high surface to volume ratio, lower mass transfer resistance for reacting with substrates, and easy way of separation from the reaction mixture by an external magnetic field, and hence reducing the loss of catalyst and increasing the reusability (Verma\u00a0et\u00a0al., 2015; Rajkumari\u00a0et\u00a0al., 2017). This makes the catalyst more profitable for industrial applications. The catalytic performances of different magnetic nanocatalysts in biodiesel synthesis are represented in Table\u00a07\n.\nChangmai\u00a0et\u00a0al.\u00a0(2021) utilized Citrus sinensis peel ash (CSPA) coated magnetic material (CSPA@Fe3O4) as the nanocatalyst for biodiesel synthesis from WCO. They prepared the catalyst by burning the dried orange (C. sinensis) peel in the air for 30\u00a0min to form ash and mixed with water and stirred for 2\u00a0h at 80 \u00b0C to extract the basic components present in CSPA. Fe3O4 nanoparticles were synthesized via the traditional co-precipitation method. Then CSPA extract was added dropwise to the mixture, stirred, and allowed to settle followed by decantation of the solution and collected the solid part, which was then washed with deionized water. At last, the catalyst was collected by evaporation of the solid portion. The chemical composition of the catalyst was investigated by X-ray photoelectron spectroscopic technique (Fig.\u00a01\n), and the major basic elements obtained were K (8.64%) and Ca (4.46%). The prepared catalyst was also characterized using SEM, XRD, FT-IR, BET and TEM (Fig.\u00a02\n). They found that the average particle size of the nanoparticle was \u0334 12\u201313\u00a0nm, the surface area was 15.55 m2\ng\u00a0\u2212\u00a01, and the pore diameter was found to be 2.45\u00a0nm. The CSPA@Fe3O4 material catalyzed transesterification produced a maximum yield of 98% biodiesel under the ORCs of 65 \u00b0C, 6:1 and 6\u00a0wt.% of temperature, MTOMR, and catalyst loading, respectively. They also reported that the catalyst was reusable for up to 9 consecutive cycles. Dantas\u00a0et\u00a0al.\u00a0(2020) demonstrated the synthesis of Ni0.5Zn0.5Fe2O4 magnetic material and utilized it as nanocatalyst for the production of biodiesel. They found that the average particle size of the Ni0.5Zn0.5Fe2O4 catalyst was 31.1\u201342.6\u00a0nm. BET surface area of magnetic nano-catalyst was found to be 50.94 m2\n\ng\n\n\u22121, and the average pore diameter and pore volume were 48.042\u00a0\u00c5 and 0.171 cm3 g\u22121, respectively. In their experiment, they found that the ORCs for yielding 99.54% of soybean oil biodiesel was 12:1 MTOMR, 2 wt% of catalyst loading, 180 \u00b0C of reaction temperature, and 1\u00a0h of reaction time. They recovered the catalyst with the help of an external magnet and reused it for up to 3 cycles.\nAli\u00a0et\u00a0al.\u00a0(2017) reported the synthesis of biodiesel from date palm oil using magnetic nanocatalyst, CaO-Fe3O4. The catalyst was prepared by the chemical precipitation method. They found that a biodiesel yield of 69.7% under the conditions of 65\u00a0\u00b0C reaction temperature, 300\u00a0min reaction time, 20:1 MTOMR, and 10\u00a0wt.% of catalyst amount. Feyzi\u00a0and Norouzi\u00a0(2016) prepared magnetic nanocatalyst, Ca/Fe3O4@SiO2 following sol-gel and impregnation methods, and utilized it in biodiesel synthesis. The surface area of the nanocatalyst was found to be 189.2 m2 g\u22121, and the average pore diameter and pore volume were 2.4\u00a0\u00c5 and 0.238 cm3 g\u22121, respectively. The nanocatalyst at the operating conditions of 8 wt% of catalyst amount, 15:1 MTOMR, 65\u00a0\u00b0C of reaction temperature, and 5\u00a0h of reaction time yielded 97% biodiesel. The catalyst could be recovered simply by using an external magnetic field and reused several times without appreciable loss of its catalytic activity. Ambat\u00a0et\u00a0al.\u00a0(2019) prepared nano-magnetic K impregnated ceria and demonstrated it in biodiesel synthesis. The BET surface area, pore volume, and pore size were found to be 72.84 m2 g\u22121, 0.18 cm3 g\u22121, and 9.99\u00a0nm, respectively. A biodiesel yield of 96.13% could be achieved under the reaction conditions of 4.5 wt% catalyst amount, 7:1 MTOMR, and 120\u00a0min of reaction time at 65 \u00b0C. They observed that the catalyst was stable up to five cycles without considerable loss of activity.\nLiu\u00a0et\u00a0al.\u00a0(2017) reported biodiesel synthesis from soybean oil using a nano-magnetic solid catalyst (K/ZrO2/\u03b3-Fe2O3). They found that the yield of biodiesel was above 93.6% at ORCs of 5\u00a0wt% of catalyst amount, and MTOMR of 10:1 at 65 \u00b0C in 3\u00a0h. They also found that the catalytic activity was maintained up to six cycles and after that, the yield of biodiesel decreased due to the loss of nano-powder and alkaline sites in the recycling process of the catalyst. Hazmi\u00a0et\u00a0al.\u00a0(2021) reported the preparation of bifunctional magnetic nano-catalyst (RHC/K2O-20%/Ni-5%) from rice husk char (RHC) for the production of biodiesel. It was revealed that the surface area, pore diameter, and pore volume of the catalyst were 32.40 m2 g\u22121, 5.8355\u00a0nm, and 0.0966 cm3 g\u22121, respectively. They reported 98.2% of biodiesel at 4\u00a0wt.% of catalyst loading and 12:1 MTOMR at 65 \u00b0C within 2\u00a0h of reaction time. They also reported that the catalyst could be reused up to the 5th reaction cycle and it was noticed that the major drop of the biodiesel yield was observed from the 4th (81.8%), 5th (71.0%), and 6th (45.9%) cycles. Hazmi\u00a0et\u00a0al.\u00a0(2020) also studied the synthesis of nano-bifunctional super magnetic material (RHC/K2O/Fe) from rice husk and application as heterogeneous nanocatalyst for biodiesel synthesis from WCO. The surface area, average pore diameter, and total pore volume of the catalyst were 57.89 m2 g\u22121, 4.70\u00a0nm, and 0.0588 cm3 g\u22121, respectively. They achieved 98.6% of biodiesel under the ORCs of 4 wt% of catalyst amount, and 12:1 of MTOMR in 4\u00a0h at 75 \u00b0C. They also reported that the catalytic performance of the nano-catalyst was maintained for five consecutive cycles.On investigating a promising alternative to the homogeneous catalyst, some chemically modified natural materials and rocks were assessed for the synthesis of low-cost, abundantly available, and ecofriendly heterogeneous catalysts for the transesterification reaction (Rabie\u00a0et\u00a0al., 2019; Abukhadra\u00a0and Mostafa,\u00a02019). Zeolite has wide applications in biodiesel synthesis due to its microporous structure, high surface area, high stability, high mechanical strength, and high cation exchange capacity (Liu\u00a0et\u00a0al., 2018; You\u00a0et\u00a0al., 2017). By controlled functionalization of the surfaces of natural and synthetic zeolites by an acidic or basic group, biodiesel yield can be enhanced (Manique\u00a0et\u00a0al., 2017; Du\u00a0et\u00a0al., 2018). Several studies demonstrated that the activation of natural zeolite by alkali metal ions can enhance the catalytic property due to an increase in their basicity, which is a vital factor for the transesterification process (Ballotin\u00a0et\u00a0al., 2016; Abukhadra\u00a0and Sayed,\u00a02018). Faujasite zeolite (NaX) was reported to be beneficial for biodiesel synthesis due to its high surface area and a huge amount of basic sites, which is attributed to its aluminum content (Davis,\u00a02003). The catalytic performance of nano-zeolite catalysts in biodiesel synthesis is shown in Table\u00a08\n.\nDehghani\u00a0and Haghighi\u00a0(2019) reported biodiesel production from WCO by the use of cerium-doped MCM-41 as a catalyst. They prepared the catalyst by hydrothermal method in which cetyl trimethyl ammonium bromide, tetraethyl orthosilicate, and cerium nitrate were mixed in distilled water. To this mixture, NaOH solution was added with continuous stirring. The suspension was put into an autoclave, dried and the sample was filtered, washed, and dried at 110 \u00b0C. The sample was irradiated with the solution of magnesium nitrate and support solution and sonicated. The mixture was filtered and dried, and finally, the power was calcined at 600 \u00b0C for 3\u00a0h. The particle size of the catalyst was about 17.3\u00a0nm. They observed that the surface area of the catalyst was 1200 m2 g\u22121. They reported 94.3% of biodiesel yield using 5 wt% of catalyst loading, 9:1 MTOMR at 70 \u00b0C within 6\u00a0h of reaction time. The catalyst could be reused up to 7 times and after the end of the 7th cycle, the biodiesel yield was found to be 88.7%. Alkali trapped zeolite composite was prepared by AbuKhadra\u00a0et\u00a0al.\u00a0(2020) and utilized as the basic catalyst for biodiesel preparation from WCO. They prepared four types of alkali-modified clinoptilolite (K, Na, Ca, and Mg) which were extracted from green tea. The SEM images of the studied samples (Fig.\u00a03\n) showed different morphological characteristics, and varied elemental compositions were revealed from the EDX investigations (Fig.\u00a04\n). They found that the average pore size of the clinoptilolite, K/clinoptilolite, Na/clinoptilolite, Ca/clinoptilolite and Mg/clinoptilolite nanoparticles were 18.3\u00a0nm, 17.6\u00a0nm, 17.3\u00a0nm, 15.4\u00a0nm, 19.6\u00a0nm, and the surface areas were 258 m2 g\u22121, 263 m2 g\u22121, 312.7 m2 g\u22121, 252.4 m2 g\u22121, and 342.5 m2 g\u22121, respectively. The biodiesel yields achieved with the modified catalyst were 93.6%, 95.2%, 96.4%, and 98.7% for K/clinoptilolite, Na/clinoptilolite, Ca/clinoptilolite, and Mg/clinoptilolite, respectively under the ORCs of 70 \u00b0C of temperature, 16:1 MTOMR, and 4\u00a0wt.% of catalyst loading. The reaction time taken was 120\u00a0min, 120\u00a0min, 180\u00a0min, and 150\u00a0min for K/clinoptilolite, Na/clinoptilolite, Ca/clinoptilolite, and Mg/clinoptilolite catalysts, respectively. They also reported that the reusability of catalyst for up to five times and after the 5th cycle of reaction, the catalytic activity decreased due to the coating of the byproducts on the active's sites of the catalysts.\nLuz\u00a0Martinez et\u00a0al.\u00a0(2011) demonstrated the preparation of CaO nanoparticles/NaX zeolite for the transesterification of sunflower oil. They reported a 93.5% yield of biodiesel using this catalyst under the ORCs of 16 wt% of catalyst amount and 6:1 MTOMR at 60 \u00b0C in 6\u00a0h. Saeedi\u00a0et\u00a0al.\u00a0(2016) prepared KNa/ZIF-8 (Zeolite imidazolate framework, ZIF-8 doped with K) material following sol-gel method and investigated as a catalyst for biodiesel production from soybean oil. They reported that the surface area, pore volume and pore diameter of the catalyst was 1195 m2 g\u22121, 0.527 cm3 g\u22121, and 1.21\u00a0nm, respectively. A biodiesel yield of 98% was obtained under the ORCs of 0.0125 wt% of catalyst loading and 10:1 MTOMR at 100 \u00b0C within 3.5\u00a0h of reaction time. Dehghani\u00a0and Haghighi\u00a0(2020) studied the preparation of sono-enhanced CaO-dispersed over Zr-doped MCM-41 nanocatalyst for the synthesis of WCO biodiesel. From the BET analysis, it was revealed that the surface area and the pore size of the nanocatalyst were 350 m2 g\u22121 and 5\u00a0nm, respectively. From FESEM analyses (Fig.\u00a05\n) and size distribution histogram (Fig.\u00a06\n), they found that the average particle size of the nanocatalyst was 15.9\u00a0nm for Ca/ZM-U (Si/Zr\u00a0=\u00a010) sample, and the size was found in the range of 20\u201380\u00a0nm in the case of non-sonicated sample Ca/ZM-I (Si/Zr\u00a0=\u00a010). The EDX analyses of the studied nanocatalysts are displayed in Fig.\u00a07\n. They reported a yield of 88.5% biodiesel under ORCs of 5 wt% of catalyst amount, 9:1 of MTOMR, and reaction time of 6\u00a0h at 70 \u00b0C. They also reported that the catalyst was reused up to the fifth reaction cycle and after that, the catalytic activity decreased which was mainly due to the blockage of the active site on the catalyst.Hydrotalcites are anionic clays that can be prepared by coprecipitation method and they have a common notation of [M2+\n1-XM3+(OH)2]\nX\n\n+[A\nn\n\n\u2212]X/N.yH2O, where M2+ and M3+ are representing divalent and trivalent metals, and An\u2212(CO3\u2212, SO4\n2\u2212, Cl\u2212, NO3\u2212) is an n- valent anion (Helwani\u00a0et\u00a0al., 2009; Endalew\u00a0et\u00a0al., 2011). These are considered heterogeneous base catalysts and their basic strength depend on the ratio of Mg/Al. Nano-hydrotalcites are used as catalysts in biodiesel synthesis and their catalytic performances are shown in Table\u00a09\n. Deng\u00a0et\u00a0al.\u00a0(2011) investigated the preparation of biodiesel from jatropha oil with the help of nanocatalyst derived from hydrotalcite with Mg/Al following the coprecipitation method. The physical-chemical properties such as Mg/Al molar ratio, surface area, pore volume, and pore diameter for the catalyst were found as 2.78, 218 m2 g\u22121, 0.17 cm3 g\u22121, and 3.9\u00a0nm, respectively. They reported that 95.2% yield of biodiesel was found at 1 wt% of catalyst loading and 4:1 MTOMR at 45 \u00b0C within 1.5\u00a0h of reaction time. They also reported that the catalyst was reusable up to 8 times after removing the glycerol. They found that the catalyst could yield 89.1% of biodiesel at the 8th cycle and the 9th cycle, the yield decreased to 43.7%, and this was due to the blocking of active sites of the catalyst by the glycerol.\nDias\u00a0et\u00a0al.\u00a0(2012) demonstrated soybean oil biodiesel synthesis using Ce modified Mg-Al hydrotalcite. They found that 90.2% biodiesel yield could be achieved under the ORCs of 5 wt% of catalyst loading and 9:1 MTOMR at 67 \u00b0C within 4\u00a0h of reaction time. Nano-hydrotalcite (Mg-Al) was prepared by Obadiah\u00a0et\u00a0al.\u00a0(2012) and applied as the catalyst for biodiesel synthesis from pongamia oil. Under the ORCs of 5 wt% of catalyst and 6:1 of MTOMR at the reaction temperature of 65 \u00b0C in 4\u00a0h, the yield of biodiesel could reach above 90.8%. Gao\u00a0et\u00a0al.\u00a0(2010) prepared biodiesel from palm oil using KF/Ca-Mg-Al hydrotalcite base catalyst. High biodiesel of 99.6% could be achieved at ORCs of 12:1 MTOMR, 5 wt% of catalyst, a reaction time of 5\u00a0min, and a reaction temperature of 65 \u00b0C. They reported a decrease in the biodiesel yield during reusability of catalyst and it was due to the absorption of by-products on the surface of the catalyst. Chelladurai\u00a0and Rajamanickam\u00a0(2014) demonstrated neem oil biodiesel synthesis using a nano-Zn-Mg-Al hydrotalcite catalyst. This catalyst yielded 92.5% of biodiesel at 7.5\u00a0g of catalyst loading, and 10:1 MTOMR at 65 \u00b0C within 4\u00a0h of reaction time.The catalytic performances of some other nano-catalysts used in the synthesis of biodiesel are presented in Table\u00a010\n. Abdullah\u00a0et\u00a0al.\u00a0(2022) demonstrated the synthesis of biodiesel from WCO using activated carbon as a catalyst prepared from empty fruit bunch. They prepared the nano-catalyst following the hydrothermal technique wherein carbonization was performed at 600 \u00b0C for 3\u00a0h. They reported that the BET surface area, pore volume, and pore diameter were 4056.17 m2 g\u22121, 0.827 cm3 g\u22121, and 5.42\u00a0nm, respectively. The particle size of the catalyst was 58\u00a0nm. They found a high yield of 97.1% biodiesel at ORCs of 12:1 MTOMR, 5\u00a0wt.% catalyst amount, a reaction time of 2\u00a0h, and a reaction temperature of 70 \u00b0C. They also concluded that the nano-catalyst was reusable up to 5th reaction cycles. In the 5th cycle, the yield was 85% and at the end of the 6th cycle, the biodiesel yield decreased to merely 61.7%, which was due to the decrease in the number of activated components during the calcination process.\nAbdullah\u00a0et\u00a0al.\u00a0(2020) demonstrated the synthesis of bifunctional nanocatalyst from waste palm kernel shell for the preparation of WCO biodiesel. The surface area, pore volume, and pore diameter of the catalyst were 438.08 m2 g\u22121, 0.3674\u00a0mm3 g\u22121, and 3.8\u00a0nm, respectively. Under ORCs of 5 wt% of catalyst amount, 12:1 of MTOMR, 4\u00a0h of reaction time, and reaction temperature of 80 \u00b0C, a 95% biodiesel yield could be found. Abdullah\u00a0et\u00a0al.\u00a0(2021) also studied the synthesis of bifunctional nanocatalyst from palm kernel shell by carbonization technique and applied the material in biodiesel production from WCO.\u00a0In this study, the surface area, pore volume, and pore diameter of the nanocatalyst were found to be 3368.60 m2 g\u22121, 2.36\u00a0mm3 g\u22121, and 5.17\u00a0nm, respectively. The FESEM images (Fig.\u00a08\n) revealed the successful impregnation of K2CO3 and CuO active components showing the formation of irregular shaped nanomaterials. The EDX mapping of the impregnated nanomaterial is shown in Fig.\u00a09\n. They reported that the yield of biodiesel could reach about 95.36% under ORCs 4 wt% of catalyst amount and MTOMR of 12:1 within 2\u00a0h at 70 \u00b0C. They also reported that the catalyst could be reused up to the 5th reaction cycle. At the end of the 6th cycle, yield decreased to 57.5% which was due to poisoning by the by-products such as unreacted oil and glycerol. Kuniyil\u00a0et\u00a0al.\u00a0(2021) reported the application of ZnCuO/N-doped graphene (NDG) as a catalyst for biodiesel synthesis from WCO. The structure of the prepared nanocomposite catalyst was studied by HRTEM technique (Fig.\u00a010\n), and this revealed the formation of nanoparticles (ZnCuO) deposited on N-doped graphene sheets. The size of the particles was found in the range of 12\u201318\u00a0nm. They found 97.1% of biodiesel at ORCs of 15:1 MTOMR, 10 wt% of catalyst amount, a reaction time of 8\u00a0h, and a reaction temperature of 180 \u00b0C. They also reported that the catalyst was successfully reused up to six cycles.\nIbrahim\u00a0et\u00a0al.\u00a0(2022) recently prepared magnetic bifunctional nanocatalysts from fruit bunch and employed in WCO biodiesel synthesis. The preparation of catalyst and biodiesel synthesis is shown in Fig.\u00a011\n. The surface morphological structures and metal oxide distribution on the surface of the catalyst are displayed in FESEM images (Fig.\u00a012\n). The catalyst, CaO (10%)-Fe2O3 (10%)/AC (AC-activated carbon), could yield 98.3% of biodiesel under the reaction conditions of 18:1 of MTOMR and 3 wt% of catalyst loading at 65 \u00b0C in 3\u00a0h of reaction. The reusability studies showed a good catalytic activity (biodiesel yield > 80%) even at the 6th consecutive cycle. However, after the 6th catalytic cycle of the reaction, the catalyst showed leaching of active sites and changes in the surface morphological characters as revealed by FESEM analysis (Fig.\u00a013\n).\nBet-Moushoul\u00a0et\u00a0al.\u00a0(2016) prepared Ag/bauxite nanocatalyst and utilized it in sunflower oil biodiesel synthesis. The SEM analyses of uncalcined bauxite, calcined bauxite (850 \u00b0C), bauxite/Ag nanocomposite, and recycled bauxite/Ag nanocomposite are displayed in Fig.\u00a014\n. This study revealed the formation and a nice dispersion of Ag nanoparticles on the bauxite surface. They reported a yield of 94% biodiesel under the conditions of 3\u00a0h of reaction time, 67 \u00b0C of reaction temperature, 9:1 of MTOMR, and 0.3 wt% catalyst loading. They also reported that the catalyst was successfully reused up to the 8th cycle and the biodiesel yield decreased to 71.79%, which might be due to the leaching of catalyst's active sites, Fig.\u00a014(D). Rashtizadeh\u00a0et\u00a0al.\u00a0(2014) demonstrated the synthesis of Sr3Al2O6 nanocatalyst and applied it in soybean oil biodiesel synthesis. The catalyst could produce 95.7% of biodiesel at the ORCs of 1.3 wt% of catalyst and 25:1of MTOMR at 60 \u00b0C in 1\u00a0h. They reported that the catalyst was successfully reused up to 4th time. Foroutan\u00a0et\u00a0al.\u00a0(2022) also reported the reaction of sunflower oil to biodiesel using waste chalk/CoFe2O4/K2CO3 as the nanocatalyst. The nanocatalyst was prepared following the co-precipitation chemical method. They found that the surface area, pore volume, and pore size of the catalyst were 5.839 m2 g\u22121, 0.0118 cm3 g\u22121, and 8.08\u00a0nm, respectively. A maximum of 99.27% biodiesel yield was obtained at 2.95\u00a0h of reaction time, 80 \u00b0C of reaction temperature, 15:2 of MTOMR, and 2.65 wt% of catalyst loading. They also reported that the catalyst was successfully reused for 6 consecutive cycles.Due to eco-friendly nature and comparable combustion properties with fossil-based fuel, biodiesel can act as a clean and substitute to fossil fuel. It is found that biodiesel is sometimes auspicious compared to fossil fuel due to its properties such as better lubricity, high cetane number, smaller carbon footprint, and easy biodegradability. Though it is eco-friendly nature, it has to satisfy the term and conditions (Table\u00a011\n) established by the United States Environmental Protection Agency (EPA) and the American Society of Testing and Materials (ASTM) (ASTM,\u00a02003). According to ASTM standard D6751, the maximum water content is 0.05% volume because high water contents can cause cruces such as microbial growth in fuel handling, storage, and transportation equipment (Van\u00a0Gerpen,\u00a02005). The density is also an important property of fuel and according to ASTM, the specific density of fuel at 40 \u00b0C should lie between 0.82\u20130.90\u00a0g cm\u22123 because high density can cause difficulties such as imperfect combustion and particulate matter emission (Blangino\u00a0et\u00a0al., 2008).Viscosity indicates the ability of a material to flow and according to ASTM (ASTM,\u00a02003), it should lie within the range of 1.9\u20136 mm2 s\u22121. Fuel with higher viscosity can form large droplets in the injection and as a result, high energy is required to pump fuel, and poor combustion leads to the emission of greenhouse gasses. The acid value is the quantity of KOH in milligram required to nullify one gram of oil (ASTM,\u00a02003). Since the oil contains FFAs, the acid value can be related to the number of carboxylic groups present in the oil. According to ASTM (ASTM,\u00a02003), the acid value of fuel should be less than 0.5\u00a0mg KOH g\u22121 because a higher acid value can erode the engine and fuel tank. Similarly, cetane number is also an important property for fuel since it measures the delay of the ignition of diesel fuel in a compression ignition engine, and the cetane number of biodiesel fuel should be greater than 47 (ASTM,\u00a02003). Similarly, there are various properties such as boiling point, flashpoint, cloud point, pour point, ash content, and sulfur content. These are also important for fuel, and according to EPA and ASTM, they also have specified values (Table\u00a011).Biodiesel is the FAME produced via catalytic esterification and transesterification processes from biological sources like edible oil, non-edible oil, animal fats, algae oil, etc. Though biodiesel is ecofriendly, it is not extensively commercialized till today due to its high price. The higher price of biodiesel is mainly due to the cost of raw-feedstocks and devices utilized in the preparation process (Skarlis\u00a0et\u00a0al., 2012). Using cheaper oil feedstocks, the production cost can be reduced but these feedstocks have some disadvantages such as high FFAs which lead to the formation of soap (Abdullah\u00a0et\u00a0al., 2017). By developing a suitable technology and enhancing the catalytic activity with an increased yield of biodiesel, the production cost can be reduced.It is known that the application of homogeneous catalyst is not feasible as it causes soap formation and complication during separation (Basumatary\u00a0et\u00a0al., 2018). The heterogeneous catalyst has more advantages because of easy separation and reusable property which has a great contribution in decreasing the cost involved in biodiesel production (Konwar\u00a0et\u00a0al., 2014a; Konwar et\u00a0al. 2014b; Kondamudi\u00a0et\u00a0al., 2011). In this 21st century, the nanocatalysts have attracted everyone's attention due to their large surface area, efficient catalytic activity, easy preparation process, and reusable properties. To reduce the biodiesel cost, researchers nowadays use different non-edible feedstocks such as dairy scum, WCO, animal fat, and tea kernel oil which are some of the cheapest feedstocks. The nanocatalyst preparation methods are also simple and easy. Nanocatalysts can be prepared from wastes and cheap materials such as eggshell, snail shell, coconut shell, scallop waste shell and waste biomass (Tables\u00a01,7,10). The least cost and the reusable properties of these catalysts would bring economic advantages for the production of biodiesel.In conclusion, compared to exhaustible fossil fuels, biodiesel is more reliable due to its non-toxic, renewable, and biodegradable properties. In this review, the meticulous discussion on biodiesel production via transesterification reaction with the help of different nanocatalysts has been systematically presented. Homogeneous, heterogeneous, and enzyme catalysts are employed in the transesterification for the synthesis of biodiesel. Each catalyst has its own merits and demerits, nonetheless the main focus of biodiesel synthesis is the reusability of the catalyst, easy separation and purification process, and minimal waste production, which makes the whole biodiesel production process economically worthwhile. Many researchers reported that heterogeneous catalyst satisfies the above-mentioned requirements but it needs some modification to improve its performance such as providing better selectivity and generating a highly active reaction site. In this review, nanocatalyst types such as metal oxides, nano-hydrotalcite, magnetic nanocatalyst, and nano-zeolite, and other nanocatalysts were comprehensively discussed. These nanocatalysts are considered as better and more efficient for yielding of higher biodiesel and selectivity compared to homogeneous and enzyme catalysts.The applications of magnetic materials as catalysts prepared from the waste-biomass sources have favourable potentials among the present trends of catalysts for biodiesel synthesis. This is due to their better catalytic performance, environment friendliness, rapid and easy separation using external magnetic field, and significant reusability for several cycles of reaction. These characters can reduce the overall time required in processing and energy consumption, and eventually decreases the overall cost of produced biodiesel. Bifunctional nanocatalyst and magnetic bifunctional catalyst from the waste biomass also showed significant efficacy and reusability characters for several reaction cycles, and have high prospective for the ecologically benign and economical synthesis of biodiesel from the low-cost WCO having high FFA content. The most important common factors influencing the biodiesel synthesis in terms of efficiency and economy are the type of oil feedstock, catalyst, and the reactor used. Besides these, other factors such as reaction time and temperature, MTOMR, and catalyst dosage have impacts on the synthesis of biodiesel leading to the overall production cost. In a nutshell, the usage of nanocatalyst in biodiesel synthesis can provide a cheap and clean renewable energy and thus it will become a strong contender for the global industry in the future.In the days to come, low-cost feedstocks such non-edible oil, WCO, waste animal fat, and waste municipal/sewage sludge should be considered for the synthesis of biodiesel by developing a potential and cost-effective catalyst that can simultaneously perform both esterification and transesterification reactions. Utilization of mixed oil feedstocks (hybrid oils) along with the effective reactor and machine learning techniques will certainly help in overall cost reduction of biodiesel. A chain system for continuous-supply of raw oil feedstocks for biodiesel synthesis needs to be strategized to meet the demand of large-scale production.\nShamim Islam: Conceptualization, Investigation, Writing \u2013 original draft. Bidangshri Basumatary: Writing \u2013 review & editing. Samuel Lalthazuala Rokhum: Validation, Writing \u2013 review & editing. Prince Kumar Mochahari: Writing \u2013 review & editing. Sanjay Basumatary: Conceptualization, Supervision, Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors are thankful to the authority of Bodoland University, Kokrajhar, India for the facilities in carrying out this study.", "descript": "\n                  Energy consumption is increasing day by day, thereby depleting the fossil fuel reserve at an alarming rate. The fossil-based fuels have many adverse effects on the environment and cause global warming due to emission of greenhouse gases. Biodiesel produced via the transesterification process is an alternative, eco-friendly, and renewable fuel. Transesterification is carried out using homogeneous, enzyme, and heterogeneous catalysts. Heterogeneous catalysts can resolve the issues faced by the homogeneous and enzyme catalysts during biodiesel synthesis. At the same time, heterogeneous nanocatalysts have much more potential due to their higher surface area, more selectivity, and stronger catalytic activity. In this review, various nanocatalysts such as metal oxides (CaO, MgO, ZnO, Ti2O, CuO, and ZrO2), magnetic nanocatalyst, nano-zeolite catalyst, and nano-hydrotalcite catalysts were studied. In addition, catalyst preparation methods, physical properties of catalyst along with various reaction parameters such as reaction temperature and time, methanol to oil molar ratio (MTOMR), catalyst loading, and biodiesel yield were highlighted and discussed. In short, biodiesel synthesis using nanocatalyst can provide a cheap and clean energy and thus the nanocatalyst can be further developed as a strong candidate for the global energy industry in the future.\n               "}
{"full_text": "Environmental pollution, resource shortages, and increased energy demand are becoming increasingly serious problems [1,2]. As potential solutions, fuel cells, water electrolyzers, and metal\u2013air batteries convert and store renewable clean energy through related electrochemical reactions, including the oxygen reduction reaction (ORR), oxygen evolution reaction (OER), and hydrogen evolution reaction (HER) [3\u20136]. Currently, precious metal catalysts are acknowledged as superior catalysts\u2014for example, Pt for the ORR [7,8] or HER [9,10], IrO2 and RuO2 for the OER [11,12], and so forth. However, precious metals are, by definition, rare and expensive, restricting their large-scale commercial application.In recent years, single-atom catalysts (SACs) with low metal support have improved the utilization of atoms and been used in various reactions due to their high activity and selectivity, and relatively low cost [13,14]. Transition metal-embedded nitrogen-doped graphene (MNx-G) SACs are considered as promising non-precious metal catalysts for electrochemical reactions [15]. Some theoretical studies have been undertaken [16\u201318] to improve the activity of MNx-G SACs successfully prepared in experiments [19\u201322]. However, most of those studies are about Fe/Co/NiN4-G or N2-G, with other metals and coordination largely ignored. Variances in method, model size, and so forthothers also make it difficult to compare results from different studies, even for the same MNx-G. For example, the combined theoretical and experimental works [23,24] showed that FeN2-G was more active than FeN4-G for the ORR, which agreed with the findings from a previous experiment [19]. However, according to the theoretical work by Chen et\u00a0al. [25], the adsorption of O2 on edge FeN2-G was too strong, leading to a low activity. The calculation from Kattel et\u00a0al. suggested that FeN4-G had higher stability and performance than FeN2-G [26]. Recently, Zhang et\u00a0al. found that the ORR activity of CoN2-G was better than that of CoN4-G [27], while Yang and co-workers reported that CoN4-G was the best candidate for the ORR and OER [28]. Hence, a computational study that applies uniform standards is much needed to establish agreement in this area.Notably, Xu et\u00a0al. [29] recently investigated the performance of MN4-G SACs, and Yang's team studied the activity of CoNx-G (x\u00a0\u200b=\u00a0\u200b1\u20134) [28], but there is still a lack of information about coordination between other metals and nitrogen. Although Lin et\u00a0al. [30] applied a machine learning (ML) model to predict the activity of MNxCy-G SACs, the unreliable training data (see Supporting Information Note 1 and Table\u00a0S1) makes the ML model unconvincing. For example, they failed with respect to ZnNx-G's outstanding ORR activity [31,32], NiN4-G's poor HER activity [33], as well as Fe/TcC3-G's HER activity [34]. All these divergences arose from ignorance of the basic principles of quantum mechanics for spin states. A comprehensive study of MNx-G with different metals, nitrogen coordination, and reasonable spin states is therefore necessary to gain an overview of these materials and develop a consensus on them.In this work based on density functional theory (DFT) calculations, the electrocatalytic performance of 3d transition metal single atoms with various forms of nitrogen coordination is systematically explored using unified calculation standards and reaction mechanisms. The adsorption free energies of reaction intermediates, the scaling relationships between them, and the volcano curves are calculated. The activity for the ORR, OER, and HER is investigated to find the optimal SAC configuration. Furthermore, ML is used to unveil possible factors affecting activity, thereby providing solid guidance for improving catalytic performance.All spin-polarized theory calculations were performed using the Vienna Ab-initio Simulation Package (VASP) with the projector augmented wave (PAW) pseudopotentials [35\u201337]. The cutoff energy for the plane-wave basis set was 520\u00a0\u200beV, and the Perdew-Burke-Ernzerhof (PBE) functional was used to describe the exchange-correlation interaction [38]. The MNx-G was modeled on the p(6\u00a0\u200b\u00d7\u00a0\u200b6) graphene unit cell. Geometry optimization was not accomplished until the maximum force per atom was less than 0.02\u00a0\u200beV\u00a0\u200b\u00c5\u20131 and the energy difference between two adjacent electronic steps was less than 10\u20135\u00a0\u200beV. The 5\u00a0\u200b\u00d7\u00a0\u200b5\u00a0\u200b\u00d7\u00a0\u200b1 Gamma-centered k-point grid was employed to sample the Brillouin zone, and a 15\u00a0\u200b\u00c5 vacuum layer was adopted to avoid interaction between surfaces due to periodicity. The Gibbs reaction free energy change (\n\n\u0394\nG\n\n) is estimated by the relation:\n\n(1)\n\n\n\u0394\nG\n=\n\u0394\nE\n+\n\u0394\n\nE\nZPE\n\n\u2212\nT\n\u0394\nS\n\n\n\nwhere \n\n\u0394\nE\n\n is the energy difference of products and reactants, T is room temperature (298.15\u00a0\u200bK), and \n\n\u0394\n\nE\nZPE\n\n\n and \n\n\u0394\nS\n\n are the changes in the zero-point energy and the entropy, respectively. The latter were calculated from the vibrational frequencies for adsorbed species. For gas-phase molecules such as H2, entropy was obtained from the NIST database [39]. The recommended PAW potentials in the VASP manual were applied (Table\u00a0S2).To reduce the influence of different spin configurations on energy, several spin states of each model were considered, and the lowest energy was used for the free energy calculations. The CoN3-G SAC is shown in Table\u00a0S3 as a demonstration. This treatment obeys the fundamental law of Pauli's Exclusion Principle, guaranteeing the reasonableness of the data from the calculations with respect to physics (Supporting Information Note 1).The ML model based on the eXtreme Gradient Boosting (XGBoost) algorithm [40] was achieved using the scikit-learn package [41]. The training data was a subset of the whole data; 80% of the whole data was randomly selected as the training set, while the rest was used as the test set. Three indices were employed to estimate the prediction accuracy: coefficient of determination (R\n2), mean squared error (MSE), and Pearson correlation coefficient (r). The impact of each feature was investigated using the Gini importance method [42]. The specific details about free energy calculation and machine learning analysis are presented in the Supporting Information.In this work, a 3d transition metal single atom (Sc to Zn) is located in various nitrogen coordination structures: double vacancy surrounded by one nitrogen atom (MN1-G in Fig.\u00a01\na), by two nitrogen atoms (MN2-G in Fig.\u00a01b), by three nitrogen atoms (MN3-G in Fig.\u00a01c), and by four nitrogen atoms (MN4-G in Fig.\u00a01d). The MN2-G structure was initially considered to contain three configurations: MN21-G, MN22-G, and MN23-G (Fig.\u00a0S1). For most metals, MN21-G has a lower formation energy and is more stable than the MN22-G and MN23-G structures (Table\u00a0S4). Although the MN22-G structures of Sc, Ti, Cr, and Zn are somewhat stable, it is noteworthy that the energy differences between the MN21-G and MN22-G structures for Sc, Ti, and Cr are tiny. Therefore, to simplify the discussion, this study focuses on the MN21-G structure (Fig.\u00a01b) for MN2-G.The binding energies (\n\n\nE\nb\n\n\n) of single metal atoms on a defective graphene support, the cohesive energies (\n\n\nE\nc\n\n\n) of metal atoms in crystals, and the energy differences \n\n\u0394\n\nE\n\nb\n\u2212\nc\n\n\n=\n\nE\nb\n\n\u2212\n\nE\nc\n\n\n are calculated to determine the stability of the SACs. \n\n\nE\nb\n\n\n and \n\n\nE\nc\n\n\n are obtained according to the following equations:\n\n(2)\n\n\n\nE\nb\n\n=\n\nE\n\nM\n+\nsubstrate\n\n\n\u2212\n\nE\nsubstrate\n\n\u2212\n\nE\nM\n\n\n\n\n\n\n\n(3)\n\n\n\nE\nc\n\n=\n\nE\n\nM\n\u2212\nbulk\n\n\n/\nN\n\u2212\n\nE\nM\n\n\n\n\nwhere \n\n\nE\n\nM\n+\nsubstrate\n\n\n\n is the total energy of the metal atom at the support, \n\n\nE\nsubstrate\n\n\n is the energy of nitrogen-doped defective graphene, \n\n\nE\n\nM\n\u2212\nbulk\n\n\n\n is the energy of the metal crystal, N is the number of metal atoms in the bulk cell, and \n\n\nE\nM\n\n\n is the energy of the isolated metal atom. Negative values for \n\n\nE\nb\n\n\n and \n\n\u0394\n\nE\n\nb\n\u2212\nc\n\n\n\n indicate that the binding between a single metal atom and the substrate is thermodynamically more favorable than the aggregation of metal. All 40 SACs meet the above two conditions (Table\u00a0S5), which suggests MNx-G is stable and matches the observations in experimental and other theoretical works [22,34].First, we study the catalytic performance for the ORR, OER, and HER under acidic conditions (pH\u00a0\u200b=\u00a0\u200b0). For the possible reaction pathway of the ORR, there are mainly two mechanisms: dissociation and association. Since the former, O2 dissociation into two separate O atoms, needs to overcome higher barrier than the latter, O2 hydrogenation to form OOH [43], the more favorable path is chosen:\n\n(4)\n\n\u2217\n+\n\n\n\n\n\u00a0\u200bO\n2\n\n\ng\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\u2194\n\u2217\nOOH\n\n\n\n\n\n(5)\n\n\n\u2217OOH+H\n+\n\n+\n\ne\n\u2212\n\n\u2194\n\u2217\nO\n+\n\nH\n2\n\nO\n\n\nl\n\n\n\n\n\n\n(6)\n\n\u2217\nO\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\u2194\n\u2217\nOH\n\n\n\n\n\n(7)\n\n\u2217\nOH\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\u2194\n\u2217\n+\n\nH\n2\n\nO\n\n\nl\n\n\n\nwhere \u2217 indicates the active sites on the catalyst surface, and (g) and (l) denote gas and liquid phases, respectively. The adsorption of reaction intermediates such as \u2217OOH, \u2217O, and \u2217OH is studied, and their adsorption free energies (\n\n\u0394\n\nG\n\u2217OOH\n\n\n, \n\n\u0394\n\nG\n\u2217O\n\n\n, and \n\n\u0394\n\nG\n\u2217OH\n\n\n) are obtained with H2O and H2: \u2217 \n\n+\n2\n\nH\n2\n\nO\n\u2194\n\n\u2217 \n\nOOH\n+\n3\n/\n2\n\nH\n2\n\n\n, \u2217 \n\n+\n\n\u00a0\u200bH\n2\n\nO\n\u2194\n\n\u2217 \n\nO\n+\n\nH\n2\n\n\n and \u2217 \n\n+\n\n\u00a0\u200bH\n2\n\nO\n\u2194\n\n\u2217 \n\nOH\n+\n1\n/\n2\n\nH\n2\n\n\n. The zero-point energy and entropy of the gas molecules are given in Table\u00a0S6. Taking the relaxed structures of CoN3-G (Figs.\u00a01e\u2013h) as an example, the adsorption structures for the other metal atoms and coordination are similar to them.Generally, the potential-determining step\u2014namely, the step with the largest reaction free energy change\u2014has the highest overpotential, and it thus may also be the rate-determining step (RDS). To identify the RDS in the reaction pathway, the reaction free energy changes of Equations (4)\u2013(7) are calculated using the following formulae, and the results are shown in Tables\u00a0S7\u2013S10.\n\n(8)\n\n\u0394\n\nG\n1\n\n=\n\u0394\n\nG\n\u2217OOH\n\n\u2212\n4.92\n\n\n\n\n\n(9)\n\n\u0394\n\nG\n2\n\n=\n\u0394\n\nG\n\u2217O\n\n\u2212\n\u0394\n\nG\n\u2217OOH\n\n\n\n\n\n\n(10)\n\n\n\u0394\n\nG\n3\n\n=\n\u0394\n\nG\n\u2217OH\n\n\u2212\n\u0394\n\nG\n\u2217O\n\n\n\n\n\n\n\n(11)\n\n\n\u0394\n\nG\n4\n\n=\n\u2212\n\u0394\n\nG\n\n\u2217\nOH\n\n\n\n\n\n\nAs the reverse reaction process of the ORR, the OER possesses the same reaction path, which is the opposite of Equations (4)\u2013(7). The corresponding reaction free energies (\n\n\u0394\n\nG\n5\n\n\n to \n\n\u0394\n\nG\n8\n\n\n, Tables\u00a0S7\u2013S10) are calculated as follows:\n\n(12)\n\n\u0394\n\nG\n5\n\n=\n\u0394\n\nG\n\u2217OH\n\n\n\n\n\n\n(13)\n\n\u0394\n\nG\n6\n\n=\n\u0394\n\nG\n\u2217O\n\n\u2212\n\u0394\n\nG\n\u2217OH\n\n\n\n\n\n\n(14)\n\n\u0394\n\nG\n7\n\n=\n\u0394\n\nG\n\u2217OOH\n\n\u2212\n\u0394\n\nG\n\u2217O\n\n\n\n\n\n\n(15)\n\n\u0394\n\nG\n8\n\n=\n4.92\n\u2212\n\u0394\n\nG\n\u2217OOH\n\n\n\n\n\nFig.\u00a02\n shows the free energy diagrams of the ORR and OER on MNx-G structures. Overall, along the selected ORR path, most of the metal reaction processes are exothermic and proceed in a favorable direction. Compared to the MN1-G system (3 downhill), more SACs in MN2-G (6 downhill), MN3-G (5 downhill), and MN4-G (4 downhill) present a downhill situation in the ORR, suggesting single N coordinated cases may be not good for the ORR. In the free energy diagram of the ORR, it can be observed that the last step of \u2217OH reduction to the second H2O molecule is uphill and holds the maximum free energy change in terms of most systems, implying that it is the RDS. For individual metals, the process of O2 hydrogenation to \u2217OOH is also the RDS. Unlike the ORR, the OER process is endothermic, requiring external energy to promote the reaction. The OER is mainly determined by the process of \u2217O becoming \u2217OOH or the transformation of \u2217OH to \u2217O.We also studied the HER on MNx-G catalysts and the free energy diagram of the HER is shown in Fig.\u00a03\n. When the HER takes place in an acidic environment (pH\u00a0\u200b=\u00a0\u200b0), first protons and electrons interact to form adsorbed H, and then this desorbs into H2 molecules. Based on previous research [33], the following path is taken:\n\n(16)\n\n\u2217\n+\n\n\n\n\n\u00a0\u200bH\n+\n\n+\n\ne\n\u2212\n\n\u2192\n\u2217\nH\n\n\n\n\n\n(17)\n\n\u2217H\n+\n\n\u00a0\u200bH\n+\n\n+\n\ne\n\u2212\n\n\u2192\n\u2217\n+\n\n\u00a0\u200bH\n2\n\n\ng\n\n\n\n\nThe HER activity depends on the adsorption free energy of H (Table\u00a0S11). The \n\n\u0394\n\nG\n\u2217H\n\n\n value of the ideal catalyst for the HER should be close to zero to balance the processes of hydrogen adsorption and desorption. In Fig.\u00a03a, CoN1-G has the minimum \n\n\u0394\n\nG\n\u2217H\n\n\n value of \u20130.51\u00a0\u200beV, and ScN1-G has the maximum \n\n\u0394\n\nG\n\u2217H\n\n\n value of 1.05\u00a0\u200beV. In Fig.\u00a03b, VN2-G has the lower \n\n\u0394\n\nG\n\u2217H\n\n\n value of \u20130.17\u00a0\u200beV, and CuN2-G has the higher \n\n\u0394\n\nG\n\u2217H\n\n\n value of 1.39\u00a0\u200beV. In Fig.\u00a03c, the \n\n\u0394\n\nG\n\u2217H\n\n\n values vary from \u20130.63\u00a0\u200beV for TiN3-G to 0.89\u00a0\u200beV for ZnN3-G. In Fig.\u00a03d, the \n\n\u0394\n\nG\n\u2217H\n\n\n value of TiN4-G is \u20130.58\u00a0\u200beV, and the \n\n\u0394\n\nG\n\u2217H\n\n\n value of CuN4-G is 1.75\u00a0\u200beV. The maximum \n\n\u0394\n\nG\n\u2217H\n\n\n value of MN2/N4-G is more positive than that of MN1/N3-G, which hinders the adsorption process and reduces the HER activity. Taking \n\n\n|\n\n\u0394\n\nG\n\u2217H\n\n\n|\n\n\n < 0.5\u00a0\u200beV as the reference range, we also calculate the variance (\n\n\nS\n2\n\n\n) of the adsorption free energy of H on MNx-G SACs relative to the expected value of 0\u00a0\u200beV. The \n\n\nS\n2\n\n\n values of the SACs are 0.044 for MN1-G, 0.045 for MN2-G, 0.027 for MN3-G, and 0.086 for MN4-G. The small \n\n\nS\n2\n\n\n value for the MN3-G configuration means there are more metal systems around \n\n\u0394\n\nG\n\u2217H\n\n=\n\n 0\u00a0\u200beV, displaying better HER catalytic activity.Since the RDS is related to the adsorption free energy of the reaction intermediates, we further explore the relationship between \n\n\u0394\n\nG\n\u2217OOH\n\n\n, \n\n\u0394\n\nG\n\u2217O\n\n\n, and \n\n\u0394\n\nG\n\u2217OH\n\n\n. Taking \n\n\u0394\n\nG\n\u2217OH\n\n\n as an independent variable, \n\n\u0394\n\nG\n\u2217OOH\n\n\n and \n\n\u0394\n\nG\n\u2217O\n\n\n are plotted versus \n\n\u0394\n\nG\n\u2217OH\n\n\n for all the investigated SACs. By linearly fitting these points, the relationship between oxygenated intermediates is determined to be as follows:\n\n(18)\n\n\u0394\n\nG\n\n\u2217\nOOH\n\n\n=\n0.85\n\u00d7\n\u0394\n\nG\n\u2217OH\n\n+\n3.28\n\n\n\n\n\n(19)\n\n\n\u0394\n\nG\n\u2217O\n\n=\n1.27\n\u00d7\n\u0394\n\nG\n\u2217OH\n\n+\n0.98\n\n\n\n\nIn Fig.\u00a04\na, the data points of \n\n\u0394\n\nG\n\u2217OOH\n\n\n\nvs.\n\n\n\u0394\n\nG\n\u2217OH\n\n\n are well fitted. The slope and intercept of the curve are close to those in previous studies (\n\n\u0394\n\nG\n\u2217OOH\n\n=\n\u0394\n\nG\n\u2217OH\n\n+\n3.2\n\n [11]). Although the linear fit for the data points of \n\n\u0394\n\nG\n\u2217O\n\n\n\nvs.\n\n\n\u0394\n\nG\n\n\u2217\nOH\n\n\n\n displays a weak correlation in Fig.\u00a04a, a scaling relationship with a slope of about 1 has also been proposed in other studies [44,45]. The weak linear relationship between \n\n\u0394\n\nG\n\u2217O\n\n\n and \n\n\u0394\n\nG\n\u2217OH\n\n\n may be related to the RDS of individual SACs. The transformation of \u2217O to \u2217OOH or of \u2217OH to \u2217O is the main RDS of the OER. But for individual systems such as Sc/TiN2-G, their RDS is the process from \u2217OOH to O2. After removing these points, as shown in Fig.\u00a0S2b, the linear relationship between \n\n\u0394\n\nG\n\u2217O\n\n\n and \n\n\u0394\n\nG\n\u2217OH\n\n\n is improved, with R\n2 changing from 0.74 to 0.81. Given that \n\n\u0394\n\nG\n\u2217O\n\n\n, \n\n\u0394\n\nG\n\u2217OOH\n\n\n, and \n\n\u0394\n\nG\n\u2217OH\n\n\n are related to each other, we can simplify the description of ORR/OER/HER activity by using the adsorption free energy of only one intermediate. The catalytic performance of MNx-G SACs is evaluated using the overpotential (\n\n\u03b7\n\n):\n\n(20)\n\n\n\n\u03b7\nORR\n\n=\nmax\n\n{\n\n\u0394\n\nG\n1\n0\n\n,\n\n\u0394\n\nG\n2\n0\n\n,\n\n\u0394\n\nG\n3\n0\n\n,\n\n\u0394\n\nG\n4\n0\n\n\n}\n\n/\ne\n+\n1.23\n\nV\n\n\n\n\n\n\n(21)\n\n\n\n\u03b7\nOER\n\n=\nmax\n\n{\n\n\u0394\n\nG\n5\n0\n\n,\n\n\u0394\n\nG\n6\n0\n\n,\n\n\u0394\n\nG\n7\n0\n\n,\n\n\u0394\n\nG\n8\n0\n\n\n}\n\n/\ne\n\u2212\n1.23\n\nV\n\n\n\n\n\n\n(22)\n\n\n\n\u03b7\nHER\n\n=\n\u2212\n|\n\u0394\n\nG\n\u2217H\n\n|\n/\ne\n\n\n\n\nFor the ideal ORR or OER catalyst, the equilibrium potential \n\n\nU\n0\n\n\n is equal to 1.23\u00a0\u200bV (4.92\u00a0\u200bV/4\u00a0\u200b=\u00a0\u200b1.23\u00a0\u200bV) versus the standard hydrogen electrode (SHE). Therefore, the more the overpotential tends to zero, the better the catalytic activity is.Here, \n\n\u0394\n\nG\n\u2217OH\n\n\n is used as an activity indicator for the ORR, and \n\n\n\u03b7\nORR\n\n\n as a function of \n\n\u0394\n\nG\n\u2217OH\n\n\n is plotted in Fig.\u00a04b. The calculated results are compared with the activity of Pt (111). On the left side of the volcano curve, as the adsorption of \u2217OH weakens, \n\n\n\u03b7\nORR\n\n\n gradually decreases and the ORR activity is promoted. However, on the right side, \n\n\n\u03b7\nORR\n\n\n decreases as the adsorption of \u2217OH strengthens, which indicates that if \u2217OH adsorption is too strong or too weak, the ORR activity will be reduced. When the value of \n\n\u0394\n\nG\n\u2217OH\n\n\n is about 1\u00a0\u200beV, the volcano top has a minimum \n\n\n\u03b7\nORR\n\n\n value, and the catalytic performance is optimal for the ORR. The \n\n\n\u03b7\nORR\n\n\n of the MNx-G systems (Table\u00a0S12) gradually increases in the following order: CoN3-G (0.84\u00a0\u200beV, 0.39\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bCoN4-G (1.07\u00a0\u200beV, 0.42\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bZnN2-G (0.84\u00a0\u200beV, 0.47\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bZnN3-G (0.85\u00a0\u200beV, 0.49\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bZnN4-G (0.89\u00a0\u200beV, 0.54\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bCoN2-G (0.67\u00a0\u200beV, 0.56\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bFeN4-G (0.65\u00a0\u200beV, 0.58\u00a0\u200bV). With a \n\n\u0394\n\nG\n\u2217OH\n\n\n of 0.84\u00a0\u200beV and an \n\n\n\u03b7\nORR\n\n\n of 0.39\u00a0\u200bV, the CoN3-G catalyst exhibits optimal ORR activity. The \n\n\n\u03b7\nORR\n\n\n value of some systems approaches or even exceeds 0.43\u00a0\u200bV of Pt (111) surface, suggesting that these structures may be promising for replacing precious metal materials as SACs for the ORR. In our study, CoN4-G (FeN4-G) has a lower overpotential than CoN2-G (FeN2-G), which is consistent with the results of Yang's group [28]. Unlike previous reports [29,30], ZnNx-G displays excellent ORR activity in our work (Fig.\u00a04b and Table\u00a0S12), agreeing with recent experimental observations on the activity of ZnNx-G catalysts [31,32].The catalytic activity for the OER is described by \n\n\u0394\n\nG\n\u2217O\n\n\n and \n\n\u0394\n\nG\n\u2217OH\n\n\n. The volcano relationship between \n\n\n\u03b7\nOER\n\n\n and (\n\n\u0394\n\nG\n\u2217O\n\n\u2212\n\u0394\n\nG\n\u2217OH\n\n\n) is shown in Fig.\u00a04c. The points deviating from the trend lines are caused by Sc and Ti metal atoms (Fig.\u00a0S2c). As already mentioned, the RDS of Sc and Ti metal atoms is different from the RDS of most other SACs for the OER. After we remove the points with different RDSs, as shown in Fig.\u00a0S2d, the volcanic activity curve of OER is significantly enhanced. When the difference between \n\n\u0394\n\nG\n\u2217O\n\n\n and \n\n\u0394\n\nG\n\u2217OH\n\n\n is around 1.5\u00a0\u200beV, the \n\n\n\u03b7\nOER\n\n\n reaches the minimum value, meaning the highest OER activity. The \n\n\n\u03b7\nOER\n\n\n of the MNx-G systems (Table\u00a0S13) gradually increases in the following order: CoN4-G (1.56\u00a0\u200beV, 0.33\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bNiN2-G (1.71\u00a0\u200beV, 0.51\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bCoN3-G (1.05\u00a0\u200beV, 0.79\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bNiN3-G (2.05\u00a0\u200beV, 0.82\u00a0\u200bV)\u00a0\u200b<\u00a0\u200bCoN2-G (1.11\u00a0\u200beV, 0.86\u00a0\u200bV). Among these, CoN4-G, with an overpotential of 0.33\u00a0\u200bV, shows outstanding catalytic performance for the OER, comparable to the activity of IrO2 (0.65\u00a0\u200bV). According to the activity analysis for the ORR and OER, more of the metals in MN4-G have an overpotential of less than 1\u00a0\u200bV than in other configurations, indicating that N4 coordination seems to be more active than the others. Moreover, CoN4-G possesses high activity in the ORR and OER, suggesting the CoN4-G system is a potential bifunctional SAC for the ORR and OER. An experimental study by Lv et\u00a0al. [46] also showed the high ORR/OER catalytic performance of the CoN4 site.Similarly, the data points for \n\n\n\u03b7\nHER\n\n\n\nvs.\n\n\n\u0394\n\nG\n\u2217H\n\n\n present the volcano curve relationship shown in Fig.\u00a04d. When \n\n\u0394\n\nG\n\u2217H\n\n\n is about 0\u00a0\u200beV, the catalytic performance for the HER peaks, which is consistent with the theoretical analysis. MNx-G systems with excellent HER performance are listed in Table\u00a0S14; their activity gradually decreases in the following order: NiN3-G (\u20130.02\u00a0\u200bV) \n\n\u2248\n\n CuN3-G (\u20130.02\u00a0\u200bV)\u00a0\u200b>\u00a0\u200bCoN3-G (\u20130.04\u00a0\u200bV)\u00a0\u200b>\u00a0\u200bCrN1-G (\u20130.05\u00a0\u200bV) \n\n\u2248\n\n CoN2-G (\u20130.05\u00a0\u200bV)\u00a0\u200b>\u00a0\u200bTiN1-G (\u20130.07\u00a0\u200bV). The HER activity of individual structures, particularly Ni/CuN3-G, is similar to that of Pt(111), with an \n\n\n\u03b7\nHER\n\n\n of 0.09\u00a0\u200bV. Notably, NiN4-G exhibits poor HER activity in our work, with a \n\n\u0394\n\nG\n\u2217H\n\n\n value of 1.68\u00a0\u200beV, which is close to previous experimental [33] (\n\n\u0394\n\nG\n\u2217H\n\n\n\u00a0\u200b=\u00a0\u200b1.62\u00a0\u200beV) and theoretical studies (\n\n\u0394\n\nG\n\u2217H\n\n\n\u00a0\u200b=\u00a0\u200b1.20\u00a0\u200beV [47] and 1.35\u00a0\u200beV [48]). The trend of nitrogen coordination affecting the catalytic performance of the HER is in line with a recent study by Song et\u00a0al. [49].The above discussion has revealed the optimal SAC configurations for the ORR, OER, and HER. The favorable adsorption strength of the reaction intermediates is significant for catalytic performance. We therefore go on to study the intrinsic factors affecting the binding strength between the intermediates and metal atoms.The ML method is applied to find the possible internal factors that affect the activity of MNx-G SACs. By continuously adjusting the ML model and algorithm, we screen out a series of important features to describe \n\n\u0394\n\nG\n\u2217OH\n\n\n (for the ORR), \n\n\u0394\n\nG\n\u2217O\n\n\u2212\n\u0394\n\nG\n\u2217OH\n\n\n (for the OER), and \n\n\u0394\n\nG\n\u2217H\n\n\n (for the HER); these features include the valence electron occupancy of the d orbitals of metal atoms (\n\n\nd\nVe\n\n\n), covalent radius (\n\n\nr\nM\n\n\n), the electronegativity of metal atoms (\n\n\n\u03c7\nM\n\n\n), the coordination number of nearest-neighbor N and C atoms for metal atoms (\n\n\nn\n\nN\n/\nC\n\n\n\n), and the bond length between metal atoms and intermediates (\n\n\nd\n\nM\n\u2212\nO\n/\nH\n\n\n\n). Comparison of the \n\n\u0394\n\nG\n\u2217OH\n\n\n, (\n\n\u0394\n\nG\n\u2217O\n\n\u2212\n\u0394\n\nG\n\u2217OH\n\n\n), and \n\n\u0394\n\nG\n\u2217H\n\n\n values predicted by the XGBoost algorithm with those calculated by DFT are shown in Figs.\u00a05\na, 5c, and 5e, respectively. The trend predicted by the XGBoost algorithm is in good agreement with the value calculated using DFT. A high R\n2 (r) and a low MSE indicate that the ML model is effectively trained for prediction. The Gini importance method is used to explore the impact of each intrinsic feature, as shown in Figs.\u00a05b, 5d, and 5f.The catalytic activity of the MNx-G structures displays varying degrees of correlation with 6 features, where \n\n\nd\nVe\n\n\n plays the most important role in the catalytic activity for the ORR, OER, and HER. The linear relationship between the number of metal valence electrons and the adsorption energy, as revealed in the work of Su et\u00a0al. [50], also proves the importance of \n\n\nd\nVe\n\n\n. Apart from electronic properties, geometric structures such as the bond length \n\n\nd\n\nM\n\u2212\nO\n/\nH\n\n\n\n also affect the activity. In addition, Bader charge analysis shows that more electrons of metal atoms are taken away as the number of N atoms increases (Table\u00a0S15), which affects the adsorption strength and catalytic performance of the MNx-G. As stated in Section 3, the adsorption of an optimal catalyst to intermediates should be moderate. One may modulate the electronic states and adsorption strength of the SACs by adjusting the coordination number of N atoms to improve the catalytic activity, and a recent study on FeN5-G proves this idea [15]. The selected features are simple and easily available physical variables, which is beneficial for screening to find other efficient MNx-G SACs. We consider four elementary reactions, and the RDS of the ORR is usually different to that of the OER. The RDS of the ORR is the last step of \u2217OH reduction to H2O or the process of O2 hydrogenation to \u2217OOH, while the OER is mainly determined by the process of \u2217O to \u2217OOH or the transformation of \u2217OH to \u2217O. As a result, the importance of the key features affecting the ORR/OER activity is dissimilar, though the OER is apparently reversible with the ORR.Inspired by the excellent ORR activity of Co and Zn metal atoms in the above 3d study, we calculate for Rh and Cd, which have the same valence electrons as Co and Zn, to demonstrate the conclusion of the ML method (Table\u00a01\n). As expected, the RhN3-G, RhN4-G, and CdN1-G structures are also potential ORR catalysts with a low overpotential, and RhN3-G may be more active than CoN3-G.The catalytic activity of SACs based on MNx-G (M\u00a0\u200b=\u00a0\u200bSc to Zn, x\u00a0\u200b=\u00a0\u200b1\u20134) for the ORR, OER, and HER has been systematically investigated. Different spin configurations of each model were considered, and the lowest energy was adopted to calculate subsequent reaction free energy. When the corresponding volcano curves were also considered, the optimal SACs for the studied electrochemical reactions were revealed to be CoN3-G for the ORR, CoN4-G for the OER, and Ni/CuN3-G for the HER. In addition to the commonly acknowledged Fe/Co/NiN4 or N2 moiety, other metals or nitrogen-coordinated MNx-G systems, including ZnNx-G and CoN3-G, also displayed outstanding performance. Catalysts with N2, N3, and N4 coordination were found to be more active for the ORR and OER than those with N1 coordination. More metals in the MN3-G configuration\u00a0\u200bwith \n\n\u0394\n\nG\n\u2217H\n\n\n appoaching to zero have superior HER activity. This trend suggests that controlling the concentration of the N component may be more flexible when preparing ORR/OER catalysts for the four coordination cases, while it is better to use as much control as possible with the MN3-G structure when preparing HER catalysts. Furthermore, the ML method revealed that the high catalytic activity of the MNx-G structures can be ascribed to the valence electron occupancy of the d orbitals, the covalent radius, the electronegativity, the coordination number, and the adsorption bond length of the metal atoms. Among these features, the valence electron occupancy of the d orbitals has the greatest influence on the activity for the ORR, OER, and HER. This research paves the way for the development and design of high-performance non-precious metal SACs.Z.P. Hu proposed the concept. C.Y. Zheng performed the density functional theory calculations. X. Zhang and Z. Zhou performed the machine learning analysis. C.Y. Zheng, Z.P. Hu, X. Zhang, and Z. Zhou co-wrote the manuscript. All authors participated in data analysis and manuscript discussion.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Natural Science Foundation of China (No. 21933006, 21773124), the Fundamental Research Funds for the Central Universities Nankai University (No. 63213042) and the Supercomputing Center of Nankai University (NKSC). We thank L.F. Zhang and Q. Gao for the fruitful discussions.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.esci.2022.02.009.", "descript": "\n                  Electrochemical reactions are essential in the processes of energy storage and conversion, and performance is tightly dependent on the electrocatalysts. Herein, we systematically investigate the activity of 3d transition metal embedded nitrogen-doped graphene (MNx-G) for single-atom catalysts (SACs) in the oxygen reduction reaction (ORR), oxygen evolution reaction (OER), and hydrogen evolution reaction (HER). The calculated volcano curves reveal the optimal SAC configuration for each reaction to be CoN3-G for the ORR, CoN4-G for the OER, and Ni/CuN3-G for the HER. Analysis based on the machine learning method suggests that high catalytic performance is dominated by the number of valence electrons occupying the d orbitals, the covalent radius, the electronegativity, the ratio of nearest-neighbor N and C atoms for the metal atoms, and the bond length between metal atoms and adsorbates. This work may shed some light on further studies of the ORR, OER, and HER with non-precious metal SACs.\n               "}
{"full_text": "With the increasing global energy demand, non-renewable fossil fuels play an important role. But the overuse of fossil fuels has worsened the environment. Hence, developing clean energy is a significant direction for sustainable development [1]. Hydrogen is considered as a green energy carrier to eliminate our dependence on fossil fuels [2]. However, hydrogen almost does not exist in nature, the actual use of hydrogen is prepared by artificial reaction. Almost all hydrogen comes from industrial steam reforming of natural gas, which consumes a large amount of energy and emits a lot of carbon dioxide. At the same time, hydrogen produced by this method contains a trace amount of carbon monoxide, which can poison platinum catalysts in fuel cells. In contrast, electrolysis of water can produce clean hydrogen [3]. Although platinum-based catalysts still have the best hydrogen evolution performance in the field of electrolytic water, precious metal-based catalysts are not only too expensive, but also have very limited reserves on the earth. It is difficult to be applied to the actual industrial production of hydrogen on a large scale [4]. Hence, seeking non-noble metal catalysts with excellent hydrogen evolution performance plays an important role in entering the hydrogen energy society [5]. Recently, many reports have shown that oxide [6], carbide [7] and chalcogenide [8] possess a good hydrogen evolution properties. In the study of non-noble metal electrocatalysts for hydrogen evolution reaction (HER), transition metal sulfides have attracted attention because of their rich natural reserves and high electrocatalytic activity for hydrogen evolution. Nickel sulfide has excellent hydrogen evolution performance because of its similar structure to the hydrogenase activity site [9]. Compared with oxygen, sulfur has lower electronegativity and can form different nickel sulfide phases (NiS, NiS2, Ni3S2) after coupling with nickel [10]. Moreover, nickel sulfides have metal-like band structure and excellent electrical conductivity [11]. It has a great potential in the field of electrolytic water. Compared with carbon materials such as carbon cloth, carbon nanotubes and graphene, carbon spheres can be fabricated by a facile hydrothermal method, which is an ideal electrocatalyst carrier material.Herein, by using a simple hydrothermal method, we use carbon spheres as the support and grow Ni3S2 and NiS on its surface as the catalytic active materials. And then we analyze the crystal structure, microstructure and electrochemical hydrogen evolution properties of the composites. Besides, hydrogen evolution reaction is the key half reaction of water splitting, which is a clean but high energy-consuming technology. Hence, developing electrocatalysts with good catalytic activity to reduce energy consumption is essential.Glucose monohydrate (C6H12O6\u00b7H2O) was bought from Macklin Co., Ltd. Cetyl Trimethyl Ammonium Bromide (CTAB) and Nickel sulfate hexahydrate (NiSO4\u00b76H2O) were furnished by Keshi Co., Ltd. Thioacetamide (C2H5NS) was supplied by Bidepharm Co., Ltd.XRD-6100 was used to study the crystal structure of the catalysts. The scanning rate was 4\u00b0\u00b7min\u22121 and 2\u03b8 were ranging from 10\u00b0-80\u00b0. Using a step-scanning mode with a step size of 0.02\u00b0, the data for the Rietveld method (by Fullprof software) was obtained. By using SEM, the surface morphology of the carbon spheres, nickel sulfides and nickel sulfides supported by carbon spheres were gained. The surface element composition was probed by EDS. The chemical state of the catalysts was analyzed by XPS. Adsorption isotherms and pore size distribution were obtained by N2 on the Autosorb-3B automatic physical adsorption instrument in 77\u00a0K.The hydrothermal method was used to prepare carbon spheres. 3.5\u00a0g of glucose monohydrate (C6H12O6\u00b7H2O) and 0.5\u00a0g of Cetyl Trimethyl Ammonium Bromide (CTAB) were blended with 65\u00a0ml deionized water and stirred to get a homogeneous solution. Then the solution was poured into the autoclave. The temperature was set up to 180\u00a0\u00b0C and the reaction time was 6\u00a0h. The samples were cleaned several times with deionized water and ethanol. The products were put into the oven to dry for use.The carbon spheres (0.025\u00a0g) were adding into a beaker containing 30\u00a0ml deionized water and 30\u00a0ml DMF. With the assistance of ultrasonication, the solution was mixed evenly. Nickel sulfate hexahydrate (NiSO4\u00b76H2O, 0.263\u00a0g) and thioacetamide (C2H5NS, 0.075\u00a0g) were then added into the solution under stirring. The solution was poured into a 100\u00a0ml autoclave. Besides, the temperature was set up to 200\u00a0\u00b0C and the reaction time was 24\u00a0h. The as-obtained product was cleaned by deionized water and ethanol to wash away impurity ions and dried in the oven for use (denoted as C25-M). C-M materials were also prepared by altering the relative ratios of carbon spheres (0.010\u00a0g and 0.040\u00a0g for C10-M and C40-M, respectively). Ni3S2 and NiS were prepared under the same condition without adding carbon spheres into the beaker (denoted as NixSy).By using a three-electrode system (the glassy carbon electrode coated with catalysts as the working electrode, Ag/AgCl electrode as the reference electrode and graphite rod as the counter electrode), the hydrogen evolution performance of the products were researched. In the experiments, we used a magnetic stirring device for to remove bubbles produced on the electrode surface during the catalytic process. By using a magnetic stirring device during the test, the solution is fully stirred to make the electrode surface concentration basically the same as the bulk concentration. In this case, the effect of mass transport could be minimized.The potential values were switched to the reversible hydrogen electrodes (RHE) according to the Nernst equation [12]:\n\n\n\nE\n\n\n\n\n\nV\n\nv\ns\n.\n\nR\nH\nE\n\n\n\n\n=\nE\n\n\n\n\n\nV\n\nv\ns\n.\n\nA\ng\n/\nA\ng\nC\nl\n\n\n\n\n+\n0.197\n+\n0.059\n\u2217\np\nH\n\n\n\n\nThe working electrode prepared by this procedure are as follows: 1) Taking 5\u00a0mg of catalyst and mixng with 500\u00a0\u03bcL deionized water and 500\u00a0\u03bcL ethanol with ultrasonication for 30\u00a0min; 2) Dropping 5\u00a0\u03bcL of the mixed ink on the electrode; 3) Dropping 2\u00a0\u03bcL of 0.5% Nafion solution on the electrode. The electrolyte was degassed by nitrogen for 30\u00a0min. Linear sweep voltammograms (LSV) were gained in N2-saturated 0.5\u00a0M H2SO4 solution. The potential range was set up to 0 to\u00a0\u2212\u00a00.9\u00a0V vs. RHE at a scanning speed of 5\u00a0mV\u00b7s\u22121. Tafel slopes could been derived from the resulting polarization curves using the Tafel formula. To evaluate the durability, i-t curve (i-t) was recorded over C25-M under a working potential of \u22120.4\u00a0V vs. RHE for 18\u00a0h. The double layer capacitance (Cdl) and electrochemical impedance spectra (EIS) were measured to assess the electrochemical activity. All the data were gained without IR compensation.From the XRD scheme (Fig. 1\n), carbon spheres merely possess a wide diffraction peak at 22\u00b0, which is coincided with the former literature [13]. The peaks at 21.8\u00b0, 31.1\u00b0, 37.8\u00b0, 44.3\u00b0, 49.7\u00b0, 50.1\u00b0 and 55.2\u00b0 are corresponding to (101), (110), (003), (202), (113), (211) and (122) lattice planes of Ni3S2 (JCPDF#44\u20131418). Other diffraction peaks at 18.4\u00b0, 30.3\u00b0, 32.2\u00b0, 35.7\u00b0, 40.5\u00b0, 48.8\u00b0, 52.6\u00b0, 57.4\u00b0 and 59.7\u00b0 are from the (110), (101), (300), (021), (211), (131), (401), (330) and (012) crystallographic planes of NiS (JCPDF#12\u20130041). The peaks illustrated that Ni3S2 and NiS possess good crystallinity and high purity. By using Rietveld calculation, we conclude that the weight fraction of NiS is 37.47% and the weight fraction of Ni3S2 is 62.53%. Using the Scherrer formula, the calculated crystal size of C25-M is 26.8\u00a0nm.\nFig. 2\na exhibits the microstructure of nickel sulfides. From Fig. 2b, it can be illustrated that the even diameter of carbon spheres is between 300 and 400\u00a0nm. As depicted in Fig. 2c, Ni3S2 and NiS generated on the carbon spheres and distributed evenly. The diameter of composite increased owing to the growth of nickel sulfides. After introducing the carbon spheres, the Ni3S2 and NiS particles are more refined, which helps to prevent aggregation, migration and structural destruction of nickel sulfides during the catalytic reaction process. The elemental distribution of C25-M is revealed by EDS mapping images (Fig. 3\na-e). From the scheme, Ni and S atoms distribute homogenously over carbon spheres and other impurities are not detected. In Fig. 4\n, adsorption isotherms and pore size distribution were provided. The surface area of C25-M is 67.12\u00a0m2/g, which is lager than carbon spheres and NixSy. It can also be seen that the pore sizes of C25-M mainly distributed at 1.867\u00a0nm.Furthermore, Fig. 5\n shows X-ray photo-electron spectroscopy (XPS) of C25-M. In the Ni 2p region of the composite (Fig. 5a), the peaks situated at 872.98 and 855.18\u00a0eV could be ascribed to 2p1/2 and 2p3/2 of Ni2+ and peaks situated at 875.18 and 857.48\u00a0eV are bound up with 2p1/2 and 2p3/2 for Ni3+. The peaks of Ni2+ are contributed by NiS and Ni3S2 and peaks of Ni3+ are related to Ni3S2. There are two peaks at 860.78 and 879.18\u00a0eV, which are from the satellite peaks of Ni 2p3/2 and Ni 2p1/2\n[14]. From S 2p diagram, the peaks at 169.18 and 170.28\u00a0eV are bound up with the sulfur in nickel sulfides [15]. The analysis mentioned above illustrates that the nickel sulfides emerge in the products.\nFig. 6\n exhibits the electrochemical data. In the LSV curve, overpotential is an important parameter to appraise the hydrogen evolution performance. From Fig. 6a, carbon spheres exhibits a very poor catalytic activity. In contrast, all C-M samples show lesser overpotential than nickel sulfides, which demonstrates the advantage of growing Ni3S2 and NiS on the surface of carbon spheres. To explore the optimal proportion of carbon spheres, we used 10, 25 and 40\u00a0mg carbon spheres as the support and grow nickel sulfides on them, respectively. Among all the C-M samples, C25-M displays the optimal hydrogen evolution activity. It is strange that with more carbon spheres adding into the autoclave, the catalytic capability of the composite is not always improving. Although C40-M has the largest amount of carbon spheres, its overpotential is larger than C25-M, which is probably caused by excessive carbon spheres resulting in a abatement of active sites on the carbon spheres. In addition, we conduct linear sweep voltammograms performance test on the commercial Pt/C catalyst. The overpotential of commercial Pt/C at a current density of 10\u00a0mA\u00b7cm\u22122 is 57\u00a0mV. And the overpotential of C25-M is negatively shifted by 259\u00a0mV compared to the Pt/C catalyst.To analyze the mechanism of the reaction, Tafel slopes are obtained by plotting overpotential against the logarithm of current density. As depicted in Fig. 6b, Tafel slopes of carbon spheres, nickel sulfides and C25-M are 156\u00a0mV/dec, 120\u00a0mV/dec and 82\u00a0mV/dec, respectively. It can be illustrated that in virtue of carbon spheres, the Tafel slope of the composite is evidently reduced. This signifies that the C25-M can obtain a higher catalytic current at the same increment of overpotential, thus improving its catalytic activity. In the acid solution, hydrogen evolution reaction involves two pathways [16]. The first step is that the proton in the solution binds to the electron to adsorb and form the adsorbent hydrogen atoms (Hads) at the active site of the catalyst surface (Volmer step). The second step is that the active hydrogen atoms (Hads) desorb to form hydrogen. This step may have two reactions due to the reaction dynamics and nature of different catalysts: adsorbent hydrogen atoms (Hads) at two adjacent catalytic sites combine to generate hydrogen (Tafel step) or adsorbent hydrogen atoms (Hads) bind to a proton in the solution to produce hydrogen (Heyrovsky step).\n\nVolmer step: H+ + e- \u2192 Hads\n\n\n\n\n\nHeyrovsky step: Hads\u00a0+\u00a0H+ + e- \u2192 H2\n\n\n\n\n\nTafel step: Hads\u00a0+\u00a0Hads\u00a0\u2192\u00a0H2\n\n\n\nEach reaction step has an important impact on the hydrogen evolution process. The reaction mechanism is determined by the Tafel slope. C25-M displays a smaller Tafel slope (82\u00a0mV/dec) and the Tafel slope of this electrocatalyst is between the theoretical values of Volmer step (120\u00a0mV/dec) and Heyrovsky step (40\u00a0mV/dec), so the reaction mechanism of C25-M is Volmer-Heyrovsky mechanism [17]. In Table 1\n, the Tafel slopes of similar sulfur-rich HER catalysts are listed. Compared with similar sulfur-rich HER catalysts, the tafel slope of C25-M is lower than others, which means that proton adsorption is easier than others. Using the Scherrer formula, the calculated crystal size of NiS in reference [9] is 28.6\u00a0nm, which is larger than that of C25-M (26.8\u00a0nm). This means that a smaller crystal size can expose more active sites, thus to improve the hydrogen evolution reaction performance. The electrocatalytic performance of Mo-Ni3S2/NF [18] and Co-Ni3S2\n[19] is enhanced by element doping. And the hydrogen evolution performance of NiSx-3 [20] catalyst is improved by changing the molar ratio of NiS and Ni3S4. We use carbon spheres as support to enable the substantial and effective exposure of the catalytic active material (nickel sulfides). Moreover, the good stability and high electron conduction properties of the carbon spheres promote the hydrogen evolution process, thus to reduce the Tafel slope of C25-M.To better understand the reasons why each electrode material has different catalytic properties, the electrochemical active surface area (ECSA) is appraised by testing the double-layer capacitance (Cdl) of the C25-M electrode and other contrast electrode materials in the non-Faradaic voltage region (\u22120.043\u00a0~\u00a00.057\u00a0V vs. RHE). The cyclic voltammetry curves at different scanning speeds (20, 40, 60, 80, 100 and 120\u00a0mV/s) of C25-M and other contrast data was achieved by applying CV technique. As depicted in Fig. 6c and d, C25-M has the largest electrochemical double-layer capacitance value, which is exceeding the values of carbon spheres and Nickel Sulfides.Electrochemical impedance spectroscopy (EIS) can further reflect the electrode kinetic characteristics of electrocatalysts during catalytic hydrogen evolution. Fig. 6e reveals the Nyquist diagram of different catalytic active materials and the corresponding equivalent fitting circuit. Compared with carbon spheres and Nickel Sulfides, C25-M exhibits a smaller semicircle diameter, indicating that compositing carbon spheres and sulfides could decrease charge transfer resistance. This result illustrates that combing carbon spheres and sulfides together is propitious to the electron conduction between the active sites.For an ideal electrocatalyst, it needs not only a small HER overpotential, but also a good catalytic hydrogen evolution stability. Consequently, we further explore the durability of C25-M (Fig. 6f). The durability of C25-M is tested at a voltage of \u22120.4\u00a0V vs. RHE. After 18\u00a0h, the composite material still maintains high hydrogen evolution capacity. Further, we conduct stability tests on the Pt/C catalyst and C25-M (Fig. 7\n). The Chronoamperometric experiment of C25-M and Pt/C is tested at a voltage of \u22120.04\u00a0V vs. RHE. With time increases, the current of Pt/C catalyst and C25-M both decreases. After 18\u00a0h, the current of Pt/C drops more than that of C25-M. This means that the C25-M catalyst still maintains higher hydrogen evolution capacity than the commercial catalyst Pt/C. Accordingly, the synthesized catalyst C25-M has a long-range electrochemical catalytic stability.In summary, Ni3S2 and NiS supported on carbon spheres were successfully fabricated via a facile hydrothermal method. The composite has been testified to actualize a current density of 10\u00a0mA\u00b7cm\u22122 at a lesser overpotential compared with Nickel Sulfides. The outstanding hydrogen evolution capacity of this composite is ascribed to the well-distributed dispersion of Ni3S2 and NiS nanoparticles by carbon spheres, which effectively promotes the exposure of hydrogen active sites. Moreover, carbon spheres support and immobilize Ni3S2 and NiS nanoparticles, which effectively alleviates the migration and aggregation of active substances during hydrogen evolution. These results demonstrate its potential as an electrocatalyst for hydrogen evolution and we are convinced of the fact that this experimentation will open up a way to rationally design electrocatalyst in the future.\nTong Gao: Writing - original draft. Ming Nie: Writing - review & editing. Jin Luo: Writing - review & editing. Zhi Huang: Data curation. Hai Sun: Conceptualization. Peitao Guo: Data curation. Zhenhong Xue: Visualization. Jianming Liao: Investigation. Qing Li: Supervision, Validation. Liumei Teng: Data curation.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This study is supported by Chongqing Municipal Training Program of Innovation and Entrepreneurship for Undergraduates (Project No: S202010635145), Zeng Sumin Scientific Research Program (Project No: zsm20190629), Fundamental Research Funds for the Central Universities (XDJK2020B004), Chongqing Key Laboratory for Advanced Materials & Technologies of Clean Energies (JJNY202002), Chongqing Graduate Student Research Innovation Project (CYS19106) and Chongqing University Key Laboratory of Micro/Nano Materials Engineering and Technology (KFJJ2015, KFJJ2009).", "descript": "\n                  Ni3S2 and NiS supported on carbon spheres are successfully synthesized by a facile hydrothermal method. And then a series of physical characterizations included XRD (X-ray diffraction), EDS (energy dispersive spectroscopy), FESEM (field emission scanning electron microscopy) and XPS (X-ray photo-electron spectroscopy) were used to analyze the samples. XRD was used to confirm that Ni3S2 and NiS were successfully fabricated. FESEM indicated that Ni3S2 and NiS disperse well on carbon spheres. Electrochemical tests showed that nickel sulfides supported by carbon spheres exhibited excellent hydrogen evolution performance. The excellent catalytic activity is attributed to the synergistic effect of carbon spheres and transition metal sulfides, of which the carbon spheres act to enhance the electrical conductivity and the dispersion of Ni3S2 and NiS, thus providing more active sites for the hydrogen evolution reaction.\n               "}
{"full_text": "No data was used for the research described in the article.In order to understand the processes occurring on the surface of adsorbents and catalysts and thus be able to modify and optimize them, we need to study their surface properties. In the case of supported metal catalysts, the metal surface must be exposed to these processes, in other words the metal must be dispersed in order to be accessible. This process can be performed during the early stages of the material preparation method [1,2]. Therefore, being able to analyze and characterize a metal that is dispersed in small particles on a surface is of vital importance. In general, the surface structure of a metallic crystal varies greatly depending on the size of the particles and, especially, on how they are found on the surface of a support and their size distribution. Likewise, the surface of the support may affect the properties of the particles, making them different to those of isolated particles. Similarly, the acid\u2013base properties of catalytic supports and catalysts are also important as they may affect catalyst preparation [1,2], and play a role in the mechanism of the reactions [1]. In order to understand the catalytic and adsorption behavior, evaluating the properties of such materials is essential.The dispersion of a supported metal is defined as the fraction of metal atoms found on its surface and can be related to the number of metal centers accessible to reactants and products [1]. Understanding this dispersion is crucial to interpret the kinetic data of a catalytic reaction and to compare catalysts of the same family; the Turn Over Frequency (TOF), defined as the number of molecules reacting per active site per second or the number of molecules reacting per surface metal atom per second, is key for this purpose. As such, several techniques have been proposed and used to measure the TOF, including transmission electron microscopy and X-ray diffraction; the most widely used, however, is surface and selective gas adsorption [2\u20135]. Chemisorption methods are particularly important for highly dispersed catalysts given the difficulty in estimating their dispersion using other techniques, such as X-ray diffraction or electron microscopy. The most frequently used gases in chemisorption are H2, CO, O2 and even N2O. Other gases are used in specific cases, such as NO, H2S, CS2, C6H6, etc. Several organizations (e.g., American Society for Testing and Materials - ASTM; International Union of Pure and Applied Chemistry-IUPAC) have provided guidance for analyzing the surface of metals, generally recommending hydrogen as an adsorbent [6]. Hydrogen has the advantage of being mainly chemisorbed on the metallic part of the surface, and the amount retained on the non-metallic part is relatively small and weak in many cases. Hydrogen is physically adsorbed on metallic and non-metallic parts, but when measuring at room temperature and pressure, the contribution of the physically adsorbed layer can be neglected due to the very small adsorption enthalpy of hydrogen (less than 8\u00a0kJ/mol). However, an exception may be necessary for activated carbon and MOFs as supports, especially when this material has a high specific surface area. Indeed, these materials, which can have a surface area of 5000\u00a0m2/g and sometimes more, have been used in H2 storage [7]. These values can be corrected by measuring the physical adsorption of hydrogen on the metal-free support separately. CO adsorption also allows the determination of metallic surfaces, since there is a specific reaction between the gas molecule and the metal. The main drawback of CO chemisorption is determining the CO/metal stoichiometry, since the CO molecule can form various types of bonds with the metal, as well as polynuclear complexes [8]. Transmission electron microscopy, in turn, has the great advantage of providing a direct view of the particles to be analyzed. This analysis allows the size distribution and all its characteristic parameters to be obtained, and shows whether the particles formed are large, what shape they have, and it is even possible to determine their crystal structure. In the case of X-ray diffraction, measurement of the size of metal crystallites is based on the presence of diffraction lines provided the particles are sufficiently small. Quantitatively, it can be determined using the Scherrer equation, which relates the average diameter of the crystals to the broadening of the diffraction peaks. The disadvantage of this technique is that it can only be applied to samples presenting a diffraction line, which means that it cannot be used for catalysts with a very low metal loading (less than 1 % by weight, although this value depends on the metal). Crystalline supports or supports presenting diffraction lines can interfere with the determination of the diffraction lines. In general, glass particles with a size of between 5 and 50\u00a0nm can be determined. In conclusion, most authors/researchers indicate that selective gas chemisorption is the best method for characterizing/determining the active surface of a metal catalyst, due to several reasons, especially to its easy accessibility, however, other techniques may be necessary to corroborate the presence of very large crystal materials.Acid catalysts are very important in alkylation, dealkylation, cracking, hydrocracking, isomerization and reforming reactions [1], all of which are used in petroleum refinery processes. Two types of acid centers can be distinguished: Lewis and Br\u00f6nsted. Lewis centers accept a pair of electrons from the adsorbed species to form a coordination bond between the adsorbed molecule and the solid surface. Br\u00f6nsted centers however, provide a proton to the adsorbed molecule to form an ion-dipole interaction between the adsorbed species and the solid surface. Both types of centers are found in alumina, for example [1], where aluminum acts as the Lewis acid center and the OH groups on the surface as the Br\u00f6nsted centers. Characterization of surface acidity is normally performed using techniques such as infrared spectroscopy, NMR spectroscopy, thermal analysis and titration with basic molecules such as ammonia or pyridine (C5H5N), although numerous studies have rather focused on temperature-programmed desorption (TPD). Base catalysis is less widespread than acid catalysis\u2014as it provides lower yields\u2014although it is more selective than the latter. Lewis bases act as an electron donor and can be related to the surface lattice oxygen, O2-. Pyrrole (C4H5N), deuterated chloroform (CDCl3), H2S and CO2 have been suggested as probe molecules [9,10]. Other applications in which the surface nature of the materials may also be important\u2014such as the use of silicates as anticorrosive agents, preventing the deterioration of washing machines, zeolitic materials as ion exchangers, activated carbons with a hydrophilic character applied as adsorbents\u2014are also worth mentioning.A several of factors that can affect gas chemisorption measurements have been reported, for instance the presence of impurities in the catalysts, surface reconstruction due to sintering during adsorption, the nature of metal/support interactions, spillover of the H2 molecule, and even the contamination of gases used in chemisorption [11,12]. In catalytic supports such as TiO2, V2O3, CeO2\u2014which exhibit a high degree of reactivity with metals\u2014supported metal particles and other metal species originating from the \u201cstrong metal\u2013support interaction\u201d (SMSI) can occur; this phenomenon reduces the adsorption capacity of the metal [13]. The SMSI state enables the metallic particles to present reversible characteristics, that is, reduction of the particles at low temperatures allows them to be in their metallic state, while the same process at high temperature favors the SMSI state. This state causes the support to develop semiconductor and even metallic properties that differ from those presented initially. Several models have been proposed to explain the SMSI phenomenon. One of these is the formation of metallic alloys, and another proposes that the reduced species on the support can present high mobility and are capable of coating the metallic particles, thereby blocking their adsorption capacity [13]. In the case of the spillover phenomenon, this involves the transport of active species adsorbed or formed in a first phase (structure) to another in which they are not generated directly under the same conditions. The most common example is hydrogen adsorbed from the gas phase onto a metal (Pt, Pd, Ni, etc.), where it dissociates into atomic hydrogen. The dissociated hydrogen can subsequently be transported to the support. This phenomenon causes more hydrogen to be adsorbed than is necessary in the chemisorption process, thereby interfering with monolayer volume determinations. Thus, concludes into erroneous dispersion results.The techniques and procedures presented below are often routine in many laboratories, since they allow the evaluation and determination of the surface properties of materials through chemisorption processes. The aim of this work is to review them and include the updates published by several researchers, who mostly aim to explain the results of bifunctional metallic and acid\u2013base catalytic behavior.Two phenomena can be observed in the adsorption process: physisorption and chemisorption [14]. In general, differentiating between these two processes is not easy, especially since intermediate behaviors can occur. Interactions between the adsorbent surface and the adsorbate are generally relatively weak via coulombic and dispersion forces, although defection at the atomic level or atoms with the availability to form bonds may be present on the surface. In such a case, chemical bonds can be formed and the process is known as chemisorption. This process often occurs at temperatures higher than the critical temperature of the adsorbate. Chemical adsorption is often irreversible, at least under mild conditions, and is characterized by large interaction potentials that lead to high adsorption heats, although this factor is not the only aspect that differentiates physisorption from chemisorption.Physical and chemical adsorption are usually characterized by the following properties:In physical adsorption, the gas molecules interact with the solid surface via van der Waals-type forces. This type of interaction determines the characteristics of the adsorption:\n\n\u2013\nphysical adsorption involves a weak interaction between gas molecules and the surface of the solid. As such, no surface modifications occur during adsorption measurements;\n\n\n\u2013\nphysical adsorption is an exothermic process: the interaction forces are attractive, and the heat released is similar to the enthalpies of condensation of the adsorbed substance (20\u201340\u00a0kJ/mol). As this process is exothermic, physisorption increases with decreasing the adsorbent temperature or increasing the adsorbate pressure;\n\n\n\u2013\nthe physisorbed molecule maintains its identity, since the energy is insufficient to break the bond, although its geometry can be distorted.\n\n\n\u2013\nphysisorption is a non-specific process, since the forces involved are not specific either. Molecules do not usually interact with specific adsorption centers.\n\n\n\u2013\nphysisorption occurs in multilayers, meaning that another layer can be adsorbed on top of a layer of adsorbed molecules. The first adsorbed layer is formed by direct interaction with the surface, while the successive ones are interactions between molecules, like the condensation process. However, the difference between these two processes is not so clear and we often find intermediate situations, especially when the chemisorption process is weak.\n\n\nphysical adsorption involves a weak interaction between gas molecules and the surface of the solid. As such, no surface modifications occur during adsorption measurements;physical adsorption is an exothermic process: the interaction forces are attractive, and the heat released is similar to the enthalpies of condensation of the adsorbed substance (20\u201340\u00a0kJ/mol). As this process is exothermic, physisorption increases with decreasing the adsorbent temperature or increasing the adsorbate pressure;the physisorbed molecule maintains its identity, since the energy is insufficient to break the bond, although its geometry can be distorted.physisorption is a non-specific process, since the forces involved are not specific either. Molecules do not usually interact with specific adsorption centers.physisorption occurs in multilayers, meaning that another layer can be adsorbed on top of a layer of adsorbed molecules. The first adsorbed layer is formed by direct interaction with the surface, while the successive ones are interactions between molecules, like the condensation process. However, the difference between these two processes is not so clear and we often find intermediate situations, especially when the chemisorption process is weak.In chemical adsorption, the gas molecules interact with the solid surface through chemical bonds. Similarly, to physical adsorption, this type of strong interaction conditions the characteristics of the adsorption:\n\n\u2013\nin chemical adsorption, the interaction forces are attractive, and the heats released are similar to the enthalpies of formation of a chemical bond (100\u2013500\u00a0kJ/mol). In chemisorption, both bond formation and bond breakage can occur, so the values of these enthalpies can be both positive and negative.\n\n\n\u2013\nas there is a strong interaction\u2014bond formation\u2014between the molecule and the adsorption center, chemical adsorption is only defined in a monolayer. In the rest of the layers, physical adsorption may occur.\n\n\n\u2013\nif a chemical bond is formed, the chemisorbed molecule does not maintain the same structure as in the gas phase.\n\n\n\u2013\nchemisorption is specific. There are certain centers on the surface of the solid at which interaction occurs whereas at others it does not.\n\n\nin chemical adsorption, the interaction forces are attractive, and the heats released are similar to the enthalpies of formation of a chemical bond (100\u2013500\u00a0kJ/mol). In chemisorption, both bond formation and bond breakage can occur, so the values of these enthalpies can be both positive and negative.as there is a strong interaction\u2014bond formation\u2014between the molecule and the adsorption center, chemical adsorption is only defined in a monolayer. In the rest of the layers, physical adsorption may occur.if a chemical bond is formed, the chemisorbed molecule does not maintain the same structure as in the gas phase.chemisorption is specific. There are certain centers on the surface of the solid at which interaction occurs whereas at others it does not.The two adsorption processes (physical and chemical) can be illustrated by representing the evolution of the potential energy of a gaseous diatomic molecule in the vicinity of a surface, where attractive and repulsive forces may appear (\nFig. 1) [15]. This figure includes the option of diatomic adsorption or bond cleavage and atomic adsorption. If adsorption occurs, the potential energy decreases, thus implying that the concentration of the gas will be higher on the surface than inside the gas, due to the adsorption phenomenon. In this situation, if the gas molecule is very close to the surface of the adsorbent, the potential increases again because of the repulsion effect. The figure illustrates how molecule B2 approaches the surface of a material at a distance r. The first interaction process is physical adsorption of the molecule on the surface of the solid. The equilibrium situation is represented by the potential minimum or adsorption potential well, which is characterized by a negative energy value (exothermic process). Below is an endothermic process in which an energy E must be overcome. If this energy value is exceeded, a new equilibrium situation can be reached but, in this situation, dissociation of the diatomic molecule occurs. Each of these situations depends on the adsorbate/adsorbent system and the temperature of the adsorption process. Three situations are represented in the figure. In the first (a), the molecule is more strongly adsorbed (its equilibrium state has a lower energy) than in the dissociated state. This the preferred form of adsorption and could represent physical adsorption. In the second case (b), the dissociated situation has a higher adsorption energy than the diatomic molecule, but there is an energy barrier to overcome. If this barrier is high enough, we would have the first case. Finally, the third case (c) is similar to the previous one, but with a very low energy barrier, so dissociated adsorption normally occurs.The bond between the chemisorbed molecule and the adsorption center is often very energetic, even though the net heat of adsorption may be low. The requirement to overcome an activation energy in chemisorption explains the low heat of adsorption and also why such a phenomenon can be relatively slow. Since chemisorption is often an activated process, the net heat of adsorption is small at low temperatures and large at high temperatures. This situation means that physisorption predominates at low temperatures and chemisorption at higher temperatures.In contrast to measurement of the specific surface area, the surface area of a catalyst component, usually the metal surface, can be measured using selective adsorption (chemisorption). The principle of selective surface area measurement by chemisorption is similar to specific surface area measurement by physisorption. As such, it will be necessary to make a series of assumptions: the metallic surface is free from other adsorbates such as carbon or other poisons that prevent or affect the gas\u2013solid interaction; metal atoms must be in its normal metal state (normally zero) that allows interaction; and the stoichiometry of the interaction must be known and be independent of the size of the metal crystal [11,12]. As such, preparation and pretreatment of the sample have to be more rigorous than when characterization is carried out by physisorption. Intrinsic to each of the techniques to be used, the kinetics and strength of the adsorption are important aspects that must be evaluated. The techniques that allow the chemical adsorption process to be analyzed and, therefore, the active sites (acid-basic and metallic adsorption centers) to be characterized, can be divided into three categories: volumetric static, gravimetric, and dynamic flow methods (isothermal or programmed temperature) [11,12].Several studies have been performed by static volumetric and the unit descriptions have been published by various authors [11,12]. The materials to be characterized can present a wide distribution of active centers, either in terms of acid\u2013base strength or metal particle sizes. These characteristics call for a technique that allows this analysis, thus meaning that the adsorbate gas must be added in small quantities in a controlled manner. Under these conditions, a technique such as the static volumetric method can guarantee low-pressure dosing of adsorbate gas, thus, could identify the different adsorption layers. The chemisorption isotherm is described as the variation in the amount of gas adsorbed as a function of pressure at equilibrium while maintaining the sample at a constant temperature. In a previous step, the surface of the sample must be cleaned with a vacuum; in many cases, pre-treatment with a cleaning gas current is preferable or even\u2014if the chemisorption is to be conducted on a metal catalyst\u2014reduction of the oxides so that it is in the form of a metal as this is the sensitive phase for the adsorption of gases such as H2, CO, etc. Chemisorption isotherms are expressed in terms of amount adsorbed at normal conditions (NTP) versus absolute pressure, rather than amount adsorbed versus relative pressure, as in the case of physical adsorption. The static volumetric technique generally produces an experimental adsorption isotherm similar to that shown in \nFig. 2, which involves a combination of physisorption, spillover and chemisorption. Hence, it is not a purely type-I isotherm with an adsorption plateau (constant amount adsorbed) as pressure increases. To differentiate the contribution of chemisorption from that of physisorption, the sample is evacuated after completion of the initial run, thus removing only reversibly adsorbed gas. The analysis is then repeated under the same conditions as the original analysis, except that during the second analysis, the active area of the sample is already saturated with chemisorbed molecules. Some authors have criticized the application of this second isotherm given that a greater amount of gas can be desorbed than the purely thermodynamic one; therefore, identical vacuum conditions to those used in the first isotherm and treatment time of up to 30\u2009min [16]. The adsorbed volume data for the first adsorption isotherm A are a combination of physical and chemical adsorption (reversible and irreversible, respectively). Isotherm B is the result of repeated analysis, where only reversible physisorption occurs. The isotherm represented by the dashed line C is generated mathematically by subtracting the adsorbed volume data for isotherm B from that for isotherm A. The result is the amount of active gas irreversibly absorbed by the sample.As in physisorption, the adsorption isotherm allows qualitative characterization of the material. For quantitative characterization, the volume of the monolayer (V\n\nm\n) chemisorbed on an active surface is determined. One way to determine this volume is by extending a line tangential to the plateau of the initial adsorption isotherm to the zero pressure axis. This is the procedure proposed in the ASTM D 3908\u201388 method to determine the amount of H2 adsorbed on a Pt catalyst supported on alumina previously reduced at 450\u2009\u00b0C and the adsorption capacity evaluated at 25\u2009\u00b0C [6]. The pressure range for the adjustment is between 100 and 300 torr. It has also been proposed to subtract the (reversible) physisorption isotherm from the combined isotherm as described above, and then extend a line tangential to the plateau of that isotherm to the zero pressure axis. Both methods should give approximately the same results, as long as the same analysis conditions are maintained. This value gives the amount adsorbed by weight of adsorbate. In the case of NH3 adsorption, the ASTM D 4824\u201393 method proposes the adsorbed volume as that obtained at a pressure of 150 torr and a temperature of 175\u2009\u00b0C [17] to minimize physisorption of ammonia. Additionally, repeated measurements at various temperatures can be used to calculate heats of adsorption (see next sections for details).In the case of the previous procedure, the surface of the solid to be analyzed must initially be free from any type of substance, that is, an initial heat treatment must be applied to clean the surface. Next, consecutive small amounts of adsorbate are added to allow the adsorption isotherm to be built, which means that a system that allows for the dosed volumes to be measured, as function of the increasing pressure. All this is performed under equilibrium conditions. Another possible procedure involves the solid sample being subjected to a stream of an inert gas to clean the surface and, subsequently, a known volume of the adsorbate gas being injected into this inert stream. This procedure has the following advantages: the measurements are fast compared to volumetric measurements; the weak bonds between adsorbent and adsorbate are not detected; the dead-volume need not to be measured; and the measurement can be easily tuned for small amount of samples. Thus, the adsorption centers can retain the adsorbate gas until the surface is completely covered, that is, until it is saturated (see \nFig. 3) [18]. If the adsorption isotherm is previously constructed, then, and, after calibration of the signal, the amount adsorbed will be obtained (V\n\nm\n, see Eq. 1). This case requires a system for detecting adsorbate gas in a gas stream, which is normally achieved by using a thermal conductivity detector (TCD).\n\n(1)\n\n\n\n\nV\n\n\nm\n\n\n=\n\n\n\n\n\u2211\n\n\ni\n=\n1\n\n\ni\n=\nn\n\n\n\n(\n\n\nh\n\n\nsaturation\n\n\n\u2212\n\n\nh\n\n\ninjected\n\n\n)\n\n\n\n\n\nh\n\n\nsaturation\n\n\n\n\n\u00b7\n\n\nV\n\n\ninjected\n\n\n\n\n\nwhere V\n\nm\n is the volume of the chemisorbed monolayer, expressed in cm3 at standard temperature and pressure (STP), V\n\ninjected\n corresponds to the loop volume previously calibrated and its volume is continuously monitored by the system for any temperature and pressure change in order to deliver a corrected number of moles at each injection, h\n\ninjected\n is the peak area corresponding to the injected volume. h\n\nsaturation\n corresponds to the injected volume that produce same peak area, and indicate saturation or end of the analysis is reached. Some practical advice can be suggested for this method: the relation between the amount of adsorbate gas injected and the sample mass should be adjusted to ensure at least one the injected dose to be completely adsorbed by the sample, the interval of time between the pulses should be constant and long enough to allow for the TCD signal to return to base line, and consecutive pulses should be injected until no increase of the signal area for consecutive pulses can be detected.In the two previous procedures, the working temperature remains constant. However, there is the possibility of repeating the analyses under other temperature conditions, which may allow additional information regarding the heat of adsorption to be obtained, or a method that enables the temperature to be increased to obtain information about the strength of adsorption to be used. This would be the case for the temperature-programmed desorption (TPD) procedure. In this case, upon sample saturation with a specific adsorbate, desorption can be carried under a specific ramping rate. If this analysis is repeated and desorbed at a different ramping rate, say (3, 5, 10, 15 and 20\u2009\u2103/min), thus would yield information about the strength of the adsorption centers (\nFig. 4). This is the case for the adsorption of bases such as NH3 or other amine molecules, as well as CO2\n[10]. As a gas stream that is in continuous contact with the solid is required, the detector used could also be a TCD. If no re-adsorption of gas takes place during desorption, and provided the molecules are adsorbed on a homogeneous surface without mutual interactions, the maximum temperature peak (T\n\nm\n) can be related to the activation energy of desorption (E\n\nd\n), see Equation 2\n[19]:\n\n(2)\n\n\n2\n\n\nln\nT\n\n\nm\n\n\n\n\u2212\n\nln\n\n\u03b2\n\n=\n\n\n\n\nE\n\n\nd\n\n\n\n\nR\n\u00b7\n\n\nT\n\n\nm\n\n\n\n\n+\nln\n\n\n\n\n\n\n\nE\n\n\nd\n\n\n\u00b7\n\n\nV\n\n\nm\n\n\n\n\nR\n\u00b7\n\n\nk\n\n\nd\n\n\n\n\n\n\n\n\n\n\nwhere \u03b2 is the rate of linear temperature increase, V\n\nm\n is the amount adsorbed at saturation, and k\n\nd\n is the pre-exponential factor in the expression for the desorption rate. If the kinetics of desorption are first order, it is possible to calculate E\n\nd\n. In the case of the presence of surface heterogeneities (large surface areas and microporosity), deviations could be found.The acidic or basic nature of the centers cannot be determined by this method, although it is possible to calculate the change in desorption activation energy with surface coating. If it is not possible to measure the amount of base adsorbed, or the amount that remains after desorption, the method can only give qualitative or semi-quantitative information (which can be obtained from the TPD profile).As a chemical bond forms between the adsorbate molecule and a specific center on the material surface, the number of sites can be determined by measuring the amount of chemisorbed gas. Although this may appear to be an easy and simple process, it should be noted that, depending on the nature of the metals and gases concerned and the operating conditions (temperature, pressure, measurement method), chemisorption could be partially reversible. The terms reversibility and irreversibility only have an operational meaning and are more important in the case of dynamic methods. In metallic catalysts, the active center is often a metal atom, with examples of this including nickel and platinum for the hydrogenation of unsaturated carbon-carbon bonds [1]. However, several important metal oxides and other non-metal catalysts must also be considered. As an example, we can cite the case of iron, to which other promoters are added to favor the synthesis of ammonia. The metallic atoms are found forming islands or clusters, rather than being distributed individually, on an inert porous material that acts as a support and favors their dispersion and stability. In several cases, this situation is not clear. The size of these islands and clusters depends on the nature of the metal and the support, as well as the method used to deposit it (preparation method). In such a case, the exposed active centers can be determined by the gas adsorption method. For supported metal oxides, the same gases used in selective chemisorption on metals (H2, CO, O2 and N2O) are not compatible, since they adsorb weakly on these surfaces: CO is only weakly adsorbed on metal oxides, and all exposed surface sites cannot be evaluated; H2 adsorption involves reaction with the surface and subsurface lattice oxygen; and O2/N2O are not adsorbed on oxidized surfaces [8]. Adsorption of H2 and O2 at sub-ambient temperatures has been attempted to avoid the participation of subsurface lattice oxygen and lattice oxygen vacancies, respectively, but was unsuccessful in avoiding the participation of these species [20]. However, small alcohols are adsorbed on dehydrated and/or evacuated oxides and allow the number of active surface sites to be quantitatively and selectively assessed. Thus, methanol is a highly reactive molecule that has been reported to be chemically adsorbed on oxides and allows quantitative determination of the number of surface active (Ns) centers. It has been observed that methanol follows several routes of chemisorption in oxides [21], depending on the nature of the metal oxide, and some of these reactions can occur and allow the quantification of adsorption centers:\n\n\n\nC\n\n\nH\n\n\n3\n\n\nOH\n+\nM\n\u2212\nOH\n\u2192\nM\n\u2212\nOC\n\n\nH\n\n\n3\n\n\n+\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\n(M is a metal cation site)\n\n\n\nC\n\n\nH\n\n\n3\n\n\nOH\n+\nM\n\u2212\nO\n\u2212\nS\n\u2192\nM\n\u2212\nOC\n\n\nH\n\n\n3\n\n\n+\nS\n\u2212\nOH\n\n\n\n\n(S is the oxide support cation site) by breaking open hetero-bonds\n\n\n\nC\n\n\nH\n\n\n3\n\n\nOH\n+\n\n\n\n*\n\n\n\n\u2192\nC\n\n\nH\n\n\n3\n\n\nOH\n\u2212\n\n\n\n*\n\n\n\n\n\n\n([*] is a coordinatively unsaturated Lewis acid site).It has been reported that the typical number of active surface sites on oxides is about 0.7\u2009\u00d7\u20091015 sites/cm2, which is about half the value for metals (1.2\u2009\u00d71015 sites/cm2), because the surface density of sites on oxides is less than on metals. The number of active surface sites on MoO3, V2O5 and ZnO is significantly lower (0.1\u2009\u00d71015/cm2) due to the presence of much less active exposed surface planes due to the presence of coordinatively saturated sites [22].\nStoichiometry: Knowing the relationship (stoichiometry) between an exposed metal atom and an adsorbate gas molecule is an important factor in this type of determination as many polyatomic gas molecules do not adsorb to a single active site. This is the case, for example, for the hydrogen molecule (H2). It has been reported that hydrogen adsorbs dissociatively, that is, it separates into two atoms, each of which reacts with a single metal atom. Thus, a gas molecule has bound to two metal atoms (this is the case of Pt, Pd, Rh, Ru, Ir and Ni). As such, the stoichiometry is said to be two (2) for this surface reaction. Similarly, a molecule of adsorbate gas could associate with more than one metal atom without dissociating. This is the case for carbon monoxide (CO), which is normally expected to bind in a one-to-one ratio (Me-CO) but could form a bridge between two metal atoms (Me-(CO)-Me). This situation would also result in a stoichiometry of two. Cases in which an excess of adsorption would result in a stoichiometry of less than one are not implausible. This is the case for the formation of hydrides (for hydrogen) and carbonyls (for carbon monoxide). These latter situations should be controlled and avoided in whatever way possible. In the case of O2 the O-Metal stoichiometry is 1.0, although the possible formation of metal oxides and bulk metal oxides may modify this relation.The value of the stoichiometric factor X\n\nm\n can be determined, for example, by chemisorption measurements using metal powders with known specific surface areas. In general, the number of atoms per unit area for polycrystalline metal surfaces is not known. For hydrogen chemisorption up to full coverage, X\n\nm\n, the average number of surface metal atoms associated with the adsorption of an adsorbed hydrogen molecule is assumed to be 2. However, some uncertainty also exists in this regard. Fundamental studies on hydrogen chemisorption on Ni yield solid evidence that strongly chemisorbed hydrogen atoms are attached to, or just below, so-called C8 sites, which are the holes formed by a cluster of three densely packed Ni atoms above an octahedral interstice. The number of C8 sites is equal to the number of Ni atoms in the (111) plane and thus, for the (111) plane of free metals, X\n\nm\n =\u20092 is a realistic choice. The total adsorbate uptake, n\n\nm\n, is also subject to uncertainties. In many group VIII metals (Ni, Pt, Pd, Ru, Rh), the H2 chemisorption isotherm has the form shown in Fig. 2. The highest amount of hydrogen is adsorbed (strong chemisorption) at a pressure of less than 133.22\u2009Pa (1 torr). Above that pressure, weakly chemisorbed hydrogen adsorption occurs, mostly of the order of 20\u201325 % of the strongly retained monolayer. The difficulty is that the transition pressure between strongly chemisorbed hydrogen and weakly chemisorbed hydrogen is not clearly defined. As such, the X\n\nm\n value of 2 refers to strongly chemisorbed hydrogen only.Several factors affect the accuracy of chemisorption methods. These include factors associated with the stoichiometric factor, the crystallographic heterogeneity of the surface, the presence of a support that theoretically does not chemisorb, the possible absorption or dissolution of the adsorbate gas in the metal, reconstruction of the surface atoms during the process of chemisorption, as well as contaminants adsorbed on the surface. The stoichiometric factor is usually not a problem when H2 is used as the adsorbate, since it generally dissociates by adsorbing on catalytically important transition metals and chemisorbs with a stoichiometric factor of 2 (based on the H2 molecule). For the other gases mentioned, obtaining an exact and constant stoichiometric factor may be difficult as adsorption of such molecules will highly depend on the surface of the adsorbent. Thus, for example, in the adsorption of CO on Pd/SiO2, if analyzed by IR, two adsorption bands that correspond to the Pd-(CO)-Pd bridge and to the linear form PdCO are observed. For particle sizes less than 10\u2009nm, the geometry of the surface and, therefore, the stoichiometric factor depend on the size of the particle. Thus, in supported Pt catalysts, for small particles, the stoichiometric factor for CO adsorption can vary between 1 and 2. For metal particles larger than 10\u2009nm, this effect disappears, and it can be considered constant.From an experimental point of view, the amount of gas adsorbed is measured. Therefore, it is essential to establish the stoichiometry involved, knowing the nature of the adsorbate gas and the active site. This information can be obtained from the literature on catalysts or by direct measurement (see \nTable 1) [1].\nMonolayer coverage: Once the amount of gas adsorbed by the sample (the adsorption isotherm) has been determined, the number of active centers can be calculated from the capacity of the monolayer, V\n\nm\n. A number of graphical and numerical methods can be applied for that purpose, and the most widely used are described below. In the case of the volumetric dynamic procedure, the adsorbed volume (V\n\nm\n) would be obtained directly (see Eq. 1).\nExtrapolation. This method involves plotting points on the adsorption isotherm until the plateau is reached (e.g, ASTM method D 3908\u201388 for H2 adsorption on Pt/Al2O3 catalyst) [6]. In this region, the surface has become saturated with the adsorbate and monolayer formation has been ensured. If the pressure and the amount of gas dosed are increased, only additional physical adsorption occurs. The contribution of this physisorption can be explained by assuming that it is zero at zero pressure. If the line joining the points of the plateau is extrapolated to the value of zero pressure (intercept with the OY axis, the value of the monolayer is obtained. This value of V\n\nm\n represents the total amount of chemisorbed gas irrespective of the exact nature of the bonding type (strong or weak; see Fig. 2).\nIrreversible isotherm: Some applications require that only strong chemisorption centers be determined and physisorption or weaker chemisorption centers excluded. In these cases, it is necessary to obtain a second adsorption isotherm. After acquisition of the first isotherm, the sample is evacuated at the analysis temperature to desorb loosely bound gas molecules. Strongly adsorbed molecules remain bound to active centers on the sample surface. A second adsorption analysis is repeated to produce a second isotherm that would provide information on weak chemisorption and physisorption and is obtained in the same way as the first. The difference between the two isotherms at any given pressure represents the amount of chemisorbed gas. Alternatively, the plateau of the irreversible isotherm can be extrapolated to zero pressure to determine V\n\nm\n graphically (see Fig. 2).The above methods try to describe a simple (or pure) chemisorption process, although in some cases the interference of the catalytic support can be considered as it may have its own adsorption centers that can interfere with the process. This may be the case, for example, for the so-called spillover process in which the hydrogen that is dissociatively adsorbed on the metal (normally Pt) migrates to the surface and the bulk of the support. In cases where there is spillover (or at least there may be), two isotherms must be measured to determine the adsorption capacity: one for the supported metallic catalyst and the other for the support only (normally is known as blank), without the active metallic phase. The first isotherm yields adsorption data consisting of strong chemisorption at the active sites, weaker chemisorption, physisorption at the active sites and on the exposed support surface, plus active site spillover. The second isotherm simply consists of physisorption on the support. The net amount of chemisorption, including spillover, can be easily calculated by subtracting the second data set from the first.If the stoichiometric factor of chemisorption is known, it is possible to calculate the accessible number of surface atoms (N\n\nS\n) of the component (generally metal) from the amount of adsorbed gas using Eq. 3:\n\n(3)\n\n\n\n\nN\n\n\ns\n\n\n=\n\n\n\n\nV\n\n\nm\n\n\n\u00b7\n\n\nN\n\n\nA\n\n\n\u00b7\n\n\nX\n\n\nm\n\n\n\n\n\n\nV\n\n\nmol\n\n\n\n\n\n\n\nwhere V\n\nm\n is the volume of the chemisorbed monolayer, expressed in cm3 at standard temperature and pressure (STP); V\n\nmol\n is the molar volume of adsorbate (22414\u2009cm3 occupied by one mol of gas at STP); N\n\nA\n is Avogadro\u2019s number (6.022\u2009\u00d71023); and X\n\nm\n is the average stoichiometric factor. X\n\nm\n indicates the number of surface atoms of the component that are covered by an adsorbate molecule after chemisorption.In many cases, the small metallic crystallites are firmly attached to the support via chemical bonds. As a result, the distribution of the crystallographic planes on the surface is, in most cases, different to the equilibrium distribution that would correspond to a free particle. Therefore, the value of N\n\nS\n is strongly affected by support/particle interactions. The presence of the SMSI (strong metal/support interaction) effect can even completely suppress any form of hydrogen chemisorption. In this case, there would be no metallic species on the surface sensitive to chemisorption and therefore this cannot be evaluated.The specific metallic surface area, A\n\nm\n, is determined as the product of the number of exposed metal atoms, N\n\nS\n, by the cross-sectional area of each atom (see \nTable 2), A\n\nX\n, and per unit mass, W (see Eq. 4):\n\n(4)\n\n\n\n\nA\n\n\nm\n\n\n=\n\n\n\n\nN\n\n\ns\n\n\n\u00b7\n\n\nA\n\n\nX\n\n\n\n\nW\n\n\n\n\n\n\nIt can also be expressed per gram of metal in the catalyst if the experimentally determined metal content (%) is included (see Eq. 5).\n\n(5)\n\n\n\n\nA\n\n\nm\n\n\n=\n\n\n\n\nN\n\n\ns\n\n\n\u00b7\n\n\nA\n\n\nX\n\n\n\n\nW\n\u00b7\n\n\n%\nmetal\n\n\n100\n\n\n\n\n\n\n\n\nAnother factor to consider when calculating the metallic surface (m2 of metal/g) from chemisorption measurements is a lack of information about the heterogeneity of the crystallographic surface of the dispersed metal particles. In such a case, the number of accessible metal atoms on the surface can be calculated using Eq. 3. However, the calculation of the metallic surface requires information about the number of atoms per unit surface. This value is clearly defined in the ideal plane of a single crystal, but not for the case of metallic particles with surfaces exposing several crystallographic planes. To avoid this difficulty, the three most prominent planes\u2014(111), (100) and (110) for cubic face-centered and (110), (100) and (211) for cubic body-centered\u2014are generally considered to be present in equal numbers. As such, the number of atoms per m2 of surface for face-centered metals (Ni, Pd) is 1.91\u2009\u00d7\u20091018/\na\n\n2 and (Fe, W) is 1.35\u2009\u00d7\u20091018/\na\n\n2 for body-centered metals, where \na\n is the lattice constant. The specific metal surface of a supported metal catalyst can be calculated using Eqs. 4 or 5, where N\n\ns\n is the number of accessible atoms on the metal surface per gram of catalyst.In the case of supported metal catalysts, it is important to know what fraction of the active metal atoms is exposed and available to catalyze a reaction. This is a surface phenomenon as the atoms inside the metal particles do not participate in surface reactions. Hence, these atoms must be dispersed as widely as possible. Dispersion is defined as the percentage of all metal atoms in the sample that are exposed at the surface. As the total amount of metal in the sample can be determined by chemical analysis of the sample, if the weight of the metal in the catalyst is known, the degree of dispersion D(%), that is, the ratio of atoms on the surface (N\n\nS\n) with respect to the total number of atoms (N\n\nT\n, atoms on the surface and in volume), of the metal can be calculated (see Eq. 6).\n\n(6)\n\n\nD\n\n\n\n%\n\n\n\n=\n\n\n\n\nN\n\n\ns\n\n\n\n\n\n\nN\n\n\nT\n\n\n\n\n=\n\n\nV\n\n\nm\n\n\n\u00b7\n\n\n\n\nX\n\n\nm\n\n\n\u00b7\n\n\nM\n\n\n\u00e1\ntomoMetal\n\n\n\n\n\n\nV\n\n\nmol\n\n\n\u00b7\n\n\n%\n\n\nmetal\n\n\n\n\n100\n\n\n\n\nLogically, if gas-adsorption techniques are used, the atoms on the surface will be those that can be evaluated by chemisorption, and it is precisely those atoms that can participate in gas-solid reactions. This property is important since it can affect both the selectivity and catalytic performance in supported metal catalysts.If both the mass of metal in the catalyst and its density are known, the volume of metal can be estimated. If the metallic surface area (A\n\nm\n) is already known, the equivalent particle diameter, d, can be estimated by assuming a shape factor for the particle (see Eq. 7).\n\n(7)\n\n\nd\n=\n\n\n6\n\n\n\n\nA\n\n\nm\n\n\n\u00b7\n\n\n\u03c1\n\n\nmetal\n\n\n\n\n\u00b7\n(\n%\n\nr\n\neduction\n)\n\n\n\n\nThis diameter is assumed to correspond to a hemisphere in contact with the surface of the catalytic support. The geometric factor (in this case 6) is identical if it is a totally spherical geometry. These two geometries have been reported as the most frequent for supported metal catalysts.Metal\u2013support interactions and metal particle shape play an important role in determining particle size by gas chemisorption. A hemispherical shape is usually assumed, but can give misleading results of up to one order of magnitude. In such a case, the metal particle sizes are underestimated when the metal strongly interacts with the support and overestimated when there is a weak metal\u2013support interaction. The assumption of spherical shapes always underestimates the size of the particles, with this error being considerably smaller with regular geometries than that associated with the effect of the metal\u2013support interaction due to its effect on the shape of the particle. Therefore, some authors have introduced a particle\u2013support interaction factor when determining particle size by chemisorption.As indicated in the Introduction, clusters and particles have unique chemical and physical properties that depend largely on their size. In the case of heterogeneous catalysis, a relationship between the size of the metal particle and its performance and selectivity for multiple systems is acknowledged, and the particle size can even determine whether or not a system is active.High-resolution transmission electron microscopy (TEM) provides qualitative and semi-quantitative information on the size and shape distribution of metal particles, as well as their dispersion in the support. In this technique, the contrast depends on the ratio of the atomic numbers of the metal and the support, with small particles having a lower contrast than large ones. Particles with diameters smaller than 1\u20131.5\u2009nm are considerably more difficult to detect, thereby limiting accurate quantification of the particle-size distribution. Although these instrumental limitations have been resolved in recent years to be able to quantify particle sizes on the sub-nanometric scale, this technique is still not commonly used because of its low availably. It is also possible to obtain information using X-ray diffraction (XRD), in this case regarding the crystal size from the broadening of the diffraction line. As in the case of electron microscopy, limitations appear for the smallest particles and for those that do not exhibit crystallinity. Gas chemisorption, typically using H2 and CO as probe molecules, is widely used in combination with TEM and XRD to quantify the particle-size distribution, or alone to estimate the metallic surface area accessible to the molecule probe. As reported previously, this technique consists of measuring the number of probe molecules adsorbed on the metallic surface of a material. Knowledge of the stoichiometric factor for the number of adsorbed probe molecules per metal surface atom allows the metal surface area, mean particle size and metal dispersion to be calculated. It is widely accepted that one of the main limitations of gas chemisorption as a particle-size determination technique is the precise determination of the aforementioned stoichiometric factor, which largely depends on the arrangement of the surface atoms. Indeed, the probe molecule can form linear, double or triple adduct bridges, therefore its value ranges between 0.5 and 2 for a given metal. It has been reported that the effect of the interaction of the metal and a support (the contact angle between the two) on the determination of the resulting average particle size may be greater than the effect of the stoichiometric factor due to the conventional assumption of the hemispherical shape of the particle.A well-accepted fact in the field of heterogeneous catalysis is that the method of metal deposition affects not only the resulting particle size and distribution, but also the metal\u2013support interaction. For example, the deposition-precipitation method generally produces hemispherical metal particles in which the flat planes of the metal are attached to the support, while impregnation methods produce spherical particles with very weak interactions with the support. The type of metal\u2013support interaction (strong or weak) can have a key effect on the catalytic behavior. It has also been possible to demonstrate, by means of high-angle annular dark field (HAADF) images taken in a STEM, that when interaction with the support is very strong, the morphology of the particles can be more similar to two-dimensional plates rather than three-dimensional particles. Thus, the conventional assumption of metal particles with a hemispherical geometry for the calculation of average metal sizes by gas adsorption characterization can give misleading results if the metal particle is not hemispherical in shape. In fact, the metal\u2013support interaction and, consequently, the resulting metal\u2013support contact angle must be taken into account for an accurate estimate of the mean metal size. Particle sizes are slightly overestimated when their contact angle is >\u200990\u2009\u00b0 (low interaction with the support); however, particle sizes are greatly underestimated when their contact angle is <\u200990\u00b0 (high interaction with the support) (see \nFig. 5) [23].When selecting the adsorbent gas to be used when using chemisorption measurements as part of the experimental method, it should be taken into account that the stoichiometric relationship that allows the quantity of metallic atoms on the surface to be determined should be known, thereby preventing the support from being able to adsorb or interact with the adsorbent gas [3,5]. Therefore, an initial study, including the operating conditions, is required for each metal to be analyzed to determine the most suitable conditions and adsorbents in order to determine the metallic atoms on the surface.\nPt catalysts: supported platinum catalysts, and how their dispersion is measured, are perhaps the most widely studied systems due to their widespread applications. The adsorption of hydrogen on Pt has been studied by several authors, who found that it is dissociative, that is, the H2 molecule breaks and each atom binds to a different Pt atom. Adsorption is normally carried out at temperatures of between 0 and 35\u2009\u00b0C. To try to clarify how H2 is retained on the surface of Pt catalysts, these authors have conducted studies of hydrogen desorption at programmed temperature and found the presence of up to four states: a) hydrogen weakly adsorbed in a non-dissociative manner (\u201373\u2009\u00b0C); b) hydrogen atoms adsorbed on the surface Pt atoms (130\u2009\u00b0C); c) reversibly adsorbed hydrogen (180\u2009\u00b0C); d) hydrogen spillover (480\u2009\u00b0C; see \nFig. 6) [11]. Of these states, it appears that option b) may have the highest possibility of being related to chemisorption. The possible contributions of the other states would cause errors in the determination of the amounts adsorbed. The stoichiometry accepted by most authors working with Pt catalysts is H2:Pt =\u20091:2, although deviations from this stoichiometry may exist in the case of highly dispersed catalysts.CO chemisorption has also been used in the characterization of Pt catalysts [24\u201326]. The main problems in this case are: a) the possibility of CO chemisorption in a linear (Pt-CO) or bridged (Pt-CO-Pt) form and, b) the possibility of formation of volatile carbonyls, and even other forms of triple bonds and dissociated molecules have been described [27\u201329]. The fact that one form or another predominates can cause the stoichiometry to be 1 or 2. The problem worsens because the relative proportion of these two forms depends on the particle size (the linear form predominates in high dispersions and the dotted form for particle sizes above 5\u2009nm [30]). In general, it is considered that the two forms predominate, therefore a CO:Pt =\u20091:1.15 ratio is normally used. This situation is more common in the case of metallic catalysts containing Ni, Co, Ru, Mo, W, etc.If the results obtained upon the adsorption of CO and H2 on Pt are compared, the additional H2 consumption observed can be explained by a spillover effect, which increases at high dispersions in which the metal\u2013support interfaces increase. There may also be differences between the two measurements if the Pt is not fully reduced, in which case CO is adsorbed rather than H2.One alternative that has been proposed to increase the sensitivity to H2 adsorption is H2-O2 titration reactions [31]. This method was proposed based on the chemisorption of H2 and O2 on Pt atoms on the surface, as well as on the reaction of H2 with oxygen chemisorbed on Pt, and on the reaction of O2 with hydrogen chemisorbed on Pt [32,33]. All these reactions are carried out at room temperature:\n\n\n\n\nPt\n\n+\n\n\n1\n\n\n2\n\n\n\n\nH\n\n\n2\n\n\n\u2192\n\nPt\n\n\u2212\nH\n,\n\nhydrogen\n\n\n\nchemisorption\n\n\n\n(\n\nHC\n\n)\n\n\n\n\n\n\n\n\n\n\nPt\n\n+\n\n\n1\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\u2192\n\nPt\n\n\u2212\nO\n,\n\noxygen\n\n\n\nchemisorption\n\n\n\n(\n\nOC\n\n)\n\n\n\n\n\n\n\n\n\n\nPt\n\n\u2212\nO\n+\n\n\n3\n\n\n2\n\n\n\n\nH\n\n\n2\n\n\n\u2192\n\nPt\n\n\u2212\nH\n+\n\n\nH\n\n\n2\n\n\nO\n,\n\n\nhydrogen\n\n\n\ntitration\n\n\n\nof\n\n\n\noxygen\n\n\n\ncovered\n\n\n\nsurface\n\n\u2062\n\n\n(\n\nHT\n\n)\n\n\n\n\n\n\n\n\n\n2\n\nPt\n\n\u2212\nH\n+\n\n\n3\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\u2192\n2\n\nPt\n\n\u2212\nO\n+\n\n\nH\n\n\n2\n\n\nO\n,\n\noxygen\n\n\n\ntitration\n\n\n\nof\n\n\n\nhydrogen\n\n\n\ncovered\n\n\n\nsurface\n\n\n\n(\n\nOT\n\n)\n\n\n\n\n\nAn HC:OC:HT:OT stoichiometry of 1:1:3:3 was initially proposed and the sensitivity of H2-O2 titration was found to be three times greater than for direct H2 or O2 chemisorption. More recently, some authors indicated that the results of the titration depend on pretreatment of the catalyst and on the titration procedure [34,35].\nPd catalysts: As in the case of Pt catalysts, CO adsorption can be used to characterize the metal surface of these catalysts [24\u201326]. The linear bond usually predominates, although it is necessary to control the conditions to ensure that this is the case [36\u201338]. Nevertheless, an average value of close to 2 was found for any support with dispersed Pd, although the measurements were performed using a pulse flow technique [39]. If hydrogen chemisorption is used to measure Pd dispersion, hydrogen absorption must be avoided. Thus, for example, exposure of supported Pd to a hydrogen atmosphere at room temperature results in the formation of \u03b2-Pd-Hx, where x decreases as Pd dispersion increases [40,41]. Starting from a 30 % dispersion, and heating above 70\u2009\u00b0C, the absorption of hydrogen decreases considerably. Despite the absorption of hydrogen, the H:Pd ratio is considered to be 1:1 [42\u201345].In catalysts of this family, the H2/O2 (or O2/H2) titration sequence has also been used as this technique has the main advantage that it allows the amount of adsorbed gas to be increased in catalysts with low dispersion. The reactions in this case would be [35]:\n\n\n\n2\nPd\n\u2212\nH\n+\n1.5\n\n\nO\n\n\n2\n\n\n\u2192\n2\nPd\n\u2212\nO\n+\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\n\n\n\n\nPd\n\u2212\nO\n+\n1.5\n\n\nH\n\n\n2\n\n\n\u2192\nPd\n\u2212\nH\n+\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\n\nRh catalysts: in catalysts of this type, it has been reported that the CO:Rh stoichiometry can be 2:1, 1:1 and 1:2 [46], and even 1:3. In the case of supported Rh catalysts, 1:1 or 1:2 is proposed. If H2 adsorption is used, it has been proposed that there is a 1:1 ratio, which is confirmed for low dispersions. The stoichiometry is 2:1 in the case of high dispersions [47\u201351].\nNi catalysts: the first drawback that can occur in this family of catalysts is that Ni is not completely reduced to the metal [52]. Although it can be assumed for the previous catalysts that all the metal is reduced, in the case of Ni catalysts this may not be the case [53]. A non-reduced phase can be found between the support and the reduced metal particle, therefore this effect should be taken into account when calculating the metal dispersion [54]. In the case of some supports, such as alumina, this phase can be incorporated into the support by the formation of a spinel [55].The formation of up to four Ni(CO)x complexes has been described, with the stoichiometry depending on the degree of dispersion and the adsorption temperature; therefore, the use of CO is not recommended when characterizing Ni catalysts [53]. The best method to characterize Ni-containing catalysts is the chemisorption of H2 at temperatures of between 0 and 35\u2009\u00b0C and at a pressure of up to 10\u201320 kPa. The stoichiometric factor in this case is 2 [53,56,57].\nCu catalysts: for this type of catalyst, H2 chemisorption is not a good option due to its low sensitivity at low temperature. CO chemisorption cannot be used either, since it can be confused with physical adsorption. Alternatively, the adsorption of O2 at \u2212136\u2009\u00b0C has been proposed. Under these conditions, the process is not activated and the stoichiometric factor is 4. However, the main drawback involves reaching the adsorption temperature. As an alternative, the adsorptive decomposition of N2O at 90\u2009\u00b0C is proposed:\n\n\n\n\n\nN\n\n\n2\n\n\nO\n\n(\n\ngas\n\n)\n\n+\n2\n\nCu\n\n\u2192\n\n\nCu\n\n\n2\n\n\nO\n+\n\n\nN\n\n\n2\n\n\n(\n\ngas\n\n)\n+\n\n(\n\nE\nx\nc\ne\ns\ns\n\n\no\nf\n\n\n\n\nN\n\n\n2\n\n\nO\n\n\nt\nh\na\nt\n\u2062\n\n\nh\na\ns\n\u2062\n\n\nt\no\n\u2062\n\n\nb\ne\n\u2062\n\n\nt\nr\na\np\np\ne\nd\n)\n\n\n\n\nAs the pressure remains constant during the process, nitrogen can be measured by assuming one N2 molecule per two Cu atoms on the surface [58,59]. This method is also proposed for Ag and Ru. Although this method is rather difficult to be determined by the dynamic technique due to the fact that the TCD is not capable to differentiate between the peak of N2 produced by the surface oxidation of Cu by N2O and the excess of N2O that does not react, Alternatively, a cold trap at \u2212\u200980\u2009\u2103 is recommended to trap the excess of N2O before reaching the TCD and allows the N2 peak to pass on. Another useful alternative is to adapt a separation column that enables the separation of the N2 peak from those corresponding to N2O, thus the N2 peak will arrive and be detected by the TCD before the delayed peak of N2O reaches the TCD. In this case, the method becomes available to properly compute the amount of N2 and to be related to the amount of Cu on the surface. This phenomenon of adversity can be easily resolved if a mass spectrometer is connected at the exhaust of the instrument. Example of this analysis is shown in \nFig. 7.\nBimetallic catalysts: the presence of two metals makes it more complex to characterize the superficial metallic centers, specifically to know the stoichiometric relationship between the adsorbate gas and the metal. In these cases it is normally necessary to use other characterization techniques such as DRX or TEM. The simplest case for the use of chemisorption in bimetallic systems is when only one of the system components chemisorbs the adsorbate gas. For example, Ru-Cu and Os-Cu systems can be analyzed, since copper atoms do not adsorb hydrogen [60]. In Pt(Re,Ir,Ru) systems, selective chemisorption is performed by means of O2/H2 titration as it allows Pt and Re on the surface to be determined. The chemisorbed oxygen in Pt can be reduced by hydrogen at 25\u2009\u00b0C, and a second titration with oxygen allows the Re atoms to be estimated by difference. This procedure can be used if the formation of alloys between metals does not occur. In the case of the Pt-Ru system, a titration using O2 and CO is used following the same previous strategy [61].In the case of acid centers, the nature of the surface must be taken into account, as well as the strength and number of centers [62]. First of all, it should be possible to differentiate between Br\u00f6nsted- and Lewis-type acidity. In the former, a proton is brought into play as a Br\u00f6nsted acid center is one capable of transferring a proton from the solid surface to an adsorbed molecule. This type of acidity can be generated when a trivalent ion is present in tetrahedral coordination with oxygen, with the most common example being aluminum [63]. When all the tetrahedral oxoanions are shared with two cations, a negative charge is created on cations with a charge of less than 4. This is the case, for example, in aluminosilicates [1]:\n\nTable\n\n\n\n\n\n\n\nImage 1\n\n\n\n\n\n\n\nWhen the excess of negative charge is compensated with protons, silanol groups are formed, which can be presented as:\n\nTable\n\n\n\n\n\n\n\nImage 2\n\n\n\n\n\n\n\nThis is also a Br\u00f6nsted center. In this case, the oxygen does not have a trigonal structure and is only represented as such to indicate that both Si and Al retain their tetrahedral coordination. This center is best detected by treatment with a basic molecule (e.g., an olefin) and subsequently observing the equilibrium:\n\nTable\n\n\n\n\n\n\n\nImage 3\n\n\n\n\n\n\n\nDepending on the strength of the Br\u00f6nsted center, this balance can be displaced. The acidic surface is therefore dynamic and depends on both the chemical nature of the adsorbed base and the solid.In Lewis-type acidity, the surface accepts an electron pair from the adsorbed molecule, forming a coordinate bond. In the case of silica-alumina, this could be represented as [1]:\n\nTable\n\n\n\n\n\n\n\nImage 4\n\n\n\n\n\n\n\nIn the particular case of clays in which a dehydration point has not been reached and the exchange centers are occupied by cations such as Na+, Ca2+, Mg2+, etc., the main Lewis centers are due to Fe(III) in the structure and the octahedral Al(IV) located on the edges of the particles [64]. Interactions between Br\u00f6nsted and Lewis centers may also occur. Thus, for example, in a clay at 300\u2009\u00b0C, the structural OH begins to be eliminated, forming trigonal Al(III) and H2O. Dehydration processes accompany the formation of Lewis centers. Synergistic interactions between the Br\u00f6nsted and Lewis centers may also occur. For example, an electron-deficient Al(III) (when in tetrahedral coordination) exerts an inducing effect on a neighboring silanol group, thus favoring H+ mobility.Depending on the nature of the surface, all materials, have an acid type and strength. The most representative acidic materials include alumina, silica-alumina, and zeolites, amongst others [65]. However, given the importance of this property, a series of solids known as superacids have been developed in recent years [66]. Treatment of activated carbons with acids (H2SO4, HNO3) or other oxidants (H2O2, Cr3O7\n2-, MnO4\n-) also creates acidic surface groups. Given the hydrophobic nature of the surface, these new centers will increase their hydrophilic character. The groups that have been proposed to exist on the surface of oxidized carbons are shown in \nFig. 8\n[67].A solid of an acidic nature will not usually have a single class of acidity and will normally present a large distribution of acid centers. This may be due to a heterogeneity in the composition of the solid or the existence of a small range of interactions or surface structures. Both Br\u00f6nsted and Lewis centers are often found in the solid at the same time. As such, it will be necessary to use methods that allows to differentiate and characterize a surface in terms of the nature, number, and strength of acid centers.The titration of acid/basic centers using dynamic methods is carried out by injecting pulses of ammonia/carbon dioxide into a gas stream that allows the ammonia/carbon dioxide to pass through the adsorbent or catalyst bed at atmospheric pressure. This procedure has already been described above and allows the amount adsorbed, which is related to the capacity of the monolayer, to be determined. Similarly, it is possible to work in TPD mode. This method is the most commonly used in solid acid catalyst due to its simplicity and low cost and its ability to determine both the number of acid sites and their strength [68,69]. However, the use of ammonia presents limitations. Thus, ammonia is a small molecule that is able to penetrate the smallest pores of the material. However, these very small pores make a very small contribution to the catalytic behavior, therefore their contribution to the acidity of the material can be neglected. Additionally, it should be noted that ammonia is a base that can react with relatively weak acid centers and does not contribute decisively to the overall catalytic behavior. Typically, ammonia-TPD curves show two peaks (see Fig. 4), which may be related to the existence of at least two types of acid sites. The first peak (A) is related to desorption of weakly bound ammonia and was found to be of no catalytic relevance (it is relevant for gas-sensing applications); the other peak (B) reflects the desorption of ammonia probably from the Br\u00f6nsted acid sites, which determines, for example, the acidic properties of zeolites. A classification related to weak, medium and strong acidic sites, related to the temperature of desorption peaks centered in the ranges 25\u2013200, 200\u2013400 and over 400\u2009\u00b0C is also proposed [68], although there are currently no standardized criteria. Other larger molecules, such as pyridine or tert-butyl amine, isopropylamine, etc, are preferred because they only penetrate the largest pores and these are the ones that contribute most to the catalytic behavior observed. However, these molecules present operational problems since they can condense under operating conditions. The TPD of amines has recently been reported as technique for measuring Br\u00f6nsted acid site concentrations. This method is based on the formation of alkylammonium ions from the adsorbed alkyl amines that are protonated by Br\u00f6nsted sites, which decompose into ammonia and olefins in a range of temperatures. Typically, amine-TPD curves show two peaks (see also Fig. 4). The first peak is related to desorption of weakly bound amine and the other peak reflects the decomposition of amines at the Br\u00f6nsted acid sites. In the case of isopropyl amine, propylene and ammonia would be obtained. The CO2-TPD method allows analysis of the nature of basic sites. As in the case of ammonia, the strength of basic sites may be classified according to their different CO2 desorption temperatures. In this case, the temperatures of desorption peaks below 400\u2009\u00b0C, between 400 and 600\u2009\u00b0C, and over 600\u2009\u00b0C are related to weakly, medium and strongly basic sites.\nAdsorption of the probe molecule and analysis by IR spectroscopy. There are numerous studies on the interaction of surfaces and basic molecules such as pyridine by IR spectroscopy [70,71]. Pyridine (C5H5N) is the preferred molecule to study Br\u00f6nsted and Lewis acidity separately as these interactions can be easily distinguished from the IR spectra [72\u201374]:\n\nTable\n\n\n\n\n\n\n\nImage 5\n\n\n\n\n\n\n\nOther proposed molecular probes include acetonitrile (CH3CN), benzonitrile (C6H5CN), CO, H2 and NO. Direct measurement of the intensity of the frequencies of the OH groups does not provide information on the acid strength of the Br\u00f6nsted centers and shifts in the frequencies of these vibrations by interaction via hydrogen bonds with adsorbed molecules provide more information. This interaction can be quantified as [75]:\n\n(8)\n\n\n\u2206\n\u03b3\n=\n\n\n3\nqE\n\n\n4\nr\n\n\n(\n2\n\u03bc\n)\n\n\n1\n/\n2\n\n\n\n\nD\n\n\n1\n/\n2\n\n\n\n\n\n\n\nwhere \u2206\u03b3 is the frequency shift of the hydroxyl group involved in the hydrogen bond interaction, q is the dipole charge, E is the electric field across the O-H axis, \u03bc is the reduced mass, and D is the dissociation energy of the O-H bond. The values of \u2206\u03b3 can be estimated, thus giving the Br\u00f6nsted-type acid strength. The strength of the acid centers can also be studied from the evolution of these bands under different conditions of temperature and vacuum.Pyridine is the most widely used molecule, since it is a weak Br\u00f6nsted base (pkb = 9) that only interacts with the strong protonic centers, that is, with the interesting ones from a catalytic point of view. The absorption bands of adsorbed pyridine are fine and allow Br\u00f6nsted centers to be distinguished from Lewis centers. Information can be obtained on:\n\n\u2013\nThe types of acid centers, as identified by the characteristic absorption frequencies.\n\n\n\u2013\nTheir strength, from the variations in intensity of the bands upon desorption at increasing temperature.\n\n\n\u2013\nThe reactivity of the OH groups with respect to the base, as seen from both the variation of intensities of the absorption maxima \u03c5(OH) during adsorption and desorption, and by the positions of the reagent bands.\n\n\n\u2013\nThe density of acid centers, from a plot of normalized absorbance (IR) against the amount of pyridine adsorbed.\n\n\nThe types of acid centers, as identified by the characteristic absorption frequencies.Their strength, from the variations in intensity of the bands upon desorption at increasing temperature.The reactivity of the OH groups with respect to the base, as seen from both the variation of intensities of the absorption maxima \u03c5(OH) during adsorption and desorption, and by the positions of the reagent bands.The density of acid centers, from a plot of normalized absorbance (IR) against the amount of pyridine adsorbed.In this case, information about the nature, density, and strength of the acid positions on the surface can be obtained. The nature of the interaction can be determined by assigning the frequencies of the physisorbed and chemisorbed pyridine bands in the 1400\u20131700\u2009cm\u22121 region of the IR spectrum (see \nTable 3). The strength of these acid positions can be evaluated by exposing the sample to vacuum treatments and at several temperatures. The reference spectrum (baseline) corresponds to the sample prior to contact with pyridine. To evaluate the density of acid centers, a magnitude known as the normalized absorbance of the intensities must first be defined.\n\n\n\n\n\n\n\nabsorbance\n\n\n\n\u00b7\n\n\n\nn\n\u00ba\nof\nwavelength\n\n\n\n\u00b7\n(\n\n\nmm\n\n\n2\n\n\nIR\nbeam\nsection\n)\n\n\n(\ng\nof\nabsorbent\n)\n\n\n\n\n\nNormalized absorbance represents an acid number that reflects the number of adsorbed species per unit area. It must be assumed that each acid center retains one adsorbed molecule. The sample is prepared in the form of a pellet, which is placed in a cell equipped with NaCl windows in which the sample can be degassed under vacuum and at a temperature of between 400 and 500\u2009\u00b0C (depending on the previous treatment to which the catalyst has been subjected). After a desorption time, the sample is cooled to room temperature before being brought it into contact with pyridine for a short period of time. The sample is desorbed under vacuum at room temperature for 30\u2009min to remove the physisorbed pyridine. Subsequently, it is subjected to vacuum and at several temperatures. At the end of each treatment, the IR spectra are recorded in the range 1300\u20134000\u2009cm\u22121. The spectra obtained upon subtracting the spectrum of the sample before pyridine adsorption and after each desorption are analyzed in the region from 3200 to 3700\u2009cm\u22121 (O\u2013H vibration) and in the region from 1400 to 1700\u2009cm\u22121 (vibration of adsorbed pyridine). Various types of OH groups can be observed in the O\u2013H vibration region (e.g., in PILC: structural hydroxyls and those related to pillared species) [64].The frequencies assigned in Table 3 suggest that the adsorption bands located around 1620, 1575, 1490 and 1450\u2009cm\u22121 are associated with coordinated pyridine (PyL), thus characterizing the Lewis-type acidity. In contrast, the bands at 1640, 1540 and 1490\u2009cm\u22121 are due to the presence of pyridinium ions formed by the interaction with the protonic positions (Br\u00f6nsted acidity). The band at 1545\u2009cm\u22121 is most characteristic of the Br\u00f6nsted-type acidity. The band around 1450\u20131455\u2009cm\u22121 corresponds to a Lewis-type acidity if the sample has been previously degassed, since physisorbed pyridine exhibits a characteristic band at 1440\u20131445\u2009cm\u22121. The bands at around 1490 and 1620\u2009cm\u22121 contain a contribution from both types of acidity. In addition, the band at 1620\u2009cm\u22121 provides information on the strength of the Lewis-type positions: a shift towards higher frequencies (even above 1626\u2009cm\u22121) indicates the presence of a high Lewis-type acidity, whereas if the band moves towards lower frequencies (below 1615\u2009cm\u22121), the acidity of the centers is weaker.It is also well known that CO can reach the Br\u00f6nsted and Lewis acid sites of microporous zeolites due to its small size [76,77]. This molecule allows determination of the oxidation state and environment of the metal cations on the surface and the amount and strength of Br\u00f6nsted and Lewis acid sites.\nQuantitative analysis: from the ratio of the absorbances of the bands due to pyridine adsorbed at a Lewis-type acid position and a band corresponding to pyridine adsorbed at a Br\u00f6nsted position, the ratio of the Lewis and Br\u00f6nsted-type acid positions multiplied by a constant K can be obtained (\n\n\n\nL\n\n\nB\n\n\n\u00b7K\n\n).This expression comes from application of the integrated form of Beer\u2019s law [1]:\n\n(9)\n\n\nB\n=\nC\n\u00b7\nL\n\u00b7\n\n\u222b\n\n\u03b3\n\n\n\n\n\n\u2208\n\n\n\u03b3\n\n\na\n\n\n\u00b7\nd\n\u03b3\n\n\n\n\nwhere B is the peak area (absorbance/cm), C is the concentration of the adsorbed species (mol/g), L is the tablet thickness (g/cm), \u03b3 is the wavenumber (cm\u22121), and \n\n\n\u222b\n\n\u03b3\n\n\n\n\n\n\u2208\n\n\n\u03b3\n\n\na\n\n\n\u00b7d\n\u03b3\n\n\n, is the integrated apparent extinction coefficient (cm/mol).The concentration of species at an IR absorbance maximum can be calculated assuming that the integrated apparent extinction coefficient is known. It can be determined if [1]:\n\n(10)\n\n\n\n\n\n\n\n\n\n\u222b\n\n\u03b3\n\n\n\n\n\n\u2208\n\n\n\u03b3\n\n\na\n\n\n\n\n\n\n\nL\n\n\n\n\n\n\n\n\n\n\u222b\n\n\u03b3\n\n\n\n\n\n\u2208\n\n\n\u03b3\n\n\na\n\n\n\n\n\n\n\nB\n\n\n\n\n=\n\n\n2\n\n\n\n\n\nB\n\n\nL\n\n\n\n\nT\n\n\n1\n\n\n\n\n\u2212\n\n\nB\n\n\nL\n\n\n\n\nT\n\n\n2\n\n\n\n\n\n\n\n\n\n\n\nB\n\n\nB\n\n\n\n\nT\n\n\n2\n\n\n\n\n\u2212\n\n\nB\n\n\nB\n\n\n\n\nT\n\n\n1\n\n\n\n\n\n\n\n\n\nwhere the subscripts L and B refer to a specific IR band for pyridine (e.g., 1450 and 1550\u2009cm\u22121) and T\n\n1\n and T\n\n2\n to two treatment temperatures for the catalyst. The concentration relationship between the Lewis and Br\u00f6nsted positions will be [1]:\n\n(11)\n\n\n\n\n\n\nL\n\n\n\n\n\n\n\nB\n\n\n\n\n\n=\n\n\n\n\nB\n\n\nL\n\n\n\n\n\n\nB\n\n\nB\n\n\n\n\n\n\n\n\n\n\n\n\u222b\n\n\u03b3\n\n\n\n\n\n\u2208\n\n\n\u03b3\n\n\na\n\n\n\n\n\n\n\nL\n\n\n\n\n\n\n\n\n\n\u222b\n\n\u03b3\n\n\n\n\n\n\u2208\n\n\n\u03b3\n\n\na\n\n\n\n\n\n\n\nB\n\n\n\n\n\n\n\n\nIn the case described initially [1]:\n\n(12)\n\n\n\n\n\n\nL\n\n\n\n\n\n\n\nB\n\n\n\n\n\n=\n\n\nL\n\n\nB\n\n\n\u00b7\nK\n,\nsi\nL\n\u2261\n\n\nB\n\n\nL\n\n\ny\nB\n\u2261\n\n\nB\n\n\nB\n\n\n;\nK\n\u2261\n\n\n\n\n\n\n\n\u222b\n\n\u03b3\n\n\n\n\n\n\u2208\n\n\n\u03b3\n\n\na\n\n\n\n\n\n\n\nL\n\n\n\n\n\n\n\n\n\n\u222b\n\n\u03b3\n\n\n\n\n\n\u2208\n\n\n\u03b3\n\n\na\n\n\n\n\n\n\n\nB\n\n\n\n\n\n\n\n\nThe best bands are 1450\u2009cm\u22121 (19b vibrations of the coordinated pyridine) and 1344\u2009cm\u22121, although it must be considered the proximity of the band to 1447\u2009cm\u22121 due to the presence of pyridine linked via a hydrogen bond that may affect the validity of the measurement. The band at 1610\u2009cm\u22121, which is also assigned to coordinated pyridine, can also be used, but in this case, there is likely to be a contribution of the band at 1639\u2009cm\u22121 due to the pyridinium ion. Hence, in general, only the relationship between the sum of the Lewis positions (plus the protonic H positions due to the OH on the surface) and the Br\u00f6nsted acid positions can be determined.Other alternative methods, such as thermogravimetry and pyridine thermo-desorption, have been proposed to quantify the number of acid centers, depending on their strength. Thus, the evolution of the band at 1445\u2009cm\u22121 can be evaluated as a function of the desorption temperature and quantified by representing the amount of pyridine adsorbed per mass of solid as a function of the absorbance per mass of solid.\nApplications. A generic description of pyridine adsorption and its use in the characterization of acid centers in adsorbents and catalysts is difficult, so its study usually involves specific examples. Hence, herein it has been decided to use intercalated/pillared clays as study materials.The acidity and nature of the acid centers (Br\u00f6nsted and Lewis) depend on the cations exchanged, the method of preparation, and the nature of the clay [64,78\u201382]. In the case of aluminum-intercalated clays, Lewis-type acidity is related to two types of acid centers, both of which arise due to the aluminum present in the tetrahedral layer of the clay (LPy, 1641\u2009cm\u22121) and to the aluminum in the pillars (LPy, 1621\u2009cm\u22121) [83]. This latter center is the one that is usually related in the literature to Lewis acidity. In contrast, the origin of the Br\u00f6nsted acid centers in the intercalated clays is not clear. These centers have been related to the structural hydroxyl groups of the clay layer, which in turn are related to the exchange centers, in other words, the protons of the oligomeric cations that form the pillars after heating, and to a synergistic phenomenon between the Si layer of the clay and the pillars [78,80\u201387].When characterizing a hectorite intercalated with pillars of Al, ZrAl and Zr, Occelli observed that, after adsorption of pyridine and being subjected to vacuum (10\u22126 torr) at 300\u2009\u00b0C, the natural hectorite presents both Br\u00f6nsted and Lewis centers [84]. The pillars introduced affect the acidity observed in the initial clay. Thus, with only Al pillars, the characteristic PyH+ bands (1638, 1547 and 1490\u2009cm\u22121) practically disappear or are significantly reduced in intensity, whereas with Zr and mixed ZrAl pillars, the pyridine is retained at both Br\u00f6nsted and Lewis centers, even after degassing under vacuum and at 400\u2009\u00b0C. At 300\u2009\u00b0C and under vacuum in an intercalation with Al2O3, the pyridine is first removed from the Br\u00f6nsted centers. In contrast, pyridine adsorbed at Lewis centers remains practically unchanged above 400\u2009\u00b0C. The presence of Zr increases the Br\u00f6nsted acidity in the intercalated hectorite. It is clear that the absolute intensities of the intercalated hectorite bands increase due to an increase in surface area.In a study on montmorillonites intercalated with aluminum, the same author observed that, after being subjected to a vacuum at 400\u2009\u00b0C, the pyridine continues to be found as PyH+ and PyL. Proton acidity must be responsible for the instability of inorganic pillars at high temperature. Thus, when the pillars are formed by dehydration of the interlayer polymeric cation, protons are generated:\n\n\n\n2\n\n\n\n\n\n\nAl\n\n\n13\n\n\n\n\nO\n\n\n4\n\n\n\n\n(\nOH\n)\n\n\n24\n\n\n\n\n(\n\n\nH\n\n\n2\n\n\nO\n)\n\n\n12\n\n\n\n\n\n\n7\n+\n\n\n\n\n\u2192\n\n\u0394\n\n\n\n13\n\n\nAl\n\n\n2\n\n\n\n\nO\n\n\n3\n\n\n+\n14\n\n\nH\n\n\n+\n\n\n+\n41\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\nAt high temperature, these protons are able to react with the aluminum in the pillars in the same way that acids extract the aluminum from the zeolite structure.\n\nTable\n\n\n\n\n\n\n\nImage 6\n\n\n\n\n\n\n\nWhen this reaction occurs, the pillars decrease in size, and if Al3+ extraction continues, collapse occurs.Fripiat et al. [88] reported that the most important acidity of the montmorillonite surface is due to hydrated H2O molecules, which means that progressive dehydration can occur. Hence, comparing the pyridine and IR adsorption data obtained, these authors considered that both the Br\u00f6nsted centers and the derivative of acid centers in intercalated montmorillonites decrease rapidly at 200\u2009\u00b0C.Despite the above, the actual nature of the acid centers in the pillars remains unknown. Assuming that the intercalated species is Al13\n7+, currently there are no information about the nature of its transformation after thermal activation at 300\u2009\u00b0C, although the formation of a bayerite or gibbsite-type structure has been proposed [64]. In any case, protons from different sources must be the source of the acidity of pillared clays. These authors conclude by proposing that more than 90 % of the acid centers in both intercalated montmorillonites and calcined intercalated beidellites are of the Br\u00f6nsted type, which are able to protonate pyridine to PyH+ (band at 1640\u2009cm\u22121). Some examples of the adsorption and desorption spectra of pyridine adsorbed on different samples (montmorillonite, alumina, silica, and aluminum-intercalated montmorillonite) are shown in \nFig. 9\n[73].\nAdsorption isotherm of the probe molecules. Acidic and basic centers can also be characterized using the static volumetric procedure described in Section 3.1. The amount of adsorbed gas (probe molecule) is obtained as a function of the equilibrium pressure at a constant adsorption temperature (see \nFig. 10). The probe molecules used are those that characterize the acidic or basic properties of the adsorbent/catalyst listed above, such as CO2, NH3, pyridine (C5H5N), acetonitrile (CH3CN), benzonitrile (C6H5CN), CO and NO, amongst others.It is possible to quantify the adsorption capacity from the volume adsorbed at a given pressure and temperature. In the case of NH3 adsorption, the ASTM D 4824\u201393 method proposes the adsorbed volume to be representative of that obtained at a pressure of 150 torr and at a temperature of 175\u2009\u00b0C [17]. However, other parameters that allow the properties of adsorbents and catalysts to be characterized can also be calculated. Thus, Henry\u2019s constant is an important characteristic of adsorption because it provides an indication of the strength of adsorption and the isosteric heat of adsorption at low pressure. Although there are several possibilities for calculating Henry\u2019s constant [89], when it is obtained directly from the isotherm, this method is more accurate that others if sufficient data are available in the low pressure region.The heat effects produced during adsorption processes can be described by the isosteric heat of adsorption and can be determined from the amount of gas adsorbed at several temperatures. The isosteric heat (q\n\nst\n) defines the energy change resulting from the phase change of an infinitesimal number of molecules at constant pressure and temperature and a specific adsorbate loading. One method for calculating the isosteric heat of adsorption involves application of the Clausius\u2013Clapeyron equation [89], which relates the isosteric heat to the pressure change of the bulk gas phase as a consequence of a temperature change for a constant amount adsorbed [89]:\n\n(13)\n\n\n\n\nq\n\n\ns\nt\n\n\n=\n\u2212\nR\n\u00b7\n\n\n[\n\n\n\n\u2202\n\nln\n\np\n\n\n\u2202\n(\n1\n/\nT\n)\n\n\n\n]\n\n\nn\n\n\n\n\n\nwhere p (kPa) is the equilibrium pressure, n is the amount of gas adsorbed at temperature T (K), and R (kJ/mol\u00b7K) is the universal gas constant. The isosteric heat can be obtained from the experimental isotherms at various temperatures by plotting ln\u2009(p) versus 1/T for a constant loading n. The isosteric heat corresponds to the slope of the amount adsorbed by the materials, and the dependence of the isosteric heats of adsorption on the amount adsorbed can indicate the effect of surface loading. Indeed, in some cases, a maximum can be observed in the isosteric heats of adsorption in the presence of such a loading (see Fig. 10) [90]. This behavior can be related to the coating of the surface and subsequent formation of multilayers. Similarly, the limiting heat (q\n\nst\n\n\n0\n) can also be obtained from the temperature dependence of Henry\u2019s constant (H\n\ni\n) by applying the Clausius\u2013Clapeyron equation in the low-pressure region, where the isotherm obeys Henry\u2019s law.\n\n(14)\n\n\n\n\nq\n\n\ns\nt\n\n\n0\n\n\n=\n\u2212\nR\n\u00b7\n\n\n[\n\n\n\nd\n\nln\n\n\n\nH\n\n\ni\n\n\n\n\nd\n(\n1\n/\nT\n)\n\n\n\n]\n\n\nn\n=\n0\n\n\n\n\n\n\nThe isosteric heats obtained from this last equation, and the values found from the isosteric heats at zero coverage, should be similar [90].The techniques and procedures presented in this work allow the characterization, evaluation, and determination of the qualitative and quantitative surface properties of adsorbents and supported metal catalysts by way of selective chemisorption processes.The reaction behavior of a supported metal catalyst depends on the metal surface, the size of the metal particles, and how these particles are distributed on the surface of the catalyst support. Measurement of these properties using a chemisorption or selective adsorption technique requires careful selection of the operating conditions. Once established, however, chemisorption can be considered to be a method for routine measurement of the dispersion of supported metal catalysts. However, to measure the dispersion and particle size from the amount of an adsorbed gas, a series of assumptions are required, and it depends on the preparation and pretreatment conditions of the catalyst. A good practice, if possible, would be to use several adsorbates (H2, CO, O2), as well as to combine O2/H2 cycles and compare the results obtained. It will also be necessary to determine the possible effects of spillover, SMSI, presence of contaminants, and reversible adsorption. Among the techniques proposed, the isothermal dynamic procedure is the most popular since it allows faster measurements compared to the time needed to perform the volumetric measurements. In addition, in this case it is not necessary to volumetrically calibrate the equipment before or after the measurements. However, it has the drawback of only evaluating the centers where there is a strong interaction between the adsorbent gas molecule and the adsorption center.Adsorbents and catalysts are characterized by having acid and basic centers that are involved in a large number of processes related to petroleum refining processes, amongst others. Two types of centers can be distinguished: Lewis and Br\u00f6nsted. The most common technique to qualitatively characterize this type of center is to adsorb an acidic or basic gas molecule (NH3 or CO2) and perform its desorption in a programmed temperature ramp. However, the types of acid or basic adsorption centers cannot be differentiated using this procedure, therefore characterization is only qualitative. It is possible to characterize the desorption forces from the activation energy of desorption by modifying the heating rate. To be able to differentiate between adsorption centers, and even perform quantification, it is necessary to adsorb a molecule (pyridine, acetonitrile, benzonitrile, etc.) and conduct an analysis using IR spectroscopy. To characterize this type of center, it is increasingly common to use the static volumetric procedure, which allows the amount of adsorbed gas as a function of the equilibrium pressure at a constant adsorption temperature to be obtained. In addition to being able to quantify the adsorption capacity from the volume adsorbed at a given pressure and temperature, it is possible to obtain Henry\u2019s constant and the isosteric heat of adsorption. The dependence of the isosteric heats of adsorption on the amount adsorbed can indicate the effects of surface loading.\nA.Gil: is the only author of this work. Conceptualization; Formal analysis; Investigation; Methodology; Resources; Supervision; Validation; Visualization; Writing - original draft; Writing - review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The author express its gratitude to Dr Sim\u00f3n Yunes for valuable discussions and critical reading of the manuscript. The author is grateful for financial support from the Spanish Ministry of Science and Innovation (AEI/MINECO) and Government of Navarra through projects PID2020-112656RB-C21 and 0011-3673-2021-000004. Open access funding provided by Universidad P\u00fablica de Navarra. AG also thanks Santander Bank for funding via the Research Intensification Program.", "descript": "\n                  The adsorption phenomenon has been used extensively to achieve and explain solid-state reactions, control contamination, and purify liquids and gases. This process implies the use of a porous medium or a material with specific adsorption centers where the interactions with the reagents occur. Determination of the properties of adsorbent or catalyst materials that do not contain specific adsorption sites by physical gas adsorption is a well-established procedure in most research and quality-control laboratories. However, characterizing the specific centers by selective adsorption\u2014chemisorption\u2014remains an open question for discussion and study. The specific centers involved are often acidic/basic and metallic; in most cases, reagents are adsorbed and desorbed in these centers, whose determination allows controlling the processes and comparing the materials. The techniques and procedures presented herein facilitate the evaluation and the qualitative and quantitative determination of the surface properties of the materials using chemisorption processes for metallic and acidic/basic sites. The aim of this work is to review these techniques and procedures, including the updates published by several researchers, who mostly strive to explain the results of bifunctional metallic and acid\u2013base catalytic behavior.\n               "}
{"full_text": "Concerns on the increasing demand for energy and strict environmental regulations have ignited interests in producing renewable chemicals and fuels [1]. Biomass, as the sole source of renewable organic carbon, has captured widespread attention because of its ability to be converted into various valuable chemicals [2,3]. As a platform chemical derived from biomass, 5-hydroxymethylfurfural (HMF) could be converted into diverse chemicals (biofuels, functional macromolecular polymers) through different reaction pathways [4,5]. Notably, the oxidation product, 2,5-furandicarboxylic acid (FDCA) which is one of the top-12 value-added chemicals has attracted enormous interests, because it is a crucial building block for the production of bio-based polymers [3,6\u201310].Given the multitude of published references, metal species of heterogeneous catalysts is the main active site [11\u201314], and molecular oxygen (O2) serves as the oxidant, offering advantages of availability and benignity to the environment [6]. Owing to the difficulty in oxygen activation, noble metals [15\u201320], especially Au-based catalysts [21\u201324] are widely used in HMF oxidation on account of their excellent activity and product selectivity under relatively mild conditions. As reported by Corma [25], the catalytic performance of Au catalysts is strongly affected by the support: higher FDCA yields could be achieved over TiO2 and CeO2 supported Au catalysts. Yield of 99% was achieved over Au/CeO2 and Au/TiO2 (Au loading amount: 2.6\u00a0wt% in Au/CeO2 and 1.0\u00a0wt% in Au/TiO2), with 4 equiv of NaOH and 10\u00a0bar air in 5\u00a0h at 65\u00a0\u00b0C [25]. Xu and coworkers performed the reaction at 60\u00a0\u00b0C, 0.3\u00a0MPa O2 with 4 equiv of NaOH and achieved 85% FDCA yield over Au/TiO2 (1.5\u00a0wt% Au) catalyst after 6\u00a0h [26]. Besides the supporting material, the amount of base also plays a significant role in determining the performance of Au-based catalysts. As reported by Riisager [27], 71% FDCA yield could be obtained over Au/TiO2 (1.0\u00a0wt% Au) with 20 equiv of NaOH. However, when low amount of base (less than 5 equiv) was applied, the main product changed from FDCA to HMFCA. Besides, the FDCA yield was only 1% under base-free conditions.Despite the great progresses, there are still severe challenges for Au-based catalysts during HMF oxidation. First, the high loading amount of Au limits their industrial application, rendering maintaining and even promoting the performance of Au/TiO2 catalysts with low Au loading amount an urgent demand. Second, HMF oxidation over Au-based catalysts are usually carried out in the presence of excessive base, leading to both economic and environmental issues. Correspondingly, base-free catalysis attracted more attention [28,29]. In the elegant work of Zhang [30], it is well demonstrated that the Fe\u2013Zr\u2013O exhibited a 60.6% FDCA yield under base-free conditions, which is a excellent result for HMF to FDCA using molecular oxygen as an oxidant. Although base-free oxidation of HMF could be achieved when solid base is utilized as the supports, such as alkaline hydrotalcite (HT), severe leaching of Mg2+ ions from HT occurs inevitably, resulting from the chemical interaction between the basic support and the generated FDCA [31]. Third, the knowledge of the intrinsic active sites and the active oxygen species are still under controversy. It is reported that hydroxide ions in alkaline solution facilitate the activation of the C\u2013H and H\u2013O bond in the alcoholic group to form the formyl intermediate [7,31,32], while Yu and coworkers propose that radicals, instead of hydroxide ions promote the alcohol oxidation step [33].Therefore, from the viewpoint of green and sustainable chemistry, it is of significant importance to prepare catalysts with low Au loading amount and high catalytic performances. Besides, it is greatly imperative to take a deep insight into the catalytic mechanism with regard to the active oxygen species and intrinsic active sites.In this work, a series of metal oxide modified MO\nx\n-Au/TiO2 (M\u00a0=\u00a0Fe, Co, Ni) catalysts were synthesized and evaluated in HMF oxidation. There may be three possible advantages of the as-prepared MO\nx\n-Au/TiO2 catalysts: first, low loading amount of Au (0.5\u00a0wt%) and high catalytic performance, being promising for industrial applications; second, satisfying FDCA yield could be obtained under base-free conditions, complying with the green and sustainable chemistry principles; third, addition of transition metal oxides promotes electron transfer and generation of Au\n\u03b4\u2212\u2013Ov\u2013Ti3+ interface, accelerating the adsorption and activation of the reactants. In order to take a deep insight into the cause for the differences in catalytic activities, the adsorption properties, kinetic study, active sites for the rate-determining step, and the active oxygen species were investigated.Ni(NO3)2\u00b76H2O, Co(NO3)2\u00b76H2O, Fe(NO3)3\u00b79H2O, titanium dioxide (TiO2), HAuCl4\u00b73H2O (99.9\u00a0wt% analytical purity), 5-hydroxymethylfurfural (HMF, 98.0\u00a0wt% analytical purity), 5-hydroxymethyl-2-furancarboxylic acid (HMFCA, 98.0\u00a0wt% analytical purity), 5-formyl-2-furancarboxylic acid (FFCA, 98.0\u00a0wt% analytical purity) and 2,5-furandicarboxylic acid (FDCA, 98.0\u00a0wt% analytical purity) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. All items were used as received without further purification.The details about the catalyst preparation are given in the Supporting Information.XRD, AC-HAADF-STEM with energy dispersive spectroscopy (EDS), inductively coupled high-frequency plasma (ICP), X-ray photoelectron spectra (XPS), temperature programmed reduction (TPR) with H2, temperature programmed desorption (TPD) with O2 were used, and the relational details are depicted in the Supporting Information.The HMF oxidation was evaluated in a high-pressure stainless-steel autoclave equipped with a stirring function and the heating function. Detailed operating procedures and calculation methods are described in the Supporting Information.XRD patterns of Au/TiO2 and MO\nx\n-Au/TiO2 are shown in Fig.\u00a01\n. The pure TiO2 displays three peaks at 25.3\u00b0, 37.8\u00b0, and 53.9\u00b0 (PDF#21-1272), corresponding to the characteristic peaks of anatase crystal phase. No obvious peaks for Au or MO\nx\n are observed, implying that Au and MO\nx\n may be highly dispersed on the surface of TiO2. The XRD lattice parameters of MO\nx\n-Au/TiO2 are intensely similar to those of Au/TiO2, indicating that the transition metal species did not enter the TiO2 crystal lattice and no solid solution was formed between them.High-angle annular dark field scanning transmission electron microscope (HAADF-STEM) was conducted to further investigate the morphological and structural characteristics of the catalysts. As shown in Fig.\u00a02\n, Au and Co species are uniformly dispersed on the surface of TiO2, in agreement with the XRD result. The content of Au was determined by ICP, and the result is listed in Table S1 in the supporting information.In order to fully understand the electronic structure and local environment of the catalysts, they were characterized by XPS, Raman and H2-TPR. As shown in Fig.\u00a03\nA, except for Au/TiO2, the binding energies of Au 4f7/2 for the other three catalysts all shift towards the lower direction. Au 4f7/2 could be deconvolved into two peaks, the one at \u223c83.5\u00a0eV corresponds to the Au0 species [34], and the other at \u223c82.6\u00a0eV is ascribed to the electron-rich Au species (Au\n\u03b4\u2212). After modification with different transition metal oxides, the Au\n\u03b4\u2212 content increases from 0 (Au/TiO2), 31.6% (FeO\nx\n-Au/TiO2), 48.7% (NiO\nx\n-Au/TiO2) to 58.1%\uff08CoO\nx\n-Au/TiO2\uff09(Table 1\n). The Ti 2p3/2 spectrum in Fig.\u00a03B was deconvolved into two peaks of \u223c458.5 and \u223c457.9\u00a0eV, corresponding to Ti3+ and Ti4+ [35], respectively. After modification by transition metal oxides, the Ti3+ contents rise obviously, reaching the highest value of 58.5% in CoO\nx\n-Au/TiO2. Fig.\u00a03C, D and E show the XPS spectra of Fe 2p, Ni 2p and Co 2p, respectively. Ni and Co species exist on the catalyst in the form of both divalent and trivalent ions, while Fe species exists in the form of trivalent and tetravalent ions [36\u201340]. This result confirms that electrons transfer from the transition metal species (Fe, Ni and Co) to Au and Ti species. Fig.\u00a03F shows the XPS spectra of O 1s, there are four peaks around 532.1, 531.0, 530.1 and 529.4\u00a0eV, which could be assigned to surface hydroxyl groups, surface hydroxyl groups, lattice oxygen and oxygen vacancy (Ov), respectively [41,42].\nFig.\u00a04\n shows the Raman spectra. The five peaks located at 146, 197, 390, 513, and 639\u00a0cm\u22121, correspond to Eg, Eg, B1g, A1g (or B1g) and Eg modes respectively [43,44], which are characteristic peaks of anatase TiO2. The peaks of FeO\nx\n, CoO\nx\n and NiO\nx\n are not detected, confirming their high dispersion on TiO2, in line with the XRD results. After modification with transition metal oxides, the intensities of all the five characteristic peaks attributed to Au/TiO2 reduce significantly d, which may result from the generation of Ov, since part of the Ti\u2013O\u2013Ti is replaced by Ti\u2013Ov\u2013Ti [45]. In addition, the Eg mode in the range of 100\u2013200\u00a0cm\u22121 is presented in Fig.\u00a04B. After addition of transition metal oxides, the peak moves to a higher frequency direction, besides, the full width half maximum (FWHM) increases slightly. CoO\nx\n-Au/TiO2 features the highest frequency and the largest FWHM, suggesting the largest amount of Ov [43]. Similar results have also been reported by Han [11], Ov concentration could be significantly improved via doping. Given that Ov could directly serves as adsorption sites for reactant molecules [46], the largest amount of Ov in CoO\nx\n-Au/TiO2 would lead to the best adsorptive performances for O2 and HMF.Besides XPS and Raman, H2-TPR was also conducted to evaluate the surface oxygen reducibility of the catalysts (Fig.\u00a05\n). As demonstrated in previous studies [47\u201349], TiO2-supported VIII group metals (Fe, Co, Ni) exhibited characteristic peaks of hydrogen consumption with a maximum over 300\u00a0\u00b0C. This indicates the two peaks could be assigned to the reduction of Au2O3 (50-200\u00a0\u00b0C) and Ti-MO\nx\n (500-600\u00a0\u00b0C) [50], respectively. The incorporation of MO\nx\n leads to significant shifts of peak maximums towards lower temperature, which could be attributed to massive oxygen adsorption on catalysts. Au\n\u03b4\u2212 may facilitates activation adsorption of small molecule gas such as O2. With more Au\n\u03b4\u2212 formation, the hydrogen consumption of Co dopped catalysts is higher than Ni and Fe, and all these samples spend more hydrogen than Au/TiO2, which is consistent with the XPS result.Combining the XPS, Raman and H2-TPR results, it is reasonable to conclude that addition of the transition metal oxides contributes to electron transfer in the catalyst, generating the Au\n\u03b4\u2212\u2013Ov\u2013Ti3+ interface.The catalytic performances for HMF oxidation were investigated, as shown in Fig.\u00a06\nA, 100% HMF conversion could be achieved over all the as-prepared catalysts. However, there is blatant differences in FDCA yield: Au/TiO2 exhibits the FDCA yield of only 53.9%, which is the lowest among the four catalysts. FDCA yield of the catalyst modified by FeO\nx\n, NiO\nx\n, and CoO\nx\n is 71.6%, 82.5% and 90.2%, respectively. The CoO\nx\n-Au/TiO2 catalyst was recycled, and the result is depicted in Fig.\u00a06B. After five recycles of HMF oxidation, the catalytic performance of CoO\nx\n-Au/TiO2 remained basically unchanged, showing excellent stability (Fig.\u00a06B). The content of Au and Co are lower than 10\u00a0ppm in the liquid, confirming the good stability of the as-prepared catalysts. Given that HMF is a notoriously labile compound and readily converts into degradation by-products at high temperature under alkaline environment, lacking active sites for HMF oxidation would result in low FDCA selectivity and yield. This explains the suboptimal FDCA yield of Au/TiO2, which may be due to the low Au loading amount and insufficient active sites for HMF oxidation. With the modification of FeO\nx\n, NiO\nx\n, and CoO\nx\n, higher FDCA yields are obtained, and the reasons will be discussed in detail in the following part.Considering that the adsorption of reactants plays a significant role during heterogeneous catalysis, the adsorption properties for both HMF and O2 over different catalysts were investigated firstly. As shown in Fig.\u00a07\nA, catalysts modified by transition metal oxides exhibit significantly enhanced HMF adsorption capacity in the order of CoO\nx\n-Au/TiO2\u00a0>\u00a0NiO\nx\n-Au/TiO2\u00a0>\u00a0FeO\nx\n-Au/TiO2. Fig.\u00a07B illustrates the O2-TPD profile, which could be deconvoluted to three peaks via a Gaussian peak fitting method according to the desorption temperature: adsorbed oxygen species (<300\u00a0\u00b0C), lattice oxygen species (300\u2013400\u00a0\u00b0C), and bulk lattice oxygen species (>400\u00a0\u00b0C) [51\u201353]. As listed in Table 1, similarly, the as-prepared catalysts present the same trend for the adsorption capacities for O2. CoO\nx\n-Au/TiO2 shows both the highest adsorption capacity of 266.8\u00a0\u03bcmol/g and the ratio of adsorbed oxygen species to all oxygen species (24.9%). Among the three kinds of oxygen species, only the adsorbed oxygen species can directly participate in and play a vital role in the reaction [52]. The promoted O2 and HMF adsorption capacity after transition metal oxides may result from generation of Au\n\u03b4\u2212 and Ov: O2 adsorption and activation could be enhanced over negatively charged metal species through the donation of electrons from the metal to the antibonding \u03c0\u2217 orbital of O2 [54\u201356], at the same time, as demonstrated by first-principle calculations and experimental results, Ov could directly serve as the adsorption sites for O2 and alcohols [46,57\u201360]. The outstanding adsorption property for HMF and the large ratio for adsorbed oxygen species over CoO\nx\n-Au/TiO2 may contribute to its excellent catalytic performance, owing to more reactants gathering around the active site and accelerating the production of FDCA.In order to take a deep insight into the reactive sites, the time course of HMF oxidation on different catalysts were investigated, and they follow the same reaction pathway (Fig.\u00a08\n). As depicted in Scheme 1\n, there are three steps for FDCA production form HMF: first, oxidation of the formyl group, generating HMFCA with a rate constant of k\n1, second, oxidation of the hydroxyl group in HMFCA to produce FFCA (k\n2), third, transformation of formyl group of FFCA to carboxyl group with the rate constant of k\n3. The content of FFCA is low over the four catalysts, revealing that FFCA could be quickly converted into FDCA under the reaction conditions. This reaction path is consistent with the previous reports using Au catalysts [61,62]. After immobilization of transition metal oxides, the consumption rate of HMFCA is accelerated distinctly, accompanied with the remarkable increase in FDCA production rate.Rate constants for each step are fitted by quasi-first order reaction kinetics. As depicted in Table 2\n, k\n2 is much smaller than k\n1 and k\n3, providing clear evidence that oxidation of HMFCA into FFCA is the rate determining step of the whole oxidation process. After modification with FeO\nx\n, CoO\nx\n and NiO\nx\n, the rate constant of each step grows obviously, especially k\n2 rises from 0.02 min\u22121 (Au/TiO2) to 0.30 min\u22121 (CoO\nx\n-Au/TiO2), demonstrating that oxidation of the hydroxyl group in HMF is intensively enhanced after modification by transition metal oxide (Table 3\n).Given that oxidation of the hydroxyl group in HMF is the rate-determining step, researches about the catalytic mechanism (active oxygen species and active sites) and the underlying cause for different performances of the four as-prepared catalysts would be carried out aiming at the step of HMFCA\u2192FFCA in the following part.Despite the extensive researches, it is still controversial whether base is the key factor for oxidation of hydroxyl group in HMF [33]. Therefore, experiments have been designed to reveal the effect of base. As shown in Fig.\u00a09\n, in the first 30\u00a0min, the HMFCA conversion rates under base and base-free conditions are roughly the same (difference is less than 1%), but the concentrations of FFCA and FDCA are quite different. Under base conditions (Fig.\u00a09A), FFCA converts into FDCA as soon as being generated, so only a small amount of FFCA could be detected during the continuous sampling process, which is consistent with the results of time course study. While when there is a trace amount base or no base in the solution (Fig.\u00a09B and C), the generated FFCA transforms into FDCA at a slow rate. The difference in FFCA conversion rate may be caused by the weak solubility of FDCA in trace alkali and non-alkaline solution. The above experiment undisputedly substantiating that base is not the key factor for hydroxyl group oxidation, in good agreement with results reported by Yu [33].This scenario motivates us to go a step further by focusing our interest in identifying the active oxygen species. As shown in Fig.\u00a010\n, EPR was used to detect hydroxyl radicals (OH\u2212) and superoxide radicals (O2\n\u2212). The four characteristic peaks with the intensity ratio of 1:2:2:1 attributes to the signal of DMPO-OH (Fig.\u00a010A), and the characteristic peak with a signal intensity ratio of 1:1:1:1 corresponding to DMPO-O2 (Fig.\u00a010B) [33]. The EPR signals of the two adducts are clearly observed over the four samples, indicating that both of the two oxygen-containing free radicals are generated. Among the four catalysts, CoO\nx\n-Au/TiO2 exhibits the highest concentration for both hydroxyl radicals and superoxide radicals.In order to further investigate which free radical governing the catalytic performance during oxidation of the hydroxyl group, a selective poisoning experiment was carried out. Isopropanol and p-benzoquinone are added as the scavenger for hydroxyl radicals and superoxide radicals, respectively. The FDCA yield is basically unchanged, regardless of the amount of isopropanol (Fig.\u00a011\nA), on the other hand, the FDCA yield is sensitive to the addition of p-benzoquinone (Fig.\u00a011B), illustrating that superoxide radical is the exclusive factor determining the catalytic performance. Oxygen is first adsorbed on the Au\n\u03b4\u2212 and Ov\u2013Ti3+ sites of the catalyst in the reaction system, and transforms into superoxide radicals (O2\n\u2212) after being activated by the catalyst, and then combines with H2O on the catalytic interface generating OOH\u2212 and OH\u2212 (Eqs (1) and (2)). Similar results have also been reported in the work of Liu and Yu [17,33]. According to DFT calculations, active oxygen species promote the reaction (from HMFCA to FFCA) by reducing the energy barrier of hydroxyl dehydrogenation. It is well accepted that the incorporated O atoms in FDCA are provided by H2O [63], while much less is known with regard to the detailed contribution of O2, which has been exactly unveiled in this work.\n\n(1)\nO2 + e\u2212 = O2\n\u2212\n\n\n\n\n\n(2)\n\nO2\n\u2212 + H2O + e\u2212 = OOH\u2212 + OH\u2212\n\n\n\nConsidering that superoxide radicals stem from oxygen adsorbed on catalysts, O2-TPO was used to evaluate the reactivity of oxygen species (Fig.\u00a0S1). The peak can be divided into two parts: 50\u2013200\u00a0\u00b0C and higher than 200\u00a0\u00b0C, which are attributed to the ultimate oxygen storage capacity (OSC) and the desorption of lattice oxygen caused by high temperature [64,65]. As shown in Table 4\n, after the modification with transition metal oxide, the OSC value of the catalyst has been significantly improved, with the order of CoO\nx\n-Au/TiO2 (62.8\u00a0\u03bcmol/g)\u00a0>\u00a0NiO\nx\n-Au/TiO2 (48.0\u00a0\u03bcmol/g)\u00a0>\u00a0FeO\nx\n-Au/TiO2 (36.6\u00a0\u03bcmol/g). According to previous reports, the adsorption strength of O2, CO2 and other small molecule gases on the metal surface is related to the electron density [49,66,67]. The strong activation ability of O2 on Au\n\u03b4\u2212 and Ov\u2013Ti3+ is one of the main factors to ensure the high activity of the catalysts.Considering that TiO2 is widely applied as photocatalyst, photocatalytic oxidation of HMF was also conducted over CoO\nx\n-Au/TiO2 catalysts. As depicted in Table 5\n, The FDCA yields obtained over the CoO\nx\n-Au/TiO2 (Au 8.0\u00a0wt%) catalyst under irradiation for 2\u00a0h and dark reaction conditions for 24\u00a0h are basically the same, which means that the reaction under light illumination is probably ten times faster than under dark conditions. Besides, after optimization in the amount of both catalyst and base, FDCA yields of 13% and 3% were achieved over CoO\nx\n-Au/TiO2 catalysts with the Au loading amount of 0.5\u00a0wt% and 8.0\u00a0wt%, respectively. Combined with the EPR characterization and the result of radical scavenger test, these experiments confirm that HMF oxidation follows the radical mechanism over CoO\nx\n-Au/TiO2 catalysts. The detailed mechanism will be carried out in our following work.For an environmentally friendly industrial application, inspired by Fu and Wang [18,19], we carried out HMF oxidation in a base-free environment (Fig.\u00a012\n), and a FDCA yield of 71.2% was obtained after prolonging the reaction time. This promotes the process of green chemistry, which demonstrates a great significance to environmental protection.The elegant work of Zhou and coworkers [68] give us the inspiration that the interface between Au and the support may be the active sites for the hydroxyl group in HMF. In order to clarify the relationship between the Au\n\u03b4\u2212\u2013Ov\u2013Ti3+ interface and the catalytic activity clearly, the rate constant of the rate-determining step (k\n2) is correlated with the surface Au\n\u03b4\u2212/(Au\n\u03b4\u2212\u00a0+\u00a0Au0) ratio and Ti3+/(Ti3++Ti4+) ratio, respectively. As illustrated in Fig.\u00a013\nA and B, Au\n\u03b4\u2212/(Au\n\u03b4\u2212\u00a0+\u00a0Au0) ratio and Ti3+/(Ti3++Ti4+) ratio display a concave and convex monotonic increase trend with k\n2, respectively, suggesting that Au\n\u03b4\u2212 and Ov\u2013Ti3+ govern the catalytic activity cooperatively. It strongly confirms that the Au\n\u03b4\u2212\u2013Ov\u2013Ti3+ interface site acts as the intrinsic active center toward oxidation of the hydroxyl group. Au\n\u03b4\u2212 site enhances O2 adsorption through the donation of electrons from the metal to the antibonding \u03c0\u2217 orbital of O2 [54\u201356]. At the same time, Ov\u2013Ti3+ site plays multiple roles: first, it serves as the adsorption site for both O2 and the alcohol [46,57\u201360], the adsorbed oxygen combines with electrons transforming into O2\n\u2212 and the adsorption of HMFCA enhances the oxidation of the hydroxyl group, similar result has also been reported by Wang [22]. Second, the activation and dissociation of H2O would be accelerated over the Ov\u2013Ti3+ site [69\u201371], more superoxide radicals (O2\n\u2212) combine with dissociated H2O to accelerate the formation of strong oxidizing species (\u2217H2O2) and promotes the rate-determining step, namely oxidation of the hydroxyl group.Based on previous research [33,72], a mechanism for HMFCA oxidation on MO\nx\n-Au/TiO2 catalyst is proposed (Scheme 2\n). The oxygen vacancy on the catalyst promotes the adsorption of both H2O and HMFCA molecules. O2 combines with the electrons on Ov\u2013Ti3+ or Au\n\u03b4\u2212 to form superoxide radicals (O2\n\u2212) and reacts with the H2O dissociated on the oxygen vacancy to generate OOH\u2212 and OH\u2212. After that, OOH\u2212 is further combined with H2O to form strong oxidizing species (\u2217H2O2) [17], which may aid the cleavage of C\u2013H and H\u2013O bond in HMFCA, generating FFCA and H2O.In this work, we have prepared Au/TiO2 catalysts modified with transition metal oxides MO\nx\n (M\u00a0=\u00a0Fe, Co, Ni) for HMF oxidation. Physical characterizations confirm the electron transfer from the transition metal species to Au and Ti, generating Au\n\u03b4\u2212\u2013Ov\u2013Ti3+ interface. The kinetic study reveals that the oxidation of hydroxyl group is the rate-determining step during FDCA production from HMF. The selective poisoning experiment demonstrates that superoxide free radicals stem from O2 instead of base is the dominant factor governing the catalytic activity. On this basis, HMF oxidation under base-free conditions has been carried out, achieving a FDCA yield of 71.2%. Studies on structure\u2013performance unveil that the Au\n\u03b4\u2212\u2013Ov\u2013Ti3+ interface is the active sites for hydroxyl group oxidation: Au\n\u03b4\u2212 sites enhance O2 adsorption and activation on the catalysts surface, and Ov\u2013Ti3+ sites act as the role of \u201ckilling two birds with one stone\u201d: enhancing adsorption of both HMF and O2, and accelerating the activation and dissociation of H2O. Therefore, this work demonstrates the synergic catalysis during HMF oxidation and achieves a better understanding of the reaction mechanism, which would be constructive for rational design of other heterogeneous catalysts.The authors declare no competing financial interest.We gratefully acknowledge the support of State Key Laboratory of Chemical Engineering (No. SKL-ChE-20A02), and the support of International Clean Energy Talent Program by China Scholarship Council.The following is/are the supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.gee.2021.09.006.", "descript": "\n                  Despite wide applications of noble metal-based catalysts in 5-hydroxymethylfurfural (HMF) oxidation, promoting the catalytic performance at low loading amounts still remains a significant challenge. Herein, a series of metal oxide modified MO\n                        x\n                     -Au/TiO2 (M\u00a0=\u00a0Fe, Co, Ni) catalysts with low Au loading amount of 0.5\u00a0wt% were synthesized. Addition of transition metal oxides promotes electron transfer and generation of the Au\n                        \u03b4\u2212\u2013Ov\u2013Ti3+ interface. A combination study reveals that the dual-active site (Au\n                        \u03b4\u2212-Ov-Ti3+) governs the catalytic performance of the rate-determining step, namely hydroxyl group oxidation. Au\n                        \u03b4\u2212 site facilitates chemisorption and activation of O2 molecules. At the same time, Ov-Ti3+ site acts as the role of \u201ckilling two birds with one stone\u201d: enhancing adsorption of both reactants, accelerating the activation and dissociation of H2O, and facilitating activation of the adsorbed O2. Besides, superoxide radicals instead of base is the active oxygen species during the rate-determining step. On this basis, a FDCA yield of 71.2% was achieved under base-free conditions, complying with the \u201cgreen chemistry\u201d principle. This work provides a new strategy for the transition metal oxides modification of Au-based catalysts, which would be constructive for the rational design of other heterogeneous catalysts.\n               "}
{"full_text": "No data was used for the research described in the article.With the increasing concern of climate change, the demand of renewable and carbon\u2013neutral energy is also arising. The IEA (International Energy Agency) scenarios approaching the carbon neutral policy advocated the need of electrification but also steady requirement of conventional liquid fuel in the transportation sector. Even the usage of renewable energy is the highest in NZE (Net Zero Emissions by 2050) scenario, the market share of EVs (Electric Vehicle) is predicted to reach only 60\u00a0%. Carbon neutral energy such as hydrogen, ammonia and biofuels should redeem the shortage. Among the biofuels, biodiesel is anticipated to be a carbon neutral liquid fuel that can fulfill the surplus energy demand [1].Biodiesel can be obtained from transesterification of triglyceride (TG) with alcohol in the presence of base catalyst. Generally, the commercial biodiesel production is based on homogeneous base catalysts such as sodium or potassium hydroxide due to their high reaction rate and reasonable cost. Unfortunately, those catalysts are difficult to be extracted and recovered from the product. On the contrary, heterogenous catalysts can mitigate these problems by simplifying the separation between the product and catalyst. Various heterogeneous base catalysts have been studied for biodiesel production for instance alkali-doped materials, alkaline earth metal oxides, transition metal oxides and natural clays [2] to obtain sufficient initial activity, stability and life time. The results of previous studies on the literature are summarized in Table 1\n. Particularly, the catalyst stability and life time can be simulated by the sustained conversions over the repeated batch reactions with the recovered catalysts from previous run.Alkali and alkaline earth metal oxides have been widely used for biodiesel production. Na doped aluminate was the first alkali metal applied to biodiesel production. Highly dispersed Na in Na/NaOH/\u03b3-Al2O3 performed almost the same activity as that of homogeneous NaOH catalyst [37]. Li et al. impregnated Li over NaY zeolite [5]. The catalyst showed initial biodiesel yield of 98.6\u00a0% which declined to below 80\u00a0% after 6 repeated cycles due to the hydrolyzation of active components. Even though some of the catalytic activity could be recovered through the additional calcination, approximately 10\u00a0% loss of initial activity was inevitable. Lani et al. prepared SiO2 impregnated CaO as heterogeneous catalyst for biodiesel production [18]. The yield of biodiesel reached 90\u00a0% in the first run with a remarkable reduction from fifth to tenth use of catalysts dropping to 52.5\u00a0%. The deactivation of the catalyst might be resulted from the bond breaking leading to leaching of active metal and the formation of inactive intermediate species such as Ca(OH)2\nand calcium diglyceroxide. Dai et al. calcinated Li2CO3 with TiO2 from 700 to 1000\u00a0\u00b0C producing active Li-O-Ti species present in the new crystal structure. Li present in the crystal structure could reinforce the activity and prevent leaching of the species. The conversion showed 98.2\u00a0% at the first batch run which decreased to 80\u00a0% over 10 repeated cycles [15].With the significant efforts, the deactivation caused by leaching and deformation of active components is still remained to be solved to develop active and stable heterogeneous catalysts. In this study, we introduced Na on the graphitic carbon nitride (GCN) as a novel heterogenous base catalyst suitable for transesterification of soybean oil. GCN was known to have graphene-like structure comprised of C and N atoms with H impurities [38]. The N atoms having lone-paired electrons could donate them performing as Lewis base sites. There have been few studies using GCN for biodiesel production focusing mainly on the initial activity of the catalysts prepared by multi steps using expensive precursors [31,39]. The present Na co-polymerized GCN showed stronger basicity compared to the bare GCN, which could be prepared through a single step using cost-effective precursors. The catalyst showed initial activity of 90\u00a0% conversion which was retained over 10 consecutive batch cycles at 70\u00a0\u00b0C. The catalyst was characterized using FT-IR, XRD, XPS and CO2-DRIFT. The characterization and DFT calculation results showed that strong electron dislocation and rigid bond between Na and N induced during the single step synthesis contributed to catalytic activity and the leaching resistance, respectively. The developed Na co-polymerized GCN catalysts might be widely applied to the newly constructed biodiesel production facilities.Pristine graphitic carbon nitride (GCN) was synthesized by the co-thermal polymerization of melamine (Sigma Aldrich) at 550\u00a0\u00b0C (10\u00a0\u00b0C/min) for 4\u00a0h in a crucible with a cover. The pale-yellow powder obtained was ground and sieved, and particles between 300 and 400\u00a0\u00b5m were collected.Suitable amounts of NaOH and melamine were ball-milled and the white solid solution obtained was ground and thermally polymerized using the same procedure used for Pristine GCN preparation. The catalysts were labeled as Na-GCN-n-copol, where n indicates the gram of NaOH introduced when the weight of melamine is 1\u00a0g.For comparison, Na impregnated catalyst was prepared by adding pristine GCN to NaOH aqueous solution. The mixture was sonicated for 1\u00a0h and transferred to rotary evaporator followed by calcination under N2 flow at 550\u00a0\u00b0C for 2\u00a0h. The concentration of NaOH solution was controlled to obtain Na impregnated catalyst with the identical amount of Na present in Na-GCN-0.25-copol catalyst assuming that all the Na in the solution was deposited on the pristine GCN. Thus, obtained catalyst was denoted as Na/GCN-imp.29.1\u00a0g of soybean oil, 71.3\u00a0g of methanol and 11\u00a0wt% of catalyst were transferred to a round-bottom flask and stirred magnetically. The assemblies were placed on a hotplate and heated to 70\u00a0\u00b0C while mixing continuously at a rate of 375\u00a0rpm. No weight difference was observed during the reaction.Following the reaction, the final liquid products was filtered using a glass vacuum filtration apparatus. The liquid products were diluted with methanol prior to the GC injection. For the reusability test, the used catalysts were collected and recycled after washing them with hexane (100\u00a0ml/g) and then drying them at 60\u00a0\u00b0C.The amount of fatty acid methyl ester (FAME) was measured by gas chromatography (GC, Agilent 6890, Agilent) equipped with an INNOWax capillary column (30\u00a0m\u00a0\u00d7\u00a00.32\u00a0mm\u00a0\u00d7\u00a00.5\u00a0\u03bcm, Agilent) and flame ionization detector (FID). The injector was set at 250\u00a0\u00b0C with an injection volume of 1\u00a0\u03bcl and a split ratio of 80:1. The oven temperature was held constant at 210\u00a0\u00b0C for 9\u00a0min, increased to 230\u00a0\u00b0C at a speed of 20\u00a0\u00b0C/min and then kept constant for 10\u00a0min. The biodiesel production yield was calculated using the following equation (1):\n\n(1)\nYield [%] = (FAME product [g])/(soybean oilfeed [g])\u00a0\u00d7\u00a0100\n\n\nFAMEproduct [g] =.\n\n\u2211\n\n\n\n\n\n(\nF\nA\nM\nE\n)\n\n\nconc\n.\n\n\n\u00d7\n\n\n\nM\nW\n\n\n\n\u00d7\n\n\n\nt\no\nt\na\nl\n\np\nr\no\nd\nu\nc\nt\n\nv\no\nl\nu\nm\ne\n\n\n\n\n\n\n\n\n\n\n\n\n\n(\nF\nA\nM\nE\n)\n\n\nconc\n.\n\n\n\n: FAME molar concentration measured by GC for each methyl esters contained in soybean oil.\n\n\nMW\n\n: molecular weight of each methyl esters.The determination of biodiesel yield may be calculated following the internal standard method using methyl heptadecanoate (C17) as described in the European Standard EN 14103:2020 and in many references [40,41]. The internal standard method ensures proper accuracy when the analyzing sample contains ester content of FAME greater than 90\u00a0%. In the present work, depending on the catalysts, the ester content in the samples varied from 25 to 95\u00a0%. Hence, direct measurement of FAME concentration was adopted as mentioned in many recent publication [21,29,35,37,42\u201344]. The calibration curves of the standard solution of the five main methyl esters contained in soybean oil were obtained as shown in Fig. S2, which confirmed linearity and precise accuracy. The total weight of the produced FAME was calculated by adding the product of FAME concentration, molecular weight and the total product volume as given by Eq.(1). The molecular weight of methyl palmitate, methyl stearate, methyl oleate, methyl linoleate and methyl linolenate was 270.5, 298.5, 296.5, 294.5 and 292.5\u00a0g/mol, respectively.FT-IR spectra were obtained using a Nicolet iS50 spectrometer (Thermo Fisher Scientific). Prior to measurement, the catalyst sample was diluted with KBr powder and pressurized at 400 bars for 30\u00a0min. In situ DRIFT spectra were collected under CO2 flow (30\u00a0ml/min) in order to analytically confirm the basicity of the catalysts. Prior to the characterization, the sample cell was fully purged with Ar followed by the heat treatment at 110\u00a0\u00b0C for 1\u00a0h to remove the moisture adsorbed on the sample and cooled to 30\u00a0\u00b0C. Then the flow gas was switched to CO2 and maintained until saturation. The temperature of the sample cell was raised to the 70\u00a0\u00b0C to measure the CO2 adsorption peak. The XRD was measured with D/MAX 2500\u00a0V (Rigaku) using Ni-filtered Cu K\u03b1 radiation (\u03bb\u00a0=\u00a00.154\u00a0nm) at 40\u00a0kV and 200\u00a0mA. The interlayer distance and crystal size were calculated using the Bragg and Scherrer equations, respectively [45]. The Brunauer\u2013Emmett\u2013Teller (BET) surface area was obtained by nitrogen sorption experiments conducted at \u2212196.15\u00a0\u00b0C using a Micrometrics instrument. The sample was heated at 200\u00a0\u00b0C for 2\u00a0h at 5\u00a0\u00b0C/min before analysis in liquid nitrogen. X-ray photoelectron spectroscopy (XPS) measurements were performed using an AXIS Ultra with a delay-line detector (DLD) (Kratos Analytical) and monochromatic Al K\u03b1 (1486.6\u00a0eV) X-ray radiation. The value of the C1s core level (284.6\u00a0eV) was used as the standard peak to calibrate the chemical shift. The amount of Na present was obtained using a liquid ICP AVIO500 (PerkinElmer). The identified amount of liquid produced was vigorously mixed with DI water and then allowed to settle until the two layers were clearly distinct. The volume of the water layer was measured and used for liquid ICP analysis.In this study, all first-principle calculations were performed based on the Kohn\u2013Sham density functional theory (KS-DFT) as implemented in the Vienna Ab initio Simulation Package (VASP) [46]. The generalized gradient approximation within the Perdew\u2013Burke\u2013Ernzerhof (PBE) functional form was used to exchange the correlation energy [47,48]. Plane-wave basis sets with a kinetic energy cutoff of 400\u00a0eV were used to expand the valence electron wave functions. For all structural relaxations, the convergence criterion for the energy in electronic SCF iterations, and the Hellmann\u2013Feynman force in ionic step iterations, was 1.0\u00a0\u00d7\u00a010-5 eV and \u22120.05\u00a0eV\u00a0\u00c5\u22121. To reduce the interaction between neighboring layers, a large vacuum space of at least 15\u00a0\u00c5 was introduced along the z-axis. A Monkhorst\u2013Pack special k-point mesh of 1\u00a0\u00d7\u00a01\u00a0\u00d7\u00a01 was used to sample the first irreducible Brillouin zone. Furthermore, we analyzed the electronic structure properties, which may provide insight into the catalytic process of the reaction trajectories. In this regard, we predicted the energetic stability of the reactants and products by computing the adsorption energies of methanol, methoxyl and hydrogen on a 3\u00a0\u00d7\u00a03\u00a0\u00d7\u00a01 supercell of g-C3N4 with 126 atoms using the following equation:\n\n(2)\nEad\u00a0=\u00a0ET \u2013 Eorganic \u2013E GCN\n\n\n\nFrom the adsorption energies of the reactant and product, we calculated the reaction energy as follows:\n\n(3)\n\u25b3E\u00a0=\u00a0Ead(CH3O)\u00a0+\u00a0Ead(H)\u00a0\u2212\u00a0Ead(CH3OH)\u00a0\u2212\u00a0E(slab)\n\n\nwhere \u25b3E, Ead(CH3O), Ead(H), and Ead(CH3OH)\n were the reaction energy, adsorption energies of methoxyl, hydrogen and methanol, respectively, and E(slab)\n was the slab energy.FTIR spectra of the catalysts were shown in Fig. 1\n. The spectra of pristine GCN was revealed to be similar with those of previous studies containing distinct CN heterocycles and tri-s- heptazine structure at 1150 to 1700 and 808\u00a0cm\u22121, respectively. The broad bands at 3000\u20133300\u00a0cm\u22121 were attributed to amine and hydroxyl groups derived from uncondensed melamine moiety and oxygen from air during polymerization. The intensity of CN heterocycles and amine was found to be reduced on both Na copolymerized and impregnated catalysts. Furthermore, this phenomenon was notably accelerated as the Na content was increased. The bands at 1100, 2900 and 2710\u00a0cm\u22121 were assigned to alkyl groups and cyano groups. During the preparation of KOH modified GCN, OH\u2013 and K+ ions were reported to interact with amine groups and break the CN heterocycle on the edge of the pristine GCN plane generating cyano groups during the copolymerization [49,50].The structural reconstruction according to the addition of Na was confirmed by XRD analysis as shown in Fig. 2\n. The pristine GCN exhibited its typical diffraction pattern at 13.3\u00b0 and 27.4\u00b0, corresponding to the (100) and (002) planes, respectively. The diffraction pattern of (100) plane could be understood as the GCN structure consisted of tri-s-triazine motifs, while the (002) pattern corresponded to the periodic stacking of GCN planes in c-axis [51]. The (100) peak of Na copolymerized catalysts became broader with the increasing of Na content and shifted toward lower angle, which suggested the partial cleavage of CN bonds on GCN planes. On the other hand, the for Na/GCN-imp catalyst, the diffraction pattern of (100) was not observed, which implied the full opening of tri-s-triazine structure comprising the GCN plane.Similarly, with the increase of Na content, the (002) peak of Na copolymerized catalysts became broader at the fixed position, indicating the continuous decrease of crystal size of 27.9, 9.1 and 3.2\u00a0nm for pristine GCN and Na-GCN-0.1-copol, 0.25-copol, respectively, with the constant interlayer distance of 0.32\u00a0nm. On the other hand for Na/GCN-imp catalyst, another peak at 25.2\u00b0 was additionally observed, which could be resulted from the intercalated Na present between GCN layers [52] with the interlayer distance of 0.35\u00a0nm. Due to the structural modification, the BET surface area tended to decrease with the addition of Na. Pristine GCN showed BET surface area of 6.9\u00a0m2/g while that of Na-GCN-0.1-copol, 0.25-copol and Na/GCN-imp decreased to 3.8, 0.7 and 3.29\u00a0m2/g, respectively.The chemical bonding of pristine GCN and Na introduced catalysts was further investigated using XPS analysis as presented in Table 2\n and Fig. 3\n. The total amount of Na in the catalysts was found to be 3.5, 12.8 and 8.5\u00a0wt% for Na-GCN-0.1-copol, 0.25-copol and Na/GCN-imp, respectively. According to the previous work on literature [53], peaks at 1071\u00a0eV and 1072\u00a0eV from Na1s XPS spectra could be assigned to Na-N and Na-O bonds, respectively. As summarized in Table 2, the content of Na-N and Na-O species present in fresh catalysts was found to be 2.0, 1.5\u00a0wt% and 8.6, 4.2\u00a0wt% and 4.9, 3.6\u00a0wt% in Na-GCN-0.1-copol, Na-GCN-0.25-copol and the Na/GCN-imp, respectively. In the case of used catalysts, the residual Na was detected mainly as Na-N with the complete loss of Na-O. The Na-N content of the used catalysts was preserved to be 8.3 and 3.8\u00a0wt% for Na-GCN-0.25-copol and the Na/GCN-imp.Generally, the growth of GCN crystals is known to proceed with the bonding between strands of \u2013NH2 groups on melamine-derivative intermediate species such as melam, melem, and melon [54] as illustrated in Scheme 1\n.During the synthesis of Na copolymerized catalysts, the introduced NaOH was first dehydrogenated to Na2O, as shown in Eq.(4). The formed Na2O moiety might eliminate the NH2 groups and form Na-N bond as described in Eq. (5).\n\n(4)\nNaOH\u00a0\u2192\u00a0Na2O\u00a0+\u00a0H2O\n\n\n\n\n(5)\nNa2O\u00a0+\u00a0\u2013NH2\u00a0+\u00a01/2 O2\u00a0\u2192\u00a02Na\u2013N\u00a0+\u00a02 H2O\n\n\nConsidering that the polymerization and crystal growth started from the condensation of NH2 species, the loss of those groups provoked the inhibition of further polymerization [52]. This phenomenon could explain the withdrawal of amine groups and the suppression of crystal size growth with Na content as previously discussed with the FTIR and XRD results. When the content of NaOH was further increased up to 50\u00a0wt% in the precursors (NaOH\u00a0+\u00a0melamine) mixture, the formation of GCN structure was totally inhibited generating only sodium cyanate (NaOCN) as shown in Fig. S1. With the proper amount of Na and melamine under mild oxidation condition, Na2O might be bonded to the NH2 strands in the intermediate species, generating NaN bonds, as described in Eq.(5). On the Na-GCN-0.1-copol and 0.25-copol, the tri-s-triazine and stacking structure were preserved as confirmed from XRD analysis results, although the elimination of NH2 species could terminate the polymerization steps generating smaller GCN crystals. This hypothesis explained the opposite inclination between the crystal size of GCN and the amount of Na in the Na co-polymerized GCN catalysts.However, during the preparation of Na/GCN-imp, NaOH solution diffused into the pre-existing stacked GCN layers with the aid of sonication. During the calcination, introduced NaOH could be converted into Na-N by forming bonding with amine group as described in Eq.(5) or coordinated with the pyridinic N atoms in the tri-s-triazine hole of the GCN plane [52]. The tri-s-triazine ring along the GCN layers could be opened through the Na introduction, which was previously confirmed by the disappearance of (100) diffraction pattern as shown in Fig. 2. The uncoordinated Na atoms located in the outer plane and between interlayers could remain as Na2O species and enlarge the interlayer distance because of the larger atomic radius of Na (\u223c100\u00a0pm) than N and C (70 and 60\u00a0pm) [53,55].The inherent basicity of Na-N species was further characterized using CO2-DRIFT analysis as shown on Fig. 4\n. Liu et al. reported the hydrogenation of CO2 to formate over a Schiff base mediated gold nanocatalyst. CO2 was adsorbed to the Schiff base forming the carbamate (NCOO\u2013) zwitterion, which could be measured at 1712\u00a0cm\u22121 on CO2-DRIFT analysis [56]. Similarly, in the present study, the catalysts showed adsorption peak between 1687 and 1706\u00a0cm\u22121, as shown in the Fig. 4. The basicity of the catalysts could be confirmed from the formation of carbamate species.To better understand the enhanced basicity of the Na modified GCN catalyst and the possible effect on biodiesel production, density functional theory (DFT) calculations were performed. For this purpose we examined the methanol decomposition into methoxy anion and proton (CH3OH\u00a0\u2192\u00a0CH3O\u2013 + H+) on the heterogeneous catalyst surface of pristine GCN and Na modified GCN catalysts, which is believed to be the first step of the transesterification reaction [57\u201359]. As shown in Table 3\nand Fig. S3, the reaction energy of the Na modified GCN catalyst was found to be lower than that of the pristine GCN, suggesting that the Na modified GCN catalyst had higher activity on biodiesel production. This enhancement was related to electron charge redistribution by the addition of Na to GCN. According to Bader charge analysis, the Na atom (\u0394\u03c3 = +0.825e) lost electrons to the neighboring N atoms (\u0394\u03c3\u00a0=\u00a0\u00a0\u2212\u00a01.126e), making Na and N atoms Lewis acid and base sites, respectively. This led to the increase of binding strength of CH3O\u2013 (Ead\u00a0=\u00a0\u00a0\u2212\u00a05.87\u00a0eV) and H+ (Ead\u00a0=\u00a0\u00a0\u2212\u00a06.64\u00a0eV) on the surface of the Na modified GCN catalyst compared to the pristine GCN one (Ead\u00a0=\u00a0\u00a0\u2212\u00a04.54\u00a0eV for CH3O\u2013 and\u00a0\u2212\u00a03.30\u00a0eV for H+) and, in turn, boosted the CH3OH dissociation reaction.The transesterification of soybean oil with methanol using GCN catalysts was carried out under identical reaction conditions which was presented in Fig. 5\n. The pristine GCN had almost no transesterification activity under our reaction conditions due to low basicity. Na-GCN-0.1-copol, \u22120.25-copol and Na/GCN-imp showed biodiesel yield of 26.8, 90.6 and 46.1\u00a0%, respectively at the first batch run. As mentioned in Section 3.2, the base catalyst is known to activate methanol by the cleavage into the methoxide anion (CH3O\u2013) and proton (H+). The methoxide ion, a strong base, attacks the carbonyl carbon of triglycerides, producing tetrahedral alkoxy carbonyl intermediates. CO cleavage on the tetrahedral intermediate yields methyl esters and diglycerides, respectively. Subsequently, diglyceride is further converted to monoglyceride through the nucleophilic attack of the methoxide ion, eventually producing three moles of methyl esters and one mole of glycerol. The generation of methoxide anion by base catalysts from methanol is known to be the rate-determining step, which can be accelerated with the sufficient basicity of the catalysts. The strong basic sites in the Na-GCN-0.25-copol could be expected to effectively promote the generation of methoxide anions.The catalyst reusability and stability were further investigated under 10 repeated cycle tests, as shown in Fig. 6\n. The Na-GCN-0.25-copol showed constant activity over 90\u00a0% biodiesel yield under 10 cycles. From the deconvolution of Na1s XPS spectra for the used catalysts as shown in Table 2, the total amount of Na in Na-GCN-0.25-copol decreased from 12.8\u00a0% to 8.3\u00a0wt%, while Na/GCN-imp from 8.5 to 4.0\u00a0wt%. Interestingly, Na-O species present in the fresh catalysts was almost not detected in the used ones [16]. Considering the catalyst preparation condition, the introduced NaOH was believed to be converted to Na2O as previously mentioned in Eq.(4). Since the amount of Na2O measured by XPS analysis was less than 3\u00a0%, distinctive diffraction pattern of Na2O crystallite was not observed for all catalysts. However, Na2O is known to be a strong base to activate methanol generating sodium methoxide (CH3O\u2013 Na+) as Eq.(6), which will further attack the triglyceride molecule during the biodiesel production reaction.\n\n(6)\nNa2O\u00a0+\u00a02 CH3OH\u00a0\u2192\u00a02 CH3O-Na+ + H2O\n\n\nAs the introduced Na was transformed to ionic Na dissolved into reaction medium, Na2O species could be easily leached out from the catalyst surface [60]. From the ICP analysis of the liquid medium after the first batch run of Na-GCN-0.25-copol, a considerable amount of Na was detected, which corresponded to 94\u00a0% of its initial Na-O content in the fresh one. As a result, Na2O species contributed as a pseudo-homogenous catalyst only at the first batch run. This phenomenon was dominant on the Na/GCN-imp with relatively high Na2O content. The presence of a high amount of Na cation in the product fuel might cause metal corrosion as well as saponification of the biodiesel phase.However, the amount of Na-N species in the used catalysts was preserved as shown in Table 2. 96.5 and 77.6\u00a0% of Na-N content in fresh Na-GCN-0.25-copol and Na/GCN-imp were retained on the used catalysts. This result asserted the high stability of Na-N species at the current reaction conditions. As mentioned in section 3.2, the introduced Na enhanced the electron charge dislocation of the GCN structure forming the coordinate covalent bond with the neighboring N atoms. The chemical bond between Na and N atoms not only increased essentially the basicity of the N atoms but also granted the resistance against leaching of the alkali metals toward the liquid reaction medium, which was reported as the main obstacle in developing active and stable heterogeneous catalyst for commercial biodiesel production.Taking all phenomena into account, both Na2O and Na-N species were presumed to activate the biodiesel production reaction on the first batch run. As Na2O was leached out to the reaction medium and removed from the catalyst surface, Na-N species was the only available basic site for repeated batch cycles.Discriminatively for Na/GCN-imp catalyst, since the stacked structure of GCN layers were retained, Na-N species generated inside the tri-s-triazine hole in inner layers were assumed to be unaccessable as the bulky TG molecules might not be able to diffuse within the layers spacing of 0.35\u00a0nm. As a result, both fresh and used Na/GCN-imp catalysts showed low yield of biodiesel considering the relatively large amount of present Na-N species. Hence, rapid deactivation of Na/GCN-imp catalyst over repeated batch cycles was observed as Na2O species were not available due to leaching, eventhough considerable amount of Na-N species were present.Herein, a novel and economic Na-modified GCN catalyst with enhanced basicity was firstly fabricated and applied to the transesterification of soybean oil. Copolymerization of melamine and NaOH was found to be effective in generating stable and active basic heterogeneous catalysts. The Na atoms were believed to transfer electrons forming rigid bond with neighboring N atoms, which enhanced the Lewis basicity and the stability of the Na-GCN-copol catalysts. While the majority of Na species in impregnated catalysts existed in the form of Na-O, which was easily leached out to the reaction medium showing rapid catalyst deactivation over the repeated cycles. As a matter of fact, the Na-GCN-copol catalyst maintained over 90\u00a0% of biodiesel yield during 10 repeated cycles with the aid of its leaching resistance property.\nSung Eun Kim: Methodology, Data curation, Writing \u2013 original draft, Formal analysis. Ji Hu Kim: Formal analysis, Methodology. Deog Keun Kim: Supervision, Project administration. Hyung Chul Ham: Supervision, Data curation, Writing \u2013 original draft. Kwan-Young Lee: Supervision. Hak Joo Kim: Writing \u2013 review & editing, Conceptualization, Project administration.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was financially supported by a Grant-in-Aid from the National Research Council of Science & Technology and Korea Institute of Energy Research (Project C2-2431, Development of hetero-catalytic system for multi-feedstock biodiesel and platform chemical production).Supplementary data to this article can be found online at https://doi.org/10.1016/j.fuel.2023.127548.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  Na-modified graphitic carbon nitrides were utilized for the transesterification of soybean oil and methanol. Graphitic carbon nitrides have not yet been widely applied in biodiesel production, despite their chemical stability and basicity. The catalysts were obtained via the co-thermal polymerization of NaOH and melamine. Catalyst prepared using conventional impregnation method was applied for comparison. The copolymerized catalyst with the optimum Na content showed over 90% biodiesel yield for 10 repeated cycles. The prepared catalysts were characterized by Fourier transform infrared spectrometry, X-ray diffraction, scanning electron microscopy/energy dispersive X-ray spectroscopy, CO2-diffuse reflectance infrared Fourier transform spectroscopy, and X-ray photoelectron spectroscopy. The elaborate rigid bonds between Na and N contributed to the leaching-resistance and catalytic activity. While the majority of Na species in impregnated catalysts existed in the form of Na-O, which was easily leached out to the reaction medium showing rapid catalyst deactivation over the repeated cycles. The basicity derived from the electron transfer from the Na to N atoms was confirmed from the density functional theory and CO2-diffuse reflectance infrared Fourier transform spectroscopy.\n               "}
{"full_text": "Fast pyrolysis of biomass with a rapid heating rate (>500\u00a0\u00b0C/s) to intermediate temperatures (400\u2013600\u00a0\u00b0C) is a promising way to generate bio-oil from the fast decomposition of biomass in the absence of oxygen, by which the short vapor residence time can lead to high bio-oil yield with less other products like gas and solid char [1,2]. For example, flash pyrolysis or very fast pyrolysis with rapid heating rate (>1000\u00a0\u00b0C/s) has been proven to provide high yield of bio-oil as well as high conversion efficiency even over 70% [3]. However, the obtained bio-oil obtained from the fast pyrolysis always has a dark brown appearance with a distinctive smoky smell. The physical properties of bio-oils have been reported by many researchers [4,5], which are determined by the chemical compositions in the pyrolysis bio-oils. Usually, there are several hundreds of organic compounds in the raw bio-oils, which mainly include phenols, acids, aldehydes, alcohols, esters, ketones, and other macromolecules and can be classified to three major groups: (I) small carbonyl compounds including carboxylic acids, hydroxyaldehydes, hydroxyketones, acetone, acetaldehyde and acetic acids; (II) sugar-derived compounds including levoglucosan, furfural, anhydrosugars and furan/pyran ring-containing compounds; and (III) lignin-derived compounds, majorly including guaiacols and phenols [6]. Besides, oligomers with molecular weights in a range of 900\u20132500 also exist in the bio-oil with a large amount [7,8]. These compounds distribution is mainly determined by biomass type and pyrolysis route, which is related to the physicochemical properties of bio-oil [9,10]. The basic properties of raw bio-oil and the petroleum fuel oil are compared in Table\u00a01\n [11,12]. Obviously, the bio-oil contains much more oxygen and H2O contents, which leads to lower heating value. Higher heating values (HHV, MJ/kg) of the pyrolysis bio-oils from wood are usually ranged from 16 to 19\u00a0MJ/kg, which are just about half of the petroleum fuel oil (40\u00a0MJ/kg) even in the highest case. It can be concluded that the low quality of raw bio-oil is resulted from its high oxygen and H2O contents with some solids components, high viscosity, low pH value, thermal instability with poor combustion property [13]. Especially, the high oxygen content is the main reason for its low heating value. In addition, the unsaturated components such as phenols and aldehydes in it are unstable, which can easily transform into macromolecules through polymerization, particularly in those acid conditions, increasing the viscosity and reducing liquidity. Thus, the application of it is still limited by these characterizations. However, despite these shortcomings, the bio-oil also behaves some advantages like less toxicity, easier biodegradation and better lubricity than the petroleum fuel. Hence, it is desired to improve the quality of the bio-oil so that it can replace fossil fuels in the future [14].Generally, there are two types of operation modes for the bio-oil upgrading [2,15\u201317]. One is in situ catalytic pyrolysis, in which the biomass and catalysts are thoroughly mixed. In this case, the catalysts are in situ exposed to the pyrolysis vapor, where the pyrolysis vapor diffuses promptly into the catalyst pore, undergoing a series of reaction processes including cracking, deoxygenation, aromatization and condensation [2,15,16]. However, only the pyrolysis vapor passes through the catalyst layer, which is separated from the biomass on the upside by quartz wool. The name given to this process is in situ catalytic upgrading. The schematic diagram of the experimental setup is shown in Fig.\u00a01.\nThe other mode is ex situ upgrading, in which biomass and catalysts are located separately in one reactor with two reaction zones or the biomass pyrolysis and catalytic bio-oil upgrading are performed in two reactors separately [17]. In this case, the temperatures for biomass pyrolysis and catalytic bio-oil upgrading can be regulated individually to achieve the best operation conditions for the two processes, which allows for well control of product distribution and selectivity.Cracking of bio-oils over porous solid catalysts such as zeolite-based catalysts at ambient pressure is considered one of effective ways for the bio-oil upgrading, especially in which hydrogen gas is not necessary. The key for the catalytic cracking of bio-oil is the development of high-performance catalysts. Herein, zeolite-based catalysts for the upgrading of pyrolysis bio-oils are critically reviewed. The effects of porous structure, acidity and other parameters including biomass type, catalyst amount and reaction temperature on cracking activity, selectivity, stability and deactivation are summarized. While, the proposed mechanisms on the bio-oil upgrading over the zeolite-based catalysts for raising the contents of hydrocarbons like benzene, toluene and xylenes (BTXs) and hindering the generation of those by-products like coke and polyaromatics are discussed. Furthermore, the main strategies such as metal modification, construction of zeolites with a hierarchical structure and synthesis of special morphologies with hollow structure or core/shell structure for the improvement of deoxygenation property performance are introduced. It is expected to provide a guidance for the design and fabricate more excellent zeolite-based catalysts and their application for production of high-quality bio-oil from the fast pyrolysis of biomass.Zeolites are microporous aluminosilicate solids named as \u201cmolecular sieves\u201d, which can accommodate various cations such as Na+, K+, Ca2+ and Mg2+ and adsorb those molecules with comparable sizes corresponding to their pore window sizes. Zeolites have been widely used as commercial ion-exchange materials, adsorbents and catalysts. Zeolites have been applied to catalyze various reactions owing to their special acidity-basicity as well as shape selectivity. Herein, the shape selectivity could be resulted from either the transition state effect or mass transfer [18]. The various micropore sizes (0.5\u20131.2\u00a0nm) can affect the mass transfer, thereby excluding certain reactant molecules and limiting the formation of products larger than the zeolite pore size. While, the confined space in the zeolite pore can restrict certain transition states, thereby influencing the reaction routes. In addition, zeolites also have \u201csolvent effect\u201d or \u201cconfinement effect\u201d, where some reactants inside the zeolite pore have higher concentrations than those outside the pores [18]. To date, various zeolites including Beta zeolite, Y zeolite and SSZ-55 with large pores, ZSM-5, ZSM-23, ZSM-11, IM-5, TNU-9, Ferrierite with medium pores and ZK-5, SAPO-34 with small pores have been investigated for the bio-oil upgrading [18]. It is found that most of these zeolites could enhance the formation of aromatics during the upgrading process, and some of them, especially the protonated ZSM-5 (HZSM-5) as well as Beta zeolite (H\u03b2) always gave higher aromatics yields. Since the effects of various parameters for the bio-oil upgrading over various zeolite catalysts are similar, herein, the bio-oil upgrading over HZSM-5 is mainly reviewed.HZSM-5 is a microporous aluminosilicate zeolite with characteristics suitable for aromatics and olefins production in the petrochemical industries due to its adequate combination of acid strength and shape selectivity. Similarly, HZSM-5 based catalysts have been widely employed in the upgrading of bio-oil obtained from the biomass pyrolysis by enhancing those reactions relating to deoxygenation including decarboxylation, decarbonylation, dehydration, isomerization, and aromatization [19]. In our previous study [20], in-situ catalytic upgrading of bio-oil derived from fast pyrolysis of lignin over different zeolite catalysts was investigated, and found that HZSM-5 was more active for the improvement of bio-oil quality in terms of the highest selectivity towards monoaromatic hydrocarbons. Simultaneously, the utilization of ZSM-5 resulted in the highest yield of light oil and the lowest yield of coke among all the applied zeolite catalysts. Engtrakul et\u00a0al. [21] studied the catalytic pyrolysis of pine wood biomass in a fluidized bed reactor at 450\u00a0\u00b0C utilizing various zeolite catalysts including Beta, Y, ZSM-5, and Mordenite. By the using of ZSM-5, lower acid and alcohol contents were contained in the liquid products. While, coke deposition on ZSM-5 appeared to be lower than that on other zeolites. It is considered that adequate balance between acid strength and shape selectivity of ZSM-5 should be beneficial for the conversion of biomass-derived oxygenates into aromatic hydrocarbons.The acid strength and acid site density of HZSM-5 catalysts always vary with the SiO2/Al2O3 ratio, which can determine the catalytic activity, deactivation, and product distribution [21]. There are two types of acid sites on HZSM-5, i.e., Br\u00f8nsted acid site and Lewis acid site. High acidity with low Si/Al ratio is controlled especially by the Br\u00f8nsted acid site, which becomes more active in the cracking process, leading to the production of more aromatics such as benzene, toluene and xylene (BTX) and light olefins with the reducing of heavy oil fraction [22,23]. However, the higher acidity will result in further secondary reactions to produce more polyaromatics and even form coke on the zeolite surface, thereby causing the deactivation of the catalyst. Polyaromatic hydrocarbons (PAH) have been reported as the precursor of coke. With the decrease in the SiO2/Al2O3 molar ratio of zeolite, more polyaromatic hydrocarbons (PAH) and water will be generated with the reduced amount of light bio-oil yield [22,24]. Thus, proper adjustment of acidity to make HZSM-5 suitable for the aromatic production has been proposed by using several techniques such as ion exchange, desilication, and dealumination [25\u201327.Surface area is one of main factors affecting catalytic efficiency. Typically, various HZSM-5 zeolites have specific BET surface areas in the range of 350\u2013450\u00a0m2/g [28]. The surface area is strongly correlated with the acidity, thereby affecting catalytic activity. In general, ZSM-5 is composed of the external surface and internal surface. The external surface area corresponds to the number of the entrances to the pores, which determines the effective internal surface area, and acidic sites existing on the surface [29]. The selectively chemical reactions catalyzed by zeolites always tend to occur within the pores, and thus the pore dimension and structure of zeolite have shape selectivity to the product. For bio-oil upgrading, catalytic cracking of large molecules tends to occur on the acid sites of the external surface, where the acid sites on the internal surface cannot effectively catalyze macromolecular cracking due to the pore size limitation [30]. In this case, the surface area is related to the formation of coke on both the exterior and internal surfaces of the zeolite and especially, the production of coke on the interior surface could lead to a greater deactivation rate of zeolite by the covering of acidic surface and blocking of pores. Therefore, zeolite catalysts with a smaller outer surface area but a larger internal surface area may have a higher rate of deactivation [29]. On the other hand, those large molecules in the bio-oil can readily diffuse into the pore and access more active sites within the pores when the surface area and pore size improve simultaneously. As a result, they can be rapidly converted to hydrocarbons and consequently reducing the formation of coke. Especially, HZSM-5 catalyst with a high surface area could be more easily modified by metal loading with less significantly reduction of the surface area [31]. In addition, a large surface area could delay the deactivation of HZSM-5 catalysts caused by coke formation and deposition on the exterior surface. It should be noted that the coke deposition within the zeolite pores could inhibit the capillary and diffusion flow of reactants, finally lowering the aromatics yielding reaction rate [32]. Thus, it is better to avoid the occurring of the coke deposition within the zeolite pores. The HZSM-5 catalyst has a higher surface area and a larger pore size, with a shorter diffusion length, resulting in reactions over it, such as biomass depolymerization to produce more monocyclic aromatics like BTX with less coke. It is reported that the nanosheet ZSM-5 catalysts had better textural properties with higher BET surface area, larger mesopore volume, and higher concentration of external Br\u00f8nsted acid sites than the conventional HZSM-5, thereby resulting in better performance for catalytic cracking reactions [33].The mass transfer ability of bulky reactants and products to pass through the micropore in the zeolite is determined by the pore size, thereby significantly affecting the catalytic activity. The chemical reactions catalyzed by zeolites are thought to occur mainly within the internal pore of the zeolite. As a result, a reactant with a molecule size larger than that of the zeolite pore cannot diffuse into the zeolite pore. Similarly, the product formed within the zeolite pores cannot be larger than the pore size. Therefore, the pore size and structure of the zeolite have a substantial influence on the production of products in the pores and the diffusion of products from the pores. HZSM-5 has micropores consisting of two intersecting three-dimensional channels of 10-membered rings with one straight channels (5.1\u00a0\u00c5\u00a0\u00d7\u00a05.5\u00a0\u00c5) and sinusoidal channels (5.3\u00a0\u00c5\u00a0\u00d7\u00a05.6\u00a0\u00c5) [34,35]. When the cracking reaction occurs within the pore, if the formed product can only diffuse slowly out of the pore, a small amount of product will be obtained. For those molecules in the raw bio-oil larger than the pore size of HZSM-5 such as levoglucosan and 5-hydroxymethyl furfural, they could be more likely converted to coke outside the pores since they cannot enter the pores [36]. Benzene, toluene, indene, ethylbenzene, and p-xylene have been reported to be pyrolyzed to intermediate intermediates that can easily diffuse into the pore due to the typical kinetic diameter of HZSM-5 [18]. Thus, the mass transfer will be improved by increasing the pore size in the HZSM-5 based catalysts [37,38]. However, it has also been reported that the zeolites with large pores have low stability, leading to the generation of undesirable compounds, which is also the primary cause of the deactivation [12].Zeolite is a crystalline substance with a structure characterized by a framework of linked tetrahedra (i.e., AlO4 and SiO4), each consisting of four O atoms surrounding a cation. This framework contains open cavities in the form of channels and cages which are large enough to allow the passage of those suitable guest species [39]. In general, there are several zeolite structures which are classified according to pore diameter and ring size and are given different names as shown in Table\u00a02\n. These dimensional properties are important for their role in determining aromatic yield and especially, the \"n\" membered ring shown in Fig.\u00a02\n can determine the zeolite structure and pore shape.The pore shape of zeolite is one factor that greatly influences aromatics production.. Micro-pore zeolites such as SAPO-34 with a CHA structure connected by 8-ring channels always cannot produce any aromatics but generate CO, CO2 and coke, which show the performances similar to pyrolysis bio-oil derived by non-catalyst pyrolysis. In contrast, those zeolites such as ZSM-5 and ZSM-11 with medium pore sizes are very suitable for biomass conversion into aromatics because they have moderate pore openings (0.5\u20130.6\u00a0nm) which are favor of aromatics production [40]. However, although ZSM-5 and ZSM-11 have similar framework and pore size, they also have some differences. ZSM-5 zeolite has an MFI crystal structure consisting of two perpendicularly intersecting channels of 10-membered rings (straight channels and zigzag or sinusoidal channels) (Fig.\u00a03\n) [41,42]. These channels are connected at right angles to straight channels at zigzag angles to form 3 intersections per unit cell. Whereas, ZSM-11 has only straight channels along a- and b-axis, the absence of opening channels along the c-axis in ZSM-11 limits reactant molecules entering into the channels [43]. Thus, these two zeolites with similar physicochemical properties in pore size and acidity display different molecular shape-selective properties due to their difference in channel tortuosity. Jae et\u00a0al. [18] studied the conversion of glucose to aromatics over various zeolites such as SAPO-34, ZSM-11, ZSM-5, ferrierite, Beta and Y zeolites with different pore shapes and sizes. The results confirmed that ZSM-5 provided the highest yield of aromatics with the lowest coke formation.On the other hand, those zeolites with larger pores such as BEA and FAU (beta, X, Y zeolite) have a three-dimensional (3D) pore system of 12- and 12-membered ring channels with increased accessibility of more intermediates generated from biomass pyrolysis to the pores. However, the secondary reactions in the pores could cause pore blockage and coke deposition, consequently resulting in low aromatic yield [44]. It is reported that among zeolite catalysts of H-\u03b2, HY, H-USY and HZSM-5, HZSM-5 is the most effective catalyst in promoting the yield of aromatics for the upgrading of bio-oil from lignin due to its well-balanced acidity and shape selectivity [42]. As stated above, the conversion of biomass-derived oxygenates to aromatics over the zeolite catalysts depends on the reactants entering the pore, converting in the pore, and pore shape selection of the generated products to diffuse out of the zeolite pores [45]. Thus, poor shape selectivity and diffusion limitation could hinder the reactants to access the catalytic active sites of the zeolite, which is a common issue with zeolite catalysts for the desired products formation in terms of quantity and quality. That is, only the oxygenated components in the pyrolysis bio-oil captured by the narrow pores of HZSM-5 can convert to aromatic hydrocarbons due to their shape selectivity.Generally, the upgrading of bio-oil could be influenced by not only the properties of the ZSM-5 based catalysts in terms of shape selectivity, pore size, surface area and acidity, but also the biomass type, catalyst to biomass ratio and reaction temperature.The kind of biomass has great impact on the compositions of the raw bio-oil since different kinds of biomass has different compositions of cellulose, hemicellulose, lignin, ash contents with different inorganic minerals especially alkali and alkaline earth metal (AAEM) species (Table\u00a03\n). Biomass can be classified into hardwood (e.g., oak, beech), softwood (e.g., pine, cedar, corn stalk), and grasses (e.g., barley straw, bagasse). In general, softwood contains more lignin than hardwood while grass biomass contains less lignin than woody biomass but has a higher ash content [46]. The decompositions of cellulose and hemicellulose occur at relatively lower temperatures than that of lignin. Lignin is a more stable component, which normally begins to decompose at a temperature above 200\u00a0\u00b0C and leaves 40% of residual solid product at the end of pyrolysis. In addition, lignin is known to contain phenolic compounds, which is the most abundant source of aromatic hydrocarbons produced from the biomass [47]. Thus, it should be more suitable to select biomass with more lignin for the production of bio-oils with more aromatic hydrocarbon contents when using HZSM-5 based catalysts. Terry et\u00a0al. [48] reviewed bio-oil production from pyrolysis of oil palm biomass by using the different parts of oil palm. It is found that the different bio-oil compositions obtained from the different parts of oil palm. For example, oil palm trunk (OPT) has relatively high cellulose and lignin contents, which contributes to the formations of acids and phenols. While, palm kernel shell (PKS) has a higher lignin content, which contributes to the formation of phenolic compounds. Especially, with the catalysis of HZSM-5, those compounds such as acids and phenols will be converted into aromatic compounds.For the upgrading of bio-oils from in-situ pyrolysis of biomass, it is always critical to choose a suitable catalyst-to-biomass ratio for the upgrading of bio-oil. When the proportion of the catalyst used is too small, the in-situ generated raw bio-oil cannot completely contact the catalyst. Especially, a part of catalysts could have already deactivated during the initial stage of pyrolysis by cracking of previous pyrolysis vapor. As such, the obtained bio-oil will consist of both upgraded oil and non-upgraded oil, resulting in under-desired bio-oil. Thus, it is important to optimize the catalyst-to-biomass ratio during the bio-oil upgrading process.The total products obtained from the pyrolysis process are heavily influenced by the reaction temperature. Pyrolysis is typically performed at temperatures in the range of 350\u2013650\u00a0\u00b0C. At a lower temperature (<350\u00a0\u00b0C), it is always difficult to completely devolatilized the volatile compounds in the biomass. Moreover, a certain increase in vapor residence time as the temperature rises from 350 to 500\u00a0\u00b0C will result in a higher bio-oil yield. As the reaction temperature rises, more bio-oil will be generated. However, with an increase in temperature over 650\u00a0\u00b0C, some secondary reactions also occur, which will decrease the generation of more bio-oil since the secondary reactions predominate with the continuous increasing of gaseous products as shown in Fig.\u00a04\n\n[57]\nThe stability of HZSM-5 catalysts is of great importance to their performances. The deactivation of HZSM-5 catalysts is always caused by coke-induced blockage of reactants and products on active sites and/or the sintering of active species on the catalyst surface [58]. Thus, the high-performance HZSM-5-based catalysts must be durable and capable of preventing self-accumulating coke or sintering, or even in the case that coke deposition or sintering does occur, it has little effect on the catalytic activity and still has good catalytic efficiency.During catalytic bio-oil upgrading, various reactions occur, of which the coke formation is mainly resulted from the polymerization of aromatics and olefin, leading to the blockage of pore opening and finally the deactivation of HZSM-5 catalysts [59]. Herein, the acidity affects not only the catalytic activity but also the deactivation by the promoting of coke deposition. The HZSM-5-based catalysts with high stability could be achieved by tuning the acid sites by adjusting the aluminum speciation on the external surface [60,61]. For example, dealumination from HZSM-5 structure can be realized by treatment with acid solutions (e.g., HCl, HNO3 and HF solutions) [60,62,63]. The dealumination can decrease Br\u00f8nsted acidity, resulting in slower polymerization and less coke formation [63]. While, the introduction of mesopore into the structure of ZSM-5 catalysts by post-treatment with alkaline solution is also one way to reduce the coke formation. In this case, the bulk Si/Al ratio, micropore volume, and crystallinity can be decreased while the Lewis acidity is increased to improve the coking resistance [64].For those active metal modified HZSM-5 based catalysts, the agglomeration of metal species on the HZSM-5 surface, i.e., catalysis species sintering, during the upgrading process is always one of the reasons for the deactivation. It is reported that some noble metal species doped on the HZSM-5 catalysts are easily accumulated, causing the deactivation since the sintering results in the losing of active sites and blockage of the zeolite pores [63]. Especially, excessive amount of metal loading and low metal species dispersion could lead to the high accumulation of metal species during the reaction, leading to the decreases in mass transfer as well as catalytic activity.There are three common types of deoxygenation reactions over the zeolite catalysts: dehydration, decarboxylation and decarbonylation. The bio-oil always contains a considerable number of components with -OH group. The dehydration is a process to remove oxygen from those oxygenated compounds in the form of water. When the dehydration occurs on the zeolite catalysts, Br\u00f8nsted acid sites play the primary role, that is, they can donate protons to the hydroxyl group of oxygenates to generate water. The decarboxylation is a process to remove oxygen from those fatty acids and fatty acid methyl esters in the form of carbon dioxide. The decarbonylation is a process to remove oxygen in the form of carbon monoxide, in which the carbonyl groups can be removed from those aldehydes and ketone compounds. Either the decarboxylation or the decarbonylation could be affected by the acidity of zeolite. Meanwhile, some carbon and hydrogen could be lost by coke formation and/or the generation of gaseous hydrocarbons by the bio-oil vapor cracking over the zeolite catalysts [65]. There are also other reactions involved in the upgrading of bio-oils, such as cracking, aldol condensation, ketonization and aromatization. Such a combined complex upgrading process will finally achieve the conversion of oxygenated compounds to hydrocarbons, thereby enhancing bio-oil quality [66].While, the aromatization on the zeolite catalysts also occurs, by which those olefins and low molecular weight oxygenates (e.g., acids, aldehydes, alcohols, esters, furans and ethers) will convert to aromatic hydrocarbons. During the aromatization process, other reactions may occur simultaneously, such as cracking, dehydrogenation, oligomerization and cyclization. Aromatization generally takes place within the pores of zeolites. Two typical examples of aromatization are: the combination of propylene and furan to produce toluene or the reaction of benzene with furan to the form naphthalene via Diels-Alder condensation reaction [32]. Fig.\u00a05\n shows the possible reactions in bio-oil upgrading over catalysts.\nFig.\u00a06\n shows a possible reaction network for upgrading of bio-oil derived from the pyrolysis of biomass over the catalysts [67]. In order to achieve a good upgraded products distribution, it is important to choose a proper catalyst with excellent selectivity and high activity. During the pyrolysis of cellulose and hemicellulose, anhydrosugars are firstly generated by the depolymerized with the deoxygenation in the forms of CO2, CO and H2O by cracking and dehydration. Then, these generated sugars are dehydrated and re-arranged, resulting in the formation of furans and some small oxygenated species [68]. These intermediates can be further deoxygenated to hydrocarbons with the assistance of catalyst. While, for the pyrolysis of lignin, those phenol alkoxy compounds can be generated by cracking, dehydration and depolymerization [69]. Here, it should be noted that the char is always more easily produced from the pyrolysis of lignin than that from cellulose or hemicellulose. Similarly, those oxygenated compounds from the pyrolysis of lignin can be deoxygenated on the active sites of catalysts. Herein, stronger acidity is more beneficial for the cleavages of C-C and C-O bonds prior to deoxygenation, thereby increasing the generation of more small hydrocarbons like BTXs. On the other hand, further polymerization and aromatization of those small hydrocarbons could occur in the presence of catalysts with higher acidity, leading to the coke formation on the active sites. Therefore, it is important to adjust the acidity of catalysts in order to enhance the performance and avoid coke formation.Despite that HZSM-5 is the most popular zeolite among other types for deoxygenation of bio-oils to make aromatics since it has suitable acidity, excellent heat tolerance, strong selective cracking ability with well isomerization property, the small micropore structure of the ZSM-5 restricts the mass transfers of the reactant and product in the pore, which makes carbon deposition and catalyst deactivation easier. More crucially, the location of a large amount of acid sites inside the pore affected catalytic efficiency, always resulting in a lower yield of the desired product. To resolve these issues, the parent HZSM-5 can be restructured by introducing relatively larger mesopore or added new active sites by metal loading to increase the resistance to carbon deposition. Moreover, great efforts have been made to modify the Si/Al ratio framework by dealumination/alumination or desilication as it is directly related to the catalytic performance.Although the parent HZSM-5 is the most efficient catalyst for the bio-oil upgrading, the coking rate over it is also quite high, leading to a significant deactivation problem. As such, the aromatics generation rate is always significantly low. Thus, the parent HZSM-5 catalyst needs to be further modified to improve the upgrading performance. Currently, the performance of parent HZSM-5 catalyst is usually improved by metal loading, which is an easy way not only for the catalyst performance improving but also the coking resistance due to the simple preparation procedure and high ability to alter the acidity of HZSM-5 for achieving the optimal upgrading result [70]. To date, the types of doping metal have been widely investigated. By doping of transition metal or noble metal on HZSM-5, it is feasible to boost deoxygenation capacity and produce more carbon oxides with less water, resulting in more hydrogen available for incorporation into hydrocarbons [71]. While, alkaline earth metals such as Mg and Ca doping on HZSM-5 can behave as bases and their metal cations could function as Lewis acid sites, which allow tailoring the zeolite activity to avoid excessive cracking of the bio-oil, and in turn result in a higher yield of the deoxygenated compounds in the upgraded bio-oil with the decreasing of the formation of undesired polyaromatic hydrocarbons and coke [65].Transition metal modified HZSM-5 catalysts for bio-oil upgrading have been widely reported. For example, Ni- and Co-modified HZSM-5 catalysts especially the HZSM-5 modified by Ni can effectively increase the aromatic hydrocarbons content in the upgraded bio-oil [71]. While, the HZSM-5 modified by 2\u00a0wt% of Zn resulted in the increase in the strong acid site content, thereby increasing the BTX yield during the bio-oil upgrading process. However, further doping of Zn (e.g.,10\u00a0wt.%) reduced the acidity and physical characteristics of the catalyst, leading to poor reactant and product diffusions in the zeolite pore, thereby reducing the BTX yield [72]. Razzaq et\u00a0al. [73] modified ZSM-5 by various metal species, i.e., Co, Ni, Zn and Fe, and found that the metal modification, particularly Fe-modified ZSM-5, improved the catalytic selectivity towards monoaromatic hydrocarbons (MAHs). Sun et\u00a0al. [74] also confirmed that Fe-modified ZSM-5 catalyst had better deoxygenation activity than the parent ZSM-5 due to the formation of new active sites and inhibiting of repolymerization, leading to a larger amount of aromatics and less coke formation with a higher BTX selectivity. Yung et\u00a0al. [75] reported that the Ga-modified ZSM-5 can result in the increase of hydrocarbons by about 30% over the parent one in the upgrading of bio-oil since the incorporation of Ga increased the dehydrogenation activity. While, Zheng et\u00a0al. [70] also reported that Ga modified ZSM-5 catalysts can lead to a higher bio-oil yield with a lower amount of coke when compared with the parent one. In addition, when the HZSM-5 was modified by 1\u00a0wt.% of Mo had the potential to produce a higher yield of aromatic hydrocarbons than the unmodified one [76]. In our recent study [77], a commercial HZSM-5 zeolite with a Si/Al molar ration of 24 was modified by Cu species with a wet impregnation method, and used for the in-situ upgrading of bio-oil from the fast pyrolysis of biomass with a biomass to catalyst weight ratio of 1:1(Fig.\u00a07\n). It is observed that Cu/HZSM-5 with low Cu modification amounts can maintain the parent HZSM-5 crystalline structure and its acid sites. The 0.5\u00a0wt.% Cu/HZSM-5 showed a highest catalytic performance with a highly relative aromatic hydrocarbons amount of 73.2% and specific aromatic hydrocarbons yield of 56.5\u00a0mg/g-biomass (d.a.f), which are much higher than those by the parent HZSM-5 (55.0% and 26.0\u00a0mg/ g-biomass (d.a.f)). Besides, the 0.5\u00a0wt.% Cu/HZSM-5 catalyst also exhibited excellent catalytic reusability and regeneration property. Herein, the suitable acidity and best textural properties for the deoxygenation of the bio-oil should be attributed to the optimum Cu loading amount. Thus, transition metal modification should be an effective way for the improvement of HZSM-5 catalyst performance. However, it should be noted that the selection of metal species and its doping amount is also important.\nTable\u00a04\n summarizes the reported HZSM-5 and transition metal modified HZSM-5 catalysts for the bio-oil upgrading, in which selectivity towards aromatic hydrocarbons and oil yield are two key parameters for the evaluation of the catalyst performance. It is found that most metal modified HZSM-5 catalysts had low oil yields (< 30%) and/or low selectivity to aromatic hydrocarbons (<60%).In general, HZSM-5 modified by alkaline metal can change acid strength by increasing Lewis acid sites and decreasing in Br\u00f8nsted acid sites, which could effectively prevent excessive cracking of the bio-oil, resulting in a higher yield of the deoxygenated compounds in the upgraded bio-oil with the extension of catalyst lifetime. AAEMs including Na, Mg and Ca have been applied to modify HZSM-5 for the upgrading of bio-oils. For example, when Mg was loaded on HZSM-5 in the form of MgO, the obtained catalysts exhibited better selectivity towards monocyclic aromatics due to the creation of new Lewis acid sites and the reducing of Br\u00f8nsted acid sites [65]. Ca loaded on HZSM-5 also altered acid strength by reducing of strong acid sites and increasing of weak acid sites. Due to the obvious acid strength change and BET surface area decrease, it promoted the production of xylenes but lowered BTX production in comparison to the parent HZSM-5 [67]. Williams and Horne [89] investigated catalytic upgrading of bio-oil derived from biomass pyrolysis over Na modified HZSM-5 (Na-ZSM-5) and compared with the parent HZSM-5, and found that the yield of single ring aromatic compounds in the upgraded bio-oils, especially BTX, increased from 15.9\u00a0wt.% for the HZSM-5 catalyst to 21.3\u00a0wt.% for the Na-ZSM-5 catalyst. Thus, AAEM modification should be also an effective way for the improvement of HZSM-5 catalyst performance.As stated above, the catalytic performance of HZSM-5 can be improved by metal modification in the upgrading of bio-oils. There are various methods for the preparation of metal modified HZSM-5 catalysts, and the preparation method could influence the physicochemical properties and catalytic upgrading performance of the obtained catalysts even using the same metal for the modification. The typical preparation methods include impregnation, ion exchange, precipitation, sol-gel, hydrothermal and the temperature programmed reaction methods. Especially, the impregnation and precipitation methods have been extensively utilized. The impregnation method can be classified as wet impregnation, in which an excessive amount of metal solution is used, and the dry impregnation, in which the volume of metal solution equals to the entire pore volume of HZSM-5 catalysts. It is reported that impregnation time and the following drying way could affect the metal modification efficiency on the HZSM-5 catalysts [77,90]. The short impregnation and drying time may cause weakly adsorbing metal species, leading to the metal deposition mainly on the zeolite surface rather than within pores. In the ion-exchange method, the excess metal precursor after ion-exchange process will be washed out by deionized water so that only the exchanged metal ions remain in the zeolite, which can result in well dispersion of metal species on the zeolite and decrease the aggregation of active species during the bio-oil upgrading process [91]. Furthermore, co-impregnation, a process for manufacturing bimetal modified HZSM-5 zeolites, is more challenging than the single metal impregnation since different metal species with different solubility and diffusivity could lead to different degrees of precipitation within the pores. However, the suitable bimetal modification could obtain more excellent zeolite catalysts for the bio-oil upgrading than the single metal modification [92]. While, due to the high dispersion ability of the precipitation method, it is more suitable for the preparation of metal-modified HZSM-5 catalysts with a high metal loading amount [77,90]. Herein, it should be noted that the nucleation and growth of metal particles will be induced by the supersaturation of precursor solution. Therefore, it is important to select suitable metal modification method for the metal-modified HZSM-5 catalyst preparation in the bio-oil upgrading.Hierarchical zeolites are characterized by the presence of a bi- or multimodal porous structure, especially containing micro-, meso- and macropores together in the HZSM-5 based catalysts. The exact definition of hierarchical structure of catalysts is a pore system of bi- or multimodal pores with different pore size where the large pores connect the small pores, i.e., small pores branch off from a continuous large pore [93]. Because of their unique properties, such materials have been attracted great attentions. Hierarchical HZSM-5 differ significantly from the conventional one in terms of improved diffusion, promoted mass transfer, enhanced resistance to deactivation [94]. The classical methods for introducing multimodal porous structure into the HZSM-5 catalysts include single templating, double templating with soft/hard-template and post-treatment.In general, the single templating method can only lead to the generation of micropores during the zeolite synthesis. Typically, tetrapropylammonium hydroxide (TPAOH) is applied as a structure-directing agent as well as a micropores template for the preparation of ZSM-5. In the case of the addition of TPAOH alone during ZSM-5 synthesis, to obtain hierarchical structured ZSM-5, it requires base etching as the post-treatment process to generate mesopores as stated in the following Section\u00a04.2.3\n[95].In the synthesis of zeolites with mesopores, two types of templates are generally selected, in which one should have a three-dimensionally structured mesopore system and it is mixed with the precursor chemicals including structure-directing agent of ZSM-5. In this case, mesopores can be produced during the crystallization stage, and the template will be decomposed by a calcination stage after the hydrothermal synthesis process [30]. It is reported that using a small amount of mesopore template during the synthesis process always results in insufficient mesopore generation whereas using a large amount of template will hinder the crystal nucleation and growth [96]. Thus, it is needed to search the optimum mesopore template amount during the synthesis. In the production of hierarchical zeolites, both hard template such as carbon black, carbon fibers, aerogels, and polymer aerogel and soft template such as cationic polymers, amphiphilic organosilane surfactants, and silylated polymers will be applied [97,98], by which the morphology of mesopores may be accurately regulated. Compared to the chemical etching (acid or base treatment) for the generation of mesoporous or microporous structure as as stated in the following Section\u00a04.2.3, the employing of a secondary template could raise the cost and a calcination process to eliminate the template is also necessary.In comparison to direct synthesis ways, the post-treatment method is more convenient, simple, and cost-effective for introducing secondary meso- and/or macro-pores in the zeolite catalyst structure by extraction of elements on the framework. Generally, an acid solution was employed for extracting Al atom while a base solution was employed for extracting Si atom. Both of which can generate mesopores in the zeolite structure. The disadvantage of this method is the loss of micropore structure after the generation of mesopore porosity, or the collapse of structure to lose the surface area in the case of over extraction. The first method is the use of base or alkaline solution, which can be called \u201cdesilication/alkaline etching\u201d. The preparation of hierarchical HZSM-5 catalysts by the alkaline etching method could result in improved surface properties. Since the small pores of the parent ZSM-5 always limit the mass transfer and diffusion of reactants and products, by introducing mesopores into the structure, the accessibility of acid sites can be improved, the diffusion length can be shortened and finally the catalyst lifetime can be enhanced [99]. Popular alkaline solutions used for the etching of HZSM-5 include sodium hydroxide (NaOH), tetrapropylammonium hydroxide (TPAOH), tetrabutylammonium hydroxide (TBAOH), sodium carbonate (Na2CO3) with NaOH solutions [100]. During the alkaline etching, some Si-Al bonds may be disrupted, causing the zeolite to lose its location adjacent to Al and decrease the acidity. While, the obtained catalysts usually have a wide pore size distribution. For the bio-oil upgrading, it is reported that the ideal Si/Al ratio in the original ZSM-5 for the construction of a hierarchical structure by using alkaline solution should be 25\u201350 [101]. Herein, since the excessive aluminum content in ZSM-5 can surround the silicon to prevent the desilication for the mesopore creation. On the other hand, when the aluminum content is too low, a large amount of Si can be also dissolved due to the instability of structure, resulting in low yield of zeolite and too many large mesopores, which always negatively affect the catalytic cracking activity. Organic hydroxides such as TPAOH and TBAOH have lower ability for the Si dissolution than those inorganic hydroxides, which can better control Si species dissolution so that only a little change occur in the acid property after the alkaline treatment [102]. Furthermore, it is reported that the hierarchical HZSM-5 catalysts by post-treatment using a mixture of inorganic and organic base solutions can result in a better control of the mesopore formation since the assistance of the organic base solution prevented the excessive extraction to maintain a better zeolite structure as shown in Fig.\u00a08\n\n[103].In our previous study [104], hierarchical HZSM-5 zeolites were fabricated via desilication of conventional HZSM-5 in various NaOH solutions in the presence of TPAOH for the bio-oil upgrading. It is found that the hierarchical HZSM-5 prepared by using 0.2 M NaOH with 0.25 M TPAOH etching had the most excellent catalytic performance with detected aromatics yield as high as 45.2\u00a0mg/g-bio-oil. Herein, in the presence of 0.25 M TPAOH, the mesopore formation can be well controlled, leading to the rise in surface area but maintaining enough acidity. Furthermore, to increase the aromatic hydrocarbons as well as the coking resistance, the hierarchical HZSM-5 was further modified by various metals. It is observed that 0.25\u00a0wt.% Cu/HZSM-5 resulted in the increase of the aromatic hydrocarbons yield (\u223c54.5\u00a0mg/g-bio-oil) with a better coking resistance performance (Fig.\u00a09\n).The second method is the use of acid solution, which is called \u201cdealumination/acid leaching\u201d. Dealumination is the process of removing Al from the zeolite framework using an acid solution. It is most typically employed for the post-treatment of parent HZSM-5 zeolites with a low Si/Al ratio (2.5-5), which usually have high acid density and strong acid strength and favor the coke generation, thereby causing catalyst deactivation [105,106]. After the dealumination using acid solution, the Si/Al ratio in the zeolite framework will be increased, and a hierarchical structure with existing mesoporosity will be generated. In this case, the adjustment of zeolite acidity becomes more controllable. In addition, even if the zeolite is extremely dealuminated, partial extra-framework Al can be reinserted into the structure using the hydrothermal technique while the zeolite structure is preserved better than the case employing the alkaline etching method [107,108]. However, the use of strong acid concentration will cause a decrease of Br\u00f8nsted acid sites due to a significant decrease in the framework Al content, and an increase in the Lewis acid site due to the detached Al which becomes an extra-framework Al floating in the environment as an extra-framework Al atom [109]. The most commonly used acids in the dealumination are hydrochloric acid (HCl), nitric acid (HNO3), sulfuric acid (H2SO4) and hydrofluoric acid (HF) [110]. Herein, chemical etching with fluoride ions generated from NH4-HF has been reported to create secondary porosity and increase the exterior surface area of hierarchical ZSM-5. However, the use of extremely concentrated HF solution could result in the decrease of Br\u00f8nsted acidity due to the indiscriminate extraction of Al and Si [111]. Fig.\u00a010\n presents an example of an SEM image of hierarchical ZSM-5 treated with an acid solution.As stated above, owing to its acidity and shape-selective property of HZSM-5, it is considered as one of the most effective catalysts for the deoxygenation of bio-oil, especially in the production of light aromatic hydrocarbons such as BTXs. However, one of the major drawbacks of HZSM-5 is the fast deactivation due to the coke formation. In the upgrading of bio-oils, the mass transfer of large intermediate oxygenates in the bio-oil is restricted by the small pore diameter of HZSM-5 so that it is difficult to diffuse into the pore for reaction, which will result in the coking to lock the inner pore and finally cause HZSM-5 catalysts to be deactivated. Thus, HZSM-5 catalysts with special morphologies have been designed to reduce the diffusion length or increase the diffusion efficiency. To date, hollow ZSM-5 catalysts, core-shell ZSM-5/mesopore composite catalysts, nanosheet ZSM-5 catalysts have been developed.To improve the catalytic effectiveness and keep the shape-selective ability of micropores in the zeolites, interior hollow structured zeolites have been designed and fabricated. In this case, the hollows within the zeolite catalysts will assist to shorten the diffusion distance with a decrease in the unutilized quantity within the center of zeolites. Two ways have been applied to fabricate hollow zeolites [112]. The first one is to assemble zeolite crystals with polycrystalline shells using a template followed by calcination at a high temperature to remove the template as shown in Fig.\u00a011\n\n[112]. Herein, the zeolite crystals are aggregated at random to form an interparticle cavity. Since the obtained shell thickness is only a few microns, the reactants and products could be more easily pass through the shell. Fig.\u00a012\n shows one obtained hollow zeolite catalysts by using mesoporous silica spheres as the sacrifice template, on which zeolite nanoparticles subsequently grow during the synthesis process [113]. Herein, the inner part of the particle is a jujube-like mesoporous silica sphere as Si and pre-seeded source for the preparation of hollow zeolite.The another approach is to employ chemical etching to create hollow cavities in the zeolite catalysts. This method involves two-step: (i) using sol-gel method to grow large single zeolite crystal from a single nucleus; (ii) constructing the hollow structure in the center of the obtained zeolite crystal with a shell by etching using various chemical concentrations. Fig.\u00a013\n shows ZSM-5 zeolites with an internal hollow structure, where the parent ZSM-5 crystal was desilicated by TPAOH [114].In our previous study [115], hollow HZSM-5 catalysts with a mesoporous shell were fabricated by a hydrothermal process combined followed with TPAOH etching. When it used for in-situ catalyst upgrading of bio-oil from rapid pyrolysis of biomass, the obtained best hollow HZSM-5 catalyst led to aromatic hydrocarbon yields in the range of 78.49\u201378.67% for cellulose and hemicellulose, which is much higher than those using the parent HZSM-5 (61.06\u201368.26%). While, when the biomass(cedar)/catalyst weight ratio was 1:2, up to 80.16% of the aromatic hydrocarbon yield was achieved. Besides, this catalyst had good reusability and regeneration property since the accessibility to the acid sites in the prepared hollow HZSM-5 catalysts was greatly enhanced, which effectively increased the reaction rate as well as the coking resistance (Fig.\u00a014\n).Preparation of HZSM-5/mesoporous material composite with a core/shell structure is another way to improve the performance of HZSM-5 based catalysts, which is considered to have following advantages: (i) the mesoporous structure on the shell is beneficial for the better diffusion and can hinder the coke formation on the surface of ZSM-5 catalysts; (ii) the acidity and shape selectivity of ZSM-5 suitable for aromatic hydrocarbons production in catalytic upgrading process can be maintained [116\u2013118]. Fig.\u00a015\n shows an illustration for the preparation of the core/shell ZSM-5/mesoporous material composite [112], which is a ZSM-5 catalyst coated with a thin mesoporous MCM-41 layer. When it was applied for the in-situ and ex-situ catalytic fast pyrolysis of biomass, the equivalent hydrocarbon yield as that from the pure ZSM-5 was obtained, but the MCM-41 shell worked as a barrier layer for coke deposition, which effectively protected the ZSM-5 from the severe coke formation.The nanosheet HZSM-5 has a hierarchical pore structure with a thin sheet structure, which can overcome the disadvantages of the small and long porous diffusion path in the conventional HZSM-5 zeolite [119\u2013121]. Due to its size (only 5\u201310\u00a0nm) as small as the size of a unit cell, the diffusion length can allow the reactants and products to diffuse in the HZSM-5 more easily, which could improve catalytic activity and lifetime of catalysts [119\u2013121]. A mesoporous HZSM-5 nanosheet was synthesized by using a surfactant as the structure-directing agent and applied to upgrade the bio-oil in the catalytic pyrolysis of cellulose. In this case, although the obtained HZSM-5 nanosheet resulted in similar aromatic hydrocarbons and olefins yields as those by using conventional HZSM-5 catalysts, it demonstrated a longer lifetime even though the coke content was also higher than those cases using the conventional HZSM-5 catalysts since the mesopores still enabled better accessibility to active acid sites. Furthermore, because such a nanosheet zeolite has a larger surface area, it could be an outstanding player for loading metal nanoparticles to create a multifunctional zeolite catalyst. With increasing metal loading amounts, the metal nanoparticles could distribute uniformly within zeolite crystals and prevent aggregation during the reaction.As is known, biomass is the only renewable energy source and is considered a potential alternative to fossil fuels. Efficient biomass utilization is full of challenges, and among the available technologies, fast pyrolysis is effective in producing bio-oil. However, pyrolysis bio-oil needs to be upgraded before use as the transport fuels or chemical feedstock due to its low quality as fuels and its low value as the chemicals. Catalytic upgrading of pyrolysis bio-oils to increase the quality is desired in recent year and the catalyst with high catalytic activity, excellent selectivity and long life-time is the foundation of this process. The HZSM-5 based catalysts are currently the most widely used and effective catalysts for biomass conversion in aromatics production due to its outstanding acidity, heat-resistant properties, adequate pore size. However, the mass transfer capacity of large reactants and products through the small pores of HZSM-5 is limited which had a significant negative impact on the catalytic performance. Due to the most cracking reaction occurring within the pore, the generated products will diffuse out slowly from the pores, resulting in a low-yield of product. In addition, the HZSM-5 based catalysts often suffer from coke formation due to its high acidity. Thus, HZSM-5 based catalysts should be more improved and modified.In this article, HZSM-5 based catalysts for the upgrading of pyrolysis bio-oils are critically reviewed. The effects of porous structure, acidity and their interactions on biou-oil upgrading activity, selectivity, stability and deactivation are summarized. The proposed mechanisms on the bio-oil upgrading over the catalysts are discussed. In particular, the main strategies including metal modification, construction of zeolites with a hierarchical structure and synthesis of special morphologies with hollow structure or core/shell and nanosheet structures for the improvement of deoxygenation property performance are introduced. In fact, each of the strategy investigated has its combination of advantages and disadvantages. It should take the advantages of each technique and continue to improve the HZSM-5 based catalyst to make it more efficient. Based on the critical review above, the diffusion path and acidity property should be moderately modified to preserve catalytic activity and meanwhile to avoid secondary reactions that lead to coke and catalyst deactivation. However, those reported studies for the development of HZSM-5 based catalysts to optimize catalytic upgrading of bio-oil remain at a laboratory scale. Therefore, it would be of great benefit to the industrial application if it is cost-effectively scaled up. Finally, the improved HZSM-5 based catalysts were not only suitable for the catalytic upgrading of bio-oil, as the original ZSM-5, they should be also served as the excellent catalysts for a variety of other processes. It is expected that this review will assist not only those who are interested in catalytic upgrading of bio-oil processes, but also those who are interested in applying my findings to other research areas (Fig.\u00a09).The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by JST Grant Number JPMJPF2104 and Hirosaki University Fund, Japan. N. Chaihad gratefully acknowledges the scholarship from the Ministry of Education, Culture, Sport, Science and Technology (MEXT) of Japan.", "descript": "\n                  Fast pyrolysis of biomass is an attractive way to produce bio-oil since it can convert most of biomass components directly into liquid fuel. However, the bio-oils obtained from such a fast pyrolysis process always have highly complex oxygenated compounds with high viscosity, serious corrosivity, and rather instability. Thus, before the raw bio-oils are used as fuel or chemical feedstock, they must be upgraded, especially deoxygenated. Cracking of bio-oils over porous solid catalysts such as zeolite-based catalysts at ambient pressure is considered one of effective ways for the bio-oil upgrading, especially in which hydrogen gas is not necessary. Herein, zeolite-based catalysts (mainly HZSM-5 based catalysts) for the upgrading of pyrolysis bio-oils are critically reviewed. The effects of porous structure, acidity and other parameters including biomass type, biomass/catalyst ratio and operation temperature on cracking activity, selectivity, stability and deactivation are summarized. While, the proposed mechanisms on the bio-oil upgrading over the zeolite-based catalysts and the possibility for the application of the developed catalysts in the industrial process are discussed. Furthermore, the main strategies including metal modification, construction of zeolites with a hierarchical structure and synthesis of special morphologies with hollow structure or core/shell structure and nanosheet structures for the improvement of deoxygenation property performance are introduced. It is expected to provide a guidance for the design and fabricate more excellent zeolite-based catalysts and their application for high-quality bio-oil production from fast biomass pyrolysis.\n               "}
{"full_text": "2-MethylindoleTetrahydro-N-ethylcarbazoleOctahydro-N-ethylcarbazoleOctahydro-2-methylindoleDodecahydro-N-ethylcarbazoleDodecahydro-N-propylcarbazolePerhydro-dibenzytolueneBenzene, toluene, and xyleneCatalytic fast pyrolysisCompressed hydrogen storageCarbon material hydrogen storageUnited States Department of EnergyLiquid ammonia hydrogen storageLiquid hydrogen storageLiquid organic hydrogen carriersMetal alloy hydrogen storageMethylcyclohexaneMethanol-to-aromaticsN-ethylcarbazoleN-propylcarbazoleThermal catalytic conversion and ammonificationTolueneUnited States of AmericaAs living standards around the world improve, there is an increasing reliance on energy, which exacerbates the global energy challenge. Renewable energy has developed into a clean and efficient alternative to existing energy sources. Furthermore, the worldwide focus on \u201cEmission Peak and Carbon Neutrality\u201d and the adoption of numerous legislations have made renewable energy the unavoidable path for the transformation of the current energy system [1\u20133]. Therefore, it is an inevitable trend to accelerate the development of renewable energy. At present, the use and storage of hydrogen is a potential route to the current development of renewable energy, as hydrogen is a clean energy source that does not produce any pollutants [4]. \u201cHydrogen economy\u201d is also a hot topic of sustainable development, and it is necessary to determine the direction and strategy of future development based on the conditions of the world. Technical methods for hydrogen production and consumption are reasonably established and well-developed in the supply chain. Hydrogen energy storage, hydrogen-powered automobiles, and hydrogen-powered ships are common applications [5,6]. Due to low volumetric density (0.0899 kg/m3), volumetric energy density (0.003 kW\u00b7h/L), and gravimetric energy density (33 kW\u00b7h/kg), hydrogen energy usage faces significant storage and transportation constraints. The flammable and explosive qualities of hydrogen at normal temperature and pressure have also hampered large-scale and commercial hydrogen energy uses [7\u20139]. For on-board hydrogen sources, the US Department of Energy (DOE) has proposed objectives of 5.5 wt% and 62 kg/m3 for gravimetric and volumetric hydrogen capacity, respectively [10].Hence, the development of efficient hydrogen storage technology is currently a popular focus of research. Improving hydrogen storage capacity and rate while reducing energy consumption are major characteristics of hydrogen storage technology. Compressed hydrogen storage (CH2), liquid hydrogen storage (LH2), liquid organic hydrogen carriers (LOHCs), liquid ammonia hydrogen storage (LAH2), metal alloy hydrogen storage (MAH2), and carbon material hydrogen storage (CMH2) have all been the subject of much investigation to address above problems [11,12]. Among them, LOHCs technology is recognized to be excellent for long-distance and large-scale hydrogen storage and transportation due to its high hydrogen storage capacity, environmental friendliness, safety, and efficiency [13,14]. According to previous research, hydrogen storage liquids are generally high-purity single aromatic or N-doped compounds. Complex refining methods are often used to create these compounds from nonrenewable fossil energy sources. For instance, aromatic compounds can be made through naphtha reforming and petroleum cracking. Potentially, LOHCs can be obtained from biomass been turned into valuable liquid fuels and aromatic compounds through a variety of conversion methods. The lignocellulosic biomass is composed of cellulose (40%-50%), hemicellulose (25%-30%), and lignin (15%-20%) [15,16]. Take lignin as an example, it is a unique, renewable natural polymer with aromatic structures. Through the thermochemical conversion pathway, lignin can be depolymerized into aromatic compound intermediates, and can also be used to produce small molecular compounds such as biomass fuels and light aromatics [17]. Theoretically, these compounds can be used for the storage and release of hydrogen via a pair of reversible reactions in LOHCs technology. Thermal conversion conditions of biomass are expected to optimize the target product as a major component of organic hydrogen storage liquids, including targeted deconstruction, nitrogen doping, better catalysts, reaction condition, etc.In all, this article provides an overview of LOHCs technology, including basic principles, technical approaches, and applications. This review also proposes the concept of biomass-derived renewable LOHCs to demonstrate the potential of biomass as a carbon-neutral energy carrier for hydrogen storage. Combined with the current thermochemical conversion technologies of biomass, the preparation and development of aromatic hydrocarbons and N-doped compounds are briefly summarized and exhibited. In addition, the technological route, feasibility, and challenges of biomass-based LOHCs were evaluated.Based on the reaction principle, the categories of hydrogen storage are mainly composed of physical hydrogen storage, physical adsorption hydrogen storage, and chemical adsorption hydrogen storage. CH2\n[18] and LH2\n[19] are two types of physical methods. The physical adsorption of hydrogen [20] is usually composed of carbon materials, zeolite, and metallic organic framework materials. LOHCs, LAH2\n[21], electrochemical hydrogen storage [22], and MAH2\n[23] are examples of chemical adsorption storage methods. It is worth mentioning that hydrate hydrogen storage is also a physicochemical method, where H2 capture occurs via the formation of a hydrate shell with hydrogen bonds between water molecules, with H2 being kept by the topology of the cavity [24]. The characteristics, advantages, and disadvantages of several common types of existing main hydrogen storage technologies are summarized in Table 1\n. Pure steel metal (17.5\u223c20 MPa), steel liner fiber wound (26.3\u223c30 MPa), aluminum liner fiber wound (30\u223c70 MPa), and plastic liner fiber wound (>70 MPa) are the four pressure categories for gas cylinders [25]. The application of some countries is relatively mature. For instance, the all-steel bottle container produced by JFE in Japan and the carbon fiber-wound hydrogen storage container with steel liner developed in the USA have been used for hydrogen refueling stations. Currently, bottle leakage, liner and interface sealing, and the design of transportable hydrogen cylinders for transportation are all issues that researchers are grappling with.CH2 technology is relatively easy to industrialize and offers fast charge and discharge rates, it is widely used for onboard hydrogen storage in new energy vehicles. Compared with CH2 technology, the storage capacity of LH2 technology has been greatly improved. LH2 technology has a density of 70.85 kg/m3, which is 1/800 of the volume of gaseous hydrogen, making it easier to transport large volumes over long distances, considerably improving transport efficiency [26]. This type of technology is universally used for cryogenic rocket propellants [27,28]. Although LH2 technology has sufficient advantages in terms of storage and transport capacity, it is based on a liquid-phase state formed at extremely low temperatures (< -253\u00b0C). As a result of its peculiar working conditions, LH2 technology has some drawbacks. On one hand, energy consumption is extraordinarily high as a result of the need to transform gaseous hydrogen into liquid through a series of technical means. Liquid hydrogen, on the other hand, absorbs heat continuously to form evaporative gases, which necessitate high insulation in storage facilities [29]. LAH2 technology tends to utilize ammonia as the hydrogen carrier, and a high hydrogen storage capacity is obtained (17.8 wt%), 1.7 times higher than that of LH2. Due to the high stability, liquid ammonia can meet the need for energy storage in time and in space [30]. However, LAH2 technology should be considered for the need of high energy input and toxicity, and potential hazards to equipment, the human body, and the and environment during long-term and long-distance storage and transportation [31].In contrast with traditional hydrogen storage technologies, LOHCs technology has following major advantages: (1) it has a prominent capacity for hydrogen storage, as well as excellent performance (meeting DOE index requirements); (2) most of the substances have a high boiling point and low melting point, allowing them to maintain a stable liquid phase at room temperature while remaining nonvolatile; (3) this system has stable catalytic hydrogenation and dehydrogenation processes, and the reactants and catalysts can be recycled; (4) the storage, transportation, and maintenance of hydrogen-storage materials are safer and more convenient, allowing for large-scale and long-distance distribution; (5) current gasoline and diesel delivery techniques and gas station buildings may be immediately implemented. However, there are also several disadvantages, mainly including: (1) the reaction requires professional hydrogenation and dehydrogenation equipment, which has high investment costs; (2) the dehydrogenation reaction must be carried out at high temperature, which is likely to result in catalyst coking and deactivation; (3) the reaction process consumes large amounts of energy, and the performance decreases after several cycles; (4) the hydrogen produced by the dehydrogenation reaction is not of high purity, and inappropriate conditions and reactants are more likely to induce side reactions; (5) the initial cost to purchase LOHCs materials is extremely high.What's more, some studies have found a higher energy demand of LOHCs than CH2, LH2, MAH2, and LAH2 based on the 0-dimensional simulation [32]. However, in terms of energy storage of regenerative hydrogen in the cell system, LOHCs technology showed efficiency with the increase of the energy storage cycle, confirming the suitability for long-term hydrogen storage. Assuming that large-scale storage and transport across oceans are targeted for applications in the field of hydrogen storage, then there will be a wider market for LOHCs technology. In addition to the characteristics of high hydrogen storage capacity, carbon cycle, and suitability for long-term utilization, the safety performance is also superior to that of other hydrogen storage methods. Since the concept of LOHCs was proposed, the technology has also been continuously optimized. In other words, LOHCs technology is exactly promising and marketable for hydrogen storage today.LOHCs technology is based on reversible hydrogen storage and release reactions using unsaturated liquid organics (e.g., toluene, naphthalene, and N-ethylcarbazole) as hydrogen storage agents and the corresponding saturates (e.g., methylcyclohexane, decalin, and dodecahydro-N-ethylcarbazole) as hydrogen carriers [7,37]. The fundamental principle of the reactions in LOHCs technology is shown in Fig. 1\n. The hydrogenation process is an exothermic reaction in which the organic hydrogen storage liquid is mixed with raw hydrogen in the reactor. The system is then heated to a specific temperature under the influence of the catalyst to form the corresponding saturated hydride. The products of the hydrogenation reaction are called hydrogen carriers (Hx-LOHCs). Essentially, it is a catalytic mechanism that uses hydrogen to transform unsaturated bonds into saturated ones [14,26]. From the perspective of chemical equilibrium, both low temperatures and high pressures are more favorable to hydrogenation. As the inverse of hydrogenation, dehydrogenation is an endothermic reaction. In the presence of the catalyst, hydrogen is extracted from Hx-LOHCs in the dehydrogenation device. The process involves continual absorption of external heat due to the energy difference between the energy required for the dissociation of hydrogen atoms and the activation energy of the C-H bond [38]. It is necessary to focus on the large energy consumption caused by the difference in temperature during the reaction, as well as the reduction of catalytic activity. Then, for long-distance transportation of hydrogen carriers, existing liquid fuel transportation (pipelines, ships, trucks, etc.) can be employed. It was found that unsaturated aromatics and corresponding hydrides can be hydrogenated and dehydrogenated without destroying the main structure of the carbon ring [39].To increase the conversion rate and selectivity, as well as the recycling efficiency and activity of the catalysts, the composition of the compounds and conditions can be tweaked. Consequently, a high-performance LOHCs system should have the following performance indices [40]: (1) low melting point and high boiling point; (2) large hydrogen storage capacity (volume: >56 kg/m3, gravity: >6 wt%); (3) high stability of ring chain during dehydrogenation and high purity of hydrogen discharged; (4) low heat uptake and mild dehydrogenation conditions; (5) cheap cost and readily available materials; (6) long cycle life and selective dehydrogenation; (7) enhanced stability during use and transportation, low toxicity, and environmental friendliness.LOHCs technology was first proposed as a non-cryogenic approach in 1975. Such technologies are more inclined to use aromatic compounds such as benzene and toluene for vehicle fuels as hydrogen storage carriers. As research progressed, scholars discovered that the enthalpy and temperature can be effectively reduced when an appropriate number of heteroatoms, such as N atoms, intervene and replace C atoms. Furthermore, as the number of N atoms in an organic compound grows, the dehydrogenation temperature also increases [42\u201344]. The most investigated organic compounds at the moment are aromatic and N-doped compounds. Simultaneously, a growing number of organic compounds are being discovered. Table 2\n lists the physicochemical parameters and reaction equations for several regularly used LOHCs systems. The main systems found in the literature are toluene (TOL)/methylcyclohexane (MCH) [9], N-ethylcarbazole/dodecahydro-N-ethylcarbazole [45,46], naphthalene/decalin [47], dibenzyltoluene/perhydro-dibenzyltoluene [48], biphenyl/bicyclohexyl, and diphenylmethane/dicyclohexylmethane. Many scholars have studied the reaction mechanism of these hydrogen storage systems through molecular dynamics, nuclear magnetic resonance and other methods, which are summarized in Fig. 2\n. It is discovered that the steric effect will be strongly impacted by the existence of molecular size, methyl, heteroatoms, etc., thereby affecting the priority of bond hydrogenation and dehydrogenation.Furthermore, a few corporations in wealthy countries such as Germany and Japan have already commercialized many systems. For example, dibenzyltoluene, which has a maximum hydrogen storage capacity of 57 kg/m3, has been the subject of research by the German business \u201cHydrogenious LOHC Technologies GmbH\u201d. Japanese enterprises, such as \u201cChiyoda Chemical Construction\u201d, have already integrated and developed in a variety of disciplines, including ocean-going hydrogen transport, miniaturization of hydrogenation and dehydrogenation, hydrogen refueling stations, and distributed energy delivery. Overall, Japanese industries have concentrated their research and development efforts on toluene/methylcyclohexane. Some Chinese firms have entered the LOHCs market. The most established company is \u201cHynertech\u201d in Wuhan, where the N-ethylcarbazole/dodecahydro-N-ethylcarbazole system is the primary research focus. They have been concentrating on the creation of \u201chydrogen oil\u201d, a liquid organic hydrogen storage solution that combines the benefits of safety and stability with high hydrogen storage capacity. It's worth noting that \u201cHynertech\u201d has demonstrated high-temperature waste gasification to \u201chydrogen oil\u201d and the hydrogen energy business.Although olefins, alkynes, and aromatic hydrocarbons can all be employed as hydrogen storage liquids, studies have confirmed that aromatic compounds are the best choice for hydrogen storage. Aromatic compounds can be applied as LOHCs because of the unique resonance interaction of aromatic rings which makes them more likely to hydrogenate and dehydrogenate than other organic molecules [65]. The first system studied was benzene/cyclohexane [65,66], which has a high gravimetric hydrogen storage capacity of 7.2 wt% and a volumetric hydrogen storage capacity of 55.9 kg/m3. The major disadvantage, however, is the high dehydrogenation temperature. The dehydrogenation reaction needs to be completed at a temperature of close to 300\u2103, resulting in considerable energy consumption. Fig. 3\na illustrates the hydrogenation and dehydrogenation reactions of the benzene/cyclohexane system [65]. In a catalyzed hydrogenation process, benzene and hydrogen are introduced into the reactor at a pressure of 4 MPa and a temperature of 150\u00b0C to create saturated cyclohexane. During the reaction, approximately 68.8 kJ/mol of energy is produced. At 300\u00b0C and 0.1 MPa, cyclohexane absorbs roughly 68.8 kJ/mol of energy, with the gradual release of hydrogen and eventually dehydrogenates to benzene [67]. To find a better reaction system, Itoh et al. [68] used Pt/Al2O3 as the catalyst to compare the conversion and dehydrogenation temperatures of cyclohexane and methylcyclohexane, two types of hydrides: benzene and toluene. Because of the presence of methyl as an alkyl group, the dehydrogenation temperature was lower than that of cyclohexane. The mixture of methylcyclohexane and cyclohexane has also been shown to be a hybrid chemical hydride, which can be researched for hydrogen storage [68].The reversible process of the toluene/methylcyclohexane system is shown in Fig. 3b, where a reduced storage capacity is accompanied by a lower dehydrogenation temperature. Overall, the benzene/cyclohexane and toluene/methylcyclohexane systems exhibited excellent hydrogen storage capacities. However, a higher dehydrogenation temperature remains a critical factor affecting the energy consumption of the reaction. Naphthalene, a cheap and simple condensed aromatic hydrocarbon with excellent hydrogen storage capacity, was shown to be equally suitable for hydrogen storage [69\u201371]. The naphthalene/decalin system reaction is shown in Fig. 3c [58]. As naphthalene is prone to many side reactions during hydrogenation, there are distinctions between the resulting naphthenic compounds such as trans-decalin and cis-decalin [72]. Many factors influence the rate of reaction, as well as the separation and purification of the product, such as the number of rings opened, organic solvents, and catalysts. Terribly, the high temperature and vapor pressure during the dehydrogenation of decalin will lead to ring-opening and cracking after several cycles. This will make more tar and coke adhere to the equipment which is difficult to remove [72]. In addition, to ensure the reactivity and state of the liquid phase, it is necessary to equip the transport with heating facilities, which increases the complexity and cost of storage and transportation.Dibenzyltoluene (DBT) is the principal component of \u201cthermal conductive oil\u201d in daily production, in addition to the three systems mentioned above. The physical and chemical qualities of \u201cthermal conductive oil\u201d are stable, and it is distinguished by its high boiling point, low melting point, and low toxicity. The dibenzyltoluene/perhydro-dibenzytoluene system has become a hot topic within current aromatic hydrogen storage liquid compounds [59,73]. As illustrated in Fig. 3d, DBT interacted with hydrogen in the presence of the catalyst to produce perhydro-dibenzyltoluene under certain conditions. What's more, DBT is a prominent research target for aromatic organic hydrogen storage liquids due to its exceptional physicochemical features. DBT may be recycled numerous times when used as the hydrogen storage carrier, and its fixed cost for hydrogen storage is lower than that of liquid chemical plants. However, the use of DBT has drawbacks of high energy consumption and slow reaction rate in dehydrogenation, and H2 needs to be purified when released [42]. In addition, the technical bottlenecks in developing efficient and low-cost dehydrogenation catalysts also limit the application of DBT in liquid hydrogen storage to some extent.In addition to the systems mentioned above, there are several studied on LOHCs utilizing biphenyl/bicyclohexyl and diphenylmethane/dicyclohexylmethane as the research subjects. Biphenly and diphenylmethane are particularly popular as hydrogen carriers because of the excellent hydrogen capacity, stability and economic performance [74]. However, the solid physical state under ambient environment has limited the application for hydrogen storage. Hence, a growing number of studies have proposed the low-eutectic mixture (such as the combination of biphenyl and diphenylmethane) as potential hydrogen carrier, realizing liquid state at room temperature and atmospheric pressure [75,76]. It has been demonstrated that an optimum composition of biphenyl (C12H10, 35 wt%) and diphenylmethane (C13H12, 65 wt%) can be formed and the hydrogen storage capacity can be maximized (6.9 wt% and 69.1 gL-1) [77].It has been confirmed that the incorporation of heteroatoms such as N, P, and O into aromatics can significantly affect dehydrogenation thermodynamics [43]. To better understand the relationship between the heterocyclic structure and the enthalpy of dehydrogenation, Clot et al.\n[63] investigated how the substituted positions of different heteroatoms affect the dehydrogenation reaction. They eventually discovered that substituting heteroatoms at the ring's 1-position can effectively lower the dehydrogenation temperature of the hydrogen storage carriers. In addition, the addition of heteroatoms to the five- and six-membered rings can reduce the enthalpy of dehydrogenation to a significant extent. Recently, an increasing number of studies have discovered that hydrides in N-doped systems of N heterocyclic compounds (e.g., indoles [78,79], pyridines [80,81], and carbazoles) exhibit lower dehydrogenation temperatures than cyclic olefins. N-ethylcarbazole (NECZ)/dodecahydro-N-ethylcarbazole (12H-NECZ) [63] is the most frequent in this sort of system. The dehydrogenation temperature gradually lowered to around 200\u2103 as the number of heteroatoms in the organic matter increased. The hydrogen storage and release reactions of the NECZ/12H-NECZ system are shown in Fig. 4\na. Under catalytic action, hydrogenation of NECZ can be achieved within 150\u00b0C - 200\u00b0C, resulting in the formation of 12H-NECZ. At a pressure of 0.1 MPa and a temperature around 200\u00b0C, 12H-NECZ could be reduced to NECZ. When entirely hydrogenated, the gravimetric hydrogen storage capacity of 6.7 wt% for carbazole and 5.8 wt% for NECZ are lower than that of the aromatic-alkane system. However, there was a significant decline in dehydrogenation temperature. The difference in temperature and enthalpy of the reaction between carbazole and NECZ is rather modest. The addition of ethyl in NECZ slightly decreases the hydrogen storage capacity. However, the passivation effect of N atom on the catalyst is effectively reduced to maintain the activity [82]. Moreover, the melting point of NECZ is just 69\u00b0C, significantly decreasing the number of additional operating steps owing to the higher melting point. Because carbazole and NECZ are solids at room temperature, some organics, such as hot ethanol and ether, need to be added to dissolve the solid or lower the melting point to enable a smooth reaction. For the dehydrogenation of 12H-NECZ, an inadequate reaction produces additional by-products. Taking 4H-NECZ and 8H-NECZ as examples, these substances decrease the efficiency of transformation. In many cases, the presence of stereoisomers of semi-hydrogenated products with different reactivity results in reduced and incomplete conversion, even though the thermodynamic is favorable [83]. Accordingly, suitable catalysts for hydrogenation and dehydrogenation must be selected to address the problem of low efficiency.In addition to carbazole, several N-doped aromatic heterocyclic derivatives (e.g., indole [79,84], pyridine [44,81], and pyrrole) can be hydrogenated and dehydrogenated. Despite the related studies being less frequent than those of carbazole, more studies need to be conducted. Fig. 4b and Fig. 4c show the relative reaction of 2-methylindole (2-MID)/octahydro-2-methylindole (8H-2-MID) and N-propylcarbazole (NPCZ)/dodecahydro-N-propylcarbazole (12H-NPCZ) systems [14]. When used as the organic hydrogen storage liquid, 2-MID has a hydrogen storage capacity of 5.76 wt% and can be hydrogenated to 8H-2-MID with catalysts such as Ru/Al2O3. Thus, 8H-2-MID can be completely dehydrogenated to form 2-MID. Similarly, NPCZ can be catalytically hydrogenated to 12H-NPCZ at 120\u00b0C-150\u00b0C with a hydrogen storage capacity of 5.43 wt%. The class of N-doped compounds has been the subject of much research, including the examination of some of the compounds produced during the hydrogenation process.Similar to mixed aromatic systems such as biphenyl/diphenylmethane, owing to the limitations of the physical and chemical properties of pure substances, some studies have explored the effects of hybrid systems of N-doped compounds [68]. Compared to a single compound, the lower freezing point of the mixture made it more beneficial for use in colder climates. The melting points of combinations containing various alkyl compounds were examined by Stark et al. [85]. They realized that the melting points might be lowered by using the appropriate combination of N-alkylcarbazole. Shuang et al. [86] investigated mixed liquid hydrogen storage systems and the impact of mixing in different proportions on the performance of the system. Consequently, they discovered that the melting point of the mixture dropped to 25\u00b0C when the system was mixed with 40 wt% 2-MID, 36 wt% NPCZ, and 24 wt% NECZ. With a solvent-free catalytic process and a high hydrogen storage rate, the hydrogen storage capacity could reach 5.64 wt%.Taking various promising LOHCs as the research object, the dehydrogenation performance differs significantly in technical, economic, and environmental aspects [87]. For the production rate of H2, TOL and DBT tended to be better than NECZ. For the economic feasibility, the use of NECZ took the higher investing cost to a large extent than DBT and TOL, with almost unit H2 production cost of 264.47 $ kgH2\n-1 while 54.94 $ kgH2\n-1 for DBT, and 19.94 $ kgH2\n-1 for MCL. Meanwhile, compared with DBT and MCH, the CO2 emissions per produced H2 were revealed to be showing environmental drawbacks of NEZ. Hence, the techno-economic performance of LOHCs should be comprehensively considered.When comparing the strengths and drawbacks of the LOHCs technology processes, it becomes obvious that a number of reasons can limit large-scale commercialization. The key obstacles include the high energy consumption, the difficulty of developing dehydrogenation catalysts, and the decrease of hydrogen storage performance as the number of cycles increases [88\u201390]. As a result, current research focuses on lowering energy consumption and developing high-performance catalysts. Major dehydrogenation catalysts are supported metal catalysts and tend to be prepared by impregnation [91], deposition precipitation [92,93], one-pot [94], and sol-gel methods [95], etc. Generally, these catalysts are loaded on carbon-based materials, Al2O3, TiO2, zeolite, and other carriers [96,97]. Since the hydrogenation and dehydrogenation of hydrogen storage liquids are reversible reactions, catalysts with high hydrogenation activity also perform well in dehydrogenation reactions. Catalysts are often categorized based on how metals are mixed. Monometallic catalysts (noble metal catalysts, non-noble metal catalysts [98\u2013100]), polymetallic catalysts [101], and other catalysts (such as boron nitride and metal complexes) are examples of different types of catalysts.(1) Monometallic catalystsMonometallic catalysts are normally categorized into metallic and non-metallic catalysts based on metal activity [89]. Palladium (Pd), platinum (Pt), ruthenium (Ru), rhodium (Rh), nickel (Ni), molybdenum (Mo), and copper (Cu) have all been extensively researched. The choice of the catalyst has a critical influence on the products and activity of the reaction. The proportion of different forms of d-electrons has been demonstrated to be related to the active component. In the case of noble metals, for example, Pd, Pt, Ru, and Rh all occupy 40% of the d-electrons in their atoms, and their activity is higher [102,103]. With high activity, Pd and Pt can significantly decrease the dehydrogenation temperature. The activity of Pd was maintained after a few cycles, demonstrating a distinct advantage in all types of LOHCs systems. Wang et al. [91] performed a comparative experimental study on the dehydrogenation of 12H-NECZ with graphene (rGO)-supported noble metal catalysts. The catalytic activity was found to follow the order Pd/rGO > Pt/rGO > Ru/rGO > Au/rGO in the experiments. The authors also compared Pd/rGO with the commercial catalyst Pd/Al2O3. Pd/rGO was shown to be superior to Pd/Al2O3 in many ways, and its catalytic activity remained unchanged within a certain range of reaction temperature. Sharma et al. [104] employed ruthenium metal as the active component of the catalyst. They discovered that at 120\u00b0C and a partial pressure of H2 of 60 atm for 2 hours, 100% conversion of benzene with 100% selectivity could be achieved. Despite the benefits of noble metal-supported catalysts in terms of activity, reaction rate, and service life, there is a demand for low-cost, high-efficiency catalysts for LOHCs.For the storage and release of hydrogen from LOHCs systems, Ni-based catalysts also exhibit outstanding catalytic activity. These catalysts are generally utilized in the form of supported or skeletal Ni for reactions [105]. Raney-Ni is widely utilized for skeletal Ni and shows a high activity due to its large specific surface area. Nevertheless, the conversion rate tended to drop gradually as the number of catalytic cycles and time increase. When it comes to the supported Ni, the catalysts Ni/SiO2\n[67], Ni/Al2O3-TiO2\n[106], and Ni2P/Al2O3\n[67] are commonly utilized. In addition to some supported metals such as Ni, Mo have also been studied as a non-noble metal, and its loading effect has a significant impact on the surface properties of the catalyst. However, Mo frequently exhibits a decrease in activity and produces more coking materials. Generally, non-noble metals have disadvantages in terms of reactivity, temperature, and service life when compared with noble metals, but they offer a remarkable economic advantage.(2) Polymetallic catalystsFor the features obtained by the different physicochemical qualities of single components of monometallic catalysts, the use of polymetallic or transition-metal catalysts can completely exploit the strengths of each component. This sort of catalyst can be improved to stimulate the breakdown of C-H bonds, improve the activity of intermediates, speed up the reaction, and even increase the selectivity of products. The addition of Ni, Cu, Sn, and other second components to noble metal catalysts can take advantage of the involved synergistic effects, such as AuxPdy/rGo [107] and Pd-M/TiO2\n[92]. Doping with polymetallic materials can considerably reduce the number of noble metals and economic pressure, while maintaining high reactivity and cycle stability. Wang et al. [94] studied the dehydrogenation performance of 12H-NECZ by using rGO as a carrier to prepare Pd and Cu bimetallic catalysts via a one-pot method. The Pd1.2Cu/rGO catalyst had the maximum catalytic activity, with 100% conversion of 12H-NECZ after 1 hour of reaction and 100% product selectivity after 7 hours. Yuki et al. [108] developed the binary alloy for the dehydrogenation of TOL/MCH, and they finally found that Pt3Fe/SiO2 acts as a highly active and durable heterogeneous catalyst with excellent toluene selectivity (>99%) and long-term durability. The core is that the excellent performance was derived from the synergy of each element (C-H activation of Pt, decoking of Fe, and toluene desorption of Zn). Numerous investigations have shown that the introduction of other metals into monometallic materials can result in positive synergy, improved catalyst activity and stability, and cost benefits [109\u2013112]. However, the characteristics of the catalysts are strongly affected by the processing procedures and types of the carriers. For the hydrogenation and dehydrogenation of various LOHCs systems, the introduction of the second or third metal should serve as the reference object. Above all, it is necessary to consider whether the addition of metals and the preparation procedures can interact positively with the hydrogen carriers.In addition to the active component, the metal-supported carrier is a crucial element that affects the performance of the catalyst. Several studies have discovered that catalysts with metal nanoparticles on the surface of carbon-based materials (e.g., activated carbon, graphene, and carbon nanotubes) can have strong hydrogenation and dehydrogenation activity [113,114]. Catalyst activity is increased when metal links are modified, allowing researchers to better regulate the reaction process and facilitate research for industrial applications. Above all, it is essential to assess whether the different metals and preparation processes used in the catalysts could lead to positive system synergy. Apart from supported catalysts, a growing number of other types of catalysts, such as boron nitride and metal complexes, have gradually become research hotspots in recent years. However, many of these catalysts have challenging production procedures and are difficult to apply in industrial settings.Overall, the reactivity of hydrogen storage qualities is affected by the types of metal, support, and carriers. In both monometallic and bimetallic systems, noble metals exhibit excellent catalytic characteristics. However, it is still necessary to balance costs and benefits. Bimetallic and polymetallic catalysts are currently recognized as the most promising strategies to improve the process. The possibility of lowering the quantity of noble metals while retaining catalytic activity holds a lot of promise. In addition to changing the catalysts in terms of the active component, carriers, and supporting mode, many researchers employ sulfur in conjunction with catalysts modification techniques in the petrochemical industry to optimize the catalytic activity [115]. In order to prepare the Pt/Al2O3 dehydrogenation catalyst in the dibenzyltoluene/perhydro-dibenzyltoluene system, Wasserscheid's team added a specific amount of sulfide. They found that sulfur occupied a low coordination sites of the supported Pt nanoparticles, which significantly boosted the dehydrogenation efficiency and lowered the production of byproducts [116]. Furthermore, by increasing the dispersion of active components, lowering the surface charge density, and maximizing the H2 spillover effect, the catalytic activity in the reaction of polycyclic aromatic hydrocarbons and heteroatom-doped compounds can be boosted [117\u2013119].Hydrogenation and dehydrogenation reactors are key components of the whole system, with dehydrogenation reactors receiving more research and development. The reactor has more rigorous pressure requirements since the hydrogenation process necessitates greater pressure. Stainless steel autoclaves, along with fixed-bed reactors and other laboratory equipment, are more commonly utilized for easy operation [65,83,120]. When the reaction is accompanied by tandem side reactions, the fixed-bed reactor allows continuous hydrogenation and effective interaction between the reactants and catalyst, resulting in high selectivity. Individual reactor units are utilized for batch hydrogenation in stainless steel autoclaves, which are relatively straightforward to operate. To maintain the reaction at a consistent pressure, a hydrogen reservoir was attached to the reactor during the reaction. According to the classification of reactants and reaction conditions, dehydrogenation reactors can be divided into two types: steady-state and non-steady-state [121,122]. Dehydrogenation reactions are typically categorized into three groups based on the phase of organic liquid hydride: gas-phase, liquid-phase, and \u201cwet-dry\u201d multi-phase dehydrogenation. The advantages and disadvantages of several types of reactors are listed in Table 3\n. Among what has been mentioned in the list, principal dehydrogenation reactors at the laboratory-research stage are \u201coil bath pot - three mouth flask\u201d [114,123], fixed-bed reactor [47], pulse jet reactor, and membrane catalytic reactor [124]. The \u201coil bath pot - three mouth flask\u201d apparatus is a batch reactor. The hydrogen-rich liquid is injected into the container after it has been heated to the temperature of the dehydrogenation reaction in the oil bath. The hydrogen removed by the reaction was cooled, separated by a serpentine condenser, and eventually collected. \u201cOil bath pot - three mouth flask\u201d is a simple-structured device of the reaction which enables the effective separation of reactants and products. However, the temperature of the dehydrogenation reaction is frequently a limiting factor, and it is more commonly utilized in systems with lower dehydrogenation temperatures, such as NECZ. Fixed-bed [125\u2013127], membrane reactors, and stainless steel autoclaves [128] are more often employed for systems with higher dehydrogenation temperatures and requirements for continuity. In a fixed-bed reactor, similar to hydrogenation, dehydrogenation is conducted by heating the catalyst to a given temperature and then transferring the organic hydrogen carrier into the unit. In the case of MCH, after being heated in the gasification chamber, it traveled through the catalyst bed in a gaseous condition and then reacted. Finally, the hydrogen was extracted and gathered.Therefore, a suitable reactor must be selected for the phase states of different reactants and mechanisms of dehydrogenation. There are numerous methods for optimizing and upgrading the reactor on a regular basis. To improve the dehydrogenation rate and purity of the produced hydrogen, multiphase dehydrogenation can be used instead of single-phase dehydrogenation. Microreactors can also be utilized to keep a consistent dehydrogenation temperature [121]. Some bottleneck difficulties in pulse injection reactors [129] and membrane-catalyzed reactors remain addressed, and commercial applications are still a long way off. Fixed-bed reactors offer a high conversion efficiency and can be utilized for continuous feeding. However, the diffusion of gas, which easily causes coke production and catalyst degradation, currently limits the reaction rate of fixed-bed reactors. During the dehydrogenation, it is necessary to take into account the activity of the catalyst, temperature, and reaction efficiency to control completion of the reaction. When the reaction is insufficient, it is also critical to avoid a reverse reaction.The development of several types of reactors for independent hydrogenation and dehydrogenation processes is progressing. Under ideal conditions, a high level of cycling performance can be reached by considering the full response system holistically. LOHCs technology is a three-step closed-cycle reaction that includes hydrogenation reactions, hydrogen carrier conveyance, and dehydrogenation reactions. Among numerous nations with relatively advanced and widespread hydrogen energy development, Japan has essentially completed the demonstration project of hydrogen storage and release, addressing hydrogenation generation and Gas-to-Liquid technologies and benefiting from hybrid fuel cells [133]. The integrated system for hydrogen storage and release is depicted in Fig. 5\n, which is a flow chart at the level of laboratory or small and medium-sized hydrogenation and dehydrogenation equipment. As shown in the Fig. 5, when completely hydrogenated, the product is condensed by employing a condenser, which discharges the excess exhaust gas and produces hydrogen storage carriers. Fully hydrogenated organic liquids can be stored and transported efficiently because of their high stability and cyclic performance. When the hydrogen is required, LOHCs are transferred by liquid storage tanks or other equipment to the dehydrogenation unit. The dehydrogenation is then carried out under the action of a specific condition. Finally, hydrogen is discharged from the hydrogen outlet of the unit and stored in a gas storage tank for later use. The organic liquid can be processed, transported, and recovered after dehydrogenation for further hydrogen storage.Owing to the advantages of safety, compatibility, and high hydrogen storage capacity of current liquid storage facilities, multiple energy sources can be transported and preserved. These features distinguish LOHCs technology from conventional hydrogen-storage systems. This aligns with the present call for a firm guarantee for the development of clean, low-carbon, safe, and efficient energy systems. This is expected to alleviate the problem of uneven distribution of energy in space and time.Since the reversible processes of LOHCs technology require specific reaction conditions, this ensures a high degree of stability for liquid organic hydrogen storage carriers. Because of the flammable and explosive characteristics and low density of hydrogen, its transport across oceans or other large-scale applications is particularly challenging. Liquid hydrogen, liquid ammonia hydrogen, and LOHCs are the three main technologies for bulk hydrogen transport now in use [10,134]. For the delivery of hydrogen, traditional LH2 technology necessitates absolute temperature control. The storage vessel is insulated as a result of the conditions of use, which can will increase the cost greatly. In the case of LAH2 technology, traces of ammonia remaining in the hydrogen after dehydrogenation cause severe degradation of the performance. In summary, high-pressure and liquid hydrogen are more suitable for short-haul transportation, but both systems have large upfront expenditures and administrative restrictions. As for LOHCs technology, hydrogen carriers can be stored and transported with existing oil and gas transmission pipelines, tankers, and storage tanks. The usage of LOHCs technology aims to store hydrogen through organic hydrogen storage liquids from raw hydrogen resources. Hydrogen storage tankers, tankers, pipelines, and ships are used to carry the carriers to their final destination [10,135,136]. Finally, catalytic dehydrogenation units release hydrogen for use in fuel cells, hydrogen refueling stations, and industrial production. At the same time, after cooling, the hydrogen-leaved liquids used for hydrogen storage can be recycled and stored for future use.\u201cChiyoda Chemical Construction\u201d in Japan has carried out the research and engineering test of large-scale LOHCs technology based on TOL/MCH, with the dehydrogenation conversion of MCH over 99.9%, selectivity of TOL over 99.9%, and catalyst life over 10,000 hours. As the representative enterprise, \u201cChiyoda Chemical Construction\u201d imported a total of 210 metric tons of hydrogen from Brunei to realize the transfer of hydrogen across the oceans in 2020. The first project of the worldwide hydrogen-energy supply chain has completed its demonstration phase. This chain was based on the production of hydrogen through natural gas reforming at the Brunei plant, employing a stable chemical as the carrier and then using conventional transportation to convey hydrogen to Japan across long distances. The goal of this project is to feed hydrogen to turbines to create electricity.In addition to being used for the bulk and transoceanic transport of hydrogen energy, LOHCs technology can also be applied to existing hydrogen energy vehicles [73,137]. It is possible to load hydrogenated cyclohexane straight into the vehicle and dehydrate it using an onboard dehydrogenation device in the case of hydrogenated cyclohexane. \u201cHynertech\u201d in Wuhan, the representative firm among the LOHCs technology-related enterprises in China has demonstrated a project for 1,000-ton NECZ plant using \u201cNECZ/12H-NECZ\u201d. The business has proposed the world's first fuel cell model that relied on liquid organic hydrogen as a source of energy in 2016. As mentioned in section 3.2, the hydrogen storage liquid developed in \u201cHynertech\u201d of is known as \u201chydrogen oil\u201d. Ultra-high-temperature gasification technology has been investigated for the production of \u201chydrogen oil\u201d in accordance with Chinese policy. This technology makes use of municipal and industrial solid waste, which is gasified at extremely high temperatures to produce hydrogen gas and \u201chydrogen oil\u201d. This not only reduces the environmental impact of hazardous waste, but it also improves energy efficiency and allows biomass to be combined with other emerging energy industries.In the context of carbon neutrality, the renewable energy industry in China is gaining traction and helping to lessen the reliance on imported energy. When it comes to long-distance transportation, conventional energy storage technologies (such as liquid fuels, electricity, and thermal energy) have limited by storage periods, high energy consumption, and low safety. As a result, positive energy storage technology research is required to realize the exploitation, transfer, and storage of important renewable energy supplies. In the case of hydrogen energy, the production and preparation of hydrogen usually consist of three categories: \u201cblue hydrogen\u201d, \u201cgrey hydrogen\u201d and \u201cgreen hydrogen\u201d [138]. Hydrogen is a prospective energy carrier that can be employed as a conversion medium for a variety of energy sources. LOHCs technology allows for efficient hydrogen energy storage and transmission, and it can also be used to collect renewable energy, large-scale distributed generation, and hydrogen. This one-of-a-kind connection makes it suited for long-term and large-scale storage and usage, allowing collaborative connectivity between diverse components of the energy network [135]. For the production of hydrogen, there are an increasing number of studies on renewable energy methods such as water electrolysis/solar photolysis, biomass fermentation, bioethanol reforming, and biomass chemical cracking [139]. Electrolysis of hydrogen from renewable energy sources and the processing of biomass feedstock have been shown in a growing number of studies to be key sources of hydrogen for increasing the amount of hydrogen produced [41,135]. When comparing the investment expenses of CH2, LH2, and LOHCs, the latter is only 32% of the former [140]. Germany has proposed the \u201cGET H2\u201d in 2020. The project seeks to develop industrial-scale green hydrogen production in regions with abundant wind and solar energy resources, and connect with downstream application. Fig. 6\n shows one of the methods in which energy is stored in a renewable system by LOHCs technology in the form of a hydrogen supply chain.Aromatics and N dopants are the critical systems available for LOHCs technology. Existing hydrogen storage liquids are commonly prepared from coal and petroleum. Pure chemicals such as triphenylbenzene [benzene, toluene, and xylene, (BTX)] tend to obtained from coal by thermal fractional distillation, and purification in turn [141,142]. When prepared from petroleum, the fractionated products contained fewer aromatic compounds and a large number of alkanes. As a result, numerous bond-breaking reforming and aromatization procedures must be used to transform them into aromatic hydrocarbons [143]. Traditional fossil energy sources utilized as raw materials are not infinite, and the cost is considerably large. To create aromatic and N-doped compounds for chemical or other applications, economical and sustainable sources must be investigated. Biomass resources, which are primarily made up of elements like carbon, hydrogen, and oxygen, offer a lot of promise for development as carbon-neutral renewable energy sources [144,145]. The three primary components of biomass, cellulose, hemicellulose, and lignin, have varied structural qualities and hence require different conversion methods and applications. Within cellulose and hemicellulose, equipped with a five-membered or six-membered cyclic polysaccharide structure, hemicellulose acts as a molecular binder bound between the cellulose and lignin [146]. Lignin, a three dimensional reticulated aromatic ring structure wrapping and reinforcing cellulose and hemicellulose, is a kind of nature organic compound made up of interlaced C-C and C-O bonds with a complex structure [147]. The aromatic and furan rings found in direct pyrolysis products of lignin can be exploited to make biomass fuels, light aromatics, and other small-molecule compounds [148,149]. Benzene is a basic carbon frame structure of biomass, consisting of a \u201cbenzene ring\u201d encircled by six C atoms [150]. The basic structure of almost all of the products produced by various biomass conversion processes is benzene rings [151]. As a result, using biomass as the source of reaction carrier in the LOHCs system has a certain amount of practicality and economic benefit.Based on the technical principles of aromatics extraction from biomass, it can be divided into direct and indirect routes for preparation. Direct preparation refers to the direct conversion of biomass into aromatic products in a reactor without further processing. This is the biomass-catalyzed thermal cracking method that is currently being explored and used the most. The conversion of biomass into intermediate products via a variety of techniques, followed by the manufacture of aromatics, is referred to as indirect preparation technology [152]. The second method has the benefit of allowing for the targeted conversion of intermediates to aromatics using existing conversion technologies with high yields and efficiency. However, the production method is lengthy, resulting in significant raw material waste and energy usage.When using intermediates in the form of syngas (CO, H2), there are three major pathways for the preparation of aromatics: methanol/dimethyl ether, Fischer-Tropsch synthesis, and direct preparation using aromatization catalysts. Methanol-to-aromatics (MTA) technology [155\u2013158] is one of the ways listed above that has been developed and applied in industry. The aromatic compounds were selective up to about 80% during the process, and the methanol is almost completely converted. The main products of syngas synthesis by Fischer-Tropsch are alkanes and olefins, with lesser yields of aromatics. Another technology for the direct preparation of aromatics from syngas is mainly obtained through improving the catalysts. The products can be directly converted to aromatics by selecting aromatized catalysts in the above two preparation procedures [157]. As indicated in Fig. 7\n, the synthesis of aromatic hydrocarbons entails various processes. First of all, through methods such as hydrolysis, hydrogenation, or fermentation of biomass, oxygenated compounds, such as sugars, aldehydes, and alcohols are produced by the action of microorganisms. Subsequently, a series of operations including reform, dehydrogenation, and cyclization, were conducted to produce aromatics. This is attributable to the fact that the three primary components of biomass can be hydrodeoxygenated or enzymatically broken down into tiny molecules like alcohols, furans, and phenolic aldehydes. Under appropriate reaction circumstances, these molecules can be continually transformed into aromatic compounds [153]. Technologies of bioFormingTM and bio-based isobutanol to aromatics, which produce aromatic compounds from 100% renewable plants and sugars, are already commercially viable in this field. Overall, the method of producing indirect aromatics from biomass is time-consuming, with additional intermediate steps in the reaction resulting in poorer product yields, higher reaction energy consumption, and feedstock waste. To put it another way, commercial applications are extremely tough to implement, necessitating further refinement and optimization.Fig 8.\n\nAromatic products like BTX and olefins can be synthesized from biomass feedstock (e.g., wood, agricultural products, organic solid waste, and fiber waste) via catalytic thermal cracking technology, which involves a series of complex reactions like depolymerization, isomerization, and polymerization. The products of catalytic pyrolysis of biomass include bio-oil, coke, and combustible gases. Depending on the reaction conditions, the proportion of products obtained varied. Catalytic slow pyrolysis mostly produces coke, while catalytic medium-speed pyrolysis generates combustible gas and catalytic fast pyrolysis (CFP) generates bio-oil [159\u2013161]. The manufacture of aromatic compounds by direct catalytic pyrolysis is primarily performed through the CFP of biomass due to the high amount of aromatic products with the most basic benzene ring structure.Owing to the different structures and types of biomass feedstock, the contents of the three main components vary accordingly. Lignin has a higher selectivity for aromatic compounds in catalytic cracking products than the other two primary components of biomass due to its high H/C ratio. During pyrolysis, cellulose is the first to undergo bond breakdown, resulting in aromatic compounds and olefins with a high added value. As a result, biomass feedstock with high lignin and cellulose content offers a better potential for aromatic chemical production employing catalytic pyrolysis procedures [162]. It has been reported that the cellulose content of biomass feedstock affects the bio-oil content. More bio-oil can be produced with a greater cellulose ratio, which contributes to the generation of aromatics. Zhang et al. [163] chose ZSM-5 as a catalyst to evaluate the yield of CFP aromatic compounds products with varying pine-to-alcohol ratios. They discovered that the yield of aromatic compounds could be enhanced with the increase of the H/C ratio of the reactants. The conversion of organic matter to aromatic substances within the bio-oil can be facilitated by an increased amount of hydrogen. However, because the type of raw material has a considerable influence on the product, numerous studies have been conducted on model compounds or single components [164]. In addition, model compounds such as furan, furfural, and glucose are commonly employed to research pyrolysis mechanisms, products, and reaction routes. In the hydrodeoxygenation of lignin, Diao et al. [146] prepared MoCo9S8/Al2O3 as a catalytic material with a balance between economy and temperature stability. The easy deactivation of sulfide catalysts was focused on to achieve an efficient one-step conversion of lignin to aromatic compounds, with 99.8% conversion and 91.0% yield of benzene.Alkali metals, alkaline earth metals, metal oxides, and molecular sieves are the most typical catalysts employed in biomass catalytic pyrolysis reactions. Molecular sieve catalysts are widely used in the field of catalytic pyrolysis of biomass to prepare aromatic compounds. Because of its unique internal aperture structure and acidic sites on the surface, ZSM-5, for example, can reduce the carbon build-up problem and promote selectivity of the target product to some extent [159]. The aperture structure, acidic sites, silica to aluminum (Si/Al) ratio, and particle size are also critical factors that affect catalytic activity. More studies on modification approaches, such as increasing metal supported [165,166], employing mesoporous catalysts [162,167], and modifying the acid-base treatment order, are needed in the future.Thermal catalytic conversion and ammonification is the process of making N-doped chemicals from biomass using catalytic pyrolysis in the presence of a nitrogen donor (TCC-A). N-doped compounds include amines, nitro substituted, nitrile, and some N heterocyclic compounds with aromatic structures, such as indoles, pyridines, and pyrroles [149]. Nitrogen donors can be divided into three types: gaseous ammonia, high nitrogen biomass, and solid ammonia sources. Additionally, the ammonium salt is a nitrogen donor since it creates ammonia gas when heated. In LOHCs technology, N-heterocyclic aromatic compounds generated by heat catalysis and ammonification can be seen as hydrogen storage agents.The indoles have a chemical structure that is similar to that of furans. Therefore, by introducing an external ammonia source, catalytic pyrolysis can be used to prepare indoles to catalyze the formation of furans from biomass. For the production of indoles, several furan derivatives such as 2-methylfuran and 2-methylfurfural can be employed as intermediates or source materials. TCC-A technology has been claimed to be capable of directly converting biomass-derived furans and furfurals to indoles for the manufacture of N-doped chemicals from biomass [168]. Xu et al. [149] experimentally discovered that TCC-A technology could transform raw biomass with complex compositions into N-doped compounds and N-containing biochar. By altering the reaction conditions, the percentage of pyridines and indoles could be selectively controlled. Yao et al. [169] exploited furfural, which is obtained from biomass, as the feedstock for thermal catalytic transformation and zeolite ammonification to convert furfural to indole compounds. Ultimately, they concluded that the conversion pathway was \u201cfurfural-'furfural-imine' furan-pyrrole-indole\u201d. Lactose, acetylation, and furan amination are the most common processes for making pyrroles. Pyrroles prepared from biomass can be obtained by rapid in situ pyrolysis of cellulose under ammonia atmosphere using HZSM-5 [170\u2013172]. In industrial settings, pyrrole is usually prepared from furans derived from petroleum under the catalysis of solid acid catalysts. Yao et al. [173] found that the catalytic fast pyrolysis of cellulose in a low-temperature ammonia atmosphere is an effective way to selectively prepare pyrroles. When the reaction temperature was controlled at 400\u00b0C, they took \u03b3-Al2O3 as the catalyst and a selectivity of up to 89.5% was achieved with a catalyst/cellulose ratio of 2. The classic method of \u201ccatalytic synthesis\u201d was used to make pyridine compounds, and the raw material used was primarily glycerol. Pyridine tends to be synthesized by aldehydes, alcohols, and unsaturated hydrocarbons in an ammonia-rich environment. As a result, when pyridines are utilized as target products, biomass can be first converted into intermediate products such aldehydes, ketones, and alcohols.Therefore, by utilizing the current technologies of biomass conversion, suitable and prospective hydrogen carriers can be purposefully and selectively prepared for LOHCs technology in the future. Overall, the utilization of biomass in the synthesis of aromatic and N-heterocyclic compounds to be used as organic hydrogen storage liquids has a great prospect. However, further studies are required to create breakthroughs in terms of enhancing the stability and selectivity of the reaction.The discrepancies between the raw biomass pyrolysis products and target carriers of the LOHCs system should be compared to directionally regulate the parameters of each process. Furthermore, according to the type of selected biomass, different catalytic reaction routes and pretreatment modes are necessary to optimize the yield of the targeted product. During the catalytic pyrolysis processes of biomass, it is important to focus on the pathways (e.g., purification and ammonification) that break bonds to form monomeric compounds for subsequent operations. On the basis of the positive or negative feedback of the reactions of biomass-based organic hydrogen storage liquids, it is necessary to adjust the conditions to optimize the characteristics within the operating range. At present, the technology for the catalytic pyrolysis of biomass for the preparation of aromatic and N-doped compounds is relatively advanced. Major studies are more likely to explore the influences of the feedstock type, reaction conditions, catalysts, and other factors. Meanwhile, the advancement of LOHCs technology has been steadily advancing, with more research currently concentrated on lowering the dehydrogenation temperature, improving catalysts, and reducing the energy consumption. In addition, certain businesses around the world have already succeeded in commercializing and trading LOHCs technology. This pattern suggests that the technology has reached a high level of maturity and usability. However, the technologies producing LOHCs have abundant disadvantages in terms of sustainability or material expense. Therefore, it is necessary to explore economic and sustainable renewable carriers. Through the review of previous researches, there are few studies on the generation of hydrogen carriers through biomass conversion technology. Combined with the reaction principles of the various LOHCs systems, biomass can be applied to LOHCs through existing methods of conversion. Future research in this field should improve the catalysts and other factors through targeted modifications to produce organic hydrogen storage liquids.LOHCs technology can effectively avoid some of the shortcomings of conventional hydrogen storage technologies, which is aimed at meeting the requirements of long-term and large-scale hydrogen storage in an ambient environment. Fortunately, this technology possesses excellent compatibility with existing equipment for oil and gas production, has tremendous development prospects, and is latent. Because of the unique feature of carbon neutrality and the structure of its basic components, biomass is widely used in the clean energy sector. The most common use of biomass is to make bio-liquid fuels with aromatics as the most basic structure, which can be prepared by a variety of depolymerization processes. The products can also be converted into other chemicals through advanced processes to realize high-value utilization. Petroleum, coal, and other raw materials are commonly used to make hydrogen storage liquids in existing LOHCs systems. However, because of the limited amount of traditional energy, newer manufacturing methods should be investigated. Based on the above review, there are still several challenges for the development LOHCs technology. (1) The temperature of dehydrogenation and the energy consumption are difficult to decrease. (2) Stable, efficient, and economic catalysts used for hydrogenation and dehydrogenation need further development. (3) The initial cost to purchase LOHCs materials is extremely high. (5) It is necessary to comprehensively consider the technical, economic, and environmental performance of the system.Taking advantage of the fact that benzene is the fundamental unit in biomass, coupling LOHCs with technologies related to the catalytic pyrolysis of biomass is promising. Switching the orientation and research ideas is critical for biomass-based organic hydrogen storage liquids. To prepare this type of liquid according to the specific reaction equipment, appropriate technologies must be established. Therefore, there are numerous difficulties to solve in the utilization of renewable biomass-based hydrogen carriers.\n\n(1)\nAccording to the findings of this paper, there are numerous challenges to overcome when combining catalytic pyrolysis of biomass for the manufacture of aromatic and N-doped chemicals with LOHCs technology. Due to the coverage of active sites caused by coking, the blockage of channels, and the agglomeration of active metals, catalysts are easy to inactivate. Hence, it is necessary to prepare high-performance catalysts to ensure the directional, efficient and stable conversion of biomass pyrolysis.\n\n\n(2)\nDistinct components of biomass have different conversion processes. On one hand, Multifunctional catalysts should be designed to achieve multiple conversion paths. On the other hand, to separate components without significantly affecting the structure, advanced pretreatment technology is necessary.\n\n\n(3)\nA few components in the products prepared by catalytic pyrolysis may not have the ability to store hydrogen, which requires to explore the technology of further separation and purification.\n\n\n(4)\nMost of the existing catalysts are used for LOHCs of single substance, which may not be applicable to the LOHCs prepared by catalytic pyrolysis of biomass. Therefore, it is necessary to design suitable and efficient hydrogenation and dehydrogenation catalysts according to the characteristics of the products.\n\n\nAccording to the findings of this paper, there are numerous challenges to overcome when combining catalytic pyrolysis of biomass for the manufacture of aromatic and N-doped chemicals with LOHCs technology. Due to the coverage of active sites caused by coking, the blockage of channels, and the agglomeration of active metals, catalysts are easy to inactivate. Hence, it is necessary to prepare high-performance catalysts to ensure the directional, efficient and stable conversion of biomass pyrolysis.Distinct components of biomass have different conversion processes. On one hand, Multifunctional catalysts should be designed to achieve multiple conversion paths. On the other hand, to separate components without significantly affecting the structure, advanced pretreatment technology is necessary.A few components in the products prepared by catalytic pyrolysis may not have the ability to store hydrogen, which requires to explore the technology of further separation and purification.Most of the existing catalysts are used for LOHCs of single substance, which may not be applicable to the LOHCs prepared by catalytic pyrolysis of biomass. Therefore, it is necessary to design suitable and efficient hydrogenation and dehydrogenation catalysts according to the characteristics of the products.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Natural Science Fund for Excellent Young Scholars (China) (Grant No. 51822604).", "descript": "\n                  Hydrogen has attracted widespread attention as a carbon-neutral energy source, but developing efficient and safe hydrogen storage technologies remains a huge challenge. Recently, liquid organic hydrogen carriers (LOHCs) technology has shown great potential for efficient and stable hydrogen storage and transport. This technology allows for safe and economical large-scale transoceanic transportation and long-cycle hydrogen storage. In particular, traditional organic hydrogen storage liquids are derived from nonrenewable fossil fuels through costly refining procedures, resulting in unavoidable environmental contamination. Biomass holds great promise for the preparation of LOHCs due to its unique carbon-balance properties and feasibility to manufacture aromatic and nitrogen-doped compounds. According to recent studies, almost 100% conversion and 92% yield of benzene could be obtained through advanced biomass conversion technologies, showing great potential in preparing biomass-based LOHCs. Overall, the present LOHCs systems and their unique applications are introduced in this review, and the technical paths are summarized. Furthermore, this paper provides an outlook on the future development of LOHCs technology, focusing on biomass-derived aromatic and N-doped compounds and their applications in hydrogen storage.\n               "}
{"full_text": "Rising global warming due to CO2 emission from fossil fuel combustion and increasing energy demands have led researchers all over the world to focus on developing technologies for clean energy production and storage, such as fuel cells and batteries. Unitised regenerative fuel cells (URFCs) represent systems that can in one instance work as electrolysers (splitting water molecules to produce hydrogen fuel and oxygen) and in the next as fuel cells (chemically \u201ccombusting\u201d H2 fuel to produce electricity) [1,2]. For URFCs to produce high currents, high-performing bifunctional catalysts for the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) are of great importance [3\u20135]. However, producing catalysts that can sustain this transition from one operating mode to another has proven to be very difficult.Currently, the most efficient OER catalysts are considered to be ruthenium (Ru) and iridium (Ir) oxides, while platinum (Pt) has been regarded as the best performing catalyst for the ORR [6,7]. However, their high cost and low natural abundance rule out these metals for applications in renewable energy technologies in the long term.ORR is considered as the main limiting step in the fuel cell mechanism due to its sluggish kinetics and the high overpotential required [8]. The thermodynamic potential of O2 reduction versus the normal hydrogen electrode is 1.23\u00a0V. Electrode materials undergo oxidation at such a high potential, so that the electrode surface is no longer composed of a pure metal catalyst, but also of metal oxide. Thus, the Pt surface at high potentials is a mixture of Pt and PtO, resulting in an open circuit potential lower than 1.23\u00a0V, depending on the Pt to PtO ratio. The catalytic activity of Pt towards the ORR strongly depends on the O2 adsorption energy, the dissociation energy of the O-O bond, and the binding energy of OH to the Pt surface.Due to the high cost of Pt-based catalysts, researchers have focused their attention on nickel (Ni), a metal from the same group as Pt, but much more naturally abundant [9]. In recent years, Ni-based electrocatalysts in different forms (e.g., nanoparticles (NPs), foams, alloys, oxides, phosphides, metal\u2013organic frameworks) have been proposed as efficient substitutes for expensive metals as catalysts for the OER, the ORR, and the hydrogen evolution reaction (HER) [10]. Thus, Ni-based catalysts have been reported to exhibit high susceptibility toward surface adsorption of O2, a crucial step in the ORR. The work that pointed out the potential of Ni and encouraged Pt alloying was carried out by Stamenkovi\u0107 et al. [11]. A density functional theory study of the adsorption of OH and H2O on Pt3Ni(111) surface showed that in the presence of a Ni sublayer with 50 at.% Ni, adsorbed OH reacted with H+ to form H2O with a positive shift of 0.1\u00a0V in the reversible potential.In the present study, we investigate the possibility of increasing the catalyst\u2019s activity while simultaneously reducing its cost by lowering the amount of Pt and replacing it with Ni. Introducing another metal to Pt can have a strong effect on the electronic structure of the Pt catalyst and the Pt\u2013Pt interatomic distance, thus changing the electrochemical activity of the Pt catalyst. Metal NPs have been supported on three different metal oxides in order to overcome the problem of support degradation, another bottleneck in the commercialisation of URFCs. Pt-based catalysts are typically supported on carbon black, but advanced carbon materials, as well as non-carbon supports including transition metal oxides, carbides, and nitrides, have been recently suggested to improve the support/catalyst\u2019s durability [12].Details of the preparation and characterisation of the catalysts are given in the Supplementary information (SI). The nominal Pt loading on the supports was set to 20\u00a0wt% and for the bimetallic catalyst Pt:Ni the weight percentage was set to 10:10. The obtained loadings were confirmed by inductively coupled plasma mass spectrometry (ICP-MS) and the composition, structure, and morphology were further examined by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and transmission electron microscopy (TEM). Details of the electrochemical measurements are also provided in the SI.ICP analysis results reveal that the metal loading was close to the targeted nominal values (Table S1). The higher content of Ni in the PtNi/Mn2O3-NiO catalyst can be attributed to the Ni content in the support material itself.XRD patterns reveal the diffraction peaks of spinel Mn2O3 (Fig. 1\n), with additional details being given in the SI. Diffraction peaks of anatase TiO2 and NiO were observed for Mn2O3-TiO2- and Mn2O3-NiO-supported catalysts, respectively. The characteristic peak of Pt was clearly observed for the Pt-based catalysts, while no Ni peaks were observed for the PtNi-based catalysts.XPS analysis (Fig. 2\n) revealed sharp peaks at binding energies of 643.5, 655.8, and 530\u00a0eV, attributed to the Mn2p3/2, Mn2p1/2, and O1s regions, respectively [13,14]. The presence of Pt is confirmed by the peaks at 314 and 331\u00a0eV as Pt4d5/2 and Pt4d3/2 shake-up (satellite) peaks, and also at 71 and 74\u00a0eV as Pt4f7/2 and Pt4f5/2 shake-up (satellite) peaks, respectively [15,16]. The Ni2p peaks were observed between 855 and 879\u00a0eV [17] with their intensity increasing upon addition of NiO to the Mn2O3 support. PtNi alloy formation was confirmed by the positive chemical shifts of Pt4d5/2 and Pt4d3/2 when compared to the unmodified Pt catalysts (Fig. 2\nb,d,f) [18].Formation of Pt and PtNi nanoparticles of size\u00a0<\u00a04\u00a0nm was confirmed by TEM analysis (Fig. 3\n). Somewhat bigger nanoparticles (4 to 13\u00a0nm) were observed in the case of Pt/Mn2O3-TiO2.Cyclic voltammograms (CVs) recorded in O2-saturated 0.1\u00a0M KOH reveal a clear peak corresponding to O2 reduction (Fig. 4\n\n, S1, S2, S3). The expected enhancement of the activity of the catalysts towards the ORR, in terms of lower onset potential, E\nonset, and higher current density, could be observed upon grafting Pt NPs (20\u00a0wt%) on the oxide supports [19]. Further comparison of catalysts with Pt NPs (20\u00a0wt%) with those where 10\u00a0wt% of Pt was replaced with 10\u00a0wt% of Ni, showed similar diffusion-limited current densities. Rotating disc electrode voltammograms indicated the highest activity for the ORR in the case of Mn2O3-NiO-supported Pt and PtNi NPs in terms of the highest current density at a given rpm. The activity of metal NPs supported on metal oxides has been reported to depend on the size of the NPs, the oxide crystal phase/morphology (being in an appropriate oxygen adsorption mode) [20] and the support-induced modification in the electronic properties of the metal NPs [21].Furthermore, these two catalysts showed promising results when compared to the reference Pt/C (40\u00a0wt% Pt) sample, as diffusion-limited current densities of \u22126.44, \u22124.48, and \u22124.32\u00a0mA\u00a0cm\u22122 were recorded at 1800\u00a0rpm for Pt/C, PtNi/Mn2O3-NiO, and Pt/Mn2O3-NiO, respectively (Table 1\n\n), with the catalysts studied herein containing two or four times less Pt than the commercial Pt/C.ORR Tafel analysis (Fig. 5, S4-S6\n) revealed comparable Tafel slope, b, values for Pt- and PtNi-based catalysts, comparable with or even lower than that of Pt/C (Table 1). The lowest b value was found in the case of PtNi/Mn2O3-NiO, indicating its somewhat higher activity compared to the other catalysts. Note that dual Tafel slope values were observed, indicating different reaction mechanisms at different potentials. Detailed analysis of the changes in Tafel slopes conducted by Shinagawa et al. [27] showed that the slope depends on the geometry of the sample as well as the rate-determining step.The number of electrons exchanged, n, during ORR at the studied catalysts was determined by Koutecky-Levich analysis to be between 3 and 4 (Fig. 5, S4-S6, and Table 1), suggesting that the direct 4e--mechanism (see SI) was the preferred one.The double-layer capacitance, C\ndl, determined from the CV study (Fig. 6\n\n, S7), was found to be the highest for Pt/C (3.10 mF cm\u22122), reflecting its largest electrochemical active surface area, followed by PtNi/Mn2O3-NiO (2.67 mF cm\u22122). The C\ndl values of PtNi/Mn2O3 (0.58 mF cm\u22122) and PtNi/Mn2O3-TiO2 (0.29 mF cm\u22122) were found to be notably lower.Although all catalysts showed good stability, i.e., constant current density with time, it is worth noting that PtNi/Mn2O3-NiO produced the highest current density seen throughout the entire study (Fig. 7\n).Samples containing only metal oxides showed high activity for the OER (Fig. 8\n), as suggested by Song et al. [28]. Pt-based catalysts showed slightly decreased catalytic activity compared to the pure oxides. Damjanovic et al. [29] have established that in acidic media, oxide films forming on the Pt surface effectively reduce the catalytic activity of Pt for OER in an exponential manner. Oxide film formation is also present in alkaline media at potentials over the OER onset potential, thus decreasing the activity of Pt. The obtained results suggest that NiO is the active site for the OER, unlike the ORR where PtNi NPs provide the active sites. The high OER activity of Ni and NiMn oxides, surpassing that of Pt/C, has been demonstrated previously [9,30,31].OER Tafel regions are shown in Fig. 8\n(d,e,f); the slight curvature at higher current densities indicates where IR becomes significant [6]. The Tafel slopes were found to be comparable for the catalysts studied, with the exception of PtNi/Mn2O3-TiO2 (Table 2\n), indicating their similar activity for OER. The Mn2O3-NiO-supported catalysts stand out for the lowest values of b and E\nonset. Furthermore, these catalysts have lower overpotential values at a current density of 10\u00a0mA\u00a0cm\u22122\n, \u03b710, compared with the Mn2O3- or Mn2O3-TiO2-supported ones. Using the Mn2O3-NiO samples, the current density at an overpotential of 0.4\u00a0V, j\n400, reached values a few times higher than those obtained using the other oxides. The highest j\n400 was recorded using the PtNi/Mn2O3-NiO catalyst.The EIS study showed that Pt/Mn2O3-NiO had the lowest R\nct (78\u00a0\u03a9), followed by PtNi/Mn2O3-NiO (106\u00a0\u03a9) (Fig. 9\n and Table S2). PtNi/Mn2O3-TiO2 had a significantly higher value of R\nct \u2248 770\u00a0\u03a9. For reference, commercial Pt/C has R\nct of ca. 433\u00a0\u03a9. The electrolyte resistance, R\ns, (45\u201362\u00a0\u03a9) values reflect small variations in the electrode distances and cell geometry.The stability tests of the Pt- and PtNi-based catalysts, as well as Pt/C, showed a decrease in OER current densities with time (Fig. 10\n\na). Still, it is worth noting that this decrease was the least pronounced in the case of PtNi/Mn2O3-NiO (43%). Interaction between metal NPs and the metal oxide support has been reported to prevent migration or agglomeration of metal NPs, or their detachment from the support [12]. A drop in current density as high as 87% was recorded in the case of commercial Pt/C.To simulate the two operation modes of a URFC, an experiment switching between OER and ORR modes was run with PtNi/Mn2O3-NiO, where the O2 generated during the OER mode was subsequently reduced during the ORR mode (Fig. 10\nb). The ORR currents were steady, but a decrease in the OER current density in the first hour was observed. Still, after the initial drop, the OER current density was relatively stable.Nine different samples of Pt and Pt Ni NPs supported on binary metal oxides were tested as bifunctional electrocatalysts for the ORR and the OER, for possible use in URFC technology. The introduction of Ni resulted in higher electrocatalytic activity at a lower cost, while the introduction of metal oxide supports led to improved stability. Although all samples showed good activity for both reactions, the PtNi/Mn2O3-NiO sample showed the highest activity for the ORR in terms of the lowest onset potential, the highest diffusion-limited current density, and the lowest Tafel slope. This catalyst also showed the highest peak current density corresponding to O2 reduction and the highest stability during the ORR. The results indicated a diversity of active sites for O2 reduction and evolution; while PtNi NPs act as highly active catalytic sites for the ORR, the NiO active sites boost catalyst activity for the OER. All samples showed a decrease in stability with time under OER conditions. Nevertheless, PtNi/Mn2O3-NiO showed the lowest current drop, so further studies should focus on improving its stability in the electrolysis mode.\nDu\u0161an Mladenovi\u0107: Investigation, Formal analysis, Visualisation, Writing - original draft. Diogo M.F. Santos: Conceptualisation, Visualisation, Writing - review & editing. Gamze Bozkurt: Investigation, Formal analysis, Writing - original draft. Gulin S.P. Soylu: Investigation, Formal analysis, Writing - original draft. Ay\u015fe B. Yurtcan: Conceptualisation, Investigation, Writing - original draft. \u0160\u0107epan Miljani\u0107: Conceptualisation, Supervision. Biljana \u0160ljuki\u0107: Conceptualisation, Writing - review & editing, Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to thank the Ministry of Education, Science and Technological Development of Republic of Serbia (contract no. 451-03-68/2020-14/200146), as well as Funda\u00e7\u00e3o para a Ci\u00eancia e a Tecnologia, Portugal, for a research contract in the scope of programmatic funding UIDP/04540/2020 (D.M.F. Santos) and contract no. IST-ID/156-2018 (B. \u0160ljuki\u0107).Supplementary data to this article can be found online at https://doi.org/10.1016/j.elecom.2021.106963.The following are the Supplementary data to this article:\n\nSupplementary Data 1\n\n\n\n", "descript": "\n                  Three different metal oxides based on Mn2O3 with TiO2 or NiO were synthesised. Pt or PtNi nanoparticles were anchored on each support, creating a set of nine samples that were tested for the oxygen reduction reaction (ORR) and the oxygen evolution reaction (OER). PtNi/Mn2O3-NiO showed the most promising results for ORR as evidenced by the lowest Tafel slope, the highest diffusion-limited current density and number of electrons exchanged, along with the highest stability. The best performance of PtNi/Mn2O3-NiO reflects its highest electrochemical surface area and the lowest charge-transfer resistance. Furthermore, this catalyst showed high activity for the OER as evidenced by the low Tafel slope and high current density at an overpotential of 400\u00a0mV. The present study indicated different active sites for the two reactions, i.e., PtNi NPs for the ORR and NiO for the OER.\n               "}
{"full_text": "Carbon-free on-site hydrogen production and hydrogen storage technologies have recently attracted much attention for fuel cell applications that require high-purity hydrogen to generate environmentally clean electricity [1\u20134]. In general, hydrogen produced from on-site natural gas reforming contains carbonaceous substances (CO, CO2) that degrade cell performance even at extremely low concentrations [5\u20138]. Therefore, a subsequent hydrogen purification system is essential after the reformer. Alternatively, the production of COx-free hydrogen via catalytic ammonia decomposition, which only produces hydrogen and nitrogen gases, is considered a more desirable option for direct use as a fuel in proton exchange membrane fuel cells (PEMFC). Furthermore, ammonia has significantly higher gravimetric and volumetric hydrogen storage capacities (17.7\u00a0wt% and 108 g\u22c5L\u22121, respectively, at 20\u00a0\u00b0C and 857 kPa), making it readily stored and transported, which is beneficial for on-site hydrogen production [9,10].In general, ammonia decomposition requires a high energy input (>400\u00a0\u00b0C) owing to thermodynamic limitations [11]. Thus, a well-designed catalyst must be utilized to increase ammonia conversion and hydrogen yields at relatively low reaction temperatures. Many noble and non-noble metals such as Ru [12\u201316], Ni [17\u201321], Rh [22,23], Co [10,24], Ir [25,26], and Fe [27,28] have been studied for ammonia decomposition. Among them, Ru-based catalysts, such as Ru/Al2O3\n[29,30], Ru/CNT [13,31\u201333], Ru/zeolite\u00a0Y [9], Ru/SiO2\n[2,34], and Ru/TiO2\n[28] have shown promising catalytic activity. In particular, the Ru/CNT catalyst exhibited the highest ammonia conversion; however, low thermal stability under a hydrogen atmosphere is an issue, which can be attributed to the methanation reaction of the carbon sources of the support itself [13,35].In addition to incorporating different types of metals, numerous studies have been conducted to adjust the basicity of the support surface to further improve catalytic activity. This is because the well-known recombinative nitrogen desorption step, which is considered the rate-determining step (RDS) for ammonia decomposition over Ru-based catalysts [36], is strongly dependent on the basicity of the catalyst support [37]. For example, adding promoters such as La, Ce, K, and Mg to the catalyst contributes to an increase in the rate of desorption of nitrogen, which is attributed to the increase in surface basicity [38\u201340]. The increase in the electron density of Ru due to the electron-donating catalyst surface weakens the Ru-N binding energy, thereby increasing catalytic activity.In our previous studies, Ru impregnated on a La-doped Al2O3 catalyst was developed in powder form, and was highly active and stable under ammonia decomposition conditions [41]. A significant increase in ammonia conversion was obtained when more than 10\u00a0mol% of La was incorporated into the Al2O3 support. This was predominantly attributed to the formation of the LaAlO3 phase, which strongly interacted with the Ru active sites (a strong metal-support interaction), thus limiting the agglomeration of Ru particles and enhancing the catalytic activity. Al2O3 has been extensively utilized as a commercial catalyst support for ammonia decomposition because of its high surface area, thermal stability, and chemical resistance under high temperatures conditions [29,42]. Ammonia decomposition is a structure-sensitive reaction; the formation of well-known active sites of Ru (B5 sites) is reported over the Al2O3 support [15]. Recent research on Al2O3-based catalysts has been conducted extensively on the non-noble metals such as Ni, Co and Fe [43\u201346]. Despite the advantages, moderate acidity of Al2O3 surface is an issue where higher basicity of the support is required to facilitate nitrogen desorption. With respect to Ru/La-Al2O3 catalyst, these acidic sites of the Al2O3 support were covered by bulky La particles, which increased the basicity of the Ru/La-Al2O3 catalyst. The catalyst was stable for more than 120\u00a0h under a gas hourly space velocity (GHSV) of 10,000\u00a0mL/gcat\u22c5h and a reaction temperature of 550\u00a0\u00b0C. The catalyst was further pelletized in a follow-up study in which it was utilized in a COx-free 1\u00a0kW-class hydrogen power pack, including a dehydrogenation reactor, an adsorbent tower, and a 1\u00a0kW-class polymer electrolyte membrane fuel cell [47]. The system was tethered to a drone with a flight time of over 2\u00a0h. For this application, the weight loading of La was optimized to 20\u00a0mol% (tested for 0, 10, 20, 30, 40 and 50\u00a0mol%) where a volcano-shaped activity trend was observed. The results indicate that a significant loss of Ru active sites occurs at high La loadings.In the case of the pelletized catalyst, the metal nanoparticles tended to penetrate the bulk support during the synthesis process, making them inaccessible to the incoming reactant molecules especially under high space velocity conditions. For the La-Al2O3 support, penetration of Ru from the catalyst surface into the bulk La-rich Al2O3 center poses a problem. A recent study conducted by Li et al. aiming to overcome this issue reported a modified synthesis strategy in which the surface of powder Al2O3 was coated by La2O2CO3 during La impregnation [48]. The authors stated that the formation of an La2O2CO3 phase over the catalyst prevented Ni metal from entering the bulk Al2O3 by instead forming NiAl2O4. This led to a higher number of exposed Ni active sites for the dry reforming reaction. The formation of La2O2CO3 has been observed in several other studies [49\u201352], where this species significantly enhanced catalytic activity by strong metal-support interactions and reductions in coke formation during dry and steam reforming as well as CO2 methanation.In the current study, Ru supported on an La carbonate-rich Al2O3 catalyst (Ru/La2O2CO3-Al2O3) was synthesized in the form of catalyst beads, and its reactivity for ammonia decomposition was examined. The catalytic activity of Ru/La2O3-Al2O3 and Ru/Al2O3 beads was also investigated to elucidate whether the La oxycarbonate species (La2O2CO3) acts as a structural stabilizer to prevent the loss of Ru metal to inner the bulk region of Al2O3 catalyst beads. Several analytical techniques were utilized to understand the physical and chemical properties of the catalysts, including Brunauer\u2013Emmett\u2013Teller (BET) physisorption, X-ray diffraction (XRD), scanning transmission electron microscopy (STEM), and CO chemisorption. The electronic properties of the catalysts were analyzed by X-ray photoelectron spectroscopy (XPS), and H2-TPR (temperature-programmed reduction). Finally, the cross-sectional area of the Ru/La2O2CO3-Al2O3 beads was examined by scanning electron microscopy-energy dispersive spectrometry (SEM-EDS) to determine the distribution of Ru molecules between the surface and bulk phases of the catalyst beads.Aluminum oxide (Alfa Aesar, 1/8\u2033, beads), lanthanum (III) nitrate hydrate (Sigma Aldrich, 99.9%), and ruthenium (III) chloride hydrate (Sigma Aldrich, 99.98%) were purchased commercially. All chemicals and raw materials were used as received without further purification.The La-Al2O3 beads were prepared using the wet impregnation method. First, 3.5408\u00a0g of La precursor solid (10\u00a0mol%) was dissolved in 25\u00a0mL of deionized water. The solution was then added to dried aluminum oxide beads (10\u00a0g) and heated to 40\u00a0\u00b0C in a vacuum oven for 2\u00a0h under mild vacuum conditions (0.8\u00a0bar). The suspension was then rigorously stirred and evacuated under a 0.3\u20130.6\u00a0bar vacuum, where the temperature was heated to 100\u00a0\u00b0C at a ramp rate of 15\u00a0\u00b0C/h. The suspension was dried at 100\u00a0\u00b0C for an additional 3\u00a0h, and the resulting solid beads were calcined at 600\u00a0\u00b0C for 3\u00a0h in air to yield the La2O3-Al2O3 catalyst. La2O2CO3-Al2O3 was prepared using the same impregnation method, except that calcination was performed in a CO2 atmosphere as reported by Li et al. [48].Ru metal was loaded on both La2O3-Al2O3 and La2O2CO3-Al2O3 bead supports using the wet impregnation method. Ruthenium(III) chloride hydrate (0.4576\u00a0g, 2\u00a0wt%) was first dissolved in 25\u00a0mL of deionized water. The Ru solution was mixed with 10\u00a0g of each bead support. The drying procedures were identical to those described above for the loading of La. The resulting catalyst beads were dried at 100\u00a0\u00b0C for 12\u00a0h in air to obtain the Ru/La2O3-Al2O3 and Ru/La2O2CO3-Al2O3 catalyst beads.The Ru/Al2O3 catalyst beads were also prepared using the wet impregnation method. Prior to the loading of Ru, the Al2O3 beads were calcined under a CO2 atmosphere at 600\u00a0\u00b0C for 3\u00a0h. Ru was then impregnated on the Al2O3 beads following the identical synthesis procedure used for the Ru/La-based Al2O3 catalyst beads. All the ex-situ reduced Ru-based catalyst beads were treated at 500\u00a0\u00b0C for 1\u00a0h under a 75% H2/N2 flow (\nScheme 1).All catalyst characterization techniques were conducted with a sample of whole, unground beads. The only exceptions were XRD, in which a bead was ground until a homogeneous mixture was achieved, and STEM/TEM, in which the outer surface of the bead was scraped off and collected for the sample.The surface area, pore volume, and pore size distribution of the samples were determined via nitrogen physisorption at \u2212\u2009196\u2009\u00b0C (Micromeritics ASAP 2000 volumetric adsorption system). Prior to nitrogen adsorption, 100\u2009mg of each sample was degassed under vacuum at 250\u2009\u00b0C for 8\u2009h to remove moisture and impurities from the sample surface. The specific surface areas were calculated using the BET method. The Barrett\u2013Joyner\u2013Halenda (BJH) method was used to obtain the pore volume and pore size distribution from the desorption branch of the nitrogen isotherm.The amounts of Ru and La in the samples were measured using inductively coupled plasma-optical emission spectrometry (ICP-OES; Agilent 5100). For the elemental analysis, a digestion process was employed using a mixture of HNO3 and HCl with a microwave oven.Pulse CO chemisorption of the samples was conducted using a chemisorption analyzer (Micromeritics AutoChem II 2920). Approximately 0.2\u2009g of sample was placed in a sample tube. The sample was heated to 500\u2009\u00b0C at a ramping rate of 5\u2009\u00b0C/min and maintained for 1\u2009h under pure He (99.999%) to remove moisture and impurities from its surface. The sample was then reduced in 10% H2/Ar at 500\u2009\u00b0C for 1\u2009h. When the reduction was completed, the temperature was decreased to 50\u2009\u00b0C under degassing conditions. 10% CO/He gas was then injected in pulses into the sample tube every 5\u2009min (cycle of injection-evacuation) until the sample surface was saturated with CO molecules. Ru dispersion was calculated based on the amount of adsorbed CO molecules. The stoichiometric molar ratio of adsorption was assumed to be Ru/CO =\u20091\n\n(1)\n\n\nR\nu\n\n\nD\ni\ns\np\ne\nr\ns\ni\no\nn\n\n\n\n\n(\n\n%\n\n)\n\n\n=\n\n\nM\n\noles\n\n\n\nof\n\n\n\nadsorbed\n\n\n\nCO\n\n\n\nM\n\noles\n\n\n\nof\n\n\n\nRu\n\n\n\nin\n\n\n\nthe\n\n\n\ncatalyst\n\n\n(\n\nICP\n\n)\n\n\n\u00d7\n100\n\n\n\n\nThe powder XRD patterns of the samples were recorded using an X-ray diffractometer (Shimadzu, XRD-6100) with Cu K\u03b1 radiation (\u03bb\u2009=\u20091.5418\u2009\u00c5), with the accelerating voltage and current set to 40\u2009kV and 30\u2009mA, respectively. A continuous scan at a rate of 2\u03b8 =\u20092\u00b0/min was conducted in the range of 10\u201380\u00b0.TEM (Thermo Scientific\u2122 Talos F200X) was employed to visualize the Ru particles and estimate their size and morphology. The sample was first suspended in ethanol for 30\u2009min under ultrasonic treatment, and the solution was then deposited onto carbon film-coated copper grids. High-angle annular dark-field imaging (HAADF) in STEM mode and elemental mapping images measured by EDS were used to identify Ru, La, and Al dispersions on the sample surface. The particle size distribution of Ru was determined using ImageJ software to measure the size of 80 Ru particles from different locations across the sample surface.The XPS spectra of the samples were collected using an Thermo Scientific K-Alpha+ equipped with monochromatic Al K\u03b1 radiation (1486.6\u2009eV, 12\u2009kV and 1.16\u2009mA). The binding energy calibration of the XPS spectra was performed using C\u00a01s (284.8\u2009eV) prior to XPS fitting (Avantage software program). Spectral regions corresponding to the C\u00a01s, O\u00a01s, La\u00a03d, and Ru\u00a03p core levels were recorded for each sample.The cross-sectional distribution of the Ru concentration in the sample bead was examined using FE-SEM/EDS (FEI Inspect F). Prior to the analysis, a sample bead was split in half, and the cross sections were covered with Pt by electrically conductive coating (15\u2009mA, 60\u2009s) for optimal imaging and analysis.TPR experiments were performed using a BELFAT-M chemisorption analyzer (MicrotracBEL Corp.) as described previously [9]. First, the sample (100\u2009mg) was pretreated at 300\u2009\u00b0C for 1\u2009h under pure Ar (99.999%) gas to remove moisture and impurities inside the pores. The temperature was then decreased to 50\u2009\u00b0C and maintained until the thermal conductivity detector (TCD) signal was stabilized. After switching the gas to 10% H2/Ar, the temperature was increased to 500\u2009\u00b0C at a rate of 3\u2009\u00b0C/min. The outlet gas stream was continuously measured using the TCD. A water trap (zeolite 13X) was placed between the sample tube and the detector to remove the water formed during the measurement.The catalytic activities of Ru/La2O2CO3-Al2O3, Ru/La2O3-Al2O3, and Ru/Al2O3 catalyst beads for ammonia decomposition were evaluated in a fixed-bed quartz reactor under a flow of pure NH3 at atmospheric pressure. A quartz frit was placed in the middle of the reactor to hold the catalyst beads. The reaction temperature was measured using a thermocouple located near the catalyst bed. The gas flow rates of NH3, H2 and N2 were controlled using a mass-flow controller. Prior to the reaction, the catalyst beads (0.180\u2009g) were placed in the reactor and reduced in situ in a 75% H2/N2 flow at 500\u2009\u00b0C for 1\u2009h. The reaction was conducted at a GHSV of 10,000\u2009mLNH3/gcat\u22c5h in the range of 350\u2013500\u2009\u00b0C where the temperature was decreased at intervals of 50\u2009\u00b0C (maintained for 1\u2009h at each temperature). The effluent gas was analyzed using an online gas chromatograph (GC, Agilent 7890\u2009A) equipped with two thermal conductivity detectors (TCDs) and two columns. The front TCD (carrier gas: He) was connected to a CP-Volamine column to identify NH3 and N2. The second carrier gas (Ar) was linked to a 19091\u2009P-MS4 J&W PoraPLOT Amines column for H2 detection. The ammonia conversion was calculated using (Eq. (2)):\n\n(2)\n\n\n\nConversion\n\nNH\n3\n\n\n\n%\n\n=\n\n\n\n\nC\n\nN\n\nH\n3\n\n,\ni\nn\n\n\n\u2212\n\nC\n\nN\n\nH\n3\n\n,\no\nu\nt\n\n\n\n\n\nC\n\nN\n\nH\n3\n\n,\ni\nn\n\n\n\n\u00d7\n100\n\n\n\nwhere \n\n\nC\n\n\n\n\nNH\n\n\n3\n\n\n,\nin\n\n\n and \n\n\nC\n\n\n\n\nNH\n\n\n3\n\n\n,\nout\n\n\n refer to the concentrations of NH3 in the feed and product gas, respectively. The ammonia conversion at each temperature was obtained by averaging the conversion values over the corresponding hour of constant reaction temperature. The increase in the total gas volume as a result of the increase in the total number of moles of the reaction (2NH3 \u2192\u00a0N2 +\u00a03H2) was reflected in the final ammonia conversion calculation. No by-products were detected other than N2, H2 and unreacted NH3.N2 adsorption isotherms at \u2212\u2009196\u2009\u00b0C were obtained to investigate the structural properties of the as-prepared La2O2CO3-Al2O3, La2O3-Al2O3, and Al2O3 and ex-situ reduced Ru/La2O2CO3-Al2O3, Ru/La2O3-Al2O3, and Ru/Al2O3 catalyst beads. Physisorption was conducted using whole, unground catalyst beads. The isotherms and the corresponding BJH pore size distribution curves are shown in \nFig. 1\n. The calculated BET surface areas and pore volumes are summarized in \nTable 1. All isotherms showed type IV with H1 hysteresis, indicating mesoporous and open-pore structures [53\u201355]. The BJH pore size distribution curve reaffirmed the range of pore widths for the mesopores. When La was added to the Al2O3 catalyst beads, a narrower range of pore size distribution was obtained. Additionally, the BET surface area for the starting material, a commercial Al2O3 catalyst bead, was SBET =\u2009227.1\u2009m2/g and the surface area, total pore volume, and average pore diameter decreased slightly after impregnation with La species. This is because the surface of Al2O3 was covered by La species with a low specific surface area (La2O3 SBET = 21.8\u2009m2/g [56], La2O2CO3 SBET = 47\u2009m2/g [57]) [41]. However, all the isotherms of the La2O2CO3-Al2O3, La2O3-Al2O3, and Al2O3 catalyst bead supports were similar, indicating that the structural properties of the support were maintained after the loading of La species on Al2O3. After the impregnation of Ru particles, the surface area increased slightly, possibly attributable to the metallic Ru dispersed on the outer surface of the catalyst beads [54]. The estimated total pore volume was very similar for all the Ru-based catalysts implying that the Ru particles were uniformly dispersed on the alumina support and that no significant sintering occurred, thus maintaining the porosity of the catalyst beads.CO chemisorption measurements were taken to determine the degree of Ru dispersion over the catalyst beads (Table 1). A higher metal dispersion in the catalyst generally indicates a greater number of active metals that can participate in the reaction; thus, it is often employed as an important factor in evaluating the performance of the catalyst. From the CO chemisorption results, the initial dispersion values were obtained as 33.3%, 23.3% and 30.6% for Ru/Al2O3, Ru/La2O3-Al2O3 and Ru/La2O2CO3-Al2O3, respectively. These dispersion values were corrected based on the actual Ru weight loading from the ICP results later on. The highest Ru dispersion was obtained for the Ru/La2O2CO3-Al2O3 catalyst beads (32.2%). It should be noted that a similar Ru weight loading was achieved for all Ru-based catalyst beads (approximately 1\u2009wt%), with the error between measurements sufficiently small to yield a fair comparison of catalytic activity. Furthermore, the La weight loading of both La-loaded samples was approximately 12\u2009wt%.XRD patterns were obtained for the ex-situ reduced Ru-based catalyst beads (in powder form) to acquire structural information (\nFig. 2). All XRD patterns presented a wide diffraction peak centered at 2\u03b8 =\u200967.5\u00b0, which was assigned to amorphous \u03b3-alumina (JCPDS 10\u20130425). For the Ru/La2O3-Al2O3 catalyst, characteristic diffractions of La2O3 (JCPDS 05\u20130602) were not observed. The peaks related to the formation of La2O3 had very low intensities, possibly because of the strong interaction between La2O3 and Al2O3\n[48,58]. This resulted in a high dispersion of La species over the catalyst surface which was later confirmed by HAADF-STEM. In contrast, for Ru/La2O2CO3-Al2O3, the peaks centered at 13.1\u00b0, 25.3\u00b0, 30.9\u00b0 and 34.0\u00b0 were assigned to La2O2CO3 with a monoclinic structure (JCPDS 48\u20131113). Unlike the Ru/La2O3-Al2O3 catalyst, La2O2CO3 showed high crystallinity and formed through calcination under CO2 conditions. Notably, the La2O2CO3 phase was stable throughout the La and Ru impregnation processes that involved high temperature treatment at 600\u2009\u00b0C. Metallic Ru was observed in both the Ru/La2O3-Al2O3 and Ru/La2O2CO3-Al2O3 catalysts. However, no peaks were identified for Ru/Al2O3 likely attributable to the low average Ru particle size leading to the absence of long-range order of Ru.STEM-EDS images of the ex-situ reduced Ru-based catalyst beads were collected to estimate the dispersity of the metal elements (Ru, Al, and La) over the catalyst surface. For sample preparation, the outer surface of the bead was scraped off and collected. The HAADF-STEM and elemental mapping images in \nFig. 3 clearly indicate highly dispersed Ru and La particles over all the catalysts. Furthermore, TEM images were taken over all three catalysts (\nFig. 4, Ru/Al2O3, Ru/La2O3-Al2O3 and Ru/La2O2CO3-Al2O3). Similar to the STEM-EDS images, Ru particles indicated as black dots were highly dispersed on the amorphous Al2O3 or mixed La-Al2O3 support. The Ru particle size distribution obtained from the TEM images show narrow distribution curves between the 0.4 and 4.0\u2009nm for all the catalysts. Among the catalysts, Ru/Al2O3 possessed the smallest average particle size (0.96\u2009nm) which was in good agreement with the XRD results. In contrast, for the La-containing catalysts (Ru/La2O3-Al2O3 and Ru/La2O2CO3-Al2O3), a larger average Ru particle size was observed compared with that of Ru particles impregnated solely on the Al2O3 support. These results may be attributable to the presence of large La particles on the catalyst surface, resulting in a lower specific surface area and smaller average pore diameter, and thus, a greater extent of Ru agglomeration.XPS spectra of the as-prepared La2O3-Al2O3 and La2O2CO3-Al2O3 bead supports were obtained to verify the formation of La2O2CO2 over the Al2O3 surface. The X-ray beam was focused on the whole catalyst bead (without grinding it). The full-range spectrum confirmed the chemical purity of the La2O2CO3-Al2O3 sample, consisting of Al, La, C, and O (\nFig. 5\n(a)). As shown in Fig. 5\n(b), the La2O2CO3-Al2O3 support primarily contained peaks for the 3d orbits of the La oxide species. The electronic states of carbon and oxygen in both samples were clearly distinguishable because of the differences in their carbon-containing structures. In the C\u00a01s spectra (Fig. 5\n(c-d)), a C-C peak was detected for both samples because of adventitious carbon contamination during the exposure to the atmosphere. However, the peak centered at 289.3\u2009eV only existed in the La2O2CO3-Al2O3 bead support and was assigned to the O-CO bond induced by (CO3)2- resonance [59]. In addition, the difference in the electron state of oxygen between both samples was more apparent than that of the carbon electron state. As shown in Fig. 5\n(e-f), the appearance of two peaks was confirmed: one at 531.8\u2009eV and the other centered at 528.8\u2009eV. The peak at 528.8\u2009eV was concluded to be related to lattice oxygen in La2O3 and Al2O3 and the other was assigned to the adsorbed oxygen-like surface (CO3)2-. It should be noted that the predominant peak at 531.8\u2009eV was only detected for the La2O2CO3-Al2O3 support [60] indicating that the La2O2CO3 structure formed well on the Al2O3 support during CO2 calcination.Furthermore, the XPS spectra of the Ru 3p region clearly indicated that the Ru surface concentration of the Ru/La2O2CO3-Al2O3 catalyst beads was higher than that of the Ru/La2O3-Al2O3 beads, evidenced by the larger Ru peak intensity (3.1 times higher; \nFig. 6). This seems to suggest that the La2O2CO3 interface can prevent Ru from entering the support, thereby making Ru metals more likely to be located on the surface of the bead support.A transverse section of each catalyst bead sample was prepared to validate the actual surface distribution of Ru metals over the ex-situ reduced Ru/La2O3-Al2O3, Ru/La2O2CO3-Al2O3 and Ru/Al2O3. The Ru concentration in the cross-sectional interior of the catalyst was observed by SEM-EDS line scanning. As shown in \nFig. 7\n, the highest Ru concentration was achieved on the bead surface for all samples. The concentration of Ru decreased towards the inside of the bead core, but the distribution trends differed for the three types of catalyst beads. To accurately compare the ratio of surface/bulk Ru concentration, the \u2018surface\u2019 was defined from the edge up to 100\u2009\u00b5m in depth, and the \u2018bulk\u2019 was limited to the region with a diameter of 500\u2009\u00b5m measured from the catalyst center. The average Ru concentration over each region was calculated to obtain the surface/bulk Ru concentration ratio. The results were 1.8, 2.4 and 3.8 for Ru/Al2O3, Ru/La2O3-Al2O3 and Ru/La2O2CO3-Al2O3, respectively. For Ru/Al2O3, the average Ru concentration in the surface region was the lowest (5.6\u2009wt%) compared to the other catalyst beads, while the average concentration in the bulk region was the highest (3.2\u2009wt%). In contrast, the average concentrations of Ru in the surface regions of Ru/La2O3-Al2O3 and Ru/La2O2CO3-Al2O3 were 7.2\u2009wt% and 8.4\u2009wt%, respectively.These results serve as compelling evidence for the pivotal role that the La carbonate-rich Al2O3 oxycarbonate species plays in preventing the penetration of Ru particles into the catalyst bead core and thus minimizing the loss of Ru metals to the bulk phase. It is speculated that there could be a steric hindrance effect in the macropore diffusion of Ru chloride species during catalyst synthesis, where the higher crystallinity and bulkier matrix of the La2O2CO3-Al2O3 structure aid the formation of Ru particles on the surface of the catalyst bead.H2-TPR experiments were performed to understand the reduction behavior of the non-reduced Ru-based catalyst beads. These types of experiments are useful to estimate the extent of the metal-support interaction by considering shifts in the reduction temperature [9]. As shown in \nFig. 8\n, the reduction of Ru/Al2O3 occurred as a single broad peak at 118\u2009\u00b0C assigned to the reduction of RuOx to metallic Ru [41,61] indicating that the support was composed of a single component of alumina. The Ru/La2O3-Al2O3 catalyst exhibited two major reduction peaks at 131 and 209\u2009\u00b0C. The reduction of Ru at temperatures higher than that of Ru/Al2O3 was ascribed to strong metal-support interactions between Ru particles and the La-Al species [41] meaning that the Ru particles located in the bulk phase of the catalyst beads required a higher temperature for complete reduction to metallic metal.The characteristic reduction peaks of Ru/La2O2CO3-Al2O3 were observed at 109 and 135\u2009\u00b0C. The reduction peak for Ru shifted to a lower temperature than that of the Ru/La2O3-Al2O3 catalyst, indicating a higher concentration of Ru on the bead surface weakly interacting with the support [46,49,62]. Similar TPR results were observed by Li and coworkers [48] where higher extent of Ni reduction in the lower temperature region was obtained over the La2O2CO3-Al2O3 in comparison to the Al2O3 indicating higher concentration of the surface Ni active sites for dry reforming of methane. Ultimately, this catalyst characterization supported previous analysis in showing that the incorporation of the La oxycarbonate species leads to a higher number of Ru active sites on the surface of Ru-based catalyst beads.The Ru-based catalyst beads prepared using three different synthesis methods were investigated with respect to catalytic activity for low temperature (350\u2212500\u2009\u00b0C) ammonia decomposition at a GHSV of 10,000\u2009mLNH3/gcat\u22c5h. As shown in \nFig. 9, ammonia conversion, an exothermic reaction, increased with an increase in reaction temperature for all catalyst beads. The catalytic performance ranked from best to worst was: Ru/La2O2CO3-Al2O3 >\u2009Ru/La2O3-Al2O3 >\u2009Ru/Al2O3 with a margin of error of \u00b1\u20093%. It should be noted that although a higher Ru dispersion was observed in Ru/Al2O3 compared with Ru/La2O3-Al2O3 beads, Ru/Al2O3 displayed lower catalytic activity. This suggests a promotion effect of La-addition to an Ru/Al2O3 catalyst for ammonia decomposition. Electron donation from La to Ru due to electronegativity differences increases the electron density of Ru, significantly increasing the kinetics of the ammonia decomposition reaction [41,63], specifically speeding up the cleavage of the N-H bond and the recombinative desorption of N2\n[63].As expected, the highest ammonia conversion was obtained for the Ru/La2O2CO3-Al2O3 catalyst beads over the full range of reaction temperatures tested. The high Ru surface concentration and Ru dispersion of La-containing catalysts resulted in a higher ammonia decomposition activity and faster reaction rate. Overall, it is likely that the formation of Ru particles was more favorable on the surface of the catalyst beads when a La oxycarbonate-rich surface (La2O2CO3) was present in comparison to the La oxide surface (La2O3). In other words, the higher structural density of La2O2CO3 prevented the Ru salts from traveling from the bead surface into the bulk region. We believe that an La2O2CO3 surface-coating can be utilized for catalyst synthesis, particularly for beads, pellets, monoliths, and other structured catalysts where a high surface concentration of metal is essential.In this study, we investigated the effect of La oxycarbonate (La2O2CO3) on the catalytic activity of Ru/La-based Al2O3 beads for ammonia decomposition. Three Ru-based catalyst beads were prepared: Ru/Al2O3 with a boundary layer of an Ru-Al mixed oxide phase and Ru/La2O2CO3-Al2O3 and Ru/La2O3-Al2O3 with a Ru-La-Al ternary layer including La2O2CO3 and La2O3, respectively. The Ru/La2O2CO3-Al2O3 catalyst beads showed the highest activity for ammonia decomposition (80.1% conversion) compared with Ru/La2O3-Al2O3 (72.1%) and Ru/Al2O3 (51.4%) at a reaction temperature of 500\u2009\u00b0C. According to the catalyst characterization results, the highest Ru concentration and dispersion was observed on the surface of La2O2CO3-Al2O3, which was attributed to the boundary layer of La2O2CO3 preventing the loss of Ru metals to the bulk phase of the Al2O3 catalyst beads. In other words, the formation of metallic Ru particles on the surface of the La2O2CO3-Al2O3 beads was more favorable than that on La2O3-Al2O3 because of the density of the macroscale structure of La2O2CO3-Al2O3 (acting as a structural stabilizer) and appropriate metal-support interactions, limiting the penetration of Ru particles into the catalyst bead core. The addition of an La layer between the Ru and Al2O3 beads promoted the catalytic activity for ammonia decomposition, believed to be the result of an increase in the Ru electron density, thereby leading to the faster recombinative desorption of N2.\nAh-Reum Kim: Writing - original draft, Investigation. Junyoung Cha: Writing - original draft, Investigation. Jin Su Kim: Validation. Chang-Il Ahn: Formal analysis. Yongmin Kim: Methodology. Hyangsoo Jeong: Data curation. Sun Hee Choi: Resources. Suk Woo Nam: Conceptualization Chang Won Yoon: Project administration. Hyuntae Sohn: Writing - review & editing, Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Research Foundation of Korea (NRF), funded by the Government of Republic of Korea (Ministry of Science and ICT, Hydrogen Energy Innovation Technology Development Program [No. 2019M3E6A1064611, No. 2019M3E6A1104113]), and the Korea Institute of Science and Technology (KIST) Institutional Program (No. 2E31872).", "descript": "\n                  Three ruthenium-supported catalyst beads (Ru/Al2O3, Ru/La2O3-Al2O3 and Ru/La2O2CO3-Al2O3) were synthesized and tested for ammonia decomposition. The catalytic activity of the Ru/La2O2CO3-Al2O3 beads was significantly higher than that of the Ru/Al2O3 and Ru/La2O3-Al2O3 beads. This was primarily attributed to the addition of La, which encouraged electron donation from the bead surface to the Ru particles, increasing the rate of N2 desorption. In particular, a higher Ru surface concentration was achieved over the boundary layer of La2O2CO3 compared with La2O3. This is thought to be a result of steric hindrance, with the crystalline surface of La2O2CO3 acting as a structural stabilizer to significantly limit the penetration of Ru particles into the catalyst bead core. SEM-EDS line scanning of transverse sections of the catalyst beads confirmed a higher Ru concentration on the surface of the catalyst beads for Ru/La2O2CO3-Al2O3 compared with Ru/La2O3-Al2O3 and Ru/Al2O3. In fact, the ratio of surface/bulk Ru concentration in Ru/La2O2CO3-Al2O3 was more than twice that of Ru/Al2O3 at equal Ru loadings. The favorable properties of an La2O2CO3 surface-coating can benefit industrial catalyst synthesis, increasing the surface metal concentration compared with traditional La-based Al2O3 beads and pellets.\n               "}
{"full_text": "Sustainable and economically-efficient ways to produce and store energy are crucial for the transition to a society independent of fossil fuels. Hydrogen as energy carrier is expected to play an important role in this transformation, finding applications in powering vehicles, electronic devices or homes [1,2]. Moreover, the combustion of H2 is a clean process releasing only water unlike the currently used fossil fuels. A sustainable, environmentally friendly and scalable production of hydrogen could be achieved via alcohol electrolysis where, in addition to hydrogen gas, valuable chemicals could be obtained in the anodic reaction [3]. Examples of alcohols that attract attention are methanol, ethanol and particularly glycerol, once these can be obtained from different biomass sources and industrial processes, such as wood-based Kraft pulping [4\u20138]. This strategy enables the valorization of the pulp waste product or glycerol from biodiesel production through transformation into economically valuable products in addition to storing energy in hydrogen gas.The methanol oxidation reaction (MOR) serves as a prototype to understand the interactions of alcohols and metal electrocatalysts. In general, formaldehyde, formic acid and CO2 are the main products obtained during the methanol oxidation reaction (Scheme\u00a01). A greater H2 production rate (in electrolysis) might be obtained for the full oxidation of methanol to CO2. This is due to the larger number of electrons involved for the full reaction (6 e\u2212) as compared to the other products (2 e\u2212 and 4 e\u2212, respectively). Platinum has been identified as a remarkable catalyst for methanol oxidation, but its high cost is an obstacle to large-scale usage. Furthermore, platinum shows some level of CO poisoning during the methanol oxidation reaction, interrupting the electrocatalytic activity [9]. This has motivated investigations to seek alternative catalysts with low production cost to improve the viability for large-scale applications. In this context, a wide range of possible single-metal catalysts, like Cu, Ni and Pd [10\u201313] or bimetallic catalysts, Pt-Ru, Pt-Ni, Pt-Pd, Pt-Cu and Pt-Sn, have been studied for methanol oxidation. [14\u201318]. Interestingly, monometallic Pd has shown to be less active for methanol oxidation than Pt catalysts [19]. However, bimetallic Pd catalysts seem to enhance the catalytic performance towards this reaction, and thus, they may become competitive with state-of-the-art Pt catalysts [20\u201327]. Bimetallic alloys may benefit from the bifunctional effect \u2013 possible oxidative removal of CO from the catalytic surface, [28] which would act positively for the methanol oxidation reaction [29\u201331]. Among these, Pd-Ni alloys provide acceptable catalytic performance for the oxidation of different alcohol-based compounds, such as methanol [20\u201322], ethanol [32,33], glycerol [22] and lactic acid [8].Rationalizing the effects of alloying Pd-Ni for the oxidation of alcohols is crucial for the development of novel catalysts. Among the published works on alcohol oxidation on Pd-Ni, Carvalho et al.\n[22] reported an increased catalytic activity for methanol oxidation as a function of the Ni content in the employed catalyst while Qiu et al.\n[34] showed that the composition, structure and surface morphology of PdNi play important roles during the oxidative reaction. The strategy of having PdNi supported on TiO2 nanotubes for methanol oxidation in a direct methanol fuel cell has been tested which resulted in higher electroactivity than the state-of-the-art PtRu/C catalyst. [35] Miao et al.\n[36] employed the framework of density functional theory (DFT) to understand the effects of the Ni position (if beneath Pd or exposed on the surface) for ethanol oxidation. When Ni is exposed on the catalyst surface, lower activation barriers were reported for C-C cleavage and higher barriers for C-O bond breaking, hence enhancing the performance of PdNi for the electrooxidation of ethanol. Other experimental and theoretical studies have also reported favorable effects of Pd-Ni over pure Pd for ethanol oxidation where the central hypothesis explored uses the oxophilic property of Ni that provides a greater number of OHads species on the catalyst surface together with an improved tolerance against CO poisoning [37\u201339].Several experimental and theoretical works have reported on Pd-Ni alloys providing important accumulated knowledge for application to catalytic electrooxidation of alcohols. Yet, a systematic investigation unveiling the effects of Pd-Ni alloying and linking that to the CO poisoning phenomenon and also the methanol oxidation mechanism itself has not been reported, to the best of our knowledge. Thus, to shed light on the effects on the methanol electro-oxidation when using Pd-Ni alloys as catalyst, experimental measurements and theoretical calculations were performed to obtain atomistic level insight into: i) the nature of the Pd-Ni surface morphology and coverage; ii) the effect of Ni concentration on the performance of the catalyst; iii) the effects of CO poisoning; iv) the effects of OH\u2212 concentration in the electrolyte (pH change) on the methanol oxidation activity and, finally, v) linking these to the greater observed activity delivered by the Pd-Ni alloy.More specifically, the present study combines experiments and DFT calculations to clarify the methanol electrooxidation activity using Pd, Pd3Ni and PdNi catalysts. Energy-dispersive X-ray spectroscopy (EDS) and X-ray diffraction (XRD) were employed to characterize the morphology of the catalysts. Cyclic voltammograms (CVs) were performed to elucidate the electrochemical performance of each studied catalyst towards methanol electrooxidation and also to clarify the CO poisoning phenomenon and effect of pH on the reaction. In conjunction with the CVs, Pourbaix diagrams were calculated to predict the catalytic surface coverage. Moreover, the methanol oxidation reaction pathways were explored using DFT and the climbing image nudged elastic band method (CI-NEB) for barriers in combination with measurements of product distributions using High-Performance Liquid Chromatography (HPLC) to elucidate the most likely oxidative paths on the studied catalysts. Based on these experiments and calculations, we propose an explanation for the higher activity toward methanol electrooxidation displayed for the alloys compared to the pure Pd catalyst.Density functional theory has been employed within the projected augmented wave method as implemented in the Vienna Ab Initio Simulation Package (VASP) [40,41]. The optB86b-vdW functional was selected to describe the exchange and correlation term of the Kohn-Sham equation as well as an estimate of the van der Waals contribution [42\u201344]. This functional has been shown to, e.g., properly describe the adsorption of methanol on gold [45]. A cutoff energy of 450\u00a0eV was used for the plane-wave expansion with a grid for the sampling of the Brillouin zone that depended on the supercell size. For bulk calculations, a Monkhorst\u2212Pack grid of 15\u00a0\u00d7\u00a015\u00a0\u00d7\u00a015 was employed, while for the slabs, a grid of 2\u00a0\u00d7\u00a02\u00a0\u00d7\u00a01 was employed due to the large supercell size. For the gas phase molecule, the \u0393 point was used. The spin-polarized approach was used with the electron partial occupancies obtained within the Methfessel-Paxton scheme of order 2 and a smearing parameter of 0.2\u00a0eV. Vibrational modes were computed through the finite difference approximation and the self-consistent field energies were corrected for zero-point energies (ZPE). Moreover, the rotational, translational and vibrational contributions to the entropy and enthalpy were considered for gas-phase species where we furthermore set \n\nPV\n=\n\nk\nB\n\nT\n\n (see eq.\u00a06), where P and V are pressure and volume, respectively, while T and kB are temperature and the Boltzmann constant, respectively.Pd was modeled in a fcc unit cell while the Pd3Ni and PdNi alloy-like structures were constructed by replacing Pd atoms by Ni atoms in the Pd fcc lattice. Starting from pre-optimized bulk structures the surface models were built as follows: The (111) facet of the bulk structures investigated were modeled as a four-layer slab with a p(4\u00a0\u00d7\u00a04) supercell and 20\u00a0\u00c5 vacuum to avoid interactions between the periodic images (Fig.\u00a01\n). We have allowed atomic relaxations of the top two layers and fixed the bottom two layers of the slabs. Gas-phase molecules were modelled in a 20\u00a0\u00c5 cubic cell.The computational hydrogen electrode approach, as proposed by N\u00f8rskov [46], was applied to model the electrochemical reactions. This approach assumes a coupled electron-proton transfer simplifying the demanding calculation of solvation energies of ionic species. The formation energy of the intermediates (Ead) was calculated as:\n\n(1)\n\n\n\nE\n\na\nd\n\n\n=\n\nE\n\na\nd\ns\no\nr\nb\na\nt\ne\n\n*\n\n\u2212\n\nE\n*\n\n\u2212\n\n\u2211\ni\n\n\nn\ni\n\n\n\u03bc\ni\n\n\n\n\nwhere \n\nE\n\na\nd\ns\no\nr\nb\na\nt\ne\n\n*\n\n is the self-consistent-field (SCF) energy of the adsorbed intermediate corrected by the zero-point energy (ZPE) of the adsorbate, E* is the SCF energy of the pure slab and ni\n is the number of species i with chemical potential \u03bci\n. Moreover, \u03bcH\n, \n\n\u03bc\n\n\nH\n2\n\nO\n\n\n, \u03bcO\n and \u03bcC\n are the chemical potentials of hydrogen, water, oxygen and carbon, respectively, that are obtained as:\n\n(2)\n\n\n\n\u03bc\nH\n\n=\n\n1\n2\n\n\nE\n\nH\n2\n\n\n+\ne\n\nU\n\nS\nH\nE\n\n\n\u2212\n\nk\nB\n\nT\nl\nn\n\n(\n10\n)\n\n\u00d7\np\nH\n=\n\n\n1\n2\n\n\nE\n\nH\n2\n\n\n+\ne\n\nU\n\nR\nH\nE\n\n\n\n\n\n\n\n\n(3)\n\n\n\n\u03bc\n\n\nH\n2\n\nO\n\n\n=\n\nE\n\n\nH\n2\n\nO\n\n\n\n\n\n\n\n\n(4)\n\n\n\n\u03bc\nO\n\n=\n\n\u03bc\n\n\nH\n2\n\nO\n\n\n\u2212\n2\n\n\u03bc\nH\n\n\n\n\n\n\n\n(5)\n\n\n\n\u03bc\nC\n\n=\n\nE\n\nC\n\nH\n3\n\nO\nH\n\n\n\u2212\n\n\u03bc\nO\n\n\u2212\n4\n\n\u03bc\nH\n\n\n\n\n\n\n\n(6)\n\n\n\n\n\n\nE\n\n\nH\n2\n\n,\n\n\nH\n2\n\nO\n,\nC\n\nH\n3\n\nO\nH\n\n\n=\n\nE\n\ns\nc\nf\n\n\n+\nZ\nP\nE\n+\n\n(\n\n\nH\n\nv\ni\nb\n\n\n+\n\nH\n\nt\nr\na\nn\ns\n\n\n+\n\nH\n\nr\no\nt\n\n\n\n)\n\n\n\n\n\n\n\n\u2212\nT\n\n(\n\n\nS\n\nv\ni\nb\n\n\n+\n\nS\n\nt\nr\na\nn\ns\n\n\n+\n\nS\n\nr\no\nt\n\n\n\n)\n\n+\nP\nV\n\n\n\n\n\n\n\nHere, pH dependence is incorporated in the potential of the proton electron transfer where URHE is the potential measured against the reversible hydrogen electrode (RHE) and USHE is the potential versus the standard hydrogen electrode (SHE). \n\n\nE\n\nC\n\nH\n3\n\nO\nH\n\n\n,\n\n\n\n\nE\n\n\nH\n2\n\nO\n\n\n, \n\nE\n\nC\n\nH\n3\n\nO\nH\n\n\n are the gas-phase Gibbs free energies computed as shown in eq.\u00a06. Escf\n is the SCF energy, H refers to the enthalpic thermal contribution and S to the entropic thermal contributions.We define coverage as the ratio between the number of adsorbates and the number of surface atoms.Activation barriers (Ea) for the deprotonation reaction were computed using the climbing image Nudged Elastic Band method (CI-NEB) where the number of intermediate images varies from four to seven depending on the studied reaction [47]. Forces on the atoms were minimized to 0.05\u00a0eV/\u00c5. Minimum energy paths (MEP) of the deprotonation reaction were modeled by adding four water molecules (explicit solvation model). The structure of the four-water cluster was obtained by checking different configurations on the surfaces. The lowest-energy structure is then used for the oxidation reaction where intermediates are added to the surface plus water structure. This model ensures that the formation of O-H polar bonds stabilizes the electron transfer process and possibly lowers activation barriers through water-assisted H transfer. In contrast, in the case of C - H breaking the presence of water tends to produce higher activation barriers [48]. The energies of the transition states have been corrected to mimic the effect of the electrode potential on the obtained activation barriers [49]. Here, we assumed an adiabatic charge transfer along the MEP and the transition state energies were corrected accordingly to: \n\n\nE\na\n\n\n(\nU\n)\n\n=\n\nE\na\n\n\u2212\n\u03bb\ne\nU\n,\n\n where \u03bb is the reaction symmetry factor - we approximate \u03bb to 0.5 [50], U is the electrode potential, e is the (positive) elementary charge and Ea\n is the activation barrier. The reaction barrier of the OH coupling assume that an OH is removed from the catalytic surface and coupled to the intermediate resulting in no effect of the electrode potential on the activation barrier.Palladium (II) chloride (> 59.0% Pd; >99.9%, metal basis), nickel chloride hexahydrate, sodium citrate, sodium hydroxide, sulfuric acid (HPLC grade) were purchased from VWR. Sodium borohydride and Nafion solution (5%\u00a0wt) were purchased from Sigma-Aldrich. Isopropanol, methanol, formaldehyde and formic acid were obtained from Merck. Super P conductive carbon was obtained from Timcal Graphite & Carbon. Ultrapure water obtained with a Millipore DirectQ3 purification system from Millipore was used throughout this work.Catalysts (Pd, Pd3Ni and PdNi) were prepared following a method previously reported [51]. Briefly, defined amounts of PdCl2 and NiCl2\u20226H2O to have a specific Pd:Ni molar ratio (1:0, 3:1, 1:1) were dissolved in 30\u00a0mL ultrapure water with a total metal content of 0.2\u00a0mmol. Then, sodium citrate (0.35\u00a0mmol) as stabilizer was dissolved into the same solution. N2 was bubbled to remove dissolved O2 from the solution (for 15\u00a0min) and was left continuously running to prevent new O2 entering the solution during the reaction. Then, 20\u00a0mL of 0.1\u00a0M NaBH4 (2\u00a0mmol) was added slowly (dropwise) to the mixture, and the solution was kept stirring for 1\u00a0h. The solid product was decanted, washed with ultrapure H2O and ethanol several times, and dried at 60\u00b0C for 30\u00a0min. The catalyst ink was prepared by dispersing 2\u00a0mg of the catalyst product in 50\u00a0\u00b5L Nafion (5%), 1500\u00a0\u00b5L isopropanol, and 450\u00a0\u00b5L ultrapure water in an ultrasonic bath (2\u00a0mg/mL as catalyst concentration). When the catalyst was well dispersed in this solution, 3\u00a0mg of carbon black was added, and the solution was kept in the ultrasonic bath for 1\u00a0h. A glassy carbon electrode (GCE) was modified by adding 10\u00a0\u00b5L of the catalyst ink twice and left to dry before rinsing it with ultrapure water.Electrochemical measurements were performed using a PAR273A potentiostat/galvanostat from Ametek in a 100\u00a0mL glass three-electrode cell (15\u00a0mL when product analysis was performed). A Pt mesh was used as counter electrode and a Hg/HgO electrode as reference (RE-A6P, Bio-Logic, 1 M NaOH). A glassy carbon electrode (GCE) of 0.6\u00a0cm diameter (~0.28\u00a0cm2 geometric area) was used as working electrode after addition of the catalyst ink. Current densities are presented normalized by the electrochemical surface area (ECSA). All measurements were performed at controlled temperature (25\u00a0\u00b1\u00a01\u00b0C) using a water bath and in a N2-saturated solution. Potentials were converted to the reversible hydrogen electrode (RHE) scale using the following equation:\n\n(7)\nEvs. RHE (mV)\u00a0=\u00a0Evs. Hg/HgO\u00a0+\u00a00.059 pH\u00a0+\u00a00.140\n\nwhere pH was considered to be 13, 13.7 and 14 for 0.1, 0.5 and 1\u00a0M NaOH solutions, respectively. Electrodes were cycled in 1\u00a0M NaOH between 0.05\u00a0V and 0.95\u00a0V (vs RHE) at a scan rate of 10\u00a0mV/s in order to achieve a stable electrochemical response (5 cycles). Overpotentials are given with respect to the standard potential for methanol full oxidation to CO2 (Scheme\u00a01), which is +0.02\u00a0V (vs RHE at pH 14) [52]. For the CO stripping experiments, CO pre-adsorption was carried out by bubbling CO gas into a 1\u00a0M NaOH solution with the working electrode at +0.1\u00a0V (vs RHE) for 20\u00a0min. Then, the solution was purged with argon for another 20\u00a0min to remove dissolved CO gas before the stripping experiment was carried out.The catalysts were characterized by Energy-dispersive X-ray spectroscopy (EDS) and X-ray diffraction (XRD) in order to corroborate their elemental composition (Pd:Ni atomic ratio) and the crystalline structure, which agreed well with the data previously reported following the same preparation method [51]. EDS was carried out using the integrated detector of a JEOL JSM-7000F instrument at an acceleration voltage of 15\u00a0kV. Fig. S1 shows the EDS spectra obtained for the different catalysts (Pd, Pd3Ni, PdNi) and Table S1 summarizes the calculated values for the elemental compositional ratio. XRD patterns were recorded with a PANalytical PRO MPD diffractometer in Bragg-Brentano geometry with 1.5406\u00a0\u00c5 Cu K\u03b11 radiation, using a 2\u03b8 range of 14\u00b0\u201390\u00b0 and a step size of 0.016\u00b0. The XRD patterns (Fig. S2) show that all the catalysts present the fcc Pd crystalline structure, similar to other PdNi catalysts reported previously [51,53,54] and the shifting of the 2\u03b8 peaks with increasing amount of Ni suggests successful alloy formation between the two metals [53].Identification and quantification of the oxidation products generated during the methanol oxidation reaction (MOR) for the different catalysts were carried out by High-Performance Liquid Chromatography (HPLC) on an Agilent 1260 Infinity II system with an Agilent Hi-Plex H column (250\u00a0\u00d7\u00a04.6\u00a0mm) and a refractive index detector (Agilent 1290 Infinity II RID) set on positive polarity. A sample volume of 10\u00a0\u00b5L was injected into the column using the autosampler. Eluent was 5\u00a0mM HPLC-grade H2SO4 at a flow rate of 0.4\u00a0mL min\u22121. Column and detector temperatures were 30\u00b0C. The counter electrode was separated from the anodic compartment by a glass frit to avoid side reactions of the generated products. To calculate the product selectivity, the number of moles of generated products (formaldehyde and formic acid) and the number of moles of methanol consumed were obtained by HPLC after calibration using standard solutions. Since the HPLC method could only be able to detect formaldehyde and formic acid, but not CO3\n2\u2212, the amount of CO2 formed during the MOR was estimated by difference assuming that the number of moles of methanol consumed is the same as the number of moles produced for all the three products:\n\n(8)\nn (methanol consumed)\u00a0=\u00a0n (formic acid produced)\u00a0+\u00a0n (formaldehyde produced)+ n (CO2 produced)\n\n\n\nFirstly, the surface electrochemistry of the investigated catalysts is studied by employing cyclic voltammograms and surface coverage analysis (section\u00a03.1). Subsequently, adsorption energies of the intermediates are discussed as a function of the Ni content in the catalyst (3.2). Catalytic activity towards methanol electrooxidation on Pd, Pd3Ni and PdNi and HPLC are given in Section\u00a03.3. CO stripping has been investigated to check catalyst tolerance (3.4). The MOR mechanism is elucidated for each case in section\u00a03.6. Section\u00a03.7 is dedicated to showing the effects of the solution pH and, finally, in section 3.8 a general discussion follows.The surface electrochemistry of the different catalysts (Pd, Pd3Ni and PdNi) is studied by recording cyclic voltammograms in 1\u00a0M NaOH (Fig.\u00a02\n (a)). Several processes are observed in the forward and backward voltammetric sweeps as a result of the typical surface reactions of Pd-based electrodes in alkaline media. Moreover, stable states of the catalytic surface with H, OH, O and mixed states with H2O are calculated using DFT as a function of the electrode potential and displayed in the form of Pourbaix diagrams [55] for the three investigated catalysts Pd, Pd3Ni and PdNi (Fig.\u00a02 (b)) on the (1 1 1) facet. The molecular states were averaged as a function of the electrode potential U using a Boltzmann distribution to properly illustrate the surface coverage and obtain insights on the onset potential where coverages change with the electrode potential (Fig.\u00a02 (c)).In the Pourbaix diagrams (Fig.\u00a02b), the thermodynamically stable phases are found for the curves showing the lower formation energy as function of the electrode potential U. As expected, and also shown in the voltammograms (Fig.\u00a02), at low potentials we find the catalyst surfaces mostly covered by hydrogen. The Boltzmann distributions on Pd, Pd3Ni and PdNi display a shift down with increasing Ni content of the onset potential at which H is the dominating component (black line in Fig.\u00a02b) with values of, respectively, 0.48, 0.39 and 0.28\u00a0V vs. RHE.With increasing electrode potential, the adsorption of OH and also H2O-OH mixtures gets more likely (blue and yellow lines in Fig.\u00a02b). Ni as an oxophilic element provides more OH on the surfaces. This effect correlates with the computed adsorption energies of OH (Table\u00a01\n), showing that a higher concentration of Ni in the catalyst produces stronger OH adsorption. Therefore, between the potentials 0.48 and 0.78\u00a0V vs. RHE, the pure Pd(111) catalyst must be mostly covered by OH, OH-H2O or OH-O-H2O mixtures, while for the Pd3Ni(111), it goes between 0.39 and 0.61\u00a0V vs. RHE and, finally, for PdNi(111) the interval is from 0.28 to 0.81\u00a0V vs. RHE. A wider interval is, hence, obtained for PdNi as compared to the Pd(111) and Pd3Ni(111) counterparts. Empirically assigning the OH adsorption process in the voltammetric curve is challenging for polycrystalline electrodes, but OH adsorption has also been reported to occur on Pd-based electrodes and this is usually attributed to an anodic process observed between the H- and O-regions [56].Oxygen-involved processes (O-region) are observed at higher potentials in the measured CV's, with the formation of Pd oxides occurring during the anodic sweep, and PdO reduction occurring during the cathodic sweep leading to a sharp peak at a potential near +0.78\u00a0V vs. RHE (Fig.\u00a02c). The performed calculations are in good agreement where oxygen emerges as the dominant component on the catalytic surfaces after 0.78\u00a0V vs. RHE for Pd and Pd3Ni and after 0.81\u00a0V vs. RHE for PdNi.Overall, the number of empty sites for the methanol electrooxidation (red lines in Fig.\u00a02c) is similar for all three catalysts. The interesting point here is that, on the alloys, H2O, O and OH adsorb preferentially on Ni sites (Table\u00a01) and this would facilitate the methanol catalytic oxidation on the free Pd sites. To illustrate this fact, the coverage containing 0.5\u00a0ML OH+H2O is displayed for the three cases in Fig.\u00a03\n. On the Pd(111) surface, the molecular state tends to form a hexagonal structure that maximizes the formation of hydrogen bonds. This geometry is broken on Pd3Ni due to the presence of Ni. Furthermore, a chain of water and OH molecules following the Ni sites is obtained for the PdNi catalyst, thus leaving free Pd sites for further methanol oxidation.Evaluation of the surface electrochemistry allows determining the ECSA. In this regard, measuring the charge involved during the reduction of PdO is the most employed method to determine the ECSA for Pd-based electrodes in alkaline media [57] considering that the reduction of a PdO monolayer takes 0.405\u00a0mC cm\u22122\n[58]. Table S1 shows the ECSA values calculated for the different catalysts in terms of electrode area and Pd mass. The ECSA was larger for the monometallic catalyst, but this is mainly due to the presence of a larger amount of Pd metal since the mass-normalized ECSA was similar (5.1\u20135.6\u00a0cm2 g\u22121 Pd) for all the catalysts. Accordingly, any differences observed in the MOR activity (vide infra) should be related to the influence of Ni on the reaction and not to a change in the electrochemical surface area.Here, the adsorption energies of methanol and the stable intermediate states along the methanol decomposition are computed as in eq.\u00a01 for potential 0 vs. RHE and shown in Table\u00a01 \u2013 more negative values mean stronger adsorption. The symmetric sites considered for the adsorption are: two top sites (TPd, TNi), three bridge sites (B2Ni, B2Pd, BPdNi), two fcc (F2NiPd, F2PdNi) and two hcp hollow sites (H2PdNi, H2NiPd); the superscripts indicate the atomic character of each site. Moreover, the last column of Table\u00a01 shows the trends of Ead vs. the Ni concentration where the \u21d1 symbol means that adsorption strength (Ead) increases with increasing Ni content while \u21d3 indicates a decrease and the first symbol refers to the changes from Pd to Pd3Ni and the second from Pd3Ni to PdNi. This provides insights into which species are more stable and which become less stable with the Ni alloying.The computed adsorption energy of CH3OH showed the top site as the most stable position for Pd(111), agreeing with other investigations using a different DFT functional [59,60] and also with cases like Pt(111) and Au(111) [45,49]. When considering the catalysts Pd3Ni(111) and PdNi(111), methanol prefers to adsorb on-top of Ni atoms instead of Pd atoms, but with lower adsorption energy as compared to the adsorption on Pd(111). Methanol bonds to the catalyst surface through a dative bond involving the oxygen lone pairs which explains the preference for binding to Ni instead of Pd. Furthermore, comparing Ead of methanol on Pd3Ni(111) and PdNi(111) a less negative value is obtained for the case of PdNi - weaker adsorption. Bader population analysis of Ni atoms in the investigated catalysts showed that the lower Ni concentration of Pd3Ni(111) leads to higher charge transfer from Ni to Pd atoms as compared to the PdNi case, i.e. Ni atoms are more positive on Pd3Ni(111), see Table S2. This produces a stronger interaction between the CH3OH and the metal surface. The same effect is also observed for water and formic acid (Table\u00a01).Another effect to be pointed out is the difference in the adsorption energy of a water molecule and a methanol molecule. During the electrochemical process, the first step to account for is the molecular adsorption of methanol on the catalytic surface. However, methanol and water come with similar adsorption energies leading to competitive adsorption with methanol interacting slightly more strongly with the metal surface by 0.02, 0.06 and 0.04\u00a0eV, for Pd, Pd3Ni and PdNi, respectively. The alloying process thus slightly favors the adsorption of methanol over water which is a positive sign for the electrocatalytic reaction.In general, adsorption energies show that intermediates with O bonding to the metal surface have their Ead strengthened when Ni is alloyed with Pd (including H2O, OH and O). For intermediates like CH2OH and CHOH the opposite trend is observed \u2013 CH2OH, for instance, prefers to adsorb at B2Pd positions highlighting the preference of such intermediates to interact with Pd instead of Ni. This configuration favors the formation of a \u03c3 bond between the C and the Pd [60]. The stabilization of O-containing species, and also the lowering of C interaction strength with the metals, have been reported for Sn alloyed with Pt and Ir alloyed with Pt [29,61]. Moreover, reduction of the CO adsorption energy is obtained where the CO bond strength is reduced when comparing Pd(111) and PdNi(111) catalysts.The MOR was experimentally studied on the different catalysts by recording cyclic voltammograms with 0.5\u00a0M CH3OH in\u00a01\u00a0M NaOH (Fig.\u00a04). The response normalized by the ECSA clearly shows the intrinsic activity of the catalysts under these experimental conditions. Significantly higher peak current densities were recorded with the bimetallic catalysts, which demonstrates the positive role of Ni to increase the MOR activity. Peak currents of 0.30, 1.31 and 1.88\u00a0mA cm\u22122 were obtained for the Pd, Pd3Ni and PdNi catalysts, respectively. A 6X increase was thus obtained with the bimetallic PdNi catalyst versus the monometallic Pd catalyst. Similarly, the overpotential required to produce the MOR at a specific current density decreased by addition of Ni. For instance, overpotentials of 845, 811 and 778\u00a0mV were obtained at 0.3\u00a0mA cm\u22122 (ECSA-normalized) for Pd, Pd3Ni and PdNi, respectively. This behavior agrees well with other studies reported in the literature, where bimetallic Pd catalysts have shown increased MOR activity [22].Moreover, product analysis was carried out and determined by HPLC as shown in\u00a0Fig.\u00a05. The electrochemical experiments were performed at a constant potential (+0.85\u00a0V) until reaching the same accumulated anodic charge (12\u00a0C) for all three catalysts, which took longer for the monometallic Pd catalyst as a result of its lower MOR activity. HPLC measurements were performed to determine the amount of formic acid and formaldehyde generated and methanol consumed during the reaction. The amount of CO2 generated was estimated by difference as described in the experimental section. Figure S3 (a) shows a typical chromatogram obtained for the reaction products, while Figure S3 (b) shows standard chromatograms of formic acid and formaldehyde employed to identify the products in the sample solution. Fig.\u00a05 shows the MOR product distribution obtained for the different catalysts. In all cases, the main products were formic acid and CO2, while formaldehyde was found in lower amounts. Formaldehyde selectivity was increased for the monometallic Pd catalyst while CO2 selectivity became higher for the Pd-Ni alloys. This suggests that the presence of Ni results in products of higher state oxidation, which would involve the transfer of a larger number of electrons, and could be one reason why larger currents are observed for the bimetallic catalysts [20,62].CO stripping experiments were carried out for the Pd, Pd3Ni and PdNi catalysts and the results are displayed in\u00a0Fig. 6\n\u00a0in the form of voltammograms. Slightly different CO oxidation responses were observed for Pd, Pd3Ni and PdNi. Sharp stripping peaks were obtained for CO oxidation using Pd3Ni and PdNi catalysts, with just a small shifting of the peak potential to +0.83 and +0.80\u00a0V, respectively. This difference would likely not have a significant effect on the MOR since the peak potential is lower than that observed for the MOR and, therefore, the CO oxidation in these catalysts during a MOR experiment is likely occurring efficiently. For the monometallic Pd catalyst, although the onset potential was similar, a larger value of the peak potential is observed (+0.87\u00a0V), and most interestingly, the CO stripping peak was broader. This fact suggests that CO stripping on the Pd catalyst is less efficient than on the PdNi catalysts, and the broader peak also indicates that the full CO oxidation is shifted to higher anodic potentials.The first effect explaining the higher tolerance towards CO-poisoning shown by the alloys is correlated to the adsorption energies of the intermediates and the electrode potential effect on the adsorption. Therefore, the effect of the electrode potential on the adsorption energies of selected intermediates is investigated for Pd(111), Pd3Ni(111) and PdNi(111) and shown in Fig.\u00a07\n, where the most thermodynamically stable products for each potential can be found.CO is found to be the most stable reaction product for potentials below 0.63\u00a0V vs. RHE for the pure Pd catalyst. Just above that, the formation of CO2 becomes more likely providing full oxidation of the reactant. A small downward potential shift is calculated for Pd3Ni and PdNi with values of 0.57\u00a0V vs. RHE and 0.56\u00a0V vs. RHE, respectively. This difference is mostly a result of the balance between the CO adsorption energies (Table\u00a01) and the CO2 adsorption energies. CO is 0.08\u00a0eV more strongly bound on Pd compared to Pd3Ni, and 0.01\u00a0eV more strongly on Pd3Ni compared to PdNi. On the other hand, the CO2 interaction strength is higher for the cases with Ni which changes the balance between the more likely products towards CO2. This agrees well with the peak potential shifts shown in Fig.\u00a06 resulting in a greater shift when going from Pd to Pd3Ni and a smaller shift when going from Pd3Ni to PdNi. Moreover, CO shows shorter bond distance for Ni-Cco than Pd-Cco in PdNi (Fig.\u00a08\n). This indicates that CO may bind close to Ni sites and possibly leaving Pd sites more available for the further oxidation processes.CHOO is also obtained as the most likely intermediate for PdNi for potentials between 0.54 and 0.56\u00a0V vs. RHE. In fact, highly oxidized intermediates like CHOO and COOH display stronger adsorption at higher potentials on PdNi compared to Pd. For instance, at 0.85\u00a0V vs. RHE (the potential used for the HPLC experiment), CHOO is the second most energetically stable intermediate for Pd3Ni and PdNi while, on the other hand, CO is the second most likely intermediate for Pd. Moreover, COOH emerged as the third most stable intermediate for PdNi. This emphasizes the tendency to obtain highly oxidized intermediates on the alloys providing greater number of electrons to the electrochemical reaction (in agreement with the voltammograms, Fig.\u00a04) and previous works [20].The solution pH (OH\u2212 ion concentration) could play a significant role in the reaction since the OH\u2212 ion seems to be involved in several intermediate steps and particularly important for the generation of products of high oxidation states, such as formic acid and CO2, where it also acts as a source of oxygen atoms required to generate the final product. Therefore, the role of OH\u2212 in the MOR activity for the different catalysts was also evaluated. Figure S4 shows the voltammograms recorded for 0.5\u00a0M methanol at different NaOH concentrations (0.1, 0.5 and 1\u00a0M) for the Pd and PdNi catalysts. MOR activity generally increased with the OH\u2212 concentration, demonstrating that OH\u2212 is involved in the reaction (scheme\u00a01), and it plays a positive role to obtain larger current densities. Notably, the ratio between anodic peak currents for the MOR between the PdNi and Pd catalysts (Figure S5) increased with increasing NaOH concentration in solution. This shows that the PdNi catalyst seems to be more tolerant or further benefitting from increased OH\u2212 concentrations.To determine the mechanism behind the higher activity of Pd-Ni alloys towards MOR, all elementary steps of the methanol oxidation reaction network are evaluated by DFT calculations and shown in Fig.\u00a09\n at 0\u00a0V vs. RHE for the investigated catalysts. The lower energy pathways are shown by the red arrows. Here we consider scission of C-H and O-H bonds while neglecting C-O cleavage and formation of methane due to the high barrier associated with this process [61]. The most stable configuration (site position) of the intermediates was used to build up these reaction networks and are shown in Figure S6. Moreover, activation barriers are calculated for specific reaction steps that we believe are the bottlenecks for the MOR and shown in Table\u00a02\n. The MOR mechanisms, considering also activation energies, are shown in green and blue color in Fig.\u00a09.The initial decomposition of methanol is characterized by the competition between losing a hydrogen from the O-H bond and forming methoxy (CH3O) or breaking a C-H bond and forming H2COH (hydroxymethyl). As discussed earlier, oxygen binds strongly to the catalysts containing Ni, hence, the formation of methoxy (CH3O) is preferable on Pd3Ni(111) and PdNi(111). This has also been reported for PtSn(111) and other compounds like Cu(111) and Ni(111) [61,63,64]. Pd(111) shows preference for C-H cleavage similar to Pt(111) [64]. Hence, based on thermodynamics, the methanol oxidation pathways on Pd(111) and on Pd-Ni alloys are different such that the subsequent intermediates on Pd(111) might be HCOH and COH while the alloys must form CH2O (formaldehyde) and HCO (formyl). In all cases, CO appears as the next deprotonation (abstraction from COH or CHO) and, hence, indicating an indirect oxidation path for the investigated catalysts towards the formation of CO2.\nCH3OH\u00a0\u2192\u00a0CH2OH\u00a0+\u00a0H or CH3O\u00a0+\u00a0H on Pd: The results provided by the HPLC show a reasonable quantity of formaldehyde formed in case of the Pd catalyst by applying 0.85\u00a0V vs. RHE. This indicates that checking only the thermodynamics of the reaction does not ensure the correct oxidative mechanism, since CH2O is not an intermediate on the minimum energy path for Pd(111) (highlighted in red in Fig.\u00a09a). The decomposition of methanol on Pt(111) has been experimentally investigated by Kruse et al.\n[65] using static secondary ion mass spectrometry (SSIMS), XPS, and pulsed-field desorption mass spectrometry (PFDMS). They proposed that the first H lost is based on the O-H bond breaking and forming methoxy (CH3O) that must further decay to form CH2O. Davis et al. [66] have used high-resolution electron energy loss spectroscopy (HREELS) to detect the presence of CH2O. These experimental investigations are in agreement with our HPLC results. Moreover, Jiang et al. [60] have used DFT to calculate the minimum energy pathway for methanol oxidation on Pd(111) and suggested that the activation barriers play a role. Yang et al. [59] studied methanol deprotonation in alkaline media and showed that the barrier of O-H bond breaking is reduced through the assistance of a hydroxyl group. Our experiments are performed in alkaline media where water is the main solvent. Therefore, H-bond formation between water molecules and adsorbed methanol is expected. Such a bond stabilizes the proton transfer (similar to a Grotthus mechanism) lowering its activation barrier. C-H bond breaking is also affected, but leading to increased activation barriers [48]. Hence, calculation of the activation barriers of the reactions CH3OH\u2192 CH2OH and CH3OH\u00a0\u2192\u00a0CH3O are performed using an explicit solvation model, as described in the computational details. The reaction CH3OH\u2192CH3O+H yielded a barrier of 0.5\u00a0eV barrier at 0.85\u00a0V while the C-H cleavage showed a barrier of 0.9\u00a0eV at 0.85\u00a0V (Table\u00a02), thus indicating a methanol oxidation path through methoxy and further going to CH2O (green line in Fig.\u00a09 (a)).\nThe OH coupling with CH2O, CHO and CO: As previously mentioned, the oxidation of intermediates plays an important role towards MOR. We have studied the activation barrier of the intermediates CH2O, CHO and also CO coupling to OH since these are the three possible channels where MOR can proceed to finally form CO2. The computed activation barriers (no solvation waters included) for the CO oxidation reaction \u2013 formation of COOH \u2013 are 1.1\u00a0eV, 1.2\u00a0eV and 1.5\u00a0eV for Pd, Pd3Ni and PdNi, respectively. Thus, CO removal emerges as a bottleneck to achieve the full deprotonation and oxidation to CO2. The oxidation of CH2O to CH2OOH showed barriers of 0.4\u00a0eV, 0.4\u00a0eV and 0.8\u00a0eV for Pd, Pd3Ni and PdNi, respectively, while the reaction CHO+OH\u2192 CHOOH displays a barrier of 0.2\u00a0eV for all studied catalysts.\nDeprotonation of CH2O and CHO: The deprotonation of CH2O and CHO to form CHO and CO competes with the OH-coupling during MOR (see Fig.\u00a09) and, hence, could affect the general reaction mechanism. The reaction CH2O\u2192CHO+H shows barriers of 0.3\u00a0eV, 0.8\u00a0eV and 1.2\u00a0eV for Pd, Pd3Ni and PdNi, respectively. The differences among the obtained barriers of the oxidation (OH coupling) and deprotonation of CH2O are-0.1\u00a0eV, 0.4\u00a0eV and 0.4\u00a0eV, for Pd, Pd3Ni and PdNi (with positive values meaning lower barrier for OH coupling), hence, it highlights that the oxidative reaction plays a more important role for Ni-containing catalysts. The next is the CHO deprotonation to form the poisoning CO intermediate. Spontaneous reactions were obtained for Pd and Pd3Ni, while PdNi displayed a 0.4\u00a0eV barrier. For PdNi, the OH coupling with CHO has a barrier of 0.2\u00a0eV and, hence, the oxidative reaction is more feasible than the deprotonation reaction leading to less CO poisoning on PdNi while the spontaneous deprotonation reaction must lead to a higher susceptibility to CO poisoning for Pd and Pd3Ni.In summary, the obtained activation barriers for Pd have redirected the MOR mechanism towards the formation of formaldehyde. Further, the deprotonation of CH2O proceeds as the more likely reaction until the formation of CO (green line in Fig.\u00a09 (a)). CO could, further, be oxidized to CO2, but, with the high activation barrier obtained for the reaction CO+OH \u2192COOH (1.1\u00a0eV), it is more likely that CO sticks to the catalytic surfaces forming a poisoning product. The oxidation of CH2O on Pd3Ni emerged as an alternative channel (blue arrow in Fig.\u00a09 (b)) that partially circumvents the formation of CO. On the other hand, any remaining CHO intermediate during MOR must directly go to CO, still forming CO with a high activation barrier for the oxidative removal to happen, even though in lesser quantity than Pd. The activation barriers computed on PdNi show that the oxidation of CH2O and CHO are more likely than the deprotonation (Table\u00a02) \u2013 differently from Pd and Pd3Ni (Fig.\u00a09 (c) green and blue). This yields less poisoning compared to the Pd and Pd3Ni counterparts and also a higher current density on the CVs due to the more effective reaction with OH occurring via two channels (also important to highlight that PdNi displays more OHads than Pd or Pd3Ni \u2013 see Fig.\u00a02).The aspects of the MOR mechanism on Pd, Pd3Ni and PdNi, studied based on a combination of experiments and DFT calculations, have indicated specific characteristics able to provide insights into their MOR activity and selectivity. Here, we discuss these features and the links between the experimental results and the DFT calculations.\nPd: The Pd catalyst showed lower tolerance towards CO poisoning compared to Pd3Ni and PdNi. Moreover, the reaction mechanism indicated in Fig.\u00a09 (a), and also the effect of potential on the adsorption energy of the intermediates, have confirmed the higher tendency of finding less oxidized intermediates on pure Pd as compared to the bimetallic alloys. These trends result in the higher selectivity of Pd towards formaldehyde, as also found in the HPLC experiment (formaldehyde is less oxidized than formic acid or CO2). Another effect that might play a role for the Pd higher selectivity towards formaldehyde is the CO poisoning. During the MOR reaction, CO must partially poison the Pd catalyst to a greater extent than the Pd-Ni alloys. This causes steric interactions between the adsorbed CO with the intermediates blocking the reaction towards CO2. Thus, earlier products like CH2O might appear in greater quantities on this catalyst compared to on the Pd-Ni alloys.\nPd3Ni: Pd3Ni has an alternative MOR route towards the formation of CO2 that is free of CO formation \u2013 oxidation of CH2O forming CH2OOH. That means a higher tolerance towards CO poisoning compared to Pd and, hence, a higher activity due to the higher number of electrons involved in the reaction (see Scheme\u00a01). The HPLC experiment found a lower quantity of CH2O on Pd3Ni compared to on Pd together with a higher amount of CO2, which might be due to the lower probability for formation of CO on the catalytic surface. It has also been shown that the oxidative removal of CO on Pd3Ni is more effective than on Pd. However, the high adsorption energy of CO governs the kinetics of the CO oxidation together with the availability of OHads species [9]. Due to the strong CO adsorption, a high activation barrier for the reaction CO+OHads\u00a0\u2192\u00a0COOH is found and, hence, this step might occur slowly. This indicates that the high activity displayed by Pd3Ni compared to Pd is, in fact, more related to the direct route towards the formation of CO2 and less related to the CO oxidation process. Jones et al.\n[67] proposed a direct non-CO MOR mechanism on a Pt-Ru alloy when the Ru concentration increases and leading to formation of formate (CHOO). The results shown here also indicate that the higher concentration of Ni changes MOR from an indirect to a direct mechanism.\nPdNi: Three main features must be emphasized for the PdNi catalyst: 1) PdNi has a strong tendency to form highly oxidized intermediates like CHOO (Fig.\u00a07 (c)). 2) The deprotonation reaction CHO\u2192CO+H exhibits higher barrier than the OHads coupling with CHO, hence providing an alternative route towards the formation of CO2. 3) Similar to for Pd3Ni, CH2O oxidation is a possible mechanism for the MOR on PdNi. The two possible direct routes (green and blue in Fig.\u00a09) indicate a higher tolerance against CO poisoning for this catalyst as compared to Pd and Pd3Ni. Finally, the fact that the highest activity towards MOR is displayed by PdNi is due to the two direct non-CO routes towards CO2 and also the tendency towards formation of highly oxidized intermediates as also confirmed by the HPLC experiment. It is important to highlight that, based on the Pourbaix diagram, this catalyst is the one with more adsorbed OH species, which effectively fuels the oxidation reactions of CH2O and CHO.The obtained results of the DFT calculations indicate the non-CO routes displayed by PdNi as the main reason for the higher activity towards MOR presented by this bimetallic alloy vs. pure Pd. The OH\u2212 concentration experiment supports this hypothesis since PdNi must be benefited by the OH\u2212 concentration to a greater degree than Pd due to the effective fueling of the non-CO paths on PdNi and, contrarily, the CO poisoning must act to block the MOR on Pd. Thus, the higher OH\u2212 concentration should enhance the MOR activity on PdNi more effectively than on Pd. The experimental results displayed a ratio between anodic peak currents for the MOR between the PdNi and Pd catalysts that increases with the OH\u2212 concentration, as expected. This, therefore, corroborates our DFT results and, moreover, shows that the enhanced coverage by adsorbed OH species leads to even greater performance of the bimetallic alloys towards MOR activity.The used models to describe the MOR contain simplifications that need to be considered. For instance, the structure of the surfaces, as shown in Fig.\u00a01 for Pd, Pd3Ni and PdNi may be expected to be subjected to variations due to strong interaction with intermediates like CO and thus the Ni/Pd distribution at the interface could be different from the one used in our model. Moreover, the experiments were performed with polycrystalline catalysts, which may add some differences to what we obtain with our model. Even if such differences can be of importance, the obtained results are in good agreement with the experimental data, which we assign to the local atomic environment dominating the interaction and different facets (with similar occurrence in the different samples) not significantly affecting the relative energies in the comparison of Pd, Pd3Ni and PdNi.Combining experiments and DFT calculations, we have successfully elucidated the benefits of a Pd-Ni bimetallic electrocatalyst towards the methanol electrochemical oxidation. We have firstly evaluated the catalytic surface coverages based on cyclic voltammograms and Pourbaix diagrams. The results revealed a shift of the onset potential where oxidation of the catalyst surface becomes the more likely process when Pd-Ni is considered. Moreover, the amount of OHads species on the catalytic surface is increased on PdNi as compared to Pd and Pd3Ni. This provides an effective OH fueling for the intermediate steps of the MOR involving OH. Higher activity and selectivity towards CO2 were also obtained for the bimetallic alloys in the cyclic voltammetry and HPLC experiments. These results are attributed to: i) different MOR mechanisms for Pd, Pd3Ni and PdNi, where Pd-Ni alloys emerged with non-CO routes for the methanol oxidation while Pd displayed a direct reaction. Notably, the PdNi catalysts showed that OHads can easily react to CH2O and CHO opening two routes towards CO2 that are free from CO. For Pd3Ni, differently, there is no barrier for the CHO\u00a0\u2192\u00a0CO+H+ reaction making the coupling of OHads with CHO less likely. ii) The oxidative removal of CO on Pd is inefficient and, therefore, this catalyst suffers CO poisoning to a greater extent than Pd-Ni. iii) The oxophilic property of Ni that makes highly oxidized intermediates more likely during MOR. Finally, the solution's pH \u2013 higher OH\u2212 concentration - benefits PdNi more than Pd due to the more important role found of the OH coupling for this electrocatalyst. Overall, alloying Pd with Ni shows to be an effective strategy towards an alternative electrocatalyst that delivers good performance towards alcohol oxidation at a reasonable price.RBA and ECdS performed the calculations and DMY the experiments. AC and LGMP conceived and supervised the study. All authors contributed to the writing of the manuscript.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the Swedish Foundation for Strategic Research (SSF) through grant number EM16-0010 and by the Swedish Energy Agency (Project 44666-1). The computations were enabled by resources provided by the Swedish National Infrastructure for Computing (SNIC) at the PDC and NSC centers partially funded by the Swedish Research Council through grant agreement no. 2016-07213.Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.electacta.2020.136954.\n\n\nImage, application 1\n\n\n\n", "descript": "\n                  Amongst promising available technologies enabling the transition to renewable energy sources, electrochemical oxidation of alcohols, in a direct fuel cell or in an electrolysis reaction (H2 production), can be an economically and sustainable alternative to currently used technologies. In this work, we highlight the advantages of a Pd-Ni bimetallic electrocatalyst for methanol electrooxidation - a convenient choice due to the low cost of Ni combined with the observed acceptable catalytic performance of Pd. We report a synergistic effort between experiments and theoretical calculations based on density functional theory to provide an in-depth understanding - at the atomistic level - of the origin of the enhanced electrochemical activity of methanol electrooxidation using the bimetallic catalysts Pd3Ni and PdNi over pure Pd. Cyclic voltammograms and High-Performance Liquid Chromatography (HPLC) demonstrate higher activity towards methanol electrooxidation with increased Ni concentration and, furthermore, higher selectivity for CO2. These effects are understood by: 1) changes in the methanol oxidation reaction mechanism. 2) Mitigation or suppression of CO poisoning on the Pd-Ni alloys as compared to the pure Pd catalyst. 3) A stronger tendency towards highly oxidized intermediates for the alloys. These findings elucidate the effects of a bimetallic electrocatalyst for alcohol electrooxidation as well as unambiguously suggest PdNi as a more cost-effective alternative electrocatalyst.\n               "}
{"full_text": "No data was used for the research described in the article.Since the 1950s, the amount of municipal waste generated by the manufacturing sector and home consumption has been increasing faster due to the expanding population and increased human activities. Common municipal waste includes plastics, used tires, old clothing, kitchen waste, paper, etc. [1]. The average weight of waste plastic, the polymerization or polycondensation of monomers, creates macromolecular; it makes up 10% of the annual worldwide trash production [2]. One of the essential plastic wastes is PSW, which makes a wide range of consumer goods. The petroleum sector typically provides the monomers, as in the case of the ethylene-based synthesis of PSW. Therefore, the production of polyethylene, polypropylene, polystyrene, polyethylene terephthalate, polyvinyl alcohol, and polyvinyl chloride has increased more than other plastics. PSW plastic is frequently used in items that demand clarity, such as food packaging and laboratory equipment, as it is a complex, solid plastic. Recent scientific reports claim that 7\u20138 billion tons of solid waste are created annually around the globe, posing a significant problem for the scientific community due to the numerous negative impacts on the environment and human health [3]. In waste-to-energy plants, very little of the plastic that we discard every day gets recycled. A large portion of it is disposed of in landfills, where it might take up to 1000 years to degrade and release potentially harmful materials into the soil and water. Burning, another way to dispose of plastic waste, is detrimental to the environment since burning releases hazardous chemicals into the air. Thus, finding the right recycling solution is necessary since burning and burying plastic waste plastic is extremely harmful to the environment.In view of this scenario, waste plastic treatment means must be applied to improve their valorization rate. The thermochemical methods are among the most promising for application, and growing interest is being shown in the thermal treatment alternatives of pyrolysis, steam reforming, and combined pyrolysis/reforming systems as feasible alternative environmental and financial solutions for plastic waste processing. Compared to landfilling or conventional waste incineration, the in-situ pyrolysis-catalytic steam reforming reaction method provides various benefits, such as conversion plastic wastes and producing H2 and liquid fuels rather than harming the environment. Accordingly, recycling of PSW is substantial for environmental remediation and moves towards a more circular plastic economy. Pyrolysis manages plastic waste sustainably while producing solid char, gases, and liquid oil as energy sources [4]. Complex compounds or long-chain hydrocarbons are thermally broken down into simpler molecules or shorter-chain hydrocarbons. Additionally, using raw bio-oil from pyrolysis is difficult since it has a high oxygen concentration that concurrently lowers its energy content [5]. Furthermore, PSW temperatures seldom reach the levels required for thermal deterioration; hence this phenomenon is uncommon [6]. As a result, dissolving PSW in a dissolving agent may be a great answer to PSW problems and the creation of clean and renewable energy. As previously investigated [7,8], phenol is an effective dissolving agent since it is acidic to certain plastics, rubber, aluminium, its alloys, and lead. Phenolic constituents often result from the production of petrochemical by-products [9] and makeup around 38% of the unwanted pyrolysis oil ingredient [10]. In addition to the practical liquid fuel generation, the employment of phenol can also allow H2 generation from in situ pyrolysis-catalytic steam reforming reaction. The chemical H2 is essential for many industrial processes and may one day serve as a source of clean energy. Thus, using PSW plastics as feedstocks for valuable liquid goods and phenol as a source for chemicals like H2 will encourage the recycling of plastic waste, stop the difficulties created by the waste plastics, and act as an alternative supply of chemicals, enabling a circular economy.In the pyrolysis-catalytic steam reforming reaction, catalysts are frequently employed to improve product dispersion and raise product selectivity. At the in-situ catalytic pyrolysis processes, the catalyst and feedstock are mixed, and the pyrolysis and vapour catalytic reforming/cracking processes take place in the same reactor; therefore, the capital and operating costs are reduced. Additionally, catalysts have been used to upgrade pyrolysis products such that the hydrocarbon distribution is improved and has characteristics comparable to traditional fuels like diesel and gasoline [11,12]. For the sustainable production of H2 and liquid fuel, large-scale commercial use of noble metals such as Pd [13\u201315], Rh [16,17], and Pt [18,19] was employed and demonstrated the best catalytic activities; however, they are prohibitively costly. As an alternative, lots of studies on the design of non-precious metal-based catalysts for the H2 generation, such as bimetallic Ni-Co [10,20\u201324], Al [25,26], and Fe [27,28], have been investigated. In contrast, significant challenges such as unstable changes in morphology after the reaction, low stability, and selectivity remain. Specifically, as appealing substitutes to traditional noble metal-based catalysts, noble metal-free catalysts could have a promising future in initiating reforming and cracking reactions. Recently, the Ti@TiO2 core-shell nanoparticles were stated as a precious metal-free photocatalyst for the photothermal H2 production from aqueous glycerol solutions [29]. Yancheng et al. [30] used porous copper (Cu) foam as catalyst support for H2 production from methanol steam reforming reaction. Though its overall efficiency was lower, the microreactor utilizing Cu foam covered with 0.6\u00a0g of catalyst had more excellent methanol conversion and H2 generation rates per catalyst weight than when foams with higher catalyst loading were utilized. The previous investigation on the addition of Cu to Ni/Al2O3 catalyst found that low loadings of Cu served to lessen the alloying impact brought on by Cu enrichment, which helped prevent the production of carbon on the catalyst [31]. Additionally, it has been claimed that the low-valence Cu species (Cuo or Cu+) in the majority of Cu catalysts are active species [32]. Developing and preparing effective precious metal-free catalysts for H2 production from the in-situ pyrolysis-catalytic steam reforming reaction of PSW dissolved in phenol is the most difficult challenge in this sector. There is a sizable possibility for carbon to develop and be deposited on the catalyst's surface since this process requires the removal of H2 from phenol and liquid fuel hydrocarbons from PSW. Therefore, additional precious metal-free catalysts must be developed to present potential candidates as precious metal catalyst substitutes.To the best of our knowledge, studies are deficient for explaining the effect of the chemical and physical properties on the selectivity and coking resistance of the Ti-Cu nano-catalyst in in-situ pyrolysis-catalytic steam reforming conditions. Herein, we report the facile synthesis and characterization of a precious metal-free Ti-Cu nano-catalyst and their favourable catalytic properties towards H2 and liquid fuel generation from in-situ pyrolysis-catalytic steam reforming reaction of PSW dissolved in phenol. Catalysts before and after reaction were characterized by several methods correlating their structural and textural characteristics with catalytic activity. In the present work, three precious metal-free nano-catalysts, Ti, 4Ti3Cu (4\u00a0g Ti with 3\u00a0g Cu) and 3Ti3Cu (3\u00a0g Ti with 3\u00a0g Cu), were prepared by hydrothermal and impregnation methods to research the influences of Cu addition on the catalytic performance of the catalysts. The crystallinity and synergistic effect of calcined Ti\u2013Cu on the inhibition of materials sintering and the surface area along with pore size distribution of the fresh catalysts were characterized by X-ray diffraction (XRD) and the Brunauer, Emmett and Teller theory (BET), respectively. The basicity study was conducted by the temperature programmed desorption of carbon dioxide (CO2-TPD), pyrrole-differential thermogravimetric analysis (DTG) curves, and pyrrole-FTIR spectra. Pyridine FTIR spectra and pyridine - DTG curves of the fresh catalysts were used to illustrate the Br\u00f8nsted and Lewis acid sites. Transmission electron microscopy (TEM) and H2 temperature-programmed reduction (H2-TPR) were used to demonstrate the benefits of Cu doping on the dispersion and reducibility of Ti. Fourier-transform infrared spectroscopy-potassium bromide (FTIR-KBr) was used to examine functional groups present in the synthesized catalysts. Catalysts were tested in a fixed-bed reactor that is modified for higher PSW plastic waste to be reacted compared with our previous research [14,33\u201335], and the optimum catalyst base on the highest phenol conversion and H2 yield was tested at 500\u2013800\u00a0oC and 45\u00a0h on stream. The produced liquid product samples were characterized by gas chromatography/mass spectrometry (GC/MS) and FTIR systems. Used catalysts after experiments were also collected and characterized by thermogravimetric analysis (TGA), FTIR-KBr, BET, TEM and CHNS.Nano catalysts Ti, 4Ti3Cu, and 3Ti3Cu were synthesized by hydrothermal technique with the Ti to Cu mass ratio of 1, 3:3, and 4:3, respectively. The starting reagents of titanium and copper were titanium (IV) oxide (TiO2, with the purity of 99.8%) and copper (II) nitrate trihydrate (Cu(NO3)2.3H2O, with the puriss. p.a. grade of 99\u2013104%) which were acquired from Sigma-Aldrich and the synthesis stages are shown in Fig. S1. In accordance with our previous exploration [36,37], nano-sized Ti and Cu catalysts were separately gone through hydrothermal treatment, in which those materials were first stirred with 100\u00a0mL of deionized water at room temperature. 5\u00a0M of sodium hydroxide (NaOH) was gently dissolved with the solution to improve the nucleation and growth rates of the nanoparticles [38] and stirred for three hours at room temperature to form a clear mixture. Then, the mixture was transferred into a 100\u00a0mL Teflon-lined autoclave reactor and was hydrothermally treated at 160\u00a0\u00b0C for two days. The solid precipitate was repeatedly centrifuged (400\u00a0rpm) to separate the solid products from the liquid phase, filtered and washed with deionized water 15 times via filter paper on a Buchner funnel that was sealed with a rubber bung on the top of a side arm conical flask. The side arm of the flusk was connected with a vacuum pump to speed the filtration and washing process of the samples, followed by drying at 110\u00a0\u00b0C overnight and then calcination for three hours at 800\u00a0oC. In order to cure and harden catalysts for industrial use, remove impurities, and drive out chemically bonded moisture, calcination is a crucial step in the process of making catalysts. The prepared nano-sized Ti and Cu particles then went through the conventional impregnation method for synthesizing 4Ti3Cu and 3Ti3Cu. The detail of preparation is explained in our previous research [23,24]. In brief, a specific quantity of the calcined Ti was mixed in 150\u00a0mL of deionized water and stirred for an hour at 90\u00a0oC, and then the calcined Cu was introduced into the mixture. After vigorous stirring for a few hours, the liquid was evaporated, and a slurry was produced and dried overnight in an oven at 110\u00a0oC. Lastly, the acquired dried solid was calcined in an oven (Model Ney Vulcan D-130) at 800\u00a0\u00b0C for 3\u00a0h (30\u00a0\u00b0C\u00a0min\u22121).XRD curves were obtained employing D8 ADVANCE Bruker X-ray diffractometer operated at 40\u00a0mA and 40\u00a0kV with Cu Ka radiation at 2 theta of 10\u2013100\u00b0. The crystalline phases were classified by JCPDS (Joint Committee on Powder Diffraction Standards) using X'Pert Highscore Plus software, and crystal sizes were estimated from diffraction line widths using the Scherrer equation. Nitrogen (N2) adsorption\u2013desorption performances for fresh and used samples were obtained in a Beckman Coulter SA3100\u2122 apparatus using liquid N2 at \u2212\u00a0196\u00a0\u00b0C. Each catalyst was degassed at 200\u00a0oC under a vacuum for 3\u00a0h before the adsorption experiments. The BET technique was used to evaluate the specific surface area. At the same time, the average pore size was calculated using the Barrett-Joyner-Halenda (BJH) technique utilizing the adsorption curve to get the total pore volume at relative pressure P/Po =\u00a00.99. The fresh and used TEM images were acquired using a JEOL JEM-1011 microscope that functioned at 80\u00a0kV. TEM specimens were equipped by dispersing the catalyst powder in acetone with sonication and dropping it onto an ultrathin carbon-coated Cu grid. TGA-DTG analysis of the used catalysts was carried out using a Shimadzu TG-50 thermogravimetric analyzer via the flow of N2 to heat the samples from 30\u00b0 to 800\u00b0C with a heating rate of 20\u00a0\u00b0C\u00a0min\u22121. The H2-TPR was accomplished on a Micromeritics Chemisorb 2720 apparatus, and the analysis was carried out in a pure H2 at a flow rate of 30\u00a0mL/min, and the temperature was increased from room temperature to 900\u00a0\u00b0C with a heating rate of 20\u00a0\u00b0C\u00a0min\u2212\u00a01. CO2-TPD was also conducted on the same device to detect the basicity of the catalyst. The samples were put into a quartz tube and pretreated in a Helium (He) flow at 250\u00a0\u00b0C for 1\u00a0h and then cooled down to room temperature naturally. The catalyst samples were exposed to the CO2 environment at 110\u00a0oC after the pre-treatment step until their surface sites reached their saturation state. After attaining saturation, the samples were flushed with inert gas He. Further, the temperature was increased to 900\u00a0oC with a ramp rate of 20\u00a0oC/min to determine the quantity of desorbed CO2 from the surface basicity sites using a thermal conductivity detector (TCD). The elemental linkage information of the fresh and used samples was studied by FTIR curves detailed via a Shimadzu IR-Prestige-21 model spectrometer using pure KBr as a reference background record with a scanning range of 400\u20134000\u00a0cm\u22121. The KBr pellet was prepared by mixing the catalyst with KBr with a mass ratio of 100:1, and the excellently prepared combination was pressed to procedure a 13\u00a0mm diameter pellet. The same apparatus was used to determine the functional cluster presented in the liquid products in addition to the GC/MS (Agilent 7890B).The reaction performance of the prepared Ti, 4Ti3Cu, and 3Ti3Cu was investigated via the combination of a fixed-bed reactor and online mass spectrometry. The case length was 300\u00a0mm, and the internal diameter was 8\u00a0mm at atmospheric pressure; the diagram of the experimental apparatus is illustrated in \nFig. 1. 0.2\u00a0g of the catalysts were located inside the reactor, and the temperature of the catalyst bed was measured and controlled by a K-type thermocouple, which was linked to a temperature controller. The catalyst was reduced in place for one hour at 600\u00a0oC using 30\u00a0mL/min of pure H2 after flushing the catalyst bed with N2 at 300\u00a0oC. The water was fed into the pre-heater using a high-performance liquid chromatography pump (HPLC Bio-RadTM, Series 1350) to inject the fuel with 0.36\u00a0mL/min before mixing with carrier N2 (30\u00a0mL/min). In our previous research [14,33\u201335], we used a very small amount of plastic waste dissolved in phenol to avoid the blockage of the line before the reactor and experimental limitation. To increase the feasibility of the reaction and the amount of plastic waste in the reaction, we modified the experimental rig with a Parr Benchtop Reactor. Herein, the slurry of phenol and PSW plastic with the volumetric ratio of 5:1 was mixed with water vapour molecules and fed to the reactor using a pressurized Parr Benchtop Reactor that kept the PSW-phenol slurry in an aqueous phase at 70\u00a0oC, and transfer pipes were swathed using glass fibre heating tape and preheated at 200\u00a0oC. With the volumetric ratio of water to PSW-phenol solution of 10:1 with two mass flow controllers individually equipped for phenol line and water line that precisely monitored the flowrates of reactants; the water to PSW-phenol vapour was pumped into the reactor. For activity testing, all catalysts were teated at 500\u00a0oC. The optimum catalysts were tested in reaction temperatures ranging from 500\u00a0oC to 800\u00a0oC with a gap of 100\u00a0oC, and the relevant performance results were collected in steady-state situations. Likewise, the constancy examination of the optimum catalyst was conducted at 500\u00a0oC for 45\u00a0h. After the reactor, a condenser was installed and connected with a circular cooling system at 10\u00a0oC to liquefy the condensable liquid molecules, followed by a liquid gas separator. The components present in the gas products were analyzed online employing a GC-TCD (Agilent 6890\u00a0N), and the liquid product was analyzed using a GC-FID (HP 5890 Series II) equipped with a 0.53\u00a0mm\u00a0\u00d7\u00a030\u00a0m CP-Wax capillary column and GC/MS (Agilent 7890B). Each run was repeated at least six times to ensure accuracy and reproducibility. The result analyses, such as phenol conversion, produced gas composition in yield, were calculated following our previous research [14] and as shown in Eqs. (1), (2), (3), and (4).\n\n(1)\n\n\nPhenol\nconversion\n(\n%\n)\n=\n\n\n\n\n[\nPhenol\n]\n\n\nin\n\n\n\u2212\n\n\n[\nPhenol\n]\n\n\nout\n\n\n\n\n\n\n[\nPhenol\n]\n\n\nin\n\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n(2)\n\n\n\n\nH\n\n\n2\n\n\nyield\n(\n%\n)\n=\n\n\nmoles\nof\n\n\nH\n\n\n2\n\n\nobtained\n\n\nmoles\nof\n\n\nH\n\n\n2\n\n\nstoichiometric\n\n\n\u00d7\n100\n\n\n\n\n\n\n(3)\n\n\nCO\nyield\n\n\n\n%\n\n\n\n=\n\n\nmoles\nof\nCO\nobtained\n\n\nmoles\nof\nCO\nstoichiometric\n\n\n\u00d7\n100\n\n\n\n\n\n\n(4)\n\n\n\n\nCO\n\n\n2\n\n\nyield\n\n\n\n%\n\n\n\n=\n\n\nmoles\nof\n\n\nCO\n\n\n2\n\n\nobtained\n\n\nmoles\nof\n\n\nCO\n\n\n2\n\n\nstoichiometric\n\n\n\u00d7\n100\n\n\n\n\nThe quantity of chemicals that react for the reaction to be fully catalyzed is known as the stoichiometric moles. So, for example, Eq. 5 represents the balancing steam reforming equation.\n\n(5)\n\n\n\n\n\n\nC\n\n\n6\n\n\nH\n\n\n5\n\n\nOH\n+\n11\n\n\nH\n\n\n2\n\n\nO\n\u2194\n6\nC\n\n\nO\n\n\n2\n\n\n+\n14\n\n\nH\n\n\n2\n\n\n\n\n\u0394\nH\n\n\no\n\n\n=\n463.65\n\n\nkJ\n\n/\n\nmol\n\n\n\n\n\n\n\n\n(6)\n\n\n\n\n\n\nC\n\n\n6\n\n\nH\n\n\n5\n\n\nOH\n+\n5\n\n\nH\n\n\n2\n\n\nO\n\u27f6\n6\nCO\n+\n8\n\n\nH\n\n\n2\n\n\n\n\n\u0394\nH\n\n\no\n\n\n=\n710.91\n\n\nkJ\n\n/\n\nmol\n\n\n\n\n\n\n\n\n(7)\n\n\nCO\n+\n\n\nH\n\n\n2\n\n\nO\n\u2194\nC\n\n\nO\n\n\n2\n\n\n+\n\n\nH\n\n\n2\n\n\n\n\n\u0394\nH\n\n\no\n\n\n=\n\u2212\n41.15\n\n\nkJ\n\n/\n\nmol\n\n\n\n\n\n\n\n\nTable 1 depicts the surface area, pore volume and pore diameter of Ti, 4Ti3Cu, and 3Ti3Cu catalysts, defined by N2 adsorption\u2013desorption investigation. As can be seen in Table 1, introducing Cu causes produce different surface areas, pore sizes and pore volumes. The BET surface area of Ti, 4Ti3Cu, and 3Ti3Cu is 15.912\u2009m2/g, 5.374\u2009m2/g, and 3.73\u2009m2/g, respectively. Obviously, higher Cu contents nano-catalysts cause to decrease in the surface area, partial collapse of ordered mesoporous structure and the rise of the pore diameter (Dp). Typically, in comparison to bare Ti catalyst, 3Ti3Cu catalyst only displayed a surface area of 3.73\u2009m2/g, and its pore volume was low. The sintering of nanoparticles during the calcination process might be significantly delayed by Ti components having a mesoporous structure, preserving a greater specific surface area. This decrease is caused by the inclusion of Cu, which partially covers the catalysts' surfaces and blocks some of their pores. The catalyst made of Cu may be simple to sinter at 600\u2009\u00b0C in a reducing atmosphere, which leads to the growth and agglomeration of the active particles and a reduction in the specific surface area [39]. \nFig. 2 displays the pore size distribution and the N2 isotherms for the fresh catalysts. Ti and 4Ti3Cu isotherms can be categorized as type-IV ascribed to the mesoporous structure [40]. However, the Ti and 4Ti3Cu samples' distribution pore size curve displayed a peak in the area above 50\u2009nm, indicating the presence of macropores in their structure. The 3Ti3Cu material, on the other hand, exhibits a type IV isotherm and an H4 hysteresis curve, both of which are typical of slit-shaped pores [41]. This phenomenon is caused by developing tiny CuO micro crystallites that partially obstruct mesopore access.The XRD outlines of as-synthesized materials are displayed in \nFig. 3; estimated crystallite sizes are included in Table 1, and the crystal sizes are diverse from 67 to 93\u2009nm for all nano-catalysts. Diffraction peaks are prominently located at about 39.2\u00b0 (200), 70.6\u00b0 (122) and 76.5\u00b0 (202) and marked with blue stars, matched with the standard XRD pattern of rutile Ti2O4 (JCPDS, No. 96\u2013900\u20137433) and match with the crystallite sizes of 80.8\u2009nm. The characteristic diffraction peaks of spinel phases at 2\u03b8 of 25.7\u00b0, 37.5\u00b0, 38.4\u00b0, and 69.04\u00b0 were observed (marked with red hearts) and can be ascribed to representative peaks of (101), (103), (004), and (116) crystal phases and corresponding to the JCPDS number of 96\u2013101\u20130943 for anatase Ti4O8, and equal with the crystal size of 42.3\u2009nm. Meanwhile, the peaks appearing at 48.6\u00b0, 54.3\u00b0, 55.7\u00b0, 63\u00b0, 75.9\u00b0 and 82.9\u00b0 could be assigned to the characteristic peaks of 202, 023, 151, 061, 151 and 402 crystal phases and correspond to the JCPDS number of 96\u2013900\u20139088 for orthorhombic phase structure of brookite Ti8O16, and correspond with the crystallite size of 98.2\u2009nm. After introducing CuO to the TiO2, multiple new diffraction peaks for the tenorite (Cu4O4) (JCPDS, No. 96\u2013110\u20130029) were detected (marked with green trefoil shapes) at 35.6\u00b0, 49.1\u00b0, 58.7\u00b0, 61.9\u00b0, 66.1\u00b0, and 72.9\u00b0 which could be assigned to the diffraction peaks of 002, 202, 202, 113, 022 and 311 monoclinic phase structures, respectively, and equal to the crystallite size of 47\u2009nm. The decrease in intensities after introducing the CuO to TiO2 reveals that the Cu is highly dispersed in the catalysts or causes a crystal size reduction. The XRD patterns of 4Ti3Cu and 3Ti3Cu catalyst resulted in two more new peaks at 2\u03b8 of 27.6 and 68.6, consistent with 110 and 126 crystal planes, which can be attributed to the monoclinic [JCPDS 96\u2013153\u20139683] structure of TiO2 and marked with purple triangles (\u223c70\u2009nm). The catalysts synthesized by the hydrothermal method showed in addition to obvious copper oxide and titanium dioxide phases, the bimetallic oxide catalyst Cu\u2013Ti was also successfully synthesized in this experiment. This phenomenon can be approved by the appearance of the peak with the blue circle at 32.9\u00b0 (\u223c60.6\u2009nm) and can also be attributed to the diffraction peaks of 101 tetragonal phase structures of CuTi3 alloy. However, for the 3Ti3Cu catalyst, the diffraction peaks' intensity is low, meaning that the Ti might be highly dispersed on the Cu. It might be the \"combustion\" process that gives the catalyst with enhanced Ti dispersion and smaller crystal sizes. Considering the low intensity of XRD peaks of the 3Ti3Cu catalyst, one can say that Ti and Cu are present mainly in the amorphous phase in the synthesized catalysts. This also could suggest that the hydrothermal method (for Ti catalyst) facilitated the crystal growth during material preparation. In the impregnation method (for 4Ti3Cu and 3Ti3Cu), the metals are distributed over the lower layer of the support of the catalyst; this causes the intensity of the peak to decrease compared to that prepared by the single component (Ti) hydrothermal method.In this work, the TEM technique was employed to identify the position of the particles, size and morphologic features in the 3Ti3Cu nano-catalyst; corresponding results are shown in \nFig. 4. The nano-sized catalysts constituted a rectangular-like shape of TiO2 with an average diameter of \u223c200\u2009nm, indicating that the TiO2 core is highly crystallized with better dispersion. However, Cu presented an uneven element form with less particle scattering. It also comprises segregated particles and appears in spherical shapes in mostly amorphous phases that are in good agreement with XRD results. This finding might be because of the less BET surface area and XRD crystal size of 3Ti3Cu compared with Ti catalyst, which was reported in previous sections. Less crystal size plays a substantial part in minimizing the coke production and deposition and enhancing the catalyst lifetime throughout the in-situ pyrolysis-catalytic steam reforming reaction process. Furthermore, as shown from the TEM micrograph, the contact between Ti and Cu particles was lower than those Ti particles. This suggests that Ti-Cu ensembles may be formed due to the interaction between Ti and Cu. And it is also confirmed that the highly dispersed nano-catalysts could be obtained with inexpensive materials via this simple hydrothermal-impregnation method.As shown in \nFig. 5, the FTIR spectrum utilizing the KBr pellet approach was obtained in the wavenumber range of 4000\u2013400\u2009cm\u22121 to analyze the practical clusters in the manufactured catalysts. The phenyl ring vibrations, such as \u03b3(C\u2212C\u2212C), are identified at 1265\u2009cm\u22121\n[42]. This peak was shifted to 1203\u2009cm\u20131 after introducing a CuO component to the TiO2. Giuseppe et al. [43] mentioned that the 1265\u2009cm\u22121 peak could be assigned to the C\u2013O vibration in guaiacyl rings that clearly shifted to a pronounced broad and strong band around 1203\u2009cm\u20131 for the 4Ti3Cu and 3Ti3Cu catalysts due to the bending vibrations of amino acids side chains [44] and attributed to the asymmetric stretching vibrations of C\u2013O\u2013C in all samples [45]. A weak peak corresponding to CO is observed at 617\u2009cm\u20131 for the bare Ti catalyst that can be assigned to \u03c9O1-Ti3-O2\n[46]. After introducing the Cu, the intensity of this peak is slightly increased and matches the metal oxide stretching, i.e., Cu\u2013O bond in the monoclinic phase, which specifies CuO nanoparticles formation [47]. The band at 1018\u2009cm\u22121 (symmetric O\u2013C\u2013O stretching [48]) belongs to the C\u2212H and N\u2009\u2212\u2009H in-plane deformation vibrations [49], and might also attribute to alkoxy groups attached to titanium ions in the catalysts [50]. FTIR bands at 1689\u2009cm\u22121 are attributed to CN vibrations modes [51]. Likewise, this band could also correspond to the \u03bd (C\u2212O) mode of a carbonyl compound formed after adsorption [52].H2-TPR characterization was implemented on all calcined samples to study the effects of adding Cu to Ti on the catalyst reducibility and the interaction between metal and support. It had been reported that the TPR profiles of Ti-based catalysts were affected by the interaction between Ti and Cu. The reduction profiles of all catalysts are shown in \nFig. 6(a), and the H2 consumption are listed in Table 1. The less H2 consumption corresponds to the poor reducibility of the catalyst. Compared to 4Ti3Cu and 3Ti3Cu samples, only a small trace of H2 uptake in the Ti sample was detected, which might be associated with the reduction of the remaining TiOx species in deficient concentration. The fact that the mixed metal oxides had solidified into a solution and the synergetic effects had boosted the reducibility is another potential explanation for the absence of pure Ti's reduction peaks. The observation of the low-temperature shoulders (281\u2009\u00b0C and 315\u2009\u00b0C), detected in reduction profiles of 4Ti3Cu and 3Ti3Cu, is possibly ascribed to the reduction of Cu2+ to Cu0 in aggregated copper oxide species. These characteristics specify a highly distributed Cu2+ species in 4Ti3Cu and 3Ti3Cu catalysts, and the presence of these species can assign to the strong interaction among Cu and Ti. The two reduction peaks at 443\u2009\u00b0C and 447\u2009\u00b0C could be the reduction of monomeric Cu+ to Cu0\n[53,54]. As a result, the species are reduced at much greater temperatures than Cu2+ species, which are associated with titanium copper alloy. As a result, we may conclude that the impregnated catalyst mainly comprises the copper oxides copper oxide and Cu-Ti. Due to the limited quantity of alloy or the high metal dispersion on the support, the Cu-Ti alloy peak is very weakly visible in the XRD examination. As presented in the figure, two main reduction peaks of the 4Ti3Cu catalyst were detected at about 443\u2009\u00b0C, and 545\u2009\u00b0C, which might be attributed to the reduction of surface oxygen and bulk oxygen of Cu, respectively, or may account for two overlapping reductions steps of copper oxide into Cuo. Noticeably, compared with the 3Ti3Cu catalyst, the reduction peaks of the 4Ti3Cu catalyst obviously shifted to a lower-temperature direction. These results reveal that the Ti species over 3Ti3Cu and 4Ti3Cu catalysts are more reducible, indicating that the reducibility of catalysts is promoted with higher Ti content. The lower the peak temperature, the better the reduction performance of the catalyst is. However, the peak area of the 3Ti3Cu catalyst was the largest, which can be concluded that the introduction of the Cu component promotes the reduction of the 3Ti3Cu catalyst, has the best redox capacity, largest H2 consumption, oxygen storage capacity and inexplicable interactions among Ti and Cu as proven by the reduction peak at 600\u2009\u00b0C.The CO2-TPD technique was employed to rationalize surface basicity and investigate basic sites' strength and distribution. The basicity curve is shown in Fig. 6(b). The quantitative analysis of surface-adsorbed CO2 was conducted using the total peak area under the curves, and the findings are reported in Table 1. Nuanced surface characterization of porous materials is made possible by the adsorption of probe molecules, and molecules with certain characteristics (such as basic or acidic) can interact with the surface active sites that are inside the pores or between the layers. Therefore, we further characterized the basic sites by using pyrrole as a probe molecule in the DTG curve (Fig. 6(c)) and FTIR-KBr curve (Fig. 6(d)). The proportion of basic sites may be used to gauge the total basicness, leading to the following pattern: 3Ti3Cu >\u20094Ti3Cu >\u2009Ti. This trend proves that the pure Ti had the weakest desorption peaks of CO2, indicating its basicity was very weak and more basic sites are in the sample with higher Cu containing, while the basic strength of the sites is nearly the same. As shown in Fig. 6(b), the CO2-TPD curves can be separated into three sorts of peaks equivalent to weak (50\u2013250\u2009\u00b0C), moderate (250\u2013600\u2009\u00b0C) and strong (>600\u2009\u00b0C) basic phases for the catalysts. By moving the basic site peaks from 675\u2009oC to the stronger area at 698\u2009oC, the addition of Cu2+ to TiO2 not only increases the value of overall basicity but also alters the distribution of basic sites. Catalysts with higher Ti concentrations (Ti and 4Ti3Cu) occupied strong basic strength within the weak, medium, and strong areas. However, the 3Ti3Cu catalyst had almost the same peaks at weak and medium regions, but one big intensity in the high-temperature range after 600\u2009\u00b0C. This finding implies that because of the fundamental properties of the catalysts, the impregnation of Cu and Ti may promote additional surface sites for CO2 adsorption.The Pyrrole-DTG profile (Fig. 6(c)) illustrated that the peak intensities observed at the 200\u2013500\u2009\u00b0C region match the trend of CO2-TPD and Pyrrole-FTIR spectra (3Ti3Cu > 4Ti3Cu > Ti). Pyrrole forms an adsorption layer on the Br\u00f8nsted sites via the cycle's two subsequent carbon atoms, and these layers have substantially lower adsorption energies than Lewis sites. Pyrrole is an amphoteric molecule that may function as a proton acceptor through its \u03c0 electron orbital or a proton donor to interact with basic sites on the surface. Surface basic sites interacting with pyrrole are caused by the stretching and bending vibrational modes of surface formate (both \u2212CH and \u2212COO) and carbonate species over CuO and TiO2 constituents, according to the Pyrrole-FTIR spectra. Bands at 1165\u2009cm\u22121, 1481\u2009cm\u20131, and 1736\u2009cm\u20131 were attributable to the C\u2013CO\u2013C stretch and bending, absorption of the phenyl ring, and C\u2550O groups in amorphous Cu, respectively. There is a small shoulder at 1442\u2009cm\u20131, which can be proven that the Cu cause to increase in the Lewis basicity of the catalysts. This observation could explain that the increase in the basicity determined by CO2-TPD with higher Cu charges is attached to the higher contribution of Lewis basic sites. The prevailing consensus is that Lewis basic sites correspond to oxygen anions with poor coordination formed as basic sites following the calcination stage. The coordination of the Lewis sites linked to O\u22122 anions determines their basicity. The oxygen atoms in the crystal corners have to be more basic than the oxygen atoms on the crystal faces or the edges [55]. The samples with smaller \"crystal sizes\" (as displayed in Table 1) should consequently have larger concentrations of Lewis sites with a low coordination number, and as a result, the basicity should be higher. Both Lewis acidic sites and Br\u00f8nsted acidic sites may interact with pyrrole, as shown by the adsorption of this molecule on catalysts in its H+ form and on components in its alkali cation form, however, only the framework oxygen atoms can connect with this molecule at basic sites [56]. Hence, we further studied the Pyridine-FTIR and Pyridine-DTG for the acidity analysis.However, it is noteworthy that Br\u00f8nsted and Lewis's acid sites detected upon Pyridine-FTIR spectra and Pyridine-DTG curves exhibited low acidity compared to Ti profile, probably due to the acidic \u2013OH on the catalyst surface. It can be seen that the band at 1542\u2009cm\u22121 for the Ti profile (\nFig. 7(a)) verified the existence of weak Br\u00f8nsted acidic sites by developing pyridinium ions which also proved the existence of surface hydroxyl clusters [57]. Meanwhile, the incorporation of Cu on the Ti by impregnation technique increases the number of basic sites, weakening the acid possessions of 3Ti3Cu and 4Ti3Cu catalysts (Fig. 7(a) and (b)). The impregnation procedure resulted in the insertion of the Cu into the Ti structure and the creation of tiny Cu particles, which affects the number of acid sites due to the contact interface with other oxides, according to XRD data (Fig. 3). This weak Br\u00f8nsted acidity hinges on the character of Cu phase species and influences plastic cracking reactions, including C\u2013N bond cleavage. Another possibility is that the collapse of Cu's layered structure decreased the total acidity, which originated from the interlayer protons [58]. Thus, it is anticipated that the basic molecule's characteristics, the type of edge, and Ti's presence would affect how the proton transfers.In a fixed-bed reactor setup, all of the reforming tests were conducted. Compared to other methods for producing H2 fuel, in-situ pyrolysis-catalytic steam reforming is more complicated, primarily due to many coke precursors and carbon in plastic and phenol. All catalysts were first applied to a continuous reaction. The conversions of phenol and the production of H2, Co and CO2 (in yield and mole percent) are shown in \nFig. 8. The properties of the pyrolysis products for the 3Ti3Cu catalyst are listed in Fig. S2. Based on these observations, the conversion of phenol could be attributed to the transformation of the phenol molecule into H2 formation via steam reforming reaction (Eq. 5). Low reforming performance was achieved under catalyst-free conditions (not shown). The H2 yield and mole percent of pyrolysis-catalytic steam reforming after adding the bare Ti catalyst was 56.8% and 68.4%, respectively. H2 production of the 4Ti3Cu catalyst was higher than the Ti sample, indicating that the Cu component played a key role in pyrolysis-catalytic steam reforming for H2 production. At the same time, by associating the catalytic performance of Ti, 4Ti3Cu, and 3Ti3Cu, it can be found that the phenol conversion for the bare Ti catalyst was 84.9%, which was lower than those of Cu -added catalysts. The reason behind this might be due to the detection of Ti-Cu alloy as shown in the XRD analysis. Ashish and Qiang [59] mentioned that the bimetallic alloys cause higher catalytic effectiveness than their monometallic complements, owing to strong interaction among the metals. Therefore, for the 4Ti3Cu catalyst, the phenol conversions increased significantly compared to the pure Ti. Meanwhile, 3Ti3Cu showed the highest phenol conversion (92.6%) and H2 yield (67.8%), indicating that higher Cu loading has a catalytic activity in the in-situ pyrolysis-catalytic steam reforming process. This heightened activity of reducible 3Ti3Cu catalyst could be because of the strong metal-support interaction, higher reducibility, strong basicity and higher amount of sites analyzed by H2-TPR and CO2-TPD, respectively. Additionally, the phenol conversion result indicates that the 3Ti3Cu catalyst with a smaller crystallite size quickly adsorbs and activates phenol molecules compared to the catalyst with a large crystallite size. It means that catalysts with large crystallite sizes (as shown in Table 1 from XRD analysis) are not preferred because they cause higher coke formation [60,61] and cause catalyst deactivation. The higher basic site results in increasing the reaction rate [62]. The results as mentioned above suggest that the strong basic sites should be catalytically active sites in this base-catalytic reaction. When Cu was added to the catalyst, more surface basic sites were created, which improved CO2 adsorption and carbonate synthesis. The carbonate was then hydrogenated by H2 that was adsorbed on and activated by the Ti metallic sites. Strong basic sites created by the introduction of Cu species activated the hydroxyl group in phenol, increasing phenol conversion and H2 selectivity. In the H2-TPR, the Cu2+ site was formed in the 3Ti3Cu catalyst due to the reducible characteristics of the Cu material. This resulted in a strong electron-donating property, a typical strong metal support interaction effect that was beneficial to the catalytic activity. Previous research [63,64] stated that strong metal support interaction effect in good decoration of metal on the support, which is favourable in catalytic performance. Since basicity was shown to grow throughout the 3Ti3Cu sample, it can be inferred from the data above that basicity and activity are directly related. In fact, significant basicity that correlated with the most active catalyst was detected at high Cu concentrations. Therefore, Cu addition modifies the catalyst's acid function, which is primarily responsible for the carbon production, and positively impacts activity and catalytic stability by preventing carbon deposition on the catalyst's surface. Fig. 14 displays the results of a TGA study of the utilized catalysts, further supporting this claim. We performed the influence of temperature and time on stream tests, as shown in \nFig. 9 and \nFig. 10, believing that the increase in stability is predicted given the combination of the chemical and physical features of the 3Ti3Cu nano-catalyst.The effects of reaction temperature on the catalytic performance of 3Ti3Cu are shown in Fig. 9. The H2, CO and CO2 yields, mole percent and phenol conversions were obtained after a 6\u2009h evaluation for each temperature (6 runs and each run around 1\u2009h) as a function of temperature. The slightly reverse water gas shift reaction (Eq. 7: CO2 +H2\u2192CO + H2O) was promoted by increasing the temperatures. The situation that CO contents after reforming also increased with increasing temperatures could be related to the slight promotion of endothermic reactions. However, CO and CO2 yield generally does not change appreciably compared to H2 yield within the whole temperature range. The enhancement of phenol steam reforming reaction (Eq. 6: C6H5OH+5\u2009H2O\u27f66CO+8\u2009H2) by temperature can be observed as phenol conversion, and H2 yield increased from 92.6% and 67.7% at 500\u2009oC to 99.7% and 97.90% at 800\u2009oC, respectively. This result suggests that the reaction process was dominated by phenol steam reforming reaction (PSR) (Eq. 6), which controls the final product distribution. Fig. 9 and Fig. 9 show that the reaction temperature of the 3Ti3Cu catalyst for the reforming of phenol may be reduced and that improved catalytic activity can be achieved, resulting in a decrease in the cost of catalyst manufacturing. In light of this, it is clear that phenol steam catalytic reforming for the generation of H2 is advantageous and that low-temperature catalytic reforming is feasible. Hence, this study confirms the feasibility of H2 generation from the in-situ pyrolysis-catalytic steam reforming reaction of PSW dissolved in phenol at temperatures less than 600\u2009\u00b0C, which is practically more sustainable and requires less energy.To further investigate the stability of the precious metal-free Ti-Cu nano-catalyst, the endurance experiments were conducted over the 3Ti3C catalyst at 500\u2009\u00b0C for 45\u2009h. The changes in phenol conversion and H2 yields and mole percent as a function of time on stream during the in-situ pyrolysis-catalytic steam reforming reaction of PSW dissolved in phenol are shown in Fig. 10. The catalyst shows a highly stable behavior for all time ranges in terms of mole percentage. The yields of CO and CO2 hardly changed over 3Ti3C, evidencing better stability. In contrast, H2 yield and phenol conversion decreased from 67.8% and 92.7% at the first hour to 59% and 89.9% after 25\u2009h time-on-stream. The rapid decrease of the conversion into the gas phase indicates the deactivation of the 3Ti3C catalyst; consequently, this pattern during a 25 run has prompted us to conduct a longer-term evaluation of the catalyst's resistance to the reaction. In spite of the severe reaction conditions and decrease in catalytic performance after 25\u2009h of time-on-stream, an improvement can be observed at 30\u201345\u2009h with a slightly decreased in H2 yield and phenol conversion, which may be related to an occurrence of slight deactivation of catalysts by carbon deposits [65,66]. The article will go into more detail on how the deactivation of the catalyst is related to the deposition of coke, the sintering and aggregation of active metal particles, and other factors.The main products analyzed from the prominent peaks in the GC/MS chromatogram are listed in Fig. S2. The chromatogram of GC\u2013MS technique for the 3Ti3Cu nano-catalyst confirmed that three value-added components were produced upon the pyrolysis reaction of PSW-phenol. Nevertheless, the mass spectrometer's limitations prevent it from detecting low molecular weight gases. The catalytic pyrolysis products were classified into aromatic compounds such as ethylamine and oxygenated aromatics such as tert-butyl hydroperoxide and benzene, (1,1-dimethylethoxy) (BDE). Tert-butyl hydroperoxide (TBHP), an alkyl hydroperoxide with a tert-butyl group, was found to be the major liquid result of the pyrolysis process of PSW-phenol. It is frequently employed in several oxidation processes. It functions as an oxidizing agent and an antibacterial agent. For example, to produce chain-elongated peroxides, Chuan et al. [67] described a practical Fe-catalyzed decarbonylative alkylation-peroxidation of alkenes using aliphatic aldehydes and TBHP. To produce \u03b1-ketoamides, Xiaobin and Lei [68] described a brand-new and effective TBHP/I2-promoted oxidative coupling process of acetophenones with amines. Therefore, in addition to a small amount of BDE, value-added components such as TBHP can also be found in the pyrolysis liquid product of PSW dissolved in phenol. The pyrolytic products were further analyzed by FTIR analysis, and the results are shown in \nFig. 11.The produced liquid component from the in-situ pyrolysis-catalytic steam reforming reaction of PSW dissolved in phenol using 3Ti3Cu nano-catalysts was also straight analyzed by FTIR to categorize the dominating functional clusters, and the results are shown in Fig. 11. FTIR can be used to analyze the chemical composition and the optical properties of the material and to better understand the release characteristics of liquid products after the reaction. All samples showed a broad band spectrum at 401, 424, 447, 864, 1049, 1165, 1296, 1620, and 2955\u2009cm\u22121 wavenumber and FTIR results with the peak intensities are shown in Fig. 13. Below 3000\u2009cm\u22121, one band is detected at 2955\u2009cm\u22121, which is attributed to aliphatic vibrations \u2212CH2\u2212 [69,70]. The absorption bands at 1620\u2009cm\u20131 correspond to the CC stretching mode of carbonyls and indicate the presence of compounds containing aromatic rings [71,72] and benzene (C\u2550C) in specific [73]. Also, aliphatic C\u2013O stretching was observed at 1296\u2009cm\u20131\n[74]. This peak is also ascribed to the detection of the aldehydes, alkanes, and ethers [75]. The band at 1165\u2009cm\u22121 was assigned to the stretching vibration of C\u2013O\u2013C [76]. In accordance with the previous research [77], C\u2013H stretch in methyl, methylene, and methyne groups can be confirmed at the spectral range around the band at 1049\u2009cm\u20131. Rainer et al. [78] stated that this peak could also be assigned to the C\u2013O\u2013C symmetric stretching in aliphatic groups and acid derivatives. In addition, there were four autocorrelated peaks at 864, 447, 424, and 401\u2009cm\u20131 that confirmed the formation of out-of-plane C\u2013H bending vibration peak for the aromatic ring [79], \u03bdCN stretching band (\u03bdCN) [80], \u03c4Ring (ring torsion) and \u03b3CN (out-of-plane bending or wagging) [81], and deformation modes [82] (which is connected with torsions and bending of benzene ring [83]), respectively. The C\u2013C group is hardly ever present throughout the pyrolysis process, indicating that the C\u2013C break is exceedingly challenging. Heating rates can impact the CO and C\u2013H groups.Several characterizations, including BET surface area, N2 adsorption-desorption isotherms and pore size distribution, TGA, DTG, CHNS, and TEM, were performed to investigate the deposited carbon and metal sintering over the used catalysts. An evaluation of the textural characteristics of fresh and spent catalysts (surface areas, pore volume, and average pore size) is presented in \nTable 2. N2 adsorption\u2013desorption isotherms and pore size distribution of spent catalysts are also illustrated in \nFig. 12. The monolayer-multilayer adsorption on the material's interior surfaces is responsible for the initial portion of the curve (at low P/Po). It is possible to explain the sharp rise in isotherm slope for high P/Po over 0.90 by capillary condensation inside the pores, followed by saturation when the pores get saturated with liquid [84,85]. All the spent catalysts present almost the same Type-IV isotherm and an H1 hysteresis loop to the Ti and 4Ti3Cu and double-hysteresis loops for the 3Ti3Cu. This means the mesoporous nature of the catalysts after the reaction is maintained, and the reactions did not substantially impact the main pore structures of the catalysts. The pore size distribution for spent Ti and 4Ti3Cu catalysts was positioned in the region of 5\u2009\u2212\u200915\u2009nm, while it was 17\u2009\u2212\u200935\u2009nm for the 3Ti3Cu catalyst. The decrease in specific surface area of Ti is large (> 70%) with the smallest diameter pores (20\u2009nm), which indicated the surface and pores of the Ti catalyst were blocked with ashes, carbon, residues or poison reactants. The weakened surface area of the Ti catalyst in comparison with the fresh catalyst might also be because of the collapsing of microchannels in the catalyst structure by the collision of energetic particles during the reaction procedure [86,87]. The surface area and total pore volume of the 4Ti3Cu and 3Ti3Cu catalysts increased after the reaction; probably, an acidic treatment by phenol compound occurred during the reaction, which resulted in the removal of blockage metal on pores. This opens up the pores in the used catalyst that were closed by dangerous metals, increasing the specific surface area of the used catalyst in comparison to the new catalyst. This would increase the access of reactant molecules to the active site of the catalyst within the pores and increase the activity. The better performance of the 3Ti3Cu catalyst could be due to their enhanced basicity strength and the stable catalytic performance compared to the Ti, and 4Ti3Cu catalysts could be attributed to their structural stability during the reforming reaction.One of the traditional techniques in FTIR spectroscopy has been around from the beginning and includes pressing or grinding a tiny quantity of a dry solid sample with powdered IR-transparent substances like halide salts, with KBr being the most popular [88]. \nFig. 13 shows the outcomes of using FTIR as a surface analysis approach to detect functional groups in the used catalyst, focusing on unsaturated hydrocarbons (CC linkages) and aromatics. All samples of spent catalyst were pelletized with KBr following the same procedure described in Section 2.2. The FTIR peak of the spent catalysts in 3757\u2009cm\u22121 fits to the surface-linked hydroxyl groups; this peak is steadily enhanced with higher Cu content, proving the catalysts have been regenerated. Also, the samples present a slender but distinct vibration near 2376\u2009cm\u22121, which is characteristic of the \u03bd3 elongation mode of linearly adsorbed CO2 on Cu2+. This peak can also be assigned to stretching vibrations of the reactants' secondary \u2013N\u2009=\u2009H\u2013 groups [89]. Due to the C\u2550O stretching of the carboxyl groups, the spent catalysts displayed distinct bands at 1628\u2009cm\u22121. As the intensities increase, more carboxyl groups are produced. The fact that this peak has a higher intensity indicates that Cu 's incorporation causes the creation of basic sites, which is compatible with its amphoteric capabilities. These bands could be related to species of H2 carbonates created when CO2 interacts with hydroxyl groups. A very weak sharp absorbance peak at 926\u2009cm\u22121 peaks could be ascribed to the asymmetric stretching vibrations of Ti\u2013O\u2013Ti bonds that increased in intensity by additional Cu. The absorbance peaks placed at 548\u2009cm\u20131 and 494\u2009cm\u22121 corresponded to the TiO2 lattice and signified the presence of functional groups like metals [90] and related to symmetric stretching vibrational modes [91], respectively. The 548\u2009cm\u20131 band might also be associated with O\u2013Ti\u2013O and Ti\u2013O\u2013Ti modes overlapping with O\u2013Cu and Cu\u2013O\u2013Ti bonds because it has high intensities for the, spent 3Ti3Cu and 4Ti3Cu catalysts. The two absorbance peaks at 525 and 656\u2009cm\u20131 for the bare Ti and 4Ti3Cu spent catalysts correspond to the bending of Ti\u2212O vibrations ( and could also be related to the detection of V\u2550O bonds [92]) and the bending of \u2013OH in the \u2013C\u2013OH group [93], respectively. In the 3Ti3Cu spent sample, two strong absorption peaks at 849\u2009cm\u22121 (probably related to the vibrations of the aromatic cycle [94]) and 710\u2009cm\u22121 (can be ascribed to rocking modes of \u2212(CH2)n\u2013 alkyl chains with n\u2009\u2265\u20094 [95]) were detected which have not appeared for the Ti and 4Ti3Cu catalysts. The bands at 849\u2009cm\u20131 are involved in Ti\u2550O stretching vibrations of orthorhombic Ti8O16 and could also represent the symmetric Ti\u2013O stretching \u03bd1. The band at 710\u2009cm\u22121 might also be originated from in-plane bending of the CO3\n2\u2212 group [96].Large organic compounds that contain C\u2212H or C\u2212C bonds might be trapped in the cages of acidic catalysts, which invariably results in catalyst deactivation under reaction circumstances. The in-situ pyrolysis-catalytic steam reforming process generates large organic compounds. Therefore, one crucial economic goal for the industrial use of these catalysts is to comprehend coke production. Herein, we analyzed the coke formation after reaction by TGA, DTG and CHNS techniques. Thermogravimetric analysis (TGA) is a thermal analytical technique extensively used to understand and measure the amount of coke deposited on the catalyst after the reaction. DTG gives information about the rate at which these coke are removed concerning time or temperature. The elemental analyzer (CHNS) is a technique used for quantitative determination in organic samples containing carbon and other basic elements of nature. The TGA and DTG curves of the spent catalysts are shown in \nFig. 14, and quantitative weight loss data (taken from TGA and CHNS) is shown in Table 2. The weight loss processes could be separated into three phases and mentioned by weight loss (WL). The vaporization of moisture caused the first phase below about 200\u2009\u00b0C, the second phase (WL2) between 200 and 600\u2009\u00b0C was attributed to the burning of deposited coke, and the third phase (WL3) above about 600\u2009\u00b0C belonged to the decomposition of remaining residues and heavy carbonaceous species. The in-situ pyrolysis-catalytic steam reforming reaction of PSW-phenol over Ti catalyst formed the highest amount of coke (8.2\u2009wt%) with 56.2% of the total weight loss. The most significant mass drop for the Ti spent sample was at 77\u2009oC, caused by the release of phenol and water molecules. The coke analysis results obtained for the Ti catalyst revealed that there is a significant interaction between Ti catalyst pore structure (pore size and shape) (as analyzed by pore size distribution in Fig. 2(b), XRD crystallinity in Fig. 3 and Table 1). The intensity of acid sites (Fig. 8) greatly affects Ti coke formation and deposition in converting PSW-phenol to H2. In the first stage, the weight loss for the 4Ti3Cu and 3Ti3Cu took place at the temperature of 91.7\u2009oC and was 7.6% and 6.6%, respectively. The DTG curve of the 4Ti3Cu catalyst showed a broad peak at 752\u2009\u00b0C, which was associated with light and heavy composites inside the 4Ti3Cu catalyst. The pyrolysis of phenol and the production of tiny derivatives from the fracture of PSW should yield light chemicals. Nevertheless, the bio-oil derivatives or the big aromatics made from the PSW structure were given credit for the heavier compounds. In both WL2 and WL3 phases, there was no coke development on the surface of the TGA curves of the Ti and 3Ti3Cu samples. The weight rise in this catalyst profile may have been caused by the oxidation of the metallic, active sites [10,23]. The DTG curve of the 3Ti3Cu sample shows almost straight like in all stages, indicating that the sample was stable in the entire temperature range. The lack of mass loss percentage in the experiment with 3Ti3Cu can be due to its low BET surface area (see Table 1) and slit-shaped pores structure. The slit-shaped pores structure has an advantage in contribution to the diffusion of oxygen during coke combustion [97]. This is perhaps due to the fact that the high content of Cu increases the basic site on the catalyst surface too much, which is not conducive to the phenol reforming reaction. The lowest weight loss for the 3Ti3Cu might also be due to the strongest metal-support interaction and strongest basic sites. The carbon content of the spent catalyst also follows the catalyst reducibility and basicity (Fig. 6) characterizations. The lowest carbon content has been taken place for the catalyst with the highest H2 consumption and CO2 uptake (Table 1). As a result of the intrinsic fundamental properties of 3Ti3Cu, which operate as a sponge for CO2 absorption, local gradients in gas concentration are generated, which promotes improved process efficiency. In fact, relative to rates of cracking side reactions, basic sites are thought to produce substantially higher rates of products and reagents desorbing from the catalyst surface [98]. The aforementioned characteristics of 3Ti3Cu made it an optimum catalyst for the in-situ pyrolysis-catalytic steam reforming reaction of PSW dissolved in phenol.Using the TEM method, we investigated the coke deposition over the employed catalysts in this study. The three forms of carbon that are deposited on the catalysts are amorphous (T\u2009\u2264\u2009570\u2009\u00b0C), filamentous (570\u2009\u00b0C < T\u2009<\u20091000\u2009\u00b0C), and graphitic (\u22651000\u2009\u00b0C). At the lowest temperature range of T570\u00b0C, amorphous carbon is produced, then filamentous carbon at T570\u00b0C to T1000\u00b0C, and lastly, graphitic carbon at T1000\u00b0C [99]. The morphology results of the spent 3Ti3Cu catalyst are depicted in \nFig. 15, which illustrates that the reaction produced carbon with random shapes and morphologies, but mainly filamentous carbon or carbon nanofibers (CNFs) on the catalyst surface were observed. The diameters of CNFs are in the range of 20\u201380\u2009nm, and pores have widened due to the pores filling. The 3Ti3Cu catalyst exhibits a similar tubular shape and sphere-like geometry with particle sizes of 20\u201350\u2009nm. The catalyst revealed the existence of Cu particles on the surface, which were corroborated by well-dispersed, small particle TEM images. In contrast, the grey forms on the surface are associated with TiO2 components. The CNFs and filamentous coke did not block catalyst particles, although its progressive deposition may hinder the contact between reactants and catalyst active sites. Additionally, due to the strong Cu\u2013Ti interaction, the Cu was not removed by the growing CNFs, and very low fragmented particles were observed. A catalyst may get fragmented during the in-situ pyrolysis-catalytic steam reforming reaction process as a result of internal catalyst reactions where carbons are first created and then developed. Smaller metal particle sizes likely helped to generate the thinner CNFs, and the thinner CNFs had greater surface energies, making them less stable. This makes it very simple to remove CNFs made with smaller Cu particles, which may be why the 3Ti3Cu catalyst exhibits the least carbon deposition in TGA. The figure clearly illustrates a nearly perfect core\u2013shell morphology was obtained. The different shell thicknesses with spherical cores with around 30\u2009nm diameters result from the reaction. This reaction produced a core shell-like structure with an approximate shell thickness of 8.2\u2009nm. The key benefits of core-shell nanoparticles are that the core material's qualities, such as reactivity, may be reduced, and its thermal stability can be altered, increasing the core particle's overall stability and dispersibility. The shell materials, which can provide surface chemistry for further modifying and functionalizing the nanoparticles, are another advantage of core-shell structures. In conclusion, the in-situ pyrolysis-catalytic steam reforming process of PSW dissolved in phenol may be sustained by catalysts with increased basicity, and the catalyst can prevent complex carbon deposition.To summarize, nano-sized precious metal-free Ti-Cu catalysts were successfully synthesized by direct hydrothermal conditions and impregnation method and innovatively applied in the in-situ pyrolysis-catalytic steam reforming reaction of PSW liquefied in phenol. This study is expected to offer an essential research reference for designing precious metal-free Ti-Cu and producing H2 and valuable liquid fuel from the abovementioned reaction. As-prepared samples were analyzed by XRD, BET, CO2-TPD, pyrrole-DTG, pyrrole-FTIR, pyridine-FTIR, pyridine-DTG, TEM, H2-TPR, FTIR-KBr and the used samples were characterized by TGA-DTG, FTIR-KBr, BET, TEM and CHNS. The liquid product was also analyzed by GC/MS and FTIR systems. The morphology study suggested that Ti-Cu ensembles may be formed due to the interaction between Ti and Cu that lead to the structure of CuTi3 alloy, as confirmed by XRD analysis. Samples with higher Cu contents cause to decrease in the surface area, crystal sizes, partial collapse of ordered mesoporous structure and the growth of the average pore diameter. However, Cu results in higher reducibility, metal support interaction and basic sites of the catalysts with highly dispersed Cu2+ species in 4Ti3Cu and 3Ti3Cu samples. Cu2+ insertion into TiO2 changes the distribution of basic sites as well as the total basicity value. The 3Ti3Cu performed the best catalysis performance, highest phenol conversion, H2 production and lowest coke formation due to strong metal-support interaction, higher reducibility, strong basicity and higher amount of sites analyzed by H2-TPR and CO2-TPD, respectively. The H2 selectivity rose, and the coke quantity dropped from 4Ti3Cu and 3Ti3Cu due to decreased acid sites and increased basic sites, providing increasingly continuous catalytic activity. The catalytic pyrolysis products were classified into aromatic compounds, such as ethylamine, and oxygenated aromatics, such as tert-butyl hydroperoxide and benzene, (1,1-dimethylethoxy). These results may help in the development of more effective solid base catalysts for a variety of base-catalyzed processes.W. Nabgan First author, contents development, writing. H. Alqaraghuli Experimental section. B. Nabgan Writing, editing. T.A. Tuan Abdullah Supervision, English editing. M. Ikram Writing (Section 2), editing. F. Medina Comments and editing. R. Djellabi Writing, Revision, response to reviewer, proofreading and English editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The principal author, Walid Nabgan, is thankful for the support from Universitat Rovira i Virgili under the Maria Zambrano Programme (Reference number: 2021URV-MZ-10), Proyectos de Generaci\u00f3n de Conocimiento AEI/MCIN (PID2021-123665OB-I00), and the project reference number of TED2021\u2013129343B-I00. The authors are also grateful for the support given by Universiti Teknologi Malaysia (UTM) allocation budget in Pusat Pengurusan Makmal Universiti (PPMU) laboratory.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.apcatb.2022.122279.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  In-situ pyrolysis-catalytic steam reforming reaction of polystyrene waste (PSW) liquefied in phenol can generate hydrogen from phenol and valuable liquid fuel from the PSW and thus has been studied recently. However, due to the complexity of phenol compounds and plastic waste, this reaction suffers from high energy consumption and coking. Herein, Ti, 4Ti3Cu and 3Ti3Cu nano-catalysts were facilitatively prepared using hydrothermal and impregnation methods and the physical and chemical properties of the fresh and used samples were deeply characterized. The experimental results show that almost complete phenol conversion with 97.90% H2 yield was achieved at 800\u00a0oC using 3Ti3Cu nano-catalyst. The catalytic pyrolysis products were ethylamine, tert-butyl hydroperoxide (TBHP) and benzene (1,1-dimethylethoxy) (BDE). The correspondence of the preparation, morphology and catalytic activity in this research elucidates the synthesis of anti-coking and stable nano-catalysts for in-situ pyrolysis-catalytic steam reforming reaction.\n               "}
{"full_text": "The authors declare that the data supporting the findings of this study are available within the article and the supplemental information. The data and results supporting the present study are available from the lead contact upon request.Because of the rapid expansion of the world\u2019s population and industrialization, which cause rapid increases in the demand for energy, new energy sources are desired that should be of great abundance and have less of an impact on the environment.\n1\n\n,\n\n2\n\n,\n\n3\n As a green, promising protocol, energy conversion via the electrochemical route is attracting more and more attention.\n4\n\n,\n\n5\n\n,\n\n6\n Electrocatalysts that are high performing and cost economic are critical for large-scale implementation of various electrochemical processes like water splitting and energy transformation in fuel cells. For the latter, unfortunately, the sluggish kinetics of the oxygen reduction reaction (ORR) on the cathode are a major limiting factor for the electrocatalytic cycles.\n7\n Up until now, Pt-based catalysts have proven to be the best choice in mediating the ORR process, while use of noble metal prevents these catalytic systems from industrial applications.\n8\n\n,\n\n9\n Thus, great effort has been made in development of low-Pt-based or non-precious metal electrocatalysts in efficient ORR processes.Transition metal-based M-N-C catalysts possess significant ORR activities with their definite coordination structures for active sites and highly adjustable electronic states and are regarded as the most promising alternatives to Pt-based precious-metal catalysts.\n10\n\n,\n\n11\n Among them, 3d transition metal-based catalysts, such as Fe, Co, Ni, and Cu, as active sites, hosted by carbon frameworks represent the frontier design of electrocatalysts.\n12\n Active metal sites with unoccupied 3d orbitals efficiently interact with oxygen intermediates for ultrahigh intrinsic activity, and\u00a0conductive and porous carbon skeletons promote electronic conduction.\n13\n However, the formation of two-electron reaction products with lower efficiency as well as the fast performance degradation because of leaching of 3d transition metals have not been completely resolved.\n14\n In contrast, M-N-C catalysts based on 4d and 5d transition metal centers exhibit better long-term stability because of the stable arrangement of outer shell electrons, while their unsatisfactory sluggish ORR kinetics activity is caused by the coordination and electronic structure of the active metal site.\n15\n\nAs a 5d Group 6 transition metal with a valency between \u22122 and\u00a0+6, tungsten shows on the left side of the volcano plot, possessing strong oxygen binding energy.\n5\n\n,\n\n16\n Hence, tungsten (W)\u00a0affords unusual performance in the ORR because the O and OH species are difficult to desorb from the W center.\n17\n In fact, W-N-centered single-atom catalysts (SACs) have been reported previously to be promising in the ORR. For example, Chen et\u00a0al.\n18\n reported that the\u00a0highest\u00a0ORR catalytic activity could be obtained by a coordination number 5 when\u00a0W coordinates only with N. Bisen et\u00a0al.\n19\n prepared an efficient and durable ORR electrocatalyst with a WN2C2 center, which adsorbs oxygen and the reaction intermediates with moderate binding energies. In addition, Jiang et\u00a0al.\n17\n also reported a W-N-C catalyst prepared via pyrolysis of W-doped ZIF-8 material. At present, the development of W-SACs regarding the ORR faces two challenges: (1) precise regulation of the coordination environments for atomic W and (2) substantial improvement of the catalytic activity relative to the Pt-based catalysts.\n20\n\nOn the other hand, transition metal nitrides (TMNs) may serve as alternative functional materials in energy storage and electrocatalysis because of their excellent physicochemical properties.\n21\n\n,\n\n22\n Wang et\u00a0al.\n23\n reported a facile and in situ nitriding method to obtain a unique W nitride cluster loaded on a 2D conductive g-C3N4 material (WN@g-C3N4-750); here, the cubic-phase WN and the small clusters over the g-C3N4 layer are responsible for the excellent ORR performance of the catalyst. Indeed, the electronic structure of W atoms can be tuned by nitrogen atoms, and the catalytic activity be promoted by increasing the number of W-N bonds.\n24\n\n,\n\n25\n In spite of this, most TMNs show invalid catalytic activity because of their poor intrinsic activity and low density of active sites.\n22\n The intrinsic activities of nitrides mainly rely on their crystalline phase, which is related to the crystal orientation and the particle size.\n26\n\n,\n\n27\n\nHere, we report an elaborately designed ORR catalyst with a concerted W-N single-atom site and WN nanoclusters. By complexation with phthalonitrile, W atoms are first anchored in a phthalocyanine-typed material (WPc). g-C3N4 synthesized from melamine, the most common raw material, was mixed and pyrolyzed at 700\u00b0C. The W ions are first chelated by phthalonitrile and then anchored onto g-C3N4 support after further pyrolysis destroys the WPc structures; consequently, the W-N single-atom sites and WN nanoclusters are afforded and anchored on the 2D N-doped carbon materials. Such a material exhibits performance comparable with commercial Pt catalysts while benefitting from low cost and high durability. The origins of the excellent performance of the W-N catalysts are discussed.A large number of MN4-macrocyclic compounds, such as metallo-porphyrin, metallo-corrole, and metallo-phthalocyanine complexes, have been used in electrocatalysis.\n28\n\n,\n\n29\n Transition-metal phthalocyanine complexes (MPcs) possess a highly conjugated macro-cyclic planar structure that exhibits excellent chemical stability in electrolysis.\n30\n\n,\n\n31\n Further, the delocalized \u03c0 electrons over the entire macrocycle participate in MPc-mediated oxidation and reduction, exhibiting unique activity.\n31\n Thus, The MPc species may serve not only as excellent catalysts but also as raw materials to prepare structurally well-defined W-SACs. With a defective 2D structure, g-C3N4 is widely used as a photocatalyst and in energy conversion devices.\n32\n The six-fold cavities of g-C3N4 can not only anchor single atoms or clusters but also metal complexes.\n33\n\n,\n\n34\n\n,\n\n35\n Further, although pure g-C3N4 is of low conductivity, a certain degree of decomposition of g-C3N4 at suitable temperatures produces in situ amorphous, graphite-like structures, thus enhancing its conductivity. Our conception originates from the combination of the characteristics of WPc and g-C3N4.First, WCl6, phthalonitrile, DBU, and 1-pentanol were added to a hydrothermal kettle along with nitrogen at 0.5 MPa and heated to 220\u00b0C for 4\u00a0h (Figure\u00a01\nA), yielding a dark-green dyestuff substance. After precipitation overnight, green W species (WPc) were obtained via filtration and washing with methanol. The control sample without metal W (denoted NPc) was also synthesized by an identical procedure as WPc, without adding W. As shown in Figures\u00a01B and 1C, the Fourier transform infrared (FTIR) spectrum of this substance shows a peak at 950.3\u00a0cm\u22121, indicative of a W-N bond,\n36\n whereas the X-ray diffraction (XRD) pattern of WPc is rather different from that of NPc. The W 4f spectrum (Figure\u00a01D) in the X-ray photoelectron spectroscopy (XPS) shows that W resides in an unsaturated coordination environment, pointing to the presence of WPc. The g-C3N4 was prepared by heating melamine to 550\u00b0C and holding it for 4 h. WPc and g-C3N4 were mixed thoroughly in methanol at a certain ratio for 4\u00a0h to form a light-green homogeneous W-nitrogen precursor (WPc/gC3N4). The mixture was next placed in a tube furnace and heated under the protection of argon to obtain the final product, named WNPc/gC3N4-T (T is the pyrolysis temperature).The XRD patterns of WNPc/gC3N4-T (T\u00a0= 600, 700, and 800) and other relevant reference samples are shown in Figure\u00a02\nA. The pyrolysis precursors WPc/gC3N4 and WNPc/gC3N4-600 have two distinct diffraction peaks with angles of 13.7\u00b0 and 27.5\u00b0, respectively. The diffraction peak at 13.7\u00b0 is caused by repetition of the 3-s-triazine structure in the graphite-phase carbon nitride structure, while the sharp peak at 27.5\u00b0 is caused by stacking of the (002) crystal plane in the graphite-phase carbon nitride conjugate plane without W. It shows that the W species is uniformly distributed on the g-C3N4 carrier, and aggregation does not occur. In addition, the main structure of W2NPc/gC3N4-600 is proven stable at 600\u00b0C despite the mass loss, as indicated by the thermogravimetric (TG) analysis (Figure\u00a02B). With the continuous increase in temperature, g-C3N4 was severely decomposed and graphitized; meanwhile, the so-released carbon- and nitrogen-containing fragments interact with the WPc raw material to form a brand-new structure. WNPc/gC3N4-700 shows distinct peaks arising from graphitic carbon compared with WPc/gC3N4 and WNPc/gC3N4-600. It is worth noting that faint peaks appeared around 36.76\u00b0, 62.42\u00b0, and75.22\u00b0, closely following the WN (PDF#75-1012). Interestingly, these peaks are lower than the standard pattern of WN (PDF#75-1012) because of the extreme distortion of the atomic clusters and lattices via strong interaction with the carriers.\n37\n Another small peak at 39.62\u00b0 may belong to the W2C (PDF#35-0776), corresponding to a rather low content.\n38\n The loading content of W in WNPc/gC3N4-700 is around 4.22 wt %, as analyzed by inductively coupled plasma mass spectrometry (ICP-MS). Part of the formation of W carbide is inevitable because of the possibility of reaction between W and carbon atoms at high temperature. Data on appropriate precursor mixing ratios are presented in Table\u00a0S1 and Figure\u00a0S1, which is crucial to get the perfect WNPc/gC3N4-700. Different mixing ratios of WPc and g-C3N4 affect the W content of the product after high-temperature pyrolysis. Only at a specific ratio can the best WNPc/gC3N4-700 be obtained. Because of the high mass loss in the pyrolysis steps, too low or too high a mixing ratio will result in high W content in the final product, leading to aggregation of W elements, subsequently affecting catalytic performance. Ammonia treatment can further modify the coordination environment of W\n18\n and facilitate formation of WNs, as indicated by comparison of the spectra of WNPc/gC3N4-NH3-700 and WNPc/gC3N4-700. When the temperature continues to rise, the peaks around 31.62\u00b0, 35.78\u00b0, 48.5\u00b0, 64.38\u00b0, 65.88\u00b0, 73.74\u00b0, 76.12\u00b0, and 77.44\u00b0 in WNPc/gC3N4-800 indicate that the W atoms are completely transformed into WN (PDF#25-1256). In the case of WPc-700 obtained via direct pyrolysis of WPc, there were more diffraction peaks, indicating that addition of g-C3N4 facilitates the dispersion of WNPc/gC3N4-700 samples that were prepared by direct heating of g-C3N4 and phthalocyanine without W, and only graphitized peaks at 26\u00b0 and 44\u00b0 were observed in XRD spectra.The double D band (1,342\u00a0cm\u22121) and G band (1,599\u00a0cm\u22121) in Raman spectra of WNPc/gC3N4-700 and NPc/gC3N4-700 (Figure\u00a02C) confirm partial graphitization of the carbon support, and the intensity ratio (ID/IG\u00a0= 1.08 versus 1.13) also indicates a higher defect level.The nitrogen adsorption-desorption isotherms of WNPc/gC3N4-700 and NPc/gC3N4-700 are depicted in Figures\u00a02D and 2E. The specific surface areas, calculated using the BET method, are 177.173 and 455.672 m2/g, respectively. They exhibit similar type IV isotherms and a slightly different pore size. This suggests that the higher temperature and metal phthalocyanine lead to shrinkage or disappearance of the pore of g-C3N4.\n32\n\n,\n\n37\n\n,\n\n39\n\nXPS studies were carried out to analyze the surface chemical composition and elemental valence states. The peaks belonging to the different orbitals of C, O, N, and W can be distinguished from each other (Figure\u00a03\nA). WNPc/gC3N4-700 shows a high ratio of N originated from the raw materials; O comes from the residual alcohol solvent, and the presence of air or free oxygen in the instrument is also responsible for the predominance of C and O elements in the sample.\n23\n The C 1s spectrum of WNPc/gC3N4-700 (Figure\u00a03C) reveals sharp peaks corresponding to 283.60 eV (W-C), 284.8 eV (C-C), 285.82 eV (C\u2212N), 287.57 eV (C=N), and 289.60 eV(C=O), indicating successful doping of N in the sample and generation of W carbide.\n19\n\n,\n\n40\n The N 1s spectrum (Figure\u00a03D) can be divided into four peaks located at 398.34 eV (pyridinic N), 399.49 eV (W-N), 400.73 eV (graphitic N), and 403.23 eV (oxidized N), respectively. Graphitic N can greatly increase the limiting current density, while pyridinic N might convert the ORR mechanism from 2e\u2212- to 4e\u2212-dominated processes.\n41\n The difference between WNPc/gC3N4-700 and NPc/gC3N4-700 indicates a binding energy of 399.49 eV, which is attributed to W-N coordination.\n23\n\n,\n\n42\n It is widely accepted that the pyridinic N and pyrrolic N moieties are prone\u00a0to coordinate with W to form WN\nx\n complexes.\n19\n The W 4f spectrum of WNPc/gC3N4-700 (Figure\u00a03B) features four pronounced peaks: the peaks at 32.4 eV and 34.47 eV are mainly indicative of W-C bonding and attributed to W carbide impurities.\n43\n By contrast, the peak aera of the W-N bond is greater than that of W-C (86.54% versus 13.46%), while the peaks at 35.57 eV and 37.64 eV are responsible for the W-N site. In addition, the W-N bonding in WNPc/gC3N4-700 is slightly lower than that in WPc, indicative of more unpaired electrons on coordinatively unsaturated sites.\n18\n\nFurther, the structure and morphology of WNPc/gC3N4-700 were investigated by scanning electron microscopy (SEM). The morphology of the sample shown in Figure\u00a04\nC is a graphene-like substance after partial pyrolysis at high temperature, and the large number of sheet-like folds shows that pyrolysis results in deformation and shrinkage of the surface.\n32\n\n,\n\n37\n Atomic-resolution high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) was used to observe the samples. A HAADF-STEM image (Figures\u00a04D\u20134F) at higher magnification shows a large number of bright atomic dots and gathered light spots, indicating the presence of single atoms (yellow circle) and nanoclusters (red circle, diameter size in the range of 0.52\u20131.42\u00a0nm). The selected area electron diffraction (SAED) pattern proves the absence of a crystal phase. Well-dispersed W-N single atom and cluster sites are thus justified. We also characterized WNPc/gC3N4-NH3-700 that was obtained via ammonia treatment at high temperature (Figures\u00a04G and 4H); the dispersed W started to agglomerate, showing a thin WN-related crystal, as indicated by the XRD pattern.X-ray absorption near-edge structure (XANES) and X-ray absorption fine structure (XAFS) spectroscopy were performed to study the electronic structure and local chemical configuration of the WNPc/gC3N4-700 sample. The XANES spectra of WNPc/gC3N4-700 were different from those of WPc, WO3, WC, and W powder (Figure\u00a04A). Furthermore, the Fourier-transformed (FT) k3-weighted \u03c7(k)-function of the EXAFS spectra in R-space suggested that the W species were W-N and W-N-W in WNPc/gC3N4-700 (Figure\u00a04B). Among them, the peak at nearly 1.78\u00a0\u00c5 is attributed to the W-N in W single-atom sites (W-SACs) and WN nanoclusters (WN-NPs), whereas the peak at nearly 2.67\u00a0\u00c5 suggested W-N-W in WN-NPs. Least-squares curve fitting was conducted to acquire quantitative structural parameters around W atoms (Figure\u00a0S2); the corresponding fitting results are shown in Table\u00a0S2. XAFS spectra of WNPc/gC3N4-700 with different W content are shown in Figure\u00a0S3. These characterizations prove that the W atoms exist as mononuclear and multinuclear centers.Cyclic voltammetry (CV) measurements for 20% Pt/C and WNPc/gC3N4-700 in alkaline electrolytes were performed to clarify the catalytic performance, as illustrated in Figure\u00a05\nA. Both electrodes display an ORR peak around 0.80\u00a0V in 0.1\u00a0M KOH solution, indicating high ORR activity. The ORR performance of different samples was evaluated systematically in O2-saturated 0.1\u00a0M KOH at 1,600\u00a0rpm and a scan rate of 10\u00a0mV/s using rotation disk electrode (RDE) measurement. As shown in Figure\u00a05B, the pyrolysis temperature is an influential factor, and the activity of WNPc/gC3N4-700 is better than that of WNPc/gC3N4 (T\u00a0= 600\u00b0C, 800\u00b0C). Notably, WNPc/gC3N4-700 exhibits extraordinary catalytic activity: the half-wave potential E1/2 (0.835V) is comparable with that of 20% Pt/C, while the mass diffusion-limited current (0.596 mA/cm2) is even better than that of 20% Pt/C that reaches the upper limit in alkaline solutions. At the same temperature, the activity of WPc-700 is much lower than that of WNPc/gC3N4-700, indicating the important role of the g-C3N4 carrier. Compared with previous reports on ORR processes mediated by W-SAC or WNs, the synergy of atomic W and small WN clusters is most likely responsible for the rather high activity of WNPc/gC3N4-700 toward the ORR. The reduced activity of the WNPc/gC3N4-NH3-700 sample can be attributed to the loss of single-atom sites and small WN clusters. The reason why it still has relatively good catalytic activity is that there still exist a small number of monatomic active sites in the catalyst, and the crystal-phase WN also has certain activity.\n23\n The double-layer capacitance (Cdl) evaluated by CV at different scan rates of 20\u2013100\u00a0mV/s in the non-faradic region is followed by the respective electrochemical active surface area (ECSA) and roughness factor (Rf), and it turned out that the combination of WPc and g-C3N4 at high temperature favors generation of active areas (Figure\u00a0S4; Table\u00a0S3). The similar ECSA before and after NH3 treatment also indicated that differences in catalytic activity arise from changes in the active site rather than the active surface area.Usually, the low Tafel slope means that the rate-limiting step following the first electron transfer and easy-to-achieve high current at low overpotential are beneficial for the ORR.\n18\n\n,\n\n44\n Interestingly, as shown in Figure\u00a05C, the Tafel slope of WNPc/gC3N4-700 is 40.47\u00a0mV dec\u22121, which is lower than that of other samples and is almost half that of 20% Pt/C (75.52\u00a0mV/dec), confirming the O2 adsorption/desorption step as the kinetically fastest process in the ORR at WNPc/gC3N4-700.\n45\n\nRotating ring disk electrode (RRDE) measurement was conducted to investigate the ORR mechanism of the samples (Figures\u00a05E and S5); this method has been used to evaluate the electron transfer and selectivity in electrochemical reactions.\n18\n The electron transfer number of WNPc/gC3N4-700 reaches 3.90 at 0.25\u20130.8\u00a0V with a yield of peroxide species (HO2\n\u2212) below 5%, only slightly lower than that of Pt/C (3.90\u20133.95) at the same range; this indicates a dominant 4-electron process with high energy conversion efficiency. The electron transfer number of WNPc/gC3N4-700 was calculated by the K-L equation, being 3.92 at 0.3\u20130.7\u00a0V (Figure\u00a0S6), suggesting that the 4-electron transfer process dominates the electrochemical reaction.Durability and stability are also important metrics for a catalyst. Figure\u00a05F shows the ORR polarization curves of the best-performing electrocatalyst WNPc/gC3N4-700 and commercial Pt/C before and after 5,000-cycle voltammograms. There is an almost similar negative shift of polarization curves. On the other hand, chronoamperometry measurement was adopted to test the durability of WNPc/gC3N4-700. As shown in Figure\u00a05G, the WNPc/gC3N4-700 electrode exhibits better stability, with only 12.52% attenuation of activity compared with 42.25% current decay of 20% Pt/C after 60,000 s. Given the methanol tolerance problems of commercial Pt/C, WNPc/gC3N4-700 and 20% Pt/C were subjected to a methanol tolerance test. As shown in Figure\u00a05H, in contrast to the sharp current density decay for the Pt/C electrode, the current density in the experiments with WNPc/gC3N4-700 is unaffected by methanol injection, indicating that WNPc/gC3N4-700 shows good methanol tolerance in the ORR process. Thus, WNPc/gC3N4-700 outperforms 20% Pt/C in terms of stability and durability.Furthermore, the ORR activities of these catalysts are also evaluated in acidic electrolyte (0.1\u00a0M HClO4) at a rotation speed of 1,600\u00a0rpm. As shown in Figure\u00a05D, the performance of WNPc/gC3N4-700 is close to that of 20% Pt/C. It is noteworthy that the difference in polarization curves between WNPc/gC3N4-NH3-700 and WNPc/gC3N4-700 under acidic compared with basic conditions is much higher, further illustrating the importance of W single-atom sites and small WN clusters.Next, to justify the origins of the excellent performance of WNPc/gC3N4-700, the reaction energy profiles and the associated electronic structures were interrogated with theoretical calculations, as displayed in Figures\u00a06\n, S7, and S8. Based on the characterization results, two models were proposed as the possible reaction center: the W single-atom sites with pyrrolic-N2 coordination (W-SAC) and a WN particle size of \u223c1nm (WN-NP). As shown in Figure\u00a06, the ORR process may undergo six coordinates: (1) \n\n4\n\n\n(\nH\n\n+\n\n+\n\ne\n\u2212\n\n)\n+\n\nO\n2\n\n+\n\u2217\n\n, (2) \n\n4\n\n\n(\nH\n\n+\n\n+\n\ne\n\u2212\n\n)\n+\n\n\nO\n2\n\n\n\n\u2217\n\n\n, (3) \n\n3\n\n\n(\nH\n\n+\n\n+\n\ne\n\u2212\n\n)\n+\n\n\nO\nO\nH\n\n\n\n\u2217\n\n\n, (4) \n\n2\n\n\n(\nH\n\n+\n\n+\n\ne\n\u2212\n\n)\n+\n\n\nO\n+\n\nH\n2\n\nO\n\n\n\n\u2217\n\n\n, (5) \n\n\n\n(\nH\n\n+\n\n+\n\ne\n\u2212\n\n)\n+\n\n\nO\nH\n+\n\nH\n2\n\nO\n\n\n\n\u2217\n\n\n, and (6) \n\n\u2217\n+\n\n\n2\nH\n\n2\n\nO\n\n. In general, W-SAC and WN-NP matter for different steps of the ORR process. According to the computational results, WN-NP outperforms W-SAC on the first two steps; i.e., adsorption of O2 and initial formation of the O\u2013H bond. For example, at the WN-NP site, the adsorption energy of O2 is much higher, while the barrier for generating \u2217OOH is much lower. A reversal takes place when coming to the last two steps: in formation of \u2217OH from \u2217O, the associated barrier at W-SAC is lower compared with the one at WN-NP, but the difference is tiny; finally, the reduction of \u2217OH to release water at W-SAC is energetically more favorable. Considering that the step \n\n\n\nO\nH\n+\n\ne\n\u2212\n\n\u2192\n\n\nO\nH\n\n\u2212\n\n\n\n\n\u2217\n\n\n is barrierless and reversible, WN-NP most likely facilitates the reduction of O2 to \u2217OH, and W-SAC accelerates the final water release step.In summary, a polymorphic W system has been elaborately constructed using WPcs and g-C3N4 as the precursors under accurately controlled temperature. The so-prepared W-N catalyst performs encouragingly in catalyzing the ORR process. The\u00a0WNPc/gC3N4-700 species exhibits efficiency comparable with the commercial Pt/C catalyst but is much more durable. This excellent performance has been attributed to the synergy of the W-SACs) and WN-NPs). Our work may point out an interesting possibility to design and prepare active, stable, non-noble-metal ORR catalysts. Further exploration is indicated to fine tune the SAC:NP ratio for commercial and industrial availability.Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Shaodong Zhou (szhou@zju.edu.cn).This study did not generate new unique materials.In a typical synthesis, W chloride (1.0 g), phthalonitrile (1.29 g), and 1.0\u00a0mL of 1,8-diazabicyclo(5.4.0)undec-7-ene were dissolved in 20\u00a0mL of 1-pentanol in a hydrothermal kettle (the volume was 100\u00a0mL). Before the reaction, nitrogen was introduced into the hydrothermal kettle to replace the air, and the nitrogen pressure in the kettle was controlled at 0.5 MPa. Then, the mixture was slowly heated for 50\u201360\u00a0min from room temperature to 220\u00b0C, and the temperature was then held for a further 180\u00a0min. The mixture of products was then cooled, transferred to a large flask with added excess methanol, and stayed overnight. The powder precursor was separated by suction filtration and washed several times with methanol until the color of the filtrate turned light green. Finally, the product was dried into a green powder at 60\u00b0C.In a typical synthesis, melamine (10 g) was heat-treated at 550\u00b0C with a heating rate of 3\u00b0C under flowing Ar gas for 240\u00a0min to obtain the g-C3N4.In a typical synthesis, 0.03\u00a0g WPc and 0.65\u00a0g g-C3N4 are mixed in methanol solution at 60\u00b0C for 4 h, evaporated, and dried at 70\u00b0C.The WPc/gC3N4 precursor was pyrolyzed at T (T\u00a0= 600\u00b0C, 700\u00b0C, 800\u00b0C) with a heating rate of 6\u00b0C under flowing Ar gas for 120\u00a0min to obtain the WNPc/gC3N4-600, WNPc/gC3N4-700, and WNPc/gC3N4-800.The chemical information and preparation of other samples are shown in the supplemental experimental procedures.XRD patterns were conducted on a Rigaku Ultima IV using nickel-filtered Cu K\u03b1 radiation of wavelength 1.5406\u00a0\u00c5 with a scanning angle (2\u03b8) of \u223c10\u00b0\u201380\u00b0, operated at 40\u00a0kV and 40 mA.XPS measurements were carried out on a Thermo Scientific K-Alpha equipped with an Al anode (Al K\u03b1\u00a0= 1,486.6 eV), operated at 12\u00a0kV and 6 mA. Energy calibration was carried out using the C1s peak of adventitious C at 284.80 eV.The Raman spectra were recorded on a Horiba HRE Volution with an excitation laser of 532\u00a0nm.FTIR measurements were carried out by Thermo Scientific Nicolet iS20, and the scans (number of scans, 32) were collected with a resolution of 4\u00a0cm\u22121 from 4,000 to 400\u00a0cm\u22121.SEM measurements were performed with Sigma300 field-emission scanning electron microscope.The weight ratios of W elements were measured by ICP-MS measurements on an Agilent 7700(MS).TG was performed on a TA-Q500 at a temperature range from 50\u00b0C\u2013850 \u00b0C at a heating rate of 10\u00b0C min\u22121.BET surface area analysis was performed using an AUTOSORB-IQ2-MP BET surface analyzer.The HAADF-STEM images were obtained in FEI Titan G2 80-200 ChemiSTEM operated\u00a0at an acceleration voltage of 200\u00a0kV, which was equipped with a high-brightness field-emission gun (X-FEG), double spherical aberration corrector, and monochromator.Electrochemical characterization for linear scan voltammetry was performed on a CHI760D electrochemical station (Shanghai Chenhua, China) using a standard three-electrode system in 0.1\u00a0M KOH/HClO4 solution. Typically, the work electrode was a glassy-carbon rotating disk electrode (RDE; 0.196\u00a0cm2) and a glassy-carbon RRDE (0.247\u00a0cm2) with a Pt ring, and the counter and reference electrodes were graphite rod and Ag/AgCl electrodes (3.5\u00a0M KCl solution), respectively.A dispersion including 3\u00a0mg non-precious catalysts, 1.5\u00a0mg carbon black (Super-P), 0.490\u00a0mL H2O, 0.490\u00a0mL ethanol, and 0.02\u00a0mL 5 wt % Nafion was sonicated for 30\u00a0min to get a homogeneous ink. Then the prepared catalyst ink with a loading of 0.3\u00a0mg cm\u22122 was transferred to the glassy carbon electrode and dried naturally. The RDE polarization curves were recorded with a scan rate of 10\u00a0mV/s at 1,600\u00a0rpm in O2-saturated 0.1\u00a0M KOH/HClO4 electrolyte. The net current was calculated by subtracting the background capacitance measured in Ar-saturated solution. All potentials were recorded with iR compensation and then referred to the RHE (\n\n\nE\n\nR\nH\nE\n\n\n=\n\nE\n\nA\ng\n/\nA\ng\nC\nl\n\n\n+\n0.059\n\u2217\np\nH\n+\n0.2046\nV\n\n). The electron transfer number (n)\u00a0and hydrogen peroxide yield (H2O2%) can be calculated by the following equations, where N is the collection efficiency of the Pt ring (N\u00a0= 0.37), and Id and Ir are the disk and ring current, respectively:\n\n(Equation\u00a01)\n\n\nn\n=\n4\n\u00d7\n\n\nI\nd\n\n\n\nI\nd\n\n+\n\n\nI\nr\n\nN\n\n\n\n\n\n\n\n\n\n(Equation\u00a02)\n\n\n\nH\n2\n\n\nO\n2\n\n\n(\n%\n)\n\n=\n200\n\u00d7\n\n\n\nI\nr\n\nN\n\n\n\nI\nd\n\n+\n\n\nI\nr\n\nN\n\n\n\n\n\n\n\nThe kinetic current density (jK) was calculated based on the following Koutecky-Levich equations:\n\n(Equation\u00a03)\n\n\n\n1\nj\n\n=\n\n1\n\nj\nL\n\n\n+\n\n1\n\nj\nK\n\n\n=\n\n1\n\nB\n\n\u03c9\n\n1\n/\n2\n\n\n\n\n+\n\n1\n\nj\nK\n\n\n\n\n\n\n\n\n(Equation\u00a04)\n\n\nB\n=\n0.62\nn\nF\n\nC\n0\n\n\n\nD\n0\n\n\n2\n/\n3\n\n\n\n\u03b3\n\n\u2212\n1\n/\n6\n\n\n\n\n\n\nwhere j is the measured current density, jL is the diffusion-limiting current density, \u03c9 is the angular velocity of the disk electrode (rpm), F is the Faraday constant (96,485 C mol\u22121), n is the electron transfer number in the reaction process, C0 represents the bulk O2 concentration (1.2\u00a0\u00d7\u00a010\u22123 mol L\u22121), D0 is the diffusion coefficient of O2 in electrolyte (1.9\u00a0\u00d7\u00a010\u22125 cm2 s\u22121), and \u03b3 is the kinematic viscosity of 0.1\u00a0M KOH (0.01\u00a0cm2 s\u22121).The accelerated stability testing was measured in O2-saturated 0.1\u00a0M KOH solution at a scan rate of 50\u00a0mV/s and rotation speed of 1,600\u00a0rpm.The accelerated durability testing was performed in O2-saturated 0.1\u00a0M KOH solution using the chronoamperometric method at 0.60\u00a0V for 60,000 s.The ECSA and Rf was calculated by the CV curves with different scanning rates. The non-faradic current measured was plotted as a function of the scan rate to obtain Cdl. Then, the ECSA was calculated according to the equations\n\n(Equation\u00a05)\n\n\nE\nC\nS\nA\n=\n\n\nC\n\nd\nl\n\n\n\nC\ns\n\n\n\n\n\n\n\n\n(Equation\u00a06)\n\n\n\nR\nf\n\n=\n\n\nE\nC\nS\nA\n\nS\n\n\n\n\nwhere Cdl is the capacitance for the sample, the value of specific capacitance (Cs) is 0.04 mF/cm2 in alkaline solution, and S is the geometric area of GCE (0.196\u00a0cm2).The X-ray absorption fine structure spectra (W L3-edge) were collected at the 1W1B station in the Beijing Synchrotron Radiation Facility (BSRF). The storage rings of the BSRF were operated at 2.5 GeV with an average current of 250 mA. Using a Si (111) double-crystal monochromator, the data collection was carried out in transmission/fluorescence mode using an ionization chamber. All spectra were collected under ambient conditions. The acquired EXAFS data were processed according to standard procedures using the ATHENA module implemented in the IFEFFIT software package. The k3-weighted EXAFS spectra were obtained by subtracting the post-edge background from the overall absorption and then normalizing with respect to the edge-jump step. Subsequently, k3-weighted \u03c7(k) data of W L3-edge were Fourier transformed to real (R)\u00a0space using a Hanning window (dk\u00a0= 1.0\u00a0\u00c5\u20131) to separate the EXAFS contributions from different coordination shells. To obtain the quantitative structural parameters around central atoms, least-squares curve parameter fitting was performed using the ARTEMIS module of the IFEFFIT software package. A k3 weighting, k-range of 2.8\u201312.8 \u01fa\u22121, and R range of 1.35\u20133.35\u00a0\u00c5 were used for the fitting. The four parameters coordination number, bond length, Debye-Waller factor, and E0 shift (CN, R, \u03c32, and \u0394E0, respectively) were fitted without any one being fixed, constrained, or correlated.\n46\n\n,\n\n47\n\n,\n\n48\n\nThe structural optimization and frequency analysis were performed at the GFN2-xTB level of the xTB package (v.6.2).\n49\n\n,\n\n50\n More accurate single-point energies were obtained at the r2SCAN-3c\n51\n level of theory using the ORCA 5.0 package.\n52\n Stationary points were optimized without symmetry constraint, and their nature was confirmed by vibrational frequency analysis. Unscaled vibrational frequencies were used to correct the relative energies for zero-point vibrational energy (ZPVE) contributions. The PDOS analysis was performed using the Multiwfn 3.7 package.\n53\n\nWe acknowledge financial support from the National Natural Science Foundation of China (21878265). We are also grateful to the Institute of High Energy Physics, Chinese Academy of Science, for XAFS measurements and help with XAFS measurements collected at the 1W1B station in the Beijing Synchrotron Radiation Facility (BSRF), China.Conceptualization, C.Z. and S.Z.; methodology, C.Z., W.A., and L.Y.; formal analysis, C.Z., S.Z., and L.Z.; investigation, C.Z., W.A., and L.Y.; resources, S.Z., C.Q., and M.L.; writing\u00a0\u2013 original draft, C.Z., S.Z., and L.Z.; writing\u00a0\u2013 review\u00a0& editing, C.Z. and S.Z.; funding acquisition, S.Z., C.Q., M.L., and L.Z.; supervision, S.Z.The authors declare no competing interests.Supplemental information can be found online at https://doi.org/10.1016/j.xcrp.2023.101288.\n\n\nDocument S1. Figures\u00a0S1\u2013S8, Table\u00a0S1, and supplemental experimental procedures\n\n\n\n\n\nDocument S2. Article plus supplemental information\n\n\n\n", "descript": "\n                  The oxygen reduction reaction (ORR) constitutes an important transformation process in fuel cells. Here we report a polymorphic tungsten catalyst with a concerted tungsten single-atom site (W-SAC) and tungsten nitride nanoclusters (WN-NPs) for mediating the ORR\u00a0processes. The catalyst is prepared via pyrolysis of phthalocyanine-typed material (WPc) with g-C3N4. Synergy of the W-SAC and WN-NP sites over the designed 2D g-C3N4 layer is responsible for the competitive ORR performance of the catalysts. Consequently, the WNPc/gC3N4-700 sample gives a half-wave potential of 0.835 V, a Tafel slope of 40.47mV dec\u22121, and a mass diffusion-limited current of 0.596 mA/cm2; such results are comparable with\u00a0and even better than that of 20% Pt/C. Furthermore, WNPc/gC3N4-700 also outperforms 20% Pt/C in terms of stability and durability. Our work may point toward a design strategy for active, stable, non-noble-metal ORR catalysts.\n               "}
{"full_text": "Data will be made available on request.Faced with the forthcoming depletion of fossil resources and the huge environmental impact related to their consumption, there is an urgent demand for an irreversible transition where these fossil resources are replaced by more sustainable compounds. One of the fundamental axes to achieve this green revolution is based on the transformation of residual biomass, based on oxygenated feedstock, into biocompounds that partially replace oil and petrochemical resources [1\u20133]. Within the wide variety of biocompounds [4], bioalcohols are easily produced with high yields and offer diverse and attractive functionalities, positioning themselves as one of the most important biomass derivatives blocks [1]. In recent years, four carbon linear alcohols, specially the straight-chain alcohol with the -OH group in the terminal position, 1-butanol (hereafter referred just as butanol or BuOH), is attracting growing attention due to its advantages as a fuel compared to ethanol, the bioalcohol with the highest implanted production. For instance, butanol has a higher energy density, assimilable to that of gasoline, higher heating value, lower volatility, is less hygroscopic, less corrosive and has a higher viscosity than ethanol. Furthermore, it is compatible with the currently used ignition engines which means it is easier to implement with some safety, sustainable and economic advantages compared to bioethanol [5,6]. Therefore, in recent years global scientific purposes have merged to overcome the higher production costs compared to ethanol when lignocellulosic feedstocks are used as raw materials [7]. These efforts have been mainly focused on the optimization of the metabolic bioengineering of the ABE (Acetone-Butanol-Ethanol) fermentation process in order to upgrade the tolerance of the microorganism to a higher butanol concentration [8\u201312]. Other efforts have been devoted to improving the sustainable butanol yields in the catalytic chemical routes [13], standing out the ethanol Guerbert condensation reaction [14,15].Moreover, butanol exhibits interesting applications not only in the fuel industry but also for the generation of added-value products [16,17]. On one hand, some valorization reactions where the oxygen functionality is retained allow the synthesis of aldehydes, esters, ethers or fatty acids such as butyraldehyde [18], butyric acid or the corresponding linear ether, with wide applications in the solvents, polymers or lubricant markets [17]. On the other hand, the most promising route for butanol upgrading implies the intermolecular dehydration giving rise to 1-butene, which is easily isomerised to a C4 olefins mixture, either directly or going through the di-n-butyl-ether (DNBE) as an intermediate (\nScheme 1). Nowadays, the extensively used C4 olefins (n-butenes, i-butene, butadiene) are mostly obtained as by-products in FCC units and classified in function of its composition in the so called C4 raffinates. Some predictions point to a further demand rise outpacing the supplies growth from fossil feedstocks [19], making it increasingly profitable to source via a sustainable route. In addition, this butanol transformation is the first step in the alcohol-to-jet-fuel process; once butanol is dehydrated into C4 olefins these can be further oligomerized and hydrogenated to obtain jet fuel cuts [20]. Nowadays, the C4 stream is usually subjected to a separation process to maximize the olefin composition to n-butenes, i-butene or butadiene. However, when butanol is used as a feedstock, the olefins composition can be tailored by modifying the reaction conditions. For instance, the single-step production of i-butene is thermodynamically limited to high temperatures (>350\u00a0\u00b0C) and requires stronger acid sites than those needed for linear olefins [21]. While an oxidizing atmosphere and a dehydrogenation functionality favour the butadiene yield from butanol [22].As it is generally accepted, alcohol dehydration reactions are catalyzed by acidic sites. In the case of butanol dehydration to butenes (1-butene and the two isomers of 2-butene), it is usually performed under atmospheric pressure and mild temperature conditions. Although enough temperature is required to restrict the DNBE yield, which is thermodynamically favoured at lower temperatures [23\u201325]. Different catalysts have been tested, either based on Lewis or Br\u00f8nsted acid sites, being both active but with a clear different catalytic behaviour. Grouping the catalysts by the nature of their acid sites, those mainly formed by Lewis or Br\u00f8nsted acid sites, the model studied catalysts are represented by \u03b3-alumina [26\u201328] or the acidic form of the zeolite ZSM-5, respectively. Although the former exhibits satisfactory reaction stability, much further contact times and higher reaction temperatures are usually needed to achieve high butenes selectivity [24,29\u201331]. However, this required elevated temperatures can lead to a decrease in butenes selectivity due to the formation of lower molecular weight compounds such as methane, ethylene or propylene [32]. Although catalysts based on Br\u00f8nsted acid sites are much more active [30,33], they frequently exhibit high deactivation rates, accompanied by a loss of butenes selectivity due to secondary reactions [32,34]. For example, Palla et al. [34] obtained a full conversion (WHSV=15\u2009h\u22121, 250\u2009\u00b0C) but with a moderate selectivity towards butene (75 %). In this context, some authors have addressed this issue by combining both types of acid sites in the catalyst\u2019s formula. De Reviere et al. [29] obtained slightly lower conversion values employing a hybrid catalyst based on nano H-ZSM5/Al2O3 in comparison with the same zeolite (60 % vs 80 %) at 240 \u00b0C, but with substantial improvements in terms of stability. Other researchers have modified the catalyst surface by the introduction of a higher quantity of Lewis acid sites, some of them with a greater strength [32]. Almost underrated in the literature, catalyst\u2019s deactivation plays a critical role to be considered in the industrial implementation of 1-butanol dehydration process. Butanol is a less reactive molecule in comparison to its C4 isomers and shorter chain alcohols, therefore higher catalyst\u2019s acidity is required to be activated. This fact, along with its higher molecular weight, multiplies the tendency to form coke precursors that lead to significant deactivation rates during the dehydration reaction [35,36].Acidic zeolites are by far the most studied materials for this dehydration reaction with a widespread use of the acidic form of the zeolite ZSM-5 [30,36\u201341]. Furthermore, the larger molecular size of butanol in comparison to lower alcohols, has revealed clear influences in terms of selectivity and stability when different zeolite topologies are used, obtaining the best performance with a MFI zeolite [38,42]. In addition to the examples already given, other acidic catalysts for the synthesis of butenes through butanol dehydration have been explored, such as amorphous aluminosilicate [41], CeO2-TiO2/carbon composites [43,44], tungstated zirconium [33], modified mesoporous silica [30] or phosphate modified carbon nanotubes [45]. A group of solid acid catalysts that have attracted much attention in recent years, due to their outstanding structure and strong Br\u00f8nsted acidity, are the polyoxometalates (POM), also known as heteropolyacids (HPA). Among them, those with Keggin structure stand out for their extraordinary performance in several industrial acid-catalyzed reactions and oxidation catalytic processes [46,47]. The Keggin structure contains an heteropolyanion stabilized by acidic protons with the formula [XM12O40]n-, standing X for the heteroatom (usually P5+,Si4+) and M for the addenda atom (Mo6+, W6+). In detail, the heteropolyanion is composed of a central tetrahedron XO4 surrounded by 12 MO6 octahedra. These Keggin units in their hydrated form are coordinated with crystallization water molecules forming a body-centered cubic structure with the Keggin anions at the lattice points and acidic H2O5\n+ bridges along the faces [48]. On one hand, the main advantages offered by this complex structure are strong and uniform acid sites, as well as their ease of tuning by modifying their constituents. On the other hand, the major shortcomings of HPAs are related to their very low surface area and poor thermal stability, unable to cope with the highly demanding regeneration treatments typically required in these acid-catalyzed reactions. Several strategies have been devoted to overcome this matter related to coke formation, ranging from the development of thermally resistant HPAs to the performance of structural modifications to avoid coke formation or to enhance coke combustion [49].One of the unique features of HPAs is their so-called pseudoliquid behavior, which allows polar molecules to enter the crystal structure and react in the bulk. In a series of articles [48,50\u201352], Gaigneaux's group deeply explored this pseudoliquid performance, resulting not only in the exploiting of the HPAs bulk but also favouring the coke inhibition and reversing its deposition thanks to a smart pre-treatment strategy. They applied this approach in the gas phase methanol dehydration to dimethylether (DME), which involved subjecting the HPA to proper conditions to partially replace the water of crystallization with methanol. Unfortunately, the same authors stated that this treatment was not extensible to butanol dehydration, due to the comparative increase in hydrophobicity and molecule dimensions. Accordingly, they attempted to partially substitute the acidic protons with NH4\n+ to activate the bulk of the HPAs [53]. Although they achieve to exploit the bulk of the HPA, the strategy was just totally successful at low temperature, thus only producing the intermolecular dehydration product, DNBE. Consequently, it seems that to develop a sufficiently active catalyst for butanol dehydration, the remaining and plausible strategy to exploit most of the acid sites in HPAs is still their dispersion on a suitable support. The use of HPAs in butanol dehydration have been mainly focused on obtaining DNBE, showing activity and selectivity values similar to the model catalysts, Amberlite [54,55]. The HPAs have been also used unsupported, taking advantage of its uniform strong Br\u00f8nsted acidity to modulate the reaction mechanism [56\u201358]. Recently, the effect of the HPA loading have been studied supported on silica, reaching almost completely conversion and selectivity to linear butenes (300\u2009\u00b0C, WHSV of 37,4\u2009h\u22121\n)\n[59] and on TiO2\n[60], showing high improvements of the catalytic activity in a photoassisted catalytic reaction.Here we report the catalytic performance \u2212in terms of activity, selectivity, stability, and regeneration ability\u2212\u2009of a series of the synthesized catalysts, emphasizing the role of HPA-support interactions, during the gas phase dehydration reaction of butanol to butenes. Two HPAs namely H4SiW12O40 (STA) and H3PW12O40 (TPA) were deposited on two commercial carbonaceous supports: an activated carbon (AC) and a high surface area graphite (HSAG). The STA was also dispersed over silica, alumina and zirconium oxide, for comparative purposes. The loading of the HPAs was as low as 15 % to evaluate the HPA-support interactions. The catalyst activity was compared against a zeolite HZSM-5 with a Si/Al ratio=\u200923. Our results evidence that graphite-STA interactions efficiently tailored the acidity resulting in an active regenerable catalyst.Heterogeneous HPAs based catalysts have been synthesized by incipient wetness impregnation from a EtOH/H2O solution with a 1:1 volumetric ratio containing the appropriate polyoxometalate quantity. The HPAs employed in this work were those with a remarked high acid nature, the Tungstophosphoric acid (TPA) (H3PW12O40\u00b7nH2O; Sigma-Aldrich) and the Silicotungstic acid (STA) (H4SiW12O40\u00b7nH2O; Sigma-Aldrich). The minimum amount of solvent to fill the pores of each support was used to solve the amount of HPA, reaching a nominal content of 15\u2009wt%. The solution was carefully deposited dropwise on the two carbonaceous materials. The commercial activated carbon (AC, SBET= 1190\u2009m2/g) was produced from olive stones by Oleicola el Tejar, C\u00f3rdoba (Spain) and pretreated with hydrochloric acid to remove the inorganic impurities. The high surface area graphite was provided by Timcal (HSAG, SBET=400\u2009m2/g). These carbonaceous supports were impregnated with both HPAs obtaining the following catalysts: TPA/HSAG, TPA/AC, STA/HSAG and STA/AC. The support effect was studied by immobilizing the STA on different oxide supports, that is: alumina purchased from Degussa (Al2O3, SBET= 173\u2009m2/g), zirconium oxide supplied by Melcat (ZrO2, SBET=100\u2009m2/g) and silica from Fluka (SiO2, SBET= 433\u2009m2/g), giving rise to the following catalysts: STA/Al2O3, STA/ZrO2 and STA/SiO2, respectively. Upon impregnation, these samples were dried in an air oven at 110\u2009\u00b0C overnight to drive off the volatile components within the solution. In order to obtain the acidic form of the ZSM-5 zeolite, the ammonium form (CB 2314, Zeolyst international) was subjected to thermal treatment under static air (5\u2009h, 550\u2009\u00b0C).Textural properties of the supports and catalysts were evaluated from N2 adsorption-desorption isotherms. The measurements were carried out at 77\u2009K using a Micromeritics 2020 ASAP equipment, before the measurement the samples were outgassed in vacuum at 423\u2009K for 16\u2009h. The Brunauer-Emmet-Teller method was used to calculate the mesoporous external specific surface area, while the microporous surface was evaluated using t-plot method. The total pore volume was estimated from the adsorbed amount at a relative pressure of 0.97 and the average pore size was determined by the BJH method using the desorption branch.X-ray diffraction analysis (XRD) were performed on a Polycristal X\u2009\u00b4\u2009Pert Pro Pananalytical diffractometer with Ni-filtered Cu/K\u03b1 radiation (\u03bb\u2009=\u20090.1544\u2009nm) operating at 45\u2009kV and 40\u2009mA. In each measure, Bragg\u00b4s angles between 4 and 90\u00ba were scanned at a rate of 0.04\u2009deg/sec. Structural properties were also characterized by Fourier transform infrared spectroscopy in attenuated total reflectance disposition (FTIR-ATR) with a JASCO FT/IR 4800 spectrometer equipped with a DTGS detector and a germanium crystal. A total of 170 scans per spectrum were recorded between 400 and 4000\u2009cm\u22121 with a resolution of 4\u2009cm\u22121.Acidity characterization of the samples was done in a Micromeritics Autochem II 2920 equipment. The samples were pre-treated in helium flow at 623\u2009K in situ prior to the analysis. Subsequently, total number of acid sites were established by NH3 pulse chemisorption analysis, at 393\u2009K to avoid physisorption artefacts, until complete saturation of the sample. Finally, the samples were flushed in helium for 1\u2009h, then the temperature was raised to 623\u2009K and hold for 1\u2009h whilst NH3 desorption were recorded (NH3-TPD). Acid site strengths were classified into weak (393\u2013523\u2009K), moderate (523\u2013623\u2009K) and strong acid sites (>623\u2009K).The 3,3-dimethyl-1-butene (33DM1BN) isomerisation model reaction was carried out to evaluate the Br\u00f8nsted acidity of the catalysts. In a typical experiment, 100\u2009mg of the catalyst (sieved between 0.35 and 0.5\u2009mm) was fixed with glass wool in a U-shape glass reactor and pre-treated in situ at 275 \u00baC during 30\u2009min under continuous N2 flow (30\u2009cm3/min). Thereafter, the catalyst was stabilized at the reaction temperature (150\u00baC) performing the reaction with the same N2 flow saturated after bubbling into pure 33DM1BN at 0\u00baC (P33DM1BN=20.4 KPa). Not converted reactant and the isomerized products were analyzed by online gas chromatography (Varian 3400) equipped with a flame ionization detector (FID) and a 20 % BMEA S/Chrom p.80/100 column. The catalytic activity (A) normalized per gram of active phase (per gram of HPA or zeolite) was calculated as below:\n\n\n\nA\n\n\n\n\n\n\n\nmmol\n\n\n33\nDM\n1\nBN\n\n\n\n\nmin\n\u2219\n\n\ng\n\n\na\n.\np\n.\n\n\n\n\n\n\n\n=\n\n\n\n\nC\n\n\n33\nDM\n1\nBN\n\n\n\u2219\n\n\n\n\n\n\nF\n\n\n33\nDM\n1\nBN\n\n\n\n\n\n\nin\n\n\n\n\n100\nx\n\n\ng\n\n\na\n.\np\n.\n\n\n\n\n\n\n\nwhere \n\n\n\n\n\n\nF\n\n\n33\nDM\n1\nBN\n\n\n\n\n\n\nin\n\n\nand \n\n\ng\n\n\na\n.\np\n.\n\n\n stands for the 33DM1BN inlet flow expressed in mmol/h and the HPA or zeolite weight in each sample (active phase), respectively. The initial activity values were obtained at time=\u20090 after linearization of the catalytic activity versus time. \n\n\nC\n\n\n33\nDM\n1\nBN\n\n\n is the conversion calculated using the equation:\n\n\n\n\n\nC\n\n\n33\nDM\n1\nBN\n\n\n(\n%\n)\n=\n\n\n\n\n\n\n\n\nF\n\n\n33\nDM\n1\nBN\n\n\n\n\n\n\nin\n\n\n\u2212\n\n\n\n\n\n\nF\n\n\n33\nDM\n1\nBN\n\n\n\n\n\n\nout\n\n\n\n\n\n\n\n\n\n\nF\n\n\n33\nDM\n1\nBN\n\n\n\n\n\n\nin\n\n\n\n\n\u2219\n100\n\n\n\nwhere \n\n\n\n\n\n\nF\n\n\n33\nDM\n1\nBN\n\n\n\n\n\n\nout\n\n\n is the 33DM1BN molar flow at the outlet.Thermogravimetric analysis of the fresh and used catalysts were done using a SDT Q600 (TA instruments) under synthetic air atmosphere (Fair = 100\u2009mL/min) from 30\u00baC to 800\u00baC and a 10\u00baC/min heating ramp.The typical experiment was carried out into a stainless steel fixed-bed reactor. The reaction was tested at 275 \u00baC under atmospheric pressure, using 30\u2009mg of catalyst and a constant flow of inert gas (He, 100\u2009mL/min) and reactant (liquid BuOH, 0.04\u2009mL/min injected by HPLC pump). The weight of the catalyst was adjusted in the selectivity studies performed at different temperatures, aiming to obtain similar conversion values. Before reaction all catalysts were treated at the same reaction temperature (275 \u00baC) during 30\u2009min. The gas products were analysed online by means of a gas chromatograph (450-GC) equipped with thermal conductivity (TCD) and flame ionization (FID) detectors. The compounds with high boiling point (BuOH and DNBE) were condensed and collected every reaction hour. These liquid samples were weighted and analysed using the same gas chromatograph, but in this case only the FID detector was used.The butanol conversion \n\n\n\nC\n\n\nBuOH\n\n\n(\n%\n)\n\n\u00a0and selectivity to a product i (\n\n\n\nS\n\n\ni\n\n\n\n\n\n%\n\n\n\n\n were estimated employing the following equations, respectively:\n\n\n\n\n\nC\n\n\nBuOH\n\n\n(\n%\n)\n=\n\n\n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nin\n\n\n\u2212\n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nout\n\n\n\n\n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nin\n\n\n\n\n\u2219\n100\n\n\n\n\n\n\n\n\n\n\nS\n\n\ni\n\n\n\n\n\n%\n\n\n\n=\n\n\n\n\nn\n\n\ni\n\n\n\n\nF\n\n\ni\n\n\n\n\n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nin\n\n\n\u2212\n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nout\n\n\n\n\n\u00b7\n100\n\n\n\nwhere \n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nin\n\n\n represents the molar flow of butanol fed to the reactor, \n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nout\n\n\n is the butanol molar flow at the reactor exit expressed in mmol/min, \n\n\nF\n\n\ni\n\n\n is the molar flow of the product i and \n\n\nn\n\n\ni\n\n\nthe stoichiometric factor of the product i relative to BuOH. The catalytic activity (A) was defined as:\n\n\n\nA\n\n\n\n\n\n\n\nmmol\n\n\nBuOH\n\n\n\n\nmin\n\u2219\n\n\ng\n\n\na\n.\np\n.\n\n\n\n\n\n\n\n=\n\n\n\n\nC\n\n\nBuOH\n\n\n\u2219\n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nin\n\n\n\n\n100\nx\n\n\ng\n\n\na\n.\np\n.\n\n\n\n\n\n\n\n\nThe carbon mass balance (CB) was determined as:\n\n\n\nCB\n\n\n\n%\n\n\n\n=\n\n\n\n\n\u03a3\n\n\ni\n\n\n\n\nn\n\n\ni\n\n\n\n\nF\n\n\ni\n\n\n+\n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nout\n\n\n\n\n\n\n\n\n\n\nF\n\n\nBuOH\n\n\n\n\n\n\nin\n\n\n\n\n\u00b7\n100\n\n\n\n\nThe nitrogen adsorption-desorption isotherms at 77\u2009K for the bare supports and the synthesized catalysts are depicted in Figs. S1-S5. A summary of the textural properties, including the pore volume and the average pore size of the supports and catalysts, is listed in \nTable 1. Although the diverse nature of the supports induced great differences in the shape of their hysteresis loops, the isotherms of all the supports can be classified as type IVa (IUPAC), characteristic of mesoporous materials, except for the AC. The latter constitutes a typical type I isotherm although there is also a contribution of a type IVa isotherm due to the existence of mesoporosity, as denoted by the presence of the H4 type hysteresis loop. These textural features are distinctive of this category of micro-mesoporous activated carbons [61]. As for the HSAG, exhibited a type H3 loop typical of materials where the mesoporosity is caused by plate shape sheets, while the oxide supports, Al2O3, ZrO2 and SiO2, generate a H2 type loop. All supports denoted relatively high specific surface areas, ranging from 100\u2009m2/g for the zirconia to 1025\u2009m2/g for the activated carbon (Table 1). Even though the AC surface seems to stand out, the mesoporosity contribution was in the same range as the rest of the supports. After HPAs deposition, apart from the reduction of the microporous branch in the AC-based catalysts, the shape of the isotherms remained like those of the supports (Figs. S1-S5), although significant changes were detected in the textural parameters, summarized in Table 1. All samples manifested some reduction of the specific surface area, being especially remarkable for the microporous surface contribution of the TPA/AC catalyst. This loss of porosity suggests the partial blockage of micropores by HPA particles. As for the plate-shape material, HSAG, the incorporation of the HPA prompted a reduction in the BET surface area from 396\u2009m2/g to 286\u2009m2/g (TPA/HSAG) and to 272\u2009m2/g (STA/HSAG). Although less obvious, the mesoporous oxides (SiO2, Al2O3 and ZrO2) likewise exhibited a reduction in the surface area after HPA impregnation, coherent with the partial coverage of their mesopores.The X-ray diffractograms of the carbon-supported HPA catalysts are represented in \nFig. 1, along with the pristine HPA and either the AC (Fig. 1a) or the HSAG (Fig. 1b) support. In general, the XRD patterns of the HPA-catalysts did not show the diffraction maxima characteristic of the HPAs except in the case of TPA/AC (Fig. 1.a). This latter displayed some distinctive reflections of the TPA, which suggests the presence of larger crystallite particles of this heteropolyacid. In every case a broad and diffuse peak centered around 2\u019f=\u20097\u00b0 is appreciated, typical of the HPA with a raised level of hydration [31]. The same trend is observed in the X-ray diffraction patterns of the oxide-supported HPAs (Figs. S6-S8), where no diffraction lines characteristic neither the HPAs nor other species, such as those of the former oxides e.g. WO3 (with reflection lines at 2\u019f= 23\u00ba, 24\u00ba, 34, 42\u00ba), are appreciated. Hence, these diffractograms point to a finely dispersion of the HPAs thorough the supports. This is in concordance with previous reported studies, where no crystallinity is reported until high HPA loadings are reached; for example, above 20\u2009wt% on silica [62] or greater than 40\u2009wt% if a carbon support is employed [63].FTIR-ATR spectroscopy is a sensitive technique to detect the vibrations of the molecular bonds, thus was performed aiming to confirm the retention of the Keggin structure in the synthesized catalysts. A magnification of the characteristic fingerprint region for the TPA-based catalysts (700\u20131200\u2009cm\u22121) is represented in \nFig. 2.a. The bands that appeared at 1071, 970, 900, 761\u2009cm\u22121 in the spectrum of the pristine TPA are in good agreement with the literature findings which assign them to the stretching vibrations of P-O, WOd, W-Ob-W (corner sharing), W-Oc-W (edge sharing), respectively, characteristics of the Keggin structure. The TPA/AC and TPA/HSAG spectra showed the preservation of the Keggin structure in both samples, although a slight shift is detected for the TPA-graphite supported. This displacement towards higher wavenumbers may be a consequence of a strong HPA-support interaction, typically observed with carbon materials [64\u201367]. This can be interpreted as a charge transfer between the basal planes or the functional groups of the graphite and the HPA that ultimately stabilizes the charge that would otherwise be stabilized with the neighbouring polyanions [68]. This is reinforced by the lack of crystallinity which can ultimately lead to a high dispersion even at high loadings on carbon materials, and specifically for the HSAG-supported catalysts. As for the STA spectra (Fig. 2.b), the fingerprint bands obtained for the pristine HPA at 1016, 978, 906 and 736\u2009cm\u22121 are assigned to the typical antisymmetric stretching vibrations WO, Si-Od, W-Ob-W (corner sharing) and W-Oc-W (edge sharing), respectively [69]. Fig. 2.b and Fig. 2.c represent a magnification of this region (700\u20131200\u2009cm\u22121) for the STA-carbon catalysts and STA-oxides, respectively. In the STA and the STA-based catalysts spectra the characteristics vibration bands of the HPA bonds are appreciated, and no signals of other species are detected. As already mentioned for the TPA-HSAG sample, the interaction effects between the STA and the graphite support are also evidenced in its corresponding spectrum (Fig. 2.b). Except for the STA/Al2O3 sample, whose spectrum does not clearly evidence the characteristic Keggin vibration bands (Fig. 2.c), the FTIR analysis confirms the preservation of the Keggin structure in the synthesized catalysts.Hence, these FTIR-ATR studies along with the XRD patterns and the nitrogen physisorption isotherms, suggest the HPA are not deposited on the surface as crystalline or isolated species but rather forming clusters of a few highly dispersed units that allow the conservation of a large part of the specific surface area of the supports.The structure, physicochemical properties and therefore the catalytic performance of HPAs are strongly dependent of the temperature. For instance, the pristine TPA, the most stable Keggin HPA, loses the crystallization water from the stable hexahydrate structure within the temperature range of 100\u2013280 \u00baC, while the total deprotonation to the unstable anhydride form (hence, the Keggin structure decomposition), starts around 370 \u00baC [46]. These processes take place at lower degrees for the pristine STA [46] and can shift to upper or lower temperatures according to HPA-support interactions [70]. This means that HPA decomposition is accompanied by water desorption resulting in chemisorbed NH3 release at lower than expected temperatures. Consequently, acidity characterization by classical NH3-TPD is challenging and may be misinterpreted, since the strength of the acid sites could be underestimated.Bearing in mind the decomposition temperatures above described, the total amount of acid sites was measured following the next procedure. After a pre-treatment stage at 350 \u00baC, the sample was subjected to a NH3 pulse chemisorption process at 120 \u00baC. The temperature was then raised recording the desorbed NH3 with time. The number of strong acid sites was estimated from the subtraction of the total number established by NH3 pulse chemisorption from those registered by the TPD. The same procedure was applied to the bare supports. NH3-TPD profiles of the HPA catalysts are represented in \nFig. 3 while those of the bare supports are depicted in Figs. S9-S12. The total number of acid sites and their distribution according to the strength degree are summarized in \nTable 2.In terms of their acidic strength, the different nature of the used supports provides some properties into the catalysts: while carbon materials and silica contribute with a limited amount of acid sites, the Al2O3 and ZrO2 acidic oxides significantly incorporate weak-moderate strength centers (Table 2). The developed acidity by a supported HPA is the result of the combination of various parameters: (i) the number of acid sites, (ii) the acid-base reaction between the HPA and the support and (iii) the aggregation degree of the Keggin clusters which ultimately provide the acidic protons [71]. The larger the HPA crystallites formed, the higher similarity to the pristine HPA, thus the greater the strength. The extent of these acidic protons has been measured quantitatively, being highly dependent on the superficial properties of the support in a wide range of nominal monolayer coverage [72]. Therefore, the combination of these variables allows to finely tune the acidity of the catalyst based on the requirements of the reaction [72].As expected, the incorporation of the HPA gave rise to a large increase in the amount and strength of acid sites (particularly the moderate and strong ones), for all the samples except STA/ZrO2. It should be noted that although ZrO2 is the support with the lowest surface area (Table 1), the STA coverage is far from the monolayer coverage. For this catalyst, the decrease in the total number of acid sites is probably due to the reduction in the specific surface area, added to the fact that the basic sites present in this oxide tend to neutralize the more acidic protons. At the same time, the polar surface is prone to generate a homogeneous dispersion of the HPA. In any case, supporting the HPA resulted in a relative increase of the acid strength, as registered in Table 2. At the other extreme, the STA/Al2O3 catalyst has generated the highest number of acid centers after the incorporation of STA, especially those of weak-moderate strength. As for the silica, STA functionalization has generated less acidity than for carbon materials (Table 2).Establishing a comparison between the two types of heteropolyacids, TPA-based catalysts developed less amount of total acid sites but with higher strength than their STA-based counterparts. This is in concordance with a lower number of acidic protons per cluster but higher strength of the pristine TPA in comparison with the STA [46]. If the carbon supported catalysts are compared, both STA/AC and TPA/AC exhibited a slight major amount of acid sites than their graphite-supported equivalents (Table 2). Nonetheless, as revealed by the NH3-TPD profiles, the developed acidity in the HPA/HSAG catalysts is less strong or shifts to a relatively higher amount of moderate-strength acid sites.NH3 is an unspecific base that indistinctly chemisorbs in Lewis or Br\u00f8nsted acid sites. Due to the importance of Br\u00f8nsted acid sites in the butanol dehydration reaction, [33,40] the characterization of this type of acid centers was performed by the isomerization reaction of 3,3-dymethyl-1-butene (33DM1BN). The use of model reactions allows the study of the catalyst in thermal and atmospheric conditions equal to or similar to those of the reaction under study. 33DM1BN isomerization proceeds through a pure protonic mechanism due to the substitution degree of the molecule with a quaternary carbon atom which impossibilities the \u03c0-allylic intermediate formed in Lewis based isomerization reactions mechanism [73]. If the reaction is performed at temperatures below 300 \u00baC often leads to two single products, 2,3-dymethylbut-1-ene (23DM1BN) and 2,3-dymethyl-but-2ene (23DM2BN) (\nScheme 2), being successfully applied to study Br\u00f8nsted acid sites of weak to medium strength [74\u201376].However, the isomerization of 33DM1BN is a susceptible reaction to coke deposition and the catalyst undergoes deactivation from very early reaction time. Aiming to avoid this deactivation effect on the comparison among catalytic activities, conversions were measured every 4\u2009min during a brief period (20\u2009min) and the activity was then linearized and extrapolated to zero time. For comparative purposes, the widely used zeolite ZSM-5 was also evaluated in this reaction, and the obtained catalytic performance is displayed in \nFig. 4, along with those of the HPA-based catalysts. Firstly, the supported HPAs catalysts exhibited higher initial activity than the model zeolite ZSM-5, which suggests the presence of higher quantity of Br\u00f8nsted acid sites in the formers. Secondly, some interesting deductions can be established by comparing these catalytic results with the acidity characterization revealed by NH3-TPD. On one hand, both STA/AC and TPA/AC displayed somewhat higher initial activity than those supported on the HSAG, in good agreement with the highest number of acid sites measured by NH3 chemisorption. On the other hand, and contrary to that one would deduce from the NH3-TPD profiles, STA/SiO2, and STA/ZrO2 exhibited a similar catalytic performance quite comparable to that of the STA/HSAG. In addition, despite STA/Al2O3 was at the top in the number of acid sites (Table 2), it displayed the worst initial activity among the HPA-based catalysts, suggesting the presence of a lower quantity of weak-medium Br\u00f8nsted acid sites.Prior to the catalytic tests, the absence of catalytic activity in the blank and bare support tests was verified at 275\u00baC, obtaining conversion values close to 2% with the alumina, the most active support.As already mentioned in the introduction section, the 1-butanol dehydration reaction towards butene isomers corresponds to the intramolecular dehydration (Scheme 1). Whereas intermolecular dehydration gives rise to DNBE and is favoured at lower temperatures from a thermodynamic point of view. To evaluate the effect of temperature on product distribution, catalytic tests were performed on STA/HSAG using different WHSV values in order to compare selectivity values under quasi-isoconversion conditions. (\nFig. 5). Within the studied temperature range (175\u2013225 \u00baC), the catalyst was highly selective to dehydration products (>99 %) either the intermolecular or intramolecular dehydration product, obtaining a mixture of linear butenes and DNBE. The following major compound was the dehydrogenation product, butyraldehyde, with selectivity values below 1 % in all cases. No signals of permanent gases (CO, CO2, H2, CH4), skeletal isomerization to i-butene or cracking products were registered. Only trace amounts of unknown heavier molecular weight compounds were detected although not relevant as carbon balance was kept above 98 % at every run.As depicted in Fig. 5, the selectivity towards linear butenes rises with temperature from the initial 40 % at 175 \u00baC to approximately 98 % at 275 \u00baC, increasing at the same time the isomerization to 2-butenes although maintaining a similar cis to trans ratio (close to 1). As observed, the dehydration reaction is strongly dependent on temperature not only in terms of activity (very different WHSV values were required to achieve similar conversion levels) but also of selectivity.Considering the highest selectivity to butenes, 275 \u00baC and a WHSV of 64.8\u2009h\u22121 were the conditions selected to compare the catalytic performance among our different synthesized catalysts.In \nFig. 6.a the catalytic activity with time on stream of both HPAs carbon-supported is compared. Except TPA/HSAG (26.7\u2009mmolBuOH\u2219min\u22121\u2219g\u22121\na.p), the HPA/carbon catalysts showed quite similar initial catalytic activities in 1-butanol dehydration reaction (37.7, 38.7\u2009y 40.0\u2009mmolBuOH\u2219min\u22121\u2219g\u22121\na.p for TPA/AC, STA/HSAG and STA/AC, respectively), which tend to converge to more comparable values with time on stream. Taking into account the acidity characterization results (Table 2), the higher amount of Br\u00f8nsted acid sites exhibited by the STA-based catalysts (Fig. 4) seems to compensate the superior acid strength characteristic of TPA [46], resulting in rather comparable activities in the butanol dehydration reaction. As frequently reported for this reaction, all catalysts suffered from a high deactivation rate. It is noticeable the higher deactivation rate exhibited by the STA/AC catalyst compared to that of the TPA/AC. This is probably due to coke deposition on the smallest micropores which were already blocked after the synthesis stage of TPA/AC, as the textural characterization suggested (Table 1).Support\u2019s nature has a great influence on the catalytic activity as observed on Fig. 6.b, where the catalytic activity with time on stream is represented for the STA dispersed on different supports and the model ZSM-5. Except STA/Al2O3, these catalysts showed an outstanding catalytic performance clearly surpassing one of the most widely studied zeolites. This has been already reported for other heterogeneous catalytic acid-demanding reactions [46]. Bearing in mind the relative low strength and amount of acid sites determined by TPD-NH3, STA/SiO2 exhibited a surprisingly high initial activity, although Br\u00f8nsted activity already denoted by the 33DM1BN model reaction was in the same range than the rest (Fig. 4 and Table 2). On the other extreme, STA/Al2O3 displayed the lowest catalytic activity despite presenting the largest number of acidic centers (characterized by TPD-NH3), although the lowest Br\u00f8nsted acidity (as revealed by the results of the 33DM1BN isomerization reaction). This fact, along with the non-appearance of the characteristic HPA reflections in the XRD pattern (Fig. 1) and the absence of characteristic Keggin vibration bands on the FTIR spectrum (Fig. 2c) led us to think on a plausible partial decomposition of the Keggin structure, which could explain this dropping in Br\u00f8nsted acidity. This latter effect has already been stated by Pizzio et al. [77] and attributed its worse catalytic performance on the i-propanol dehydration to this fact. They reported the TPA decomposition into a polymeric anion [P2W21O71]7- and WO3 when a TPA solution is in contact with an Al2O3, which was confirmed by different techniques.Interestingly, although the initial catalytic activities of the STA/ZrO2, STA/HSAG and STA/AC catalysts are well correlated to those obtained in the 33DM1BN model reaction, the deactivation rate displayed for each catalyst differs greatly. The large activity drop exhibited by STA/AC has been above mentioned and can be easily disclosed in terms of the support textural features. On the opposite, the low deactivation rate of the STA/ZrO2 catalyst compared to the rest is striking. Since the porosity of the ZrO2 differs from the other supports, a steric hindrance effect to inhibit coke precursors cannot be completely discard, but a close look to the products distribution (\nFig. 7) points to a differentiating fact with respect of ZrO2, which will be discussed below.Different reaction mechanisms have been proposed for the dehydration of alcohols and specifically for BuOH. They can be disclosed into an E1, E1cB or E2, standing for a unimolecular, concerted or bimolecular elimination mechanism, respectively, which take place on acidic, basic, or acid-base sites [78], respectively. Although they are competitive mechanisms, the 1-butene to 2-butene isomerization ratio is supposed to be closely related to the dominant path, being lower when the active sites are of basic nature. Moreover, a 2-butene cis/trans ratio lower than 1, as it is the case for STA/ZrO2 (rcis/trans=0.69), is distinctive of basic isomerization catalysts. Indeed, 1-butene isomerization has long been used as a model test of acid-basic catalytic sites [79]. In this catalyst, HPA may be highly dispersed near to a basic center developing a strong acid-weak base interaction (otherwise if a strong base were involved the dehydrogenation to 1-butanal would be more relevant). This acid-base interaction could not only stabilize the reaction intermediate but also aiding in a fast product desorption, avoiding or shortening the contact time with the strong acid, which is ultimately responsible of the catalytic activity (as the ZrO2 itself has no activity under these conditions). Except for STA/Al2O3 and STA/ZrO2, the obtained product distributions are rather similar and characteristic of Br\u00f8nsted acid sites as is the nature of HPA. Thus, in the mentioned exceptions, the support can tailor at some extent the acid strength and interaction or desorption of the reaction intermediates but not modifying the acidity nature of the active site involved in the reaction mechanism of the alcohol dehydration reaction.Finally, the regeneration capacity of our catalysts was studied. As already indicated in the introduction section, this is a subject frequently underestimated in literature and especially in the recent objective of biobutanol valorization. However, it is a relevant topic to be discussed that can determine the implementation feasibility of a large-scale production process [80]. As shown above, all the studied catalysts exhibited a more or less pronounced activity drop which tends to stabilize after some hours on stream, which is characteristic of acid-catalyzed reactions [46]. Moreover, it is one of the major limitations when HPA catalysts are used [46,81]. We attempted to address this catalyst\u2019s deactivation matter by employing some different strategies. Initially, despite described fruitful for similar catalytic systems [82], the introduction of a metallic functionalization to oxidize coke species or testing with different gas streams in the feed (for instance, reductive atmosphere) were not successful.An interesting regeneration treatment consisted of a mild oxidation process of the deactivated catalyst in synthetic air at 400 \u00baC for 90\u2009min, and the obtained results for STA/HSAG and STA/SiO2 are depicted in \nFig. 8. As observed, while the initial catalytic activity of the STA/HSAG was fully recovered after the treatment, the silica-supported catalyst not only did not recuperate its initial activity, but it decreased even more. It must be mentioned that the rest of catalysts subjected to this regeneration process (not shown for the sake of brevity) exhibited a similar behaviour than that of STA/SiO2; thus, being STA/HSAG the only regenerable catalyst. It should be noted that the recovery of the catalytic activity for STA/HSAG is not due to an effect of increase in the intrinsic activity in the oxidation process, because the same treatment to the fresh catalyst did not rise the initial activity already obtained with the untreated sample (not shown for the sake of brevity).The already STA-graphite interactions demonstrated (displacement of the HPA signal towards higher wavenumbers in the FTIR-ATR spectrum), the optimal degree of HPA crystallites agglomeration and the consequent decrease in the acid sites strength in comparison to that of silica, may be the reasons to explain the feasible regeneration of this catalyst.In order to get more insights to explain the different behaviour detected among our catalytic systems after the regeneration treatment, thermogravimetric analysis of fresh and used samples was performed and plotted in \nFig. 9. As for the fresh STA/SiO2 (Fig. 9.a), apart from the weight loss characteristic of physiosorbed water until 100 \u00baC, the thermogravimetric analysis profile registered a weight loss starting a few degrees above 350 \u00baC that corresponds to the 1.5\u2009H2O molecules release that gives rise to the decomposition of the Keggin structure [46]. For the used catalyst, the derivative curve evidences an asymmetric peak, whose maximum shifted a few degrees to higher temperatures in comparison to the fresh catalyst (from 435 to 462\u2009\u00b0C). This contribution may be clearly deconvoluted into two peaks: one characteristic of the Keggin structure decomposition and the other one of the strongly adsorbed coke oxidation reaction. This clearly evidences the concurrence of the HPA decomposition and coke oxidation processes, which explains the unsuccessful application of the thermal regeneration treatment for this catalyst (Fig. 8). Interestingly, a weight loss at around 260\u2009\u00b0C, very close to that of the BuOH dehydration reaction, which is absent in the fresh catalysts, is also observed. This can be attributed to the desorption of reaction intermediates, more precisely to DNBE. Thus, the fast activity drop produced in the first reaction hours can be ascribed to the time needed to reach a steady-state equilibrium of the different species implied in the reaction pathway, which blocked the active sites. These species are removed after the regeneration treatment and the initial fast deactivation takes place again.Curiously, the fresh and used STA/HSAG catalysts exhibited a totally different thermogravimetric analysis profile (Fig. 9.b). HPA decomposition commences at temperatures around 510 \u00baC, once the combustion of the graphitic support has already started, either for the used or the fresh catalyst. Indeed, this decomposition generates two different combustion rates of the HSAG. The combustion of the formed carbon deposits on the spent catalyst was confirmed from the TGA curve, which shows a weight loss at about 350\u2009\u00b0C, absent in that of the fresh catalyst. This explains the total regeneration of this catalyst with the employed thermal treatment at 400\u00baC. An intermediate behaviour occurred for STA/ZrO2 fresh and used catalyst (Fig. 9.c), which finally means an incomplete recover of the catalytic activity, depicted in Fig S14. In the same manner, the oxidation treatment performed on STA/AC was not successful and the catalytic activity did not recover (Fig. S15).Hence, although SiO2 is a widespread employed support for HPAs, due to its relatively inertness which allows to maintain the former HPA acidity, when applied in the BuOH dehydration reaction an important high deactivation is observed. In our case the low decomposition temperature observed for STA/SiO2 seems not to be operative for the catalyst reactivation. Contrarily, taking advantages of the strong carbon-HPA interactions [83,84], and the thermal stability of graphite under oxidative conditions, it is possible to regenerate the softer coke and to restore the catalytic properties of STA/HSAG catalyst.The successful incorporation and high dispersion of the HPAs on the carbon supports was confirmed by means of N2 physisorption, X-ray diffraction and FTIR-ATR, which evidenced the slight and coherent decrease in the surface area, the absence of crystallinity and the preservation of the Kegging structure, respectively. The spectroscopic technique also revealed a strong interaction between the HPA and the graphitic support on HPA/HSAG samples.As expected, and as revealed by the NH3-TPD profiles, the incorporation of HPAs led to a huge increase in the amount and strength of acid sites, displaying the TPA-based catalysts less amount of total acid sites but higher strength than their STA-based counterparts. Comparison between the explored carbon supports discloses that HPA catalysts supported on AC exhibit a somewhat major amount of acid sites than those over HSAG. This was also evidenced by the higher initial activity displayed by the formers during the 33DM1BN isomerization, which is a sensible reaction to Br\u00f8nsted acid sites of weak or moderate strength.During the gas phase dehydration reaction of butanol, all the explored catalyst systems were highly selective to the target products, linear butenes, and at the same time all were susceptible to deactivation by coke deposition. However, the support\u2019s nature has shown to modulate at some extent the catalytic activity, isomers distributions, stability degree, and regeneration ability. On one hand, STA/SiO2 developed the highest catalytic activity but could not be regenerated by means of combustion. On the other hand, thanks to the strong HPA-carbon interaction, the thermal stability of graphite and HPA along with the deposition of a softer coke, STA/HSAG resulted to be a highly selective and fully regenerable catalyst following well selected oxidation treatments.\nJ.M. Conesa: Conceptualization, Methodology, Investigation, Writing \u2013 original draft. M.V. Morales: Writing \u2013 original draft, Supervision. N. Garc\u00eda-Bosch: Investigation. I. Rodr\u00edguez-Ramos: Writing \u2013 review & editing, Supervision, Project administration, Funding acquisition. A. Guerrero-Ruiz: Conceptualization, Supervision, Funding acquisition. All authors have given approval to the final version of the manuscript.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The Financial support from the Spanish Agencia Estatal de Investigaci\u00f3n (AEI) and the EU (FEDER) (projects PID2020-119160RB-C21 and -C22) is gratefully acknowledged. J.M. Conesa gratefully acknowledge the funding provided by UNED to carry out his Ph.D. (EIDUNED; jconesa61@alumno.uned.es).Supplementary data associated with this article can be found in the online version at doi:10.1016/j.cattod.2023.01.024.\n\n\n\nSupplementary material.\n\n\n\n.", "descript": "\n                  1-butanol dehydration reaction has recently emerged as a sustainable route to produce butenes which can be further oligomerized to be applied as jet fuel. However, the high catalyst deactivation rates observed during this reaction due to coke deposition is still a pending matter. As promising catalysts for this reaction, we have supported two heteropolyacids (HPA), i.e. H4SiW12O40 (STA) and H3PW12O40 (TPA), on two commercial carbon materials: an activated carbon (AC) and a high surface area graphite (HSAG). Aiming to evaluate the role of HPA-support interactions, the STA was also dispersed over metallic oxides of different acidic nature, namely SiO2, Al2O3 and ZrO2. An exhaustive physicochemical characterization revealed that after the HPA dispersion thorough the support, the Keggin structure was maintained and an increase in the amount and strength of acid sites was provoked, but in different degree according to the HPA type and support\u2019s nature. While the TPA-based catalysts developed less quantity of total acid sites, but higher strength than their STA-carbon counterparts, the STA/AC and TPA/AC samples exhibited a slight major amount of acid sites than STA/HSAG and TPA/HSAG. The HPA-support interactions have ultimately modulated to some extent the activity, selectivity, stability and regeneration ability of the synthesized catalysts, when applied in the gas phase butanol dehydration reaction at 275\u00a0\u00b0C. The higher STA decomposition temperature prompted by the graphitic support, among other factors, allowed the total regeneration of the highly active (39\u00a0mmolBuOH\u2219min\u22121\u2219ga.p) and n-butenes selective (>98 %) STA/HSAG catalyst by means of combustion of carbon deposits at 400\u00a0\u00b0C.\n               "}
{"full_text": "Ceramic matrix composites (CMCs) are a frequently used category of materials in nowadays high-tech industries [1,2]. Ceramic-metals, also called cermets, are a prominent subcategory of CMCs, which can be a solution for ceramic-property limitations such as brittle failure, low fracture toughness, and sensitivity to thermal shock [3,4]. They are widely used in space vehicles, gas turbines, automotive parts, slide bearings, etc. [5\u20137]. Alumina is one of the commonly used ceramics in the fabrication of CMCs [8], which is proven to encompass impressive behavior at elevated temperatures, excellent thermal conductivity, high hardness, significant corrosion, wearing and abrasion resistance, and relatively low density (nearly half of that of steel) [9]. Alumina is an ideal material for high-temperature insulators, substrates, and circuit boards [10]. Also, the high hardness and wear resistance of alumina or alumina-based composites, makes them ideal for use in cutting tools or high-resistance coatings, especially when heat resistance is required [11]. Moreover, alumina is biocompatible, which means it can be used in medical implants without causing an immune response. Porous alumina-based implants are used in hip and knee replacements, dental implants, and other medical applications [12,13]. Porous alumina-based materials are also widely used as catalyst converters or to purify air or water [14,15].On the other hand, alumina is an excellent candidate to combine with steel, considering the importance of various types of steel in industries [16]. For instance, SiC/steel composites are almost impossible to fabricate, as the oxide coatings on SiC are not protective enough to shield the ceramic against corrosion caused by the melted metal. Therefore, the dissolution of SiC in the melt occurs. Consequently, alumina as an oxide ceramic is a well-deserved option to compound with high-melting alloys (e.g., steels or Ni-based superalloys). In such composites, the interface of steel-alumina has proven to show satisfactory bonding, with no cracks [17\u201319]. Koopmann et al. [20] also demonstrated that the good interface behavior of alumina-steel can be observed even for additively manufactured samples. Moreover, for high-tech applications, there is an increasing necessity to develop novel materials specifically designed for exclusive purposes [21,22]. To this end, various innovative materials, particularly composites, were introduced through different fabrication techniques [23\u201325]. Considering the broad utilization of alumina in industries, a favorable category of composites can be achieved by developing alumina/metallic composites. It enables us to obtain superior material behavior, encompassing selected properties of alumina and metals (e.g., metals act as the reinforcement phase or a coolant agent in the structure, and alumina adds high abrasion resistance to it, etc.).Such composites can also be produced through powder metallurgy (PM) [26], hot pressing (HP) [27], hot isostatic pressing (HIP) [28], spark plasma sintering (SPS) [29], and casting methods [30]. However, fabrication of the CMCs via conventional techniques can be costly and geometrically limited. Besides, the aforementioned processes sometimes prolong the path to the desired parts. To overcome these barriers, additive manufacturing (AM) techniques can be introduced as an alternative solution to produce CMCs [31]. In these methods, the components are produced through a layer-upon-layer material deposition until the desired shape is made [32]. It is required to study the potential of AM to produce CMCs. The additively manufactured CMC samples were also compared to those fabricated through conventional techniques [33\u201335]. One major AM category is laser powder-bed fusion (LPBF), in which a laser beam is utilized to selectively melt or sinter and fuse the powder particles [36]. The principle of this method is that a thin layer of powder is uniformly deposited on the printing bed through a blade coater. The laser beam scans the selected area via the intended parameters and the desired scanning pattern. The adjacent powders are then fused together under the influence of the heat induced by the laser beam. Subsequently, the printing bed moves downwards on the z-axis, and another layer of powder is coated on the top. This procedure is repeated until the final geometry is obtained. The schematic illustration of the LPBF process is shown in Fig. 1\n. Despite considerable advancement in this area, the LPBF technique is still in the early stages of development. A considerable number of researches were carried out to develop new materials for the LPBF, including iron-based alloys [37,38], aluminum-based alloys [39\u201341], metal-ceramic composites [42\u201344], and metal matrix composites [45\u201348].During the past few years, a limited number of researches are conducted on AM of metal-reinforced CMCs, most of which are associated with AM of WC-Co composites. In literature, LPBF [49\u201351], direct laser deposition [52], binder jetting [53,54], and laser engineering net shaping (LENS) [55,56] AM techniques have been used to fabricate CMCs. Among the mentioned techniques, LPBF includes most of the research. A major problem in LPBF of the metal-reinforced CMCs is the formation of cracks due to the thermal gradients [57]. Grigoriev et al. [50] used LPBF to additively manufacture WC-6 wt% Co samples and reported crack formation, which is attributed to the significant difference between the melting point of the ceramic and the metal. Khmyrov et al. [58]} studied the formation of various phases during the LPBF of WC with different Co contents (6, 50, 72.4, and 75 wt%). They reported the presence of WC and the formation of W2C in the structure for samples with 6 wt% Co. For any other sample, they observed a full dissolution of WC. A small amount of W4Co2C was revealed in samples containing 75 and 72.4 wt% Co, while samples with 50 wt% Co contained W3Co3C.On the other hand, some difficulties are faced in the laser processing of ceramics related to their low optical absorption coefficient and, subsequently, the problematic absorbing ratio of the laser energy during the process. Compounding the ceramic with metal particles is an effective solution to enhance light absorption. Besides, ceramics usually have high melting points and do not melt during laser processing. As a result, they do not have the proper flowability to fill the pores during printing. Adding metallic particles to ceramics can reduce the porosity ratio, as the metals have the potential to melt, flow, and fill the gaps at lower melting temperatures.With regard to the reasons mentioned above, this research aimed to additively manufacture the alumina/standard EOS steel alloy (DS20) composite through the LPBF technique. It is worth noting that parameters such as flowability, density, size, shape, and conductivity of the powder are critical characteristics in the broad implementation of the technique. After accomplishing the manufacturing phase, the samples' microstructure variation, porosity analysis, hardness testing, and compressive strength were addressed. Alumina-based materials have a broad range of industrial applications, especially alumina/metal composites. Previous research by Koopmann et al. [20] reported zirconia-alumina ceramic coating on a steel substrate. However, to our knowledge, this study presents the first development of an alumina/steel ceramic matrix composite using LPBF, demonstrating the feasibility of printing cermets. The advantages of additive manufacturing, including design flexibility, energy efficiency, and reduced post-processing and machining requirements, make this breakthrough approach a promising option for the fabrication of alumina-based materials. This research offers new avenues for the additive manufacturing of ceramic-based materials in general.To prepare the composite powder, an aluminum oxide powder (GF54503557, Aldrich, United States) and a standard EOS steel alloy (DS20, EOS GmbH, Germany) powders were carefully sieved through a mesh size of < 45 \u03bcm to eliminate any agglomerated particles. Using a dry ball milling setup, the powders were mixed at a ratio of 80 wt% alumina and 20 wt% steel for 4 hours. The composition of steel powder was investigated by the inductively coupled plasma\u2013optical emission spectrometry (ICP-OES) analysis through the Sciex Elan 6100 (PerkinElmer, USA) system.A direct metal laser sintering (DMLS) machine (M250 XTENDED DMLS 3D printer, EOS GmbH, Germany) was used to fabricate alumina/steel cylindrical specimens with a diameter and height of 5 and 15 mm, respectively. The optimal printing parameters (Table 1\n), including laser power, laser scanning speed, layer thickness, scanning pattern, and hatch space, were adjusted, starting with the values derived from previous studies [59]. The as-printed specimens were then sintered in a graphite furnace under an inert gas (argon) atmosphere to improve the bonding within the structure. The sintering heating program is presented in Fig. 2\n.After the sintering process, a polymeric resin (Dichtol WTF 1532, Metaplastic, Germany) was used to infiltrate into the cracks and pores of the specimens. According to the data provided by the resin manufacturer, Dichtol polymer has a service temperature of -40 \u00b0C to +300 \u00b0C. In addition, the low viscosity of the resin at room temperature could enhance the infiltration process. For the infiltration process, specimens were immersed in the liquid resin for 60 min and then cured for 4 hours at room temperature according to the guideline provided by the manufacturer. This process was repeated three times to ensure complete penetration of the resin.Simultaneous thermal analysis (STA 504, B\u00c4HR-Thermoanalyse GmbH, Germany) up to 1600 \u00b0C and in an argon atmosphere was used to investigate the steel powder's thermal properties and determine a suitable sintering temperature.Microstructural studies were performed on both as-received powders and processed samples (as-printed, as-sintered, and polymer-impregnated). Field-emission scanning electron microscopy (FE-SEM) (Mira3, TESCAN, Czech Republic) was used to study the polished cross-sections and the fracture surface of the fabricated specimens. In addition, an energy-dispersive x-ray spectrometer (EDS) was utilized to characterize their chemical compositions.Mercury porosimetry distribution analysis was conducted in two different ranges. To investigate the distribution of pores below D: 10 \u03bcm and pores above D: 10 \u03bcm, a Pascal 440 and a Pascal 140 Mercury Porosimeters (Thermo Fisher Scientific, United States) were used, respectively.To evaluate the mechanical properties, cylindrical samples with a diameter and height of 5 and 7.5 mm, respectively, were prepared and subjected to the compression test using a universal testing machine (2000KPX, Instron, United States); the test was conducted in accordance with the ASTM-C1424 standard at a 1 mm/min rate for all samples. The as-fractured samples were studied to investigate the fracture behavior. To check the hardness of the specimens, a Vickers microhardness indenter was used (MMT-X, Matsuzawa, Japan) under a force and duration of 50g and 20 seconds, respectively. The microhardness test was carried out at different locations along the surface of the specimens with a spacing distance of 500 microns between adjacent indentation points.Powder characteristics, including morphology and particle size, were studied using FE-SEM (Fig. 3\n). The as-received alumina particles showed an elongated polygonal shape, and most of the particles were in the range of 30-45 \n\n\u03bc\nm\n\n (Fig. 3(a)). The DS20 steel particles exhibited a spherical morphology with an average particle size below 20 \u03bcm. Some agglomerates could be noted in the initial steel powder, which could be removed by ball milling and sieving (Fig. 3(c)). The relatively small diameter of the steel particles and their large specific surface area led to the enhancement in light absorption properties during DMLS, which raised the heat of the particles and facilitated the sintering properties. Furthermore, the spherical morphology of the metal particles could lead to a smoother and more uniform layer during the powder deposition [60].The composition of DS20 steel powder obtained from ICP-OES analysis is listed in Table 2\n.The homogenous distribution of steel particles in the dominant Al2O3 powder was found using SEM of the as-milled powder mixtures (Fig. 4\n). The even distribution was necessary to increase the laser absorption rate and enhance the mixture's flowing characteristic during the printing process. It could also improve liquid wetting characteristics and reordering of the particles.Different powder ratios were investigated to obtain the optimum mass ratio of each component in the fabricated specimens. Any defective ones with visible cracks or large porosities were ruled out. The printed specimen with the composition of 80 wt% of Al2O3 powder and 20 wt% of the DS20 steel particles showed no visible defects. The DMLS process parameters were also determined. Different parameters, including the laser scan pattern, laser energy density, and heat treatment program, were selected as variables, and optimum parameters were obtained based on the printed specimens' characteristics, which are reported in Table. 2. Then, the DMLS was successfully used to manufacture cylindrical specimens with the mentioned composition. The fabricated cylindrical specimens are shown in Fig. 5\n.The sintering process was then conducted to improve the strength of the as-printed cylinders and relieve their associated residual stresses that formed due to fast localized cooling during printing [61]. The sintering temperature for alumina was determined in different works to be about 1300 \u00b0C to 1600 \u00b0C [62]. However, due to the existence of the metallic component in the CMC structure, the melting point of this component played an important role in determining its sintering temperature. To select a suitable maximum temperature for the sintering phase, the STA analysis was performed on the steel powder, presented in Fig. 6\n. Based on the signals, an endothermic peak is evident at 1370 \u00b0C, possibly associated with the liquid formation. Accordingly, this temperature (1370 \u00b0C) was selected for the sintering process.\nFig. 7\n displays backscattered electron (BSE) SEM micrographs of the polished surface of the samples. The as-printed sample showed uniform distribution with no accumulation of the steel particles over the Al2O3 matrix. Besides, the diffusion of the steel alloys in the Al2O3 matrix is not evident from the images (Fig. 7(b)). The as-printed body is a porous structure and contains micro-cracks mainly formed between the porosities. From the fracture cross-section (Fig. 8\n(a)), it can be observed that these cracks can propagate through the structure when the load is applied. The formation of micro-cracks in the sample is attributed to the brittleness of the alumina matrix and the differences in melting points and thermal expansion coefficients of alumina and steel components [51,63]. Due to the weak adhesion between steel and alumina matrix, steel particles were removed during polishing in some areas; these are distinguished in dark circles in the micrograph. It also can be understood from Fig. 7(c) that some of the alumina particles were melted during the printing process. A similar effect was noted on the fracture surface of the bars, as highlighted in Fig. 8(a). Besides, a long crack formed due to thermal gradients during solidification. Fig. 7(d-f) illustrates BSE images of the sintered sample.The images of the as-printed and as-sintered samples show no discernible difference between them. Due to the existence of the steel component in the fabricated composite and considering the steel area being melted at 1370 \n\n\u00b0C\n\n, liquid phase sintering must have occurred during the heat-treatment to achieve full densification. However, the mentioned phenomenon did not happen because of the poor wettability of the steel and alumina, the high interfacial energy between steel and alumina, and the presence of a thin layer of oxide on the surface of the steel, which prevents melted steel from moving between alumina particles [64\u201366]. From the fracture surface analysis (Fig. 8(c) & (d)), spherical steel particles can still be observed, confirming the stated phenomenon. It is evident from the SEM images of the polished cross-section that the melted metallic particles did not rearrange the alumina particles due to capillary action. On the other hand, it could be noticed that the as-sintered samples have relatively higher densification than the as-printed body. The reason could be associated with the densification of the alumina component during sintering. The onset of sintering temperature for alumina is around 1000 \u00b0C. Hence, alumina particles started to densify at this temperature through the solid-state sintering mechanism. Since the sintering temperature of the CMC was selected to be at 1370 \u00b0C, which was 370 \u00b0C higher than the onset of the sintering temperature of alumina, more densification was expected. However, the complete densification of the CMC was not achieved since the ideal sintering temperature of alumina is between 1500 to 1700 \u00b0C, which was impossible to implement due to the presence of the metal in the structure. Considering the STA analysis of the metallic phase (fig. 6), the formation of gas at temperatures above 1370 \u00b0C could introduce defects into the structure. The incomplete densification of alumina may also be linked to the sintering atmosphere chosen. This is because argon has a lower solubility in alumina than oxygen, which may lead to less efficient sintering [67]. It is worth noting that a noble gas was chosen as the atmosphere due to the presence of metallic particles, and the risk of oxidation during sintering, which could have had a detrimental effect. It should be considered that solidification occurred so fast when the alumina particles were melted during the printing process. So, densification could not be achieved entirely. Therefore, the particles only stuck together. By comparing the steel and alumina interaction for the as-printed body and the sintered samples in Fig. 7(b) & (e), it could be concluded that the pores inside the steel area are removed after sintering.Despite performing the sintering process and the resultant relative increase in the strength of the areas, the samples remained brittle due to the porosities in their structures. So, they could not be used under relatively high mechanical loads. Therefore, an attempt was made to remove the cracks and reduce the porosity in the samples by penetrating a suitable polymer into the matrix to improve their strength. The specimens were immersed in the Dichtol WFT 1532 resin. This polymeric solution was specifically designed to fill the pores (0\u20130.1 mm) and subsequently cause an improvement in the material\u2019s mechanical and corrosion behavior. The micrographs of the samples immersed in the resin, \u201cthe polymer-impregnated sample\u201d, are shown in Fig. 7(g-i). From the polished cross-section, almost difference can be noted between the images of the samples before and after the polymer infiltration process. However, as highlighted in Fig. 8(f), the polymeric solution could close some micro-cracks due to the low viscosity and easy penetration characteristics. Larger cracks and porosities cannot be filled in, but a thin layer on the defects' walls can be formed, improving the properties. It is worth noting that while the sintering phase was unable to eliminate large pores or voids, it was a crucial step that could not be omitted. Laser Powder Bed Fusion (LPBF) only provided initial adhesion between the particles to form the desired geometry. Therefore, the as-printed samples had little strength, and sintering was necessary to enhance the particle bonding before polymer impregnation.EDS was also employed to reveal elemental analysis and distribution of the elements within the microstructure after sintering the printed sample. The EDS analyses in three different regions of the sintered sample are shown in Fig. 9\n. In region A (Fig. 9), the highest weight percentages belong to Fe and Ni elements, which confirms the presence of DS20 steel alloy. This region also has C, O, Al, Si, P, and Cu in low percentages. About 90 wt% consists of Al and O in region B, proving this area comprises alumina. The other local elements are C, Si, P, Fe, Ni, and Cu, which possess 10 wt% of the composition. In region C, which is created during the printing process or due to the removal of steel areas during the polishing of samples, C and O have the highest weight percentages. For polishing, samples were cold-mounted using polymeric materials. This polymer can smear off from the cold mount during the polishing and penetrate through the porosities and cracks. This can be the reason for the existence of high contents of C and O in dark areas. All in all, EDS analyses of the different regions of the surface further revealed the poor diffusion of the molten steel inside the alumina during the printing and sintering processes.EDS x-ray maps of the sintered sample are shown in Fig. 10\n. These maps show the distribution of the elements in the microstructure. As it is clear from the images, the steel alloy elements are distributed randomly and uniformly in the matrix. Furthermore, the semi-dendritic regions are enriched with alumina, which is caused by the non-equilibrium solidification of the composite during the DMLS process.\nFig. 11\n presents the cumulative volume vs. the pore diameter for the as-printed and as-sintered samples before and after the polymer impregnation. The test was conducted in two ranges porosities of >10 \u03bcm and <10 \u03bcm. After sintering, the cumulative volume of the pores below 10 \u03bcm reduced by up to 16%. However, the cumulative volume of pores above 10 \u03bcm increased up to 17%, attributed to the higher thermal expansion of the metal components and causing new defects in the structure. Only a slight change in the overall percentage of porosities could be noted upon sintering the as-printed body, which further ensured the poor wettability of alumina by steel alloy. Fig. 11 also displays that the cumulative volume of porosities below 10 \u03bcm decreased up to 34% after immersing the samples in the polymeric resin, as the tiny pores and microcracks were filled during the infiltration. Moreover, polymer impregnation decreased the porosity related to the pores larger than 10 \u03bcm by 31%. To summarize, the overall porosity percentage in the structure decreased from 36% to 27% after polymer infiltration. It should be mentioned that the goal of the polymer infiltration was to improve the fabricated composite's mechanical properties and corrosion behavior. Polymer impregnation did not aim to remove porosities entirely, as it would significantly increase the composite\u2019s weight. Besides that, it could noticeably reduce the heat transfer coefficient, which is considered a disadvantage as regards the application, requiring wear resistance and could lead to a considerable temperature increase in such processes. Finally, the main reason for the existence of porosities may relate to the irregular shapes of the initial alumina powder, which was also shown by Chen et al. [49]. The consequence of this problematic factor could be partially compensated if the wettability of alumina and the molten steel were sufficient to fill the matrix pores with steel.Compression and Vickers microhardness tests were conducted to investigate the mechanical properties of the printed samples. Fig. 12\n presents the microhardness test results of the as-printed and as-sintered samples. The as-sintered sample recorded higher microhardness values than the as-printed one, which is attributed to defining a proper sintering scenario. Owing to the elevated temperatures during the sintering process, the bonding of the ceramic particles happened suitably. This improved bonding could consequently lead to the reduction of the small pores and an increase in the hardness of the composite.Moreover, the main reason the as-printed sample had more microcracks in the ceramic-metal interaction areas was their lower microhardness values [68]. Besides, the distribution of fine phases in the as-sintered composite was another critical contributing factor to the higher hardness values [57]. The results were consistent with other research works [69]. As it is clear from Fig. 12, the distribution of the microhardness value along the investigated line in the sintered sample is more uniform. This could be attributed to holding the sample at an elevated temperature for a long time in the sintering process [70]. Due to the much lower hardness of the infiltrating polymer and its poor effect on the microhardness of the CMC, the microhardness values were not presented for the polymer-impregnated sample.The results of the uniaxial compression test for both as-sintered and polymer-impregnated samples are presented in Fig. 13\n. As predicted, the polymer-impregnated sample showed a higher compressive strength. The slope of the compression-elongation curves in the elastic region is almost the same for both samples, indicating no considerable change in the compressive elastic modulus. The matrix-reinforcement interfaces and the porosity are two determinative factors for the compressive strength in CMCs. By immersing the as-sintered sample in the polymer, the compressive strength increased from 56 MPa to 120 MPa, indicating a sharp increase (more than twice). Polymer's presence improved the mechanical behavior of the composite by reducing the porosity, voids, and micro-cracks. The process of polymer impregnation managed to enhance the bonding between the ceramic grains by filling the gaps with the polymer, resulting in a stronger structure. The polymer was able to reinforce the ceramic material by providing extra strength and toughness. The polymer acts as a stress-relieving layer that disperses and absorbs stress, preventing cracks from spreading and improving the fracture toughness of the ceramic porous samples [71,72]. The polymer\u2019s role in enhancing ductility, toughness (the area under the stress-strain curve), and elongation is well-known.In summary, this study aimed to develop a composite material of alumina and Fe-Ni (steel) alloy using laser powder bed fusion additive manufacturing technology. The microstructural analysis showed a homogenous distribution of steel particles in the alumina matrix, demonstrating the effectiveness of the mixing strategy. Sintering the samples at 1370\u00b0C improved the Vickers microhardness from approximately 1475 to 1960 HV, indicating enhanced mechanical properties due to better particle bonding. Despite this improvement, the samples still contained porosity and microcracks after sintering. By utilizing polymer impregnation, the overall porosity was reduced from 36 to 27%, microcracks were eliminated, and the compressive strength increased sharply from 56 to 120 MPa, without any considerable weight gain or decrease in thermal isolation.This research presents a practical method for manufacturing alumina-based materials, which have broad applications in areas like the fabrication of electronic components, cutting tools, biomedical implants, and catalyst converters due to their biocompatibility, low density, high hardness, and corrosion and wear resistance. The promising properties of the developed samples suggest that ceramic matrix composites reinforced by particulate metallic materials, in general, could be a promising research direction for materials development in additive manufacturing. Future research could explore the corrosion resistance of these samples and the possibility of scaling up production for industrial use.The authors declare that they have no conflict of interest.", "descript": "\n                  Additive Manufacturing (AM) plays a key role in meeting the vital demands of Industry. The AM industry needs the range of applicable materials to be expanded by conducting research on novel ones. In the present investigation, alumina/Fe-Ni (steel) ceramic matrix particulate composite was fabricated employing laser powder bed fusion (LPBF) additive manufacturing (AM) technology. The quality of the printed samples was associated with the LPBF process parameters, which were optimized for this process. In general, the fabricated samples showed a microstructure of alumina matrix with uniform distribution of steel (Fe-Ni) particles. The as-printed samples exhibited pores. Thus, they were subjected to a sintering heat treatment cycle under an inert atmosphere. Although the sintering cycle considerably increased the average Vickers hardness, pores were not eliminated entirely. Therefore, polymer impregnation of the as-sintered samples was carried out to reduce porosities and microcracks. The mercury porosimeter showed that the porosity decreased sequentially after sintering and polymer impregnation. In addition, mechanical investigations revealed that the polymer impregnation improved the compressive strength of the sintered samples (from 56 to 120 MPa). Alumina-based materials find wide applications in various fields, including the manufacturing of electronic components, cutting tools, biomedical implants, and catalyst converters, owing to their low density, high hardness, wear and corrosion resistance, and biocompatibility. This study presents a viable approach for the fabrication of these materials, with developed samples exhibiting promising properties. The study emphasizes the potential of additive manufacturing as an approach for the fabrication of ceramic matrix composites reinforced with metallic particulates in future research.\n               "}
{"full_text": "Meta data for XRD, PDF and XPS is available at DOI:10.17028/rd.lboro.20170784.The consumption of fossil fuels, such as oil and gas, has led to the release of large amounts of greenhouse gases which is resulting in severe environmental problems [1]. Therefore, the use of a clean energy carrier such as hydrogen is important, due to reduced CO2 emissions from its combustion and utilisation. However, this is dependent on the production method of said hydrogen. This has led to the description of hydrogen and its production methods by colours. The colour chosen is dependent on the CO2 emissions of the process, with steam methane reforming classed as grey hydrogen, due to the release of excess CO2 [2]. Potentially, renewable hydrogen can be produced from sustainable biomass sources [3]. Whilst debated, hydrogen production from biomass can be considered as green hydrogen due to the carbon neutrality of the overall process [2,4]. One such production method, which has generated much interest, is the aqueous phase reforming (APR) process, developed by Dumesic and co-workers [5\u20137].APR utilises a range of waste aqueous phase oxygenates derived from biomass sources, such as ethanol, methanol, ethylene glycol and glycerol [8\u201310]. APR is advantageous when compared to traditional methane reforming and waste oxygenate steam reforming due to the low operating temperature (200\u2013250\u00a0\u00b0C), intermediate pressures (15\u201350\u00a0bar), and no requirement to vaporise the solvent which results in a lower energy demand and thermodynamically favours the water gas shift (WGS) reaction [11\u201313]. The latter reaction results in low CO concentrations in the effluent, as required for many industrial applications, when compared to the traditional methane steam reforming production method [14]. Other sustainable methods that reduce the CO concentration in the effluent include sorption enhanced glycerol steam reforming with in-situ CO2 removal [15,16].Many active metals have been applied to the APR process, with Pt being a promising choice of active metal due to its high activity for the WGS reaction, C\u2013C bond scission, and low-methanation activity [17\u201319]. Further control over reaction pathways and increased hydrogen selectivity can be achieved through catalyst design and development of stable support materials [20\u201322]. Basic materials have been shown to facilitate the WGS reaction, which leads to higher hydrogen selectivity. However, these materials suffer from low hydrothermal stability and undergo restructuring and phase changes resulting in catalyst deactivation [23,24]. In contrast, acidic materials such as zeolites, favour dehydration pathways and increased alkane production [25,26]. Therefore, designing stable and favourable pathway promoting support materials is key for the viability of the APR process.Perovskites with the structural formula AMO3 have high thermal stability and wide structural versatility, which has led to a range of applications [27\u201329]. Previously, we have applied Pt/LaMO3 (where M\u00a0=\u00a0Al, Cr, Mn, Fe, Co, Ni) catalysts in the APR of glycerol, and most of the materials, apart from Pt/LaCrO3, were found to be undergo phase transformation under the reaction conditions [30]. The formation of hexagonal LaCO3OH phases, alongside M site oxides, decorated with Pt nanoparticles, were observed in these catalysts and these phases were found be active and stable catalysts in their own right. Mao and co-workers verified this finding through an extended APR reaction under flow conditions with methanol as the feedstock [31]. Therefore, the perovskite can be considered as a precursor to prepare stable and active catalysts [32]. Interestingly, these findings are contradictory to previous reports of perovskite phase stability in Ni/LaAlO3 catalysts in APR, with comparable glycerol concentration and reaction temperatures [33,34].Given this discrepancy, we have investigated the effect of the phase purity of the LaAlO3 support through alteration of the calcination conditions required to produce it. The catalytic performance and subsequent stability of the materials when applied to APR, hydrothermal conditions and acidic reaction intermediates provided insight into the importance of phase purity.Al(NO3)3.9H2O (99+% Acros Organic), La(NO3)3.6H2O (99.9% Alfa Aesar), PtCl4 (99.99% Alfa Aesar), glycerol (99% Fisher Chemical), lactic acid (\u226588% Fisher Chemical), LaB6 (99.5% Thermo Scientific), citric acid monohydrate (99.9% VWR Chemicals), ammonia (32% v/w VWR Chemicals). All chemicals were used without further purification.LaAlO3 perovskite materials were prepared by sol-gel combustion method [35]. La(NO3)3.6H2O (6.073\u00a0g) and Al(NO3)3.9H2O (5.262\u00a0g) were used in stoichiometric amounts. Citric acid (11.790\u00a0g) was then added in a 2:1 ratio to metal nitrates and dissolved in deionised water (15\u00a0mL). The pH of the resultant solution was adjusted to 7 using aqueous ammonia solution (3\u00a0M) and aged at 130\u00a0\u00b0C until gel formation. The gel was then combusted at 400\u00a0\u00b0C for 10\u00a0min and further calcined for 2\u00a0h at 700, 900 or 1100\u00a0\u00b0C and labelled LaAlO3-700, LaAlO3-900, and LaAlO3-1100.Preparation of 1\u00a0wt% Pt/LaAlO3-c (c=700, 900, 1100) catalysts by a conventional wet impregnation method was as follows: Prepared PtCl4 solution (Pt content: 4.8\u00a0mg/mL; 2.082\u00a0mL) and deionised water was dispensed to give an overall solution of 16\u00a0ml. The mixture was stirred (800 RPM) at 60\u00a0\u00b0C. The support (0.99\u00a0g) was periodically added slowly over a period of 10\u00a0min. The resulting slurry was stirred for a further 15\u00a0min before heating to 95\u00a0\u00b0C and dried overnight. The dried powder was then ground and calcined in air at 450\u00a0\u00b0C (2\u00a0h, ramp rate: 10\u00a0\u00b0C/min).X-ray powder (XRD) patterns of the materials and catalysts were recorded with an LaB6 internal standard (33\u00a0wt%) using a Bruker d8 discover operating at 35\u00a0kV and 40\u00a0mA with a monochromated Co source (\u03bb\u00a0=\u00a01.79\u00a0\u00c5) and a Vantec detector (scan range: 20\u2013100\u00b0; step size: 0.014\u00b0; step count: 1 s unless specified). Patterns were matched to ICDD PDF database patterns, the list of database patterns is given in Table\u00a0S1. The refinement of the fresh Pt/LaAlO3-\n\nc\n\n was performed by the Rietveld method, using TOPAS v5 software and ICSD database crystal patterns given in Table\u00a0S2. Atom positions, occupancies, and thermal parameters were not refined.N2 adsorption experiments were performed at\u00a0\u2212196\u00a0\u00b0C using a Micromeritics Gemini VII to obtain surface areas determined by BET method. Before measurements, the required amount of sample was measured and degassed under vacuum overnight at 90\u00a0\u00b0C.Thermogravimetric analysis (TGA) of the combusted gel of the LaAlO3 precursor and LaAlO3-\n\nc\n\n was carried out using a TA SDT Q600 to investigate the formation temperature of the perovskite. The samples were heated in air (ramp rate: 10\u00a0\u00b0C/min) to 1200\u00a0\u00b0C and 1000\u00a0\u00b0C respectively.CO chemisorption measurements were performed using an Altamira AMI-300Lite. Approximately 100\u00a0mg of catalyst sample was loaded in-between quartz wool and reduced under 5% H2/Ar flow (50 SCCM) at 240\u00a0\u00b0C (2\u00a0h, ramp rate: 10\u00a0\u00b0C/min). For analysis, conducted at RT, the sample was titrated with 10% CO/Ar by pulsing through a 574\u00a0\u03bcL sample loop.ICP-AES experiments were conducted to determine the extent of metal (La and Pt) leaching into reaction filtrates and Pt weight loadings in the fresh catalysts using an Agilent 4210\u00a0MP-AES fitted with a SPS4 autosampler. Attenuated total reflection infrared spectroscopy (ATR-IR) was collected using a Shimadzu IR Affinity-1 fitted with an ATR stage on fresh and used Pt/LaAlO3-\n\nc\n\n. The spectra were recorded between 340 and 4700\u00a0cm\u22121 with a 2\u00a0cm\u22121 resolution for 60 scans for background and spectra.XPS analysis was performed using a Thermo NEXSA XPS fitted with a monochromated Al k\u03b1 X-ray source (1486.7\u00a0eV), a spherical sector analyser and 3 multichannel resistive plate, 128 channel delay line detectors. All data was recorded at 19.2\u00a0W and an X-ray beam size of 200\u00a0\u00d7\u00a0100\u00a0\u03bcm. Survey scans were recorded at a pass energy of 160\u00a0eV, and high-resolution scans recorded at a pass energy of 20\u00a0eV. Electronic charge neutralisation was achieved using a Dual-beam low-energy electron/ion source (Thermo Scientific FG-03). Ion gun current\u00a0=\u00a0150\u00a0\u03bcA. Ion gun voltage\u00a0=\u00a045\u00a0V. All sample data was recorded at a pressure below 10\u22128 Torr and a room temperature of 294\u00a0K. Data was analysed using CasaXPS v2.3.19PR1.0. Peaks were fit with a Shirley background prior to component analysis. Line-shapes of LA (1.53243) were used to fit components.Total scattering X-ray data were collected on three LaAlO3 (700\u00a0\u00b0C fresh, 1100\u00a0\u00b0C LA/Glycerol and 700\u00a0\u00b0C LA/Glycerol/CO2) samples using the I15-1 beamline at the Diamond Light Source (Didcot, UK). The samples were loaded into 1.5\u00a0mm borosilicate capillaries with scattering collected on the sample, container and empty beamline using an energy of 76.69\u00a0keV (\u03bb\u00a0=\u00a00.161669\u00a0\u00c5). Corrections for background, fluorescence, absorption, multiple scattering and Compton scattering were performed using GudrunX [36]. The corrected scattering was Fourier transformed to obtain the pair distribution function (PDF) using a Q-range 0.4\u00a0<\u00a0Q\u00a0<\u00a030\u00a0\u00c5\u22121. PDF were refined using TOPAS v7 and ICSD database crystal patterns given in Table\u00a0S2 [37]. Instrumental parameters obtained by Rietveld refinement of a standard Si 640f yielding a dQ\u00a0=\u00a00.0605. PDF refinements were conducted using a fixed dQ and refined lattice parameters, scale, thermal parameters, and spherical dampening.A FEI Tecnai F20 field emission gun transmission electron microscope (FEG-TEM) was used to image the Pt nanoparticles in conventional high resolution TEM mode. The images were recorded using the Gatan Orius SC200 CCD camera equipped to the FEG-TEM at full CCD size of 2048\u00a0\u00d7\u00a02048 pixels. Enough frames from typical areas were recorded to ensure >300 particles are available to render a statistical representation of the Pt particle size. TEM specimens were prepared by drop casting of catalysts powder suspension in isopropanol onto standard holey carbon (300 mesh) TEM support grids.Support and Pt catalyst testing was carried out using a Parr 5500 series bench top micro reactor (50\u00a0mL) equipped with a Parr 4848 reactor controller system. The catalysts were tested using a standard procedure: 10\u00a0wt% glycerol solution (20\u00a0mL) and the catalyst (60\u00a0mg) were loaded into the autoclave and the system was purged multiple times with argon. The reaction was then carried out for 2\u00a0h at 240\u00a0\u00b0C, 42\u00a0bar, 1000 RPM. The gas products (H2, CO, CO2, and CH4) were collected and analysed using an Agilent 8860\u00a0GC equipped with a TCD detector and a Shincarbon ST column. The reactant and products in the liquid phase were analysed by HPLC using a Hitachi Chromaster equipped with an Agilent Metacarb 67H column and a refractive index detector. The liquid phase products observed were Lactic Acid (LA), Ethylene Glycol (EG), Hydroxyacetone (HA), 1,2-Propanediol (1,2-PD), 2-Propanol (2-P), 1-Propanol (1-P), and Ethanol (E). Simulated reactions without dispersed Pt species were undertaken, using the standard procedure, with LaAlO3-\n\nc\n\n (120\u00a0mg) and 10\u00a0wt% glycerol or 1\u00a0wt% LA/9\u00a0wt% Glycerol mixture (20\u00a0ml) and additional 1\u00a0bar of PCO2.Calculations for reactions were carried out as follows:\n\n\n\nConversion\n\nX\n\n(\n%\n)\n\n=\n\n\n(\n\n\n\n[\n\ng\nl\ny\nc\ne\nr\no\nl\n\n]\n\n\ni\nn\n\n\n\u2212\n\n\n[\n\ng\nl\ny\nc\ne\nr\no\nl\n\n]\n\n\no\nu\nt\n\n\n\n)\n\n\n\n[\n\ng\nl\ny\nc\ne\nr\no\nl\n\n]\n\n\ni\nn\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n\n\nConversion\n\nto\n\ngas\n\n\nX\n\ng\na\ns\n\n\n\n(\n%\n)\n\n=\n\n(\n\n\n\u03a3\n\n(\n\nm\no\nl\n\ng\na\ns\n\np\nr\no\nd\nu\nc\ne\nd\n\n)\n\n\n\n\u03a3\n\n(\n\nm\no\nl\n\nt\nh\ne\no\nr\ne\nt\ni\nc\na\nl\n\ng\na\ns\n\n)\n\n\n\n)\n\n\u00d7\n100\n\n\n\n\n\n\n\n\nTurnover\n\nfrequency\n\nT\nO\nF\n\n\n(\n\nh\n\n\u2212\n1\n\n\n)\n\n=\n\n(\n\n\n(\n\n\n(\n\n\n\n[\n\ng\nl\ny\nc\ne\nr\no\nl\n\n]\n\n\ni\nn\n\n\n\u2212\n\n\n[\n\ng\nl\ny\nc\ne\nr\no\nl\n\n]\n\n\no\nu\nt\n\n\n\n)\n\n/\nm\no\nl\n\nP\nt\n\n)\n\n\nt\ni\nm\ne\n\n\n)\n\n\n\n\n\n\n\n\nHydrogen\n\nselectivity\n\nS\n\n\nH\n2\n\n\n\n%\n\n=\n\n\n\nm\no\nl\n\n\nH\n2\n\n\np\nr\no\nd\nu\nc\ne\nd\n\n\n\u03a3\n\n\nm\no\nl\n\ng\na\ns\n\np\nr\no\nd\nu\nc\nt\ns\n\n\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n\nCarbon\n\nproduct\n\nselectivity\n\nS\n\n(\n\nC\ni\n\n)\n\n\n(\n%\n)\n\n=\n\n(\n\n\n(\n\nm\no\nl\n\nP\ns\n\n\u00d7\n\nC\nn\n\n\n)\n\n\n(\n\n\u2211\n(\nm\no\nl\n\n\nP\ni\n\n\u00d7\n\nC\nn\n\n\n)\n\n\n)\n\n\u00d7\n100\n\n\n\nwhere Ps\u00a0=\u00a0specified carbon product; Pi\u00a0=\u00a0carbon product; Cn\u00a0=\u00a0carbon number.\n\n\n\nHydrogen\n\nformation\n\nrate\n.\n\nr\n\n(\n\nH\n2\n\n)\n\n\n(\n\n\u03bc\nm\no\nl\nm\ni\n\nn\n\n\u2212\n1\n\n\n\ng\n\nc\na\nt\n\n\n\u2212\n1\n\n\n\n)\n\n=\n\n\n\u03bc\nm\no\nl\n\n(\n\nH\n2\n\n)\n\n/\nmin\n\n\ng\n\n(\n\nc\na\nt\na\nl\ny\ns\nt\n\n)\n\n\n\n\n\n\n\nTGA analysis of the combusted gel powder precursor of LaAlO3, given in Fig.\u00a01\na, was used to determine the formation temperature of the perovskite. The TG curve shows continuous weight loss (7.0\u00a0wt%) up to \u223c300\u00a0\u00b0C followed by a major weight loss (37.4\u00a0wt%) between 300\u00a0\u00b0C and 600\u00a0\u00b0C, with a final further minor weight loss (7.3\u00a0wt%) beginning at \u223c650\u00a0\u00b0C. Above 900\u00a0\u00b0C, no further weight loss is noted up to 1200\u00a0\u00b0C. The first weight loss can be identified as the loss of water through dehydration; the major weight loss can be ascribed to the decomposition of precursors and formation of an intermediate decomposition product. The final recorded high temperature weight loss suggests a removal of impurity amorphous carbonate phases (further evidence from FTIR infra vide) [38]. The amorphous carbonate materials arise from the synthesis procedure and the use of citric acid as a carbon fuel source. Calcination temperatures were then chosen either side of the final weight loss for investigation. Further TGA analysis of the LaAlO3-\n\nc\n\n materials, given in Fig.\u00a01b, shows the presence of a small amount of amorphous impurity (4.8\u00a0wt%) in the LaAlO3-700 samples that is not present in the materials (LaAlO3-900 and LaAlO3-1100) calcined at a higher temperature. However, the final weight loss of amorphous content in the LaAlO3-700 was less than the original 7.2\u00a0wt% seen from the TGA of the combusted gel precursor, suggesting that extended isotherms at 700\u00a0\u00b0C could remove more of this impurity.Powder XRD patterns of the synthesised Pt/LaAlO3-\n\nc\n\n are given in Fig.\u00a02\na. For each of the catalysts, no peaks for platinum species PtOx or Pt(0) were found, due to the low loading and the possible small and well-dispersed nature of the nanoparticles. The predominant phases in all the materials were the perovskite LaAlO3 (ICDD PDF 31-0022) phase with no other crystalline phases present, apart from the internal standard LaB6 (ICDD PDF 34-0427). The internal standard LaB6 (33\u00a0wt%) was used to determine the crystallinity of the samples as a ratio of LaAlO3(100): LaB6(110) reflections, which are given in Table\u00a01\n along with the physiochemical properties of the materials. It is evident that significant amorphous content was present in Pt/LaAlO3-700, which then notably decreases as the calcination temperature was increased, as seen from LaAlO3:LaB6 peak ratios changing from 1.1 in Pt/LaAlO3-700 to 2.3 and 2.4 for Pt/LaAlO3-900 and Pt/LaAlO3-1100, in agreement with findings from the TGA. However, it should be noted, that a decrease in crystallinity was seen in LaAlO3-700 (LaAlO3:LaB6 2.0 to 1.1) upon impregnation with PtCl4 solution, showing that the acidity of the Pt precursor solution could facilitate loss of crystallinity in the material. The decrease in crystallinity was not seen in the higher calcined materials suggesting higher stability of these materials. The analysis presented makes it clear that simple XRD database pattern matching, without an attempt to quantify amorphous content, leads to a lack of full understanding of the catalyst support. The LaAlO3-\n\nc\n\n particle size was calculated using the Rietveld method and calculation described by Balzar et\u00a0al., assuming monodisperse distribution of particles and negligible strain [39]. The calculated fits are given in Figs.\u00a0S1\u2013S3 and Tables\u00a0S3\u20135. As anticipated, the calculated particle size moderately increased with calcination temperature due to sintering of the particles at higher temperatures, further leading to a reduction in surface area (Table\u00a01).To further elucidate structural defects in the LaAlO3-700 samples, PDF refinement was used to compare known models and experimental PDF data to identify the phases present within the sample. A good agreement between the experimental PDF and the model was found and is shown in Fig.\u00a02b. For this refinement, it was found that a two-phase model, with phases LaAlO3 and La2(CO3)2OH2, was required to obtain a good fit (Rwp\u00a0=\u00a018.24%), with each phases scale allowed to refine, to obtain 52.8% and 14.6% respectively. This gives confirmation that the major phase formed in reaction is LaAlO3. However, this phase alone is not sufficient to fully explain the structure, which is in agreement with the lower crystallinity (LaAlO3:LaB6 ratio\u00a0=\u00a02.0) in the XRD patterns. Upon addition of La2(CO3)2OH2 to the refinement, the Rwp was improved by 1\u20132%, with the features in the r\u00a0=\u00a01\u201310\u00a0\u00c5 showing a much-improved fit to the peak shapes, therefore suggesting that a carbonate phase is also present in the 700\u00a0\u00b0C calcined material. Repeating the multi-phase refinement with other lanthanum carbonate containing species (LaC2O2 and La2O2CO3; Table\u00a0S6, Fig.\u00a0S4) also yields an improved model fit to the experimental data. It should also be noted that the existence of lanthanum carbonate species also suggests the existence of equivalent amounts of Al2O3 in the system, albeit at a lower % scale due to atomic mass and hence are not refined. This analysis shows that whilst PDF can confirm the presence of lanthanum carbonate phases and incomplete perovskite formation, it is difficult to deconvolute the phases and PDF alone is not a sufficient technique to definitively identify the carbonate phases present.Further evidence of the nature of the amorphous impurity was provided by IR, shown in Fig.\u00a03\n, with bands at 1405 and 1477\u00a0cm\u22121 associated with the \u03bd\n3 mode of carbonate only being seen in Pt/LaAlO3-700. The moderate splitting of the \u03bd\n3 mode (\u039472 cm\u22121) shows that the D3h symmetry of the carbonate anion, while not completely retained, has not been significantly lowered. This indicates that there is a minor interaction between the metal centre and carbonate in Pt/LaAlO3-700 suggesting carbonate species that are not free. Note as a point of comparison, the crystalline hexagonal LaCO3OH, given in Fig.\u00a0S5, has multiple stretching bands for carbonate at 1510, 1430, 1081, 870, 846, 725, and 680\u00a0cm\u22121 with \u03bd\n3 splitting (\u039480 cm\u22121) not dissimilar to the Pt/LaAlO3 sample, however, there are multiple other carbonate bands in the crystalline LaCO3OH that are not seen in Pt/LaAlO3-700 [40]. The IR therefore supports the incomplete formation of perovskite, as shown in PDF and the existence of amorphous La carbonate species.Pt dispersion measurements by CO chemisorption, show poor dispersion (3%) for the Pt/LaAlO3-700 compared to the Pt/LaAlO3-900 (48%) and Pt/LaAlO3-1100 (32%). Pt particle sizes were also determined from TEM micrographs, the values of which are given in Table\u00a01, with histograms and micrographs in Fig.\u00a0S6, and show similar particles sizes for each catalyst but with an increased standard deviation for Pt/LaAlO3-700. The similarity in particle size is in disagreement with the Pt dispersion measured in the Pt/LaAlO3-700, and this suggests that whilst there are small particles of Pt dispersed on LaAlO3-700, there are larger particles, as suggested by the large standard deviation, of Pt that are not imaged in the TEM micrographs. Alternatively, Pt particles supported on LaAlO3-700 may be highly unstable and sinter during the reduction process prior to CO chemisorption. Despite discrepancy in Pt particle dispersion, wt% loadings determined by MP-AES show all the catalysts are 1\u00a0wt% loading within error. The determined values are given in Table\u00a01.XPS analysis of the Pt 4f, Al 2p and O1s levels of fresh Pt/LaAlO3-\n\nc\n\n are given in Fig.\u00a04\n. Atom percent and component binding energy is given in Table\u00a0S7. The Al 2p region and Pt 4f region overlap in energy and it should be noted that shifts in Al 2p binding energy can be caused by oxidation number, ligand type and coordination, and therefore the chemical shifts in binding energies of oxides, hydroxides, and oxyhydroxides can be difficult to deconvolute [41,42]. All the materials show doublet peaks which correspond to Al-O, similar to those reported for LaAlO3 and Al2O3 [30,43]. The Pt/LaAlO3-700 material shows doublet peaks at 72.57\u00a0eV, which can be assigned to PtO [44]. As well as PtO, Pt/LaAlO3-900 and Pt/LaAlO3-1100 also show doublet peaks at 74.27 and 74.57\u00a0eV respectively, corresponding to PtO2 species [44]. The contribution of higher oxidation state Pt species indicates a different and stronger interaction and stabilisation of Pt species in the higher calcined LaAlO3 materials when compared to the Pt/LaAlO3-700 [45,46]. Pt/LaAlO3-700 also shows a lower at.% (0.54\u00a0at.%) when compared the Pt/LaAlO3-900 (1.51\u00a0at.%) and Pt/LaAlO3-1100 (0.77\u00a0at.%) agreeing with lower Pt dispersion and potentially larger particles in the catalyst.The O 1s levels for each of the materials can be considered as a combination of Olattice and Oads. Olattice encompasses M\u2212O, M\u2013CO3, and M\u2013OH that can be present in the lattice [47]. Binding energies for metal hydroxides and carbonates are similar and therefore are not distinguished in peak fitting due to the potential for both components in the materials. Oads can include adsorbed oxygen from various species including hydroxyl, carbonate, water and bound reaction species [48,49]. The O 1s region of the fresh Pt/LaAlO3-700 show three peaks at 529.21, 530.66, and 532.24\u00a0eV corresponding to M\u2013Ox, M\u2013O, and M\u2013CO3/M\u2013OH respectively and the assignment remains the same for the Pt/LaAlO3-900 and Pt/LaAlO3-1100. However, a higher contribution for the M-CO3/M\u2013OH peak (13.72\u00a0at.%) is seen with Pt/LaAlO3-700 than with the higher calcined materials (8.83\u20139.11\u00a0at.%), possibly due to the presence of the amorphous carbonates confirmed with IR and PDF. The La 3d levels of the fresh materials, given in Fig.\u00a0S7, have doublet peaks with a multiplet splitting of 3.9\u00a0eV which can be assigned to La(OH)3 and is reported in single oxide and perovskite materials [50,51].In summary, calcination of LaAlO3 precursor at 700\u00a0\u00b0C leads to a crystalline perovskite phase, however evidence from LaB6 doping, PDF, IR, and XPS shows incomplete perovskite formation from the precursors with amorphous lanthanum carbonate impurities. These impurities can be removed by calcination at higher temperatures albeit at the expense of surface area and LaAlO3 particle sintering. PtOx nanoparticles were successfully dispersed on the LaAlO3 support materials with potentially residual amounts of Cl\u2212 arising from the wet impregnation synthesis method, which unfortunately were difficult to quantify due to the overlap of Cl 2p and La 4d features in the XPS. The phase purity affects the dispersion and speciation of Pt nanoparticles on the surface of the support, with a reduced support interaction for Pt/LaAlO3-700.The catalytic performance of the Pt/LaAlO3-\n\nc\n\n catalysts was investigated in the APR of 10\u00a0wt% aqueous glycerol under optimised conditions in a batch reaction as determined by Subramanian et\u00a0al. for Pt/\u03b3-Al2O3 [18]. The catalyst performance over 2\u00a0h reaction times is shown in Table\u00a02\n. The catalytic active site can be regarded as the Pt nanoparticles, due to the perovskite materials LaAlO3-\n\nc\n\n not showing any activity for the APR reaction [30].The catalyst performance was found to be better for catalyst supports that had been subjected to higher calcination temperature during the perovskite synthesis, with the highest glycerol conversion being with the Pt/LaAlO3-1100 at 20.4% (TOF\u00a0=\u00a0686.4 h\u22121). The modestly higher reactivity of Pt/LaAlO3-1100 vs Pt/LaAlO3-900, whist the latter had a higher initial Pt dispersion, showed that there is no strong correlation with this parameter and activity. The Pt/LaAlO3-700 had the lowest conversion (5.4%) showing that the amorphous impurity has affected the catalyst activity. The H2:CO2 ratios for the Pt/LaAlO3-900 and Pt/LaAlO3-1100 were lower than the ideal ratio (2.33) suggesting competing hydrogen consumption reactions of unsaturated intermediates, as observed by Wawrzetz et\u00a0al. for a Pt/\u03b3-Al2O3 catalyst [52]. The higher than ideal ratio observed for the Pt/LaAlO3-700 catalyst suggest other hydrogen production reactions, such as dehydrogenation, are promoted over the reforming reaction at low conversions, which was observed for Pt/LaMO3 catalysts with low activity [30]. The combination of hydrogen production and consumption reactions lead to hydrogen formation rates that are similar for each Pt/LaAlO3 catalyst despite differences in H2:CO2 ratios.The carbon product selectivity for the observed liquid and gas phase carbon products is shown in Fig.\u00a05\n. The selectivity profile was, within error, identical between Pt/LaAlO3-900 and Pt/LaAlO3-1100, while Pt/LaAlO3-700 gave higher lactic acid (LA) and ethylene glycol (EG) selectivity at the expense of hydroxyacetone (HA) and ethanol. HA is proposed as the first intermediate from glycerol dehydration, which is a more reactive substrate than glycerol, and can be readily converted to 1,2-PD. LA is also produced from HA and the dehydrogenated first intermediate glyceraldehyde and LA is achieved in high selectivity over the Pt/LaAlO3-700, suggesting dehydration/dehydrogenation reactions are favoured at low conversions over reforming [21,52]. This agrees with the gas analysis suggesting hydrogen production reactions are favoured over reforming and agrees with previous studies on Pt/LaMO3 catalysts [30]. Minimal amounts of CO were recorded for all catalysts suggesting high WGS shift activity of the catalysts under APR conditions [11,53].MP-AES analysis of the reaction effluents, shown in Table\u00a03\n, showed limited leaching of Pt during the reaction suggesting strong metal support interaction between the particles and the support material. Leaching of La was highest for Pt/LaAlO3-700 and decreased in-line with increasing calcination of the perovskite suggesting higher stability. However, the high percentage of La leaching is still present in the Pt/LaAlO3-1100 sample, suggesting dissolution of the perovskite structure in all samples. Al leaching is not reported due to its stability under acidic media and the formation and restructuring of Al2O3 to boehmite under reaction conditions [24].XRD patterns, with 33\u00a0wt% LaB6, of the recovered Pt/LaAlO3-\n\nc\n\n after 2\u00a0h APR reaction are given in Fig.\u00a06\na. Perovskite LaAlO3 is the dominant crystalline phase in the 900 and 1100 samples, with trace LaCO3OH (ICDD PDF 26\u20130815) in the 1100 sample. The small amount of LaCO3OH can be correlated to the higher activity of the Pt/LaAlO3-1100, with higher turnover for WGS and hence increased CO2 production. The XRD pattern for the Pt/LaAlO3-700 also shows loss of perovskite phase crystallinity and the formation of hexagonal LaCO3OH phase and La2O(CO3)2. \n\nx\n\n H2O (ICDD PDF 28\u20130512) phase, which can be seen in Fig.\u00a06b. The lanthanum oxide carbonate phase is a precursor carbonate phase that, under hydrothermal reaction conditions, dissolves and recrystalises in the eventual formation of LaCO3OH. The loss of crystallinity within all samples, evidenced by the reduction of the apparent LaAlO3:LaB6 ratios (Table\u00a03), agrees with the AES results that significant leaching has occurred during the initial 2\u00a0h reaction. The minimal evidence of crystalline by-phases post reaction for the 900 and 1100 samples explains the apparent contradiction in the literature surrounding LaAlO3 stability, i. e that perovskite decomposition produces amorphous phases not detectable by XRD. However, in the IR spectrum of all the recovered catalysts, given in Fig.\u00a0S8, show bands at 1562, 1426, and 1406\u00a0cm\u22121 which are assigned to lower symmetry v\n3 stretching modes which both relate to crystalline LaCO3OH (Fig.\u00a0S5) and are indicative of the formation of this species. Importantly, it should be noted that bands observed at 3309, 3088, 1650 and 1066\u00a0cm\u22121 corresponding to O-H, C-H, C=O, and C-O stretches respectively as well as carbonate bands and this suggests binding of organic carbon species (CxHyOz) on the catalyst surface and potential blockage of sites.The Pt particle size, size distributions of which are given in Fig.\u00a0S9 and values in Table\u00a03, show an increase in Pt particle size for all samples when compared to the fresh samples. This is not without precedent, as Pt particle sintering, blockage of sites and particle migration is common under APR conditions and support material phase transformation [21]. The reduced stability of Pt/LaAlO3-700 can be shown with larger Pt particles and wider standard deviation than the other two samples. Evidence of the formation of poorly crystalline phases and subsequent Pt particle migration and are also seen in TEM images of Pt/LaAlO3-1100, given in Fig.\u00a0S10, which confirm restructuring of the catalyst is occurring within the reaction timeframe.XPS analysis of the used catalysts confirm multiple changes and reconstruction on the catalysts surface upon catalytic APR testing. The O 1s and Pt 4f/Al 2p levels are given in Fig.\u00a07\n and the surface analysis data for the used catalysts is given in Table\u00a0S8. In the Pt 4f/Al 2p region of all catalysts, a shift in the binding energy of Pt doublet peaks show reduction of the Pt species by reaction products, such as evolved H2(g), to Pt(0) from PtO and PtO2 [44,54]. A shift in binding energy is also noted for each catalyst in the Al-O doublet peaks, possibly corresponding to hydroxylation to Al species to AlO(OH) due to the acidic conditions of APR [24,55]. The O 1s region also shows higher energy peaks at 532.60\u2013533.5\u00a0eV for each sample which correspond to adsorbed oxygen species. The presence of Oads correlates with the IR spectra showing bound surface species, which can lead to site blocking and catalyst deactivation. The La 3d levels, given in Fig.\u00a0S11, show changes in the La environment of Pt/LaAlO3-700 and Pt/LaAlO3-1100 with a shift in the binding energy and reduction of multiplet splitting from 3.9 to 3.5\u00a0eV, which can be assigned as changes from La(OH)3 to La2(CO3)3 [30,56]. This is consistent with the dissolution of perovskite phase and eventual formation of LaCO3OH phase. The at.% of La also decreases and Al at.% increases consistent with La leaching in the AES.To further confirm the structural evolution of Pt/LaAlO3 into Pt/LaCO3OH-AlO(OH) under reactions conditions, Pt/LaAlO3-1100 was chosen to be tested under APR conditions at an extended reaction time (4\u00a0h). The XRD pattern, given in Fig.\u00a0S12, of the recovered catalyst mixed with 33\u00a0wt% LaB6 show clear reflections for the crystalline hexagonal LaCO3OH phase alongside little amounts of the LaAlO3 phase.To further elucidate the breakdown and restructuring of the perovskite LaAlO3 materials, simulated reactions at standard reaction conditions (240\u00a0\u00b0C, 2\u00a0h, 42 bar, 1000 RPM) without impregnated Pt nanoparticles were undertaken. Reactions were chosen to investigate the effect of reaction conditions and of acidic products. Initially, LaAlO3-\n\nc\n\n was tested with 10\u00a0wt% glycerol and the XRD patterns, given in Fig.\u00a0S13, show little changes in the 900 and 1100 samples, whereas the 700 sample has a small amount of lanthanum oxide carbonate phase present as well as LaAlO3 and LaCO3OH. This shows that the remnant carbonates within the LaAlO3-700 crystallise under reaction conditions. XPS analysis (Table\u00a0S8) of the glycerol treated materials show limited changes in the LaAlO3-900 and LaAlO3-1100 materials when compared to the fresh material indicating surface stability under hydrothermal conditions. However, for the La 3d region of LaAlO3-700, given in Fig.\u00a08\na(ii), a reduction in La at.% and multiplet splitting indicates the loss of surface La and formation of La2(CO3)3 species, in-agreement with the XRD and formation of lanthanum carbonate species. This was confirmed by MP-AES of the filtrate, given in Table\u00a04\n, which shows increased La leaching (17.9%) compared to the higher calcined materials (6\u20137%).Reactions with lactic acid were then chosen, as LA is one of the main acidic products that is formed under batch reaction conditions and can therefore contribute to the breakdown of the perovskite. The XRD of the recovered samples, shown in Fig.\u00a09\na, after standard reaction with 1\u00a0wt% LA and 9\u00a0wt% glycerol shows crystalline perovskite was retained in all the samples, with minimal evidence of crystalline LaCO3OH. However, similarly with the post APR recovered catalysts, there is a dramatic reduction in crystallinity, evidenced by reduction of the LaAlO3:LaB6 ratio (Table\u00a04), of the samples indicating the formation of amorphous phases. Elemental analysis of the effluent after the simulated reaction shows La leaching up to 86% of the original content for all samples. It is clear that the perovskite phase is highly unstable in the presence of lactic acid and in the absence of a carbonate source does not form crystalline carbonate biproducts.Confirmation of further structural changes was given by PDF, shown in Fig.\u00a010\na, by refining a model against the LaAlO3-1100 after LA and glycerol reaction, which yields a good fit with an Rwp\u00a0=\u00a021.06%. Similarly, to the fresh LaAlO3-700 refinement, two phases were needed to fully describe the data, in this instance, LaAlO3 and AlO(OH). This multiphase approach allowed identification of the bulk of the sample, which remains the LaAlO3 phase, with a scale 88.68%. This shows that the perovskite phase is retained despite the reduction in crystallinity post treatment. However, the presence of AlO(OH) phase shows the creation of this poorly crystalline phase under reaction conditions which agrees with the La leaching under acidic conditions and previous studies of alumina transformation into hydroxylated species [24].XPS analysis of the recovered materials show shifts in Al 2p peaks (Table\u00a0S8), possibly from hydroxylation of the Al species, in agreement with the PDF, and an increase in surface Al at.%. In the La 3d region, a large at.% reduction is noted in the La 3d region of all samples, indicating high La leaching. Although, the speciation only changes for the LaAlO3-700 to La2(CO3)3, as shown in Fig.\u00a08b(iii). Therefore, a clear picture emerges regarding support stability in the presence of organic acids; most of the La has dissolved into the filtrate in all samples, with the remainder locked into residual crystalline perovskite phase and the exsolved aluminium is present as a disordered AlO(OH). Again, it is important to highlight that simple fingerprinting of the XRD of post reaction materials would reveal only the perovskite phase which could easily be misinterpreted as a stable phase, when it is in fact, as supported by XPS/PDF and LaB6 doping studies, it is clearly not.CO2 is an important product in the APR of glycerol and can readily dissolve into water forming carbonic acid [57]. Simulated reactions were carried out with a partial pressure of 1\u00a0bar PCO2, which equates to 2.3% glycerol conversion assuming an ideal reforming reaction (i.e. compete product selectivity to CO2 and H2). However, as seen from the catalytic results (Table\u00a02 and Fig.\u00a05\nvide supra), selectivity towards complete reforming is low. Therefore, the PCO2 partial pressure added in simulated experiments equates to similar amounts produced in the Pt/LaAlO3-1100 batch reactions at 20% conversion. The XRD patterns of reactions with 1\u00a0wt% LA and 1\u00a0bar PCO2 are shown in Fig.\u00a09b, with crystalline phases LaCO3OH and LaAlO3 clearly present in LaAlO3-900 and LaAlO3-1100, with only LaCO3OH and no LaAlO3 remaining in the LaAlO3-700 sample. PDF refinement of LaAlO3-700 after the LA, glycerol and PCO2 reaction, shown in Fig.\u00a010b, yields a well-fitting model, with an Rwp\u00a0=\u00a021.44%, with this sample being dominated by the LaCO3OH phase with a smaller contribution from AlO(OH). This confirms that the treatment with PCO2 facilitates the phase transformation of LaAlO3 into LaCO3OH-AlO(OH) phases.The presence of Al species in the XPS in all samples (Table\u00a0S8), under acidic conditions and PCO2 atmosphere, in significant amounts (30.07\u201331.80\u00a0at.%), despite limited crystalline Al phases in the XRD patterns also suggests the formation of amorphous or poorly crystalline AlO(OH) species. The La 3d region for the samples, given in Fig.\u00a08b(iv), also confirm the change of surface speciation to La2(CO3)3. From the MP-AES (Table\u00a04), the addition of PCO2 into the reaction mixture leads to a drop of La in the filtrate, relative to the 1\u00a0wt% lactic acid solution (42\u201353% vs 85\u201386%). The increased stability of La being due to the formation of the stable LaCO3OH phase under an overpressure of CO2. However, a large percentage of La remains in solution, indicating lanthanum carbonate hydroxide phase formation is incomplete under the reaction timescale (2\u00a0h). It is important to note, that the rate of LaAlO3 decomposition is different for each of the LaAlO3-\n\nc\n\n samples, which has an impact on Pt particle migration during the APR reaction. It is clear that the amorphous content present in the LaAlO3-700 significantly affects the stability of the material more than increasing particle size arising from the different calcination temperatures. This is shown in the differences in LaAlO3:LaB6 peak ratios in Table\u00a04.Interestingly, testing of LaAlO3-900 with 10\u00a0wt% glycerol and 1\u00a0bar PCO2, in the absence of an organic acid, also formed LaCO3OH, evidenced by the XRD pattern in Fig.\u00a0S15. This suggests that carbonic acid is acidic enough to decompose the perovskite and force the formation and stabilisation of the LaCO3OH phase, whilst sequestering some of the CO2 present. The amount of CO2 in the simulated reactions, however, is no more than what is produced under real catalytic reactions, and this suggests that residence time under acidic environment, as well as amount of CO2 is also important in material stability.To summarise, upon testing under hydrothermal conditions (glycerol), LaAlO3-700 decomposed slightly with lanthanum carbonate phases present, as well as perovskite, whereas the higher temperature calcined materials remained stable. Testing with organic acidic products (lactic acid), without a carbonate source, led to a reduction in perovskite crystallinity, extensive lanthanum leaching and AlO(OH) formation in all samples (Equation (1)). This confirms the instability of perovskite in acidic media, as predicted by Pourbaix diagrams (Fig.\u00a0S16) [58\u201360].\n\n(1)\nLaAlO3(s)\u00a0+\u00a03H+\n(aq)\u00a0+\u00a0H2O(l) \u2192 La3+\n(aq)\u00a0+\u00a0AlO(OH)(s)\u00a0+\u00a02H2O(l)\n\n\n\n\n\n(2)\nCO2(g)\u00a0+\u00a0H2O(l) \u2192 HCO3H(aq) \u2192 CO3\n2\u2212\n(aq)\u00a0+\u00a02H+\n(aq)\n\n\n\nWhile other acids facilitate phase dissolution and segregation, it is the presence of a carbonate source (CO2), which dissolves as carbonic acid and dissociates to the carbonate species (Equation (2)), that facilitates the formation of lanthanum carbonate crystalline phases La2O(CO3)2. \n\nx\n\n H2O and LaCO3OH (Equations (3) and (4)). Surface terminating La(OH)3 can also be readily converted to lanthanum carbonate phases in the presence of a carbonate source (Equations (5) and (6)) [30]. The La2O(CO3)2. \n\nx\n\n H2O phase dissolves and recrystallises under hydrothermal conditions in the eventual formation of crystalline LaCO3OH (Equation (7)). It is important to note that these reactions are happening concurrently until the formation of crystalline LaCO3OH and AlO(OH). It has also been reported that glycerol can mediate the formation of LaCO3OH in combination with a carbonate source [61].\n\n(3)\n2La3+\n(aq)\u00a0+\u00a02CO3\n2\u2212\n(aq)\u00a0+\u00a0yH2O(l) \u2192 La2O(CO3)2.xH2O(s)\u00a0+zH+\n(aq)\n\n\n\n\n\n(4)\nLa3+\n(aq)\u00a0+\u00a0CO3\n2\u2212\n(aq)\u00a0+\u00a0H2O(l) \u2192 LaCO3OH(s)\u00a0+\u00a0H+\n(aq)\n\n\n\n\n\n(5)\n2 La(OH)3 (s)\u00a0+\u00a02 CO3\n2\u2212\n(aq) \u2192 La2O(CO3)2. xH2O(s)\n\n\n\n\n\n(6)\nLa(OH)3 (s)\u00a0+\u00a0CO3\n2\u2212\n(aq) \u2192 LaCO3OH(s)\u00a0+\u00a02 OH\u2212\n(aq)\n\n\n\n\n\n(7)\nLa2O(CO3)2.xH2O(s) \u2192 2 LaCO3OH(s)\u00a0+\u00a0yH2O(l)\n\n\n\nThe decomposition rates for each calcination temperature are different and this would impact Pt nanoparticle redistribution in real catalytic systems. It is also important to consider residence time in acidic products, which can be further investigated using a flow reactor.The stability of LaAlO3 perovskite supports, during the Pt catalysed aqueous phase reforming of glycerol, has been investigated with respect to the original calcination temperature used to produce the perovskite. Analysis of the perovskite precursor decomposition shows that whilst a calcination temperature of 700\u00a0\u00b0C yields the perovskite phase as the sole crystalline phase, there was significant amounts of amorphous carbonate material, as identified by PDF analysis. The amorphous carbonate could be expelled from the structure upon increasing calcination temperature above 850\u00a0\u00b0C. 1\u00a0wt% Pt/LaAlO3, prepared using a perovskite support calcined at 700\u00a0\u00b0C, had a notably poorer catalytic activity during aqueous phase reforming than catalysts prepared using a support calcination temperature of 900\u00a0\u00b0C or 1100\u00a0\u00b0C. Such low activity could be attributed to poor Pt interaction with the support containing amorphous content. The highest conversion being with the Pt/LaAlO3-1100 at 20.4% (TOF 686 h\u22121).Characterisation of the materials after batch APR reactions, and hydrothermal exposure of supports to simulated product mixtures, showed that the perovskite is replaced by amorphous content alongside La leaching under acidic conditions. Organic acids (i.e. lactic acid) attack the structure and chelate La causing leaching. This was observed for catalysts prepared at each of the three calcination temperatures, although the process was notably faster when using the perovskite containing residual carbonate (700\u00a0\u00b0C). The smaller size of the perovskite particles at lower temperature may also have some influence on stability, however, the presence of amorphous carbonate had a more pronounced effect. The support phase purity is an important factor in catalyst activity through material stability and Pt particle interaction.Formation of crystalline LaCO3OH was only observed when a ready source of carbonate, from CO2 production by reforming, was present. This phase, alongside nanocrystalline Al2O3 or AlO(OH), was consistently formed regardless of LaAlO3 calcination temperature, however, the rate of transformation was slowed with increasing calcination temperature. The effect of calcination temperature on the rate of transformation through phase purity and particle size is an important factor to consider when producing stable and active catalysts through support phase restructuring. When the crystalline LaCO3OH was formed leaching of La significantly dropped, showing that carbonic acid provides a stabilising effect vs organic acid, which enhance leaching.It is essential to note, that simple fingerprint analysis of the support by XRD would have missed the complexity and early stages of LaAlO3 decomposition. The identification of amorphous content and its structural determination by X-ray PDF and XPS analysis was required to fully understand the changes to support structure during reactions.\nDonald R. Inns: Conceptualization, methodology, investigation, formal analysis and visualization, writing \u2013 original draft; Xuetong Pei: Investigation, formal analysis; Zhaoxia Zhou: Investigation, Resources; Daniel J. M. Irving: Formal analysis and visualization, writing \u2013 original draft; Simon A. Kondrat: Conceptualization, project administration, supervision, funding acquisition, writing \u2013 review and editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.We would like to acknowledge funding from the EPSRC\nCDT \u2018fuel cells and their fuels\u2019 (EP/L015749/1). The X-ray photoelectron (XPS) data collection was performed at the EPSRC National Facility for XPS (\u2018HarwellXPS\u2019), operated by Cardiff University and UCL, under Contract No. PR16195. We would also like to acknowledge the use of the facilities within the Loughborough Materials Characterisation Centre (LMCC). Finally, we acknowledge Diamond Light Source, U.K., for access to beamline I15-1 as part of the Catalysis Hub BAG proposal (CY29757). Meta data for XRD, PDF and XPS is available at https://doi.org/10.17028/rd.lboro.20170784.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.mtchem.2022.101230.", "descript": "\n                  Aqueous phase reforming (APR) of waste oxygenates offers the potential for sustainable hydrogen production. However, catalyst stability remains elusive, due to the aggressive hydrothermal conditions employed. Herein, we show that the catalytic performance and stability of Pt supported on LaAlO3 catalysts for glycerol APR is strongly influenced by the phase purity of LaAlO3. Calcination of the support at 700\u00a0\u00b0C produces the LaAlO3 perovskite phase and an amorphous lanthanum carbonate phase, which can be removed by calcination at higher temperature. Catalysts comprised of phase pure LaAlO3 were notably more active, with a support calcination temperature of 1100\u00a0\u00b0C resulting in 20.4% glycerol conversion (TOF 686 h\u22121) in a 2\u00a0h batch reaction. Interestingly, all the catalysts, regardless of LaAlO3 phase purity, eventually transform into Pt/LaCO3OH-AlO(OH) during reaction, but only in the presence of evolved carbon dioxide, itself produced from glycerol reforming. Studies using simulated reaction products showed that organic acid products (lactic acid), in the absence of CO2, facilitated La leaching and loss of crystallinity. A carbonate source (CO2) is essential to limit La leaching and form stable Pt/LaCO3OH. Pt supported on LaCO3OH and AlO(OH) are stable and active catalysts during APR reactions. Yet, the rate of perovskite phase decomposition strongly influences the final catalyst performance, with the initially phase impure LaAlO3 decomposing too quickly to facilitate Pt redistribution. LaAlO3 calcined at higher temperatures evolved more slowly and consequently produced more active catalysts.\n               "}
{"full_text": "The authors are unable or have chosen not to specify which data has been used.The rising CO2 emissions, global warming and the risks posed by climate change have put the carbon dioxide capture and utilization (CCU) technologies in the limelight. In this scenario, the catalytic processes, especially the catalytic hydrogenation of CO2 for the production of liquid fuels and bulk chemicals, receive a great deal of attention [1\u20134]. The interest of these processes leads in their contribution to the circular carbon economy, by replacing the fossil sources by CO2 as carbon source, and using green hydrogen and renewable energy for the products with an increasing demand in the petrochemical industry [5].The processes for the direct synthesis of hydrocarbons from CO2 (in a single reactor) show several thermodynamic benefits as opposed to the indirect routes (in two stages). Moreover, lower capital investment and operating cost are required. Two main routes can be distinguished for the direct conversion of CO2 into hydrocarbons, and both of them are carried out by means of tandem catalysts [6\u20139]. In the Modified Fischer Tropsch (MFT) synthesis, CO2 reacts according to the Anderson-Schulz-Flory (ASF) mechanism, characteristic of the FT synthesis, using Fe- or Co-based catalysts. The products are in situ reformed over a zeotype providing the adequate acidity and shape selectivity to the selective production of the desired hydrocarbon fraction [10,11]. On the other hand, in the route with oxygenates (methanol/dimethyl ether (DME)) as intermediates, OX/ZEO (metallic oxide/zeotype) catalysts are employed, in which the metallic oxide is the responsible of the formation of the oxygenates, and the zeotype is used for the selective conversion of these into hydrocarbons [12,13]. In both routes, the integration of the two reaction stages helps to: (i) diminish the required investment and the operating costs in contrast to the processes with two reactors; and (ii) to benefit the thermodynamics, because of the shifting of the equilibrium of the CO2 conversion step. Guo et al. [14] made a comprehensive thermodynamic study of the hydrogenation of CO2 to alcohols and hydrocarbons (ethylene, propylene, benzene), proving the thermodynamic feasibility of these processes and pointing out thermodynamics as the necessary preliminary step to establish the appropriate reaction conditions and catalyst to maximize CO2 conversion and hydrocarbon yield.The knowledge on the catalysts and mechanisms of the routes for the direct conversion of CO2 into hydrocarbons is based on the prior knowledge of the individual stages in the indirect routes. In that way, for methanol/DME synthesis, Cu-based metallic catalysts (mainly Cu-ZnO-Al2O3) have been used in the industry (originally with syngas and more recently for CO2 hydrogenation), due to their low cost and high performance [15\u201317]. The role of the ZnO is key to increase the dispersion of Cu and to minimize its sintering [15,18]. The presence of Al2O3 as promoter provides surface area and eases the separation of Cu-ZnO sites, resulting in an increase of the stability as well as mechanical resistance [19]. The replacement of Al2O3 by ZrO2 forming Cu-ZnO-ZrO2 catalysts resulted to improve the stability of the catalyst, which is especially interesting in the CO2 hydrogenation due to the high content of H2O in the medium [20,21]. The hydrogenation of CO2 to methanol over Cu-based catalysts is considered to proceed according to the formate species (HCOO*) as first intermediate species and successive hydrogenations [22\u201324]: CO2\u00a0\u2192\u00a0HCOO*\u00a0\u2192\u00a0HCOOH*\u00a0\u2192\u00a0H2COOH*\u00a0\u2192\u00a0H2CO*\u00a0\u2192\u00a0H3CO*\u00a0\u2192\u00a0CH3OH*\u00a0\u2192\u00a0CH3OH. The smallest extent of the alternative route, i.e., the conversion of CO2 into CO by means of the reverse Water Gas Shift (rWGS) reaction and the hydrogenation of CO to formyl species, is explained by the instability of this intermediate, that is decomposed to form CO and H2.Among the catalysts developed to avoid the limitations (temperature, H2O concentration) of Cu-based catalysts in the direct synthesis of hydrocarbons from CO2, those based on In2O3 have received great attention due to their high activity for the conversion of CO2 into methanol and their stability at the temperature required (>350\u00a0\u00b0C) for the conversion of methanol/DME to hydrocarbons [6,25,26]. Moreover, In2O3 is known to suppress the rWGS reaction, avoiding the initial CO2 to CO shift taking place over the Cu-based catalysts [27]. In In2O3 catalysts, CO2 is adsorbed and activated in the oxygen vacancies, and produces formate species following the sequence [28,29]: CO2\u00a0\u2192\u00a0HCOO*\u00a0\u2192\u00a0HCOOH*\u00a0\u2192\u00a0H2COOH*\u00a0\u2192\u00a0H2COHOH*\u00a0\u2192\u00a0H2CO*\u00a0\u2192\u00a0H3CO*\u00a0\u2192\u00a0CH3OH*\u00a0\u2192\u00a0CH3OH. The incorporation of ZrO2 as promoter boosts the formation of additional oxygen vacancies and increases the stability of In2O3 [27,30,31]. The properties of In2O3-ZrO2 have been improved by incorporating Ni [32] and noble metals such as Pd [33,34], Rh [35], Pt [36] or Au [37].Zn containing oxides have also pointed out among methanol synthesis catalysts by providing high CO2 conversion and methanol selectivity, especially when combined with ZrO2 as support, which helps to increase methanol selectivity [38,39]. The properties of ZnO-ZrO2 catalysts are enhanced with the incorporation of noble metals [40]. Wang et al. [41] established that the reaction mechanism of CO2 hydrogenation to methanol over ZnO-ZrO2 based catalyst are both formate and CO reaction pathway. In addition to the great performance, this catalyst has shown to be highly stable, due to the formation of the ZrZnOx solid solution, and it does not undergo deactivation in long catalytic runs (up to 500\u00a0h).As aforementioned, in the direct synthesis of hydrocarbons from CO2, for the methanol/DME conversion into hydrocarbons (second reaction stage), zeolite-based catalysts are used. The activity, selectivity and stability of the zeolites are a direct consequence of their properties, especially of the shape selectivity and the acidity [42]. It is well established the dual cycle mechanism for the conversion of methanol/DME [43,44]. This mechanism takes place by the formation of light olefins as primary products by means of the cycles of alkylation/dealkylation of the intermediate polymethylbenzenes confined in the catalysts, and of oligomerization/cracking of the light olefins. The extent of the secondary reactions (alkylation, isomerization, condensation to aromatics) favors the formation of light paraffins (by hydrogen transfer and cracking), BTX aromatics and not aromatic C5+ hydrocarbons, especially interesting for their use as green gasoline. Therefore, the main challenge is the election of a selective catalyst for each aim. For the selective production of light olefins, SAPO-34 (CHA framework) is highly employed [42,45\u201347]. As an example of good results in the literature, Zhang et al. obtained with GamCrOx/HSAPO-34 catalyst a CO2 conversion of 11.9% and a selectivity of light olefins of 87.3% (excluding CO) at 350\u00a0\u00b0C and a selectivity of 34.5% to CO, at 350\u00a0\u00b0C and 30\u00a0bar [48]. On the other hand, the drawback of the rapid deactivation by coke deposition (assisted by the easy confinement of the polymethylbenzenes in the cages of the porous structure of SAPO-34) is lessened with the particular operating conditions (high H2 and H2O partial pressure) [49].HZSM-5 zeolite is the most studied catalyst for the production of higher hydrocarbons (such as aromatics or linear paraffins in the gasoline-range) from CO2 [13,50,51], owing to its MFI structure, that eases a major extent of the dual cycle mechanism in the conversion of methanol/DME and also of some of the secondary reactions. Moreover, its versatility towards different products in the conversion of methanol/DME by the generation of hierarchical porous structures, the adjustment of the acidity and the incorporation of metals is well established [52\u201354]. The porous structure of HZSM-5 zeolite (without cages in the intersections) facilitates the diffusion of the intermediate coke precursors, delaying their confinement and attenuating the deactivation [55]. Ticali et al. [50] related the higher interest of ZnZrO2/HZSM-5 catalyst for the production of aliphatic compounds in contrast to ZnZrO2/SAPO-34 highlighting higher conversion and stability of HZSM-5 at lower temperature and space time.After the development of the direct synthesis of hydrocarbons from syngas, the direct conversion of hybrid feeds (H2/CO2/CO) is gaining awareness [46,56], on account of the interest in terms of sustainability and joint valorization of CO2 with syngas obtained from biomass [57,58] or waste [59,60]. Moreover, syngas co-feeding partially provides the required hydrogen. Additionally, CO co-feeding also favors thermodynamically the production of methanol as compared to the hydrogenation of CO2 by attenuating the extent of the reverse Water Gas Shift reaction [14,61]. Moreover, the differences in the role of CO when it is formed by the rWGS reaction as a byproduct or when it is co-fed with CO2 has been assessed [14].In this context, the performance of three different metallic oxides (CuO-ZnO-ZrO2, In2O3-ZrO2, ZnO-ZrO2) was compared for their interest to activate the methanol synthesis step in the direct production of gasoline-range C5+ hydrocarbons from CO2 and mixtures of CO2/CO. The results with these catalysts in the synthesis of methanol are continued in this manuscript with those obtained using them in tandem together with HZSM-5 zeolite, aiming at selecting both the metallic oxide and the appropriate operating conditions for the selective production of isoparaffinic gasoline with commercial interest as a fuel. The results allow to assess the prospects of a ZnO-ZrO2/HZSM-5 catalyst for an attractive target (gasoline production), that has received less attention in the literature, and which is complementary to other goals in the catalytic CO2 valorization processes, such as the production of light olefins or aromatics. Considering the importance of the results for the decarbonization objective, attention will also be paid to the CO2 and COx conversion results, attending to the interest of also valorizing the syngas obtained from biomass or wastes.The metallic catalysts for the methanol synthesis step, i.e., CuO-ZnO-ZrO2, In2O3-ZrO2 and ZnO-ZrO2, named in a simplified way CZZ, IZ and ZZ, respectively, were synthesized following a co-precipitation method. CZZ catalyst was prepared with a Cu/Zn/Zr atomic ratio of 2/1/1 following the method described by S\u00e1nchez-Contador et al. [20]. IZ catalyst, with an atomic In/Zr ratio of 2/1, was prepared following the method described by Portillo et al. [30]. For the synthesis of ZZ, a metal nitrate solution with 6.00\u00a0g of Zn(NO3)2\u00b76H2O (Sigma-Aldrich) and 13.69\u00a0g of ZrO(NO3)2\u00b76H2O (Sigma-Aldrich) was co-precipitated over 59.12\u00a0mL of deionized water, and a (NH4)2CO3 solution (VWR Chemicals, 1\u00a0M) was added dropwise under continuous stirring to form a precipitate with an atomic Zn/Zr ratio of 1/2.5. This synthesis method was based on a previous work of Li et al. [62] and slight modifications were considered. The three catalysts were prepared at 70\u00a0\u00b0C and neutral pH. After the co-precipitation, the catalysts were aged, filtered and washed with deionized water. Afterwards, the catalysts were dried and calcined (at 300\u00a0\u00b0C for 10\u00a0h, at 500\u00a0\u00b0C for 1\u00a0h and at 500\u00a0\u00b0C for 5\u00a0h for CZZ, IZ and ZZ catalysts, respectively) in order to provide the corresponding metal oxides, according to the protocols established for each catalyst [20,30,62]. The resulting powders were pelletized to provide higher mechanical resistance, and sieved to a particle size in the 125\u2013250\u00a0\u03bcm range.As acid catalyst, a commercial HZSM-5 zeolite (Zeolyst International) with a Si/Al ratio of 140 was used. The election of the zeolite is a complex decision. This Si/Al ratio was selected to minimize cracking reactions and to increase the gasoline yield. The zeolite, provided in ammonium form was calcined at 575\u00a0\u00b0C for 2\u00a0h to obtain the acid form, pelletized and sieved to a particle size between 300 and 400\u00a0\u03bcm. These calcination temperature is appropriate for equilibrating the catalyst, so that it can recover its activity when used in reaction-regeneration cycles, after the elimination of the coke by air combustion at 550\u00a0\u00b0C [63]. The different particle size of both catalysts was selected as to ensure the easy separation after the reaction, having formerly proved that no diffusional limitations occur with these sizes. The tandem catalysts (CZZ/HZSM-5, IZ/HZSM-5 and ZZ/HZSM-5) were prepared by physical mixture of both metallic and acid catalysts, with a metal/acid mass ratio of 1/1.The physical properties of the catalysts (Table 1\n) were determined by N2 adsorption-desorption isotherms (Micromeritics ASAP 2010). For this, the samples were degassed in vacuum conditions prior to the analysis, in order to remove impurities and H2O adsorbed on the surface of the catalyst. Afterwards, N2 adsorption-desorption equilibrium stages were conducted at \u2212196\u00a0\u00b0C. It is remarkable that among the metallic catalysts, CZZ has a more favorable porous structure for the access of the reactants and diffusion of the intermediates and products, with higher values of BET surface area (SBET), pore volume and mean size of pore diameter, whereas the ZZ catalyst has the lowest values of these properties. The properties of the HZSM-5 catalyst are characteristic of this zeolite, and correspond to a mostly microporous structure, whereas the presence of mesopores is due to the pelletization step. It should be noted that the kinetic results in Section 3.2 highlight the minor importance of these properties and the fundamental role of the different activity of the active sites of the catalysts in the corresponding reaction mechanism.The chemical composition and atomic ratios were quantified by X-Ray fluorescence (XRF), by means of a PANalytical wavelength dispersive X-ray fluorescence sequential spectrometer (WDXRF), model AXIOS, equipped with a Rh tube and three detectors (gas flow, scintillation and Xe sealing). Results are shown in Table S1. The structure was determined by X-Ray diffraction (XRD) with a PANalytical Xpert PRO diffractometer, equipped with copper tube (\u03bbCuK\u03b1\u00a0=\u00a01.5418\u00a0\u00c5), a vertical goniometer (Bragg-Brentano geometry), secondary monochromator and PixCel detector. The measurement conditions were 40\u00a0kV/40\u00a0mA and the pattern was recorded in a 5\u00a0<\u00a02\u03b8\u00a0<\u00a060 range for CZZ catalyst and in a 5\u00a0<\u00a02\u03b8\u00a0<\u00a080 range for IZ and ZZ catalysts.The normalized XRD patterns of the three metallic catalysts are gathered in Fig. 1\n. According to the diffractograms, IZ comprises cubic structure for both In2O3 and ZrO2 oxides (in accordance with ICDD (International Center for Diffractional Data) #71\u20132195 and #49\u20131642, respectively), corresponding to the state with the highest catalytic activity [64]. On the other hand, hexagonal ZnO (in accordance with ICDD #36\u20131451) and cubic ZrO2 (in accordance with ICDD #49\u20131642) structures were found in ZZ catalyst. Regarding the traditional CZZ catalyst, its structure was described thoroughly elsewhere [20]. Briefly, the characteristic peaks of CuO and ZnO oxides were visible on the spectra, while ZrO2 peaks were not noticeable due to the high dispersion and small size of the crystals.H2 temperature programmed reduction (H2-TPR) analyses were carried out (Micromeritics Autochem 2920) to study the reducibility of the catalysts. For this assay, 100\u00a0mg of sample were treated previous to the reduction by sweeping with He, to remove possible impurities and H2O. The H2-TPR analysis was carried out heating the sample up to 800\u00a0\u00b0C at a 2\u00a0\u00b0C\u00a0min\n\u22121 rate in a diluted H2 stream (10% H2 in Ar). Attending to the TPR profiles (Fig. S1), CZZ is completely reduced at temperatures above 200\u00a0\u00b0C, whereas IZ and ZZ require higher temperature, so they might be in their oxide form at the beginning of the reactions. The same equipment was used for measuring the acidity by means of NH3-TPD analyses. 50\u00a0\u03bcL\u00a0min\n\u22121 NH3 were injected at 150\u00a0\u00b0C until the saturation of the sample. The desorption step was conducted with a 5\u00a0\u00b0C\u00a0min\n\u22121 rate up to 550\u00a0\u00b0C in a He stream. Fig. S2 exhibits the NH3-TPD profile of the HZSM-5 catalyst. The total acidity of this zeolite accounts for 62 \u03bcmolNH3 gcat\n\u22121, with a peak at 190\u00a0\u00b0C and a higher one at 320\u00a0\u00b0C, stating low total acidity but a great presence of strong acid sites according to the classification in the literature [65].The catalytic runs were performed in an isothermal packed bed reactor (PID Eng & Tech Microactivity Reference). The reactor is made of 316 stainless steel and has a ceramic coating to avoid the contact of the reactants with the steel and so, any possible side reaction. The internal diameter of the reactor is of 9\u00a0mm and it has an effective length of 10\u00a0cm. This equipment can operate at a pressure up to 100\u00a0bar and temperatures up to 700\u00a0\u00b0C. The catalytic bed is composed of a mixture of the catalyst and an inert solid (SiC), in order to ensure isothermal conditions of the bed, to avoid preferential flow paths and to achieve sufficient bed height when operating at low space time values.The feed and reaction product streams were analyzed on-line in a micro chromatograph (Varian CP-4900, Agilent), equipped with three analysis modules composed of TCD detectors and the following columns: (i) molecular sieve (MS-5) (10\u00a0m\u00a0\u00d7\u00a012\u00a0\u03bcm) for the quantification of H2, O2, N2 and CO; (ii) Porapak Q (PPQ) (10\u00a0m\u00a0\u00d7\u00a020\u00a0\u03bcm) for the quantification of CO2, methane, H2O, C2-C4 hydrocarbons, methanol and DME; and (iii) 5 CB (CPSiL) (8\u00a0m\u00a0\u00d7\u00a02\u00a0\u03bcm) for the quantification of C5+ hydrocarbons.The reaction runs of methanol synthesis (with the CZZ, IZ and ZZ metallic catalysts) were carried out under the following conditions: 250\u2013430\u00a0\u00b0C; 50\u00a0bar; space time, 6 gcat h molC\n\u22121; CO2/COx molar ratio in the feed, 0, 0.5 and 1; H2/COx molar ratio in the feed, 3. The reactions for the direct synthesis of hydrocarbons (with the CZZ/HZSM-5, IZ/HZSM-5 and ZZ/HZSM-5 tandem catalysts) were performed under the following conditions: 340, 380 and 420\u00a0\u00b0C; 30 and 50\u00a0bar; space time, 12 gcat h molC\n\u22121; CO2/COx ratio in the feed, 0.5 and 1; H2/COx ratio in the feed, 3. Prior to all the reaction runs, the catalysts were subjected to a partial reduction in a H2 and N2 stream (1\u00a0h at 350\u00a0\u00b0C, 2\u00a0bar and with a flow rate of 30 cm3\nH2 min\u22121 and 30 cm3\nN2 min\u22121).The conversions of CO2 (XCO2) and of COx (XCOx) were defined according to the expressions:\n\n(1)\n\n\nX\n\nC\n\nO\n2\n\n\n\n=\n\n\n\nF\n\nC\n\nO\n2\n\n\n0\n\n\u2212\n\nF\n\nC\n\nO\n2\n\n\n\n\n\nF\n\nC\n\nO\n2\n\n\n0\n\n\n\u00b7\n100\n\n\n\n\n\n(2)\n\n\nX\n\nC\n\nO\nx\n\n\n\n=\n\n\n\nF\n\nC\n\nO\nx\n\n\n0\n\n\u2212\n\nF\n\nC\n\nO\nx\n\n\n\n\n\nF\n\nC\n\nO\nx\n\n\n0\n\n\n\u00b7\n100\n\n\nwhere F0\nCO2 and F0\nCOx are the molar flow rates of CO2 and COx at the inlet of the reactor, respectively, and FCO2 and FCOx are the corresponding values at the reactor outlet stream.The yield and selectivity of each i product (Yi and Si, respectively) excluding CO2 and CO, were calculated as:\n\n(3)\n\n\nY\ni\n\n=\n\n\n\nn\ni\n\n\u00b7\n\nF\ni\n\n\n\nF\n\nC\n\nO\nx\n\n\n0\n\n\n\u00b7\n100\n\n\n\n\n\n(4)\n\n\nS\ni\n\n=\n\n\n\nn\ni\n\n\u00b7\n\nF\ni\n\n\n\n\n\u2211\ni\n\n\n\n\nn\ni\n\n\u00b7\n\nF\ni\n\n\n\n\n\n\u00b7\n100\n\n\nwhere ni is the number of carbon atoms of the i compound and Fi the molar flow rate of the i compound in the products stream in content C atoms. It should be noted that with the definition of yields in Eq. (2), XCOx is the sum of the yields.The reactions presented in this section were conducted without zeolite, with the aim of testing the metallic catalysts alone in the first stage of the gasoline production (synthesis of methanol) in the 250\u2013430\u00a0\u00b0C range. Fig. 2\n shows the effect of the temperature on the conversion (XCOx) of an equimolar mixture of CO2 and CO, and on the selectivity of methanol and other byproducts (CH4, C2-C4 paraffins and C2-C4 olefins) with the three catalysts. Comparing the results, notable differences are observed. With CZZ catalyst (Fig. 2a), high conversion was reached at low temperature. In fact, XCOx accounted for 19% at 280\u00a0\u00b0C, which corresponds with the thermodynamics prediction [14,66\u201368]. Moreover, the selectivity of oxygenates (mainly methanol, with an insignificant DME content) was 100%. Nonetheless, XCOx decreased steadily with increasing temperature, until declining to 3.5% at 430\u00a0\u00b0C, due to the thermodynamic limitation. In addition, selectivity of methanol also decreased at high temperature due to the favoring of CO formation by the rWGS reaction. These results are in accordance with the prediction of thermodynamic studies in the literature [14,66\u201368]. The presence of C2-C4 paraffins is explained by the hydrogenation of the light olefins formed from the conversion of oxygenates, and the presence of CH4 over 340\u00a0\u00b0C exposes the activity of CZZ in the methanation at this temperature.XCOx values were lower with the IZ (Fig. 2b) and ZZ (Fig. 2c) catalysts. These catalysts had similar activity, lower than that of the CZZ catalyst. The XCOx reached a maximum value of 4.7% at 340\u00a0\u00b0C with IZ catalyst, and between 340 and 370\u00a0\u00b0C with ZZ catalyst, which decreased above these temperatures due to thermodynamic limitations, that also affect to the conversion of CO [14,66,67]. It is noteworthy that methanol selectivity was higher with IZ and ZZ catalysts than with the CZZ catalyst (Fig. 2a). In this sense, the best performance corresponded to the ZZ catalyst, with a methanol selectivity of almost 100% in the 340\u2013370\u00a0\u00b0C range. Indeed, selectivity only decreased slightly with increasing temperature up to 430\u00a0\u00b0C. Considering the aforementioned results, the higher activity of CZZ catalyst for methanol production below 300\u00a0\u00b0C (with a maximum at 280\u00a0\u00b0C) is of arguable interest from the perspective of its use in the direct conversion of CO2/CO mixtures to hydrocarbons, since this reaction must be performed at higher temperature to achieve the extent of the dual cycle mechanism to obtain C5+ hydrocarbons. In this regard, the reduced methanation activity of the ZZ catalyst in the 300\u2013400\u00a0\u00b0C is of particular interest. It is noteworthy that this better performance of the ZZ catalyst with respect to the other catalysts cannot be attributed to the properties of its porous structure (Table 1), because these are less favorable for the diffusion of the reactants and products. Consequently, it should be attributed to the high activity and selectivity of the active sites of the ZZ catalyst in the methanol formation mechanism explained by Wang et al. [41] with formate ions and CO as intermediates in the reaction pathway.In Fig. 3\n the effect of the feed composition (CO2/COx ratio) on methanol yield is shown. The results correspond to 350\u00a0\u00b0C, temperature considered as limit to avoid the sintering of the Cu on CZZ catalyst [27]. This catalyst is the most active for CO hydrogenation, with a methanol yield of 21%, higher than with IZ (4%) and ZZ (2%) catalysts. At higher CO2/COx ratio, methanol yield remarkably decreased with CZZ catalyst. This trend fits with previous findings regarding the effect of the CO2 content in the feed [66,69] in high conversion conditions (high concentration of methanol), for which the presence of CO is preferable to CO2, as it eases the H2O removal by means of the WGS reaction. As could be expected, the methanol yield in the CO2 conversion is lower than that obtained in the literature with catalysts of similar composition under optimal conditions for methanol synthesis, i.e., lower temperature and higher pressure than those used [70]. On the other hand, the results with IZ and ZZ catalysts showed a similar trend. They both exhibited the highest methanol yield when the carbon source of the feed was 50% CO and 50% CO2. This concurs well with previous works in the literature with the IZ catalyst [25,56]. This occurs because the reaction mechanism lies on the creation and eradication of oxygen vacancies, and the joint feed boosts this process and, additionally, favors the preservation of the oxygen vacancies.It is also outstanding in Fig. 3 that, for the hydrogenation of CO2 (CO2/COx of 1), the obtained methanol yield was similar with the three catalysts. This result evidences the aforementioned limitation of the equilibrium conversion, and that this conversion is low in CO2 hydrogenation. This is in accordance with thermodynamic studies in the literature [14,66\u201368]. It is also observed that with IZ catalyst methanol yield was similar in CO and CO2 hydrogenation.With the purpose of assessing the performance of the metallic catalysts used in tandem, in Figs. 4 and 5\n\n corresponding to IZ/HZSM-5 and ZZ/HZSM-5, respectively, the effect of temperature (340\u2013420\u00a0\u00b0C range) and pressure (30 and 50\u00a0bar) on the conversion of COx (sum of the products yields, Eq. (2)) and CO2 and on the different products yield is shown. The results correspond in both cases to an equimolar feed of CO2 and CO (CO2/COx of 0.5) and hydrogen. It should be noted that the results for the CZZ/HZSM-5 catalyst are not shown because the sintering of Cu above 320\u00a0\u00b0C was verified. In fact, an increase of the crystal size from \u223c10\u00a0nm (fresh catalyst) to \u223c35\u00a0nm was determined by XRD analysis of the spent catalyst (Table S2). On the contrary, IZ/HZSM-5 and ZZ/HZSM-5 spent catalysts maintained constant their properties in long reaction runs at these temperatures. Consequently, the attention was focused in these two catalysts because of their stability in the required temperature range.It is also remarkable (in Figs. 4 and 5) that the C5+ hydrocarbons are the main products for the two catalysts and the oxygenates are almost completely converted. At higher pressure, the results upturned, boosting the overall COx conversion. Regarding the IZ/HZSM-5 catalyst (Fig. 4) at 30\u00a0bar, the influence of the temperature was more subdued. COx conversion did not increase >2% when rising temperature from 380 to 420\u00a0\u00b0C. Regarding the CO2 conversion, it was more affected by temperature at the lower pressure of 30\u00a0bar, rising from 8% to 23% by increasing temperature from 340 to 420\u00a0\u00b0C. For its part, at 50\u00a0bar, the CO2 conversion reached 28% at 420\u00a0\u00b0C. At 420\u00a0\u00b0C and lower pressure (30\u00a0bar), the C5+ hydrocarbons yield was of approximately 7%, with a COx conversion of 22%. Nonetheless, under a pressure of 50\u00a0bar and at the same temperature, the obtained products were highly interesting for the insight into sustainable fuels production. With almost no methane yield (<0.5% at 30\u00a0bar), and nearly complete oxygenates conversion, the remaining products were composed of C2-C4 paraffins (with a yield of 3.5%), C2-C4 olefins (2%) and C5+ heavier compounds (17.3%) at the optimal conditions. The presence of olefins was not particularly outstanding, as they are chiefly hydrogenated due to the high H2 partial pressure.For ZZ/HZSM-5 catalyst (Fig. 5) the result of C5+ hydrocarbon yield was even improved compared to IZ/HZSM-5. The CO2 conversion boosted from 8.1% to 28.3% when increasing the temperature from 340 to 420\u00a0\u00b0C (at 30\u00a0bar); and the COx conversion enhanced from 12.8% to 28.3% when rising the operating pressure from 30 to 50\u00a0bar (at 420\u00a0\u00b0C). In addition, the CO2 conversion reached 40% under 420\u00a0\u00b0C and 50\u00a0bar, since, unlike IZ, ZZ catalyst hardly inhibits the rWGS reaction. Under such conditions, besides methane and methanol (whose yield did not exceed 1%), C2-C4 paraffins, C2-C4 olefins and C5+ hydrocarbons yields accounted for 5.1%, 1.5% and 20.7%, respectively. These hydrocarbons were mainly composed by 5 and 6 carbon number isoparaffins and some cyclic hydrocarbons that will be further itemized below.\nFig. 6\n shows the CO2 conversion and the product distribution (in yield terms) achieved with each catalyst in the optimal conditions (420\u00a0\u00b0C and 50\u00a0bar) for the hydrogenation of CO2 and of an equimolar mixture of CO2 and CO. There are some remarkable aspects to highlight in these results that evidence the better performance of the ZZ/HZSM-5 catalyst. As mentioned above, CZZ was not an applicable catalyst for H2\u00a0+\u00a0CO2 valorization. At temperatures above 350\u00a0\u00b0C Cu sintered, because of both temperature and water content (especially high with CO2 content feeds). Nevertheless, the results with this catalyst are summarized in Fig. S3. The COx conversion at 420\u00a0\u00b0C and 50\u00a0bar did not reach 2.5% with the CZZ/HZSM-5 catalyst, since almost no oxygenates were produced at these conditions. On the other hand, with the hybrid feed (CO2/COx\u00a0=\u00a00.5) there was a higher content of oxygenates, which, however, were not successfully converted into C5+ hydrocarbons (<3%), as roughly all the hydrocarbons remained as C2-C4 paraffins, due to the poor synergy between sintered CZZ and the HZSM-5. This poor performance of CZZ is explained by the accumulation of unfavorable circumstances such as the sintering of Cu in the catalyst (Table S2) and the reduced activity of Cu catalysts for CO2 conversion. These circumstances further deteriorate under the used reaction conditions (unfavorable for the methanol synthesis step according to thermodynamics) [14,66\u201368]. With regard to IZ/HZSM-5 catalyst, as noted above, it showed better performance with a mixture of CO and CO2 in the feed, which is in agreement with the finding of Ara\u00fajo et al. [56] about the better preservation of the oxygen vacancies for higher CO content in the feed than for a CO2 and hydrogen feed. Besides the reduced conversion, the production of gasoline-range hydrocarbons fell sharply for the H2\u00a0+\u00a0CO2 feed, revealing that IZ might not be the best metallic catalyst for gasoline-range hydrocarbon production in these operating conditions. In fact, the conversions (XCO2 and XCOx) and C5+ hydrocarbons yield was higher with ZZ/HZSM-5 catalyst for both feeds. The values obtained for these indices with the CO2/CO mixture were of 39.7%, 28.4% and 20.7%, respectively. Additionally, ZZ/HZSM-5 was not affected by the higher content of CO2 in the feed in such manner. Actually, the COx conversion fell merely from 28% to 26% for H2\u00a0+\u00a0CO2 feed. All this evidences the powerful interest of the ZZ/HZSM-5 as a feasible industry catalyst, as it could cope adequately with the current fluctuations of the feed composition in this process.In order to assess the importance of the synergy of the tandem catalysts on the reaction mechanisms, both in the synthesis of oxygenates and in the conversion of these into hydrocarbons, in Fig. 7\n the effect of the temperature on the COx conversion for IZ and ZZ metallic catalysts and for IZ/HZSM-5 and ZZ/HZSM-5 tandem catalysts is compared. The results correspond to the hydrogenation of the equimolar mixture of CO2 and CO. As aforementioned in the synthesis of methanol (Fig. 2), the results were similar for the two catalysts above 350\u00a0\u00b0C as a consequence of the thermodynamic constraints. These constraints are removed with the presence of the HZSM-5 zeolite in the tandem catalysts, due to the shift of the equilibrium by the immediate conversion of the oxygenates. When comparing the results of the two tandem catalysts, the benefit of the synergy between the two reaction steps was more remarkable with the ZZ/HZSM-5 catalyst. At optimal conditions for the integrated process (420\u00a0\u00b0C and 50\u00a0bar), COx conversion was multiplied \u223c12 times (from 1.8% to 23%) with IZ/HZSM-5 with respect to the synthesis of methanol with IZ catalyst, whereas it increased a factor of >15 with the ZZ/HZSM-5 catalyst (from 1.8% to 28%) with respect to ZZ catalyst.\nFig. 8\n exhibits the yield of the different hydrocarbons in the products stream with IZ/HZSM-5 (Fig. 8a) and ZZ/HZSM-5 (Fig. 8b) catalysts. These results allow to compare the performance of the two catalysts from the perspective of product interest. Additionally, the comparison of the results in the hydrogenation of the equimolar mixture of CO2 and CO, and of CO2 was assessed. The majority of hydrocarbons produced with both catalysts were C6, C5 and C4 (in this order from highest to lowest). The highest yields (9.8%, 8.6%, and 4.1%, respectively) were obtained with ZZ/HZSM-5 catalyst for the equimolar mixture. In addition, with this catalyst the yield of C6 fraction was virtually the same in the hydrogenation of CO2 and of the CO2/CO mixture, which evidences that ZZ/HZSM-5 catalyst withstands in a good way the fluctuations in the feed. It is also remarkable that the gasoline fraction (C5+ with a yield of 20.7% with ZZ/HZSM-5 catalyst) was mainly isoparaffinic with both catalysts. This elevated isoparaffin content is in accordance with the well-established activity of the HZSM-5 zeolite-based catalysts for the isomerization of the corresponding linear paraffins [71]. In addition, using HZSM-5 catalysts doped with Zn (by ion exchange or isomorphically substituted) in the conversion of DME at high pressure and in the presence of H2, the high hydroisomerization activity of Zn, favored by its capacity for H2 dissociation and surface H generation, has been determined [72,73]. Because of the favorable conditions, the total yield of C5 and C6 isoparaffins (2- methylbutane, 2-methylpentane and 3-methylpentane) reached nearly 20% in these operating conditions, with almost no C4+ n-paraffin production. Besides, the high temperature and the elevated hydrogen content hinder the dehydrocyclization and aromatization reactions, resulting in low yield of cycloalkanes (2.6%) and aromatics (0.1%). On the other hand, compounds of >7 carbon atoms were not very significant (with a total yield of 2.3%). These results are a consequence of the properties of the metallic oxide and the zeolite used in the tandem catalyst. In this way, the hydrogenating activity of the Zn-based metallic catalyst in high pressure conditions and with the presence of H2 hindered the formation of aromatics [73]. This behavior of the Zn is different from that without the presence of H2 in the feed, where the presence of CO2 favors the formation of aromatics from methanol [74]. On the other hand, the moderate total acidity and the Br\u00f6nsted/Lewis ratio of HZSM-5 limited the extent of heavier hydrocarbon formation reactions [74]. Consequently, the high isoparaffin content and the low presence of linear paraffins resulted in a high Research Octane Number (RON) hydrocarbon mixture with ZZ/HZSM-5 catalyst (RON of 91.8 determined according to the method proposed by Anderson, Sharkey and Walsh [75]), indicating high quality gasoline fraction. Its characteristic composition, without the presence of aromatic compounds, is of great interest for its incorporation to the refinery gasoline pool.As concluded in preceding results in Fig. 8, comparing the catalysts, ZZ showed better performance than IZ when operating in tandem with HZSM-5. Nonetheless, for each feed composition, the trend for both catalysts was virtually the same. However, it is observed that IZ/HZSM-5 catalyst was more afflicted by alterations in feed compositions (Fig. 8a), resulting in considerably lower yield of isoparaffins with an increasing CO2/COx ratio, whereas ZZ/HZSM-5 catalyst withstood better the changes in feed, maintaining almost unchanged the production of isoparaffins (Fig. 8b).The results ratify the good performance (high activity, selectivity of methanol and stability) of In2O3-ZrO2 and ZnO-ZrO2 catalysts in methanol synthesis, especially from CO2, and also from CO2/CO mixtures, at appropriate conditions for the direct synthesis of hydrocarbons. Moreover, both catalysts showed great performance when used in tandem together with a HZSM-5 zeolite, exposing the effective synergy between the mechanisms of methanol formation and its conversion into hydrocarbons, obtaining a high yield of C5+ hydrocarbons.It is especially significant the performance of ZnO-ZrO2/HZSM-5 catalyst, with which at 420\u00a0\u00b0C, 50\u00a0bar, CO2/COx of 0.5 and H2/COx of 3, a yield of C5+ of 20.7% was obtained. Under such conditions, CO2 and COx conversions were very high, of 39.7% and 28.4%, respectively. An interesting advantage of this catalyst with respect to In2O3-ZrO2/HZSM-5 is the low dependence of the results to the CO2/COx ratio in the feed, which provides high versatility in the operation, combining the targets of valorizing CO2 and syngas derived from gasification of biomass or waste from the consumer society.The good results of gasoline production with ZnO-ZrO2/HZSM-5 catalyst from CO2 and mixtures of CO2/CO allow to value positively the interest of this route as a complementary route to others studied in the literature for the production of other hydrocarbons, such as light olefins, light paraffins or aromatics. The C5+ fraction obtained consisted mainly of C5 and C6 isoparaffins, with a yield of isoparaffins of 20% and 0.1% of aromatics. With a RON of 91.8, the obtained product had a very interesting composition for its incorporation into the refinery gasoline pool. Therefore, it can be combined with other streams which, like those derived from fluidized catalytic cracking (FCC), have a content of aromatics and olefins that exceeds legal limitations. In addition, the results can be considered pioneering for this purpose with this catalyst, and they provide good prospects for improvements in the catalyst and in the optimization of the reaction conditions.\nOnintze Parra: Conceptualization, Investigation, Writing \u2013 original draft, Writing \u2013 review & editing. Ander Portillo: Validation, Visualization, Methodology, Writing \u2013 review & editing. Javier Ere\u00f1a: Project administration, Funding acquisition. Andr\u00e9s T. Aguayo: Methodology, Resources, Supervision, Project administration, Funding acquisition. Javier Bilbao: Conceptualization, Writing \u2013 original draft, Writing \u2013 review & editing, Project administration, Funding acquisition. Ainara Ateka: Conceptualization, Writing \u2013 original draft, Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work has been carried out with the financial support of the Ministry of Science, Innovation and Universities of the Spanish Government (PID2019-108448RB-100); the Basque Government (Project IT1645-22), the European Regional Development Funds (ERDF) and the European Commission (HORIZON H2020-MSCA RISE-2018. Contract No. 823745). O. Parra is grateful for the financial support of the grant of the Basque Government (PRE_2021_1_0014) and A. Portillo is grateful for the grant from the Ministry of Science, Innovation and Universities of the Spanish Government (BES2017-081135). The authors thank for technical and human support provided by SGIker (UPV/EHU).\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.fuproc.2023.107745.", "descript": "\n                  The direct production of C5+ hydrocarbons from CO2/CO mixtures with methanol as intermediate is an attractive alternative for the production of gasoline from CO2 and syngas derived from biomass. With this purpose, the performance of CuO-ZnO-ZrO2 (CZZ), In2O3-ZrO2 (IZ) and ZnO-ZrO2 (ZZ) metallic oxides was compared by using them in tandem with a HZSM-5 zeolite. The catalysts were analyzed by means of N2 adsorption-desorption, XRD, XRF, H2-TPR and NH3-TPD. Two series of runs were performed in a packed bed reactor: (i) the methanol synthesis with the metallic oxides as catalysts, at 250\u2013430\u00a0\u00b0C; 50\u00a0bar; CO2/COx, 0\u20131; H2/COx, 3; space time 6 gcat h molC\n                     \u22121; and (ii) the synthesis of hydrocarbons with the tandem catalysts with a metallic oxide/zeolite mass ratio of 1/1, at 340, 380 and 420\u00a0\u00b0C; 30 and 50\u00a0bar; CO2/COx, 0.5 and 1; H2/COx, 3; space time 12 gcat h molC\n                     \u22121. The results were quantified in terms of yield and selectivity of the product fractions and CO2 and COx (CO2\u00a0+\u00a0CO) conversion. The higher methanol yield attained with the CZZ catalyst for the CO\u00a0+\u00a0H2 feed and its mixing with CO2 was faded by the problem of its sintering above 350\u00a0\u00b0C (minimum temperature required for the extent of methanol conversion to hydrocarbons). The IZ and ZZ catalysts were active, selective to methanol and stable both in the methanol synthesis and when used in IZ/HZSM-5 and ZZ/HZSM-5 tandem catalysts. Excellent results were obtained with the latter, which resulted in a 20.7% yield of C5+ hydrocarbon fraction at 420\u00a0\u00b0C and 50\u00a0bar, with CO2 and COx conversion of 39.7% and 28.4%, respectively. This fraction corresponded to isoparaffinic gasoline, with isoparaffin yield (mainly C5 and C6) surpassing 20% and low concentration of aromatics (0.1%) that led to a Research Octane Number of 91.8. This composition results attractive for its integration into the refineries gasoline pool. Furthermore, the changes of the CO2/COx ratio in the feed barely affected the yield and composition of the gasoline obtained with the ZZ/HZSM-5 catalyst, stating its great versatility.\n               "}
{"full_text": "Data will be made available on request.Endocrine-disrupting chemicals (EDCs) are frequently detected in the environment and are of concern due to their potentially harmful effects on human health and the ecosystem [1,2]. Recently, SO4\n\u2219\n\u2212-based advanced oxidation processes (SR-AOPs) are gaining enormous attention as an efficient technology to degrade and mineralize recalcitrant organic pollutants in water [3,4]. SO4\n\u2219\n\u2212 (E0\u00a0=\u00a02.5\u20133.1 VNHE) has higher standard redox potential and longer half-life time than \u2219OH (E0\u00a0=\u00a01.8\u20132.8 VNHE) [5,6]. Generally, SO4\n\u2219\n\u2212 can be generated through activating peroxymonosulfate (PMS) or peroxodisulfate (PS) by heating, ultraviolet radiation, electricity, and ultrasound [7]. However, these activation methods require high energy. To reduce energy consumption, transition metals (Co, Fe, Mn, Cu, etc.)-based materials were used as heterogeneous catalysts for PMS activation [8,9]. Among them, Fe-based materials have shown great potential in the degradation of organic pollutants due to their advantages of being cost-effective, high-efficient, eco-friendly characters, and easily accessible as the second most abundant metallic element of the earth's crust [10]. Most Fe-based materials are magnetically separable, making them easier to recycle [11]. Notably, nanoscale zero-valent iron (Fe0) has been recognized as an efficient catalyst for activating PMS due to its nano size and high surface reactivity [12]. Nevertheless, Fe0 has the characteristics of its high surface energies and inherent magnetism, which lead to the formation of larger particles and subsequently reduce the catalytic activity [13]. To overcome these shortcomings, many kinds of carbon materials such as active carbon (AC), biochar (BC), graphene oxide (GO), and N-doped carbon (NC) with highly porous structures have been used as favorable support materials for loading Fe0 [9,10,14]. Nevertheless, these Fe/C composites still showed limited catalytic performances due to the thermodynamic instability and aggregation of Fe0 on carbon supports [15,16]. Therefore, seeking a better Fe/C-based catalyst with adequate activity and stability is necessary.Metal-organic frameworks (MOFs) are highly ordered porous materials, with metal ions or clusters in their center and organic ligands as linkers [17]. Their structures can be precisely engineered into a variety of multilevel nanoarchitectures with desired size, porosity, and functional groups by controlling the geometry of the constituent components through various synthetic methods [18]. In recent years, pyrolysis of Fe-based MOFs (Fe-MOFs) to form core-shell-like carbon nanostructures has become one of the most promising platforms for fabricating active and stable Fe/C-based catalysts [19]. For instance, Zhao et al. prepared Fe0/Fe3C/Fe-Nx decorated NC nanotubes by pyrolysis of MIL-88B(Fe) and melamine (MM) at 900\u00a0\u00b0C in N2 atmosphere [20]. Zhang et al. reported Fe3O4/ZnO incorporated carbon spheres by pyrolysis of Zn/Fe-MOFs at 650\u00a0\u00b0C in N2 atmosphere [21]. Chen et al. synthesize Fe0/Fe3C/Fe-Nx inside NC nanofibers by pyrolysis of polyacrylonitrile (PAN) modified Fe-MIL-101 at 900\u00a0\u00b0C in N2 atmosphere [22]. It is generally accepted that altering the annealing temperature can change catalyst characteristics, including elemental composition, metallic phase, pore structure, and graphitization degree [23,24]. These affect the active site, specific surface area, and electrical conductivity, influencing the catalytic activity of the Fe/C-based materials [25]. However, most reported Fe-MOFs derived SR-AOPs catalysts were doped with nitrogen (N) or other metallic elements (Co, Cu, Ni, etc.). The effect of annealing temperature on the major elements (Fe, C) in the catalytic materials becomes unclear [22,26].To the best of our knowledge, the effect of annealing temperature on characteristics and catalytic activity of Fe-MOFs derived catalysts with neither N nor other metallic elements have not been investigated in PMS-based SR-AOPs. Thus, in this work, we synthesized composites with Fe0/Fe3C/Fe3O4 wrapped in porous carbon shell (CC-Fe/C) at different annealing temperatures (700, 800, and 900\u00a0\u00b0C). To exclude the influence of other elements, an unmodified MIL-88B(Fe) was prepared as the only precursor, and the pyrolysis was performed under an inert atmosphere (Ar) instead of N2. The effects of annealing temperature on morphology, elemental composition, crystal phase, and pore structure of the CC-Fe/C catalysts were investigated. The catalytic activity of samples for PMS activation was evaluated by the removal of bisphenol A (BPA), a typical EDC. Results suggested that an optimal annealing temperature of 800\u00a0\u00b0C led to multiple Fe-based active sites (Fe0, Fe3C, and Fe3O4) on CC-Fe/C-800 surface, synergistically promoting its catalytic activity with good reusability (up to the third cycle) and practicability (in tap water/treated wastewater). Finally, the mechanisms (including synergistic activation of PMS, dominated reactive oxygen species, and acceleration of the Fe3+/Fe2+ cycles) involved in the catalytic oxidation of BPA were proposed. Based on the progress, this work may offer a good reference to tune the physicochemical properties of the CC-Fe/C catalyst for enhancing its catalytic activity and a new insight into the underlying catalytic reaction mechanism of PMS\u00a0+\u00a0CC-Fe/C system for BPA removal.All chemicals used in this work (listed in Table S1) were of analytical grade without any purification. All solutions were prepared with deionized (DI) water, otherwise indicated. Tap water and treated wastewater were obtained from our laboratory and a wastewater treatment plant in Hangzhou, respectively.MIL-88B(Fe) was synthesized by a modified solvothermal method [27]. Typically, 0.54\u00a0g (2\u00a0mmol) of FeCl3\u22196H2O and 0.664\u00a0g (4\u00a0mmol) of terephthalic acid were added into a solution of 20\u00a0mL of N, N-Dimethylformamide (DMF) and 3.2\u00a0mL of 1\u00a0M NaOH. After stirring, the mixture was transferred into a 100-mL Teflon-lined stainless-steel reactor and heated at 100\u00a0\u00b0C for 24\u00a0h. The orange precipitates [MIL-88B(Fe)], were collected by centrifugation, consecutively washed with 50\u00a0mL pure ethanol and 50\u00a0mL DI water three times, respectively, and dried at 60\u00a0\u00b0C in a vacuum. The magnetic CC-Fe/C was synthesized by annealing MIL-88B(Fe) under an Ar atmosphere with a heating rate of 5\u00a0\u00b0C/min at different temperatures (700, 800, or 900\u00a0\u00b0C) for 5\u00a0h. Then the black products were cooled down to room temperature, collected, and labeled as CC-Fe/C-X, where CC refers to coral-like core-shell, and X refers to annealing temperature (See Scheme 1\n).The morphology of the catalyst was characterized by a field emission scanning electron microscope (FESEM, Sigma 300, Zeiss, Germany) with energy dispersive X-ray spectroscopy (EDS) and a high-resolution transmission electron microscope (HRTEM, Tecnai G2 F30, FEI, America). The thermogravimetric analysis (TGA) was performed on a HITACHI STA200 simultaneous thermal analyzer with a heating rate of 5\u00a0\u00b0C/min from room temperature to 1000\u00a0\u00b0C in Ar. The crystal phase of the catalyst was analyzed by an X-ray diffractometer (XRD, D8 advance, Bruker, Germany) equipped with a Ni-filtered Cu K radiation. The Brunauer-Emmett-Teller (BET) surface area and pore structure of the catalyst was determined by the N2 adsorption/desorption method on a Genini 2390 analyzer (Micromeritics, America). The surface composition of the materials was characterized by X-ray photoelectron microscopy (XPS, Thermo Scientific K-Alpha, Thermo Fisher, America) with Al-K\u03b1 (1486.6\u00a0eV) radiation. The microstructural properties of samples were analyzed by Raman spectroscopy (Thermo Scientific DXR 3Xi, America). The residual BPA after the catalytic reaction was analyzed in triplicates by an HPLC Agilent 1260, America) equipped with a C18 column. The mobile phase consisted of water/acetonitrile (v/v, 40:60) at a flow rate of 1.0\u00a0mL/min, and the UV detection wavelength was 276\u00a0nm. The instrument detection limit and the limit of quantification were 0.2\u00a0\u03bcg/mL and 50\u00a0\u03bcg/mL, respectively. The concentration of total iron ions leached out was measured by an inductively coupled plasma mass spectrometer (ICP-MS, NexION-300X, PerkinElmer). The reactive oxygen species (ROS) generated during PMS activation were identified by the electron paramagnetic resonance (EPR) spectra (Bruker EMXplus-6, Germany). PMS was colorimetrically determined based on the amount of iodine (\u03bb\nmax\u00a0=\u00a0352\u00a0nm) formed via the oxidation of iodide by PMS.The catalytic performance of CC-Fe/C samples was evaluated by degrading BPA, as the model pollutant. The degradation process was performed in a 250\u00a0mL conical flask on a constant temperature shaker with a speed of 250\u00a0r/min in triplicate at room temperature. In a typical run, 100\u00a0mL of 10\u00a0mg/L BPA solution was added into a 250\u00a0mL conical flask. Then, 10\u00a0mg of as-prepared catalytic materials were added to the above solution. The mixture was sonicated for 1\u00a0min and left standing for 30\u00a0min to achieve adsorption equilibrium. Afterward, 1.0\u00a0mL 10\u00a0mg/mL PMS solution was added to initiate the reaction, and the initial concentration of KHSO5 was 0.31\u00a0mM in the reaction solution. At given intervals, 1.0\u00a0mL of the reaction liquid was withdrawn, quenched with 0.1\u00a0mL (0.05\u00a0M) Na2S2O3, and filtered with a 0.22\u00a0\u03bcm membrane before analysis. To evaluate the influence of the initial solution pH on the degradation process, PMS and BPA were added to the working solution before adjusting the pH using 0.1\u00a0M HCl and NaOH. Then, the catalyst was added to start the reaction.FESEM was used to observe the morphologies of the synthesized MIL-88B(Fe) and the CC-Fe/C catalysts prepared at different annealing temperatures. Fig. 1a and e revealed the high uniformity of the MIL-88B(Fe) crystals. The MIL-88B(Fe) crystal appeared as a hexagonal prism with a truncated hexagonal cone at each end. The lateral dimension was about 500\u00a0nm. After the pyrolysis process, the MIL-88B(Fe) transformed into CC-Fe/C catalysts and its uniform crystal structure disappeared, indicating the complete structural decomposition of the MIL-88B(Fe). The CC-Fe/C-700 depicted a coral-like porous morphology (Fig. 1b), similar to the cauliflower coral (inset of Fig. 1c). The graphite-like clusters assembled the coral-like nanoarchitecture (Fig. 1f). Notably, CC-Fe/C-800 and CC-Fe/C-900 (Fig. 1c and d) exhibited coral-like structures similar to that of CC-Fe/C-700, except they had the more compact and blockier pattern. Furthermore, Fig. 1f showed visible pores on the surface of CC-Fe/C-700. The pore structure of samples with higher annealing temperatures (i.e., CC-Fe/C-800 and CC-Fe/C-900) became invisible (Fig. 1g and h). This phenomenon suggested that annealing temperature might influence the pore structure of the CC-Fe/C catalysts, which would be confirmed via the N2 adsorption-desorption isotherm (Fig. 4a). The elemental mappings (Fig. 1i-l) of CC-Fe/C-800 showed the presence of Fe, C, and O. Fe atoms in CC-Fe/C-800 were distributed uniformly, which was important for promoting the catalytic activity.The EDS results (Table S2) showed that, after annealing at 800\u00a0\u00b0C, the weight ratios of C and O in the MIL-88B(Fe) decreased from 51.61% and 26.05% to 36.85% and 3.08% in CC-Fe/C-800, respectively. This may be due to the evaporation or decomposition of organic components in the MIL-88B(Fe). Thus, TGA was conducted in an argon atmosphere to gain insight into the thermal decomposition demeanor of the MIL-88B(Fe). As exhibited in Fig. 1m, the weight loss of the MIL-88B(Fe) can be divided into two main stages. The first stage occurred between 25\u00a0\u00b0C and 300\u00a0\u00b0C, with loss of weight of about 20\u00a0wt%, can be ascribed to the evaporation and decomposition of solvent molecules (such as DMF, ethanol, and water) absorbed in the porous channel of the MIL-88B(Fe) [28]. The second 43\u00a0wt% weight loss stage located at 300\u2013650\u00a0\u00b0C was due to the complete structural decomposition of the MIL-88B(Fe) [19]. While the weight of the sample had no significant change with temperature increased from 650\u00a0\u00b0C to 900\u00a0\u00b0C. Therefore, 700, 800, and 900\u00a0\u00b0C were chosen as the annealing temperatures to ensure that CC-Fe/C catalysts were free of any residual solvents and the MIL-88B(Fe) precursor.According to XRD results (Fig. 2a), the synthesized MIL-88B(Fe) exhibited typical diffraction peaks at 9.2\u00b0, 10.2\u00b0, 16.8\u00b0, 18.3\u00b0, and 20.4\u00b0, which were consistent with the literature report [28]. After pyrolysis at 700\u2013900\u00a0\u00b0C, the diffraction peaks of MIL-88B(Fe) disappeared in the spectrum of CC-Fe/C samples. It suggested that the MIL-88B(Fe) has completely decomposed during the high-temperature pyrolysis. Meanwhile, the new diffraction peak at 26.4\u00b0 corresponded to the (002) plane of graphitic carbon, indicating the presence of graphitic structures in all three CC-Fe/C samples [29]. The graphitic peak intensity of CC-Fe/C-900 was slightly stronger than those of the other two samples due to its higher degree of graphitization resulting from elevated annealing temperature [30]. The peaks of sole Fe0 at 2\u03b8\u00a0=\u00a044.7\u00b0, 65.0\u00b0, and 82.3\u00b0 were assigned to the (110), (200), and (211) planes (JCPDF No. 06\u20130696), respectively [31]. The peaks at 2\u03b8\u00a0=\u00a037.8\u00b0, 39.9\u00b0, 41.0\u00b0, and 30.2\u00b0, 35.5\u00b0, 43.2\u00b0, 62.9\u00b0 were assigned to the crystalline phase of Fe3C (JCPDS 35\u20130772) [32], and spinel Fe3O4 (JCPDS No. 19\u20130629) [33], respectively, indicating that the Fe species in MIL-88B(Fe) were converted into different iron species during the pyrolysis. In the sample obtained at 900\u00a0\u00b0C (CC-Fe/C-900), the diffraction peaks of Fe3O4 nearly disappeared, while the peaks of Fe0 became sharper. Besides, according to the EDS results (Table S2), the atomic contents of C and O decreased with the annealing temperature rising from 700\u00a0\u00b0C to 900\u00a0\u00b0C. The ratio of the decreased C atomic content (1.07%) to the decreased O atomic content (0.88%) was 1:0.82, which was close to the stoichiometric ratio of C to O in CO (1:1). This phenomenon implied that the sharper XRD diffraction peaks of Fe0 in CC-Fe/C-900 was probably due to the reduction of Fe3O4 (Fe3O4\u00a0+\u00a04C\u00a0\u2192\u00a03Fe\u00a0+\u00a04CO) with the annealing temperature increased [34]. The above XRD results showed that the CC-Fe/C catalyst was a hybrid material composed of graphitic carbon, Fe0, Fe3C, and Fe3O4. The contents of the four components were affected by annealing temperature due to the redox reaction between these Fe and C species. The approximate weight contents of the four components in CC-Fe/C-800 sample were calculated (Text S1, Table S3). As shown in Fig. S1, the weight contents of graphitic carbon, Fe0, Fe3C, and Fe3O4 in CC-Fe/C-800 were 36.41%, 49.31%, 3.07%, and 11.02%, respectively. Finally, the magnetic CC-Fe/C catalyst can be easily recovered after using an external magnetic field (Fig. S2).Raman spectroscopy was used to evaluate the carbon matrix with graphite crystal structures in the CC-Fe/C samples. The D-band (\u223c1350\u00a0cm\u22121) and the G-band (\u223c1580\u00a0cm\u22121) were related to the disordered sp3 C atoms (or amorphous carbon) and the sp2 C atoms in both rings and chains, respectively [8]. The intensity ratio of the G to D band (I\nG/I\nD) indicates the degree of graphitic order in a carbon material [35]. According to Fig. 2b, the I\nG/I\nD ratios of CC-Fe/C-700, CC-Fe/C-800, and CC-Fe/C-900 were 0.87, 1.00, and 1.15, respectively. These results demonstrated that the graphitization degree was improved with increasing annealing temperature, which was in agreement with the XRD results (Fig. 2a). The 2D-band (\u223c2680\u00a0cm\u22121) is often used to estimate the number of graphene layers [36]. All three samples showed the characteristic 2D band. The CC-Fe/C-900 depicted the sharpest diffraction pattern, suggesting that increasing annealing temperature can promote the degree of graphitization and a high level of internal ordering.The TEM images (Fig. 3a and b) show variation in the contrast between the dark core and the light shell. As observed, the dark Fe-like cores (marked with yellow polygons) were well distributed in the carbon matrix. In the HRTEM image of CC-Fe/C-800 (Fig. 3c), the lattice distances of 0.241\u00a0nm and 0.294\u00a0nm were ascribed to the (210) and (220) planes of Fe3C and Fe3O4, respectively, which were consistent with the XRD analyses (Fig. 2a) [9,13]. According to the HRTEM images of the core-shell structure (Fig. S3), the lattice distances of 0.202\u00a0nm in the dark cores were ascribed to the (110) plane of Fe0, suggesting that the cores of CC-Fe/C-800 were Fe0 nanoparticles. Furthermore, the lattice spacing (0.335\u00a0nm) in the outer shell (Fig. 3d) was assigned to the (002) facet of graphitic carbon [32], indicating that the graphitic structure may endow Fe0 cores with an excellent protective ability which effectively prevented Fe leaching. In summary, the EDS, XRD, and HRTEM results together suggested that the main components of CC-Fe/C samples changed along with the thickness of the coating layer: the outermost shell consisted primarily of graphitic carbon; the sub-outer layer was made up of Fe3C and Fe3O4; while the interior cores were mainly Fe0 nanoparticles.The porosity and BET surface area (SBET) of the synthesized MIL-88B(Fe) and three CC-Fe/C samples were analyzed by N2 adsorption-desorption isotherms (Fig. 4a and Table 1\n). The pore size distributions were calculated by BJH (Barrett-Joyner-Halenda) method (Fig. 4b). As observed in Table 1, the pore volume of the four samples ranged from 0.089 to 0.266\u00a0cm3/g, indicating the existence of pores in them. Moreover, all four samples exhibited type IV isotherm with H3 hysteresis loops (Fig. 4a), reflecting the mesoporous structure of them [37]. While the iron-based nanoparticles with high crystallinity were likely non-porous, it suggested that the mesoporous characteristic mainly stemmed from the shell (porous graphitic carbon). The mesoporous shell can provide channels and facilitate the mass transfer of the reactants (such as PMS and BPA) from the catalyst surface to the interior cores [38]. Thus, depending on the porosity of the shell, the interior active sites (such as Fe0) can play the catalysis role directly.The SBET value (44.4\u00a0m2/g) and pore volume (0.102\u00a0cm3/g) of the MIL-88B(Fe) were small since the breathing nature of the MIL-88B(Fe) which presented closed pore structures in dry state [28]. However, after pyrolysis at 700\u00a0\u00b0C, the SBET value and pore volume of CC-Fe/C-700 increased to 135.14\u00a0m2/g and 0.266\u00a0cm3/g, respectively. It suggested that the decomposition of the MIL-88B(Fe) during the pyrolysis process was more likely to form pores, leading to an increase in SBET as well as pore volume. The SBET (46.2\u2013135.1\u00a0m2/g) and average pore width (4.7\u20135.9\u00a0nm) of the three CC-Fe/C samples were close to those of previously reported Fe/C porous materials (Table S4). Among the three catalytic samples, CC-Fe/C-900 obtained at 900\u00a0\u00b0C had fewer 2\u20133\u00a0nm and more 3\u20134\u00a0nm sized mesoporous. Besides, the CC-Fe/C-900 with the highest graphitization degree (based on the XRD and Raman results, Fig. 2) presented a lower SBET of 46.16\u00a0m2/g and a smaller pore volume of 0.089\u00a0cm3/g than CC-Fe/C-700 (135.14\u00a0m2/g, 0.266\u00a0cm3/g) and CC-Fe/C-800 (73.29\u00a0m2/g, 0.136\u00a0cm3/g). This phenomenon might be due to the increase of graphitization degree driven by the growing annealing temperature, which caused a decrease in the defects in graphitic layers and, thus, decreased SBET and pore volume [24].BPA was selected as the model pollutant to assess the catalytic activities of the four prepared samples (i.e., MIL-88B(Fe), CC-Fe/C-700, CC-Fe/C-800, and CC-Fe/C-900). After 30\u00a0min of the adsorption without PMS, the removal efficiencies of BPA were insignificant (less than 4%), indicating that all four samples were ineffective in adsorbing molecular BPA (Fig. 5a). When PMS was added, only 17.7% of BPA was removed by the PMS\u00a0+\u00a0MIL-88B(Fe) system within 60\u00a0min. The degradation efficiency was far less than those achieved by the CC-Fe/C samples (85.1 to 100%), elucidating that the annealing process significantly improved the catalytic activity of the MIL-88B(Fe) precursor. Herein, the influence of the annealing temperature of CC-Fe/C on PMS activation was evaluated. As shown in Fig. 5a, the removal rate of BPA reached 100%, 78.3%, and 59.7%, in 20\u00a0min in PMS\u00a0+\u00a0CC-Fe/C-800, PMS\u00a0+\u00a0CC-Fe/C-900, and PMS\u00a0+\u00a0CC-Fe/C-700 systems, respectively. The degradation kinetics can be described using a pseudo-second-order model (Fig. S4a). It was found that the rate constant k of CC-Fe/C-800 (1.18\u00a0mM\u22121\u00a0min\u22121) was much higher than those of CC-Fe/C-700 (0.17\u00a0mM\u22121\u00a0min\u22121) and CC-Fe/C-900 (0.35\u00a0mM\u22121\u00a0min\u22121). Among the three samples, CC-Fe/C-700 exhibited the lowest catalytic activity, although it owned the largest SBET and pore volume. Herein, the specific activity comparison among the three catalysts was undertaken. As depicted in Table 1, the specific activity of CC-Fe/C-800 (0.0161\u00a0mM\u00a0min\u22121\u00a0m\u22122) was higher than those of CC-Fe/C-700 (0.0013\u00a0mM\u00a0min\u22121\u00a0m\u22122) and CC-Fe/C-900 (0.0075\u00a0mM\u00a0min\u22121\u00a0m\u22122). It implied that the specific surface area of the catalysts was not the key factor affecting the catalytic activity. The high catalytic performance of CC-Fe/C-800 was due to the multiple Fe-based active sites (Fe0, Fe3C, and Fe3O4) generated by the appropriate annealing temperature (800\u00a0\u00b0C) rather than its relatively high SBET. Based on its excellent catalytic activity, CC-Fe/C-800 was selected in the following experiments.The effects of catalyst dosage, PMS concentration, initial BPA concentration, and solution pH on BPA removal were then investigated. Changing the CC-Fe/C-800 dosage had a negligible impact on the adsorption ability of BPA but significantly influenced the BPA degradation (Fig. 5b). Increasing the catalyst dosage resulted in a higher BPA degradation. After applying 0.06 and 0.08\u00a0g/L of CC-Fe/C-800, the removal rate was about 67.7 and 91.3% in 60\u00a0min, respectively. While at 0.10 and 0.12\u00a0g/L of catalyst dosage, BPA was completely degraded in 40 and 20\u00a0min with rate constants k of 1.18 and 7.04\u00a0mM\u22121\u00a0min\u22121 (Fig. S4b), respectively. The enhanced degradation efficiency can be ascribed to the availability of abundant active sites with increasing catalyst dosage, which ultimately activated PMS into more ROS [26]. Furthermore, the effect of PMS concentration on BPA removal was studied (Fig. 5c). With PMS concentration at 0.19\u20130.38\u00a0mM, 100% of BPA was degraded in 20\u00a0min. However, the removal efficiency decreased (78.6% in 60\u00a0min) at a PMS dosage of 0.09\u00a0mM, probably due to the total consumption of oxidant that can not continuously provide ROS to degrade BPA [35].The effect of changing the initial BPA concentrations (5 to 30\u00a0mg/L) on the degradation efficiency was shown in Fig. 5d. BPA removal rate decreased with increasing its initial concentration. For low initial concentrations (5 and 10\u00a0mg/L), complete degradation was achieved in 20\u00a0min. However, the degradation rate dropped to 79.8% and 54.1% in 60\u00a0min at initial BPA concentrations of 20 and 30\u00a0mg/L, respectively. This phenomenon might be attributed to the possible coverage of reactive surface sites by excess BPA molecules, leading to decreased PMS activation and an insufficient amount of ROS generated [35]. Due to the solution pH significantly influencing the catalytic performance of the SR-AOPs system [39], the catalytic activity at a wide pH range (3.8\u201311.4) was investigated. As shown in Fig. S5, the degradation efficiency was reduced obviously with increasing initial solution pH from acidic (pH\u00a0=\u00a03.8) to alkaline (pH\u00a0=\u00a011.4) conditions. Complete removal of BPA can be achieved in a 60\u00a0min reaction time at initial pH\u00a0=\u00a03.8, 6.7, and 9.3. However, when the initial solution pH was adjusted to 11.4, this SR-AOPs system exhibited a relatively lower removal (67.8% in 60\u00a0min) of BPA. For the strong alkaline condition (pH\u00a0=\u00a011.4), the possible hydroxylation of the catalyst surface by OH\u2212 might be unfavorable to the PMS adsorption due to the electrostatic repulsion and thus inhibited the degradation of BPA [40].The catalyst was recycled by a magnet and used in three cycles of BPA degradation under the same operating conditions. The degradation efficiencies of BPA after the first, second, and third cycles were 100%, 95.4%, and 81.7%, respectively (Fig. 6a). The used CC-Fe/C-800 was regenerated by vacuum drying (60\u00a0\u00b0C for 12\u00a0h) or thermal treatment (800\u00a0\u00b0C/h in Ar flow). As expected, the catalytic activity was partially recovered (86.9% in 60\u00a0min) by the vacuum-dried process. Significantly, the catalyst regenerated by thermal treatment in Ar degraded nearly 99.0% of BPA, showing its good renewability. According to the XRD patterns of the pristine and recycled CC-Fe/C-800 samples (Fig. S6), the diffraction peak intensities of Fe0 and Fe3C became weaker while those of Fe3O4 became stronger after the reaction. Even so, the diffraction peaks of all three iron-components (i.e., Fe0, Fe3C, and Fe3O4) still can be observed in the recycled catalyst sample. It suggested that the graphitic carbon shell of CC-Fe/C-800 may protect the interior Fe0/Fe3C/Fe3O4 nanoparticles, avoiding too much iron ion leaching from the catalyst into the solution. As expected, the ICP-MS results for Fe (Fig. S7) showed that the concentrations of leached iron ions in the catalytic system at a pH range of 5.0\u201310.8 were lower than 1\u00a0mg/L, which is within the safe limit as per the discharge standard of iron [41]. To further examine the catalyst's potential for practical application, the catalytic tests were performed in two kinds of actual water substrates (tap water and treated wastewater). As displayed in Fig. 6b, the removal rate slightly decreased from 100% to 98.8% and 95.0% using tap water and treated wastewater, respectively. The above results demonstrated that CC-Fe/C-800 depicted good reusability, stability, and a great prospect for water treatment.For identifying the possible ROS generated in the PMS\u00a0+\u00a0CC-Fe/C-800 system, an EPR test was conducted using DMPO (5,5-dimethyl-1-pyrrolidine N-oxide) as the spin trap agent. As shown in Fig. 7a, no peaks were detected in PMS alone solution in 1\u00a0min reaction, but characteristic peaks of DMPO-OH and DMPO-SO4 adducts appeared after adding CC-Fe/C-800 [42], indicating the generation of \u2219OH and SO4\n\u2219\n\u2212. Prolonging the reaction to 10\u00a0min, the peak intensities of DMPO-OH and DMPO-SO4 adducts showed little change, suggesting that CC-Fe/C-800 can continuously activate PMS. Especially, the slight decrease in intensity of DMPO-OH peaks might be attributed to the transformation of \u2219OH to 1O2 after a series of radical reactions [43]. Thus, TEMP (2, 2, 6, 6-tetra-methyl-4-piperidone) was added into the reaction solution as a sacrificial agent for 1O2 [44]. In Fig. 7b, a weak 1:1:1 triplet signal appeared in the first minute, indicating the presence of 1O2 in the SR-AOPs system. The intensity of these signals increased significantly in 10\u00a0min, which demonstrated that 1O2 was continuously generated.To further confirm the role of different ROS in the PMS\u00a0+\u00a0CC-Fe/C-800 system, a series of quenching experiments were performed by adding different concentrations of tert-butyl alcohol (TBA), methanol (MeOH), L-histidine (L-His), and p-benzoquinone (p-BQ), respectively. TBA could only effectively scavenge \u2219OH, whereas MeOH can react with both SO4\n\u2219\n\u2212 and \u2219OH [45,46]. As shown in Fig. 8a, the BPA removal efficiency decreased from 100% to 95.8%, 79.2%, and 56.5% within 60\u00a0min after adding 0.1\u00a0M, 0.3\u00a0M, and 0.5\u00a0M TBA, respectively. It suggests that \u2219OH was involved in the degradation but not the dominant reactive species. When excessive 0.1\u00a0M, 0.3\u00a0M, and 0.5\u00a0M MeOH were added (Fig. 8b), the removal efficiency of BPA decreased dramatically from 100% to 49.3%, 31.2%, and 18.7%, respectively. This significant inhibition phenomenon indicated that SO4\n\u2219\n\u2212 rather than \u2219OH would play an important role in the degradation of BPA. In the presence of L-His (a commonly used scavenger for 1O2), the reaction was also inhibited, as displayed in Fig. 8c. However, L-His greatly accelerated PMS loss in the catalytic system, inhibiting the degradation of BPA by complete consuming of PMS in 20\u00a0min (Fig. S8). Herein, another 1O2 scavenger, \u03b2-carotene, was used to evaluate the role of 1O2 in the catalytic reaction [47]. As displayed in Fig. 8d, the degradation efficiency of BPA decreased from 100% to 70.8% within 60\u00a0min in the presence of \u03b2-carotene, implying that 1O2 contributed to the BPA degradation. In addition, the degradation experiments were conducted using deuteroxide (D2O) as a solvent because 1O2 has a longer lifetime in D2O than in H2O due to the slower decaying rate as 1O2 in D2O [44]. The result in Fig. S9 showed that 92.3% of BPA was decontaminated in D2O in 10\u00a0min compared with 86.7% accomplished in H2O, which further proved the contribution of 1O2 to BPA degradation in the PMS\u00a0+\u00a0CC-Fe/C-800 system. Besides, the degradation efficiencies were slightly affected in the presence of p-BQ (scavenger of O2\n\u2212\n\u2219) with different concentrations, indicating the low generation of O2\n\u2212\n\u2219 during the PMS activation process (Fig. 8e) [48]. To explore the dominant ROS responsible for the oxidation of BPA, the inhibition rates of SO4\n\u2219\n\u2212, 1O2, \u2219OH, and O2\n\u2212\n\u2219 with BPA were compared according to the quenching test results (Fig. 8a-e) [49]. As shown in Fig. 8f, the inhibition ratio of BPA degradation by different scavengers was ranked as follows: MeOH (81.3%)\u00a0>\u00a0TBA (43.5%)\u00a0>\u00a0\u03b2-carotene (29.2%)\u00a0>\u00a0p-BQ (8.5%). It suggested that the contribution of different ROS to BPA degradation in the PMS\u00a0+\u00a0CC-Fe/C-800 system should follow the order as SO4\n\u2219\n\u2212\u00a0>\u00a0\u2219OH\u00a0>\u00a01O2\u00a0>\u00a0O2\n\u2212\n\u2219.XPS was then applied to investigate the surface state changes of the CC-Fe/C-800 after activating PMS for BPA removal. According to the survey spectrum of fresh CC-Fe/C-800 (Fig. S10a and Table S5), three elements of Fe, C, and O were contained in the composite with surface atomic contents of 3.1%, 88.8%, and 8.1%, respectively. Significantly, the Fe content (3.1 atom%) was much lower than by EDS analysis (25.0 atom%, Table S2) since XPS can only examine the outer 3\u20135\u00a0nm of samples. In contrast, EDS can probe several micron depths beneath the material. This phenomenon suggested that CC-Fe/C-800 is a composite of an iron core covered by a thin graphitic carbon layer, consistent with the previous discussion (Section 3.1). After the catalytic reaction, the surface C content slightly decreased (from 88.8% to 76.2%), whereas the Fe and O contents increased to various degrees (Table S5), implying that the carbon (such as Fe3C) might be involved in the PMS activation.It is well known that Fe species can activate PMS into ROS [8]. Thus, the high-resolution Fe 2p XPS spectra of fresh and used CC-Fe/C-800 were carried out to estimate the possible catalytic mechanism. For the fresh sample (Fig. 9a), the peaks at binding energies of 708.9, 710.6, 713.4, 723.9, and 727.6\u00a0eV were assigned to Fe3C, Fe2+ 2p3/2, Fe3+ 2p3/2, Fe2+ 2p1/2, and Fe3+ 2p1/2, respectively [9]. Two satellite peaks of Fe 2p at 718.1 and 734.0\u00a0eV were fitted. The peak of Fe0 at around 706\u00a0eV was not detected in the XPS spectrum due to the rapid formation of iron oxide on the catalyst surface, and the XPS is only limited to surface detection [50]. The Fe 2p XPS spectrum of the used CC-Fe/C-800 (Fig. 9b) revealed an increase in the content of Fe3+. The weakening of the Fe3C and Fe2+ signals indicated that part of Fe3C and Fe2+ transformed to Fe3+. However, the percentage of Fe2+ did not significantly decrease (42.1% to 40.5%) after the reaction. This phenomenon might be due to the FeC bond and graphitic carbon in CC-Fe/C-800 that facilitated the electron transfer from Fe0 to Fe3+, leading to an efficient regeneration of Fe2+ [51,52]. In other words, there was Fe2+/Fe3+ cycling on the catalyst surface during the catalytic process [53].The C 1s spectra for the fresh and used CC-Fe/C-800 showed four characteristic peaks (Fig. 9c and d). The strong peaks at about 283.7\u00a0eV were ascribed to the CFe bonds (from the Fe3C phase in the carbon support) [54], confirming the presence of the Fe3C in the CC-Fe/C catalyst. The peaks at around 284.4\u00a0eV were attributed to CC derived from the sp2 hybrid graphite, revealing that another major part of C was present in cross-lined cellular lattices [53]. Moreover, two weak peaks of CO (at 285.2\u00a0eV) and CO (at 288.3\u00a0eV) were also observed [15]. After the reaction, the CO peak intensity enhanced, indicating the catalyst surface was hydroxylated. Meanwhile, the intensity of the CFe bond decreased from 74.0% to 56.0%, implying the involvement of Fe3C in the reaction through enhancing positive charge density on the adjacent carbon atoms that improved the nucleophilic addition of PMS for 1O2 generation [55].In the O 1s spectra (Fig. 9e and f), the peaks at 529.4, 531.9, and 532.8\u00a0eV, were assigned to the lattice oxygen from metal oxides (Olat), the adsorbed oxygen (Oads), and the adsorbed water molecule (Osurf), which accounted for 43.1%, 45.9%, and 11.0% before reaction, respectively [56]. After the reaction, the contents of Olat, Oads, and Osurf changed to 34.2%, 44.4%, and 21.3%, respectively. The consumption of Olat might be due to the reduction of Fe3+ to Fe2+ by electrons donated from the Fe0 inside CC-Fe/C-800 [52], and the increase in Osurf (11.0% to 21.3%) can be assigned to the H2O adsorbed on the catalyst surface after the heterogeneous reaction.Combined with the results of EPR, quenching, and XPS tests, the possible reaction mechanisms of PMS activation on CC-Fe/C-800 can be proposed, as exhibited in Fig. 10\n. Firstly, HSO5\n\u2212 (dissolved PMS) and BPA molecules were adsorbed onto the porous carbon surface of CC-Fe/C-800. At the same time, Fe3C in the sub-outer layer could modify the electron states of the adjacent carbon regions, resulting in a higher positive charge density on the adjacent carbon atoms (Eq. (1)) [57]. Then, the positively charged carbon was prone to nucleophilic addition of HSO5\n\u2212 for generating 1O2 (Eq. (2)) [55]. Secondly, Fe3C nanoparticles could react with HSO5\n\u2212 via the transformation of Fe0 and Fe3+ and subsequently undergo a series of complex radical chain reactions to produce O2\n\u2212\n\u2219 (Eq. (3)). The latter will recombine with another O2\n\u2212\n\u2219 and generate 1O2 via Eq. (4) [58].\n\n(1)\n\n\nFe\n3\n\nC\n-\nC\n\n\n-\n\n\ne\n\u2212\n\n\u2192\n\nFe\n3\n\nC\n\u2212\n\nC\n+\n\n\n\n\n\n\n(2)\n\n2\n\nHSO\n5\n\u2212\n\n\n\u2192\n\n\nFe\n3\n\nC\n\u2212\n\nC\n+\n\n\n\n\n\nO\n\n\n1\n\n2\n\n+\n2\n\nH\n+\n\n+\n2\n\nHSO\n4\n\n2\n\u2212\n\n\n\n\n\n\n\n(3)\n\n\nFe\n3\n\nC\n\n\u2192\n\nHSO\n5\n\u2212\n\n\n\nFe\n0\n\n\nand\n\n\nFe\n\n3\n+\n\n\n\n\u2192\n\nHSO\n5\n\u2212\n\n\n\nO\n2\n\u2212\n\n\u22c5\n\n\n\n\n\n(4)\n\n2\n\nO\n2\n\u2212\n\n\u22c5\n+\n2\n\nH\n+\n\n\u2192\n\n\nO\n\n\n1\n\n2\n\n+\n\nH\n2\n\n\nO\n2\n\n\n\n\nMeanwhile, the Fe0 cores inside CC-Fe/C-800 also participated in the activation of PMS. According to the previous literature [59,60], nano Fe0 was easily oxidized by dissolved oxygen, leading to a passivating oxide layer (\u2261Fe2+) on the catalyst surface and the release of dissolved Fe2+ via Eq. (5). Subsequently, these as-formed \u2261Fe2+ and dissolved Fe2+ activated HSO5\n\u2212 to produce SO4\n\u2219\n\u2212 or \u2219OH according to Eq. (6\u20139), respectively [39]. The EPR and quenching test (Fig. 7a and Fig. 8a) showed that some of SO4\n\u2219\n\u2212 would react with H2O to produce \u2219OH via Eq. (10) [61]. In addition to Fe3C and Fe0, the \u2261Fe2+-containing Fe3O4 nanoparticles can also contribute to PMS activation through Eq. (6) and (7) [10,21].\n\n(5)\n\n\nFe\n0\n\n+\n\nO\n2\n\n+\n4\n\nH\n+\n\n\u2192\n2\n\u2261\n\nFe\n\n2\n+\n\n\n/\n\nFe\n\n2\n+\n\n\n+\n\nH\n2\n\nO\n\n\n\n\n\n(6)\n\n\u2261\n\nFe\n\n2\n+\n\n\n+\n\n\nHSO\n5\n\u2212\n\n\u2192\n\u2261\n\nFe\n\n3\n+\n\n\n+\n\n\nSO\n4\n\u2212\n\n\u22c5\n+\n\nOH\n\u2212\n\n\n\n\n\n\n(7)\n\n\u2261\n\nFe\n\n2\n+\n\n\n+\n\n\nHSO\n5\n\u2212\n\n\u2192\n\u2261\n\nFe\n\n3\n+\n\n\n+\n\n\nSO\n4\n\n2\n\u2212\n\n\n+\n\u22c5\nOH\n\n\n\n\n\n(8)\n\n\nFe\n\n2\n+\n\n\n+\n\n\nHSO\n5\n\u2212\n\n\u2192\n\nFe\n\n3\n+\n\n\n+\n\n\nSO\n4\n\u2212\n\n\u22c5\n+\n\nOH\n\u2212\n\n\n\n\n\n\n(9)\n\n\nFe\n\n2\n+\n\n\n+\n\n\nHSO\n5\n\u2212\n\n\u2192\n\nFe\n\n3\n+\n\n\n+\n\n\nSO\n4\n\n2\n\u2212\n\n\n+\n\u22c5\nOH\n\n\n\n\n\n(10)\n\n\nSO\n4\n\u2212\n\n\u22c5\n\n+\n\n\nH\n2\n\nO\n\u2192\n\nSO\n4\n\n2\n\u2212\n\n\n\n+\n\n\u22c5\nOH\n\n+\n\n\nH\n+\n\n\n\n\nFurthermore, the Fe0 cores would not only participate in activating HSO5\n\u2212 but also act as a cocatalyst to promote Fe2+ regeneration. It is generally accepted that Fe2+ can be regenerated by the reaction between Fe3+ and excess HSO5\n\u2212 (Eq. (11)), but the reaction rate is very slow [62]. However, the Fe0 in CC-Fe/C-800 can efficiently facilitate Fe3+ reduction to Fe2+ via Eq. (12) and accelerate the Fe3+/Fe2+ cycle [52], significantly improving the catalytic activity. Besides, the graphitic carbon, an excellent platform, and electron reservoir would enhance the electron transfer between the above-mentioned Fe active sites (Fe3C, Fe0, and Fe3O4) and HSO5\n\u2212, leading to the generation of more ROS. Afterward, the ROS (i.e. SO4\n\u2219\n\u2212, \u2219OH, 1O2, and O2\n\u2212\n\u2219) took part in the degradation of BPA into intermediates (such as muconic, oxalic, and malonic acids), which were finally mineralized to CO2 and H2O (Eq. (13)) [35].\n\n(11)\n\n\nFe\n\n3\n+\n\n\n+\n\n\nHSO\n5\n\u2212\n\n\u2192\n\nFe\n\n2\n+\n\n\n+\n\n\nSO\n5\n\u2212\n\n\u22c5\n+\n\nH\n+\n\n\n\n\n\n\n(12)\n\n\nFe\n0\n\n+\n\n2\n\nFe\n\n3\n+\n\n\n\u2192\n3\n\nFe\n\n3\n+\n\n\n\n\n\n\n\n(13)\n\nBPA\n\n\u2192\nROS\n\n\nintermediates\n\n\n\u2192\nROS\n\n\n\nCO\n2\n\n+\n\nH\n2\n\nO\n\n\n\nIn summary, coral-like CC-Fe/C catalysts with Fe0/Fe3C/Fe3O4 nanoparticles wrapped in porous carbon shell were synthesized for PMS activation. Characterization results showed that increasing the annealing temperature (700\u2013900\u00a0\u00b0C) would increase Fe0 (main active site) content but reduce the specific surface area of CC-Fe/C catalysts. CC-Fe/C-800 prepared at 800\u00a0\u00b0C exhibited the best catalytic activity toward a wide pH condition (5.0\u201310.8) with low Fe leaching (less than 1\u00a0mg/L). 10\u00a0mg/L BPA was completely degraded in PMS\u00a0+\u00a0CC-Fe/C-800 system in 20\u00a0min. Moreover, the catalyst can be easily recovered using a magnetic field and showed good practicability. Mechanism study indicated that the excellent catalytic activity of CC-Fe/C-800 was mainly attributed to the synergistic effect of multiple active sites (Fe0, Fe3C, and Fe3O4) on its surface. During the catalytic oxidation of BPA process, SO4\n\u2219\n\u2212 played a dominant role rather than \u2219OH, 1O2, or O2\n\u2212\n\u2219. The Fe0 cores in CC-Fe/C-800 were beneficial for Fe3+/Fe2+ recycling. This work may offer a good reference to tune the physicochemical properties of Fe-MOFs derived Fe/C materials in terms of annealing temperature for improving the catalytic activity and provide new insight into the underlying reaction mechanism of the heterogeneous SR-AOPs system.\nJie Yu: Conceptualization, Supervision, Writing \u2013 original draft, Funding acquisition. Shahzad Afzal: Investigation, Writing \u2013 original draft. Tao Zeng: Investigation, Writing \u2013 review & editing, Formal analysis. He Wang: Methodology, Validation, Writing \u2013 review & editing. Hailu Fu: Conceptualization, Writing \u2013 review & editing, Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was financially supported by the National Natural Science Foundation of China (No. 22108265, 22276172), Zhejiang Provincial Natural Science Foundation of China (No. LTGS23E080006, LR21E080001), and the Fundamental Research Funding Project of Zhejiang Province (No. 2022YW25).\n\n\nSupplementary material\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2023.106660.", "descript": "\n                  Fe0/Fe3C/Fe3O4 nanoparticles wrapped in graphite shells (CC-Fe/C) were synthesized via pyrolyzing MIL-88B(Fe) at 700, 800, and 900\u00a0\u00b0C. CC-Fe/C-800 prepared at 800\u00a0\u00b0C exhibited the best performance for activating peroxymonosulfate with complete removal of bisphenol A at 10\u00a0mg/L in 20\u00a0min, attributing to the synergistic effect of multiple active sites (Fe0, Fe3C, and Fe3O4) on its surface. Mechanism study suggested that SO4\n                     \u2219\n                     \u2212, \u2219OH, 1O2, and O2\n                     \u2212\n                     \u2219 were involved in the degradation. Fe0 cores could act as cocatalysts to promote the regeneration of Fe2+, enhancing the catalytic activity. Finally, CC-Fe/C-800 showed good reusability and practicability.\n               "}
{"full_text": "Biomass-to-GasBrunauer-Emmett-Teller modelDual fluidized bedGreenhouse gasKey performance indicatorspiping and instrumentation diagramPower-to-Gasraw synthetic natural gas after methanation/before upgradingsorption enhanced reformingstoichiometric number of the feed gassynthetic natural gastemperature-programmed oxidationtemperature-programmed reductionweight hourly space velocity in Nl/g\u00a0hdry basismean Sauter diameter in \u03bcmin the feed gas to the methanation reactorreaction enthalpy at standard conditions in kJ/molmolar flow of species i in mol/snumber of carbon atoms in species i\nin the outlet of the methanation reactorselectivity of CO towards CO2 in %selectivity of C2H6 towards C2H4 in %minimum fluidization velocitycarbon monoxide conversion in %carbon dioxide conversion in %hydrogen conversion in %methane yield in %molar fraction of species i\nbulk density in kg/m3\nMany industrial high-temperature processes and domestic residences rely on the supply of natural gas as an energy carrier [1,2]. However, the targets formulated by the European Commission will require a substantial reduction in the use of fossil fuels in the future [3]. The conversion of biogenic feedstock to renewable synthetic natural gas (SNG) offers the possibility of producing a chemically and physically almost identical gas that can be transported in the already existing gas distribution infrastructure and utilized with already established end-use technologies [4].Catalytic methanation processes have been studied and developed for more than 100\u00a0years since Sabatier and Senders first discovered that noble metals catalyze methanation reactions. In the 1970s and 1980s, the primary focus lay on the conversion of coal to SNG. Due to the rising awareness of climate change and the urgent need to reduce GHG emissions, renewable alternatives for SNG production have been developed. Biomass-to-Gas (BtG) as well as Power-to-Gas (PtG) routes gained importance [4,5]. Besides catalytic methanation concepts, biological methanation approaches attract more and more attention [6,7]. Today, various process concepts exist that aim at an optimized production of SNG. One possible production route is the dual fluidized bed (DFB) gasification of woody biomass or waste materials and consecutive fluidized bed methanation. A primary advantage of the DFB process is that it produces a nitrogen-free syngas that is well suited for downstream synthesis processes. At TU Wien, a 100 kWth advanced DFB pilot plant has been developed and extensively investigated [8]. The investigations show that the new design allows the utilization of various waste resources and significantly impacts the quality of the syngas, which in turn affects the downstream synthesis processes [9,10]. However, due to the typical composition of woody biomass, the production of a stoichiometric syngas for methanation with a H2/CO ratio of three is impracticable and thus further measures must be taken. Sorption enhanced reforming (SER) is an alternative operation mode of the advanced DFB process, where the stoichiometric ratio of H2 to CO and CO2 is influenced and can be adapted to the needs of the downstream synthesis process. Fuchs et al. showed that both the gasification temperature [11] and the bed material circulation rate [12] have a major impact on the gas composition. By utilizing this syngas in the methanation reactor, it is theoretically possible to produce grid feedable SNG without the need for a CO2 separation unit [13]. So far, the investigations on the suitability of these syngases (advanced DFB and SER) for methanation have only been of theoretical nature [13\u201315]. In a modelling approach, both Bartik et al. [13] and Brellochs [14] state optimal gasification temperatures in the range of 680\u00a0\u00b0C\u2013700\u00a0\u00b0C for the production of SNG via SER. Since these studies are of theoretical nature, an objective of this work is the experimental investigation and evaluation of syngas from the advanced DFB pilot plant in a fluidized bed methanation unit. The fluidized bed methanation unit is designed to allow an isothermal operation of the methanation process through internal particle circulation while not disturbing the bubble formation and the gas/solid contact. A more detailed description of the reactor is shown in Section 2.1. Since the DFB and the SER processes are not part of the experimental investigations in this study, literature is referred to [8,9,16\u201321].During catalytic methanation, H2 and CO react to CH4 and H2O according to Eq. 1. The water-gas shift reaction (Eq. 2) leads to the formation of CO2 and H2 if the syngas shows a low H2/CO ratio. Vice versa, if the syngas shows an overstoichiometric composition, CO2 and H2 react via the reversed water-gas shift reaction and form CH4 and H2O in combination with Eq. 1.\n\n(1)\n\nCO\n+\n3\n\nH\n2\n\n\u21cc\nC\n\nH\n4\n\n+\n\nH\n2\n\nO\n\n\n\n\n\n\n\u0394\n\nH\nR\n0\n\n=\n\u2212\n206\n\nkJ\nmol\n\n\n\n\n\n\n(2)\n\nCO\n+\n\nH\n2\n\nO\n\u21cc\nC\n\nO\n2\n\n+\n\nH\n2\n\n\n\n\n\n\n\n\u0394\n\nH\nR\n0\n\n=\n\u2212\n41\n\nkJ\nmol\n\n\n\n\nEspecially for syngas with a low H2/CO ratio, the Boudouard reaction (Eq. 3) plays an important role. Carbon may form on the catalyst surface and block or infiltrate active reaction sites [22].\n\n(3)\n\n2\nCO\n\u21cc\nC\n\nO\n2\n\n+\nC\n\ns\n\n\n\n\n\n\n\n\u0394\n\nH\nR\n0\n\n=\n\u2212\n172\n\nkJ\nmol\n\n\n\n\nOther species often found in the syngas of the DFB or SER process are hydrocarbons like ethylene (C2H4). Ethylene can be hydrogenated to ethane (C2H6) (Eq. 4) and further to methane (Eq. 5) but can also lead to coke deposits on the catalyst [23]. The behavior of ethylene very much depends on the applied conditions and the type of reactor [24].\n\n(4)\n\n\nC\n2\n\n\nH\n4\n\n+\n\nH\n2\n\n\u2192\n\nC\n2\n\n\nH\n6\n\n\n\n\n\n\n\n\u0394\n\nH\nR\n0\n\n=\n\u2212\n137\n\nkJ\nmol\n\n\n\n\n\n\n(5)\n\n\nC\n2\n\n\nH\n4\n\n+\n2\n\nH\n2\n\n\u2192\n2\nC\n\nH\n4\n\n\n\n\n\n\n\n\u0394\n\nH\nR\n0\n\n=\n\u2212\n202\n\nkJ\nmol\n\n\n\n\nAll these reaction equations are highly exothermic, and large quantities of heat need to be removed. For this purpose, reactor concepts have been developed to cope with this issue [25]. Fluidized beds are known for their high heat and mass transfer capabilities due to the movement of the particles [26]. Hence, fluidized beds have been under investigation for catalytic methanation processes since 1950, as a review by Kopyscinski describes in detail [4]. The Paul Scherrer Institute recently picked up on the developments and investigated the fluidized methanation process more closely. They applied spatially resolved concentration and temperature measurements along the height of the catalytic bed. The results show that the particle movement leads to an in-situ regeneration of the catalyst particles and therefore reduces the risk for carbon depositions even in the presence of ethylene [24,27]. Furthermore, they concluded that the mass transfer between the bubble phase and the dense phase is a limiting factor in the upper part of the bed [28]. Seemann et al. [29] utilized a 10\u00a0kW fluidized bed reactor and demonstrated the conversion of syngas from the 8\u00a0MW DFB plant in G\u00fcssing. They reached around 40 vol.-% CH4 for a period of 200\u00a0h until a sulfur breakthrough was detected. Witte et al. [30] applied the same reactor setup to convert biogas from a digester with hydrogen to SNG and showed a stable long-term operation for >1000\u00a0h. On a larger scale, Hervy et al. [31] demonstrated CO2 methanation in a 400\u00a0kW fluidized bed methanation reactor. They proved that a high conversion efficiency could be maintained despite temperature and load variations.Despite these advantages, fluidized beds impose mechanical stress on catalyst particles. Thus, the development of an attrition-resistant catalyst with a proper fluidization behavior is necessary, which has not been considered or documented in the investigations mentioned above. In general, a significant amount of research has been put into the development of methanation catalysts, as some reviews show [32,33]. However, only a few investigations focus on the application in fluidized beds or the use of \u03b1-Al2O3 as catalyst support. Typically, \u03b3-Al2O3 is used because of its high surface area and the highly dispersed metal particles, while \u03b1-Al2O3 is often disregarded because of its low surface area and weak metal-support interaction [33]. For fluidized bed applications, Cui et al. [34] added different binders to improve the attrition resistance of the produced catalyst and found that acidic silica sol showed the highest resistance. Other investigations proved the superiority of fluidized beds over fixed beds in terms of conversion rates and coking resistance in small lab-scale test rigs [35,36]. However, a holistic approach, considering the catalytic activity in combination with an optimal fluidization behavior in a representative fluidized bed reactor scale, seems to be missing.This work investigates the catalytic methanation process in a 10\u00a0kW bubbling fluidized bed methanation reactor utilizing an optimized catalyst for fluidized bed applications. In contrast to the commonly used \u03b3-Al2O3, an attrition-resistant \u03b1-Al2O3 with a high specific surface area and improved fluidization properties is utilized. Besides the determination of the reactor and catalyst performance, the proposed concept is applied to systematically investigate the methanation of premixed gases imitating syngas from the advanced DFB pilot plant. The goal is to demonstrate that (raw) SNG production via SER and fluidized bed methanation with a tailored catalyst can lead to an optimized process chain with technical and economic advantages.\nFig. 1\n shows a 3D-CAD drawing of the fluidized bed reactor setup designed for the catalytic methanation of syngas. The reactor consists of two separate reaction zones operated in the bubbling fluidization regime. Both zones can be fluidized individually via two separate wind boxes. The gas distributor consists of nozzles, which provide the necessary pressure drop for uniform gas distribution. Both reaction zones are cooled individually to manage the heat released by the exothermic reaction. An air perfused coil cools the inner reaction zone, while a cooling jacket is used to cool the annular reaction zone. Thus, an isothermal operation of the methanation reactor is ensured. At the same time, the fluidization in the two reaction zones is not disturbed by internals. The catalyst, however, can move freely between the zones through the \u2018upper gap\u2019 and the \u2018lower gap\u2019 as denoted in Fig. 1. More information on the reactor setup is documented in [37].\nFig. 2\n shows a simplified piping and instrumentation (P&I) diagram of the reactor setup. In the lower-left part of the diagram, the gases (N2/H2/CO/CO2/CH4/C2H4) are withdrawn from gas cylinders and premixed according to the volume flow set by valves and rotameters. After splitting and preheating, the gas stream enters the wind boxes. Here, water vapor can be added if the syngas composition requires so. The reaction zones are equipped with thermocouples type K to measure the axial temperature distribution along the reactor height. The gas outlet is equipped with a particle filter and downstream the raw-SNG is burnt in a flare. The gas compositions of the syngas input and the raw-SNG output are analyzed online, as described in section 2.3.The catalyst contained 20 wt.-% NiO and 2 wt.-% MgO and was produced in 6 batches, following the preparation method of Hu et al. [38]. The reagents used for this are listed in the supplementary material (chapter A). Nickel nitrate hexahydrate and magnesium nitrate hexahydrate were dissolved (approx. 300\u00a0ml) in water and afterward the support was added. The solution was heated and stirred until the excess water was evaporated. The powder was dried overnight at 120\u00a0\u00b0C and calcined for 4\u00a0h at 500\u00a0\u00b0C with a heating ramp of 5\u00a0\u00b0C/min. The used support was a Puralox SCCa-150/200 \u03b1-Al2O3 from SASOL, which is in particular designed for fluidized bed applications and thus exhibits a high level of attrition resistance. Despite the high calcination temperatures typical for \u03b1-Al2O3, the material is reported to have a high surface area [39].A MicrotracBEL Catalyst Analyzer Belcat-II was used for the temperature-programmed reduction (TPR) and the pulse chemisorption measurements of CO and H2 on the catalyst sample. N2 physisorption was performed in a Micromeritics ASAP 2020 Serial # 1455 for the measurement of the surface area using the BET model. To determine possible carbon depositions on the catalyst, temperature-programmed oxidation (TPO) experiments were carried out. More information on the experimental measurement procedure can be found in the supplementary material (chapter B).The particle size distribution and the Sauter diameter (dSV) of the catalyst are determined with a Malvern Instruments Mastersizer 2000 laser diffraction particle size analyzer. A Rosemount NGA 2000 gas analyzer is used to measure H2, CO, CO2 and CH4 concentrations in the feed gas and the raw-SNG. Another NGA 2000 module with a low measurement range (< 5000 ppmv) is used for the reliable detection of low CO concentrations in the raw-SNG. Additionally, a Perkin Elmer ARNEL \u2013 Clarus 500 gas chromatograph (GC) detects ethane (C2H6), ethylene (C2H4), acetylene (C2H2), propane (C3H8), and nitrogen (N2) quantitatively. Higher hydrocarbons (C3+) are qualitatively detected. Furthermore, the GC redundantly measures the CO, CO2, and CH4 concentrations, which are included in the data evaluation.For the fluidized bed methanation experiments, 1.6\u00a0kg of the prepared Ni/\u03b1-Al2O3 catalyst and 1.5\u00a0kg of the unimpregnated \u03b1-Al2O3 support were used in the reactor. This amounts to an unfluidized bed height of about 20\u00a0cm. The catalyst was heated up at a rate of approximately 3\u00a0\u00b0C/min and reduced at 500\u00a0\u00b0C for 6\u00a0h in a 9:1 volume-based hydrogen to nitrogen atmosphere.For the purpose of this investigation, the gas velocity and the mean temperature in the inner and the annular reaction zone were kept equal to each other for all experiments. Furthermore, the gas preheating temperature was set to 250\u00a0\u00b0C, and the pressure was equal to ambient conditions for all experiments. All temperatures given in section 3 are mean values computed from two measurements in the inner reaction zone and three measurements in the annular reaction zone, as indicated in Fig. 2.The following syngas compositions are tested:\n\n\u2022\nStoichiometric H2/CO syngas for the determination of the catalyst and reactor performance (H2/CO).\n\n\n\u2022\nTypical DFB syngas composition from the 100 kWth advanced DFB pilot plant (DFB).\n\n\n\u2022\nThe flexible SER syngas composition from the 100 kWth advanced DFB pilot plant (SER).\n\n\nStoichiometric H2/CO syngas for the determination of the catalyst and reactor performance (H2/CO).Typical DFB syngas composition from the 100 kWth advanced DFB pilot plant (DFB).The flexible SER syngas composition from the 100 kWth advanced DFB pilot plant (SER).\nTable 1\n lists the gas compositions for these cases. The exact compositions of the SER syngases are given in the supplementary material (chapter C), following the work of Fuchs et al. [11]. Additionally, the parameter variations carried out for each gas composition are depicted. A variation of the C2H4 content was investigated because the syngas of the advanced DFB gasifier exhibits varying concentrations depending on factors such as temperature, bed material, and type of biomass [40]. Because of the different gas compositions and reaction conditions, the fluidization number varies between 1.8 and 7.8 u\n\nmf\n. Especially the variation of the weight hourly space velocity (WHSV) results in the most pronounced influence on the fluidization number.All gas analysis measurements are validated by the calculation of mass and energy balances around the reactor. For this purpose, a mathematical model of the reactor is created in the process simulation tool IPSEpro. In addition to the mass and energy balances, the model also calculates the fluid dynamic properties of the fluidized bed as well as the key performance indicators (KPI's) defined in Eqs. 6\u201312. All presented results in this paper reflect the balanced solution obtained from IPSEpro and not the direct measurement values.The methane yield (Y\n\nCH\n4\n) is calculated according to Eq. 6, where \n\nn\n\u0307\n\n is the molar flow of species i and N is the number of carbon atoms in the respective gas component in the feed gas (feed) and the raw-SNG (out). Eqs. 7, 8, and 9 define the CO conversion (X\n\nCO\n), the CO2 conversion (X\n\nCO\n2\n) and the H2 conversion (X\n\nH\n2\n), respectively.\n\n(6)\n\n\nY\n\nC\n\nH\n4\n\n\n\n=\n\n\n\n\n\nn\n\u0307\n\n\nC\n\nH\n4\n\n,\nout\n\n\n\n\n\n\u2211\ni\n\n\nN\ni\n\n\n\n\nn\n\u0307\n\n\ni\n,\nfeed\n\n\n\n\n\u2217\n100\n\n\n\n\n\n(7)\n\n\nX\nCO\n\n=\n\n\n\n\n\nn\n\u0307\n\n\nCO\n,\nfeed\n\n\n\u2212\n\n\nn\n\u0307\n\n\nCO\n,\nout\n\n\n\n\n\n\n\nn\n\u0307\n\n\nCO\n,\nfeed\n\n\n\n\n\u2217\n100\n\n\n\n\n\n(8)\n\n\nX\n\nC\n\nO\n2\n\n\n\n=\n\n\n\n\n\nn\n\u0307\n\n\nC\n\nO\n2\n\n,\nfeed\n\n\n\u2212\n\n\nn\n\u0307\n\n\nC\n\nO\n2\n\n,\nout\n\n\n\n\n\n\n\nn\n\u0307\n\n\nC\n\nO\n2\n\n,\nfee\nd\n\n\n\n\n\u2217\n100\n\n\n\n\n\n(9)\n\n\nX\n\nH\n2\n\n\n=\n\n\n\n\n\nn\n\u0307\n\n\n\nH\n2\n\n,\nfeed\n\n\n\u2212\n\n\nn\n\u0307\n\n\n\nH\n2\n\n,\nout\n\n\n\n\n\n\n\nn\n\u0307\n\n\n\nH\n2\n\n,\nfeed\n\n\n\n\n\u2217\n100\n\n\n\nEqs. 10 and 11 show the calculation of the CO2 selectivity (S\n\nCO\n2\n) and the C2H6 selectivity (S\n\nC\n2\nH\n6\n), respectively. Eq. 10 is based on the assumption that CO2 is only formed from CO via the water-gas shift reaction (Eq. 2). The C2H6 selectivity only considers the formation of C2H6 via the hydrogenation of C2H4 (Eq. 4). However, the catalyst also shows a slight selectivity of CO towards C2H6 under certain reaction conditions. Therefore, the C2H6 selectivity is only depicted if the formation via CO does not occur.\n\n(10)\n\n\nS\n\nC\n\nO\n2\n\n\n\n=\n\n\n\n\n\nn\n\u0307\n\n\n\nCO\n2\n\n,\nout\n\n\n\u2212\n\n\nn\n\u0307\n\n\n\nCO\n2\n\n,\nfeed\n\n\n\n\n\n\nn\n\u0307\n\n\nCO\n,\nfeed\n\n\n\u2212\n\n\nn\n\u0307\n\n\nCO\n,\nout\n\n\n\n\n\n\u2217\n100\n\n\n\n\n\n(11)\n\n\nS\n\n\nC\n2\n\n\nH\n6\n\n\n\n=\n\n\n\n\n\nn\n\u0307\n\n\n\nC\n2\n\n\nH\n6\n\n,\nout\n\n\n\n\n\n\nn\n\u0307\n\n\n\nC\n2\n\n\nH\n4\n\n,\nfeed\n\n\n\u2212\n\n\nn\n\u0307\n\n\n\nC\n2\n\n\nH\n4\n\n,\nout\n\n\n\n\n\n\u2217\n100\n\n\n\nThe stoichiometric number (SN) is calculated according to Eq. 12. It assesses the stoichiometry of the feed gas for methanation according to the reaction equations Eqs. 1, 2 and 5.\n\n(12)\n\nSN\n=\n\n\ny\n\nH\n2\n\n\n\n3\n\n\ny\nCO\n\n+\n4\n\n\ny\n\nC\n\nO\n2\n\n\n\n+\n2\n\n\ny\n\n\nC\n2\n\n\nH\n4\n\n\n\n\n\n\n\n\n\nTable 2\n gives an overview of the properties of the \u03b1-Al2O3 support and the prepared NiO/\u03b1-Al2O3 catalyst. From a fluid dynamic point of view, both can be classified as group B particles close to the transition area to group A, according to Geldart [41]. The Sauter diameter (dSV) and the bulk density (\u03c1b) increase through the impregnation of the support with NiO. Nevertheless, the uniform and narrow particle size distribution of the support is maintained. 99% of the particles are sized between 80 and 280\u00a0\u03bcm (cf. Fig. 4). Since the particles are also approximately spherical, they are deemed well suited for an optimal fluidization behavior [26]. Additionally, a very high BET surface area was measured for the support, despite literature reports, which attribute Al2O3 in the alpha configuration a rather low surface area due to the high calcination temperature. Liu et al. [42] for example, achieved a surface area of 44\u00a0m2/g, while other supports exhibit values around 10\u00a0m2/g [43,44]. By impregnating the support with nickel, the surface area is reduced by about 22%, most likely through the blockage of pores with NiO particles [43]. Nevertheless, the resulting catalyst shows a very high surface area at around 140\u00a0m2/g, which is even in the range of commonly used \u03b3-Al2O3-based catalysts [28,38]. The average Ni particle size at 37\u00a0nm is in the upper part of the spectrum but within the expected range. The somewhat larger Ni particles may result from a weaker catalyst/support interaction and the preparation conditions [43].\nFig. 3\n illustrates the H2 consumption during the TPR of the catalyst. Two main reduction temperatures at 595\u00a0\u00b0C and 766\u00a0\u00b0C were identified. The hydrogen uptake was 1.5\u00a0mmol/g and 0.34\u00a0mmol/g for the first and the second peak, respectively. The lower temperature can be assigned to a less strongly bounded NiO on Al2O3 and the NiO that is reduced at 766\u00a0\u00b0C to Ni-aluminate spinels. MgO is known to be responsible for a higher amount of NiO being present in the form that is easier to reduce. MgO was not likely to be reduced under these conditions [45]. Additionally, the Al2O3 in the alpha configuration leads to less strongly bound NiO, which lowers the reduction temperature [43]. The H2 consumption per Ni atom was approximately 0.7, indicating a core-shell structure of Ni, where the core was still oxidized and only the shell atoms were in metallic form after reduction.In order to further evaluate the performance of the catalyst, the mechanical and chemical stability was evaluated. Fig. 4\n (left) depicts the particle size distribution and the mean Sauter diameter (dSV) of the fresh and the used catalyst. No significant attrition of the catalyst was detected during approximately 200\u00a0h of operation under fluidized bed conditions. The narrow and uniform particle size distribution of the fresh catalyst could be maintained. Only a slightly smaller Sauter diameter was measured for the used catalyst. The deviation is, however, too small to state significant attrition of the catalyst. On the one hand the measurement accuracy is lower than the deviation and on the other hand the reduction of the catalyst is not accounted for since the fresh catalyst was in the original, oxidized state when the measurement was performed. The measurement accuracy is deemed suitable for the statement of no significant attrition, especially considering the relatively high number of operating hours. For a finite statement, even longer-term experiments could show if relevant attrition occurs. Furthermore, no chemical deactivation of the catalyst occurred during the methanation experiments over a period of approximately 100\u00a0h. The chemical stability was determined by repeatedly carrying out methanation experiments under the same process conditions and comparing the raw-SNG gas composition (see supplementary material chapter B). Nevertheless, some carbon deposition was found on the catalyst by performing TPO analysis of the used catalyst (Fig. 4 right). The TPO curve suggests that only very little amounts of carbon were deposited on the catalyst. Furthermore, the CO2 peak at around 350\u00a0\u00b0C suggests amorphous carbon, which is rather weakly bound [46]. In general, literature reports under fixed bed conditions indicate that larger Ni particles on \u03b1-Al2O3 are more stable and active over time than smaller particles on low-temperature calcined supports [43].In this section, the performance of the catalyst and the reactor is investigated by carrying out stoichiometric H2/CO methanation experiments. Fig. 5\n shows the raw-SNG composition for varying temperatures and WHSV's in comparison to the maximum thermodynamic values (dotted lines). There is a clear influence of both parameters visible. At high temperatures, the experimental values approach the thermodynamic equilibrium independent of the applied WHSV. However, a small deviation remains, which is attributed to the back-mixing behavior of fluidized beds. Additionally, the results also confirm the findings of Kopyscinski et al. [28], who describe that the mass transfer between the bubble phase and the dense emulsion phase is a limiting factor. Above the surface of the fluidized bed, less reacted gas of the bubble phase mixes with gas from the dense emulsion phase, which overall results in a below-maximum conversion. At lower temperatures, kinetic limitations take over and lead to a pronounced deviation from the thermodynamic equilibrium. Thus, also the WHSV has a greater influence on the gas composition. The maximum CH4 content of 76.5 vol.-%db is reached at a temperature of 320\u00a0\u00b0C and a WHSV of 0.8 Nl/g\u00a0h. At the same time, the H2 and CO contents are minimal at 18.9 vol.-%db and 400 ppmdb, respectively. Accordingly, the maxima and minima shift towards higher temperatures for higher WHSV, following the kinetic limitation. CO2 is produced via the water-gas shift reaction (Eq. 2), yielding around 5 to 8 vol.-%db. Additionally, the amount of CO2 is higher and the CO2 selectivity increases from around 6 to 8% with a higher WHSV. Both assertions indicate that the catalyst is very active towards the water-gas shift reaction (Eq. 2). Furthermore, the catalyst shows a slight selectivity of CO towards ethane below 330\u00a0\u00b0C. Up to 0.6 vol.-%db ethane were detected.During the experiments, the temperature distribution along the height of the catalytic bed is monitored to determine the isothermal operation capabilities of the reactor. Since the stoichiometric H2/CO methanation yields the highest specific heat amount compared to the other investigated gas compositions, the maximum temperature gradients also occur in this case. A maximum gradient of 10\u00a0\u00b0C was measured at 280\u00a0\u00b0C and 1 Nl/g\u00a0h. In general, the gradient is much lower. Especially at temperatures above 320\u00a0\u00b0C a deviation of only around 2\u00a0\u00b0C was measured. Thus, an isothermal operation with the applied reactor and catalyst combination was shown to be feasible and the thermal stress on the catalyst is kept at a minimum. An additional particle mixing by induced solid circulation due to different fluidization velocities in the inner and annular reaction zones was not needed.In this section, typical DFB and SER syngas compositions from the advanced 100\u00a0kWth DFB pilot plant at TU Wien are investigated in the fluidized bed methanation reactor. The utilized syngas compositions are defined in Table 1 and the figure headings. The exact syngas compositions for the SER methanation experiments are listed in the supplementary material (chapter C).\nFig. 6\n shows the raw-SNG composition for varying temperatures and WHSV's in comparison to the maximum thermodynamic values (dotted lines). Because of the understoichiometric composition of the DFB syngas (SN\u00a0=\u00a00.26), 20 vol.-% water vapor was added to the syngas. Both the added water vapor and the water produced through the methanation reaction (Eq. 1) lead to a shift of the gas (Eq. 2) and the production of CO2. Together with the CO2 already present in the syngas, it is the component with the highest concentration in the raw-SNG. A maximum CH4 concentration of 43.4 vol.-%db and a minimum H2 and CO concentration of 8.8 vol.-%db and 0.32 vol.-%db, respectively, was measured at a temperature of 320\u00a0\u00b0C and a WHSV of 1 Nl/g\u00a0h. Similar to Fig. 5, there is an influence of temperature and WHSV visible on the gas composition. At higher temperatures and lower WHSV's, the thermodynamic equilibrium is approached. However, the influence of both parameters is less pronounced compared to the stoichiometric CO methanation experiments. Between 300 and 400\u00a0\u00b0C, the CH4 concentration only varies by 2.7 vol.-%db at 1.5 Nl/g\u00a0h. Interestingly, the CO2 and CO concentrations at high temperatures are closer to the thermodynamic equilibrium than the CH4 and H2 concentrations. Again, this could indicate that the water-gas shift reaction is favored over the methanation reaction, as Kopyscinski et al. already discussed in [24,28] for a different nickel catalyst. However, at low temperatures, Kopyscinski et al. [47] state that the water-gas shift reaction is negligible. This does not seem to be the case for the investigated catalyst since the CO2 concentrations are especially high at low temperatures. Ethylene was also added to the feed gas and is fully converted to ethane and other components like methane. The concentration of ethane strongly depends on the temperature and, to a lesser extent, on the WHSV.\nFig. 7\n visualizes the C2H6 selectivity (S\n\nC\n2\nH\n6\n) and the CO2 selectivity (S\n\nCO\n2\n). At temperatures around 400\u00a0\u00b0C, almost no ethane is formed, whereas at a temperature of 280\u00a0\u00b0C the C2H6 selectivity is as high as 65%. Additionally, a lower WHSV leads to a lower selectivity. This shows that the reaction of ethane to methane is kinetically inhibited at lower temperatures, whereas the conversion of ethylene is complete for all applied conditions. On the other hand, the concentration of ethylene in the syngas does not seem to have any influence on the C2H6 selectivity. At 360\u00a0\u00b0C, the same selectivities were achieved for all investigated concentrations. The course of the C2H6 selectivity over temperature agrees very well with the results published by Kopyscinski et al. [24], who found similar C2H6 selectivities even though they used a catalyst with 50 wt.-% NiO and a syngas composition with higher CO and C2H4 contents. However, above 350\u00a0\u00b0C, their C2H6 selectivity approaches zero, whereas Fig. 7 still shows some selectivity.Furthermore, the measurement data shows that no other hydrocarbons are formed under the applied reaction conditions. Reference experiments without ethylene in the syngas also show that CO is not selective towards C2H6 over the whole temperature range for a typical DFB syngas (see supplementary material chapter D). This is in contrast to the stoichiometric H2/CO methanation experiments in Fig. 5 where CO shows some selectivity towards C2H6. Both the added steam and the understoichiometric composition of the syngas can be responsible for the suppression of ethane formation from CO. In general, the formation of ethane to a certain amount is desirable because of the higher energy density and increased heating value of the resulting SNG. This, in turn, increases the leeway for fulfilling the specifications of the gas grid, since the minimum required heating value can be reached more easily. However, ethane and other hydrocarbons usually present in the DFB syngas can lead to coke formation. Even though this issue is less pronounced in fluidized beds, it can still lead to a coverage of the catalyst surface with carbon species and deactivation of the catalyst [48,49]. Carbon deposition is, however, strongly influenced by the reaction temperature and low temperatures (<380\u00a0\u00b0C) were found to suppress carbon formation to a large extent [24]. Within this work, carbon deposition was detected after an operation period of 100\u00a0h under different operating conditions. However, the deposited carbon did not lead to a deactivation of the catalyst (also see section 3.1.1). A more detailed investigation under specified conditions would be required to allow a final statement concerning this topic.A look at the CO2 selectivity shows that it is relatively constant over the investigated temperature range, which is again in agreement with [24]. For higher temperatures only a minor increase can be observed. A lower WHSV decreases the CO2 selectivity as is the case for the stoichiometric H2/CO methanation in section 3.1.2. This is explained by the fact that the methanation reactions gain importance at lower WHSV and lead to an increased CO conversion to CH4. The higher CO2 concentrations for lower WHSV in Fig. 6 are therefore attributed to dilution effects.The feed water content was set to 20 vol.-% to prevent carbon depositions and was chosen according to thermodynamic consideration as well as literature values. However, the amount of water also shows an influence on the raw-SNG composition since it is a reaction product of the methanation reactions but an educt of the water-gas shift reaction. Accordingly, the CH4 and CO concentrations decrease with increasing water content, while the H2 concentration and the CO2 selectivity increase as experiments (see supplementary material chapter D) and Kopyscinski et al. [24] confirm. Additionally, higher water concentrations can lead to a hydrothermal deactivation of the catalyst, which involves grain growth of the catalytic phase, especially at higher temperatures [22]. Therefore, an optimization of the feed water content depends on the reaction conditions, the syngas composition, and the catalyst and is a tradeoff between the raw-SNG composition and catalyst deactivation. In this work, no further investigations on the hydrothermal deactivation were carried out.Syngas from the SER process is very flexible in its composition, which can be taken advantage of in methanation processes. Fig. 8\n depicts the raw-SNG composition over the stoichiometric number SN (Eq. 12) of the syngas at a methanation temperature of 360\u00a0\u00b0C and a WHSV of 1.5 Nl/g\u00a0h. Additionally, the thermodynamic values are given (dotted lines). The syngas compositions range from widely understoichiometric to overstoichiometric compositions (SN\u00a0=\u00a00.4\u20131.6). Accordingly, also the raw-SNG composition changes with SN. In general, the experimentally determined values follow the course of the thermodynamic prediction well. For a high SN, the distance to the thermodynamic equilibrium decreases further, due to the diminishing amounts of CO and CO2, which need to be methanated, despite the constant WHSV. Furthermore, the CO2 concentration is very close to the thermodynamic equilibrium, which again indicates a preference of the catalyst towards the water-gas shift reaction. The highest methane concentrations are found for an almost stoichiometric composition containing 71 vol.-%db CH4. A lower SN leads to an excess of CO2 and higher CO concentrations but to a lower H2 content. In this case, not enough H2 is available to convert the CO2 in the syngas. On the other hand, a higher SN and therefore a higher H2 partial pressure in the syngas allows an almost complete conversion of CO2 and CO. No CO2 separation unit for the upgrading of the raw-SNG is necessary in this case. This is true for a SN \u2265 1.2 under the considered reaction conditions. However, excessive amounts of hydrogen are left in the raw-SNG, which need to be separated or recirculated before grid feeding. Hydrogen separation is necessary for all investigated compositions, since the limit of 10 vol.-%db according to the regulations [50] was not reached. Ethylene is again fully converted, but a certain amount remains as ethane in the raw-SNG.\nFig. 9\n depicts the key figures Y\n\nCH\n4\n, X\n\nCO\n, X\n\nCO\n2\n, X\n\nH\n2\n, and S\n\nC\n2\nH\n6\n over SN for the methanation of the SER syngas. Y\n\nCH\n4\n, X\n\nCO\n, and X\n\nCO\n2\n increase with a higher SN and reach almost 100%. The H2 conversion, on the other hand, decreases from left to right. Especially when looking at the CO2 conversion, the effect of the different syngas compositions becomes evident. While the CO2 conversion is almost zero on the left side of the diagram, it is nearly complete for the highest depicted SN. Overall, the performance improvement is most pronounced on the left side of the diagram, i.e., when increasing SN from 0.4 to 1. Higher SN lead to a lower increase of the key figures at the expense of a more pronounced decline in H2 conversion. In other words, the driving force for the reaction decreases. Interestingly, the ethane selectivity is relatively constant over the whole SN range even though the syngas composition varies considerably. Only a slight decrease is observed towards higher SN.This chapter compares the raw-SNG composition (Fig. 10\n left) and the key figures (Fig. 10 right) of the DFB and SER syngas methanation experiments. The two displayed datasets were recorded under the same reaction conditions at 360\u00a0\u00b0C and a WHSV of 1 Nl/g\u00a0h with a SN of 1.05 for the SER syngas. The CH4 content increases by more than 30 percentage points by utilizing the SER syngas, while the CO and CO2 concentrations decrease to low levels. However, the residual H2 content more than doubles. This is not because of a lower H2 conversion but due to the dilution of the DFB raw-SNG with CO2. After CO2 separation, the residual H2 concentrations are in a similar range. Nevertheless, both raw-SNG gases require an H2 and CO2 separation unit before grid feeding under the considered reaction conditions. On the contrary, the residual CO concentration of the SER raw-SNG is within the limit of 0.1\u00a0mol.-% according to the regulations [50], whereas a further reduction for the DFB raw-SNG is required. In general, the SER raw-SNG is much closer to the specifications of the gas grid and an injection to the gas grid is possible without CO or CO2 separation, as Fig. 8 shows.Thermodynamic investigations predict that grid feeding even without H2, CO and CO2 separation is theoretically possible under the right process conditions. This includes a further reduction of reaction temperature and a pressurized application. For the depicted, slightly overstoichiometric composition, this could be thermodynamically achieved at around 300\u00a0\u00b0C and 4 bara [13]. Alternatively, a two-stage methanation process could be applied, as some theoretical considerations show [51].\nFig. 10 (right) displays the key figures defined in Eqs. 6\u20139 and 11. The SER syngas methanation allows a doubling of the CH4 yield and a substantial amount of CO2 conversion. Simultaneously, the H2 and CO conversions are slightly increased as well. The DFB syngas, on the other hand, leads to the production of additional CO2 through the shift of the gas via the water-gas shift reaction. However, the ethane selectivity is lower in the case of the SER syngas. This is attributed to the high H2 partial pressure of the SER syngas, which influences the selectivity to some extent (cf. Fig. 9).Overall, it is necessary to look at the performance of the whole process chain and not only at the methanation itself. The methanation KPI's are much more favorable for the SER syngas. However, this should not create the illusion that the performance of the whole process from biomass to SNG is more favorable as well. The high methane yield for the SER syngas methanation is only possible because of the different process characteristics of the SER and the DFB operation modes of the gasifier. Therefore, a comparison of the whole process chain is necessary.This study shows experimental results of syngas methanation in a fluidized bed methanation reactor. The main focus points are the development and testing of a stable catalyst for fluidized beds, the methanation reactor design and the detailed investigation of the fluidized bed methanation process characteristics through parameter variations. Furthermore, optimized process concepts are investigated through the methanation of flexible syngas compositions from the advanced dual fluidized bed technology.The following results can be summarized:\n\n(i)\nThe synthesized catalyst performs well in terms of avoidance of mechanical attrition and chemical deactivation and, therefore, maintaining a proper fluidization behavior. No significant mechanical attrition and chemical deactivation of the catalyst was detected during 200\u00a0h under fluidized bed conditions and 100\u00a0h under methanation conditions. Simultaneously, the catalyst showed a sufficient catalytic activity and selectivity towards the methanation reactions. The results were achieved by utilizing a highly porous and inert \u03b1-Al2O3 as support.\n\n\n(ii)\nThe reactor design, in combination with the dilution of the catalyst with support material, allowed an isothermal operation of the process with temperature gradients as low as 2\u00a0\u00b0C. Between 280\u00a0\u00b0C and 420\u00a0\u00b0C, a clear transition between thermodynamic and kinetic limitations could be observed. The optimal tradeoff between these limitations was found to be in a temperature range between 320 and 360\u00a0\u00b0C.\n\n\n(iii)\nA maximum of 43 vol.-%db CH4 was reached through the methanation of a typical syngas composition from the advanced 100\u00a0kW DFB gasification pilot plant. Because of the understoichiometric composition of the syngas from the DFB process, CO2 was produced through the water-gas shift reaction and constituted the main component in the raw-SNG.\n\n\n(iv)\nUnder all conditions, a full conversion of ethylene to ethane and other components was shown. At low temperatures, kinetic limitations favored the production of ethane while higher temperatures allowed a complete conversion to other components. Neither the ethylene concentration in the syngas nor the syngas composition (DFB or SER) showed a significant influence on the selectivity of ethylene towards ethane.\n\n\n(v)\nThe SER syngas methanation was shown to yield a much more favorable composition for grid feeding and higher methane yields, while simultaneously improving the H2, CO and CO2 conversions, compared to DFB syngas methanation. A maximum CH4 content of 73 vol.-%db could be reached, which represents an increase of 30 vol.-%db compared to DFB syngas methanation.\n\n\n(vi)\nAn almost complete conversion of CO and CO2 was achieved through the methanation of an overstoichiometric SER syngas (SN\u00a0\u2265\u00a01.2), allowing grid feeding without the need for an expensive CO2 separation unit. Hydrogen separation and recirculation is, however, necessary under the investigated reaction conditions.\n\n\nThe synthesized catalyst performs well in terms of avoidance of mechanical attrition and chemical deactivation and, therefore, maintaining a proper fluidization behavior. No significant mechanical attrition and chemical deactivation of the catalyst was detected during 200\u00a0h under fluidized bed conditions and 100\u00a0h under methanation conditions. Simultaneously, the catalyst showed a sufficient catalytic activity and selectivity towards the methanation reactions. The results were achieved by utilizing a highly porous and inert \u03b1-Al2O3 as support.The reactor design, in combination with the dilution of the catalyst with support material, allowed an isothermal operation of the process with temperature gradients as low as 2\u00a0\u00b0C. Between 280\u00a0\u00b0C and 420\u00a0\u00b0C, a clear transition between thermodynamic and kinetic limitations could be observed. The optimal tradeoff between these limitations was found to be in a temperature range between 320 and 360\u00a0\u00b0C.A maximum of 43 vol.-%db CH4 was reached through the methanation of a typical syngas composition from the advanced 100\u00a0kW DFB gasification pilot plant. Because of the understoichiometric composition of the syngas from the DFB process, CO2 was produced through the water-gas shift reaction and constituted the main component in the raw-SNG.Under all conditions, a full conversion of ethylene to ethane and other components was shown. At low temperatures, kinetic limitations favored the production of ethane while higher temperatures allowed a complete conversion to other components. Neither the ethylene concentration in the syngas nor the syngas composition (DFB or SER) showed a significant influence on the selectivity of ethylene towards ethane.The SER syngas methanation was shown to yield a much more favorable composition for grid feeding and higher methane yields, while simultaneously improving the H2, CO and CO2 conversions, compared to DFB syngas methanation. A maximum CH4 content of 73 vol.-%db could be reached, which represents an increase of 30 vol.-%db compared to DFB syngas methanation.An almost complete conversion of CO and CO2 was achieved through the methanation of an overstoichiometric SER syngas (SN\u00a0\u2265\u00a01.2), allowing grid feeding without the need for an expensive CO2 separation unit. Hydrogen separation and recirculation is, however, necessary under the investigated reaction conditions.In this work, the catalytic methanation of syngas from the advanced DFB technology was experimentally investigated, combining a bubbling fluidized bed methanation reactor with an optimized methanation catalyst for fluidized bed applications. Fluidized bed reactors can be advantageously applied to SNG production because of the high heat and mass transfer capabilities. This leads to several advantages compared to the commercially utilized fixed bed methanation reactors, provided that the catalyst shows a proper fluidization behavior.To this end, the following conclusions and recommendations can be drawn from the conducted experiments:\n\n(i)\n\u03b1-Al2O3 was shown to be a viable catalyst support for methanation reactions. Especially in fluidized bed applications, it could be advantageously used as an alternative to the commonly utilized \u03b3-Al2O3 because of the high mechanical and chemical stability and a proper fluidization behavior of the prepared catalyst.\n\n\n(ii)\nThe stress on the catalyst was minimized due to nearly isothermal operating conditions. This is a result of the special reactor design and the dilution of the catalyst with inert support material.\n\n\n\u03b1-Al2O3 was shown to be a viable catalyst support for methanation reactions. Especially in fluidized bed applications, it could be advantageously used as an alternative to the commonly utilized \u03b3-Al2O3 because of the high mechanical and chemical stability and a proper fluidization behavior of the prepared catalyst.The stress on the catalyst was minimized due to nearly isothermal operating conditions. This is a result of the special reactor design and the dilution of the catalyst with inert support material.Additionally, the fluidized bed methanation reactor and the synthesized catalyst were applied to the methanation of syngas from advanced DFB gasification and the SER process. The SER process allows the production of syngas with a suitable stoichiometric ratio of H2 to CO and CO2, yielding the following conclusions for fluidized bed methanation:\n\n(iii)\nThe SER process in combination with fluidized bed methanation could lead to an improved and more cost-effective route for SNG production. No separation of excessive CO2 or CO from the raw-SNG is required for grid-feeding when selecting a suitable syngas composition. Compared to conventional DFB steam gasification, the methane yield is doubled (up to 95%) and the H2, CO and CO2 conversions are improved. Especially if no external hydrogen is available, the direct methanation of SER syngas could lead to a simpler and more efficient process route for SNG production.\n\n\nThe SER process in combination with fluidized bed methanation could lead to an improved and more cost-effective route for SNG production. No separation of excessive CO2 or CO from the raw-SNG is required for grid-feeding when selecting a suitable syngas composition. Compared to conventional DFB steam gasification, the methane yield is doubled (up to 95%) and the H2, CO and CO2 conversions are improved. Especially if no external hydrogen is available, the direct methanation of SER syngas could lead to a simpler and more efficient process route for SNG production.For a full comparison of the DFB and SER syngas methanation, further investigations on the optimal process conditions and the performance of the whole process chain are necessary. Additionally, the long-term mechanical and chemical stability of the catalyst should be examined.\nAlexander Bartik: Conceptualization, Methodology, Investigation, Writing \u2013 review & editing, Visualization. Josef Fuchs: Methodology, Validation, Writing \u2013 review & editing. Gernot Pacholik: Investigation, Writing \u2013 review & editing. Karin F\u00f6ttinger: Validation, Writing \u2013 review & editing, Supervision. Hermann Hofbauer: Conceptualization, Writing \u2013 review & editing, Supervision, Project administration, Funding acquisition. Stefan M\u00fcller: Validation, Writing \u2013 review & editing, Supervision, Funding acquisition. Florian Benedikt: Conceptualization, Methodology, Writing \u2013 review & editing, Supervision, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This study was carried out within the doctoral college CO2Refinery at TU Wien. It is also part of the research project ReGas4Industry (871732) and receives financial support from the research program \u201cEnergieforschung\u201d funded by the Austrian Climate and Energy Fund. Furthermore, the authors would like to thank Jonas Hauser for the technical assistance, Wolfgang Ipsmiller for the assistance and rental of the particle size analyzer, Johannes Schmid for the support during the conceptualization of the reactor and SASOL for the supply of the catalyst support. The authors acknowledge TU Wien Bibliothek for financial support through its Open Access Funding Programme.\n\n\n\nSupplementary material.\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.fuproc.2022.107402.", "descript": "\n                  Catalytic methanation processes allow the production of natural gas substitutes on a sustainable and renewable basis. This study investigates the catalytic methanation of syngas from dual fluidized bed steam gasification of biomass in an innovative bubbling fluidized bed methanation reactor with an optimized catalyst. Syngas from conventional gasification and a novel combination with syngas from sorption enhanced reforming were investigated. The applied fluidized bed reactor allowed an almost isothermal operation with optimal reaction temperatures between 320\u00a0\u00b0C\u2013360\u00a0\u00b0C. Simultaneously, no chemical deactivation or mechanical attrition during 200\u00a0h of operation indicates a high long-term stability of the catalyst. The methane concentration downstream the methanation reactor increased from 43 to 74 vol.-%db through the methanation of a hydrogen-rich syngas produced via sorption enhanced reforming. Simultaneously, the methane yield is doubled to 95% and the hydrogen, carbon monoxide and carbon dioxide conversions are improved. Furthermore, it could be shown that a CO2 content below 1 vol.-%db is feasible in the (raw) synthetic natural gas, allowing grid injection without CO2 separation. The results indicate that sorption enhanced reforming in combination with an optimized fluidized bed methanation can lead to technical and economic improvements in sustainable synthetic natural gas production.\n               "}
{"full_text": "In the last decades, hydrogen fuel cell technology has become an alternative to energy production systems in fixed and mobile devices [1]. The polymer-electrolyte-membranes fuel cells (PEMFCs) has turned into the most hopeful energy converters due to their high efficiency, no pollution, and energy suitability. Nowadays, H2 is produced mainly by hydrocarbons steam reforming, which inherently includes a large CO concentration (0.5%\u201310%) in the H2 stream [2]. However, high purity hydrogen is necessary since Pt anodes are very sensitive to impurities such as carbon monoxide and sulfides [2,3]. The CO presents a poisoning effect since it is adsorbed much easier than hydrogen on the surface of the Pt anode. Furthermore, CO is difficult to be eliminated from the active sites [4,5].Different strategies such as pressure swing adsorption (PSA), membrane separation, water-gas shift (WGS) reaction, selective methanation, and preferential oxidation have been proposed to minimize and remove the CO concentrations in H2 streams [2]. Among them, selective methanation of carbon monoxide (S-MET) where CO reacts preferentially to form CH4 without CO2 conversion, has been reported as a promissory strategy in hydrogen purification (Eqs. (1) and (2)) [6,7]:\n\n(1)\n\n\nC\nO\n+\n3\n\nH\n2\n\n\u2194\nC\n\nH\n4\n\n+\n\nH\n2\n\nO\n\n\u0394\n\nH\n0\n\n=\n\u2212\n206\n\nkJ\n\n\nmol\n\n-\n1\n\n\n\n\n\n\n\n\n(2)\n\n\nC\n\nO\n2\n\n+\n4\n\nH\n2\n\n\u2194\nC\n\nH\n4\n\n+\n2\n\nH\n2\n\n\n\u0394\n\nH\n0\n\n=\n\u2212\n165\n\nkJ\n\n\nmol\n\n-\n1\n\n\n\n\n\n\nIn this sense, most of the catalysts studied in this reaction can be classified as Ni- or Ru-based catalysts [8,9]. The role of the support materials and the promoters in nickel (Ni) and ruthenium (Ru) catalysts have proved to play an essential part in the enhancement of the selectivity of CO due to their direct relationship to the electron density that inherently influences the activity of CO methanation [10\u201313]. Due to this, if the electron density is enhanced, CO adsorbed on the metal surface is more easily dissociated by enhanced d\n\u03c0\n-p\n\u03c0\n\u2217 back bonding and by consequence, CO methanation is improved [14,15]. In addition, it has been reported that in Ru-based catalyst, the CO methanation activity has been improved using supports such as TiO2 compared to others such as Al2O3, CeO2, YSZ, SiO2, ZrO2, and MgO due to the synergy between the support and the electron density of the active phase [16,17]. Indeed, Tada and Kikuchi described a mechanistic study over Ru/TiO2 catalyst for selective carbon monoxide methanation where the control of the interfaces between active metals and support materials was observed to be a key step. Due to this, the choice of suitable support materials such as TiO2 and promoters are necessary to improve CO methanation [11].The addition of small amounts of metals in Ru-based catalysts has been reported as a strategy to enhance the selectivity in the hydrogenation of C\u2013O towards C\u2013C bonding through the combination of electropositive metals giving place to particles with improved redox properties [18]. This improvement can be explained in two pathways: (i) the most electropositive metal acts as a Lewis base that increases the density in the other one, therefore decreasing the bond energy of the adsorbed species, in particular, the C\u2013C bond, and favouring the C\u2013O hydrogenation towards the C\u2013C one [19]; (ii) the metal active phase act as electrophilic or Lewis acid sites for CO adsorption and activation of C\u2013O bond through the oxygen lone pair of electrons.Among the different metal promoters, rhodium (Rh) has proved to enhance catalytic activity in different reactions due to its ability to form solid solutions in a Ru matrix [20\u201322]. Additionally, Rh possess one extra electron in its electronic configuration, which can increase the electronic density to the active sites, therefore making the C\u2013O bonding breakage becomes easier due to the back-bonding effect in the adsorbed metal. In a similar route, Platinum (Pt) is also able to form alloys with Ru, adopting an hcp structure at high Pt concentrations (>80%), that is, when forming solid solution Pt-Ru [23]. Pt possesses even more electronic density than Rh, which facilitates the breakage of adsorbed CO in comparison with Rh. Both Rh and Pt have intrinsic properties as methanation catalysts [24].According to these premises, in this work we synthesize a series of active Ru/TiO2 catalysts promoted with Rh and Pt to evaluate the effect of small amounts of these metals in the catalytic performance in the S-MET reaction using a real composition H2 stream to simulate the streams outlines from the WGS and PROX units.Ru/TiO2 parent catalyst was prepared by wet impregnation method based on a similar procedure as reported elsewhere [25]. Typically, 13.2\u00a0g of nitrosyl nitrate solution (14.391\u00a0wt% of Ru, Johnson Matthey) were mixed in the necessary amount of water to impregnate 19\u00a0g of Aeroxide\u00ae TiO2 P25 (Evonik) support to obtain a nominal loading of 9.5\u00a0wt% of Ru. The solid catalyst was dried at 130\u00a0\u00b0C for 24\u00a0h and finally calcined at 400\u00a0\u00b0C for 2\u00a0h.Analogously, the bimetallic catalysts were prepared by co-impregnation of both ruthenium and platinum or rhodium precursors in order to achieve a nominal loading value of 9.5\u00a0wt% of Ru and 0.5\u00a0wt% of Pt or Rh. Rh(NO3)3 from Alfa Aesar and Pt(NH3)4(NO3)2 from Johnson Matthey were employed as metallic precursors. The fresh Rh-based catalyst was then calcined at 400\u00a0\u00b0C for 2\u00a0h while Pt-based catalyst was calcined at 350\u00a0\u00b0C for 8\u00a0h based on previous results of calcination conditions reported in the literature [26,27]. For clarity, the catalysts were designated as Ru\u2013TiO2, RhRu\u2013TiO2 and PtRu\u2013TiO2.The structural analysis of the synthesized catalysts was elucidated by powder X-ray diffraction on an X-Pert Pro PANalytical (Malvern PANalytical Ltd, Malvern, UK). X-ray diffraction patterns were collected using Cu K\u03b1 as a radiation source in the range of 2\u03b8\u00a0=\u00a010\u00b0\u201380\u00b0 with a step size of 0.05\u00b0 and a step time of 80\u00a0s.The quantitative analysis of the metal loading was performed using X-Ray Fluorescence in an Axios low-power (1\u00a0kW) wavelength dispersive XRF (WDXRF) spectrometer equipped with a Rh cathode.The textural properties of the catalyst were evaluated using N2 adsorption isotherms at 77\u00a0K in a Micromeritics ASAP 2010 instrument. Previous to analysis, the samples were outgassed under dynamic vacuum at 250\u00a0\u00b0C for 2\u00a0h to eliminate adsorbed molecules and impurities. The specific surface area of each solid was determined according to using the BET method. Pore volumes were determined by BJH desorption method.Hydrogen temperature-programmed reduction (H2-TPR) experiments were carried out placing the catalyst (0.1\u00a0g) in a U-shape quartz tube reactor and initially pretreated at 300\u00a0\u00b0C (rate 10\u00a0\u00b0C min\u22121) in Ar atmosphere for 1\u00a0h to remove adsorbed water and then cooled to 75\u00a0\u00b0C in Ar followed by heating to 700\u00a0\u00b0C at a heating rate of 10\u00a0\u00b0C min\u22121 in a flow of 10% H2/Ar. A thermal conductivity detector (TCD) was used to quantify the H2 consumption. By using the same experimental set-up, the metallic particles average size distribution was determined through hydrogen chemisorption (H2-TPD) experiments. Prior to the analysis, the samples were heated up to 300\u00a0\u00b0C under Ar flow for 45\u00a0min and then cooled down to 75\u00a0\u00b0C. Finally, the adsorption of H2 took place and then Ar was introduced to perform the TPD measurements with a temperature ramp of 10\u00a0\u00b0C min\u22121 to 300\u00a0\u00b0C. The dispersion of Ru (D) was calculated based on the volume of chemisorbed H2 using the following simplified equation (Eq. (3)):\n\n(3)\n\n\nD\n\n(\n%\n)\n\n=\n100\n\n(\n\n\n\nV\n\nS\nT\nP\n\n\n\u00b7\n\nS\ni\n\n\n22414\n\n)\n\n\n\n(\n\n\nF\ni\n\n\nP\n\nM\ni\n\n\n\n)\n\n\n\u2212\n1\n\n\n\n\n\nwhere V\n\nSTP\n denotes the total volume of H2 consumed (mL g\u22121), S\n\ni\n is the stochiometric factor of H2 to Ru, which is considered S\n\ni\n\u00a0=\u00a02 as previously reported in the literature [28,29], PM\n\ni\n is the molecular weight of the active phase (mmol g\u22121) and F\n\ni\n corresponds to the fraction of active phase per gram of sample. Finally, the crystallite size was estimated using the equation (Eq. (4)):\n\n(4)\n\n\n\nd\ni\n\n=\n\n\n6\n\u00b7\n\n\nV\ni\n\n\n\nD\n\u00b7\n\n\na\ni\n\n\n\n\n\n\nassuming the geometry of Ru particles as hemispherical, where V\n\ni\n is the average volume of a metallic bulk particle and a\n\ni\n is the exposed area of a metallic atom in the surface. For the calculation, a\n\ni\n for ruthenium particles was considered as a\n\nRu\n\u00a0=\u00a09.09\u00a0\u00d7\u00a010\u22122\u00a0nm3 instead of 6.29\u00a0\u00d7\u00a010\u22122\u00a0nm3 based on the previous work reported by Shen et\u00a0al. that considered that metallic Ru is able to expose indistinctly (100), (001) and (110) planes due to their similar superficial energy [28].X-ray Photoelectron Spectroscopy (XPS) analysis was carried out using a LEYBOLD-HEREUS model LHS-10/20 device equipped with Al-K\u03b1 radiation (1486.6\u00a0eV) and a twin crystal monochromator to produce a focused X-ray spot at 30\u00a0mA\u00a0\u00d7\u00a011\u00a0kV (400\u00a0\u03bcm major axis length of the elliptical shape). The alpha hemispherical analyzer was operated at the constant energy mode with survey scan pass energies of 200\u00a0eV to measure the whole energy band and 50\u00a0eV in a narrow scan to selectively measure specific elements. The reference binding energy was the C 1s core level at 284.6\u00a0eV.The catalytic activity was measured in a continuous flow fixed-bed stainless steel reactor (i.d. 9\u00a0mm) coupled to a Microactivity Reference Unit (PID Eng&Tech\u00ae). For each experiment, 150\u00a0mg of sieved catalyst (100\u2013200\u00a0\u03bcm) was mixed with commercial SiC (125\u00a0\u03bcm \u2013 VWR Prolabo\u00ae) up to a volume bed of 0.32\u00a0cm3. Prior to the reaction, the catalyst was activated in pure H2 flow of 60\u00a0mL\u00a0min\u22121 at 300\u00a0\u00b0C for 1\u00a0h and subsequent cooled up to reaction temperature. The reaction was conducted at atmospheric pressure increasing the temperature from 180\u00a0\u00b0C to 300\u00a0\u00b0C with heating rate of 10\u00a0\u00b0C min\u22121. A simulated mixture of real reformate stream containing H2 (50%), CO2 (15%), H2O (15%) and CO (1% or 300\u00a0ppm) balanced with N2 was fed at flow rate of 200\u00a0mL\u00a0min\u22121 (WHSV\u00a0=\u00a080\u00a0L\u00a0g\u22121\u00a0h\u22121). The effluents were on-line analyzed in a gas micro-chromatograph Varian 4900. The CO2 amount at the outlet was determined by a CO2 detector Vaisala CARBOCAP GMT220. The conversion (X\n\ni\n) and selectivity (S\n\ni\n) values were estimated according to the following equations (Eq. (5) to Eq. (7)):\n\n(5)\n\n\n\nX\n\nC\nO\n\n\n\n(\n%\n)\n\n=\n\n\n(\n\n\nF\n\nC\nO\n\ni\nn\n\n\n\u2212\n\nF\n\nC\nO\n\no\nu\nt\n\n\n\n)\n\n\nF\n\nC\nO\n\ni\nn\n\n\n\n\u2217\n100\n\n\n\n\n\n\n(6)\n\n\n\nX\n\nC\n\nO\n2\n\n\n\n\n(\n%\n)\n\n=\n\n\n(\n\n\nF\n\nC\n\nO\n\n2\n\ni\nn\n\n\n\n\n\u2212\n\nF\n\nC\n\nO\n\n2\n\no\nu\nt\n\n\n\n\n\n)\n\n\nF\n\nC\n\nO\n\n2\n\ni\nn\n\n\n\n\n\n\u2217\n100\n\n\n\n\n\n\n(7)\n\n\n\nS\n\nC\nO\n\nm\ne\nt\nh\na\nn\na\nt\ni\no\nn\n\n\n\n(\n%\n)\n\n=\n\n(\n\n\n\nX\n\nC\nO\n\n\n\u00b7\n\n\nF\n\nC\nO\ni\nn\n\n\n\n\nF\n\nC\n\nH\n4\n\n\no\nu\nt\n\n\n\n)\n\n\u00b7\n100\n\n\n\nbeing \n\n\n\nF\n\nC\nO\n\n\n\n, \n\n\nF\n\nC\n\nO\n\n2\n\n\n\n\n\n\nand\n\n\nF\n\nC\n\nH\n4\n\n\n\n\n the flow in mL min\u22121 of CO, CO2 and CH4, respectively, at the inlet (in) or the outlet (out) flow. \n\n\nC\n\ni\n,\n\no\nu\nt\n\n\n\n corresponds to the concentration of product i in the outlet and \n\n\n\u03bd\ni\n\n\n is the carbon number according to its chemical formula.In all cases, the methane selectivity estimated from Eq. (8) was higher than 95% and only small traces of ethane and ethylene were detected. The carbon balance resulted to be better than 97%.\n\n(8)\n\n\n\nS\n\nC\n\nH\n4\n\n\n\n\n(\n%\n)\n\n=\n\n(\n\n\n\nC\n\nC\n\nH\n4\n\n\no\nu\nt\n\n\n\n/\n\n\u03bd\n\nC\n\nH\n4\n\n\n\n\n\n\n\n\n\u2211\n\ni\n\n\nC\n\ni\n\no\nu\nt\n\n\n/\n\n\u03bd\n\ni\n\n\n\n\n\n)\n\n\u00b7\n100\n\n\n\n\n\n\n\nC\n\ni\n\no\nu\nt\n\n\n\n is the product i concentration in the outlet and \n\n\n\u03bd\n\ni\n\n\n\n\n is the carbon numbers according to its chemical formula.\nTable 1\n includes the wt.% metal loading values obtained by XRF analysis. As can be observed, the measured metal contents were close to the nominal values confirming that the synthesis method was successful. However, it is noticeable that Ru loading was slightly lower than that of expected one in the bimetallic catalysts. Although these differences in Ru content may seem significant, this variation can be assumed with the uncertainty of the measurement, where variations between 0.5 and 1\u00a0wt% have been reported [30].\nFig.\u00a01\na shows the XRD patterns of the as-synthesized catalysts in which are observed the typical diffraction lines of anatase (JCPDS 73-1764) and rutile (JCPDS 78-1510) phases present in the P25 titania support. Additionally, diffraction peaks ascribed to RuO2 (JCPDS 21-1172) are observed in all Ru-loaded patterns. However, after the incorporation of Rh and Pt in the Ru-loaded catalysts, no additional peaks were observed in the XRD patterns possibly due to the small amount of metal added to the catalyst but also possibly due to the high dispersion of the metal. Similarly, the XRD diffraction measurements were performed in the catalyst after a reduction process in H2 atmosphere to evaluate the structural changes taking place due to the reduction of the metals. Fig.\u00a01b presents the XRD patterns of the reduced samples where the diffraction peaks related to the metallic Ru reflection planes (100), (002) and (101) are observed close to ca. 2\u03b8\u00a0=\u00a041.4\u00b0, 43.7\u00b0 and 44.1\u00b0, which confirms the complete reduction of RuO2 at the evaluation conditions. However, the diffraction peaks related to the dopant metals are not present in the patterns mostly attributed to the high dispersion of those species. The Ru metal average crystal size of the three reduced catalysts was calculated using the Debye-Scherrer method. As summarized in Table 1, it is observed a slight increase in the crystal size in Ru-based catalyst after the incorporation of the Rh and Pt promoters which may be attributed to the interactions Ru\u2013Pt and Ru\u2013Rh able to form bimetallic alloys.The textural properties of the catalyst were evaluated through nitrogen adsorption-desorption isotherms at \u2212196\u00a0\u00b0C as displayed in Fig.\u00a02\na. It can be observed that all the isotherms present a type III shape according to the IUPAC classification, which is mainly attributed to non-porous or macroporous materials and hysteresis loop type H3 usually found in materials with a wide distribution of pore size [31]. The specific surface area and pore volume of the catalysts were calculated using the BET model and D-R model, respectively. As depicted in Table 1, the incorporation of the second metal does not promote any change in the surface area, which may confirm the high dispersion of the active phase in all catalysts. Additionally, the pore volume all catalyst remains similar. However, in the bimetallic catalysts, it is observed an increase in the average particle size as a result of the possible formation of alloys [32]. The pore distribution calculated using the BJH method is shown in Fig.\u00a02b, where a similar distribution is observed among the monometallic catalyst and the bimetallic Rh and Pt bimetallic catalysts showing a narrower pore distribution with a maximum displaced to bigger pore diameter as an effect of alloys formation.The hydrogen consumption profiles obtained in the H2-TPR measurements for all the as-synthesized catalysts are shown in Fig.\u00a03\n. In general terms, the ruthenium oxide reduction can be represented by Eq. (9):\n\n(9)\n\n\nR\nu\n\nO\nx\n\n+\nx\n\nH\n2\n\n\u2192\nR\n\nu\n0\n\n+\nx\n\nH\n2\n\nO\n\n\n\nwhere x identifies the possible different ruthenium oxides that may be present in the sample. If we assume that the oxidation state of Ru is +4 for the determination of the reducibility degree, the Eq. (10) indicates that to reduce 1\u00a0mol of ruthenium oxide, 2\u00a0mol of H2 are required.\n\n(10)\n\n\nR\nu\n\nO\n2\n\n+\n2\n\nH\n2\n\n\u2192\nR\n\nu\n0\n\n+\n2\n\nH\n2\n\nO\n\n\n\n\nBased on previous results reported in the literature related to Ru/TiO2, these types of catalysts show three TPR signals in their profiles attributed to three different reduction zones [33,34]. The signal at lowest temperature (approximately at 105\u00a0\u00b0C) has been established to be related to RuOx well dispersed amorphous species. The second reduction zone, at approximately 125\u00a0\u00b0C, is attributed to the RuOx species within the bulk. Finally, the peak at 150\u00a0\u00b0C was attributed to the reduction of the RuOx species strongly interacting with the support, being these ones the hardest to be reduced.All the catalysts present the typical reduction profiles described above. However, the bimetallic catalysts present an additional reduction zone at low temperature (<100\u00a0\u00b0C), mainly attributed to the so-called spillover effect promoted by those noble species. For instance, Kim et\u00a0al. [35] reported the synthesis of 1\u00a0wt% Pt/TiO2 calcined at several temperatures, obtaining in the reduction profile three reduction process at 100, 180 and 300\u00a0\u00b0C This agrees with the first reduction profile that can be observed in the PtRu\u2013TiO2 and RhRu\u2013TiO2 catalyst, which is attributed to well dispersed Pt species. Additionally, the RhRu\u2013TiO2 catalyst shows an extra reduction process at 500\u00a0\u00b0C. Wang and Ruckenstein reported the reduction of Rh in a 1% Rh coated MgO catalysts describing mainly two reduction processes, the first at 350\u00a0\u00b0C attributed to the reduction of MgRu2O4 species and the second one at 520\u00a0\u00b0C due to the reduction of Ru2O3 species [36]. However, the reduction process present in the RhRu\u2013TiO2 is shifted to slightly lower values likely due to the higher dispersion of the Rh2O3 species. Besides, in the noble metal on reducible supports TPR profiles, a reduction zone at 180\u00a0\u00b0C is frequently found and attributed to M-Ov-Ti3+ species related to strong metal-support interaction (SMSI) effect, being Ov oxygen vacancies [37,38]. That is a common effect for Ru, Rh and Pt. All samples presented similar reducibility degree of about 100%.\nTable 2\n shows the metal dispersion calculated from H2 chemisorption where is observed a slight decrease in the bimetallic catalyst, which may be attributed to the formation of bimetallic alloys, as was observed in above mentioned results. Komaya et\u00a0al. reported the limitations of hydrogen chemisorption for the determination of the particle applied to a Ru/TiO2 sample, where it was concluded that the dispersion could be overestimated due to a fraction of H2 adsorbed that suffers spillover to the support. As a consequence, the number of active sites can be apparently higher. Additionally, if the sample is treated at high temperature reduction treatments, the metal could be partially encapsulated by the support, underestimating the mean particle size [39]. Table 2 also shows how the introduction of little Pt diminished the total amount of chemisorbed hydrogen to form a monolayer compared to the monometallic catalyst. Aguilar-R\u00edos et\u00a0al. [40] obtained in their work similar results for Sn-modified Pt catalysts. After being doped with a ratio Sn/Pt\u00a0<\u00a01, the chemisorbed hydrogen increases. However, for Sn/Pt\u00a0>\u00a01 the H2-monolayer value diminished. That may be explained in back-bonding terms. In the monometallic Pt catalyst, the occupied \u03c3 orbital of H2 donates electronic density to the 6s Pt's orbital. For its part, the Pt donates electronic density from its 5d\n\nyz\n to the antibonding \u03c3 orbital (back bonding), destabilizing and finally dissociating the bond. The addition of Sn produces a reduction of the electronic density transferred to the H2 \u03c3\u2217 orbital, diminishing the destabilization of the H\u2013H and therefore making it less active. This explained the decrease in the H2 chemisorbed when Sn/Pt\u00a0>\u00a01 but not the increase when the ratio is smaller than 1. Aguilar-R\u00edos et\u00a0al. [40] also explained this suggesting that small amounts of Sn favours the Pt dispersion acting as anchoring sites for Pt due to the affinity that Pt has for Sn (and by extension every metal of Pt group - Ru, Rh, Pd, Os and Ir) according to theoretical calculations. However, according to these authors, despite the higher dispersion, every new created site must be less active.\nFig.\u00a04\n shows the TPD-H2 profiles for the samples after chemisorption experiments in which an H2 monolayer was firstly adsorbed. The two reduction processes observed are characteristics from transition metals and have been classified as H\u2217w (weakly adsorbed hydrogen) and H\u2217s (strongly adsorbed hydrogen). Sayari et\u00a0al. [41] correlated a larger amount of H\u2217w for particles between 0.9 and 2.2\u00a0nm, which is in good agreement with the present work for Ru and RhRu samples. This may be related to the fact that the H2 adsorption/desorption is not dissociative, according to the work of Lin et\u00a0al. where similar profiles were obtained [42]. By contrast, PtRu sample contains a large amount of H\u2217s species and metallic particles are bigger than 2.2\u00a0nm (3.1\u00a0nm in as shown in Table 2). This suggest that Pt could favors the dissociative adsorption of H2.To quantitatively and qualitatively analyze the surface species and their oxidation states before reaction, Ru, RhRu, and PtRu catalysts were ex situ reduced at 300\u00a0\u00b0C for 1\u00a0h in H2 and characterized by XPS. Fig.\u00a05\n shows the Survey spectra of all the samples, where the peaks of the main elements, titanium, oxygen and carbon, are detected. As for the ruthenium peaks, the most intense are Ru3d, followed by Ru3p. Ru3p are very weak in which Ru3p3/2 is practically masked by the intense Ti2p3/2 peak. The Ru3d peaks are located in the C1s zone [43]. The surface composition is included in Table 3\n. Apparently, the surface chemical composition is similar in all the samples.As can be expected, Fig.\u00a06\na shows that the O1s peak recorded at 530.3\u00a0eV is characteristic for oxides and is well suited with the peaks detected at 459.1 Ti 2p3/2 and 464.8 Ti 2p1/2\u00a0eV, with 2p doublet splitting of 5.7\u00a0eV, which is typical for TiO2 (Fig.\u00a06b) [43\u201345]. As seen in Fig.\u00a06c, there is no peak at 461.2(3) eV for Ru 3p3/2 in any of the spectra of the three samples [45]. This could be related to the low surface concentration of ruthenium as also reflected in the Table 3 of surface composition. Ruthenium was detected in all catalysts by the peak at Ru 3d5/2\u00a0at 279.7\u00a0eV. As shown in Fig.\u00a06c, the position of the peak is not affected by the incorporation of Rh or Pt. The regions Rh 3d and Pt 4f are also shown in Fig.\u00a06d and e, respectively. From the position of the peaks for Ru 3d5/2 (279.7\u00a0eV) as well as for Rh 3d (307.3\u00a0eV with DS 4.7\u00a0eV) and for Pt 4f (70.2\u00a0eV with DS 3.4\u00a0eV), we can conclude that these are metallic species [44\u201346].The catalysts were tested in selective CO methanation using two reaction gas compositions simulating a typical output stream of a water-gas shift (WGS) unit: H2 (50\u00a0vol%), CO2 (15\u00a0vol%), H2O (15\u00a0vol%) and CO (1\u00a0vol% or 300\u00a0ppm) balanced with N2. Fig.\u00a07\n shows the performance in terms of CO and CO2 conversion and CO methanation selectivity as a function of reaction temperature for all the catalysts when the reaction was performed with 1\u00a0vol% of CO inlet. As can be noticed, increasing the temperature leads to an exponential growth of CO conversion between 200 and 240\u00a0\u00b0C (Fig.\u00a07a). Meanwhile, Fig.\u00a07b shows that CO2 conversion was initiated at about 220\u00a0\u00b0C and increasing the temperature above 260\u00a0\u00b0C a drop in CO conversion become simultaneously evidenced (Fig.\u00a07a inlet). This negative CO conversion as temperature was increased suggests that CO is produced in parallel via reverse water gas shift (RWGS) reaction. Consequently, CO methanation is thermodynamically unfavorable at higher temperatures (Fig.\u00a07c) and the effluent CO concentration is increased with increasing temperature. Although the three catalysts presented similar activity, it can be noticed that PtRu catalyst shows minor CO conversion at high temperature (Fig.\u00a07a inlet). Furthermore, Fig.\u00a07c clearly shows that Pt addition affects negatively the selectivity of CO methanation. According to Xu et\u00a0al. [47], the reason for this low CH4 production in PtRu catalyst can be related to the alloy formation that prevents the C\u2013O bond cleavage of and subsequent hydrogenation of C to CH4. This argument is coherent with our results of characterization discussed above. On the other hand, it is also significant in Fig.\u00a07b inlet that Rh addition increases slightly the CO2 conversion although selectivity of CO methanation is hardly affected. This indicates that Rh favors CO2 methanation against reverse water gas shift. It is well known that Rh is an excellent active metal for CO2 methanation [48].Subsequently and without reactivation, the catalysts were tested in the same conditions but with a lower concentration of CO (300\u00a0ppm) balanced with N2. As shown in Fig.\u00a08\n, the effect of the metal doping only is positive in the case of Rh doped catalyst, which increases the catalytic activity and selectivity in the whole temperature range. Similarly for the three samples, CO methanation was initiated at relatively low temperatures achieving conversion about 70% at 130\u00a0\u00b0C and increasing with temperature until completed CO conversion. On the other side, the CO2 methanation initiates once the CO conversion has reached approximately the 90% in every case. It is noteworthy that CO conversion decreases at higher temperatures due to the RWGS contribution, which tend to produce CO from CO2 and hydrogen, and which is highly visible in PtRu but not appreciated in the Rh doped one.It has been tested that polymer-electrolyte-membranes fuel cells (PEMFCs) only tolerates H2 stream with CO concentration below 20\u00a0ppm. Considering that inlet steam contains 300\u00a0ppm of CO, a minimum conversion of 93% will be required to decrease the concentration below 20\u00a0ppm in the outlet stream. In order to compare the three catalysts, we have defined T93 as the required temperature to achieve CO conversion of 93%. Remarkably, Fig.\u00a08 inlet shows that RuRh bimetallic catalyst decreases the temperature T93 of selective CO methanation in absence of CO2 methanation. In comparison to monometallic Ru and bimetallic PtRu, this variation decrease the total consumption of hydrogen and it could become important in industrial-volume streams.In this work, a Ru supported on TiO2 catalyst was doped with Rh and Pt by wet impregnation method and tested in CO selective methanation using two simulated mixtures from reforming reactors. The obtained catalytic results showed that with a lower concentration of CO (300\u00a0ppm) the addition of Rh as dopant results to be advantageous in the methanation catalyst while that Pt has a negative effect since it promotes the reverse water gas shift reaction decreasing the selectivity of methanation. At higher concentration values of CO (1\u00a0vol%), Pt addition is also detrimental for CO selective methanation whereas that Ru and RuRh catalysts are practically identical. Although the positive effect of Rh observed at low CO concentrations is relatively mild and may not compensate for its use in this particular case due to the expensive nature of this metal, this approach takes one step further for a better understanding of the promotion effect of other noble metals in a Ru methanation catalyst.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Financial support for this work has been obtained from the Spanish Ministerio de Ciencia, Innovaci\u00f3n y Universidades (Grant: RTI2018-096294-B-C33) and Junta de Andaluc\u00eda project with reference US-1263288, both programs being co-funded by the European Union FEDER.", "descript": "\n                  Selective CO methanation from H2-rich stream has been regarded as a promising route for deep removal of low CO concentration and catalytic hydrogen purification processes. This work is focused on the development of more efficient catalysts applied in practical conditions. For this purpose, we prepared a series of catalysts based on Ru supported over titania and promoted with small amounts of Rh and Pt. Characterization details revealed that Rh and Pt modify the electronic properties of Ru. The results of catalytic activity showed that Pt has a negative effect since it promotes the reverse water gas shift reaction decreasing the selectivity of methanation but Rh increases remarkably the activity and selectivity of CO methanation. The obtained results suggest that RuRh-based catalyst could become important for the treatment of industrial-volume streams.\n               "}
{"full_text": "Hydrogen is considered the most promising fuel applied in fuel-cell vehicle because of the wide applicability, high application efficiency, safety, environmentally friendly and inexhaustible reserves characteristics of it [1\u20133]. An accepted view considers that the widespread utilization of clean renewable energy vehicles is an effective measure to deal with environmental pollution and excessive consumption of the limited fossil fuels because transportation consumes about a quarter of global energy while there is about 23% carbon dioxide emission originating from the combustion of fossil fuels in the world [4\u20136]. However, the chemical energy of unit volume hydrogen is low, which limits the commercialization of it. Therefore, a lot of investigations pay attention to dealing with this problem and finding the safe, suitable hydrogen storing materials that can meet the needs of using hydrogen in the field of motor vehicle and electronic products [7\u201311]. The ideal hydrogen storage methods should have the following characteristics, large capacity, low-temperature fast adsorption rate and high cycle survivability [12]. In present, hydrogen storage methods include high-pressure storage, liquid storage, physical adsorption storage and hydride storage. Generally, the gas storage needs 70\u00a0MPa pressure for reaching only 4.8\u202fwt.% H2 capacity [13]. The condition of liquid storage is low-temperature (20\u00a0K) that brings about huge energy consumption (about 30% for filling) and short boiling storage duration [14]. Solid hydrogen storage is reckoned as the best one among the various methods [15,16].Among the metal hydrides, MgH2 with high hydrogen storage capacity (7.6\u00a0wt.%), low cost, non-toxicity, abundant reserves and superior reversibility characteristics is the most promising material for hydrogen storage [17,18]. Nonetheless, the attempt of the widespread application of commercial MgH2 is retarded by its some disadvantages, such as high thermodynamic stability (i.e. the high strength of the Mg-H bonds), the complicated activation procedure, the high dissociation temperature and the sluggish hydrogenation/dehydrogenation kinetics [19,20]. To copy with the dilemma, many studies have been conducted to reduce the decomposition temperature, accelerate the adsorption kinetics and change the reaction thermodynamics by the method of grain refinement [11,21], mixing catalysts [22\u201324], ball-milling [25,26], alloying [27,28], surface modification [29,30] and the other [31]. The mechanical grinding and melt-spinning [32,33] are very efficient means to reduce the particle or grain size. Especially, the mechanical milling can realize the refinement of grains and particles simultaneously [38]. In addition, mechanical milling brings on the formation of crystal defects and modifies the superficial characteristics of alloy particles [1]. Hence, mechanical grinding is reckoned as the best way to improve the hydrogen storing performance of magnesium based materials [34]. Besides, refining the particle to very small size will introduce the capillarity effect that can ameliorate the hydrogen absorption and desorption thermodynamics in theory. The calculation results indicate that a particle radius on the order of 5\u00a0nm will reduce the dehydrogenation enthalpy of pure Mg by about 10% [35]. As well known, there is considerable energy used for the hydrogen dissociation on the magnesium surface and this segment is regarded as a rate controlling factor in the procedure of hydrogen absorption [36]. Transition metals either in their pure form (e.g.: Ni, Ti, Nb, Fe, Co, Al etc.) [37,38] or as oxides (e.g.: Nb2O5, Fe2O3, TiO2 etc.) [39\u201342], hydrides (e.g.: TiH2, ZrFe2Hx, etc.) [43,44], fluoride (e.g.: FeF3, TiF3, NiF2, NbF5 etc.) [45,46] or intermetallics [47,48], can act as the catalysts because they can weaken this dissociation energy. As studied by Du et\u00a0al. [49] and Pozzo and Alfe [50], the hydrogen dissociation energy on magnesium surface is 1.15\u00a0eV but it can be decreased to 0.03, 0.06, 0.56, 0.39\u00a0eV by the addition of Co, Ni, Cu, Pd, severally. Daryani et\u00a0al. [51] researched that adding 6 mol% TiO2 could improve the hydrogen absorbing kinetics and reduce the decomposition temperature of as-milled Magnesium hydride by 100\u00a0K. According to the research of Shahi et\u00a0al. [52], the composite MgH2 \u22125\u00a0wt.% Ni absorbs 5.0\u00a0wt.% H2 hydrogen at the temperature of 443\u00a0K in 15\u00a0min and at the temperature 613\u00a0K it starts to decompose. As investigated by Hou et\u00a0al. [53], Mg2NiH4 with the catalyst composed of MWCNTs and TiF3 has the 503\u00a0K (516.6\u00a0K for pure Mg2NiH4) hydrogen releasing temperature (T\nD) and the activation energy (E\na) of it is 53.24\u00a0kJ/mol (90.13\u00a0kJ/mol for pure Mg2NiH4). In particular, adding appropriate rare earths or their oxides can obviously make the Mg-based hydrides unstable and accelerate the rate of dehydrogenation reaction [54,56]. Lass [56] found that the Mg85Ni15-\n\nx\nM\nx\n (M= La, x\u202f=\u202f0 or 5) alloys possess a lower enthalpy change in the reaction of producing MgH2 and Mg2NiH4. On the basis of the investigation of Luo et\u00a0al. [55], the element Y is beneficial to improve the thermodynamics property of magnesium based materials and the composite Mg90In5Y5 has a lower \u0394H (about 62.9\u00a0kJ/(mol H2)) in comparison with the Mg95In5 binary alloy (about 67.9\u00a0kJ/(mol H2)) and pure Mg (about 74.9\u00a0kJ/(mol H2)). Sadhasivam et\u00a0al. [57] found that the original desorption temperature of the composite MgH2 \u22125\u00a0wt.% Mm-oxide was reduced by 76\u00a0K from 654 to 578\u00a0K. Kalinichenka et\u00a0al. [58] researched the improved reaction kinetics of Mg90Ni8RE2 (RE\u202f=\u202fY, Nd, Gd) and found that the activated Mg90Ni8RE2 could reversible absorb and release 5.5\u00a0wt.% H2 within 20\u00a0min.According to our investigation on REMg11Ni (RE\u202f=\u202fSm, Y)\u202f+\u202f5\u00a0wt.% M (M\u202f=\u00a0MoS2, CeO2) composites, the additives MoS2 and CeO2 play a catalytic role in improving hydrogen storing performance [59,60] and 5\u00a0wt.% addition of catalysts is optimal as studied in this reference [33]. It must be very interesting to compare the effects of TiO2 and La2O3 additives with high hardness on the hydrogen storing performance of ball milling magnesium based materials. Thereby, the alloys La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fNone, TiO2, La2O3) were fabricated by mechanical milling. The thermodynamics and dynamics of the experimental alloys were investigated. A comprehensive comparison of the impacts of different catalysts on the structure and hydrogen storing performances of the alloys is conducted.The La7Sm3Mg80Ni10 material was fabricated by inductive melting La, Sm, Mg, Ni (purity \u2265 99.9%) under 0.04\u00a0MPa He (purity \u2265 99.999%) to inhibit the volatilization of magnesium. To compensate the melting losses, additional magnesium (8\u00a0wt.%) and RE (RE\u202f=\u202fLa, Sm) (5\u00a0wt.%) are required. The above-mentioned materials were all provided by CISRI Corporation. A Varian Liberty 100 inductively-coupled plasma (ICP) was applied to determining the chemical composition of experimental alloys. Then the ingot was mechanically crushed and ground to the 200\u2013400 meshes powders. The obtained power with 5\u00a0wt.% TiO2 or La2O3 (purity \u2265 99.9%) catalyst was mechanically ground by a mill crusher at the speed of 350\u00a0rpm (the weight ratio of specimen and balls is 1: 40). The milling duration is set at 20\u00a0h. Thus, the chemical compositions of the as-milled powder were La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fNone, TiO2, La2O3). In order to heat dissipation and reduce the cold welding of powder in the process of milling, the working mechanism of ball mill is to stop half an hour every 3\u00a0h and the powder adhered to the milling chamber walls and grinding balls needs to be scrapped in due course of time all of these operations were operated under the protective atmosphere of Ar.X-ray diffraction (XRD) (D/max/2400) determined the phase structures and compositions of the alloys. The experimental parameters were 40\u00a0kV, 160\u00a0mA, and 2\u00b0/min with 2\u03b8 changing from 20\u00b0 to 90\u00b0. The radiation was CuK\u03b11 filtered by graphite. The particles morphology observation was completed by a scanning electron microscope (SEM) (QUANTA 400). A high resolution transmission electron microscope (HRTEM) (JEM-2100F, operated at 200\u00a0kV) was utilized to the characterization of microstructure and crystalline state.Hydrogenation and dehydrogenation kinetics curves of the as-milled specimens were tested by automatic Sieverts apparatus. Prior to measuring, the sample need to be activated by six hydriding/dehydriding cycles (633\u00a0K and original hydrogen pressure of 3\u00a0MPa for hydrogen absorption, 633\u00a0K and 1\u202f\u00d7\u202f10\u22124\u00a0MPa original pressure for hydrogen desorption). The temperature of hydrogen absorption was set as 473, 513, 533, 553, 573, 593, 613 and 633\u00a0K, severally, while 553, 573, 593, 613 and 633\u00a0K for hydrogen desorption. The setting of initial hydrogen pressure is the same as activation. The sample mass required for every determination was 300\u00a0mg. Non-isothermal hydrogen desorption property was researched by utilizing thermogravimetry (TGA) and differential scanning calorimetry (DSC) (SDTQ600) whose heating rates were 5, 10, 15 and 20\u00a0K/min.\nFig.\u00a01\n gives the X-ray diffraction of the as-milled La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) composites before and after hydrogen absorption and desorption under the condition of 633\u00a0K, 3\u00a0MPa and 633\u00a0K, 1\u202f\u00d7\u202f10\u22124\u00a0MPa, severally. ICDD (International Centre for Diffraction Data) identification of X-ray diffraction patterns shows that the as-milled M\u202f=\u202fnone specimen consists of the major phase La2Mg17 and secondary phases Mg2Ni and La2Ni3. The addition of catalysts TiO2 and La2O3 do not introduce any new phase, indicating that these additives are not involved in the reaction with alloy. Moreover, it is visible to observe the broadened diffraction peaks representing the typical nanocrystalline and amorphous structures of as-milled specimens in comparison with that of the as-cast one (XRD patterns are not show here). After hydrogen absorption, the diffraction peaks get narrow and sharp. Meanwhile there are four hydrides become visible and emerge in the specimens, including MgH2, Mg2NiH4, LaH3 and Sm3H7. The reaction relationship between the elements is as follows:\n\nLa2Mg17\u202f+\u202fH2\u202f\u2192\u202fLaH3\u202f+\u00a0MgH2\n\n\n\n\n\nMg2Ni\u202f+\u202fH2\u202f\u2192\u00a0Mg2NiH4\n\n\n\n\n\nLa2Ni3\u202f+\u202fLa2Mg17\u202f+\u202fH2\u202f\u2192\u202fLaH3\u202f+\u00a0Mg2NiH4\u202f+\u00a0MgH2\n\n\n\n\n\nSm\u202f+\u202fH2\u202f\u2192\u202fSm3H7\n\n\n\nAfter dehydrogenated, the four phases Mg, Mg2Ni, LaH3 and Sm3H7 can be found. Evidently, the LaH3 and Sm3H7 phases are not decomposed because of the high thermal steadiness of them. Hence, the hydrogen desorption reactions are summarized as the following two equations:\n\nMgH2\u202f\u2192\u00a0Mg\u202f+\u202fH2\n\n\n\n\n\nMg2NiH4\u202f\u2192\u00a0Mg2Ni\u202f+\u202fH2\n\n\n\nAccording to the above inference, we can see that in the process of hydrogenation and dehydrogenation, the reversible reactions of activated composites include\n\nMg\u202f+\u202fH2 \u2194 MgH2\n\n\n\n\n\nMg2Ni\u202f+\u202fH2 \u2194 Mg2NiH4\n\n\n\nThrough a careful observation, we find that the width of the XRD peak narrows down after dehydrogenation compared with that after hydrogen absorption, which was owing to the cell volume reduction and stress relief rendered by hydrogen desorption. As found by Montone et\u00a0al. [61], the volume of a metallic Mg atom is about 33% smaller than that of Mg atom in MgH2. It has been reported in the literature [62,63] that lattice distortion along with expansion and contraction of cell volume are inevitable in hydrogen storage materials during hydrogen absorption and desorption, which will cause many lattice defects such as vacancy and dislocation. The formation of the defects will have a beneficial effect on the hydrogen absorption and desorption property of the alloy. Mechanical milling of Mg-based alloy with TiO2 and La2O3 catalysts creates the defects on the surface and inside the magnesium matrix, which generate reactive clean surfaces and shrink the particle size of Mg. The creation of defects facilitates nucleation, the production of the reactive clean surface enhances the superficial reactivity, and the diminution of particle decreases the diffusion distances of hydrogen atoms. These effects ameliorate the hydriding and dehydriding kinetics of magnesium based alloy significantly.The morphologies of the as-milled La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) alloy powder are observed by SEM and presented in Fig.\u00a02\n. As can be observed, the alloy particles have the typical morphology of ball-milling powder and the size of them is in a range of 1\u201310\u00a0\u00b5m. After careful observation, it is failed to find the catalysts TiO2 and La2O3 particles, which indicates that the TiO2 and La2O3 particles are not appear on the superficial part of alloy, but is wrapped in them. Evidently, the agglomeration tendency of the as-milled particles of the M\u202f=\u202fTiO2 and M\u202f=\u202fLa2O3 alloys was decreased (Fig.\u00a02b and c) with smaller size than that of M\u202f=\u202fnone alloy. It means that adding a certain amount of TiO2 and La2O3 can significantly improve the efficiency of ball milling. After comparing the particles with different catalysts, we found no obvious difference in particle size, suggesting that two catalysts have similar effect on the efficiency of ball milling. As considered by Floriano et\u00a0al. [64], some catalysts with high hardness, e.g. La2O3, CeO2, TiO2 Nd2O5, etc. can act as lubricants, dispersants and/or cracking agents in the procedure of milling and are helpful to further reduce refine the particles of as-milled alloy. A very similar result also appears in the investigation of Daryani et\u00a0al. [51] and Aguey-Zinsou et\u00a0al. [65].The HRTEM micrographs and SAED (Selected Area Electron Diffraction) patterns of the as-milled La7Sm3Mg80Ni10 \u22125\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) materials are presented in Fig.\u00a03\n. We have noticed the nanocrystalline and amorphous structures of the as-milled alloys and the emergence of crystal defects. The SAED patterns also prove the existence of La2Mg17, Mg2Ni and La2Ni3 phases, and there is no any new phase caused by adding TiO2 and La2O3. After hydrogen absorption, the structures of as-milled alloy still are amorphous and nanocrystalline (Fig.\u00a03b, e and h), but there is an observably decrease in amorphous phase, meaning that the dehydrogenation promotes the crystallization reaction. Four hydrides MgH2, Mg2NiH4 LaH3 and Sm3H7 also can be identified after hydrogen desorption by SAED patterns. According to Fig.\u00a03(c), (f) and (i), we can see that the alloys after hydrogen desorption exhibit an entirely crystal structure, and the size of grain evidently increase, and Pukazhselvan et\u00a0al. [66] also had the similar report. The SAED rings of dehydrogenated alloys reflect the existence of Mg, Mg2Ni, LaH3 and Sm3H7. Apparently, it is consistent with the result of XRD, the LaH3 and Sm3H7 phases still exist after dehydrogenation. In addition, it is found from Fig.\u00a02 that the LaH3, Sm3H7, TiO2 and La2O3 nanoparticals distribute in Mg matrix dispersedly and uniformly, which is considered to be the preferred nucleation sites for hydride formation/decomposition. The phase interfaces of LaH3 (or Sm3H7, TiO2 and La2O3)/Mg (or MgH2) provide channels for the diffusion of hydrogen atoms. Therefore, the additives can be regarded as catalysts to improve the hydrogen storage performance of magnesium and Mg-based alloy [67].In this investigation, it was found that the alloy powder prepared by traditional mechanical milling can hardly absorb hydrogen because the alloy powder exposed to air easily forms an oxide film on the particle surface that blocks the contact between H2 molecules and alloy surface and prevents H2 from dissociating into H atoms. As well known, this dissociation process is the basic step in the phase transformation from metallic Mg to MgH2 and it is necessary for the incorporation of H atoms into the Mg lattice. Fortunately, it is found that when the alloy sample is kept under proper temperature and hydrogen pressure for a long time, the oxide film formed can be broken gradually, which results in exposing the fresh alloy surface and restoring the hydrogen absorption capability of the alloys. This process is called as activation. The activation performance of specimens is greater if it needs less cycle numbers. Fig.\u00a04\n demonstrates the isothermal hydrogen absorption and desorption curves of the activated La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) materials. It is found from Fig.\u00a04(a), (b) and (c), the alloys are almost fully activated after the first cycle since the activation curves of the next five cycles are almost identical. It is noted that the time spent on the first hydrogen absorption to a saturated state is long. It takes 21716 s for the M\u202f=\u202fnone alloy, 14568 s for the M\u202f=\u202fTiO2 alloy and 13340 s for the M\u202f=\u202fLa2O3 alloy to achieve the saturated capacity of 5.15\u00a0wt.%, 5.052\u00a0wt.% and 4.916\u00a0wt.%, severally, suggesting that the activation property of the alloy is considerably improved by adding TiO2 and La2O3. For a given hydrogen absorption capacity of 4\u00a0wt.%, by which the time required is 4820, 4624 and 4758\u00a0s corresponding to the as-milled M\u202f=\u202fnone, M\u202f=\u202fTiO2 and M\u202f=\u202fLa2O3 alloys, respectively. It indicates that the hydrogenation rate is in the order M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone. The improved activation ability by adding TiO2 and La2O3 was attributed to the modified particles surface state and the increased defect density of the crystals resulted from adding catalysts TiO2 and La2O3. The catalyst nanoparticles distributing on the particles surface significantly increase the dissociation rate (limiting factor of hydrogen absorption rate) of hydrogen molecules.Generally speaking, the first activation reaction is a long procedure [68] and in this process, the H atoms penetrate the formed oxide layer to form mental hydrides. What's more, the attendant mechanical stress and lattice distortion are unfavorable to the absorption of hydrogen [69]. As well known, the nucleation of MgH2 on the superficial sites of alloy is retarded by the thin oxide layers [70]. Although the operation is conducted at inert gas atmospheres the oxide layers with 3\u20134\u00a0nm thickness still can easily form [69]. The sluggish dissociation of H2 on the alloy surface is another reason to explain the slow hydrogen absorption rate [71]. The dissociation on pure Mg surface needs high energy [48]. Besides, the diffusion of H atoms in the metal hydrides is difficult [56, 57]. The growth rate of MgH2 is decided by the hydrogen pressure due to the fact that higher pressure provides the greater the thermodynamics driving force for the hydrogen absorption. Nevertheless, if the original hydrogenation process is fast enough, a superficial layer of magnesium hydride will form to retard the hydrogen permeation [71]. Because hydrogen diffuses along the interfaces but not along the Mg hydride layer [72], the MgH2 hydride grows up in the form of slow Mg/Mg hydride interface movement. When the thickness reaches a certain value (30\u201350\u00a0\u00b5m), the hydrogen absorption reaction stops [73], indicating that powdered magnesium used for hydrogenation changes into massive magnesium. So the hydrogen absorption rate is affected by the powder size [74]. The hydrogenation kinetics is markedly enhanced after the first hydrogen absorption and desorption cycle. With the increase in the cycle number, the hydrogen absorption kinetics curves have little change, which means the great activation performance of experimental alloys. Noticeably, the hydrogen absorbing capacity of all the specimens after first cycle first is no more than 4.8\u00a0wt.%, which represents a visible decline in the capacity. The formation of stable hydrides LaH3 and Sm3H7 is most likely responsible for the 0.25\u00a0wt.% capacity loss.The dehydrogenation curves of the alloys are provided in Fig.\u00a04(d), (e) and (f). It is visible that the dehydrogenation rate is fast and the dehydrogenation kinetics of the alloy was markedly improved by adding TiO2 and La2O3. In particular, the first hydrogen desorption took less time. For a given hydrogen desorption capacity of 3\u00a0wt.%, by which the time required is 193, 162 and 175\u00a0s corresponding to the as-milled M\u202f=\u202fnone, M\u202f=\u202fTiO2 and M\u202f=\u202fLa2O3 alloys, respectively. Evidently, the dehydrogenation rate is in the order M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone. The improved activation performance is deemed to be related to the decreased particle size, the surface modification and the weakening effect of Mg-H bond strength resulted from the additive TiO2 or La2O3. The improvement of thermodynamics performance is directly associated with the weakening of Mg-H bond strength. The co doping of multi elements, especially the transition element (or their compounds) and rare element (or their compounds) is beneficial to the reduction of thermal stability of MgH2 [75,76]. The reduction of particle observably enhances the decomposition rate of H2 on the particle surface and is beneficial to the H atoms diffusion thus enhance the activation performance [77]. Particularly, because TiO2 and La2O3 are high hardness particles, they are likely to cut into the alloy particles and form a new interface under the action of high impact stress in the ball-milling process, which may become the nucleation sites of hydride, acting as rapid paths for atoms diffusion [78]. So the addition of TiO2 and La2O3 not only enhance the efficiency of mechanical milling, make the particle size decrease but also modify the surface of alloy particles, make the nucleation of hydrides more easy.After the activation treatment, the hydrogen absorbing and desorbing properties of experimental composites were improved significantly. It is necessary to explore the change of structures in the process of activation. With the help of SEM, the morphological variations of the experimental composites before and after activation process are provided in Fig.\u00a05\n. Clearly, the particles show irregular morphologies with the very rough surface. After six hydriding and dehydriding cycles, the particle morphologies of the alloy have a dramatically change. It is very evident that many cracks appear on powder surface due to the lattice stress forming in the process of hydrogen absorption. When the lattice stress exceeds the fracture strength of the material, the pulverization of the alloy is inevitable and results in the improved properties. Through the above structural analysis, we believe that the activation is significant to the formation and decomposition of hydrides, the oxide film on the surface of alloy particles breaks, and along with the cracking of alloy particles, the specific surface area of the alloy is increased, thus improving the hydrogen absorbing and desorbing properties.To explore the influence of the different catalysts, the hydrogen absorption curves of the as-milled La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) composites were tested at different temperatures from 423 to 633\u00a0K and 3\u00a0MPa, as shown in Fig.\u00a06\n. It is noted that at the initial stage corresponding to the rapid formation of hydride layer near the surface, the rate is very fast and hydrogen absorption capacity can reach more than 85% of saturated capacity in less than 200\u00a0s, while in the following stage it takes long time to achieve the saturated state due to the hindrance of formed hydride layer acting on the hydrogen diffusion. Freidlmeier et\u00a0al. [79] considered that when the thickness of hydride layer reached to a certain value (100\u00a0nm), the rate of hydrogen absorption tended to 0. Aiming at investigate the hydrogen absorption kinetics more deeply, the time required to absorb 4\u00a0wt.% hydrogen was calculated and compared. As obtained from Fig.\u00a06, the spent time is 108, 67 and 56\u00a0s at 473, 513 and 533\u00a0K for the M\u202f=\u202fnone specimen, 96, 62 and 48\u00a0s for the M\u202f=\u202fTiO2 composite, and 103, 64 and 53\u00a0s for the M\u202f=\u202fLa2O3 material, respectively. Apparently, the hydrogenation rate of the composites is in order of M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone, which suggests that the added TiO2 and La2O3 notably ameliorate the hydrogen absorption kinetics, but this favorable effect decreases rapidly with hydrogen absorption temperature rising. Noticeably, the alloys almost display the same hydrogen absorption kinetics when the temperature exceeds 513\u00a0K, indicating the predominant role of temperature among many factors acting on the rate of hydrogen absorption. As we all know, there are three steps happening in the hydrogenation procedure of magnesium [80], namely a) the dissociation of superficial H2 molecules, b) the diffusing of H atoms through grain boundaries, c) the combination of H atoms and Mg atoms to form MgH2 on the Mg/catalyst interfaces. Because the hydrogen dissociation needs quite high energy, it is reckoned as the rate-controlling step [81]. As confirmed by Sakintuna et\u00a0al. [35], the additives transition metals or their oxides in magnesium can act as the catalysts to decrease the dissociation energy. Liu et\u00a0al. [82] considered that theoretically, the substitution atoms weakened the stability of Mg-H bond owing to the interaction between the valence electrons of H and the unsaturated d/f electron shell of the transition metals or oxides, improving the hydrogen absorption performance. Agarwal et\u00a0al. [47] reported that it is difficult to refine the grains of Mg by mechanical milling due to the inevitable agglomeration of particles. The additive of brittle oxides or intermetallics provides convenience for reducing the particle size of Mg. The refined particles ameliorate the hydrogen absorbing and releasing properties due to the decreased diffusion length and the larger reactive surfaces of H2 caused by particle refinement [83]. Compared with the experimental alloy, TiO2 and La2O3 have higher hardness. Therefore, the existence of TiO2 or La2O3 nanoparticles increases the brittleness of alloy and eventually makes the equilibrium between fragmentation and agglomeration change to a reduced particle size, as stated by Rafi-ud-din et\u00a0al. [84]. The shorter diffusion channels for H atoms and larger specific surface for H2 dissociation caused by particle refinement facilitate to enhance hydrogen absorption kinetics [85].To investigate the hydrogenation degree of the alloys and the phase structure changes during hydrogenation, The Rietveld refinements of the XRD patterns of as-milled La7Sm3Mg80Ni10 alloy hydrogenated at 3\u00a0MPa and 593\u00a0K are provided, as illustrated in Fig.\u00a06(d). The milled alloys cannot be fitted with the Rietveld method because their XRD detections are amorphous. After hydrogenation, the amorphous phase is completely crystallized. Thus, the Rietveld method can be used analyzed the evolution of the phases of the as-milled alloy after hydrogenation. The result reveals that the as-milled hydrogenated alloy is composed of four hydrides, viz. MgH2, Mg2NiH4, LaH3 and Sm3H7 and the relative content of each phase is 56.1, 25.9, 11.5 and 6.5%, respectively. It suggested that the alloy is in saturated hydrogenation state.In order to research the relationships between the catalysts TiO2, La2O3 and hydride stability, the temperature programmed desorption and DSC of the as-milled La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) after complete hydrogen absorption were measured at the heating rate of 5\u00a0K/min, as presented in Fig.\u00a07\n. It is observed that the adding catalyst renders an obvious effect on the hydride stability. The onset dehydrogenation temperature of La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) composites after hydrogenation is 547.4, 537.2 and 540.1\u00a0K, severally. The temperatures of endothermic peaks in DSC curves of the alloys are 553.2, 546.4 and 548.9\u00a0K, respectively. The change of initial hydrogen desorption temperature can reflect the hydrides stability. Evidently, the stability of the hydrogenated La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) composites is the following order M\u202f=\u202fnone > M\u202f=\u202fLa2O3 > M\u202f=\u202fTiO2. The decreased stability is attributed to the decline in Mg-H bond energy. The additives transition metals [86] or rare-earth elements can reduce the Mg-H bond energy and act as the catalysts owing to the electronic exchange reaction between these catalysts and magnesium hydride [57]. Pighin et\u00a0al. [87] investigated the function of various catalysts on Mg-H bond energy and found that the addition of transition metals or their oxides effectively decreased this bond energy, the magnesium hydride stability and the dehydrogenation temperature. According to the research of Abdellaoui et\u00a0al. [12], the existence of new bonds weakens the bond strength between Mg and H, which can help us to understand the system instability and the decreased hydrogen desorption temperature mentioned above.Aiming at studying the hydrogen desorption kinetics of the composites with additives TiO2 and La2O3, the plots of capacity versus time were tested at 553, 573, 593, 613 and 633\u00a0K and presented in Fig.\u00a08\n. As can be observed, the reaction temperature greatly affects the hydrogen desorption kinetics of experimental alloys. Under the high-temperature conditions, all the alloys have a very fast reaction rate. In addition, it is noted that the adding catalysts TiO2 and La2O3 generates a favorable impact on the isothermal dehydrogenation kinetics. For further making sense of the influence of adding TiO2 and La2O3 on the kinetics, the time inquired by releasing 3\u00a0wt.% hydrogen is regarded as a reference standard. As shown in Fig.\u00a08, the time required by releasing 3\u00a0wt.% hydrogen at the temperatures of 553, 573, 593, 613 and 633\u00a0K is 988, 553, 419, 227 and 152\u00a0s for the M\u202f=\u202fnone alloy, and 578, 352, 286, 188, and 112\u00a0s for the M\u202f=\u202fTiO2 alloy, and 594, 366, 301, 197 and 132\u00a0s for the M\u202f=\u202fLa2O3 specimen, respectively. Obviously, the dehydrogenation rate is in the order M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone. Based on the above data, the relationship between the time needed by desorbing 3\u00a0wt.% H2 and temperature can be constructed, as displayed in Fig.\u00a08(d). It indicates that the additives TiO2 and La2O3 markedly ameliorate the hydrogen absorption kinetics, but the positive contribution decreases rapidly with the increase of hydrogen desorption temperature, which suggests that among all the factors affecting the dehydrogenation kinetics of alloys, temperature is predominate. It has come to light that the hydrogen desorption of MgH2 is completed through three stages: (a) magnesium phase nucleates and grows, (b) hydrogen diffuses from the magnesium hydride matrix to the surfaces, and (c) two adjacent hydrogen atoms combine to form hydrogen molecules [88]. The improved hydrogen desorption kinetics of magnesium based alloys is most likely attributed to the decline in hydride stability, the diminution of the particles size and the increase of the defect density on the particle surface of the alloys caused by adding catalysts TiO2 and La2O3. As mentioned above, the rare-earth elements and transition metals or their oxides can reduce the bond energy of Mg-H, thus, weaken the magnesium hydrides stability and accelerate the hydrides decomposition [47]. High hardness catalyst particles are likely to be embedded into the interior of alloy particles under the action of repeated impact stress in the process of ball grinding, so the alloy particles broken and particle size greatly reduced, as considered by Jain et\u00a0al. [89]. Meanwhile, the additive high-hardness catalyst induces the surface defects and brings on the particle refinement of alloy in the procedure of mechanical milling [84]. The induced defects are favorable to the nucleation and increase the surface reactivity. Besides, the decreased particle size makes the length of hydrogen diffusion channels shorten. These factors definitely accelerate the hydrogenation and dehydration of the Mg-based materials [90]. Nevertheless, the improved kinetics because of adding TiO2 and La2O3 particles is not only due to the decline in particles size. Other factors such as the nature of the added oxides and the local electronic structure and the reduction of the oxide during heating should also be considered. Partially reduced oxides are expected to have different valence states and may act on ameliorating the hydrogen desorption property. Generally speaking, the main catalysis of transition metal-based catalysts is engendered by the transition metal ions and their ability to form hydrogen bonds. In this way, transition metal-based catalysts provide a faster route for H atoms diffusion. It has been reported that oxygen vacancies (also known as anoxic surfaces) on oxide surfaces also have catalytic activity [84]. Hence, we believe that the hydrogen desorption results from thermodynamically induced surface vacancies. In summary, the improved kinetics may be due to the uniform dispersion of these anoxic oxide particles, which shorten the diffusion path between reaction ions. The oxygen vacancies act as the sites for nucleation and growth of dehydrogenation products, and promote the dehydrogenation process.To investigate the dehydrogenation degree of the hydrides and the phase structure changes during dehydrogenation, The Rietveld refinements of the XRD patterns of as-milled saturated hydrides dehydrogenated at 1\u202f\u00d7\u202f10\u22124\u00a0MPa and 633\u00a0K are given, as illustrated in Fig.\u00a08(d). It reveals that the phase Mg, Mg2Ni, LaH3 and Sm3H7 exist in the dehydrogenated alloy. It is very clear that rare earth hydrides LaH3 and Sm3H7 remain undecomposed at experimental temperatures and pressures. The relative content of each phase in the alloy is 55.7, 26.1, 11.6 and 6.6%, respectively.Generally, the occurrence of gas-solid reaction needs to overcome a total energy barrier that can be reflected in terms of the apparent activation energy. Hence, when the activation energy reaches a certain requirement, the reaction can take place smoothly. The apparent activation energy of the hydrogenated La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) alloys in hydrogen desorption is evaluated by the Arrhenius and Kissinger methods. As well known, the nucleation and growth of dehydrogenation products are the crucial factors that control the hydrogen desorption reaction of magnesium based materials [91]. In general, the Johnson-Mehl-Avrami-Kolmogorov (JMAK) model can simulate this solid-state reaction [92]:\n\n(1)\n\n\nln\n\n[\n\u2212\nln\n(\n1\n\u2212\n\u03b1\n)\n]\n\n=\n\u03b7\nln\nk\n+\n\u03b7\nln\nt\n\n\n\n\n\na \u2014 the reaction fraction;\nn \u2014 the avrami index;\nk \u2014 the dehydrogenation rate constant;\nt \u2014 the reaction time.According to Fig.\u00a08, the fitting curves ln [-ln (1-\u03b1)] vs. lnt at 573, 593, 613 and 633\u00a0K can be ploted, as provided in Fig.\u00a09\n. As can be observed, the JMAK sketches are almost linear, suggesting the dehydrogenation of the composite is composed of two steps, including the first stage instantaneous nucleation and the second stage 3D growth controlled by interface [93]. The \u03b7 and \u03b7lnk values were obtained according to the slope and intercept in fitting curves at the corresponding temperature. Thus the value of k can be acquired. The apparent activation energy (\n\nE\n\na\n\nde\n\n) of dehydrogenation reaction was estimated gained by using Arrhenius formula [57]:\n\n(2)\n\n\nk\n=\nA\nexp\n\n(\n\n\n\u2212\n\nE\n\na\n\nde\n\n\n\nR\nT\n\n\n)\n\n\n\n\n\n\nA \u2014 a temperature independent coefficient;\nR \u2014 the gas constant (8.3145\u00a0J/mol/K);\nT \u2014 the absolute temperature of reaction;\nk \u2014 the dehydrogenation rate constant.The Arrhenius plots of lnk vs. 1/T of the alloys are sketched, as presented in Fig.\u00a09. The apparent activation energy \n\nE\n\na\n\nde\n\n of the as-milled alloys was acquired from the slopes of the Arrhenius plots. 68.1, 62.1 and 63.6\u00a0kJ/mol correspond to the La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) alloys, severally. Obviously, the apparent activation energy of the dehydrogenation of the alloys is in sequenced M\u202f=\u202fTiO2 < M\u202f=\u202fLa2O3 < M\u202f=\u202fnone. As reported by Rafi-ud-din et\u00a0al. [84], the catalysts TiO2 added by mechanical milling has shown further enhancement in the reaction kinetics by reducing the H2 dissociation activation energy. Mustafa and Ismail [83] considered that the improved ability of MgH2 decomposition was originated from the decline in this activation energy. Hou et\u00a0al. [53] reported that the proper catalysts were proved to be an efficient strategy to decrease apparent activation energy E\na of MgH2 hydrides.In order to compare with the JMAK model, the Kissinger method is employed to estimate the activation energy, as following equation [94]:\n\n(3)\n\n\n\n\nd\n\n[\n\nln\n\n(\n\n\u03b2\n/\n\nT\n\nP\n\n2\n\n\n)\n\n\n]\n\n\n\nd\n\n(\n\n1\n/\n\nT\nP\n\n\n)\n\n\n\n=\n\n\n\u2212\n\nE\n\nk\n\nde\n\n\nR\n\n\n\n\n\n\n\u03b2 \u2014 the heating rate;\n\n\nE\n\nk\n\nde\n\n \u2014the activation energy;\nT\nP \u2014 the absolute temperature corresponding to the maximal desorption rate,\nR \u2014 the gas constant (8.3145\u00a0J/mol/K).The as-milled La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) alloys need to absorb H2 to the saturated state before the measure of DSC. Fig.\u00a010\n demonstrates the non-isothermal dehydrogenation curves tested at the heating rates of 5, 10, 15 and 20\u00a0K/min, severally. As we can see, an endothermic peak exists in each DSC curve, suggesting the same reaction procedure of each specimen. According to Fig.\u00a010, the graphs of \n\nln\n(\n\u03b2\n/\n\nT\n\nP\n\n2\n\n)\n\nvs. 1/T\nP can be sketched, as presented in Fig.\u00a010. It is noted that \n\nln\n(\n\u03b2\n/\n\nT\n\nP\n\n2\n\n)\n\nvs. 1/T\nP plot is almost linear, so from the slopes of it, the activation energy \n\nE\n\nk\n\nde\n\n was obtained. According to the calculation, apparent activation energies of the La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) alloys are 65.5, 60.2 and 61.7\u00a0kJ/mol and in order of M\u202f=\u202fTiO2 < M\u202f=\u202fLa2O3 < M\u202f=\u202fnone. The addition of TiO2 or La2O3 distinctly reduces the apparent activation energy. Fan et\u00a0al. [95] believe that as a significant index to evaluate the hydrogen desorption property, the decreased activation energy indicates the decline in the barrier of hydrogen desorption. Thus, it can be concluded, the additives TiO2 or La2O3 can decrease the hydrogen releasing activation energy, which is the essence behind the improved dehydrogenation kinetics.Reducing the thermal stability of magnesium hydride is the main goal to improve its hydrogen storage performance and realize its practical application. To inspect the effect of adding TiO2 and La2O3 on the thermodynamics, the P-C-T curves of the as-milled specimens were tested at the temperature of 593, 613 and 633\u00a0K and given in Fig.\u00a011\n. Obviously, the pressure platform is quite flat and the hysteresis coefficient (H\nf\u202f=\u202fln (P\na/P\nd)) is small. The catalyst TiO2 or La2O3 has no evidently change in the platform characteristic reflected in the P-C-T curves of alloys. As we can observe, two pressure plateaus emerge in every P-C-T curve and the higher and lower platform pressures stand for the formation/dissociation of the Mg2NiH4 and MgH2 hydrides, severally [96,97]. According to the plateau pressures (P\na and P\nd) in Fig.\u00a011, the thermodynamics parameters enthalpy change \u0394H and entropy change \u0394S are evaluated by Van't Hoff equation [98]:\n\n(4)\n\n\nln\n\n(\n\n\nP\n\nH\n2\n\n\n\nP\n0\n\n\n)\n\n=\n\n\n\n\u0394\n\nH\n\n\nR\nT\n\n\n\u2212\n\n\n\n\u0394\n\nS\n\nR\n\n\n\n\n\n\nP\nH2 \u2014 the equilibrium hydrogen gas pressure corresponding to MgH2;\nP\n0 \u2014 the standard atmospheric pressure;\nR \u2014 the gas constant (8.3145\u202fJ/mol/K);\nT \u2014 the absolute temperature of reaction.The Van't Hoff graphs of \n\nln\n\n\nP\n\nH\n2\n\n\n/\n\nP\n0\n\n\n\n vs. 1/T for the as-milled La7Sm3Mg80Ni10\u20135\u00a0M (M\u202f=\u202fnone, TiO2, La2O3) composites can be sketched. Hence, the \u0394H and \u0394S can be calculated according to the slopes and intercepts in Van't Hoff diagrams and listed in Table\u00a01\n. It uncovers that the addition of TiO2 and La2O3 has not notably impact on the improvement of the experimental materials\u2019 thermodynamics and the reduction of corresponding hydrides stability. The addition of catalysts TiO2 or La2O3 decreases\u00a0 the stability magnesium hydride. A similar result also emerged in the investigations of Anik et\u00a0al. [99] and Bououdina et\u00a0al. [39]. Obviously, the absolute values of dehydrogenation enthalpy change \u0394H\nde of the alloys are in following order M\u202f=\u202fnone > M\u202f=\u202fLa2O3 > M\u202f=\u202fTiO2. Based on the above results, we can find that both isothermal and non-isothermal analyses reveal that the catalysts TiO2 and La2O3 weaken the magnesium hydride stability and improve the hydrogen absorption and desorption kinetics. The positive contribution to the hydrogen storing thermodynamic and dynamics of the specimens caused by two catalysts is in following order M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone.\n\n\n(1)\nThe addition of TiO2 and La2O3 has no change in the phase composition but reduces the agglomeration tendency of particles in the process of mechanical milling and make the particle size of the as-milled alloy markedly decreased. It is this modification of the microstructure that remarkably enhances the hydrogen absorption and desorption performances.\n\n\n(2)\nThe addition of TiO2 and La2O3 have obviously positive contribution to the hydrogenation and dehydrogenation kinetics of the experimental alloys, which is in order M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone. It is ascribed to the decline in the size of grains and particles, the generation of the fresh surface and the creation of the various crystal defects derived from ball milling and adding catalysts.\n\n\n(3)\nThe addition of TiO2 and La2O3 catalysts has a slightly favorable influence on the improvement of the thermodynamics of alloy and the stability of the hydride, which is in sequence M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone.\n\n\nThe addition of TiO2 and La2O3 has no change in the phase composition but reduces the agglomeration tendency of particles in the process of mechanical milling and make the particle size of the as-milled alloy markedly decreased. It is this modification of the microstructure that remarkably enhances the hydrogen absorption and desorption performances.The addition of TiO2 and La2O3 have obviously positive contribution to the hydrogenation and dehydrogenation kinetics of the experimental alloys, which is in order M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone. It is ascribed to the decline in the size of grains and particles, the generation of the fresh surface and the creation of the various crystal defects derived from ball milling and adding catalysts.The addition of TiO2 and La2O3 catalysts has a slightly favorable influence on the improvement of the thermodynamics of alloy and the stability of the hydride, which is in sequence M\u202f=\u202fTiO2 > M\u202f=\u202fLa2O3 > M\u202f=\u202fnone.We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of the manuscript entitled.This study was financially supported by the National Natural Science Foundation of China (Nos. 51901105, 51871125, and 51761032), Natural Science Foundation of Inner Mongolia, China (2019BS05005), Inner Mongolia University of Science and Technology Innovation Fund (2019QDL-B11) and Major Science and Technology Innovation Projects in Shandong Province (2019JZZY010320).", "descript": "\n                  In this investigation, mechanical grinding was applied to fabricating the Mg-based alloys La7Sm3Mg80Ni10\u202f+\u202f5\u202fwt.% M (M\u202f=\u202fNone, TiO2, La2O3) (named La7Sm3Mg80Ni10\u20135\u202fM (M\u202f=\u202fNone, TiO2, La2O3)). The result reveals that the structures of as-milled alloys consist of amorphous and nanocrystalline. The particle sizes of the added M (M\u202f=\u202fTiO2, La2O3) alloys obviously diminish in comparison with the M\u202f=\u202fNone specimen, suggesting that the catalysts TiO2 and La2O3 can enhance the grinding efficiency. What's more, the additives TiO2 and La2O3 observably improve the activation performance and reaction kinetics of the composite. The time required by releasing 3\u202fwt.% hydrogen at 553, 573 and 593\u202fK is 988, 553 and 419\u202fs for the M= None sample, and 578, 352 and 286\u202fs for the M\u202f=\u202fTiO2 composite, and 594, 366, 301\u202fs for the La2O3 containing alloy, respectively. The absolute value of hydrogenation enthalpy change |\u0394H| of the M (M\u202f=\u202fNone, TiO2, La2O3) alloys is 77.13, 74.28 and 75.28\u00a0kJ/mol. Furthermore, the addition of catalysts reduces the hydrogen desorption activation energy (\n                        \n                           E\n                           \n                              a\n                           \n                           de\n                        \n                     ).\n               "}
{"full_text": "The aqueous electrocatalytic reduction of CO2 into energy-dense industrial chemical fuels and feedstocks has been proposed as a promising strategy to mitigate the challenge of CO2-induced global warming [1\u20133]. A wide range of carbon compounds (such as CO, CH4 and HCOOH) are possible products of this process, allowing an efficient pathway to simultaneous CO2 fixation and storage of a renewable energy source under ambient conditions [4,5]. However, due to the intrinsic thermodynamic stability of CO2 and the complex reaction pathway of the CO2 reduction reaction (CRR), it is still challenging to find a cost-effective and stable electrocatalyst that can directly reduce CO2 to carbonaceous products [6\u20139]. In an attempt to solve this problem, many electrocatalysts have been considered for CRR; however, even the most well-known Au and Ag-based catalysts cannot meet the criteria, due to their high overpotential and high cost, as well as their easy deactivation during the CRR process [10,11]. Hence, rational design of non-noble metal electrocatalysts with high selectivity and stability is critical for the application of CO2 electroreduction technology.Achieving atomic-level regulation of active transition metal atoms is important in designing an efficient catalyst [12\u201314]. In this regard, single-atom catalysts (SACs) have emerged as a highly promising category of electrocatalyst owing to their optimized atomic utilization, strong metal\u2013substrate interactions and highly unsaturated coordination environment [15\u201317]. Moreover, there is the possibility of chemical potential tuning, in which the size, structure, shape and composition of materials can be controlled to alter the electronic structure of the SACs [18,19]. These characteristics thus make SACs ideally suited as electrocatalysts for a series of reactions including the hydrogen evolution reaction (HER) [20], oxygen reduction reaction (ORR) [21] and CO2 reduction reaction [22,23]. The preparation of SACs with a controllable microstructure and highly exposed active metal atoms is thus highly desirable.In this work, we demonstrate the successful synthesis of single-atom-Ni-decorated, nitrogen-doped carbon (denoted SA-Ni@NC) layers by carbonizing the precursor: layers of a two-dimensional bimetallic zeolite imidazolate framework (ZnNi-ZIF). After the selective etching of Zn atoms at high temperature, single atoms of Ni can be preserved and simultaneously immobilized on the resulting nitrogen-doped carbon layers via Ni\u2013N bonds. Unlike general bulk ZIF precursors, which are nearly one hundred nanometers in size, our 2D ZIF layers allow single metal atoms in the prepared ultrathin SA-Ni@NC to be fully exposed to the electrolyte. As a result, the SA-Ni@NC layers exhibit excellent electrocatalytic activity for CRR: a high Faradaic efficiency (FE) of 86.2% is achieved at \u22120.6\u202fV (vs. reversible hydrogen electrode (RHE)). A single-atom catalyst of this type with a highly exposed active site and high catalytic activity may open a new approach to the design of other CRR electrocatalysts.Nickel nitrate hexahydrate (Ni(NO3)2\u00b76H2O), zinc nitrate hexahydrate (Zn(NO3)2\u00b76H2O), cobalt nitrate hexahydrate (Co(NO3)2\u00b76H2O), methanol and 2-methylimidazole were purchased from Beijing Innochem Technology Co. Ltd.Typically, 900.0\u202fmg of 2-methylimidazole was added to 60\u202fmL of deionized water, producing solution A. After that, 300.0\u202fmg of Zn(NO3)2\u00b76H2O and 29.3\u202fmg of Ni(NO3)2\u00b76H2O were added to 60\u202fmL of deionized water, and the resulting solution was then added to solution A under continuous stirring (600\u202frpm) for 12\u202fh. Finally, Zn1Ni0.1-ZIF layers were obtained after washing and drying treatments. The Zn1Ni0.1-ZIF layers were then annealed at 1000\u202f\u00b0C for 3\u202fh at a heating rate of 10\u202f\u00b0C\u202fmin\u22121 under a flow of Ar/H2 (v/v\u202f=\u202f9:1) gas mixture. Finally, the black product (SA-Ni@NC) was collected after cooling to room temperature. The procedure for synthesizing NC, SA-Ni-2@NC, P-Ni-2@NC and P-Ni@NC was similar to that for SA-Ni@NC, except that the amount of Ni(NO3)2\u00b76H2O was changed to 0, 58.6, 117.2 and 293.0\u202fmg, respectively.Typically, 900.0\u202fmg of 2-methylimidazole was added to 60\u202fmL of deionized water, producing solution A. After that, 300.0\u202fmg of Zn(NO3)2\u00b76H2O and 29.3\u202fmg of Co(NO3)2\u00b76H2O were added to 60\u202fmL of deionized water, and the resulting solution was then added to solution A under continuous stirring (600\u202frpm) for 12\u202fh. Finally, Zn1Co0.1-ZIF layers were obtained after washing and drying treatments. The as-prepared Zn1Co0.1-ZIF layers were then annealed at 1000\u202f\u00b0C for 3\u202fh at a heating rate of 10\u202f\u00b0C\u202fmin\u22121 under a flow of Ar/H2 (v/v\u202f=\u202f9:1) gas mixture. Finally, the black product (SA-Co@NC) was collected after cooling to room temperature. The synthetic procedure for P-Co@NC was similar to that of SA-Co@NC, except that the amount of Co(NO3)2\u00b76H2O was changed to 293.0\u202fmg.Typically, 900.0\u202fmg of 2-methylimidazole was added to 30\u202fmL of methanol, producing solution A. After that, 300.0\u202fmg of Zn(NO3)2\u00b76H2O and 29.3\u202fmg of Ni(NO3)2\u00b76H2O were added to 30\u202fmL of methanol, and the resulting solution was then added to solution A under continuous stirring (600\u202frpm) for 10\u202fmin. The mixed solution was then transferred to 100\u202fmL Teflon-lined stainless-steel autoclaves and heated at 120\u202f\u00b0C for 6\u202fh. Zn1Ni0.1-ZIF-particles were then obtained after washing and drying treatments. After that, the Zn1Ni0.1-ZIF-particles were annealed at 1000\u202f\u00b0C for 3\u202fh at a heating rate of 10\u202f\u00b0C\u202fmin\u22121 under a flow of Ar/H2 (v/v\u202f=\u202f9:1) gas mixture. Finally, the black product (SA-Ni@3D-NC) was collected after cooling to room temperature.The morphologies and microstructures of the samples were characterized by field emission scanning electron microscopy (FE-SEM JEOL-7500) and high-resolution transmission electron microscopy (HRTEM, JEOL, NEM-2100F). Chemical states were characterized by X-ray photoelectron spectroscopy (XPS, Thermo escalab 250Xi). X-ray diffraction (XRD, Rigaku D/max2500PC) was carried out using Cu K\u03b1 radiation over the range 5\u201380\u00b0. X-ray absorption fine structure (XAFS) spectroscopy was conducted at the 1W1B station of the Beijing Synchrotron Radiation Facility. The storage rings of the Beijing Synchrotron Radiation Facility were used at 2.5\u202fGeV with a maximum current of 250\u202fmA. A Veeco/Bruker (ICON) atomic force microscope was used to measure the thickness of the sample. The Ni content of the SA-Ni@NC was measured by ICP-ASE (Optima-7000DV). Nitrogen sorption isotherms and BET surface area were measured by Qudrasorb at 77\u202fK.The fabrication of single-atom-Ni-decorated, nitrogen-doped carbon layers is schematically illustrated in Fig. 1\na. Initially, zinc\u2013nickel bimetallic ZIF layers were synthesized by mixing Zn2+ and Ni2+-containing solutions with 2-methylimidazole. This process ensures the mutual diffusion of Zn and Ni atoms in the resulting ZIF layers, and the successful preparation of the ZIF crystal was confirmed by X-ray diffraction (XRD) (Fig. 1b) [24,25]. Moreover, scanning electron microscopy (SEM) images show the ultrathin 2D morphology of the resulting ZIF layers (Figs. 1c and S1). After that, the ZnNi-ZIF layered precursor was annealed at 1000\u202f\u00b0C in a mixture of Ar/H2 (v/v\u202f=\u202f9/1). Since the saturated vapor pressure of Zn is much higher than that of Ni (Fig. S2), Zn atoms rather than Ni atoms are selectively volatilized from the ZnNi-ZIF layers during the high temperature treatment, generating ultrathin SA-Ni@NC layers (Figs. 1d and S3). Meanwhile, the Ni atoms are preserved and coordinated with the N atoms in the nitrogen-doped layers to form Ni\u2013N bonds, generating single-atom-Ni-decorated, nitrogen-doped carbon layers. It should be noted that the atomic ratio of Ni to Zn has a great influence on the product. By adjusting the atomic ratio of Ni to Zn, nitrogen-doped carbon (denoted NC, Ni/Zn\u202f=\u202f0) layers and Ni-particles-decorated, nitrogen-doped carbon (denoted P-Ni@NC, Ni/Zn\u202f=\u202f1:1) layers were also prepared (Figs. S4 and S5).The morphology and microstructure of the prepared SA-Ni@NC layers were investigated via transmission electron microscopy (TEM). Abundant thin and transparent layers were obtained, as shown in Fig. 2\na and b, indicating the ultrathin nature of the SA-Ni@NC layers. The lateral sizes of these carbon layers are typically in the range from 500\u202fnm to several micrometers. These layers have good flexibility, possibly originating from their intrinsically flexible nature and/or from the defective structures formed during the synthesis process. Moreover, as shown in Fig. 2c, the high resolution-TEM (HR-TEM) image reveals the obviously amorphous structure of the SA-Ni@NC layers, which can be further confirmed by the corresponding fast Fourier transform (FFT) patterns (inset in Fig. 2c). No distinct nanoparticles or clusters can be seen on the surface of the layers, suggesting that the Ni atoms might be present in the form of single atoms [26\u201328]. In contrast, numerous Ni particles are observed on the surface of the P-Ni@NC layers (Fig. S7c), and the existence of Ni particles in the P-Ni@NC sample can be further confirmed by the XRD patterns (Fig. 2d). The atomic force microscopy (AFM) image (Fig. 2e) and the corresponding thickness analyses (Fig. 2f) further reveal that the SA-Ni@NC layers have a thickness of \u223c2.5\u202fnm.To verify the isolated, dispersed nature of Ni atoms in the SA-Ni@NC layers, synchrotron-based X-ray absorption fine structure (XAFS) measurements were also conducted. As shown in the X-ray absorption near-edge structure (XANES) spectra (Fig. 3\na and b), the position of the blue line for SA-Ni@NC layers is located between those for the Ni foil (black line) and NiO (red line), clearly suggesting the typical electronic structure of Ni\n\u03b4\n\n+ (0\u202f<\u202f\u03b4\u202f<\u202f2) [29]. Further Fourier transforms of R space for Ni K-edge EXAFS were conducted, and compared with Ni foil, NiO and phthalein cyanide nickel (Ni\u2013Pc) as references. As shown in Fig. 3c, SA-Ni@NC layers exhibit a dominant Ni\u2013N coordination peak at 1.41\u202f\u00c5, which is nearly identical to the peak of the Ni\u2013Pc reference sample (1.45\u202f\u00c5), suggesting an interaction between Ni and N atoms in the nitrogen-doped carbon layers [30]. In contrast, the P-Ni@NC layers exhibit an obvious peak at 2.02\u202f\u00c5, showing the presence of Ni particles in the sample. Moreover, the single atom nature was further confirmed by wavelet transform (WT) of Ni K-edge EXAFS oscillation. As shown in Fig. 3d, there is only one intensity maximum at 6\u202f\u00c5\u22121 for SA-Ni@NC layers, which can be ascribed to the Ni\u2013N bonding. We also carried out least-squares curve fitting to obtain the quantitative structural parameters of Ni in the SA-Ni@NC layers, and the fitting curves are shown in Fig. 3e. According to the fitting, the coordination number of Ni is 3.3 (Table S1), indicating that the Ni atoms mainly have three-fold coordination with N atoms.Additionally, the presence of Ni\u2013N bonds was further confirmed by the XPS spectrum of the SA-Ni@NC layers. As shown in Fig. 3f, the high-resolution N 1\u202fs spectrum for SA-Ni@NC can be deconvoluted into five peaks corresponding to oxidized N (402.5\u202feV), graphitic N (401.3\u202feV), pyrrolic N (400.6\u202feV), Ni\u2013N (399.1\u202feV) and pyridinic N (398.5\u202feV) species [31]. The existence of Ni\u2013N bonds is in excellent agreement with our EXAFS analysis. These results suggest that the Ni atoms are atomically dispersed in the nitrogen-doped carbon layers through Ni\u2013N bonds. More importantly, this method can be extended to produce a series of SACs (such as single-atom-Co-decorated, nitrogen-doped carbon (SA-Co@NC) layers), verifying the generality of this simple method (Figs. S10\u2212S12).The porous nature of the SA-Ni@NC layers was validated by nitrogen physisorption measurements (Fig. S13a). A high specific surface area of 449.0\u202fm2\u202fg\u22121 was obtained for the SA-Ni@NC layers, which is much higher than that of P-Ni@NC layers (246.7\u202fm2\u202fg\u22121). Moreover, the corresponding pore-size distribution curves demonstrate the existence of both micropores and mesopores in the SA-Ni@NC layers (Fig. S13b). These micropores and mesopores originate from the inheritance of ZIF precursors and the evaporation of Zn during the annealing treatment, respectively [32]. The Ni content of the SA-Ni@NC layers was found to be \u223c1.61\u202fwt% using the inductively coupled plasma-atomic emission spectrometry analysis (ICP-AES) measurement.The electrocatalytic activity of the SA-Ni@NC layers for CRR was first examined by linear sweep voltammetry (LSV) in Ar and CO2-saturated 0.1\u202fM KHCO3 electrolytes. As presented in Fig. 4\na, under CO2-saturated conditions, the current density of SA-Ni@NC layers reached a maximum of \u221210.1\u202fmA\u202fcm\u22122 at \u22121.0\u202fV, much higher than that obtained in the Ar-saturated electrolyte (-6.7\u202fmA\u202fcm\u22122). The excess current density is ascribed to the occurrence of CO2 reduction. In addition, the SA-Ni@NC layers exhibit the greatest reduction current density in the CO2-saturated electrolyte, which was approximately 2.1 and 1.7 times higher than those for NC layers (\u22124.8\u202fmA\u202fcm\u22122) and P-Ni@NC layers (\u22125.8\u202fmA\u202fcm\u22122) catalysts at \u22121.0\u202fV, respectively (Fig. 4b).Apart from current density, selectivity is another important criterion for CRR electrocatalysts. Thus, for each sample, potentiostatic electrolysis was conducted at various potentials, and the gas and liquid products were identified by gas chromatography (GC) and 1H nuclear magnetic resonance (NMR), respectively. It is found that the only carbon-containing product detected was CO, and no liquid products were detected by NMR (Fig. S15). Fig. 4c shows the FE of CO production at various applied potentials, and it can be seen that the applied potential greatly affects the distribution of the reduction product. For the SA-Ni@NC catalyst, the production of CO starts at \u22120.3\u202fV. Furthermore, SA-Ni@NC achieves the highest FE value for the production of CO: 86.2% at a potential of \u22120.6\u202fV, which is much higher than the FE values obtained using NC (26.5%) and P-Ni@NC layers (30.2%), demonstrating the superior selectivity of the SA-Ni@NC layers. Moreover, the FE can be further improved by increasing the content of single Ni atoms in the nitrogen-doped layers. As shown in Fig. S18, SA-Ni-2@NC with a high content of single Ni atoms achieves the highest FE value for the production of CO: 98.1% at a potential of \u22120.6\u202fV vs. RHE, which is higher than that of the SA-Ni@NC catalyst (86.2%). The FE of SA-Ni@NC is also higher than that of single-atom-Ni-decorated, 3D nitrogen-doped carbon (67.9%) at \u22120.6\u202fV vs. RHE, Fig. S19, thus demonstrating the importance of designing ultrathin support materials to improve the activity of single-atom catalysts.SA-Ni@NC layers also possess good stability for the electrocatalytic reduction of CO2. As shown in Fig. 4d, both the current density and the corresponding FE for CO production show negligible decay for continuous catalysis over at least 10\u202fh. Moreover, the single atoms of Ni are also stable after electrocatalytic CRR tests (Fig. 4e), showing the good structural stability of the SA-Ni@NC layers. These results indicate that the SA-Ni@NC layers possess both excellent activity and stability towards CO2 reduction. Electrochemical impedance spectroscopy (EIS) measurements were conducted to gain further insight into the kinetics of the CO2 reduction process. As shown in Fig. 4f, SA-Ni@NC shows a much faster charge-transfer rate during the CRR process than the P-Ni@NC and NC samples, which greatly contributes to its catalytic activity.To elucidate the origin of the excellent CO2 reduction properties of the SA-Ni@NC catalyst, DFT calculations were performed on three different sites in SA-Ni@NC, i.e., pristine N-doped carbon (NC), Ni-N3-C (Ni atom with three-fold coordination with N atoms) and Ni-N4-C (Ni atom with four-fold coordination with N atoms). Fig. 5\na depicts the calculated free-energy profiles for CO2 reduction at the three sites, in which a *COOH intermediate with two transferred proton-electrons was considered. It was found that the reaction \n\n\u2217\nC\n\nO\n2\n\n+\n\n\nH\n\n+\n\n+\n\n\ne\n\n-\n\n\u2192\n\n\u2217\n\nC\nO\nO\nH\n\n is the potential limiting step for the pathways at the three sites, in which the lowest barrier of 0.78\u202feV is obtained at Ni-N3-C site, compared to a barrier of 2.90\u202feV at pristine NC and one of 1.57\u202feV at the Ni-N4-C site. Fig. 5b\u20135e show snapshots of the reaction pathway for CO2 reduction at the Ni-N3-C site. It can be seen that the Ni-N3-C site is the active site, facilitating the transformation from CO2 to CO. Local density of states (LDOS) values were calculated to further interpret the electronic origin of the high catalytic activity at the Ni-N3-C site. Fig. S20 presents the calculated local density of states (LDOS) for the atom (N in NC or Ni in Ni-N3-C/Ni-N4-C) bonded to *COOH/CO, as well as that of the group *COOH/CO. It can be seen that there is a stronger hybridization between Ni and *COOH/CO than between N and *COOH/CO, indicating a stronger interaction between Ni and *COOH/CO. Remarkably, a significant DOS distribution consisting of Ni atom at the Fermi level can be observed at the Ni-N4-C site, suggesting its low stability. In contrast, a splitting of DOS from Ni atoms at the Fermi level emerges at the Ni-N3-C site, which might be responsible for its high catalytic activity.In conclusion, we have demonstrated an effective electrocatalyst towards CRR based on single-atom-Ni-decorated, nitrogen-doped carbon layers. The synthesized SA-Ni@NC not only possesses highly active Ni-N3-C sites, but also has a unique 2D structure with a high surface area, multilevel pores and thin walls. Such features not only provide large amounts of available active sites for electrocatalytic CO2 reduction, but also favor fast mass transport during the catalysis. As a consequence, SA-Ni@NC layers exhibit high CRR activity for the production of CO, and also show good stability. Moreover, this simple method can also be used for the preparation of a series of single-atom catalysts (such as SA-Co@NC). In view of the large family of zeolite imidazolate frameworks, we anticipate that a series of single-atom-decorated, nitrogen-doped carbon layers could be fabricated with broader applications in the area of energy conversion technology.\nChao Zhang: Conceptualization, Methodology, Software, Data curation, Writing - review & editing. Zhongheng Fu: Methodology, Software. Qi Zhao: Software, Data curation. Zhiguo Du: Methodology, Software. Ruifeng Zhang: Methodology, Data curation. Songmei Li: Conceptualization, Methodology, Supervision, Validation, Project administration.This work was financially supported by National Natural Science Foundation of China (No. 51622203).The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Supplementary data to this article can be found online at https://doi.org/10.1016/j.elecom.2020.106758.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  Efficient selective electroreduction of carbon dioxide into energy-dense carbonaceous fuel products is highly desirable to mitigate environmental and energy-related problems. However, there is still a need to design an electrocatalyst with high selectivity and stability towards the CO2 reduction reaction (CRR). Here, we present the promising performance of single-atom-Ni-decorated, nitrogen-doped carbon layers (SA-Ni@NC) as an efficient electrocatalyst for CRR. In this catalyst the Ni atoms are atomically dispersed and most have three-fold coordination with the N atoms in the carbon layers. Theoretical calculations show that the Ni-N3-C site can act as a highly active site for the reduction of CO2 owing to the low energy barrier for the formation of *COOH intermediates. As a consequence, SA-Ni@NC exhibits a high Faradaic efficiency (up to 86.2%) for the production of CO at a potential of \u22120.6\u202fV versus the reversible hydrogen electrode. Moreover, this simple method can also be used to produce a range of single-atom catalysts (such as SA-Co@NC). In view of the large family of zeolite imidazolate frameworks, we anticipate that our strategy will be extended to a variety of single-atom-decorated, nitrogen-doped carbon layers with a broad range of applications in energy conversion systems.\n               "}
{"full_text": "Excessive consumption of fossil fuels and rapid population growth have led to several environmental problems, including greenhouse gas emission, SOx, NOx, acid rain, global warming, and urban smog (Abas et al., 2015; Abokyi et al., 2019; Zhang et al., 2018). Furthermore, the fluctuation of fossil fuels prices and the heavy reliance of energy and chemical sectors on fossil fuels have caused a dramatic increase in demand for alternative, renewable and sustainable energy. Biomasses stand out as a suitable renewable energy source to produce liquid fuels due to their environmental benefits, such as abundant availability, renewability, low cost and carbon neutral (Long et al., 2013). About 220 billion metric tons of lignocellulosic biomass are generated annually worldwide, making biomass the world's largest renewable source of energy (Hassan et al., 2016). Biomass-derived bio-oil can be an alternative to fossil fuels to produce value-added chemical, heat, electricity, and energy (Yaman et al., 2018). In 2016, lignocellulosic biomass constitutes about 70\u00a0% of the total primary energy supply, which was equivalent to 56.5 EJ as shown in Fig. 1\n (Global Bioenergy Statistics, 2018). Currently, numerous countries have imposed strong policies on the utilization of renewable biofuels. For example, European Union (EU) Commission demands more than 20\u00a0% of the entire automotive fuel usage to be consisted of biofuels by 2020. The U.S governmental departments also have set an aim to achieve 25\u00a0% of oil-based chemicals and 20\u00a0% of transport energy with biofuel-based alternatives by 2030 (Liang et al., 2021).Over the past two decades, increasing population and consumption have driven a massive increase in plastic demand due to its excellent characteristics of durability, light of weight, easy manufacturing, ease of use and resistance to corrosion. The global production of plastics is expected to expand from 300 million metric tons in 2015 to 1.8 billion metric in 2050 (Lee et al., 2021). In 2020, the global plastic production has reached 370 million tonnes, with Asian region contributing to about half of it (PlasticsEurope, 2020). Plastics including polypropylene (PP), polyethylene (PE), polyethylene terephthalate (PET), polystyrene (PS), and polyvinyl chloride (PVC) are extensively utilized in diverse areas, including packaging, construction, electronics, households, automobiles and others (Wang et al., 2021). The incessant growth of plastics demands has resulted in the increase of plastic solid waste (PSW) deposit every year. Municipal solid waste (MSW) accounts for around 30\u201335\u00a0% of the total plastic wastes in industrialized country (Tencati et al., 2016). At present, the traditional recycling methods, including incineration and landfills pose a serious threat to the environment via water resource pollution, air pollution and damages to marine ecosystems and terrestrial habitats (Ghayebzadeh et al., 2020). In addition, the natural degradation of plastic needs 400 to 1000\u00a0years, causing a major negative impact to the environment. Therefore, an alternative approach that can convert the abundant plastic waste into a more value-added product and protect the environment and human health needs to be explored.Co-feeding hydrogen-rich materials to the oxygen-rich biomass has recently paved the way to upgrade bio-oil quality. The co-pyrolysis process is highly similar to pyrolysis because it can deliver high quality bio-oil, but it involves the combination of two or more feed materials. This technique can compensate the flaws of biomass-derived bio-oil, and provide safe and effective waste treatment (Chen et al., 2020). Hydrogen-rich materials such as plastics, tires and lubricant oil can act as hydrogen donor, increase the hydrogen-to-carbon ratio of feedstock and induce positive synergistic interaction with biomass to enhance the oil quality. The interactions between the intermediates of lignocellulosic biomass and synthetic polymers during co-pyrolysis can produce bio-oil with high carbon yields, high calorific value, aromatic selectivity and hydrocarbon (Dorado et al., 2015; Lu et al., 2018b). Furthermore, co-pyrolysis offers economic advantages since it requires less energy than the pyrolysis of biomass and plastic alone (Chen et al., 2020). Suriapparao and Vinu, (2021) investigated the synergistic effects between biomass (rice husk, groundnut shell, bagasse, mixed wood sawdust and Prosopis juliflora) and hydrogen-rich plastics (Polyisoprene (PIP) and low-density polyethylene (LDPE)). The study deduced that co-pyrolysis significantly boosted the calorific value of bio-oil. The heating value of co-pyrolysis oil varied from 38 to 42\u00a0MJ/kg as compared to the heating value of biomass pyrolysis oil of 20 to 28\u00a0MJ/kg. In addition, the deoxygenation degree also increased due to the synergistic effects. Rahman et al. (2021) carried out pyrolysis for mixtures of pine and HDPE in a double-column staged reactor and observed that the addition of HDPE to pine could increase the pyrolysis oil yield up to threefold compared to pyrolysis oil of pine alone. In addition, the oil produced was rich in hydrocarbon with 99\u00a0% selectivity. Adding the catalyst to the co-pyrolysis process could facilitate the cracking of pyrolysis vapor and deoxygenate the oxygenated compounds via dehydration, decarbonylation and decarboxylation reactions, improving selectivity towards the desired compounds, such as hydrocarbon (Dyer et al., 2021).The integration of co-pyrolysis and microwave radiation could enhance the yield and properties of liquid fuel product with less energy input in a single step, and prevent the need of an additional upgrading reactor network. Microwave co-pyrolysis technique are advantageous compared to other pyrolysis techniques, including high heating efficiency, uniform heating, energy saving, fast response and better heat and mass transport even with large particle size (Chen et al., 2022; Suriapparao, et al., 2022). Xia et al. (2021) claimed that this technique could be one of the future trends in advanced pyrolysis technique. Microwave co-pyrolysis can resolve the limitations of conventional microwave pyrolysis of plastic, which produces highly viscous bio-oil consisting of heavy hydrocarbon (waxy paraffinic components) that results from the insufficient decomposition of long-chain hydrocarbon (Wan Mahari et al., 2022). The addition of microwave absorbent can enhance the pyrolysis efficiency since the plastic and biomass waste have weak microwave absorption capacity. The bio-oil obtained from microwave co-pyrolysis has high heating value and H/C ratio with low viscosity (Suriapparao, et al., 2020). This technique also accelerates the dehydration reactions with low energy utilization, resulting in bio-oil that has low moisture content and acidity. Nevertheless, the cracking of long-chain molecules into short-chain compounds and the conversion of oxygenated compounds which are achievable through catalytic up-gradation can enhance the product selectivity and properties.Catalytic co-pyrolysis of biomass-plastic mixture can be a more reliable method compared to the catalytic pyrolysis of single biomass due to the catalyst deactivation resulted from hydrogen deficiency properties of biomass. Catalyst offers an alternative pathway with lower energy requirement for selective product generation. During pyrolysis, catalyst can accelerate the reactions, including cracking, hydrocracking, decarboxylation, alkylation, aromatization, decarboxylation, and Diels-Alder reactions for better product selectivity and quality (Suriapparao et al., 2022). It is critical to deeply understand the characteristics of the catalyst in order to select an appropriate catalyst for an effective co-pyrolysis process. Several comprehensive reviews have summarized the catalytic co-pyrolysis of lignocellulosic biomass and waste plastics. Gin et al. (2021) summarized and discussed the impact of heating systems, experimental conditions, and synergistic effects of the co-pyrolysis of plastic and biomass wastes. In addition, the reaction pathway and the kinetics of the catalytic co-pyrolysis version of the same feedstocks were exclusively presented. In another review, Ryu et al. (2020) summarized the latest progress in catalytic co-pyrolysis of biomass and plastic in terms of feedstock pre-treatment, properties of feedstock and catalyst on the production of the biofuels and desired chemicals, such as aromatic hydrocarbon. However, to the best of the author\u2019s knowledge, a review on the influence of various types of plastic as co-reactant in co-pyrolysis with solid biomass to produce chemicals and liquid fuels is still lacking. The quality and product distribution of co-pyrolysis process depends on the biomass, plastic types and properties, and processing conditions, such as temperature, particle size, residence time, reactor type and catalyst addition. Therefore, this review paper focused on the influence of different types of plastic as the co-reactant in co-pyrolysis with solid biomass on the product distribution, synergistic effect, and quality of bio-oil. Furthermore, this review also provided concise information on the critical role of zeolite-based catalyst (microporous, mesoporous, hierarchical, and metal modified zeolite) and low-cost mineral-based catalyst in upgrading the yield and quality of liquid oil. The characteristics, synthesis methods, advantages, disadvantages, and performance of each catalyst in upgrading the bio-oil through the co-pyrolysis of biomass and plastic were compared in detail. Lastly, the potential challenges and future directions for this technique were also suggested.High density polyethylene (HDPE) is a common waste in municipal solid waste (MSW), with a H/Ceff ratio of 2 (He et al., 2021). HDPE is widely used to produce sturdy bottles, flexible pipes, toys, geomembranes, ropes, cutting boards and others. Due to its less fixed carbon with no oxygen, the addition of HDPE in co-pyrolysis of biomass can lower the formation of coke on catalyst and oxygenated compound in pyrolysis oil. Furthermore, HDPE is favourable for pyrolysis process since it possesses greater than 99\u00a0% volatile content with nil moisture (Rahman et al., 2021).Few studies have reported that the incorporation of HDPE can considerably improve the quality of pyrolysis oil in terms of olefins and monoaromatic hydrocarbon while reducing the undesired product of polycyclic aromatic hydrocarbons (PAHs), which are toxic, carcinogenic, and mutagenic (Chen et al., 2016). He et al. (2021) investigated the co-pyrolysis of HDPE and corn stalk over HZSM-5 using Py-GC/MS in the temperature range and biomass-to-HDPE ratio of 550\u2013800\u00a0\u00b0C and 1.0\u20130:1, respectively. The result revealed that the addition of HDPE to corn stalk sharply reduced the oxygenated compounds from 97.02\u00a0% to 42.03\u00a0%. In addition, significant synergistic effect on condensable volatile organic products (CVOPs) and hydrocarbon was observed as the experimental value of CVOPs and hydrocarbon was higher than the calculated value. During co-pyrolysis, the corn stalk-derived oxygenates could interact with HDPE-derived olefins to form hydrocarbon via Diels-Alder reactions, enhancing the hydrocarbon production while reducing the oxygenated compounds. In addition, the hydrogen atoms transferred from HDPE could promote the hydrocarbon production by enhancing the cracking and deoxidation (decarbonylation, decarboxylation and dehydration) reaction of corn stalk-derived oxygenates to hydrocarbon. Furthermore, the addition of HDPE to corn stalk as co-reactant could inhibit the coke formation by stabilizing the corn stalk-derived oxygenates (lignin-derived phenolic compound) and preventing it from undergoing polymerization on the surface of HZSM-5. A comparable trend was observed by Rahman et al. (2021) who attributed the improvement in gasoline range hydrocarbon to proton supplement provided by HDPE. The highest selectivity of gasoline range hydrocarbon (77\u00a0%) was found at pine-to-HDPE ratio and temperature of 50:50 and 550\u00a0\u00b0C, respectively. The authors also highlighted that the amount of oxygenated compounds reduced from 31.11\u00a0% to 0 as the pine-to-HDPE ratio surged from 100/0 to 0/100 because H/Ceff ratio increased as the HDPE ratio in the feedstock increased, promoting the cracking of oxygenated compounds such as phenolic to gasoline-equivalent hydrocarbons (C6-C12).\nYuan et al. (2018) carried out the co-pyrolysis of cellulose and HDPE at different ratios and reported that the synergistic effects in the co-pyrolysis accelerated the generation of small molecule volatiles, including H2O, CO/C2H4, and CO2. The decomposition of HDPE via chain-end and random scission can transfer hydrogen for the decomposition of cellulose-derived anhydrosugars to aldehyde and ketone while cellulose-derived oxygenated compounds, which act as acceptor, promote the scission of HDPE to alkane and alkene groups. During co-pyrolysis, aldehyde and ketone can be further decomposed to hydrocarbon. Fig. 2\n\n shows the reaction mechanism between cellulose and HDPE at different biomass-to-plastic ratio. In co-pyrolysis of biomass with HDPE, HDPE generally provides positive synergistic effect on bio-oil yield (Hassan et al., 2020; \u00d6nal et al., 2014; Rahman et al., 2021). Co-pyrolysis of discarded newspaper and HDPE produced more bio-oil and less gas than the theoretical value due to cross reaction between newspaper and HDPE which interfered with the degradation of functional groups attached to the cellulose structure of WP (Chen et al., 2016). This condition inhibited the production of gases of low molecular weight and favoured the production of oil. The highest oil yield of 68.43\u00a0wt% was achieved at newspaper-to-HDPE blend ratio of 1:2, and it was 31.59\u00a0% higher than the theoretical data based on weighted averages. Positive synergistic effects on fuel properties were also observed in terms of significant reduction of total acid number and viscosity by 216\u00a0% and 76\u00a0%, respectively. In addition, the quality of bio-oil was also enhanced with maximum hydrocarbon and alcohol yield of 85.88\u00a0%, which was obtained at WP:HDPE ratio of 50:50.LDPE represents the second biggest portion of plastic waste with the approximate consumption of 415 million in 2015. This value is expected to increase by 4\u00a0% in the following years. LDPE is widely used as plastic bags and packaging due to its excellent characteristics of flexibility, ease of processing and low cost (Duan et al., 2021). Compared to HDPE, LDPE has more branching (2\u00a0% of carbon atom) and weaker intermolecular forces. Suriapparao and Vinu, (2021) examined the co-pyrolysis of LDPE with five different biomass and found that the experimental bio-oil yield (13.2 \u2013 32.3\u00a0wt%) was less than the theoretical value (42\u201347.5\u00a0wt%). Excessive cracking of heavier molecules into lighter gases contributed to low bio-oil yield. Although the yield of bio-oil was low, the heating value of bio-oil (37.6\u201341\u00a0MJ/kg) was better than the theoretical value (32.6\u201337.8\u00a0MJ/kg) due to the interaction between oxygen transfer from condensable phase to gas phase and hydrogen release from LDPE vapours. Substantial progress has been observed in the selectivity of naphthalene derivatives, including methylnaphthalene and 2-methylnaphthalene produced from catalytic co-pyrolysis of LDPE, and cellulose and pine wood. LDPE was a high molecular weight polyolefin, and its pyrolysis and chain scission were incomplete, resulting in the production of larger molecules. Furthermore, the addition of LDPE could inhibit the coke formation during pyrolysis of biomass due to the breakdown of LDPE and biomass via free radical. Compared to LDPE, biomass could decompose earlier due to its poor thermal stability to produce primary radicals. As the temperature increased, the LDPE started to decompose to hydrocarbon that was rich in hydrogen (H) radical. The LDPE-derived H radical could promote secondary decomposition of biomass to generate volatile substances. These volatile substances prevented the LDPE from covering the biomass by melting down at high temperature. Furthermore, the coke growth was also hindered due to the inhibition of free radical polymerization by the precipitation of volatile substances (Zheng et al., 2018).\nAl-Maari et al. (2021) studied the co-pyrolysis of empty fruit bunch (EFB) and oil palm frond (PF) with LDPE for bio-oil production. The results showed positive synergistic interaction on the production of aliphatic hydrocarbons and inhibition of oxygenated compounds. The hydrogen released from LDPE enhanced the decarbonylation of carbonyls and sugar, and decarboxylation of acid to hydrocarbon due to oxygen removal via CO and CO2, respectively. In addition, significant synergistic interaction between EFB and PF with LDPE on the production of bio-oil has also been observed. The positive synergistic effect could be attributed to the secondary radical reaction, leading to the condensation of non-condensable fragments. Furthermore, LDPE that acted as the hydrogenation medium for biomass could inhibit the cross-linking reactions and polymerization of biomass, leading to greater biomass weight loss (Aboulkas et al., 2012; Yuan et al., 2018). Co-feeding LDPE and sugarcane bagasse yielded pyrolysis oil which mainly consisted of aliphatic compounds with fewer aromatic compounds as compared to individual biomass with high calorific value of 40\u00a0MJ/kg. The addition of LDPE to sugarcane bagasse enhanced the H/C ratio from 1.25 to 1.47 and boosted the formation of saturated hydrocarbon in the range of C6 \u2013 C25 (Dewangan et al., 2016).PET is the third largest thermoplastic consumed in Europe after polypropylene and LDPE. PET is usually used in a variety of consumer goods, including synthetic polyester fibres, bottles and films due to its characteristics of clear and strong thermoplastic (Choi et al., 2021; Dhahak et al., 2020). \u00d6zsin and P\u00fct\u00fcn, (2018) investigated the co-pyrolysis of PET blended with peach stones and walnut shells using a fixed bed reactor, and observed increased ester and acid compounds and decreased phenolic compound. Maximum acid and ester yield of 65.87\u00a0% and 63.11\u00a0% were achieved in co-pyrolysis of PET with walnut shells and peach stones, respectively. The liquid was dominated by benzenecarboxylic acid with more than 40\u00a0% yield for both co-pyrolysis blend. Benzenecarboxylic acid and vinyl benzoate were formed when the ester link of carboxylic group was broken via beta scission, initiating the decomposition of PET. One of the biggest challenges regarding pyrolysis oil from PET is the high acid content such as benzoic acid. The acidic characteristic of pyrolysis oil can lead to corrosiveness, depreciating the fuel quality. In addition, benzoic acid can clog the pipelines and heat exchanger, triggering issues during operation at industrial scale (Lee et al., 2017). Despite the disadvantages, it is noteworthy that benzoic acid is a valuable precursor/feedstock for various industries (\u00c7epelio\u01e7ullar and P\u00fct\u00fcn., 2014).\nChen et al. (2017) investigated the synergistic interaction effects on char morphology and thermal behaviour during co-pyrolysis of PET with paulownia wood (PAW) using TGA. Their result showed a remarkable deviation between the experimental and calculated value on volatile release. Higher char yield was obtained from the PAW/PET blends at final decomposition temperature of 530\u00a0\u00b0C with \u0394W above zero. In addition, the char yield increased as the PET blending ratio increased. With the increment of PET ratio in the feedstock, more cross-link reaction between PET-derived products and PAW-derived char occurred, leading to greater char production. The PET decomposition played a role as a limiting factor for the cross-linking reaction. Meanwhile, the addition of PET to PAW resulted in the agglomeration morphology of char. Ablative surface and granule cohesion were observed on the char topography as the PET blending ratio increased due to the reaction between PET decomposition products and PAW-derived initial char. PET typically decomposed at temperature between 370\u00a0\u00b0C and 460\u00a0\u00b0C. Non-catalytic pyrolysis of PET produced a liquid containing terephthalic acid and benzoic acid along with CO and CO2 gas whereas co-pyrolysis of PET and biomass formed mainly acid and esters. The upgraded bio-oil from catalytic co-pyrolysis of biomass and PET demonstrated high content of aromatic compounds in the range of C5-C12 of carbon number fuel range (Dyer et al., 2021).Polycarbonates (PC) are a group of thermoplastic polymers that contain carbonate groups in their chemical structures. PC is a prominent engineering plastic due to its characteristics, such as high impact strength, superb thermal resistance, and exceptional electrical insulation properties; and it is widely used in automobile industry, building and construction, and data storage devices such as compact disc and DVDs (Antonakou et al., 2014; Bai et al., 2020). In 2017, the global PC production has reached 6 million metric tons (Do et al., 2018). PC is unrecyclable due to its superior opposition to chemical attacks and difficulty to be extracted from the waste stream. Landfilling PC could pose environment threats due to the leaching of bisphenol. Bisphenol A (BPA) and diphenyl carbonate (DPC) substance contained in PC are regarded as endocrine disruptors that cause serious illnesses, including cancer, threaten adult health and interfere with infant hormones (Bai et al., 2020).\nLiu et al. (2021) researched the co-pyrolysis of pinewood blended with PC to determine the synergistic effect. The extent of synergistic effect was determined via comparison between the experimental result of co-pyrolysis of pinewood-PC mixture with the weighted average values from individual feedstock pyrolysis. Positive synergy between pinewood and PC was obtained due to the enhancement of H2, CO and total syngas yield of 33\u00a0%, 36\u00a0% and 19\u00a0%, respectively, compared to the theoretical value from individual pyrolysis. However, negative synergistic effect was noticed in the formation of CnHm. The variation in synergistic behaviour of different gas components could be attributed to the interactions between PW and PC intermediates during co-pyrolysis, producing more oxygenated compounds (alcohols, carboxylic acids, and aldehydes) with less hydrocarbons. In addition, co-pyrolysis of PW and PC remarkably enhanced the gas production yield (from 67.6\u00a0wt% to 77.2\u00a0wt%) but reduced the tar and char yield compared to the theoretical values from individual feedstock pyrolysis. This phenomenon suggested that the synergistic effects of co-pyrolysis of PW and PC involved both mutual interaction of volatile in gas phase and volatile-solid interaction which enhanced the total conversion of solid feedstock to gases. The pyrolysis of PC tends to generate more phenol via oxygen removal as CO and CO2 (Burra and Gupta, 2018). Decomposition of PC mainly occurs via chain scission mechanism which can be divided into two main reactions: primary step in cyclic oligomers production by an intramolecular exchange reaction and hydrolytic cleavage of the carbonate group, generating hydroxyl-terminated oligomers and CO2 at 400 to 500\u00a0\u00b0C temperature (Jin et al., 2016). Blending pinewood (PW) with PC could enhance this pathway, and stable phenolic intermediates could be formed with the lignin portion, enhancing the breakdown and conversion of PW to low molecular weight aromatics that exist as volatiles, and decreasing the char formation at about 10\u00a0% (Burra and Gupta, 2018). On the other hand, addition of lignin to PC pyrolysis can escalate the decomposition of PC to phenolic type compounds by enhancing the release of CO during co-pyrolysis while inhibiting the aromatic compound (Jin et al., 2016).Polyvinyl chloride (PVC) is widely used in the production of cable and wire insulation, fashion and footwear, packaging, window frames, and water pipes. PVC has a longer lifespan than other packaging plastics. About 44.3 million metric tons of PVC was produced globally in 2018, and by 2025 the world\u2019s market size of PVC is expected to grow to nearly 60 million metric tons (Statistica.com, 2021). PVC is the main source of chlorine in municipal solid waste (MSW) and one of the problematic plastics in the feed. Its presence in the feedstock is limited to less than 5\u00a0% and generally around 1 to 2\u00a0%. The release of chlorinated hydrocarbons and HCl in PVC results in corrosion in the reactor and renders the oil halogenated (Qureshi et al., 2020). As there is no public recycling system for PVC, the proportion of the recovered PVC is relatively low. Moreover, PVC needs to be treated using hydrochloride scrubber for PVC cracking as chloride is not desired in the fuels (Xue et al., 2017).\n\u00d6zsin and P\u00fct\u00fcn, (2018) analysed the synergistic effects during co-pyrolysis of PVC with two solid biomasses (walnut shell and peach stones). Negative synergistic interaction on the liquid yields were observed as the liquid yield during the co-pyrolysis (14.70 \u2013 17.60\u00a0wt%) was lower than the aggregate values (17.21 \u2013 18.64\u00a0wt%). On the other hand, positive synergistic effect was observed with higher aromaticity of tars in co-pyrolysis yields than biomass alone. 1H NMR result showed that both aromatic protons comprised of guaiacyl units (ArH and HC\u00a0=\u00a0C-(conjugated)); and \u0251-hydrogen atoms of the branched chain of aromatic ring carbons, methoxy and aliphatic hydroxyl were increased when PVC was added into the biomasses. Polyenes condensation and aromatization during PVC decomposition contributed to the enhanced formation of tars aromatic. It has been well established that chlorine radicals generated during PVC decomposition could initiate condensation reaction, cyclization and aromatization. In addition, considerable value of PAHs was observed during co-pyrolysis of PVC with walnut shell (64.40\u00a0%) and peach stones (59.06\u00a0%). The decomposition of PVC favoured aromatization reaction and creation of heavier tar compounds via dichlorination, followed by inter-molecular chain transfer; the aromatic chain scission generated two or three aromatic-ring side chain before the coke deposition. HCl release during co-pyrolysis of PVC blends escalated the progression of light tar portions to heavy portions, resulting in the generation of higher molecular weight substances, such as PAHs (Tang et al., 2018).The addition of PVC could instigate the decomposition of pinewood (PW) at lower temperature range due to the acceleration of PW decomposition by HCl from the dehydrochlorination of PVC. In addition, the co-pyrolysis of PW and PVC yielded more char and less liquid compared to the theoretical data. HCl generation from PVC at lower temperature range (230\u2013300\u00a0\u00b0C) promoted the dehydration of cellulose to aldehyde compound which was confirmed from the cleavage of glycosidic units. The hydrogen and oxygen atoms in cellulose were lessened due to the dehydration at low temperature, leading to higher char yield. Furthermore, the dehydration also reduced the tendency of depolymerization, consequently reducing the liquid yield. Furthermore, the PW-derived solid char could also act as a catalyst owing to the presence of some inorganics, such as Cao, K2O and NaO that promoted the secondary cracking of PVC oil to generate more char and gas (Lu et al., 2018a). The presence of PVC could influence the reactivity and activation energy of lignocellulosic biomass. The magnitude of reactivity of co-pyrolysis of cherry seed (CS) and PVC was nearly-two orders higher than the pyrolysis of CS at all heating rates. This observation was credited to the chemical structure of PVC which contained high electronegative chloride ions. The activation energy of co-pyrolysis of CS/PVC fell between CS and PVC value. The deviation between theoretical and experimental value of activation energy signifies the occurrence of synergistic effect between CS and PCV during co-pyrolysis (\u00d6zsin and P\u00fct\u00fcn, 2019).Generally, the addition of PS to biomass can enhance the liquid yield while decreasing the gas and char yield. (Stan\u010din et al., 2021) reported that an addition of 25\u00a0% of PS to sawdust (SD) could double the yield of pyrolysis oil from 31\u00a0% to 62\u00a0%, specifically on the expense of gas formation, indicating the occurrence of synergistic effect in the process. Moreover, blending 25\u00a0% of PS with SD could enhance the quality of bio-oil in terms of reduction of oxygenates and PAHs while promoting the aromatic hydrocarbon. However, when the ratio of PS exceeded 25\u00a0%, a higher generation of undesired benzene derivatives and toxic PAHs became noticeable due to the secondary cracking of PS-derived styrene monomer accelerated by the interaction with biomass feedstock. Benzene derivatives in bio-oil limit its further utilization since such compounds are categorized as carcinogenic.\nSamal et al. (2021) examined the co-pyrolysis of eucalyptus biomass and polystyrene waste on the physiochemical and thermal characteristic of the solid char. Two distinct physiochemical and thermal characteristics of char have been observed basically at temperature below and above 450\u00a0\u00b0C. The char generated below 450\u00a0\u00b0C has high heating value and volatile content with low fixed carbon because of the polystyrene coating on the char surface. The melting polystyrene waste could deposit over biomass at temperature below 450\u00a0\u00b0C, go through volatilization with additional increase in temperature, and be transformed to liquid oil and syngas. Solid fuels with high volatile content and low fixed carbon generally possess low ignition and burnout temperatures and a higher mass-loss rate, making them unstable. However, the increased high heating value due to the existence of waste plastic coating could ease in enhancing the combustion efficiency of the fuel. In contrast, the produced chars at temperature 450\u00a0\u00b0C and above possessed more high heating value and fixed carbon with low volatile content. This kind of solid fuel demonstrates superior combustible behaviour with broader temperature range and longer time for complete combustion, all of which signify an excellent solid fuel.The addition of PS enhances the yield and property of pyrolysis oil. In contrast to pyrolysis oil from biomass (Mahua seeds) alone, the addition of 20\u00a0wt% of PS in co-pyrolysis enhanced the liquid yield from 39.26\u00a0wt% to 45.89\u00a0wt%(Mishra and Mohanty, 2020). At 20\u00a0% blending ratio, the plastic could have a maximum synergistic interaction between particles which subsequently maximize the generation of hot volatiles that could be further transformed to liquid form. At this state, greater heat and mass interaction happened between biomass and plastic particles. However, at 10\u00a0wt% and 30\u00a0wt% blending ratios, the interaction between biomass and plastic particles created negative synergistic effect, reducing the formation of hot volatiles and the production of liquid oil. Furthermore, the higher plastic ratio in the feedstock could cause the plastic melting, which would coat the biomass surface, eventually creating resistance for the discharge of hot volatiles and reducing the liquid yield. The NMR study showed the increment of aromatic and olefinic percentage in the co-pyrolysis oil (as confirmed in the FTIR diagnostics showing the peak of 1650\u00a0cm\u22121 \u20131580\u00a0cm\u22121 attributed to CC stretching vibration). Meanwhile, the GC\u2013MS results revealed that an addition of 20\u00a0wt% of PS as co-reactant substantially enhanced the hydrocarbon compounds and reduced the oxygenate derivatives such as acid, making it attractive compared to thermal pyrolysis oil. However, further upgrading technique is needed due to higher viscosity value than diesel fraction. Van Nguyen et al. (2021) examined the co-pyrolysis of waste PS and coffee-grounds at various blending ratio of 75:25, 50:50, and 25:75. The results revealed that co-pyrolysis could accelerate the deoxygenation reaction, causing a reduction of oxygenated compounds and enhancement in carbon content. The effect was strongest at the PS ratio of 75\u00a0% with reduction of oxygen content to 5.68\u00a0wt%. This condition contributed to an improvement in the calorific value (39.66\u00a0MJ/kg) of pyrolysis oil which was comparable to the heating value of conventional fuel. Table 1\n shows the yields and quality of bio-oil obtained from co-pyrolysis of various biomasses and plastics.Employing suitable catalyst in co-pyrolysis is beneficial to the thermochemical decomposition of biomass and plastic by tailoring the products composition and lowering the activation energy of the reaction. The benefits of catalyst addition in the degradation process include shortening the reaction time, lowering the degradation temperature, promoting the extend of degradation, reducing the amount of solid residue in final products and narrowing the product distribution (Antonakou et al., 2014). In addition, the catalyst helps to direct the reaction toward the desired products via interactions between its structure, and the reaction pyrolyzates and products (Rocha et al., 2020). The effectiveness of a catalyst depends on its acidic characteristics, redox properties, and porosity. Tuning the catalyst acidity based on its density, strength, and type is vital in designing the catalyst as each of these elements have particular influence on the activity, product selectivity and reaction pathway (Antonakou et al., 2014).Zeolite is recognized as the most efficient catalyst to produce high-value chemicals because of its high acidity, high specific surface area, high adsorption ability and shape selectivity (Han et al., 2020; Ryu et al., 2020). Its unique pore structure with strong acidity favours aromatic selectivity with excellent cracking and deoxygenation ability (Hassan et al., 2016). The acidity of zeolite which is expressed by the Si/Al ratio determines their reactivity and affects the end products of pyrolysis process with low ratio, indicating high acidity (Chi et al., 2018). Generally, the introduction of microporous zeolite in the pyrolysis is usually favourable to enhance the aromatic production.It is well established that the introduction of microporous zeolite in the pyrolysis is favourable to enhance the aromatic production. Park et al. (2019b) investigated the co-pyrolysis of Quercus variabilis (Q. variabilis) and waste plastic films (PFs) over two microporous zeolites (HZSM-5 and HY) of different acidity and surface area. The acidity (SiO2/Al2O3) of HZSM-5 and HY zeolite was adjusted to 30 and 23, respectively. The result showed that HZSM-5 with higher and stronger acidity could enhance the aromatics production than HY catalyst during the co-pyrolysis at 600\u00a0\u00b0C due to higher cracking efficiency of pyrolyzates. In addition, more appropriate shape selectivity of HZSM-5 which has medium pore size, appropriate pore window size and internal pore volume together with steric hindrance characteristic could favour the production of aromatics (Jae et al., 2011). On the other hand, higher formation of coke was observed for HY catalyst due to the more space provided as it had higher surface area (780\u00a0m2/g) than HZSM-5 (425\u00a0m2/g). In contrast, Kim et al. (2016) observed greater aromatic production over HZSM-5 catalyst at high temperature and catalyst-to-reactant ratio compared to HY catalyst during catalytic co-pyrolysis of cellulose-PP/LDPE mixture. HZSM-5 which had strong acidity was advantageous for aromatic production while high catalyst-to-reactant ratio of 1:10 could provide a large number of active sites for aromatization reaction. The authors emphasized that the properties of catalyst, specifically acidity and pore size, are crucial in determining the aromatic production efficacy during catalytic co-pyrolysis reaction. On the other hand, low temperature and less catalyst-to-reactant ratio were applied for HY catalyst since the reaction intermediates could diffuse easily into its pore and make intimate contact with active sites to undergo further reaction to form aromatic.Coke deposition and limitation of mass transfer and reactant flow diffusion are among the major challenges of pyrolysis over microporous zeolite (Kim et al., 2017b). The small pore size (less than2 nm) of microporous HZSM-5 zeolite inhibits the diffusion of large biomass and plastic reaction intermediates produced during the initial stage of pyrolysis into its internal acidic sites. The large molecules of biomass and plastic pyrolysis intermediates formed during the initial stage of pyrolysis cannot pass through the inner pores and contact the active sites of HZSM-5 since their kinetic diameter is greater than the pore size of ZSM-5 (Hassan et al., 2019). Furthermore, pore blockage from polymerization and polycondensation reactions due to acidic properties of zeolite causes deactivation of the catalyst and reduces the catalyst lifetime. Shao et al. (2017) reported that the parallel side reactions of anhydrosugars, furans, and other organic molecules in the hydrocarbon pool could lead to the coke formation on the interior surface of zeolite while Custodis et al. (2014) stated that the competing side reaction of phenol repolymerization and lignin polycondensation could cause the coke deposition.Mesoporous zeolite catalysis has been recognized as an efficient approach to attenuate the diffusion restriction of bulky biomass and plastic molecules and expand the production of aromatic hydrocarbon through the larger pore size. High surface area of mesoporous zeolite provides greater access to active sites and enhance the catalytic interaction between the co-pyrolyzed reactants, causing higher conversion rate of oxygenates to aromatic hydrocarbon. Hong et al. (2017) reported the influence of microporous and mesoporous HZSM-5 during co-pyrolysis of cellulose and polypropylene on the aromatic formation efficiency. The result showed that mesoporous HZSM-5 by ZSM-5 desilication could offer better catalytic activity than microporous HZSM-5 in term of aromatic yield. Larger pore opening obtained by desilication can enhance the diffusion of bulky intermediates to active sites of catalyst to undergo further reactions to aromatics. In addition, mesoporosity can be allocated into the zeolite core-structure via post-synthesis treatments, such as steaming and leaching with acidic or basic media (Zhu et al., 2013).MCM-41 is a type of mesoporous zeolite that has bigger pore size, making it suitable for adsorption, separation and macromolecular catalysis. Its larger pore size could ease the diffusion limitation in pores. MCM-41 could provide enough active sites for adsorption and catalytic reaction due to its high specific surface area greater than 1000\u00a0m2/g. Chi et al. (2018) conducted co-pyrolysis of cellulose and PP in the presence of MCM-41 and Al-MCM-41. The cracking of oxygenated compounds was heightened by the strong acidity originated from the inclusion of Al onto the mesoporous MCM-41. The results indicated that the production of olefins and aromatics were enhanced by using Al-MCM-41, inferring that Al-MCM-41 had superior cracking and deoxygenation effect. The aromatic formation during the co-pyrolysis was governed by internal acid sites, hydrocarbon pool, and Diels-Alder reaction (Fig. 3\n). Cellulose was decomposed earlier compared to polypropylene as it had lower decomposition temperature. Numerous oxygenated compounds and penta heterocyclic furans were produced via ring cleavage and catalytic cracking to break its hexa heterocyclic, followed by dehydration and cyclization. During the catalytic co-pyrolysis, olefin was produced from direct cracking of polypropylene via carbonium ion and \u03b2-scission and deoxygenation of oxygenated compounds at acid sites via dehydration, decarbonylation, and decarboxylation reactions. These intermediates (olefins and oxygenates) participated in deoxygenation and oligomerization to form carbocation hydrocarbon pool where the aromatic and olefins were formed. Along with hydrocarbon pool mechanism, the monocyclic aromatic hydrocarbon can be formed via Diels-Alder reaction between cellulose-derived furans and polypropylene-derived olefin.\nKim et al. (2017c) investigated the impact of acidity and molecular diameter on the formation of aromatic hydrocarbon in co-pyrolysis of carbohydrates with linear LDPE. They assessed the catalytic activity of microporous and mesoporous ZSM-5 with high mesoporosity and poor acidity Al-SBA-15. Higher yield of monoaromatic hydrocarbons was obtained under catalysis of ZSM-5 due to the combination of micropores and mesopores structure. This framework is suitable for the shape selectivity of aromatic production and to improve diffusivity of bulky molecular pyrolysis intermediates into the catalyst pore. The finding of this study indicated that catalyst with higher acidity together with an appropriate structure and pore diameter was an ideal catalyst for aromatic formation in co-pyrolysis reaction. Similar trend was found in the catalytic co-pyrolysis of yellow poplar and HDPE over three types of mesoporous catalysts, including hierarchical mesoporous MFI, hierarchical mesoporous Y, and Al-SBA-15 (Rezaei et al., 2017). Hierarchical mesoporous MFI which had large mesopores and strong acidity delivered the highest yield of olefins and aromatic hydrocarbons attributable to the efficient hydrocarbon pool mechanism. The yield of solid residue (char/coke) decreased for all three types of mesoporous catalyst. The lifespan of catalyst could be enhanced by reducing the coke deposition.Metal addition could modify the textural characteristics and acid sites, and enhance thermal stability of the catalyst. This process aids in decreasing the rate of coke growth over the catalyst and enhancing the liquid production (Botas et al., 2014; Iliopoulou et al., 2012). Razzaq et al. (2019) observed that anchoring of Ni, Co, Zn and Fe oxides onto the HZSM-5 framework by wet impregnation technique could reduce the coke yield by 50\u00a0% compared to intrinsic HZSM-5 during the co-pyrolysis of wheat straw and polystyrene. This was due to the moderate acidic strength of metal-modified zeolite which was helpful in decreasing the coke formation over the zeolite. In addition, pyrolytic oil catalysed over metal-modified zeolite contained relatively higher organic phase yield instead of aqueous phase as compared to unmodified HZSM-5. The presence of metal-modified zeolite could enhance the decarboxylation and decarbonylation while inhibiting the dehydration reaction. French and Czernik, (2010) reported that incorporation of metal sites onto the zeolite framework could alter the deoxygenation pathway so that it favourably released more oxygen in the form of carbon monoxide instead of carbon dioxide and water, thereby offering more hydrogen available for aromatic production. The presence of metals boosted the aromatic selectivity towards high value mono-aromatic hydrocarbon (MAHs) and supressed the formation of oxygenated compounds.\nKim et al. (2017b) investigated HZSM-5, mesoporous MFI, Pt/mesoporous MFI and Al-SBA-16 catalyst effect for the Laminaria Japonica and PP co-pyrolysis. Pt/mesoporous MFI showed higher aromatic yield and oxygenate removal efficiency than the other catalysts due to the strong Br\u00f6nsted acid sites and large pore size as well as catalytic effect resulted from the incorporation of Pt. Pt promoted the cracking and deoxygenation of oxygenated compounds to aromatic. The authors also highlighted that the strength of acidity played more essential role than the pore size in the production of aromatic hydrocarbon. Coupling of weak acid sites and large mesopores lowered the catalytic performance of Al-SBA-16. Conversely, mesoporous Al-SBA-15 with weak acidity showed a better oxygen removal efficiency than HZSM-5, concluding that the pore size played an important role during the cracking of large oxygenate molecules.Impregnation of phosphorous onto the zeolite framework could enhance the hydrothermal stability and anti-coking properties of zeolite and ease the transformation of alkane to olefin, which was subsequently converted to aromatic. Yao et al. (2015) found that the modification of ZSM-5 with phosphorous (P) and nickel (Ni) increased the production of valuable aromatic hydrocarbons and olefins in the catalytic fast co-pyrolysis of pine wood and LDPE due to the enhanced zeolite\u2019s Lewis acid sites which acted as electron pair acceptor and which had a high tendency to accept the hydride ions generated during the conversion of alkanes to olefins. Higher content of aromatic hydrocarbon boosts the commercial value of bio-oil as the aromatic compound is vastly used as additives in transportation fuel and feedstock materials in the petrochemical industry (Kim et al., 2017a). In addition, the rate of coke-induced catalyst deactivation, which is the main concern in catalytic fast pyrolysis, has also been reduced due to the impregnation of ZSM-5 with P and P/Ni cation. The incorporation of P and Ni onto the ZSM-5 significantly decreased the strong Bronsted acid sites of zeolite, in turn reducing the coke deposition. Gallium (Ga) altered the texture characteristic and acidity of zeolite by reducing the pore volume and surface area of zeolite (Li et al., 2015). Ga was introduced into the zeolite framework via incipient wetness impregnation. The Ga decreased the density of Br\u00f6nsted acid sites due to the replacement of some Br\u00f6nsted acid sites by Ga. Ga-containing zeolite substantially increased the production of olefin and/or monoaromatic hydrocarbons at the expense of less valuable alkane during the catalytic co-pyrolysis of pine wood and LDPE. Non-framework Ga provided a new route for dehydrogenation of alkane to olefin, which is subsequently converted to aromatic.In an effort to enhance the catalytic activity of the zeolite catalysts, the incorporation of hierarchical porosity or alteration through metals and oxide supplement has been frequently reported (Han et al., 2020; Jin et al., 2016; W. Wang et al., 2019). Although the mesoporous materials are synthesized to solve the problem of diffusion limitations, it has poor surface acidity and unstable structure property, bringing about unsatisfactory activity in acid-catalysed reactions. To solve this shortcoming, researchers have combined the advantages of microporous molecular sieve and mesoporous material, producing zeolites with hierarchical micro-mesoporous composite (Talebian-Kiakalaieh and Tarighi, 2020). It works in the way that the external mesopores capture molecules in several directions and concentrate them towards the zeolite micropores. The mesoporous structure could enhance mass transfer and cracking of large molecular pyrolysis vapours, which are hard to diffuse into the microporous zeolite (Kim et al., 2019). Furthermore, every mesopore behaves as a funnel and enables the effective penetration of molecules within the narrow one-dimensional micropore system. Such a combination of the properties of both porous systems would make the hierarchical aluminosilicates a versatile material for many applications. Five different approaches to synthesize hierarchical micro-mesoporous include recrystallization of ordered mesoporous silicas, zeolite-seeding, mesoporous carbon templating during crystallization, alkaline extraction of zeolites, and combining mesostructure and microstructure-directing agents (Enterr\u00eda et al., 2014).Several studies reported that the hierarchical zeolites could substantially resolve the limitations of the conventional zeolite, such as low mass transfer problem, deactivation of catalytic activity and low activity to bulky substrates in different chemical reactions due to significant deoxygenation and excellent aromatic selectivity (Ahmed et al., 2020; Chi et al., 2018). Moreover, Song et al. (2018) mentioned that more advantages from hierarchical zeolites could be observed, such as shortened diffusion path length, abundant external acid sites and surface area, and excellent hydrothermal stability. Combination of mesoporosity and traditional zeolites of hierarchical zeolite could broaden its application in catalysis.The catalytic activity of hierarchical zeolite is mainly dependent on the synthesis method. Desilication (removing silica) and dealumination (removing aluminum) are an efficient approach to generate mesoporosity though it may result in a considerable shift in acidic properties (Ahmed et al., 2020). The alteration of zeolite structure during desilication and dealumination of zeolite is shown in Fig. 4\n. Proper acid sites distribution and mesoporosity resulted from the alteration in acidity could benefit the reaction pathway and intermediates stabilization as reported by (Hong et al., 2017). The desilicated ZSM-5 showed superior catalytic activity in term of aromatic selectivity (33.50\u00a0wt%) compared to parent ZSM-5 during co-pyrolysis of cellulose and polypropylene. The treatment enlarged the pore for better diffusion while retaining its strong acidity. The desilication enhanced the weak acid sites, thus improving the liquid products yield. In addition, the weak acid sites also fostered the deoxygenation of furan via reaction with olefins to produce more aromatics. Hierarchical zeolite has great potential in catalytic reactions related to bulky molecules due to the presence of microporous and mesoporous structure (Lv et al., 2020). A hierarchical pore structure could be created in ZSM-5 by including larger pore structures namely mesopore linked to the core microporous framework as an endeavour to inhibit the coke deposition and attain higher transformation of bulky oxygenates (Feliczak-Guzik, 2018). The mesoporous structure could enhance mass transfer and cracking of large molecular pyrolysis vapours, which were hard to be diffused into the microporous zeolite (Kim et al., 2019). Alkaline treatment is well-known and established as a post-synthetic technique comprised of fractional desilication of the zeolite structure to create secondary mesopores in ZSM-5 with bigger pore opening and outer surface area (Li et al., 2014). Apart from alkaline treatment, re-assembly aided with organic templating agent permits restructuring and redeposition of silicate and aluminosilicate fragment into the mesoporous material while maintaining the weight and/or acidity in basic medium (Chen et al., 2018).\nLin et al. (2021) developed a series of hierarchical HZSM-5 with various alkaline solutions ranging from 0.2 to 0.4\u00a0mol/L and found that low alkaline solution (\u22640.3\u00a0mol/L) accelerated the formation of monoaromatics from 63.79\u00a0% catalysed by HZSM-5 to 71.75\u00a0% for 0.3-HZSM-5 while higher alkaline solution diminished the framework of HZSM-5, leading to reduction of aromatic production. The alkaline treatment enhanced the mesoporosity of the zeolite so that the larger intermediates including oxygenated compounds and aliphatic hydrocarbons could effortlessly access the acid sites of hierarchical zeolite to form aromatics. Furthermore, the alkaline treatment reduced the polyaromatic hydrocarbons (PAHs) formation due to shorter diffusion path distance of molecules in the hierarchical HZSM-5 zeolites, retarding the secondary polymerization reactions of mono-aromatics inside the catalyst channel. Conversely, the selectivity of aliphatic hydrocarbons and oxygenated compounds were reduced as the alkaline concentration reached 0.3\u00a0mol/L, probably due to the conversion to aromatics at the catalyst pores via a series of reactions.\nLi et al. (2020) investigated the catalytic fast co-pyrolysis of waste greenhouse plastic films and rice husk over hierarchical micro-mesoporous zeolite with HZSM-5 as core and MCM-41 as shell (HZSM-5/MCM-41). The result showed that the relative content of hydrocarbons and CO2 were higher than for the non-catalytic pyrolysis, suggesting that HZSM-5/MCM-41 promoted the conversion of pyrolyzates to aromatic and decarboxylation becoming one of the routes that governed the conversion. The addition of MCM-41 mesopore around the HZSM-5 crystal particles assisted in cracking the large-molecular weight volatile to small molecular compound (Lin et al., 2021). Zhang et al. (2018) reported that hierarchical HZSM-5/MCM-41 which contained a moderate amount of mesopore was effective for pyrolysis intermediate upgrading while reducing the coke formation simultaneously.\nQian et al. (2021) synthesized a novel hierarchical zeolite with the aid of alkaline lignin in the re-organization of alkaline treatment core material. The result revealed that the yield of bio-oil and gas was enhanced at the expense of solid residue. Higher transformation of main pyrolyzates derived from co-reactant and inhibition of char was observed due to higher acidity and hierarchical pore system of the catalyst. In addition, coke yield also decreased due to the enhanced diffusion capability of the feedstock and coke precursor, and shorter diffusion path length in the ZSM-5 structure. Deactivation rate could be reduced as no secondary reaction was produced resulting from the short residence time (Serrano et al., 2013). More particularly, the abundant reactive species of pyrolyzates from biomass-plastic mixture rapidly traverse the catalyst layer by hierarchical pore structure prior to absorption, producing solid residue. Diffusion through hierarchical zeolite crystals is faster in a manner that is closely related to Knudsen regime since the diffusion through mesoporous materials proceeds by molecule-to-molecule interaction as well as molecule-to-pore interaction.Extensive efforts are being made to develop new catalyst with low-cost, good catalytic performance and environmental friendliness. The utilization of natural ore and industrial waste as low-cost and high-activity catalyst in the production of value-added bio-oil can pave ways for recycling and reusing those mineral and waste. Red mud (RM) is a waste residue generated from aluminium industries by the Bayer process of alumina production from bauxite (Wang et al., 2019). It comprises a complex mixture of metal oxides, notably iron oxides and small amounts of alkali earth metals (Das and Mohanty, 2019). Recently, there are significant interest in making use of red mud as a catalyst in pyrolysis of biomass due to its compositional properties containing metal oxides, including CaO, TiO2, Fe2O3, Al2O3, MgO and SiO2. (Chang et al., 2020) investigated the catalytic pyrolysis of palm kernel shell over red mud using a bench scale fixed bed reactor. The result indicated that the presence of RM could enhance the cleavage of oxygen-containing double bonds and functional groups in-side chains on benzene ring to phenol and aromatic. Duman et al. (2013) studied the catalytic pyrolysis of safflower oil cake over RM in a dual reactor system and found that the RM was an effective catalyst in deoxygenation reaction, enhancing the aromatic selectivity. Although the base property of RM could provide the additional cracking efficiency, high production of aromatic could not be achieved since RM did not possess strong acid sites and shape selectivity that were able to limit the diffusivity of longer chain intermediates into the active sites and to foster the secondary reactions including isomerization and aromatization to produce aromatic hydrocarbon (Kelkar et al., 2015). Therefore, the combination of low-cost alkaline catalyst and acidic catalyst is regarded as an ideal approach to achieve higher formation of aromatic hydrocarbon and enhance the zeolite lifetime. Yathavan and Agblevor, (2013) pyrolyzed pinyon\u00a0\u2212\u00a0juniper (PJ) woody biomass over HZSM-5 and RM catalyst. The addition of RM as fractional catalyst could enhance the deoxygenation process in which the oxygen was rejected via decarboxylation (CO2) process instead of decarbonylation (CO) and dehydration (H2O) process. This process could enhance the overall carbon and hydrogen efficiency and thus, more hydrogens are available for aromatic production. Furthermore, the pyrolysis oil catalysed by RM has relatively lower viscosities than HZSM-5 catalyst. In another study, in-situ RM was used in the catalytic fast co-pyrolysis (CFCP) of organosolv lignin (OL) and polypropylene (PP) over ex-situ HZSM-5. The authors reported an increase in the cracking efficiency of OL/OP intermediates as well as an enhancement of aromatic selectivity due to the effective interaction between pyrolyzates. The presence of RM in the in-situ catalytic reactor could improve the formation of selected hydrocarbon that acted as precursor to produce aromatics over ex-situ HZSM-5 in second reactor (Ryu et al., 2020a).Coal fly ash (CFA), a by-product of coal-fired thermal power plants (TPP) is often disposed in the landfill, causing environmental and economic issues. One of the key features of CFA is that it consists of aluminosilicates, such as SiO2 and Al2O3, making it appealing as a precursor to produce zeolite-based catalysts (Supelano et al., 2020). Vichaphund et al. (2019) successfully synthesized ZSM-5 from CFA (HZSM5-FA) via consecutive alkaline fusion and hydrothermal treatment (Fig. 5\n). The zeolite crystallization time was varied at 24 hr (HZSM5-FA \u221224) and 72 hr (HZSM5-FA-72). The catalytic activity of the HZSM5-FA was determined in catalytic fast pyrolysis of Jatropha waste at the temperature of 500\u00a0\u00b0C and Jatropha-to-catalyst ratio of 1:1\u20131:10. The addition of HZSM5-FA considerably enhanced the aromatic selectivity up to 97.2\u00a0% and reduced the undesired oxygenated and N-containing compounds via deoxygenation and denitrogenation reactions. HZSM5-FA promoted the cracking of large oxygenates and nitrogenated species and further converted them to olefins and aromatic via a series of reactions, including decarbonylation, decarboxylation, dehydration, cyclization, aromatization, dehydronitration, deamination, and hydrogenation. On the other hand, HZSM5-FA-72 produced low amounts of aromatics compared to HZSM5-FA-24 due to both low acidity and high mesopore volume. Low acidity zeolite had low number of active sites (Bronsted acid sites) which were responsible to convert oxygenated compounds to aromatic compounds within the framework of zeolite catalyst while high mesopore volume could limit the molecular diffusion of pyrolyzates to inner pore of zeolite to further undergo the series of reactions for aromatic formation. Based on this result, it can be concluded that the pore structure and type of acidity play an important role for aromatic formation.Steel-slag is a waste by-product derived from steel-making process which accounts about 15\u00a0% of the total crude steel output. Most of the steel slags are accumulated heavily in landfill, becoming environmental hazards due to the leaching of heavy metals, particularly mercury (Hg), lead (Pb), chromium (Cr), cadmium (Cd), and arsenic (As) (Song et al., 2021). The higher activity of Faujasite zeolite derived from steel slag in hydrocarbon production was described by Hassan et al. (2019) in the co-pyrolysis of sugarcane bagasse and HDPE. The origin of Faujasite zeolite influenced the formation of mesopore with an average size of 45\u00a0nm and a modest surface area of 39.6\u00a0m2/g. Even though the surface area of the zeolite is quite low, the NH3-TPD measurement showed a relatively strong acidity. Promotion to the hydrocarbon pool and deoxygenation reaction takes place due to the fact that the microporous structure enables intimate contact to the strong acid sites. Due to the lack of weak acidity in the catalyst, thermal condition is thought to be responsible for the decomposition and cracking of biomass and HDPE molecules. In increasing the pyrolysis temperature, they have been able to compensate the inadequacy of weak acidity while the strong acidity contributes to the upgrading of bio-oil pyrolyzates through a succession of dehydration, decarbonylation, decarboxylation, and oligomerization reactions. Nonetheless, at the reaction temperature of 500\u00a0\u00b0C and above, reverse Diels-Alder reaction would begin to occur, hindering further upgrading of the product by favouring the generation of olefins in place of aromatics.With its distinctive pore structure and high acidity, microporous zeolite has exceptional cracking and deoxygenation abilities which favour the aromatic selectivity. However, a very small pore size (0.54\u20130.56\u00a0nm) of microporous zeolite resulted to coke formation, mass transfer limitation, and slow diffusion of large molecules into its inner active sites, preventing further reactions of macromolecules to valuable aromatic. Acidity and porosity are two paramount factors that influence the catalytic activity of zeolite catalyst. Mesoporous zeolite has been recognised as an effective method for reducing the diffusion restriction of bulky biomass and plastic molecules, and increasing aromatic hydrocarbon production due to the larger pore size. Although mesoporous materials are synthesised to address the issue of diffusion limitations, they have poor surface acidity and an unstable structure, resulting in inadequate activity in acid-catalysed reactions. To solve this challenge, introduction of new mesoporosity in the micropore of zeolite produces zeolites with hierarchical micro-mesoporous composite. Hierarchical zeolites can substantially resolve the limitations of the conventional zeolite, such as low mass transfer problem, deactivation of catalytic activity and low activity to bulky substrates in different chemical reactions due to the combination of two levels of porosity. Desilication and dealumination are the effective methods to create mesoporous structure with large pore size for better diffusion of bulky intermediates to active sites of catalyst to undergo further reactions to aromatic. In addition, the acidity could also be altered to foster the deoxygenation reaction. Incorporation of metal into the zeolite could alter the pore size and total number of acidic sites, and enhance the thermal stability of the catalyst, all of which are helpful to decrease the coke formation of zeolite catalyst. With the addition of metal, new enhanced Lewis acid sites of zeolite were generated, which boosted the aromatic production. Ubiquitous, low cost and a complex mixture of metal oxides, natural ores such as red mud could be utilized as a catalyst in pyrolysis of biomass. The base properties of red mud could provide additional cracking and deoxygenation for aromatic production. However, the result was still unsatisfactory due to the lack of strong acid sites and shape selectivity as compared to the conventional zeolite catalyst. Therefore, it is advisable to combine low-cost base catalyst with acidic catalyst to achieve higher production of aromatic. Natural mineral wastes including coal fly ash and steel slag consisting of aluminosilicates, such as SiO2 and Al2O3 could be exploited as precursors for the synthesis of zeolite. However, two paramount factors that need to be considered to ensure high production of aromatics are pore structure and type of acidity. A zeolite catalyst with high acidity with an appropriate pore structure needs to be tailored to obtain high cracking and deoxygenation efficiency for the production of aromatics. The performance of different types of catalysts in the catalytic co-pyrolysis of biomass and plastic is summarized in Table 2\n.Co-feeding hydrogen-rich plastic to the oxygen-rich biomass offers a promising technique for the production of chemicals and bio-oil. Utilization of biomass and plastic waste in co-pyrolysis process could bring a positive impact to the environment and human being since a large amount of waste polymer could be reduced and value-added fuels and chemicals could be produced. However, to be able to fully exploit this technique, further research and development are required.Co-pyrolysis of biomass and various plastic mixture needs to be considered as a feedstock in the future as the waste materials generally are not collected separately according to their criterion. According to waste management situation, the separation of biomass and plastic waste from each other during the recycling stage is not feasible and uneconomical. Many studies have focused on the co-pyrolysis of binary mixtures instead of multi-component mixtures. Therefore, multi-component feedstocks with the optimal reaction conditions needs to be investigated Furthermore, the gas emission associated with the multi-component pyrolysis needs to be examined to fully optimize the pyrolysis technology to achieve high quantity and quality of bio-oil.Although co-pyrolysis of biomass with plastic remarkably supresses the coke formation as compared to pyrolysis of individual biomass, catalyst deactivation remains a great challenge. Selecting suitable catalysts that have high catalytic activity as well as stability is of a great importance. Acidity/basicity, shape selectivity, porous structure and number of active sites are paramount factors that need to be considered when designing a catalyst. Bifunctional catalyst that possesses acid and base properties should be developed. Base catalyst promotes the fragmentation of oxygenates which can easily diffuse into the pores of acidic zeolite. The oxygenates will then be converted to aromatic hydrocarbon via cracking and deoxygenation reaction induced by acid catalyst. Furthermore, the detailed catalytic pyrolysis and catalytic co-pyrolysis reaction mechanisms of pyrolyzates on the external surface and inner pores of catalyst also need to be understood.Pre-treatment of biomass such as torrefaction and hydrothermal could be a solution to enhance the physicochemical characteristics of the biomass which could lead to the enhancement of conversion efficiency, reduction of the coke formation and improvement to the aromatic production during the catalytic pyrolysis of biomass. For example, torrefaction can enhance the cellulose content and physicochemical properties of biomass including less oxygenated compounds and high heating value to produce bio-oil with low oxygenated compounds, low acidity, high energy content and high monoaromatic hydrocarbons (Boateng and Mullen, 2013; Ryu et al., 2020). Hydrothermal treatment can produce crystalline cellulose and remove the alkali and alkaline metals, especially the K and Na metals, which provide a suitable medium for aromatic formation (Wang et al., 2021).Synergistic effect mechanism in catalytic co-pyrolysis of biomass and plastic is extremely complex reaction pathway, dominated by free radical fragments at high temperature. During catalytic co-pyrolysis, different free radicals that act as reaction intermediates are participating in hundreds of parallel or continuous reaction pathways. However, the detailed knowledge on the evolution of free radicals as reaction intermediates during catalytic co-pyrolysis is limited and unclear as it is hard to be obtained by the conventional experimental methods alone. Most researchers propose the synergistic mechanism based on the weight loss and final product obtained via TGA and GC\u2013MS instead of reaction intermediate verification. Until now, there has been no solid and unequivocal hypothesis explaining the synergistic effect mechanism involved in radical-induced catalytic co-pyrolysis of biomass and plastic. Therefore, it is important to identify the type and composition of free radicals present during catalytic co-pyrolysis of biomass and plastic.The authors thankfully acknowledge the support obtained from Lotte Chemical Titan (M) Sdn. Bhd. and Universiti Sains Malaysia (Grant No: 304/PJKIMIA/6050422/L128), in the form of research grant and facilities which brought forth this article. The first and second authors also acknowledge the research grant provided by Universiti Teknologi MARA, under Research Incentive Grant (Grant No: 600-RMC/GIP 5/3 (045/2021)) that has resulted in this article.The authors confirm that the data supporting the findings of this study are available within the article. Raw data that support the findings of this study are available from the corresponding author, upon reasonable request.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.", "descript": "\n                  The increasing global fuel consumption and growing environmental concerns are the impetuses to explore alternative energy that is clean and renewable for fuel production. Converting biomass and plastic waste into high-value fuel and chemicals via co-pyrolysis technique may provide a sustainable remediation to this problem. This review critically discussed the influence of various types of plastic wastes as co-reactant in co-pyrolysis with biomass on the product distribution, synergistic effect, and quality of bio-oil. The outcome of this review revealed that the addition of plastic enhanced the yield and quality of bio-oil and inhibited the production of oxygenated compound and coke formation. Next, the critical role of zeolite-based catalyst (microporous, mesoporous, hierarchical, and metal modified zeolite) and low-cost mineral-based catalyst in upgrading the yield and quality of liquid fuel were compared and discussed. The characteristic, synthesis method, strength, and limitation of each catalyst in upgrading the products were summarized. Hierarchical zeolites can resolve the problems of mass transfer, and diffusion limitation of large molecules into active sites associated with conventional zeolite due to the combination of two levels of porosity. Finally, the potential challenges and future directions for this technique were also suggested.\n               "}
{"full_text": "Data will be made available on request.Conversion of renewable resources to biofuels and chemicals is becoming an increasingly attractive option given the depletion of fossil fuels and the environmental problems caused by the excessive use of unsustainable energy resources. Lignin is typically regarded as one of the significant sustainable biomass sources for the production of fuels and chemicals, which continues to drive human development and utilization of them from huge abundance of agriculture and forest waste [1]. In general, lignin is mainly a three-dimensional network polymer composed of the following three phenylpropane monomers: guaiacyl, p-hydroxyphenyl and syringyl. Phenolic compounds such as guaiacol, phenol and 2,6-dimethoxyphenol can be directly obtained from the fast pyrolysis of lignin [2], but their high oxygen content, poor chemical stability, corrosivity and immiscibility with hydrocarbon fuels make them unsuitable for direct use as engine fuels [3]. Therefore, for better uses of phenolic compounds derived from lignin, the process of hydrodeoxygenation (HDO) is needed.The current literatures are keen to convert the lignin model compound guaiacol into various high-valued chemicals such as cyclohexanol, cyclohexane, benzene, 2-methoxycyclohexanol, phenol, catechol, etc. [4]. Among them, cyclohexanol is a versatile petrochemical and high-quality oxygen-containing fuel, which also can be widely used in pesticides, medicines, cosmetics and other fields. Unfortunately, the selective HDO of guaiacol to cyclohexanol is difficult due to involving the hydrogenation of the CC unsaturated bond of aromatic ring, the cleavage of the CAR-OCH3 bond, while retaining the CAR-OH [5]. According to the dissociation energy of the three bonds (CAR-OH\u00a0>\u00a0CAR-OCH3\u00a0>\u00a0CARO-CH3), the CARO-CH3 bond of guaiacol is more easily dissociated (demethylation), resulting in catechol [6]. There are two pathways from the HDO of guaiacol to form cyclohexanol (Fig. S1) [7]. The 2-methoxycyclohexanol generated by Route II limits the cleavage of the CAR-OCH3 bond due to steric hindrance and electronic effects. And Route II needs to be carried out at higher temperature and H2 pressure, which is prone to generate cyclohexane from cyclohexanol. Route I, by comparison, can be conducted at lower temperature, which meets the needs of economic development and is the better way to synthesize cyclohexanol.The key factor for guaiacol HDO to cyclohexanol is to design an efficient reaction system conducive to the demethoxylation of guaiacol. The product selectivity and total activity in HDO reactions are greatly influenced by the properties of solvents used. Water is the greenest and the mostly being used solvent for guaiacol HDO, which is reported to be effective at promoting CO cleavage and reducing the undesirable thermal degradation [8].The more important point is to develop effective catalyst for guaiacol HDO to cyclohexanol. A series of metal catalysts have been developed and investigated for the HDO of guaiacol to cyclohexanol, including metal hybrids (sulfides, carbides, nitrides, etc.), non-noble metals and noble metals [9]. Metal hybrids exhibited satisfactory catalytic activity for the HDO reaction, but their reusability was not good enough. Non-noble metals such as Ni and Co usually show low activity. To obtain a high yield of cyclohexanol, a high metal loading, high temperature (> 200\u00a0\u00b0C), and high H2 pressure are needed. Noble metal catalysts Ru, Pd and Pt, especially Ru-based catalysts showed great potential in HDO reactions. The combination of Ru metal sites and a base such as MgO or MnOx could achieve high cyclohexanol yield (79% and 81%, respectively) owing to the presence of the base, which suppresses the unselective CO dissociation by Ru catalyst, and may also promote the demethoxylation step via stabilizing the produced phenol [10,11]. Although Ru-based catalyst with base sites gave a\u00a0\u223c\u00a080% yield of cyclohexanol for guaiacol HDO, a great amount of 2-methoxycyclohexanol (> 13%) was generated, which inhibits the increase in cyclohexanol selectivity. Highly selective HDO of lignin-derived phenols to cyclohexanol is still challenging. Besides base sites, suitable acid sites also facilitate cleavage of the CO bond [4]. Based on the above reported results, we postulate that acidic carrier supported Ru-based multifunctional catalyst with metal sites and acid/base sites will be more beneficial to the demethoxylation step in the selective HDO of lignin-derived phenols, thereby increasing the selectivity to cyclohexanol. As we know, \u03b3-Al2O3 is a common solid acid carrier, however, it is often accompanied by structural changes in water with significantly decreased acidity and surface area that trigger catalyst deactivation [12]. The introduction of SiO2 can effectively inhibit the hydration of \u03b3-Al2O3 support and prevent the active decay of \u03b3-Al2O3 supported catalyst. Al2O3-SiO2 composite is a carrier with excellent performance owing to its developed pore structure, large specific surface area and strong thermal stability, and its surface acidity is easy to control [13]. The structure of Al2O3-SiO2 supported catalyst affects its separation from the reaction solution. Microspheres are easy to be separated from the aqueous phase. In addition, the Al2O3-SiO2 composite microspheres with high pore volume and specific surface area are conducive to the uniform dispersion of the noble metal [14].Based on the above considerations, herein, we designed an efficient and reusable Al2O3-SiO2 acidic uniform microspheres supported RuMn multifunctional microsphere catalyst for the highly selective HDO of guaiacol and 2,6-dimethoxyphenol to cyclohexanol in water. The choice of metal Mn as the second metal component is mainly based on the following considerations. Besides base sites, MnOx have many oxygen vacancies, which facilitates the adsorption of oxygen-containing functional groups, being conducive to the deoxidation step in the hydrodeoxygenation reaction process [15]. In our recent work, we prepared MnOx modified Ni/AC catalyst and found that the catalyst exhibited excellent performance for the HDO of 5-hydroxymethylfurfural to 2,5-dimethylfuran, owing to the synergistic effect of Ni and MnOx [16]. In this work, it was observed that, compared with the reported Ru-based catalysts, our RuMn/Al2O3-SiO2 catalyst showed higher catalytic activity and selectivity for the partial HDO of guaiacol and 2,6-dimethoxyphenol to cyclohexanol under low Ru amount and mild reaction conditions. And it exhibited excellent performance for the hydrogenation of phenol to cyclohexanol. The catalyst was easy to be separated from the aqueous solution and showed good reusability. The catalyst was well characterized by different techniques and the synergistic effect among Ru, Mn and Al2O3-SiO2 was investigated. Furthermore, possible reaction pathways of guaiacol over RuMn/Al2O3-SiO2 catalyst was proposed based on the product distribution.Al2O3-SiO2 microspheres were prepared as follows. 1\u00a0g of \u03b3-Al2O3 was added to 80\u00a0mL of isopropanol solvent. The mixture was stirred for 30\u00a0min to ensure that \u03b3-Al2O3 was uniformly dispersed. Then, 8\u00a0mL of distilled water, 5\u00a0mL of ammonia water and 1\u00a0mL of tetraethylorthosilicate (TEOS) were added into the above solution in sequence, and the mixture was stirred at room temperature for 10\u00a0min and kept at room temperature for 8\u00a0h. The white precipitate obtained was subsequently filtered. The obtained solid was washed with distilled water for 5 times and then centrifuged. The resulted sample was dried in an oven at 110\u00a0\u00b0C for 12\u00a0h, followed by calcination in a muffle furnace at 500\u00a0\u00b0C for 3\u00a0h. The prepared Al2O3-SiO2 sample (Si/Al\u00a0=\u00a00.2) was ground into flour and passed through a 120-mesh sieve for use.A series of RuMn catalysts supported on Al2O3-SiO2 microspheres obtained by the above-described method was prepared by an incipient wetness impregnation method. In a typical procedure, an appropriate amount of Al2O3-SiO2 was added into the aqueous solution containing the required amount of RuCl3\u00b73H2O and C4H6MnO4\u22194H2O in a beaker. After impregnated for 24\u00a0h, the mixture was dried at 110\u00a0\u00b0C for 10\u00a0h and finally reduced at 300\u00a0\u00b0C in a tubular furnace under hydrogen flow for 3\u00a0h to obtain the target catalyst, which was denoted as RuMn(x:y)/Al2O3-SiO2, where x:y means the molar ratio of Ru to Mn. The molar ratio of Ru to Mn was varied from (4:1) to (1:2) by changing the Mn content and using a fixed amount of Ru (3.0\u00a0wt%). Two monometallic catalysts 3.0\u00a0wt% Ru/Al2O3-SiO2 and 3.0\u00a0wt% Mn/Al2O3-SiO2, and active carbon (AC), SiO2 and \u03b3-Al2O3 supported Ru or RuMn catalysts were prepared by using the same method for comparison. The details of the chemical reagents and the catalyst preparation including information on the content of Ru and Mn in the corresponding catalyst are described in the Supporting Information (Table S1).The detailed information of characterization methods including X-ray diffraction (XRD), Brunauer-Emmett-Teller (BET), transmission electron microscopy (TEM), high-angle annular dark-field scanning transmission electron microscope (HAADF-STEM), H2 temperature-programmed reduction (H2-TPR), X-ray photoelectron spectroscopy (XPS), temperature-programmed desorption of CO2 (CO2-TPD), ammonia temperature-programmed desorption (NH3-TPD), pyridine adsorbed IR spectroscopy (Py-IR) are described in the Supporting Information.The catalytic hydrodeoxygenation of phenols was performed in a 25\u00a0mL stainless-steel autoclave equipped with magnetic stirring. In a typical experiment, 30\u00a0mg of catalyst, 500\u00a0mg of guaiacol and 15\u00a0mL of solvent water was added into the reactor. The reactor was sealed, purged with H2 for five times, and pressurized with H2 to the required pressure. The autoclave was then heated to the required temperature and kept at this temperature for the required time under the continuous stirring speed of 1000\u00a0rpm. After the reaction, the autoclave was quickly cooled to room temperature, the gases were collected in a gas bag and the reaction products were separated from the catalyst by centrifugation. The liquid reaction products were extracted by ethyl acetate, and then quantitatively analyzed with an Agilent 7890A gas chromatography equipped with an HP-5 capillary column (30.0\u00a0m\u00a0\u00d7\u00a00.32\u00a0mm\u00a0\u00d7\u00a00.25\u00a0\u03bcm) and a flame ionization detector (FID) using n-dodecane as an internal standard, and identified by an Agilent 6890 GC system coupled to a mass spectrometer equipped with an Agilent 5973 quadrupole mass analyzer. Conversion of guaiacol, selectivities and yields were calculated as follows (Eqs. (1)\u2013(3)), based on the number of C6 rings in the substrate and products other than C1 products (methanol and methane). Methanol was also a major product, but because of its high volatility, its concentration is difficult to be accurately measured. In addition, the formation of methanol does not affect the quantification of products containing C6 rings. The gaseous product collected by a gas bag was analyzed using Agilent 7890A gas chromatography with FID and no methane was detected. Carbon balance\u00a0=\u00a0(moles of C6 ring in liquid products and unreacted reactant/initial moles of C6 ring in the reactant)\u00a0\u00d7\u00a0100%, and all the carbon balance were higher than 95%, so the experimental results were reliable.\n\n(1)\n\nConversion\n\n\n%\n\n=\n\n\n\ninitial conc\n.\nof guaiacol\n\u2212\nfinal conc\n.\nof guaiacol\n\n\n\ninitial conc\n.\nof guaiacol\n\n\n\u00d7\n100\n\n\n\n\n\n(2)\n\nSelectivity\n\n\n%\n\n=\n\nmoles of desired product formed\nmoles of guaiacol consumed\n\n\u00d7\n100\n\n\n\n\n\n(3)\n\nYield\n\n\n%\n\n=\nConversion\n\u00d7\nSelectivity\n\u00d7\n100\n\n\n\nThe guaiacol conversion and cyclohexanol selectivity over various carries supported Ru and RuMn catalysts were compared, and the results are listed in Table 1\n. The Ru loading amount was fixed at 3.0\u00a0wt%, and the HDO activity was measured at reaction conditions of 30\u00a0mg catalyst, 180\u00a0\u00b0C and 2\u00a0MPa H2 for 0.5 and 4\u00a0h, respectively. Methanol was also a major product for guaiacol HDO, but because of its high volatility, its concentration was difficult to be accurately measured. Therefore, we did not show the data of methanol in Table 1. Both hydrodeoxygenation and hydrogenation reactions proceeded, and the main products were cyclohexanol and 2-methoxycyclohexanol over all the Ru-based catalysts, which indicated that Ru is the active sites for guaiacol HDO. Phenol, which is a precursor of cyclohexanol, was detected in shorter reaction time (0.5\u00a0h). It can be seen that although guaiacol could be completely converted over AC, SiO2 and \u03b3-Al2O3 supported Ru catalysts for 4\u00a0h, only 66.4%, 65.4%, 74.4% and 68.6% of cyclohexanol yields were obtained, respectively, and a large amount of a saturated product of 2-methoxycyclohexanol was detected (entries 1\u20134). It was worth noting that Al2O3-SiO2 composite microspheres supported Ru catalyst gave higher selectivity to cyclohexanol (86.3%) and lower selectivity to 2-methoxycyclohexanol than Ru/AC, Ru/SiO2 and Ru/\u03b3-Al2O3 under the same reaction conditions (entry 5). Monometallic Mn/Al2O3-SiO2 showed much lower catalytic activity for the HDO of guaiacol (entry 6), indicating that Mn species are inactive for guaiacol HDO. The addition of Mn obviously improved the activity and the selectivity of Ru/Al2O3-SiO2 catalysts for the selective HDO of guaiacol to cyclohexanol and decreased the selectivity to 2-methoxycyclohexanol, which was consistent with the previous report that the additive MnOx in the catalyst could accelerate the dissociation of CO [17]. The conversion of guaiacol and the selectivity to cyclohexanol increased with the decrease in Ru:Mn molar ratio from 4:1 to 2:1 (entries 7 and 8), and the highest cyclohexanol yield of 91.3% was achieved over RuMn(2:1)/Al2O3-SiO2 catalyst after reaction 4\u00a0h, much higher than those of AC, SiO2 and /\u03b3-Al2O3 supported RuMn(2:1) catalysts (entries 11\u201313), and much higher than those of the reported Ru/C\u00a0+\u00a0MgO and Ru-MnOx/C catalytic systems (entries 14 and 15) [10,11]. Further increasing the Mn content, however, the activity and the selectivity of RuMn/Al2O3-SiO2 catalyst to cyclohexanol decreased, and thus cyclohexanol yield decreased (entries 9 and 10). Therefore, RuMn(2:1)/Al2O3-SiO2 catalyst was chosen for further investigation.The HDO of guaiacol was carried out at varied initial hydrogen pressure from 1 to 4\u00a0MPa at 180\u00a0\u00b0C to investigate the catalytic performance of the RuMn(2:1)/Al2O3-SiO2 catalyst. As shown in Fig. S2, the initial hydrogen pressure had obvious effect on both guaiacol conversion and product distribution. Under lower initial hydrogen pressure (< 2\u00a0MPa), cyclohexanone was detected, guaiacol conversion and cyclohexanol selectivity increased with the increase in initial hydrogen pressure and a maximum cyclohexanol selectivity of 91.3% with a guaiacol conversion of 100% was achieved when the reaction was conducted at an initial hydrogen pressure of 2\u00a0MPa. These results were in good agreement with Tomishige's study, which also reported that lower initial hydrogen pressure was beneficial to higher cyclohexanol selectivity over Ru/C\u00a0+\u00a0MgO catalyst system [10]. It was because that H2 was involved in the hydrogenation of the benzene ring and the cleavage of the CO bond in guaiacol. And the increase in hydrogen pressure meant the increase in the concentration of dissolved hydrogen in the solution and the hydrogen atoms adsorbed on the catalyst surface, thereby promoting the hydrogenation of the benzene ring and the cleavage of the CO bond in guaiacol [4]. While higher initial hydrogen pressure suppressed the formation of cyclohexanol and enhanced the formation of the by-product 2-methoxycyclohexanol, which was consistent with the literature report [18]. Finally, 2\u00a0MPa was selected as the optimal reaction pressure considering guaiacol conversion and cyclohexanol selectivity based on the above results.The effect of reaction temperature on the HDO of guaiacol over RuMn(2:1)/Al2O3-SiO2 catalyst was investigated at 2\u00a0MPa H2 for 4\u00a0h and the results are shown in Fig. S3. The conversion of guaiacol reached 100% at all the tested temperatures from 120 to 220\u00a0\u00b0C, while the product distribution was different. It was found that at a low temperature of 120\u00a0\u00b0C, the HDO of guaiacol gave 2-methoxycyclohexanol as a dominating product, indicating the occurrence of hydrogenation of the aromatic ring of guaiacol. Our experimental results well confirmed the report, in which low reaction temperature was beneficial to the hydrogenation of aromatic rings in the HDO of guaiacol over supported noble metal catalysts [19]. With the increase in reaction temperature, the selectivity to cyclohexanol increased gradually and arrived at its maximum (91.3%) at 180\u00a0\u00b0C, which indicated that the HDO of guaiacol to cyclohexanol might happen at relatively higher reaction temperature, being in good accordance with the work of Zhou et al. [20]. A continuous increase in the reaction temperature led to a decrease in the selectivity to cyclohexanol and an increase in the selectivity to cyclohexane. This indicated that excessive reaction temperature (> 200\u00a0\u00b0C) could promote the cleavage of the CO bond on the aliphatic ring of cyclohexanol, resulting in the formation of excessive hydrogenolysis product cyclohexane [18]. Meanwhile, intermediate cyclohexanone and unknown substances appeared and increased with the increase in reaction temperature, in good accordance with the observation by Wang et al. [21]. Based on the above results, 180\u00a0\u00b0C was finally chosen as the optimum reaction temperature considering both guaiacol conversion and cyclohexanol selectivity.To investigate the effect of catalyst dosage for the HDO of guaiacol to cyclohexanol, the time course experiments within 24\u00a0h were carried out over RuMn(2:1)/Al2O3-SiO2 with a dosage range of 10 to 70\u00a0mg (0.07 to 0.52\u00a0mol%), and the results are shown in Fig. 1\n. It can be seen that RuMn(2:1)/Al2O3-SiO2 showed good performance for guaiacol HDO even at a dosage as low as 10\u00a0mg (Fig. 1a), under which the conversion of guaiacol could reach 100% and the yield of cyclohexanol was up to 82.1% after 2\u00a0h. Prolonging the reaction time, the yield of cyclohexanol didn't obviously increase due to the forming of a large amount of saturated product of 2-methoxycyclohexanol, which was difficult to be hydrogenated to cyclohexanol. Contrary to the literature [22], under our catalytic conditions, cyclohexanol was not further dehydroxylated to cyclohexane, the over\u2011hydrogenated product. This observation is consistent with our assumption that Al2O3-SiO2 support is helpful to promote the demethoxylation and inhibit the hydrogenolysis of cyclohexanol to cyclohexane. The yield of cyclohexanol was significantly improved with the increase of catalyst dosage. With the catalyst dosage of 70\u00a0mg (Fig. 1d), the cyclohexanol yield was up to 96.8% after 4\u00a0h, and the highest yield of cyclohexanol could reach 99.9% after 24\u00a0h. Besides, RuMn(2:1)/Al2O3-SiO2 exhibited a higher TOF value (861.1\u00a0h\u22121) than that of many catalysts (Table S2).RuMn(2:1)/Al2O3-SiO2 was further applied to other two lignin-related monomers, phenol and 2,6-dimethoxyphenol, and the results are shown in Fig. S4 and Fig. S5, respectively. It was found that phenol could be converted to cyclohexanol with 100% yield at a catalyst dosage of 10\u00a0mg for 4\u00a0h (Fig. S4), which is the highest yield under the lowest dosage reported. When the dosage of the catalyst increased to 50\u00a0mg, 100% cyclohexanol yield was achieved after 40\u00a0min. Zhan et al. [23] proposed that the introduction of Lewis acid sites favored the activation of the aromatic ring and promoted the hydrogenation of phenol, thereby improving the selectivity to cyclohexanol. As shown in Fig. S5, 2,6-dimethoxyphenol could also be well converted to cyclohexanol (a yield of 81.8% at the catalyst dosage of 50\u00a0mg and 94.5% at the catalyst dosage of 100\u00a0mg). These results indicated that RuMn(2:1)/Al2O3-SiO2 catalyst was efficient for CAR-OCH3 cleavage and benzene ring hydrogenation while protecting the CAR-OH.Reusability is one of the important aspects for the practical application of the heterogeneous catalyst. Subsequently, we investigated the stability of RuMn(2:1)/Al2O3-SiO2 catalyst for the HDO of guaiacol to cyclohexanol for 1\u00a0h and 4\u00a0h, respectively. During each cycle, after a complete reaction at 180\u00a0\u00b0C and 2.0\u00a0MPa H2, the catalyst was centrifuged, washed with water for five times, and reused for the next runs. As shown in Fig. S6, the RuMn(2:1)/Al2O3-SiO2 catalyst kept its good performance for the HDO of guaiacol to cyclohexanol during recycling. The conversion of guaiacol and the yield of cyclohexanol decreased slightly when the catalyst was recycled for four runs in 1\u00a0h (Fig. S6a). And guaiacol was still completely converted and the yield of cyclohexanol was 90.1% after four runs in 4\u00a0h (Fig. S6b). XRD patterns showed that there was no obvious change in the structure of the recovered catalyst (Fig. S7). These data indicated that the RuMn/Al2O3-SiO2 catalyst was essentially reusable in the aqueous phase HDO of guaiacol.Fig. S8 illustrates the XRD patterns of the support Al2O3-SiO2 and the reduced catalysts Mn/Al2O3-SiO2, Ru/Al2O3-SiO2, and RuMn(2:1)/Al2O3-SiO2. The broad peak observed at 23\u00b0 for all the samples is assigned to amorphous SiO2 [24]. The XRD peaks that are observed at 2\u03b8\u00a0=\u00a032.4\u00b0, 37.2\u00b0, 39.5\u00b0, 45.9\u00b0, 61.1\u00b0 and 67.0\u00b0 are assigned to \u03b3-Al2O3 belonging to these hkl values (220), (311), (222), (400), (500) and (440), respectively [25]. In the cases of Mn/Al2O3-SiO2 and RuMn(2:1)/Al2O3-SiO2, no reflections due to Mn species were observed. And the XRD patterns of both Ru/Al2O3-SiO2 and RuMn(2:1)/Al2O3-SiO2 catalysts showed no Ru species reflections. These are likely due to the low metal loadings or the well-dispersion of Ru and Mn species. It can be observed that the XRD pattern of Al2O3-SiO2 was not changed after the addition of metals, indicating that the addition of Ru and Mn species did not compromise the structure of the support.Nitrogen adsorption-desorption isotherms and pore size distribution of Al2O3-SiO2 support and RuMn(2:1)/Al2O3-SiO2 catalyst are presented in Fig. S9, and the relative textural properties are listed in Table S3. Both the two samples exhibited type IV isotherms, a typical characteristic for mesoporous materials [26], which is attributed to the HDO of guaiacol [4]. A broad pore size distribution was observed in Al2O3-SiO2 and RuMn(2:1)/Al2O3-SiO2, consisting mainly of 20\u00a0nm pores, which exhibited mesoporous properties (2\u201350\u00a0nm) [27]. It can be seen from Table S3 that both Al2O3-SiO2 and RuMn(2:1)/Al2O3-SiO2 have high specific surface areas. Compared with Al2O3-SiO2, the specific surface area of RuMn(2:1)/Al2O3-SiO2 reduced from 130.8 to 117.5\u00a0m2\u00b7g\u22121 due to the addition of Ru and Mn species, and the average pore size and the total pore volume also slightly reduced, which may be due to the deformation of the mesoporous channels caused by the doping of Ru and Mn atoms in the support framework [28].The reducibility of the dried catalysts was investigated by hydrogen temperature-programmed reduction. Fig. 2\n shows the H2-TPR profiles of Mn/Al2O3-SiO2, Ru/Al2O3-SiO2 and RuMn(2:1)/Al2O3-SiO2 catalysts. Mn/Al2O3-SiO2 catalyst exhibited two broad peaks at 281 and 436\u00a0\u00b0C, which likely belonged to the reduction of MnO2 to Mn3O4 and Mn3O4 to MnO, respectively. And it was difficult for MnOx to generate metallic Mn in a hydrogen atmosphere at a temperature lower than 900\u00a0\u00b0C [29]. The Ru/Al2O3-SiO2 profile showed a strong reduction peak at 137\u00a0\u00b0C and a weaker reduction peak at 318\u00a0\u00b0C, which were attributed to the reduction of RuCl3 and RuO2 to Ru0, respectively [30]. In the case of RuMn(2:1)/Al2O3-SiO2, two peaks were observed at 176 and 316\u00a0\u00b0C, which were ascribed to the reduction of Ru species to Ru0. In comparison with the monometallic catalyst Ru/Al2O3-SiO2, the peak at low temperature moved towards high temperature and the peak intensity at high temperature was stronger, indicating a clear interaction between Ru and Mn species, and the addition of Mn promotes the reduction of RuO2. In addition, the peak intensity of RuMn(2:1)/Al2O3-SiO2 at higher temperature was stronger than that of Ru/Al2O3-SiO2, suggesting that Mn species are simultaneously reduced to some extent. Ishikawa et al. also evidenced the reduction of Mn species to MnO in Ru-MnOx/C catalyst [11].The morphologies of the catalysts were characterized by TEM (Fig. 3\n). Fig. 3a clearly showed that Al2O3-SiO2 support had a uniform spherical morphology with an average diameter of 0.36\u00a0\u03bcm. After immersing Ru and Mn, the morphology of the carrier did not change significantly (Fig. 3b, c and d). For Ru/Al2O3-SiO2 and RuMn(2:1)/Al2O3-SiO2 catalysts, the diameters of about 100 metal particles were randomly selected to obtain the corresponding particle size distributions. According to the measurement, an average Ru metal nanoparticle size of 2.26\u00a0nm is obtained on the Ru/Al2O3-SiO2 catalyst (Fig. 3c), and the average metal nanoparticle size on RuMn(2:1)/Al2O3-SiO2 is 1.25\u00a0nm (Fig. 3d), which indicates that the addition of Mn improved the dispersion of metal Ru. Fig. S10 showed the HAADF-STEM elemental mapping of RuMn(2:1)/Al2O3-SiO2. It can be seen that the RuMn(2:1)/Al2O3-SiO2 micro-area contained Al, Si, O, Mn and Ru. The signals from both Mn and Ru metals were clearly detected along the particles, indicating that the MnOx species were present nearby Ru nanoparticles.XPS technique was performed to investigate the chemical state and surface composition of RuMn/Al2O3-SiO2 catalyst. The results are shown in Fig. 4\n and XPS relative quantitative analysis are listed in Table S4. The survey spectrum (Fig. 4a) indicated the presence of elements Si, Al, O, C, Ru and Mn. It was generally known that the Ru 3d spectrum was not clear because it was often obscured by the strong C 1\u00a0s signal [31], as shown in Fig. 4b. The Ru 3p XPS spectrum was usually used to characterize the chemical states of Ru particles, and the test was performed between 450 and 500\u00a0eV (Fig. 4c). It can be seen that the Ru 3p spin split into pairs of Ru 3p3/2 and Ru 3p1/2. And the peaks at 462.2 and 465.8\u00a0eV in the Ru 3p3/2 were assigned to Ru0 and RuO2. The higher energy (484.3 and 486.5\u00a0eV) peaks in the Ru 3p1/2 were also contributed by Ru0 and RuO2, respectively [5]. The observation of RuO2 indicates that the oxidation of metallic Ru occurred during its exposure to air. The Mn 2p3/2 spectra of RuMn catalysts (Fig. 4d) showed peaks at 642.2 and 646.1\u00a0eV, which were assigned to Mn2+ and Mn4+, respectively. And the peaks at 484.6 and 487.0\u00a0eV in the Mn 2p1/2 were also contributed by Mn2+ and Mn4+, respectively, indicating that Mn mainly exists in the form of MnOx: MnO and MnO2. Compared to their monometallic counterparts, the binding energies of the Ru species in RuMn/Al2O3-SiO2 catalyst ascend by ca. 0.3\u00a0eV, while the binding energies of Mn species descend by ca. 0.3\u00a0eV.CO2-TPD of Ru/Al2O3-SiO2 and RuMn(2:1)/Al2O3-SiO2 were measured to investigate the basic character of RuMn(2:1)/Al2O3-SiO2 and the results are shown in Fig. S11. Compared with Ru/Al2O3-SiO2, the CO2 desorption peak of RuMn(2:1)/Al2O3-SiO2 at 414\u00a0\u00b0C became stronger, and a new peak at 516\u00a0\u00b0C, indicating the presence of moderate and strong base sites, which was attributed to the addition of Mn oxides on the catalyst surface [32].It has been proposed that the acid sites on catalyst surface play a crucial role in HDO reactions. Thus, NH3-TPD profiles of Ru/Al2O3-SiO2 and RuMn(2:1)/Al2O3-SiO2 were measured to investigated acidity, as shown in Fig. 5a. It could be observed that both the catalysts had acidity, and the acid site distribution of RuMn(2:1)/Al2O3-SiO2 was wider, indicating that the addition of Mn oxide can broad the acid site distribution. Two peaks at 128\u00a0\u00b0C and 245\u00a0\u00b0C were assigned to weak and moderate acid sites, respectively [33]. Typically, the peak at low temperature (< 300\u00a0\u00b0C) was attributed to ammonia coordinated to Lewis acid sites. The Lewis acidic active sites were favorable for CO bond cleavage, which was one of the reasons for the better catalytic performance of RuMn(2:1)/Al2O3-SiO2 catalysts.We further investigated the acid species and their distribution in the RuMn(2:1)/Al2O3-SiO2 catalyst by FT-IR spectra of adsorbed pyridine at 50, 150 and 250\u00b0C, respectively, and the FT-IR spectra of the samples were recorded at 1400\u20131700\u00a0cm\u22121. As shown in Fig. 5b, the bands at 1606, 1575 and 1444\u00a0cm\u22121 are assigned to Lewis acid, the band at 1643\u00a0cm\u22121 is assigned to Br\u00f8nsted acid, and the band at 1492\u00a0cm\u22121 is assigned to the common absorption peak of Bronsted and Lewis acids [34]. The acid content of the samples was calculated, and the results are listed in Table S5. The Lewis acid content in the RuMn(2:1)/Al2O3-SiO2 sample was significantly higher than the Br\u00f8nsted acid content. The Lewis acid sites detected at 150 and 250\u00b0C were attributed to moderate acid sites, which is consistent with the above NH3-TPD results. Combined with the performance of the above RuMn catalysts, we speculate that the moderately strong Lewis acid sites are the key to the selective demethoxylation of guaiacol.According to the first step of guaiacol HDO, there were four possible pathways, as shown in Fig. S12. (1) Guaiacol was first demethylated to catechol, then hydrogenated to 1,2-cyclohexanediol, and finally hydrogenolyzed to cyclohexanol (Eq. (1)). In this work, no catechol and 1,2-cyclohexanediol were observed during guaiacol HDO to cyclohexanol. In addition, the literature pointed out that the binding ability of catechol or 1,2-cyclohexanediol to the catalyst was stronger than that of guaiacol, which would hinder the reaction of guaiacol [35]. Therefore, the route (1) was unreasonable. (2) Guaiacol was first dehydroxylated to anisole, then hydrogenated to methoxycyclohexane, and finally demethylated to cyclohexanol (Eq. (2)). During the guaiacol HDO reaction over RuMn/Al2O3-SiO2, we detected no anisole and only a trace of methoxycyclohexane. Over time, methoxycyclohexane was not further converted to cyclohexanol, thus ruling out the route (2). (3) Guaiacol was first demethoxylated to phenol, and then phenol was hydrogenated to cyclohexanone and cyclohexanol in turn (Eq. (3)). Considering that a certain amount of phenol was detected in guaiacol hydrodeoxygenation at short reaction time (Fig. 1), phenol was an intermediate of cyclohexanol from guaiacol. Furthermore, under our reaction conditions, phenol can be converted to cyclohexanone and cyclohexanol (Fig. S4). Therefore, it can be inferred that guaiacol mainly converts cyclohexanol through route (3) over RuMn/Al2O3-SiO2 catalyst. And (4) guaiacol was first hydrogenated to 2-methoxycyclohexanol, and then demethoxylated to cyclohexanol (eq. 4). 2-Methoxycyclohexanol was the main by-product in the reaction of guaiacol HDO to cyclohexanol. When guaiacol was completely converted, 2-methoxycyclohexanol gradually decreased with reaction time, and the yield of cyclohexanol increased slowly. In addition, we investigated the reaction time curves of HDO of 2-methoxycyclohexanol and cyclohexanol under the optimized conditions (2\u00a0MPa H2, 180\u00a0\u00b0C), respectively, as shown in Fig. S13 and Fig. S14. The results in Fig. S13 showed that 2-methoxycyclohexanol was mainly converted to cyclohexanol. However, the conversion rate was too slow, and the yield of cyclohexanol was only 8.06% after 24\u00a0h of reaction. Cyclohexanol was hardly converted, and the conversion was only 0.27% after 24\u00a0h under the reaction conditions (Fig. S14). Therefore, a small part of cyclohexanol can be obtained through route (4), but this route was not the main route for the guaiacol HDO to cyclohexanol over RuMn/Al2O3-SiO2.According to the above analysis, we proposed the reaction pathways of guaiacol HDO to cyclohexanol over RuMn/Al2O3-SiO2, as shown in Scheme 1\n. In this mechanism, the demethoxylation of guaiacol to phenol and the hydrogenation of guaiacol to 2-methoxycyclohexanol can simultaneously proceed, and the relative rate of the demethoxylation (step (i)) to hydrogenation (step (ii)) of guaiacol critically determines the selectivity to cyclohexanol. Cyclohexanol was produced rapidly with a high yield from phenol, while the reaction rate from 2-methoxycyclohexanol to cyclohexanol was slow. This reaction mechanism is consistent with that proposed by Tomishige et al. [11]. The combination of appropriate acidity of the support Al2O3-SiO2 and the addition of Mn was beneficial to promote the demethoxylation of guaiacol to phenol, thereby improving the yield of cyclohexanol.Al2O3-SiO2 composite microspheres were prepared. Using it as support, RuMn multifunctional catalysts with metal active sites and acid/base sites were successfully prepared by a wetness impregnation method and applied to the aqueous phase selective HDO of lignin-derived guaiacol and 2,6-dimethoxyphenol and the hydrogenation of phenol to form cyclohexanol. Under optimized mild reaction conditions of 0.52\u00a0mol% Ru, 180\u00a0\u00b0C, 2.0\u00a0MPa H2 for 4\u00a0h, a cyclohexanol yield of 96.8% was achieved from guaiacol, and the yield of cyclohexanol could reach 99.9% by prolonging the reaction time to 24\u00a0h over RuMn(2:1)/Al2O3-SiO2 catalyst. Meanwhile, phenol and 2,6-dimethoxyphenol could also be converted to cyclohexanol with high yields of 100% and 94.5%, respectively. The catalyst was easily separated from the aqueous solution and can be reused 4 times without obvious loss of activity. The Al2O3-SiO2 composite support provided appropriate acid sites, large surface area and modifiable pores, and the addition of Mn improved the dispersion of Ru, and provides moderate acid-base sites, which were all attributed to the high performance of RuMn/Al2O3-SiO2. Therefore, the present strategy has great potential for application in high-quality biofuel production from renewable lignocellulosic biomass due to its high efficiency, little catalyst dosage, green solvent, and mild conditions.\nMengting Chen: Conceptualization, Methodology, Investigation, Writing \u2013 original draft, Writing \u2013 review & editing. Qifeng Zhong: Methodology, Resources, Writing \u2013 review & editing. Meihua Zhang: Investigation, Writing \u2013 review & editing. Hao Huang: Writing \u2013 review & editing. Yingxin Liu: Conceptualization, Methodology, Writing \u2013 review & editing, Visualization, Supervision. Zuojun Wei: Conceptualization, Methodology, Writing \u2013 review & editing, Visualization, Supervision, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was supported by the National Science Foundation of China (21878269, 21476211), the Zhejiang Provincial Natural Science Foundation of China (LY18B060016) and Jiangxi Qilin Chemical Industry Co., Ltd. (YX-[2012]008@).\n\n\n\nSupplementary material\n\nImage 6\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106550.", "descript": "\n                  Herein, RuMn/Al2O3-SiO2 multifunctional catalyst was prepared and efficiently converted lignin-derived phenols into cyclohexanol by aqueous partial hydrodeoxygenation (HDO). The optimal RuMn/Al2O3-SiO2 (Ru/Mn\u00a0=\u00a02) catalyst achieved 96.8% of cyclohexanol yield with 100% guaiacol conversion under mild reaction conditions (180\u00a0\u00b0C and 2.0\u00a0MPa H2) for 4\u00a0h, with a turnover frequency (TOF) of 861.1\u00a0h\u22121, which is better than most results reported. The high performance of RuMn/Al2O3-SiO2 was attributed to the synergic effect of highly dispersed Ru nanoparticles with small size (1.25\u00a0nm), the microsphere structure, the appropriate acid sites from Al2O3-SiO2 and acid-base sites from MnOx.\n               "}
{"full_text": "Energy conservation and environmental protection issues [1\u20135] and lightweight and recyclable materials have become common in the industry. Magnesium (Mg) alloy has low density, high rigidity and can improve energy efficiency in the transportation industry. The electromagnetic shielding is beneficial for 3C products that are lightweight. It has good mechanical properties but it has exceptionally high activity and is prone to severe corrosion in humid environments, which limits its application. Many surface treatments and coating systems are used to increase the corrosion resistance of Mg alloys. These methods include conversion coating [6\u201310], organic coating [11,12], electroplating [13\u201316], electroless plating [17\u201320], anodizing, micro-arc oxidation (MAO) [14,15,17] and other surface treatments [21\u201326].Electroless nickel-phosphorus (Ni\u2013P) plating is an appropriate process for forming amorphous metallic alloys. Due to its excellent corrosion and wear resistance, quickly expanding its range of applications, electroless Ni\u2013P plating has recently received wide attention and research interest [27,28]. Electroless Ni\u2013P plating is also popularly used in Mg alloys to increase their corrosion resistance and wear resistance properties [29\u201335]. However, when Mg is in electrical contact with nickel, there is dissimilar metal corrosion (galvanic corrosion) [35\u201337]. Studies show that a stable intermediate layer gives good protection for the substrate and increases the consistency of the nickel plating adhesion [17,38]. Common surface treatment methods that are used to improve corrosion resistance and stable intermediate layers are conversion coating and MAO. A manganese-vanadium conversion coating has been used for electroless Ni\u2013P plating on Mg alloys [6,39,40]. To increase resistance to corrosion properties, phosphoric acid is used for conversion coating/electroless nickel plating [41,42].MAO treatment is an environmentally friendly technology widely used to create corrosion resistant ceramic coatings on metals and alloys [43,44]. A previous study [17] by the authors showed that a treated phosphate-free-MAO/Ni\u2013P coating has significantly better corrosion resistance because there is good adhesion between the electroless Ni\u2013P plating layer and the MAO layer. Furthermore, there is a good mechanical locking between the electroless Ni\u2013P plating layer and the MAO layer, which increases corrosion resistance, mechanical and chemical properties [17].Varying the composition of electroless Ni\u2013P plating solution generates samples with amorphous, mix-structure, and nanocrystalline films [45,46]. In general, the corrosion resistance of the coating is affected by various deposited parameters of the electroless plating system [47,48]. Most studies add fluoride to the electroless plating system in order to achieve a uniform coating for the electroless Ni\u2013P plating solution because F\u2013 ions retard the corrosion of Mg alloy [49\u201353]. However, excessive F\u2013 ions reduce the growth rate of the coating, so it does not protect the substrate [50\u201352]. Few studies determine the effect of fluoride (F\u2013 ion) on the electroless Ni\u2013P plating of Mg alloys and no studies investigate the effect of adding F\u2013 ions to the MAO layer of Mg alloys during the electroless Ni\u2013P plating process. This study evaluates the effect of fluoride addition to the electroless Ni\u2013P plating solution on the MAO/Ni\u2013P composite coating properties and gives a better understanding of the growth mechanisms during electroless Ni\u2013P plating in the MAO layer.This study concerns the electroless Ni\u2013P plating on AZ31B Mg alloy with increased corrosion resistance after the substrate undergoes MAO treatment and compares the characteristic properties of chemical electroless Ni\u2013P plating solutions with/without fluoride content to form a MAO/Ni\u2013P composite coating. The microstructure is observed using SEM and TEM and the coating elements are analyzed using EDS and XRD. The corrosion resistance of the coating is tested using a potentiodynamic polarization curve, electrochemical impedance spectroscopy (EIS) and a salt spray test (SST). Finally, the effect of fluoride on the formation mechanism for the electroless Ni\u2013P plating film on MAO-coated AZ31B Mg alloy is explained in detail.Rectangular pieces (50\u00a0\u00d7\u00a050\u00a0\u00d7\u00a02\u00a0mm3) of AZ31B Mg alloy were used as the substrate, with 2.89\u00a0wt.% of Al, 0.897\u00a0wt.% of Zn, 0.282wt.% of Mn, 0.055wt.% of Si, 0.03wt.% of Fe. The remainder is Mg. Before the MAO process, all samples were ground with sandpaper of 400#, 800#, and 1200#, washed with deionized water, degreased with alcohol and dried at room temperature. The MAO process used a pulsed DC power supply with a duty cycle of 30\uff05 and a constant voltage of 400V in an alkaline solution for 7\u00a0min. The electrolytes for MAO treatment were Na2SiO36\u00a0g/L, NaOH 1.5\u00a0g/L and NaF 3\u00a0g/L. The temperature of the electrolytes was maintained at 10\u00a0\u00b1\u00a05\u00a0\u00b0C using a water cooling system.After MAO treatment, the sample was cleaned with deionized water, wiped with alcohol and dried at room temperature. The MAO samples were then immersed in a laboratory-developed catalyst ink for 10\u00a0s and dried in a 90\u00a0\u00b0C oven before electroless Ni\u2013P planting. The catalyst ink consists of a metal palladium nanoparticle that easily penetrates MAO holes and initiates electroless Ni\u2013P plating. The electrolyte composition and processing conditions for electroless Ni\u2013P plating are listed in Table 1\n.Scanning electron microscopy (SEM) using a JEOL JSM-IT100microscope equipped with Energy Dispersive Spectrometer (EDS) was used to observe the surface and cross-section morphology of the MAO and MAO/Ni\u2013P composite coatings. FIB-SII 3050SE was used as a sample cutting. The transmission electron microscope (TEM) was the Philips Tecnai F30 of National Taiwan University, which uses an acceleration voltage from a LaB6 gun of 300\u00a0keV. The chemical composition of the coating was measured using EDS and the TEM. The XRD experiments used a Bruker D2 PHASER X-ray diffractometer (\u03bb\u00a0=\u00a01.54184\u00a0\u00c5, 30\u00a0kV and 10\u00a0mA) with Cu K\u03b1 radiation. The scanning range for the diffraction angle (2\u03b8) was 10\u00b0 and 90\u00b0, with a step width of 0.05\u00b0 and a time step of 0.5\u00a0s.All Electrochemical tests used a Versa STAT 4 potentiostat/frequency to analyze the corrosion behavior of the MAO and MAO/Ni\u2013P composite coatings. A three-electrode cell, with a saturated Hg/Hg2Cl2/KCl was used as the reference electrode and a Pt flake as the counter electrode, with the sample as the working electrode (a circle with a diameter of 1\u00a0cm is the measurement area). Potentiodynamic polarization specimens were measured in 3.5\u00a0wt.% NaCl solution. Before the test, all samples were tested for 600\u00a0s at the open circuit potential (OCP). The scanning range of the polarization curve is relative to the open circuit potential from \u22120.3V to +0.5\u00a0V. The scanning rate is 0.005\u00a0V/s and the 50\u00a0mV interval between the upper and lower corrosion potentials is used for the Tafel approximation method to calculate the corrosion current density. An EIS test was performed after immersion in OCP for 10\u00a0min at frequencies of 100\u00a0kHz to 10\u00a0mHz and using a sinusoidal AC perturbation of 10\u00a0mV amplitude. A salt spray test was conducted according to ASTM B-117 [54\u201358]. The specimens were placed in a 5\u00a0wt.% NaCl solution at a pH of 6.5\u20137.2 and atomized into a mist. The heating chamber was maintained at 35\u00a0\u00b0C.The microstructure results for the MAO coating are shown in Fig.\u00a01\n. The coating consists of two layers, as shown in Fig.\u00a01(a). The outer-porous layer is located on the surface and has many pores. The second layer, called the compact layer, is between the outer layer and the substrate. The coating thickness is about 6\u20138\u00a0\u03bcm. The SEM image for the surface of the MAO coating in Fig.\u00a01(b) displays that the micro-pores size is similar or the same. The micro-pores form when the voltage penetrates through the coating, and their size is approximately 1.0\u00a0\u00b1\u00a00.3\u00a0\u03bcm.The corrosion resistance of the MAO coating is measured using an electrochemical test in a 3.5\u00a0wt.% NaCl solution. Bare AZ31B was also used in the test for comparison. Polarization curves and an EIS test for bare AZ31B and the MAO-coated AZ31B specimens are respectively shown in Figs. 2\n(a) and (b). The MAO-coated specimen has better corrosion resistance than the uncoated sample. In terms of the corrosion rate, the results indicate that the corrosion resistance of the coated specimen (1.01\u00a0\u00d7\u00a010\u22128 A/cm2) is better than that for bare AZ31B (3.66\u00a0\u00d7\u00a010\u22125 A/cm2). These coatings reduce the corrosion current by approximately 2\u20133 orders of magnitude in the potentiodynamic polarization tests.For this EIS measurement, the absolute impedance (|Z|\n\nf\u00a0=\u00a00.01\u00a0Hz\n) of the MAO-coated AZ31B specimens increases to 4.3\u00a0\u00d7\u00a0106\u00a0\u03a9\u2027cm2 from the value of 1.7\u00a0\u00d7\u00a0103 for bare AZ31B, as shown in Fig.\u00a02(b). This increase is consistent with the result for the polarization curves in Fig.\u00a02(a). The MAO-coated AZ31B specimens show superior corrosion resistance. SST was used to determine the corrosion resistance. The corrosive medium attacks weak points on the sample surface. If bare AZ31B is subject to SST for 96\u00a0h, the corroded area fraction is 100%, as shown in Fig.\u00a03\n(a). Fig.\u00a03(b) shows the overall morphology of the MAO-coated AZ31B specimens after 96\u00a0h of SST. The results show that the MAO-coated AZ31B specimens have excellent corrosion resistance and the corroded area fraction is approximately 0%.The characteristics of the initial deposition stage (30\u00a0s, 3\u00a0min, and 10\u00a0min) are studied to determine the role of fluoride in the electroless Ni\u2013P plating solution. Fig.\u00a04\n shows the substrate after MAO treatment and compares the surface morphologies for the two chemical electroless Ni\u2013P plating solutions, which are used to form a MAO/Ni\u2013P composite coating for 30\u00a0s. Fig.\u00a04(a) shows the SEM image of the MAO specimens that are treated in the fluoride-free solution for 30\u00a0s. The MAO surface has a discontinuous coating. The micro-pores are larger than the original and the inside of the discharge holes feature rosettes.\nFig.\u00a04(c) shows the backscattered electron image (BSE) of MAO samples treated in the fluoride-free solution for 30\u00a0s. The dark matrix in the SEM/BSE image is the Mg chemical compound with a low atomic number and the bright matrix is the Ni chemical compound, which has a higher atomic number. MAO specimens treated in the fluoride-containing solution for 30\u00a0s exhibit the MAO surface completely covered with a Ni\u2013P coating, shown in Figs. 4(b) and (d). Table 2\n shows the composition of the various electroless Ni\u2013P plating treatments for MAO specimens for SEM/EDS taken from the areas marked as 1, 2, 3 and 4 in Figs. 4(c) and (d). The coating formed in the presence of fluoride contains more Ni species than that formed in the absence of fluoride. There is only a minimal amount of Ni in Area 1.\nFig.\u00a05\n shows SEM surface morphology of samples that undergo electroless Ni\u2013P plating for 3min and 10\u00a0min, with and without the addition of fluoride. The surface morphology shows an incomplete coating, as shown in Fig.\u00a05(a). Due to the fluorine-free protection, the MAO surface is incomplete during the initial 30\u00a0s, so the electroless Ni\u2013P plating solution invades the substrate and the Mg alloy produces a displacement reaction [59\u201362]. As the plating time increases, the replacement film continues to form and the MAO coating at the damaged section becomes more broken. Fig.\u00a05(a) shows the disintegration of the middle MAO layer and the outer ring is a Ni\u2013P layer that grows unevenly.\nFig.\u00a05(c) shows the MAO specimens treated in the electroless Ni\u2013P plating bath with fluoride for 3\u00a0min. The Ni\u2013P coating gradually covers the MAO layer and this coating is uniform and smooth. The surface of the MAO features only some micro-pores that are not filled with the Ni\u2013P coating. When treatment time increases to 10\u00a0min, the surface morphology of the MAO specimens treated in the electroless Ni\u2013P plating bath without fluoride shows an imperfect film, as shown in Fig.\u00a05(b). The surface film peels off due to poor adhesion. MAO specimens treated in the fluoride-containing solution for 10\u00a0min are shown in Fig.\u00a05(d). The MAO surface is completely sealed and the Ni\u2013P nodules are larger than the nodules in the fluoride-free bath solution. The joints of the nodules are closely attached and the surface is complete and evenly coated.\nFig.\u00a06\n shows the cross-sectional SEM image of a sample that undergoes electroless Ni\u2013P plating for 3min and 10\u00a0min, with and without fluoride addition. Fig.\u00a06(a) shows that the MAO layer is broken and the structure is destroyed. The Ni\u2013P layer covered by the MAO upper layer is thinner and the plating layer features large undulations. The thickness is about 1.25\u00a0\u00b1\u00a00.2\u00a0\u03bcm. Fig.\u00a06(b) shows that the MAO layer has a complete structure, which is conducive to the batching of nickel-phosphorus coatings. The thickness is 1.60\u00a0\u00b1\u00a00.1\u00a0\u03bcm. The coating becomes thicker with the addition of fluorine, which is consistent with the surface topography results.For a treatment time of 10\u00a0min, the cross-sectional SEM image of MAO specimens that are treated in an electroless Ni\u2013P plating bath without fluoride is shown in Fig.\u00a06(c). There is more significant internal corrosion of the MAO and there is thinning and embedding of the Ni\u2013P layer. This is not conducive to nickel-phosphorus layer batching. The thickness of the coating is about 2.1\u00a0\u00b1\u00a00.1\u00a0\u03bcm. MAO specimens that are treated in the fluoride-containing solution for 10\u00a0min are shown in Fig.\u00a06(d). The Ni\u2013P coating is completely batched on the MAO coating and has good binding properties so the layer thickness is about 4.5\u00a0\u00b1\u00a00.2\u00a0\u03bcm.These cross-sectional characterizations show that the structure of the fluorine-containing composite coating is complete but the Ni\u2013P layer and MAO coating which are obtained from fluoride-free bath solution do not achieve a good mechanical locking force. Previous studies [17,50,63] showed that there is a degradation of corrosion resistance for MAO-coated Mg alloy that is produced by electroless Ni\u2013P plating. With the initiation of Ni\u2013P deposition, H+ ions form from the oxidation of the reducing agent that is adsorbed on the inner layer surface to dissolve MgO. Therefore, H+ ions accelerate damage to the MAO coating and reduce its corrosion resistance.If a fluorine compound is formed on the MAO coating, it gives better protection. During the process of continuous nickel reduction, the fluorine-free solution damages the plating solution for about 10\u00a0min. This results in severe damage to the MAO layer due to the absence of nickel fluoride. Loss of the structure that retains the catalyst ink causes the nuclei to fall to the bottom of the beaker of plating solution and this reaction quickly destroys the plating solution, so the experiment is terminated.The plating layer of the test piece without nickel fluoride shows significant peeling and cracking. No fluorine is added because during electroless Ni\u2013P plating, the MAO layer loses fluoride protection, so the MAO layer is not conducive to the Ni\u2013P batch coating. The plating layer peels off at the bottom of the burned back area; regarding this situation, nickel ions form in the plating solution. Therefore, the plating solution was destroyed and blackened for about 10\u00a0min, so the experiment was not continued.\nFig.\u00a07\n(a) shows the polarization curves for electroless Ni\u2013P plating for 30\u00a0s, 3\u00a0min and 10\u00a0min, with and without fluoride addition. The polarization test results show the corrosion potential (E\n\ncorr\n) and the corrosion current density (i\n\ncorr\n) in Table 3\n. Regardless of the treatment time, the corrosion resistance of the fluorine-containing treatment is far better than that which does not use fluorine. Fig.\u00a07(b) represents the Bode plots for various electroless Ni\u2013P plating samples using the same experimental parameters and process as Fig.\u00a07(a). The Bode plots show that the absolute impedance of the fluorine-containing treatment is higher than that for coatings without fluoride. The absolute impedance (|Z|\n\nf\u00a0=\u00a00.01\n\u00a0\nHz\n) results are also listed in Table 3. The SEM surface morphology is shown in Figs. 4 and 5. The fluorine-containing treatments produce fewer defects than coatings without fluoride. The NaCl solution corrodes and penetrates through the grain boundaries and reacts with the electroless Ni\u2013P plating layer. If the defects on the electroless Ni\u2013P plating layer are obvious, the electroless Ni\u2013P plating corrodes more significantly.\nFig.\u00a08\n shows the complete process for an electroless Ni\u2013P plating time of 40\u00a0min, with and without fluoride addition. Figs. 8(a) and (c) show that the MAO porous layer and the dense layer are damaged and loose and have no structure, leaving only the partially replaced nickel-phosphorus nodules interspersed in the defects. MAO specimens that are treated in a fluoride-containing solution for 40\u00a0min are shown in Figs. 8(b) and (d). The Ni\u2013P coating is intact, the surface is flat and the boundaries of the Ni\u2013P nodules are tightly connected, so this Ni\u2013P coating is uniform and balanced, with a thickness of 6.60\u00a0\u00b1\u00a00.4\u00a0\u03bcm.\nFig.\u00a09\n shows the EDS mapping results for the surface of the test piece for electroless Ni\u2013P plating for 40\u00a0min. The broken morphology of the discontinuous Ni\u2013P coating is verified by the presence of O and Mg to be a MAO layer. The Ni signal and the middle part are shown in Fig.\u00a09(a). The Ni\u2013P coating is uniformly and densely coated, no MAO layer is exposed and both Ni and P signals are shown in Fig.\u00a09(b). The surface composition shows that the non-nickel fluoride damages the MAO coating, causing the Ni\u2013P coating to exhibit a discontinuous film structure. On the other hand, the fluoride-containing treatment maintains the integrity of the MAO coating; thus, a continuous and dense Ni\u2013P coating forms.\nFig.\u00a010\n shows MAO specimens treated in an electroless Ni\u2013P plating bath for 40\u00a0min after 96\u00a0h SST. The Ni\u2013P layer is not completely deposited on the surface of the MAO coating when a fluoride-free solution is used, so there is severe corrosion. MAO specimens treated in the fluoride-containing solution for 40\u00a0min show only one corrosion spot after 96\u00a0h SST. The addition of fluorine affects the electroless Ni\u2013P plating.A previous study by the authors [17] used a pull-off test [64\u201366] to verify the adhesion of Ni\u2013P coatings on MAO-coated samples. The pull-off results are listed in Table 4\n. The fluoride-containing treatment produces better adhesion than the fluoride-free treatment and the surface morphology and the integrity of the micro-pores is maintained. The uniform initial particle coating is also conducive to the subsequent coating of the Ni\u2013P layer and a mechanical combination with the MAO coating, as shown by the microstructure results. The fluoride-containing treatment protects the MAO-coated AZ31B alloy. The addition of fluoride in an electroless Ni\u2013P plating bath gives good mechanical strength and better corrosion resistance for the MAO coating.XRD patterns were evaluated to determine the crystal structure of Ni\u2013P coatings on MAO-coated samples. The results are shown in Fig.\u00a011\n. The XRD results for the MAO coating for the different electroless Ni\u2013P plating solutions for 30\u00a0s show that the main structure of the coatings is Mg, MgO, MgF2 and NaMgF3. However, other structures appear in the MAO coating for the fluoride-containing solution, with signals for peaks at 33.48\u00b0 and 47.04\u00b0 being attributed to NaMgF3. The coating time is only 30\u00a0s, so insufficient elemental Ni is deposited to be detected in the XRD. However, the SEM/EDS and TEM/EDS analyses for all coating times verify the existence of elemental Ni. The XRD analysis shows that there is a signal for MgF2, as shown in Fig.\u00a011. The chemical reaction produces NaMgF3 and MgF2 with orthogonal, tetragonal and cubic crystal structures, respectively. The lattice parameters for NaMgF3 are a\u00a0=\u00a05.3603\u00a0\u00c5, b\u00a0=\u00a05.4884\u00a0\u00c5, and c\u00a0=\u00a07.666\u00a0\u00c5. The XRD semi-quantitative analysis shows that, due to the small number of particles that is produced in the initial reaction, the film is not completely formed, so the peak amplitude is small. The initial reaction for the fluoride-containing solution also produces a NaMgF3 cubic lattice. The EDS composition analysis also shows that the square morphology of the initial surface is MgF2 and NaMgF3 cubic crystal.In order to confirm the MgF2 and NaMgF3 that are observed by XRD, TEM was used for microstructure analysis. Fig.\u00a012\n shows the cross-sectional TEM images for MAO specimens that are treated in an electroless Ni\u2013P plating bath with fluoride for 30\u00a0s MgF2 particles and NaMgF3 cubic lattices are present on the surface and the inner wall of the hole over a large area, which grow with nickel and phosphorus particles. For a specific area of the inner wall of the hole, an EDS dot analysis was performed on the particles that are distributed in the hole.The elemental composition is shown in Table 5\n. The initial reaction containing fluorine produces MgF2 particles and NaMgF3 cubic lattice and Ni\u2013P particles. When fluorine is added in the Ni\u2013P solution, the MAO layer morphology is protected and the integrity of the discharge holes is maintained. The uniform initial particle coating is also beneficial to the subsequent coating of Ni\u2013P and allows mechanical combination with the MAO layer.The microstructural characterizations show that the fluoride ions in the electroless Ni\u2013P plating bath are crucial. The initial reaction for electroless Ni\u2013P plating produces H+ ions, which dissolve the MAO layer to produce Mg ions. The fluoride ions, sodium ions and magnesium ions in the electroless Ni\u2013P plating solution form NaMgF3, which prevent the MAO layer from continuing to dissolve, which is beneficial to the subsequent batch coating of the Ni\u2013P coating.The generated H+ ions corrode the MAO layer and MgO is converted to Mg(OH)2, so the MAO is less resistant to corrosion. Fluoride ions in the electroless Ni\u2013P plating bath form MgF2 and NaMgF3 thin films to protect the MAO layer but the low fluorine content does not allow the thin film to form quickly; consequently, there appears an uneven coating and poor adhesion of the Ni\u2013P coating. Corrosive mediums can easily penetrate the substrate through defects, so the coating lacks protection and pitting corrosion ensues. As the time for SST increases, corrosion increases the MAO layer expands. Furthermore, the electroless Ni\u2013P coating peels off and loses its ability to protect. Excessive fluorine content also leads to excessive growth of NaMgF3, so the film becomes discontinuous and breaks.Hydrogen ions continue to corrode the MAO layer through the defects. The chloride ion is a strong adsorptive anion, which easily replaces oxygen and water molecules and is preferentially adsorbed onto the surface of the electroless Ni\u2013P plating layer. Ni\u2013P ions are in a dynamic balance: they are destroyed to form soluble NiCl2, which leads to the destruction of the coating morphology and the production of corrosion points. The formula for this process is:\n\n(1)\nNi \u2192 Ni2+ + 2e\u2013\n\n\n\n\n\n(2)\nNi2+\u00a0+\u00a02Cl\u2013 \u2192 NiCl2\n\n\n\nThe schematic diagram is shown to determine the configuration of the complex and the mechanism for MAO ceramic growth in Fig.\u00a013\n. At the beginning of the electroless Ni\u2013P plating reaction, the hydrogen ions that are generated at the interface have a detrimental effect and the pH value at the interface reaction decreases, so the MAO layer corrodes and the surface morphology is destroyed. If there is no added fluorine protection in the electroless Ni\u2013P plating solution, the MAO becomes loose and the catalyst does not flow in the hole (Fig.\u00a013), so the electroless Ni\u2013P plating solution reaches the substrate and a displacement reaction. It then completely coats the electroless Ni\u2013P plating and the MAO layer. This coating adheres poorly and peels off; thus, corrosive media infiltrate from the cracks and the peeled areas on the coating. Finally, there is severe corrosion of the substrate, as shown in Table 4.If the fluorine content is insufficient in the electroplating bath, the resulting MgF2 particles and the cubic lattice of NaMgF3 are unevenly distributed and some MAO is not protected, resulting in defects. The subsequent electroless Ni\u2013P coating layer grows along the grains; accordingly, so there are undulations and uneven coating in some areas. The uneven growth of the coating of the electroless Ni\u2013P plating creates cracks, so corrosive media flow from micro-defects and corrosion occurs.This study determines the effect of fluoride on the formation of an electroless Ni\u2013P plating film on MAO-coated AZ31B Mg alloy using Ni\u2013P plating solutions with/without fluoride.Preliminary observation of the surface of electroless Ni\u2013P plating for a short period of time (30\u00a0s, 3\u00a0min and 10\u00a0min) shows that the initial surface of the MAO specimens treated in the fluoride-containing solution contains MgF2 particles and NaMgF3 cubic lattice. MgF2 particles and NaMgF3 cubic lattice distribution impact the Ni\u2013P coating during the electroless plating process. Hydrogen that is formed at the interface during the electroless coating process reduces the pH value. The fluoride-free electroless Ni\u2013P plating solution induces the MAO layer to become loose due to a lack of fluoride protection. Therefore, the MAO sample treated in a fluoride-free bath suffers peel-off of the coating and the electroless Ni\u2013P plating does not cover all of the surface due to a lack of fluoride protection.On the contrary, the MAO sample treated in the fluoride-containing solution has the best corrosion resistance and only one corrosion spot after 48\u00a0h SST. With the initiation of Ni\u2013P deposition, H+ ions are formed by the oxidation of the reducing agent which is adsorbed onto the inner layer surface and dissolve MgO. H+ ions accelerate damage to the MAO coating and reduce its corrosion resistance. This study demonstrated that the formation of a fluorine compound on the MAO coating protects the coating and the electroless Ni\u2013P plating coating completely covers the surface.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This study was financially supported by the Ministry of Science and Technology of Taiwan, Republic of China, under Grant No.MOST 109-2221-E-606 -009 -MY3.", "descript": "\n                  This study adds fluoride to the electroless nickel-phosphorus (Ni\u2013P) plating solution to prevent the deterioration of MAO-coated AZ31B Magnesium alloy after contact with an electroless plating bath. During the electroless Ni\u2013P plating process, fluoride reacts with Ni2+ ions and the MAO coating to form interphases (NaMgF3), which exhibit good bonding and corrosion resistance. NaMgF3 buffers H+ ions formed from the initiation of Ni\u2013P deposition, preventing the interface of materials from damaging the MAO coating with H+ ions. As immersion time increases, nickel is scattered over the coating.\n                  The fundamental data for MAO/Ni\u2013P coated AZ31B Mg alloy determines whether there is fluoride in the electroless Ni\u2013P plating solution. The results show that the coating for a fluoride-containing solution is more resistant to corrosion than those in fluoride-free solution. The compositions, structure and morphology of the MAO/Ni\u2013P coatings that formed for different working parameters are determined using energy dispersive spectroscopy (EDS), X-ray diffraction (XRD), scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The corrosion resistance of the MAO/Ni\u2013P coatings is calculated in a 3.5\u00a0wt.% NaCl solution using a potentiodynamic polarization test, electrochemical impedance spectroscopy (EIS) and a salt spray test (SST).\n               "}
{"full_text": "C2H2 semi-hydrogenation to C2H4 is a commonly used method to remove small quantity of C2H2 impurities in C2H4-rich stream, in which C2H2 acts as an impurity that poisons the catalysts used subsequently for the polymerization of ethylene [1\u20133]. In C2H2 semi-hydrogenation, the most effective catalyst was found to be Pd-based catalysts owing to high C2H2 conversion occurred at low temperature, however, C2H4 selectivity still needs to be improved because of C2H4 over-hydrogenation to ethane [4\u20136]. Thus, the development of Pd-based catalysts with high selectivity and activity in C2H2 semi-hydrogenation is still highly expected.Nowadays, in order to effectively utilize the active components of the catalyst, the metal catalyst is usually dispersed on the support with high specific surface area. However, only a small part of surface metal atoms participate in the catalytic reaction and the atom utilization rate is low. Thus, small size of supported metal clusters have been reported due to high atom utilization and more unique catalytic performance, however, the particle size in the subnanometer range is sensitive to its structure with particular electronic properties. For example, Liu et\u00a0al. [7] theoretically found that the size of Pd\nn\n clusters affects H2 dissociation activity on Pd\nn\n(n\u00a0=\u00a04,6,13,19,55) clusters. Mercedes et\u00a0al. [8] calculated the LUMO and HOMO orbitals of Au1\u223cAu38 clusters, and predicted that Au1, Au3 and Au38 have the best activity. Zhang et\u00a0al. [9] prepared the atomically dispersed Pt3 clusters anchored over the core\u2013shell nanodiamond@graphene, which presented excellent catalytic performance for n-butane direct dehydrogenation at a temperature as low as 450\u00a0\u00b0C. On the other hand, for C2H2 semi-hydrogenation, Gluhoi et\u00a0al. [10] observed that when Au particle size was less than 3\u00a0nm, Au/Al2O3 showed high C2H2 conversion and C2H4 selectivity. Abdollahi et\u00a0al. [11] found that Pd2 cluster have higher activity than Pd12 cluster, indicating that Pd2 cluster is more suitable. Density functional theory (DFT) studies by Xiao et\u00a0al. [12] found that H2 adsorption capacity on the graphene supported Pd\nn\n(n\u00a0=\u00a01\u20135) clusters was stronger than that on Pd (111) surface. Huang et\u00a0al. [13] prepared single atom Cu catalyst supported by the nanodiamond-graphene, C2H2 conversion is 95%, C2H4 selectivity is 98%, and the catalyst has good stability. Shi et\u00a0al. [14] experimentally synthesized Cu single atom and nanoparticles corresponding to the sizes of about 3.4, 7.3 and 9.3\u00a0nm over Al2O3 support using atomic layer deposition, indicating that a size decrease of Cu nanoparticle obviously reduces the activity of C2H2 semi-hydrogenation but gradually improves both C2H4 selectivity and durability. The experiments by Huang et\u00a0al. [15] prepared Pd1/ND@G catalyst with the atomically dispersed Pd over a defective nanodiamond-graphene (ND@G), which showed significantly high C2H4 selectivity (90%) and C2H2 conversion (100%) in C2H2 semi-hydrogenation. Thus, the studies on small size of metal cluster presenting particular active sites are of great significance in C2H2 semi-hydrogenation.Recently, graphdiyne (GDY), a new allotrope carbon material including C atoms with sp and sp2 hybrid, has attracted broad attentions [16\u201321]. The C atoms of GDY show a \u03c0\u2013\u03c0 conjugate system, which is highly delocalized in the whole planar framework [16,22,23]. Moreover, in comparison with grapheme, GDY has a unique pore structure to provide abundant adsorption site and more open storage space for molecular adsorption. As shown in Fig.\u00a01\n, the basic geometry of GDY is 6-membered ring (6\u00a0MR) and 18-membered ring (18\u00a0MR). The 18\u00a0MR structure provides a natural framework for anchoring metal clusters through a strong metal carbon covalent bond, thus forming a stable and isolated structure. For example, Lu et\u00a0al. [24] systematically examined the adsorption of single atom Pd, Pt, Rh or Ir on the GDY, and found that the single atom was favorable for embedding into 18\u00a0MR of GDY and combining with four carbon atoms. The adsorption and diffusion behavior of Au, Cu, Fe, Ni or Pt atoms on the GDY were also researched by Lin et\u00a0al. [25] using theoretical calculation, and found that the metal atoms have good thermal stability and very small overflow rate on the GDY even at 900\u00a0K. Chen et\u00a0al. [26] studied CO oxidation reaction on the GDY supported Ag38 cluster, which showed the excellent performance due to the unique combination of the cluster and GDY.Further, the metal atoms and clusters supported by GDY are also ideal catalysts, which have been widely applied for the hydrogenation reaction of unsaturated hydrocarbons. For example, Xing et\u00a0al. [27] investigated C2H2 semi-hydrogenation over the catalysts with the cluster MxN3-x (M, N = Ru, Os) supported by GDY, indicating that three atom metal clusters can be firmly anchored on the 18\u00a0MR of GDY, and effectively catalyze C2H2 semi-hydrogenation to produce C2H4; the synergistic effect between the metal cluster and GDY as a charge buffer contributes to the improvement of catalytic performance. However, up to now, few studies about C2H2 semi-hydrogenation focus on the catalysts with GDY supported small size of metal clusters PdxMy, meanwhile, the effects of cluster composition and size on the activity and selectivity are still unknown, which would provide an open space for designing highly-efficient GDY supported Pd-based catalyst in C2H2 semi-hydrogenation. Moreover, previous studies [28] showed that a Group IB metal (Cu, Ag or Au) doped into Pd to form Pd-based bimetallic alloys can well improve C2H4 selectivity in C2H2 semi-hydrogenation; meanwhile, Abdollahi et\u00a0al. [29] theoretically demonstrated that among the Nin(n\u00a0=\u00a02\u201310) nanoclusters, Ni6 nanocluster could be used as a suitable catalyst in C2H2 semi-hydrogenation. Wongwaranon et\u00a0al. [30] experimentally observed that C2H4 selectivity was improved on the Pd/Ni-modified \u03b1-Al2O3 catalysts in the presence of Ni atoms. Jin et\u00a0al. [31] experimentally claimed that PdNi catalyst possessed high selectivity and stability for C2H2 semi-hydrogenation. Thus, the bimetallic PdM(M\u00a0=\u00a0Cu, Ag, Au, Ni) catalysts can be well applied in C2H2 semi-hydrogenation to improve its catalytic performance.In this study, a large number of PdxMy/GDY catalysts using GDY supported different sizes of PdxMy (M\u00a0=\u00a0Cu, Ag, Au, Ni; x + y\u00a0=\u00a01\u20133) bimetallic clusters have been for the first time designed; Then, the underlying mechanism of the hydrogenation process of C2H2 on the PdxMy/GDY catalysts were fully investigated using DFT calculations, the obtained results were expected to illustrate the effects of cluster composition and size in the PdxMy/GDY catalysts on the activity and selectivity of C2H4 formation. This study would provide a good clue for designing and screening out the potential catalysts with GDY supported small sizes of PdxMy clusters and other metal clusters in C2H2 hydrogenation process.Dmol3 code [32,33] in Materials Studio 8.0 were carried out for the performance of all DFT calculations. The exchange-correlation functional PBE with generalized gradient approximation (GGA) [34,35] was used. The double-numeric polarized (DNP) basis set was used to expand valence electron functions [36,37]. The van der Waals correction (DFT-D) [38] method was used to correct the weak adsorption free energy underestimated by the GGA functional. The all electron and effective core potential (ECP) basis set were used to treat the non-metal atoms and the inner electrons of metal atoms, respectively [39,40]. The k-point 4\u00a0\u00d7\u00a04\u00a0\u00d7\u00a01 was considered together with a smearing of 0.005 Ha for all calculations. The complete LST/QST techniques were used to obtain transition state of an elementary step [41,42], which was confirmed by the methods of Frequency analysis and TS confirmation implemented in Dmol3 code.A lots of studies showed that the type of active metals [43], the temperature [44] and the H2/C2H2 ratio [45\u201348] in C2H2 semi-hydrogenation can all obviously affect the formation of green oil over Pd-based catalyst, suggesting that the high temperature and high H2/C2H2 ratio can prevent \u201cgreen oil\u201d formation leading to the deactivation of polymerization catalysts [6,49,50]. This study only focus on the investigations about the effect of the types of active metals including the metal cluster composition and size on the activity and selectivity of C2H2 hydrogenation process, as a result, the effect of green oil formation on the activity and selectivity of C2H2 hydrogenation process are expected to be ignored, a high temperature of 425\u00a0K and a high H2: C2H2 ratio of 10 corresponding to C2H4, C2H2 and H2 partial pressures of 0.89, 0.01 and 0.1\u00a0atm were performed. Thus, all energies in the process of adsorption and reaction were the values at 425\u00a0K in this study (see details in the Supplementary Material).For the protocell of GDY, see Fig.\u00a01, the C\u2013C bond length on the 6\u00a0MR is 1.430\u00a0\u00c5, the C\u2212C and C=C bond lengths on the 18\u00a0MR are 1.390 and 1.232\u00a0\u00c5, respectively, which agree with the reported values of 1.430, 1.390 and 1.240\u00a0\u00c5 [24,51]. For the supercell of GDY, a single-layer p (2\u00a0\u00d7\u00a02) structure with a 30\u00a0\u00c5 vacuum thickness was constructed, and the lattice constant obtained through structure optimization is 18.880\u00a0\u00c5. During the calculations, the edge C atoms of GDY denoted as the red balls in Fig.\u00a01(b) were fixed; the C atoms of two 18\u00a0MR were fully relaxed.For PdxMy/GDY catalysts, as shown in Fig.\u00a02\n, five types of single metal catalysts, Pd, Cu, Ag, Au and Ni, are used to form metal clusters with the atom numbers of one, two and three, respectively; four types of bimetallic catalysts, Cu, Ag, Au, Ni alloyed with Pd to form PdxMy bimetallic clusters with the atom numbers of two and three, respectively. As a result, there are twenty-seven kinds of PdxMy/GDY catalysts, named as Pd1/GDY, Pd2/GDY, Pd3/GDY, Cu1/GDY, Cu2/GDY, Cu3/GDY, Pd1Cu1/GDY, Pd1Cu2/GDY, Pd2Cu1/GDY, Ag1/GDY, Ag2/GDY, Ag3/GDY, Pd1Ag1/GDY, Pd1Ag2/GDY, Pd2Ag1/GDY, Au1/GDY, Au2/GDY, Au3/GDY, Pd1Au1/GDY, Pd1Au2/GDY, Pd2Au1/GDY, Ni1/GDY, Ni2/GDY, Ni3/GDY, Pd1Ni1/GDY, Pd1Ni2/GDY and Pd2Ni1/GDY, respectively. Further, the interaction between PdxMy cluster and GDY is calculated [52,53] (see details in Table S1).C2H2 semi-hydrogenation follows the continuous hydrogenation of hydrocarbons [54]. Hydrogenation of C2H2 may occur by three routes [55\u201358], as presented in Fig.\u00a03\n, the first is that C2H2(ad) is successively hydrogenated to form C2H4(ad) via C2H3(ad) intermediate, then, C2H4(ad) desorb from the catalyst surface, which is the desired route for C2H2 semi-hydrogenation defined as C2H4 desorption route. The latter two is that C2H2(ad) hydrogenation via the common C2H3(ad) intermediate produces C2H4(ad) or CHCH3(ad), which could be further hydrogenated to form ethane via C2H5(ad) intermediate; these two routes, called as C2H4 hydrogenation route and CHCH3 hydrogenation route, are expected to be suppressed to facilitate the semi-hydrogenation of C2H2 to form C2H4.For PdxMy/GDY catalysts, to identify whether C2H4 desorption route prefers to occur in C2H2 semi-hydrogenation, firstly, it is necessary to calculate the priority between C2H4 desorption and its hydrogenation, if C2H4 desorption is more favored than C2H4 hydrogenation, namely, C2H4 desorption route is superior to C2H4 hydrogenation route; Then, judging whether C2H4 desorption route also prefers to occur compared to CHCH3 hydrogenation route. Based on above two aspects of analysis, we can confirm the catalysts with better C2H4 selectivity, on which C2H4 desorption route is the dominant among three routes.C2H4 feed produced by steam-cracking process is known to contain about 0.1\u20131% of C2H2 [59], only when C2H2 adsorption is stronger than C2H4 adsorption over the catalysts, the removal of trace C2H2 in C2H4-rich feed gas could be achieved on the catalysts. The intuitional comparison between C2H4 and C2H2 adsorption energies on PdxMy/GDY catalysts is shown in Fig.\u00a04\n (see details in Table S2 and Fig.\u00a0S1). For H, C2H3, CHCH3 and C2H5 species, Fig.\u00a0S2 and Table S3 give out the adsorption energies and stable configurations on above PdxMy/GDY catalysts.For PdxMy/GDY catalysts, Ag3/GDY is seriously deformed and unstable when the C2H2 or C2H4 species were adsorbed. On the Ag2/GDY and Ni3/GDY, C2H2 is not effectively adsorbed due to the weak physisoption (4.9 and 8.7\u00a0kJ\u00a0mol\u22121). On the Ag1/GDY, Au1/GDY and Pd1Au2/GDY, the adsorption ability of C2H2 and C2H4 species are close (26.9 and 22.2\u00a0kJ\u00a0mol\u22121, 125.4 and 129.9\u00a0kJ\u00a0mol\u22121, 55.2 and 53.9\u00a0kJ\u00a0mol\u22121), namely, trace C2H2 in C2H4-rich stream cannot be sufficiently adsorbed. However, as listed in Table S2, C2H2 has stronger adsorption ability than C2H4 on twenty-one kinds of PdxMy/GDY catalysts, including Pd1/GDY, Pd2/GDY, Pd3/GDY, Cu1/GDY, Cu2/GDY, Cu3/GDY, Pd1Cu1/GDY, Pd1Cu2/GDY, Pd2Cu1/GDY, Pd1Ag1/GDY, Pd1Ag2/GDY, Pd2Ag1/GDY, Au2/GDY, Au3/GDY, Pd1Au1/GDY, Pd2Au1/GDY, Ni1/GDY, Ni2/GDY, Pd1Ni1/GDY, Pd1Ni2/GDY and Pd2Ni1/GDY, namely, trace C2H2 to participate into the subsequent hydrogenation reaction can be sufficiently adsorbed in C2H4-rich stream.Based on above analysis, six kinds of Ag3/GDY, Ag2/GDY, Ni3/GDY, Ag1/GDY, Au1/GDY and Pd1Au2/GDY catalysts are excluded in C2H2 semi-hydrogenation on the basis of the weak C2H2 physisoption or the close adsorption ability between C2H2 and C2H4 species. Further, on other twenty-one kinds of PdxMy/GDY catalysts, C2H2 has stronger adsorption than C2H4, which favors the hydrogenation of C2H2.For PdxMy/GDY catalysts with stronger adsorption ability of C2H2 than C2H4, it is needed to firstly identify the priority of C2H4 between its desorption and hydrogenation, as listed in Table 1\n (see the structures in Fig.\u00a0S3).For PdxMy/GDY catalysts, C2H4 + H\u2192C2H5 is more favored or competitive compared to C2H4 desorption on the Pd2/GDY, Pd3/GDY, Cu1/GDY, Pd1Cu1/GDY, Pd1Ag1/GDY, Pd2Ag1/GDY, Pd1Au1/GDY and Ni1/GDY catalysts, which easily leads to ethane, thus, these eight kinds of PdxMy/GDY catalysts exhibit poor C2H4 selectivity, CHCH3 hydrogenation route does not need to be considered. However, as listed in Table 1, C2H4 desorption would be superior to its hydrogenation to C2H5 on thirteen kinds of PdxMy/GDY catalysts, including Pd1/GDY, Cu2/GDY, Cu3/GDY, Pd1Cu2/GDY, Pd2Cu1/GDY, Pd1Ag2/GDY, Au2/GDY, Au3/GDY, Pd2Au1/GDY, Ni2/GDY, Pd1Ni1/GDY, Pd1Ni2/GDY and Pd2Ni1/GDY, which would be favor C2H2 semi-hydrogenation to gaseous C2H4, and suppress its over-hydrogenation to ethane.As mentioned above, C2H4 preferred to be desorption instead of its hydrogenation on thirteen kinds of PdxMy/GDY catalysts, namely, C2H4 desorption route become dominant compared to C2H4 hydrogenation route, thus, we need to further identify whether C2H4 desorption route also prefers to occur compared to CHCH3 hydrogenation route on these thirteen kinds of PdxMy/GDY catalysts (see details in Figs. S4\u2013S15). The energy profiles of C2H2 semi-hydrogenation on Pd1/GDY catalyst is shown in Fig.\u00a05\n as an example.On Pd1/GDY catalyst, C2H2 + H\u2192C2H3 has the activation barrier of 38.4\u00a0kJ\u00a0mol\u22121, and it is exothermic by 73.9\u00a0kJ\u00a0mol\u22121; starting from C2H3 intermediate, C2H4 formation is superior to CHCH3 formation in kinetics (1.0 vs. 116.4\u00a0kJ\u00a0mol\u22121); further, C2H4 desorption would be much preferred kinetically compared to C2H4 + H\u2192C2H5 (65.1 vs. 326.5\u00a0kJ\u00a0mol\u22121), suggesting that Pd1/GDY catalyst is in favor of C2H2 semi-hydrogenation to produce gaseous C2H4. Similarly, the easy formation of gaseous C2H4 also occurs on the Cu2/GDY (Fig.\u00a0S4), Cu3/GDY (Fig.\u00a0S5), Pd1Cu2/GDY (Fig.\u00a0S6), Pd2Cu1/GDY (Fig.\u00a0S7), Pd1Ag2/GDY (Fig.\u00a0S8), Au2/GDY (Fig.\u00a0S9), Au3/GDY (Fig.\u00a0S10), Pd2Au1/GDY (Fig.\u00a0S11) and Pd1Ni2/GDY catalysts (Fig.\u00a0S14). However, CHCH3 formation leading to ethane is much easier than C2H4 formation on the Ni2/GDY (Fig.\u00a0S12, 62.0 and 105.7\u00a0kJ\u00a0mol\u22121), Pd1Ni1/GDY (Fig.\u00a0S13, 70.3 and 137.9\u00a0kJ\u00a0mol\u22121) and Pd2Ni1/GDY catalysts (Fig.\u00a0S15, 40.9 and 115.2\u00a0kJ\u00a0mol\u22121), as a result, these three types of catalysts present poor C2H4 selectivity due to the formation of ethane.The energy difference between C2H4 hydrogenation and its adsorption was used to quantitatively describe the selectivity of C2H4 (\u0394G\ns) using the Eq. (1), which has been widely applied in many previous studies [6,28,60,61].\n\n(1)\n\u0394G\ns\u00a0=\u00a0\u0394G\na\u2212|G\nads|\n\nWhere G\nads and \u0394G\na correspond to C2H4 adsorption free energy and the activation barrier of C2H4 hydrogenation to C2H5, respectively; the positive and large value of \u0394G\nsel means that the catalyst exhibits better C2H4 selectivity. As mentioned above, ten kinds of PdxMy/GDY catalysts have better C2H4 selectivity, including Pd1/GDY, Cu2/GDY, Cu3/GDY, Pd1Cu2/GDY, Pd2Cu1/GDY, Pd1Ag2/GDY, Au2/GDY, Au3/GDY, Pd2Au1/GDY and Pd1Ni2/GDY.Meanwhile, according to the two-step model widely used in the previous work [56,58,62,63] (see details in the Supplementary Material), the reaction rate of C2H4 formation was calculated to evaluate the catalytic activity on these ten kinds of PdxMy/GDY catalysts.As listed in Table 2\n, the selectivity of C2H4 over ten kinds of PdxMy/GDY catalysts, Pd1/GDY, Cu2/GDY, Cu3/GDY, Pd1Cu2/GDY, Pd2Cu1/GDY, Pd1Ag2/GDY, Au2/GDY, Au3/GDY, Pd2Au1/GDY and Pd1Ni2/GDY, are 261.4, 114.5, 162.6, 309.1, 67.1, 19.3, 384.2, 54.1, 104.7 and 85.5\u00a0kJ\u00a0mol\u22121, respectively, the corresponding activity of C2H4 formation are 1.69\u00a0\u00d7\u00a0108, 2.94\u00a0\u00d7\u00a0106, 7.63\u00a0\u00d7\u00a010\u22124, 3.45\u00a0\u00d7\u00a010\u22121, 3.45\u00a0\u00d7\u00a010\u22121, 4.25\u00a0\u00d7\u00a0109, 1.55\u00a0\u00d7\u00a0108, 8.71\u00a0\u00d7\u00a010\u221228, 2.16\u00a0\u00d7\u00a010\u22121, 2.26\u00a0\u00d7\u00a0107 and 3.63\u00a0\u00d7\u00a01010 s\u22121 site\u22121, respectively.Further, H2 dissociation may affect the activity of C2H4 formation; H2 adsorption and dissociation were calculated on above ten kinds of PdxMy/GDY catalysts with better C2H4 selectivity (see Table S4 and Fig.\u00a0S19). Our results show that only on the Au3/GDY, Pd2Au1/GDY and Pd1Ni2/GDY catalysts, the activation free energies of the rate-determining in C2H4 desorption route is lower than those of H2 dissociation, namely, H2 dissociation affects the catalytic activity toward C2H2 semi-hydrogenation to C2H4. Whereas it does not affect the catalytic activity of C2H4 formation on other seven types of Pd1/GDY, Cu2/GDY, Cu3/GDY, Pd1Cu2/GDY, Pd2Cu1/GDY, Pd1Ag2/GDY and Au2/GDY catalysts. Moreover, since this study only focus on the investigations about the effect of active metal types including the metal cluster composition and size on activity and selectivity of C2H2 semi-hydrogenation process, the effect of the initial H2 dissociation activity on the catalytic performance of C2H2 semi-hydrogenation process will be considered in our next work.For GDY supported single-metal catalysts, firstly, when the supported metal is the single atom, only Pd1/GDY is favorable for C2H4 formation, while on the Cu1/GDY, Ag1/GDY, Au1/GDY and Ni1/GDY, C2H2 could not be effectively adsorbed or the over-hydrogenation of C2H4 to ethane occurs. Secondly, when the supported metal is the double and three atoms cluster, only the Cu and Au clusters (Cu2/GDY, Cu3/GDY, Au2/GDY, Au3/GDY) are favorable for C2H4 formation; while on the Pd, Ag and Ni clusters, C2H2 cannot be effectively adsorbed or C2H4 is inclined to be over-hydrogenated to ethane. Thus, only five kinds of single metal catalysts including Pd1/GDY, Cu2/GDY, Cu3/GDY, Au2/GDY and Au3/GDY are favorable for C2H2 semi-hydrogenation to form gas phase C2H4.For GDY supported bimetallic catalysts, firstly, when the supported metal is double atoms cluster, all catalysts including Pd1Cu1/GDY, Pd1Ag1/GDY, Pd1Au1/GDY, Pd1Ni1/GDY are not favorable for C2H2 semi-hydrogenation to C2H4. However, when the supported metal is three atoms cluster, only five kinds of the catalysts including Pd1Cu2/GDY, Pd2Cu1/GDY, Pd1Ag2/GDY, Pd2Au1/GDY and Pd1Ni2/GDY are favorable for C2H2 semi-hydrogenation to C2H4.\nFig.\u00a06\n shows C2H4 selectivity and its formation activity over ten kinds of PdxMy/GDY catalysts favored the formation of gaseous C2H4 (Pd1/GDY, Cu2/GDY, Cu3/GDY, Pd1Cu2/GDY, Pd2Cu1/GDY, Pd1Ag2/GDY, Au2/GDY, Au3/GDY, Pd2Au1/GDY and Pd1Ni2/GDY), among them, the catalysts with high C2H4 selectivity are Au2/GDY (384.2\u00a0kJ\u00a0mol\u22121), Pd1Cu2/GDY (309.1\u00a0kJ\u00a0mol\u22121), Pd1/GDY (261.4\u00a0kJ\u00a0mol\u22121) and Cu3/GDY (162.6\u00a0kJ\u00a0mol\u22121), accordingly, C2H4 formation activity are 8.71\u00a0\u00d7\u00a010\u221228, 3.45\u00a0\u00d7\u00a010\u22121, 1.69\u00a0\u00d7\u00a0108 and 7.63\u00a0\u00d7\u00a010\u22124 s\u22121 site\u22121, respectively, indicating that Pd1/GDY has the highest C2H4 formation activity, while Au2/GDY, Pd1Cu2/GDY and Cu3/GDY presents poor C2H4 formation activity. On the other hand, the catalysts with high C2H4 formation activity are Pd1Ni2/GDY (3.63\u00a0\u00d7\u00a01010 s\u22121 site\u22121), Pd2Cu1/GDY (4.25\u00a0\u00d7\u00a0109 s\u22121 site\u22121), Pd1/GDY (1.69\u00a0\u00d7\u00a0108 s\u22121 site\u22121) and Pd1Ag2/GDY (1.55\u00a0\u00d7\u00a0108 s\u22121 site\u22121), accordingly, C2H4 selectivity are 85.5, 67.1, 261.4 and 19.3\u00a0kJ\u00a0mol\u22121, respectively, thus, Pd1/GDY catalyst presents the highest C2H4 selectivity. As shown in Fig.\u00a0S20, generally, there is not a seesaw effect for the activity and selectivity, only both Au2/GDY and Pd1Cu2/GDY have higher C2H4 selectivity (384.2 and 309.1\u00a0kJ\u00a0mol\u22121), while these two catalysts have lower activity (8.71\u00a0\u00d7\u00a010\u221228 and 3.45\u00a0\u00d7\u00a010\u22121 s\u22121 site\u22121). On the contrary, Pd1Ni2/GDY, Pd2Cu1/GDY and Pd1Ag2/GDY catalysts have higher activity of C2H4 formation (3.63\u00a0\u00d7\u00a01010, 4.25\u00a0\u00d7\u00a0109 and 1.55\u00a0\u00d7\u00a0108 s\u22121 site\u22121), while the lower C2H4 selectivity (85.5, 67.1 and 19.3\u00a0kJ\u00a0mol\u22121).Based on above analysis, the composition and size of supported metal cluster PdxMy in PdxMy/GDY catalysts present the sensitivity toward the selectivity and activity of C2H2 semi-hydrogenation. Among them, taking C2H4 selectivity and its formation activity into consideration, GDY supported single atom Pd catalyst (Pd1/GDY) in this study should provide the best selectivity (261.4\u00a0kJ\u00a0mol\u22121) and excellent catalytic activity (1.69\u00a0\u00d7\u00a0108 s\u22121 site\u22121) for C2H2 semi-hydrogenation to gaseous C2H4.Further, the catalytic origin of Pd1/GDY catalyst with the highest activity is revealed. The high C2H4 selectivity should be attributed to the structural confinement of single atom Pd in Pd1/GDY leading to the much weaker C2H4-\u03c0 bonding interactions (65.1\u00a0kJ\u00a0mol\u22121) compared to the stronger C2H4 adsorption on the large size of Pd55 cluster (189.2\u00a0kJ\u00a0mol\u22121) [63]. The weaker C2H4-\u03c0 bonding interactions do not facilitate C2H4 activation and hydrogenation. The C3H6-\u03c0 bonding characteristics between C3H6 and V1/g-C3N4 catalyst were also obtained [64]. As a result, C2H4 hydrogenation (326.5\u00a0kJ\u00a0mol\u22121) is much difficult than C2H4 desorption (65.1\u00a0kJ\u00a0mol\u22121) on Pd1/GDY catalyst, the produced C2H4 will easily desorb from Pd1/GDY catalyst to become the dominant product. Meanwhile, compared to the large size of metal Pd55 cluster, the faster desorption rate of C2H4 on Pd1/GDY catalyst enhances C2H4 selectivity.As shown in Fig.\u00a07\n, the metal-support interaction (E\nMSI/kJ mol\u22121) of PdxMy/GDY catalyst showed that when E\nMSI\u00a0value\u00a0was weak (\u2212200\u223c\u2212400\u00a0kJ\u00a0mol\u22121) or strong (\u2212600\u223c\u2212800\u00a0kJ\u00a0mol\u22121), for example, Ag3/GDY (\u2212236.0), Au1/GDY (\u2212265.5), Ag2/GDY (\u2212321.5), Pd1Au1/GDY (\u2212322.6), Ag1/GDY (\u2212338.5), Pd1Au2/GDY (\u2212381.5), Pd1Ag1/GDY (\u2212400.7), Pd1Ni1/GDY (\u2212604.0), Ni2/GDY (\u2212654.4), Pd3/GDY (\u2212654.8), Pd2Ni1/GDY (-667.8) and Ni3/GDY (\u2212772.1), these catalysts could not adsorb C2H2 preferentially or were not conducive to C2H4 formation; however, when the values of E\nMSI were moderate (\u2212400\u223c\u2212600\u00a0kJ mol\u22121), the catalyst could adsorb C2H2 preferentially and realize C2H4 formation, for example, Au3/GDY (\u2212418.3), Cu2/GDY (\u2212440.6), Pd1Ag2/GDY (\u2212463.9), Pd1/GDY (\u2212481.1), Pd2Au1/GDY (\u2212485.5), Pd1Cu2/GDY (\u2212561.4) and Pd2Cu1/GDY (\u2212595.1) catalysts are all favorable for C2H4 production.As listed in Table 2, Bader charge indicates that the Pd, Cu, Ag, Au or Ni atoms transfer electrons to the C atom of GDY. For PdxMy/GDY with better C2H4 selectivity, when the average charge of metal atoms is small, such as Au2/GDY (0.032), Pd1Cu2/GDY (0.112) and Au3/GDY (0.135), these catalysts have low C2H4 formation activity of 8.71\u00a0\u00d7\u00a010\u221228, 3.45\u00a0\u00d7\u00a010\u22121 and 2.16\u00a0\u00d7\u00a010\u22121 s\u22121 site\u22121, respectively. When the average charge of metal atoms is large, such as Cu3/GDY (0.349) and Cu2/GDY (0.387), both catalysts also have low C2H4 formation activity of 7.63\u00a0\u00d7\u00a010\u22124 and 2.94\u00a0\u00d7\u00a0106 s\u22121 site\u22121, respectively. Only when the average charge of metal atoms is moderate, such as Pd1Ag2/GDY (0.148), Pd1Ni2/GDY (0.156), Pd2Cu1/GDY (0.177), Pd2Au1/GDY (0.228) and Pd1/GDY (0.277), these catalysts have higher C2H4 formation activity of 1.55\u00a0\u00d7\u00a0108, 3.63\u00a0\u00d7\u00a01010, 4.25\u00a0\u00d7\u00a0109 and 2.26\u00a0\u00d7\u00a0107 s\u22121 site\u22121, respectively. Thus, the average charge amount of metal atom is closely related to C2H4 formation activity, namely, the average charge amount of metal atom is less or more, C2H4 formation activity is low; whereas the average charge amount of metal atom is moderate, C2H4 formation activity is high.On the other hand, Huang et\u00a0al. [65] implied that coke formation on the single-atom Pd1/C3N4 is markedly inhibited compared to Pd NP catalysts in C2H2 hydrogenation, the geometric effect improved coking-resistance. The oligomerization of C2H2 can be avoided on Pd1/ND@G catalyst, which is attributed to the pyramidal geometry between Pd and C atoms [15]. Indeed, C2H2 polymerization to form green oil or coke requires multiple adjacent adsorption sites that cannot be available for Pd single atom, thus, coke formation is suppressed on the Pd single atoms compared to that on the Pd NP catalyst. Similarly, Pd1/GDY catalyst in the present study only has a single active site, which can suppress the green oil or coke formation due to its unique geometric effect. Further, previous DFT studies [6] have revealed that Pd (111) with surface or subsurface C atom make the shift of the d-projected density of states of the surface Pd atoms to lower energy level, which weakens C2H4 adsorption compared to those on clean Pd (111) surface, meanwhile, the activity of Pd (111) surface slightly increase in the presence of subsurface carbon species. As a result, C2H4 desorption becomes easier, and the selectivity of C2H4 increase in the presence of surface and subsurface carbon. The projected density of states (pDOS) plots for the d-orbitals of Pd atom on the Pd (111)-surface C, Pd (111)-subsurface C, Pd1/GDY and Pd (111) catalysts are calculated, as shown in Fig.\u00a0S21, similar to Pd (111) in the presence of surface or subsurface carbon species, compared to the pure Pd (111) surface, the shift of the d-projected density of states for surface Pd atoms to lower energy level also occur on Pd1/GDY, which also weaken C2H4 adsorption to increases its selectivity.To deeply illustrate the excellent activity and selectivity of Pd1/GDY catalyst, the comparisons for the activity and selectivity of C2H4 formation between Pd1/GDY and the reported catalysts in the literatures were carried out. On Pd1/GDY catalyst, C2H4 selectivity was 261.4\u00a0kJ\u00a0mol\u22121, both C2H2 + H\u2192C2H3 and C2H3 + H\u2192C2H4 reactions have the activation barriers of 36.0 and 1.0\u00a0kJ\u00a0mol\u22121, respectively; the overall barrier of C2H2 + 2H\u2192C2H4 was 38.4\u00a0kJ mol\u22121.As shown in Fig.\u00a08\n(a), for the Pd-based intermetallic compounds (IMCs), the single atom Pd active site can be completely isolated by the second metal, Zhou et\u00a0al. [66] found that PdZn IMCs had highly active and selective for C2H2 semi-hydrogenation, DFT results showed that C2H4 selectivity was 36.0\u00a0kJ\u00a0mol\u22121, meanwhile, the activation barriers of C2H2 + H\u2192C2H3 and C2H3 + H\u2192C2H4 were 55.0 and 56.0\u00a0kJ\u00a0mol\u22121, respectively. Feng et\u00a0al. [67] found that the single atom Pd active site in PdIn IMCs had C2H4 selectivity of 34.0\u00a0kJ\u00a0mol\u22121; the activation barriers of C2H2 + H\u2192C2H3 and C2H3 + H\u2192C2H4 were 36.0 and 34.0\u00a0kJ\u00a0mol\u22121, respectively. Sandoval et\u00a0al. [68] calculated that the activation barriers of C2H2 + H\u2192C2H3 and C2H3 + H\u2192C2H4 on PdGa IMCs were 70.0 and 75.0\u00a0kJ\u00a0mol\u22121, respectively. Hence, compared to Pd1/GDY catalyst in this study, the activity and selectivity of C2H4 formation over these reported Pd-based IMCs catalysts are lower.As shown in Fig.\u00a08(b), for the single atom Pd doped into metal surface, Zhang et\u00a0al. [57,58] studied the hydrogenation of C2H2 on the single atom Pd-doped Cu(111), Cu(211) or Cu2O(111) surfaces, C2H4 selectivity and the overall activation barrier of C2H2 + 2H\u2192C2H4 on Pd1/Cu(111) are 42.6 and 47.5\u00a0kJ mol\u22121, respectively; those on Pd1/Cu(211) are 36.4 and 78.8\u00a0kJ mol\u22121, respectively; on Pd1/Cu2O(111), C2H2\u00a0is easily over-hydrogenated to ethane via CHCH3 intermediate. On the other hand, for the single Pd atom doped-Cu13, Cu38 or Cu55 clusters [69], C2H4 is easily hydrogenated to ethane. Wang et\u00a0al. [70] obtained that the single atom Pd-doped Ag surface can facilitate the over-hydrogenation of C2H4 to produce ethane. Yang et\u00a0al. [6] found that when the surface coverage of Cu, Ag or Au on Pd (111) is increased to present the single atom Pd, C2H4 was prone to be over-hydrogenated to ethane. Yang et\u00a0al. [71] showed that trimetallic PdAg2Au/Pd (111) surface showed C2H4 selectivity of 24.0\u00a0kJ mol\u22121, the activation barriers of C2H2 + H\u2192C2H3 and C2H3 + H\u2192C2H4 were 41.0 and 58.0\u00a0kJ mol\u22121, respectively. Thus, the selectivity and activity of C2H4 formation over the catalysts doping the single atom Pd into metal surface are still lower than those on Pd1/GDY catalyst.As shown in Fig.\u00a08(c), for the supported single atom Pd catalysts, Wei et\u00a0al. [72] experimentally prepared the thermodynamically stable Pd1\u2013N4 structure with the single atom Pd anchored on the defects of nitrogen-doped carbon, then, DFT results show that C2H4 selectivity is 91.0\u00a0kJ mol\u22121; the activation barriers of C2H2 hydrogenation and C2H3 hydrogenation are 37.0 and 94.0\u00a0kJ mol\u22121, respectively. The experiments by Huang et\u00a0al. [15] prepared the atomically dispersed Pd over a defective nanodiamond-graphene (Pd1/ND@G catalyst), DFT results show that C2H4 selectivity is 51.0\u00a0kJ mol\u22121; the activation barriers of C2H2 + H\u2192C2H3 and C2H3 + H\u2192C2H4 are 110.0 and 85.0\u00a0kJ mol\u22121, respectively. Huang et\u00a0al. [13] experimentally prepared Cu1/ND@G catalyst, which also exhibits excellent catalytic performance for C2H2 + 2H\u2192C2H4, and DFT results show that C2H4 selectivity is only 19.0\u00a0kJ mol\u22121; the activation barrier of C2H2 hydrogenation to C2H3 is 136.0\u00a0kJ mol\u22121. Zhou et\u00a0al. [73] found that atomically dispersed Pd on nitrogen-doped graphene (Pd1/N-graphene) exhibits better activity and selectivity for C2H2 + 2H\u2192C2H4, and DFT results show that C2H4 selectivity is 88.0\u00a0kJ mol\u22121, which is much lower than that on Pd1/GDY catalyst. Further, C2H2 semi-hydrogenation on the Pd1/SVG catalyst with the single atom Pd supported by a single vacancy graphene (SVG) is calculated in this study (see details in Figs. S16 and S17), the results show that C2H4 is easily over-hydrogenated to C2H5 in kinetically instead of its desorption (10.6 vs. 43.2\u00a0kJ\u00a0mol\u22121). Similarly, the selectivity and activity of C2H4 formation over these supported single atom Pd catalysts reported in the literatures are still lower than those on Pd1/GDY catalyst.Based on above analysis, surprisingly, we found that GDY supported single atom Pd catalyst (Pd1/GDY) in this study should so far provide the best selectivity and activity toward C2H4 formation in C2H2 semi-hydrogenation in comparison with other types of the single atom Pd or Cu catalysts previously reported in the literatures. Moreover, the experiments by Qi et\u00a0al. [74] found that the direct oxidation\u2013reduction reaction of GDY and PdCl4\n2\u2212 could realize the chemical deposition of Pd nanoparticles on GDY, which provided an important guidance for the preparation of GDY supported single atom Pd catalyst in the experiment.In summary, the activity and selectivity of a series of the designed PdxMy (M\u00a0=\u00a0Cu, Ag, Au, Ni; x + y\u00a0=\u00a01\u20133) clusters anchored on GDY (PdxMy/GDY catalysts) in C2H2 semi-hydrogenation have been fully examined using DFT calculations. Our results show that the activity and selectivity of C2H4 formation in C2H2 semi-hydrogenation on the PdxMy/GDY catalysts strongly depend on the composition and size of supported metal cluster, which has the relationship with the metal-support interaction of PdxMy/GDY catalysts and electronic properties. The supported metal is single atom, only Pd1/GDY is favorable for C2H4 formation; the supported metal is two atoms cluster, only Cu2/GDY and Au2/GDY are favorable for C2H4 formation; the supported metal is three atoms cluster, seven kinds of the catalysts including Cu3/GDY, Au3/GDY, Pd1Cu2/GDY, Pd2Cu1/GDY, Pd1Ag2/GDY, Pd2Au1/GDY and Pd1Ni2/GDY are favorable for C2H4 formation. Moreover, aiming at realizing the balance between the activity and selectivity of C2H4 formation, the metal-support interaction of PdxMy/GDY catalysts and the average charge amount of metal atoms should be maintain in a moderate range.Surprisingly, the comparisons among the catalysts considered in this study and previously reported in the literature showed for the first time that Pd1/GDY catalyst exhibits the high catalytic activity and selectivity toward C2H4 formation in C2H2 semi-hydrogenation. The high activity of Pd1/GDY is ascribed to Pd inherent properties toward C2H2 hydrogenation, however, the high C2H4 selectivity is attributed to the structural confinement of single atom Pd in Pd1/GDY catalyst leading to the much weaker C2H4-\u03c0 bonding interactions, which is not favorable for C2H4 activation and hydrogenation, thus, the faster desorption rate of the produced C2H4 on Pd1/GDY catalyst than its hydrogenation rate enhanced C2H4 selectivity in C2H2 semi-hydrogenation. The obtained results could provide good clues for designing and screening out the potential catalysts with GDY supported small sizes of metal clusters for selective hydrogenation of alkanes.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work is financially supported by the National Natural Science Foundation of China (No. 21776193 and 21736007) and U.S. NSF-sponsored NCAR-Wyoming Supercomputing Center (NWSC).The following is the Supplementary data to this article:The calculations methods of metal-support interaction and Gibbs free energy, C2H\nx\n (x\u00a0=\u00a02\u20135) and H adsorption, the energy profile of C2H2 semi-hydrogenation on the PdxMy/GDY catalysts with better C2H4 selectivity and Pd1/SVG catalyst, as well as the calculations of C2H4 formation activity on PdxMy/GDY catalysts are described.\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.gee.2020.10.020.", "descript": "\n                  C2H2 semi-hydrogenation has been widely applied in industry to eliminate trace C2H2 from C2H4 feed. C2H2 semi-hydrogenation to C2H4 on a series of the newly designed catalysts, graphdiyne (GDY) as a new carbon allotrope supported different sizes of PdxMy clusters (PdxMy/GDY, M\u00a0=\u00a0Cu, Ag, Au, Ni; x+y\u00a0=\u00a01\u20133), were studied using DFT calculations. The results found that C2H2 semi-hydrogenation to C2H4 on PdxMy/GDY catalysts exhibits that both the activity and selectivity greatly depend on the composition and size of PdxMy/GDY catalysts. Surprisingly, our results for the first time discovered the Pd1/GDY catalyst with GDY supported the single atom Pd that presents the best selectivity and activity toward C2H4 formation compared to the previously reported catalysts so far in C2H2 semi-hydrogenation. This study would provide a theoretical clue for designing and screening out the potential catalysts with GDY supported small sizes of PdxMy and other metal clusters in C2H2 hydrogenation.\n               "}
{"full_text": "No data was used for the research described in the article.Brunauer-Emmet-TellerBarrett-Joyner-HalendaComputational fluid dynamicsCatalytic pyrolysisDistributed activation energyDifferential friedman\nDelonix regia\nDifferential thermogravimetryEnvironmental protection agencyHigher heating valueInternational confederation for thermal analysis and calorimetryIndices for pyrolysis performanceKissinger-akahira-sunoseKinetic factorsSodium-Y zeoliteOzawa-flynn-wallPlatinum (10\u00a0wt %) on activated carbonSurface areaStarinkThermogravimetric analyzerTitanium oxideThermodynamic parametersZinc oxideFlammability indexBurnout indexIgnition indexDevolatilization indexActivation energyConversion-dependent reaction model (differential)Conversion-dependent reaction model (integral)Change in Gibbs free energyPlanck's constantChange in enthalpyBoltzmann constantFrequency factorReaction rate constantMass of catalytic DR mixture at the final timeMass of catalytic DR mixture at the initial timeMass of catalytic DR mixture at timeReaction orderTemperature approximation (integral)Universal gas constantMaximum decomposition rateMean decomposition rateCombustion indexChange in entropyAbsolute temperatureTimeBurnout temperatureBurnout timeThermogram decomposition temperatureIgnition temperatureIgnition timeMaximum decomposition temperatureMaximum decomposition timeTemperature interval at half of R\n\np\n\nTime interval at half of the half of R\n\np\n\nTemperature at \u03b1Reaction pathwayExtent of conversionHeating rateGlobal diminishment of fossil fuels, as well as intensifying environmental pollution, necessitate exploring alternative energy sources. Biomass has received a lot of attention as the only abundant renewable resource that can be used to produce sustainable biofuels. As per the EPA [1] data, the global production of lignocellulosic biomass solid waste in 2020 was 18.1\u00a0MT. Out of this huge quantity, 17.1% is used for energy recovery, 15.7% is employed for combustion, and 67.2% is utilized for other purposes such as disposal and landfills.\nDelonix regia (DR) is a common lignocellulosic biomass found in the Fabaceae family [2], and its residue is used to produce bio-oil by several researchers [3\u20137]. However, biomass conversion to bio-oil via the pyrolysis process still faces several challenges due to the thermal instability of bio-oil, and its higher oxygen and viscosity contents. As a result, a catalyst is required to enhance the properties for a better quality of bio-oil [8]. In addition, other pyrolytic processes variables such as residence time, temperature, the ratio of biomass by a catalyst, and type of catalyst significantly influence the generation of higher yields and selectivity of bio-oil during the catalytic pyrolysis of the biomass [9\u201311].The following three approaches are used to employ a catalyst in the pyro\u2013catalytic process: first, mixing the catalyst with the raw material; second, introducing the catalyst at the top of the reactor to facilitate vapor\u2013catalyst contact; and third, placing the catalyst into a secondary reactor after primary (pyrolizer) reactor. The first two approaches are known as in\u2013situ, whereas the third is known as ex\u2013situ pyro\u2013catalytic processes; each approach has a distinct effect [12\u201314]. Thus, selecting a proper catalyst to enhance the pyrolysis process is an alternate method for minimizing the total energy utilization of the process [9\u201311]. The conversion of biomass into pyrolytic products are determined mainly by the kinetic rates of the reactions that occur during pyrolysis. It has been demonstrated that an accurate kinetic technique is vital to design an effective pyrolysis process [15,16].Thermogravimetric analysis (TGA) is an analytic tool employed for assessing the pyrolytic degradation composition of heterogeneous biomass [12\u201314]. Understanding the pyrolytic decomposition of biomass is highly significant due to the kinetic factors (activation energy, E\n\n\u03b1\n; frequency factor, k\n\no\n; and reaction pathway, Z\n\u03b1\n) being inherently associated with the degradation mechanisms [9\u201311]. TGA can quantify the mass loss caused by devolatilization during the pyrolytic disintegration of biomass at a given heating rate in relation to temperature and time [8]. Furthermore, the first derivate of the TG profiles (dm/dt), commonly termed differential thermogravimetry (DTG), can be used to calculate the maximum rate of reaction [8]. Numerous kinetic and thermodynamic investigations [17\u201324] for biomass pyrolysis employ a first order reaction method. In this method, the parameters such as the extent of conversion (\u03b1), frequency factor (k\n\no\n), and reaction rate constant k(T) are reported by using the TGA data [17\u201324]. Hence kinetic factors from TGA can be measured accurately. TGA\u2013derived kinetic factors serve as the basis for the modeling and optimization of the pyrolysis process. Consequently, kinetic factors knowledge is essential for comprehending and developing the catalytic pyrolysis process. Furthermore, the International Confederation for Thermal Analysis and Calorimetry (ICTAC) Kinetics Committee suggested multiple iso\u2013conversional techniques for evaluating kinetic factors using TGA data [25,26]. The most used iso\u2013conversional techniques throughout this work are Differential Friedman (DFM), Kissinger\u2013Akahira\u2013Sunose (KAS), Ozawa\u2013Flynn\u2013Wall (OFW), Starink (STK), and Distributed Activation Energy (DAE). Using iso\u2013conversional techniques, it is possible to precisely predict the kinetic factors and thermodynamics of the biomass pyrolysis [17\u201324,27\u201331]. Furthermore, understanding of thermodynamic properties including changes in enthalpy and Gibbs free energy are critical for establishing the feasibility and energy requirement of biomass pyrolysis [32\u201341].The literature study indicates that no studies have been published that examined kinetic factors and thermodynamic parameters for catalytic pyrolysis (CP) of DR biomass employing iso-conversional techniques. However, various feedstock, including wheat bran [42]; poplar sawdust [30]; pine needle [43]; coconut copra and rice husk [44]; and torrified bamboo [45], have been explored with different catalysts, respectively, such as ZSM-5, Pt/C and Pd/C; Fe\u2013Ni/ZSM-5; Ni/Al2O3; Ni\u2013Ce/Al2O3; and HZSM-5. Furthermore, only a few investigations on CP of DR biomass for the production of bio-oil have been reported [3\u20137]; however, they have not reported the corresponding kinetics and thermodynamics of the process. According to the aforementioned rudimentary information, the CP of DR biomass for kinetics and thermodynamic analysis has yet to be extensively researched; therefore, the current research focuses on this investigation.Briefly, this work involved the systematic exploration of kinetic factors and thermodynamics during the catalytic pyrolysis of Delonix regia (DR) biomass over three different catalysts (Na\u2013Y zeolite, 10\u00a0wt % of Pt/C, and TiO2\u2013ZnO) using a thermogravimetric analyzer (TGA). The TGA of DR biomass is conducted with each catalyst loading (30, 20, and 10\u00a0wt %) at multiple heating rates (5, 10, 20, 35, and 55\u00a0\u00b0C min\u22121), and associated kinetic factors are ascertained that includes activation energy (E\n\n\u03b1\n) determined employing five iso\u2013conversional techniques (DFM, OFW, KAS, STK, and DAE) followed by frequency factor (k\n\no\n). Further, the variations of reaction pathways are evaluated using Criado's master plot technique. Finally, the thermodynamic properties of catalytic pyrolysis of DR biomass are thoroughly examined. This report shall be critical for elucidating the influence of different types of catalysts (Na\u2013Y, Pt/C, and TiO2\u2013ZnO) for biofuel generation, supporting scientific sources for pyrolyzer modeling, and optimizing the pyrolysis process.Material including Delonix regia (DR) biomass was sourced from solid wastes at IIT Guwahati. Catalysts were procured: zeolite Na\u2013Y (SAR-5.1: 1) from alfa aesar in the USA, platinum on activated carbon (10\u00a0wt %) from sigma-aldrich in the USA, titanium oxide (>99.5\u00a0wt %), and zinc oxide (>99.9\u00a0wt %) from Sisco research laboratories in India.The methodologies of preparation and results of physico\u2013chemical properties, including grinding/crushing, washing, drying, and higher heating value, and ultimate/proximate analysis of DR material were discussed elsewhere [3\u20137,46,47].Pore size distribution such as size and volume of catalysts were determined using an N2 sorption (Model: Tristar II, Make: Micromeritics, USA) analyzer. Before the analysis catalysts (three) were degassed for 6\u00a0h at 180\u00a0\u00b0C in a vacuum to eliminate moisture and volatiles. The surface area (SA) of each catalyst was calculated utilizing the Brunauer-Emmet-Teller (BET) technique. The Barrett-Joyner-Halenda (BJH) technique was employed to calculate pore size distribution. The pore size (nm) and pore volume (cm3 g\u22121) were 4.91 and 0.07; 5.64 and 0.07; and 3.67 and 0.51, respectively, for Na\u2013Y, Pt/C, and TiO2\u2013ZnO.The catalytic pyrolysis (CP) of DR biomass experiments was conducted in a thermogravimetric (Model: TG209F1, Make: Netsch, Germany) analyzer. In all experiments, DR material and catalysts were homogenized by mechanically blending them with the aid of mortar and pestle at various catalyst ratios (30, 20, and 10\u00a0wt %). For all experiments, approximately 6\u00a0mg of sample was used and heated from temperatures ranging between 25 and 1000\u00a0\u00b0C. For all experiments, 40\u00a0mL\u00a0min\u22121 of nitrogen gas was maintained and the heating rates (\u03b2) were: 5, 10, 20, 35, and 55\u00a0\u00b0C min\u22121. TGA experimental runs were undertaken at \u03b2 ranging from 5 to 55\u00a0\u00b0C min\u22121 to explore the CP of DR biomass between slow and fast pyrolysis range.In this study, kinetic factors (KF) including activation energy E\n\n\u03b1\n (kJ mol\u22121), frequency factor, k\n\no\n (s\u22121), and pathway of reaction (Z\n\n\u03b1\n) for catalytic pyrolysis (CP) of DR biomass were evaluated. All calculations by the following techniques were conducted using MATLAB (Version: R2021a): Differential Friedman (DFM), Ozawa-Flynn-Wall (OFW), Kissinger-Akahira-Sunose (KAS), Starink (STK), and Distributed Activation Energy (DAE).Calculations of KF of catalytic pyrolysis (CP) of Delonix regia (DR) were provided below:\n\n\n\nD\ne\nl\no\nn\ni\nx\n\nR\ne\ng\ni\na\n\n\u2192\n\nN\na\n\u2212\nY\n,\nP\nt\n/\nC\n,\n\n\nT\ni\nO\n\n2\n\n\u2212\nZ\nn\nO\n\n\nV\no\nl\na\nt\ni\nl\ne\ns\n+\nB\ni\no\nc\nh\na\nr\n\n\n\n\nThe following was the conversion rate of CP of DR:\n\n(1)\n\n\n\n\nd\n\u03b1\n\n\nd\nt\n\n\n=\nk\n\n(\nT\n)\n\nf\n\n(\n\u03b1\n)\n\n\n\n\nwhere k(T) denotes reaction rate constant, and f(\u03b1) denotes conversion\u2013dependent reaction model (differential).Extent of conversion (\u03b1) of CP of DR was given by:\n\n(2)\n\n\n\u03b1\n=\n\n\n\nm\ni\n\n\u2212\n\nm\nt\n\n\n\n\nm\ni\n\n\u2212\n\nm\nf\n\n\n\n\n\n\nwhere m\n\ni\n, m\n\nf\n, and m\n\nt\n represent the mass of catalytic DR mixture at the initial, final, and any time respectively.Reaction rate constant for CP of DR was defined by Arrhenius Eq.:\n\n(3)\n\n\nk\n\n(\nT\n)\n\n=\n\nk\n0\n\n\nexp\n\n(\n\n\u2212\n\n\nE\n\u03b1\n\n\nR\nT\n\n\n\n)\n\n\n\n\nwhere k\n\no\n, R, T, and E\n\n\u03b1\n denotes frequency factor (s\u22121), universal gas constant (J mol\u22121 K\u22121), absolute temperature (K), and activation energy (kJ mol\u22121).Making use of Eqs. (1) and (3) gave following equation:\n\n(4)\n\n\n\n\nd\n\u03b1\n\n\nd\nt\n\n\n=\n\nk\no\n\n\nexp\n\n(\n\n\u2212\n\n\nE\n\u03b1\n\n\nR\nT\n\n\n\n)\n\nf\n\n(\n\u03b1\n)\n\n\n\n\n\nNow, consider introducing the rate of heating (\u03b2) as:\n\n(5)\n\n\n\u03b2\n=\n\n\nd\nT\n\n\nd\nt\n\n\n=\n\n\nd\nT\n\n\nd\n\u03b1\n\n\n\u00d7\n\n\nd\n\u03b1\n\n\nd\nt\n\n\n\n\n\n\nFrom Eqs. (4) and (5):\n\n(6)\n\n\ng\n\n(\n\u03b1\n)\n\n=\n\n\n\u222b\n0\n\u03b1\n\n\n\n\nd\n\u03b1\n\n\nf\n\n(\n\u03b1\n)\n\n\n\n=\n\n\n\n\nk\no\n\n\u03b2\n\n\n\n\u222b\n\nT\n0\n\nT\n\n\nexp\n\n(\n\n\u2212\n\n\nE\n\u03b1\n\n\nR\nT\n\n\n\n)\n\nd\nT\n\n\n=\n\n\n\nk\no\n\n\nE\n\u03b1\n\n\n\n\u03b2\nR\n\n\n\n\n\u222b\nx\n\u221e\n\n\n\nu\n\n\u2212\n2\n\n\n\n\nexp\n\n\u2212\nu\n\n\nd\nu\n\n\n=\n\n\n\nk\no\n\n\nE\n\u03b1\n\n\n\n\u03b2\nR\n\n\np\n\n(\nx\n)\n\n\n\n\nWhere, \n\nx\n=\n\n\nE\n\u03b1\n\n\nR\nT\n\n\n\n.The reaction model (integral) expression was Eq. (6), which does not have a numerical solution; nevertheless, it can be approximated using multiple iso-conversional techniques.Iso\u2013conversional techniques: DFM [48], OFW [49,50], KAS [51], STK [52], and DAE [53] were utilized for evaluating KF for CP of DR biomass.The rearranged mathematical expressions for various iso\u2013conversional techniques were as follows:\n\n(7)\n\n\nln\n\n[\n\n\u03b2\n\n(\n\n\nd\n\u03b1\n\n\nd\nT\n\n\n)\n\n\n]\n\n=\nln\n\n[\n\n\nk\no\n\n\n\nf\n\n(\n\u03b1\n)\n\n\nn\n\n\n]\n\n\u2212\n\n\nE\n\u03b1\n\n\nR\n\nT\n\u03b1\n\n\n\n\nD\nF\nM\n\n\n\n\n\n\n(8)\n\n\nln\n\n(\n\n\u03b2\n\nT\n\u03b1\n2\n\n\n)\n\n=\nln\n\n[\n\n\n\nk\no\n\nR\n\n\n\nE\n\u03b1\n\ng\n\n(\n\u03b1\n)\n\n\n\n]\n\n\u2212\n\n\nE\n\u03b1\n\n\nR\n\nT\n\u03b1\n\n\n\n\nK\nA\nS\n\n\n\n\n\n\n(9)\n\n\nln\n\u2061\n\n(\n\u03b2\n)\n\n=\nln\n\n[\n\n\n\nk\no\n\n\nE\n\u03b1\n\n\n\nR\ng\n\n(\n\u03b1\n)\n\n\n\n]\n\n\u2212\n5.331\n\u2212\n1.0516\n\n\nE\n\u03b1\n\n\nR\n\nT\n\u03b1\n\n\n\n\nO\nF\nW\n\n\n\n\n\n\n(10)\n\n\nln\n\u2061\n\n(\n\n\u03b2\n\nT\n\u03b1\n1.92\n\n\n)\n\n=\nln\n\n[\n\n\n\nk\no\n\n\nR\n0.92\n\n\n\n\nE\n\u03b1\n0.92\n\ng\n\n(\n\u03b1\n)\n\n\n\n]\n\n\u2212\n0.312\n\u2212\n1.0008\n\n\nE\n\u03b1\n\n\nR\n\nT\n\u03b1\n\n\n\n\nS\nT\nK\n\n\n\n\n\n\n(11)\n\n\nln\n\n(\n\n\u03b2\n\nT\n\u03b1\n2\n\n\n)\n\n=\nln\n\n[\n\n\n\nk\no\n\nR\n\n\nE\n\u03b1\n\n\n]\n\n+\n0.6075\n\u2212\n\n[\n\n\nE\n\u03b1\n\n\nR\n\nT\n\u03b1\n\n\n\n]\n\nD\nA\nE\n\n\n\n\nVarious approximations such as Doyle, Miura and Maki, Perez\u2013Maqueda, and Starink, \n\np\n\n(\nx\n)\n\n=\n\nx\n\n\u2212\n2\n\n\n\n\nexp\n\n\u2212\nx\n\n\n\n, \n\np\n\n(\nx\n)\n\n=\n\nexp\n\n(\n\n\u2212\n1.0516\nx\n\u2212\n5.331\n\n)\n\n\n\n, \n\np\n\n(\nx\n)\n\n=\n\nexp\n\n\n(\n\n\u2212\n1.0008\nx\n\u2212\n0.312\n\n)\n\n\nx\n1.92\n\n\n\n\n, and \n\np\n\n(\nx\n)\n\n=\n0.6075\n\u2212\nx\n\n were utilized to get Eqs. (8 - 11) respectively [52\u201355].\nReaction pathway: The reaction pathway (Z\n\n\u03b1\n) of CP of DR was analysed by Criado's technique [56]. Table 1\n included the numerous reaction pathways expressions, which comprise models of differential (f(\u03b1)) and integral (g(\u03b1)) categories. The theoretical and experimental equations (12\u201315) of Z\n\n\u03b1\n were provided below:\n\n(12)\n\n\nZ\n\n\n(\n\u03b1\n)\n\n\nT\nh\ne\no\n\n\n=\nf\n\n\n(\n\u03b1\n)\n\n\nT\nh\ne\no\n\n\n\u00d7\ng\n\n\n(\n\u03b1\n)\n\n\nT\nh\ne\no\n\n\n\n\n\n\n\n\n(13)\n\n\nZ\n\n\n(\n\u03b1\n)\n\nExp\n\n=\n\n(\n\n\nd\n\u03b1\n\n\nd\nt\n\n\n)\n\n\u00d7\nexp\n\n(\n\n\nE\n\u03b1\n\n\nR\nT\n\n\n)\n\n\u00d7\n\n\n\u222b\n\nT\n0\n\nT\n\n\nexp\n\n(\n\n\u2212\n\n\nE\n\u03b1\n\n\nR\nT\n\n\n\n)\n\nd\nT\n\n\n\n\n\n\n\n\n(14)\n\n\nZ\n\n\n(\n\u03b1\n)\n\nExp\n\n=\n\n(\n\n\nd\n\u03b1\n\n\nd\nt\n\n\n)\n\n\u00d7\n\n\nE\n\u03b1\n\nR\n\n\u00d7\nexp\n\n(\n\n\nE\n\u03b1\n\n\nR\nT\n\n\n)\n\n\u00d7\np\n\n(\nx\n)\n\n\n\n\n\n\n\n(15)\n\n\nZ\n\n\n(\n0.5\n)\n\nExp\n\n=\n\n(\n\n\nd\n\u03b1\n\n\nd\nt\n\n\n)\n\n\u00d7\n\n\nE\n0.5\n\nR\n\n\u00d7\nexp\n\n(\n\n\nE\n0.5\n\n\nR\nT\n\n\n)\n\n\u00d7\np\n\n(\nx\n)\n\n\n\n\nFor above, \n\np\n\n(\nx\n)\n\n=\n0.00484\n\nexp\n\u2061\n\n(\n\n\u2212\n1.0516\nx\n\n)\n\n\n\nThermodynamic parameters (TP) of CP of DR, comprising changes in Gibbs free energy (\u0394G), enthalpy (\u0394H), and entropy (\u0394S) were evaluated utilizing the following governing equations:\n\n(16)\n\n\n\nk\no\n\n=\n\n\n\u03b2\n\u00d7\n\nE\n\u03b1\n\n\u00d7\n\ne\n\n(\n\n\nE\n\u03b1\n\n\n\nR\nT\n\nm\n\n\n)\n\n\n\n\nR\n\nT\nm\n2\n\n\n\n\n\n\n\n\n\n(17)\n\n\n\u0394\nG\n=\n\nE\n\u03b1\n\n+\nR\n\nT\nm\n\n\u00d7\nln\n\n(\n\n\n\nK\nB\n\n\nT\nm\n\n\n\n\nh\nk\n\no\n\n\n)\n\n\n\n\n\n\n\n(18)\n\n\n\u0394\nH\n=\n\nE\n\u03b1\n\n\u2212\nR\n\nT\nm\n\n\n\n\n\n\n\n(19)\n\n\n\u0394\nS\n=\n\n\n\u0394\nH\n\u2212\n\u0394\nG\n\n\nT\nm\n\n\n\n\n\nwhere T\n\nm\n\u00a0=\u00a0thermogram decomposition temperature (K), K\n\nB\n = Boltzmann (1.381\u00a0\u00d7\u00a010\u221223\u00a0J\u00a0K\u22121) constant, and h\u00a0=\u00a0Planck's (6.626\u00a0\u00d7\u00a010\u221234\u00a0Js) constant.For the CP of DR, various indices including flammability index (C, K\u22122 min\u22121), burnout index (D\n\nb\n, min\u22124), ignition index (D\n\ni\n, min\u22123), devolatilization index (D\n\nv\n, K\u22123 min\u22121), and combustion index (S, K\u22123 min\u22122) were determined. The following equations Eq 20, 21, 22, 23 and 24 describe the definitions of these indices:\n\n(20)\n\n\nC\n=\n\n(\n\n\n\u2212\n\nR\np\n\n\n\nT\ni\n2\n\n\n)\n\n\n\n\n\n\n\n(21)\n\n\n\nD\ni\n\n=\n\n(\n\n\n[\n\n\u2212\n\nR\np\n\n\n]\n\n\n\nt\ni\n\n\u00d7\n\nt\np\n\n\n\n)\n\n\n\n\n\n\n\n(22)\n\n\n\nD\nb\n\n=\n\n(\n\n\n[\n\n\u2212\n\nR\np\n\n\n]\n\n\n\n\n\u0394\nt\n\n0.5\n\n\u00d7\n\nt\nb\n\n\u00d7\n\nt\np\n\n\n\n)\n\n\n\n\n\n\n\n(23)\n\n\nS\n=\n\n(\n\n\n\n[\n\n\u2212\n\nR\np\n\n\n]\n\n\u00d7\n\n[\n\n\u2212\n\nR\nv\n\n\n]\n\n\n\n\nT\ni\n2\n\n\u00d7\n\nT\nb\n\n\n\n)\n\n\n\n\n\n\n\n(24)\n\n\n\nD\nv\n\n=\n\n(\n\n\n\n[\n\n\u2212\n\nR\np\n\n\n]\n\n\u00d7\n\n[\n\n\u2212\n\nR\nv\n\n\n]\n\n\n\n\n\n\u0394\nT\n\n0.5\n\n\u00d7\n\nT\ni\n\n\u00d7\n\nT\np\n\n\n\n)\n\n\n\n\nwhere, R\n\np\n\u00a0=\u00a0maximum decomposition rate (wt. % min\u22121), T\n\ni\n\u00a0=\u00a0ignition temperature (K), t\n\ni\n\u00a0=\u00a0ignition time (min), t\n\np\n\u00a0=\u00a0maximum degradation time (min), \u0394t0.5\u00a0=\u00a0interval at half of the R\n\np\n, t\n\nb\n\u00a0=\u00a0burnout time (min), R\n\nv\n\u00a0=\u00a0mean decomposition rate (wt. % min\u22121), T\n\nb\n\u00a0=\u00a0burnout temperature (K), \u0394T0.5\u00a0=\u00a0temperature (K) interval at half of R\n\np\n, and T\n\np\n\u00a0=\u00a0maximum decomposition temperature (K). The logical process flow diagram of the current research was illustrated in Fig. 1\n. Whereas Fig. 2\n presented TGA and DTG distribution of DR biomass and three catalysts at 10\u00a0\u00b0C min\u22121 of heating rate. Finally, it should be noted that Eqs. (16)\u2013(19) are derived from and applicable only to unimolecular reactions but not to heteromolecular biomass samples although these equations are extensively used for the pyrolysis of biomass.\nFig. 3\n (A to C), Fig. 3 (D to F), and Fig. 3 (G to I) represents TG\u2013DTG pyrograms of catalytic pyrolysis (CP) of DR biomass using three different catalysts namely zeolite (Na\u2013Y), 10 wt % (Pt/C), and 1\u20131\u00a0wt. % (TiO2\u2013ZnO)respectively at 1:30, 1:20 and 1:10\u00a0wt %. From Fig. 3, it was noticed that the CP of DR biomass occurred in three zones. The primary zone was formed within the temperature range (25\u2013150\u00a0\u00b0C) which mainly liberates moisture and light molecular weight compounds. The second pyrolytic zone occurred at the temperature ranges of 150\u2013587\u00a0\u00b0C for 30\u00a0wt% Na\u2013Y; 150\u2013592\u00a0\u00b0C for 20\u00a0wt% Na\u2013Y; 150\u2013598\u00a0\u00b0C for 10\u00a0wt% Na\u2013Y; 150\u2013572\u00a0\u00b0C for 30\u00a0wt% TiO2\u2013ZnO; 150\u2013590\u00a0\u00b0C for 20\u00a0wt% TiO2\u2013ZnO; 150\u2013598\u00a0\u00b0C for 30\u00a0wt% TiO2\u2013ZnO; 150\u2013640\u00a0\u00b0C for 30\u00a0wt% Pt/C; 150\u2013632\u00a0\u00b0C for 20\u00a0wt% Pt/C; and 150\u2013610\u00a0\u00b0C for 10\u00a0wt% Pt/C. This was generally identified as an active pyrolytic zone with the decomposition of major compounds such as hemicellulose and cellulose [30,42,57]. Additionally, it was emphasized that lignin decomposition also starts in this zone. At the same time, a significant part of lignin decomposition was noted in the third zone at temperatures of >500\u00a0\u00b0C [46,58]. In the third stage, biomass degradation was slow and the lignin compound contributed such a trend. Furthermore, from Fig. 2, it was observed that at a heating rate (10\u00a0\u00b0C min\u22121) the catalysts were stable with 76.4\u00a0wt %, 56.2\u00a0wt %, and 98.8\u00a0wt % mass of Na\u2013Y, Pt/C and TiO2\u2013ZnO remaining, respectively, even after 800\u00a0\u00b0C.\nFig. 3 (A-C) represent the TG\u2013DTG pyrograms of CP of DR biomass with 30, 20, and 10\u00a0wt % load of Na\u2013Y. TG pyrograms between 5 and 55\u00a0\u00b0C min\u22121 showed mass loss of 53\u201351\u00a0wt % by using 30\u00a0wt % Na\u2013Y; 58\u201356\u00a0wt % by use of 20\u00a0wt % Na\u2013Y; and 62\u201360\u00a0wt % by use of 10\u00a0wt % Na\u2013Y along with DR in the pyrolyzer. DTG pyrograms by 5\u201355\u00a0\u00b0C min\u22121 showed the different maximum decomposition temperatures were 323\u2013362\u00a0\u00b0C by 30\u00a0wt % Na\u2013Y; 326\u2013366\u00a0\u00b0C by 20\u00a0wt % Na\u2013Y; and 326\u2013368\u00a0\u00b0C by 10\u00a0wt % Na\u2013Y, respectively. It was observed that as the catalyst (Na\u2013Y) loading increased, the mass loss decreased significantly and was comparable to other reports [30,59]. Moreover, an increasing heating rate causes variations of TG\u2013DTG pyrograms for non-catalytic pyrolysis, indicating an improper heat transfer between the particles [9,46]. Likewise, change in TG\u2013DTG pyrograms were observed for the catalytic pyrolysis. This change of pyrograms for a catalytic pyrolysis process also depends on the pore size characteristics, surface area and acid sites (combination of Lewis and Br\u00f8nsted sites) availability in zeolite catalyst [30,32,34,41,44,60].\nFig. 3 (D-F) represent the TG\u2013DTG pyrograms for CP of DR biomass with 30, 20, and 10\u00a0wt % of Pt/C. TG pyrograms between 5 and 55\u00a0\u00b0C min\u22121 indicated a mass loss of 55\u201353\u00a0wt % by 30\u00a0wt % Pt/C; 63\u201360\u00a0wt % by 20\u00a0wt % Pt/C; and 64\u201359\u00a0wt % by 10\u00a0wt % of Pt/C. It was observed as Pt/C ratio increased, mass loss decreased significantly. DTG pyrograms at 5\u201355\u00a0\u00b0C min\u22121 showed the different maximum decomposition temperatures: 334\u2013372\u00a0\u00b0C by 30\u00a0wt % Pt/C; 334\u2013372\u00a0\u00b0C by 20\u00a0wt % Pt/C; and 334\u2013373\u00a0\u00b0C by 10\u00a0wt % Pt/C, respectively. The significant reactions contributing to the weight loss were decarbonylation, decarboxylation, and dehydrogenation [30,42,61]. The lower catalytic activity of the Pt/c in comparison to the Na\u2013Y was due to lower surface area (55\u00a0m2\u00a0g\u22121) and lower pore volume [42,61].\nFig. 3 (G-I) represents the TG\u2013DTG pyrograms for CP of DR biomass at 30, 20, and 10 wt % of TiO2\u2013ZnO. TG pyrograms between 5 and 55\u00a0\u00b0C min\u22121 noticed a mass loss of 54\u201350\u00a0wt % by 30\u00a0wt % TiO2\u2013ZnO; 56\u201355\u00a0wt % by 20\u00a0wt % TiO2\u2013ZnO; and 63\u201362\u00a0wt % by 10\u00a0wt % TiO2\u2013ZnO respectively. It was ascertained as the catalyst (TiO2\u2013ZnO) ratio increased, the mass loss decreased significantly [30,62]. DTG pyrograms at 5\u201355\u00a0\u00b0C min\u22121 showed the different maximum decomposition temperatures: 332\u2013371\u00a0\u00b0C by 30\u00a0wt % TiO2\u2013ZnO; 332\u2013371\u00a0\u00b0C by 20\u00a0wt % TiO2\u2013ZnO; and 332\u2013370\u00a0\u00b0C by 10\u00a0wt % TiO2\u2013ZnO.At lower loadings (10 and 20\u00a0wt %), catalysts provided better thermal decomposition of biomass when compared to loading of 30\u00a0wt %. Such a trend can be attributed to the coke deposition on the surface of the catalyst [14,43]. Whereas for the non-catalytic pyrolysis process, the thermal degradation of DR biomass was higher as the remaining mass was 30.2\u201325.5\u00a0wt % against the increasing heating rates in the range of 5\u201355\u00a0\u00b0C min\u22121 [46].From surface area analysis of any catalyst, it was known that as the catalyst surface area diminishes, the number of active sites available for the reaction also reduces which states that the activity of the catalyst was less [30,63]. Additionally, the thermal decomposition of a catalyst indicates that the rigid walls of the pores offer the active sites significant strength, thus enhancing the catalyst's stability at high temperatures. This facilitates the usage of catalyst at higher temperatures (>1000\u00a0\u00b0C) without losing its properties [30,63]. Moreover, during the prolonged catalytic reactions, higher fractions (gaseous) were adsorbed over active sites by blocking pores that enhance the coke formation on the catalyst thereby reducing the number of active sites available for the reaction [30,62]. This also increases the activation energy (E\n\n\u03b1\n) required for initiating the reaction. Fig. 4\nA and Table 2\n present the variation of E\n\n\u03b1\n with zeolite Na\u2013Y at 30, 20, and 10\u00a0wt % loading. From the kinetic triplet analysis, according to DFM (Fig. 4A), it was observed that using Na\u2013Y catalyst, (mean E\n\n\u03b1\n (kJ mol\u22121) and k\n\no\n (min\u22121) factors) decreased in the range of (203 and 7.59\u00a0E+17) \u2013 (182 and 1.33\u00a0E+16) as the catalyst loading was varied from 30 to 10\u00a0wt %. When contrasted to the corresponding non\u2013catalytic pyrolysis of DR biomass at a mean E\n\n\u03b1\n of 195\u00a0kJ\u00a0mol\u22121 [46], Na\u2013Y catalyst process provided lower mean E\n\n\u03b1\n 181\u00a0kJ\u00a0mol\u22121. While the remaining four model\u2013free techniques (KAS, OFW, STK and DAE) yielded (mean E\n\n\u03b1\n (kJ mol\u22121) and k\n\no\n (min\u22121) factors) of (205 and 9.70\u00a0E+19) at 30\u00a0wt % Na\u2013Y; (191 and 1.96\u00a0E+19) at 20\u00a0wt % Na\u2013Y; and (181 and 1.83\u00a0E+17) at 10\u00a0wt % Na\u2013Y. It was observed that the mean E\n\n\u03b1\n factor decreased (181\u00a0kJ\u00a0mol\u22121) with decreasing Na\u2013Y ratio (10\u00a0wt%). Hence, Na\u2013Y with lower loading (10\u00a0wt%) was found to be suitable catalyst for catalytic pyrolysis of DR biomass.For Pt/C catalyst, the findings revealedthat mean E\n\n\u03b1\n (kJ mol\u22121) and k\n\no\n (min\u22121) factors decreased as (218 and 1.28\u00a0E+21) at 30\u00a0wt % Pt/C to (200 and 9.62\u00a0E+17) at 10\u00a0wt % Pt/C using model\u2013fitting (DFM) technique. While the remaining four model\u2013free techniques (KAS, OFW, STK and DAE) yielded (mean E\n\n\u03b1\n (kJ mol\u22121) and k\n\no\n (min\u22121) factors) as (211 and 4.73\u00a0E+18) at 30\u00a0wt % Pt/C; (204 and 1.06\u00a0E+18) at 20\u00a0wt % Pt/C; and (204 and 2.03\u00a0E+19) at 10\u00a0wt % Pt/C, respectively. It was observed that the mean E\n\n\u03b1\n factor decreased with decreasing Pt/C ratio [42,61]. Fig. 4B and Table 3\n show the alteration of E\n\n\u03b1\n with Pt/C catalyst at 30, 20, and 10\u00a0wt % loading. Further, Fig. 4C and Table 4\n demonstrate the variation of E\n\n\u03b1\n with TiO2\u2013ZnO (1\u20131\u00a0wt. %) catalyst at 30, 20, and 10\u00a0wt % loading. Findings showed that the (mean E\n\n\u03b1\n (kJ mol\u22121) and k\n\no\n (min\u22121) factors) decreased between (196 and 1.03\u00a0E+17) at 30\u00a0wt % TiO2\u2013ZnO to (191 and 7.34\u00a0E+16) at 20\u00a0wt % TiO2\u2013ZnO and increased to (201 and 3.41\u00a0E+17) at 10\u00a0wt % TiO2\u2013ZnO utilizing model\u2013fitting (DFM) technique. While remaining four model\u2013free techniques (KAS, OFW, STK and DAE) yielded (mean E\n\n\u03b1\n (kJ mol\u22121) and k\n\no\n (min\u22121) factors) of (192 and 3.72\u00a0E+17) at 30\u00a0wt % TiO2\u2013ZnO; (194 and 2.30\u00a0E+18) at 20\u00a0wt % TiO2\u2013ZnO; and (204 and 6.17\u00a0E+19) at 10\u00a0wt % TiO2\u2013ZnO. It was observed that the mean E\n\n\u03b1\n factor increased with decreasing TiO2\u2013ZnO ratio. The probable reasons for such alteration of mean E\n\n\u03b1\n factors can be attributed to the reasons stated earlier.The above findings clearly define the effect of the catalyst in determining the kinetic factors. Though the E\n\n\u03b1\n factor during pyrolysis increased at conversion (\u03b1) 0.1 to 0.4, there was a slight decrease at a conversion of 0.5\u20130.6 and then increased up to 0.7. This was due to the hemicellulose compounds degradation in the conversion range of 0.1\u20130.4, while cellulose degradation at 0.5\u20130.6 conversion and lignin degradation at conversion >0.7 [46,64]. At higher temperatures (>500\u00a0\u00b0C), lignin decomposition was prominent but increased the E\n\n\u03b1\n requirement because of the denser molecules nature of lignin. As the catalyst loading to the feedstock ratio increases, more coke deposition occurs on the catalyst's surface, leading to increased activation energy. The results of the present study follow a similar trend observed in recent works by Refs. [14,30,43,65].Considering all the catalysts used in the present study, each catalyst contributes differently to the pyrolysis reactions. The Pt/C catalyst was known for hydrogenation and higher deoxygenation reactions [42,61]. In comparison, TiO2\u2013ZnO catalyst enhances the dehydration of the alcohol to olefins and leads to lower hydrocarbon production. Furthermore, the TiO2 catalyst helps convert xylene, an intermediate of hemicellulose decomposition [66]. Additionally, the zeolite catalyst provides a higher acid site, promoting the decarbonylation and deoxygenation reactions [43]. Hence, it was apparent that each of these catalysts plays a different role in the decomposition of biomass.Reaction pathways of catalytic pyrolysis (CP) of DR biomass were analysed with Criado's technique. All graphs were obtained using Eqs. (12 \u2013 15) for conversion range (\u03b1\u00a0=\u00a00.1\u20130.7) at a heating rate of 10\u00a0\u00b0C min\u20131, shown in Fig. 5\n (A-C) and Table 5\n. Fig. 5A indicated the reaction pathways of F2, P4, 3A, F0, A4, F3 for 1:30\u00a0wt % of DR:Na\u2013Y; F2, P4, R2, F2, F1, A4 for 1:20\u00a0wt % of DR:Na\u2013Y; and F2, P4, R2, F0, R3, F4 for 1:10\u00a0wt% of DR:Na\u2013Y. Fig. 5B depicted the reaction pathways of F4, F3, F3, F4, D1, P2 for 1:30\u00a0wt % of DR:Pt/C; P4, D3, P3, F0, F2, A4 for 1:20\u00a0wt % of DR:Pt/C; and P4, D2, P3, F0, F5, F3 for 1:10\u00a0wt% of DR:Pt/C. Fig. 5C exhibited the pathway of reaction as F4, F1, F3, F4, D3, R2 for 1:30\u00a0wt % of DR:TiO2\u2013ZnO; A4, P4, P3, D3, F4, F2 for 1:20\u00a0wt % of DR:TiO2\u2013ZnO; and F2, R2, P3, F3, D3, F1 for 1:30\u00a0wt% of DR:TiO2\u2013ZnO. Therefore, the CP of DR biomass was revealed to follow a multistep reaction pathway rather than a single reaction pathway for all three catalysts of each loading [30,31,46].\nFig. 6\n (A-C), shows variations of enthalpy change (\u0394H) of catalytic pyrolysis (CP) of DR biomass with Na\u2013Y, Pt/C, TiO2\u2013ZnO respectively from the DFM technique. The change in enthalpy (\u0394H) was defined as the energy necessary to move a molecule to a higher energy level from a lower energy level [30,63,64]. From the thermodynamic parameter analysis, it was found that the \u0394H varied from 177 for 10\u00a0wt % Na\u2013Y; 195 for 10\u00a0wt % Pt/C; and 196\u00a0kJ\u00a0mol\u22121 for 10\u00a0wt % TiO2\u2013ZnO. Additionally, it varied as 179kJ\u00a0mol\u22121 for 20\u00a0wt % Na\u2013Y; 200\u00a0kJ\u00a0mol\u22121 for 20\u00a0wt % Pt/C; and 186\u00a0kJ\u00a0mol\u22121 for 20\u00a0wt % TiO2\u2013ZnO catalyst loadings. Whereas, it can be seen that the variation in enthalpy change can be noted as 198 kJ mol\u22121 for 30\u00a0wt % Na\u2013Y; 212\u00a0kJ\u00a0mol\u22121 for 30\u00a0wt % Pt/C; and 190kJ\u00a0mol\u22121 for 30\u00a0wt % TiO2\u2013ZnO. The change in enthalpy during the non-catalytic pyrolysis of DR biomass was identified as 190\u00a0kJ\u00a0mol\u22121 [46]. Results indicate that lower \u0394H factors were ascertained for 10 and 20\u00a0wt % of Na\u2013Y and TiO2\u2013ZnO catalyst loadings when compared with the pyrolysis of DR biomass without catalysts. Subsequently, it can also be seen that positive \u0394H factors signify the indigenous endothermic reactions. Though initially, during the beginning of the reaction, the \u0394H factors were noted to be lower, there was a slight increase for \u03b1\u00a0\u2265\u00a00.6 indicating that requirement of energy was more to the reaction to persist at higher conversion [30].\nFig. 6 (D-F), shows variations in change in Gibbs free energy (\u0394G) for catalytic pyrolysis (CP) of DR biomass in the presence of zeolite Na\u2013Y, Pt/C, and TiO2\u2013ZnO catalysts with feedstock to catalyst ratio of 30, 20, and 10\u00a0wt % for the DFM technique. The \u0394G was defined as an increase in the overall energy of the process to constitute an activated complex [30,63,64]. The calculations of \u0394G of the current study show that it has varied from 178,178,176\u00a0kJ\u00a0mol\u22121 for all three catalysts at catalyst loading from 30, 20, 10\u00a0wt % respectively. The \u0394G factor obtained during non-catalytic pyrolysis of DR biomass was 180\u00a0kJ\u00a0mol\u22121 [46]. Also, \u0394H, and \u0394G were found to be positive implying that additional energy was needed to perform the CP of DR biomass.\nFig. 6 (G-I), shows variations in change in entropy (\u0394S) for catalytic pyrolysis (CP) of DR biomass in the presence of zeolite Na\u2013Y, Pt/C, and TiO2\u2013ZnO, catalysts for the feedstock to catalyst ratio of 30, 20, and 10\u00a0wt % by DFM technique. \u0394S was an essential parameter that measures the disordeness of the pyrolysis process. \u0394S < 0 values were described as \u201cslow pyrolysis,\u201d and in this work, the CP of DR biomass experiences \u0394S close to zero, implying close to thermodynamic equilibrium conditions [40]. This scenario was observed for Na\u2013Y 10\u00a0wt% with decreased diorderness. Whereas \u0394S > 0 values were referred to as \u201cfast pyrolysis,\u201d which indicates that the CP of DR biomass undergoes maximal physicochemical changes with more significant reactions suggesting the process was far from thermodynamic equilibrium. This scenario was observed at the higher loading of catalysts (30\u201320\u00a0wt%) for all three catalysts with increased disorderness. The obtained results of \u0394S were also consistent with other studies [30,43].The indices of pyrolysis performance (IPP) of CP of DR biomass at three catalysts including zeolite Na\u2013Y, Pt/C (10\u00a0wt %), and TiO2\u2013ZnO (1\u20131\u00a0wt. %) at 30 to 10\u00a0wt% loading of each was shown in Table 6\n. According to the IPP findings, all parameters (C, D\n\ni\n, D\n\nb\n, S and D\n\nv\n) were increased with increasing heating rates (5\u201355\u00a0\u00b0C min\u22121) for three catalysts. Furthermore, in comparison to the other catalyst loadings, 10\u00a0wt % of zeolite Na\u2013Y exhibited the lowest IPP parameters. This might occur because the zeolite Na\u2013Y has a larger surface area and was stable at higher temperatures [67]. The higher flammability index (C\n\n\n=\n\n 8.78\n\n\u00d7\n\n10\n\n\u2212\n5\n\n\n\n), suggested a lower moisture content and larger heating factor. The higher ignition and burnout indices (D\n\ni\n\u00a0=\u00a0\n\n944\n\u00d7\n\n10\n\n\u2212\n3\n\n\n\n and D\n\nb\n\n\n\n=\n4540\n\u00d7\n\n10\n\n\u2212\n5\n\n\n\n), indicated the higher combustibility. The higher combustion index (S\n\n\n=\n26.5\n\u00d7\n\n10\n\n\u2212\n8\n\n\n\n), exhibited stronger combustion characteristics. The higher devolatilization index (D\n\nv\n\n\n\n=\n48.3\n\u00d7\n\n10\n\n\u2212\n8\n\n\n\n), demonstrated the generation of a significant quantity of volatile content during the catalytic pyrolysis (CP) of DR biomass [46,64,68].This work has investigated the catalytic pyrolysis (CP) of Delonix regia (DR) biomass using three different kinds of catalysts: Na\u2013Y, TiO2\u2013ZnO, and Pt/C, with loading ranging from 30 to 10\u00a0wt%. According to the kinetic and thermodynamic results, it has been concluded that the Na\u2013Y catalyst with a loading of 10\u00a0wt% showed better catalytic activity for CP of DR. From kinetic studies, the KAS technique at DR: Na\u2013Y (1:10\u00a0wt %) has yielded the lowest mean factors of E\n\n\u03b1\n (kJ mol\u22121) 181.29 and k\n\no\n (min\u22121) 2.10 E+16. Therefore, inexpensive, and higher thermal stability catalyst, zeolite Na\u2013Y (load of 10\u00a0wt %) should be considered further for the bio-oils production from CP of DR biomass compared to other loads of catalysts of this study. Moreover, based on the findings, it has been observed that higher loadings (30\u00a0wt %) of all three catalysts are not recommended for a bio-oil generation because of the possible accumulation of coke on the surface of the catalysts. Criado's plots have revealed that the CP of DR biomass has followed a multistep reaction pathway instead of a single reaction pathway during the process. From thermodynamic findings, a 5\u00a0kJ\u00a0mol\u22121 discrepancy between E\n\n\u03b1\n and \u0394H has been noticed from the DFM technique, which has indicated that a large quantity of energy, i.e., at least equivalent to \u0394H or more, should be provided for the pyrolysis to occur. Furthermore, \u0394G and \u0394S results have revealed that the CP of DR biomass undergoes non-spontaneous reactions. Further, more effective catalysts with various loadings should be investigated in the future to evaluate the best kinetic and thermodynamic characteristics. In addition, simulation studies such as computational fluid dynamics (CFD) also facilitate new dimensional work to analyse transportation and other insightful phenomena of such pyrolysis processes using the kinetics developed in the present work.D Rammohan: Investigation; Methodology; Validation, Writing Original Draft. N Kishore: Conceptualization; Resources; Writing Review & Editing; Supervision; Fund Acquisition. RVS Uppaluri: Supervision; Resources.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.", "descript": "\n                  Renewable energy from biomass waste shall become an effective alternative to faster-depleting fossil fuels and pyrolysis is a suitable approach to produce renewable energy from biomass. Further, kinetics parameters are essential in designing reactors for pyrolysis and this work provides kinetics of catalytic pyrolysis (CP) of Delonix regia (DR) biomass. In addition to this significance, the novelty of this work includes the utilization of catalysts of three different kinds such as zeolite (Na\u2013Y), mixed metal oxides (TiO2+ZnO), and supported noble metal (Pt/C) catalysts at varying loads of 30\u201310\u00a0wt %. All experiments were performed in a micro pyrolyzer (thermogravimetry analyzer) under non-isothermal conditions at five heating rates (5, 10, 20, 35, and 55\u00a0\u00b0C min\u22121) in a temperature range of 25\u20131000\u00a0\u00b0C. To estimate kinetic factors (KF) and thermodynamic parameters (TP), five iso-conversional techniques such as Differential Friedman (DFM), Kissinger\u2013Akahira\u2013Sunose (KAS), Ozawa\u2013Flynn\u2013Wall (OFW), Starink (STK), and Distributed Activation Energy (DAE) were employed. KAS technique yielded the lowest mean activation energy, E\n                     \n                        \u03b1\n                      (181.29\u00a0kJ\u00a0mol\u22121), and frequency factor, k\n                     \n                        o\n                      (2.10 E+16 s\u22121) factors by use of Na\u2013Y zeolite of load 10\u00a0wt % whereas the corresponding change in enthalpy is 177\u00a0kJ\u00a0mol\u22121, change in Gibbs free energy is 178\u00a0kJ\u00a0mol\u22121, and change in entropy is \u22129.58\u00a0E\u221204\u00a0kJ\u00a0mol\u22121 K\u22121. Criado's master plots confirmed the reaction pathway as: second order (F2), power-law (P4), contraction area (R2), zero order (F0), contraction volume (R3), and fourth order (F4) for 20\u00a0\u00b0C min\u22121 from DFM technique for CP of DR by using Na\u2013Y zeolite catalyst of load 10\u00a0wt %.\n               "}
{"full_text": "The oxygen evolution reaction (OER) is an essential half-reaction used in energy conversion systems, such as in an electrochemical water splitting system [1,2]. The currently used OER electrocatalysts are based on noble metals, e.g. RuO2 and IrO2; however, their application is limited because they are expensive and rare and show sluggish kinetics [3,4]. To overcome such issues, many studies have attempted to develop non-noble metal-based OER catalysts, such as those based on transition metals (e.g., Ni, Fe, and Co) because of their natural abundance, low cost, and good chemical stability [5-7]. Among the transition metals, Ni-based electrocatalysts have shown excellent catalytic performance for OER owing to their suitable bond strength with neighboring active components [8]. However, they exhibit low conductivity and insufficient active sites, which hinder their electrocatalytic behavior [9].Recently, Ni-doped carbon nanostructures have attracted attention as OER electrocatalysts as they exhibit excellent electrical conductivity, remarkable chemical stability, and high number of active sites [10]. Until now, Ni-based carbon structures have been synthesized by harsh chemical methods, such as hydrothermal and solvothermal methods, which necessitates a further purification process [4,7,11]. In addition, these methods often change the electronic structures of conductive supports, which lead to reduced electron charge transport [12]. Therefore, it is challenging but highly desirable to develop a facile and green way for synthesizing Ni-doped carbon nanostructures.An alternative to the above-mentioned harsh chemical methods for the synthesis of functional carbon nanomaterials is pulsed laser ablation (PLA) process [13-16]. PLA process is simple and environment-friendly because it does not require a harsh chemical reactant and post-purifying steps [17]. Moreover, the remarkable rapid reaction associated with PLA process at high temperatures and high pressure can lead to the formation of a unique carbon nanostructure, such as heteroatom-doped graphene quantum dots and surface-functionalized carbon nanotubes [14,18,19]. Therefore, this robust process can effectively incorporate transition metals (i.e., Ni) in the carbon framework, resulting in further enhancement of the electrocatalytic OER performance. However, there have been few reports on the use of PLA process for fabricating OER catalysts; thus, herein we develop Ni-doped multi-walled carbon nanotubes (MWCNTs; Ni-MWCNTs) as high-performance OER catalysts by PLA process. The Ni-MWCNTs were characterized by high-resolution transmission electron microscopy (HR-TEM), X-ray photoelectron spectroscopy (XPS), X-ray diffraction analysis, and linear sweep voltammetry (LSV).Pristine MWCNTs were purchased from Hanwha Chemical (Republic of Korea). High-purity ethanol (99.9%) and nickel chloride (NiCl2) were purchased from Sigma Aldrich (USA). First, 50\u00a0mg of MWCNTs and 2\u00a0g of NiCl2 were dispersed in 500\u00a0mL of anhydrous ethanol. The PLA process was performed on the mixed solution for 1\u00a0h using a Q-switch ND:YAG pulsed laser system. Simultaneously, tip-type sonication was performed for achieving homogeneous dispersion of precursors. The mixed solution (of MWCNT and NiCl2) was ablated by pulsed laser beam (355\u00a0nm, third harmonic) at a repetition rate of 10\u00a0Hz. The pulse width was 10\u00a0nm, and the pulsed laser energy was 1\u00a0J. After the PLA process, subsequent centrifuging and drying process yielded Ni-MWCNTs.HR-TEM and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) were performed using TALOS F200X (Thermo Fisher Scientific, 200\u00a0kV). XPS measurements were recorded using VG ESCALAB 200i (Thermo Fisher Scientific), where survey and high-resolution scans were obtained at pass energies of 100 and 20\u00a0eV, respectively. All electrochemical measurements (Autolab PGSTAT, Metrohm) were recorded in 1.0\u00a0M KOH (pH\u00a0\u2245\u00a013.7) electrolyte using a three-electrode electrochemical system cell with a rotating disk electrode (RDE). LSV was performed at a scan rate of 5\u00a0mV\u00a0s\u22121. The potentials were calibrated against reversible hydrogen electrode, and all the polarization curves were iR-compensated. Electrochemical impedance spectroscopy (EIS) was performed in the frequency range 0.1\u00a0Hz \u2013 100\u00a0kHz, by applying a sinusoidal voltage with amplitude of 5\u00a0mV.\nFig. 1\n shows the possible formation mechanism of Ni-MWCNTs via PLA process. Briefly, when the pulsed laser is injected into the mixed solution (i.e., MWCNT, NiCl2, and ethanol), extremely harsh environments, such as high pressure and high temperature, are created owing to multi-photon absorption ionization [17]. Subsequently, an intrinsic phenomenon (e.g., plasma plume and cavitation bubble) of the PLA process occurs in the mixed solution, which leads to the generation of Ni molecules (i.e., derived from NiCl2), oxygen, and hydrogen (i.e., derived from ethanol in this study). Simultaneously, the MWCNT structure can partially collapse by the strong pulsed laser. Finally, these Ni molecules, oxygen, and hydrogen are unstable because of their high surface energy, resulting in their incorporation into the carbon framework (Ni-MWCNTs) [18].TEM and HR-TEM images of pristine MWCNTs (P-MWCNTs) and Ni-MWCNTs are shown in Fig. 2\n. Both P-MWCNTs and Ni-MWCNTs have tubular structures and show negligible changes after the PLA process, as seen in low-resolution TEM images (Fig. 2\na,d). However, the HR-TEM images of Ni-MWCNTs show that the walls of the MWCNT structure were collapsed, overlapping with the walls, which were also partially distorted (Fig. 2\ne,f). This phenomenon usually observed when the precisely controlled pulsed laser energy injected into the MWCNT, which leads to the increased surface area of MWCNTs [14]. Although the outer wall of the MWCNTs is partially collapsed, the inner walls of the MWCNTs still retain the tubular structure (Fig. 2\nb,c). Finally, we confirmed the Ni was successfully incorporated in the entire MWCNT structure, as shown in the energy-dispersive X-ray spectroscopy mapping images (Fig. 2\ng).We investigated the chemical composition of Ni-MWCNTs using XPS. For comparison, we prepared control samples of surface-modified carbon nanotubes (SMCNTs) using the same process, but without a Ni precursor. Fig. 3\n\na-c shows the C1s peaks of P-MWCNTs, SMCNTs, and Ni-MWCNTs, respectively. Both SMCNTs and Ni-MWCNTs showed the presence of oxygen-rich functional groups, such as hydroxyl and carboxyl, whereas P-MWCNTs showed only a low-intensity hydroxyl peak. Deconvolution of the Ni2p spectrum shows two main peaks at 856 and 876\u00a0eV [20]. In addition, Ni-O peaks are observed at 853 and 872\u00a0eV. These results indicated that the PLA process modulated the structure of MWCNTs, including the incorporation of Ni.The electrocatalytic performance of Ni-MWCNTs was analyzed in 1\u00a0M aqueous KOH electrolyte using a three-electrode system with an RDE, as shown in Fig. 4\n. For comparison, the electrocatalytic performances of P-MWCNTs, SMCNTs, and the commercial catalyst RuO2 were analyzed under the same conditions (Fig. 4\na). LSV was performed on all the samples, and results showed that Ni-MWCNTs exhibited superior electrocatalytic activity for water oxidation. The overpotential (\u03b7) for delivering a current density of 10\u00a0mA\u00a0cm\u22122 (\u03b7\n10) was 320\u00a0mV for Ni-MWCNTs. Importantly, the commercial RuO2 catalyst exhibited the same current density (10\u00a0mA\u00a0cm\u22122) at 360\u00a0mV, revealing that Ni-MWCNTs had higher electrocatalytic OER activity than that of commercial RuO2 catalyst. This superior OER activity of Ni-MWCNTs is related to the oxygen-rich functional groups, such as carboxyl and hydroxyl, on their surface. The presence of abundant oxygen groups effectively promoted interactions with H-carrying OER intermediates, such as OH* and OOH*, leading to enhanced OER activity [21]. Simultaneously, the successful doping of substitutional Ni results in more efficient OER activity due to the suitable bond strength with neighboring carbon framework and three-dimensional electronic intrinsic structure [22]. In addition, the Tafel slope was determined to calculate the kinetics of the OER rate-determining step [23]. Ni-MWCNTs gave the smallest Tafel slope (30.085\u00a0mV dec-1), followed by SMCNTs (37.035\u00a0mV dec-1), RuO2 (46.16\u00a0mV dec-1), and P-MWCNTs (110\u00a0mV dec-1). The low Tafel slope of 30.085\u00a0mV dec-1 suggests that OH* absorption is favorable on the surface of Ni-MWCNTs, which leads to excellent OER activity. EIS analysis was performed for Ni-MWCNTs and the control samples to calculate the charge transfer resistance (Rct\n) between the surfaces of catalysts (Fig. 4\nc). Generally, the smaller diameters of semicircle correspond to lower Rct\n values [24]. SMCNTs show a large Rct\n value of 63\u00a0\u03a9, while low Rct\n values were obtained for RuO2 (50\u00a0\u03a9), Ni-MWCNTs (30\u00a0\u03a9), and P-MWCNTs (22\u00a0\u03a9). Thus, the highly conductive MWCNTs are believed to be one of the reasons for the increased charge transfer between the catalysts and the reactants. Although the SMCNTs exhibited a larger Rct value than P-MWCNT due to low electron transfer by the collapsed outer wall of MWCNT (Fig. 1), the SMCNTs have oxygen-rich functional groups and high-surface area after the PLA process resulting in the high number of active site (Fig. 3). Consequently, the SMCNTs showed higher electrocatalytic performance than the P-MWCNT. To further investigate the OER activity, we performed cyclic voltammetry (CV) measurements for determining the electrochemically active surface area (ECSA) and turn over frequency (TOF), as shown in Fig. 4\nd,e. Generally, the number of active sites is proportional to the ECSA value [25]. In SMCNTs, the number of active sites increased owing to enhanced surface area and formation of oxygen functional groups by PLA process, as shown in Fig. 4\nd\n[14]. Subsequently, the incorporation of Ni into the MWCNT structure exhibited the highest active site. This result indicates that Ni-MWCNTs have the highest number of active sites, which is attributed to generation of reactive centers by incorporation of Ni. The calculated TOF values of all samples are shown in Fig. 4\ne. Typically, a higher TOF value could be obtained from high density of catalytic activity, which was generally indicated the number of electrons produced per active site. The TOF for Ni-MWCNTs was found to be 0.035\u00a0s\u22121 at an overpotential of 320\u00a0mV, which is three times higher than that of commercial RuO2. In addition, the electrocatalytic durability of Ni-MWCNTs for OER was analyzed by chronoamperometry measurements. After 10\u00a0h of water oxidation at 1.54\u00a0V in 1\u00a0M KOH, Ni-MWCNTs retained 100% of initial current density, as shown in Fig. 4\nf.In summary, a highly efficient and chemically stable Ni-MWCNT electrocatalysts were synthesized by a simple and green PLA process. The Ni molecules derived from NiCl2 are successfully incorporated in the entire MWCNT structure. Partial decomposition of MWCNTs and the formation of oxygen functional groups on the MWCNT surface occurred simultaneously. As a result, we observed a significant enhancement in the OER performance of Ni-MWCNTs due to the synergetic effect of Ni doping and MWCNT structural modification. Thus, our study offers an efficient and simple way of preparing highly efficient OER electrocatalysts, which can also assist in designing new functional carbon catalysts.\nSukhyun Kang: Conceptualization, Writing - original draft. HyukSu Han: Conceptualization, Writing - original draft, Supervision, Writing - review & editing. Sungwook Mhin: . Hui Ra Chae: . Won Rae Kim: . Kang Min Kim: Conceptualization, Writing - original draft, Supervision, Writing - review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea Government (MSIT) (No. 2020R1A2C1102079).", "descript": "\n                  Transition\u2013metal-doped carbon-based electrocatalysts have attracted attention as alternatives to noble metal electrocatalysts (e.g. IrO2 and RuO2) for oxygen evolution reaction (OER) because they are inexpensive and highly efficient. However, their poor catalytic activity and time-consuming synthesis remain a challenge. Herein, we report a facile and green technique using pulsed laser ablation for preparing Ni-doped multi-walled carbon nanotubes (Ni-MWCNTs) as OER catalysts. Ni-MWCNTs exhibit high surface area, oxygen-rich functional groups (e.g., hydroxyl and carboxyl), and successful doping of Ni in the carbon framework. The as-prepared Ni-MWCNTs exhibited excellent OER catalytic performance, with an overpotential of 320\u00a0mV at the current density of 10\u00a0mA\u00a0cm\u22122 in an alkaline medium, which is lower than that of the commercial RuO2 catalyst. Furthermore, Ni-MWCNTs displayed the initial electrocatalytic activity after 10-h stability tests, demonstrating good electrochemical durability. We believe that this work provides a simple protocol for fabricating heteroatom-doped carbon nanotubes as high-performance OER electrocatalysts.\n               "}
{"full_text": "No data was used for the research described in the article.The Biginelli reaction is one of the simple and direct methods for the synthesis of tetrahydropyrimidines, originally reported by Biginelli [1]. Regarding the importance of the Biginelli reaction products, much work on improving the yield and reaction conditions has been actively pursued. For example, using Lewis acids as a catalyst such as Cu(OTf)2\n[2], Sc(OTf)3 or La(OTf)3, Yb(OTf)3\n[3], Triethylammonium hydrogen sulfate [4], BiCl3\n[5], and Mn(OAc)3\u00b72H2O [6] instead of acidic reagents significantly improved the reaction output with reduced reaction times. Some other routes have been reported for the synthesis of Biginelli reaction products using various catalysts [7\u201313]. On the other hands, metal catalysts play very important role in organic synthetic reactions, especially in Biginelli reactions which are reducing reaction time and increases product yields, some of previously reported metal catalysts as H3PO3/Pd, NH4VO3 CeCl3\u00b77H2O, Sm(ClO4)3, ((NH4)2Ce(NO3)6), NiCl2/KI, GaI3, TiO2, FeCl3\u00b76H2O Cu(NO3)2\u00b73H2O, Ce(SO4)2-SiO2, Fe(NO3)3\u00b79H2O, ZnO, copper nitrate, ZrCl4, LnCl3\u00b77H2O, ZrO2-nanopowder, and ZrOCl2\u00b78H2O or ZrCl4/neat, etc [8\u201313]. These reactions were reported by use of ionic liquids, microwave irradiation, solid phase reagents, baker\u2019s yeast, polymer-supported catalysts, zeolites, surfactants, and PEG, etc. [8\u201313].However, these previous methodologies have various drawbacks as long reaction times, work up complexity, tedious reaction conditions, highly economical with low yields. In addition, the multi-component reactions have been known to be powerful tools over conventional multi-step reactions and have emerged as a novel promotion in organic synthesis [7]. These reactions offer a direct fast route and enable the assembly of highly complicated and diversified molecules in a one-pot single-step process with enhanced atom economy [7]. The compounds of tetrahydropyrimidines and their derivatives have attracted considerable interest in medicinal chemistry due to their broad spectrum of pharmacological activities such as antibacterial [14], antifungal [15], anticancer [16], anti-inflammatory, analgesic [17], anti-HIV [18], antihypertensive [19], antimalarial activities [20]. These compounds exhibit specifically broad range of therapeutic and pharmacological properties, namely anticancer, antihypertensive, antiviral, and antifungal. tetrahydropyrimidines derivatives which are found as core units in many marine alkaloids, have been found to be potent HIV gp-120CD4 inhibitors [14\u201320]. Considering the emerging properties of tetrahydropyrimidines derivatives, the progress of advanced, clean, and uncomplicated methodologies for the efficient catalytic synthesis of these compounds with accessible reagents is of great importance.Although most of the reported routes have some benefits, however, some disadvantages are also combined with many of them using environmentally toxic organic solvents, expensive mediators, long reaction time, corrosive nature, tedious work-up, non-recyclable catalysts, limited substrate scope, high temperature, and low yields. Therefore, the development of effective and environmental benign methods is desirable for the synthesis of tetrahydropyrimidines compounds [21\u201323]. Therefore, we have opted to synthesis tetrahydropyrimidines derivatives via a green approach by using ZrO2/La2O3 catalyst as multi-component reactions. According to this scenario, several catalysts including metal oxides have been reported as nanocatalysts for the synthessis of chromenes and its derivatives as well as Biginelli products including Bi2V2O7\n[24], NiO@TPP-HPA [25], CuO [26,27], MoO3\u2013ZrO2 nanocomposite [28], MnO2\u2013MWCNT [28], Cu2O [29], Mg\u2013Al\u2013CO3 and Ca\u2013Al\u2013CO3 hydrotalcite [30], Bi2O3/ZrO2\n[31], ZrO2\u2013Al2O3\u2013Fe3O4\n[32], histaminium tetrachlorozincate [33], Alumina supported MoO3\n[34], ZrO2-pillared clay [35], Zn(l-proline)2\n[36], Fe3O4-CNT [37], TiO2-MWCNT [38], Melem@Ni-HPA [39], RuO2\n[40] and SiO2/H3PW12O40\n[41].So far, many researchers have been reported Biginelli reaction by using various metal oxide systems at different reaction conditions. The various materials, including transition metals were necessitated to enhance the qualities and efficacy of ZrO2 or La2O3 nanoparticles. The surfaces of these particles were indeed transmogrified as a consequence of transition metal oxides doping to demonstrate significantly larger functionalities such as a higher surface area, and compactness, enabling them to actively participate catalytic applications. To the best of our knowledge, there is no report on the synthesis of the discerning weight composition of ZrO2/La2O3 heterogeneous catalyst. Besides, there is no report on the catalytic application of the obtained tetrahydropyrimidine derivatives in the Biginelli reaction and the investigation of the correlation of the reaction conditions including solvent-free conditions with the catalytic application. So, it was rationalized that the hard/soft natures of the metal ions would play important role in the catalytic activity of such catalysts. The current synthetic protocol furnishes several advantages like short reaction time, high purity of the isolated products, ease of reaction handling, simple separation and high yield of the desired products along with the synthesized nanocatalyst.In a typical synthesis, dissolve Lanthanum nitrate hexahydrate (5.0 g) in 100\u00a0ml of double distilled water and 0.1\u00a0M ammonium hydroxide was added to form white precipitation. The precipitate compound was filtered using Buchner funnel and washed with ammonium hydroxide then the obtained compound dried in an oven at 150\u00a0\u00b0C for 12 h. Then, the pulverised Lanthanum (III) hydroxide (2.5 g) was dissolved in double distilled water and added zirconium(IV) oxynitrate hydrate, then heated on water bath up to dry precipitate is formation. The compound dried in an oven 150\u00a0\u00b0C for overnight and further calcinations at 650\u00a0\u00b0C for 4\u00a0h to obtain pure pulverised 5\u00a0% of ZrO2/La2O3.To estimate the surface modification of the synthesized catalysts, FTIR (KBr) spectra were recorded on a Shimadzu FT-IR-8400s spectrophotometer. The powder X-ray diffraction pattern has been recorded on a Siemens D-5000 diffract meter by using Cu K radiation source and a Scintillation counter detector. The XRD phases present in the samples were identified with the help of JCPDS data files. The BET surface area was determined by N2 adsorption\u2013desorption isotherms at liquid N2 temperature on a Micromeritics Gemini 2360 instrument. Prior to physical measurements, the synthesized compounds were dried in an oven at 393\u00a0K for 10\u00a0h and flushed with Argon gas for 1 hr. The melting points were determined in open capillary tubes and are uncorrected. The purity of the compounds was checked by TLC using pre-coated silica gel plates 60254(Merck). 1H NMR and 13C NMR spectra were recorded on Bruker Avance II 400\u00a0MHz spectrometer using tetramethylsilane as an internal standard. Mass spectra were recorded on a GCMS-QP 1000 EX mass spectrometer.In a typical procedure, a mixture of aldehydes (Ia-n, 1\u00a0mmol), acetoacetate (II, 1\u00a0mmol), urea (III, 1\u00a0mmol) and the synthesized 5.0\u00a0% ZrO2/La2O3 (30\u00a0mg) were placed in a round-bottom flask under solvent free conditions. The suspension was stirred at 80\u00a0\u00b0C for 30\u201340\u00a0min. The progress of the reaction was monitored by thin layer chromatography (TLC) [6:4 hexane:ethylacetate]. After completion of the reaction, the solid crude product was washed with deionized water to separate the unreacted raw materials. The precipitated solid was then collected and dissolved in ethanol to separate the solid catalyst. The filtrate was left undisturbed at room temperature to afford the crystals of the pure products such as tetrahydropyrimidine derivatives (IVa-n). The structures of all of the products were verified using 1H and 13C NMR spectral information.The synthesized ZrO2/La2O3 catalyst were generally characterized by the investigating their size, shape, morphology, optical band gap, and surface area by various techniques. A homogeneity in these properties results in the advancement in applications of nanoparticles.The surface and structural changes of the synthesized La2O3 and ZrO2/La2O3 catalysts were characterized by FT-IR spectra. It shows majorly-two characteristic stretching frequencies at peaks around 1460 and 856\u00a0cm\u22121 as shown in Fig. 1\n. The peak around 1460\u00a0cm\u22121 is representing to presence oxide and the peak found at 856\u00a0cm\u22121 is characterizing the crystalline La2O3. The very weak absorption bands at 3604\u00a0cm\u22121 are assigned to OH symmetric stretching vibration of water molecules which is obtained due to hygroscopic nature of lanthanum oxide. Basis on the obtained FT-IR data, it concludes that the crystalline La2O3 do not changes with ZrO2 in the synthesized ZrO2/La2O3 catalyst.The X-ray diffraction patterns for the synthesized La2O3 and ZrO2/La2O3 catalysts were illustrated in Fig. 2\n. In the XRD spectrum, it can be clearly seen that the highly narrow sharp lines for the formation of La2O3. The Braggs reflection pattern can be assigned for the formation of lanthanum oxide with cubic phase. The intense diffraction peaks obtained from XRD data include the 2\u03b8 values 15.66, 27.31, 27.98, 31.63, 39.49, and 48.62. A strong intensity peak (Fig. 2) is detected at a diffraction angle of 31.63, which is assigned to (101) plane of La2O3. The other peaks are assigned to (002), (102), and (110) lattice planes belonging to cubic crystalline phase of La2O3 and agreed with the previous report [42]. In the XRD pattern of the synthesized ZrO2/La2O3 catalyst, in addition, the strongest intensity peaks appeared at 2\u03b8 values 25.12, 30.58 and 44.05 these are associated with (111), (200), and (201) planes respectively, which indicates ZrO2 doped La2O3, which is suggested to the hexagonal phase. The average particle size was calculated using the Debye-Scherer equation D\u00a0=\u00a00.9 \u03bb/\u03b2 cos\u03b8 (where D is the average crystalline size, \u03bb is X-ray wavelength, \u03b2 is (FWHM) diffraction line and \u03b8 is the diffraction angle). The average crystalline size for La2O3 and ZrO2/La2O3 catalysts was found to be 35\u00a0\u00b1\u00a05\u00a0nm.The surface and structural morphology of prepared La2O3 and ZrO2/La2O3 catalysts were characterized by SEM. In the SEM images of La2O3 and ZrO2/La2O3 as shown in Fig. 3\n, these nanoparticles are found to be very effective to the surface area contribution and similar to each other. The average crystalline size of the particles was also found to be less than 50\u00a0nm. It indicates that the particles were uniformly distributed all over the surface and spherical in shape and this result was agreement with XRD results hexagonal phase with same crystallite size. SEM micrographs indicated that these nanoparticles are comprised uneven spheres. It is obvious that there is some aggregation occurring in these nanoparticles. The SEM micrographs further disclose the materials porosity, which is necessary for catalytic applications.The energy-dispersive spectra (EDS) was analyzed for La2O3 and ZrO2/La2O3 catalysts to confirm the elemental composition of prepared La2O3 and ZrO2 nanoparticles. The EDS spectrum of La2O3 and ZrO2/La2O3 catalysts were illustrated in Fig. 4\n, it clearly indicating the presence of elemental lanthanum, zirconium, and oxygen at 4.3\u00a0keV, 2.1\u00a0keV, and 0.3\u00a0keV, respectively, which is in good agreement with the reported status of ZrO2 doped La2O3. In the molecular formula of ZrO2/La2O3 catalyst, the stoichiometric atomic weight percentage for lanthanum to oxygen is 1:3, and in the present synthesis, the stoichiometric atomic weight percentage exactly matches with the ideal composition of ZrO2/La2O3 catalyst material.The UV\u2013visible spectra of the synthesized La2O3 and ZrO2/La2O3 catalysts were presented in Fig. 5\n. The absorption edges obtained from the plots of absorbance vs wavelength. (The interception of the tangent on the descending part of the absorption peak of the wavelength axis gives the value of diffuse absorption edge in nm). The UV\u2013visible spectrum of La2O3 shows absorption peak in visible region, the wavelength observed at 380\u00a0nm with band gap 3.26\u00a0eV (The band gap was measured using Eg\u00a0=\u00a01240/\u03bb formula, where Eg is the band gap energy and \u03bb is the wavelength of the absorption edge). The UV\u2013visible spectrum of the synthesized ZrO2/La2O3 catalyst shows a trivial red shift with compared to La2O3 at observed wavelength at 385\u00a0nm with band gap of 3.21\u00a0eV, the red shift in UV\u2013visible DRS spectrum clearly indicate incorporation of ZrO2 on La2O3 in ZrO2/La2O3 catalyst.The surface area is an important aspect for the important applications such as surface adsorption and catalytic phenomenon. The synthesized La2O3 and ZrO2/La2O3 catalysts were investigated by nitrogen adsorption/desorption isotherms for the determining the quantitative aspect such as surface area. The specific surface area of the synthesized La2O3 and ZrO2/La2O3 catalysts were found to be 7.0157\u00a0m2/g and 11.0191\u00a0m2/g, respectively as shown in Fig. 6\n. The specific surface area of ZrO2/La2O3 catalyst was superior than that of pure La2O3, the better surface area may due to impression of ZrO2 on the surface of La2O3 in the ZrO2/La2O3 catalyst. It clearly represent the ZrO2 is strongly influences the surface area of pure La2O3 in the ZrO2/La2O3 catalyst.In the application part, all the experiments were performed at the optimized concentration of catalyst. For catalytic study, the standard reaction between of benzaldehyde, urea, and acetoacetate as well as effect of solvent was investigated and summarized in Tables 1 and 2\n\n. By varying the catalyst concentration from 2 to 10\u00a0mol%, the best outcome was obtained using concentration of 5.0\u00a0mol% catalyst under ultrasonic irradiation at 50\u00a0\u00b0C. The catalyst loading and selection of solvent were the important tasks during the present study. The increase in catalyst loading was found to exert substantial effect on product yield. There was not much difference in the yield and reaction time when catalyst loading was changed from 5.0 to 10\u00a0mol%. However, in terms of catalytic efficiency, 5.0\u00a0mol% was found to be the better choice, and therefore, the remaining experiments were performed at the concentration of 5.0\u00a0mol%. During the solvent effect study, when we switched our attention is without solvent to other solvents (Table 2), there was marked effect on the yield of the product. The scope and generality of this protocol were studied by performing the experiments with broad range of aromatic aldehydes. Importantly, we found that benzaldehydes with variety of substitution pattern did not show large difference in the yield of tetrahydropyrimidine derivatives in this novel route.The prepared 5.0\u00a0% ZrO2/La2O3 catalyst was efficaciously used to orchestrate tetrahydropyrimidine derivatives under reflux conditions. A standard reaction encompassing benzaldehyde, urea, and acetoacetate was inspected for catalyst optimization as shown in Scheme 1\n. Having followed a literature review, it was encountered that ethanol was the most prevalently utilized solvent for the synthesis of a wide range of heterocyclic compounds prompting us to utilize ethanol solvent. It was encountered that a catalyst potency of 50\u00a0mg produces better performance. One of the most critical factors throughout this investigation became to optimize the catalyst loading and pick the right solvent. We actually started our exploration without a catalyst and afterward progressively augmented the dose of the catalyst. The proportion of the 5.0\u00a0% ZrO2/La2O3 catalyst dose was found to be superior in terms of catalytic proficiency and yield of the product under solvent free conditions, the remaining reactions with various aldehydes were likewise accomplished with a 5.0\u00a0% ZrO2/La2O3 catalyst loading and the obtained results were summarized in Table 1.The validity of present protocol was tested by employing the optimization conditions to synthesize broad range of tetrahydropyrimidine derivatives using aromatic aldehydes comprising various types of substituents (Table 1). All types of aromatic aldehydes furnished resulted in good yield within short reaction time without purification required. Among the heterocyclic aldehydes, the compounds of aldehydes (formula Ic, Ig, Ij, and Ii) was also used to check whether heterocyclic ring remains intact or not and to our credit here also product yield was quite higher with good stability of the resulting product. Besides all the aldehydes (formula Ib, Id, Ie, and Ih) were excellent in producing tetrahydropyrimidine derivatives without affecting the yield and reactions completed within one hour. Table 1 is completely depicts the physico-chemical data of the synthesized tetrahydropyrimidine derivatives by using the synthesized 5.0\u00a0% ZrO2/La2O3 catalyst.FTIR spectrum, \u03bd, cm\u22121: 758, 1219, 1643, 1724 and 3242; 1H NMR spectrum, \u03b4, ppm: 0.93\u20130.95 (t, 3H), 2.11 (s, 3H), 3.84\u20133.86 (q, 2H), 5.17 (s, 1H), 5.84 (s, 1H), 7.05\u20137.08 (m, 5H), 8.15 (s, 1H); M 261 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 783, 1219, 1645, 1703 and 3240; 1H NMR spectrum, \u03b4, ppm: 1.07\u20131.11 (t,3), 2.24 (s, 3H), 3.70 (s, 3H), 3.97\u20134.01 (q, 2H), 5.26 (s, 1H), 5.95 (s, 1H), 6.73\u20136.75 (d, 2H), 7.14\u20137.16 (d, 2H), 8.38 (s, 1H); M 291 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 792, 1226, 1637, 1693 and 3348; 1H NMR spectrum, \u03b4, ppm: 0.81\u20130.83 (t, 3H), 2.20 (s, 3H), 3.77\u20133.99 (q, 2H), 5.51 (s, 1H), 5.64 (s, 1H), 6.99\u20137.14 (m, 4H), 8.03 (s, 1H); M 295 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 779, 1221, 1645, 1706 and 3240; 1H NMR spectrum, \u03b4, ppm: 0.95\u20130.95 (t, 3H), 2.10 (s, 3H), 3.85\u20133.87 (q, 2H), 5.16 (s, 1H), 5.79 (s, 1H), 7.01\u20137.04 (m, 4H), 8.15 (s, 1H); M 295 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 775, 1226, 1647, 1699 and 3360; 1H NMR spectrum, \u03b4, ppm: 1.20\u20131.23 (t, 3H), 2.22 (s, 3H), 2.30 (s, 3H), 4.12\u20134.14 (q, 2H), 4.95 (s, 1H), 5.68 (s, 1H), 7.01\u20137.03 7.05\u20137.07 (d, 2H), 7.95 (s, 1H); M 275 [M\u00a0+\u00a0H]+.FTIR spectrum, \u03bd, cm\u22121: 734, 1228, 1668, 1712 and 3342; 1H NMR spectrum, \u03b4, ppm: 1.12\u20131.15 (t, 3H), 2.28 (s, 3H), 3.89\u20133.92 (q, 2H), 5.46 (s, 1H), 5.99 (s, 1H), 7.65\u20137.69 (m, 3H), 8.16 (s, 1H); M 306 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 781, 1217, 1674, 1721 and 3340; 1H NMR spectrum, \u03b4, ppm: 1.25\u20131.28 (t, 3H), 2.35 (s, 3H), 4.12\u20134.15 (q, 2H), 5.55 (s, 1H), 5.69 (s, 1H), 7.41\u20137.43 (d, 2H), 7.56\u20137.58 (d, 2H), 8.38 (s, 1H); M 306 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 787, 1223, 1651, 1726 and 3329; 1H NMR spectrum, \u03b4, ppm: 1.24\u20131.27 (t, 3H), 2.35 (s, 3H), 3.85 (s, 3H), 4.07\u20134.10 (q, 2H), 5.25 (s, 1H), 5.75 (s, 1H), 6.80\u20136.81 (d, 1H), 6.89 (s, 1H), 7.25 (d, 1H), 8.11 (s, 1H); M 307 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 787, 1212, 1651, 1725 and 3346; 1H NMR spectrum, \u03b4, ppm: 1.18\u20131.21 (t, 3H), 2.35 (s, 3H), 3.85 (s, 3H), 3.87 (s, 3H), 4.11\u20134.14 (q, 2H), 5.37 (s, 1H), 5.56 (s, 1H), 6.58 (s, 1H), 6.88\u20136.90 (m, 2H), 8.15 (s, 1H); M 321 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 792, 1222, 1648, 1718 and 3322; 1H NMR spectrum, \u03b4, ppm: 1.24\u20131.27 (t, 3H), 2.35 (s, 3H), 3.85 (s, 9H), 4.11\u20134.14 (q, 2H), 5.35 (s, 1H), 5.69 (s, 1H), 6.59 (s, 2H), 8.07 (s, 1H); M 351 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 754, 1228, 1666, 1712 and 3348; 1H NMR spectrum, \u03b4, ppm: 1.13\u20131.16 (t, 3H), 2.30 (s, 3H), 4.11\u20134.14 (q, 2H), 5.42 (s, 1H), 5.99 (s, 1H), 7.21\u20137.22 (d, 1H), 7.33\u20137.34 (d, 2H), 7.41 (s, 1H), 8.21 (s, 1H); M 329 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 785, 1232, 1658, 1720 and 3337; 1H NMR spectrum, \u03b4, ppm: 1.18\u20131.21 (t, 3H), 2.35 (s, 3H), 4.01\u20134.04 (q, 2H), 5.51 (s, 1H), 5.68 (s, 1H), 6.29 (d, 1H), 6.99\u2013702 (m, 2H), 8.00 (s, 1H); M 251 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 788, 1215, 1660, 1737 and 3328; 1H NMR spectrum, \u03b4, ppm: 1.16\u20131.19 (t, 3H), 2.35 (s, 3H), 4.12\u20134.15 (q, 2H), 5.60 (s, 1H), 5.61 (s, 1H), 6.29\u20136.31 (d, 1H), 6.41\u20136.43 (d, 1H), 7.22\u20137.35 (m, 5H), 7.98 (s, 1H); M 174 [M+H]+.FTIR spectrum, \u03bd, cm\u22121: 775, 1224, 1645, 1697 and 3346; 1H NMR spectrum, \u03b4, ppm: 1.08\u20131.12 (t, 3H), 2.26 (s, 3H), 3.98\u20134.03 (q, 2H), 5.28 (s, 1H), 5.89 (s, 1H), 6.70\u20136.79 (m, 3H), 6.89\u20136.92 (d, 2H), 7.19 (s, 1H),7.34\u20137.40 (m, 3H), 8.18 (s, 1H); M 311 [M+H]+.In conclusion, we represent multicomponent synthesis of broad range of tetrahydropyrimidine derivatives employing robust and green 5.0\u00a0% ZrO2/La2O3 catalyst under neat reaction conditions from commonly accessible aromatic aldehydes, acetoacetate, and urea. Moreover, the 5.0\u00a0% ZrO2/La2O3 catalyst was synthesized by using a very simple co-precipitation approach that furnished highly pure product with high yield. FTIR, UV\u2013vis, XRD, SEM, EDS, and BET, have been used to investigate the physical, morphological, and surface characteristics of La2O3 and ZrO2/La2O3 catalysts. The characterization study revealed that synthesized catalyst possessing hexagonal structure and having high porosity for catalytic activity. The synthesized tetrahydropyrimidine were characterized by 1H NMR and 13C NMR spectroscopic techniques to confirm their formation. The catalyst loading was optimized and we disclosed that 5.0\u00a0% ZrO2/La2O3 catalysts furnished more than 90\u00a0% product yield that too within short time span. The present protocol validates ample substrate scope comprising wide range of aldehydes and acetoacetatee in combination with urea. The benefits of this environmentally friendly framework encompass simple approach of synthesis of nanocatalyst, the heterogeneous nature of the nanocatalyst and its simple separation from the reaction mass, quick reaction times, simple isolation, high yields, and clear and simple work-up procedures.\nG. Balraj: Data curation, Investigation, Methodology, Software, Writing \u2013 original draft, Writing \u2013 review & editing. Kurva Rammohan: Data curation, Methodology, Conceptualization, Funding acquisition, Project administration, Validation. Ambala Anilkumar: Data curation, Investigation, Methodology, Software, Writing \u2013 original draft, Writing \u2013 review & editing. M. Sharath Babu: Data curation, Methodology, Conceptualization, Funding acquisition, Project administration, Validation. Dasari Ayodhya: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing \u2013 original draft, Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to thankful to the Head, Department of Chemistry, Osmania University, Telangana state, India for continuous encouragement and providing necessary facilities.Not applicable.", "descript": "\n                  The objective of the present study is focused on the facile synthesis of ZrO2/La2O3 catalyst and the catalytic performance was evaluated towards one-pot synthesis of tetrahydropyrimidne derivatives through Biginelli condensation reaction. The synthesized catalyst was characterized by using FTIR, XRD, BET, SEM, EDX and UV-DRS techniques. All catalysts perform better than the without catalyst based reactions, which is the formation of tetrahydropyrimidne derivatives (IVa-n) conversion averaging 85\u201395\u00a0% at 50\u00a0\u00b0C temperature and 3\u201340\u00a0min under identical reaction conditions. Also, it maintains good stability. The relative ease of reducing the catalyst plus its porous nature are responsible for its performance. All the products were characterized by comparing their physical and spectral data including FTIR, 1H NMR, and ESI-MS with those of authentic compounds reported in the literature.\n               "}
{"full_text": "Efficient and cheap renewable hydrogen production from water electrolysis is a crucial challenge for a sustainable society [1,2]. Anion exchange membrane (AEM) water electrolysis aims to combine the low cost of alkaline with the advantages of proton exchange membrane (PEM) electrolyzers [3,4]. Nickel is the most active and cheapest non-noble metal catalyst reported for hydrogen evolution reaction (HER) in alkaline electrolysis [5,6]. Nickel initially shows a high HER activity, but the activity deteriorates rapidly over time due to the formation of metal hydride under HER conditions [7]. Alloying nickel with other elements such as Mo, P, S, Cu increases HER activity and stability [8]. Sluggish HER kinetics in alkaline electrolytes causes high overpotentials thus remains a challenge to develop a highly active and stable HER catalyst [9,10]. The mechanism of the HER in alkaline media is usually discussed in terms of the Volmer, Heyrovsky, and Tafel reactions [11], and a fundamental understanding of the factors influencing the rates of these steps may provide important clues for catalyst design.NiCu has been reported as an active HER catalyst in alkaline electrolytes. NiCu catalysts have been synthesized using different processes such as freeze casting [12] electrodeposition [13] and powder metallurgy [14]. He et\u00a0al. obtained a current of \u221210\u00a0mA/cm2 at 117\u00a0mV with NiCu synthesized by galvanostatic deposition and with an atomic ratio of Ni to Cu equal to one [15]. Electrodeposited NiCu nanosheets exhibited enhanced HER activity with an onset potential of 48\u00a0mV vs. RHE [16]. Solmaz et\u00a0al. reported that NiCu showed higher HER activity than Ni and Cu due to the roughness effect and synergistic interaction between Ni and Cu atoms [17,18]. Oshchepkov et\u00a0al. found that high mass activity of Ni0.95Cu0.05/C is due to the electronic influence of Cu on Ni [19]. However, there is a deficiency in the literature of results demonstrating NiCu catalyst as a cathode in real alkaline electrolyzers.Mixed metal oxide (Ni/NiO-transition metal oxide (TMO)) composite structures exhibit superior HER activity [20\u201322] NiO attracts OHads while the metallic Ni attracts Hads intermediate during HER, thus lowering the free energy of the first step in the HER, viz. the formation of adsorbed hydrogen through the Volmer step of the reaction. The other TMO oxide, such as Cr2O3 or Fe3O4 appears to stabilize the composite NiO component under HER conditions [20\u201323]. Therefore, it is beneficial to design HER catalyst containing both a catalytic element and elements (or oxides) stabilizing mixed oxidation states for the catalytic element. Below we suggest NiCu mixed metal oxide (MMO) to catalyze the HER so that the catalyst contains both Ni (which has an affinity for Had) and NiO (which has an affinity for OHad), the simultaneous presence of these being stabilized by CuO under HER conditions.Ionomers are frequently employed in electrochemical testing is to promote ink uniformity and coating quality [24]. The presence of the ionomer may also increase ionic conductivity and minimizes mass-transport limitations related to the diffusion of the ionic species [24]. The literature has reported that catalytic layers containing Nafion ionomer result in higher HER activity compared to catalytic layers with other anion exchange ionomers [4,21]. The activity difference has been attributed to several factors, such as ionomer head groups, and ionomer backbone chemistry [4,21]. However, tuning ionomer to catalyst ratio is required for optimum catalyst utilization [25]. Not only its activity and stability but also the interaction of NiCu metal mixed oxide catalyst with ionomers and effects of an aqueous electrolyte, is therefore also important if this catalyst is to play a role in upscaling to AEM water electrolyzer devices, preferably including not only experiments in aqueous cells but also electrolyzer testing.In this work, we investigate the HER activity of various nickel-copper catalysts such as NiCu alloy, NiCu oxide, and NiCu mixed metal oxide (MMO) synthesized by chemical reduction. As we will show below, NiCu MMO shows an exceptional HER activity in alkaline media. In situ Raman spectroscopy under HER conditions was carried out to investigate the state of copper and nickel species present and how these states vary over time for various nickel copper catalysts and correlate this with their activity for the HER. The electrochemical activity of NiCu MMO was further optimized in terms of the type of ionomer binder, KOH concentration in the aqueous electrolyte, and the ionomer to carbon ratio. An AEM water electrolyzer based on membrane electrode assembly (MEA) of NiCu MMO at the cathode and Ir black at the anode was fabricated, tested, and compared to Pt/Ir MEA. The NiCu MMO/Ir MEA shows comparable performance to Pt/Ir MEA which indicates that it could replace scarce and expensive Pt catalyst.NiCu alloy and NiCu oxide were synthesized by mixing 10\u00a0mmol of nickel nitrate hexahydrate Ni(NO3)2.6H2O (97.0%, Sigma Aldrich) and 10\u00a0mmol Copper(II) sulfate pentahydrate (98.0%, Sigma Aldrich) in 500\u00a0ml water (18.2 M\u03a9 cm, 3 ppb TOC, Milli-Q ultrapure water). The precursor mixture was stirred for 15\u00a0min at 750\u00a0rpm. 200\u00a0ml of 0.15\u00a0M NaBH4 (98%, Sigma Aldrich) was added dropwise while bubbles were observed. The solution mixture was stirred for another 1\u00a0hour to ensure the complete chemical reduction of precursors. The resulting precipitate was centrifuged 5 times at 8000\u00a0rpm for 6\u00a0min and cleaned with water and ethanol three times. The produced precipitate was dried in a vacuum oven at 80 \u00b0C overnight. The dried powder was annealed in an air atmosphere to obtain NiCu oxide or 5%H2/Ar to obtain NiCu alloy. The annealing was done at 500 \u00b0C for 6\u00a0h with a ramping rate of 10 \u00b0C/min.In order to make NiCu MMO, 100\u00a0ml of 1\u00a0M Na2CO3 (\u226599.5%, Sigma-Aldrich) were added to nickel-copper precursors solution until the solution became milky and pH reached 10. The mixture was then stirred for another 15\u00a0min, followed by the addition of NaBH4 dropwise. The produced catalyst was subjected to the same procedure for cleaning and drying as above. The dried powder was annealed in 5%H2/Ar atmosphere at 500 \u00b0C for 6\u00a0h with a ramping rate of 10 \u00b0C/min.For catalyst supported on carbon, Ketjen black EC-600JD (AkzoNobel) was dispersed in the precursors' solution mixture to get (60\u00a0wt% catalyst supported on carbon) and stirred for another 1\u00a0hour before adding NaBH4 and complete the chemical reduction step.Scanning electron microscopy (SEM, Carl Zeiss supra 55) and energy dispersive X-ray (EDX) spectroscopy in the SEM device were used to study the morphology and elemental composition of catalysts. The catalyst morphology was further studied using Hitachi S-5500 via scanning transmission electron microscopy (STEM) mode. Bruker D8 A25 DaVinci X-ray device (Cu-K\u03b1 radiation with a wavelength of 1.5425\u00a0\u00c5) was used to examine the crystalline characteristics of catalysts. X-ray diffraction (XRD) patterns were taken between 15 [2\u03b8] and 75 [2\u03b8] using a step size of 0.3 [2\u03b8]. WITec alpha300 R Confocal Raman device with a 532\u00a0nm laser was used to collect the Raman vibrational characterstics of catalyst powders. X-ray photoelectron spectroscopy (XPS) was done via an Axis Ultra DLD instrument (Kratos Analytical) equipped with Al X-ray monochromatic source.Electrochemical investigation of the catalysts was carried out in a three-electrode cell using a rotating disk electrode (Pine Research,) with an (Ivium-n-Stat) potentiostat. Carbon paper (Toray 090, Fuel cell store) was used as the counter-electrode while Hg/HgO electrode (Pine Research) was served as the reference electrode. The working electrode was catalyst deposited on glassy carbon (GC) electrodes (5\u00a0mm diameter, Pine Research). The GC electrode was polished using alumina suspension (5 and 0.05\u00a0\u03bcm, Allied High-Tech Products, Inc.) on polishing pads. The GC electrode was then washed, sonicated in 1\u00a0M KOH for 5\u00a0min, and finally rinsed with water. The catalyst ink was prepared by dispersing 10\u00a0mg catalyst powder in 1.0\u00a0mL of a solution [500 \u03bcL water, 500 \u03bcL isopropanol]. The ionomer used was either Nafion (5\u00a0wt%, Alfa Aesar) or anion exchange ionomer Fumion FAA-3 (10 wt% fumatech) with an ionomer to catalyst weight ratio of 0.2. The Nafion ionomer to catalyst weight ratio in the ink was then optimized from a selection of weight ratios equal to 0.1, 0.3, 0.5, 0.7, and 0.9. The ink was then sonicated for 30\u00a0min in an ice bath. Catalyst loading on the GC surface was kept 250\u00a0\u00b5g/cm2.The catalyst ink was spin-coated on a GC electrode turned upside down and rotated to assure a homogenous catalyst distribution. A water drop was deposited on the electrode before immersed in the electrolyte to prevent air bubbles from forming at the electrode surface. All the electrochemical measurements were conducted in N2-saturated 1\u00a0M KOH electrolyte at room temperature (20\u00a0\u00b1\u00a02). The electrolyte was purged for 30\u00a0min with N2 gas before using and during the experiment to remove any dissolved gasses during electrochemical measurements. The electrolyte was prepared by using KOH (Sigma Aldrich, 85%), and water (18.2 M\u03a9 cm, Milli-Q\u24c7 Integral ultrapure water). The electrolyte was purified according to the procedure reported by Trotochaud et\u00a0al. [26].The working electrode underwent electrochemical activation by cycling between \u22120.8 to \u22121.5\u00a0V vs Hg/HgO at a scan rate of 100\u00a0mV/s for 50 cycles. The linear sweep voltammetry (LSV) polarization curves were recorded in a potential range of \u22120.8 to \u22121.5\u00a0V vs. Hg/HgO at 1\u00a0mV/s sweep rate under continuous stirring at 1600\u00a0rpm to avoid the accumulation of gas bubbles over the GC electrode. The electrochemical impedance spectroscopy (EIS) measurements were collected at specific overpotentials (\u2212100 to \u2212250\u00a0mV) in a frequency range of 0.1\u00a0\u2212\u00a0105\u00a0Hz with an amplitude of 10\u00a0mV alternative current (AC) perturbation. In this work, ohmic resistance (IR) drop was compensated at 85% of high-frequency resistance, which was measured by the EIS technique. The potential was compensated by the following equation:\n\n(1)\n\n\n\nE\ncompensated\n\n=\n\nE\nmeasured\n\n\u2212\ni\nR\n\n\n\nwhere E\ncompensated and E\nmeasured are compensated and measured potentials, respectively.The Hg/HgO potentials were converted to RHE by measuring the voltage at zero current of the HER curve in a hydrogen-saturated electrolyte on Pt electrodes. The Hg/HgO reference electrode potential was converted to RHE in 1\u00a0M KOH using the following equation:\n\n(2)\n\n\n\nE\n\nvs\nRHE\n\n\n\n=\n\nE\n\nvs\nHg\n/\nHgO\n\n\n+\n0.9\n\n\n\n All the reported current densities were normalized to the geometric area of the electrode.The electrochemical active surface area (ECSA) was measured by the electrochemical double-layer capacitance method. Then capacitance from 0.9 to 1\u00a0V vs RHE at scan rates of 50, 100, 150, 200, 250\u00a0mV/s. The CV used for electrochemical double-layer capacitance (Cdl) calculation was acquired in a potential window where no Faradaic process occurred. To derive the Cdl, the following equation was used:\n\n(3)\n\n\n\nC\n\nd\nl\n\n\n\n=\n\n\nI\nc\n\n\u03bd\n\n\n\n\n\nwhere Cdl\n is the double-layer capacitance (mF/cm2) of the electroactive materials, Ic\n is the charging current (mA/cm2), and \u03bd is the scan rate (mV/s).Chronoamperometry was measured at a fixed potential (\u22120.4\u00a0V vs. RHE) for 30\u00a0h. The stability of the catalyst material was also evaluated using an accelerated stress test (AST). AST was carried out by cycling the electrode between \u22120.8 to \u22121.3\u00a0V at a scan rate of 100\u00a0mA/cm2 for 5000 cycles. The Hg/HgO reference electrode was calibrated versus a reversible hydrogen electrode (RHE) in 1 and 0.1\u00a0M KOH. The electrochemical data shown are average data from 3 inks from every powder for each catalyst.In situ Raman measurements were carried out with a lab-made Teflon cell. The catalyst deposited on GC (pine research), a carbon paper (fuel cell store), and Hg/HgO (Pine Research) was used as a working, counter, and reference electrode, respectively as in Fig.\u00a01\n. In situ Raman spectra were collected using a WITec alpha300 R Confocal Raman microscope [532\u00a0nm laser with a power of 5.0\u00a0mW] coupled with Zeiss EC Epiplan 10x objective and G1: 600\u00a0g/mm BLZ=500\u00a0nm grating. The GC surface was polished with \u03bcm-sized alumina powders, then sonicated in 1\u00a0M KOH for 5\u00a0min and then rinsed with water and dried in air. The experiments was carried out using purified N2-saturated 1\u00a0M KOH electrolytes. The laser is emitted on the working electrode through a transparent quartz glass window that reduces contamination and interference. All the experiments were conducted at room temperature (20\u00a0\u00b1\u00a02 \u00b0C). All the data points were processed using origin software.In situ Raman-chronoamperometry study was done at \u22120.4\u00a0V vs. RHE for 30,000\u00a0s for NiCu catalysts. The Raman spectra were collected at the applied potential in 1\u00a0M KOH every 10 sweeps (10\u00a0s/sweep) from 100 to 2000 cm\u22121. The spectrum shift of silicon wafer Raman peak at 520.7 cm\u22121 was used for calibration.Catalyst inks were fabricated by mixing catalyst powder with water: isopropanol (1:1), and ionomer (Fumion FAA-3-SOLUT-10 (Fuel Cell Store)). The solution was sonicated for 30\u00a0min to ensure fine and well-dispersed ink. Cathode catalysts loadings were 1\u00a0mg/cm2 for Pt/C (60 wt% metal on support, Alfa Aesar) and 5\u00a0mg/cm2 for 60 wt% NiCu MMO/Ketjen black. An Ir black benchmark catalyst with a loading of 3\u00a0mg/cm2 (Alfa Aesar) was used at the anode for all MEAs. Catalyst layers were sprayed at 60 \u00b0C using a Coltech airbrush (0.35\u00a0mm nozzle) on Toray 090 carbon paper (25 cm2, Fuel Cell Store) for the cathode, and Ti felt (Bekaert Inc.) coated with Au for anode as catalyst coated substrates (CCSs). The area of carbon paper equals the area of Ti felt and represents the electrode surface area (25 cm2). The Ti felt was pretreated by etching in HCl (37 wt%, Sigma Aldrich) for 2\u00a0min to remove the non conductive surface oxide and then sonicated for 5\u00a0min in water and ethanol before being sputter-coated with Au using an Edwards sputter coater to reduce interfacial contact resistance (ICR) within the cell. The coating was carried out at a vapor deposition pressure of 0.15 atm at 20\u00a0mA for 2\u00a0min on each side. The ionomer content amounted to 25 and 7\u00a0wt% of the total solids in ink for cathode and anode, respectively. The membrane, Fumapem-3-PE-30, was sandwiched between cathode and anode gas diffusion electrodes as in Fig.\u00a02\n. The MEAs were conditioned and exchanged to the OH form in 1\u00a0M KOH overnight. The AEM water electrolyzer setup consisted of a 5\u00a0L Teflon tank with heaters and a peristaltic pump. Tests were conducted at T\u00a0=\u00a050 \u00b0C. The concentrations of KOH employed were 1 and 0.1\u00a0M KOH (ACS reagent, \u226585%, pellets, Sigma Aldrich). The flow rates of the pumps were 250\u00a0ml/min.A high-current potentiostat (HCP-803, Biologic) was used to control cell voltage and measure impedance in the single-cell measurements. The polarization curve was recorded galvanostatically, ramping the current from 0 to 2 A/cm2 at a rate of 80\u00a0mA/cm2 per minute. Electrochemical impedance spectroscopy (EIS) was employed to determine the cell resistances and performed at different current densities, such as 0.2 A/cm2, in the AC frequency range of 100 kHz\u20131\u00a0Hz. The NiCu MMO catalytic layers were post analyzed by SEM and EDX.SEM and STEM images of the nickel-copper catalyst synthesized by chemical reduction with the addition of Na2CO3 and annealed in 5% H2/Ar (NiCu MMO) are shown in Fig.\u00a03\n. The Figs.\u00a03a and 3b show that NiCu MMO catalysts have dense areas of agglomerated nanosheet morphology. The STEM image in Fig.\u00a03c displays that NiCu MMO nanosheets are loaded on the carbon support (Ketjen black EC-600JD) with a dark thick region of NiCu nanosheets. Fig.\u00a03d confirmed the loosely stacked nanosheets morphology of NiCu MMO catalysts. Similar catalyst morphology produced by chemical reduction by sodium borohydride has given various names from nanocotton [29], nanosponges [30\u201332, and nanosheets [33\u201339] and in this work, we will refer to these catalysts as nanosheets. During the chemical reduction process, sodium borohydride reacts quickly with transition metal cations to precipitate metal boride MxBy species [39\u201342]. In the case of NiCu MMO, Na2CO3 was added during the synthesis process to precipitate oxide species [21]. We investigated another nickel-copper catalyst without the addition of Na2CO3 and annealed the resulted powder in the air (NiCu oxide) and 5% H2/Ar (NiCu alloy) and they exhibited also an agglomerated nanosheets morphology similar to NiCu MMO as seen in (Fig. S1). Energy dispersive x-ray spectroscopy (EDX) of NiCu MMO is shown in Fig.\u00a04\na. The EDX spectrum displays peaks corresponding to Ni, Cu, O, and C with Ni: Cu weight percentage as 52.3:47.7, which is in good agreement with precursors percentage. The EDX spectrum displays peaks corresponding to Ni, Cu, O, and C. Impurities or remaining elements from the synthesis process appear to be absent.The XRD pattern of NiCu MMO in Fig.\u00a04b shows peaks at 2\u03b8 values of 32.5\u00b0, 35.6\u00b0, 37.2\u00b0, 38.9\u00b0, 43.2\u00b0, 44.56\u00b0, 48.9\u00b0, 51.93\u00b0, and 62.8\u00b0. The diffraction peaks at 2\u03b8 values of 44.5\u00b0 and 51.93\u00b0 are associated with Ni (111) and Ni(200) crystal planes of nickel face-centered cubic (FCC) structure with (JCPDS card No. #04\u20130850) [43]. The peaks at 2\u03b8 values at 37.2\u00b0, 43.2\u00b0, and 62.8\u00b0 correspond to (111), (200), and (220) diffraction planes of NiO (JCPDS card no. #47\u20131049) [44]. The diffraction peaks at 2\u03b8 values of 32.5\u00b0, 35.6\u00b0, 38.9\u00b0, 48.9\u00b0 values correspond to CuO crystal structure (JCPDS card no. #80\u20130076) [45].NiCu alloy shows peaks at 2\u03b8 values of 44.5\u00b0 and 51.93\u00b0 that correspond to pure Ni (JCPDS No. 04\u20130850) [43] while the peaks at 44\u00b0 and 51.2\u00b0 correspond to pure Cu (JCPDS No. 04\u20130836) [46]. While NiCu oxide shows peaks at 2\u03b8 values of 37.2\u00b0, 43.2\u00b0, and 62.8\u00b0 correspond to NiO (JCPDS card no. #47\u20131049) and peaks at 2\u03b8 values of 32.5\u00b0, 35.6\u00b0, 38.9\u00b0, 48.9\u00b0 of CuO crystal structure (JCPDS card No. #80\u20130076) [44,45]. The NiCu MMO vibrational modes were characterized by Raman spectroscopy in Fig.\u00a04c. The Raman spectrum in Fig.\u00a04c shows Raman peaks at 490, 606, 810, 1020, and 1100 cm\u22121 respectively. The Raman peak at 490 cm\u22121 corresponds to Cu(OH)2 while the Raman peak 606 cm\u22121 corresponds to the Bg Raman mode of CuO [47\u201350. The Raman peaks at 810, 1020, and 1100 cm\u22121 correspond to two-phonon (2P) NiO vibrational modes [51\u201355.XPS analysis provides sensitive information about the surface chemical composition of NiCu MMO catalyst. NiCu MMO survey spectrum is shown in Fig.\u00a04d. The survey spectrum indicates the presence of Ni, Cu, B, O, and C peaks. Ni 2p high-resolution XPS spectrum is shown in Fig.\u00a05\na and 5b. The Ni 2p XPS spectrum is divided into two main peaks (Ni-2p1/2 and Ni-2p3/2) due to the spin-orbit effect and two oxidation states for nickel (Ni0 and Ni2+) can be deconvoluted. The XPS peaks at 853.8\u00a0eV and 871.4\u00a0eV can be assigned to Ni 2p3/2 and Ni 2p1/2 of Ni0\n[20,56]. The XPS peaks located at 855.4\u00a0eV with a satellite at 860.9\u00a0eV correspond to Ni 2p3/2 of Ni2+. The peak at 872.5\u00a0eV with a satellite at 879.4\u00a0eV can be attributed to Ni 2p1/2 of Ni2+\n[20,56] Cu-2p high-resolution spectrum is shown in Fig.\u00a05c and 5d. The XPS peaks at 932.6\u00a0eV and 952.4\u00a0eV correspond to Cu 2p3/2 and Cu 2p1/2 of Cu0\n[57]. The XPS peak at 933.7\u00a0eV corresponds to CuO [58]. The peaks at 934.8 and 954.4\u00a0eV are associated with Cu(OH)2\n[58]. Cu(OH)2 appears to form due to CuO reaction with chemisorbed water on the catalyst surface.The high-resolution Ni 2p XPS spectrum in NiCu alloy exhibits peaks at 852.4 and 869.5\u00a0eV which correspond to Ni 2p3/2 and Ni 2p1/2 peaks of metallic Nio) Fig. S2a ([59]. The Cu 2p spectrum shows two peaks at 932.5 and 952.3\u00a0eV which are assigned to Cu 2p3/2 and Cu 2p1/2 of metallic Cu0 (Fig. S2b) [16].The high-resolution XPS spectrum of Ni 2p in NiCu oxide shows that the Ni 2p3/2 main peak and its satellite at 854 and 862\u00a0eV, and the Ni 2p1/2 main peak and its satellite at 872 and 879\u00a0eV, respectively confirming the presence of Ni+2 state (Fig. S2c) [60]. The high-resolution XPS spectrum of the Cu 2p spectrum of NiCu oxide shows peaks at 933.7, 943.1, 954.3, 962.9\u00a0eV. The peaks at 933.7 and 954.3\u00a0eV correspond to the Cu 2p3/2 and Cu 2p1/2, respectively. Also, there are two satellite peaks centered at about 943.1 and 962.9\u00a0eV, demonstrating the presence of Cu+2 state (Fig. S2d) [61].Based on the structural characterization. NiCu mixed metal oxide (MMO) nanosheets have Ni, NiO, CuO phases and hydroxide species such as Cu(OH)2 which can be beneficial for HER in alkaline electrolytes [62] as we will see from the electrochemical measurements. NiCu alloy contains pure Ni and pure Cu phases while NiCu oxide contains NiO and CuO phases.\nFig.\u00a06\na shows linear sweep voltammetry (LSV) curves of NiCu alloy, NiCu MMO, and NiCu oxide in 1 and 0.1\u00a0M KOH. All catalyst loadings were equal to 250\u00a0\u00b5g/cm2. NiCu MMO has the highest HER activity in 1\u00a0M KOH by achieving \u221210\u00a0mA/cm2 at \u2212200\u00a0mV compared to the \u2212250 and \u2212300\u00a0mV for NiCu alloy and NiCu oxide, respectively, to obtain the same current density. As seen from Fig.\u00a06a, the current density normalized to geometric surface area for NiCu MMO at \u22120.35\u00a0V vs RHE in 1\u00a0M KOH is five times higher than 0.1\u00a0M KOH. However, the activity trend for the nickel-copper catalysts is the same in 0.1\u00a0M KOH. Fig.\u00a06b shows a comparison between the NiCu MMO HER activity and data from the literature. The NiCu MMO shows one of the best mass active HER catalytic activities reported in Table S1, Table S2, and Fig.\u00a06b.The LSV curves in Fig.\u00a06a show that the HER activity increases with increasing KOH electrolyte concentration, which in agreement with literature [63,64]. Lasia et\u00a0al. found that the rate constants of Volmer and Heyrovsky reactions depend on the bulk OH\u2212concentrations [65]. An appropriate rise of the KOH electrolyte concentration increases hydroxide ion activity [64\u201367]. Recently Wang et\u00a0al. showed that the high HER activity at high KOH concentration is due to H3O+ intermediates generated on nanocatalyst surface [68].Electrocatalytic active surface area (ECSA) measurements were carried out to evaluate the intrinsic catalytic activity of nickel-copper catalysts. The ECSA was estimated by measuring the electrochemical double-layer capacitance (Cd\nl) from cyclic voltammograms at various scan rates over a non-faradaic (totally polarized) potential range, as in Fig. S3 in the Supplementary Information. The NiCu MMO catalysts exhibit the largest double-layer capacitance Cdl of 9.16 (mF/cm2) compared to those of the NiCu alloy (6.58 mF/cm2) and the NiCu oxide (3.80 mF/cm2), showing that a larger ECSA of NiCu MMO allows more exposed active sites to promote HER performance. The specific surface area of the NiCu catalysts was also investigated with Brunauer\u2013Emmett\u2013Teller (BET) measurement (Figure S3, ESI\u2020). The specific surface area has a similar trend as ECSA. NiCu MMO possesses a surface area of 156 m2/g which is far higher than that of NiCu alloy (112 m2/g) and NiCu oxide (92 m2/g). When normalized to electrochemical surface area (Fig. S3, ESI\u2020) the differences in catalyst activity become less, especially in the lower potential range. However, NiCu MMO still possesses the highest intrinsic activity.The linear regions of Tafel plots in Fig.\u00a06c are fitted to the Tafel equation, yielding Tafel slopes of 120, 130, and 195\u00a0mV/dec for NiCu MMO, NiCu alloy, and NiCu oxide respectively. The kinetic parameters for the nickel-copper catalysts (jo and b) presented in Table\u00a01\n were derived from the Tafel equation:\n\n(4)\n\n\n\u03b7\n=\na\n+\nb\nlog\nj\n\n\n\nWhere \u03b7 (V) is the applied overpotential, j (mA/cm2) is the current density, b (V/dec) is the Tafel slope, and a (V) is the intercept.The exchange of current density jo can be obtained by extrapolating the Tafel plots to the x-axis or assuming \u03b7 is zero.\n\n(5)\n\n\n\n\na\n\n\n=\n\n\n\n\n(\n2.3\nRT\n)\n\n/\n\n(\n\u03b1\nF\n)\n\nlog\n\nj\no\n\n\n\n\n\n\nb\n\n\n=\n\n\n\n\n(\n2.3\nRT\n)\n\n/\n\n(\n\u03b1\nF\n)\n\n\n\n\n\n\n\nwhere R is the gas constant (8.314\u00a0kJ mol\u22121\nK\u00a0\u2212\u00a01), T is the temperature in K, \u03b1 is the charge-transfer coefficient, and F is the Faraday constant (96,485 C mol\u22121).\nTable\u00a01 summarizes the kinetic parameters for nickel-copper catalysts. NiCu MMO shows the lowest Tafel slope and highest charge transfer coefficient and exchange current density over NiCu alloy and NiCu oxide which confirms the superior activity of NiCu MMO.In view of the Tafel slope being close to 120\u00a0mV/dec, it is likely that the charge transfer coefficient represents the symmetry factor of the Volmer step in this case. The Tafel slopes reflect an intensive property of the HER catalysts from which some indication about the reaction mechanism of the HER and the rate-determining step (rds) can be obtained. The Volmer reaction involves the electroreduction of water molecules with hydrogen adsorption as in Eq.\u00a0(6), while the Heyrovsky's reaction involves electrochemical hydrogen desorption eq\u00a0(7). The Tafel reaction involves chemical desorption Eq.\u00a0(8)\n[65].\n\n(6)\n\n\n\nM\n\n+\n\n\nH\n\n2\n\n\nO\n\n+\n\n\n\ne\n\n\n\u2212\n\n\u2194\n\nM\n\n\n\nH\n\nads\n\n+\n\nO\n\n\n\n\nH\n\n\n\u2212\n\n\n\nVolmer\n\n\n\n\n\n\n(7)\n\n\n\n\nMH\n\nads\n\n+\n\n\nH\n\n2\n\n\nO\n\n+\n\n\n\ne\n\n\n\u2212\n\n\u2194\n\n\nH\n\n2\n\n+\n\nM\n\n+\n\nO\n\n\n\n\nH\n\n\n\u2212\n\n\nHeyrovsky\n\n\n\n\n\n\n(8)\n\n\n\n\nMH\n\nads\n\n+\n\n\nMH\n\nads\n\n\u2194\n\n\nH\n\n2\n\n+\n\nM\n\n\nTafel\n\n\n\nA detailed analysis shows that rds for the HER at NiCu MMO is the Volmer reaction, then a Tafel slope in the order of 120\u00a0mV would result. Whether the next step in the reaction sequence is the Heyrovsky or Tafel step [65] cannot be determined by this analysis, however.\nFig.\u00a06d shows impedance complex plane plots of NiCu alloy, NiCu oxide, and NiCu MMO in 0.1 and 1\u00a0M KOH at an applied potential of \u2212250\u00a0mV vs. RHE after subtracting ohmic resistance. In the complex plane plots, only one semicircle is observed, which can be attributed to a charge transfer process related to the HER [70\u201372]. The charge transfer resistance (Rct) is represented by the diameter of the semicircle. The radius of the semicircle decreases at higher KOH concentration, signifying a lower charge transfer resistance (Rct) and a higher rate of hydrogen evolution. NiCu MMO exhibit Rct value of 6.96 \u03a9 at an applied potential of \u2212250\u00a0mV compared to NiCu alloy (10.38 \u03a9) and oxide (13.81 \u03a9) which further confirm the superior activity, faster reaction kinetic and high electron transfer efficiency of NiCu MMO [72].The equivalent circuit for the NiCu alloy, NiCu MMO, and NiCu oxide in Fig.\u00a06d is characterized by a single time constant, and we modeled the impedance by a series resistance (Rs, related to ohmic solution resistance), in series with one parallel circuit consisting of a charge transfer resistance (Rct) and a constant phase element (CPE) related to the double-layer capacitance. This equivalent circuit has previously been used in literature to describe HER on polycrystalline Ni and Ni-based materials [73]. (The constant phase element (CPE) was used instead of capacitance due to frequency dispersion and the appearance of depressed semicircles in the impedance plane plots). The charge transfer resistance (Rct) represents the kinetics of the HER at the electrode/electrolyte interface. The absence of Warburg impedance indicates that mass transport is rapid enough so that the reaction is kinetically controlled [70,71,74]. The impedance complex plane plots for different applied overpotentials are shown in Fig. S4a in the ESI\u2020, and these show that the Rct decreases with increasing potential as it would if the current-voltage relationship is described by Eq.\u00a0(4) above. (The lower Rct value at higher potential reflects the exponential dependence of the current on the overpotential and thus the accelerated electron transfer and higher rates of the HER at higher overpotential [70,71,74]. As can be shown by a simple differentiation of Eqn.\u00a04 above, the Tafel slope may be obtained from plots of potential vs. log (Rct)\u22121. From our plots, we obtain 120\u00a0mV/dec, see Supplementary Information, Figure S4b, which is the same as that obtained from the LSV curves. The same Tafel slope being obtained with impedance spectroscopy thus validates the iR-corrected linear-sweep voltammograms.\nFig.\u00a07\na shows the current vs. time recorded in chronoamperometric measurements performed by applying a constant potential of \u2212400\u00a0mV for 30\u00a0h on NiCu alloy, NiCu MMO, and NiCu oxide. For all samples but NiCu MMO in 0.1\u00a0M KOH, the current density decreases (i.e. the activity decreases) rapidly during the first few minutes. For NiCu MMO in 0.1\u00a0M KOH and NiCu oxide in 1\u00a0M KOH, the current density then levels off and remains constant at approximately \u221211\u00a0mA/cm2 and \u221250\u00a0mA/cm2, respectively. NiCu oxide shows stable performance at \u221250\u00a0mA/cm2 for 30\u00a0h. For NiCu alloy in 1\u00a0M KOH, the current density slowly increases (activity increases) with time from approximately \u221290\u00a0mA/cm2 to \u2212100\u00a0mA/cm2 after 30\u00a0h. For the NiCu MMO sample in 1\u00a0M KOH, the current density reaches a minimum activity after approximately 30\u00a0min at which the current density is approximately \u2212180\u00a0mA/cm2, and then slowly increases to a little below \u2212200\u00a0mA/cm2 at 30\u00a0h. The chronoamperometric measurements confirm the higher HER activity of NiCu MMO in 1\u00a0M KOH than in 0.1\u00a0M KOH and over NiCu alloy and NiCu oxide in the same electrolyte. The current densities observed from chronoamperometry are in good agreement with those observed in the LSVs.\nFig.\u00a07b shows in-situ Raman spectra of NiCu MMO under an applied potential of \u22120.4\u00a0V vs RHE at different time intervals. All the spectra in Fig.\u00a07b display peaks at 292, 530, 1060, 1350, and 1585 cm\u22121. The peak at 292 cm\u22121 can be assigned to copper hydroxide Cu(OH)2 species [47] while the peak at 530 cm\u22121 can be assigned to nickel hydroxide Ni(OH)2\n[75]. The peak at 1060 cm\u22121 can be assigned to carbonates [76] while the peaks at 1346, and 1585 cm\u22121 correspond D band, and G band peaks of carbon respectively [77,78]. The spectra show clear peaks of Ni(OH)2 and Cu(OH)2 at the beginning of HER. However, the Cu(OH)2 peaks decreased more significantly than the Ni(OH)2 peaks. In other words, whereas both Cu(OH)2 and Ni(OH)2 both exist during the entire period of 30,000\u00a0s, both Cu and Ni hydroxides get slowly reduced, but Cu more so than Ni. The reduction of these elements is consistent with the Pourbaix diagrams of Cu and Ni which predict that both Ni(OH)2 and Cu(OH)2 would be reduced to metallic nickel and copper at this potential [79].The NiCu oxide also displayed peaks corresponding to Ni(OH)2 and Cu(OH)2, but these peaks disappeared completely after 15,000\u00a0s, as shown in Fig.\u00a07c. However, the HER activity of the NiCu oxide is much lower than that of NiCu MMO. This confirms the importance of the presence of metallic species on the catalyst surface, as found by Danilovic et\u00a0al. [80], for superior HER activity.Finally, the Raman spectrum for the NiCu alloy catalyst also shows peaks related to Ni(OH)2 and Cu(OH)2 surface species when the catalyst is immersed in KOH, which confirmed the hypothesis that Ni metal will convert to oxide species once in contact with KOH [80,81]. The hydroxide species on the NiCu alloy gets reduced rather rapidly (< 5000\u00a0s) at the surface, as shown in Fig.\u00a07d, compared to NiCu MMO and NiCu oxide and did not lead to exceptional activity compared to NiCu MMO.NiCu MMO thus showed the best HER activity in the alkaline electrolytes with a Tafel slope of 120\u00a0mV/dec. The bifunctional system of NiCu MMO catalyst includes Ni metal, NiO, and CuO oxides, and provide a rapid Volmer step and thus rapid overall HER reaction kinetics. The improved HER kinetics of the NiCu MMO can be attributed to the presence of both Ni and NiO where NiO sites to facilitate water dissociation and bind OHad while Ni metallic binds Hads and CuO stabilizes NiO under HER conditions. Similarly, Bates et\u00a0al. found that the synergistic HER enhancement of Ni/NiO is due to NiO content and Cr2O3 appears to stabilize NiO under HER conditions [82]. The in situ Raman results show that the presence of both metal and oxide phases is essential to sustain a high HER activity, the performance of NiCu MMO relative to that of NiCu alloy or NiCu oxide. We relate this to the in situ Raman data showing that copper hydroxide gets reduced and nickel hydroxide is to some extent preserved under HER conditions.We attribute the rapid decay in electrocatalytic activity in all samples to an initial and rapid adjustment of the surface state of all catalysts, whereas the long-term behavior is more complex. For the NiCu oxide, there is no further change in the surface state after 15000s (Fig.\u00a07c), and the electrocatalytic activity remains the same as that immediately after the initial transient. The current transient is thus fully consistent with the Raman spectra for NiCu oxide in Fig.\u00a07c. Since the Raman spectra of the NiCu alloy (Fig.\u00a07d) indicate a surface at which hydroxides are completely absent after 5000\u00a0s, however, the slow increase in catalytic activity with time in Fig.\u00a07a for this catalyst may be related to a slow change in the composition or surface area, i.e. to catalyst instability. For NiCu MMO in 1\u00a0M KOH, the initial transient is followed by a slower increase in catalytic activity. A correspondingly slow change in the surface state, c.f. the Raman spectra in Fig.\u00a07b, appear to persist throughout the chronoamperometry experiment.We relate this difference to the synthesis. For the NiCu alloy, only a thin layer of the hydroxides will form as the NiCu alloy is exposed to the KOH solution. This layer is rapidly reduced as the catalyst is subjected to a negative potential. However, since there is no indication of any metal phase in the NiCu oxide in the diffractograms, we may assume that during exposure to negative potentials these catalysts will be reduced continuously until the entire catalyst is converted to metal. For the NiCu MMO, this seems to have combined behavior (mixed metal oxide (Ni-NiO-CuO) catalyst), since the oxidation due to the annealing is not complete, c.f. the diffractograms in Fig.\u00a04b which displays a substantial peak corresponding to Ni(111). This catalyst heterogeneity of metallic and oxide phases will cause mixed behavior of a continuous but slow change in the surface state throughout the experiments, which may be related to a slow diffusion-limited process in the sample. The surface is therefore also slowly reorganized and will consist of a mix of phases and a slowly changing activity.NiCu MMO also showed good stability during an accelerated stress test consisting of 5000 potential cycles from between \u22120.8 to \u22121.3\u00a0V at a scan rate of 100\u00a0mV/s. The LSV for NiCu MMO before and after the procedure showed only a 20\u00a0mV difference in the potential required to achieve - 100\u00a0mA/cm2 as shown in Fig. S5 ESI\u2020.\nFig.\u00a07e shows The HER activity using Nafion and anion exchange ionomer (Fumion ionomer) of NiCu alloy, NiCu MMO, and NiCu oxide. The activity for the HER of nickel-copper catalysts decreased if Fumion ionomer replaced Nafion in the catalyst ink, and resulted in a potential shift of 30\u00a0mV at \u2212100\u00a0mA/cm2 as compared to the Nafion ionomer. Catalyst inks with Nafion resulted in higher HER activity compared to catalyst inks with Fumion ionomer. We assign the difference in activity between catalysts in inks with Nafion and those with Fumion ionomers to the nature of the ionomer backbone and its chemistry (ammonium-, imidazolium-, phosphonium-based compounds in anion exchange ionomers such as Fumion, or sulphonic acid groups (SO3\n\u2212) in Nafion) [83,84]. The SO3\n\u2212 moiety in Nafion interacts only weakly with the catalyst surface, and the effect of SO3\n\u2212 adsorption on electrocatalyst performance is expected to be negligible, particularly in the HER region where the negative charge on the catalyst surface would repel sulfonate species [21]. The quaternary ammonium (QA) functional group used for OH\u2212 transport in anion exchange ionomer (AEI), on the other hand, appears to poison NiCu MMO catalyst and block active catalyst sites. Fumion ionomer shows higher total polarization resistance than Nafion as shown in the impedance complex plane plot of NiCu MMO using Nafion and Fumion ionomers (Fig. S6.a). The small semicircle at the low-frequency region for Fumion ionomer (Fig. S6.a) has been suggested to correspond to quaternary ammonium adsorption [85]. The results show that the anion ionomer not only serves as a binder but also affects the electrocatalyst's HER activity [4].We consequently investigated the impact of the Nafion ionomer content to find the composition at which the HER performance peaks for NiCu MMO. The results are shown in Fig.\u00a07f. The HER activity thus increases with increasing Nafion ionomer to the catalyst weight ratio (I/C), and a maximum appears at a weight ratio of I:C of 0.5. The NiCu MMO at I/C\u00a0=\u00a00.5 achieves \u221210\u00a0mA/cm2 at 170\u00a0mV, which indicates better catalyst utilization, lower total polarization resistance, and optimum HER performance as shown in Fig. S6b and S6c. The low performance with a low ionomer content is attributed to the poor dispersion of the ink. At high ionomer content, the HER activity is small due to increased aggregation of Nafion and the associated blocking of mass transport and active sites [25]. The moderate I/C ratio indicates that Nafion improves the catalyst dispersion and distribution and reduced transfer resistance. The optimized ionomer content provides an efficient pathway for OH\u2212 (in the aqueous electrolyte) and electrons and forms a stable reaction interface [86].To test the activity of NiCu MMO in an actual AEM electrolysis environment, NiCu MMO and Ir black MEAs were fabricated and mounted in an AEM water electrolysis cell as explained in the Supplementary Information and Fig. S7 ESI\u2020. Two types of MEAs will be mentioned in Results Pt/C cell and NiCu MMO cell for NiCu MMO-Ir and Pt/C-Ir cells respectively. Fig.\u00a08\n shows the impedance complex-plane plot at 0.2 A/cm2 for NiCu MMO cells (Fig.\u00a08a) and Pt/C cells (Fig.\u00a08b) in 0.1 and 1\u00a0M KOH. The impedance complex-plane plots appear to consist of two partly overlapping and depressed semicircles. The ohmic resistance of the cell was determined from the high-frequency resistance (HFR), i.e., from the intercept with the real (Re) axes [87].In Fig.\u00a08, we show the equivalent circuit that is used to fit the impedance data taken at 0.2 A/cm2 in both NiCu MMO and Pt/C cells. We assign the low-frequency arc to mass transport [87,88] and the high-frequency arc to electrode kinetics contributions to the cell voltage from the NiCu MMO and Pt/C cathodes. The fitted electrical circuit is comprised of a series combination of two parallel circuits each consisting of a resistance and a constant phase element (CPE), in series with a resistor, R\u03a9. The R\u03a9 corresponds to the ohmic resistance of the cell (catalyst layer, current collectors and membrane). The Rct describes the charge transfer resistance of the cathode and anode. CPE1 is the constant phase element that represents the electrode roughness. The circuit has an additional RC combination, constant phase element, and the resistance (CPE2 and R1), which is suggested to describe the mass transport related to bubble formation at the electrode-electrolyte interface [88]. All parameters extracted from the fitting of the impedance data to are presented in Table S3. For 1\u00a0M KOH, NiCu MMO cell has an HFR of 0.195 \u03a9.cm2 while Pt/C based AEMWE cell has an HFR of 0.115 \u03a9.cm2. NiCu MMO cell displays an HFR of 0.295 \u03a9.cm2 while Pt/C achieves 0.225 \u03a9.cm2 in 0.1\u00a0M KOH. NiCu MMO (5\u00a0mg/cm2) higher loadings resulted in thicker catalyst layers and higher HFR compared to Pt/C. The HFR increases as KOH concentration decreases to 0.1\u00a0M KOH. This HFR increase with decreasing KOH concentration may indicate insufficient ionic conductivity of the membrane [26].The impedance data were converted to Tafel impedance. The Tafel slope can be estimated from the Tafel impedance, for a kinetically limited process, as the diameter of the impedance arc [89]. The Tafel impedance shown in Fig. S8 ESI\u2020 is the impedance multiplied with the steady-state current density at which it was obtained. We thus estimate the Tafel slope in 1\u00a0M KOH to be 40\u00a0mV for Pt and 65\u00a0mV for NiCu MMO at 0.2 A/cm2. The Tafel slope from the impedance data is in the range of 50 millivolts, whereas the slopes from the polarization curve are twice this value (see Fig. S9 ESI) suggested that the polarization curves are dominated by the ohmic resistance. Fig.\u00a08c and 8d show the potentiostatic polarization curves of both HFR-corrected and uncorrected voltages for the AEMWE at different KOH concentrations for NiCu MMO and Pt/C cells.\nFig.\u00a08c and 8d show the AEM electrolyzer performance of NiCu MMO and Pt/C cathode catalysts in 1 and 0.1\u00a0M KOH at 50 \u00b0C using Ir black as an anode. In 1\u00a0M KOH, with NiCu MMO a cell performance of 1.85 A/cm2 at 2\u00a0V achieved, which may be compared to Pt /Ir cell that delivers 2 A/cm2 at 2\u00a0V in 1\u00a0M KOH while both cells achieved 1 A/cm2 at 2\u00a0V in 0.1\u00a0M KOH. The increase in KOH electrolyte concentration leads to a higher AEM electrolyzer performance.\nFig.\u00a08d showed that NiCu MMO cell exhibits higher performance than Pt/C catalyst when HFR-corrected. NiCu MMO (5\u00a0mg/cm2) shows higher resistance than Pt/C (1\u00a0mg/cm2) in 1 and 0.1\u00a0M KOH (Fig.\u00a08a and 8b). Since the cell hardware, components, electrolyte, temperature is the same and the only difference is the cathode catalyst, the origin of high resistance is the higher loading and the presence of oxide species in NiCu MMO (Ni-NiO-CuO). This leads to a higher resistance in the NiCu MMO catalytic layer itself as compared to the Pt/C, with its lower loading and metallic conductivity.The results suggest that the differences in the activity of the samples (Fig.\u00a08c) are not merely due to their different intrinsic activities, but also partly due to low electronic resistance in the catalytic layer. This contribution to the resistance will be particularly significant for poorly conducting oxides such as those of NiCu MMO. The high-frequency resistance (HFR) corrected polarization curves in Fig.\u00a08d confirm that the electronic resistance of the cathode catalyst layer significantly affects cell performance. Similar results can be found in the literature. Yu et\u00a0al. [90] showed that for catalysts with widely different conductivity the ranking depends on whether iR compensation is applied or not. Xu et\u00a0al. [91] referred the differences in AEM electrolyzer performance partially to differences in the OER catalyst phases electrical conductivity. Finally, D. Chung et\u00a0al. [92] showed that poorly conductive MoS2 HER activity is affected by the ohmic losses and recommend that electrical conductivity should be considered when designing active catalysts for water electrolysis.The NiCu MMO/Ir MEA activity shows a good reproducibility for three different MEAs in 1\u00a0M KOH at 50 \u00b0C as in Fig. S10 ESI\u2020. The post-mortem analysis of NiCu MMO catalytic layers shows no visible cracks which prove the stability of catalytic layers during AEM water electrolysis as indicated in Fig. S11. Energy dispersive X-ray (EDX) mapping of NiCu MMO catalytic layers and cross-section shows the presence of nickel, copper, carbon, and a thin potassium layer after the electrolysis experiment Fig. S12, and S13 ESI\u2020.The excellent performance of 1.85 A/cm2 at 2\u00a0V in 1\u00a0M KOH obtained for the NiCu MMO hydrogen catalyst outperforms most of those summarized in (Fig. S14 and Table S4 ESI\u2020) allows for an active and cheap catalyst for AEM water electrolysis operation on a commercial scale [93,94] and comparable to the state of the art performance of PEM electrolysis as summarized by Ayers et\u00a0al. [95].NiCu mixed metal oxide (MMO) nanosheets synthesized by chemical reduction showed an exceptional activity for the HER compared to NiCu alloy and NiCu oxide catalysts, with higher performance in 1\u00a0M KOH than 0.1\u00a0M KOH. The improved HER kinetics of the NiCu MMO bifunctional system can be attributed to the presence of both Ni and NiO where NiO sites to facilitate water dissociation and bind OHad while Ni metallic binds Hads and CuO stabilizes NiO under HER conditions. In situ Raman spectroscopy at the NiCu MMO catalysts showed that a substantial fraction of in situ formed nickel hydroxide remained after 30,000\u00a0s at HER conditions, which may explain why the NiCu MMO is able to maintain its very high activity as compared to that of NiCu alloy and NiCu oxide over longer periods of time. Despite that anion exchange ionomers would be expected to be suitable ionomers in an AEM environment, the application of anion exchange ionomers in catalytic layers resulted in a lower HER activity as compared to catalytic layers with Nafion as the ionomer. Using Ir black as an anode catalyst, cells with NiCu MMO nanosheets as cathode catalyst achieved AEM electrolyzer performance of 1.85 A/cm2 at 2\u00a0V in 1\u00a0M KOH at 50 \u00b0C.\nAlaa Y. Faid: Conceptualization, Methodology, Investigation, Writing - original draft, Writing - review & editing. Alejandro Oyarce Barnett: Funding acquisition, Supervision, Writing - review & editing. Frode Seland: Supervision, Writing - review & editing. Svein Sunde: Funding acquisition, Supervision, Writing - review & editing.\u201cThere are no conflicts to declare.\u201dThis work was performed within HAPEEL project \u201cHydrogen Production by Alkaline Polymer Electrolyte Electrolysis\u201d financially supported by the Research Council of Norway-ENERGIX program contract number 268019 and the INTPART project 261620. The Research Council of Norway is acknowledged for the support to the Norwegian Micro- and Nano-Fabrication Facility, NorFab, project number 245963/F50.Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.electacta.2021.137837.\n\n\nImage, application 1\n\n\n\n", "descript": "\n                  We report on the optimization of nickel-copper catalysts for superior performance as a cathode catalyst in anion exchange membrane (AEM) water electrolysis. The bifunctional system of NiCu mixed metal oxide (MMO) nanosheets includes Ni metallic, NiO, and CuO oxides provide rapid kinetics for the hydrogen-evolution reaction (HER) of the Volmer step. In-situ Raman spectroscopy for NiCu MMO proved that nickel hydroxide was sustained under HER conditions for at least 30,000\u00a0s, which may explain why the exceptional activity of NiCu MMO as compared to other nickel-copper catalysts is maintained over time. The activity of the NiCu MMO for the HER activity in alkaline electrolytes increased as KOH concentration raised from 0.1\u00a0M to 1\u00a0M. The NiCu MMO nanosheets showed superior stability under alkaline HER conditions for 30\u00a0h. The use of Nafion ionomer in the ink resulted in a higher HER current density as compared to inks with a Fumion anion ionomer. The maximum HER performance was achieved at a Nafion ionomer to catalyst weight ratio of 0.5. Using Ir black as the anode, the NiCu MMO cathode gave an AEM electrolyzer performance of 1.85 A/cm2 at 2\u00a0V in 1\u00a0M KOH at 50 \u00b0C. The NiCu MMO catalyst developed here delivers AEM performance comparable to PEM water electrolysis.\n               "}
{"full_text": "The environmental benefit of hydrogen (H2) as an energy vector stems from its zero carbon footprint [1\u20134]. However, more than 95% of H2 is currently produced from fossil fuels worldwide, leading to even more production of carbon dioxide (CO2) than that produced from the direct use of fossil fuels [5]. Hence, it is necessary to minimize the dependence on fossil fuels and to shift toward renewable and clean resources for hydrogen production [6,7]. Ethanol (C2H5OH) is the most widely used liquid fuel made from renewable biomass and has a relatively high H/C ratio, which is desirable for hydrogen production [8]. Ethanol can be reacted directly with water through steam reforming to produce a H2-rich gas over 3d transition metal catalysts (Eq. (1)) [9]. This process utilizes the raw product of bioethanol, which avoids the energy-consuming distillation separation of the ethanol\u2013water mixture [9\u201311].\n\n(1)\n\n\n\n\nC\n\n\n2\n\n\n\n\nH\n\n\n5\n\n\nOH\n+\n3\n\n\nH\n\n\n2\n\n\nO\n\u2192\n2\n\n\nCO\n\n\n2\n\n\n+\n6\n\n\nH\n\n\n2\n\n\n,\n\n\n\n\n\u0394\n\n\nH\n\n\n298\nK\n\n\n\u2296\n\n\n=\n173\n\nkJ\n\u00b7\n\n\nmol\n\n\n-\n1\n\n\n\n\n\n\nHowever, ethanol steam reforming is a strongly endothermic reaction. Chemical looping steam reforming (CLSR), as a process intensification technology, can be employed to promote the efficiency of the steam reforming process [12,13]. In the CLSR of ethanol, the oxygen carrier (OC) is first reduced by ethanol in a reforming reactor. For example, when NiO is used as the OC, the redox reaction between C2H5OH and NiO is carried out as shown in Eq. (2).\n\n(2)\n\n\n\nC\n2\n\n\nH\n5\n\nOH\n+\n6\n\nNiO\n\n\n\n\n\ns\n\n\n\n\u2192\n2\n\nCO\n2\n\n+\n3\n\nH\n2\n\nO\n+\n6\n\nNi\n\n\n\n\n\ns\n\n\n\n,\n\n\n\n\n\u0394\n\nH\n\n298\nK\n\n\u2296\n\n=\n150\n\nkJ\n\u00b7\n\n\nmol\n\n\n-\n1\n\n\n\n\n\n\nThe Ni2+ in NiO is reduced to metallic nickel (Ni) with the depletion of oxygen (O). Next, ethanol steam reforming occurs with the catalysis of metallic Ni (Eq. (1)). The thermal decomposition of ethanol is also carried out on the surface of the Ni when the steam-to-carbon ratio (S/C) is low (Eq. (3)):\n\n(3)\n\n\n\nC\n2\n\n\nH\n5\n\nOH\n\u2192\n\nC\n\n\n\n\n\ns\n\n\n\n+\nCO\n+\n3\n\nH\n2\n\n,\n\n\n\n\u0394\n\nH\n\n298\nK\n\n\u2296\n\n=\n136\n\nkJ\n\u00b7\n\n\nmol\n\n\n-\n1\n\n\n\n\n\n\nAll of Eqs. (1)\u2013(3) are endothermic. Ni is then re-oxidized by air in a regeneration reactor (Eq. (4)). The deposited carbon (C) formed during ethanol steam reforming is also gasified (Eq. (5)).\n\n(4)\n\n\n\nNi\n\n\n\n\n\ns\n\n\n\n+\n0.5\n\nO\n2\n\n\u2192\n\nNiO\n\n\n\n\n\ns\n\n\n\n,\n\n\n\n\n\u0394\n\nH\n\n298\nK\n\n\u2296\n\n=\n-\n187\n\nkJ\n\u00b7\n\n\nmol\n\n\n-\n1\n\n\n\n\n\n\n\n\n(5)\n\n\nC\n\n\n\ns\n\n\n+\n\n\nO\n\n\n2\n\n\n\u2192\n\n\nCO\n\n\n2\n\n\n,\n\n\n\u0394\n\n\nH\n\n\n298\nK\n\n\n\u2296\n\n\n=\n-\n395\n\nkJ\n\u00b7\n\n\nmol\n\n\n-\n1\n\n\n\n\n\n\nThe heat required for endothermic steam reforming can be supplied from the oxidation of OCs (Eq. (4)) and deposited carbon (Eq. (5)) in the regeneration reactor. Therefore, the excess heat required from an external burner can be minimized. The overall reaction of the CLSR of ethanol can be regarded as the sum of ethanol steam reforming and the complete oxidation of ethanol (Eq. (6)).\n\n(6)\nC2H5OH\u2009+\u2009xH2O\u2009+\u2009(1.5\u2009\u2212\u20090.5x)O2\u00a0\u2192\u00a02CO2\u2009+\u2009(3\u2009+\u2009x)H2\n\n\n\nThe OCs, which are normally reducible metal oxides, play essential roles in the CLSR of ethanol. The use of metal oxides instead of gaseous oxygen (O2) as the OCs help to avoid safety risks during operation [8]. The extra oxygen from the OCs can remarkably reduce the S/C of the CLSR, which may lead to autothermal hydrogen production from renewable feedstock with an appropriate ratio of ethanol to OC. The OCs in chemical looping processes must meet a number of criteria for practical applications [14,15]. They must exhibit long-term redox stability and provide oxygen species with suitable activity [16]. NiO, as an outstanding candidate, has been investigated for use as the OC in various chemical looping processes [17]. Jiang et al. [18] applied NiO/montmorillonite in the CLSR of ethanol and achieved greater than 60% H2 selectivity in 20 cycles. However, the oxygen release of bulk NiO is too drastic, and the dispersion of the Ni derived from bulk NiO is relatively inadequate for the activation of reactive species and long-term operation, which limits the stability of Ni-based OCs [19\u201321]. The regulation of the reduction kinetics is the key to obtaining highly dispersed Ni and further promoting the performance of Ni-based OCs.Due to the similar atomic sizes of Ni2+ (69\u00a0pm) and Mg2+ (72\u00a0pm), a substitutional Ni\nx\nMg1\u2212\n\nx\nO solid solution in any proportion (0\u00a0\u2264\u00a0x\u00a0\u2264\u00a01) can be formed by means of an adequate calcination temperature [22\u201324]. The Ni\u2013Ni boundary is isolated by the Mg2+ in Ni\nx\nMg1\u2212\n\nx\nO; thus, the rapid movement of Ni2+ is inhibited in the solid solution [25]. The reduction of the solid solution is related to the rate of bulk Ni2+ diffusion and can be tuned by the concentration of Ni2+ in Ni\nx\nMg1\u2212\n\nx\nO [26]. Huang et al. [27] designed Mg\u2013Ni\u2013Al\u2013O OCs with a solid solution structure and achieved excellent performance in chemical looping combustion. Ni\nx\nMg1\u2212\n\nx\nO shows great potential for chemical looping processes, although the applicability of Ni\nx\nMg1\u2212\n\nx\nO solid solutions as OCs in the CLSR of ethanol remains unclear.In this work, Ni\nx\nMg1\u2212\n\nx\nO solid solutions with different chemical compositions were synthesized as OCs to investigate the modulation effect of Mg2+ on the CLSR of ethanol. With the introduction of Mg2+, the oxygen release of Ni-based OCs was tunable. The relationship between the structural evolution of Ni\nx\nMg1\u2212\n\nx\nO and the mechanism of the surface reaction was investigated. Ethanol\u2013water pulse and H2-temperature-programmed reduction (TPR) experiments were applied to explore the oxygen release of Ni\nx\nMg1\u2212\n\nx\nO. An in situ diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) experiment was also carried out to determine the changes in intermediates during the CLSR of ethanol.A series of Ni\nx\nMg1\u2212\n\nx\nO (x\u00a0=\u00a00.2, 0.4, 0.6, and 0.8) solid solutions was synthesized using a co-precipitation method. Typically, a mixture of Mg(NO3)2\u00b76H2O (98%, J&K Scientific Co., Ltd., China) and Ni(NO3)2\u00b76H2O (99%, Aladdin Biological Technology Co., Ltd., China) was dissolved in 150\u00a0mL of deionized water (18.25 M\u03a9) with a total metal molarity of 2\u00a0mol\u00b7L\u22121. Then, 100\u00a0mL of as-prepared 6\u00a0mol\u00b7L\u22121 NaOH (99%, Aladdin Biological Technology Co., Ltd.) solution was used as the precipitant. The formed precipitate was aged for 12\u00a0h, and the products were filtered and washed thoroughly with hot water to remove sodium. The obtained samples were dried in an oven at 125\u00a0\u00b0C for 24\u00a0h, and then calcined at 700\u00a0\u00b0C in air for 4\u00a0h with a heating rate of 10\u00a0\u00b0C\u00b7min\u22121. NiO and MgO were also synthesized by the precipitation method for reference.The crystalline structures of the samples were measured using powder X-ray diffraction (XRD; Bruker Corp., USA) with a Bruker D8 Focus equipped with Cu K\u03b1 radiation (\u03bb\u00a0=\u00a01.54056\u00a0\u00c5, 1\u00a0\u00c5\u00a0=\u00a010\u221210\u00a0m). The diffraction angle 2\u03b8 ranged from 20\u00b0 to 80\u00b0 with a scan speed of 8\u00b0 per minute. The texture and morphology of the samples were acquired from transmission electron microscopy (TEM) characterization on a JEM-2100F transmission electron microscope (Japan Electronic Materials Corp., Japan) operated at 200\u00a0kV. The samples for TEM analysis were sonicated in ethanol and subsequently supported on copper grid-supported transparent carbon foil. The transmission electron microscope was also equipped with an energy-dispersive X-ray spectroscopy (EDS) detector (Ultim Max, Oxford Instruments, UK) for elemental analysis.The specific surface area and pore volume of the OCs were measured by nitrogen (N2) adsorption\u2013desorption at \u2212196\u00a0\u00b0C using a Micromeritics Tristar II 3020 analyzer (Micromeritics Instrument Corp., USA), based on the Brunauer\u2013Emmett\u2013Teller (BET) and Barrett\u2013Joyner\u2013Halenda (BJH) methods, respectively. Before the tests, all the materials were degassed at 300\u00a0\u00b0C for 3\u00a0h. Elemental compositions of the prepared OCs were determined by means of inductively coupled plasma optical emission spectroscopy (ICP-OES) (VISTA-MPX, Varian, UK). Prior to the measurements, the samples were dissolved in HNO3 solutions.The reduction behavior of the OCs was determined by H2-TPR measurement. The experiments were performed on a Micromeritics Autochem II 2920 instrument equipped with a thermal conductivity detector (TCD; Micromeritics Instrumen Corp., USA). In a typical experiment, the sample (100\u00a0mg) was pretreated at 300\u00a0\u00b0C for 1\u00a0h under flowing argon (Ar; 30\u00a0mL\u00b7min\u22121). After the sample had cooled to 100\u00a0\u00b0C, the analysis was carried out in a mixture of 10\u00a0vol% H2 in Ar (30\u00a0mL\u00b7min\u22121) from 100 to 950\u00a0\u00b0C at 10\u00a0\u00b0C\u00b7min\u22121.To determine the transfer of oxygen species during the CLSR of ethanol, the C2H5OH\u2013H2O mixture and O2-pulse experiments were measured on a Micromeritics Autochem II 2920 instrument equipped with a Hiden QIC-20 mass spectrometer (Hiden Analytical, USA). Prior to the experiments, all the samples were pretreated in situ using a flow of Ar (30\u00a0mL\u00b7min\u22121) at 300\u00a0\u00b0C for 1\u00a0h. Subsequently, pulses of the mixture of C2H5OH and H2O in Ar or 2% O2 in helium (He) were admitted to the reactor. The loop volume was 0.5031\u00a0mL, and the time interval between different pulses was 3\u00a0min, excluding the interference of contiguous pulses. The reactor effluent was continuously monitored by the mass spectrometer, and the gas-phase composition was calculated from the mass spectrometer signal at mass-to-charge ratios (m/z) of 44, 31, 29, 28, 27, 18, 16, and 2 for CO2, C2H5OH, acetaldehyde (CH3CHO), carbon monoxide (CO), ethene (C2H4), water (H2O), methane (CH4), and H2, respectively.To detect the transformation of intermediates in the CLSR process, in situ DRIFTS experiments were performed on a Nicolet iS50 spectrometer (Nicolet iS50, Thermo Scientific, USA) equipped with a Harrick Scientific diffuse reflection accessory and a mercury\u2013cadmium\u2013telluride (MCT) detector cooled by liquid N2. All samples were pretreated at 600\u00a0\u00b0C under an Ar flow for 0.5\u00a0h, followed by purging with Ar for 1\u00a0h, and were then cooled to 400\u00a0\u00b0C to obtain a background spectrum. This collected spectrum was then subtracted from the sample spectrum for each measurement under CLSR conditions.The carbon formation on the OCs was characterized by thermogravimetric analysis (TGA; TGS-2A, Yuanbo, China) and temperature-programmed oxidation (TPO). The TGA experiment was carried out by filling 20\u00a0mg of OC into an alumina crucible. The temperature and weight change were then recorded when the temperature was increased from 50 to 900\u00a0\u00b0C with a heating rate of 10\u00a0\u00b0C\u00b7min\u22121 under air flow (50\u00a0mL\u00b7min\u22121). The TPO profiles of the spent OCs were obtained from the same apparatus, as described for the C2H5OH-pulse experiment. The OC (50\u00a0mg) was pretreated at 300\u00a0\u00b0C for 0.5\u00a0h under flowing Ar (30\u00a0mL\u00b7min\u22121). After the OC was cooled to 50\u00a0\u00b0C, a flow rate of 30\u00a0mL\u00b7min\u22121 of 10 vol% O2/He was used for oxidation, and the temperature was increased linearly from 50 to 900\u00a0\u00b0C. The CO2 in the effluent was monitored and recorded online using a mass spectrometer.CLSR tests were conducted in a stainless-steel tubular fixed-bed reactor with an internal diameter of 20\u00a0mm and a length of 400\u00a0mm. Two grams of OC (20\u201340 mesh) was used for the CLSR reaction. Prior to the test, the OCs were pretreated at 600\u00a0\u00b0C for 1\u00a0h under pure N2 (200\u00a0mL\u00b7min\u22121). After purging with N2, the bed was subsequently adjusted to the designed temperature. An ethanol\u2013water mixture with a flow rate of 0.03\u00a0mL\u00b7min\u22121 and a specific S/C of 1 was fed through a pump (P230, Elite, China) into a heated chamber (200\u00a0\u00b0C), where the mixture was completely evaporated in a stream of N2 (100\u00a0mL\u00b7min\u22121) to start the CLSR reaction for 1\u00a0h. Then, the reactor was heated to the desired oxidation temperature under air flow (200\u00a0mL\u00b7min\u22121) to regenerate the OC for 10\u00a0min. The gaseous products were analyzed online by an Agilent 490 Micro gas chromatograph. The gas chromatograph consisted of two different channels for gaseous product analysis. Channel 1 was equipped with a 10\u00a0m Molecular Sieve 5A column, with Ar used as the carrier gas for the quantification of permanent gases except for CO2 (H2, N2, CO, and CH4). Channel 2 was equipped with a 10\u00a0m PoraPlot Q column, with He used as the carrier gas for the detection of CO2 and C1\u2013C3 hydrocarbons. All the gaseous products were quantified using the micro-machined thermal conductivity detectors (\u03bcTCDs) included in each channel. Liquid products were collected and analyzed over an Agilent 7890A gas chromatograph equipped with a flame ionization detector (FID). Possible liquid products including C2H5OH, CH3CHO, and acetone (CH3COCH3) were quantified over the FID with a Porapak-Q column using N2 as the carrier gas. The selectivities (Si\n) of the carbon-containing products were calculated by the following:\n\n(7)\n\n\n\nS\ni\n\n\n=\n\n\n\n[\ni\n]\n\n\n\n\n\nCO\n2\n\n\n\n+\n\n\n\nCO\n\n\n\n+\n\n\n\n\nCH\n4\n\n\n\n\n\n\n\n\u00d7\n\n100\n%\n\n\n\nwhere i represents the different species in the products, and [i] represents the molar concentration of i in the products.The H2 selectivity (\n\n\nS\n\nH\n2\n\n\n\n) was calculated as follows:\n\n(8)\n\n\n\n\nS\n\n\n\n\nH\n\n\n2\n\n\n\n\n=\n\n\n\n\nF\n\n\n\n\nH\n\n\n2\n\n\n\n\n\n\n\n\nF\n\n\n\n\nC\n\n\n2\n\n\n\n\nH\n\n\n5\n\n\nOH-in\n\n\n\n\n\n\n\nwhere \n\n\nF\n\nH\n2\n\n\n\n represents the molar flow rate of hydrogen in the products, and \n\n\n\nF\n\n\n\n\nC\n\n\n2\n\n\n\n\nH\n\n\n2\n\n\nOH-in\n\n\n\n represents the molar flow rate of ethanol in the reactants.Product distributions (Pi\n) were calculated as follows:\n\n(9)\n\n\n\n\nP\n\n\ni\n\n\n=\n\n\n\ni\n\n\n\n\n\n\nH\n\n\n2\n\n\n\n+\n\n\n\n\nCO\n\n\n2\n\n\n\n\n+\n\n\nCO\n\n\n\n+\n\n\n\n\nCH\n\n\n4\n\n\n\n\n\n\n\u00d7\n100\n%\n\n\n\n\nThe physicochemical properties of the as-prepared Ni\nx\nMg1\u2212\n\nx\nO are shown in Table 1\n. The specific surface area characterized by the BET method is in the range of 15\u201330\u00a0m2\u00b7g\u22121, and the pore volume is in the range of 0.03\u20130.06\u00a0cm3\u00b7g\u22121. The XRD patterns of NiO, MgO, and Ni\nx\nMg1\u2212\n\nx\nO are shown in Fig. 1\n(a). NiO, MgO, and Ni\nx\nMg1\u2212\n\nx\nO possess a rock salt structure. The crystalline sizes of Ni\nx\nMg1\u2212\n\nx\nO, as calculated by the Scherrer equation, are similar. To show the influence of the content of Ni on the lattice parameter, the XRD patterns in the range of 40\u00b0\u201346\u00b0 are provided in Fig. 1(b). The (200) peak of Ni\nx\nMg1\u2212\n\nx\nO shifts from 42.8\u00b0 to 43.2\u00b0 with increasing Ni content (i.e., from MgO to NiO). The lattice parameter of Ni\nx\nMg1\u2212\n\nx\nO can be calculated from the peak position based on Bragg\u2019s law (Table 1). When the lattice parameter of Ni\nx\nMg1\u2212\n\nx\nO is correlated with the content of Ni in Ni\nx\nMg1\u2212\n\nx\nO, a linear relationship can be verified (Fig. 1(c)), which indicates the formation of Ni\nx\nMg1\u2212\n\nx\nO solid solutions in the corresponding Ni/Mg proportions [22].Ni0.4Mg0.6O is selected as an example to observe the morphology of the solid solution. TEM images of Ni0.4Mg0.6O are given in Figs. 1(d)\u2013(f). According to Figs. 1(d) and (e), the particle size of Ni0.4Mg0.6O is in the range of 10\u201320\u00a0nm. No segregated NiO crystals are observed. The (200) plane of Ni\nx\nMg1\u2212\n\nx\nO with a lattice spacing of 4.215\u00a0\u00c5 can also be measured in Fig. 1(f), which is in accordance with the results of the XRD characterizations. EDS mapping was applied to probe the elemental dispersion. According to Figs. 1(g)\u2013(i), the distribution of Ni, Mg, and O in Ni\nx\nMg1\u2212\n\nx\nO is homogeneous, indicating the formation of a substitutional solid solution of Ni\u2013Mg oxide.To achieve efficient hydrogen production, 400\u00a0\u00b0C was chosen as the temperature for the CLSR reaction (Fig. S1 in Appendix A). The S/C was set to 1. The selectivities of the carbon-containing products and H2 are given in Fig. 2\n(a). As the content of Ni increases, more CH4 is generated, which is detrimental to H2 selectivity. This phenomenon can be attributed to the poor dispersion of Ni (Table 1). CO selectivity over Ni0.2Mg0.8O is the highest among various Ni\nx\nMg1\u2212\n\nx\nO solid solutions. The generation of CO hinders the purity of H2. In this study, Ni0.4Mg0.6O presents the maximum H2 selectivity of 4.72\u00a0mol H2 per mole ethanol.We further studied the properties of Ni0.4Mg0.6O. The results from the time-on-stream test of ethanol CLSR over Ni0.4Mg0.6O at 400\u00a0\u00b0C in a single cycle are given in Fig. 2(b). The CLSR of ethanol can be generally divided into three stages based on the changes in the distribution of products. In stage I (from the start of the reaction to 6\u00a0min), CO2 is the main product. Ethanol is completely oxidized by the surface oxygen of Ni0.4Mg0.6O. In stage II (from 6 to 33\u00a0min), as more Ni2+ ions are gradually reduced to metallic Ni, the decomposition of ethanol occurs over the surface of Ni to produce H2 and CH4. The selectivity of the gaseous products is then steadily maintained. The selectivity of H2 reaches its maximum and the CO concentration is suppressed to 1% in stage II. In stage III (after 33\u00a0min) the conversion of ethanol and the selectivities of H2 and CO2 decrease with the generation of more CO and CH4. The deactivation of Ni0.4Mg0.6O occurs in this stage.A cyclic stability test was carried out on Ni0.4Mg0.6O. After the CLSR reaction at 400\u00a0\u00b0C, the reduced Ni0.4Mg0.6O was re-oxidized and the carbon was combusted in the air at 600\u00a0\u00b0C for 10\u00a0min. This process is referred to as the \u201cregeneration step\u201d in our study. One cyclic test constituted 60\u00a0min of the CLSR of ethanol and 10\u00a0min of regeneration. The performance of Ni0.4Mg0.6O in the cyclic test is shown in Fig. 2(c). The selectivity of H2 over Ni0.4Mg0.6O only drops by about 3% in 30 cycles, indicating that the regeneration can recover the Ni0.4Mg0.6O. The structure of Ni\nx\nMg1\u2212\n\nx\nO after 30 cycles was characterized by TEM and XRD (Table 1, Fig. 2(d), and Appendix A Fig. S2). The morphology and crystal structure of Ni0.4Mg0.6O remained the same after the long-term test. The solid-solution OC exists in the form of particles, without the occurrence of sintering. The crystalline size of Ni0.4Mg0.6O after the stability test was 14.2\u00a0nm, which is similar to that of fresh Ni0.4Mg0.6O. These results verify the recovery of Ni0.4Mg0.6O in the regeneration step and demonstrate the superior stability of this solid solution in the CLSR of ethanol.A pulse experiment with an ethanol\u2013water mixture (S/C\u00a0=\u00a01) over Ni0.4Mg0.6O at 400\u00a0\u00b0C was conducted in order to explore the oxygen-release process of Ni\nx\nMg1\u2212\n\nx\nO (Fig. 3\n(a)). During the first five pulses, the peaks of H2 were not obvious and CO2 was the main product. This phenomenon indicates that the redox reaction between Ni0.4Mg0.6O and ethanol is dominant in this period, which corresponds to the stage I observed in the time-on-stream test of Ni0.4Mg0.6O (Fig. 2(b)). Afterward, the H2 peaks were enlarged and remained stable. CO2 became the dominant carbonaceous product, which represents the characteristics of stage II.XRD was applied to detect the change in the composition of Ni0.4Mg0.6O during the pulse experiment. Since there may be a diffraction peak of metallic Ni at 44\u00b0 near the peak, corresponding to the (200) plane of Ni\nx\nMg1\u2212\n\nx\nO, the second strongest peak for the (220) plane of Ni\nx\nMg1\u2212\n\nx\nO was analyzed. The XRD patterns in the range of 60\u00b0\u201364\u00b0 for Ni0.4Mg0.6O after different pulses of the ethanol\u2013water mixture are given in Fig. 3(b). The lattice parameter of the reduced Ni0.4Mg0.6O was calculated according to the peak position. If the distribution of Ni2+ and Mg2+ in Ni\nx\nMg1\u2212\n\nx\nO is homogeneous, then the Ni content, x, of such a solid solution can be calculated according to Vegard\u2019s law [28]:\n\n(10)\n\n\n\n\na\n\n\n\n\nNi\n\n\nx\n\n\n\n\nMg\n\n\n1\n-\nx\n\n\nO\n\n\n\n=\n\nx\n\n\na\n\n\nNiO\n\n\n\n+\n\n\n(\n1\n\n-\n\nx\n)\n\n\n\na\n\n\nMgO\n\n\n\n\n\nwhere \n\n\na\n\n\nNi\nx\n\n\nMg\n\n1\n-\nx\n\n\nO\n\n\n\n is the lattice constant of Ni\nx\nMg1\u2212\n\nx\nO, and the lattice constants of NiO (a\nNiO) and MgO (a\nMgO) were obtained from pure oxides (powder diffraction file (PDF) No. 47\u20131049 for NiO and PDF No. 45\u20130946 for MgO). Based on the calculated lattice constants of the reduced Ni0.4Mg0.6O, we obtained the Ni contents and degree of reduction of Ni0.4Mg0.6O (Figs. 3(c) and (d)).The change in the degree of reduction of Ni0.4Mg0.6O is in accordance with the findings from the pulse experiment. In stage I, the degree of reduction of Ni0.4Mg0.6O increases rapidly. The complete oxidation of ethanol is dominant, with the generation of CO2. In stage II, H2 is formed consistently in the last three pulses. Simultaneously, oxygen release continues according to the change in the degree of reduction of Ni0.4Mg0.6O. In comparison with stage I, the rate of oxygen release in stage II drops, indicating that the oxygen from Ni0.4Mg0.6O participates in the reaction between ethanol and water to produce H2. H2 selectivity is increased due to the occurrence of water gas shift. The stoichiometric S/C in ethanol steam reforming (Eq. (1)) is 1.5, which is larger than the S/C in the CLSR and pulse experiment. Therefore, additional oxygen is necessary for stable hydrogen production in stage II. Stage II in the CLSR is carried out as follows:\n\n(11)\nC2H5OH\u00a0+\u00a02H2O\u00a0+\u00a0[O]\u00a0\u2192\u00a02CO2\u00a0+\u00a05H2\n\n\nwhere [O] represents the oxygen from Ni\nx\nMg1\u2212\n\nx\nO.When the active oxygen from Ni0.4Mg0.6O is depleted, the low S/C provides insufficient oxidation capacity for the steam reforming, resulting in decreased selectivity toward H2 and CO2 (stage III in the CLSR test). Meanwhile, ethanol is decomposed to carbon, which covers the surface of the OC and leads to deactivation. TGA and O2-TPO experiments were conducted to verify this process (Fig. S3 in Appendix A). The mass increase at the beginning of the TGA analysis of the reacted Ni0.4Mg0.6O after one cycle can be attributed to the oxidation of Ni (Fig. S3(a)). The subsequent mass loss is in accordance with the peak position of CO2 in the O2-TPO, which corresponds to the gasification of the deposited carbon (Fig. S3(b)). The carbon deposition is considered to be the cause of deactivation in stage III. The results also show that the coke generated in the CLSR of ethanol can be eliminated at 600\u00a0\u00b0C in the regeneration step.To further investigate the modulation effects of Mg2+ on Ni\nx\nMg1\u2212\n\nx\nO, H2-TPR experiments were performed to detect the reactivity of different oxygen species in the solid solution (Fig. 4\n(a)). No reduction peak was observed over pure bulk MgO up to 900\u00a0\u00b0C. The reduction peak of NiO is very broad at 200\u2013400\u00a0\u00b0C. The H2-TPR profiles of Ni\nx\nMg1\u2212\n\nx\nO mainly consist of a low-temperature reduction peak at around 300\u00a0\u00b0C and a large reduction peak at 400\u2013800\u00a0\u00b0C, indicating the existence of two types of oxygen species with different reactivities. Based on the reduction profile of NiO and the structure of Ni\nx\nMg1\u2212\n\nx\nO, the low-temperature reduction peak of Ni\nx\nMg1\u2212\n\nx\nO can be attributed to the release of surface oxygen. The large peak in the high-temperature range corresponds to the reduction of Ni2+ in the bulk of Ni\nx\nMg1\u2212\n\nx\nO [29]. The temperatures of the different reduction peaks of Ni\nx\nMg1\u2212\n\nx\nO are summarized in Fig. 4(b). The reactivity of the surface oxygen is enhanced with the increase of Ni concentration in Ni\nx\nMg1\u2212\n\nx\nO. The reducibility of the metal oxide is related to the band gap between the valence and conduction bands [30]. Closer valence and conduction bands make metal oxides more easily reduced [30]. Previous research indicates that, when the Ni content\u00a0x\u00a0is greater than 0.074, the band gap of Ni\nx\nMg1\u2212\n\nx\nO decreases linearly with x\n[31]. Therefore, the oxygen-release process of Ni\nx\nMg1\u2212\n\nx\nO is inhibited with increased Mg2+ content, which aligns with the results from H2-TPR. Moreover, the coefficient of the Ni2+\u2013Mg2+ interdiffusion increases exponentially with the concentration of Ni2+ in the air [32]. In conclusion, Ni2+ diffusion in Ni\nx\nMg1\u2212\n\nx\nO is suppressed by the lattice confinement of Mg2+. Therefore, the reactivity of bulk oxygen decreases with the enrichment of Mg2+ in Ni\nx\nMg1\u2212\n\nx\nO, which can be reflected by the increased reduction temperature of bulk oxygen. For the Ni0.2Mg0.8O sample, the reduction temperature for oxygen in the bulk is slightly lower than that of Ni0.4Mg0.6O. MgO formed in the surface layer prevents the deeper reduction of bulk Ni0.2Mg0.8O, lowering the apparent reduction temperature of bulk oxygen and resulting in a lower degree of reduction (Table 1).To investigate the reaction pathway over Ni\nx\nMg1\u2212\n\nx\nO, in situ DRIFTS experiments were carried out (Fig. S4 in Appendix A). The spectra collected at different times during the reaction were divided into three distinct stages. To observe the changes of the C-containing surface species over Ni0.4Mg0.6O, the in situ DRIFTS spectra in the range from 2400 to 800\u00a0cm\u22121 were obtained, and are presented in Fig. 5\n.At the beginning of the reaction, the infrared (IR) peaks of gaseous CO2 at 2350\u00a0cm\u22121 and CO3\n2\u2212 at 1510 and 1240\u00a0cm\u22121 were observed [33]. The generation of CO2 and CO3\n2\u2212 can be attributed to the complete oxidation of ethanol by surface oxygen, corresponding to stage I observed in the time-on-stream test. As the reaction proceeds, CO is generated, according to the appearance of the peak at 2170\u00a0cm\u22121. The C\u2013O bond in CH3CH2O* at 1030\u00a0cm\u22121 can be seen [34]. The peaks at 1740 and 1580\u00a0cm\u22121 are assigned to the C=O bond in CH3COO*, which is a characteristic intermediate over Ni-based catalysts in ethanol steam reforming, corresponding to stage II [35]. The IR peaks indicate that the decomposition of ethanol into CH3CH2O* occurs over metallic Ni, and the CH3CH2O* is further oxidized to CH3COO*. According to the evidence of the changes in the degree of reduction, water may work collaboratively with the bulk oxygen of Ni\nx\nMg1\u2212\n\nx\nO to oxidize ethanol in stage II. In stage III, the CO3\n2\u2212 peak disappears gradually, and the intensity of the acetate peak increases. Moreover, the peak at 880\u00a0cm\u22121 for the C\u2013H bond in gaseous CH4 also appears in this stage. The multiple peaks in the range from 1600 to 1400\u00a0cm\u22121 correspond to the C\u2013H vibration of the deposited carbon [36]. The changes in the intermediates indicate the occurrence of the decomposition of ethanol to generate CH4 and carbon in this stage. Due to the low S/C, the oxidation capacity of water is insufficient to convert the surface C-containing species to CO2. The proposed surface reaction pathway of ethanol for the CLSR of ethanol over Ni\nx\nMg1\u2212\n\nx\nO is in line with the structural evolution of solid solution (Fig. 6\n).Ni\nx\nMg1\u2212\n\nx\nO solid solution was applied as a novel OC in the CLSR of ethanol for hydrogen production. The oxygen release of Ni\nx\nMg1\u2212\n\nx\nO is regulated with the lattice confinement by Mg2+. As a result, the optimum OC, Ni0.4Mg0.6O, was found to exhibit a robust performance toward hydrogen production (4.72\u00a0mol of H2 per mole of ethanol), with an S/C of 1. A three-stage reaction mechanism of the CLSR process was proposed. In stage I, ethanol is completely oxidized by the surface oxygen of Ni\nx\nMg1\u2212\n\nx\nO. After the depletion of the surface oxygen and the formation of surface Ni sites, ethanol is oxidized by H2O and the bulk oxygen from Ni\nx\nMg1\u2212\n\nx\nO collaboratively, achieving the maximum efficiency for hydrogen production in stage II. Without the participation of oxygen species, ethanol steam reforming becomes the dominant process in stage III. The CLSR of ethanol using Ni\nx\nMg1\u2212\n\nx\nO as the OC could potentially reduce the S/C in comparison with conventional steam reforming and achieve renewable hydrogen production from biomass with a minimum external heat supply. This research provides a feasible strategy for the design of a novel OC in diverse chemical looping processes with improved performance and structural stability.This work was supported by National Natural Science Foundation of China (U20B6002, 51761145012, and 21525626) and the Program of Introducing Talents of Discipline to Universities (BP0618007) for financial support.Hao Tian, Chunlei Pei, Sai Chen, Yang Wu, Zhjian Zhao, and Jinlong Gong declare that they have no conflict of interest or financial conflicts to disclose.Supplementary data to this article can be found online at https://doi.org/10.1016/j.eng.2020.08.029.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  The chemical looping steam reforming (CLSR) of bioethanol is an energy-efficient and carbon-neutral approach of hydrogen production. This paper describes the use of a Ni\n                        x\n                     Mg1\u2212\n                     \n                        x\n                     O solid solution as the oxygen carrier (OC) in the CLSR of bioethanol. Due to the regulation effect of Mg2+ in Ni\n                        x\n                     Mg1\u2212\n                     \n                        x\n                     O, a three-stage reaction mechanism of the CLSR process is proposed. The surface oxygen of Ni\n                        x\n                     Mg1\u2212\n                     \n                        x\n                     O initially causes complete oxidation of the ethanol. Subsequently, H2O and bulk oxygen confined by Mg2+ react with ethanol to form CH3COO* followed by H2 over partially reduced Ni\n                        x\n                     Mg1\u2212\n                     \n                        x\n                     O. Once the bulk oxygen is consumed, the ethanol steam reforming process is promoted by the metallic nickel in the stage III. As a result, Ni0.4Mg0.6O exhibits a high H2 selectivity (4.72\u00a0mol H2 per mole ethanol) with a low steam-to-carbon molar ratio of 1, and remains stable over 30 CLSR cycles. The design of this solid-solution OC provides a versatile strategy for manipulating the chemical looping process.\n               "}
{"full_text": "Experimental data relating to the figures and tables presented in this manuscript have been deposited at Zenodo under the https://doi.org/10.5281/zenodo.7199360 and are publicly available as of the date of publication. Supporting DFT datasets have been deposited at ioChem-BD\n45\n under the https://doi.org/10.19061/iochem-bd-1-258 and are publicly available as of the date of publication. This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon reasonable request.The carbon dioxide (CO2) electroreduction (eCO2R) toward simple products like carbon monoxide\n1\n and formate\n2\n has attracted wider research efforts and more recently the interest of chemical industries. However, more reduced products cannot be generated yet at promising yields under relevant operating conditions, with the single exception of ethylene.\n3\n\n,\n\n4\n Besides the initial proof of concept for recently discovered inorganic Ni oxygenates,\n5\n the only catalysts sustaining more\u00a0than two proton-electron transfers beyond trace rates\n6\n are copper-based (Cu-based) materials. However, the tendency of Cu-based electrocatalysts to promote numerous products has driven an intense quest for strategies enabling selectivity control. For this purpose, nanostructure and compositional modifications have been suggested. Early observations\n7\n\n,\n\n8\n revealed that halogen-modified catalytic systems display leveraged selectivity toward ethylene, which triggered follow-up efforts to explore their potential.\n9\n Two strategies based on materials (copper halide catalysts) and electrolytes (halogen-containing) have been practiced so far.Regarding halogen-containing electrolytes, the positive effect of adding chlorinated, brominated, or iodinated salts like KX (X\u00a0= Cl, Br, I) to the aqueous medium has been often reported. Most works claim selectivity increases toward C2+ products,\n10\n\n,\n\n11\n\n,\n\n12\n though a few claimed exclusively larger activity toward carbon monoxide.\n13\n\n,\n\n14\n In spite of some rationalization efforts, the influence of the halide anion coming from the electrolyte is not well understood, as it is hard to deconvolute from that of its compensating cation, which affects the medium conductivity and double layer structure as well as the CO2 activation itself.\n9\n\n,\n\n15\n Alternatively, copper halide catalysts (CuX, X\u00a0= Cl, Br, I, with special focus on CuCl) represent the most pursued route through a variety of synthetic strategies, with wet bench methods\n16\n\n,\n\n17\n and electrodeposition\n8\n\n,\n\n18\n\n,\n\n19\n\n,\n\n20\n being the two most commonly used. Fluorine-containing halides have been less extensively studied.\n21\n\n,\n\n22\n Overall, reported C2+ Faradaic efficiencies outcompete halogen-free copper systems, reaching up to 80% when operated under optimized conditions,\n20\n\n,\n\n21\n though noticeably different performances have been reported over seemingly similar systems.\n10\n\n,\n\n14\n It\u00a0is commonly accepted that copper halides act as catalyst precursors. A common\u00a0preparation route is exposure of copper halides to reductive potentials to obtain metallic copper matrices with enhanced performance.\n8\n\n,\n\n20\n\n,\n\n23\n\n,\n\n24\n\n,\n\n25\n Alternatively, an intermediate step transforming copper halides into copper oxides has been introduced to increase the population of defects present upon reduction.\n17\n\n,\n\n19\n Similarly to the case of modified electrolytes, the role of halogens in the lattice is still unknown, despite pioneering efforts correlating copper halide structures with catalytic performance.\n26\n Halogen atoms remaining in the copper structure after reconstruction have been proposed to enhance the stability of oxidic Cu centers,\n21\n\n,\n\n22\n while other simulations predict that the adsorption of leached halide ions may alter surface electronic properties\n20\n or hinder the parasitic\u00a0hydrogen evolution reaction.\n9\n On the other hand, most reports consider halogens as structure-directing agents toward specific copper facets\n17\n\n,\n\n23\n or nanostructures.\n18\n\nIn this context, a key pending task is determining synthesis-structure-performance relations linked to the halogenation degree especially for the most widely explored chlorine-based systems. Chlorine exhibits advantageous features compared with bromine, the second most commonly studied halogen to this end. Its more sustainable character and frequent use in industry makes it a preferential target for practical applications. Closely related, bromination of copper phases is much faster than chlorination, making the controlled incorporation of the latter extremely challenging.\n27\n However, finding a synthetic method enabling controlled chlorination of copper has still remained challenging due to the strong affinity between the two elements, a well-known cause of catalyst deactivation in the Deacon process.\n28\n As a result, the inability to modulate chlorine content has limited the study of this family of materials to copper chloride, precluding effective catalyst design. This work develops a gas-phase treatment based on exposure to HCl at different temperatures enabling fine control of chlorine incorporation. Partially chlorinated copper systems (Cu, Cu2O, CuO) were synthesized to reveal a maximal promotion toward highly reduced products (HRPs, i.e., those requiring more than 2e\u2212 transfer and comprising CH4, C2, and C3 compounds) for a Cu2O-based catalyst containing roughly equal Cu2O and CuCl, underscoring the relevance of selecting the copper phase and chlorination degree. A direct correlation between Faradaic efficiency for HRPs and post-reaction surface chlorine content emerged. Simulations show that the origin of the stability and enhanced performance relies on a large material reconstruction, generating partially chlorinated metallic surfaces and copper oxychloride ensembles showing reduced barriers toward HRPs. These insights thus provide tools\u00a0and guidelines for the design of enhanced chlorine-promoted copper catalysts.The first step toward copper catalysts with controlled degree of chlorination was to develop two oxidation treatment protocols (Figure\u00a01\nA, conditions in Table\u00a0S1) to obtain bulk Cu2O or CuO starting from Cu foils. The reason for targeting the two oxidic phases is their different reactivities toward HCl due to the mechanism of lattice O-HCl interaction in the gas phase.\n29\n Exposure of pretreated Cu foils to the harsher oxidation conditions of 50 vol\u00a0% O2/He at 673 K for 30\u00a0min was sufficient to predominantly form surface CuO domains as indicated by the presence of (002) and (111) reflections at 35.6\u00b0 and 38.8\u00b0 2\u03b8 respectively in the X-ray diffraction (XRD) profiles (Figure\u00a0S1), in addition to small traces of partially oxidized Cu2O(111) at 36.7\u00b0 2\u03b8. Milder exposure to 20 vol\u00a0% O2/He at 553 K for 5\u00a0min formed primarily surface Cu2O as well as traces of over-oxidized CuO domains, as indicated by the Cu2O(111) reflection at 35.6\u00b0 2\u03b8 shifting toward lower angles compared with the reference due to lattice strain (Figure\u00a0S1).Preliminary tests suggested temperature as the variable enabling better control of the chlorination degree. Chlorination treatment of clean Cu and CuO\nx\n in 2 vol\u00a0% HCl/He for 30\u00a0min at various temperatures resulted in systems with variable copper chlorides content, most notably CuCl as indicated by the (111) reflection at 28.6\u00b0 2\u03b8 in the XRD profiles of chlorinated samples (Figure\u00a01B). So-prepared samples are coded according to the initial copper phase and chlorination temperature CuO\nx\n-HCl(T). While the relative proportion of bulk chloride to oxide phases steadily increases with T for Cu2O-HCl(T) (Figure\u00a01C), two distinct regimes are observed for CuO-HCl(T), separated by the onset temperature of approximately 450 K, above which copper oxychloride phases form (Figures\u00a0S2 and S3). Full chlorination of both oxides occurs above 523 K. The overall low bulk CuCl content in the Cu-HCl(T) family does not display a strong correlation with chlorination temperature, corroborating the less marked reactivity of the metallic phase. The relative proportion of surface Cl quantified using X-ray photoelectron spectroscopy (XPS) Cl 2p signals (Table\u00a0S2 and Figure\u00a0S4) followed the general trend in the bulk chlorination degree of the freshly chlorinated CuO-HCl(T) and Cu2O-HCl(T) systems. Similarly, assignment of the most intense Cu LMM Auger signals for Cu2O-HCl(T) identified a relative decrease and increase of Cu2O and CuCl, respectively, with increasing chlorination temperature (vide infra).Once the set of copper catalysts with controlled chlorination degrees was available, their catalytic performance was evaluated in a two-compartment cell containing CO2-saturated 0.1\u00a0M KHCO3 at \u22120.8\u00a0V vs. RHE, a mild potential at which complex products are not favored on the reference metallic Cu surface.\n30\n Extended details can be found in the supplemental experimental procedures.\nFigure\u00a02\n displays product distributions ordered by chlorination temperature for the three catalyst families (see Tables\u00a0S3\u2013S5 for numerical values and Figure\u00a0S5 for a typical chronoamperometry profile). Figure\u00a02A reveals that the reference family Cu-HCl(T), showing a low and uniform chlorination degree (Figure\u00a01C), displayed a very mild promotional effect. Formate production was slightly enhanced at most temperatures with no discernible pattern. Modest ethylene formation was observed around Cu-HCl(323), aligned with reports claiming enhanced ethylene production on copper halide surfaces.\n9\n The set of CuO-HCl(T) catalysts yielded Faradaic efficiencies toward HRPs seemly independent from the chlorination temperature and similar to that of unmodified CuO, as can be seen in Figure\u00a02B. Faradaic efficiency toward simple compounds, and especially to formate, was favored at intermediate chlorination temperatures. Contrary to the case of metallic copper, the CuCl content varied from 0 to 100% in this set of samples (Figure\u00a01C), suggesting that the initial chlorination degree of CuO-based systems does not influence the ability of the catalyst to produce HRPs.A different picture emerged for the case of Cu2O-HCl(T) catalysts. Pristine Cu2O exhibited predominant formation of carbon monoxide and formate with only trace HRPs. Formate production was sharply favored for the Cu2O-HCl(323) sample. As chlorination temperature rises, HRPs become increasingly predominant up to Cu2O-HCl(373), for which HRP Faradaic efficiency peaks (ca. 14%, Figure\u00a02A). Of note, production of all compounds gathered under the HRP acronym (methane, C2 and C3) increase with chlorination temperature, though the surge of methane must be highlighted. At higher chlorination temperatures the promotional effect is still observable, though to a lesser extent, and it becomes approximately temperature independent. According to the quasi-linear CuCl content-temperature dependence registered for Cu2O (Figure\u00a01C), this result hinted to an optimized CuCl content of ca. 40% in the system prior to testing. The survey of other potentials suggested that the promotional effect was optimal at potentials around \u22120.8\u00a0V vs. RHE (Table\u00a0S6), arguably due to the insufficient overpotential available at less cathodic potentials, as well as the likely instability of adsorbed Cl at highly cathodic potentials. Additional experiments upon bromination at room temperature of Cu2O and CuO shown in Figure\u00a0S6 disclosed a milder promotional effect toward HRPs than their chlorinated counterparts showing similar degree of halogenation, reinforcing the interest of chlorine promotion.These distinctive behaviors already disclosed the relevance of the copper source, and that complete chlorination may not be associated with optimal promotional effect in copper materials. The next sections are devoted to elucidating the effect of chlorine on the materials\u2019 physico-chemical properties and developing mechanistic insights supporting observed patterns.The first step was to investigate the influence on the promotional effect of surface morphologies introduced by the different synthesis conditions of the catalysts. Differences in surface roughness, which may impact the local chemical environment and thus selectivity,\n3\n\n,\n\n31\n were visualized by scanning electron microscopy (SEM, Figure\u00a03\nA). Micrographs of CuO-HCl(T) and Cu2O-HCl(T) catalyst surfaces chlorinated at low, moderate, and high temperatures did not show notably different microscale features. In parallel, changes in electrochemically active surface area quantified from double-layer capacitance measurements did not show any discernible correlation with FE\nHRP (Figure\u00a03B). Since surface morphology may not be a decisive factor explaining the promotional effect, the distinctive trends observed between the three catalyst families may exhibit a chemical ground. Predictably, the density of defects, such as undercoordinated atoms or grain boundaries, or the relative population of facets could also have a relevant role to determine differences in catalytic performance among these systems. Due to the highly dynamic nature of copper surface under operation conditions,\n25\n its operando monitoring would be required, which still represents an experimental challenge. For the case of porous materials like catalytic layers in gas diffusion electrodes, a parallel study considering porosity should be considered.The concentration and nature of chlorinated species at the reaction interface, which evolved from the fresh structure upon exposure to eCO2R reaction conditions, were primarily analyzed and quantified using XPS and Auger signals of relevant Cl and Cu regions of used catalyst samples. For all CuO-HCl(T) and Cu2O-HCl(T) catalysts, small but quantifiable amounts of Cl reaching up to ca. 0.7 atom % and 1.8 atom %, respectively, remained after reaction (Figures\u00a04A and S7). For Cu2O-HCl(T), a narrow range of temperature with increasing values around the maximum at 373 K for both FE\nHRP (Figure\u00a02) and surface Cl content (Figure\u00a04B) is evident, suggesting that the choice of treatment temperatures crucially affects both inter-linked phenomena. These observations are combined in Figure\u00a04C, where an apparent linear correlation between FE\nHRP and surface Cl content is shown for the Cu2O-HCl(T) family.Note that the dependences of the Cl content with chlorination temperature in fresh (Figure\u00a01C and Table\u00a0S2) and used materials (Figure\u00a04B) are notably different. Indeed, mild initial chlorination degrees are associated with the largest capability of retaining Cl during reaction. Cu LMM Auger spectra of fresh materials (Figure\u00a05\nA) show the expected increase of the CuCl signal with temperature. Fresh Cu2O-HCl(373) exhibits comparable intensities for both Cu2O and CuCl, in line with XRD analysis (Figure\u00a01C). After eCO2R, metallic Cu is the predominant species with small shoulders assigned to CuCl appearing upon increased sputtering time of the samples. Only Cu2O-HCl(373) features stronger signals pertaining to CuCl and Cu2O, suggesting that the presence of both phases leads to a distinct (sub)surface state upon exposure to reaction conditions. However, the detection of Cu2O formed during the exposure of the sample to air cannot be discarded at this point. The surface analysis was complemented by XRD observations (Figure\u00a05B). Since no reflections could be assigned to CuCl in Cu2O-HCl(T) after reaction, it is reasonable to expect that Cl is present as part of smaller surface and subsurface domains without crystalline order or strongly adsorbed on the surface. We thus postulated that in situ reduction of Cu oxide and chloride results in the stabilization of Cl around the reduced, defective oxidic surface ensembles formed during eCO2R,\n3\n which we denote in general as \u201ccopper oxychloride-like ensembles.\u201dThe parallel analysis for CuO-HCl(T) and Cu-HCl(T) catalysts showed different results. They exhibited a very mild dependence of surface Cl content after reaction with chlorination temperature (Figure\u00a04B), with FE\nHRP largely unchanged compared with their untreated counterparts (Figure\u00a02), leading to no performance-Cl content correlation (Figure\u00a04C). Despite the clear formation of copper chloride and oxychloride phases in the bulk of the used samples that left detectable crystallites as measured by XRD for CuO-HCl(423) and CuO-HCl(473) (Figure\u00a05B), the surface Cl content kept fairly constant around 1 atom % (Figure\u00a04B). This suggests that most of the Cl could be locked in bulk chloride-containing phases outside the XPS measurement depth. The nature of such species and that of surface \u201ccopper oxychloride-like ensembles\u201d would conceivably be different for the CuO-HCl(T) family, as reflected by their selectivity patterns.Overall, these results suggest structural differences among families upon exposure to reaction conditions (Figure\u00a05C). For Cu-HCl(T), the low Cl content initially incorporated may lead to very sparse Cl coverage after reaction with no significant effect on HRP formation. For CuO-HCl(T), stabilization of inactive subsurface (oxy)chlorides may be responsible for the lack of promotion effect toward HRPs. Finally, for Cu2O-HCl(T), moderate initial chlorination degrees and subsequent exposure to eCO2R conditions favor higher Cl contents that may be stabilized at the defective, oxygen-containing copper matrix with distinctive electronic features for HRP formation. The combination of the optimum Cu phase and the accuracy of the chlorination procedure are thus crucial.Density functional theory (DFT) simulations contributed to unravel the structural and compositional features affecting HRP formation. The consensus mechanisms in eCO2R starts with CO2 adsorption promoted via electron transfer, followed by protonation to \u2217COOH and a proton-coupled electron transfer to produce water and \u2217CO.\n3\n From there, CO can either desorb or diverge toward C1 and C2 HRP formation. Thus, product distribution toward CO, methane, and C2+ products is controlled by the energy of \u2217CO desorption, \u2217CO protonation to \u2217COH, and \u2217CO-CO coupling,\n32\n\n,\n\n33\n\n,\n\n34\n which are taken as descriptors for selectivity.The nature of the active sites in native copper catalysts under eCO2R is controversial due to the deep structural changes depending on the reaction conditions and the initial oxidation state of the material.\n35\n\n,\n\n36\n\n,\n\n37\n Recent studies\n25\n\n,\n\n38\n suggest that the existence of metallic Cu0, oxidic Cu+, and polarized Cu\n\u03b4+ might be responsible for the C-C coupling ability. Similarly to oxygen, chlorine can also act as modifier in similar terms due to their similar electronegativity. Therefore, the ability to present variable oxidation states in the chlorinated materials may be a favorable feature for enhancing reactivity.Thermodynamic properties of relevant intermediates were analyzed on different Cl-modified copper surfaces. First, the stability and electronic properties of metallic copper surfaces showing increasing chlorine coverages was explored considering the desorption of chlorine atoms on the lowest surface energy facets Cu(111), Cu(100), and Cu(110),\n39\n as well as on Cu(211) step models. Although electrolyte-metal interface studies\n40\n have shown the key role of halide adlayers on Cu surfaces, they might be modestly stable under eCO2R conditions.\n41\n However, these structures constitute the simplest model to understand the electronic effect exerted by Cl on its nearest Cu atoms. Given the maximum XPS-derived Cl coverage of ca. 0.20 ML after reaction (Figure\u00a04),\n42\n simulations covered the 0.00\u20130.33 ML range at \u22120.80\u00a0V vs. RHE with implicit solvation. The results on Cu(100) show that chlorine desorption energies are independent of coverage within this range (Table\u00a0S7). Chlorine adatoms on Cu(100) are marginally stable (Tables\u00a0S8 and S9) since desorption as aqueous Cl\u2212 is close to thermoneutrality (\u0394G\ndes\u00a0= \u22120.23 eV, U\u00a0= \u22120.80\u00a0V vs. RHE), similar to the case of Cu(110) and Cu(211) (\u0394G\ndes\u00a0= \u22120.21 eV and \u22120.13 eV, respectively), while Cl\u2212 on Cu(111) is even less stable (\u0394G\ndes\u00a0= \u22120.54 eV). Desorption energies of Cl from Cu(100) and ab initio thermodynamics-derived adsorption energies per unit area (Figure\u00a0S8) both indicate reduced stability at potentials more negative than U\u00a0= \u22120.57\u00a0V vs. RHE. As \u2217Cl-\u2217Cl lateral interactions are negligible, effects of chlorine incorporation are localized, modifying the electronic density of the nearest neighboring Cu atoms as shown by the Bader charge of 0.10\u20130.15 |e\u2212| (Figure\u00a06\nA and Table\u00a0S10) and supported by d-band analyses (Figure\u00a0S9 and Table\u00a0S11). Therefore, surface chlorination may generate polarized Cu\u03b4+ species, known to have an important role in eCO2R.\n43\n\n,\n\n44\n As charging creates asymmetric Cu-Cu\u03b4+ pairs and \u2217CO adsorption energy on Cu sites close to Cl atoms is smaller than on regular Cu sites (Figure\u00a06B), the \u2217CO-CO coupling energy is lowered compared with clean Cu without significantly altering other steps, enhancing C2 HRP formation (Figure\u00a06C). For Cl coverages above 0.17 ML, the decreasing frequency of these pairs diminishes this effect, promoting CO desorption instead (Figure\u00a06B). \u2217H adsorption strength was however found to be unaffected with respect to Cl coverage (Figure\u00a0S10, E\n\u2217H,ads\u00a0= \u22120.90 eV), suggesting that the competing hydrogen evolution reaction remains unaltered. Energy profiles leading to different HRPs (Figures\u00a0S11 and S12) demonstrate that the applied potential is crucial in achieving C-C bond formation. However, this modification does not lead to enhanced formation of C1 HRPs (namely methane) (Figures\u00a0S13\u2013S16).To discover stable structures able to generate asymmetric Cu-Cu\n\u03b4+ pairs and promote C1 HRP formation, a heuristic approach involving model systems mirroring the structural complexity resulting from the chlorination treatment protocol was devised. Noting the optimal surface Cl content for Cu2O-HCl(373) where both oxidic and chloride phases coexisted prior to eCO2R (Figure\u00a05), a total of 68 structural models representing a wide compositional range were built upon Cu2O(111) as a reference, with certain oxygen atoms removed and (H)Cl incorporated, to assess stable copper (hydr)oxychloride phases (see supplemental experimental procedures and Figure\u00a0S17 for full details).\n25\n All structures were optimized via DFT, and their computed energies were compared with a predicted energy dependent on only the stoichiometry, using a general equation with the form of Equation\u00a0S21. Multivariate linear regression and subsequent refinement of the variable selection resulted in a final regression model iteration that uses the numbers of Cl, H, and O atoms, as well as \u2217Cl adatoms, as independent variables (see supplemental experimental procedures and Table\u00a0S12). The independence of variables precluding interaction-dependent terms implies that structures with DFT energies (E\nDFT) lower than these predicted by the regression model (E\npred) show synergetic effects that render them more stable (Figure\u00a0S18A). The cooperative effects (up to \u22121.76 eV) appear to be associated with structural motifs that locally resemble Cu2OCl2-like (Figure\u00a0S18B) or CuOHCl-like (Figure\u00a0S18C) bulk structures.\n25\n\nThe stability of these copper (hydroxy)chloride ensembles was ascertained by calculating energies of chlorine desorption to Cl\u2212 for the three most stable structures identified. Over the Cu2OCl2-like ensemble, the energy of both subsurface Cl atoms is \u0394G\ndes\u00a0= \u22120.26 eV at U\u00a0= \u22120.80\u00a0V vs. RHE, and kinetic trapping due to poor Cl\u2212 diffusion to the surface is unlikely (process is endergonic by 0.2 eV). This suggests that Cl atoms are strongly stabilized if ensembles are generated at the surface, as their desorption Gibbs free energy is much higher (1.0\u20132.5 eV more endergonic) than those of other subsurface Cl atoms in other structures and also higher (0.2\u20131.5 eV more endergonic) than surface-stabilized Cl atoms (Table\u00a0S13). Despite the suggested stability of CuOHCl-like ensembles by the heuristic model, the desorption energy of Cl subsurface atoms on the structures (\u0394G\ndes\u00a0= \u22121.27 eV) suggests their low stability under applied potential. The unique stability of the Cu2OCl2-like ensembles mirrors that of their bulk counterparts as detected by XRD (Figure\u00a0S3) and XPS (Figure\u00a05A) measurements of used CuO-HCl(T) catalysts, while CuOHCl initially present disappears under reaction conditions (Figure\u00a0S3). Cu2OCl2-like ensembles induce changes in the neighboring Cu d-band centers with respect to Cu atoms located further away, by \u0394\u03b5\n\nd-band\u00a0= 0.32 eV (Figure\u00a0S19), in opposition to Cu(100). Moreover, the most stable Cu2OCl2-like ensemble showed a significant promotional effect toward CH4 formation (Figure\u00a06D), as CO2 adsorption is more favorable by \u22120.57 eV compared with over Cu(100). Subsequent protonation and dehydroxylation steps leading to \u2217C are also slightly more favored by \u22120.10 eV and \u22120.35 eV, respectively, with a stronger CO adsorption by 0.73 eV. Thus, Cu2OCl2-like ensembles could be assigned as sites responsible for C1 HRPs, while for the metastable CuOHCl-like ensembles, no promotion is found (Figure\u00a0S20). Also, Bader charge analysis demonstrates the generation of surface Cu-Cu\n\u03b4+ pairs in Cu2OCl2-like ensembles able to promote C2+ products (Figure\u00a0S21 and Table\u00a0S14). From a wider perspective, the nature of herein suggested active sites for the promotion of methane explains why chlorine-modified copper catalysts reported in the literature have exclusively been claimed to favor multi-carbon products. Since all previous reports were based on copper chloride due to lack of a method to control chlorination degree, the virtual absence of oxygen in the initial material precludes the formation of Cu2OCl2-like ensembles.In summary, this work developed a method to control the chlorination degree of copper electrocatalysts and applied it to Cu, CuO, and Cu2O to investigate structure-performance correlations. Surface chlorine content upon reaction correlates with FE\nHRP and could be maximized for mildly chlorinated Cu2O materials, revealing the importance of both the copper phase and the synthesis procedure. This work also reveals the potential of chlorine-promoted copper catalysts for the production of methane. Computational studies predict two types of sites to explain observed performance patterns. Chlorine incorporation both on Cu surfaces and in stable copper oxychloride phases showed that \u2217CO-CO coupling could be enhanced by mild Cl contents giving rise to asymmetric Cu-Cu\n\u03b4+ pairs with higher reactivity, while methane formation is thermodynamically favored over copper oxychloride-containing structures. These tools and fundamental insights gathered are expected to contribute to the design of the next generation of technical chlorine-promoted copper electrocatalysts.Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Javier P\u00e9rez-Ram\u00edrez (jpr@chem.ethz.ch).This study did not generate new unique materials.Polished Cu foils were oxidized in a flow of dilute oxygen at tailored conditions, followed by chlorination in a gas flow of 2 vol\u00a0% hydrogen chloride diluted in helium at temperatures between 298 K and 523 K, to obtain three families of catalysts: Cu-HCl(T), Cu2O-HCl(T), and CuO-HCl(T). Further details on catalyst preparation are provided in the supplemental experimental procedures with the conditions applied in the oxidation and chlorination treatments fully detailed in Table\u00a0S1.XRD analysis was used to identify and quantify bulk phases and investigate their crystallinity. XPS measurements were carried out to identify copper phases and quantify chlorine content. SEM disclosed structural features at the microscale. Further details for each of the characterization techniques are provided in the supplemental experimental procedures.Catalyst evaluation was performed in a gas-tight two-compartment cell, mounted with a microporous carbon layer as counter electrode and Ag/AgCl 3\u00a0M as reference electrode at \u22120.8\u00a0V vs. RHE in CO2-saturated 0.1\u00a0M KHCO3. On-line gaseous products quantification was performed by gas chromatography every 15\u00a0min, whereas liquid products were analyzed by proton nuclear magnetic resonance (1H-NMR) spectrometry after reaction. An extended description can be found in the supplemental experimental procedures.DFT calculations were performed with the Vienna Ab initio Simulation Package (VASP)\n46\n\n,\n\n47\n using the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional\n48\n\u00a0with dispersion included through the DFT-D2 method.\n49\n Inner electrons were represented through projector augmented wave (PAW) pseudopotentials\n50\n with a plane-wave energy cutoff of 450 eV. Mechanistic studies were performed on the chlorine adatom and heuristically computed copper (hydr)oxychloride models using the computational hydrogen electrode (CHE) model\n51\n\u00a0and solvation.\n52\n Further details on DFT parameters, model generation, and energy calculations are provided in the supplemental experimental procedures.The authors acknowledge financial support from the Swiss National Science Foundation through the National Center of Competence in Research NCCR Catalysis (grant 180544), ETH grant ETH-47 19-1, and from the Spanish Ministry of Science and Innovation (PRE2021-097615, PID2021-122516OB-I00, Severo Ochoa Center of Excellence CEX2019-000925-S 10.13039/501100011033). The Barcelona Supercomputing Centre-MareNostrum (BSC-RES) is acknowledged for providing generous computational resources. T.Z. thanks the Agency for Science, Technology and Research (A\u2217STAR) Singapore for support through a graduate fellowship. The authors thank Thaylan P. Ara\u00fajo and Dr. Simon B\u00fcchele for assistance with XPS measurements and Shibashish Jaydev for assistance with visualizations.T.Z. and F.L.P.V., methodology, data curation, investigation, visualization, writing\u00a0\u2013 original draft. E.I.-A, data curation, investigation, visualization, writing\u00a0\u2013 original draft. R.G.-M. and G.Z., methodology. A.J.M., methodology, investigation, visualization, validation, supervision, writing\u00a0\u2013 original draft. N.L., conceptualization, funding acquisition, supervision, project administration, writing\u00a0\u2013 review\u00a0& editing. J.P.-R., conceptualization, funding acquisition, supervision, project administration, writing\u00a0\u2013 review\u00a0& editing.The authors declare no competing interests.Supplemental information can be found online at https://doi.org/10.1016/j.xcrp.2023.101294.\n\n\nDocument S1. Supplemental experimental procedures, Figures\u00a0S1\u2013S21, and Tables\u00a0S1\u2013S14\n\n\n\n\n\nDocument S2. Article plus supplemental information\n\n\n\n", "descript": "\n                  Chlorinated copper catalysts have shown promise for electroreduction of carbon dioxide to complex products, but the challenging control of chlorination keeps shaded the potential of chlorine as a selectivity promoter. This work develops a gas-phase chlorination strategy based on exposure to diluted hydrochloric acid at different temperatures to study the effect of chlorine content in copper (II) oxide (CuO), copper (I)\u00a0oxide (Cu2O), and metallic copper (Cu) foils. Contrary to CuO and Cu, chlorination of Cu2O enhances the formation of highly reduced products (those requiring more than two electron transfers). Faradaic efficiency toward these products (0%\u201314% at -0.8 V vs. the reversible hydrogen electrode) correlates with the surface chlorine content after reaction (0 to 1.8 atom % chlorine), which is maximized for mild initial chlorination degrees (Cu2O:CuCl\u223c1). Experimental and computational studies suggest metallic copper surfaces with moderate chlorine coverage and oxychloride-like clusters are active sites responsible for the promotional effect. These findings may facilitate structure-performance relationships, forwarding the next generation of this family of catalysts.\n               "}
{"full_text": "Data will be made available on request.The nanomaterial displays an exceptional property such as surface to volume ratio, desirable electrical and optical possessions versus their bulk counterparts. Additionally, due to these unique possessions, the NPs have potential applications in diverse fields with highly promising efficiency, i.e., biomedical, photocatalysis, food industry, energy and environment etc [1\u20134]. Among the different kinds of materials explored, ferrites are a progressive material based on their incredible response and application in diversified fields. Ferrites based on the bandgap energy have phenomenal light absorption efficacy and have high capability of charge shifting. Thus, they are auspicious for application in photocatalysts, sunlight-based gadgets, and electronic devices [5\u20137]. Hence, these materials have been emerged a one of important class of materials for application in photocatalysis, magnetic, optical and electronic materials. Depending on the magnetization, these have two types, i.e., hard and soft ferrites. Based on geometry, these materials have spinel, orthogonal, garnet, and hexaferrites types and are classified into\u00a0X,\u00a0Y, Z, W, U, M, and R types depending on the crystal structures [8,9]. Among the different categories of ferrites, the Sr-hexaferrites, SrFe12O19, are the most interesting and auspicious materials in response of the unique properties. These properties make Sr-hexaferrites appropriate for application as microwave absorbers, recording media, supercapacitors, in spintronics devices, solar based cells, and radiofrequency gadgets [10,11]. Sr-hexaferrites are presented as MFe12O19, having M a divalent cation. Structurally, Sr-hexaferrites including a spinel and R block comprise of oxygen, iron, and strontium ions on tetrahedral, octahedral, and interstitial sites and have RSR*S* sequence in the c-axis [5,12,13]. Doping with di and tri-valent metal cations is regarded as a useful way to deal with the alteration of electrical, optical, magnetic, photocatalytic properties of Sr-ferrites. The optoelectronic properties of Sr-ferrites are mainly dependent on their electronic structures, crystallite size, composition, annealing temperature, dopant type and fabrication routes [14\u201317]. Different synthetic routes have been employed for the fabrication of Sr-hexaferrites, like the facile micro-emulsion route, co-precipitation route, hydrothermal method, sol\u2013gel method, and green synthesis approaches. These synthetic approaches have their own advantages and disadvantages [5,6], which have been applied in various fields and responses reported were highly promising. In this scenario, the pollution a serious issue due to dyes and other toxic pollutants and need to be tackled efficiently. The photocatalytic treatment is one of advanced techniques, which uses photocatalytic process and degrade the organic toxic pollutant non-selectively and convert into non-toxic end products like H2O, CO2 and inorganic ions, which offer various advantages versus conventional wastewater treatment approaches [18\u201321].Based on aforesaid facts and due to the simplicity, easy approach, versatility and cost-effectiveness, the micro-emulsion strategy was used for the fabrication of Zn and Ni doped Sr-hexaferrites in the current study. The doping effect was investigated on the structural, optical, dielectric and photocatalytic properties basis. The photocatalytic activity of Sr1-xZnxFe1-yNiyO19 and SrFe12O19 was appraised for MG dye under solar light exposure.Analytical grade Fe(NO3)3\u00b79H2O (\u226598\u00a0%), Sr(NO3)2 (99.99\u00a0%) and Zn(NO3)2\u00b76H2O (98\u00a0%) were obtained from Sigma Aldrich. The Ni(NO3)2\u00b76H2O (\u226598\u00a0%) and MG dye were obtained from Merck, while C19H42BrN and ammonia solution (35\u00a0% by weight) were acquired from AnalaR.Sr-hexaferrites Sr1-xZnxFe12-yNiyO19 NPs were prepared via micro-emulsion route. Sr-hexaferrites Sr1-x Znx Fe12-y NiyO19 NPs with composition of\u00a0\u00d7\u00a0\u00a0=\u00a00.0, 0.15, 0.30, 0.45, 0.60 and y\u00a0=\u00a00.0, 0.2, 0.3, 0.4, 0.5 were synthesized by micro emulsion procedure. The salts amounts (stoichiometric) were mixed in de-ionized water and stirred on a hot plate at 45\u201355 0C for 2\u00a0h. The CTAB solution was added into each composition. The pH of the mixtures was adjusted at about 11\u201312 with the help of ammonia solution. The mixtures were stirred for 6\u20137\u00a0h. The precipitates of Sr1-x ZnxFe12-yNiyO19 NPs formed are rinsed with de-ionized water repeatedly till neutral pH and drying was done at 150 \u2103 in an oven for about 5\u00a0h. The crystals of Sr1-xZnxFe12 yNiyO19 NPs were ground to fine powder and annealed at 900 \u2103 for 7\u20138\u00a0h in Vulcan A-550 furnace A schematic illustration of the synthesis protocol of Sr1-xZnx Fe12-yNiyO19 NPs is given in Fig. 1\n.The PCE of Sr1-xZnxFe12-yNiyO19 NPs was estimated for MG dye under the exposure of visible light. A 3\u00a0mg of the as-synthesized Sr1-xZnxFe12-yNiyO19 photocatalyst material was added in 50\u00a0mL of MG dye solution having 10\u00a0mg/L concentration. The mixture with MG dye was stirred for half an hour in the dark, which then, was exposed to visible light (200Watt Argon lamp with cutoff filter 420\u00a0nm). For A given specified time, a 5\u00a0mL sample was taken from the mixtures, filtered and then the absorption was recorded at 632\u00a0nm and MG dye removal (%) was estimated as depicted in Eq. (1), where At and A0 are the absorbance values of MG dye solution after and before the irradiation, respectively.\n\n(1)\n\n\nD\ne\ng\nr\na\nd\na\nt\ni\no\nn\n\n\n\n%\n\n\n=\n\n1\n-\n\n\nA\nt\n\n\nA\n0\n\n\n\n\n\n\n\n\n\nThe powder XRD of Sr1-xZnxFe12-y NiyO19 NPs was carried out using Philips x-pert (PRO-3040/60) X-ray diffractometer at room temperature within a wavelength range of 2\u03b8\u00a0=\u00a020-80\u00b0 by CuK\u03b1 \u03bb\u00a0=\u00a00.15406\u00a0nm. FTIR spectra of the sample was recorded using Alpha-Bruker-ATR using OPUS-Mentor software in the range of 390\u20134200\u00a0cm\u22121. Dielectric properties of Sr1-x Znx Fe12-y NiyO19 NPs were measured using 4287-A RF LCR-meter. UV visible was accomplished using SHIMADZU-3101 spectrophotometer.The composition, crystallinity and phase of SrFe12O19 (SF-1), Sr0.85Zn0.15Fe11.8Ni0.2O19 (SF-2), Sr0.7Zn0.3Fe11.7Ni0.3O19 (SF-3), Sr0.55Zn0.45Fe11.6Ni0.4O19 (SF-4) and Sr0.4Zn0.6Fe11.5Ni0.5O19 (SF-5) were investigated by XRD analysis (Fig. 2\na). The diffraction peaks show the hexagonal structure of SrFe12O19 NPs, according to standard pattern (JCPDS# 00\u2013033-1340) [22]. As depicted from the XRD pattern, after doping zinc (Zn2+) and nickel (Ni3+) ions in SrFe12O19 NPs with crystal structure, the peak position at 32.9\u00b0 for (107) plane shifted towards the lower angle of diffraction [23]. The XRD pattern of SF-2 that might be due to growing larger ionic radii of Zn2+ ions (Zn2+ = 88\u00a0pm) in strontium ferrite lattice Fig. 2(b). In the XRD pattern of SF-3, the peak position at 32.9\u00b0 for (107) plane shifted towards the larger angle of diffraction, which again might be due to growing smaller ionic radii of Ni3+ ions (Ni3+ = 74\u00a0pm) in strontium ferrite lattice [24]\nFig. 2(b). The XRD pattern of SF-4 was similar to that of SF-2 in which the peak position was shifted to lower diffraction angle. However, the XRD pattern of SF-5 was quite similar to that of SF-1. The unit cell parameters such as cell volume and side lengths \u2018a\u2019 & \u2018c\u2019 and crystallite size were found to be in the range from SF-1 to SF-5\u00a0V\u00a0=\u00a0673.315 to 663.512 (\u00c5)3, a\u00a0=\u00a05.8702 to 5.8601\u01fa and c\u00a0=\u00a022.5629 to 22.3111\u01fa and 21.023 to 14.318\u00a0nm, respectively Table 1\n. It was observed from the Table 1 that the lattice constants \u2018a\u2019 & \u2018c\u2019 and crystallite size decreases which was attributed because of the substitution of larger host (Sr and Fe) ions with smaller dopant (Zn and Ni) ions Fig. 3\n(a-b). The crystallite size is appraised as per Eq. (2).\n\n(2)\n\n\nD\n=\nk\n\u03bb\n/\n\u03b2\nc\no\ns\n\u03b8\n\n\n\n\nWhere \u2018k\u2019 is the Scherrer factor (\u223c0.99). \u2018\u03bb\u2019 is the beam wavelength of X-ray, \u2018\u03b2\u2019 is the FWHM and \u2018\u03b8\u2019 represents angle (Bragg\u2019s) [25]. All the peaks were well indexed at 2\u03b8 value of 21.87\u00b0, 23.82\u00b0, 25.62\u00b0, 28.53\u00b0, 32.98\u00b0, 35.33\u00b0, 38.66\u00b0, 40.61\u00b0, 42.41\u00b0, 49.08\u00b0, 53.79\u00b0 and 57.40\u00b0, indicating (103), (104), (105), (106), (107), (103), (201), (116), (205), (206), (213), (300) and (1112) planes of crystal respectively. With increase in the amount of doped metals, the sharpness of the peak at about 32.9\u00b0 angle was more strengthened.The FTIR spectra of SrFe12O19 (SF-1), Sr0.85Zn0.15Fe11.8Ni0.2O19 (SF-2), Sr0.7Zn0.3Fe11.7Ni0.3O19 (SF-3), Sr0.55Zn0.45Fe11.6Ni0.4O19 (SF-4) and Sr0.4Zn0.6Fe11.5Ni0.5O19 (SF-5) were observed in 400\u20134000\u00a0cm\u22121 wavenumber range (Fig. 4\n). For strontium ferrites, four different types of absorption peaks in 450\u20133800\u00a0cm\u22121 range were observed; two bands were of high frequency abbreviated as f1 and f2, while the other two vibrational modes were of low frequency represented as f\u03011 and f\u03012. The vibrational modes of high-frequency, f1 and f2 are attributed to tetrahedral and octahedral stretching vibration of M\u00a0\u2194\u00a0O, while the remaining two have low vibrational modes, f\u03011 & f\u03012 are attributed to lattice vibrations [26]. Fig. 3 depicts that the vibrational mode of 469\u00a0cm\u22121 is represented as f2 (Mocta\u00a0\u2194\u00a0O), while the band at 568\u00a0cm\u22121 represented as f1 (Mtetra\u00a0\u2194\u00a0O). The tetrahedral sites show variation in band position due to the doping of zinc and nickel ions in SrFe12O19 lattice. By comparing the SrFe12O19 FTIR spectra with some [24,27], there is peak shifting from region of low to high frequency when SrFe12O19 is doped by zinc and nickel ions. The reason is that grain size is reduced by doping zinc and nickel ions. For NPs, it is a common phenomenon that there is a change in the characteristic vibrational frequencies in functional groups due to small variation in the environment. Due to decrease in the size of grain, there is an increase in vibrational frequency [13].The electrical properties of SrFe12O19 (SF-1), Sr0.85Zn0.15Fe11.8Ni0.2O19 (SF-2), Sr0.7Zn0.3Fe11.7Ni0.3O19 (SF-3), Sr0.55Zn0.45Fe11.6Ni0.4O19 (SF-4) and Sr0.4Zn0.6Fe11.5Ni0.5O19 (SF-5) were measured with current voltage (I\u2013V) analysis at normal temperature, as shown in Fig. 5\n. The curves show the semiconducting behavior of SrFe12O19 semiconductors has been tailored in their Sr1-xZnxFe12-yNiyO19 ferrites. The curve of current\u2013voltage related to Sr1-xZnxFe12-yNiyO19 ferrites demonstrates a rectifying behavior, and shows the probable progress of diode heterojunction for Sr1-xZnxFe12-yNiyO19 ferrites [28]. The value of DC resistivity for SrFe12O19 NPs and Sr1-xZnxFe12-yNiyO19 ferrites was appraised as per Eq. (3).\n\n(3)\n\n\nP\n=\nR\n\u00d7\nA\n/\nl\n\n\n\n\nWhere \u201c\u03c1\u201d represents the value of DC resistivity, \u201cR\u201d is the resistance, pellets thickness and the area of SrFe12O19 NPs and Sr1-xZnxFe12-yNiyO19 ferrites is represented by \u201cA\u201d and \u201cl\u201d respectively. The values of DC resistivity for SrFe12O19 NPs and Sr1-xZnxFe12-yNiyO19 ferrites were 9.59\u2a2f107, 5.11\u2a2f104, 8.11\u2a2f106, 1.11\u2a2f105, 1.06\u2a2f106 \u03a9cm respectively. These results clearly show that the value of resistivity decreased to 5.11\u2a2f104, 8.11\u2a2f106, 1.11\u2a2f105, 1.06\u2a2f106 \u03a9 cm in Sr1-xZnxFe12-yNiyO19 ferrites. The decrease in the resistivity value observed for Sr1-xZnxFe12-yNiyO19 ferrites may be due to the doping of Zn2+ and Ni3+ ions in the interstitial regions of the crystal lattice formed by SrFe12O19 that generated structural deformation. The DC resistivity values for SrFe12O19 NPs and Sr1-xZnxFe12-yNiyO19 ferrites were changed into conductivity values. Finally, the values of electrical conductivity for SrFe12O19 NPs and Sr1-xZnxFe12-yNiyO19 ferrites were found to be 1.4\u2a2f10-8, 1.95\u2a2f10-5, 1.23\u2a2f10-7, 9.0\u2a2f10-6, 9.4\u2a2f10-7 Scm\u22121, respectively. The increase in the conductivity value of Sr1-xZnxFe12-yNiyO19 ferrites may be due to the following reasons, (i) increase in scattering of crystallite boundary and (ii) an increase in the electron motion due to development in crystallite size [29]. The values of slope, thickness, area, resistivity and conductivity of the samples are provided in Table 2\n.\nFig. 6\n(a) shows the dielectric constant versus frequency of Zn and Ni-doped SrFe12O19 NPs. The change in dielectric constant value versus frequency shows that it decreases by increasing the frequency range up to a certain frequency region, after which it becomes constant. This behavour of the dielectric constant was explained by Maxwell Wagner and Koop\u2019s phenomenal theory [30]. According to this theory, the dielectric nature of the ferrite materials was based on two layer models, the grain boundaries consisting of high and low conducting properties. Electrons enter these grain boundaries from the varoius interstitial sites causing a barrier by hopping conduction mechanism. At these interstitial sites (boundaries) the resistance offered results in polarization which increases the dielectric constant. While at the high-frequency region the induced polarization decreases due to the decrease in the movement of electrons towards these boundaries, which decreases the dielectric constant at high-frequency region [31,32]. By adding the dopants (Zn and Ni) in the doped material, it was observed that the dielectric constant increased as the concentrations of the dopants increased in SrFe12O19 NPs, as depicted in Fig. 6(b). This might be attributed to the composition of the lattice structure and its changes from the octahedral to tetrahedral sites. This reduces the host (Sr and Fe) content ultimately decreasing the hoping conduction from Fe3+ to Fe2+ ions in the doped materials [33].\nFig. 7\n (a-b) depicts tangent and dielectric loss versus shows the frequency of the SrFe12O19 NPs and Zn and Ni doped ferrites materials. The tangent and dielectric loss factor show a similar trend as depicted by the dielectric constant at different frequency regions. It was observed that the tangent and dielectric loss was high at low frequency region and shows decreasing trend at high frequency region. The decrease of the tangent and dielctric loss with frequency was ascribed to the reduction in the domain wall motion, magnetization and space charge polarization. These, resulted in the decrease of tangent and dielectric losses at high frequency range [5]. The effect of dopant content on the variation of tangent and dielectric loss reveals that the doping of Zn and Ni in Sr-hexaferrites decreases the dielectric and tangent loss. It was observed that the tangent and dielectric loss was low for the doped material compared to Sr-hexaferrites. The very low loss value for the doped material was attributed to smaller oxygen vacancies, low resistance offered, hopping conduction mechanism, grain interfaces and the grain boundaries. The Zn and Ni doped Sr-hexaferrites material exhibits low resistivity, which needs small amount of energy to exchange the electrons from Fe3+ to Fe2+ ions resulting in the reduction of the dielectric and tangent loss values. The doped Sr-hexaferrites material with smaller values of tangent and dielectric loss exhibit a very small leakage current. This makes this material to be potentially suitable for energy storage, electronic, microwave and energy storage devices [34].The material possesses AC conductivity due to electrons transfer in various valance states. Fig. 8\n(a) depicts AC conductivity of Zn and Ni doped and undoped Sr-hexaferrites. AC conductivity value was enhaced as the concentration of the dopants (Zn, and Ni) was increased as well as frequency. This increase in the AC conductivity values was attributed to the high conduction power of dopant metal ions which can donate electrons easily compared to the host metals. This results in increase in polarization, which might be due to relaxation and reorientational phenomenon, and dispersion effects as illustrated by the jumping relaxation model [13]. The electrical resistivity measurements of the Zn and Ni doped and undoped Sr-hexaferrites materials are illustrated in Fig. 8(b). The electrical resistivty values declined with frequency and concentrations of the dopants. The resistivity value was higher in low-frequency region and decreased onwards with increased frequency. The decreased values of electrical resistivity upon doping might be attributed to reduction in the resistance offered by the dopant ions in the movement charge carriers species, that is electron which results in increase in the hoping conduction mechanism. The material with small values of electrical resistance at ordinary conductions are the potentail candidates for electrical and electronic devices applications [5].UV\u2013Visible absorption spectrum of SrFe12O19 (SF-1), Sr0.85Zn0.15Fe11.8Ni0.2O19 (SF-2), Sr0.7Zn0.3Fe11.7Ni0.3O19 (SF-3), Sr0.55Zn0.45Fe11.6Ni0.4O19 (SF-4) and Sr0.4Zn0.6Fe11.5Ni0.5O19 (SF-5) NPs was analyzed in 200\u2013800\u00a0nm range. Obviously, all the samples absorb both in UV and visible range and four distinct absorption bands (Fig. 9\n). One absorption band was in the visible light range and the remaining three were observed in UV light region; however, the position of these absorption bands was changed by changing the concentration of the dopants Zn and Ni metal ions in SrFe12O19 lattice Table 3\n. Hence, the variation observed in the absorption bands of pure & zinc- nickel doped strontium ferrites NPs might be due to changing some aspects such as oxygen deficiency, crystallite size and structural defects [35].The value of band gap energy of all samples were calculated using Eq. (4).\n\n(4)\n\n\n\n\n\n\n\u03b1\nh\n\u03c5\n\n\n\n2\n\n=\n\nK\n\n\n\n\nh\n\u03c5\n-\n\nE\ng\n\n\n\n\nn\n\n\n\n\n\nWhere \u2018h\u028b\u2019, \u2018Eg\u2019, \u2018\u03b1\u2019, \u2018K\u2019 and \u2018n\u2019 are indicating photon energy, band gap energy, coefficient of absorption, constant and transition type (indirect, direct, forbidden, and allowed), respectively. The value direct band gap energy (n\u00a0=\u00a02) for pure and doped strontium ferrite NPs was calculated by extra-plotting (\u03b1h\u028b)2 verses h\u028b (Fig. 9) and band gap energy values are provided in Table 3. The band gap energy values of SF-1, SF-2, SF-3, SF-4 and SF-5 were calculated to be 1.24(eV), 1.21(eV), 1.18(eV), 1.18(eV) and 1.09(eV), respectively. It was observed from the Table 3 that the bandgap values decreases from un-doped SF-1 to the highly doped SF-5. This variation in bandgap energy might be ascribed to the formation of sub-energy levels and the defects caused by doping Zn and Ni ions in Sr-ferrite [5].The PCE of the synthesized photo-catalyst Sr1-xZnxFe12-yNiyO19 NPs material with highly dopant content Zn, and Ni (x\u00a0=\u00a00.6, y\u00a0=\u00a00.5) was evaluated using MG dye, as depicted in Fig. 10\n. From the absorption spectrum it was observed that the as-fabricated Sr1-x Znx Fe12-y NiyO19 NPs doped material degraded almost 72.23\u00a0% of the MG dye in one hour under visible light-irradiation. The rate constant for degrading the MG dye by the as-fabricated Sr1-xZnxFe12-yNiyO19 NPs was observed to be 0.03163\u00a0min\u22121 as illustrated in Fig. 11\n (a-b).The photodegradation mechanism of dye by the as-fabricated Sr1-xZnxFe12-yNiyO19 NPs is described in Fig. 12\n. Photo-catalytic mechanism involves the activation of active sites at the surface of the photo-catalyst materials exposed to visible light irradiation. As a result of light exposure, e- are excited from VB to CB generating an electron-hole pair. It converts the water molecules for generating O\u1e22 radical, which has oxidizing nature [13]. The (O\u1e22) radical thus formed oxidizes MG dye to non-toxic small inorganic and organic end-products. On the other hand, oxygen takes up electrons forming super-oxide (O2\n\u2013) anionic radical, which after being protonated forms H2O2. The H2O2 produced further associates and produce O\u1e22 radical, which interacts with dye structure and causes the degradation of MG dye molecule to simple non-toxic degraded end-products [5], as depicted in Eqs. 5\u20138. Table 4\n shows a comparison of photo-catalytic activities of related photocatalytic material versus present study and analysis revealed that Sr1-xZnxFe1-yNiyO19 NPs has promising photocatalytic activity and can be employed to the remedy of coloring agents in the textile effluents under visible light exposure and and under the current situation of water contamination with diverse type of pollutants [36\u201340], there is need to develop and adopt eco-benign methods for wastewater treatment and photocatalytic processes under solar light irradiation is one of promising in this regard.\n\n(5)\n\n\nP\nh\no\nt\no\nc\na\nt\na\nl\ny\ns\nt\n+\nI\nr\nr\na\nd\ni\na\nt\ni\no\nn\n\u2192\n\n\nh\n\n+\n\n\n\nC\nB\n\n\n+\n\n\ne\n\n-\n\n\n\nC\nB\n\n\n\n\n\n\n\n\n(6)\n\n\n\n\ne\n\n-\n\n+\n\nO\n2\n\n\u2192\n\nO\n\n2\n\n\n-\n\u2219\n\n\n\n\n\n\n\n\n(7)\n\n\n\n\n\nh\n\n+\n\n\n\nC\nB\n\n\n+\n\n\nO\nH\n\n-\n\n\u2192\n\n\nO\nH\n\n\u2219\n\n\n\n\n\n\n\n(8)\n\n\n\n\nO\nH\n\n\u2219\n\n+\n\nO\n\n2\n\n\n-\n\u2219\n\n\n+\nM\nG\n\nd\ny\ne\n\u2192\nd\ne\ng\nr\na\nd\ne\nd\n\np\nr\no\nu\nc\nt\ns\n\n\n\n\nA photocatalyst's stability is crucial for commercial application. As a result, the reusability and stability of Sr1-xZnxFe12-yNiyO19 photocatalyst was tested four times using repeated cycle runs. Following each experiment, the photocatalyst was removed from the MG aqueous solution by ultracentrifugation and dried at 50\u00a0\u00b0C. The recycling studies, givenin Fig. 13\n, show that there was slight decline in the photocatalytic activity of Sr1-xZnxFe12-yNiyO19 after each run which can be attributed to catalyst loss due to deactivation of active sites and during separation process for next cycle and aggregation or leaching of the photocatalyst's surface as a result of subsequent heat treatment [45].Sr1-xZnxFe12-yNiyO19 NPs were successively synthesized by micro-emulsion method. Analysis of the synthesized Sr1-xZnxFe12-yNiyO19 NPs was performed by different techniques along with photocatalytic application studies. The XRD analysis confirmed the growth of SrFe12O19 with hexagonal crystal lattice. The doping of Zn and Ni metals in SrFe12O19 crystal lattice was confirmed by shifting of peaks in the XRD pattern of Sr1-xZnxFe12-yNiyO19 NPs. The dielectric properties of the doped and undoped material reveal the potential uses in electronic and electrical devices. The study of the optical properties of pristine SrFe12O19 and doped Sr1-xZnxFe12-xNiyO19 NPs proved the change in the absorption due to, the engineering of bandgap energies. The minimum bandgap energy of 1.09\u00a0eV was achieved for SF-2 by controlling Zn and Ni concentrations. The PCA of Zn and Ni-doped material was evaluated for MG dye and 72.23\u00a0% dye was degraded in 60\u00a0min under visible light irradiation. The doped material might have potential applications as photocatalyst under visible light irradiation, which will make the process economical for application at pilot scale.This research was funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project (PNURSP2022R26), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. The authors extend their appreciation to the Deanship of Scientific Research at King Khalid University, Saudi Arabia for funding this work through Research Groups Program under Grant No R.G.P.2: 33/43.\nIsmat Bibi: Conceptualization, Supervision. Shahid Iqbal: Investigation, Writing \u2013 original draft. Shagufta Kamal: Validation. Qasim Raza: Methodology. Mongi Amami: Investigation, Data curation. Khadijah M. Katubi: Funding acquisition, Data curation, Resources. Norah Alwadai: Software, Resources. Munawar Iqbal: Writing \u2013 review & editing, Writing \u2013 original draft, Visualization.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors express their gratitude to Princess Nourah bint Abdulrahman University Researchers Supporting Project (PNURSP2022R26), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. The authors extend their appreciation to the Deanship of Scientific Research at King Khalid University, Saudi Arabia for funding this work through Research Groups Program under grant number R.G.P.2: 33/43.", "descript": "\n                  A series of Sr1-xZnxFe1-yNiyO19 with composition of x\u00a0=\u00a00.0, 0.15, 0.30, 0.45, 0.60 and y\u00a0=\u00a00.0, 0.2, 0.3, 0.4, 0.5 were prepared using micro-emulsion approach and characteristics were studied by XRD, FTIR and UV\u2013vis techniques. The effect of doping was investigated based on the optical, dielectric, structural and photocatalytic properties of Sr1-xZnxFe1-yNiyO19. XRD analyses confirmed the hexagonal crystal structure of pure and doped SrFe12O19 NPs. The lattice parameters, X-ray density, porosity, bulk density and crystallite size of Sr1-xZnxFe1-yNiyO19 and SrFe12O19 were compared. The FTIR analysis confirmed that there was shifting of peaks from region of lower to higher frequency in Sr1-xZnxFe1-yNiyO19, due to the reduction of the size of grains after doping. The current\u2013voltage I-V data shows that there is gradual increase in the conductivity of as prepared doped-material, but the reverse was the case for resistivity, which exhibit good conducting behavior of the fabricated material. Evaluation of the dielectric properties of the samples shows that the AC conductivity and dielectric constant enhanced, while tangent, dielectric loss and resistivity decreased with concentration of the dopants (Zn and Ni). These excellent properties of doped material can be used for the fabrication of various electronic, electrical, microwave and high frequency devices applications. The photocatalytic activity of Sr1-xZnxFe1-yNiyO19 and SrFe12O19 was tested against methyl green (MG) dye. Results revealed that Sr1-xZnxFe1-yNiyO19 degraded 72.23\u00a0% of MG dye in 60\u00a0min under the exposure of solar light. The recyclability and usability tests revealed that there was a minute loss after successive four runs. Sr1-xZnxFe1-yNiyO19 might have potential applications as a photo-catalyst under solar light irradiation.\n               "}
{"full_text": "Functional amines, among different kinds of chemicals, are considered as the highly valuable precursors widely applied in many fields of biology, medicine, and material [1\u20133]. For instance, the top-selling drugs reported in 2020 usually contain the nitrogen and/or amino groups as the integral parts that play an important role in their activities [3,4]. Various strategies for the synthesis of functional amines are thus developed, including but not limited to the reduction of functional groups containing nitrogen, the alkylation of ammonia and amines, the Gabriel synthesis, and the reductive amination of carbonyl compounds [1,5]. Among them, the last one represents the most resourceful method adopted in both academic laboratory and industry as it uses molecular hydrogen to be the reducing agent, which is beneficial for the atomic balance and efficiency [4,6,7]. To our knowledge, reductive amination is a cascade reaction [1], where the initial step forms the carbinol amine that loses water to offer imine or iminium ion; subsequently, these intermediates will be further reduced to produce the specific amines. Screening appropriate catalysts in this kind of reaction is thus important because they have to selectively convert the imine without considerably affecting the primary aldehyde or ketone or other available reducible groups.In this regard, Fernandes et\u00a0al. [8] reported a strategy for the direct reductive amination of aldehydes by the catalysis of various high valence oxorhenium (V and VII) complexes. Shortly afterward, Fischmeister et\u00a0al. [9] also described a zwitter-ionic iridium complex catalyst with a 2, 2\u2032-dipyridylamine ligand as effective for the reductive amination of lactic acid (LA) with 4-methoxyaniline. However, Liu et\u00a0al. [10,11], Beller et\u00a0al. [12], and Huang et\u00a0al. [13] clearly stated that the synthesis and application of homogeneous catalysts are usually limited by the disadvantages in terms of environmental safety, stability, and recyclability. For the advancement of sustainable and cost-effective processes, it is preferable to develop non-noble metal-based heterogeneous catalysts, whose catalytic activity is mostly characterized by the support features and metallic phases. Herein, the carbon-based supports, especially for those derived from biomass [14\u201317], have recently received growing attention as they are easily prepared and abundant in nature. Some strategies are also adopted to enhance the interaction of metallic nanoparticles with carbonaceous structures for better catalytic activity, and one of the common ways is the encapsulation of metal nanoparticles into the carbonaceous structure [11,12,18]. Yan et\u00a0al. [19] suggested the synthesis of carbon-encapsulated iron nanoparticles using pyrochar from fast pyrolysis of pine wood as the supporting materials, which possesses a high activity for the conversion of biomass-derived chemicals to liquid hydrocarbons. Similarly, several groups have demonstrated that the carbonized cellulose or other nanostructured carbonaceous materials fabricated with metallic nanoparticles exhibit the desirable performance in the reaction of reduction hydrogenation [12,16]. The coating of metallic nanoparticles on such carbonaceous materials is prepared by the surface precipitation of metal precursor to the hard templates, and the loading process heavily relies on the porous carbonaceous materials. This process technically belongs to the physical adsorption, and the whole procedure is limited by the tedious and complex preparation processes because a series of treatments such as pyrolysis, loading, and reduction is necessary.Very recently, the hydrochar prepared by the facile and mild hydrothermal process of lignocellulose has been used as another sustainable carbonaceous support for the applications of material science [14,15,17,19\u201321]. In contrast to the pyrochar, the hydrochar is easily prepared and characterized by the oxygenated functional groups, even though its specific surface area is relatively low. Biradar et\u00a0al. [14] utilized a straightforward route of synthesis for the flower-like nanoparticles originated from waste bagasse, and this catalyst facilitates the reductive amination of aldehydes with nitroarene. Similar hydrochar-supported catalyst was also reported by Gai et\u00a0al. [17], who uses pinewood sawdust as the carbon resource. Furthermore, Titirici et\u00a0al. [20], Ravi et\u00a0al. [14] and Hu et\u00a0al. [21] stated that such carbonaceous support (i.e., hydrochar) can cooperate with a metal precursor to enhance the dispersion and promote simultaneous/mutually reactions because of the presence of various functional groups, especially for the oxygenated complexes (e.g., hydroxyl, carboxyl, carbonyl). These functionalities can improve the access of metal solutions into carbonaceous matrix as its decrease in hydrophobicity [22]; in addition, its surface functionalities can also provide the anchorage sites for metal precursor and act as the active centers in multifunctional catalysts caused by their acid-base or red-ox properties [23,24]. The loading process of metal precursor in this carbonaceous support intrinsically belongs to the chemical adsorption, but unfortunately, there is still a large uncertainty in this field. The relationship between the craft conditions and the catalytic properties of prepared materials, such as the hydrothermal severity, the loading process, and the post-treatment, have been rarely reported so far. And also, the selection of real biomass in previous studies is another obstacle, since the natural and complex components make it difficult to control the structural features of hydrochar-supported catalysts.In this study, we prefer to use biomass-derived glucose as the carbon source for the preparation of support via the hydrothermal process; meanwhile, the exploration on the influence of synthesis routes or conditions towards the performance of hydrochar-support catalysts are identified. Simple handling of the catalyst and separation from the medium is possible with the applied support, and the preparation process is sustainable since the applied support is very common and available in large quantities. The structural features of catalysts associated with its catalytic activity are characterized by several techniques, and the selected catalysts are next applied to the reductive amination of benzaldehyde as a simple model reaction under varied conditions. Subsequently, with the optimized conditions in hand, we conduct the reductive amination protocol to gram-scale synthesis as well as the lifecycle performance in a batch reactor to prove the potential in industrial application, which may provide an available preparation of simple but highly efficient catalyst for the reductive amination in near future.The chemicals and solvents used in this study were purchased from the certified companies registered in the China Academy Science On-line market systems. For instance, nickel nitrate, cobalt acetate, and ferric nitrate were purchased from Sigma-Aldrich Co., Ltd., whose purification was higher than 99.5%; Methanol (AR, 99.7%), and ethanol (GC, 99.9%) were obtained from Aladdin (Shanghai) Chemical Technology Co., Ltd.; Glucose (AR, 99%) and citric acid (AR, 99%) were purchased from Shanghai Macklin Biochemical Co., Ltd. More specifically, the metal salts and glucose were used as the precursor and carbon source, respectively, while the citric acid was adopted as the inducer which can improve the carbonization degree of carbonaceous support; all of them were applied without any further purification after receiving.Biomass-derived glucose was applied to produce hydrochar as the support for metallic nanoparticles, which was heterogeneously loaded on the carbonaceous matrix via two protocols, as explained below. The catalysts prepared by the impregnation process (method A) and one-pot synthesis (method B) were labeled as x@HCIm-y or x@HCOp-y, respectively. In here, x was the type of metal while y was the hydrothermal temperatures, and the procedure was illustrated in Fig.\u00a01\n for reference.Hydrothermal process was conducted on a stainless steel autoclave reactor (as shown in Fig.\u00a0S1), and the detailed experimental procedure can be referred to our previous literatures [10,11,25]. Briefly, 3\u00a0g of glucose, 0.15\u00a0g of citric acids, and 30\u00a0mL of deionized water were loaded into the reactor to keep the solid-liquid ratio at approximately 1:10; afterward, the reactor was sealed and a high-purity N2 flow (99.999%) was injected into the reactor through the inlet pipeline to create an inert atmosphere for subsequent carbonization. The hydrothermal temperatures were in a range of 180\u2013240\u00a0\u00b0C (with an interval of 30\u00a0\u00b0C), while the holding period and heating rate were fixed at 16\u00a0h and 5\u00a0\u00b0C min\u22121, respectively, to avoid the effects caused by secondary factors. Additionally, the magnetic stirrer was operated at a constant rate of 300\u00a0rpm throughout the whole hydrothermal process to attain the homogeneous heating. Once the carbonization stage was finished, the reactor was rapidly cooled down, and the hydrochar was separated from the resultant mixture by vacuum filtration and dried in an oven at 105\u00a0\u00b0C for at least 12\u00a0h.In the impregnation method, around 5\u00a0g of the hydrochar was impregnated into a 50\u00a0mL solution with a metal concentration of 0.50\u00a0mol\u00a0L\u22121. Subsequently, the aqueous system was kept at 55\u00a0\u00b0C for slow evaporation, meanwhile, the samples were magnetically stirred to promote the uniform deposition of metal ions into carbonaceous matrix. Importantly, EtOH was used as the solvent in this aqueous system, which not only leads to a well-distributed of metallic atoms but also makes the present protocol safer than other in which highly dangerous solvent was used. Following the impregnation step, the samples were ground and calcined at 600\u00a0\u00b0C for 3\u00a0h, with an increasing rate of 5\u00a0\u00b0C min\u22121 and a constant flow of N2 (99.999%) at 60\u00a0mL\u00a0min\u22121. The residual solid was washed with deionized water and then vacuum freeze-dried at \u221241\u00a0\u00b0C overnight to obtain the hydrochar-supported catalyst, which was stored in a vacuum vessel waiting for analysis.In the one-pot synthesis, the general procedure was similar to that of the hydrochar preparation described in Section 2.2.1, but the metal precursor was added into the aqueous system (0.50\u00a0mol\u00a0L\u22121) prior to the hydrothermal process, which might induce the chemical linking between the metal ions and the functional groups during the synthesis of hydrochar. Herein, the metal precursor can be in-situ reduced and anchored by the functional groups in hydrochar, thereby leading to the formation of inner-sphere surface complexes with metallic nanoparticles. Afterward, the obtained solids were filtered from the resultant mixture and washed with deionized water to remove the unloaded metals. Finally, the prepared catalysts were vacuum freeze-dried and stored under identical conditions as introduced in method A.Generally speaking, the performance of heterogeneous catalysts is characterized by two aspects: one is relevant to the structural features of carbonaceous support, and another is closely related to the metallic nanoparticles. In an attempt to explore the comprehensive information, both aspects were carefully analyzed with the help of several techniques as described below.For the structural features of carbonaceous support: 1) the morphological information was measured using a high resolution transmission electron microscopy (HRTEM), and another high-angle annular darkfield scanning transmission electron microscope (HAADF-STEM; JEM-2100F, Japan) was coupled with an energy dispersive X-ray spectrometry (EDS; Thermo Scientific, Waltham, MA) to measure the distribution of metallic atoms. At the same time, the particle surface was investigated by scanning electron microscopy (SEM, S-4800, HITACHI, Japan). The surface area and pore structure were also analyzed by a Micromeritics Gemini VII 2390 gas-adsorption analyzer according to the N2 isothermal adsorption/desorption at \u2212186\u00a0\u00b0C in the relative pressure (P/P\n\no\n) between 0.01 and 0.99; 2) the crystal structure was explored by XRD (PANalytical, X'Pert PRO, Netherlands) with a Cu K\u03b1 radiation (\u03bb\u00a0=\u00a00.15406\u00a0nm), and the scanning conditions were adjusted from 20\u00b0 to 80\u00b0 in 2\u03b8 range with 0.0167\u00b0 step interval. In addition, given the correlation of Raman spectra with the parameters calculated from XRD, a Raman spectrometer (LabRAM HR800-LS55, France) was applied for the specific microstructure information of metal-based catalysts. The light source was provided by Nd-YAG at 532\u00a0nm, and the scanning region was selected between 1800\u00a0cm\u22121 and 1000\u00a0cm\u22121. XPS analysis was adopted to characterize the surface chemistry, especially for the metallic states and carbon functionalities, of solid samples to a depth of around 0.1\u20131\u00a0nm. XPS spectra were obtained on a Thermo Scientific ESCALAB 250Xi spectrometer, which was equipped with Al (K\u03b1) X-ray radiation source (hv\u00a0=\u00a01486.6\u00a0eV) at a 20\u00a0eV pass energy, a 0.1\u00a0eV energy step, and 0.1\u00a0s dwelling time.For the catalytic activity of metallic nanoparticles: 1) the total acidity and acid strength distribution were determined by temperature-programmed desorption of ammonia (NH3-TPD) with an ASIQACIV200-2 automatic physical/chemical adsorption analyzer (Quantachrome, U.S.). The sample (150\u00a0mg) was loaded into the U-tube quartz reactor and heated to 200\u00a0\u00b0C for 30\u00a0min (heating ramp of 10\u00a0\u00b0C min\u22121) under 30\u00a0mL\u00a0min\u22121 of He flows, with the purpose of moisture removal. The system was cooled back to 80\u00a0\u00b0C, and the 8% NH3/He mixed gas was switched for the chemical adsorption of NH3 on the sample, following with another He flow of 30\u00a0mL\u00a0min\u22121 to remove the physically adsorbed NH3. Next, the system was heated to 800\u00a0\u00b0C with a heating ramp of 10\u00a0\u00b0C min\u22121 under the same He flows; 2) the distribution of Bronsted (B) and Lewis (L) acidity was examined by the pyridine-absorbed Fourier transform infrared spectrometer (Py-FTIR; Nicolet 6700, Orlando, FL). A 100\u00a0mg sample was vacuum-activated (1\u00a0\u00d7\u00a010\u22124\u00a0mmHg) at 200\u00a0\u00b0C for 60\u00a0min to remove the moisture, and the background spectrum was recorded after the sample was cooled back to 50\u00a0\u00b0C. Subsequently, the sample was exposed to the pyridine (Aldrich, GC, purity \u226599.5%) vapor for adsorption within 15\u00a0min, and the experimental spectrum was recorded several times after the extra pyridine was completely removed by vacuum-pumping.The reductive amination of benzaldehyde was conducted in a stainless autoclave reactor (NS6-20D-SS1, Anhui Kemi Instrument Co., Ltd.), coupled with six independent channels for different substrates. In a typical run, 0.5\u00a0mmol benzaldehyde, 10\u00a0mg catalyst, and 5\u00a0mL 7\u00a0mol\u00a0L\u22121\u00a0NH3 (in MeOH) were accurately weighed into one of the channels; next, the reactor was flushed with H2 several times to remove air and charged 2\u00a0MPa H2 at the final time. The reactor was then heated up to the target temperature for a given period, and the magnetic stirrer was rotated at a constant speed of 300\u00a0rpm throughout the whole period to ensure a homogeneous reaction. Afterward, the products in the mixture were filtered and detected by gas chromatography to primarily judge the feasibility of the reactions. The target products were further identified by GC-MS (Thermo Trace 1300-ISQ QD, USA) and 1H NMR (Avance III 400\u00a0MHz NMR, Bruker, Germany) to calculate the corresponding conversion and selectivity, where the 1, 3, 5-trimethoxybenzene was used as the internal standard.It is generally agreed that the catalytic activity of heterogeneous catalysts is the result of a complex interplay among multiple factors, including not only the structural features of carbonaceous matrix but also the specific activities of metallic phases in samples [26]. First of all, the SEM images exhibit the morphologies of the prepared catalyst, and it is obvious to find that upon hydrothermal carbonization of glucose, micrometer-sized carbon spheres are observed in all samples. In the present case, the addition of citric acid induces a change in the particle morphology, that is, the microspheres are formed out of small aggregated particles. This observation is confirmed by Titirici et\u00a0al. [27] who stated that citric acid seems to stabilize the first formed small droplet, thereby preventing them from further growth, as might occur in the pure glucose case. Later on in the process, the primary polymerized particles assemble into a micrometer-sized \u201craspberry\u201d-like structure, as depicted in Fig.\u00a02\n. And also, the type of the loaded metals affected the surface structures to a different extent by chemically changing the carbonization degree, as the Ni@HCIm-180 possess relatively smoother spheres than that of Co@HCIm-180 and Fe@HCIm-180. This situation might be relevant to the specific amount of metal coupling with the functional groups, as their concentration in the catalyst systems is Ni\u00a0>\u00a0Co\u00a0>\u00a0Fe, reaching 74.9\u00a0wt%, 17.0\u00a0wt%, and 5.4\u00a0wt%, respectively (as shown in Fig.\u00a02(a)\u2013(c)). In addition, concerning the cobalt-loaded catalysts under different temperatures, it can also be found that the \u201craspberry\u201d-like structure was slowly destroyed as a result of the serious decomposition [18,28], leading to more porous microstructures in Co@HCIm-210 (Fig.\u00a02(d)) and Co@HCIm-240 (Fig.\u00a02(e)). In contrast to Co@HCIm-180 prepared by the impregnation method (method A), a smoother structure in the nanoparticles prepared by one-pot synthesis (method B) was clearly observed in Fig.\u00a02(f), which may inhibit the interactions between the active sites and the reactants during catalytic reactions.In addition, the TEM images were in good agreement with the above finding and further confirms the solid characteristic. As shown in Fig.\u00a03\n(a), Co-based nanoparticles are well-dispersed and surrounded by a combination of some graphitic layers and short-range ordered graphitic spherical shells with an average size of 6.03\u00a0nm, while Ni-based (Fig.\u00a0S2) and Fe-based (Fig.\u00a0S3) nanoparticles are easily sintered when operating at the same temperature (i.e. 600 \u00b0C). An aggregation of Co@HCIm into larger particles can be observed when hydrothermally treated at higher temperatures, increasing to 8.47\u00a0nm and 12.11\u00a0nm in Co@HCIm-210 and Co@HCIm-240, respectively. However, some of the large particles with size over approximately 100\u00a0nm can also be observed in Fig.\u00a03(a), which might be attributed to the slight agglomeration of partial nanoparticles on the surface. Their catalytic activity is limited by the particle size [22,26]; as a rule, the formation of large particles of the active phase in catalysts is undesirable because of the low reactivity, which redounds in detrimental economic consequences. Not only that, a high degree of dispersion of the active phase is essential since it allows the efficient diffusion, transportation, and transfer of reactants to the catalytic active sites [29]. In this consideration, hydrochar is used as support because its oxygenated functional groups allow the preparation of nanomaterial with good dispersion of the active phase. These groups can serve as bridges or coordinating sites to capture and anchor metal ions, followed by in-situ metallic nanoparticles formation after the calcination in N2, which leads to the robust fixation and high dispersion of metallic nanoparticles [17,20], as depicted in Fig.\u00a03(c)\u2013(f). Meanwhile, Fig.\u00a03(b) demonstrated that the Co2+ and Co0 metal existed in one plane, which reveals the presence of the (1 1 1) plane of metallic Co with an inter-planar spacing of 0.20\u00a0nm and (2 2 0) plane of CoO with an inter-planar spacing of 0.21\u00a0nm. The corresponding elemental mapping images of the Co@HCIm also illustrate the coexistence of carbon, oxygen, and cobalt with a mass fraction of 72.8\u00a0wt%, 9.1\u00a0wt%, and 18.1\u00a0wt%, respectively, which could be caused by the superficial oxidation of cobalt in air, consistent with the observations on metallic Co-based materials previously reported [29,30].On the other hand, the morphology of the Co@HCOp catalysts was also characterized by the TEM technique. As shown in Fig.\u00a04\n, the amount of micro-sized particles with nearly spherical morphology are observed in the sample; however, there is no cobalt nanoparticles can be found on the carbon surface although the elemental mapping did reflect the existence of cobalt. This phenomenon can be explained by the reason that the relatively mild conditions provided by the hydrothermal process are insufficient for the crystallization of metallic atoms, where the metal precursor prefer to form the amorphous states during one-pot synthesis [31]. As reported by Gai et\u00a0al. [17], the catalysts prepared by the one-pot synthesis have a more uniform size and spherical shape than those prepared by the impregnation method. Hu et\u00a0al. [21] also stated that metal ions can effectively accelerate the hydrothermal process of carbohydrates to form carbonaceous spheres, while the agglomeration and sintering of metallic nanoparticles may destroy the catalytic activity. Thus, we presume that during the one-pot synthesis, hydrochar can serve as the stabilizing ligands for coating of the out layer of the metallic nanoparticles to prevent the agglomeration and facilitate the formation of highly dispersed metallic nanoparticles on the surface, but further studies are necessary to measure the catalytic activity of metallic cobalt in amorphous states.Subsequently, the Brunauer -Emmett -Teller (BET) and Barrett -Joyner -Halenda (BJH) methods were applied to analyze the different pretreated carbon catalysts, which allows reliable pore size, porosity, and surface area characteristics to be calculated in the pore width range from micro-to mesoporous structures [18,29], as shown in Fig.\u00a05 and Table 1\n. Following with the N2 adsorption-desorption isotherms at \u2212196\u00a0\u00b0C, it can be calculated that the specific surface areas of the prepared catalysts are in a sequence of Fe@HCIm-180 (338.17\u00a0m2\u00a0g\u22121)\u00a0>\u00a0Ni@HCIm-180 (331.09\u00a0m2\u00a0g\u22121)\u00a0>\u00a0Co@HCIm-180 (290.46\u00a0m2\u00a0g\u22121). In addition, all of them present a type IV isotherm as a capillary condensation step in the adsorption branch is clearly observed, which is attributed to the presence of slit-shaped pores [18]. This situation corresponds to an irregular but well-developed porous structure, as a large amount of N2 is adsorbed in the entire relative pressure range, which is also confirmed by their higher surface areas and larger pore volume. Interestingly, some early studies demonstrated that the catalyst containing porous structures facilitates the rapid transportation of products and reactants, thus playing a key role in regulating product distribution [29,32]. For the series catalysts of Co@HCIm, the value of specific surface area and pore volume gradually enlarge with the increased hydrothermal severity, but it will also cause the agglomeration of cobalt atoms at the same time. Concerning the Co@HCOp, a great decrease of the porous structure is found with only 12.02\u201314.41\u00a0m2\u00a0g\u22121 of specific surface area and 0.014\u20130.024\u00a0cm3\u00a0g\u22121 of total pore volume, which can be explained by the formation route of hydrochar during one-pot synthesis as discussed above [27,28]. Furthermore, Fig.\u00a05\n also exhibits the pore size distribution of the prepared catalysts. It is noted that these catalysts display a peak centered at around 2\u201310\u00a0nm, indicating that the hydrochar is a mesoporous material that can enhance catalytic activity and provide efficient diffusion [17]. The average pore diameters of Ni@HCIm-180, Fe@HCIm-180, and Co@HCIm-180 are calculated to be 3.432\u00a0nm, 3.814\u00a0nm, and 3.075\u00a0nm, while their corresponding total pore volumes are 0.169\u00a0cm3\u00a0g\u22121, 0.114\u00a0cm3\u00a0g\u22121, and 0.065\u00a0cm3\u00a0g\u22121, respectively. And also, the pore diameter and pore volume of the Co@HCIm-180 is larger than those of the Co@HCOp-180, which is supported by their trends in specific surface area.To further investigate the structure-activity relationship, XRD and Raman techniques were adapted together as they can provide comprehensive insight into the structural features [28], such as the crystallization and aromatization of samples. As can be seen from Fig.\u00a06\n(a), the types of metal precursors will cause differences in XRD patterns. 1) For Ni@HCIm-180, nickel nitrate will decompose into metallic Ni when treating with high temperature without any additional reductant, which is possibly caused by the in-situ reduction of nickel oxide during the carbonization of hydrochar in N2 [10,11]. As a result, the strong peaks of Ni (111), Ni (200), and Ni (220) are presented at the same time. 2) For Co@HCIm-180, the diffractions at 2\u03b8\u00a0=\u00a044.3\u00b0, 51.4\u00b0 and 75.8\u00b0 belong to the diffraction of (111), (200) and (220) facets of metallic Co, respectively, while another peak associated with 2\u03b8\u00a0=\u00a036.6\u00b0 assign to the (111) planes of CoO, meaning that Co2+ and Co0 metal are co-existed in the samples [4,12,29,30]. Furthermore, with the increase in hydrothermal temperatures, the intensities of the metallic phase increased slightly, indicating the increase of the crystallinity, as depicted in Fig.\u00a06(b); 3) For Fe@HCIm-180, there are several prominent diffraction peaks located at 2\u03b8\u00a0=\u00a035.4\u00b0, 62.6\u00b0, and 44.7\u00b0, which can be well indexed to characteristic (311) and (440) reflections of Fe2O3 and (110) reflections of metallic Fe, respectively [17,19]. An interstitial compound of iron carbide is also observed, as the diffraction peaks around 42.9\u00b0 correspond to (211) planes of Fe3C. The XRD spectra of all samples exhibit the strong and broad C (002) peak [10,32], which indicates that the metallic nanoparticles have been adsorbed and co-exist with the carbonaceous matrix on the support. In contrast, no obvious peaks or differences are observed among the Co@HCOp catalysts due to the amorphous states of metallic phase and carbonaceous matrix, as depicted in Fig.\u00a06(c), corresponding with the statistical analysis of the elemental mapping in TEM.The carbonaceous structure of the prepared catalysts was further characterized with the help of Raman spectra, whose results are illustrated in Fig.\u00a06(d). All of the hydrochar-supported catalysts exhibit a D band around 1360\u00a0cm\u22121 due to the disordered arrangement and low symmetry of graphite lattice structure, while another G band at 1580\u00a0cm\u22121 is a scattering peak attributed to the stretching of all sp\n\n2\n atomic pairs on carbon ring or long carbon-chain in graphite [27,28]. The intensity of the D band to G band (ID/IG) of the prepared catalysts is calculated to be 1.98\u20132.26 (in method A) and 2.43\u20132.88 (in method B), suggesting much more defects are found in the graphitic network. This situation confirms the result of SEM image that the materials obtained via the hydrothermal carbonization of glucose in the presence of citric acid are highly carbonized. The value of La (R\u00a0=\u00a0ID/IG, La\u00a0=\u00a044/R) was also calculated to characterize the degree of aromatization level [33], reaching 22.2, 19.5, and 21.6 for Ni@HCIm-180, Co@HCIm-180, and Fe@HCIm-180, respectively. This difference indicates the type of metal precursor might affect the aromatization of carbonaceous matrix. In addition, it is noted that both intensities in the Co@HCIm and Co@HCOp increase gradually at higher hydrothermal temperatures, indicating that carbon accumulation occurs and leads to the much higher number of covering graphitic layers on the surface.XPS was employed to investigate the elemental compositions and chemical states of the specific species on the samples. As can be seen from Fig.\u00a07\n(a), (d), and (g), the types of metal precursors will cause differences in the patterns, but a similar trend can be observed: both of the metallic and oxidized states co-existed on the surface of the catalyst, but the former occupies the dominant parts. For instance, the chemical state of Co 2p\n\n3/2\n is mainly in metallic form with the binding energy at around 777.8\u00a0eV, while that of Co 2p\n\n3/2\n\u00a0at 780.5\u00a0eV belongs to the oxidized form of Co2+; a strong Co 2p\n\n3/2\n satellite peak at 783.7\u00a0eV and Co 2p\n\n1/2\n satellite peak at 799.8\u00a0eV are found, which has been used for the identification of cobalt species [34,35]. By calculating and comparing the corresponding peak areas of different Co species in the series samples of Co@HCIm and Co@HCOp, the Co species are mainly in the form of metallic cobalt in Co@HCIm rather than Co@HCOp, but the increase of hydrothermal temperatures might convert part of Co0 into Co2+, as depicted in Table S1. In the C 1s spectra (Fig.\u00a07(b), (e), and (h)), the wide peak ranging from 282 to 292\u00a0eV can be resolved into five individual peaks corresponding to the CH bonds CC sp\n\n3\n graphene bonds, CC sp\n\n3\n graphene oxide bonds, CO bonds, and OCO bonds, respectively [28,36]. Among them, the graphitic structure of CC occupies the dominant role while the graphene oxide bond ranks second, suggesting that the prepared catalysts are highly carbonized but might undergo surface oxidation during the preparation [29]. Further evidence of the surface oxidation can be found through the peaks of O2, O3, and O4 in the O 1s spectra (Fig.\u00a07(c), (f), and (i)); not only that, the O1 peak at 529.3\u00a0eV is representative of a metallic oxide network [30], which supports the results of elemental mapping and XRD.In an attempt to identify the effect of acidity on catalytic activity, the relevant types and total amount of acid sites were collaboratively examined by Pyridine adsorption and NH3-TPD, whose profiles are depicted in Fig.\u00a0S4. In general, the Py-FTIR spectra of prepared catalysts are detected between 1400 and 1650\u00a0cm\u22121. The reflection bands at approximately 1450\u00a0cm\u22121 and 1590\u00a0cm\u22121, which is ascribed to the adsorption of pyridine on the Lewis acid sites, are all observed for Ni@HCIm-180, Fe@HCIm-180, and Co@HCIm-180; at the same time, only a tiny reflection bond at 1540\u00a0cm\u22121 is found, which indicates that the Br\u00f8nsted acid sites can almost be negligible [37,38]. The Br\u00f8nsted acid sites can be attributed to the hydroxy groups on the support and easy to be dissociated at high temperatures, while the reduced metal species can act as Lewis acid sites or electrophilic sites to polarize and facilitate the cleavage of C\u2013O bond [33]. It is reasonable that most of the acid sites on the surface of prepared catalysts belong to Lewis acid sites, which is the dominant species of these catalysts regardless of the type of metal precursors. Afterward, Co@HCIm-180 is taken as an example to further determine the specific acid sites by NH3 adsorption, which can be classified into weak (<250\u00a0\u00b0C), medium (250\u2013450\u00a0\u00b0C), and strong (>400\u00a0\u00b0C) acidic sites based on the procedure temperatures [36,38]. As a result, it is found from the NH3-TPD profile that the catalyst has a relatively lower amount of medium acid site and a much high amount of strong acid sites. Meanwhile, Lv et\u00a0al. [36] also reported that the number of acid sites identified by NH3 adsorption will be decreased with the increase of hydrothermal temperatures, indicating that the higher hydrothermal severity reduces the surface acidity.All of the prepared catalysts as well as some commercial catalysts were tested by the reductive amination of benzaldehyde in the presence of ammonia solution and H2 to produce the benzylamine, which is presented as a structural motif in several bioactive molecules [4,39]. According to the primary experiments, as shown in Table 2\n, Co@HCIm-180 (Entry 3) exhibits the best catalytic activity and selectivity towards 1a when compared to most commercial catalysts (Entry 8\u201313) and other catalysts prepared by different metal precursors (Entry 1: Ni@HCIm-180; Entry 2: Fe@HCIm-180), hydrothermal temperatures (Entry 4: Co@HCIm-210; Entry 5: Co@HCIm-240) or procedures (Entry 6: Co@HCOp-180), reaching a satisfying conversion and yield of approximately 99% and 82%, respectively.First of all, Gould et\u00a0al. [40] identified that the catalytic ability of transition metals in terms of reductive amination follow an order of Fe\u00a0<\u00a0Ni\u00a0<\u00a0Co, which confirms the present results described above (Entry 1\u20133). Besides, the hydrothermal carbon in this catalyst systems provides abundant adsorption sites for metallic nanoparticles but interacts weakly when compared to other oxide supports, thereby overcoming the drawbacks associated with the formation of inactive mixed oxides [41]. The cobalt oxide can be auto-reduced upon heating under an inert environment, thus leading to a smaller cobalt particle size than that reduced under H2. This auto-reduction is related to two aspects [41,42]: one is the volatiles released by hydrochar during thermal decomposition, which contains CO, CH4, or H2 that can reduce Co2+ to Co0, and another is the formation of smaller cobalt nanoparticles under an inert environment, which is attributed to the rapid diffusion of oxygenated functionalities that induces the migration of cobalt atoms on the carbon surface at higher temperatures. As a result, it is easy to understand that the catalytic selectivity towards 1a decreases with the increase of hydrothermal temperatures as most of the volatile matters are previously decomposed at the hydrothermal stage [43]. Additionally, it is found that the O/C ratio is weakened by increasing the hydrothermal temperatures, leading to a downtrend in the density of carboxylic and carbonyl groups. Less metal precursor is thus impregnated in Co@HCIm-240 when compared to Co@HCIm-180.Second, the one-pot synthesis also provide suitable conditions for the in-situ reduction caused by the oxygenated functional groups in the hydrochar [44], which is already proven by the XRD and XPS results. This reduction may proceed prior to the intermolecular dehydration and aldol condensation during the hydrothermal process, so that the metallic nanoparticles tend to be preferentially in-situ dispersed in the hydrophobic core of the hydrochar. However, only amorphous metals are observed on the prepared catalysts of Co@HCOp, which is probably caused by the fact that the hydrothermal temperatures below 240\u00a0\u00b0C is insufficient for the crystallization of metal. This difference in the chemical states of cobalt provides a plausible explanation for the better catalytic performance on the hydrochar-supported catalysts prepared by the impregnation method (Entry 3) rather than the one-pot synthesis (Entry 6).With the best catalyst (i.e., Co@HCIm-180) in hand, we next optimized the reaction conditions to get the highest yield of target products; meanwhile, the reaction route of the reductive amination of benzaldehyde was also explored to investigate the influence of specific parameters. As is known to all, side reactions are one of the main reasons for carbon loss [25,45], as illustrated in Fig.\u00a08\n(a). In addition to benzylamine, the condensation product of N-benzylidenebenzylamine (i.e., 1c) is identified to be one of the intermediates for the reductive amination of benzaldehyde. The hydrogenation of N-benzylidenebenzylamine can easily produce the by-product of dibenzylamine (i.e., 1d), while the subsequent thermal cyclization to 2, 4, 5-triphenyl-4, 5-dihydro-1H-imidazole (i.e., 1e) has also been achieved. Interestingly, the hydrogenation of benzaldehyde into phenylcarbinol (i.e., 1b) is not detected in our catalyst system, which might be owing to the fact that the cobalt-based catalysts are inactive towards the hydrogenation of benzaldehyde [40].Following this route, the effect of reaction temperature is firstly studied under 2\u00a0MPa H2 and 5\u00a0mL of 7\u00a0mol\u00a0L\u22121\u00a0NH3 solution (in MeOH), and the results of yield are depicted in Fig.\u00a08(b). As can be seen, benzaldehyde can be completely converted even at room temperature, but no target product is observed. By increasing the temperature to the boiling point of the solvent or over it, the yield of 1a steadily grow up to 89.5% as the molecular reactions are accelerated at gas-liquid mixed state; moreover, higher temperatures also provide more heat to overcome the activation energy required by reductive amination. On another hand, the influence of reaction period exhibits the similar trend, where the zero hour can only catalyze a few of the reactant but the 4\u00a0h are sufficient enough to convert most of the benzaldehyde into benzylamine. It seems that the reaction temperature and period are the primary factors, as the influence of the NH3 concentration, the H2 pressure, and the amount of catalyst are relatively less important towards the yield of 1a, ranging from approximately 84.4%\u201393.7%.After developing the successful synthesis of primary amines by Co@HCIm-180, we hope to apply this method in the future industry. And here, we are delighted to show that the conversion of benzaldehyde to benzylamine under three gram-scale tests, including 0.25-g scale, 0.50-g scale, and 1-g scale, are well performed in this catalyst systems based on the optimal reaction conditions (i.e., 90\u00a0\u00b0C, 4\u00a0h, 7M NH3 solution, 2\u00a0MPa H2, and 10\u00a0mg catalyst). They are similar to those microgram-scale tests, and the excellent yields of 1a are obtained for all the tested scales, as depicted in Fig.\u00a0S5(a). Furthermore, we also evaluated the catalyst stability and recycle ability using the benchmark reaction under similar experimental conditions. As expected, in comparison to the previously reported catalysts, the hydrochar-supported catalyst can be recycled at least 5 times without the significant loss of activity, although the downtrend become more serious after the 5th cycle (Fig.\u00a0S5(b)). More importantly, owing to the magnetic properties offered by the metallic cobalt, these catalysts can be easily separated from the aqueous systems by a magnetic bar [10,29], which is considered as an advantage in industrial application.In summary, we have successfully prepared the hydrochar-supported catalysts through two sustainable routes, i.e., the impregnation method and the one-pot synthesis, by using glucose as carbon source, citric acids as inducer, and different metal salts as the catalytic actives. According to the structural analysis, differences between two types of catalysts were carefully revealed, which affects its catalytic performance to a large extent. The impregnation method at the atmospheric pressure might favor the electrostatic attraction with the outer hydration shell of the metal cations, leading to the formation of outer-sphere surface complexes with well-distributed metallic nanoparticles. By contrast, the one-pot synthesis might provide suitable conditions for the formation of inner-sphere surface complexes as the chemical adsorption of metal cations on hydrochar proceed prior to the intermolecular dehydration and aldol condensation. Unfortunately, as a result of the relatively lower carbonization temperature, its catalytic activity on the reductive amination was limited when compared to the former type of catalyst, which is reflected in two aspects: the inferior porous structures and the amorphous state of metal atoms. Subsequently, the catalyst prepared by the impregnation method was adopted to analyze the influence of reaction conditions, which indicates that the reaction temperature and period are the primary factors. And also, this protocol could exhibit a similar reactivity during the gram-scale and had a long lifecycle as it can be easily recycled and reused up to five times without the significant loss of catalytic activity and selectivity. By and large, the above findings confirm an available and sustainable preparation of simple but highly efficient catalysts on the production of functional amines, and future work will be directed towards the identification of catalytic mechanisms on various substances with different functional groups.J. G. L. and L. L. M supervised and designed the research. X. Z. Z. performed most of the experiments and wrote the original paper. J. G. L. reviewed and corrected the original manuscript. All authors discussed the results and assisted during manuscript preparation.Correspondence and requests for materials should be addressed to L. L. M or J. G.L.The authors declare no competing financial interests.This work was supported financially by the National Key R&D Program of China (2018YFB1501500), and National Natural Science Foundation of China (51976225).The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.gee.2022.01.012.", "descript": "\n                  Since the utilization of abundant biomass to develop advanced materials has become an utmost priority in recent years, we developed two sustainable routes (i.e., the impregnation method and the one-pot synthesis) to prepare the hydrochar-supported catalysts and tested its catalytic performance on the reductive amination. Several techniques, such as TEM, XRD and XPS, were adopted to characterize the structural and catalytic features of samples. Results indicated that the impregnation method favors the formation of outer-sphere surface complexes with porous structure as well as well-distributed metallic nanoparticles, while the one-pot synthesis tends to form the inner-sphere surface complexes with relatively smooth appearance and amorphous metals. This difference explains the better activity of catalysts prepared by the impregnation method which can selectively convert benzaldehyde to benzylamine with an excellent yield of 93.7% under the optimal reaction conditions; in contrast, the catalyst prepared by the one-pot synthesis only exhibits a low selectivity near to zero. Furthermore, the gram-scale test catalyzed by the same catalysts exhibits a similar yield of benzylamine in comparison to its smaller scale, which is comparable to the previously reported heterogeneous noble-based catalysts. More surprisingly, the prepared catalysts can be expediently recycled by a magnetic bar and remain the satisfying catalytic activity after reusing up to five times. In conclusion, these developed catalysts enable the synthesis of functional amines with excellent selectivity and carbon balance, proving cost-effective and sustainable access to the wide application of reductive amination.\n               "}
{"full_text": "Energetic crisis is a critical global issue, particularly for the development of hydrogen generation technologies, since molecular hydrogen has a relevant energy density but is also considered as an environmentally friendly fuel, with theoretically zero-emissions of greenhouse gases, allowing to also address the nowadays environmental sustainability concern [1]. In this context, ammonia borane (AB) is an excellent hydrogen storage molecule, being considered as solid-state hydrogen, mainly because of its high hydrogen gravimetric content (near 19.6\u00a0%) and good stability under normal conditions [2]. Hydrogen can be generated from this molecule by a hydrolytic reaction using a catalytic material to accelerate its cleavage. Nevertheless, most catalysts employed for efficient AB hydrolysis are based on noble metal nanoparticles such as Pt [3,4], Pd [5], Rh [6], and Ru [7]. Within this context, it is well-known that noble metals are non-abundant and expensive materials, which could limit their practical applications. In this sense, a facile approach to reduce the cost of catalytic materials consists in combining highly active noble metals with earth-abundant metals into nanostructured systems, such as alloyed bimetallic nanoparticles. Additionally, the incorporation of metals into nanostructured systems has promoted several improvements, especially in catalysis, due to synergistic enhancements. This has been observed, for instance, in CuAu nanoparticles for the hydroxylation of aromatic compounds [8] or CoPd nanoparticles for CO oxidation, which has also showed a composition-dependence selectivity [9].Therefore, it results important to select catalytically active earth-abundant metals for this purpose and avoid a dramatic decrease in the catalytic efficiency of the bimetallic material. An excellent candidate for this task is nickel since it has been demonstrated to have an active performance in hydrogen generation reactions [10,11]. Besides, nickel is the fourth most abundant transition metal in nature, after Fe, Ti, and Zr, and is much cheaper than noble metals such as platinum. The latter is the most active metal of its periodic group for hydrogen generation reactions among several catalytic processes, but it is also almost 1600 times more expensive than nickel. In this sense, the combination of both metals, nickel and platinum, into nanostructured materials is of interest since both elements have been demonstrated to be active towards hydrogen generation reactions and share similar properties for being in the same periodic group, enabling them to be easily alloyed [12,13]. Therefore, the development of bimetallic nanomaterials has undergone significant growth, with the appearance of different metallic combinations and compositions, in order to tune and improve the catalytic properties of the final material [14,15]. In this context, combining a catalytically active metal, like platinum, with earth-abundant metals, such as nickel, would primarily allow an overall material cost reduction. Furthermore, NiPt alloyed bimetallic systems tend to present synergic behaviors, especially for catalytic applications due to modifications to the electronic, structural or textural properties of the material [16,17], making bimetallic systems an attractive and versatile option to carry out catalytic processes.Additionally, the use of different supports is a simple and efficient strategy to improve the stability and reusability of catalytic materials. In this regard, biopolymers are highly abundant in nature, biodegradable, non-toxicity, and have a wide diversity of functional groups that are crucial for stabilizing nanostructured systems [18]. However, despite these promising properties, there is a lack of information on the implementation of biopolymers for the synthesis of alloyed bimetallic systems for hydrogen generation purposes. In this context, alginate constitute an interesting biopolymer, commonly extracted from brown algae, that displays outstanding interactions with divalent metal ions. An excellent example is the combination of alginate with calcium ions, which drives the facile formation of alginate hydrogel in different formats such as spherical beads [19\u201321]. Additionally, the use of hydrogel-based systems derived from biopolymers is appealing due to the low amounts of the biopolymer necessary to generate a 3D network upon swelling in large amounts of water (e.g., up to 90\u00a0% w/w of the final material) [22]. Due to the above, alginate-based materials are interesting candidates to assist the formation of bimetallic nanoparticles, acting as a support system for their formation, as well as for the conservation of the properties of the nanostructures for recyclability.With this in mind, the aim of this work was to establish: i) a simple method for the synthesis of catalytically active NiPt alloyed bimetallic nanoparticles on alginate hydrogel beads, and ii) the use of this hybrid system as a heterogeneous catalyst in the hydrolysis of AB for the efficient production of hydrogen. Herein, we wish to highlight bimetallic nanoparticles supported by non-toxic hydrogels, derived from renewable feedstocks, as an attractive option for sustainable technological applications such as catalytic hydrogen evolution.Sodium alginate (MW\u00a0=\u00a0380.000\u00a0g/mol, G:M\u00a0=\u00a025:75), nickel chloride (II) (NiCl2, 98\u00a0%, Sigma-Aldrich), potassium tetrachloroplatinate (K2PtCl4, 99.9\u00a0%, Sigma-Aldrich), sodium borohydride (NaBH4, 98\u00a0%, Merck), ammonia borane complex (NH3BH3, 97\u00a0%, Sigma-Aldrich), and calcium chloride (CaCl2, 99\u00a0%, Merck) were used as purchased with no further purification. Milli-Q water 18.2\u00a0M\u03a9\u00a0cm\u22121 was used for all experiments.Alginate beads were prepared following a previously reported method [23]. Briefly, 10\u00a0mL of a 3.0\u00a0%\u00a0w/v sodium alginate solution was slowly added dropwise employing a syringe pump at 0.3\u00a0mL/min provided with a G25 syringe needle on 90\u00a0mL of a CaCl2 5\u00a0%\u00a0w/v solution as a gelling medium. As the sodium alginate solution was added, the calcium solution was gently stirred in order to conserve the formed alginate beads. Once the alginate solution was completely added, the hydrogel beads were kept in the gelling medium without stirring for 1\u00a0h to allow the formation and proper maturation of the beads, which were subsequently washed water (3\u00a0\u00d7\u00a0100\u00a0mL) to remove the excess of calcium ions. Finally, so-obtained alginate hydrogel beads were stored in water until their use on-demand.Alginate beads loaded with mono- and bimetallic nanoparticles were synthesized by a two-stage procedure that consisted of the adsorption of metallic precursors on hydrogel beads, followed by the reduction into the corresponding nanoparticles. The first stage was carried out by immersing 2\u00a0g of wet alginate hydrogel beads into 50\u00a0mL of water, in which aliquots of the nickel and platinum solutions, 74.8\u00a0mM and 24.9\u00a0mM respectively, were added (aliquot amounts are specified in Table S1) and kept under continuous stirring for 48\u00a0h. Once the metal adsorption stage was completed, the beads were washed with water (3\u00a0\u00d7\u00a0100\u00a0mL) to remove the excess of ions.The reduction step was subsequently performed by immersing 500\u00a0mg of hydrogel beads loaded with metallic ions into 10\u00a0mL of water in a sealed Schlenk tube, which was purged with argon for 10\u00a0min. After this time, 500\u00a0\u03bcL of a NaBH4 60\u00a0mg/mL solution was slowly added to the reaction vessel, which was placed in an oil bath at 60\u00a0\u00b0C and stirred at 700\u00a0rpm for 2\u00a0h. After this time, the beads were removed, washed with abundant water and stored prior to characterization and catalysis testing.Hydrogen generation reactions were performed using a sealed Schlenk flask connected to a gas burette system. In a typical experiment, 80\u00a0mg of catalyst were added to a sealed Schlenk flask containing 2\u00a0mL of water. Then, 400\u00a0\u03bcL of a 10\u00a0mg/mL AB aqueous solution was injected to start the reaction. The hydrogen generation kinetics were monitored by the volume displacement observed in the gas burette. All experiments were performed at room temperature unless otherwise stated.Turn over frequency (TOF) values were calculated using Eq. (1), considering the amount of hydrogen produced, the metal content incorporated to the reaction and the time t at which each catalyzed reaction reached a plateau on its hydrogen evolution profile.\n\n(1)\n\nTOF\n=\n\n\nmmol\n\n\nH\n2\n\n\n\n\nmmol\n\ncat\n\nx\n\nt\n\n\n\n\n\nInfrared spectroscopy was performed using a PerkinElmer UATR spectrometer by directly inserting the sample in the ATR probe. Spectra were recorded between 4000 and 400\u00a0cm\u22121, with a resolution of 1\u00a0cm\u22121. Thermogravimetric analyses (TGA) were performed in a Mettler thermogravimetric analyzer with a nitrogen flux of 20\u00a0mL/min. Rheologic measurements were performed in a discovery hybrid rheometer HR 20 with a 40\u00a0mm stain plate, by homogeneously dispersing 500\u00a0mg of alginate beads on the testing plate. The rheological experiments were conducted at room temperature and a constant shear frequency of 10\u00a0s\u22121, using an oscillatory strain window between 0.001 and 100\u00a0%. The metallic content in the beads was determined with an Agilent atomic absorption spectrophotometer (AAS) GTA 120 and the data were processed using a SpectrAA 240Z. TEM images were obtained with a JEOL JEM 1010 microscope, with a resolution of 0.4\u00a0nm, placing the samples on copper grids. TEM images were processed using ImageJ software. Finally, XPS analyses were measured in a \u201cHippolyta\u201d Devi-sim (SPECS) near ambient pressure X-ray photoelectron spectrometer under ultra-high vacuum equipped with a PHOIBOS NAP-150 analyzer and a 2D-DLD detector, in which the samples were irradiated with a monochromatized Al source (K\u03b1 h\u03bd\u00a0=\u00a01486.7\u00a0eV) and a flood gun for charge compensation. Data treatment was carried out using XPSPEAK software version 4.1, adjusting a Shirley baseline for each region and referencing the signals with respect to the C 1s peak at 284.8\u00a0eV.Synthesis of the hydrogel beads loaded with mono- and bimetallic NiPt nanoparticles was carried out employing an easy procedure that allowed the obtention of the nanocomposites in a controllable and reproducible way (Fig. 1\n). Firstly, the adsorption step was performed taking advantage of the good interaction of alginate chains with divalent ions [18], which favors the adsorption of Ni (II) and Pt (II) ions. Subsequently, the reducing step was performed employing a strong reducing environment, with the addition of a concentrated solution of NaBH4, causing a notable color change in the material going from pale brown to black beads, indicating the formation of the desired nanoparticles. The conditions were selected due to the difficulty of reducing nickel ions into a zero-valence state, because of its negative reducing potential (E0\u00a0=\u00a0\u22120.257\u00a0V). [24] In this case, based on the Marcus theory and taking into consideration the mismatch between the reduction potential of Pt (II) and Ni (II), a strongly reducing environment should be required to kinetically favor the formation of alloyed bimetallic nanoparticles [25,26]. With this synthetic procedure we expect that the nanoparticles are formed with an adequate availability on the outer surface of the bead to work as a supported catalyst. Nevertheless, some nanoparticles may also diffuse to the inner part of the beads and contribute to the potential catalysis, at least to some extent.The presence of the nanoparticles was confirmed by TEM analysis (Fig. 2\n), in which the size distribution of the synthesized nanoparticles was between 2.72\u00a0\u00b1\u00a00.69\u00a0nm and 5.24\u00a0\u00b1\u00a01.14\u00a0nm. These results confirmed that the proposed reaction protocol allowed the obtention of nanoparticles with narrow size distributions. The metallic content in the material was calculated by atomic absorption spectroscopy (AAS), revealing values around 1.87\u00a0\u00b1\u00a00.03\u00a0\u03bcmol of metal per gram of hydrogel, values that are also in good agreement with the initial targeted compositions for each system (Fig. 3\n).Subsequently, hydrogels loaded with Ni (II), Pt (II) and the corresponding mono- and bimetallic nanoparticles were characterized by FT-IR, in order to gain an insight into the role that alginate functional groups plays over the adsorption and stabilization of the metallic ions and nanoparticles (Fig. 4\n). The spectra showed the main signals attributed to the alginate backbone, that is the most abundant part of the nanocomposite in comparison with the metallic load of the material. Bands centered at 3320, 2920, 1598, 1412 and 1019\u00a0cm\u22121 correspond to the stretching vibrations of OH, CH, asymmetric and symmetric CO stretching, and to the CO vibrations of the pyranose rings. After the reduction steps, the main signals corresponding to the alginate chains remained in the spectra of all materials, indicating that alginate structure remains unaltered after the reductive process. Nevertheless, there was an increment in the intensity of the signals of OH and CO bonds, which was more noticeable as they were proportionally compared with the intensity of the CH band at 2900\u00a0cm\u22121, indicating the particular importance of these groups during the adsorption of metallic ions and stabilization of metallic nanoparticles [27,28]. The spectral changes observed in the FT-IR suggest that functional groups of alginate were adequately preserved during the protocol, maintaining a number of hydroxyl and carboxylic groups still available to interact with metallic ions and adsorb them onto the hydrogels, which represents a key step in this protocol.Considering that the functional groups of alginates interact in a different way with metallic ions and nanoparticles, they may also influence other properties of the materials such as the mechanical properties of the beads. For this reason, samples of the material at the different stages of the synthetic process were analyzed by oscillatory rheometry. Recently, Posbeyikian et al\n[29] applied a detailed rheometric method to deeply understand the crosslinking process during the formation of alginate beads. Following that method, we carried out rheological analysis on pristine synthesized alginate beads, beads loaded with adsorbed metallic ions, and beads loaded with reduced metallic nanoparticles (Fig. 5\n). Rheological profiles showed a dominant solid-like behavior during a large part of the analysis, with a storage modulus (G') almost 10 times higher than the loss modulus (G\") in all cases. Tan (\u03b4) plots of the materials showed that all materials behave similarly in terms of damping properties, maintaining a phase angle close to zero, typical of solid-liked materials, which demonstrates a notable increase at higher oscillating stress values, and suggesting the presence of a yield stress, accusing the start of the viscous region. However, there is a notable decrease in the plateau of G' (material strength) once the reduction reaction was performed in comparison to the G' values obtained for the pristine alginate beads or beads with adsorbed ions. In this regard, a reduction of the strength of the material could be related to a decrease in the interaction strength between the polymer chains that conform the 3D polymer network. This phenomenon can be related to the stabilization of the formed nanoparticles by the alginate functional groups, since the stabilization process requires their interactions with the surface of the metallic nanoparticles in order the avoid agglomeration or leaching.Further characterizations indicate that the thermal stability of the hydrogel beads loaded with mono- and bimetallic nanoparticles was limited by the evaporation of the water that forms part of the 3D hydrogel network (Fig. 6\n). This was distinguished by a significant weight loss stage centered approximately at 100\u00a0\u00b0C, confirming that the material is composed of a high amount of water, followed by a moderate second weight loss stage approximately at 200\u00a0\u00b0C corresponding to the degradation of the remanent polymer content. A similar thermal\u2013behavior has been observed in previous reports, where a first degradation step taking place at temperatures below 150\u00a0\u00b0C has been previously assigned to the evaporation of the water content present in alginate beads [30,31]. Despite the rapid weight loss experienced by the materials, these can still be useful to perform reactions employing water as a reaction medium and near room temperature, which is highly desirable for sustainable applications.After the complete characterization of the materials, they were tested as catalysts for the hydrolysis of AB as a model reaction for the generation of hydrogen (Fig. 7a). It is important to highlight that hydrogen generation from the hydrolytic reaction was not detected at room temperature in the absence of alginate beads loaded with metallic nanoparticles (Fig. S2). The catalytic properties were determined using a metal loading near 0.1\u00a0% with respect to the initial AB amount in each run. In order to compare the released equivalents of hydrogen against the initial equivalents of AB, the obtained data were normalized by the initial amount of AB placed into the reactor. The hydrogen generation profiles showed that the reaction started immediately after the AB addition, without any induction period, suggesting a fast activity of the evaluated catalysts (Fig. 7b). Furthermore, almost all prepared nanocatalysts were highly active, reaching conversion values around 3.0 equivalents at room temperature. This value corresponds to a theoretical complete hydrolysis of 1.0 equivalent of AB, except for the case of alginate beads loaded with NiPt 3:1 and pure monometallic Ni nanoparticles. The latter being the least active in the series of evaluated catalysts.The obtained data were fitted using the well-known pseudo-order models (zero, first and second order, Fig. S3) aiming to have better insight into the kinetic mechanism that governs the reaction. The best fit was achieved using the pseudo-zero order model with a R2 value near 0.996, indicating a zero order of AB in the rate law (Eq. (2)).\n\n(2)\n\n\u2212\n\n\nd\n\nAB\n\n\ndt\n\n=\n\nk\napp\n\n\u2192\n\n\nAB\n\nt\n\n=\n\n\nAB\n\n0\n\n\u2212\n\nk\napp\n\nt\n\n\nin which, [AB]\n\nt,\n[AB]\n\n0\n, k\n\napp\n and t are the time-dependent and initial concentrations of ammonia borane, apparent kinetic rate constant, and reaction time, respectively. To confirm the zero order of AB in the hydrogen generation reaction, different initial concentrations of AB were evaluated. However, the kinetic profiles (Fig. 7c) showed different reaction rates as the initial concentration varied from 39 to 56\u00a0mM, being opposite to the expected tendency from the previous model, since a change in the initial concentration should not modify the reaction rate. In this regard, Figen et al\n[32] previously reviewed the most common models used to fit the kinetic data of hydrogen generation from AB hydrolysis, in which zero order is the most used. Nevertheless, some articles highlight that the catalyzed reaction has little dependence on the AB concentration, disagreeing from the zero order kinetics [33].With these considerations in mind, the Langmuir-Hinshelwood model was applied, taking into account the heterogeneous nature of the catalyst [34]. This model is commonly used for bimolecular reactions that take place at the surface of catalytic materials, in which both reactants are adsorbed on neighboring sites of the catalyst to subsequently accomplish a surface chemical reaction between them, which is usually the rate limiting step of the reaction. The model ends with the desorption of the generated products [35]. The general Langmuir-Hinshelwood expression that explains this phenomenon is given by Eq. (3):\n\n(3)\n\n\u2212\n\n\nd\n\nAB\n\n\ndt\n\n=\n\n\n\nk\nr\n\nK\n\nAB\n\n\n\n1\n+\nK\n\nAB\n\n\n\n\n\nin which k\n\nr\n and K are the superficial chemical reaction and the equilibrium adsorption constant of AB onto the catalyst, respectively [36]. This equation can be integrated as following (Eq. (4)):\n\n(4)\n\nLn\n\n\n\n\nAB\n\n0\n\n\nAB\n\n\n\n+\nK\n\n\n\n\nAB\n\n0\n\n\u2212\n\nAB\n\n\n\n=\n\nk\nr\n\nKt\n\n\n\nHowever, the integrated Langmuir-Hinshelwood equation was difficult to use, due its non-linear nature. For that reason, instead of applying the integrated method, Eq. (3) was linearized into the following expression:\n\n(5)\n\n\n1\nr\n\n=\n\n1\n\nk\nr\n\n\n+\n\n1\n\n\nk\nr\n\nK\n\nAB\n\n\n\n\n\nin which r represents the rate of the reaction. The data adequately fit into Eq. (5), in which k\n\nr\n and K could be extracted from the reciprocal values of the slope and the intercept of the fitted Langmuir-Hinshelwood curves, respectively. Results are summarized in Table 1\n.The obtained values revealed notable changes for both k\n\nr\n and K parameters as the metal composition of nanoentities was modified. Firstly, K[AB] values were calculated considering the initial concentration of AB loaded in the reaction vessel. This was carried out considering the importance of K[AB] as a determining factor, because of its influence in the establishment of limiting conditions for the model. Mathematically, when K[AB]\u00a0>\u00a01 the general Langmuir-Hinshelwood equation can be approximated to a rate law equation only affected by the k\n\nr\n of the reaction. On the other hand, when K[AB]\u00a0<\u00a01, the rate law can be expressed as a first order expression on AB concentration. In this regard, all K[AB] values were in better agreement with the first limit case, which, from a mechanistic point of view, suggests an initial fast adsorption of AB molecules on the surface of the material, closely related to the Langmuir-Hinshelwood mechanism. These results indicate that the rate limiting step of the hydrogen generation is the surface reaction that takes place on the catalyst, characterized by k\n\nr\n, which shows a notable increase by the formation of the alloyed systems, a phenomenon related to a synergistic effect promoted by the presence of both metals forming part of the catalysts.Further testing of the materials on recyclability and leaching studies were performed, which are especially important to evaluate the efficiency of the bio-based support to maintain the catalytic properties of the nanoparticles and to avoid their desorption into the reaction medium. To accomplish this, the reaction was performed in multiple catalytic cycles using alginate beads containing NiPt 1:3 (Fig. 8a). The reaction profiles showed minor differences throughout the evaluated cycles, with a slight decrease on the kr which was below 10\u00a0% during the evaluated cycles. Additionally, leaching of catalytically active nanoparticles form the hydrogel was also evaluated (Fig. 8b). Initially, a cycle of the reaction was performed using hydrogels loaded with NiPt 1:3 nanoparticles as an example. Then, once the reaction was completed, the beads were removed from the medium and the reaction was performed only with the supernatant of the previous reaction. This demonstrated that the supernatant of the first reaction did not show a water displacement in the gas burette, suggesting that there is no presence of an active material able to perform the hydrogen generation reaction, and that the nanoparticles are well retained by the alginate beads sustaining its use as an adequate support for these systems, facilitating the easy recovery of the material from the reaction medium, and its subsequent utilization without further purification processes, in contrast with other colloidal dispersed systems [37].Additionally, temperature effect on the reaction was evaluated for beads loaded with NiPt 1:3 nanoparticles (Fig. 8c). The results confirmed a gradual increase in the reaction rate with the temperature. This was in good agreement with Arrhenius model (Fig. 8d), in which activation energy of the reaction using this catalyst was extracted with a value of 50.24\u00a0kJ/mol, which is comparable to the value reported for other monometallic systems such as Pd [35] and Pt [38] catalysts, with Ea values of 41.50 and 50.35\u00a0kJ/mol, respectively.Then, TOF values were calculated to provide a more accurate comparison between the obtained catalysts and other systems already reported in the literature (Fig. 9\n). The hydrogels containing alloyed systems presented higher TOFs than their monometallic counterparts, showing a volcano-shaped trend over the complete composition range studied. These results are in good agreement with the synergic behavior for the catalytic hydrogen generation [10,39,40].In order to gain a better insight into the catalytic synergy obtained by the combination of both metals, XPS analysis was carried out on the NiPt 1:3 system, since it had the best performance among the evaluated materials. Fig. 10\na-c show the XPS spectra of the regions corresponding to carbon 1s, platinum 4f and nickel 2p regions. Fig. 10a shows carbon 1s region of several peaks at 284.8, 286.6, 287.9 and 289.0\u00a0eV regarding to several carbon atoms in different environments with increasing electronegative surroundings, related to carbon atoms in alginate chains bonded to hydroxyl groups (COH) or forming part of carboxylate moieties (OCO). Fig. 10b reveals the presence of two pairs of signals, with its 4f7/2 peaks at 70.7 and 72.8\u00a0eV, each with its characteristic spin-orbit split of 3.3\u00a0eV. This result suggests the presence of platinum atoms in two oxidation states at the surface of the material, mostly in the form of Pt(0) and Pt(II), being the latter related to PtO moieties. Finally, Fig. 10c illustrates the signals related to the nickel 2p region, with a Ni 2p3/2 at 854.1\u00a0eV, with a spin-orbit split of 17.6\u00a0eV, and a shake-up satellite peak at 871.7\u00a0eV. The observed binding energy probably indicate the presence of nickel atoms in a high oxidation state, which could be attributed to the presence of Ni(OH)2 species at the surface of the nanoparticles due to the observed chemical shift. Previously, Fu et al\n[10] evaluated the activity of NiPt bimetallic nanoparticles increasing the alkalinity of the reaction medium, in which the hydrolytic reaction was enhanced varying the pH of the media. Furthermore, Zhao et al\n[41] previously modeled water adsorption on nickel oxide species at the surfaces, evidencing a dissociative pathway for its adsorption, accompanied by an notable exothermic \u0394H. In this regard, a plausible mechanism to explain the surface enhancement of AB hydrolysis could be related to the presence of preformed hydroxylated species at the surface of the catalyst, preferentially on nickel atoms as suggested by XPS analysis. This would facilitate the surface cleavage of AB molecules that are preferentially adsorbed on platinum atoms, a phenomenon that might explain the greater adsorption constant calculated by the Langmuir-Hinshelwood model in platinum-rich materials.At this point, the TOF value calculated for alginate beads containing NiPt 1:3 catalyst was compared other supported and also unsupported catalysts from the literature (Table 2\n). Our system showed a competitive behavior against other similar bimetallic systems based on the combination of noble Pt atoms with earth abundant metals such as Ni or Co. The beads containing NiPt 1:3 bimetallic nanoparticles showed a TOF above some non-supported systems such as Pt0.65Ni0.35 or Pt0.01Ni0.99. These systems tend to present higher activity in comparison with supported materials since they behave more similar to homogeneous catalysts, demonstrating a better performance in comparison to unsupported bimetallic nanoparticles formed by only noble metals such as Ag@Pd nanoparticles [42]. Additionally, the biobased support afforded a good dispersity of the catalytic nanoparticles, exhibiting a TOF value similar to that obtained for CoPt bimetallic nanoparticles on nanoporous graphene sheets [43], a system characterized by its wide surface area.Biohydrogel beads loaded with NiPt bimetallic nanoparticles were successfully synthesized using an easy and cost-effective method, achieving homogeneously distributed mono- and bimetallic nanoparticles. These materials constitute a promising bio-based candidate for hydrogen generation reactions, being highly active for the hydrolysis of AB, reaching a quantitative hydrogen generation employing a catalyst concentration near to 0.1\u00a0% with respect to the AB loading. The reaction mechanism consists of an initial fast adsorption of reactants at the surface of the nanoparticles, followed by a surface reaction which is enhanced by the combination of nickel and platinum atoms. This is attributed to the synergic behavior of NiPt nanoalloys, which affords a TOF value of 84.1\u00a0min\u22121 for alginate beads containing NiPt 1:3 bimetallic nanoparticles. Finally, the synthesized bio-based materials are also easily recovered thanks to the properties of the biohydrogel support.\nOscar Ram\u00edrez: Investigation, Conceptualization, Methodology, Formal analysis, Writing \u2013 original draft. Sebastian Bonardd: Supervision, Methodology, Writing \u2013 review & editing. C\u00e9sar Sald\u00edas: Writing \u2013 review & editing. Macarena Kroff: Resources, Formal analysis. James N. O'Shea: Resources, Formal analysis. David D\u00edaz D\u00edaz: Supervision, Project administration, Writing \u2013 review & editing, Funding acquisition. Angel Leiva: Supervision, Project administration, Writing \u2013 review & editing, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.O. Ram\u00edrez thanks Beca doctorado nacional\nANID 21191002. S. Bonardd thanks MINECO for a Juan de la Cierva \u2013 Formaci\u00f3n contract FJC2019-039515-I. C. Sald\u00edas thanks Fondecyt Project 1211022. M. Kroff thanks to Beca doctorado nacional\nANID 21180627. J. N. O'Shea acknowledges and thanks Innovate UK through the Energy Research Accelerator, the Engineering and Physical Sciences Research Council (EPSRC), and the University of Nottingham Propulsion Futures Beacon for funding. A. Leiva thanks to FONDECYT\n1211124 and FONDAP\n15110019 projects for the financial support of the research. D. D. D\u00edaz thanks financial support from the Spanish Government for the Senior Beatriz Galindo Award (BEAGAL18/00166) and the project PID2019-105391GB-C21/AEI/10.13039/501100011033. The authors thank NANOtec, INTech, Cabildo de Tenerife and ULL for laboratory facilities.\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.ijbiomac.2022.11.106.", "descript": "\n                  Alginate hydrogel beads were loaded with bimetallic NiPt nanoparticles by in situ reduction of the respective polymer matrix containing precursor metallic ions using a NaBH4 aqueous solution. The alginate hydrogel beads loaded with NiPt nanoparticles were characterized by TEM, AAS, FT-IR, TGA, XPS, and oscillatory rheometry. The prepared hybrid hydrogels were proven to be effective as catalytic materials for the hydrolysis of ammonia borane (AB) for quantitative hydrogen generation using catalytic loadings of 0.1\u00a0mol%. In addition, the reaction mechanism of the hydrolytic reaction using NiPt loaded alginate hydrogel beads was determined by Langmuir-Hinshelwood model. The experimental results showed that the reaction mechanism consisted of an initial fast adsorption of reactants at the surface of the nanoparticles, followed by a rate-limiting surface reaction. The NiPt nanoalloys exhibited an enhanced behavior for hydrogen generation with a maximum TOF of 84.1\u00a0min\u22121, almost 71\u00a0% higher compared to monometallic platinum atoms, and likely related to a synergistic interaction between both metals. Finally, the hydrogel matrix enabled the material to be easily recovered from the reaction medium and reused in further catalytic cycles without desorption of active nanoparticles from the material.\n               "}
{"full_text": "The advent of industrialization over the last 200 years gives rise to many anthropogenic activities that lead to the destruction of the natural environment. Increased levels of carbon dioxide (CO2) in the atmosphere caused by the consumption of fossil fuels play a leading role in the greenhouse effect and subsequent global warming. Over the last few decades, extensive research has been carried out to explore alternative energy carriers such as hydrogen that can alleviate the role of CO2 in the atmosphere [1\u20136]. Different strategies have been explored to produce H2 in a sustainable and renewable way such as solar water splitting, but the storage and transportation of hydrogen results in additional costs for the commercialization of this technology. As a promising alternative, the future hydrogen economy can be enhanced by the utilization of ammonia (NH3) as a carrier due to its high hydrogen content (1.4 % greater mass fraction than methanol) and volumetric energy density as compared to liquid hydrogen [7]. Also, ammonia is an important commodity in the food and energy supply chains with high annual consumption of 200 million tons. In addition to the production of fertilizers, ammonia is also a feedstock in the production of explosives, plastics, resins, synthetic fibers, and refrigerants (Fig.\u00a01\n).In nature, nitrogen fixation at ambient conditions is carried out by the nitrogenase enzyme present in microorganisms, comprised of a molybdenum-iron (MoFe) protein (Equation 1) [8\u201310]. This process is carried out by a proton-coupled electron transfer (PCET) reaction with a significant input of energy by the hydrolysis of adenosine triphosphate (ATP) into adenosine diphosphate (ADP) and inorganic phosphate (Pi) with intrinsic hydrogen production. In industry, the Haber\u2013Bosch (HB) process utilizes iron (Fe) and ruthenium (Ru) based catalysts to reduce nitrogen with hydrogen (from steam reforming) at high temperature (400\u2013500\u00a0\u00b0C) and pressure (130\u2013150\u00a0bar) to produce ammonia [11]. The high temperature and pressure are required to overcome thermodynamic constraints related to the bond dissociation energy of the nitrogen molecule (911\u00a0kJ\u00a0mol\u22121) and for an increased rate of production. Moreover, the production, purification, compression, and transportation of reactant gases add to the high cost and energy consumption of this process [4,12]. Therefore, there is a strong demand to explore less energy-intensive, eco-friendly, and economically viable alternative strategies for facilitated ammonia production under ambient conditions.\n\n(1)\nN2 + 8\u00a0H+ + 8 e\u2212 + 16 ATP \u2192 2 NH3 + H2 + 16 ADP + 16 Pi\n\n\nFor green and sustainable ammonia production, different strategies have been employed in the recent past including, molecular catalysis, electrochemical and photo-electrochemical reduction of nitrogen to produce ammonia [9,14\u201317]. Inspired by the nitrogenase enzymes, different transition-metal-based molecular catalysts have been developed and assessed for ammonia production. The mechanism of nitrogen reduction catalyzed by molybdenum and iron-based phosphine and trisamidoamine complexes has been studied [18\u201322]. The mechanism's elucidation, as well as the catalyst's stability and regeneration, are issues that need to be addressed. Photo-electrochemical nitrogen reduction and nitrogen oxidation into ammonia and nitrates, respectively, have also been explored using the visible light spectrum. Photosensitive materials of various sorts have been used, primarily titania (TiO2), but also transition metal-based semiconductors, noble metals, chalcogenides-supported biomaterials, and polymeric materials [20,23\u201326]. The use of this method on a wide scale has been limited due to the limitations of different photocatalytic materials such as low solar spectrum absorption and significant charge recombination.Electrochemical reduction of nitrogen to ammonia (eNRR) at ambient conditions is recognized as one of the most promising approaches toward a CO2-free product due to the abundance of carbon-free reactants (nitrogen and water) and the benefits of heterogeneous catalysis along with the use of clean and sustainable energy sources (Equations 2-4) [27].\n\n(2)\nCathode: N2 + 6H2O + 6e\n\u2212 \u2192 2NH3 + 6OH\u2212 \u22120.77\u00a0V vs. SHE\n\n\n\n\n(3)\nAnode: 2OH\u2212\u00a0+\u00a0H2 \u2192 2H2O\u00a0+\u00a02e\n\u2212 +0.82\u00a0V vs. SHE\n\n\n\n\n(4)\nOverall: N2\u00a0+\u00a03H2 \u2192 2NH3 +0.06\u00a0V\n\n\nModern research on eNRR opens new possibilities for reducing copious atmospheric nitrogen to ammonia using a range of efficient electrocatalysts (Table 1\n) [28\u201331]. Electrochemical ammonia production may be implemented as a small-scale decentralized system, which promises to be a breakthrough for the industry of underdeveloped countries. The facile transportation and storage of liquid ammonia further make it a promising low-cost technology [32,33].Several factors mitigate against an efficient eNRR. The availability of dinitrogen at the electrode is limited by its low solubility in aqueous solutions. Without a suitable catalyst, the strong non-polar bonding (E\nd\u00a0=\u00a0911\u00a0kJ\u00a0mol\u22121) and high activation barrier for dissociation (941\u00a0kJ\u00a0mol\u22121) of N\u2261N result in a low rate of ammonia production, and competition from the electrochemical hydrogen evolution reaction (HER, Equation 5) lead to a low faradaic efficiency (FE) [34]. Therefore, the development of an efficient and selective catalyst that can promote the reduction of nitrogen to ammonia while suppressing the HER is of great importance to the modern world.\n\n(5)\n2H2O + 2e \u2192 H2 + 2 OH\u2212\n\n\n\nTo attain this purpose, rational catalyst design is essential and must address both the intrinsic and extrinsic catalytic activity of the electrocatalyst. The intrinsic catalytic activity of an electrocatalyst directly affects the overall performance of the reaction. Electronic structures dictate the intrinsic catalytic properties of different metal and non-metal surfaces that can be further altered by doping of heteroatoms, alloying with other metals, core-shell structures formation, facet engineering, creation of defects, and intercalation of different anions or cations [35\u201339]. Extrinsic catalytic properties are determined by structural and morphological parameters that can affect the kinetics of the reaction. The shape, size, and morphology of the catalytic material dictate its active site density, which should be finely engineered to enhance overall catalytic performance (Fig.\u00a02\n). Moreover, rational electrode design could also improve the mass transport and charge transfer of the reaction by an enhancement in the surface area, the development of porous structures, and the use of different support materials [16,35].Theoretical insights into the mechanistic approach provide a framework for the rational design of a catalyst with effective catalytic activity towards eNRR. A systematic account of the structure-to-performance relationship of a catalyst for eNRR is presented in this review. This review summarizes the latest work in rational catalyst design strategies and mechanistic analyses for eNRR in ambient settings by taking theoretical to experimental approaches. Systematic illustrations of recent theoretical and experimental data on catalyst design will provide a full account of important challenges and constraints of this technique, as well as the ability to envision many potential future pathways. A framework that is supported by theoretical studies that have led to the optimal selection and development of eNRR catalysts in the recent past will be described. Following a discussion of recently-developed electrocatalysts for eNRR in ambient conditions, we outline various recently-published experimental methodologies to improve the intrinsic and extrinsic catalytic activity of advanced electrocatalysts for eNRR. Finally, we discuss some key findings and projections for this developing field.Electrochemical processes are generally driven by reactant adsorption, followed by activation, polarization, and transformation of reactants into products via various PCET reactions. Theoretical and experimental elucidation of several mechanistic paths for eNRR is first discussed in this context, as well as the analysis of thermodynamic and kinetic variables during the process.Molecular nitrogen is one of the most thermodynamically stable and inert molecules attributed to the short bond length (109.67 pm) and high dissociation energy (911\u00a0kJ\u00a0mol\u22121) of the triple bond. The large energy gap (10.82\u00a0eV) between the highest occupied and lowest unoccupied molecular orbitals (HOMO, LUMO) of nitrogen with its low polarizability, impedes electron transfer and reaction kinetics for ammonia production. Also, thermodynamic constraints related to various intermediates (NH, NH2, NH3) of this process add to the challenge. This process under ambient conditions is hampered by the competitive hydrogen evolution reaction. The affinity of available catalysts toward the hydrogen atom is greater than the non-polarizable nitrogen molecule. Due to the high redox potentials of intermediates and high activation energy requirement, this process is energy-intensive. However, the total Gibbs energy for ammonia formation favors the eNRR over hydrogen evolution. Also, an intrinsic challenge related to the low solubility of nitrogen (0.66\u00a0mmol\u00a0L\u22121) and high concentration of water (55\u00a0mol\u00a0L\u22121) in aqueous media limits the interaction of reactant nitrogen molecules with the active surface of the catalyst resulting in the low selectivity for ammonia formation. In this regard, active and selective catalyst design is a priority to overcome the above-mentioned challenges.Mechanistic investigations of the chemical process are highly desirable to develop an active selective and stable catalyst. Electrochemical NRR is governed by several PCET reactions and can be explained by different pathways. Firstly, the adsorption of nitrogen molecules or adatoms on the surface of the catalyst is the first step in the process that is followed by the PCET reaction, and after several PCET reactions, the desorption of the product takes place.Electrochemical NRR is generally believed to proceed via two types of reaction mechanisms, i.e., associative, and dissociative mechanisms (Fig.\u00a03\n) [40\u201342]. In the associative mechanism, the hydrogenation of adsorbed nitrogen molecules on the surface of the electrocatalyst takes place with two nitrogen atoms bound together. After two successive hydrogenations on the distal nitrogen atom, the final cleavage of the N\u2261N bond takes place resulting in the release of the first ammonia molecule (distal pathway). The second ammonia molecule is released after successive hydrogenations of the second adsorbed nitrogen atom. Moreover, successive alternating hydrogenations could occur on each nitrogen atom resulting in the hydrogenation of both N-atoms before N\u2261N bond dissociation (alternating pathway). Different types of adsorption symmetry of nitrogen molecules on the surface of the active site decide the pathway for the associative mechanism. If the adsorption of a nitrogen molecule is end-on, then the associative distal pathway is possible, and if the adsorption is side-on, then the alternating pathway is possible. For the dissociative mechanism, N\u2261N bond dissociation, a step that requires a large amount of energy, takes place before the hydrogenation of the nitrogen atom.For catalysis by transition metal nitrides, ammonia is proposed to form by hydrogenation of a nitrogen atom in the transition metal nitride structure; the Mars\u2013van Krevelen mechanism (MVK) [43,44]. This results in the creation of a nitrogen vacancy in the structure that is then occupied by N from the applied nitrogen gas. Successive hydrogenations result in the formation of ammonia. Further exploration with experimental evidence is important for an improved understanding of this process particularly the rate of vacancy creation and consumption.To achieve the above-mentioned mechanistic pathways, an enhanced interaction of nitrogen with an active site is required. However, despite the weak interaction between nitrogen and noble metal catalysts (Pd, Au) at low potentials, reasonable performance is seen with these catalysts. For these catalysts, surface hydrogenation is considered an important factor in the formation of ammonia [45]. Density functional theory (DFT) calculations revealed that the adsorption of nitrogen is outcompeted by the reduction of proton (H+) in the first step providing high coverage of adsorbed hydrogen (\u2217H) (Fig.\u00a04\n). Nitrogen can be activated by this hydrogenated surface to form ammonia at low overpotentials. A synergistic role of surface hydrogenation and the activation energy on the surface of the catalyst has been elucidated and the reduction of H+ is confirmed as the potential-determining step.Mechanistic attributes of the eNRR process can be better understood by using in-situ and operando characterization tools. Yao et\u00a0al. recently conducted a spectroscopic study of eNRR on gold (Au) and platinum (Pt) nano surfaces and proposed an associative mechanism for eNRR on Au thin films using results from surface-enhanced infrared absorption spectroscopy (SEIRAS) [46]. N2H species were detected by SEIRAS at potentials below 0\u00a0V vs. RHE which is a clear indication of the associative pathway for eNRR. Moreover, no intermediate was detected for Pt under the same eNRR conditions that demonstrated the highest activity of Pt as a HER catalyst. Similarly, rhodium (Rh) surfaces were also explored as an eNRR catalyst by SEIRAS, and a new reaction mechanism was proposed [47]. Firstly, the electrochemical formation of N2H2 by a two-electron pathway was observed and then the formation of ammonia after successive hydrogenation was completed. The IR bands for the hydroxyl group are observed at 3250\u00a0cm\u22121 and 1612\u00a0cm\u22121 which can be attributed to the 1st layer of water (Fig.\u00a05\na). At potentials from 0.2 to \u22120.4\u00a0V vs. RHE, the N=N stretching band at 1997\u20132036\u00a0cm\u22121 is allotted to adsorbed N2H\nx\n (0\u2264 x\u00a0\u2264\u00a02). Furthermore, surface hydrogenation, which varies with applied potential, was also confirmed by a weak band at 1865\u00a0cm\u22121 due to adsorbed hydrogen atoms. Lai et\u00a0al. reported superior eNRR performance for rhenium sulfide (ReS2) when doped with Fe on the surface of N-doped carbon which was verified by in-situ attenuated total reflectance infrared (ATR-IR) analysis (Fig.\u00a05b) [48]. A positive shift in the IR bands for O\u2013H stretching was observed from 0 to \u22120.3\u00a0V vs. RHE which was ascribed to the change in the adsorption configuration of water. Furthermore, the stability of the catalyst was also corroborated by SEIRAS spectra at \u22120.2\u00a0V vs. RHE with similar absorption results.Fu et\u00a0al. demonstrated the enhancement in the selectivity of this process by a dual atom catalytic system [49]. Re2MnS6 nanosheets were developed and the mechanism was evaluated for eNRR by in-situ Raman analysis. The presence of a Raman mode at 658\u00a0cm\u22121 under applied voltage in the presence of N2 confirmed the formation of NH3 (Fig.\u00a05c). Maximum production of ammonia was obtained at \u22120.30\u00a0V vs. SHE and no ammonia was detected in argon (Ar) on a Fe doped ReS2 surface. A weak band for ReS2 was observed with a blue shift to 709\u00a0cm\u22121 that was attributed to the synergistic effect of dual active sites (Fig.\u00a05d). Dual sites provided enhanced interaction with nitrogen due to the delocalized electronic structure and different adsorption sites. Several other authors have employed similar methodologies to explore the eNRR mechanism by different in-situ techniques [50\u201356]. By providing evidence of the mechanism these techniques are helpful in the proper selection of material in rational catalyst design for eNRR.To screen promising candidates, understand reaction processes, and optimize catalysts, ab initio simulations such as Hartree\u2013Fock, and multi-reference approaches can be utilized. DFT compared to other ab initio calculations, offers information about the energy, structure, and electronic configurations of a particular set of compositions with a lesser computational cost and higher accuracy [57]. Herein, the most relevant DFT results for electrochemical nitrogen reduction under ambient conditions, such as Volcan plots, Gibbs free energy diagrams, the density of states (DOS), and charge analysis are discussed. The utilization of these results to understand experimental results or predict novel catalyst compositions is highlighted. The framework provided by theoretical studies is envisaged to facilitate the proper selection of material for advanced electrocatalyst design.Extensive theoretical studies have been carried out for the exploration of the eNRR reaction mechanism on a variety of different surfaces [58\u201362]. Based on free energy changes and adsorption strength of possible intermediates for the eNRR process, different endergonic steps have been identified that limit the rate of reaction. A careful analysis of adsorption energies of different intermediate leads to the development of a volcano plot that gives a basic selection criterion for different types of suitable surfaces. It is a quantitative illustration of the Sabatier principle [3] in terms of the optimum interaction of different reactants, intermediates, and products with the catalyst surface. A good catalyst has optimized binding energies for both reactants and products that direct the overall process towards spontaneity. Moreover, free energy profiles provide basic insight into the mechanism and kinetics of the reaction by elucidation of rate-limiting steps. These described approaches are robust and powerful in the screening of a broad database and using thermodynamic properties as descriptors.DFT calculations on first and second-row transition metals of two different surfaces (flat and stepped) have been conducted for eNRR, and a volcano plot was established [63]. The volcano plot in Fig.\u00a06\na illustrates the free energy relationship of adsorbed N atoms on both flat and stepped surfaces of different transition metals with the applied potential difference. Metals on the right side are limited by the adsorption of nitrogen molecules on the surfaces with the first PCET reaction (an associative mechanism) and nitrogen dissociation (a dissociative mechanism) as the rate-limiting steps. However, the left side of the plot depicted the same rate-determining step for both associative and dissociative mechanisms but different reaction pathways depending on the metal surfaces. For flat surfaces the second PCET reaction and for stepped surfaces, the final PCET reaction or desorption of ammonia are the rate-determining steps. Moreover, early transition metals were predicted to have an optimum binding for N-adatoms as compared to H-adatoms, they are preferred over late transition metals for eNRR at operating potentials of around \u22121.0 to \u22121.5\u00a0V vs. SHE. Therefore, the low rate of ammonia formation could be attributed to the competing HER reaction on the late transition metals. By this approach, one can predict the most suitable transition metal for eNRR at ambient conditions based on optimal nitrogen binding and rate-determining steps.Montoya et\u00a0al. explained the linear scaling relationship of adsorption energies of eNRR intermediates on pure noble metals (Ag, Ru, Re) [64]. They employed a two-variables description of theoretical overpotentials that resulted in the exploration of the fundamental limitations of this process. It helped in the optimization of required overpotentials for different steps in the eNRR process that led to the selective design of electrocatalysts for eNRR. For eNRR, NH2 and N2H have relatively large adsorption energies that lead to the scaling relation with an overpotential of 0.5\u00a0V [61]. In this regard, strategies have been proposed to enhance eNRR activity including selective stabilization and destabilization of N2H and NH2 at different catalyst surfaces, functionalization of the surfaces by co-adsorbed molecules, and the solvation effect [65]. This can be achieved by rational catalyst design strategies that tune the catalyst surface according to the reaction requirements.DFT calculations have been also conducted to envisage the role of metal oxides in the activation of nitrogen [66]. As illustrated in Fig.\u00a06b, most of the metal oxides have endergonic nitrogen binding energies except tantalum oxide (TaO2) and rhenium oxide (ReO2), so the first step in the eNRR process is rate-limiting. Moreover, the effect of HER was also observed by examining the adsorption energies for hydrogen and nitrogen on various metal oxides. Except for ReO2 and TaO2, all metal oxides are more susceptible to adsorb hydrogen than nitrogen, making them less suitable for the eNRR process. As adsorption energies dictate various steps to be the rate-limiting step for any reaction, so the respective energy profiles will give a better understanding of reaction mechanisms. For instance, ammonia formation on Ru@C2N was theoretically analyzed by free energy profiles for different mechanisms [67]. In the associative mechanism by distal pathway, nitrogen adsorption is the first step on the electrocatalytic surfaces that resulted in the decrease in free energy for conversion from gas to the chemisorbed molecule. At applied bias U\u00a0=\u00a00\u00a0V, the first hydrogenation of the nitrogen molecule requires energy making it an endergonic step in this case. This endergonic step can become spontaneous when the bias of U\u00a0=\u00a0\u22120.96\u00a0V is applied (Fig.\u00a06c). All other steps become spontaneous with no change or decrease in free energy. Moreover, in the dissociative mechanism, the most endergonic step is the dissociation of nitrogen molecules on the surface of the catalyst as it is restricted by the high free energy barrier (2.6\u00a0eV). Therefore, the associative mechanism is more plausible for eNRR on this surface. In addition, four pathways are proposed for the dissociative mechanism including (A) the direct dissociation of the N\u2261N bond caused by the first PCET reaction on the \u2217NNH species, (B) the dissociation of the \u2217NNH2 species into \u2217N and \u2217NH2, (C) the decomposition of the \u2217NHNH2 into \u2217NH and \u2217NH2; and (D) the \u2217NH2NH2 separating into two \u2217NH2. For Ru@C2N-DM, after the adsorption of the nitrogen molecule and first proton transfer, the free energy difference is small in pathways C and D which drives the reaction downhill after every proton transfer under a bias of \u22120.96\u00a0V (Fig.\u00a06d). After the formation of NHNH2 species, the dissociation of nitrogen molecules takes place which further lowers the energy. Therefore, a careful design of the catalyst surface is necessary to make it more favorable for nitrogen adsorption as compared to the H adsorption.DOS is a non-self-consistent computation that is used to determine the density of occupied electronic energy levels following structural optimization and static calculation. The DOS is a semi-quantitative indicator of a catalyst\u2019s electronic conductivity, particularly for catalysts with similar compositions and structures. The lower bandgap between the valence and conduction bands, as well as the greater DOS near the Fermi level, indicate a higher charge carrier concentration, which favors higher electronic conductivity.Wu et\u00a0al. investigated the nitrogen reduction reaction by heteronuclear metal-free double-atom catalysts [68]. A set of 36 catalysts was evaluated for the activity and selectivity of the nitrogen reduction process by using DFT. Firstly, four different non-metallic substrates were used to evaluate the performance of this process. Among carbon nitrides (g-C3N4, g-CN, g-C2N), and boron phosphides (BP1, BP2, and BP3), Boron-based substrates were found to be more active in nitrogen reduction. A synergistic role of boron and silicon was predicted to be a suitable composition for this process. B\u2013Si@BP1 and B\u2013Si@BP3 were the most active catalysts based on the adsorption energies for different reaction intermediates. The conductive behavior of these catalysts was elucidated by the density of states. As illustrated in Fig.\u00a07\na the Fermi level was traversed by the 2\u03c0\u2217 orbital of adsorbed dinitrogen on B\u2013Si@BP1, showing that this component of the antibonding orbital of nitrogen was easily filled by electrons from the catalyst, resulting in the activation of dinitrogen. However, for B\u2013Si@ BP3, the electronic density location of 2\u03c0\u2217 was at a slightly higher energy than that of adsorbed nitrogen on BSi@BP1, which confirmed the ease in side-on adsorption than end-on adsorption to feedback electrons to the antibonding orbital of nitrogen (Fig.\u00a07b).Li et\u00a0al. systematically investigated several transition metal single-atom catalysts on tungsten sulfides (WS2) monolayers for nitrogen reduction by DFT studies and kinetic modeling [69]. N2H adsorption was chosen as the activity descriptor and found three catalysts to have the highest activity, Re@WS2, Os@WS2, and Ir@WS2. The presence of vacant and occupied 5d orbitals in these catalysts facilitated donations and back donation of electronic density, resulting in enhanced nitrogen activation. The introduction of transition metals on the monolayer WS2 induced spin polarization with the mid-gap states formed by the 5d states of the transition metal. These orbitals hybridized with the W5d and S3p to form a strong metal-support interaction. This visualization of electronic states promises a facile approach for the effective screening of catalysts.Li et\u00a0al. also demonstrated the doping effect of carbon in hexagonal boron nitride ribbons by DOS investigations [70]. A decrease in the overall nitrogen overpotential from 1.14 to 0.39\u00a0V was observed with a modulation in the adsorption energy of different reaction intermediates after the introduction of carbon atoms, which was attributed to the changes in the electronic structure of boron nitride. To further the investigation, \u2018zigzag\u2019 and \u2018armchair\u2019 nanoribbons were evaluated for the electronic state modulation with nitrogen adsorption. Fig.\u00a07c shows confirmation of the presence of a partial wave at the Fermi level for the zigzag configuration with improved conductivity that reflected the affected electron transfer in this case. A similar trend was observed for the armchair configuration that demonstrated both configurations as the active materials for nitrogen reduction (Fig.\u00a07d). However, the zigzag configuration showed magnetism due to the unsymmetrical nature of the spin-up and spin-down density of states. This spin-polarized magnetism in the zigzag configuration provided localized charge density sites on the surface of the catalysts that improved the interaction of reactants with the active site. This result demonstrated the theory-assisted rational catalyst design is a desirable approach for the development of an active, selective, and efficient catalyst for nitrogen reduction.The spatial distribution of electron density in a catalytic structure is determined as a charge density distribution. On the catalyst's active sites, which are nothing more than the localized or delocalized charge density sites, numerous surface reactions occur during electrocatalytic processes. The formation of the active site and the successive adsorption of the reactant moiety are directly related to the charge density sites in the overall structure. So, the investigation of these charge distributions is desirable for understanding the role of the active site with the plausible mechanism of a process. The charge density difference can be used to study electronic interactions and redistributions when new bonds, interfaces, or heterojunctions occur.Bader charge analysis provides evidence of the charge distribution on the surface of the catalyst [71]. This analysis demonstrates the charge distribution due to any modulation of the catalyst structure or a chemical change during a process. For instance, Zhang et\u00a0al. evaluated Fe2 clusters on molybdenum sulfides (Fe2/MoS2) for nitrogen reduction by Bader charge analysis in an enzymatic pathway [72]. The variation of the Bader charge was calculated for every elementary step of this process on three different moieties including the adsorbed N\nx\nH\ny\n (x\u00a0=\u00a01, 2; y\u00a0=\u00a01\u20133) species, the Fe2 cluster, and the MoS2 substrate (Fig.\u00a08\na). The charge fluctuation on the Fe2 cluster is less than on N\nx\nH\ny\n and the MoS2 substrates. During the eNRR process, the MoS2 substrate serves as an electron reservoir, while the Fe2 cluster serves as a charge transmitter between the adsorbed N\nx\nH\ny\n species and the MoS2 substrate.Arachchige et al. employed Bader charge analysis to investigate the role of dual metal atoms on the surface of graphdyine [73]. The Bader charge analysis for cobalt-nickel on graphdyine (CoNi@GDY) during the distal mechanism revealed a charge distribution throughout this mechanism (Fig.\u00a08b). The positive charge of CoNi@GDY was increased due to electron donation to nitrogen during the adsorption step. The charge plot for the first and second PCET revealed electron accumulation on CoC4 and electron exchange from NiC4 to adsorbed N\nx\nH\ny\n. Then, for the remaining steps, both CoC4 and NiC4 demonstrated identical charge distribution, apart from the fifth PCET, when charge fluctuation was in the other direction. The synergistic effects of two metal atoms, Co, and Ni, were established by such charge change, underscoring the necessity of a double atom catalytic site.Theoretical studies provide a solid framework for rational catalyst design and the proper selection of efficient and stable catalysts. Different surfaces have different capabilities to activate nitrogen molecules for eNRR. Identification of rate-limiting steps and scaling relations between different intermediates on catalytic surfaces help in the better design of a catalyst [60,68,74,75]. Moreover, kinetic barriers that are directly affecting the rate of reaction could be removed by an enhancement in the extrinsic properties of the catalyst. However, linear scaling relations depend mainly on the type of catalytic surface under study. Therefore, in addition to these approaches, the gradient ascent method and microkinetic modeling should be employed to assess the overall turnover frequency of a catalyst [39,71].In recent years, a variety of materials have been investigated for nitrogen reduction at ambient conditions. To have a comprehensive description of different categories, herein, we have divided electrocatalysts into four types based on composition. These include (i) noble metal-based, (ii) non-precious metal-based, and (iii) metal-free electrocatalysts.Noble metal catalysts have been employed as active catalysts for a variety of applications due to their high conductivity, active crystalline surfaces, enhanced interaction, activation, and polarization of different reactants [76]. These versatile characteristics make them a superior class of materials in catalysis. Various noble metal-based catalysts have also been explored in the recent past for eNRR [9,77\u201379]. Herein, we discuss some representative reports based on ruthenium, gold and palladium (Ru, Au, Pd) metals for eNRR at ambient conditions.Corroborated by theoretical studies, Ru is one of the most active materials and was first evaluated for eNRR at ambient conditions by Kordali et\u00a0al. [80]. They electrochemically deposited nanosized Ru on carbon felt and evaluated the electrode for eNRR. At \u22121.10\u00a0V vs. Ag|AgCl in a nitrogen atmosphere, 20\u00a0\u00b0C, and atmospheric pressure, a rate of 0.21\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122 and a current efficiency of 0.28 % for ammonia production were achieved. Similarly, a nanosized Ru-based electrocatalyst was synthesized by oleate-mediated synthesis and evaluated for eNRR in an acidic environment [81]. A greater yield of ammonia (5.5\u00a0\u03bcg\u00a0h\u22121\u00a0m\u22122) with a FE of around 5.4 % at 0.01\u00a0V vs. RHE was obtained. This selective ammonia conversion at low overpotential could be attributed to well-dispersed nanosized Ru particles on the surface of carbon fiber paper (CFP). DFT calculations highlighted the enhanced interaction of nitrogen with Ru (001) surfaces that resulted in the exergonic adsorption of nitrogen (Fig.\u00a09\na). In a recent report by the same group, Ru/MoS2 was demonstrated as a selective catalyst for eNRR [82]. MoS2 polymorphs were employed to control the HER and a FE of 17.6 % with a high yield rate of ammonia (1.14\u00a0\u00d7\u00a010\u221210\u00a0mol\u00a0s\u22121\u00a0cm\u22122) is achieved. Theoretical calculations confirmed a hypothesis that nitrogen activation took place at Ru nanoclusters and a synergistic coupling effect with S-vacancies on the 2H\u2013MoS2 resulted in the hydrogenation of N adatoms.For effective electrocatalytic ammonia production, ruthenium-doped defect-rich tin oxide nanoparticles on carbon cloth (Ru\u2013SnO2/CC) were developed by Sun et\u00a0al. [83]. In 0.1\u00a0mol\u00a0L\u22121 sodium sulfate (Na2SO4), it produced 4.83\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122 of ammonia with a FE of 17 %. The developed synergy by the combination of Ru species, SnO2, and oxygen vacancies resulted in a system with enhanced Ru stabilization and suppression of HER, while oxygen vacancies in the SnO2 lattice improved nitrogen adsorption and enhanced the activity of the Ru active center (Fig.\u00a09b).Shi et\u00a0al. developed Au sub-nanoclusters supported by TiO2 for nitrogen reduction by the tannic acid reduction method [84]. Au sub-nanoclusters (0.5\u00a0nm) decorated TiO2 demonstrated a high production yield (21.4\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121) of ammonia with a FE of 8 % at \u22120.2\u00a0V vs. RHE. This efficient activity of isolated Au nanoparticles could be attributed to the sub-nanometre dimensions that provide the enhanced exposure of active sites. Besides intrinsic catalytic activity improvement at sub nanometric dimensions, the dispersion of these isolated metal atoms on oxide supports also improved the active site density and stability of nanoparticles. In a recent report, the micelle-assisted electrodeposition approach was used for the direct fabrication of porous Au on Ni foam [85]. This catalyst exhibited a FE of 13 % with an improved rate of ammonia production (9.42\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122). Enhanced performance could be attributed to the porous nature of the catalyst that increased the interaction of reactants by nanoconfinement. Moreover, exposure to more active sites resulted in improved Au intrinsic activity.Au nanoparticles dispersed on bismuth telluride (Bi2Te3) nanosheets to form two-dimensional (2D) heterojunction. Au\u2013Bi2Te3 nanosheets have shown activity for the eNRR [79]. This could be attributed to the enhanced active site density due to the good dispersion of Au nanoparticles and the synergistic effect of the heterojunction composite. An ammonia yield rate of 32\u00a0\u03bcg\u00a0h\u22121 mg\u22121 with a FE of 20 % in a 0.1\u00a0mol\u00a0L\u22121 Na2SO4 electrolyte was obtained. The proposed method for building high-performance heterojunction electrocatalysts for electrochemical ammonia production is a viable option for rational catalyst design. Similarly, Wang et\u00a0al. studied the effect of hydrogenation on Au/TiO2 composite for the eNRR [86]. A hydrogen plasma was used to create defects and vacancies in the composite. The treated sample (H\u2013Au/TiO2) demonstrated a FE of 2.7 % at \u20130.1\u00a0V vs. RHE that could be attributed to disordered patches on the surface of TiO2 nanoparticles that emerged after hydrogenation treatment, and a substantial number of oxygen vacancies were developed into TiO2 crystalline structures.Pd-based catalysts have been explored for nitrogen reduction due to hydride formation at certain voltages, which promotes different hydrogenation reactions at the Pd surface. Nano-sized Pd decorated on carbon was synthesized by the polyol reduction method [87]. In nitrogen-saturated phosphate buffer solution (PBS), a yield of 4.5\u00a0\u03bcg\u00a0h\u22121 mg\u22121 with a high FE of 8 % was obtained at \u20130.1\u00a0V vs. RHE with lower overpotential. This is due to the dispersion of active sites with localized charge density, and the Grotthuss-like hydride transfer mechanism that controls the rate-limiting step of this process. DFT calculations corroborated the proposed hydride mechanism on the Pd surface. PCET or direct hydrogenations are thermodynamically less favorable as compared to the in-situ formed palladium hydride (\u03b1-PdH) that allows the activation of a nitrogen molecule on the Pd surface. Similarly, strong interfacial interactions were developed between PdO\u2013Pd that enhanced the nitrogen interaction and eNRR performance [88]. Laser-irradiated PdO/Pd heterojunctions on CNTs exhibited a high FE (11 %) with a high rate of ammonia production (18.2\u00a0\u03bcg\u00a0h\u22121 mg\u22121) at 0.1\u00a0V vs. RHE. A synergistic effect of PdO and Pd decreased the proton transfer rate and reduced the overpotential.In a recent report by Chen et\u00a0al. electrodeposited Pd/PdO electrocatalysts were developed to regulate oxygen levels in various gas atmospheres [89]. The inclusion of an oxygen atom into a pure Pd catalyst modulated the electron density of the Pd/PdO heterojunction that increased the adsorption energy for nitrogen and hydrogen, as corroborated by theoretical calculations. Experimental data revealed that a moderate oxygen content led to improved performance, with an ammonia yield of 11\u00a0\u03bcg\u00a0h\u22121 mg\u22121 and a FE of 22 %. This was ascribed to the moderate adsorption of nitrogen on the Pd surface along\u00a0with hydrogen suppression due to the formation of defects.Due to the high cost and scarcity of noble metals, non-precious metal-based catalysts must be investigated. Similar electronic properties of transition metals and their surface chemical bonds make non-precious metals a good choice for gas-phase reactions and as an alternative for noble metals. In the recent past, enormous research has been done on non-precious metal-based electrocatalysts for various applications [90\u201392].Based on theoretical studies, Fe is predicted to be a good catalyst candidate for eNRR at ambient conditions. Moreover, its interactions with nitrogen and other intermediates promise feasible pathways for eNRR. Different Fe-based nanocatalysts have been investigated for eNRR at ambient conditions. Zhou et\u00a0al. achieved a FE of 60 % for ammonia on a nanostructured Fe-based electrocatalyst [93]. They employed phosphonium-based ionic liquids with high nitrogen solubility to improve its availability in the system. The selectivity can be improved by controlling the water content in the reaction media that provides hydrogen for the competing HER. At around \u22120.8\u00a0V vs. NHE, FE of more than 60 % and a rate of 4.7\u00a0\u00d7\u00a010\u221212\u00a0mol\u00a0s\u22121\u00a0cm\u22122 was achieved for Fe-FTO catalyst in [P6,6,6,14][eFAP] ionic liquid with low current densities. Fe\u2013N3 sites were explored as active sites for eNRR at ambient conditions [37]. Zeolitic imidazole framework (ZIF) and carbon nanotubes (CNT) derived Fe\u2013N/C-CNT hybrid exhibited a FE of 9.2 % with a 34\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 rate of ammonia production at \u22120.2\u00a0V vs. NHE. As corroborated with DFT calculations, Fe\u2013N3 sites were active centers for enhanced nitrogen activation and polarization.Due to its high electron-donating capability and empty 6d orbital, bismuth (Bi) has been explored as an active material for different gas-phase reactions where the activation and polarization of reactants are conducted by transfer of electron density back and forth [94]. Its semiconducting nature provides localized density states near the Fermi level that act as localized charge sites that help in the donation of p-electrons to nitrogen for activation. In this regard, different Bi-based catalysts have been explored for eNRR. 2D mosaic Bi nanosheets (BiNS) were developed for eNRR in neutral media [95]. The in-situ electrochemical reduction process resulted in the formation of mosaic BiNS that exhibited enhanced eNRR performance. A FE of 10 % with an ammonia rate of 13\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved at \u22120.8\u00a0V vs. RHE. The 2D structure provided the enhanced exposure of active sites that synergistically improved the intrinsic catalytic property of Bi by effective p-electron delocalization in BiNS. In another report, a Bi nanosheet array was developed electrochemically and evaluated for eNRR at ambient conditions [94]. A FE of 10 % with an ammonia production rate of 6.89\u00a0\u00d7\u00a010\u221211\u00a0mol\u00a0s\u22121\u00a0cm\u22122 at \u22120.5\u00a0V vs. RHE in acidic media. DFT calculations revealed the enhanced activation of nitrogen on Bi and an associative alternating pathway was proposed for nitrogen reduction.Xue et al. employed a wet-chemical approach using sodium citrate as a stabilizing agent to grow ultrafine tin (Sn) nanoparticles on carbon black [96]. This catalyst exhibited a FE of 22 % with an ammonia yield rate of 17\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 in the 0.1\u00a0mol\u00a0L\u22121 Na2SO4 electrolyte which was attributed to the small size of the particle improved the active site density and intrinsic catalytic activity of the material. The role of Sn as a sacrificial species for the hydrogen evolution reaction on the surface of phosphorene was explored by Liu et al. [97]. This improved the overall conservation and separation of nitrogen adsorption sites on the surface of phosphorene which resulted in the improved selectivity toward ammonia formation. At a low overpotential, Sn-phosphorene obtained a FE of 36 % and a notable ammonia production rate of 26.9\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121. After doping with Sn, DFT calculations revealed that different adsorption sites for water and nitrogen were available, with water adsorbing preferentially onto the Sn sites. Similarly, antimony (Sb) based catalysts were also evaluated for nitrogen reduction and found to be active at ambient conditions. The coupling of Sb and metallocene (Nb2CT\nx\n) developed a localized electron-rich interface as confirmed by the DFT calculations [98]. This resulted in the modulation of the Sb band structure which improved nitrogen activation and hydrogenation. An ammonia production rate of 49.8\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved with a FE of 27 % at ambient conditions.Owing to the better intrinsic catalytic activity of different metals as suggested by theoretical approaches, the derivatization of these metals opens a new window for enhanced properties. For instance, Mo is an active entity in the HER process but to improve its intrinsic catalytic properties variety of derivatives have been developed including phosphides, carbides, and sulfides [99]. Modification of the electronic properties of the parent metal resulted in better interaction with reactant molecules (H2O, N2, O2, CO2) due to the effective polarization of electronic density by the incoming group. In this regard, different transition metal derivatives have been explored in the recent past and can be categorized into i) transition metal oxides, ii) transition metal nitrides, iii) transition metal carbides, and iv) transition metal sulfides. Representative candidates of these types are illustrated in the sections below.Due to lesser conductivity than the parent metals, transition metal oxides tend to have lower performance in some electrochemical processes. Theoretical calculations revealed that early transition metals have a high tendency to interact with nitrogen as compared to late transition metals, resulting in the ease of nitrogen activation and polarization. To investigate the catalytic ability of metal oxides for eNRR at ambient conditions, Skulason and co-workers conducted DFT calculations of rutile-type transition metal oxides [66]. A stability diagram for each transition of metal oxide was developed by adsorption of different species on its (110) facet, and then the adsorption energies were evaluated as a function of applied potential. Thermodynamic aspects were also elucidated by free energy diagrams of the eNRR process. Based on the binding energies of N2H species on different transition metal oxides, a volcano plot was developed reflecting the catalytic activity and selectivity of different surfaces for eNRR. The most promising transition metal oxides for nitrogen reduction are NbO2, ReO2, and TaO2 with low onset potentials. Iridium oxide (IrO2) was found to be the most active oxide for this process with an onset potential of \u22120.36\u00a0V vs. SHE, but it is prone to be poisoned by hydrogen atoms, giving a lower selectivity towards eNRR.Rational catalyst design leads to improved intrinsic catalytic properties of a catalyst [100]. Fe/Fe-oxide catalysts were prepared by the oxidation of Fe foil at high temperatures and then electrochemically reduced to different interfaces containing Fe\u2013Fe3O4 and Fe\u2013Fe2O3 hybrids. Due to the less conductive nature of Fe-oxides, suppression of HER takes place with enhancement in the intrinsic catalytic property of Fe towards eNRR at ambient conditions. At \u22120.3\u00a0V vs. RHE, a FE of 8 % was obtained with an ammonia yield rate of 0.20\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122. Multishelled hollow chromium oxide (Cr2O3) microspheres were synthesized by the hydrothermal approach and evaluated as an efficient eNRR electrocatalyst\u00a0[101]. A FE of 6 % in 0.1\u00a0mol\u00a0L\u22121 Na2SO4 was obtained with a total ammonia production yield of 23.5\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121\ncat at \u22120.9\u00a0V vs. RHE. This performance was attributed to the rational design of this non-noble metal catalyst with a hollow texture. This hollow structure facilitated the diffusion of nitrogen and desorption of products that enhance the mass transport with the improved kinetics of the reaction.Niobium oxide (Nb2O5) nanofibers showed excellent performance towards eNRR [102]. In 0.1\u00a0mol\u00a0L\u22121 hydrochloric acid (HCl), an average yield of 43.6\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 with a FE of 9 % at \u22120.55\u00a0V vs. RHE was achieved. Intrinsic catalytic properties of Nb due to the enhanced interaction with nitrogen in an exergonic step are the source of the improved performance. Also, DFT calculations predicted that Nb edge atoms were involved in the activation and polarization of the dinitrogen molecule, and charge transfer takes place between the surface Nb atom and dinitrogen molecule that was compensated by the neighboring Nb atom by back donation, hence, weakening the N\u2261N bond. Interestingly, NbO2 exhibited better nitrogen reduction performance than Nb2O5 [103]. At \u22120.6\u00a0V vs. RHE in acid media, a FE of 32 % was achieved, and at \u22120.65\u00a0V vs. RHE, a high rate of ammonia production (11.6\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121) was obtained that could be assigned to the different linkage of Nb atoms in the crystal structure. In NbO2, there is an availability of empty d-orbitals that caused the enhanced interaction of nitrogen molecules on the surface by receiving the electronic density from nitrogen. Moreover, a single d-electron participated in the activation of N\u2261N by back donation resulting in the high performance of the material towards nitrogen reduction.Similarly, layered 2D perovskites have anomalous electronic properties that make them an interesting choice for different electrochemical applications. Atomically disassembled nanosheets of La2Ti2O7 were developed by the hydrothermal method and their electrochemical nitrogen reduction performances were investigated [104]. A high ammonia yield of 25\u00a0\u03bcg\u00a0h\u22121 mg\u22121 with a FE of 4.5 % was achieved at \u22120.55\u00a0V vs. RHE in the acid electrolyte with good stability. It could be attributed to the 2D layered structure of La2Ti2O7 that enhanced the interaction of reactants with catalysts by exposing more active sites. In a recent report, doped-LaFeO3 was developed by the sol-gel method and evaluated for nitrogen reduction [105]. The ammonia formation rate of 13\u00a0\u03bcg\u00a0h\u22121 mg\u22121 was achieved at a cell voltage of 2.4\u00a0V, while, a FE of 1.9 % was obtained at 1.8\u00a0V. Enhanced eNRR performance could be attributed to the Cs and Ni doping in the LaFeO3 structure along with oxygen vacancies.Modified electronic structures of metal nitrides along with N-vacancy tailored structures provide enhanced interaction of nitrogen with metal nitride surfaces. Extensive theoretical studies have been carried out on transition metal nitrides to evaluate the eNRR. Abghoui et\u00a0al. extended this approach by investigating the reaction mechanism and the dissociation barrier of nitrogen on transition metal nitrides [44,106]. Four promising candidates were identified including vanadium nitride, zirconium nitride, chromium nitride, and niobium nitride (VN, ZrN, CrN, NbN). These metal nitrides were proposed to follow the Mars-van Krevelen mechanism in which an ammonia molecule is formed by the reduction of surface nitrogen of a nitride creating a N-vacancy that was filled by dissolved nitrogen from the electrolyte. They proposed that low-index facets of these transition metal nitrides were stable towards poisoning, decomposition, and suppression of activity by adsorbed oxygen and hydrogen molecules and were expected to produce ammonia with high current densities. Computation by the hydrogen electron method determined that the Mars-van Krevelen mechanism (MvK) was most plausible for eNRR on transition metal nitrides (Fig.\u00a010\na and b) [107]. Cobalt molybdenum nitride (Co3Mo3N) is reported to be the most active catalyst for nitrogen reduction at 400\u00a0\u00b0C [108]. Various theoretical studies were conducted on the understanding of its mechanism at the atomic level by Yazdi et\u00a0al. [109]. Dispersion-corrected DFT calculations were performed over Co3Mo3N surfaces with defects to elucidate the mechanism of ammonia formation at different temperatures and the MvK mechanism was confirmed.The activity of VN nanoparticles towards nitrogen reduction was studied in a membrane electrode assembly in a fuel cell arrangement with hydrogen fed at the anode as the source of protons [110]. At \u22120.1\u00a0V vs. RHE, a FE of 6 % with an ammonia production rate of 3.3\u00a0\u00d7\u00a010\u221210\u00a0mol\u00a0s\u22121\u00a0cm\u22122 was achieved and 15N isotope labeling experiments confirmed the MvK mechanism. X-ray photoelectron spectroscopy (XPS) analysis of fresh and spent catalysts observed the phase transformation from VN to VN1-x\nO\nx\n and confirmed the active site for this process. Structural N-atoms with adjacent O-atom were susceptible to hydrogenation which leads to ammonia formation. In this regard, in-situ and operando characterizations are crucial for the explanation of structure performance relationships. In a recent report, titanium oxynitride (TiON) was evaluated as an efficient catalyst for eNRR [111]. Commercial titanium nitride (TiN) was etched by a plasma-enhanced approach and a TiON phase formed that enhanced the overall performance. This catalyst exhibited a FE of 9 % with an ammonia production rate of 3.32\u00a0\u00d7\u00a010\u221210\u00a0mol\u00a0s\u22121\u00a0cm\u22122 at \u22120.6\u00a0V vs. RHE in neutral media.Despite various theoretical studies that predict transition metal nitrides as active nitrogen reduction catalysts, experimental evidence is lacking in this regard. Recently, MacFarlane and coworkers evaluated Nb4N5 and VN for nitrogen reduction at ambient conditions and found interesting results [43]. According to their deductions after various experiments, VN and Nb4N5 were inactive for eNRR but they were producing ammonia by reductive decomposition of lattice nitrogen. Nb and V hydroxide precursors were developed by hydrothermal methods and their further nitridation was done at high temperature in the presence of an ammonia environment. After evaluating nitrogen reduction in different control experiments (dinitrogen, Ar, open circuit), no great difference was observed in ammonia production. However, soaking the electrode in acid resulted in the production of a great amount of ammonia which was attributed to the non-catalytic reductive decomposition of lattice nitrogen. Different control experiments were also conducted to elucidate the exact origin of ammonia on these transition metal nitrides. It was confirmed that the produced ammonia was from the decomposition of lattice nitrogen by the formation of irrecoverable nitrogen vacancies. In another recent report, similar results were obtained for Mo2N at ambient conditions [113]. Different control experiments highlighted the decomposition of Mo2N under these reductive conditions and proved to be inactive for eNRR at ambient conditions. These results illustrated that careful insight is required towards ammonia production by N-containing materials.Like 2D nitrides, 2D transition metal carbides such as M3C2 also have the potential to capture and reduce nitrogen to ammonia. DFT calculations were conducted by Azofra et\u00a0al. to explore metal carbides for nitrogen reduction [112]. The efficient stabilization of nitrogen molecules on metal carbides leads to its activation by elongation and weakening of its bonds, resulting in the high performance of ammonia formation. Moreover, the first PCET reaction was the rate-determining step for nitrogen reduction that required low activation barrier, for instance, vanadium carbide (V3C2) required 0.64\u00a0eV. Metal carbides of d2, d3, and d4 are proposed to be more active for the stabilization of nitrogen as indicated by the spontaneous chemisorption energies (Fig.\u00a010c). In a recent report, molybdenum carbide (Mo2C) nanodots were developed by molten salt synthesis and evaluated for eNRR [114]. At \u22120.3\u00a0V vs. RHE, and ammonia yield of 11\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved with a FE of 7.8 % on a hydrophilic substrate. Similarly, Mo2C nanorods also exhibited eNRR activity at ambient conditions [115]. A Mo-based precursor (Mo3O10(C6H8N)2\u00b72H2O) was pyrolyzed at high temperatures to form Mo2C nanorods. A FE of 8 % with an ammonia yield rate of 95.1\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was obtained at \u22120.3\u00a0V vs. RHE in acidic media.2D transition-metal carbide provide the metallic properties of carbides along with the hydrophilic nature of the termination (OH, O) that make them a suitable class of compounds for versatile applications [116,117]. 2D Ti3C2T\nx\n (T\u00a0=\u00a0functional group termination like F, OH, x\u00a0=\u00a01\u20133) nanosheets were developed by the delamination technique and evaluated for nitrogen reduction [118]. At \u22120.4\u00a0V vs. RHE in acidic media, a FE of 9 % with an ammonia yield of 20\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was obtained which is attributed to the improved selectivity for nitrogen reduction. Enhanced chemisorption capabilities of nitrogen on Ti3C2Tx were confirmed by DFT calculations that resulted in the activation and polarization of the N\u2261N bond. High specific surface area and abundant active sites were the key improvements by nanostructuring. Similarly, a few-layered MoSe2 on Ti3C2Tx MXene (MoSe2/Ti3C2Tx) was fabricated via a hydrothermal synthesis and thermal annealing as an active electrochemical nitrogen reduction catalyst [119]. DFT studies elucidated Mo atoms as active centers for nitrogen reduction with the distal pathway for ammonia production. MXene provided the platform for the exposure of active sites by improved dispersion and conductivity. The developed heterostructure exhibited an ammonia yield rate of 56\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 at \u22120.50\u00a0V vs. RHE and a FE of 14 % \u22120.35\u00a0V vs. RHE.Like other derivatives, TMS also have been investigated for nitrogen reduction at ambient conditions. Fe-atom decorated 2D MoS2 sheets for ammonia formation at ambient conditions were investigated by theoretical studies [10]. Like the FeMo cofactor in nitrogenase, the Fe\u2013Mo linkage is supposed to have high selectivity for nitrogen, which facilitates the nitrogen reduction endergonic steps. DFT calculations demonstrated that the stabilization of the nitrogen molecule on a Fe site could be ascribed to the exchange of mutual electronic density between Fe and N atoms activating the N\u2261N bond. A low activation barrier of 1.02\u00a0eV was calculated for the first PCET reaction which was the most endergonic step on the Fe\u2013MoS2 surface. Also, the lower binding energy between ammonia and this surface resulted in the ease of product desorption that regenerated the catalytic active sites.Recently, MoS2 was synthesized by the hydrothermal method on carbon cloth as an electrode for nitrogen reduction [1]. DFT calculations confirmed that positively charged Mo edges were the main catalytic sites for eNRR as they stabilized nitrogen by receiving electronic density. Also, the first hydrogenation step was the rate-determining step that requires 0.68\u00a0eV free energy in the absence of applied potential. A rate of 8.8\u00a0\u00d7\u00a010\u221211\u00a0mol\u00a0s\u22121\u00a0cm\u22122 for ammonia production was achieved with a FE of 1.1 % in 0.5\u00a0mol\u00a0L\u22121 Na2SO4 at \u22120.5\u00a0V vs. RHE. Wu et\u00a0al. developed a unique sub-monolayer MoS2\u2212x\n structure that selectively adsorbed, activated, and dynamically hydrogenated nitrogen [120]. This selective interaction improved the selectivity of the catalyst by controlling the scaling relation. Experimental and theoretical results demonstrated that the developed surfaces modulated the surface binding of nitrogen intermediates which resulted in improved ammonia formation. By the rational catalyst design, a FE of 24.7 % with an ammonia formation rate of 17\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved with enzymatic side-on nitrogen adsorption.Nitrogen-doped carbons are viewed as an efficient class of material due to their effective role in various catalytic reactions. Due to the redistribution of electronic density in the structure after N-doping, improvement in the number and type of active sites for enhanced interaction of reactant nitrogen molecules is observed [121]. Specifically, a higher content of pyridinic and pyrrolic N-sites contributes to the performance. A wide interest has been developed in the derivatization of carbon from different porous materials like metal-organic frameworks (MOFs) and Zeolitic imidazole frameworks (ZIFs) as they inherit the parent porosity of material to the formed carbon. These porous structures improve mass transport and increase the charge density of the material. Also, high conductivity and large potential window favor their applicability in nitrogen reduction.Porous nitrogen-doped carbon was synthesized by the carbonization of ZIF-8 and evaluated for nitrogen reduction at ambient conditions [122]. Ammonia production rate of 1.4\u00a0mmol\u00a0g\u22121\u00a0h\u22121 was achieved at \u22120.9\u00a0V vs. RHE in an aqueous electrolyte. DFT calculations proposed an associative alternative pathway for ammonia formation by porous N-doped carbon where successive hydrogenations take place on alternate N atoms. Similarly, nanoporous nitrogen-doped carbon derived from ZIF-8 was evaluated for nitrogen reduction [123]. An improved ammonia production rate of 3.4\u00a0\u00d7\u00a010\u22126\u00a0mol\u00a0s\u22121\u00a0cm\u22122 was obtained with a FE of 10 % at \u22120.3\u00a0V vs. RHE. Pyridinic-N sites adjacent to C-vacancy were more prone to interact with nitrogen molecules which lead to the dissociation of N\u2261N bond with the subsequent hydrogenations, as revealed by DFT.Biomass-derived porous carbons are gaining much interest due to the large abundance of different types of natural sources, their non-toxic nature, and their low cost. Also, the presence of other heteroatoms (N, B) improved their intrinsic catalytic properties. Cicada sloughs were used to develop porous NDC by a high-temperature annealing process [124]. These developed carbons were assigned as an effective and stable nitrogen reduction catalyst with an ammonia yield of 15.7\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 and a FE of 1.4 % at \u22120.2\u00a0V vs. RHE. This could be attributed to the synergistic effect of doping and the high surface area of developed carbons (1547\u00a0m2\u00a0g\u22121) that increased the number of active sites with improved mass transport. Similarly, using tannin as a precursor, biomass-derived O-doped carbon nanosheets were developed and their nitrogen reduction performance was investigated [125]. At \u22120.6\u00a0V vs. RHE, a FE of 4.9 % was obtained with an ammonia yield rate of 20\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 with high electrochemical stability.Different types of active sites for enhanced interaction with nitrogen were achieved by boron doping. B-doped 2D graphene was synthesized by high-temperature treatment in the presence of a B-precursor and evaluated for eNRR at ambient conditions [126]. The redistribution of electronic density in the graphene structure resulted in the formation of Lewis acid sites that could interact well with the nitrogen (Lewis base). Moreover, an empty orbital in B readily accepted the electronic density of nitrogen resulting in the activation and polarization of the nitrogen molecule. At \u22120.5\u00a0V vs. RHE, a FE of 10.8 % with an ammonia yield of 9.8\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122 was observed with 6.2 % B-doping in the graphene structure.Song et\u00a0al. developed N-doped carbon nanospikes for eNRR in aqueous media as a physical catalyst [127]. At \u22121.19\u00a0V vs. RHE, a FE of more than 11 % with a 97\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122 rate of ammonia production was obtained. Due to the absence of any metal, performance could be attributed to the confined electric charge developed at the surface of nanospikes as proven by the comparison of performance with O-etched blunt carbon tips that produced very low ammonia. Also, HER was suppressed by the formation of a dehydrated cationic layer near the surface of tips that allowed only nitrogen to pass on and excluded water.The catalytic activity of any material depends on its chemical nature and the composition of the catalyst which dictate the electronic properties that in return validate it as a good or a bad catalyst. The systematic investigation of intrinsic catalytic activity can provide new insight into a high-performance catalyst design. The rate of a chemical reaction depends on the exposed active sites as the collisions and interactions of reactant molecules with these active sites result in the formation of products. Towards the improvement in reaction kinetics, an increase in active site density is a prerequisite for enhanced catalytic activity. It has been confirmed that multifold enhancement in the performance of material with fast kinetics has been achieved by tuning the number of active sites. Herein, we describe recent strategies for rational catalyst design for eNRR by theoretical and experimental approaches. We shed light on (i) nanostructuring of materials, (ii) single atom formation, (iii) electronic structure modifications, and (iv) surface modifications.The reduction of particle size from bulk to the nano level modifies the fundamental properties of the material including physical, electronic, and chemical properties. Confinement of particles in nano dimensions introduces different energy levels that show the inclusive alteration in the electronic structure of the material. Due to greater exposure to a total number of atoms in nano dimensions, materials behave differently based on the size and number of particles due to the increased availability of active centers. Further, certain areas with localized charge density can act as sites for enhanced interaction with the reactant moiety [128]. Due to the highly conductive nature and the availability of free (d orbital) electrons, nanosized noble metals have excellent catalytic activity towards a variety of reactions related to energy applications. Small molecules (N2, H2, O2, CO2, etc.) can be easily stabilized by the transfer of electronic density from d-orbitals of metals to the vacant orbitals (antibonding) of these small molecules.Nanosized non-noble transition metals have also been employed as catalysts for different applications as they can have similar electronic properties as noble metals [61,62,129]. Improvement in the active site density and mass transport properties by nanostructuring during nitrogen reduction has led to enhancement in FE towards ammonia production with high yield. Within the nanoscale, further optimization of the shape and morphology of a catalyst along with the exposure of specific facets inclined toward enhanced catalytic activity. Herein, we highlight these strategies from recent literature for enhanced eNRR performance.The peculiar properties of metal and non-metal catalysts are observed in different shapes and sizes, and it is a fundamental goal of material science to tailor the shape of particles [130\u2013133]. Anisotropic metal nanoparticles (NPs) are important in the fabrication of different smart devices including a wide range from electronics to biological applications. Different shapes have various surfaces that interact differently with the reactant molecules as the specific electronic environment is available for a specific shape.Nazemi et\u00a0al. developed and evaluated Au hollow nanocages (AuHNCs) for the eNRR (Fig.\u00a011\na and b) [134]. In 0.5\u00a0mol\u00a0L\u22121 lithium chlorate (LiClO4) solution, at \u22120.4\u00a0V vs. RHE, a FE of 30 % was obtained and the highest yield of ammonia production (3.9\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122) was achieved at \u22120.5\u00a0V vs. RHE. This high performance was attributed to the hollow structure that provided high surface area and the confinement effect resulted in improvement in mass transport and active site density. Further to this approach, the same group elucidated the effect of pore size and density on the walls of AuHNCs on eNRR [135]. In a recent report by the same group, the role of oxidation of silver (Ag) in Au\u2013Ag HNCs was elucidated [136]. Resultant Ag2O\u2013Au nanocages exhibited a FE of 23.4 % at \u22120.4\u00a0V with an ammonia production rate of 2.14\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122. This study emphasized the need for an O2-free environment in electrochemical nitrogen reduction for stable ammonia formation. Moreover, the role of Ag in bimetallic Au\u2013Ag nanocages for the improved selectivity and activity towards nitrogen reduction was confirmed to be vital.Cobalt phosphide (CoP) hollow nanocages (HNCs) were developed by the self-assembly of MOF-derived ultrathin CoP nanosheets (Fig.\u00a011c and d) [137]. The shape-controlled growth of these nanoarchitectures shows a synergistic effect for nitrogen reduction with coordinatively unsaturated catalytic sites of phosphide. A FE of 7.3 % was achieved at 0\u00a0V vs. RHE with an ammonia yield of 2\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 in 1\u00a0mol\u00a0L\u22121 potassium hydroxide (KOH). However, an exponential increase in ammonia rate was observed until \u22120.4\u00a0V vs. RHE, and a rate of ammonia formation of 10.78\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was obtained. It was ascribed to the inhomogeneous surfaces with charge-separated sites on the CoP catalyst due to hollow nanocages.Tuning the morphology of material with hierarchical structures at the nanoscale also improved the active site density and confinement effect. Also, enhancement in the mass transport of material was achieved with an improved rate of reaction. Recently, hierarchical Au flower-like microstructures were synthesized by a soft templating method using gum Arabic as a capping agent and evaluated for nitrogen reduction [138]. As compared to spherical Au particles, a high ammonia yield (25\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121) with a FE of around 6 % was achieved at \u22120.2\u00a0V vs. RHE which was ascribed to the enhanced exposure of active sites in a hierarchical structure. Similarly, atomically distributed nanosheets nano-assemblies of Rh metal were synthesized by cyanogel-assisted method and evaluated for nitrogen reduction [139]. High specific surface area and modified electronic structure by 2D-nanosheet structure were the attributes of improved performance. In alkaline media, an ammonia production rate of 23\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 and improved selectivity were obtained at low potential (\u22120.2\u00a0V vs. RHE). Different shapes and morphologies improved the exposure of active sites in a catalytic material. In aqueous nitrogen reduction, challenges related to the low solubility of the nitrogen could be overcome by the confinement of reactants in hollow and porous structures. This results in the shortened mass and electron pathway with the enhancement in the rate of a catalytic process.The development of nanocrystals with specific facets provides coordinatively saturated and unsaturated active sites and surfaces that can be tailored according to the framework provided by the computational approach. The lattice mismatch at different scales by facet engineering develops strain in the structure of the material. This strain can propagate in the structure and gradually fades away from the interface. This resulted in the improved energetics of the catalytic reaction at the interface. High index and low index facets of material interact differently with a reactant. Low index facets are energetically more stable as compared to high index facets but show less intrinsic catalytic activity [140,141]. So, a necessary insight by theoretically assisted experimental studies towards facet control synthesis of a variety of materials for nitrogen reduction is outlined here.Theoretically, nitrogen reduction on the molybdenum nitride (\u03b3-Mo2N) surface with different facets was evaluated by Metanovic et\u00a0al. [142]. Free energy profiles confirmed the eNRR process proceeds by both (associative and dissociative) pathways, with a series of reactivity decreasing in the order of (111)\u00a0>\u00a0(101)\u00a0>\u00a0(100)\u00a0\u223c\u00a0(001) with a range of overpotential between \u22120.7\u00a0V to \u22121.4\u00a0V. Due to the high interaction of (111) surface with the nitrogen as compared to H-adatoms and side-on adsorption of nitrogen molecule resulted in the higher activity of this facet.Facet control synthesis of Mo-nanofilm was achieved by combined electrochemical anodization and reduction process and was further evaluated for ammonia formation [143]. At an overpotential of \u22120.14\u00a0V vs. RHE, a FE of 0.72 % was obtained with the 3.09\u00a0\u00d7\u00a010\u221211\u00a0mol\u00a0s\u22121\u00a0cm\u22122 rate of ammonia production. Similarly, tetrahexahedral Au (THH) nanorods with a high index facet (730) were prepared and had a 1.6\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122 rate of ammonia production at \u22120.2\u00a0V vs. RHE [144]. Moreover, a FE of around 4 % was obtained for ammonia along with the production of hydrazine. Briefly, the interaction of nitrogen with the highly coordinatively unsaturated surfaces (730) on terraces was the first step of this mechanism. Later, the nitrogen bond dissociated and the N atom was chemisorbed. Successive PCET reactions lead to the complete reduction process. The improved eNRR performance on the high index facet is attributed to the facet engineering that provided the coordinatively unsaturated active sites.Single atoms due to a small size (below 1\u00a0nm) and localized electronic density behave in a different pattern (like the molecular catalyst) as compared to agglomerated nanostructures for a catalytic process. Enhanced atomic-level interactions of the catalyst with reaction intermediates and localized electronic densities of atomically dispersed transition metals result in effective activation and polarization of small molecules with an improvement in the selectivity of the reaction.DFT calculations suggested the atomically distributed Mo on boron nitride (BN) monolayer to be an efficient catalyst for eNRR [145]. Exploration of different other transition metals (Sc to Zn, Ru, Rh, Pd, and Ag) as a single atom on a BN monolayer as a nitrogen reduction catalyst had also been done. Single Mo atoms supported by a BN monolayer are found to be highly active for nitrogen reduction with an overpotential of 0.19\u00a0V vs. RHE, which is much lower than other investigated Mo-based catalysts. This high activity could be attributed to the high spin-polarization, selective stabilization of N2H, or destabilization of NH2 intermediate on this surface.Au single atomic sites were developed on N-doped porous carbon (NDPC) by a template-assisted, impregnation method followed by reductive annealing at high temperatures [146]. The NDPC exhibited improvement in overall mass transport and metal catalyst stabilization. Electronic polarization by Lewis acidic and basic sites resulted in the activation of nitrogen and improved efficiency. At \u22120.2\u00a0V vs. RHE, a FE of 12.3 % was achieved with a yield rate of 2.32\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122 for ammonia production. Single-atom Ru sites on NDPC were also developed by thermal treatment of MOF-based precursors [147]. eNRR performance in aqueous media at \u22120.1\u00a0V vs. RHE revealed a FE of 21 % with an ammonia yield rate per mg of Ru of 3.66\u00a0mg\u00a0h\u22121. Importantly, the addition of zirconium oxide (ZrO2) suppressed the hydrogen evolution reaction. Being an early transition metal, it exhibited high interaction with nitrogen which improved the overall ammonia production.Similarly, isolated single-atom Fe anchored on N-doped carbon exhibited nitrogen reduction activity [148]. DFT calculations confirmed the stabilization of Fe by N in the Fe\u2013N4 configuration which is considered an active site for nitrogen activation. At \u22120.4\u00a0V vs. RHE, this catalyst achieved a FE of 18 % with an ammonia yield rate of 62.9\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121. In another report, a positive shift in the onset potential of ammonia was achieved on single atom Fe anchored on N\u2013C derived from pyrrole [149]. At 0.193\u00a0V vs. RHE over 56.55\u00a0% FE was achieved which could be attributed to the enhancement in active sites and the synergistic effect of N\u2013C. Molecular dynamic simulations revealed the enhanced access to nitrogen on the single-atom surfaces that resulted in the suppression of HER. Similarly, a cost-effective Mo-based single-atom catalyst was developed and evaluated for eNRR [150]. Single-atom Mo anchored on nanoporous carbon (NPC) achieved a FE of 14.6 % with a high rate of ammonia yield of 34\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121. Moreover, this catalyst exhibited long-term durability and electrochemical stability.The fundamental shift in energy level and electron structure is the effect of reducing the size of nanomaterials to cluster- or atom-level. Theoretical and experimental results show that single atoms or smaller clusters have superior catalytic activity or selectivity than larger clusters. This catalytic design is facing a few challenges that need more attention for rational catalyst design. For instance, the role of support in the confinement of single atoms and its interaction with the metal moiety to control the reaction progress need to be elucidated. It is highly desirable to develop a multi-functional single-atom catalyst as most electrocatalytic reactions are multi-step processes and are affected by the scaling relation. Different adsorption sites in a catalytic structure could provide a tandem-like structure that results in a domino effect. Importantly, single-atom catalysts are less stable and hard to produce at a large scale which dictates the need for the development of high-throughput methods for their development [151].For rational catalyst design, improvement in the intrinsic catalytic activity of the material is a fundamental requirement. The electronic structure of a catalyst dictates its intrinsic catalytic activity. Therefore, the electronic properties of the material could be modified after considering of reaction mechanisms suggested by the theoretical framework. Thermodynamic and kinetic reaction parameters can be well optimized on different types of catalytic surfaces. Enhanced interaction of reactants and products with the surface of the catalyst is vital for the spontaneity of a reaction. In this regard, different types of active sites on the surface of the catalyst will improve the overall interaction of reactants and products with the catalyst surface. Electronic properties could be improved by i) doping, ii) alloying and ii) core-shell structures. Herein, the exploitation of these strategies for nitrogen reduction is highlighted.The deliberate introduction of heteroatoms into a parent material to vary its electrical and structural properties is known as doping. It modifies the charge density of the parent material by the inclusive effect of the electronic density of the dopant and the creation of defects. Metal and non-metal doping of material produce localized charge sites that help in the enhanced interaction of catalysts with the reactant molecule. For instance, doping of carbon or carbon-based material with the more electronegative element (N) distributed the electronic density of carbon away from it resulting in the creation of positively charged carbon sites [152]. These positively charged sites are available to receive an electronic density of reactant molecules (N2, O2) that causes their stabilization.Aluminium-doped graphene was explored as a catalyst for eNRR by DFT calculations [153]. Aluminium provided adsorption and binding sites for the reactants and graphene facilitated efficient electron transport. The proposed mechanism described the hydrogenation of adsorbed nitrogen moiety on the aluminium site like the internal hydrogen transfer in homogeneous catalysis. Interestingly, B-doped TiO2 exhibited eNRR performance in neutral media [154]. In neutral electrolyte, an ammonia formation rate of 14.4\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was obtained with a FE of 3.4 % at \u22120.8\u00a0V vs. RHE. This performance could be attributed to the positive charge developed on the TiO2 due to the electron-deficient nature of B, which resulted in the enhanced interaction with nitrogen. Similarly, fluorine-doped iron double hydroxides (\u03b2-FeOOH) nanorods displayed an improved nitrogen reduction performance [155]. At \u22120.6\u00a0V vs. RHE, a FE of 9 % with an ammonia yield rate of 42\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved. DFT calculations revealed that substitution of the hydroxide (OH) group with fluorine (F) helped in the lowering of adsorption energy that improved the kinetics of the reaction.Various efforts have been employed for the alteration of carbon structure with heteroatom doping. Sulfur-doped carbon nanospheres were developed by hydrothermal method followed by annealing in Ar using glucose as a carbon source [156]. At \u22120.7\u00a0V vs. RHE, a FE of 7 % with an ammonia rate of 19\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved in 0.1\u00a0mol\u00a0L\u22121 Na2SO4. A similar effect has been explained by O-doped and biomass-derived N-doped porous carbons towards nitrogen reduction [157,158].To exploit the improved electronic environment produced by the co-doping strategy, B\u2013N-rich defective carbon nanosheets were synthesized by thermal treatment of graphene oxide and boric acid in an ammonia environment (Fig.\u00a012\na) [159]. DFT calculations elucidated the B\u2013N pair as a trigger and the adjacent C atom as an active site for the eNRR process. Due to the presence of extended active sites, enhanced nitrogen reduction performance was achieved with a FE of 13\u00a0% and an ammonia production rate of 7.7\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 at \u22120.3\u00a0V vs. RHE. Accurate quantification of ammonia was obtained by rigorous control experiments (Fig.\u00a012b). In another report, N\u2013P co-doped carbon was developed by thermal treatment of polyaniline (PANi) in phytic acid [160].Hierarchical nanocarbon foams were obtained and evaluated for nitrogen reduction in acidic electrolyte. A FE of 4.2 % with an ammonia yield rate of 0.97\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved at \u22120.2\u00a0V vs. RHE. In-situ FTIR confirmed the associative alternating pathway for ammonia formation on this surface, as hydrazine formed during this process. This performance was attributed to the synergy developed by the co-doping.Modified electronic properties of different metals could be achieved by alloy formation. Theoretical studies persuaded the utilization of more than one metal as the active site due to the difference in the interaction capability of reactant molecules with different metals. Importantly, ligand and strain effect due to the formation of hetero-atom bond and altered bond length improved the electrocatalytic properties of catalysts. Moreover, alloying an expensive metal with a non-precious metal could be cost-effective.The effect of alloy formation on eNRR was studied by Manjunatha et\u00a0al. by employing RuPt alloys as nitrogen reduction catalysts at ambient conditions [161]. The synergistic effect of two noble metal catalysts resulted in the improved interaction with the nitrogen molecule by subsequent electron addition to antibonding orbitals of nitrogen. An improved activity with an ammonia production rate of 5.1\u00a0\u00d7\u00a010\u22129\u00a0g\u00a0s\u22121\u00a0cm\u22122 and a FE of 13 % at \u22120.123\u00a0V vs. RHE was achieved. Similarly, a PdCu/RGO system was developed by the facile reduction method for eNRR [162]. The focus was on the improvement of intrinsic catalytic properties of non-noble transition metal by alloying with a small quantity of noble metal. The improved efficiency towards ammonia formation was achieved by Pd0\u00b72Cu0.8/RGO system at \u22120.2\u00a0V vs. RHE with a FE of less than 1 % and an ammonia yield rate of 2.80\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 and attributed to the enhanced intrinsic catalytic property by alloying. Bimodal palladium-copper (PdCu) alloys were also developed recently for eNRR evaluation (Fig.\u00a012c and d) [163]. Due to the interconnected porous structure and appropriate Pd/Cu ratio, an ammonia yield rate of 40\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved. Palladium-ruthenium (PdRu) with tripod structure was investigated for eNRR and a high yield of ammonia was achieved in 0.1\u00a0mol\u00a0L\u22121 KOH electrolyte [164]. The tripod structure along with bimetallic composition improved the overall ammonia production. Exposure of active sites in specific crystal orientations enhanced the intrinsic property of a catalyst.High-entropy alloys (HEAs) are a new class of multicomponent alloys that have found widespread application as electrocatalysts [59,129,165,166]. HEA design has already overcome the limitations of primitive alloy materials. Because of their variable element compositions, HEAs open a plethora of possibilities for the design of electrocatalysts (particularly multifunctional electrocatalysts). Zhang et\u00a0al. developed RuFeCoNiCu HEA nanoparticles as an electrocatalyst for eNRR [167]. The developed HEA nanoparticles demonstrated an ammonia yield of 57.1\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 with a FE of 38.5 %. Furthermore, these electrocatalysts demonstrated excellent electrochemical activity in other electrolytes as well as excellent stability for 100\u00a0h. This improved performance is attributed to a multi-site cooperative catalytic mechanism on the surface of HEA alloys.Creating defects and different vacancies results in the formation of coordinatively unsaturated sites in the structure of the catalyst that is more active due to localized charge density and free electrons [168]. Free electrons available on these sites could be back donated to the reactant molecules to activate and polarize them. For instance, oxygen vacancies (Vo) created in metal oxides provide a localized charge site that stabilizes and activates small molecules by accepting electronic density. Herein, we shed light on the role of vacancies and defects on the surface of the catalyst for nitrogen reduction.Improvement in surface oxygen vacancies of ferric oxide (Fe2O3) was achieved by calcination in an inert environment [169]. These modified structures were further evaluated for nitrogen reduction and an improvement in the eNRR performance was observed as compared to the non-modified sample. A FE of 6 % was achieved at \u22120.9\u00a0V vs. Ag/AgCl with an average ammonia production rate of 0.46\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122. Oxygen vacancy-dependent nitrogen reduction performance was also evaluated by Lv et\u00a0al. They synthesized ceria-assisted amorphous bismuth vanadate (Bi4V2O11) hybrid by electrospinning method (Fig.\u00a013\na and b) [170]. Due to the proper alignment of electronic bands, interfacial electron transfer was more plausible in this structure which enhanced the electrochemical performance of the material. Intrinsic oxygen vacancy (Vo), due to reduced vanadium provided the localized electron density that was proposed to be donated back to the nitrogen antibonding molecular orbitals for enhanced activation and polarization. A high average yield of ammonia (23\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121) was achieved with FE >10 % at \u22120.2\u00a0V vs. RHE. This could be attributed to the enhanced interaction of nitrogen with the catalyst surface and special active sites on the V atom due to its low coordination and high spin polarization after the creation of oxygen vacancies. In a recent report, Vo-derived tantalum oxide (Ta2O5) nanorods were proven to be an efficient catalyst for nitrogen reduction with an ammonia yield rate of 15.9\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 and a FE of 8.9 % [171]. As corroborated by the DFT studies, Vo supported the activation and polarization of reactant molecules by the transfer of electron density, hence improvement in the kinetics of the reaction obtained.Interestingly, atomic layers of Vo-derived MoO2 were developed by the CVD approach [172]. This Vo-modified catalyst exhibited a high selectivity and efficiency towards ammonia formation with a yield rate of 12.20\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121and a FE of 8.2 % at \u22120.15\u00a0V vs. RHE. Based on adsorption experiments and theoretical studies, enhanced interaction of nitrogen was observed. Tuning of Vo at the atomic level improves the performance by a complete utilization of active sites on the surface. Similarly, Vo-derived Cr-doped CeO2 was developed by a hydrothermal approach and evaluated for nitrogen reduction in 0.1\u00a0mol\u00a0L\u22121 Na2SO4 [173]. At \u22120.7\u00a0V vs. RHE, an ammonia yield of 16\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 with a FE of 4 % was obtained.In a recent report, Vo-derived TiO2 in situ grown on Ti3C2Tx was evaluated for the nitrogen reduction reaction [176]. Abundant surface defects and high conductivity with a large surface area were the main attributes of ammonia production. An ammonia yield of 32.17\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved at \u22120.55\u00a0V vs. RHE with a FE of 16 % at \u22120.45\u00a0V vs. RHE. DFT calculations confirmed the role of Ti edge atoms and Vo on eNRR performance in an associative distal mechanism. Similarly, Vo-derived CeO2 nanorods also exhibited improved eNRR performance as compared to a bulk counterpart in neutral media [177]. An ammonia yield rate of 16.4\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was obtained at \u22120.5\u00a0V vs. RHE, while a FE of 3.7 % was achieved at \u22120.4\u00a0V vs. RHE.Oxygen vacancy enhances the localized charge density on the surface of the catalyst and contributes to the overall performance of the catalyst. However, the role of oxygen vacancy concentration and stability in the improved activity of the catalyst is lacking in the literature. It is recommended to evaluate the structural evolution of a material with oxygen vacancy during the process to understand more about its role.The defect engineering strategy produced localized charge density sites due to electronic density redistribution that facilitated the enhanced nitrogen confinement at N-vacancy sites. Defective polymeric carbon nitride (PCN) was developed and evaluated for nitrogen reduction (Fig.\u00a013c and d) [174]. A FE of 11 % with an ammonia yield of 8\u00a0\u03bcg\u00a0h\u22121\u00a0cm\u22122 at \u22120.2\u00a0V was achieved. High performance could be attributed to the strong activation of nitrogen on the N-vacancy site as corroborated well with DFT calculations. Like transition metal catalysts, back donation of charges found in this system caused the activation of N\u2261N and rendered nitrogen reduction. Similarly, defective fluorographene was developed by a hydrothermal approach and evaluated for eNRR [178]. At \u22120.7\u00a0V vs. RHE, a FE of 4.2 % with an ammonia yield rate of 9.3\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 was achieved.Defect-rich MoS2 nanoflowers were also developed and evaluated for nitrogen reduction [179]. At \u22120.4\u00a0V vs. RHE, a FE of 8.3 % was achieved with a yield rate of 29.28\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121. As corroborated well with the DFT calculations, defects were the active sites for eNRR. Based on the Hydrogen model, due to the presence of defects, a bare Mo atom could activate a nitrogen molecule with an associative distal hydrogenation step that led to the production of ammonia. However, defects also showed strong interactions with ammonia making the desorption step more endothermic. In another theoretical-assisted experimental approach, a defective TiO2 catalyst was developed for eNRR [180]. It exhibited improved eNRR performance as compared to non-defective counterparts with an ammonia formation rate of 1.24\u00a0\u00d7\u00a010\u221210\u00a0mol\u00a0cm\u22122\u00a0s\u22121 and a FE of 9 % at \u22120.15\u00a0V vs. RHE.Overwhelming HER competition could be minimized by a variety of approaches. Importantly, surface modification was done by a hydrophobic coating of ZIFs on the noble metal catalyst (Ag\u2013Au) that suppressed the HER process and improved the performance of noble metals (Fig.\u00a013e and f) [175]. The hydrophobic nature and high gas sorption ability of ZIFs were the main attributes of high ammonia selectivity (90\u00a0%) over hydrogen production. Deposition of Ag-nanocubes on Au electrodes and subsequent coating of hydrophobic ZIFs by wet chemical deposition was performed. Confinement of electroactive species and the inhibition of water access by ZIFs film resulted in the 10\u00a0pmol\u00a0cm\u22122\u00a0s\u22121 rates of ammonia production at \u22122.9\u00a0V vs. Ag/AgCl with a FE of 18 % in an aprotic solvent (THF) containing lithium triflouromethanesulfonate as the electrolyte. Despite this approach, eNRR performance is still impeded by the HER process. Moreover, the role of an aprotic solvent, Li-ion incorporation, and flouro-based additives in the electrolyte need to explore more in this regard.Core-shell type nanostructures provide two different interaction sites depending on the core and the shell. The synergistic effect of core-shell morphology and composition resulted in the unique properties of the material [181\u2013184]. Mostly, the stability of reactive cores could be tuned up by a coating of the shell. Moreover, these core-shell type of structures also provides localized charge density sites for enhanced interaction with the reactant molecules.Core-shell type of \u03b1-Fe nanorods@Fe3O4 was grown on carbon fiber paper by the hydrothermal approach and evaluated for ammonia formation in the aprotic solvent-ILs mixture [185]. An enhancement in the eNRR activity was achieved with an ammonia production rate of 2.35\u00a0\u00d7\u00a010\u221211\u00a0mol\u00a0s\u22121\u00a0cm\u22122 and a FE of 32 %. A smoother energy profile was obtained by DFT calculations, highlighting the associative distal pathway as the most plausible mechanism for these structures. Au@CeO2 core-shell structures were developed by room temperature spontaneous redox approach with 3 % loading of Au [183]. Au cores below 10\u00a0nm were the main active sites for nitrogen reduction, but oxygen vacancies created on CeO2 also enhanced the localized charge density. In acidic media, this catalyst exhibited a FE of 9.5 % with an ammonia yield rate of 10.6\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 at \u22120.4\u00a0V vs. RHE.Electrochemical nitrogen reduction is regarded as a possible avenue for storing renewable energy in chemical bonds, as well as an environmentally acceptable alternative to the century-old Haber\u2013Bosch process. For the last few years, scientists have been working hard to improve the rate and efficiency of this reaction, and they have achieved significant progress. In this review, recent theoretical and experimental studies for eNRR are summarized to provide a fundamental basis for the rational design of an electrocatalyst. A detailed theoretical framework is explained that helps in the rational catalyst design based on different plausible mechanisms for eNRR. This framework involves the selection of an active material based on the reactant adsorption and desorption kinetics along with the elucidation of scaling relation by the adsorption and desorption of reaction intermediates. Furthermore, explorations related to the development of new catalytic compositions for nitrogen reduction could be achieved by the evaluation of the density of states and charge distribution. It has been deduced from theoretical studies and verified experimentally, that electrocatalytic nitrogen reduction on the surface of transition metal catalysts requires higher overpotentials than the hydrogen evolution reaction, which results in low FE and ammonia production. However, transition metal derivatives and single-atom catalysts are predicted by theoretical studies to be more active than their parent bulk metals for nitrogen reduction. Based on the intrinsic and extrinsic catalytic properties of a catalyst, several strategies are suggested by consideration of a theoretical-directed experimental approach. Different compositions designed by rational catalyst design showed enhanced FE for ammonia including transition metal oxides, carbides, sulfides, and metal-free carbon-based materials. Synergism between theoretical and experimental studies will lead to a better understanding of the mechanism that eventually leads to rational catalyst design.Based on our understanding of this reaction, herein, we present several outlooks for the development of this field.\n\n(1)\n\nCatalyst development methods: In the literature, there is a lack of reports of highly active and efficient catalysts for this process. This may be attributed to the limitations related to the catalyst development method. For instance, achieving homogeneity in the catalyst structure is the biggest challenge in several catalytic methods. Importantly, electrochemical reactions are surface-based interactions of reactant moieties with the catalyst surface on flat electrodes that limit access to deeper catalytic content. Several methods are hard to scale up for industrial-scale production of the catalyst. Moreover, challenges related to the optimization of reaction parameters hamper the production of active and efficient catalysts. In this regard, the vital paradigm shift in the chemistry of this process is possible by theory-guided rational catalyst design. Thus, elucidation of the mechanism by theoretical studies is required, especially those related to heterogeneous catalyst surfaces and atomic-level electrocatalysis. More precise control of the content, structure, and active sites is required. Although external dopants, intrinsic defects, and alloy formation have shown to be useful in altering the electronic structure, accurately controlling the appropriate doping site/type, defect quantity, and alloy content, which is important to determine the catalytic activity, remains a challenge. It is desirable to have new catalyst development methods that are facile and sustainable to control the desired target catalytic properties. Moreover, advanced strategies for catalyst development that have been employed at the lab scale need optimization to work at a larger scale. Finally, artificial intelligence and machine learning might improve the search for new efficient electrocatalysts.\n\n\n(2)\n\nSelectivity and rate of ammonia production: Selectivity is one of the challenges reported for eNRR in aqueous media that is attributed to the enhanced tendency of transition metals for the adsorption of hydrogen as compared to the nitrogen moiety. Moreover, the competition for adsorption on the surface of the catalyst between reactants and different reaction intermediates resulted in low rates of ammonia production. Though an improved FE had been reported by rational catalyst design strategy, the overall rate of ammonia production was low (>10\u22129\u00a0mol\u00a0s\u22121\u00a0cm\u22122). It is highly desirable to put effort into the improvement of the rate of this process after the selection of an efficient catalyst. Furthermore, catalyst stability and durability for long-time operation are also important to make this process economically feasible. For instance, surface reconstruction of the catalyst has gained attention in the recent past during the electrolysis at high reductive/oxidative potentials [186\u2013190]. As a result, during electrocatalytic activities, the surface structures and compositions of catalysts are dynamically reconstructed. With advances in in-situ and operando techniques, it has been discovered that during electrolysis, electrocatalysts undergo surface reconstruction to form the actual active species, accompanied by a change in their oxidation state. As a result, establishing unambiguous structure-composition-property relationships in the pursuit of high-efficiency electrocatalysts requires a thorough understanding of the surface reconstruction process.\n\n\n\n\n(3) In-situ characterization and product quantification: It is highly desirable to employ a rigorous protocol for the measurement of the amount of ammonia produced from dinitrogen by eNRR to confirm the exact origin of the N-source. Andersen et\u00a0al. reported a benchmarking protocol for the eNRR evaluation at ambient conditions [5]. As illustrated in Fig.\u00a014\n, the amount of ammonia and other nitrogen-containing compounds in the setup due to background contamination needs screening. If the contamination levels measured are within an order of magnitude of the ammonia produced different measures are required to eliminate the interference including cleaning of membrane, electrochemical setup, and the developed material. Moreover, quantitative isotope-sensitive measurements of produced ammonia are recommended to validate the exact origin of the N-source. In-situ and operando characterization tools are crucial to closing the gap between mechanistic understanding and the performance of the catalyst for eNRR. Real-time mechanistic evaluations are recommended for this process to understand challenges and explore solutions. This will improve the deeper understanding of this process in real-time and a correlation between the structure to performance of the catalyst will be developed. Elucidation of electrochemical reactions during nitrogen reduction regarding kinetics and solid-liquid interface is the need of time. Moreover, ammonia detection techniques are required for the rigorous validation of this process. In this regard, validated electrochemical, spectroscopic, and chromatographic techniques with high sensitivity are required.\n\n\n(4)\n\nDevice fabrication: The sluggish rate of NH3 production needs improvement with high current densities to be used as a commercial process for eNRR at ambient conditions. Such kinetic studies related to mass and electron transfer mechanisms are particularly significant. For the large-scale application of this process, a device is required that can work in harsh conditions to produce high rates of ammonia. In this regard, a gas diffusion electrode-based flow cell type eNRR electrolyzer is expected that can operate at high current density.\n\n\n\nCatalyst development methods: In the literature, there is a lack of reports of highly active and efficient catalysts for this process. This may be attributed to the limitations related to the catalyst development method. For instance, achieving homogeneity in the catalyst structure is the biggest challenge in several catalytic methods. Importantly, electrochemical reactions are surface-based interactions of reactant moieties with the catalyst surface on flat electrodes that limit access to deeper catalytic content. Several methods are hard to scale up for industrial-scale production of the catalyst. Moreover, challenges related to the optimization of reaction parameters hamper the production of active and efficient catalysts. In this regard, the vital paradigm shift in the chemistry of this process is possible by theory-guided rational catalyst design. Thus, elucidation of the mechanism by theoretical studies is required, especially those related to heterogeneous catalyst surfaces and atomic-level electrocatalysis. More precise control of the content, structure, and active sites is required. Although external dopants, intrinsic defects, and alloy formation have shown to be useful in altering the electronic structure, accurately controlling the appropriate doping site/type, defect quantity, and alloy content, which is important to determine the catalytic activity, remains a challenge. It is desirable to have new catalyst development methods that are facile and sustainable to control the desired target catalytic properties. Moreover, advanced strategies for catalyst development that have been employed at the lab scale need optimization to work at a larger scale. Finally, artificial intelligence and machine learning might improve the search for new efficient electrocatalysts.\nSelectivity and rate of ammonia production: Selectivity is one of the challenges reported for eNRR in aqueous media that is attributed to the enhanced tendency of transition metals for the adsorption of hydrogen as compared to the nitrogen moiety. Moreover, the competition for adsorption on the surface of the catalyst between reactants and different reaction intermediates resulted in low rates of ammonia production. Though an improved FE had been reported by rational catalyst design strategy, the overall rate of ammonia production was low (>10\u22129\u00a0mol\u00a0s\u22121\u00a0cm\u22122). It is highly desirable to put effort into the improvement of the rate of this process after the selection of an efficient catalyst. Furthermore, catalyst stability and durability for long-time operation are also important to make this process economically feasible. For instance, surface reconstruction of the catalyst has gained attention in the recent past during the electrolysis at high reductive/oxidative potentials [186\u2013190]. As a result, during electrocatalytic activities, the surface structures and compositions of catalysts are dynamically reconstructed. With advances in in-situ and operando techniques, it has been discovered that during electrolysis, electrocatalysts undergo surface reconstruction to form the actual active species, accompanied by a change in their oxidation state. As a result, establishing unambiguous structure-composition-property relationships in the pursuit of high-efficiency electrocatalysts requires a thorough understanding of the surface reconstruction process.\n\n(3) In-situ characterization and product quantification: It is highly desirable to employ a rigorous protocol for the measurement of the amount of ammonia produced from dinitrogen by eNRR to confirm the exact origin of the N-source. Andersen et\u00a0al. reported a benchmarking protocol for the eNRR evaluation at ambient conditions [5]. As illustrated in Fig.\u00a014\n, the amount of ammonia and other nitrogen-containing compounds in the setup due to background contamination needs screening. If the contamination levels measured are within an order of magnitude of the ammonia produced different measures are required to eliminate the interference including cleaning of membrane, electrochemical setup, and the developed material. Moreover, quantitative isotope-sensitive measurements of produced ammonia are recommended to validate the exact origin of the N-source. In-situ and operando characterization tools are crucial to closing the gap between mechanistic understanding and the performance of the catalyst for eNRR. Real-time mechanistic evaluations are recommended for this process to understand challenges and explore solutions. This will improve the deeper understanding of this process in real-time and a correlation between the structure to performance of the catalyst will be developed. Elucidation of electrochemical reactions during nitrogen reduction regarding kinetics and solid-liquid interface is the need of time. Moreover, ammonia detection techniques are required for the rigorous validation of this process. In this regard, validated electrochemical, spectroscopic, and chromatographic techniques with high sensitivity are required.\nDevice fabrication: The sluggish rate of NH3 production needs improvement with high current densities to be used as a commercial process for eNRR at ambient conditions. Such kinetic studies related to mass and electron transfer mechanisms are particularly significant. For the large-scale application of this process, a device is required that can work in harsh conditions to produce high rates of ammonia. In this regard, a gas diffusion electrode-based flow cell type eNRR electrolyzer is expected that can operate at high current density.The authors declare no conflict of interest.The study is supported by Australian Research Council (DP210103892). C.Z. also thanks Australian Research Council for the award of Future Fellowship (FT170100224).", "descript": "\n                  Ammonia (NH3), a carbon-free hydrogen carrier, is an important commodity for the food supply chain owing to its high energy capacity and ease of storage and transport. The Haber\u2013Bosch process is currently the favored industrial method for large-scale ammonia production but requires energy-intensive and sophisticated infrastructure which hampers its utilization in a sustainable and decentralized system of manufacture. The electrochemical nitrogen reduction reaction (eNRR) at ambient conditions holds great potential for sustainable production of ammonia using electricity generated from renewable energy sources such as solar and wind. However, this approach is limited by a low rate of ammonia production with high overpotential and the competing hydrogen evolution reaction (HER). For a better understanding and utilization of eNRR as a sustainable process, insight into rational catalyst design and mechanistic evaluations by a theoretically-directed experimental approach is imperative. Herein, recent insights into rational catalyst design and mechanisms, based on intrinsic and extrinsic catalytic activity are articulated. Following the elucidation of basic principles and mechanisms, a framework supplied by theoretical studies that lead to the optimal selection and development of eNRR catalysts is presented. Following a discussion of recently developed electrocatalysts for eNRR, we outline various recently-used theoretical and experimental methodologies to improve the intrinsic and extrinsic catalytic activity of advanced electrocatalysts. This review is anticipated to contribute to the development of active, selective, and efficient catalysts for nitrogen reduction.\n               "}
{"full_text": "The photocatalytic process is one of the cleanest technologies from an environmental perspective. It has been investigated intensively over the last two decades because of its potential application in waste treatment and the production of new sources of energy [1,2]. TiO2 is one of the most attractive UV-photocatalysts with application in many fields, especially wastewater treatment [3,4]. However, the photocatalytic efficiency of TiO2 is prone to decline over time, mainly due to its limited light-gathering capacity and the easy recombination of electron\u2013hole pairs [3,5].Using sunlight as the energy source for the photocatalytic reaction is a key target in studies of photocatalysis. The development of small band gap photocatalysts is a promising way to approach this target and has been the subject of considerable interest in recent years. A suitable high-efficiency semiconductor for visible light photocatalysts needs a band gap (E\nG\u00a0<\u00a03.0\u00a0eV) that is sufficiently narrow to harvest visible light but is large enough (E\nG\u00a0>\u00a01.23\u00a0eV) to provide energetic electrons [6]. Further, a significant reduction in the band gap energy enhances the recombination of electrons and holes, dictating an optimum value for band gap energy [7]. Other important parameters determining photocatalytic efficiency are the morphology, crystal structure, and particle size of the photocatalyst [8]. It has been reported that the photocatalytic activity of semiconductors can be improved by increasing crystallinity and enhancing the specific surface area [9].The main ways of improving the performance of existing materials are: ion-doping or defect modification [10], fabrication of a composite material consisting of a photocatalyst and a highly conductive material to suppress the recombination of electron\u2013hole pairs by enabling fast photogenerated electron transfer [11], and creation of heterojunctions between two semiconductors with different band structures to inhibit electron\u2013hole recombination and improve photocatalytic performance [12]. Recent studies have used different approaches to improving the photocatalytic performance of UV\u2013TiO2 catalysts, including doping with metal ions (Ag, Fe, V, Au, Pt, Ni, Co, Cu, Nb) [13,14], non-metal (N, S, C, B, P, I, F) [13,15], mixed oxides with p\u2013n junction characteristics [16,17] and combining TiO2 with smaller band gap energy semiconductors, such as WO3, SiC, Cu(OH)2, CuOx, Ni(OH)2, NiO, Si, CdS, and SrTiO3 [18\u201320].Because of their high level of photocatalytic activity under UV irradiation and visible light, perovskite-type oxides, such as tantalates and titanates, have recently attracted much attention [21,22]. Ti-based materials, and especially perovskite titanate ATiO3 (A\u00a0=\u00a0Ba, Sr, Fe, and Ni), are promising new photocatalysts with notable advantages [23] that have been the subject of intensive research [24,25]. Among perovskite titanates, NiTiO3 has an ilmenite-type crystal structure in which both Ni and Ti are in octahedral coordination, with alternating cation layers occupied by Ni and Ti [26]. NiTiO3 has attracted considerable attention due to its superior photocatalytic and electro-optical properties and low dielectric constant [25,27]. NiTiO3 is an n-type semiconductor material with antiferromagnetic properties [28] is highly stable in an oxidizing environment and under light irradiation. In addition, NiTiO3 has an optical absorption spectrum with band gap energy of around 2.2\u00a0eV, which means that it has excellent potential for use in photocatalytic applications in visible light [28]. There have been numerous reports on the investigation of the photocatalytic activity of NiTiO3, including degradation of Tergitol, Safranine T [29], and Rhodamine B [30]. The photobleaching of methyl orange by NiTiO3 confirms its good photocatalytic properties in visible light [31]. However, its low band gap energy reduces its quantum efficiency when used as an individual photocatalyst [32].An effective strategy for improving the photocatalytic performance of TiO2 is to combine it with other semiconductors to form a heterostructured photocatalyst [18,33]. TiO2-based heterostructure systems, i.e. SrTiO3\u2013TiO2 and BaTiO3\u2013TiO2, have been shown to promote cation charge and hole transport, with a narrowing of the band gap energy of TiO2 and enhanced electron\u2013hole pair separation [20,34], which improve photochemical efficiency. However, both BaTiO3 and SrTiO3 have relatively large band gap energy of about 3.4\u00a0eV, limiting the utilization of sunlight sources. TiO2-coupled nickel titanate has recently been reported as an efficient photocatalyst in visible light for the decomposition of methylene blue [35]. NiO/NiTiO3 composites present a higher degradation rate than pure NiTiO3 regardless of the amount of NiO present [36].The doping of TiO2 with NiTiO3 is a promising solution for improving the photocatalytic performance of TiO2 and expanding absorption to the visible light region [27]. However, few studies have used NiTiO3-doped TiO2 catalysts for wastewater treatment, especially the treatment of persistent organic compounds. Therefore, it is desirable to develop NiTiO3-doped TiO2 photocatalysts for this purpose. In this study, TiO2 catalysts doped with various NiTiO3 concentrations were prepared by combining the sol\u2013gel and the hydrothermal method using water as an eco-friendly solvent in the synthesis process. The physicochemical and photochemical properties of doped catalysts were determined using modern physico-chemical analysis techniques. In addition, using cinnamic acid (CA) as the representative persistent compound, we further demonstrated that the modified NiTiO3\u2013TiO2 catalysts are excellent UV-A-photocatalysts for the photodegradation of persistent organic pollutants.NiTiO3 perovskite was synthesized by the sol\u2013gel method. First, 2.91\u00a0g of nickel nitrate hexahydrate (Ni(NO3)2\u00b76H2O, Merck, >99%) and 2.10\u00a0g of citric acid monohydrate (C6H8O7\u00b7H2O, Merck, >99%) were dissolved with 5\u00a0mL of ethanol absolute (C2H5OH, Merck, >95%) and stirred at 300\u00a0rpm for 1\u00a0h to form a homogeneous mixture. Next, 3\u00a0mL of titanium (IV) isopropoxide (Ti(OC3H7)4, Sigma-Aldrich, >97%) was added drop by drop while stirring and then held for 1\u00a0h to form a transparent sol mixture. The synthetic sol was heated slowly to 60\u00a0\u00b0C and dried for 24\u00a0h to produce a bright green puffy porous gel. Finally, the porous gel was heated at 700\u00a0\u00b0C for 2\u00a0h to form the required NiTiO3 nanoparticles (NTO).NiTiO3\u2013TiO2 mixed catalysts with different NiTiO3 concentrations were synthesized in an eco-friendly neutral medium. First, 3\u00a0mL of Ti(OC3H7)4 was added drop by drop to the distilled water and stirred for 30\u00a0min at 300\u00a0rpm. Next, m grams of NiTiO3 synthesized in the above step were added to the solution. The mixture was then hydrated in a steel-lined Teflon container at 160\u00a0\u00b0C for 12\u00a0h. The solids were then filtered and washed three times with distilled water and ethanol. Subsequently, the solids were dried at 60\u00a0\u00b0C for 12\u00a0h to obtain NiTiO3\u2013TiO2 mixed catalysts. The catalysts were denoted as xNTO\u2013Ti, representing the NTO content of the mixed catalysts, x\u00a0=\u00a00.5, 1.0, 1.5, and 2.0\u00a0wt.%.The physicochemical characteristics of the samples were studied using \u03a7-ray diffraction (\u03a7RD), Raman spectroscopy, scanning electron microscopy (SEM), Fourier-transform infrared spectroscopy (FT-IR), Brunauer\u2013Emmett\u2013Teller adsorption (BET), energy dispersive X-ray (EDX), UV\u2013vis absorption spectra techniques, and point of zero charges (PZC). The methods are detailed in our previous study [37].In this study, CA, a recalcitrant phenolic acid, was selected as the representative agent for the persistent organic matter to be researched. The photocatalytic activity of the samples was studied using the batch method for the photodecomposition of CA, as described in our previous study [37]. The effects of the NiTiO3 loading (0.5, 1.0, 1.5, and 2.0\u00a0wt.%) in the modified samples on CA photodegradation were investigated, along with the operating parameters including, the airflow rate (0.2, 0.3, 0.4, and 0.5\u00a0L/min), the catalyst dosage (0.75, 1.00, 1.25, 1.50 and 1.75\u00a0g/L), and the initial pH of the solution (3.8, 5.0, and 7.0). The reaction temperature was fixed at 25\u00a0\u00b0C by heat exchange with cyclic cold water. The initial pH of the solution was adjusted using a buffer solution pH\u00a0=\u00a010\u00a0\u00b1\u00a00.01.The XRD patterns of the catalysts are presented in Fig.\u00a01\n. The XRD pattern of the TiO2 synthesized in the neutral medium of water contained the characteristic peaks of the anatase phase at 2\u03b8\u00a0=\u00a025.6, 38.2, 48.3 and 54.7\u00b0 (JCPDS 21-1272), with the strongest intensity at 2\u03b8\u00a0=\u00a025.6\u00b0 [37]. The XRD patterns of the NTO\u2013Ti samples contained the characteristic peaks of the high crystallinity anatase phase at 2\u03b8\u00a0=\u00a025.4\u00b0; 38.1\u00b0; 48.3\u00b0; 54.3\u00b0; and 62.6\u00b0. In addition, a small amount of rutile (110) at 2\u03b8\u00a0=\u00a027.4 (JCPDS card 21-1276) [38] and brookite crystalline phases of TiO2 appear at 31.3\u00b0 [39]. On samples 0.5NTO\u2013Ti and 1.0NTO\u2013Ti, almost no characteristic peaks of NiTiO3 appear because of the relatively low concentration; this proves that NiTiO3 is well dispersed on the TiO2 surface of these samples (as can be seen from SEM images in Fig.\u00a0S1). However, the diffraction planes (102), (104), (110), (116), (214) and (300) situated at 2\u03b8\u00a0=\u00a023.9\u00b0; 32.8\u00b0; 35.4\u00b0; 53.4\u00b0; 61.9\u00b0; and 63.5\u00b0, respectively, reveal the single ilmenite phase NiTiO3 [38] corresponding to JCPDS 75-3757 [40] and were observed with the 1.5NTO\u2013Ti and 2.0NTO\u2013Ti samples that had higher concentrations of NiTiO3. No peaks corresponding to the NiO (111), (200), (220) crystal plane, respectively at 2\u03b8\u00a0=\u00a037.2\u00b0, 43.3\u00b0, and 62.8\u00b0 (JCPDS card no. 47-1049) [36,38] were observed, indicating that Ni exists solely in the structure of the NiO6 octahedra of NiTiO3. In the NTO\u2013Ti modified sample, no shift in the peak characteristics of the phases of TiO2 relative to the single TiO2 was observed, whereas a right-shift in the XRD peaks of NiTiO3 compared to pure NiTiO3 can be clearly seen; this is evidence of the fusion of NTO and TiO2.Based on the XRD results at 2\u03b8\u00a0=\u00a025.4\u00b0 of the (101) anatase TiO2 plane, the average crystal size of the catalysts is calculated by the Debye\u2013Scherrer equation with K\u00a0=\u00a00.94 to be 34.8\u00a0nm, 8.6\u00a0nm, 9.0\u00a0nm, 9.0\u00a0nm and 8.7\u00a0nm for TiO2 and the 0.5NTO\u2013Ti, 1.0NTO\u2013Ti, 1.5NTO\u2013Ti, and 2.0NTO\u2013Ti samples, respectively. It appears that the addition of NTO significantly reduces the crystal size of TiO2 and simultaneously enables the transfer of a small amount of the anatase phase to rutile and brookite phases under hydrothermal conditions at 160\u00a0\u00b0C.Additional information regarding the crystalline structure was obtained through Raman spectroscopy (Fig.\u00a02\na). On the bare TiO2 sample synthesized in the neutral medium of water (denoted as Ti(w)), the peaks of the anatase phase are observed at 147, 402, 517 and 639\u00a0cm\u22121. For the NTO\u2013Ti samples, the characteristic peaks of TiO2 anatase still appear, with a clear shift relative to pure TiO2, specifically 2\u03b8\u00a0=\u00a0146, 198, 320, 398, 515, 640\u00a0cm\u22121. The peaks at 146,198, 640\u00a0cm\u22121 with the strongest signal intensity at 146\u00a0cm\u22121 are characteristic of the E\ng spectrum; 398\u00a0cm\u22121 is attributed to the B\n1g spectrum, and 515\u00a0cm\u22121 is assigned to A\n1g. E\ng is the asymmetric stretching vibration, while B\n1g and A\ng correspond to the asymmetric and symmetric bending vibration of TiO2 anatase. In addition, Raman spectra also detected the E\ng asymmetric stretching vibration of the rutile crystalline phase at wavenumber 238\u00a0cm\u22121 [41]. Meanwhile, peaks at wavenumbers 246, 284, 351, 552, and 706\u00a0cm\u22121 are typical of the appearance of NiTiO3 [42]. A weak Raman peak is observed at around 706\u00a0cm\u22121 in the NTO\u2013Ti modified catalysts, which can be ascribed to the NTO phase with a hexagonal structure [43]. The absence of wavenumber 547\u00a0cm\u22121 originating from Ni\u2013O bonds in the Raman spectrum suggests that almost all the Ni had been utilized to form NiTiO3 nanocrystallites [44], confirming the results of the XRD analysis. The Raman spectra indicated that adding of NTO to TiO2 leads to a shift in the oscillation peaks of TiO2 and NTO relative to bare TiO2 and NTO. These findings indicate a strong interaction between NTO and TiO2, resulting in the fusion of NTO into the structure of TiO2.The FTIR spectra of the NTO\u2013Ti samples recorded in the wavenumber region 400\u20134000\u00a0cm\u22121 are shown in Fig.\u00a02b. It can be observed that there are three main absorption regions for all the samples, in the ranges 2500\u20133600\u00a0cm\u22121, 1500\u20131650\u00a0cm\u22121, and 400\u2013800\u00a0cm\u22121, respectively. The peaks at 2500\u20133600\u00a0cm\u22121 with the highest intensity at 3405\u00a0cm\u22121 are typical of stretching vibration of \u2013OH in the hydroxyl group [45]. It follows from Fig.\u00a02b that the addition of NTO into TiO2 leads to a reduction in the intensity of the OH groups; further, the greater the amount of perovskite added, the greater the production in the hydroxyl groups. The characteristic bands discovered at 1500\u20131650\u00a0cm\u22121 are H\u2013OH groups adsorbed on the catalyst surface. Previous studies have shown that the presence of OH and H\u2013OH groups has an essential role in photocatalysis since the hydroxyl group present on the catalyst surface can react with holes to form hydroxyl radicals [46]. The peaks at about 1560 and 1395\u00a0cm\u22121 are due to the N\u2013O bond vibration of NO3\n\u2212 and carboxyl vibration, respectively [47]. The typical bands in the 400\u2013800\u00a0cm\u22121 range correspond to the vibrations of the metal\u2013oxygen bond [40]; the absorption peak near 440\u00a0cm\u22121 corresponds to the Ti\u2013O\u2013Ni bond [48] and the strong absorption bands at 450 and 565\u00a0cm\u22121 correspond to the stretching vibrations of Ti\u2013O and bending vibrations of O\u2013Ti\u2013O, respectively. The absorption bands of the Ti\u2013O octahedral appear at 675 and 500\u00a0cm\u22121, corresponding to the formation of the NiTiO3 phase [47]. The characteristic bands of the carbonate (867 and 1067\u00a0cm\u22121) do not appear in the spectrum, which means that the synthesized crystals are carbonate-free [49]. Therefore, this result demonstrates the successful synthesis of the NTO-mixed TiO2 catalyst.On the SEM image (Fig.\u00a0S1(a)), the clumps of spherical-like TiO2 particles size of 10\u201340\u00a0nm can be observed in pure TiO2 sample, while NiTiO3 (Fig.\u00a0S1(b)) appears in the form of bigger bright spherical granules, 20\u201350\u00a0nm in size. On mixed samples the interspersing dark and light particles of few nm in size can be seen. As the concentration of NTO in the mixture catalyst increased, the density of light-colored particles, attributed to NTO, increased. However, the particle sizes of the three samples containing 0.5\u20131.5% NTO were approximately the same, at about 5\u201310\u00a0nm. From the SEM analysis results, it can be suggested that in the mixed samples TiO2 and NiTiO3 exist in form of smaller separate particles, which is consistent with the results of the XRD analysis above.The N2 adsorption/desorption isotherms of Ti(w) and the 1.0NTO\u2013Ti samples (Fig.\u00a03\n) exhibit a type IV(a) isotherm curve, based on the classification by the International Union of Pure and Applied Chemistry [50]. In this type, the initial monolayer\u2013multilayer adsorption on the mesopore walls is followed by pore condensation. In the adsorption isotherm of the TiO2 sample, the adsorption branch contains a low slope region, which is associated with multilayer adsorption on pore walls and a narrow hysteresis loop, which are indicative of a narrow distribution of uniform mesopores and limited networking effects [50]. Further, the adsorption branch ends with a plateau region, indicating that mesopores are completely filled and macroporosity is non-existent. Based on these features, the isotherm for the bare TiO2 sample is type IV(a) with an H1 hysteresis loop [51]. The H1 hysteresis has been found in networks of ink-bottle pores where the width of the neck is similar to the width of the pore/cavity [50] or corresponds to cylindrical pores with openings at both ends [52].The shape of the adsorption/desorption isotherm of the 1.0NTO\u2013Ti sample is consistent with types H2(b) or H3. However, there is no appearance of the sharp step-down of the desorption branch, typical of the H3 type, so it can be concluded that the adsorption/desorption isotherm in this case is type H2(b) [50]. The type H2(b) loop is associated with pore blocking and indicates that the sample contains ink bottle-like pores; however, adsorption of N2 in this sample takes place to some extent in the monolayer region and is much stronger at higher p/p\no values (p/p\no\u00a0>\u00a00.6), indicating that the neck size is large. Thus, in both samples, there are networks of ink-bottle pores where the widths of the neck and the pore/cavity are similar. The pore diameters of the bare TiO2 and NTO\u2013Ti mixed catalysts were determined to be around 20\u00a0\u00c5 (Fig.\u00a03 and Table\u00a01\n). The existence of ink bottle-like pores in TiO2 and the NTO\u2013Ti samples is beneficial to the adsorption of oil and gas [52].The hysteresis loop type H2(b) of the NTO\u2013Ti mixed sample [51] exhibits more complex pore structures in which network effects are important. Type H2 is characterized by a more random distribution of pores and an interconnected pore system in the heterostructure sample [53] that affects the textural properties of the NTO\u2013Ti catalysts, such as surface area, pore volume, and pore size. Indeed, the specific surface area of the NTO\u2013Ti samples ranges from 142.8 to 163.2\u00a0m2/g, while the pore volume is in the range 0.086\u20130.182\u00a0cm3/g and pore size is in the range 17.6\u201323.4\u00a0\u00c5. It appears that the size and dispersion of the nanostructures obtained by the hydrothermal process depend on the NiTiO3 loading. Of the catalysts investigated, the largest pore size (23.4\u00a0\u00c5) and pore volume (0.182\u00a0cm3/g) are found in the 1.0NTO\u2013Ti sample, which provides superior adsorption capacity compared to the other catalysts. The values of specific surface area (159.7\u00a0m2/g), pore volume (0.178\u00a0cm3/g), and pore size (22.8\u00a0\u00c5) of pure TiO2 are within the variation range of the corresponding quantities of the NTO\u2013Ti samples. The textural properties of the 1.0NTO\u2013Ti and bare TiO2 catalysts are approximately the same, showing that the structure distribution in the 1% NTO\u2013Ti sample is the most favorable. The pore volume and pore size of the 1.0NTO\u2013Ti sample are also comparable to those of TiO2 and are higher than those of the other two mixed samples, as can be seen in Table\u00a01. The relatively high specific surface area of the samples may be related to the loose assembly of the primary nano-sized spherical-like particles to form porous secondary particles 30\u2013100\u00a0nm in size, as observed in the FE-SEM images (Fig.\u00a0S1).The elemental compositions of the 1.0NTO\u2013Ti sample were determined by EDS analysis. The EDX results (Fig.\u00a0S2) reveal the presence of Ti, O, and Ni in the NTO\u2013Ti sample. This result indicates the formation of a pure 1.0NTO\u2013Ti powder without impurities. Fig.\u00a0S2a shows that the distributions of the Ti, O, and Ni elements in this sample are synchronous. Meanwhile, Fig.\u00a0S2b shows the appearance of Ti at energy levels 0.28, 0.42, 4.51, and 4.92\u00a0keV; Ni at 0.95 and 5.62\u00a0keV; and O at the highest energy level of 0.53\u00a0keV. The analytical mass percentages of O, Ti, and Ni elements are 39.01%, 60.75%, and 0.24%, approximating to their theoretical mass ratios (O: 39.83%; Ti: 59.45%; Ni: 0.72%). The results confirm the formation of the NTO\u2013Ti material. The mass percent of Ni is lower than the theoretical value because more Ni2+ ions are lost due to the higher solubility of Ni-containing precipitate and the filtering and washing processes [44].The optical absorption characteristics of the NTO\u2013Ti catalysts were estimated by UV\u2013vis diffuse reflectance spectroscopy at ambient temperature; the values for band gap energy (E\nG) were determined using the Tauc formula [54]. The band gap energy of NiTiO3 was found at 2.18\u00a0eV (corresponding to visible light at 560\u00a0nm) due to Ni2+\u00a0\u2192\u00a0Ti4+ charge-transfer bands [55]. The addition of NiTiO3 to TiO2 reduced the band gap energy from 3.2\u00a0eV (for anatase TiO2) to 3.02\u20133.08 depending on the NiTiO3 loading, corresponding to the extension of light absorption to the visible region (404\u2013412\u00a0nm versus 385\u00a0nm) (Fig.\u00a04\n). This effect can be explained as follows. Adding NTO to TiO2 makes the phase composition of the TiO2 to be changed changes, comprising anatase (E\nG\u00a0=\u00a03.2\u00a0eV), rutile (E\nG\u00a0=\u00a03.1\u00a0eV) [1] and brookite (E\nG\u00a0=\u00a01.86\u00a0eV) [56] crystalline phases simultaneously, as shown in the XRD analysis, while bare TiO2 exists solely in the anatase phase. In addition, the absorption band at 404\u2013412\u00a0nm is also due to the O2\u2212\u00a0\u2192\u00a0Ti4+ charge transfer band of NiTiO3 [55]. The energy bands of the NiTiO3 align with the TiO2 to construct a heterojunction, as noted by Yue-Ying Li [41]. When NiTiO3 and TiO2 are excited by light illumination, the heterojunction promotes the photogenerated electrons of the conduction band to migrate from NiTiO3 to TiO2, with a simultaneous flow of holes from TiO2 to NiTiO3. As a result of these factors, the band gap energy of the mixed TiO2 sample reduces from 3.2\u00a0eV to a value of 3.02\u20133.08\u00a0eV. As shown in Fig.\u00a04, increasing the NiTiO3 content from 0.5 to 1.0\u00a0wt.% leads to a redshift in the absorption band from 404 to 412\u00a0nm and reduces the band gap from 3.08\u00a0eV to 3.02\u00a0eV, approaching the upper threshold of the desired band gap range (3.0\u00a0eV) [6]. Meanwhile, the absorption wavelength reduces to 407\u00a0nm and the band gap energy increased slightly to 3.05\u00a0eV when the NiTiO3 concentration is increased further, to 1.5\u00a0wt.%. This result is consistent with a prior study [14]. Clearly, the enhanced absorption of the visible region may lead to higher utilization performance during the photodegradation process for organic compounds.The conversion of CA during a 120\u00a0min reaction using NTO\u2013Ti catalysts with different NTO concentrations is shown in Fig.\u00a05\n. It was found that pure NiTiO3 exhibits very low activity in the photodegradation of CA, with 120-min\u00a0removal efficiency (X120) of 3.8%, while the value of bare TiO2 is 68.7%. Similar results were obtained in the study of Li [44], which reported that NiTiO3 exhibits much lower catalytic activity for methylene blue photo-degradation than Degussa P-25 (10% methylene blue degraded in 80\u00a0min on NTO versus 100% in 20\u00a0min on P-25). TiO2 consists of TiO6 octahedra, and both valence band and conduction band consist of hybridized O 2p and Ti 3d orbitals. Thus, under UV light, most charge-transfer transitions in TiO2 are O2\u2212\u00a0\u2192\u00a0Ti4+. TiO2 absorbs little visible light but exhibits high photocatalytic activity under UV light [57]. NiTiO3 has a narrower band gap than TiO2, but the crystal structure of NiTiO3 consists of alternating NiO6 and TiO6 layers and induces a wide energy gap from the hybridized Ni 3d and O 2p orbitals to the predominant Ti 3d orbitals, blocking both Ni2+\u00a0\u2192\u00a0Ti4+ and O 2p\u00a0\u2192\u00a0Ti 3d charge-transfer transitions. This effect leads to the low photocatalytic performance of NiTiO3 [44].The results in Fig.\u00a05 show that by increasing NTO concentrations from 0.5 to 1.0\u00a0wt.%, CA removal efficiency increases significantly; the decomposition at 30\u00a0min increased from 32.5% to 49.2% and rises from 59.9% to 82.8% after 120\u00a0min. However, when the NTO concentration is increased to 1.5% and 2.0\u00a0wt.%, CA treatment efficiency decreases sharply; the 30\u00a0min conversion of CA is only 38.5 and 41.9%, respectively, while CA conversion after 120\u00a0min is 65.2 and 69.7% in the same conditions. Among the NTO\u2013Ti catalysts, 0.5NTO\u2013Ti has the least activity and is the sample with the smallest specific surface area, pore volume, and pore dimension. In addition, this sample also exhibits the lowest intensity of OH groups and adsorbed water and the greatest band gap energy, which is unfavorable in the photocatalytic reaction [58]. The superior properties, namely the largest pore diameter and pore volume, the smallest particle size, the highest density of OH groups and adsorbed water and the smallest band gap energy, were responsible for the outstanding performance of the catalyst contained 1.0% NTO. The CA conversion efficiency for the 1.0NTO\u2013Ti sample was also much higher than that of pure TiO2 (X120\u00a0=\u00a082.8% versus 68.7%). Therefore, adding NTO to TiO2 significantly improves the physicochemical and photochemical properties of the photocatalyst by reducing the crystal size, narrowing the band gap energy, enhancing UV-A light absorption and converting part of the TiO2 anatase phase to rutile that enhances the photocatalytic performance of the heterostructure catalyst. However, due to the very low photocatalytic activity of NiTiO3, the high loading of NTO content (1.5% and more) leads to a decrease in the activity of the resulting catalyst.The effects of the reaction conditions on CA photocatalytic degradation for the 1.0NTO\u2013Ti catalyst are shown in Fig.\u00a06\n. It can be seen from Fig.\u00a06a that airflow rate greatly influences catalytic activity in CA photodegradation. CA conversion increases sharply as airflow increases from 0.2 to 0.3\u00a0L/h. The 120-min\u00a0removal efficiency of CA increased from 68.4 to 82.8%. However, CA conversion decreased with increasing airflow up to 0.4 and 0.5\u00a0L/h, and the value of X120 fell to 79.6 and 68.8%, respectively. Dissolved oxygen plays an essential role in supporting the generation of free radicals such as OH and HO2\n [59]. At the same time, O2 has a critical role in capturing electrons and limiting electron\u2013hole recombination [60]. Therefore, the airflow entering the reactor increases, leading to an increase in dissolved oxygen content, which improves the efficiency of CA decomposition. However, an increase in oxygen concentration also increases the level of HO2\n which reacts with OH radicals in the solution reducing the efficiency of CA treatment [59]. In addition, foaming occurs when there is a high level of aeration. This phenomenon interferes with UV light absorption by the catalyst particle and causes a substantial disturbance to the catalyst particles moving out from the reaction volume to the solution surface, resulting in a reduction in the catalyst concentration directly involved in the reaction; consequently, the efficiency of CA oxidation reduces [61]. Thus, the most suitable airflow rate is 0.3\u00a0L/h.When increasing the catalyst dosage from 0.75 to 1.5\u00a0g/L, CA treatment efficiency increases significantly. CA decomposition at 120\u00a0min\u00a0increased from 70.9% to 82.8% (Fig.\u00a06b). Catalyst concentrations continued to rise to 1.75\u00a0g/L, and CA treatment efficiency decreased slightly, being 80.6% after 120\u00a0min. The greater the concentration of catalyst, the larger the number of active sites obtained, resulting in higher conversion efficiency. However, the agglomeration of particles at high catalyst concentrations leads to a decrease in the contact surface area. In addition, collisions between particles can inactivate the catalyst, reducing the number of active sites and the reaction efficiency. On the other hand, a greater density of particles in the solution leads to an increase in turbidity and a decrease in light transmittance into the solution [62]. Therefore, the best catalyst dosage was chosen as 1.5\u00a0g/L.The pH of the solution is an essential parameter in photocatalytic reactions because it determines the surface charge characteristics of the catalyst. The highest CA conversion rate was achieved at pH 3.8, at 82.8% after 120\u00a0min (Fig.\u00a06c). The point of zero charges (pHPZC) of the 1.0NTO\u2013Ti sample is 6.3 (Fig.\u00a0S3). Thus, the 1.0NTO\u2013Ti surface is positively charged in acidic media (pH\u00a0<\u00a06.3). In addition, the pKa value of CA is 4.4 [63]. At a pH lower than its pKa value, CA is deprotonated to form the anion C6H5C2H2COO\u2212. Then, at pH\u00a0<\u00a06.3, strong CA adsorption on the 1.0NTO\u2013Ti particles is attained since electrostatic attraction results in positively charged NTO\u2013Ti with the anion C6H5C2H2COO\u2212. In addition, at a lower solution pH, more H+ ions are generated and more HO2\n free radicals are produced from the combination of H+ ions with O2\n\u2212 free radicals [64], enhancing CA decomposition performance. At a pH higher than pHPZC (>6.3) the catalyst surface is negatively charged. A base solution with Coulomb repulsion was created between the negatively charged surface of the catalyst and OH\u2212 ions that reduced OH radical formation by reaction between the ions (OH\u2212) and holes (h+), reducing CA conversion. Further, an alkaline environment significantly decelerates the transmission of ions in the reactive solution and reduces the possibility of beneficial free radicals forming [65], also resulting in a reduction in catalytic activity. Therefore, the positive charges of the surface of the 1.0NTO\u2013Ti sample at pH\u00a0<\u00a06.3, especially at pH\u00a0=\u00a03.8, were favorable in attracting anions in the reaction solution, performing the CA decomposition reaction on the surface. Furthermore, 3.8 is the inherent pH value of the CA solution, facilitating its decompose without pH adjustment.Ilmenite-structural NiTiO3 consisting of alternating layers of NiO6 and TiO6 octahedra not only strongly absorbs ultraviolet light (wavelength\u00a0<\u00a0360\u00a0nm) but also selectively absorbs visible light, mainly in a wavelength range 420\u2013540\u00a0nm and above 700\u00a0nm [66] and so could be a potential photocatalyst. There are numerous reports of investigations into the photocatalytic activity of NiTiO3, including degradation of methyl orange [31], methylene blue [66], humic acid [67], Tergitol, Safranine T, and Rhodamine B [29,30]. In some cases, NiTiO3 does not exhibit obvious photocatalytic activity in the degradation of contaminants, for example, methylene blue in water [66], due to the relatively low mobility of charges [68]. Many attempts have been made to improve the photocatalytic activity of NiTiO3. In a recent publication [66], Khan et\u00a0al. noted that a heterostructure catalyst system based on the combination of NiTiO3 nanofibers and porous gC3N4 sheets removed 97% of methylene blue molecules after 60\u00a0min exposure to visible light irradiation. Further, Pham et\u00a0al. in [69] showed that removal of methylene blue by molybdenum (Mo)-doped NiTiO3/g-C3N4 composite photocatalysts under visible light increases by a factor of 6.5 compared with that of pristine NiTiO3. Bi4NbO8Cl modified with NiTiO3 (10% in weight) showed improved activity in the photocatalytic degradation of the organic dye Rhodamine B under ultraviolet-visible irradiation. The enhanced photocatalytic performance can be ascribed to the formation of intimate interfacial contact and type-II band alignment between NiTiO3 and Bi4NbO8Cl [70]. The heterostructure catalyst NiO/NiTiO3 showed a higher degradation rate than pure NiTiO3, on which 92.9% of the Rhodamine B was degraded after 60\u00a0min illumination under a UV\u2013 full visible spectrum [36].Currently, the use of NiTiO3\u2013TiO2 heterostructure catalysts and their photocatalytic properties is rarely reported. TiO2-coupled NiTiO3 nanoparticles exhibit suitable photocatalytic activities under visible light irradiation and perform markedly better than that commercial P25, as reported by Xin Shu et\u00a0al. [35]; with TiO2-coupled NiTiO3 nanoparticles, 73% decomposition of methylene blue in aqueous solution was achieved after 6\u00a0h against around 20% by P25 under visible light irradiation. Further, Bellam et\u00a0al. [38] reported that almost equal levels of efficiency were achieved after 2.5\u00a0h. Binary phase TiO2 coupled NiTiO3 prepared by a co-precipitation method showed 75% tetracycline degradation and 35% total organic carbon (TOC) removal after 2\u00a0h [25]. Up until now, there have been no published reports that allow comparison of the activity of NTO-modified TiO2 catalysts for the treatment of the persistent organic matter, such as CA.This study provides a design strategy for incorporating small band gap nickel titanate into TiO2 semiconductors to form a highly active photocatalyst. Through combining the sol\u2013gel and the environmentally friendly hydrothermal methods the small band gap TiO2-based photocatalyst was successfully fabricated by coupling TiO2 with NiTiO3 perovskite. The addition of a small amount of NiTiO3 to TiO2 causes the reduction of crystal size and band gap energy that promotes a red shift of absorbed light to the visible region. The textural properties and band gap of NTO\u2013TiO2 heterostructure catalysts can be controlled by the amount of NiTiO3 loading. The TiO2 contained 1.0\u00a0wt.% of NiTiO3 sample was shown to be the most effective catalyst thank to its progressive textural properties and smallest band gap energy (3.02\u00a0eV), which is close to the upper threshold of visible-light catalysts. The reaction conditions have a significant influence on the performance of the photocatalyst. At the most favorable situation, the heterostructure catalyst 1.0NTO\u2013Ti exhibits superior activity compared to the parent TiO2 and NiTiO3 in photodegradation of a persistent phenolic acid.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was supported by Vietnam Academy of Science and Technology under the grant No. \u0110LTE00.09/20-21.The following is the supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.jsamd.2021.100407.", "descript": "\n                  Mesoporous TiO2 mixed with NiTiO3 at various concentrations was synthesized by combining sol\u2013gel and eco-friendly hydrothermal methods. The properties of the NiTiO3\u2013TiO2 (NTO\u2013Ti) photocatalyst were characterized using \u03a7-ray diffraction, Raman spectroscopy, scanning electron microscopy, Fourier-transform infrared spectroscopy, Brunauer\u2013Emmett\u2013Teller adsorption, energy dispersive X-ray, and UV\u2013vis absorption spectra techniques. The photocatalytic activity of NTO\u2013Ti catalysts was assessed by way of the photodegradation of cinnamic acid (CA) under UV-A irradiation. The effects of the operating parameters, including catalyst dosage, airflow, and initial solution pH on the photodecomposition efficiency of CA were also investigated. Research results confirm that NTO\u2013Ti heterostructure catalysts are synthesized in the crystalline phase with high crystallinity. Compared with pure TiO2, the NTO\u2013Ti catalysts have a smaller particle size and average crystallite size (8.6\u20139.0\u00a0nm versus 34.8\u00a0nm) and lower band gap energy (3.02\u20133.08\u00a0eV versus\u00a03.20\u00a0eV). The catalysts also enable a redshift in the absorption band from UV (\u03bb\u00a0=\u00a0385\u00a0nm) to UV-A light (\u03bb\u00a0=\u00a0404\u2013412\u00a0nm). The study showed that the physicochemical and photochemical properties and the photocatalytic performance of the NTO\u2013Ti catalysts are controlled by the NiTiO3 loading. NTO\u2013Ti with NiTiO3 1.0\u00a0wt.% was found to maximize CA photodegradation. Under the most favorable conditions, CA removal of 82.8% was obtained after 120\u00a0min, which is higher than for pure TiO2 (68.7%) and NiTiO3 (3.8%) catalysts under the same conditions.\n               "}
{"full_text": "Data will be made available on request.Over the past years, 3D printing or so-called additive manufacturing (AM) gained more and more public interest and inspired researchers to elaborate novel solutions in various application fields such as automotive/aerospace industry, bioprinting, medical/dental application or in arts. [1\u20135] This wide range of applications is based on the plethora of printable materials ranging from metals and plastics even to ceramics. [1,6] In addition, several AM techniques are conceivable for most processible materials, each of which featuring different advantages and disadvantages. [1] For example, in the field of ceramic 3D printing, the processes binder jetting, fused deposition of ceramics, stereolithography, selective laser sintering, or direct ink writing (DIW, also called robocasting) have proven to be suitable options. [7\u201311] Amongst others they have also been successfully used for ceramic 3D printing in heterogeneous catalysis, whereat direct ink writing is one of the more frequently used techniques. [12\u201316] The printed catalysts were shown to be active in different catalytic processes e.g. Tub\u00edo et al. printed monolithic Cu/Al2O3 structures for successful use in Ulman reactions [17], Stuecker et al. compared wash-coated printed alumina monoliths and directly printed material to commercial (wash-coated) monolithic structures for the combustion of methane [18], Middelkoop et al. investigated the activity of directly printed Ni/Al2O3 monolithic structures for the CO2 methanation reaction [19], and Xu et al. reported printed Al2O3 tubes with Pd immobilized in the porous inner Al2O3 ceramic to be used as continuous flow reactor for the reduction of 4-nitrophenol [20]. Compared to conventional shaping techniques such as tableting, granulation or extrusion, 3D printing offers access to tailor-made geometries as well as more complex shapes including e.g. lateral channels. The geometrical shape of a catalyst it influences important process parameters such as pressure drop as well as heat and mass transport within the catalyst bed and is therefore crucial for technical realization. [10,21\u201325]When combining additive manufacturing and heterogeneous catalysis, a distinction must be made between whether a mold is printed and later burned off, the catalytically active material or one of its precursors is printed itself, or a carrier is printed first which is afterwards loaded with the active component. [7,12,26\u201329] The latter two main approaches are comparable with the existing technical production methods for heterogeneous catalysts, where extrusion, tablet pressing, and granulation are the common shaping processes and either the active component itself or a carrier is used. [25,30,31] One example of unsupported catalysts is the mixture of iron oxide, potassium oxide and alumina for ammonia synthesis [32], whereas platinum (doped with rhenium) on alumina is an industrially applied supported catalyst for reforming. [30,32] Two main techniques are used to impregnate the carrier with the active component: Dry impregnation (incipient wetness impregnation) describes the method in which a quantity of impregnating solution corresponding to the exact pore volume of the support material is applied. However, if the volume of the impregnating solution exceeds the pore volume by a multiple, it is called wet impregnation. [32\u201335] The oxidic supports contain surface hydroxyl groups which are positively or negatively charged, depending on the pH value of the impregnation solution relative to the point of zero charge, and interact with the dissociated metal salts. [33,36] As-prepared materials are so-called egg-shell catalysts, in which the active component is in the outer layers of the shape. Competitive adsorption with other anions can be used to influence the penetration depth of the metal precursor into the catalyst leading to egg-yolk or egg-white distributions. Hereby, the respective active component is not directly on the surface but further inside the catalyst particle, which prevents catalyst poisoning. [32,33]One reaction in which platinum catalysts impregnated on alumina are used is the dehydrogenation reaction of perhydro dibenzyltoluene (18H-DBT). [37\u201339] This reaction has attracted more and more interest in the last years, as it is an integral part for the use of liquid organic hydrogen carriers (LOHC) since its first introduction by Br\u00fcckner et al. in 2014. [37,40\u201342] With the help of aromatic or heteroaromatic compounds such as methylcyclohexane, dodecahydro-N-ethyl carbazole or perhydro dibenzyltoluene, it is possible to store hydrogen chemically. [42,43] For this purpose, the corresponding materials are hydrogenated at a hydrogen surplus and dehydrogenated at hydrogen demand. For the system 18H-DBT and its corresponding dehydrogenated form 0H-DBT (dibenzyltoluene) a non-toxic compound, a hydrogen storage capacity of 6.2\u00a0wt% is reported and thus proving to be well suited. [42] In literature, perhydro dibenzyltoluene dehydrogenation reactions are carried out using e.g. Pt/AlOx catalysts such as the egg shell catalysts EleMax-102D or EleMaxD 101 from Clariant which usually are mortared or ball-milled prior to reaction. [44\u201346] However, also different metals (Pt, Pd, Ru) and carriers (SiO2, CeO2, C, TiO2) have been investigated as catalysts. [41,47,48] As this reaction is highly endothermic, [49] it is therefore, prone to heat and mass transport limitations when performed with commercial shaped catalysts. These could be overcome by using AM as a novel shaping technique that allows tailor-made catalyst shapes for improved flow behavior within the reactor. To the best of our knowledge, so far, no 3D printed catalyst has been investigated in the dehydrogenation reaction of perhydro dibenzyltoluene.The aim of this work was to investigate the influence of shapes manufactured by the 3D printing technique direct ink writing on catalyst carriers and their subsequent wet impregnation with a platinum precursor. The effect of tailored geometries and thus surface-to-volume ratios, as well as calcination temperatures and target loadings on the impregnation with platinum was examined in detail using BET analysis, light microscopy, TEM and \u03bcCT, providing useful insight into the impregnation process. Using the prepared catalysts, the catalytic activity for the dehydrogenation reaction in a semi-batch set-up was investigated.To produce a printable paste based on boehmite (AlOOH), 45.5\u00a0wt% Disperal 60 (Sasol Germany GmbH), 19.5\u00a0wt%. Pural SB (Sasol Germany GmbH) and 35\u00a0wt% acetic acid (pH\u00a0=\u00a03, Sigma Aldrich) were mixed twice by means of a SpeedMixer\u00ae (Hauschild GmbH & Co. KG) for 2\u00a0min at a maximum speed of 3500\u00a0rpm with 3\u00a0min cooling time after each mixing cycle. The printing itself was carried out with a self-constructed lab-scale DIW printer. For printing of 4\u00a0\u00d7\u00a04 mm cylinders the nozzle size was 0.41\u00a0mm, whereas for printing the monolithic structures with an overall size of 23\u00a0\u00d7\u00a04 mm and a 4\u00a0mm central hole, 0.51\u00a0mm nozzles were used. To simplify the removal of the print bed for the bigger shapes, it was covered with a thin layer of Formentrenn\u00f6l C (Clariant Produkte Deutschland GmbH). The structures were dried on the print bed for 24\u00a0h at room temperature prior to thermal post-treatment.Calcination was carried out in a WiseTherm\u00aeFHP-12 (witeg Labortechnik GmbH) muffle furnace with a heating rate of 1\u00a0K/min and an 1\u00a0h isothermal step at the target temperatures 1000\u00a0\u00b0C or 1100\u00a0\u00b0C respectively.Wet impregnation of the calcined carriers was performed by adding the desired amount of platinum sulfite acid (Clariant Produkte Deutschland GmbH) to the pre-wetted shapes in bidistilled water. The cylinders were placed directly in the water whereas the monolithic structures were immersed hanging to minimize contact areas. Impregnation was carried out for 3\u00a0h before removing excessive solution. A second calcination step was carried out in a WiseTherm\u00aeFHP-12 (witeg Labortechnik GmbH) muffle furnace by heating to 60\u00a0\u00b0C at 1\u00a0K/min for 1\u00a0h, followed by further temperature increase to 120\u00a0\u00b0C at a rate of 2\u00a0K/min for 3\u00a0h. The final temperature of 400\u00a0\u00b0C was reached with a rate of 2\u00a0K/min and held for another 3\u00a0h. Consecutive reduction was carried out at 400\u00a0\u00b0C for 3\u00a0h in a tube furnace R50/250/12 (Nabertherm GmbH) under constant flow of forming gas (H2/N2\u00a0=\u00a05 /95, Westfalen GmbH) at a heating rate of 1\u00a0K/min.Size and mechanical stability of printed and calcined shapes were measured using a MultiTest 50 (Dr. Schleuniger Pharmatron). Uniaxial compression tests vertically to the cylinder axis allowed calculation of the side crushing strength \u03c3\ncrush from the fracture load F, the cylinder diameter d and its height h by using the equation given by Timoshenko and Goodier [50]:\n\n(1)\n\n\n\u03c3\ncrush\n\n=\n\n\n2\nF\n\n\n\u03c0\ndh\n\n\n\n\n\nN2 physisorption measurements were performed on a NOVAtouch analyzer (Quantachrome Instruments) at 77\u00a0K. Prior to measurement the samples were degassed under vacuum at 120\u00a0\u00b0C for 3\u00a0h. The specific surface area SBET was calculated according to the method of Brunauer, Emmett, and Teller (BET) between p/p\n0\u00a0=\u00a00.05 and 0.3. According to the method of Barrett, Joyner, Halenda (BJH), the desorption branch of the isotherm was used to determine the pore size distribution.Powder X-Ray diffraction (XRD) measurements were performed on a PANalytical Empyrean diffractometer (Malvern) using Cu K\u03b1 radiation with a voltage of 45\u00a0kV and a monochromator. The powders were scanned in the range of 5\u00b0 to 90\u00b0 (2\u0398) with a step size of 0.0065\u00b0. Obtained data was processed using HighScore Plus.For infrared spectroscopy (IR) of adsorbed pyridine, the catalyst carrier was pelleted into a thin wafer and heated to 450\u00a0\u00b0C with a rate of 10\u00a0K/min for 1\u00a0h activation under vacuum. After cooling down to 150\u00a0\u00b0C, the apparatus was filled with pyridine until the sample was fully saturated and equilibrated for 1\u00a0h prior to another evacuation for 1\u00a0h. Scans were taken using a Nicolet 5700 FT-IR spectrometer after activation and after outgassing.Detailed images of the catalysts were taken with a MZ8 microscope (Leica) equipped with a MicroCam II (Bresser). To determine the penetration depth of platinum into the cylinders, they were embedded in epoxy resin (EpoFix and TekMek, Struers) and later polished using a Beta Grinder Polisher (Buehler).Micro-computed tomography (\u03bcCT) measurements (v|tome|x s 240, phoenix/GE) were performed to enable non-destructive analysis of the internal structure of printed cylinders as well as their impregnation behavior. The direct tube xs 240 D was operated at 70 kVp and 60\u00a0\u03bcA. To minimize beam hardening artifacts, a 0.5\u00a0mm aluminum filter was used for all measurements. The X-ray detector was a DXR-250RT with 200\u00a0\u03bcm \u00d7 200\u00a0\u03bcm pixel size on a 1000\u00a0\u00d7\u00a01000 pixel matrix of amorphous silicon directly coupled to a CsI scintillator. The distance of the X-ray focus to the detector was 812.0\u00a0mm and the focus to object distance was 18.8\u00a0mm for all measurements. This results in a magnification of 43.25 and an effective voxel size of the reconstructed volume of 4.62\u00a0\u03bcm. 1600 projections were taken for 360\u00b0 rotation of the sample. The reconstruction was done with the software xaid (MITOS, Germany). \u03bcCT images were processed using Fiji ImageJ.Even low local concentrations of platinum in a matrix of Al2O3 can very well be visualized by X-ray computed tomography because of the much stronger photo electric absorption of platinum. Fig. 5 shows XY, XZ and YZ slices through \u03bcCT scans of impregnated cylinders at different calcination temperatures and different platinum loadings. The diffusion of platinum into a homogeneous Al2O3 matrix follows Fick's second law. The diffusion constant can be determined by fitting the attenuation constant due to the position dependent platinum concentration depending to its nearest distance to the surface of the sample. Due to the many sintered particles within the Al2O3 matrix, this distance cannot be determined by the length of a straight line between the considered voxel and the surface of the sample. Instead, we used the open source software imageJ/Fiji and its macro capabilities for this task. [51] Our macro was inspired by a similar one written by O. Burri [52] and changed for our needs. The wand tracing tool of imageJ and a suitable threshold value served for tracing the sample boarder of 2D slices of the reconstructed 3D volume of each sample. With the freehand lines tool of imageJ a line was drawn between the middle of a sample and its boarder, carefully avoiding to draw through sintered particles. With the plugin \u201cExact Signed Euclidean Distance Transform (3D)\u201d the shortest distance of each pixel on the freehand line to the tracing line at the boarder was determined and saved along with the grey value of the reconstructed slice at this pixel position.These data then were fitted with Fick's second law density distribution for platinum plus a constant mean grey value for the Al2O3 matrix. Fitting was done with the module lmfit 0.9.2 in python 3.6.9. Hereby, c(x,t) is the concentration of platinum in dependence of the distance x from the outer surface and the time t. N\n0 is the number of particles in an infinitesimal small area A at x,t\u00a0=\u00a00, and D the diffusion coefficient. Eq. (2), Fick's second law for a diffusion from a boarder into a semi-infinite space, can be transformed to Eq. (3) leading to the fitting parameters a and b for the diffusion and the constant background c\n0. [53,54]\n\n(2)\n\nc\n\nx\nt\n\n=\n\n\nN\n0\n\n\nA\n\u2219\n\n\u03c0Dt\n\n\n\n\u2219\n\ne\n\n\u2212\n\n\nx\n2\n\n\n4\nDt\n\n\n\n\n+\n\nc\n0\n\n\n\n\n\n\n(3)\n\nc\n\nx\nt\n\n=\n\na\n\n\u03c0b\n\n\n\u2219\n\ne\n\n\u2212\n\n\nx\n2\n\n\n4\nb\n\n\n\n\n+\n\nc\n0\n\n\n\n\nArchimedes buoyancy method by means of a Jolly balance was used to determine bulk densities \u03c1\nbulk, apparent solid densities \u03c1\napp and porosities \u03d5 for cylinders and monoliths according to the DIN EN623\u20132 standard. [55] Three different masses of each sample were determined: the mass of the dry sample m\ndry, the mass of the completely impregnated sample m\ndamp, and the mass when suspended in water m\nsuspended. With the density of water \u03c1\nwater the following calculations were performed [55,56]:\n\n(4)\n\n\n\u03c1\nbulk\n\n=\n\n\nm\ndry\n\n\n\nm\ndamp\n\n\u2013\n\nm\nsuspended\n\n\n\n\u2219\n\n\u03c1\nwater\n\n\n\n\n\n\n(5)\n\n\n\u03c1\napp\n\n=\n\n\nm\ndry\n\n\n\nm\ndry\n\n\u2013\n\nm\nsuspended\n\n\n\n\u2219\n\n\u03c1\nwater\n\n\n\n\n\n\n(6)\n\n\u03d5\n=\n\n\n\nm\ndamp\n\n\u2013\n\nm\ndry\n\n\n\n\nm\ndamp\n\n\u2013\n\nm\nsuspended\n\n\n\n\n\n\nInductively coupled plasma optical emission spectrometry (ICP-OES) was carried out on an Aglient 700 Series ICP Optical Emission Spectrometer to determine the amount of platinum on the impregnated shapes. Therefore, the respective catalysts were grounded and dissolved in aqua regia (hydrochloric acid: nitric acid (both Sigma Aldrich)\u00a0=\u00a03:1\u00a0vol%/vol%). The samples were diluted with bidistilled water and filtered via 0.45\u00a0\u03bcm PTFE syringe filters (VWR). For preparation of the metal standards, several concentrations ranging from 1\u00a0ppm to 50\u00a0ppm were prepared using a platinum AAS standard (Sigma Aldrich). For concentration determinations the wavelength 214.42\u00a0nm was used.Metal particle size as well as dispersion were examined by means of transmission electron microscopy (TEM). Grounded samples were suspended in absolute ethanol (Sigma Aldrich), dropped on Holey Multi A grids (Quantifoil Micro Tools GmbH) and dried. The measurements were performed on a JEOL JEM 1400 plus instrument at an acceleration voltage of 120\u00a0kV. Using the software Fiji ImageJ, the metal particle diameter d\nPt was determined manually by measuring at least 300 particles each. Based on d\nPt the metal dispersion D\nPt can be calculated as following: [57].\n\n(7)\n\n\nD\nPt\n\n=\nK\n\u2219\n\n\nV\nPt\n\n\n\nS\nPt\n\n\u2219\n\nd\nPt\n\n\n\n\n\nwith the constant K reflecting the particle shape (K\u00a0=\u00a06 for spherical particles), V\nPt the volume per metal atom and S\nPt the average surface area of metal particles per metal atom.Dehydrogenation test reactions were carried out semi-batch-wise in a 100\u00a0mL flask equipped with two flow breakers and a thermocouple. The required amount of the reactant perhydro dibenzyltoluene (18H-DBT; Hydrogenious LOHC Technologies GmbH) was stirred and heated up to 325\u00a0\u00b0C under argon atmosphere using a heating mantle (Winkler AG). After reaching the set temperature the catalyst (nPt/n18H-DBT\u00a0=\u00a00.175 mmolPt/mol18H-DBT) was added to the reactant and the temperature was reduced to the reaction temperature of 310\u00a0\u00b0C. The higher starting temperature was chosen to counterbalance the strong temperature drop at the start of the reaction caused by the strong endothermicity of the dehydrogenation. The mass of the catalyst with 0.9\u00a0wt% loading was only 1/3 of the catalyst's mass with a loading of 0.3\u00a0wt%, while the fine tuning of the Pt to reactant ratio was carried out adjusting the reactant amount. For full particle test reactions, the cylinders and monoliths were placed in a wire cage or on a wire (V4A stainless steel) as catalyst holder, respectively. Intrinsic test reactions with catalyst powder were carried out by adding the ground catalyst powder to the reactant. Sieving of the samples was not performed to ensure that the entire sample was used and to avoid sieving out alumina or Pt, which would then affect the concentration of active species in the reaction. The reaction was monitored for 6\u00a0h using proton nuclear magnetic resonance spectroscopy (1H NMR) in acetone\u2011d\n6 carried out using a Bruker Ascend spectrometer at 400\u00a0MHz (300K). All spectra are referred to the solvent residual signal and chemical shifts are given in \u03b4-values (ppm). Based on the 1H NMR data, the degree of hydrogenation (DH) was calculated according to Do et al. [35] and Preuster [58] with the ratio x of the integral of the aromatic protons (7.5\u20136.6\u00a0ppm) to the integral of all protons (7.5\u20136.6\u00a0ppm, 4.8\u20133.6\u00a0ppm, 2.6\u20132.1\u00a0ppm, and 2.0\u20130.4\u00a0ppm):\n\n(8)\n\nDH\n=\n1.3945\n\nx\n6\n\n\u2212\n4.9037\n\nx\n5\n\n+\n5.6287\n\nx\n4\n\n\u2212\n5.207\n\nx\n3\n\n+\n4.0098\n\nx\n2\n\n\u2212\n2.9217\nx\n+\n1\n\n\n\nFor better comparison of the dehydrogenation activity, the productivity P is defined as the ratio of the mass of hydrogen m\nH2 evolved per mass Pt m\nPt and time [59,60]:\n\n(9)\n\nP\n=\n\n\nm\n\nH\n2\n\n\n\n\nm\nPt\n\n\u2219\nt\n\n\n=\n\n\n\u2206\nDH\n\u2219\n\nm\n\n18\nH\n\u2212\nDBT\n\n\n\u2219\n\n\nM\n\nH\n2\n\n\n\nM\n\n18\nH\n\u2212\nDBT\n\n\n\n\u2219\n\n\n\u03bd\n\nH\n2\n\n\n\n\u03bd\n\n18\nH\n\u2212\nDBT\n\n\n\n\n\n\u2206\nt\n\u2219\n\nm\ncat\n\n\u2219\n\n\u03c9\nPt\n\n\n\n\n\n\nHerein, \u0394 DH is the difference in degree of hydrogenation during the time \u0394 t. m\ncat is the catalyst mass, M\ni and \u03bd\ni are the molar mass and the stoichiometric coefficient of the respective component i and \u03c9\nPt is the platinum loading as determined via ICP-OES. For better comparison, P should be compared at the same \u0394 DH, therefore it was always examined between a degree of hydrogenation of 90% and 40%. Moreover, the productivity per bulk volume V\nbulk of the catalyst is defined as P\n\nVbulk:\n\n(10)\n\n\nP\n\nV\nbulk\n\n\n=\n\n\nm\n\nH\n2\n\n\n\n\nV\nbulk\n\n\u2219\nt\n\n\n=\n\n\n\u2206\nDH\n\u2219\n\nm\n\n18\nH\n\u2212\nDBT\n\n\n\u2219\n\n\nM\n\nH\n2\n\n\n\nM\n\n18\nH\n\u2212\nDBT\n\n\n\n\u2219\n\n\n\u03bd\n\nH\n2\n\n\n\n\u03bd\n\n18\nH\n\u2212\nDBT\n\n\n\n\n\n\u2206\nt\n\u2219\n\nV\nbulk\n\n\n\n\n\n\nThe bulk volume V\nbulk of the monoliths is calculated as the volume of a cylinder with the respective diameter and height of the monoliths used. V\nbulk of the cylinders is determined by multiplying the wire cage base area with the measured filling height of the cylinders.In a first step, two different catalyst shapes were printed using direct ink writing of a paste consisting of boehmites and acetic acid as cheap organic binder. [61] The first ones were cylinders with a height and diameter of 4\u00a0mm whereas the second shape was a monolithic shape sized 23\u00a0\u00d7\u00a04 mm including a 4\u00a0mm hole in the center (Fig. 2), herein referred to as monolith. While cylinders can be fabricated by extrusion or tableting, such monoliths including lateral holes are not accessible by commercial shaping techniques. Addition of these lateral holes compared to extrusion-based monoliths improves the flow tortuosity and thus catalytic activity by improving mass and heat flow. [18] Afterwards, thermal post treatment was performed to transform the printed aluminum oxide hydroxide to alumina, which acted as final carrier material. 1000\u00a0\u00b0C and 1100\u00a0\u00b0C were chosen as calcination temperatures (T\ncalc), as within this temperature range the phase transition from \u03b3- to \u03b1- via \u03b8-alumina takes place [25,30,32] causing strong changes in surface area and stability [9]. While for other test reactions like the oxidation of ethanol, formation of the more stable \u03b1-Al2O3 phases is suited, [62] pure \u03b1-Al2O3 is not considered reasonable herein as it has an extremely low surface area. It emerges around 1100\u00a0\u00b0C, thus higher calcination temperatures are not used. [25,30,32,63] The formation of the desired phases has been confirmed via powder XRD measurements of the printed shapes before and after calcination (Fig. S1).Mostly due to the drying on the print bed but also to some extent caused by calcination, the final shapes turned out slightly smaller than initially aimed (Table S1). For the cylinders using common formulas, the surface-to-volume ratio (S/V) was calculated to 1.64\u00a0mm\u22121 and 1.72\u00a0mm\u22121 for T\ncalc\u00a0=\u00a01000\u00a0\u00b0C and T\ncalc\u00a0=\u00a01100\u00a0\u00b0C, respectively. When calculating the surface-to-volume ratio of the monoliths, the strand diameter is the most important characteristic and leads to S/V ratios of 9.70\u00a0mm\u22121 for T\ncalc\u00a0=\u00a01000\u00a0\u00b0C or 10.11\u00a0mm\u22121 for T\ncalc\u00a0=\u00a01100\u00a0\u00b0C (Table 1\n). For the mathematical calculations, perfect geometrical shapes were assumed leading to small inaccuracies of the obtained values. However, these inaccuracies are considered negligible small compared to the differences between cylindrical and monolithical shapes. The distance between two strands in the monoliths has been calculated to be 0.33\u00a0mm for the monoliths calcined at 1000\u00a0\u00b0C and 0.29\u00a0mm for the monoliths calcined at 1100\u00a0\u00b0C. As the overall shrinkage was increased at higher temperatures for both, monoliths and cylinders, the shapes at lower calcination temperatures had higher surface-to-volume ratios. Additionally, the surface-to-volume ratio for monoliths was about 5.9 times higher compared to the cylinders independent of the corresponding calcination temperature.Using Archimedes buoyancy method by means of a Jolly balance the overall porosity \u03d5 of the shapes was examined revealing values ranging from 58.4% (cylinders, 1100\u00a0\u00b0C) to 77.9% (monoliths, 1000\u00a0\u00b0C). The higher porosity at lower calcination temperature could be explained by temperature-dependent shrinkage behavior as well as the decrease of small pores. As the weight during phase transition remained constant, higher shrinkage led to a smaller porosity. Differences between cylinders and monoliths could be explained as the porosity \u03d5 determined via Jolly balance is defined as the fracture of the volume of open pores to the sum of the volume of open and closed pores as well as the solid density itself. [56] The higher surface-to-volume ratio of monolithic structures increased the amount of open pores and subsequently increases the porosity.Regarding the mechanical stability, the crushing strength increased from 0.7\u00a0MPa by a factor of six when changing from 1000\u00a0\u00b0C to 1100\u00a0\u00b0C calcination temperature (Table 2\n). This is in accordance with the results from Ludwig et al. [9] Generally, at both calcination temperatures shapes with sufficient crushing strength for catalytic applications can be obtained. The crushing strength can be calculated from the pressure required for compression using literature formulas based on the geometrical shape. [50] As no such formulas are present for the respective monolithic shapes, no values for them could be determined. By means of N2 physisorption, the specific surface area S\nBET was determined quantifying a decrease from 55 m2g\u22121 to 22 m2g\u22121 and hereby showing a similar trend as the total pore volume with increasing the calcination temperature. Surface area and pore volume are herein considered as material characteristics and thus primarily based on the calcination temperature rather than the shape printed. Overall pore size distributions determined via BJH method (Fig. 3) showed only mesopores and a bimodal curve for both calcination temperatures with pore radii of approximately 5\u00a0nm and 20\u00a0nm. However, sintering at higher temperatures led to a decreased amount of both, smaller and bigger pores with a much more prominent decline of the smaller pores. This finding is in accordance with common literature, showing sintering of smaller pores first. [31,64,65]\u03bcCT measurements provided information about the inside of the printed and calcined cylindrical carriers (Fig. S2). XZ and the YZ slices of the shapes (as depicted in Fig. 1\n) showed a rough outer surface area caused by the layer-wise manufacturing technique independently if the uncalcined so-called green part or the calcined material were scanned. Further, all samples exhibited small cracks on the flat bottom side of the cylinder that are most likely caused by anisotropic shrinkage due to drying on the print bed at room temperature directly after printing. Some small dark spots from entrapped air could be caused either by printing inaccuracies, especially at the outer layer, or form during drying and sintering. As the white spots, that could be seen for both calcined samples, are not present in the green body, these are caused by the thermal treatment and therefore sintering of the alumina particles. In general, the amount of white sintered particles is slightly higher for the samples calcined at 1100\u00a0\u00b0C.\n\n\n\n\n\n\nIn a next step, wet impregnation with platinum sulfite acid solution of cylinders and monoliths, both calcined 1000\u00a0\u00b0C as well as at 1100\u00a0\u00b0C, was carried out. With 0.3\u00a0wt% and 0.9\u00a0wt% there are two target platinum loadings for each of the four different carriers resulting in a set of eight different impregnated catalysts (Table 3\n).Again, surface area measurements were performed of the impregnated samples. For cylinders calcined at 1000\u00a0\u00b0C as well as 1100\u00a0\u00b0C and both target Pt loadings (0.3\u00a0wt% and 0.9\u00a0wt%), physisorption measurements were carried out after impregnation and calcination of the platinum sulfites to platinum oxides as well as after the subsequent reduction. Comparison of the specific surface area S\nBET as well as the pore volume V\np showed that those characteristics are not changed by impregnating and reduction of the carrier (Fig. S3). Only the total pore volume decreased slightly most likely caused by the fact that platinum clusters are now partly filling the pores. When comparing the pore size distributions for carriers calcined at 1000\u00a0\u00b0C, one can see the bimodality was maintained as pores with a radius of approximately 5\u00a0nm as well as the ones with 20\u00a0nm radius decrease both, which indicates that the platinum was deposited in all pores.To determine the exact amount of platinum deposited on the catalyst carrier, ICP-OES measurements were performed. The results showed that the target loading was not achieved for any of the samples as the actual loading is lower (Table 3). The achieved loading for 0.9\u00a0wt% target loadings is lower relative to the 0.3\u00a0wt% target loadings, as a higher amount of platinum was supposed to impregnate on the same overall surface. Further, the loading for monoliths was decisively higher as about 78% to 97% of the targeted loading could be obtained whereas the loading of the cylinders ranged between 37% (1100\u00a0\u00b0C, 0.9\u00a0wt%) and 82% (1000\u00a0\u00b0C, 0.3\u00a0wt%). This can be explained by the fact, that the surface-to-volume ratio of monolithic shapes is about six times greater than for cylinders. When comparing the loading of the shapes at the calcination temperatures, it is remarkably that the relative loading was higher when calcining at 1000\u00a0\u00b0C, even though the S/V ratio was higher for shapes calcined at 1100\u00a0\u00b0C. However, not only the external surface-to-volume ratio must be taken into account but also the specific surface area S\nBET of the carrier itself as determined via N2 physisorption. Here, the specific surface area of shapes calcined at higher temperatures was lower and therefore explaining why only lower loadings could be achieved. These findings in general correlate with literature. [66]Transmission electron microscopy (TEM) measurements were carried out to determine the diameter d\nPt of the platinum clusters on the catalyst (Fig. S4, Fig. S5). In accordance with the previous results from ICP-OES measurements showing that a higher calcination temperature and therefore a lower specific surface area resulted in lower loadings, TEM measurements revealed that the platinum particle diameter was generally higher for higher calcination temperatures. The lower specific surface area as well as a lower amount of surface hydroxyl groups as derived from IR spectra of adsorbed pyridine (Fig. S6) and in accordance to literature [64,67\u201370] allowed only a limited number of particles to form which consequently get bigger. Despite the fact of the specific surface area, also the S/V ratio is important for the impregnation as the monoliths show smaller mean values of the platinum cluster size than the cylinders. Interestingly, the platinum cluster diameter decreased with increased loading. One reason for that might be, that if impregnated with higher amounts of platinum, the probability of ion exchange during impregnation is higher leading to more nucleation and thus overall, slightly smaller particles. Longer impregnation times might lead to loadings close to the target loading and similarly to bigger metal particles. This effect was more prominent for the cylinder samples; however, the achieved loading for these is maximum doubled with three-fold targeted loading. However, it has to be noted that the standard deviation is relatively high, so the values have to be treated with caution. As the dispersion D\nPt is inversely proportional to the metal particle diameter d\nPt, it showed opposing trends ranging from 74% (cyl. 1100\u00a0\u00b0C, 0.3\u00a0wt%) to 94% (monol. 1000\u00a0\u00b0C, 0.9\u00a0wt%).When examining enlarged images of the monolithic cross section derived from light microscopy (Fig. 4), it is notable that at a calcination temperature of 1000\u00a0\u00b0C the overall surface did not seem smooth but shows dark spots that were more prominent for 0.9\u00a0wt% but could also be observed at 0.3\u00a0wt%. As the metal cluster size determined via TEM was in the same range for all the samples, it can be assumed that these dark spots were caused by the higher porosity of samples calcined at 1000\u00a0\u00b0C. One possible explanation is the higher number of small pores that were observed which might cause an optical illusion and hereby just appear to be darker. However, the presence of larger Pt particles might also be a explanation thereof. Microscopic analysis of the cylinders (Fig. S7) revealed a similar coloring like the monolithic structures.Light microscopy was used additionally to examine the impregnation and its depth into the cylinders and monolithic structures. In general, a higher target loading resulted in a deeper platinum penetration for all carriers. Due to the lower specific surface area, platinum penetrated deeper into the shape at calcination temperatures of 1100\u00a0\u00b0C. Further, the increased surface-to-volume ratio of the monolithic structures led to a decreased penetration depth of the platinum. When comparing cylinder shapes with each other, the penetration depth of cylinders calcined at 1000\u00a0\u00b0C and a Pt loading of 0.9\u00a0wt% was lower than that of those calcined at 1100\u00a0\u00b0C with 0.3\u00a0wt% loading. For monoliths however, shapes calcined at 1100\u00a0\u00b0C with 0.3\u00a0wt% loading showed a smaller penetration depth compared to those calcined at 1000\u00a0\u00b0C with 0.9\u00a0wt%. As previously discussed, two characteristics, namely the surface-to-volume ratio and the specific surface area influence the impregnation with platinum and lead to opposing trends for these sets. Apparently, the high surface-to-volume ratio (compared to the platinum loading) of the monolithic structures seems to dominate penetration over the specific surface area. For the cylinder on the other hand, the surface-to-volume ratio for both calcination temperatures are so small that the decreased specific surface seems to be predominant at elevated calcination temperatures. According to literature, the layer thickness of the active material in egg-shell catalysts for the dehydrogenation reaction of perhydro dibenzyltoluene should not exceed 90\u00a0\u03bcm in order to prevent mass transport limitations. [59] This requirement was only fulfilled for monolithic structures calcined at 1000\u00a0\u00b0C with a target loading of 0.3\u00a0wt% (50\u00a0\u03bcm penetration depth) and for monolithic structures calcined at 1100\u00a0\u00b0C with a target loading of 0.3\u00a0wt% (90\u00a0\u03bcm penetration depth) (Table 3, Fig. 4).To gain a deeper understanding of the platinum penetration depth into the catalyst carrier on a macroscopical scale, \u03bcCT imaging analysis and fitting of the platinum solution diffusion and the concentration decrease into the carrier has been performed (Fig. 5). Due to the considerably higher atomic number of platinum compared to aluminum, the X-ray absorption of Pt is much higher and thus impregnated areas are depicted brighter than unimpregnated centers of the cylinders. \u03bcCT images showed that a higher calcination temperature and a higher loading increased the penetration depth, which is in accordance with the overall trends that were observed via light microscopy. Proving a homogeneous impregnation, one can see that the impregnation occurred on all outer surface areas of the cylinders. Still, the penetration did not only occur lateral but also diagonal, explaining why the rough outer surface area could not be seen as of the boundary layer within the cylinder caused by the platinum impregnation. However, the optical determination of the impregnation depth was in general more difficult for carriers calcined at 1100\u00a0\u00b0C, due to a reduced overall contrast as the concentration of platinum is more widely spread. Further, no platinum could be observed along the cracks at the bottom side of each cylinder. One explanation for this is that as the catalyst was pre-wetted prior to impregnation, capillary forces do not play an important role but only diffusion processes influence the impregnation. [33] However, it is also possible, that the capillary forces in general were not strong enough to completely fill the cracks with impregnation solution at all. [33]The penetration behavior itself was further investigated by means of a grey scale analysis of the \u03bcCT scans. Carefully avoiding the white sintered particles, a line was drawn from the center to the edge of the cylinder. The resulting grey value of each pixel as well as its shortest distance to the edge of the cylinder was recorded. These values were then fitted according to the Fick's second law for a semi-infinite cylinder (2) (Fig. S8) and for better comparison only the stacked and fitted grey values are visualized (Fig. 6). Regarding the courses of the grey value, it is obvious that cylinders with equal calcination temperatures show similar trends to each other. In accordance with the microscopic images, one can clearly see that the amount of platinum at the outer surface of the cylinders calcined at 1000\u00a0\u00b0C is higher but decreases rapidly. On the other hand, for shapes calcined at 1100\u00a0\u00b0C the overall course is flatter but the penetration deeper. This higher course at lower calcination temperatures can be explained as similar or even higher amounts of platinum are impregnated on a smaller area resulting in higher loadings. As described previously, these differences in penetration depth and loading are caused by the varying amount of surface hydroxyl groups allowing a higher platinum loading. Additionally, \u03bcCT scans show that the penetration depth increases with higher target loadings. In general, the penetration depth obtained via light microscopy approximately corresponds to a platinum density of one tenth of the value at the surface of the cylinder as measured via \u03bcCT (Fig. S9). This shows that \u03bcCT analysis as non-destructive technique is a useful tool to assess impregnation behavior and penetration depth of the catalytically active species, allowing detailed understanding of the impregnation for the preparation of highly active catalysts.Catalytic test reactions completed the evaluation of the influence of catalyst shape, calcination temperature and Pt impregnation. Therefore, the catalysts were tested for the dehydrogenation of 18H-DBT (Fig. 7). As platinum is the active component, the mass of catalyst was adjusted to the reactant mass, so that a constant Pt to 18H-DBT ratio was maintained throughout all reactions.To exclude diffusion limitations, intrinsic test reactions were executed with powdered catalysts first (Fig. 8a). These revealed that almost all catalyst samples showed the same dehydrogenation activity independently on catalyst shape, calcination temperature and Pt loading. The productivities of all catalysts are very similar and range from 6.6 gH2\u00b7gPt\n\u22121\u00b7min\u22121 to 8.9 gH2\u00b7gPt\n\u22121\u00b7min\u22121 (Table 4\n). Small differences were most likely caused by inaccuracies of the 1H NMR examination. After a reaction time of 1\u00a0h the degree of hydrogenation had already decreased from 98% to approximately 25%. A minimum degree of hydrogenation at around 7% seemed to be reached with all samples after 3.5\u00a0h. No further decline in the degree of hydrogenation was observed when extending the reaction time. There were only two curves that are not perfectly in line to other reactions, namely the monoliths calcined at 1000\u00a0\u00b0C and 1100\u00a0\u00b0C with a target loading of 0.3\u00a0wt%. Especially between 0.5\u00a0h and 3\u00a0h they showed a slightly better dehydrogenation activity. This results in higher productivities with values of 8.9 gH2\u00b7gPt\n\u22121\u00b7min\u22121 and 8.3 gH2\u00b7gPt\n\u22121\u00b7min\u22121 for 1000\u00a0\u00b0C and 1100\u00a0\u00b0C, respectively. One possible explanation is that the overall distribution of platinum on those shapes is the best, as the monolithic structures have the highest S/V ratios and a target loading of 0.3\u00a0wt% is rather low.When executing full particle test reactions, the cylinder catalysts were put in a stainless-steel wire basket whereas the monoliths were placed on a stainless-steel wire. After reaching the set temperature, the catalyst was lowered into the reaction solution and fixated, starting the reaction. The overall courses in Fig. 8 clearly show that the dehydrogenation activity of the powdered catalyst is higher than the respective activity in the full particle reactions. This can be confirmed as the productivities for the full particle test range from 1.3 gH2\u00b7gPt\n\u22121\u00b7min\u22121 to 4.0 gH2\u00b7gPt\n\u22121\u00b7min\u22121. The reactions with full particles consequently exhibit only between 16% (cyl. 1100\u00a0\u00b0C, 0.9\u00a0wt%) to 48% (monol. 1100\u00a0\u00b0C, 0.3\u00a0wt%) of the respective powdery productivities (Table 4). A degree of dehydrogenation of 25% required reaction times of approximately 3 to 4\u00a0h or even up to 6\u00a0h (cyl. 1000\u00a0\u00b0C and 1100\u00a0\u00b0C, 0.9\u00a0wt%). The enhanced performance of catalyst powder can be attributed to diffusion limitations of full particle catalysts. Throughout the whole reaction all printed shapes remain intact due to their sufficiently high crushing strength enabling easy separation of the catalyst and the reaction mixture afterwards.Comparing the activities of either the cylinders or the monolithic structures in full particle reactions to each other (Fig. 8c and Fig. 8d) it is remarkable that the performance and productivity of catalysts calcined at 1100\u00a0\u00b0C is higher than of those calcined at 1000\u00a0\u00b0C (Table 4). This effect could be observed even though the penetration depth of the platinum is higher at 1100\u00a0\u00b0C and one would expect that this hinders the diffusion leading to slower reaction speeds. However, another aspect regarding the activity is coming into play here as 18H-DBT is a relatively large molecule and hence a minimum pore diameter of around 26\u00a0nm is beneficial for the reaction, making the bigger pores more important. [59,60] As discussed previously, the pore size distributions of the two carriers (Fig. 3) differed as the number of pores with radii of 5\u00a0nm was significantly higher when calcining at 1000\u00a0\u00b0C. As a homogeneous distribution of platinum over all pores is expected, it is likely that there was some platinum present in the smaller pores and thus inaccessible for the reactant. This led to more inaccessible platinum which lowers the overall activity. Yet, when grounding the catalysts for intrinsic reactions this diffusion limitation to smaller pores seemed to be no longer prominent. Nevertheless, the trend that the calcination temperature of 1100\u00a0\u00b0C with otherwise identical parameters led to a higher activity is not applicable for the cylinders with a loading of 0.9\u00a0wt%. These showed a similar but also the lowest activity of all samples examined as well as a productivity of only 1.3 gH2\u00b7gPt\n\u22121\u00b7min\u22121 regardless of the calcination temperature. Even after a reaction time of 6\u00a0h the dehydrogenation was lower than for all other systems and the degree of hydrogenation achieved was at around 35%. Compared to the other catalysts, the general uptake of platinum during impregnation was the lowest with values of 56\u00a0wt% (cyl. 1000\u00a0\u00b0C, 0.9\u00a0wt%) and even lower 37\u00a0wt% (cyl. 1100\u00a0\u00b0C, 0.9\u00a0wt%). This leads to the assumption that even though the platinum particle size is comparable to the other samples, the general accessibility of platinum is hindered due to lower S/V ratios and high target loadings.For all further examinations, it is important to keep in mind that the bulk volume during the reaction was as identical as possible for cylinders and monoliths when comparing same target loadings. For 0.3\u00a0wt% V\nbulk was in average 3.3\u00a0\u00b1\u00a00.2\u00a0cm3 whereas for 0.9\u00a0wt% it was 1.4\u00a0\u00b1\u00a00.4\u00a0cm3. Due to higher measurement inaccuracies for cylinders with a 0.9\u00a0wt% target Pt loading as well as the relatively lower loadings compared to the target loadings, the standard deviation for the latter one is high, and values should be regarded with caution. Soley for monoliths with a target loading of 0.9\u00a0wt%, V\nbulk can be calculated to 1.0\u00a0\u00b1\u00a00.1\u00a0cm3. On average, the bulk volume for the monoliths with a loading of 0.9\u00a0wt% was only 32% compared to the respective lower loaded cylinders. In general, at the same calcination temperature and target loadings the monolithic structures showed higher activities and productivities than the cylinders. For this trend several factors came into play. As the S/V ratio of monoliths was about six times as high as for cylinders, the same amount of platinum was spread over a larger surface. In consequence, the overall penetration depth was lower and platinum particle diameters were smaller. These factors are important for the reaction, as they influence the accessibility of the platinum for the reactant molecules. Still, it cannot be distinguished how important which of the effects is because of the overlap of these effects. Beside the differences in platinum distribution, it is also likely that the overall fluid dynamics within the semi-batch reactor must be considered. The monoliths with several small channels most likely enabled better mass and heat flow through the catalyst bed than the cylinder bed and hereby also influenced the catalytic reaction beneficially.Lastly, the influence of the different loadings at similar shapes and calcination temperatures was examined. For all tested full particle reactions, the catalysts with a target loading of 0.9\u00a0wt% showed a reduced activity and lower productivities compared to those with 0.3\u00a0wt%. On first sight, this seems counterintuitive, but all reactions were carried out with the same platinum to reactant ratio. Therefore, an increased loading resulted in a decreased mass of catalyst in the reaction and therefore a reduced bulk volume. Another factor to be considered is that higher loadings also resulted in deeper penetration of the platinum and therefore were likely to increase reactant diffusion limitation to the active platinum center leading to lower activities. This effect could not be balanced by the fact that the platinum particle sizes are smaller for higher loadings as discussed previously. In general, the best activity and productivity within the full particle tests were observed for the monolith 1100\u00a0\u00b0C, 0.3\u00a0wt% with a value of 4.0 gH2\u00b7gPt\n\u22121\u00b7min\u22121 followed by the monolith with the same loading at the lower calcination temperature (3.5 gH2\u00b7gPt\n\u22121\u00b7min\u22121).Interestingly, the comparison of monolithic structures with a loading of 0.9\u00a0wt% to cylinders with a loading of 0.3\u00a0wt% (Fig. 8b) at same calcination temperatures revealed similar Pt-based productivities. However, a significant difference within this set of catalysts was the bulk volume since for similar platinum to reactant ratios the catalyst mass was reduced for the approximately three-fold increased loading. Hereby, the monoliths had only about 32% of the bulk volume compared to the respective cylinders. Only slightly better Pt-based productivities of cylinders than the respective monolithic structures were achieved (Table 4). The differences in bulk volume changed heat and mass transport within the semi-batch reactor and led to the minimally decreased activities compared to the respective cylinders, even though the platinum particle size or the penetration depth present at the monolithic samples would lead to the assumption of higher activities.The productivity in respect to the bulk volume P\n\nVbulk it is in general higher for monoliths than for cylinders. The values for the cylinders range from 0.0029 gH2\u00b7gPt\n\u22121\u00b7min\u22121\u00b7cm\u22123 (cyl. 1100\u00a0\u00b0C, 0.9\u00a0wt%) to 0.0050 gH2\u00b7gPt\n\u22121\u00b7min\u22121\u00b7cm\u22123 (cyl. 1000\u00a0\u00b0C, 0.3\u00a0wt%). Due to the overall low activity of the cylinders with higher loading, P\n\nVbulk for them is slightly smaller than for the ones with 0.3\u00a0wt% loading. However, when comparing the monolithic samples regarding their P\n\nVbulk, a different trend can be observed. The higher loaded monoliths exhibit an approximately two-fold higher volume-based productivity. These contrary trends can be attributed to the differences in surface-to-volume ratio of the monoliths compared to the cylinders and the resulting differences in wet impregnation. When comparing P\n\nVbulk for the cylinders 1000\u00a0\u00b0C and 1100\u00a0\u00b0C at 0.3\u00a0wt% (0.0050 gH2\u00b7gPt\n\u22121\u00b7min\u22121\u00b7cm\u22123 and 0.0044 gH2\u00b7gPt\n\u22121\u00b7min\u22121\u00b7cm\u22123) and the respective monolith values of 0.0139 gH2\u00b7gPt\n\u22121\u00b7min\u22121\u00b7cm\u22123 and 0.0147 gH2\u00b7gPt\n\u22121\u00b7min\u22121\u00b7cm\u22123 which exhibit similar Pt based productivities, the three fold average higher P\n\nVbulk for the monoliths becomes apparent. This shows that by means of increased surface-to-volume ratios the activity and productivities P and P\n\nVbulk of a catalyst can be easily increased just by improving the impregnation behavior. Hereby, AM and especially DIW are suited techniques to obtain these varied and more complex shapes with increased surface-to-volume ratios. Furthermore, advanced shape optimization is likely to increase the flow behavior, leading to even improved catalytic activities.Herein, alumina catalyst carriers were fabricated by direct ink writing, namely cylinders and monolithic structures. These shape variations and respective changes in external surface area as well as two calcination temperatures (1000\u00a0\u00b0C and 1100\u00a0\u00b0C) resulting in different specific surface areas influenced the wet impregnation behavior and consequently the activity of full particle catalysts in the dehydrogenation of 18H-DBT. The prepared shapes have in been analyzed in depth using a combination of various techniques, including BET, light microscopy, TEM and \u03bcCT. By evaluating the catalysts, the use of \u03bcCT as an advanced analysis technique offers unique advantages for the preparation of 3D printed, heterogeneous catalysts. Using these techniques, some conclusions can be drawn, which provide useful guidance for the impregnation of catalyst carriers. In general, a higher surface-to-volume ratio of the carrier resulted in a higher loading relative to the target loading as well as in smaller platinum particles and lower penetration depth. Similar trends could be observed at lower calcination temperatures and therefore higher specific surface areas as well. Higher targeted loadings on the other side cause a decrease of the platinum loading as well as the average particle size but increase the penetration depth. These deductions can be used for the preparation of well-defined catalysts by means of impregnation with an active species. Especially when focusing on very complex structures accessible via 3D printing, such findings help understanding the preparation processes, enabling tailor-made impregnation of such advanced structures.Resulting from the intrinsic and full particle catalytic dehydrogenation test reactions one can conclude that fabrication of catalysts by direct ink writing is beneficial for the catalytic reaction in general. As a higher exposed surface of the catalyst is beneficial for the catalytic performance, variation of the geometries by DIW or even printing e.g. a continuous flow reactor itself might influence our catalytic activity even further. [20,62] Overall, higher calcination temperatures were beneficial for the reaction as the pore size distribution of the carrier is enhanced hereby. This seemed to be more important than a higher penetration depth or larger platinum particle sizes. When impregnating monoliths and cylinders with the same target loading and working with similar reactor volumes, the monolithic structures showed significantly higher activities for dehydrogenation of perhydro dibenzyltoluene. Further, when aiming for similar activities with monolithic structures one can either reduce the platinum loading keeping the reactor volume constant or keep the same Pt amount but reduce the reactor volume. Both is beneficial, as it either requires a reduced amount of the very expensive noble metal platinum or decreased reactor sizes and therefore decreased operating costs. This reveals the potential that additive manufacturing and especially direct ink writing show when fabricating catalyst carriers.Based on the requirements of your journal, we want to hereby list a detailed breakdown of all authors contributions:Paula F. Gro\u00dfmann: conceptualization, writing - original draft, data curation, formal analysis, investigation, visualization, writing - review & editing.Markus Tonigold: conceptualization, formal analysis, writing - review & editing, resources.Norman Szesni: formal analysis, visualization, writing - review & editing.Richard W. Fischer: conceptualization, formal analysis, writing - review & editing, project administration.Alexander Seidel: formal analysis, visualization, writing - review & editing, resources.Klaus Achterhold: formal analysis, visualization, data curation, writing - review & editing.Franz Pfeiffer: supervision, writing - review & editing, project administration.Bernhard Rieger: supervision, project administration, funding acquisition, resourcing, writing - review & editing.The financial support of the Bayerische Forschungsstiftung (BFS) is gratefully acknowledged. K. Achterhold and F. Pfeiffer acknowledge financial support through the DFG Cluster of Excellence Munich-Centre for Advanced Photonics (MAP, DFG EXC-158), the DFG Gottfried Wilhelm Leibniz Program and the Center for Advanced Laser Applications (CALA).The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.P. F. Gro\u00dfmann would like to thank Max Koch for carrying out the ICP-OES measurements and Dr. Carsten Peters and Roland Weindl for their help when carrying out TEM measurements. Furthermore, P. F. Gro\u00dfmann would like to thank Larissa Sommer, Marlene Viertler, Jan Meyer, Stefanie Pongratz and Mira Eggl for their help during printing and in carrying out various measurements and Moritz Kr\u00e4nzlein for his scientific input. Special thanks to the MuniCat team and especially Hanh My Bui for the fruitful discussions.\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nDetailed dimensions and densities for printed shapes, physisorption results of the impregnated cylinders, metal particle distributions as examined via TEM analysis, microscopic images of impregnated cylinders. Supplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2023.106610.", "descript": "\n                  Direct ink writing as additive manufacturing technique was used to print two different boehmite based shapes, cylinders and monoliths, serving as catalyst carriers. These were wet impregnated targeting 0.3\u20130.9\u00a0wt% platinum loadings. ICP-OES, \u03bcCT and microscopy revealed dependencies from calcination temperature, geometry and platinum loading. Dehydrogenation reactions of perhydro dibenzyltoluene as liquid organic hydrogen carrier were performed examining the catalytic performance. Differences when executing full particle measurements led to the conclusion that direct ink writing as shaping technique for catalyst carriers and the respective impregnation is highly beneficial as more complex shapes can be obtained, resulting in higher activities.\n               "}
{"full_text": "Data will be made available on request.Zeolite is a microporous material comprising of aluminosilicates that can be discovered naturally in nature [1\u20133] or synthesized [4\u20137], with superior properties such as large surface area, molecular shape selectivity, and high thermal and chemical stability [8\u201311]. These properties immensely satisfy catalyst criteria. Thereby, zeolite becomes one of the most widely used heterogeneous catalysts in various applications, especially in the petrochemical industry [12\u201315]. However, the restricted access of the small micropore could limit the reactant and product diffusion, leading to slower catalytic reaction and accelerating the pore blocking formation [16\u201321]. Furthermore, it also hinders the use of bulkier reactants. Thus, the catalytic process is not appropriately utilized [22\u201324].A large number of studies have been focused on overcoming the shortcoming of micropore zeolite catalysts [25\u201327]. The strategy is to modify the zeolite structure by introducing a hierarchical structure that can be synthesized through a bottom-up or top-down process [17,28\u201330]. Hierarchical zeolite is defined as a zeolite with at least has two pore types. Generally, the second pore is mesopore type (2\u201350\u00a0nm) [31\u201333]. Another strategy is to extremely reduce the size of the zeolite crystals into nanosized. However, the synthesis of nanosized zeolite faces serious issues, i.e., low yield and difficulty in separation [34].Zeolites with two-dimensional structures (nanosheet) were intensively developed and demonstrated remarkable catalytic activity relative to conventional zeolite. The term of two-dimensional or nanosheet zeolite could be used if one of the dimensions of zeolite crystal is significantly less than several nanometers (about one- or two-unit cells). Typically, it contains a remarkable accessible external surface per mass/volume unit compared to the 3D one. Their open structure, as well as surface-exposed active sites, could address the mass transport limitation issues that occurred in zeolite catalysis [35,36]. Furthermore, it also has a high pore volume enabling bulk molecules to undergo catalytic reaction and shortening the diffusion pathway. Hence, the catalytic lifetime could be more prolonged [8\u201310]. Besides, this 2D zeolite also showed high thermal stability and strong acidity properties, which are necessary for the catalytic reaction [32].The enhanced catalytic performance in terms of activity, selectivity, and stability of nanosheet zeolite has been reported for several reactions, e.g., the benzylation of toluene [37], methanol to propylene [38], isomerization [39], and pyrolysis of biomass reaction [40]. Furthermore, for some particular reactions, the metal sometimes should be introduced in the zeolite system to supply suitable acidic strength. Thus, the reaction could be directed to generate the desired product. In this case, the nanosheet structure provides the exceptional surface area in which the metal could be well-distributed on the zeolite surface, enhancing the catalytic performance. For instance, W-substituted nanosheet zeolite shows a lower coke deposition and higher selectivity to propylene product than that of the conventional W-substituted zeolite [38]. In another experiment reported by Liu et al. [41] nanosheet zeolite could generate strong interaction with Pd nanoparticles, then restraining the Pd leaching. Moreover, it also promotes a better Pd dispersion and higher Br\u00f8nsted acid sites than the conventional zeolite. In result, the higher formation rate of H2O2 product (1.9 times) and longer catalytic stability (almost 3 times) than that of the Pd-conventional zeolite are observed.From the catalyst design point of view, several synthesis strategies have been documented. For instance, Ji et al. [42] reported the successful synthesis of MFI nanosheet zeolite by employing bifunctional organic surfactant resulted in higher resistance to coke formation. Tian et al. [43] reported that pillaring by using seed-assisted secondary growth could alleviate the negative effect of calcination resulting in nanosheet zeolite with high connected mesopore which is highly selective to light olefins and minimizes the aromatic byproduct [43]. Other than those strategies, ADOR (Assembly-Disassembly-Organized-Reorganized) method was also mentioned as a potential strategy to achieve the nanosheet structure from the existing zeolite structure [44,45].From those significant progresses that have been reached regarding the nanosheet zeolite, several reviews have been published, in which it covers the discussion of 2D zeolite using different frames of reference [46,47]. Most of the publications are well-written by Roth and his group. Particularly, they have summarized and classified both the synthesis procedure and their catalytic opportunities, i.e., petrochemistry, oxidation reaction, fine chemical synthesis, and organometallic [48\u201350]. Very recently, wang et al. [51] also published the nanosheet zeolite for catalysis, which especially focused on the zeolite that has been synthesized from a sustainable approach. In this review, we present a comprehensive overview of the recent development of nanosheet zeolites comprising i) synthesis strategies, ii) physicochemical properties, and iii) their catalytic applications. Particularly, the relationship between the characteristic of the catalyst and the catalytic performance will be highlighted. This review also contains a broader catalytic application, e.g., the conversion of biomass and photocatalysis, in which it is essential to direct the development of the energy future.Synthesis of nanosheet zeolite has been widely reported, especially zeolite with MFI and MWW framework types. Nanosheet zeolites with other types of frameworks have also been successfully synthesized, such as FAU [37,52] and MOR [53]. Mainly, the synthesis of 2D or nanosheet zeolite could be obtained by the three different synthesis procedure, i.e., (i) the hydrothermal synthesis, in which the zeolite form is as a layered precursor that produce 3D zeolite framework upon the calcination treatment. So, the layered precursor zeolite must undergo the subsequent post-treatment for obtaining the 2D structure, e.g., pillaring, delamination. (ii) Surfactant-assisted synthesis, for example, the MFI nanosheet structure. (iii) the transformation from the 3D structure of the zeolite. It is also so-called the ADOR process, in which nanosheet zeolite was generated from existing zeolite [44,45]. The two former methods could be classified as bottom-up method. Meanwhile, the latter is the bottom-up method. Other methods, including seed-induced method and chemical etching, are also elaborated.Basically, the quaternary ammonium surfactant comprised a long chain alkyl group as a tail, and the quaternary ammonium groups as a head. The latter act as a structure-directing agent for zeolite framework formation. Meanwhile, the former control the growth of the zeolite crystal, leading to the formation of nanosheet structure [45,54]. This concept was introduced by Ryoo and Co-workers [54]. They succeeded in synthesizing multilamellar MFI nanosheet zeolite with a thickness of 2\u00a0nm using bifunctional diquaternary ammonium-type surfactant C22H45\u2013N1(CH3)2\u2013C6H12\u2013N1(CH3)2\u2013C6H13 in the bromide form (C22-6-6Br2) under crystallization temperature of 150\u00a0\u00b0C for five days. The diammonium head group directed the formation of MFI zeolite, while the hydrophobic tail induces the formation of the micelle, then created ultrathin nanosheet by inhibiting the excessive crystal growth along the b-axis as already mentioned previously in subsection 2.1. Unilamellar MFI nanosheet was also obtained by employing surfactant C22-6-6(OH)2 and reducing the Na\u00a0+\u00a0concentration in the synthesis mixture with a more prolonged hydrothermal time (11 days). Na+ concentration was assumed to be a crucial factor during the crystal growth in a\u2013c axis and layer stacking along the b-axis, producing different forms of MFI nanosheet.Furthermore, Na et al. [55] reported that silica sources and ethanol strongly affect the formation of nanosheet zeolites. The presence of a small amount of ethanol could increase the surfactant movement accelerating crystal growth. Zou et al. [33] prepared a ZSM-5 nanosheet using [C18H37\u2013N+(CH3)2\u2013C6H12\u2013N+(CH3)2\u2013C6H13]Br2 (denoted as C18-6-6Br2) as a template with a similar procedure reported by Choi et al. [54]. In this case, the Na-ZSM-5 nanosheets were then ion-exchanged with Cu cations resulting in a Cu-ZSM-5 nanosheet with a large surface area (410\u00a0m2/g) and Vmeso (0.45\u00a0cm3/g) compared to conventional zeolite (381\u00a0m2/g and 0.12\u00a0cm3/g). Zhang et al. [32] also synthesized MFI nanosheet with a slightly modified procedure [54] by optimizing the amount of ethanol. The obtained MFI nanosheets showed a highly enhanced specific surface area (366\u2013665 m2g-1) compared to the conventional MFI zeolite (293 m2g-1). Accordingly, these MFI nanosheets exhibited a greater number of micropores, mesopores, and macropores than the conventional MFI zeolite [32]. Ethanol can serve as a cosolvent, promoting the crystal growth process of nanosheet zeolites by adjusting the surfactant micelle structure. Moreover, a larger volume of ethanol can produce a disordered structure of nanosheets [55].Wei et al. [15] proposed the mechanism of self-interlocked ordered nanosheet stack (SI-ONS) in C22-6-6Br2 system with variation of anions and cations of precursors as described in Fig. 1\n; the mechanism consists of 4 steps; specifically, i) formation of micelles from the surfactant, ii) formation of silicate micelles where Br\u2212 ions are replaced by silicate ions, iii) aggregation of silicate micelles through condensation due to lower silicate surface charge, and iv) the silicate micelle aggregates turn into a solid mesophase then the transformation and crystallization occurs simultaneously. In addition to that, Wei et al. [15] also explored the effect of precursor cations and anions on the textural and acidity properties of self-interlocked ordered nanosheet stack. The result showed that variation of cations and anions influence the formation of inter-crystalline mesopore by cation order of Na+> K+> Rb+> Cs\u00a0+\u00a0and anion order SO4\n2\u2212> NO3\n\u2212> Cl\u2212. Besides, XRD analysis showed improvement of crystallinity with the increase of cation size. Based on the NH3-TPD result, anions and cations of precursors showed negligible effect on the number and strength of strong acid sites in the resulting 2D MFI products.Despite the use of diquaternary ammonium surfactant, the monoquaternary has also been reported to synthesize the zeolite nanosheet by Xu et al. [56]. A single-crystalline mesostructured MFI zeolite nanosheets (SCZN) was produced by using surfactant-containing aromatic group, i.e., C6H5\u2013C6H4\u2013O-(CH2)m-N+(CH3)2C6H13 (Br-) coded as CPh-Ph-m-6. The variation of carbon chain length (m) from 6 to 10 results in the various interlamellar d-spacing of 1.7\u20132.1\u00a0nm. Biphenyl groups can interact with each other through \u03c0-\u03c0 interaction, stabilizing lamellar micelles and aligning the MFI nanosheet structure. Therefore, \u03c0-\u03c0 interaction and is considered an important function in preserving the shape of the nanosheet [57].However, the limitation of using bifunctional structure-directing agents (SDAs) is the reduction of ordered mesopores during the calcination process. An alternative strategy to hinder the destruction of MFI zeolite layers is by introducing pillars in the interlayer space to support the ordered mesoporous structure in the zeolite. In 2010, Ryoo and co-workers [55] reported a method to preserve the ordered multilamellar MFI nanosheet zeolites by intercalating silica pillars in the interlayer spacing of the MFI nanosheet.Despite of the MFI, the use of the quaternary ammonium surfactant was also reported for SAPO-34 nanosheet synthesis. In the study reported by Chen et al. The nanosheet morphology was assembled as the results from the existence of [3-(trimethoxysilyl)propyl] octadecyl dimethyl-ammonium chloride (TPOAC) as mesoscopic aggregation. Meanwhile, the formation of CHA structure was assisted by tetraethylammonium hydroxide. Notably, the surfactant functionedas a stabilizer of the energy surface of the morphology. The amount of TPOAC modulates the formation of the unique aggregation morphology. Furthermore, the hydrothermal synthesis condition are also strongly related to the crystal formation [58].The structure of mesopores is possibly destroyed at high-temperature forming new Si\u2013O\u2013Si bonds between adjacent nanosheets. The formed 2D crystalline framework leads to a partial loss of uniform mesopores and surface area by interlayer condensation. Alternatively, the pillaring method evidently reduces the calcination effect issue resulting in more preserved lamellar nanosheets after high temperatures treatment [26,55,59].Na et al. [55] introduced a nanosheet zeolite pillar through hydrolysis by adding TEOS (tetraethyl orthosilicate) into the non-calcined as-synthesized nanosheet zeolite and transforming it into silica pillars. As shown in Fig. 2\na, the surfactants form spherical micelles with a hexagonal phase which then transform into lamellar mesophase; with the process involved, the head of the two SDA facing each other will enter in a straight channel on the crystal axis-b. The resulting pillared nanosheet zeolites showed a high surface area (615\u00a0m2/g) with a large pore volume (0.44\u00a0cm3/g) as a result of the maintenance of a regular multilamellar structure during SDA removal. Apparently, the introduction of silica during the pillaring process covered the Al sites; thus, pillared nanosheet zeolite showed reduced external acid sites compared to non-pillared one. However, the reduction in external acid sites by pillaring is highly tolerable as it still showed greater external acid sites than the conventional MFI.On the other hand, a large molecule can reach zeolite micropores from any direction. Moreover, the MFI framework remains 3D pore connectivity along the crystal axis and affords a high molecular diffusion efficiency. Besides, the interlayer spacing can be easily adjusted by surfactant tail length. Another necessary aspect of this pillared zeolite is the mesopore surface, described as the nanosheet zeolite external surface, terminated with a great number of silanol (-Si-OH) groups. Thereby, the surfaces can be functionalized through a silylation process to change surface features such as hydrophilic or hydrophobic properties [55].Using the same pillaring precursor TEOS, Ali et al. [59] also prepared a self-pillared MFI nanosheet. The procedure of the synthesis process is depicted in Fig. 2b. Based on XRD measurement, the self-pillared nanosheets exhibited a very sharp peak at (501). Thus, the characteristic peak of the pillared nanosheet shows the same peak as the nanosheet that appears in the h0l plane. This occurrence is because the crystals grow in the a-c-plane direction. In addition, the characterization using high-angle annular dark-field (HAADF) STEM result exhibited that the nanosheet zeolite has a wide a-c plane with a mesospace-interlayer and thickness of \u223c2\u00a0nm. Compared with conventional zeolite MFI, the unique structure of the nanosheet self-MFI nanosheet pillar showed enhanced catalytic activity, selectivity, and hydrothermal stabilityTian et al. [26] reported comparison results of pillared HZSM-5 nanosheets prepared by two different pillaring methods, dual-template and Si intercalation, by using TPAOH and TEOS as pillaring precursors, respectively. Both pillaring methods efficiently protect the mesopores of HZSM-5 nanosheets from collapse during the calcination. However, the dual-template method showed superior channel connection and more preserved acidity.Chang et al. introduced a new type of triblock SDA (N3-POn-N3) with linker tripropylene oxide (PO3) for direct synthesis of highly-branched and pillared MFI nanosheet. The branching of the MFI nanosheet was a result of the short PO6 linker of N3\u2013PO6\u2013N3 that limited the movement of the SDA. Accordingly, the degree of branching (density of orthogonal stacking of bundles of nanosheets) can be adjusted by the length of the POn blocks of the SDAs [60]. The use of longer linker N3\u2013PO68\u2013N3 led to a reduction of branching and partially generated random assembling of the stackings of nanosheet. Furthermore, adjusting the linker length of triblock SDA can control the textural properties of nanosheets, including the degrees of branching and the distance between adjacent nanosheets [61].In addition, pillaring of zeolite nanosheet to produce more open structure could be carried out through vapor phase pillarization (VPP) method as reported by Wei et al. [62]. In their work, they integrate the three typical processes in zeolite pillarization, i.e., intercalation, hydrolysis, and calcination, into single operation using only one apparatus in which the quartz \u201cU\u201d-shaped tube was placed in a furnace. The liquid alkoxide as a pillaring precursor was added dropwise into Teflon cup and evaporated for intercalation. The hydrolysis of the intercalated alkoxides was also performed in the same set-up. Meanwhile, the 2D zeolite was placed in a separated glass tube container. The result showed that this method required less alkoxide about ten times than the typical method. Furthermore, almost 100% efficiency and TEOS utilization for the formation of pillared zeolite were obtained with the reduction of liquid waste. As inert spacers, the use of SiO2 as pillars is associated to the partial blockage of the active site which present between the external lamellar surface. In this case, Schwanke et al. [63] has reported that the addition of active species, i.e., Niobium Oxide as a mixed pillars could improve the catalytic activity of a pillared lamellar MWW zeolites.Other than those strategies of pillaring zeolite, the epitaxially growing the layered zeolite on the bulk one has also been reported. In this case, the bulk zeolite act as support to avoid the condensation and disordering the interlayers during the calcination process. Furthermore, it also provides the shape selectivity properties and the acid site, which are very important for the reactions. Therefore, this hybrid material possesses a good catalytic performance on the methanol to propylene reaction [64,65]. However, despite pillaring effectively preserve the mesopore structure of nanosheet zeolite during the SDA removal, the silica pillars quickly disintegrated under ambient conditions. Besides, the synthesis process is complicated and time-consuming. Thus, the synthesis method delimits its practical application.Zeolite MFI nanosheet has been successfully synthesized by Jeon et al. [66] through a bottom-up process using MFI seeds with a diameter of 30\u00a0nm. In this process, nanosheet zeolite is formed through single rotational growth after the seeds reached a specific size and shape [66]. The schematic of the hierarchical ZSM-5 (Hi-ZSM-5) nanosheets can be seen in\u00a0Fig. 3As shown in Fig. 3, Liu et al. [67] reported synthesizing the nanosheet zeolite using the seed method assisted hexadecyl trimethyl ammonium bromide (CTAB) as the second template. Concerning the first step, the ZSM-5 seed was destroyed to form sub-nanocrystals following the aging process. The distinct approaches for nanosheet zeolite synthesis processes included adding CTAB into the mixture to produce hierarchical ZSM-5 (Hi-ZSM-5) at high temperatures. Zeolite Hi-ZSM-5 exhibits a particle size of 2\u20134\u00a0mm with a honeycomb-like morphology.Shang et al. [68] prepared nanosheet zeolite by applying ZSM-5 seed in the presence of di-quaternary ammonium surfactant (C18-6-6Br2). The seed-assisted strategy provides a positive result in decreasing the crystallization time and the particle size. For instance, the crystallization times were reduced linearly from 120\u00a0h to 24 has the seed content increased from 0 to 30\u00a0wt%. Meanwhile, increasing the ZSM-5 seed content also led to a reduction in particle size. Besides, the acidity and textural properties of nanosheet zeolites are easily tuned through this method (by variation of number seed)Xing et al. [69] synthesized nanosheet zeolite through seed-induced with the addition of urea as a crystal growth inhibitor. The seed crystals promote the formation of nuclei to form small particle sizes. Different types of seeds, spherical and sheets-like ZSM-5, contributed to alterations in acid properties, pore structure, and defect sites of the nanosheet zeolites. The use of ZSM-5 sheet-like seed produced nanosheet zeolite with slightly higher surface area (437\u00a0m2/g) and total acidity 123\u00a0\u03bcmol/g compared to that synthesized from spherical ZSM-5 seed (416\u00a0m2/g, 114\u00a0\u03bcmol/g).Chemical etching is one of the top-down methods for the synthesis of nanosheets. Prech et al. [70] synthesized layered zeolite by using fluoride etching (NH4F). The concentration and strength of the acid sites and textural properties of the layered zeolite can be altered by fluoride etching. The external surface area of the as-synthesized layered zeolite increased by hardness treatment (from 86\u00a0m2/g became 92\u00a0m2/g after 30\u00a0min). In contrast, the external surface area of the calcined samples did not change with both mild and hard treatment. The fluoride treatment can remove all species that are not an integral part of the zeolite structure. Consequently, the materials treated after calcination restrain the Al framework. The addition of CTAB surfactant can increase the Lewis acid site concentration in proportion to the amount of extra Al frameworkBesides, Zhou et al. [71] introduced chemical etching using alkaline media, i.e., TPAOH, to transform the bulk microcrystal 3D ZSM-5 zeolites into hierarchical nanosheet zeolites (Fig. 4\n). Typically, a bulk zeolite crystal was treated with TPAOH solution, sonicated, and finally conducted in a hydrothermal treatment at 170\u00a0C for 18\u00a0h under a rotational convection oven. During this hydrothermal treatment, the dissolution-recrystallization mechanism may occur as indicated by declining crystallinity. As a result, a hierarchical zeolite crystal with a hollow structure was observed. Furthermore, the dissolved silicon and aluminum species may diffuse toward the external surface of hollow zeolites, and the subsequent secondary nucleation and crystal growth may exist on their a/c faces. Then, a longer etching period allows the thinning of the crystal walls progressively in all directions. Notably, the largest crystal faces in ZSM-5 zeolites were the faces perpendicular to the direction of the b-axis; therefore, only [0 k 0] faces of bulk zeolites survived, resulting in the formation nanosheet structure in 18\u00a0h of the etching period. However, the thickening of the nanosheet structure was observed in the prolonged etching period at 24\u00a0h, owing to the unbalanced rate between the dissolution and recrystallization rate.The modification of layered zeolites, such as delamination, was usually preceded by the breakage of interlayer bonds and the expansion of interlamellar space called swelling. A high concentration of surfactant and hydroxide (OH\u2212) combined with elevated temperature are often needed for performing this process. Corma et al. [72] have applied this technique to generate the delaminated zeolite ITQ-2 from the MWW-type structure zeolite precursor, i.e., MCM-22. The precursor was firstly swollen by refluxing the slurry of the solid sample (mixture of MCM-22\u00a0+\u00a0water) with an aqueous solution of 29\u00a0wt% hexadecyltrimethylammonium bromide and 40\u00a0wt% of tetrapropylammonium hydroxide for 16\u00a0h at 353\u00a0K. After the subsequent ultrasound, acid treatment, and calcination, a new structure of nanosheet zeolite ITQ-2 was obtained. The much higher well-defined external surface area in the ITQ-2 provides a more accessible acid site. Thus, it exhibits a better catalytic activity than MCM-22 or MCM-36, in which they are its parent and its pillared form of MCM-22, respectively [73]. The schematic illustration of the different synthesis technique was islustrated in Fig. 5\n.Regarding the severe condition needed for the swelling process in conventional delamination zeolite, recently, the fluoride or chloride anion-promoted exfoliation is expected to realize the delaminated zeolite under milder conditions. Eilertsen et al. [74] reported the synthesis of UCB-2 zeolite from the delamination of zeolite layered precursor MCM-22 using a mixture of cetyltrimethylammonium bromide, tetrabutylammonium fluoride, and tetrabutylammonium chloride in N,N-dimethylformamide (DMF) as solvent. Further treatment using concentrated acid at room temperature could result in delaminated zeolite, i.e., UCB-2. They suggest that the delamination may be facilitated by the organic solvent in a certain manner since it was reported to successfully exfoliate the layered materials such as hydrotalcites.The use of exfoliation method has also been reported for SAPO-34 nanosheet, which is generated from lamellar SAPO-34. Practically, the lamellar zeolite was exfoliated in mild condition through the strategy of the solvent-mediated freeze-thaw process. Firstly, the lamellar SAPO-34 was dispersed in hexane and frozen in a liquid nitrogen bath. Then, subjected into a sonication process. After repeated this process for 20 times, the sonicated samples was mixed with the excessive ethanol and aged for 12\u00a0h. Notably, the unexfoliated crystal will be settled down in bottom, in which it is separated by centrifugation process. Meanwhile, the upper suspension was separated and dried for 30\u00a0min. SAPO-34 nanosheet suspension was obatined by dispersing the dried sample in ethanol and subjected to a sonication process for 2\u00a0h. Finally, the nanosheet with 4\u00a0nm of thickness is generated [75].Basically, the ADOR process contains the disassembly of the previously assembled zeolite selectively and controllably into layered building units (Fig. 6\n). Then, organize them into a suitable orientation and finally reassemble again into a new zeolite structure through a condensation. This method could be used to obtain a certain novel zeolite framework that could not be prepared through direct hydrothermal synthesis. For instance, the IPC-2 (OKO) and IPC-4 (PCR) could only be obtained through the ADOR process of germanosilicate UTL zeolite [76,77].Cejka and Coworkers [44] demonstrated the seriously altered XRD characteristic of zeolite UTL after the hydrolysis process (room temperature to 100\u00a0\u00b0C of hydrothermal treatment at pH neutral to acidic (0.1\u00a0M HCl)), which indicate the significant structural change assigned to the transformation of the 3D to 2D zeolite (also confirmed by TEM). During the hydrolysis process, the interlayer space was contracted depending on the boron content, liquid media (water or acid solution), and temperature. In addition, the D4R units and their connectivities to the original layer were broken. Finally, the nitrogen isotherm characterization of the calcined hydrolyzed UTL showed that this new material has a typical microporous profile volume of 0.095\u00a0cm3/g and BET 270\u00a0m2/g with no mesoporous character.Nanosheet zeolite has ultrathin sheet morphology with thickness ranging from 2 to 100\u00a0nm [79\u201381]. A model of the nanosheet zeolite structure has been proposed by Ryoo et al. [54] The model depicts that the surfactant molecule aligns along the channels of the MFI framework. An assembly is built along the b-axis to form a multilamellar structure or a random assembly to form a unilamellar structure [54]. This unique structure impacts the distribution of the active sites. Different from 3D zeolite which the active sites mainly existed in the micropore, most active sites in 2D zeolites exist on the external surface [82]. Thus, a solid understanding of the structural and chemical properties of nanosheet is indispensable to enhance catalytic performance.Numerous investigations on the morphology of nanosheet zeolites have been reported and are depicted in Fig. 7\n. It should be noted, the term morphology which is used in this section was related to both of nanosheet itself and the morphology of the assembled zeolite nanosheets. A distinct difference between the conventional and nanosheet form of MFI zeolites (Fig. 7a\u2013g) is clearly observed. Conventional ZSM-5 (Fig. 7a) shows a typical hexagonal-prismatic or coffin shape morphology [83,84]. Kadja et al. [85] reported a nearly perfect spherical morphology for conventional ZSM-5 zeolite synthesized through the low-temperature synthesis method (LTS). Similar thickness in three different directions was observed, indicating that this spherical morphology is formed from three-dimensional crystal growth. Therefore, the morphology of conventional ZSM-5 zeolite is strongly influenced by crystal growth [86]. Meanwhile, the MFI nanosheet zeolite shows a thin sheet morphology with various assemblies (Fig. 7b\u2013g). Hu et al. [87] obtained an MFI nanosheet zeolite with particle size around 4\u00a0\u03bcm and confirmed that the obtained nanosheet zeolite has a lamellar stacking morphology with the nanosheets intergrowth in a three-dimensional direction (Fig. 7b). This morphology was achieved by using the bifunctional organic surfactant, i.e., [C18H37\u2013N+(CH3)2\u2013(CH2)6\u2013N+(CH3)2\u2013C6H13] Br2\nIn the seed-assisted synthesis of nanosheets, the particle size of the obtained zeolite is greatly influenced by the seed content. The use of 5\u201330\u00a0wt% seed in the gel resulted in particle sizes from 0.8 to 1.2\u00a0\u03bcm to\u223c500\u00a0nm (Fig. 7c) [68]. Using the similar approach, Fang and coworkers [88] demonstrated that the nanosheet assemblies could be tuned by adjusting the Si/Al ratio. In this case, the spherical morphology of the nanosheet zeolite with a house-of-cards structure was obtained when Si/Al ratio is 31 (Fig. 7d). The lower Si/Al ratio resulted in the formation of an irregular stack, whereas the higher Si/Al ratio generated a nanosponge-like morphology. Besides, Hao et al. [89] customize the nanosheet thickness by arranging the proportion of surfactant and tetraethyl orthosilicate (TEOS) (Fig. 7g). An enormous amount of SDA leads to a denser morphology. Further, the thickness of the nanosheet structure was related to the ratio between mesopore and micropore, in which it strongly affected their catalytic activity. A similar result was also reported by Xu et al. [81], in which the thickness of aluminosilicate ferrierite (FER) nanosheet zeolites (6\u2013200\u00a0nm) is strongly correlated with the amount of N,N-diethyl\u2010cis\u20102,6\u2010dimethylpiperidinium (DMP) as a structure directing agent in the starting gel.Very recently, Li et al. [93] demonstrated the synthesis of ultrathin FER zeolite nanosheets named SCM-37 zeolite using octyltrimethylammonium chloride (OTMAC) and 4-dimethylaminopyridine (4-DMAP) as dual templates. A study using 13C MAS NMR revealed the involving oh those two organics in the zeolite formation. Particularly, the ammonium head group of OTMA+ is located in FER cages, 8-MR, and 10-MR channels. Meanwhile, no 4-DMAP is found at those location. Thus, FER layers might be the location of the 4-DMAP, which then inhibiting the crystal growth in the a-direction.Multilamellar structure with a petal-like morphology (Fig. 7e) was observed for the MFI nanosheet zeolite synthesized from CPh-Ph-m-6 surfactant [56]. Park et al. [18] investigated the effects of the multi-quaternary ammonium surfactant structure (tail-N+(CH3)2\u2013{spacer\u2013N\u00a0+\nR\n2}\nn\u22121\u2013R*) on the synthesis of MFI nanosheet zeolites including the spacer length between the ammonium, alkyl groups in the terminal ammonium (R), length of surfactant tail, and the number of ammonium groups. The variation of \u2013C3H6-, \u2013C6H12-, and \u2013C8H16- spacers (CiH2i) in the C22\u2013i\nN2 [i.e., C22H45\u2013N+(CH3)2\u2013C\ni\nH2i\n\u2013N+(CH3)2\u2013C6H13] leads to the formation of micrometer-sized bulk crystals, multilamellar mesostructure, and disordered nanosheets, respectively. In addition, alkyl group moieties in the terminal ammonium also greatly influence zeolite crystallization. The decrease in the hydrophilicity of the alkyl group tends to hinder the SDA function. Thus, in the formula of C18\u20136N2(R) surfactants, the MFI nanosheet zeolites are only formed when Me2, Et2, and Pr3 are used as the alkyl moieties (R). In reverse, the use of Pr2 and But2 lead to the formation of an unknown silicate phase. In the case of the length of the surfactant tail, C8 and C6 are considered not hydrophobic enough to form an MFI nanosheet. Meanwhile, the thickness of the MFI nanosheet zeolite is reportedly increased with the number of ammoniums.Several researchers have also revealed nanosheet zeolites for other framework types. FAU nanosheet shows a ball-like morphology as depicted in Fig. 7h []. Salakhum et al. [90] reported that the morphology of the FAU nanosheet could be affected by the amount of surfactant, 3-(trimethoxysilyl)propyl octadecyl dimethyl ammonium chloride (TPOAC). A high amount of TPOAC could inhibit the crystal growth obstructing the morphological assembly of nanosheet. Meanwhile, the textural properties can be controlled by the synthesis conditions such as crystallization temperature and the amount of surfactant. Typical morphology of FAU nanosheet was observed uniformly at a crystallization temperature of 85\u00a0\u00b0C and a TPOAC molar fraction of 0.030 []. Ferdov et al. [91] also reported an FAU nanosheet assembly with ball-like morphology and a diameter of 2\u00a0\u03bcm\u00a0at a crystallization temperature of 65\u00a0\u00b0C for 96\u00a0h (Fig. 7i). By employing silane surfactant, Fu et al. [52] obtained a Y nanosheet with a flower-shaped cards-like morphology, the thickness of 50\u00a0nm, and particle size of 2\u20135\u00a0\u03bcm (Fig. 7j).Zhou et al. [92] reported a disordered 2D MWW zeolite with noticeable uniform thickness (Fig. 7k) through the one-pot synthesis in the presence of long-chain surfactant cetyltrimethylammonium (CTA). This surfactant CTA could adjust the Al position without impacting the final product of the Si/Al ratio. This technique can be applied to layered zeolites to produce high surface area nanosheets []. In the case of template-free synthesis, the nanosheet mordenite was successfully synthesized with a nanosheet stack morphology (Fig. 7l) with a thickness of about 50\u2013100\u00a0nm assembled. A higher H2O/SiO2 molar ratio contributed to the transformation of a single plate of the building unit with a thickness of around 3\u00a0\u03bcm into a bundle of nanosheets about 50\u2013100\u00a0nm, suggesting that the decrease in basicity (higher H2O/SiO2 ratio) favors the formation of nanosheet structure [53].For the M-substituted zeolite, the MFI nanosheet with various subtituate metals, i.e., Al, Ga, and Fe was successfully obtained by Ji et al. [42] using diquaternary ammonium-type surfactant. The as-synthesized nanosheet zeolite exhibited a petal-like morphology composed of stacked intergrown crystals (Fig. 7m-o). Those inter-grown crystals can act as a pillar that prevents the structure collapse when the SDA is removal by calcination. Besides, M-substituted MFI nanosheets with other metals, i.e., Mn, Ce, W were also reported by Hadi et al. [38] showing disordered multilamellar structures composed of interlinked ultrathin MFI type zeolite nanolayers (Fig. 7p). Shown in the image that the thicknesses of the nanolayers are relatively uniform along the crystal axis perpendicular to the nanosheet layers.TEM and HRTEM images of nanosheet zeolites are shown in Fig. 8\n, showing the ultrathin morphology of the nanosheet zeolite. Multilayer piles observed in the TEM image indicate the presence of mesopores between the gaps of the nanosheet [37]. In several cases, the lamellar stacking of nanosheet zeolite was clearly observed enabling us to further identify its component. For instance, Ryoo and co-workers [54] found that the stacking was comprised of the layer of MFI framework with a thickness of 2\u00a0nm and surfactant micelle with a thickness of 2.8\u00a0nm. The former is assigned to a single unit cell (three pentasil sheet) dimension along the b axis (b\u00a0=\u00a01.9738\u00a0nm) [54]. Figure 8a shows MFI zeolite nanosheets with a thickness of 2.0 nm on the b axis and a width of 60 nm on the a-c axis [130]. A similar MFI-surfactant layer stacking was also observed by several researchers (Fig. 8b\u2013d). According to Liu et al. [23], the thickness of the surfactant layer was affected by the surfactant used. In this case, the C6-12-diphe (\n\n\u223c\n\n 2.5\u00a0nm) results in the thicker layer than that of C6-6-diphe (\n\n\u223c\n\n 2.0\u00a0nm). Furthermore, the molecular structure of surfactant strongly affected the morphology of MFI nanosheet zeolite. Variation of packing parameters between surfactants contributed to the formation of different mesophases. The packing parameter (g) is expressed as the volume of the surfactant chain (V) to the effective area of headgroup (a\n\n0\n) and length of the surfactant (l\n\nc\n) [55,94]. Mainly, the surfactant packing parameters can manage the growth of the 2D structure of the zeolite. Furthermore, the properties of MFI nanosheet zeolites, including textural and morphologies, can be tunned with different alkyl spacers of surfactants [23].In the study conducted by Park et al. [18], spacers between ammonium in the diaquaternary ammonium surfactant affected the morphology of the MFI nanosheet. The multilamellar MFI nanosheets with a thickness of 2\u00a0nm in Fig. 8c were generated by employing surfactant C22\u20136N2. Meanwhile, when surfactant with longer alkyl linkage such as C22\u20138N2 was used, a disordered MFI nanosheet was generated [18]. This result is probably can be described by packing parameters. Moreover, the multilamellar morphology of the MFI nanosheet is also confirmed by another report, as depicted in Fig. 8d [31]. Furthermore, by using the procedure reported by Ref. [54], Zou et al. [33] reported a lamellar stacking with lettuce-leaf-like morphology (the thickness around 20\u201330\u00a0nm) with overall particle size of 500\u00a0nm for the MFI-Cu nanosheet (Fig. 8e). In contrast, conventional zeolite exhibits a sphere particle consisting of small particles with a diameter around 250\u2013500\u00a0nm and a bigger particle (1\u00a0mm) [33].Other than the stacking of the nanosheet structure, TEM images could also reveal that nanosheet zeolite provides the compatible surface for the metal to be well-dispersed. For instance, Pd was successfully dispersed on the outer of the surface Y nanosheet (Fig. 8f, red arrow indicates the Pd particles (1\u20133\u00a0nm) spread on the surface nanosheet) [52]. Besides, Yutthalekha et al. [37] revealed that the FAU nanosheet zeolite which obtained by using TPOAC (3-(trimethoxysilyl)propyl octadecyl dimethyl ammonium chloride) as surfactant is not only contain the multilayer stacking of nanosheet of nanosheets with the interstitial mesopores but also the mesopores cavity as clearly shown in Fig. 8g.In the case of titanium silicalite-1 (TS-1) nanosheets prepared by using a di-quaternary ammonium template (C22-6-6Br2), a cylinder-like bulk was observed as shown in Fig. 8h with a thickness of \u223c200\u00a0nm, an average width of \u223c250\u00a0nm, and an interlayer distance of 3.0\u00a0nm [95]. Lu et al. [96] reported the MOR nanosheet with a highly ordered layer and 11\u00a0nm of thickness (Fig. 8i) was achieved by using a bifunctional SDA, Gemini-type amphiphilic surfactant. The benzyl quaternary ammonium cations directed the formation of MOR topology, while the hexadecyl tailing group possibly formed a hydrophobic barrier in the micelles preventing the continued crystal growth along the b-axis. In contrast, the TEM image of the MWW nanosheet (Fig. 8j) synthesized by Zhou et al. [92] shows randomly oriented nanosheets. In addition, ITQ-2 prepared by delamination of MCM-22 zeolite precursor resulted in a random aggregation of nanosheets with a thickness of 5\u201310\u00a0nm (Fig. 8l), while MCM-22 exhibits disc-shaped particles with thickness of 20\u201330\u00a0nm (Fig. 8k) [97].X-ray Diffraction is the primary tools to characterize the structural properties especially the crystallinity of both 3D and 2D zeolites. Different from 3D zeolite, 2D Nanosheet zeolite has a diffraction pattern that are often broad. A sharp reflection is appeared in the h0l lattice plane, reflecting the 2D order of the layers, in which crystal growth occurs only in the a-c plane [45]. Also, quite similar to zeolite nanoparticles, the wider peak which is observed is correlated to the smaller particle size compared to the conventional 3D zeolite [98]. Fig. 9\na demonstrates the diffraction pattern of the 2D and 3D structures of Cu-ZSM-5. A sufficient sharp of the peak was only detected for h0l reflections confirming that the thickness along the b-axis is extremely small or broadened for Cu-ZSM-5 nanosheet Meanwhile, for the conventional one, all the hkl reflections are clearly observed [33].Furthermore, the low angle x-ray diffraction could also be used to identify the presence of the periodic interlamellar structural order as reported by Xu et al. [57]. The layered MFI nanosheet structure was indexed by the peak at 2\u03b8 1.84\u03bf and 3.65\u03bf (Fig. 9bi), respectively assigned to the first and second-order reflection of layered MFI. In this case, the new unit cell constant of B\u00a0=\u00a04.8\u00a0nm, whereas the A and C parameters are still the same as MFI unit. This technique was also applied by the same group to evaluate the various single head ammonium surfactant in generating the MFI nanosheet zeolite (Fig. 9bi,ii). Actually, from the high angle XRD pattern, it was already observed that the 1 and 2 samples did not produce the lamellar structure indicated by the sharp peak of Bragg reflections in all directions. For sample 3, identification of the pattern is still elusive. In this case, applying the low angle XRD could help to further confirm the presence of the nanosheet structure, although it also could be obtained by performing the SEM or TEM characterization [56].Corma and Coworkers [73] also performed XRD analysis to investigate the structural evolution of the layered precursor MWW structure during the post-synthesis process (Fig. 9c). As the precursor was swollen using CTMA+, the XRD pattern was significantly changed. Only the peak at 18\u03bf \u2212 28 \u03bf was sufficiently observed. Also, the shift of the peak at 3\u20137 to the lower 2 theta followed by the increase of intensity is indicated the increase in the distance between the layers from 2.7\u00a0nm to 4.5\u00a0nm. Furthermore, pillaring those swollen materials results in the change again in XRD pattern due to the formation of the new material which has been known as MCM-36. On the contrary, the delamination of that swollen material removes the peak at 2 theta 3\u20137. Moreover, the high angle peak was much broader, and the peak intensity was much smaller compared with its MWW precursor, suggesting that both crystal size and its previous long order are the dramatically reduced. Then, a new material called ITQ-2 was generated from this process. A significant structural change also occurred on the ADOR processes. The alteration of the XRD pattern of UTL zeolite and its transformation during the hydrothermal treatment at room temperature to 100 \u03bfC was depicted in Fig. 9d. The abundance peak of 3D UTL zeolite was dramatically reduced to several low-intensity peaks, in which the dominant 2 theta position was assigned to the 1.2 \n\n\u00b1\n\n 0.1\u00a0nm\u00a0d\u2212spacing. Furthermore, the interlayer reflections of UTL framework (hkl) are disappeared and the low intensity peaks was resulted form the reflection of (0\u00a0kl) indices. Lately, the lamellar materials called IPC-1P was identified as the product form the structurally modification of 3D zeolite materials [44].In the case of porosity, the N2 adsorption-desorption was usually performed to evaluate the pore volume and surface area. Typically, nanosheet zeolite contains a higher mesopore volume and external surface area. For instance, numerous researchers reported that MFI nanosheets exhibited type IV isotherms [23,32,98\u2013100]. In the study reported by Zhang et al. [32], the obtained MFI nanosheets showed a hysteresis loop with the capillary condensation at the relatively high pressure from 0.4 to 0.9, indicating the presence of larger pores (mesopore and macropore). The specific surface areas of the obtained MFI nanosheet zeolite were varied in the range 366\u2013665\u00a0m2\u00a0g\u22121, which are greater than that of conventional MFI zeolite (293\u00a0m2\u00a0g\u22121). In line with that, enhanced total volume, mesopore volume, and micropore volume of MFI nanosheet were observed in the pore-size distributions calculated from Barrett\u2013Joyner\u2013Halenda (BJH) method. The calcined MFI nanosheets also exhibited broad distribution of mesopore diameter, suggesting partial condensation of MFI nanosheets during the calcination [23,32].\nFig. 10\na displays nitrogen adsorption for ZSM-5 nanosheet prepared by Xiao et al. [100]. Compared to the conventional 3D one, the ZSM-5 nanosheet demonstrate a highly mesoporous structure. The surface area for the latter(515\u00a0m2/g) is much greater than conventional (307\u00a0m2/g). Besides, they also reported an external surface area of ZSM-5 nanosheets (260\u00a0m2/g). Similar character was also observed for pillared MFI nanosheet (Fig. 10c) [62]. A study by Yutthalekha et al. [37] shows that the textural properties of nanosheet zeolite (in this case, for FAU zeolites) is strongly affected by the amount of TPOAC (3-(trimethoxysilyl) propyl octadecyl-dimethylammonium chloride) as a template. The total surface area and micropore volume decrease with the increase of the TPOAC/Al2O3 ratio from 0.01 to 0.04, i.e., from 734 to 566\u00a0m2/g, and from 560 to 407\u00a0m2/g, respectively. Meanwhile, the micropore volumes of nanosheets also decrease from 0.22 to 0.16\u00a0cm3/g. Other than that, the nanosheet zeolite obtained by the pillaring method was also reported to provide a much higher mesopore volume and mesoporous surface area that of the conventional one. Moreover, it is observed that V\n\nmeso\n of MFI nanosheet prepared by intercalation method (0.398 cm3g\u22121) is higher than the dual template method (0.369 cm3g-1). Meanwhile, the microporous surface area (S\n\nmicro\n) of the nanosheet prepared by the dual template method (221 m2g-1) is 1.38 times higher than that of the intercalation method (160 m2g-1). These results show that the intercalation method is more efficient in generating mesoporous structures, while the dual template method produces more microporous structures [26].Corma et al. [73] demonstrated the transformation of the textural properties during the post-synthesis of layered precursor MCM-22 zeolite (Fig. 10b). The progressive increase of the total pore volume and surface area was observed, in which the nanosheet zeolite obtained from the delamination process (ITQ-2) exhibited the highest value compared to the pillared one (MCM-36) and its layered precursor (MCM-22). For the ADOR synthesis process, the effect of molarity of the hydrolysis media was investigated by \u010cejka and Co workers [76]. By choosing the suitable molarities, the surface area and pore volume in the range of 150\u2013590 m2g-1 and 0.06\u20130.22 cm3g-1, respectively, could be adjusted (Fig. 10d).Acidity property in the zeolite active sites is the most crucial part of catalytic reaction. The acidity of the zeolite is determined by the Si and Al content in the zeolite structure. Several characterization techniques have been reported to investigate the acidity properties of nanosheet zeolites, including NH3-TPD, 27Al MAS NMR, and FTIR. The acidic strength and densities of nanosheet zeolites can be examined by NH3-TPD, in which the weak and strong acidity could be observed in the peaks of 190\u2013230\u00a0\u00b0C and 405\u2013475\u00a0\u00b0C, respectively. Al coordination in the zeolite structure can be determined using 27Al Mas NMR. The peak at 54\u00a0ppm is assigned to the tetrahedrally coordinated Al in the zeolite framework. Meanwhile, the type of acidity/basicity could be analyzed by FTIR. The hydrogen atom is observed at peaks of 1445 and 1600\u00a0cm\u22121. The Lewis acid sites (LAS) can be detected at 1454, 1487, and 1625\u00a0cm\u22121. Meanwhile, the Br\u00f8nsted acid sites (BAS) could be observed at 1487, 1540, and 1634\u00a0cm\u22121) [61].Acidity properties could be improved by increasing the crystallinity of the zeolite (under the same SiO2/Al2O3 ratio). It is because higher crystallinity means higher internal-framework Al sites and lower extra-framework Al sites. Meanwhile, the internal-framework Al sites contribute more significantly to the improvement of acid strength since internal-framework Al sites have better tetrahedral geometry [32]. Based on research conducted by Verheyen et al. [31], the amount of acid for nanosheet zeolite and conventional zeolite is almost the same when both 2D and 3D zeolite have the same amount of Al [].\nFig. 11\na displays two desorption peaks for nanosheet zeolite (denoted NS), The seed-fused ZSM-5 nanosheets (CNS-x), x refers to the number of seeds, and B-incorporation sample (B\u2013CNS-5) showing weak and strong acid at 180\u2013250\u00a0\u00b0C and 405\u2013475\u00a0\u00b0C, respectively. The addition of seed seemingly leads to weaker acidity since the strong acid peaks of CNS-x samples shift to a lower temperature, and boron incorporation further lowers the acid strength. Pyridine-adsorbed FT-IR (Py-IR) in Fig. 11b reveals the presence of BAS and LAS at around 1450\u00a0cm\u22121 and 1545\u00a0cm\u22121, respectively, for all resulting nanosheet zeolites. Besides, the pyridine adsorptions on both BAS and LAS are observed at the peak of 1490\u00a0cm\u22121. The non-seed-fused nanosheet zeolite showed a remarkably enhanced BAS/LAS ratio (5.57) compared to the nanosheet zeolite with seed addition. Furthermore, boron incorporation generated more total acidity but less BAS number, suggesting that boron incorporation affects the Al distribution and increases the Si/Al ratio[68]. It can be concluded that the synthesis approach of the nanosheet zeolite strongly affects Al species coordination, whereas the more content of Al in the framework can generate more Br\u00f8nsted acidity.Saenluang et al. [103] measured 27Al NMR for hierarchical nano spherical ZSM-5 nanosheets and resulted in two signals at 55\u00a0ppm attributed to the presence of Al species in the tetrahedral zeolite framework and 0\u00a0ppm ascribed from an extra framework of Al species [103]. The variation of synthesis precursors contributes to different intensities at 0\u00a0ppm. The use of pure silica Nanobeads precursor generated much higher intensity at 0\u00a0ppm compared to aluminosilicate nanobeads precursor. These results indicate that pure silica nanobeads precursor, which does not provide the aluminum source, favors the formation of extra-framework aluminum. In contrast, the presence of silicon and aluminum in the aluminosilicate nanobeads precursor assists the formation of internal framework aluminum.Jung et al. [104] investigated the acidity of hierarchical zeolites synthesized through desilication and structure-directed synthesis using 31P NMR spectra (Figure 11c). The former one generated hierarchical 3D MFI zeolite, and the latter produced an MFI nanosheet (denoted SD-MFI). Although both zeolites exhibited comparable textural properties and a total number of external acid sites, nanosheet zeolites surprisingly showed a higher concentration of strong acid sites than conventional MFI (denoted MFI-0.05). Thus, weaker BAS in the desilicated-MFI zeolite may be caused by the altered rearrangement of Al\u2013O\u2013Si bonds during the desilication process.The acid site properties of zeolite are also determined by substituting heteroatom in zeolite. In this case, Ji et al. [42] investigated the acidity properties of isomorphous MFI nanosheet zeolites (MFI (M), M\u00a0=\u00a0Al, Ga, and Fe). These isomorphous MFI nanosheets showed weaker acidity compared to conventional ZSM-5. The strength of acidity for isomorphous MFI nanosheet zeolites increased in the following order: MFI-Fe\u00a0<\u00a0MFI-Ga\u00a0<\u00a0MFI-Al. Furthermore, the result of pyridine-absorbed FTIR of isomorphous MFI nanosheet zeolites shows that the amount of Lewis acid sites was lower than that of Br\u00f8nsted acid sites. This indicates that most aluminum, gallium, and iron atoms existed in the framework of zeolite forming Br\u00f8nsted acid sites. The presence of these metals in the MFI nanosheet framework increases the ratio of LAS/BAS ratio, which is favorable for catalytic cracking of hydrocarbon .Compared to the 3D structure, the enhanced acid sites accessibility is more expected to the 2D one. For instance, in the pillared 2D zeolite, the expanding of the two adjacent of zeolites layer could enhance results in more exposed surfaces, leading to the accessibility improvement. Typically, it is closely related to the catalytic performance. Therefore, quantification of the type of acid sites is also crucial. In this case, there are range of organic molecules that has been utilized as a probe molecule, e.g., CO, pyridine, DTBP, TPP, and TMPO. The density of total acid sites was determined from the uptake of small organic base molecules. Meanwhile, the uptake of the larger one was used to calculate the external acid concentration. Finally, the accessibility was determined by calculating the ratio between adsorbed large molecules to the small ones [82].Wu et al. [101] used the pyridine and 2,4,6-collidine as probe molecule to evaluate the total and external acidity of MFI zeolite through IR spectroscopy. Result showed that nanolayered zeolite contains the higher amount of external Br\u00f8nsted acid site and Lewis acid sites than that of the bulk MFI zeolite. Meanwhile, the more defective nature on the unilamellar zeolites results in the higher of external Br\u00f8nsted sites, Lewis sites and silanol sites than multilamellar zeolites. Thus, unilamellar MFI showed a lower deactivation rate although there is no correlation between external BAS with the total turnover of methanol . Choi and Coworkers [102] also evaluated the accessibility of nanosheet zeolite compared with the 3D zeolite structure. ITQ-2 and the ultrathin zeolite samples (AS-8) show a higher accessibility factor than other samples, indicating the benefits of higher mesopore surface area. It corresponds to the two-dimensional structure which provides enough surface to the higher concentration of external acid sites.Nanosheet zeolite is widely used in various catalytic applications, including methanol conversion, aromatization, isomerization, oxidation/reduction. Other reactions are also described in this section. The summary of applications of nanosheet zeolites as heterogeneous catalysts are tabulated in Table 1\n.Methanol can be obtained by converting syngas (H2, CO) from natural gas, coal, or pyrolysis and gasification of biomass [105]. Methanol can be converted into a hydrocarbon to produce fuel, and the process is known as methanol to hydrocarbons (MTH), such as methanol-to-propylene (MTP) or methanol-to-olefin (MTO) [106]. Those processes usually involve catalysts to accelerate the conversion rate. Generally, the catalysts used in methanol conversion are ZSM-5 and SAPO zeolite. Many researchers focus on modifying zeolite to improve the catalytic process, such as acidity, stability, and selectivity. There are three main steps in the methanol conversion, (i) dehydration of methanol to form dimethyl ether (DME), (ii) equilibrium state consisting of a mixture of methanol, DME, and water, and (iii) conversion into olefins through dehydration [107].Hu et al. [87] reported that the nanosheet zeolite exhibited excellent catalytic activity compared with conventional zeolite. The use of zeolite MFI nanosheet catalyst in the MTP process resulted in nearly 100% conversion after a 240\u00a0h reaction and 51% selectivity toward propylene (Fig. 11a). The remarkably improved catalytic performance of MTP over nanosheet catalysts is ascribed to its unique morphology. Due to its large dimensions in the a-c plane, as previously mentioned, the nanosheet zeolite crystals end up with a higher surface fraction (010) than their conventional counterparts. Accordingly, the substrate is more feasible to reach the pore of the MFI nanosheet through the (010) surface corresponding to a straight channel. Besides, the ultrathin nanosheet provides shortened diffusion paths (especially the straight channels), hindering the formation of a larger molecule such as naphthene and aromatics that could block the zeolite channels. Consequently, the favored light olefin could be diffused out before the secondary reaction occurs, and the deactivation rate of the catalyst could be depressed. In addition, Hadi et al. [38] also reported the better performance of W-substituted MFI nanosheet zeolites. It exhibits propylene selectivity of 55.7%, with the total selectivity to light olefins reached 88.04%, and the catalyst is stable until 81\u00a0h (Figure 12b). It is higher than that of the W-substituted conventional zeolite. It corresponds to the unique morphology of the MFI nanosheet which provide a shorter diffusion path. In this case, once the product was formed, it could conveniently reach the outer, restraining the occurrence of the secondary reaction such as aromatization and hydrogen transfer.Xing et al. [69] reported the different catalytic performances of hydrothermally treated ZSM-5 nanosheet (HT-ZSM-5-SM/FM) prepared by different types of seeds, spherical (SM) and sheet-like (FM) ZSM-5. The hydrothermally treated ZSM-5 nanosheet prepared from spherical seed showed enhanced catalytic activity compared to that from sheet-like seed. According to TG analysis, the amount of coke loss in the range 200\u2013700\u00a0\u00b0C for hydrothermally treated ZSM-5 nanosheet prepared from spherical and sheet-like seed are 32.8% and 22.6%, respectively. These demonstrated that the use of spherical seed for the synthesis of ZSM-5 nanosheet catalyst favors higher carbon capacity.The selectivity and catalyst lifetime is remarkably enhanced after hydrothermal treatment [69]. According to Lu et al. [131] MOR nanosheets exhibit greater selectivity for light olefins than traditional MOR zeolite, as demonstrated in Figure 12 c. The ethylene selectivity decreased as the MOR thickness increased. These findings suggest that ethylene is formed via an aromatic-based mechanism on particular acid sites in the MOR framework.Furthermore, Kim et al. [108] studied the catalyst performance of MFI nanosheet zeolite with varied thicknesses (2.5\u00a0nm and 7.5\u00a0nm) and different Si/Al ratios (100\u2013700) in MTP reaction. Based on Fig. 12d, all samples possess a thickness of 2.5\u00a0nm except the H-ZSM-5 sample and NS-MFI-500 (7.5\u00a0nm). The nanosheet zeolite's thickness and Si/Al ratio can influence catalytic activity, while a higher propylene selectivity and longer lifetime were reached by nanosheet zeolite with thicknesses 2.5\u00a0nm (Fig. 12d). The nanosheet with thicknesses >7.5\u00a0nm increases the catalyst's diffusion resistance, making the reactant and product transfer more difficult. Therefore, the nanosheets with the thickness lower than 2.5\u00a0nm may lead to higher selectivity to propylene and a more robust tolerance to coke formation. In addition, the Si/Al ratio of the nanosheet zeolite is related to the amount of acidity for the MTP reaction. Besides, the number of cokes can also increase linearly with the number of acid sites. However, when the number of acid sites is too low, it is not adequate to convert higher molecules into light olefins [].Light olefin production by steam cracking is remained challenging due to several restrictions, including high reaction temperature (more than 800\u00a0\u00b0C), expensive construction materials, high energy consumption, the inflexibility of the product, and, in particular, low propene/ethene (P/E) ratio. Conversely, catalytic cracking can operate at much lower temperatures (often at 500\u2013600\u00a0\u00b0C) with consequent energy savings. Besides, the P/E ratio can easily be tuned by appropriate catalyst design and operating variables. Therefore, catalytic cracking makes a necessary contribution to satisfy the growing demand for light olefins today. Catalyst (generally zeolite) is a key factor to achieve high activity and selectivity to light olefins [109].The crucial aspects of product selectivity are the textural properties and acidity of the catalyst. The acidity of the catalyst is the primary determining factor of conversion. Fig. 13\n\na shows the conversion of n-decane by using nanosheets zeolite (ZN-2), HZSM-5 nanosheet zeolite with silica-pillared (PZN-2), HZSM-5 nanosheet zeolite with dual-template synthesis method (DZN-2), and conventional HZSM-5 zeolite (CZ-500) (Fig. 13a). The olefin selectivity of the dual-template HZSM-5 nanosheet zeolite is almost two times higher than that of the nanosheets zeolite and conventional HZSM-5 zeolite. As for silica-pillared HZSM-5 nanosheet zeolite, although the B acid is reduced and decreases the n-decane conversion, the selectivity towards light olefins is significantly higher than that of the non-pillared nanosheet zeolites and conventional HZSM-5 zeolite. This indicates that the pore structure of zeolite catalyst plays a crucial role in light olefin selectivity. The deactivation rate of conventional HZSM-5 zeolite reached 68.24%, which is\u00a0>\u00a011 times higher than HZSM-5 nanosheets zeolite (5.81%) (Fig. 13a). Besides, HZSM-5 nanosheets zeolite achieved the highest yields of light olefin (\u223c35%) with a slight reduction after 16\u00a0h of reaction (Fig. 13b) [26].\nFig. 13c displays the conversion of n-dodecane using nanosheet zeolites synthesized with a different number (n) of TPAOH, (10(C22-6-6)/n (TPAOH). Among the catalyst, nanosheet zeolite with n\u00a0=\u00a04 shows superior catalytic activity with a conversion rate of 76.8%, TOF value of 130.92 s\u22121, and deactivation rate of 9.11%. The inter-connected mesopores structure may inhibit the secondary reactions such as aromatization, leading to high olefin selectivity and low selectivity of aromatic compounds. The catalytic cracking of n-dodecane was tested in a flowing reactor at 4\u00a0MPa and 550\u00a0\u00b0C, as depicted in Fig. 13d [43].In the case of the aromatization process, the pore structure and strength of acid sites are the main factors responsible for high conversion and selectivity [97]. Kim et al. [39] studied the effect of Pt/MFI thickness from bulk to nanosheet scale (300\u20132\u00a0nm). The catalytic result is shown in Fig. 14\na. The catalytic conversion of n-C7 was first plotted as a reaction temperature function. It can be seen that the thickness of zeolite does not give a significant effect on the conversion of n-C7 since all Pt/MFI zeolites show a relatively similar S-curves. Meanwhile, the MFI zeolite crystal thickness becomes smaller, the conversion of i-C7 was higher than others. Moreover, B-300 was used as support for Pt NPs, and the catalyst achieved the maximum i-C7 mole percent 22\u00a0mol%, C-40 resulted in 29\u00a0mol%, NC-10 reached 42%, and NS-2 given 48%. However, reduction of thickness generated a higher i-C7 mole percent. Fig. 11b shows a correlation between the total conversion and the mol percent of i-C7 for a given catalyst, as measured over the range of reaction temperatures (200\u2013300\u00a0\u00b0C). As shown in Fig. 14b, the i-C7 mole percent increases by the order B-300\u00a0<\u00a0C-40\u00a0<\u00a0NC-10\u00a0<\u00a0NS-2 with the maximum value of 48%. It can be inferred that nanosheet zeolite is more favorable for i-C7 isomer production. The extremely thin morphology of the nanosheet facilitates the branched isomer product to migrate out before further reaction occurs, yielding a more highly branched isomer [].\nFig. 14c shows the n-pentane conversion of both conventional and nanosheet forms of Ga-ZSM-5. Ga-ZSM-5 nanosheets (GaExcZS-5-NS-X, with X, refers to the ratio of Si/Al) exhibit remarkably improved conversion of n-pentane compared to the conventional one (GaExcZS-5-CON-X). The BTEX selectivity for both Ga-ZSM-5 nanosheets with Si/Al ratios of 69 and 38 is over 40 and 43%, respectively. The catalytic performance on Ga embedded in ZSM-5 nanosheets is associated with the highly dispersed Ga species. With the increasing acidity in Ga-MFI (Fig. 14d), the catalytic activity was also increased because the high acid density can facilitate the catalytic breakdown of n-pentane in the first step of the reaction. It also activated oligomerization and further cyclization. Conversely, even at low Si/Al ratios, conventional structures still have a low conversion, low aromatic selectivity, and fast deactivation of the catalyst. This investigation confirms that an increase in the aromatic yield depends not only on the Si/Al ratio but also on the zeolite pore structure [110].Alkylation is one of the most important reactions in organic synthesis, especially in the interconversion of alkylbenzenes. Therefore, a great number of studies have been focused on the alkylation reaction [111]. Liu et al. [53] synthesized MOR nanosheets by varying the amount of water in the hydrothermal process. The sample was denoted H-MOR-x, where x refers to either the water amount (H2O/SiO2 ratio), symbol A (leaching), or C (commercial), as shown in Fig. 15a and b. Among the obtained zeolites, H-MOR nanosheet with a water amount of 26 (H-MOR-26) exhibits the highest conversion. Meanwhile, the order of the conversion for all catalyst are H-MOR-26\u00a0>\u00a0H-MOR-20\u00a0>\u00a0H-MOR-11\u00a0>\u00a0H-MOR-11-A\u00a0>\u00a0H-MOR-40\u00a0>\u00a0HMOR- C. This order is related to the number of strong Br\u00f8nsted acid sites. Each of the various catalysts has the following Br\u00f8nsted acidities 1.004, 0.893, 0.833, 0.623, 0.527, 0.140, respectively. Meanwhile, all MOR nanosheets show similar selectivity to methyl acetate (\u223c98%) and produce a small amount of ethanol and methanol. The DME conversion of MOR zeolite remains stable after 12\u00a0h except for H-MOR-11, which shows an extreme reduction from 39% to 9%. It is reasonable since H-MOR-11 possesses a small external surface area promoting coke deposition inside the micropore and decrease the catalyst stability.Feng et al. [112] studied the carbonylation of DME reaction to produce methyl acetate using commercially conventional H-ZSM-35 (CZ35), H-ZSM nanosheets (NZ35), and ZSM-35 hierarchical nanosheets prepared with various concentrations of NaOH 0.2\u00a0M (Hi-NZ350.2), NaOH 0.4\u00a0M (Hi-NZ350.4), and NaOH 0.6\u00a0M (Hi-NZ350.6). As shown in Fig. 15c, the catalyst ability in the carbonylation reaction increases with the order Hi-NZ350.6\u00a0<\u00a0CZ35\u00a0<\u00a0NZ35\u00a0<\u00a0Hi-NZ350.2\u00a0<\u00a0Hi-NZ350.4 with the conversion of 14.9, 22.4, 26.2, 31, and 42%, respectively. The low catalytic activity of the Hi-NZ350.6 catalyst may occur because of the damaged structure, lowest surface area, and pore volume, among other catalysts. The selectivity of all catalysts is relatively comparable, with a selectivity value over 90% to MA. The by products such as CO2 and CH4 are produced in small amounts caused by the WGSR (water-gas shift reaction) reaction. Furthermore, nanosheet zeolite shows high stability due to excellent diffusion resistance suppressing the deactivation of the catalyst [].Saenluang et al. [103] evaluated the effect of Si/Al ratio and Al distribution of hierarchical nanospherical ZSM-5 nanosheet on the catalytic performance in the alkylation of benzene. The catalytic activity of hierarchical ZSM-5 nanosheet with uniform Al distribution denoted as Hie-SZSM-5-AS (Low) for Si/Al\u00a0=\u00a059 and Hie-SZSM-5-AS (High) for Si/Al ratio\u00a0=\u00a0112, hierarchical ZSM-5 nanosheet with less uniform Al distribution denoted as Hie-SZSM- 5-PS, and conventional ZSM-5 denoted as Con-ZSM-5. The catalytic activity increases by the order Hie-SZSM-5-AS (Low)\u00a0>\u00a0Hie-SZSM-5-PS\u00a0>\u00a0Con-ZSM-5> Hie-SZSM- 5-AS (High). High Si/Al ratio (112) in hierarchical nanopsherical ZSM-5 nanosheet led to lower acid density, thereby decreasing the catalytic performance that was even lower than conventional ZSM-5. Noticeably, Al distribution also contributed to the altered catalytic performance. Hierarchical ZSM-5 nanosheet with less uniform Al distribution (Hie-SZSM- 5-PS) shows lower benzene conversion than the more uniform one (Hie-SZSM-5-AS (Low)) despite both catalysts possess similar acidity and textural properties. The different Al distribution means different acid distribution. As previously discussed, the different Al distribution could be attributed to different starting materials, where the use of aluminosilicate precursor for the synthesis of Hie-SZSM-5-AS (Low) sample favors the formation of tetrahedrally coordinated Al species generating Br\u00f8nsted acid sites. Meanwhile, the use of pure silica precursor promotes the formation of octahedrally coordinated Al species or extra-framework Al species. The low catalytic performance of conventional ZSM-5 compared to both SZSM-5-AS (Low) and Hie-SZSM- 5-PS demonstrates the contribution of mesopore in the hierarchical ZSM-5 nanosheet, which promotes more facile diffusion of the reactant to the active sites. In case of selectivity, all catalysts are more selective to ethylbenzene. Although the Hie-SZSM- 5-AS (High) shows the lowest benzene conversion, the selectivity to higher aromatics is the lowest among the catalysts , possibly due to the insufficient acidity to undergo further reaction.Friedel\u00a0\u2212\u00a0Crafts alkylation reaction of benzyl alcohol (BA) with benzene was reported by Zhou et al. [92] using MWW nanosheet catalysts. The MWW nanosheets prepared by using cetyltrimethylammonium (CTA) (Hd-MWWx, x refers to %wt CTA), and MWW nanosheet prepared by post-synthetic exfoliation method (ITQ-2). The catalytic ability of those zeolite decreases by the order H-d-MWW8.0\u00a0>\u00a0H-MCM-22\u00a0>\u00a0H-d-MWW5.5\u00a0>\u00a0H-ITQ-2. The excellent catalytic activity of H-d-MWW8.0 is related to the high external surface area (359\u00a0m2/g) and high site Al density. Meanwhile, the low catalytic activity of H-ITQ-2 can be caused by the damaged structure during the exfoliation process. This demonstrates that the post-synthetic exfoliation method adversely affects the catalytic performance of 2D layered MWW zeolite in Friedel\u00a0\u2212\u00a0Crafts alkylation reaction [92]. Liu et al. [132] reported a hierarchically structured MFI zeolite nanosheet for benzylation reactions. Zeolite nanosheets are labeled as MZA-n with the Si/Al molar ratio (n = 15, 31, and 50), while commercial zeolite ZSM-5 is labeled as CZSM-5. The relationship between the Bronsted acid/Lewis acid ratio (B/L) and benzyl alcohol conversion and alkylation/etherification selectivity is shown in Figure 15d. The Bronsted acid/Lewis acid ratio is a critical factor for evaluating benzylation activity and selectivity in the catalyst studied, according to the linier trend between Cc/2CE and B/L. Bronsted acid and Lewis acid sites\u00a0can promote benzyl alcohol conversion and alkylation product selectivity.The application of zeolites in oxidation or reduction reaction was usually coupled with metals. In this case, nanosheet zeolites provide a high surface area for metals to uniformly disperse onto their surfaces. Zou et al. [33] examined the performance of Cu-ZSM-5 nanosheets for N2O decomposition by comparing it with conventional Cu-ZSM-5. Cu-ZSM-5 nanosheet shows excellent catalytic activity and stability, as shown in Fig. 16\na. The N2O conversion of Cu-ZSM-5 nanosheets is quite stable around 80% after a 50\u00a0h reaction at 475\u00a0\u00b0C. In contrast, a sharp drop in N2O conversion from 74% to 54% is observed for conventional Cu-ZSM-5 after 50\u00a0h. The difference in the performance of the two types of catalysts is attributed to the differences in structure and properties where Cu-ZSM-5 nanosheets possess a higher surface area of 410\u00a0m2/g compared to conventional Cu-ZSM-5 (381\u00a0m2/g). Besides, from the O2-TPD data, the desorption of O2 from the Cu sites in the Cu-ZSM-5 nanosheet was more easily desorbed than the conventional one. High desorption of O2 is favorable for N2O decomposition. Thus, the Cu-ZSM-5 nanosheet showed improved catalytic activity.Structured Pt @ ZSM-5 nanosheets were prepared by Liu et al. [113] and applied in catalytic combustion of VOC. The catalytic process is carried out by loading Pt metal on three varied supports. The Pt ion was loaded on the nanosheet zeolite ca. 2\u00a0nm without calcination (Pt/PZN-2), with calcination (Pt/ZN-2), and conventional ZSM-5 ca. 500\u00a0nm (Pt/CZ-500) by impregnation method. Based on Fig. 16b, all types of catalysts reached 100% toluene conversion. However, the toluene conversion for Pt/ZN-2 catalyst and Pt/CZ-500 decreases after 180\u00a0h and 72\u00a0h, respectively. Pt particles are possibly aggregated under such conditions, particularly for Pt/ZN-2 and Pt/CZ-500 catalysts. Pt/ZN-2 was prone to collapse after 360\u00a0h of toluene combustion. Interestingly, Pt/PZN-2 remains stable after 360\u00a0h, showing the excellent hydrothermal stability of the sandwich-structures.Li et al. [114] compared the performance of 3 types of catalysts, including conventional TS-1 (CTS-1), mesoporous TS-1 (MTS-1), and hierarchical TS-1 nanosheets (HTS-1 50) with Si/Ti ratio of 50. They were applied in the catalytic epoxidation of cyclic olefins (cyclohexene and cyclooctene). As shown in Fig. 16c, the order of cyclohexene and cyclooctene conversions are HTS-1-50\u00a0>\u00a0MTS-1\u00a0>\u00a0CTS-1. This could be explained by the lower external surface area, which only generates a small portion of the active site. In contrast, HTS-1, which has a high external surface area, shows superior catalytic performance. All catalyst shows high selectivity toward epoxides. HTS-1-50 exhibited epoxide selectivity of 75.4% for cyclohexene. Meanwhile, the epoxide selectivity from cyclooctene is 98.3%, 96.9%, and 92.6% for CTS-1, HTS-1-50, and MTS-1, respectively. The high epoxide yield from cyclooctene could be caused by the higher electrophilicity of the carbon double bond in cyclooctene, which leads to a more stable cyclooctene epoxide. Furthermore, the HTS-1-50 catalyst showed higher epoxide yields of 17.6% and 17.4% for cyclohexene oxide and cyclooctene oxide, respectively. Meanwhile, the epoxide yield of the CTS-1 catalyst was only 3.5% and 2.5% for cyclohexene oxide and cyclooctene oxide, respectively. The more enhanced catalytic activity of hierarchical TS-1 nanosheet over the conventional TS-1 is attributed to the presence of larger porosity and a 2-dimensional form of hierarchical TS-1 nanosheet.Meng et al. [115] reported Fe/ZSM-5 nanosheet zeolite catalysts for benzene oxidation to produce phenols. The Fe/ZSM-5 nanosheets were synthesized using different SDAs, C22-6-3Br2 and C22-6-6-6-3 Br4. As shown in Fig. 16d, the samples are denoted as Fe/ZSM-5 (xN, y), where x is the number of quaternary ammonium ions (2 and 4) and y is the Si/Fe ratio (180, 360, 720). The suffix -st refers to steamed treatment. Steaming is an effective method to increase the amount of Fe2+that also acts as the active site. Thus, the steamed zeolite shows higher conversion and selectivity due to the high density of Fe2+ and FexOy aggregates in the steamed zeolite. However, steamed zeolite is more prone to deactivation than bulk zeolite. Steamed Fe/ZSM-5 nanosheet synthesized using C22-6-3Br2 (with a thickness of 3\u00a0nm) showed the highest catalytic activity by producing 185\u00a0mmol\u00a0g\u22121 phenol after 24\u00a0h on stream. Meanwhile, the bulk zeolites exhibited lower phenol selectivity and underwent faster deactivation with the increase in the amount of Fe. Likewise, the higher the Fe content for the nanosheet zeolite, the more coke production, and the faster it is deactivated. The initial catalytic activity of the nanosheet zeolite was about 50% higher than that of bulk zeolite. This occurrence is due to the diffusion resistance of the micropore structures in the bulk zeolite. Characterization of the texture of the spent catalyst demonstrated that the carbon coke in the deactivated bulk zeolite is assembled in the micropore. In contrast, the coke of the deactivated nanosheet zeolites was mainly collected in the mesopores.Besides the thermal catalytic reduction process, zeolites have also been applied in photocatalysis reactions, e.g., the reduction of CO2 into various products such as CO [118] and CH4 [119], the H2O splitting [117], the degradation of methylene blue [120] and methyl orange [121]. It has been known that combining metal oxide into porous materials such as zeolites could generate the isolated metal oxides acting as single site photocatalysts (Fig. 17\na) [116], in which the photocatalytic activity is not related to the transfer of electron and hole in the valence band and conduction band, respectively to the photocatalyst surface as observed in the semiconductor materials [119], but it correlates to the formation of the (M(n\u22121)+) species as a consequence of the charge transfer from ligand to metals (LMCT)). In that sense, this confined structure provides a localized environment with a quite short distance, thus, the charge transfer process in the photocatalyst materials could be enhanced, leading to extraordinary photocatalytic activity [116].Compared to conventional zeolites, nanosheet zeolites have a higher photocatalytic activity. For instance, Liu et al. [117] reported a higher H2 evolution rate (2152,7\u00a0\u03bcmol/h) of CdS/Pt nanoparticles supported on a porous lantern-like MFI zeolite composed of 2D nanosheets (NL-MFI) than those of supported on conventional MFI zeolites (1079.3\u00a0\u03bcmol/h) and unsupported Pt/CdS (515.0\u00a0\u03bcmol/h) (Fig. 17b and c). Moreover, the Pt/CdS-NL-MFI sample also exhibits higher stability (Fig. 17d). This was assigned to the promoting effect of nanosheet structure in absorbing visible light, separating the photogenerated electron-hole pairs, and enhancing the interaction between the water molecules and the photocatalyst.Moreover, other reports suggested that the remarkable photocatalytic activity of nanosheet zeolites was accounted for their ability to facilitate the high dispersion of metal nanoparticles [120] and the abundance of exposed Al atoms as alkaline sites, which allowed the high adsorption capacity of acidic molecules like CO2. Moreover, it also offers more abundance of surface active sites, allowing the excited state electrons to be quickly captured and transferred, thus resulting in a high photocatalytic activity [119]. Furthermore, for the reaction involving bulk molecules such as methylene blue, the existence of a mesopore in the nanosheet structure accommodates the molecule to conveniently adsorbed and react on the catalyst surface [121]. Moreover, it shortens the diffusion path length of the molecules, thus improving mass transfer.Although several positive results of the nanosheet zeolite were associated with their high external surface properties, a different result was reported by Ji\u0159i \u010cejka and his group [118]. They demonstrated that 3D TS-1 zeolites show a higher product formation rate (3.29\u00a0\u00d7\u00a010\u22124) than 2D TS-1 zeolites (3.04\u00a0\u00d7\u00a010\u22124), with a selectivity of 30.8% and 30.7%, respectively. Notably, the external surface area for these two samples was 6\u00a0m2/g and 94\u00a0m2/g. It suggests that the photocatalytic reduction of CO2 was not merely driven by the specific or the external photocatalyst surface, which produces a stronger and higher interaction between the molecules and the catalyst. Yet, hydrophilicity is also strongly affected in the efficiency and the selectivity of the catalyst. This result was confirmed with the H2O TPD results, which demonstrated the best correlation between the amount of the adsorbed H2O with the hydrogen production.Lee et al. [40] reported the catalytic performance of unilamellar mesoporous MFI nanosheets (UMNs) and Al-SBA-15 for pyrolysis of lignocellulosic biomass (cellulose, xylan, and lignin). For catalytic pyrolysis of cellulose, both UMNs and Al-SBA-15 converted the cellulose into various products, with the highest produced products are oxygenates (ketone, aldehyde, alcohol, cyclo-compounds, furans, and levoglucosan). The strong Br\u00f8nsted acid in UMNs catalyst promoted the conversion of levoglucosan into furan and aromatic compounds. The different catalysts resulted in different cyclo compounds. The UMNs strong acid sites are considered to facilitate the production of cyclo compounds with lower oxygen content, 2-cyclopentene-1-one. Meanwhile, Al-SBA-15 mainly produced 2-cyclopentene-1,4-diones.Furthermore, the xylan catalytic process also produced high oxygenated products (esters, ketones, aldehydes, alcohols, cyclo compounds, and furans). The increase in mono-aromatic yield occurs with the addition of UMNs. The degree of crystallinity of hemicellulose (xylan) is lower than cellulose. Thereby, xylan is easier to break down and converted into aromatics. Furthermore, the catalytic pyrolysis of lignin yields a highly phenolic product such as 2-methoxy-phenol. The yield of mono-aromatic was much higher when using UMNs than Al-SBA-15. This occurrence is feasible because Al-SBA-15 has a weaker acidity that can convert heavy compounds into light phenolics but cannot further convert light phenolics into aromatics.UMNs, which have strong acidity, showed better catalytic performance than Al-SBA-15. A tremendous amount of oxygenating has been removed, and the aromatic yield is three times higher, leading to more outstanding quality bio-oil production [40].Liu et al. [41] prepared Pd loaded on HZSM-5 nanosheets (Pd/ZN) for direct synthesis of hydrogen peroxide. The study was conducted by comparing the Pd loaded on nanosheet zeolite (Pd/ZN-50) and conventional zeolite (Pd/CZ-50) with a Si/Al ratio of 50. The catalytic performance of Pd/ZN-50 exhibited 82.3% conversion, whereas Pd/CZ-50 reached 25.7% conversion at 0.5\u00a0h. Meanwhile, the H2O2 selectivity of the catalyst Pd/ZN-50 (35.1%) is much lower than Pd/CZ-50 (62.6%). Based on the conversion and selectivity values, the productivity of H2O2 from the Pd/ZN-50 (28.9\u00a0mmol\u00a0H2O2 (g cat)\u22121. h\u22121) was greater than that Pd/CZ-50 (20\u00a0mmol\u00a0H2O2 (g cat)\u22121. h\u22121) after 0.5\u00a0h. The catalyst performance depends on the concentration of H+, crystal phase, Pd dispersion, and Pd0 content [122]. Pd/ZN-50 shows lower selectivity due to a higher decomposition rate compared to Pd/CZ-50. This result could also be due to the Pd+-Z- structure formed by the stronger metal-support interaction between the Pd particles and the acid sites. The higher selectivity of Pd/CZ-50 toward H2O2 is due to the lower surface charge density of the Pd atom inhibiting the dissociation of the O\u2013O bonds in H2O2. Moreover, catalytic performance (including H2 conversion, H2O2 selectivity, and H2O2 formation rate) is enhanced with a lower Si/Al ratio. In addition, the smaller Pd nanoparticle size in the Pd/ZN-50 catalyst offers a strong ability to dissociate the O\u2013O bonds in H2O2.Feng et al. [123] evaluated Ni/ZSM-5 nanosheet catalysts for hydroconversion of oleic acid to aviation-fuel-range-alkanes (AFRAs) by alternate the Si/Al ratio (100, 200, or 300). In addition, the external surface of nanosheet zeolite ZSM-5 was modified through CLD (Chemical liquid deposition) method. The catalytic test showed that all catalysts reached 100% oleic acid conversion. The Ni/ZSM-5 nanosheet catalyst showed high deoxygenation activity due to improved access to the active site. In addition, the decrease in the concentration of strong Br\u00f8nsted acid for CLD-modified catalysts led to higher OLP (organic liquid products) production than the pristine ones. Low external Br\u00f8nsted acid concentration promotes cracking reaction of the primary long-chain deoxygenated product into linear AFRA. Conversely, the high external Br\u00f8nsted acid concentration supports secondary cracking of the deoxygenated products generating short-chain alkanes. Besides, low or moderate internal Br\u00f8nsted acid concentration increases the isomerization of AFRA and decreases the secondary cracking.Kim et al. [80] compared the performance of three types of catalysts, specifically MFI nanosheet, amorphous silica, and bulk silica zeolite, in the gas-phase Beckmann rearrangement of cyclohexanone oxime. MFI nanosheets showed high catalytic activity with cyclohexanone oxime (CHO) conversion of 77%, a selectivity to lactam of 92%, and a long catalytic period up to 100\u00a0h with no observable change in selectivity. On the other hand, the bulk silica zeolite catalyst showed an initial conversion of 48% and underwent more rapid deactivation. The conversion considerably dropped to 15% after 20\u00a0h reaction. The longer catalytic lifetime of the MFI nanosheet can be explained by the significant difference in the coke content between deactivated nanosheet and bulk MFI (4 and 7\u00a0wt%, respectively). It suggested that the MFI nanosheet defeated the polymer species that produce coke and lead to the deactivation of the catalyst. The high selectivity of the nanosheet zeolite toward lactam is attributed to the silanol group's surface located in the crystal plane (010).Chang et al. [61] also synthesized silicalite-1 nanosheet for vapor-phase Beckmann rearrangement of cyclohexanone oxime (CHO) to produce lactams. The SDA was prepared by using poly (propylene glycol) bis(2-aminopropyl ether) (NH2(PO)nNH2) with an average molecular weight of \u223c400 (\u00f16) and \u223c2000 (\u00f133). Silicalite-1 samples synthesized with N3-POn-N3 are denoted S1-n-A. The catalytic result showed that the S1-6-A sample could convert 76% of CHO at the beginning of the reaction. It decreased to 34% after 50\u00a0h reaction time with e-caprolactam (CL) selectivity reaching 96%. Moreover, the S1-33-A catalyst arose conversion of CHO to 92% at the beginning of the reaction and decreased to 51% after 50\u00a0h, with CL selectivity reaching 97%. High CHO conversion of both catalysts can be assigned to large inter-lamellar and inter-bundle porosity that facilitate mass transport of substrate. Noticeably, a longer POn linker contributes to the improved pore volume. In addition, the S1-33-A sample has a large pore volume (0.76 cm3g\u2212 1) compared to S1-6-A (0.34\u00a0cm3\u00a0g\u22121), so that S1-33-A shows higher CHO conversion. The high CL selectivity for S1-6-A and S1-33-A can be attributed to the binding sites in the hierarchical silicalite-1 nanosheet, such as the silanol groups on the external surface.Ali et al. [59] prepared a self-pillared MFI nanosheet catalyst for acrolein production through the glycerol dehydration process. Self-pillared MFI nanosheet achieved 92% selectivity toward acrolein at WHSV 4\u00a0h \u22121. The stability analysis shows fresh and regenerated pillared MFI nanosheets resulting in comparable catalytic activity, which means the catalyst exhibited good thermal stability. The zeolite possesses a Br\u00f8nsted acid, which is the active site utilized in glycerol dehydration. Since the Br\u00f8nsted acid conducts the catalytic reaction in glycerol, the acidity of the catalyst is crucial that can impact the activity, selectivity, and stability of the catalyst. Besides, the catalyst structure is also an important factor that could affect diffusion and coke capacity. Pillared MFI nanosheet showed higher TON than nanocrystalline ZSM-5 due to high catalyst stability, although they have the same crystal size. The poor selectivity of nano-ZM-5 is correlated with restricted reactant diffusion. Therefore, a self-pillared MFI nanosheet catalyst can inhibit coke formation and increase the selectivity of acrolein by reducing diffusion restrictions.To date, research on the nanosheet zeolite is under the spotlight due to the remarkable advantages offered by its two-dimensional structure. From the catalyst design point of view, several synthesis strategies have been extensively studied. Generally, the preparation of nanosheet zeolite could be conducted through bottom-up and top-down methods. The former is usually performed by hydrothermal synthesis process, in which the structure-directing agent is usually involved. Typically, it produces the so-called layered precursor, which could condense upon the calcination, e.g., MCM-22. In this case, several post-synthesis modifications could be performed to prevent the structure collapse, including pillarization, delamination, swelling, and exfoliation. Furthermore, another strategy like the use of surfactant as multifunctional template has also been reported to directly synthesize the nanosheet structure. In this case, the variation in the synthesis parameter, such as type of the surfactant as SDA, SDA amount, Si/Al ratio, crystallization temperature, and time could affect the morphology, thickness, acidity, and textural properties of nanosheet zeolite. In the case of SDA, almost every part of the structure, such as to N\u2013N spacer length, alkyl groups of SDA, hydrophobic tail length, and the number of ammonium groups, also could influence the characteristic of nanosheet zeolite. On the other hand, the ADOR method is considered as the bottom-up strategy, in which the 2D structure was generated from the existing 3D structure. In addition, chemical etching, the top-down method, has also been reported using NH4F. The acidity and textural properties of the obtained nanosheet zeolite could be tailored by the fluoride etching.In the structural and textural analysis of nanosheet zeolite, a similar characterization technique with conventional zeolite is usually performed, i.e., XRD, FTIR, NMR, SEM, TEM, and N2 adsorption-desorption. Typically, the XRD pattern of nanosheet zeolite is broader, in which the sufficient sharp enough peak dominantly appeared for h0l indices. Moreover, the peak observation at a low angle position could also be utilized to evaluate the interlayer space of nanosheet zeolite. For the direct observation of the nanosheet morphology formation, the SEM and TEM analyses are the mandatory techniques that should be performed. In general, nanosheet zeolite exhibits a more open structure with the ultra-thin of crystal thickness, in which the composition of the stacking layers could be determined by observing the d-spacing between the layers observed in the TEM images. Since the space of the interlayers is in the mesopore scale, the nanosheet zeolite usually shows a type IV isotherm, indicating the presence of the mesopore with a large external surface. It benefits the catalysis process, which needs an accessible acidic site to facilitate the reaction. Furthermore, when the product is formed, it will easily diffuse to the outer, restraining the secondary reaction. For the metal-substituted catalyst, nanosheet zeolite provides a remarkable surface for metal to be well-distributed. Thus, the appropriate amount, strength, and distribution of acid sites could be obtained, leading to improved catalytic performance. In the case of the acid properties evaluation, the FTIR, NH3-TPD, as well as the MAS-NMR characterization was usually employed. In addition, the organic base molecule has also been reported as a probe molecule to determine the accessibility of the acid site.The catalytic activity of nanosheet zeolites in several reactions such as methanol conversion, alkylation, isomerization, cracking, oxidation/reduction exhibited excellent performance compared to conventional zeolite. Nanosheet zeolites show high conversion and selectivity as well as high stability. Due to its large dimensions in the a-c plane, as previously mentioned, the nanosheet zeolite crystals end up with a higher surface fraction (010) than their conventional counterparts. Accordingly, the substrate is more feasible to reach the pore of the nanosheet zeolite. The ultrathin nanosheet also provides shortened diffusion paths (especially the straight channels) in nanosheet zeolite catalysts. The crucial aspects for product selectivity are texture properties and acidity of the catalyst. Furthermore, nanosheet zeolite is more resistant to deactivation because the formed coke is stored in the external surface so that the active site is not easily covered by coke.\nTable 1 demonstrates several zeolites that have been reported to be applied in several reactions. As can be seen, MFI is the most common topology that has been applied, in which the surfactant assisted method is the most used synthesis strategy. Notably, some modification with metals is also applied to adjust the catalyst acidity and/or facilitate the redox reaction. Although the presence of the nanosheet morphology has improved the catalytic activity, other properties shall be carefully adjusted. For instance, the Si/Al ratio for the common methanol to propylene reaction (MTP) is usually higher than 200, meanwhile, it is lower for the aromatization and isomerization reaction (<100). Besides, other zeolite types such as FER, MOR, and FAU nanosheet is still limited, suggesting the availability of more opened areas to design and applied those type of zeolites. In the catalysis point of view, the application of nanosheet zeolite to support the green and sustainable process begin to be emphasized, i.e., the water splitting reaction, biomass conversion, and the CO2 reduction. The research focused on the broader area should be engineered owing to the potential capability of nanosheet zeolites.Despite nanosheet zeolites have advantages over conventional zeolite, the nanosheet zeolites still face several synthesis challenges, which are time-consuming, costly materials, and requires several steps; thus, limiting their practical applications. The diquaternary ammonium-based SDAs are not commercially available, and should be synthesized through complex, multistep organic reactions. In this sense, the seed-induced method might be the more preferred route due to its less complexity and relatively lower cost. Extensive studies should be pursued to engineer the crystal growth parameters to obtain the nanosheet zeolites with desired properties (thickness, lateral dimension, chemical composition, etc.). The use of small molecules (e.g., urea) could induce the crystal growth-inhibiting effect in particular directions; thus, promoting the nanosheet morphology. Moreover, the synthesis of nanosheet zeolites using low-cost, renewable precursors, e.g., silica extracted from agricultural waste, or natural sources, could be promising, as have been implemented in the synthesis of hierarchical zeolites [28].Finally, from the significant progress of nanosheet zeolite, several insights could be constructed. The design of a more affordable bifunctional template and the utilization of the renewable source for nanosheet zeolite synthesis should be considered. Moreover, the in-situ characterization coupled with the computational studies could be performed to precisely determine a significant factor dictating both the nanosheet formation and the catalysis reaction. Further development using machine learning techniques also allows the rationalization of physicochemical, and structural insights into the chemistry of zeolite synthesis and catalysis, leading to the understanding of empirical knowledge, classification of synthesis records, discovery of novel materials and efficient reaction route. Moreover, It could also speed up understanding the synthesis\u2212structure relationship. From a technological point of view, applying nanosheet zeolite for catalytic membrane reactors would bring a remarkable breakthrough to the catalysis industry since its unique structure demonstrates an outstanding performance. To this end, the industrial upscale of the material synthesis, as well as its application, could drive a breakthrough in energy and fine chemical industries.Grandprix T. M. Kadja: Conceptualization, Writing \u2013 original draft, Writing \u2013 review and editing, Supervision, Funding acquisition. Azhari, Noerma J. Azhari: Writing \u2013 original draft, Writing \u2013 review and editing, Validation, Visualization. St. Mardiana: Writing \u2013 original draft, Writing \u2013 review and editing, Validation, Visualization. Neng T. U. Culsum: Writing \u2013 review and editing, Validation, Formal analysis. Ainul Maghfirah: Writing \u2013 review and editing, Validation, Formal analysis.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work is supported by Hibah PDUPT 2022 (first year) (Contract no. 083/E5/PG.02.00.PT/2022) from the Ministry of Education, Culture, Research and Technology of the Republic of Indonesia.", "descript": "\n                  Recently, zeolites with two-dimensional structures, so-called nanosheet zeolites, have been intensively advanced owing to their excellent catalytic performance compared to conventional zeolite. This extremely thin nanosheet structure of zeolite provides more accessible active sites contributing to shortened diffusion pathways and primarily enables bulky molecules to undergo catalytic reactions. Several nanosheet zeolites, such as MFI, FAU, MOR, MWW, TS-1, have been successfully synthesized. Generally, they are obtained by direct synthesis method involving surfactant as structure-directing agent (SDA) and or seed. Interestingly, almost every part of the synthesis parameter conditions can alter the characteristics of this nanosheet zeolite. Thus, 2D zeolites also offer easily tunable characteristics. Up to now, nanosheet zeolites have been extensively studied as a catalyst in methanol conversion, catalytic cracking, isomerization, alkylation, oxidation/reduction, lignocellulose conversion, and other reactions. Furthermore, the modification of nanosheet zeolite could be pursued to enhance its catalytic performance. Pillaring is evidently effective in improving the textural properties of nanosheet zeolite by protecting the mesopore structure during the SDA removal. Herein, we comprehensively present the recent progress on the development of nanosheet zeolites. This review concludes with a summary of a discussion of remaining challenges and outlook for nanosheet catalyst technologies.\n               "}
{"full_text": "Recently, due to declining oil resources, research on alternative transport fossil fuels has grown rapidly in both industry and academia [1,2], and fossil fuel customers have turned to biofuels [3]. For such reason, researchers have paid special attention to hydrogen as a promising alternative to fossil fuels to overcome the energy crisis and environmental pollution, because the product of hydrogen and oxygen combustion is steam [4,5]. One of the strengths of this valuable energy carrier is its abundance [6], renewability [7], and its non-pollution and high energy density [8]. Electrochemical reduction of molecular hydrogen from water is a general process in producing pure and very clean hydrogen, but it requires considerable electrical energy to break the hydrogen and oxygen bond, which is due to the high voltage of hydrogen. Therefore, minimizing the cathode voltage and its economic aspect is very important [9]. In the last few decades, many efforts have been made to develop transition metals such as cobalt and nickel as inexpensive high-efficiency catalysts [10]. Nickel selenide possess unique electron states and exhibit excellent performance for HER [11]. Cobalt selenide is one of the catalysts that has been considered recently and due to the position of its electron layers and the inherent nature of the charge transfer material from the electrode surface to the bulk is very fast, which can create superior electrocatalytic properties [12]. For electrocatalysts, high specific surface area in nanostructured materials can increase hydrogen release activity [13,14]. Nickel foam is an attractive raw material for electrocatalytic application due to its low cost, high electrical conductivity, high porosity, and structural integrity [15], which used as substrate for catalytic applications. In previous researches to synthesis nickel\u2013cobalt selenide CoCl2\u00b76H2O, Co(NO3)2, Co(OH)2, Co(CH3COO)2 Co3 was used as precursor for cobalt, NiCl2\u00b76H2O, Ni(NO3)2, Ni(OH)2 and Ni(CH3COO)2 as precursor for nickel and Se powder, Na2SeO3, Se(CH3)2 and SeO2 as precursor for selenium, which synthesized on various substrate; for example Bhat et\u00a0al. [16] synthesized nickel selenide nanosheets on nickel foam using nickel hydroxide and Se powder by hydrothermal method, Kwak et\u00a0al. [17] synthesized nickel\u2013cobalt selenide nanocrystals using Co (NO3)2, Se (CH3)2, Ni (NO3)2, Ge (CH3)4 and Nafion precursors using a laser process and layered on the silicon wire substrate, Liu et\u00a0al. [18] using Ni(CH3COO)2, Co(CH3COO)2 and SeO2 precursors, nickel\u2013cobalt selenide nanoparticles were synthesized by electrochemical deposition on titanium sheet, Ming et\u00a0al. [19] synthesized (Ni, Co)Se2/C-HRD nanoparticles using CO (NO3)2, Ni (NO3)2 and Se powder precursors on ZIF-67 by heating in a quartz tube space, Rezaei et\u00a0al. [20] reported a two-step synthesis of nickel\u2013cobalt selenide using CO(NO3)2, Ni (NO3)2, Na2SeO3 and N2H4 precursors on reduced graphene by hydrothermal process; but, in previous research, there has been no report on the synthesis of nickel and cobalt selenide using CoCl2\u00b76H2O, NiCl2\u00b76H2O, Na2SeO3 and N2H4 precursors on nickel foam using one-step hydrothermal synthesis; in addition to contradictory compounds have been reported for the most suitable nickel\u2013cobalt\u2013selenide composition. Bin et\u00a0al. [21] and Liu et\u00a0al. [18] reported the combination of Co0.11Ni0.89Se2 and Co0.13Ni0.87Se2, respectively, while Zheng Xin et\u00a0al. [22] reported the combination of Co0.8Ni0.2Se as the one with the most activity in the hydrogen evaluation; Therefore, this research is based on the synthesis of nickel\u2013cobalt selenide using CoCl2\u00b76H2O, NiCl2\u00b76H2O, Na2SeO3, and N2H4 precursors on nickel foam using one-step hydrothermal method. The purpose of this study was to investigate the structures of nickel selenide, cobalt selenide, and nickel\u2013cobalt selenide as a cathode material in the process of water molecule breakdown, or in other words, hydrogen reduction in the water electrolysis process, and investigated the effect of different ratios of nickel and cobalt and the effect of hydrothermal temperature on electrocatalytic efficiency of these structures.Distilled water was used in all experiments in this study. The chemicals used included nickel chloride hexahydrate (NiCl2\u00b76H2O), cobalt chloride hexahydrate (CoCl2\u00b76H2O), sodium selenide (Na2SeO3), hydrazine (N2H4), hydrochloric acid (HCl), and nickel foam, which were purchased from Merck.Commercial nickel foam with 3\u00a0mm thickness cut into 10\u00a0mm\u00a0\u00d7\u00a030\u00a0mm pieces, then to remove the surface oxides and impurities and activate the nickel foam surface, in a solution containing 3\u00a0M hydrochloric acid became ultrasonic in an ultrasonic bath for 15\u00a0min and washed several times with acetone and distilled water, then dried at room temperature.Two series of hydrothermal experiments were performed at 150 and 180\u00a0\u00b0C. Sodium selenide and hydrazine were constant in all experiments and 10\u00a0ml (100\u00a0mM) and 2\u00a0ml were used, respectively. Another variable parameter is the ratio of nickel chloride to cobalt chloride. The solution was prepared by first pouring 10\u00a0ml of nickel chloride solution (100\u00a0mM) into the beaker for the first sample (NiSe-150) and then 10\u00a0ml of sodium selenite (100\u00a0mM) was added while stirring to prevent the solution from clotting during the addition of the solutions, the beaker was stirred, then 2\u00a0ml of hydrazine (as reducing agent) was added to the solution during stirring, and by adding 78\u00a0ml of distilled water, the total volume of the solution was increased to 100\u00a0ml. The solution was stirred by a magnetic stirrer for 15\u00a0min to obtain a uniform solution without light pink clots. The solution with a piece of pre-activated nickel foam was poured into a Teflon-lined autoclave and thoroughly sealed. The container containing the solution was placed in the oven and it was set at 5\u00a0h and a temperature of 150\u00a0\u00b0C. Subsequent experiments were performed by different Ni+2:Co+2\u00a0mol ratios of 8:2 (Ni8Co2Se-150), 4:6 (Ni4Co6Se-150), 6:4 (Ni6Co4Se-150), 2:8 (Ni2Co8Se-150), and finally cobalt chloride 0:10 (CoSe-150) instead of nickel chloride with the conditions of the first experiment, in which a total of 6 samples were made. Then 6 samples were synthesized again with the previous ratios at 180\u00a0\u00b0C (NiSe-180, Ni8Co2Se-180, Ni6Co4Se-180, Ni4Co6Se-180, Ni2Co8Se-180, and CoSe-180). After reaction in the autoclave and cooling to ambient temperature, the chamber was ejected from the oven and deposited precipitation with precipitation growing on the nickel foam substrate was washed several times with distilled water and dried in a vacuum at 60\u00a0\u00b0C for 5\u00a0h. The precipitated powder was separated from the solution in an autoclave process by centrifugation and perform XRD analysis with a device model Rigaku ivultim to determine the structural phases, FESEM with a device model Mira3 TESCAN-xmu and TEM with a device model Philips CM30 to evaluate the morphology and FT-IR with a device thermos-avatar model to determine the functional groups. To evaluate the amount of hydrogen release activity, the produced samples (powder precipitated on nickel foam) were electrochemically examined. For this purpose, a Vertex model Ivium potentiostat device was used in a three-electrode system with a platinum electrode as a counter, a silver electrode (Ag|AgCl/saturated with KCl) as a reference electrode, and deposited thin films for a working electrode to which the test specimen is attached and 1\u00a0M KOH solution at 25\u00a0\u00b0C was used for the electrolyte of all electrochemical tests.The results of the electrochemical measurements were recorded in the form of Nyquist, Phase, and Bode plots and the Zview software was used to determine equivalent circuits for this test. All of the LSV diagrams shown in this activity were iR modified. All the potentials measured experimentally against Ag|AgCl were shown using the Nernst equation, the same as Equation (1), converted to a reversible hydrogen electrode (RHE).\n\n(1)\n\n\nERHE\n=\nEAg\n|\nAgCl\n+\n0.059\npH\n+\nE\u00b0Ag\n|\nAgCl\n\n\n\n\nHere, the standard potential of Ag|AgCl at 25.1\u00a0\u00b0C is 0.197\u00a0V versus RHE.The electrochemical kinetics of the (NiCo) Se HER catalysts related to the overpotential (\n\n\u03b7\n\n) with current density (j) have been performed to calculate the Tafel slop using Equation (2).\n\n(2)\n\n\n\u03b7\n\u00a0\n=\n\u00a0a\u00a0\n+\n\n\n\n2.3\nR\nT\n\n\n\u03b1\nn\nF\n\n\n\nlog\n\u00a0\n\n(\nj\n)\n\n\n\n\nwhere \n\n\u03b7\n\n is the overpotential, j is the current density and other\u00a0symbols have their usual meaning. The Tafel slope is\u00a0\n\n2.3\nR\nT\n/\n\u03b1\nn\nF\n\n. Tafel slopes were calculated from the polarization test with a scanning speed of (10\u00a0mV\u00a0s\u22121) in 1\u00a0M KOH solution.X-ray diffraction analysis of synthesized samples at 150\u00a0\u00b0C from hydrothermal process, can be seen in Fig.\u00a01\n(a). According to the diffraction pattern of NiSe and CoSe samples, the peaks show high intensity, which indicates the crystalline structure of the electrodes, and the formation of the crystallinity structure leads to stability in long-term performances and it is a determinant factor in hydrogen evolution activity. In diffraction pattern of NiSe-150 and CoSe-150, the peaks generated in the 2\u03b8 specific correspond to the pages specified in Fig.\u00a01(a). Due to the similar atomic diameters of nickel and cobalt, there will be no change in the plans where cobalt replaces nickel [18], finally, there will be no significant change in the angle of the peaks; e.g. peak in NiSe-150 which 2\u03b8\u00a0=\u00a033.50, in CoSe-150, the same peak is in 2\u03b8\u00a0=\u00a033.4 which clearly shows the overlap of the peaks. The crystallinity size calculated by Scherrer method; crystalline size was 27, 33, 31, 45, 24 and 39 for NiSe-150, Ni8Co2Se-150, Ni6Co4Se-150, Ni4Co6Se-150, Ni2Co8Se-150 and CoSe-150 respectively. Fig.\u00a01(b) also shows the FTIR spectroscopy pattern of synthesized samples at 150\u00a0\u00b0C and confirms X-ray diffraction, and shows that at low wavelengths metal bonds are formed [23]. EDS analysis was used for Ni8Co2Se-150 to shows the amount of each existing element in the sample, the results of which can be seen in Fig.\u00a01(c) and shows that all three elements nickel, cobalt, and selenium are presented in the structure. To investigate the surface morphology of the Ni8Co2Se sample, which was introduced as the optimum sample, TEM analysis was used and the results presented in Fig.\u00a01(d) which shows that the powders are in the form of spheres on the surface of that structures. From the TEM pattern, it can be seen that the material is polycrystalline. This pattern indicates that the material is amorphous if the light points be a complete circle, but if it is a series of dotted points around a circle perimeter, it indicates that the material is polycrystalline.FESEM results of the synthesized samples at 150\u00a0\u00b0C are shown in Fig.\u00a01(e\u2013j). The resulting NiSe-150(free cobalt) image Fig.\u00a01(e) shows that it has grown as a nanoparticle, the synthesis of nickel selenide by the hydrothermal process is usually arranged as a nano-grid [24]. These images show that the nanoparticles are in the form of plates and placed next to each other and have formed a nano-plates structure. Fig.\u00a01(f) shows a sample of Ni8Co2Se-150 growing in the form of regular and continuous nano-plates due to the addition of a small amount of cobalt in the nickel selenide structure, which is called hydrangea structure. Porous spaces have grown uniformly in all directions [25]. Such nanostructured particles are very useful for electrocatalytic activity [26,27]. One of the factors that improves the electrocatalytic activity of electrodes is to increase the specific surface area of the electrode, because the empty space between the nano-plates is a good location to place ions in the electrolyte and increase the reaction surface in the hydrogen release. In sources, this principle is known as the roughness factor, so that the higher specific surface area of an electrode (roughness factor), active points on the surface also increases and exhibits better kinetics.Then, by increasing the ratio of cobalt to nickel, the morphology is taken out of the regular lattice state and also reduces the specific surface area of the catalyst. In Fig.\u00a01(g), which is related to the Ni6Co4Se-150 sample, the structure is shown as irregular nano-plates with double growth of some plates, which is discontinuous and nanoparticles have grown along with it. This reduces the catalytic properties of the electrode, which can be proved by electrochemical tests, which will be described below. In the Ni4Co6Se samples and Ni2Co8Se, respectively, shown in Fig.\u00a01(h and i), it will grow in the same way and will decrease the specific surface area with the roughness factor. In the CoSe-150 sample Fig.\u00a01(j), regular structure growth appears again, which is visible as a sphere with a latticed surface. The nano-plate thicknesses were measured between 20 and 30\u00a0nm. In the following, in Fig.\u00a01o, the sample of map analysis is shown, which in Fig.\u00a01(k and n) shows the distribution of elements in the structure. This proper dispersion results in the uniform electrocatalytic activity of the sample at different stages of the electrochemical tests.The most important quantity studied in the Tafel test is the slope of the Tafel cathodic plot (bc), because the Tafel slope is directly related to the kinetics of the catalysts reaction [28], which is the angle coefficient of the logarithmic variation of the observed current density according to the applied overpotential. If other conditions are constant, the smaller Tafel slope of the test electrode exhibits higher electrocatalytic activity of the electrode in the hydrogen evolution reaction. Steady-state polarization diagrams (Tafel diagrams) obtained from the surface of the electrodes in 1\u00a0M KOH solution at 25\u00a0\u00b0C are shown in Fig.\u00a02\n(a and b). In these two diagrams number of Tafel slope is be able seen on each plot. That the Ni8Co2Se-150 sample with 61.3\u00a0mV shows the lowest slope among all samples. Increasing the cobalt/nickel ratio in the structure, increases the Tafel slope, for this reason, samples Ni6Co4Se-150, Ni4Co6Se-150, Ni2Co8Se-150 has higher slopes at both synthesis temperatures of 150 and 180\u00a0\u00b0C respectively. The higher slope indicates lower electrocatalytic activity and lower reaction kinetics. As mentioned before, one of the factors that affect the density of the exchange current is the roughness factor, and this parameter does not affect the reaction mechanism, but its effect is observed on the current density and the Tafel curve is performed without changing the slope or reaction mechanism leads to a higher current density, and as the electrode surface area increases, this principle becomes more pronounced: that attention to Tafel slope the samples is quite obvious. The Tafel slope also shows that the samples containing cobalt have the lowest potential, which confirms the increase in nickel activity by adding cobalt to a certain amount. According to the Tafel test shown in Fig.\u00a0(2a and b), the activity of hydrogen evolution increases by 20% with the addition of cobalt, and with increasing the cobalt to nickel ratio over than 0.2, we see an increase in the cathodic slope. This can be attributed to the reduction in hydrogen evolution activity, which can be attributed to the reduction of surface active sites due to the reduction of the continuous structure arrangement of the nano-plates [21].\nFig.\u00a02(c and d) show the graph of the LSV and the activation potential for the synthesized samples which syntheses at 150 and 180\u00a0\u00b0C respectively, at the applied potential of 0 to \u22120.5\u00a0V with a sweep rate of 10\u00a0mV\u00a0s\u22121, and Table 1\n also presents the activation potential numerically. Since the evolution activity may vary in different current densities, this parameter was investigated in three current densities, start-up activity, \u221210 and \u221220\u00a0mA\u00a0cm\u22122. In the initial potential or in some way the hydrogen reduction potential of the Ni8Co2Se-150 sample with \u221262\u00a0mV has the lowest value, which indicates better activity of this sample than other samples. In continuation, this sample maintains its activity trend and has the best activity in \u221210\u00a0mV current density with \u2212149\u00a0mV over potential. In summary, the Ni8Co2Se-150 sample has a minimum of \u027310 for hydrogen evolution in alkaline environments among selenide-based studies. For example: for nickel selenide nano-plates with carbon plates at \u027310 overpotential 184\u00a0mV [29], for nano-forest nickel selenide with nickel foam substrate at \u027310 over potential 203\u00a0mV [30], for cobalt selenide at \u027310 overpotential 472\u00a0mV [31], for molybdenum selenide\u2013nickel selenide at \u027310 overpotential 210\u00a0mV [32], for cobalt selenide with nickel-double layer hydroxide foam substrate and graphene at \u027310 overpotential 260\u00a0mV [33] have reported.Along with the confirmation of the Tafel test, the hydrogen evolution activity decreases with increasing the amount of cobalt over than 20%, as in the Ni2Co8Se sample with \u2212149\u00a0mV overpotential in the current density of \u221210\u00a0mA, we see an activity of almost half of the sample activity. By adding the amount of cobalt to the structure of electrode, we see higher activity of this catalyst, which can be due to the structure with high electrical conductivity of the catalyst and the addition of alternate cobalt at the electrode surface causes better absorption of protons and water molecules [21].In Fig.\u00a02(e, f) the vertical axis is the imaginary impedance (z\u2033) and the horizontal axis is the real impedance (z\u2032). In this figure the distance of the circle center from the origin is Rs\u00a0+\u00a0Rp/2. In the Nyquist diagram, the lower final resistance (Rct) leads to greater hydrogen evolution activity, which the Ni8Co2Se sample has the lowest final resistance and higher electrocatalytic activity, which confirms the LSV diagram and Tafel slope polarization. Also, other samples such as CoSe, Ni2Co8Se, NiSe, Ni4Co6Se, and Ni6Co4Se have higher resistance, respectively, which indicates a lower catalytic property than the Ni8Co2Se sample. This final resistance trend also applies to the synthesized samples at 180\u00a0\u00b0C, which can be attributed to the change in crystal structure from hexagonal to rhombohedral. Also, Zhang et\u00a0al. reported that the catalyst crystal structure changes when temperature or time varies in hydrothermal process [34]. Also, in the equivalent circuit for these curves using Z-view software, we see two condenser capacity and three resistors, which indicates the fracture of the layer and the penetration of active ions into the substrate by the electrolyte. Other parameters in this equivalent circuit are: Rs is the soluble resistance (uncompensated resistance), Rf is the layer resistance, Cdl is the capacitance, Cdp is the imaginary capacitance, and Rct is the final resistance [35]. Numerical quantities obtained from the electrodes using ZView software are shown in Table 2\n. Another diagram that can be obtained from the EIS test is the bode diagram. This diagram consists of two separate curves, one is phase angle variations in terms of frequency logarithm, Fig.\u00a02(g and h) and the other related to impedance logarithm variations in terms of frequency logarithm, Fig.\u00a02(i and j). In the diagram of phase angle changes shown in Fig.\u00a02((g) at 150\u00a0\u00b0C and (h) at 180\u00a0\u00b0C), we see two time-constant indicating existence of two capacitors in the equivalent circuit. The phase angle transfer at the maximum frequency at higher values close to 90\u00b0 indicates less electrocatalytic activity in hydrogen evolution [36]. In this diagram, the Ni8Co2Se-150 sample has the lowest phase angle transfer value, which indicates the high activity of this sample in hydrogen evolution reaction. In the bode curve shown in Fig.\u00a02(i and j), the impedance at the highest frequency on the right side of the curve is equivalent to the soluble resistance and the impedance at the lowest frequency on the left side of the curve is equivalent to the resistance of the whole system. As expected, the Ni8Co2Se sample at both synthesis temperatures shown in Fig.\u00a02((i) at 150\u00a0\u00b0C and (j) at 180\u00a0\u00b0C) has the lowest final resistance, which is confirmed by other diagrams obtained from electrochemical tests. Another noteworthy point in this diagram is the slope between the soluble resistance and the final resistance, which is equivalent to the capacitive region. Increased capacitive region indicates the inhibitory properties against the entry of electrolytes and corrosive agents into the coating and substrate.Since the Ni8Co2Se-150 sample synthesized at 150\u00a0\u00b0C, which was known as the optimum sample in the hydrogen evolution process according to other electrochemical tests, was subjected to chronoamperometry stability test. As can be seen in the diagram in Fig.\u00a02(k), after 12\u00a0h at a potential of 150\u00a0mV, the sample does not show a significant drop in current, indicating the catalyst stability in an alkaline environment for long periods. To compare the catalytic properties of the Ni8Co2Se-150 electrode, after stability test for 12\u00a0h, a linear sweep voltammetry test (LSV) was performed on the sample and the results of the LSV test shown in Fig.\u00a02(l). Failure to move the LSV diagram on the horizontal axis (potential axis) indicates no dispersion and ultimately a decrease in the catalytic properties of the electrode. The fact that no significant decrease in hydrogen evolution potential occurred after 12\u00a0h indicates that this sample will remain stable for longer periods.One of the applications of cyclic voltammetry tests is to measure the reversibility of the electrode. The greater difference between the anode and cathode peaks, the electrode is less reversible. All samples were scanned at 10, 20, 30, 40, 50, 70, and 100\u00a0mV\u00a0s\u22121 scan rates. The cyclic graph for the synthesized electrodes at 150\u00a0\u00b0C is shown in Fig.\u00a03\n(a\u2013f) and 150\u00a0\u00b0C is shown in Fig.\u00a03(i\u2013n). The reactions performed during the test include the oxidation reaction (related to the anodic peak) and the reduction reaction (related to the cathodic peak). An anode peak and a cathode peak appeared in all samples. As shown in Fig.\u00a03, the anode peak current increases with increasing scan rate, indicating the high electrocatalytic activity of the electrodes in the redox reactions. As it is known, with increasing the scan rate without any apparent change in the electrode, the current density has increased, which indicates the stability of the electrode material. The peak symmetry indicates the excellent reversibility of the electrodes. The displacement of the peaks towards positive or negative currents indicates the property of electron transfer and electrical conductivity, that is, the farther apart the anode and cathode peaks, the higher catalytic property of the sample, which separates with increasing scanning rate of the anode and cathode peaks, and the difference between the anodic and cathodic peaks is linearly proportional to the square root of the scanning rate. These behaviors suggest that faradaic reactions to store energy are a process controlled by the penetration of electrolyte ions, which may be as following:\n\n(3)\n\n\n\n(\nNi,Co\n)\n\n\nSe\n2\n\n+\n2\n\nOH\n\u2013\n\n\u21cc\nNiSeOH\n+\nCoSeOH\n+\n2\n\ne\n\u2013\n\n\n\n\n\n\n\n(4)\n\n\nCoSeOH\n+\n\nOH\n\u2013\n\n\u21cc\nCoSeO\n+\n\nH\n2\n\nO\n+\n\ne\n\u2013\n\n\n\n\n\n\n\n(5)\n\n\nNiSeOH\n+\n\nOH\n\u2013\n\n\u21cc\nNiSeO\n+\n\nH\n2\n\nO\n+\n\ne\n\u2013\n\n\n\n\n\nThe cyclic diagram of the synthesized electrodes at 150\u00a0\u00b0C is shown in Fig.\u00a03(g, h) and for the synthesized electrodes at 180\u00a0\u00b0C in Fig.\u00a03(o, p). Scan rates of 10\u00a0mV\u00a0s\u22121\u00a0(g and o) and 100\u00a0mV\u00a0s\u22121\u00a0(h and p) were compared in all samples and showed that by adding some cobalt to the nickel selenide structure, the area under the CV graph increases, and along with it the difference between the anodic and cathodic peaks also increases, which in both 150 and 180\u00a0\u00b0C indicates an increase in current density and thus an improvement in catalytic properties, which the area increase in the Ni8Co2Se sample is quite clear. Then, by increasing the cobalt to nickel ratio over than 20%, we encounter a decrease in the anodic and cathodic peak differences, which indicates that increasing the cobalt to nickel ratio over than 20% causes a decrease in the catalytic properties of the electrodes, which is confirmed by other electrochemical tests.\nTable 3\n shows the comparison of the difference between anodic and cathodic peak currents (\u0394J) on the CV diagram of the synthesized samples at 150 and 180\u00a0\u00b0C at a scan rate of 10\u00a0mV\u00a0s\u22121. This table shows that all samples synthesized at 150\u00a0\u00b0C have a higher \u0394J number than samples synthesized at 180\u00a0\u00b0C and the highest value is related to the Ni8Co2Se-150 sample.We further used the Comparative chart in Fig.\u00a04\n to demonstrate the preference of electrocatalytic activity of the Ni8Co2Se sample in compared previous research which show Ni8Co2Se sample with Release activity 187 mv Contains more optimal activity than some other catalysts.In summary, to synthesize the optimal NiCoSe composition, we synthesized different compositions at two different temperatures and examined the hydrogen evolution. We showed that by slightly increasing the amount of cobalt in this composition, the hydrogen evolution activity increases so that in Ni8Co2Se sample with an overpotential of \u2212185\u00a0mV at current density of \u221210\u00a0mA\u00a0cm\u22122 and \u2212201\u00a0mV at a current density of \u221220\u00a0mA\u00a0cm\u22122\u00a0has the highest activity in hydrogen evolution. Also, the mentioned sample has maintained its activity after 12\u00a0h in the chronoamperometry test, which is an indication of the proper conditions of this electrocatalyst.No funding was received for this work.All listed authors meet the ICMJE criteria. We attest that all authors contributed significantly to the creation of this manuscript, each having fulfilled criteria as established by the ICMJE.No conflict of interest exists.", "descript": "\n                  Designing of complex nanostructure with a high specific surface area using transition metal selenide seems to be a necessity in response to the challenges of the hydrogen release process for renewable energy. Herein, we synthesized the electrocatalyst nickel foam-supported with nanostructured nickel\u2013cobalt selenide using the hydrothermal process at 150 and 180\u00a0\u00b0C and has been investigated in hydrogen evolution reaction. Synthesis of nickel\u2013cobalt selenide nanostructure on nickel foam carried out with different Ni+2:Co+2\u00a0mol ratios of 0:10, 2:8, 6:4, 4:6, 8:2, and 10:0 in the structure of electrodes. The XRD results indicate the formation of the nickel\u2013cobalt selenide phase in different ratios. The FESEM and TEM results show the formation of mesoporous three-dimensional nano-lactic with petal thickness in the range of 20\u201330\u00a0nm. Also, to evaluate the properties and electrocatalytic efficiency, electrochemical tests of Tafel slope, cyclic voltammetry, linear sweep voltammetry (LSV), and electrochemical impedance spectroscopy (EIS) were used. The Tafel result test shows a cathodic slope of \u221261.3\u00a0mV and an exchange current density of 0.86\u00a0mA\u00a0cm\u22122 in the Ni8Co2Se-150 (Ni+2:Co+2\u00a0=\u00a08:2) sample, which LSV shows \u2212185\u00a0mV at a current density of \u221210\u00a0mA\u00a0cm\u22122. EIS test shows a polarization resistance value of 2744\u00a0\u03a9 for mentioned sample and was selected as the best sample with the highest catalytic activity.\n               "}
{"full_text": "Hydrogen peroxide (H2O2) is one of the cornerstones of the chemical industry, with a wide range of applications in areas such as pulp bleaching, chemical synthesis, and disinfection. The demand for H2O2, considered a green oxidant that generates only O2 and H2O as by-products, is constantly increasing.\n1\n\n,\n\n2\n The global H2O2 market has a value of $3.1 billion and is predicted to reach over $4 billion by 2027 at an annual growth rate of 4.2%.\n3\n The history of H2O2 production dates back to the early 19th century, when it relied on the chemical reaction of BaO2 with HCl and electrolysis of H2SO4 solution.\n4\u20136\n Later, the autooxidation process became a major route for H2O2 synthesis,\n7\n followed by the emergence of the anthraquinone process in 1939, which remains a mainstay in the industry.\n8\n The anthraquinone process is currently implemented at a large scale and supplies approximately 95% of global demand.\n1\n\n,\n\n2\n\n,\n\n9\n However, this process requires pressurized H2 gas and Pd-based hydrogenation catalysts and involves energy-intensive distillation steps. Along with intensive energy requirements, the process further suffers from the generation of organic wastes by side reactions. Furthermore, the manufacture of highly concentrated products is more economically viable because of its centralized plant system, which, however, imposes additional costs for transportation. In addition, H2O2 decomposition occurs faster at high concentrations, which compels the use of stabilizing agents such as sodium pyrophosphate and chelating agents. Therefore, alternative processes for H2O2 synthesis are currently being widely investigated. Direct synthesis is a straightforward method in which H2 and O2 gases react to produce H2O2.\n10\n However, the high thermodynamic activity of this reaction obliges H2 gas to be diluted with CO2 or N2, and Pt-group-metal (PGM)-based catalysts are required to control the reactivity and reaction pathways. This method is also plagued by relatively low H2O2 selectivity and the potential risk of explosion.In this regard, the electrochemical production of H2O2 has attracted significant attention.\n11\u201318\n This method enables small-scale, on-site, and continuous production of H2O2. It is also safe and environmentally benign because no carbon-involving side products are generated, and aqueous electrolytes (H2O and H+) serve as hydrogen sources. There are two major pathways for H2O2 electrosynthesis: oxygen reduction reaction (ORR) and water oxidation reaction (WOR). The ORR can take place in several electrochemical and chemical reaction pathways, among which the desired H2O2 generation occurs by the two-electron pathway ORR (2e\u2212 ORR; Equation\u00a01).\n\n2\n\ne\n\u2212\n\n\nORR\n:\n\n(Equation\u00a01-1)\n\n\nO\n2\n\n+\n2\n\nH\n+\n\n+\n2\n\ne\n\u2212\n\n\u2192\n\nH\n2\n\n\nO\n2\n\n\nacidic\n\n\n\n\n\n\n(Equation\u00a01-2)\n\n\n\nO\n2\n\n+\n\nH\n2\n\nO\n+\n\n\n2\ne\n\n\u2212\n\n\u2192\n\nHO\n2\n\u2212\n\n+\n\nOH\n\u2212\n\n\n(\nalkaline,\n\npH\n\n>\n\n11.7\n)\n\n\n\n\n\nHowever, H2O2 production via 2e\u2212 ORR is thermodynamically unfavorable compared with 4e\u2212 ORR (Equation\u00a02). The generated H2O2 is unstable, leading to further reduction to H2O (Equation\u00a03) or chemical decomposition (Equation\u00a04).\n\n\n4\n\ne\n\u2212\n\nORR\n\n:\n\n(Equation\u00a02-1)\n\n\n\nO\n2\n\n+\n4\n\nH\n+\n\n+\n4\n\ne\n\u2212\n\n\u2192\n2\n\nH\n2\n\nO\n\n(\nacidic\n)\n\n\n\n\n\n\n\n(Equation\u00a02-2)\n\n\n\nO\n2\n\n+\n2\n\nH\n2\n\nO\n+\n4\n\ne\n\u2212\n\n\u2192\n4\n\nOH\n\u2212\n\n\n(\nalkaline\n)\n\n\n\n\n\nElectrochemical H2O2 reduction:\n\n(Equation\u00a03-1)\n\n\n\nH\n2\n\n\nO\n2\n\n+\n2\n\nH\n+\n\n+\n2\n\ne\n\u2212\n\n\u2192\n2\n\nH\n2\n\nO\n\n(\nacidic\n)\n\n\n\n\n\n\n\n(Equation\u00a03-2)\n\n\n\nHO\n2\n\u2212\n\n+\n\nH\n2\n\nO\n+\n2\n\ne\n\u2212\n\n\u2192\n3\n\nOH\n\u2212\n\n\n(\nalkaline\n)\n\n\n\n\n\nH2O2 disproportionation:\n\n(Equation\u00a04)\n\n\n2\n\nH\n2\n\n\nO\n2\n\n\u2192\n2\n\nH\n2\n\nO\n+\n\nO\n2\n\n\n\n\n\nTherefore, the development of active and selective electrocatalysts for the 2e\u2212 ORR is crucial for the successful implementation of electrochemical H2O2 production technology. The same is applied to H2O2 electrosynthesis via WOR, as it presents similar issues regarding multiple reaction pathways and controlling reaction selectivity.This paper provides a comprehensive review of selective electrocatalysts for the 2e\u2212 ORR and a current understanding of the electrocatalytic process and reactivity-determining factors. First, we present the recent progress in the design of H2O2 electrosynthesis catalysts via 2e\u2212 ORR and in the understanding of the nature of the active sites and their impact on the activity and selectivity. The electrocatalyst section is categorized according to the chemical composition of the catalysts: PGM-based atomically dispersed catalysts (ADCs), non-PGM-based ADCs, and metal-free heteroatom-doped carbon catalysts. The H2O2 electrosynthesis activity of each class of catalysts is benchmarked to understand the current status of advancement and to provide guidelines for future studies. Interfacial factors and phenomena that regulate the H2O2 production activity and selectivity are also introduced. We suggest guidelines for the accurate measurement of the catalysts\u2019 performance on H2O2 electrosynthesis, which have been largely overlooked in the current literature. Finally, reactors designed for high-current-density operation and systems that utilize electrosynthesized H2O2 are presented.A mechanistic understanding of a specific reaction is essential for the rational design of efficient electrocatalysts. For the ORR, two major mechanisms, the associative mechanism and the dissociative mechanism, are suggested (Figure\u00a01\n). In the associative mechanism, O2 adsorption and subsequent \u2217OOH formation commonly occur regardless of the final product, according to the following equation:\n\n(Equation\u00a05)\n\n\n\nO\n2\n\n+\n\u2217\n+\n\n(\n\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n)\n\n\u2192\n\nOOH\n\n\n\u2217\n\n\n\n\n\nIf the binding strength of the \u2217OOH intermediate on a surface site is medium or weak, further proton and electron transfer occur, leading to the formation of H2O2 via the 2e\u2212 pathway (Equation\u00a06).\n\n(Equation\u00a06)\n\n\u2217\nOOH\n+\n\n\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n\n\u2192\n\nH\n2\n\n\nO\n2\n\n+\n\u2217\n\n\n\nHowever, when the \u2217OOH intermediate is strongly adsorbed, the O\u2013O bond dissociates to form the \u2217O intermediate. As a result, sequential 2e\u2212 processes complete the 4e\u2212 ORR (2e\u2212\n\n\n\u00d7\n\n 2e\u2212 pathway) to generate H2O (Equation\u00a07).\n\n(Equation\u00a07-1)\n\n\n\u2217\nOOH\n+\n\n(\n\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n)\n\n\u2192\n\nO\n\n\n\u2217\n\n+\n\nH\n2\n\nO\n\n\n\n\n\n\n(Equation\u00a07-2)\n\n\n\nO\n\n\n\u2217\n\n+\n\n(\n\n2\n\nH\n+\n\n+\n2\n\ne\n\u2212\n\n\n)\n\n\u2192\n\nOH\n\n\n\u2217\n\n+\n\n(\n\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n)\n\n\u2192\n\nH\n2\n\nO\n+\n\u2217\n\n\n\n\nIn the dissociative mechanism, a pair of adjacent surface sites strongly adsorb O2 molecules, which are broken down into \u2217O intermediates. Only \u2217O and \u2217OH intermediates are involved in this mechanism without the formation of the \u2217OOH intermediate, leading to H2O product via the 4e\u2212 pathway (Equations 8 and 7-2).\n\n(Equation\u00a08-1)\n\n\n\nO\n2\n\n+\n\n2\n\u2217\n\n\u2192\n\n2\n\u2217\n\nO\n\n\n\n\n\n\n(Equation\u00a08-2)\n\n\n\n2\n\u2217\n\nO\n+\n\n(\n\n2\n\nH\n+\n\n+\n2\n\ne\n\u2212\n\n\n)\n\n\u2192\n\n2\n\u2217\n\nOH\n\n\n\n\nTherefore, high H2O2 product selectivity from the ORR can be achieved using electrocatalysts that favor the associative mechanism. Another important factor is the \u2217OOH binding energy of the catalytic active site. Active sites with low \u2217OOH binding energies are unfavorable for the first O2 adsorption, according to the linear scaling relation of the reaction intermediates, reactants, and products. The weak binding thus leads to high H2O2 selectivity but a low conversion rate (high overpotential). Likewise, the scaling relation also suggests that the active sites with high affinity for \u2217OOH intermediates also show high binding strength with \u2217OH and \u2217O intermediates, resulting in low H2O2 selectivity. According to the binding energy criteria, inert PGMs such as Au and Ag with low \u2217OOH binding energies would have high H2O2 selectivity and low activity, whereas oxophilic transition metals such as Fe, Co, Ni, and Mn would show low 2e\u2212 ORR selectivity and activity. Pt and Pd, with optimal \u2217O binding energies, favor the 4e\u2212 ORR.\n11\n\nThe arrangement of catalyst surface atoms affects the adsorption geometry and binding strength and thus the reaction pathways. For instance, ab initio calculations revealed that the different adsorption strengths of oxygenated reaction intermediates depend on the adsorption sites on crystalline Pt surfaces.\n19\n Therefore, controlling the atomic arrangement can serve as an effective way of catalyst design, leading to different O2 adsorption modes and reaction mechanisms. For a geometrically isolated site, O2 is adsorbed in the end-on mode, rendering preference for the associative mechanism, whereas an ensemble site can adsorb O2 in a side-on geometry favoring the dissociative mechanism (Figure\u00a01). In summary, ORR selectivity can be adjusted by the chemical composition and geometric structure of the active sites. Various classes of electrocatalysts, including monometallic PGMs (Au and Ag),\n20\n\n,\n\n21\n PGM-based alloys (Au\u2013Pd, Pt\u2013Hg, Pd\u2013Hg, and Pd\u2013Au),\n22\u201324\n PGM-based ADCs (Os, Ir, Rh, Ru, Pt, and Pd),\n25\u201331\n non-PGM-based ADCs (Fe, Co, Ni, and Mn),\n32\u201346\n metal compounds (oxides, sulfides, borides, etc),\n47\u201350\n and metal-free doped-carbon catalysts,\n51\u201388\n have been explored as selective electrocatalysts for the 2e\u2212 ORR. In this review, we focus on low- and non-PGM catalysts: PGM-based ADCs, non-PGM-based ADCs, and metal-free doped-carbon catalysts, which will be reviewed in the next section.In the early stage of PGM-based selective electrocatalysts for the 2e\u2212 ORR, Pt- and Pd-based alloys diluted with less catalytically active elements (i.e., Hg and Au) were demonstrated to exhibit excellent H2O2 selectivity. This reactivity can be explained by the peculiar atomic arrangements of active sites that inhibit O\u2013O bond breakage and stabilize the \u2217OOH intermediate.\n22\u201324\n However, only a small fraction of the precious metals in the alloy nanoparticles are exposed on the surface, and thus, their utilization efficiency is low. In addition, the use of toxic or other precious metals as secondary metals raises concerns from environmental and economic points of view. PGM-based ADCs allow full utilization of PGM with atomically dispersed active sites, which significantly reduces the amount of expensive PGM. Hence, considerable effort has been devoted to the development of PGM-based ADCs on appropriate supports with optimized anchoring sites.\n25\u201331\n Carbon is the most commonly used support material. However, carbon itself possesses a very low capacity for generating single atomic sites owing to its deficient anchoring sites. In this regard, electron-rich heteroelements such as N, S, and O have been deliberately introduced into the carbon lattice, as these dopants have the propensity to form coordination bonds with the metal atoms. In a notable example, Choi and co-workers developed S-doped microporous carbons as supports for atomically dispersed Pt.\n25\n Because S has a strong affinity for most PGMs, the S-doped carbon sustains a high loading of Pt atoms of up to 5 wt\u00a0%. Unlike Pt nanoparticles, which are favorable for the 4e\u2212 ORR, the atomically dispersed Pt coordinated to S exhibited high 2e\u2212 ORR selectivity.The electrocatalytic activity and selectivity of ADCs strongly depend on the composition and structure of their active sites. Therefore, a synthesis platform that can produce ADCs across a broad compositional range without any clusters or particles is in high demand, which allows the unraveling of the trends of 2e\u2212 ORR activity and selectivity. Joo and co-workers established a general strategy for PGM-based ADCs with various types of metal centers (Os, Ru, Ir, Rh, and Pt) to study their 2e\u2212 ORR performance.\n26\n The \u201ctrapping-and-immobilizing\u201d strategy developed in this work could effectively prevent the agglomeration of PGM atoms during impregnation and reductive activation steps. In this process, a carbon support was coated with an ionic-liquid-derived carbonaceous layer containing electron-rich heteroatoms. During impregnation and drying, such heteroatoms can trap metal precursors via electrostatic interactions. A sacrificial silica layer was then coated on the metal-precursor-impregnated carbon to immobilize the metal species and thereby to mitigate their agglomeration during reductive thermal activation. Thus, the synthesized ADCs (denoted as M1) generally showed higher 2e\u2212 ORR selectivity than the corresponding PGM nanoparticle (MNP) catalysts, which favored 4e\u2212 ORR. Among the M1 catalysts, Rh1 showed the highest H2O2 production activity (lowest overpotential), and Pt1 showed the highest H2O2 selectivity. The difference in H2O2 selectivity between the ADCs and NP catalysts can be explained by different reaction mechanisms; the M1 catalysts follow an associative mechanism (end-on adsorption) in which the OOH\u2217 intermediates are preserved, whereas the ORR over the MNP catalysts proceeds via a dissociative mechanism due to the presence of metal ensemble sites. The trend for H2O2 production among the M1 catalysts can be explained by density functional theory (DFT) calculations. The difference between the binding energies of the \u2217O and \u2217OOH species (\u0394G\nO \u2013 \u0394G\nOOH), which describes the thermodynamic stability of \u2217OOH intermediates, showed a linear relationship with the H2O2 yield (Figure\u00a02A). The oxygen-binding-energy trend was experimentally verified by an O2 temperature-programmed desorption (TPD) experiment, where the O2 desorption temperature had an inverse linear relationship with the H2O2 selectivity. The M1 catalysts exhibited higher onset potentials for the 2e\u2212 ORR than the MNP catalysts, and a volcano-shaped plot was observed between the onset potential and oxygen binding energy of the M1 catalysts. Among them, Rh1, which has an appropriate oxygen binding energy, had the highest onset potential for H2O2 production. In contrast, the MNP catalysts generally exhibited low onset potentials for 2e\u2212 ORR, which had a relatively poor correlation with the \u2217O binding strength (Figure\u00a02B).The chemical identity of the coordinating atoms is another critical factor that regulates the catalytic properties. Because the preparation of ADCs typically involves high-temperature reactions of precursor mixtures, the control of the coordinating atoms can be mostly achieved by using a precursor comprising targeted elements. Joo\u2019s group demonstrated reversible modification of the coordination environment of PGM-based ADCs by gas-phase ligand exchange reactions.\n27\n When the as-synthesized PGM-based ADCs (described above) were subjected to heat treatment under CO or NH3 gas, the ligands of the as-prepared ADCs were substituted with the corresponding gas molecules. Depending on the type of ligand, the H2O2 production activity and selectivity could be noticeably tuned; CO-coordinated Rh ADCs exhibited a higher onset potential but lower selectivity toward the 2e\u2212 ORR, and the reverse trend was observed for the NH3-coordinated catalyst. Importantly, ligand exchange was reversible, which consequently modulated the oxidation state of the metal center and its electrocatalytic activity in a reversible manner (Figure\u00a02C). DFT calculations revealed that the switching behavior of the ORR selectivity originates from the change in the electronic structure and \u2217OOH binding energy with respect to the ligand type.These examples highlight the importance of the chemical identities of the metal centers and coordinating atoms. Although carbon is usually utilized as a support for ADCs, other conductive materials have great potential, although many remain unexplored.\n28\n\n,\n\n29\n Lee and co-workers investigated TiN and TiC as supports for atomically dispersed Pt (Pt1/TiN and Pt1/TiC, respectively).\n29\n Pt1/TiC showed higher activity and selectivity for H2O2 production than Pt1/TiN (Figure\u00a02D). DFT calculations suggested that the high oxygen affinity of TiN could facilitate the side-on adsorption of O2 molecules on Pt and Ti atoms, decreasing the 2e\u2212 ORR selectivity. In contrast, on Pt1/TiC, O\u2013O bonds were preserved because of favorable end-on adsorption on the Pt atom, leading to higher H2O2 selectivity (Figure\u00a02E). Li and co-workers developed a redox-based ion-exchange method to anchor Pt atoms on a CuS\nx\n support, where the strong Pt\u2013S coordination enabled high loading of atomically dispersed Pt up to 24.8 wt\u00a0%.\n30\n The optimized catalyst exhibited >90% H2O2 selectivity over a wide potential range of acidic electrolytes (Figure\u00a02F). In contrast, the Pt NP on the CuSx catalysts and CuSx support itself showed low H2O2 selectivity, suggesting that the Pt\u2013S\nx\n sites were active motifs.Non-PGM-based ADCs (also known as M\u2013N/C catalysts because N-doped carbons are usually used as supports) have been studied mainly as 4e\u2212 ORR catalysts for application as hydrogen fuel-cell cathodes. In this regard, M\u2013N/C catalysts exhibiting high 2e\u2212 ORR selectivity are undesirable for the fuel-cell application because 2e\u2212 ORR decreases the conversion efficiency and the generated H2O2 accelerates the deterioration of the fuel cell. However, the rapidly increasing interest in H2O2 electrosynthesis has shed light on non-PGM ADCs with a high 2e\u2212 ORR selectivity. In M\u2013N/C catalysts, three major structural factors critically influence ORR activity and selectivity. These include the type of metal center, type of coordination atom and geometry (first coordination shell), and presence of heteroatom species adjacent to the active sites (second coordination shell) (Figure\u00a03A). The impact of the metal center has been intensively studied because its control is relatively simple. Liu and co-workers investigated the H2O2 activity and selectivity trends of M\u2013N/C catalysts with controlled metal centers (Mn, Fe, Co, Ni, and Cu).\n32\n DFT calculations predicted that Co\u2013N/C would exhibit the highest performance for 2e\u2212 ORR because of the optimal \u2217OOH oxygen binding energy of the Co\u2013N active sites. The experimental activity trends in terms of overpotential were found to be Co\u00a0>\u00a0Fe\u00a0>\u00a0Ni\u00a0>\u00a0Mn\u00a0>\u00a0Cu (Figure\u00a03B), whereas the selectivity was in the order of Co\u00a0>\u00a0Mn\u00a0>\u00a0Ni\u00a0>\u00a0Cu\u00a0>\u00a0Fe (Figure\u00a03C), which is consistent with the DFT calculations. Similar trends have been verified in other studies;\n33\n\n,\n\n34\n however, considerable discrepancies in the obtained results have been found. This difference could originate from a change in the coordination environment depending on the catalyst preparation conditions.Therefore, the catalytic effect of the first coordination shell was examined.\n35\u201341\n Wang and co-workers prepared a set of Fe-based ADCs with different coordinating atoms: one with typical N-coordination (Fe-N-CNT) and the other with C- and O-coordination (Fe-CNT).\n35\n The two catalysts showed dramatically different ORR selectivity trends, where Fe-N-CNT followed the 4e\u2212 ORR, whereas Fe-CNT favored the 2e\u2212 ORR (Figure\u00a03D). This was further substantiated by demonstrating the selectivity shift from 4e\u2212 to 2e\u2212 ORR by reductive heat treatment of Fe-N-CNT, which transformed the 4e\u2212-ORR active Fe\u2013N center to the 2e\u2212-ORR active Fe\u2013C\u2013O center (Figure\u00a03E). The functional groups adjacent to the atomically dispersed active site, even if not directly bonded to it, can also significantly impact electrocatalytic processes.The control of the so-called second coordination shell has recently emerged as a simple way to tune activity.\n41\u201343\n Hyeon, Sung, and co-workers investigated the activity of electrochemical H2O2 production of Co\u2013N/C catalysts with neighboring oxygen functionalities in the second coordination shell as activity and selectivity modifiers.\n42\n DFT calculations revealed that the epoxy groups in the second coordination shell would optimize the \u2217OOH binding energy of the Co-based active center and improve the 2e\u2212 ORR activity of the Co\u2013N/C catalyst. In contrast, the raw Co\u2013N sites are expected to exhibit a 4e\u2212 pathway preference. Experimentally, a Co\u2013N/C catalyst was prepared on graphene oxides rich in epoxy groups (denoted as Co1\u2013NG(O)). As a comparative sample, the Co1\u2013NG(O) catalyst was annealed at high temperatures under inert conditions, resulting in a Co1\u2013NG(R) catalyst. Spectroscopic analyses identified a higher content of epoxy groups in Co1\u2013NG(O) than in Co1\u2013NG(R). Co1\u2013NG(R) showed a positively shifted disk onset potential and higher diffusion-limited disk current density than Co1\u2013NG(O), whereas Co1\u2013NG(O) showed a higher ring onset potential and diffusion-limited ring current density than Co1\u2013NG(R) (Figure\u00a03F). These results suggest that Co1\u2013NG(R) and Co1\u2013NG(O) catalyze 4e\u2212 ORR and 2e\u2212 ORR, respectively. This was verified by the H2O2 selectivity results, where Co1\u2013NG(O) exhibited a high H2O2 selectivity (>80%) over a wide potential range (Figure\u00a03G). Although the promotion effect of oxygen groups in M\u2013N/C catalysts has been shown by other groups, various combinations of the metal centers and functionalities should be further explored to obtain the best synergistic effect.Metal-free carbon-based catalysts are perhaps the most attractive class of non-PGM catalysts because of their low price, high electrical conductivity, and abundance. As such, carbon-based catalysts have been actively studied for many electrochemical reactions in the past few decades. For the 2e\u2212 ORR, a surge of research interest in carbon-based electrocatalysts has emerged after promising results were demonstrated by Cui\u2019s and McCloskey\u2019s groups.\n51\n\n,\n\n52\n Both revealed that oxygen functional groups on carbon nanomaterials play critical catalytic roles in H2O2 electrosynthesis from O2. Cui and co-workers showed that the 2e\u2212 ORR selectivity and activity increased linearly with the oxygen content, and these correlations have since been further confirmed with other carbon nanomaterials.\n51\n DFT calculations suggested that C\u2013O\u2013C (ether-type) and O\u2013C=O (carboxylate-type) moieties are possible active sites. McCloskey and co-workers prepared few-layered mildly reduced graphene oxide (GO) catalysts by thermal reduction of GO at controlled temperatures (F-mrGO(X), X\u00a0= temperature).\n52\n\nF-mrGO(600) exhibited a higher H2O2 electroproduction activity than F-mrGO(300) and F-mrGO (Figure\u00a04A). The high activity of F-mrGO(600) containing a lower O content than the others indicates that the annealing step transformed the less-active O groups in F-mrGO into active species. Spectroscopic analyses suggested that the basal ethers were transformed into active edge ether groups at high temperatures (Figure\u00a04B). Subsequent studies have suggested several oxygen functional groups, including ether, epoxy, carbonyl/quinone, and carboxyl, as active sites for the 2e\u2212 ORR. In an effort to identify reactive oxygen species, Joo and co-workers investigated the intrinsic H2O2 electrosynthesis activity of respective carboxyl, carbonyl, and phenol groups, which are typically generated during the oxidative doping process.\n53\n Oxidized graphitic ordered mesoporous carbon (O-GOMC) catalysts containing the three oxygen functional groups were treated with benzoic anhydride (BA), phenyl hydrazine (PH), and 2-bromo-1-phenylethanone (BrPE), which can selectively block phenol, carbonyl, and carboxyl groups, respectively, allowing assessment of the activity of the respective functional groups. The site-blocked catalysts exhibited a considerable decrease in activity compared with the pristine O-GOMC catalysts but to a different extent depending on the blocked oxygen functionality (Figure\u00a04C). To calculate the turnover frequency (TOF), which represents the intrinsic catalytic activity per active site, the number of each functional group was determined by quantifying the number of blocking molecules. The resulting TOF values revealed that the intrinsic activity was higher in order of the carboxyl, carbonyl, and phenol groups (Figures\u00a04D and 4E). Liu and co-workers also performed a chemical blocking experiment using O-doped carbon-nanosheet catalysts, where C=O groups were identified as the active species.\n54\n There are still inconsistencies in the proposed active sites, which mainly originate from the fact that many different types of functional groups are simultaneously generated during uncontrolled doping procedures.Nitrogen doping is one of the most common methods to promote the activity of carbon catalysts; however, H2O2 electrosynthesis with N-doped carbons has been relatively less explored than with O-doped carbons. Strasser and co-workers prepared a set of N-doped mesoporous carbons using a hard-templating method by changing the synthesis conditions, such as the number of precursors and the carbonization temperatures. They found that the activity trends of the prepared carbons were not well explained by the zeta potential or defect site density. Instead, a volcano-type activity behavior with the N content was found, suggesting that excessive N active species can decrease H2O2 selectivity, although its origin remains elusive.\n56\n Qiao and co-workers investigated the ORR activity and selectivity trends of N-doped carbon model catalysts called N-mFLG-X (X\u00a0= mass ratio of melamine to glycine).\n57\n The precursor composition was found to control the pyrrolic-N content with a fixed relative amount of pyridinic N and graphitic N (Figure\u00a04F). The H2O2 electrosynthesis efficiency increased as the quantity of pyrrolic N in the catalysts increased (Figures\u00a04G and 4H). In situ X-ray absorption near-edge structure (XANES) analysis indicated that N-mFLG-8, the best catalyst, contained a larger amount of OOH\u2217 intermediates and fewer O\u2217 intermediates than N-mFLG-16, which demonstrated that N-mFLG-8 favored the 2e\u2212 ORR, whereas N-mFLG-16 favored the 4e\u2212 ORR. The electrocatalytic properties for the 2e\u2212 ORR of heteroatom-doped carbons other than O- and N-doped carbons have rarely been reported. Wang and co-workers screened various heteroatom-doped carbons (B, N, P, and S) for H2O2 electrosynthesis.\n58\n Doping was performed by the thermal reaction of O-doped carbon black with precursors containing the desired heteroelements. All the catalysts possessed similar structural properties, except for the type of heteroatom dopant, enabling the study of model catalysts. The doped carbons generally exhibited high 2e\u2212 ORR activity with H2O2 selectivity of over 60%, and the trends were in the order of B, N, S, and P (Figures\u00a04J and 4K). N-doped carbon showed a relatively high onset potential but a relatively low H2O2 selectivity, whereas P-doped carbon exhibited the opposite trend. DFT calculations suggested that the thermodynamic stability of the \u2217OOH intermediate on B-doped carbon may explain the high H2O2 selectivity.In addition to heteroatom doping, the H2O2 production activity of carbon electrocatalysts can be boosted by structural tuning. Structural modification includes pore structure control, which can increase the surface area and enhance the mass transport of carbon catalysts, and the associated formation of defective carbon sites and surface curvatures, which can bring changes in the electronic structure of carbons. Nabae and co-workers prepared various N-doped carbons by hard templating using mesoporous silica KIT-6 with controlled surface area and porosity.\n59\n They found that mesoporous carbons exhibited approximately double 2e\u2212 ORR activity than microporous carbons, highlighting the importance of facilitated mass transport in the mesopores (Figures\u00a05A and 5B). Bao and co-workers demonstrated similar results by comparing the electrocatalytic activities of mesoporous and microporous carbons (MesoCs and MicroCs, respectively) and highly ordered pyrolytic graphite (HOPG).\n60\n MesoCs exhibited a higher onset potential than MicroCs; however, both catalysts possessed a similar H2O2 selectivity of approximately 70% (Figure\u00a05C). In addition to the effect of porosity-dependent mass transport, DFT calculations suggested the possibility of enhancing the activity of porous carbons through the formation of defect sites (Figure\u00a05D). Optimal defect structures were identified based on the binding strength of the \u2217OH intermediates. Joo and co-workers exploited a highly curved surface structure in MesoCs to enhance H2O2 electrosynthesis activity.\n61\n They prepared GOMCs using a hard-templating method with highly aromatic carbon precursors. Periodically arrayed graphitic carbon nanorods in GOMCs, which were composed of vertically stacked graphene nanosheets, allowed maximum exposure of active graphitic carbon edges (Figure\u00a05E). The H2O2 electrosynthesis activity of the edge-site-rich GOMC was 28 times higher than that of the basal plane-rich CNT, whereas both catalysts showed similarly high H2O2 selectivity of approximately 90% regardless of the number of edge sites (Figures\u00a05F\u20135H). Oxidative treatment of GOMC selectively installed oxygen functional groups at the edge and achieved an additional 3.5-fold increase in activity.Electrocatalysis is essentially an interfacial process that occurs at the boundary between an electrocatalyst and electrolyte. Therefore, in addition to the active-site structure of a catalyst, understanding the role of interfacial species and phenomena is equally important for tuning reaction selectivity. Joo and co-workers synthesized a set of OMCs doped with various combinations of heteroatoms (N, S, and O).\n62\n Kelvin probe force microscopy (KPFM) was used to correlate the ORR activity and selectivity trends of the doped OMCs with the nanoscale surface charge density (Figure\u00a06A). The electrocatalytic ORR activity increased in the order N,S,O-OMC\u00a0>\u00a0N,O-OMC\u00a0>\u00a0S,O-OMC\u00a0>\u00a0O-OMC\u00a0>\u00a0OMC, whereas the H2O2 selectivity showed the reverse order (Figure\u00a06B). Interestingly, the work functions of the doped OMC catalysts, as assessed by KPFM measurements, were linearly correlated with their H2O2 selectivities and reaction rates (Figure\u00a06C). The results indicate that a doped OMC with a lower work function has a lower energy barrier for donating electrons from its surface to the adsorbed oxygen, which facilitates the formation of OOH\u2217 species and consequently promotes the 4e\u2212 ORR, and vice versa. Hence, the work function of a doped carbon and its capability to promote the first electron transfer to generate OOH\u2217 species can be used as an important descriptor for designing advanced carbon-based ORR electrocatalysts. In this context, Joo\u2019s group prepared edge-rich O-doped GOMCs with tuned O contents and measured their heterogeneous electron transfer (ET) rate constants (k\nobs\n0) using the Nicholson method and an outer-sphere redox species, [Fe(CN)6]3\u2212/4\u2212 (Figure\u00a06D).\n61\n The H2O2 electrosynthesis activity, ET rate, and reciprocal of the charge-transfer resistance exhibited similar volcano-like trends depending on the amount of O in the catalysts (Figure\u00a06E). The k\nobs\n0 values and mass activity (j\nm) of a series of O-doped GOMCs showed a linear relationship with the 2e\u2212 ORR (Figure\u00a06F).In typical aqueous electrolytes, the catalyst surface is charged under an applied potential, and ions are distributed to generate an electrical double layer at the interface. The type and concentration of ions modify the double-layer structure and electrocatalytic properties. Sa and co-workers investigated the effect of ion type and concentration on the 2e\u2212 ORR activity of O-doped carbons.\n63\n Cation was found to be a major ionic species at the interface based on potential of zero charge (PZC) measurements. ORR measurements in alkaline electrolytes with various cations (Cs+, K+, and Li+) revealed a cation-dependent activity trend, Cs+\u00a0>\u00a0K+\u00a0>\u00a0Li+, whereas no change in the activity was observed by the type of the anion. In addition, a higher cation concentration was effective for increasing the 2e\u2212 ORR activity. These activity trends correlated with the ET kinetics at the interface, which is known to be influenced by the structure of the cation hydration shell.The interfacial structure can also be influenced by the presence of bulky surfactant molecules, as investigated by Guo and co-workers (Figure\u00a06G).\n64\n The linear sweep voltammetry (LSV) curves of carbon black (CB) for the ORR in 0.1\u00a0M KOH indicated that the addition of a cationic surfactant (cetyltrimethylammonium bromide [CTAB]) increased the H2O2 selectivity of CB, whereas the presence of an anionic surfactant (sodium dodecyl sulfate [SDS]) exerted an adverse effect on selectivity, with both surfactants having a marginal impact on the catalytic activity (Figure\u00a06H). These phenomena could be attributed to the attractive or repulsive Coulombic forces between the product HO2\n\u2212 (alkaline form of H2O2) and the cationic or anionic surfactant, respectively; the interaction with CTAB was suggested to induce faster peroxide desorption kinetics, whereas that with SDS promoted further reduction of the peroxide. Furthermore, the authors conducted a kinetic model analysis of the carboxyl (\u2013COO) and carbonyl (C=O) functionalities using O-doped CB. The ratios of the HO2\n\u2212 desorption to HO2\n\u2212 electroreduction rate constants of the O-CB with or without CTAB were calculated, which indicated that an increased \u2013COO/C=O ratio led to better H2O2 production selectivity. The density of the surface carboxyl groups affected not only the rate of H2O2 production but also the rate of peroxide desorption, which collectively facilitated the formation of H2O2. In addition, electrokinetic calculations revealed a 4-fold higher HO2\n\u2212 desorption rate of the CB catalyst with CTAB than that of pristine CB (Figure\u00a06I).Tuning the hydrophobicity/hydrophilicity of electrocatalysts may control their local mass-transport behavior and generate a special interface environment. Sun and co-workers developed honeycomb carbon nanofibers (HCNFs) with high porosity and superhydrophilicity (Figure\u00a06J).\n65\n The HCNFs exhibited higher activity and selectivity for H2O2 production than non-porous solid CNFs (SCNFs). The enhanced performance of the HCNFs was attributed to the superhydrophilicity resulting from rich oxygen functionalities and the surface topography of the carbon matrix. This structure allowed the effective wetting of the catalyst by an aqueous electrolyte and sufficient interaction between the surface and electrolyte. Furthermore, the honeycomb-like pore structure entrapped O2 inside the pores, increasing the local O2 concentration (Figures\u00a06K and 6L).As described above, significant advances have been made in H2O2 electrosynthesis over the past few years. In this section, we summarize the 2e\u2212 ORR activity of selected, high-performance catalysts, which would serve as important guidelines for assessing newly developed catalysts (Figure\u00a07\n; Table\u00a01\n). As key indicators, we selected O2-to-H2O2 mass activity (MA) and site-normalized activity, which are current per catalyst mass and turnover number per second per active site, respectively. The MA and site-normalized activity values were calculated by normalizing apparent activity with the catalyst loadings and concentrations of active species in the developed materials. MA and site-normalized activity represent the device-oriented and intrinsic activities, respectively.H2O2 electrosynthesis activity and selectivity are typically evaluated using a rotating ring disk electrode (RRDE) (discussed in detail in the next section). The two voltametric curves obtained from the RRDE measurements show the reaction rates for different processes: the disk current is related to the total O2 conversion rate, regardless of the product, and the ring current represents the O2-to-H2O2 conversion rate. To calculate the MA of the catalyst, the ring current was extracted and corrected using the collection efficiency. Next, the kinetic ring current (actual reaction rate) was obtained by removing the effect of diffusion-limited current. Finally, the MA value was calculated by dividing the kinetic current by the catalyst loading. A summary of the MA values of PGM-based ADCs, non-PGM-based ADCs, and metal-free heteroatom-doped carbon catalysts as a function of applied potentials is shown in Figures\u00a07A\u20137D, which are categorized by the electrolyte pH. The MA values would be higher for catalysts with (1) active sites with higher intrinsic activity, (2) a larger number of active sites, (3) greater specific surface area, and (4) better mass transport. It should be noted that the MA for H2O2 production generally decreases at lower pH, i.e., acidic and neutral electrolytes (Figures\u00a07A and 7B), which presumably arises from the pH-dependent reaction kinetics. However, in alkaline H2O2 production, H2O2 is relatively unstable and prone to chemical decomposition; hence, the operation of the H2O2 electrosynthesis system under neutral and acidic conditions is preferable for practical applications. Therefore, the development of efficient electrocatalysts for H2O2 synthesis requires an understanding of the pH-dependent reaction mechanism, which has rarely been investigated.TOF provides the most accurate information on the intrinsic activity of an individual active site. However, accurate quantification of the number of accessible active sites and thus obtaining true TOF values are challenging tasks, which arise from two major difficulties: (1) identification of genuine active sites and (2) estimation of accessibility of active sites under reaction conditions. Recently, some groups have assessed the number of active sites based on poisoning methods for calculating the exact TOF value. These methods consist of blocking the active sites via the adsorption of poisoning molecules on a catalyst and analyzing the amount of the molecules that are either attached to the active sites or remained unreacted. The blocking molecules should (1) deactivate the active sites when combined, (2) react with only a single type of active site in the same molar ratio, and (3) be stably attached under the reaction conditions. The poisoning strategy has been particularly successful for ADCs because there are several probe molecules available that have high selective affinity to metal centers. Strasser and co-workers utilized CO pulse chemisorption and subsequent TPD to estimate the accessible active-site density of non-PGM ADCs.\n89\n Because CO binding strength with the metal center is insufficient, the chemisorption experiment should be conducted at a low temperature (193 K). Kucernak and co-workers found that NO2\n\u2212 can selectively bind the active sites of Fe\u2013N/C catalyst and form an adduct that is stable under open circuit potential in electrolytes.\n90\n Poisoning with the NO2\n\u2212 and the subsequent reductive stripping provided the quantitative insight. The stripping charge and the degree of deactivation were correlated to evaluate the intrinsic activity of Fe\u2013N/C catalysts. Although the NO2\n\u2212-based method is versatile, the stability of the ion is pH sensitive, which necessitates a buffered electrolyte for the reproducible measurement, and the optimum pH (5.2) is less relevant to the practical operation conditions. Choi and co-workers developed a CN\u2212 blocking method where poisoning catalysts with CN\u2212 and determining the residual CN\u2212 concentration resulted in quantitative information about the metal center.\n91\n The CN\u2212-based method not only gave a consistent result with the above-mentioned two methods but also could be used to determine the active-site density of a wide variety of catalysts including non-PGM ADCs, a Pt-based ADC, and Pt NPs.For metal-free heteroatom-doped carbon catalysts, the active-site quantification method has rarely been reported, except for the organic reaction-based site-blocking method discussed earlier in the metal-free carbon-based catalysts section. We also note that it is still challenging to estimate TOF values precisely when a catalyst contains two adjacent active sites that show a synergistic activity boost. Because the amount of accessible active sites is undetermined in most literature, we summarized the intrinsic activity of reported catalysts by normalizing with the total metal contents (metal-normalized activity [MeNA]) for ADCs and total heteroatom contents (heteroatom-normalized activity [HNA]) for metal-free doped carbon catalysts instead of TOF value (Figure\u00a07). Because this calculation assumes that every metal atom or heteroatom is accessible and equally active, the summarized MeNA and HNA values represent the lower bound limit of a TOF.\nFigures\u00a07E and 7F show the MeNA values of ADCs in acidic and neutral electrolytes. In acidic media, Pt- and Rh-based PGM ADCs exhibited the highest intrinsic 2e\u2212 ORR activity (Figure\u00a07E). Among the non-PGM ADCs, only the Co-based ADC exhibited significant activity for acidic 2e\u2212 ORR. For the 2e\u2212 ORR in neutral media, a limited number of catalytic activities have been reported only for non-PGM ADCs (Figure\u00a07F). We note that ADCs normally contain metal active centers as well as metal-free heteroatom dopant sites; the latter species can probably contribute significantly to the apparent activity of ADCs measured in alkaline media. The MAs of metal-free heteroatom-doped carbons surpass those of many non-PGM ADCs (Figures\u00a07C and 7D). In this regard, the MeNA values of the ADCs in alkaline media are not shown to prevent any misleading conclusion. For metal-free carbon-based catalysts, although no robust methods for the quantification of active sites are established, it is plausible that the activity is generally proportional to the amount of heteroatom dopants (Figures\u00a07G\u22127I). The HNA values provide several insights. First, like the ADC catalysts, the carbon-based catalysts also exhibit higher activity at alkaline conditions than at neutral and acidic conditions. Second, O-doped carbons show generally better H2O2 electrosynthesis activity than N-doped carbons, as the latter have a propensity for promoting 4e\u2212 ORR. The comparative study on the intrinsic dopant-dependent activity of doped carbon containing each heteroatom (O, N, S, B, etc.) will be of great scientific importance.Establishing robust evaluation protocols for H2O2 electrosynthesis is important for reliably assessing a newly developed catalyst. In this section, the best practices and pitfalls are suggested for the accurate laboratory measurements of H2O2 electrosynthesis. Currently, two major methods are widely used: RRDE and bulk electrolysis (Figure\u00a08A and 8B).RRDE is the most convenient method for the rapid measurement of the kinetics and selectivity of a catalyst. Dissolved O2 is transported to the disk by forced convection of the rotational movement of the electrode and reduced to H2O2 and H2O at the disk. The products then diffuse out to the outer Pt ring, where an appropriate potential is applied to selectively oxidize H2O2 (1.2\u20131.4\u00a0V versus reversible hydrogen electrode [RHE]). The H2O2 selectivity calculated according to the following equation: \n\n(Equation\u00a09)\n\n\n\nH\n2\n\n\nO\n2\n\n\nSelectivity\n\n\n(\n%\n)\n\n=\n\n200\n\n1\n\n+\n\n\n\nN\n\n\u00d7\n\n\ni\nd\n\n\n\ni\nr\n\n\n\n\n\n\n\nwhere i\nd, i\nr, and N are the disk current, ring current, and collection efficiency, respectively. Although both the H2O2 selectivity and H2O2 Faradaic efficiency (|i\nr/N\u00d7i\nd|) represent how the catalysts selectively convert O2 to H2O2, the former is always higher.\n92\n The collection efficiency relates the ratio of the produced amount on the disk to the detected amount on the ring. This value depends primarily on the dimensions of the disk and ring electrodes. Although the N value is given by the manufacturer, it must be determined experimentally, as it can be influenced by the measurement conditions.\n93\n The collection efficiency was measured by the fast redox reaction of [Fe(CN)6]3\u2212/4\u2212 while rotating the electrode.Measurements are usually conducted in an inert gas-saturated electrolyte with 2\u00a0mM K3[Fe(CN)6] using chronoamperometry. During the measurement, the disk and ring overpotentials should be sufficiently large to guarantee diffusion-limited conditions. For every measurement, it is recommended to use the same experimental conditions, including electrolyte, electrode rotation speed, and catalyst loading. In particular, the catalyst loading on the disk electrode critically affects both the collection efficiency and H2O2 selectivity and thus can be a source of error. Ideally, a larger amount of catalyst contains a larger number of active sites, which should increase the apparent activity, while the selectivity remains constant. However, the higher catalyst loading results in a thicker catalyst layer in which the product is trapped, inhibiting diffusion to the ring electrode. This trapping effect often underestimates the collection efficiency and H2O2 selectivity. In addition, the trapped H2O2 has a higher chance of undergoing further reduction to H2O in the presence of H2O2 reduction sites (2e\u2212\n\n\n\u00d7\n\n 2e\u2212 ORR pathway).\n94\n On the contrary, insufficient catalyst loading may lead to a substrate effect where the apparent activity mainly originates from the substrate itself (detailed discussion presented below). In this regard, exploration of the loading effect of the catalysts and optimization the catalyst loading are important.To demonstrate the effect of the catalyst loading, we measured the collection efficiency of a commercial RRDE with our previously developed catalyst (O-GOMC) at loadings of 0, 50, 100, 300, and 600\u00a0\u03bcg cm\u22122. The collection efficiency (value provided by the manufacturer of 37%) decreased from 40% without the catalyst to 30% with the catalyst loading of 600\u00a0\u03bcg cm\u22122 (Figure\u00a08C). We then measured the loading-dependent ORR activity and selectivity in 0.1\u00a0M KOH. As the loading increased, the onset potential shifted positively because of the enhanced reaction kinetics with a larger number of active sites (Figure\u00a08D). However, owing to the trapping effect, the ring current decreases with increasing catalyst loading. This resulted in a lower H2O2 selectivity at higher catalyst loadings when a fixed N value (0.37) was used. This underestimation (or overestimation) can be corrected by applying a loading-dependent collection efficiency in the calculation of H2O2 selectivity. After the correction, the loading-dependent selectivity difference decreased substantially, yet discrepancy remained at low potentials (Figure\u00a08E). This phenomenon can be attributed to the mixed reaction pathways, including H2O2 chemical decomposition in the thick catalyst layer, electrochemical reduction of produced H2O2 by multitudinous catalytic sites (2e\u2212\u00a0\u00d7\u00a02e\u2212 pathway) in high catalyst loading, and complex mass-transport behavior. Hence, for H2O2 electrosynthesis selectivity assessments using the RRDE, (1) the collection efficiency should be determined under the same experimental conditions where the ORR measurements are conducted, and (2) a catalyst loading below 100\u00a0\u03bcg cm\u22122 is preferable.In addition, H2O2 oxidation at the ring should be measured in a diffusion-controlled manner (i.e., fast kinetics), for which the Pt ring should be physically and electrochemically cleaned before every measurement. Physical cleaning can be performed by traditional polishing with an alumina suspension, and the residual alumina particles on the electrode are removed by ultrasonication. The Pt surface was electrochemically cleaned via repeated potential cycling. Although this is typically carried out for the measurement of Pt-based electrocatalysts, the effect of Pt-ring cleaning on H2O2 selectivity has not been investigated. Herein, we propose a Pt-ring-cleaning protocol, which consists of applying potential cycles on the Pt ring in an N2-saturated electrolyte between 0.05\u20131.20\u00a0V (versus RHE) at a scan rate of 500\u00a0mV s\u22121. Generally, the hydrogen underpotential deposition and desorption peaks are restored during cycling (Figures\u00a08F\u22128H). We also observed that the Pt ring was severely contaminated when a typical alcoholic solvent, such as isopropanol or ethanol, was used to wipe the RRDE before the electrochemical cleaning step. This effect can be attributed to the adsorption of the solvent molecule and/or its decomposition products from catalysis by Pt. To optimize the Pt-ring-cleaning protocol, the number of cycles was first controlled, and the H2O2 selectivity after 10 potential cycles was compared to that without cycling (Figure 8I). However, excessive cycling led to lower H2O2 selectivity because the prolonged potential cycles caused Pt dissolution into the electrolyte, and trace Pt ions were redeposited on the catalyst during the ORR measurement. The redeposited Pt acted as an efficient 4e\u2212-ORR catalyst and decreased the H2O2 selectivity. To remove the dissolved Pt ions, the electrolyte was replaced with a fresh electrolyte after 50 potential cycles. An electrochemically cleaned Pt ring and a Pt-free electrolyte led to the highest H2O2 selectivity. In the same context, after cyclic voltammetry of the Pt ring, the ORR measurement should begin preferably within 10\u00a0min. Pt is spontaneously oxidized in an aqueous electrolyte, and a passivation layer is formed, which deteriorates the H2O2 oxidation kinetics of the Pt ring. As the Pt-cleaning step is followed by O2 purging of the system, at least 3\u00a0min of sparging time is necessary for O2 saturation. We tested the effect of Pt-ring passivation times of 3, 5, and 10\u00a0min after cleaning. A considerable decrease in H2O2 selectivity was observed after 10\u00a0min, whereas similar selectivities were observed after 3 and 5\u00a0min (Figure\u00a08I). Overall, the best way to evaluate the H2O2 electrosynthesis activity is to (1) determine the collection efficiency using a low amount of catalyst-loaded RRDE in the desired electrolyte and (2) electrochemically clean the Pt ring before the ORR measurements, and O2 bubbling time should be limited up to 5\u00a0min after the Pt-ring cleaning. Finally, obtaining and reporting the H2O2 selectivity data can be supplemented by Koutecky\u2013Levich analysis, which requires rotation-speed-dependent LSV measurements.Another issue concerns the evaluation of the reaction kinetics. The onset potential of an active 2e\u2212 ORR catalyst is usually above the standard equilibrium potential of 0.70 V. Such a phenomenon originates from the absence of H2O2 in the electrolyte until right before the 2e\u2212 ORR initiates, increasing the equilibrium potential at around 0.8 V, according to the Nernst equation. This situation may cause a fluctuation or drift of the equilibrium potential owing to the change in the local concentration of H2O2 produced in situ during electrocatalysis. We assumed that such changes could inhibit the appropriate selection of a linear region in Tafel analysis, which is typically used to extract kinetic information. Therefore, to unravel the impact of the in-situ-generated H2O2, the ORR measurements were conducted in the electrolyte with a controlled concentration of H2O2 (0\u20130.1 M) using the O-GOMC-5.5 catalyst. Following the Nernst equation, the LSV curves were negatively shifted as the H2O2 concentration increased (Figure\u00a08J). The corresponding Tafel plots (Figure\u00a08K) reveal that the selection of the linear region below the kinetic current of 0.05 mA is not appropriate, as the Tafel slopes change significantly and randomly with respect to the H2O2 concentration. Instead, selecting the region above the kinetic current of 0.1 mA is more reasonable in a kinetic sense because the Tafel slopes change monotonically with H2O2 concentration in this range. The Tafel slope of this catalyst was 73\u00a0mV dec\u22121 without H2O2 and increased to 110\u00a0mV dec\u22121 with 0.1\u00a0M H2O2. If the LSV curve measured without H2O2 is available, the apparently linear region between 0.78\u20130.84\u00a0V (versus RHE) may be selected, which results in the Tafel slope of 40\u00a0mV dec\u22121 and possibly leads to wrong conclusions.Finally, when reporting the 2e\u2212 ORR activity, the substrate activity should also be considered. This is particularly important when testing electrocatalysts using carbon-based substrates, mostly glassy carbon (GC) electrodes, in alkaline media. The GC electrode exhibited a significant 2e\u2212 ORR activity and selectivity. In addition, it becomes electrochemically oxidized during multiple uses, resulting in the introduction of O species in the GC that can promote catalytic activity. Figures\u00a08D and 8E present the H2O2 electrosynthesis activity of a polished GC electrode (denoted as catalyst loading \u201c0\u201d) measured in 0.1\u00a0M KOH. Interestingly, the GC exhibited a considerably high onset potential (0.62\u00a0V versus RHE) and a maximum H2O2 selectivity of 92%.Bulk electrolysis is a more accurate method for assessing the H2O2 electrosynthesis activity of a catalyst that can quantify accumulated H2O2 in the electrolyte and respective Faradaic efficiency (FE). For H2O2 quantification, the generated H2O2 must not be consumed or decomposed during the bulk electrolysis. Therefore, the counter electrode, where an arbitrary oxidation potential is applied, must be separated from the working electrode; otherwise, the produced H2O2 is readily oxidized. An H cell is suitable for this purpose as its cell design separates the working and counter electrode chambers using an ion-exchange membrane (usually Nafion). In addition, the RRDE with a Pt ring is not suitable for H-cell experiments because Pt can act as a very efficient catalyst for chemical H2O2 decomposition. Because FE is calculated based on the ratio of the produced amount of H2O2 to the theoretical production, it is crucial to accurately measure the produced H2O2 accumulated in the electrolyte.Three major methods were used to quantify the accumulated H2O2: ultraviolet-visible (UV-vis) spectrophotometry, titration, and colorimetric test strips.\n95\n\n,\n\n96\n In the UV-vis spectrophotometry, the Beer-Lambert law describing the linear dependence of the absorbance of a target material at a specific wavelength with the concentration is used. Because H2O2 barely exhibits its color, chemicals such as cobalt(II) carbonate, I\u2212, and Ti4+ are added to react with H2O2 to exhibit strong UV-vis absorbance. In the cobalt(II) carbonate method, Co2+ ions react with H2O2 in solution, generating a Co3+ carbonate complex that absorbs a wavelength of 260\u00a0nm. Iodometric analysis relies on the oxidation of I\u2212 by the generated H2O2, and the resulting I3\n\u2212 shows absorbance at 351\u00a0nm. In the titanium method, a peroxotitanium complex is formed, which absorbs a wavelength of 400\u00a0nm under acidic conditions. Spectroscopic H2O2 quantification has the advantages of a low detection limit of 50\u2013100\u00a0\u03bcg/L and high reliability. Hence, detection can be achieved in a short time of measurement: it takes approximately 1\u00a0min at a fixed current of 1 mA with an electrolyte volume of 100\u00a0mL (assuming 100% H2O2 FE).Second, redox titration is a simple method, as it requires neither spectroscopic devices nor calibration curves. However, its measurements take more time because of a higher detection limit (0.1\u20131 wt%). Similar to the spectroscopic method, titration requires compounds that change their colors by the redox reaction with H2O2, such as I\u2212, MnO4\n\u2212, and Ce4+. Iodometric titration has similar chemical principles to the above-mentioned iodometric analysis, but starch must be added as an indicator that dramatically changes color at the endpoint. In permanganate titration, MnO4\n\u2212 is reduced by H2O2, losing its characteristic purple color and thus making the use of an indicator unnecessary. However, because KMnO4 is not a primary standard, additional standardization of the KMnO4 titrant is required for this method. Cerium titration was performed using the reduction reaction of Ce4+ by the generated H2O2, forming colored Ce3+. The reliability of this method can only be guaranteed under acidic conditions and at temperatures below 10\u00b0C.Finally, colorimetric strip tests were performed by dipping the test strip into an H2O2 solution of arbitrary concentration, drying, and checking the color. Although this method is fast, simple, and intuitive, the detection limit is very high, the error is very large, and the reliability of data decreases as time passes after the measurement.There are several points to be considered in H2O2 quantification experiments. First, O2 dissolved in the electrolyte should be removed before starting the quantification. Residual O2 may react with the reagents used in the titration reaction, which leads to an inaccurate determination of the H2O2 concentration. The second is related to the selection and amount of H2O2 stabilization agents added to the electrolyte. H2O2 is an unstable compound that must be often stabilized during bulk electrolysis where the measurement is conducted for a long time, particularly in high-pH electrolytes. Notably, the additive itself undergoes or initiates redox reactions within a certain potential range. Hence, the stability and potential effects of the additive should be checked before its use in H2O2 quantification.In the development of new catalysts, the above-mentioned measurements using the RRDE are extremely useful because they allow the rapid screening of candidate materials. However, the RRDE can output a maximum current density of only a few mA\u00a0cm\u22122, which corresponds to an H2O2 generation rate of a few mg h\u22121 per 1\u00a0cm2. This low production rate is not industrially viable. Furthermore, the rotation system or any agitation tool for convection increases the volume of the device and production cost. A high current density of up to a few tens of mA cm\u22122 can be obtained using a porous carbon-paper-based electrode in an H cell. The O2 diffusion and transport behavior at these electrodes can be improved by controlling the internal pore structure and hydrophobicity. However, the production rate remains far from acceptable ranges for industrial applications. This limitation originates from the low concentration (1\u00a0mM) and low diffusion coefficient of aqueous O2.Therefore, the design of an electrochemical reactor that performs high-rate conversion with gaseous reactants is necessary to achieve current densities over hundreds of mA cm\u22122. Yamanaka and co-workers demonstrated an electrochemical H2O2 synthesis reactor in which a membrane-electrode assembly (MEA), consisting of a 2e\u2212 ORR cathode, an oxygen evolution anode, and a solid polymer electrolyte (SPE; Nafion in this case), was used to separate the cathode and anode chambers filled with neutral electrolytes.\n97\n Electrolysis performed using aqueous O2 as the reactant was\u00a0inefficient for H2O2 production with low rates and current efficiencies below 10\u00a0\u03bcmol h\u22121 and 2%, respectively (Figure\u00a09A). In contrast, when the cathode chamber was half filled with the electrolyte, gaseous O2 was supplied from the exposed part of the electrode, allowing much faster H2O2 generation of 132\u00a0\u03bcmol h\u22121 (Figure\u00a09B).Because similar issues are found in the area of electrochemical CO2 reduction, intensive efforts have been made to develop high-current-density reactors. For CO2 reduction, pioneering work by Kenis and co-workers demonstrated a membrane-free flow reactor (microfluidic reactor), where the gas chamber and catholyte chamber were separated by a catalyst-coated carbon-paper electrode (or gas-diffusion electrode [GDE]).\n100\n The Jaramillo group utilized this reactor for H2O2 electrolysis, where O2 gas was supplied from the backside of the electrode and reacted at the gas-catalyst-electrolyte boundary (Figure\u00a09C).\n98\n As this system allows the reaction of gaseous O2, the limitations regarding the low concentration and diffusion coefficient of O2 can be overcome. In addition, as the catholyte flow system inhibits the transportation of the produced H2O2 to the anode, membraneless operation is possible, which can reduce the Ohmic loss. Using this reactor, a 50-mA operation was demonstrated at a total cell voltage of 1.6 V.Although the flow reactor enables high-current-density operation, the generated H2O2 is dissolved in the electrolyte, requiring a separation process to obtain pure H2O2. To address this problem, Wang and co-workers devised an electrolyte-free reactor using a porous solid electrolyte that enabled H2O2 electrosynthesis in deionized water.\n99\n A typical MEA is composed of a cathode, membrane electrolyte, and anode, and these components are hot pressed for close contact. In porous solid-electrolyte design, the MEA consists of a cathode, anion-exchange membrane (AEM), porous solid electrolyte, cation-exchange membrane (CEM), and anode. The 2e\u2212 ORR at the cathode and hydrogen oxidation reaction at the anode produced HO2\n\u2212 and H+, respectively. The ions were then transported through the AEM and CEM to the porous solid electrolytes, in which deionized water flowed. Finally, the ions were dissolved in deionized water flowing through the solid electrolyte to produce a ready-to-use pure H2O2 solution with a maximum concentration of 20 wt % at a current density of 200 mA cm\u22122 (Figure\u00a09D).The efficiency of the flow reactor and MEA-type reactor critically depends on the number of gas-catalyst-electrolyte boundaries (or triple-phase boundaries) and the transport of reactants and products. Therefore, the architecture and surface properties of GDEs should be meticulously designed, and some operation variables, such as O2 and electrolyte flow rates, should be carefully controlled.The integration of high-current reactors and renewable energy sources will enable low-cost, on-site H2O2 production systems. Electrochemically produced H2O2 can be directly utilized in many areas, such as drinking-water cleaning in developing countries (Figure\u00a09E),\n98\n disinfection, and bleaching applications.\n35\n\n,\n\n50\n Wang and co-workers developed a highly selective Fe-CNT catalyst for H2O2 production, where an H2O2-containing solution was produced for 210\u00a0min of operation.\n35\n Despite its low concentration (1,613 ppm), the reaction solution could be directly used for bacterial disinfection without further purification (Figure\u00a09F). Chen and co-workers prepared a tested electrolyte containing 309\u00a0mM H2O2 for 3\u00a0h by applying a current density of 35.4 mA cm\u22122 with a high maximum FE of 91.9% in an H cell.\n50\n With traces of Fe2+ ions, H2O2 is decomposed to hydroxyl radicals by the Fenton reaction, which can successfully bleach various types of dye solutions with approximately 99% removal efficiency. These examples of the utilization of electrosynthesized H2O2 should be preceded by adequate accumulation of H2O2 in the electrolyte after a few hours of operation.Recently, Jang, Joo, and co-workers developed all-in-one devices that carry out bias-free photoelectrochemical production of H2O2 and its direct in situ utilization for chemical valorization. They first demonstrated a three-compartment photo-electro-biochemical reactor for lignin conversion (Figure\u00a09G).\n44\n The system was composed of three separate chambers, each containing a reduced TiO2 photoanode (H:TiO2), a Co\u2013N/C electrocatalyst-based cathode for H2O2 production, and a lignin peroxidase enzyme for lignin upgrading. When exposed to light, the H:TiO2 photocatalyst invoked photocatalytic oxygen evolution, and the generated electrons were transported to the Co\u2013N/C electrocatalyst, which received electrons to promote O2-to-H2O2 conversion. The produced H2O2 diffused into the adjacent enzyme chamber, enabling lignin peroxidase to catalyze lignin depolymerization with a conversion and selectivity of 93.7% and 98.7%, respectively. The authors expanded the unbiased H2O2 generation and utilization system to H2O2-involving heterogeneous catalysis (Figure\u00a09H).\n45\n They applied this reactor for propylene epoxidation to propylene oxide (PO), an important chemical in the plastic industry. A Co\u2013Pi/BiVO4 photoanode and a Co\u2013N/C cathode electrocatalyst were used for photoelectrochemical H2O2 production. Titanium silicate (TS-1) then catalyzed the propylene epoxidation reaction using in-situ-generated H2O2 as an oxidant with a high production rate (10.6\u00a0mmol h\u22121 for 5 h) and a PO selectivity of 97.6%, which was maintained for 24 h. It should be noted that the integrated photo-electro-heterogeneous catalytic system enabled propylene epoxidation under ambient conditions using only sunlight and O2. As photo-electrochemical H2O2 generation was successfully demonstrated over a wide pH range (2\u20138), the integrated system is envisioned to be applicable for other H2O2 utilization reactions.Electrochemical H2O2 production is a highly promising carbon-free technology that enables continuous on-site H2O2\u00a0production and has recently emerged as a promising alternative method to the current anthraquinone process. This review summarized recent design strategies for efficient electrocatalysts for the 2e\u2212\u00a0ORR, including PGM-based ADCs, non-PGM-based ADCs (M\u2013N/C), and metal-free doped carbons. Remarkable progress has been made with these catalysts over the last few years in terms of their activity and selectivity. However, several critical issues remain to be addressed. The origin of the activity of these catalysts remains largely elusive, as the chemical processes for doping of active moieties in such catalysts induce high heterogeneity of the active-site distribution in the resulting catalysts. Therefore, identification of the active-site structure and selective generation of desired active sites are essential for future progress. The role of the interfacial species should be clarified, which is hampered by the difficulty in characterizing the electrical double layer. With an improved understanding of the double-layer structure, the rational design of interfacial additives or co-catalysts can be facilitated.The summarized MA and site-normalized activity of reported catalysts suggest that future electrocatalyst research should focus on achieving low overpotentials in neutral and acidic electrolytes for practical applications of H2O2 electrosynthesis technology. This objective can be achieved by the discovery of new electrocatalysts and new insights into the pH-dependent reaction mechanism and double-layer structure. In the course of catalyst development, the ORR activity and selectivity of the developed catalysts from different laboratories should be compared on a standardized protocol for rapid advancement. In this regard, a measurement protocol for the accurate assessment of electrocatalytic properties should be established, particularly for H2O2 selectivity. Although RRDE is the most prevalent method for evaluating H2O2 selectivity, the measured value is affected not only by the intrinsic properties of the catalysts but also, to a significant degree, by the experimental conditions, including catalyst loading and cleanliness of the electrode. The critical impacts of each source of error are demonstrated in this review, which is envisioned as a measurement guideline for future research. We also emphasize that the efficiency of the developed catalysts should be tested by bulk electrolysis measurements using an H cell, which has more practical relevance.Industrial-scale reactors using GDE and MEA, which have been widely used in the CO2 electroreduction field, have recently been employed for high-rate H2O2 electrosynthesis. The acquired knowledge from previous studies should be considered to obtain a high conversion rate and long-term stability. At this stage, most of the high-current-density reactor research focuses on the optimization of the electrode architecture to maximize the triple-phase boundary and prevent electrode flooding, which deteriorates the activity. However, the effects of local pH must be studied, as it is one of the most important issues in CO2 conversion. Because the 2e\u2212 ORR is a proton-consuming process, the accumulation of OH\u2212 at the interface under the operating conditions may be utilized to enable efficient H2O2 electrosynthesis even using acidic and neutral electrolytes. Characterization of the interfacial pH and molecular dynamics simulations will aid in understanding such phenomena. Finally, the advanced design of the reactor can contribute to the economic and efficient utilization of electrosynthesized H2O2, which implies direct usage without further purification or concentration. This includes an electrolyte-free reactor producing a pure H2O2 solution and the integration of the H2O2 electrosynthesis reactor and utilization modules.This work was supported by the National Research Foundation (NRF) of Korea (NRF-2019M3E6A1064521, NRF-2019M3D1A1079306, NRF-2019M1A2A2065614, and NRF-2021R1A2C2007495 to S.H.J.; NRF-2020R1C1C1006766 to Y.J.S.).S.H.J. and Y.J.S. conceptualized this review. J.S.L. drafted the manuscript. Y.J.S. and S.H.J. revised the manuscript.The authors declare no competing interests.Supplemental information can be found online at https://doi.org/10.1016/j.xcrp.2022.100987.\n\n\nDocument S1. Note S1\n\n\n\n\n\nDocument S2. Article plus supplemental information\n\n\n\n", "descript": "\n                  The electrosynthesis of H2O2 via a two-electron pathway oxygen reduction reaction (2e\u2212 ORR) has emerged as a promising way of carbon-free and on-site production of H2O2. Active and selective electrocatalysts for the 2e\u2212 ORR are essential for achieving high O2-to-H2O2 conversion efficiency. In this review, we present the recent progress in the development of 2e\u2212 ORR electrocatalysts including Pt-group-metal (PGM) and non-PGM atomically dispersed catalysts and metal-free heteroatom-doped carbons. The impact of the active sites and interface structures on the electrocatalytic process is summarized. Benchmarking of the electrocatalytic activities in terms of O2-to-H2O2 mass activity and site-normalized activity is presented to understand the current status of advancement and\u00a0to provide an insight into possible future research directions. In addition, some guidelines and pitfalls in typical laboratory measurements for assessing 2e\u2212 ORR performance are proposed. Finally, recent advances in high-current-density H2O2 electrosynthesis reactors and devices that exploit electrosynthesized H2O2 are introduced.\n               "}
{"full_text": "Transition metal sulfide nanoparticles have been identified as a family of very promising earthabundant and low-cost electrocatalysts, which can rival those of metal selenides, nitrides, phosphides, and carbides nanoparticles [1\u20133]. Due to their electrochemical properties, metal sulfide that demonstrate a potential catalyst to substitute noble metals, are expected to reduce polluted 4-nitrophenol (4-NP) to eco-friendly 4-aminophenol (4-AP) [4\u20137]. In addition, incorporating heteroatoms into catalytic materials, the catalytic performance will be considerably enhanced by improving the electron transfer [8\u201310]. In contrast, synergistic effect brought by multiple heteroatoms (such as N-, P-, S-) co-doping can verify directly with the electronic structures and polarities of the catalysts [11,12].Metal organic frameworks (MOF) possess intramolecular pores which constituted by the self-assembly of metal ions and organic linkers [13\u201315]. Hu et al. prepared Cu-BMOF using two organic ligands by a hydrothermal method [16]. Based on the high affinity of N, S heteroatoms with noble metal ions, the two-ligand Cu-BMOF displayed auspicious adsorption capacity of 933\u00a0mg/g for removing Au (III) from aqueous solutions. Unfortunately, most of the MOF framework easily destroyed in acidic or basic media, and will be strictly limited to practical applications. Therefore, a large number of derivatives transformed from MOF are proven to be ideal materials, which not only retain some important advantages of MOF, but also significantly generate prominent feature [17\u201319]. Pan et al. synthesized porous carbon derivatives through pyrolyzing the two-ligand bimetallic MOF (CoxZny-JUC-160) [20]. Their derivatives possesses uniformly dispersed active sites, hierarchical porous structure, open pore network. Actually, the multiple heteroatom-doped (such as N-, P-, S-) materials could be synthesized by one-step controlling the well-designed two-ligand MOF [21].Due to self-assembly of metal ions and organic linkers, the mixed ligand MOFs provide a practical platform to in-situ synthesize multiple heteroatoms co-doping materials with enhanced catalytic performance. In this work, a known structural [M6(TDC)6(hmt)2(DMF)6(H2O)3] (JUC-85Ni) (M\u00a0=\u00a0Ni; TDC\u00a0=\u00a02,5-thiophenedicarboxylic acid; hmt\u00a0=\u00a0hexamethylenetetramine) is synthesized through a simple solvothermal method. The JUC-85Ni precursors are further investigated to synthesize metal sulfide nanoparticles dispersed N, S-codoped carbon materials (NiS\n\n2\n\n@NSC). The NiS2@NSC nanocomposites provide well-balanced N, S-codoped structure, good physicochemical stability and high loading of catalytic sites (Scheme 1\n). Herein, the route adopted two-ligand MOF to synthesize catalysts who rich in metal sulfide, N and S functional groups is convenient and effective. Thus, with accurate designing, the NiS2@NSC nanocomposites can be exploited as excellent catalytic center for hydrogenation reduction of 4-NP. Eventually, the proposal catalytic mechanism for NiS2@NSC is proposed, which is of great significance toward the future design of catalysts for 4-NP reduction.The final structure and crystallinity of the as-synthesized JUC-85Ni precursors and their derivatives are investigated by PXRD. We compare the typical peaks of JUC-85Ni with the simulated one, whose metal center is cadmium [22]. Fig. S1 indicates Ni- and Cr-based two-ligand MOF are isomorphic three-dimensional structure with loh1 topology. After solvothermal sufidation reaction with TAA at 120\u00a0\u00b0C for 4\u00a0h, JUC-85Ni precursor is converted to NiS2, N, S enriched porous carbon nanocomposites. In Fig. S2, two board peaks observed at 2\u03b8\u00a0=\u00a024o and 44o are indexed to porous carbon [23]. The obvious NiS2 peaks appear at 2\u03b8\u00a0=\u00a027.2\u00b0, 31.6\u00b0, 35.6\u00b0, 38.8\u00b0, 45.30\u00b0, 48.0, 53.6\u00b0, 56.2\u00b0, 61.2\u00b0 and 68.1\u00b0, corresponding to (111), (200), (210), (211), (220), (211), (311), (222), (321) and (410) plane of cubic vaesite NiS2 (PDF # 11-0099) [7]. Noticeably, the composition of product can be affected by the sulfidation reaction time. As shown in Fig. S3, the crystal structure of JUC-85Ni-2\u00a0h (The product of the reaction time for 2\u00a0h) is still maintained, while the decreased intensity indicate that the crystallinity change. In order to confirm the formation of NiS2 nanoparticles, sulfidation reaction time should be extended to 4\u00a0h.N2 adsorption/desorption isotherms show a type-III curve and a characteristic steep increase at high relative pressures, which demonstrate the sample has characteristic microporous structure (Fig. S4). The BET surface area of NiS2@NSC catalyst is 13.4\u00a0m2\u00b7g\u22121. The average pore size is 8.5\u00a0nm for NiS2@NSC, measured from the Density-Functional-Theory (DFT) method (Fig. S5).The morphology characterization of materials are explored by SEM and TEM at different magnification (Fig. 1\n, Fig. S6). The JUC-85Ni precursors appear as regular nanocube shape (Fig. S6a). However, the NiS2@NSC couldn't keep the original appearance, showing irregular shape (Fig. 1). SAED images (the upper right corner) indicate lattice fringes distance is 0.254\u00a0nm, which corresponds to the (210) plane of NiS2. It is noticed that the NiS2 nanoparticles are all wrapped by a thin layer of carbon, which are likely to avoid becoming large particles (Fig. S6b). The TEM elemental mapping images reveal carbon, nitrogen, oxygen, sulfur and nickel elements are uniformly distributed in NiS2@NSC (Fig. S7).The chemical composition and coordination of NiS2@NSC are analyzed by XPS instrument. The binding energies around 289, 402, 532, 164 and 860\u00a0eV in the survey spectra are corresponding to C, N, O, S and Ni, respectively (Fig. S7). The N 1\u00a0s XPS spectra show thtee peaks at 399.6, 400.5, 401.7\u00a0eV, suggesting the existence of pyridinic-N, pyrrolic-N, graphitic-N (Fig. 2a). For S 2p XPS spectra, the two peaks (163.1 and 164.3\u00a0eV) are assign to Metal-S 2p 3/2 and Metal-S 2p 1/2, and match well with XRD data (Fig. 2b). The peaks at 161.0 and 162.0\u00a0eV could be attributed to the bond of C-S-C, indicating the presence of S in the carbon matrix [11,24]. The observed peaks at 168.6 and 169.8\u00a0eV corresponds to sulfate species from unavoidable oxidation [5]. Four sub-peaks are observed in Ni 2p XPS spectra, wherein the peaks at 853.2\u00a0eV (Ni 2p 3/2) and 870.4\u00a0eV (Ni 2p 1/2) assign to Ni2+, and the fitting peaks at 857.1\u00a0eV (Ni 2p 3/2) and 875.2\u00a0eV (Ni 2p 1/2) are Ni3+ (Fig. 2c) [25]. Added to this, there are other two satellite peaks (identified as \u201cSat.\u201d) located at 861.5 and 880.4\u00a0eV [26]. The C 1\u00a0s, O1s XPS spectra are also analyzed in Fig. S8.Considering its simplify operation and mild experimental condition, the catalytic activity of NiS2@NSC are evaluated by the reduction of 4-NP to 4-AP in NaBH4 aqueous solution [5,6]. The UV\u2013vis absorbance spectroscopy is used to monitor the change of 4-NP during the reaction. With the adding of NaBH4, the 4-NP transform to 4-nitrophenolate ions, demonstrating by absorption peak shift from 317 to 400\u00a0nm. When a small amounts of NiS2@NSC catalyst is added into the reaction solution, a peak of 4-AP appear at around 300\u00a0nm, which indicate that 4-nitrophenolate ions are reduced. When NiS2@NSC catalysts are used to reduce 4-NP, the UV\u2013vis absorption spectra varied with time between 10\u00a0\u00b0C and 40\u00a0\u00b0C are showed in Fig. S9a-d. The catalyst behave excellent catalytic property at 40\u00a0\u00b0C with a reduction time of 45\u00a0s. However, the reduction time will be higher than 60\u00a0min or more without a catalyst (Fig. S10).To investigate the reduction of 4-NP to 4-AP is mainly through degradation, the 4-NP adsorption test is measured (Fig. S11) [27]. The 4-NP removal rates is only 4.6% at 5\u00a0min, which has low adsorption potential because of a small surface area. Compared to the catalytic reaction time of 225\u00a0s, the adsorption process plays an unimportant role in removal of 4-NP.Fig. S9e-f present the plot of C\nt/C\n0 against reduction time (reduction time is \u201ct\u201d, C\nt is the corresponding concentration at \u201ct\u201d, C\n0 is initial concentration of 4-NP) using NiS2@NSC catalyst under varying temperatures from 10\u00a0\u00b0C to 30\u00a0\u00b0C. The value of C\nt/C\n0 decline quickly with reduction time, revealing the 4-NP conversion increases rapidly. As the excess concentrated NaBH4, pseudo-1st order kinetic is assumed to determine the apparent rate constant (k\napp\u00a0=\u00a0\u2212ln(C\nt/C\n0)/t). The k\napp acquired from linear correlation are calculated to be 0.0201, 0.0355 and 0.0550\u00a0s\u22121 at 10\u00a0\u00b0C, 20\u00a0\u00b0C and 30\u00a0\u00b0C, and summarized in the Table S1. Additionally, as temperature increases, it could be observed that the k\napp increases. According to our statistics (Table 1\n), the NiS2@NSC catalyst possess the higher kapp than other published, indicating that the catalytic activity is superior to noble metal, transition metal and metal sulfide [5,6,8,19,28\u201330]. These k\napp values are further adopted to calculate thermodynamic parameters [31].The turn over frequency (TOF) of NiS2@NSC catalyst catalyst can better indicates the catalytic performance compared with k\napp, which can be calculated by the mass of 4-NP reduced per mass of NiS2 per reaction time. The catalyst possess uniformly heteroatom-doped structure, small size and high catalytic sites, the interfacial electron will transfers faster between the catalyst and 4-NP, causing TOF value of 188.6\u00a0h\u22121 for NiS2@NSC catalyst at 40\u00a0\u00b0C.The thermodynamic data are conducive to deeply understand the catalytic reaction pathway. Activation energy (Ea), enthalpy (\u0394H), entropy (\u0394S) and gibbs free energy (\u0394G) can be obtained from following equations.Arrhenius Eq. (1) reflects a relationship between T and k\napp, applying to calculate activation energy (Ea):\n\n(1)\n\nln\n\n\n\nk\napp\n\n\n\n=\n\n\n\n\u2212\n\nEa\nR\n\n\n\n\n\n1\nT\n\n\n+\n\nln\n\n\nA\n\n\n\n\nWhere R is molar gas constant (8.314\u00a0J\u00b7K\u22121\u00b7mol\u22121). A straight line of lnk\napp\nversus 1/T is given in Fig. S12 (red lines), from which the Ea value is 35.74\u00a0kJ\u00a0mol\u22121 for the NiS2@NSC catalyst.Eyring Eq. (2) is used to calculate activation enthalpy (\u0394H) and entropy (\u0394S):\n\n(2)\n\nln\n\n\n\n\nk\napp\n\nT\n\n\n\n=\n\n\n\n\u2212\n\n\n\u0394\nH\n\nR\n\n\n\n\n\n1\nT\n\n\n+\n\nln\n\n\n\n\nK\nB\n\nh\n\n\n\n+\n\n\n\n\u0394\nS\n\nR\n\n\n\n\nGibbs free energy (\u0394G) is calculated by Eq. (3):\n\n(3)\n\n\u0394\nG\n\n=\n\n\u0394\nH\n\n\u2212\n\nT\n\u0394\nS\n\n\n\nThe constant applied in calculation: K\nB is boltzmann constant (1.381\u00a0\u00d7\u00a010\u221223\u00a0J\u00b7K\u22121), h is Planck constant (6.626\u00a0\u00d7\u00a010\u221234\u00a0J\u00b7k\u22121\u00b7mol\u22121), T is absolute temperature.Fig. S12 (blue lines) reveal fitted straight lines of ln(k\napp/T) versus 1/T, and \u0394H and \u0394S value for the 4-NP reduction can be obtained from slope and intercept of the lines. The positive \u0394H value catalyzed by NiS2@NSC catalyst are determined to be 33.40\u00a0kJ\u00b7mol\u22121, showing the endothermic nature of the catalytic reduction process. A negative \u0394S value corresponding to NiS2@NSC catalyst is calculated to be \u2212158.8\u00a0kJ\u00a0mol\u22121\u00a0K\u22121, suggesting a disordered system. \u0394G values for 4-NP reduction exhibits equivalent values of 83.10\u00a0kJ\u00b7mol\u22121 at 40\u00a0\u00b0C in case of NiS2@NSC. The positive \u0394G reflects the reduction of 4-NP is not feasible and spontaneous without a catalyst.In terms of above analysis results and reference, the proposal mechanism for the 4-NP reduction reaction catalyzed by our catalysts is proposed in Scheme 2\n. According to Langmuir\u2013Hinshelwood model (L-H model), BH4\n\u2212 and 4-nitrophenolate ions are absorbed on the surface of porous N,S-codoped carbon layer where is enriched with electrons [32,33]. NiS2 nanoparticles should be active of BH bond to produce surface hydrogen species (NiH species). On the other hand, the e\u2212 are transferred from BH4\n\u2212 to the Ni atom. Then NO bond of 4-nitrophenolate ions will be broken by active hydrogen species to form NH bond [34], finally generating the reduced 4-AP. Simultaneously, the 4-NP accept the e\u2212 to reach the charge conservation. The L-H model regard the step of 4-nitrophenolate reacted with active hydrogen species as rate determining step due to the relatively slow speeds [31]. Eventually, the product 4-AP desorb from carbon matrix into the solution, and catalyst is ready for next cycle. The NiS2@NSC catalyst exhibit excellent catalytic activity, which cloud be summarized as follows: (I) synergetic effect between NiS2 and N, S-doped carbon facilitate the electron relaying to establish electron-enhanced areas, and more 4-NP preferentially adsorb on. (II) the dispersed and high amounts of NiS2 catalytic sites are benificial to generate active hydrogen species. (III) relatively higher porosity thin N,S-doped carbon layer support express entry to 4-NP interaction and the 4-AP release.The recycling performance is a factor to evaluate its applicability as important as catalytic activity. The reusability of the NiS2@NSC catalyst can be performed in five repeated experiments. Fig. S13 shows a catalytic efficiency of the NiS2@NSC catalyst as high as 93.0% after fifth cycles. Additionally, the TEM image show no aggregation of the used NiS2@NSC catalyst in comparison with those of the newly prepared one (Fig. S14). The BET surface area of uesd NiS2@NSC catalyst is 14.7\u00a0m2\u00b7g\u22121 (Fig. S15). Moreover, no obvious difference can be viewed from the XPS spectra of the used and newly prepared catalysts (Fig. S16), demonstrating a good recyclability and stability. By virtue of spatially separation of NiS2 nanoparticles in the thin NSC layer, the NiS2@NSC catalyst exhibit not only the excellent catalytic activity but also high stability, making it a promising candidate for the reduction of nitroaromatic compounds.In our previous work, a synthesized NiS nanocatalyst exhibit catalytic activity toward 4-NP hydrogenation as well as methyl orange (MO) hydrogenation [5]. Wang et al. prepared Co@NC catalyst for 4-NP and dye (RhB, MB) catalytic reduction in aqueous solution [19]. To verify the universality of Co@NC catalytic performance, simulated experiment for purification of the 4-NP and MO containing water pollutants is done according to the relevant literature [35], 50mL of 4-NP (0.12mM), MO (0.02\u00a0mM) and NaBH4 (12\u00a0mM) mixed solution was stirred, and then 0.5mL of NiS2@NSC catalyst suspension (4\u00a0mg/mL) was added into the above solution. Afterwards, the mixture was quickly filtered through an injector with nylon syringe filter. Fig. S17 demonstrate the absorbance peak of 4-NP and MO was absent, indicating NiS2@NSC catalyst can be hopful regarded as an efficient catalyst for 4-NP reduction and dye.As has been stated, we employee S-containing ligands and N-containing ligands to synthesize N, S co-doping mixed-ligand metal organic framework (JUC-85Ni). Next, the NiS2@NSC catalyst was synthesized by sulfidation of JUC-85Ni precursors with thioacetamide (TAA). When evaluating as a catalyst for 4-NP reduction, the excellent characteristics endows NiS2@NSC catalyst outstanding catalytic activities with a apparent rate constant of 0.0550\u00a0s\u22121 at 30\u00a0\u00b0C and a catalytic efficiency of 93.0%. According to Langmuir\u2013Hinshelwood model, the proposal catalytic mechanism of NiS2@NSC catalyst has been proposed in our paper. We hoped these high-effective catalysts will demonstrate universal application prospects in removing pollutants from wastewater.\nGuozhu Zhang: Investigation, Data curation, Visualization, Formal analysis, Writing \u2013 original draft. Yuhe Wang: Investigation, Validation, Formal analysis. Fei He: Investigation, Validation. Lixin He: Investigation. Haixia Li: Writing \u2013 review & editing, Funding acquisition. Dan Xu: Conceptualization, Methodology, Visualization, Resources, Writing \u2013 original draft, Writing \u2013 review & editing, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was financially supported by the Hainan Provincial Natural Science Foundation of China (219QN219), the Finance Science and Technology Project of Hainan Province (220RC618). Additionally, we thank the assistance from Sub-center of The Environment and Plant Protection Institute, CATAS Precision instruments Sharing Center.\n\n\n\nSupplementary material: Scheme 1. Schematic illustration for the synthesis of JUC-85Ni precursors and NiS2@NSC.\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106454.", "descript": "\n                  Herein, two-ligand nickel-based MOF (JUC-85Ni) have been prepared by choosing high N-containing and S-containing ligand. On this basis of the research, the NiS2 nanoparticles dispersed N and S co-doped carbon materials (NiS2@NSC) are fabricated through the sulfidation of JUC-85Ni precursors with thioacetamide (TAA). The rationally designed NiS2@NSC catalysts are conducted a test for catalytically reducing 4-nitrophenol (4-NP) to environment friendly 4-aminophenol (4-AP) under NaBH4 alkaline condition. The superior features endows the catalyst with outstanding catalytic activity and recycling performance, and support it great potential for wastewater treatment.\n               "}
{"full_text": "Ammonia (NH3) is a crucial chemical feedstock for the production of fertilizers, pesticides and many other chemicals. Moreover, due to its unique characteristics of a high energy density, clean combustion, and convenience for storage and transportation, NH3 is also regarded as a promising alternative clean and sustainable energy storage carrier in the future [1]. Currently, the conventional method for industrial-scale NH3 synthesis, namely the Haber-Bosch (HB) process, requires high pressures (20\u201340\u00a0MPa) and high temperatures (400\u2013600\u00a0\u00b0C) in the presence of an Fe-based catalyst. As a result, the HB process consumes \u223c2% of the global primary energy supply and produces \u223c300 million tons of CO2 per year [2,3]. Great efforts have been devoted to developing greener and more sustainable alternatives for ammonia production at lower pressures and temperatures, including biochemical and electrochemical processes. More recently, the potential applications of decentralized ammonia (NH3) production using green hydrogen on small scales have attracted increasing interest since the process could be driven by renewable energy sources such as wind and solar power [4,5].Non-thermal plasma (NTP) is regarded as a promising and emerging technology for NH3 production from N2 and H2 at low temperatures and ambient pressure. Plasma processes can be switched on and off instantly due to fast plasma-chemical reactions, thus offering great flexibility that can be coupled with renewable energy sources especially intermittent renewable energy for decentralized NH3 production. Different types of NTP have been investigated for plasma synthesis of NH3, including dielectric barrier discharge (DBD), microwave plasma, radio-frequency plasma, etc. DBD plasma has attracted great attention in NH3 synthesis due to its system compactness and scalability, mild operation conditions and simplicity of plasma-catalyst integration, which could generate a plasma-catalytic synergy to greatly improve the performance of chemical reactions [6\u20138]. The enhancement mechanisms of plasma-catalysis have been attributed to the synergistic interactions between plasma and catalysts, as the packed catalysts could affect the discharge characteristics, which in turn change the chemical reactions and alter the reaction kinetics.It is widely recognized that catalysts and surface reactions play an important role in determining the reaction performance of plasma-catalytic chemical processes including CO2 conversion, oxidation of volatile organic compounds (VOCs) and NOx abatement [9\u201311]. More recently, various catalysts have been investigated in plasma-catalytic NH3 synthesis. Shah et\u00a0al. investigated the effect of 11 transition metals and low-melting-point metals on the plasma NH3 production. Ni, Sn and Au showed superior energy efficiency in the plasma-catalytic NH3 synthesis [12]. Wang et\u00a0al. studied the effect of transition metal catalysts (M/Al2O3, M\u00a0=\u00a0Cu, Ni and Fe) on the plasma-catalytic NH3 synthesis. Ni/Al2O3 showed outstanding catalytic activity on the NH3 production, with a 15.2% higher NH3 synthesis rate than that using Fe/Al2O3 [2]. In addition to transition metal-based catalysts, noble metal (e.g., Ru)-based catalysts have also attracted much attention in plasma-catalytic NH3 synthesis due to their high activity. Patil et\u00a0al. evaluated the effect of a wide range of supported metal catalysts on the plasma-catalytic NH3 synthesis and found that the activity of these catalysts followed the order of Ru\u00a0>\u00a0Rh\u00a0>\u00a0Ni\u00a0>\u00a0Co\u00a0>\u00a0Fe\u00a0>\u00a0Pd\u00a0\u226b\u00a0Mo [21]. Kim et\u00a0al. studied the plasma-catalytic NH3 synthesis over promoted Ru/AC catalysts and reported the effect of metal promoters followed the order of Mg\u00a0>\u00a0K\u00a0>\u00a0Cs\u00a0>\u00a0no promoter [13]. Mizushima et\u00a0al. reported the use of Ru, Pt, Ni and Fe as the catalytic active phase improved the NH3 yield by 40\u2013100% in a plasma reactor, while the presence of Ru showed the highest NH3 yield [14]. However, most previous studies focused on improving NH3 production by changing the active metals of the catalysts, while far less research has been conducted to understand the effect of different supports on the plasma-catalytic NH3 synthesis. Xie et\u00a0al. reported that using L-MgO supported Ru catalysts one can reach an NH3 synthesis rate of 1.04\u00a0g s\u22121, \u223c8% higher than that using Ru/Al2O3 at a relatively low temperature (300\u00a0\u00b0C) [15]. Gorky et\u00a0al. found the presence of zeolitic imidazolate frameworks of ZIF-8 and ZIF-67 supports significantly enhanced the NH3 synthesis rate by 30\u201360% compared to Beta, 5A and SAPO zeolites [16]. However, the fundamental understanding of different catalyst supports and their physicochemical properties on the plasma-catalytic NH3 synthesis is still limited.In this work, we have investigated NH3 synthesis from N2 and H2 over supported Ru catalysts in a co-axial DBD plasma reactor. Activated carbon (AC), \u03b1-Al2O3, ZSM-5 and SiO2 are chosen as the supports for Ru catalysts. The effect of these supports on the discharge characteristics, NH3 concentration and energy yield of the process was investigated at different operating conditions. Catalyst characterization including N2 adsorption-desorption, X-ray diffraction (XRD) and temperature-programmed desorption of CO2 (CO2-TPD) was performed to understand the structure-performance relationship between the catalysts and NH3 synthesis. The key reaction performance of the plasma-catalytic synthesis of NH3 in this work is compared with the results reported in the literature.In this work, an ultrasonic-enhanced wet impregnation method was used for the preparation of the Ru-based catalysts. All chemicals were of analytical grade. To prepare the Ru-based catalysts, a weighed amount of RuCl3\u22193H2O was firstly dissolved in deionized water and magnetically stirred for 1\u00a0h to form a transparent solution. Then, a desired amount of the support was added into the solution and treated by ultrasonication for 3\u00a0h at room temperature. After that, the mixture was heated and vigorously stirred in a water bath (80\u00a0\u00b0C) for 3\u00a0h, followed by drying in an oven at 110\u00a0\u00b0C for 12\u00a0h. The samples were then calcined in a nitrogen gas stream at 500\u00a0\u00b0C for 5\u00a0h, then crushed and sieved to 40\u201360 meshes. The obtained samples were reduced in a 5\u00a0vol.% H2/Ar gas stream at a total gas flow rate of 100\u00a0mL\u00a0min\u22121 at 500\u00a0\u00b0C for 5\u00a0h. The catalysts are denoted as Ru/M where M is the catalyst support (M\u00a0=\u00a0AC, \u03b1-Al2O3, ZSM-5 and SiO2). The loading amount of Ru was 1\u00a0wt.% in this work.N2 adsorption-desorption experiments were conducted at 77\u00a0K to obtain the textural properties of the Ru/M catalysts (M\u00a0=\u00a0AC, \u03b1-Al2O3, ZSM-5 and SiO2) using a Micromeritics ASAP 2010 instrument. Each sample was degassed at 200\u00a0\u00b0C for 5\u00a0h before the measurement. The specific surface area (SBET) and pore size of the Ru/M catalysts were obtained using the Brunauer-Emmett-Teller (BET) equation, while the average pore diameter and pore volume of the samples were calculated based on the Barret-Joyner-Hallender (BJH) method. The XRD patterns of the Ru/M catalysts were obtained using a Rikagu D/max-2000 X-ray diffractometer with a Cu-K\u03b1 radiation source. All samples were scanned in the 2\u03b8 range of 10\u00b0\u201380\u00b0 with a step size of 0.02\u00b0. The basicity of the Ru/M samples was measured by CO2-TPD. During the test, each catalyst (100\u00a0mg) was pre-treated and degassed at 250\u00a0\u00b0C in an Ar flow for 1\u00a0h before being cooled down to 50\u00a0\u00b0C. The sample was then saturated with 5\u00a0vol.% CO2/Ar at a flow rate of 40\u00a0mL min\u22121 for 1\u00a0h, followed by purging with pure Ar at a flow rate of 40\u00a0mL min\u22121 to remove any weakly adsorbed CO2. Finally, the TPD measurement was performed by heating the sample from 50\u00a0\u00b0C to 800\u00a0\u00b0C at a heating rate of 10\u00a0\u00b0C min\u22121 in a pure Ar flow at a flow rate of 40\u00a0mL min\u22121. The CO2 desorption amount was determined by integrating the CO2-TPD profile.\nFig.\u00a01\n shows a schematic diagram of the experimental setup. The plasma reactor consisted of a quartz tube, two polytetrafluoroethylene (PTFE) seals and two electrodes. The quartz tube had an outer diameter of 10\u00a0mm with a wall thickness of 1\u00a0mm. The two PTFE seals were placed at both ends of the quartz tube. Aluminum foil with a length of 30\u00a0mm was wrapped tightly around the quartz tube and served as a ground electrode. A stainless steel (SS) rod (high voltage electrode) with a diameter of 4\u00a0mm was placed in the center of the quartz tube and held by the PTFE seals. The DBD plasma reactor was connected to an AC high voltage power supply (CTP-2000K, Suman, China). The Ru/M catalyst (200\u00a0mg, 40\u201360 meshes) was packed into the discharge region and held by glass wool in each test. The DBD reactor was fan-cooled during the experiments, and the temperature on the outer wall of the reactor was around 100\u2013110\u00a0\u00b0C as measured by an infrared (IR) temperature camera (UTi165A, UNI-T, China).In this work, N2 (99.99%) and H2 (99.99%) were used as the reactants and their flow rates were controlled by mass flow controllers (D07-B, Sevenstars, China). N2 (99.99%) and H2 (99.99%) were pre-mixed in a mixing chamber before flowing into the DBD reactor. The gas flow rate after the plasma reaction was measured using a soap-film flowmeter. A high voltage probe (Tektronix P6015A, 1000:1, USA) was used to measure the applied voltage, while a non-source voltage probe (Tektronix TPP0500, USA) was applied for measuring the voltage drop across a capacitor (0.47\u00a0\u03bcF). All signals were recorded by a Tektronix DPO2014 digital oscilloscope. The discharge power was determined using the Q-U Lissajous figure method (Eq. (1)).\n\n(1)\n\nP\n\nW\n\n=\nf\n\u00d7\n\nC\nm\n\n\u00d7\nA\n\n\nwhere Cm is the measuring capacitance (0.47\u00a0\u03bcF), f is the frequency (10.1\u00a0kHz in this work) and A is the area of the Lissajous diagram.The specific input energy (SIE) of the plasma process is expressed as follow:\n\n(2)\n\nSIE\n\n\nkJ\u00a0L\n\n-\n1\n\n\n\n=\n\n\nP\n\nW\n\n\n\nQ\n\n\nmL\u00a0min\n\n-\n1\n\n\n\n\n\n\u00d7\n60\n\n\nwhere Q is the total flow rate.The NH3 concentration was measured online using gas chromatography (Fuli 9720, China) equipped with a thermal conductivity detector (TCD). All data was measured three times and the average value was presented in this work.The energy yieldEY) of the plasma NH3 synthesis process is calculated according to Eq. (3):\n\n(3)\n\nEnergy\n\nYield\n\u00a0\n\n\ng\u00a0kWh\n\n-\n1\n\n\n\n=\n\n\nM\n\u00d7\n\nC\nout\n\n\u00d7\n\nQ\nafter\n\n\nP\n\n\n\nwhere M denotes the molar mass of NH3, Cout is the NH3 concentration measured at the reactor outlet and Qafter is the gas flow rate after the reaction measured by a soap-film flowmeter.\nTable\u00a01\n shows the textural properties of the Ru/M catalysts based on the N2 adsorption-desorption experiment [17]. The physical properties of the Ru/M catalysts show a significant difference due to the presence of different supports. Ru/AC has the highest SBET of 1333.8\u00a0m2 g\u22121, followed by the ZSM-5 (358.5\u00a0m2 g\u22121), \u03b1-Al2O3 (6.5\u00a0m2 g\u22121) and SiO2 (8.3\u00a0m2 g\u22121) supported Ru catalysts. The Ru/AC catalyst shows the highest porous volume of 0.79\u00a0cm3 g\u22121, which is about 3 times of the pore volume of Ru/ZSM-5 and Ru/\u03b1-Al2O3. Ru/SiO2 has an average pore size of 14.2\u00a0nm, much larger than that of Ru/ZSM-5 and Ru/AC (2.3\u20132.4\u00a0nm). Similar findings regarding the natures of the catalyst supports were reported in the synthesis of SiO2, ZrO2 and YSZ supported Ag catalysts [18].\nFig.\u00a02\n shows the XRD patterns of the Ru/M catalysts. The XRD patterns of all the Ru/M catalysts show the typical diffraction peaks of the supports, namely orthorhombic ZSM-5 (JCPDS No. 44-0003), hexagonal \u03b1-Al2O3 (JCPDS No.75-1864), hexagonal SiO2 (JCPDS No. 05\u20130490), and hexagonal activated carbon (JCPDS No. 50\u20131086). No identical diffraction peaks of Ru species were observed for any of the Ru/M catalysts, which could be ascribed to the low Ru loading amount or the high dispersion of Ru particles on the catalyst surface [19].\nFig.\u00a03\n presents the CO2-TPD profiles of the Ru/M catalysts. The CO2-desorption peaks below 250\u00a0\u00b0C are attributed to the weak basic sites, while the peaks located between 250\u00a0\u00b0C and 500\u00a0\u00b0C are associated with the medium basic sites [20]. Moreover, the peaks that appeared above 500\u00a0\u00b0C can be ascribed to the presence of strong basic sites [21]. For Ru/AC, a strong CO2 desorption peak is observed at 726\u00a0\u00b0C, while a weak desorption peak is located at 389\u00a0\u00b0C. Ru/ZSM-5 shows two small peaks at 166\u00a0\u00b0C and 756\u00a0\u00b0C, indicating the co-existing of strong and weak basic sites on the surface of the Ru/ZSM-5 catalyst. For Ru/\u03b1-Al2O3 and Ru/SiO2, only faint CO2-TPD desorption peaks are observed in the tested temperature range. The CO2 desorption amount of the Ru/M catalysts is associated with the basicity of the catalysts and is determined by the CO2-TPD profiles (Table\u00a01). The Ru/AC catalyst shows the highest CO2 desorption amount of 2.71\u00a0mmol g\u22121. The desorption amount of the catalysts follows the order of Ru/AC (2.71\u00a0mmol g\u22121)\u00a0>\u00a0Ru/ZSM-5 (0.33\u00a0mmol g\u22121)\u00a0>\u00a0Ru/SiO2 (0.17\u00a0mmol g\u22121)\u00a0>\u00a0Ru/\u03b1-Al2O3 (0.16\u00a0mmol g\u22121), indicating that Ru/AC has the highest basicity in this work.The effect of the N2/H2 molar ratio on the plasma-catalytic NH3 synthesis over the Ru/M catalysts is presented in Fig.\u00a04\n. The NH3 concentration is between 102\u00a0ppm and 251\u00a0ppm in the plasma reaction without a catalyst, while the highest NH3 concentration is obtained at the optimal N2/H2 molar ratio of 1:1. The presence of the Ru/M catalysts in the DBD reactor significantly enhances the NH3 concentration regardless of the N2/H2 molar ratio. It is reported that the generation of N radicals is crucial for NH3 synthesis since the dissociation energy (9.75\u00a0eV) of the N\u2261N triple bond is more than twice the dissociation energy of H2 molecules (4.52\u00a0eV) [13]. In an N2-rich condition, the probability of effective collisions between energetic electrons and N2 molecules could be increased, which may further accelerate the generation of N radicals and thus NH3 synthesis in the plasma environment. An early study by Bai et\u00a0al. reported a favorable N2/H2 molar ratio of 9:10 in an MgO coated DBD reactor for NH3 synthesis [22]. Shah et\u00a0al. also reported that the highest NH3 synthesis rate of 1.4\u00a0\u03bcmol min\u22121 was achieved at an N2/H2 ratio of 1:1 despite the presence of a 5A zeolite in the plasma reactor [23].\nFig.\u00a05\n shows the effect of gas flow rate on the plasma-catalytic NH3 synthesis at a discharge power of 9\u00a0W and an N2/H2 molar ratio of 1:1. For all cases, the NH3 concentration decreases with the increase of the total gas flow rate. Using the Ru/AC catalyst shows the highest NH3 concentration of 1544\u00a0ppm at 50\u00a0mL min\u22121, while further increasing the gas flow rate to 150\u00a0mL min\u22121 decreases the NH3 concentration to 439\u00a0ppm. Similarly, for the case of plasma only, the NH3 concentration decreases from 331\u00a0ppm to 175\u00a0ppm when raising the flow rate from 50\u00a0mL min\u22121 to 150\u00a0mL min\u22121. A higher gas flow rate reduces the residence time of reactants in the plasma-catalytic system. As a result, the possibility of effective collisions for 1) the generation of N radicals and H radicals between electrons and carrier gas molecules, and 2) the recombination of N radicals and H radicals for NH3 synthesis would be decreased in a given plasma reactor under the same reaction conditions regardless of the catalyst type, leading to a lower NH3 concentration. Similar phenomena were widely reported in plasma processes for NH3 synthesis, CO2 decomposition and oxidation of VOCs, etc. [24].\nFig.\u00a06\n shows the Lissajous figures of the discharge with and without the Ru/M catalysts at a constant discharge power of 9\u00a0W. The shape of the Lissajous figure changes from a parallelogram shape to an oval shape when the Ru/M catalyst is packed in the plasma reactor, indicating the variation of discharge mode in the presence of the Ru/M catalysts. Kim et\u00a0al. reported that the presence of a supported noble metal catalyst in the plasma reactor could expand the discharge region and the discharge mode could be shifted from typical filamentary micro-discharge to a combination of surface discharge and weak micro-discharges [25,26]. At a fixed discharge power, the peak-to-peak (pk-pk) applied voltage of the DBD reactor without packing is 12.6 kVpk-pk, while it increases to 13.4 kVpk-pk for the DBD reactor packed with Ru/ZSM-5, Ru/\u03b1-Al2O3 or Ru/SiO2, indicating a decreased current in the presence of these catalysts at a fixed discharge power. This phenomenon could be ascribed to the increased dielectric constant of the plasma reactor packed with the Ru/ZSM-5, Ru/\u03b1-Al2O3 and Ru/SiO2 catalysts compared to the non-packed DBD reactor [27]. The differences of relative dielectric constants between \u03b1-Al2O3, SiO2 and AC are quite small as listed in Table\u00a02\n and the dielectric constant of ZSM-5 is around 100. In the present work, the Lissajous figures for the Ru/ZSM-5, Ru/Al2O3 and Ru/SiO2 packed-DBD reactors are almost similar, suggesting the presence of these materials provides no significant effect to the electrical characteristics of plasma. It is interesting to note that the applied voltage of the DBD coupled with Ru/AC is only 11.6 kVpk-pk, significantly lower than the other cases in this work. The lower applied voltage for the Ru/AC packed DBD reactor could be ascribed to the electrical conductivity of the activated carbon support, which could contribute to the charge transfer in the plasma environment, and consequently decrease the applied voltage. Hong et\u00a0al. also reported that the charge transfer was enhanced by around 80% in a diamond-like carbon-coated Al2O3 packed plasma reactor compared to a bare-Al2O3 packed reactor at a fixed applied voltage [28].The effect of discharge power on NH3 synthesis over the Ru/M catalysts is shown in Fig.\u00a07\n. The NH3 concentration increases monotonically with the increasing discharge power for all cases. In the plasma ammonia synthesis without a catalyst, the NH3 concentration ranges from 31\u00a0ppm to 437\u00a0ppm when increasing the discharge power from 5\u00a0W to 18\u00a0W. The presence of the Ru/M catalysts in the plasma reactor considerably improves the reaction performance compared with the reaction using plasma alone. When packing Ru/AC into the DBD, the NH3 concentration is varied between 151\u00a0ppm and 1788\u00a0ppm in the same discharge power range, and the highest NH3 concentration is achieved at 18\u00a0W. The Ru/ZSM-5, Ru/\u03b1-Al2O3 and Ru/SiO2 catalysts show lower ammonia concentrations compared to Ru/AC. The energy dissipated into the plasma reactor was recognized as the driving force of plasma-induced NH3 synthesis since it could contribute to the generation of energetic electrons and consequently chemically reactive species including N and H radicals, excited N2 species and N2\n+ ions [2]. The increase of discharge power could increase the number density of filamentary micro-discharges and expand the discharge region, resulting in higher possibilities of effective collisions between the plasma species, enhancing the production of NH3 [31]. The energy yield of the NH3 synthesis process in the plasma-catalytic system increases within the discharge power range of 5\u00a0W\u20139\u00a0W, then the energy yield decreases when further increasing the discharge power. The highest energy yield of 0.64\u00a0g kWh\u22121 is achieved at 9\u00a0W over Ru/AC, followed by Ru/ZSM-5, Ru/SiO2 and Ru/\u03b1-Al2O3, as shown in Fig.\u00a07b. This phenomenon could be ascribed to the dynamic equilibrium between NH3 decomposition and recombination of N and H radicals at a high discharge power. Similar trends have been reported by Peng et\u00a0al. using an MCM-41 support for plasma-induced NH3 synthesis [24], and our previous work on catalyst screening for NH3 synthesis in a plasma reactor [3].The performance of the plasma-catalytic NH3 synthesis shows a distinct enhancement over the Ru/M catalysts. In the plasma-catalytic systems where the catalysts are directly in contact with the discharge, the local and average electric fields would be enhanced due to the higher dielectric constant of the catalysts, especially in the regions near the contact points between the catalyst pellets and reactor walls [32]. The intensified electric field could contribute to the generation of N and H radicals in the gas phase of the plasma region, contributing to the formation of NH\nx\n (x\u00a0=\u00a01 or 2) intermediates and NH3 molecules. Moreover, the reactions on the surfaces of the Ru/M catalysts also play a crucial role in the plasma-induced process as the radicals and intermediates could be transported and adsorbed on the catalyst surfaces and undergo a series of complex surface reactions for NH3 generation [33]. The physicochemical properties of the Ru/M catalysts may significantly affect the surface reactions in the plasma region. As shown in Table\u00a01, the Ru/AC catalyst possesses the highest SBET, followed by Ru/ZSM-5, Ru/\u03b1-Al2O3 and Ru/SiO2. A higher SBET value could offer more adsorption sites for the reactants and intermediates including N and H radicals, excited N2 species, etc. Thus, the residence time of these species would be prolonged on the surface of Ru/AC compared to the other Ru/M catalysts, resulting in higher possibilities of effective collisions for NH3 formation. The CO2-TPD profiles of the Ru/M catalysts show two major desorption peaks except for Ru/SiO2. Ru/AC shows the highest CO2 desorption amount, indicating it has the strongest basicity among the tested Ru/M catalysts. Previous work reported that the weak basic sites are associated with the Br\u00f8nsted basicity of the lattice-bond OH groups, while the medium and strong basic sites could be related to the Lewis basicity from three- or four-fold-coordinated O2\u2212 anions, showing stronger electron-donating capacity compared with the Br\u00f8nsted basic sites [34]. As a result, the presence of more basic sites, particularly medium and strong basic sites, could provide electrons to Ru species during the reaction and contribute to the dissociation of N2 molecules [35]. Previous work also confirmed that N2 dissociation could be enhanced over catalysts with lower electronegativity [36]. It is worth noting that materials with a higher electronegativity tend to accept electrons during the catalytic reactions, which may inhibit N2 dissociation and reduce the formation of N species for NH3 synthesis. The adsorbed N species could react with the H radicals in the gas phase and on the catalyst surfaces to form NH3 molecules. The order of basicity of the Ru/M catalysts is in accordance with the activity of the plasma-catalytic NH3 synthesis, indicating that the basicity of the catalysts is a very important factor to tune the reaction performance of the plasma-catalytic NH3 synthesis.\nFig. 8\n summarizes a comparison of energy yield in the process of plasma NH3 synthesis over various catalysts. The energy yield of NH3 synthesis in the cases of plasma only ranges from 0.11\u00a0g kWh\u22121 to 0.28\u00a0g kWh\u22121 in previous works. Clearly, introducing a catalyst in the plasma reactor could considerably improve the energy yield of NH3 synthesis, and the composition of catalysts is critical to determine the reaction performance. The energy yield of NH3 synthesis over bare supports (without a metal) is much lower compared to the supported catalysts. For example, Patil et\u00a0al. reported an energy yield of 0.34\u00a0g kWh\u22121 in the plasma-catalytic NH3 synthesis over BaTiO3 at an SIE of 1.3\u00a0kJ L\u22121 [37]. The presence of a supported active metal phase can significantly improve the energy yield of ammonia. Mehta et\u00a0al. reported that the energy yield of ammonia production using Ni/Al2O3 was 0.89\u00a0g kWh\u22121, almost twice that of the bare Al2O3 support at the same SIE of 6.0\u00a0kJ L\u22121 [5]. The results show that the doping of active metals can significantly enhance the energy yield of the NH3 synthesis. A similar finding was also reported by Xie et\u00a0al. using Ru/L-MgO which gave a higher energy yield of 1.14\u00a0g kWh\u22121 compared to that (\u223c0.3\u00a0g kWh\u22121) using plasma only [15]. In this work, the energy yield of ammonia production is about 200% higher than that of the reaction using plasma only. This significant enhancement can be attributed to the promotion of the dissociation of N2 and H2 molecules on the catalyst surface through the Eley\u2013Rideal mechanism and Langmuir-Hinshelwood mechanism with the presence of Ru, which accelerates the reaction of ammonia synthesis.In this work, the highest energy yield of 0.63\u00a0g kWh\u22121 was achieved at an SIE of 5.4\u00a0kJ L\u22121 using Ru/AC, which is 21.2% higher than that of using the Ru/ZSM-5 catalyst. Our results show that the type of catalyst support could directly affect the performance of the plasma-catalytic ammonia production. It is worth noting that the value of energy yield is a bit lower when compared with our previous work. The difference could be a result of the different loading of active metal Ru since a significantly higher amount of Ru (5\u00a0wt.%) was used in our previous work [20]. However, the energy yield achieved in this work could still be optimized further. The performance of the ammonia synthesis rate in the plasma environment not only depends on the components of the catalyst but also various parameters, such as reactor configuration [38,39] and operation conditions [40,41], etc. For example, increasing the ammonia synthesis performance by increasing the amount of catalyst used does not seem economically viable as it leads to added costs for the catalyst and reactor size. To sum up, the balance between the energy yield and NH3 concentration should be taken into account in the stated studies, while the optimization of the amount of catalyst used and catalyst compositions should be considered for further development and optimization of the plasma-catalytic NH3 synthesis process.In this work, the effect of various catalyst supports on the plasma-catalytic NH3 synthesis over the Ru/M catalysts was studied in a DBD plasma reactor. The NH3 concentration and energy yield of the plasma-catalytic process were significantly affected by the different supports. Compared with the reaction using plasma alone, the presence of the Ru/M catalysts improved the NH3 concentration by 163.4%\u2013387.6% at an SIE of 5.4\u00a0kJ L\u22121, and the energy yield of ammonia production was increased by 163.1%\u2013387.0%. The reaction performances followed the order of Ru/AC\u00a0>\u00a0Ru/ZSM-5\u00a0>\u00a0Ru/\u03b1-Al2O3\u00a0>\u00a0Ru/SiO2. The results also showed that the optimum N2/H2 molar ratio for NH3 synthesis was 1:1 in this work, and lower gas flow rates benefitted NH3 production. The catalyst characterization showed that the enhancement in NH3 synthesis in the plasma reactor over the Ru/AC catalyst could be attributed to the larger specific surface area, pore volume and stronger basicity of the Ru/AC catalyst.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Natural Science Foundation of China (No. 51976093) and K. C. Wong Magna Fund at Ningbo University. X. Tu acknowledges the support from the Engineering and Physical Sciences Research Council (No. EP/V036696/1) and the British Council Newton Fund Institutional Links Programme (No. 623389161).", "descript": "\n                  In this work, we have investigated the effect of different supports (activated carbon (AC), \u03b1-Al2O3, ZSM-5 and SiO2) on the plasma-catalytic synthesis of ammonia (NH3) from N2 and H2 over Ru-based catalysts in a dielectric barrier discharge (DBD) plasma reactor. Compared with the NH3 synthesis using plasma alone, the presence of the Ru-based catalysts in the DBD reactor significantly enhanced the NH3 production and energy yield by 163%\u2013387.6% with a sequence of Ru/AC\u00a0>\u00a0Ru/ZSM-5\u00a0>\u00a0Ru/\u03b1-Al2O3\u00a0>\u00a0Ru/SiO2. The effect of different operating parameters on the plasma-catalytic NH3 synthesis over Ru/AC was also examined. N2 adsorption-desorption experiment, X-ray diffraction analysis and temperature-programmed desorption of CO2 were performed to get insights into the structure-performance relationships between the plasma-catalytic NH3 synthesis and Ru-based catalysts with different supports. Both textural properties and the basicity of the Ru/AC catalyst contributed to the enhanced NH3 production in the hybrid plasma-catalytic system.\n               "}
{"full_text": "Water electrolysis, powered by renewable energy (e.g., solar, wind), has been identified as presently the most favorable way of producing high-purity green hydrogen (H2), which can substitute conventional fossil fuels to decarbonize different sectors of our economy [1]. It is predicted that the surge in the demand for green H2 will significantly boost the installation of many gigawatt (GW) electrolyzers worldwide by 2030 [2]. To turn this blueprint into reality, it is important to keep developing new materials and components composed of inexpensive and earth-abundant materials that can be integrated into electrolyzers to improve performance and lower H2 production cost. Particularly, there is a pressing need to replace the costly and scarce platinum group metal (PGM) catalysts with earth-abundant transition metal (TM)-based alternatives to promote the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER). To this end, substantial effort has been made in the past decade, and a variety of TM-based electrocatalysts, such as phosphides [3], chalcogenides [4], nitrides and carbides [5], were reported, having shown HER and/or OER performance comparable to their PGM counterparts. Among these emerging catalysts, transition metal tellurides (TMTs) have recently attracted considerable attention, given that tellurium (Te) shows lower electronegativity and greater metallic character compared to other chalcogens like selenium (Se), sulfur (S), and oxygen (O) [6], which can give rise to enhanced electrical conductivity and a higher degree of covalency in the metal\u2013chalcogen bonds in TMTs. Such a covalent character can lead to a favorable electronic band structure, facilitate the alignment of the valence and conduction band edges with the water oxidation/reduction potentials, and also promote the redox reactions of the transition metal center, contributing to electrocatalytic performance improvement [7\u20139].Silva et\u00a0al. lately proposed that TMTs would show better electrocatalytic activity for OER compared to other chalcogenides, because the coordination environment around the TM atoms largely influences the TM's pre-oxidation; with a decreased electronegativity of the anion (i.e., S\u00a0>\u00a0Se\u00a0>\u00a0Te), the OER onset potential can be effectively lowered. They further experimentally verified their hypothesis using Ni chalcogenides (i.e., Ni3Te2, Ni3Se2 and Ni3S2) synthesized under similar conditions [10]. Although an opposite activity trend was observed for Chevrel-phase Mo6X8 (X\u00a0=\u00a0S, Se, Te) electrocatalysts toward the HER [11], using the hydrogen adsorption free energy (\u0394GH) as the activity descriptor Lee et\u00a0al. recently demonstrated that many TMTs (e.g., ZrTe2, TiTe2, MoTe2, and VTe2) with a proper density of anion vacancies are situated on top of the volcano plot exhibiting much better HER activity compared to the corresponding transition metal sulfides and selenides [12].This opinion article provides a short account of the advances made in the last 3\u20134 years in developing TMT-based electrocatalysts for use in HER, OER, and overall water splitting, with a focus on major strategies developed so far to improve the electrocatalytic performance, including nanostructure engineering, composition engineering, and heterostructuring/hybridization. An outlook about future research on the design and development of TMT catalysts is also outlined.According to previous studies [13\u201315], bulk TMTs only exhibit mediocre electrocatalytic activity. To improve the catalytic performance, nanostructuring has turned out to be an effective strategy because it not only allows more active sites to be exposed but also facilitates mass transfer of electrolyte and gaseous products [10,15,16]. For instance, MoTe2 nanosheets (NSs) prepared by liquid exfoliation demonstrated a notably enhanced HER activity in 0.5\u00a0M\u00a0H2SO4, compared to the bulk MoTe2 [14]. Hence, much effort has been devoted in the last few years to developing various TMT nanostructures, such as hollow NiTe2 nanotubes (NTs) [17], core\u2013shell CoTe2@NC nanoparticles (NPs) [18], and hierarchical CoTe2 nanowires (NWs) [19].Ananthara et\u00a0al. synthesized NiTe2 nanostructures with two distinct morphologies, for example, NWs and nanoflakes (NFs) by a short-time hydrothermal treatment of Ni foam in the presence of Te powders and NaHTe, respectively. They found that NiTe2 NWs showed better HER activity than NiTe2 NFs in both acidic and alkaline conditions (Figure\u00a01\na), and meanwhile exhibited a Tafel slope similar to (in 0.5\u00a0M\u00a0H2SO4) or smaller than (in 1\u00a0M KOH) that of commercial Pt/C benchmarks, revealing favorable reaction kinetics [20]. They ascribed the improved performance to the high charge-transfer ability of NiTe2 NWs and their large electrochemically accessible surface area (ECSA). Nanostructuring can help effectively expose catalytically active sites. To this end, Zhang et\u00a0al. recently found that although the pristine chemical vapor deposited (CVD) 1T\u2032-MoTe2 ultrathin films showed an inconspicuous HER activity, after ion-beam etching the HER performance of the 1T\u2032-MoTe2 films was significantly enhanced, able to achieve a current density of 100\u00a0mA\u00a0cm\u22122 at an overpotential (\u03b7) of 296\u00a0mV and a small Tafel slope of 44\u00a0mV dec\u22121 in 0.5\u00a0M\u00a0H2SO4 [15]. Moreover, the ion-beam etched sample revealed an improvement in catalytic stability, retaining 87% of the initial current density when compared with the pristine sample which only had 40% of the initial current density after continuous electrolysis for 3600\u00a0s [15]. The enhancement was attributed to the largely exposed active edge sites, which was confirmed by conductivity measurement, visualized copper electrodeposition, and density functional theory (DFT) calculation. While the authors claimed that the ion-beam etching method can be extended to increase the active sites of other materials, it is arguably time-consuming and perhaps economically unviable for massive production of catalysts.Metal\u2013organic framework (MOF) has been extensively used to prepare nanostructured electrocatalysts given its spatially-ordered microstructure, large specific surface area and high nanoporosity. Using zeolitic imidazolate framework (ZIF)-67 as a template, Wang et\u00a0al. developed a composite catalyst comprising CoTe2 NPs encapsulated in nitrogen-doped carbon nanotube frameworks (CoTe2@NCNTFs, Figure\u00a01b) [21]. This MOF-derived catalyst possesses a large surface area, high conductivity, and open channels for effective gas release; moreover, it enables fast electron transport presenting a Tafel slope much smaller than that of bulk CoTe2. Consequently, CoTe2@NCNTFs was reported to show good catalytic performance, requiring overpotentials of 330 and 208\u00a0mV to achieve 10\u00a0mA\u00a0cm\u22122 for the OER and HER in 1.0\u00a0M KOH. When used as bifunctional catalysts for overall water splitting, CoTe2@NCNTF could afford 10\u00a0mA\u00a0cm\u22122 at a cell voltage of 1.67\u00a0V (Figure\u00a01c). Using a similar approach, Wang et\u00a0al. further demonstrated that the composition of cobalt telluride in MOF-derived nanostructures can be easily adjusted [22]. The obtained optimal Co1\n\u00b7\n11Te2/C catalyst had more reducible Co species and higher surface dispersion of Co ions, compared to the CoTe/C and CoTe2/C references, leading to notably enhanced HER performance (178\u00a0mV@10\u00a0mA\u00a0cm\u22122 in 1\u00a0M KOH). The observed high activity of Co1\n\u00b7\n11Te2/C was explained by DFT calculations, where Co1\n\u00b7\n11Te2/C shows an optimal Gibbs free energy (Figure\u00a01d).Besides, electrocatalysts in-situ grown on current collectors, which form self-supported ready-to-use electrodes, also drew considerable attention lately. Previous studies have demonstrated that such an architecture can prevent the collapse and agglomeration of nanostructured catalysts, rendering long-term stability of the electrodes [23\u201326]. Additionally, efficient charge transfer and mass diffusion can be achieved given the intimate, binder-free adhesion of catalysts to the current collector. In this respect, a number of self-supported TMT-based HER/OER electrodes were reported recently, for example, CoTe2 NWs array on Ti mesh [19], CoTe2 NPs on Co foam [27], Cu7Te4 arrays on Cu foil [28], NiTe2 NWs on Ni foam [20], NiTe2 NS array anchored on Ti mesh [16], and FeTex NSs on Fe foam [29], which cannot be exhaustively elaborated in this short review article. However, it is worth noting that under OER conditions, TMTs usually show unfavorable electrochemical stability. The surface Te species tend to become soluble and the TMT will be eventually converted into the corresponding metal oxyhydroxide, which serves as the real catalytically active species, as reported before [30] and verified by the CoTe nanoarrays grown on Ni foam reported recently by Yang et\u00a0al. [31].While nanostructure engineering can enhance the catalytic activity by tuning catalyst's physical morphology and structure to expose more catalytically active sites, in order to further boost the performance the intrinsic catalytic activity of materials must also be improved, which can be enabled by composition engineering of catalysts. Tuning the composition of TMT catalysts may lead to changes of local coordination environment and chemical properties, enhancing the catalytic activity through the ligand and/or ensemble effects. To this end, heteroatom doping has been widely adopted to regulate the electronic structure, and several transition metal (e.g., Ni, Co, Fe) [32\u201334] and nonmetal (e.g., P, S) [35\u201337] elements were already successfully doped into TMTs to improve catalytic performance. For example, by doping Fe into Mo/Te nanorods (NRs), He et\u00a0al. demonstrated that the catalytic stability of Mo/Te NRs could be largely improved, because Fe-doping promoted the formation of high valence state Mo species and induced strong electronic state modification [38]. Additionally, after Fe doping, the Tafel slope and charge transfer resistance of Fe\u2013Mo/Te became smaller, and the Tafel analysis revealed that the main kinetics pathway involves a mixed step of the M\u2212O or M\u2212OOH formation [38]. Moreover, Fe-doping was also reported to be able to boost the catalytic performance of Co1\n\u00b7\n11Te2 NPs encapsulated in nitrogen-doped carbon nanotube frameworks (NCNTF) [39]. He et\u00a0al. found that the Fe\u2013Co1\n\u00b7\n11Te2@NCNTF obtained by tellurization of Fe3+-etched ZIF-67 in H2/Ar gas showed a blue-shift in binding energy in the Co2p XPS spectrum, relative to the pristine Co1\n\u00b7\n11Te2@NCNTF (Figure\u00a02\na), resulting in Co species with decreased electron density and a higher intensity of Co3+ components. This rationally explained the better HER and OER performance of Fe\u2013Co1\n\u00b7\n11Te2@NCNTF, which presented TOF values ten times higher than those of undoped Co1\n\u00b7\n11Te2 for both reactions [39]. Besides, Pan et\u00a0al. demonstrated that doping Mn into 1T-VTe2 helped stabilize the 1T-phase and develop nanosheet-like structure with a high surface area and porosity, which substantially enhanced the HER and OER performance, compared to the undoped 1T-VTe2 (Figure\u00a02b and c) [40]. Moreover, the performance could be further boosted when hybridizing Ni nanoclusters (NiNCs) with Mn-doped 1T-VTe2 NSs (NiNCs-1T-Mn-VTe2 NS) which showed markedly improved reaction kinetics among all catalysts investigated [40].Previous study disclosed that the number of dopants can influence the electrocatalytic performance [41]. This was also demonstrated in TMT-based catalysts. Based on DFT calculations, Gao et\u00a0al. found that co-doping of Co and Ni into MoTe2 can readily trigger the 2H-to-1T' phase transition, compared to the monoatom doping (Figure\u00a02d) [33]. They further experimentally proved the markedly enhanced HER performance for the Co/Ni co-doped MoTe2.Aside from cation doping, anion incorporation can also effectively improve the electrocatalytic activity. Wang et\u00a0al. reported that S-doping can turn the electrocatalytically inactive 2H\u2013MoTe2 into an active catalyst, due likely to the ligand effect induced electronic structure changes, upon which electrons accumulate on the surface S atoms such that the S sites can adsorb H\u2217 intermediate more readily, promoting the HER [42]. Similarly, Chen et\u00a0al. demonstrated the activation of CoTe2 for OER by doping secondary P anions into Te vacancies to trigger a structural transformation from the hexagonal to the orthorhombic phase (Figure\u00a02e) [43]. This allowed a current density of 10\u00a0mA\u00a0cm\u22122 to be achieved at \u03b7\u00a0=\u00a0241\u00a0mV, lower than the hexagonal CoTe2 (\u03b7\u00a0=\u00a0308\u00a0mV@10\u00a0mA\u00a0cm\u22122). Besides, other mixed tellurides and anion-doped TMTs, such as Ni1-xFexTe2 hierarchical nanoflake arrays [44], free-standing CoNiTe2 NSs [45], MoSxTey/Gr [46] and MoSe0\n\u00b7\n12Te1.79 solid solutions [47], were also explored recently as HER and OER catalysts. Most of these reports suggest that the stoichiometry between two metals or two chalcogen elements is a key factor of regulating the intrinsic catalytic activity through electronic structure modulation.Besides nanostructure and composition engineering, heterostructuring or hybridization of the active catalyst with other active components has also turned out to be an effective approach to boosting the catalytic performance. Typically, such heterostructuring/hybridization can introduce abundant interfaces, which allows for engineering the electronic structure and thus the selectivity and reactivity of the catalysts. Moreover, the ensemble effect may come into play in the exposed hetero-interfaces through the migration of adsorbed reaction species from one component to the other, unlocking unprecedented catalytic reaction pathways and thereby promoting the overall reaction rate [48,49]. In this regard, a number of TMT-based heterostructured catalysts have been recently reported to show improved HER/OER performance, such as TMT nanostructures composited with a secondary TMT [50\u201352], an oxide/hydroxide [53\u201355], a chalcogenide [56,57] or a phosphide [58,59]. For instance, Xu et\u00a0al. demonstrated that Ni3Te2\u2013CoTe hybrids grown on carbon cloth in a single-step hydrothermal process (Figure\u00a03\na) showed better OER activity than each individual component (i.e., Ni3Te2, CoTe), capable of affording a current density of 100\u00a0mA\u00a0cm\u22122 at \u03b7\u00a0=\u00a0392\u00a0mV [60], with a low Tafel slope of 68\u00a0mV dec\u22121 indicating that the chemisorption of hydroxyl groups on the catalyst surface is the rate-determining step. The authors proposed that the improved performance results from the incorporation of CoTe that helps expose more Ni3Te2 active sites, reflected by the double-layer capacitance measurements (Figure\u00a03b and c) and the high density of states near the Fermi level in the Ni3Te2 component, as suggested by DFT analysis. Additionally, Xue et\u00a0al. reported a NiTe/NiS heterojunction fabricated by coupling NiS nanodots (NDs) on hydrothermally-synthesized NiTe nanoarrays in an ion-exchange process (Figure\u00a03d) [57]. The NiS NDs were found to decorate on the surface of NiTe with a high density (Figure\u00a03e), and the interface between NiS and NiTe could be clearly seen under high-resolution transmission electron microscopy (HRTEM) examination (Figure\u00a03f). The introduced NiTe/NiS nanointerfaces led to notable electronic structure modulation, thus optimizing the binding energy of the \u2217OOH intermediates. This could be explained by the ligand effect, given that the d-band center of Ni in NiTe/NiS shifts to low-energy level with respect to NiTe due to the triggered electron transfer from Ni to S, which decreases the binding strength of intermediates on catalyst surface, resulting in a low reaction barrier. Consequently, the hybrid catalyst only needed an overpotential of 257\u00a0mV to deliver 100\u00a0mA\u00a0cm\u22122 and showed a Tafel slope of 49\u00a0mV dec\u22121 for OER in 1.0\u00a0M KOH, much lower than pure NiTe and NiS. Moreover, the catalyst also exhibited good stability of over 50\u00a0h at 50\u00a0mA\u00a0cm\u22122 (Figure\u00a03g), with a potential increase of around 6%. Besides, Sun et\u00a0al. managed to couple RuO2 and NiFe layered double hydroxide (NiFe-LDH) on NiTe NR surfaces, forming NiTe@RuO2 and NiTe@NiFe-LDH heterostructures, which were used as cathode and anode, respectively, for overall water splitting [61]. The assembled device delivered a current density of 200\u00a0mA\u00a0cm\u22122\u00a0at a voltage of 1.63\u00a0V and could be powered by a 1.5-V solar cell for continuous water electrolysis. This result favorably compares to the Pt/C||NiTe@FeOOH electrode pair reported by the same group [62], which showed a voltage greater than 1.6\u00a0V for the same current density.Our group recently developed heterostructured dual-phase CoP\u2013CoTe2 NWs with abundant interfaces, which exhibited good HER and OER performance in acidic/alkaline solutions [59]. The dual-phase CoP\u2013CoTe2 NWs were used as bifunctional catalysts for bipolar membrane water electrolysis (BPMWE). The use of a BPM allows HER and OER to take place simultaneously in their respective kinetically favorable acidic and alkaline electrolytes. Particularly, when used in the forward-bias configuration, that is, the cation-exchange layer (CEL) faces the alkaline anolyte and the anion-exchange layer (AEL) faces the acidic catholyte, electrochemical neutralization between OH\u2212 and H+ ions happens, which will assist water electrolysis by lowering the external electrical energy needed. We demonstrated that using CoP\u2013CoTe2 as bifunctional electrocatalysts, the BPM electrolyzer in the forward-bias configuration could deliver 10\u00a0mA\u00a0cm\u22122 at a low cell voltage of merely 1.01\u00a0V (Figure\u00a03h), and it could operate stably for 100\u00a0h without notable degradation, presenting performance better than that of anion-exchange membrane water electrolysis (AEMWE) using the same CoP\u2013CoTe2 electrode pairs (Figure\u00a03i). The forward-bias BPMWE represents a promising alternative to the conventional proton-exchange membrane water electrolysis (PEMWE) and AEMWE technologies, enabling hydrogen production with minimized electrical energy consumption.Transition metal tellurides have recently emerged as a class of promising electrocatalysts for both hydrogen and oxygen evolution reactions. Although bulk TMTs usually only show inferior catalytic activity, particularly for the HER, various strategies including nanostructure engineering, composition engineering and interface engineering, have been developed to improve TMT's electrocatalytic performance, as outlined in this article. Table\u00a01\n summarizes the HER, OER and overall water splitting performance of most TMT-based nanocatalysts reported in the last couple of years. Notwithstanding remarkable achievements in catalyst design and synthesis, further improvement in electrocatalytic performance is still critically needed. In particular, most TMT catalysts reported so far were only tested at low current densities relevant to solar water splitting (e.g., 10\u201320\u00a0mA\u00a0cm\u22122), and the viability of using TMTs for water electrolysis under industry-relevant conditions (e.g. current density higher than 500\u00a0mA\u00a0cm\u22122)\u00a0has not been assessed yet. Additionally, the catalytic stability of TMT-based catalysts should be substantially improved. Furthermore, TMT-based catalysts should also be validated in membrane electrode assemblies (MEAs) or single cells, instead of only in an aqueous model system. This will help evaluate the commercial viability of TMT catalysts. Many TMTs show metallic character and possess high electrical conductivity, which is advocated to facilitate charge transfer. However, high electrical conductivity is not the only determining factor of electrocatalysis. It is crucially important to tune the electronic structure of TMTs through composition and/or interface engineering to lower reaction barriers and regulate the binding energy of reaction intermediates. An in-depth understanding is yet to be achieved in this regard, and the combination of theoretical predictions with delicate experimental studies, especially using advanced in-situ and operando spectroscopic and microscopic characterization techniques, will provide fundamental new insights into the catalytic/degradation mechanisms of TMT catalysts, contributing to the rational design of TMTs with significantly improved electrocatalytic performance.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.L. Liu acknowledges the financial support of National Innovation Agency of Portugal through the Mobilizador programme (Baterias 2030, Grant No. POCI-01-0247-FEDER-046109). I. Amorim is thankful to Funda\u00e7\u00e3o para a Ci\u00eancia e Tecnologia (FCT) for the support of PhD grant No. SFRH/BD/137546/2018, co-financed by the Fundo Social Europeu (FSE) through the Programa Operacional Regional Norte (Norte 2020) under Portugal 2020.", "descript": "\n                  Renewable energy powered electrochemical water splitting has been recognized as a sustainable and environmentally-friendly way to produce green hydrogen, which is an important vector to decarbonize the transport sector and hard-to-abate industry, able to contribute to achieving global carbon neutrality. For large-scale deployment of water electrolyzers, it is essential to develop efficient and durable electrocatalysts\u2014one of key components determining the electrochemical performance, based on cheap and earth-abundant materials. To this end, transition metal tellurides (TMTs) have recently emerged as a promising alternative to the conventional platinum group metals for both hydrogen evolution reaction (HER) and oxygen evolution reaction (OER). This review article provides a brief account of the latest development in TMT-based HER and OER catalysts, with a focus on various strategies developed to improve the catalytic performance, such as nanostructure engineering, composition engineering, and heterostructuring/hybridization. Perspectives of future research on TMT-based catalysts are also shortly outlined.\n               "}
{"full_text": "Catalysts are applied to chemical processes within the commodity, specialty, fine, and pharmaceutical industries to decrease production costs, reduce waste generated, and improve reaction yield. On an industrial scale, heterogeneous catalysts are typically used due to their implementation versatility, ease in separation, and extensive customizability. Transition metal heterogeneous catalysis innovation focuses on active site availability, broad reaction applicability, and process activity through optimizing material combinations of metal and support systems. Furthermore, new or existing catalysts could be applied to a broad range of chemical systems to achieve previously inaccessible outcomes. Research areas within this field include synthesizing precision single-site catalysts, designing support structures, and defining single or multiple metal and metal oxide active catalytic components. A common support material found in a variety of catalysts is activated carbon, which can be produced from a variety of organic feed materials. However, various new allotropes of carbon have been discovered with great potential as catalyst supports. One support material, graphene, has received increasing attention due to its many properties: 1) native catalytic properties enabling co-catalytic functions [1\u20134], 2) bolstered catalytic activity of supported metals [5\u201313], 3) chemical functionalization to customize catalytic properties [2,14-16], 4) tough lattice structure and electric conductivity, and 5) specific enhancement of electron donor/acceptor properties of the metal [17,18].Graphene, a 2D allotrope of carbon, consists of sp2-hybridized carbon atoms packed in a hexagonal honeycomb lattice. These hybridized sp2 orbitals allow the carbon to form three covalent \u03c3-bonds, separated by a distance of 0.142\u00a0nm, attributing to the toughness of graphene's lattice structure. The fourth unused electron of carbon atoms form a \u03c0-bond, perpendicular to the \u03c3-bonds, enabling efficient electron transfer between atoms [19]. The \u03c0-bond is attributed to the electronic properties of graphene [15,19-24]. Fig.\u00a01\n below represents the 2D and 3D structure of graphene due to the hybridized carbon bonds.Similar to graphene, carbon nanotubes (CNTs) are essentially graphene sheets rolled into a 1D tube called single-wall CNTs (SWCNTs) or into a 3D structure with multiple tubes encased within each other called multi-wall CNTs (MWCNTs). CNTs possess exceptional mechanical, chemical, electronic, and optical properties which make these materials excellent candidates for catalytic applications [25\u201330]. A distinguishable factor between graphene and CNTs is the aspect ratio, or ratio between the length and width or diameter. CNTs typically have an aspect ratio in excess of 1000 with minimal defects or distortions along the length of tube [31]. Fig.\u00a02\n below represents the differences between graphene, SWCNTs, and MWCNTs.While the invention of CNTs was a great achievement for the materials research community, small volume production and excessive cost due to processing complications restrict CNTs from further development and adoption in commercial applications [32]. The presence of impurities, nonuniformity in morphology and chemical structure, limited control over length and chirality, and the discrepancy between quality and yield limit CNT implementation as a support material for commercial catalytic applications [33\u201336]. However, active research on CNTs as a catalyst support is still ongoing and hopefully in the future these complications will be overcome.Heterogeneous catalyst supports are an essential component of supported metal catalysts. The support-metal interaction is typically the defining feature of specific support-metal combinations and are often application dependent. The support can improve the native catalytic activity of metal nanoparticles through these interactions by increasing electron transport for the active catalyst center or by promoting better metal dispersion and narrow particle size distribution resulting in high catalytic performance [37,38]. Furthermore, support materials can behave as metal particle stabilizers which prevent agglomeration and aggregation of metal nanoparticles during catalysis, promoting catalytic activity for the desired reaction pathway and preventing deactivation. Mechanical strength, external surface area, narrow pore size distribution and internal surface area, strong thermal and chemical stability, high resistance towards metal sintering, and low-cost are among the important criteria for support selection. Carbon materials can be designed and synthesized to meet a variety of performance criteria based the desired application, which make carbon one of the most promising support materials in heterogeneous catalysis [39,40]. Table\u00a01\n below summarizes important physical properties characteristic to carbon supports.\nTable\u00a01 demonstrates how different carbon material types possess different physical properties which make certain forms better suited for different applications. Carbon can be fabricated into a variety of geometric forms such as granules, extrudates, pellets, fibers, powder, and cloth. The flexibility in manufacturing of carbon materials means carbon can be synthesized to meet a variety of specifications such as pore size distribution for selective reactions, chemical modification by introduction of certain elements or reactive groups, and external surface area. Furthermore, the relative inertness, ability to adsorb metals, high thermal stability (>1000\u00a0K in inert, \u223c500\u00a0K in oxygen, and \u223c700\u00a0K in hydrogen), relative abundance and low cost often result in the choice of carbon over other conventional supports such as silica and alumina [39]. Graphene, and CNTs, can be synthesized and customized to meet any specific application for heterogeneous catalysts. A material property summary for both graphene and CNTs are in Table\u00a02\n below.Pristine graphene is considered an efficient electron conductor due to its hybridized bonds resulting in high electron mobility (\u223c200,000 cm2/Vs). This is the highest value compared to conventional semiconductor materials such as silicon (\u223c1400 cm2/Vs), indium antimonide (77,000 cm2/Vs), and carbon nanotubes (>100A,000 cm2/Vs) [45\u201347]. Graphene exhibits superior thermal conductivity (\u223c5000 Wm\u22121K\u22121), [48] a high surface area (2630 m2/g) [61], an excellent Young's modulus (1.0 TPa) [53], and high optical transmittance (97.7%) [52]. This electronic network, along with graphene's mechanical properties, makes graphene a robust material able to withstand repeated applications of chemical or mechanical stress [22-24,62-67].Through study of both the fundamental nature and manipulation of graphene's properties, the potential of graphene as a support system is expanding the field of transition metal heterogeneous catalysis. To catalyze future graphene research, this review focuses on the following features 1) modification of graphene to create unique structures, 2) improvement of graphene's support function due to the aforementioned modifications, 3) synthesis methods for leveraging these effects, and 4) statistical analysis focusing on future opportunities to optimize applications for graphene based catalysts.Metastable states for any material exhibit unusual behavior compared to classically stable states. This behavior is often associated with \u201cdefects\u201d in the material that forms a kinetic, metastable state and account for the observed properties to differ from a thermodynamically stable phase. These defects, which exist as atomic vacancies and intrusions, molecular rearrangements, and grain boundary changes, are caused by applied stresses which enable graphene's properties to be tuned for specific applications. These properties are dependent on the nature of applied stress and how graphene re-hybridizes to perform repairs [23,65,66]. Graphene is typically synthesized by chemical vapor deposition of methane or other carbon sources or by exfoliating graphite [22-24,64,66,68-76]. These techniques yield either pristine or defective graphene depending on feed source, synthesis conditions, and forces applied during synthesis. Graphene defects can be characterized as point, line, and interlayer defects (Fig.\u00a03\n) [63,65,77-79].0D point defects are single atom substitutions or additions, vacancies, and reconstructions resulting from an inserted or displaced atom. Common 0D point defects include vacancies, adatoms and substitutions, and Stone-Wales (SW) defects. Vacancies result when one or more carbon atoms are displaced from their normal position within the hexagon lattice, leading to form a variety of polygons (Fig.\u00a04\na) [80\u201382]. Adatoms (same-type atom) and substitutions (different-type atom) occur when an atom is inserted into the carbon network, displacing the atoms and effectively doping graphene (Fig.\u00a04b). SW defects are caused by the rotation of the CC bond without any loss or gain of carbon atoms. A single CC rotation transforms four adjacent hexagons into two separate pentagons and two heptagons, which share the rotated bond (Fig.\u00a04c) [80\u201383]. In addition to 0D defects, 1D line defects and 2D bilayer and multilayer defects occur when applied stresses cause decreased dimensionality in graphene. 1D line defects appear as atomic dislocations or grain boundary changes where atoms are anomalously organized (Fig.\u00a04d). 2D bilayer and multilayer defects are the manifestations of 0D and 1D defects when defect-containing graphene sheets are stacked on one another (Fig.\u00a04e).Experiments investigating graphene bond re-hybridization have been conducted to determine how graphene re-hybridizes after applied external chemical and physical stresses. These experiments shed light on how the resulting metastable state could be tuned to enhance specific properties, such as electron transport, metal-binding strength, and durability [22-24,62-66]. Computational experiments were able to predict and explain the mechanism behind the formation of these thermodynamically metastable structures. The geometrical change of the graphene layer due to vacancy movement at high temperature is shown in Fig.\u00a05\n where a single vacancy proceeds toward another single vacancy by successive jumps that eventually form a stable 555\u2013777 SW like defect (more details in [67]).The most common method of graphene functionalization is to introduce non-metallic elements, usually oxygen, nitrogen, phosphorous, and sulfur, to graphene through chemical modification. Functionality refers to when multiple atoms of a non-carbon element is added to graphene in a non-specific manner. Graphene oxide (GO) is a prime example of oxygen-functionalized graphene (Fig.\u00a06\n).GO was first developed by Hummers and Offeman in 1958 in an effort to introduce oxygen functionality and exfoliate graphite [7,84]. The Hummers method involves subjecting graphite to a concentrated solution of sulfuric acid, sodium nitrate, and potassium permanganate at room temperature [24,64,84-89]. The graphite simultaneously becomes oxidized and exfoliated to form GO from this process. Due to the oxygen functionality, the GO sheets can no longer restack and thus remain as separate sheets. Characterization of the specific functional groups on GO is challenging but imperative to understanding the nature and degree of functionalization when GO is synthesized [90]. For this reason, GO is often used as a precursor for producing exfoliated graphene from graphite rather than used directly as a support. Graphene is often produced by reducing GO to form reduced GO (rGO) from graphite following an oxidation procedure much like the Hummers method. RGO is typically referred to as defective graphene since the oxidation and reduction processes tend to leave the sheet crumpled and with other defects. Graphene has also been functionalized with nitrogen [91], phosphorous [92], and sulfur [93].Graphene doping is the process of adding a single atom, typically heteroatoms such as N, B, O, S, and P, of a non-carbon element to graphene's carbon lattice. If functionalization is considered a non-selective technique, then doping is considered a precision technique. When performed correctly, dopants are typically added into graphene's defect sites, often substituting a carbon atom or filling a vacancy site. Doping into defect sites can improve electron transport across the structure, which provides the strongest binding between the doped site and defected graphene lattice [94]. While functionalization can also generally improve graphene's electron transport properties, doping specifically targets properties to enhance for a variety of applications.Graphene modification by functionalization or doping often is performed to achieve improved results for specific applications. Recently, a focus in catalysis has been to use modified graphene as a support for a variety of reactions. Metallic dopants are of particular interest due to the ability to act as catalysts or co-catalysts during reactions. Metals are excellent catalysts and single-atom catalysis is a promising field compared to traditional particle catalysis due to material savings potential. For example, Xi et Al. has recently demonstrated single-atom Pd doped graphene catalysts exhibiting high catalytic performance for two applications [95,96]. Thus, metal doped graphene can be promising single-atom catalysts or co-catalysts for a variety of applications (some examples present in Section\u00a04).By inducing defects on graphene, its catalytic properties and the stability of metal-support binding are enhanced; controlling the defect formation mechanism is imperative to improve the catalytic potential [97\u2013107]. Defects allow the active metal catalyst more direct access to graphene's electronic \u201chighway\u201d which can be tuned to modify the electron donor/acceptor property [17,18,108]. This is attributed to the bond formed between the metal particle at the defect site. Engineering graphene defect sites should be an important consideration when considering graphene as a support. As described earlier, there are a variety of modifications which can achieve desired results in a modified catalyst support. But how do those modifications change the physical nature of the material?The properties of graphene's electronic highway are derived from its chiral band structure. The 2D band structure of pristine graphene is confined by the carbon atoms and exhibits an effective bandgap of zero [15,22,23,56]. For catalysis, having a small bandgap is ideal since theoretically this means the resistance to electron excitation, or flow, is at a minimum. However, the act of depositing metal particles onto the surface changes graphene's band structure due to material interactions. Additionally, maintaining pristine graphene throughout the catalyst synthesis process is difficult and impractical. Modified graphene, containing defects, functionality, or dopants, is a far more practical starting support material for prepared metal nanoparticle catalysts.Modified graphene inherently has a different band structure compared to pristine graphene. For defective graphene, the modifications are the inherent defects which change how electrons move through the \u03c0-network. The type of defect, functional group, and dopant control or direct the band structure and ultimately change the band gap [7,15,22,63]. For example, GO has increased affinity to metal ions compared to pristine graphene, leading to increased metal uptake and potential particle nucleation sites based on the relative surface density of oxygen-containing functional groups [24,64,109-114]. This is true for all forms of modified graphene. Dopants are advantageous due to the potential for precision site engineering of band gap structures and particle nucleation site dispersion. The stresses applied during the production of modified graphene create structures and introduce chemical species that create sites to produce heterogeneous catalysts with desirable particle sizes, particle dispersion, and chemical activity [23,24,40,62,64,67,70,95,96,109,115-117].Considering the relatively low native catalytic activity of GO and rGO, a more practical endeavor is using these materials as catalytic supports. The most sought-after supports typically have high surface-areas and contain micropores, providing an abundance of deposition sites for metals and helping to control metallic particle sizes, respectively [118\u2013120]. Defect sites and oxygen functionality typically increase surface area leading to higher metal uptake, as well as mimic micropores, limiting particle size. The synthesis of these materials and catalysts can be thought of as the means to making the next generation of materials. Ultimately, the desired material changes must be made by a controlled synthesis method. How can these materials be made, and which method is optimal for the application? These questions will be addressed in the following section where specific focus will be made to address how the synthesis method itself plays a role in the creation process of modified graphene and the final catalyst product.Carbon is widely used as a support for metal nanoparticle catalysts due to its unique properties discussed above. Graphene forms of carbon have been used as a specialty support to take advantage of its structural properties. Specific applications are discussed in a later section. The synthesis methods used for preparing supported metal nanoparticles on graphene do not differ extensively from synthesis on regular carbon supports. Some of these common methods are discussed in this section.The most common way of preparing supported metal catalysts is by simple impregnation methods. The support is immersed in an aqueous solution containing a precursor of the metal, usually a dissolved salt (Fig.\u00a07\n). Immersion can be done with excess solution, called wet impregnation, or in incipient wetness mode, also known as pore-filling or dry impregnation [121,122]. Impregnated supports are thereafter recovered, dried, and thermally treated to generate the metal nanoparticles usually by heating under flowing gas that can be oxidizing or reducing. While simple impregnation methods can be cost effective in making graphene catalysts these typically produce large particles or agglomerates due to the poor interaction between metal precursor and support surface [123,124], thus having poor dispersion \u2013 the availability of metal surface sites relative to the amount of metal used in the catalyst. Wide size distribution of nanoparticles is also a usual feature of simple impregnation due to the mobility of the metal during nucleation. In some cases, the solution conditions during simple impregnation can result in high dispersion due to enhanced electrostatic adsorption of the metal precursor [125].The method of strong electrostatic adsorption (SEA) improves the metal dispersion in the product catalyst by enhancing precursor-support interaction [125,126]. SEA, essentially a special impregnation method, is an industrially applicable process easily done by soaking the support in a solution containing the metal precursor at the appropriate pH. A surface potential, or charge, is imparted on the support by protonation or deprotonation of surface functional groups with the pH of the impregnating solution at an optimum value, far from the support point of zero charge (PZC) (Fig.\u00a08\n). Oppositely charged precursor ions are then electrostatically attracted to the surface while retaining hydration sheaths [127,128]. Carbon supports are abundant in surface groups that can be protonated or deprotonated [129,130]. Graphene rings on carbon can accept protons when in contact with an acidified precursor solution, creating a positively charged surface which can adsorb anionic metal precursor complexes. These rings however cannot be deprotonated from neutral charge; thus, the surface cannot be negatively charged when in contact with a basified solution [125]. Terminating groups that can be protonated (or deprotonated) can be created by functionalization of carbon surface [130,131]. In cases where impregnation without intentional SEA synthesis on carbon support resulted in highly dispersed metal particles, the impregnating solution was highly acidic and far from the PZC of carbon, promoting electrostatic adsorption of the precursor [125].The use of SEA for synthesizing carbon or graphene supported catalysts has been demonstrated for efficient utilization of expensive noble metals, ensuring most of the metal atoms are available on the surface [132\u2013135]. In electrochemical applications, highly dispersed noble metal catalysts are sought as it corresponds to high electrochemical surface area which correlates with improved activity as shown in recent work using SEA for polymer electrolyte membrane fuel cell (PEMFC) [136,137]. In that work, catalysts prepared by SEA performed better than a comparable commercial catalyst due to higher metal dispersion shown by TEM imaging (Fig.\u00a09\n). In Suzuki cross coupling using SEA prepared Pd/graphene oxide catalysts, the strong interaction between metal and support was cited as beneficial to catalyst performance [18].Another method of supported nanoparticle catalyst preparation is by deposition-precipitation (DP) synthesis. Similar to wet impregnation, the support is immersed in a solution containing the metal precursor but with the addition of a precipitating agent (e.g. urea, sodium hydroxide, etc.) that causes particle nucleation [138]. The precipitation is induced by an increase in pH causing highly insoluble metal hydroxide to form and adhere to the support surface (Fig.\u00a010\n). Monitoring and controlling the pH is necessary to regulate the growth of particles. Depending on the pH and functional groups on the surface of the support, metal ion adsorption can happen and these can serve as nucleation sites for DP [139]. Much like SEA, the metal and support surface interaction during DP synthesis is credited for catalyst stability during evaluation [140].Continuous manufacturing technology presents a new frontier for heterogeneous catalysis synthesis wherein issues identified in batch processes can be minimized or eliminated. Batch-to-batch variability of continuous processes are significantly reduced compared to batch equivalent processes [141]. Product quality and consistency is paramount when synthesizing a product with a specified performance. In fact, the given tolerance associated with product performance can usually be narrowed when a continuous process is implemented. Heterogeneous catalysts have been difficult to completely produce in a single continuous manufacturing process due to limitations in handling slurries, controlling metal deposition and nucleation, and efficient product separation. Significant engineering considerations must be made in order to control parameters critical for manufacturing a heterogeneous catalyst.Examples of adapted thermochemical deposition methods for continuous synthesis have been populating journals over the past five years. Thermochemical methods are techniques suited for continuous processes due to their dependence on concentration, rapid mass- and heat-transfer, and time-dependent reaction progress. Continuous systems which contain plug-flow reactors and a laminar flow fluid are best suited for heterogeneous nanoparticle synthesis. While heterogeneous nucleation, nucleation of metal particles onto a material of different elemental composition, remain a challenge, controlled particle nucleation and growth and subsequent deposition onto graphene can be achieved. Smith, et Al. successfully synthesized core-shell particles for use in Fischer-Tropsch reactions by adapting a solvothermal procedure to a continuous plug-flow reactor [142]. A traditional plug-flow type reactor was used in the experiment to synthesize the copper based nanocomposite particles. Fig.\u00a011\n below is the schematic of the synthesis process.New reactor technology could further improve upon existing or enable new synthesis methods used for manufacturing heterogeneous catalysts at scale. One such system is the spinning disk reactor (SDR). SDRs are unique reactor systems which are rotor-stator setups with a spinning disk housed inside a stationary reactor. These reactors are typically characterized by their fast residence time, high sheer forces, and plug-flow configuration. They are an instrument for process intensification applied for pharmaceutical, particle, and polymer reactions to improve reaction kinetics at small volumes [143\u2013145]. Computational fluid dynamic and classical models demonstrate the micromixing efficiency and fluid hydrodynamic effects found within the reactor to demonstrate the enhanced mass- and heat-transfer associated with accelerated reaction kinetics. For nanoparticle synthesis reactions, SDRs can be applied to accelerate monomer formation to favor rapid nucleation and small particle formation. An example of one such setup can be found in Fig.\u00a012\n below.Rapid nucleation of metal monomer is desirable in these systems since the resulting particles or clusters tend to be very small [146,147]. The rate of nucleation can be controlled by the disk spin speed, disk gap spacing, and overall throughput (flow rate and concentration). Recently, examples of catalyst particle synthesis have been successfully demonstrated, such as copper particle catalysts for methanol synthesis [148]. The authors were able to control the particle size by varying the spinning speed of the reactor and each condition exhibited tight size distribution. Work being performed by the Gupton group with Pd has also demonstrated that the SDR can produce small Pd particles supported on graphene.Reduction treatments are often necessary after metal deposition occurs since a variety of methods do not simultaneously or sequentially deposit and nucleate the metal particles in the same reaction solution. Typically, reduction is handled by a tube furnace flowing hydrogen or using a reducing agent, such as hydrazine, in solution to reduce the metal ions on the carbon surface or in solution to form metal nanoparticles. Microwaves offer an alternative method for reduction since the heating mechanism is fundamentally different from conventional heaters. The heating rate of a material is directly proportional to the material's (molecule, particle, or atom) dipole moment [149\u2013152]. Metals in particular have high dipole moments, resulting in a bulk metal's surface temperature reaching close to 1000\u00a0K under ambient microwaving conditions [151,153-156]. Metal nanoparticles and dissolved metal ions or atoms will absorb microwave energy, often providing sufficient energy required for particle nucleation and growth.Microwaves also enable exploiting unique properties of materials undergoing transition states. Microwaves have been used to exfoliate pristine graphene from GO in solution [157]. In the case of graphene supported nanoparticles, microwaving graphene samples containing metal nanoparticles can form defect sites while also exfoliating the graphene. This effect has been demonstrated for palladium nanoparticular systems where the palladium catalyzes the formation of defect sites on the graphene [150]. The in situ defect formation allows the palladium to bind to a defect site allowing the metal better access to graphene's electronic network [17]. DFT calculations performed in that study corroborated this effect and explained this enhancement in catalytic activity [17,18,108]. Due to the stable bind between palladium and graphene, the leached metal content reduced compared to catalysts prepared by non-microwave methods [18,150]. Microwaves are a proven method to enhance catalytic activity as well as form solid-state ligand structures between metals and graphene. Microwaves have also been successfully applied in a flow application to synthesize heterogeneous materials [158].The unique properties of graphene and other specialty sp2 carbon that have been discussed in the previous chapters can be utilized when implemented as a catalyst or catalyst support in a reactive application. Enhanced catalytic performance can be attained by using an appropriate preparation method, to induce interaction between the specialty carbon support and the catalytic metal, or other heteroatom-doped element. The various applications cited in the succeeding sections provide examples of benefits brought about by these properties and interactions. These concepts may be expanded to other similar reactive applications.The oxygen reduction reaction (ORR) is the most crucial reaction for energy conversion devices such as fuel cells and metal-air batteries. However, the sluggish kinetics of ORR limits effective energy conversion, highlighting the necessity of active and stable electrocatalysts to enhance ORR performance [159]. Although Pt or Pt-alloy nanoparticles supported on graphitic carbons (Pt/C) more effectively catalyze ORR, the excessive cost, limited reserves, and low stability of Pt are still impeding the progress of fuel cells and metal-air batteries toward commercialization [160]. To reduce the catalyst cost, recent researches have focused on the reduction or replacement of Pt electrodes for ORR. Several experimental investigations revealed metal-free doped-graphene as promising electrocatalysts for ORR [16,161-164]. The rotating-disk (RDE) voltammograms studies by Qu et\u00a0al. showed that the specific current density at metal-free N-doped graphene electrodes is about three times higher relative to Pt/C electrodes at the potential between 0.4\u00a0V and 0.8\u00a0V (see Fig.\u00a013\na) [165]. In another ORR study, S-doped graphene exhibited 76% resistance to catalyst performance decay at the end of 6500\u00a0s stability test, which performed in O2-saturated KOH solution using current- time chronoamperometric response at \u22120.2\u00a0V, while graphene and Pt/C catalyst showed about 74% and 46% resistance respectively (see Fig.\u00a013b) [162]. Both these investigations indicate that heteroatom doped graphene materials have high activity as well as stability towards ORR.To eliminate the uses of Pt electrodes for ORR, many researchers are also developing non-precious metals (M: Co, Fe, Ni, Mo, Al, Cu, Sc, etc.) doped or M-X co-doped graphene materials to effectively catalyze ORR. One of the examples of such material is Fe and N co-doped 3D graphene (Fe-N/R3DG), which exhibits higher onset potential (Eonset\u00a0=\u00a00.98\u00a0V) and half-wave potential (E1/\n2\u00a0=\u00a00.82\u00a0V) compared to Pt/C catalyst (Eonset\u00a0=\u00a00.97\u00a0V and E1/\n2\u00a0=\u00a00.82\u00a0V) (see Fig.\u00a013c). Additionally, the durability of the Fe-N/R3DG catalyst was found higher since the current retention of Fe-N/R3DG catalyst was 89% after 20,000\u00a0s of continuous stability test relative to that of Pt/C catalyst (62%) (see Fig.\u00a013d) [166]. Similar results were obtained for Co and N co-doped [167\u2013169], Cu and N co-doped [170], Al and N co-doped [171], and Sc and N co-doped [172] graphene materials at ORR conditions.Fischer-Tropsch Synthesis (FTS) and other selective hydrogenation reactions represent important large, industrial scale hydrocarbon transformations. Much like common carbon supports (i.e. activated carbon, carbon black), graphene has been utilized as an effective support for metal catalyzed hydrocarbon synthesis. Graphene offers advantages over traditional carbon supports with its high specific surface area, inertness, and ease of functionalization. Cobalt supported on carbon has been widely studied for low temperature FTS, with deactivation of the catalyst being a major problem. When Co is supported on high surface area graphene, Co can be better dispersed thus producing smaller particles and giving higher availability of surface active sites [123]. Using graphene also decreased internal mass transfer limitations compared to carbon nanotube supported catalysts, with higher contact surface of the catalyst nanosheets, resulting in better activity and selectivity for FTS (Fig.\u00a014\n). Nitrogen functionalization of graphene surface with ammonia, prior to addition of Co, created nucleation sites for nanoparticle growth [124]. The increased anchoring strength between Co and graphene reduced the mobility of Co particles on the graphene which prevented sintering and deactivation. High activity, C8+ selectivity, and stability for FTS have also been reported on Fe/rGO catalyst prepared by microwave synthesis [173]. Similar to the aforementioned Co studies, the formation of defect sites during chemical reduction of GO as well as during microwave treatment provided nucleation sites for nanoparticle anchoring that enhanced the Fe/rGO catalyst stability. The electronic interaction of Fe with GO was cited as a factor contributing to enhanced activity affecting adsorbate binding and catalyst fouling.High selectivities under mild conditions have been reported for graphene supported catalysts applied to hydrogenation reactions. In one study, atomically dispersed Pd on graphene made by atomic layer deposition was applied to hydrogenation of 1,3-butadiene at 50\u00a0\u00b0C giving 100% selectivity to butenes with 95% conversion (Fig.\u00a015\n) [174]. High durability of the catalyst was recorded up to 100\u00a0h. The high selectivity was attributed to lack of Pd ensembles that fully hydrogenate butenes through a secondary reaction. In another study, atomically dispersed Pd/N-graphene was prepared via freeze-drying mediated synthesis [175]. Photothermal hydrogenation of acetylene to ethylene was achieved with a remarkable 99% conversion and 93.5% ethylene selectivity at 125\u00a0\u00b0C. Anchoring of the Pd atoms on the N-doped graphene was also credited for the high durability. Another study looked at the liquid phase hydrogenation of cinnamaldehyde using Pt/graphene catalyst [176]. Ethanol solvent was deemed best for the chemoselective reduction of the C=O bond, forming cinnamyl alcohol, achieving 73.9% conversion with 83.2% selectivity to the alcohol. This high selectivity was largely attributed to the high dispersion of the Pt catalyst on the graphene.Carbon-carbon cross-coupling transformations represent an important class of reactions for the preparation of complex organic molecules. Particularly, Pd catalyzed Suzuki couplings are of interest for pharmaceutical applications due to the mild reaction conditions and broad application across a wide substrate scope [17,18,108,150,177]. Heterogeneous Pd/graphene (Pd/G) systems have demonstrated a marketable competitive edge against commercial heterogeneous Pd/C catalysts and even some homogeneous Pd catalysts with added ligands [18]. For pharmaceutical applications, low metal contamination in the final product and high catalyst recyclability or lifetime is critical. Pd/G catalysts ensure little to no metal leaching off the catalyst and maintain high catalyst recyclability or lifetime compared to homogeneous catalysts [17]. This system achieves this through a unique set of synergistic properties, often dependent on the graphene and synthesis method.A Pd/G catalyst was prepared from co-reducing PdNO3 and GO with hydrazine in an aqueous solution using a microwave reactor. The catalyst yielded turn-over-frequencies (TOF) over 100000 hr\u22121 for a model Suzuki coupling reaction [9,17,18]. This high activity was attributed to the defect sites formed on graphene's surface during synthesis. This process was further optimized by changing to PdCl2 and performing SEA followed by microwave treatment. SEA was performed first to uptake the metal salt onto the support, then the dried material was microwaved to form defect sites and reduce the metal onto the support. The resulting material was highly active for cross-coupling reactions, demonstrating TOF over 200000 hr\u22121\n[150]. These results were remarkable on their own, however computation provided greater insight into why these materials exhibited such high activity.Density Functional Theory (DFT) calculations were performed to better understand the stability of these catalysts how these defect sites played a role in the catalytic mechanism. It was found that metals were strongly immobilized and stabilized on the graphene support when the metal was anchored to a defect site (see Fig.\u00a016\nA). The strongly anchored metal particles on defect sites were found to exhibit relatively high electron charge transfer properties compared to non-bound particles (see Fig.\u00a016B). Further, these strongly immobilized and stabilized metal particles were found to lower the activation energies of each step in the catalytic cycle compared to the non-supported metal particles (see Fig.\u00a016C). The significance of these findings demonstrates how metal catalysts behave as charge donors and acceptors to facilitate the catalytic mechanism [18,108]. When these particles are supported on defective graphene, the charge-transfer capabilities of the metal particles are enhanced which decrease their overall reaction activation energy barrier. This concept can be applied to other reactive applications other than Suzuki reactions.Statistical design of experiments (DoE) has long been a tool available to optimize industrial processes, minimize the number of conducted studies and trials, and avoid testing or analysis bias. DoE is typically used to determine causation where the relationships between the tested variables and measured responses may be difficult to ascertain. Conducting DoE effectively requires constructing simple, informative experiments about a known system such that the results are absent of systematic error, valid across a broad range of conditions, and estimates uncertainty to assert statistical significance. Additionally, DoE can enable advanced understanding of how these conditions can synergize together in a final effect that is more than the sum of the parts. The basic principles of any DoE are: 1) true replication of experimental conditions such that repeatability is verified and the random error variance is estimated; 2) randomized sampling and test order such that systematic errors are minimized from the study; and 3) blocking or partitioning into test subsets such that the precision and range of validity are maximized. This tool however has not been adopted widely by practitioners of heterogeneous catalysis and nonexistent in the development of new graphene catalysts. Fig.\u00a017\n below displays the rate of increase of mentions of \u201cheterogeneous catalysis\u201d in publications ranging from 1905 to 2021, and compares that with the subset of those papers that also mention \u201cDoE.\u201dIt is clearly seen that DoE comprises a very small portion, only 0.2% in 2020, of the methodologies utilized in the study of heterogeneous catalysis. Prior to 1980, there were no mentions of DoE in publications focused on heterogeneous catalysis. Looking towards the future, DoE principles should be applied to the field of heterogeneous graphene supported metal catalysis. This represents a fertile area of investigation considering DoE can be used to reduce development time for new catalysts, optimize catalyst synthesis conditions for desired property, improve material performance and robustness, and evaluate catalytic performance in industrial process. It is important to consider that any statistical based analysis must be compared with non-statistical based knowledge to ensure the results are meaningful to this field.Ali et\u00a0al. utilized central composite design (CCD) to optimize the one-step preparation of a reduced graphene oxide-titanium carbon nanotube (rGO-TNT) visible light catalyst [178]. CCD is a specific type of design that uses a polynomial model. The percent degradation of methylene blue dye (MB) was optimized by investigating the anodization time (hr) and voltage (V) with the ranges of 1\u20133 hr and 30\u201360\u00a0V, respectively. The design space included 13 experiments, and the model was fit to the second-order polynomial Eq.\u00a0(1) with coded factors\n\n(1)\n\n\ny\n=\n\nc\n0\n\n+\n\nc\n1\n\n\nx\nt\n\n+\n\nc\n2\n\n\nx\nV\n\n+\n\nc\n3\n\n\nx\nt\n\n\nx\nV\n\n+\n\nc\n4\n\n\nx\nt\n2\n\n+\n\nc\n5\n\n\nx\nV\n2\n\n\n\n\n\nwhere xt\n and xV\n represent the anodization time and voltage, respectively, and the remaining factors represent interacting variables. The study and statistical analysis determined all factors, including interacting factors, were significant and retained in the final model for the CCD experimental analysis [178]. Eq.\u00a0(1) was used to determine predicted MB degradation percentages and the values well matched the experimental values obtained. The data was then used to perform a response surface methodology (RSM) analysis in order to find a predicted set of optimal conditions. RSM analysis projects an experimentally derived dataset to find a predicted optimal outcome which may not be contained in the space bounded by the original designed experiments. These analyses generate contour and response surface plots which demonstrate how the factors affect the desired response (MB% degradation) across a continuum (see Fig.\u00a018\n).The 3D plot (Fig.\u00a018) demonstrated a maxima of MB% degradation response generated by the experimental design model from varying the voltage and time. Different colors in the plot indicated different ranges for the response where the orange levels show upwards of 90%. The contour plot (Fig.\u00a018) represents a 2D projection of the 3D response surface plot (Fig.\u00a018). The red and tan points on the plots represent the conditions performed in each individual experiment. The RSM analysis indicated that the optimal set of conditions a reaction with a duration of 2.06 hr and a voltage of 47.74\u00a0V would achieve a predicted MB removal of 90.1%. This value had good agreement with the experimentally determined value of 91.1% at the same conditions [178]. The good agreement between predicted and experimental values demonstrate the potential of DoE to accurately and precisely predict response values when a good experimental design was conducted.While the authors give no physical basis for the significance of an interaction between anodization potential and time, it seems reasonable that any interacting effect between voltage and time may be a result of the imparted strain on the RGO-TNT catalyst substrate. The generation of electron-hole pairs was identified as the limiting step in the reaction, and therefore the use of RGO in the catalyst should increase the rate of this step. This electron-hole pair was responsible for generating the radical species in the photocatalytic reaction to degrade the organics into oxidized products [178]. As discussed previously, imparting stress into graphene to cause strain will cause the electrons within the pi-structure to shuffle which can generate a charge. Upon removal of this stress, the graphene can rehybridize to return to its initial base state, much like electron-hole pair combinations. Under photocatalytic conditions in this experiment, it is reasonable to attribute the electron-hole generation necessary for MB degradation to this inherent property of RGO-TNT and why it is superior to native TNTs in this application.DoE can capture a system remarkably well and enable researchers to predict desirable conditions accurately and precisely for an experiment rather than laboriously performing an excessive number of experiments. However, DoE is currently underutilized in the field of heterogeneous catalysis as a whole and nonexistent for designing new graphene catalysts. The consequence of statistical analysis for independent factors and their potential interactions are important to discuss within the context of the mechanisms pertinent to the chemical system and will add value to a study. In the case of omission of a factor or interaction from a model due to statistical insignificance, the researchers need to explicitly explain their reasoning for the omission to improve the quality of results. Conversely, the researchers must scientifically rationalize why specific variables or interactions are statistically significant with reference to an underlying physical mechanism.Physical interactions, particularly those that are strong enough to give statistical signals, can make a traditional scientific analysis much more difficult, sans statistical quantifications of these interactions. One example of this complexity in heterogeneous supported-metal catalysis is the physical interaction between the metal and support, and analysis of statistical interactions can help elucidate the nature of this metal-support interaction in each system. There will also be situations where either an unexpected statistical interaction may be present, or where there is no statistical indication that there is an interaction where one expects there to physically be, such as the interaction or lack thereof between the support and the metal particle. Therefore, statistical analysis must be complimented by physical and chemical knowledge to obtain significant and meaningful results. This will elevate the discussion of both the statistics and the studied material or process and lead to new physical insights.The investigations of metal catalysts supported on graphene, especially functionalized or defective graphene, demonstrate the enormous potential for leveraging enhanced catalytic properties. In this review, we have discussed the recent advances in this field, elaborated on the nature of functionalized and defective graphene, how to synthesize these materials and create metal supported catalysts, and demonstrated real world examples leveraging these properties. Graphene functionalization and defect creation represent a critical area of opportunity to produce the next generation of highly active and selective catalysts. Through advanced synthesis techniques, such as SEA and the utilization of microwaves, the rational synthesis of metal-based catalysts on functionalized and defective graphene is achievable. To date, considerable advances in this area have been made.However, limitations of these materials still exist particularly in the area of graphene doping, functionalization, and defect creation at specific sites in a controlled manner. A considerable amount of work remains to better synthesize site specific functionalization and defect sites on graphitic carbon for their use as catalyst supports. Specific attention to advancing these methods used to create these special graphene materials in a manner which creates site specific modifications should be investigated. Further, advanced characterization methods in collaboration with computational methods are necessary to elucidating the effect of graphene functionality and defects on catalytic performance of metal particles. All the above work must be done to better understand how the metal-support interaction drives chemical catalysis. The nature of the support including dopants and defects, the metal type and quantity, preparation technique, and the application as a catalyst are all dependent on the metal-support interaction. This interaction will be inherently unique to any catalyst prepared due to the complexity involved in each step of the process when designing a catalyst. The future of designed graphene containing specific functional groups and defect sites is promising. Especially in the field of chemical catalysis, these materials used as supports could provide a significant advantage to activity and selectivity in industrial applications and replace current industrial catalyst in the coming years.M.B.B. and F.B.A.R. contributed equally to the preparation of this review article. M.B.B., F.B.A.R., J.M.M.T., and E.H.C. wrote this Review. J.M.M.T, E.H.C., J.R.R., and B.F.G. were involved in editing the manuscript.The authors declare no competing financial interest.The authors are grateful for financial support by the Center for Rational Catalyst Synthesis an Industry/University Cooperative Research Center funded in part by the National Science Foundation [Industry/University Collaborative Research Center grant IIP1464595]; Nanomaterials Core Characterization Facility at Virginia Commonwealth University.", "descript": "\n                  Transition metal-based heterogeneous catalysts are widely used across many industries. The prevalence of these materials across so many domains has inspired research into many different types of solid supports, the nature of which can affect catalytic performance. One support receiving increased attention because of its many desirable features is graphene. These features include 1) native catalytic properties enabling co-catalysis, 2) enhanced catalytic activity when both metal atoms and nanoparticles are supported, 3) chemical functionalization to tune catalytic properties, 4) tough lattice structure and high electric conductivity, and 5) specific solid-state ligand bond formation augmenting electron transport between graphene and the metal to name a few. Although graphene shows tremendous applicability in heterogeneous catalysis, researchers are still tuning the structure to improve its catalytic performance, such as by incorporating defects or dopants into its morphology. Another important consideration is the interaction between the graphitic support and metal catalyst particle, which in turn is highly dependent upon the nature and quality of the catalyst preparation technique. This work reviews the modification of graphene structure along with the applications of different modified graphene-supported catalysts. It also discusses some of the most used and efficient catalyst preparation techniques for both batch and continuous modes. Various examples of applications that highlight graphene properties and catalytic interactions are discussed. To strengthen our reviews, a set of statistical analysis is included.\n               "}
{"full_text": "The surge in global energy consumption, the depletion of reserves of fossil fuels and the growing concern about their harmful effects on the environment necessitates the search for alternative energy sources [1]. One of the best candidates among the possible alternative renewable and sustainable fuels is bio-oil, which can be obtained through the pyrolysis of biomass. Bio-oil is considered carbon-neutral: the carbon dioxide produced by burning bio-fuels is absorbed by plants, from which biomass and bio-oil can be obtained again; thus, the greenhouse gas emissions are significantly reduced compared to fossil fuels. However, bio-oil contains high amount of oxygen and unsaturated compounds, such as aldehydes, ketones, organic acids, phenols and their derivatives, which are thermally and chemically unstable [2]. In addition, these compounds have a low calorific value, high corrosiveness and can polymerize during storage, which limits the direct use of bio-oil as a fuel [3]. Therefore, bio-oil must be upgraded to be used as a substitute fuel or valuable chemical feedstock.One of the common ways to improve the quality of bio-oil is hydrodeoxygenation (HDO). Usually HDO is carried out in the presence of conventional hydrotreating catalysts \u2013 sulfides of CoMo and NiMo [4,5]. However, higher yields of deoxygenation can be obtained with noble-metal catalysts [6], including palladium (Pd), platinum (Pt), ruthenium (Ru), and rhodium (Rh), which retain their high activity at mild temperatures and pressures. Also, due to complexity of bio-oil composition, most studies are focused on the HDO of bio-oil model compounds \u2013 guaiacol, eugenol, phenol, anisole and vanillin [7]. Among them, vanillin occupies a special place, since it contains three oxygen-containing groups, namely: a hydroxy-, a methoxy- and a carbonyl-group. This makes vanillin an interesting substrate for studying the activity of catalysts in hydrodeoxygenation. Thus, it is possible to carry out the selective hydrogenation of vanillin to vanillin alcohol [8], or convert it to p-creosol via hydrogenation and hydrodeoxygenation [9], and obtain also cyclohexanol or cyclohexane.In addition to the nature of the metal used, the activity and selectivity of catalysts is significantly affected by the support used [10]. In particular, the chemical composition and morphology of the surface of support affect the properties of the metal nanoparticles, and hence the performance of the catalyst. As an example, one of the most active hydrodeoxygenation catalysts is metal nanoparticles supported on a carrier with Br\u00f8nsted acid sites [11], which facilitates hydrogenolysis of CO bond [12]. Such carriers can include: zeolites, metal-organic frameworks (MOFs), phosphonic acid modified porous materials, and N-doped carbons. The introduction of this bifunctionality is usually carried out by the modification of catalysts supports with acidic group [13], or using the materials which already had acidic properties [12]. At the same time, another urgent task is the selective hydrogenation of vanillin to vanillin alcohol \u2013 \u0430 valuable intermediate in the synthesis of novel flavorings and fragrances, and a potential building block in the synthesis of epoxy resins [14]. In this case, on the contrary, the hydrogenolysis of the CO bond must be avoided, which can be achieved either by changing the hydrogenation reaction conditions [15], or by changing the composition of the catalyst.A convenient tool for controlling the properties of the catalyst is the modification of the support with functional groups. Thus, phosphonic acids were used [16] to modify Pd/Al2O3 towards the application of a low-temperature liquid-phase vanillin hydrodeoxygenation. Modification allowed to enhance the yield of creosol from 2.5 to 87% at 50\u00a0\u00b0C due to the creation of metal/acid bifunctional sites. In another work, the authors carried out a modification of metal-organic framework UiO-66 with amino-groups. The obtained Pd@NH2-UiO-66 had superior performance in vanillin HDO due to the cooperation between the metallic Pd sites and the amine-functionalized MOF [17].Among other supports, porous carbon polymers have attracted an increasing level of researcher's interest. These materials can integrate the advantages of both porous materials and polymers. Their main advantages over porous carbons, which were used by other authors for vanillin hydrodeoxygenation [8,9], are the greater variety of synthesis and modification methods. Porous polymers have high chemical and physical customization provided by the versatility of organic chemistry. Moreover, various functional groups can be incorporated into porous polymers by pre-synthetic and post-synthetic modifications. Compared with zeolites and MOFs that are relatively highly sensitive to acidic or basic conditions [16,17], porous polymers generally have high chemical stability. An example is the porous aromatic frameworks (PAFs), which are widely used in catalysis [18]. PAFs have rigid structure consisting of aromatic rings connected to each other [19]. They are attracting more and more the attention of researchers due to their high surface area, the possibility of varying their pore size, high thermal and mechanical stability. The aromatic structure of PAFs contributes to the stabilization of metal nanoparticles, and also makes it relatively easy to modify catalysts with functional groups [20]. Catalysts based on metal nanoparticles in the pores of PAFs are promising for the conversion of lignin and related molecules. In the current work, we investigated vanillin hydrogenation on palladium nanoparticles, supported on three porous aromatic frameworks: PAF-30 without functional groups; PAF-30-SO3H, modified with sulfo-groups; and PAF-30-TEA modified with thiethylamino-groups.The following reagents were used in the present work: Methanol CH3OH (Component-Reactiv, high-purity grade); Ethanol C2H5OH (Component-Reactiv, high-purity grade); Isopropyl alcohol (CH3)2CHOH (Component-Reactiv, high-purity grade); Tetrahydrofuran C4H8O (Component-Reactiv, high-purity grade); Diethyl ether (Component-Reactiv, high-purity grade); Dichloromethane CH2Cl2 (Component-Reactiv, high-purity grade); Acetic acid CH3COOH (Ruskhim, high-purity grade); Chlorosulfonic acid HSO3Cl (99%, Sigma-Aldrich); Triethylamine (Sigma\u2013Aldrich, St. Louis, MO, USA, 98%); Sodium hydroxide NaOH (Reakhim, 99%); Palladium (II) acetate Pd(OAc)2 (Sigma\u2013Aldrich, 98%); Potassium hexachloropalladate (IV) K2PdCl6 (Sigma\u2013Aldrich, 98%); Sodium borohydride NaBH4 (Aldrich, 98%); Vanillin C8H8O3 (Rushim, 99%); Salicylaldehyde C7H6O3 (ACROS Organics, 99%); p-Anisaldehyde C8H8O2 (ACROS Organics, 99%), p-toluenesulfonic acid (Sigma-Aldrich, 99%); tetramethylammonium hydroxide pentahydrate (Sigma-Aldrich, \u226597%).Porous aromatic framework PAF-30 was prepared by Suzuki cross-coupling reaction between tetrakis-(p-bromophenyl)methane and 4,4\u2032-biphenyldiboronic acid according to previous published procedure [21]. Sulfonation of PAF-30 was carried out using a solution of chlorosulfonic acid in dichloromethane [22]. Chlorosulfonic acid (167\u00a0\u03bcl) was added dropwise to the suspension of PAF-30 (500\u00a0mg) in dichloromethane (25\u00a0ml) at 0\u00a0\u00b0C, and the resulting mixture was stirred at room temperature for 24\u00a0h. After completing the reaction, the suspension was poured into ice, the solid product of PAF-30-SO3H was filtered, washed with water, THF, diethyl ether and finally dried in vacuum.Synthesis of PAF-30-TEA was performed in two steps. First, chloromethylation of PAF-30 was carried out according to the method reported earlier [23]. The resulted material PAF-30-CH2Cl (300\u00a0mg) was refluxed with triethylamine (30\u00a0ml) for 3\u00a0days. The solid product was then collected by filtration, washed with 1\u00a0M NaOH (50\u00a0mL), water, ethanol, and finally dried in vacuum. The obtained material PAF-30-[CH2NEt3]+OH\u2212was named PAF-30-TEA.Immobilization of palladium on supports included the impregnation of materials with metal salts and their reduction with NaBH4. The choice of palladium salt and solvent depended on the type of carrier used. Thus, impregnation of PAF-30 was carried out from a solution of Pd(OAc)2 in chloroform, PAF-30-SO3H \u2013 from a solution of Pd(OAc)2 in methanol, and PAF-30-TEA \u2013 from a solution of K2[PdCl6] in methanol. By the general procedure, 100\u00a0mg of PAF was stirred with 10\u00a0mL of 1.9\u00a0mmol/L solution of palladium salt at room temperature for 24\u00a0h. Then, a solution of 80\u00a0mg of sodium borohydride in 5\u00a0mL of water-methanol mixture (1:1\u00a0mL) was added dropwise to the resulting suspension. The reaction mixture darkened, and then stirred for 12\u00a0h. After reaction, the resulting solid product was collected by centrifugation and washed with ethanol (2\u00a0\u00d7\u00a050\u00a0mL), water (2\u00a0\u00d7\u00a050\u00a0mL) and THF (2\u00a0\u00d7\u00a050\u00a0mL). Also, the catalyst based on PAF-30-SO3H was pre-washed with acetic acid (50\u00a0mL) to remove residual Na+ cations.Nitrogen adsorption isotherms were measured on a Micromeritics Gemini VII 2390 instrument (Micromeritics, Norcross, GA, United States). All samples were degassed at 120\u00a0\u00b0C for 8\u00a0h before analysis. The specific surface area (SBET) was calculated using the Brunauer\u2013Emmett\u2013Teller (BET) method based on adsorption data in the relative pressure range of P/P0\u00a0=\u00a00.05\u20130.25. The total pore volume (Vtot) was determined by the amount of nitrogen adsorbed at the relative pressure of P/P0\u00a0=\u00a00.965.IR spectra were recorded with a Nicolet IR200 (Thermo Scientific) instrument using multiple distortion of the total internal reflection method with multi-reflection HATR accessories, containing a 45\u00b0 ZnSe crystal for different wavelengths with a resolution of 4\u00a0cm\u22121 in the range of 4000\u2013400\u00a0cm\u22121. All spectra were recorded by averaging 100 scans.Chemical composition (compositional weight percentage of carbon, hydrogen, palladium, sulfur and nitrogen) was determined using a CHNS elemental analyzer (Thermo Flash 2000) located in the Center for Collective Usage \u201cAnalytical Center for the Problems of Deep Refining of Oil and Petrochemistry\u201d of A.V. Topchiev Institute of Petrochemical Synthesis, RAS.Transmission electron microscopy (TEM) images were obtained using a JEOL JEM-2100/Cs/GIF microscope (JEOL, Tokyo, Japan) with a 0.19\u00a0nm lattice fringe resolution and an accelerating voltage of 200\u00a0kV. The processing of the micrographs and the calculation of the average particle size were conducted using the ImageJ software program. Palladium and nitrogen localization in the catalyst Pd-PAF-30-TEA was investigated using energy dispersive X-ray (EDX) analyzer (EX-24065JGT).The acidity measurement was carried out by acid-base titration. The acid catalyst was dispersed in a standard solution of NaCl (0.01\u00a0mol/L), and then in a standard solution of NaOH (0.01\u00a0mol/L) as titrant. Acid-base potentiometric titration was carried out using a PH-009(II) pH meter with the following characteristics: pH measurement ranges from 0.00 to 14.00; resolution 0.01 pH; accuracy \u00b10.01 pH.Hydrodeoxygenation of vanillin, salicylaldehyde and anisaldehyde was carried out in a stainless-steel batch reactor, equipped with glass tube and magnetic stirrer. An amount of 2\u00a0mg of the catalyst and 1.5\u00a0ml of 0.173\u00a0M substrate solution in isopropyl alcohol or water were loaded in the tube, which was then placed in the reactor. Reactions were carried out for 0.25\u201318\u00a0h at a hydrogen pressure of 10\u00a0bar and in the temperature range of 60\u201380\u00a0\u00b0C. After completion of the reaction, the autoclave was cooled to room temperature and depressurized. Reaction products were analysed by gas chromatography. All experiments were performed at least twice, the experimental error doesn't exceed 5%.The surface of the PAF-30 was modified with -SO3H and \u2013[CH2NEt3]OH groups, resulting in PAF-30-SO3H and PAF-30-TEA materials as shown in Fig. 1\n. Introduction of functional groups reduced the free volume of mesopores, which could be seen by the decrease of the distance between adsorption and desorption curves in modified materials (Fig. 2\n). Thus, the surface area and pore volume of PAF-30 decreased from 489 m2/g and 0.241\u00a0cm3/g to 428 m2/g and 0.192\u00a0cm3/g, respectively, in PAF-30-SO3H, and to 427 m2/g and 0.135\u00a0cm3/g in PAF-30-TEA (Table 1\n), respectively. Adsorption isotherms for all materials are typical for porous aromatic frameworks, contain steep N2 uptakes at low pressures (P/P0\u00a0=\u00a00\u20130.05), and a hysteresis loop between adsorption and desorption curves. This indicates that all PAFs have micro-mesoporous nature structure.According to the elemental analysis (Table 1), PAF-30-SO3H contains 7.39\u00a0wt% of sulfur, which corresponds to an acidity of 2.41\u00a0mmol/g. PAF-30-TEA contains 0.55\u00a0wt% of nitrogen, and the concentration of \u2013[CH2NEt3]+ was 0.39\u00a0mmol/g. FTIR spectrum of the PAF-30-SO3H material (Fig. S1, ESI) contains characteristic absorption bands at 1370, 1135\u20131221, 1034, 901, 221,610\u00a0cm\u22121 for sulfo-groups [22,24], which confirms the successful modification of PAF-30. In contrast, the FTIR spectrum of PAF-30-TEA is almost identical to the spectrum of parent PAF-30 material. On the one hand, this might be due to a lower concentration of functional groups in PAF-30-TEA compared to PAF-30-SO3H. On the other, the absorption band at 1637\u00a0cm\u22121, characteristic for -NEt3\n+ groups [25,26], has low intensity.The immobilization of palladium nanoparticles was performed by the wetness impregnation method using a solution of Pd(OAc)2 or K2[PdCl6] with subsequent metal reduction with sodium borohydride. According to the ICP-AES, the palladium content in Pd\u2013PAF-30-TEA was 1.41\u00a0wt%, while in Pd\u2013PAF-30 and Pd-PAF-30-SO3H it was almost two times less \u2013 0.84 and 0.74\u00a0wt%, respectively. This could be due to the presence of triethylamine groups in PAF-30-TEA which can readily bind tetrachloropalladate species by an ion exchange mechanism [27,28].The size of palladium nanoparticles (Pd NPs) and their distribution were studied by TEM (Fig. 3\n). For the Pd-PAF-30 catalyst, Pd NPs have a broad size distribution with main maxima at 1\u20131.5\u00a0nm and 3\u20134\u00a0nm. Smaller particles were encapsulated in the pores of the support, while larger particles with an average diameter of 4\u20134.5\u00a0nm were distributed mainly on its external surface. Nonetheless, the number of small particles is greater than that of large ones. In contrast, particle size distribution curves for Pd\u2013PAF-30-SO3H and Pd-PAF-30-TEA are close to normal and have maxima at 3.5\u00a0nm and 4\u00a0nm, respectively. Energy dispersive X-ray spectroscopy (EDX) of Pd-PAF-30-TEA also confirms the uniform distribution of both the metal and functional groups in the catalyst (Fig. 4\n).All catalysts were tested in the reaction of vanillin hydrodeoxygenation (Fig. 5\n), which includes sequential hydrogenation of the carbonyl group and hydrogenolysis of the CO bond in the resulting -CH2OH group. In the case of Pd-PAF-30, the complete conversion of vanillin into products, vanillyl alcohol (75%) and p-creosol (25%), was observed already after 30\u00a0min of the reaction. With an increase in the reaction time, the yield of creosol increased significantly. For 1\u00a0h of reaction, it was found to be 93%, and complete conversion of vanillin into creosol occurred within 2\u00a0h. Such high activity of Pd-PAF-30 can be associated with the small size of palladium nanoparticles (1\u20131.5\u00a0nm).Pd-PAF-30-SO3H showed the greatest vanillin hydrodeoxygenation activity, where the creosol yield for 30\u00a0min on Pd-PAF-30 was 25%, and on Pd-PAF-30-SO3H was 85% (Fig. 6\n). The presence of acidic -SO3H sites accelerated deoxygenation to produce the desired p-creosol, which had also been observed in many studies [16,29,30]. The exact mechanism by which acidic groups are involved is still debated. It should be mentioned that the rate of hydrogenolysis of CO bond reaction depends not only on the presence and amount of Br\u00f8nsted acids [17,31], but also on reaction temperature, solvent polarity [15,32], metal dispersion [33], concentration of oxygen functionalities of support (for carbon supports) [8], and electronic properties of palladium surface [34]. However, we are close to the point of view of the authors of work [31], according to which acidic groups can affect the adsorption of [H] on the support near the metal-support interface, transfer charge to Pd particles, or acid groups are directly involved in the hydrogenolysis of the CO bond.In contrast, the reaction over Pd-PAF-30-TEA gave vanillyl alcohol as the main product, and the hydrogenation rate was lower than over Pd-PAF-30 and Pd-PAF-30-SO3H even though the palladium content was higher. The reason for the such high selectivity of Pd-PAF-30-TEA for vanillyl alcohol might be associated with the blocking of active sites on palladium nanoparticles with alkylammonium groups [35], or with the interaction of counterions of \u2013CH2[NEt3]+ groups with the surface of palladium nanoparticles with changing the adsorption of substrate and hydrogen. Also, this could be related with the reduced acidity of this catalyst.In order to confirm the influence of functional groups, we conducted experiments with Pd-PAF-30 and monomeric analogues of PAF-30-SO3H and PAF-30-TEA \u2013 p-toluenesulfonic acid and tetraethylammonium hydroxide (Table 2\n). Reaction time was short (15\u00a0min) for Pd-PAF-30-SO3H catalyst in order to achieve a high content of vanillyl alcohol in the reaction products, and to clearly show the effect of sulfo-groups. On the contrary, for Pd-PAF-30-TEA catalyst, reaction time was long (2\u00a0h) to show a decrease in the rate of deoxygenation of vanillin alcohol in the presence of alkylammonium ions.Vanillin hydrogenation on Pd-PAF-30 for 15\u00a0min gives 75% of vanillyl alcohol and 24% of creosol. However, even a small amount of p-TsOH accelerates deoxygenation of vanillyl alcohol to creosol, leading to a formation of 80% of creosol and 20% of vanillyl alcohol. The same reaction products \u2013 15% of vanillyl alcohol and 85% of creosol \u2013 were obtained on Pd-PAF-30-SO3H catalyst, which proves the effect of sulfo-groups in PAF-30 on the hydrodeoxygenation activity of the catalyst. Nevertheless, there seems to be no direct interaction between Pd nanoparticles and nearby acidic sites in the Pd-PAF-30-SO3H catalyst, since the addition of p-TsOH also increases the creosol yield observed.When the reaction was carried out for 2\u00a0h on the Pd-PAF-30 catalyst, complete conversion of vanillin into creosol was observed. The introduction of tetraethylammonium hydroxide completely changed the composition of the reaction products: vanillyl alcohol became the predominant product (85%), while creosol was present only in trace amounts. Similar results were observed for the catalyst Pd-PAF-30-TEA. The lower selectivity of the formation of vanillyl alcohol might be associated with an insufficient number of \u2013[CH2NEt3]+ groups in the material, and the presence of palladium nanoparticles on the surface of which these groups are absent.Pd-PAF-30-TEA was also tested in the hydrogenation of vanillin at different temperatures (Fig. 7\n). As expected, an increase in the process temperature led to a growth in the rates of both hydrogenation and hydrogenolysis reactions. Thus, an increase in temperature by 10\u00a0\u00b0C reduced the time of complete hydrogenation of vanillin by about 1.5\u20132 times. The appearance of the kinetic curves of vanillin consumption is characteristic for first-order reactions, and is typical for this reaction. Creosol accumulation curves obey more complex kinetic relationships: at 60\u00a0\u00b0C the reaction rate is near to be constant, and at 70 and 80\u00a0\u00b0C it is initially high and then decreases and also becomes constant. However, despite the slowdown of vanillyl alcohol hydrodeoxygenation reaction, this does not cease: the yields of creosol after 68\u00a0h of reaction at 60, 70 and 80\u00a0\u00b0C were 45, 51 and 61%, respectively.The stability of all catalysts was studied in reuse experiments (Figs. 8 and 9\n\n). Pd-PAF-30 and Pd-PAF-30-SO3H catalysts gradually lost their activity, probably due to metal leaching from the external surface of PAF particles or from their pores, which is confirmed by a decrease in the content of palladium to 0.41 and 0.38\u00a0wt%, respectively. Also, in the case of Pd-PAF-30-SO3H, a slight decrease in the concentration of sulfo-groups (<0.5%) was observed after 6 recycling experiments, and isopropyl-vanillyl ether was present in the reaction products starting from the second cycle. In contrast, Pd-PAF-30-TEA shows great stability at least six times without significant loss of activity and selectivity to vanillyl alcohol. The decrease in palladium content was insignificant, and final metal content was 1.37\u00a0wt%.We have also investigated the activity of catalysts in the hydrogenation of some other aromatic carbonyl compounds \u2013 anisaldehyde and salicylaldehyde (Fig. 10\n). As before, during the reaction the substrates undergo hydrogenation and further deoxygenation. Also, in the case of vanillin, the formation of a small amount of vanillyl-isopropyl ether was observed.Complete conversion of all substrates was observed on Pd-PAF-30-SO3H and Pd-PAF-30 catalysts (Table 3\n). As expected, mainly deoxygenation products were formed on the Pd-PAF-30-SO3H catalyst due to the presence of Br\u00f8nsted acid sites. In the case of Pd-PAF-30, hydrogenation of both vanillin and salicylic aldehyde led to the formation of approximately equal amounts of hydrogenation and hydrodeoxygenation products, whereas the reaction with anisaldehyde gives p-methoxytoluene as the only product. When Pd-PAF-30-TEA was used as a catalyst, the main product of anisaldehyde hydrogenation was also p-methoxytoluene, while hydrogenation of vanillin and salicylaldehyde gave the corresponding alcohols with high selectivity. However, the conversion of these substrates did not reach 100%. Due to the fact that all these aromatic aldehydes contain electron-donating groups (EHOMO values are \u22120.2453, \u22120.2482 and\u00a0\u2212\u00a00.2551 for vanillin, anisaldehyde and salicylic aldehyde respectively [34]), we suppose that this difference in catalytic activity is largely related to the different strength of adsorption of molecules on the surface of palladium nanoparticles, and the rate of their interaction with adsorbed hydrogen. Thus, the rates of hydrogenation and deoxygenation reactions over Pd-PAF-30-SO3H are so high that after the adsorption of substrates on the palladium surface, their conversion into deoxygenated products occurs very quickly. In the case of Pd-PAF-30, the substrates are rapidly hydrogenated to the corresponding alcohols, and then desorbed from the palladium surface. The rate of further deoxygenation depends on how easily alcohol molecules will be re-adsorbed on the surface of nanoparticles. Anisaldehyde, due to the absence of steric hindrance, diffuses to palladium nanoparticles and is easily adsorbed on their surface, which leads to a high yield of the deoxygenation product. Salicylaldehyde molecule contains -OH group in the ortho position, and hydrogen atom in it forms a hydrogen bond with the oxygen of the carbonyl group. This makes the adsorption of salicylaldehyde on the surface of palladium nanoparticles more difficult, due to which the hydrogenation and hydrodeoxygenation rates of this substrate become lower. Vanillin is the most \u201cbulk\u201d molecule, and its diffusion to the active sites of the catalyst is expected to be slower than for the other substrates. This also applies for the Pd-PAF-30-TEA catalyst: hydrogenation of anisaldehyde proceeds much faster than vanillin and salicylic aldehyde due to the absence of diffusion or adsorption restrictions and the resulting p-methoxybenzyl alcohol, than being converted into p-methoxytoluene. In contrast, hydrogenation of vanillin and salicylic aldehyde gives 50\u201380% conversion with >95% selectivity to the corresponding alcohol.Considering that water is a desirable green solvent for chemical transformations, we also investigated the activity of the catalysts and results are given in Table 4\n. Pd-PAF-30-SO3H exhibits the same activity in water as in isopropyl alcohol, while Pd-PAF-30-TEA demonstrates even higher deoxygenation activity. However, the activity of Pd-PAF-30 in water decreased \u2013 probably due to lower hydrophilicity of aromatic PAF-30 and poorer dispersion of the catalyst in reaction media. Nevertheless, we can conclude that the synthesized catalysts could be used in environmentally friendly process of bio-components processing.Three supported palladium catalysts based on porous aromatic frameworks with different composition of the surface were tested towards vanillin hydrogenation. All catalysts showed high hydrogenation activity, but the other properties \u2013 stability and activity in deoxygenation \u2013 depend on the structure of the PAF. Thus, the Pd-PAF-30 exhibited both moderate deoxygenation activity and stability, producing vanillin alcohol and creosol as reaction products. Modification of PAF-30 with sulfo-groups greatly enhanced the activity of the catalyst in deoxygenation, but reduces its stability. In contrast, modification of PAF with alkylammonium groups inhibited the catalytic activity in deoxygenation, making the Pd-PAF-30-TEA catalyst selective in the hydrogenation of aromatic aldehydes to alcohols. Apparently, this selectivity is associated more with the creation of steric restrictions for the adsorption of substrate molecules or with the reduced acidity of this catalyst. The observed effects of the influence of sulfo- and alkylammonium groups were also confirmed by reactions with PAF-30 and toluenesulfonic acid and tetraethylammonium hydroxide. The simplicity of PAFs modification methods and the very high chemical stability, makes it easy to tune the properties of the obtaining catalysts without loss of stability and activity. The results of catalytic tests showed that using PAFs as a support, one can easily tune the activity and selectivity of supported palladium catalysts by changing the surface composition of the porous aromatic framework.\nM.A. Bazhenova: Investigation, Writing \u2013 original draft. L.A. Kulikov: Project administration, Writing \u2013 review & editing, Visualization. Yu.S. Bolnykh: Investigation, Visualization. A.L. Maksimov: Formal analysis, Methodology. E.A. Karakhanov: Supervision.The authors declare no competing financial interest.This work was financially supported by the Russian Science Foundation (RSF) grant (project \u2116 20-19-00380).\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106486.", "descript": "\n                  The aim of the current work was to study the activity of supported palladium catalysts based on porous aromatic frameworks (PAF) in vanillin hydrogenation, and to develop the methods for controlling their activity and selectivity by modifying the PAF structure with various functional groups. Using unmodified PAF-30, its sulfonated derivative PAF-30-SO3H, and that of PAF-30-TEA modified with CH2N+(C2H5)3 groups, palladium catalysts were synthesized and tested for HDO of lignin components. This investigation provides one potential route for the development of efficient catalysts for hydrodeoxygenation by the adjustment of the composition of the catalyst and its appropriate derived properties.\n               "}
{"full_text": "Data will be made available on request.As is well-known, carbon dioxide (CO2) is the major cause of global climate change, mainly originating from the burning of fossil fuels such as coal and petroleum [1\u20133]. Meanwhile, CO2 is an abundant resource, economically attractive, nontoxic, and a renewable C1 carbon source [4\u20136]. As shown in \nFig. 1, chemical transformation of CO2 is an exciting way to reduce CO2 concentration.The structure of DMC contains some functional groups such as methyl (CH3-), methoxy (CH3O-), carbonyl (-C(O)-) and carbonyl methoxy (CH3O-C(O)-) [8,9], which can substitute toxic dimethyl sulfate and phosgene in methylation and carbonylation reactions. It is also widely used as a fuel additive [10], organic solvent [11] and electrolyte in lithium-ion battery [12], etc. Various DMC synthesis routes include methanol phosgenation, oxidative carbonylation of methanol, transesterification [13], alcoholysis of urea, direct synthesis from CO2 and methanol, etc. However, some of these processes use toxic, corrosive, flammable and explosive gases such as phosgene, hydrogen chloride, carbon monoxide. Due to the cheap raw material, the avoidance of high toxic reagents, as well as the direct utilization of greenhouse gas CO2, the direct synthesis of DMC from CO2 and methanol is promising.As previously reported, a number of homogeneous and heterogeneous catalysts including ionic liquids [14], alkali carbonates [15], transition metal oxides [16,17], heteropoly acid catalysts [18], and supported catalysts [19] have been investigated. Studies over the past two decades have provided important information on metal oxides; particularly about CeO2 and ZrO2 are the dominating catalysts for the direct synthesis of DMC. Bell et al. [20,21] investigated the mechanism of DMC for formation over zirconia using in situ infrared spectroscopy, and then established that the acid-base sites of the catalyst surface played a decisive role in the direct synthesis of DMC from CO2 and methanol. However, its catalytic performance is still hampered by low specific surface area and insufficient exposure of active sites.Metal-organic frameworks (MOFs) owing inherent large specific surface areas, high porosity and the flexibility of structure and properties design, have been widely applied to catalytic processes [22\u201326]. In previous studies, a Zr-based metal-organic frameworks catalyst UiO-66-X (X refers to the molar equivalent of trifluoroacetic acid modulator relative to terephthalic acid) was effective for the synthesis of DMC from CH3OH and CO2\n[2]. The addition of trifluoroacetic acid (TFA) could increase the amount of active sites (Lewis acidic site, Lewis basic site and terminal hydroxyl), and provide higher specific surface area (>1479\u2009m2 g\u20131) and highly developed pore structure. The UiO-66\u201324 catalyst showed excellent catalytic activity (0.17\u2009mmol\u2009g-cat\u20131 h\u20131) compared with the ZrO2 (0.03\u2009mmol\u2009g-cat\u20131 h\u20131) for direct synthesis of DMC. On this basis, Zr-based metal organic frameworks catalyst MOF-808-X (X refers to the molar ratio of ZrOCl2\u00b78H2O/1,3,5-benzenetricarboxylic acid) was synthesized and used as the catalyst for this reaction [27]. MOF-808\u20134 showed the best activity with almost no redundant BTC or zirconium clusters, which outperformed previously reported Zr-based metal-organic frameworks catalyst UiO-66\u201324 [28]. Based on the results, HPW encapsulated inside the micropore of MOF-808 matrix (HPW@MOF-808) and HPW mainly aggregated on the outside surface of MOF-808 matrix (HPW/MOF-808) were prepared and used for the direct synthesis of DMC from CH3OH and CO2\n[2]. HPW@MOF-808 exhibited higher activity than UiO-66\u201324. Hence, MOFs exhibited excellent catalytic for the reaction system. However, the hydrothermal stability of MOF-808 was not reliable.In order to quantitatively regulate the composition of Lewis acid-base sites on the catalyst surface UiO-66, a series of cerium modified Ce-UiO-66-X (X refers to the millimole of Ce doping) catalysts for the synthesis of DMC from CO2 and CH3OH were synthesized through cationic modification, which has not been reported in the literature. Then, the relationship between the Ce doping amount and the amount of acid and base sites of the catalyst was explored. Furthermore, the effect of 2-cyanopyridine dehydrating agent was investigated at the optimal reaction conditions. Finally, a possible reaction mechanism was deduced based on the characterization results.Zirconium chloride (ZrCl4, 98%), 1,4-benzenedicarboxylic acid (BDC, 99%), cerium nitrate hexahydrate (Ce(NO3)3\u00b76\u2009H2O, 99.5%) were purchased from Aladdin Industrial Inc. (Shanghai, China). N,N-dimethylformamide (DMF, 99.5%), methanol (99.5%), 1-Pentanol (99%) and dimethyl carbonate (DMC, 99%) were obtained from Deen Chemical Co. (Tianjin, China). Carbon dioxide (CO2, 99.9%), ammonia (NH3, 99.9%) and helium (He, 99.99%) were purchased Yuanzheng Gas Co. (Henan, China). All chemicals were used without further purification.UiO-66 was synthesized by a hydrothermal method described in the literature [26,29], with slight modification. Brie\ufb02y, 1,4-benzenedicarboxylic acid (0.830\u2009g, 5\u2009mmol) and zirconium (IV) chloride (1.165\u2009g, 5\u2009mmol) were added in N,N\u2010dimethylformamide (30\u2009mL). The mixture was stirred at room temperature for 1\u2009h. Then the obtained mixture was transferred to a Teflon-lined autoclave and heated at 120\u2009\u00b0C for 24\u2009h. After cooling in air to room temperature, the white solid was filtered off, washed sequentially three times with DMF and methanol, and finally dried at 150\u2009\u00b0C.The Ce-UiO-66 catalysts synthesized in the same way as described above, except that a certain amount of Ce(NO3)3\u00b76\u2009H2O (1.5\u2009mmol, 2\u2009mmol, 3.5\u2009mmol or 5\u2009mmol) was added to the above precursor ZrCl4/BDC solution and stirred at room temperature for 1\u2009h. The Ce/Zr content is 1.5\u2009mmol/3.5\u2009mmol (Ce-UiO-66\u20131.5), 2\u2009mmol/3\u2009mmol (Ce-UiO-66\u20132), 3.5\u2009mmol/1.5\u2009mmol (Ce-UiO-66\u20133.5), 5\u2009mmol/0\u2009mmol (Ce-UiO-66\u20135) respectively. The remaining solvothermal and cleaning processes were identical to that of UiO-66.Meanwhile, in order to compare the effect of different synthesis methods on the catalytic performance, the Ce-UiO-66\u20132-IM and Ce-UiO-66\u20132-IE catalysts with the same Ce/Zr molar ratio as Ce-UiO-66\u20132 (ICP-AES analysis) were synthesized by impregnation and ion-exchange methods respectively (Supporting Information).The X-ray diffraction (XRD) patterns were recorded on a Bruker D8 Advance diffractometer operated at 40\u2009kV and 40\u2009mA with Cu-K\u03b1 radiation. The scan of diffraction angle was in a range of 5\u201380\u00b0 (2\u03b8) at a scanning speed 0.05\u00b0/s.The specific surface area was measured by nitrogen adsorption at 77\u2009K using the Brunauer-Emmett-Teller (BET) method on a Autosorb-iQ-MP-C. Prior to measurements, the samples were desorbed at 150\u2009\u00b0C for 12\u2009h using a Belprep vacuum instrument. The pore size distribution of the samples was evaluated by the Density Functional Theory (DFT) method. The total pore volume (Vp) was estimated from the volume of nitrogen adsorbed at a relative pressure of 0.99.The infrared analysis of as-prepared materials has been performed using Fourier transform infrared spectroscopy (FT-IR), Vertex 70, in the range of 400\u20134000\u2009cm\u22121.Scanning electron microscopy (SEM) analysis was performed using a JSM-7900\u2009F equipment. Transmission electron microscopy (TEM) and energy dispersive X-ray spectroscopy (EDS) measurements were performed on a JEM-2100\u2009F equipment operated at 200\u2009kV.Thermogravimetric analysis (TGA) was performed using a NETZSCH thermogravimetric analyzer with a heating rate of 10\u2009\u00b0C\u2009min\u22121 from 25 to 700\u2009\u00b0C under air environment.The acid-base properties of the samples were studied by temperature-programed desorption of ammonia (NH3-TPD) and temperature programed desorption of carbon dioxide (CO2-TPD). Typically, 100\u2009mg of samples were used for each measurement. The sample was pretreated at 300\u2009\u00b0C in the helium gas flow for 30\u2009min to eliminate impurities. After cooling to 40\u2009\u00b0C, the sample was saturated with NH3 or CO2, then weakly adsorbed NH3 or CO2 was subsequently removed by purging with He for 60\u2009min. The desorption process was conducted in He from room temperature to 400\u2009\u00b0C. The desorbed NH3 or CO2 was detected by TCD detector.X-ray photoelectron spectroscopy (XPS) spectra were obtained using a Thermo ESCALAB 250 spectrometer equipped with Al Ka radiation (1486.6\u2009eV) under a binding energy was referenced to the C1s line (284.8\u2009eV).The element content of Zr and Ce of the catalysts was determined by an inductively coupled plasma atomic emission spectrometer (ICP-AES, Thermo iCAP6300, Thermo Fisher, USA).A 60\u2009mL autoclave batch reactor was used for the direct synthesis of dimethyl carbonate (DMC) from CO2 and CH3OH. Typically, 0.1\u2009g catalyst and 6.4\u2009g CH3OH were added into the autoclave, then 5.5\u2009MPa CO2 was introduced into the autoclave at 25\u2009\u00b0C. After reacting for a desired time, the reactor was immediately cooled by water bath and depressurized. Then, the catalyst and liquid product were separated by centrifugation. The compositions of the reaction product were identified with a Haixin GC-950 gas chromatograph equipped with a flame ionization detector. The amount of DMC formed was determined by internal method with n-amyl alcohol as the internal standard substance. For all catalytic reactions, no other substances were detected except DMC in both gas and liquid phase samples. Therefore, the selectivity of DMC was considered as 100%.The DMC yield, DMC formation rate and the turn over frequency (TOF) of DMC were calculated using the following formulas [28]:\n\n(1)\n\n\nDMC\n\nyield\n=\n\n\nAmount\n\nof\n\nDMC\n\nformed\n\n(\nmmol\n)\n\u00d7\n2\n\n\nAmount\n\nof\n\nmethanol\n\nadded\n\n(\nmmol\n)\n\n\n\u00d7\n100\n%\n\n\n\n\n\n\n(2)\n\n\nDMC\n\nformation\n\nrate\n=\n\n\nAmount\n\nof\n\nDMC\n\nformed\n\n(\nmmol\n)\n\n\nAmount\n\nof\n\ncatalyst\n\n(\ng\n)\n\u00d7\nReaction\n\ntime\n\n(\nh\n)\n\n\n\n\n\n\n\n\n(3)\n\n\nTOF\n=\n\n\nAmount\n\nof\n\nDMC\n\nformed\n\n(\nmmol\n)\n\n\nAmount\n\nof\n\nmetal\n\nin\n\nadded\n\ncatalyst\n\n\nmmol\n\n\u00d7\nReaction\n\ntime\n\n(\nh\n)\n\n\n\n\n\n\nThe crystal structures of synthesized catalysts were investigated by XRD patterns. As shown in \nFig. 2, the typical characteristic diffraction peaks of the as-synthesized UiO-66 were observed at 7.3\u00b0, 8.5\u00b0, 12.0\u00b0, 14.8\u00b0, 17.0\u00b0, 22.2\u00b0, 25.7\u00b0 and 30.7\u00b0, which corresponded to the (111), (002), (022), (113), (004), (115), (224) and (046) planes respectively. All the diffraction peaks were found to be well-matched with the reported literatures [30,31], confirming that Zr-based metal-organic framework UiO-66 was successfully synthesized. After the addition of Ce, with the cerium content increasing from 0 to 2\u2009mmol, the peak intensity of the XRD patterns increased, indicating the increased crystallinity of Ce-UiO-66-X. However, no new characteristic diffraction peaks of Ce-UiO-66\u20131.5 and 2 catalysts were observed, implying that Ce doping did not change the crystal structures of UiO-66. Based on this result, it might be deduced that the Zr element was partially substituted by Ce, and Ce atoms were successfully inserted into UiO-66 structure successfully. A further increase in Ce content to 5\u2009mmol, the intensity of the characteristic diffraction peak gradually weakened and finally disappeared [32], and a new diffraction peak at 2\u03b8 =\u20099.4\u00b0, corresponding to cerium tris (3,5-diaminobenzoatel) hydrate (PDF 41\u20131744) appeared. Moreover, the intensity of the new diffraction peak intensity increased with increasing cerium content, which implied that the MOFs structure formation was hampered. That is to say, it was necessary to control the dosage of Ce to ensure formation of the crystalline lattices of UiO-66 MOFs [33]. Obviously, the diffraction peak intensity of the Ce-UiO-66\u20132 catalyst was the largest highest in the synthetic catalysts, indicating the highest crystallinity.The nitrogen adsorption-desorption isotherms and pore size distribution curves of the different catalysts are displayed in Fig. S1. UiO-66 (Fig. S1a) exhibited a typical Type \u2160 adsorption isotherm curve, indicating that the sample presented microporous structure [34]. After adding cerium, the Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132 still maintained a microporous structure, implying Ce element replaced part of the Zr atoms by one-step synthesis method successfully. A further increasing in the content of cerium, the isotherm curve of Ce-UiO-66\u20133.5 displayed a type IV curve with a hysteresis loop at P/P0 =\u20090.45\u20130.95, demonstrating that the material possessed a typical mesoporous structure [35]. The appearance of this mesoporous structure was probably due to the fact that excessive cerium dosage hampered the formation of UiO-66 MOFs structure.\n\nTable 1 lists the specific surface area and total pore volume of the UiO-66 and Ce-UiO-66-X catalysts. Ce-UiO-66\u20132 had a maximum specific surface area (1281.3\u2009m2 g\u22121) and pore volume (0.724\u2009cm3 g\u22121). Among these catalysts, the specific surface area and total pore volume of Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132 were larger than those of UiO-66, suggesting that the introduction of Ce was beneficial to enlarge the specific surface area to some extent, which were probably because of their high crystallinity. Whereas the specific surface area and total pore volume of Ce-UiO-66\u20133.5 and Ce-UiO-66\u20135 were both smaller than those of UiO-66. Their specific surface area and total pore volume decreased sharply may be due to the formation of a new crystalline phase. Combined with XRD results, it can be concluded that a decrease in crystallinity could result in a decrease in specific surface area.FT-IR spectra of UiO-66, Ce-UiO-66\u20131.5, Ce-UiO-66\u20132, Ce-UiO-66\u20133.5 and Ce-UiO-66\u20135 catalysts are showed in \nFig. 3. The broad band around 3227\u2009cm\u22121 was ascribed to -OH stretching vibration from the surface adsorbed water [36]. The band at 1710\u2009cm\u22121 corresponded to the uncoordinated -COOH of free BDC [27,37], corroborating the existence of unreacted BDC trapped in the micropores of Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132. Obviously, the band for 1710\u2009cm\u22121 didn\u2019t emerge in FT-IR spectra of Ce-UiO-66\u20133.5 and Ce-UiO-66\u20135, indicating that unreacted BDC was washed out completely. The band at 1657\u2009cm\u22121 was assigned to CO stretching in the carboxylic acid present in BDC [25,38]. The band at 1579\u2009cm\u22121 was assigned to CC vibration in the aromatic ring of the organic linker [39]. The strong band at 1400\u2009cm\u22121 was linked to C-O stretching vibrations in carboxylic acid [39].It was clear that the peaks of CC and C-O bonds shift to a lower wavenumber in Ce-UiO-66\u20133.5 and Ce-UiO-66\u20135 catalysts, which was attributed to the formation of a new crystalline phase, these were consistent with XRD results. The peaks at 668, 744, 810\u2009cm\u22121 were the mix of O-H band and C-H vibration in the BDC ligand [40,41]. The characteristic adsorption peaks at 549\u2009cm\u22121 and 485\u2009cm\u22121 corresponded to Zr-(OC) asymmetric stretch and Zr-O vibration in the UiO-66, Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132 respectively [30,42]. It is worth noting that the diffraction peaks at 549\u2009cm\u22121 and 485\u2009cm\u22121 showed the maximum diffraction peak intensity in Ce-UiO-66\u20132 catalyst. With the further increase of Ce content, the peaks at 549\u2009cm\u22121 and 485\u2009cm\u22121 disappeared and a new peak at 509\u2009cm\u22121 appeared corresponding to the bond asymmetric stretching of Ce-(OC) in the spectrum of Ce-UiO-66\u20133.5 and Ce-UiO-66\u20135 [39]. Obviously, the Ce-UiO-66\u20132 catalyst showed the maximum diffraction peak intensity in the FT-IR spectra, revealing that the Ce-UiO-66\u20132 catalyst had the most regular UiO-66 MOFs structure. These results were consistent with the XRD results.The surface morphology of the samples was investigated by SEM (\nFig. 4). As depicted in Fig. 4a, it was clear that the UiO-66 catalyst showed the irregularly spherical crystal morphology. This morphology was in line with that reported in the literature [31,43]. Compared to the UiO-66 catalyst, the particles of Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132 catalysts became smaller and the crystal structure became more complete gradually. Especially, Ce-UiO-66\u20132 catalyst (Fig. 4c) had the most regular UiO-66 MOFs structure. However, a further increase in amount of Ce, large lumps appeared on SEM images of Ce-UiO-66\u20133.5 and Ce-UiO-66\u20135 catalysts, as shown in Fig. 4d and e. This may be because that a large amount of Ce caused the protonation of terephthalic acid ligands and then restrained the formation of UiO-66 MOFs structure, finally appeared big block morphology. The results indicated that excessive cerium content formed a new crystalline phase, which was consistent with the XRD observation.Furthermore, the TEM characterization of UiO-66, Ce-UiO-66\u20132, Ce-UiO-66\u20135 catalysts are also presented in \nFig. 5. As provided in Fig. 5a and b, it showed that the particle sizes of the crystalline UiO-66 and Ce-UiO-66\u20132 were uniform, 150\u2013200\u2009nm approximately. Moreover, a big block morphology was observed from the TEM image in Fig. 5c, which was consistent with the SEM observation. The elemental mapping images of the UiO-66, Ce-UiO-66\u20132 and Ce-UiO-66\u20135 materials were observed, including C, Ce, O and Zr. No conglomeration of Ce element was observed, suggesting the good dispersion of Ce on the Ce-UiO-66\u20132 catalyst surface.To investigate the thermal stability of Ce-UiO-66-X and UiO-66 catalysts, the materials were measured by thermal gravimetric analysis (TGA). As shown in Fig. S2, three distinct mass-loss regions were observed. A minor mass-loss below 150\u2009\u00b0C was observed due to the removal of H2O and methanol, and the weight loss within 150\u2013400\u2009\u00b0C was related to the evaporation of DMF solvent and dehydroxylation of zirconium/cerium oxo-cluster [44,45]. Also, a sharp weight decline in the range of 400\u2013540\u2009\u00b0C (UiO-66, Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132 samples) and 340\u2013440\u2009\u00b0C (Ce-UiO-66\u20133.5 and Ce-UiO-66\u20135 samples) were corresponded to the decomposition of organic bridging ligands (BDC) [28]. In addition, no additional weight loss was observed above 550 \u00b0C. The TGA results indicated that the UiO-66 possesses highly stable structure and the Ce-UiO-66\u20131.5 as well as Ce-UiO-66\u20132 display no obvious change on the thermal stability, which was maybe due to the Ce incorporated into the framework of UiO-66. However, the Ce-UiO-66\u20133.5 and Ce-UiO-66\u20135 samples decreased obviously, which could be related to incomplete UiO-66 MOFs structure due to the introduction of excessive cerium in the synthesis process.The number of BDC linkers in Zr6 formula unit was calculated following the method reported by Shearer [46]. In the work, the number of linkers in the ideal Zr6 formula unit is 6, and the actual number of linkers in the defective Zr6 formula unit is 5.13 for UiO-66. As displayed in \nFig. 6a, the experimental weight loss (%) of BDC linkers for UiO-66 and Ce-UiO-66\u20132 were 42.17% and 44.51% respectively. Therefore, Ce-UiO-66\u20132 possessed more ideal structural units, which was consistent with the results of XRD.As shown in \nFig. 7, the NH3-TPD and CO2-TPD were performed to probe the acidic and basic properties of all catalysts. Based on the area of NH3 and CO2 desorption peaks, the number of different acidic and basic sites were summarized in \nTable 2. Obviously, all samples showed three types of acidic sites with different intensities in Fig. 7a, which could be attributed to weak (\u03b11 and \u03b12 peak 50\u2013200\u2009\u00b0C) and medium (\u03b2 peak 200\u2013330\u2009\u00b0C) acidic sites, respectively, similar results were also reported by Zhang et al. [47]. The acidity of UiO-66 was originating from the missing of BDC-linker in the Zr6 clusters and the exposed Zr4+ of Zr6 node [34,48\u201351]. As can be seen, the number of total acidic sites in UiO-66 was 3.10\u2009mmol\u2009g-cat\u22121. After Ce modification, the amount of the total acidic sites increased from 3.10\u2009mmol\u2009g-cat\u22121 to 5.02\u2009mmol\u2009g-cat\u22121 for Ce-UiO-66\u20132, which probably originated from the insertion of Ce3+ ions into the zirconium clusters [44]. However, as the cerium content increased to 5\u2009mmol, the amount of the total acidic sites decreased from 5.02\u2009mmol\u2009g-cat\u22121 to 0.48\u2009mmol\u2009g-cat\u22121 for Ce-UiO-66\u20135. This was probably because the excessive cerium dosage hampered the formation of UiO-66 MOFs structure as indicated by the XRD results.\nFig. 7b shows the CO2 desorption profiles of Ce-UiO-66-X and UiO-66 catalysts, which could be assigned to weakly (peak \u03b1 200\u2013310\u2009\u00b0C) and moderately (peak \u03b2 310\u2013400\u2009\u00b0C) basic sites, respectively. Peak \u03b2 could be attributed to medium (\u03b2 peak) basic sites arising from the unsaturated lattice oxygen anion (Zr-O and Ce-O) and unsaturated terminal oxygen anion in zirconium/cerium cluster (Zr-O\u2212 and Ce-O\u2212) [28]. As can be seen in the Table 2, the Ce-UiO-66\u20132 showed the largest amount of medium basic sites of 1.04\u2009mmol\u2009g-cat \u22121. And Ce-UiO-66\u20135 showed no basicity virtually, which was probably attributed to the irregular structure, as indicated by the TEM and SEM results.As indicated by NH3-TPD and CO2-TPD results, compared with UiO-66, Ce modulation increased the amount of acid and basic sites in Ce-UiO-66\u20132 significantly. The total acidity and basicity decreased in the order of Ce-UiO-66\u20132\u2009>\u2009Ce-UiO-66\u20131.5\u2009>\u2009UiO-66\u2009>\u2009Ce-UiO-66\u20133.5\u2009>\u2009Ce-UiO-66\u20135.To investigate the surface compositions of Ce-UiO-66-X and UiO-66 catalysts, X-ray photoelectron spectroscopy (XPS) was employed and the results are shown in \nFig. 8. As presented in Fig. 8a, three main peaks corresponding to Zr 3d, C 1\u2009s, and O 1\u2009s were observed in the XPS spectra of UiO-66. After the Ce3+ doping (Fig. 8b), new peaks appeared at 882.5, 885.5\u2009eV and 900.8, 903.8\u2009eV of Ce-UiO-66\u20133.5 corresponding to Ce 3d5/2 and Ce 3d3/2\n[52,53], respectively, which indicated the incorporation of Ce3+. However, no characteristic diffraction peaks for Ce 3d appeared on the Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132 catalysts, which was mainly because the highly dispersed Ce or low cerium content (below the limit of detection by the XPS) on the catalyst surface.In the Zr 3d XPS spectra of UiO-66 catalysts (Fig. 8c), the doublets in the binding energies at 182.9\u2009eV and 185.3\u2009eV belonged to the Zr 3d5/2 and Zr 3d3/2, respectively. The Zr 3d binding energy increased first and then decreased with the increase of cerium content, demonstrating that the bonding environment of Zr nodes was changed after the introduction of Ce3+. In addition, the result indicated that Zr4+ in the structure of UiO-66 might be partially replaced by Ce3+\n[44,54].It is worth noting that there was a negative shift in the binding energy of O 1s spectra from 532.0 to 531.7\u2009eV when the Ce3+ doping amount reached 5\u2009mmol, indicating that a new chemical state of O1s emerged. This result was consistent with the XRD analysis. Furthermore, the O 1s (Fig. 8d) spectra can be fitted by three peaks. The peak at \u223c530.3\u2009eV could be attributed to unsaturated lattice species of Zr-O and Ce-O bonds [55]. The peak at \u223c531.9\u2009eV was ascribed to surface hydroxyl oxygen species O-H [28]. The peak at \u223c533.3\u2009eV was assigned to oxygen component in the O-CO [56]. The curve fitting results of all five samples were listed in \nTable 3, unsaturated lattice species (Zr-O and Ce-O) have the maximum content of 15% in Ce-UiO-66\u20132 catalyst, indicating it had more intermediate alkaline sites, which was consistent with the CO2-TPD results.To investigate the effect of different preparation methods on the catalytic performance, the Ce-UiO-66\u20132-IM and Ce-UiO-66\u20132-IE catalysts with the same Ce/Zr molar ratio as Ce-UiO-66\u20132 (ICP-AES analysis) were synthesized and evaluated in \nFig. 9. The Ce-UiO-66\u20132 catalyst has the highest DMC yield. This was mainly due to the high dispersion of Ce into skeleton lattice of Ce-UiO-66\u20132 catalyst. Therefore, one-step synthesis is the best catalyst preparation method.\n\nTable 4 summarizes the catalytic performances of UiO-66 MOFs catalysts with different contents of cerium. The DMC yield first increased and then decreased with the content of Ce increasing, and it reached the maximum value when X\u2009=\u20092\u2009mmol. In addition, it should be noted that DMC was not detected in the absence of catalyst, which showed the importance of catalyst in this reaction. Among the catalysts tested, the UiO-66 (none Ce) exhibited low catalytic activity. After Ce modification, the average DMC formation rate increased from 0.171\u2009mmol\u2009g-cat\u22121 h\u22121 for pristine UiO-66 to 0.335\u2009mmol\u2009g-cat\u22121 h\u22121 for Ce-UiO-66\u20132, indicating that the Ce modification significantly enhanced the catalytic activity of UiO-66. Moreover, the Ce-UiO-66\u20135 catalyst (pure Ce) also showed no activity for this reaction, which could be related to irregular UiO-66 structure due to the introduction of excessive cerium in the synthesis process as indicated by the XRD and SEM results.In order to eliminate the influence of the crystal structure of Ce-based UiO-66 catalyst on the catalytic performance, the UiO-66(Ce) catalyst (pure Ce) with complete UiO-66 structure was synthesized according to the method used in literature [57,58], and evaluated for the synthesis of DMC from CO2 and CH3OH, and the results are shown in Table 4. As can be seen, the DMC yield over UiO-66(Ce) was 0.047%, which has a higher catalytic activity than Ce-UiO-66\u20135. The results indicated that the crystal structure of Ce-based UiO-66 catalyst really do has effect on the activity of the catalyst in the aforementioned reaction. However, the catalytic activity of UiO-66 (Ce) was still lower than that of other catalysts. In addition, the Ce-UiO-66\u20132 catalyst exhibited the best catalytic performance (0.335\u2009mmol\u2009g-cat\u22121 h\u22121) in this work, which could be due to the maximum amount of acidic and basic sites and the most regular UiO-66 MOFs structure.As shown in \nFig. 10, the effects of different reaction parameters (reaction temperature, reaction time, and catalyst dose) on the DMC yield over Ce-UiO-66\u20132 catalyst were investigated.As we know, the reaction is not occurred spontaneously in nature [59]. So, the influence of the reaction temperature on the DMC yield is shown in Fig. 10a, the temperature was varied between 100 and 150\u2009\u00b0C. There was a significant increase in the yield of DMC as the reaction temperature increased from 100\u2009\u00b0C to 140\u2009\u00b0C. As the temperature increased to 150\u2009\u00b0C, the DMC yield decreased. So, it can be seen that the optimum reaction temperature is 140\u2009\u00b0C.\nFig. 10b shows the effect of the catalyst dose on the catalytic performance. The yield of DMC increased with catalyst dose, and the highest yield of DMC was 0.209% at 0.25\u2009g of the catalyst. With a further increase in the catalyst dose >\u20090.25\u2009g, the DMC yield decreased instead.\nFig. 10c indicates the effect of the reaction time on DMC yield, the time was varied between 1 and 48\u2009h. It can be seen that the DMC yield increased from 0.126% to 0.295% with the reaction time increased from 1 to 12\u2009h. After 12\u2009h, DMC yield was constant. This was probably due to the fact that the reaction has reached equilibrium when the reaction time is 12\u2009h.So, the optimum reaction conditions on the Ce-UiO-66\u20132 catalyst are as follows: reaction temperature was 140\u2009\u00b0C, catalyst weight was 0.25\u2009g; reaction time was 12\u2009h.\n\nFig. 11 shows the catalytic results of Ce-UiO-66\u20132 for the synthesis of DMC from methanol and CO2 with six times reused. After each run, the catalyst was recovered by centrifugation, and further washed with methanol and dried at 150\u2009\u00b0C for 12\u2009h. The regenerated sample was then used for the next run under the same reaction conditions. The observed DMC yield slightly decreased with an increase in reuse cycles. While, the XRD patterns (Fig. S3) suggested that the structure of Ce-UiO-66\u20132 was retained after used for six cycles. Furthermore, ICP-AES analysis (Table S4) showed that only 0.2% of the total amount of Zr and Ce was leached from Ce-UiO-66\u20132 after six cycles. The results showed that the Ce-UiO-66\u20132 catalyst was stable and effective for DMC synthesis in our reaction system.Bell et al. [20,60] had pointed out that amphoteric Zr-OH hydroxyl groups and coordinately unsaturated Zr4+O2\u2212 sites as Lewis acid-base pairs played a key role in the synthesis of DMC from CO2 and CH3OH over ZrO2. Wang et al. [61] reported that spindle-like ceria exhibited the best catalytic performance due to exposing active planes and a large amount of acid-base sites. Kumara et al. [62] believed that manganese modification could enrich the amount of acid-base sites on CeO2, then, methanol could be activated to form H+ and CH3O\u2212 due to basic sites, and to form CH3\n+ and OH\u2212 in the presence of weak/moderate acidic sites. The catalyst based on Ce1-Mn0.125 showed the best catalytic activity with the maximum number of weak and moderate acidity and basicity. Thus, tuning the amount of acid-base sites of catalysts was an effective strategy to improve the catalytic activity of the catalyst. \nFig. 12 shows the relationship between the catalytic activity and the amount of acid and medium basic sites of Ce-UiO-66-X and UiO-66 samples. It could be found that the yield of DMC increased with the number of acid and medium basic sites (Ce-UiO-66\u20132\u2009> Ce-UiO-66\u20131.5\u2009> UiO-66\u2009> Ce-UiO-66\u20133.5\u2009> Ce-UiO-66\u20135). The Ce-UiO-66\u20132 catalyst presented the best catalytic activity compared to other catalysts, which was due to the fact that the higher amount of acidic and medium basic sites are responsible for the formation of methyl carbonate.Xuan et al. [28] reported that the enlargement of the pore size and pore volume were also contributed to the enhancement of catalytic activity, since the mass transfer resistance in the channel with larger pore size was lower. So, in addition to the acid-base properties, the effect of the textural properties on the catalytic activity should also be considered [2]. As shown in Table 2, after Ce modification, Ce-UiO-66\u20132 and Ce-UiO-66\u20131.5 provided more active sites (acid sites: 5.02\u2009mmol\u2009g-cat\u22121 and 4.97\u2009mmol\u2009g-cat\u22121, medium basic sites 0.61\u2009mmol\u2009g-cat\u22121 and 0.53\u2009mmol\u2009g-cat\u22121) than pure UiO-66 (acid sites: 3.1\u2009mmol\u2009g-cat\u22121, medium basic sites: 0.36\u2009mmol\u2009g-cat\u22121) for this reaction. In addition, the average DMC formation rate increased from 0.171\u2009mmol\u2009g-cat\u22121h\u22121 for pristine UiO-66 (Table 4) to 0.315\u2009mmol\u2009g-cat\u22121h\u22121 and 0.335\u2009mmol\u2009g-cat\u22121h\u22121 for Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132, respectively. It is worth noting that the difference between average DMC formation rate of Ce-UiO-66\u20132 and Ce-UiO-66\u20131.5 is 0.02 (0.335\u2009\u2212 0.315\u2009= 0.02), the difference between acid sites of Ce-UiO-66\u20132 and Ce-UiO-66\u20131.5 is 0.05 (5.02 \u2212 4.97\u2009= 0.05), and the ratio of 0.02\u20130.05 is 0.4. However, the ratio of the difference between average DMC formation rate of Ce-UiO-66\u20131.5 and UiO-66 to the difference between acid sites of Ce-UiO-66\u20131.5 and UiO-66 is 0.077 [(0.315 \u2212 0.171)/ (4.97 \u2212 3.1) =\u20090.077] (Fig. 12). That is to say, it is not linear correlation completely between the acid sites and the average DMC formation rate, the same is true for medium basic sites. In conclusion, the acid-base properties are not the only factor affecting catalytic activity. According to the N2 adsorption results, the specific surface area and pore volume of Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132 were larger than that of pristine UiO-66, and the catalytic activity of Ce-UiO-66\u20131.5 and Ce-UiO-66\u20132 were higher. This showed that the enlargement of the specific surface area and pore volume in Ce-UiO-66-X may also contributed to the catalytic enhancement, since the large specific surface area may also contribute to provide higher accessibility for the reactant to the active sites located in the micropores.In summary, the increased amount of active sites as well as the enlargement of specific surface area originated from the addition of a suitable amount of cerium contributed to the enhanced activity of Ce-UiO-66-X.The catalytic activity of Ce-UiO-66\u20132 catalyst for the direct synthesis of DMC was compared with the reported catalysts under different reaction conditions (\nTable 5). As can be seen in Table 5, the Ce-UiO-66\u20132 (Entry 1) catalyst showed the DMC yield of 0.134%. Additionally, using 2-cyanopyridine as dehydrating agent (Entry 2), DMC yield can be increased by about ten times. It can be seen that the addition of dehydrating agent has increased the yield of DMC to a certain extent, which is consistent with previous literature [27,28,63].However, the TOF over Ce-UiO-66\u20132 (Entry 2) was lower than that over CeO2 catalyst using 2-cyanopyridine as a recyclable dehydrating agent (Entry 12). This is because CeO2 can improve the hydration of 2-cyanopyridine, thus rapidly remove the by-product water and increase the formation of DMC. Besides, the Ce-UiO-66\u20132 (Entry 1) catalyst showed higher TOF than those reported without the addition of dehydrating agent. It should be noted that the TOF over 5%Cu-Ni/ZIF-8 (Entry 16) was higher than that over Ce-UiO-66\u20132 catalyst (Entry 1). But in the reusability test experiments, the yield of DMC over 5%Cu-Ni/ZIF-8 [70] catalyst dramatically decreased from 1.71\u2009mmol\u2009g-cat\u22121 to 0.57\u2009mmol\u2009g-cat\u22121 after four cycles. In contrast, Ce-UiO-66\u20132 was used for six times with DMC yield slightly decreased from 1.18\u2009mmol\u2009g-cat\u22121 to 1.16\u2009mmol\u2009g-cat\u22121, indicating the better stability of Ce-UiO-66\u20132 catalyst.A possible reaction mechanism for the direct synthesis of DMC from CO2 and methanol over Ce-UiO-66\u20132 catalyst is shown in \nScheme 1. Ce-UiO-66\u20132 contains Lewis acid site (exposed Zr4+/Ce3+) and Lewis basic sites (unsaturated O2\u2212 anion in Zr-O, Ce-O, Zr-O\u2212 and Ce-O\u2212).One molecule of methanol and CO2 is activated on Lewis acid-base pair of sites. Methanol binds to Lewis acidic site of metal node in Ce-UiO-66\u20132 to form Zr/Ce-OCH3 and releases an H atom, which then reacts rapidly with a surface OH group to form H2O. The activated CO2 is then inserted into the Zr/Ce-O bond of the Zr/Ce-OCH3 species to produce m-CH3OCOO-Ce/Zr. In addition, methanol can also be activated into methyl group and hydroxyl group by the acidic sites. The methyl group is transferred to the terminal O atom of methyl carbonate species to produce DMC, while the hydroxyl group can react with exposed Zr4+/Ce3+ to form terminal hydroxyl. The reaction mechanism is similar to some reported literature [2,19,27,28].In this study, a series of Ce-doped Zr-based metal-organic frameworks UiO-66 were synthesized via a modified hydrothermal method and investigated for the direct synthesis of DMC from CO2 and methanol. Among all the Ce-UiO-66-X catalysts, Ce-UiO-66\u20132 catalyst showed the highest DMC yield of 0.295% at 12\u2009h, 140\u2009\u00b0C, 11\u2009MPa. The results demonstrated that the addition of cerium to UiO-66 could influence the growth of UiO-66 MOFs, and thus increased the amount of the surface acidic and moderately basic sites, which then facilitated the activation of methanol and CO2. Besides, the enlargement of the specific surface area was also contributed to the enhancement of catalytic activity. This study offers strategies to design metal doping modification of MOFs materials and provides a novel catalyst for the synthesis of dimethyl carbonate.\nLinmeng Huo: Conceptualization, Writing \u2013 review & editing, Data curation. Lin Wang: Software, Investigation, Data curation. Jingjie Li: Formal analysis, Visualization, Software. Yanfeng Pu: Methodology, Resources, Review, Supervision. Keng Xuan: Software, Investigation. Congzhen Qiao: Supervision, Resources. Hao Yang: Methodology, Investigation.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was funded by the Technology Research Project of Henan Province (Grant No. 212102210210), the First-class Discipline Construction Project of Henan University (No. 2019YLZDCG01), the Technology Research Project of Kaifeng City (Grant No. 2001003).Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jcou.2022.102352.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  The disadvantage of Ce/Zr metal oxide as catalyst for direct synthesis dimethyl carbonate (DMC) from CO2 and methanol is low DMC formation rate because of its low specific surface area and active sites. In this work, it can be improved by highly uniform dispersion cerium doped Zr-based metal-organic frameworks (UiO-66 MOFs), which were synthesized via a modified hydrothermal method. The as-prepared catalysts have been extensively characterized using XRD, BET, FT-IR, SEM, TEM, TGA, NH3-TPD, CO2-TPD, XPS techniques. Experimental evaluation results indicated that the highly uniform dispersed Ce-doped UiO-66 MOFs exhibited markedly improved catalytic performance than traditional Ce/Zr metal oxide catalyst. The highest yield of DMC catalyzed by Ce-UiO-66\u20132 was 0.295% (reaction time was 12\u00a0h; reaction temperature was 140 \u00b0C; reaction pressure was 11\u00a0MPa). Then, it was found that doping of Ce atoms into the zirconium lattice to produce UiO-66 MOFs could effectively increase the number of acidic and medium basic sites than UiO-66 (none Ce), thus could greatly enhance the catalytic performance. Moreover, using 2-cyanopyridine as dehydrating agent, the DMC yield could be further raised greatly. Last, based on reported literature and our results, a possible reaction mechanism over Ce-UiO-66-X was proposed.\n               "}
{"full_text": "We are nearing the point where hydrogen energy is rapidly developing and aimed to replace the traditional fossil fuel with green and renewable energy [1,2]. Electrocatalytic water splitting has been proven to be a promising clean energy technology for hydrogen production. Still, its large-scale implementation is hindered by the high-cost and scarce resource of noble-metal catalysts (e.g., RuO2 and IrO2) commonly used on the anode to overcome the sluggish kinetics of oxygen evolution reaction (OER) [3]. Therefore, searching for alternative electrocatalysts with analogous OER activities but low cost and high abundance is an attractive research topic [4,5]. During recent decades, researchers have developed numerous non-noble metal-based electrocatalysts with low overpotential, small Tafel slope, and good stability, such as metal oxides [6], hydroxides [7,8], chalcogenides [8,9], nitrides [10], phosphides [11,12], and carbon-based materials [13,14].Metal-organic frameworks (MOFs) possess large specific surface area, tunable porosity, ordered structure, and flexible metal-linker compositions, accordingly, demonstrate high potential in constructing the advanced OER electrocatalysts [15\u201317]. Firstly, MOFs can be directly applied as anode catalysts to boost OER activities, such as MIL-88 [18], MIL-53 [19], MIL-100 [20], NiFe-MOF [21] and ZIF-67 [22,23]. Moreover, the MOF derived materials can afford unique nanostructures with enhanced catalytic performance while avoiding the intrinsic chemical instability of MOFs, such as hollow structured metal oxides [24], metal phosphides [25], and carbon-coated metal alloys [26,27]. To further optimize the activity, the second or third metal ions were deliberately incorporated into the MOF matrix to modify the coordination environments and electronic properties of active sites [28,29]. Fe-Co-Ni based ternary materials are commendable cases in the preparation of MOF-derived electrocatalysts compared with monometallic derivatives owing to the following multiple advantages: facile electronic structure manipulation to reach the optimal oxygen binding affinity; the synergetic effects between different metal sites for multistep OER process; and improvement in the electrical conductivity. For example, Lang et al. synthesized Fe-Co-Ni-MIL-53 by the solvothermal method. They acquired a current density of 10\u2009mA cm\u22122 with a low overpotential of 219\u2009mV, the electronic environment of active metal sites is efficiently modulated under the synergy effect of the ternary metals to favor enhancing the catalytic OER process [30]. Chen et al. designed a hollow multivoid nanocuboidal catalyst based on ternary Ni-Co-Fe MOF precursor, different ionic reaction rates of [Co(CN)6]3- and [Fe(CN)6]3- in the MOF precursor with S2- are exploited to produce internal interconnected voids, heteroatom doping, and a favorable electronic structure, thus generating dual-functionality toward OER and hydrogen evolution reaction (HER) [31].Generally, the one-pot synthesis method, in which the multimetal-organic frameworks are prepared by introducing all metallic sources into the organic precursors, is a conventional strategy to build electrocatalysts [32]. However, the in-situ generated multimetal-organic frameworks always possess the uniform structure as the monometal-organic frameworks, which may suffer from the chemical and structural instability during electrochemical test or post-treatment, causing severe damage to the original well-designed nanostructures, such as the aggregation of metal site, decline of surface area and destroy of pore structure [33]. For example, Zhang et al. observed that the stability of in-situ generated FeCo-MIL-88B declined seriously (retain \u223c75 % of its initial activity) after 40\u2009h electrocatalysis, while the anion-exchange (such as W and Se) treated products can retain \u223c90 % of its initial activity at the same reaction conditions [18]. Han et al. found there are some large metal oxide nanoparticles forming when calcinating the in-situ generated Co-Fe MOF at 550\u2009\u00b0C in N2 gas, and the particle size becomes larger obviously when increasing the calcination temperature [34].In this study, we report an effective strategy to build a hierarchical MOF structure through ion exchange between Fe-MIL-101-NH2 and Co, Ni ions, where 2-D ternary metal MOF layers encapsulate 3-D MOF octahedral crystals. The original octahedral skeleton structure of MOF precursor can be maintained after air calcination treatment, resulting in hierarchically structured CoNiFe spinel oxide-carbonitrides hybrids, simplified as CoNiFeOx-NC. The obtained catalyst achieved an excellent OER activity with a low overpotential of 265\u2009mV at 50\u2009mA cm\u22122 and held excellent stability of more than 40\u2009h OER catalysis at around 12\u2009mA cm\u22122. A combination of multiple characterization techniques and density functional theory (DFT) calculations was employed to investigate the structure-performance relationships. The results revealed that the enhanced performance of CoNiFeOx-NC can be attributed to the highly dispersed metal oxides nanoparticles, large surface area, and unique electronic structure modulation of Co-Ni-Fe oxide active sites. Finally, the strategy developed here may open up more novel and versatile approaches to developing non-noble electrocatalysts with high performance.2-aminaterephthalic acid (BDC-NH2, 98 %, Aladdin); Iron (III) chloride hexahydrate (FeCl3\u00b76H2O, 99 %, Aladdin); Nickel (II) chloride hexahydrate (NiCl2\u00b76H2O, 99 %, Aladdin); Cobaltous (II) chloride hexahydrate (CoCl2\u00b76H2O, 99 %, Aladdin); N, N-Dimethylformamide (DMF, \u226599 %, Aladdin); Ethanol (C2H5OH, \u226599.7 %, Sinopharm Chemical Reagent Co., Ltd); Methanol (CH3OH, \u226599.5 %, Aladdin); Nafion (5\u2009wt.%, Sigma-Aldrich). All chemicals were purchased from commercial sources and used without further treatments.The 3-D Fe-MIL-101-NH2 was prepared following previously reported procedures [35]. Typically, BDC-NH2 (0.45\u2009g, 2.5\u2009mmol) was dissolved in DMF (15\u2009mL), then a mixture of FeCl3\u00b76H2O (1.35\u2009g, 5.0\u2009mmol) and DMF (15\u2009mL) was added into the above solution. After stirring for one hour at room temperature, the resultant mixture was transferred into a 50\u2009mL Teflon-lined autoclave and placed in a 115\u2009\u00b0C oven for 20\u2009h. After cooling to room temperature, the powder products were centrifuged, washed with DMF and methanol several times, and finally dried under vacuum for 24\u2009h.Co and Ni co-doped Fe-MIL-101-NH2 (labeled as CoNiFe-MOF) was synthesized by an ion-exchange method. The Fe-MIL-101-NH2 precursor (0.27\u2009g) was dispersed in DMF solvent (15\u2009mL). The mixture of NiCl2\u00b76H2O (0.67\u2009g, 2.8\u2009mmol) and CoCl2\u00b76H2O (1.34\u2009g, 5.6\u2009mmol) dissolved in DMF solvent (15\u2009mL) was added into the above solution under continuous magnetic stirring. After stirring for one hour at room temperature, the resultant mixture was transferred into a 50\u2009mL Teflon-lined autoclave and placed in a 90\u2009\u00b0C oven for 48\u2009h. After cooling to room temperature, the powder products were centrifuged, washed with DMF and methanol several times, and finally dried under vacuum for 24\u2009h.Co-doped Fe-MIL-101-NH2 (labeled as CoFe-MOF) and Ni-doped Fe-MIL-101-NH2 (labeled as NiFe-MOF) were synthesized in the similar process as CoNiFe-MOF but changing the cation precursor to CoCl2\u00b76H2O (2.00\u2009g, 8.4\u2009mmol) for CoFe-MOF and NiCl2\u00b76H2O (1.99\u2009g, 8.4\u2009mmol) for NiFe-MOF.\nIn-situ generated ternary-metal MOF (denoted as IS-CoNiFe-MOF) was synthesized in the similar process as Fe-MIL-101-NH2 but changing the cation precursor to the mixture of FeCl3\u00b76H2O (2.5\u2009mmol), NiCl2\u00b76H2O (2.0\u2009mmol) and CoCl2\u00b76H2O (0.5\u2009mmol).CoNiFeOx-NC was obtained through heating CoNiFe-MOF in the air. Firstly, CoNiFe-MOF frameworks were taken in a tube furnace. Then the temperature was raised to the target temperature (i.e., 200, 300, 400, and 500\u2009\u00b0C) at a rate of 5\u2009\u00b0C\u2009min\u22121 and maintained for 60\u2009min. Finally, the pyrolysis products were naturally cooled down and collected.CoFeOx-NC, NiFeOx-NC, IS-CoNiFeOx-NC were synthesized in the same process as CoNiFeOx-NC catalyst by heating CoFe-MOF, NiFe-MOF and IS-CoNiFe-MOF, respectively.X-ray powder diffraction (XRD) was carried out on a Philips X\u2019pert pro MPD Super diffractometer equipped with Cu K\u03b1 radiation (\u03bb\u2009=\u20091.5418\u2009\u00c5). Transmission Electron Microscope (TEM), High-Resolution TEM (HRTEM), and Selected Area Electron Diffraction (SAED) images were acquired on JEM-2100 UHR at an acceleration voltage of 200\u2009kV. X-ray photoelectron spectra (XPS) were recorded by Thermo Fisher ESCALAB 250 analyzer with Al K\u03b1 radiation. The element component analysis was characterized by inductive coupled plasma-atomic emission spectrometry (ICP-AES) using Agilent ICPOES 730 spectrometer. Thermogravimetric (TG) analysis was performed using a PerkinElmer Diamond TG-DTA set-sys-evolution instrumentation. IR spectra were recorded with a Bruker Alpha Platinum ATR in the 400\u20134000 cm\u22121 region. The pore properties and Brunauer-Emmett-Teller (BET) surface area of samples were characterized by a Micromeritics ASAP 2020 nitrogen adsorption apparatus at 77\u2009K.All the electrochemical performance tests were undertaken in a typical three-electrode system using CHI660E electrochemical workstation. The carbon paper electrode coated with catalysts, Ag/AgCl, and graphite rod acted as the working electrode, reference electrode, and counter electrode, respectively. The working electrode was prepared as follows: 5\u2009mg\u2009as-prepared catalyst and 20\u2009\u03bcL Nafion solution were dispersed in 500\u2009\u03bcL ethanol to form a homogeneous slurry by sonication, then 100\u2009\u03bcL of the acquired slurry was uniformly dispersed onto the 1.0 cm2 of carbon paper and dried at room temperature. Before the measurements, Ar gas was purged into the electrolyte solution (1.0\u2009M KOH) for at least 30\u2009min.All the measured potentials (E\nAg/AgCl) were calibrated to the reversible hydrogen electrode (RHE) potentials based on the equation: E\nRHE = E\nAg/AgCl + 0.0591\u00d7pH\u2009+\u20090.198. 90 % iR-correction was done using the instrument\u2019s available function for the linear sweep voltammetry (LSV). The electrochemical double-layer capacitance (C\ndl) of as-prepared samples were calculated by the CV method at different scan rates (5\u221250\u2009mV s\u22121) within a non-faradic potential range (1.107\u20131.207\u2009V).DFT\u2009+\u2009U calculations were performed using Vienna Ab-initio Simulation Package (VASP). The interactions between the valence electrons and ionic cores were described using the projector augmented wave (PAW) method. The electron exchange-correlation energy terms were evaluated using the Perdew-Burke-Ernzerhof (PBE) functional with on-site Coulomb repulsion U term on Co, Ni, and Fe 3d electrons [36]. According to the previous study, the values U(Co)\u2009=\u20093.0\u2009eV, U(Ni)\u2009=\u20095.5\u2009eV, and U(Fe)\u2009=\u20093.5\u2009eV are found to provide an appropriate description for the properties of transition metal oxides [37]. The plane-wave cut off energy for the system was optimized as 400\u2009eV. The Brillouin zone was sampled with a 3\u2009\u00d7\u20093\u00d71 k integrations mesh. The final forces and energy convergence criteria were set as 0.03\u2009eV \u00c5\u22121 and 1\u2009\u00d7\u200910-4 eV, respectively.The \u03b3-Fe2O3(110), CoFe2O4(110), and NiFe2O4(110) surface models were built with 15\u2009\u00c5 of vacuum along the z-axis. For the Co-doped NiFe2O4(110) surface model, one Ni Oh atom on the NiFe2O4(110) surface was substituted by a Co atom. During the geometry optimization, the uppermost two layers of atoms were relaxed, and the other bottom atoms were fixed to represent the bulk position. All the pristine surface models were optimized to minimal energy before the adsorption calculation. For the OER process calculation, a standard four-electron reaction mechanism in alkaline condition was considered for the calculation of Gibbs free energy change according to the previous study [38]:\n\n(1)\n\n\n\n\n\n\nO\n\nH\n\u2013\n\n+\n*\n\u2192\n*\nOH+\n\ne\n\u2013\n\n\n\n\n\n\u0394\n\nG\n1\n\n=\n\u2009\n\u0394\n\nG\n\nO\nH\n\n\n\u2212\n\u2009\ne\nU\n\u2009\n+\n\u2009\n\u0394\n\nG\n\nH\n+\n\n\n\n\np\nH\n\n\n\n\n\n\n\n\n\n\n\n\n(2)\n\n\n\n\n\n\n*\nOH+O\n\nH\n\u2013\n\n\u2192\n*\nO+\n\nH\n2\n\nO+\n\ne\n\u2013\n\n\n\n\n\n\u0394\n\nG\n2\n\n=\n\u2009\n\u0394\n\nG\nO\n\n\u2212\n\u2009\n\u0394\n\nG\n\nO\nH\n\n\n\u2212\n\u2009\ne\nU\n\u2009\n+\n\u2009\n\u0394\n\nG\n\nH\n+\n\n\n\n\np\nH\n\n\n\n\n\n\n\n\n\n\n\n\n(3)\n\n\n\n\n\n\n*\nO+O\n\nH\n\u2013\n\n\u2192\n*\nOOH\n+\n\ne\n\u2013\n\n\n\n\n\n\u0394\n\nG\n3\n\n=\n\u2009\n\u0394\n\nG\n\nO\nO\nH\n\n\n\u2212\n\u2009\n\u0394\n\nG\nO\n\n\u2212\n\u2009\ne\nU\n\u2009\n+\n\u2009\n\u0394\n\nG\n\nH\n+\n\n\n\n\np\nH\n\n\n\n\n\n\n\n\n\n\n\n\n(4)\n\n\n\n\n\n\n*\nOOH+O\n\nH\n\u2013\n\n\u2192\n*\n+\n\nO\n2\n\n+\n\nH\n2\n\nO\n+\n\ne\n\u2013\n\n\n\n\n\n\u0394\n\nG\n4\n\n=\n\u2009\n4.92\n\u2009\n\u2212\n\u2009\n\u0394\n\nG\n\nO\nO\nH\n\n\n\u2212\n\u2009\ne\nU\n\u2009\n+\n\u2009\n\u0394\n\nG\n\nH\n+\n\n\n\n\np\nH\n\n\n\n\n\n\n\n\n\nwhere U is the potential measured against normal hydrogen electrode (NHE) at standard conditions. The theoretical overpotential is then readily defined as:\n\n(5)\n\n\u03b7\n=\n\nmax\n(\n\n\u0394\n\nG\n1\n\n,\n\u2009\n\u0394\n\nG\n2\n\n,\n\u2009\n\u0394\n\nG\n3\n\n,\n\u2009\n\u0394\n\nG\n4\n\n)\n/\ne\n-\n1.23\n\n\n\nThe Gibbs free energy of adsorbed species is defined as:\n\n(6)\n\n\u0394\nG\n=\n\u0394\n\nE\n\na\nd\ns\n\n\n+\n\u0394\n\nE\n\nZ\nP\nE\n\n\n-\nT\n\u0394\n\nS\n\na\nd\ns\n\n\n\n\nwhere \n\u0394\n\nE\n\na\nd\ns\n\n\n is the adsorption energy, \n\u0394\n\nE\n\nZ\nP\nE\n\n\n is the zero point energy difference between adsorbed and gaseous species, and \n\u2009\nT\n\u0394\n\nS\n\na\nd\ns\n\n\n is the corresponding entropy difference between these two states (T was set to be 298\u2009K). The calculation results of zero point energy and entropy of the OER intermediates were listed in Table S1. \n\u0394\n\nE\n\na\nd\ns\n\n\n was calculated relative to H2O and H2 (at U\u2009=\u20090 and pH\u2009=\u20090) as:\n\n(7)\n\n\u0394\n\nE\n\nO\nH\n\n\n=\n\u2009\n\nE\n\nO\nH\n*\n\n\n-\n\u2009\n\nE\n*\n\n-\n(\n\nE\n\n\nH\n2\n\nO\n\n\n-\n\u2009\n\n1\n2\n\n\nE\n\n\nH\n2\n\n\n\n)\n\n\n\n\n\n(8)\n\n\u0394\n\nE\nO\n\n=\n\u2009\n\nE\n\nO\n*\n\n\n-\n\u2009\n\nE\n*\n\n-\n(\n\nE\n\n\nH\n2\n\nO\n\n\n-\n\nE\n\n\nH\n2\n\n\n\n)\n\n\n\n\n\n(9)\n\n\u0394\n\nE\n\nO\nO\nH\n\n\n=\n\u2009\n\nE\n\nO\nO\nH\n*\n\n\n-\n\u2009\n\nE\n*\n\n-\n(\n2\n\nE\n\n\nH\n2\n\nO\n\n\n-\n\u2009\n\n3\n2\n\n\nE\n\n\nH\n2\n\n\n\n)\n\n\n\n\nFig. 1\n depicts the synthesis scheme of multi-metal MOFs derived hybrids via the ion-exchange method. Firstly, Fe-MIL-101-NH2 was synthesized using BDC-NH2 and FeCl3\u00b76H2O according to the previously reported procedures [35]. The as-prepared Fe-MIL-101-NH2 displays uniform octahedron morphology with a smooth surface and an average diameter of \u223c250\u2009nm (Fig. 2\na). Secondly, Co2+ and Ni2+ were introduced into Fe-MIL-101-NH2 at 90\u2009\u00b0C in the DMF solvent through a solvothermal method. ICP-AES results (Table S2) show that the content of Ni and Co is about 7.69\u2009wt.% and 1.76\u2009wt.% respectively in the as-prepared CoNiFe-MOF, and Fe content decreases from 14.96\u2009wt.% of the original Fe-MIL-101-NH2 to 7.39\u2009wt.%. XRD results (Fig. 2c) demonstrate that the diffraction pattern of CoNiFe-MOF is similar to that of Fe-MIL-101-NH2, suggesting the pristine MOF phase remained after ion exchange treatment. Their FT-IR spectra (Fig. S1) clearly show the adsorption band of the \u03bdas(-COO-) linking to the metal atom at 1657\u2009cm\u22121 [20], indicating no free BDC-NH2 ligand in the CoNiFe-MOF structure [39]. However, the microstructure of CoNiFe-MOF has changed concerning Fe-MIL-101-NH2, as shown in Fig. 2b, the original smooth surface has turned into rough and fluffy nanosheets.Finally, the CoNiFe-MOF was pyrolyzed in the air at 300\u2009\u00b0C in a tube furnace. As shown in Fig. 2e, the obtained sample (CoNiFeOx-NC) almost preserves the original morphology of CoNiFe-MOF. XRD results (Fig. 2f) reveal the formation of CoNiFe oxides phase with spinel structure (NiFe2O4, JCPDS no. 54-0964; CoFe2O4, JCPDS no. 22-1086; \u03b3-Fe2O3, JCPDS no. 39-1346). The diffraction peaks of Fe-MIL-101-NH2 disappeared completely in CoNiFeOx-NC after air heating treatment. The HRTEM image of CoNiFeOx-NC (Fig. 2h) further confirms the microstructure of CoNiFe spinel oxides, where the lattice fringe with the distance of 0.213, 0.241, and 0.222\u2009nm can be assigned to (400), (222) facet of NiFe2O4 and (321) facet of \u03b3-Fe2O3 respectively. Furthermore, the compositional distribution of a typical CoNiFeOx-NC was investigated through STEM coupled with EDX-mapping (Fig. 2j). It can be observed that Co and Ni elements display a uniform distribution. Interestingly, the Fe element exists in the core of octahedron and appears in the nanosheets shell. This indicates that the Fe-MIL-101-NH2 precursor was etched by CoCl2 and NiCl2 during the ion exchange process [24,40], and the released Fe ions regrowth on the surface of octahedron together with Ni and Co ions, forming solid ternary solutions (CoNiFe-MOF). The broad peak at about 22\u00b0 in XRD pattern of CoNiFeOx-NC (Fig. 2f) should be the characteristics peak of amorphous C-N compounds (carbonitrides) [41], which is generated from the part decomposition of organic ligands. The considerable amount of C and N elements detected by the EDX (Fig. 2i) and STEM (Fig. S2) verifies the presence of carbonitrides in the sample, which agrees with XPS results (Table S3). Besides, the substrate surrounding the metal oxide nanoparticles presents an amorphous state in the HRTEM images (Fig. S3), generally indicating the presence of amorphous carbon material. However, The FeOx-NC sample derived from pristine Fe-MIL-101-NH2 possesses a featureless morphology, including numerous large nanoparticles (Fig. 2d). The component of these nanoparticles was determined to be \u03b3-Fe2O3, according to the XRD patterns (Fig. 2f) and HRTEM images (Fig. 2g).To further investigate the impact of Co and Ni ions on the microstructure of CoNiFe-MOF, control samples were synthesized by solely introducing Co or Ni element into Fe-MIL-101-NH2 and compared with the as-prepared ternary metal-based MOF. ICP-AES results (Table S2) verify the successful incorporation of Co and Ni in the CoFe-MOF and NiFe-MOF, respectively. As TEM images displayed in Fig. 3\na, the introduction of Ni has little influence on the morphology of Fe-MIL-101-NH2. XRD patterns further confirm the unvaried topological structure for NiFe-MOF (Fig. 3b). As a previous study revealed, the FeO6 geometries in the MIL-53 frameworks can be partly substituted by the NiO6 ones, indicative of the similar coordination ability of Ni and Fe ions with carboxylate groups [30]. However, the introduction of Co changed the original octahedral morphology to concaved octahedrons (Fig. 3a). This is also reflected in the XRD patterns of CoFe-MOF (Fig. 3b), where the diffraction peaks are distinguished from those of Fe-MIL-101-NH2. This could arise from the strain-stress induced by the Co2+ with a larger ionic radius (0.0745\u2009nm) than Fe3+ (0.0645\u2009nm) and Ni2+ (0.0690\u2009nm) [42,43] or the different coordination preference with carboxylate groups between Fe and Co [44,45], in the enlightenment of the previous declaration that the substitution process in MOF is significantly metal ion-dependent [46]. Interestingly, the monometallic incorporation of Ni or Co could not afford the analogous hierarchical structure of CoNiFe-MOF, suggesting the indispensable role of mixed Co and Ni precursor during the metal ion exchange process.Fig. S4a-d show the TEM images of CoNiFe-MOF with different ion exchange times. It was observed that as the exchange time increased, more ultrathin nanosheets formed on the surface of pristine Fe-MIL-101-NH2, and the octahedral core part gradually diminished in size. Fig. S5 gives the TEM images of CoNiFe-MOF with various Co precursor amounts (0.6, 1.4, 2.8, and 5.6\u2009mmol), which shows that decreasing Co2+ amount inhibits the formation of nanosheets. More evidences were given in a controlled ion-exchange process without Co ions addition but different amounts of HCl addition (0.9, 1.7, 3.5, and 6.0\u2009mmol). As shown in Fig. S6, the inhibited nanosheets formation further certify the important role of Co2+ on the formation of nanosheets shell. Based on the above results, we deduce the possible growth mechanism of the hierarchical structured CoNiFe-MOF as following (Fig. 3c): During the ion exchange process, Ni ion substituted part of Fe atoms in the frameworks of Fe-MIL-101-NH2, while Co ion etched the surface of Fe-MIL-101-NH2, providing the favorable conditions for the second building of metal-organic frameworks, then the dissolved Fe ions re-coordinated with Ni, Co ions and carboxylate groups to form tri-metallic MOF nanosheets [40,47]. During the growth of tri-metallic MOF nanosheets, the similar coordination ability with carboxylate groups between Ni and Fe ions facilitates the preferred coordination of Fe ions with Ni ions and carboxylate groups rather than Co ions. As a result, the as-synthesized tri-metallic MOF nanosheets are mainly composed of Fe and Ni solid solutions with limited amount of Co dopant. Similar dissolution-coordination phenomenon has also been reported by other researchers. Xun Wang et al. revealed the in-situ transformation from Zn/Ni-MOF-5 nanocubes to Zn/Ni-MOF-2 nanosheets with pre-formed nanocubes acting as supporting template without any surfactants [48]. David Lou et al. observed a cooperative etching-coordination-reorganization process when introducing the guest metal salt as a Lewis acid into ZIF-67 [40]. It should be emphasized that there is no prominent diffraction peak corresponding to layered double hydroxide (LDH) found in the XRD patterns of CoNiFe-MOF even after 48\u2009h of ion exchange treating (Fig. S4e). Besides, the DMF solvent that we used in the ion exchange process is not favorable for the formation of LDH.The OER electrocatalytic performances of FeOx-NC, CoFeOx-NC, NiFeOx-NC, and CoNiFeOx-NC were evaluated in Ar-saturated 1.0\u2009M KOH solution. Firstly, the linear sweep voltammetry (LSV) was recorded in the range of 1.05\u20131.85\u2009V vs. RHE at 5\u2009mV s\u22121 scan rate (Fig. 4\na). According to the iR-corrected LSV results, the overpotential of FeOx-NC, CoFeOx-NC, NiFeOx-NC, and CoNiFeOx-NC catalyst at the current density of 50\u2009mA cm-2 is 526, 361, 289, and 265\u2009mV respectively. CoNiFeOx-NC catalyst behaves the best catalytic activity than the other samples with nearly 100 % Faradaic efficiency (Fig. S7) towards OER, which is also superior to the commercial IrO2 catalyst and among the best in recently reported advanced OER catalysts (Fig. 4f). There is a prominent oxidation peak at 1.40\u2009V vs. RHE in the LSV curve of CoNiFeOx-NC, which should be attributed to the oxidation peak of the remaining MOF in CoNiFeOx-NC because of the protection of ternary-metal based shell during the air annealing process. Then the Tafel plots were recorded to study their OER reaction kinetics (Fig. 4b). The Tafel slope was calculated to be 99.9, 64.0, 47.6, and 64.1\u2009mV dec\u22121 for FeOx-NC, CoFeOx-NC, NiFeOx-NC, and CoNiFeOx-NC, respectively, demonstrating the enhanced reaction kinetics of CoNiFeOx-NC after Co and Ni incorporation. Meanwhile, electrochemical impedance spectroscopy (EIS) experiments display that CoNiFeOx-NC electrode possesses the smallest charge transfer resistance (Rct\u2009=\u20091.65 \u03a9) compared with FeOx-NC (35.17 \u03a9), CoFeOx-NC (19.42 \u03a9), NiFeOx-NC (4.77 \u03a9), and IrO2 (2.24 \u03a9) electrodes (Fig. 4c). To estimate their electrochemical active surface area (ECSA), the double-layer capacitance (C\ndl, Fig. 4d) was measured by recording the cyclic voltammograms (CVs) at scan rate from 5 to 50\u2009mV s\u22121 within the non-faradic voltage range (1.107\u20131.207\u2009V vs. RHE, Fig. S8). According to the equation ECSA = C\ndl / C\ns, the ECSA is directly proportional to the C\ndl because C\ns is constant as the specific capacitance of per unit area of material with an atomically smooth surface. Therefore, IrO2 catalyst displays the highest ECSA, and CoNiFeOx-NC displays much higher ECSA than FeOx-NC, CoFeOx-NC, and NiFeOx-NC.Additionally, the effect of synthesis conditions on the OER performance of CoNiFeOx-NC was investigated by varying some crucial factors, such as the cation precursor ratio (NiCl2/CoCl2, Table S4) in the ion-exchange process and the calcination temperature (200, 300, 400, and 500\u2009\u00b0C) in the pyrolysis process. As shown in Fig. S9, the NiCl2/CoCl2 ratio significantly affects the formation of nanosheets shell outside the MOF core, thus endowing different OER performances (Fig. S10). As shown in Fig. S12, CoNiFeOx-NC samples obtained at other calcination temperatures (200, 400, and 500\u2009\u00b0C) behave the inferior OER activities than the as-prepared CoNiFeOx-NC sample at 300\u2009\u00b0C. XRD and TEM results (Figs. S14 and S15) indicate that the high calcination temperature could result in nanoparticle agglomeration due to the collapsing of 3-D structure. This also suggests the significant role of carbonitrides part in CoNiFeOx-NC, which is providing a 3-D framework that could afford large specific surface area for CoNiFeOx active sites distribution. Nevertheless, in this catalyst, the carbonitrides part of CoNiFeOx-NC is not the main active site for OER reaction as revealed by the performance outcomes of carbonitrides itself (Fig. S18).The durability properties of electrocatalyst is another crucial factor for renewable energy application systems. Firstly, the polarization curve of the CoNiFeOx-NC electrode after 1500 consecutive CV scans was compared with the initial one (Fig. 4e), which shows a slight decrease of only 7\u2009mV at the current density of 50\u2009mA cm\u22122. Then the long-term chronoamperometry method was applied to test the stability of CoNiFeOx-NC anode at the constant voltage of 1.49\u2009V. As shown in the inset of Fig. 4e, the CoNiFeOx-NC catalyst can retain 96.6 % of its initial activity after 40\u2009h electrocatalysis, while the commercial IrO2 catalyst losses about 71.3 % of its pristine activity. Besides, the stability of counterpart obtained from in-situ generated CoNiFe-MOF (donated as IS-CoNiFeOx-NC) was evaluated at the constant voltage of 1.61\u2009V (Fig. S22), and the current density of IS-CoNiFeOx-NC also deteriorates quickly during the long-term test (maintained 70.4 % activity after 40\u2009h electrocatalysis). The crystal phase and microstructure of CoNiFeOx-NC after 40\u2009h stability tests were analyzed by XRD and HRTEM characterizations (Fig. S23), which indicates that the spinel metal oxides still exist after the stability test. Meanwhile, the XPS spectra results of post-OER CoNiFeOx-NC also display similar information with the original CoNiFeOx-NC sample (Fig. S24). All of the above proofs confirm that the as-prepared CoNiFeOx-NC catalyst is considerably stable for the OER reaction.The excellent OER actiity of CoNiFeOx-NC could be attributed to its unique structure concerning FeOx-NC, CoFeOx-NC, and NiFeOx-NC. Firstly, the formed 2-D MOF layers encapsulating 3-D MOF octahedral cores prevents the collapse of the frameworks during air calcination. The stable 3-D framework composed of plenty of carbonitrides compound contributes to the generation of highly dispersed CoNiFeOx nanoparticles in CoNiFeOx-NC (Fig. 2h). As shown in Fig. S25, for NiFe-MOF and CoFe-MOF, the original octahedron morphology has almost disappeared after air heating treatment. Meanwhile, TG curve results (Fig. S26) also show that CoNiFe-MOF is more heat resistant than Fe-MIL-101-NH2 under the air atmosphere. Secondly, the ion exchange induced dissolution-recoordination process between Fe, Co, and Ni triggered a fluffy nanosheets structure with rich channels and large surface area. The porosity properties of FeOx-NC, CoFeOx-NC, NiFeOx-NC, and CoNiFeOx-NC were characterized by N2 adsorption-desorption isotherms. As shown in Fig. 5\na-b, all of them displayed typical IV isotherm characteristic of mesoporous materials, but CoNiFeOx-NC has the largest pore volume and BET surface area (104\u2009m2\u2009g\u22121). This result echoes the much higher ECSA surface area of CoNiFeOx-NC than that of FeOx-NC, CoFeOx-NC, and NiFeOx-NC (Fig. 4d). The expansion of pore size distribution and specific surface area in CoNiFeOx-NC may contribute to the exposure of active sites to electrolyte and effective mass transport.Thirdly, the unique modulation of electronic structure through electron transfer between Co, Ni, and Fe could optimize the adsorption energy of oxygen species on the catalyst surface. Fig. 5c shows the deconvoluted XPS spectra of the Fe 2p signals for FeOx-NC, CoFeOx-NC, NiFeOx-NC, and CoNiFeOx-NC samples. The 2p3/2 peaks at around 710.7 and 713.1\u2009eV, along with the 2p1/2 peaks at around 724.0 and 726.0\u2009eV should be indexed to the Fe3+ from the different binding environment (octahedral and tetrahedral sites). This is because CoFe2O4 and NiFe2O4 possess an inverse spinel structure with Co2+ or Ni2+ in octahedral sites and Fe3+ equally distributed between octahedral (Oh) and tetrahedral (Td) sites of the O2\u2212 fcc cell [26,49]. The peaks at around 718.6 and 732.6\u2009eV are satellite peaks induced by the charge transfer from Fe to O [50]. The Fe 2p peaks shift to higher binding energy after Ni or Co incorporation, revealing the charge transfer from Fe to Co or Ni.Moreover, the binding energy of Ni 2p peaks in CoNiFeOx-NC is lower than that of NiFeOx-NC (Fig. 5d), suggesting the charge transfer from Co to Ni, which is further confirmed by the higher shift of Co 2p peaks in CoNiFeOx-NC in contrast to that of CoFeOx-NC (Fig. 5e). Therefore, a general charge transfer trend of Fe\u2192Co\u2192Ni was observed in the as-prepared materials. This is consistent with the electronegativity order of Fe\u2009<\u2009Co\u2009<\u2009Ni as the more electronegative Ni element attracts more electrons [51]. It should be stated that Ni and Co elements are present as the divalent form in CoNiFeOx-NC and NiFeOx-NC, corresponding to the tetrahedral binding sites in the spinel structure. However, trivalent Co was found in the Co 2p spectra of CoFeOx-NC. This is due to the formation of Co3O4 species in CoFeOx-NC sample, as seen in XRD patterns (Fig. S27). The high-resolution XPS spectra for C, O, and N are also given in Figs. S28\u2013S31, confirming the existence of carbonitrides and metal oxides in all the samples. Additionally, as reflected in the O1s spectra of CoNiFeOx-NC (Fig S28d), the area ratio of CO and CO species is much larger than the metal oxides species in comparison to the other three samples. This also proves that the MOF core was finely protected by the outside 2-D layers in CoNiFeOx-NC during the air heating treatment.To further get insight into the electronic structure modifications and the corresponding effects on the OER activity, DFT\u2009+\u2009U calculations were deliberately performed to investigate the relationships between the electronic structure and OER mechanism in different spinel metal oxides. The spinel oxide crystal was chosen as the catalyst model since all the catalysts displayed spinel crystal structures according to Figs. 2 and S27, which is \u03b3-Fe2O3 (110), CoFe2O4 (110) and NiFe2O4 (110) surface model for FeOx-NC, CoFeOx-NC and NiFeOx-NC catalyst. The content of Co in CoNiFeOx-NC is slight compared with the contents of Ni or Fe, and the lattice of spinel oxide in CoNiFeOx-NC expanded (Fig. 2), suggesting the incorporation of larger heteroatom such as Co. Therefore, the Co doped NiFe2O4 (110) surface model was adopted to represent CoNiFeOx-NC catalyst. Firstly, the electronic structure of Co-NiFe2O4 was analyzed by comparing it with NiFe2O4. Fig. 6\na shows the contour charge maps of the charge density distributions for NiFe2O4 and Co-NiFe2O4 along the 110 facet, in which the yellow and blue regions correspond to the electrons accumulation and depletion respectively. It\u2019s found that the electron-deficient state around the Fe site is apparently intensified in contrast to the corresponding Ni or Co site. In Co-NiFe2O4, the evident electron depletion was also observed around Co site, whereas the electron accumulation is more pronounced around Ni than that in NiFe2O4, indicating the partial electron transfer from Co site to Ni site via O bridge. These simulation results agreed well with the XPS results where the electron transfer follows the trend of Fe\u2192Co\u2192Ni, and the charge density distribution of CoFe2O4 is also consistent with the above trend (Fig. S32).The diversity of electronic structure in metal oxides will lead to the different adsorption conditions for the reaction intermediates. Herein, the adsorption geometries and adsorption energies of three primary OER intermediates (OH*, O*, and OOH*) on the \u03b3-Fe2O3 (110), CoFe2O4 (110), NiFe2O4 (110) and Co-NiFe2O4 (110) catalysts surface were calculated and compared. On each catalyst surface, both Fe Td site and Fe/Co/Ni Oh site (referring to Fe Oh site for \u03b3-Fe2O3, Co Oh site for CoFe2O4, Ni Oh site for NiFe2O4, and Ni-Co coordinated Oh site for Co-NiFe2O4, respectively) were chosen as the adsorption sites. The optimized adsorption geometries were shown in Fig. S33, and the calculated adsorption energies were displayed in Fig. S34 (Td site) and Fig. 6b (Oh site). The adsorption energy results indicate that the oxygen species prefer to binding with Oh site with larger binding energies rather than Td site for all the catalysts, and the Gibbs free energy results (Table S5) in the following discussion also demonstrate a lower reaction energy barrier on Oh site than Td site. Therefore, the Oh site (Fe Oh site for \u03b3-Fe2O3, Co Oh site for CoFe2O4, Ni Oh site for NiFe2O4, and Ni-Co coordinated Oh site for Co-NiFe2O4, respectively) was received as the preferred adsorption site for oxygen species in this system. As shown in Fig. 6b, the adsorption energies for three types of oxygen species on the Oh site follow the trend of \u03b3-Fe2O3 (110)>CoFe2O4 (110)>Co-NiFe2O4 (110)>NiFe2O4 (110). According to the above electronic structure results, the electron-deficient state in spinel structure is in the order of Fe\u2009>\u2009Co\u2009>\u2009Ni. Consequently, the binding strength of Fe Oh site on \u03b3-Fe2O3 (110) with oxygen species is the strongest, followed by the Co Oh site on CoFe2O4 (110), and the Ni Oh site on NiFe2O4 (110) is the weakest. Impressively, after Co incorporation, the Ni-Co coordinated Oh site on Co-NiFe2O4 (110) can provide moderated oxygen-binding strength (e.g. EOH* = \u22120.83\u2009eV) compared with the original Ni Oh site (e.g. EOH* = \u22120.39\u2009eV) and Co Oh site (e.g. EOH* = \u22121.01\u2009eV). This binding strength modulation optimized the adsorption of oxygen species on Co-NiFe2O4 (110), hence causing OER intermediates transformation multi-steps more readily accessible on CoNiFeOx-NC. This assumption was further verified by the theoretical energy barrier results of four catalysts towards OER. Fig. 6c displays the Gibbs free energy diagrams along the proposed OER pathway on different catalysts. For \u03b3-Fe2O3 (110), CoFe2O4 (110), and NiFe2O4 (110), the third elementary reaction step has the largest energy barrier of 2.44, 2.31, and 2.02\u2009eV respectively, which can be considered as the rate-determining step. However, for Co-NiFe2O4 (110), the energy barrier of the third step decreases significantly to 1.82\u2009eV, and the rate-determining step turns to the last step (1.92\u2009eV). This means Co-NiFe2O4 (110) requires a lower overpotential to drive water oxidation. Meanwhile, the backward rate-determining step is commonly corresponding to the acceleration of reaction kinetics [52]. The theoretical energy barrier results agree well with the experimental results, where the overpotential at the 50\u2009mA cm\u22122 is in the order of FeOx-NC\u2009>\u2009CoFeOx-NC\u2009>\u2009NiFeOx-NC\u2009>\u2009CoNiFeOx-NC. Finally, the typical OER reaction mechanism on Co-NiFe2O4 (110) was depicted in Fig. 6d.In summary, we have demonstrated a hierarchically structured CoNiFeOx-NC catalyst with the superior performance of the OER rection, derived from core-shell structured MOF. Co and Ni ions display the indispensable and cooperative role for the formation of 2-D ternary metal MOF layers encapsulating 3-D MOF octahedral crystals in the ion exchange process. Co ions provide favorable environments for the second building of metal-organic frameworks by etching the surface of Fe-MIL-101-NH2, while the dissolved Fe ions prefer to re-coordinating with Ni, Co ions, and carboxylate groups to form tri-metallic MOF nanosheets. The formed nanosheets shell not only protects the frameworks from being destroyed during the air calcination treatment but also helps maintain a large surface area and porous structure of the materials. More importantly, an electron transfer trend of Fe\u2192Co\u2192Ni was found in the as-prepared CoNiFeOx-NC material based on the experimental and DFT studies. This optimized electronic structure results in moderated oxygen-binding strength on Ni-Co coordinated Oh site in contrast to the original Ni Oh site and Co Oh site, thus lowering the theoretical energy barriers for OER. Benefitting from the above features, the well-designed CoNiFeOx-NC catalyst delivers high efficiency in the alkaline OER reaction.\nChen Chen: Conceptualization, Methodology, Investigation. Yongxiao Tuo: Software, Formal analysis, Writing - original draft. Qing Lu: Validation, Investigation. Han Lu: Validation. Shengyang Zhang: Validation. Yan Zhou: Funding acquisition. Jun Zhang: Supervision, Funding acquisition, Writing - review & editing. Zhanning Liu: Data curation. Zixi Kang: Writing - review & editing. Xiang Feng: Writing - review & editing. De Chen: Supervision, Writing - review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was partially supported by the Taishan Scholar Project of Shandong Province, the National Natural Science Foundation of China (No. 21805308), the Key Research and Development Project of Shandong Province (No. 2019GSF109075), the Fundamental Research Funds for the Central Universities (No. 18CX06065A, No. 20CX06022A).Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.apcatb.2021.119953.The following is Supplementary data to this article:\n\n\n\n\n", "descript": "\n                  Metal-organic frameworks (MOFs) have recently emerged as promising precursors to construct efficient non-noble metal electrocatalyst for oxygen evolution reaction (OER). Herein, a Co-Ni-Fe spinel oxide-carbonitrides hybrids (CoNiFeOx-NC) electrocatalyst with hierarchical structure was synthesized from Fe-MIL-101-NH2 through a unique ion-exchange based strategy. The ion exchange of Fe-MIL-101-NH2 with both Ni and Co ions induced a hierarchically structured 2-D ternary metal MOF shell layer encapsulated 3-D octahedral MOF crystals as a core. This prevents the collapse of MOF frameworks during the air calcination process and affords highly porous structure and large surface area. Additionally, the unique combination of Co-Ni-Fe in spinel oxides derived from calcination of the hierarchically structured core-shell MOF provides a favorable electronic environment for the adsorption of OER intermediates, which was further verified by the XPS characterizations and DFT calculations. DFT study revealed the Ni-Co coordinated Oh sites in the MFe2O4 reverse spinel structures as the main active sites, which tuned the binding strength of oxygen species with a catalyst through electron transfer of Fe\u2192Co\u2192Ni, thereby lowered the energy barriers for OER. As a result, the rationally designed CoNiFeOx-NC catalyst manifests superior OER performance with a low overpotential of 265\u2009mV at 50\u2009mA cm\u22122 and a decent Tafel slope of 64.1\u2009mV dec-1. The ion-exchange based strategy may serve as a versatile platform for rational design and synthesis of multi-metallic MOF derived electrocatalysts.\n               "}
{"full_text": "", "descript": "\n                  A process for the selective extraction and separation of vanadium and nickel from spent-residue oil hydrotreating catalysts by a direct acid leaching\u2212solvent extraction method was studied. The extraction and separation of vanadium(IV) and nickel(II) are divided into two stages: acid coleaching of vanadium and nickel and solvent extraction. In the acid coleaching stage, the leaching ratios of vanadium and nickel reach 88.07% and 75.58%, respectively, which can realize highly effective coleaching. In the solvent extraction stage, countercurrent experiments show that the extraction ratio of vanadium can reach 99.21% after a three-stage extraction with P204 as the high-efficiency extractant of vanadium in the acidic environment, while nickel and iron are not extracted. After the anti-extraction solution is pretreated by aluminum precipitation, the extraction ratio of nickel reaches 99.79% after a three-stage extraction with LIX84-I as a high-efficiency extractant of nickel in ammonia medium. A process flow for the recovery of vanadium and nickel is proposed, which not only can realize the separation and recovery of vanadium and nickel but also can realize the recycling of reagents.\n               "}
{"full_text": "Generally, catalysts include homogeneous catalysts and heterogeneous catalysts. Homogeneous catalysts have relatively uniform active centers, higher activity and selectivity, and fewer side reactions, but they are hard to separate, recover, and regenerate from the reaction system. Heterogeneous catalysts are easily separated from the system and can be used repeatedly, but the activity and selectivity of heterogeneous catalysts are often worse than that of homogeneous catalysts. Importantly, single-atom catalyst (SAC), a new type of catalyst, was developed, which is considered to be a bridge between homogeneous catalysts and heterogeneous catalysts due to the distinguished selectivity, catalytic activity, and easy to separation (Chen et\u00a0al., 2018c; Mao et\u00a0al., 2019; Sun et\u00a0al., 2019b; Yang et\u00a0al., 2013). In recent years, many research have contributed to the developments of single atoms catalysis, but the understanding of single atoms from atomic and electronic insights is still inadequate due to the deficiency of characterization techniques. Therefore, it is very important to comprehend the development process, synthesis methods, and coordination regulation approaches thoroughly for SACs.In 1995, Thomas and colleagues studied isolated single atom of Titanium as the active site of heterogeneous catalyst (Maschmeyer et\u00a0al., 1995). In 2000, the presence of single atoms was discovered when size-selected Pdn (1\u00a0\u2264 n \u2264 30) cluster supported on MgO were prepared by using a mass separation soft landing technique (Abbet et\u00a0al., 2000). In 2003, single-site Au species on ceria-based catalyst for water-gas shift were reported by Fu and colleagues (Fu et\u00a0al., 2003). In 2007, mesoporous Pd/Al2O3 with single sites was prepared by impregnation method for selective aerobic oxidation of allyl alcohol (Hackett et\u00a0al., 2007). With the development of characterization techniques, the concept of \u201csingle atom\u201d was first proposed by Zhang and colleagues in 2011 (Qiao et\u00a0al., 2011). The isolated single Pt atoms fabricated on the surfaces of iron oxide (Pt1/FeOx) displayed high activity and selectivity in CO oxidation. In recent years, the design and preparation of atomically dispersed catalysts have attracted extensive research interests in plenty of applications, such as photocatalysis, organic catalysis, electrocatalysis, and environmental and energy aspect. Meanwhile, because of the high surface free energy of single atoms, it is still a major challenge to increase the loading capacity of single atoms. What\u2019s more, the coordination environments, including the coordination number, the coordination atom, and the distance between the center atoms and the neighboring atoms, greatly influence the catalytic activity of single-atom catalysts (Cook and Borovik, 2015; Mao et\u00a0al., 2019; Sun et\u00a0al., 2019b; Tao et\u00a0al., 2020). Therefore, how to systematically regulate the coordination environments is of great significance to the screening of efficient single-atom catalysts (Lang et\u00a0al., 2019; Liang et\u00a0al., 2019; Liu et\u00a0al., 2019a; Qiao et\u00a0al., 2015).The coordination environment and the loading of metal atoms are closely related to the catalytic performance of SACs. In order to further study the regulation of coordination environment and loading, a variety of materials are used as supports for single-atom catalysts, including metal and metal oxide (Cao and Lu, 2020; Ma et\u00a0al., 2021), sulfide (Feng et\u00a0al., 2018; Li et\u00a0al., 2022), phosphide (Jiang et\u00a0al., 2020), zeolites (Sun et\u00a0al., 2019a), metal-organic frameworks (MOFs) (Zhou et\u00a0al., 2021a), covalent organic frameworks (COFs) (Liu et\u00a0al., 2020), and carbon-based materials (Guo et\u00a0al., 2021), such as graphene, graphdiyne, and hexagonal boron nitride. Among these carriers, carbon-based materials are regarded as the promising candidate materials for large-scale production of SACs due to low cost, superb conductivity, tunable physicochemical property, and high specific surface area (Georgakilas et\u00a0al., 2015; Shaik et\u00a0al., 2019; Su et\u00a0al., 2013). Therefore, the overview and summary of carbon-based-material-supported SACs are necessary for the improvement of future work.In this review, we summarized the synthesis methods of SACs supported on carbon-based materials and then highlighted the great significance to guide the coordination regulation of single atoms and improve the loading of SACs. Then, we introduced the advanced characterization techniques, including ex situ and in situ technologies, which is vital to learn about the SACs from atomic and electronic levels. Most important of all, the applications of carbon-based-material-supported SACs in electrocatalysis are discussed by combining calculations and experiments, and the coordination environment and metal loading of the SACs are emphasized, involving hydrogen evolution reaction (HER), oxygen evolution reaction (OER), oxygen reduction reaction (ORR), carbon dioxide reduction reaction (CO2RR), and nitrogen reduction reaction (NRR). In the end, the challenges and development opportunities of SACs are fully discussed. We believed that this review could provide essential information to rationally construct SACs in the future.SACs usually display superb catalytic activities for electrochemical reactions, hydrogenation reactions, and so on. However, the highly dispersed single-atoms tend to migration and agglomerate during the synthesis process due to the high surface energy. How to synthesize stable and high-loaded single-atom catalysts requires continuous exploration. In this section, the methods of atomic layer deposition (ALD), impregnation strategy, electrochemical deposition, chemical vapor deposition (CVD), balling milling, and pyrolysis strategy are summarized. Notably, impregnation strategy and pyrolysis strategies are beneficial to the large-scale preparation of SACs in industry.Atomic layer deposition (ALD) has a promising future in the synthesis of small size catalysts. In the ALD process, the substrate is alternately exposed to different reactions of precursor vapors. The material could be deposited in atomic layers by sequential and self-limiting surface reactions. Previous studies have shown that ALD can control the morphologies of deposited metals, from tiny discrete nanoparticles to continuous films, by surface chemistry strategy (King et\u00a0al., 2008; Liu et\u00a0al., 2009). Meanwhile, ALD can precisely control the size and distribution of particles on the substrate (George, 2010; Marichy et\u00a0al., 2012). In 2013, single Pt atoms on graphene nanosheets with low coordination were prepared by using ALD technique for the first time (Sun et\u00a0al., 2013). The container for graphene was placed directly on the heated stage of ALD. MeCpPtMe3 acted as the precursor keeping at 65\u00b0C to provide a steady state flux to the reactor (\u223c800 mTorr). Gas lines were heated to 100\u2013150\u00b0C to avoid precursor condensation. High purity O2 and N2 were used as the counter reactant and purging gas, respectively. The reaction took place at 250\u00b0C with the pulse of 1\u00a0s of MeCpPtMe3, 20\u00a0s of N2 purge, and 5\u00a0s of O2 in an ALD cycles. By increasing the number of ALD cycles, the size of prepared catalysts can be precisely regulated, including single atoms, sub-nanometer clusters, and nanoparticles, while affecting the coordination number of the metal center atom. Similarly, isolated Pt single-atoms and clusters were fabricated on the nitrogen-doped graphene nanosheets through the same ALD synthesis approach expect for the carrier (Cheng et\u00a0al., 2016; Figure\u00a01A). There are numerous individual Pt atoms and very small clusters in the ALD50Pt/NGNs catalysts, whereas some Pt clusters grow into nanoparticles after 100 cycles. Meanwhile, the HER catalytic performance decreased with the increased number of ALD cycles, and ALD Pt/NGN catalysts showed higher catalytic activity and long-term stability in comparison with the commercial Pt/C catalysts in 0.5\u00a0M H2SO4. What\u2019s more, the Pd1/graphene with Pd-C2 coordination were prepared by anchoring the palladium hexafluoroacetylacetate (Pd(hfac)2) and formalin at 150\u00b0C on a viscous flow reactor (Yan et\u00a0al., 2015a). The (Pd(hfac)2) was held at 65\u00b0C to get a sufficient vapor pressure. Likewise, the 99.999% of purity N2 act as carrier gas at a flow rate of 200\u00a0mL min\u22121. The manifold was kept at 110\u00b0C to avoid precursor condensation. It showed outstanding activity, selectivity, and durability than the Pd-NPs/graphene in selective hydrogenation of 1, 3-butadiene with about 100% butenes selectivity at 95% conversion at about 50\u00b0C. ALD can also be used to prepare carbon-based-material-supported nonnoble metal single-atom catalysts. The isolated Co atoms anchored on graphene (Co1/G) with tunable high metal loading and a six-coordination of Co1-O2C4 were fabricated by ALD (Yan et\u00a0al., 2018b). The amount of metal loading can be modulated by simply controlling the number of turns of the ALD. As shown in Figure\u00a01B, the first cycle of Co ALD was carried out on thermally reduced graphene oxide by exposing the carrier to CoCp2 vapor. Subsequently, the molecule O3 is injected into the cavity to remove the ligand. Although the new active site is reconstructed, another batch of Co atoms is loaded in the subsequent ALD cycle. The load density of Co1/G catalyst was precisely adjusted by controlling the cycles of Co ALD. A series of Co1/G catalysts with loading capacities of 0.4, 0.8, 1.3, 2.0, and 2.5 wt% were obtained by 1, 2, 3, 4, and 5 ALD cycles, respectively. The Co1/G SACs exhibited high activity and selectivity for the hydrogenation of nitroarenes to produce azoxy aromatic compounds.By adjusting the number of cycles, order, and type, ALD can achieve atomically fine control over the structure of catalyst active sites, providing a bottom-up strategy for precise and controllable catalyst synthesis. However, the high cost and low yield are the primary reasons for limiting the industrial application of ALD method.Impregnation method, as one of wet-chemical methods, is considered to be the most promising route for mass production due to the low price and easy operation. Impregnation method is widely used in the preparation of supported catalysts, especially low-content noble-metal-supported catalysts. Yin et\u00a0al. realized coordination regulation through impregnation method and confirmed better HER performance in low coordination environment (Yin et\u00a0al., 2018). Pt single-atom supported on graphdiyne (GDY) was prepared by wet-chemical route. The coordination environment is controlled by controlling facile annealing step. React GDY with K2PtCl4 aqueous solution at 0\u00b0C for 8\u00a0h (named as Pt-GDY1) and wash with plenty of water. Then annealing in Ar atmosphere at 200\u00b0C for 1\u00a0h (named as Pt-GDY2) (Figure\u00a01C). The Pt-GDY2 with four-coordinated C2-Pt-Cl2 exhibits higher mass activity up to 3.3 times than Pt-GDY1 with five-coordinated C1-Pt-Cl4. Higher total unoccupied density of states of Pt 5d orbital and close to zero |\u0394GPt\nH\u2217| value makes Pt have higher HER catalytic activity. Hetero-atom doping modification of the carrier affects the coordination environment. For example, atomically dispersed electrocatalysts (ADCs) with Ru-C5 single atoms and Ru-O4 nanoclusters were fabricated in S-doped carbon black by using impregnation strategy in room temperature (Cao et\u00a0al., 2021). Activated carbon and 2, 2-bithiophene were grinded fully and then calcined in a tube furnace at 800\u00b0C for 2\u00a0h under N2 atmosphere to obtain the S-doped carbon material. Thirty milligrams of S-doped carbon material and 20\u00a0mL water were mixed intensively in a beaker; 0.05\u00a0mmol RuCl3\u00b7xH2O dissolved in 5\u00a0mL deionized water was dropped into the above solution and stirred for another 6 h. Then, the mixture was centrifuged and dried in a vacuum drying oven at room temperature to obtain the atomically dispersed Ru catalyst. Meanwhile, dual-site Ir, Rh, Pt, Au, and Mo ADCs can also be prepared by this method. The Ru ADCs show enhanced HER performance in alkaline solution due to the synergic effect between single-atoms and sub-nanoclusters. What\u2019s more, Cu-SA/SNC with low-valence Cu(+1)-N4-C8S2 was constructed by impregnation method with single copper atoms embedded in a sulfur and nitrogen-modified carbon support (Jiang et\u00a0al., 2019). Na2S\u00b79H2O and S powder were dissolved in deionized water by ultrasonic dissolving for 5\u00a0h at ambient conditions. Then, the solution was heated at 80\u00b0C for 12\u00a0h in a Teflon autoclave to obtain S precursor. CuPc, DCDA, and trimesic acid were dissolved in deionized water. Then S precursor was dropped into the above solution for continuously stirring and drying at 80\u00b0C. Next, the mixture was annealed at 900\u00b0C for 2\u00a0h under N2 atmosphere, and the samples were leached in 0.5\u00a0M H2SO4 solution at 80\u00b0C for 24\u00a0h to remove the free-standing metallic residues. Synthesizing single atom catalysts by the impregnation method is simple, without complicated and expensive equipment (Sun et\u00a0al., 2020). Therefore, this method is very suitable for large-scale synthesis of single-atom catalysts. Yang et\u00a0al. successfully synthesized a series of M-SACS catalysts (M\u00a0= Ni, Mn, Fe, Co, Cr, Cu, Zn, Ru, Pt, and their combinations) by complexing a series of metal cations with 1, 10-phenantholine and loading them on commercial carbon black (Yang et\u00a0al., 2019a). The synthetic approach enables large-scale (>1\u00a0kg) production of single-atom catalysts with high metal loadings. The synthesized Ni single-atom catalyst exhibits excellent activity in the electrochemical reduction of carbon dioxide to carbon monoxide. It provides an important approach for large-scale preparation of SACs by impregnation method.Electrochemical methods were regarded as the effective strategies to synthesize high-purity SACs because of low cost and simple applicability. The standard three-electrode system can quickly prepare the target sample and accurately control the catalyst preparation process by adjusting the workstation parameters, which has obvious advantages over the traditional wet chemical method.Atomic dispersed Ru-doped ultrathin Co(OH)2 nanosheet arrays (CoRu@NF) was fabricated by electrochemical deposition method (Zhu et\u00a0al., 2021; Figure\u00a02A), which shows excellent catalytic performance of OER in 1.0\u00a0M KOH and 0.1\u00a0M KOH solution. A standard three-electrode system was carried out for electrochemical deposition. The nickel foam (NF) was immersed 2\u00a0cm below the liquid surface; saturated calomel electrode (SCE) and the carbon rod were used as the working electrode, reference electrode, and counter electrode. Co(NO3)2\u22196H2O aqueous solution was poured into a 100\u00a0mL electrolytic cell. Cyclic voltammetry with the scanning potential of 0 \u223c \u22121.2\u00a0V (versus SCE), the scanning rate of 100\u00a0mV s\u22121, and the scanning cycles of 40 times was used. The introduction of Ru reduces the thickness of the nanosheets, exposing more active sites. Besides, single-atoms Ru were anchored on the surface of MoS2 nanosheets array supported by a carbon cloth with 3D porous structure based on theoretical predictions (Wang et\u00a0al., 2019; Figure\u00a02B). The MoS2/CC acts as the working electrode. Atomically Ru was electrodeposited by cycling MoS2/CC substrate from \u22120.5 to 0.4\u00a0V versus SCE at the sweep rate of 20\u00a0mV s\u22121 in the electrolyte containing RuCl3 and H2SO4 for 20 cycles. Finally, the Ru-MoS2/CC was taken out, washed with deionized water and dried by nitrogen flow. The catalyst displays HER catalytic performance comparable to commercial Pt/C under pH-universal conditions. What\u2019s more, Zeng\u2019s group reported the fabrication of SACs by electrochemical deposition method in a wide range of metals and supports (Zhang et\u00a0al., 2020). The cathodic voltage was from 0.10 to \u22120.40 V, the anodic voltage was from 1.10 to 1.80 V, and the scanning rate was 5\u00a0mV s\u22121. The processes were repeated for 10 times and 3 times in cathodic and anodic deposition, respectively. The experimental results showed that SACs displayed different electronic states due to different redox reactions between the cathodes and the anodes. More than 30 different SACs can be successfully fabricated from cathodic or anodic deposition only by varying different metal precursors and supports. Interestingly, the SACs deposited by cathode have higher activity for hydrogen evolution reaction, whereas the SACs deposited by anode have higher activity for oxygen evolution reaction.In general, the species and coordination environment of SACs can be changed by varying the supports and metal precursors. However, there are still few studies on the preparation of single atoms on carbon-based materials by electrodeposition due to the influence of electrodeposition equipment.As a kind of \u201ctop-down\u201d method, CVD is often used to synthesize single-atom catalysts. The research of CVD began in the late nineteenth century. Its principle is to introduce the reaction agent vapor and other gases required into the reaction chamber, by increasing the temperature, or other forms of energy, so that they have chemical reactions on the substrate surface to generate new solid substances deposited on the surface (Drosos and Vernardou, 2018; Zhang et\u00a0al., 2019).The CVD method consists of the following four steps: (1) The reaction gas diffuses to the surface of the material; (2) the reaction gas is adsorbed on the surface of the material; (3) the chemical reaction occurs on the surface of the material; (4) the gaseous by-products are separated from the surface of the material. Due to the nucleation or growth at the molecular level, CVD is more suitable for the formation of dense and uniform thin films on the surface of irregularly shaped substrates, and the deposition speed is fast and the quality of the film is very stable (Gardecka et\u00a0al., 2018). Some special films also have excellent optical, thermal, and electrical properties and thus easy to achieve mass production (Liu et\u00a0al., 2019b). For example, Miroslav and colleagues (Kettner et\u00a0al., 2019) synthesized Pd-Ga alloy supported on highly ordered pyrolytic graphite (HOPG) by vapor deposition under ultra-high vacuum, which is shown in Figure\u00a02D. Pd was deposited using a commercial electron beam evaporator from a Pd wire onto the HOPG substrate that was kept at room temperature, and Gallium was evaporated from a pyrolytic boron nitride crucible in a second electron beam evaporator at an angle of approximately 45\u00b0 with respect to the sample normal. The evaporation rates were 1.5\u00a0\u00c5/min for Pd and 0.5\u00a0\u00c5/min for Ga, respectively. Through STM/AFM characterization results, it can be seen that the HOPG-rich Pd-Ga alloy was prepared on HOPG. The Pd-Ga alloys of Ga exhibit superior Pd single-atom site properties and excellent stability. In addition, Mohammad and his team (Tavakkoli et\u00a0al., 2020) successfully synthesized N, Co, and Mo single-atom-decorated highly graphitized graphene nanoflake-carbon tube (CNT) composites by a one-step reactive vapor deposition method. As shown in Figure\u00a02E, the method first obtains the CoMo mixed catalyst by heating and calcination. Then, acetonitrile was added in the mixed atmosphere of H2/CH4 for N doping, and the carbon material was grown on the catalyst at 1000\u00b0C. A high specific surface area mesoporous material obtained by this method is favorable for the oxygen mass transfer process and exhibits high catalytic activity and stability (basic conditions) for OER and ORR. Through STEM images, single metal atoms can be clearly identified in the multilayer graphite films, as shown in Figure\u00a02F. However, the deposition temperature of CVD is usually very high, generally between 900\u00b0C and 2000\u00b0C, so it is usually used on carbon materials. However, high temperature can easily cause great damage to common materials, such as nickel foam, which limits the choice of substrates and deposition layers. At present, two aspects of medium, low temperature and high vacuum, are the main development directions of CVD (Malarde et\u00a0al., 2017).Ball milling can cut and reconstruct the chemical bonds of materials/molecules and is widely used in the preparation of carbon-based-material-supported single atoms. Moving balls with kinetic energy apply their energy to the materials, causing a single metal atom to be embedded on the surface of the carbon substrate (Yang et\u00a0al., 2020b). A series of graphene-embedded FeN4 (FeN4/GN) catalysts with different Fe content were prepared via high-energy ball milling (Deng et\u00a0al., 2015). Firstly, 2.0\u00a0g graphite flake and 60\u00a0g steel balls were placed in a hardened steel cylinder in a glove box, cleaned with high-purity argon for 20\u00a0min, and sealed. Various ratios of 2.0\u00a0g FePc and GN composites and 60\u00a0g steel balls (1\u20131.3\u00a0cm in diameter) were operated like before. Ball milling was agitated with 450\u00a0rpm for 20 h. A series of FeN4/GN samples with different Fe content were obtained. Coincidentally, single Fe atoms anchored on graphene nanosheets (FeN4/GNs) were fabricated by ball milling method for the direct conversion of methane to C1 oxygenated products at room temperature (Cui et\u00a0al., 2018). Other transition metals were also prepared by this method, including Mn, Fe, Co, Ni, and Cu. A series of M-N4-coordinated SACs were obtained by simply regulating the type of metal precursor salt. Fe and/or Co atomically dispersed within the 2D conjugated aromatic networks (CAN) were synthesized with the assistance of ball milling (Yang et\u00a0al., 2019b). Firstly, PMDA, urea, NH4Cl, (NH4)6Mo7O24\u00b74H2O, and a certain amount of metal chloride (FeCl3 and CoCl2\u00b76H2O) were mixed in a crucible. The metal polyphthalocyanine weas prepared by heating the mixture in a muffle furnace at 220\u00b0C for 3 h; 0.2\u00a0g above polyphthalocyanine and 15\u00a0mL deionized water were transferred into a zirconium dioxide capsule containing zirconium dioxide balls (0.5\u00a0mm in diameter). Ball milling was carried out at 1000\u00a0rpm for 1 h. The single-metal-atom-site density up to 10.7 wt% without agglomeration. CAN-Pc (Fe/Co) with Fe-N4 and Co-N4 coordination displays superior performance to benchmark Pt/C for ORR and Zn-air batteries. In one study, a rapid one-step mechanochemically induced self-sustaining reaction was proposed (Jin et\u00a0al., 2021b). Nitrogen-doped-carbon-supported single Co atoms were prepared by direct ball milling of cobalt (II) 5,10,15,20-tetrakis-(4\u2032-bromophenyl) porphyrin (Co-TPP-Br) and calcium carbide without the pretreatment of carbon support and further pyrolysis procedure (Figure\u00a02C). The mechanochemical energy can ignite and propagate a self-sustaining exothermic process, leading to the direct formation of carbon matrix to stabilize metal sites. The sample Co-BM-C with\u00a0CoN4 configuration prepared by ball milling (BM) showed excellent HER (\u03b710\u00a0= 126\u00a0mV) and OER\u00a0(\u03b710\u00a0= 240\u00a0mV) performance in 1.0\u00a0M KOH, showing great potential in overall water splitting (1.60\u00a0V\u00a0@ 10 mA cm\u22122).The pyrolysis strategy shows the merits of low price, environmental friendliness, and simplification in the synthesis procedures. The different types of metal atoms can be controlled by modulating the parameters in the synthesis process. Metal nodes in metal-organic framework (MOFs) are known to be atomically dispersed and have a well-coordinated environment, making them ideal precursor types for building SACs (Wang et\u00a0al., 2018a). The single tungsten atoms supported on MOF-derived N-doped carbon matrix was achieved successfully for HER applications (Chen et\u00a0al., 2018a). The W-SAC and MOF were prepared by pyrolysis strategy. Tungsten precursor (WCl5) was encapsulated in the skeleton of MOF (UiO-66-NH2) and then pyrolyzed at 950\u00b0C (Figure\u00a03A). The excess zirconia is removed by hydrofluoric acid solution to obtain W-SAC. It is important to note that the uncoordinated amines in UIO-66-NH2 play an important role in preventing the aggregation of W species. The catalyst displays 85\u00a0mV at a current density of 10 mA cm\u22122 in 0.1\u00a0M KOH, where HER catalytic performance is close to that of commercial Pt/C. In addition, zeolitic-imidazolate frameworks (ZIFs) also are often used to design templates and precursor for single-atom catalysts due to its flexibility and ultrahigh surface area (Xia et\u00a0al., 2015, 2016). ZIFs can be converted into amorphous or graphite-carbon frames by pyrolysis synthesis, thus providing a rich platform for the design of functional custom materials for electrocatalytic applications (Tang et\u00a0al., 2015; You et\u00a0al., 2015; Zheng et\u00a0al., 2014). Co\u2013Nx/C nanorod array derived from 3D ZIF nanocrystals was prepared through Zn2+ clusters that react with methylimidazole/PVP ligand to form ZIF nanocrystals, which catalyze the structural evolution of nanorods (Amiinu et\u00a0al., 2018) (Figure\u00a03B). Due to the synergistic effect of the chemical composition and abundant active sites of the nanorods, the catalysts show excellent ORR and OER performance compared with commercial Pt/C and IrO2.In order to achieve coordination regulation, Wang and colleagues prepared a series of single-Co-atoms catalysts with different nitrogen coordination numbers and studied their catalytic performance for CO2 reduction (Wang et\u00a0al., 2018b). Co/Zn ZIFs were synthesized at room temperature firstly, and then Zn would be evaporated away during the pyrolysis process. The Co ions would be reduced to single Co atoms anchored on nitrogen-doped porous carbon. SACs with coordination number from 2 to 4 were prepared by controlling volatile C-N fragments to adjust the number of N around central Co site through bimetallic Co/Zn ZIFs at 1000\u00b0C, 900\u00b0C, and 800\u00b0C of pyrolysis temperatures, respectively (Figure\u00a04A). Co nanoparticles were also prepared by pyrolysis of ZIFs containing pure Co. As can be seen from Figure\u00a04B and C, the Co-N2 catalyst maintained the initial ZIF morphology. EDX spectrum indicates that Co atoms are uniformly distributed throughout the structure (Figure\u00a04D). Meanwhile, atomically dispersed Co atoms can be directly observed from AC-HAADF-STEM (Figures\u00a04E and 4F). What\u2019s more, SAED with ring pattern indicates that the crystallinity of Co-N2 catalyst is poor (Figure\u00a04G). The optimum selectivity and activity are shown when Co is coordinated with two N atoms with 94% CO formation Faradaic efficiency and a current density of 18.1 mA cm\u22122 at an overpotential of 520\u00a0mV. Meanwhile, the turnover frequency (TOF) value of the CO formation is up to 18200 h\u22121. The results of experiments and theoretical calculations show that Co-N2 sites can promote the formation of CO2 into CO2\n\u2212 intermediates, thus enhancing the CO2RR performance. Moreover, improving the loading of SACs is another factor to promote the industrial application of single atom catalysis. Atomically dispersed transition metals anchored on nitrogen-doped carbon nanotubes (MSA-N-CNTs, where M\u00a0= Ni, Co, NiCo, CoFe, and NiPt) with high loading were fabricated through a multi-step pyrolysis strategy (Cheng et\u00a0al., 2018). Take the NiSA-N-CNTs for example, Ni(acac)2 was dispersed with dicyandiamide C2H8N2 in 100\u00a0mL solution and stirred for 10 h, followed by drying and grinding. Subsequently, the mixture was heated at 350\u00b0C and 650\u00b0C for 3\u00a0h under Ar atmosphere, respectively. Finally, the as-prepared yellow powder was heated in a selected temperature range of 700\u2013900\u00b0C to obtain the target production. From the SEM and TEM image, it can be observed that the average CNT diameter is around 31\u00a0nm without metallic nanoparticles (Figure\u00a04H and 4I). Meanwhile, the uniform distribution of N and Ni can be seen from STEM-EDS mapping (Figure\u00a04J). In AC-STEM, bright spots corresponding to Ni atoms were uniformly distributed in CNTs, and individual Ni atoms were located on the walls of a CNT (Figure\u00a04K and 4L). In addition to normal C6 carbon rings, C5, C7, and other nonC6 carbon rings were also formed in CNT, combining with the results of Raman spectroscopy (Figure\u00a04M). NiSA-N-CNTs with a load of up to 20 wt% showed the best selectivity and activity for the electrochemical reduction of CO2 to CO, with TOF values two orders of magnitude higher than those of Ni nanoparticles loaded on CNTs. Han and colleagues reported the single Cu atoms dispersed on graphene through a unique confined self-initiated dispersing protocol (Han et\u00a0al., 2019a). The GO/DICY was prepared by stirring dicyandiamide (DICY) graphene oxide dispersion and then freeze-drying. Then the mixture was added into a quartz boat that was tightly wrapped by a piece of Cu foil. The quartz boat was pyrolyzed at 600\u00b0C for 2\u00a0h and 800\u00b0C for 1\u00a0h under Ar atmosphere in a tube furnace. The production was treated with 0.5\u00a0M H2SO4 and then pyrolyzed at 300\u00b0C again. This novel in situ dispersion protocol produces highly reactive gaseous copper-containing intermediates that essentially circumvent the large-scale agglomeration of metal atoms in traditional processes and facilitate Cu dispersion on graphene (Figure\u00a04N). The catalyst with the loading of 5.4 wt% showed an outstanding performance for ORR due to the abundant and highly dispersive Cu single atoms. Zhao and colleagues prepared a series of M\u2212NC (M\u00a0= Mn, Fe, Co, Ni, Cu, Mo, Pt, etc.) SACs with metal loadings up to 12.1 wt% through a cascade anchoring strategy (Zhao et\u00a0al., 2019). Firstly, the metal ions are chelated by chelating agent and anchored onto oxygen-species rich porous carbon support. Then the complex bound carbon and melamine were put into a tube furnace and heated to 800\u00b0C for 2\u00a0h under Ar flow to obtain the M-NC. Furthermore, the scale-up synthesis can be achieved in parallel by the same synthesis route except for increasing the amounts of materials.In general, the precise and controllable preparation of SACs can be achieved by ALD, however, its expensive equipment and low yield limited the development of this technology. From a practical point of view, wet-chemistry synthetic methods for SACs are more desirable approaches because of its ease of operation and feasibility of large-scale manufacturing. From the methods introduced earlier, it can be seen that both impregnation method and pyrolysis method can be used to achieve the scale-up preparation of carbon-based-material-supported SACs. In particular, the impregnation method does not require complex and expensive equipment and displays the characteristics of low cost, simple operation, and easy synthesis, so it shows great potential in mass preparation.The research progress of SACs is closely related to the development of characterization technology. Advanced characterization techniques help to understand the coordination environments and electronic\u00a0structures of SACs, which directly affect the catalytic performance. Hitherto, the main applied characterization methods included high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), atomic force microscope (AFM), scanning tunneling microscopy (STM), X-ray absorption spectroscopy (XAS), and Raman measurements. Furthermore, in situ characterization has also been introduced as a powerful tool for studying the real active sites, structural changes during the catalysis process, and reaction mechanisms of SACs. Thus, in order to fully understand the structure, composition, and coordination environment of SACs, it is necessary to combine a variety of advanced characterization methods. In addition, density functional theory (DFT) simulation provides an unprecedented opportunity to discover catalytic reaction mechanisms, enabling the rational design of materials with personalized activity.Nowadays, transmission electron microscopy (TEM) is a powerful tool to obtain the fine structure of nanomaterials. From the Figure\u00a05A (Ren et\u00a0al., 2020), we can see that the received signals are mainly transmitted electron beam and scattered electron in the range of \u03b83. In the range of \u03b82, Bragg scattered electrons are signals. When the received signal is \u03b81, the HAADF image can be obtained. Especially, supported single atoms can be directly observed through aberration-corrected HAADF-STEM, and even the atomic structure information could be obtained. The phase contrast of atoms is directly proportional to their relative atomic number Z. Because the Z value of an isolated atom is different from that of the host atom, the relative bright dots of single metal atoms can be observed clearly. Therefore, theoretically speaking, the higher the Z value of the atom on the hosts, the better the image will be. For example, the as-prepared Co SAs/N-C structure shows rhombododecahedral shape (Yin et\u00a0al., 2016), as revealed by TEM and HAADF-STEM (Figures\u00a05B and 5C). The aberration-corrected HAADF-STEM (Figures\u00a05D and 5E) were carried out to elucidate the form of the Co atoms with sub-angstrom resolution. Due to the different Z contrast between Co and C, the isolated heavier Co SACs can be identified in the carbon support.The STM is a scanning probe microscopy tool that enables the observation and localization of individual atoms with a higher resolution than its atomic force microscope counterparts. The atomic-level sharp metal tip in STM scans the surface atomic structure based on the electron quantum tunneling effect of the tip-sample nano-gap. The effect causes the tunnel current to show an exponential relationship with the size of the gap, and an atomic-level sample surface topography characteristic image is obtained. It is commonly used in surface scientific research to examine model catalysts, such as single crystals with well-defined surface structures. What\u2019s more, the STM can directly observe whether the surface atoms of the material have periodic surface structure features, surface reconstruction, and structural defects. However, STM cannot detect deep information and observe insulators. Atomically dispersed platinum (Pt) was synthesized by photochemical reduction method (Wei et\u00a0al., 2017). The atomically dispersed Pt on ultrathin carbon films can be directly observed through STM (Figure\u00a06A). At atomic resolution, single Pt atoms appear as single-peak protrusions with a diameter of about 0.2\u00a0nm and a height of about 0.3\u00a0nm. Besides, Wu and colleagues synthesized high-density Cu(I)-N active sites in an N-doped graphene matrix via pyrolysis of copper phthalocyanine and dicyandiamide (Wu et\u00a0al., 2016). HAADF-STEM and the corresponding element mappings show the uniform distribution of Cu, N, and C (Figure\u00a06B). From atomic resolution STM image (Figure\u00a06C), the obvious bright spots can be observed, which indicates that the Cu atoms are atomically dispersed. STM simulations further revealed the atomic structure of this catalyst, in which atomically dispersed Cu-N2 centers are embedded in the graphene lattice (Figure\u00a06D).AFM can not only measure the surface morphology of the sample (close to the atomic resolution) but also detect the force between the atoms on the surface, the elasticity, plasticity, hardness, adhesion, friction, and so on. From topographic image, the height of the nanoparticle or the surface roughness of the sample can be seen clearly. Reduced graphene oxide (rGO) has stable anchor sites for metal single atoms, but the anchor sites are sparse, making it difficult to prepare high-load metal single-atom catalysts. Therefore, combining rGO with two-dimensional materials with abundant connecting atoms, such as carbon nitride, is an effective strategy to deal with this challenge. Therefore, metal single atoms (Pd, Pt, Ru, Au) were fabricated on porous carbon nitride/reduced graphene oxide (C3N4/rGO) foam (Fu et\u00a0al., 2020). Among these catalysts, Pd1/C3N4/rGO showed enhanced catalytic activity over its NPs counterpart for Suzuki-Miyaura reaction. From the Figures\u00a06E\u20136G, the C3N4/rGO layer with an onion-like microstructure with orderly organization can be observed. The thickness of GO sheet building blocks was measured by AFM (Figure\u00a06H). The sheet height of about 0.7\u00a0nm corresponds to the height of a single GO layer. Meanwhile, the AC-HAADF-STEM was applied to observe the samples (Figure\u00a06J). It can be seen from the element mapping diagram (Figure\u00a06I) that Pd element is evenly distributed in the matrix containing N and C. As shown in isolated single Pd atoms, sites can be seen clearly on the 2D C3N4/rGO sheet without aggregated Pd nanoparticles or clusters.XAS is an element-specific technique used to obtain the properties of absorbing atoms and their surroundings, resulting in a comprehensive understanding of the chemical state and structure of catalysts. It is the main technique used to characterize different coordination structures, which can be used to gain insight into the local atomic and electronic structure of single atoms. XAS includes X-ray absorption near edge structure (XANES) spectrum and extended X-ray absorption fine structure (EXAFS) spectrum. The energy of XANES spectrum ranges from the absorption edge to 30\u201350 eV above the absorption edge, and it is sensitive to the charge state and orbital occupancy of single metal atoms. The EXAFS spectrum represents the spectral region where the energy above absorption edge ranges from 30\u201350 eV to 1000 eV or more. Through Fourier transform (FT) analysis of EXAFS, the coordination number and distance between the central atom and adjacent atoms can be extracted. Wavelet transform can distinguish backscattered atoms and provide strong resolution in k and R space, which is the perfect complement to FT. Single Cu atoms coordinated with both S and N atoms in MOF-derived hierarchically porous carbon (S-Cu-ISA/SNC) was reported by atomic interface regulation (Shang et\u00a0al., 2020). To better analyze the chemical state and atomic structure of the sample, synchrotron-radiation-based soft XANES and XAFS was carried out. From the analysis of Cu L-edge spectrum (Figure\u00a07A), carbon K-edge spectrum (Figure\u00a07B), N K-edge spectrum (Figure\u00a07C), S L-edge, and K-edge spectra of S-Cu-ISA/SNC combining the XPS results, bonds between atoms can be obtained. Meanwhile, the interface structure at atomic scale, like the average oxidation state of Cu, can be obtained from Cu K-edge XANES spectra of S-Cu-ISA/SNC and the references (Cu foil, CuS, and CuPc) (Figure\u00a07D). The scattering of Cu-N and Cu-S was detected by FT peaks in FT-EXAFS spectra for S-Cu-ISA/SNC, and no Cu-Cu bond was found (Figures\u00a07E and 7G). Cu K-edge wavelet transform (WT)-EXAFS has also been applied to study the atomic configuration and the Cu-N and Cu-S contributions of S-Cu-ISA/SNC due to the strong resolution of k and R spaces (Figure\u00a07F). These results strongly prove the existence of Cu single atoms. Based on the above analysis, the first coordination number of the central copper atom is 4, including one metal-sulfur and three metal-nitrogen bonds, in which bond lengths corresponds to 2.32 and 1.98\u00a0\u00c5, respectively (Figure\u00a07H).Raman spectroscopy is a nondestructive analysis technique based on the interaction of light and chemical bonds in materials. It can provide detailed information about the chemical structure, phase and morphology, crystallinity, and molecular interaction of the samples. A Raman spectrum is usually composed of a certain number of Raman peaks. Each Raman peak represents the wavelength position and intensity of the corresponding Raman scattered light. Every peak corresponds to a specific molecular bond vibration, which includes not only a single chemical bond, such as C-C, C=C, N-O, and C-H, but also the vibration of a group composed of several chemical bonds, such as the benzene ring breath vibration, long polymer chain vibration, and lattice vibration. Raman measurements are used to further analyze the structural information of SACs. Raman spectra also show the D band and G band, which can be distinguished allotropes of carbon in carbon materials. The D band represents the disordered carbon atoms and sp2-hybridized carbon atoms (Li et\u00a0al., 2012, 2018; Pan et\u00a0al., 2013; Zhang et\u00a0al., 2017b), whereas G band is related to the tangential stretching mode of sp2 carbon atoms, indicating the existence of crystalline carbon in the carbon material (Deng et\u00a0al., 2017). Wei and colleagues prepared N-decorated carbon-encapsulated W2C/WP heterostructure as an efficient HER electrocatalyst in acid and alkaline solutions (Wei et\u00a0al., 2021b). The samples prepared with different precursors of (NH4)10H2(W2O7)6/NH4H2PO4, 1:0, 1:1, 1:2, 1:4, and 1:12, were labeled as W2C/W@NC, W2C/WP@NC-1, W2C/WP@NC-2, W2C/WP@NC-4, and WP@NC. The two main peaks located at 697 and 803\u00a0cm\u22121 in the Raman spectrum correspond to the stretching vibration of W-C (Figure\u00a08A). The ID/IG ratios value of W2C/WP@NC-2 is 0.96 (Figure\u00a08C), which is smaller than that of W2C/WP@NC-1 (Figure\u00a08B) and W2C/WP@NC-4 (Figure\u00a08D). The results showed that W2C/WP@NC-2 illustrated high conductivity and quick charge-transfer rate.\nEx situ techniques are used to establish the relationship between electrochemical performance and the properties of materials. However, in situ characterization can not only provide plenty of valuable information during the dynamic change process but also assess the coordination environment of the active site accurately. Nowadays, various in situ characterizations have gradually emerged with the continuous in-depth study of single atoms.Infrared spectroscopy can directly detect the interaction between adsorbed molecules and the supporter surface, and time- and temperature-resolved Fourier transform infrared spectroscopy (FTIR) can be used to detect catalytic intermediates. By detecting the vibration frequency and intensity of the model, the characteristics of the active center can be inferred after appropriate correction. Selection of appropriate probe molecules, such as CO, NH3, pyridine, and so on, can be used to analyze the overall catalyst, which is an important strategy to analyze the SACs. This paper mainly introduces the application of infrared spectroscopy in the characterization of SACs using CO as probe molecule. FTIR measurement was performed using CO as probe molecule to analyze the dispersion and oxidation state of Pt in the sample (Qiao et\u00a0al., 2011) (Figure\u00a09A). In sample B, there are three vibration bands in 2030\u00a0cm\u22121, 1950\u00a0cm\u22121, and 1860\u00a0cm\u22121. The main peak at 2030\u00a0cm\u22121 is the linear adsorption of CO at Pt0 site, whereas the weak vibration band at 1950\u00a0cm\u22121 and 1860\u00a0cm\u22121 is caused by the adsorption of CO on the bridge of two Pt atoms and the interface between Pt clusters and the support. That is, bridge-bonded CO indicates the presence of dimer or Pt clusters. These results indicate that Pt clusters and single atoms coexist in the samples. What\u2019s more, Hu and colleagues analyzed the existence state of Pt in Pt-SA/CsPbBr3 NCs (Hu et\u00a0al., 2021). The strong vibration peak at 2058\u00a0cm\u22121 indicates the linear adsorption of CO at the Pt\u03b4+ sites, which proves the existence of single Pt atoms (Figure\u00a09B). The absence of CO bridge adsorption peak indicates that Pt atoms may not have formed Pt nanoparticles or massive Pt atoms agglomeration.\nIn situ XAS can be used to analyze the evolution of the coordination environments during the catalytic process. Xiong et\u00a0al. reported isolated single-atom Rh anchored on N-doped carbon (SA-Rh/CN) for formic acid oxidation (Xiong et\u00a0al., 2020). The in situ XANES spectra of SA-Rh/CN were collected at Rh K-edge during chronoamperometry (CA) to investigate the change of oxidation state for Rh atom (Figure\u00a09C). The results showed that the intensity of the main absorption peak at \u223c 23250 eV gradually increased with the extending of reaction time, indicating that the oxidation state of Rh atoms became higher and higher, which may be caused by the formation of oxides in the process of high potential reaction. Similarly, the structural evolution and atomic interface structure of isolated Cu sites were collected by Cu K-edge XANES (Figures\u00a09E and 9F) and EXAFS during ORR (Shang et\u00a0al., 2020). In situ XAS was carried out in electrochemical cell set-up (Figure\u00a09D). The energy at the edge decreases gradually, along with the intensity of the white line from 1.05\u00a0V to 0.75 V. The in situ spectroscopic analysis shed light on the evolution of the electronic and atomic structures of the Cu-S1N3 moiety of S-Cu-ISA/SNC, revealing that the low-valence (+1) Cu-N-bond-shrinking HOO-Cu-S1N3, O-Cu-S1N3 and HO-Cu-S1N3 may contribute to ORR activity (Figure\u00a09G). However, in situ XANES is not yet prevalent because of the extremely limited resources of synchrotron radiation. With the construction of more advanced synchronous light sources, in situ XANES will play an increasingly important role in scientific research.\nIn situ Raman is a powerful analytical tool for revealing the reaction route and analyzing the reaction mechanism due to the high temporal and spatial resolution. Surface-enhanced Raman scattering (SERS) is caused by electromagnetic and the charge transfer mechanism, which means that when the analyte is adsorbed on rough metal surface, its Raman signal will be enhanced (Cialla et\u00a0al., 2012). Sun and colleagues employed the in situ SERS to monitor the adsorbate-substrate interaction in the process of ORR on the Au@Pd@Pt core/shell nanoparticles, which provided the direct evidence of \u2217OOH intermediate (Sun et\u00a0al., 2022). Furthermore, it is proved that the introduction of Pd shells affects the strain and electronic effect, leading to enhanced ORR activity. The relationship between ORR performance and strain/electron effect was illustrated by detecting intermediates from in situ SERS technique. What\u2019s more, time-resolved SERS (TR-SERS) was applied to reveal the dynamics of carbon dioxide (CO2) reduction reaction intermediates on Cu electrodes (An et\u00a0al., 2021). The results showed the surface reconstruction of Cu and the dynamic CO surface intermediates. This technique is of great significance for understanding the dynamic information of the surface reaction during CO2 electrolysis.However, SERS is limited to metal substrates with nanostructures. Shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS) is applicable to the surfaces of various materials and nanostructures (Li et\u00a0al., 2013). SHINERS are composed of a plasma gold or silver core and an inert shell such as SiO2 and Al2O3. These shell-isolated nanoparticles are easy to manufacture and can be dispersed on the surface of analytes with different composition and morphology, which are employed to enhance the Raman vibration signal of nearby molecules (Ding et\u00a0al., 2016). Wei and colleagues adopted in situ SHINERS to monitor the catalytic reaction process and kinetics of hydrogenation of nitro compounds and to characterize the structure of Pd single atoms for the first time (Wei et\u00a0al., 2021a). Pd SAs were anchored on the surfaces of TiO2 or Al2O3 shells of Au-shell-isolated nanoparticles. It also revealed the nucleation process of Pd species from single atoms to nanoparticles. This work provided a new spectroscopic tool for the in situ study of SACs, especially the solid-liquid interface.With the development of science and technology, researchers have developed numerous high-performance electrocatalysts, and the understanding about these reactions is gradually deepening. Density functional theory (DFT) calculations are widely used to study the free energies of intermediates and further reveal the mechanism of enhanced reactivity. The development of DFT theoretical models and advanced characterization techniques has greatly enhanced the understanding of electrocatalyst reaction mechanisms, such as the identification of active sites and the theory design of catalysts. Pt/C is considered to be the best catalyst for HER and ORR. RuO2 and IrO2 show the best catalytic performance toward OER. However, due to the scarcity and high price of precious metals, it is urgent to develop new type catalysts to reduce the production cost. SACs display great potential in the realization of efficient and selective electrocatalytic processes because of unique electronic structure and coordination environment. The surrounding coordination atoms of the central metal atom show important effects on their catalytic activity, selectivity, and stability, which are significant indicators of catalysts. In this section, the characterization of the existence of single atoms and coordination environments and their related catalytic performance will be discussed in detail.With the consumption of fossil fuels, a series of environmental problems have aroused people\u2019s attention. Hydrogen is considered to be the most likely alternative to fossil fuels due to its high energy density, no carbon emission, and without pollution. The method of electrolyzing water to produce hydrogen has attracted wide attention because of high efficiency, no need to consume fossil energy, and high product purity. Since Nicholson and Carlisle proposed the concept of water electrolysis in the 18th century, electrochemical water splitting has been developed for more than 200 years (Kreuter and Hofmann, 1998). Hydrogen evolution reaction (HER) is represented by the chemical formula as: 2H+\u00a0+ 2e\u2013 \u2192 H2, which is a multi-step electrochemical process that occurs on the electrode surface to generate gaseous hydrogen at the cathode (Zheng et\u00a0al., 2015). The reaction mechanism of HER is different in acidic and alkaline solutions, but both can be divided into two elementary reactions (Bockris and Potter, 1952; Sheng et\u00a0al., 2010). The first step is electrochemical hydrogen adsorption, which is called Volmer reaction. The second step is electrochemical desorption (Heyrovsky reaction) or chemical desorption (Tafel reaction). The mechanism under acidic conditions can be expressed as (Lasia, 2010)\n\n(Equation\u00a01)\n\n\n\n\n\nH\n+\n\n+\nM\n+\n\ne\n-\n\n\u21ccM\n-\n\nH\n\u2217\n\n\n\nVolmer\n\nreaction\n\n\n\n\n\n\n\n\n\n\n(Equation\u00a02)\n\n\n\n\nM\n\u2212\n\nH\n\u2217\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\u21cc\nM\n+\n\nH\n2\n\n\n\nHeyrovsky\n\nreaction\n\n\n\n\n\n\n\n\nor\n\n(Equation\u00a03)\n\n\n\n\n2\nM\n\u2212\n\nH\n\u2217\n\n\u21cc\n2\nM\n+\n\nH\n2\n\n\n\nTafel\n\nreaction\n\n\n\n\n\n\n\n\nThe mechanism under alkaline conditions can be expressed as\n\n(Equation\u00a04)\n\n\n\n\n\n\n\nH\n2\n\nO\n+\nM\n+\n\ne\n\u2212\n\n\u21cc\nM\n\u2212\n\nH\n\u2217\n\n+\n\nOH\n\u2212\n\n\n(\nVolmer\n\nreaction\n)\n\n\n\n\n\n\n\n\n\n\n\n(Equation\u00a05)\n\n\n\n\n\n\nM\n\u2212\n\nH\n\u2217\n\n+\n\nH\n2\n\nO\n+\n\ne\n\u2212\n\n\u21cc\nM\n+\n\nOH\n\u2212\n\n+\n\nH\n2\n\n\n(\nHeyrovsky\n\nreaction\n)\n\n\n\n\n\n\n\n\n\nor\n\n(Equation\u00a06)\n\n\n\n\n\n\n2\nM\n\u2212\n\nH\n\u2217\n\n\u21cc\n2\nM\n+\n\nH\n2\n\n\n(\nTafel\n\nreaction\n)\n\n\n\n\n\n\n\n\nwhere H\u2217 designates a hydrogen atom chemically adsorbed on an active site of the electrode surface (M).As explained in Conway and Tilak in detail, the Tafel slope value obtained by the HER polarization curve can be used to judge the possible rate-determining step during the reaction process (Conway and Tilak, 2002). From these elementary reactions, we can see that the chemical adsorption and desorption of H atoms is a complicated process. According to the different pathways of H atom desorption, the HER reaction mechanism can be divided into Volmer-Heyrovsky mechanism and Volmer-Tafel mechanism (Li et\u00a0al., 2016).The bonding ability of the active sites in the catalyst and H\u2217 should neither be too strong nor too weak according to the Sabatier principle (Sabatier, 1920). When the bonding ability of M-H\u2217 is too strong, it is not conducive to the breaking of the bond and the release of hydrogen. On the contrary, it is not beneficial to the process of proton-electron transfer. However, the bond energy between the active site of the catalyst and H\u2217 cannot be measured directly, so it is difficult to establish a direct correlation between the intermediate H\u2217 and the electrochemical reaction rate (Markovi\u0107 and Ross, 2002). From the perspective of physics and chemistry, the adsorption free energy of H\u2217 (\u0394GH\u2217) can be used to evaluate the capacity of H adsorption and the release of H2 (N\u00f8rskov et\u00a0al., 2005). Therefore, \u0394GH\u2217 is a key parameter used to evaluate the reaction rate. \u0394GH\u2217 close to zero could be used to evaluate the efficiency of the catalyst, but not a requirement. More importantly, it is concluded that the experimental exchange current density j0 value and \u0394GH\u2217 have a \u201cvolcano curve\u201d correlation through DFT calculation (Sk\u00falason et\u00a0al., 2010).As can be seen from the volcano diagram (Zeradjanin et\u00a0al., 2016), Pt is located at the top of the volcano diagram, indicating appropriate adsorption energy and the highest current density, which explains the optimal performance of Pt in HER (Figure\u00a010A). However, due to the scarcity of precious metals, the application of SACs in HER has attracted wide attention from researchers in order to improve atomic utilization rate and reduce catalyst cost. Meanwhile, the development of high-active non-noble-metal catalysts is also an effective solution.The introduction of heteroatoms into carbon substrate can not only form new coordination environments but also improve metal loading, which has attracted extensive attention of researchers. The general coordination number between metal atom and N atom is 4, which may be related to the valence state and electronic structure of the central atom. For example, Ye and colleagues reported single Pt atoms anchored on aniline-stacked graphene with a Pt loading of 0.44 wt% through microwave reduction method (Ye et\u00a0al., 2019). The Pt SASs/AG with unique Pt-N4 coordination not only displays high HER activity with the overpotential of 12\u00a0mV at 10 mA cm\u22122 in 0.5\u00a0M H2SO4, better than 20 wt% Pt/C, but also the mass current density of 22,400 AgPt\n\u22121 at the overpotential of 50\u00a0mV is 46 times higher than commercial 20 wt% Pt/C. Moreover, stability is also a significant issue in the development SACs. Cyclic stability and long-term stability can be assessed by cyclic voltammetry sweeps and chronopotentiometric measurements. The Pt SASs/AG shows negligible decay by comparing the LSV curves before and after 2000 cycles and displays outstanding long-term stability over 20 h. From the XPS spectrum of Pt 4f (Figure\u00a010B), it shows the strong interaction of Pt and aniline-formed Pt\u03b4+ XAFS, and DFT calculations show that the isolated Pt is coordinated with the N of four aniline molecules (Figures\u00a010C\u201310E), which optimizes the electronic structure of Pt. The modulation of the d-band center and density of states (DOS) near the Fermi level of Pt atoms by aniline caused the single Pt sites to have appropriate hydrogen adsorption energy and finally enhances HER activity. In addition, other noble-metal-based SACs have also been studied for HER. The unsaturated coordination between the central Ru atom and the surrounding N atom can also significantly enhance HER performance. For example, C3N4-Ru were fabricated by thermal treatment of graphitic C3N4 nanosheets and RuCl3 in water, which shows apparent HER activity in acidic media, and the HER activity is positively correlated with loading of Ru (Peng et\u00a0al., 2017). The test results showed charge transfer from C3N4, a unique functional scaffold, to the Ru center (Figure\u00a010G). The formation of unsaturated coordination Ru-N2 moieties as effective active sites facilitated the adsorption of hydrogen from the DFT calculation (Figure\u00a010F). Besides, single Ruthenium atoms were anchored amorphous phosphorus nitride nanotubes (Ru SAs@PN) through strong coordination interactions between the d orbitals of Ru and the lone pair electrons of N located in the HPN matrix (Yang et\u00a0al., 2018b). The SACs in Ru-N3.8 coordination environment were prepared by impregnation method. The Ru SAs@PN showed excellent HER activity with overpotential of 24\u00a0mV at 10 mA cm\u22122 and robust long-term stability over 24\u00a0h in 0.5\u00a0M H2SO4. DFT calculations showed that the Gibbs free energy of adsorbed H\u2217 over the Ru SAs on PN is much closer to zero compared with the Ru/C and Ru SAs supported on carbon and C3N4 (Figure\u00a010H). Therefore, adjusting the number of coordination atoms of metal centers is considered to be a very effective way to optimize HER performance.In recent years, non-noble-metal-based SACs have also attracted extensive attention. However, compared with Pt based catalysts, their performance still needs to be further improved. Non-noble-metal-based SACs with low coordination environment and unique electronic structure have the potential to replace Pt-based catalysts (Chen et\u00a0al., 2017a; Zhang et\u00a0al., 2018). Graphdiyne (GD) is a two-dimensional carbon material with monatomic thickness, which has natural uniform pores, rich triple bonds, and strong reduction ability (Li et\u00a0al., 2010, 2014). It has been used in various research fields, so GD may also be an excellent candidate as support. Isolated nickel/iron atoms anchored on graphdiyne were fabricated by electrochemical synthesis method with Ni-C and Fe-C coordination, respectively (Xue et\u00a0al., 2018) (Figure\u00a011A). From the results of ICP-MS, the loading of Ni in Ni/GD is 0.278 wt% and the loading of Fe in Fe/GD is 0.680 wt%. Fe/GD with higher metal laoding exhibits the overpotential of 66\u00a0mV at 10 mA cm\u22122, which is smaller than Ni/GD (88\u00a0mV). And their performances are superior to the most state-of-the-art bulk nonprecious catalysts. Meanwhile, the Fe/GD displays more superior durability through 5000 cycling tests. The strong chemical interaction and electronic coupling between single atoms Ni/Fe and GD allow a high charge transfer between the catalytic active center and the support, so the performance of HER is improved.In addition to coordinating with the same type of surrounding atom, the central metal atom can also coordinate with different atoms simultaneously. Mo-based catalysts such as carbide, nitride, and sulfide have attracted a lot of attention because of good performance in HER. Chen and colleagues designed single Mo atoms supported on N-doped carbon for the first time (Chen et\u00a0al., 2017a), which shows high HER performance with the overpotential of 132\u00a0mV at 10 mA cm\u22122 and onset overpotential of 13\u00a0mV in 0.1\u00a0M KOH. From AC-STEM and XAFS, it can be seen that Mo1N1C2 was formed by single Mo atom immobilized with one nitrogen atom and two carbon atoms (Figures\u00a011B\u201311F). More importantly, the active sites Mo1N1C2 showed higher catalytic activity than Mo2C and MoN due to the lowest absolute value of \u0394GH\u2217 in Mo1N1C2 compared with the Mo2C and MoN from the DFT calculation results. Further DOS calculations revealed that the DOS of Mo1N1C2 near the Fermi level was much higher than that of Mo2C and MoN, which was favorable for the charge transfer during the HER process because of higher carrier density. Moreover, tungsten-based catalysts, including WCx (Gong et\u00a0al., 2016; Wang et\u00a0al., 2020), WNx (Yan et\u00a0al., 2015b), WPx (Wang et\u00a0al., 2016; Xing et\u00a0al., 2015), WSx (Lin et\u00a0al., 2014; Lukowski et\u00a0al., 2014; Merki and Hu, 2011; Voiry et\u00a0al., 2013), and so on, also have outstanding properties in HER. However, in order to achieve industrial application, it is necessary to further improve HER activity and stability of the tungsten-based materials. W-SAC with W-N1C3 sites supported on MOF-derived N-doped carbon was prepared for HER in both alkaline and acidic media (Chen et\u00a0al., 2018a). The W-SAC showed excellent stability without attenuation after 10,000 CV cycles. It is determined by HAADF-STEM and XAFS that the atomically dispersed W1N1C3 act as the active site, which plays a significant role in enhancing the HER performance as proved by DFT calculation results. The DOS of W-SAC near the Fermi level is much higher than of WC and WN, leading to a larger carried density for promoting charge transfer in HER (Figures\u00a011G and 11H). Furthermore, the DOS near the Fermi level in W-SAC was mainly contributed by the W d-orbital, whereas the contributions of C and N p-orbital were negligible; this suggested that the single W dispersion as well as unique electronic structure could efficiently enhance the d-electron domination near the Fermi level and enhance the HER catalytic performance. The coordinating atom N is partially replaced by other atoms, which shows a great influence on the local chemical environment of the central atom. Thus, it can be used to optimize the electronic structure of the central atom and is an important method to enhance the activity of SACs by modulating the N coordination to form dual-atom coordination.Oxygen evolution reaction (OER) is another important half-reaction that occurs at the anode during the water electrolysis process, involving four coordinated proton-electron transfer steps. However, there is still a long way to go in the mechanism understanding and material design of OER catalysts. As shown in Figure\u00a012A (Chen et\u00a0al., 2021), the generally accepted OER reaction mechanisms are the traditional adsorbate evolution mechanism (AEM) and lattice oxygen mechanism (LOM).In AEM, the reaction mechanism of OER is similar to HER, involving the steps of adsorption and desorption. The mechanism can be expressed as follows (Man et\u00a0al., 2011; Rossmeisl et\u00a0al., 2007):\n\n(Equation\u00a07)\n\n\n\n\n\n\n\u2217\n\n+\n\nH\n2\n\nO\n\n1\n\n\u2192\n\nOH\n\u2217\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n\n\n\n\n\n\n\n(Equation\u00a08)\n\n\n\nOH\n\u2217\n\n\u2192\n\nO\n\u2217\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n\n\n\n\n\n(Equation\u00a09)\n\n\n\n\n\n\nO\n\u2217\n\n\n\n\n+\n\nH\n2\n\nO\n\n(\n1\n)\n\n\u2192\n\nOOH\n\u2217\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n\n\n\n\n\n(Equation\u00a010)\n\n\n\n\n\n\nOOH\n\u2217\n\n\n\n\n\u2192\n\nO\n2\n\n\n(\ng\n)\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n\n\n\nSimilarly, OER also has a volcanic curve based on oxygen adsorption energy. Noble metal oxides, such as IrO2 and RuO2, showed the best OER performance due to low overpotential. Based on this mechanism, the OER performance can be effectively improved by modifying the catalyst supports, such as introducing heteroatoms through synergistic effects to optimize the electronic structure, creating defects to redistribute surface charges and so on.LOM has attracted much attention in recent years. The lattice oxygen participates in the proton-electron transfer in the reaction process (Huang et\u00a0al., 2019). The mechanism can be expressed as follows (Rong et\u00a0al., 2016):\n\n(Equation\u00a011)\n\n\n\n\n\n\nOH\n\u2217\n\n\n\n\n\u2192\n\n\n(\n\n\nV\nO\n\n+\n\nOO\n\u2217\n\n\n)\n\n\u2020\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n\n\n\n\n\n(Equation\u00a012)\n\n\n\n\n\nV\nO\n\n+\n\nOO\n\u2217\n\n\n\n\u2020\n\n+\n\nH\n2\n\nO\n\n1\n\n\n\u2192O\n2\n\n\ng\n\n+\n\n\n\n\nV\nO\n\n\n+OH\n\u2217\n\n\n\n\u2020\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\n\n\n\n\n(Equation\u00a013)\n\n\n\n\n\nV\nO\n\n+\n\nOH\n\u2217\n\n\n\n\u2020\n\n+\n\nH\n2\n\nO\n\n1\n\n\u2192\n\n\n\n\nH\n\nO\n-\nsite\n\n\u2217\n\n+\n\nOH\n\u2217\n\n\n\n\u2020\n\n+\n\nH\n+\n\n+\n\ne\n-\n\n\n\n\n\n\n(Equation\u00a014)\n\n\n\n\n(\n\n\nH\n\no\n-\nsite\n\n\u2217\n\n\n\n+\nOH\n\n\u2217\n\n\n)\n\n\u2020\n\n\n\u2192OH\n\u2217\n\n\n\n+\nH\n\n+\n\n\n\n+\ne\n\n\u2212\n\n\n\n\n\u2020 Parentheses indicate that adsorbates are calculated in the same supercell.In early studies, the involvement of lattice oxygen in OER process under acidic conditions was confirmed by isotope labeling and differential electrochemical mass spectrometry (DEMS) techniques (K\u00f6tz et\u00a0al., 1984; Willsau et\u00a0al., 1985; Wohlfahrt-Mehrens and Heitbaum, 1987). The O in RuO2 participates in the reaction to form soluble RuO4, which is detected in the solution. Pan and colleagues reported a model system of Si-incorporated strontium cobaltite perovskite electrocatalysts in alkaline solution with similar surface transition metal properties but different oxygen diffusion rates (Pan et\u00a0al., 2020). The correlation of intrinsic OER activity with oxygen ion diffusion rate and oxygen vacancy diffusion rate are shown in Figures\u00a012B and 12C. The evolution of oxygen correlates with the contribution of the LOM mechanism at different degrees that closely relates to the oxygen ion diffusivity (Figure\u00a012D). This work provides a reference for designing more stable perovskite surfaces to further optimize electrocatalysts. A series of perovskite OER catalysts were also tested through in situ\n18O isotope labeling mass spectrometry (Grimaud et\u00a0al., 2017). The results showed that the O2 generated from the lattice oxygen for some highly active oxides. In combination with experiments and DFT calculations, catalysts with lattice oxygen exchange exhibited pH-dependent OER activity, whereas those without lattice oxygen exchange displayed pH-independent OER activity. LOM shows higher OER activity than the conventional AEM as proved for the ABO3 (A\u00a0= lanthanum or strontium, B\u00a0= transition metal) perovskites (Yoo et\u00a0al., 2018). Activity volcano plots for AEM and LOM of perovskite systems have been established by a simulation work. Furthermore, the LOM is preferred for achieving bifunctional catalysts for OER and ORR.In the HER and OER of water splitting, OER is the core of electrochemical energy conversion. However, OER displays high overpotential during the reaction because of sluggish kinetics, which is the main step of energy consumption. Therefore, high-efficiency electrocatalysts are particularly important to OER (Chen et\u00a0al., 2019; Liu et\u00a0al., 2021; Zhou et\u00a0al., 2019). RuO2 and IrO2 exhibit high catalytic activity for OER in a wide pH value and are often used as benchmarks for OER catalyst evaluation. It was found that the coordination environments of single atom may undergo changes during the process of reaction. Ru single atoms anchored on nitrogen-carbon support (Ru-N-C) were synthesized with Ru1-N4 sites (Cao et\u00a0al., 2019). The catalyst showed an efficient and durable electrocatalyst for acidic OER with overpotential of 267\u00a0mV at the current density of 10 mA cm\u22122, mass activity of 3571 A gmetal\n\u22121, and TOF of 3348 O2 h\u22121. The dissolution rate of Ru is less than 5% in acid solution due to the outstanding structural stability. The Ru-N-C was employed to measure the overall water splitting in a two-electrode system to mimic the PEMWE, showing superior activity and stability. The dynamic pre-adsorption of single oxygen atom into the formation of O-Ru1-N4 structure with more charge donations of Ru through in situ XAFS and FTIR was also identified (Figures\u00a013A\u201313F). Theoretical calculations showed O-Ru1-N4 with higher Ru oxidation state as the real active site for the high OER activity in acidic solution. The formed O-Ru1-N4 moieties under operando state exhibited a low barrier of O-O coupling to form the OOH\u2217 intermediate. It is also an effective way to prepare SACs by creating defects on the support to regulate the coordination environment. Atomically dispersed Ni catalyst on defective graphene (a-Ni@DG) with four-coordination Ni-C4 structure was fabricated through an incipient wetness impregnation method and subsequent acid leaching (Zhang et\u00a0al., 2018). The Ni loading of a-Ni@DG was around 1.24 wt% by this facile and inexpensive strategy. XAS and DFT calculation revealed that the diverse defects in graphene can induce different local electronic DOS of a-Ni, which suggested that aNi@defect serves as an active site for unique electrocatalytic reactions (Figures\u00a013J\u201313L). For example, aNi@G5775 and aNi@G585 are responsible for HER and OER with low overpotential and high TOF values and stability, respectively. HAADF-STEM not only confirmed the uniform distribution of single Ni atoms but also observed that aNi trapped in the Di-vacancy provided direct evidence for the Ni-C4 configuration (Figures\u00a013G\u201313I). Diverse defects can induce different local electronic density of states. Creating specific defects on the support forming various active sites can achieve good catalytic effects for different reactions at the same time.What\u2019s more, bimetallic center SACs have gradually attracted extensive attention from researchers. The synergistic effect of bimetallic centers can optimize the adsorption and desorption of intermediates and reduce the reaction energy barrier. For example, atomically dispersed binary Co-Ni sites embedded in N-doped hollow carbon nanocubes (CoNi-SAs/NC) are synthesized for bifunctional OER and ORR (Han et\u00a0al., 2019b) (Figure\u00a013M). The rechargeable process of Zn-air batteries is realized efficiently and with low potential and robust reversibility. DFT calculation showed that the uniformly dispersed single sites and synergistic effect of adjacent Co-Ni bimetallic centers optimized the adsorption and desorption process, reduced the overall reaction energy barriers, and finally promoted the reversible oxygen electrocatalysis.In the reported literature, the activity and stability of HER and OER were measured in standard three-electrode system of lab scale. Nonetheless, the catalyst\u2019s performance in the lab scale is somewhat different from that in practical electrolyzers. The practical electrolyzers usually require higher current density and voltage, and the operating environment is more intense. Therefore, the practical electrolyzers put forward higher requirements for the activity and stability of the catalysts.Currently, the electrolyzers used for water electrolysis include alkaline water electrolyzers (AWE), proton exchange membrane water electrolyzers (PEMWE), and anion exchange membrane water electrolyzers (AEMWE). Although traditional AWE has been fully industrialized, it is limited by its environmental friendliness and purity of hydrogen production. PEMWE and AEMWE have received extensive attention due to the high efficiency, high purity of hydrogen produced, and low energy consumption. Influenced by the properties of the membrane and the local pH, PEMWE and AEMWE are suitable for catalysts in acidic and alkaline environments, respectively. The current development of PEMWE is relatively mature in commercial-scale water electrolysis for long-term operation, but the progress of durable catalysts other than Ir-based noble metal OER catalysts in acidic environments remains a great challenge. As more OER catalysts display better performance in alkaline environment, AEMWE shows certain advantages, but the stability of membrane and the design of electrolyzer still need to be further improved to meet the requirements of long-term electrolysis.The membrane electrode assemblies (MEA), composed of the catalytic layer and the proton exchange membrane, is the main site for material transport and electrochemical reaction in the entire electrolytic cell. The characteristics and structure of the MEA directly affect the performance and life of the PEMWE. Hao and colleagues applied the prepared grain boundaries (GB)-TaxTmyIr1-x-yO2-\u03b4 nanocatalysts to PEMWE as anode in an acidic condition (Hao et\u00a0al., 2021). The pretreated carbon paper- (CP) and platinum-plated titanium foam were applied as cathode and anode gas diffusion layers (GDLs), respectively. The MEA were constructed by placing the catalyst-supported Nafion 117 membrane between CP and Pt-plated Ti foam GDLs (Figure\u00a013N). The polarization curves were measured at 50\u00b0C, showing the cell voltage of GB-Ta0.1Tm0.1Ir0.8O2-\u03b4 is 1.766V to reach the current density of 1 A cm\u22122 (Figure\u00a013O). Moreover, the PEMWE using GB-Ta0.1Tm0.1Ir0.8O2-\u03b4 displayed the outstanding stability at 1.5 A cm\u22122 for at least 500\u00a0h without obvious attenuation (Figure\u00a013P). However, the research of SACs on PEMWE still needs to be further explored.Oxygen reduction reaction (ORR) occurs at the cathode of electrochemical energy equipment through either a 4-electron pathway (O2+4H++4e-\u21922H2O) or a 2-electron pathway (O2+2H++2e-\u2192H2O2). The four-electron reaction mechanisms are shown as follows (N\u00f8rskov et\u00a0al., 2004):\n\n(15)\n\n\n1\n2\n\n\nO\n2\n\n\n+\n\u2217\n\n\u2192\n\nO\n\u2217\n\n\n\u2192\n\n\ne\n-\n\n+\n\nH\n+\n\n\n\n\nHO\n\u2217\n\n\n\u2192\n\n\ne\n\u2212\n\n+\n\nH\n+\n\n\n\n\nH\n2\n\nO\n\n+\n\u2217\n\n\n\n\nor\n\n(16)\n\n\nO\n2\n\n\n+\n\u2217\n\n\n\u2192\n\n\ne\n-\n\n+\n\nH\n+\n\n\n\n\nHOO\n\u2217\n\n\n\u2192\n\n\ne\n-\n\n+\n\nH\n+\n\n\n\n\nH\n2\n\nO\n+\n\nO\n\u2217\n\n\n\u2192\n\n\ne\n-\n\n+\n\nH\n+\n\n\n\n\nHO\n\u2217\n\n\n\u2192\n\n\ne\n-\n\n+\n\nH\n+\n\n\n\n\nH\n2\n\nO\n\n+\n\u2217\n\n\n\n\nThe two-electron reaction mechanisms are shown as follows:\n\n(17)\n\n\nO\n2\n\n\n+\n\u2217\n\n\n\u2192\n\n\ne\n-\n\n+\n\nH\n+\n\n\n\n\nHOO\n\u2217\n\n\n\u2192\n\n\ne\n-\n\n+\n\nH\n+\n\n\n\n\nH\n2\n\n\nO\n2\n\n+\n\nO\n\u2217\n\n\n\nwhere \u2217 represents the catalytically active sites.The Equations (7) and (8) represent dissociative and associative mechanisms, respectively. Whether the O-O bond breaks during ORR process determines the selectivity of H2O2 or H2O. The 2-electronic process is clean and pollution free, and the hydrogen peroxide produced is an important fine chemical. The 4-electronic process produces water directly, which is mainly used in fuel cells and metal-air batteries (Liu et\u00a0al., 2018; Tong et\u00a0al., 2021; Zhou et\u00a0al., 2021b). However, these two reaction pathways often occur at the same time, resulting in reduced selectivity of the desired product. Thus, the application of SACs in ORR has been extensively studied to improve selectivity, stability, and activity.The precious-metal-group (PMG) catalysts exhibit high-efficiency electrocatalytic performance for ORR. Doping heteroatoms into carbon materials can affect the coordination environment of the metal center atoms, which in turn shows an impact on the catalytic performance. M-N-C catalysts with M-N4 coordination structure are considered as potential catalysts for replacing Pt-based materials in ORR. Single Pt atoms supported on carbon black were fabricated with carbon monoxide/methanol tolerance for ORR (Liu et\u00a0al., 2017). DFT calculations were used to study the synergetic effect between single Pt atoms and doped-N and the intrinsic activity of the active sites on Pt1-N/BP (carbon black) (Figures\u00a014A and 14B). The results showed that the main effective sites are single-pyridinic-nitrogen-atom-anchored single-Pt-atom centers, displaying highly active and making it one of the most promising sustainable, large-scale alternatives to conventional Pt-NP-based electrocatalysts. The catalyst was employed as a cathode in acid single cell, and the power density reached 680 mW cm\u22122 at 80\u00b0C. The current of Pt1-N/BP still maintains more than 70% when used as cathode in fuel cell after working for 200\u00a0h continuously. It shows good stability compared with other non-noble-metal catalysts. Other precious metal-based SACs have also been used in ORR studies. Ir-N-C SAC with Ir-N4 configuration fabricated by host strategy exhibited orders of magnitude higher ORR activity than Ir NPs (Xiao et\u00a0al., 2019) (Figure\u00a014C). The Ir-SAC was applied in a H2/O2 fuel cell as cathode, showing a higher open circuit voltage of 0.0955V and a power density of 932 mW cm\u22122. The SEM and HAADF-STEM displayed the morphology and nanostructure of Ir-SAC, and the EDS mapping showed the existence of Ir embedment within the carbon matrix (Figures\u00a014D\u201314H). The bright spots in Figure\u00a014I corresponded to the atomically dispersed Ir atoms. Atomic structure characterization results and DFT calculations showed that the high activity of Ir-SAC was attributed to the moderate adsorption energy of the Ir-N4 moiety. Single Ru atoms supported on N-doped graphene oxide for ORR in acidic solution were prepared through NH3 atmosphere annealing (Zhang et\u00a0al., 2017a). The Ru/N-doped graphene showed excellent four-electron ORR activity, stability, and tolerance. Combing the DFT calculations, the Ru-oxo-N4 moieties during the oxidative electrocatalytic condition are responsible for the ORR catalytic activity (Figures\u00a014J and 14K).Recently, transition-metal-based catalysts have emerged as promising alternatives to PMG materials due to the adjustable electronic structure. Fe-N4 is considered to be the best performing non-noble-metal catalyst in ORR. Chen and colleagues performed adjustment of O coordination on Fe SACs to enhance ORR performance. Single iron atoms anchored on N-porous carbon with Fe loading of 2.16 wt% were fabricated through a cage-encapsulated-precursor pyrolysis strategy (Chen et\u00a0al., 2017b). It can be seen from XANES that Fe is positively charged (Figure\u00a015A). According to the analysis results of FT-and WT-EXAFS, Fe exists in the form of a single atom in the catalyst (Figures\u00a015B\u201315E). Meanwhile, the corresponding EXAFS R space fitting curves of Fe-ISAs/CN is shown as Figure\u00a015F. The coordination number of Fe is 5 with four N atoms and one O atom. So, the atomic structure model is shown in Figure\u00a015G through further analysis. The Fe as isolated atoms with N coordination showed higher ORR activity than commercial Pt/C and most non-precious-metal materials, which was attributed to the high capability of the single Fe atoms in transferring electrons to the adsorbed OH species demonstrated by first principle calculations. Meanwhile, Fe-ISAs/CN showed superb durability with little change in E\n1/2 for 5000 CV sweeps.What\u2019s more, changing the metal central atom is a direct means to regulate the coordination environment, which also affects the catalytic performance. Single Co atom and N co-doped carbon nanofibers with CoN4-G coordination were reported for ORR in both acidic and basis medium (Cheng et\u00a0al., 2017). TEM and HRTEM showed that the diameter of the prepared catalyst was about 150\u00a0nm, and the substrate was amorphous carbon (Figures\u00a015H and 15I). No obvious bright regions were observed in HAADF-STEM (Figure\u00a015J), further indicating the absence of cobalt-containing particles. EDX mapping results showed that Co, N, and C were evenly distributed in CNFs (Figure\u00a015K), and the Co corresponding to bright spots existed as a single atom in AC-HAADF-STEM (Figures\u00a015L and 15M). The Co-N/CNFs displays desirable ORR performance and high stability with negligible decrease of E\n1/2 after 10,000 CV cycles. Moreover, the catalyst as cathode reached a power density of 16 mW cm\u22122 and an outstanding stability with more\u00a0than 200 h, showing the potential of application. What's more, a series of M-N-C materials (M\u00a0= Mn, Fe, Co, Ni, and Cu) with atomically dispersed M-Nx sites were investigated the trends in electrochemical H2O2 production from molecular first principles to bench-scale electrolyzers operating at industrial current density (Sun et\u00a0al., 2019c). Co-N-C catalyst showed outstanding ORR activity and selectivity to H2O2 and more than 4\u00a0mol peroxide gcatalyst\n\u22121 h\u22121 at a current density of 50 mA cm\u22122. The relationship of activity-selectivity and the trend of M-N-C materials was further analyzed by DFT calculations, providing a molecular scale understanding of the experimental volcanic trend of four-electron and two-electron ORR (Figure\u00a015N). Meanwhile, the binding free energy of HO\u2217 intermediate placed Co-N-C close to the top of the two-electron volcano, retaining catalytic activity while promoting two-electron pathway selectivity.The increasing global environmental crisis has aroused people\u2019s attention to greenhouse gas emissions, conversion, and storage (Anagnostou et\u00a0al., 2016; Mun et\u00a0al., 2018; Obama, 2017). CO2 is considered to be the main cause of the greenhouse effect. Electroreduction of CO2 into high value-added products, such as CO, HCOOH, CH4, CH3OH, C2H4, and so on, is a promising route, which can mitigate environmental problems. Because water acts as the medium of CO2RR, HER inevitably becomes the side reaction of this reaction (Li et\u00a0al., 2017; Yang et\u00a0al., 2018a). Therefore, efficient catalysts in CO2RR should reduce HER activity and enhance CO2RR activity at the same time (Lin et\u00a0al., 2019; Yan et\u00a0al., 2018a).Currently, the key obstacle to the development of efficient CO2RR catalysts is the lack of a basic understanding of surface-mediated electrochemical reactions. There are many possible products in CO2RR, involving electron transfer numbers ranging from CO (2e\u2212) and HCOOH (2e\u2212) to CH3CH2CH2OH (18e\u2212), so the interpretation of the reaction mechanism is more demanding (Lu and Jiao, 2016). Some typical multi-electron reactions in neutral medium are shown as follows:\n\n(Equation\u00a018)\n\n\n\nCO\n2\n\n\n(\ng\n)\n\n\n\n+\n2\nH\n\n+\n\n\n\n+\n2\ne\n\n-\n\n\u2192CO\n\n(\ng\n)\n\n\n\n+\nH\n\n2\n\nO\n\n\n\n\n\n\n(Equation\u00a019)\n\n\n\nCO\n2\n\n\n(\ng\n)\n\n\n\n+\n2\nH\n\n+\n\n\n\n+\n2\ne\n\n-\n\n\u2192HCOOH\n\n(\n1\n)\n\n\n\n\n\n\n\n(Equation\u00a020)\n\n\n\nCO\n2\n\n\n(\ng\n)\n\n\n\n+\n4\nH\n\n+\n\n\n\n+\n2\ne\n\n-\n\n\u2192HCOOH\n\n(\n1\n)\n\n\n\n+\nH\n\n2\n\nO\n\n\n\n\n\n\n(Equation\u00a021)\n\n\n\n\n\n\n\nCO\n2\n\n\n(\ng\n)\n\n\n\n+\n6\nH\n\n+\n\n\n\n+\n6\ne\n\n-\n\n\n\u2192CH\n3\n\nOH\n\n(\n1\n)\n\n\n\n+\nH\n\n2\n\nO\n\n\n\n\n\n\n\n\n\n\n(Equation\u00a022)\n\n\n\nCO\n2\n\n\n(\ng\n)\n\n\n\n+\n8\nH\n\n+\n\n\n\n+\n8\ne\n\n-\n\n\n\u2192CH\n4\n\n\n(\ng\n)\n\n\n\n+\n2\nH\n\n2\n\nO\n\n\n\n\nSACs show high activity in many reactions with the highest atomic utilization rate, especially the unsaturated coordination between the central metal atom and the surrounding atoms, which significantly enhances the catalytic performance. Meanwhile, the uniform active site and geometry structure enhance the interaction between the mental centers and the support, which helps to improve the selectivity of the catalyst (Chen et\u00a0al., 2018c). Therefore, SACs display great application potential in CO2RR.Reducing the coordination number between metal center and N leads to form unsaturated coordination, which is helpful to optimize the catalytic performance. Zhao and colleagues prepared Ni atoms anchoring on N-doped porous carbon with Ni-N3 coordination by ZIF-assisted strategy for the first time in CO2RR (Zhao et\u00a0al., 2017) (Figure\u00a016A). The SAC displayed outperforming current density of 10.48 mA cm\u22122 at an overpotential of 0.89\u00a0V with a high turnover frequency (TOF) of 5273 h\u22121 and Faradaic efficiency (FE) for CO production of over 71.9%. Besides, Zheng and colleagues fabricated an unsaturated coordination copper with nitrogen sites anchored into graphene matrix (Cu-N2/GN) for CO2RR to CO production (Zheng et\u00a0al., 2019). The catalyst showed higher activity and selectivity with the maximum FE of 81% at a low potential of \u22120.50\u00a0V and an onset potential of \u22120.33\u00a0V than the atomically dispersed Cu-N anchored on carbon materials reported previously. From a practical point of view, the Cu-N2/GN was applied in rechargeable Zn-CO2 battery with a peak power density of 0.6 mW cm\u22122, and the battery charging process can be powered by natural solar energy. Theoretical calculations showed that the moderate free energy of Cu-N2 sites promote the adsorption of CO2 molecules at the Cu-N2 site (Figures\u00a016B\u201316E). The adsorption state of H2O, CO2, COOH, and CO with Cu-N2-based DFT electron density was hybridized with surface states of Cu-N2 (Figures\u00a016F\u201316I). The short bond length of Cu-N2 sites caused the accelerated charge transfer from Cu-N2 site to \u2217CO2, which enhanced effectively the formation of \u2217COOH and CO2RR performance. Adjustment of N coordination was also applicable for controlling the coordination environment of CO2RR SACs. Bifunctional catalysis of Co and N co-doped hollow carbon for CO2RR and HER has been reported (Song et\u00a0al., 2018). The loading of Co single-atoms was around 3.4 wt%, and Co-C2N2 moieties acted as the major active sites during the process of CO2 reduction. The catalyst was prepared through high temperature pyrolysis (900\u00b0C) to remain the high content of Co single atoms and prevent the loss of nitrogen. The Co-HNC possessed better catalysis performance than Co NP-SNC in 0.1\u00a0M KHCO3 (Figure\u00a016J). The Co-HNC showed a nearly 100% FE and high formation rate of around 425\u00a0mmol g\u22121 h\u22121 at 1.0 V, with the product ration of CO/H2 approximating ideal 1/2 in the potential range from \u22120.7 to \u22121.0\u00a0V (Figure\u00a016K). Meanwhile, the catalyst displayed the long-term stability for 24\u00a0h with negligible degradation of current density (Figure\u00a016L) and the almost identical resistance to the Co NP-SNC (Figure\u00a016M). Potassium thiocyanate (KSCN) poisoning experiment was carried out to confirm the selectively functioning of Co SAs and N-C groups for CO2RR and HER. The CV curves and formation rate change were recorded in Figures\u00a016N and 16O, which showed the gaseous products increased significantly but the CO selectivity of Co-HNC decreased sharply to 9.8%. What\u2019s more, the flow-cell electrolyzer effectively solves the limitation of CO2 dissolution and diffusion in the traditional test device, realizing the highly selective conversion of CO2 under the high current density and accelerating the industrial application of CO2RR technology (Jin et\u00a0al., 2021a). Yuan et\u00a0al. designed single Cu atoms anchoring on the graphediyne (Yuan et\u00a0al., 2022). In situ Raman and DFT calculations revealed that the presence of Cu-C bonds leads to the formation of CH4, more facile during the process of CO2RR. The catalyst also showed high activity and CH4 FE and partial CH4 current density in flow-cell electrolyzer.Ammonia (NH3) is not only a key raw material for main agricultural fertilizers but also shows important applications in chemical engineering and pharmaceutical and synthetic fiber fields (Galloway et\u00a0al., 2004, 2008; Yang et\u00a0al., 2020a; Zamfirescu and Dincer, 2008). Currently, the industrial synthesis of NH3 commonly depends on the Haber-Bosch method under high temperature and pressure conditions (300\u2013500\u00b0C, 15\u201330 MPa), consuming more than 1% of the global energy supply annually (Guo et\u00a0al., 2018; Song et\u00a0al., 2019; van der Ham et\u00a0al., 2014a). Moreover, the thermodynamically limited conversion is only \u223c15%. Using N2 as raw material, electrocatalytic nitrogen reduction reaction (NRR) realizes the synthesis of ammonia at room temperature and pressure, which exhibits the merits of low energy consumption and without pollution. It provides a green and low-carbon technical route for ammonia synthesis industry. NRR involves a 6e\u2212 transfer process:\n\n(Equation\u00a023)\n\n\n\n\n\n\n\nN\n2\n\n\n(\ng\n)\n\n\n\n+\n6\nH\n\n+\n\n\n\n+\n6\ne\n\n-\n\n\n\n\u2192\n2\nNH\n\n3\n\n\n(\ng\n)\n\n\n\n\n\n\n\n\n\nHowever, the high bond energy of N\u2261N (940.95\u00a0kJ mol\u22121) is a major obstacle to the NRR process, so it is necessary to develop efficient electrocatalysts, especially SACs, to reduce the reaction energy barrier and accelerate the generation of NH3 (Chen et\u00a0al., 2018b; van der Ham et\u00a0al., 2014b; Wang et\u00a0al., 2017). What\u2019s more, the adsorption of N2 on the catalyst surface is usually not satisfactory, which is not conducive to the formation of intermediates and limits the selectivity and yield of NH3 (Tao et\u00a0al., 2019). Although many metal-based catalysts have been researched, most metals are too weakly bonded to achieve efficient N2 adsorption and activation, which is generally considered a rate-limiting step for NRR. Meanwhile, NH3 yield and faradaic efficiency (FE) are still far below the requirements of practical application.Carbon-based-material-supported SACs show great application potential toward NRR due to the abundant exposed active sites and high catalytic activity. Single Ru atoms anchored on nitrogen-doped carbon (Ru SAs/N-C) were fabricated by facile pyrolysis method. The Ru SAs/N-C achieved a recorded-high activity in NRR, which possessed an FE of 29.6% for NH3 production with partial current density of \u22120.13 mA cm\u22122 (Geng et\u00a0al., 2018). More importantly, the yield of the SACs reaches 120.9\u00a0\u03bcgNH3 mgcat.\n\u22121 h\u22121, well above the highest number ever reported (Figures\u00a017A\u201317D). The stability of Ru SAs/N-C displayed less than 7% attenuation of NH3 yield rate after 12\u00a0h potentiostatic measurement. DFT calculation showed that Ru SAs/N-C promoted N2 dissociation, resulting in increased activity relative to Ru NPs/N-C. In addition to noble metals, non-noble-metal SACs have also been studied in NRR. Copper single atoms attached in porous N-doped carbon network (Cu SAC) with Cu-N2 active sites as pH-universal catalyst showed outstanding NH3 yield rate and FE under 0.1\u00a0M KOH and 0.1\u00a0M HCl conditions (Figures\u00a017E\u201317H) (Zang et\u00a0al., 2019). Meanwhile, the Cu SAC also displayed excellent stability over 12\u00a0h with little current attenuation. The combination of experiment and first-principles calculations revealed that Cu-N2 coordination acts the effective active sites in NRR catalysis.Up to now, SACs have attracted extensive research interests in a wide range of catalytic fields, including photocatalysis, organic catalysis, electrocatalysis, and environmental aspect. The primary target of the rapid-developing SACs field is reducing the using of precious metals while keeping the catalytic activity. Carbon-based-material-supported SACs display great application prospect because of its low cost, high efficiency, and robustness. In this review, we introduced the synthesis methods and the advanced characterization techniques used in the identification of SACs, mainly concerning X-ray-derived spectroscopy and in situ techniques, which showed important guiding significance for coordination regulation and coordination environment recognition of SACs. In addition, the applications of carbon-based-material-supported SACs were discussed in electrocatalysis fields, including HER, OER, ORR, CO2RR, and NRR. To date, some progress has been made in enhancing catalytic performance of SACs. However, there are still many opportunities and challenges for the prospect of single atoms in the future.Firstly, the low loading of single atom in the SACs prepared by the existing synthesis strategies restricts the development of SACs. The sluggish reaction kinetics need to be overcome through exposing more active sites in catalysis. Low loading SACs may lead to the accumulation of intermediates during the reaction process, resulting in side reactions and reduced selectivity, which is not suitable for industrial scale applications. However, when the metal atom loading increases, the migration and agglomeration of single atoms tends to form nanoclusters or nanoparticles due to its high surface free energy. Therefore, it is imperative to develop SACs with high loading active sites for industrial production. In addition, it is very essential to study the interaction between metal single atoms and support, because the support shows an effect on the loading and electronic structure of single atom. For example, Xia and colleagues used graphene quantum dots as carbon substrates, which were modified with -NH2 groups to improve the coordination activity for metal ions (Xia et\u00a0al., 2021). The as-prepared transition metal single-atom material achieved a loading of up to 40 wt% and excellent thermal stability. Besides carbon-based materials, two-dimensional material transition metals, such as the sulfides, selenides, phosphides, and so on, have also been studied as carriers for SACs. The electron transfer between metal and carrier can be directly regulated by electronic metal-support interaction (EMSI), thereby regulating the electronic state of the supported metal. Therefore, it is of significance to develop the novel supports of SACs with superior catalytic performance for energy conversions.Secondly, the coordination environments show a great influence on the electronic and geometric structure of the central metal atoms, which plays an important role in the catalytic properties of SACs. Nonmetal heteroatomic doping (N, O, S, P, etc.) is one of the main strategies to regulate coordination environments. However, other elements, such as SE, Te, and halogen elements, are rarely studied and may display unexpected catalytic properties. In addition, the asymmetric distribution of charges may lead to superior performance. Thus, it is imperative to study the dual or more metal center sites. In a word, rationally constructing coordination environments of SACs is significant to boost the catalytic activity, which provides a direct way to understand the intrinsic activity of SACs.Thirdly, the characterization techniques are the significant fundament for the recognition of SACs. At present, the identification of coordination environments relies heavily on XAS, whereas the technique is bulk sensitive and only provides bulk average information. Therefore, it is very important to improve the spatial resolution of characterization technology. Furthermore, in order to determine the active sites of SACs and dynamic changes during the reaction, it is necessary to combine in situ characterization techniques. The dynamic changes of coordination structures and oxidation states of SACs during the catalytic process is worthy of further exploration because it is closely related to intrinsic activity.Finally, the step process and reaction mechanism of single-atom catalytic reaction are still in the preliminary exploration stage. Constructing a reliable structure-activity relationship of catalytic reactions is crucial for designing high-performance SACs. Theoretical simulation is conducive to understanding the structure-activity relationship of catalysts at atomic level. DFT calculation is a powerful tool to explore the atomic structure and intrinsic active sites. In addition, the reaction free energy of each elementary step and the adsorption energies of the intermediates can be obtained from DFT calculation, which is of great significance to the understanding of reaction mechanism. However, some of the proposed mechanisms do not match well with experimental results. More accurate models should be developed to reflect rational catalytic processes. What\u2019s more, DFT calculation combined with machine learning can predict efficient SACs, which shows a positive effect on the prospect of electrocatalysis. We believe that this work can promote the development of single-atom catalysis and deepen readers' understanding of single atoms.This work is supported by the National Key Research and Development Program of China (2021YFA1500500), National Natural Science Foundation of China (Grant Nos. 21822801 and 22005025), and China Postdoctoral Science Foundation (2021M700352).D. Cao and D.J. Cheng supervised the preparation of this review article. D. Cao, D.J. Cheng and H.M. Zhang conceived the topic. H.M. Zhang contributed to the most of the writing, and W.H. Liu contributed to some content and figures. H.M. Zhang and W.H. Liu revised the manuscript. D. Cao and D.J. Cheng revised and finalized the manuscript. All author approved the final version of the manuscript.The authors declare no competing interests.", "descript": "\n                  In recent years, single-atom catalysts (SACs) with unique electronic structure and coordination environment have attracted much attention due to its maximum atomic efficiency in the catalysis fields. However, it is still a great challenge to rationally regulate the coordination environments of SACs and improve the loading of metal atoms for SACs during catalysis progress. Generally, carbon-based materials with excellent electrical conductivity and large specific surface area are widely used as catalyst supports to stabilize metal atoms. Meanwhile, carbon-based material-supported SACs have also been extensively studied and applied in various energy conversion reactions, such as hydrogen evolution reaction (HER), oxygen evolution reaction (OER), oxygen reduction reaction (ORR), carbon dioxide reduction reaction (CO2RR), and nitrogen reduction reaction (NRR). Herein, rational synthesis methods and advanced characterization techniques were introduced and summarized in this review. Then, the theoretical design strategies and construction methods for carbon-based material-supported SACs in electrocatalysis applications were fully discussed, which are of great significance for guiding the coordination regulation and improving the loading of SACs. In the end, the challenges and future perspectives of SACs were proposed, which could largely contribute to the development of single atom catalysts at the turning point.\n               "}
{"full_text": "Data will be made available on request.Imines have wide health and industrial applications, for example, imines show anti-inflammatory and anticancer potential in health treatment [1] and serve as nitrogen-containing building blocks for various industrial chemicals [2]. Traditionally, imines are manufactured via condensation of primary amines and carbonyl compounds (e.g., aldehydes or ketones) [3]. Because of the disadvantages including heavy odor, easy deterioration, difficulty in conservation, handling and transportation, etc., the conventional process is not widely popularized.To overcome those issues, several new approaches have been proposed such as primary amines tethering [4], oxidative alcohols-amines coupling [5,6], secondary amines dehydrogenation [7], and amines' N-alkylation [8]. Among them, the alcohols-amines oxidative coupling method is considered as most promising [5,9]. Compared to aldehydes and ketones, alcohols are cheaper, widely available, more stable and less toxic. The reaction operates at mild temperature and ambient pressure and uses inexpensive molecular oxygen as the oxidant. Water is its sole by-product. Thus, the process is greener and more environmentally benign. So far, various supported heterogeneous catalysts such as Au [10], Pd [11], Pt [12], and Ru [13] have been examined for this reaction. Although, these noble-metal-based catalysts exhibited good performance in alcohols conversion and imine production, the high price and low availability of noble metals is a concern for large-scale application in imine synthesis. Developing transition-metal-based catalysts with comparative performance is much desired.Recently, manganese oxide (MnO\nx\n) has received increasing attention for imine synthesis largely because of its suitable physicochemical properties and relatively low cost [14\u201320]. Blackburn and Taylor [14] applied manganese oxides for imine synthesis from alcohols along with 4A zeolite for dehydration. Sithambaram and coworkers [15] developed a more efficient catalytic process by using OMS-2 as the bifunctional catalyst [15]. Other manganese-oxide-based catalysts such as MnO\nx\n/HAP [5], MnCo2O4\u2013500 [16], Mn1Zr0.5O\ny\n [17], and \u03b1-MnO2/GO [18] have also been reported. However, relatively high reaction temperature and long reaction time are still required, due to their relatively low activities.One promising approach to enhance the catalytic activity of manganese oxide catalysts is to alter their crystal structures and surface properties by doping transition metal ions into the framework or pore channels of manganese oxides [21\u201323]. OMS-2 is an allotrope of MnO2, consisting of a one-dimensional tunnel structure. Its 2\u00a0\u00d7\u00a02 edge is shared with MnO6 octahedra. The size of its tunnel opening is about 0.46\u00a0nm (see Fig. 1\n). Within OMS-2, Mn exists mainly as Mn4+, along with a small number of Mn3+ and Mn2+ ions, exhibiting an average oxidation state of \u223c3.8 [15]. OMS-2 shows good catalytic performance in some oxidation reactions due to its mixed valence manganese ions, large surface area and opening tunnel structures [15,24]. Doping low-valent metal ions (e.g., M2+, M3+) into the framework of OMS-2 has been widely studied, whereas only a few reported the incorporation of high-valent ions (e.g., M5+, M6+) [25,26]. High-valent metal ions have larger ionic radius and stronger Lewis acidity, which may be in favor of forming active oxygen species on the surface of manganese dioxides. For example, incorporating V5+ ions into OMS-2 enhanced the activity for aldehyde and methane combustion at low temperatures, which was attributed to the increased oxygen vacancies, Lewis acid sites and redox properties [27]. Consequently, vanadium doped OMS-2 may also work well for imine synthesis from oxidative coupling of amine and alcohols.Herein, we present our study for the first time on using vanadium doped cryptomelane-type manganese oxides (V-OMS-2) for imine synthesis via oxidative coupling of benzyl alcohol and aniline. Various characterization techniques such as X-ray diffraction powder (XRD), nitrogen physisorption (BET), scanning electron microscopy (SEM), transmission electron microscopy (TEM), temperature- programmed reduction (H2-TPR) and temperature-programmed desorption (NH3-TPD) were employed to characterize the catalyst structure. By combining with the reaction results, we were able to establish the correlation of catalyst structure and catalytic activity. The effects of vanadium doping amount and vanadium precursors on the physicochemical properties and catalytic performance of V-OMS-2 catalysts were systemically studied, and their relationship was discussed. The stability and recyclability of the catalyst was also investigated.The undoped OMS-2 was synthesized with a modified reflux method [28]. Typically, 9.9\u00a0g of MnSO4\u00b7H2O (>99%, Sinopharm) and 3.4\u00a0mL of HNO3 (68\u00a0wt%, Sinopharm) were dissolved in 35\u00a0mL of deionized water (solution-A). Then, 9.9\u00a0g of KMnO4 (>99%, Sinopharm) was dissolved in 120\u00a0mL of deionized water (solution-B). Next, solution-B was tardily added to solution-A under magnetic agitation. A brown slurry was formed immediately, which was further refluxed at 100\u00a0\u00b0C for 24\u00a0h. After naturally cooling down, the obtained brown-black precipitate was filtered, washed and recovered, and then dried at 110\u00a0\u00b0C for 12\u00a0h. The obtained solid was then treated at 250\u00a0\u00b0C in air for 2\u00a0h to completely remove adsorbed water from the pores.As for the preparation of vanadium doped OMS-2 catalysts, the same procedure was applied. A certain amount of vanadium pentoxide (V2O5, >97%, Sinopharm) or sodium metavanadate (NaVO3, >99%, Sinopharm) was added right after the completion of the solution-B addition. It should be pointed out that the presence of excess nitric acid (\u223c50\u00a0mmol) in the solution-A can ensure the complete dissolution of vanadium pentoxide (0.5\u20136\u00a0mmol) or metavanadate (\u223c3\u00a0mmol). The amount of vanadium pentoxide or sodium metavanadate was adjusted to get 1\u201312\u00a0mol% of V/Mn in the synthetic solution (i.e., vanadium concentration was 0.007\u20130.08\u00a0mol/L). The obtained vanadium doped catalysts are labeled as x%V-OMS-2(y), where x stands for the molar percentage of V over the input Mn in the synthetic solution and y represents the type of vanadium precursor, i.e., vanadium pentoxide (y\u00a0=\u00a01) or sodium metavanadate (y\u00a0=\u00a02). For comparison with the literature, the as-synthesized 3%V-OMS-2(1) was also calcined at 400\u00a0\u00b0C in air for 3\u00a0h, and the obtained catalyst was named as 3%V-OMS-2(1)-400.For comparison, low-valent transition metal (M) ions including Cu2+, Ni2+, Co2+, Fe3+, Cr3+ were also doped into the OMS-2 structure using the same procedure, and the molar percentage of M/Mn in the synthetic solution was fixed at \u223c3\u00a0mol%.The conventional wet impregnation method was also employed to prepare vanadium-doped OMS-2 catalyst at 3\u00a0mol%. Typically, 2.0\u00a0g of OMS-2 was immersed in a 10\u00a0mL aqueous ammonium metavanadate solution (\u223c 0.069\u00a0mol/L), which was stirred continuously for 2\u00a0h, then stood for 24\u00a0h, followed by 10\u00a0h of drying at 120\u00a0\u00b0C and another 3\u00a0h of roasting at 250\u00a0\u00b0C. The obtained catalyst is then termed as 3%V/OMS-2.The nitrogen adsorption-desorption isotherm was measured at \u2212196\u00a0\u00b0C with a Micromeritics ASAP 2020 system. Before analysis, each sample was degassed at 250\u00a0\u00b0C for 3\u00a0h. The standard Brunauer-Emmett-Teller (BET) method was used to calculate the specific surface area, and the Barrett-Joyner-Halenda (BJH) method was employed to evaluate the pore size distribution. The crystal structure of each catalyst sample was analyzed on a Bruker D8 Advance X-ray powder diffractometer using Cu K\u03b1 radiation (\u03bb\u00a0=\u00a00.15064\u00a0nm) at 40\u00a0kV and 30\u00a0mA. The SEM images and EDX spectra were taken on a HITACHI SU8010 electron microscope to analyze the morphology and chemical component of each catalyst sample. Transmission electron spectroscopy (TEM) images were obtained on a JEOL JEM-2100 electron microscope operated at 200\u00a0kV. A Perkin Elmer ELAN DRC-e inductively coupled plasma mass spectrometer (ICP-MS) was used to obtain the elemental content of catalyst sample. X-ray photoelectron spectra (XPS) were recorded on a ESCALAB 250Xi (Thermo-Fisher) spectroscope equipped with Al K\u03b1 as the exciting X-ray source. The C1s photoelectron peak at 284.8\u00a0eV was used as the reference to correct the binding energies of other elements. Hydrogen temperature-programmed reduction (H2-TPR) was performed on a home-assembled flow system equipped with a TCD detector. Typically, 20\u00a0mg of sample was loaded in a quartz reactor and treated with air at 250\u00a0\u00b0C for 1\u00a0h, followed by purging with pure N2 (30\u00a0mL\u00b7min\u22121) for 0.5\u00a0h. After cooling to 30\u00a0\u00b0C, the 5% H295% Ar mixture gas (30\u00a0mL\u00b7min\u22121) was introduced into the reactor, and the H2-TPR profile was recorded at a heating rate of 10\u00a0\u00b0C\u00b7min\u22121 up to 800\u00a0\u00b0C. NH3-TPD profiles were obtained on a Micromeritics Autochem 2920 II chemisorption device equipped with a TCD detector.The oxidative coupling of benzyl alcohol with aniline was carried out in a 50\u00a0mL three-necked round-bottom flask assembled with a condenser and an air balloon. In a typical run, benzyl alcohol (BA, 0.5\u00a0mmol), aniline (0.5\u00a0mmol), catalyst (50\u00a0mg), and solvent toluene (2\u00a0mL) were added into the flask and heated to 80\u00a0\u00b0C under continuous stirring for a certain time. As shown in Fig. S1, the mass transfer limitation can be excluded when the agitation speed is 800\u00a0rpm and higher. Therefore, all experiments were carried out at the stirring speed of 800\u00a0rpm. After reaction, the reactor was cooled down immediately and the solid catalyst was recovered by centrifugation. The obtained liquid products were analyzed by a gas chromatograph (GC-7890B) assembled with a FID detector and a Rtx\u00ae-5 type capillary column (30\u00a0m\u00a0\u00d7\u00a00.32\u00a0mm\u00a0\u00d7\u00a00.25 um) in the presence of dodecane as the internal standard. The calculated carbon balance (C-balance) was above 97% in all the reactions. The benzyl alcohol conversion, imine selectivity [17], yield, and catalyst activity (the formation rate of imine) were computed via the following equations:\n\n(1)\n\nConversion\n\n\n%\n\n=\n\n\n\nn\nBA\n0\n\n\u2212\n\nn\nBA\nt\n\n\n\nn\nBA\n0\n\n\n\u00d7\n100\n%\n\n\n\n\n\n(2)\n\nImine Selectivity\n\n\n%\n\n=\n\n\nn\nimine\nt\n\n\n\nn\nBA\n0\n\n\u2212\n\nn\nBA\nt\n\n\n\n\u00d7\n100\n%\n\n\n\n\n\n(3)\n\nYield\n\n\n%\n\n=\nconversion\n\u00d7\nimine selectivity\n\u00d7\n100\n%\n\n\n\n\n\n(4)\n\nAct\n.\n\n\n\n\u2212\nW\n\n\n=\n\n\nn\nimine\nt\n\n\n\nW\ncat\n\n\u00d7\nt\n\n\n\n\n\nmmol\n\u00b7\n\ng\n\n\u2212\n1\n\n\n\u00b7\n\nh\n\n\u2212\n1\n\n\n\n\n\n\n\nweight\n\u2212\nbased Activity\n\n\n\n\n\n\n\n(5)\n\nAct\n.\n\n\n\n\u2212\nA\n\n\n=\n\n\nn\nimine\nt\n\n\n\nS\ncat\n\n\u00d7\nt\n\n\n\n\n\nmmol\n\u00b7\n\nm\n\n\u2212\n2\n\n\n\u00b7\n\nh\n\n\u2212\n1\n\n\n\n\n\n\n\nsurface area\n\u2212\nbased Activity\n\n\n\n\nwhere n\n\nBA\n\n0 is the moles of benzyl alcohol in the feed (mmol), n\n\nBA\n\n\nt\n is the moles of benzyl alcohol (mmol) and n\n\nimine\n\n\nt\n is the amount of imine (mmol) in the liquid at the reaction time of t (h), W\n\ncat\n is the mass of catalyst used for the reaction (g), S\n\ncat\n is the total surface area of catalyst used for the reaction (m2).The N2 adsorption-desorption isotherms of V-OMS-2(1) samples with different V/Mn ratios are shown in Fig. 2\n. Both the OMS-2 and 1%V-OMS-2(1) samples exhibited the IUPAC defined type-II isotherm, while it became less distinct for the 3%, 6% and 12%V-OMS-2(1) samples. Generally, the hysteresis loop in the type-II isotherm appears at the relative pressures of 0.8\u00a0<\u00a0P/P0\u00a0<\u00a01.0. Its shape represents the difference in the relative proportions of voids or pores between particles. Such a hysteresis loop for the 3%, 6% and 12%V-OMS-2(1) became much smaller. Meanwhile a new hysteresis loop formed at the relative pressure range of 0.4\u00a0<\u00a0P/P0\u00a0<\u00a00.8, indicating the emergence of mesoporous structure in these samples. Their corresponding BJH pore size distribution curves are shown in Fig. S2.\nTable 1\n displays the physical properties of each sample including surface area (SA), pore volume and pore size. The specific surface area of undoped OMS-2 is about 86 m2\u00b7g\u22121, of which \u223c16% is related to micropores (14\u00a0m2\u00b7g\u22121). Its micropore volume is only 0.0067\u00a0cm3\u00b7g\u22121, which is <2% of the total pore volume (0.43\u00a0cm3\u00b7g\u22121), suggesting that the contribution of micropores is negligible. Doping vanadium into OMS-2 increased the surface area, especially the non-microporous surface area. The surface area of V-OMS-2(1) samples reached to the highest at 6\u00a0mol%\u00a0V doping (SA\u00a0=\u00a0241\u00a0m2\u00b7g\u22121), then decreased slightly at the vanadium content of 12\u00a0mol% (SA\u00a0=\u00a0206\u00a0m2\u00b7g\u22121). On the other hand, the pore volume and average pore size of V-OMS-2(1) samples decreased gradually with the increase of V/Mn ratio. A remarkable decrease was observed when the V/Mn ratio was higher than 6\u00a0mol%. This may be correlated with the decrease in the particle size and the crystallinity of these samples [28].\nFig. 3\n compares the N2 adsorption-desorption isotherm and pore size distribution of the undoped OMS-2, 3%V-OMS-2(1), 3%V-OMS-2(2) and 3%V/OMS-2 samples. All samples display the type-II isotherm, except the 3%V-OMS-2(1) sample which exhibits the feature of type-IV isotherm. Compared to the undoped OMS-2, the pore size of 3%V-OMS-2(2) and 3%V/OMS-2 became slightly larger (Fig. 3B), whereas their surface areas and pore volumes including micropore surface area and micropore volume were smaller (Table 1\n, entry 7 and 8). Over the 3%V-OMS-2(1) sample, the pore size distribution became much narrower, showing a significant decrease in the pore size (20\u00a0\u2192\u00a09.2\u00a0nm). On contrast, its surface area remarkably increased (86\u00a0\u2192\u00a0179\u00a0m2\u00b7g\u22121), whereas the change in the pore volume was little (0.43 to 0.41\u00a0cm3\u00b7g\u22121).The XRD patterns of the OMS-2 and V-OMS-2(1) samples with different V/Mn ratios are presented in Fig. 4\n. They all were alike to the tetragonal structure of natural cryptomelane-type manganese oxide (OMS-2, JCPDS 20\u20130908), confirming the success in obtaining and preserving the 2\u00a0\u00d7\u00a02 tunnel structured manganese oxide phase. The diffraction peaks at 2\u03b8\u00a0=\u00a012.8, 18.1, 28.8, 37.6, 42.0, and 50.1o are assigned to (110), (200), (310), (211), (301), and (411) of the OMS-2 crystal phase, respectively. With V/Mn ratio increasing, these diffraction peaks became broader while the intensities dropped significantly. It indicates that V5+ ions are embedded into the OMS-2 skeleton, resulting in a decrease in the OMS-2 crystallinity. When the V/Mn ratio was higher than 6\u00a0mol%, only one broad diffraction peak at 2\u03b8\u00a0=\u00a038.7o was detected, showing a significant loss of the OMS-2 crystal structure. It suggested that doping large amount of vanadium could impede the formation of OMS-2 structure.It should be highlighted that no additional diffraction peaks related to any vanadium species were detected in all the samples studied, suggesting the doped vanadium species were well-dispersed either on the surface or within the framework of OMS-2 [26].\nFig. 5\n compares the XRD patterns of the OMS-2 and 3%V-OMS-2 samples prepared with different vanadium precursors. All samples exhibited similar XRD patterns, matching well with the standard tetragonal structure of cryptomelane-type MnO2 (JCPDS 20\u20130908). The intensity of diffraction peaks decreases as follows: OMS-2, 3%V/OMS-2\u00a0>\u00a03%V-OMS-2(2) > 3%V-OMS-2(1). The weaker diffraction peaks suggest that more vanadium ions were incorporated into the OMS-2 structure, resulting in the reduction of the OMS-2 crystallinity. Such a phenomenon was also reported over Ce-doped OMS-2 catalysts [29].TEM and high-resolution TEM (HRTEM) images of the OMS-2, 3%V-OMS-2(1), 3%V-OMS-2(2) and 6%V-OMS-2(1) samples are presented in Fig. 6\n. The OMS-2 showed a nanorod-like morphology with the length of 200\u2013400\u00a0nm. The 3%V-OMS-2(1) and 3%V-OMS-2(2) samples both exhibited similar nanorod-like morphology as the OMS-2, but with a much shorter length of 50\u2013100\u00a0nm and a slight larger width. Similar morphological changes can also be observed in the SEM images of these samples (Fig. S3). Through the HRTEM images, we can observe the well-defined lattice planes of these samples. Over OMS-2, the lattice fringe is 0.69\u00a0nm, which can be attributed to the (110) planes. In the 3%V-OMS-2(1) and 3%V-OMS-2(2) samples, the fringe spacing was 0.31\u00a0nm, which is assigned to the (310) plane.As for the 6%V-OMS-2(1) sample, nanoparticles sized from 5 to 20\u00a0nm were formed and served as building blocks to construct the flake or spherical agglomerates (Fig. 6D). The nano crystallites of 6%V-OMS-2(1) sample were randomly oriented, indicating extremely short-range order of lattice structure (Fig. 6D1). The lack of diffraction rings or spots in the selected area of electron diffraction pattern (Fig. 6D, inset) confirmed the poor crystal phase of this sample, which is consistent with the XRD result. No other separated phases were identified in both TEM and SEM images of all the samples, further confirming the high dispersion of vanadium species.Mn 2p, V 2p and O 1\u00a0s XPS spectra of the OMS-2 and V doped OMS-2 samples are shown in Fig. 7\n. The Mn 2p3/2 spectra of all samples (Fig. 7A) are very broad and asymmetrical, suggesting the coexistence of various manganese species. Thus, we conducted the peak deconvolution based on the literature [30] and obtained three characteristic peaks at 640.5, 641.7 and 643.0\u00a0eV, corresponding to surface Mn2+, Mn3+ and Mn4+ species, respectively. Then their composition was computed accordingly and is presented in Table 2\n\n. The undoped OMS-2 contains mainly Mn4+ on the surface (79.5%) along with 17.2% Mn3+. Doping vanadium results in the increase of Mn3+ composition. It is 21.3% Mn3+ in the 3%V-OMS-2(1) and further to 27.3% in the 6%V-OMS-2(1) sample, along with the decrease of both Mn4+ and Mn2+. It means the replacement of Mn with V favors the formation of Mn3+ ions, which may be beneficial to the catalytic activity, as the Mn3+/Mn4+ pair is important for catalyzing oxidation reaction [28].The binding energy of V 2p3/2 was 517.0\u00a0eV for all V-OMS-2 samples (Fig. 7B), close to the binding energy of vanadium pentoxide reported in the literature [31], indicating that the doped vanadium species are mainly in V5+ state. The O 1\u00a0s spectra can be deconvoluted into three peaks at 529.5\u00a0eV (\u03b1 peak), 531.0\u00a0eV (\u03b2 peak) and 533.0\u00a0eV (\u03b3 peak), corresponding to the lattice oxygen (Osat), the surface-adsorbed oxygen (Ounsat) and the oxygen in the surface-adsorbed water, respectively [32]. As shown in Table 2, the ratio of Ounsat/(Osat\u00a0+\u00a0Ounsat) for the OMS-2 was 0.293. The value increased with the increase of vanadium doping, implying that the more vanadium atoms were incorporated into the framework of the OMS-2, the more adsorbed oxygen species on the surface. The Ounsat/(Osat\u00a0+\u00a0Ounsat) ratio of 6%V-OMS-2(1) (0.391) was higher than that of 3%V-OMS-2(1) (0.339), indicating that the 6%V-OMS-2(1) catalyst possessed more surface-adsorbed oxygen species.EDX spectra of the undoped OMS-2, 3%V-OMS-2(1), 3%V-OMS-2(2) and 3%V/OMS-2 samples are shown in Fig. 8\n. The K, Mn and O peaks are clearly observed over the undoped OMS-2 (Fig. 8A). A new peak emerged at 4.95\u00a0keV over the V doped OMS-2 samples, confirming that V was successfully introduced into OMS-2. Compared to 3%V-OMS-2(2), 3%V-OMS-2(1) had less K, O and V, but higher Mn content (Table S1), indicating a chemical composition difference when using a different vanadium precursor.The bulk V/Mn ratios for all 3%V-OMS-2 samples measured by ICP and the surface V/Mn ratios for the 3%V-OMS-2(1) and 3%V-OMS-2(2) samples estimated by XPS are summarized in Table 1. The bulk V/Mn ratios of all samples were very close to the input for the synthesis. It suggests that all vanadium from the feedstock were incorporated into OMS-2. On the other hand, the 3%V-OMS-2(2) catalyst showed identical V/Mn ratio on the surface to that in the bulk value (3.0 vs. 3.2), implying that the doped vanadium in this sample was uniformly dispersed. For the 3%V-OMS-2(1) sample, however, its V/Mn ratio on the surface was only 1.9, much smaller than that in the bulk. It indicates that more vanadium atoms located inside the channels/framework of OMS-2. Replacing framework Mn atoms with V could inhibit the growth of OMS-2 crystals, resulting in poor crystallinity and more structural defects, which is in a good agreement with XRD and TEM results. Additionally, the 3%V-OMS-2(1) catalyst showed higher Ounsat/(Osat\u00a0+\u00a0Ounsat) ratio (0.339) than that of 3%V-OMS-2(2) sample (0.303). It implies that vanadium pentoxide is a better vanadium precursor in terms of generating more active surface oxygen species by adsorbing oxygen molecules from air on the defective sites [33,34]. More active surface oxygen species could promote the oxidative coupling of benzyl alcohol and aniline to imine.The redox properties of the OMS-2 and V-OMS-2(1) samples with different V/Mn ratios were examined by H2-TPR, which are shown in Fig. 9\n. The OMS-2 exhibited three reduction peaks at the temperature ranged from 250 to 400\u00a0\u00b0C, which is labeled as \u03b1 (at 335\u00a0\u00b0C), \u03b2 (at 359\u00a0\u00b0C) and \u03b3 (at 376\u00a0\u00b0C), respectively. The \u03b1 peak is attributed to the reduction of the surface oxygen species [35]. Whereas the \u03b2 and \u03b3 peaks are attributed to the sequential reduction of MnO2\u00a0\u2192\u00a0Mn3O4 and then Mn3O4\u00a0\u2192\u00a0MnO, respectively [36]. Doping vanadium changes the reduction behavior of the OMS-2 material (see Fig. 9 and Table 3\n). The 1%V-OMS-2(1) samples exhibited a similar H2-TPR profile to OMS-2, with a slight shift of both \u03b2 and \u03b3 reduction peaks to a lower temperature. While over the 3, 6 and 12%V-OMS-2(1) samples, the H2-TPR profiles changed significantly. First, the \u03b1 reduction peak gradually shifted to lower temperature and became larger with the increase of vanadium content, which means the increase in the surface oxygen species with doping vanadium into OMS-2 [28], in good agreement with the XPS results. Lower reduction temperature also suggests the increase in the mobility of oxygen species [28,37]. Second, the \u03b2 and \u03b3 peaks became distinctly separated, due to a considerable shift of the \u03b3 reduction peak to a higher temperature. It indicates that with the presence of vanadium, the reduction of Mn3O4 species becomes harder. On the other hand, the \u03b2 reduction peak became relatively smaller, which may suggest the decreased amount of MnO2 in OMS-2 with vanadium doping. In fact, such a change is consistent with XPS results shown in Table 2, where the Mn4+ content decreased and the Mn3+ content increased with the increase of vanadium doping. Third, a new reduction peak was observed at 550\u00a0\u00b0C over the 6%V-OMS-2(1) sample and it became larger over the 12%V-OMS-2(1) sample. Since V\u2212O bond is stronger than MnO bond, a higher temperature is required for the reduction of vanadium oxide [37,38]. Thus, this new peak can be ascribed to the reduction of vanadium species, although they were not detected by XRD. Furthermore, the presence of vanadium oxide species on the external surface may hinder the reduction of surface-adsorbed oxygen species. As a result, only a broad reduction peak at 321\u00a0\u00b0C (i.e., merged \u03b1 and \u03b2 peaks) was obtained over the 12%V-OMS-2(1) sample.Based on H2-TPR profiles, the H2 consumption was estimated and is listed in Table 3. The H2 consumption was 10.9\u00a0mmol\u00b7g\u22121 for the undoped OMS-2, which is slightly smaller than the theoretical value for the complete reduction of MnO2 to MnO (11.5\u00a0mmol\u00b7g\u22121). Over the V-OMS-2(1) samples, the H2 consumption decreased with the increase of vanadium doping, being 10.7, 9.8, 9.3 and 8.5\u00a0mmol\u00b7g\u22121 for 1, 3, 6 and 12%V-OMS-2(1) samples, respectively. The decreased H2 consumption could be attributed to the formed Mn3+ species induced by the introduction of vanadium, as suggested by XPS.The influence of vanadium precursor on the redox property of prepared V-OMS-2 catalyst was also studied by H2-TPR. The results are shown in Fig. 10\n. Although all four samples consumed similar amount of hydrogen during the reduction (see Table 3), their H2-TPR profiles are distinguishably different. The 3%V/OMS-2 sample prepared with wet impregnation method exhibited the most different reduction behavior. All the three reduction peaks (i.e., \u03b1, \u03b2, and \u03b3 peaks) are delayed to higher temperatures and the \u03b2 and \u03b3 peaks are even merged into one peak at 450\u00a0\u00b0C. Considering that the reduction of vanadium oxides generally occurs above 500\u00a0\u00b0C [38], it indicates that the vanadium species introduced by impregnation method are mostly located on the external surface of OMS-2, which retards the reduction of manganese oxides in OMS-2.Compared to 3%V-OMS-2(1), 3%V-OMS-2(2) exhibited the \u03b1 peak at a relatively higher temperature, but slightly lower than that of OMS-2. It suggests that the mobility of surface oxygen species over 3%V-OMS-2(2) is better than OMS-2 but worse than 3%V-OMS-2(1). The reduction temperature of both \u03b2 and \u03b3 peaks shifted to a higher temperature, whereas the \u03b3 peak changed more, resulting in a relative more separation of the \u03b2 and \u03b3 peaks in comparison to the OMS-2 sample, but still less than that in the 3%V-OMS-2(1) sample. It implies that the formation of Mn3+/Mn4+ pair in the 3%V-OMS-2(2) sample is more than that of OMS-2, but less than that of 3%V-OMS-2(1) sample, which is consistent with XPS results (Fig. 7 and Table 2).It is reported that weak acid sites of a catalyst play an important role in the imine synthesis in oxidative coupling of benzyl alcohol and aniline [15,39]. This is because the weak acid sites can interact strongly with aniline, which is a weak base. Consequently, the NH3-TPD experiments were carried out to determine the amount and strength of the weak acid sites on the vanadium doped OMS-2 catalysts. The obtained NH3-TPD profiles are presented in Fig. 11\n and the estimated weak acid sites are listed in Table 3 (the value for individual peak is shown in Table S2). All the profiles can be deconvoluted into four peaks named as A, B, C and D. According to literature [40,41], the A and B peaks are attributed to weak Br\u00f8nsted acid sites and the C and D peaks are related to Lewis acid sites.As shown in Table 3, the OMS-2 sample has more Lewis acid sites (0.50\u00a0mmol\u00b7g\u22121) than the weak Br\u00f8nsted acid sites (0.15\u00a0mmol\u00b7g\u22121), which is consistent with that reported in literature [42]. After doping vanadium, both weak Br\u00f8nsted acid sites and Lewis acid sites significantly increased as compared with the original OMS-2. The total weak acid sites were 1.21, 1.45, 1.40 and 1.16\u00a0mmol\u00b7g\u22121 for 1, 3, 6 and 12% V-OMS-2(1) samples, respectively. The maximum amount of acid sites was obtained over the 3%V-OMS-2(1) sample, which is more than double of the acid sites on the undoped OMS-2. As indicated by XRD and TEM, doping vanadium higher than 3\u00a0mol% could result in worse crystallinity and smaller particles. It may be the reason for the decrease in the acid sites over the 6 and 12% V-OMS-2 samples.We also compared the effect of vanadium precursor on acid properties of V-OMS-2 catalysts by NH3-TPD. Fig. 12\n shows the NH3-TPD profiles of the undoped OMS-2, 3%V-OMS-2(1), 3%V-OMS-2(2) and 3%V/OMS-2 samples. Their corresponding acid sites are computed and listed in Table 3. The total acid sites decrease as 3%V-OMS-2(1) (1.45\u00a0mmol\u00b7g\u22121)\u00a0>\u00a03%V-OMS-2(2) (0.79\u00a0mmol\u00b7g\u22121)\u00a0>\u00a0OMS-2(1) (0.65\u00a0mmol\u00b7g\u22121)\u00a0>\u00a03%V/OMS-2 (0.57\u00a0mmol\u00b7g\u22121). 3%V-OMS-2(1) sample has the acid sites 80% more than the 3%V-OMS-2(2). Whereas the 3%V/OMS-2 sample shows even less acid sites than the undoped OMS-2, suggesting that loading vanadium by wet impregnation has negative impact on the acidity of OMS-2. The NH3 desorption on the 3%V-OMS-2(1) sample appears at lower temperatures among the four samples, implying that the acidity on 3%V-OMS-2(1) is the weakest. The results indicate that more and weaker acid sites can be obtained when vanadium pentoxide is used as the precursor.The catalytic performances of various transition metal doped OMS-2 catalysts have been examined for the imine synthesis from oxidative coupling of benzyl alcohol and aniline with air. The results are shown in Table S3. Compared to OMS-2 catalyst, only the V doped OMS-2 showed a significant promotion in the benzyl alcohol conversion and imine yield, while doping other transition metals including Cu, Ni, Co, Fe and Cr all resulted in a decrease in the catalytic performance. Thus, the vanadium doped OMS-2 catalysts are investigated more in detail for this reaction.As shown in Table 4\n (Entry 1\u20138), doping vanadium enhanced catalytic activity of the V-OMS-2(1) catalysts for oxidative coupling of benzyl alcohol and aniline to imine. The benzyl alcohol conversion increased with V/Mn ratio, reached the maximum at the V/Mn ratio of 3\u00a0mol%, then dropped gradually with the further increase of vanadium content. The weight-based activity (Act.-W) also followed the same trend. The high activity of the V-OMS-2(1) catalyst could be attributed to the improvement in the surface area, surface oxygen species (or Mn3+/Mn4+ pair), weak acid sites, and relatively smaller particle size as the characterization results indicated (see Section 3.1).In order to assess the contribution of catalyst surface area, the surface-area-based activity (Act.-A) was calculated and is also presented in Table 4. The Act.-A of the undoped OMS-2 was 0.020\u00a0mmol\u00b7m\u22122\u00b7h\u22121. When 1 and 3\u00a0mol%\u00a0V was doped to OMS-2, the Act.-A became 0.024 and 0.016\u00a0mmol\u00b7m\u22122\u00b7h\u22121. The results exhibited a different trend from that observed with the weight-based activity. It suggests that the surface area may play an important role in determining the catalyst activity for the imine synthesis reaction. Although the 1%V-OMS-2(1) catalyst showed a slightly better Act.-A than the 3%V-OMS-2(1) catalyst, the 3%V-OMS-2(1) catalyst exhibited a much higher benzyl alcohol conversion, imine yield and Act.-W, which can be attributed to its significantly larger surface area (179 vs. 95 m2\u00b7g\u22121 in Table 1).In fact, we did obtain same surface-area-based activity over the V-OMS-2(1) catalysts when the reaction was conducted over the catalysts with the same surface area (please see the discussion in Section 3.2.3).When more vanadium was doped, a decrease in the Act.-A was observed. It was 0.011 and 0.009\u00a0mmol\u00b7m\u22122\u00b7h\u22121 for the 6 and 12%V-OMS-2(1) catalyst, respectively, which is smaller than that of OMS-2 catalyst, and much smaller than that of the 1%V-OMS-2(1) catalyst. However, both the 6 and 12%V-OMS-2(1) catalysts showed better performance than OMS-2 catalyst, especially the 6%V-OMS-2(1) catalyst, which was even better than the 1%V-OMS-2(1) catalyst in terms of benzyl alcohol conversion, imine yield and Act.-W. Considering the significantly higher surface area of 6%V-OMS-2(1), it can be further concluded that the surface area plays an important role in the imine synthesis over the V-doped OMS-2 catalysts.It should be pointed that the 6%V-OMS-2(1) catalyst possessed higher surface area and more surface-adsorbed oxygen species and similar acid sites than the 3%V-OMS-2(1) catalyst as suggested by BET, XPS, H2-TPR and NH3-TPD characterizations, however, its catalytic activity was less than the 3%V-OMS-2(1) catalyst. Same phenomenon was also observed on the 12%V-OMS-2(1) catalyst versus the 1%V-OMS-2(1) catalyst. The XRD results clearly showed the loss in the OMS-2 crystal structure of the 6%V-OMS-2(1) and 12%V-OMS-2(1) catalysts (Fig. 4). It indicates that the preservation of the OMS-2 structure is also crucial for this reaction.To understand the effect of vanadium precursor, the synthesized 3%V-OMS-2(2) and 3%V/OMS-2 catalysts were also examined for the oxidative coupling of benzyl alcohol with aniline at the same reaction conditions, which are also presented in Table 4\n(Entry 9\u201310). Over the undoped OMS-2 catalyst, the benzyl alcohol conversion was 60% and the imine yield was about 52%. Over the impregnated 3%V/OMS-2 catalyst, the catalytic activity dropped significantly to 23%. The characterization results showed that on this catalyst, the vanadium species are mainly located on the external surface of OMS-2, resulting in a decrease in the surface area. Moreover, the impregnated vanadium decreased the acid sites of the catalyst as indicated by the NH3-TPD characterization. As a result, the benzyl alcohol conversion decreased much. It indicates that the impregnated vanadium species have no promotion for the imine synthesis from oxidative coupling of benzyl alcohol and aniline. On the contrast, both 3%V-OMS-2(1) and 3%V-OMS-2(2) exhibited higher catalytic activity for the reaction, the benzyl alcohol conversion of which was 94% and 70%, respectively. Compared to the 3%V-OMS-2(2) catalyst, the 3% V-OMS-2(1) catalyst prepared with vanadium pentoxide possessed much higher catalytic activity, which could be associated with its larger surface area, better mesopore structure, shorter nanorod morphology, the substantially increased acid sites, more substitution of framework Mn species with vanadium and greatly enhanced surface reactive oxygen species as suggested by BET, XPS, SEM, TEM, H2-TPR and NH3-TPD.Since our catalysts were only treated at 250\u00a0\u00b0C, a conventionally calcined catalyst, the 3%V-OMS-2(1)-400 catalyst was also prepared by calcination at 400\u00a0\u00b0C and evaluated for the oxidative coupling of benzyl alcohol and aniline. Compared to the 3%V-OMS-2(1) catalyst, both benzyl alcohol conversion and imine yield of the 3%V-OMS-2(1)-400 catalyst were significantly lower. Although 3%V-OMS-2(1)-400 showed an improved crystal structure than 3%V-OMS-2(1) (Fig. S4), its acid sites were much less (Fig. S5). It suggests that high temperature calcination could reduce the catalyst acid sites, which may be responsible for the decrease in the catalytic activity of the 3%V-OMS-2(1)-400 catalyst.As discussed above, the 3%V-OMS-2(1) catalyst showed the best performance among the catalysts studied. Consequently, we conducted the reaction condition optimization on this catalyst. We first studied the influence of solvent. As shown in Table 5\n, the solvent plays a crucial role in this reaction. Non-polar and aprotic toluene is particularly efficient in providing both high benzyl alcohol conversion and imine selectivity. Although a high benzyl alcohol conversion was obtained when acetonitrile was used, the selectivity of imine was low (Entry 7). Dichloroethane led to a high imine selectivity, but the benzyl alcohol conversion was poor (Entry 8). Other solvent, such as THF, isopropanol and dioxane, were also examined. Both the benzyl alcohol conversion and the imine selectivity were low (Entry 9\u201311).Second, the effect of reaction temperature was investigated. After 24\u00a0h of reaction at 30\u00a0\u00b0C, the benzyl alcohol conversion was only 29% and the selectivity of imine was 86% (Table 5\n, Entry 3). With increasing temperature, both the benzyl alcohol conversion and the selectivity of imine improved greatly. For example, after 12\u00a0h of reaction at 60\u00a0\u00b0C, the benzyl conversion and the selectivity of imine were 97% and 93%, respectively (Table 5\n, Entry 2). When the temperature was raised to 80\u00a0\u00b0C, 99% of benzyl alcohol conversion and 93% of imine selectivity were achieved with only 4\u00a0h of reaction (Table 5\n, Entry 1).We have also examined the effect of aniline-to-BA ratio. When the aniline:BA ratio was changed from 1:1 to 1.5:1, the benzyl alcohol conversion increased from 94% to 97%, and the selectivity of imine increased from 90% to 95% (Table 5\n, Entry 4 and 5). Further increasing the aniline:BA ratio to 2:1, the benzyl alcohol conversion and the imine selectivity reached 99% and 100%, respectively (Table 5\n, Entry 6). Although adding more aniline is beneficial, the improvement in the conversion and imine selectivity is not substantial.\nTable 6\n compares the catalytic performance of our 3%V-OMS-2(1) catalyst with those reported in literature [5,10,13,16,39,43]. The 3%V-OMS-2(1) catalyst considerably outperforms the other catalysts including some noble-metal-based catalysts. The reaction time to achieve the similar benzyl alcohol conversion and imine yield is remarkably shorter. Its Act.-W value is significantly better than the other catalysts (Table 6\n, Entry 3\u20137). Over the 3%V-OMS-2(1) catalyst, the highest benzyl alcohol conversion (\u223c99%) was attained after 4\u00a0h of reaction, where the imine yield was as high as 92% (Table 6\n, Entry 2). The highest Act.-W of 2.83\u00a0mmol\u00b7g\u22121\u00b7h\u22121 (Table 6 Entry 1) was achieved at the reaction time of 3\u00a0h in this work, which is more than double of the best Act.-W (1.24\u00a0mmol\u00b7g\u22121\u00b7h\u22121 [16]) for the Mn-based catalyst reported in literature. Although the Ru/Zn1@Ui-66 catalyst showed a higher Act.-W (4.12\u00a0mmol\u00b7g\u22121\u00b7h\u22121 [43]), it consumes more aniline and uses pure oxygen for the reaction.As discussed above, the incorporation of vanadium can induce the change in the surface area, acid sites, Mn3+ component and active surface oxygen species of the OMS-2, thus resulting in a difference catalytic performance in oxidative coupling of benzyl alcohol and aniline to imine. To clarify which factor plays a more important role, we further studied the effect of the catalyst amounts on the reaction performance over the OMS-2 and 3%V-OMS-2(1) catalysts. As shown in Table 4 (entry 1 and 2), when the amount of OMS-2 increased from 50\u00a0mg to 104\u00a0mg (so that the total surface area of OMS-2 catalyst is the same as that of 50\u00a0mg 3%V-OMS-2(1) catalyst), the conversion of benzyl alcohol increased from 60% to 83%, and the yield of imine increased from 52% to 73%. Its surface area-based activity (Act.-A) was 0.014\u00a0mmol\u00b7m\u22122\u00b7h\u22121, which is very close to that of the 3%V-OMS-2(1) catalyst at the same surface area level for the reaction. On the other hand, when we used 24\u00a0mg of 3% V-OMS-2 (the same surface area as that of 50\u00a0mg OMS-2) for the reaction, the benzyl alcohol conversion dropped to 69% and the corresponding imine yield was 60%. The Act.-A of the 3% V-OMS-2(1) catalyst became 0.023\u00a0mmol\u00b7m\u22122\u00b7h\u22121, which is also close to that of the OMS-2 catalyst (0.020\u00a0mmol\u00b7m\u22122\u00b7h\u22121,) at the same surface area level for the reaction. The results disclose that the improved catalytic performance of the 3%V-OMS-2(1) catalyst mainly related to its increased surface area. In other words, the increased surface area induced by vanadium doping plays a major contribution to the enhanced activity of the V-OMS-2 catalyst.We have further investigated the relationship between the catalytic activity of V-doped OMS-2 catalysts and their redox property (Mn3+ component and active surface oxygen species) and acidity. Fig. 13\n shows the surface area-based activity (Act.-A), the total acidity, the fraction of surface Mn3+ ions and Ounsat species of the catalysts as a function of V/Mn ratio in the V-doped OMS-2 catalysts. Both the activity and the total acidity showed the same trend with the increase of V/Mn ratio (Fig. 13A), showing higher Act.-A associated with higher acidity, i.e., a proportional relationship. It suggests a positive correlation between the catalyst activity and the total acidity. On the contrast, we observed a tradeoff between the catalyst activity and the fraction of surface Mn3+ ions and Ounsat species (Fig. 13B). Both the activity and the fraction of surface Mn3+ ions and Ounsat species increased with the addition of vanadium. When the V/Mn ratio was higher than 3 mol%, although the fraction of surface Mn3+ ions and Ounsat species kept increasing, the catalyst activity decreased sharply instead. The XRD and TEM results showed a degradation of OMS-2 structure at higher V/Mn ratio. Thus, it indicates that preserving the OMS-2 structure is critical for achieving positive impact of the fraction of surface Mn3+ ions and Ounsat species on the catalyst activity.To understand the reaction pathway over the 3%V-OMS-2(1) catalyst\uff0cthe dependence of catalytic activity on the reaction time was investigated. As shown in Fig. 14A, at the beginning of the reaction, the selectivity of imine was relatively low, whereas the selectivity of benzaldehyde was much higher. The initial reaction rate was about 7.1\u00a0mmol-BA\u00b7g\u22121\u00b7h\u22121. With prolonging the reaction time, the selectivity of imine increased, while the selectivity of benzaldehyde decreased. After 2\u00a0h of the reaction, the selectivity of imine and benzaldehyde changed little. The conversion of benzyl alcohol reached 99% in 4\u00a0h with the corresponding imine yield of 92%. We have also tracked the conversion of aniline and the production of related products from aniline conversion (Fig. 14B). The yield of the imine (i.e., N-benzylideneaniline, PhC=NPh) was almost overlapped with the aniline conversion, along with the detection of small amount of azobenzene (PhN=NPh) byproduct (< 3%), which is generated from the self-coupling of aniline reactant. The results indicate that imine (i.e., N-benzylideneaniline) is the main product, and the benzaldehyde is a primary product generated from the selective oxidation of benzyl alcohol.According to our experimental results and previous reports [15,18,39], the air-oxidative coupling reaction between benzyl alcohol and amine over 3%V-OMS-2(1) possibly proceeds with two consecutive steps, that is, oxidation of benzyl alcohol to benzaldehyde, followed by condensation with aniline to form imine, as depicted in Scheme 1\n. In this process, the 3%V-OMS-2(1) serves as a bifunctional catalyst to catalyze these two distinct steps (i.e., selective oxidation and condensation), which is closely related to the catalyst's surface area and acid sites as discussed in the structure-activity relationship section.In order to provide further mechanistic insights, three key control experiments including (i) aerobic oxidative of benzyl alcohol without aniline, (ii) aniline without benzyl alcohol, and (iii) condensation of benzaldehyde with aniline were carried out under the same reaction conditions and the results are presented in Fig. 15\n\n. In the control experiment (i), the benzyl alcohol was readily converted without the presence of aniline and the conversion increased with the increase of reaction time. The initial reaction rate was 5.7\u00a0mmol\u00b7g\u22121\u00b7h\u22121. Benzaldehyde was the sole product at the selectivity of 100%. The BA conversion in this experiment was lower than that with the presence of aniline (conversion in Fig. 14A). The result suggests that the presence of aniline is indispensable to the formation of N-benzylideneaniline and can promote the conversion of BA to benzaldehyde by consuming benzaldehyde to produce imine.Without the presence of benzyl alcohol, aniline can also be readily converted, but at a much slower reaction rate (control experiment (ii)). The initial reaction rate was about 2.0\u00a0mmol\u00b7g\u22121\u00b7h\u22121. The aniline conversion gradually increased within 2\u00a0h and became stable at about 25% with the reaction time suggesting a reaction equilibrium under the conditions studied. More important, the reaction product was azobenzene (100% selectivity), indicating a self-coupling reaction of aniline.Compared to single reactant reactions in control experiment (i) and (ii), we observed that the condensation of benzaldehyde and aniline to imine proceeded much faster (the control experiment (iii)). The conversion of benzaldehyde reached >90% in 1\u00a0h. The initial conversion rate was about 28.9\u00a0mmol\u00b7g\u22121\u00b7h\u22121. With prolonging the reaction time, the conversion of was stabilized at about 92%, suggesting a reaction equilibrium was reached. The imine selectivity was 100% throughout the reaction time in this experiment.Based on the above results, a reaction scheme with competing pathways has been proposed, which is illustrated in Scheme 2\n. In the oxidative coupling of benzyl alcohol and aniline over the 3% V-OMS-2(1) catalyst, imine is the main product (92% selectivity), whereas benzaldehyde and azobenzene are the by-products, the selectivity of which was 6% and 2%, respectively. The initial conversion rate for the condensation of benzaldehyde and aniline reaction (\u223c 28.9\u00a0mmol\u00b7g\u22121\u00b7h\u22121, Step 2 in the\nScheme 1) is >5 times faster than that of the benzyl alcohol oxidation reaction (\u223c 5.7\u00a0mmol\u00b7g\u22121\u00b7h\u22121, Step 1 in the\nScheme 1), and is >14 times faster than that of aniline self-coupling reaction. The results clearly display that the oxidation of benzyl alcohol to benzaldehyde, i.e., the Step 1 in the\nScheme 1 is the rate-determining step to the formation of imine.Since both the initial conversion rates in the oxidation of benzyl alcohol and the condensation of aldehyde with aniline reactions are much greater than that of aniline self-coupling reaction, it is reasonable that only trace amount of azobenzene was detected in the oxidative coupling of benzyl alcohol with aniline. It should also be noted that we did not observe any evidence related to disproportionation.As shown in Scheme 1, benzaldehyde is the primary product (step 1), which then quickly reacts with aniline to form imine via a fast condensation reaction (step 2). Therefore, appropriately increasing the ratio of aniline: BA at the beginning of the reaction can change the reaction equilibrium and promote the synthesis of imine. As confirmed by the results in Table 5\n(entry 4\u20136), the imine yield increased from 85% to 99% when the aniline:BA ratio increased from 1:1 to 2:1. The azobenzene yield changed little as the reaction rate of aniline self-coupling is much slower, requiring longer time to reach the equilibrium. Reaction temperature can largely affect the reaction rate. As a result, a low reaction temperature normally requires a longer reaction time to achieve the same level of imine yield (Table 5\n, entry 2\u20134).To further understand the role of the surface oxygen species and air atmosphere, we have performed additional experiment on the oxidative coupling of benzyl alcohol with aniline over 3%V-OMS-2(1) catalyst under an N2 atmosphere. The results are shown in Fig. 16\n. The quick conversion of benzyl alcohol and the production of imine was observed for first hour of reaction. Then, the conversion of benzyl alcohol significantly slowed down and stabilized at about 45% after 3\u00a0h of reaction. The oxidation process can be speeded up again by replacing N2 with air. The benzyl alcohol conversion rapidly increased to 81% within 1\u00a0h by introducing air. The results suggest that surface oxygen species play an important role in the rate-determining step of the imine synthesis as the oxidation occurs in the Step 1. The surface-activated oxygen species on the 3%V-OMS-2(1) can oxidize benzyl alcohol to benzaldehyde under an inert atmosphere until complete consumption. The presence of molecular oxygen in air can interact with the defective sites of 3%V-OMS-2(1) to re-generate surface-activated oxygen species which keeps the oxidation reaction ongoing continuously. It proves that the regenerable surface-activated oxygen species on 3%V-OMS-2(1) are the active sites for the oxidation coupling of benzyl alcohol with aniline and the acid sites are responsible for the imine formation via benzaldehyde-aniline condensation, which is basically consistent with other manganese-based catalyst reported in literature [17].After reaction, the used 3%V-OMS-2(1) catalyst was recovered by filtration and washed sequentially with a small amount of dichloroethane and ethyl acetate for two times in turn, followed by drying overnight in a vacuum oven at 50\u00a0\u00b0C. After being treated at 250\u00a0\u00b0C in air for 2\u00a0h again, the regenerated catalyst was then used in the next run. As shown in Fig. 17\nA, no obvious change in the benzyl alcohol conversion and imine yield was observed in the repeated four runs. Furthermore, the following experiment was conducted to verify whether there was a V and Mn leaching during the reaction process: the solid catalyst was removed by filtration after 1\u00a0h of reaction and then kept the filtrate to continue for another 3\u00a0h under the same reaction conditions. As shown in Fig. 17B, the reaction halted right away after the catalyst was removed. No appreciable loss of V and Mn was detected in the solution by ICP-MS.The spent catalyst was analyzed by N2 physisorption. As shown in Fig. S6, the fresh and spent catalysts possessed almost identical N2 adsorption-desorption isotherm and hysteresis loop, indicating that the pore structure of the catalyst was not affected during the reaction. As compared with the fresh catalyst, only a slight decrease in the specific surface area of the spent catalyst was observed, whereas the pore volume and the pore size were almost the same (Table 1). The decrease in the surface area may be caused by the adsorption of some products or/and reactants on the catalyst surface. The XRD patterns shown in Fig. S7 suggest that the characteristic crystal structure of the fresh catalyst was well preserved after reaction. Mn 2p, V 2p and O 1\u00a0s XPS spectra of the fresh and spent catalysts are shown in Fig. S8. The similar spectral shapes and binding energies suggest that the valence and bond states of Mn, V and O did not change after the reaction. Moreover, the molar ratios of various elements on the surface of the spent catalyst remained the same as compared with the fresh catalyst (see \nTables 1 and 2\n). The above results confirm that the 3%V-OMS-2(1) catalyst is highly stable and recyclable for the reaction, which makes it promising for practical application in imine synthesis from oxidative coupling of alcohols-amines with air.The effect of vanadium doping amount on the structure and catalytic activity of the V-OMS-2 catalyst for imine synthesis from oxidative coupling of benzyl alcohol and aniline was studied in this work. Doping proper amount of vanadium not only improves the surface area and acid sites, but also increases the amount of Mn3+ component, the mobility and amount of active oxygen species on the OMS-2 surface. As a result, the catalytic activity for imine synthesis from benzyl alcohol and aniline was greatly enhanced. The high specific surface area is found to be the key contributor for the high catalytic activity of 3%V-OMS-2(1). Additionally, the crystallinity of OMS-2 may also play an important role in the imine synthesis. Doping large amount of vanadium (> 6 mol% of V/Mn) hindered the formation of OMS-2 crystal phase, leading to a drop in the catalytic activity for the imine synthesis.Furthermore, the influence of vanadium precursor on the structure and catalytic performance of 3%V-OMS-2 catalyst was also investigated. Compared to sodium metavanadate, vanadium pentoxide was a better precursor for the 3%V-OMS-2 catalyst, which exhibited much higher catalytic activity for the reaction. The enhancement could be attributed to its larger surface area, the substantially increased acid sites, unique mesoporous structure and increased active surface oxygen species.The developed 3%V-OMS-2(1) catalyst exhibited exceptional catalytic performance for imine synthesis. The high imine yield of 92% was achieved at the 99% of benzyl alcohol conversion with 4\u00a0h of reaction. The highest activity was 2.83\u00a0mmol\u00b7g\u22121\u00b7h\u22121, much better than those reported in literature. In addition, the catalyst showed high stability and good recyclability for the reaction. Thus, it is promising for practical application in the imine synthesis via air oxidative coupling of benzyl alcohol and aniline.Xiaohui Guo contributed to data curation and writing - original draft;Mengke Li contributed significantly to investigation;Peizheng Zhao contributed to methodology and resources;Xiaoxing Wang contributed to formal analysis and writing - review & editing;Qinghu Tang contributed to the conceptualization, supervision, project administration, funding acquisition and visualization.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The financial support from the key research project funded by the Department of Education of Henan Province (19A150030) is greatly acknowledged.\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106540.", "descript": "\n                  A series of vanadium doped cryptomelane-type manganese oxide (V-OMS-2) catalysts were prepared by a simple, low-cost reflux method, and investigated for one-pot imine synthesis from oxidative coupling of benzyl alcohol and aniline with air. The physicochemical properties of the V-OMS-2 catalysts were characterized by various techniques including XRD, BET, SEM, TEM, XPS, H2-TPR and NH3-TPD. It was found that the surface area, Lewis acid sites, the amount of Mn3+ component and active surface oxygen species were much improved with vanadium doping. Consequently, the activity of V-OMS-2 catalyst for oxidative coupling of benzyl alcohol and aniline to imine was enhanced. The highest conversion and the imine yield were obtained over the 3\u00a0mol%\u00a0V-OMS-2 catalyst, being \u223c99% and 92%, respectively. Higher vanadium doping (\u2265 6\u00a0mol%), however, hindered the preservation of OMS-2 crystal structure, leading to a drop in the catalytic performance. The high specific surface area was suggested to be the key contributor to the high catalytic activity of 3% V-OMS-2(1) catalyst. Among the vanadium precursors studied, the catalyst prepared with vanadium pentoxide exhibited a much higher catalytic activity, which can be attributed to its larger surface area, unique mesoporous structure, increased Lewis acid sites and more readily available surface oxygen species. In addition, the stability and recyclability of the catalyst were also studied, and the reaction mechanism was discussed.\n               "}
{"full_text": "In the last two decades, Metal-Organic Frameworks (MOFs) have revolutionized the application fields of nanoporous materials in both academic and even industrial contexts [1,2]. In the case of heterogeneous catalysis, the discovery of MOFs ended (at least, potentially) the limitations of having metal centers of any nature, in any proportion and in practically any chemical environment within porous networks, closing the gap between homogeneous and heterogeneous catalysis [3\u20135], and found correlation between their structural features and their catalytic performance [6,7]. The other great limitation of the porous solid catalysts versus homogeneous catalysts is the reactant and products diffusion [8,9]. One of the most successful strategies to combat the diffusion problems is a drastic reduction of crystal size [9,10].For so many reasons, MOF-74 is one of the most widely studied MOF materials. From a catalytic point of view, the most interesting properties of this material are the existence of open metal sites, its versatility in metal composition [11\u201314] and the possibility of being prepared with the smallest crystal size ever described for a porous material [10]. Fig.\u00a0S1 shows the perpendicular view of the hexagonal shaped pores of Zn- and Cu-MOF-74 materials having a diameter of approximately 1\u00a0\u200bnm. The nanocrystalline form of the MOF-74 materials have shown to have much higher catalytic performance than their micron-sized counterparts prepared by conventional solvothermal methods [15,16]. The synthesis methodology of such nanocrystalline M-MOF-74 materials is an important advance with respect to the conventional ones in terms of economic and energetic sustainability, as it is prepared practically instantaneously, at room temperature and with high atomic economy (high yield and the metal/linker ratio coinciding with the stoichiometry found in MOF-74) [10,17]. However, the \u2018Achilles heel\u2019 of such methodology is the nature of the solvent, which is the non-volatile N,N-dimethylformamide (DMF), and which should be removed/exchanged by tedious washing procedures. Other attempts to prepare M-MOF-74 at room temperature in a solvent as sustainable as water (although in the presence of an stoichiometric amount of a base for deprotonating the organic ligand) led either to non-nanocrystalline MOF-74 [18] or to the impossibility of preparing some M-MOF-74 material such as Cu probably because of trend of Cu(II) to form Cu(OH)2 phases, even at moderate pHs [19]. In this context, more sustainable preparation methodologies of nanocrystalline M-MOF-74 materials are demanded. Fortunately, MOF-74 is also very versatile in its preparation media. For instance, it is not so common that a give MOF material could be synthesized in four different solvents: DMF, water, THF [20,21] and methanol [21,22].The use of methanol as the unique solvent for the preparation of MOF-74 at room temperature has been described only for the Cu-based material [22], whereas it has been used as a co-solvent (only for metal source) in the preparation of different M-MOF-74 at very low temperature (\u221278\u00a0\u200b\u00b0C) [21]. Compared with DMF, the most conventional solvent in the synthesis of MOF-74 [11,23,24], including at room temperature [10,15,17], methanol possesses some very attractive sustainable properties such as much lower boiling point, much lower price, much higher availability, etc. Compared with water, which is the reference solvent in terms of sustainability, methanol is more volatile (easier to be activated) and, more importantly, MOF-74 can be prepared in methanol without adding any chemical specie beyond the essential metal and linker sources, whereas the assistance of a base acting as a deprotonating agent is compulsory in the synthesis in aqueous solution. Not less, methanol is the universal solvent used in the washing protocols, which normally lasts six days, making washing procedure very much unsustainable than the synthesis procedures themselves. The preparation of MOF-74 in methanol would minimize the effort in the washing/activation protocols. In this work, attempts of preparing M-MOF-74 (M\u00a0\u200b=\u00a0\u200bMg, Mn, Co, Ni, Cu, Zn and Cd) materials at room temperature in methanol are described. To make the procedure ever more sustainable, a metal/linker ratio of 2:1, which is the MOF-74 stoichiometry, was used instead of the conventional 2.6:1 ratio. The successfully-prepared Cu-, Co- and Zn-based MOF-74 materials (at room temperature, after really short times, with high yields and with small crystal size) were fully characterized and catalytically tested in the styrene oxidation to benzaldehyde.Benzaldehyde is one of the most industrially demanded aromatic aldehydes, as it is used in so many applications such as flavoring agent in the food industry, reagent for the pharmaceutical industry or an intermediate for the production of perfumes and dyes or industrial solvent [25]. It is commonly obtained as a byproduct of the oxidation of toluene in the synthesis of benzoic acid or by hydrolysis of benzylidene chloride [26]. Some recent works have focused on the production of benzaldehyde by the oxidation of the olefin styrene with different peroxide-based oxidants like H2O2 [27] o tert-butylhydroperoxide (TBHP) [28]. This reaction gives superior yields to benzaldehyde and has the extra key advantage of being heterogeneously catalyzed, which implies ease for recovery, reactivating and reusing the catalysts. Some MOF-74 materials, such as the Cu- and the Co-based ones, have been used to catalyze this reaction with O2 as the oxidant, with selectivities of 100 and 35% for benzaldehyde and conversions of 0.6 and 47% respectively [29]. On the other hand, Mn-MOF-74 gave a conversion of 95% and selectivity to benzaldehyde of 55% using TBHP as an oxidant [28]. Therefore, the influence of the nature of the metal on the catalytic activity seems to be evident, whereas it is expected that other variables such as reaction time and temperature, type and amount of oxidant, type of solvent and amount of catalyst have also marked influence. In this work, we have tested the catalysts Co-, Cu- and Zn-MOF-74 prepared at room temperature in methanol in this reaction. After optimizing different reaction conditions, the activity of Cu-MOF-74 surpassed that of their counterparts, with notable styrene conversion (57%) and selectivity to benzaldehyde (65.4%). Moreover, unlike the structural transformation of the catalyst Cu-MOF-74 described in some other Fine Chemistry reactions [15,22,30], diffraction techniques indicate that the structure is preserved after reaction.The attempts of preparing M-MOF-74 (M\u00a0\u200b=\u00a0\u200bCd, Cu, Co, Mn, Mg, Ni and Zn) materials were carried out following a methodology already described for Cu-MOF-74 [22] at room temperature, in methanol (MeOH) as the unique solvent, and a metal/2,5-dihydroxyterphthalic acid (H4dhtp) ratio of 2, which is the stoichiometry found in the final structure of MOF-74. A solution of H4dhtp (0.20\u00a0\u200bg, 1.0\u00a0\u200bmmol)) in MeOH (6.67\u00a0\u200bg, 208\u00a0\u200bmmol) was added dropwise over a solution of M(CH3COOH)2\u00b7xH2O (2.0\u00a0\u200bmmol) in MeOH (3.33\u00a0\u200bg, 104\u00a0\u200bmmol) during a period of 10\u00a0\u200bmin with constant agitation at room temperature (23\u00a0\u200b\u00b0C). Such mixture of solutions provokes the immediate appearance of a precipitate. The agitation was kept for 20\u00a0\u200bh, at room temperature, after which the crystalline solid was recovered by centrifugation and washed with 10\u00a0\u200bmL of MeOH five times. The solid was kept immersed in 10\u00a0\u200bmL of MeOH for 6 days and exchanged for the same amount of fresh MeOH three times.Powder X-ray diffraction (PXRD) data were collected under ambient conditions on Bruker D8 Discover diffractometer using Cu K\u03b11 (\u03bb\u00a0\u200b=\u00a0\u200b1.5406\u00a0\u200b\u00c5) and with Bragg-Brentano configuration that operates at a voltage of 40\u00a0\u200bkV and a current of 40\u00a0\u200bmA. It has a fast multichannel LYNXEYE XE-T detector that discriminates fluorescence. Attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectra were recorded using a Thermo Nicolet Nexus 670 FTIR spectrometer equipped with a SensIR Technologies Duras-amplIR horizontal ATR accessory and a liquid nitrogen-cooled MCT detector. Thermogravimetric analysis (TGA) was registered from 25 to 900\u00a0\u200b\u00b0C with a heating rate of 20\u00a0\u200b\u00b0C/min under air flow using a Perkin-ElmerTGA7 instrument. Morphology studies were carried out in an ultrahigh resolution Philips XL 30 SEM instrument with a tungsten filament. N2 adsorption/desorption isotherms were performed on a Micromeritics ASAP 2420 device at \u2212196\u00a0\u200b\u00b0C; previously, the samples were activated at 150\u00a0\u200b\u00b0C for 18\u00a0\u200bh under high vacuum. The surface area was estimated by applying the BET method to the experimental adsorption points registered at low p/p0. Micropore size distributions were estimated applying Hovarth-Kawazoe (cylinder geometry) method to a different N2 isotherm, being registered only the adsorption branch in a Micromeritics ASAP 2020 device a \u2212196\u00a0\u200b\u00b0C from very low p/p0 (10\u22127) up to p/p0\u00a0\u200b=\u00a0\u200b0.1.Before initiating the catalytic experiment, the M-MOF-74 material was activated at 150\u00a0\u200b\u00b0C and 10\u22123\u00a0\u200bbar for 1\u00a0\u200bh. The styrene oxidation reactions were carried out in a batch glass reactor of 50\u00a0\u200bmL, under atmospheric pressure and continuous stirring. The reagents were added in the following order: activated MOF-74 (100\u00a0\u200bmg), acetonitrile (10\u00a0\u200bmL), styrene (5\u00a0\u200bmmol, 0.52\u00a0\u200bg) and finally oxidant (10\u00a0\u200bmmol of tert-butylhydroperoxide, TBHP, 5.5\u00a0\u200bM in n-decane). Different aliquots were taken from the reaction media at different reaction times (between 0 and 4\u00a0\u200bh). Finally, the reaction mixture was centrifuged to recover the catalyst, which was dried and analyzed by PXRD. For specific catalytic tests, all of them carried out with Cu-MOF-74 as catalyst, H2O2 (50\u00a0\u200bwt% in aqueous solution) was used as the oxidant, TBHP/styrene molar ratio was varied to 1:1 and 3:1, the amount of catalyst was 0\u00a0\u200bmg, 50 or 150\u00a0\u200bmg, and the reaction temperature was modified to 45 or 82\u00a0\u200b\u00b0C.The catalyst-free aliquots were analyzed by gas chromatography (GC) using an Agilent 6890 HP instrument with a Carbowax column (25\u00a0\u200bm\u00a0\u200b\u00d7\u00a0\u200b0.25\u00a0\u200bmm and 0.25\u00a0\u200b\u03bcm), He flow of 2.5\u00a0\u200bmL/min, FID detector at 250\u00a0\u200b\u00b0C and heating ramps of 8\u00a0\u200b\u00b0C/min from 40 to 180\u00a0\u200b\u00b0C and of 10\u00a0\u200b\u00b0C/min from 180 to 210\u00a0\u200b\u00b0C.The mixture of the metal acetate solution and the H4dhtp solution, both in methanol, led to the immediate formation of a solid. The XRD patterns of the resultant solids after maintained the mixture under magnetic stirring overnight (20\u00a0\u200bh) are plotted in Fig.\u00a01\n. The diffractograms of the samples having MOF-74-like structure are plotted on Fig.\u00a01-left, whereas Fig.\u00a01-right shows the XRD patterns of the samples with either impure MOF-74 or those samples that do not contain MOF-74 phase. The XRD pattern of the Zn-based sample matches well with that generated from a Zn-MOF-74 cif file. The XRD pattern of the sample Cu-MOF-74 is substantial different to these of the other M-MOF-74 due to the structural differences arisen from the notable Jahn-Teller effect in the octahedral coordination of the Cu(II) within the framework [31,32]. Such Jahn-Teller effect, which is evident in Fig.\u00a0S1, is a geometric distortion of a non-linear molecular system that reduces its symmetry and energy, being typical of octahedral coordination, particularly in Cu environments. The matching of the XRD pattern of our Cu-MOF-74 with that of the theoretical Cu-MOF-74 (Fig.\u00a0S2) is also very good. The diffractogram of Co-MOF-74 scarcely contains a few reflections, which are very broad and very little intense, but they also match well with the pattern of the simulated Co-MOF-74 (Fig.\u00a0S2). The reason behind such so \u2018poor\u2019 diffraction is simply the extremely nanocrystalline nature of this sample and it is not necessarily related to low quality of the sample, in good agreement with what has been described for this material prepared in DMF at room temperature [10]. The nanocrystallinity degree of these MOF-74 samples prepared at room temperature seems to depend on the nature of the metal, as the same order in crystal/domain size (Co\u00a0\u200b<\u00a0\u200bCu\u00a0\u200b<\u00a0\u200bZn) was found either in DMF [10,15] or in methanol (this work).The Mg-, Mn-, Ni- and Cd-dhtp samples, whose XRD patterns are shown in Fig.\u00a01-right, were formed by other phases. Mg-dhtp phase is amorphous, Mn-dhtp is a non-identified poor crystalline phase and Cd-dhtp is an unknown crystalline phase. That is why these samples were discarded for further characterization and for testing them as catalysts. Only Ni-dhtp sample contains an important proportion of MOF-74 phase, perhaps it is even pure. This Ni-based sample was characterized by FTIR spectroscopy, TGA and N2 adsorption/desorption isotherm (non-shown) under the same conditions than these used for Co-, Cu- and Zn-MOF-74 (see below), and all the achieved results agrees the possibility of such sample could be a very nanocrystalline (pure) Ni-MOF-74. Nevertheless, because of the very low resolution of its XRD pattern together with the presence of some reflections of doubtful origin (but they could be perfectly due to a very ordered coordination of methanol to the open metal sites), we have decided not to go further with this sample in order to avoid, for instance, any misinterpretation of its catalytic behavior.Some attempts to obtain the formation of Cu-MOF-74 were also carried out in other common alcohols, particularly ethanol and isopropanol, but the phase MOF-74 was not detected. All remaining characterization is only shown for Co-, Cu- and Zn-MOF-74 samples, which were also the only ones tested in the catalytic oxidation of styrene to benzaldehyde.\nFig.\u00a02\n shows the 700-1000\u00a0\u200bcm\u22121 region of the FTIR spectra of these three samples compared with that of the organic linker in its acidic form H4dhtp. Such region is quite sensitive to the conformational and/or local environment of organic molecules, in such a way that it could be considered as a fingerprint region. Because of the different environment of the organic linker dhtp when it is tetraprotonated (H4dhtp, 2,5-dihydroxyterephthalic acid) and when it is forming the MOF material (anion dhtp4\u2212), the corresponding FTIR spectra are radically different. Moreover, the FTIR spectra of the samples Zn-MOF-74, which was unequivocally identified as MOF-74 by XRD pattern (Fig.\u00a01), and Co-MOF-74, whose structural identification by diffraction techniques generates reasonable doubts (Fig.\u00a01), have practically the same pattern, confirming that both samples have the same local linker environment; in other words, both have the same short-range structure. The only real difference between these two spectra is the broadening. Co-MOF-74 is formed by so small crystal size that affects the IR peak width, in spite of infrared spectroscopy provides information at relatively short range [33]. On the other hand, the FTIR spectrum of the sample Cu-MOF-74 somehow reminds these of the Zn- and Co-MOF-74 samples but, at the same time, it possesses marked differences. Such features must be due to the structural differences between them, as a consequence of the marked Jahn-Teller effect found in Cu-MOF-74 [31,32], which also makes completely different the XRD patterns beyond the two most intense and lowest angle reflections (Fig.\u00a01 and S2).Thermogravimetric analysis (TGA) profiles of the M-MOF-74 prepared in methanol are shown in Fig.\u00a03\n. The first weight loss is due to methanol (which was both synthesis and washing solvent) within the pores. The main weight loss, which is attributed to removal/combustion of the linker dhtp, is directly related to the MOF decomposition. According to such assignment, the order of the thermal stability of the three MOFs is Cu\u00a0\u200b<\u00a0\u200bCo\u00a0\u200b<\u00a0\u200bZn, which is in good agreement with the literature [10,34]. There is also good agreement on the temperature values at which the different M-MOF-74 decomposition takes places as well as on the shape of the TGA curves and for the linker/residual weight ratio. The thermal stability of any of these samples (decomposition of the less stable Cu-MOF-74 starts above 225\u00a0\u200b\u00b0C under air flow) should in principle be enough for being used as catalysts in the conversion of styrene to benzaldehyde under the reaction conditions of this work (below 82\u00a0\u200b\u00b0C).One of the most valued physicochemical properties of nanoporous catalysts is their textural properties. Fig.\u00a04\n compares the N2 adsorption/desorption isotherms of these sustainable M-MOF-74 materials. All of them possess outstanding (micro)porosity, reaching BET surface areas of the same order to these published in the literature for high-quality MOF-74. In addition, it is well known that the nanocrystallinity entails a decrease of the microporosity in porous materials [10,35]. In particular, the estimated BET surface area was 702, 925 and 1013 m2g-1 for Zn-, Co- and Cu-MOF-74, respectively. For some unknown reasons, it is quite common that Zn-MOF-74 material has lower surface area compared to its counterparts based on other divalent metal ions of similar atomic weight and prepared under the same experimental conditions [10,11]. That is also the case of this series of samples. Furthermore, Cu- and especially Co-MOF-74 samples, but not Zn-based one, have certain mesoporosity, as evidenced by the presence of notable hysteresis loops in their isotherms. Such mesoporosity must be of intercrystalline nature, which is in good agreement with other nanocrystalline M-MOF-74 materials prepared in DMF also at room temperature [10,15]. The so small crystals and/or crystalline domains forming these two samples are unstable in an isolated form, and then they are aggregated (rather than agglomerated) in very consistent samples (see SEM images in Fig.\u00a05\n), leaving meso-holes of relatively homogeneous size, which provokes the appearance of the hysteresis loops in the mesoporous region of their isotherms (Fig.\u00a04). Supporting this interpretation, the order of the amount of mesoporosity (Co\u00a0\u200b>\u00a0\u200bCu\u00a0\u200b>\u00a0\u200bZn) follows the inverse order to that found for crystal size (Zn\u00a0\u200b>\u00a0\u200bCu\u00a0\u200b>\u00a0\u200bCo). Another N2 isotherm of the sample Co-MOF-74 (and also of Zn-MOF-74 for comparison purposes) was registered at low range p/p0 (10\u22127 \u2013 10\u22121) in order to further support/deny the microporous nature of this sample that is so hard to structurally characterize by diffraction techniques (Fig.\u00a01 and S1). Fig.\u00a0S3 shows such isotherms and makes clear that the Co-dhtp is undoubtedly of microporous nature and that its pore diameter (centered at ca. 9.3\u00a0\u200b\u00c5) is quite close to that expected. The relatively large width of this peak must be again attributed to their very small crystalline domains.Representative SEM images of the three M-MOF-74 are shown in Fig.\u00a05. It is obvious that the observed particles in the samples Co- and Cu-MOF-74, whose size is well below micrometer scale, are composed by agglomerates/aggregates formed by a large number of nanocrystals. This is in good agreement with samples similarly prepared in DMF as solvent [10,15], as suggested by the large broadening of the XRD reflections (Fig.\u00a01 and S1) and even of the FTIR bands (Fig.\u00a02), and by the hysteresis loops in the mesopore region of the N2 isotherms (Fig.\u00a04). In contrast, the Zn-MOF-74 material is formed by isolated needle-like crystals, with length of a few micrometers. The morphology of the sample Zn-MOF-74 is associated to two characterization features seen previously: (i) the relatively sharp reflections found in the XRD pattern of this sample (Fig.\u00a01); and (ii) the absence of any adsorption in the mesoporosity region, leading to a type-I N2 isotherm at \u2212196\u00a0\u200b\u00b0C (Fig.\u00a04), since these isolated crystals cannot generate any intercrystalline porosity. This singularity of Zn-MOF-74 somehow is also related to a couple of precedents. On the one hand, large crystals of Zn-MOF-74 are formed in the synthesis at room temperature in water using NaOH as deprotonating agent [18]. On the other hand, in spite of the samples Zn-MOF-74 prepared in DMF at room temperature forms agglomerates like the rest of M-MOF-74 samples, it is formed by the largest crystals of series M-MOF-74 and its crystals are scarcely fused, unlike for instance the sample Co-MOF-74, in which the nanocrystalline domains are completely fused in very large particles [10].Only the three samples undoubtedly formed by MOF-74 phase and whose characterization have been described in detailed above, that is, Co-, Cu- and Zn-MOF-74 materials, were catalytically tested. The formation of benzaldehyde from styrene requires an oxidant and ideally catalytic redox centers, which are provided by the catalyst. (A scheme in the Supplementary Information -Scheme S1- shows the most accepted mechanism on the synthesis of benzaldehyde through the styrene oxidation using TBHP as an oxidant and MOF-74 or relative materials as catalysts [27,28,36]). Therefore, it is expected that the Zn-based MOF-74 sample is not active in this process, as Zn does not have redox nature. In any case, it could be used as a kind of blank experiment.Given the interest of this work in the catalytic synthesis of benzaldehyde, results and discussion of the catalytic performance will focus on the yield to benzaldehyde rather than in styrene conversion or selectivity. Before studying and comparing the catalytic activity of the three M-MOF-74 in the synthesis of benzaldehyde, the reaction conditions were optimized using Cu-MOF-74, which was selected based on its good catalytic performance in different oxidation reactions of organic compounds [15]. Fig.\u00a06\n shows the kinetics of yield to benzaldehyde under different reaction conditions. (The detailed data is given in Table\u00a0S1). Thus, the nature of the oxidant (Fig.\u00a06A) has marked influence on the yield even though the comparison was carried out between two very similar oxidant species, hydrogen peroxide H2O2 and tert-butylhydroperoxide (TBHP). Using the latter as the oxidant, a notable yield of 37.3% was achieved, which is five times higher than the yield reached when H2O2 was the oxidant. Since both species possess the same oxidant group (peroxide), we believe that the reason behind so dramatic difference in catalytic performance must be related to the \u2018solvent\u2019 of the peroxide sources, in particular, to the presence/absence of water. These peroxides must be stabilized with a solvent; otherwise they spontaneously react. The used TBHP in this work is basically water-free as it is diluted in n-decane whereas H2O2 is diluted in water (50\u00a0\u200bwt % H2O2). Similar behavior has been found in the oxidation of other olefin, cyclohexene, with the same family of catalysts (but prepared in DMF at room temperature) [15]. Polar H2O molecules (and presumably not these of n-decane) could be serious competitor of reactants (styrene and H2O2) to be coordinated to the Lewis acid open metal sites of the MOF-74. The negative influence of water in the reaction media of oxidation reactions has been also made clear in the oxidation of olefins catalyzed by Ti-containing nanoporous materials [37,38].Once optimized the nature of the oxidant, its content in the reaction media was also studied (Fig.\u00a06B). A higher concentration of TBHP does not necessarily produce higher yield to benzaldehyde. A TBHP/styrene molar ratio of 2 improves the yield to benzaldehyde versus a ratio of 1, but the yield decreases when the ratio is increased to 3. In spite of the oxidation mechanism of this reaction could be not completely clear, there is a general agreement about the need of both reactants to reach a given open metal site in order to the reaction occurs; therefore, it could make sense that a disproportionate excess of one of the reactants could be counterproductive for catalytic performance purposes.\nFig.\u00a06C evidences notable influence of the amount of catalysts on the yield to benzaldehyde. Increasing the amount of the catalyst is beneficial to some extent (for instance, in the catalyst content range from 0 to 100\u00a0\u200bmg). However, further increase induces the contrary effect (for instance, from 100\u00a0\u200bmg to 150\u00a0\u200bmg). It must be highlighted the different shape of the kinetics curves in the presence and in the absence of any catalyst. When Cu-MOF-74 is present in the reaction media, practically all benzaldehyde is formed at the very beginning of the reaction (only during the first 30\u00a0\u200bmin), whereas benzaldehyde production continues growing for at least 2\u00a0\u200bh in the blank experiment. In other words, Cu-MOF-74 indeed accelerates the formation of benzaldehyde until it somehow becomes inactive.The last optimized parameter was the reaction temperature (Fig.\u00a06D). Once again, an increase of this parameter is initially favorable, in such a way that the yield to benzaldehyde is significantly enhanced at 75\u00a0\u200b\u00b0C in comparison with that achieved at 45\u00a0\u200b\u00b0C. However, when the reaction temperature was increased further, up to 82\u00a0\u200b\u00b0C, the maximum temperature as it is the boiling point of the solvent acetonitrile, the yield decreases, due to the selectivity to other undesired products in this work (such as styrene oxide or phenylacetaldehyde) is favored.Once some of the reaction parameters were optimized (Fig.\u00a06), the three M-MOF-74 and a blank experiment (with no catalyst) were compared under such optimized conditions (TBHP as oxidant, in a molar ratio of 2 with respect to styrene, with 100\u00a0\u200bmg of catalysts and at 75\u00a0\u200b\u00b0C as reaction temperature) (Fig.\u00a07\n). The most active catalyst is in the synthesis of benzaldehyde is Cu-MOF-74. Indeed, after 4\u00a0\u200bh, which is the longest tested reaction time, it is the only catalyst that surpasses the blank experiment in the reached yield to benzaldehyde. As expected, the Zn-MOF-74 catalyst, which is free of any redox center, acts as an inhibitor rather than as a catalyst [15]. On the other hand, although Co-MOF-74 could be also considered as an inhibitor because its yield to benzaldehyde is lower than that given by the blank experiment, this material, unlike Zn-MOF-74, shows indications of its real catalytic role. At reaction time shorter than 30\u00a0\u200bmin, its yield to benzaldehyde is significantly higher than the yield detected in the experiment with no catalyst. It is important to note that the reaction conditions have been optimized for Cu-MOF-74, so it should not be ruled out that Co-MOF-74 give higher yield to benzaldehyde than the blank under its own optimized conditions. In any case, it seems clear that Cu-MOF-74 is the best MOF-74-based catalyst for this reaction, just like it was found in the related reaction oxidation of cyclohexene with TBHP.In order to value the quality of these results, Fig.\u00a08\n compares the yield to benzaldehyde obtained in this work with these achieved by other M-MOF-74 published elsewhere [28,29]. After analyzing the outstanding influence of certain reaction parameters in the final yield to benzaldehyde, the comparison shown in Fig.\u00a08 must be taken cautiously. The conditions of the reaction published elsewhere were: (i) 80\u00a0\u200b\u00b0C, 20\u00a0\u200bh, no solvent, O2 as the oxidant for Co- and Cu-MOF-74 [29], and (ii) 75\u00a0\u200b\u00b0C, 6\u00a0\u200bh, no solvent, TBHP as the oxidant for Mn-MOF-74, whereas our conditions were: 75\u00a0\u200b\u00b0C, 4\u00a0\u200bh, acetonitrile as solvent, and TBHP as the oxidant for Co- and Cu-MOF-74. Anyway, both MOF-74 materials that can be compared with the literature data, Co- and Cu-based ones, resulted to be more active in our case. It is particularly significant the difference in yield to benzaldehyde in the case of Cu-MOF-74. Even though, the yield given by our Cu-MOF-74 is below that given by Mn-MOF-74. Unfortunately, we could not get Mn-MOF-74 by our sustainable methodology.Probably the most controversial property of MOFs when applied as heterogeneous catalysts is their low both thermal and chemical stability. This is especially true for Cu-MOF-74, the most active catalyst in this work, as it has been reported their structural transformation/decomposition during the redox [15,22] and acidic-catalyzed [30] reactions. Fig.\u00a09\n shows the XRD patterns of the three samples before and after being tested in the synthesis of benzaldehyde from the oxidation of styrene with TBHP. In all cases, the structure of MOF-74 is preserved, which is particularly important in the case of Cu-MOF-74.M-MOF-74 materials (M\u00a0\u200b=\u00a0\u200bCu, Co, Zn and possibly Ni) were successfully prepared at room temperature using methanol as the unique solvent. The method is simple, quick, sustainable, gives high yields, and carried out in methanol as solvent, which is the solvent almost universally used for washing/activating this material. The resultant Cu- and especially Co-MOF-74 materials were so nanocrystalline that the phase MOF-74 by diffraction techniques could not be undoubtedly identified. However, the combination of all characterization techniques used in this work (XRD, FTIR spectroscopy, TGA, N2 isotherms and SEM) certified that these materials are (nanocrystalline) M-MOF-74. Apart from microporosity, Co- and Cu-MOF-74 materials possess intercrystalline mesoporosity given by the aggregation of nanocrystals in relatively large particles, whereas Zn-MOF-74 is formed by needle-like isolated crystals and lacks any significant mesoporosity. These three M-MOF-74 were catalytically tested in the oxidation of styrene to benzaldehyde using peroxides as oxidants. When the oxidant was H2O2 in aqueous solution, there was practically no reaction probably because of water interference, but the reaction took place when TBHP was used as the oxidant. Under the optimized conditions (amount of catalyst, TBHP/styrene ratio or reaction temperature), the highest catalytic activity was given by Cu-MOF-74 whereas Zn-MOF-74 behaved as an inhibitor rather than as a real catalyst. The yield to benzaldehyde catalyzed by the sustainable Cu-MOF-74 is competitive with the best M-MOF-74 catalysts reported so far for this reaction, and, not less, this catalyst preserved intact their structure during the reaction process.This work has been partially financed by the Spanish State Research Agency, the European Regional Development Fund (FEDER) through the Project MAT2016-77496-R (AEI/FEDER, UE), 2019AEP076 (CVCSIC-AEPP-Ayudas Extraordinarias para preparaci\u00f3n de proyectos 2019), PAPIIT UNAM Mexico (IN101517) and CONACyT (1789) projects. This work has been also financed by the CONACyT project A1-5-30646.\nJ. Gabriel Flores: Formal analysis, carried out most of the experimental work. In addition, he contributed to the analysis, interpretation of the results, and to preparation of the manuscript. Manuel D\u00edaz-Garc\u00eda: contributed to part of the experimental work, in particular, in the synthesis of materials. Ilich A. Ibarra: took part in the design of the research. This work has been partially financing by his projects. Julia Aguilar-Pliego: took part in the design of the research and coordinated the catalytic part. This work has been partially financing by her projects. Manuel S\u00e1nchez-S\u00e1nchez: Formal analysis, Writing \u2013 original draft, designed and coordinated the project, contributed to the analysis of the results, and wrote and submitted the manuscript. This work has been partially financing by his projects.The authors declare no conflict of interest.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.jssc.2021.122151.", "descript": "\n                  Metal-Organic Framework (MOF) materials are promising heterogeneous catalysts in different areas including Fine Chemistry, mainly if they possess open metal sites and if they are nanocrystalline. In this work, a new sustainable methodology to obtain nanocrystalline M-MOF-74 materials at room temperature, without any energy input and in methanol as the unique solvent, is described. Amongst the seven divalent metal tested, pure MOF-74 phase was achieved in the case of Co, Cu and Zn (and maybe Ni), but not in the case of Mg, Mn and Cd. The formation of these MOFs is taken place as soon as the metal source (metal acetate) and the organic linker 2,5-dihydroxyterehpthalic acid are put together. The so-prepared Co- and Cu-MOF-74 (but not Zn-MOF-74) resulted to be nanocrystalline and having high external surface area and intercrystalline mesoporosity. They were tested as heterogeneous catalysts in the synthesis of benzaldehyde from styrene using tert-butylhydroperoxide as an oxidant. Different experimental parameters like reactants ratio, amount of catalysts or reaction temperature were optimized. Cu-MOF-74 gave the highest benzaldehyde yield and interestingly it maintained intact its structure after reaction, unlike the same catalyst used in some other catalytic processes under similar relatively mild conditions.\n               "}
{"full_text": "Heavy metals are natural elements presenting high atomic mass and density (> 5\u00a0g\u00a0cm\u22123). Some of these metals are essential for animals, with indispensable functions for human metabolism [62]. However, several studies indicate that some heavy metals are likely to be carcinogenic (hexavalent chromium, arsenic, cobalt, nickel, antimony, vanadium and mercury), mutagenic (arsenic and vanadium), teratogenic (arsenic), allergenic (nickel) or endocrine-disrupting (silver, copper, zinc and selenium). Low levels of nickel result in reduced growth in intrauterine development, and its deficiency can reduce iron absorption, leading to anemia [51]. The main adverse effects caused by exposure to compounds containing this metal are skin allergies, lung fibrosis and lung cancer, depending on their ability to enter cells [10,98]. Despite its effects, nickel(II) is largely used in the manufacturing process of stainless steel, metallic alloys and batteries [80]. The release of this metal into the environment may occur from various industries, viz., nickel plating, zinc-based casting industry and storage batteries, silver refinery, mining and metallurgy of nickel [5,39].The presence of nickel in drinking water can also occur due to corrosion of pipes containing nickel in their composition or even to the poor removal of this metal by water treatment systems [60]. Regulatory environmental agencies establish concentration limits for nickel, owing to the risks presented by its existence in drinking water and wastewater. For instance, the World Health Organization establishes as a guideline a value of 0.07\u00a0mg\u00a0L\u22121 for the concentration of nickel in drinking waters (WHO/SDE/WSH/07.08/55). The concentration of nickel may range from 0.5\u00a0mg\u00a0L\u22121 to 192\u00a0mg\u00a0L\u22121 in wastewater effluents [53]. Adsorption on several carbon-based adsorbents [80,95,96] has resulted in efficient processes for the removal of Ni. However, many studies report high uptake capacities, since the removal of Ni from waste waters is normally studied considering high loads of the heavy metal (> 50\u00a0mg\u00a0L\u22121) [24,53,59]. The feasibility of the adsorption of nickel on carbonaceous adsorbents should also be explored at low nickel concentrations.An efficient scenario allowing to decrease the costs of the adsorption process and to reach a circular economy approach consists in the development of technologies to valorize wastes by their transformation into suitable adsorbents [25,84,89]. In this sense, the scientific community has been putting a great effort into the development of carbon-based adsorbents from biomass wastes coming from agro-industrial activities, such as fruit peels [25], shell of nuts [42], bagasses [16], among others. By using biomass waste as a carbon precursor, different carbon-based adsorbents can be obtained, viz. pyrochars, hydrochars, or activated carbons, depending on the carbonization processes applied [16]. A pyrochar (PC) is obtained through the thermal treatment of the precursor at 400\u20131000\u00a0\u00baC in an inert or oxygen-limited environment [97]. Hydrochars (HCs) can be prepared by hydrothermal carbonization (HTC), which consists in a thermochemical conversion in the presence of water at temperatures ranging from 150 to 350 \u00baC and autogenous pressure [71]. HTC is interesting because of its technical simplicity, low cost and energy efficiency. Activated carbons (ACs) are typically obtained through two steps: activation and carbonization. Activation can be conducted using chemical (treatment of the precursor with oxidants) or physical (steam, CO2 and air) methods [94]. As an activation step, HTC also works as an efficient process to obtain a suitable precursor (HCs) for the production of ACs [16].The chemical activation to prepare ACs from biomass waste has been studied with different activating agents, such as inorganic acids, bases or salts [2,94]. However, there are scarce studies on HTC of biomass wastes using additives to improve the physicochemical properties of the resultant HCs [16,66,71,85]. The use of chemical agents in HTC can be exploited to introduce improved surface chemistry for adsorption applications of the resultant HCs or ACs. Furthermore, chemical agents in HTC can also act as structure-directing agents to prepare carbonaceous spheres [9]. Among them, iron (III) chloride has proved to be an excellent activating agent for the preparation of carbonaceous materials [3,55,81] and as a metal doping for the adsorption of heavy metals from aqueous solution [11,27,53]. In fact, for carbonaceous adsorbents, the metals and functional groups on their surfaces, with acid or base character, play an important role in the adsorption process [80]. In this sense, HCs are rich in functional groups that can greatly improve chemical reactivity [36]. Because of this, many scientists have been testing HCs as adsorbents for the removal of heavy metals, pesticides, and drug residues [35]. However, the influence of the adsorbent\u2019s characteristics (e.g. functionalities, morphology, or textural properties) on the adsorption of Ni has not been deeply studied so far.The properties of the carbon-based materials not only depend on the type and operating conditions of the carbonization process but also on the carbon precursor selected for their preparation. The materials obtained under the same conditions can present significant differences in their characteristics when other carbon precursors are used [16]. Therefore, the biomass waste used for the preparation of adsorbents should be carefully selected. In this sense, citrus fruit peels have shown to be efficient precursors for preparing carbon-based materials [23,25]. As the precursor contains citric acid, interesting carbon-based materials may be obtained, since citric acid is used as catalyst to develop this type of material [85,87].Citrus fruits are one of the largest fruit crops in the world. Similarly, the citrus industry is also the second largest fruit processing industry, surpassed only by the grape industry, which mainly produces wine [38]. Approximately one-third of the citrus fruits are processed for juice production, resulting in 50\u201360\u00a0% of organic waste, typically constituted by the peel, seeds and leaf residues [75]. It is noteworthy that due to the amount of organic matter present in citrus fruit peels, the disposal of this type of residue directly in the soil can cause damage, given its ability to change the physicochemical characteristics of the soil [79]. Currently, land space occupation and pollution with phenolic compounds due to dumping of waste are becoming problematic [26]. For this reason, the development of techniques to valorize the large amount of waste generated in the citrus juice processing industry is required. In this sense, the production of biochars from diverse citrus peels has become interesting as a low-cost alternative to obtain high-value products, avoiding the pollution of waste dumping [65,78,81,83].This work deals with the preparation of activated carbon, pyrochar and hydrochar materials using tangerine peels as carbon precursor and their assessment in the removal of Ni(II) by adsorption. Hydrochars (HCs) are prepared by HTC assisted with FeCl3, known as a catalyst of carbonization processes [55] and later used as a precursor for the preparation of activated carbons (ACs) by pyrolysis at the same conditions of pyrochar (PC) directly prepared from the tangerine peels. The different properties of the ACs, PC and HCs and how they affect the adsorption of Ni(II) are analyzed, and the kinetic and equilibrium adsorption of Ni(II) on them is modeled. To the best of our knowledge, there is a scarcity of studies dealing with the valorization of tangerine peels, as is the case of other peels, especially considering FeCl3-assisted HTC. Similarly, few studies assess ACs, PCs and HCs prepared from the same source to be applied to the adsorption of a heavy metal at similar operating conditions.The modelling of the adsorption process is invaluable, not only for the prediction of the solute adsorption onto the adsorbent at different operating conditions, but also for a better understanding of the adsorption mechanism occurring on a system [20,28]. Adsorption isotherms data (quantification of adsorbed solute per unit mass of adsorbent at a constant temperature for different solute concentrations in solution at the equilibrium) can be processed for a deep understanding of the interaction between the solute (Ni(II) in this work) and the adsorbent. The constants obtained from the different models provide important information on the affinities of the adsorbent for the removal of the pollutant and on the mechanisms of adsorption. The application of kinetic adsorption is also useful in studying the dynamics of the adsorption mechanism in terms of the order of the adsorption rate constant. Additionally, the parameters obtained as results of the fitting kinetic models allow to assess the time required to remove Ni(II) on the selected adsorbent [49,76].The Langmuir equation is a well-known isotherm model that assumes that adsorption occurs on a homogeneous surface of an adsorbent containing sites that are equally available for adsorption [52]. The separation factor (R\n\nL\n) is an important parameter of the Langmuir isotherm typically used to verify whether the adsorption under study is unfavourable (R\n\nL\n > 1), linear (R\n\nL\n = 1), favourable (0\u00a0< R\n\nL\n < 1) or irreversible (R\n\nL\n = 0). Langmuir equation and R\n\nL\n are expressed by Eqs. (1\u20132),\n\n(1)\n\n\n\n\nq\n\n\ne\n\n\n=\n\n\n\n\nq\n\n\nm\n\n\n\u00b7\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\n1\n+\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\n,\n\n\n\n\n\n\n(2)\n\n\n\n\nR\n\n\nL\n\n\n=\n\n\n1\n\n\n1\n+\nK\n\u00b7\n\n\nC\n\n\n0\n\n\n\n\n,\n\n\n\nwhere q\n\ne\n and C\n\ne\n refer to the solute adsorbed per mass of adsorbent (mg g\u22121) and adsorbate concentration in aqueous media (mg L\u22121) at equilibria stage, q\n\nm\n and K are constants (two-parameter model) measured in mg g\u22121 and L mg\u22121, respectively, R\n\nL\n is the separation factor (dimensionless quantity) and C\n\n0\n is the initial concentration of the adsorbate.Freundlich isotherm is an empirical equation (Eq.(3)) widely applied for heterogeneous systems with interaction between the adsorbate, representing suitably non-asymptotic adsorption curves between uptake capacity (q\n\ne\n) and equilibria concentration (C\n\ne\n) in the aqueous media [69]. The heterogeneity factor (n) can be employed to indicate if the adsorption is linear, chemical or a physical adsorption process (n\u2009=\u20091, n\u2009\n                           <\u2009\n                           1 or n\u2009\n                           >\u2009\n                           1, respectively). This two-parameter model is represented by Eq. (3),\n\n(3)\n\n\n\n\nq\n\n\ne\n\n\n=\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n1\n/\nn\n\n\n,\n\n\n\nwhere K is the constant of Freundlich measured in L1/n mg\u22121/n and n is the exponent.The Sips isotherm model (Eq. (4)) is a combination of the Langmuir and Freundlich isotherms [82]. At high adsorbate concentrations, the equation provides the adsorption capacity in the monolayer, typical of the Langmuir isotherm. At low adsorbate concentrations, the Sips equation is reduced to the Freundlich equation. In the literature, it is possible to find the Sips model named as Koble-Corrigan model [28,77,93], but Koble and Corrigan used the Sips model indeed, as they described [46]. For this reason, the Koble-Corrigan model was not object of study in this work. This three-parameter model is represented by Eq. (4),\n\n(4)\n\n\n\n\nq\n\n\ne\n\n\n=\n\n\n\n\nq\n\n\nm\n\n\n\u00b7\nK\n\u00b7\n\n\n\n\nC\n\n\ne\n\n\n\n\nn\n\n\n\n\n\n\n1\n+\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\nn\n\n\n\n\n,\n\n\n\nwhere q\n\nm\n and K are constants measured in mg g\u22121 and Ln mg-n, respectively, and n is an exponent (three-parameter model).To improve the fitting of Langmuir and Freundlich equations, Redlich and Peterson developed their model [68], which is mathematically equal to the Radke and Prausnitz isotherm model developed in the adsorption of solutes from dilute aqueous solutions on activated carbon [20,67]. Redlich and Peterson isotherm model is typically expressed as Eq. (5),\n\n(5)\n\n\n\n\nq\n\n\ne\n\n\n=\n\n\nA\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\n\n\n1\n+\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\nn\n\n\n\n\n,\n\n\n\nwhere A and K are constants measured in L mg\u22121 and Ln mg-n, respectively, and n is an exponent.The General Isotherm Equation (GIE) proposed by T\u00f3th for all types of isotherms was developed to consider the heterogeneity, and the lateral and vertical interaction energies of the adsorbed molecules. The T\u00f3th isotherm model usually applied in the modelling of adsorption systems for the wastewater treatment is the solution of the GIE when the dynamic equilibrium adsorption is higher for the monolayer than the subsequently formed layers [88] and it is expressed by Eq. (6),\n\n(6)\n\n\n\n\nq\n\n\ne\n\n\n=\n\n\n\n\nq\n\n\nm\n\n\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\n\n\n\n\n1\n/\nK\n+\n\n\nC\n\n\ne\n\n\nn\n\n\n\n\n\n\n1\n/\nn\n\n\n\n\n,\n\n\n\nwhere q\n\nm\n is a constant measured in mg g\u22121, K is a constant (Ln mg-n), and n is an exponent, which can take values in a wide range (>0), allowing to suitably predict the adsorption isotherms.The isotherm model of Khan was developed for studying the adsorption of aromatic compounds on activated carbons from multi-component aqueous phase solutions [43\u201345]. The generalized model for a single solute could be formulated according to Eq. (7),\n\n(7)\n\n\n\n\nq\n\n\ne\n\n\n=\n\n\n\n\nq\n\n\nm\n\n\n\u00b7\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\n\n\n\n\n1\n+\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\n\n\nn\n\n\n\n\n,\n\n\n\nwhere q\n\nm\n and K are constants measured in mg g\u22121 and L mg\u22121, respectively, and n is the exponent.The model Vieth-Sladek was first proposed to model adsorption of gases in glassy polymers. Owing to the remarkable resemblance of the studied application with the adsorption on porous solids [92] it has been also used to model the adsorption of model pollutants [47,91]. The model may be expressed by Eq. (8),\n\n(8)\n\n\n\n\nq\n\n\ne\n\n\n=\n\n\n\n\nq\n\n\nm\n\n\n\u00b7\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\n1\n+\nK\n\u00b7\n\n\nC\n\n\ne\n\n\n\n\n+\nn\n\u00b7\n\n\nC\n\n\ne\n\n\n,\n\n\n\nwhere q\n\nm\n, K and n are constants measured in mg g\u22121 for q\nm and L mg\u22121 for both K and n.Brouers and Sotolongo proposed a Weibull distribution as a possible empirical isotherm model [8] that has been used to predict pollutants adsorption on carbon-based materials [91]. The Brouers and Sotolongo equation is formulated as Eq.(9),\n\n(9)\n\n\n\n\nq\n\n\ne\n\n\n=\n\n\nq\n\n\nm\n\n\n\u00b7\n\n\n\n1\n\u2212\nexp\n\n\n\n\u2212\nK\n\u00b7\n\n\nC\n\n\ne\n\n\nn\n\n\n\n\n\n\n\n\n,\n\n\n\nwhere q\n\nm\n and K are constants measured in mg g\u22121 and Ln mg-n, respectively, and n is the exponent.The Jovanovi\u0107 model consists of two equations developed for the physical adsorption on monolayer and multilayer adsorption. Initially, this model was developed for adsorption in the gas phase [41], but it is largely used in the adsorption of solutes from aqueous media solutions [91,93].The pseudo-first-order equation describes the adsorption rate based on the monolayer adsorption capacity [33] and it is typically represented by Eq. (10):\n\n(10)\n\n\n\n\nq\n\n\nt\n\n\n=\n\n\nq\n\n\ne\n\n\n\u00b7\n\n\n\n1\n\u2212\nexp\n\n\n\n\u2212\nk\n\u00b7\nt\n\n\n\n\n\n\n,\n\n\n\nwhere q\n\nt\n and q\n\ne\n refer the solute adsorbed (Ni(II)) per mass of adsorbent (mg g\u22121) at a time of contact t (min) and at the equilibria stage, respectively, and k represents the rate constant of the adsorption process (min\u22121).The pseudo-second-order model [58], also found as an hyperbolic model [20], is typically used to describe adsorption processes controlled by chemisorption, involving valence forces through sharing or exchange of electrons between the adsorbent and the adsorbate. Eq. (11) represents this model,\n\n(11)\n\n\n\n\nq\n\n\nt\n\n\n=\n\n\n1\n\n\n\n\n1\n\n\n\n\n\nk\n\u00b7\n\n\n\n\nq\n\n\ne\n\n\n\n\n2\n\n\n\n\n\n\n\n\u00b7\n\n\n1\n\n\nt\n\n\n+\n\n\n1\n\n\n\n\n\n\n\nq\n\n\ne\n\n\n\n\n\n\n\n\n\n,\n\n\n\nwhere the rate constant k is measured in g mg\u22121 min\u22121.Bangham is a pore diffusion model expressed by Eq. (12):\n\n(12)\n\n\n\n\nq\n\n\nt\n\n\n=\nk\n\u00b7\n\n\nt\n\n\n1\n/\nm\n\n\n,\n\n\n\nwhere the rate constant k is measured in mg g\u22121 min\u22121/m and m is an exponent.The Elovich equation is a model based on chemical adsorption [70], typically used in the simplified form obtained by Chien and Clayton [20]. In this work, the integrated form of Elovich equation was used, as shown in Eq. (13),\n\n(13)\n\n\n\n\nq\n\n\nt\n\n\n=\n\n\n1\n\n\n\u03b2\n\n\n\u00b7\nln\n\n\n\n\u03b1\n\u00b7\n\u03b2\n\u00b7\nt\n+\n1\n\n\n\n,\n\n\n\nwhere \u03b1 and \u03b2 (two-parameter model) are the Elovich constants measured in mg g\u22121 min\u22121 and g mg \u22121, respectively.The D\u00fcnwald-Wagner intraparticle diffusion model is typically expressed as shown in Eq. (14)\n[69],\n\n(14)\n\n\n\n\nq\n\n\nt\n\n\n=\n\n\nq\n\n\ne\n\n\n\u00b7\n\n\n1\n\u2212\nexp\n\n\n\n\u2212\nk\n\u00b7\nt\n\n\n\n\n\n,\n\n\n\nwhere the rate constant k is measured in min\u22121 and m is an exponent.Weber-Morris equation is another mechanistic model typically found as shown in Eq. (15),\n\n(15)\n\n\n\n\nq\n\n\nt\n\n\n=\nk\n\u00b7\n\n\nt\n\n\n+\nm\n,\n\n\n\nwhere the rate constant k is measured in mg g\u22121 min\u22121/2 and m is a parameter measured in mg g\u22121.The Avrami kinetic model was developed considering possible changes of the adsorption rates as a function of the initial concentration and the adsorption time, as well as the determination of fractionary kinetic orders [56] and it is expressed as in Eq.(16),\n\n(16)\n\n\n\n\nq\n\n\nt\n\n\n=\n\n\nq\n\n\ne\n\n\n\u00b7\n(\n1\n\u2212\nexp\n\n\n\n\n\n\u2212\nk\n\u00b7\nt\n\n\n\n\nm\n\n\n\n)\n,\n\n\n\nwhere the rate constant k is measured in min\u22121 and m is an exponent (only three parameter-kinetic adsorption model used in this work).Tangerine peels (TP) were obtained after domestic use. 99.995\u00a0% nitrogen was supplied from Praxair. 97\u00a0% iron (III) chloride hexahydrate (FeCl3.6\u2009H2O) was supplied from Panreac, 95\u00a0% nickel(II) chloride hexahydrate (NiCl2.6\u2009H2O), 98\u00a0% sodium hydroxide (NaOH) and 37\u00a0% hydrochloric acid (HCl) were obtained from Fisher chemicals. All reagents were used as received without further purification, and distilled water was used throughout the research.TP was first dried in oven at 100\u2009\u00baC for 24\u2009h, and then grinded and sieved to obtain particle sizes between 106 and 250\u2009\u00b5m using two sieves with metallic mesh (CISA) according to ISO 3310.1 and ASTM E-11\u201395 (N\u00ba 140 and 60, respectively). Hydrochar microspheres were then produced adapting the methodology described elsewhere [15,16]. Briefly, a suspension of 2.5\u2009g of the dried and sieved TP was prepared with 20\u2009mL of FeCl3 solution (2.5, 1.0 and 0.5\u2009M) in a 125\u2009mL high-pressure autoclave (Model 249\u2009M 4744\u201349, Parr Instrument co., USA), heated to 200 \u00baC for 3\u2009h under autogenous pressure. The recovered hydrochar microspheres were labelled as HCMS-2.5, HCMS-1.0, and HCMS-0.5, according to the concentration of FeCl3 solution used in the HTC.A pyrochar (PC) and activated carbons microspheres (ACMS-2.5, ACMS-1.0, ACMS-0.5 from HCMS-2.5, HCMS-1.0, and HCMS-0.5, respectively) were produced by pyrolysis of the TP and hydrochars, respectively, under N2 continuous flow (100 Ncm3 min\u22121) at 800 \u00baC, for 4\u2009h, using a tubular furnace (Therm Concept).The compositions of the solid materials (PC, HCMSs and ACMSs) were determined by elemental analysis (Carlo Erba Instrument EA 1108) to know the weight percentages of carbon, nitrogen, hydrogen and sulfur. To determine ashes, the carbonaceous materials were weighted before and after calcination, conducted in static air (muffle) at 800\u2009\u00baC for 4\u2009h.The textural properties of the carbonaceous materials were determined from the analysis of N2 adsorption-desorption isotherms at 77\u2009K, obtained in a Quantachrome NOVA TOUCH LX4 adsorption analyzer. Degasification was conducted for 16\u2009h at 120 \u00baC. BET, Langmuir, external and microporous surface areas (S\n\nBET\n, S\n\nLangmuir\n, S\n\next\n and S\n\nmic\n, respectively), micropore volume (V\n\nmic\n) and total pore volume (V\n\nTotal\n), were determined as described elsewhere [63].Scanning electron microscopy (SEM) images of the TP-based carbonaceous materials were obtained using a FEI Quanta 400FEG ESEM/EDAX Genesis X4Minstrument equipped with an Energy Dispersive Spectrometer (EDS).Functionalities were studied through X-ray photoelectron spectroscopy (XPS) and Fourier-transform infrared spectroscopy (FT-IR). XPS analysis was conducted in a PHI-5701 of Physical Electronics, whereas FT-IR spectra were obtained with a Perkin Elmer FT-IR spectrophotometer UATR Two with a resolution of 1\u2009cm\u22121 and scan range of 3000\u2013450\u2009cm\u22121.Acidity and basicity of the carbon-based materials were determined by acid-base titration of an acid or base solution after keeping in contact with the adsorbents for 48\u2009h, as detailed in the literature [18,73]. Surface acidity (SA) and basicity (SB) were determined considering the BET surface area of each adsorbent.Equilibrium adsorption isotherms of Ni(II) on the ACs, PC and HCs were determined by means of the equilibrium method [19]. First, 0.125\u2009g of adsorbent were added into 50\u2009mL of nickel(II) chloride solutions at different concentrations (5, 10, 20, 50, 80 and 100\u2009mg\u2009L\u22121 of Ni(II)). The mixtures were stirred at 240\u2009rpm for 72\u2009h.Kinetic adsorption of Ni(II) on the TP-based materials was conducted using 0.125\u2009g of adsorbent and 50\u2009mL of a 5\u2009mg\u2009L\u22121 nickel(II) chloride solution. The adsorption was conducted at 240\u2009rpm. Then, different samples were withdrawn from the Erlenmeyer at the following selected times: 15, 30, 60, 120, 240 and 1440\u2009minThe effect of pH in the adsorption of Ni(II) onto the TP-based adsorbents was assessed at pH ranging from 3 to 9. For each run, 50\u2009mL of the 100\u2009mg\u2009L\u22121 Ni(II) chloride solution was used and 0.125\u2009g of the adsorbent was added. The pH of the solution was adjusted using 1\u2009mol\u2009L\u22121 HCl and 1\u2009mol\u2009L\u22121 NaOH during all runs. After 72\u2009h, the samples were filtered in order to separate the adsorbent from the liquid fraction and the concentration of Ni (II) in the filtrate was determined.Samples withdrawn during the adsorption experiments were filtered to separate the adsorbent, and the liquid samples were analyzed by atomic absorption spectrophotometry (Varian SpectrAA 220, Steinhausen, Switzerland) to determine Ni(II) in the aliquots.The amount of Ni(II) adsorbed on the TP-based materials was determined by application of Eq. (17),\n\n(17)\n\n\n\n\nq\n\n\nt\n\n\n=\n\n\n\n\n\n\n\n\nC\n\n\nNi\n\n\n\nII\n\n\n\n,\n0\n\n\n\u2212\n\n\nC\n\n\nN\ni\n\n\n\nII\n\n\n\n,\nt\n\n\n\n\n\n\nV\n\n\n\n\nW\n\n\nadsorbent\n\n\n\n\n,\n\n\n\nwhere q\n\nt\n refers to the amount of Ni(II) adsorbed per unit mass of TP-based material at time t (mg g\u22121), C\n\nNi(II),0\n is the initial Ni(II) concentration in the aqueous solution (mg L\u22121), C\n\nNi(II),t\n is the concentration of Ni(II) in the solution at the adsorption time t (mg L\u22121), W\n\nadsorbent\n refers to the mass (g) of TP-based material and V is the volume of the aqueous solution (L).Kinetic and isotherm adsorption models were obtained using non-linear regression since better-fitted equations are obtained than using linearized equations [49], consisting of successive numerical iterations to minimize the least sum of squared errors (SSE) of q\n\nt\n (cf. Eq. (18)), as detailed in previous works dealing with modeling methods [17,20],\n\n(18)\n\n\nSSE\n=\n\n\u2211\n\ni\n=\n1\n\n\nn\n\n\n\n\n\n\n\n\n\nq\n\n\nt\n,\nexp\n,\ni\n\n\n\u2212\n\n\nq\n\n\nt\n,\nmodel\n,\ni\n\n\n\n\n\n\n2\n\n\n\n,\n\n\n\nwhere q\n\nt,exp,i\n (mg g\u22121) is the amount of Ni(II) adsorbed per unit mass of TP-based adsorbent at time t in the measured adsorption experiments (q\n\nt\n being expressed as q\n\ne\n for equilibrium runs), q\n\nt,model,i\n (mg g\u22121) the respective calculated values given by the model, i representing each value up to n values obtained in each experiment.Alternatively, different error functions were used, viz. the sum of the square of the errors (SSE), the sum of absolute errors (SAE), the hybrid error function (HYBRYD), the Marquard\u2019s percent standard deviation (MPSD), and the average relative error (ARE) [42] to assure the good fitness of the models. SAE, HYBRYD, MPSD, and ARE error functions are respectively described by Eqs. (19\u201322),\n\n(19)\n\n\nSAE\n=\n\n\u2211\n\ni\n=\n1\n\n\nn\n\n\n\n\n\n\n\nq\n\n\nt\n,\nexp\n,\ni\n\n\n\u2212\n\n\nq\n\n\nt\n,\nmodel\n,\ni\n\n\n\n\n\n,\n\n\n\n\n\n\n(20)\n\n\nHYBRYD\n=\n\n\n100\n\n\nn\n\u2212\np\n\n\n\n\u2211\n\ni\n=\n1\n\n\nn\n\n\n\n\n\n\n\n\n\n\n\nq\n\n\nt\n,\nexp\n,\ni\n\n\n\u2212\n\n\nq\n\n\nt\n,\nmodel\n,\ni\n\n\n\n\n\n\n2\n\n\n\n\n\n\nq\n\n\nt\n,\nexp\n,\ni\n\n\n\n\n\n,\n\n\n\n\n\n\n(21)\n\n\nMPSD\n=\n100\n\n\n\n\n1\n\n\nn\n\u2212\np\n\n\n\n\u2211\n\ni\n=\n1\n\n\nn\n\n\n\n\n\n\n\n\n\n\n\nq\n\n\nt\n,\nexp\n,\ni\n\n\n\u2212\n\n\nq\n\n\nt\n,\nmodel\n,\ni\n\n\n\n\n\n\nq\n\n\nt\n,\nexp\n,\ni\n\n\n\n\n\n\n\n\n2\n\n\n\n\n\n,\n\n\n\n\n\n\n(22)\n\n\nARE\n=\n\n\n100\n\n\np\n\n\n\n\u2211\n\ni\n=\n1\n\n\nn\n\n\n\n\n\n\n\n\n\nq\n\n\nt\n,\nexp\n,\ni\n\n\n\u2212\n\n\nq\n\n\nt\n,\nmodel\n,\ni\n\n\n\n\n\n\nq\n\n\nt\n,\nexp\n,\ni\n\n\n\n\n\n\n\n,\n\n\n\nwhere p refers to the number of parameters for each model (remaining parameters as above described for SSE).The models were also evaluated by the determination factor (r\n\n2\n) and the adjusted determination factor (r\n\n2\n\n\nadjust.\n), to take into account the degrees of freedom or the number of parameters from each model equation [28].\n\nTable 1 summarizes the carbon, hydrogen, nitrogen, sulfur and ash contents for the precursor (TP) and for the prepared carbonaceous materials. Compared to the raw waste (42.1\u2009wt\u00a0% and 6.17\u2009wt% of C and H, respectively), it is possible to observe that the carbon content increases (64.7\u201393.1\u2009wt%) and the hydrogen composition decreases (1.06\u20135.21\u2009wt%) for all TP-based materials prepared, resulting in an increment of the C/H ratio (from 6.8 in the TP precursor to 12.4\u201382.8 in the prepared materials). The effect is more evident in materials subjected to pyrolysis (ACMSs and PC) since the thermal process causes the release of volatile compounds, such as water and low molecular weight hydrocarbons, and the carbonization of the sample [12,16].The increase in the C/H ratio observed for hydrochars (12.4\u201313.9) with respect to TP (6.82) is due to aromatization, condensation or polycyclization reactions during the carbonization of TP [12,16]. ACs (ACMS-2.5, ACMS-1.5 and ACMS-0.5) prepared by sequential HTC and pyrolysis show the highest C/H ratios (77.0\u201382.8), due to the effect of both processes. Among the ACMSs and HCMSs, ACMS-2.5 and HCMS-2.5 show the highest C/H ratios (82.8 and 13.9), evidencing the role played by the iron catalyst during the carbonization processes.The ash content in the TP and in the carbon-based materials is a measure of the inert, inorganic and probably unusable part of the material whose presence may modify the interaction between the surface of the carbon material and the adsorbate [6]. The TP-based materials show values of ashes ranging from 1.4 to 8.0\u2009wt%. The highest value was found for PC since pyrolysis leads to the volatilization of the organic compounds of the precursor TP, as also observed in works dealing with the production of carbon materials from other sources [16]. The same effect was observed for ACMSs (2.2\u20135.1\u2009wt% of ashes) prepared by pyrolysis from HCMS (1.4\u20134.8\u2009wt% of ashes). On the other hand, HTC leads to a decrease in the ash content, likely due to the leaching of alkali and alkaline earth metals present in the TP, promoted by the contact of the solid with the high temperature liquid solution, as was observed in previous works dealing with HTC or using acid solutions during the activation of ACs [16,17,71]. Only one hydrochar (HCMS-2.5) shows a slight increase of ash content (from 4.4 to 4.8\u2009wt%), which was ascribed to the impregnation of the material with iron during HTC, as evidenced by the tendency of increasing ash content with increasing iron concentration: 4.8, 3.2 and 1.4\u2009wt% for hydrochars prepared with 2.5, 1.0 and 0.5\u2009M of FeCl3, respectively.The remaining content (different from C, H, N, S and ashes) is typically associated with other heteroatoms, such as oxygen. As observed, its content decreases after either pyrolysis or HTC due to the carbonization processes (those elements are released).The N2 adsorption isotherms of the studied TP-based adsorbents are depicted in Fig. S1 and the textural properties obtained through the calculation methods described in the methodology are summarized in \nTable 2. As observed, nitrogen adsorption isotherms show higher quantities of volume adsorbed for ACMSs, followed by PC and HCMSs, whose uptake capacity is considerably lower. Similar trend has been reported in studies dealing with the preparation of ACs, PCs and HCs and from different precursors [16].As expected, ACMSs have the highest BET (238\u2013287\u2009m2 g\u20131) and Langmuir (330\u2013391\u2009m2 g\u20131) surface areas and total pore volumes (162\u2013282\u2009mm3 g\u20131), among all adsorbents prepared. ACMSs show similar results to those reported in the literature regarding the synthesis of carbon materials prepared by diverse activation and carbonization methods of rice husk (171\u2013280\u2009m2 g\u20131) [13,64], palm shell (260\u2013266\u2009m2 g\u20131) [14,31], or coconut shell (183\u2009m2 g\u20131) [14].The PC sample, obtained by pyrolysis of TP without any other treatment, reaches values considerably lower than ACMSs (104 and 146\u2009m2 g\u22121 of BET and Langmuir specific surface areas, respectively and 66\u2009mm3 g\u22121 of total pore volume), evidencing that HTC works as activation process for the development of ACMSs. The PC sample presents a significant microporosity (S\n\nmic\n = 94\u2009m2 g\u20131 and V\n\nmic\n = 50\u2009mm3 g\u20131), showing that citric peels are interesting precursors for the development of carbon-based adsorbents. Obviously, microporosity increases when HTC is used prior to pyrolysis, obtaining ACs instead of PC, resulting in materials with a microposity two times higher than that in PC (S\n\nmic\n = 198\u2013217\u2009m2 g\u20131 and V\n\nmic\n = 101\u2013116\u2009mm3 g\u20131).Moreover, the differences found among the ACMSs evidence the effect of the quantity of iron impregnating the HCs as an active catalyst for the carbonization and development of porosity, since specific surface areas and total pore volume increase in the following order ACMS-0.5\u2009<ACMS-1.0\u2009<ACMS-2.5. Similar values are found in the literature [16,71] for the specific surface areas and pore volume of HCs that were determined according to the nitrogen adsorption isotherms (Fig. S1).Preliminary tests related to the adsorption of Ni(II) (in 5\u2009mg\u2009L\u20131 aqueous solution) over the TP-based carbonaceous materials (2.5\u2009g\u2009L\u20131 of adsorbent) at room temperature reveal that is possible to remove an amount of the heavy metal from the aqueous solution per unit mass of the adsorbent ranging from 0.19 to 1.99\u2009mg per g of adsorbent (cf. Table 2), being the highest uptake capacity of Ni(II) obtained with the PC adsorbent, i.e., the pyrochar. Although ACs show more interesting textural properties for this application (highest specific surface areas and total pore volumes); however, lower adsorption capacities are obtained with ACMSs (1.43\u20130.40\u2009mg\u2009g\u22121) when compared to the pyrochar (1.99\u2009mg\u2009g\u20131). On the other hand, HCMSs show some unexpected uptake capacity (0.60\u20130.19\u2009mg\u2009g\u20131), when considering their textural properties. In fact, the capacity of adsorption per surface area (Q\n\ne\n) of these materials is significantly higher (27.8\u201354.8\u2009\u03bcg\u2009m\u20132) than that of PC (19.1\u2009\u03bcg\u2009m\u20132) and ACMSs (1.7\u20135.0\u2009\u03bcg\u2009m\u20132).It is noteworthy that the largest uptake capacities among ACMSs (1.43\u2009mg\u2009g\u22121) and HCMSs (0.60\u2009mg\u2009g\u22121) were obtained for the TP-based adsorbents prepared with the highest concentration of FeCl3 during the HTC (ACMS-2.5 and HCMS-2.5). For this reason, only these materials were selected for further investigation.Texture and morphology of three selected TP-based adsorbents (HCMS-2.5, PC and ACMS-2.5) were determined through SEM (cf. \nFig. 1). The microphotographs confirm the production of microspheres by HTC using FeCl3. For HCMS-2.5, it is possible to observe particle sizes ranging from 1858 to 3615\u2009nm at low image magnification (Fig. 1a), but many carbon microspheres were found with sizes between 50 and 350\u2009nm (Fig. 1b).The material prepared by pyrolysis of TP without further physical-chemical treatment (PC) presents higher particle sizes (105.7\u2013238.1\u2009\u00b5m) than HCMS-2.5 and ACMS-2.5 (Fig. 1c). The particle size of PC is in accordance with the sieving performed to prepare the adsorbents (sieves with a metallic mesh to separate particle sizes between 106 and 250\u2009\u00b5m). The implication is that the particle size of the precursor is not strongly affected by pyrolysis, as it is by HTC with FeCl3, which led to decrease more than 100 times the particle size. Regarding the morphology, PC shows irregular particle shape and a rough surface, likely due to the porosity developed during the pyrolysis to prepare PC.The pyrolysis of HCMS-2.5 to obtain ACMS-2.5 allowed to keep the particle size of microspheres, although it is possible to observe the sintering of some microspheres of HCMS-2.5, leading to agglomeration of activated carbon microspheres (Fig. 1e). The surface of this material is also slightly rough, probably due to the porosity development during the thermal treatment, as observed for PC.Selected regions of micrographs were analysed by EDS to determine the elemental content of HCMS-2.5, PC and ACMS-2.5 (Fig. S2). As observed, the materials mainly consist of carbon. HCMS-2.5 (Fig. S2a) and ACMS-2.5 (Fig. S2c) show a slight iron content. On the other hand, the EDS of PC adsorbent confirms the presence of alkali and alkaline earth metals (Na, Mg, K and Ca) from the precursor. The presence of this metal and the highest signal in oxygen for PC compared to ACMS-2.5 is in accordance with the CHNS-elemental analysis presented above. Alkali and alkaline earth metals were not detected for HCMS-2.5, nor ACMS-2.5, meaning that HTC allows to obtain more purified carbonized adsorbents.The adsorptive properties of carbonaceous adsorbents are influenced not only by their textural properties but also by their surface chemistry. TP-based adsorbents may contain heteroatoms such as oxygen, nitrogen and phosphorus on their surface that form organic functional groups, such as carbonyls, carboxylic acids, lactones, ethers, phenols, aldehydes, amines and phosphates, which can be neutral, acidic or basic [40]. The surface chemistry depends on the composition of the precursor and on the method of carbonization [7,30]. \nFig. 2 shows FT-IR spectra obtained for the TP precursor and the selected adsorbents (ACMS-2.5, PC and HCMS-2.5). The bands were assigned to probable bounds according to FT-IR spectra of diverse carbon-based materials [90].As observed, the TP precursor presents more bands than the respective TP-based carbon materials due to the complex composition of the biomass. As a C-rich material, some bands can be ascribed to the presence of C-H (2930\u20132850\u2009cm\u22121) and CC (1635\u2009cm\u22121) bonds. However, this material stands out for the high content of surface oxygen groups (SOGs). Hydroxyl groups (-O-H) are found at 1635\u20131625\u2009cm\u22121. Carbonyls (-CO) can be identified at 1385\u2009cm\u22121 and the presence of epoxy groups are demonstrated by bands at 1050, 1075 and 1150\u2009cm\u22121. The carbonized samples (HCMS-2.5, PC and ACMS-2.5) show a strong decrease in the bands representative of SOGs. In fact, PC and ACMS-2.5 only present a significant band at 1630\u2009cm\u22121 that can be assigned to double bounds between carbon atoms. The absence of significant bands for CH bonds and SOGs demonstrated the success of the carbonization and a poor functionalized surface. The lower content of O and H agrees with the elemental composition found by elemental analysis. On the other hand, HCMS-2.5 show significant bands at 2930\u20132850\u2009cm\u22121 (CH bonds) and the regions of diverse SOGs, as expected for the hydrothermal treatment at low temperatures, also agreeing with the results for elemental analysis.\n\nFig. 3 shows the XPS spectra and the relative atomic percentage content of elements on the surface of each TP-based adsorbent. As expected, HCMS-2.5 corresponds to hydrochar microspheres rich in surface oxygen groups (SOGs), since 17.73\u00a0% of O1s was obtained. Surprisingly, PC also presents a high content of oxygen on the surface and this technique also confirms the presence of metals (K and Ca). The values obtained are close to those obtained by SEM/EDS and also similar to the elemental composition obtained by CHNS-AE. K and Ca were not detected in ACMS-2.5 and HCMS-2.5, as HTC can remove those elements, as corroborated by SEM/EDS. ACMS-2.5 shows the lowest content of oxygen and the highest content of C among the selected materials, agreeing with the results obtained by elemental analysis. As observed, most of iron was removed from the microsphere materials (HCMS-2.5 and ACMS-2.5) after washing. On the other hand, the low iron content observed for PC most likely comes from the raw TP used as a precursor.To deepen knowledge of the surface chemistry, O1s peaks of XPS spectra analysis were assessed for the materials without significant quantities of other metals, i.e., HCMS-2.5 and ACMS-2.5. O1s deconvolution curves are represented in Fig. S3 and the assignments to the relative content of SOGs are summarized in \nTable 3. As observed, HCMS-2.5 presents mainly hydroxyl groups (14.7\u00a0% out of 17.7\u00a0%), whereas ACMS-2.5 shows lower content of hydroxyl groups with a more predominant presence of carbonyl groups (1.5\u00a0% out of 3.9\u00a0%). The reduction of hydroxylic groups in ACMS-2.5 can be ascribed to the release of weak hydroxyl groups during the pyrolysis at 800 \u00baC of HCMS-2.5 to obtain ACMS-2.5.The results obtained in terms of acidity and basicity of the selected adsorbents (ACMS-2.5, PC and HCMS-2.5) are shown in \nTable 4. As can be seen, acid-base properties show significant differences among the adsorbents, hydrochar HCMS-2.5 being more acid (0.32\u2009mmol\u2009g\u20131) than ACMS-2.5 (0.23\u2009mmol\u2009g\u20131) and, especially, than PC (0.01\u2009mmol\u2009g\u20131). Taking into account the specific surface area of the materials, the same order is observed for the surface acidity: HCMS-2.5 (45.7 \u03bcmol m\u20132) >\u2009ACMS-2.5 (0.78 \u03bcmol m\u20132) >\u2009PC (0.10 \u03bcmol m\u20132). On the opposite, the basic character of the adsorbents follows a different order when measured per mass or per surface area, highlighting the basicity of PC (1.83\u2009mmol\u2009g\u20131) and HCMS-2.5 (30.4 \u03bcmol m\u20132). Interestingly, the same order was found for the basicity (Table 4) and the uptake capacity of the materials (Table 2) per gram (PC > ACMS-2.5\u2009> HCMS-2.5) and per surface area (HCMS-2.5\u2009> PC > ACMS-2.5), evidencing the strong role played by the basicity in the adsorption of Ni(II).The highest acidity of HCMS-2.5 can be ascribed to the chloride precursor (FeCl3) used as activating agent during HTC. The strong acid and basic character of this material compared to the others are due to the carbonization in the presence of water that leads to a material highly functionalized with oxygen-containing surface groups [4]. This can be expected by the oxygen content of hydrochars according to the remaining content presented in Table 1 (24.0\u201328.0\u2009wt% for HCMSs, whereas 2.5\u20139.6\u2009wt% was found for PC and ACMSs). The basicity of PC may be ascribed to the fact of having the highest content of ashes (8.0\u2009wt%, whereas less than 5.1\u2009wt% is found for other TP-based adsorbents), consisting mainly of alkali and alkaline earth metals, as observed both in SEM/EDS and XPS analysis.\n\nFig. 4 shows the isotherm adsorption of Ni(II) on the selected adsorbents: (a) HCMS-2.5, (b) PC and (c) ACMS-2.5. As observed, PC shows the highest uptake capacity of Ni(II), reaching values of 13.9\u2009mg\u2009g\u20131 at the highest tested initial concentration of Ni(II) (C\n\nNi(II),0\n =\u2009100\u2009mg\u2009L\u20131). In contrast, the highest adsorption capacity of ACMS-2.5 and HCMS-2.5 was 5.18 and 4.88\u2009mg\u2009g\u20131, respectively, at the same operating conditions.The shape of the isotherms is also considerably different among the TP-based adsorbents. Giles classification is usually used to distinguish the isotherm adsorption curves according to their characteristic shapes between four isotherm classes: high affinity (H), Langmuir (L), constant partition (C) and sigmoidal-shaped (S) [29]. H and L isotherms have a convex shape. However, the slopes of H isotherms reach higher values than L isotherms because the sorption affinity of H isotherms strongly increases with decreasing concentration. S isotherms have a concave shape at low concentrations, while C isotherms are defined by a constant sorption affinity [34].Accordingly, the isotherm curve obtained with ACMS-2.5 may be classified as L1 since it describes a medium affinity at low concentrations of Ni(II) and equilibrium uptake capacities (q\n\ne\n) gradually describe an asymptotic curve for C\n\ne\n higher than approximately 20\u2009mg\u2009L\u20131. The isotherm curve obtained with PC may be classified as H1, since it describes a similar curve, but it shows a strong affinity at a low concentration of Ni(II). The isotherm curve obtained with HCMS-2.5 may be classified as C1, because of the line trend described by the equilibrium uptake capacities upon the initial concentration of Ni(II). These differences may be ascribed to the different textural and acid-base properties of the adsorbents. PC should present the highest affinity and uptake capacity for the combination of a significant porosity and basicity, whereas HCMS-2.5 has not enough specific surface area and ACMS-2.5 shows limited functionality.The isotherms for adsorption of Ni(II) onto the TP-based adsorbents were evaluated by 10 models: 7 with three parameters (Sips, Redlich-Peterson, T\u00f3th, Khan, Vieth-Sladek, Brouers-Sotolongo and Jovanovi\u0107 for multilayer adsorption), and 3 with two parameters (Jovanovi\u0107 for monolayer adsorption, Langmuir and Freundlich) in their functions [20]. Those models were fitted by a non-linear regression method, since model parameters may be distorted by linear regressions [49,54] and adjusted determination factor (r\n\n2\n\n\nadjust.\n) used to take into account the number of parameters from each model function [20]. The equations, the parameter values and the most significant statistical data obtained from the fitting of the isotherm adsorption models to the experimental data are compiled in \nTable 5 (isotherm curves obtained are depicted in Fig. 4). As observed, most of the isotherm models accurately fit (r\n\n2\n = 0866 \u2013 0.996 and r\n\n2\n\n\nadjust.\n = 0.809 \u2013 0.993) to the experimental data obtained with the TP-based adsorbents. The fitting with all models considered the minimization of different error functions (SSE, SAE, HYBRYD, MPSD and ARE), since other studies reported about the importance of the objective function in the fitting with the models [42,48]). However, non-significant differences were observed among the models for the parameter and statistical data (r\n\n2\n), as exemplified by the q\n\nm\n values predicted using SSE, SAE, HYBRYD, MPSD, and ARE for each isotherm model in Fig. S4. For this case, the maximum difference for the predicted value of q\n\nm\n was 2.8\u2009mg\u2009g\u22121 found for the T\u00f3th model (simplified function). It is noteworthy that the isotherm models keep the same order of values for the predicted q\n\nm\n regardless of the error functions used for the fitting, as follows: T\u00f3th >\u2009Sips >\u2009Brouers-Sotolongo >\u2009Langmuir >\u2009Jovanovi\u0107 (monolayer) >\u2009Vieth-Sladek >\u2009Redlich-Peterson >\u2009Jovanovi\u0107 (multilayer) >\u2009Khan >\u2009Freundlich. A similar trend was observed in a previous work [20], so it may be expected to predict higher values of monocape uptake capacity (q\n\nm\n) with the models of T\u00f3th (simplified model), Sips, Brouers-Sotolongo, Langmuir and Jovanovi\u0107 (model for monolayer adsorption). Among the TP-based materials tested, PC leads to the highest values of both K and q\n\nm\n, evidencing it as the material with the highest affinity and uptake capacity.The isotherm adsorption of Ni(II) on ACMS-2.5 is best represented (r\n\n2\n = 0986 \u2013 0.989, and r\n\n2\n\n\nadjust.\n = 0.977 \u2013 0.983) by hyperbolic isotherm adsorption models (Sips, Redlich-Peterson, T\u00f3th, Khan, Vieth-Sladek and Langmuir), which is in agreement with the sorted L1 type isotherm curve according to Giles classification. Taking into account the degrees of freedom in the fitting (r\n\n2\n\n\nadjust.\n), Langmuir is the best model representing the curve of ACMS-2.5 with q\n\nm\n =\u20095.44\u2009mg\u2009g\u22121, K =\u20090.334\u2009L\u2009mg\u22121, r\n\n2\n\n\nadjust.\n =\u20090.983, as also evidenced by the value taken by parameter n (close to 1 as the exponent, and close to 0 for Vieth-Sladek) in the other models. The values of the q\n\nm\n and K for the Langmuir model obtained through the minimization of the different error functions (SSE, SAE, HYBRYD, MPSD, and ARE) were similar (5.44\u20135.50\u2009mg\u2009g\u22121, and 0.289\u20130.334\u2009L\u2009mg\u22121 for q\n\nm\n and K, respectively).The adimensional Langmuir separation factors (R\n\nL\n) predicted from the Langmuir model for each TP-based adsorbent are depicted in Fig. S5. As can be seen, the values of R\n\nL\n obtained for ACMS-2.5, PC, and HCMS-2.5 range from 0 to 1. Hence adsorption is favorable in all cases, but R\n\nL\n values show great differences among the TP-based adsorbents. HCMS-2.5 leads to values close to 1, characteristic of linear type isotherms, as is described by this adsorbent (Fig. S1). On the opposite, the adsorbent presenting the highest adsorption capacity (PC) shows values of R\n\nL\n close to 0, expected for irreversible adsorption.The isotherm adsorption of Ni(II) obtained with PC is also well described using hyperbolic models (r\n\n2\n = 0.930 \u2013 0.986, and r\n\n2\n\n\nadjust.\n = 0.912 \u2013 0.977) being best represented by Khan (q\n\nm\n = 5.62\u2009mg\u2009g\u22121, K = 27.7\u2009L\u2009mg\u22121, n\u2009=\u20090.880, r\n\n2\n\n\nadjust. = 0.977) and Vieth-Sladek models (q\n\nm\n = 9.26\u2009mg\u2009g\u22121, K = 12.8\u2009L\u2009mg\u22121, n\u2009=\u20090.082\u2009L\u2009g\u22121, r\n\n2\n\n\nadjust. = 0.977). The differences obtained by fitting with different error functions were also negligible.For the isotherm adsorption curve of Ni(II) on HCMS-2.5, the model developed for Jovanovi\u0107 for multilayer adsorption (q\n\nm\n = 1.53\u2009mg\u2009g\u22121, K = 0.137\u2009L\u2009mg\u22121, n\u2009=\u20090.014\u2009L\u2009mg\u22121) considerably fitted better (r\n\n2\n = 0.996 and r\n\n2\n\n\nadjust.\n = 0.993) than all other models evaluated. In a previous work, Jovanovi\u0107 multilayer isotherm was found to satisfactorily predict the adsorption of a pollutant onto activated carbonaceous materials [20] with the same magnitude for both K and m constant values.Considering the BET surface area of each TP-based adsorbent, the uptake capacity on the monolayer of HCMS-2.5 could be considered the highest (Q\n\nm\n = q\n\nm\n /S\n\nBET\n = 139\u2009\u03bcg\u2009m\u20132). In this case, the values of the parameters obtained using different error functions were also similar.The adsorption of Ni(II) from aqueous solution (C\n\nNi(II),0\n =\u20095\u2009mg\u2009L\u22121) upon time of contact with 2.5\u2009g\u2009L\u22121 of ACMS-2.5, PC, and HCMS-2.5 is represented in \nFig. 5. Although HCMS-2.5 shows the lowest adsorption capacity, the adsorption of Ni(II) on HCMS-2.5 is faster compared to ACMS-2.5 and PC, likely due to the absence of microporosity. In contrast, internal diffusion may be hindered in PC and ACMS-2.5. As a consequence, kinetic adsorption on HCMS-2.5 shows an asymptotic trend that is quickly reached, and the equilibrium uptake capacity should be achieved fast.The assessment of the kinetic adsorption of Ni(II) on the TP-based adsorbents was evaluated using pseudo-first-order, pseudo-second-order, Bangham, Elovich, D\u00fcnwald-Wagner, Weber-Morris, and Avrami kinetic models. The equation of these models, the kinetic constants and the statistical data resulting from their fitting to the experimental data obtained in the adsorption of Ni(II) on ACMS-2.5, PC, and HCMS-2.5 are summarized in \nTable 6. Those models were also fitted by a non-linear regression method [49,54] (r\n\n2\n\n\nadjust.\n was not presented, since only one model has three parameters \u2013 Avrami \u2013 and it is not the model best able to predict the data obtained).The kinetic adsorption curves on the TP-based adsorbents predicted by the models are represented in Fig. 5. As observed, most of the kinetic models are capable of accurately predicting the experimental data obtained and, except for the Weber-Morris for ACMS-2.5 and HCMS-2.5 (r\n\n2\n = 0.532\u20130.647), well fitness is available for the kinetic models to predict the kinetic adsorption curves of the TP-based adsorbents (r\n\n2\n = 0.864\u20130.999). As expected, all models predict higher values of equilibrium adsorption capacity for PC, followed by ACMS-2.5 and HCMS-2.5.Kinetic constants (k) for pseudo-first-order and pseudo-second-order are higher for the hydrochar (HCMS-2.5), followed by ACMS-2.5 and PC, evidencing that surface chemistry is not so determinant for the adsorption rate, as was found for the uptake capacity. The specific surface area and pore volume of the TP-based adsorbents do not show apparent relation with the kinetic constants, so the adsorption rate of Ni(II) on the ACMS-2.5, PC and HCMS-2.5 is ruled by the combination of different factors, such as textural properties and surface chemistry.In the case of HCMS-2.5, the highest adsorption rate of Ni(II) may be ascribed to the surface chemistry and to the absence of porosity (all Ni(II) is adsorbed on its external surface, and there is no internal diffusion). The model able to better predict the kinetic adsorption of Ni(II) on HCMS-2.5 was the pseudo-second-order (q\n\ne\n = 0.553\u2009mg\u2009g\u20131, k\u2009=\u20090.270\u2009g\u2009mg\u20131 min\u20131, r\n\n2\n = 0.999), due to the fast adsorption because of the affinity of HCMS-2.5 with Ni(II). ACMS-2.5 shows values of kinetic constant higher than PC, likely due to the highest specific surface area available (S\n\next\n = 70\u2009m2 g\u20131 and S\n\nmic\n = 217\u2009m2 g\u20131) and moderate affinity of ACMS-2.5 according to the surface chemistry (SA = 0.78 \u03bcmol m\u20132 and SB = 2.44 \u03bcmol m\u20132). The kinetic adsorption of Ni(II) on ACMS-2.5 was also better described by pseudo-second-order (q\n\ne\n = 1.07\u2009mg\u2009g\u20131, k\u2009=\u20090.0469\u2009g\u2009mg\u20131 min\u20131, r\n\n2\n = 0.978).It is typically assumed that the pseudo-second-order model fits well adsorption processes controlled by chemisorption, involving valence forces by sharing or exchange of electrons that may happen between Ni(II) and functional groups on the surface of HCMS-2.5 [58]. It is also reported that the adsorption of Ni on biomass-based adsorbents is well predicted by pseudo-second-order [50,84]. For the sample with the highest uptake capacity (PC), the pore diffusion model of Bangham was the kinetic model able to better predict the kinetic adsorption of Ni(II) on PC (k\u2009=\u20090.471\u2009mg\u2009g\u20131 min\u20131/m, m = 6.54, r\n\n2\n = 0.996), revealing that the process may be strongly controlled by the internal diffusion of Ni(II) inside pores of PC.The parameter m of the kinetic model of Bangham may be used as an indicator of the intensity in the adsorption of Ni(II) [70]. In this sense, the value of 49.9 for HCMS-2.5 reveals a strong affinity for the Ni(II) on this adsorbent, as expected for the fast adsorption observed. The value is considerably higher when compared with previous works [20].It is noteworthy that the best kinetic models predicting the adsorption of Ni(II) on each TP-based adsorbent was the same when other error functions were considered (SAE, HYBRYD, MPSD, and ARE) to fit the kinetic models, obtaining values for the parameters close to those presented minimizing SSE.In the same sense that acid-base functionalities of carbon-based adsorbents affect the adsorption of Ni(II), the pH of the aqueous solution also plays a key role in adsorption. The alteration of the pH media may cause dissociation of acid-base groups on the surface of the adsorbent and cause a significant improvement or worsening of the efficiency of removal of Ni(II) ions. In \nFig. 6 is shown the effect of pH on nickel adsorption with the TP-based adsorbents evaluated.It is notable that there is an increase in Ni adsorption at alkaline pH, as reported in literature [32,95,96]. At pH 3 there was no adsorption on any TP-based adsorbent. The adsorption of Ni is improved at a higher pH, due to the lower number of H+ ions, while at low pH there is a competition between hydronium cations and metals, reducing the adsorption of metal ions [1,57]. When the pH is higher, the concentration of H3O+ ions decreases and the sites on the surface of the carbon turn mainly into dissociated forms and can exchange H3O+ ions with metal ions in solution [86].According to the above results related to the characterization of the TP-based adsorbents and their performance, it can be concluded that textural properties and mainly surface chemistry plays a significant role in Ni(II) adsorption. HCMS-2.5 adsorbent shows a significant uptake capacity of Ni taking into account its BET surface (11\u2009m2 g\u22121), reaching the highest Ni adsorption capacity per surface area (54.6\u2009\u03bcg\u2009m\u22122). FT-IR and XPS spectra revealed that HCMS-2.5 is rich in SOGs mainly consisting of hydroxyl groups. PC sample, which was prepared by pyrolysis of the raw precursor at higher temperatures than HCMS-2.5, also presents a significant content of oxygen on its surface. Its functionalization, coupled with a higher porosity, make it a standout adsorbent for Ni removal. The pyrolysis of HCMS-2.5 to obtain ACMS-2.5 results in activated carbon microspheres with the highest surface area, but the adsorbent does not display a surface chemistry suitable for adsorption of Ni. A proper combination of SOGs allows adsorbing Ni by cation exchange (hydroxyl and carboxylic groups), electrostatic attraction (ester, carbonyl groups) or complexation (epoxy, ether groups) [37,96]. Taking into account that pseudo-second order model is the best kinetic model able to predict Ni adsorption on these adsorbents, it is expected that the process is controlled by chemisorption, involving valence forces by sharing or exchange of electrons because of the functional groups on their surface.\n\nTable 7 gives a comparison of the maximum Ni(II) adsorption capacity on different carbonaceous materials reported in literature. As observed, TP-based adsorbents show values of uptake capacity similar or higher than others reported taking into account the diverse operating conditions in the adsorption of Ni(II) found in literature (C\n\nNi(II),0\n =\u200930\u2013150\u2009mg\u2009L\u22121, C\n\nads\n =\u20090.6\u201320\u2009g\u2009L\u22121, 20\u201340 \u00baC, pH 3\u20138, 3\u201324\u2009h of contact time).The successful preparation of hydrochar and activated carbon microspheres from tangerine peels has been proved by hydrothermal carbonization with FeCl3 followed by pyrolysis at mild and moderate conditions, respectively, evidencing that biomass waste, such as citrus fruit peels, may be turned into high-added value materials in the context of a circular economy approach. The presence of chemical agents, such as FeCl3 in hydrothermal carbonization, has resulted in an effective tool for tuning the morphology, surface chemistry and increased carbonization yields in hydrochars and further preparation of activated carbons from those hydrochars. Furthermore, it was demonstrated that the shape and size of the microspheres obtained by hydrothermal carbonization are maintained even after heating to obtain activated carbon microspheres with higher porosity.The role played by textural properties and surface chemistry of carbonaceous adsorbents has been elucidated, concluding that the functionality of carbon-based materials is determinant in their performance for the adsorption of Ni (the highest uptake capacity of Ni was found with pyrochar, which shows a basicity of 1.83\u2009mmol\u2009g\u22121 and a S\n\nBET\n of 104\u2009m2 g\u20131, whereas a poor adsorption capacity was observed with the activated carbon, which has a S\n\nBET\n of 287\u2009m2 g\u20131).The best isotherm and kinetic models predicting the adsorption of Ni on the carbonaceous materials were different among them due to their unique characteristics. Langmuir, Khan, and Jovanovi\u0107 are models that best represent the adsorption of Ni on the activated carbon, pyrochar and hydrochar, respectively. On the other hand, kinetic adsorption of Ni was well predicted by the pseudo-second order model for activated carbon and hydrochar. In contrast, the adsorption of Ni on pyrochar was represented by the Bangham model.\nJose L. Diaz de Tuesta: Investigation, Conceptualization, Methodology, Formal analysis, Writing \u2013 original draft preparation, Visualization. Fernanda F. Roman: Investigation, Validation. Vitor C. Marques: Investigation, Writing\u00a0\u2013\u00a0original draft preparation. Adriano S. Silva: Investigation. Ana P.F. Silva: Investigation. Assem A. Shinibekova: Investigation. Sadenova Aknur: Investigation. Marzhan S. Kalmakhanova: Supervision. Bakytgul K. Massalimova: Supervision. Margarida Arrobas: Investigation. Tatiane C. Bosco: Supervision. Adri\u00e1n M.T. Silva: Supervision, Writing \u2013 review & editing, Funding acquisition, Project administration. Helder T. Gomes: Supervision, Conceptualization, Writing\u00a0\u2013\u00a0reviewing & editing, Funding acquisition, Project administration.The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Helder Teixeira Gomes reports financial support was provided by European Regional Development.The authors are grateful to the FCT (Foundation for Science and Technology, Portugal) and FEDER (European Regional Development Fund) under Programme PT2020 for financial support to CIMO (UIDB/00690/2020). We would also like to thank the scientific collaboration under Base-UIDB/50020/2020 and Programmatic-UIDP/50020/2020 funding of LSRE-LCM, and LA/P/0045/2020 funding of ALiCE, funded by national funds through FCT and\u00a0MCTES (Minist\u00e9rio da Ci\u00eancia, Tecnologia e Ensino Superior, Portugal) by PIDDAC (Programa de Investimentos e Despesas de Desenvolvimento da Administra\u00e7\u00e3o Central, Portugal). Fernanda F. Roman and Adriano dos Santos Silva acknowledge the national funding by FCT and MIT (Massachusetts Institute of Technology, USA), and the ESF (European Social Fund) for individual research grants with reference numbers of SFRH/BD/143224/2019 and SFRH/BD/151346/2021, respectively.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jece.2022.108143.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  The presence of heavy metals in the environment as a consequence of human activity is an issue that has caught the attention of researchers to find wastewater treatment solutions, such as adsorption. In this work, hydrochars and activated carbon microspheres are prepared from tangerine peels as carbon precursor and FeCl3 as activating and structure-directing agent in the hydrothermal carbonization, allowing to obtain hydrochar microspheres ranging from 50 to 3615\u00a0nm. In addition, a pyrochar was prepared by pyrolysis of the same precursor. The activated carbon shows the highest surface area (S\n                     \n                        BET\n                      up to 287\u00a0m2 g\u20131), but the basicity of the pyrochar (1.83\u00a0mmol\u00a0g\u22121, S\n                     \n                        BET\n                      = 104\u00a0m2 g\u20131) was determinant in the adsorption of Ni, being considered the carbon-based material with the highest uptake capacity of Ni. Isotherm and kinetic adsorption of Ni on the most representative activated carbon microsphere, pyrochar and hydrochar microsphere are assessed by 10 and 7 models, respectively.\n               "}
{"full_text": "There is rising global concern on climate change phenomenon and the eventual depletion of non-renewable fossil fuels [1]. Industrial processes at present rely on steam reforming of methane for generation of synthesis gas for power generation or hydrogen production. Methane, i.e. natural gas is the simplest hydrocarbon molecule, having an energy density of 55.5\u202fMJ/kg [2]. However, concerns over rising anthropogenic carbon dioxide emission into the atmosphere by consumption of fossil fuel reserves including natural gas, oil or coal has made it important to look for alternate fuel sources that are renewable and carbon neutral. Recent years have seen much work devoted to reforming of renewable resources such as biomass and biomass derived liquid fuels which are emerging as important research fields. There are several advantages of using biomass as a renewable source as it is abundantly available in many forms since it is less dependent on location and climate. As biomass absorb CO2 through photosynthesis in the daylight, the use of biomass for energy conversion via biomass gasification results in net neutral CO2 emission [3]. In light of these advantages, there is increased interest in developing biomass gasification as a solution to using biomass as a renewable source for electricity generation.Gasification of solid biomass at high temperature converts it into syngas which is a mixture of CO and H2. The gaseous syngas mixture can be catalytically transformed into industrially useful chemical intermediates and products such as methanol, dimethyl ether, olefin and paraffins or be used for power generation [4\u20137]. There are several configurations to perform biomass gasification such as updraft gasifier, downdraft gasifier and fluidized bed gasifier [8]. These different configurations face a common problem which is the production of tar, a mixture of polycyclic aromatic hydrocarbons compounds due to the incomplete gasification of biomass into syngas. Tar production can lead to severe maintenance problems such as clogging of reactor outlet and other downstream equipment. Moreover, formation of tar as a byproduct causes a loss in energy conversion efficiency of the gasification process since tar consists of significant amounts of energy in the form of mixture of aromatics. Tar also contains polycyclic aromatic hydrocarbons compounds which can be hazardous for human health if disposed into water sources such as rivers and underground water. These constraints have been identified as a hindrance to widespread commercialization of biomass gasification as a commercially viable alternative for sustainable energy production [6,9]. The technical and commercial feasibility of biomass-mediated syngas route for chemical intermediates and product manufacturing hinges on the effectiveness of tar elimination/conversion.For tar removal/conversion, various approaches are presently being reported in the literature, including physical removal, thermal cracking and catalytic conversion [10,11]. The physical removal technique uses either cyclone or ceramic filters to separate the particulate matter. However, after a specific duration of time the problem of pressure drop arises due to blockage of filters by the particles and build-up of pressure drop on the filter cake. Although ceramic filters are useful techniques for the removal of particles, it shows the poor performance for the removal of tar. Under hot gas filtration condition technique, the tar remains in gaseous form and escapes the filter. Therefore, ceramic filtration alone is not a promising technique for hot gas cleaning. However, it can be combined with other separation techniques such as thermal and catalytic cracking/conversion. In few instances wet scrubbers are also used for physical tar removal by impingement of tar on the water droplets. Here, the tar and liquid flows through the decanter and the tar is separated from the aqueous phase. The temperature of the liquid varies from 35 to 60\u202f\u00b0C. Moreover, the energy content of the tars is usually wasted in this process, reducing the overall efficiency of this process, while the post disposal of collected tars is one of the disadvantages of this process. Although physical removal process offers a reliable and effective solution for tar elimination, it is more energy consuming and more expensive to operate compared to other processes. In the thermal cracking process, the tar is converted to smaller non-condensable gases at a high temperature (1000\u20131300\u202f\u00b0C), and the complete conversion to small molecule depends upon the residence time and cracking temperature. Here, a higher air-to-fuel ratio is used to maintain the high temperature at the oxidation zone. The higher concentration of oxygen reduces the tar content in the product gas. Additionally, the heating value of the burnable gas is also reduced. Generally, downdraft gasifier is used for this process due to high temperature requirement. The gas composition after thermal cracking is although very high yet the tar conversion is lower even below the acceptable range. Therefore, there is a need to develop an efficient tar conversion technology to obtain clean syngas. In contrast with thermal cracking reaction, catalytic cracking/conversion of tar at high temperatures (750\u2013900\u202f\u00b0C) is considered as the efficient technology to achieve complete tar conversion and higher product gas composition.To eliminate tar compounds before transporting the syngas to downstream processes, the syngas stream can be either treated within the gasifier itself which is known as primary process, or it can be treated at a separate unit which is known as secondary process. For the primary process, tar is treated with catalyst within the gasifier with steam additive; thus, syngas of higher calorific value can be attained at gasifier outlet. As for secondary process, tar contaminated syngas stream is sent to an absorption tower to absorb tar compound and reduce tar content. Tar contaminated syngas stream can be alternatively sent to a catalytic steam reformer unit to convert tar into syngas which also enhances syngas calorific value. Considering the complexity of real tar, low-cost and disposable catalysts are more likely to be applied in practical gasification plants. Dolomite and olivine mineral are typically used as catalyst in primary and secondary tar elimination process due to high availability and affordability [8]. However, the catalytic stability and activity of dolomite and olivine catalyst have much to improve on. Thus, there has been substantial research on developing synthetic catalyst to attain higher tar conversion and higher catalytic stability. There are works in the literature dealing with the steam reforming of real biomass tars [10,12]. However, the complexity of tar composition makes it difficult to ascertain both the reaction mechanism and the main species responsible for catalyst deactivation. In a typical biomass tar benzene (37.9%), toluene (14.3%), naphthalene (9.6%) and other monocyclic (13.9%) and bicyclic (7.9%) aromatic hydrocarbons are major components, however the composition of individual molecules can be varied with the nature of biomass and gasification conditions [10,12]. Therefore, it is important to focus on the behavior of individual model molecules, usually toluene, benzene, phenol or naphthalene during steam reforming on metal catalysts. Furthermore, since tar is a mixture of organic compounds with different structure and molecular weight affecting product distribution and coke nature, a deeper understanding on the behavior of tar main components and their mixture will be helpful to understand the catalytic tar reforming process.There are numerous review articles has been reported on biomass gasification and catalytic tar reforming process in recent years. For instance, Tomishige et al. had presented a comprehensive review on the development of metal catalysts for steam reforming of tar [13]. Lasa et al. reviewed biomass gasification processes as well as Ni-based catalysts for steam gasification of biomass [14]. Furthermore, the background on tar evolution, main tar precursors and models that simulate tar formation and evolution was also reviewed and according to temperature increments, tar is divided into primary (acetols, and acetic acid), secondary (phenols, and toluene) and tertiary tars (naphthalene) [15]. In another review, the various types of tar reduction catalysts are reviewed by classifying them into minerals and synthetic catalysts according to their production methods [16]. Yung et al. [17] organized and discussed the investigations of catalytic conditioning of biomass-derived syngas with various catalyst formulations and also discussed the roles of catalyst additives. Guan et al. presented a myriad of catalyst support and active metals used for steam reforming of tar as well as some elucidation on the mechanistic aspect of tar reforming [18]. Xiong et al. summarized enhanced performance for the selective CC and CO/CO bond-scission reactions of bimetallic and metal carbide catalysts for biomass derived oxygenate conversion reactions [19]. Liu et al. had presented a comprehensive review on the application of catalysis-plasma system for tar reforming in biomass gasification system [20]. Recently, Hu et al. reviewed various biomass pyrolysis process and the effect of the process design, reactors and catalysts on pyrolysis process was highlighted [21]. In this review, we will be covering recent developments in the catalyst technology for steam reforming of tar model compounds with respect to the influence of their catalytic properties to the catalysis. Also highlighted are the kinetics and reaction mechanism insights for steam reforming of toluene reaction.The seminal concept of adsorption kinetic in heterogeneous catalysis, first introduced by Irwin Langmuir and refined by Sir Cyril Hinshelwood and Sir Eric Rideal and Daniel Eley, has provided the conceptual bedrock for clearer understanding and description of reaction mechanisms on catalyst surface for rational catalyst design [22\u201324]. With clear mechanistic understanding of the reaction, an intrinsic kinetic equation can be first derived, and effectiveness factor can be factored to account for diffusion limitations for more accurate reactor design which is crucial for safe and economical chemical process design. A good example of such application could be found in a highly cited paper by Xu and Froment where 220 experimental runs were performed for catalytic steam reforming of methane reaction. Twenty kinetic models were derived and subjected to non-linear model fitting, based on the results from 220 runs. Out of the twenty kinetic models, only one was found to be statistically significant and thermodynamically consistent [25]. The second part of his work involved determining the catalyst effectiveness factor which was factored in for simulation of steam reforming of methane reactor [26]. However, the reactive intermediate may not be included in the intrinsic reaction mechanism. A microkinetic model can be developed to include the reactive intermediate in a set of elementary reactions that are thought to be relevant in the catalytic chemical transformation. A rate equation in every elemental step contain rate constant for the forward and backward reactions. This rate constant can be determined using DFT calculations under transition state theory. Once the rate constants are determined, the instantaneous concentration of the species can be solved using numerical methods as these elementary steps are often expressed as a system of ordinary non-linear differential equations [27,28].In addition to having a fundamental understanding of the reaction mechanism, an awareness of the state of catalyst surface under actual reaction condition has to be developed. Experimental spectroscopic techniques such as X-ray absorption spectroscopy (XAS), x-ray photoelectron spectroscopy (XPS) and DRIFTS can be utilized ex-situ, in-situ or operando to evaluate the state of catalyst surface. In conjunction with the use of in-situ or operando spectroscopic technique, computational simulation techniques such as density functional theory (DFT) and molecular dynamic (MD) have been extensively utilized to provide theoretical framework to analyze the experimental results obtained from spectroscopic techniques and gain scientific insights on reaction mechanisms and catalytic trends [29,30]. However, establishing structure-function relationship of tar reforming catalyst based on microkinetic modeling, spectroscopic techniques and computational simulation using actual tar is a stimulating challenge as tar is a mixture of polyaromatic compounds. Hence, aromatic compounds such as benzene, toluene and phenol are frequently used as tar model compounds. Due to the complexity of the tar model compounds, several reactions are possible to occur simultaneously. Some of them are as follows:\n\n(1)\n\nSteam reforming\n\n\n\nS\n/\nC\n=\n1\n\n\n:\n\nC\nx\n\n\nH\ny\n\n+\n\nxH\n2\n\nO\n\u2192\nxCO\n+\n\n\nx\n+\ny\n/\n\n\n2\n\n\n\n\nH\n2\n\n\n\n\n\n\n(2)\n\nSteam reforming\n\n\n\nS\n/\nC\n=\n2\n\n\n:\n\nC\nx\n\n\nH\ny\n\n+\n\n2xH\n2\n\nO\n\u2192\n\nxCO\n2\n\n+\n\n\n2x\n+\ny\n/\n\n\n2\n\n\n\n\nH\n2\n\n\n\n\n\n\n(3)\n\nWater\n\u2212\ngas\n\nshift\n:\nCO\n+\n\nH\n2\n\nO\n\u2192\n\nCO\n2\n\n+\n\nH\n2\n\n\n\n\n\n\n(4)\n\nDry\n\nreforming\n:\n\nC\nx\n\n\nH\ny\n\n+\n\nxCO\n2\n\n\u2192\n2xCO\n+\ny\n/\n\n2H\n2\n\n\n\n\n\n\n(5)\n\nSteam reforming of methane\n:\n\nCH\n4\n\n+\n\nH\n2\n\nO\n\u2192\nCO\n+\n\n3H\n2\n\n\n\n\n\n\n(6)\n\nCarbon formation\n:\n\nC\nn\n\n\nH\nx\n\n\u2192\nnC\n+\n\n\nx\n/\n2\n\n\n\nH\n2\n\n\n\n\nIn addition to above reaction, below reactions are also possible for steam reforming of toluene reaction.\n\n(7)\n\nHydrodealkylation\n:\n\nC\n7\n\n\nH\n8\n\n+\n\nH\n2\n\n\u2192\n\nC\n6\n\n\nH\n6\n\n+\n\nCH\n4\n\n\n\n\n\n\n(8)\n\nSteam dealkylation\n:\n\nC\n7\n\n\nH\n8\n\n+\n\nH\n2\n\nO\n\u2192\n\nC\n6\n\n\nH\n6\n\n+\nCO\n+\n\n2H\n2\n\n\n\n\n\n\n(9)\n\nSteam dealkylation\n:\n\nC\n7\n\n\nH\n8\n\n+\n\n2H\n2\n\nO\n\u2192\n\nC\n6\n\n\nH\n6\n\n+\n\nCO\n2\n\n+\n\n3H\n2\n\n\n\n\nSwierzcynski et al. initially hypothesized that toluene was dealkylated from the methyl group to form benzene as the main product. The adsorbed methyl group was then reformed to CO\u202f+\u202fH2. The adsorbed benzene underwent CH bond scission and CC bond scission to form linear hydrocarbon fragments which subsequently transform into CO\u202f+\u202fH2 [31].Benzene, a non-polar compound, can experience intermolecular forces when adsorbed on the metal surface. Recent DFT and surface study have shown that the discrepancy between simulated and experimental values can be reduced when intermolecular forces are accounted for in DFT simulations [32]. Mei et al. performed combined experimental and DFT study for steam reforming of benzene on Rh/MgAlO4 and Ir/MgAlO4 catalysts [33]. The TOF of Rh/MgAlO4 was reported to be higher than Ir/MgAlO4 catalyst due to reduced CC and CH bond scission while a smaller Ir nanoparticle size only affects CC bond scission. The DFT simulation performed in that study suggested that aliphatic CH and CC bonds are both competitive on Rh (111) surface due to similar reaction energy which leads to phenyl and linear C6H6 species. The CH bond scission of phenyl at the meta position is more favorable than para position. The CH bond breaking preference order for ring-opening of linear C6H6 species goes in the order of decreasing endothermicity as follows: ortho\u202f>\u202fmeta\u202f>\u202fpara. CC bond scission of the linear hydrocarbon chain leads to formation of C4H4 and C2H2 acetylene species [33].The findings from benzene as a tar model are also equally applicable to oxygenated aromatic compound such as phenol which is also a major constituent in tar mixture. Spencer et al. experimentally performed phenol adsorption onto Ni (111) surface and compared to DFT results. Heat of adsorption of phenol reduced as surface coverage which indicates that multi-layer coverage is formed on the nickel surface. The measured bond energy is compared with the ones obtained from DFT. The bond energy estimated by DFT is significantly smaller than the experimentally determined values by 87\u201394\u202fkJ/mol. Reasons for significant underestimation are attributed to the presence of van der Waals intermolecular force and possibly the functionals used in DFT calculation. The extent of underestimation is reduced when van der Waals force is included in the DFT calculation to account for long-range intermolecular effects but the magnitude of underestimation remains significant [34]. Nonetheless, this study is an important example on including intermolecular forces in surface DFT calculation and the need to keep improving the DFT method for more accurate estimation of surface bond properties.Comparatively, toluene as a model for tar is much investigated than other model compounds because it represents the stable aromatic structure and apparent in tar formed with high-temperature processes. Quang et al. comparatively investigated toluene decomposition behavior on Ni (111) and boron-doped Ni (111) surfaces and proposed a toluene decomposition mechanism as shown in Fig. 1\n [35]. The \u03c0 electrons in toluene aromatic ring are transferred to metal cations which stabilize the aromatic ring and result in toluene adsorption on the catalyst surface [35,36]. The first step in toluene decomposition starts with CH bond activation of the methyl group (CH3) as the activation energy and reaction energy for this step is lower than aromatic CH dissociation [35]. This is shown in step 1 as CH3 is dehydrogenated into CH2*. Such facile CH bond dissociation of the methyl group is also observed on stepped and stepped-kinked Ni surface under ultra-high vacuum (UHV) condition [37]. This is followed by further dehydrogenation of the CH2* to C*, from step 2 to step 3. CH bond dissociations at the ortho or/and meta position occurs at step 4. From step 5 onwards, ring opening via aryl CC bond cleavage and more CC cleavages occurs to generate shorter hydrocarbon chains which become energetically possible after the ring opening steps.From these computational DFT investigations by other co-authors, toluene decomposition on catalyst surface is clearly a complex phenomenon. However, the effect of water addition on the metal surface, in the presence of toluene, cannot be ignored. Mukai et al. performed in-situ DRIFTS for steam reforming of toluene over Ni/La0.7Sr0.3AlO3\n\u2212\n\u03b4 catalyst in an attempt to elucidate the reaction mechanism for steam reforming of toluene [36]. The rate-determining step is suggested to change at 400\u202f\u00b0C as the absorbance profile at that temperature between the wavenumber 1250 and 1750\u202fcm\u22121 remains unchanged after steam introduction which may suggest that the peaks belong to a reactive intermediate. To determine the identity of reaction intermediate with wavenumber between 1250 and 1750\u202fcm\u22121, different probe molecules such as benzene, n-heptane, ethylene and benzaldehyde are introduced to the catalyst surface. Based on the DRIFTS results after dosing with different listed molecules, the authors concluded that toluene was decomposed into C2 species to form reactive intermediate on Ni/La0.7Sr0.3AlO3\n\u2212\n\u03b4 surface in the presence of steam.Usman et al. performed in-situ DRIFTS and temperature-programmed surface reaction using steam and toluene to elucidate steam reforming of toluene reaction mechanism and identify the reactive intermediate [38]. The presence of aldehyde species is detected from respective CO and CH stretching signals at 1760\u202fcm\u22121 and 2820\u20133170\u202fcm\u22121 and is assumed to be the reactive intermediate. Based on the presence of aldehyde (CHO) intermediate, the reaction mechanism and Langmuir-Hinshelwood type kinetic model for steam reforming of toluene over La0.8Sr0.2Ni0.8Fe0.2 catalyst was proposed and experimentally validated using non-linear regression fitting as shown:\n\n(R1)\n\n\nC\n7\n\n\nH\n8\n\n+\n\ns\n1\n\n\n\n\u2194\n\nK\n1\n\n\n\n\nC\n7\n\n\nH\n8\n\n\u00b7\n\ns\n1\n\n\n\n\n\n\n(R2)\n\n\nC\n7\n\n\nH\n8\n\n\u00b7\n\ns\n1\n\n+\n\ns\n1\n\n\n\n\u2194\n\nK\n2\n\n\n\n\nC\n6\n\n\nH\n6\n\n\u00b7\n\ns\n1\n\n+\nC\n\nH\n2\n\n\u00b7\n\ns\n1\n\n\n\n\n\n\n(R3)\n\n\nC\n6\n\n\nH\n6\n\n\u00b7\n\ns\n1\n\n+\n2\n\ns\n1\n\n\n\n\u2194\n\nK\n3\n\n\n\n3\n\nC\n2\n\n\nH\n2\n\n\u00b7\n\ns\n1\n\n\n\n\n\n\n(R4)\n\n\nH\n2\n\nO\n+\n\ns\n2\n\n\n\n\u2194\n\nK\n4\n\n\n\nO\n\u00b7\n\ns\n2\n\n+\n\nH\n2\n\n\n\n\n\n\n(R5)\n\n\nC\n2\n\n\nH\n2\n\n\u00b7\n\ns\n1\n\n+\nO\n\u00b7\n\ns\n2\n\n\n\n\u2194\n\nK\n5\n\n\n\n\nC\n2\n\n\nH\n2\n\nO\n\u00b7\n\ns\n1\n\n+\n\ns\n2\n\n\n\n\n\n\n(R6)\n\n\nC\n2\n\n\nH\n2\n\nO\n\u00b7\n\ns\n1\n\n+\n\ns\n1\n\n\n\n\u2194\n\nK\n6\n\n\n\nC\n\nH\n2\n\nO\n\u00b7\n\ns\n1\n\n+\nC\n\u00b7\n\ns\n1\n\n\n\n\n\n\n(R7)\n\nC\n\nH\n2\n\n\u00b7\n\ns\n1\n\n+\nO\n\u00b7\n\ns\n2\n\n\n\n\u2194\n\nK\n7\n\n\n\nC\n\nH\n2\n\nO\n\u00b7\n\ns\n1\n\n+\n\ns\n2\n\n\n\n\n\n\n(R8)\n\nC\n\nH\n2\n\nO\n\u00b7\n\ns\n1\n\n+\n\ns\n1\n\n\n\n\u2194\n\nK\n8\n\n\n\nCHO\n\u00b7\n\ns\n1\n\n+\nH\n\u00b7\n\ns\n1\n\n\n\n\n\n\n(R9)\n\nCHO\n\u00b7\n\ns\n1\n\n+\n\ns\n1\n\n\n\n\u2194\n\nK\n9\n\n\n\nCO\n\u00b7\n\ns\n1\n\n+\nH\n\u00b7\n\ns\n1\n\n\n\n\n\n\n(R10)\n\nCO\n\u00b7\n\ns\n1\n\n+\nO\n\u00b7\n\ns\n2\n\n\n\n\u2194\n\nK\n10\n\n\n\nC\n\nO\n2\n\n\u00b7\n\ns\n2\n\n+\n\ns\n1\n\n\n\n\n\n\n(R11)\n\nCHO\n\u00b7\n\ns\n1\n\n+\nO\n\u00b7\n\ns\n2\n\n\n\n\u2194\n\nK\n11\n\n\n\nC\n\nO\n2\n\n\u00b7\n\ns\n2\n\n+\nH\n\u00b7\n\ns\n1\n\n\n\n\n\n\n(R12)\n\nC\n\u00b7\n\ns\n1\n\n+\nC\n\nO\n2\n\n\u00b7\n\ns\n2\n\n\n\n\u2194\n\nK\n12\n\n\n\n2\nCO\n\u00b7\n\ns\n1\n\n+\n\ns\n2\n\n\n\n\n\n\n(R13)\n\nCO\n\u00b7\n\ns\n1\n\n\n\n\u2194\n\nK\n13\n\n\n\nCO\n+\n\ns\n1\n\n\n\n\n\n\n(R14)\n\nC\n\nO\n2\n\n\u00b7\n\ns\n2\n\n\n\n\u2194\n\nK\n14\n\n\n\nC\n\nO\n2\n\n+\n\ns\n2\n\n\n\n\n\n\n(R15)\n\n2\nH\n\u00b7\n\ns\n1\n\n\n\n\u2194\n\nK\n15\n\n\n\n\nH\n2\n\n\u00b7\n\ns\n1\n\n+\n\ns\n1\n\n\n\n\n\n\n(R16)\n\n\nH\n2\n\n\u00b7\n\ns\n1\n\n\n\n\u2194\n\nK\n16\n\n\n\n\nH\n2\n\n+\n\ns\n1\n\n\n\n\n\n\n\nr\n=\n\n\n\n\nk\nf\n\n\n\n\n\n\n\nC\n7\n\n\nH\n8\n\n\n\n\n1\n3\n\n\n\n\nCO\n2\n\n\n\n\n\n\nH\n2\n\nO\n\n\n\n4\n3\n\n\n\n\n\n\nCO\n\n\n4\n3\n\n\n\n\n\nH\n2\n\n\n\n8\n3\n\n\n\n\n\u2212\n\nk\nb\n\n\n\nCO\n\n2\n\n\n\n\n\n\n1\n+\n\nK\nCO\n\n\u00b7\n\np\nCO\n\n+\n\nK\n\nH\n2\n\n\n\np\n\nH\n2\n\n\n+\n\nK\ntoluene\n\n\np\ntoluene\n\n\n\n\n\n1\n+\n\nK\n\nCO\n2\n\n\n\np\n\nCO\n2\n\n\n+\n\n\n\nK\n\n\nH\n2\n\nO\n\n\n\np\n\n\nH\n2\n\nO\n\n\n\n\np\n\nH\n2\n\n\n\n\n\n\n\n\n\n\nThe reaction between the aldehyde species and oxygen species to produce CO2\n(Eq. (R11)) was found to be the rate-determining step for steam reforming of toluene based on close fitting of the parity plot. The apparent activation energy of 109.64\u202fkJ/mol was reported which was attributed by the author to the presence of lattice oxygen [38].Recently, Du et al. has performed steam reforming of toluene over Ni nanoparticles supported on pyrolyzed carbon at 600\u202f\u00b0C and has observed that the TOF for smaller Ni nanoparticles is more than the TOF for larger nickel nanoparticles [39]. The authors assume that there are more stepped surfaces on smaller nanoparticle than larger nanoparticle. DFT simulations are performed by the authors which demonstrate that the toluene adsorption on stepped Ni (111) surface are stronger than flat Ni (111) surface as there are more un-saturated coordinated sites on stepped Ni (111) surface. The authors also proposed that toluene adsorbs parallel to the surface, followed by dehydrogenation of the methyl group in a step-wise manner. Aromatic CH bond cleavage and aromatic ring opening occurs. The ring-opened aliphatic hydrocarbon undergoes further CC bond cleavage to form C3 and C4 fragments.Novel methods for the preparation of heterogeneous catalyst systems endowed with preferred properties are important fundamental research areas. In general, the metal surface constitutes the main active site for the catalytic reforming reaction. Moreover, the surface properties of the catalyst support also play an integral role in defining the reaction mechanism [40\u201343]. For steam reforming of tar, the catalyst should be able to activate both the reactants that are tar and steam. The balance between these two reactions controls the activity and stability of a catalyst. Since tar is heavy aromatics the major catalyst deactivation is identified to be active metal sintering in high steam environment and encapsulation of carbon over active metal centers. In order to address these issues several approaches in developing efficient catalyst has been reported in the literature. By considering the recent reports on steam reforming of tar model reaction, in this section we have categorized the catalysts as role of support, oxygen mobility, basicity and alloying with second metal with respect to catalytic conversion and carbon suppression.Support is the backbone of any catalytic material offering several advantages including strongly binding the metal with the support which prevents leaching out of active metal, provides acidic, basic centers, oxygen vacancy and also improves the performance. Therefore, the choice of support material is crucial for steam reforming of biomass tar reactions. As discussed before, the major challenge in tar model steam reforming reactions includes metal sintering and coke deposition. Under this section, we will be discussing about the role played by different properties of support to overcome the challenge along with discussion on the current development in the new types of catalyst support. Table 1\n summarizes catalytic systems reported for steam reforming of biomass tar model reactions.Alumina is one of the most commonly used catalyst support for nickel catalysts for tar reforming reaction in view of its good mechanical strength and chemical/physical stability, and availability in dispersion of the active metal phase. \u03b1- and \u03b3-Al2O3 are the commonly used alumina phase support for toluene steam reforming [44]. He et al. studied the effect of two phases of alumina (\u03b1 and \u03b3) as a catalyst support for Ni on hydrogen production and formation of carbon nanotubes during the steam reforming of toluene. This study found that the performance of Ni/\u03b3-Al2O3 was superior than Ni/\u03b1-Al2O3. Due to enhanced metal support interaction between Ni particle and \u03b3-Al2O3 led to the base-growth mechanism of CNTs on Ni/\u03b3-Al2O3 in which Ni particles located at the bottom of CNTs, thus the growth of CNTs covered Ni active sites and decreased catalytic activity of Ni/\u03b3-Al2O3 [45]. In order to improve hydrothermal stability under aqueous phase reforming condition, silylation of metal oxide surfaces has also emerged as a promising route to tune the surface properties of inorganic materials, such as silica, zeolites and alumina [46,47]. In other study Artetxe et al. studied steam reforming activities of various biomass tar model compounds using Ni/Al2O3 catalysts and found that the highest carbon conversions and H2 potential for anisole and furfural, while methyl naphthalene gave the lowest reactivity [48]. Among all the model compounds, the amount of coke was higher with oxygenate reactants due to their higher reactivity favoring unwanted reactions that promoted its formation. F. Liu et al. modified Pt/\u03b3-Al2O3 catalyst with silica to prevent the hydrolytic attack, as shown in Fig. 1. \u03b3-Al2O3, silylation followed by high temperature calcination could block the surface Lewis acid Al sites that serve as initial hydration sites for boehmite formation via coordinative saturation and the formation of Al-O-Si bonds, inhibiting water adsorption [49]. The improvement in the thermal stability of Al2O3 with SiO2 doping is also observed by others [49\u201353]. Similarly, M.A. Adnan et al. synthesized Fe2O3/SiO2 doped Al2O3 by one-pot synthesis method and the toluene conversion of about 76% was reported for 10% Fe2O3/SiO2 doped Al2O3 catalyst at 600\u202f\u00b0C. The Fe2O3 accelerates the \u03b3-Al2O3 collapse and transform into other phases, leading to the loss in surface area. And presence of Fe2O3 generates new strong acid sites in Fe2O3/SiO2 doped Al2O3 catalysts. And they have observed that the strong acidic sites promote tar cracking and toluene conversion reactivity. Therefore, the toluene conversion of a catalyst is controlled by the balance between the specific strong acidic sites and surface area of the catalyst [54].Yet in another study, Castro et al. studied the effect of bare Al2O3 and CeO2 doped Al2O3 as a support for Pt metal. The addition of Ce decreases the density of acidic sites as observed from ammonia TPD data due to coverage of acidic sites of Al2O3. Due to this Ce doped Al2O3 showed highest formation of CO2 in comparison to CO due to promotion of higher WGS reaction (Fig. 2\n) [55]. In another study, the Ce/Zr ratio supported on Al2O3 was varied. The study found that for all the catalyst the toluene conversion decreased tremendously during the 22\u202fh operation with no significant coke formation. The decrease in Ce/Zr ratio increased the acid site density and the carbon formation rate. Ce present on the catalyst surface covered the acidic centers of Al2O3 and covered the Lewis acidic centers which inhibit higher coke formation [56]. Many studies were also carried out on hexa-aluminates as catalyst support for various number of reactions [57]. A different approach to modified alumina based catalyst is the development of hexaluminates based support. La, La/Ce and Ca hexa-aluminates were synthesized using co-precipitation method. Among the three doped hexa-aluminates, La/Ce showed the superior performance. For La/Ce the nickel dispersion was independent of Ni loading and showed stable performance for 18\u202fh operation. The Ce promoted catalyst exhibits reduction in the interaction between metal and support thus promotes more metal dispersion [58].Hydrotalcites are another class of support materials which has gained huge interest. Hydrotalcites, when calcined above 450\u202f\u00b0C forms mixed metal oxides which exhibits various advantageous properties for catalytic applications such as large surface, higher metal dispersion and synergistic effect between metal and metal oxide support [59,60]. One of the major issue when Al2O3 alone was used as a support is higher coke deposition and metal sintering. And addition of alkaline earth metals such as MgO and CaO in the form of hydrotalcite precursor have shown improved resistant towards coke formation and metal sintering. K. Tomishige group has studied various mono- and bi-metallic catalysts derived from parent Mg-Al-Ox hydrotalcite precursor for biomass derived tar reforming [61,62]. D. Li et al. studied the effect of composition and reduction pretreatment condition of Ni/Mg/Al catalyst for steam reforming of biomass derived tar. The characterization results showed a nanocomposite was formed between Ni metal surface and Mg(Ni,Al)O particles comparatively smaller Ni particles are formed and lower coke deposition [63]. In conjugation, Ashok et.al studied the hydrotalcites derived from NiO-CaO-Al2O3 oxides and observed the established synergy between NiO and CaO species. The better performance at catalyst composition of NiO-CaO-Al2O3 (8:62:30) was attributed to high basic strength of support, metal dispersion and higher resistance towards metal sintering [64]. Likewise, Lertwittayanon et al. [65] studied the effect of promoting CaZrO3 nanoparticles on Ni/\u03b1-Al2O3. The study varied the loading amount from 0 to 15% and found that the loading of 15% showed the highest activity. Mayenite (Ca12Al14O33) type of support which also consists of CaO has also been an active and stable support for steam tar reforming and CO2 capture.Overall, for steam reforming of tar reactions, Al2O3 was considered as preferred support materials by many researchers due to its availability and ease of catalysts synthesis with various catalyst formulations. The doping of Al2O3 with various base metal oxides such as MgO, CaO, SrO, CeO2, ZrO2 and so on were investigated to suppress or neutralize the Lewis acidic centers which helps for the activation of CH bond in the hydrocarbons and polymerize the carbonaceous species to deposited as coke on the catalyst. Furthermore, hydrotalcites derived catalyst possess great potential in terms of stability and better metal dispersion due to enhancement in the basic properties of the catalysts. Similarly, modifying Al2O3 with redox metal oxides (CeO2, ZrO2) helps to improve the oxygen vacancy in the support thereby suppress the coke formation. Finally, Al2O3 is one of the widely accepted supported material for most of the steam reforming applications.Silica being the inert support offers several advantageous properties such as high surface area, thermal stability and metal sintering resistance due to the ease in the formation of metal silicate [66\u201369] [70\u201372]. There are several modifications adapted in the preparation method to improve the metal support interaction [73\u201376]. A very promising class of silicate, phyllosilicate, is one of the widely accepted class of mineral which unique property playing a dynamic role in catalysis. Phyllosilicate structure derived Ni/SiO2 catalyst has been reported for steam reforming of ethanol (SRE) reactions [77]. The stronger interaction between Ni and SiO2 facilitates partial reduction of Ni species, which in turn provided necessary hydroxyls species to enhance SRE activity and suppressed carbon formation. Recently, there has been extensive research based on utilization of natural minerals materials as a catalyst support being the low cost materials. Ni-phyllosilicate catalyst, this type of support has been deeply investigated by our group [78\u201381]. This study includes investigating the shape effect of phyllosilicate in terms of metal support interaction, coke resistant and metal sintering. Core-shell catalyst was investigated elucidating the effect of silica shell in preventing metal sintering and thus enhancing the stability and performance [82\u201386]. Recently, Li et al. reported the NiCo@NiCo phyllosilicate@CeO2 hollow core shell for steam reforming of toluene as a biomass tar model, as shown in Fig. 3\n. The catalyst showed good catalytic activity and stability for 45\u202fh on stream. The activity was due to strong NiCo interaction and also with the support CeO2. The synergistic effect between Ni and Co also prevented the formation of carbon on the catalyst surface [87].Mesoporous silica in the form for SBA-15 has been widely used for various applications due to high surface area and ordered, uniform hexagonal pores. Modification of SBA-15 with other lanthanide series elements such as La, Ce has been reported to increase the oxygen mobility, inhibits cracking of larger molecules into smaller and improves stability of the support. Usman et al. studied the series of La2O3 modified Ni/SBA-15 catalysts prepared by organic acid assisted synthesis method. As compared to Ni/SBA-15, La doped Ni/SBA-15 showed lower coke formation by forming oxy-carbonates thus making the catalyst stable for >30\u202fh of operation (Fig. 4\n) [88]. In other study Ni support SBA-15 was compared with Ni/Al2O3 catalyst for steam reforming of guaiacol as a tar model. It was reported that the deactivation in Ni/Al2O3 catalyst is due to the encapsulation of amorphous carbon, while the carbon nanotubes grown on the tip of Ni particle and had less impact on catalyst deactivation for Ni/SBA-15 catalyst [89]. Another type of mesoporous support was used in the form of MCM-41. Nickel was impregnated on MCM-41 using ethylene glycol (EG) assisted co-impregnation method. An enhanced metal dispersion with higher metal support interaction was observed when compared with Ni/MCM-41 prepared by conventional wetness impregnation method. The catalyst also showed a good reusability up-to 5\u202fcycles. In terms of H2 yield, the value increased by 8% when using 20\u202fwt% Ni/MCM-41-EG as compared with 20\u202fwt% Ni/MCM-41 for the steam reforming of tar.Besides Al2O3, SiO2 is one of the most commonly used catalyst support for various high temperature reforming applications due to several advantages such as availability, high surface area and thermal stability by forming stable phase such as metal silicates. One of the major issue for silica-based catalysts in high temperature steam reforming applications is the leaching of silica material in high steam reaction environment. Therefore, for commercial steam reforming applications the utilization of silica based catalysts is limited to lower SiO2 content. And there are limited studies were reported for steam reforming of biomass tar model reforming reaction using silica-based materials as catalysts. Recently, Ni/Co-phyllosilicate structured silica based materials seem to be more interesting for steam reforming applications due to the presence of hydroxyl species as structural moieties of the phyllosilicates. And the involvement of hydroxyls species in activating steam during high temperature reforming reaction makes these materials gave stable catalytic performance with reduced carbon deposition. However, the structural stability of silica groups in steam reforming environment is required detailed investigation. Until now, silica based materials are fall behind Al2O3 support for biomass tar reforming applications.Similar to hydrotalcite materials, the oxides derived from perovskite materials are also used as a catalyst and/or catalyst support for steam/CO2 reforming reactions. These kinds of materials known to offer high oxygen vacancies, presence of lattice oxygen which enhances oxidation of hydrocarbons adsorbed on metal, redox property and thermal stability [90]. Mukai et al. carried out series of studies to investigate the Ni support on perovskite catalyst for steam reforming of toluene [91]. However, highest conversion and H2 was achieved by LaAlO3 as compared to other combinations of A and B site elements. Although the performance was highest, the coke deposition on the catalyst was significant. Thus, in order to suppress the coke formation, when A-site was partially substituted with Sr the activity enhanced and the coke deposition decreased approximately 3 times that of un-doped LaAlO3. The Sr doping increased the metal support interaction which was obtained by calcining the catalyst at different temperatures. After calcination at higher temperature the percentage dispersion was decreased which suppressed the catalytic activity.Perovskite structure also consists of active metal ions such as Ni/Co/Fe in the B-site of the crystal. Therefore, it is important to study the effect of perovskite as only support with no active metal present on its structure. Laosiripojana et al. studied the effect of palygorskite, MgO\u2013Al2O3, La0.8 Ca0.2CrO3, and La0.8Ca0.2CrO3/MgO\u2013Al2O3 for Ni and NiFe as active metal sites. Among all the catalyst tested NiFe supported on La0.8Ca0.2CrO3/MgO\u2013Al2O3 shows the highest value of H2 yield and resistance towards coke formation [92]. Alumina is one of the most prominently used B-site elements for reforming reaction due to its acidic nature which helps to break the CC bond. Takise et al. extensively investigated numerous perovskite as a support for cobalt metal including La0.7Sr0.3AlO3\n\u2212\nx (LSAO), La0.7Ca0.3AlO3\n\u2212\nx (LCAO), La0.7Ba0.3AlO3\n\u2212\nx (LBAO), LaAlO3, Sr/LaAlO3, LaAl0.7Zn0.3O3\n\u2212\nx, SrTiO3, SrTi0.7Fe0.3O3\n\u2212\nx, SrZrO3, and SrCe0.5Zr0.5O3\n\u2212\nx [93\u201395]. In this study, the catalysts are categorized at La ion and Sr ion incorporated perovskite oxide. Co/LSAO and Co/LCAO showed highest activity and stability for 300\u202fmin of operation. In Sr-ion category, Co/SrZrO3 performed highest among all other catalyst. In TPR study, the reduction peak for Co/LSAO was larger than Co/LaAlO3 ensuring the higher metal support interaction (CoO); higher dispersion of metal ion was also observed. Another study reported the Ni/La0.7Sr0.3AlO3\n\u2212\nx which showed highest toluene conversion of 67.5% at 650\u202f\u00b0C. The remarkable enhancement in catalytic activity was due to the insertion of Sr which caused the lattice distortion and also helped in suppressing the coke deposition when compared with Ni/La2O3 and Ni/LaAlO3 [96].Perovskite based support has great potential in terms of providing the oxygen vacancy for the steam activation as well as gasifies the carbon formed during the reaction. However, due to low surface area and cost of the synthesis makes it difficult for commercialization. The further modification in the perovskite type of support may be focus on the improvement of surface area and finding the facile and inexpensive route for the synthesis will make this system more viable for the gasification process.Biochar, one of the by-product of biomass gasification process was reported as a potential candidate as support material for tar reforming reaction. Due to high surface area, higher pore volume, thermal stability and availability of surface functional group has greatly attracted attention towards investigating biochar as a catalyst for tar reforming reaction [97]. The char can be activated to carbon and used as a catalyst support, as the activated carbon is considered to be stable under both acidic and basic condition; it is indeed flexible enough in terms of change in textural and chemical properties. The metal dispersion and interaction between metal and carbon also affect the catalytic performance of the carbon based catalysts [98\u2013100]. Liu et al. used biochar as catalyst for tar reforming reaction and it was reported that the tar destruction ability of biochar is influenced by O-containing functional groups in biochar [101]. Qian et al. utilized gasification derived Red cedar char as a support. Before impregnating it with metal the char was activated, then followed by impregnation with nickel precursor. The study includes the effect of nitrate precursor and pre-treatment method. Catalyst prepared by impregnating nitrate precursor showed the highest performance of around 80% conversion at 700\u202f\u00b0C. However, the long term stability test any of the catalyst was not shown. The surface area of spent catalyst decreased and large portion of nickel appeared on the surface due to structural destruction of activated char during reforming reaction [102]. Furthermore, Zhen-Yi et al. prepared Ni/Biochar via simple one-step pyrolytic approach and investigated for steam reforming of toluene at relatively lower temperature of 600\u202f\u00b0C (Fig. 5\n). A highly dispersed nickel particles were obtained on the supported catalyst. This was shown due to the presence of high porosity and also the presence of abundant amount of surface oxygen on the raw biomass. The nickel particle size played crucial role in enhancing the catalytic activity and stability which showed the smaller nickel particle (4.2\u202fnm) gave a turn over frequency of 1.64\u202fs\u22121 [103]. Additionally, the pyrolysis temperature played more significant role in forming smaller nickel particle sizes [39].Another strategy to improve the mechanical stability and chemical resistibility of activated char was studied by J.-P. Cao et al. [104]. In this paper, a novel porous carbon catalyst was prepared by inserting metallic nickel in D151 resin through ion exchange. The result showed that Ni dispersion depends on two parameters, pH and carbonization temperature. Ni/resin char showed higher performance as compared to commercially available catalyst Ni/Al2O3. As compared with other types of char such as bio-char and lignite char, in Ni/RC (resin char) the metal is embedded in a definite chemical structure and thus showed higher potential to be used in catalytic biomass gasification [104]. Bimetallic (NiRe) supported on sewage sludge char was prepared and investigated its performance cracking and reforming of biomass tar. At optimum temperature of 800\u202f\u00b0C, the 3%Re-7%Ni/Char showed the highest conversion and H2 yield, and the performance was stable for 18\u202fh of operation with only 5% coke deposition [105].A different type of char studied is the hydro-char. It is considered as an ideal support for the synthesis of nickel based nanomaterials derived from hydrothermal carbonization of carbohydrate rich biomass [106]. C. Gai et al. proposed and synthesized one pot hydrothermal carbonization of metal on hydrochar to obtain smaller size Ni and NiFe metal, as shown in Fig. 6\n. With the help of characterizations such TEM, the author elucidates factors that affect the structure, composition, and size of the nanoparticles [107].Biochar based materials as catalysts for tar reforming reaction is an emerging area which is the least expensive material and versatile route to modify the catalyst. Owing to the high surface area, this type of support increases the availability of active sites for the reaction. Yet, the thermal and chemical stability of char is questionable which decreases the activity during the reaction. Recently, many researchers are focussing on this class of catalyst materials especially for commercial biomass gasification processes.Apart from mostly studied Al2O3, SiO2, perovskite and char based support used for this reaction, there are several other types of support (mixed oxide, core-shell, natural minerals) which have shown promising behavior will be discussed in this section. It has been reported in the literature that Ni/TiO2 catalyst formed a mixed oxide (NiTiO3) when calcined at high temperatures (i.e. 800\u202f\u00b0C) [108]. After activation, NiTiO3 system enabled the coexistence of smaller particles which were more active and resistant to deactivation and sintering during SRE reaction. Similarly, the strong metal-support interactions could also be obtained between Ni species and sol-gel synthesized Al for Ni/Al2O3 catalysts which were calcined at high temperatures [109].Core shell like catalyst has been extensively investigated for various thermal reactions including reforming reactions [110]. In one study [111], confinement effect observed in Ni@ZrO2 core-shell catalyst prevented metal sintering, enriched surface active oxygen content and widened metal\u2013support interfacial perimeter, thereby aiding the removal of carbon deposits during SRE reaction. Core shell like structure, Ni@Al2O3 was studied by Qian et al. (Fig. 7\n). The stable shells provide the unique environment around active sites and the strong interaction between Ni and the core\u2013shell supports seems to be responsible for the catalyst activity and stability in toluene steam reforming [112].Similarly, X. Zou et al. studied the inexpensive and earth abundant materials known as Palygorskite (Pal). Fe3Ni8 was impregnated on this support Fe3Ni8/Palygorskite catalyst with high dispersion was successfully prepared and exhibited superior catalytic performance compared with those of the monometallic catalysts (Fe3/Palygorskite and Ni8/Palygorskite) and the bare Palygorskite. Olivine is a natural occurring mineral in the form of (MgxFe1\n\u2212\nx)SiO4. This support has been widely used for biomass steam reforming [113,114]. From all these studies, it was found that the presence of olivine in the reactor offers following advantages, decreases the tars content, increases syngas yield, and a decrease in the amount of CH4. Another preferred aspect of Olivine is due to the high sintering resistance compared with other nature ore catalysts. J. Meng et al. reported the performance of iron supported on olivine prepared via thermal fusion reaction. The effect of thermal fusion was studied and it was found that a higher thermal fusion temperature (1400\u202f\u00b0C) enhanced the interactions between Fe and olivine supports. This lead to the formation of a new phase, (Mg, Fe) Fe2O4 which promoted the increase in a high H2 yield and also showed higher resistance towards coke formation [115]. This author reported a study where Ni was used as a major element supported on Olivine prepared via thermal fusion technique where the toluene conversion of 99.6% was achieved [116]. This study was assisted by various range of characterization to understand the physicochemical properties of both support and catalyst. Only Fe/Olivine showed higher toluene conversion whereas Fe-NI/Olivine prevented coke formation with the highest toluene conversion of 98.44% and carbon formation of 7% after 45\u202fh of reaction. Yet in another study, Fe/Olivine [117,118] strong interaction between iron and olivine, led to the attainment of an equilibrium between Fe0/Fe2+/Fe3+ in partially reduced catalyst that was proposed to be responsible for the catalyst activity and stability.Mixed oxides were also reported for biomass gasification process. Herein, various types of supports were investigated. The enhancement in catalytic performance of Ni species in Ni supported over MgO-Al2O3 [119,120], MgO-CaO [121,122], and CeO2-ZrO2 [123,124] was mainly attributed to intimate interaction between Ni and supports. In another study, cheaper and ecofriendly source of support materials can be derived from Municipal solid waste (MSW) incineration process, which consists of wide range of metal and metal oxides to be used as catalyst [125]. According to them upon hydrothermal treating with 3\u202fm NaOH, the incinerator bottom ash (IBA) act as better catalytic support than un-treated IBA. And the nature of the support is greatly influenced by the time of hydrothermal treatment.Naturally occurring and waster derived materials are available in abundance promising properties as a catalyst or support. This type of supports can be modified by means of synthesis method or treatment under higher temperature. Few drawbacks of these materials include low surface area, and complex metal/metal oxide composition. The determination of exact properties of these materials which leads to catalytic activity is a challenge. Therefore, the availability of in-depth characterization techniques is in scarce.The oxygen storage capacity and oxygen mobility of the catalyst support are important parameters that affect the activity and stability in reforming reactions. The presence of oxygen vacancies and mobile oxygen species in the support structure can enhance the activation of steam in steam reforming reaction and increase the transport and supply of oxygen to the active sites to oxidize active carbon species to CO. Greater amount of oxygen vacancy and higher oxygen mobility are crucial for surface carbon gasification since they promote the prevention of carbon deposition. In fact, oxides with high surface oxygen mobility aid in H2O activation, promotion of water gas shift reaction as well as the oxidative elimination of surface bound carbon present on Ni nanoclusters [40]. This ultimately improves the selectivity, activity and stability of the catalyst. Moreover, for higher hydrocarbons such as tar, the undesirable coke deposition is due to adsorption and desorption of unsaturated hydrocarbons and tars on the catalyst surface [38]. Reducible supports like perovskite oxides and ceria are known for their lattice oxygen mobility. In perovskite (ABO3) oxides, the partial substitution of A or B site metals by other elements with different valence or cationic radius gives rise to structural defects in the lattice, that generates oxygen vacancies and mobile oxygen species [38,96,126]. On the other hand, CeO2 can undergo redox cycles between CeO2 and CeO2\n\u2212\nx under the steam reforming atmosphere due to its unique property of being able to stabilize both Ce4+ and Ce3+ ions in its fluorite structure. The formation of partially reduced CeO2\n\u2212\nx gives rise to oxygen vacancies that can activate steam and absorb the active oxygen species, which can thereafter diffuse to the active metal-ceria interface and participate in the reaction. In the following section, the effect of oxygen mobility of supports in steam reforming of tar will be discussed with respect to type of materials.CeO2 is well-known for its redox nature and oxygen storage capacity (OSC). The CeO2 support can undergo rapid redox cycles and accept and release lattice oxygen in a reaction [127,128]. Addition of CeO2 to other supports has hence shown improved activity and coke resistance in steam reforming of tar. For example, Ashok and Kawi [129] reported that addition of appropriate amount of CeO2 to Ni/Ca-Al catalysts enhances activity in SRT reaction and suppresses carbon deposition by promoting the oxidation of carbon precursors deposited on nickel surface by supplying necessary oxygen species. It is also worth mentioning that the redox nature and OSC of CeO2 makes it highly active for water gas shift reaction, which in turn produced higher values of H2/CO for Ni/Ca\u2013Al\u2013Ce catalysts than Ni/Ca\u2013Al catalyst.The incorporation of ZrO2 in CeO2-ZrO2 mixed oxides has been shown to increase the concentration of oxygen vacancies and mobile oxygen species in the support, and promote the support reducibility, leading to higher stability in reforming reactions [127]. Maia et al. studied the effect of doping 25%, 50% and 75% Zr in Ni/CexZr1\n\u2212\nxO2 and observed that the highest stability was observed on Ni/Ce0.5Zr0.5O2 in steam reforming of glycerol. OSC measurements revealed that a doping of 25\u201350% Zr in the CeO2 structure resulted in highest oxygen storage capacity, which could be correlated with the higher stability and lower coke formation on these catalysts [130]. Doping ceria with rare earth metal oxides like samaria, gadolinia have also been shown to increase activity and coke resistance in tar reforming by increasing the oxygen storage capacity. For instance, Laobuthee et al. observed that doping a small amount of Sm in CexSm1\n\u2212\nxO2 (without any metal doping) increases the reducibility and oxygen storage capacity of the oxide and increases the hydrogen yield from steam reforming of toluene from 20.2% on CeO2 to 32.8% on Ce0.85Sm0.15O2 [131]. Higher Sm doping leads to a fall in the oxygen storage capacity, possibly due to the inability to form a pure solid solution structure, and a concurrent drop in activity is observed. In another study [132], it was observed that incorporation of Mn into Ce0.75Zr0.25O2 mixed oxide has the potential to modify its redox properties by introducing more structural defects and oxygen vacancies and increasing the oxygen mobility. Upon impregnation with Ni, the presence of Mn was found to vividly decrease carbon deposition in the steam reforming of tar (using naphthalene as the model compound) by promoting surface carbon gasification and/or water gas shift reaction. Castro et al., however, reported a different trend for Pt/CeO2-ZrO2/Al2O3 catalysts where an increase on oxygen vacancies and oxygen mobility due to the addition of ZrO2 in CeO2 did not benefit the coke resistance of the catalyst. Instead, the doping of ZrO2 increased the core formation tendency in steam reforming of toluene, which was attributed to an increase in acid sites that catalyzed the oligomerization of toluene into heavier aromatics and carbon deposits. The authors inferred that the increase in oxygen storage capacity by zirconia addition did not have a beneficial effect in suppressing coke formation [133].Addition of Fe in Ni catalysts has also been shown to improve activity and coke resistance in steam reforming of tar by increasing the amount of lattice oxygen and adsorbed oxygen on the catalyst surface. Fe in the NiFe alloy can easily be converted to FeOx by steam, which can then supply lattice oxygen to oxidize carbon and other intermediates on Ni to form CO [134,135]. Ashok et al. showed that surface enrichment of iron species on nickel in Ni/Fe2O3-Al2O3 catalyst enhanced the conversion of toluene and suppressed coke formation by increasing the coverage of oxygen species on the catalyst active sites [136]. Similarly, Sun et al. studied the reaction process of different oxygen species in a series of Ni-Fe/Al2O3 catalysts during steam reforming of toluene. Based on a H2O/toluene pulse study, they proposed that adsorbed oxygen from iron species participates in oxidation of intermediates from toluene decomposition to produce CO/CO2 through a Langmuir-Hinshelwood mechanism and the lattice oxygen in FeOx oxidizes the carbon deposition or intermediates following a Mars-van Krevelen mechanism [137].Perovskite oxides are well-known for their oxygen transport capabilities through the mobility of lattice oxygen, spurring their application in oxygen permeable membranes. It has been shown that introduction of heteroatoms in the A or B sites of ABO3 perovskite oxides cause a lattice distortion and increases oxygen vacancies, leading to an improvement in redox ability and oxygen mobility. Oemar et al. studied the effect of Sr doping in the A site of LaAlO3 supported Ni catalysts and observed that Ni/La0.8Sr0.2O3 catalyst showed superior catalytic performance both in terms of activity and coke resistance in steam reforming of toluene [96]. Using XRD, TPR, XPS and O2-TPD, they showed that the superior catalytic performance of the Ni/La0.8Sr0.2AlO3 catalyst was a result of lattice distortion caused by strontium doping, which produced a higher concentration of oxygen vacancies on the catalyst surface. This lowered the activation energy of the migration of lattice oxygen, enhancing the mobility of lattice oxygen species which favored the direct partial oxidation of toluene, and also improved the adsorption abilities of gas phase oxygen species. Sekine et al. has reported that the suppression of coke and enhanced activity is possible when A-site in LaAlO3 perovskite is partially substituted with Sr oxide. The decrease in coke formation is approximately 3 times that of un-doped LaAlO3. Sr doping enhanced the mobility of lattice oxygen on the perovskite as shown by transient response test using H2\n18O [138]. The Sr role and mechanism was further investigated by transient isotopic technique [95]. The role of lattice oxygen and metal support interaction was obtained by calcining the catalyst at different temperature and it was found that higher the temperature for calcination the percentage dispersion was smaller and also the catalyst activity decreased from 72.9% (750\u202f\u00b0C) to 15.5% (1100\u202f\u00b0C). Similarly, Mukai et al. demonstrated with transient response techniques using H2\n18O that the surface lattice oxygen of La0.7Sr0.3AlO3\u2212\u03b4 worked as active oxygen by redox mechanism at 600\u202f\u00b0C and oxidized the adsorbed CHx species on the Ni surface directly, forming CO in steam reforming of toluene [139]. Arrhenius plots showed that the rate-determining step of the SRT reaction changed at a certain temperature at which the lattice oxygen was able to contribute to reaction over Ni/La0.7Sr0.3AlO3\u2212\u03b4 and Ni/LaAlO3. The same authors also showed that the lattice oxygen mobility of the La0.7Sr0.3AlO3\u2212\u03b4 support could be enhanced by increasing the Ni particle perimeter or the Ni/support interface area. By changing the Ni particle size through different calcination temperatures, it was observed that the catalytic activity of the catalyst could be correlated to the Ni particle perimeter on the support instead of the metallic surface area, and carbon decomposition increased concomitantly with a decrease in this Ni perimeter [91]. The presence of oxygen vacancies in the La0.7M0.3AlO3\u2212\u03b4(M\u202f=\u202fSr, Ca, Ba) support was also postulated to create an anchoring effect on Co nanoparticles that stabilized them and prevented sintering under toluene reforming conditions [140]. The substitution of alkali metals was shown to create a more reductive La ion in the lattice, and the fixed oxygen species near reductive La is expected to be effective in Co metal anchoring. High oxygen release rate was measured on the alkali metal substituted perovskites (Fig. 8\n).Substitution of lanthanide based perovskites with redox elements such as ceria in the perovskite lattice structure can also enhance oxygen storage capacity which ultimately leads to suppression of sintering and formation of char on the surface [141,142]. For instance, toluene steam reforming activity was enhanced due to Ce substitution in La0.6Ce0.4NiO3 catalyst [143]. Recently, the high oxygen vacancy and mobility of Ce substituted perovskites have also been shown to increase the catalyst stability and resistance to sulfur poisoning, which is an important challenge in biomass tar processing [144]. In a series of LaxCe1\n\u2212\nxCo0.5Ti0.5O3 catalysts, it was observed from XPS that La0.8Ce0.2Co0.5Ti0.5O3 possessed the highest percentage of mobile oxygen species on the surface. The La0.8Ce0.2Co0.5Ti0.5O3 catalyst showed the highest toluene reforming activity and the best regeneration capacity after exposure to H2S. Sulfur poisoning in steam reforming of toluene mainly occurs by the conversion of the active metal into inactive sulfides and the catalytic activity can be partially restored after stopping sulfur in feed depending on the stability of the sulfide formed. It was observed that compared to other compositions, the La0.8Ce0.2Co0.5Ti0.5O3 catalyst could achieve almost 90% of its original activity after exposure to H2S, which is possibly due to its better O2 mobility nature, which makes it possible to oxidize the sulfides to respective sulfur oxide species during SRT reaction post exposure to H2S.Besides oxygen mobility, catalyst basicity is also an important property which is desirable for the breaking of CC bonds in the hydrocarbon as well as suppressing carbon formation. In steam reforming reactions, basic sites also activate steam to generate hydroxyls and these hydroxyls have the ability to reduce the carbon deposition over metal centers. Despite the fact that acidic supports such as alumina have been widely used for reforming reactions owing to their low cost and high surface area, catalyst deactivation remains a prevalent issue on alumina supported catalysts. This is inevitable due to the acidic nature of alumina which favors undesirable parallel side reactions such as the dehydration of oxygenates to olefin intermediates, eventually causing carbon formation on the surface of the catalyst [1,145]. Improving the catalyst basicity are both essential and beneficial in minimizing these side reactions. Besides, presence of basic sites can promote non-oxidative cleavage of CH bonds [146]. In order to improve the basicity of a catalysts several approaches such as change in the synthesis method to establish interactions between metal and basic oxide supports, doping with base metal oxides and so on are reported. Among the several approaches, doping the metal and/or support with base metal oxides is much explored for various reforming reactions especially steam reforming of tar model compounds. The most commonly added metal oxide can be either from alkali or alkali earth metal oxides series. Besides, optimum amount of doping with other metal oxides such as CeO2, ZrO2, MnO2 and Y2O3 also showed enhanced basicity property of a material [147]. With this scenario in this section, we will discuss categorically the influence of change in the basicity of a catalyst for steam reforming of tar model reactions with respect to nature of dopants. Most commonly reported base oxide promoted catalytic systems for biomass tar model reforming reaction are detailed in Table 2\n.Alkali metals are metal belonging to group 1A of the periodic table such as lithium (Li), sodium (Na), potassium (K). Many studies proved that alkali metal catalysts are very effective in steam reforming of tar and can improve the quality of gaseous product [148,149]. However, the major disadvantage of these catalysts is their evaporation during the reaction and difficulty in recovery. In one work K-doping to Ni/dolomite catalyst was compared with Ca and MnO2 doping for steam reforming of toluene activity. Comparatively, doped catalysts exhibited higher toluene conversion with improved carbon suppression than reference Ni/dolomite catalysts [150]. In another work, Moud et al. reported the effect of K coverage on Ni/MgAl2O4 catalyst for sulfur-laden tar reforming reaction [151]. According to them the reason for improved reforming activity for both methane and tar reforming was related to K adsorption, lowering the surface coverage of S at active sites due to weakening of NiS bonds. Furthermore, in another study, K. Yip et al. prepared biochar from pyrolysis of leaf, wood and bark in a fixed bed reactor at 750\u202f\u00b0C. The result indicated that the presence of Na, K, and Ca in bio-chars played key role in enhancing the catalytic activity with the order of K\u202f>\u202fNa\u202f>\u202fCa present during the steam gasification of biochar [152]. Similarly, Feng et al. reported K and Ca-loaded biochar catalysts for reforming of various model compounds (such as toluene, naphthalene and phenol). At 800\u202f\u00b0C, the release of K from biochar samples is nearly twice of Ca during the tar model reforming with 15% H2O or pure CO2 [153]. More O-containing functional groups are formed on K-loaded biochar than on Ca-loaded and H-form biochars. Recently, Ashok et al. reported the acidity of incineration bottom ash (IBA) derived from municipal solid waste can be neutralized and Lewis basic centers can be generated by hydrothermally treating with 3\u202fM NaOH solution for 8\u201324\u202fh. The treated IBA act as better support material for Ni catalysts for steam reforming of toluene reaction. Both toluene conversion and carbon suppression was enhanced significantly than un-treated Ni/IBA catalyst [125]. Furthermore, as mentioned in previous sections, production of higher H2/CO values than stoichiometric values for steam reforming of tar is inevitable due to the involvement of water-gas-shift (WGS) reaction. And it is well reported in the literature that the doping of small amounts of alkali metal oxides to supported catalysts significantly enhances WGS activity and suppressed methane formation [154\u2013156]. Thus the role of alkali metal oxide in promoting steam reforming of tar activity is greatly associated with their activity towards WGS reaction [157]Among all the base metal oxides, the alkali earth metal oxides such as MgO and CaO are extensively investigated for steam reforming of biomass tar reaction [158\u2013160]. It is due to the availability of these oxides, which are also major components of many naturally available minerals such as olivine, dolomite, calcite, phyllosilicates and so on. Furthermore, these oxides are used as dopants to improve the basicity of the material and stabilize the metal and support materials during high temperature reforming environment. Besides, SrO and BaO are also used as other alkaline earth metal oxide dopants for reforming applications [161]. By considering the uniqueness of these meatal oxides, the researcher selectively investigated them for various types of catalytic materials. For instance, the role of MgO in tar reforming activity is majorly studied using Mg-Al-Ox hydrotalcite derived catalyst precursors, where Mg2+ species was partially or completely replaced with other catalytically active M2+ species such as Ni2+, Cu2+ and Co2+, and Al3+ with Fe3+ species [135,162]. A uniformly distributed mono and bimetallic catalytic systems can be obtained via decomposition of hydrotalcite precursors.Similarly for steam reforming of toluene reaction, Ashok and Kawi [163] found that doping CaO to a NiFe alloy (at the optimal molar ratio of 1.5:1:2) supported on iron-alumina catalyst, resulted in activation of the water molecules at lower temperatures, resulting in enhancement of steam reforming of toluene at lower temperatures as well as suppression of carbon deposition. The function of CaO in the hydrotalcite structured NiO-CaO-Al2O3 was also explored by Ashok et al. [64] for the steam reforming of toluene reaction. CO2 sorption studies were conducted in order to elucidate the basic strength of the catalysts, and Ni-Ca-Al catalysts with the composition of Ni:Ca:Al\u202f=\u202f8:62:30, was found to display substantially stable CO2 sorption behavior for up to 10 carbonation and de-carbonation cycles. The authors correlated the synergism between the active Ni phase and CaO phase in Ni-Ca-Al catalysts as the main factor which resulted in promising performance for the steam reforming of tar reaction. Similarly, catalytic bi-functional material Fe/CaO-Ca12Al14O33 was also reported for steam reforming and WGS reaction, where the iron phases favor the H2 production and the CaO-Ca12Al14O33 phase simultaneously captures CO2 during several cycles of carbonation-decarbonation [164,165]. In another study promotional effect of CaO doping to NiFe bimetallic catalyst was reported for steam reforming of toluene reaction. According to authors, the defect sites in CaO could be helping in H2O adsorption, which can react with adjacent carbon on Ni surface at their interface [166]. Likewise, enhancement in the Ni activity and stability from the individual roles of CaO and CeO2 in 2.5\u202fwt% of Ni20Ca60Ce20 catalyst was also reported for reforming reaction, where CeO2 can act as redox component that provides for a higher H2O dissociation leading to carbon oxidation and the CaO can enhance the basicity of the catalyst as well as provide for a higher dispersion of Ni metal [167].In another study on Ni/CaAlOx for steam reforming of biomass, the authors varied Ca/Al ratio and observed increase in Ca loading increased the CO formation and also reduced CO2 formation, however the Ni particle size was sacrificed [168]. Effect of calcium addition was recently studied by Jin et al.; the catalyst was modified by incorporating Ca into the Ni-MgO-Al2O3 catalyst which was effective for enhancing hydrogen production. This was explained to be due to enhanced adsorption of CO2 on the CaO promoting the WGS reaction [169].Likewise, Oemar et al. [126] found that substitution of a small amount of Sr with La in the LaxSr1\n\u2212\nxNi0.8Fe0.2O3 (LSNFO) perovskite structure for steam reforming of toluene, especially at low steam/carbon ratio, led to good catalytic activity and stability (Fig. 9\n). This was attributed to the inherent property of Sr in the perovskite structure which can strongly adsorb water. Similar behavior of enhanced catalytic performances at low steam/carbon ratio using basic additives like CaO [170], SrO [95,171\u2013173], MgO [174], SmO [131], BaO [58,175] was also reported for steam reforming of toluene reaction.Furthermore, Takise et al. extensively investigated numerous perovskite as a support for cobalt metal including La0.7Sr0.3AlO3\n\u2212\nx (LSAO), La0.7Ca0.3AlO3\n\u2212\nx (LCAO), La0.7Ba0.3AlO3\n\u2212\nx (LBAO), LaAlO3, Sr/LaAlO3, LaAl0.7Zn0.3O3\n\u2212\nx, SrTiO3, SrTi0.7Fe0.3O3\n\u2212\nx, SrZrO3, and SrCe0.5Zr0.5O3\n\u2212\nx\n[\n\n93\u201395\n\n]\n. In this paper, the catalysts are categorized at La ion and Sr ion incorporated perovskite oxide. In La ion category, Co/LSAO and Co/LCAO showed highest activity and stability for 300\u202fmin of operation. Whereas, in Sr-ion category, Co/SrZrO3 performed highest among all other catalysts. The anchoring effect between Co and support was investigated using STEM. From H2\n18O SSITKA, the Co/LSAO showed higher lattice oxygen release rate than that of Co/LCAO or Co/La0.7Ba0.3AlO3\n\u2212\nx. Also, BaO substituted Ba\u2013Ni-hexaaluminate (BaNixAl12\n\u2212\nxO19) was tested for steam-reforming of 1-methylnaphthalene. The BaNi substituted hexaaluminates show ~90% conversion at 900\u202f\u00b0C for tar cracking, and also showed activity for water\u2013gas-shift reaction [176,177].In addition to alkaline and alkaline earth metals, the doping with other metal oxides also gave improved basicity. For instance Abou Rached et al. [178] observed that the addition of cerium to Ni2Mg2Al4 enhanced the formation of surface carbonates which is due to the enhanced the redox nature, facilitating the CO2 transport to Mg oxides/hydroxides basic sites. In another study, the Ce/Zr ratio supported on Al2O3 was varied. The study found that for all the catalyst the toluene conversion decreased tremendously during the 22\u202fh operation with no significant coke formation. The decrease in Ce/Zr ratio increased the acid site density and the carbon formation rate. Ce present on the catalyst surface covered the acidic centers of Al2O3 and covered the Lewis acidic centers which inhibit higher coke formation [56]. In another study Mazumde and De Lasa [179] observed that the addition of La2O3 up to 5\u202fwt% to Ni/Al2O3 catalyst improved surface area, CO2 adsorption capacity, Ni reducibility and metal dispersion, as well as reduction in support acidity and improves basicity. Mayenite (Ca12Al14O33) type of support which also consists of CaO has also been as an active and stable support of CaO catalysts for steam tar reforming and CO2 capture [180]. E. Savuto et al. investigated the effect of Ce addition to the mayenite. The Ce doping on the catalyst was very useful as the activity for undoped Ni/Mayenite was better even in the presence of sulfur. However, the Ce doped catalyst showed lower coke deposition [181]. Likewise, Lertwittayanon et.al [65] studied the effect of promoting CaZrO3 nanoparticles on Ni/\u03b1-Al2O3. The study varied the loading amount from 0 to 15% and found that the loading of 15% performed the highest. This was attributed to oxygen vacancies present on the catalyst surface which enhanced the steam adsorption and desorption; this step led to the formation of additional amount of H+. This additional H+ increased the H2 yield and prevented coke formation by gasification process.Improvement in metallic dispersion is an important factor to constrain metal sintering which generally occur at high temperature reforming activities. Nickel based catalysts and especially bimetallic catalysts are found to exhibit higher activities than monometallic catalysts and this phenomenon is attributed to the co-existence of the well dispersed metals [182]. In view of this, alloying or bimetallic catalyst synthesis can effectually enhance resistance to metal sintering [20,183\u2013186]. The physicochemical properties of the alloyed metals generally correlate to their composition, atomic ordering and particle size [40]. In general, volcano-type relationships between the catalyst compositions and the correlating catalytic performances have been observed. However, it is rather challenging to control composition and size of each particle in alloy nanoparticle system [187]. In this section, we mainly discuss the effect of bimetallic alloy on the catalytic performance in steam reforming of biomass tar and its model compounds in detail (Table 3\n). Throughout the literatures, Ni, Fe, and Co are the most common and promising alloying elements utilized in steam reforming of biomass tar due to their high catalytic activity and coke resistance resulting from the synergistic effect.Among bimetallic catalysts, NiFe alloy catalysts have been extensively studied for steam reforming of biomass tar and its model compounds. Oemar et al. [188] reported the synergistic interaction between Ni and Fe in forming bi-metallic NiFe particles in the perovskite oxide catalyst which conferred high activity and stability for the steam reforming of toluene. The presence of LaNi0.8Fe0.2O3 (LNFO) catalyst resulted in about 30\u201350% higher hydrogen production compared to the monometallic-based LaNiO3 (which showed a continual decreasing trend in hydrogen production). Moreover, the carbon deposition rate of the bimetallic LNFO catalyst was almost half of the monometallic LNO catalyst. The authors attributed the enhanced catalytic performance of the Fe substituted LNFO catalyst to the synergy between the Ni and Fe atoms on the smaller NiFe bimetallic particles. Moreover, the strong interaction formed between the metal and support further prevented metal sintering. Furthermore, Ashok and Kawi [136] reported a toluene conversion of >90% for a period of 26\u202fh over Ni/Fe2O3\u2013Al2O3 catalyst at 650\u202f\u00b0C. According to XRD analysis, NiFe alloys were formed and stable throughout the reforming reaction. The surface Fe species played the role of co-catalysts by increasing the coverage of oxygen species during the reforming reaction to enhance the reaction of toluene and suppress coke formation (Fig. 10\n).Tomishige et al. [187] reported that NiFe alloy particles with uniform compositions could be produced via hydrotalcite-like precursors in Ni-Fe/Mg/Al catalysts. An optimum alloy composition of supported catalysts (Fe/Ni\u202f=\u202f0.25) generally showed better catalytic performance over individual metal supported catalysts for biomass tar steam reforming reaction. Furthermore, they investigated the regenerability of hydrotalcite-derived Ni-Fe/Mg/Al bimetallic catalysts [61]. The behavior of NiFe alloy nanoparticles on Ni-Fe/Mg/Al catalyst during the oxidation-reduction treatment is illustrated in Fig. 11\n. Upon the oxidation, NiFe alloy nanoparticles are oxidized and incorporated into the near surface of Mg(Ni, Fe, Al)O periclase. Upon the subsequent reduction, the uniform NiFe alloy nanoparticles are regenerated. In a follow-up study [189], the same group also examined the catalytic performance of Ni-Fe/Mg/Al catalysts derived from hydrotalcite towards the steam reforming of tar model compounds including benzene, toluene and phenol. Similarly, Ni-Fe/Mg/Al (Fe/Ni\u202f=\u202f0.25) catalyst showed greater activity and effectively suppressed the carbon deposition than the one without Fe addition, specifically in the case of steam reforming of toluene and benzene. Nonetheless, carbon deposition over Ni-Fe/Mg/Al catalyst in steam reforming of phenol was relatively high due to the strong adsorption of phenol on both Fe and Ni sites and its successive decomposition to carbonaceous species.In Fig. 12\n, environmentally friendly NiFe alloy catalysts supported on olivine were also investigated in steam reforming of phenol and naphthalene as tar model compounds [190,191]. By using thermal fusion (TF) method, Fe and Ni were partly fused into the structure of olivine existing in the form of (Mg, Fe) Fe2O4 and Ni2SiO4 (NiFe2O4), respectively. After reduction, active metal Fe, Ni and NiFe alloy particles uniformly dispersed on Si, Mg, O phases. NiFe alloy particles sizes on TF-Ni/Fe/olivine were thus smaller compared to those on Ni/Fe/olivine prepared by wetness impregnation method. The NiFe bimetallic catalysts supported on olivine performed well in phenol steam reforming and exhibited good stability in the initial stages of stability test. The stable active sites were attributed to strong interaction between NiFe alloys and olivine support. It was also found that the deposited carbon from phenol steam reforming consisted of C\u03b1, which was easily eliminated by steam. In contrast, C\u03b2 and C\u03b3 with higher graphitization degrees were detected in the carbon deposits from naphthalene steam reforming. They were more difficult to react with steam, and hence naphthalene mostly underwent the catalytic cracking process.A majority of NiFe bimetallic catalysts are acquired from specific precursor structures such as perovskite, hydrotalcite, spinel and olivine. The synergistic effect between Ni and Fe plays an important role in catalyst performance during tar reforming. Ni sites have high ability to activate CH and CH bonds in hydrocarbon molecules, while Fe sites can promote the activation of water molecules, providing adsorbed oxygen. The incorporation of Fe to Ni catalysts can also increase the coverage of oxygen species due to the higher oxygen affinity of Fe than that of Ni. These available oxygen atoms during tar reforming can quickly react with carbonaceous species on the adjacent Ni sites, thus suppressing the coke formation. The key requirement for these bi-metallic catalysts is the formation of uniform NiFe alloy structures and the composition of the elements. The optimization of the catalysts composition is significantly depending on the catalyst synthesis method and the uniformity of the alloys helps in improving the selectivity towards desired product such as reduced carbon formation.Co-based catalysts have also been reported to be effective for tar removal. One study from Tomishige's group investigated the synergistic effect of Ni-Co/Al2O3 catalysts on steam reforming of tar from cedar wood pyrolysis [192]. Compared to monometallic catalysts, Co/Al2O3 and Ni/Al2O3, Ni-Co/Al2O3 (Co/Ni\u202f=\u202f0.25 as an optimum alloy composition) exhibited superior activity and possessed greater coke resistance. Additionally, toluene was also used as the model compound in this work. In contrast to the result observed in steam reforming of oxygenated tar, Co/Al2O3 was highly active and stable with smallest amount of carbon deposition for steam reforming of this aromatic hydrocarbon. It was reported that monometallic Co catalyst could effectively suppress toluene decomposition and CO disproportionation, which are the main route for coke formation. In another study by Nabgan et al. [193], bimetallic NiCo catalysts supported on ZrO2 were investigated for catalytic steam reforming of phenol, which is the main component of tar formed following steam gasification of wood biomass. Monometallic Ni and Co catalysts possessed higher acidity sites compared to bimetallic catalysts, leading to lower activity towards phenol steam reforming and higher coke deposition. The existence of Co in NixCoy/ZrO2 (x\u202f=\u202f0, 1, 2, 3, 4 where x\u202f+\u202fy\u202f=\u202f4) not only neutralized the acidity but also caused a decrease in the crystal size and reducibility of the catalysts. Among bimetallic catalysts, Ni3Co1/ZrO2 catalyst showed highest basic site and the presence of tetragonal (t-ZrO2) phase structure was still observed. While the amount of t-ZrO2 phases in Ni2Co2/ZrO2 and Ni1Co3/ZrO2 catalysts was found to decrease. It has been reported that t-ZrO2 phase is more stable and active for chemical reactions than monoclinic (m-ZrO2) and cubic (c-ZrO2) phases [194,195]. Thus, Ni3Co1/ZrO2 catalyst presented superior catalytic activity in terms of phenol conversion and hydrogen yield with high coke resistance.The similarity in atom radius of Ni and Co would be favored for the formation of NiCo alloy nanoparticles, leading to the synergistic effect between these two metals. Specifically, the Co addition to Ni catalysts can enhance the resistance to coke deposition since the Co sites can effectively suppress coke formation reaction such as tar decomposition and CO disproportionation. In most of the studies, the deactivation in NiCo alloy catalysts is preferably oxidation of metallic Co species into Co oxides than coke deposition. Therefore, the amount of Co doping to Ni is critical to obtain stable catalytic performance with reduced coke deposition. As highlighted before for NiFe alloy catalysts, the catalyst synthesis method is important to obtain uniform NiCo alloy supported catalysts and reduction gas enriched reaction conditions helps in minimizing deactivation due to Co oxidation.Steam reforming of biomass tar and/or model compound was also conducted on CoFe bimetallic catalysts. Wang et al. [196] modified the Co/Al2O3 with Fe by co-impregnation method and reported the increase in toluene steam reforming activity due to the synergy between Co and Fe at the appropriate composition. However, the activity of Co-Fe/Al2O3 catalyst considerably decreased as a function of time while that of Co/Al2O3 catalyst remained stable. It is noteworthy that with the presence of H2 in the reactant gas, the stability of Co-Fe/Al2O3 catalyst was improved remarkably. This improvement may be attributed to the strong interaction of H2 on the bcc CoFe alloy surface, which in turn restrained the oxidation of Fe in the bcc CoFe alloy. The preservation of the bcc CoFe alloy thus contributed to high activity in steam reforming of toluene. Koike et al. [197] further unraveled the effect of the H2 addition to steam reforming of toluene. The reaction mechanism over Co\u2013Fe/\u03b1-Al2O3 catalyst was proposed as shown in Fig. 13\n. The hydrogen species was hypothesized to facilitate the activation of toluene via its methyl group.Thereafter, Wang et al. [198] attempted the synthesis of Co-Fe/Mg/Al catalysts prepared from hydrotalcite-like precursors. In this way, nanocomposite structure of the bcc CoFe alloy on MgAl2O4-based solid solution was obtained (Co/Fe/Mg/Al\u202f=\u202f10/10/40/40) with more uniform composition than that on Co\u2013Fe/\u03b1-Al2O3 (Fe/Co\u202f=\u202f0.25) catalyst prepared through conventional co-impregnation method. Compared to Co\u2013Fe/\u03b1-Al2O3 (Fe/Co\u202f=\u202f0.25), Co/Mg/Al and Ni-Fe/Mg/Al (Fe/Ni\u202f=\u202f0.25) catalysts, (Co/Fe/Mg/Al\u202f=\u202f10/10/40/40) catalyst was more active and resistant to coke deposition. Moreover, the regeneration ability of (Co/Fe/Mg/Al\u202f=\u202f10/10/40/40) catalyst was higher than that of Co\u2013Fe/\u03b1-Al2O3 (Fe/Co\u202f=\u202f0.25) catalyst, although its catalytic performance decreased after the repeated use.Finally, in CoFe alloy catalysts, Fe itself have low reforming activity; however, the synergistic effect between Co and Fe, especially the bcc CoFe alloy phase, can enhance tar reforming activity and suppress coke formation. Co sites can activate tar molecules, while neighboring Fe sites can supply oxygen atom to the carbonaceous intermediate.Apart from NiCo and NiFe bimetallic catalysts, alloying Ni with Cu has also been applied to enhance the catalytic performance in terms of activity and resistance to coking [199]. In one study [200], Ni-Cu/Mg/Al bimetallic catalyst was derived from hydrotalcite-like compounds through the calcination and reduction. At the optimum composition of Ni-Cu/Mg/Al (Cu/Ni\u202f=\u202f0.25) gave almost total conversion of tar at 550\u202f\u00b0C. It also exhibited higher activity and lower yield of coke as compared to monometallic Ni/Mg/Al and Cu/Mg/Al catalysts as well as Ni-Fe/Mg/Al (Fe/Ni\u202f=\u202f0.25) catalyst under the same conditions in steam reforming of biomass tar reported in previous work [61]. Furthermore, high stability of Ni-Cu/Mg/Al (Cu/Ni\u202f=\u202f0.25) could be achieved over 2\u202fh of testing without aggregation and change in structure of NiCu alloy particles observed. Following this work, Li et al. [201] evaluated the same Ni-Cu/Mg/Al hydrotalcite-like catalyst in steam reforming of 1-methylnaphthalene so as to elucidate the effect of Cu/Ni. The volcano-type relationship between the catalyst compositions and the correlating catalytic performances was also observed in this system, in which Cu/Ni\u202f=\u202f0.25 composition provided the highest reforming activity and satisfying selectivity towards CO\u202f+\u202fCO2. Based on the kinetic study, it was also found that 1-methylnaphthalene had strong interaction with both Ni and NiCu catalysts, whereas steam adsorbed more in the presence of NiCu alloy and possibly dissociated to create more adsorbed oxygen species. Ashok and co-workers [199] have explored the catalytic performance of Ni-Cu/SiO2p catalysts derived from phyllosilicate structures for biomass tar steam reforming using cellulose as biomass model compound. At optimum molar composition of Cu/Ni\u202f=\u202f0.15 in 30Ni-5Cu/SiO2p catalyst, NiCu alloys could be formed along with the existence of the phyllosilicate structure. It was shown that 30Ni-5Cu/SiO2p catalyst gave the highest biomass conversion to gaseous products and possessed better stability for longer reaction times at 600\u202f\u00b0C.Dagle et al. [202] presented the study of steam reforming of hydrocarbons from biomass gasifier (i.e. benzene and naphthalene) over MgAl2O4-supported transition metals. Particularly, novel bimetallic IrNi catalyst was examined, focusing on the theoretical modeling study to understand the nature of IrNi alloy structure, its resistance to coking and activity towards reforming reaction. As shown in Fig. 14\n, Ni50 wetted the surface markedly with direct NiO contacts, whereas Ir50 exhibited fewer direct IrO contacts. A well-mixed alloy of the gas-phase Ir5Ni45 cluster was obtained, where three It atoms were lifted and occupied the surface of Ir5Ni45 particle. The relative energetics of the last reaction during methane dissociation were also measured to investigate the propensity of the cluster to form coke. It was found that small Ir clusters in IrNi alloy particles showed coke resistance ability. Additionally, small Ir (~2\u20133 atoms) on the surface of Ni-rich particles provided electron-rich Ir sites, enhancing activity and durability for steam reforming as compared to only small Ir clusters or Ni particles.Noble transition metal Pt has also been added into Ni-based catalysts to build up the bimetallic systems for steam reforming reactions [203]. Mukai et al. [204] reported the effect of Pt addition to Ni/La0.7Sr0.3AlO3\u2212\u03b4 perovskite (Pt/Ni/LSAO) catalyst on toluene steam reforming. Even with no pre-reduction, high catalytic activity and low coke formation were achieved using Pt/Ni/LSAO catalyst. These results might be related to the enhancement of Ni reducibility due to the additive Pt as well as the formation of adjacent or alloy structure between Pt and Ni on Pt/Ni/LSAO catalyst. It was also discovered that different sequence of Pt and Ni impregnation resulted in different structure of supported metals, and hence affected the catalytic activity and the amount of coke formation. Since the interface between Ni and LSAO perovskite was crucial for exchanging lattice oxygen, the impregnation of Pt to Ni/LSAO catalyst was preferable to the inverse impregnation. Trace amount of noble metal Pd was also discovered to promote activity and stability of hydrotalcite-derived Ni/Mg/Al catalyst in oxidative steam reforming of biomass tar from pyrolysis [205]. With the same noble metal/Ni molar ratio, Pd showed higher catalytic performance than Pt, Au, Ru, Rh and Ir, which could be due to the enhancement of Ni reducibility and the formation of highly dispersed Pd atoms on the surface of Ni particles (Fig. 15\n). The presence of trace Pd could also prevent the oxidation of active Ni metal, resulting in more stable catalytic performance during two-hour stability test as compared to monometallic Ni catalyst.Rare transition metal rhenium (Re) was also utilized as the promoter in Ni/Char catalysts for cracking/reforming of naphthalene and biomass tar [105]. The addition of Re could improve dispersion and prevent agglomeration of Ni particles supported on char, resulting in the catalytic activity improvement. Hydrogen yield was also increased with the presence of Re probably because it facilitated water dissociation, which subsequently enhanced water gas shift activity.Higher oxygen affinity can also be obtained through the formation of small NiCu alloy nanoparticles. The higher oxygen affinity can promote the activation of steam and subsequently facilitate the reaction of oxygen atoms with activated tar molecules. Alloying Ni with other second metals can also provide synergistic effect, which increases reducibility, yields small particle size of alloys and/or enhances metal dispersion. The ample surface metal particles thus offer the large number of active sites for tar dissociation and steam activation.This review summarizes the recent efforts that have been made for the development of catalysts for steam reforming of biomass tar reactions using tar model compounds. Research efforts have been more focused on the design and influence of material properties on the catalysis with respect to reactant conversion, stability and carbon suppression. The catalytic performance of a material is also highlighted for change in the mechanism of steam reforming reactions. On a common note, the role of an efficient catalyst is pivotal for economically feasible of catalytic biomass gasification technology. The nature and properties of the catalysts such as redox or acid-base properties play an important role in determining the catalytic activity as well as selectivity towards the desired product. In general, an excellent catalyst should have high tar conversion efficiency, low cost, easy regenerability and be environment friendly. For an efficient catalyst both nature of active metal and support material plays a crucial role. With respect to availability and activity, Ni as an active metal component is widely accepted for this reforming reaction. Furthermore, the use of synergistic bimetallic catalysts has been one of the promising approaches to increase the catalytic performance and enhance coke resistance during steam reforming of biomass tar model reaction. This strategy can provide (i) the significant increase in the coverage of oxygen species by alloying either Ni or Co with Fe (ii) higher oxygen affinity through the formation of small NiCu alloy particles (iii) the enhancement of reducibility along with higher surface metal concentration and/or greater dispersion due to the synergistic effect between Ni and second metals, such as Co, Ir, Pt, Pd and Re.Based on the current state of art in the development of catalyst for biomass gasification process, the key criteria for the future modification are as follows: (i) increasing the basic strength of catalyst to achieve higher syngas yield via water gas shift reaction, (ii) incorporation of oxygen vacancies by using redox metal oxides such as CeO2, FeOx or perovskite-based support as this will help to suppress the coke formation, (iii) development of core-shell structure with porous shell to prevent the metal sintering, coke formation and improve the mass transfer of reactants and products; specifically formation of oxide shell with high oxygen mobility such as CeO2, mixture of CeZr and perovskite shell will enhance the coke resistance ability; additionally using hydrotalcite as a shell with porosity will help to enhance the basic properties of the catalyst, and (iv) enhancing the surface area of natural minerals and waste-derived catalysts by modification with structure-directing agents or optimizing the hydrothermal conditions such as the choice of base, time and temperature for the hydrothermal treatment.For the choice of nature of support, so far, most of the works considered alumina-based support is an effective support which promotes the breaking of CC bond in the hydrocarbons and enhances conversion. However, the challenge is to improve the hydrothermal stability of alumina and also to prevent coke deposition. Modification of alumina support with alkaline earth metal oxides such as MgO, CaO and inert metal oxides SiO2 helps in improving the coke deposition and improved the stability of the catalyst. For SiO2 based catalyst, phyllosilicates type of support has emerged as a promising catalyst. SiO2-based catalyst has been designed in the form of the core-shell catalyst which improves the confinement effect and prevents metal sintering problem under the high-temperature reaction condition. Synthesis of catalysts via hydrotalcites precursors also have great potential to fine tune and achieve higher activity and longer stability. In future, the core-shell structure of catalyst can be explored with CeO2, ZrO2, and mixed oxides. Another possible route to improve the coke resistance of the catalyst is by utilizing the oxygen vacancy of the support such as provided by perovskite-based support etc.So far, the commercially available catalysts for the gasification process are natural minerals (olivine, dolomite), alkali metal catalysts, and transition metal-based catalysts. With the use of natural mineral as a catalyst, the gaseous product formed needs to be improved. Moreover, this leads to the also additional gas-cleaning step as the quality of the end product is inadequate. In contrast, a wide range of studies was conducted and reported in the literature using commercial nickel-based catalysts in biomass gasification to promote steam-reforming, water\u2013gas shift reactions and to eliminate tar. When nickel is used as a catalyst, the quality of the gaseous product can be enhanced. Ni-based catalyst is more economically attractive as both gasification and gas clean-up occur simultaneously. There are several Ni-based catalyst systems commercialized for the biomass gasification, for example, BASF developed Ni supported on CaO\u2013Al2O3\u2013SiO2\u2013K2O and MgO\u2013CaO\u2013Al2O3\u2013SiO2\u2013K2O referred as G1-25/1 and G1-50, respectively [206,207]. These catalysts showed the toluene conversion of 89\u201399% can be achieved in the temperature range of 660\u2013850\u202f\u00b0C. Baker and Li et al. studied the commercial nickel catalyst G-90C as a primary catalyst for biomass gasification [208,209]. Both of these studies elucidated the advantages and disadvantages of using the commercial Ni-based catalyst. The catalyst active the quality of gaseous product and also decreased the overall tar yield. However, the catalyst suffers from deactivation after several cycles. Also, the leading cause was the coke deposition on the catalyst. In a review paper by Chan et al. described a summary and detailed description of the commercially available primary and secondary catalyst for biomass gasification process [210]. Therefore, a tremendous effort has been made in the past few years to improve the transition metal based catalyst to achieve with high tar conversion and improved resistant towards coking.A cheaper and effective catalyst will make the biomass reforming process more economical; therefore, eco-friendly materials, char based supports were also investigated. A tremendous effort has been implemented to improve the thermal and chemical resistance of the char based support for toluene steam reforming. However, the performance still needs to be improved to be comparable with other types of support. Therefore, modification of the synthesis method to improve the acidic and basic centers, improving the surface area of the char will be helpful to achieve excellent performance for biomass gasification processes. Several natural minerals and waste materials such as incinerator bottom ash are some of the emerging types of support for this steam reforming reaction. Although the catalyst showed appreciable thermal stability, the catalytic activity was inferior to that of other catalytic system. Therefore, further improvement in the catalyst surface area, fine-tuning the acidity and basicity and improving the metal dispersion via modification of the synthesis method is crucial to achieving higher catalytic activity.In the context of actual biomass tar conversion, there is a pressing demand to develop high performance, low cost and robust catalyst, with the use of harsher temperatures and feedstock that contains potential catalyst poisons such as sulfur, nitrogen and chlorine, the issues of catalytic deactivation through thermal sintering and poisoning (e.g. sulfur, nitrogen, chlorine and etc.) needs to be solved through development of thermally-stable and sulfur-tolerant catalysts for actual tar reforming applications [211]. Beside the development of efficient catalyst for tar reforming, there are some alternative approaches which are reported for efficient tar conversion. For instance, Wang et al. [212] showed integration of perovskite based membrane with catalyst and performed steam reforming of toluene reaction. The perovskite based membrane selectively separates oxygen from air and promotes partial oxidation of tar reaction. By this way the catalyst showed enhanced catalytic performance with respect to toluene conversion and carbon suppression ability. However, the stability of perovskite based membrane in the presence of poisoning compound needs to be investigated further [213]. Another approach for efficient tar conversion is performing steam reforming tar using electro chemical reforming technique [20,214,215]. Herein the tar reforming is performed in the presence of mild electric current [216]. It can be performed between 100\u202f\u00b0C and 800\u202f\u00b0C temperature conditions. The reports showed the catalytic performance was remarkably increased in the presence of electric current. These two processes required much more research in order to confirm their technical feasibility in biomass gasification processes. From this comprehensive review on the progress of the performance of the catalyst for steam reforming of biomass tar model, it can be established that different properties of support such as acidic sites, oxygen vacancy, and basicity, and alloying active metal improve the interaction of metal with support which eventually helps in making a robust catalyst.The authors gratefully thank the financial support from National University of Singapore, National Research Foundation-Prime Minister's office, Republic of Singapore, the National Environment Agency - Singapore under the Waste-to-Energy Competitive Research Program (WTE CRP 1501 103, WBS No. R-279-000-491-279), Agency for Science, Technology and Research (AME-IRG A1783c0016, WBS No. R-279-000-509-305) and Ministry of Education - Singapore (MOE2017-T2-2-130, WBS No. R-279-000-544-112).The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.", "descript": "\n                  This review describes recent advances in development of catalysts for steam reforming of biomass tar model reactions, using toluene, benzene and naphthalene as tar model compounds. Catalytic systems have been categorized based on their catalytic properties. The material properties such as oxygen mobility and basicity of catalysts showed great influence in their effectiveness in tar reforming. Changes in the properties such as oxygen mobility and basicity with various metal and/or support modifications and their influence on catalytic behavior with respect to reactant conversion and coke inhibition is comprehensively discussed. The activity of the catalysts derived from various synthesis methods is also introduced. The changes induced in the pathways of steam reforming reactions by catalyst modification is also highlighted together with changes in catalytic properties. Reaction pathways for steam reforming of toluene by experimental studies together with insights gained from computational DFT studies is also presented.\n               "}
{"full_text": "Waste incineration is developing rapidly in China, since this process can reduce waste volume as well as generate heat energy (Zhu et al., 2018). However, it cannot be avoidable to have pollutants, such as NO\nx\n and chlorine-containing volatile organic compounds (CVOCs) from flue gas of municipal solid waste incinerators, and the emissions are increasing in recent years in China (Kulkarni et al., 2008; Wu et al., 2016; Li et al., 2017; Shao et al., 2021). Meanwhile, NO\nx\n and CVOCs are the critical precursors of PM2.5 and O3 (Ni et al., 2009; Wu et al., 2016; Ye et al., 2022), which can cause severe environmental problems (Alzaky and Li, 2021). Therefore, how to effectively control emissions of NO\nx\n and CVOCs has been a focus of air pollution control.Selective catalytic reduction of NO\nx\n by hydrocarbons (HC\u2010SCR) has attracted much attention over the past two decades (More et al., 2018; Xu et al., 2020). The transition metal oxides display the good reactivity with inexpensive costs, which are becoming potential catalysts in HC-SCR, especially Mn-based catalysts. For example, Liu et.al. (Liu et al., 2019a) obtained up to 100\u00a0% NO conversion at 130\u2103 with MnOx/AC in C2H4-SCR.In the last few decades, it has been emerging various strategies for VOCs elimination, such as adsorption (Zhao et al., 2022), plasma (Yu et al., 2020), biodegradation (Khan et al., 2018), catalysis (Chen et al., 2022), photocatalysis (Chen et al., 2022), Photothermal Catalysis (Yang et al., 2022b). Catalytic oxidation is recognized as one of the most efficient technologies for low concentration CVOCs removal due to high removal efficiency and mineralization rate (Fang et al., 2019; He et al., 2019; Liu et al., 2019c). Interestingly, transition metal oxides-based (MnOx) catalysts with variable valance states also exhibit the potential of high reactivity in CVOCs decomposition (Kim and Shim, 2010; Santos et al., 2010; Piumetti et al., 2015; Ye et al., 2021). Sphere-Shaped Mn3O4 exhibited a complete conversion of methyl\u00a0\u2212\u00a0ethyl\u00a0\u2212\u00a0ketone to CO2 at 200\u00a0\u00b0C (Pan et al., 2017). Cheng et. al. (Cheng et al., 2017) also obtained 90\u00a0% dimethyl ether conversion at 238\u00a0\u00b0C on \u03b1-MnO2.Hence, MnOx catalysts shows the bifunctional ability to reduce NO\nx\n and oxidize CVOCs. However, to the best of our knowledge, the promising catalyst of \u03b1-MnO2 in these two different reaction mechanisms was seldom reported when applying for NO\nx\n reduction in HC\u2010SCR and CVOCs oxidation separately. It is worthy investigating the relationship between structure properties and reactivity of \u03b1-MnO2, and to discuss the physicochemical characterizations. In this present work, a series of \u03b1-MnO2 catalysts with the different physical\u2013chemical characteristics were prepared and applied for the catalytic removal of both NO\nx\n and DCE (as the typical pollutant of CVOCs). Catalyst characterization and reactivity were investigated to illustrate the relation between properties and reactivity of \u03b1-MnO2 catalysts.All \u03b1-MnO2 samples were synthesized by hydrothermal method using Potassium permanganate (KMnO4) and Manganous acetate ((CH3COO)2Mn) as precursors. For \u03b1-MnO2-1, 1.5\u00a0g (CH3COO)2Mn and 2.5\u00a0g KMnO4 were added into 160\u00a0mL distilled water with magnetic stirring for 30\u00a0min at room temperature. Next, the mixed solution was poured into two Teflon-lined stainless-steel autoclave (100\u00a0mL), and then heated at 160\u00a0\u00b0C for 12\u00a0h. The obtained colloids were washed with deionized water and ethanol, then dried in vacuum at 80\u00a0\u00b0C for 12\u00a0h. The dried samples were calcined at 450\u00a0\u00b0C for 4\u00a0h. Finally, the catalyst was pelleted, crushed, and sieved to 40\u00a0\u223c\u00a060 mesh granules before use.To modify the physical\u2013chemical characteristics of \u03b1-MnO2 catalysts, the mass ratio of (CH3COO)2Mn and KMnO4 was 4.9:3.2, and the temperature of hydrothermal treatment was 140\u00a0\u00b0C. The obtained product was donated as \u03b1-MnO2-2. Also, according to the synthesis plan of \u03b1-MnO2-1, the mass ratio of (CH3COO)2Mn and KMnO4 was 9.7:6.3, and the temperature of hydrothermal treatment was 90\u00a0\u00b0C. The obtained product was donated as \u03b1-MnO2-3. To obtain \u03b1-MnO2-4, the mass ratio of (CH3COO)2Mn and KMnO4 was 0.8:3.0, and the hydrothermal treatment was carried out at 240\u00a0\u00b0C.The catalytic activity test of HC-SCR was carried out in a stainless-steel tubular (i.d.10\u00a0mm). 600\u00a0mg catalyst was evaluated under a typical feed gas included 800\u00a0ppm NO, 600\u00a0ppm C3H8, 6.5\u00a0vol% O2, and N2 as balance. The reaction temperature was conducted from 150 to 550\u00a0\u00b0C at a total flow rate of 450\u00a0mL/min, corresponding to a weigh hourly space velocity of 19 000\u00a0mL\u00b7g\u22121\u00b7h\u22121. The NO\nx\n concentration of the inlet and outlet were continuously measured by an infrared gas analyzer (Xi\u2019an Juneng Corporation, China). NO\nx\n conversion was evaluated by Eq. (1):\n\n(1)\n\n\nN\n\nO\nx\n\n\nconversion\n\n\n\n\\%\n\n\n\n=\n\n\n\n\n\nN\n\nO\nx\n\n\n\n\nin\n\n\n-\n\n\n\nN\n\nO\nx\n\n\n\n\nout\n\n\n\n\n\n\nN\n\nO\nx\n\n\n\n\nin\n\n\n\n\u00d7\n100\\%\n\n\n\n\nwhere [NO\nx\n]in and [NO\nx\n]out are the inlet and outlet concentration of NO\nx\n, respectively.The catalytic activity of DCE combustion was measured using a stainless-steel tubular (i.d.10\u00a0nm). 500\u00a0mg catalyst was selected for activity test. The reaction was conducted from 100 to 450\u00a0\u00b0C at a flow rate of 400\u00a0mL/min of feeding gases with 500\u00a0ppm DCE and 21\u00a0vol% O2, corresponding to a gas hourly space velocity (GHSV) of 48 000\u00a0mL\u00b7g\u22121\u00b7h\u22121. DCE and products (CO and CO2) were measured by an on-line gas chromatograph (GC9890) with ECD and FID. An on-line Cl2 and HCl detectors (PN\u20132000, China) was used to analyse the concentrations of Cl2 and HCl. DCE conversion, CO\nx\n (CO2 and CO) yield, HCl yield and Cl2 yield were evaluated by the following Eqs. (2) to (6), respectively.\n\n(2)\n\n\nN\n\nO\nx\n\n\nconversion (\\%)=\n\n\n\n\n\nN\n\nO\nx\n\n\n\n\nin\n\n\n-\n\n\n\nN\n\nO\nx\n\n\n\n\nout\n\n\n\n\n\n\nN\n\nO\nx\n\n\n\n\nin\n\n\n\n\u00d7\n100\\%\n\n\n\n\n\n\n(3)\n\n\nC\n\nO\n2\n\n\nyield\n\n\n\n\\%\n\n\n=\n\n\n\n\nC\n\nO\n2\n\n\n\nout\n\n\n2[DC\n\nE\nin\n\n]\n\n\n\u00d7\n100\\%\n\n\n\n\n\n\n(4)\n\n\nCO yield\n\n\n\n\\%\n\n\n=\n\n\nC\n\nO\nout\n\n\n\n2[DC\n\nE\nin\n\n]\n\n\n\u00d7\n100\\%\n\n\n\n\n\n\n(5)\n\n\nHCl yield\n\n\n\n\\%\n\n\n=\n\n\nHC\n\nl\nout\n\n\n\n2[DC\n\nE\nin\n\n]\n\n\n\u00d7\n100\\%\n\n\n\n\n\n\n(6)\n\n\nC\n\nl\n2\n\n\nyield\n\n\n\n\\%\n\n\n=\n\n\n\n\nC\n\nl\n2\n\n\n\n\nout\n\n\n\n[DC\n\nE\nin\n\n]\n\n\n\u00d7\n100\\%\n\n\n\n\nwhere [DCE]in is the DCE inlet concentration of DCE. [DCE]\nout\n, [CO]\nout\n, [CO2]\nout\n, [HCl]\nout\n and [Cl2]\nout\n are the outlet concentration of DCE, CO, CO2, HCl and Cl2, respectively.Fourier infrared spectrum (FT-IR) spectra was recorded in the range 400\u00a0\u223c\u00a02000\u00a0cm\u22121 with a resolution of 4\u00a0cm\u22121 on a thermo scientific nicolet iS20 using the KBr pellet technique.X-ray powder diffraction patterns (XRD) were obtained by Panalytical X'Pert'3 Powder diffractometer, equipped with Cu K\u03b1 X-ray radiation (\u03bb\u00a0=\u00a00.15406\u00a0nm). All the catalysts were scanned at 2\u03b8 range between 10\u00b0 to 80\u00b0 (rate of 2\u00b0/min).The surface areas of the synthesized materials were determined by the Brunauer\u00a0\u2212\u00a0Emmett\u00a0\u2212\u00a0Teller (BET) method using ASAP 2460 3.01 instrument. Nitrogen physisorption experiments were carried out at 77\u00a0K after initial pretreatment of the samples by degassing at 300\u00a0\u00b0C for 2\u00a0h.Scanning electron microscopy (SEM) images have been studied using Schottky (ZEISS Gemini 300) equipment with a resolution of 10\u00a0kV and 50\u00a0kV.Transmission electron microscopy (TEM) images were obtained from JEM-F200 electron field emission transmission electron microscope (JEOL, Japan) under 200\u00a0kV acceleration voltage.X-ray photoelectron spectra (XPS) were carried out on thermo scientific system with Al K\u03b1 radiation. The binding energy scale was corrected for surface charging by use of the C 1\u00a0s peak of contaminant carbon as reference at 284.8\u00a0eV.Hydrogen temperature programmed reduction (H2-TPR) experiment was conducted on an AutoChem1 II 2920 instrument equipped with a thermal conductivity detector (TCD) to measure the consumption of H2. Before detection by the TCD, a 50\u00a0mg sample was pretreated under N2 stream (40\u00a0mL\u00b7min\u22121) at 300\u2103 for 1\u00a0h, and then cooled to 50\u00a0\u00b0C. A mixed stream with a 10\u00a0vol% H2/Ar mixture (50\u00a0mL\u00b7min\u22121) was introduced into the sample, and the sample was heated from room temperature to 800\u00a0\u00b0C.The NH3 temperature-programmed desorption (NH3-TPD) was performed on an AutoChem1 II 2920 instrument. The catalyst was (100\u00a0mg) was pretreated in a N2 (50\u00a0mL\u2219min\u22121) of at 300\u2103 for 1\u00a0h, and then cooled to 50\u2103. Sample was treated with 10\u00a0% NH3 diluted in N2 (30\u201350\u00a0mL\u2219min\u22121) for 1\u00a0h to achieve adsorption saturation. The gas was switched back to He (30\u00a0mL\u2219min\u22121) for 1\u00a0h to purge the physically adsorbed species. Finally, the catalyst was heated from 50\u2103 to 800\u2103 at a rate of 10\u00a0\u00b0C\u2219min\u22121 in high purified N2 (30\u00a0mL\u2219min\u22121).\nFig. 1\n showed the different catalytic performance of \u03b1-MnO2 samples for NO\nx\n reduction and DCE oxidation at the range of 100\u00a0\u223c\u00a0450\u00a0\u00b0C. For NO\nx\n reduction, activity performance decreased as follows: \u03b1-MnO2-3 (63.5\u00a0%)\u00a0>\u00a0\u03b1-MnO2-4 (53.8\u00a0%)\u00a0>\u00a0\u03b1-MnO2-1 (45.2\u00a0%) \u2248 \u03b1-MnO2-2 (42.1\u00a0%) at 250\u00a0\u00b0C. For DCE oxidation, T50 was chosen to compare the activity of these samples. T50 of DCE oxidation follows the order of \u03b1-MnO2-3 (276.4\u00a0\u00b0C)\u00a0<\u00a0\u03b1-MnO2-4 (310.9\u00a0\u00b0C)\u00a0<\u00a0\u03b1-MnO2-2 (337.3\u00a0\u00b0C) \u2248 \u03b1-MnO2-1 (348.6\u00a0\u00b0C). Thus, \u03b1-MnO2-3 exhibited the highest catalytic activity for both NO\nx\n reduction and DCE oxidation.As exhibited in Fig. 2\n, in yield of CO2 and CO, for \u03b1-MnO2-3, it is remarkable to reveal 84.4\u00a0% of C3H8 was decomposed into CO\nx\n. While for \u03b1-MnO2-1, \u03b1-MnO2-2 and \u03b1-MnO2-4, C element in C3H8 was about 36.8\u00a0%, 41.6\u00a0% and 51.6\u00a0% converted to CO2, and about 12.8\u00a0%, 14.6\u00a0% and 15.9\u00a0% converted to CO, respectively. In yield of HCl and Cl2, for \u03b1-MnO2-3, 53.8\u00a0% of HCl and 22.7\u00a0% of Cl2 were observed. That is, 76.5\u00a0% of DCE was total oxidized into inorganic chlorine products using \u03b1-MnO2-3. However, for \u03b1-MnO2-1, \u03b1-MnO2-2 and \u03b1-MnO2-4, Cl element in DCE were about 10.0\u00a0%, 8.6\u00a0% and 17.7\u00a0% converted to Cl2, and about 29.9\u00a0%, 34.2\u00a0% and 34.9\u00a0% converted to HCl, respectively. There are about 60.0\u00a0%, 57.1\u00a0% and 47.4\u00a0% of chlorine remained as organic chlorine. Above all, \u03b1-MnO2-3 displayed the best products selectivity.FT-IR spectra of the synthesized materials are shown in Fig. 3\na. The peaks in low wavenumbers between 800\u00a0cm\u22121 and 400\u00a0cm\u22121 are assigned to Mn-O lattice vibration (Yuan et al., 2009; Wang et al., 2019). Our samples showed the well-defined absorption peaks of MnO2 at 469\u00a0cm\u22121, 526\u00a0cm\u22121, and 720\u00a0cm\u22121 as well as the weak defined shoulder at 597\u00a0cm\u22121 (King'ondu et al., 2011; Chen et al., 2015; Liu et al., 2019b). Fig. 3b shows the XRD patterns of the as-prepared manganese oxide samples. Comparing to the XRD patterns of the standard \u03b1-MnO2 (JCPDS 44\u20130141) (Cheng et al., 2017; Gao et al., 2017), it can deduce that all of the four samples could be well corresponding to the tetragonal \u03b1-MnO2 phase. The diffraction peaks at 2\u03b8\u00a0=\u00a012.9\u00b0, 18.2\u00b0, 25.8\u00b0, 28.9\u00b0, 37.6\u00b7\u00b0, 42.0\u00b0, 49.9\u00b0, 56.4\u00b0, 60.3\u00b0, 65.1\u00b0, and 69.7\u00a0\u00b0could be attributed to the (110), (200), (220), (310), (211), (301), (411), (600), (521), (002) and (541) plane, respectively. \u03b1-MnO2 presented three main peaks at 12.9\u00b0, 28.9\u00b0 and 37.6\u00b0, which were assigned to the (110), (310) and (211) planes of \u03b1-MnO2 (JCPDS 44\u20130141). The peak intensity of (110) plane order decreased as follows: \u03b1-MnO2-3\u00a0>\u00a0\u03b1-MnO2-2\u00a0>\u00a0\u03b1-MnO2-4\u00a0>\u00a0\u03b1-MnO2-1. The intensity order of (310) plane is as follows: \u03b1-MnO2-2\u00a0>\u00a0\u03b1-MnO2-3\u00a0>\u00a0\u03b1-MnO2-1\u00a0>\u00a0\u03b1-MnO2-4. Interestingly, the peak intensity of (211) plane follows the order as \u03b1-MnO2-3\u00a0>\u00a0\u03b1-MnO2-4\u00a0>\u00a0\u03b1-MnO2-2\u00a0>\u00a0\u03b1-MnO2-1, which is similar with the activity order of four samples. It indicates that (110) plane, (310) plane and (211) plane may be exposed active planes, in agreement with the findings of TEM. Besides, no diffraction peaks of other phases, such as \u03b2-MnO2 and \u03b3-MnO2, are detected, implying that each catalyst is composed of \u03b1-MnO2 phase. This result is consistent with the result of FT-IR.The average pore size, pore volume and specific surface area of catalysts were measured by N2 adsorption\u2013desorption, and the results are presented in Fig. 4\n and Table 1\n. N2 adsorption\u2013desorption isotherms of the samples type II characteristics with well-developed H3 type hysteresis loops, confirming that the samples have mesoporous characteristics (Fig. 4a) (Sing, 1982). The samples possessed a mesopore distribution in the range of 2\u00a0\u223c\u00a035\u00a0nm in Fig. 4b. \u03b1-MnO2-1 and \u03b1-MnO2-4 presented a wide peak centered at from 5\u00a0nm to 35\u00a0nm, while \u03b1-MnO2-2 and \u03b1-MnO2-3 presented the peak centered at ca. 2.5\u00a0nm. In Table 1, the surface area order decreased as follows: \u03b1-MnO2-4 (104.6\u00a0m2\u00b7g\u22121)\u00a0>\u00a0\u03b1-MnO2-1 (72.6\u00a0m2\u00b7g\u22121)\u00a0>\u00a0\u03b1-MnO2-3 (44.1\u00a0m2\u00b7g\u22121)\u00a0>\u00a0\u03b1-MnO2-2 (34.6\u00a0m2\u00b7g\u22121). The pore volume of order was as follows: \u03b1-MnO2-4 (0.6\u00a0cm3\u00b7g\u22121)\u00a0>\u00a0\u03b1-MnO2-1 (0.2\u00a0cm3\u00b7g\u22121)\u00a0>\u00a0\u03b1-MnO2-3 (0.1\u00a0cm3\u00b7g\u22121)\u00a0=\u00a0\u03b1-MnO2-2 (0.1\u00a0cm3\u00b7g\u22121). The difference in preparation condition leads to a big difference in surface area of the \u03b1-MnO2 samples.To further analyze the morphologies and surface structures of the catalysts, the SEM and TEM images of four samples are shown in Fig. 5\n. It should be noted that the wire-like morphology can be differentiated from the rod-like morphology in terms of the bending or straight shape (Wang et al., 2012). Fig. 5 (a) and (d) showed that both \u03b1-MnO2-1 and \u03b1-MnO2-4 are presented as stacking-nanowires, while Fig. 5 (b) and (c) exhibited nanorod-like appearance of \u03b1-MnO2-2 and \u03b1-MnO2-3 with uniform distribution. It should be noted that the surface area of wire-like morphology can be much larger than the rod-like morphology. The well-identified periodic lattice fringes of 2.40\u00a0\u00c5, 3.10\u00a0\u00c5 and 6.94\u00a0\u00c5 are corresponding to the interplanar distance of (211), (310) and (110) facets of \u03b1-MnO2, respectively. Whereas, severe blurring of the lattice fringes were also detected (highlighted by red rectangles) in \u03b1-MnO2-3. It is worth noting that large amount of point defects on \u03b1-MnO2 could obscure the distorted lattice fringes, which may result from the existence of oxygen vacancies on catalyst surfaces(Huang et al., 2018).The morphologies of samples are consistent with the findings of XRD.XPS measurements were carried out to identify the surface species of \u03b1-MnO2 samples. Fig. 6\na illustrated Mn 2p spectra of four samples. Peaks at 642.7, 641.7 and 640.4\u00a0eV can be attributed to Mn4+, Mn3+ and Mn2+, respectively (Si et al., 2015; Ma et al., 2017; Zhang et al., 2022). In Table 1, the proportion of low valence Mn (Mn3+ and Mn2+) followed the order (Table 1): \u03b1-MnO2-3 (0.7)\u00a0>\u00a0\u03b1-MnO2-4 (0.6)\u00a0>\u00a0\u03b1-MnO2-2 (0.5)\u00a0=\u00a0\u03b1-MnO2-1 (0.5). Low valence Mn content is an indicator of surface oxygen vacancies (Yang et al., 2020). Additionally, Mn2+-O and Mn3+-O bonds are weaker than Mn4+-O (Zhang, 1982). Large proportion of low valence Mn results in longer and weaker Mn-O bonds on the surface of \u03b1-MnO2-3 (Yang et al., 2020). It suggests that oxygen atoms on its surface are more likely to be released to participate in oxidation. In addition, the existence of surface low valence Mn would promote dissociation and activation of circumambient oxygen atoms (Yang et al., 2020).The XPS spectra of O 1\u00a0s of the samples are shown in Fig. 6b. As reported previously, the peak around 529.0\u00a0\u223c\u00a0530.0\u00a0eV is typical for Olatt in a coordinatively saturated environment, while the peak around 531.0\u00a0\u223c\u00a0532.0\u00a0eV can be attributed to the Oads in a low-coordinated environment (Tang et al., 2010; Wang et al., 2011; Yang et al., 2022a). As shown in Table 1, the Oads/Olatt molar ratio for all \u03b1-MnO2 catalysts follows the order of \u03b1-MnO2-3 (0.6)\u00a0>\u00a0\u03b1-MnO2-4 (0.5)\u00a0>\u00a0\u03b1-MnO2-2 (0.4)\u00a0=\u00a0\u03b1-MnO2-1 (0.4), which is consistent with the catalytic activity results. As we known that Oads performs high activities and makes an important impact in SCR reaction because of its higher mobility than Olatt (Zhang et al., 2020). On the basis of the Mars-van Krevelen mechanism, the emergence and annihilation of oxygen vacancies is the key step of VOC oxidation. Adsorbed oxygen species participate in the redox cycle of the vacancies from gaseous-adsorbed oxygen transformation (Huang et al., 2015). Surface adsorbed oxygen are relevant to the formation of Mn3+ and Mn2+ and are more active than lattice oxygen at low temperatures (Wang et al., 2012). Thus, \u03b1-MnO2-3 might be highly active in DCE oxidation owing to large numbers of low valence Mn cations and adsorbed oxygen species.H2-TPR measurement is carried out to analyze the reducibility of different \u03b1-MnO2 samples and the results are presented in Fig. 7\n. As shown in Fig. 7a, Two reduction peaks (\u2160, \u2161) could be due to the reduction of Mn4+ to Mn3+ and Mn3+ to Mn2+, respectively (Yang et al., 2020). The initial H2 consumption rate was calculated to better evaluate the reducibility of these samples, as depicted in Fig. 7b. It was clearly seen that the initial H2 consumption rates of the samples decreased in the order of \u03b1-MnO2-3\u00a0>\u00a0\u03b1-MnO2-4\u00a0>\u00a0\u03b1-MnO2-1\u00a0>\u00a0\u03b1-MnO2-2. The lower reduction temperature and the larger initial H2 consumption rate indicate a better low-temperature redox ability (Chen et al., 2017; Gong et al., 2017).NH3-TPD experiments are taken to analyze the acidities of the four types of \u03b1-MnO2 and the results are shown in Fig. 8\n and Table 1. As shown in Fig. 8, NH3-TPD curves of \u03b1-MnO2 samples exhibited two desorption peaks (labeled as I and II). The desorption peak I at low temperature is attributed to the desorption of NH3 from weak acid sites and the desorption of physisorbed NH3, the desorption peak II at middle temperature is assigned to the desorption of NH3 from middle strong acid sites (Fang et al., 2013; Yao et al., 2017). It is worth noting that two desorption peaks of \u03b1-MnO2-3 and \u03b1-MnO2-4 belong to weak acid site and middle strong acid site at below 450\u2103. The quantitative analysis data of NH3-TPD was summarized in Table 1. It was reported that the quantity of the desorption peak was proportional to the strength of acid site (Zhang et al., 2020). Hence, as seen in Fig. 8, the quantity of peaks can be ranked by \u03b1-MnO2-3\u00a0>\u00a0\u03b1-MnO2-4\u00a0>\u00a0\u03b1-MnO2-1\u00a0>\u00a0\u03b1-MnO2-2, implying the order of acid site numbers. \u03b1-MnO2-3 catalyst not only shows two acid sites at 145\u2103 and 350\u2103, but also presents the largest amount of acid sites among these catalysts, which is basically consistent with the activity test of NO\nx\n reduction.The possible properties-reactivity relationship of four \u03b1-MnO2 samples was illustated in Fig. 9\n. In this work, Oads/Olatt, low valence Mn content and total acidity were positively related to activity of the samples. However, surface area, pore structure and redox properties are not the key factors in our study.\u03b1-MnO2-3 catalyst presented the best performance among a series of \u03b1-MnO2 for both the catalytic reduction of NO\nx\n reduction and DCE oxidation. The Oads/Olatt, the proportion of low valence Mn content and total acidity are the crucial factors for the activity of \u03b1-MnO2. The maximum conversion of NO\nx\n achieved 63.5\u00a0% at 250\u2103 and DCE achieved 80\u00a0% at 338\u2103 on \u03b1-MnO2-3 catalyst, respectively. As for the yield of carbon and chlorine, \u03b1-MnO2-3 also exhibits highest yield, which implies that \u03b1-MnO2-3 may be a potential catalyst for removal of NO\nx\n and VOCs.The authors declare that they have no know competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was financially supported by Zhejiang Provincial Natural Science Foundation of China (LGF22B070005) and Innovation and Entrepreneurship Talent Project of Jiangsu Province (JSSCRC2021236).", "descript": "\n                  The reduction ability of NO to N2 and the oxidation performance of 1,2-dichloroethane (DCE) over \u03b1-MnO2 catalysts were investigated. The results show that \u03b1-MnO2-3 exhibited the highest catalytic activity in 63.5\u00a0% conversion of NO\n                        x\n                      reduction by C3H8 at 250\u00a0\u00b0C, and 80\u00a0% conversion of DCE combustion by O2 at 338\u00a0\u00b0C. It is revealed the active phase of \u03b1-MnO2-3 is tetragonal \u03b1-MnO2 with the selectively exposed plane of (211). It was proposed the high DCE decomposition of \u03b1-MnO2-3 was ascribed to the redox properties. The overall characterization results revealed that \u03b1-MnO2-3 catalyst preserves more active sites of low valence Mn and higher surface adsorbed oxygen (Oads) /lattice oxygen (Olatt) at the outermost layers, and lower reduction temperature in H2-TPR profiles than that of other catalysts. Meanwhile, NH3-TPD profile of \u03b1-MnO2-3 also shows a large number of acid sites promote NO\n                        x\n                      reduction.\n               "}
{"full_text": "As a class of the rarest elements on the planet, platinum group metals (PGMs) are indispensable materials in many critical technologies attributed to their unique physical and chemical properties. In particular, platinum, palladium and rhodium are utilized as the essential ingredients for vehicle-use catalysts (Nassar, 2015). The traditional application is in emission after-treatment devices for internal combustion engine vehicles (Shelef and McCabe, 2000; Gandhi et al., 2003; Li et al., 2010). The emerging application is in fuel cell vehicles (FCVs), which have attracted global attentions in the context of transport sector decarbonization (Kampker et al., 2020). Platinum is a core ingredient for the catalyst layer, which is indispensable in proton exchange membrane FCVs (Hao et al., 2019). Global consumption of PGMs for producing vehicle-use catalysts reached 400\u00a0t in 2020, accounting for over 60% of total consumption (Market Report, 2021).There are numerous driving factors that are expected to significantly change future PGM supply and demand (Rasmussen et al., 2019). The enforcement of stricter emission regulations will result in an increase in the demand for ICEV-related PGM catalysts. These regulations are gradually implemented around the world, such as the Low Emission Vehicle (LEV) LEV-III automotive emission regulations in North America and the national stage \u2165 emission regulations in China (Xun et al., 2020). Meanwhile, breakthroughs have been made in the commercialization of FCVs. More than twenty thousand FCVs have been produced around the world in 2020 (Hart and Jones, 2020). Approximately 0.175\u00a0g/kW of platinum is used in the engine of the Toyota Mirai FCV (Toyota, 2020), which is higher than that contained in the catalytic converters of conventional vehicles (about 0.05\u00a0g/kW for a Euro VI gasoline engine) (Hao et al., 2019). Under such circumstances, the proportion of PGMs needed for vehicle related catalyst production to the total PGM production will remain at a high level. Recycling is another important factor for PGM supply. According to Graedel et al. (Graedel et al., 2011), the end-of-life recycling rates of PGMs were above 50% in 2011. In some developed countries, they can reach a level above 60% (Xu et al., 2018; Saidani et al., 2019). However, they are relatively lower in many developing countries (Xun et al., 2020). There is still great improvement potential for the global recycling situation, which can replace the primary supply of PGMs to a certain extent.Furthermore, a big geo-political, labor or health related incident could have a significant impact on the PGM catalyst supply chain. For instance, the COVID-19 pandemic has magnified the concerns over the vulnerability of the supply chains of many goods and services to disruptions (Roskill. , 2020). Mining companies in South Africa can only operate at 50% capacity for the moment, causing a sharp drop in mine production (WPIC, 2020; USGS, 2021). Numerous suppliers of refined PGMs and PGM catalysts have halted production (WPIC, 2020). Jowitt (Jowitt, 2020) provided an overview of the effects of COVID-19 mitigation on the mining sector, indicating that it is more likely that a decreased demand leads to further lower metal prices with negative economic impacts on mining operations. As for downstream commodities of the supply chain, the pandemic caused declined production of both ICEV-related PGM catalysts and FCV-related PGM catalysts (Market Report, 2021; Hart and Jones, 2020). The global supply chain structures of PGM catalyst may face certain risks. Thus, understanding of the historical trajectory of the entire supply chain provides the basis to reduce such supply risks.Under such a circumstance, securing the raw material supply of PGMs and assessing associated supply risks received wide attentions (Schrijvers et al., 2020). Numerous studies have evaluated supply risks and vulnerabilities of PGMs (Rasmussen et al., 2019; Graedel et al., 2015; Hayes and McCullough, 2018; Mudd and Jowitt, 2017; Jowitt et al., 2018). Mudd et al. (Mudd et al., 2018) presented a global assessment of PGM resources and analyzed the key mining trends, indicating that key problems are not geological or resource depletion, but social, economic and environmental in nature like recent social issues in South Africa and volatile global economic conditions. Jowitt et al. (Jowitt et al., 2020) analyzed global metal reserves and found that primary PGM supply will not be exhausted within several decades. They suggested similar points of view that environmental, social, and governance factors are likely to be the main source of risk in metal supply. Yuan et al. (Yuan et al., 2020) introduced an assessment framework to analyze the criticality of platinum from 1975 to 2015, showing that the supply risk of platinum is strongly influenced by South Africa\u2019s socio-political status and dominance over global supply and reserves. All these studies show that PGM supply risks cannot be neglected which are mainly caused by the governance and environmental factors.Existing studies have laid a solid foundation for analyzing supply risks of PGM supply chain. However, they are mainly focused on the upstream stages including mining and refining. Considering manufacturing technical barriers, global production of PGM catalysts has concentrated in specific countries and regions, with significant barriers for capacity shifting (Islam et al., 2018). Identifying the supply risks of each stage along the supply chain becomes more crucial. The difference between the supply of the ICEV-related PGM catalysts and FCV-related PGM catalysts at the manufacturing stage needs more attention. Hence, this study provides a perspective of the entire supply chain, systematically assesses the global supply structures and risks of major stages throughout the production processes of PGM catalysts used in vehicles for the period 2010\u20132020. Traditional and new applications of PGM catalysts in the automotive industry are both considered, including ICEV-related PGM catalysts and FCV-related PGM catalysts. The government management and environmental factors are also considered in each stage. At the refining stage, this study takes the recycling into full consideration. The results indicate that significant supply risks exist in the PGM catalyst supply chain, especially in the mining stage of PGMs and the manufacturing stage of FCV-related PGM catalysts. Relevant policy suggestions are put forward to mitigate the supply risks of the supply chain.This study chooses the whole world as the spatial boundary. The main PGM mining, refining and catalyst production countries or regions with related explanations are shown in the supplementary document. The temporal boundary is set to 2010\u20132020. This period is chosen to reflect the status quo of the supply risks sufficiently.The entire PGM catalyst supply chain considered in this study covers three stages: mining, refining, and manufacturing. From the raw material perspective, PGMs are a major concern. Mining stage refers to the process from natural PGM resource to minerals. The mining activities always exert significant pressure on environmental and ecological systems (Kosai et al., 2021). Refining stage refers to the process from PGM minerals to intermediate products. The specific processes mainly include milling, concentrating, smelting, converting, separating and refining (Rasmussen et al., 2019). Manufacturing stage refers to the process from intermediate products to final products for end-use purpose, specifically referring to the process from refined PGMs to ICEV-related PGM catalysts and FCV-related PGM catalysts in this study (Sun et al., 2019).The supply risk normally refers to the probability of material supply disruption. The most widely used indicator is the diversity of supplying countries or regions, measured by the Herfindahl-Hirschman-Index (HHI) (Schrijvers et al., 2020). HHI is a comprehensive index to measure the degree of industrial concentration, which is widely used by economists and government regulatory departments (Silberglitt et al., 2013; Brown, 2018). In this study, we combine the HHI with the Worldwide Governance Indicator (WGI) and the Environmental Performance Index (EPI) respectively to quantitatively evaluate supply risks, focusing on the governance and environmental influence.The HHI is calculated by summing the squares of the market shares of fifty largest supply entities in a given market, as shown in Eq (1). It should be noted that, the supply entity refers to the country in this study.\n\n(1)\n\n\nHHI\n=\n\n\u2211\ni\n\n\nS\n\ni\n\n2\n\n\n\n\nwhere,\n\n\n\nS\ni\n\n\nis the market share of country i (in percentage unit).The WGI is an aggregate indicator for measuring six dimensions of governance for over 200 countries and regions over the period 1996\u20132019, including Voice and Accountability (VA), Political Stability and Absence of Violence (PV), Government Effectiveness (GE), Regulatory Quality (RQ), Rule of Law and Control of Corruption (RC) (World Bank, 0000). All the sub-indicators range from \u20132.5 (bad governance performance) to 2.5 (good governance performance). Among these dimensions, PV has a significant correlation with the stability of supply structures (van den Brink et al., 2020; Nassar et al., 2012; Nuss et al., 2014). The stability of supply structures significantly affects the probability of supply disruptions. The value of WGI_PV is scaled to 0\u20131 by using Eq (2).\n\n(2)\n\n\n\n\nW\nG\nI\n_\nP\nV\n\n\nscaled\n\n\n=\n-\n0.2\n\u00d7\nW\nG\nI\n_\nP\nV\n+\n0.5\n\n\n\n\nThe aggregate indicator of HHI and WGI_PV is defined as HHI-WGI. It takes both the diversity and stability of supplying countries into account. The HHI-WGI is calculated by Eq (3) (Mudd et al., 2018). The higher HHI-WGI is, the higher supply risk can be identified.\n\n(3)\n\n\nHHI\n-\nW\nG\nI\n=\n\n\u2211\ni\n\n\nS\n\ni\n\n2\n\n\u00d7\n\n\nW\nG\nI\n_\nP\nV\n\n\ni\n,\ns\nc\na\nl\ne\nd\n\n\n\n\n\nwhere,\n\n\n\nS\ni\n\n\nis the market share of country i (in percentage unit);\n\n\n\n\nW\nG\nI\n_\nP\nV\n\n\ni\n,\ns\nc\na\nl\ne\nd\n\n\n\n is the scaled WGI_PV value of country i.The EPI aggregates 24 indicators covering environmental health and ecosystem vitality, and measures environmental performance of 180 countries and regions (Wendling et al., 2018). It ranges from 0 (bad environmental performance) to 100 (good environmental performance). In the production processes, unsatisfactory environmental performance can cause supply risks (van den Brink et al., 2020). The reason is that such processes may lead to a level of environmental damage that society does not considers acceptable (Graedel et al., 2012). In such a situation, the production processes can be disrupted. The value is scaled to 0\u20131 by using Eq (4).\n\n(4)\n\n\n\n\nEPI\n\n\nscaled\n\n\n=\n\n\n100\n-\nE\nP\nI\n\n100\n\n\n\n\n\nThe aggregate indicator of HHI and EPI is defined as HHI-EPI. It considers the diversity of supplying countries and the environmental risk in the supply processes. The HHI-EPI is calculated by Eq (5). The higher HHI-EPI is, the higher environmental risk can be identified.\n\n(5)\n\n\nHHI\n-\nE\nP\nI\n=\n\n\u2211\ni\n\n\nS\n\ni\n\n2\n\n\u00d7\n\n\nEPI\n\n\ni\n,\ns\nc\na\nl\ne\nd\n\n\n\n\n\nwhere,\n\n\n\nS\ni\n\n\n is the market share of country i (in percentage unit);\n\n\n\n\nEPI\n\n\ni\n,\ns\nc\na\nl\ne\nd\n\n\n\n is the scaled EPI value of country i.To estimate the HHI, data are needed on PGM mining, refining and catalyst manufacturing at the national or regional level. Data on production of platinum and palladium in the mining stage are obtained from annual mineral yearbooks for the period 2010\u20132017, and mineral commodity summaries 2020 and 2021 by USGS (USGS, 2021; USGS, 2018; USGS, 2020). Data on production of rhodium in the mining stage are obtained from market reviews for the period 2010\u20132013, and PGM market report February 2021 of Johnson Matthey (Market Report, 2021; Market Report, 2013). In the refining stage, PGM primary and secondary supply data are supported by Steve Forrest Associates (SFA, 2020). The production of PGMs in this stage refers to the primary supply. When taking recycling into account, the PGM production refers to the total supply that can be calculated by Eq (6).\n\n(6)\n\n\n\n\nTS\n\ni\n\n=\n\n\nPS\n\ni\n\n+\n\n\nSS\n\ni\n\n\n\n\nwhere,\n\n\n\n\nTS\n\ni\n\n\nis the total supply of refined PGMs of country i;\n\n\n\n\nPS\n\ni\n\n\nis the primary supply of refined PGMs of country i;\n\n\n\n\nSS\n\ni\n\n\nis the secondary supply of refined PGMs of country i.This study uses the PGM mass contained in catalytic converters to represent the production of ICEV-related PGM catalysts in the manufacturing stage. Data on the global production distribution of ICEV-related PGM catalysts are calculated by Eq (7)-(10).\n\n(7)\n\n\n\n\nSA\n\ni\n\n=\n\n1\nn\n\n\u2217\n\n\u2211\nn\n\n\n\n\nAP\n\n\ni\n\nk\n\n\n\n\u2211\ni\n\n\n\nAP\n\n\ni\n\nk\n\n\n\n\n\n\n\n\n\n(8)\n\n\n\n\nAP\n\n\ni\n\nk\n\n=\n\n\nVP\n\n\ni\n\nk\n\n+\n\n\u2211\nj\n\n\n(\n\n\nNE\n\n\ni\n,\nj\n\nk\n\n-\n\n\nNI\n\n\ni\n,\nj\n\nk\n\n)\n\n\n\n\n\n\n\n(9)\n\n\n\n\nNE\n\n\ni\n,\nj\n\nk\n\n=\n\n\nEX\n\n\ni\n,\nj\n\nk\n\n\u2217\n\n\u03b2\n\nj\n\nk\n\n\n\n\n\n\n\n(10)\n\n\n\n\nNI\n\n\ni\n,\nj\n\nk\n\n=\n\n\nIM\n\n\ni\n,\nj\n\nk\n\n\u2217\n\n\u03b2\n\nj\n\nk\n\n\n\n\nwhere,\n\n\n\n\nSA\n\ni\n\n\n is the share of global ICEV-related PGM catalyst production for country i;\n\n\nn\n\n is the number of elements considered in this study;\n\n\n\n\nAP\n\n\ni\n\nk\n\n\n is the mass of element k contained in ICEV-related PGM catalysts produced in country i;\n\n\n\n\nVP\n\n\ni\n\nk\n\n\n is the mass of element k contained in vehicles produced in country i;\n\n\n\n\nNE\n\n\ni\n,\nj\n\nk\n\n\n is the mass of element k contained in commodity j exported from country i;\n\n\n\n\nNI\n\n\ni\n,\nj\n\nk\n\n\n is the mass of element k contained in commodity j imported to country i;\n\n\n\n\nEX\n\n\ni\n,\nj\n\nk\n\n\n is the mass of commodity j which contain element k exported from country i;\n\n\n\n\nIM\n\n\ni\n,\nj\n\nk\n\n\n is the mass of commodity j which contain element k imported to country i;\n\n\n\n\u03b2\n\nj\n\nk\n\n\n is the proportion of the mass of element k in commodity j of the international trade.In this study, data on PGM mass contained in vehicles are obtained from SFA\u2019s reports (SFA, 2020). The international trade data of relevant commodities which contain PGMs are mainly obtained from the United Nations Commodity Trade Database from 2010 to 2020 (UN, 2021). It should be noted that, when two countries report different trade data, the data from the country with high WGI_PV are used in this study. Information about these commodities and corresponding proportions of PGM mass are shown in Table 1\n.This study uses the platinum equivalent to represent the production of FCV-related PGM catalysts in the manufacturing stage. Country-level supplies are calculated by using a bottom-up approach as shown in Eq (11). Due to low production of FCVs, this study does not consider the stocks of these commodities.\n\n(11)\n\n\n\n\nTS\n\nj\n\n=\n\n\u2211\ni\n\n\n\nPV\n\n\ni\n,\nj\n\n\n\u2217\n\n\nSP\n\n\ni\n,\nj\n\n\n\u2217\n\n\nCD\n\n\ni\n,\nj\n\n\n-\n\n\nIM\n\nj\n\n\n\n\nwhere,\n\n\n\n\nTS\n\nj\n\n\n is the total supply of FCV-related PGM catalysts in all FCV types in country j;\n\n\n\n\nPV\n\n\ni\n,\nj\n\n\n\n is the production volume of FCV type i in country j;\n\n\n\n\nSP\n\n\ni\n,\nj\n\n\n\n is the fuel cell system power of FCV type i in country j;\n\n\n\n\nCD\n\n\ni\n,\nj\n\n\n\n is the demand of FCV-related PGM catalysts per unit power of FCV type i in country j;\n\n\n\n\nIM\n\nj\n\n\n is the total import of FCV-related PGM catalysts in country j.The major FCV suppliers and system powers are obtained from annual reports of numerous companies and literatures, which are shown in our previous study and the latest report (Xun et al., 2021; Yu, 2021). As for the demand of platinum per unit power of FCVs, this study estimates that it remained unchanged from 2015 to 2017, and from 2019 to 2020. Data on the demand from 2017 to 2019 can be found in technical reports presented by Xun (Xun et al., 2021). Relevant information about international trade are obtained from the same source.To estimate the HHI-WGI and HHI-EPI, data on scores of countries or regions over the period 2010\u20132020 are needed. They are obtained from open sources published by authorities (World Bank, 0000; Wendling et al., 2018). Detailed regional material production data, original and scaled WGI_PV and EPI index values with corresponding explanations are shown in the supplementary document.\nFig. 1\n shows the supply structures of each stage along the PGM catalyst supply chain in 2020, from which the reason for the status of supply risks can be found. The main producing countries for platinum are South Africa and Russia. For palladium, Russia and South Africa together occupy the main market supply. For rhodium, South Africa dominates the market, which means that the supply sources are more concentrated. According to the WGI_PV assessment results, South Africa and Russia are mid-ranking countries (ranked 138th and 153rd out of 214 countries and regions) (World Bank, 0000). As for EPI, Russia is at the relatively high level (ranked 52nd out of 180 countries and regions), while South Africa ranks much lower (ranked 142nd out of 180 countries and regions) (Wendling et al., 2018). Most ores are uneconomical for transportation because of low concentration (Reith et al., 2014). Smelters are mainly constructed in areas proximal to major mines like Noril'sk in Russia and the Bushveld in South Africa (USGS, 2018). The smelters and refiners could use both primary concentrates and recycled materials to make PGM powder or bars. The global supply structures for PGM refining are similar to those for mining.While in the manufacturing stage, the global supply structure for ICEV-related PGM catalysts is more diversified than supply structures of upstream materials. The main producing countries of ICEV-related PGM catalysts are the United States, China, Japan and Germany, which are also major producers of vehicles. The global supply structure of FCV-related PGM catalysts is very different to that of ICEV-related PGM catalysts. The main producing countries are also major producers of FCVs. The production in Korea accounts for more than 70% of global production, while the United States accounts for 13%. Other producing countries include China, Japan and Germany. These countries have better performance than South Africa and Russia in terms of the WGI_PV and EPI performances, thus the supply risks of the manufacturing stage are lower obviously.\nFig. 2\n shows global production situation of FCV-related PGM catalysts. Since the FCV industry is still at the initial stage of development, the gross production of FCV-related PGM catalysts is much smaller than that of ICEV-related PGM catalysts. The market-dominant country changed from Japan to Korea, which triggered such a tremendous change in supply risk of FCV-related PGM catalysts. The production in Korea had increased more than 150 times, from less than 2\u00a0kg platinum equivalent in 2015 to 326\u00a0kg in 2020. The production in Japan had fallen by 70%, from 47\u00a0kg to 14\u00a0kg during the period 2015\u20132020. The production in the United States, China and Germany surged from 2015 to 2019, while they suffered a sharp decline in 2020 affected by the COVID-19 pandemic.\nFig. 3\n(A) shows the HHI of the PGM catalyst supply chain from mining to manufacturing stage during the period 2010\u20132020. A larger HHI represents higher supply concentration, which implies a higher probability of supply disruption. It should be noted that, an HHI above 2500 indicates a highly concentrated market, while an HHI between 1500 and 2500 indicates a moderately concentrated market (Silberglitt et al., 2013). In the mining stage, the HHI for platinum, palladium and rhodium had been declining slowly in the recent decade, from 6103, 3490 and 7462 to 5198, 2689 and 6105, respectively. The reason could be that the Lac des Iles mine located in Canada have come onstream. Commercial production of PGM began in 1993 from the deposit known as the Roby zone, which until recently was exploited exclusively via open pit mining (Market Report, 2021). The rising production of PGM mine in countries other than South Africa has reduced supply concentration. The HHI for platinum, palladium and rhodium in the refining stage showed a similar trend, dropping from 6173, 3544 and 7144 in 2010 to 5291, 2999 and 5839 in 2020, respectively. While considering the secondary supply, there was a significant decline of HHI. The average HHI values for platinum, palladium and rhodium during the period dropped down by 39%, 32% and 43% respectively. It should be noted that, there was a sharp fall of HHI in 2020. The main reason could be that the COVID-19 led to a sharp drop in mine production of South Africa, with 10%, 13% and 28% decline for platinum, palladium and rhodium, respectively.In the manufacturing stage, the HHI were significantly different from those of the supply chain upstream. As for ICEV-related PGM catalysts, they were much lower and showed a downward trend, changing from 1433 in 2010 to 1079 in 2020. The HHI of FCV-related PGM catalysts were not present from 2010 to 2014 due to little global production. They decreased from 6760 in 2015 to 3185 in 2018, and then increased to 5937 in 2020. The main reason for such a trend was that the production in Korea had increased sharply from 2015 to 2020, while the production in Japan had experienced a significant decline. The production in other countries rose steadily.\nFig. 3(B) shows the HHI-WGI of the PGM catalyst supply chain in recent years. A larger HHI-WGI represents higher supply concentration with lower governance capability of supply countries, which implies a higher chance that supplies from these countries are cut-off (Nassar et al., 2020). Similarly to HHI, the HHI-WGI for platinum had decreased from 3122 to 2839 and from 3156 to 2888 in the mining and refining stages during the last decade respectively. The HHI-WGI for palladium in the mining and refining stages were approximately the same. It should be noted that, the values in the refining stage fell about 40% when taking the secondary supply into account. So were the HHI-WGI for rhodium, the values in the refining stage nearly equaled to those of the mining stage. In the manufacturing stage, the HHI-WGI values of ICEV-related PGM catalysts decreased by 15%. The HHI-WGI of FCV-related PGM catalysts dropped from 1972 to 1312 from 2015 to 2018, and rose up to 2407 in 2020.\nFig. 3(C) shows the HHI-EPI of the PGM catalyst supply chain in the last decade. A larger HHI-EPI represents higher supply concentration with higher environmental risk of supply countries, which implies a higher probability of causing negative environmental impacts in the supply processes. In the mining stage, the values for platinum, palladium and rhodium varied from 3339 to 2837, from 1579 to 1151, and from 4107 to 3348 in the last ten years respectively. The values in the refining stage were almost the same. In the manufacturing stage, the values of ICEV-related PGM catalysts were much lower than those in the upstream stages, changing from 591 to 351 during the period 2010\u20132020. The values of FCV-related PGM catalysts changed from 1725 to 944 during the period 2015\u20132018, and went up to 2220 in 2020. The HHI-EPI in the refining stage were much lower when recycling is considered, exactly as the HHI and the HHI-WGI.In the upstream stages of the PGM catalyst supply chain, the risk of rhodium was the highest among PGMs as Fig. 4\n shows. Considering the mean value of HHI from 2010 to 2020, the value for rhodium was 24% and 115% higher than that for platinum and palladium respectively in the mining stage, while 21% and 112% higher in the refining stage. As for the HHI-WGI, the value for rhodium was over 20% and 90% higher than that of platinum and palladium respectively in the mining and refining stages. And the average value of HHI-EPI for rhodium was higher than that for platinum and palladium in these stages. In the downstream stages of the PGM catalyst supply chain, the supply risk of ICEV-related PGM catalysts remained stable in the last decade. However, the supply risk of FCV-related PGM catalysts changed drastically, due to the change of global production situation as shown in Fig. 2. In the last decade, the production of FCV-related PGM catalysts was initially concentrated in Japan and gradually moved to Korea. Although the global productions of FCV-related PGM catalysts were significantly less than those of ICEV-related PGM catalysts, the supply risk could not be ignored. On the one hand, the higher technical threshold of FCV-related PGM catalysts led to the more concentrated supply structure. Many countries devote great effort to developing FCV-related PGM catalysts that have not been commercialized yet. On the other hand, building production capacity needs the initial investment of time and money. The rapid development of FCVs will increase the demand of fuel-cell-related PGM catalysts. In the event of a supply shortage, only a few countries are able to respond promptly and increase the production of them.Based on the above analysis, great supply risks are found in the entire PGM catalyst supply chain. There are countries that play an important role in this supply chain, producing substantial commodities at each stage. The inaccuracy of production statistics can have a significant impact on the results of supply risk assessments. Taking the United States and South Africa as an example, this study presents the sensitivity analysis of the production on the HHI, HHI-WGI and HHI-EPI of the major stages in the PGM catalyst supply chain as shown in Fig. 5\n. The United States is a typical country with strong production technology, while South Africa has abundant natural resources.This study takes the commodity outputs mentioned above along the PGM catalyst supply chain in 2020 as the baseline. The results show that the production of the United States mainly influence the supply risks of commodities in the downstream of the supply chain, while the production of South Africa mainly have impacts on upstream commodities. For instance, when assuming 10% higher production of ICEV-related PGM catalysts in the United States, the estimated HHI, HHI-WGI and HHI-EPI of ICEV-related PGM catalyst manufacturing stage change by 4%, 4% and 3% respectively. When assuming 10% lower production, the estimations change by \u22124%, \u22124% and \u22123% respectively. As for the upstream stages like refining, the results indicate that the assumption of 10% higher production of primary refined platinum in South Africa causes 5%, 5% and 5% changes of the estimated HHI, HHI-WGI and HHI-EPI respectively. While the assumption of 10% lower production causes \u22125%, \u22125% and \u22125% changes of those supply risk indicators respectively.As mentioned above, extensive studies have discussed whether the PGM resources are facing the problem of short supply. The balance between supply and demand has huge impact on supply risks. If the global demand of PGM catalysts used in vehicles cannot be met by the total PGM supply, supply risks are inevitable. So far, the consensus is that the primary PGM supply will most likely not to be the restriction for future development of the catalyst used in vehicles. However, as we found in our previous study, there could be significant supply risks due to resource location. The distributions of PGM reserves and demands are highly mismatched (Hao et al., 2019). The countries in need of PGMs are facing with the serious challenge of securing PGM supply. A high concentration of processing capacity leads to considerable potential supply risks, for the reason that the productions of commodities in the supply chain of PGM catalysts have higher technical threshold and are difficult to be replaced by alternatives once supply disruption occurs (CsA and Hausmann, 2009). Besides, the production processes are under restrictions of Environmental, Social and Governance (ESG) (Jowitt et al., 2020). The possibility of supply disruption can be reflected through the ESG performance of suppliers, which can be quantified by WGI_PV and EPI. Furthermore, the transport processes are seriously influenced by the ESG performance of suppliers. A sophisticated transportation logistics system had been built and maintained around the world. Although the system remains stable most of the time, it is also confronted with potential risks because the country with unsatisfactory ESG performance is more likely to interrupt transport.According to the results shown in Fig. 1, the global supply structures for PGM mining are highly concentrated. In the refining stage, the situation remains unchanged only considering the primary supply. While taking the secondary supply into consideration, the supply structure for palladium becomes moderately concentrated. As for the manufacturing stage, the supply structure for ICEV-related PGM catalysts becomes much less concentrated. But the supply structure of FCV-related PGM catalysts is highly concentrated. The supply risks along the entire supply chain of PGM catalysts used in vehicles cannot be ignored. Based on the analysis, several suggestions could be propounded with the aim of mitigating the risks of the PGM catalyst supply chain for countries seeking to secure the future supply of it.First, optimizing the supply structure of the upstream stages along the PGM catalyst supply chain could effectively reduce supply risks. The global supply structure for PGM mining is hard to change, for the reason that mining production distribution is the map of reserve and resource endowment to some extent (Ali et al., 2017). The richest mineral resources of PGMs are concentrated in South Africa and Russia. It is difficult to build up PGM ore stocks for the major consumers of the PGM catalyst including the United States, China, Japan, Korea and Europe. The total amount of PGMs in these countries or regions is relatively small and the ore grade is very low. While they can pay attention to overseas investment which can effectively reduce regional supply risks (Sun et al., 2019). For example, as one of the world\u2019s largest PGM mining companies, Anglo American Platinum Limited provided 34%, 21% and 36% of the global mining production of platinum, palladium and rhodium respectively. In fact, it is incorporated in the United Kingdom (Anglo, 2019). In addition to overseas investment, establishing national stocks of refined PGM products can also help these economically advanced countries or regions to reduce import dependences.Second, striving to develop secondary supply of PGMs could be another way to face the challenge. As the results show, the values of supply risk indicators for PGM refining decrease over 30% with the secondary supply considered. In fact, the global in-use stocks of PGMs have very promising recycling potential, accounting for more than 10% of natural resources in 2015 (Nassar, 2015). Meanwhile their geographical distribution is more consistent with the distribution of demand. However, the current recycling situation of PGM materials is not optimistic. The world is facing the challenges of low end-of-life recycling rates, especially in developing countries (Xun et al., 2020). The state-of-the-art recovery rates of PGMs are relatively high, while the collection rates are much lower due to complex factors, including management at the national level and recycling profits (Jha et al., 2013; Maes et al., 2016; Hagel\u00fcken et al., 2009). In order to improve recycling actuality, further measures need to be adopted, including strengthening the supervision of the recycling processes, constructing the recycling infrastructures and so on.Third, developing low-PGM catalysts used in vehicles could address the challenge of supply risks along the supply chain. Considerable efforts have been dedicated to the exploration in catalytic converters since the last century (Li et al., 2010; Kapteijn et al., 1993). As for the future demand, many countries have already made national plans to reduce the PGM loadings of fuel cells. For instance, the United States, China and Japan published the technology roadmap, setting the targets of reaching 0.1\u00a0g/kW by 2025 (Doe, 2017), 0.125\u00a0g/kW by 2030 (SAE-China, 2016), and 0.05\u00a0g/kW by 2030 (Nedo, 2017) respectively. At the same time, countries that are striving to develop FCVs need to improve the localization rate of PGM catalyst. They can consider to take specific measures, such as increasing policy and financial support to R&D, enhancing demonstrations and applications of new low-PGM catalyst technologies, stepping up infrastructure construction, and so on.Last but not the least, appropriate optimization of the PGM catalyst supply chain plays a positive role in reducing supply risks. However, risk reduction cannot be the sole objective. Measures like the re-shoring of the manufacturing sector help reduce supply risks, but they should be assessed cautiously. In the context of economic globalization, many companies have been moving their production capacity abroad through acquisitions and building plants overseas (Sun et al., 2019). These actions help to overcome many risks, including the lack of domestic natural resource and the restriction of human resource (Dachs et al., 2019). Under the circumstance of the COVID-19 pandemic, the importance of manufacturing capacity is fully exhibited before the world, which could accelerate the re-shoring of manufacturing sector. Although the re-shoring help to increase domestic employment and reduce national supply risks, it may affect the efficiency of production due to diseconomies of scale in the industry (Delis et al., 2019; Bailey and De Propris, 2014). Protecting the resilience of supply chains is not simply to bring all production home, but to play a large role in the global supply chains. Whether to adopt measures for the re-shoring of manufacturing sector needs to be based on the endowment of domestic relevant resources and product performances (Li and Zobel, 2020).This study provides insights into supply risks along the entire supply chain of the PGM catalyst used in vehicles during the period 2010\u20132020. The results show that there are significant supply risks embodied in current supply chain, especially in the mining stage. The HHI of platinum, palladium and rhodium in the mining stage changed from 6103, 3490 and 7462 in 2010 to 5198, 2689 and 6105 in 2020 respectively. Enhancing recycling of PGMs can effectively reduce supply risks in the refining stage, bringing about over 30% falls in the supply risks. As for the manufacturing stage, the HHI of FCV-related PGM catalysts decreased from 6760 in 2015 to 3185 in 2018, and then increased to 5937 in 2020. Comparing to that of ICEV-related PGM catalysts, the supply risk of FCV-related PGM catalysts was apparently higher. The results are of high relevance and importance to policy makers seeking to secure the future supply of conventional vehicles or FCVs.One limitation of this study is that a high concentration of processing capacity does not necessarily mean great supply risk. A clear mechanism between the selected indicators and a resulting supply risk cannot be provided by our model. Another is that since the FCV industry is still at the initial stage of development, the analysis of supply risks along the supply chain of FCV-related PGM catalysts does not necessarily reflect future development trends. Further efforts are needed to fill this research gap and provide more accurate estimations that can lead to more relevant policy implications.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This study is sponsored by the National Key R&D Program of China (2019YFC1908501), National Natural Science Foundation of China (72122010, 71774100).", "descript": "\n                  Platinum group metals (PGMs) are important catalytic materials for producing emission after-treatment devices in internal combustion engine vehicles (ICEVs) and fuel cells in fuel cell vehicles (FCVs). There are high resource and technical barriers along the supply chain of PGM catalysts, for which assessing the associated supply risks is essential to understand the potential challenges. This study maps the supply chains of PGM catalysts used in ICEVs and FCVs for the period of 2010\u20132020, and compares the associated supply risks. The results show that significant supply risks can be identified in both the upstream and downstream stages of PGM catalyst supply chains for ICEV and FCV uses. Specifically, the ICEV-related PGM catalyst supply chain maintained high levels of supply risks with slight decline over the assessed period. With recycling considered, supply risks in the refining stage can be reduced by over 30%. Supply risks embodied in the FCV-related PGM catalyst supply chain were relatively higher, with significant fluctuations during the period 2015\u20132020. Relevant suggestions are propounded aiming to mitigate supply risks of PGM catalysts used in ICEVs and FCVs.\n               "}
{"full_text": "Carbon Balance (%)Molar flow at the reactor inlet of the component i (\u03bcmol min-1)Molar flow at the reactor outlet of the component i (\u03bcmol min-1)Selectivity towards CH4 (%)CO2 stored (\u03bcmol g-1)Time (min)Temperature (\u00b0C)Sample mass (g)CH4 production (\u03bcmol g-1)CO production (\u03bcmol g-1)H2O production (\u03bcmol g-1)CO2 Capture and SequestrationCO2 Capture and UtilizationDual Function MaterialEnergy Dispersive SpectroscopyIntegration of CO2 Capture and UtilizationInductively Coupled Plasma Atomic Emission SpectroscopyLaNiO3 perovskiteNanoparticlesSynthetic Natural GasScanning Transmission Electron Microscopy-High Angle Annular Dark FieldThermal Conductivity DetectorTemperature Programmed DesorptionTemperature Programmed ReductionX-Ray DiffractionThe increase in global energy demand has led to a rapid growth in the fossil fuels consumption. As a result, the emission of greenhouse gases has been constantly increasing during the last decades, contributing to global warming and ultimately to climate change [1]. CO2, mainly emitted from the power generation sector and the industrial and transportation vehicles, is a major contributor to global warming due to its huge emission amounts [2,3]. Thus, the reduction of the CO2 emissions to the atmosphere is essential to limit global warming. In this context, carbon capture and sequestration (CCS) from industry and energy related sources as well as the increase in the efficiency of industrial processes and the widespread implementation of renewable energies, are expected to play an important role in overcoming this increasing problem [4]. However, CCS technology requires captured CO2 purification and transport to storage places and its isolation, which increases drastically the cost and the energy consumption of the process [5,6].During the last years, there is keen interest in the integration of CO2 capture and its utilization (ICCU), since this technological alternative allows reducing the cost of the overall process by eliminating transportation and storage of CO2 by its conversion to fuels or value-added chemicals [7]. Farrauto et al. [8,9] have patented in 2015 the use of dual function materials (DFMs) to convert the captured CO2 from diluted exhaust gases into methane in a single reactor. Such ICCU process can be even greener when is carried out with hydrogen obtained by the electrolysis of water using surplus renewable energies, contributing at the same time to store excess electrical energy in the form of methane. Therefore, ICCU-methanation technology reduces CO2 emissions to the atmosphere and contributes to solve the problem of intrinsic intermittency of renewable sources [10]. Moreover, the cycling operation, in contrast to the observed for the continuous hydrogenation of CO2, can be directly applied to an effluent gas without the necessity of additional heat input to perform the CO2 capture and does not need purification steps, which reduces the global costs of the process [11].The selected DFM should selectively capture CO2 from steam- and O2-containing flue gas at different temperatures (200\u2013550\u00a0\u00b0C), depending on the application and effluent gas properties; and then hydrogenate the adsorbed species to methane with H2 in a carbon neutral cycle. The overall CO2 adsorption-hydrogenation process follows the stoichiometry of the Sabatier reaction [12], which is thermodynamically favoured at low temperature due to its strong exothermicity.\n\n(1)\n\n\n\nCO\n2\n\n\n\n+\n\n\n\n\n4H\n2\n\n\u21c4\n\nCH\n4\n\n\n\n+\n\n\n\n\n2H\n2\n\nO\n\n\n\n\n\n\n\n\n\u0394\n\n\nH\n\n0\n\n=\n-\n164\n\nkJ\n\n\n\nmol\n\n\n\n- 1\n\n\n\n\n\n\nHowever, the stable electronic structure of the CO2 molecule makes its activation difficult under mild conditions, such as low temperature and low pressure. The use of high reaction temperatures favours the kinetics of CO2 to CH4 conversion; however, it contributes to increase the equipment investment as well as the operational cost, which is undesirable for large-scale industrial utilization. Therefore, a high-performance catalyst that can activate CO2 and promote the reaction rate under relative low temperatures and pressures is vital for its widespread implementation. Based on the characteristics of dual operation, these catalysts require the presence of a storage material for CO2 capture and an active site for H2 activation and CO2 hydrogenation to methane. Regarding to the CO2 adsorption functionality in DFMs, a wide variety of alkali/alkaline-earth phases have been proposed as CO2 storage material, mainly Na [13,14], Ca [8,15], Mg [4,16] or K [4,17]. These storage components should be capable to reversibly operate at intermediates-high temperatures (200\u2013450\u00a0\u00b0C) [18]. On the other hand, the catalytic and hydrogenation sites are usually based on Ni [15,19\u201321], Ru [13,14,22,23] or Rh [24] metals. Among them, Ni presents the best cost to activity ratio, which makes this alternative most suitable for industrial applications. Finally, both phases are usually dispersed on a high surface area carrier in order to increase the methane production. In this sense, previous studies reported that \u03b3-Al2O3 is the most appropriate support among other materials [19,25].Ni-based catalysts prepared by conventional preparation methods often lead to large and heterogeneous particle size distribution, which limits the control over the interaction between the metal nanoparticles (NPs) and the support [26]. Hence, Ni-based catalysts present limited catalytic activity at low temperatures and can be easily deactivated due to the metal sintering occurring at high temperatures. Furthermore, Ni-based DFMs have been considered only for process at intermediate-high temperatures, since Ni can be readily oxidized during the CO2 adsorption period but not easily reduced back during the hydrogenation step at low temperatures [11,19]. Intense efforts have been made in order to design and improve Ni-based catalysts for their application as dual function materials (DFMs). The catalytic behaviour of the Ni-based materials depends on several factors such as the type of support, Ni loading, addition of a second metal and preparation method [27,28,29].Largely based on the pioneering research of Daihatsu and Toyota, the ex-solution of active metal NPs from an oxide host, such as perovskite-type lattice, has been identified as a simple way to achieve a homogeneous active sites distribution, with good reversibility and controlled interactions between metal and the support [30]. This concept has been already explored for controlling Ni particle sizes and distribution of the catalysts used in the stationary CO2 methanation process [31\u201333]. Specifically, LaNiO3-type perovskites, partially doped with different components (i.e., Ce, K or Ca), have been proposed as promising host materials to carry out the inside-outside ex-solution of Ni NPs. Nevertheless, non-supported perovskites exhibited rather low surface areas, which could limit the active sites dispersion and the reaction intermediates diffusion. To address the aforementioned limitations, Li et al. [34] and Wang et al. [35] distributed LaNiO3-type perovskites on silica supports. The obtained catalysts have shown improved CO2 methanation efficiency. Based on the well-known promoting effect observed for ceria in its application to the conventional Ni/CeO2 catalyst for the stationary CO2 methanation [36\u201338], we recently explored the viability of ceria-supported LaNiO3 perovskites as precursor of highly active and stable materials for the continuous CO2 methanation [39]. These catalysts present notably higher methane production than the conventional Ni/CeO2 catalyst and that obtained from the bulk LaNiO3 in the stationary CO2 to CH4 hydrogenation process. Nevertheless, to the best of the authors' knowledge, the use of LaNiO3 perovskite as precursor of highly efficient DFM material for CO2 capture and hydrogenation to methane has not been published to date.Considering this background, the aim of this work is to evaluate for the first time in the scientific literature the applicability of supported LaNiO3 perovskites, as precursors of efficient dual function materials for CO2 adsorption and in-situ hydrogenation to methane. For that, the previously developed 30% LaNiO3/CeO2 formulation as well as others here synthesized over conventional DFM supports, such as Al2O3 and La-Al2O3, are evaluated in cycles of CO2 adsorption and hydrogenation to CH4. Taking into account the characterization results, the interrelationships between physico-chemical properties, activity, and stability are discovered.Prior to the preparation of perovskite-based formulations, different supports were obtained. On the one hand, the ceria support was obtained by direct calcination of the Ce(NO3)3\u00b76H2O (Sigma Aldrich, 99.9%) precursor at 500\u00a0\u00b0C for 4\u00a0h in static air. On the other hand, the 5\u00a0wt% La-Al2O3 support was obtained by wetness impregnation method over previously calcined \u03b3-Al2O3 (650\u00a0\u00b0C, 2\u00a0h). For that, the amount of La(NO3)2\u00b76H2O (Merck, 99.0%), neccesary to obtain a 5\u00a0wt% of La2O3 over the support, was incorporated onto \u03b3-Al2O3 (Saint Gobain, SA6173) inside a rotary evaporator (vacuum and 35\u00a0\u00b0C).Once different supports (Al2O3, La-Al2O3 and CeO2) were obtained, supported perovskites were prepared by combining citric acid and impregnation methods, as reported elsewhere [39]. For that, nominal perovskite loading of 30\u00a0wt% was impregnated over these supports. The adopted nomenclature for the fully formulated samples was the following: LNO, 30% LNO/CeO2, 30% LNO/Al2O3 and 30% LNO/La-Al2O3. Note that this nomenclature corresponds to the precursors of the catalysts for CO2 methanation reaction. In order to obtain different DFMs, these precursors were in-situ reduced in the reaction bench.X-ray diffraction (XRD) analyses of the fresh and used samples were carried out using a Philips PW1710 diffractometer. For that, all samples were subjected to Cu K\u03b1 radiation in a continuous scan mode in the 2\u03b8 range 5\u201370\u00b0 with 0.02\u00b0 per second sampling interval. PANalytical X\u2018pert HighScore and Winplotr profile fitting software were used for data treatment. ICDD (International Centre for Diffraction Data) database cards were used for comparative purposes to identify the phases present in the samples.Scanning Transmission Electron Microscopy - High Angle Annular Dark Field (STEM-HAADF) images were taken for the samples after reduction and CO2 methanation reaction with a Cs-image-corrected Titan (Thermofisher Scientific). This equipment operated at a working voltage of 300\u00a0kV, and was equipped with a CCD camera (Gatan) and a HAADF detector (Fischione). The instrument has a normal field emission gun (Shottky emitter) equipped with a SuperTwin lens. Alternatively, the TEM apparatus was also equipped for X-ray Energy Dispersive Spectroscopy (EDS) experiments with an Ultim Max detector (Oxford Instruments). A 2\u00a0k\u00a0\u00d7\u00a02\u00a0k Ultrascan CCD camera (Gatan) was positioned before the filter for TEM imaging, using an energy resolution of 0.7\u00a0eV. The acquisition time for the analysis was 50\u00a0ms per spectrum and the used energy dispersion was 0.2\u00a0eV pixel\u22121. Prior to these experiments, the samples were sonicated in ethanol and dropped onto a holey, amorphous carbon film supported on a copper grid.Textural properties of the fresh and used samples, that is, after controlled reduction and CO2 methanation, were determined by N2 adsorption\u2013desorption isotherms at \u2212196\u00a0\u00b0C, using a Micromeritics TRISTAR II equipment. All samples were pretreated with flowing N2 on a Micromeritics SmartPrep instrument at 300\u00a0\u00b0C for 10\u00a0h.La, Ni, Al and Ce contents were quantitatively determined by Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP\u2013AES). The analyses were carried out using a mass spectrometer with a plasma source (Q-ICP-MS XSeries II model of Thermospectrometer). Prior to the analysis, the samples were digested at 120\u00a0\u00b0C with an acid mixture (HNO3:HCl\u00a0=\u00a01:3) inside a microwave furnace.The redox behaviour of the fresh samples was examined by means of Temperature Programmed Reduction with a 5% H2/Ar mixture (H2-TPR) in a Micromeritics AutoChem II equipment. For that, the quartz tube reactor was loaded with 0.1\u00a0g of sample, which was pretreated in 30\u00a0mL\u00a0min\u22121 of 5% O2/He mixture at 600\u00a0\u00b0C for 30\u00a0min. Then, the sample was cooled down to 35\u00a0\u00b0C under inert conditions, and finally the temperature was increased from 35 to 950\u00a0\u00b0C in a 5% H2/Ar mixture (30\u00a0mL\u00a0min\u22121) using a heating rate of 10\u00a0\u00b0C\u00a0min\u22121. Water generated during samples reduction was removed by condensation in a cold trap placed before TCD detector. The outlet gas stream was continuously monitored with a Hiden Analitical HPR-20 EGA mass spectrometer.The basicity of the samples was evaluated by Temperature Programmed Desorption of CO2 (CO2-TPD) experiments, which were carried out in a Micromeritics AutoChem II equipment. The quartz tube reactor was loaded with 0.15\u00a0g of the fresh samples. Aiming to obtain a catalytic material similar to that analyzed in the activity test, bulk perovskite and ceria- and alumina-supported samples were completely reduced in a 5% H2/Ar mixture (50\u00a0mL\u00a0min\u22121) at 650, 550 or 800\u00a0\u00b0C (2\u00a0h), respectively. Then, the reduced samples were cooled down to 40\u00a0\u00b0C, under inert conditions (He flow stream). Once this temperature was reached, the adsorption of CO2 was performed by exposing the samples to a 5% CO2/He flow stream (50\u00a0mL\u00a0min\u22121) for 60\u00a0min. Finally, the samples were heated from 40 to 900\u00a0\u00b0C at 10C min\u22121 in He (50\u00a0mL\u00a0min1) and the desorbed gases were continuously monitored with a Hiden Analitical HPR-20 EGA mass spectrometer.CO2 adsorption and hydrogenation cycles were carried out in a vertical stainless steel tubular reactor inside a 3-zone tube furnace. The reactor was filled with 1.0\u00a0g of pelletized (0.3\u20130.5\u00a0mm) fresh formulation, where the operating temperature was continuously measured through a thermocouple placed in the centre of the catalytic bed. Prior to the catalytic test, fresh samples were in-situ reduced with a stream composed of 10% H2/Ar leading to the conformation of the final DFM due to the controlled reduction of perovskite-based formulation. With that aim, the temperature was progressively increased from room temperature to 650, 550 or 800\u00a0\u00b0C (2\u00a0h) for bulk and ceria- and alumina-supported samples, respectively. Note that the reduction temperature for each support was already optimized in a previous study [39].Once the DFM was obtained, CO2 adsorption and hydrogenation experiments were carried out, increasing the reaction temperature progressively from 280 to 520\u00a0\u00b0C, in steps of 40\u00a0\u00b0C. During the adsorption period (60\u00a0s), the feed composition was 10% CO2/Ar. Then, this step was followed by a purge with Ar (120\u00a0s) to remove weakly adsorbed CO2 and prevent mixing of streams. Finally, CO2 was replaced by a 10% of H2 during the hydrogenation (methanation) period (120\u00a0s). Before starting the following CO2 adsorption period, the catalyst and the system were again purged with Ar for 60\u00a0s. The catalytic tests were carried out with a total flow rate of 1200\u00a0mL\u00a0min\u22121. This flow corresponds to space velocities of around 45,000 and 140,000\u00a0h\u22121 for ceria- and alumina-supported samples, respectively. CO2, CH4, CO and H2O were continuously quantified by a MKS MultiGas 2030 FT-IR analyser.The amount of CO2 stored was calculated from Eq. (2). With that aim, the amount that leaves the reactor must be subtracted from the amount fed. To determine the amount of CO2 fed, the stream from the feed system was led directly to the analyser. The obtained profile corresponds to the actual CO2 input that was fed to the reactor.\n\n(2)\n\nS\nT\n\n\nO\nCO\n\n2\n\n\n\n\n\u03bcmol\n\n\ng\n\n-\n1\n\n\n\n\n=\n\n1\nW\n\n\n\u222b\n0\nt\n\n\n\n\nF\n\nCO\n2\n\nin\n\n\nt\n\n-\n\nF\n\nCO\n2\n\nout\n\n\nt\n\n\n\n\nd\nt\n\n\n\nOn the other hand, the CH4, CO and H2O productions were calculated from the following expressions:\n\n(3)\n\n\nY\n\nCH\n4\n\n\n\n\n\n\u03bcmol\n\n\ng\n\n-\n1\n\n\n\n\n=\n\n1\nW\n\n\n\u222b\n0\nt\n\n\nF\n\nCH\n4\n\nout\n\n\nt\n\nd\nt\n\n\n\n\n\n(4)\n\n\nY\nCO\n\n\n\n\n\u03bcmol\n\n\ng\n\n-\n1\n\n\n\n\n=\n\n1\nW\n\n\n\u222b\n0\nt\n\n\nF\nCO\nout\n\n\nt\n\nd\nt\n\n\n\n\n\n(5)\n\n\nY\n\n\nH\n2\n\nO\n\n\n\n\n\n\u03bcmol\n\n\ng\n\n-\n1\n\n\n\n\n=\n\n1\nW\n\n\n\u222b\n0\nt\n\n\nF\n\n\nH\n2\n\nO\n\nout\n\n\nt\n\nd\nt\n\n\n\nCH4 selectivity is determined by relating the CH4 and CO productions since they were the only two products that were detected:\n\n(6)\n\n\n\nS\n\nCH\n4\n\n\n\n\n%\n\n\n=\n\n\nY\n\nCH\n4\n\n\n\n\nY\n\nCH\n4\n\n\n+\n\nY\nCO\n\n\n\n\u00d7\n100\n\n\n\n\nFinally, the carbon balance check was carried out from the following expression:\n\n(7)\n\nC\nB\n\n%\n\n=\n\n\n\n\nY\n\nCH\n4\n\n\n+\n\nY\nCO\n\n\n\nS\nT\nO\n_\nC\nO\n2\n\n\n\n\u00d7\n100\n\n\n\n\nFig. 1\n includes XRD patterns of 30% LaNiO3/CeO2, 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3 samples before (a) and after (b) CO2 methanation reaction. The corresponding XRD diffractograms of ceria (CeO2) and alumina (Al2O3) supports as well as of bulk perovskite (LNO) are also included as reference.Regarding to fresh samples (Fig. 1a), intense diffraction peaks (\u25cb) at 28.6, 33.1, 47.5, 56.3 and 59.1 \u00b02\u03b8 are observed for ceria support, whereas wide peaks (+) at 7.6, 45.9 and 67.0 \u00b02\u03b8 are identified for alumina support. These reflections are characteristic of a cubic highly crystalline ceria and an amorphous cubic alumina phases, respectively. On the other hand, the bulk perovskite (LNO) shows three mains diffraction peaks (\u0394) at 32.9, 47.4 and 58.7 \u00b02\u03b8, which are characteristic of a rhombohedral LaNiO3 phase. Furthermore, this sample also shows additional peaks in form of impurities, characteristic of hexagonal La2O2CO3 (\u25cf), cubic NiO (\u25a1) and tetragonal La2NiO4 (\u25b2) phases, respectively. Their presence suggests that a fraction of Ni2+ and La3+ is not inserted inside the perovskite structure during perovskite structure conformation, due to a limited stability of LaNiO3 oxide.Supported samples (30% LaNiO3/CeO2, 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3) show an intermediate diffractogram to that described for the bulk perovskite and the corresponding support. However, it is worth mentioning that the peaks of the ceria support overlap those of the perovskite in the case of the 30% LaNiO3/CeO2 sample. In any case, the small displacement of the peak situated at 33.1 \u00b02\u03b8 to lower 2\u03b8 positions with respect to bare ceria support suggests the coexistence of both phases in the supported samples (Figure S1). Finally, it can be noticed that the relative intensity of the characteristic diffraction peaks of La2O2CO3 (\u25cf), NiO (\u25a1), and La2NiO4 (\u25b2) impurities, increases for Al2O3- and La-Al2O3-supported samples with respect to bulk perovskite and CeO2-supported samples. These results suggest that the LaNiO3 conformation is partially limited over alumina-supported samples, especially for bare Al2O3 support.Once LaNiO3 perovskite is conformed, Ni should be ex-solved from the perovskite host with controlled size to conform the desired DFM. In order to confirm the Ni nanoparticles (NPs) ex-solution, XRD measurements were carried out for the samples used in cyclic CO2 adsorption and in-situ hydrogenation (Fig. 1b). Note that these samples were in-situ reduced prior to the catalytic test at the temperatures specified in Section 2.3. As can be observed, all samples show intense diffraction peaks of corresponding supports (CeO2 or Al2O3), which confirm their high stability. In contrast, no diffraction peaks are discernible for LaNiO3, NiO and La2NiO4 phases. These results confirm the complete reduction of NiO, LaNiO3 and La2NiO4 phases, leading to cubic Ni0 (\n\n) and La2O3 formation. However, an increase in the intensity of La2O2CO3 diffraction peaks is observed, instead of La2O3 phase identification. In agreement with that reported in previous works [40,41], this fact is due to CO2 adsorption on La2O3 sites during CO2 methanation. Hence, XRD results demonstrate the controlled ex-solution of Ni0 nanoparticles from the LaNiO3 during the controlled reduction process. Ultimately, DFMs are obtained with the following general formulation: Ni-La2O3/support (with support\u00a0=\u00a0Al2O3, La-Al2O3 or CeO2).The Ni0 crystallite sizes are determined by applying the Scherrer equation to the peak located at 51.8 \u00b02\u03b8 in the used samples (Table 1\n). As can be observed, the crystallite size of the 30% LNO/CeO2 sample is 7.0\u00a0nm, whereas it increases to 12.4 and 11.5\u00a0nm for 30% LNO/Al2O3 and 30% LNO/La-Al2O3 samples, respectively. In agreement with XRD results (Fig. 1a), the ceria support favours the formation of a higher proportion of LaNiO3 perovskite, instead of impurities; as a consequence, this fact increases the Ni3+ available to be ex-solved, in form of smaller Ni0 NPs, during the reducing step. In any case, these values are significantly lower than that observed for the bulk perovskite (31.7\u00a0nm). Thus, supporting the LaNiO3 perovskite over different nature supports seems to be an efficient way to promote the ex-solution of Ni NPs with smaller crystallite size.Different phase\u2019s distribution of 30% LaNiO3/CeO2, 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3 used samples were analyzed by STEM-HAADF images and representative EDS elemental maps (Fig. 2\n). Furthermore, their Ni particle sizes distribution, estimated by measuring the size of at least 100 particles identified in these images, is also included in the form of a histogram. As a general trend, Ce (dark blue colour) or Al (light blue colour) and La (green colour) elements coexist with homogeneous distribution in all analyzed areas. Furthermore, small-sized Ni NPs (red colour) uniformly distributed on La and Ce or Al surface can be identified irrespective the analyzed support. Nevertheless, the Ni particle\u2019s agglomeration is higher for the alumina-supported samples (Fig. 2b and c), which leads to a more heterogeneous Ni particle size distribution in the right side histogram. In agreement with the Ni size estimated by XRD experiment (Table 1), the lowest average size (5.0\u00a0nm) corresponds to 30% LaNiO3/CeO2 sample (Fig. 2a), whereas both alumina-supported samples show a Ni average size above 10\u00a0nm. These observations evidenced that the utilization of ceria as support favours the formation of a DFM with smaller Ni NPs. As previously suggested by XRD diffractograms (Fig. 1), the LaNiO3 perovskite formation is favoured with respect to the formation of impurities (i.e. La2O2CO3, NiO and La2NiO4), favouring the ex-solution of Ni0 NPs with smaller size from the perovskite host. Furthermore, the ceria-supported sample requires lower reduction temperature (550\u00a0\u00b0C vs. 800\u00a0\u00b0C) to completely ex-solve the Ni0 NPs from the different Ni-based phases, which limits their sintering during reduction step. Finally, these aspects seem to favour a more homogenous distribution of the La-based phases for ceria-supported samples.The analysis of the main textural properties of the used samples was carried out by isothermal (\u2013196\u00a0\u00b0C) N2-adsorption\u2013desorption. As expected, all perovskite-based formulations show type IV isotherms (Figure S2) according to the IUPAC classification, which are characteristic of mesoporous materials. Table 1 summarizes the corresponding specific surface areas (S\nBET) and pore volumes (V\np) for the used and fresh samples (in brackets). 30% LaNiO3/CeO2, 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3 samples present S\nBET of 37, 103 and 105\u00a0m2 g\u22121, respectively. This trend is ascribed to the progressive overlap of the pores and the support\u0301s surface by the deposition of the LaNiO3 perovskite, together with pores narrowing due to calcination as well as reduction at high temperatures, in line with the proportional decrease in pore volume observed. In any case, these values are much higher than that of the bulk perovskite (12\u00a0m2 g\u22121), which contributes to the ex-solution of smaller Ni NPs.In order to investigate the redox properties of the samples, Fig. 3\n shows the hydrogen consumption profiles, normalized per sample mass unit, for fresh 30% LaNiO3/CeO2, 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3 samples. For comparative purposes, the H2\u2013TPR profiles of bulk LaNiO3 perovskite (LNO) and CeO2 support are also included.In agreement with the results in our previous work [39], the ceria-supported perovskite shows an intermediate reduction profile between the CeO2 support and bulk LaNiO3 perovskite. Specifically, two main H2 consumption regions can be identified, i.e. below and above 650\u00a0\u00b0C. As observed for the bulk LaNiO3, the low temperature region presents three main peaks centred at 225, 375 and 475\u00a0\u00b0C, which are ascribed to the progressive reduction of NiO, LaNiO3 and La2NiO4 phases following the stoichiometry of Eqs. (8\u201310). Note that the Ce4+ at the surface is also reduced in this temperature region. Meanwhile, the peak above 650\u00a0\u00b0C is ascribed to the final reduction of the bulk CeO2 (Eq. (11)) [42].\n\n(8)\n\n\n\n4LaNiO\n3\n\n\n\n+\n\n\n\n\n2H\n2\n\n\u2192\n\nLa\n4\n\n\nNi\n3\n\n\nO\n10\n\n\n\n+\n\n\n\n\n\nNi\n\n0\n\n\n\n\n+ 2H\n\n2\n\nO\n\n\n\n\n\n\n(9)\n\n\n\nLa\n4\n\n\nNi\n3\n\n\nO\n10\n\n\n\n+\n\n\n\n\n3H\n2\n\n\u2192\n\nLa\n2\n\n\nNiO\n4\n\n\n\n+\n\n\n\n\n\n2Ni\n\n0\n\n\n\n+\n\n\n\n\nLa\n2\n\n\nO\n3\n\n+\n\n\n2H\n2\n\nO\n\n\n\n\n\n\n(10)\n\n\n\nLa\n2\n\n\nNiO\n4\n\n\n\n+\n\n\n\n\nH\n2\n\n\u2192\n\n\nNi\n\n0\n\n\n\n+\n\n\n\n\nLa\n2\n\n\nO\n3\n\n+\n\n\nH\n2\n\nO\n\n\n\n\n\n\n(11)\n\n\n\n2CeO\n2\n\n\n\n+\n\n\n\n\nH\n2\n\n\u2192\n\nCe\n2\n\n\nO\n3\n\n\n\n+\n\n\n\n\nH\n2\n\nO\n\n\n\n\nThe reduction profiles of alumina-supported samples (30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3) differ from that observed for ceria-supported sample. On the one hand, the reduction of NiO, LaNiO3 and La2NiO4 phases (Eqs. 8\u201310) takes place more progressively. On the other hand, a new contribution, centred around 750\u2013800\u00a0\u00b0C, is observed. Taking into account that Al2O3 and La-Al2O3 support are no reducible, this peak is associated with the reduction of highly stable NiAl2O4 phase formed during calcination step [43]. These results denote a high interaction between the Ni and alumina support, which limits the conformation of the perovskite oxide and favours the presence of impurities, as previously suggested by XRD analysis (Fig. 1).As can be observed, the reducibility is clearly enhanced for ceria-supported samples with respect to those observed for the bulk perovskite and the ceria support. In contrast, this shift is limited for the alumina-supported samples, especially for the perovskite deposited onto bare alumina. However, it is worth to mention that species below 250\u00a0\u00b0C are more easily reduced for alumina-supported samples than for ceria-supported and bulk perovskites. As previously suggested, the reduction of NiO, not inserted in the perovskite lattice, also occurs at this temperature region and is favoured by the higher specific surface area of alumina-supported samples with respect to the ceria-supported one. In any case, the concentration of the species reduced in the low temperature region (below 650\u00a0\u00b0C) is significantly higher for the ceria-supported samples. These results evidence that the redox properties of the samples are favoured with ceria as support, which leads to an easier reduction of Ni-based species as well as of ceria support [44]. Thus, synergetic effects between LaNiO3 and ceria phases are evidenced, which promote the accessibility of the former and the reducibility of the latter due to spill-over effect of activated H2\n[32].To gain insight on the hydrogen consumption occurred during H2-TPR experiments, the effluent gas was analyzed by mass spectroscopy for ceria- and alumina-supported samples (Figure S3). As can be observed, a noticeable methane formation can be identified in both cases due to the hydrogenation (CO2\u00a0+\u00a04H2\u00a0\u21c4\u00a0CH4\u00a0+\u00a0H2O) of the CO2 released due to La2O2CO3 decomposition on Ni0 sites, specie formed during the perovskite reduction at lower temperatures. As a result, this process implies the consumption of additional H2. Thus, the H2 consumption observed between 250 and 600\u00a0\u00b0C is not only due to the reduction of reducible species but also due to the methane formation of the adsorbed CO2 at the surface. Among different supports, the 30% LNO/CeO2 sample shows the highest CH4 production below 400\u00a0\u00b0C. This trend suggests that the activation of CO2 methanation takes place at lower temperatures for the DFMs obtained from the ceria-supported sample, in line with the higher reducibility observed during H2-TPR experiments (Fig. 3).\nTable 2\n shows the integrated area related to the reduction of the different species per gram of sample. Based on the reduction steps described in Eqs. (8\u201310), 1.5\u00a0mol of hydrogen are consumed per 1\u00a0mol of LaNiO3 perovskite, whereas 0.5\u00a0mol of hydrogen are consumed in the reduction of CeO2 support (Eq. (11)). In contrast, no hydrogen consumption is expected due to the La-Al2O3 or Al2O3 supports reduction. Moreover, noticeable hydrogen consumption is related to carbonates reduction, in line with the high CH4 production identified in Figure S3. In fact, this contribution should be more relevant for ceria-supported samples. As a result, the overall H2 uptake decreases from 7507\u00a0\u00b5mol H2 g\u22121 for ceria-supported samples to values below 4745\u00a0\u00b5mol H2 g\u22121 for alumina-supported samples.It is worth to mention that the increase in H2 consumption for the 30% LNO/CeO2 sample is especially remarkable below 400\u00a0\u00b0C (Table 2). In order to explore in more detail this aspect, Fig. 4\n plots the evolution of the H2 uptakes below 250\u00a0\u00b0C related to Ni content for different supported samples and bulk perovskite. Note that this hydrogen consumption was previously assigned to the partial reduction of Ni3+ in the perovskite lattice together with the reduction of highly dispersed NiO nanoparticles, since no CH4 formation is observed in this temperature region (Figure S3). As can be observed, the H2/Ni ratio progressively decreases from 0.82 for 30% LNO/CeO2 sample to 0.23 for 30% LNO/La-Al2O3 sample. This result confirms that the H2 activation during CO2 methanation is promoted at lower temperatures by the use of ceria support due to the increase in the concentration of highly reducible Ni-based species in LaNiO3 perovskite.To investigate the interaction between the CO2 molecule and the catalyst surface, the CO2\u2013TPD profiles of the LaNiO3, 30% LaNiO3/CeO2 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3 samples, as well as those of the corresponding supports (CeO2, Al2O3 and La-Al2O3) were compared (Fig. 5\n). Each sample was prereduced following a similar procedure to that achieved prior to catalyst test (Section 2.3). On the one hand, ceria support shows a single desorption peak centred at 100\u00a0\u00b0C, which is assigned to the CO2 decomposition arising from bridged and bidentate carbonates adsorbed onto ceria surface. Meanwhile, the weak signal, observed above 400\u00a0\u00b0C, is assigned to the decomposition of carbonates not eliminated during calcination and reduction steps [45]. On the other hand, bare Al2O3 and La-Al2O3 supports present an asymmetric desorption peak at 100\u00a0\u00b0C, which is assigned to CO2 desorption from weak Br\u00f6nsted OH\u2013 groups [43]. It is worth to note that the shoulder at higher temperatures is slightly higher for La-Al2O3 support, due to the CO2 desorption from monodentate carbonates adsorbed on highly dispersed La2O3 formed on catalytic surface.Regarding bulk perovskite, three main desorption peaks can be observed: below 200\u00a0\u00b0C, between 200 and 550\u00a0\u00b0C and above 550\u00a0\u00b0C. In increasing order of temperature, these peaks are assigned to the decomposition of weakly adsorbed CO2 on Ni0 sites [46], and decomposition of monodentate carbonates linked to highly dispersed and bulk-like La2O3 species in the form of La2O2CO3\n[27,44], respectively. As expected, supported perovskites show an intermediate CO2 desorption profile to that observed for the bulk LaNiO3 perovskite and the corresponding support. Nevertheless, two main differences can be identified. On the one hand, the desorption of the different adsorbed species takes place at lower temperatures with respect to bulk perovskite. On the other hand, a more progressive CO2 decomposition during the whole temperature range is favoured for ceria-supported sample. As previously discussed, the impregnation of the LaNiO3 perovskite on a high surface area supports limits its agglomeration during calcination. This fact promotes a more homogeneous distribution of the La2O3 phases at the surface, which favours the formation of monodentate carbonates of different stability on the La2O3 sites. However, this process is partially limited for alumina-supported samples due to the formation of NiAl2O4 in detriment of LaNiO3 perovskite conformation, in line with XRD results (Fig. 1). This fact limits the ex-solution of La2O3 from the perovskite, favouring its sintering during the reduction at high temperatures, which ultimately leads to more heterogeneous distribution of the La2O3 phase than in ceria-supported samples (Fig. 2).According to desorption temperature or chemical bond strength, basic sites can be classified into weak (T\u00a0<\u00a0150\u00a0\u00b0C), medium (T\u00a0=\u00a0200\u2013550\u00a0\u00b0C) and strong (T\u00a0>\u00a0550\u00a0\u00b0C). Based on this distribution, the concentration of the different species was determined after the deconvolution and integration of the different peaks identified in the desorption profile (Table 3\n). Comparing supported samples, the lower concentration of the weak basic sites corresponds to the reduced 30% LaNiO3/CeO2 sample. As observed in Table 1, this fact is ascribed to its lower specific surface area, which leads to almost total coverage of the ceria surface. In contrast, this sample shows significantly higher medium and strong basic sites concentration than alumina-supported samples (30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3). Specifically, the concentration of medium and strong basic sites ranges from 120.6 and 87.1\u00a0\u00b5mol CO2 g\u22121 for the 30% LNO/CeO2 sample to 29.0 and 58.5\u00a0\u00b5mol CO2 g\u22121 for the 30% LNO/Al2O3 sample, respectively. As suggested in our previous work [39], the controlled reduction of 30% LaNiO3/CeO2 sample leads to the formation of additional medium strength basic sites (NiO-CeO2 interface) for CO2 adsorption with respect to alumina-supported samples and bulk perovskite. Furthermore, the high interaction of Ni with the support leads to the presence of NiAl2O4 phase (Fig. 3), which limits the LaNiO3 perovskite conformation, increases the proportion of impurities (La2O3 and NiO) for fresh samples (Fig. 1a), and needs a higher reduction temperature (800\u00a0\u00b0C vs. 550\u00a0\u00b0C). Ultimately, these aspects limit the ex-solution of highly dispersed La2O3 basic sites from LaNiO3 host, leading to a significant decrease of the concentration of medium strength basic sites.In order to gain insight on this issue, Table 3 shows the ratio of desorbed moles of CO2 per mol of La2O3. Note that for the accurate determination of the amount desorbed from the La2O3 adsorbent, the moles of CO2 desorbed from the corresponding support were subtracted. According to decomposition reaction of lanthanum carbonate (La2O2CO3\u00a0\u21c4\u00a0La2O3\u00a0+\u00a0CO2) 1\u00a0mol of CO2 should be desorbed per mol of La2O3, if this compound was completely carbonated during saturation step. However, this ratio is below 1 for all samples. Among them, the lowest value (0.08) corresponds to LaNiO3 perovskite, whereas supported perovskites show values more than twice of that of bulk counterpart. Although ceria supported sample show significantly lower specific surface area (Table 1), it presents the highest value (0.28). This trend confirms the higher accessibility of La2O3 sites and the presence of additional CO2 adsorption sites in the NiO-CeO2 interface (180\u2013360\u00a0\u00b0C). Finally, this fact leads to a significant increase of the surface density of medium basicity species, from 0.39\u00a0\u03bcmol CO2 m\u22122 for alumina-supported samples to 3.26\u00a0\u03bcmol CO2 m\u22122 for ceria-supported one. Thus, these results confirm that supporting LaNiO3 perovskite over ceria support increases the accessibility and the concentration of CO2 adsorption sites with respect to alumina-supported samples.Aiming to introduce the basic principles of the operation, Fig. 6\n displays the evolution with time of the outlet concentration of CO2, CH4, CO and H2O during an entire CO2 adsorption and hydrogenation cycle at 400\u00a0\u00b0C. Although these results correspond to the DFM derived from 30% LaNiO3/CeO2 formulation, the global reaction evolution is similar for alumina-supported samples and bulk perovskite.During the adsorption cycle (1\u00a0min) a gas stream composed of 1.4% CO2/Ar is fed. To estimate the amount of CO2 adsorbed on the catalyst, the CO2 concentration profile when the reactor is bypassed is also included. As can be observed, CO2 concentration is almost negligible at the beginning of the adsorption period; in fact, no CO2 signal at the reactor outlet is detected during the first 35\u00a0s. Following, it increases rapidly achieving almost the inlet concentration at the end of the storage period. This trend reveals the progressive CO2 adsorption on storage sites, mainly La2O3 phase [47] and, in minor extent, on NiO\u2013CeO2 interface, up to their total saturation through the following reaction:\n\n(12)\n\n\n\nLa\n2\n\n\nO\n3\n\n\n\n\n+ CO\n\n2\n\n\n\u21c4\n\n\n\nLa\n\n2\n\n\nO\n2\n\n\nCO\n3\n\n\n\n\n\nFew seconds delayed, an increasing H2O signal is detected at the reactor outlet. The identification of this compound during the adsorption period reveals that CO2 is progressively displacing pre-adsorbed H2O due to its competitive adsorption on La2O3 storage sites through Eq. (13). However, it can be concluded that the CO2 adsorption preferentially occurs onto free La2O3 sites, since H2O is detected quite delayed with respect to CO2 identification. Once La2O3 adsorption sites are completely carbonated (Eq. (12)), the storage of CO2 is transferred to La(OH)3 sites (Eq. (13)).\n\n(13)\n\n\n2La\n\n\n\nOH\n\n\n3\n\n\n\n\n+ CO\n\n2\n\n\n\u21c4\n\n\n\nLa\n\n2\n\n\nO\n2\n\n\nCO\n3\n\n\n\n\n+ 3H\n\n2\n\nO\n\n\n\n\nNote that the desorption of a small fraction of H2O stored on ceria or alumina supports in form of hydroxyls cannot be ruled out, which can conform bicarbonates during CO2 adsorption period. However, it is well-known that their stability is limited at working temperatures, which makes this adsorption route minority with respect to that expressed by Eqs. (12\u201313), especially for the ceria-supported sample [38].From these data, the amount of CO2 adsorbed onto the catalyst is calculated by Eq. (2) (Table 4\n). In order to assess the stable behavior of the DFM, the corresponding values to 3 consecutive cycles is included, which results in values between 86.4 and 90.7\u00a0\u00b5mol CO2 g\u22121. Furthermore, an almost negligible CO peak is observed during the adsorption period, which value is determined by Eq. (4) and summarized in Table 4 (around 9\u00a0\u00b5mol\u00a0g\u22121), which is related to the incomplete hydrogenation of adsorbed CO2 with H2 chemisorbed on the Ni0 sites during the previous hydrogenation period following the reverse water gas shift reaction (RWGS, Eq. (14)). Alternatively, other authors related CO formation to the progressive decomposition of adsorbed formate species [48].\n\n(14)\n\n\n\nH\n2\n\n\n\n\n+ CO\n\n2\n\n\n\u21c4\n\n\n\nCO + H\n\n2\n\nO\n\n\n\n\nOnce the adsorption period is completed, the CO2 is removed from the feed stream and a constant Ar flow rate is fed during 2\u00a0min, in order to purge the catalyst as well as the reaction system. As a result, CO2 and H2O signals progressively decrease practically to zero during this period.Then, the hydrogenation period (2\u00a0min) begins with the admission of a gas stream composed of 10% H2/Ar. Immediately after the injection of 10% H2/Ar mixture, a sudden CH4 production is observed with a long tail extended during the rest of the period. Besides, H2O formation is detected around 10\u00a0s delayed from CH4 detection. This process can be described by the following reaction scheme:\n\n(15)\n\n\nStep\n\n1\n:\n\nLa\n2\n\n\nO\n2\n\n\nCO\n3\n\n\u21c4\n\n\n\nLa\n\n2\n\n\nO\n3\n\n\n\n\n+ CO\n\n2\n\n\n\n\n\n\n\n(1)\n\n\nStep\n\n2\n:\n\n\n4H\n2\n\n\n\n\n+ CO\n\n2\n\n\u21c4\n\n\n\nCH\n\n4\n\n\n\n\n+ 2H\n\n2\n\nO\n\n\n\n\n\n\n(16)\n\n\nStep\n\n3\na\n:\n\nLa\n2\n\n\nO\n3\n\n\n\n\n+ 3H\n\n2\n\n\nO\n\n\n\u21c4\n\n\n2La\n\n\n\n\nOH\n\n\n3\n\n\n\n\n\n\n\n(17)\n\n\nStep\n\n3\nb\n:\n\nCe\n2\n\n\nO\n3\n\n\n\n\n+ 3H\n\n2\n\n\nO\n\n\n\u21c4\n\n\n2Ce\n\n\n\n\nOH\n\n\n3\n\n\n\n\n\nFirstly, lanthanum oxide carbonate is decomposed to form gaseous CO2 (Eq. (15)). Then, the CO2 released reacts with hydrogen to form methane and water following Sabatier reaction (Eq. (1)). Taking into account the stoichiometry of Eq. (1), 2\u00a0mol of H2O should be detected per mol of CH4; nevertheless, the experimental ratio during hydrogenation period ranges between 1.38 and 1.40, which reveals that part of H2O is stored on the surface La2O3 sites (Eq. (16)) or ceria support (Eq. (17)). As we already reported in previous work for conventional Ni/CeO2 catalysts [49], the water adsorption on ceria sites is limited due to its high oxygen mobility, which favours water desorption during the hydrogenation period. Indeed, the H2O/CH4 ratio is significantly higher than that observed for conventional Ru-CaO/Al2O3 and Ru-Na2CO3/Al2O3 DFMs (H2O/CH4\u00a0<\u00a01.14) [13]. Finally, a small fraction of CO (around 1\u00a0\u00b5mol\u00a0g\u22121) is also detected during the hydrogenation period due to RWGS reaction (Eq. (14)).If the entire CO2 adsorption and hydrogenation cycle is considered, a H2O/CH4 ratio ranging between 2.00 and 2.03 is obtained, that is close to the stoichiometry value (H2O/CH4\u00a0=\u00a02) defined by Sabatier reaction (Eq. (1)). With the aim of giving more reliability to the results obtained, the carbon balance was also determined (Eq. (7)). As can be observed in Table 1, the amount of CO2 stored during the adsorption period is around 88\u00a0\u00b5mol\u00a0g\u22121, whereas around 80\u00a0\u00b5mol\u00a0g\u22121 of CH4 and 8\u00a0\u00b5mol\u00a0g\u22121 of CO are produced during CO2 hydrogenation period. Thus, carbon balance closed within\u00a0\u00b1\u00a05% since CH4 and CO are the only products detected by FTIR during the reaction.Catalytic activities of LaNiO3-derived DFMs are evaluated by analyzing the evolution of CH4 and CO production per cycle with reaction temperature (Fig. 7\n). These parameters were estimated applying Eqs. (2\u20133) for the data obtained from similar CO2 adsorption and hydrogenation experiments to that reported in Fig. 6. Aiming to mimic an effluent gas from a combustion process, the CO2 concentration during storage period was increased from 1.4 to 10% in these experiments, in which the carbon balance closed with an error below 5%.As can be observed in Fig. 7a, the evolution of CH4 production with reaction temperature is influenced by the type of perovskite-based formulation used as precursor of the corresponding DFM in each experiment. As expected, methane production increases up to 440\u00a0\u00b0C for DMFs obtained after the reduction of bulk LaNiO3 and 30% LaNiO3/CeO2 formulations. Above this temperature, CO2 conversion slightly decreases due to a destabilization of adsorbed carbonates. In contrast, DFMs obtained from alumina-supported perovskites (30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3) achieve their maximum CH4 production at 280\u00a0\u00b0C. Then, a progressive decrease in the amount of CH4 produced is observed at increasing temperatures. As previously observed, weak strength basic sites are predominant for alumina-supported DFMs (Table 3). Thus, the maximum CH4 production observed at 280\u00a0\u00b0C is related to a more efficient CO2 adsorption on weak strength basic sites, main adsorption sites at this temperature range. On the contrary, as the reaction temperature increases, the adsorbed CO2 on weak basic sites become less stable, limiting their hydrogenation for alumina-supported samples. Meanwhile, the presence of higher strength basic sites for non\u2013 and ceria-supported samples favours decomposition of a major quantity of the adsorbed CO2 species to be hydrogenated to CH4 at higher temperatures. Thus, Ni-La2O3 interface higher accessibility can be considered as a key parameter to maximize CO2 adsorption and in-situ hydrogenation at this temperature range.Regarding to CO formation (Fig. 7b), all samples show an increasing CO production with reaction temperature. This trend is ascribed to the promotion of the RWGS reaction (Eq. (17)) during the CO2 hydrogenation step. In any case, the CO production is below 31\u00a0\u00b5mol\u00a0g\u22121 for all samples, which remarks the high selectivity towards methane of here developed materials.Among different samples, the DFM obtained after the reduction of the LaNiO3 formulation exhibits the highest CH4 production (117\u00a0\u00b5mol\u00a0g\u22121), in line with the higher density of medium basic sites identified in Table 3. However, the DFM derived from 30% LaNiO3/CeO2 precursor maintains the highest CH4 production, if the whole temperature range is considered. Furthermore, this sample shows a CO production 3 times lower (8 vs. 31\u00a0\u00b5mol\u00a0g\u22121) than the DFM obtained from bulk perovskite, which is ascribed to the higher strength of CO2 adsorbed species [50,51].To better understand the differences in the catalytic behaviour, Fig. 8\n plots CH4 concentration profiles during complete CO2 adsorption and hydrogenation cycles at 280, 400 and 520\u00a0\u00b0C for DFMs obtained from 30% LaNiO3/CeO2, 30% LaNiO3/La-Al2O3 and LaNiO3 precursors. Profiles corresponding to the 30% LaNiO3/Al2O3 precursor have not been included since they are similar to those of La-Al2O3-supported sample. In general, the evolution of CH4 is significantly affected by DFM composition, especially at intermediates-high temperatures. The maximum CH4 production is observed at initial times for the alumina-supported sample, whereas this process is delayed and takes place more progressively for the DFM obtained from bulk perovskite. On the other hand, the ceria-supported sample shows an intermediate CH4 production profile. As previously observed in Table 1, the specific surface area was significantly higher for supported samples with respect to that observed for the DFM derived from bulk perovskite, which leads to the exsolution of Ni NPs with significantly lower average particle size than bulk counterpart (31.7\u00a0nm), especially for ceria-supported sample. Furthermore, this sample shows the higher proportion of medium basic sites with respect to strong basic sites. Taking into account that the close contact between storage component and the Ni0 NPs is regarded as the key factor to efficiently transfer of dissociated H to desorb, and subsequently to hydrogenate, adsorbed CO2, these facts explain the wider temperature window of the DFM derived from the 30% LaNiO3/CeO2 formulation. In contrast, the stability of adsorbed species is limited for alumina-supported sample, favouring only the CH4 production at the beginning of the hydrogenation period and low temperatures.To sum up, the 30% LaNiO3/CeO2 emerges as the optimal catalytic precursor, resulting in a dual function material with high efficiency to adsorb CO2 and in-situ hydrogenate it to methane. This fact is ascribed to an proper balance of different basic sites concentration, where CO2 adsorption takes places (Ni-CeO2 interface as well as highly dispersed and bulk-like La2O3), and higher accessibility of active sites for H2 activation and CO2 methanation (Ni0 NPs). Ultimately, this fact also favours a higher selectivity towards methane. Furthermore, this sample is able to produce a high fraction of methane at initial period of the hydrogenation cycle. As suggested in our previous works [52,53], the duration of the hydrogenation period should be enough to ensure high CH4 production but not too long to limit hydrogen conversion. Hence, optimal hydrogenation time will provide a best balance between more efficient use of reductant agent and CH4 production. As a result, the faster kinetics discovered with the DFM derived from 30% LaNiO3/CeO2 catalytic precursor, makes it a first-class alternative as promotes the joint optimization of H2 conversion and CH4 production. The last aspect to consider is that the Ni content is around 70% lower with respect to bulk LaNiO3, which reveals a superior intrinsic activity of this sample.In order to have a more realistic view of the relevance of the reported results, the CH4 and CO productions obtained with this DFM where compared to those obtained with 15% Ni-15% CaO/Al2O3 model DFM (Figure S4) [13], showing comparable CO2 adsorption and in-situ hydrogenation to CH4. These results remark that here developed DFMs can be considered as promising novel materials for CO2 methanation technology. The still limitation of higher CH4 production at higher temperatures is actually under study in our labs with the use of other alkaline or earth-alkaline adsorbents, such as Ca, Ba, Na and K.The real-world applicability of the DFM obtained after the controlled reduction of 30% LaNiO3/CeO2 precursor was more deeply analyzed by subjecting this DFM to long-term CO2 adsorption/hydrogenation experiments under hard operational conditions. This study was completed by evaluating the influence of the presence of O2 during adsorption period on its CO2 adsorption and hydrogenation efficiency.\nFig. 9\n shows the evolution of CH4 and CO productions with the number of CO2 adsorption/hydrogenation cycles for 30% LaNiO3/CeO2-derived sample at 520\u00a0\u00b0C. As can be observed, CH4 and CO productions as well as selectivity towards methane remain almost stable irrespective of time elapsed. Specifically, the CH4 production slightly decreases from 80\u00a0\u00b5mol\u00a0g\u22121 to 78.4\u00a0\u00b5mol\u00a0g\u22121, whereas CO production keeps at 9.3\u20139.4\u00a0\u00b5mol\u00a0g\u22121. This catalytic behaviour reveals the high stability of the developed DFM towards CO2 adsorption and hydrogenation in consecutive cycles. The close contact between Ni0 NPs, La2O3 and CeO2 phases (Fig. 2), formed after the controlled reduction of 30% LaNiO3/CeO2, prevents the thermal agglomeration of Ni NPs during calcination and CO2 methanation reaction processes, in line with the observed by Wang et al. [35] for Ni-La2O3/SBA-15 catalyst.The developed DFM should also selectively capture CO2 from O2-containing flue gas at relatively high temperatures and then, hydrogenate the adsorbed species to methane with H2. With the aim of evaluating the influence of the presence of O2 during storage period a 10% of O2 is jointly fed with a 10% of CO2 during adsorption cycles. Fig. 10\n plots the evolution of CH4 and CO productions with the number of CO2 adsorption/hydrogenation cycle at 400\u00a0\u00b0C for 30% LaNiO3/CeO2-derived sample. Note that the cycles 1\u20132 and 8\u20139 were carried out in the absence of O2 in the feed stream, whereas this compound was fed in cycles 3\u20137.Comparing cycles 1\u20132 with cycles 3\u20137, a negative effect on CH4 production can be detected for the experiments in the presence of O2 during the adsorption period. Indeed, the CH4 yield immediately decreases from 99 to 47\u00a0\u03bcmol\u00a0g\u22121 from the 2nd to 3rd cycle, whereas no significant changes are observed for cycles 4\u20137. Zheng et al. [54] justified the loss of activity for O2-containing experiments by the oxidation of the active metallic phase during the adsorption step. In agreement with their results, a small CO2 signal and a significant decrease in methane production is observed at the beginning of the hydrogenation period for the oxygen-containing experiment with respect to oxygen-free experiment (Figure S5). This fact reveals that some carbonates, adsorbed during the storage period, are released without being hydrogenated due to the absence of enough Ni0 active sites to reduce them towards CH4. In any case, the decrease in methane production for O2-containing experiments is significantly lower to that observed for 10% Ni-6.1% NaO/Al2O3, where no methane formation was observed when the sample was exposed to O2 and H2O during the CO2 capture step [19]. On the other hand, it is worth to mention that, in contrast to that observed in our previous work for conventional 10% Ni\u201310% Na2CO3/Al2O3\n[11], CO production remains invariable after inclusion of O2 in the feed stream. This trend discards the promotion of RWGS reaction (Eq. (14)) due to a partial oxidation of Ni0 to NiO.During the next cycles (i.e. from the 8th to the 9th), CO2 adsorption/hydrogenation cycles were again carried in an oxygen-free environment. Remarkably, the CH4 production is recovered immediately after the oxygen is removed from the feed stream. Note that the 8th and 9th cycles show similar CH4 and CO productions than 1st and 2nd cycles. Thus, these results reveal that the here discovered DFM has a high ability to restore activity once O2 is not fed during CO2 adsorption, which it is one of the main limitations of the conventional Ni-based formulations [19,55,56]. In agreement with the H2-TPR results, the high reducibility of different Ni species implies that Ni can be easily reduced back during the hydrogenation step at low temperature. Therefore, 30% LaNiO3/CeO2-derived DFM can be considered a superior candidate for real conditions process at intermediate-high temperatures.In summary, the confinement of Ni NPs on La2O3 or La-Ce-O interfaces prevents them from thermal agglomeration and favours their redox properties. Hence, ceria-supported LaNiO3 perovskites can be considered as an efficient precursor of highly stable and versatile dual function materials for the cyclic CO2 adsorption and hydrogenation to methane technology.This work is focused on the analysis of the viability of LaNiO3-based formulations as precursor of active, stable and versatile dual function materials for CO2 adsorption and hydrogenation into CH4. In particular, the following perovskite-based formulations are prepared by combining citric acid and wetness impregnation: LaNiO3, 30% LaNiO3/CeO2, 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3. The prepared formulations are widely characterized before and after controlled reduction process. XRD experiments reveal the presence of LaNiO3 as well as of impurities in form of La2O2CO3, NiO and La2NiO4 for all calcined samples. The concentration of these impurities increases for 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3 samples with respect to LaNiO3, 30% LaNiO3/CeO2 ones, whereas the specific surface area follows the opposite trend. Among different samples that supported on ceria oxide (30% LaNiO3/CeO2) shows the higher reducibility in H2-TPR experiments, whereas the redox properties are limited for 30% LaNiO3/Al2O3 and 30% LaNiO3/La-Al2O3 due to the higher perovskite-support interaction, which favours the NiAl2O4 formation instead of LaNiO3 conformation.After the controlled reducing process, Ni is exsolved from perovskite host in developed formulations, leading to the formation of a mix between Ni0 nanoparticles and La2O3 phase. However, the interaction of Ni nanoparticles with the La2O3 phase and support is controlled by the characteristic of the precursor used. The best compromise between specific surface area and LaNiO3 stability observed for ceria-supported sample (30% LaNiO3/CeO2) promotes the formation of smaller Ni0 nanoparticles and a more homogeneous La2O3 distribution. As a result, the obtained DFM presents a stronger interaction between Ni0 NPs and the La2O3 and CeO2 phases and a more modulated basicity. Finally, this fact favours the transfer of dissociated H to hydrogenate near-adsorbed CO2 at the studied operating temperature. As a result, the material obtained after the reduction of the 30% LaNiO3/CeO2 formulation exhibits the highest CH4 production, if the whole temperature range is considered. Specifically, its maximum CH4 production per cycle is 104\u00a0\u00b5mol\u00a0g\u22121 (440\u00a0\u00b0C) and its selectivity towards CH4 formation is above 90% in the whole temperature range. Furthermore, this DFM also emerges as promising approach to promote the joint optimization of H2 conversion and CH4 production, since its present quite fast kinetics during CO2 hydrogenation to methane.The 30% LaNiO3/CeO2-derived DFM also shows promising properties for the real-world applicability. On the one hand, it demonstrates high stability during long-terms experiments under hard reactions conditions, even above than other conventional Ni-based DFMs. On the other hand, although, the presence of O2 during the CO2 capture step has a detrimental effect on CH4 production, the decrease is lower than that reported for other conventional Ni-based catalysts. Indeed, the activity recovery capacity is higher when the system comes back to an oxygen-free environment due to the enhanced redox properties of this novel DFM. Thus, ceria-supported LaNiO3 perovskites emerge as promising precursors of highly active, versatile and stable novel dual function materials for CO2 adsorption and hydrogenation to methane under wide variety of operational conditions.\nJon A. Onrubia-Calvo: Conceptualization, Methodology, Validation, Writing \u2013 original draft. Alejandro Bermejo-L\u00f3pez: Methodology, Investigation. Sonia P\u00e9rez-V\u00e1zquez: Investigation. Be\u00f1at Pereda-Ayo: Conceptualization, Methodology, Visualization, Writing \u2013 review & editing. Jos\u00e9 A. Gonz\u00e1lez-Marcos: Methodology, Data curation, Supervision, Funding acquisition. Juan R. Gonz\u00e1lez-Velasco: Conceptualization, Supervision, Funding acquisition, Project administration.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Support for this study was provided by the Spanish Ministry of Science and Innovation (Project PID2019-105960RB-C21) and the Basque Government (Project IT1297-19). One of the authors (JAOC) acknowledges the Post-doctoral research grant (DOCREC20/49) provided by the University of the Basque Country (UPV/EHU).Supporting information includes detailed information on characterization of fresh and used samples, i.e. Enlargement of XRD results, N2 adsorption-desorption isotherms and CH4 and CO2 mass spectroscopy signals during H2-TPR experiments. A comparison of CH4 and CO productions with respect to conventional 10% Ni-15% CaO/Al2O3 during CO2 adsorption and hydrogenation cycles is also included as reference. Supplementary data to this article can be found online at https://doi.org/10.1016/j.fuel.2022.123842.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  The valorisation of CO2 through its capture and in-situ hydrogenation to methane, using dual function materials (DFMs), emerges as promising alternative to reduce CO2 emissions to atmosphere and the global cost of current CO2 Capture and Utilization (CCU) technology. This work investigates the viability of LaNiO3-derived formulations as precursors of DFMs for CO2 capture and in-situ conversion to CH4. For this purpose, a set of DFMs obtained from 30% LaNiO3/CeO2, 30% LaNiO3/Al2O3, 30% LaNiO3/La-Al2O3 and LaNiO3 precursors were synthesized and systematically characterized before and after a controlled reduction process. Results of XRD analysis, STEM-EDX images, H2-TPR and CO2-TPD experiments reveal that the DFM obtained after reduction of 30% LaNiO3/CeO2 formulation shows the smallest Ni0 particle size (7\u00a0nm) and the highest medium-strong basic sites concentration. In fact, this DFM widens operation window with methane production ranging between 80 and 103\u00a0\u00b5mol\u00a0g\u22121 and maintains a selectivity towards methane above 90% in the range of 280\u2013520\u00a0\u00b0C. The best catalytic behaviour is related to a better contact between the different nature basic sites and the homogenously distributed Ni0 sites, which favours a fast spill-over of dissociated H to near CO2 adsorption sites. The applicability of this formulation is further evidenced by a highly stable CH4 production during long-term experiments and a promoted Ni0/NiO reversibility in the absence/presence of O2 during the CO2 adsorption period, which allows a fast and complete recovery of CH4 production in absence of O2. These aspects favour a versatile application of the 30% LaNiO3/CeO2-based DFM formulation to convert CO2 outlet streams from combustion flue gases of different nature.\n               "}
{"full_text": "Data will be made available on request.Lower Olefins (C2-C4) are indispensable intermediates in the chemical industry. [1] Approximately 20% of the yearly production revenue of the chemical industry is related to the olefin-based polymer industry. [2] Traditionally, olefins are produced via steam cracking of fossil feedstocks, like crude oil. [1] Due to the growing social and political awareness regarding sustainable production, non-fossil based production routes and the shift towards a circular economy gain great interest. [3,4] While part of the olefine based polymers can be recycled mechanically [5] a large portion still needs to be recycled by other means. [6] Direct chemical recycling to the olefin monomers is more or less impossible as selective breaking of the CC bond and dehydrogenation is difficult. [7] A possible way to close the carbon cycle is i) gasification or incineration of the polymer waste and ii) synthesis of ethylene and propylene through hydrogenation of the resulting CO/CO2 or solely CO2 streams. [8] While the incineration of plastic waste is commercialized [9] and recently the gasification of these waste is studied intensively [10,11], the synthesis of Olefines from the CO/CO2 streams remains the more critical challenge in this carbon cycle. Direct electrochemical reduction of CO/CO2 to ethylene is an attractive process but still from a technical point of view not mature enough. Methanol synthesis and subsequent Methanol-to-Olefin (MTO) conversion is a possible route, while for the MTO process no large-scale production plant with a capacity of >1 MTPA is existing. The Fischer-Tropsch-Synthsis (FTS) in contrary is commercialized on big scales and it is known that the product selectivity for CO as feedstock can be tuned towards high olefine yields and in a single process. [1,12] Fischer-Tropsch-to-Olefins (FTO) can thus be an important, scalable and readily available process to close the carbon cycle for polypropylene and polyethylene recycling. CO2 as feedstock, in combination with green hydrogen, would be especially interesting, as it cannot only stem from plastic waste incineration, but also other sustainable resources as well as from direct air capture. [3]As CO2 is not converted into Fischer-Tropsch-products directly, the reverse water-gas shift reaction (RWGS) to CO has to be carried out prior. The combination of this endothermic equilibrium reaction and the highly exothermic Fischer-Tropsch synthesis (FTS) can be realized either in two separate or in one combined reactor unit. Due to their opposite reaction enthalpies and process intensification aspects, the direct heat integration is one main advantage of coupling both reactions in one reactor. Additionally, as CO is converted to FTS-products in a series of irreversible consecutive reactions, a high CO2 conversion by RWGS can be obtained even at low temperatures. However, the coupling of RWGS and FTS demands a catalyst system exhibiting activity for both the shift-reaction and the FTS, which is met by using Fe as an active metal. [4]In this context, especially carbon supported iron catalysts are of increasingly growing interest. [13\u201315] Compared to the traditionally used oxidic support materials, such as SiO2 and Al2O3, iron species on carbon materials prove to be reducible more easily, since there is no strong metal support interaction (SMSI). This facilitates the conversion of Fe species into FT-active carbides, as the carbidisation of elemental iron is facilitated on carbon supports. [16,17] Despite the less pronounced interactions between support and active metal, carbon-supported iron systems, as demonstrated by the use of carbon nanotubes (CNT) as support, showed excellent resistance to sintering. [18,19]Regarding the utilization of carbon supported Fe in a CO2 based FTO process, Oschatz et al. reported sulfur and sodium as promising promotors for the production of olefins using ordered mesoporous carbon (CMK-3) as support route. [20\u201323] They found that the undesired methane formation is inhibited by sulfur doping, while the desired \u03b2-hydride elimination to short chained olefins is preferred at the same time. Addition of sodium supports chain growth and reduces selectivity to methane while inhibiting olefin hydration to paraffins. [24]Most of the existing studies dealing with Fe/C-catalysts in the FTS were carried out using research carbon materials like CNT's or CMK-3. Facing the challenges of an industrial FTS process, these materials reveal disadvantages for commercialization of derived catalysts like expensive and complicated synthesis, as well as handling on the industrial scale of these small sized powdered materials and the resulting severe pressure drop for technical fixed bed reactors. Also, so far, results are mainly presented for some hours time on stream and rarely long term stability studies for several hundred hours. Overall, both hinders an application of Fe/\u2212catalysts in FTO strongly.The following study presents the use of beaded carbon blacks (CBs) as support for iron based CO2-FT catalyst. CB represents a material which is widely available in industrial scale at low price with additional formulation procedures to reduce pressure drop for fixed bed reactors e.g. through beading also being carried out on an industrial scale. Commercial beaded CBs varying in properties (e.g. specific surface areas from 36 m2\u00b7g\u22121 to 380 m2\u00b7g\u22121) were employed for catalyst synthesis. The catalyst performance in direct olefin synthesis for CO2 hydrogenation was assessed and for the most promising catalyst a long-term study proved stability for 170\u00a0h time on stream.Five commercially available beaded Carbon Blacks (Orion Engineered Carbons: Printex G, Printex 35, Printex 60, Printex 85, Printex 90) were used as supports in this study. The preparation strategy was modified according to Oschatz et al. [15] Carbon supported catalysts were prepared by incipient wetness impregnation, using a solution of ammonium\u2011iron-citrate (1.787\u00a0g; 16.5\u201318.5% Fe; Acros Organics), tri\u2011sodium-citrate dihydrate (0.038\u00a0g; 99+ %, Fisher Scientific), and iron sulfate heptahydrate (0.026\u00a0g; 99+ %; Acros Organics) in 6\u00a0mL bidistilled water. Concentrations were chosen so that a loading of 10\u00a0wt% Fe, 0.3\u00a0wt% Na and 0.1\u00a0wt%\u00a0S with respect of the carbon support results. 3\u00a0g of dried carbon support (100\u00a0\u00b0C over night) was impregnated with this solution in six to eight steps depending on the capacity of the support material. Every impregnation step was followed by a two hour drying step at 100\u00a0\u00b0C combined with the homogenization of the sample in an achate mortar. After the last drying step, the raw catalyst was calcinated in a tube furnace for 5\u00a0h at 500\u00a0\u00b0C under constant nitrogen flow at atmospheric pressure. This catalyst is referred to as pristine catalyst.Optical emission spectrometry with inductively coupled plasma (ICP-OES, Optima 2000DV (Perkin Elmer)) was used to determine the loading of Fe, Na and S on the carbon supports. Sample preparation included the oxidation of the carbon support at 500\u00a0\u00b0C in a muffle furnace, followed by aqua regia digestion of the residues. N2-physisorption measurements (Quadrasorp-MP-30 (Quantachrome Instruments)) are carried at \u2212196\u00a0\u00b0C, with samples degassed at 350\u00a0\u00b0C and 0.1\u00a0mbar for 18\u00a0h. The multi-point Brunauer-Emmett-Teller method (MBET) is used to determine the specific surface area. [25] The pore volume is determined by means of density functional theory (DFT) using the adsorption isotherm. Oil adsorption is used to determine the degree of branching of the carbon blacks studied. The determination is made by Orion Engineered Carbons according to ISO 4656 respectively ASTM D 2414 utilizing dibutyl phthalate as adsorbate. [26] TPD-MS was carried out in a STA 409 PC Luxx thermo-balance (NETZSCH) coupled with an online mass spectrometer (Omnistar, Pfeiffer Vacuum GmbH). For this purpose, 150\u00a0mg of the carbon black were initially heated to 80\u00a0\u00b0C at a flow rate of 30 NmL min\u22121 He with a heating rate of 5\u00a0\u00b0C\u00a0min\u22121 and dried for 30\u00a0min. The sample was then heated from 80\u00a0\u00b0C to 1000\u00a0\u00b0C at a heating rate of 5\u00a0\u00b0C\u00a0min\u22121. X-Ray diffraction (XRD) measurements have been performed in a Stadi-P (Stoe GmbH) using a Cu-K\u03b1,1-source. Elemental analysis was carried out in a Vario EL III analyzer (Elementar Analysensysteme GmbH). All samples were measured five times. The determination of the primary particle size was carried out by means of dynamic light scattering by Orion Engineered Carbons. [27] The ISO 1125 respectively ASTM D 1506 standard was used to determine the ash content of the carbon blacks used by Orion Engineered Carbons. [28] Scanning transmission electron microscopy (STEM) was carried out in a JEM 2100F (JEOL) with an accelerating voltage of 200\u00a0kV and a spot diameter of 0.7\u00a0nm. STEM images were used for the determination of the iron particle size distribution, by analyzing a minimum of 50 particles per sample. Scanning electron microscopy (SEM) was carried out using a Philips XL30 FEG electron microscope with an acceleration voltage of 10\u201315\u00a0kV. X-ray diffraction (XRD) was performed using a Rigaku Miniflex exquipped with a D/tex Ultra detector (CuK\u03b1, 40\u00a0kV, 0.03\u00a0mm Ni-filter). Temperature programmed reduction (TPR), CO-Chemisorption as well as CO- and CO2-temperature programmed desorption (TPD) were carried out in a 3Flex (Micromeritics) chemisorption unit. TPR is performed up to 400\u00a0\u00b0C with a heating rate of 5\u00a0\u00b0C\u00a0min\u22121 using 0.5% H2 in Argon. Following an evacuation, static CO-chemisorption is performed at 30\u00a0\u00b0C between 100\u00a0mbar and 800\u00a0mbar. Afterwards CO-TPD is realized between 30\u00a0\u00b0C and 400\u00a0\u00b0C obtaining a heating rate of 5\u00a0\u00b0C\u00a0min\u22121. Next a second TPR is performed and the sample is again evacuated. After the CO2-saturation of the sample using 4% CO2 in He at 30\u00a0\u00b0C, CO2-TPD is performed at a temperature up to 600\u00a0\u00b0C using a heating rate of 5\u00a0\u00b0C\u00a0min\u22121.Catalyst reduction and CO2-FTS activity measurements were carried out in the same u-shaped fixed bed reactor. 0.5\u00a0g of the pristine catalyst were placed into the reactor and fixed with the aid of two glass wool plugs. Catalyst reduction was carried out at 300\u00a0\u00b0C for 5\u00a0h at 30\u00a0bar using pure hydrogen. For FTS, a premixed H2:CO2-mixture (3:1\u00a0mol\u00a0mol\u22121) was used as feed at a mass flow of 0.03\u00a0g\u00a0min\u22121 (GHSV\u00a0=\u00a0\u223c8000\u00a0h\u22121). 325\u00a0\u00b0C were set as standard temperature. Carbon monoxide and carbon dioxide were detected via FT-IR (Bruker alpha), hydrocarbons were analysed in a Shimadzu GC-2014 gas chromatograph equipped with two FID's. Separation of C1-C6-hydrocarbons took place at a Rt-QS-BOND Plot column (Restek) while higher hydrocarbons were separated at a Rtx-column (Restek). Resulting \u03c3i values represent the fraction of the respective substance class i of the total amount of hydrocarbons and are standardized to one. All conversions, selectivities and product distributions shown result from the mean value of three measurements after 30\u00a0h TOS.For the long-term measurement,1.5\u00a0g of the pristine catalyst (Printex 60) and 3.0\u00a0g Printex 60 carbon black as inert dilution were placed into the reactor and fixed with the aid of two glass wool plugs. Catalyst reduction was carried out at 300\u00a0\u00b0C for 5\u00a0h at 30\u00a0bar using hydrogen in helium (1:2\u00a0mol\u00a0mol\u22121). A stoichiometric H2:CO2:CO-mixture (2.5:0.5:0.5\u00a0mol\u00a0mol\u00a0mol) was used for preconditioning of the catalyst for the first 48\u00a0h TOS at 325\u00a0\u00b0C. After preconditioning, the feed was set to a stoichiometric H2:CO2-mixture (3:1\u00a0mol\u00a0mol\u22121) using 80.7 NmL min\u22121 H2 and 3.2\u00a0g\u00a0min\u22121 CO2 (GHSV\u00a0=\u00a0\u223c3200\u00a0h\u22121). Product analytics were performed using an Agilent GC6890N gas chromatograph equipped with one FID and one TCD. Separation of C1-C3-hydrocarbons as well as CO and CO2 took place at a Rt-QS-BOND Plot (Restek) column and a Rt-Alumina (Restek) column while higher hydrocarbons were separated at a Rtx-column (Restek). Resulting \u03c3i values represent the fraction of the respective substance class i of the total amount of hydrocarbons and are standardized to one.Carbon dioxide (X\nCO2) conversion is calculated dividing the difference between the initial quantity (\n\n\nn\n\u0307\n\n\n0\n,\n\nCO\n2\n\n\n\n)\n and the resulting quantity of carbon dioxide (\n\n\nn\n\u0307\n\n\nCO\n2\n\n\n)\n through the initial quantity (eq. 1).\n\n(1)\n\n\nX\nCO2\n\n=\n\n\n\n\nn\n\u0307\n\n\n0\n,\n\nCO\n2\n\n\n\n\u2212\n\n\nn\n\u0307\n\n\nCO\n2\n\n\n\n\n\nn\n\u0307\n\n\n0\n,\n\nCO\n2\n\n\n\n\n\n\n\nSelectivity to carbon monoxide (S\nCO,CO2) build out of carbon dioxide is calculated dividing the resulting quantity of CO (\n\n\nn\n\u0307\n\nCO\n\n)\n through the difference between the initial quantity (\n\n\nn\n\u0307\n\n\n0\n,\n\nCO\n2\n\n\n\n)\n and the resulting quantity of carbon dioxide (\n\n\nn\n\u0307\n\n\nCO\n2\n\n\n)\n (eq. 2).\n\n(2)\n\n\nS\n\nCO\n,\nCO2\n\n\n=\n\n\n\nn\n\u0307\n\nCO\n\n\n\n\n\nn\n\u0307\n\n\n0\n,\n\nCO\n2\n\n\n\n\u2212\n\n\nn\n\u0307\n\n\nCO\n2\n\n\n\n\n\n\n\n\nSelectivity to hydrocarbons (S\nHC,CO2) build out of carbon monoxide is calculated dividing the product of the resulting quantity of HC (\n\n\nn\n\u0307\n\nHC\n\n)\n and its number of carbon atoms (N\nC, HC) through the difference between the initial quantity (\n\n\nn\n\u0307\n\n\n0\n,\n\nCO\n2\n\n\n\n)\n and the resulting quantity of carbon dioxide (\n\n\nn\n\u0307\n\n\nCO\n2\n\n\n)\n (eq. 3).\n\n(3)\n\n\nS\n\nHC\n,\nCO2\n\n\n=\n\n\n\n\nn\n\u0307\n\nHC\n\n\u2219\n\nN\n\nC\n,\nHC\n\n\n\n\n\n\nn\n\u0307\n\n\n0\n,\n\nCO\n2\n\n\n\n\u2212\n\n\nn\n\u0307\n\n\nCO\n2\n\n\n\n\n\n\n\nYield of product P (Y\nP,CO2) is calculated multiplying the selectivity of product P with the carbon dioxide conversion (eq. 4).\n\n(4)\n\n\nY\n\nP\n,\nCO2\n\n\n=\n\nS\n\nP\n,\n\nCO\n2\n\n\n\n\u2219\n\nX\n\nCO\n2\n\n\n\n\n\nUsing the selectivities for hydrocarbons (S\nHC,r), the product fractions \u03c3HC,r are determined. This calculation (eq. 5) is based on the normalization of the product spectrum and serves the separate consideration of the resulting FT product distribution neglecting the WGS equilibrium.\n\n(5)\n\n\n\u03c3\n\nHC\n,\nr\n\n\n=\n\n\nS\n\nHC\n,\nr\n\n\n\n\n\u2211\nHC\n\n\nS\n\nHC\n,\nr\n\n\n\n\n\n\n\nSince the focus of this work is on the production of short-chain olefins, the olefin fraction in the C2-C6 fraction of the product spectrum is introduced as a further measure (eq. 6).\n\n(6)\n\n\nC\n2\n\n\u2212\n\nC\n6\n\n\u2212\nolefin fraction\n\n=\n\u0302\n\n\n\nx\n\n\nC\n2\n\n\u2212\n\nC\n6\n\n,\nolefins\n\n\n\nx\n\n\nC\n2\n\n\u2212\n\nC\n6\n\n,\ntotal\n\n\n\n\n\n\nUsing the logarithmic expression shown in eq. 7 and via linear regression, the chain growth probability \u03b1 is determined graphically. For this purpose, log x\nHC is plotted against the carbon number N\nC,HC.\n\n(7)\n\nlog\n\nx\nHC\n\n=\n\nN\n\nC\n,\nHC\n\n\n\u00b7\nlog\n\u03b1\n+\nlog\n\n\n\n1\n\u2212\n\u03b1\n\n\n\u03b1\n\n\n\n\nCarbon blacks are known to be hierarchically build from primary particles, forming stable aggregates and loose agglomerates. The employed carbon blacks exhibit a wide variety of properties, owing to variations in their microstructure. As shown in Fig. 1A, all carbon blacks exhibit different primary particle sizes. [27] Starting with Printex G, the primary particle size decreases from 51\u00a0nm to 14\u00a0nm for Printex 90. The aggregates built from these primary particles were characterized by oil adsorption carried out with dibutyl phthalate (DBP), with the resulting DBP number representing a measure for the degree of branching (Fig. 1A). [27] With 48 mLDBP 100\u00a0g\u22121 and 42 mLDBP 100\u00a0g\u22121, Printex 35 and 85 have by far the lowest DBP number, and exhibit therefore the most linear structure. Printex 60 with 115 mLDBP 100\u00a0g\u22121 is the carbon black with the highest degree of branching. Due to the different primary particle sizes and structuring of the aggregates, varying textural properties result. The N2-physisorption isotherms (Fig. 1B) of Printex G, 35 and 60 represent isotherms of type III according to IUPAC nomenclature. They include almost no uptake in the relative pressure range of 0 to 0.1, and furthermore lack desorption hysteresis indicating the absence of meso- and microporosity. Since the majority of nitrogen uptake occurs at relative pressures above 0.8, and no plateau could be observed, macroporosity is dominant in these materials. In contrast, the isotherms of Printex 85 and Printex 90 show a desorption hysteresis (type H3) and a low nitrogen uptake in the relative pressure range of 0\u20130.1 can be observed, resulting in a classification as a mixture of Type II and Type IV. [25] The specific surface area (SSA, determined by BET) increases with decreasing primary particle size from 36 m2\u00a0g\u22121 (Printex G) to 380 m2\u00a0g\u22121 (Printex 90). With high probability, the SSA results from the geometric surface of the particles, with particle interstices as the origin of porosity. In consequence, the degree of branching plays an important role, as is illustrated by Printex 85 and Printex 90, which both have a similar primary particle size, but a clearly different structuring and thus a very different SSA.These findings are clearly supported by scanning electron microscopy (SEM) and transmission electron microscopy (TEM, Fig. 2\n). SEM imaging of Printex G and Printex 90 illustrates the different morphologies that result from the difference in primary particle size. Printex G, which exhibits a large primary particle size of 54\u00a0nm, shows clear macroporosity at the agglomerate level, while Printex 90, which has an average primary particle diameter of 14\u00a0nm, appears to largely lack porosity in the \u03bcm range (Fig. 2A, B). TEM imaging shows that the aggregates of Printex G and Printex 90 exhibit clear differences in terms of particle interstices that reach up to several tens of nanometers for Printex G, while they are significantly smaller in case of Printex 90 (Fig. 2C, D).In addition to their structure, the composition of the carbon black supports is of interest, which can be described by the carbon content, the amount of ash and amount of volatile components (Fig. S1). Carbon content of the different carbon blacks ranges from 97\u00a0wt% for Printex 85 to up to 99.7\u00a0wt% for Printex 35. In addition to the lowest carbon content, Printex 85 reaches the highest values in case of volatiles (1.2\u00a0wt%) and ash (0.8\u00a0wt%). Volatile components are often present in the form of oxygen surface groups on the carbon surface. In order to be able to identify carbon surface oxides, temperature-programmed desorption measurements with coupled mass spectrometry (TPD-MS) were carried out (Fig. S2). During TPD to 1000\u00a0\u00b0C, Printex 35 shows no detectable mass loss, while the largest loss (1.7\u00a0wt%) is observed in case of Printex 85. Regarding the detected CO, CO2 and H2O emission profiles it must be emphasised that the interpretation must be done with caution due to the very low amount of surface groups on the soot surface. In addition, due to the wide temperature ranges of the desorption of different surface species, it is generally difficult to make assignments. For Printex 85, three H2O desorption maxima can be observed at 180\u00a0\u00b0C, 300\u00a0\u00b0C and 710\u00a0\u00b0C. The detection of water might be assigned to condensation reactions, for example of two hydroxyl groups to an ether. The CO2 emission profile shows desorption maxima in comparable temperature ranges. CO2 emission in the range between 200\u00a0\u00b0C and 400\u00a0\u00b0C can be assigned to the decarboxylation of carboxylic acids and anhydrides, while high temperature CO2 evolution (>500\u00a0\u00b0C) indicates the decomposition of lactones. CO emission is detected above 600\u00a0\u00b0C. This desorption range is characteristic for the presence of hydroxyl groups, ethers, ketones or aldehyde groups. [29\u201331]The carbon-supported iron catalysts were prepared by incipient wetness impregnation (IWI). As described in detail in the experimental section, the target loading was 10\u00a0wt% iron (Fe), 0.3\u00a0wt% sodium (Na) and 0.1\u00a0wt% sulfur (S). Fe, Na and S loading was determined by means of optical emission spectroscopy with inductively coupled plasma (ICP-OES) (Fig. 3A). The mean loading of the samples was found to be 9.75\u00a0wt% (Fe), 0.66\u00a0wt% (Na) and 0.16\u00a0wt% (S). It should be emphasised that there are clear outliers in the form of the P_85-Fe/Na/S (sodium 0.93\u00a0wt%) and P_90-Fe/Na/S (sulfur 0.27\u00a0wt%) catalysts. On average, the iron loading is slightly below the target value of 10\u00a0wt%. At the same time, the sodium and sulfur loadings clearly exceed the respective target value. With regard to the sulfur loading, weighing error might play a role, but also the influence of sulfur impurities in the pristine carbon black is conceivable, since the industrial carbon blacks used are produced on the basis of crude oil. The sulfur content in the pristine carbon black determined by ICP-OES is up to 0.02\u00a0wt%. This form of contamination is also conceivable in the case of sodium loading. The sodium content in the pristine carbon black is <0.025% by weight. An important source of error arises regarding the sodium content in the sodium citrate dihydrate used as a precursor for Na loading. According to the manufacturer, this amounts to 16.5\u201318.5\u00a0wt%. In control samples examined by ICP-OES, the content was clearly higher at around 24\u00a0wt%.The pristine carbon black and final catalyst were additionally characterized by X-ray powder diffraction (XRD) (Fig. S3). Due to the high content of amorphous domains in the samples, an intensive background noise can be observed. The diffractogram of the Printex 60 carbon black shows two significant reflections at diffraction angles of 25\u00b0 and 43\u00b0 that can be assigned to the 002 and 100 or 101 lattice planes of graphitic domains. [32] As expected, for the support loaded with iron, sodium and sulfur, additional reflexes appear, that are characteristic for iron oxides consisting of different species. Due to the very broad reflections, which may be caused by the small crystallite sizes of the supported iron oxide particles, an exact assignment of the reflections is not possible. The most dominant reflections are at 35\u00b0, 43\u00b0 and 62\u00b0. These can be assigned to the 311, 400 and 440 reflections of magnetite (Fe3O4) [33] or the 111, 200 and 220 reflections of w\u00fcstite (FeO) [34]. The low symmetry of the reflexes at 35\u00b0 and 43\u00b0 is also an indication of the superposition of at least two reflections of different species. Furthermore, the reflex intensities of the powder diffractograms shown do not correspond to the reflex ratios of phase-pure Fe3O4 or FeO species. [35]As a result of Fe, Na and S loading, changes in texture can occur resulting from the deposition of the active phase on the carbon blacks and from the restructuring of the agglomerates due to the water impregnation process. A comparison of the specific surface area and the pore volumes before and after Fe, Na and S loading is shown in Fig. S4. Both SSABET and VPore show comparable trends. Printex G and 35 supported catalysts show slight increases in SSABET and VPore (SSABET: P_G to P_G-Fe/Na/S (by 50% from 36 to 54 m2\u00a0g\u22121), P_35 to P_35-Fe/Na/S (by 22% from 59 to 72 m2\u00a0g\u22121); VPore: P_G to P_G-Fe/Na/S (by 66% from 0.06 to 0.10\u00a0cm3\u00a0g\u22121), P_35 to P_35-Fe/Na/S (by 69% from 0.16 to 0.27\u00a0cm3\u00a0g\u22121)). P_60 and P_60-Fe/Na/S include almost constant SSABET (\u223c116\u00a0m2\u00a0g\u22121) as well as slightly increasing VPore (by 25% from 0.20 to 0.25\u00a0cm3\u00a0g\u22121). The values of P_85 and P_90 behave differently, as the initial specific surface areas decrease significantly (P_85 to P_85-Fe/Na/S (by 23% from 195 to 150 m2\u00a0g\u22121), P_90 to P_90-Fe/Na/S (by 24% from 389 to 297 m2\u00a0g\u22121)). In addition to that, the loss regarding pore volume is also significant, especially for P_85-Fe/Na/S (by 47% from 0.47 to 0.25\u00a0cm3\u00a0g\u22121).Transmission electron microscopy was employed to determine the iron particle sizes distribution (Fig. 3B and Fig. 4\n). All materials show a narrow distribution of nanoscale iron oxide particles in the range of 2 to 15\u00a0nm. This is particularly evident from the low standard deviation of 2.4\u00a0nm in the case of P_35-Fe/Na/S. Fe particles larger than 20\u00a0nm were not found in any of the samples examined. The maximum of the Fe particle size distribution for P_90-Fe/Na/S is 3.7\u00a0nm. All other supports show average particle diameters in the range of 6.5\u00a0nm to 7.5\u00a0nm. These observations correspond well with results of Oschatz et al., who reported narrow size distributions of Fe particles deposited on CMK-3 and carbon black supported systems with comparable metal loadings and dopants. Average particle sizes were reported to be 5.58\u00a0nm and 4.2\u20134.7\u00a0nm for CMK-3 and carbon black supported catalysts, respectively, with standard deviations of 1.22\u00a0nm and 1.0\u00a0nm. [15,22]H2-TPR measurements (Fig. 3C) lead to comparable trends. All four examined samples include a broad reduction peak with a connected shoulder (Onset-Temperature\u00a0\u223c\u00a0200\u00a0\u00b0C). The peak temperatures lie in a range between 353\u00a0\u00b0C (P_G-Fe/Na/S and P_90-Fe/Na/S) and 388\u00a0\u00b0C (P_85-Fe/Na/S). Chew et al. and Ma et al. report two reduction peaks for Fe/CNT- respectively Fe/AC-systems in the same temperature range. The first peak between 200\u00a0\u00b0C and 300\u00a0\u00b0C is related to the reduction of Fe2O3 to Fe3O4. This would indicate that P_90-Fe/Na/S includes the highest amount of Fe2O3. The second peak between 300\u00a0\u00b0C and 400\u00a0\u00b0C can be assigned to the reduction of Fe3O4 to FeO. [36,37] A third reduction peak representing the reduction of FeO to Fe in the temperature range up to 700\u00a0\u00b0C is to be expected but is not investigated here. [37] The presence of a low-temperature shoulder due to the absence of sharply separated reduction peaks in the case of this study can be related to impoverishment effects due to the low H2 concentration of 0.5%. The observed differences in the temperature of the H2 consumption maxima and the corresponding peak intensities are most likely a consequence of differences in the metal support interactions between Fe/Na/S and the different carbon black supports. In this context it is well-known that support properties influence the reducibility as well as the ratio of FeOx species that are obtained after calcination. [36]To probe the influence of the carbon support on CO2 and CO adsorption properties, CO- as well as CO2-TPD was carried out. Reduced catalysts were used in this context, it should be noted, however, that due to instrumental limitations the reduction conditions differ from those used for the catalytic activity tests. Regarding CO2-TPD of CO2-saturated iron catalysts, studies of Cheng et al., Xu, Wang et al. and Xu, Zhai et al. show strong dependencies of the support material used. [38\u201340] All three studies report weakly and strongly chemisorbed CO2. This is also observable regarding the Fe/C-systems used in this study (Fig. 5A). Strongly chemisorbed CO2, desorbed in the temperature range between 200\u00a0\u00b0C and 300\u00a0\u00b0C, is related to the influence of highly basic Na2O-species. [40] Desorption in lower temperature regions can also be assigned to CO2 adsorbed on plain iron. [40] The resulting CO2 binding strength of the carbon black supported catalysts lies between graphene oxide and silica supported iron systems. [38,39] In addition to CO2-TPD, also CO-TPD was carried out, and shows the influence of alkali metals (Fig. 5B). [41] Strong CO chemisorption, caused by the already mentioned effect of Na2O leads to a broad shoulder (100\u2013150\u00a0\u00b0C) in addition to a desorption maximum at \u223c82\u00a0\u00b0C. This effect can be seen most significantly in case of P_90-Fe/Na/S. A reason for that is delivered by the static CO-chemisorption (Fig. S5) conducted prior the CO-TPD. P_90-Fe/Na/S includes, by far, the highest amount of reversible chemisorbed CO (4.6 mmolCO gFe\n\u22121, Fig. 3D). Due to the high amount of irreversible chemisorbed CO in case of P_60-Fe/Na/S (7.3 mmolCO gFe\n\u22121), the highest specific iron surface area (SSA\nFe) results (27.1\u00a0m2 gFe\n\u22121). The large difference between reversible (0.6 mmolCO gFe\n\u22121) and irreversible chemisorbed CO might be caused by carbonylation and therefore iron leaching at low temperature (30\u00a0\u00b0C). [34] In contrast to that P_90-Fe/Na/S exhibits a SSA\nFe of 12.5\u00a0m2 gFe\n\u22121 while P_G-Fe/Na/S (1\u00a0m2 gFe\n\u22121) and P_35-Fe/Na/S (1.9\u00a0m2 gFe\n\u22121) show much lower accessible iron surfaces. Table 1\n gives a comprehensive summary of the characterization of the pristine catalysts.Following the comprehensive characterization of the catalysts (Table 1) they were tested regarding their CO2-FTS activity. Considering both CO2 conversion (Fig. 6A) and CO yield (Fig. 6B), the five catalyst systems show considerable differences. While P_G-Fe/Na/S and P_35-Fe/Na/S show a FTS-typical start-up behaviour, with an increase of the conversion in the first 30\u00a0h from 22% and 15% to 25% and 21%, respectively, the catalyst supported on Printex 85 deactivates instantly after the start of the reaction. P_90-Fe/Na/S also shows a slight decrease in CO2 conversion (X\nCO2) over the course of the reaction time. Only P_60-Fe/Na/S shows a true steady state with an almost unchanged degree of CO2 conversion of around 35% and a constant CO yield of 12.5%. The Printex G supported catalyst reaches a steady state in terms of CO yield (Y\nCO,CO2) While P_35-Fe/Na/S and P_90-Fe/Na/S show largely identical Y\nCO,CO2 curves with an increase from 12.5% to around 17%, P_85-Fe/Na/S reaches a maximum at 10\u00a0h TOS. Beyond activity, the product spectrum also differs between the individual catalyst supports (Fig. 6C and D). The catalyst supported on Printex 85 shows a hydrocarbon yield of only 1.4%, followed closely by P_35-Fe/Na/S (4.3%) and P_G-Fe/Na/S (6.5%). The catalytic performance of P_G-Fe/Na/S and P_35-Fe/Na/S differs only minimally, both in terms of the conversion and yield values as well as in relation to the product classes formed, with selectivities to methane of 62.7% and 63.0%, to alkanes of 24.0% and 23.1% to olefins of 6.2% and 8.5% and to alcohols of 7.0% and 9.2% with Y\nHC,CO2 values of 15% and 20%, respectively, P_90-Fe/Na/S and P_60-Fe/Na/S show the highest FT yields. Strikingly, the two systems differ significantly in their product spectrum, which is reflected in both the \u03b1-values achieved and regarding the olefin fraction. The catalyst supported on Printex 90 with \u03b1-value of 0.32 tends to form significantly shorter carbon chains compared to the system supported on Printex 60 with 0.43. At the same time, almost exclusively saturated hydrocarbons are formed (\u03c3Met,CO2\u00a0=\u00a048.8%, \u03c3Par,CO2\u00a0=\u00a041.2%). Only P_60-Fe/Na/S, with an olefin content in the C2-C6 fraction of 40%, combined with the lowest selectivity to methane (24.6%) and the highest conversion of CO2 (33.5%), shows true potential for the use as catalyst in the Fischer-Tropsch-to-olefins-reaction (FTO). This is also confirmed by the chain length distribution of the different product classes (Fig. S6) as all three substance classes, C2+ alkanes, alkenes and alcohols, exhibit maxima at low chain length between C2 and C4. Above a chain length of C8, only very negligible amounts of product are detected. Table 2\n gives a comprehensive overview of the catalytic performance of all carbon black supported Fe/Na/S catalysts.In order to explore changes in catalyst properties, post mortem characterization was carried out by N2 physisorption, TEM and XRD. Regarding P_G-Fe/Na/S, P_35-Fe/Na/S and P_85-Fe/Na/S no changes in specific surface area can be observed. In contrast, SSABET of P_60-Fe/Na/S and P_90-Fe/Na/S decrease slightly (from 113 m2\u00a0g\u22121 to 89 m2\u00a0g\u22121 for P_60-Fe/Na/S and 343 m2\u00a0g\u22121 to 294 m2\u00a0g\u22121 for P_90-Fe/Na/S). A possible cause of this decrease is the sintering of the supported iron particles, as a lower number of particles with increased diameter leads to a lower specific surface area. Regarding the pore volumes according to the DFT method only in case of P_35-Fe/Na/S and P_60-Fe/Na/S a slight decrease can be observed (from 0.24\u00a0cm3\u00a0g\u22121 to 0.19\u00a0cm3\u00a0g\u22121 for P_35-Fe/Na/S and 0.22\u00a0cm3\u00a0g\u22121 to 0.19\u00a0cm3\u00a0g\u22121 for P_60-Fe/Na/S). The decrease in pore volume as a result of pore filling by long-chain FTS products is known from the literature, especially for cobalt-catalysed low-temperature FTS. [42] Due to the low average chain length of the products formed in the case of iron-catalysed HT FTS, this explanation is not convincing (Fig. S6). In addition, the decrease in pore volume would be more pronounced with an increasing rate of hydrocarbon formation, which is not the case here. It appears likely that the textural properties of the examined supports play a leading role in the activity of the corresponding FTS catalysts, as they significantly influence the resulting specific iron surface. The Printex G and Printex 35 supported catalysts have the lowest specific surface area, 54 m2\u00a0g\u22121 and 72 m2\u00a0g\u22121 respectively, and at the same time the lowest SSA\nFe (P_G-Fe/Na/S: 1 m2 gFe\n\u22121) leading to a low catalytic activity. In contrast, a higher catalytic activity is observed in the case of P_60-Fe/Na/S and P_90-Fe/Na/S. These materials also include a higher SSA\nBET of 113 m2\u00a0g\u22121 and 343 m2\u00a0g\u22121 resulting in a higher SSA\nFe of 27.1\u00a0m2 gFe\n\u22121 and 12.5\u00a0m2 gFe\n\u22121. Only P_85-Fe/Na/S represents an exception, as the specific surface area of 150 m2\u00a0g\u22121 lies between that of P_60-Fe/Na/S and P_90-Fe/Na/S whereas this system shows low SSA\nFe (1.9\u00a0m2 gFe\n\u22121) as well as the lowest FT activity.In this context, the specific surface area of the carbon black and its degree of branching appear to be the key for the observed differences. A high specific surface area in combination with a high degree of branching appears to facilitate a good initial dispersion and high specific surface areas of FeOx nanoparticles on Printex 60 as well as on Printex 90. High specific surface areas of the support, a high degree of branching as well as a high specific surface area of Fe nanoparticles all contribute to a large area of contact between the carbon black support and the FeOx nanoparticles. This large contact area might facilitate the carburization of FeOx in the first hours TOS to the FT active iron carbide phase, and thus the high hydrocarbon productivities of Fe/Na/S supported on Printex 60 and Printex 90. The other carbon black supports exhibit either a low degree of branching or comparatively low specific surface areas and do not facilitate the formation of well dispersed, accessible FeOx nanoparticles. In consequence, the area of contact between FeOx and carbon black support might be much smaller, leading to a low degree of carburization with the large remaining fraction of iron oxide species accounting for a high selectivity to CO.\nOschatz et al. list sintering as the main reason for the deactivation of carbon-supported iron catalysts in the FTS. [15] TEM images of the used catalysts confirm the sintering of Fe nanoparticles over the course of 30\u00a0h of FTS (Fig. 7A, B). A broadening of the particle size distribution as well as a shift of the maximum of the distribution towards higher particle sizes is observed (Fig. 7C). This is reflected both in the mean value but also in the standard deviation. The largest increase of particle size is recorded in case of P_90-Fe/Na/S with an increase of 350% (Fig. 7D). Likewise, the averaged Fe particle size of P_60-Fe/Na/S increases by 299%. This behaviour correlates with the observations of Torres Galvis et al., who report increases in Fe particle size of 110\u2013270% for an Al2O3-supported Fe/Na/S system after 120\u00a0h TOS. [23] It should be noted that in this study the time on stream was only 30\u00a0h, which led to a similar increase in Fe particle size. This observation is probably a consequence of different metal-support interactions, with the utilization of carbon black as a weakly interacting support leading to an increased sintering tendency. Oschatz et al., using carbon supports, also report significant increases in Fe particle sizes after 120\u00a0h or 140\u00a0h TOS. They increase on average to 25\u201328\u00a0nm and 16\u201318\u00a0nm, respectively, depending on the support material, [15,22] which translates to increases of 400% on CMK-3 and 280% on carbon black. It is also noticeable that Fe particle growth seems to correlate with catalytic activity: particle growth is significantly less pronounced regarding P_G-Fe/Na/S, P_35-Fe/Na/S and P_85-/Fe/Na/S, correlating directly with the lower FT activity of these systems.According to XRD measurements of the used catalysts two significant differences are observed (Fig. 7E, F). Additional weakly pronounced reflexes appear in the range of 40\u00b0 / 2\u03b8 to 60\u00b0 / 2\u03b8. Although challenging to assign to individual Fe phases, most likely iron carbides such as the Eckstrom-Adcock carbide (Fe3C7), cementite (\u0398-Fe3C) or a poorly defined iron carbide structure (FexCy) are responsible for these new reflexes. [43] Furthermore, it is striking that the iron oxide reflexes at 30\u00b0 / 2\u03b8, 35\u00b0 / 2\u03b8, 43\u00b0 / 2\u03b8, 57\u00b0 / 2\u03b8 and 63\u00b0 / 2\u03b8 are clearly narrower and more intense in the case of the used catalyst. Due to the intensity ratios, a phase pure FeOx species is still not present, however, the ratios known from Fe3O4 are clearly approached. Possible reasons for this observation are the transformation of FeO to Fe3O4, Fe0 as well as FeCx, while the narrowing of reflexes can be attributed to the growth of the crystallites, [36] thereby indicating sintering.In summary, P_60-Fe/Na/S emerges from the support variation as the most suitable industrial carbon black. Accordingly, this material is used as the catalyst support for the long-term stability test.The Printex 60 supported system was additionally tested for its long-term-stability over 170\u00a0h TOS (Fig. 8A, B). The conversion of CO2 reaches a steady state of about 30% directly with the setting of the operating conditions after the preconditioning of the catalyst at 48\u00a0h TOS. Over the entire test period of 170\u00a0h TOS, X\nCO2 fluctuates only within the range of 28.2% to 31.5%, while no deactivation is observed. A similar behaviour for K/Mn/Fe catalysts on nitrogen-doped CNTs is reported by Kangvansura et al.. Following the first 40\u00a0h of TOS, which roughly corresponds to the duration of the preconditioning performed in the present work, they report steady-state CO2 conversions of 25.4% to 31.8% for the following 20\u00a0h of TOS at 25\u00a0bar and 360\u00a0\u00b0C. The group attributes the deactivation during the induction period to the growth of the iron particles. [43] It should be noted that catalyst deactivation might also be caused by re-oxidation of FT-active iron carbide species to FeOx species. However, in case of CO2 hydrogenation to olefins this connection is non-trivial, as the presence of iron oxide species are integral to the desired reaction pathway as they catalyse the conversion of CO2 to CO via the RWGS equilibrium. As CO2 conversion as well as the yield of CO and hydrocarbons remains stable over 170\u00a0h TOS, we assume that after preconditioning a steady state between iron carburization and oxidation is established which results in a constant ratio of FeOx and FeCx species. The steady-state yields of hydrocarbons and CO are \u223c16% in the case of Y\nHC,CO2 and\u00a0\u223c\u00a014% for Y\nCO,CO2, corresponding to selectivities of 53% and 46%, respectively. Thus, the CO selectivity of 48.5% published by Kangvansura et al. on a Mn/Fe/NCNT system is achieved. [43] Significant changes over the reaction time occur exclusively with regard to the products formed. In this context, from 80\u00a0h TOS, an increase in the proportion of methane from 22% to 38% can be observed. At the same time, the proportion of C2+-alkanes decreases from 23% at the beginning to 16%. Alkenes and alcohols show similar trends. Starting from initial values of 33% and 15%, respectively, both proportions increase to 42% and 18%, respectively, up to a TOS of 62\u00a0h, and then decrease.Between 100\u00a0h and 175\u00a0h TOS, both values fluctuate in very narrow ranges between 31% and 33% in the case of the alkenes and between 13% and 14% in the case of the alcohols. Consequently, regarding the whole reaction time, no significant changes for these to product classes exist. The observations described are also reflected in the chain growth probabilities and olefin fractions in the C2-C6 fraction shown in Fig. 5B. As a result of the increasing fraction of methane formed, the chain growth probability drops slightly from 0.48 to 0.45. However, it should be noted that the plotted \u03b1-values are subject to a considerable scatter in the range of 0.5 to 0.43. At the same time, due to decreasing alkane-content, the share of olefins on the C2-C6 fraction rises steadily from 42% to almost 50% from a TOS of 90\u00a0h onwards. At that point additionally studies are necessary to explain the changes regarding the product selectivities.Within the present study, iron-based Fischer-Tropsch catalysts were prepared with sodium and sulfur promoters supported on five industrially available, beaded carbon blacks. The carbon blacks exhibited different structural parameters, with specific surface areas ranging from 36 m2\u00a0g\u22121 to 380 m2\u00a0g\u22121. As a result of testing these systems regarding their suitability as CO2-FTS-catalysts with the target product short-chain olefins, clear correlations emerge between the structure of the materials and their catalytic activity. Thus, the combination of aggregates with a high degree of branching and a specific surface area in the range of 150 m2\u00a0g\u22121 proves to be desirable. Using the most promising system (X\nCO2\u00a0=\u00a035%, \u03c3C2-C6-Ole,CO2\u00a0=\u00a040%), the long-term stability of the Fe/C catalysts over 170\u00a0h TOS was demonstrated, whereas no decrease in CO2 conversion was observed. In this study we could demonstrate that inexpensive, industrial available beaded carbon blacks can be utilized as supports to prepare easy-to-handle Fe based FTS catalysts exhibiting high stability along with similar catalytic activity and selectivity compared to considerably more expensive carbon nanomaterials such as CNT's.\n\nAci\n\nacids\n\nAlc\n\nalcohols\n\nCB\n\ncarbon black\n\nCMK-3\n\nordered mesoporous carbon\n\nCNT\n\ncarbon nano tubes\n\nDBP\n\ndibutyl phtalate\n\nDFT\n\ndensity functional theory\n\nFID\n\nflame ionisation detector\n\nFT-IR\n\nfourier-transform infrared\n\nFTO\n\nfischer-tropsch-to-olefins\n\nFTS\n\nfischer-tropsch-synthesis\n\nGC\n\ngas chromatograph\n\nGHSV\n\ngas hourly space velocity\n\nSTEM\n\nscanning transmission elektron microskopy\n\nICP-OES\n\noptical emission spectroscopy with inductiv coupled plasma\n\nIWI\n\nincipient wetness impregnation\n\nHC\n\nhydrocarbons\n\nMBET\n\nmulti-point brunauereEmmett-teller-methode\n\nMet\n\nmethane\n\nMTO\n\nmethanol to olefins\n\nO/P\n\nolefin\u2212/paraffin ratio\n\nOEC\n\norion engineered carbons\n\nOle\n\nolefins\n\nPar\n\nparaffins\n\nSMSI\n\nstrong metal support interaction\n\nTCD\n\nthermal conductivity detector\n\nTPD/TPR\n\ntemperature programmed desorption / reduction\n\n(R)WGS\n\n(reverse) watergas-shift-reaction\n\nXRD\n\nx-ray diffraction\n\n\nacidsalcoholscarbon blackordered mesoporous carboncarbon nano tubesdibutyl phtalatedensity functional theoryflame ionisation detectorfourier-transform infraredfischer-tropsch-to-olefinsfischer-tropsch-synthesisgas chromatographgas hourly space velocityscanning transmission elektron microskopyoptical emission spectroscopy with inductiv coupled plasmaincipient wetness impregnationhydrocarbonsmulti-point brunauereEmmett-teller-methodemethanemethanol to olefinsolefin\u2212/paraffin ratioorion engineered carbonsolefinsparaffinsstrong metal support interactionthermal conductivity detectortemperature programmed desorption / reduction(reverse) watergas-shift-reactionx-ray diffraction\n\n\u03b1\n\nchain growth propability, \u2013\n\n\nd\nP,cCB/iron\n\n\ncarbon black / iron diameter, nm\n\n\nQ\nCO,Ads,rev/irrev\n\n\nquantity of reversible / irreversible adsorbed CO, mmol gFe\n\u22121\n\n\n\nS\np,r\n\n\nselecivity to product p from reactant r, mol\u00a0mol\u22121\n\n\n\n\u03c3\nHC,r\n\n\nshare of hydrocarbon HC related to all HC's, mol\u00a0mol\u22121\n\n\n\nS\nHC,r\n\n\nselecivity to hydrocarbon HC from reactant r, mol\u00a0mol\u22121\n\n\n\nSSA\nBET\n\n\nspecific surface area using BET-method, m2\u00a0g\u22121\n\n\n\nTOS\n\n\ntime on stream, h\n\n\nV\nPore\n\n\npore volume, cm3\u00a0g\u22121\n\n\n\nw\nFe,Kat\n\n\nweight fraction of iron related to total catalyst mass, g\u00a0g\u22121\n\n\n\nX\nr\n\n\nconversion of reactant r, mol\u00a0mol\u22121\n\n\n\nY\np,r\n\n\nyield of product p out of reactant r, mol\u00a0mol\u2212\n\n\n\nchain growth propability, \u2013carbon black / iron diameter, nmquantity of reversible / irreversible adsorbed CO, mmol gFe\n\u22121\nselecivity to product p from reactant r, mol\u00a0mol\u22121\nshare of hydrocarbon HC related to all HC's, mol\u00a0mol\u22121\nselecivity to hydrocarbon HC from reactant r, mol\u00a0mol\u22121\nspecific surface area using BET-method, m2\u00a0g\u22121\ntime on stream, hpore volume, cm3\u00a0g\u22121\nweight fraction of iron related to total catalyst mass, g\u00a0g\u22121\nconversion of reactant r, mol\u00a0mol\u22121\nyield of product p out of reactant r, mol\u00a0mol\u2212\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Stephan Schultheis acknowledges a fellowship from the Darmstadt Graduate School of Excellence Energy Science and Engineering. Felix Herold acknowledges a fellowship within the Walter-Benjamin-program of the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation, project number 471263729). The authors acknowledge funding from DFG within CRC 1487 (Iron, upgraded!; project number 443703006). Orion Engineered Carbons is acknowledged for providing carbon black samples.\n\n\n\nSupplementary material: Carbon black characterization by TPD-MS, XRD and N2 physisorption. Fe/Na/S catalyst characterization by, CO chemisorption, ICP-OES, XRD and N2 physisorption.\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2023.106622.", "descript": "\n                  The Fischer-Tropsch-to-Olefins process allows to convert waste stemming CO2 with green hydrogen to olefins. Iron can catalyse both core reactions: 1) reverse-water-gas-shift as well as 2) Fischer-Tropsch. Carbon supported catalysts were reported to be highly attractive in this context, but until now mainly non technically applicable research carbons like nanotubes or ordered mesoporous carbons were studied and long term stability studies are missing. Here, beaded carbon blacks, were studied as available and inexpensive support materials for Fe catalysts in CO2-based FTO. The most promising support yielded selectivities towards olefins of almost 40% and showed for 170\u00a0h high stability.\n               "}
{"full_text": "Energy consumption in the world is rapidly increasing due to economic developments. Approximately 80% of the world's primary energy is currently provided from fossil fuels [1,2]. This excessive dependence on fossil fuels contributes to severe environmental degradation such as water, air, and soil pollution. Therefore, biomass is considered a promising feedstock for producing renewable fuels and chemicals [3,4].A potentially efficient and cost-effective method for converting biomass to liquid fuel (bio-oil) is fast pyrolysis [1,4\u20137]. Fast pyrolysis is a thermochemical process that decomposes biomass into bio-oil, bio-char, and gaseous products [8]. Compared to conventional petroleum-derived fuels, pyrolysis oil has acidic and corrosive properties, high water content, and a relatively low energy density, making it challenging to utilize as transportation fuels [9,10]. In addition, the high oxygen content in crude bio-oils, usually 20 to 50\u00a0wt%, results in a low heating value, poor stability and volatility, high viscosity, and corrosiveness [10\u201312]. Thus, bio-oil upgrading is required to lower its water and oxygen contents for further applications [11\u201314].Catalytic hydrodeoxygenation (HDO), one of the most effective methods for bio-oil upgrading, involves the stabilization and selective removal of oxygen from untreated bio-oil [11,13,15\u201318]. Identifying highly active and stable catalysts for upgrading bio-oil in pilot scale has been the objective of several studies [13,17,18]. Jahromi et al. [11] investigated hydrotreating of guaiacol (GUA) using red mud-supported nickel and commercial Ni/SiO2-Al2O3 catalysts at different reaction temperatures (300, 350, and 400\u00a0\u00b0C) and initial hydrogen pressures (4.83, 5.52, and 6.21\u00a0MPa). They found that the major products of the hydrotreating process were catechol, anisole, phenol, cyclohexane, hexane, benzene, toluene, and xylene. Ly et al. [14] examined various catalysts such as CoMoP/\u03b3-Al2O3, Co/\u03b3-Al2O3, Fe/\u03b3-Al2O3, and HZSM-5 to upgrade bio-oil from Saccharina japonica alga. Among the catalysts, Co/\u03b3-Al2O3 showed higher HDO performance than others (HHV of bio-oil was 34.41\u00a0MJ/kg). The kerosene-diesel fraction (C12\u2013C14) increased from 36.17 to 38.62\u201348.92\u00a0wt% by catalytic HDO.Activated carbon (AC) has been used as a promising catalyst support due to its high surface area, which allows for the high dispersion of metal species [19\u201322]. Unlike common activated carbon made from other biomasses such as coconut shell, palm, or coal, bamboo-based activated carbon has relatively large pores suitable for the adsorption of large molecules due to the coarse texture of the raw bamboo. Kim et al. [23] showed that the pore size structure of bamboo-based activated carbon has a larger BET surface area (1329\u00a0m2/g), as compared to coconut-based activated carbon (1199\u00a0m2/g) and carbon cyrogel (639\u00a0m2/g). This characteristic made bamboo-based activated carbon more attractive catalyst support than those from other biomass materials. It was also reported that AC reduces the reactivity while improving the selectivity in HDO products [19,20]. Jin et al. [19] investigated \u201chydrogen-free\u201d HDO of GUA in a high-pressure batch reactor. They found that the activity of Ru/C catalyst is superior to other studied catalysts (i.e., Au/C, Pd/C, and Rh/C). Using 10\u00a0wt% Fe/AC at 300\u00a0\u00b0C and atmospheric pressure, Tran et al. [20] successfully hydrodeoxygenated GUA into cresol and 1,2-dimethoxybenzene. Higher selectivity but lower HDO product diversity was obtained by Fe/AC, compared to Ni/Al2O3. Jin et al. [21] studied in-situ HDO of GUA over Ni-based nitrogen-doped activated carbon-supported catalysts (Ni/PANI-AC), improving the conversion of GUA by 8 % compared to Ni/AC catalyst. This improvement can be attributed to the acid-base properties and modified electronic properties, which promote the C-O cleavage and enhance the dispersion of Ni particles on the surface of the catalyst.Transition metals, such as Co and Fe, are often used as catalysts in HDO because they are less expensive than precious metals (e.g., Pt, Pd, Ru, Rh, Ir, etc.) [20,24\u201326]. Among these catalysts, Co was proposed to promote catalytic activity in HDO of 2-furyl methyl ketone (FMK) as a model compound in bio-oil from pyrolysis of Saccharina Japonica alga [27]. The conversion of FMK was found to be in the following order: Co\u00a0>\u00a0Mo\u00a0>\u00a0Ni. Using Ni-Co/\u03b3-Al2O3 catalysts, Raikwar et al. [25] obtained 98.9 % conversion of GUA along with 35.2 % selectivity of benzene and 59.1 % selectivity of cyclohexane at 302\u00a0\u00b0C. Since Fe has been reported to have lower activity in benzene hydrogenation than other transition or precious metals (Ni, Co) [28,29], it is expected to provide a good balance between activity and selectivity [30,31]. Ly et al. [24] investigated HDO of FMK over 5\u00a0wt% Fe2P/\u03b3-Al2O3 catalyst and obtained the highest conversion of 92.6 % into 2-allyl furan (79.34%) and methylcyclohexane (13.26%) at 400\u00a0\u00b0C.The present study aims to determine inexpensive and effective catalysts for upgrading bio-oils. Another objective of this study is to control the reaction pathways during bio-oil upgrading to increase the selectivity of valuable chemicals, especially methyl phenol derivatives, in the liquid products. Methyl phenol derivative, one of the most common components of lignocellulosic bio-oils, is an important chemical intermediate and is essential to produce various chemicals and materials such as phenolic resins, alkylphenols, and more. It is also used as a precursor to other compounds and materials, including plastics, pesticides, pharmaceuticals, and dyes [32]. However, the current technologies for producing methyl phenol derivatives are challenging because of their high cost and increased pollution to the environment [33]. For this reason, in this study, AC-supported Co, Fe, and bimetallic (Co-Fe) catalysts were investigated for HDO of pyrolysis oil from wood pallet sawdust (WPS) in an autoclave using various techniques. In addition, the performance and deactivation of various AC-supported catalysts were systematically investigated under different operating conditions.The wood pallet sawdust (WPS) was provided by Hanssem Co. LTD. (Korea). The samples were subjected to drying at 105\u00a0\u00b0C overnight to remove the moisture before the experiments. The bio-oils (organic phase) were obtained by fast pyrolysis in a fluidized bed reactor (pilot-scale 20\u00a0T/day) at 450\u00a0\u00b0C with a fluidization velocity of 2.5\u00a0\u00d7\u00a0Umf (Umf\u00a0=\u00a04.5\u00a0L/min), using silica sand as bed material. The thermogravimetric analysis (TGA N-1000, SINCO) was carried out under a nitrogen flow rate of 20\u00a0mL/min to understand the thermal decomposition of the biomass and bio-oil. During this analysis, the temperature was increased from 20 to 700\u00a0\u00b0C at a heating rate 10\u00a0\u00b0C/min (Figure S1).In this study, the AC support was produced from bamboo through two steps, i.e., carbonization and activation. These process details are described elsewhere [20,22]. After cooling to room temperature, the AC was sieved to homogenize the particle size between 75 and 180\u00a0\u00b5m. The catalysts were prepared by an incipient wetness impregnation method with a 10\u201330\u00a0wt% metal loading (Co and Fe) on AC. The calculated amounts of cobalt(II) nitrate hexahydrate (Co(NO3)2\u00b76H2O, 98%, Sigma-Aldrich) and iron(III) nitrate nonahydrate (Fe(NO3)3\u00b79H2O, 98%, Sigma-Aldrich) were hydrolyzed with deionized water. Following impregnation, the catalysts were dried at 105\u00a0\u00b0C for 24\u00a0h. The catalysts on AC were pretreated with 250\u00a0mL/min of N2 at 500\u00a0\u00b0C (Co catalyst) and 550\u00a0\u00b0C (Fe catalyst) at atmospheric pressure due to high oxygen content on the surface of AC [22]. For the preparation of the bimetallic Co-Fe catalyst, 20\u00a0wt% Co-Fe with various ratios was chosen to investigate the HDO process. These pretreated catalysts are denoted as xCo-yFe, in which x and y are the weight ratio of Co and Fe (i.e., 1:1, 2:1, 3:1, and 4:1) loaded on AC, respectively.The crystallographic structure of the catalysts was examined using an X-ray diffractometer (MAC-18XHF, Rigaku, Japan) with a CuK\u03b1 radiation source (\u03bb\u00a0=\u00a01.54 A), which was operated at a scanning rate of 5\u00b0/min in 2\u03b8 range of 10\u00b0 to 80\u00b0. The textural properties of the catalysts were analyzed using a N2 porosimeter (Tristar, Micromeritics, USA). The specific surface area was calculated via the multipoint Brunauer\u2013Emmett\u2013Teller (BET) method within a relative pressure (P/Po) range of 0.05\u20130.25. The morphology of the prepared catalysts was analyzed using field-emission scanning electron microscopy (FE-SEM; Leo-Supra 55, Carl Zeiss STM, Germany), while Transmission Electron Microscopy (JEM 2100, JEOL, Japan, accelerating voltage: 20\u00a0kV) was used for confirmation of carbon deposit on spent catalyst after the reaction. Apart from TEM, a thermogravimetric analyzer (TGA-Scinco 1000, Korea) was used to determine the amount of carbon on the spent catalyst, which was washed with acetone, and dried at 105\u00a0\u00b0C to remove the contaminants after the reaction. For temperature program reduction (TPR) analysis, the catalysts were treated under H2 and Ar (2:8 mL/min) at atmospheric pressure in a fixed bed reactor. The reduction process continued up to 800\u00a0\u00b0C, then cooled and re-passivated for a catalytic deoxygenation reaction [20,24,32]. Gas chromatography (GC) with a thermal conductivity detector (TCD) was employed to measure the hydrogen consumption of the catalysts. The C, H, and N contents of the AC from bamboo were determined using an elemental analyzer (Flash EA 1112, CA Instrument, USA), and the O content was determined by the difference.Temperature-programmed desorption (TPD) of ammonia was investigated using a mass spectrometer (HPR20. Hyden Analytical) to evaluate the total acidity of the catalysts [20,27]. The catalysts were treated with He gas, followed by ammonia adsorption at 100\u00a0\u00b0C with 4\u00a0vol% of NH3 in He gas. The catalyst samples were flushed with He gas to remove the adsorbed NH3. The amount of desorbed NH3 under He flow was measured using gas chromatography with a TCD. X-ray photoelectron spectroscopy (XPS, K-alpha X-ray Photoelectron Spectrometer, Thermo Scientific) measurement was also carried out to determine the oxide states of metal dopants (Co, Fe). The carbon C (1s) line was used as a reference with a binding energy value at 284.6\u00a0eV. The reduced catalyst was passivated in a 3% air in Ar at room temperature for 2\u00a0h before exposing it to the air [33].Upgrading of WPS bio-oil was carried out in an autoclave reactor. As shown in Figure S2, the system has been reported in our previous research [14]. The reactor was heated from 300 to 350\u00a0\u00b0C under different initial hydrogen pressures, ranging from 25 to 60\u00a0bar. With the autoclave submerged in the molten salt bath, a catalyst/bio-oil ratio of 1.5/10 (i.e., 3\u00a0g of catalyst with 20\u00a0g bio-oil sample) was employed in the experiments. In each experiment, the residence time was fixed at 60\u00a0min, and the reactor was removed from the bath and cooled to room temperature after each run. After HDO, the product yield was calculated from the amounts of the product and the bio-oil fed to the system. The gaseous product was collected with a gas sampling bag, and the gas yield was determined by measuring the weight difference of the bag before and after the collection. The liquid and solid products were separated using a solvent extraction technique with a microfilter paper (pore size: 0.45\u00a0\u03bcm). Then, the solid yield was calculated by weighing the solid after drying, while the liquid yield was given by difference. For all the presented calculations, each data point was an average of at least three experimental results. The elemental compositions of the upgraded bio-oils were characterized by Flash EA1112, CE Instrument [14,34], while the moisture content was measured by a Karl-Fischer (CA-200, Mitsubishi, Japan). The gas compositions were identified by gas chromatography (YL 6500GC, YL Instrument, Korea) equipped with dual detectors, a flame ionization detector (FID) with a Porapak N column to analyze hydrocarbon gases (C1\u2013C4) and a TCD with a Molecular sieve 13X column for H2, CO, CO2, and CH4. The FID was operated at 250\u00a0\u00b0C and a flow rate of 20\u00a0mL/min N2 (ultra-high purity, 99.999%) as a carrier gas, while the TCD detector at 150\u00a0\u00b0C under the same flow rate argon (ultra-high purity, 99.999%). The gas compositions were also analyzed by a gas chromatograph/mass spectrometry (GC/MS, 7890A, Agilent, USA), with a capillary column of HP-5MS (30\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm\u00a0\u00d7\u00a00.25\u00a0\u03bcm) under a constant flow rate of helium (1.0\u00a0mL/min). The oven temperature was programmed at 40 to 90\u2009\u2103 at a heating rate of 10\u2009\u2103/min, followed by 40\u2009\u2103 /min to 250\u2009\u2103 while the detector temperature was set to 280\u2009\u2103. A 13C NMR spectrometer at 300\u00a0MHz was employed to characterize functional groups of the bio-oils dissolved in dimethyl sulfoxide-d6 (DMSO\u2011d6). The high heating value (HHV) of the bio-oil was measured using a bomb calorimeter (SDC715 Calorimeter, Sundy). The degree of deoxygenation (DOD) describes the effectiveness of oxygen removal, indicating the quality of the produced bio-oil. The degree of deoxygenation is calculated by the following equation [35]:\n\n(1)\n\nD\ne\ng\nr\ne\ne\n\no\nf\n\nd\ne\no\nx\ny\ng\ne\nn\na\nt\ni\no\nn\n\n\n\nD\nO\nD\n\n\n=\n\n\n1\n-\n\n\nwt\n%\n\nO\n\n\ni\nn\n\nu\np\ng\nr\na\nd\ne\nd\n\nb\ni\no\n-\no\ni\nl\n\n\n\n\nwt\n%\n\nO\n\n\ni\nn\n\nr\na\nw\n\nb\ni\no\n-\no\ni\nl\n\n\n\n\n\n\n\u00d7\n100\n%\n\n\n\n\nTable 1\n shows the characteristic of WPS and WPS pyrolysis bio-oil. The moisture and ash contents of the WPS were measured to be 22.99\u00b10.17 and 2.43\u00b10.47\u00a0wt%, while those of the WPS pyrolysis bio-oil were 3.02\u00b10.11, and 0.34\u00b10.04\u00a0wt%, respectively. Elemental analysis shows that the C, H, N, and O contents of the WPS were 49.95\u00b10.11, 6.07\u00b10.03, 0.61\u00b10.02, and 43.37\u00b10.24\u00a0wt%, respectively. For the WPS bio-oil, the contents were 61.27\u00b10.14, 6.45\u00b10.22, 0.29\u00b10.02, and 31.99\u00b10.09\u00a0wt%. The HHV of WPS and WPS bio-oil were 18.43\u00b10.05 and 25.69\u00b10.14 MJ/Kg.The textual properties of the AC-supported Co and Fe catalysts are presented in Table 2\n. The specific surface area (SBET), pore volume, and average pore size of the AC support pretreated at 550\u00a0\u00b0C were determined to be 604.4300\u00a0m2/g, 0.0431\u00a0cm3/g, and 2.4500\u00a0nm, respectively [20]. As the Fe content increased from 10 to 30%, the SBET decreased from 111.5020 to 44.0120\u00a0m2/g, while pore size increased from 21.2101 to 9.5506\u00a0nm, and pore volume increased from 0.0297 to 0.993 and then decreased to 0.0424\u00a0cm3/g. While increasing the Co content from 10 to 30%, the SBET, pore size, and pore volume increased from 9.1974 to 19.0026\u00a0m2/g, 4.3157 to 21.2101\u00a0nm, and 0.0087 to 0.0387\u00a0cm2/g, respectively. With increasing the ratio of Co:Fe from 1:1 to 4:1, the SBET of bimetallic Co-Fe/AC decreased from 350.8678 to 117.7973\u00a0m2/g, whereas pore size and volume increased from 2.6370 to 3.2530\u00a0nm and from 0.0303 to 0.0885\u00a0cm2/g, respectively.The X-ray diffraction (XRD) patterns of the AC-supported catalysts are shown in Fig. 1\n. The diffraction peaks at 2\u03b8\u00a0=\u00a024.0, 28.8, 30.0, 31.2, 33.9, 40.5, 50.0, and 73.7\u00b0 represent the formation of KHCO3 in the AC. It is likely due to the high content of K in bamboo [20]. The diffraction peaks at 16.16 and 20.71\u00b0 are attributed to the formation of CoCl2\u00b7H2O. The peaks assigned to CoCO3, CoO, and Co are observed at 32.59, 36.81, and 43,97\u00b0, respectively [38\u201340]. The XRD pattern of Fe/AC and Co-Fe/AC catalysts exhibit peaks assigned to Fe3O4 (35.63, 57.29, and 62.93\u00b0) and Fe (44.81\u00b0). The formation of spinel CoFe2O4 (35.40 and 62.54\u00b0) was observed from the spent catalyst (S-20\u00a0wt% 4Co-1Fe/AC).The TPR analysis was conducted to investigate the reduction of the metal oxides to metals by contact between metal species and the supports [36]. As shown in Fig. 2\n, an increase in Co loading from 10 to 30% increased the reduction temperature of Co catalyst from 492 to 533\u00a0\u00b0C. A similar trend was observed with an increase of Fe loading, leading to an increase of reduction temperature from 640 to 688\u00a0\u00b0C. The increase of Co loading in Co-Fe/AC catalyst led to the increase of reduction temperature slightly. Reduction behaviors of bimetallic catalysts were influenced by their compositions (i.e., Co:Fe ratio). For example, the reduction peaks of the Co-Fe/AC catalyst were observed to shift to higher temperatures with increasing the Co:Fe ratio (from 1:1 to 4:1).The profiles for temperature-programmed desorption of ammonia (NH3-TPD) are presented in Fig. 3\n. The desorption temperature illustrates the potency of the catalyst's active sites. Based on the desorption temperature, the acid sites can be classified into weak (<250\u00a0\u00b0C), moderate (250\u2013500\u00a0\u00b0C), and strong acid sites (>500\u00a0\u00b0C) [27,37]. The loading of metal on AC resulted in changes in the distribution of active sites. The Co catalyst only has strong acid sites, while the Fe catalyst has both weak and strong acid sites. The increased loading of Co species in Co/AC catalyst resulted in a slight increase of strong acid sites. However, with changing the Fe loading amount in Fe/AC catalyst, the number of active sites was in the following order: 20\u00a0wt% Fe/AC\u00a0>\u00a030\u00a0wt% Fe/AC\u00a0>\u00a010\u00a0wt% Fe/AC. The amount of both weak and strong acid sites increased with the Co ratio in Co-Fe/AC catalyst. The acidity density of bimetallic catalysts was higher than that of the monometallic catalysts, suggesting that the bimetallic catalysts would have higher catalytic activities. It was found that 20\u00a0wt% 4Co-1Fe/AC showed the most elevated acidity among the tested catalysts.The SEM-EDS of activated carbon in Fig. 4\n shows a mixture of smaller irregular geometrical (20\u00a0wt% Co/AC, 20\u00a0wt% Fe/AC), cubic (20\u00a0wt% 4Co-1Fe/AC) with Co, Co3O4, Fe, and Fe3O4 particles agglomerated together on the surface of activated carbon support (>10\u00a0\u03bcm). It is possibly due to the different amorphous phases in the sample, as confirmed by the earlier XRD analysis. The growth of the particles in the solution during the synthesis was facilitated by the Ostwald ripening, during which the smaller particles grew in size while decreasing in number as highly soluble small particles dissolved and re-precipitated. In addition, a particle with negative curvature exhibited lower solubility than another particle with positive curvature. Therefore, the latter inclined to precipitate on the surface of the former, leading to growing of the particle\u2019s neck, strengthening the particle\u2013particle cluster during agglomeration [38]. The TEM images of selected catalysts are shown in Figure S3 (a-e). The dark spots in Figure S3 (a-c) are attributed to the presence of metal oxides (Co3O4 and Fe3O4), which is in good agreement with the XRD analysis [39].As shown in Fig. 5\n, X-ray photoelectron spectroscopy (XPS) analysis was carried out to investigate the valence state of Fe and Co in the catalysts [33]. Fig. 4 (a) shows that the XPS spectrum for Co2p can be deconvoluted into four and five peaks for 20\u00a0wt% Co/AC and 20\u00a0wt% 4Co-1Fe/AC, respectively. Two peaks centered at 780.2 and 784.6\u00a0eV correspond to Co2p3/2 and Co2p3/2sat, while the other peaks at 796.3 and 802.3\u00a0eV are attributed to Co2p1/2 and 2p1/2sat. Other observed peaks show Co2p3/2 and 2p1/2 spin\u2013orbit doublet due to the formation of Co-oxide phases (i.e., CoO or Co3O4) [39\u201341]. For fresh and spent 20\u00a0wt% 4Co-1Fe/AC catalysts, the shake off satellites were detected, which occur when the valence electron is ejected from the ion completely (to the continuum). It appears as a broadening of the core level peak or contribution to the inelastic background. In addition, unfilled shell containing unpaired electrons might cause multiplet splitting, which result in two broad peaks around 785\u00a0eV. Similarly, the XPS spectrum of Fe2p is deconvoluted into three peaks, as shown in Fig. 4 (b). Peaks corresponding to Fe2p3/2 and Fe2p3/2sat were observed at binding energies of 709.8 and 714.3\u00a0eV, whereas a peak at 723.4\u00a0eV is assigned to Fe2p1/2. The peak at 709.8\u00a0eV seems to correspond to the metallic phase of Fe, while other peaks centered at 714.3 and 723.4\u00a0eV seem to be due to the formation of the iron oxide phase (i.e., Fe3O4) [39\u201341]. These results are in good agreement with the those from the XRD analysis. Fig. 4 clearly shows that the XPS spectrum of both Co2p and Fe2p changed after reactions, indicating that the ratio of metal and metal oxide species was altered by the reaction. These changes in the chemical state and in the oxygen species of the catalyts lead to the deactivation of the catalysts.The WPS bio-oil was upgraded under different temperatures and hydrogen pressures in the presence of AC. The product yields and elemental analysis of the upgraded oil are summarized in Table 3\n. As the reaction temperature increased, the liquid yield decreased, while the solid and gas yields increased. With an increase in temperature from 300 to 350\u00a0\u00b0C at 25\u00a0bar, the liquid yield of the bio-oil decreased from 80.06 to 55.36\u00a0wt%, while the solid and gas yields increased from 13.83 to 37.18\u00a0wt% and from 6.11 to 7.46\u00a0wt%, respectively. It was also found that with increasing hydrogen pressure, the gas and liquid yields increased, but the solid yield decreased. For example, when the hydrogen pressure increased from 25 to 60\u00a0bar at 350\u00a0\u00b0C, the solid yield decreased from 37.18 to 26.79\u00a0wt%, while the gas and liquid yields increased from 7.46 to 12.53\u00a0wt% and from 55.36 to 60.68\u00a0wt%.The HHV of bio-oil is also influenced by the operating conditions. As seen in Table 3, hydrogen pressure and reactor temperature enhance the HHV of the upgraded bio-oils. Consequently, among the conditions examined in this study, HDO at 350\u00a0\u00b0C and 60\u00a0bar provided the highest HHV of 33.72\u00a0MJ/Kg, which is higher than 25.69\u00a0MJ/Kg from the raw bio-oil.The gaseous products mainly consist of CO2, CO, CH4, and small amounts of hydrocarbons (C2-C4). H2 in the gas products was not measured in this study because it is generated, while consumed simultaneously through the bio-oil upgrading process. The CO2, CO, and CH4 were produced by decarboxylation, decarbonylation, and demethylation, while other hydrocarbon gases were formed by a cracking reaction. By increasing the reaction temperature and hydrogen pressure, the selectivity of CO2 was slightly decreased from 84.14 to 80.93%. On the other hand, the selectivity of CO and CH4 increased from 4.52 and 3.99% to 9.23 and 8.15%. It is most likely that as temperature and pressure increase, CO production is further promoted through a secondary cracking of volatiles, followed by a reduction of CO2, while other hydrocarbons are stable [42].\nTable 4\n presents the liquid compositions of upgraded bio-oil under different operating conditions. In the raw WPS bio-oil, the main components were found to be phenolic compounds and benzenediol derivatives. In particular, the area% of phenol, methyl phenol, and benzenediol derivatives in the raw WPS bio-oil were measured to be 5.92, 15.67, and 35.09%, respectively. After the HDO at 350\u00a0\u00b0C and 60\u00a0bar, the area% of phenol and methyl phenol increased to 6.99 and 19.51%. In contrast, the area% of benzenediol derivatives decreased to 8.95%. These results indicate that a hydroxyl group (\u2013OH) in the benzenediol derivatives is cleaved to form the phenolic compound, especially methyl phenol derivatives. The mechanism of hydroxyl group cleavage has also been elucidated using a bio-oil model compound such as guaiacol in our previous studies [10,11,20,21] and phenol [16,26]. During the upgrading process, the cracking reaction produces lighter components such as acetic acid (9.10%) and butanoic acid (3.44%). Besides, some transfers of alkyl groups indicate that alkylation proceeded during the bio-oil upgrading. The Me-O bond in the bio-oil components could be transformed into a methyl radical remaining attached to the catalyst surface [43], which allows a transalkylation in which the methyl radical is attached to the phenol in an ortho position via electrophilic substitution. This result agrees with the findings reported by Zhao et al. [44] and Bui et al. [43,45], who proposed a pathway for the formation of alkyl phenols. The moisture content of the upgraded oil was 17.30%, which is much higher than that of raw bio-oil (3.25%). Accordingly, the DOD of bio-oil increased from 40.21 to 63.66% after the HDO, suggesting that effective dehydration occurred during the bio-oil upgrading process.The yields and the distribution of HDO products for different catalysts are shown in Table 5\n. For single-metal atom catalysts, 10, 20, and 30\u00a0wt% of Co or Fe were loaded on AC while 20\u00a0wt% of Co-Fe dual atoms with different ratios of Co/Fe were loaded for Co-Fe bimetallic catalysts.Among the mono-metal catalysts, a maximum liquid yield of 70.46\u00a0wt% was obtained at 20\u00a0wt% Co, while an increase of Fe loading decreased the liquid yield. For the mono Fe catalysts, a maximum liquid yield (59.33\u00a0wt%) was observed at 10\u00a0wt% Fe. As the Fe loading increased from 10 to 30\u00a0wt%, the bio-oil (liquid) yield decreased sharply from 59.33 to 36.78\u00a0wt%, while the solid and gas yields increased from 36.66 to 56.83\u00a0wt% and from 4.01 to 6.93\u00a0wt%, respectively. Excessive loading of Fe could lead to the agglomeration of the metal species and the reduction of the active sites on the catalysts. In addition, as shown in Table 2, the pore size decreased with the amount of Fe loading onto AC, which is probably due to the pore blockages caused by the growth of metal ions or the formation of bulk metal oxides during the impregnation and/or calcination step. Tran et al. [20] reported similar findings from HDO of GUA over Fe/AC and Ni/\u03b3-Al2O3 catalysts synthesized by the impregnation method. Abu and Smith [46] also obtained comparable results using Co-added MoP and Ni2P catalysts in hydrodesulfurization of 4,6-dimethyldibenzothiophene. In this study, 20\u00a0wt% Co and 10\u00a0wt% Fe were found to be optimal by facilitating the uniform dispersion of metal atoms on the AC surface. Among the tested mono-metal catalysts, the bio-oil upgraded using 20\u00a0wt% Co/AC offered the highest HHV (34.22\u00a0MJ/Kg). Differences in the textural properties of catalysts could lead to these results.Based on the results from the mono-metal catalysts, 20\u00a0wt% of Co was chosen for bimetallic Co-Fe catalysts, and the effect of Co/Fe ratio (1 to 4) was further studied. By increasing the ratio of Co to Fe, both bio-oil and gas yields increased from 65.76 to 68.85\u00a0wt% and from 3.30 to 7.15\u00a0wt%. However, the solid yield decreased from 30.94 to 24.00\u00a0wt%. A ratio of 4Co-1Fe seems to be optimal in this study, based on the HHV (34.16\u00a0MJ/Kg) and yield (68.85%) of bio-oil. In the presence of mono- or bi-metallic catalyst, the DOD of the upgraded bio-oils ranged from 44.61 to 79.81%.According to the analysis in the composition of gas products, the gas selectivities from Co/AC and Fe/AC were quite different. In particular, when using Co/AC catalysts, CO, CO2, and CH4 were mainly produced, whereas CO production was significantly reduced with Fe/AC catalysts. With increasing Co/AC ratio in bimetallic catalysts from 1 to 4, the CO and CH4 contents in the gas products increased, while the CO2 decreased. When Fe-based catalysts were used, the moisture content in the upgraded bio-oil was 32.96\u201334.21%, which is higher than that of Co-based catalysts (25.86 to 26.90%). The higher moisture contents in the use of Fe-based catalysts can be explained by dehydration as well as the gas-phase reactions such as CO/CO2 methanation and reverse water\u2013gas shift reactions presented below [15,47].\n\n(2)\n\n\n\n\nC\nO\n\n2\n\n+\n\nH\n2\n\n\u2194\nC\nO\n+\n\nH\n2\n\nO\n\n\n\n\n\n\n(3)\n\n\nC\nO\n+\n\n\n3\nH\n\n2\n\n\u2194\n\n\nC\nH\n\n4\n\n+\n\nH\n2\n\nO\n\n\n\n\n\n\n(4)\n\n\n\n\nC\nO\n\n2\n\n+\n4\n\nH\n2\n\n\u2194\n\n\nC\nH\n\n4\n\n+\n\n\n2\nH\n\n2\n\nO\n\n\n\n\n\nTable 6\n illustrates the liquid compositions of upgraded bio-oil obtained using different AC-supported catalysts at 350\u00a0\u00b0C and 60\u00a0bar. It is well known that phenol and benzenediol derivatives are favorable as liquid components in bio-oils due to their high stability and HHVs. From this context, the 20\u00a0wt% 4Co-1Fe/AC seems to be the most desirable catalyst among the tested catalysts. Table 6 clearly shows that the production of the components mentioned above improved significantly with 20\u00a0wt% 4Co-1Fe/AC. The area% of phenol, methyl phenol, and benzenediol derivatives were calculated to be 9.58, 26.40, and 8.95%, respectively. Besides, the area% of naphthalene derivatives were enhanced with 20\u00a0wt% 4Co-1Fe/AC. The Retro-Diels\u2013Alder reaction might proceed during the HDO process. These trends have been reported in the ex-situ catalytic upgrading of bio-oil vapors over HZSM-5 and microwave-assisted pyrolysis of waste olefins [48\u201350].\nFig. 6\n shows the carbon number distribution of raw bio-oil and upgraded bio-oil using different AC-supported catalysts at 350\u00a0\u00b0C and 60\u00a0bar. For the raw bio-oil, for example, the light (C5-C11), diesel (C12-C18), and heavy (C19-C38) fractions were 49.99, 32.30, and 17.66\u00a0wt%. The carbon numbers of upgraded bio-oil were mainly distributed in C5\u2013C11 fractions, which were calculated to be 46.79, 46.17, and 50.66\u00a0wt% for 20\u00a0wt% Co/AC, 20\u00a0wt% Fe/AC, and 20\u00a0wt% 4Co-1Fe/AC, respectively. With 20\u00a0wt% 4Co-1Fe/AC, the highest proportion of light fraction (C5-C11) was achieved, especially with the C8 component (20.40\u00a0wt%). The C8 component of the light fraction might belong to the methyl phenol derivatives, one of the most favorable components of bio-oils. Methyl phenol derivative is an important chemical intermediate and is essential to produce various chemicals and materials, such as phenolic resins, alkylphenols, etc. However, most of the methyl phenol derivatives are currently produced from benzene through cumene process, which consumes a large amount of fossil fuels and causes environmental pollution [33]. In this study, the highest proportion of light fraction (C5-C11) was obtained with 20\u00a0wt% 4Co-1Fe/AC, suggesting that this catalyst is most promising for upgrading WPS bio-oils.Using the NMR spectroscopy, functional groups in bio-oils were identified [8,12] based on the chemical shift regions. As shown in Fig. 7\n, the 13C NMR spectra of bio-oils were divided into several chemical shifts regions, 0\u201355, 55\u201395, 95\u2013165, and around 200\u00a0ppm. The raw bio-oil mainly consisted of aromatic compounds, aliphatic hydrocarbons, and a small amount of carbohydrates, alcohols, esters, and phenolic methoxy groups. The upgraded bio-oil with Co/AC and 4Co-1Fe/AC produced high intensity of aromatic compounds, CO groups, and aliphatic hydrocarbons, whereas Fe/AC led to generate carbohydrates, alcohols, esters, and phenolic methoxy groups. Generally, aldehydes and ketones are not found in 13C NMR spectra. Relatively high intensity of aromatic compound groups observed at 90\u2013165\u00a0ppm might be due to cleavage of the hydroxy and methoxy groups via dehydration and demethoxylation, leading to lower oxygen contents. It was found that aromatic compounds and carboxylic acids of the bio-oil with Fe/AC were less than those for Co/AC and Co-Fe/AC catalysts. This difference can be attributed to the formation of CO2 via decarboxylation, which occurs in the presence of iron oxides. It is in good agreement with the GC/MS analysis and the results from our prior work, where iron oxide catalysts were employed in the pyrolysis of spent coffee waste for upgrading sustainable bio-oil in a bubbling fluidized-bed reactor [12]. In this study, the 2D NMR HSQC spectra also was collected (Figure S4) to determine the proton-carbon single bond correlation and predict small molecular components in the bio-oil. A strong signal attributed to the aromatic carbon was observed at around 7.5\u00a0ppm, and signals at the region of 7.08-6.59\u00a0ppm appear to be contributed by the benzene ring. The chemical shift by hydroxy and ketone groups can be seen at ~2\u00a0ppm, while the peak at ~3.77\u00a0ppm shifted downfield represents the methoxy group. These observations are also consistent with the 13C NMR results in Fig. 7.Catalyst deactivation is primarily caused by the carbonaceous formation on internal and/or external surfaces [51\u201356]. Furthermore, the reduction in catalytic activity can be caused by blocking the active sites of catalysts when intermediate products are adsorbed onto the catalyst surfaces during the HDO process [52,54,56].In this study, the deactivation of catalysts was investigated by characterizing the fresh and spent catalysts. After HDO, as shown in Table 2, the specific surface area of Co/AC, Fe/AC, and 4Co-1Fe/AC catalysts significantly decreased to 7.1634, 12.3916, and 8.1094\u00a0m2/g from 11.1108, 93.2195, and 117.1773\u00a0m2/g. In addition, the pore volume was reduced to 0.0066, 0.0130, and 0.007\u00a0cm3/g. It is likely that the formation of coke in catalysts reduced both the specific surface area and pore volume.Thermogravimetric analysis was utilized to quantify the coke formation by measuring the difference in weight loss between fresh and spent catalysts (Fig. 8\n). The oven temperature was increased from 20 to 700\u00a0\u00b0C at a heating rate 10\u00a0\u00b0C/min under the air atmosphere. The weight loss differences for Co/AC, Fe/AC, and 4Co-1Fe/AC catalysts were 35,1, 17.5, and 46.0\u00a0wt% at 700\u00a0\u00b0C, respectively. The weight loss curve by differential thermalgravimetric (DTG) analysis was also prepared to determine the combustion temperature of the AC support in the catalyst, as shown in Fig. 8 (b). The DTG peaks of fresh catalysts confirm that the AC is burned out in the range of 380\u2013450\u00a0\u00b0C. Hou et al. [51] reported that the weight loss of the spent HZSM-5 catalyst in the range of 400\u2013700\u00a0\u00b0C is due to the carbon combustion. Except for the peaks in the combustion temperature range of AC, the spent catalysts showed additional DTG peaks at different temperature ranges. For the spent Fe/AC and Co-Fe/AC, the peaks were observed in the range of 225\u2013266\u00a0\u00b0C and 330\u2013500\u00a0\u00b0C, while Co/AC was at a relatively higher temperature (524\u00a0\u00b0C). The mass loss before 350\u2009\u2103 is mainly ascribed to the volatilization of the non-desorbed products, while other losses are due to the combustion of coke [52].Depending on the location and type of coke, its combustion temperature varies. For example, the internal coke is burned at a higher temperature than the external coke [52]. In addition, weakly bonded cokes such as linked-ring compounds are more likely to burn at lower temperatures than strongly adsorbed cokes such as condensed-ring compounds [54]. Besides, Figure S3 shows the shape and type of carbonaceous formation on the catalyst surface. Based on our thermogravimetric analysis, the carbonaceous compounds that caused the deactivation were formed on the external catalyst surface. Aside from the carbonaceous materials, the formation of spinel CoFe2O4 during HDO, which was confirmed by XRD analysis, might lead to the deactivation of catalysts [27].Pyrolysis of biomass is a promising process for producing bio-oils and valuable chemicals. Unfortunately, bio-oils produced by pyrolysis are corrosive and chemically unstable due to high water and oxygen contents. Catalytic hydrodeoxygenation (HDO) is one of the most effective technologies to improve the quality of bio-oils. In this study, HDO of bio-oils produced by pyrolysis of wood pallet sawdust (WPS) was systematically investigated using mono- (Co or Fe) and bi-metallic (Co-Fe) catalysts supported on activated carbon (AC). Based on the overall results, 20\u00a0wt% 4Co-1Fe/AC catalyst was found to be optimal among the tested catalysts. Using this catalyst, high HHV (34.16\u00a0MJ/kg) and liquid yield (68.85%) were obtained at 350\u00a0\u00b0C and 60\u00a0bar. Furthermore, the production of desirable components such as phenol derivatives was improved. Finally, with 20\u00a0wt% 4Co-1Fe/AC, the highest proportion of light fraction (C5-C11) was achieved, especially with the C8 component (20.40\u00a0wt%). The catalysts were deactivated by the formation of carbonaceous compounds on the external surface, oxidation of metal species, and blocking of active sites on catalysts. Our results suggest that the upgraded bio-oils by HDO can be used as transportation fuels or great sources for alternative fuels and valuable chemicals. Along with the results presented here, the feasibility studies, including techno-economic analysis and energy balance for the entire pyrolysis-HDO process, need to be carried out in the future.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the (NRF) grant funded by the Korea government(MSIT) (No. 2020R1A2B5B01097547). This work was supported by the Engineering Research Center of Excellence Program of the Korea Ministry of Science, ICT & Future Planning (MSIP)/National Research Foundation of Korea (NRF) (Grant NRF-2021R1A5A6002853).", "descript": "\n                  The catalytic hydrodeoxygenation (HDO) processes for upgrading pyrolysis bio-oils from wood pallet sawdust (WPS) were studied using activated carbon (AC) as a support of mono- (Co/AC and Fe/AC) and bi-metallic (Co-Fe/AC) catalysts. The effects of the reaction temperature and hydrogen pressure on products and high heating value (HHV) were systematically investigated. At 350\u00a0\u00b0C and 60\u00a0bar, 20\u00a0wt% Co/AC showed the highest liquid yield (70.46\u00a0wt%) along with HHV of 34.22\u00a0MJ/Kg. Among the tested bimetallic catalysts, comparable liquid yield (68.85\u00a0wt%) and HHV (34.16\u00a0MJ/kg) were achieved with 20\u00a0wt% 4Co-1Fe/AC catalyst. Methyl phenol derivatives were found to be the main component in upgraded bio-oil. The carbon number of the upgraded bio-oil was mainly distributed in C5\u2013C11 fraction, especially with the C8 component (20.40\u00a0wt%). The catalysts were deactivated by the formation of carbonaceous compounds on the external surface, oxidation of metal species, and blocking of active sites on catalysts.\n               "}
{"full_text": "The oxygen evolution reaction (OER) is pivotal because of its multiple prospective energy storage devices, including water electrolyzers and rechargeable metal-air batteries;\n1\u20134\n however, the OER pathways are complex and generally impose significant kinetic bottlenecks with large overpotentials (usually more than 350\u00a0mV).\n5\n The benchmark Ru/Ir catalysts can hardly satisfy scaled up practical utilization because of their scarcity, prohibitive cost, and detrimental environmental effects.\n6\n Nowadays, earth-abundant transition-metal-based compounds have comparable performance and, thus, have attracted extensive attention as alternatives.\n7\u20139\n However, most of these compounds undergo surface reconstruction under operating conditions, which makes it difficult to capture the dynamic structure and recognize real catalyst information.\n10\n For example, Ni-based OER electrocatalysts,\n11\n\n,\n\n12\n such as chalcogenides, nitrides, and phosphides, are thermodynamically less stable than oxides in strongly oxidative environments. The electro-derived oxidation on their surface induces reconstruction to form target OER-active oxides/(oxy)hydroxides (Table S1), which have been identified as real active species.\n13\n These post-OER catalysts generally show a near-surface reconstruction structure at the nanoscale, such as a core-shell structure, containing a large percentage of inactive atoms in the core.\n14\n Because of their limited near-surface reaction region,\n15\n\n,\n\n16\n the evolving catalysts usually exhibit incompletely developed catalytic activity. The compositional complexity of partially reconstructed catalysts greatly hinders an insightful understanding of catalytic origins.\n17\n Therefore, it raises curiosity and challenges in fundamental mechanistic research of the reconstruction chemistry of pre-catalysts, including origins of limited reconstruction degrees and dynamic reconstruction mechanisms.The reconstruction results of reconstruction layer thickness (RLt) and the smallest size in one dimension based on reported OER pre-catalysts are summarized in Figure\u00a01\n (see also Table S2). The RLt values for most reported pre-catalysts are less than 10\u00a0nm. Deep/complete reconstruction will maximize the number of active sites in reconstructed catalysts and thus endow high-mass-activity catalysis. To date, very few studies have focused on completely reconstructed (CR) catalysts.\n18\u201320\n Because of the fast reconstruction processes, deep comprehension of reconstruction is crucial but difficult. Despite its importance, the general synthesis method and underlying mechanism of CR catalysts have not been reported. More importantly, most current reports investigate catalyst performance in low-concentration alkali (0.1\u20131\u00a0M KOH) rather than in industrial-concentration 20\u201330 wt % KOH.\n18\u201321\n Harsh conditions may result in different reconstruction results for pre-catalysts. Evaluating catalysts in 20\u201330 wt % KOH for alkaline water electrolysis (AWE) can promote their commercial application.\n22\n Therefore, it is meaningful for performance evaluation under realistic operating conditions.As illustrated in Schemes 1A\u20131C, we present etching-leaching-reconstruction engineering to achieve complete reconstruction of bulk hydrate pre-catalysts. As a proof of concept, various in/ex situ technologies were employed to capture the time-resolved structural/phase evolution of NiMoO4\u00b7xH2O under electro-oxidation conditions. The reconstruction steps generally consist of complete collapse of hydrates with co-leaching of crystal water and MoO4\n2\u2212 and then its reconstruction to NiOOH via electro-oxidation. Intrinsically, the loose reconstruction layer caused by co-leaching is the key to promote deep penetration of solution and, thus, complete reconstruction. However, for NiMoO4, the entire crystal structure could not be disintegrated by alkali etching, and the formed dense layer prevents electrolyte penetration for further etching and/or reconstruction. Therefore, the formed, partially reconstructed NiMoO4@NiOOH features an inert core part because of near-surface catalysis properties (Schemes 1D\u20131F). More importantly, the unique interconnected structure of CR catalysts endows them with ultrastable catalysis and potential AWE applications under realistic conditions.NiMoO4\u00b7xH2O nanowires (NWs) grown on conductive nickel foam (NF) were fabricated using a low-temperature hydrothermal method that has demonstrated manufacturing amplification capability (>250\u00a0cm2 in one-pot synthesis) toward commercial applications (Figure\u00a0S1). These nanowires are monocrystalline with a smooth surface, confirmed by transmission electron microscopy (TEM) imaging and the corresponding selected area electron diffraction (SAED) pattern (Figure\u00a0S2). Its crystal structure was determined via atomic substitution and structure optimization because the powder X-ray diffraction (XRD) pattern is similar to that of the reported analog CoMoO4\u00b70.75H2O.\n23\n The loss of lattice water (LW) does not induce collapse of the NiMoO4\u00b7xH2O framework, and the obvious phase evolution occurs when the calcination temperature is higher than 400\u00b0C, which is attributed to loss of coordination water (CW) (Figure\u00a0S3).Nickel foam cannot maintain its flexibility and toughness for direct non-binder catalysis when heated to more than 600\u00b0C. Therefore, a proper calcination temperature of 550\u00b0C was chosen to fabricate the anhydrous NiMoO4 nanowire arrays.Cyclic voltammetry (CV) activation was conducted at 0.924\u20131.724\u00a0V versus reversible hydrogen electrode (VRHE), resulting in two different geometric/phase structures of NiMoO4\u00b7xH2O and NiMoO4. For NiMoO4\u00b7xH2O, CR-NiOOH forms in a stable state during OER (Figure\u00a0S4). High-angle annular dark-field scanning TEM (HAADF-STEM) imaging clearly demonstrates its morphological characterization, which is represented schematically in an inset in Figure\u00a02\nA. Such a nanowire is interconnected by sub-5-nm ultrasmall nanoparticles (NPs), resulting in visible interspaces with \u223c5-nm nanopores accessible to electrolytes. All interplanar spacings of NPs within the nanowire are well indexed to the planes of NiOOH (Figure\u00a02B). The exposed interplanar spacings of 0.158, 0.208, 0.213, 0.240, and 0.248\u00a0nm can be well assigned to the (220), (210), (111), (011), and (101) planes of orthorhombic NiOOH (Joint Committee on Powder Diffraction Standards [JCPDS] 27-956), respectively. It displays low-crystalline and polycrystalline characteristics confirmed by the SAED pattern (Figure\u00a0S4D). Only two Raman peaks at 474 and 554\u00a0cm\u22121 belonging to the eg bending and the A1g stretching vibration of Ni-O in NiOOH\n24\n are observed (Figure\u00a0S4E), and the O/Ni atomic ratio is 2.05 from the corresponding energy-dispersive X-ray (EDX) spectroscopy spectrum (Figure\u00a02C), which further demonstrates the pure phase of CR-NiOOH. In addition, negligible content (0.1 atomic percentage [at.%]) of the Mo element suggests its absence within the nanowire. The color of CR-NiOOH is black (inset in Figure\u00a02C), which is the typical color of nickel (oxy)hydroxide. Tomographic data were further analyzed to show its three-dimensional structure from multiple perspectives (Figures 2D\u20132G; Video S1). Electron tomography was conducted at consecutive rotational angles from \u221260\u00b0 to 48\u00b0, and HAADF-STEM images were collected simultaneously. High homogeneity of ultrasmall NPs and no agglomerated large particles are observed from various\u00a0rotational angles. Particularly, CR catalysts that featured an ultrasmall NP-interconnected structure with evenly distributed gas-permeable pores were reported first.\n\n\nVideo S1. Electron Tomography Video of a Single CR-NiOOH Nanowire\n\n\n\nDifferent from NiMoO4\u00b7xH2O, NiMoO4 undergoes surface reconstruction occurs after 1-day continuous electro-oxidation, which results in core-shell NiMoO4@NiOOH nanowires. Only an \u223c7-nm-thick NiOOH layer forms, and the inner continuous lattice fringes covering dozens of nanometers are indexed to the planes of NiMoO4 (JCPDS 86-0361) (Figure\u00a02H; Figures S5A and S5B). STEM element mapping further confirms the core-shell structure as undetected Mo signals in the shell region (Figure\u00a02I). No diffraction signals are assigned to NiOOH from the SAED pattern (Figure\u00a0S5C), and Raman spectra of NiMoO4 before and after activation remain almost unchanged (Figure\u00a0S5D). The undetected Raman/SAED signals of NiOOH are attributed to its ultrathin layer structure. The reported Ni-Mo nitride after activation also occurred the partial reconstruction just like NiMoO4.\n25\n\n\nEx situ high-resolution TEM (HRTEM) characterizations during potential-controlled CV measurements were carried out to uncover the morphological evolution of NiMoO4\u00b7xH2O. An \u223c2.66-fold current density at 1.724 VRHE is achieved when comparing the 20th cycle with the initial one, with the area of the closed curves increasing gradually (Figure\u00a03\nA). Only 20 cycles with 640-s duration are required to achieve complete reconstruction, indicating a fast reconstruction rate. The redox peak currents are gradually stabilized in the 15th\u201320th cycles (Figure\u00a0S6), which indicates that the pre-catalyst reached the steady state. Ex situ electrochemical impedance spectroscopy (EIS) results from the same electrode show that the charge transport resistance (Rct) decreased significantly from 34 to 6.5 ohm (\u03a9), indicating faster charge transfer of CR-NiOOH (Figure\u00a03B). To visualize the dynamic reconstruction process, the microstructures of intermediates at different stages were analyzed (Figures 3C\u20133F). After the first CV to 1.23 VRHE, which is the equilibrium potential for OER, the surface of monocrystalline NiMoO4\u00b7xH2O becomes amorphous (see Figure\u00a0S7 for detailed characterization). After an anodic scan to 1.60 VRHE, where the evolution of O2 happens, the distinct three-layer region appears (Figure\u00a0S8A). The outermost layer consists of ultrasmall low-crystalline NPs with an amorphous transitional interlayer and innermost NiMoO4\u00b7xH2O layer. Because of the oxygen-evolving process, the surface Ni species are oxidized to OER-stable high-valence species. The generated (oxy)hydroxide is transformed in situ from surface species rather than by a dissolution-deposition process (Figure\u00a0S9). Elemental distributions and content analyses depict the uniform distribution of Ni and O elements, whereas Mo shows a gradient distribution and decreases gradually from inner to outer (Figures S8D\u2013S8F). When back at 0.924 VRHE to achieve a complete CV cycle, the reconstruction degree deepens (Figure\u00a0S10). A rough region with \u223c50-nm thickness is clearly visible in TEM images. Benefiting from the continuous co-leaching of crystal water and Mo species, NiMoO4\u00b7xH2O is completely reconstructed only after 20-cycle CV. Based on the above results, the geometric/phase evolution is illustrated in Figures 3G\u20133J. Furthermore, the reconstruction mechanism is also proposed from the point of view of the crystal structure (Figure\u00a03K). The reconstruction processes include bond breakage, co-leaching of crystal water and Mo species, and OH\u2212 contact and electro-oxidation, which will be discussed further below.\nIn/ex situ technologies were further utilized to gain insights into the reconstruction mechanism of NiMoO4\u00b7xH2O. With the ex situ XRD measurement, disappearance of the representative peaks at \u223c27.3\u00b0 and 29.8\u00b0, which belong to NiMoO4\u00b7xH2O, demonstrates its structural crack and amorphization (Figure\u00a0S11A). An in situ electrochemistry-Raman coupling system was applied to understand electrocatalytic reactions in liquid electrolytes because of the high molecular specificity and non-interference of water of Raman signals.\n26\n\n,\n\n27\n At potentials below 1.324 VRHE, the peak intensities of NiMoO4\u00b7xH2O decrease with increased potential, indicating gradual destruction of its crystal structure (Figure\u00a04\nA). At 1.424 VRHE, two well-defined bands at 474 and 554\u00a0cm\u22121 appear that belong to NiOOH, and such a potential is attributed to the Ni(II)/Ni(III) oxidation peak (\u223c1.37 VRHE).\n28\n Here, the oxide phase forms below 1.424 VRHE, which will be discussed later. The Raman peaks for NiOOH are kept as the applied bias voltage increases, indicating that it serves as an OER-stable catalytic species. Furthermore, the new peak at 900\u00a0cm\u22121 is assigned to MoO4\n2\u2212 in alkaline solution,\n29\n which originates from dissolution of Mo species. Contrary to NiMoO4\u00b7xH2O, NiMoO4 shows unchanged Raman peaks under the same test conditions (Figure\u00a04B). The undetected Raman peaks of NiOOH are attributed to its thin layer on the NiMoO4 surface, as shown in Figure\u00a02H.It should be noted that the reconstruction process for NiMoO4\u00b7xH2O is complete and exhaustive rather than forming core-shell NiMoO4\u00b7xH2O@NiOOH as the final product. Generally, the reconstruction process is very common for the reported Ni-based OER catalysts. However, such a process is partial in these compounds, and only a thin layer of NiOOH/Ni(OH)2 forms on their surface. Here the origins of the complete reconstruction of NiMoO4\u00b7xH2O are analyzed. After soaking in 1\u00a0M KOH, NiMoO4\u00b7xH2O was gradually etched by alkali. As shown in Figure\u00a04C, the intensity of three Raman peaks at 800\u20131,000\u00a0cm\u22121 assigned to the Mo-O-Ni stretching vibration\n30\n decreases and almost disappears after soaking for 1 h, suggesting breakage of the Mo-O-Ni bond. The newly formed peak at 900\u00a0cm\u22121 is assigned to MoO4\n2\u2212. The peak at 355\u00a0cm\u22121 assigned to MoO4 vibration shifts to a lower wave number after soaking for 800 s, which may be associated with its vibrational environment. The Mo species are dissolved during reconstruction of NiMoO4\u00b7xH2O (Figures S12A\u2013S12C), and such a phenomenon also happens to the Ni-Mo nitride OER electrocatalyst reported by Yin et\u00a0al..\n25\n The NiMoO4\u00b7xH2O nanowire evolves to a nanosheet-assembled nanowire morphology after alkali etching (Figure\u00a0S12D), and the etching reaction is further demonstrated by ex situ XRD patterns (Figure\u00a04D). As a result, the washed product after soaking is Ni(OH)2, whereas the K2MoO4 phase is detected for the product without washing. Therefore, when NiMoO4\u00b7xH2O serves as pre-catalyst measured in 1\u00a0M KOH, the spontaneous etching reaction happens simultaneously. The multicomponent co-leaching results in a loose reconstruction layer and triggers its complete reconstruction.As discussed above, the several-nanometer-thick reconstruction layer is observed for the NiMoO4 pre-catalyst. Because the reconstruction reaction involves interaction with an alkaline solution, the limited reconstruction depth could be attributed to limited electrolyte penetration. A high-magnification STEM image of NiMoO4 after soaking in 1\u00a0M KOH verifies our speculation because it shows the dense etching layer of \u223c3\u00a0nm (Figure\u00a04E). However, for NiMoO4\u00b7xH2O, the fast etching rate makes the surface loose (Figure\u00a04F), which facilitates electrolyte penetration for further etching. Therefore, we guess that the intrinsic properties of materials etched by alkali, which results in different etching structures, either loose or dense, are responsible for the two different results above. Even for the microns of Fe-doped cobalt molybdate hydrate, the loose etching structure enables its complete etching (Figure\u00a0S13). It is not only difficult for alkaline solution to pass through the dense surface layer but also difficult for solution to pass through the crystal structure of the surface layer. Density functional theory (DFT) calculations confirm this difficulty by showing the high-energy barrier of 2.2 electron volt (eV) for OH\u2212 to pass through the NiOOH layer (Figure\u00a04G). Besides, for reported pre-catalysts such as Ni2P\n11\n and Co4N,\n14\n their post-OER products show the dense reconstruction layer, which supports our points.CR catalysts feature an ultrasmall NPs-interconnected structure with evenly distributed gas-permeable pores, and the key is that etching and electro-oxidation reconstruction happen simultaneously. If NiMoO4\u00b7xH2O is soaked in 1\u00a0M KOH prior to electro-oxidation, CR-NiOOH with a nanosheet-assembled nanowire structure can be obtained (Figures S14A and S4B, denoted CR-NiOOH\u2217). This is because the etching reaction induces formation of monocrystalline/high-crystalline Ni(OH)2 nanosheets (Figures S14C and S14D), which further evolve to (oxy)hydroxide during the subsequent electro-oxidation. Because the smaller-size catalyst is endowed with\u00a0more exposed active sites, CR-NiOOH shows much better OER catalysis (Figures S14E and S14F). Ni(OH)2 nanosheet arrays grown on nickel foam were also prepared but with poor OER activity (Figure\u00a0S15), suggesting the effectiveness of structure engineering for better catalysis. To explain the unique structure of CR-NiOOH, identifying the formed amorphous intermediates shown in Figure\u00a03D is important. The voltage range below the theoretical decomposition voltage is analyzed, which is helpful for understanding the effects of voltage bias during alkali etching. After CV at 0.924\u20131.224 VRHE, the nanowires mainly consist of \u223c5-nm polycrystalline NiO NPs (Figure\u00a0S16F). Furthermore, the Raman peak at 460\u00a0cm\u20131 is assigned to the Ni-O stretching mode of NiO, and other peaks may be assigned to molybdenum oxides, indicating phase separation and formation of amorphous oxide intermediates (Figures S16G and S16H). These results suggest that the simultaneous etching-reconstruction processes facilitate formation of an ultrasmall NP-interconnected structure with nanopores.The prerequisites for forming ultrasmall NP-interconnected CR catalysts can be summarized as follows: (1) achieving complete collapse of bulk materials, (2) promoting deep penetration of electrolytes for inner electro-oxidation as the loose reconstruction layer dominates, and (3) simultaneous reconstruction and etching. To demonstrate the universality of these results, other bulk alkali-sensitive pre-catalysts, such as NiMoO4\u00b7xH2O nanosheets, Ni-BTC (BTC\u00a0= 1,3,5-benzene tricarboxylate) metal organic framework microspheres, CoMoO4\u00b70.75H2O nanowires, and Co(CO3)0.5(OH)\u00b70.11H2O nanowires, were also investigated (Figure\u00a0S17). All of them can completely evolve to their corresponding hydroxides after alkali etching, which guarantees complete reconstruction of CR catalysts. Consequently, the CR catalysts of the (oxy)hydroxide phase are obtained with the ultrasmall NP-interconnected multilevel structure. Therefore, the complete reconstruction mechanism can be widely extended to various bulk alkali-sensitive pre-catalysts.Using a standard three-electrode system, NiMoO4\u00b7xH2O NWs/NF was directly employed as a binder-free working electrode to achieve activation and acquire the (oxy)hydroxide arrays. Before performance evaluation, the purity of KOH solution was examined. KOH reagent purity (Fe content of\u00a0< 0.001%) and undetected Fe 2p XPS signals using an Mg source confirmed the absence of Fe impurities in the solution used.\n19\n CV curves of NiMoO4\u00b7xH2O ink coated on the carbon cloth for the initial cycles are provided in Figure\u00a05\nA. The small shift (\u223c20\u00a0mV) of the Ni(II)/Ni(III) oxidation peak is much below 50\u00a0mV, negating the influence of iron according to Klaus et\u00a0al.\n31\n The chronopotentiometric response of fresh NiMoO4\u00b7xH2O shows a gradually decreased potential, but the potentials of NiMoO4 are almost unchanged, indicating fast reconstruction on the NiMoO4\u00b7xH2O surface during activation (Figure\u00a0S18A). The simplex nickel foam shows the increased potentials, implying that the activity enhancement of NiMoO4\u00b7xH2O is independent of the substrate. After activation, the obtained CR-NiOOH was tested in 1\u00a0M Fe-free KOH, whereas surface-reconstructed NiMoO4@NiOOH and commercial IrO2/C served as control samples. Linear sweep voltammetry (LSV) curves were normalized by geometric area, electrochemically active surface area (ECSA), and catalyst mass, respectively (Figures S18B\u2013S18D). CR-NiOOH has much higher OER activity than NiMoO4@NiOOH. For example, CR-NiOOH requires the lowest overpotential at 10 mA cm\u22122 (\u03b710 of 278.2\u00a0mV), which is much lower than that of NiMoO4@NiOOH (353.6\u00a0mV). To reveal the advantages of the CR catalyst for the OER, the mass activity-related overpotential \u03b710, m (\u03b710, m is calculated as the ratio of \u03b710 and mass of the loading catalysts) is compared. The \u03b710, m of CR-NiOOH (289\u00a0mV mg\u22121) is even lower than that of the commercial IrO2/C (325.9\u00a0mV mg\u22121), indicating that it can serve as a superior catalyst. However, the \u03b710, m value of NiMoO4@NiOOH is as high as 393.2\u00a0mV mg\u22121. In addition to a high-mass-activity OER, the CR catalyst also possesses the advantage of 97.4% Faradaic efficiency (Figure\u00a0S18E). The 2.6% loss may be due to the dissolved gas in the solution and gas adsorbed on the electrode.\n32\n\nThe reasons why CR-NiOOH is superior to NiMoO4@NiOOH were analyzed. First, CR-NiOOH possesses more OER-active species, whereas an \u223c7-nm-thin layer of NiOOH on NiMoO4 serves as a catalytic species; thus, the former can achieve a higher mass activity toward surface-catalyzed reactions. Our as-prepared NiOOH is reconstructed completely and in situ from NiMoO4\u00b7xH2O, and the whole catalyst could perform the catalytic reactions with abundant catalytic sites. The higher intensity of the oxidation peak at \u223c1.37\u00a0V for CR-NiOOH (Figures S18B\u2013S18D) suggests larger amounts of active phases as well.\n21\n Second, CR-NiOOH possesses a sub-5-nm NP-interconnected structure. This unique multilevel structure is endowed with numerous pores accessible to electrolytes and conductive to gas diffusion. Although NiOOH has been shown to not be very active in the OER,\n31\n the low-crystalline characteristics and abundant defects in catalysts have been reported to accelerate the OER kinetics.\n33\n Therefore, the newly developed CR-NiOOH shows great potential as an IrO2-substituted oxygen evolving system. More importantly, CR-NiOOH exhibits a negligible change in potential after 1,350\u00a0h in the durability test, which demonstrates its potential for ultrastable electrolytic applications (Figure\u00a05B). Electron microscope characterization after the durability test displayed an unchanged morphology and retained microstructure, suggesting a robust and stable nature of NiOOH in the face of corrosion (Figure\u00a0S19). Its excellent stability is mainly attributed to the robust (oxy)hydroxide NP-interconnected structure with evenly distributed gas-permeable pores. To show the multiple applications of CR-NiOOH, the urea oxidation reaction (UOR) was also measured, which is essential for urea electrolysis and also has sluggish kinetics.\n34\n As a result, it shows decreased overpotential of 106\u00a0mV at 10 mA mg\u22121 compared with that of NiMoO4@NiOOH and good durability for 110\u00a0h at 0.48 VHg/HgO (Figure\u00a0S20). In addition, CR-NiOOH can also provide stable OER catalysis at a high temperature of 52.4\u00b0C for 120 h, with an overpotential increase of only \u223c10\u00a0mV (Figure\u00a0S21).Fe impurities in the testing solution greatly enhance the OER activity of NiOOH,\n35\n which is attributed to formation of highly active nickel-iron (oxy)hydroxide. Here, the advanced iron-incorporated nickel (oxy)hydroxide (denoted Fe-NiOOH) was fabricated in situ in 1\u00a0M KOH solution containing a trace of Fe by adding an iron source. To achieve 10\u00a0mg cm\u22122, a small overpotential of 248\u00a0mV is required (inset in Figure\u00a05C). The enhanced OER catalysis is attributed to its electronic tuning by Fe incorporation and enhanced electron conductivity. During a 20-day long-term chronopotentiometry measurement of Fe-NiOOH, the potential is almost unchanged, with a small potential change of only 10\u00a0mV (Figure\u00a05C). These results confirm the well-retained ultrastable property after optimizing its electronic structure. To demonstrate the water electrolysis application, our reported heterostructured MoO2-Ni NWs/NF were chosen as hydrogen evolution reaction (HER) electrodes.\n36\n As shown in Figure\u00a05D, the MoO2-Ni NWs/NF featured by interface catalysis exhibits excellent HER activity with a small decay of 0.155\u00a0mV h\u22121. When pairing Fe-NiOOH with MoO2-Ni arrays in a two-electrode alkaline water electrolyzer, the electrolyzer delivers 10 mA cm\u22122 at 1.48\u00a0V for over 580\u00a0h under fast-moving fluid condition produced by rapid stirring of fresh magneton (Figure\u00a05E). Its electrolysis durability is superior to that in previous reports (Figure\u00a05F; Table S3 ). These results highlight the potential of CR catalysts for ultrastable and high-efficiency\u00a0catalytic applications.Most industrial alkaline electrolyzers are operated in a strong alkaline KOH solution (>20 wt %; Figure\u00a05G). Therefore, we evaluated the half-reaction catalysis and AWE performance based on the abovementioned array system in a two-electrode cell in 30 wt % KOH, as illustrated schematically in Figure\u00a05H. Under such harsh operating conditions, the Fe-NiOOH anode still performs stable OER catalysis at \u223c0.5 VHg/HgO for over 210\u00a0h with activity decay of 0.075\u00a0mV h\u22121 (Figure\u00a05I). Fe-NiOOH maintains its structural and component characterization after testing under such harsh conditions (Figure\u00a0S22). For the cathodic MoO2-Ni array, it can catalyze HER for 300\u00a0h with activity decay of 0.21\u00a0mV h\u22121 (Figure\u00a0S23). Here, the 10\u00a0\u03bcL solution containing 0.6\u00a0mg Fe(NO3)3\u00b79H2O was also added to ensure the same test environment as for OER testing. As expected, the Fe-NiOOH//MoO2-Ni array system operated for 260\u00a0h (Figure\u00a05J), indicating its potential practical applications. Performance evaluation of reported catalysts in industrial-concentration alkali has also been provided for comparison (Table S4).For the NiMoO4 pre-catalyst, the alkali etching rate on its surface is slow in 1\u00a0M KOH, and a dense reconstruction layer forms via surface reconstruction. This leads to quick termination of reconstruction. However, under harsh conditions of 30 wt % KOH, alkali etching is accelerated, and leaching of Mo species is promoted. This could result in a porous structure of the reconstruction layer and promote deep reconstruction in concurrent electro-oxidation processes. Therefore, NiMoO4 can also be completely reconstructed to (oxy)hydroxide in 30 wt % KOH, which is reflected by decreased potentials because of the enhanced number of active species (Figure\u00a0S24). This result suggests that some pre-catalysts (such as phosphides, nitrides, and chalcogenides) may also be completely reconstructed to stable catalytic species in industrial alkali. Therefore, understanding reconstruction chemistry and evaluating performance under realistic conditions are necessary and meaningful, especially for pre-catalysts involved in reconstruction.In summary, we discovered different reconstruction results for hydrate/anhydrous molybdate pre-catalysts at oxidized potentials in 1\u00a0M KOH; i.e., complete/surface reconstruction. Such a difference depends on the microstructure characteristics (dense or loose) of the reconstructed layer, caused by different leaching species from pre-catalysts. The proposed reconstruction mechanism can be extended to other bulk alkali-sensitive pre-catalysts, resulting in a series of electrochemically formed CR catalysts. These CR catalysts display a unique structure interconnected by ultrasmall NPs endowed with abundant defects and pores accessible to electrolytes. Such an interconnected structure allows CR-NiOOH to perform ultrastable catalysis for 1,350 h. After iron incorporation, the obtained Fe-NiOOH exhibits remained structure and ultrastable catalysis. The coupled Fe-NiOOH and MoO2-Ni system was confirmed with excellent water electrolysis performance in 1\u00a0M and 30 wt % KOH. Furthermore, different reconstruction results of anhydrous NiMoO4 in industrial alkali were obtained, suggesting the importance of evaluating catalysts under realistic conditions. This work highlights fundamental reconstruction chemistry, CR catalysts with a unique structure and ultrastable catalytic properties, and different reconstruction phenomena in low-concentration and industrial alkali.Requests for further information and resources and reagents can be directed to the Lead Contact, Prof. Liqiang Mai (mlq518@whut.edu.cn).This study did not generate new unique reagents.The authors declare that the data supporting the findings of this study are available within the article and the Supplemental Information. All other data are available from the Lead Contact upon reasonable request.First, NiMoO4\u00b7xH2O nanowire/nanosheet arrays on nickel foam were fabricated following previous reports.\n36\n\n,\n\n37\n Next, the anodic oxidation process on NiMoO4\u00b7xH2O precursor was carried out in the standard three-electrode system in 1\u00a0M KOH and operated on an CHI760E electrochemical analyzer. A piece of NiMoO4\u00b7xH2O served as a working electrode, and the graphite rod and the unused Hg/HgO electrode served as a counterelectrode and a reference electrode, respectively. After carrying out CV tests in 0.924\u20131.724 VRHE at a scan rate of 50\u00a0mV s\u22121 for 30 cycles, black CR-NiOOH nanowire/nanosheet arrays were obtained with a mass loading of \u223c1.2\u00a0mg cm\u22122. In addition, after calcination of NiMoO4\u00b7xH2O nanowire arrays at 550\u00b0C, NiMoO4 nanowire arrays with a mass loading of \u223c2.6\u00a0mg cm\u22122 were fabricated.24\u00a0mmol Ni(NO3)2\u00b76H2O and 24\u00a0mmol Na2MoO4\u00b72H2O were dissolved into 360\u00a0mL deionized water and formed a transparent green solution. The solution was then transferred into a 500-mL Teflon-lined autoclave, and four pieces of nickel foam were added. After reaction at 120\u00b0C for 6 h, the nickel foam samples were taken out, washed, and vacuum dried, and NiMoO4\u00b7xH2O arrays were obtained.First, CoMoO4\u00b70.75H2O nanowires,\n38\n Co(CO3)0.5(OH)\u00b70.11H2O nanowires,\n39\n and Ni-BTC microspheres\n40\n were fabricated according to previous reports. Next, these compounds were transformed into the corresponding CR catalysts after CV activation at 0.924\u20131.724 VRHE at 50\u00a0mV s\u22121 for more than 30 cycles.Scanning electron microscope (SEM) images were collected with a JEOL-7100F microscope at an acceleration voltage range of 15\u201325 kV. Microscopy images, SAED patterns, elemental mapping, and linear scanning analysis were collected on JEM-2100F and Thermo Fisher Scientific Titan G260-300 scanning/transmission electron microscopes. Ex situ XRD patterns were obtained using a Bruker D8 Advance X-ray diffractometer with Cu K\u03b1 radiation. In situ XRD patterns for alkali soaking experiments were recorded using a Bruker D2 Phaser X-ray diffractometer. Raman spectra and in situ Raman spectra for alkali soaking experiments were recorded using a HORIBA HR EVO Raman system. XPS measurements were carried out using an ESCALAB 250Xi instrument. Element content was determined using an inductively coupled plasma atomic emission spectrometer (ICP-AES) on a PerkinElmer Optima 4300DV spectrometer.All electrochemical measurements were carried out in fresh KOH (1\u00a0M or 30 wt %) on a CHI 760E electrochemical station using a standard three-electrode system. The test samples grown on the substrates (nickel foam or carbon cloth) served as a working electrode, and an unused Hg/HgO electrode was applied as a reference electrode; a graphite rod served as a counterelectrode. EIS was recorded in a frequency range of 0.01\u2013100,000\u00a0Hz. Homogeneous ink was prepared by dispersing 8\u00a0mg commercial IrO2/C and 2\u00a0mg Vulcan XC-72R in 250\u00a0\u03bcL deionized water, 700\u00a0\u03bcL isopropyl alcohol, and 50\u00a0\u03bcL Nafion solution (5 wt %). Next, 9\u00a0\u03bcL ink was coated on glassy carbon with an area of 0.07069\u00a0cm2 for catalytic tests. All chronopotentiometric measurements were carried out by applying a constant current density of 10 mA cm\u22122. In 1\u00a0M KOH, the iR-corrected potentials were referenced to RHE based on the following equation:\n\n\n\n\nE\nRHE\n\n=\n\nE\n\nHg\n/\nHgO\n\n\n+\n0.059\n\u00d7\npH\n+\n\nE\n\nHg\n/\nHgO\n\no\n\n\u2212\niR.\n\n\n\n\nIn situ electrochemistry-Raman measurements were recorded using a HORIBA HR EVO Raman system (633\u00a0nm laser) and an electrochemical workstation (CHI760E). The potential-dependent in situ Raman spectra were recorded with 150-s duration and a 50-s interval, and the LSV measurements were carried out in 0.924\u20131.824 VRHE at 0.25\u00a0mV s\u22121 in 1\u00a0M KOH during in situ Raman testing. Time-dependent in situ Raman spectra were recorded at 10 mA cm\u22122 with an interval time of 150 s.All DFT simulations were performed using Vienna ab initio simulation package (VASP) software.\n41\n The exchange-correlation interactions were described by generalized gradient approximation (GGA)\n42\n within the Perdew-Burke-Ernzerhof (PBE) function.\n43\n A plane wave basis set was adopted with a cutoff of 500 eV. Gaussian-type smearing with an energy window of 0.05 eV was used for optimization and frequency calculation. The energy convergence tolerance was 0.01 millielectron volt (meV). The force tolerance for the optimization task was 0.05 eV/\u00c5. All calculations were performed with spin unrestricted, and initial magnetic moments of 2\u00a0Bohr magneton (\u03bcB) for Ni, 1\u00a0\u03bcB for K, and 0\u00a0\u03bcB for O and H were set. 1\n\u00d7\n1\n\u00d7\n1 K point was sampled. The GGA with Hubbard U parameter (GGA+U) method for Ni species was adopted with an Hubbard effective parameter (U-J) value of 6.6 eV, the same as in previous reports.\n44\n The DFT-D3 method was adopted for all calculations. The linear mixing parameter was set to 0.06, and the cutoff wave vector for the Kerker mixing scheme was set to 0.0001 to make electron state converge more stable than default settings.This work was supported by the National Natural Science Foundation of China (51521001 and 21890751), the National Key Research and Development Program of China (2016YFA0202603), the National Innovation and Entrepreneurship Training Program for College Students (WUT: 20191049701034), and the Foshan Xianhu Laboratory of the Advanced Energy Science and Technology Guangdong Laboratory (XHT2020-003). The S/TEM work was performed at the Nanostructure Research Center (NRC), supported by the Fundamental Research Funds for the Central Universities (WUT: 2019III012GX and 2020III002GX).X. Liu and L.M. conceived the idea. X. Liu, J.M., L.M., and R.G. designed the experiments, analyzed the results, and wrote the manuscript. X. Liu, B.W., and R.G. performed the experiments and analyzed the results. X. Liu, K.N., and X.W. performed the DFT computations and theoretical analyses. D.Z., J.W., and L.M. provided helpful suggestions and refined the manuscript. All authors read and commented on the manuscript and approved the final version of the manuscript.A patent application related to this work has been submitted in China (application number 201811648828.2).Supplemental Information can be found online at https://doi.org/10.1016/j.xcrp.2020.100241.\n\n\nDocument S1. Figures S1\u2013S24 and Tables S1\u2013S4\n\n\n\n\n\nDocument S2. Article plus Supplemental Information\n\n\n\n", "descript": "\n                  Fundamental investigations of reconstruction of oxygen evolution reaction (OER) pre-catalysts and performance evaluation under realistic conditions are vital for practical water electrolysis. Here, we capture dynamic reconstruction, including the geometric/phase structure, of hydrate molybdates at oxidized potentials. Etching-reconstruction engineering endows the formed NiOOH with a sub-5-nm particle-interconnected structure, as revealed by multi-angle electron tomography. The key to complete reconstruction is the multicomponent co-leaching-induced loose reconstruction layer, conductive to solution penetration and mass transport. This unique structure avoids particle agglomeration in catalysis and promotes complete exploitation of the catalyst with 1,350\u00a0h of durability to meet industrial requirements. Upon addition of iron during reconstruction, mainstream Fe-NiOOH with a retained structure forms. Coupled with MoO2-Ni arrays in a membrane-free and two-electrode cell, it achieves stable electrolysis in industrial-concentration KOH for 260 h. This work highlights the reconstruction chemistry of hydrate oxygen-evolving systems and their performance evaluation under industrial conditions.\n               "}
{"full_text": "The environmental pollution brought by automobiles emission is sharply increasing. The exhaust gas generated by the combustion of sulfur compounds in gasoline and diesel is a major source of pollution to acid rain and haze weather. To produce ultra-low sulfur transportation oil, one of the key measures is to develop efficient hydrodesulfurization catalyst for sulfur removal from the fuels (Sun and Prins, 2010; Escobar et\u00a0al., 2018; Vatutina et\u00a0al., 2016).Traditional crystalline alumina has excellent mechanical properties, suitable acid properties, and low price, especially the most widely used \u03b3-Al2O3 (Moser et\u00a0al., 2010; Santolalla-Vargas et\u00a0al., 2015). But the specific surface areas of these alumina are usually less than 300\u00a0m2/g, and the pore size distributions are also relatively wide range (generally 3\u201315\u00a0nm). In recent years, researchers have focused on the synthesis of ordered mesoporous alumina materials with specific surface area greater than 300\u00a0m2/g, ordered pore channels and narrow pore distribution. The ordered mesoporous alumina (OMA) material is expected to be widely used in the fields of catalysis, adsorption and separation (Yuan et\u00a0al., 2008; Wu et\u00a0al., 2011; Liu et\u00a0al., 2016). At present, the research on application of ordered mesoporous alumina is still in the experimental stage and the synthesis method is still complex and difficult. Thus, it is of great significance to develop and study new ordered mesoporous alumina material.It is generally believed that the interaction between the metal and the support plays an important role in sulfidation degree of the active component (Hu et\u00a0al., 2020). The strong interaction between active metal Mo (W) and the traditional \u03b3-Al2O3 support would result in the low sulfidation degree of MoS2 sulfide phase. Moreover, the metal Ni and \u03b3-Al2O3 is easy to form the non-active spinel phase, leading to the low utilization rate of metal Ni and the formation of undesired active type I \"Ni\u2013Mo\u2013S\" active phase (Van Veen et\u00a0al., 1993). Therefore, optimizing the active-support interaction is the focus of researchers to improve the catalytic activity of the bimetallic catalysts in the HDS reaction.Chelating agents, such as citric acid (CA) (Zhang et\u00a0al., 2017; Li et\u00a0al., 2011), ethylenediaminetetraacetic acid (EDTA) (Ortega-Dom\u00ednguez et\u00a0al., 2017; Badoga et\u00a0al., 2012), aminotriacetic acid (NTA) (L\u00e9lias et\u00a0al., 2009), and cyclohexanediaminetetraacetic acid (Cy-DTA) (Hiroshima et\u00a0al., 1997) are commonly used as additives during the impregnation process to modify the active metal species. The hydrogenating catalysts prepared by using chelating agents have displayed excellent activity and stability, and the effects of complexing agents on the hydrogenating catalysts include the following three aspects: 1) To delay the vulcanization of the auxiliary metal. 2) To weaken the strong interaction between the active component and the support, thus contributing to form more highly active type II \"Ni\u2013Mo\u2013S\" active phases; 3) To adjust the dispersions of the active phases on the surface of the catalyst support. The MoS2 phase with a suitable sheet length and stacking number can effectively improve the HDS activity of the catalysts (Cattaneo et\u00a0al., 2001; Nikulshin et\u00a0al., 2014; Asadi et\u00a0al., 2019).Based on above information, this research aims at synthesizing ordered mesoporous alumina (OMA) with high specific surface area, concentrated pore size distribution and outstanding hydrodesulfurization performance. The NiMoE/OMA series catalysts were prepared by adding different amounts of chelating agent EDTA. Dibenzothiophene (DBT) was served as the model reactant to evaluate the hydrodesulfurization activity of NiMoE/OMA series catalysts.Hydrothermal precipitation method is used to prepare ordered mesoporous alumina materials and the specific preparation steps are listed as follows: a certain amount of template PEG was added into the 0.6\u00a0mol/L aluminum nitrate solution. Then, 1.2\u00a0mol/L ammonium carbonate solution was dropwise added to the above mixed solution at a stirring rate of 400 r/min. Continue to stir for 3\u00a0h and then age for 6\u00a0h in a 70\u00a0\u00b0C water bath to obtain a white sol. After filtering, it needs to be dried at 70\u00a0\u00b0C for 12\u00a0h and calcined in a N2 atmosphere at 350\u00a0\u00b0C for 3\u00a0h, then heated up to 550\u00a0\u00b0C at a rate of 3\u00a0\u00b0C/min for 6\u00a0h. Finally, the series ordered mesoporous alumina materials were obtained through changing the ratio of PEG/Al (0.05, 0.1, 0.2), which were named as OMA-0.05, OMA-0.1 and OMA-0.2, respectively.The supported NiMo catalyst precursor was prepared by incipient wetness impregnation method with the mass fraction loadings of 3.5\u00a0wt% NiO and 15\u00a0wt% MoO3 respectively. NiMo/OMA catalyst precursor was obtained by dried at 80\u00a0\u00b0C for 5\u00a0h and calcined at 550\u00a0\u00b0C for 6\u00a0h.The modified catalysts were prepared the same as above, except that different proportions of EDTA chelating agent are added during the impregnation process. The synthesized catalysts were named as NiMoE(X)/OMA, where X represents the ratio of EDTA/Ni.X-ray powder diffraction (XRD) was used to perform phase analysis on the samples by Shimadzu X-6000 X-ray powder diffractometer (Cu K\u03b1 radiation, 40\u00a0kV tube pressure, and 30\u00a0mA tube current). Scanning range of 2\u03b8 is 0.5\u00b0\u20135\u00b0; wide angle range of 2\u03b8 is 5\u00b0\u201380\u00b0.Fourier Infrared Spectroscopy (FTIR) was used to analyze the skeleton structure of the sample by the DIGILAB FTS-3000 Fourier Infrared Spectrometer of Tianmei Technology Company. The sample and KBr were mixed and compressed at a mass ratio of 1: 100 with a resolution of 2\u00a0cm\u22121. Wavelength range: 400\u20131200\u00a0cm\u22121.Pyridine adsorption infrared spectrometer (Py-FTIR) was used to carry out qualitative and quantitative analysis of acidity in the samples by Digilab FT-IR from Bole Pacific Company. A vacuum (1\u00a0\u00d7\u00a010\u22123\u00a0Pa) was used to purify the sample for 2\u00a0h at 350\u00a0\u00b0C, and then cooled to room temperature.Scanning Electron Microscope (SEM) was used to analyze the micro-morphology of the samples by the Quanta 200F scanning electron microscope of the Dutch company FEI.High resolution transmission electron microscopy (HRTEM) was used to calculate the dispersity, average length (L\nav) and stack number (N\nav) of the active phases on the sulfided catalysts by the JEM 2100 transmission electron microscope of Japan JEOL Company. The stacks number and average length of the MoS2 active phases on each catalyst were statistically calculated by at least 300 sticks according to the equations (a) and (b) (L\u00f3pez-Ben\u00edtez et\u00a0al., 2017):\n\n(a)\n\n\nL\nav\n\n=\n\n\n\n\u2211\n\ni\n=\n1\n\nn\n\n\nn\ni\n\n\nl\ni\n\n\n\n\n\u2211\n\ni\n=\n1\n\nn\n\n\nn\ni\n\n\n\n\n\n\n\n\n(b)\n\n\n\nN\nav\n\n=\n\n\n\n\u2211\n\ni\n=\n1\n\nn\n\n\n\nn\ni\n\n\nN\ni\n\n\n\n\n\n\u2211\n\ni\n=\n1\n\nn\n\n\nn\ni\n\n\n\n\n\n\nwhere l\n\ni\n represents the length of MoS2 microstrips, n\n\ni\n is the numbers of MoS2 microstrips, N\n\ni\n refers to the layers stack of MoS2 active phases.The fraction of Mo atoms located on the edges of MoS2 clusters, which was denoted as f\nMo, was calculated using Mototal (the total number of Mo atoms) and Moedge (the number of Mo atoms located on the edges of MoS2 particles). The values of f\nMo were calculated by the equations (c) and (d).\n\n(c)\n\n\n\nf\n\nM\no\n\n\n=\n\n\nM\n\no\n\ne\nd\ng\ne\n\n\n\n\nM\n\no\n\nt\no\nt\na\nl\n\n\n\n\n=\n\n\n\n\u2211\n\ni\n=\n1\n\nt\n\n\n(\n6\n\nn\ni\n\n\u2212\n6\n)\n\n\n\n\n\u2211\n\ni\n=\n1\n\nt\n\n\n(\n3\n\n\nn\ni\n\n2\n\n\u2212\n3\n\nn\ni\n\n+\n1\n)\n\n\n\n\n\n\n\n\n\n(d)\n\n\n\nn\ni\n\n=\n\nL\n6.4\n\n+\n0.5\n\n\n\nWhere f\nMo represents the ratio of the number of Mo atoms at the edge to the total number of Mo atoms on the active phase, n\n\ni\n refers to the number of Mo atoms on the side of the MoS2 crystal strips.Low temperature N2-adsorption and desorption experiment (BET) was used to determine the specific surface area, pore volume and pore size distribution of the samples by using Micromeritics ASAP 2010 adsorption analyzer. The sample was pretreated at 100\u00a0\u00b0C for 1\u00a0h under vacuum at a pressure of 15\u00a0\u03bcm Hg, and purified and degassed at 350\u00a0\u00b0C for 3\u00a0h, and then subjected to static adsorption analysis at\u00a0\u2212196\u00a0\u00b0C with N2.Raman Spectrum was used to study the state of the metal species on the surface of the sample by using Renishaw's Raman spectrometer. The laser light source has a wavelength of 325\u00a0nm and a power of 8\u00a0MW.X-ray Photoelectron Spectroscopy (XPS) was used to analyze the surface Mo species of the catalyst samples reduced by CS2. The XPSPEAK 4.1 software was used to Gaussian fit the Mo 3d XPS spectrum, and the relative content of Mo species in different valence states on the catalyst surface was calculated. The degree of sulfuration of Mo species and the utilization rate of Ni metal were calculated according to formula (e) and (f) respectively.\n\n(e)\n\n\nM\n\no\nsulfurization\n\n=\n\n\nMo\n\n4\n+\n\n\n\n\nMo\n\n4\n+\n\n\n+\n\nMo\n\n5\n+\n\n\n+\n\nMo\n\n6\n+\n\n\n\n\n\n\n\n\n\n\n(f)\n\n\nN\n\ni\nNiMoS\n\n=\n\nNiMoS\n\nNiMoS\n+\n\nNiS\nx\n\n+\nNiO\n\n\n\n\n\n\n1g of sieved 40\u201360 mesh catalyst is charged into the constant temperature section of the reactor. The two ends of the bed are filled with 20\u201340 mesh quartz sands, and each bed is separated by quartz cottons. 3.0\u00a0wt% CS2 cyclohexane solution was used to presulfide the oxidation state of the series catalysts. Firstly, the flow rate of the presulfiding solution was maintained at 60\u00a0mL\u00a0h\u22121 for 20\u00a0min, and then adjusted to 5\u00a0mL\u00a0h\u22121. After 4\u00a0h of presulfiding, the device was washed three times with pure cyclohexane, and the remaining presulfiding solution was washed out. After pretreatment, a 500\u00a0ppm DBT cyclohexane solution was passed as the raw material under the conditions of the hydrogen pressure of 4\u00a0MPa, the temperature of 340\u00a0\u00b0C, and the volume ratio of H2/Oil of 200.\nFig.\u00a01\n(A) and (B) are the small-angle XRD and wide-angle XRD spectra of OMA materials prepared with different PEG/Al ratios respectively. As shown in Fig.\u00a01 (A), a relatively strong diffraction peak appears in the range of 0.5-1\u00b0, indicating that the pore structure of the synthesized alumina material is ordered. It can be seen from Fig.\u00a01(B) that there are no characteristic peaks of crystalline alumina, confirming that the synthesized alumina is an amorphous alumina material.\nFig.\u00a02\n shows the pore size distribution (A) and N2 adsorption-desorption isotherm (B) of the alumina material under different PEG/Al ratios. It can be seen from Fig.\u00a02(A) that the pore sizes of all the synthesized alumina materials are concentrated in the range of 5\u201310\u00a0nm, proving the relatively ordered pore size distribution of the support. Fig.\u00a02(B) shows that the N2 adsorption-desorption isotherm of the synthesized OMA material belongs to a characteristic type IV adsorption equilibrium curve with a H2 hysteresis loop in the range of P/P0\u00a0=\u00a00.5\u223c0.9. Table\u00a01\n summarizes the textural properties of OMA the series materials, among which the OMA-0.1 material shows an excellent pore structure with high specific surface area (328\u00a0m2\u00a0g\u22121), large pore volume 0.74 (cm3\u00b7g\u22121) and big average pore size (8.1\u00a0nm). Therefore, OMA material with PEG/Al ratio of 0.1 is chosen as the ideal support for hydrodesulfurization catalysts.\nFig.\u00a03\n shows the wide-angle XRD patterns of NiMoE/OMA catalysts modified with different EDTA/Ni molar ratios. As can be seen from Fig.\u00a03, NiMoE/OMA the series catalysts show a characteristic peak at 2\u03b8\u00a0=\u00a067.30\u00b0, indicating that the support exists in amorphous structure rather than crystalline alumina. Moreover, no peaks of metal oxides in the spectrum are detected, which indicates that there are no large particle aggregations of Mo and Ni species.\nFig.\u00a04\n shows the Raman diagrams of NiMoE/OMA oxidized state catalyst with different EDTA/Ni molar ratios. It can be seen from Fig.\u00a04 that the diffraction peaks at 339\u00a0cm\u22121 and 850\u00a0cm\u22121 are weak, which are attributed to the characteristic bending vibration peak of the MoO bond of MoO4\n2\u2212 tetrahedral species. The MoO4\n2\u2212 tetrahedral species are the products of the strong interaction between the support and the active metal, which are difficult to be vulcanized into the MoS2 active phases in the pre-vulcanization stage (Xiao et\u00a0al., 2018). In Fig.\u00a04, the series NiMoE/OMA catalysts show obvious wide peaks at 953\u00a0cm\u22121 attributed to the stretching vibration of the MoO bond of two-dimensional polymer Mo7O24\n6\u2212 species, which is easy to be vulcanized into MoS2 active phase in the pre-vulcanization stage (Parola et\u00a0al., 2002; Wang et\u00a0al., 2015). Therefore, the Raman result shows that the NiMoE/OMA catalysts modified by EDTA possess an optimal support-metal interaction force. The characteristic peak of the modified NiMoE(1.0)/OMA catalyst at 953\u00a0cm\u22121 is stronger than that of other catalysts, meaning that the MoS2 active phases are easier to be formed on the NiMoE(1.0)/OMA catalyst when a moderate amount of EDTA is added.To analyze the acid types and acid amounts, the Py-TR spectra of NiMoE/OMA series catalyst is shown in Fig.\u00a0S1. Fig.\u00a0S1(A) is the Py-TR spectra at 200\u00a0\u00b0C, representing the total acid amount, while Py-TR spectra at 350\u00a0\u00b0C represent the strong and medium strong acids in Fig.\u00a0S1(B) (Barzetti et\u00a0al., 1996; Zhang et\u00a0al., 2008). It can be found that the diffraction peak intensity of EDTA-modified NiMoE/OMA series catalysts increases firstly and then decreases followingly as the increase molar ratio of the EDTA/Ni. Table\u00a02\n shows the calculated acid amounts of NiMoE/OMA series catalysts. The order of the quantities of total acidities and the medium and strong acidities is NiMoE(1.0)/OMA\u00a0>\u00a0NiMoE(1.5)/OMA\u00a0>\u00a0NiMoE(0.5)/OMA\u00a0>\u00a0NiMoE(0.2)/OMA\u00a0>\u00a0NiMo/OMA.\nFig.\u00a0S2 shows the Mo 3d XPS spectra of the series sulfided NiMoE/OMA catalysts. According to the peak splitting data in Fig.\u00a0S2, the binding energy peaks at 228.9\u00a0\u00b1\u00a00.1\u00a0eV and 232.0\u00a0\u00b1\u00a00.1\u00a0eV are assigned to the Mo 3d5/2 and Mo 3d3/2 of Mo4+, and the binding energy peaks at 230.5\u00a0\u00b1\u00a00.1\u00a0eV and 233.6\u00a0\u00b1\u00a00.1\u00a0eV are ascribed to the Mo 3d5/2 and Mo 3d3/2 of Mo5+, whereas the binding energies of Mo 3d5/2 and Mo 3d3/2 of Mo6+ are 232.5\u00a0\u00b1\u00a00.1\u00a0eV and 235.6\u00a0\u00b1\u00a00.1\u00a0eV, respectively (Jos\u00e9 et\u00a0al., 2002). According to Table\u00a03\n, the sulfidation degree of the NiMoE/OMA catalysts modified by EDTA is all higher than 0.6 compared to the unmodified NiMo/OMA catalyst. With the increasing molar ratios of EDTA/Ni, the sulfidation degrees of the corresponding catalysts are also increased (Li et\u00a0al., 2019).\nFig.\u00a0S3 shows the spectra of Ni 2p XPS of the series sulfided catalysts. In Fig.\u00a0S3, the existence forms of Ni metals in the support are mainly NiMoS, NiSx and NiO compounds, among which the corresponding binding energies are 856.2\u00a0\u00b1\u00a00.1eV, the 853.2\u00a0\u00b1\u00a00.2eV and 861.2\u00a0\u00b1\u00a00.2eV respectively (Lai et\u00a0al., 2013). It can be found that Ni metals are mainly formed as NiMoS active phases, whereas the peak of NiSx is almost non-existent over NiMoE(0.5)/OMA and NiMoE(1.0)/OMA catalysts. As shown in Table\u00a04\n, the proportion of NiMoS phase increases from 0.54 over NiMo/OMA to 0.72 over NiMoE(1.5)/OMA, indicating that the addition of EDTA could significantly promote the formation of NiMoS active phases.HRTEM images of NiMoE/OMA sulfided catalysts and the distributions of stacking layers of MoS2 active phases are shown in Fig.\u00a05\n. The average length (L\nav), average number of layers (N\nav) and dispersion (f\nMo) of MoS2 stacks are calculated and listed in Table\u00a05\n. With the increase of EDTA/Ni ratios, the average lengths of NiMoE/OMA series catalysts decreases from 3.5\u00a0nm of NiMoE(0.2)/OMA to 3.2\u00a0nm of NiMoE(1.0)/OMA and then increases to 3.3\u00a0nm of NiMoE(1.5)/OMA. Among all the catalysts, the NiMoE(1.0)/OMA catalyst exhibits short average length (3.2\u00a0nm) and suitable stacking layers (2.8). The f\nMo value is closely related to the layer number and the average length of active phases, and the follows trend of NiMoE(1.0)/OMA\u00a0>\u00a0NiMoE(0.5)/OMA\u00a0>\u00a0NiMoE(0.2)/OMA\u00a0>\u00a0NiMoE(1.5)/OMA.\nFig.\u00a06\n shows the DBT HDS performance of EDTA modified catalysts at different WHSVs. The HDS activity of NiMoE/OMA catalysts show a gradually increasing trend with the decrease of WHSVs from 100 h\u22121 to 20 h\u22121. The DBT HDS conversions of the series catalysts are in order of NiMoE(1.0)/OMA\u00a0>\u00a0NiMoE(1.5)/OMA\u00a0>\u00a0NiMoE(0.5)/OMA\u00a0>\u00a0NiMoE(0.2)/OMA, among which the highest desulfurization rate of NiMoE(1.0)/OMA catalyst is 97.7% at 20 h\u22121 WHSV. Compared with the unmodified NiMo/OMA catalyst, NiMoE/OMA catalysts modified with EDTA display higher hydrodesulfurization activities at each WHSVs.The cyclohexylbenzene (CHB) and biphenyl (BP) are the main products of DBT HDS detected from GC-MS, which are the main products of HYD and DDS reaction route respectively [24]. The product distributions of the series NiMoE/OMA catalysts are shown in Fig.\u00a07\n. As can be seen, the CHB selectivity over NiMoE/OMA catalysts increases as the ratio of EDTA/Ni increases from 0.0 to 1.0, whereas the BP selectivity decreases as the increasing amount of EDTA. Therefore, the appropriate addition of EDTA has great influence on the promotion of HYD reaction route of DBT HDS to some extent.Through homogeneous precipitation method, the synthesized ordered mesoporous alumina (OMA) has ordered mesoporous channels (8.1\u00a0nm), the high surface area (328\u00a0m2\u00a0g\u22121) and the large pore volume 0.74 (cm3\u00b7g\u22121), which could eliminate the diffusion resistance of the reactants and products through the mass transfer process and increase the accessibility of the reactants to the active metals. NiMoE/OMA catalysts synthesized by EDTA post-modification method show higher HDS activity of DBT compound compared with the unmodified NiMo/OMA catalyst. The HDS evaluation of DBT compound shows that as the molar ratios of EDTA/Ni increase, the DBT hydrodesulfurization conversions of the series catalysts increase at first and then decrease gradually, among which NiMoE(1.0)/OMA catalyst displays high DBT hydrodesulfurization activity, reaching 97.7%. The addition of EDTA has several following effects that are beneficial to the catalytic activity.The physicochemical properties of the OMA support contributes a lot to the diffusion process of DBT molecules. Through adjusting the ratio of PEG/Al during preparing ordered mesoporous alumina materials, the OMA with PEG/Al ratio of 0.1 exhibits high surface area (328\u00a0m2\u00a0g\u22121), large pore volume 0.74 (cm3\u00b7g\u22121) and big average pore size (8.1\u00a0nm). The well-ordered mesoporous structure and open pore channels of the OMA support can largely reduce the diffusion resistances of reactants and products compared to the traditional alumina materials. In addition, it can be seen from Fig.\u00a03, no peaks of metal oxides are detected in the XRD spectra, which indicates that the outstanding pore structure of OMA supports can also avoid the aggregation of active metals.Secondly, the addition of EDTA is beneficial to the formation of Ni\u2013Mo\u2013S active phases. The XPS results show that NiMoE(1.0)/OMA and NiMoE(1.5)/OMA catalysts have the highest sulfidation degree of 0.65. The higher the sulfidation degree of the catalyst, the easier to form the type II \u201cNi\u2013Mo\u2013S\u2033 active phase, which is more conducive to the HDS performance of the catalyst. As the molar ratio of EDTA/Ni increases, the proportion of NiMoS also increases, as shown in Table\u00a04. It is well-known that NiMoS active phase is formed at the edge of MoS2 layer structure by horizontal bonding (Liu et\u00a0al., 2020). And some non-active phases are easily formed between Ni and alumina support, which leads to low utilization rate of Ni active metal. Coulier et\u00a0al., (2001) has reported that EDTA chelating ligand could stabilize nickel against sulfidation by forming a stable (NiEDTA)2- complex with the Ni promoter. The interaction between EDTA and Al cations acts as a driving force for decomposition of the (NiEDTA)2- complex so as to retard sulfidation of Ni to temperatures where Mo is completely sulfided in the form of MoS2 (Al-Dalama, Stanislaus, 2011). And Ni metallic species are combined with MoS2 at the corner positions of hexagonal configurations to form the highly NiMoS active phase, which greatly enhance the utilization rate of Ni active metals. Therefore, the presence of EDTA chelating ligands can achieve a good level of interaction between Ni and the MoS2 crystallites (Van Veen et\u00a0al., 1993; Zepeda et\u00a0al., 2016).Moreover, the chelating agent EDTA promotes the dispersion of both Ni(II) and Mo(VI) on alumina. It has been reported that the chelating ligands EDTA tend to isolate Ni from the environment, thus avoiding the formation of excessive bulk Ni sulfide (Rana et\u00a0al., 2007; Zhao et\u00a0al., 2006). Meanwhile, EDTA had a higher coordinating constant with the surface Al cations than Mo(VI). Free EDTA can compete with Mo(VI) on the alumina surface sites. Therefore, chelating agent EDTA can also limits strong interaction between the metal Mo ions and alumina. According to the HRTEM images, The NiMoE/OMA-1.0 catalyst has a short length of MoS2 active phase (3.2\u00a0nm), a moderate number of stacking layers (2.8), and a maximum dispersion degree (0.35). The Raman characterization results show that the stretching vibration of the MoO bond of polymer Mo7O24\n6\u2212 species (which is easy to vulcanize) over EDTA-modified NiMoE/OMA series catalysts is higher than that of unmodified NiMo/OMA.The addition of EDTA can also change the acid properties of the modified catalysts so as to enhance DBT HDS activity. As shown in Py-FTIR characterization results, with the addition of EDTA increasing, the total acidities of NiMoE/OMA catalysts increase firstly and then decrease, among which the NiMoE(1.0)/OMA catalyst possesses the highest acid amount. It is well known that acidity plays an important role in improving the adsorption of DBT via its aromatic rings, thus finally enhancing HDS activity (Wang et\u00a0al., 2016). The NiMoE(1.0)/OMA catalyst with the maximum total acid amount exhibits the high selectivity of CHB product, which indicates that the acid properties of the NiMoE/OMA catalysts have great influence on HYD and DDS pathways during the DBT HDS process.Based on the above analysis, the introduction of an appropriate amount of EDTA can promote the HDS performance of DBT and HYD reaction route due to the combination effect of the weak interaction between the metal support, the high sulfidation degree of NiMoS phase and suitable acid properties on HDS.Ordered mesoporous alumina (OMA) was successfully synthesized by homogeneous precipitation with aluminum nitrate as the inorganic aluminum source, ammonium carbonate as the precipitant and polyethylene glycol (PEG) as the template. The chelating agent EDTA was added to adjust the interaction between the support and the metal, so as to improve the sulfidation degree of Mo2S and enhance utilization rate of Ni metal. NiMoE/OMA catalysts synthesized by EDTA post-modification method displayed higher HDS activities of DBT compound compared with the unmodified NiMo/OMA. And among all the synthesized catalysts, NiMoE(1.0)/OMA catalyst showed the highest desulfurization rate of 97.7% and more HYD hydrogenation products of CHB.This research is financially supported by the National Natural Science Foundation of China (No. 21878330, 21676298) and the National Key R&D Program of China (2019YFC1907602) and the CNPC Key Research Project (2016E-0707).The following is/are the supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.petsci.2021.11.005.", "descript": "\n                  In this paper, ordered mesoporous alumina (OMA) support with the high surface area (328\u00a0m2\u00a0g\u22121) and the large pore volume 0.74 (cm3\u00b7g\u22121) was synthesized by homogeneous precipitation method. And the influence of EDTA on the physical and chemical properties of the modified catalysts was also studied. The characteristic results showed that the addition of EDTA could adjust the metal-support interaction and improved the acidity of the corresponding catalyst. Combined with the catalytic performance results, the EDTA-modified NiMoE(1.0)/OMA catalyst displays the highest DBT hydrodesulfurization conversion (97.7%).\n               "}
{"full_text": "Data will be made available on request.Polymerizations in aqueous heterogeneous media are characterized by good heat transfer, low amount of volatile organic solvents, low viscosity, and low toxicity [1], making them advantageous for commercial applications. [2] Moreover, conducting radical polymerizations in heterogeneous systems promotes easier access to high molecular weight (MW) polymers, because radical segregation and compartmentalization limit radical termination events. [3\u20136] In these dispersed and segregated systems, the small polymerization loci perform as \u201cnanoreactors\u201d [7] that enable the preparation of various nanoobjects. Polymerizations in microemulsion, miniemulsion, emulsion, and dispersion have been used to prepare polymer-based nanoparticles with various morphologies (e.g., core-shell, microcapsules, and multilayered particles), which have found applications in catalysis, coatings, and in the biomedical and diagnostic fields. [1,8\u201311].Reversible-deactivation radical polymerizations (RDRPs) are robust and versatile techniques for the synthesis of polymers with predetermined MW and low dispersity (\u0110), starting from a wide range of monomers. [12,13] Among RDRP techniques, atom transfer radical polymerization (ATRP) [14\u201318], reversible addition-fragmentation chain-transfer (RAFT) polymerization [19], nitroxide-mediated polymerization (NMP) [20], and organotellurium mediated polymerization (TERP) [21] have been successfully developed in both homogeneous and heterogeneous media [22\u201329]. While limited advances in NMP in dispersed media were reported over the past 15 years [26], progress in TERP and particularly RAFT polymerization in dispersed media predominantly focused on exploiting the self-assembly of amphiphilic polymer chains prepared through emulsion polymerization processes, to produce nanoobjects with controllable morphologies. On the other hand, the implementation of ATRP in dispersed media has considerably advanced in the past 15 years, driven by the development of novel polymerization components and strategies to mediate ATRP systems [30], giving access to a broader range of building blocks and polymer architectures. In contrast to other RDRPs that require a stoichiometric amount of chain transfer or trapping agents, ATRP is a catalytic process, and the continuous evolution in catalyst design has enabled to prepare polymers with minimal catalyst contamination, which can additionally be removed through various methods.ATRP is based on radical generation by the \u201cactivator\u201d form of a catalyst, which is typically a Cu complex with a polydentate amine ligand (L) in its lower oxidation state, i.e., CuI/L (Scheme 1\n). The CuI/L complex activates an alkyl halide initiator (R\u2013X) or dormant chain end (Pn\u2013X) via an inner sphere electron transfer (ISET) process, forming a propagating radical and a higher oxidation state, \u201cdeactivator\u201d complex, X\u2013CuII/L. [31\u201333] The high concentration of dormant species minimizes the fraction of terminated chains, in contrast to conventional radical polymerization. The fast initiation and rapid activation-deactivation equilibrium ensure that all chains grow concurrently, resulting in polymers with low dispersity and high chain-end functionality. [34].To compensate for the accumulation of CuII deactivator, which is generated by unavoidable radical termination, traditional ATRP methods required high concentration of Cu species. However, this caused issues related to catalyst solubility and removal. In the past decade, the loading of Cu catalysts was drastically decreased from over 10,000\u202fparts per million (ppm, expressed as molar concentration relative to the monomer) to hundreds or less ppm, by implementing a variety of methods for the continuous regeneration of the CuI/L activator. These methods include the addition of a reducing agent in the polymerization system, as in activators re-generated by electron transfer (ARGET) ATRP [35,36], the addition of a thermal radical initiator as in initiators for continuous activator regeneration (ICAR) ATRP [37], and the use of metallic Cu in supplemental activator and reducing agent (SARA) ATRP [38\u201340], as well as the use of external stimuli such as electrical current, light, and ultrasounds in electrochemically mediated ATRP (eATRP) [41,42], photoATRP [43,44], and mechanoATRP, respectively. [45,46] These techniques are collectively called \u201cATRP with activator regeneration\u201d or \u201clow-ppm ATRP\u201d, and they can provide polymeric materials with complex architectures, including decorated nanoparticles, networks and gels. [15] The residual small amount of catalyst could be left in the product, or removed by column filtration, electrodeposition, or other purification techniques [47\u201350], achieving a sufficiently low Cu contamination for most applications.During the past 15 years, low-ppm ATRP methods have been successfully implemented in dispersed media. Most initial works were carried out in miniemulsion, due to the advantage of conducting polymerizations in a \u201cmini-bulk\u201d environment, with minimal migration of polymerization components into the continuous phase. More recently, ATRP was expanded to ab initio emulsion, thanks to an improved understanding and engineering of catalyst location during the heterogenous polymerization process.This minireview describes the development of low-ppm ATRP in dispersed media. The outline and structure of this minireview is presented in Scheme 2\n. Section 2 provides relevant background on the topics of polymerization in dispersed media and low-ppm ATRP. Section 3 discusses synthetic strategies involving engineered polymerization components, particularly surfactants and catalysts, as well as the different external stimuli that were used to trigger ATRP in (mini)emulsion systems. The unique features resulting from the combination of heterogeneous polymerizations and low-ppm ATRP enabled to prepare a broad variety of well-defined polymer architectures, which are presented in Section 4. Finally, relevant applications are reviewed in Section 5, while conclusions and perspectives are provided in Section 6.Heterogeneous polymerization processes involve multiphase systems where the starting monomer(s) and/or the resulting polymer are dispersed in an immiscible liquid. Typically, dispersed media are generated by using a surfactant and an external force to form a kinetically and hydrodynamically stable mixture, although the dispersion can also be thermodynamically stable in some cases. The dispersed droplets have a spherical shape, which enables to minimize the surface-to-volume ratio, and thus the surface energy. [51] Typically, oil-in-water (O/W) systems are employed, where \u201coil\u201d refers to any water-insoluble liquid (monomer/polymer), and water is the continuous phase. Water-in-oil (W/O) systems are also possible.Heterogeneous polymerizations are generally categorized as suspension, emulsion, miniemulsion, microemulsion, dispersion, or precipitation. [51] However, the nomenclature and definitions are sometimes ambiguous in the literature. In this review, the different techniques of polymerization in dispersed media are distinguished by considering four features: (i) the initial state of the polymerization mixture; (ii) the kinetics of polymerization; (iii) the mechanism of particle formation; and (iv) the size of the final polymer particles. The following paragraphs describe the most relevant features of the different heterogenous polymerization techniques, which are then summarized in Table 1\n.\nSuspension polymerization. To perform a suspension polymerization, the initiator is first dissolved in the monomer phase, which is then dispersed in the aqueous phase to form droplets. Water is a nonsolvent for both the monomer and the polymer. The mixture is stirred in the presence of a droplet stabilizer or suspension agent, such as poly(vinyl alcohol). Via thermal polymerization, the monomer droplets are converted directly to the polymer microbeads with no significant size change. Suspension polymerization is typically used to produce polymer beads with a size of 20\u202f\u03bcm - 2\u202fmm. [51].\nPrecipitation polymerization. This technique starts from a homogeneous solution of monomer, initiator, and stabilizers; the system quickly turns into a heterogeneous one, because the generated polymer is not soluble in the reaction medium beyond a critical MW. For example, in the polymerization of polyacrylonitrile (PAN), the polymer is insoluble in its own monomer (acrylonitrile), therefore it precipitates after reaching a critical chain length. After the first phase of particles nucleation, the monomer swells the particles, so the polymerization continues in the droplets/colloids. The polymer particles have a size in the range of 1\u201315\u202f\u03bcm. [22].\nDispersion polymerization. This method is a subclass of precipitation polymerizations, with final particle dimension <10\u202f\u03bcm, i.e., colloidal dimensions. The smaller particle size of a dispersion polymerizations is achieved by applying more effective dispersing agents in larger amounts than in typical precipitation polymerizations.\nEmulsion polymerization. A traditional emulsion polymerization system, also known as ab initio emulsion polymerization system, consists of a hydrophobic monomer, a water-soluble initiator, a surfactant, and water. At the beginning of the polymerization, a large portion of the monomer resides in large, surfactant-stabilized monomer reservoirs (\u226b 1\u202f\u03bcm), and only a small fraction of monomer molecules is in water and in surfactant-formed micelles (<10\u202fnm). The initiator molecules are decomposed in water, where they initiate the growth of oligomeric radicals. Upon reaching a critical chain length, the hydrophobic oligomeric radicals enter the micelles rather than the monomer droplets, due to the much higher surface area of the smaller and numerous micelles relative to the larger but fewer monomer droplets. The monomer molecules diffuse from the reservoirs into the aqueous phase, and enter the micelles to propagate the radical chains, resulting in the formation and growth of polymer particles (Scheme 3\n). The final polymer particle size is 50\u2013500\u202fnm. [52].\nMiniemulsion polymerization. Before polymerization, the hydrophobic monomer, oil-soluble initiator, and a hydrocarbon co-stabilizer form a macroscopic organic phase, while the surfactant is dissolved in a macroscopic aqueous phase (Scheme 4\n). The miniemulsion is generated by a vigorous homogenization process, such as by employing probe ultrasonication or a microfluidizer. The monomer droplets have a size of 50\u2013500\u202fnm, and they are stabilized by the surfactant and the co-stabilizer, which strongly limit the mass transfer of monomer during the polymerization. Thus, the polymerization proceeds in each droplet like in a \u201cmini-bulk\u201d polymerization and, as a result, the size of the final polymer particles is similar to that of the initial monomer droplets. [53].\nMicroemulsion polymerization.\n By using a large amount of an appropriate surfactant, the interfacial tension of a dispersed media can be ultra-low, leading to particle size <50\u202fnm, and an optically transparent and thermodynamically stable system with high interfacial area. This system is termed microemulsion [54], and the polymerization proceeds similarly to a miniemulsion system.Other specialty configurations of polymerization in dispersed media are possible, some of which are described in the following paragraphs.\nSelf-assembled non-spherical nanoparticles. Molecules of polymeric surfactants can self-assemble into structures that are not necessarily globular, resulting in a combination of spherical micelles, cylindrical micelles, vesicles, and even bicontinuous planar interfaces. Polymerization occurring inside these particles can lead to the formation of non-spherical latex particles. [55].\nPickering emulsion. A Pickering emulsion is an emulsion that is stabilized by solid particles (for example colloidal silica or proteins) which adsorb onto the interface between the water and oil phases.\nHigh Internal Phase Emulsions (HIPEs). Typically, for any polymerization mechanism, the volume percentage of the internal (i.e., dispersed) phase is 55% or less. Higher content of internal phase often results in very high viscosity. In fact, the maximum volume percentage occupied by uniform, non-deformable spheres packed in the most effective way is 74%. High internal phase emulsions (HIPEs) represent a special case where the internal (droplet) phase exceeds 74% of the total volume of the system. HIPEs are generally stabilized by large amounts of surfactants. Because of its high-volume fraction, the dispersed phase forms non-uniform interconnected spheres or polyhedral shapes. In such systems, the continuous phase is loaded with the monomer(s) and crosslinker(s), and polymerized to yield, upon purification, a highly porous material with interconnecting voids, which is called polyHIPE or pHIPE (Scheme 5\nA). This emulsion templated method is convenient for the synthesis of porous polymers, as it can provide a wide variety of highly interconnected, highly porous monolithic systems (Scheme 5B). [56\u201358].The equilibrium constant of ATRP, K\nATRP, is expressed by Equation (1), where P\nn\n\n\u2981 and P\nn\n-X are, respectively, the active and dormant chains.\n\n(1)\n\n\n\n\n\n\n\nK\n\nA\nT\nR\nP\n\n\n=\n\n\n\n[\n\nP\nn\n\u2981\n\n]\n\n\n[\n\nX\n\u2212\n\n\n\nC\nu\n\n\nI\nI\n\n\n/\nL\n\n\n]\n\n\n\n\n[\n\n\nP\nn\n\n\u2212\nX\n\n]\n\n\n[\n\n\n\nC\nu\n\nI\n\n/\nL\n\n]\n\n\n\n\n\n\n\n\n\n\n\nThus, the rate of ATRP polymerization, R\np, can be expressed by Equation (2) (where M is the monomer, and k\np is its propagation rate constant), and it depends on the relative amount of CuI/L activator and X\u2013CuII/L deactivator.\n\n(2)\n\n\n\n\n\n\n\nR\np\n\n=\n\nk\np\n\n\n[\nM\n]\n\n\n[\n\nP\nn\n\u2981\n\n]\n\n=\n\nk\np\n\n\nK\n\nA\nT\nR\nP\n\n\n\n[\nM\n]\n\n\n[\n\n\nP\nn\n\n\u2212\nX\n\n]\n\n\n\n[\n\n\n\nC\nu\n\nI\n\n/\nL\n\n]\n\n\n[\n\nX\n\u2212\n\n\n\nC\nu\n\n\nI\nI\n\n\n/\nL\n\n\n]\n\n\n\n\n\n\n\n\n\nThe dispersity \u0110\u202f=\u202fM\nw/M\nn of the resulting polymer decreases by increasing the equilibrium concentration of the X\u2013CuII/L deactivator, according to Equation (3), where DP\nn\n is the degree of polymerization, k\ndeact is the deactivation rate constant, and p is the conversion.\n\n(3)\n\n\n\n\nM\nw\n\n\nM\nn\n\n\n=\n1\n+\n\n1\n\n\nD\nP\n\nn\n\n\n+\n\n(\n\n\n\nk\np\n\n\n\n[\n\nR\nX\n\n]\n\n0\n\n\n\n\nk\n\nd\ne\na\nc\nt\n\n\n\n[\n\nX\n\u2212\n\n\n\nC\nu\n\n\nI\nI\n\n\n/\nL\n\n\n]\n\n\n\n)\n\n\n(\n\n\n2\np\n\n\u2212\n1\n\n)\n\n\n\n\n\nIn low-ppm ATRP techniques, the polymerization kinetics follows the steady-state in radical concentration. [30,59] Thus, R\np is expressed by Equation (4), whereby the numerator corresponds to the rate of CuI/L (re)generation.\n\n(4)\n\n\n\nR\np\n\n=\n\nk\np\n\n\n[\nM\n]\n\n\n[\n\nP\nn\n\u2981\n\n]\n\n=\n\nk\np\n\n\n[\nM\n]\n\n\n\n\nR\n\n\n\nC\nu\n\nI\n\n/\nL\n\nr\ne\ng\ne\nn\ne\nr\na\nt\ni\no\nn\n\n\n\nk\nt\n\n\n\n\n\n\n\nIn ARGET ATRP, the CuI/L activator is (re)generated by means of a chemical reducing agent, such as the water-soluble ascorbic acid (AsAc) or oil-soluble SnIIR2 compounds. [35,36] The rate of polymerization depends on the amount or feeding rate of the reducing agent. In ICAR ATRP, a thermal radical initiator such as the water-soluble 2,2\u2032-azobis[2-(2-imidazolin-2-yl)propane] dihydrochloride (VA-044) or the oil-soluble azobisisobutyronitrile (AIBN) are employed to exploit their slow thermal decomposition to induce the reduction of the X\u2013CuII/L deactivator. [30] The slow decomposition of the thermal radical initiator can cause the generation of a small fraction of new chains, thus ICAR ATRP can be non-ideal for the synthesis of well-defined block copolymers. SARA ATRP uses zero-valent copper as both supplemental activator and reducing agent [39,60], and the rate of polymerization is affected by the ratio of the surface area of Cu0 and the volume of the reaction. Metallic Cu can be reused for multiple polymerizations or can be periodically lifted from the reaction mixture to achieve temporal control. [61].Electrochemically mediated ATRP (eATRP) takes advantage of an applied current or voltage to continuously reduce CuII species. [41] No external chemicals are needed in eATRP. Since the rate of polymerization is affected by the ratio of CuI and CuII species, it can be tuned by the applied potential, according to the Nernst equation. [59] Temporal control over polymerization can also be achieved by adjusting the applied potential or current. eATRP was recently scaled up to a pilot scale. [62].In photoATRP, UV or visible light promotes the photo-excitation of X\u2013CuII/L complexes followed by the reduction of excited X\u2013CuII/L* to CuI/L in the presence of an electron-donor species, typically an aliphatic amine, which can be an excess of ligand. [43,44] Thus, the rate of polymerization is influenced by the ratio of ligand (or other electron donor) to copper. Similar to eATRP, the use of light as an external stimulus enabled temporal control over the polymerization. PhotoATRP can also be performed by employing other metal photocatalysts, which typically operate in the absence of an electron donor, as well as organocatalysts, such as phenothiazines, phenazines, and phenoxazines. [63,64] A procedure for scale up of photoATRP has also been recently disclosed. [65].\nMechanoATRP and sonoATRP employ ultrasounds to (re)generate the activator. In the first technique, the sonication of piezoelectric materials induces an electron transfer to the CuII species [45,46], while in the second case, the application of ultrasound in aqueous media produces hydroxyl radicals from water molecules. [66] The hydroxyl radicals react with monomer or with alcoholic solvent, forming carbon based radicals that start the propagation.The implementation of ATRP in dispersed media requires the partitioning of each species among the aqueous and organic phases prior, throughout, and at the end of the polymerization. In a typical free radical polymerization in dispersed media the only components are the monomer and radical initiator, besides the surfactant and/or eventual co-surfactant, (co)stabilizers or dispersing agents (Table 1). Conversely, a low-ppm ATRP requires an alkyl halide initiator, the catalyst, i.e., a Cu-halide salt and the ligand, and either a reducing agent, a thermal radical initiator, metallic Cu, or an external trigger. All these elements should be located in the appropriate phase (water, oil, or their interphase) at any stage of the polymerization. For instance, at the onset of a miniemulsion ATRP, the catalyst and RX must reside in the surfactant-stabilized monomer droplets, and they should remain in the hydrophobic phase for the whole duration of the process. An eventual reducing agent or thermal radical initiator should also reach the hydrophobic phase, and the eventual external stimulus must be conveyed to the hydrophobic phase. Therefore, the design of low-ppm ATRP systems in dispersed media necessitates tuning the hydrophilicity of the catalyst, RX and eventual reductants, as well as appropriately selecting the surfactant and/or stabilizers. Low-ppm ATRP techniques have been effectively performed in various types of dispersed media by carefully designing the catalytic systems, as it will be explained in Section 3.An important advantage provided by the use of heterogeneous media is the possibility to reduce or suppress bi-radical termination events. In homogeneous ATRP, bi-radical termination is unavoidable and can result in decreased rate of polymerization, low chain-end functionality, and even gelation if a branching point exists. In dispersed media, radicals located in different particles are unable to terminate with each other (segregation effect), which is particularly relevant for the design of complex polymer architecture. [3] On the other hand, the confined space effect can result in enhanced reaction rate between two radicals located in the same particle as the particle size decreases. However, compartmentalization in dispersed media ATRP also enhances the rate of deactivation. Thus, depending on the catalytic system, lower polymerization rates and improved control were observed for particle volumes below a threshold value. [7,67] The latter is dependent on the particular system, and it generally increases with increasing the targeted degree of polymerization, i.e., decreasing the amount of initiating molecules (and thus of growing radicals) confined within each particle. [67].In dispersed media, each droplet acts as a \u201cnanoreactor\u201d, allowing the preparation of nanoobjects that cannot be easily prepared in other media, including crosslinked nanoparticles, nanocapsules, and core-crosslinked hairy nanoparticles. In Pickering emulsion system, Janus platelets with either a polymer grafted on a single side or different polymers on each side were synthesized. By adding the monomer and crosslinker to the continuous phase, porous polymer monoliths could be fabricated. The different polymer architectures prepared by low-ppm ATRP in dispersed media will be presented in Section 4, whereas their most relevant applications will be discussed in Section 5.Seminal ATRP in dispersed media with high loading of Cu catalyst generally employed non-ionic or cationic surfactants, which stabilized the latex without interfering with the cationic Cu complexes and thus with the ATRP equilibrium. [68,69] In contrast, anionic surfactants can interact with the Cu complexes used as catalysts, modifying their stability and catalytic activity. [22] The introduction of low-ppm ATRP methods was concomitant to the engineering of the surfactant, resulting in the use of anionic, ionic liquid, and reactive surfactants.The most important drawback of conventional surfactants is their tendency to remain in the final polymer, negatively affecting its electric, photonic, and surface properties. In conventional emulsion free radical polymerization, this issue was overcome by designing soap-free emulsion systems. In such systems, traditional surfactants are replaced by initiator or monomer molecules capable of anchoring onto the surface of the latex particles, leading to improved colloidal stability and the absence of surfactant leaching from the produced latexes. This immobilization strategy eliminates the need for surfactant removal after polymerization. In ATRP, soap-free emulsion polymerizations were developed by exploiting the concept of \u201creactive surfactant\u201d [70], which is a multifunctional molecule that combines the function of a surfactant with an initiator, monomer, or catalyst/ligand. [71,72].For example, hydrophilic or amphiphilic macroinitiators prepared by RDRP were employed as reactive surfactants to simultaneously initiate polymerizations and stabilize micelles, forming amphiphilic polymers that behave as macro-emulsifiers (Table 2\n). In fact, poly(ethylene oxide) homopolymer with terminal \u03b1-bromoisobutyrate moiety (PEO-Br), and poly(ethylene oxide)-b-polystyrene (PEO-b-PSt-Br) block copolymer made by ATRP were used as macroinitiators and stabilizers in miniemulsion AGET ATRP of butyl acrylate (BA), generating polymer latexes with narrow particle size distribution. [73].Besides macroinitiator surfactants, small-molecule surfactants can be used to initiate an ATRP. When a polymerization initiating site is introduced in surfactant molecules the resulting construct is termed an \u201cinisurf\u201d, i.e., initiator-surfactant. [74] Both anionic [75,76], and cationic [77,78] inisurfs were employed for ATRP in dispersed media. Dextran derivatives bearing a phenoxy hydrophobic group were modified to introduce \u03b1-bromoisobutyrate sites for ATRP initiation, forming a multifunctional inisurf. Nanoparticles with hydrophobic cores and hydrophilic shells were formed during polymerization. [9] By grafting polymer from the inisurf, a \u201csterically-stabilized latex\u201d was obtained, which was particularly resistant toward destabilization induced by high shear force, electrolyte addition, and freeze-thaw.The various types of reactive surfactants provided latexes with increased stability and eliminated the need for removing the surfactant after polymerization. Alternatively, surfactant monomers have been employed, that copolymerized within the main chain. Unsaturated molecules such as a methacrylic ester [79] or a cardanol ether [80], were covalently attached to a tetraalkylammonium cationic surfactant, forming surfactant-monomers exhibiting drastically decreased CMC values and enhanced stabilization capabilities.A complementary strategy consists of incorporating multidentate nitrogen groups into the surfactant to form a surfactant-ligand (SL). In ab initio emulsion ATRP, a SL \u201clocked\u201d the CuII on the surface of the droplet, eliminating the escape of CuII to water. [81,82] However, the immobilization of the catalyst on the surface of polymerizing droplets can restrict its diffusion rate and result in decreased degree of control. The addition of conventional ligand, dNbpy (4,4\u2032-dinonyl-2,2\u2032-bipyridine), narrowed the molecular weight distribution. [83].Ionic liquid surfactants represent a useful alternative to conventional ionic or neutral surfactants as they can be easily separated and reused. N-tetradecyl-N-methyl-2-pyrrolidonium bromide was employed in microemulsion AGET ATRP of methyl methacrylate (MMA). The ionic liquid surfactant had low toxicity and was recycled and reused for up to 5 times. [84].Besides ionic liquids, insoluble solids that could be partially wetted by both phases could significantly decrease the surface energy and stabilize the emulsion, giving, in this case, a so called Pickering emulsion. [85] An example of Pickering agent is cellulose nanocrystals (CNCs), which have been used for a photoATRP catalyzed by Eosin Y. [86] CNCs could be recycled and reused for Pickering emulsion polymerization multiple times. CNCs modified with \u03b1-bromoisobutyrate moieties could also stabilize inverted (W/O) or double (W/O/W) emulsions, and surface initiated ATRP (SI-ATRP) occurred from the surface of the CNCs. Thus, capsules, filled beads, and microporous polymers were directly prepared. [87] Similarly, silica nanoparticles modified with \u03b1-bromoisobutyrate groups stabilized Pickering emulsions and acted as initiators in SI-ATRP. [88,89] Biphasic grafting from both the aqueous and organic phase resulted in Janus particles, since the in situ formation of amphiphilic particles restricted their own rotation. [90].Common anionic surfactants were previously considered incompatible with ATRP, since they could displace the halide ion from the X\u2013CuII/L deactivator, forming a CuII species that cannot deactivate radicals. [22,91] However, the detrimental displacement of X\u2212 was minimized by adding an excess of bromide ions in miniemulsion ATRP systems. This has opened up the possibility of employing inexpensive, effective, and readily-available surfactants, such as sodium dodecyl sulfate (SDS), instead of more costly, neutral surfactants, e.g., Brij 98, that were previously used in miniemulsion ATRP. [91].Catalysts are selected according to the type of dispersed media. In miniemulsion polymerizations, the hydrophobicity of the catalyst influenced the partition of the activator and deactivator between the organic and aqueous phases, thus affecting the rate of polymerization and degree of control. [92] The prevalent strategy toward well-controlled ATRP in miniemulsion involved the design of ligands for sufficiently hydrophobic Cu catalysts that resided within hydrophobic monomer droplets, where the polymerization proceeds. [93] Therefore, highly hydrophobic and highly active catalysts bearing hydrophobic alkyl chains in the polydentate amine were synthesized and effectively employed in low-ppm amounts in miniemulsion ATRP. [94,95] Only recently, the need for designing specific, hydrophobic ligands was overcome by demonstrating that the commercially available tris(2-pyridylmethyl)amine (TPMA) is a suitable ligand for miniemulsion ATRP when used in combination with an anionic surfactant, such as SDS.Interestingly, the strong interaction between dodecyl sulfate ions (DS\u2212), and Cu-based ATRP catalysts was later exploited to identify a Cu complex that, in the presence of SDS, could lead to enhanced polymerization control. The interaction between SDS and Cu/TPMA formed a catalyst that was conveniently partitioned between aqueous phase, surface of hydrophobic monomer/polymer droplets, and inside the hydrophobic droplets to better control the polymerization within the hydrophobic phase. The presence of SDS affected the localization of hydrophilic Cu/TPMA that would have been otherwise present almost exclusively in the water phase. It was demonstrated that in miniemulsion systems composed of BA droplets stabilized by SDS, only 4% of Cu/TPMA was located in the continuous aqueous phase, while most of it (95%) was located at the interface of the droplets. Moreover, a small portion of Cu complex (1%), formed ion-pairs with DS\u2212 capable of entering the polymerizing particles (Scheme 6\n). This small amount of hydrophobic ion pair has high mobility inside the particle and can effectively deactivate growing radicals. Thus, the interaction between DS\u2212 and Cu/TPMA provided an \u201cintelligent\u201d catalyst that could control radical propagation from the interface and the inside of hydrophobic droplets. [96] In addition, after polymerization, a simple dilution of the system with water followed by centrifugation induced the migration of the hydrophilic Cu/TPMA back to the water phase, recovering polymers with residual Cu content as low as 0.3\u202fppm. [97].Beyond Cu-catalyzed systems, Fe catalysts, including N,N-butyldithiocarbamate ferrum [98,99], Fe/N,N,N\u2032,N\u2032-tetramethyl-1,2-ethanediamine [99,100], and Fe/ethylene diamine tetraacetic acid [101], were employed for ATRP in dispersed media, either in miniemulsion or microemulsion systems. Metal-free ATRP employing 10\u2010phenylphenothiazine as a photocatalyst was successful in microemulsion ATRP. [102].The monomers used in oil-in-water dispersed media ATRP are generally hydrophobic. A slight change in hydrophobicity (or hydrophilicity) results in different polymerization rates and control. The hydrophilicity of molecules, including monomers, is typically quantified by considering its partition coefficient (logP) in an octanol-water mixture. In emulsion ATRP, the monomer diffuses to the micelles under moderate stirring, thus, the solubility of the monomer affects the rate of diffusion, which in turn impacts the rate of polymerization and consequently the degree of control. Emulsion ATRPs of MMA, ethyl methacrylate (EMA), butyl methacrylate (BMA) and lauryl methacrylate (LMA) were carried out. Monomers with lower logP values, i.e., more hydrophilic monomers, revealed faster rate of polymerization yet lower degree of control. Additionally, more hydrophilic monomers and polymers migrated between the droplets, leading to reduced colloidal stability. [103] On the other hand, the most hydrophobic monomer, LMA, showed negligible conversion because it could not diffuse though water to reach polymerizing micelles. [104].Hydrophilic monomers were polymerized in inverse emulsion systems, i.e., water-in-oil emulsions. For example, oligo(ethylene glycol)methyl ether methacrylate (OEOMA) was polymerized in inverse miniemulsion and inverse microemulsion ATRP, forming well-defined brush-like structures. [105\u2013107].ATRP is a versatile technique that enables the polymerization of monomers bearing various functional groups. For instance, glycidyl methacrylate (GMA) [108] and 2,2,3,3,4,4,4-heptafluorobutyl acrylate were polymerized in ab initio emulsion ATRP, with keeping intact the epoxy and fluorine functionalities, respectively. [109].As discussed in Section 2.1, traditional emulsion polymerization requires an aqueous initiation and nucleation phase, followed by polymerization within hydrophobic particles. Thus, in emulsion ATRP the Cu complex should \u201cfollow\u201d the radicals from the aqueous phase to the organic phase to control the entire dynamic ATRP process. Thus, the partition of the catalysts is critical. However, several common ATRP catalysts are highly hydrophilic, resulting in the deactivating species, X\u2013CuII/L, leaving the oil phase, negatively affecting control over the polymerization. At the same time, specifically designed hydrophobic catalysts also performed poorly because they mostly resided in the monomer reservoir and therefore could not control the polymerization in the aqueous phase. During the past decade, three main approaches were developed to perform well-controlled emulsion ATRP: (i) a microemulsion (or miniemulsion) \u201cseed\u201d approach, (ii) the use of a phase-transfer catalyst, and (iii) the engineering of the surfactant and ATRP catalyst.The seeded emulsion approach enabled to \u201cbypass\u201d the aqueous initiation and nucleation phase by introducing an initial microemulsion (or miniemulsion) step (Scheme 7\na). [110,111] Thus, first microemulsion (or miniemulsion) ATRP was performed using a tiny amount of monomer with the aim of encapsulating the hydrophobic catalyst complex into the polymer particles. Then, the latter served as \u201cseeds\u201d that were swelled by the addition of a large batch of monomer, during which the catalyst remained located in the particles. This two-step approach yielded well-controlled polymers and relatively uniform latex particles. Moreover, it could be further used to form a variety of structures, including block copolymers, hairy nanoparticles [112], and onion-like structures [113,114], as it will be described in Section 4. Seeded emulsion polymerization tend to have particle size with low batch-to-batch variability.Similar to the case of miniemulsion polymerizations, highly hydrophobic ligands were employed in emulsion systems to limit the escape of CuII deactivators from growing polymer particles. However, a large fraction of these hydrophobic catalysts resided in the oil phase rather than in the micelles during the first stage of the process, leading to a miniemulsion-like mechanism and poor polymerization control in the initial nucleation stage. The location of the catalyst could be modulated by using phase transfer catalysts in combination with shuttle molecules. Shuttle molecules were polar organic molecules, such as acetone, that was mixed with water to aid the solubility of several components in the continuous aqueous phase in an emulsion/miniemulsion ATRP. [115] Phase-transfer catalysts, typically organic ions, were used to transport the catalysts within the phases. For example, tetrabutylammonium bromide, TBAB, was shown to favor the mobility of the catalyst as well as the initiator and halide ions to the hydrophobic polymerizing particles (Scheme 7b). [115].To localize the catalyst within the micelles and surfactant-stabilized hydrophobic particles, a surfactant-ligand (SL) compound was specifically designed. The SL comprised a multidentate amine-based ligand for Cu centers, attached to an hydrophobic moiety (Scheme 7c) [83]. However, the diffusion of CuII species coordinated to the SL was relatively slow, likely due to the steric hindrance within the SL. Thus, the addition of a hydrophobic catalyst (e.g., CuBr2/dNbpy) was needed to promote sufficiently fast diffusion within the micelles.In a complex medium such as an emulsion, the key for excellent polymerization control is the presence of a dynamic catalyst that can react with propagating radicals throughout the life of a polymer chain, i.e., from aqueous nucleation to hydrophobic propagation inside monomer droplets. A simpler and scalable approach consists of exploiting the combination of interfacial and ion-pair catalysis provided by the Cu/TPMA-SDS system described in Section 3.1 to achieve ATRP in true ab initio emulsion. [27] The hydrophilicity of Cu/TPMA and the use of a hydrophilic alkyl halide initiator enabled initiation of the polymerization within the aqueous phase, which then seeded SDS-micelles (Scheme 8\n). The much higher total surface area of the micelles relative to the monomer reservoir caused the Cu/L complexes interacting with the surfactant to preferably reside on the surface or within the micelles. The anionic surfactant acted as a shuttle for the catalyst, promoting the localization of the catalyst at the interface of the hydrophobic particles, and to a lower extent inside the particles. This emulsion ATRP technique is facile, scalable, and it was successfully adapted to photoATRP. [104] Note that in this emulsion ATRP technique, pre-partitioning of catalyst and initiator was prevented by avoiding pre-mixing of the oil and water phases prior to starting the polymerization (Scheme 8). [27] A pre-emulsified monomer could potentially be fed as monomer reservoir, but this is not been tested yet.Prior to the development of low-ppm ATRP techniques, normal, reverse, simultaneous reverse and normal initiation (SR&NI), and activator generated by electron transfer (AGET) ATRP, were performed in various dispersed media using high loadings of Cu complexes. Low-ppm techniques such as ICAR and ARGET ATRP enabled to reduce the loading of Cu in dispersed media polymerization to hundreds of ppm. ATRP techniques based on external stimuli [116] such as light, electrical current/potential and ultrasound opened new possibilities for low-ppm ATRP in dispersed media, providing additional tuning of the polymerization rate and features by simple manipulation of the external stimulus.The main challenge hindering the implementation of eATRP in miniemulsion systems was identified as the physical disconnection of the working electrode (i.e., the electrode that provides the electrons for the reduction of CuII species) from the organic phase. [117] The electrode was instead in contact with the continuous aqueous phase. This hampered the regeneration of the hydrophobic catalysts generally employed in miniemulsion ATRP that dissolved in the organic phase. In fact, due to its \u201cmini-bulk\u201d feature, the mass transport in miniemulsion polymerizations is negligible. To solve this issue, a dual-catalyst approach was developed, whereby a water-soluble catalyst CuII/Laq and a hydrophobic catalyst CuII/Lorg were used at the same time, so that the first could shuttle the external stimulus from the working electrode to the droplets\u2019 interface, while the second was controlling the ATRP equilibrium inside the hydrophobic phase (Scheme 9\n). By comparing several combinations of CuII/Laq and CuII/Lorg, it was determined that the hydrophilicity/hydrophobicity of the catalysts played a more important role on polymerization control relative to the activity of the catalysts (i.e. its standard reduction potential). Later, the development of the catalytic system based on Cu/TPMA interacting with SDS enabled to simplify miniemulsion eATRP by eliminating the need for a dual-catalyst system. [96].An electrochemically mediated ATRP approach was also used to prepare molecularly imprinted polymer (MIP) nanoparticles, through a precipitation polymerization system. eATRP of 4-vinylphenylboronic acid in the presence of ethylene glycol dimethacrylate (EGDMA) as crosslinker and sialic acid as template was conducted in water/methanol (1/4 v/v), catalyzed by Cu/TPMA. By tuning the applied potential, the polymerization yielded nanoparticles with hydrodynamic diameter ranging from 160 to 330\u202fnm, capable of recognizing the sialic acid template. [118].The possibility to perform polymerizations in dispersed media through light irradiation could offer improved reaction efficiency, lower energy consumption, and increased safety. [119] However, photochemistry in dispersed-phase polymerizations is challenging because of the limited light penetration in turbid (mini)emulsion systems. Light absorption and scattering phenomena in miniemulsion photopolymerizations have been extensively studied by aid of theoretical modeling [120] and actinometry [121]. Droplet size played a crucial role, with smaller particles reducing the scattering coefficient and thus resulting in improved light penetration. [122] Despite these limitations, photoinduced miniemulsion free radical polymerizations have been successfully performed and even employed for the encapsulation of pigments in UV-cured nanoparticles. [10,123].After the application of the Cu/TPMA-SDS catalytic system (see Section 3.1) in miniemulsion eATRP and ARGET ATRP, photoATRP was investigated. Despite the turbidity of the heterogeneous media. photomediated miniemulsion ATRP was performed over a broad range of solid contents and particle sizes, achieved by tuning the surfactant amount. [124] In addition, excellent temporal control was achieved upon switching the UV light on and off multiple times. PhotoATRP was applied in ab initio emulsion polymerization of various methacrylate monomers, also by using an enzymatic degassing procedure. [104].Metal-free photoATRP was also conducted in heterogeneous media. For example, by using Eosin Y as a photocatalyst and triethylamine as electron donor, well-controlled PMMA was prepared via photoinduced electron transfer (PET)-ATRP of MMA in Pickering emulsions, stabilized by cellulose nanocrystals. [86].ATRP in dispersed media was performed by means of microwave irradiation and ultrasound. The use of microwave irradiation in ATRP generally results in significantly increased reaction rates and yields in comparison with other techniques. [125] The acceleration is attributed to the increased temperature and mass transport. In emulsion polymerization, microwave irradiation afforded nanoparticles with smaller average size and narrower size distribution compared to conventional heating, which typically yielded nanoparticles with size >100\u202fnm. Conversely, combining microwave irradiation and emulsion ATRP, PEG-b-PSt nanoparticles with diameters in the range 30\u201350\u202fnm were produced. [126].The use of ultrasound is a powerful strategy that enabled miniemulsion polymerization without chemical radical initiators or co-stabilizers, while maintaining fast polymerizations. [127] Acoustic cavitation provided radicals that sustained the polymerization, while the high shear forces limited the impact of Ostwald ripening. The interfacial and ion-pair catalyst system composed by Cu/TPMA-SDS was applied to miniemulsion sonoATRP. [128] Radicals generated in the aqueous phase by sonication effectively initiated and sustained the polymerization, which could be temporally controlled by switching ultrasound on and off.RDRP systems typically require a physical deoxygenation process, such as nitrogen bubbling or freeze-pump-thaw. The coupling of chemical deoxygenation reactions with RDRP simplified the reaction setup and allowed for conducting polymerizations in open-to-air conditions. [129] Glucose oxidase (GOx)-catalyzed oxygenation of glucose consumes oxygen efficiently, and thus it was applied to aqueous RAFT polymerization and ATRP. [130\u2013135] The GOx-deoxygenation system was also implemented in miniemulsion and emulsion ATRP, leading to well-controlled synthesis of hydrophobic polymers. [104,136] Circular dichroism measurements demonstrated that the structure of GOx remained intact in the presence of anionic surfactant. [136] Lignin nanoparticles (LNPs) coated with chitosan and GOx enabled efficient stabilization of Pickering emulsions and simultaneous in situ enzymatic degassing of ATRP, without requiring hydrogen peroxide scavengers. [137] The enzymatic degassing eliminated the possibility of monomer evaporation during traditional degassing; moreover, the low cost of the stabilizer and deoxygenation reagents can favor the implementation in industrial settings, especially in the case of emulsion polymerizations.The biphasic nature of heterogeneous polymerizations can simplify catalyst removal thanks to the large surface area of the organic/water interface. In fact, in miniemulsion systems, the large interfacial surface facilitates the mass transport of the catalyst from polymer particles to the aqueous phase, provided that a sufficiently hydrophilic catalyst is used. At the end of a miniemulsion ARGET ATRP of BMA, the product was precipitated into methanol/water (1/1 by v/v) and analyzed by inductively coupled plasma-mass spectrometry (ICP-MS). Using the highly hydrophilic catalyst Cu/TPMA, the residual Cu in the polymer could be as low as 300 part per billion (ppb), which was 10 times less than the residual Cu obtained using the hydrophobic Cu/BPMODA* (BPMODA*\u00a0=\u00a0bis[2-(4-methoxy-3,5-dimethyl)pyridylmethyl]octadecylamine) catalyst. [97].Another strategy for Cu removal consists of destabilizing the ligand. After Cu/HMTA (HMTA\u202f=\u202fhexamethylenetetramine) catalyzed emulsion ATRP of MMA, HMTA was decomposed into NH3 and HCHO under acidic conditions (the optimal pH was 4\u20135) and CuII was released to the aqueous phase. [138].Moreover, Cu can be conveniently removed by electrolysis. This method is especially efficient in polymer latexes due to the large organic/water interface area. Electrolysis was used to remove Cu from emulsion ATRP systems that employed a surfactant-ligand (SL, see Section 3.1) After polymerization, >98% of Cu was collected by electrolysis of the system that employed the SL, while only 50% Cu was recovered from a similar system employing a conventional hydrophobic catalyst. The improved efficiency was attributed to the anchoring of Cu/SL on droplets\u2019 surface, which decreased the diffusion resistance. Importantly, the latex stability was retained during the electrolysis. Purified polymer showed higher strength and higher antiaging performance than the untreated counterpart, as demonstrated in tensile tests. [50].Polymers and copolymers with different compositions and topologies were prepared by ATRP, owing to the retention of chain-end functionality and the possibility of introducing multifunctional moieties (i.e., crosslinkers, mono/multi-functional (macro)initiators, inimers, macromonomers). In dispersed media, the polymerization confinement in droplets of limited size contributes to decreased chain-termination events and can be exploited to tune the polymer architecture.Linear polymer chains represent the most common polymer architecture in ATRP. By tailoring the composition of monomer feed(s) and reaction conditions, statistical, block, and gradient copolymers could be synthesized in ATRP in dispersed media. Here we provide examples of block and gradient copolymers.In ATRP, block copolymers are typically prepared by first polymerizing the monomer that gives a more active chain end (e.g., polymethacrylates or polystyrene macroinitiators), followed by polymerization of less active monomers (e.g., acrylates), in order to assure a good initiation efficiency for the second block. However, this sequence cannot be followed when preparing certain A-B-A triblock copolymers. Therefore, halogen exchange is typically used as an efficient way to chain-extend from a less active macroinitiator (MI) to a more active monomer. With halogen exchange, the limitation of mismatching monomer reactivity can be circumvented by switching from C\u2013Br to less active C\u2013Cl chain ends. This has been achieved by using CuICl/L in equimolar amount to Pn-Br MI in the chain-extension step. [139\u2013141] However, this approach cannot be effectively applied in systems based on activator regeneration, since they operate with ppm amounts of catalysts. Thus, catalytic halogen exchange (cHE) was developed [142] and later implemented in miniemulsion ARGET ATRP to chain-extend a less active PBA-Br MI with a more active MMA monomer, using a catalytic amount of Cu (Scheme 10\n). [143] Addition of 0.1\u202fM NaCl or tetraethylammonium chloride to ATRP of MMA initiated by methyl 2-bromopropionate led to high initiation efficiency and polymers with low dispersity. Similar conditions were then employed in chain extension of PBA-Br MI with MMA to prepare P(BA-b-MMA) and P(MMA-b-BA-b-MMA). This technique allows for building various block copolymers with different structures and functionalities.Moreover, block copolymers could be synthesized via either in situ or stepwise chain extension. After miniemulsion eATRP of BA reached 78% conversion, in situ chain extension was achieved by dispersion of tBA into the miniemulsion under ultrasonication followed by N2 sparging and electrolysis, resulting in PBA-b-P(BA-co-tBA). [96] In stepwise chain extension, a solution of PBA-Br MI was isolated by precipitation from the miniemulsion system, then tBA was used in the organic phase of a second miniemulsion. The subsequent ATRP resulted in P(BA-b-tBA). [96].Copolymerizations of monomer mixtures by ATRP can result in statistical or gradient copolymers, depending on monomer reactivity [144,145], but also feeding rate, and hydrophilicity. For monomers with different reactivity, such as acrylates and methacrylates, spontaneous gradient copolymers could be prepared by miniemulsion ATRP. [146] Conversely, for monomers with similar reactivity it was necessary to feed one monomer into miniemulsion polymerization media to produce \u201cforced\u201d gradient copolymers. [146,147] On the other hand, in emulsion ATRP one can exploit the different water solubility of monomers with similar reactivity to obtain spontaneous gradient copolymers. The monomer with higher water solubility will diffuse more rapidly through the aqueous phase, therefore being incorporated first into the copolymer in comparison with the less water-soluble monomer. Thus, P(MMA-grad-BMA) and P(BMA-grad-LMA) (LMA\u202f=\u202flauryl methacrylate) could be prepared by emulsion ATRP of the corresponding monomer mixtures, with no need for adopting a feeding strategy. [104,148].Compared to linear polymers, the distinct architecture and multiple chain-terminal groups of branched polymers endow higher solubility, lower solution/melt viscosity, less deformability, and more chain-end functionality. [149] In bulk and solution polymerizations, especially at high monomer concentrations, the increased number of initiation sites favors the occurring of crosslinking reactions, which can lead to macroscopic gelation. By switching from homogeneous systems to dispersed media, cross-termination reactions can be greatly decreased owing to the segregation of growing polymer chains (Scheme 11\na), thus high conversions and low polymer dispersity can be achieved more easily. This feature is especially beneficial for the synthesis of bottlebrush, star, and hyperbranched polymers (Scheme 11b), as it will be described in the next paragraphs.Molecular brushes, also known as bottlebrushes, comprise densely grafted side chains, allowing for decreased intermolecular entanglement and for the presence of multiple functionalities in the side chains. [150\u2013153] This unique structure makes molecular brushes suitable for application as lubricants and surfactants, among many others. There are three methods to synthesize bottlebrush polymers by ATRP: \u201cgrafting-through\u201d, \u201cgrafting-from\u201d, and \u201cgrafting-onto\u201d.\u201cGrafting-through\u201d refers to a process where an oligomer/polymer chain bearing a vinyl group at one end is polymerized by ATRP into a bottlebrush structure. For example, AGET ATRP of OEOMA475 (OEOMA with average MW 475) was initially conducted in aqueous solution. [105] Continuous feeding of AsAc and increasing monomer concentration resulted in higher conversion, but final polymers showed bimodal distribution of MW caused by bimolecular termination. Significantly improved polymerization control was obtained upon switching to an inverse miniemulsion system, where chain segregation effectively reduced the chances of termination reactions. The resulting polymers had desired MW and low dispersity.The \u201cgrafting-from\u201d method employs a polymer backbone with multiple initiation sites (i.e., a multifunctional macroinitiator). ATRP was initiated from these sites to form densely packed side chains. However, in normal and AGET ATRP in solution, gelation typically occurred at 20\u201330% monomer conversion. Conversely, in miniemulsion systems eventual crosslinking occurs within the latex particles, with limited effect on the fluidity of the miniemulsion system even when monomer conversion reaches >80%. [97].Finally, \u201cgrafting-onto\u201d could be conducted by attaching clickable functional groups to the backbone, followed by performing a click reaction. This approach was not yet used in dispersed media.Star polymers represent a class of branched architectures with linear \u201carms\u201d connected to a central branching point, typically referred to as the \u201ccore\u201d. [154\u2013156] Similarly to the \u201cgrafting-from\u201d approach for the synthesis of bottlebrushes, star polymers have been mainly prepared via the \u201ccore-first\u201d approach. Compounds with multiple hydroxyl groups (e.g., cyclodextrin, glucose, tannic acid) can be transformed into (multi)functional ATRP initiators by substituting hydroxyl groups with C-X functionalities, typically a-bromoisobutyrate groups. Cyclodextrin-based ATRP initiators with 14\u202fC\u2013Br sites were used to prepare stars with PBA and PBMA arms in miniemulsion via ARGET ATRP, using the Cu/TPMA-SDS catalytic system. [97].Highly branched polymers, including dendrimers and hyperbranched polymers, possess highly compact structure, high branching density in the backbone, and numerous periphery groups, leading to many interesting properties, such as high solubility, low viscosity, high functional group density, and potential for cargo loading and release. [157] Dendrimers are regularly branched polymers with a dendritic, tree-like structure. [158,159] Hyperbranched structures do not necessarily have regular branches, however, in contrast to dendrimers, they are prepared via inexpensive one-pot synthesis while retaining highly branched architectures, high solubility, low viscosity and high functional group density. Hyperbranched structures with controlled number of branching sites were prepared by ATRP, through the copolymerization of an inimer (a monomer with an initiating group) with one or more conventional monomers. [160\u2013162] The preparation of hyperbranched polymers via microemulsion ATRP enabled faster kinetics and the generation of polymers with higher MWs, narrower MW distributions and through faster polymerization processes than in typical solution polymerizations, where high dilution and low monomer conversions were needed to avoid gelation. [102].Microemulsion ATRP was also exploited to prepare hyperstar polymers, i.e., core\u2013shell structured star polymers that contain a highly branched polymer as the core and densely grafted radiating arms. This required the polymerization of an inimer to form an hyperbranched core with abundant C\u2013Br initiating sites for the subsequent growth of radiating arms through the addition of second monomer, in a one-pot microemulsion ATRP process. [163,164] When hydrophobic BA was employed as second monomer, BA molecules diffused into the latexes and swelled the hyperbranched polymers to form a seeded emulsion, minimizing hyperstar-hyperstar coupling events (Scheme 12\n). On the other hand, the use of a zwitterionic monomer, cysteine methacrylate, as second monomer enabled to stabilize the growing hyperbranched stars, owing to the electrostatic repulsion between charged arms and stars, which also avoided coupling at high conversion.The heterogenous nature of dispersed media has proven to be an excellent platform to produce nanostructured materials, generated either via (i) self-assembly of block copolymer surfactants, (ii) crosslinking polymers inside the dispersed phase, and (iii) exploiting the peculiar properties of the oil-water interphase.Post-polymerization self-assembly of pre-formed amphiphilic block copolymers with incompatible blocks typically occurs in highly diluted copolymer solutions to form microphase separated structures. Instead, polymerization induced self-assembly (PISA) in aqueous dispersion with relatively high solid content (up to 50\u202fwt%) is characterized by simultaneous polymerization and self-assembly by using a soluble macroinitiator that also acts as a stabilizer as the polymerization of a second soluble monomer proceeds. Upon reaching a critical chain length, the second block becomes insoluble, driving the reorganization of the block copolymer into a variety of nanoobjects via a dispersion polymerization approach. [165] RAFT polymerization has been frequently coupled to a PISA approach. [166] In contrast, the application of ATRP in PISA is limited by the partitioning of Cu complexes in the complex dispersed system.ICAR ATRP with low Cu concentration was carried out using POEOMA-Br as the macroinitiator and stabilizer and poly(benzyl methacrylate) (PBnMA) as a core-forming block. [167] The system was homogenous in ethanol, but phase separated upon polymerization. Distinct architectures were obtained either at room temperature or at 65\u202f\u00b0C, i.e., below and above the T\ng of PBnMA, respectively (Scheme 13\n). Another core-forming monomer, glycidyl methacrylate, underwent ring-opening reaction during PISA ATRP, allowing for the in situ formation of crosslinked nanoparticles. [168] Protein-polymer conjugates could be synthesized during ATRP induced self-assembly. The aqueous dispersion polymerization of 2-hydroxypropyl methacrylate (HPMA) from a hydrophilic protein (human serum albumin, HSA) modified with ATRP initiating groups yielded HSA-PHPMA nanospheres and vesicles. During the PISA process, a model green fluorescent protein (GFP) was encapsulated in situ, and the polymeric architecture enabled enhanced intracellular GFP delivery. [169].Multi-layer polymer particles could also be prepared by means of hydrophobic multiblock copolymers with incompatible compositions prepared in dispersed media. Sequential miniemulsion ATRP and in situ seeded emulsion ATRP yielded poly(isobutyl methacrylate)-b-polystyrene P(iBMA-b-St) exhibiting a core-shell structure. Upon subsequent alternating addition of iBMA and St, an onion-like, alternating multilayered morphology was observed. [113,114] Similar polymer-vesicle latex particles were prepared by starved feeding of MMA to Br-modified vesicles particles. [170].Chemically crosslinked nanoparticles are more stable against external stimuli and mechanical processing compared to self-assembled structures. The most straightforward recipe to prepare chemically crosslinked nanonetworks via polymerization in dispersed media is to incorporate a crosslinker, e.g., EGDMA. 2-Hydroxyethyl methacrylate (HEMA), 4-vinyl pyridine (4-VP) and PEGDMA were copolymerized in an inverse emulsion, forming a pH-sensitive hydrogel that underwent multiple swelling-deswelling cycles. [171] Similar PHEMA-POEOMA nanogels with uniform and controllable sizes were prepared in inverse miniemulsion by AGET ATRP, where the hydroxy groups from the HEMA moiety allowed for introducing photoinitiation sites for subsequent photopolymerizations. [172].In comparison with crosslinked nanoparticles prepared by conventional free radical polymerization (FRP), those made by ATRP had narrower molecular weight distribution, sharper glass transition [173], and improved loading efficiency. [174] Degradable nano-networks were prepared by ATRP utilizing a disulfide-containing crosslinker. [175\u2013177] Cross-linkable monomers such as allyl methacrylate were incorporated into a block copolymer, where the allyl group could crosslink under UV irradiation, forming network structures. [178] Alternatively, other techniques such as ring opening metathesis polymerization (ROMP) and click-chemistry were conducted simultaneously with heterogeneous ATRP, forming crosslinked networks in one pot. [179\u2013181].Nanoparticles consisting of a solid core and some polymer chain \u201chair\u201d attached to the core surface are termed \u201chairy nanoparticles\u201d. Common core materials include metal/metal oxide, silica nanoparticles, and polymer networks. Polymer network nanoparticles, generally prepared by emulsion/miniemulsion polymerization and functionalized with ATRP initiating groups were re-dispersed in monomer solution to form hairy nanoparticles by SI-ATRP. For example, ATRP initiating groups were immobilized onto P(St-co-HEMA) microspheres, enabling the grafting of 2,2,6,6-tetramethyl-4-piperidyl methacrylate. [182] Hydrophilic polymer chains were grafted from hydrophobic cores, yielding amphiphilic hairy nanoparticles with poly(ethylene glycol) or polyzwitterionic-based corona (Scheme 14\n). [183].Nanocapsules have attracted much interest due to the encapsulation and controlled release capability provided by their hollow structure. [184,185] Nanocapsules are typically prepared in miniemulsion polymerizations, by confining the polymerization within the water/oil interface. [186] PEO-PBMA-Cl, an inisurf bearing an ATRP initiating group in the hydrophobic fragment, was used to initiate miniemulsion ATRP of BMA at the water/monomer droplet interface, with polymer chains slowly growing inward in a controlled manner. In the presence of a crosslinker and an organic solvent, polymeric nanocapsules with high stability and good dispersibility in organic solvents were prepared. [187] The outer surface of the nanocapsules was further modified when using a difunctional reactive surfactant. Using N3-PEO-PBMA-Cl yielded functional nanocapsules with N3 moieties on the surface, which allowed for attaching organic dye probes or additional polymers via click-chemistry. [188].Similar nanocapsules were fabricated via Pickering emulsion ATRP, initiated at the oil-water interface. Materials with different morphologies were formed via either FRP or ATRP in Pickering inverse emulsion stabilized by CNCs and initiator-stabilized CNCs, respectively (Scheme 15\n). [189] FRP formed beads, whereas ATRP resulted in hollow structures due to the initiation solely occurring from the oil/water interface. Similarly, when ATRP was performed on silica nanoparticles functionalized with initiator and immobilized at the interface of Pickering emulsions, a polymer network grew inwards, although it did not fill the whole available space of the polymerizing internal phase. [190] Therefore, the resulting polymer-inorganic hybrid has the shape of hollow capsules, which could be processed as semipermeable membranes that served as microdevices for drug or cell delivery.Polymer-inorganic composites and hybrid materials combine the functionality and flexibility of polymers with the high strength of inorganic materials. Polymer-inorganic composites are systems where polymer and inorganic phases are mixed, and they are typically both present in aggregates of large dimensions. However, simple blending of inorganic materials and hydrophobic polymers often leads macroscopic phase separation, leading to difficulties in processing, inadequate structure heterogeneity, and poor mechanical properties. [191].Conversely, polymer-inorganic molecular hybrids are materials in which chemical bonds are established between the constituents, so that mixing is effectively occurring at the molecular level, and aggregation and surface defects are minimized. [192] Polymer-inorganic composites have unique properties, such as good mechanical and thermal stability, gas barrier performance, and flame retardancy. [193\u2013195].Polymer-inorganic hybrids can be formed via miniemulsion ATRP from the surface of inorganic materials, which are well-dispersed in monomer droplets thanks to high-shear sonication; this allows for encapsulating inorganic materials in polymer latexes via covalent bonds. [193,196\u2013198] SI-ATRP is a powerful tool to fabricate polymer-inorganic hybrids with improved properties, however polymerizations in solution are generally limited to low monomer conversion (\u223c10%) and/or conducted in the presence of sacrificial initiators to avoid interparticle crosslinking and macroscopic gelation. [11,199,200] In contrast, the segregation of polymerization loci in miniemulsion SI-ATRP enables to reach high conversion without gelation. Polymers were grafted from CdS quantum dots [196], montmorillonite nanoclay [201], and silica nanoparticles [202] via miniemulsion SI-ATRP, yielding well-defined particle brushes. In emulsion ATRP, the electrostatic interaction between a negatively charged P(AA-co-BA)-Br macroinitiator and positively charged Gibbsite platelets facilitated the good dispersion and alignment of the platelets in the resulting polymer matrix, forming \u201cmuffin\u201d-like encapsulated Gibbsite structures. [203\u2013205].Janus structures are another example of complex polymer architecture, which consist of particles whose surfaces display two or more distinct physical properties. [206] Janus structures can be applied as particulate surfactants, imaging nanoprobes, and self-propelled colloidal materials capable of \u201csmart\u201d motion.Self-assembly of incompatible (co)polymers in emulsion is one strategy to prepare Janus nanoparticles (Scheme 16\na). For example, non-functional PMMA and functional P(St-BIEM) (BIEM\u202f=\u202f2-(2-bromoisobutyryloxy)ethyl methacrylate, an ATRP initiator) were emulsified together and self-assembled into Janus composite particles during solvent evaporation. P(St-BIEM) accumulated on one side of the particle. Subsequently, poly(2-(dimethylamino)ethyl methacrylate) was grafted from the surface area occupied by localized C\u2013Br initiation sites of P(St-BIEM), forming \u201cmushroom\u201d particles with controllable morphology. [207].In a Pickering emulsion, the Pickering agent itself can be transformed into a Janus particle via ATRP. Graphene oxide (GO) platelets, modified with an ATRP initiator, served as Pickering agents in a toluene/water emulsion. ATRP of toluene-soluble 2-(acryloyloxy)ethyl ferrocenecarboxylate (MAEFc) occurred only on the side of the platelet exposed to the monomer solution. Subsequently, grafting of polydopamine led to Janus GO nanosheets. [209] In an emulsion stabilized by silica-Br nanoparticles and containing an hydrophilic and an hydrophobic monomer in the two phases (Scheme 16b), different polymer brushes were simultaneously grafted from the nanoparticles via SI-ATRP, forming amphiphilic Janus colloids with advanced emulsification properties. [90,208] The stable colloid structure prevented the particles from rotating during polymerization, giving rise to a clean Janus morphology.Highly porous polymer monoliths have high surface area and thus can find application in separation, catalysis, and extraction. [210] pHIPE is a representative example of highly porous polymer monolith prepared through an emulsion templating approach. [57,58] The highly interconnected porosity and high surface area renders pHIPE materials suitable as liquid droplet elastomers and templates for molecular recognition materials. [211,212] Various technologies have been employed for pHIPE formation, including FRP, RDRP, step-growth polymerization, click reactions, etc. Among these approaches, RDRP (including ATRP and RAFT polymerization) provides more homogeneous network structures. [58].Typical HIPEs are water-in-oil emulsions, where the hydrophobic monomer(s) is dissolved in the continuous oil phase, while the internal phase represents over 74% of the total volume. The type and locus of ATRP initiation affects the macromolecular structure of pHIPEs. [213] When a conventional oil-soluble ATRP initiator was used by dissolving it in the monomer phase, the final pHIPEs presented rather spherical voids. When nanoparticles functionalized with ATRP initiators were employed, Pickering HIPEs were obtained, in which the type and locus of initiation affected the porous and macromolecular structure of the resulting pHIPEs. If a highly organic-soluble nanoparticle ATRP initiator was used, then a pHIPE with rather spherical voids was produced and no preferential diffusion of monomer molecules to the interface was observed. Conversely, when a water-soluble conventional radical initiator was introduced in the system, then interfacial polymerization occurred, and polyhedral voids were formed. During interfacially initiated polymerization, the monomer diffused toward the oil/water interface and the polymerizing macromolecules \u201clocked-in\u201d the nanoparticles at the interface, affecting both the wall and pore structure.Low density and degradable pHIPEs are also of interest. Materials with very low density (0.06\u202fg/cm3) were prepared by using special star polymer surfactants by ATRP via an arm-first approach. The surfactants were active at a very low loading (<0.1\u202fwt%). [214] Degradable pHIPEs were prepared by incorporating disulfide crosslinkers in the network. [215] Bis(2-methacryloylxyethyl)disulfide (DSDMA) crosslinker was copolymerized with 2-ethylhexyl methacrylate by AGET ATRP catalyzed by the hydrophobic catalyst CuBr2/BPMODA (BPMODA\u202f=\u202fbis[2-pyridylmethyl]octadecylamine). The material had a uniform crosslinked structure which was degraded by tributylphosphine (Bu3P), and the resulting degraded product had M\nn\u202f=\u202f30,500 and \u0110\u202f=\u202f1.6. The low molecular weight and relatively low dispersity indicated the good degradability of the pHIPE.Similarly, ATRP was conducted in a medium internal phase emulsion (MIPE) system (where the internal phase is between 30% and 70%). The MIPE was stabilized by the synergy of Pluronic F127 and amphiphilic diblock glycopolymers. [216] The latter comprised a glycopolymer-block and a Br-terminated PSt block, thus serving as inisurf for AGET ATRP of styrene. The resulting pMIPE exhibited a biocompatible, homogeneous structure with bimodal pore size distribution, showing potential for use in catalysis and biomedical applications.The versatility of ATRP in dispersed media and the preparation of polymers with a broad range of architectures have opened the door to numerous applications. Polymers and copolymers made by ATRP in aqueous dispersed media have been tested for bio-related applications and for coatings. In the following paragraphs, several potential applications will be reviewed, including drug delivery, molecular recognition, bio-quantification, as well as the design of polymer coatings and films with electrocatalytic properties. These applications were supported by synthesizing smart (bio)materials with stimuli-responsive functionalities.Several complex polymer structures, (e.g., vesicles, nanogels, and hydrogels) hold promise as drug carriers and other biomaterials. Limited biocompatibility and hydrophobicity could be overcome by PEGylation of the nanoobjects. For example, upon incorporation of POEOMA chains by emulsion ARGET ATRP, the cytotoxicity of cationic nanogels was greatly decreased without compromising their antibacterial activity. [217].Polymer nanoparticles, including nanocapsules, nanospheres, and nanogels, are suitable materials for drug delivery. [218] Hydrophilic nanogels prepared by inverse miniemulsion/microemulsion ATRP have uniform network structure, and thus higher swelling ratios and better colloidal stability in comparison to analogous nanoobjects made by FRP. Moreover, they benefit from controlled degradability upon incorporation of functionalities with desired responsiveness. Nanogels made by ATRP in inverse mini/microemulsions could be loaded with star-branched polymer nanoparticles (via in situ covalent incorporation), with carbohydrates and proteins (via in situ physical incorporation), with fluorescent dyes, anticancer drugs, and gold nanoparticles (via physical incorporation). The versatility of cargos and the biocompatibility of the nanogels imparted great potential for targeted drug delivery applications. [219].On the other hand, hydrophobic molecules could be loaded into polymer particles made in O/W systems. Dispersion polymerization of PEG, 2-(diethylamino)ethyl methacrylate (DEAEMA), and tert-butyl methacrylate (tBMA) led to nanoparticles, where each monomer had a different function: the short hydrophilic PEG chains provided biocompatibility to the outer surface of nanoparticles, the hydrophobic PtBMA core facilitated the loading and release of hydrophobic fluorescein molecules, and the positively charged PDEAEMA could uptake negatively charged siRNA. Thus the cationic nanoparticles served as carriers for both nucleic acids and hydrophobic drugs. [173].The delivery of drugs could be achieved by incorporation of stimuli-responsive moieties, including thermo-responsive polymers, pH-responsive groups, degradable bonds, and magnetic Fe3O4. [220] Di(ethylene glycol) methyl ether methacrylate (M(EO)2MA) is a water-insoluble monomer, while its polymer has a lower critical solution temperature (LCST) of 25\u202f\u00b0C. The transition temperature of thermoresponsive P(M(EO)2MA-OEOMA-EGDMA) microgels could be tuned by changing the network composition. [221] The drug-releasing profile was controlled by either tuning temperature or by chemical reduction of disulfide bonds in DSDMA (Scheme 17\n), used in place of EGDMA. The magnetically loaded microgels were guided to particular body parts for the delivery of anesthetic drugs. [222].The recognition of molecules through devices with high reusability, high selectivity, and low ageing is essential for sensing and removal of toxic compound. Molecularly imprinted polymers (MIP)s were used for effective molecular recognition. [223] MIPs were prepared by copolymerization of functional monomers and crosslinkers in the presence of a template (i.e., the target molecule or a dummy molecule with a similar structure to the target molecule), followed by the removal of template molecules to generate tailor-made recognition sites, which resemble the shape, size and functionality of the template. The preparation of MIPs in dispersed media polymerization is advantageous over surface/film imprinting and surface graft imprinting because of the low toxicity, good dispersibility, and large adsorption capacity in dispersed media.Indole MIP was synthesized by emulsion ATRP via copolymerization of 4-VP and vinyl modified SiO2 nanoparticles in the presence of EGDMA as crosslinker. Indole was incorporated by hydrogen bonding with 4-VP. The resulting indole MIP had high specific area with an equilibrium adsorption capacity of 34.5\u202fmg/g, showing promise in removing indole from fuel oil. [224] Emulsion ATRP was used to synthesize superparamagnetic molecularly imprinted nanoparticles for selective recognition of tetracycline molecules from aqueous medium. Acrylate-modified Fe3O4 was copolymerized into the MIP, allowing for the recognition of tetracycline with a capacity as high as 12.1\u202fmg/g; furthermore, the material could be reused multiple times. [225] MIPs prepared via emulsion ATRP from multifunctional initiation sites of yeast presented fast recognition of ciprofloxacin, high adsorption, and high reusability. [226] Precipitation eATRP of 4-vinylphenylboronic acid in the presence of sialic acid as template molecule yielded nano-MIPs with morphology and size tuned by varying the applied potential. [118].Glycopolymers bearing carbohydrates are highly effective in protein recognition due to the multifunctional carbohydrate side chains. Glycopolymer-modified colloidal particles prepared by emulsion ATRP methods allowed for the recognition of specific proteins. Chlorine-modified PSt nanoparticles were synthesized by emulsion polymerization, followed by SI-ATRP, to incorporate glycol-modified PSt chains on the surface. The resulting particles coagulated in the presence of lectins, including Concanavalin A (Con A) or peanut agglutinin (PNA) through specific glycol recognition sites. [92] Glycopolymer-grafted nanoparticles were also casted into a film to recognize proteins on a surface. For example, a glycosylated amphiphilic block copolymer P(HEMAGl-b-BMA)-Cl was used as an inisurf (see Section 3.1) for BMA emulsion polymerization to prepare glycolsilated core-shell particles. The resulting latexes were cast into polymer films with bioactive surface owing to the presence of PHEMAGl capable of specific binding of Con A. [227].ATRP in dispersed media was also applied to the preparation of highly selective filtration material. Hydrophilic Fe3O4 particles used as Pickering agents were modified by ligand exchange with Br-containing carboxylic acid on one side, and subsequent ATRP of MMA led to amphiphilic, superparamagnetic Janus nanoparticles with excellent performance in oil purification. [228].Precipitation polymerization of N-isopropylacrylamide (NIPAAm) has been used to quantitatively detect hemoglobin, which acted as a catalyst for the ATRP of NIPAAm. The polymerization was conducted at 37\u202f\u00b0C (i.e., above the LCST of PNIPAAm), and simple measurement of the turbidity could reflect the amount of hemoglobin in the system (Scheme 18\na). The viability of the hemoglobin dose-turbidity formation rate assay both in solution and in physiological fluids proved the versatility of this method, which is also environmentally friendlier than established chemical assays for hemoglobin based on toxic reagents. [229].ATRP in dispersed media was translated from solution to surface polymerization for the preparation of polymer brushes. [231] A surface was functionalized with ATRP initiator and put in contact with an aqueous solution containing catalyst and hydrophobic monomer aggregates. Remarkably fast brush growth was observed, ascribed to the formation of monomer aggregates in the aqueous phase and to the beneficial role of hydrogen-bonding by interfacial water.Self-healing coatings via microencapsulation exploit the release of flowable material from a microcapsule to restore cracks or damages. These approaches were based on mixing ATRP components in a microreactor (i.e., the microcapsule) that induced polymerization upon release of its constituents. [232,233] Preparation of solvent-based microcapsules via Pickering emulsion templated interfacial ATRP (PETI-ATRP) involved the electrostatic deposition of a polyanionic ATRP initiator onto cationic nanoparticle surfaces, which yielded modified nanoparticles (Scheme 18b). [230] Subsequently, the microcapsules were synthesized by PETI-ATRP of N,N\u2032-methylene bisacrylamide to form the shell wall. The method allowed encapsulation of core solvents (xylene, hexadecane, and perfluoroheptane) with different solubility properties, and the microcapsule wall-forming chemistry afforded the use of different vinyl monomers.In another application, switchable latexes were prepared and employed in coating compositions that responded to changes in the environment. Using 1,1-(diethylamino)undecyl 2-bromo-2-methylpropanoate as an amine-bearing inisurf (Scheme 19\n), the resulting PMMA polymer latexes could switch between aggregated and dispersed states using CO2 and argon as triggers. [234].Crosslinked nanoparticles obtained in dispersed media were used to prepare catalytic films on electrodes. For example, surface-protected P(AN-b-BA) polymer nanoparticles were pyrolyzed into individual nanoporous carbon spheres with electrocatalytic properties. [235] These PAN-based nanoparticles were cast as an electroactive material with high surface area, for application as supercapacitors or for the oxygen reduction reaction. Surface-protected P(AN-b-BA) self-assembled polymer nanoparticles prepared by miniemulsion ATRP were pyrolyzed into individual nanoporous carbon spheres with better performance for CO2 capture with a higher CO2/N2 selectivity, and increased efficiency in catalytic oxygen reduction reactions, as well as improved electrochemical capacitive behavior, as compared to regular nanocarbon monoliths. [235,236].The last 15 years have witnessed important developments and increasing interest in performing ATRP in aqueous dispersed media for preparing well-defined polymers and nanoobjects. The availability of low-ppm ATRP techniques triggered the development of novel synthetic approaches for heterogenous ATRP that used lower Cu loadings, facilitating the removal of Cu from the final latexes. Moreover, the design of reactive surfactants and catalytic systems capable of overcoming the limitations dictated by the partitioning of most Cu complexes among different phases enabled improved polymerization control. Noteworthy is the development of effective approaches for ATRP in emulsion, including the seeded-emulsion strategy, phase-transfer catalyst, surfactant-ligand complex, and a combination of ion-pair and interfacial catalysis via hydrophilic Cu complexes and anionic surfactants. Traditional ab initio emulsion is much easier to employ at large scale compared to miniemulsion or microemulsion methods, which require high shear forces and high surfactant loading, respectively. Moreover, oxygen scavenging strategies and catalytic halogen exchange simplified the reaction setup for a broader range of polymer compositions and architectures.When performing ATRP in dispersed media, the segregation of polymerization loci into droplets of 50\u2013500\u202fnm diminished interparticle radical coupling reactions, thus high viscosity and macroscopic gelation were avoided, enabling to reach higher monomer conversion compared to homogeneous ATRP. This benefited the preparation of polymer bottlebrushes, hyperbranched structures, and inorganic-polymer hybrids with excellent control over polymer dispersity even at high monomer conversion. Various polymer architectures and topologies were achieved by exploiting self-assembly or crosslinking approaches, and by introducing inimers or multi-functional surfactants. The tunable composition of polymers allowed for their use for molecular recognition and bio-quantification. The hydrophobic environment within polymer nanocapsules or nanogels and the incorporation of stimuli-responsive functionalities enabled the loading and delivery of hydrophobic drugs. Nanocapsules were designed for potential application in self-healing coatings.Innovations in dispersed media ATRP are expected to come from improved mechanistic understanding, as recently illustrated by the design of the Cu/TPMA-SDS catalytic system, which outperformed most catalysts in terms of polymerization control and ease of product purification. This advancement eliminated the need for developing hydrophobic Cu complexes or dual-catalyst systems. In-depth studies of catalyst partitioning and interactions with other components can guide the rational design of reactive surfactants, ligands, or phase-transfer complexes.Mechanistic analysis could be promoted by computational studies and simulations, although examples of computational studies of dispersed media polymerizations are still quite rare. PREDICI, a kinetic-based modeling tool, was successfully used for modeling ATRP under homogeneous conditions [237,238], even for bottlebrush preparation. [239] However, only one report used PREDICI for the simulation of (semi)batch emulsion ATRP of styrene. [240] A Design of Experiment (DoE) approach was used to assess the influence of five independent variables (catalyst, initiator, temperature, reducing agent and surfactant loadings) on monomer conversion, polymer average molecular weights, and dispersity in AGET emulsion ATRP. Analysis of 5 fractional factorial experiments showed that temperature was the most influential factor. [241].The translation of academic research on ATRP in dispersed media into industrial applications depends on the multidisciplinary collaboration among chemists and engineers for reaction scale-up, as well as with biologists and medical researchers for the preparation of biomaterials and efficient drug-delivery systems. Roughly half of commercial coatings are prepared by dispersed media polymerizations, and several polymer-based products, including paints, creams, and medical treatments are sold in a dispersed state. The industrial use of heterogenous ATRP systems is promoted by simple and low-cost reactions setups, as in the case of ab initio emulsion ATRP, which could be readily integrated into existing reactors for emulsion FRP. Indeed, emulsion ATRP was already scaled up to the 2\u202fL volume. [242] Upon optimizing the reaction conditions, including temperature and reagent ratios, PMMA with dispersity of 1.17 was produced, without appreciable coagulation. [242,243] Despite the complexity of heterogenous emulsion ATRP systems, empirical models showed that temperature strongly affects the rate of polymerization, M\nn and \u00d0. Moreover, the surfactant amount and stirring speed affected the rate of emulsion ATRP, while the nature and amount of ligand mostly influences M\nn and \u00d0 of the formed polymer. [244] Improved understanding of the influence of various parameters, achieved through modeling and high-throughput experimentations, as well as continuous advancement in reaction design, is expected strongly promote the industrialization of ATRP in dispersed media in the coming years.The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: K Matyjaszewski reports financial support was provided by National Science Foundation.Fundings from the National Science Foundation (CHE 2000391 and DMR 2202747), the European Union \u2013 NextGenerationEU, and the University of Padua under the 2021 STARS Grants@Unipd programme \u201cPhoto-e-cat\u201d are gratefully acknowledged.", "descript": "\n                  Polymerizations in aqueous dispersed media benefit from good heat transfer, low viscosity, low content of volatile compounds, and established industrial use. Over the past two decades, the implementation of ATRP in dispersed media highlighted the possibility of achieving improved livingness and producing a variety of macromolecular architectures. With the introduction of several methods for activator regeneration in ATRP, catalyst loading has been greatly diminished, and the reaction setup has been simplified. The availability of ATRP techniques employing low-ppm (part per million) catalyst loadings enabled to access an even larger variety of polymer architectures and functional polymer particles, which have been applied for molecular recognition, drug delivery, bio-quantification, and advanced coatings. This minireview presents innovative synthetic approaches, polymer architectures, and relevant applications, as well as the challenges that remain to be overcome to promote the industrialization of ATRP in dispersed media.\n               "}
{"full_text": "Definition UnitDeactivation factor -Archimedes number -Carbon yield gC/gcat\nWeisz-Prater criterion -Effective diffusivity m/2Diameter of the catalyst particle mGravitational acceleration m/s2\nActivation energy of kinetic parameters kJ/molKinetic rate constant \n\n\n\n\nmol\n\n\n\n\nCH\n\n\n4\n\n\n\n\n/\n\n\natm\n\n\n\n\nCH\n\n\n4\n\n\n\n\n/\n\n\ng\n\n\ncat\n\n\n/\nmin\n\n\nAdsorption rate constant of hydrogen 1/atm1.5\nAdsorption rate constant of methane1/atmKinetic rate constants of deactivation factor Dependent on equationEquilibrium constant atmLifetime of the catalyst minConcentration of the reactant in bulk of gas \n\n\nmol\n/\n\n\nm\n\n\ngas\n\n\n3\n\n\n\n\nOrder of reaction -Partial pressure of methane atmPartial pressure of hydrogen atmInitial reaction rate \n\n\n\n\nmol\n\n\n\n\nCH\n\n\n4\n\n\n\n\n/\n\n\ng\n\n\ncat\n\n\n/\nmin\n\n\nActual reaction rate \n\n\n\n\nmol\n\n\n\n\nCH\n\n\n4\n\n\n\n\n/\n\n\ng\n\n\ncat\n\n\n/\nmin\n\n\nReynolds number at \n\n\n\nu\n\n\nmf\n\n\n\n\n-Time minTemperature \u00b0 CSuperficial velocity of the gas \n\n\n\n\nm\n\n\ngas\n\n\n3\n\n\n/\n\n\nm\n\n\nreactor\n\n\n2\n\n\n/\ns\n\n\nMinimum fluidization velocity m/sSpace Velocity Ln/ gcat/ minEffectiveness factor -Density of the gas \n\n\n\n\nkg/m\n\n\ng\n\n\n3\n\n\n\n\nDensity of the particle \n\n\n\n\nkg/m\n\n\np\n\n\n3\n\n\n\n\nViscosity of the gas Pa.sHeat of reaction \n\n\n\n\nkJ/mol\n\n\n\n\nCH\n\n\n4\n\n\n\n\n\n\nHeat of adsorption of components kJ/molThe destructive consequences of the climate change crisis followed by ongoing efforts toward emission-free technologies have instigated a growing interest in low CO2 and CO2-free hydrogen production (Amin et al., 2011; Borghei et al., 2010; Ashik et al., 2017; Nezam et al., 2021; Ra et al., 2020). Various approaches such as chemical looping reforming, steam methane reforming integrated with Carbon Capture and Sequestration processes (CCS), water splitting, thermal and thermocatalytic decomposition of methane have been studied for this purpose. Among those, thermocatalytic decomposition of methane (TCD) is one the most promising (Hadian et al., 2021). A major advantage of TCD is the potential capability of producing highly valuable carbon nanomaterials instead of CO2, next to hydrogen. The intrinsic characteristics of carbon nanomaterials make them suitable for many industrial applications such as building materials, semiconductors, catalytic materials and energy storage (Ashik et al., 2015; Ashik et al., 2017; Douven et al., 2011; Saraswat and Pant, 2013). In addition, TCD requires less complex down stream purification or separation units than conventional processes. These advantages make TCD an environmentally and economically attractive approach for CO2-free hydrogen production (Hadian et al., 2021).Methane is thermally decomposed to solid carbon and gaseous hydrogen in the absence of a catalyst or oxidizing agents at temperatures above \n\n1300\n\u00b0\nC\n\n (reaction 1). Alternatively, in TCD, a catalyst facilitates the same reaction at a much lower temperature (500\u00a0\u00b0C-950\u00a0\u00b0C) with formation of nano-structured carbon materials. The structure of this material depends on operating conditions and foremost on the catalyst properties that are employed (Hadian et al., 2021). Nickel, iron, copper and carbon are the most studied active sites of the catalyst and among them nickel on silica support, Ni-SiO2, showed the highest methane decomposition activity (Pudukudy et al., 2016; Wang et al., 2000; Reshetenko et al., 2003; Avdeeva et al., 1996; Li et al., 2000; Guevara et al., 2010).\n\n(1)\n\n\n\n\nCH\n\n\n4\n\n\n(g)\n-\n>\nC(s)\n+\n2\n\n\nH\n\n\n2\n\n\n(g)\n\n\n\n\n\n\n\u0394\n\n\nH\n\n\n(\n298\nK\n)\n\n\n=\n+\n74.52\n\nkJ\n/\nmol\n\n\n\nA considerable amount of literature has been published on the preparation of single or bimetallic or carbonaceous catalysts. Their performance in very small lab-scale units and under mild reaction conditions have been established. Srilatha et al. (2017) and Ashik et al. (2015, 2017) reviewed and compared these studies using carbonaceous and metallic catalysts, most of which have been employed in small-scale fixed bed reactors with up to 0.5g of catalyst at limited space velocities and low concentrations of methane. Since the size of the catalyst particle in TCD is increasing over time due to carbon build-up, fixed bed reactors suffer from serious drawbacks such as a high probability of clogging, particle crushing, increasing pressure drop and fracturing the body of the reactor. Therefore, fluidized bed reactors are preferred over fixed bed reactors for large-scale TCD. Indeed, there have been few studies that use fluidized bed reactors or high space velocities (ratio of the total flow rate at normal conditions per gram of catalyst initially loaded), SV. For instance Torres et al. (2012) performed experiments with 20g of fine catalyst particles in a fluidized bed; however, the SV did not exceed 0.2Ln/ gcat/ min. Suelves et al. (2009) used higher SV (2Ln/ gcat/ min) in a fixed bed reactor that contained no more than 0.05g of catalyst.Alongside experimental parametric studies on the performance of the reaction, kinetic studies on TCD in a fixed bed reactor and mild conditions and the mechanism investigations of reaction 1 over metallic catalysts have been performed. These studies revealed that the actual rate of TCD is not constant over time and can be described by Eq. 2, where \n\n\n\nr\n\n\n0\n\n\n\n is the maximum reaction rate and \n\na\n(\nt\n)\n\n is defined as the deactivation factor (Borghei et al., 2010; Amin et al., 2011; Douven et al., 2011; Latorre et al., 2010). In an earlier contribution, the authors summarized the kinetic studies and the proposed kinetic models, including the maximum reaction rate and deactivation factor of the catalyst (Hadian et al., 2021). Several researchers (Amin et al., 2011; Saraswat et al., 2016) proposed a mechanism based on the molecular adsorption of methane followed by step-by-step dehydrogenation reactions until separate adsorbed atoms of carbon and hydrogen are obtained. The first dehydrogenation reaction was found to be the rate-limiting step. The remaining carbon atom of methane on the surface of metal active site, passes through the metal by diffusion and forms nano layers of carbon on the other side. If the decomposition step occurs faster than diffusion and construction rate of carbon nano-structures, carbon atoms accumulate on the surface of metal active site and deactivate it by encapsulation (Toebes et al., 2002; Henao et al., 2021). The maximum reaction rate is modelled by a Langmuir\u2013Hinshelwood type equation that accounts for the thermodynamic equilibrium and the competition between hydrogen and methane adsorption over the active sites, as represented by Eq. 3. The semi-empirical deactivation factor expression is obtained from a species balance on the active sites of the catalyst, resulting in Eq. 4.\n\n(2)\n\n\nr\n(\nt\n)\n=\na\n(\nt\n)\n\u00d7\n\n\nr\n\n\n0\n\n\n\n\n\n\n\n\n(3)\n\n\n\n\nr\n\n\n0\n\n\n[\n\n\nmol\n\n\n\n\nCH\n\n\n4\n\n\n\n\n/\n\n\ng\n\n\ncat\n\n\n/\nmin\n]\n=\n\n\nk\n(\n\n\nP\n\n\n\n\nCH\n\n\n4\n\n\n\n\n[\natm\n]\n-\n\n\nP\n\n\nH\n2\n\n\n2\n\n\n[\natm\n]\n/\n\n\nK\n\n\np\n\n\n)\n\n\n\n\n\n\n1\n+\n\n\nK\n\n\n\n\nH\n\n\n2\n\n\n\n\n\n\nP\n\n\n\n\nH\n\n\n2\n\n\n\n\n1.5\n\n\n[\natm\n]\n+\n\n\nK\n\n\n\n\nCH\n\n\n4\n\n\n\n\n\n\nP\n\n\n\n\nCH\n\n\n4\n\n\n\n\n[\natm\n]\n\n\n\n\n2\n\n\n\n\n\n\n\n\n\n\n(4)\n\n\na\n=\n\n\n\n\n\n\n1\n\n\n1\n-\n0.5\n\n\nk\n\n\nd\n\n\n\n\n\n\n\nk\n\n\nd\n,\nC\n\n\n+\n\n\nk\n\n\nd\n,\n\n\nCH\n\n\n4\n\n\n\n\n\n\nP\n\n\n\n\nCH\n\n\n4\n\n\n\n\n+\n\n\nk\n\n\nd\n,\n\n\nH\n\n\n2\n\n\n\n\n\n\nP\n\n\n\n\nH\n\n\n2\n\n\n\n\n0.83\n\n\n\n\n\nt\n\n\n\n\n\n\n-\n0.8\n\n\n\n\n\nAlthough extensive research has been performed on catalyst preparation and reaction mechanism of TCD, very little is currently known about the feasibility of the TCD process in large scale industrial fluidized beds at the harsh operating conditions encountered. The performance of the catalyst and the fluidized bed reactor need to be thoroughly investigated by performing experimental and numerical studies. This performance can be expressed in terms of the maximum reaction rate, \n\n\n\nr\n\n\n0\n\n\n\n, the life-time of the catalyst, LT, and carbon yield (the ratio of the mass of produced carbon to the initial mass of catalyst used, Eq. 5), CY. The outline of this paper is as follows: in Section 2 we describe the experimental setup, materials as well as the adopted procedures. In Section 3 we introduce the reactor model used for the interpretation of the experiments, whereas the results and discussion is given in Section 4. Finally in Section 5 the conclusions are presented.\n\n(5)\n\n\ncarbon\n\nyield\n\n(\nCY\n)\n=\n\n\nmass\n\nof\n\nproduced\n\ncarbon\n\n(\ng\n)\n\n\ninitial\n\nmass\n\nof\n\ncatalyst\n\n(\ng\n)\n\n\n\n\n\n\nIn the experiments a commercial catalyst made by BASF (Ni 5256 E RS) was used. The catalyst is originally designed as a hydrogenation fixed bed catalyst that contains 56% nickel on silica support and was supplied as extrudates and in reduced and passivated state. Table 1\n shows the characteristics of the fresh catalyst. All of the gases used in this study were at least 99.995% pure, supplied by Linde.The experiments were performed in a cylindrical quartz fluidized bed reactor equipped with a spherical free-board section, see Fig. 1\n. The inner diameter and the height of the cylindrical part are 1cm and 10cm, respectively and the diameter of the free-board is 7.5cm. The spherical free-board reduces the chance of entrainment by lowering the gas velocity and at the same time acts as an expansion space for the growing catalyst particles. The reactor was placed in an electric oven and the desired feed gas composition and flow rate were adjusted by calibrated Bronkhorst mass flow controllers. The local temperature in the reactor can be measured with the help of thermocouples. The outlet gas is transferred to a SICK gas analyzer model GMS815P (three measuring modules: Thermor, Oxor-P and Multor) for gas composition measurement after cooling down, Fig. 2\n.In each experiment 1g of crushed and sieved (500-600\n\n\u03bc\n\nm) catalyst was loaded into the reactor, unless otherwise stated (to examine the effect of the parameter). Prior to activation of the catalyst, the air in the porous catalyst particle was extracted by flowing 2Ln/ gcat/ min nitrogen to the reactor for 30min. Subsequently the temperature of the reactor was increased to 250\u00a0\u00b0C by a ramp of 5\u00b0 C/min using 10vol.\n\n%\n\n H2 in the feed to actuate the catalyst. The catalyst was further reduced in 100vol.% H2 with the same ramp of temperature up to 500\u00a0\u00b0C. Afterwards, the catalyst was heated up further in nitrogen. Once the reaction temperature was reached, 4.5Ln/ gcat/ min of the predefined gas composition of methane, hydrogen and the inert gas (nitrogen) was fed to the reactor, and the outlet composition was measured typically until the catalyst was fully deactivated. Finally, the reactor was cooled down and the product was collected and weighed.BET surface area and pore volume measurements have been carried out for the used catalyst using Thermo Fisher Scientific Analyzer model Surfer. Oxidation temperature of the produced carbon was measured in air via Thermo Gravimetric Analysis (TGA) to characterize the products. Surface composition of the fresh and used catalyst was analyzed by X-ray Photoelectron Spectroscopy (XPS) measurements on Thermo Scientific K-Alpha XPS with an Al source (1486.6\u00a0eV). The structure of carbon nanomaterials were obtained by performing Transmission Electron Microscopy (TEM) imaging of the samples of the products on a FEI cryo TEM TITAN 300\u00a0kV.The composition of the gas entering the reactor is known and the same as the predefined values for each experiment. However, as the gas passes through the reactor methane is consumed and hydrogen is produced. Therefore, the composition of the gas, and consequently the reaction rate is varying along the height of the bed whereas only the outlet conversion can be compared with the experimental data. This conversion is calculated by a Plug-Flow Reactor (PFR) model, Eq. 6. See Fig. 3\n-a.\n\n(6)\n\n\n\n\ndX\n\n\ndw\n[\n\n\ng\n\n\ncat\n\n\n]\n\n\n=\n\n\n-\n\n\nr\n\n\n0\n\n\n[\nmol\n/\nmin\n/\n\n\ng\n\n\ncat\n\n\n]\n\n\n\n\nF\n\n\n\n\nA\n\n\n0\n\n\n\n\n[\nmol\n/\nmin\n]\n\n\n\n\n,\n\n\n\nX\n(\n0\n)\n=\n0\n\n\n\nwhere X is conversion, \n\n\n\nF\n\n\n\n\nA\n\n\n0\n\n\n\n\n\n is the molar flow rate of methane to the reactor, and dw is a fraction of the bed, such that the reaction can be considered constant over the fraction. \n\n\n\nr\n\n\n0\n\n\n\n is replaced by Eq. 3. This differential equation is numerically integrated by Runge\u2013Kutta fourth-order method (RK4). The equilibrium constant at temperature of reaction, \n\n\n\nK\n\n\np\n\n\n\n in Eq. 3, is calculated by Eq. 7 proposed by Kuvshinov et al. (1998),Zavarukhin and Kuvshinov (2004).\n\n(7)\n\n\n\n\nK\n\n\np\n\n\n[\natm\n]\n=\n5.0215\n\u00d7\n\n\n10\n\n\n5\n\n\nexp\n\n\n\n-\n\n\n9.12\n\u00d7\n\n\n10\n\n\n4\n\n\n\n\nRT\n\n\n\n\n\n\n\n\nThe local partial pressures of methane and hydrogen are updated by Eq. 8 and 9. Finally, the kinetic parameter of Eq. 3 was fitted by comparing the conversion of the last section of the fluidized bed with the maximum conversion obtained from each of the experiments.\n\n(8)\n\n\n\n\nP\n\n\n\n\nCH\n\n\n4\n\n\n\n\n=\n\n\n(\n1\n-\nX\n)\n\n\nP\n\n\n\n\nCH\n\n\n4\n\n\n\n\n\n\n1\n+\nX\n\n\n\n\n\n\n\n\n(9)\n\n\n\n\nP\n\n\n\n\nH\n\n\n2\n\n\n\n\n=\n\n\n2\n\n\nXP\n\n\n\n\nCH\n\n\n4\n\n\n\n\n+\n\n\nP\n\n\n\n\nH\n\n\n2\n\n\n\n\n\n\n1\n+\nX\n\n\n\n\n\n\nOver time as the carbon products are being formed, the catalyst particles grow in size and weight with different rates. This growth can reach a point where some of the particles become too heavy to be fluidized by the available gas flow rate. Therefore, they settle down at the bottom of the reactor. These segregated particles are exposed to the fresh feed entering the reactor with a higher concentration of methane and lower concentration of hydrogen compared to the upper parts of the reactor. As a result, they grow faster and they also deactivate quicker than the rest of the particles. Over time more and more particles are segregated and deactivated until the whole bed is deactivated.In order to predict the deactivation of the catalyst and determine the parameters of Eq. 4 it is crucial to model this complex behavior over time. In order to obtain a predictive model for the deactivation, the model described in Section 3.1 is run for each time step starting from the beginning of the reaction until full deactivation. At the end of each time step, the total consumed methane and produced hydrogen and solid carbon are calculated and used to update the particle size. Then the minimum fluidization velocity of the particles is calculated for fine and coarse particles with Eqs. 10 and 11, respectively (Kunii and Levenspiel, 2013).\n\n(10)\n\n\n\n\nRe\n\n\np\n,\nmf\n\n\n=\n\n\n[\n\n\n33.7\n\n\n2\n\n\n+\n0.0408\nAr\n]\n\n\n1\n/\n2\n\n\n-\n33.7\n\n\n\n\n\n\n(11)\n\n\n\n\nRe\n\n\np\n,\nmf\n\n\n=\n\n\n[\n\n\n28.7\n\n\n2\n\n\n+\n0.0494\nAr\n]\n\n\n1\n/\n2\n\n\n-\n28.7\n\n\n\nwhere \n\n\n\nRe\n\n\np\n,\nmf\n\n\n\n is the Reynolds number of the particles, Eq. 12, and Ar is the Archimedes number calculated by Eq. 13.\n\n(12)\n\n\n\n\nRe\n\n\np\n,\nmf\n\n\n=\n\n\n\n\n\u03c1\n\n\ng\n\n\n\n\nu\n\n\nmf\n\n\n\n\nd\n\n\np\n\n\n\n\n\u03bc\n\n\n\n\n\n\n\n\n(13)\n\n\nAr\n=\n\n\n\n\nd\n\n\np\n\n\n3\n\n\n\n\n\u03c1\n\n\ng\n\n\n(\n\n\n\u03c1\n\n\np\n\n\n-\n\n\n\u03c1\n\n\ng\n\n\n)\ng\n\n\n\n\n\u03bc\n\n\n2\n\n\n\n\n\n\n\nBy growing the particles the minimum fluidization velocity increases and the ratio of gas velocity to the minimum fluidization velocity, \n\nu\n/\n\n\nu\n\n\nmf\n\n\n\n decreases. The moment that \n\nu\n/\n\n\nu\n\n\nmf\n\n\n\n is not sufficient to maintain fluidization (\n\nu\n/\n\n\nu\n\n\nmf\n\n\n<\n1.2\n\n this ratio is dependent on the reactor and particles properties), the bottom part of the bed is separated from the rest of the reactor (Fig. 3-b), and the particles are not mixed with the top part anymore. The inflow of gas to the top part is higher due to production of 2\u00a0mol hydrogen for each mole of consumed methane in the segregated section. Therefore, the ratio \n\nu\n/\n\n\nu\n\n\nmf\n\n\n\n can be high enough for fluidization of the particles in the upper sections of the bed. Due to further growth of the particles, the segregated zone propagates along the reactor and eventually the entire bed becomes segregated as shown in Fig. 3-c.This is a phenomenological 1D model representing the evolving reactor and therefore radial difference, wall effect, channelization, and bubble formation are neglected. Before the particles in the first section start segregating, all the particles are fluidized and well mixed in the reactor. Therefore, the particles grow with the same rate at this stage. Segregation only occurs if the carbon yield is high enough (mostly in cases with lower temperature or if hydrogen is added to the feed). Due to segregation, particles are not mixed any more. The growth rate is higher at the bottom of the reactor but there the deactivation starts earlier.Many experiments were conducted by systematically altering the settings of operating temperature, gas concentrations, catalyst particle size and WHGV. All results were confirmed with at least one duplicate experiment. In these experiments depending on the settings lifetime varied from 5min to longer than 12h where the obtained carbon yield ranged between 0gC/gcat to more than 70gC/gcat (at 550\u00a0\u00b0C and 70vol.% CH4-5vol.% H2).The max. conversion of the reactor was limited to about 20% because of the very high SV. It was observed that although at lower SV fluidization occurs with fresh catalyst particles (\n\n\n\nu\n\n\nmf\n\n\n\u2248\n0.1\nm\n/\ns\n\n), the heavier and larger particles including the produced carbon (\n\n\n\nu\n\n\nmf\n\n\n\n depends on the CY and in can exceed \n\n2\nm\n/\ns\n\n for the largest particles) cannot be fluidized and therefore leads to breaking the reactor. Therefore, the gas flow rate (and as a result SV) is chosen to be high enough for mobilization of the grown catalyst particles even after hours of accumulation of carbon on them.Since the diameter of the reactor is relatively small and the consumed heat by reaction is small compared to the heat supplied by the oven, temperature drop along the reactor was limited to a maximum of 17\u00a0\u00b0C (at maximum reaction rate at 600\u00a0\u00b0C and with a feed of 100% CH4). Fig. 4\n shows that the maximum reaction rate increases as the temperature is increased as expected. On the other hand, as can be seen in Fig. 5\n-bottom, a high temperature has a negative effect on the lifetime of the catalyst. These findings are in agreement with literature findings (Hadian et al., 2021; Amin et al., 2011). The carbon yield is a parameter that integrates both effects of maximum reaction rate and the lifetime of the catalyst. Therefore, as shown in Fig. 5-top from 550\u00a0\u00b0C to 650\u00a0\u00b0C a shorter lifetime can overcome the higher reaction rate and carbon yield is significantly lower. However, at lower temperatures the carbon yield is more affected by lower maximum reaction rate and there is an optimum temperature for carbon yield between 500-550\u00a0\u00b0C, balancing initial reaction rate and lifetime of the catalyst.\nFig. 6\n shows that the maximum reaction rate is directly dependent on the volumetric concentration of methane as the only reactant of the reaction. What stands out in this figure is that at 550\u00a0\u00b0C and lower, the maximum reaction rate slightly declines with an increase in methane concentration to 90vol.%. A possible explanation for this might be that adsorption of methane is the dominating step at this temperature and saturation of the active sites allow less neighboring active sites to facilitate the detachment of hydrogen molecules.One unanticipated finding was that the lifetime of the catalyst is shorter when the concentration of methane (and therefore the reaction rate) is lower, Fig. 7\n, while some other researchers observed the opposite behavior (Latorre et al., 2010; Henao et al., 2021). This effect is stronger at lower concentrations or at higher temperatures. We believe that the key difference is the scale of the reactor. Small reactors can be considered as a differential reactor and all the catalyst particles are in an environment with the same concentrations as the feed, while in the larger reactors such as in this work, except for the small portion next to the gas inlet, catalyst particles are in contact with a gas mixture containing the produced hydrogen. Another difference is that the methane concentration range in this study is much higher than most of the literature studies where only mild conditions were tested (max. methane concentration of 7.5% and 42.9% was tested by Latorre et al. (2010) and Henao et al. (2021) respectively). High concentrations lead to higher reaction rate and larger temperature drop along the reactor that is in favor of longer lifetime. At higher methane concentrations, larger amounts of hydrogen are also produced and are present in the reactor. As mentioned in Section 4.1.3, the addition of hydrogen changes the chemical potential and enhances the reverse reaction and converts the accumulated carbon on the surface of active sites back to methane. This phenomenon prevents encapsulation of the active sites and renews them, which boosts the lifetime of the catalyst significantly. Fig. 7 also reveals that carbon yield is decreased as the methane concentration is lowered by dilution with nitrogen. This was confirmed by using argon instead of nitrogen that led to almost the same effect. Analyzing the gas outlet with a mass spectrometer and also the solid products with XPS tests confirmed that neither nitrogen nor argon are involved in any reaction and are indeed inert.Adding a small fraction of hydrogen to the feed decreases the maximum reaction rate by promoting the reverse reaction by Le Chatelier\u2019s principle, Fig. 8\n. This leads to higher refresh rate of the surface of active sites and as a result a higher lifetime of the catalyst. Fig. 9\n shows that also the carbon yield is improved by introducing small amounts of hydrogen to the reactor. This behavior was also observed by other researchers such as Latorre et al. (2010).Three different catalyst particle sizes were tested to investigate the importance of mass transfer limitation in TCD process. It can be seen from Fig. 10\n that the effect of changing the particle size from the average diameter of 350\n\n\u03bc\n\nm to 800\n\n\u03bc\n\nm on the maximum reaction rate is very small. This confirms previous findings in the literature (Hadian et al., 2021; Saraswat et al., 2016; Borghei et al., 2010). However, over time the catalyst particle become larger and the effect of mass transfer limitation becomes more important by decreasing both diffusion length scale and pore volume due to carbon formation (See Table 1). These findings are also confirmed by Weisz-Prater criterion, see the appendix. Fig. 11\n shows that the carbon yield is directly affected by the initial size of the catalyst particle. Because, mass transfer limitation keeps the concentrations of methane and hydrogen inside the particles compared with smaller particles lower and higher respectively. Therefore, the lifetime of the catalyst is boosted and as the result carbon yield is also increased.SV was changed in the experiments while maintaining the volumetric flow rate at 4.5Ln/min by changing the amount of the catalyst in the reactor. Lowering the volumetric flow rate would change the fluidization regime and can lead to breaking the quartz reactor due to defluidization of the grown catalyst particles and increasing it would turn the reactor at the beginning of the reaction into a pneumatic riser. Fig. 12\n shows that the reaction rate does not linearly increase with SV because the local reaction rate decreases along the height of the reactor. An increase in the reaction rate is due to larger amounts of methane and smaller amounts of hydrogen available per unit of catalyst. This also explains the shorter lifetime and lower carbon yield of the catalyst at higher SV, Fig. 13\n.BET measurement results are presented in Table 1 and reveal a sharp decrease in the specific surface area and pore volume compared with the initial catalyst, suggesting the pores are filled up with carbon. The material density of the produced carbon including the catalyst is much lower than the fresh catalyst. As a result, since carbon is less dense than the catalyst material containing nickel the bulk density is also reduced by about 32%. XPS analyses of the deactivated catalyst showed that all the nickel was in the metallic form and covered by carbon graphene layers. It suggests that deactivation is happening due to encapsulation of the active sites which makes them inaccessible for the methane molecules. It was found that even in the case of less pure gases (99.5%) no byproducts (either solid or gas) were formed.The fraction of carbon that is produced in the form of desired graphene layer structures can be determined by evaluating the Derivative ThermoGravimetric (DTG) curve of oxidation temperature. The DTG is defined as the derivative of a TGA curve of the corresponding oxidation temperature. The highest temperature limit for oxidation of amorphous carbon reported in literature is 410\u00b0C (Luxembourg et al., 2005). However, Hu et al. (2003) and Li and Zhang (2005) reported 365\u00b0C and 350\u00b0C respectively for the oxidation of amorphous carbon and the carbon that is oxidized in temperatures above these limits is generally accepted to be nano-structured carbon. Fig. 14\n illustrates the TGA and DTG curve of the carbon produced from a feed of 100% CH4 at 550\u00a0\u00b0C. Even with considering 410\u00a0\u00b0C as the limit temperature of oxidation of amorphous carbon, the lowest purity of the different tested samples was about 96% nano-structured carbon.\nFig. 15\n shows a TEM image of a cluster of produced carbon nanofibers, CNFs, in 100% CH4 at 550\u00a0\u00b0C. The diameters are in the range of 15-80nm. CNFs produced at the different operating conditions were in the form of stacked cones, see Fig. 16\n. Stacked cones are also called a fish bone structure and were obtained also in the literature on the nickel catalyst supported by silica (Toebes et al., 2002; Lehman et al., 2011).The lifetime of the catalyst for some of the conditions at 650\u00a0\u00b0C is so short that the maximum reaction rate could not be measured reliably. At 500\u00a0\u00b0C the reaction was limited by thermodynamic equilibrium in some of the operating conditions. Therefore, only the experimental data obtained at 550\u00a0\u00b0C, 575\u00a0\u00b0C and 600\u00a0\u00b0C were used to find the exact kinetic parameters by the model described in Section 3.1. Table 2\n presents the kinetic parameters of the TCD reaction rate, Eq. 3. The average and maximum relative error of Eq. 3, using parameters from Table 2, do not exceed 11% and 22% respectively.The deactivation factor was fitted to the experimental data at 550\u00a0\u00b0C, 575\u00a0\u00b0C and 600\u00a0\u00b0C and Table 3\n shows the obtained values. The total carbon yield obtained from each experiment and the model were compared. The average and maximum relative difference were 13.0% and 28.7% at the highest. Figs. 17,18\n\n illustrate three examples of the performance of the catalyst over time in the experiments and predicted by the model at different operating conditions. Segregation of the reactor as it is described in Section 3.2 is clearly visible by drops of the deactivation factor.In this study, the thermocatalytic decomposition of methane in a fluidized bed reactor was studied and the corresponding reaction kinetics were established. A commercial nickel on silica catalyst was used in this study and carbon yields of up to and in excess of 70gC/gcat were obtained. The carbon produced was mainly in the form of carbon nano fibers. Its purity was characterized by TEM, TGA and XPS tests. The produced carbon had at least 96% purity of fish bone structures.The effect of operating conditions has been investigated and it was found that at lower temperature, a larger amount of carbon was produced despite that the maximum reaction rate was lower. This was due to the delayed deactivation of the catalyst at lower temperature. Lowering the concentration of methane lowered the maximum reaction rate, lifetime and carbon yield. Our study has also revealed that although the presence of hydrogen decreases the maximum reaction rate, a higher carbon yield is achieved due to longer lifetime of the catalyst.A kinetic model describing the maximum reaction rate and the deactivation factor was developed. This model describes the reaction rate of TCD as a function of time in the temperature range of 550-600\u00a0\u00b0C with a reasonable accuracy and averaged error in initial kinetic rate of 10% and deactivation factors up to 17 %. This model together with the corresponding commercially available catalyst can be used for further study of TCD and reproduction of data. This kinetic model provides the basis to simulate a fluidized bed reactor for TCD with more detailed (i.e. CFD-based) models to facilitate reactor development and optimization.A very high SV was chosen in this study to facilitate the mobility of catalysts that have grown due to large depositions of carbon, and to prevent breaking the reactor\u2019s wall. As a result, the conversion of the gas was limited. In an industrial scale, this can be overcome with a proper continuous reactor design and partial recycle of the gas stream.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work is part of the Advanced Research Center for Chemical Building Blocks, ARC CBBC, which is co-founded and co-financed by the Netherlands Organisation for Scientific Research (NWO) and the Netherlands Ministry of Economic Affairs.The authors acknowledge Rick Joosten (Center For Multi-scale Electron Micoscopy in Eindhoven University of Technology) for TEM imaging of the samples, and also Rim van de Poll (Inorganic Materials & Catalysis research group form Eindhoven University of Technology) for conducting XPS measurements.According to Weisz-Prater criterion, \n\n\n\nC\n\n\nWP\n\n\n\n, the mass transfer limitation can be neglected compared with reaction kinetics limitations, if Eq. 14 holds (Vannice, 2005).\n\n(14)\n\n\n\n\nC\n\n\nWP\n\n\n=\n\n\nRate\n.\n\n\n(\ndp\n/\n2\n)\n\n\n2\n\n\n\n\n\n\nD\n\n\neff\n\n\n.\n\n\nM\n\n\nb\n\n\n\n\n\u2a7d\n3\n\u03b2\n\n\n\n\n\n\n(15)\n\n\n\u03b7\n=\n1\n-\n\n\nn\n\u03b2\n\n\n4\n\n\n\n\n\nwhere Rate is the observed rate, \n\n\n\nD\n\n\neff\n\n\n\n is effective diffusivity, \n\n\n\nM\n\n\nb\n\n\n\n is the concentration of the reactant outside the porous particle, \n\n\u03b2\n\n maximum decrease in concentration gradient in pores, \n\n\u03b7\n\n is effectiveness factor, and n is the order of reaction. The reaction is considered first order (following Eq. 3) and and the observed rate of reaction is obtained from experimental results. Using the other parameters as is presented in Table 4\n, it is found:\n\n\n\n\n\nC\n\n\nWP\n\n\n=\n0.42\n\u2a7d\n0.6\n\n\n\nthat confirms that the diffusional mass transfer limitation can just be neglected. However, with larger particles, it can become important as it has effect on the carbon yield and lifetime of the catalyst, see Section 4.1.4.", "descript": "\n                  ThermoCatalytic Decomposition of methane (TCD) offers an interesting route to convert natural gas into hydrogen and functional carbon. In this study the reaction kinetics of TCD is studied for a nickel supported catalyst using a special fluidized bed reactor. The effect of operating conditions such as temperature, concentrations of methane and hydrogen and space velocity (SV) was studied on a commercial nickel catalyst on a silica support. The performance of the catalyst was evaluated in terms of three parameters: maximum reaction rate, lifetime and carbon yield. Values up to and in excess of 70gC/gcat and 12h (at 550\u00a0\u00b0C and 70vol.% CH4-5vol.% H2) have been achieved for carbon yield and lifetime, respectively. The carbon product has fish bone structure. Our study has revealed that at lower temperatures and in the presence of small amounts of hydrogen (\n                        \n                           \u2264\n                           10\n                           %\n                        \n                     ) a higher carbon yield is obtained. Lower concentration of methane (higher concentration of the inert) lowers the reaction rate, the lifetime and therefore the carbon yield. A dual kinetic approach has been adopted to determine maximum reaction rate and the associated deactivation factor. The kinetic parameters were estimated for the temperature range of 550-600\u00a0\u00b0C.\n               "}
{"full_text": "Data will be made available on request.Global warming due to emissions of the greenhouse gases (GHGs) is a major environmental challenge today and carbon dioxide CO2 is the most potent contributor [1\u20133]. Consumption of fossil fuels worldwide in various sectors is the predominant cause of CO2 emissions and at a rate of 1.5 to 3\u00a0ppm per anum; from 2006 to 2020 alone, CO2 concentration has risen from 380\u00a0ppm to around 412\u00a0ppm globally [4\u20137]. Hence, a sustainable means of CO2 mitigation is required for the reduction of global warming as well as the clean production of renewable fuels. Indeed, carbon capture and storage (CCS) can be employed to capture CO2 from the atmosphere and subsequently store the captured CO2 [8,9]. However, the storage of large amounts of CO2 will require a lot of space in addition to increasing cost. In this regard, catalytic conversion of captured CO2 into methanol is ultimately more preferable and is highly desirable not only from environmental but also from economic perspectives [10]. This is because methanol can either be used directly as fuel additive (gasoline, dimethyl ether, methyl tertbutyl ether) and/or can be further converted to various useful chemicals, such as olefins, aromatics, formaldehyde and acetic acid [11,12]. Thus, creating a huge opportunity for utilization of CO2 for various finished products and commodities, such as plastics, polymers; consequently resulting in longer-term abatement of CO2 or into relatively greener fuels [13]. Carbon Recycling International operating in Iceland is a good example of focusing on methanol production from catalytic conversion of CO2 [14].Carbon dioxide hydrogenation to methanol is surface catalyzed reaction and occurs according to the following exothermic reaction:\n\n(1)\n\n\nCO\n2\n\n+\n\n3H\n2\n\n\u2192\n\nCH\n3\n\nOH\n+\n\nH\n2\n\nO\n,\n\n\u0394H\no\n\n\n\n\n275\n\u00b0\nC\n\n\n=\n\u2212\n59\n\nkJ\n\nmol\n\n\u2212\n1\n\n\n\n\n\nHowever, the above reaction is accompanied by reverse water gas shift reaction (RWGS) and methanation reaction as given below,\n\n(2)\n\n\nCO\n2\n\n+\n\n4H\n2\n\n\u2192\n\nCH\n4\n\n+\n\n2H\n2\n\nO\n,\n\n\u0394H\no\n\n\n\n\n275\n\u00b0\nC\n\n\n=\n\u2212\n176\n\nkJ\n\nmol\n\n\u2212\n1\n\n\n\n\n\n\n\n(3)\n\n\nCO\n2\n\n+\n\nH\n2\n\n\u2192\nCO\n+\n\nH\n2\n\nO\n,\n\n\u0394H\no\n\n\n\n\n275\n\u00b0\nC\n\n\n=\n+\n39.5\n\nkJ\n\nmol\n\n\u2212\n1\n\n\n\n\n\nHydrogenation of CO2 to methanol has been subject of intensive study with various reports on noble and noble metal-based catalysts. Amongst these copper, indium and palladium-based catalysts have shown promise with relatively better activity, selectivity and stability [12,15\u201317]. Copper based ternary catalyst consisting of oxides of Cu, Zn and Al due to its lower cost and its use in commercial methanol synthesis via syngas have been extensively investigated; furthermore Cu catalyst have been supported and promoted with other metal oxide systems, such as Zn, Zr, Ce, Si, V, Ti, Ga, B, Cr and Mn, with the aim to improve overall activity and selectivity for CO2 to methanol [18]. However, due to thermodynamic and kinetic limitations as well as due to debatable reaction mechanism of CO2 hydrogenation until now no single catalyst system has been industrially optimized for methanol synthesis from pure CO2 feed. The key factor for the success of CO2 hydrogenation to methanol requires the development of highly active, selective and stable Cu-based catalysts.Indeed, it is generally established that particle size, particle size distribution, surface defects, composition of the catalyst and oxygen vacancies have considerable influence on the catalytic activity of CO2 hydrogenation catalysts [19,20]. This is because CO2 hydrogenation over Cu-based catalysts is believed to be structure sensitive in nature and with smaller CuO particle more active for CO2 hydrogenation process [21]. In spite of this, there is a limited reported literature on the effects of Cu-particle size on the catalytic performances for CO2 hydrogenation to methanol. Despite the presence of some relevant reports, there seems to be a lack of a systematic investigation of the dependence of CO2 hydrogenation over copper nanocrystallite size and a considerable debate and contradiction in conclusions of whether or not MeOH is structure sensitive exists. For example, according to some reports, CuO particles size in the size range of 8.5\u00a0nm\u201337.5\u00a0nm did not affect MeOH formation rates. Whereas, CuO particles with size bigger than >38-200\u00a0nm favored MeOH formation. Similarly, CuO with particle sizes lower than <6\u00a0nm, particles favored CO formation. The above mentioned behavior has been reported for methanol synthesis over Cu based catalysts from CO2/H2 mixture [22] and co-feed of CO2/CO/H2 [23]. By contrast, other reports suggest that the catalysts with CuO particle sizes in the range between 10\u00a0nm\u201327\u00a0nm, enhanced the formation rate of methanol [24]. Additionally, the catalytic studies of Cu/Zn/Zr catalysts with Cu particle sizes in the range between 2\u00a0nm\u201325\u00a0nm and with low percentage dispersion promoted higher methanol formation turnover frequencies [25]. In this regards, the preparative methodology have been reported to exhibit an enormous influence on the catalytic activities of Cu-based catalysts [26,27]. In the last decade, various synthesis methods such as sol-gel, co-precipitation, impregnation and strong electronic adsorption method have been reported for the preparation of Cu-based catalysts resulting in catalysts with desirable properties and believed to have a high catalytic activity for methanol synthesis from CO2 hydrogenation. However, to further increase the activity and selectivity of these catalysts the synthesis methodology should be improved. Amongst the synthesis techniques, the solution combustion synthesis (SCS) method has gained a lot of interest in recent years. This is because of attracting an attention for the bulk production of high quality nanomaterials and nanocomposites for various applications with synthesis of catalytic materials at the heart of it [28\u201330]. The SCS preparative methodology involves an exothermic and self-sustained reaction between oxidants and reductants/fuel. Oxidants are mainly nitrates of the metal/metals precursor salts whereas reductants/fuel used are glycine, hydrazine, citric acid etc. The SCS method results in the production of highly porous nanomaterials, with small and highly dispersed metal particle size and large active surface area. This is due to exhaust of large volume of gases during synthesis. In our previous work, we reported development of an active and stable nickel based catalysts for dry reformation of methane synthesized via the SCS method [29,31]. The Ni-based catalysts synthesized by the SCS method were comparatively more active and superior in stability than the catalysts prepared by conventional impregnation method. However, the physicochemical properties of the final products depends mainly on factors such as the amounts of fuel (glycine) and oxidant (nitrates) in the reaction mixture, type of the fuel and calcination temperatures [32,33]. For example, Bedekar and co-workers have demonstrated preparation of various phases of yttrium and chromium oxides at various fuel to oxidants ratios. The fuel enriched combustion system were reported to result in the formation of YCrO3 phase, by contrast the fuel deficit system favored the formation of YCrO4 [34]. Fuel deficit combustion system has also been utilized for the synthesis of single-phase oxide compositions of zinc, nickel, copper, iron and cobalt oxides. By varying the reductants to oxidants (F/O) ratios mixed oxide spinels of the aforementioned metal oxides were also synthesized [35]. However, it is worth to mention that even though SCS is an exothermic reaction, but due to heat losses and large volume of gases released the actual flame temperature in most cases has been found to be lower than the theoretically calculated values, thus making it difficult to theoretically determine properties of the final powder [36].It is clear from the above discussion that the development of efficient Cu-based catalysts for the hydrogenation of pure CO2 into value added chemicals remains a key factor and a challenging topic for researchers. Although, amongst the preparative methods, the SCS preparative method is regarded as a potential route to produce nanocatalysts with excellent catalytic properties. However, the SCS synthesis variables have greater influence in controlling various physicochemical properties, such as the phase, crystal size, crystal structure, oxidation states, mixed oxide phases and surface defects of the materials prepared. Thus, the aim of this study is to develop active and selective Cu-based catalysts for reduction of CO2 to methanol using SCS method. The work covers a systematic investigation of various key parameters of the SCS synthesis technique, such as glycine to nitrates ratio, effects of calcination temperature and effects of activation temperature. A correlation between the physicochemical properties of the Cu-based catalysts prepared at various G/O ratios and catalytic performance was carried out.The Cu-based catalyst, with composition of 30wt%CuO-49.65wt%ZnO/20.35wt%Al2O3, was synthesized via the preparative methodology of solution combustion synthesis (SCS). The samples were prepared at various glycine to nitrates ratios (nG/nO) of 0.105, 0.206, 0.258, 0.309, 0.618, 0.747, 0.804, 0.927, and 1.236 and were denoted as C1\u20130.1.5, C1\u20130.206, C1\u20130.258, C1\u20130.309, C1\u20130.618, C1\u20130.747, C1\u20130.927, and C1\u20131.236, respectively. Briefly, in this SCS method, calculated amounts of the precursor salts, Copper (II) nitrate hexahydrate (Cu(NO3)2.6H2O, BDH \u226598.5%%), Zinc nitrate hexahydrate (Zn(NO3)2\u00b76H2O, Sigma Aldrich \u226598%%), and aluminum nitrate nonaahydrate (Al(NO3)3\u00b79H2O, Sigma Aldrich, \u226599.9%) were dissolved in 100\u00a0ml of deionized water in a 500\u00a0ml beaker and stirred well to get a homogeneous mixture. The required amounts of glycine (Sigma Aldrich, 98.5%) as a fuel was then added to this solution under continuous stirring. The mixture was heated on a hotplate to a final temperature of 150\u00a0\u00b0C for combustion. The solution turned into soft gel before a self-sustained combustion reaction initiated. The synthesized powder was calcined in muffle furnace using static air at 600\u00a0\u00b0C at a heating and cooling rate of +1\u00a0\u00b0C/min and -1\u00a0\u00b0C/min for three hours. For studying the effects of calcination temperature, the C1\u20130.206 catalyst was calcined at 400\u00a0\u00b0C, 500\u00a0\u00b0C, 600\u00a0\u00b0C, and 800\u00a0\u00b0C. These catalysts were denoted as C1\u2013400, C1\u2013600, C1\u2013400, and C1\u2013800, respectively.The analytical technique of X-ray diffraction (XRD) analysis was used to study the bulk phase and crystallinity of the calcined catalysts. The XRD measurements were performed on a desktop X-ray diffractometer (Rigaku, MiniFlexII) having CuKa as radiation source at 30\u00a0kV and 15 m\u0410. Morphology of the calcined catalysts in terms of average metal oxide particle size and particle size distribution was investigated by using a high-resolution transmission electron microscope (HRTEM). Samples for HRTEM analysis were prepared by dispersing around 10\u00a0mg of calcined powder into n-propanol by sonication. A drop of the dispersed sample was then placed onto a copper grid with a mesh size of 150-mesh. The HRTEM analysis was carried out using a JEO; 2010\u00a0F high-resolution microscope at an operating voltage of around 200kv. Chemical surface composition and oxidation states of various copper species over the surface of the calcined samples were studied by using X-ray Photoelectron Spectrometer (AXIS Ultra DLD, KRATOS). Prior to XPS analysis, the surface each sample was cleaned from adventitious carbon using the ion gun. Temperature programmed reduction (H2-TPR) analyses was used to study the reduction behavior of the catalysts using ChemiSorb2750 (Micromeritics) equipped with a thermal conductivity detector (TCD). For TPR analysis, around 100\u00a0mg of the calcined sample was loaded in the U-shaped quartz tube and degassed for two hours at 200\u00a0\u00b0C in presence of 30\u00a0ml/min of Ar. This step was followed by decreasing the furnace temperature to 40\u00a0\u00b0C in inert flow. The TPR was then recorded by switching the flow 30 cm3min\u22121 of 10vol%H2/Ar and heating from 40\u00a0\u00b0C to 900\u00a0\u00b0C at 5oCmin\u22121.The catalytic performances of the catalysts for CO2 hydrogenation was evaluated in a high-pressure lab-scale test unit (PID, Effi, Micromeritics). The unit is equipped with three gas lines controlled by high accuracy mass flow controllers and operates with a hastelloy fixed bed reactor (ID: 9.3\u00a0mm). The reactor tube is externally heated with a three-zone electric furnace. The exit stream of the reactor was cooled via a paltrier cold trap where liquid and gaseous products were separated. The reaction temperature was monitored with a thermocouple inserted in the catalytic bed. The gaseous products were analyzed on-line, with an online GC-TCD. The liquids were collected in a trap (5\u00a0\u00b0C) and were analyzed offline. The analysis was performed with a GC Agilent 7890A equipped with FID detectors. For catalytic tests around 0.5\u00a0cm3 of the palletized catalyst with sieve size of 75-150\u00a0\u03bcm was diluted with 1.0\u00a0cm3 of quartz powder with similar grain size was loaded into the reactor. Prior to reaction, the catalysts were activated at the required temperature (350\u00a0\u00b0C or 500\u00a0\u00b0C) in a stream of 30\u00a0ml/min of pure H2 with a dwell time of three hours. The reactor temperature was then cooled down to 250\u00a0\u00b0C followed by switching the reactants mixture (CO2/H2/N2) and increasing the pressure to the initial set point (60bars).\n\n(4)\n\n\nCO\n2\n\n\nConversion\n\n\n%\n\n=\n\n\n\n\n\nF\nCO\n\n2\n\n\nin\n\n\u2212\n\n\nF\nCO\n\n2\n\n\nout\n\n\n\n\n\nF\nCO\n\n2\n\n\nin\n\n\n\n\nx\n\n100\n\n\n4\n\n\n\n\n\n\n(5)\n\n\nCH\n3\n\nOH\n\nSelectivity\n\n\n%\n\n=\n\n\n\n\n\nn\nCH\n\n3\n\nOH\n\n\nn\nTotal products\n\n\n\nx\n\n100\n\n\n5\n\n\n\n\n\n\n(6)\n\nCO\n\nSelectivity\n\n\n%\n\n=\n\n\n\nn\nCO\n\n\nn\nTotal products\n\n\n\nx\n\n100\n\n\n\n\n\n\n(7)\n\n\nCH\n4\n\n\nSelectivity\n\n\n%\n\n=\n\n\n\n\nn\nCH\n\n4\n\n\nn\nTotal products\n\n\n\nx\n\n100\n\n\n\n\n\n\n(8)\n\n\nCH\n3\n\nOH\n\nYield\n=\n\n\n\ng\n\nCH\n3\n\nOH\n\n\nwt\n\nof catalyst\n\n\ng\n\nx\n\nh\n\n\n\n\n\n\n\n\n\n(9)\n\nCO\n\nYield\n=\n\n\n\ng\nCO\n\n\nwt\n\nof catalyst\n\n\ng\n\nx\n\nh\n\n\n\n\n\n\n\n\n\n(10)\n\nCarbon balance\n\n\n%\n\n=\n\n\n\n\n\nn\nCO\n\n2\n\n\nin\n\n\n\n\n\n\n\nn\nCO\n\n2\n\n+\nn\n\nCH\n3\n\nOH\n+\nnCO\n+\n2\nnC\n2\nH\n5\nOH\n+\n3\nnC\n3\nH\n8\nOH\n\n\nout\n\n\n\nx\n100\n\n\n\n\nWhere, F is flow of the gases in NL/h and n is number of moles.In order to study the effects of G/O ratio on the physicochemical properties of the catalyst, a series of the 30wt%CuO50wt%ZnO/Al2O3 (C1) catalysts with various glycine to nitrates (G/O) ratio were synthesized by the SCS method. The G/O ratio represents ratio of moles of glycine to the number of moles of total nitrates in the combustion mixture. Photographs of the as synthesized powder of different batches are shown in Fig. 1\n. In general the sample prepared at a G/O ratio of <0.618 were porous, whereas samples prepared at a G/O ratio of >0.804 turned less porous and bulky. Moreover no combustion occurred for the catalyst prepared at a G/O ratio\u00a0<\u00a00.105. As shown in Fig. 1d, the combustion with a G/O ratio of 0.618 occurred in such a violent manner that almost all powder was blown away due to the aggressive combustion and exhaust of large volume of gases. This is because ratio of fuel to oxidants ratio (\u03d5) was close to the stoichiometry and resulted in making the combustion violent. As shown in Fig. 1a and Fig. 1b, for the catalysts prepared at a G/O ratio of 1.23 and 0.927, a controlled combustion was observed. However, the synthesized powders were comparatively bulky and less porous. This behavior was attributed to the presence of excess fuel in the reaction mixture making oxidants (nitrates) limited. As can be seen in Fig. 1f, the catalyst prepared at a G/O ratio of 0.206 proceeded with a slow combustion. Although no real flame was observed but a large volume of gases kept exhausting until a complete combustion occurred. The as produced powder was more porous with some unburnt gel remaining on the walls of the beaker. As will be discussed in later section. This nanocatalyst exhibited exceptionally high catalytic performance in comparison with other catalysts tested for CO2 hydrogenation reaction. For sample synthesized with a G/O ratio of 0.105, since the combustion mixture was fuel deficit therefore negligible combustion occurred and almost no gases were exhausted. Interestingly, the combustion mixture with G/O ratio of <0.105 no combustion happened even upon increasing the hotplate temperature (Fig. 1h). This behavior was attributed to the unavailability of sufficient fuel to trigger combustion.XRD analysis of the uncalcined samples and after calcination at 600\u00a0\u00b0C were recorded to study a correlation between G/O ratio and textural properties. Average particle size was calculated using Scherer's equation and XRD diffraction patterns are shown in Fig. 2\n. As shown in Fig. 2a, in the diffraction patterns of the unclaimed C1\u20130.206, C1\u20130.258 and C1\u20130.308, samples, a diffraction line corresponding to carbon at 2\u03b8 value of 15\u00b0 was observed. This was expected as these catalysts were prepared at fuel deficit condition and small amounts of the uncombusted gel material remained. The XRD peaks of the uncalcined C1\u20130.206 and C1\u20130.258 samples were comparatively broader and distinct diffraction lines corresponding to various metal components were not observed. Broadening of the diffraction lines was attributed to smaller and/or well dispersed CuO nanoparticles. Average CuO particle size of uncalcined C1\u20130.206 and C1\u20130.258, samples were 3.5\u00a0nm and 5.6\u00a0nm, respectively. Diffraction patterns of the C1\u20130.309 were comparatively sharper and the average metal particle size was larger (12.6\u00a0nm). It is important to highlight that the diffraction lines of uncalcined samples prepared at G/O of 0.6 and\u00a0>\u00a00.6 were sharper and more intense. Especially for the catalysts samples prepared at G/O ratios of 0.8, 0.9 and 1.23, sharper and distinguishable diffraction lines were recorded. The average particle size of the catalysts also increased with larger particle size recorded for C1\u20131.23 (24.2\u00a0nm). The 2\u03b8 values at 32.5o, 35.9o, 38.6o, 48.6o, 53.3o, 58.2o and 61.4o were attributed to the presence of CuO particles [37]. Whereas the diffraction peaks located at 31.9\u00b0, 34.6\u00b0, 36.4\u00b0, 47.5\u00b0, 56.69\u00b0, 62.71\u00b0, and 68.13\u00b0, were assigned to the presence of ZnO nanoparticles [37,38]. A possible explain was that for the samples prepared with G/O ratio of >0.6 i.e. either with stoichiometric amounts of fuel to oxidants and/or fuel enriched systems a violent combustion occurred presumably with high flame temperature. Metal oxide particles might have agglomerated resulting in larger metal particles with less porosity. As shown in Fig. 2b, after calcination for the catalysts C1\u20130.206, C1\u20130.258, and C1\u20130.309, the diffraction lines corresponding to surface carbon disappeared indicating successful removal of surface carbon. As can be seen in Fig. 2c, in general the average particle of CuO increased with increase in G/O ratio from 0.105 to 0.804 and then remained constant at a G/O ratio of >0.6018. Moreover, for all samples metal oxide particles of the calcined samples were comparatively larger than that of the uncalcined samples.The results of the HRTEM analysis of the calcined samples affirmed the findings of XRD analysis results. Representative TEM images of the samples prepared with G/O ratio of 0.206, 0.6018 and 1.236 are demonstrated in Fig. 3\n. The sample C1\u20130.206 exhibited smaller metal oxide particle size with narrow particle size distribution whereas the sample C1\u20130.618 was found to have larger metal oxide particles with broader particle size distribution. The average metal oxide particle sizes for samples C1\u20130.206, C1\u20130.6018 and C1\u20131.236 were16.14\u00a0\u00b1\u00a05.16, 21.46\u00a0\u00b1\u00a06.9, and 36.6\u00a0\u00b1\u00a011.70, respectively.The effects of G/O ratios on the surface chemical composition of the calcined catalysts was studied by means of X-ray photoelectron (XPS) analysis and the Cu2p3/2 core level region of the catalysts are represented in Fig. 4\n. The XPS of the Cu(2p) core level region of the C1\u20130.206 catalyst (Fig. 4a) showed three Cu(2p3/2) peaks. The binding energies at 934.8\u00a0eV and 933.3\u00a0eV were assigned to Cu(2p3/2) strongly associated bCu2+ and loosely dispersed aCu2+ species, respectively [39,40]. The third peak at binding energy of 936.6\u00a0eV was assigned to highly ionized cCu2+ ions which were either more ionized and/or were experiencing stronger interaction than those associated with CuO. The appearance of high binding energy peak could possibly be due to the insertion of copper species into the ZnO/Al2O3 lattices resulting in the formation of surface defects and oxygen vacancies. The relative percentages (e.g. a/(a\u00a0+\u00a0b\u00a0+\u00a0c)) of the aCu2+, bCu2+ and cCu2+ ions were 25.6%, 41.3% and 33.1%, respectively. As shown in Fig. 4b, the Cu2p core level XPS spectrum of the C1\u20131.236 catalyst consisted of Cu species similar to that of the C1\u20130.206 catalyst. However, the relative percentage of highly ionized cCu2+ ions for C1\u20131.236 catalyst decreased to 27.7%. It was worth noticing the XPS of both the catalysts predominantly consisted of bCu2+ and cCu2+ ions whereas aCu2+ species for C1\u20130.206 and C1\u20131.236 were only 25.6% and 14.9%, respectively. The C1\u20130.206 catalyst revealed the highest percentage of highly ionized cCu2+ species and previous studies have reported positive influences of these species on the catalytic performance during CO2 hydrogenation to methanol [41]. As will be discussed in later section, the C1\u20130.206 catalyst due to smaller metal oxide particle size and presence of high percentage of cCu2+ species exhibited comparatively high catalytic activity than catalyst prepared at other G/O ratios.The calcination and/or annealing temperature is one of the most important parameter affecting various physicochemical properties of Cu-based catalysts and in turn affects the catalytic performance. Thus, identification of proper calcination conditions and/or temperature might results in catalysts with the desired properties. In order to study the effects of calcination temperature on textural properties as well as to make a correlation with the performance, the batch of the C1 catalyst prepared at a G/O ratio of 0.206 was divided into four portions and subsequently calcined at temperature of 400\u00a0\u00b0C, 500\u00a0\u00b0C, 600\u00a0\u00b0C and, 800\u00a0\u00b0C. These samples were and denoted as C1\u2013400, C1\u2013500, C1\u2013600 and C1\u2013800, respectively.As shown in Fig. 5a, the uncalcined sample revealed presence of diffraction lines at 2\u03b8 value of 15o corresponding to diffraction lines of carbon. As discussed in section 3.1, the C1\u20130.206 catalyst was prepared under fuel deficit conditions resulting in incomplete combustion and small amounts of the gel material remained even after combustion. XRD peaks of the uncalcined sample were broader and distinct diffraction lines did not appear. This is because the average metal oxide particle size was around 3.87\u00a0nm. In addition, broader diffraction lines also indicated that either the nanoparticles were well-dispersed or mixed oxide and/or induced phases were formed. The sample calcined at 400\u00a0\u00b0C exhibited diffraction lines very similar to that of the uncalcined samples. However, it is important to highlight that the diffraction lines corresponding to carbon disappeared. The average CuO particle size of the C1\u2013400 catalyst was 4.46\u00a0nm suggesting that the calcination at 400\u00a0\u00b0C effectively removed surface carbon without having a significant impact on the metal oxide particle size and metal dispersion.It is worth to mention that for the C1\u2013500 catalyst, the diffraction peaks became comparatively sharper and an increase in peak intensities was recorded. For the C1\u2013600 and C1\u2013800 catalysts sharper and distinguishable diffraction lines of CuO, ZnO and Al2O3 were observed. Especially, for the C1\u2013800 catalyst, the diffraction peaks were not only intense but also clearly distinguishable from one another. For the C1\u2013800 catalyst, the 2\u03b8 values at 32.5o, 35.9o, 38.6o, 48.6o, 53.3o, 58.2o and 61.4o confirmed the presence of copper oxide particles [37]. Whereas the diffraction lines located at 2\u03b8 degree of 31.9\u00b0, 34.6\u00b0, 36.4\u00b0, 47.5\u00b0, 56.69\u00b0, 62.71\u00b0, and 68.13\u00b0, were assigned to the presence of zinc oxide nanoparticles [37,38]. The average particle size of the catalyst C1\u2013600 and C1\u2013800 was 7.4\u00a0nm and 18.1\u00a0nm, respectively, which was larger than the other catalysts. Indeed, as shown in Fig. 5b, this behavior strongly suggested that at calcination temperature\u00a0<\u00a0500\u00a0\u00b0C the metal particle size was smaller with good dispersion. However, for samples calcined at temperature\u00a0>\u00a0500\u00a0\u00b0C agglomeration of nanocrystallites occurred resulting in larger copper oxide particles and presumably poor dispersion.HRTEM analysis results of the catalysts calcined at various temperatures affirmed the findings of the XRD analysis results. Representative TEM images of the samples are shown in Fig. 6\n. The C1\u2013400 catalyst exhibited smaller metal oxide particle size with narrow particle size distribution whereas the sample C1\u2013600 was found to have larger metal oxide particles with broader particle size distribution. The average metal oxide particle sizes for samples C1\u2013400 and C1\u2013600 were 4.4\u00a0nm and 7.5\u00a0nm, respectively.The effects of calcination temperature on the surface chemical composition of the calcined catalysts was studied by means of X-ray photoelectron (XPS) analysis and the results are represented in Fig. 7\n. As can be seen in Fig. 7a, the XPS of the Cu(2p) core level region of the C1\u2013400 catalyst revealed presence of three Cu(2p3/2) peaks. The binding energies at 933.58\u00a0eV and 934.94\u00a0eV were assigned to Cu(2p3/2) strongly associated bCu2+ and loosely dispersed aCu2+ species, respectively [42]. The third peak appeared at binding energy of 936.44\u00a0eV and was assigned to highly ionized cCu2+ ions. The relative percentages (a/(a\u00a0+\u00a0b\u00a0+\u00a0c)) of aCu2+, bCu2+ and cCu2+ species were 42%, 30% and 28%, respectively. As can be seen in Fig. 7(b), the Cu2p core level XPS spectrum of the C1\u2013600 catalyst revealed a shift in the binding energies to higher values. Moreover, the relative percentage of the bCu2+ and cCu2+ species also increased to 41.3% and 33.1%, respectively. It is worth noticing the XPS of the C1\u2013800 catalyst predominantly consisted of bCu2+ (50%) and cCu2+ (42%) ions whereas the highly ionized aCu2+ species were only 7%.As shown in Fig. 8\n, increase in calcination temperature resulted in proportional increase in the highly ionized cCu2+ ions as well as metal oxide particle size. For example, upon increasing the calcination temperature from 400\u00a0\u00b0C to 800\u00a0\u00b0C, metal oxide particle size increased from 4.4\u00a0nm to 18.1\u00a0nm and at the same time for the C1\u2013400 and C1\u2013800 catalysts, the relative percentages of cCu2+ ions were 28%, and 42%, respectively.Three catalysts i.e. the catalyst synthesized under fuel deficit conditions (C1\u20130.206), the catalyst prepared under stoichiometric amount of fuel (C1\u20130.618) and the catalyst prepared under enriched fuel conditions (C1\u20130.804) were tested for CO2 hydrogenation. Methanol, water, carbon monoxide and methane were the main products of the reaction. Traces of higher alcohols (ethanol and propanol) were also detected. As can be seen in Fig. 9\n, the results indicated that the catalyst prepared at G/O ratio of 0.206 significantly improved the overall catalytic performance. As demonstrated in Fig. 9a, the CO2 conversion over the C1\u20130.206 catalyst was exceptionally higher than the C1\u20130.618 and C1\u20130.804 catalysts. For example at a reaction condition 1 (250\u00a0\u00b0C, 60bars, H2/CO2\u00a0=\u00a03.43 and GHV\u00a0=\u00a07000\u00a0h\u22121) the CO2 conversion over the catalysts C1\u20130.6 and C1\u20130.8 was \u223c3.1% and\u00a0\u223c\u00a02.2%, respectively, whereas the CO2 conversion over the C1\u20130.206 catalyst was \u223c11.2%. This was higher by a factor of \u223c5 than that of the C1\u20130.618 and C1\u20130.804 catalysts. In general, CO2 conversion over all catalysts increased with increase in reaction temperature and pressure. The maximum CO2 conversion of \u223c30% over the C1\u20130.206 catalyst was obtained at temperature of 300\u00a0\u00b0C, P\u00a0=\u00a085b and a space velocity of 7000\u00a0h\u22121. Whereas, under these conditions the CO2 conversion over C1\u20130.616 and C1\u20130.804 was 17.60% and 16.4%, respectively.It is worth to mention that at reaction condition 1 (250\u00a0\u00b0C, 60\u00a0bar) and reaction condition 2 (250\u00a0\u00b0C, 85\u00a0bar), the methanol selectivity over the C1\u20130.206 catalyst was lower than the C1\u20130.618 and C1\u20130.804 catalysts. Whereas, the MeOH selectivities for the C1\u20130.618 and C2\u20130.804 catalysts were similar. For example, at 250\u00a0\u00b0C MeOH selectivities for the C1\u20130.206, C1\u20130.6 and C1\u20130.8 catalysts were 53%, 72% and 70%, respectively. Similarly CO selectivity for the C1\u20130.206 catalyst under these conditions was higher than that of the other two tested catalysts. Indeed, the reaction temperature had a remarkable impact on the products selectivity. Generally, for all three tested catalysts an increase in CO selectivity and a decrease in MeOH selectivity was recorded. However, increase in temperature from 275\u00a0\u00b0C and 300\u00a0\u00b0C resulted in a drastic decrease in MeOH selectivity. For example with increase in reaction temperature from 250\u00a0\u00b0C to 275\u00a0\u00b0C, MeOH selectivities over the C1\u20130.618 and C1\u20130.804 catalysts decreased by 19.1% and 37.5%, respectively. Whereas a decrease of only 7.8% in the MeOH selectivity over the C1\u20130.206 catalyst was recorded. It was interesting to note that with further increase in reaction temperature to 300\u00a0\u00b0C MeOH selectivities for all catalysts decreased. However, MeOH selectivity at an operating temperature of 300\u00a0\u00b0C for C1\u20130.206 catalyst was higher than that of the other catalysts. CO selectivity followed a similar trend where an increase in percentage selectivity with increase in reaction temperature was observed. The increase in CO selectivity and decrease in MeOH selectivity with increase in reaction temperature was due to enhancement of competing reverse water gas shift (RWGS) reaction. A similar behavior has also been reported by other researchers [43\u201345]. As shown in Fig. 9d and Fig. 9e, CO yield and MeOH yield over the C1\u20130.206 catalyst was higher than the C1\u20130.618 and C1\u20130.804 catalysts. The maximum MeOH yield of 0.24gMeOH/g-cat.h over the C1\u20130.206 catalyst was recorded under condition 3. The MeOH yield for the C1\u20130.618 and C1\u20130.804 catalysts was very similar. The higher yields of CO and MeOH was attributed to high CO2 conversions over the C1\u20130.206 catalyst. However, when reaction temperature was further increased to 300\u00a0\u00b0C a drop in the yield of MeOH due to decrease in MeOH selectivity was recorded.The results revealed that the C1\u20130.206 catalyst exhibited exceptionally high catalytic performance than the other catalysts. This exceptional high performance of the C1\u20130.206 nanocatalyst was related to three main differences;\n\n(a)\nComparatively smaller and uniformly distributed metal oxide nanoparticles. As revealed from the analysis results of XRD and TEM, the C1\u20130.2 catalyst had an average metal oxide particle size of 6.4\u00a0nm. This was smaller than that of the C1\u20130.6 and C1\u20130.8 catalysts where CuO average particle size was 21.3\u00a0nm and 22.1\u00a0nm, respectively. It is clear that high catalytic performance was due to smaller CuO crystallite size of the C1\u20130.2 catalyst. Since the catalysts C1\u20130.6 and C1\u20130.8 had almost similar particle size, therefore to no surprise a quite similar behavior of catalytic activity was recorded. Indeed, Cu-based catalysts have been extensively studied for CO2 hydrogenation reaction and the structure sensitive nature of the catalysts for CO2 hydrogenation to methanol is still debatable [46\u201348]. The illustrated results with our catalysts provide a direct evidence that the CO2 hydrogenation reaction was structure sensitive in nature and the C1\u20130.2 nanocatalyst with competitively smaller metal oxide particles exhibited exceptionally high activity.\n\n\n(b)\nThe high catalytic performance over C1\u20130.206 nanocatalyst might be related to the better dispersion of CuO nanoparticles as revealed by the XRD analysis, the diffraction lines of the C1\u20130.206 catalyst were broader with undistinguishable diffraction lines. By contrast, the XRD spectrum of C1\u20130.618 and C1O.804 were comparatively sharper with distinguishable diffraction lines of CuO and ZnO and alumina possibly due to agglomeration of the crystallites with poor dispersion. TEM analysis affirmed the findings from XRD analysis results. These findings were in accordance with previously reported literature [49\u201351].\n\n\n(c)\nThe high performance can also be attributed to the presence of surface defects, oxygen vacancies and induced phases. For example, the XPS analysis results the C1\u20130.206 catalyst revealed presence of around 34% of highly ionized and/or induced copper species (cC2+) whereas for the C1\u20131.26 catalyst the percentage of cC2+ ions decreased to around 26% only evidencing a direct relation of the presence of induced copper species with catalytic performance. Indeed, with increase in the number of induced copper species CO2 conversion and methanol production was expected to increase as previous studies [52\u201354] have also reported a similar correlation where the active sites of the CO2 hydrogenation were believed to be along with the Cu and support interferences and increase in induced phases improved the catalytic activity of the catalysts.\n\n\nComparatively smaller and uniformly distributed metal oxide nanoparticles. As revealed from the analysis results of XRD and TEM, the C1\u20130.2 catalyst had an average metal oxide particle size of 6.4\u00a0nm. This was smaller than that of the C1\u20130.6 and C1\u20130.8 catalysts where CuO average particle size was 21.3\u00a0nm and 22.1\u00a0nm, respectively. It is clear that high catalytic performance was due to smaller CuO crystallite size of the C1\u20130.2 catalyst. Since the catalysts C1\u20130.6 and C1\u20130.8 had almost similar particle size, therefore to no surprise a quite similar behavior of catalytic activity was recorded. Indeed, Cu-based catalysts have been extensively studied for CO2 hydrogenation reaction and the structure sensitive nature of the catalysts for CO2 hydrogenation to methanol is still debatable [46\u201348]. The illustrated results with our catalysts provide a direct evidence that the CO2 hydrogenation reaction was structure sensitive in nature and the C1\u20130.2 nanocatalyst with competitively smaller metal oxide particles exhibited exceptionally high activity.The high catalytic performance over C1\u20130.206 nanocatalyst might be related to the better dispersion of CuO nanoparticles as revealed by the XRD analysis, the diffraction lines of the C1\u20130.206 catalyst were broader with undistinguishable diffraction lines. By contrast, the XRD spectrum of C1\u20130.618 and C1O.804 were comparatively sharper with distinguishable diffraction lines of CuO and ZnO and alumina possibly due to agglomeration of the crystallites with poor dispersion. TEM analysis affirmed the findings from XRD analysis results. These findings were in accordance with previously reported literature [49\u201351].The high performance can also be attributed to the presence of surface defects, oxygen vacancies and induced phases. For example, the XPS analysis results the C1\u20130.206 catalyst revealed presence of around 34% of highly ionized and/or induced copper species (cC2+) whereas for the C1\u20131.26 catalyst the percentage of cC2+ ions decreased to around 26% only evidencing a direct relation of the presence of induced copper species with catalytic performance. Indeed, with increase in the number of induced copper species CO2 conversion and methanol production was expected to increase as previous studies [52\u201354] have also reported a similar correlation where the active sites of the CO2 hydrogenation were believed to be along with the Cu and support interferences and increase in induced phases improved the catalytic activity of the catalysts.The catalyst pre-treatments parameters have been reported to improve the physicochemical properties of the catalyst surface [55,56]. Therefore, investigation of the correct pretreatment and/or activation parameters is very important step for the development of efficient hydrogenation catalysts. In this regards, the effects of the activation temperature prior to reaction over the C1\u20130.206 catalyst was also studied. In this study, the catalyst was activated either at 500\u00a0\u00b0C (Activation 2) or at 350\u00a0\u00b0C (Activation 1) in presence of 30\u00a0ml/min of pure hydrogen and the activity results are displayed in Fig. 10\n. As can be seen, a significant performance difference between the activities of two activation treatments was recorded. As shown in Fig. 10a, there was a significant improvement in CO2 conversion with decrease in the activation temperature from 500\u00a0\u00b0C to 350\u00a0\u00b0C. For example at condition 1, an improvement of \u0360 17.9% in CO2 conversion was recorded for the catalyst activated at 350\u00a0\u00b0C compared to the activation at 500\u00a0\u00b0C. It is worth noticing that with the activation temperature at 350\u00a0\u00b0C a significant improvement in MeOH selectivity was also recorded. As shown in Fig. 10b and Fig. 10c, at an operating temperature of 250\u00a0\u00b0C with a pressure of 60 bars, selectivity for MeOH over the catalyst activated at 500\u00a0\u00b0C was 52.3%. Whereas, under similar reaction conditions for sample activated at 350\u00a0\u00b0C MeOH selectivity improved by 14%. At an operating temperature of 275\u00a0\u00b0C MeOH yield for the catalyst activated at 350\u00a0\u00b0C was 0.37 gMeOH/g-cat.h, which was notably higher than the catalyst activated at 500\u00a0\u00b0C (0.32 gMeOH/g-cat.h).Indeed, as was discussed in section 3.2, in light of the XRD and TEM analysis results it can be concluded that the activation temperature of around 350\u00a0\u00b0C resulted in reduction of Cu2+ to Cu0 without effecting the particle size. The decrease in overall activity with increase in the activation temperature to 500\u00a0\u00b0C was presumably because the increase in temperature during the pretreatment might result in agglomeration of the copper particles and thus giving rise to the increase in metal particle size and dispersion. This was in accordance with the results obtained for the effects of calcination temperature on metal particle size. It has been suggested that comparatively bigger copper particles have lower tendencies for hydrogen chemisorption and spill over and are therefore disadvantageous to methanol synthesis. Various researchers have reported similar trends of the effects of pretreatment on the catalytic activity during CO2 hydrogenation reaction [57\u201359].Previous studies have shown that calcination temperature exhibited profound effects on the catalytic performances of Cu-based catalysts [16,60\u201362]. Although there is a contradiction in the reported results, but in general, decrease in calcination temperature have been reported to result in an increase in overall catalytic performance. Moreover, as was discussed in section 3.2, calcination at a temperature of 400\u00a0\u00b0C was suitable to remove the residual surface carbon deposits with negligible impact on the metal oxide particle size. In addition, the catalyst activated at 350\u00a0\u00b0C was exceptionally more active than the catalyst activated at 500\u00a0\u00b0C. One can assume that the calcination temperature\u00a0<\u00a0600\u00a0\u00b0C might further improve the overall catalytic performance. Inspired by the above mentioned insights the effects of calcination temperature on the activity of the C1\u20130.206 catalyst was also investigated. In this study, portions of the C1\u20130.206 catalyst were calcined at a temperature of 400\u00a0\u00b0C and 600\u00a0\u00b0C. These catalysts were denoted as C1\u2013400 and C1\u2013600 and were investigated for catalytic tests.As can be seen in Fig. 11a, the C1\u2013400 catalyst exhibited exceptionally higher catalytic performance than that of C1\u2013600 catalyst. Compared to C1\u2013600, a significant improvement in CO2 conversion was recorded over the C1\u2013400 catalyst. For example at condition 1 (250\u00a0\u00b0C, 60bars pressure), CO2 conversion over the C1\u2013600 catalyst was 10% whereas for the C1\u2013400 catalyst CO2 conversion increased by \u0360 100%. It is worth to mention that for the C1\u2013400 catalyst a noticeable increase in MeOH selectivity and decrease in CO selectivity was also recorded. For example, at an operating temperature of 250\u00a0\u00b0C with a pressure of 60 bars, selectivities for MeOH and CO over the catalyst C1\u2013600 were 47% and 52%, respectively. Whereas, under similar reaction conditions, the MeOH selectivity over the C1\u2013400 catalyst increased to 57%, whereas CO selectivity decreased to 42%. At an operating temperature of 300\u00a0\u00b0C and pressure of 85bars, MeOH yield over the C1\u2013400 catalyst was 0.52gMeOH/g-cat.h. This was significantly higher than that of the C1\u2013600 catalyst where a MeOH yield of 0.35gMeOH/g-cat.h was recorded. Indeed, these findings were in accordance with the earlier test results presented in section 3.3. Since, the CuO particle size for C1\u2013400 was 4.48\u00a0nm which smaller than that of the C1\u2013600 catalyst where particle CuO was 7.4\u00a0nm and the CO2 hydrogenation over Cu-based catalysts is structure sensitive in nature [63\u201365], expectedly, the C1\u2013400 catalyst out performed C1\u2013600 catalyst during CO2 hydrogenation reaction. Moreover, C1\u2013400 exhibited the highest number of surface defects and oxygen vacancies. Thus, it will be reasonable to conclude that the synergy between smaller CuO particle size and presence of higher number surface defects of C1\u2013400 catalyst was responsible for higher catalytic activity.The effects of SCS synthesis variables such as G/O ratio, calcination and activation temperature on the catalytic performance of the Cu-based catalysts synthesized by SCS method was thoroughly investigated. The results suggest that the fuel deficit combustion mixture which was prepared at a G/O ratio of <0.618 resulted in porous materials with smaller metal oxide particle size, whereas the fuel enriched samples prepared at a G/O ratio of >0.804 turned less porous and bulky. Particularly, the catalyst prepared at a G/O ratio of 0.206 proceeded with a slow combustion, with no real flame but a large volume of gases kept exhausting until a complete combustion occurred. The highest catalytic performance for CO2 hydrogenation to methanol was achieved for the C1\u20130.206 catalyst prepared at G/O ratio of 0.206, calcined at 400\u00a0\u00b0C and activated in pure hydrogen at 350\u00a0\u00b0C. At an operating temperature of 300\u00a0\u00b0C, pressure of 85\u00a0bar, H2/CO2 ratio of 3.43 and GHSV of 7000\u00a0h\u22121 the CO2 conversion, CO selectivity and methanol selectivity over this catalyst were 30%, 38.60%, and 61.4%, respectively; whereas, methanol and CO production were 0.52gMeOH/g-cat.h and 0.33gCO/g-cat.h, respectively. The exceptional high catalytic performance of the C1\u2013400 catalyst attributes to the smaller CuO particle size, better dispersion and to the presence of interfaces and/or oxygen vacancies.\nSardar Ali: Conceptualization, Methodology, Data curation, Writing \u2013 original draft, Investigation. Dharmesh Kumar: Supervision, Writing \u2013 review & editing, Validation, Investigation. Kartick C. Mondal: Writing \u2013 review & editing. Muftah H. El-Naas: Supervision, Writing \u2013 review & editing.The authors have no competing interests to declare.This research work was made possible by the Qatar Shell Research and Technology Center (QSRTC) funded project QUEX-CENG-QSRTC18/19. The statements made herein are solely the responsibility of the authors.", "descript": "\n                  This work investigates the effects of solution combustion synthesis (SCS) variables on the performance of copper-based catalysts for CO2 hydrogenation to methanol. The catalyst with a composition of 30wt%CuO50%ZnO/Al2O3 was prepared at various glycine to nitrates (G/O) ratios in the range between 0.1 and 1.23. A correlation of the effects of calcination and activation temperatures with catalytic activity was also studied. The catalyst synthesized at a G/O ratio of 0.206, calcined in air at 400\u00a0\u00b0C and activated in a stream of pure hydrogen at a temperature of 350\u00a0\u00b0C resulted in a significant improvement in the performance of the catalyst for CO2 hydrogenation. The exceptionally high catalytic performance of the catalyst was attributed to the synergic effects between small well-dispersed CuO nanoparticles and high number of induced copper phases. The highest activity of the catalyst was recorded at an operating temperature of 300\u00a0\u00b0C, a pressure of 85\u00a0bar and GHSV of 7000\u00a0h\u22121. The CO2 conversion, CO selectivity and methanol selectivity under these conditions were 30%, 38.60%, and 61.4%, respectively; whereas, methanol and CO yields were 0.52gMeOH/g-cat.h and 0.33gCO/g-cat.h, respectively.\n               "}
{"full_text": "The presence of pharmaceutical chemicals in natural water is a well-known phenomenon that poses a major threat to aquatic biota\u00a0(Hong et al., 2021). Due to a lack of research on the environmental fate and behavior of lesser-known medicines, exposure assessments can be erroneous. Metoclopramide (MCP), a substituted benzamide and 4-aminobenzoic acid derivative well known as a dopamine antagonist and utilized as an antiemetic and analgetic, particularly in gastroenterology circumstances\u00a0(Dabi\u0107 et al., 2022), is one such chemical. Despite the fact that MCP is generally removed from the human body in its original form, its low biodegradability indicates that it is more sensitive to abiotic degradation in the aquatic environment\u00a0(Wielens\u00a0Becker et al., 2020).Over the past few decades, heterogeneous photocatalysis based on semiconductors has been recognized as the most powerful advanced oxidation process (AOPs) for efficiently degrading numerous detrimental organic chemicals and is an environmentally friendly, cost-effective, benign, and green process. Typical AOP techniques employ the generation of hydroxyl (\u2022OH) and other free radicals as powerful oxidizing agents that are able to mineralize organic pollutants into CO2, \n\n\nH\n\n\n2\n\n\nO, and degrade some inorganic species\u00a0(Husain\u00a0Khan et al., 2022). For example, photo-excited semiconductors can produce electron/hole (e\n\n\n\n\u2212\n\n\n/h\n\n\n\n+\n\n\n) pairs, which then participate in reduction\u2013oxidation (redox) reactions with dissolved oxygen and water molecules, to produce superoxide and hydroxyl radicals, respectively. These reactive radicals can degrade organic compounds ultimately leading to carbon dioxide and the mechanistic pathways have been discussed in the literature in detail\u00a0(Aliyan et al., 2013; Fazaeli et al., 2014).Layered double hydroxides (LDHs) with the general formula [M\n\n\n\n\n2\n+\n\n\n\n\n\n1\u2212x\n\n\n\nM\n\n\n\n\n3\n+\n\n\n\n\n\nx\n\n\n\n(OH)2][A\n\n\n\nn\u2212\n\n\n]\n\n\n\nx/n\n\n\n zH 2O (M \n=\n metal, A \n=\n anion) represent a class of two-dimensional anionic clays having attractive features that include large specific surface area, high stability, tunable layer elements and environmental friendliness\u00a0(Liang et al., 2015). The layered structures of these hydrotalcite materials are known to incorporate di- and trivalent metal cations, a wide range of organic or inorganic anions (e.g.\u00a0OH\n\n\n\n\u2212\n\n\n, CO\n\n\n\n\n3\n\n\n\n\n\n2\n\u2212\n\n\n\n\netc.) and water molecules between the slabs. They have been carefully investigated in the context of catalysis\u00a0(Sels et al., 1999), electroactive/photoactive materials\u00a0(Lee et al., 2011) and molecular sieves\u00a0(Villegas et al., 2003). The importance and versatility of LDHs and modified LDHs for the removal of differing organic pollutants from aquatic environments is well-recognized.\u00a0Karim et al. (2022) Specifically, [NiFe]-LDH materials consist of sheets of edge-shared nickel oxide octahedral, with varying amounts of ferric iron substituting at nickel sites\u00a0(Hunter et al., 2016). The excess positive charges of Fe3\uff0b substituting for Ni\n\n\n\n2\n+\n\n\n are balanced by interlayer anions and the hydroxide groups extend into the interlayer space, which also contain water\u00a0(Duan and Evans, 2006). LDHs possess semiconductor properties, which facilitate the transfer of the photogenerated electrons on the surface of the photocatalyst and could provide great potential in many applications\u00a0(Pau\u0161ov\u00e1 et al., 2015).LDHs have recently been coupled with magnetic nanoparticles (MNPs), such as Fe3O4, to exploit their combined characteristics\u00a0(Pengcheng et al., 2018). These adaptable composite materials are appearing in a variety of applications including targeted drug delivery, magnetic resonance imaging, photocatalysis and environmental remediation.A promising route to boost the photocatalytic activity of LDHs is to decorate the surface of these materials with metal nanoparticles. This process can improve the catalytic activity of LDHs and, by providing a support for the nanoparticles, address the issue of catalyst recyclability\u00a0(Chaturvedi et al., 2012). Like other supports,\u00a0Zhen and Sheng (2011) LDHs have been documented for successful immobilization of gold nanoparticles\u00a0(Varade and Haraguchi, 2012) and bimetallic NPs, (gold-palladium: Au-Pd)\u00a0(Sobhana et al., 2016), to develop efficient catalyst for organic reactions or photodegradation processes. In particular, plasmonic noble metals have been intensively studied for loading on semiconductors due to their unique characteristics\u00a0(She et al., 2016). After combining with semiconductors, the strong surface plasmon resonance (SPR) of the noble metal nanoparticle can enhance the absorption of visible light\u00a0(Samanta et al., 2014). Such visible-light-driven Z-scheme photocatalysts for degradation of aqueous pollutants\u00a0(Hassani et al., 2021) Simultaneously, the photogenerated electrons can be transferred from the noble metal and trapped in the conduction band of semiconductor due to the Schottky barriers formed at the metal\u2013semiconductor interface, while the holes can remain on the valence band\u00a0(Wang et al., 2011). This combination can lead to the detachment of photogenerated electrons from the excitation site of both semiconductors and metals and prevent the recombination of charge carriers to provide a better opportunity for their consumption in photodegradation. Plasmonic metallic nanoparticles, particularly Pd, have attracted tremendous attention due to the highly efficient photodegradation achieved through the combination of Pd with g-\n\n\nC\n\n\n3\n\n\nN4. Previous studies have demonstrated that Pd particles can serve as sinks to transfer electrons from the conduction band of photocatalysts under simulated solar light irradiation\u00a0(Li et al., 2016). Moreover, Pd particles can also exhibit local surface plasma resonance (LSPR) phenomenon, enabling enhanced visible light adsorption capability\u00a0(Yin et al., 2020).The present work reports a new composite material Pd-Fe3\n O4/NiFe-LDH, displaying a LDH with a combination of supported magnetic and noble metal nanoparticles. This novel material has been characterized by a broad range of structural and spectroscopic techniques. Furthermore, this material functions as a catalyst for the photodegradation of organic pollutant and is specifically effective for the photocatalytic degradation of metoclopramide (4-amino-5-chloro-N-(2-(diethylamino)ethyl)-2-methoxybenzamide, MCP), a model for pharmaceutical industry effluent.X-ray diffraction (XRD) patterns of the prepared samples were determined using an X-ray diffractometer model X\u2019PertPro (with Ni-filtered Cu-Ka radiation source at 1.5406\u00a0\u00c5, 40 kV, i 30\u00a0mA; Netherland). Fourier transform infrared (FTIR) spectrophotometer (Perkin Elmer Spectrum 65) was used for recording FTIR spectra. Ultraviolet\u2013visible (UV\u2013Vis) diffuse reflectance spectra (DRS) were recorded by a JASCO V 670 model instrument, Japan, against BaSO4 as reference. FESEM images were recorded by a MIRA3LMU scanning electron microscope (TESCAN Co., Czech Republic). A pH-meter (Jenway model 3505) was used for pH adjustment. A transmission electron microscope (TEM) Philips EM 208\u00a0s (100 kV) was used for recording the TEM images. A TOC analyzer (Analytik Jena, Germany) was used for the analysis of MCP samples before and after the photodegradation process. Brunauer\u2013Emmett\u2013Teller\u00a0(BET) specific surface areas and pore volumes of the catalysts were determined by nitrogen adsorption\u2013desorption at liquid nitrogen temperature using a Micromeritics TriStar II Series instrument. The samples were degassed at 623 K for 12\u00a0h at reduced pressure of 10\u22124 Pa before carrying out the adsorption measurements. A BAHR-STA-504 instrument was used to perform thermogravimetric analysis (TG) and differential thermal analysis (DTA). Thermal analyses were performed in the 25\u2013800\u00a0\u00b0C range at a heating rate of 10 K\u00a0min \u22121.The hydrothermal method was used to synthesize NiFe-LDH, however the hydrothermal method is energy-intensive due to requirement of elevated temperature. In a typical experiment, NiCl2. 6H2O and FeCl2. 6H2O in a 3:1 molar ratio were dissolved into a mixed solution of 10 mL of deionized water and 10 mL of absolute ethanol (metal-ion concentration 0.2 M) with 500 mg arginine; the resulting mixture was stirred continuously for 10\u00a0min. Then, 1\u00a0ml of NH3\n\n\n\nH\n\n\n2\n\n\nO was added to the above mixture and further stirred for 10\u00a0min. The mixture was then transferred to a Teflon-lined steel autoclave and heated at 180\u00a0\u00b0C for 12\u00a0h in an oven. Once cooled down to room temperature, the resulting precipitate was collected and washed with deionized water and ethanol several times and subsequently dried at 60\u00a0\u00b0C to yield the hierarchical porous NiFe-LDH\u200b microspheres.In a 25 mL round-bottom flask, 20 mg of Pd(OAc)2 (0.05 mmol) and 0.1\u00a0g of NiFe-LDH microspheres were combined along with 5\u00a0ml of ethanol and the mixture heated to reflux for 24\u00a0h. At this point, 0.02 mmol of NaBH4 was added to the reaction mixture, which was again heated to reflux for 4\u00a0h. The reaction mixture was cooled to ambient temperature, filtered and washed with ethanol to remove any unreacted Pd complex and finally dried at 80\u00a0\u00b0C in vacuo (Fig.\u00a01).\n\nThe photocatalytic experiments were carried out in a photocatalytic reaction chamber using a moderate pressure Hg lamp (35\u00a0W, Philips, type G-line with maximum emission at 435.8\u00a0nm) which was positioned 10\u00a0cm above the reaction beaker that was equipped with a magnetic stirrer (100 rpm). The degradation of MCP was monitored by UV\u2013Visspectroscopy. Direct photolysis was also studied under similar conditions on a solution in the absence of catalyst. The reaction suspension was centrifuged (\n>\n13,000 rpm) at regular times and the absorbance of the clear solution was recorded (at \n\u03bb\nmax \n=\n 212\u00a0nm for MCP). The values of the recorded absorbance, (\n\n\nA\n\n\n0\n\n\n and A, for the samples before and after irradiation process, respectively) were used in Eq.\u00a0(1) to calculate the change in concentration at time t and the percent decomposition, D. \n\n(1)\n\n\nD\n\n\n%\n\n\n=\n\n\n\n\n\n\nA\n\n\n0\n\n\n\u2212\nA\n\n\n\n\nA\n\n\n0\n\n\n\n\n\n\n\u00d7\n100\n=\n\n[\n\n\n\n\nC\n\n\n0\n\n\n\u2212\nC\n\n\n\n\nC\n\n\n0\n\n\n\n\n]\n\n\u00d7\n100\n\n\n\n\nThe LDH starting material was prepared using a water/ethanol solution of NiCl 2 and FeCl2, in a 3:1 stoichiometric ratio, with the addition of arginine. Addition of aqueous ammonia produced a basic reaction mixture that was heated to 180\u00a0\u00b0C in a Teflon-lined steel autoclave for 12\u00a0h to yield a precipitate of NiFe-LDH microspheres\u00a0Fu et al. (2017). The arginine plays a crucial role in this preparation by coordination to the metal ions and controlling the nucleation rate of the NiFe LDH (Fig.\u00a01b); The interaction between arginine molecules through hydrogen-bonding and electrostatic interactions has also been proposed to be influential in the self-assembly of primary nanocrystals, as is illustrated in Fig.\u00a01b. The Pd-nanoparticle functionalized NiFe-LDH was prepared by first adsorbing Pd(OAc)2 onto the NiFe-LDH microspheres in an ethanol suspension and then reducing the Pd(II) using NaBH4 in refluxing ethanol. Cooling, filtering, washing and drying the resulting solid yielded the Pd-Fe3\n O4/NiFe-LDH. At this stage, Fe3O\n\n\n\n4\n\n\n nanoparticles (NPs) were formed on the NiFe-LDH microspheres as evidenced by the following characterization.The NiFe-LDH precursor displayed characteristic features of layered material, as observed in the PXRD patterns of NiFe-LDH precursor sample (Fig. S1), with strong, symmetric lines at low 2\n\u03b8\n values and weak, less symmetric lines at high 2\n\u03b8\n values. The average crystallite size and the\u00a0d-spacing between the crystal lattice for NiFe-LDH are D \n=\n 12.36\u00a0nm and d \n=\n 2.41\u00a0\u00c5, respectively.The powder X-ray diffraction pattern (PXRD) for the Pd-Fe3\n O4/NiFe-LDH, along with the corresponding reflections for NiFe-LDH (ICDD-40-0216), Pd (ICDD-01-087-0641) and Fe3O4 (ICDD-01-089-4927) is shown in Fig.\u00a02a The assignment of the individual components if based on the comparison of the sample PXRD with the standard materials. For example, the diffraction peaks at 2\n\n\u03b8\n=\n11\n\n.6\u00b0, 23.4\u00b0, 34.4\u00b0, 39.0\u00b0, 46.5\u00b0, 60.5\u00b0 and 61.2\u00b0 were assigned to the (003), (006), (012), (015), (018), (110) and (113) reflections of the layered structure of NiFe-LDH\u00a0(Carvalho et al., 2015). Characteristic PXRD peaks for both Pd and Fe3O\n\n\n\n4\n\n\n are also displayed.\nThe surface features and porous nature of the parent FeNi-LDH and the Pd-Fe3\n O4/NiFe-LDH samples was measured by \n\n\nN\n\n\n2\n\n\n adsorption\u2013desorption isotherms and are shown in Fig.\u00a02b. In accordance with the International Union of Pure and Applied Chemistry (IUPAC) nomenclature, both samples show typical IV type isotherms and H1 type hysteresis looped at high relative pressures\u00a0(Ross, 2019). The inset to Fig.\u00a02b reveals that both samples exhibit pores of different sizes ranging from micro (\n<\n2\u00a0nm), meso (2\u221250\u00a0nm) to macro pores (\n>\n50\u00a0nm) and suggesting a hierarchical porous structure\u00a0(Barret and Massalski, 1980). As expected, the pore size and pore volume of the bulk NiFe-LDH are generally larger than those of the Pd-Fe3\n O4/NiFe-LDH (Table S2).The thermal analysis of NiFe-LDH and Pd-Fe3\n O4/NiFe-LDH was carried out and the results are shown in Fig.\u00a02c and Fig. S2. Thermogravimetric/differential thermal analyses measured over the range of 25\u20131000\u00a0\u00b0C revealed three events that all appear to be directly resulting from NiFe-LDH in the composite. The first (76\u2013194\u00a0\u00b0C, about 6.16%), was assigned to the removal of water from internal gallery surfaces and the external non-gallery surface of LDH carbonate. The second event (194\u2013283\u00a0\u00b0C, about 2.07%), was ascribed to the dehydroxylation of the brucite-like sheets and removal of interlayer anions with decomposition of arginine\u00a0(Weiss et al., 2018). The third event (238\u2013638\u00a0\u00b0C, about 9.54%) seems to correspond to several separate mass losses that have been assigned to loss of interlayer water, interlayer hydroxyl removal and interlayer anion decomposition (i.e.\u00a0CO\n\n\n\n\n3\n\n\n\n\n\n2\n\u2212\n\n\n\n decomposition/release of CO2)\u00a0(Luo et al., 2017).In order to probe the magnetic behavior of the Pd-Fe3\n O4/NiFe-LDH composites, magnetic measurements were carried out and compared with similar measurements of Fe3O4\u00a0(Luo et al., 2016) as shown in Fig.\u00a02d. At room temperature the saturation magnetization (Ms) values were 136.8 and 19.88 emu g\u22121 for Fe3O4 and Pd-Fe3\n O4/NiFe-LDH, respectively. The smaller saturation magnetization of the Pd-Fe3\n O4/NiFe-LDH nanocomposites was attributed to the existence of larger content of non-magnetic LDH phase\u00a0(Yan et al., 2015).The FT-IR spectrum of the NiFe-LDH and Pd-Fe3\n O4/NiFe-LDH are shown in Fig.\u00a02e and is similar to those generally reported for hydrotalcite compounds\u00a0(Pan et al., 2018). The bands around 3422\u00a0cm\n\n\n\n\u2212\n1\n\n\n was ascribed to the stretching mode of the OH group with hydrogen bonding and of interlayer water molecules\u00a0(Cheng et al., 2010). The band observed at 1630\u00a0cm\n\n\n\n\u2212\n1\n\n\n was attributed to the bending mode of crystalline water\u00a0(Sahu et al., 2013). The bands at approximately 829 and 530\u00a0cm\n\n\n\n\u2212\n1\n\n\n arise from the metal\u2013oxygen bond (M\u2013O, M\u2013O\u2013M and M\u2013OH) vibrations in the LDHs\u00a0(Wang et al., 2014). The four weak bands at 830, 1024 and 1392\u00a0cm \u22121 are characteristic of carbonate ions\u00a0(Hou et al., 2005). In addition, the preparation of this NiFe-LDH employed arginine and, therefore, the stretching modes of N-H from arginine were observed around 3150\u00a0cm \n\n\n\n\u2212\n1\n\n\n\u00a0(Kumar and Rai, 2010).The charge transfer between NiFe-LDH and the Pd and Fe3O4 particles was investigated by using photoluminescence spectroscopy (PL). The PL spectra analysis is essential to reveal the migration, transfer and separation efficiency of the photogenerated electron and hole pairs in various semiconductors\u00a0(Xu et al., 2009). Fig.\u00a02f shows the PL spectra of pure FeNi-LDH in comparison with the Pd-Fe3\n O4/NiFe-LDH composite. The pure LDH material (NiFe-LDH) showed one strong PL emission peaks at around 359\u00a0nm, due to the trapping of charge carriers in the form of excitons on its surface, and another peak at 728\u00a0nm is PL emission due to the defect sites of LDH\u00a0(Van\u00a0Vugt et al., 2005). If the combination of Pd-Fe3\n O4 with FeNi-LDH improves the e\n\n\n\n\u2212\n\n\n/h\n\n\n\n+\n\n\n separation compared to the parent material, the photoinduced recombination rate would be reduced, which leads to a reduced peak intensity for the FeNi-LDH nanocomposite. The observed intensity reduction confirms that the photogenerated holes and electrons undergo less recombination and are well-separated after Pd-Fe3\n O4 introduction.The scanning electron microscopy (SEM) image of the NiFe-LDH precursor shows a typical hexagonal nanoplatelet morphology with the edge length of about 100\u00a0nm (Fig.\u00a03a,b and Fig. S3). The SEM images of the Pd-Fe3\n O4/NiFe-LDH samples are shown in (Fig.\u00a03c,d and Fig. S4) and revealed a hierarchical porous architecture with numerous nanosheets as the building blocks. Fig.\u00a03e shows the particle size distribution of this material and indicated that the most common particle size in this composite material is \n<\n50\u00a0nm. The scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDX) spectrum of Pd-Fe3\n \u200b O4/NiFe-LDH samples is shown in Fig.\u00a03f and Fig. S5, which provided evidence of the expected elements Ni, Fe, and Pd in this composite. Interestingly, the peaks of Pd are distinct at 2.838 and 2.984 keV corresponding to Pd L\n\n\n\u03b1\n\n\n1\n\n\n and Pd L\n\n\n\u03b1\n\n\n2\n\n\n and confirmed the presence of metallic Pd consistent with Pd-NPs deposited on the NiFe-LDH surface. In addition, element mapping images of the Pd-Fe3\n O4/NiFe-LDH material is shown in Fig.\u00a03g and revealed a material that displayed uniform Pd, Fe and Ni distribution across the sample. The transmission electron microscopy (TEM) images and particle size distribution of the Pd-Fe3\n \u200b O4/NiFe-LDH particles, elucidated in Fig.\u00a03h and i (different magnifications are shown in Fig. S6) suggested that Pd nanoparticles, with average size of 7\u00a0nm, are homogeneously distributed on NiFe-LDH.\nNiFe-LDH and Pd-Fe3\n O4/NiFe-LDH showed absorption bands in both the UV and visible regions (Fig.\u00a04a) and these absorption bands can be broken into three characteristic regions: 200\u2013300, 300\u2013600, and 600\u2212800\u00a0nm. The intrinsic absorption band within 200\u2212300\u00a0nm in the UV region were assigned to the LMCT from the O 2p \n\u2192\n Ni 3d orbital, while the bands within the 300\u2212800\u00a0nm region corresponded to d-d transitions and are characteristic geometry of Ni\n\n\n\n2\n+\n\n\n ions in an octahedral field\u00a0(Wei et al., 2017). Another intense absorption band found at 476\u00a0nm was assigned to charge transfer for the transition of Ni\n\n\n\n2\n+\n\n\n\u2013O\u2013Fe3\uff0b to Ni\n\n\n\n+\n\n\n\u2013O\u2013Fe4\uff0b, originating from the induced MMCT for the oxobridged bimetallic linkage\u00a0(Nayak et al., 2015b). The absorption bands located at 383 and 750\u00a0nm corresponded to spin-allowed transitions of 3A\n\n\n\n2g\n\n\n(F) \n\u2192\n\n3T\n\n\n\n1g\n\n\n(P) and 3A\n\n\n\n2g\n\n\n(F) \n\u2192\n\n3T\n\n\n\n1g\n\n\n(F), which result from the characteristic d\n\n\n\n8\n\n\n configuration geometry of Ni\n\n\n\n2\n+\n\n\n ions in an octahedral field\u00a0(Rudolf et al., 2014). Nanosized Pd in Pd-Fe3\n O4/NiFe-LDH samples (Fig.\u00a04b) was also reported to exhibit a LSPR, but it is mainly located in the UV region of spectrum (300\u00a0nm), which impedes the observation of this feature\u00a0(De\u00a0Marchi et al., 2020).\nUsing a Tauc plot the UV\u2013Vis analysis can provide a means to determine the band gap energy (Eg) of the crystalline Pd-Fe3\n O4/NiFe-LDH by calculation using by Eq.\u00a0(2)\u00a0(Zeng et al., 2018): \n\n(2)\n\n\n\n\n(ah\u03c5)\n\n\n1\n/\nn\n\n\n= A(h\u03c5 \u2212 Eg)\n\n\n\n where a, h, v , Eg and A are the absorption coefficient, Planck\u2019s constant, light frequency, band gap energy, and a constant, respectively. In the case of an indirect optical transition of a semiconductor \n\nn\n=\n1\n/\n2\n\n\u00a0(Nayak et al., 2015a). Therefore, the plot of (\n\u03b1\nh\n\u03c5\n)2 vs. h\n\n\u03c5\n (Kubelka-Munk function as a function of light energy) gives the value of Eg \n=\n 3.56 eV corresponding to Pd-Fe3\n O4/NiFe-LDH is shown in Fig.\u00a04b.Photodegradation studies were carried out using metoclopramide (MCP) as a model pharmaceutical effluent. MCP is a benzamide derivative and is used as an anti-emetic in the treatment of some forms of nausea and vomiting and to increase gastrointestinal motility and this compound contains two substituents particularly reactive in radical reactions (chlorine, and amino group)\u00a0(Herrero et al., 1998). The chemical structure and physico-chemical characteristics of MCP are summarized in Fig. S6(a). Hydrochloric acid (0.1 mol L \u22121) and sodium hydroxide solutions (0.1 mol L\u22121) were used for the adjustment of water pH-value. Fig. S6(b) presents typical UV\u2013Vis spectra showing the three absorption bands for MCP (10 mg l\u22121) located at 212, 272, and 308\u00a0nm\u00a0(Chaabanea et al., 2013).Before examining the ability of the nanocomposite Pd-Fe3\n O4/NiFe-LDH to photocatalytically remove MCP several background measurements were carried out. The role of surface adsorption or dark reactions by the as-synthesized Pd-Fe3\n O4/NiFe-LDH composite on removal of MCP were first investigated. When the nanocomposite was added to a solution of MCP and held in the dark, there was an approximate 6% removal of MCP molecules during a 7\u00a0min time interval with no significant further changes with increased time. As a result, prior to all photodegradation experiments, suspensions of Pd-Fe3\n O4/NiFe-LD were shaken with MCP solutions at the dark for 7\u00a0min to allow the adsorption/desorption process to reach equilibrium (Fig.\u00a05a). The potential for direct photolysis of MCP solutions in the absence of any added catalyst was next examined and resulted in a decrease of about 9% of MCP molecules over a 30\u00a0min. time interval. While this background effect confirmed that irradiation alone is not sufficient to degrade MCP, it was an important background correction for the subsequent photodegradation experiments\nThe ability of the individual components of the Pd-Fe3\n O4/FeNi-LDH composite to degrade MCP was next examined. Fig.\u00a05a displays not only the background measurements but also the catalytic abilities of Fe3O\n\n\n\n4\n\n\nNPs, the support material FeNi-LDH, and two physically mixed materials Fe3O4/FeNi-LDH and Pd/FeNi-LDH. Each of these materials did demonstrate some level of catalysis for the destruction of MCP. However, substantially superior degradation efficiency was observed when the Pd-Fe3\n O4/NiFe-LDH nanocomposite was used. The effect of Pd loading on the photocatalyst performance was measured and is shown in Fig.\u00a05b. A clear and significant improvement of degradation was observed between 1 and 5% Pd loading. There was little improvement between 5 and 8% Pd loading. Further measurements were carried out with 7%Pd-Fe3\n O4/FeNi-LDH (Fig.\u00a05b).We examined a series of scavenging agents in order to probe the principle active species responsible for the photocatalytic degradation of MCP (Fig.\u00a06a). Specifically we examined the effect of the following added reagents: isopropanol (IPA: an \n\n\n\n\u2022\n\n\nOH scavenger); disodium salt of ethylenediaminetetraacetic acid (EDTA: a h\n\n\n\n+\n\n\n scavenger); benzoquinone (BQ: an \u2022O\n\n\n\n\n2\n\n\n\n\n\n\u2212\n\n\n\n scavenger); and \n\n\nH\n\n\n2\n\n\nO2 (an electron acceptor). The photodegradation efficiency was only marginally reduced when isopropanol (IPA) was added to the reaction cell as a \n\n\n\n\u2022\n\n\nOH scavenger, demonstrating that the \n\n\n\n\u2022\n\n\nOH radicals play no part in the degradation process. One reason may be that the holes in the photocatalyst VB were unable to oxidize OH\n\n\n\n\n\u2212\n\n\n/\n\n\nH\n\n\n2\n\n\n\nO to \n\n\n\n\u2022\n\n\nOH (OH\n\n\n\n\u2212\n\n\n\n\n+\n h\n\n\n\n+\n\n\n\n\n\u2192\n\n\n\n\n\n\u2022\n\n\nOH, E\n\n\n\no\n\n\n: 2.6 V vs. NHE). The addition of the other scavenging agents (i.e.\u00a0EDTA, BQ, and \n\n\nH\n\n\n2\n\n\nO2) reduced the degradation efficiency, showing the importance of superoxide and holes as reactive species in the degradation of MCP molecules\u00a0(Deng et al., 2017). We propose that the effectiveness of the full composite system is due to matching of the standard potentials between conduction (\n\n\nC\n\n\nb\n\n\n) and valence (\n\n\nV\n\n\nb\n\n\n) bands of the Pd and Fe3O4 with NiFe-LDH levels. A proposed mechanism for photocatalytic degradation of MCP by Pd-Fe3\n O4/NiFe-LDH composite and the transfer pathway of charge carriers is schematically illustrated in Fig.\u00a06b. Although in some particles both Fe3O\n\n\n\n4\n\n\n and NiFe-LDH semiconductors can produce e\n\n\n\n\u2212\n\n\n/h\n\n\n\n+\n\n\n pairs, they may recombine and act as independent semiconductors. However, in some particles, the photogenerated electrons in Fe3O\n\n\n\n\n4\n\n\n\u2212\n\n\nC\n\n\nb\n\n\n\n level migrate to that of NiFe-LDH because of its more negative potential\u00a0(Li et al., 2018). This internal reduction\u2013oxidation process extends the lifetime of the photogenerated e\n\n\n\n\u2212\n\n\n/h\n\n\n\n+\n\n\n pairs and enhances the degradation efficiency (Fig. 15). Given the difference in the work function of metals and semiconductors, a Schottky barrier could be formed between Pd and Fe3O4\u2013NiFe-LDH under UV\u2013Vis illumination. In this model the Pd nanoparticles absorbed the resonant photons and the photogenerated electrons due to SPR would be transferred from Pd to NiFe-LDH until the two levels reached equilibrium to form a new Fermi energy level\u00a0(Su et al., 2012). These photogenerated electrons would be transferred to the \n\n\nC\n\n\nb\n\n\n of NiFe-LDH. The complex heterojunctions between the two components facilitated the transfer of photogenerated electron\u2013hole pairs during the course of photocatalysis\u00a0(Seery et al., 2007).\n\n\n\nKinetic analysis of this catalytic process was consistent with the Langmuir\u2013Hinshelwood (L-H) model\u00a0(Rezaei and Nezamzadeh-Ejhieh, 2020). Catalyst performance was quantified in terms of the apparent rate constant \n\n\nk\n\n\na\np\np\n\n\n derived from the slope of ln(\n\n\nC\n\n\n0\n\n\n\n/C) versus time. \n\n\nC\n\n\n0\n\n\n\n/C was obtained from the absorbance at \n\n\n\u03bb\n\n\nmax\n\n\n\n (\n\n\nA\n\n\n0\n\n\n\n/A), where A represents the MCP absorbance at time t. Fig.\u00a07a presents the data as a function of catalyst dose on the rate of the degradation reaction. The plot of ln(\n\n\nC\n\n\n0\n\n\n/C) versus time revealed a linear dependence consistent with first order kinetics (Fig.\u00a07a, inset). The rate constants as the function of the catalyst dose were obtained from the slopes of these linear curves and summarized in Table\u00a02. The results illustrate an increase in the rate when the amount of the catalyst was increased from 0.1 to 0.4 g/L (k from 0.0196 to 0.0263\u00a0min \u22121) and thereafter decrease in the rate was observed. The increase in the rate was simply attributed to having more active sites of the catalyst as the amount of catalyst is increased. At higher doses, agglomeration and light scattering effects appear to lead to a decrease in rate of degradation, as has been illustrated in detail in the literature\u00a0(Norouzi et al., 2021). Thus, the dose of 0.4 g/L was selected for the next experiments. As shown in Fig.\u00a07b and Table\u00a01, an increase in the MCP from 5 to 20 mg/L lead to an increase in the MCP photodegradation rate consistent with a first order dependence on the concentration of this reactant. Interestingly, at a high concentration of the MCP pollutant the degradation rate decreased\u00a0(Y\u0131lmaz et al., 2015). This is likely due to a screening effect of MCP molecules resulting in a decrease in the photoexcitation of the catalyst. Thus, the MCP solutions with a concentration of 20 g/L and were selected for the next experiments. The effects of initial pH of MCP solution on the degradation efficiency are shown in Fig.\u00a07c, for pH values from 3.5 to 9.5. The highest rate of degradation of MCP was obtained at pH 6.5. Estimation of pH\n\n\n\nPZC\n\n\n\u00a0(Dianat, 2018) (Fig.\u00a07d) showed that the Pd-Fe3\n O4/NiFe-LDH photocatalyst has a pH\n\n\n\npzc\n\n\n of 3.8\u20134.0. At pH values <4.0, the surface of the catalyst was positively charged and repels the protonated dye molecules. This suggests that the negative charge of Fe3O4/NiFe-LDH photocatalyst at pH \n>\n 4.0, results in the protonated functional groups adsorbing to the surface of the catalyst and the attractive force between the cationic functional groups and the negatively charged catalyst adsorb MCP species and the degradation efficiency tends to increase\u00a0(Dimitrakopoulou et al., 2012).\n\nThese optimal conditions of catalyst dosing, MCP loading pH, were used for a set of photodegradation experiments with irradiation times ranged from 0 to 30 min. Based on the disappearance for the recorded absorbance for MCP, a typical Hinshelwood plot was constructed as shown in Fig.\u00a07e. The apparent first order behavior yielded a rate constant of 0.0265\u00a0min \u22121 corresponding to a half-life of 26.25\u00a0min for the MCP photodegradation. The extent of mineralization of these solutions was determined by analysis of the chemical oxygen demand (COD)\u00a0(Aliyan et al., 2013). The residual COD values are presented in Fig.\u00a07f which confirmed a decrease in the COD from the photodegradation process. The inset of Fig.\u00a07f shows the Hinshelwood plot obtained from the COD results. The COD measurements yielded a rate constant, k, of 0.0420\u00a0min \u22121 corresponding to \n\n\n\nt\n\n\n1/2\n\n\n=\n16\n.\n50\n\n min. The mineralization of MCP is about 1.59 times faster than its photodegradation extent. Generally, mineralization is a slower process associated with the formation of persistent transformation by-products. However, process performance is affected by several factors, namely irradiation time, photocatalyst type and loading, solution pH and the water matrix\u00a0(Omrani and Nezamzadeh-Ejhieh, 2020).\nThermodynamic functions\nThe effect of the temperature on the MCP photocatalytic degradation by the Pd-Fe3\n O4/NiFe-LDH nanocomposite was evaluated under the optimized conditions in the temperature range of 298\u2013323\u00a0K. Fig.\u00a08a shows a plot of ln(\n\n\nC\n\n\n0\n\n\n/C) for various reaction temperatures during the time interval of 5\u201330\u00a0min. From this data the apparent rate constant (\n\n\nk\n\n\napp\n\n\n) as a function of the reaction temperature was extracted and presented in Table\u00a02. From a plot of ln \n\n\nk\n\n\napp\n\n\n versus 1/T\u00a0(Harbourne et al., 2008) (Fig.\u00a08b) the activation energy was obtained and is given in Table\u00a02. The other thermodynamic parameters were calculated (Table\u00a02) using the activation energy and apparent rate constant\u00a0(Garsoux et al., 2004). Then, the plot of ln (k/T) versus 1/T was constructed (Fig.\u00a08c) and the values of \n\u0394\nH\n\u2021\n and \n\u0394\nS\n\u2021\n was calculated from the slope and intercept, respectively. Chen and Ray (1998) reported that the increase in rate constant is most likely due to the increasing collision frequency of molecules in the solution that increases with increasing temperature. As shown, the MCP photodegradation by the Pd-Fe3\n O4/NiFe-LDH photocatalyst was accompanied by a relatively high positive activation enthalpy and the Gibbs free energy values, indicating that a highly hydrated transition state complex was produced. These positive values also confirmed that for reaching this transition state complex needs enough energy and it cannot produce at ambient conditions\u00a0(Gupta et al., 2015).The performance of the photocatalyst in different real water samples was tested and the results are shown in Fig. S8. The best photocatalytic activity was observed in the distilled water sample. Regarding other real examples, it can be said that the relative reduction of degradation is possibly due to the presence of other mineral compounds. The lowest rate of MCP degradation was obtained in the sewage water sample.\nRecyclability of the catalyst\nTo further evaluate the performance of the Pd-Fe3\n O4/NiFe-LDH composite, this material was subjected to five catalysis cycles. Before each run, the recycled catalyst was dried at 100\u00a0\u00b0C for 30\u00a0min to remove the adsorbed species, the specific performance (Table S3) revealed good stability and reusability.Moreover, the filtrate was examined by ICP-MS analysis, and after hot filtration, no trace of Pd was detected in the filtrate. In addition, FT-IR and XRD results of catalysts before and after recycling are shown in Fig. S9. No obvious change was observed, the spectral patterns of catalysts before and after use were identical, which is strong evidence of the stability of catalysts. However, after the fourth run, during successive uses of the catalyst, a small amount of leaching of Pd from the catalyst surface (approx. 0.05%, according to ICP analysis) was observed and may decrease the efficiency of catalyst.\nConclusion\nIn summary, a novel catalyst, Pd-Fe3\n O4/NiFe-LDH, was prepared under hydrothermal conditions and characterized by XRD, BET, TG-DTG, VSM, FTIR, PL, SEM/EDX, TEM, and DRUV analysis. The photocatalytic activity of Pd-Fe3\n O4/NiFe-LDH for the photodegradation of MCP was investigated under visible light irradiation. The results showed that the photodegradation efficiency of MCP by Pd-Fe3\n O4/NiFe-LDH was 95.2% in 80\u00a0min, and the photodegradation rate was higher than that of pure NiFe-LDH or Fe3O4/NiFe-LDH. Furthermore, Pd-Fe3\n O4/NiFe-LDH maintained good photocatalytic recyclability. It appears that the SPR effect of the Pd NPs accelerated the separation of photoexcited e\n\n\n\n\u2212\n\n\n/h\n\n\n\n+\n\n\n couples. Furthermore, the porous interior cavities of this material may have created many reflections of the arriving photons, greatly lengthening the action time. Simultaneously, the ordered mesoporous opening structure of the catalyst, which greatly boosts the photocatalytic activity of the Pd-Fe3\n \u200b O\n\n\n\n4\n\n\non the surface of the mesoporous silica, may efficiently enable the transfer of reactant molecules. The superparamagnetic Pd-Fe3\n O4/NiFe-LDH particles may allow an avenue for\u00a0easily separating\u00a0the catalysts from the reaction medium using external\u00a0magnetic\u00a0fields. Finally, the use of this composite photocatalyst made of Pd-Fe3\n O4/NiFe-LDH for the treatment of mineral processing wastewater, may lead to the development of photo-, photoelectro-systems for the treatment of genuine mineral processing of wastewater.\nForouzan Shabib: Data curation, Writing \u2013 original draft, Visualization, Investigation, Software. Razieh Fazaeli: Conceptualization, Methodology, Supervision, Software, Writing \u2013 review & editing, Validation. Hamid Aliyan: Data curation, Writing \u2013 original draft, Visualization, Investigation, Software. Darrin Richeson: Conceptualization, Methodology, Supervision, Software, Writing \u2013 review & editing, Validation.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.We gratefully thank Shahreza Branch, Islamic Azad University for financial support.Supplementary material related to this article can be found online at https://doi.org/10.1016/j.eti.2022.102515.The following is the Supplementary material related to this article. \n\nMMC S1\n\n\n\n", "descript": "\n                  With the primary objective to develop efficient, environmentally benign, visible-light-driven heterogeneous catalysts for the degradation of metoclopramide (MCP), a waste-water pollutant, a newly prepared heterogeneous composite Pd-Fe3O4/NiFe-LDH catalyst is reported. This material has been characterized with a wide range of analysis methods (i.e. XRD, SEM-EDS, BET, FTIR, TG-DTG, DTA, DRUV and TEM analysis). When applied to aqueous solutions of MCP, Pd-Fe3O4/NiFe-LDH displayed a photocatalytic degradation of MCP with an efficiency of 95.2% in 80 min. The photocatalytic degradation rate for this composite material was higher than that of pure NiFe-LDH or Fe3O4/NiFe-LDH. The best photocatalytic activity was obtained at pH 6.5, with 0.4 g/L of the catalyst. Application of the Arrhenius equation yielded an activation energy for this process of 13.4 kJ/mol. A negative activation \n                        \u0394\n                     S\u2021\u00a0(-0.25 kJ/mol) with the positive \n                        \u0394\n                     H\u2021\u00a0and \n                        \u0394\n                     G\u2021\u00a0values were obtained for this MCP photodegradation.\n               "}
{"full_text": "Data will be made available on request.Facing one of the acute environmental crises, i.e., global warming that is attributed mainly to the massive CO2 emission via the combustion of fossil fuels, more and more efforts from both industrial and academic researchers have been made to reduce atmospheric CO2 or, even better, to transform it into value-added chemicals and thus advocate a circular carbon economy. CO2 can be utilized as a promising carbon feedstock for various functional molecules, e.g., methanol [1,2], higher alcohols [3,4], dimethyl ether [5] and C2+ hydrocarbons [6\u20138]. The transformation of CO2 into methanol, which is well recognized as an important bulk chemical in industries [9,10], is of interest in this study. In general, the activation of CO2 molecules remains a challenge in CO2 conversion due to its notorious high thermodynamic stability (\n\n\u0394\n\nH\n\nf\n\n0\n\n\n\u00a0=\u00a0\u2212393.5\u00a0kJ\u00a0mol\u22121) [11]. Moreover, the methanol synthesis from CO2 is known from the literature to undergo a competition with the reverse water gas shift (RWGS) reaction forming CO and possibly the conversion of CO into methanol [12]. While RWGS is endothermic, the methanol synthesis is exothermic (\n\n\u0394\n\n\nH\n\n0\n\n\n\u00a0=\u00a0\u201349.5\u00a0kJ\u00a0mol\u22121) and thermodynamically unfavored at high temperatures [10,13]. Therefore, catalysts are crucial to lower the activation energy and consequently enhance the yield of methanol in the thermocatalytic conversion of CO2. Alternatively, CO2 can be converted using various approaches, including molecular catalysis [14], photocatalysis, electrocatalysis [15] and hybrid approaches [16], which are out of the scope of this study.There have been numerous attempts to improve the catalytic activity via catalyst design. Regarding the active metals, three main groups have been studied extensively, namely, Au-, Pd- and Cu-based catalysts, besides other metals (Pt, Ni, Co, Ru, \u2026) [5,12,14]. The selection of catalysts is further extended to bimetallic [17] and trimetallic catalysts [18] in order to advance the synergy of the multicomponent catalytic systems. Irrespective of active metals employed, the particle size and the dispersion of active metals were found crucial to the catalytic performance in the hydrogenation of CO2. In particular, the increase of Cu particle size (5\u201325\u00a0nm), in association with gradual raising Cu content (5\u201325\u00a0wt.%) supported on CeO2-ZrO2 materials, leads to a reduction in the number of active Cu sites exposed to reactants, as indicated by lower H2 consumption during temperature programmed reduction, and thus lower methanol formation activity (from 2.92 to 0.46 \n\n\ng\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\u2219\n\n\n\n\n\ng\n\nCu\n\n\n\u2219\nh\n\n\n\n\n-\n1\n\n\n\n) [19]. Similar observations are also recorded for Cu-based catalysts supported on various types of materials, e.g., CeO2\n[20], Al2O3\n[21], ZnO [22], and ZrO2\n[23]. Au-based catalyst is another classic example of the strong dependence between Au particle size and the corresponding catalytic activity [24\u201326]. The highly dispersed Au nanoparticles (\u22481 nm) supported on amorphous ZrO2 via deposition precipitation method provided a much higher methanol formation rate than that of Au/ZrO2 (d\nAu\u224850\u00a0nm) synthesized by impregnation, i.e., 2.1\u00a0>\u00a00.4 \n\n\ng\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\u2219\n\n\n\n\n\ng\n\ncat\n\n\n\u2219\nh\n\n\n\n\n-\n1\n\n\n\n in the methanol synthesis from CO2 (240\u00a0\u00b0C, 40\u00a0bar) [27]. It is suggested that the particle size of active metal is strongly associated to the number of interfacial surface area, at which the methanol synthesis has been repeatedly reported to take place [28,29]. Hence, the high specific surface area of interfacial sites can promote the hydrogenation of CO2 into methanol and thus lower the formation of CO.Besides, there are other factors affecting the catalytic performance of solids, in particular, reducibility indicated by the temperature at which most of the active metal is reduced, the chemical nature of active sites and the corresponding role of each site, and active sites dispersion determined by the specific surface area of support materials as well as the interaction between the metals and supports. Therefore, one of the approaches to improve the catalytic performance is via support materials.Regarding the support materials, there are a plethora of choices, among these, \u03b3-Al2O3, ZnO, CeO2, ZrO2, SiO2, TiO2, and mixed oxides of various rare-earth and transition metals are widely studied in the hydrogenation of CO2. One of the key roles of these materials is providing high surface area and thus improving the dispersion of the active metals. Besides, the acid-base properties of the catalyst support are also required to influence the catalytic activity of CO2 conversion greatly. The catalysts supported on acidic materials (TiO2, ZrO2) are beneficial for CO2 conversion [30,31], which is contributed mainly by the RWGS reaction forming CO, whereas methanol formation is favored over catalysts exhibiting strong basic properties (CeO2, ZnO) coupled with low CO2 conversion [31\u201333]. The different behaviors were speculated to associate with the high basicity, which reinforced the adsorption of CO2 and CO, thereby stabilizing the formate intermediates to methanol and possibly hindering the production of by-product CO, respectively [34]. Moreover, acidity caused by Cu cations adjacent to oxygen vacancies was found to be linear with the methanol formation rate [35]. However, further understanding, particularly the influencing manner of these basic/acidic sites in CO2 hydrogenation, remains debatable. Furthermore, the combination of different metal oxides, particularly CeO2-ZrO2 mixed oxides, has shown numerous benefits, including greater dispersion, reducibility, and interaction between metallic sites, e.g., Cu and support [33,36\u201338]. Taking into consideration, for example, the Au-based catalysts on TiO2, due to the electronic polarization between Au and the support only observed in the presence of CeO2, both CO2 adsorption and activation were greatly improved [39]. Additionally, the surface hydroxyl groups available on metal oxides were found to participate in the surface chemical reaction and affect the selectivity of CO2 conversion into methanol. These groups were found to promote the adsorption of CO2 at the isolated Cu sites in proximity resulting in bidentate formate, the primary reactive intermediate of methanol synthesis, for urea-assisted hydrothermally synthesized Cu/SiO2\n[40]. Another study combining experimental and theoretical results for SiC quantum dots revealed that the hydroxyl groups lowered the barrier to form formate intermediates and thus increased the methanol productivity [41]. This phenomenon was also observed in various catalytic systems, e.g., Rh-based catalysts supported on TiO2\n[42] and Cu/\u03b3-Al2O3\n[43]. These are a few factors important for methanol synthesis, apart from the morphology of the supports [23,44\u201348], which are not discussed in detail. Nevertheless, the changes in catalytic activity seem to be under the influence of multiple factors. Thus, one should consider all possible reasons and their corresponding gravity to obtain a full spectrum based on which solid conclusions can be drawn.In our study, the objective was to exploit the advantages of Ce, i.e., great basicity and surface oxygen vacancy, in Au catalysts supported on amorphous ZrO2 (a-ZrO2) for CO2 hydrogenation. Owing to the great number of uniform acidic sites mainly contributed by Lewis acid sites and a fraction of surface hydroxyl groups as reported in [49], a-ZrO2 was chosen over monoclinic (m-) and tetragonal (t-) ZrO2, two commonly found phases of ZrO2. Furthermore, CeO2-ZrO2 mixed oxides were obtained by a simple and low-energy cost coprecipitation method, which is highly desired amid the current energy crisis [27]. Accordingly, a series of Au catalysts supported on CeO2-ZrO2 mixed oxides with a gradually increased molar ratio of Ce to Zr (from 0 to 0.1, with a step of 0.025) was prepared. The gradual increase of Ce content offers a systematic approach to investigating the influence of Ce on the properties of the resulting catalysts based on a-ZrO2. Furthermore, the catalytic activity of the Au-based catalysts supported on CeO2-ZrO2 mixed oxides in the hydrogenation of CO2 into methanol was examined and elucidated by considering various influential factors, e.g., reducibility, acid-base properties, chemical environment (of Au, Ce, Zr), and Au particles size of the catalysts.The ZrO2-CeO2 nanocomposites were synthesized following a precipitation method in a previously published work [27] with slight adjustments. Typically, 4\u00a0g of zirconium (IV) oxynitrate (ZrO(NO3)2\u00b7xH2O, Sigma Aldrich) and a predetermined amount of cerium (III) nitrate hexahydrate (Ce(NO3)3\u00b76H2O, Sigma-Aldrich) was dissolved in 50\u00a0mL deionized water resulting in an aqueous solution of 0.1\u00a0M Zr4+ and 0.1x M Ce3+, respectively, with\u00a0x\u00a0denoted for the Ce to Zr molar ratios (x\u00a0=\u00a0n\nCe/n\nZr). To obtain high dispersion of Ce in ZrO2, x was varied in the low range from 0.000 to 0.025, 0.05, 0.075 and 0.100. The mixture was stirred at room temperature (RT) for 30\u00a0min. Subsequently, the precipitation was conducted by adding 5\u00a0wt.% ammonia solution (pH\u00a0=\u00a012) dropwise (approximately 10\u00a0mL) to the Zr4+ and Ce3+ aqueous solutions (pH\u00a0=\u00a01\u20131.5) until complete precipitation (pH \u2248 9). The mixtures were then placed in a dryer at 80\u00a0\u00b0C under static air for 12\u00a0h. Afterwards, the obtained mixtures were washed and filtered with deionized water. The solids were dried at 90\u00a0\u00b0C overnight and subsequently calcined at 300\u00a0\u00b0C under static air for 4\u00a0h. The obtained products are labelled as ZrCex with\u00a0x\u00a0referring to the nominal n\nCe/n\nZr ratio.In order to obtain small Au particle size, which is favored at low Au content, a nominal Au content of 1\u00a0wt.% was chosen for Au introduction. Specifically, 1\u00a0g of ZrCex was added into 50\u00a0mL of 1\u00a0mM gold (III) chloride trihydrate (HAuCl4\u00b73H2O, Sigma-Aldrich) aqueous solution. The suspension was stirred at room temperature (RT) for 30\u00a0min. The deposition precipitation was conducted by adding 5\u00a0wt.% ammonia solution (approximately 3\u00a0mL) in droplets until complete precipitation (pH\u00a0=\u00a07). The suspension was then transferred to a dryer, and the temperature was kept at 80\u00a0\u00b0C under static air for 6\u00a0h. Afterwards, the solids were filtered and washed with water. The obtained materials were further dried at 120\u00a0\u00b0C for 2\u00a0h and named as Au/ZrCex (x\u00a0=\u00a0n\nCe/n\nZr\u00a0=\u00a00, 0.025, 0.05, 0.075 and 0.1).The structural properties of all investigated catalysts were characterized by a PANalytical X\u2019Pert PRO MPD X-ray diffractometer (Cu K\u03b11\u00a0=\u00a00.154\u00a0nm). The powder X-ray diffraction patterns were recorded at room temperature in the 2\u03b8 range from 5\u00b0 to 90\u00b0 with a step of 0.033\u00b0.The elemental content of Au, Ce, and Zr in synthesized catalysts was determined by optical emission spectroscopy with inductively coupled plasma (ICP-OES) using a Varian 715-ES ICP Optical Emission Spectrometer. In preparation for the analysis, 10\u00a0mg of the samples were dissolved in 10.0\u00a0mL HF and 1.0\u00a0mL HClO4 and diluted to obtain 50.0\u00a0mL aqueous solutions.N2 sorption isotherms were recorded on a volumetric adsorption analyzer Tristar 3000 (Micromeritics, Norcross (GA), USA) at 77\u00a0K. Prior to the measurements, the samples were evacuated at 180\u00a0\u00b0C for 10\u00a0h. The specific surface area (A\nBET) and total pore volume (V\nP) were determined using the Brunauer\u2013Emmett\u2013Teller (BET) model and single point (p/p0\n\u00a0=\u00a00.98) method, respectively.Temperature-programmed reduction with H2 (H2-TPR) was performed using a Micromeritics AutoChem II 2920 chemisorption analyzer. 60\u00a0mg of the catalysts was inserted into a U-shaped quartz tube and oxidized at 400\u00a0\u00b0C under 5\u00a0vol% O2 in argon (Ar) (25\u00a0mL\u00a0min\u22121) for 10\u00a0min with a heating rate of 10\u00a0K\u00a0min\u22121. Subsequently, before the reduction of the samples, the gas line was switched to Ar (25\u00a0mL\u00a0min\u22121) for O2 evacuation (50\u00a0\u00b0C, 10\u00a0min) followed by reducing the catalysts using the mixture of 5\u00a0vol% H2 in Ar (25\u00a0mL\u00a0min\u22121) at a constant heating rate of 10\u00a0K\u00a0min\u22121 up to 350\u00a0\u00b0C. Hydrogen consumption was determined using a TCD detector and a calibration using Ag2O (from Micromeritics) as a reference.Temperature-programmed desorption with CO2 (CO2-TPD) analyses were carried out using the aforementioned AutoChem II 2920 chemisorption analyzer (Micromeritics, USA) coupled with a mass spectrometer (Pfeiffer Vacuum, model ThermoStar). The samples were placed into a U-shaped quartz tube and reduced at 300\u00a0\u00b0C under 5\u00a0vol% H2 in Ar (25\u00a0mL\u00a0min\u22121). The samples were then cooled down to 50\u00a0\u00b0C and flushed with Ar (25\u00a0mL\u00a0min\u22121) for 15\u00a0min. At the same temperature (50\u00a0\u00b0C), the reduced samples were saturated with 80\u00a0vol% CO2 in Ar via 20 pulses of 0.532\u00a0mL. The average peak area of the last 15 pulses was used as the calibration for the quantification of CO2 desorption. Subsequently, the samples were heated with a heating ramp of 10\u00a0K\u00a0min\u22121 up to 400\u00a0\u00b0C under Ar (25\u00a0mL\u00a0min\u22121) for CO2 desorption. The amount of desorbed CO2 was determined via integration of the total area under the MS fragment m/z\u00a0=\u00a044.The pyridine adsorption-desorption experiments were carried out using a Perkin Elmer Pyris 1 TGA instrument to determine the acid site density of reduced catalysts. The as-synthesized catalysts were reduced externally in a tubular oven at 350\u00a0\u00b0C in H2 flow (50\u00a0mL\u00a0min\u22121) for 4\u00a0h. Subsequently, the samples were placed in a sample pan and pretreated at 350\u00a0\u00b0C in nitrogen flow (50\u00a0mL\u00a0min\u22121) for 30\u00a0min, then cooled down to 50\u00a0\u00b0C at which the sample weight was recorded and referred to as m0 (g). The pyridine saturation step was carried out by flushing with pyridine vapor in nitrogen flow (50\u00a0mL\u00a0min\u22121) until recording a constant weight. The excess pyridine was removed by purging with nitrogen flow (50\u00a0mL\u00a0min\u22121) at 50\u00a0\u00b0C for 2\u00a0h. The temperature was then increased to 450\u00a0\u00b0C, with a heating rate of 20\u00a0K\u00a0min\u22121. The desorption of pyridine was recorded via the sample weight loss as a function of temperature/time. The weight loss during the desorption of pyridine while heating from 50\u00a0\u00b0C to 450\u00a0\u00b0C is \u0394m (mg). Assuming the stoichiometry between pyridine molecule and acid site is equal to 1, the acid site density (ASD) was, therefore, calculated as follows \n\nASD\n=\n\n\n\u0394\nm\n\n\n\n\nM\n\npyridine\n\n\n\u2219\nm\n\n0\n\n\n\n (mmol\u22c5g\u22121) with Mpyridine\u00a0=\u00a079.1\u00a0g/mol.The carbon content of the fresh and spent catalyst samples were determined using a 2400 series II CHNS elemental analyzer (Perkin Elmer, USA).The crystallography, phase composition, morphology, and size of the samples containing Au nanoparticles were analyzed by transmission electron microscope (TEM, JEM-2100, JEOL), operating at 200\u00a0kV and equipped with a high-resolution slow-scan CCD camera (Orius SC1000, Gatan). The powdered samples pre-reduced in a tubular oven (in the H2 flow (40\u00a0mL\u00a0min\u22121) at 350\u00a0\u00b0C for 4\u00a0h) were dispersed in absolute ethanol and sonicated to prevent agglomeration. The suspension was transferred onto Cu-supported amorphous carbon grids. The maximum Feret diameter was used as the size descriptor of Au nanoparticles, which was manually outlined and determined using ImageJ. To obtain a representative overview of Au particles, the TEM micrographs were recorded at various sites of interest for each sample, and the number of particles detected is up to 150 particles. Before the TEM investigation, the microscope image and diffraction mode were calibrated by the MAG*I*CAL\u00ae reference standard through all of the magnification ranges, with the overall uncertainty on the calibrated values \u0394t\u00a0<\u00a01.0 %.X-ray photoelectron spectroscopy (XPS) measurements were performed with a Supra plus instrument (Kratos Analytical, Manchester, UK) equipped with an Al K\u03b1 excitation source and a monochromator. The measurements were performed with a spot size of 700\u00a0\u00d7\u00a0300\u00a0\u00b5m. Pass energy of 160\u00a0eV and 20\u00a0eV were used to obtain a survey and high-resolution spectra, respectively. The binding energy scale was corrected using the C-C/C-H peak at 284.8\u00a0eV in the C 1s spectrum. The neutralizer was on during the spectrum acquisition. Data were acquired and processed using ESCApe 1.4 (Kratos, Manchester, UK). The background of the high-resolution spectra was subtracted according to the method of Shirley [50]. Reported atomic concentrations at the surface were normalized to 100.0 %.The hydrogenation of CO2 was carried out in a fixed-bed reactor (Microactivity Reference MA-Ref reactor from PID Eng&Tech, Madrid, Spain). Typically, an amount of ca. 200\u00a0mg of the catalysts was packed and sandwiched with quartz wool in a tubular reactor (I.D.\u00a0=\u00a09\u00a0mm, L\u00a0=\u00a0305\u00a0mm) made of Hastelloy. During the packing step, the reactor was tapped frequently to ensure reproducible packing state of the catalyst beds. The catalysts were reduced internally using a mixture of H2 (30\u00a0mL\u00a0min\u22121) and N2 (10\u00a0mL\u00a0min\u22121) at 350\u00a0\u00b0C for 4\u00a0h. Afterwards, the reactor pressure was increased to 40\u00a0bar, and the temperature was reduced to 250\u00a0\u00b0C. The reactant mixture comprised of CO2 (24\u00a0vol%) and H2 (72\u00a0vol%) balanced in N2 (4\u00a0vol%). The gas flow rate was 40\u00a0mL\u00a0min\u22121 (GHSV\u00a0=\u00a048000\u00a0h\u22121). The steady state was typically reached after 1\u00a0h marking the start of the catalytic experiments. In each experiment, the product mixture was sampled every 20\u00a0min in 2\u00a0h. The remaining reactants and gas products in the discharged gas stream were analyzed using an online Agilent 7890A chromatograph equipped with Porapak Q, HayeSep Q and molecular sieve 5A columns. The CO2 conversion (\n\n\nX\n\n\nCO\n\n2\n\n\n\n) and selectivity to methanol (\n\n\nS\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n) and CO (\n\n\nS\n\nCO\n\n\n\n) were calculated using the following equations:\n\n(1)\n\n\n\nX\n\nC\n\nO\n2\n\n\n\n=\n\n\n\nn\n\nC\n\nO\n2\n\n,\ni\nn\n\n\n-\n\nn\n\nC\n\nO\n2\n\n,\no\nu\nt\n\n\n\n\nn\n\nC\n\nO\n2\n\n,\ni\nn\n\n\n\n\u2219\n100\n%\n\n\n\n\n\n\n(2)\n\n\n\nS\n\n\n\nCH\n\n3\n\nO\nH\n\n\n=\n\n\nn\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n\nn\n\nC\n\nO\n2\n\n,\ni\nn\n\n\n-\n\nn\n\nC\n\nO\n2\n\n,\no\nu\nt\n\n\n\n\n\u2219\n100\n%\n\n\n\n\n\n\n(3)\n\n\n\nS\n\nCO\n\n\n=\n\n\nn\n\nCO\n\n\n\n\nn\n\nC\n\nO\n2\n\n,\ni\nn\n\n\n-\n\nn\n\nC\n\nO\n2\n\n,\no\nu\nt\n\n\n\n\n\u2219\n100\n%\n\n\n\nwhere \n\n\nn\n\nC\n\nO\n2\n\n,\ni\nn\n\n\n\n and \n\n\nn\n\nC\n\nO\n2\n\n,\no\nu\nt\n\n\n\n refer to the input and output molar amount of CO2. \n\n\nn\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n and \n\n\nn\n\nCO\n\n\n\n are the molar numbers of methanol and CO, respectively, in the product mixtures. Based on the 6 collected data points, the average and standard deviation of \n\n\nX\n\n\nCO\n\n2\n\n\n\n, \n\n\nS\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n and \n\n\nS\n\nCO\n\n\n\n were determined. Besides, the methanol formation rate (\n\n\nr\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n) was also used as a measure to evaluate the catalytic activity of the studied catalysts in the hydrogenation of CO2 into methanol. The methanol formation rate (\n\n\nr\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n) was calculated as follows:\n\n(4)\n\n\n\nr\n\n\n\nCH\n\n3\n\nO\nH\n\n\n=\n\n\n\n\n\nQ\n\n\nCO\n\n2\n\n\n\u2219\nX\n\n\nC\n\nO\n2\n\n\n\n\u2219\n\nS\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n\n\u2219\nM\nW\n\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n\nm\n\nAu\n\n\n\n\n[\n\ng\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\u2219\n\n\n\n\n\ng\n\nAu\n\n\n\u2219\nh\n\n\n\n\n-\n1\n\n\n]\n\n\n\n\nwhere \n\n\nQ\n\n\nCO\n\n2\n\n\n\n, \n\n\n\nMW\n\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n and \n\n\nm\n\nAu\n\n\n\n denote the total flow rate of the reaction mixture, the molecular weight of methanol, and Au mass of the catalysts, respectively.The elemental composition of synthesized solids, namely, the content of Au, Ce, and Zr, was determined by the ICP-OES technique, and the results are displayed in Table 1\n. All the catalysts exhibit a comparable Au content of 0.7\u00a0wt.% lower than the nominal Au content (1 %), which is probably associated with the hydration of hygroscopic Au precursor. Additionally, the Ce content gradually increased up to 4.6\u00a0wt.% rendering a gain in the Ce to Zr molar ratio (n\nCe/n\nZr) from 0 to 0.09, which agrees well with the nominal n\nCe/n\nZr. The chlorine content was determined in Au-containing catalyst samples, after digestion, by using in-house ion chromatography method. In the analyzed samples, the content of chlorine was below the level of quantification, i.e. <0.5\u00a0mg/g.The structural properties of all the catalysts were characterized by XRD analysis, from which the results are shown exemplarily for Au/ZrCex samples in Fig. 1\n. Independently of Ce and Au content, all the samples exhibit almost identical XRD patterns featuring 2 broad signals centered at 2\u03b8 of 31\u00b0 and 55\u00b0, which can be ascribed to a-ZrO2. No other phases, e.g., cerium oxides and Au-containing phases, were detected. This might be explained by the low content of Ce and Au, or they are present in very fine particles below the detection limit of XRD. The latter was later excluded by TEM analysis. Specifically, the deposited Au is evidenced by TEM and selected area electron diffraction (SAED), which are exemplarily shown for Au/ZrCe0.05 in Fig. 2\n. The SAED pattern of the Au particle shows sharp and continuous rings corresponding to individual crystal planes. On the other hand, the ZrO2-based support material exhibits a halo ring typical for amorphous materials and blurred rings indicating a short-range order (Fig. 2 and Fig. S1). Additionally, separate images of ab-initio simulations of SAED included in Fig. S2, present the main differences between 2-phases, i.e., ZrO2-CeO2, and solid solution ZrO2/CeO2 (Fig. S2). The experimental SAED pattern of Au/ZrCe0.05 matches well with the corresponding simulated pattern, and thus confirms the formation of ZrO2/CeO2 solid solution. The presence of residual Ce species existing in the form of amorphous CeO2 is, however, not excluded.Furthermore, the morphology of Au particles visualized by TEM shows irregular shapes for all the Au particles (Fig. 3\n). The distribution of Au nanoparticle size is slightly uniform for Au/ZrCe0 catalyst and centered at ca. 35\u201340\u00a0nm, which accounts for 15\u201317 % of Au particles. In the meantime, a rather broad distribution was recorded for the particle size of Au particles in all the Ce-containing catalysts, which ranges from 10 to 150\u00a0nm. Noticeably, at the highest Ce content, Au/ZrCe0.1 catalyst exhibits the highest fraction of Au particles of ca. 70\u00a0nm, i.e., 10 %, as compared to 0\u20135 % for other samples (Fig. 3). This indicates that the introduction of Ce using the coprecipitation method did not improve Au dispersion in the amorphous ZrO2 support materials.Regarding the textural properties, minor changes were recorded by the N2 adsorption analyses as displayed in Fig. 4\n. A Ib-type isotherm with a gradual N2 uptake over the low relative pressure range (p/p0\u00a0<\u00a00.4) was recorded for all the samples suggesting a complex pore structure consisting of wider micropores and narrow mesopores [51]. Further analysis using non-local density functional theory model revealed the pore widths ranging from below 2\u00a0nm to 8\u00a0nm. Additionally, all the samples exhibit an H2b type hysteresis loop with a gradual delay on desorption branch indicative of a broader neck width distribution compared with the pore width distribution. The values of specific surface area and total pore volume of Au/ZrCex samples are listed in Table 1. The Ce-free catalyst Au/CeZr0, despite the slightly higher N2 uptake at the relative pressure of 0.6, exhibits a similar porous structure in comparison with the Ce-containing catalysts, ca. 200\u00a0\u00b1\u00a010\u00a0m2 g\u22121 specific surface area (A\nBET) and ca. 0.13\u00a0\u00b1\u00a00.01\u00a0cm3 g\u22121 total pore volume (V\nP).Concerning the redox properties, the results from H2-TPR profiles (Fig. 5\n) provide the very first influence of the introduction of Ce in the Au-based catalyst series. While no signal is visible in the TPR profile of Ce-free catalyst (Au/ZrCe0), the H2 consumption peak centered in the temperature range of 166\u2013176\u00a0\u00b0C is recorded for all Ce-containing catalysts. Additional H2-TPR profiles were recorded for all supports (Fig. S3), in which the peak of interest is absent. The peak is, therefore, associated with the reduction of Au\u03b4+ to Au0\n[26,52], which is facilitated in the presence of Ce. This might be explained by the electron redistribution often known for metals supported on CeO2 due to the formation of surface oxygen vacancies, generated when Ce4+ cations are reduced to Ce3+ rendering to the oxygen transfer process between the metals and CeO2 support [53]. The interaction between Au and Ce-containing supports gradually increased as suggested by the linear correlation found between the Ce content and the H2 consumption shown in Table 2\n. The absence of any reduction peak in Au/ZrCe0 sample is an indication of no positively charged Au in the sample [52,54], and the majority of the Au species is available in the metallic form, which is stable in the oxidation step conducted internally prior to the TPR measurements (see section 2.3). These observations suggest that the Au0 formation is more favored in pure ZrO2 sample, i.e., Au/CeZr0 catalyst, as reported in [54].Moreover, the synthesized solids were evaluated by XPS to obtain additional information related to the chemical environment of the catalysts. The survey spectra for the surfaces of the samples include O-, Zr-, Au-, and C-containing species indicated by the corresponding signals in Fig. 6\n. The signal for Ce, i.e., Ce 3d, was detected and thus confirmed the presence of Ce in all the modified samples except for Au/ZrCe0 catalyst.To investigate the surface chemistry of the synthesized solids in detail, high-resolution XPS spectra were measured and are shown in Fig. 7\n. The relative quantification of surface Au, Ce and Zr species obtained from XPS is displayed in Table S1.The surfaces of all samples consisted of C-containing species, i.e. C-C/C-H, C-O, and COO\u2013/COOH located at the dashed lines designated in the C 1s spectra (Fig. 7a). These species originate from the adventitious carbonaceous species adsorbed on the surface after sample preparation and transport to the spectrometer. The presence of oxidized carbonaceous species is also confirmed in O 1s spectra, i.e., the spectral feature seen as a shoulder located at dashed line 2 in the O 1s spectra (Fig. 7b). The main peak at dashed line 1 in the O 1s spectra represents metal oxides. The doublet in the Zr 3d (the Zr 3d5/2 main peak at approximately 182\u00a0eV) and Au 4f (the Au 4f7/2 main peak located at 84.0\u00a0eV) spectra correspond to ZrO2 and Au, respectively. The Zr 3d and Au 4f spectra for all samples tested do not show significant changes in the binding energy (E\nB) and the shape of the spectra. This indicates that the Zr and Au environment is similar in all the samples. The observation of similar Au species for all the samples is, however, different from TPR results. The presence of H2 consumption for all Ce-containing samples, but not the bare ZrO2, is likely associated with changes in electronic properties induced by the addition of Ce in the ZrO2 structure.The spectra of Ce 3d are more complex because they contain many features. It is known that Ce(IV) contains the spectral feature at about 917\u00a0eV, which is absent in Ce(III) [55]. The spectra illustrated in Fig. 7e show the feature at approximately 917\u00a0eV marked with dashed line in Fig. 7e for all samples except for Au/ZrCe0 catalyst. Therefore, it is confirmed that Ce(IV) was present on the surface in all Ce-containing samples. On this basis, the spectra were fitted for the presence of CeO2 using the positions of the peaks from the previously reported reference spectra [55]. The Ce 3d spectra for CeO2 consist of two pairs of multiplets (2 pairs of 3 peaks) corresponding to spin-orbit splitting. However, the measured spectra do not fit well with this model probably due to partial reduction of Ce(IV) forming Ce(III) under vacuum conditions (ultra-high vacuum in XPS spectrometer) which is known for CeO2\n[56]. Thus, the spectra of Ce-containing samples were fitted for both Ce(IV) and Ce(III) species (with two doublets [55]). The fitted results match well with all measured spectra suggesting the presence of these two oxidation states on the surface. The fitted Ce 3d spectrum for the sample Au/CeZr0.1 is exemplarily shown in Fig. 8\n.The basic properties of all Au-containing catalysts were evaluated using CO2-TPD analysis and the obtained results are shown in Fig. 9\n. The CO2 desorption of Au/ZrCe0 catalyst features a signal centered at ca. 100\u00a0\u00b0C, and end at ca. 280\u00a0\u00b0C, which is contributed mainly to the presence of weak basic sites. This peak was also recorded for the Ce-containing catalysts, but it is visible with a considerable broadening and coupled with an additional peak at higher temperatures above 300\u00a0\u00b0C, which is likely associated with the presence of Ce providing higher basicity than pristine ZrO2. However, no clear trend is observed for basic site density (BSD) when the catalysts gradually increases Ce content (Table 2). This suggests that in addition to Ce introduction, the changes in the number of basic sites on the surface of ZrO2 support are likely affected by other factors, e.g., thermal treatment causing dehydroxylation and formation of acid-base Zr4+/Ce4+-O2\u2013 pairs at various extents. Similarly, the influence of Ce introduction on the acid site density (ASD) occurs in a random fashion despite the gradual rise in Ce content as shown in Fig. S4. Interestingly, it was found that the acidic and basic properties of the Au-based catalyst series are closely related, as deducted from a linear regression observed between the acid and basic site density (Table 2). This might be related to the fact that a-ZrO2 is known to possess mainly coordinatively unsaturated Lewis acid-basic Zr4+/O2\u2013 pairs as well as surface hydroxyl groups [57].The Au-based catalysts supported on ZrO2-CeO2 mixed oxides were tested in the methanol synthesis from CO2 under various temperatures from 250 to 320\u00a0\u00b0C as displayed in Table 3\n. In addition to CO2 conversion (\n\n\nX\n\n\nCO\n\n2\n\n\n\n), methanol selectivity (\n\n\nS\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n) and methanol formation rate (\n\n\nr\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n) are employed as measures to evaluate the activity for the methanol synthesis over the Au-based catalysts. Over the studied temperature range, CO2 conversion varies between 1 % and 11 %, with methanol selectivity up to 27 %, amounting to a methanol formation rate of 1.5 to 4.6 \n\n\ng\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\u2219\n\n\n\n\n\ng\n\nAu\n\n\n\u2219\nh\n\n\n\n\n-\n1\n\n\n\n. Apart from methanol, CO and trace amounts of coke (except for Au/ZrCe0.1 catalyst), no other carbon-containing gas product was found in the discharged gas mixture, and the carbon mass balance is closed up to 85 % to 98 %. The catalytic activity obtained in this study is rather low, in particular, compared with the commercial methanol synthesis. Interestingly, the methanol selectivity of Ce-free Au/ZrCe0 catalyst is significantly lower, i.e., 20 % vs. 72.5 %, in relation to the Au/ZrO2 published in a previous work form our group despite the analogous Au content (0.7\u00a0wt.% and 0.5\u00a0wt.%, respectively) as well as reaction conditions (240\u2013250\u00a0\u00b0C, 40\u00a0bar). This might be explained by the drastic discrepancy in the average Au particle size often reported to play a key role in the hydrogenation activity, i.e., \u224840\u00a0nm vs. 1.1\u00a0nm, respectively. The large size of Au particles might lead to low interfacial Au-support contact and thus facilitates the formation of undesired competive product CO generated by reverse water gas shift reaction. Moreover, the difference in other properties of the catalysts, e.g., phase composition of ZrO2-based support, acidic/basic properties, textural and redox properties, may also play a role.The reaction temperature, as expected, exhibits a positive effect in the CO2 conversion independent of Ce content but at the cost of the selectivity to methanol. This agrees well with literature reports [35,58,59] and is attributed to the higher sensitivity to temperature of the CO formation rate in relation to the methanol synthesis, which is evidenced by higher apparent activation energy calculated from Arrhenius plots, e.g., \n\n\nE\n\nA\n,\nC\nO\n\n\n\n=64\u00a0kJ\u00a0mol\u22121 > \n\n\nE\n\n\n\nA\n,\nC\nH\n\n3\n\nO\nH\n\n\n\n=27\u00a0kJ\u00a0mol\u22121, as observed in the case of Au/ZrCe0.025 catalyst. The values of EA for other catalysts are provided in Fig. S5. Besides, all studied catalysts exhibit similar apparent activation energy for methanol formation, suggesting that the introduction of Ce did not alter the hydrogenation mechanism of CO2 into methanol [49]. Consequently, the methanol synthesis via formate intermediate, as proposed in a previous study on Au/a-ZrO2 catalysts [27], was assumed for the Au-based catalysts in this study. Accordingly, the reaction starts with adsorption and activation of CO2 and H2 onto ZrO2-based support and metallic Au, respectively, which later migrate to the Au-support interface. The dissociated H species react with the activated CO2 forming mono- and bidentate formates, the two primary intermediates of methanol. Subsequently, the formate species undergo hydrogenation to H2COO*, and further terminal protonated to H2COOH*, which then cleaves at the C-O bond releasing OH* and H2CO*. The latter participates in hydrogenation, forming H3CO* and finally methanol. Among these steps, CO2 is simultaneously involved in the RWGS process forming undesired product CO.Furthermore, while the Ce-free catalyst Au/ZrCe0 exhibits the highest CO2 conversion (2.3 % at 250\u00a0\u00b0C) coupled with a 20 % selectivity to methanol, a slightly lower CO2 conversion was recorded for all the Ce-containing catalysts (Table 3). The decrease in CO2 conversion is more profound with increasing Ce content. Considering the comparable Au content as well as the chemical environment of Au, one of the other possible reasons for the decline in CO2 conversion most likely lies in the large size of Au particles generally accepted to be crucial for its catalytic activity, which was found larger for Ce-containing samples. Besides, catalyst deactivation caused by coke deposition in association with the presence of Ce, as reported by Pojanavaraphan et al. [52], might explain the lower activity in Ce-containing catalysts. However, it is noted that in comparison to the current study, the Ce content was much higher (\u226525\u00a0wt.%) than that presented in this study, and consequently, the negative influence of Ce via boosting carbonaceous species is less likely. In fact, the drop in CO2 conversion is only considerable for Au/ZrCe0.1 catalyst exhibiting the highest Ce content (4.6\u00a0wt.%) and the highest extent of coke deposition (Table 3). On the other hand, the Ce-containing materials, as reported in the literature [25 30,32], exhibit a higher and stronger affinity towards CO2 molecules, which might lead to active sites poisoning. Additionally, it is worth mentioning that the addition of Ce makes H2O more favorably adsorbed on the surface [60] and contributes to a blockage of active sites and thus a decline in CO2 conversion.Unlike CO2 conversion, no clear trend was found for methanol selectivity in relation to the presence of Ce. Thus, the next part focuses on the influence of the acidic/basic properties on the methanol selectivity.In order to examine relation between methanol selectivity and the catalyst properties, the methanol selectivity \n\n\nS\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n was plotted against the density of base sites (BSD) and acid sites (ASD) as shown in Fig. 10\n.Interestingly, volcano-like shaped dependencies were recorded for both BSD and ASD in relation to the methanol selectivity, which exhibited a critical point of 120\u00a0\u03bcmol\u00a0g\u22121 (BSD) and 600\u00a0\u03bcmol\u00a0g\u22121 (ASD) at which the highest methanol selectivity of 27 % is reached. The positive effect of the acidity of catalysts in the methanol synthesis was reported in previous studies using Cu/ZrO2\n[55,56]. However, the correlation of the methanol selectivity with gradual changes in acidic/basic properties was not studied. To gain insight into this matter, the nature of acidic and basic sites should be considered. The majority of the acidity and basicity of the investigated catalysts is contributed to the Lewis acid-base pairs, i.e., coordinatively unsaturated Zr4+-O2\u2013, which can reinforce the adsorption of CO2, particularly, the interaction between O and C atoms of CO2 molecules with Zr4+ and O2\u2013, respectively [49]. Moreover, in the presence of Ce, the adsorption of CO2 forming monodentate carbonates, which was not observed for pristine ZrO2, is facilitated as suggested by density functional theory (DFT) results obtained for t-ZrO2 incorporated with ca. 2\u00a0wt.% Ce [60]. Monodentate carbonate is an important precursor to the more stable bidentate carbonate, which can be further converted into bidentate formate and subsequently to methanol via the formate reaction pathway. Besides, these Lewis acid-base pairs can also participate in the dissociation of water molecules generating surface hydroxyl groups, which can promote methanol synthesis via reacting with carbonate species, forming mono-/bidentate formates [61]. The eased formation of formates might hinder the competitive process, RGWS, as suggested by various studies [34,62], and thus increase the methanol selectivity. Alternatively, the methanol synthesis can also be facilitated via the hydrogenation of CO [57], which is, however, unlikely in this study as the CO adsorption was suggested to be hardly affected by Ce incorporation at low content [60]. Thus, the increase of the acid-base pairs in numbers, as well as strength, probably renders to the improved methanol selectivity (Fig. 10). Nevertheless, the interface between the multiple components of the catalysts, i.e., metallic sites, hydroxyl groups and the acid-base pairs, seemingly plays a crucial role in the methanol synthesis as depicted in Fig. 11\n, which might explain the decrease in methanol selectivity when further increasing the number of acid/basic sites. The surface catalytic sites probably suffer catalyst deactivation due to blockage of active sites by a multilayer of adsorbed CO2 or intermediates.Considering the highest methanol formation rate among the Ce-containing Au-based catalysts, Au/ZrCe0.025 sample was selectively investigated in the catalytic stability carried out throughout 93\u00a0h under reaction conditions (280\u00a0\u00b0C and 40\u00a0bar) (Fig. 12\n). The obtained results suggest a stable performance through the whole experiment with a CO2 conversion of 3.5 % \u00b1 0.2 % rendering to a methanol formation rate of 3.32\u00a0\u00b1\u00a00.25 \n\n\ng\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\u2219\n\n\n\n\n\ng\n\nAu\n\n\n\u2219\nh\n\n\n\n\n-\n1\n\n\n\n. Noticeably, with respect to the start of the experiment, there is a slight loss of catalytic activity by 0.6 % in CO2 conversion and 0.47 \n\n\ng\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\u2219\n\n\n\n\n\ng\n\nAu\n\n\n\u2219\nh\n\n\n\n\n-\n1\n\n\n\n in (\n\n\nr\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\n) occurring in the first 3\u00a0h, which might be due to the considerable formation of water via RWGS causing saturation of the support surface due to high affinity to water in the presence of Ce [60]. Nevertheless, the reactivity of the catalyst (Au/ZrCe0.025) was steady for the next 90\u00a0h. Despite the long reaction time, the coke deposition is negligible as deducted from the marginal discrepancy in the consumed catalyst's C content, i.e., 0.7\u00a0wt.% (spent) vs. 0.6\u00a0wt.% (fresh).In summary, this study presents a systematic investigation of the influence of Ce in the Au catalysts supported on amorphous ZrO2 prepared via the simple coprecipitation method. The newly introduced Ce was present in the form of both Ce(III) and Ce(IV), which does not alter the textural properties compared to the pristine ZrO2 materials with a Ce content up to 4.6\u00a0wt.%. The Ce-containing catalysts show a slightly lower Au dispersion evidenced by large particles (ca. 40\u00a0nm) and a broad distribution ranging from 5 to 125\u00a0nm. Besides, the formation of metallic Au is less favored and coupled with increased Au cations with rising Ce content, probably due to the electron exchange between Au and CeO2. These findings might explain the decrease of CO2 conversion with increasing Ce content of Au catalysts supported on ZrO2-CeO2 mixed oxides in the hydrogenation of CO2. In addition to methanol formed via formate intermediates, only CO was observed as a side product under the studied reaction conditions (typically at 40\u00a0bar and 250\u2013320\u00a0\u00b0C), irrespectively of the catalysts employed. It was proven that acid-base properties play a role in CO2 adsorption/activation and thus in tailoring the product distribution of the CO2 hydrogenation. The selectivity of the desired product methanol was found to be closely associated with the number of both acidic and basic sites of the catalysts in a volcano-shape fashion. In particular, there is a critical point of acid/base site density, at which the methanol selectivity reaches a maximum 27 % (at 40\u00a0bar and 280\u00a0\u00b0C), which is 120 and 600\u00a0\u03bcmol\u00a0g\u22121, respectively. The Au catalysts on CeO2-ZrO2 mixed oxides exhibit excellent catalytic stability, for example, the methanol formation rate of 3.32 \n\n\ng\n\n\n\nCH\n\n3\n\nO\nH\n\n\n\u2219\n\n\n\n\n\ng\n\nAu\n\n\n\u2219\nh\n\n\n\n\n-\n1\n\n\n\n recorded for Au/ZrCe0.025 catalyst at 250\u00a0\u00b0C and 40\u00a0bar remained for up to 93\u00a0h TOS.\nHue-Tong Vu: Investigation, Methodology, Visualization, Writing \u2013 original draft, Writing \u2013 review & editing. Matja\u017e Fin\u0161gar: Funding acquisition, Investigation, Writing \u2013 review & editing. Janez Zava\u0161nik: Investigation, Writing \u2013 review & editing. Nata\u0161a Novak Tu\u0161ar: Funding acquisition, Supervision, Writing \u2013 review & editing. Albin Pintar: Funding acquisition, Project administration, Conceptualization, Supervision, Visualization, Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was funded by the Slovenian Research Agency, grant numbers J7-3151, P1-0418 and P2-0118. The project is co-financed by the Republic of Slovenia, the Ministry of Education, Science and Sport and the European Union under the European Regional Development Fund. The authors thank Mojca Opresnik, Edi Kranjc, \u0160pela Bo\u017ei\u010d, Iris \u0160tucin, Matev\u017e Ro\u0161kari\u010d, and Gregor \u017derjav from the National Institute of Chemistry (Ljubljana, Slovenia) for N2 sorption, XRD, CHN, ICP-OES analyses and their support in the labs, respectively. Janez Zava\u0161nik acknowledges the support via ARRS P1-0417.Supplementary data to this article can be found online at https://doi.org/10.1016/j.apsusc.2023.156737.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  To exploit the potential of both ZrO2 and CeO2, the mixed oxides with Ce content up to 5\u00a0wt.% were prepared via the simple coprecipitation method, subsequently loaded with Au catalysts, and investigated in the hydrogenation of CO2. The obtained catalysts, namely Au/ZrCex (x\u00a0=\u00a0n\n                     Ce/n\n                     Zr\u00a0=\u00a00.0\u20130.1), exhibit similar Au content (0.7\u00a0wt.%), structural and textural properties, however, considerably different acidic/basic properties and great surface oxygen vacancy. The introduction of Ce leads to slightly decreased CO2 conversion, which was found proportional to the Ce content. Interestingly, it is evidenced that the methanol selectivity is closely related to the acidic/basic properties of the catalysts employed. Over the broad range of acid site density (ASD) ranging from 400 to 700\u00a0\u03bcmol\u00a0g\u22121, the highest methanol selectivity of 27 % was recorded at 250\u00a0\u00b0C and 40\u00a0bar over the catalyst exhibiting an ASD of 600\u00a0\u03bcmol\u00a0g\u22121, which decreased with further increasing ASD. The volcano-shape trend was also discovered for the base site density showing the critical point at 120\u00a0\u03bcmol\u00a0g\u22121 over the range from 90 to 210\u00a0\u03bcmol\u00a0g\u22121. These findings suggest that the acidic/basic properties can be tuned, e.g., via thermal treatment, to tailor the product distribution of the CO2 hydrogenation. Above all, the Au catalysts supported on ZrO2\u2013CeO2 mixed oxides exhibit excellent catalytic stability, e.g., a methanol formation rate of 3.32 \n                        \n                           \n                              g\n                              \n                                 \n                                    \n                                       C\n                                       H\n                                    \n                                    3\n                                 \n                                 O\n                                 H\n                              \n                           \n                           \u2219\n                           \n                              \n                                 \n                                    \n                                       \n                                          g\n                                          \n                                             A\n                                             u\n                                          \n                                       \n                                       \u2219\n                                       h\n                                    \n                                 \n                              \n                              \n                                 -\n                                 1\n                              \n                           \n                        \n                      is remained over the reaction course of 93\u00a0h at 250\u00a0\u00b0C and 40\u00a0bar for Au/ZrCe0.025 sample.\n               "}
{"full_text": "Due to the industrial growth in emerging economies, extensive utilization of fossil fuels has been the major source of environmental pollution and energy crisis. Thus, transforming solar energy into chemicals through photosynthesis has been considered the favorable approach to tackle the situation [1,2]. Particularly, PEC water splitting effectively generates fuels without any harmful carbon dioxide emissions [3,4]. Material selection for PEC water splitting is critical, provided that photoelectrodes absorb incident photons and generate the charge carriers and need to be economical for practical applications [5,6,7\u201310]. Numerous n-type semiconductor metal oxides photocatalysts have been widely studied (TiO2, ZnO, Fe2O3, BiVO4, Bi2O3 etc.) [11-15]. Titanium dioxide (TiO2) is frequently utilized due to its chemical steadiness and photocatalytic capabilities [16-18]. However, due to the short-wavelength cutoff qualities of TiO2, only 6% of the solar energy reaching the earth's atmosphere can be exploited to drive photovoltaic and photocatalytic effects. More importantly, TiO2 has the latent capacity for full-water splitting reaction owed to its fortunate band-edge position [19,20]. In this context, two complicated multi-electron half-reactions are involving in the PEC water-splitting process (2H2O\u00a0\u2192\u00a0O2\u00a0+\u00a04H+\u00a0+\u00a04e\u2212, E\u00a0=\u00a01.23 VRHE; 4H+\u00a0+\u00a04e\u2212\u00a0\u2192\u00a02H2, E\u00a0=\u00a00 VRHE). According to earlier reports, the reactions among photoinduced holes and water molecules primarily restrict the features of PEC reactions, as it usually happens at a substantially higher potential to remove 4 electrons and 4 protons from 2 water molecules to produce an O2 molecule [21-23]. However, the photoinduced valence band (VB) holes in TiO2 are considered to be kinetically inefficient, and thereby, it requires an additional anodic bias before water can be oxidized [24]. In this context, the combination of water oxidation catalysts with a photon-absorbing substrate represents an appropriate route to reduce the driving force of the electrolysis chemistry of the catalyst and thereby enrich efficiency. Significant effort has been put into emerging an appropriate oxygen evolution (OE) rection electrocatalyst for photoelectrode [25-31].In recent years, transition metal-based catalysts, such as oxides and hydroxides of Ni, Co, Fe, and Mo, as well as oxyhydroxides and phosphates, have been used as OER co-catalysts [32-35]. Though, all such co-catalysts hurt from low electrical conductivity [36]. Because of its abundant earth reserve properties, low cost, and ability to function under benign conditions, transition-metal phosphides (TMPs), namely FeP, CoP, MoP, and NiP is a superior catalyst for OE associated with precious metals (Ir, Ru) and metal oxides (RuO2, IrOx) catalysts [37-41]. Among the existing metal phosphates, Ni-based TMPs materials have been broadly examined for various kinds of electrochemical uses [42,43]. Notably, Ni-based TMPs show a substantial part in the electrochemical features, whereas P impacts a stable structure [44]. However, fabricating a trustworthy approach for incorporating TMPs with oxide-based electrodes in a convenient, scalable, and economical approach is still required. Liu et al. and Bu et al. integrated Fe-integrated Co2P and Ni2P nanoparticles as operative co-catalytic materials on Fe2O3-based electrodes to boost the water splitting system[45,46]. Thus, extensive research effort on the exact mechanisms of TMP-based electrocatalytic material for water oxidation reaction is still required. Ruifeng et al. reported that the NiPi modified of Pi-Fe2O3 photoanode showed enhanced photoelectrochemical activity for glycerol oxidation. With the addition of NiPi, the PEC features of Pi-Fe2O3 were increased by approximately twofold at 1.5 VRHE\n[47]. Schipper et al. have recently demonstrated the first successful use of FeMnP with rutile TiO2 as a co-catalyst for PEC solar water-splitting reactions [48].We demonstrated the PEC water splitting using NiPi nanoparticle-modified TNT array semiconductor photoanodes using simple two-step anodization and electrodeposition process.Fabricated TNTs comprised vertically stacked nanotubes arrays with a high surface area that allows high integration with NiPi nanoparticles. The TNTs/NiPi photoelectrodes displayed superior photocurrent density (0.759\u00a0mA/cm2) at 1.23 VRHE, representing nearly 3-fold enhancements than TNTs. Furthermore, the TNTs/NiPi demonstrated well-separated electron-hole pairs and improved the charge transfer process at the interface between the electrode and electrolyte, indicating the NiPi certainly accelerates the water oxidation rate and reduces the needed overpotential. These findings emphasized the multifunctional role of decoration of TNTs arrays with NiPi in enhancing PEC solar water splitting.TNT arrays were obtained via a two-step electrochemical anodization process involving Ti foil. Before reaction, a 0.25\u00a0mm Ti foil (>99.5%, Alfa Aesar) was ultrasonically washed with acetone and deionized (DI) water. Subsequently, Ti foil was assembled in a 2-electrode assembly with a Pt foil as the counter electrode in 0.12\u00a0M NH4F in a 5/100 (w/w) mixed solution of DI water and ethylene glycol (EG), at a continuous potential of 60\u00a0V. All through the anodization process, the foil was exposed to the solution, the first step was for 10\u00a0min, then ultrasonicated for 5\u00a0min, the second step was anodization for 20\u00a0min, and then cleaning by sonicated in ethanol for 10 sec then washed very well with DI water and then subjected to an annealing process at 450\u00a0\u00b0C for 2\u00a0h.TNTs/NiPi electrodes were fabricated by electrochemical deposition through an electrochemical bath comprised of 20\u00a0mM NiCl2 in DMSO. The deposition was executed in a 3-electrode cell comprising TNTs, Pt, and Ag/AgCl as working, counter, and reference electrodes. Afterward, the electrodeposition was carried out at \u22122.0\u00a0V vs. Ag/AgCl, and an optimization process was tuned by tuning the deposition charges varying from 1 to 10 mC/cm2. Then, the obtained film was exposed to an annealing at 450\u00a0\u00b0C for 1\u00a0h in the air (2.0\u00a0\u00b0C/min). Thus the optimal charge density was estimated to be 5 mC/cm2. Fig. S1 shows the different phases of the synthesis approach employed to obtain the TNTs/NiPi photoanodes.X-ray diffraction (XRD) investigations were done via X-ray diffractometer (Rigaku Miniflex 600). The UV\u2013visible diffuse reflectance spectra (DRS) were acquired via Shimadzu UV-2600. Also, the surface morphologies of electrodes were performed through a field-emission scanning electron microscope (FE-SEM, JSM7600F, JEOL, USA). X-Ray Photoelectron (XPS) analyses were executed on the photoanodes with JEOL XPS-9030.All the PEC analyses were evaluated through an AutoLab potentiostat PGSTAT30. All the experiments were measured in a 0.1\u00a0M PBS (pH 7.5) solution with and without 1.0\u00a0M Na2SO3 as a hole scavenger. The applied bias photon-to-current efficiency (ABPE) of the TNTs, and TNTs/NiPi photoelectrodes under illumination were evaluated from the linear sweep voltammogram curves using the Eq. (1).\n\n(1)\n\n\nA\nB\nP\nE\n\n\n\n\n\n%\n\n\n\n\n=\n\n\n\n\n\n\n\nI\n\u00d7\n\n\n\n1.23\nv\n-\n\nV\nb\n\n\n\n\n\n\n\n\nP\n\ntot\n\n\n\n\n\n\n\u00d7\n100\n%\n\n\n\nwhere I\u00a0=\u00a0current density at certain potential Vb (mA/cm2) under illumination conditions (100\u2009mW\u2009cm\u22122), and Ptot\u00a0=\u00a0power density of the incident light.In order to explore the source of PEC features of TNTs electrodes after optimal Ni incorporation, complete optical, crystalline, and morphological characteristics were carried out. Initially, the optimal Ni content over the bare TNTs was evaluated after evaluating different Ni incorporation by tuning the applied charge throughout the deposition method. Notably, the obtained results are presented in Fig. S2, assessing that the optimal charge agrees to 5 mC/cm2. Further, the crystalline nature of the TNTs, TNTs/Ni, and TNTs/NiPi composite nanotube arrays was dogged by XRD analysis (Fig. S2a, and Fig. 1\n). Also, the obtained diffraction peaks for all samples were at 25.3\u00b0, 46.7\u00b0, and 54.8\u00b0, respectively. The diffraction peaks belonging to the (101), (200), and (211) plans of the anatase phase have been identified [49]. This specifies that the TNTs crystal phase changed from amorphous to anatase only after annealing, but no rutile phase was observed. Moreover, an XRD analysis of TNTs decorated with different Ni loadings was performed (Fig. S2a,b). Remarkably, increasing the Ni incorporation over TNTs systematically increased the lattice constant, which is consistent with the bigger Ni2+ ion. It can be evidenced that the diffraction pattern of optimized TNTs/Ni, and TNTs/NiPi was identical to that of TNTs, then the intensity of the peak was comparatively decreased (Fig. 1a), due to Ni, and NiPi presence on TNTs. It must be noted that, due to the smaller quantity of Ni, the observed peak shifts in the XRD are not straightforward.Using Raman spectroscopy, we further examine the structure of unmodified TNTs, TNTs/Ni, and TNTs/NiPi samples. The spectra of bare and Ni/NiPi loaded TNTs photoanodes are shown in Figs. 1b, and S2c. The six different modes of anatase TNTs, such as 1A1g, 2B1g, and 3Eg, can be represented by the following Raman active modes: (1A1g\u00a0+\u00a02B1g\u00a0+\u00a03Eg) [50]. There are three waves with the numbers 143, 196, and 637\u00a0cm\u22121 associated with Eg mode in the anatase phase.The B1g mode is present at 396 and 516\u00a0cm\u22121, while the A1g mode can be found at 516\u00a0cm\u22121.A1g and B1g modes have a wavelength of 516\u00a0cm\u22121, hence they are unresolved doublet modes. In the resulting Raman pattern, a shift toward higher wavenumbers appears as the Ni concentration increases Fig. S2c, and NiPi is loaded (Fig. 1b). Similar trends are also seen in the XRD pattern (Fig. S2b, and Fig. 1a), where peak intensities above 10 mC shift towards high 2\u03b8. The peak intensity increases as Ni and/or NiPi are deposited on TNTs. These findings were in concordance with our XRD measurements.In order to examine the impact of both additives (Ni and/or NiPi) on the behavior of the photoelectrodes under PEC, a detailed optical and structural characterization was performed (Fig. 2\na, Fig. S3). Fig. 2a shows the DRS methods applied to measure the bandgap and absorption of photoelectrodes.As can be seen, incorporating Ni and/or NiPi nanoparticles enhanced the visible-light optical density of the TNTs photoanodes by providing electron transitions at the band edges of anatase-scheelite phase TNTs. The band-to-band transition in TNTs causes them to show an intense band at wavelengths shorter than 400\u00a0nm. TNTs/Ni photoanode showed a red-shift compared to the undoped TNTs (Fig. 2b). Impurity levels can cause the photoexcitation energy to decrease and contribute to a substantial red-shift of the absorption edge [51]. Comparatively, in TNTs/Ni photoanode, there was an enhanced light absorption on the TNTs/NiPi photoelectrodes. The enriched light absorption is accredited to the excitation of electrons in the bandgap by localized P 2p. TNTs/NiPi samples exhibit a dramatic radiation absorption between 400 and 760\u00a0nm due to the additional crossovers between the Ti4+, VB, and conduction band (CB).\nFig. 3\n displays top-view and cross\u2010sectional FE-SEM photographs of two-step anodized TNTs, TNTs/Ni, and TNTs/NiPi photoanodes after the anodic oxidation process and annealing at elevated temperatures. As shown in Fig. 3, morphological observations reveal a homogenous distribution of highly ordered, vertically aligned, hollow TNTs in all samples (TNTs, TNTs/Ni, and TNTs/NiPi) with a diameter of around 95\u2013115\u00a0nm, length of around 522\u00a0nm and a wall thickness of 18\u00a0nm. The TNTs still kept uniform nanotube arrays after the Ni (Fig. 3c, d), and NiPi modification process (Fig. 3e and f). No noticeable accumulation of NiO was observed on the surface of the TNTs; only tiny fragments made by cracking of the tube walls were present. Fig. 4\n displays HR-TEM photographs of the fabricated TNT/NiPi electrodes. The TEM images in Fig. 4a and b specify the homogeneity and arrangement of the obtained nanotubes morphology in the TNT/NiPi electrodes. Furthermore, the acquired TNTs were greatly identical, with an outer diameter of 165\u00a0\u00b1\u00a02\u00a0nm and a 28\u00a0\u00b1\u00a02\u00a0nm wall thickness. Also, the characteristic lattice fringes of 0.351\u00a0nm seen in the photographs in Fig. 4c matching to the (101) plane of anatase TiO2 (JCPDS # 21-1272), suggesting the anatase phase of the TNTs. The TNTs/NiPi films were defined by the presence of crystalline particles in the range of 5\u201310\u00a0nm over the surface of TNTs (Fig. 4d). Further, the EDS spectrum of the TNT/NiPi in Fig. S4, where Ti, O, P, and Ni signals are observed, demonstrated a successful addition of NiPi layers over TNTs.The surface composition nature of the TNTs, and TNTs/NiPi composite electrode were explored by XPS, as presented in Fig. 5\n. The survey XPS spectrum of TNTs/NiPi composite electrodes reveals the existence of all the elements (Ti, Ni, P, and O) without obvious contaminations (Fig. 5a). Further, Fig. 5b presents the deconvolution of the Ti 2p spectrum for TNTs, and TNTs/NiPi photoanodes. The Ti 2p spectra show the characteristic doublet of spin\u2013orbit coupling (2p3/2, 2p1/2) for both photoanodes, with the highest intensity peak of the Ti 2p3/2 component at approximately 458.95\u00a0eV binding energy (BE) due to the presence of Ti4+ ions in TiO2 and a lower intensity peak at ca 460.10\u00a0eV due to the existence of Ti3+ ions in Ti2O3\n[52,53]. This validates that both TiO2 and Ti2O3 are created in the TNTs crystals from the anodization route. It can be envisioned that the existence of Ti3+ in photoelectrodes could be a benefit for PEC uses under illumination since the ionic radius of Ti3+ (0.81 *) is near to Ni2+ (0.83\u00a0\u00c5) than Ti4+ (0.75\u00a0\u00c5) [48,54]. Fig. 5c shows the O1s spectra of different photoanodes. The high-resolution O 1s spectra show a more substantial peak at \u223c530.18\u00a0eV and an observed shoulder peak at around 531.19\u00a0eV, implying the existence of two dissimilar O chemical states, with the crystalline lattice oxygen and hydroxyl oxygen with upsurging BEs [55,56]. Interestingly, in Ni 2p spectrum (Fig. 5d), the obtained peaks at 855.96\u00a0eV and 873.3\u00a0eV are allocated to Ni 2p3/2 and 2p1/2 of Ni2+, correspondingly [56-58]. Moreover, other peaks at 849.89\u00a0eV are allotted to Ni 2p3/2 for metallic Ni [59]. As seen in Fig. 5e, the BE of P 2p is at \u223c133.53\u00a0eV, representing P in the phosphate group, endorsing that P subsists as the nature of the phosphate group [31].The PEC features of the unmodified TNTs, TNTs/Ni, and TNTs/NiPi electrodes toward water oxidation were examined under chopped and constant solar light irradiation in a 0.1\u00a0M PBS solution (pH \u223c7.5). Beforehand the observed photocurrent of the TNTs/Ni photoanodes was optimized by tuning the total deposition charge for NiO from 1 to 10 mC/cm2 as shown in Fig. 6\na. Interestingly, the obtained photocurrent density upsurged from 0.533 to 0.621\u00a0mA/cm2 as the deposition charge of the Ni was upsurged from 2 to 10 mC/cm2, owing to boosted light absorption and, subsequently, improved carrier photo-generation. Though, the photocurrent density reduced when the deposition charge of the Ni extended by about 10 mC/cm2 owed to an increased carrier recombination rate (Fig. 6b). Thus, the optimal photocurrent of 0.621\u00a0mA/cm2 was succeeded at a charge density of 5 mC/cm2 and 1.23 VRHE under standard conditions. This can be simplified by the electrocatalytic OE reaction at the interface between the NiPi and electrolyte, and when NiPi integration at a higher deposition charge (>0.5 mC/cm2), the photoinduced holes must be moved amongst several NiPi molecules and NiPi/electrolyte interface, whereas considerably affected the reduced kinetics of the hole transfer, and successively, a lower photocurrent response is observed [60]. In addition to the deposition charge of the Ni, the morphological features and nanoarchitecture of the surface influenced the PEC features of photoelectrodes. Fig. 6c displays the chopped linear sweep voltammograms (LSV) plots of the TNTs, TNTs/Ni, and TNTs/NiPi electrode at 5\u00a0mV/s in 0.1\u00a0M PBS solution (pH \u223c7.5). Upon irradiation conditions, the obtained photocurrent response of the electrodes at 1.23 VRHE declined as TNTs/NiPi (0.76\u00a0mA/cm2)\u00a0>\u00a0TNTs/Ni (0.621\u00a0mA/cm2)\u00a0>\u00a0TNTs (0.243\u00a0mA/cm2). The LSV plots for the PEC water oxidation in Figure 6d confirm that the optimized TNTs/NiPi photoanodes exhibited enhanced photocurrents than bare TNTs. The evaluated photocurrents, which we assign to OE reaction [61,62], are superior upon NiPi addition over TNTs photoanodes. Notably, the photocurrent density improved considerably with the applied bias and obtained \u223c0.764\u00a0mA/cm2\nat 1.23\u00a0VRHE, which agrees to a nearly 3.2-fold development associated with the bare TNTs(Table 1\n). After incorporating NiPi, the substantial reduction in onset potential and upsurge in photocurrent response revealed rapid water oxidation. Also, enhanced PEC features are caused by effective carrier separation and reduced rate of carrier recombinations. Notably, the recombination might occur in the bulk or the surface of the TNTs electrodes. Fig. 6e reveals the plots of the ABPE with respect to the applied bias. The bare TNTs electrodes display an optimal ABPE of 0.253% at \u223c0.35 VRHE. Considerably, the TNTs/NiPi electrodes realize the highest ABPE of 0.46% at a quite lower potential of \u223c0.31 VRHE (Table 1). Furthermore, >1.85 times enhanced ABPE at a smaller external bias directly determines that the addition of NiPi over TNTs is an accessible route to enrich the PEC features of TiO2. As debated above, the continued charge separation and transfer process of NiPi are the key factors for the enhanced PEC performance of TNTs/NiPi photoanode. The fabricated electrodes' comparative electrochemical surface area (ECSA) was assessed by the capacitive part of the cyclic voltammogram (CV). As presented in Fig. S5(a,b), CVs were executed at numerous sweep rates in the region of 10\u2013100\u00a0mV/s in 0.1\u00a0M PBS. The ECSA was then evaluated by assessing the capacitive current related to double-layer charging from the sweep rate requirement of the CV. The double-layer capacitance (Cdl) was assessed from the association between \u0394J = (Ja-Jc) of RHE at 0.66 VRHE and the scan rate. As displayed in Fig. 6f, the linear slope is equivalent to twice the Cdl and can be applied to exemplify the ECSA (Table 1). Moreover, the linear slope of the TNTs/NiPi electrode is 2.98 times that of the TNTs electrode, which further proves that loading NiPi increases the specific surface area and enriches the active site.To better evaluate the efficiency of NiPi on the surface recombination, we applied Na2SO3 as a hole scavenger to exclude the injection barrier for holes [63]. The PEC measurements were executed in 0.1\u00a0M PBS mixed with 1.0\u00a0M of Na2SO3. Both TNTs and TNTs/NiPi electrodes show higher photocurrent density (Fig. 7\na) in the presence of Na2SO3, due to sulfite oxidation. Fig. 7b plots displayed the photocurrent produced by TNTs/NiPi electrode in the presence or absence of Na2SO3. Also, it clearly shows the TNTs/NiPi demonstrated an apparent upsurge of photocurrent and reduction of onset potential in the hole scavenger, signifying that Na2SO3 eliminated the surface carrier recombination and boosted the injection of holes to the electrolyte than TNTs. Moreover, to better recognize the charge transfer dynamics, surface charge transfer efficiency (\u03b7surface) of TNTs and TNTs/NiPi at various potentials were assessed in Fig. 7a, and the results are shown in Fig. 7c. Indeed, TNTsphotoanodes yield only <40%\n\u03b7surface\n, even at potentials as higher potentials, at which the larger electric field impedes surface carrier recombination. After incorporating NiPi, \u03b7surface\n of the TNTs/NiPi photoanodes is boosted to \u223c80% at 1.23\u2009VRHE, suggesting enriched charge transfer kinetics.As shown in the above PEC measurement results, sparse NiPi decoration can be used to enrich the PEC features of TNTs in neutral electrolytes. Additionally, we tested this positive electrocatalytic result with acidic and alkaline electrolytes. By adding concentrated H2SO4 or KOH to aqueous solutions, the pH values of four same 0.1\u00a0M PBS electrolytes were tweaked to 1, 4, 10, and 14, respectively, to maintain a similar ionic environment. Fig. 7d shows that all the photocurrent densities of the modified TNTs/NiPi photoelectrodes in various pH value electrolytes improved in the low bias potential region. In these studies, NiPi catalysts successfully promoted the PEC features of TNTs over a varied pH, ranging from 1 to 14. Since the NiPi modification can enhance light absorption across a wide pH range, it suggests applications for added light-absorbing substrates in numerous electrolytes.Finally, the durability of the TNTs, and TNTs/NiPi photoanodes were investigated using chronoamperometry at 1.23\u00a0VRHE acquired in 0.1\u00a0M PBS, under irradiation (Fig. 7e). The photocurrent-time profile of TNTs/Ni revealed better durability than bare TNTs; after 4\u00a0h of testing, 78.50% of its initial features were upheld (0.59 vs. 0.749\u00a0mA/cm2). This validated the part of the NiPi in boosting the durability of the TNTs by decreasing the carrier recombination or through the fast and complete OE reaction [64,65]. In order to explore the mass loss of TNTs throughout the J-t plots, the XRD patterns and FE-SEM photographs of TNT/NiPi photoanodes were acquired at four hours. As can be seen, the TNTs/NiPi had no discernible changes in XRD pattern or FESEM image (Figs. 7f and S6) after 4\u00a0h compared with samples obtained before 4\u00a0h. The PEC features of the TNTs/NiPi photoanode are considerably higher than the recently reported TNTs-based photoanodes (Table 2\n). The higher PEC activity of the TNTs/NiPi than TNTs could be explained by light absorption, the active sites, recombination of charge carriers, charge-transfer efficiency at the electrode/electrolyte interface, and oxidation kinetics of the water molecule on the surface of the electrode [61,66-80].To better recognize the interfacial charge transfer behavior in the TNTs and TNTs/NiPi photoanodes, their electrochemical impedance spectra (EIS) were evaluated. Fig. 8\na displays the Nyquist curves of the TNTs/NiPi and the TNTs electrodes evaluated in both dark and illumination response at 1.0 VRHE, and the corresponding equivalent circuit. As is evident on the EIS Nyquist plot, the diameter of arc radius under both dark and irradiation response was obviously much smaller for the TNTs/NiPi composite photoanodes than for bare TNTs, demonstrating rapid interfacial charge transfer across the interface, more active separation of photogenerated charge carriers, subsequent in superior PEC features [80]. Remarkably, NiPi-decoration improved the charge-carrier density and electronic conductivity, thereby reducing the resistance (Table S1). These results confirmed that the NiPi enhanced the separation of carriers in the TNTs/NiPi photoanodes, thereby contributing to their superior PEC features.The Mott-Schottky (M-S) plots of the TNTs/NiPi and bare TNTs electrodes are shown in Fig. 8b. M-S curves of unmodified TNTs, and TNTs/NiPi exhibited a positive slope, as anticipated for semiconductors in the n-type regime.Evidently, the obtained M-S plots noticeably verified that the TNTs/NiPiphotoanodes had a smaller slope than its bare TNTs, demonstrating the enhanced donor density and conductivity of the former [60]. We estimated the donor density of TNTs and TNTs/NiPi electrodes from the M-S plots. The estimated donor density ofTNTs/NiPiwas 2.4\u00a0\u00d7\u00a01019 cm\u22123, superior to bareTNTs(7.90\u00a0\u00d7\u00a01018 cm\u22123). Moreover, the greater donor density was attributed to the incorporation of NiPi, which decreased the recombination of carriers and thus contributed to the higher photocurrent response of the TNTs/NiPi photoanode. Also, by inferring the M-S plots to the potential axis, as revealed in Fig. 8b, the flat band potentials (EFB) of the bare TNTsandTNTs/NiPi electrodes were estimated to be 0.048 and 0.235 VRHE, respectively (Table 1). The EFB ofTNTs/NiPiwas more positively shifted and smaller than that ofTNTs, matching the anodic change of the overpotential OE reaction. The apparent positive shift of EFB inTNTs/NiPi enriched the band bending at the TNTs/NiPi and electrolyte interface, thus reducing the recombination of the photogenerated charge pairs and the overpotential in the OE kinetics of theTNTs/NiPi photoelectrode.Based on the earlier PEC examinations, a probable PEC splitting mechanism of TNTs/NiPi was proposed in Fig. 9\n. The NiPi catalysts were loaded over the surface of and inside the TNTs. Interestingly, loaded NiPi catalysts in TNTs assist reaction sites and accelerate the transfer of photoinduced holes [74]. Also, the incorporation of NiPi enabled TNTs to absorb visible light. Under irradiation response, induced electrons in the VB of TNTs were agitated to form photoinduced carriers, and the electrons were moved to the CB of the tubes. Bearing in mind the Ni species of NiPi is an effective electrocatalyst for water oxidation reaction, a probable mechanism for the improved charge transfer is shown in\nFig. 9, which comprised of the hole oxidation of Ni2+\nto a higher valence state (Ni3+/4+), subsequently, it is reduced back to Ni2+\nstate with the instantaneous water oxidation to produce O2. The NiPi co-catalyst, as a greatly effective catalytic material for OE reaction, boosted the hole-trapping features to quicken the separation of photoinduced carriers.Herein, we have demonstrated a catalyst based on NiPi nanoparticles decorated over TNTs photoanode surface by anodization and photoelectrodepostion process. Decoration of the NiPi catalyst over TNTselectrodes revealed exceptional reaction features for OE reaction to produce oxygen gas. Thephotocurrentdensity was upsurged to 0.76\u00a0mA/cm2 with 3-fold enhancements at 1.23\u00a0VRHE for TNTs/NiPi related to the bare TNTs photoanode. The onset potential of TNTs/NiPi decreased to 0.06\u00a0VRHE, and the ABPE was 0.45% at 1.23\u00a0VRHE. The EIS spectrum and Cdl categorization specified that the fabricated NiPi particles delivered greater concentrations of the catalytically active region for the water splitting system. This study offers a new paradigm for designing nanostructured photoanode/donor density/hole transfer composite photoanodes for effective and highly stable solar fuel production, and the proposed material design strategy can be applied to other photoanodes.This project was funded by the National Plan for Science, Technology, and Innovation (MAARIFAH), King Abdulaziz City for Science and Technology, Kingdom of Saudi Arabia, Award Number (3-17-02-001-0011).All data generated or analyzed during this study are included in this published article (and its Supporting Information files).The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Supplementary data to this article can be found online at https://doi.org/10.1016/j.jscs.2022.101484.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  We present exemplary fabrications of controlled Nickel phosphate (NiPi)/TiO2 nanotubes arrays (TNTs) in phosphate buffer for boosted photoelectrochemical (PEC) water splitting. The TNTs/NiPi composite electrodes revealed a considerably enhanced photocurrent density of 0.76\u00a0mA/cm2, up to 3-time enhancements than bare TNTs, mostly because of the enhanced charge separation, decreased carrier recombination, and improving kinetics of the water oxidation. Also, we demonstrated that the NiPi can assist the PEC features of TNTs over a varied region of pH values from 1 to 14. Incorporation of NiPi over the TNTs surface advances the light absorption features of the electrode, resulting in an enhanced photogenerated charge carrier; and promotes the reactive sites for water oxidation, which was proved by the double-layer capacitance. The TNTs/NiPi photoelectrode exhibited excellent photostabilization under continuous illumination for 5\u00a0h, and the photoconversion efficiencies were 0.45%, 3-fold enhancements than with bare TNTs under the illuminations. Overall, this work might offer an innovative approach to fabricating and designing efficient electrodes with superior contact interfaces among photoanodes and numerous co-catalysts.\n               "}
{"full_text": "In an electrolyser device, the catalyst layer can contribute to losses in the overall running of the electrolyser due to inactive OER sites and low conductivity of the materials [1,2]. To achieve next generation inexpensive OER electrolyser catalysts, the catalysts themselves must be electrically conductive, mechanically and chemically stable under operating conditions, exhibit a high electrochemical surface area, and contain a high concentration of active sites for the evolution of O2. This has not been achieved to date for Proton exchange Membrane (PEM) and Alkaline Anion Exchange Membrane (AAEM) water electrolysis. One avenue to explore to make a catalyst that possesses all these characteristics would be to essentially combine different materials that exhibit these properties individually and make a \u2018super\u2019 catalyst.MXenes are a family of 2D materials, which are made up of transition metal carbides and nitrides, produced from MAX phases by various etching and delamination processes, Figure\u00a01\na [3,4]. A MAX phase has the general formula of Mn+1AXn where the M is an early transition metal, the A is an element from group 13 and 14 of the periodic table, and the X represents a carbon or nitrogen [5]. During the etching process, done in a fluoride ion based solution, the element from group 13/14 is removed from the MAX structure causing the carbide layers to become terminated by OH\u2212, O\u2212, Cl\u2212 or F\u2212 groups which are subsequently called \u2018surface groups or edge sites\u2019 [5]. The resulting structure is known as a \u2018MXene\u2019 [5,6].MXenes are known to be highly conductive, hydrophilic, and tuneable which are all advantageous properties that could lead to improving pure metal oxides when combined in a catalyst layer for OER [6]. As metal oxide materials lack high conductivity, which adds to the overall losses in an electrolytic cell, the addition of MXene materials could provide a high conductive support network making the hybrid material into a superior OER catalyst. Additionally, the hydrophilic nature of the MXenes will allow for the full coverage of OH\u2212 ions on the surface of the catalyst from the electrolyte which should help in the formation of O2. Finally, being able to tune the various MXene materials could significantly improve the conductivity and the hydrophilic nature of the MXene, hence further improving the ability to evolve O2. However, to date MXenes are not known to contain active sites for the OER, as no MXenes with metals for promoting the OER (e.g. Ni, Ru, Ir, Co, Mn or Fe) have been successfully synthesised (however MXenes containing Mn and Fe have been theoretically reported) [7,8]. Interestingly, for the opposite water splitting reaction, the hydrogen evolution reaction (HER), MXene materials have been proven to be promising through computation calculations and experimentally methods [9\u201311].On the other hand, Transition Metal Oxides (TMOs) are an exciting group of materials, that possess various intriguing physical properties that can change depending on the oxidation state of the material and are known to be active OER catalysts, Figure\u00a01b [12,13]. However, these TMO materials exhibit instability and dissolution during operation which renders these materials unsuitable for deployment in large-scale electrolyser devices [14].By combining inexpensive, active TMO catalysts with the conductive MXenes, most of the characteristics of the \u2018super\u2019 catalysts described previously, i.e. active site density due to the metal oxides and high conductivity due to the MXenes, in theory can be achieved [15]. The MXene materials may also provide a high surface area network for the TMO materials, similar to what has been attempted with carbon nanotubes (CNT) for OER composite materials, in order to improve electron transfer properties [16]. However, CNT based materials are known to contain metal impurities that enhance the OER (unaware to most) and corrodes under anodic OER potentials therefore are not the ideal composite materials with TMOs for the OER [17,18]. This high surface area network of a TMO/MXene composite will hopefully improve the operational stability of the catalyst layer due to improved mechanical properties.Finally, combining MXenes and the TMOs may even result in a lower loading of the TMOs, further lowering the catalyst costs. Hence, the combination of TMOs and MXenes into heterostructured layers or functionalising the MXenes with TMOs is a new and exciting avenue to be explored for the generation of highly active OER catalysts in PEM and AAEM electrolysers [19].To date only a hand full of papers have been published in this area, Figure\u00a01c and Table\u00a01\n. All of these papers have the same overarching conclusion that the addition of MXenes to transition metal oxides improves the initial OER performance compared to the TMO or the MXene alone.For example, Lu et\u00a0al. synthesised a hybrid MXene composite with Co3O4 decorated on Ti3C2Tx MXene flakes by a solvothermal reaction at 150\u00a0\u00b0C for 3\u00a0h [22]. The resulting Co3O4/Ti3C2Tx hybrid exhibited an OER overpotential of \u223c300\u00a0mV at a current density of 10\u00a0mA\u00a0cm\u22122 (the current benchmark used in literature when reporting the performance of OER materials) from linear sweep voltammetry measurements, Figure\u00a02\na [22]. Under the same OER conditions, the authors reported that a Co3O4 only material reached the same current density at overpotentials of 390\u00a0mV, while the Ti3C2Tx can be deemed OER inactive as the current density at an overpotential of \u223c400\u00a0mV is virtually zero [22]. Furthermore, in this study, Lu and co-workers investigated the effect of the ratio of metal oxide:MXene on the OER. The four ratios of metal oxide:MXene prepared were 1:0.1, 1:0.4, 1:1 and 1:10. The results showed that the lowest amount of MXene to metal oxide (i.e. 1: 0.1) exhibited the best OER results in terms of overpotential at a current density at 10\u00a0mA\u00a0cm\u22122.Benchakar and co-workers have observed a similar phenomenon for a Co layered double hydroxide (LDH)/Ti3C2Tx material fabricated by a polyol and solvothermal process [23]. In this particular study, the Co LDH/Ti3C2Tx outperformed the unsupported Co LDH for the OER by 50\u00a0mV, Figure\u00a02b [23]. The authors have also reported that the MXene structure can be preserved for oxidation during synthesis and the OER by the well distributed Co LDHs on the surface of the MXene. Interestingly, the authors also reported that this Co LDH/Ti3C2Tx hybrid synthesized by chemical routes outperformed a material which consisted of the Co LDH and the Ti3C2Tx mechanically mixed. This increase in performance of the chemically synthesized hybrid compared to the mechanically mixed catalyst was hypothesised to be due to the higher charge transfer resistance due to the close proximity of the Co LDHs and the Ti3C2Tx or the lower amount of active sites in the mechanically mixed material [23].Additionally, the integration of MXene materials into a composite with bimetallic TMOs has been shown to be advantageous for the OER. Yu et\u00a0al. reported that the synthesis of an FeNi-LDH/Ti3C2Tx material, fabricated by the co-precipitation of Fe2+ and Ni2+ from metal salts with already exfoliated MXene flakes under reflux, which also outperformed its individual counterparts under OER conditions, Figure\u00a02c. The authors attributed the superior OER performance of the composite material to the increase in the charge transfer properties from electrochemical impedance measurements, a shift in the Ni redox peaks to more anodic potential that may induce the OER earlier and an enhancement in the O binding strength of the FeNi LDH due to an electron extraction as a result of the coupling with the MXene [24].It is evident from literature that MXenes do significantly enhance the initial performance of TMO catalysts for the OER. However, MXene materials in a water-based solution are known to be unstable in air, Figure\u00a03\na [28]. The edge sites/surface groups of the MXene materials will oxidise first (due to deoxygenated species) to produce TiO2 which will then lead to the whole flake becoming oxidized and hence decreasing the conductivity of the materials, Figure\u00a03a. This can potentially be a huge problem for any electrochemical energy application including electrolysis.Unfortunately, the electrochemical instability of hybrid TMO/MXene materials have been already observed in the small amount of literature published in the area to date. For example, Lu et\u00a0al. multi-cycled their Co3O4/Ti3C2Tx hybrid in the OER region for 2000 times and observed a significant decrease in activity over time, Figure\u00a03b [22]. Furthermore, a FeNi-LDH/Ti3C2Tx material on Ni foam, synthesized by Yu and co-workers, also exhibited a decrease in activity over time during a chronopotentiometry test at 10\u00a0mA\u00a0cm\u22122 for 60\u00a0h, Figure\u00a03c.The instability of these materials at such low current densities presents a major drawback in the potential for MXenes to be incorporated as a component of an OER catalyst layer. If these hybrid materials are to be employed in large scale electrolysis devices, the materials must remain stable at high current densities of 1\u00a0A\u00a0cm\u22122 and higher. In order to alleviate major instabilities in these hybrid materials, investigations using in-situ or operando measurements in conjunction with electrochemical techniques need to be carried out in order to determine the reasoning for the instability of the materials, which is likely to be related to the degradation of the MXene to TiO2 under the extreme oxidative conditions present during the OER. The controlled synthesis of hybrid materials that can hinder the instability of the MXene component by covering the edge sites with active materials for the OER is another avenue which can be explored. Benchakar and co-workers have reported that their polyol/solvothermal synthesis method did fabricate a TMO/MXene hybrid material that is OER stable by covering the MXene edge sites with Co LDHs, however no stability tests were conducted to determine the long-term stability, which is needed to learn about possible TMO/MXene instabilities if superior such composite materials are to be designed in the future [23].A second route which could be undertaken to improve the OER performance and stability of TMO/MXene hybrids is to explore the (continuously expanding) world of MXenes materials. In literature the only MXene which has been utilised to synthesise TMO/MXene materials is Ti3C2Tx. There are numerous MXene materials now available and all exhibit different chemical and physical properties [3,29,30]. These other MXenes may be more stable and/or active for the OER when compared to the most common MXene, Ti3C2Tx. For example, it is well known that the incorporation/presence of Fe based materials/impurities into metal oxides improves the parent oxide towards the OER [31]. Therefore if the theoretically proposed Fe2CT2 MXene materials could be synthesized [32], this may lead to an enhancement in the OER activity, while also attaining high conductivity and hydrophilicity of the combined TMO/MXene materials compared to when a non-Fe MXene (e.g. Ti3C2Tx) is present in the hybrid material.A third route which could be undertaken to improve the performance of the metal oxide/MXene hybrid is to further investigate what ratio of TMO:MXene is optimum for the OER. Lu and co-workers showed that a ratio of 1:0.1 metal TMO:MXene is the best ratio for the OER in their study. However, this 1:0.1 TMO:MXene material contained the lowest ratio of MXene. Therefore, further studies into lower amounts of MXene in the hybrid need to be undertaken to determine if even less MXene is favourable for the OER when combined with metal oxides.Furthermore, Gogotsi and co-workers have recently reported that MXene materials containing more than one metal can be fabricated with different stoichiometries and can all be tuned in respect to the chemical properties they exhibit [3]. If the chemical properties of the aforementioned mixed metal MXenes can be tuned, this opens up a huge space to investigate the suitability of these TMO/MXene hybrid materials for the OER.Due to the large volume of TMO/MXene combinations, Density Functional Theory (DFT) will also play a vital role in the screening of the most promising TMO/MXene OER catalysts, as it has done for the HER [11]. DFT calculations, such as free energy and surface Pourbaix diagrams, could be utilised to predict stable surface terminations under OER conditions which could help experimental synthesis of stable TMO/MXene OER catalysts.Finally, for these hybrid TMO/MXene catalysts to reach a stage of commercialisation, the TMO/MXene materials would need to outperform the current state-of-the-art materials in terms of activity and stability in actual electrolyser devices and not just in a conventional three-electrode cell used in OER studies throughout academia. The reason behind this, is that in a conventional three-electrode cell, it has been shown that OER catalysts do not behave the same as in an electrolyser device [21,33,34]. However, to reach a stage of testing TMO/MXene catalysts in an electrolyser device, first more fundamental issues must be tackled including the oxidation of MXenes under OER potentials, the optimum metal oxide to MXene ratio for OER and improving the overall activity of the hybrid catalysts by synthetic/in-situ characterisation feedback mechanisms.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.M.P.B. would like to acknowledge the European Union's Horizon 2020 research and innovation programme under the Marie Sk\u0142odowska-Curie grant agreement No. 884318 (TriCat4Energy). V.N acknowledges the support of SFI AMBER (12/RC/2278_P2) and the ERC CoG 3D2DPrint (681544).", "descript": "\n                  MXenes are a class of 2D/layered materials which are highly conductive, hydrophilic, have a large electrochemical surface area and are easily processible into electrodes for energy applications. Since the discovery of MXenes over ten years ago, these materials have been mainly used in the preparation of electrodes for batteries and supercapacitors. However, due to their aforementioned properties, MXenes could potentially be utilised as a component in the catalyst layer for the Oxygen Evolution Reaction (OER). This opinion piece will discuss some of the recent literature in the area of hybrid catalysts consisting of various Transition Metal Oxides (TMOs) and MXenes for the OER. We will also discuss current drawbacks and future outlook in this new area of research.\n               "}
{"full_text": "The majority of energy today is produced through the combustion of fossil fuels such as coal and petroleum. However, energy production is accompanied by the emission of greenhouse gases, including methane and carbon dioxide, resulting in global warming and other international environmental issues. Hydrogen, which offers high energy density, has emerged as an alternative fuel [1]. Methane has come under the spotlight as a resource for clean fuel production on the basis of its high H/C ratio, making it advantageous in hydrogen production, as well as large availability in natural gas, shale gas, landfills, and byproduct gases.Despite its less significant greenhouse effect, methane has a global warming potential that is about 25 times that of CO2. As such, many research groups have attempted to reduce carbon by using methane in reforming or direct decomposition [2].Among the various methane reforming reactions, steam methane reforming (SRM) is popular around the world as it is an affordable solution that offers high hydrogen yield. However, this method produces byproduct gases such as CO and CO2, adding to the cost of processing, and again contributing to greenhouse gases [3,4].The above issue was addressed through the thermal decomposition of methane (TDM), which produces hydrogen and solid carbon without any emission of CO and CO2\n[5,6]. The cost of producing 1000 Nm3/h of hydrogen is about USD 2167 to 3764 under TDM, and USD 2639 under SRM, which highlights the need for the former to secure cost competitiveness [7].Catalytic chemical vapor deposition (CCVD) is recognized as a low-cost approach to the TDM process. Methane is thermally decomposed at temperatures higher than 1000 \u2103, but its decomposition temperature can be lowered to 600\u2013850 \u2103 with catalysts. In addition, carbon of relatively low grades, such as amorphous CB, is obtained using the TDM process, but crystalline carbon nanotube(CNT) can be derived by applying the CCVD method [8\u201310].Recently, CNTs have been extensively studied as a conductive agent for the cathodes of lithium ion batteries [11\u201313], and they have vast applications covering sensors [14] to fuel cells [15]. With the increase in demand for CNTs, there has been active research on related production processes, among which the CCVD method has drawn considerable attention [16].The catalysts used in the CCVD method can be largely divided into carbon-based and metal-based catalysts. Carbon-based catalysts are affordable and stable at high temperatures, and do not require separate refinement of carbon products. Some examples of such catalysts are active carbon [17] and carbon black [18].Commonly used metal-based catalysts are transition metals with high carbon solubility, such as Co [19,20], Ni [21], and Fe [22,23] and bimetallic catalysts composed of Mo, Mn, and Cu. According to past studies, the Co-Mo catalyst offers higher catalytic activity compared to other bimetallic catalysts. Gamal, Ahmed et al. [24] examined the catalytic activity of cobalt catalysts including different promoters such as Fe, Mo, Cu, and Ni. A higher methane conversion rate of 6.3 % was observed when Mo was added as a promoter and allowed to react for 0.5\u00a0h. Chai, Siang-Piao et al. [25], who studied the effects of adding Cu, Fe, Mo, and Ni to Co, reported a methane conversion rate of 6.3 % and a carbon yield of 281 % for Mo with a reaction time of 1.5\u00a0h.Metal oxide supports such as Al2O3\n[26\u201329] SiO2\n[30], ZrO2\n[31], and MgO [32\u201335] are also used to enhance the structural stability of catalysts, thereby maintaining catalytic activity. MgO supports are more easily removed in acidic solutions, minimizing the damage to products caused by acidic exposure [36,37]. The electrical conductivity of CNT is another important factor to be considered when using them as a conductive agent for cathodes of lithium-ion batteries. The use of MgO supports is favored since electrical conductivity is negatively affected by the introduction of a functional group to CNT under acidic conditions during catalyst removal [38].Catalytic activity is also influenced by the surface properties of supports [39]. Supports having a larger specific surface area (SSA) tend to have better catalytic activity due to the higher dispersion of active sites, which translates to a larger reactive surface area. In methane catalytic decomposition (MCD) reactions, the SSA of supports is expected to cause differences in carbon properties and catalytic activity.For this reason, various methods of synthesis of high surface area MgO have been examined [40\u201342]. Most methods require processing of precursors in various stages, as well as post-processing using expensive or toxic solvents.Meanwhile, Bartley, Jonathan K., et al. [43] proposed a method of synthesizing high surface area MgO via the thermal decomposition of different precursors without requiring separate post-processing, but research on the use of thermochemical processes as catalytic supports is still insufficient.This study examined the MCD of Co-Mo/MgO catalysts and changes in properties of carbon products in relation to the surface properties of supports. The temperature of synthesis of MgO supports was varied to fabricate supports having different SSA, and Co-Mo/MgO catalysts were obtained through impregnation of Co and Mo in the same ratio. To assess the effects of changes in surface properties of MgO supports on MCD, this study analyzed and evaluated the fabricated catalysts, MCD reactions, and properties of carbon products.Magnesium carbonate (MgCO3, Sigma-Aldrich, Lot No. BCCC2599), as MgO precursor was oxidized amount of 7\u00a0g for 2\u00a0h in a muffle furnace at varying temperatures of 400 \u2103, 500 \u2103, and 600 \u2103, thereby preparing MgO supports having different surface properties. After oxidation treatment, the obtained MgO supports were 3.8\u00a0g, 3.6\u00a0g, and 3.2\u00a0g depending on the 400\u00a0\u00b0C, 500\u00a0\u00b0C, and 600\u00a0\u00b0C temperature, respectively. Commercial MgO powder (Sigma-Aldrich, Lot No. MKCH6857), which has very little SSA, was used as a control. The MgO specimens were named 4MgO, 5MgO, and 6MgO depending on the temperature of synthesis, and commercial MgO as C-MgO.Cobalt (\u2161) nitrate hexahydrate (Co(NO3)2\u00b76H2O, SAMCHUN, Lot No. 012,920) and ammonium molybdate tetrahydrate ((NH4)6Mo7O24\u00b74H2O, Alfa Aesar, Lot No. 10,223,334) were used as a precursor of cobalt and molybdenum, respectively. First, 5.1\u00a0g and 1.9\u00a0g, corresponding to 25 wt% Co and 25 wt% Mo, was respectively dissolved in distilled water, and impregnated via a rotary evaporator in four types of supports: C-MgO, 4MgO, 5MgO, and 6MgO. The catalysts were dried at 80 \u2103 for 12\u00a0h and calcined at 400 \u2103 under atmospheric conditions for 4\u00a0h, resulting in CM/C-MgO, CM/4MgO, CM/5MgO, and CM/6MgO.To synthesize CNT through MCD, 0.2\u00a0g of catalysts was placed in a horizontal reactor, having an internal diameter of 6\u00a0cm and length of 120\u00a0cm, equipped with a quartz boat. The temperature was increased to 800 \u2103 (10 \u2103/min, Ar 100\u00a0cc/min), and the active metal in the oxidized state was reduced at a flow rate of H2:Ar (20\u00a0cc:80\u00a0cc) for 30\u00a0min. Next, Ar (100\u00a0cc/min) purging was performed for 40\u00a0min to create inert conditions. After performing MCD at 50\u00a0cc/min for 2\u00a0h using a gas mixture with a mole ratio of CH4:N2\u00a0=\u00a095:5, the catalysts were cooled to room temperature under argon. The reactor set up used in this study is described in Fig. S1.The used catalysts were denoted as U-CM/C-MgO, U-CM/4MgO, U-CM/5MgO and U-CM/6MgO.The physical and chemical properties of supports, synthesized catalysts, and used catalysts were characterized through various analytical techniques.The surface properties of the four MgO supports, synthesized catalysts, and used catalysts were examined by conducting an N2-adsorption desorption (TRISTAR 3020) analysis at 77\u00a0K after heat treatment at 474\u00a0K over 8\u00a0h.The crystal structure of the synthesized catalysts was analyzed by X-ray diffraction (Rigaku Ultima \u2163), with Cu K\u03b1 (\u03bb\u00a0=\u00a01.5418\u00a0\u00c5) as the X-ray source, current of 40\u00a0mA, voltage of 40 Kw, and 2\u03b8\u00a0=\u00a010\u201390\u00b0. crystallite size was calculated by the Scherrer equation using the full width at half maximum (FWHM) of XRD peaks.An Inductively Coupled Plasma Mass Spectrometer (Agilent ICP-MS 7700S) was used to analyze the amount of impregnated active metal in catalysts.The distribution of surface elements on the catalysts was analyzed by X-ray photoelectron spectroscopy (Axis Supra) with monochromatic Al-K\u03b1 (\n\nh\n\u03c5\n\n=\n\n1486.6\ne\nV\n\n) as the X-ray source.To determine the amount of hydrogen consumed in the reduction of metal oxides within catalysts and the extent of reduction, the specimens were pre-processed at 200 \u2103 under Ar 50\u00a0ml/min for 1\u00a0h and analyzed by H2-TPR (AUTOCHEM \u2161 2920) using 10 % H2/Ar gas.After MCD, the catalysts were heated to 900 \u2103 at 5 \u2103/min under air conditions, and the oxidation temperature of carbon products and decrease in nitrogen were measured with a Thermo Gravimetric Analyzer (Thermo plus EVO \u2161).To compare the crystallinity of carbon products after MCD, a Raman Spectroscopy (Nanophoton Ramanforce) analysis was performed. The wavelength was 532.04\u00a0nm and the slit width was 50 \u339b.SEM (Scanning Electron Microscope, TESCAN MIRA3 LMU) analysis was carried out at an accelerating voltage of 10.0\u00a0kV to examine the shape and size of catalysts and products.TEM (Transmission electron microscopy) images were recorded using a FEI Talos F200X (ThermoFisher Scientific, USA) operating at a voltage of 200\u00a0kV.The electrical conductivity characteristics of the samples subjected to MCD were investigated using a powder resistance meter (HPRM-FA2, HANtech Co. Ltd., Korea), while increasing the pressure from 400 to 2000kgf at intervals of 400kgf.Lastly, to analyze the catalytic activity and hydrogen selectivity over time, the gases produced were analyzed using a Gas Chromatography (YL6500 GC). N2, H2 and CH4 were analyzed with a thermal conductivity detector (TCD) equipped with a packed column (ShinCarbon ST 80/100, RESTEK).\nFig. 1\n shows the N2 adsorption-desorption results for the four types of supports: C-MgO, 4MgO, 5MgO, and 6MgO.As can be seen from the N2 adsorption-desorption isotherms (Fig. 1(a)), 4MgO and 5MgO fall under IUPAC isotherm type \u2161, while 6MgO is type III [44]. Types \u2161 and III are characterized by hysteresis loops, which occur due to capillary condensation within mesopores [45]. A hysteresis loop, classified as H3 by IUPAC, arises from the adsorption of non-polar gases associated with slit-shaped particles [46]. On the other hand, C-MgO is composed of particles that are almost non-porous, and thus does not exhibit N2 adsorption or desorption. Also, developed pore structure was observed in case of the synthesized MgO in TEM images (see Fig. S2).\nFig. 1(b) shows the pore size distribution of the MgO supports: 4MgO has a bimodal distribution, comprised mostly of pores of 3 to 4\u00a0nm, while 5MgO pore size was around 7\u00a0nm. And the pore size of 6MgO were distributed from 10 to 100\u00a0nm. As a result, it was confirmed that 4MgO, 5MgO had mesopore structure and 6MgO had mesopore with macropore structure. Lastly, C-MgO support had non-porous structure. And the pore diameter of each support is presented in the Fig. S3.\nTable 1\n presents the SSA, pore volume of C-MgO, 4MgO, 5MgO, and 6MgO supports. The SSA was found to be 0.481\u00a0m2/g, 170\u00a0m2/g, 131\u00a0m2/g, and 50.4\u00a0m2/g for C-MgO, 4MgO, 5MgO, and 6MgO, respectively. The specimens arranged in increasing order of SSA are as follows: 4MgO\u00a0>\u00a05MgO\u00a0>\u00a06MgO\u00a0>\u00a0C-MgO.The differences in surface properties of the synthesized 4MgO, 5MgO, and 6MgO assumed to the amount of oxidation of functional groups, such as \u2013H2O, \u2013CO2 remaining in precursors according to the oxidation temperature during MgO synthesis. Meanwhile, Bartley, Jonathan K., et al. [43] showed TGA analysis under an air atmosphere of MgCO3, as MgO precursor and found that H2O was removed at 250\u2013400 \u2103 and CO2 was removed at 325\u2013550 \u2103. Also, the SSA of supports were decreased with increasing MgCO3 oxidation temperature while the crystallinity was increased.\nFig. 2\n presents the TGA and DTG analysis results of the used catalysts obtained under atmospheric conditions after MCD. The weight decrease in air of the U-CM/C-MgO, U-CM/4MgO, U-CM/5MgO, U-CM/6MgO catalysts were 43.76 %, 60.12 %, 58.43 %, and 48.67 %, respectively. This decrease resulted from the oxidation of carbon products following MCD. Each catalyst produces a different amount of carbon products because of the varying dispersion of active metals, impregnated in the four supports, contributing to different surface areas involved in reactions.The DTG results showed that used catalysts experienced peaks in a temperature range of 500 to 600 \u2103 due to oxidation of deposited carbon. Amorphous carbon is generally oxidized at temperatures below 400 \u2103 [47], and MWCNT (Multi-Walled Carbon Nanotube) in the range of 500 to 700 \u2103 [48]. As such, we can deduce that the carbon product obtained from the four types of catalysts after MCD is MWCNT. This was verified through the SEM images in Fig. 4.In addition, the weight of catalyst after MCD reaction and weight loss of used catalyst by TGA analysis were described in the Table S1.\nFig. 3\n(a) shows the results of Raman spectroscopy of the used catalysts. D, G band peaks were observed for all used catalysts. The D band peak at around 1350\u00a0cm\u22121 is caused by disordered carbon, such as amorphous carbon or defective graphite sheets [49]. The TGA results (Fig. 2) verified that the D band is induced by the disordered structure of MWCNTs, rather than amorphous carbon. The G-band peak at 1580\u00a0cm\u22121 corresponds to CC stretching vibrations that are characteristic of graphite [50]. The intensity ratio (IG/ID) of G band and D band is used to compare the crystallinity of carbon materials, where a higher ratio indicates higher crystallinity [51]. The intensity ratios IG/ID of MWCNTs on used catalysts were 2.15\u00a0\u00b1\u00a00.04, 1.72\u00a0\u00b1\u00a00.05, 1.40\u00a0\u00b1\u00a00.06 and 1.41\u00a0\u00b1\u00a00.04 for U-CM/C-MgO, U-CM/4MgO, U-CM/5MgO, and U-CM/6MgO, respectively. Since the IG/ID ratios of MWCNT grown on the used catalyst were higher than 0.98\u00a0\u00b1\u00a00.04 of conventional MWCNT (Sigma-Aldrich, Lot No. MKCM1457), it is considered that the synthesized MWCNTs using Co-Mo/MgO catalysts have high crystallinity.Deposited carbon on the U-CM/C-MgO catalyst showed higher crystallinity than that of MWCNTs produced by the supported catalysts on synthesized MgO. The formation of Carbon Nano Sheets (CNS) is deduced to have influenced the IG/ID ratio [51]. The growth of CNS can be seen TEM image. (Refer to Fig. S4.).In addition, it has been reported that the IG/ID ratio is closely related to the MWCNT diameter [52]. Based on this, Fig. 3(b) shows the correlation between the IG/ID ratio and the diameter of MWCNT according to SSA of support. In the case of U-CM/C-MgO, it is difficult to correlate due to the formation of a small number of CNS, but the IG/ID ratio and CNT diameter trend were consistent with the remaining catalysts mainly produced by CNTs. This was also consistent with the trend of the mean diameter of CNTs measured in the SEM images in Fig. 4\n(a\u2013d).\nFig. 4 presents SEM images revealing the shapes of carbon products of each used catalyst. A diameter distribution histogram was derived from a bar graph of 100 CNT diameter values. The average diameter was 48.3\u00a0\u00b1\u00a012.9\u00a0nm for (a) the U-CM/C-MgO catalyst, 20.5\u00a0\u00b1\u00a07.8\u00a0nm for (b) the U-CM/4MgO catalyst, 34.3\u00a0\u00b1\u00a013.9\u00a0nm for (c) the U-CM/5MgO catalyst, and 36.5\u00a0\u00b1\u00a08.2\u00a0nm for (d) the U-CM/6MgO. In addition, active metals and CNT diameter of U-CM/C-MgO and U-CM/4MgO were confirmed through TEM images. (Refer to Fig. S5.).The mechanism by which methane is decomposed in contact with the Co0 surface (Sco) is as follows;\n\ni)\nCH4\u00a0+\u00a02SCo\u00a0\u2194\u00a0CH3SCo\u00a0+\u00a0HSCo,\n\n\nii)\nCH3SCo\u00a0+\u00a0SCo\u00a0\u2194\u00a0CH2SCo\u00a0+\u00a0HSCo,\n\n\niii)\nCH2SCo\u00a0+\u00a0SCo\u00a0\u2194\u00a0CHSCo\u00a0+\u00a0HSCo,\n\n\niv)\nCHSCo\u00a0+\u00a0SCo\u00a0\u2194\u00a0CSCo\u00a0+\u00a0HSCo,\n\n\nv)\nCSCo\u00a0\u2194\u00a0CCo\u00a0+\u00a0S,\n\n\nvi)\n2HSCo\u00a0\u2194\u00a0H2\u00a0+\u00a02SCo\n\n\n\nCH4\u00a0+\u00a02SCo\u00a0\u2194\u00a0CH3SCo\u00a0+\u00a0HSCo,CH3SCo\u00a0+\u00a0SCo\u00a0\u2194\u00a0CH2SCo\u00a0+\u00a0HSCo,CH2SCo\u00a0+\u00a0SCo\u00a0\u2194\u00a0CHSCo\u00a0+\u00a0HSCo,CHSCo\u00a0+\u00a0SCo\u00a0\u2194\u00a0CSCo\u00a0+\u00a0HSCo,CSCo\u00a0\u2194\u00a0CCo\u00a0+\u00a0S,2HSCo\u00a0\u2194\u00a0H2\u00a0+\u00a02SCo\nSo, the difference in diameter of MWCNT produced by each catalyst after methane decomposition is affected by dispersion rate of active metal due to the surface properties of MgO supports. Fig. S6 shows the results of XPS analysis in the reduced state of CM/C-MgO and CM/4MgO. Through this analysis, it was confirmed that Co0 was formed in the reduced catalyst.In addition, the electrical conductivity at 2000 kgf of before and after acid treatment of used catalysts having different surface properties of the support are showed at Table S2. Before acid treatment of each used catalyst, U-CM/C-MgO, U-CM/4MgO, U-CM/5MgO, U-CM/6MgO, respectively, showed electrical conductivity values of 25.89, 55.54, 51.82, and 39.57 S/cm. Also, after acid treatment, 30.19, 67.55, 49.19, and 47.61 S/cm. A decrease in MWCNT diameter shows high crystallinity. As diameter becomes thinner, crystallinity and electrical conductivity are higher. Also, all used catalyst was increased electrical conductivity after acid treatment. It can be seen that U-CM/4MgO, which has high electrical conductivity, has a relatively thinner MWCNT [54]. Super P (Imerys Graphite & Carbon), which is used as a conductive additive for a cathode material for a lithium ion battery, exhibited an electrical conductivity of 29.43 S/cm under the same conditions.Through this, it was confirmed that the carbon material produced in the MCD reaction has electrical conductivity that can be used as the conductive additive. In addition, it was confirmed that the electrical conductivity characteristics of the carbon derived from MCD increased with increasing SSA of the MgO support.\nFig. 5\n shows the methane conversion over time on stream obtained from online GC. The initial methane conversion at 4\u00a0min on-stream was 68.9 %, 85.7 %, 84.3 %, and 83.4 % for U-CM/C-MgO, U-CM/4MgO, U-CM/5MgO, and U-CM/6MgO, respectively. After 108\u00a0min, the methane conversion rate was 0 %, 21.4 %, 17.6 %, and 14.9 %. The U-CM/C-MgO catalyst became inactive most rapidly, while the catalytic activity of U-CM/4MgO, U-CM/5MgO, and U-CM/6MgO decreased more gradually over time (Fig. 5).MgO supports with a larger SSA and greater dispersion have a broader area of contact between methane and active metal, which allows for a higher conversion rate. A larger SSA facilitates metal dispersion, enhancing interaction with supports and suppressing sintering.The difference in gas selectivity was also examined. The initial hydrogen selectivity was 72.76 %, 87.92 %, 86.94 %, and 86.32 % for catalysts synthesized using C-MgO, 4MgO, 5MgO, and 6MgO supports, respectively. All four catalysts showed a gradual decrease with reaction time (refer to Fig. S7).\n\n(1)\n\n\n\n\nCH\n\n4\n\n\u2192\nC\n+\n2\n\nH\n2\n\n\n\n\nWhere, 1 mole of methane is decomposed to produce 1 mole of carbon and 2 moles of hydrogen. Therefore, the higher the methane conversion rate, the higher the carbon production (Fig. 2) and hydrogen selectivity. And the hydrogen selectivity was also reduced according to the deactivation of the catalyst.CM/C-MgO, CM/4MgO, CM/5MgO, and CM/6MgO catalysts were fabricated with supports having different surface properties, and the catalytic activity was compared in relation to MWCNT properties after MCD reactions. The MWCNT yield and catalytic activity increased in the order of SSA of supports, namely, 4MgO\u00a0>\u00a05MgO\u00a0>\u00a06MgO\u00a0>\u00a0C-MgO. The resulting MWCNT also had a thinner diameter. A comparison of crystallinity with commercial MWCNT based on IG/ID ratio showed that MWCNT produced from the four catalysts had better crystallinity.\nFig. 6\n shows the XRD analysis of the fresh Co-Mo/MgO catalysts. The major diffraction peaks of MgO (JCPDS no. 45-0946) and CoO (JCPDS no. 65-0902) were observed for all catalysts at 2\u03b8\u00a0=\u00a036.9\u00b0, 42.8\u00b0, 62.3\u00b0, 74.7\u00b0, and 78.6\u00b0 [53,54]. In addition, the diffraction peaks of MgMoO4 and CoMoO4 (JCPDS no. 72-2153, no. 21-868) were observed at 2\u03b8\u00a0=\u00a026.3\u00b0 [55,56]. The results verify that the same crystal structure had formed in the four catalysts. As can be seen from the N2 adsorption-desorption of each support (Table 1), the impregnation of catalysts with C-MgO had sharp peaks, whereas those impregnated with porous supports such as 4MgO, 5MgO, and 6MgO had broader peaks, indicating that the crystals were smaller [57]. This was compared with the cobalt oxide crystallite size calculated at 42.8\u00b0 using the Scherrer equation (2) for the peaks exhibited by the four catalysts. Where L is crystallite size (nm), \u03bb is the X-ray wavelength in nanometer (nm), \u03b2 is full width at half maximum (FWHM) and K is a constant related to crystallite shape, normally taken as 0.9. The \u03b8 can be in degrees or radians, since the cos\u03b8 corresponds to the same number.\n\n(2)\n\n\nL\n\n=\n\nK\n\u03bb\n/\n(\n\u03b2\n\nc\no\ns\n\u03b8\n\n)\n\n\n\n\nThe cobalt oxide crystallite size of catalysts was 162.17\u00a0nm, 11.74\u00a0nm, 13.94\u00a0nm, and 14.38\u00a0nm for CM/C-MgO, CM/4MgO, CM/5MgO, and CM/6MgO, respectively. In the case of 4MgO, 5MgO, and 6MgO porous supports, the crystallite size was relatively smaller due to the impregnation of active metals in pores allowing stronger interactions, thus resulting in greater dispersion. The cobalt oxide crystallite size of CM/C-MgO, which has smallest SSA, was much bigger than that of catalyst supported on synthesized MgO due to the clustering by weak interaction with support.\nFig. 7\n shows the results of a H2-TPR analysis, conducted to examine interactions between supports and active metals, and the temperature of reduction of compounds within catalysts. First, the fresh catalysts showed four peaks at similar temperature ranges. The first peak in the range of 300 to 360 \u2103 is due to the reduction of Co3+ of Co3O4 to Co2+. Peaks at 500 to 750 \u2103 are due to the reduction of Co2+ of Co3O4 to Co0 and compounds such as CoMoO4 and MoO3 in the catalysts [55,58]. The peaks at temperatures higher than 850 \u2103 were due to the reduction of the CoO-MgO solid solution and compounds such as MgMoO4, MgCo2O4, and MoO2\n[55,59].Compared to the catalyst prepared with C-MgO, the catalysts using synthesized MgO saw the reduction temperature shift to a higher range. Because the impregnation of active metals on porous supports reinforced the interaction between the active metals and supports. In case of the catalysts using synthesized MgO, the reduction temperature was decreased with increasing SSA of MgO support. It can be deduced that the dispersion rate of active metals was increased with increasing the SSA of MgO support. The intensity and area of CM/C-MgO reduction peaks were higher than the catalysts using synthesized MgO because of bigger size of active metal crystal. The tendency of the dispersion rate of active metal according to the SSA of the support was consistent with the crystallite size calculated from XRD.\nTable 2\n shows the proportion of metals existing on the surface of catalysts based on the surface/bulk ratio of active metals impregnated in the fresh catalysts. And this formula was calculated from the XPS and ICP data. The CM/C-MgO which is supported on non-porous MgO had the highest Surface/Bulk ratio of metals. And also, Surface/Bulk ratio of fresh catalysts was decreased with increasing SSA of MgO supports. It can be inferred that the decrease of the Surface/Bulk ratio is due to the increase of active metal dispersion into the pores by the increase of interaction with support. XPS and ICP data of each catalyst are refer to Table S3.Next, the surface properties were evaluated after impregnation of active metals and MCD reaction. C-MgO and 6MgO, which had smaller SSA, saw an increase in SSA after active metals were impregnated. 4MgO and 5MgO, with larger SSA of 170\u00a0m2/g and 131\u00a0m2/g respectively, underwent a decrease in SSA. The increase in SSA of C-MgO and 6MgO can be traced to the active metals being impregnated on the external surface of supports. In case of CM/4MgO and CM/5MgO, the SSA was decreased because of blockage by active metals impregnated into pores. Specifically, the pore volume of the 4MgO support decreased from 0.4\u00a0cm3/g to 0.32\u00a0cm3/g, and that of the 5MgO support from 0.5\u00a0cm3/g to 0.32\u00a0cm3/g. Also, used catalysts showed an increase in SSA, attributed to the greater adsorption and desorption of N2 following the growth of CNT after MCD. This result could be confirmed through the data that each used catalyst was oxidized at 700 \u2103 to remove the deposited carbon. (Table 2).\nFig. 8\n Shows the MCD behavior of Co-Mo catalysts supported on non-porous MgO (C-MgO; Fig. 8(a)) and porous MgO (4MgO, 5MgO, 6MgO; Fig. 8(b)). In the case of non-porous MgO, the active metals are present in large crystallite sizes with low dispersion on the support. The lowly dispersed large Co-Mo particles induced low conversion due to the low contact rate with CH4 gas. And also, the carbon products on large crystals were formed CNS or thick fibrous type by the carbon deposition mechanism on metal catalysts. In the case of a catalyst with a low dispersion of active metals, CNS or thick CNT were produced due to the horizontal carbon growth. On the other hand, in the case of a highly dispersed catalyst, a large number of CNT having a thin diameter were generated due to the vertical growth of carbon [47,51].So, comparatively high dispersed Co-Mo catalyst induced high methane conversion and thin MWCNT because of high rate of CH4 contact and vertical carbon growth. In conclusion, a catalyst with highly dispersed active metal can induce high methane conversion and carbon productivity, and can produce high crystalline/conductivity MWCNTs with a thin diameter.In this study, the porous MgO with high porosity was prepared by simplified oxidation process of MgCO3 with varying temperature. The SSA and pore size distribution of the prepared MgO were controlled by the temperature of the oxidation process. The improved pore structure of MgO enhanced the dispersibility of the Co active metal, thereby improving methane conversion and MWCNT production. The dispersion of supported active metal was increased with increasing porosity of MgO due to the enhanced improved interaction. The CM/4MgO catalyst using 4MgO support prepared by 400 \u2103 oxidation, which had the largest SSA of 170\u00a0m2/g, showed highest methane conversion of 85.7 % with 60.12 % of MWCNT production. The produced MWCNT on CM/4MgO had highest crystallinity and electrical conductivity with average diameter of 20.5\u00a0\u00b1\u00a07.8\u00a0nm. Consequently, the 4MgO support induced the highest dispersilbility of Co, and the methane decomposition behavior and MWCNT properties were affected by dispersibility of active metal.The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: This work was supported by the Technology Innovation Program (20010853, Development of natural gas based carbon material on graphite structure for high crystalline conductivity) funded By the Ministry of Trade, Industry & Energy (MOTIE, Korea).This work was supported by the Technology Innovation Program (20010853, Development of natural gas based carbon material on graphite structure for high crystalline conductivity) funded By the Ministry of Trade, Industry & Energy (MOTIE, Korea).Supplementary data to this article can be found online at https://doi.org/10.1016/j.jiec.2022.05.008.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  To investigate the effects of porous properties of supports on the methane decomposition of Co-Mo/MgO catalysts, supports having different porous properties were prepared by varying the synthesis temperature of MgO. Co-Mo/MgO catalysts were prepared using the impregnation method. The CM/4MgO catalyst using 4MgO support, which had the largest specific surface area of 170\u00a0m2/g, showed highest methane conversion of 85.7 % with 60.12 % of MWCNT production. The produced MWCNT on CM/4MgO had highest crystallinity and electrical conductivity with average diameter of 20.5\u00a0\u00b1\u00a07.8\u00a0nm. Overall, the improved pore structure of MgO prepared by the oxidation process enhanced the dispersibility of the Co active metal, thereby improving methane conversion and MWCNT production.\n               "}
{"full_text": "Carbon nanotubes (CNTs) have been known as the miracle material in the material engineering sectors ever since been discovered in 1991 [1]. CNTs have been utilised in many sectors such as toughening agents in ceramic and metals [2,3], wastewater treatment [4,5], micro- and nano-electronics [6,7], energy scavenging [8,9], and drug delivery carrier [10,11]. CNTs were first synthesised through the arc-discharge technique [1], while today, CNTs can be mass synthesised through different methods, i.e. laser ablation [12], hydrothermal [13], electrolysis [14\u201317], gas phase growth [18] and chemical vapour deposition (CVD) [19\u201326].CVD is the most practical synthesis option for researchers and manufacturers to mass-produce CNTs at a lower operation cost. The supplied thermal energy from the heating filament catalysed the carbon nucleation on the catalyst surface or carbon dissolution into the catalyst core in the CVD system. Parameters control in the CVD system, e.g. the gas flow rate, is critical and it determines the CNTs\u2019 quality and quantity. Lin et\u00a0al. optimized the concentration of Ni catalyst (10\u00a0w.t.%) and the methane flow rate (100 sccm) during the CNTs growth on the TiB2 substrate [27]. Mamat et\u00a0al. also presented a deposition temperature study on the CNTs growth using silica-supported NiFe2O4 catalyst. They claimed the highest CNTs yield with the best quality requiring 700\u00a0\u00b0C with the flowing H2/C2H4 gas mixture at a 50:50 ratio [28].Furthermore, Wang et\u00a0al. produce MWCNTs with symmetric walls through an optimized fluidised CVD system [29]. They purged the gasified polyethylene to react with the fluidized NiCl2 catalyst (metallic salt) in an inert gas flow environment (Ar/H2) at 750\u00a0\u00b0C and grew symmetrical MWCNTs with diameters ranging between 30 and 40\u00a0nm on the catalyst. The ID/IG ratio of the fabricated MWCNTs is 1.12. Karaeva et\u00a0al. derived Ni from nickelocene (metallocene) and grew straight CNTs, with resistivity 6.7\u00a0\u00d7 10\u22123 \u03a9\u2022m, on Ni catalyst using floating catalyst CVD process at 1150\u00a0\u00b0C [30]. Researchers have made extensive refinements and optimization in the CVD system to produce large quantity CNTs with higher quality. However, there is always room for parameter optimization for the CVD technique to understand the best mechanism for obtaining a low-cost and straightforward technique to achieve mass production of CNTs.This study investigated the effect of the precursor flowing path (namely distance between precursor and catalyst) and the catalyst size, subjected to the synthesised CNTs\u2019 quality and quantity from the CVD system. NiO particles and ethanol were selected as the catalyst and carbon precursor, while argon and hydrogen gas as the carrier gas and reducing agent, respectively.25.0\u00a0g nickel nitrate hexahydrate (Ni(NO3)2\u20226H2O, >97.0% purity, Systerm) was stir-mixed with 250\u00a0mL ethanol (C2H5OH, 95.0% purity, Systerm) homogenously, in which the stirring speed was set at 1300\u00a0rpm. The mixture was then dried at 100\u00a0\u00b0C (72\u00a0h) and oxidised at 400\u00a0\u00b0C (4\u00a0h) in the air. The initial oxidation temperature for bulk Ni is around 600\u00a0\u00b0C; however, the Ni powder oxidised at a lower temperature (398\u00a0\u00b0C) [31]. The oxidised powder was milled with different milling duration (0, 4, and 7\u00a0h) using high-energy ball milling (HEBM), labelled as M1, M2, and M3 (see Table\u00a01).To investigate the effect of precursor flowing path (D) on the CNTs growth, M1 was fixed in the centre of the tubular furnace to conduct CVD (see Fig.\u00a01a). The stainless steel tube acted as the carbon precursor channel, with a fixed internal diameter of 4\u00a0mm. The precursor flowing path (D) was adjusted by replacing the steel tube with different lengths (see Fig.\u00a01b). The sample was labelled as N1, N2 and N3 after the CVD process, subjected to the distance, where D = 2, 7 and 12\u00a0cm. Ar gas was flowed into the furnace for 15\u00a0min before heating the metal catalyst to ensure an inert environment during the deposition. The metal catalyst was heated from room temperature to 600\u00a0\u00b0C at the heating rate of 10\u00a0\u00b0C/min, with the Ar flowing at 50 sccm. At 600\u00a0\u00b0C, the flowing Ar gas and H2 gas carried the vaporised ethanol (\u2248150\u00a0\u00b0C) into the CVD system. The deposition took 30\u00a0min (with flowing H2 gas) and cooled down to room temperature in the Ar only environment. The median cooling rate of this furnace is 1.94\u00a0\u00b0C/min (see the detailed furnace setting in Table\u00a02\n). To further determine the effect of catalyst size onto the CNTs growth, the CVD process was repeated with samples M2 and M3 at the D = 12\u00a0cm, which showed the highest carbon yield. The abbreviations of the samples were listed in Table\u00a03\n.The crystal phases and compositions of the calcined powder and the sample were analysed with an X-ray diffractometer (XRD, PW 3040/60 MPD X'pert High Score Pro PANalytical, Philips, Almelo, Netherlands). The field-emission scanning electron microscope determined the morphology and element distribution of the samples (FESEM, NOVA NANOSEM 230, FEI, Oregon, United States of America) coped with energy-dispersive X-ray spectroscopy (EDX, Max 20, Oxford Instruments, Oxfordshire, England). The cross-section of synthesised CNTs was viewed under the transmission electron microscopy (TEM, JEM-ARM200F, JEOL Ltd., Tokyo, Japan), with 200.0\u00a0kV accelerating voltage and vacuum chamber pressure less than 2.5\u00a0\u00d7 10\u22125\u00a0Pa. Raman spectroscopy (NRS 3300, JASCO Inc, Japan) was employed to determine the quality of the synthesised carbon structure. The temperature profile of the tube furnace was measured using Infrared Thermometer (IR-750, Amprobe, Washington, United States of America).The computational fluid dynamics (CFD) within the tube furnace was performed using COMSOL Multiphysics 5.3a. The simulation uses creeping flow model instead of turbulence model. The stationary CFD equations were expressed as follows (see Eq.\u00a0(1) and Eq.\u00a0(2)).Continuity Equation of fluid flow [32]\n\n\n(1)\n\n\n\u03c1\n\u2207\n\u00b7\n\n(\nu\n)\n\n=\n0\n\n\n\nwhere \u03c1 and u are the density of gas and velocity vector.Momentum equation [33]\n\n\n(2)\n\n\n0\n=\n\u2207\n\u00b7\n\n[\n\n\u2212\np\nl\n+\n\u03bc\n(\n\u2207\nu\n+\n\n\n(\n\u2207\nu\n)\n\nT\n\n\n]\n\n+\nF\n+\n\u03c1\ng\n\n\n\nwhere p, l, T, \u03bc, F and g are the gas pressure, length, gas temperature, dynamic viscosity, applied body force and gravitational acceleration.\nFig.\u00a02 shows the FESEM micrographs of the oxidised and milled NiO powder, together with the size distribution. The M1 sample showed NiO particles with average size 98.99\u00a0\u00b1 42.79\u00a0nm. For sample M2, the particles most likely experienced agglomeration due to the generated heat during the milling. The kinetically generated thermal energy melted the NiO particles partially and caused them to fuse with neighbour particles. For sample M3, it showed that the 7-hour milling process caused damage and crack on the NiO particles and agglomeration.\nFig.\u00a03 showed\n\n the XRD spectra of the oxidised and milled powder. The calcined powder showed the bunsenite crystal structure (NiO, cubic \n\nR\n\n3\n\u00af\n\nm\n\n, ICSD code: 61318). The cubic structure from XRD phase identification agreed to the structure observed under FESEM. The chemical conversion from Ni(NO3)2 to NiO was shown in Eq.\u00a0(3).\n\n(3)\n\n\nN\ni\n\n\n(\nN\n\nO\n3\n\n)\n\n2\n\n\u00b7\n6\n\nH\n2\n\nO\n+\n\nC\n2\n\n\nH\n5\n\nO\nH\n\n\u27f6\n\n\u0394\n\n\n\nN\ni\nO\n+\n\nC\n2\n\n\nH\n4\n\n+\n7\n\nH\n2\n\nO\n+\n2\nN\n\nO\n2\n\n+\n\nO\n2\n\n\n\n\n\nFrom the XRD spectra, both samples shared an identical plane but different FWHM (see detailed in Table\u00a04\n). The crystallite size (\u03c4) of the sample was determined by Scherrer equation, as presented in Eq.\u00a0(4)\n[34], where k is the shape factor (0.9), \u03b2 is the full width half maxima in radian, \u03bb is the X-ray wavelength (1.542\u00a0\u00c5), and \u03b8 is the Bragg angle in\n\n degree. The M3\u2019s XRD spectra showed broader FWHM as the crystals were milled, and the resulted crystallite size were 45.36\u00a0nm (M1) and 27.05\u00a0nm (M3) [35].\n\n(4)\n\n\n\u03c4\n=\n\n\nk\n\u03bb\n\n\n\u03b2\ncos\n\u03b8\n\n\n\n\n\n\n\nFig.\u00a04 displayed the XRD spectra of the samples N1, N2 and N3. Nickel (Ni, cubic fm3m, ICSD code: 646090) and graphite (C, hexagonal \n\nP\n63\n/\nmmc\n\n, ICSD code: 76767) was detected in the samples. It agreed to the reduction of NiO to Ni with the presence of H2 at high temperatures. The graphite (002) plane was barely detected at 2\u03b8\u00a0=\u00a026 \u00b0 in the samples, and the FWHM values were narrowed down to 0.3819 \u00b0. Fig.\u00a04 also showed the graphite grain in sample N1 achieving the largest crystallite size yet lower intensity than the Ni phase (rel. int.\u00a0=\u00a00.46%). The flowing hydrocarbon (CxHx) crystallised rapidly once deposited on the Ni particles (for D = 2\u00a0cm), regardless of the number of deposited carbon atoms or the optimised precursor flowing path. Fig.\u00a05 displayed the FESEM and TEM micrographs of samples N1, N2 and N3, with the stated average diameter of MWCNTs. The TEM image revealed the size of Ni particles decreased after the H2 etching, which agreed to Sheng et\u00a0al. findings [36]. Worm-like nanostructure was identified as multi-walled CNTs with hollow structures. The external and internal diameters of the CNTs (dext and dint) were measured and further calculated for the average number of walls (Nwall) through Eq.\u00a0(5)\n[37], displayed in Table\u00a05. Ever since the diameter of the NiO catalyst is constant, the Nwall of the MWCNTs is in the range from 38 to 47 layers. This finding agreed to Ali et\u00a0al., where the dint of MWCNTs was about 5 \u2013 7\u00a0nm at the growth temperature between 600 and 700\u00a0\u00b0C [38].\n\n(5)\n\n\n\nN\n\nw\na\nl\nl\n\n\n=\n\n1\n\n0.34\nn\nm\n\n\n\n(\n\n\n\nd\n\ne\nx\nt\n\n\n\u2212\n\nd\nint\n\n\n2\n\n)\n\n+\n1\n\n\n\n\n\nFig.\u00a06 showed the XRD spectra of the as-synthesised samples with different catalyst sizes. Compared to the sample N1 \u2013 N3, the relative intensity of graphite (002) plane in the samples N4 and N5 are higher (N4: 10.60% and N5: 24.36%). The intensity increment of the (002) plane indirectly indicated that the carbon nucleation and crystallization on the NiO catalyst had increased. However, the calculated \u03c4graphite from XRD spectra is much lower than samples N1 \u2013 N3. It was attributed to the existence of the nano-sized amorphous carbon, forming a mixed state with the crystalline graphite phase.\nFig.\u00a07 showed the morphology, fast Fourier Transform (FFT) pattern, and the line profile of the samples N4 and N5. The FESEM micrograph showed that the diameter of the nanotubes is in the range of 82.29\u00a0nm. The FFT image (Fig.\u00a07a(iii)) showed the Ni diffraction, agreed with the XRD spectra of the hydrogen reduced NiO. Ni has the high oxidation rate (KP\u00a0=\u00a02.52\u00a0\u00d7 10\u221213 cm2/s, at 425\u00a0\u00b0C), meaning that the Ni oxidised easily at high temperature [39,40]. The graphite had encapsulated the Ni particles and prevent the particles from oxidizing. Subash et\u00a0al. presented a similar work on the carbon encapsulated metallic Li to reduce the rapid oxidation [41]. Also, the line profile (Fig.\u00a07a(iv)), representing the selected region in the TEM image, showed the interplanar spacing is 0.353\u00a0nm which agreed to the graphite (002) plane. In Table\u00a05, sample N4 achieved the highest number of the wall (116 carbon layers) due to the smallest size of Ni catalyst (338.11\u00a0nm). Sample N5 achieved the smallest tube diameter in the range of 28.33\u00a0nm, which may contribute to the NiO cracked structure. The NiO cracked structure (see Fig.\u00a02 M3) provided a smaller nucleation site during the carbon deposition, and it limited the graphite from rapid crystal growth. In the TEM micrograph (Fig.\u00a07b(ii)), a blurry region (iii) was found and reconstructed into an FFT pattern. The pattern showed the amorphous carbon overlapped the Ni catalyst, where the Ni catalyst appeared as a darker background in the TEM graph due to high density. The line profile (iv) also showed the spacing between the planes in the range 0.345\u00a0nm, which also agreed with the graphite plane.\nTable\u00a06 showed the carbon yield (or quantities) of the as-synthesised sample (N1 to N5). The carbon yield was calculated through Eq.\u00a0(6), where Mi and Mf are the catalyst mass and the final mass of samples. The flowing Ar and H2 gas carried the vaporised ethanol (or the carbon precursor) into the system. When the gases flowed into the quartz tube, the thermal energy supplied the gas with an upthrust force, yet the gravity slowed down the carbon precursor and limited the upthrust force. It was suggested that a flow path for the ethanol gas at 600\u00a0\u00b0C, i.e. an approximately 12\u00a0cm parabolic flying path at the current gas flow before it deposited on the Ni catalyst.\n\n(6)\n\n\nY\ni\ne\nl\nd\n=\n\n\n\nM\nf\n\n\u2212\n\nM\ni\n\n\n\nM\ni\n\n\n\u00d7\n100\n%\n\n\n\n\nThe quartz tube's real-time temperature profile and CFD simulation have been performed to support the statement mentioned earlier on the precursor-flowing path. Fig.\u00a08a showed the temperature profile along the quartz tube during the deposition. We segmented the 1-metre quartz tube into three regions: the region exposed to the outer environment (L0 \u2013 L1 and L13 \u2013 L14), the region with Al2O3 blocks (L1 \u2013 L3 and L11 \u2013 L13), and the heating region (L3 \u2013 L11). Position L4 \u2013 L6 were the carbon feedstock introduced into the CVD reactor respected to D = 12, 7, and 2\u00a0cm. L7 is the position of the NiO catalyst. From the graph, the temperature increased sharply in the heating region (from 243.7 to 587.1\u00a0\u00b0C) after the position L3 (Al2O3 thermal insulation blocks). The temperature of L4 \u2013 L6 (or D = 12 to 2\u00a0cm) were slightly lower than the centre of the heating zone (587.1\u00a0\u00b0C), resulting in 507.4\u00a0\u00b0C (L4), 543.2\u00a0\u00b0C (L5), and 565.0\u00a0\u00b0C (L6). Fig.\u00a08b showed the CFD simulation of carrier gas and carbon precursor within the tube at 600\u00a0\u00b0C. The gas entrance was estimated to achieve the maximum velocity, which was 4.165\u00a0\u00d7 10\u22122\u00a0cm/s, similar to the outlet. However, the gas that flew along the quartz tube was slower than the entrance (average velocity\u00a0=\u00a02.3\u00a0\u00d7 10\u22123\u00a0cm/s) due to the relatively larger diameter of the stainless steel tube and the boundary effect [44]. As a result, the gas flew slower for D = 12 and 7\u00a0cm (2.8\u00a0\u00d7 10\u22123\u00a0cm/s) before reaching the catalyst. Hence, the Ar/H2-carried carbon precursor was expected to have sufficient time to deposit on the Ni catalyst.\nFig.\u00a09\n showed the Raman spectra of the samples. Sample N5 achieved relatively better quality in terms of the intensity ratio between D-band and G-band (ID/IG\u00a0=\u00a01.069), where D-band and G-band of the carbon layers located at 1349 and 1601 cm\u22121, respectively. The defect on the graphitic layers can be caused by H2 etching, which increases the ID/IG ratio [42]. The crystallite size of the synthesised carbon structure was calculated via Eq.\u00a0(7), where \u03bb is the laser wavelength during the analysis (532.08\u00a0nm) [43].\n\n(7)\n\n\n\nL\na\n\n=\n\n(\n2.4\n\u00d7\n\n10\n\n\u2212\n10\n\n\n)\n\n\n\u03bb\n4\n\n\n\n(\n\n\nI\nD\n\n\nI\nG\n\n\n)\n\n\n\u2212\n1\n\n\n\n\n\n\n\nFig.\u00a010\n displayed the element distribution and the crystallite size of the samples. It showed that the carbon deposition for samples N3, N4 and N5 is higher than 90%, according to the carbon yield in Table\u00a06. Furthermore, the ID/IG-derived crystalline size increased with the milling time. This phenomenon is due to the fractured surface of the Ni particles lowered the activation energy by introducing a larger surface area. By considering the balance of quality and quantity of the as-synthesised MWCNTs in this work, the optimized precursor path is 12\u00a0cm and the milling time for the NiO catalyst was 7\u00a0h.The growth of CNTs on the NiO particles by CVD were demonstrated with ethanol as the carbon precursor. We identified two factors affecting the CNTs' growth by performing several parameter changes during CVD: the precursor flowing path and catalyst size. The precursor-flowing path (D = 12\u00a0cm) determined the highest carbon yield with the fixed gas flow rate. We suggested the precursor flowing path is a function of furnace temperature and gas flow rate. Also, the size of the synthesised NiO fluctuated as the milling time increased. The 7-h milled NiO particles resulted in a more fractured site, and it provided a smaller nucleation site and larger surface area for carbon deposition. Smaller nucleation sites catalysed the crystallisation of the CNTs. In this study, the optimal deposition conditions for MWCNTs growth on NiO particles were as follows: 12\u00a0cm distance between catalyst and precursor and 7\u00a0h milled NiO. Through the optimization of the parameter, the carbon yield of as-synthesised MWCNTs reached >50% and the ID/IG ratio is 1.069. Thus, our findings are noteworthy because they provide more information to researchers working to improve fabrication procedures to mass-produce high-quality MWCNTs.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This study was supported by Malaysian Ministry of Education (MoE) through Fundamental Research Grant Scheme (FRGS/1/2018/STG07/ UPM/02/3) No. 5540132 and Nanotechnology Platform Program <Molecule and Material Synthesis (S-19-NI-0041)> of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.", "descript": "\n                  Carbon nanomaterials have been found to have promising performance in various applications. However, the complexity and high operation cost during the fabrication still limit the mass production. In this study, multi-walled carbon nanotubes (MWCNTs) were grown on nickel oxide (NiO) via chemical vapour deposition (CVD) with ethanol as the carbon precursor. The NiO catalyst was fabricated from a nickel nitrate \u2013 ethanol mixture. The particle size of NiO was altered through high-energy ball milling for 0, 4, and 7\u00a0h. The influence of precursor flowing path (D) and NiO catalyst size during the MWCNTs growth have been investigated. The Raman spectra showed that the crystallite size of MWCNTs (La) increased from 16.97 to 18.00 nm as the NiO milling time increased. Furthermore, NiO-catalysed MWCNTs at D = 12\u00a0cm achieved the highest carbon yield (80.54%), with an ID/IG ratio of 1.134. Also, SEM and TEM revealed that the larger size of NiO catalyst produced fewer layers of MWCNTs. These findings are significant to aid researchers and manufacturers in optimising the CVD process towards large-scale MWCNTs fabrication.\n               "}
{"full_text": "Data will be made available on request.Solid oxide fuel cells (SOFCs) are considered as one of the most efficient electrochemical energy-conversion devices due to their high energy efficiency, low pollutant emission and high flexibility to utilize various fuels such as hydrogen, syngas and hydrocarbons [1\u20133]. Readily available hydrocarbon fuels with a low cost and good security are considered as promising alternatives to conventional H2 fuel for fuel cells [4]. Furthermore, liquid renewable hydrocarbon fuels such as methanol has high volume energy densities (1.6\u00a0\u00d7\u00a0104\u00a0kJ\u00a0m\u22123), which is beneficial to fuel storage and the mobile applications at ambient pressure and temperature. Moreover, unlike other heavier alcohols like ethanol with the CC bond in the molecular structures, the cleavage of CH bonds is much easier for methanol through methanol thermal decomposition or steam reforming of methanol [5\u20137]. However, the utilization of hydrocarbon fuels in SOFC is hindered by serious carbon deposition on anode catalyst and the sluggish anode kinetics [8].Ni-based cermet with a high electrical conductivity is the most widely used anode material in SOFC research. Besides external and internal reforming, which first transforms the hydrocarbon feed to syngas, much effort has been focused on enhancing the activity and the resistance to coking of the Ni-based anode materials for the direct electrochemical conversion of hydrocarbon fuels. Cu with a high coking resistance has been used to replace Ni fully or partially, which leads to a low catalytic activity [9]. Yoon and Manthiram [10] found that the incorporation of 1 atom% W to Ni brings in surface hydroxyl groups through the reaction with water vapor, which facilitates the oxidation of carbon deposited. The addition of BaO and NbOx has similar effects [11\u201313]. The alloying of Ni with other metals, such as Co [14,15], Mo [16] and Fe [17,18], improves the anode activity and suppresses carbon deposition. Among the possible alternative alloys, Ni-Sn is potentially an excellent candidate for the anode catalyst [19,20]. The computation models suggest that Ni-Sn is more carbon-tolerant than Ni [21]. However, Li et al. [22] found that the addition of Sn in Ni decreases the anode activity slightly while improving the coking resistance remarkably with CH4 as fuel. Notably, intermetallic compounds (IMCs) gradually attracted more attention because they exhibit unique catalytic properties. Cabot et al. [23] prepared NiSn NPs with controlled stoichiometry and achieved excellent performance towards methanol oxidation reaction, meanwhile, significantly improved stability compared to single metal nickel. However, the effect of NiSn IMCs, as active sites on the anode of methanol fueled SOFC, on the catalytic performance of the anode has not been reported.The recent research on cermet anodes mainly focuses on the metal catalysts. Nevertheless, the ceramic support providing oxygen ions also have remarkable influence on the activity and coking resistance of the cermet anodes with hydrocarbon fuels [24,25]. The activity of a Cu-based anode with doped ceria as the support is much higher than that with yttria stabilized zirconia (YSZ) as support when CH4 is used as the fuel, attributing to the high oxygen storage capacity (OSC) of ceria [9]. The activity and coking resistance of cermet anode with Ce0.8Sm0.2O1.9 (SDC) as support are enhanced when Sm is partially substituted with other rare earth metals such as La, Pr and Nd due to the improved activity of surface oxygen species [26,27]. Among the very few examples in the literature, Sn doped CeO2 was used as a support in CO oxidation [28,29]. Sn improves the reducibility and OSC of CeO2, bringing about a high catalytic activity. However, Sn doped SDC has not yet been tested as anode material in SOFC.In this work, Sn doped SDC (Ce0.8\u2212xSnxSm0.2O2\u2212\u03b4, x \u00a0=\u00a00\u20130.15, SSnDC) is examined as the support in Ni-based cermet anode. We found that excessive Sn exsolves partially from the oxide phase and forms intermetallic compounds with Ni after reduction. The effects of Sn in both metal and the oxide phases are investigated. The dual-modified Ni-SDC anode material with Sn doping shows enhanced performance and stability with methanol as the fuel.SDC, Ce0.75Sn0.05Sm0.2O2\u2212\u03b4 (SSn5DC), Ce0.70Sn0.10Sm0.2O2\u2212\u03b4 (SSn10DC) and Ce0.65Sn0.15Sm0.2O2\u2212\u03b4 (SSn15DC) powders were synthesized via a hydrothermal procedure [30,31]. All of the chemicals (A.R. in purity) were purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd. Stoichiometric Ce(NO3)3\u00b76H2O, SnCl4\u00b75H2O and Sm(NO3)3\u00b76H2O were dissolved in deionized water to form a solution with a total metal ion concentration of 0.15\u00a0mol\u00a0L\u22121. Urea was subsequently added to the solution under constant stirring to reach a concentration of 1\u00a0mol\u00a0L\u22121. Then the precursor solution was transferred into a hydrothermal reactor and kept at 140\u00a0\u00b0C for 5\u00a0h. The precipitate was washed with deionized water until no Cl- was detected with 0.1\u00a0mol\u00a0L\u22121 AgNO3 solution. The powder obtained was dried at 100\u00a0\u00b0C for 12\u00a0h and finally calcined in air at 700\u00a0\u00b0C for 2\u00a0h.NiO-SSnxDC (x\u00a0=\u00a00, 5, 10 and 15) powders with a Ni loading amount of 10\u00a0wt% were prepared through an incipient wetness impregnation technique with a Ni(NO3)2\u00b76H2O aqueous solution. Then the powders were dried at room temperature overnight and calcined subsequently at 700\u00a0\u00b0C in air for 2\u00a0h. The sample was reduced in pure H2 at 700\u00a0\u00b0C for 2\u00a0h for characterization and are marked as Ni-SSnxDC. For comparison, a sample, denoted as Sn@Ni-SDC, was prepared as follows. Sn, equivalent to 50\u00a0wt% Sn in Ni-SSn10DC, was added into the NiO-SDC powder via impregnation with an aqueous solution of SnCl4\u00b75H2O.A D8 Focus diffractometer (Bruker Corp., Cu K\u03b1 radiation, 40\u00a0kV and 200\u00a0mA) X-ray was used to record the diffraction (XRD) patterns at a scanning rate of 1\u00b0\u00a0min\u22121. The microstructure of the samples was observed using a transmission electron microscope (TEM, JEM-200F, JEOL Inc., Japan). High-angle annular dark-field scanning TEM (HAADF-STEM) and energy-dispersive X-ray (EDX) analysis were carried out to mapping the element distributions. The surface property of the samples was analyzed with an ESCALAB 250 Xi X-ray photoelectron spectrometer (XPS, K-Alpha+, Thermo Fisher Scientific) using Al-K\u03b1 (h\u03bd\u00a0=\u00a01486.6\u00a0eV) as the X-ray source. The spectra obtained were referenced to the C 1\u00a0s binding energy (284.8\u00a0eV).The activity of oxygen species in the samples was evaluated with CH3OH temperature-programmed surface reaction (CH3OH-TPSR) in a quartz tube reactor. 80\u00a0mg of powder was reduced in hydrogen at 700\u00a0\u00b0C for 2\u00a0h. After cooling to 150\u00a0\u00b0C, the sample was treated in pure O2 (30\u00a0mL\u00a0min\u22121) for 30\u00a0min and was purged subsequently with Ar (30\u00a0mL\u00a0min\u22121) at the same temperature for another 30\u00a0min to sweep the weakly absorbed oxygen. Then about 10\u00a0vol% gasified CH3OH was added in Ar by bubbling Ar through liquid CH3OH at room temperature. The reactor was heated from 150 to 800\u00a0\u00b0C at a heating rate of 5\u00a0\u00b0C\u00a0min\u22121, and the oxidation of CH3OH was monitored with measuring the amount of CO2 produced with an online mass spectrometer (HPR20, Hiden Analytic Ltd.). The reduced anode powder was treated with gasified anhydrous methanol (150\u00a0mL\u00a0min\u22121, STP) at 700\u00a0\u00b0C for 4\u00a0h. The carbon deposited was observed with a scanning electron microscope (SEM, S-4800, Hitachi, Japan), and its amount was measured with thermogravimetric analysis (TGA) in oxygen atmosphere using a thermal analyser (NETZSCH STA449, Germany).70\u00a0wt% SDC and 30\u00a0wt% (Li0.67Na0.33)2CO3 composite was used as the electrolyte material [32], which was uniaxially pressed into disk-shaped pellets at 500\u00a0MPa and then sintered at 700\u00a0\u00b0C for 2\u00a0h. The diameter and the thickness of the electrolyte layer were 13\u00a0mm and 500\u00a0\u03bcm, respectively. The anode powder was mixed with an organic binder (V006, Heraeus Ltd.) to make a slurry, which was screen-printed on both sides of the electrolyte pellets, and subsequently calcined at 700\u00a0\u00b0C for 2\u00a0h to form symmetric cells. The geometrical area of the electrode was 0.64\u00a0cm2, and Ag paste was used as the current collector. The electrode layers were reduced in H2 at 700\u00a0\u00b0C for 2\u00a0h, and then the electrochemical impedance spectra (EIS) of the symmetric cell were recorded with an electrochemical workstation (VERSASTAT 3, Ametek) under various H2 partial pressures (with Ar as the balance gas).Electrolyte-supported single cells were fabricated via a similar screen-printing process [18]. The cathode consisted of 70\u00a0wt% composite electrolyte and 30\u00a0wt% lithiated NiO, which was screen-printed on one side of the electrolyte pellet. The anode slurry was printed on the other side of the pellet, followed by calcined at 700\u00a0\u00b0C for 2\u00a0h to form single cells. The anode was reduced in H2 at 700\u00a0\u00b0C for 2\u00a0h, and then the performance of the cell was measured using the electrochemical workstation with dry hydrogen and gasified anhydrous methanol (100\u00a0mL\u00a0min\u22121, STP) as the fuels and O2 (30\u00a0mL\u00a0min\u22121, STP) as the oxidant.The I-V and I-P characteristics of the single cells with dry H2 as the fuel at 700\u00a0\u00b0C are presented in Fig. 1\na. The open circuit voltages (OCV) of the cells are between 1.00 and 1.12\u00a0V. The maximum power density (P\nmax) of the cells with Ni-SDC, Ni-SSn5DC, Ni-SSn10DC and Ni-SSn15DC anodes are 0.53, 1.15, 1.93 and 1.62\u00a0W\u00a0cm\u22122, respectively. When the temperature drops to 650, 600 and 550\u00a0\u00b0C, the P\nmax of the cell with the Ni-SSn10DC anode decreases to 1.56, 0.97 and 0.61\u00a0W\u00a0cm\u22122, respectively (Fig. S1a). When methanol is used as the fuel, the OCVs are in the range of 1.00\u20131.08\u00a0V (Fig. 1b), and the cell with the Ni-SSn10DC anode shows the highest P\nmax of 2.11\u00a0W\u00a0cm\u22122, much higher than that of the cell with the Ni-SDC anode (0.78\u00a0W\u00a0cm\u22122). When the temperature drops to 650, 600 and 550\u00a0\u00b0C, the P\nmax of the cell with the Ni-SSn10DC anode decreases to 1.74, 1.18 and 0.83\u00a0W\u00a0cm\u22122, respectively (Fig. 1c), which is similar to the trend of P\nmax obtained at different temperatures with dry H2 as the fuel. The corresponding impedance spectra at the open circuit condition are presented in Fig. S1b, c. The good performance of the cell in the intermediate temperature range is attributed to the high catalytic activity and the low activation energies (E\na) of the Ni-SSn10DC anode. Fig. 1d shows the short-term stability of the single cells with methanol as the fuel at a constant current density of 0.2\u00a0A\u00a0cm\u22122 at 700\u00a0\u00b0C. The cell with the Ni-SDC anode exhibits a steady output voltage in the first 4\u00a0h, which then drops gradually probably due to the coking on the anode. On the contrary, the output voltage of the cell with the Ni-SSn10DC anode is stable for more than 12\u00a0h attributed to the improvement of the resistance to carbon deposition of the anode with Sn doping.The EIS curves of the symmetric cells with H2 and methanol at both sides at 700\u00a0\u00b0C are presented in Fig. 1e and Fig. S1d. The data are fitted with the equivalent circuit R\n0(R\n1\nQ\n1)(R\n2\nQ\n2), in which R\n0, R\n1 and R\n2 are resistances, while Q\n1 and Q\n2 are constant phase elements. The ohmic resistances of the cells shown as the intercepts of the Nyquist curves on the real axis in high frequency region are similar. The anode polarization resistances (R\np) reflected by the arcs show an order of Ni-SDC\u00a0>\u00a0Ni-SSn5DC\u00a0>\u00a0Ni-SSn15DC\u00a0>\u00a0Ni-SSn10DC.The Bode plots of the cells are shown in Fig. 1f, which can be roughly devided into a high frequency (HF, 104\u2013101\u00a0Hz) region and a low frequency (LF, 101\u201310\u22122\u00a0Hz) region. The R\np in the LF region decreases remarkably with the doping of Sn in the anode support. Ni-SDC, Ni-SSn5DC and Ni-SSn10DC anodes exhibit similar R\np in the HF region, while that of the Ni-SSn15DC anode is slightly larger.The EIS results of the symmetric cells under various hydrogen partial pressures (\n\n\nP\n\nH\n2\n\n\n\n) are plotted in Fig. S1e, in which the ohmic resistances are deduced for a better comparison. The R\np of all the anodes increase with the decrease of \n\n\nP\n\nH\n2\n\n\n\n, and linear relationships between LogR\np and Log \n\n\nP\n\nH\n2\n\n\n\n are observed (Fig. 1g). The slope values for Ni-SSnxDC are 0.24\u00a0\u223c\u00a00.31.The Arrhenius plots of the R\np of the anodes are presented in Fig. 1h. The E\na of the electrochemical oxidation of H2 on Ni-SDC and Ni-SSn5DC anodes are 0.61\u20130.65\u00a0eV. The Ni-SSn10DC anode shows the lowest E\na of 0.32\u00a0eV attributed to the acceleration of the surface steps, while the E\na of Ni-SSn15DC increases to 0.43\u00a0eV due to the suppression of the bulk conduction of O2\u2013.The cross-sectional SEM images of the single cell with Ni-SSn10DC anode before test are shown in Fig. 2\na and Fig. S2. The thicknesses of the anode, the electrolyte and the cathode layers are about 30, 500 and 40\u00a0\u03bcm, respectively. The anode surface exhibits a fine and uniform porous microstructure (Fig. 2b).The hydrothermally synthesized SDC powder shows a face-centered cubic fluorite structure (JCPDS#075\u20130158, Fig. S3a). With the partial substitution of Sn for Ce, the fluorite structure is maintained. Meanwhile, the XRD peaks shift gradually to higher angles, indicating the contraction of SDC lattice with the incorporation of Sn since Sn4+ (0.81\u00a0\u00c5) is smaller than Ce4+ (0.97\u00a0\u00c5) [33,34]. Furthermore, the characteristic peaks of SnO2 (JCPDS#041\u20131445) is observed in the XRD pattern of SSn15DC, implying a limited solubility of Sn in SDC [35,36]. The XRD results without obvious SnO2 peaks probably are caused by the low amount of SnO2 phase and its uniform distribution on the support. SDC and Ni (JCPDS#087\u20130712) phases are found in Ni-SDC after reduction (Fig. 3\na). With the addition of Sn in the SDC phase, the Ni peaks are weakened remarkably, while Ni3Sn and Ni3Sn2 phases are observed, indicating the partial exsolution of Sn and the formation of intermetallic compounds during the reduction. Ni-SSn10DC has the strongest Ni3Sn peaks, while more Ni3Sn2 phase is found in Ni-SSn15DC. Ni3Sn and Ni3Sn2 phases are also formed in the Sn@Ni-SDC after reduction. Rietveld refinement of XRD data for the Ni-SSnxDC anode is carried out and the results are shown in Fig. 3b and Fig. S2c\u2013f. Fig. S4 depicts the unit cell structure of SSnxDC, with a cubic fluorite structure, drawn using the VESTA software. All materials have a tetragonal structure (space group I4/mmm). The lattice parameters of Ni-SSn10DC are a\u00a0=\u00a0b\u00a0=\u00a08.655\u00a0\u00c5 and c\u00a0=\u00a05.500\u00a0\u00c5. These parameters are slightly smaller than those of Ni-SDC, which were a\u00a0=\u00a0b\u00a0=\u00a08.752\u00a0\u00c5 and c\u00a0=\u00a05.536\u00a0\u00c5. These refinement results demonstrates the lattice volume shrinkage after Sn doping.The TEM micrograph of the NiO-SSn10DC anode powder before reduction is shown in Fig. 3c, revealing the nanoparticles with the size of 10\u201320\u00a0nm. The lattice fringes corresponding to NiO, SDC and SnO2 are observed in the HRTEM image (Fig. 3d). After reduction, the nanoparticles of Ni, Ni3Sn and Ni3Sn2 are formed (Fig. 3e\u2013g). The HAADF-STEM image and the corresponding EDX elemental mappings of the reduced Ni-SSn10DC anode is shown in Fig. 3h. Ni exists in all the nanoparticles, while Ce is found only in the SDC phase. On the contrary, Sn is distributed in the whole anode, further proving the partial exsolution of Sn from SDC phase and the formation of Ni-Sn intermetallic compounds. The TEM results are consistent with the XRD characterization. The structure of the Ni-SSnxDC anode material is illustrated in Fig. 3i.The surface chemical states of the anode samples were further investigated with a XPS technique. Fig. 4\n shows the XPS survey spectra and the corresponding fitting curves of the Ni 2p and Sn 3d spectra after reduction. The deconvoluted peaks at about 852.3 and 854.0\u00a0eV are attributed to Ni0 and Ni2+, respectively (Fig. 4a) [37,38]. Ni2+ is probably formed from the quick reoxidation of metallic Ni on the surface of the reduced anodes before the XPS test [39,40]. Meanwhile, the binding energy of Ni0 decreases slightly with the addition of Sn, indicating the electron transfer from Sn to Ni in Ni3Sn and Ni3Sn2\n[41,42]. The Sn 3d spectrum of the Ni-SSnxDC in Fig. 4b could be deconvoluted into Sn0, Sn2+ and Sn4+components, revealing the coexistence of Sn, Sn2+ and Sn4+ species on the surface of Ni-SSnxDC.The CH3OH-TPSR results of the SSnxDC composite oxide powders are presented in Fig. 5\na. The weak CO2 peaks at about 300\u00a0\u00b0C are due to the reaction between CH3OH and oxygen species adsorbed weakly on the surface of the samples, while the strong peaks in 400\u2013700\u00a0\u00b0C correspond to the oxidation of CH3OH by lattice oxygens [26]. The oxidation of CH3OH on SDC reaches the highest rate at about 666\u00a0\u00b0C. With the increase of Sn content in SDC, the oxidation temperature of CH3OH decreases gradually, indicating that the partial substitution of Sn for Ce enhances the activity of lattice oxygen in SDC. The CH3OH-TPSR curves of the Ni-SSnxDC anode powders after reduction are shown in Fig. 5b. The oxidation temperature of CH3OH on Ni-SDC (546\u00a0\u00b0C) is much lower than that on SDC, and the impregnation of Sn (Sn@Ni-SDC) further decreases the oxidation temperature significantly, demonstrating that the formation of Ni-Sn intermetallic compounds improves the catalytic activity towards CH3OH oxidation. These results suggest that the dual-modified Ni-SDC with Sn are more favorable to the CH3OH oxidation process.The SEM images of the anode powders after carbon deposition are shown in Fig. 6\n. The surface of Ni-SDC is closely packed by filamentous carbon (Fig. 6a), while negligible carbon is found on Ni-SSn10DC (Fig. 6b). The TGA curves of the samples in the oxygen atmosphere are presented in Fig. 6c. The weight loss in 400\u2013650\u00a0\u00b0C reflects the oxidation of carbon deposits. The weight losses of Ni-SDC, Ni-SSn5DC, Ni-SSn10DC and Ni-SSn15DC are 27, 4.5, 2.4 and 1.8\u00a0wt%, respectively. The addition of Sn results in the enhancement of oxygen activity of the support and the formation of Ni3Sn and Ni3Sn2, both bring about the remarkable improvement of the resistance to carbon deposition [43\u201345].The performance of the single cell, with a P\nmax of 2.11\u00a0W\u00a0cm\u22122 at 700\u00a0\u00b0C, is a significant improvement on previously reported cells operated below 800\u00a0\u00b0C [18,22,42\u201345]. It also exhibits superior operational stability when compared with similarly structured SOFCs in the literature [15,46], [47]. Previously reported data can be categorized based on two different strategies: Ni-alloy anodes and oxide doping to the anode support. Our strategy involves coupling the NiSn intermetallic compounds and Sn doped SDC as a composite anode. This composite approach is responsible for the excellent performance. This approach provides an effective example for high-performance hydrocarbon-fueled SOFCs, especially under intermediate operating temperature conditions.To design more efficient Ni-based SOFC anodes, it is essential to identify the underlying rate-limiting step of the fuel conversion process. The CH3OH-TPSR results demonstrate that the formation of Ni-Sn intermetallic compounds improves the catalytic activity towards CH3OH oxidation. However, the reaction mechanism of methanol fuel at SOFC anode is complicated, which involves many reactions such as CH3OH decomposition, partial/full oxidative reforming, steam reforming, dry-reforming, and electro-oxidation reactions [5]. Therefore, it is diffiticult to clarify the anode processes with CH3OH as fuel. Notably, the performance of the single cells exhibit a consistent trend for various anodes when fueled with CH3OH and H2. Therefore, a symmetric cell was examined in detail and the anode reaction mechanism are discussed with using H2 as the fuel.The electrochemical oxidation of H2 at the anode of SOFC starts from the dissociative adsorption of H2 on the active sites (step (1)), followed by the surface diffusion of the adsorbed H to the reaction site, i.e., the three-phase boundary (TPB, step (2)). Meanwhile, oxygen ions also transfer to TPB in the anode through the ceramic phase (step (3)), which then reacts with the adsorbed H, forming H2O and releasing electrons (step (4)) [18].\n\n(step 1)\n\n\n\nH\n2,g\n\n\u2194\n\n2H\nad\n\n\n\n\n\n\n\n(step 2)\n\n\n\nH\nad\n\n\u2194\n\nH\nTPB\n\n\n\n\n\n\n\n(step 3)\n\n\n\nO\n\nO,bulk\n\nx\n\n\n\n\n+ V\n\n\n\nO, TPB\n\n\n\n\n\u00b7\n\u00b7\n\n\n\u2194\n\nO\n\nO,TPB\n\nx\n\n\n\n\n+ V\n\n\n\nO, bulk\n\n\n\n\u00b7\n\u00b7\n\n\n\n\n\n\n\n\n(step 4)\n\n\n\n2H\nTPB\n\n\n\n\n+ O\n\n\n\nO,TPB\n\n\n\nx\n\n\u2194\n\nV\n\n\nO, TPB\n\n\n\n\u00b7\n\u00b7\n\n\n\n\n\n+ H\n\n2\n\n\nO\nTPB\n\n\n\n\n+ 2e\n\n\n\n-\n\n\n\n\n\n\n\nThe slope of the fitting straight line should be close to \u22121 if step (1) is the rate-determining step (RDS) in the anode steps, which will change to \u22120.5 when the RDS is step (2). The R\np will not change with \n\n\nP\n\nH\n2\n\n\n\n if step (3) or (4) is the RDS. The Ni-SDC anode exhibits the lowest slope of \u22120.31, and the R\np is more prominent in the LF region of the EIS result (Fig. 1f), implying that the rate of the anode reaction could be codetermined by the surface diffusion of the adsorbed H (step (2) and the surface reaction between lattice oxygen and the adsorbed H (step (4).The electron cloud density of Ni increases with the addition of Sn (Fig. 5a), which may weaken the adsorbing strength of H and accelerate the surface diffusion of H [18], resulting in the decrease of the R\np in the LF region (Fig. 1f) and the increase of the slope (Fig. 1g). Meanwhile, the substitution of Sn for Ce in SDC improves the activity of lattice oxygen (Fig. 5a) and thus accelerates step (4). On the other hand, the doping of Sn may suppress the conduction of O2\u2013 in SDC (step (3), leading to a higher R\np in the HF region (Ni-SSn15DC, Fig. 1f) and the further increase of the slope (Fig. 1g).Based on above analysis, Sn doped Ni-Ce0.8Sm0.2O2\u2212\u03b4 anode clearly displayed improved performance and stability over the Ni-based anode. This fact has two different explanations: (1) The presence of Sn atoms within the Ni structure, forming Ni3Sn and Ni3Sn2 phase, certainly modifies the electronic density of Ni states, thus affecting its chemistry, which accelerates the surface diffusion of H. (2) The substitution of Sn for Ce in SDC improves the activity of lattice oxygen, which react with the adsorbed H, forming H2O and releasing electrons. Both the effects of Sn doping in metal and in oxide phases improve the carbon resistance.SSnxDC is hydrothermally synthesized and investigated as a catalyst precursor of a Ni-based cermet anode. Sn exsolves partially from the ceramic phase after reduction, and Ni3Sn and Ni3Sn2 intermetallic compounds are formed, in which the electrons transfer from Sn to Ni, weakening the adsorbing strength and facilitating the diffusion of H species on the surface of the anode. Meanwhile, the activity of lattice oxygen in the SDC phase, as the support in the composite anode, is also improved with the doping of Sn.A cell with the Ni-SSn10DC anode yields record high P\nmax values, e.g., 1.99 and 2.11\u00a0W\u00a0cm\u22122 at 700\u00a0\u00b0C with H2 and methanol as fuels, respectively. This remarkable performance is superior to the ever-reported Ni-based anodes. Meanwhile, it has been confirmed that the dual modified Ni-SSnxDC anode are highly resistant to carbon deposition. Such a strategy of anode design may have the great potential of application in the anode material design for non-hydrogen fueled SOFCs, which encounters the great challenge of carbon deposition when the temperature lowers to around 500\u00a0\u00b0C, around which selective and stable metal catalyst may be feasible for long term stability.The surface diffusion of the adsorbed H and the reaction between lattice oxygen and the surface H species are the probable RDS of the anode process, both of which are accelerated with the incorporation of Sn in both the support and metal phases. Ni-SSn10DC anode shows the lowest Ea of 0.32\u00a0eV for the electrochemical oxidation of H2.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The financial support from the Program of Introducing Talents to the University Disciplines under file number B06006 and the support of the Program for Changjiang Scholars and Innovative Research Teams in Universities under file number IRT 0641 are gratefully acknowledged. The work has been also supported by the Start-up Fund of Suzhou University of Science and Technology.Supplementary data to this article can be found online at https://doi.org/10.1016/j.cej.2022.140692.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  A crucial challenge in the commercialization of Ni-based materials as the anode of solid oxide fuel cell is the fast voltage drop due to carbon deposition and structural degradation during cell operation. Herein, Sn-doped Ce0.8Sm0.2O2\u2212\u03b4 (SDC) supported Sn-Ni alloy anode is rationally designed and prepared, via a simple and convenient dual-modification strategy. The substitution of Sn of Ce in the oxide phase enhances the mobility of lattice oxygen in SDC. Meanwhile, Sn exsolves partially from the oxide phase and forms Ni3Sn and Ni3Sn2 intermetallic compounds with Ni after reduction. The composite anode thus formed achieves unprecedent activity in the electrochemical oxidation of H2 and CH3OH. The maximum power densities of a cell supported by 500\u00a0\u03bcm-thick Ce0.8Sm0.2O2\u2212\u03b4-carbonate electrolyte layer with the Ni-Ce0.7Sn0.1Sm0.2O2\u2212\u03b4 (Ni-SSn10DC) anode reach 1.99 and 2.11\u00a0W\u00a0cm\u22122 at 700\u00a0\u00b0C, respectively for using H2 and methanol as fuels. The doping of Sn also remarkably enhances the coking resistance of the anode. This work opens a path on the design of high-performance SOFC anode.\n               "}
{"full_text": "Volatile organic compounds (VOCs) as major atmospheric pollutants are typically defined as compounds with boiling points lesser than 260 \u00b0C and greater than 50 \u00b0C at atmospheric pressure by WHO. These compounds usually originated from indoor and outdoor sources. The emissions from outdoor are mainly composed of various industries, such as coal chemistry, organic chemicals, petrochemicals, painting, solvent, dyeing, and so on. And most indoor come from organic solvents for interior decoration materials, adhesives, cosmetics, cleaning agents, household products, etc. It has been widely recognized that VOCs are highly harmful to human health due to their toxic, carcinogenic, mutagenic and teratogenic effect. Besides, the VOCs are also highly contributed to the formation of ozone, photochemical smog and secondary aerosols, which are destructive to the environment. Thus, the restriction of VOCs emission has become one of the most important tasks for the protection of human health and environment, since the rapid urbanization and industrialization recently lead to the growth of VOCs pollutions. It is of great significance and urgency to employing efficient technology to reduce VOCs emissions.Currently, the existing manifold technologies can be mainly divided into two categories: adsorption (condensation, adsorption, absorption and membrane separation) and destruction (incineration, catalytic oxidation, photocatalytic degradation, and plasma technology). There are plenty of limitations to adsorption techniques. For example, the condensation and absorption efficiencies are mainly restrained by the concentration of VOCs, as well as the adsorption technology is restricted by the limitation of adsorption adsorbent and the potential risk of secondary contamination. Owing to the advantage of simple operation, the membrane separation exhibited good potential, but it was still limited by its expensive cost and maintenance charges. Adversely, the destruction technologies have wider applicability. Taking into removal efficiency, energy-saving and environment-friendly, catalytic oxidation seems to be an optimal strategy among these technologies because the VOCs can be fully degraded into CO2 and H2O at low temperatures. Therefore, the fabrication of suitable catalysts for VOCs elimination is crucial, but still a big challenge.Recently, numerous investigations have been considered on the development of catalysts for VOCs oxidation and these catalysts can be divided into noble-metal catalysts and transition-metal oxides. In general, the noble metal catalyst is an optimal choice for VOCs destruction due to its superior activity, but the nature of scarcity and poor stability still limit its further applications. On the other hand, owing to the abundant reserves and superior stability, more and more transition-metal oxides (including oxides of Co, Mn, Fe, Cu, Ni) are investigated to replace noble metal catalysts for VOCs purification. Although the catalytic activity of transition-metal oxides is not as good as noble metal catalysts so far, it has drawn increasing attention and deemed to be dominant in the future. Among these transitional metals, cobalt-based (Co-based) oxides usually exhibit promising performance in air purification reactions due to their excellent redox property, high activity and stability, which is one of the most promising candidates for VOCs oxidation.To date, lots of studies have been reported to optimize the catalytic activity of Co-based catalysts, and breakthroughs have been made for catalytic removal of VOCs at low-temperature (as low as room temperature for some VOC). Thus, based on recent researches, we prefer to give a comprehensive review to summarize Co-based oxides with various compositions for VOCs removal to offer a logical and systematic picture of this area. To be specific, single cobalt oxides involving different morphologies, crystal phase, structures and valence; cobalt-based compounds such as binary composites, multiple composites; common and unusual strategies to improve catalytic performance including acid treatment and doping strategy, have been coherently reviewed. Subsequently, the detailed mechanism of relevant catalysts is also summarized. Finally, considering the enormous challenges and opportunities in this field, we make some perspectives for future development directions for VOCs oxidation. We believe that this review will be instructive to establish a reference for catalyst design for the removal of VOCs.Cobalt oxides (CoOx) are one of the most active low-cost transition metal oxides, which were widely applied in the field of air purification. The superior activity of CoOx can be attributed to the oxidation states, various crystal phases, weak CoO bond strength, numerous structures and morphologies. Besides, the physicochemical property of CoOx can be also modified by the engineering of crystal facets, resulting in different catalytic performances. A brief summarize of the effect of these factors on catalytic performance was displayed in Table\u00a01\n.Owing to the multivalent oxidation state of cobalt species (Co2+, Co3+), the CoOx can be divided into various types, such as CoO, Co2O3, Co3O4. Simultaneously, the catalytic performance of VOCs oxidation over CoOx is mainly influenced by the adsorption, activation and breaking the CH bond, which is strongly related to the oxidation states of cobalt ions. For example, Zhao and Ye [10] reported Co3O4 nanosheets with rich Co3+sites exhibit excellent catalytic performance for toluene oxidation, which is even comparable to the state-of-the-art noble metal catalyst. And the enhanced catalytic activity can be assigned to the activation of CH bond by Co3+ sites. Similarly, Zhong and Ye [1] prepared Co3O4 catalyst with hierarchical morphology via alkali etching technology and explore its catalytic activity for toluene combustion, which also confirms the high low-temperature activity and selectivity are contributed by Co3+species. On the contrary, other researchers have found a good relationship between the catalytic performance and concentration of surface Co2+species during VOCs oxidation (propane, o-xylene, benzene) [11\u201313]. Thus, it seems that the influence of the valence state of different CoOx on catalytic performance is highly related to the type of VOCs. Taking toluene as an example, Jiang et\u00a0al. [2] have studied the catalytic activity of Co2+and Co3+ for benzene oxidation by using the metal ion substitution method, as shown in Fig.\u00a01\n. The octahedrally coordinated Co3+ and tetrahedrally coordinated Co2+ sites were replaced with inactive ion species, such as Zn2+, Al3+, and their results indicated that the Co3+ is beneficial for breaking the benzene ring, leading to a superior catalytic activity. At the same time, partial Co2+ species located at octahedral sites were more easily oxidized to Co3+ species, which also contributed to catalytic performance.As reported, the different catalytic behavior was strongly associated with the activation of oxygen species, and that was originated from the formation of crystalline phases [12,14,15]. For instance, Li's group [3] used the MOF-templated approach to prepare M-Co3O4 catalysts with different crystallinity and explored their catalytic performance in toluene oxidation. It has been found that the calcination temperature was the main influence factor in the formation of structure and peculiarity, as well as the M-Co3O4\u2013350 catalyst which calcined at 350 \u00b0C possess smaller nanosized crystal size and superior activity (T90 of toluene conversion is about 239 \u00b0C, 1000\u00a0ppm toluene with GHSV\u00a0=\u00a020,000\u00a0mL g\u22121 h\u22121). Such outstanding performance can be attributed to the smaller nanosized crystals formed, which raises the concentration of defect sites and active oxygen species. Similar results were also revealed by other researchers, the poor crystallinity offers a great number of structural defects and oxygen vacancies, thereby promoting the replenishment of oxygen species and the deep oxidation of VOCs [16\u201318]. Besides, the phase transformation of CoOx catalyst also plays a vital role in VOCs purification [19]. As displayed in Fig.\u00a02\na-b, Jiang et\u00a0al. [4] synthesized CoOCo3O4 mixed-phase catalyst by using citric acid complexation and C2\nC5 diols, the relevant characteristic demonstrated the formation of mixed-phase could effectively reduce the strength of CoO bond, which elevated the mobility of oxygen species, resulting in best propane conversion activity (T90 =227 \u00b0C, 0.3 vol% propane with GHSV\u00a0=\u00a030,000\u00a0mL g\u22121 h\u22121). Such improvement by mixed phased construction could strengthen the interaction between different CoOx phases and bring better catalytic performance for VOCs abatement. However, the comprehensive building of the mixed-phase CoOx catalyst based on metal\u2212oxide interaction was, to our knowledge, still pretty rare.The selective synthesis of Co3O4 with dominating crystal facets is of great significance for catalytic reactions [20\u201323], especially for air purification and practical applications [24,25]. In most cases, the high-energy facets with expected properties were buried inside the catalyst during rapid growth, which accompanied by the exposure of stable facets on the external surface, resulting in dissatisfactory catalytic performance [26]. Thus, exploring the exposure of reactive crystal facets seems to be a promising strategy to affiliate more active species for catalytic reactions. For instance, Liang et\u00a0al. [5] developed a strategy for controllable construction of CoOx catalyst with numerous morphologies and specific crystal facets for propane combustion. The result demonstrated that the book-shaped Co3O4-B sample dominated with (110) facet, which endows favorable active oxygen species and good low-temperature reducibility, thus promoting the oxidation of propane (T90\u00a0=\u00a0278 \u00b0C, 0.2 vol% propane, GHSV\u00a0=\u00a0120,000\u00a0mL g\u22121 h\u22121). Similarly, the catalytic activity of Co3O4 materials with different exposed crystal facets ((110), (100), (111)) was compared by He and coworkers [6], and the corresponding results also indicated the (110) facet with high-energy was favorable for the formation of oxygen vacancies and low coordination Co atoms, eventually induce the enhancement of propane oxidation (2500\u00a0ppm propane was fully converted into CO2 at 205\u00a0\u00b0C under WHSV\u00a0=\u00a030,000\u00a0mL g\u22121 h\u22121), as depicted in Fig.\u00a02c. Obviously, the facet engineering of (110) plane appears to be an effective technical approach for VOCs removal.As a matter of fact, even for catalysts with the same components, the changing in microstructure always results in differences in the catalytic activity, which was caused by the variations in the exposure of active sites, the concentration of active species, low-temperature reducibility, and so on. Besides, the pore structure also plays an important role in VOCs oxidation, just as the 3D CoOx usually shows better catalytic performance than 1D catalyst due to its large specific surface area and rich porous structures, thereby promoting the mass transfer of reactant and providing more contact opportunities through complex channels. In an early study, Ye et\u00a0al. [7] have prepared 1D-Co3O4 nanoneedle, 2D-Co3O4 nanoplate and 3D-Co3O4 nanoflower by using the template-free hydrothermal method. As illustrated in Fig.\u00a02d, compared with 1D-Co3O4, the 3D nanoflower displayed much better catalytic activity, where T90\u00a0=\u00a0238 \u00b0C (WHSV\u00a0=\u00a048,000\u00a0mL g\u22121 h\u22121). A detailed investigation has revealed the large specific surface area with abundant active Co3+ sites and oxygen species were responsible for the improved oxidation ability. In addition to the three-dimensional transformation, the subtle structural variations also significantly affect the performance of the catalyst. Recently, Ren and Ye [27] fabricated 4 kinds of 3D hierarchical Co3O4 and explore their catalytic performance for toluene oxidation, where the sheet-stacked fan-shaped Co3O4 exhibits high efficiency due to its large specific surface area, highly defective structure and rich high valence Co species. Meanwhile, the CoOx catalyst with an orderly structure demonstrated great potential for VOCs abatement. 3DOM (Three-dimensionally ordered microporous) Co3O4 with surface area as high as 22.4 m2 g\u2212\n1 was prepared by PMMA template methods [28]. Compared with commercial 3D cobalt oxides, the 3DOM Co3O4 possesses rich adsorbed oxygen species, good low-temperature reducibility, leading to specific catalytic properties.The destruction of VOCs over cobalt oxides is dependent on the redox cycle of Co ions, which is likely to be influenced by its morphology (such as rods, tubes, spheres, sheets, flowers, cage, etc.), affecting the activity in total oxidation reactions [18,29\u201332]. The meso\u2011Co3O4 catalyst with urchin-like (U-80), shale-like (S-160) and mixture (US-120) morphologies were obtained and employed in toluene oxidation in our group [8], as displayed in Fig.\u00a03\n. It has been found that the morphologies can be tuned from urchin-like to shale-like by simply adjusting the hydrothermal temperature from 80 to 160 \u00b0C, and the S-160 sample possessed the best catalytic performance, where achieved the T50 of toluene conversion at 234\u00a0\u00b0C (toluene\u00a0=\u00a0500\u00a0ppm GHSV= 60,000\u00a0mL g\u00a0\u22121 h\u22121). And detailed investigation revealed the enhanced catalytic performance could be assigned as the rich active oxygen species, good redox property and dominated (110) planes, and these factors are closely related to its specific morphology. Apart from catalysts with common morphologies prepared by simple methods, there are plenty of CoOx with complex morphologies that have been obtained in recent years and possessed superior catalytic performance in VOCs abatement [33,34]. For instance, Lin et\u00a0al. [9] reported the Co3O4 catalyst with nanorod interweaved lamellose structure by using MOF as a template, which exhibited better catalytic performance (toluene conversion T90\u00a0=\u00a0188 \u00b0C, under the conditions: toluene\u00a0=\u00a03000\u00a0ppm, WHSV\u00a0=\u00a030,000\u00a0mL g\u22121 h\u22121) than Co3O4 nanofiber and nanosheet. And such superior activity originated from the higher concentration of Co3+/Co2+ redox couples and abundant defects.As mentioned, it seems that the influence of the different designing strategies on catalytic performance is highly related to the type of VOCs. Thus, to compare the influence of different factors and consequently provide guidelines for the fabrication of highly effective catalysts, a preliminary analysis was conducted by using toluene as an illustration. As shown in Fig.\u00a04\n, several single cobalt oxides as mentioned before were selected and their catalytic activities were normalized as toluene conversion rates. It can be seen that all these single cobalt oxides follow the chemical formula of Co3O4, and this implies that the crystalline phase of Co3O4 favors the catalytic degradation of VOCs. Compared to bulk Co3O4 (M-Co3O4), the Co3O4 with well-designed 3D structure exhibited higher catalytic activity, which indicates the construction of 3D structure is beneficial for the transportation of VOCs molecules and the exposure of active sites, resulting in better catalytic performance. Besides, the Co3O4 with rich Co3+ species displayed superior catalytic activity and lower T90 temperature than bulk M-Co3O4, which also implies the exposure of Co sites with high valence states is of great help to prepare effective co-based catalysts. Besides, the modification of morphologies also plays a crucial role in the redox cycle of Co3+/Co2+, thereby promoting the catalytic performance, which can be certified by the excellent toluene conversion rates and lowest T90 temperature. Therefore, it can be concluded that designing Co3O4 catalyst with specific morphologies, three-dimensional structure and abundant cobalt sites with high valence states is promising in the field of VOCs elimination. In addition, all these works highlight the influence of oxygen mobility, oxygen vacancy density and the concentration of oxygen species, indicating modification of oxygen species is also important.Perovskite oxides (ABO3), as a promising catalyst for heterogeneous catalysis, contains the merit of relatively low price, excellent redox performance, high oxygen mobility, and good thermal stability [39]. Due to the modification of preparation conditions, the A and B metallic cations could exist as different valence states, leading to the redistribution of redox cycles, resulting in the alteration of the redox properties and catalytic activity [40]. Among these perovskites, the cobalt perovskite (ACoO3) is widely applied in air purification due to their nonstoichiometric composition and multivalent nature of cobalt species [41,42], and significant efforts have been made in current science to further improve the catalytic performance, as displayed in Table\u00a02\n. Representatively, the surface etching seems to be an efficient strategy to enhance the catalytic behavior of ACoO3 by creating more oxygen vacancies and providing more active sites. For example, Li et\u00a0al. [35] reported well-designed LaCoO3 for toluene oxidation by employing the citrate sol-gel method and acid etching strategy. Compared to bulk LaCoO3 (LCO-0) catalyst (T90\u00a0=\u00a0263\u00a0\u00b0C), the catalytic activity of LaCoO3 treated with acetic acid (LCO-1) was significantly improved, where the T90 is approximately 223\u00a0\u00b0C under a WHSV of 60,000\u00a0mL g\u22121 h\u22121. The acetic acid etching was favorable to generate more small nanoparticles with large surface area and improve the Oads proportion, resulting in better catalytic performance. Simultaneously, the precisely controlled acid treatment also remains the structure of perovskite phase, allowing superior stability even after thermal treatment at 500 \u00b0C. Similarly, Chen and coworkers [43] prepared the Co-enriched LaCoO3 oxides with abundant surface defects by using selective acid etching. This approach is based on the La ions located at A-site being preferentially dissolved than cobalt ions located at B-site, which induces the re-dispersion of Co species on the surface of ACoO3. And such unique structure facilitates the electron transfer in the Co3+/Co2+ redox cycles and accelerate the activation of oxygen molecules, thus achieving optimum catalytic property (T90\u00a0=\u00a0206 \u00b0C under the condition of 1000\u00a0ppm toluene with GHSV\u00a0=\u00a015,000\u00a0mL g\u22121 h\u22121) and durability (remains stable over 10\u00a0h under the condition of 5 vol% water).Besides, the highly stable nature of ACoO3 also makes the appearance of structural defects possible via cation substitution, which produces more oxygen vacancies to enhance the catalytic activity in VOCs oxidation [39]. Thus caused the vast majority of works are focused on metal ion substitution over A or B site of perovskite to yield more defective structures. For instance, Liang et\u00a0al. [44] had conducted A site substitutions over LaCoO3 catalyst by using Ag cations, which effectively raised the proportion of surface oxygen and low-temperature reducibility, hence promoting the catalytic performance. Similar conclusions were also found by Weng's group [45], in which the oxygen mobility and the redox properties of perovskite can be affected by the introduction of heteroatoms. The experimental results demonstrate the introduction of Ca2+and Mg2+ induces abundant oxygen vacancies and generates more Co cations with high valence states, respectively. Such synergistic effect introduced by dual-site substitution shows a relatively positive effect on toluene oxidation. Similarly, the combination of different modifications also exhibits unexpected catalytic properties. Wei et\u00a0al. [46] fabricated a series of LaxSr1-xCoO3-\u03b4 catalysts via A site substitution together with acid treatment, as displayed in Fig.\u00a05\n. The investigation found that the combination of these two methods was not only conducive to generating more active oxygen species, but also favorable to eliminating the side effect on the catalytic performance that brought by the generation of SrO, which have resulted in catalytic activity improvements. As such, Liang et\u00a0al. [36] employed two modified methods (Ca substitution and citric acid etching) to further improve the La-Co perovskites for toluene oxidation. The results show that the T90 value of Ca-substituted LaCoO3 and acid-etched was 220 \u00b0C and 215 \u00b0C, respectively, which are substantially lower than that bulk LaCoO3 (T90\u00a0=\u00a0305 \u00b0C). Such excellent catalytic behavior can be further improved by the combination of these two strategies, resulting in lower T90 value (202 \u00b0C). Corresponding characterization reveals that these modifications provide higher specific surface area, fluffy morphology, smaller crystallite size, more oxygen vacancies, less basic sites, thereby displaying the highest catalytic activity. All these studies have determined that the synergistic effect of different strategies is beneficial to design promising catalyst for practical application.In general, spinel-type catalyst displays better stability than single metal oxides and mixed metal oxides, thereby attracted more and more attention in the field of catalytic reactions. The typical AB2O4 spinel contains the A and B metals located in the tetrahedral and octahedral positions, respectively, and its catalytic performance can be optimized via the efficient charge transfer between adjacent cations (A and B) [38,75], thus the selection of appropriate A, B cations was really mattered. Among various spinel catalysts, the Co-based spinel (CoCo2O4) is usually regarded as the most functional material for VOCs abatement, which attracted extensive attention because of its high intrinsic activity [76]. Recent investigation reveals that to further improve the catalytic activity, the spinel with larger surface area, plenty of defects and abundant surface-active species was urgently needed, which is especially applicable to govern the spinel with Co cations as B sites [75,77]. For instance, Wang et\u00a0al. [47] reported the solvothermal alcoholysis approach for the preparation of spinel MCo2O4 (M\u00a0=\u00a0Ni, Co, Cu) oxides with hollow mesoporous spherical structures towards acetone oxidation. The experimental results demonstrated that CuCo2O4 exhibits outstanding catalytic performance owing to the enriched surface Co3+ cations and abundant oxygen species, which is originated from the cation-substitution effect of Cu. While Han and Wang [48] also synthesized the spinel-type NiCo2O4 nanosheet by using hydrothermal method and discovered that E-NiCo2O4 (taking ethanol as solvent) shows satisfying catalytic performance (T99 =256 \u00b0C under the condition of SV\u00a0=\u00a010,000\u00a0mL h\u00a0\u2212\u00a01 with 2538\u00a0ppm toluene) because of its unique mesoporous structure and specific crystal plane, enlarging the Co3+ active sites located at the octahedral position. Jiang et\u00a0al. [37] proposed a facile time-saving method to design CuCo spinel catalyst by treatment of oxalic acid with different addition amounts and evaluated their catalytic activity of toluene elimination. The characterizations indicated that the appropriate addition of oxalic acid plays a crucial role in the enlargement of SBET and pore-creating via its thermal decomposition into CO2 and H2O, and it is also contributed to exposure of Co3+ cations and mobility of oxygen species by dissolving Cu2+ into Co3O4 lattice. According to the above investigations, it can be concluded that the intrinsic activity of Co3O4 spinel can be further elevated by A-site substitution method, as well as other surface embellish strategies.Besides, the fabrication of CoA2O4 spinel catalyst in which Co cations are located at A position has also attracted extensive attention. Li et\u00a0al. [78] introduced Co3-xMnxO4 (x\u00a0=\u00a00.75, 1.0, 1.5) spinel with different Co proportion by using a controlled template-free route for toluene abatement, in which the porous flower structure provided an accessible approach for the reactant to get access to active sites. The Co2.25Mn0.75O4 catalyst with rich Co composition possessed the best catalytic performance (T100\u00a0=\u00a0239 \u00b0C, under the condition of 1000\u00a0ppm toluene with the total flow rate of 33.4\u00a0mL min\u22121), and such excellent property was raised from the integrated effects of rich surface oxygen vacancies, large surface area and porous structures via the formation of Mn-Co interaction. Dong et\u00a0al. [38] compared the catalytic performance of CoMn2O4 (prepared by oxalic acid sol-gel method), single Co3O4, MnOx and mixed Co3O4/MnOx towards toluene combustion. It has been found that the CoMn2O4 sample with a large surface area and rich cationic defects demonstrate the highest catalytic activity and lowest activation energy (35.5\u00a0kJ mol \u2212\n1) compared to other metal oxides, where the 100% conversion of toluene is achieved at 220\u00a0\u00b0C even in the moisture situation (2.0 vol% water vapor). The well-designed TP experiments concluded the oxygen vacancy induced by Co substitution can accelerate the oxygen circulation, thereby elevating the rate of this reaction. Additionally, the in situ DRIFTS analysis also revealed the spinel-type catalyst has a strong ability to destroy the aromatic ring to generates anhydride species, which results in the rate-controlling step of toluene oxidation over CoMn2O4 is different from Co3O4/MnOx, thereby promoting its catalytic behavior. Hence, these investigations highlight the Co-based spinel was an efficient candidate which showing great promise in the field of VOCs elimination.The catalytic performance of Co-based perovskite and spinel catalyst was cross-compared by taking toluene oxidation as an example, as shown in Fig.\u00a04. Compared to bulk Co3O4 (M-Co3O4), most of these Co-based perovskites and spinel catalysts displayed higher catalytic activity and lower T90 temperature, implying the fabrication of Co3O4 with specific structures is quite efficient to improve the catalytic property. It can be seen that the toluene conversion rates of these catalysts follow the order of CuCoOx > La0.9Ca0.1CoO3\u00a0=\u00a0LaCoO3 > CoMn2O4, which is quite different from their rank of T90 temperature: CuCoOx > CoMn2O4 > LaCoO3 > La0.9Ca0.1CoO3. This result demonstrated the perovskite structure seems to be more reactive than spinel, endowing superior catalytic performance at relatively low-temperature. Therefore, we think it makes sense to use cobalt-based perovskite as the substrate for the forthcoming investigationsAlthough the single metal oxides possess the merit of both high reactivity and low cost, the downside of poor stability (both in chemical and thermal) usually causes particle aggregation, restraining its further application [26]. Thus, owing to the intrinsic nature of CoOx, it remains a challenge to improve the catalytic performance. Recently, the multi-metal combination provides a promising strategy to develop an efficient Co-based composite for VOCs elimination by instituting the synergistic effects between different metal oxides. This synergistic effect brings out new features by taking the complementary advantage of different metals that can promote the surface characteristic, electronic property and stability [79]. Specifically, the Co-based composite were shown to be more active than the individual cobalt oxides by similar methods, which can be mainly ascribed as the promotion of surface oxygen mobility, the improvement of low-temperature reducibility, the higher surface area, etc. [80\u201383]. Thus, much efforts have been devoted in the preparation of various composites, such as binary metal oxides (contains heterostructure of cobalt oxide, incorporation of CoOx with perovskite, the combination of cobalt oxides with other metal oxides and cobalt oxides coupled with carbon materials), multi-metal oxides and so on, which exhibits different catalytic properties. The catalytic performance of these composites was summarized in Table\u00a02.To realize a better catalytic performance, fabricating transition metal-based heterostructures have emerged as an efficient strategy due to the optimal electronic properties and tunable structural morphologies [84]. Especially, for the compound oxides, the rationally designed interface of heterostructure enables to expose more defects and active sites, thereby facilitating the catalytic activity [85]. Thus, it is very important to choose appropriate cobalt oxides to construct the functional interface. For instance, Ye et\u00a0al. [49] designed a series of Co3O4-based hetero-structured catalysts with monolithic core-shell structure by introducing different elements (Mn, Cu and Co) and evaluated their catalytic activity towards the co-oxidation of CO and toluene. Compared to single Co3O4 materials, the Co-based heterostructure catalyst displayed much better catalytic performance. While for composite oxides, the catalytic activity for the degradation of CO and toluene follows the order of Co3O4@Co3O4 > Co3O4@Co2CuO4 > Co3O4@Co2MnO4 > Co3O4@MnO2 > Co3O4, where the T99 value of Co3O4@Co3O4 reaches 230 \u00b0C under mixture conditions (1000\u00a0ppm toluene and 1.0 vol% CO, SV\u00a0=\u00a010,000 h\u00a0\u2212\u00a01). Notably, the combination with catalytic and characterization results demonstrated that the functional interface with strong interaction was fabricated in the core-shell Co3O4@Co3O4 heterostructure, which promotes the exposure of Co3+ active species and causes rich lattice defects, contributing to the superior catalytic activity even in moisture environment.Besides, textural and redox properties of heterostructure catalyst were also shown to be affected by morphology transformation, the ratio of metal oxide with different crystal forms, etc. While these influencing factors have been extensively explored by taking manganese oxide as candidates. For example, Qu's group [86] developed a simple hydrothermal method to prepare \u03b1@\u03b2-biphases materials and displayed higher catalytic performance than single MnOx, which can be further improved by adjusting the ratio of \u03b1-MnO2 and \u03b2-MnO2. However, even though the catalyst with heterostructure possessed a bright prospect in the field of VOCs elimination, such investigation for CoOx heterostructure, to our knowledge, is still pretty rare. Thus, for cobalt oxides with heterostructure, in my opinion, the exploring of feasible preparation method, probing detailed interfacial mechanism toward VOCs abatement seems to be highly desired.As mentioned, the industrial use of single cobalt oxides is still not satisfied, mainly because of its poisoning and sintering trouble during the reactions. While the perovskite oxides possessed satisfactory thermal stability and excellent anti-poisoning ability due to their high temperature aged structure, which attracted widespread attention [87,88]. But for the same reason, compared to individual metal oxides, the catalytic activity of perovskite oxides is largely confined by its surface enrichment of lanthanide series cations and low specific surface area, which only provides limited actives sites. Thus, engineering binary metal oxides via the combination of perovskite oxide and cobalt oxide are proposed as an effective strategy to overcome these deficiencies. In fact, related researches demonstrated that the synergistic effects derived from CoOx and perovskite oxide are extremely useful to improve the catalytic performance [51]. For example, the Co3O4 with different content supported by three-dimensionally ordered microporous (3DOM) La0.6Sr0.4CoO3 was prepared via in situ PMMA-templating strategies by Dai and coworkers [50]. Of the catalysts researched, the 8 wt%Co3O4/La0.6Sr0.4CoO3 was found to be the most active catalyst for toluene oxidation, achieving 90% toluene conversion at 227 \u00b0C (reaction conditions: 1000\u00a0ppm toluene, GHSV\u00a0=\u00a020,000\u00a0mL g\u22121 h\u22121). Such superior performance was linked with the 3DOM architecture and the strong interaction between Co3O4 and La0.6Sr0.4CoO3, leading to high specific surface area, enriched oxygen adspecies and promoted low-temperature reducibility. Similarly, He et\u00a0al. [89] fabricated 3D-ordered meso\u2011macroporous Co3O4/La0.7Sr0.3Fe0.5Co0.5O3 materials for 1,2-dichloroethane (DCE) oxidation via a one-step method by taking PMMA as template. The systematic investigation demonstrated the introduction of residual Co3O4 nanoparticle increase the generation of oxygen vacancies, which plays a vital role in the circulation of oxygen species and chlorine poisoning resistance. Along with the special meso\u2011macroporous structure of as-prepared perovskite favors the migration of 1,2-DCE to contact with the surface-active sites, further improving the total oxidation of 1,2-DCE. To reveal the synergistic effects and working principles between these two materials, LaOx-Co3O4 with varied La/Co ratios were synthesized via precipitation means by our group [90], as displayed in Fig.\u00a06\n. We found that the incorporation of La cations into Co3O4 facilitates the formation of LaCoO3 perovskite, as well as increases the concentration of cation defects and adsorbed oxygen species, all of which was beneficial to the low-temperature reducibility. Further investigation also demonstrated the variation of La/Co ratios have a great influence on the catalytic properties and the 80Co-20La catalyst present optimal catalytic efficiency, where the T90 of toluene oxidation was around 242 \u00b0C together with an impressive specific reaction rate (Rs\u00a0=\u00a02.0\u00a0\u00d7\u00a010\u22123\u00a0mmol h\u00a0\u2212\u00a01\nm\u00a0\u2212\u00a02). It is also revealed that the introduction of La cations in appropriate amounts could increase the interaction between CoOx and LaCoO3, promoting the re-dispersion of Co3O4 and decrease its crystal sizes, which result in the formation of lattice defects, leading to the improvement of catalytic activity. In addition to the common synthetic strategies, it is well known that the surface engineering strategy is able to tailor the textural and surface properties of catalyst, which is conducive to VOCs oxidation. For instance, Xiao et\u00a0al. [51] proposed a convenient strategy to fabricate CoOx/LaCoO3 composites by using in situ H2O2 treatment over LaCoO3 materials, which showed a higher efficiency towards propane combustion than that of bulk catalyst counterparts, achieving T90 temperature around 312\u00a0\u00b0C (reaction condition: 1 vol% propane with the WHSV of 100,000\u00a0mL g\u22121 h\u22121). The authors declared that acidic treatment by H2O2 over LaCoO3 substrate not only result in porous structure but also leads to the re-dispersion of Co3O4 particles on the surface of LaCoO3, as well as causing the formation of oxygen vacancies via the selective removal of La cations, as a consequence, the propane combustion efficiency is greatly improved. Compared to powder counterpart, the monolith catalyst can offer better dispersion of active species and unhindered mass transfer, thus showing prospective in practical employment. For instance, CoOx/LaCeO2 powder was washcoated onto cordierite honeycomb by Vidal et\u00a0al. [91] and applied in VOCs oxidation. The results indicated that the compositional property was well preserved upon being supported on cordierite monolith, while the catalyst powder was homogeneously dispersed, demonstrating good activity in the oxidation of toluene and ethyl acetate. This strategy is simple, reliable and sheds a light on catalyst designing for practical use.The shortcoming of individual cobalt oxides, such as limited absorption capacity, poor dispersity, low specific surface area, and cost issues limited its further application. In this regard, combining the cobalt oxides with non-metallic materials offer the possibility to fabricate new composites to overcome these drawbacks. Recently, various non-metallic materials with different chemical compositions, variable morphologies, and adsorption properties, such as carbon materials [92\u201396], silicon carbide [97], carbon nitride [52], silicon oxide [98] have been employed to modify Co-based oxides, which provides abundant adsorption sites, improved surface area and stronger stability, harvesting enhanced catalytic activity. Among these materials, carbon is being preferred due to its distinctive physicochemical properties, various structures and low price [92,99]. For instance, Huang et\u00a0al. [92] employed the carbon modification strategy to prepare Co-based catalysts with small grain size and evaluate their feasibility of formaldehyde abatement. The result revealed that the as-prepared carbon/Co3O4 nanocomposite (CCo3O4) demonstrate stable removal efficiency (90%) for HCHO oxidation (1\u00a0ppm), by contrast, the bulk Co3O4 was rapidly deactivated. And the enhanced activity is linked to the formation and interaction between carbon and Co3O4 interface, which caused abundant lattice disorders, promoting the generation of oxygen vacancies, thereby improving the catalytic performance. In addition, the oxygen vacancy enriched surface is also conducive to continuously convert O2 and H2O into reactive oxygen species (ROS), consequently, giving the fast and deep oxidation of intermediates and preventing the carbonate cumulation, finally result in excellent stability. The carbon material is commonly used as support due to its low price and availability. Wey and coworkers [100] coated transition metal oxides (Cu, Co, Fe, Ni) on activated carbon (AC) via polyol method for VOCs elimination and revealed that the AC support shows positive effects on the catalytic performance due to its tremendous surface area. To screen the appropriate activated carbon (AC) support with low cost, Khorasheh et\u00a0al. [99] selected three typical agricultural wastes including Iranian almond shells, walnut shells and apricot stones as precursors to prepare porous activated carbons. Compared to walnut and apricot stones approaches, the activated carbon prepared via the almond route displayed strong thermoresistance, rich oxygen groups and high hydrophobic surface area. These special properties of AC support offer multiple benefits for establishing efficient metal oxides/AC hybrids, such as enhanced surface area, high dispersion of active metal oxides, strong adsorption to VOCs and good water resistance, leading to better catalytic performance for the removal of toluene and cyclohexane. Fan et\u00a0al. [97] adopted a two-step method to support CoOx on silicalite-1/SiC foam for isopropanol removal and found that the silicalite-1 layer could not only ensure the uniform growth of Co3O4 nanoflower but also provides acid sites to absorb oxygen species and isopropanol molecules, contributing to the improvement of catalytic efficiency.Additionally, charge transfer engineering has been noticed as a promising approach to improve catalytic activity. Ren et\u00a0al. [52] well designed a series of supported CoOx catalysts by using different support, such as g-C3N4, SBA-15, \u03b3-Al2O3 and AC to investigate the influence of support on the catalytic combustion of toluene. Compared to other formulations, the 10%CoOx/g-C3N4 exhibits the optimum catalytic performance, achieving T90 of toluene conversion at 279 \u00b0C (reaction condition: 1000\u00a0ppm toluene with a total flow rate of 130\u00a0mL min\u22121), as well as excellent durability (at least 36\u00a0h). Comprehensive digging revealed that the electron-rich nature of g-C3N4 and the well-formed interfaces enhance the charge transfer and promote the generation of active Co3O4 phase, which facilitates the reducibility of Co3+, along with the high surface Co3+ content and high density of active oxygen species, responding to the enhanced catalytic performance. Thus, it makes g-C3N4 good support for its intrinsic nature and deserves further development.The integration of Fe, Cu, Mn and other metal oxides with Co-based oxides have been previously investigated in the field of VOCs elimination and show significant improvement in catalytic performance [53,101\u2013103]. And such enhancement mainly come from the synergistic effects of different elements, which usually promote elevated charge transfer between the multiple available energy levels of the metals and their associated oxygen anions, consequently resulting in better low-temperature reducibility and lavished oxygen species.For this reason, our group subsequently prepared a series of Co-M (M\u00a0=\u00a0La, Mn, Zr, Ni) composites using a co-precipitation strategy and found that their catalytic activity for toluene combustion varied quite significantly [104]. The performance observed for as-prepared catalyst follows the order as below: Co-La > Co-Mn > Co-Zr > Co-Ni, which is aligned with their low-temperature reducibility, specific surface area and pore volume results. Besides, among these materials, the Co-La sample possessed the most Co3+ active sites, excellent oxygen storage capacity (OSC) and sufficient oxygen species. All these feathers contribute to the optimal catalytic efficiency, where the 90% toluene conversion was obtained at 243 \u00b0C (reaction condition: 500\u00a0ppm toluene with a total flow rate of 100\u00a0mL min\u22121). And these results indicated the physicochemical property and catalytic performance are strongly influenced by the interaction of Co-M elements. Jia et\u00a0al. [53] conducted the catalytic efficiency of CuO/Co3O4 binary oxide derived from in situ pyrolysis of Cu2+ partly-substituted ZIF-67 template. It has been found, compared to bulk CuO, Co3O4 and mix-CuO/Co3O4 (mechanical mixture), the CuO/Co3O4 composites were highly active and 90% of toluene could be fully converted into CO2 at 229 \u00b0C (the promotion is more than 40\u00a0\u00b0C under 1000\u00a0ppm toluene with the WHSV of 20,000\u00a0mL g\u22121 h\u22121). Meanwhile, it also demonstrated good stability and strong water resistance to high humidity (10 vol%) gas flow. The superior activity observed could be assigned to the porous structure, high Co3+/Co2+ ratios, good low-temperature reducibility and abundant oxygen vacancies, and all these characteristics originated from the formation of Co-Cu interaction. This strong interaction was closely related with the preparation process, the Co2+ cations of ZIF-67 frameworks was partly replaced by Cu+ ions, thus resulting in the enhanced interaction between CuO and Co3O4 in pyrolysis process.Introducing special morphology furnishes an efficient approach to synthesis bimetal oxides with considerable catalytic performance towards VOCs oxidation. For instance, well-structured CeO2@Co3O4 core-shell material was prepared by Chen and his coworkers employing a ZIF-67 as sacrificial template, while its feasibility of toluene abatement was also investigated [54]. The CeO2@Co3O4 was determined to display a higher catalytic activity (T90\u00a0=\u00a0225\u00a0\u00b0C) towards toluene oxidation than that of pure CeO2 and Co3O4, which was strongly linked with its unique morphology. The unique core-shell structure and the hierarchically wrinkled surfaces strengthen the synergistic effect between cobalt and cerium elements, which result in better low-temperature reducibility, endowing the improved catalytic performance. Similarly, Yan et\u00a0al. [55] synthesized a series of Co3O4/CeO2 hybrids by using CeO2 with specific morphologies (plate, rod and cube) as substrate and studied their catalytic performance for dibromomethane (CH2Br2) elimination. They found the rod-like Co3O4/CeO2 material exhibited significantly improved catalytic performance than that of plate-like Co3O4/CeO2 and cube-like Co3O4/CeO2, where the T90 of CH2Br2 conversion is approximately 312 \u00b0C (under the condition of 500\u00a0ppm CH2Br2 with the GHSV of 75,000\u00a0mL g\u22121 h\u22121). The as-formed rod structure is not only conducive to the exposure of (100) and (110) planes with numerous Co3+ sites but also benefits to the improvement of surface adsorbed oxygen species, as well as the enhanced redox properties, making it superior for CH2Br2 abatement.Recently, compared to the construction of simple binary oxides, interfacial engineering has drawn extensive attention because the interface between different metal oxides usually possesses rich oxygen vacancies, which provides abundant active sites and a large number of oxygen species. For example, Ye et\u00a0al. [56] designed a new bottom-down strategy to fabricate \u03b1-MnO2@Co3O4 nanocomposite via the in situ growth of ZIF-67 over 1D \u03b1-MnO2 substrate. Of the catalyst researched, the \u03b1-MnO2@Co3O4 demonstrated great catalytic activity with 90% of toluene conversion is fulfilled at 229 \u00b0C (1000\u00a0ppm toluene with WHSV of 48,000\u00a0mL g\u22121 h\u22121), which is 28 \u00b0C and 47 \u00b0C lower than those of pure Co3O4 and \u03b1-MnO2, respectively. It has been proposed the synergistic effect between Co3O4 and MnO2 is mainly reflected in their coupled interface, and the constructed interface is contributed to improve the content of surface adsorbed oxygen species, accordingly promote the mobility of oxygen species and accelerate the redox cycles of Mn and Co cations, finally result in enhanced catalytic efficiency. Besides, it was determined that the gaseous oxygen species is more tend to be activated into adsorbed oxygen species on the surface of \u03b1-MnO2@Co3O4, giving a better oxygen supplementation for the oxidation. To investigate the influence of establishing different interfaces on catalytic performance, Co3O4 was anchored on the surface of \u03b1-MnO2, \u03b2-MnO2 and \u03b3-MnO2 by using a secondary hydrothermal method to prepare Co3O4/MnO2 binary oxides in our recent study [105], as displayed in Fig.\u00a07\n. We found that the catalytic performance increased in the order of Co3O4/\u03b2-MnO2 < Co3O4/\u03b3-MnO2 < Co3O4/\u03b1-MnO2, where the toluene can be fully converted into CO2 at 260 \u00b0C over the most effective Co3O4/\u03b1-MnO2 material, revealing the existence of a remarkable interfacial effect. Subsequent work suggested such strengthened MnO2\nCo3O4 interact not only favorable for the improvement of redox properties but also facilitate the generation of oxygen vacancies, both of these play crucial factors in the catalytic performance. Therefore, it can be concluded that interfacial engineering might provide a new way to prepare Co-based catalysts with good catalytic efficiency.As mentioned earlier, the integration of cobalt oxides and other metal oxides usually affect the structure, redox property, oxygen mobility, acid-base property of the composite, as a result, it offers a useful way to design catalysts with better catalytic activity. In this respect, the introduction of other metal elements to fabricate ternary or multicomponent composites is proposed to further enhance the capacity for VOCs oxidation. Thus, it is quite critical to understand the role of each element and the interaction between different elements to design efficient multicomponent catalysts. For the past few years, lots of elements, such as Fe, Mn, Cu, Ni, Ce, La have been employed to construct Co-based polymetallic oxides, in an attempt to improve the oxidative activity for VOCs removal [57\u201359].For instance, a highly efficient material with the composition of CuOCo3O4\nCeO2 (Cu: Co: Ce\u00a0=\u00a010: 45: 45) was prepared by Domen's group [106], seeking proper means to control the large-scale VOCs emissions. It has been found that CeO2 and CuOCo3O4 are beneficial to reduce the risk of sintering and maintain structural strength during the oxidation process, respectively. As a result, the ternary oxides exhibited excellent activity and stability for the elimination of industrial exhaust gasses (including toluene, xylene, ethylbenzene, butyl alcohol and formaldehyde), where the most of VOCs was degraded to below 0.5\u00a0ppm (under the condition of SV =16,600\u00a0mL h\u00a0\u2212\u00a01, reaction temperature\u00a0=\u00a0280 \u00b0C, catalyst volume =188\u00a0mL). Jir\u00e1tov\u00e1 et\u00a0al. [107] prepared the Co-Mn-Al ternary composite and employed K doping strategy to further improve its catalytic behavior towards toluene and ethanol oxidation. It was determined that the low K additions (1%) could increase the acidity of the catalyst, as well as induce the generation of more active Co3+ species, resulting in catalysts with enhanced activity. Wang and his coworkers [57] reported the fabrication of CuyCo3-yFe1Ox catalyst with adjustable oxygen vacancies via LDHs precursors as the efficient catalyst for toluene elimination. Comparison with Co3Fe1Ox (289\u00a0\u00b0C) and Cu3Fe1Ox (304\u00a0\u00b0C) binary oxides, the 100% of both toluene conversion and mineralization over ternary Cu1Co2Fe1Ox are realized at approximately 241 \u00b0C, which is closely related to the as-obtained multi-phase interfaces. The systematic investigation demonstrated that good dispersion and close contact between different metal phases are conducive to the reinforcement of synergistic effect, which is responsible for the enhanced oxygen vacancies than binary oxides. Besides, it has been found that the ternary Cu1Co2Fe1Ox exhibited excellent stability and water resistance, and these factors are very crucial for practical application.Additionally, the precise design of multi-metal composite with specific structures was also investigated. Chen and coworkers [58] prepared a novel nanocage with the formula of MnCeO\u03b4/Co3O4 by using ZIF-67 as a sacrifice template. Compared with bulk Co3O4 and Co3O4 nanocube, the as-prepared MnCeO\u03b4/Co3O4 catalyst demonstrated the lowest apparent activation energy (56.10\u00a0kJ mol\u22121) and highest catalytic activity, where the T95 of toluene oxidation is around 230 \u00b0C (1000\u00a0ppm toluene, WHSV\u00a0=\u00a040,000\u00a0mL g\u22121 h\u22121). It can be concluded that the well-designed structure boosts the interactions between MnCeO\u03b4 and Co3O4, thereby providing an adequate specific surface area (100.40 m2\ng\u00a0\u2212\u00a01), high concentration of surface adsorbed oxygen and Co3+ species, excellent low-temperature reducibility, endowing the superior catalytic property. Similarly, the low-temperature elimination of 1,2-Dichlorobenzene (o-DCB) was investigated over 3D hollow nanoflower ball-like NiO@MnMOx (M\u00a0=\u00a0Co, Cu and Fe) materials by Zha and Tang [59]. They found the application of interfacial reaction over the surface of hollow substrate is an effective approach to prepare polymetallic oxide with specific structures and adjustable properties. The unique hollow flower structure of NiO@MnCoOx not only offers abundant adsorption sites, but also shows the merits of numerous active species (such as Co3+, Ni2+, and Mn4+), enriched active oxygen species, as well as the vast of acid sites. Based on these favorable conditions, the optimal NiO@MnCoOx exhibited outstanding catalytic activity and superior reusability towards the destruction of o-DCB.Creatively, a recent study of Sun's group [60] points out a new way to acquire polymetallic oxides by using spent lithium-ion cobaltate batteries. As displayed in Fig.\u00a08\n, the cathode of the spent lithium-ion batteries was treated by citric acid, subsequently, filtered out and finally calcined to reach the LiCoM (M includes Ni, Mn, Cu and Al elements) composite. Besides, the Li and Al ions were wiped out from the leaching solution via the treatment of oxalic acid, and the as-obtained solution was further calcined and marked as Co3-xMxO4 (M contains Ni, Mn, Cu). To make a comparison with the individual cobalt oxide doping by different elements, the corresponding catalysts were also prepared and named CoM-X. Of the catalyst prepared, the catalytic activity for toluene oxidation follows the trend of Co3-xMxO4 > Co3O4 > LiCoM, indicating the presence of aluminum and lithium elements is disadvantageous to the catalytic performance. Conversely, the introduction of manganese and copper displayed a positive effect on catalytic performance, which is conducive to the formation of stronger weak acid sites, abundant Co3+ and Mn4+ active sites, favorable Olatt/Oads ratios, together with the promotion of low-temperature reducibility. Thus, it shed a light on the employment of waste materials as precursors to prepare efficient catalysts for air purification.Differ from the integration of CoOx with other materials, the surface engineering strategy is another way to modify the redox property and the surface oxygen vacancy concentration of CoOx, which can facilitate oxygen mobility and consequentially improve the catalytic performance. Recently, among the different surface engineering strategies, doping method and acid treatment are two commonly studied strategies, which exhibits promising prospects in VOCs purification [26,69,72,73].Dopant was widely used to modify the surface oxygen property by varying the electronic and geometric nature of the host metal oxides. Generally speaking, the doping strategy can be divided into three categories according to the employment of different dopants, mainly including noble metal, non-noble metal, and non-metal dopant. As mentioned before, noble metal catalysts faced the disadvantage of sintering, poisoning and high cost, nonetheless, their outstanding catalytic efficiency is nonnegligible. Thus, the doping of noble metal (including Ag, Au, Pd, Pt, Ru, etc.) is commonly used to improve the catalytic performance of Co-based metal oxides [61,62,64,108].For this reason, Ge and Yu [61] prepared Ag/Co3O4 catalyst by using a one-pot solvothermal method and explored its catalytic behavior towards benzene elimination. It has been found that the as-obtained 2%Ag/Co3O4 displayed superior catalytic activity compared to the 2%Ag/Co3O4-I synthesized by simple impregnation, where the 90% of benzene can be oxidized into CO2 at 201 \u00b0C (under the condition of 100\u00a0ppm benzene with GHSV\u00a0=\u00a0120,000\u00a0mL g\u22121 h\u22121). The characterization results demonstrated the doping of Ag by solvothermal strategy can induce more surface Co3+ species and oxygen vacancies, subsequently promoting the low-temperature reducibility and the amount of active oxygen species, respectively, contributing to the benzene elimination. He et\u00a0al. [108] reported the catalytic oxidation of o-xylene over Pd supported on Co3O4 catalysts, and explore the influence of the Co3O4 supporter and Pd dopant. Depending on the carrier employed, the catalytic performance for o-xylene oxidation are as follows: Pd/Co3O4 (3D) > Pd/Co3O4 (B) > Co3O4 (3D) > Co3O4 (B). It was determined that the ordered mesoporous 3D Co3O4 support not only exhibit porous structure, but also benefits to the exposure of PdO active sites, and all these features would be responsible for its outstanding catalytic activity, resulting in the T90 value of o-xylene combustion was achieved around 249 \u00b0C (150\u00a0ppm o-xylene and WHSV\u00a0=\u00a060,000\u00a0mL g\u22121 h\u22121). Among the commonly used noble metal catalyst, the Pt have attracted extensive attention, which is attributed to its high efficiency at low temperature. It has been reported the introduction of Pt species over Co3O4 substrate may not only provides more active sites, but also increase the proportion of oxygen vacancies via its affinity to the 3d orbital of Co atoms [109]. Ye et\u00a0al. [62] utilized a novel metal-organic templated conversion method to prepare Pt-Co3O4 catalyst and evaluated its catalytic performance for toluene oxidation. It was determined that the conversion from rhombic dodecahedron to nanosheet is favorable for the exposure of Pt nanoparticles, as well as the enhancement of strong metal-support interaction (SMSI), both of these have substantial effects on the catalytic activity, thereby achieving the optimal catalytic performance over Pt-Co(OH)2-O catalyst (T90\u00a0=\u00a0167 \u00b0C under the condition of 1000\u00a0ppm toluene with WHSV\u00a0=\u00a060,000\u00a0mL g\u22121 h\u22121). Besides, the further investigation indicates the existing SMSI effect could facilitate the electron transfer and weaken the CoO bond, which were conducive to catalytic activity.In addition, the co-doping of different noble metals also provides an efficient approach to further improve the catalytic efficiency. For example, Dai et\u00a0al. [63] reported the preparation of 3DOM Co3O4 and x%AuPd/3DOM Co3O4 catalyst via polymethyl methacrylate-templating and polyvinyl alcohol-protected reduction routes, respectively. Of the catalyst researched, the 3DOM Co3O4 coupled with Au-Pd alloy performed much better than supported individual Au or Pd materials, and the best catalytic activity was achieved on 1.99%AuPd/3DOM Co3O4, where the T90 of toluene conversion is approximately 168 \u00b0C (GHSV\u00a0=\u00a040,000\u00a0mL g\u22121 h\u22121). It was found that compared to the supported single Au or Pd catalyst, the AuPd/3DOM Co3O4 possessed the lowest apparent activation (33\u00a0kJ/mol) and strongest ability in the activation of oxygen and toluene, thereby showing the better catalytic performance. Based on this research, they further prepared a meso\u2011Co3O4 support ternary AgAuPd alloy nanoparticles by using KIT-6 templating and NaBH4 reduction strategy [110]. It was found that the supported AgAuPd material outperformed the Au, Pd, Ag sample, which can be attributed to the strong interactions between alloy nanoparticles and meso\u2011Co3O4, resulting in the highest oxygen vacancy density, thereby improving the catalytic performance.However, owing to the high proportion of noble metal being used, the industrial application of these catalysts is still restricted by economic considerations. Thus, a lot of efforts have been made to tackle this problem. Recently, it has been found that the single-atom catalyst (SACs) is a promising way to reduce the cost of the noble metal catalyst, as well as improve the catalytic activity. For instance, a series of single atom Pt-Co/HZSM-5 composites were synthesized and used for the catalytic combustion of dichloromethane (DCM) by Liu and Han [64], as displayed in Fig.\u00a09\na-b. Compared to individual Co/HZSM-5 material, the doping of trace amount of Pt species (0.01\u00a0wt.%) over Co/HZSM-5 support could result in tremendous promotion in catalytic performance. Further investigation revealed the construction of Co-Pt interaction was beneficial to anchor Pt atoms, which promote the high dispersion of single-atom Pt species, thereby increasing the concentration of oxygen vacancies and the redox properties of Co3O4, accelerating the deep oxidation of DCM, in turn, protect the Pt species from being poisoned. He et\u00a0al. [109] presented an atomically dispersed 0.02%Pt1-Co3O4 catalyst and found that was exceptionally efficient for methanol oxidation. By combining the experimental investigation and density functional theory calculations, it has been found the single Pt atoms was anchored on the (111) planes of Co3O4, which exhibited high occupied electronic states, demonstrating significant electron transfer between Pt and Co, sequentially accelerating the regeneration of oxygen vacancies, leading to the effective oxidation of VOCs at last. Thus, it seems that the design of Co-based SACs catalyst is a promising way to fabricate catalyst with excellent activity and high cost-effective.Without the disadvantage of high cost and instability, the non-noble metal dopant has been extensively studied to improve the catalytic activity of individual cobalt oxides. For instance, Achraf et\u00a0al. [65] synthesized cobalt spinel film doped with a trace amount of Cu by using a novel pulsed-spray evaporation chemical vapor deposition method to study its catalytic effect on C3H6 elimination. The as-formed catalyst displayed superior activity, and it is even comparable with the supported noble metal catalyst, which can be attributed to the enriched oxygen species and well-dispersed Cu particles. Mn decorated cobalt oxides with abundant surface defects were successfully prepared via a MOF-templated route by Ma and Yu [66]. It was determined that the incorporation of Mn into the surface lattice of cobalt oxide could cause the formation of lattice distortion, which is likely to expose more surface defects than bulk defects. Besides, owing to the efficient electron transfer between Co-Mn couples, more active Co3+ species were also exposed on the surface of catalyst. Both of these factors were conducive to the promotion of low-temperature reducibility and lattice oxygen mobility, contributing to the enhanced catalytic behavior rather than single Co3O4 catalyst. Fendler and coworkers [67] pointed out that the different metal dopants adopted for the preparation of M-Co composite had a significant influence on their activity for VOCs elimination. As depicted in Fig.\u00a09c-d, the catalytic performance for toluene oxidation of these catalysts follows the order of NiCo > CuCo > MnCo > Co3O4 > FeCo, while their catalytic behavior towards propane combustion was quite different, in the order of MnCo \u2248 Co3O4 > FeCo > NiCo > CuCo, indicating the doping of Mn seems to be ideal way to prepared Co-based catalyst with strong universality.In addition to the commonly used transition metal dopant (Cu, Mn, Fe, Ni, etc.), other metals, including rare-earth metal, alkali metal, etc. were also widely doped into Co-based materials to improve the activity. Recently, our group reported the Ce doped on Co3O4 lower the T90 temperature (238 \u00b0C- 257 \u00b0C) for toluene combustion as the Ce: Co ratio increased from 0:1 to 0.05:1 [68]. It has been revealed the introduction of modest cerium provide more surface Ce3+ species, which was highly related to the formation of oxygen vacancies, further improving the removal efficiency. Similarly, Xiao et\u00a0al. [111] pointed out that the doping of La could induce the lattice distortion of Co3O4, thereby promoting the generation of surface defects, demonstrating improved performance. It is worth mentioning that the incorporation of La is not only conducive to the adsorption of toluene molecule, but also inhibits the accumulation of carbonate intermediates, as a result, contributing to the long-term reactions. Guo and Fendler [69] investigated propane combustion over Zr doped Co3O4 materials and found that the 1%Zr-Co3O4 exhibited the best activity, with T90 of propane conversion achieved at 241 \u00b0C. The performance of this material was dependent on several factors, including the smaller Co3O4 grain size, rich Co2+ and oxygen species, enhanced low-temperature reducibility, and all these features were triggered by the formation of Co-O-Zr species via Zr entering Co3O4 lattice. Schwank et\u00a0al. [70] reported the indium with a large cation radius could affect the chemical status of oxygen species around Co cations, which demonstrated that the surface lattice oxygen tend to be more easily abstracted by C3H6 during the reactions, thus promoting its catalytic performance.However, doping technology was not always effective in improving catalytic properties. Gao et\u00a0al. [71] investigated the doping effect of alkali metal (Na, K, Li) into Co3O4 catalyst, which revealed the introduction of alkali metals could significantly delay the VOCs combustion reactions. Such poisoning effect can be divided into aspects, including the locking effect on oxygen species and the growing adsorption of CO2, which result in poor oxygen mobility and diminished active sites, respectively.Alongside the traditional metal doping technology, the doping of nonmetallic elements has been employed in catalytic reactions, such as photocatalysis or electrocatalysis. It has been reported the introduction of nonmetallic dopants can optimize eg electron filling and improve the conductivity of oxygen ions, thereby raising the oxygen evolutions [112]. For this reason, Qian et\u00a0al. [72] synthesized Co-based perovskite doped with P to investigate the catalytic effect on propane combustion. As shown in Fig.\u00a09e-f, it was determined that the doping of P had a significant influence on the catalytic performance, where the T90 temperature for propane oxidation was achieved at 376 \u00b0C over the optimal LaCo0.97P0.03O3 catalyst under the condition of 0.8% C3H8 with GHSV\u00a0=\u00a060,000\u00a0mL g\u22121 h\u22121. The experiment results revealed the doping of P has several advantages, such as regulating the valence states of Co cations, improving active oxygen density, enhancing surface acidity, all of which results in better oxidation properties. Although this study provides a new strategy for designing efficient Co-based catalysts in the field of VOCs elimination, but as far as we know, similar work is still pretty rare.As mentioned in the section on Co-based perovskite, several researchers reported the mechanisms for instituting Co3+-rich perovskites via acid treatment, indicating the acid etching is quite effective in surface reconstruction. Thus, the influence of acid treatment on Co-based materials has been investigated in the field of VOCs degradation and found that was contributed to modify the valence states of Co ions, subsequently facilitating the oxidation reactions. For example, a series of mesoporous Co3O4-n (n represent the concentration of acid) catalyst was fabricated by using acid etching technology over Co3O4-P that prepared via hydroxycarbonate precipitation method [73], as displayed in Fig.\u00a010\n. Among these materials, an obvious enhanced catalytic performance was observed over Co3O4\u20130.01 catalyst, which exhibited outstanding activity (T90 =225 \u00b0C) and strong water resistance during the oxidation of toluene. It was concluded that after the treatment of acid, the Co3O4-n demonstrated rich Co2+ and adsorbed oxygen species, together with large surface area and more weak acidic sites, benefitting the improvement of catalytic behavior.Similarly, acid etching was applicable to the regulation of vacancy (including cation and anion) density. For instance, Liu et\u00a0al. [113] reported the fabrication of LiCoO2 catalyst for benzene degradation via acid treatment of spent cathodes. They pointed out the Co3+ and Li+ cations were leached out due to the disproportion reactions caused by nitric acid treatment, inducing the formation of Lo, Co and O vacancies. And the existence of Li and O vacancies could promote the adsorption and activation of benzene species, while the presence of Co and O vacancies boosted the formation of abundant active oxygen species, accordingly, the catalytic performance of as-prepared LiCoO2 was significantly improved. Moreover, the recent investigation of Guo and coworkers [74] have proposed that the surface defects induced by HF modification can be utilized to strengthen the interaction between Co3O4 substrate and active RuOx, thus improving the dispersity and stability of active RuOx species, leading to the enhanced activity for the oxidation of vinyl chloride. All these works indicate the acid etching is an efficient method that can help to develop catalysts for VOCs elimination.The catalytic performance of as mentioned Co-based composite was normalized as toluene conversion rates to compare the effects of these strategies for the fabrication of highly effective catalysts. As shown in Fig.\u00a011\n, it was found that CeO2@Co3O4 showed better catalytic activity than catalysts prepared via the simple combination of cobalt oxide and other materials (perovskite, non-metal material), implying that Co-based binary oxides could be an effective path for VOC oxidation, where the choice of the second metal is very important. Besides, it can be observed the \u03b1-MnO2@Co3O4 exhibited extraordinary performance, even comparable to some noble catalyst at low temperature, which indicates the interfacial engineering strategies used in this work is incredibly effective, and this strategy may become one of the hot spots for future researchCompared with common cobalt-based binary oxides, the catalytic performance of cobalt-based ternary oxides (MnCeO\u03b4/Co3O4) is not inferior or even better, suggesting that combining the advantages of multi-metals provides a useful way to design catalysts with better catalytic activity. In addition, as evidenced in Fig.\u00a011, it is worth mentioning that surface engineering techniques are also an effective way to improve catalytic performance, and the doping strategy is of particular interest, where the toluene conversion rates of noble metal doping catalyst (Pt-Co3O4, AuPd/3DOM Co3O4) can reach high levels (2.68 and 1.79\u00a0mmol g\u22121 h\u22121) at very low temperatures (167 and 168 \u00b0C). Therefore, the future directions of cobalt-based noble metal catalyst development should be reducing the noble metal loading and improving the stability of noble metal by adopting noble metal alloy or single atom noble catalyst, so as to increase its industrial application potential. to investigate the catalytic effect on propane combustionWater, as one of the by-products often present in the flue gasses emitted by various industries, is also one of the products of catalytic oxidation of VOCs. The impact of water vapor on the activity of catalysts depends on a variety of factors, such as the type of VOCs, catalyst composition, reaction conditions, etc., which have been extensively studied and reported. Generally, the presence of water vapor is able to compete for the adsorptive sites of catalyst with VOCs molecules, leading to inadequate oxidation of VOCs. For instance, Li et\u00a0al. [73] claimed that the water vapor had a significant negative effect on the performance of Co3O4 catalyst in toluene oxidation (toluene conversion decreased from 90% to 61% in the presence of 5\u00a0vol.% of water vapor at 225 \u00b0C), and this effect completely disappeared after removing the feeding of water vapor. Thus, the catalyst with high water resistance is one of the promising research directions. Zhan et\u00a0al. [114] prepared a series of LaCoO3 catalysts via a citric acid sol-gel method to investigate the catalytic effect on propane combustion and CO oxidation. They claimed that the introduction of water inhibited the oxidation of CO over LaCoO3 catalyst due to a decreased activity under the humid atmosphere (3\u00a0vol.%) compared to the dry reaction conditions. Besides, they also found the acid etching was able to improve the water resistance of the catalyst, according to the stable catalytic behavior of LaCoO3-AE (treated by acid etching) in the humid reaction atmosphere. Hao and coworkers [115] reported the amorphous Co1Mn3Ox displayed great water resistance capacity (propane conversion slightly changes from 82% to 79% in the presence of 3.1 vol% water vapor), indicating the appropriate combination of bimetal will provide excellent water vapor tolerance. Similarly, Chen et\u00a0al. [116] pointed out the construction of CuxCo3-xAl mixed metal oxides catalysts delivered strong water tolerance, which remains 93.8% of benzene conversion (94.5% in dry condition) with the presence of 1.5\u00a0vol.% water.Although water is generally considered to be a hindrance to VOCs removal, in some cases, the presence of water vapor may be beneficial. For example, the positive effect of water vapor on the catalytic removal of formaldehyde was reported by Huang and Shen [117]. They found that the Co@NC catalyst showed higher catalytic performance for formaldehyde removal in humid environments (relative humidity\u00a0=\u00a025%) than in dry conditions. Zhan et\u00a0al. [118] proposed that the existence of water in the feed gas shows an obvious negative influence on the activity of Ru/CoANS (cobalt-doped alumina nanosheets) in propane oxidation attributed to the competitive adsorption of H2O and C3H8/O2 on the RuOx active sites. On the contrary, for Pd/CoANS catalyst, although the propane conversion instantaneously drops from 82% to 52% when water is first introduced, subsequently the propane conversion gradually increases to 100%, which is much higher than the initial conversion under dry conditions. Besides, such enhancement of activity could remain at least 8\u00a0h after the water vapor was switch off, demonstrating the re-activation phenomenon by water vapor. Further investigation revealed that the removal of Cl species remaining on the surface during catalyst synthesis by H2O was responsible for the re-activation effect. This phenomenon has also been reported for the catalytic elimination of CVOCs, which demonstrated the aggressive role of water in removing Cl\u2212 from the active site, preventing the catalyst from being deactivated. Zhang and Zhao [119] studied the catalytic durability of a RuCoOx/Al2O3 catalyst for vinyl chloride elimination and revealed that the introduction of 1 vol% water leads to a tremendous increase in VC conversion, which derives from the removal of surface Cl by H2O, leading to an increase in HCl yield. And after stopping the addition of water vapor, the conversion of VC gradually decreased to stabilization, further confirming the speculation.In conclusion, the contribution of water vapor in the catalytic removal of VOCs is quite complex. Therefore, the effect of water vapor should be considered while designing application-based catalysts.Generally, the mechanism of VOCs combustion can be divided into three categories, including Eley-Rideal (E-R), Langmuir-Hinshelwood (L-H) and Mars-van-Krevelen (MVK) models. And the applicability of each mechanism is strongly linked to the catalyst nature, as well as the properties of VOCs. Thus, the investigation of catalytic mechanisms over Co-based catalyst have been extensively researched, and it was demonstrated that the MVK model is the commonest for VOCs elimination [33,34,82].As shown in Fig.\u00a012\n, the MVK mechanism can be interpreted as a two-step redox model: Firstly, the absorbed VOCs molecules react with active lattice oxygen, which caused the partial reduction of metal oxide, resulting in the generation of oxygen vacancies; Subsequently, the as-formed oxygen vacancy was immediately replenished by gas-phase oxygen species in the airflow or oxygen from the bulk. The balance between oxidative and reductive rates is quite important for catalytic reactions. Thus, it indicates the improvement of oxygen mobility and defect density can be useful to design efficient Co-based catalysts.L-H model is also widely used in VOCs combustion, which assumed the reaction occurs between the absorbed VOCs molecules and the absorbed oxygen species [120]. Accordingly, the controlling step of this mechanism is the reaction rate between these two species. For example, Jiang et\u00a0al. [82] reported the degradation of benzene over ACo2O4 spinel catalyst proceeded via both MVK and L-H mechanisms, in which the lattice oxygen takes part in the generation of carboxylates intermediate species and the absorbed oxygen species favor the oxidation of carboxylates species to final products. Thus, the L-H mechanism provides a reasonable explanation for the fact that those Co-based catalysts containing rich adsorbed oxygen species usually exhibit elevated catalytic performance.Unlike MVK and L-H model, the E-R mechanism demonstrated that the catalytic reaction happens between absorbed VOCs molecules and oxygen species in gas phase (or between the absorbed oxygen species and VOCs molecules in gas phase) [121]. The E-R model has very limited applications toward VOCs oxidation, which was only applicable to catalysts with inert or less reactive carriers (such as zeolites, activated carbon, etc.). For instance, Khorasheh et\u00a0al. [122] found the E-R model was appropriate to elucidate the reaction kinetic of cyclohexane combustion over Co/AC catalysts.Cobalt catalyst is one of the most promising catalysts for VOCs purification in air, which is far more economical than noble metal catalysts, offering the possibility to perform the reaction at low temperatures. This article is aimed to provide a comprehensive overview on the catalytic oxidation of VOCs over the past few years, and the advantage of different Co-based catalysts was also summarized. Besides, the kinetic models of VOCs elimination over cobalt catalyst were all-sidedly discussed and summarized based on corresponding researches, which is beneficial for the discovery of VOCs theoretical degradation process over Co-based catalysts.Concerning single cobalt oxides, based on the present research, it can be concluded that the catalytic performance towards VOCs oxidation is strongly related to the oxygen vacancy density, bulk oxygen mobility, the exposure of reactive (110) facets, the concentration of surface oxygen species and active Co3+ sites, three-dimensional structure, as well as the specific morphologies.Regarding Co-based composites, a conclusion can be drawn that the preparation methods, the intrinsic nature of other components, and the application of surface decoration strategies seem to affect the interaction between cobalt oxide and other components, regulating the structural and textural properties of the catalyst. In specific, this interaction works as a function of adjusting the redox properties and constructing surface defects of the composites, contributing to significant improvement of catalytic efficiency. However, the internal mechanism of the interaction has not been completely addressed in recent studies. Hence, more work needs to be done in the future to understand more about the effect of interactions in Co-based composites.From the current results, it can be inferred that although much progress has been made on the elimination of different VOCs, but there are still lots of problems that need to be solved to meet the emission standards of VOCs. As concerns future perspectives, we thought the research direction of Co-based catalyst may have the following aspects:\n\n(1)\nTo investigate the preparation method that satisfies the characteristics of simple methods, accurately controllable, large preparation scale, etc.\n\n\n(2)\nDesigning efficient catalysts with highly dispersed active sites, abundant oxygen vacancies, highly exposed crystal planes, specific structures and morphologies.\n\n\n(3)\nExploring the employment of cobalt oxides to support single-atom noble metal seems to be a promising way to reducing the cost of the noble metal catalyst, as well as improving the catalytic activity.\n\n\n(4)\nDeveloping simple and efficient surface modification strategies to further improve the performance of Co-based catalysts.\n\n\n(5)\nDeriving a greater understanding of the mechanism of interaction among three or more component catalytic systems, to establish the correlation between the interaction and its catalytic behavior.\n\n\n(6)\nExploiting new technologies or integrating existing strategies, such as photothermal catalytic oxidation, non-thermal plasma catalytic oxidation, pre-adsorption-catalytic oxidation, to improve the catalytic efficiency of VOCs purification, simultaneously, reducing the cost of this reaction.\n\n\n(7)\nDeveloping efficient Co-based catalyst towards the degradation of multiple VOCs mixtures, furthermore, to reveal the different catalytic behaviors of the catalyst in the oxidation of single VOC and VOCs mixtures.\n\n\nTo investigate the preparation method that satisfies the characteristics of simple methods, accurately controllable, large preparation scale, etc.Designing efficient catalysts with highly dispersed active sites, abundant oxygen vacancies, highly exposed crystal planes, specific structures and morphologies.Exploring the employment of cobalt oxides to support single-atom noble metal seems to be a promising way to reducing the cost of the noble metal catalyst, as well as improving the catalytic activity.Developing simple and efficient surface modification strategies to further improve the performance of Co-based catalysts.Deriving a greater understanding of the mechanism of interaction among three or more component catalytic systems, to establish the correlation between the interaction and its catalytic behavior.Exploiting new technologies or integrating existing strategies, such as photothermal catalytic oxidation, non-thermal plasma catalytic oxidation, pre-adsorption-catalytic oxidation, to improve the catalytic efficiency of VOCs purification, simultaneously, reducing the cost of this reaction.Developing efficient Co-based catalyst towards the degradation of multiple VOCs mixtures, furthermore, to reveal the different catalytic behaviors of the catalyst in the oxidation of single VOC and VOCs mixtures.The authors declare no competing financial interest.This work is supported by the National Natural Science Foundation of China (No.21872096), the Educational Department of Liaoning Province (LZ2019002, LQ2020011)", "descript": "\n                  As the main contributor to air pollution, lots of volatile organic compounds (VOCs) were emitted into the atmosphere due to the rapid urbanization and industrialization, threatening environmental safety and human health. Catalytic oxidation has been verified as an efficient approach for VOCs elimination from industrial waste gas streams. Owing to the merits of cost-effective and high activity, cobalt-based catalysts have been considered as one of the most promising candidates for VOCs degradation. This review systematically summarized the developments achieved in the design of cobalt-based catalysts for VOCs removal over the past decade. Specifically, the fabrication of single cobalt oxides, cobalt-based binary oxides and cobalt-based composites, as well as the modified cobalt-based oxides by the surface engineering strategies, such as doping technology and acid etching method are coherently reviewed. Subsequently, the corresponding kinetic models and mechanisms are also discussed. Finally, considering the enormous challenges and opportunities in this field, the perspective with respect to future research on cobalt-based catalysts is proposed.\n               "}
{"full_text": "Data will be made available on request.The acceleration of industrialization and excessive use of carbon-rich fossil fuel, oil, coal, and natural gases, the atmospheric CO2 concentration has reached an unprecedented high level (420\u00a0ppm), which results in global warming [1]. The advice of climate change experts on stabilizing surface temperature escalation below 2\u00a0\u00b0C compared to the preindustrial level in the 21st century to increase the CO2 emission approximately 450\u00a0ppm by 2100 represents the most stringent Representative Concentration Pathway (RCP) [2,3]. As a result, the environment and energy have emerged as the most pressing challenges of the 21st century. Due to this, researchers focus shifted to environmental mitigation and increasing the demand for greenhouse gas (CO2) utilization to value-added chemicals and fuels production [4,5]. However, the abundantly available CO2 can be used as a feedstock material to produce different kinds of important chemicals like formic acid, urea, methanol, and ethers. Atmospheric CO2 adsorbs and is used for valuable feedstock chemical synthesis [6]. Unfortunately, CO2 is a very rigid molecule that is kinetically and thermodynamically stable, with high bond energy (806\u00a0kJ\u00a0mol\u22121). This led to a big challenge encountered in CO2 valorization [7]. Exploring and developing a heterogeneous catalyst for efficient CO2 upgrading have become one of the most active fields in catalysis [8]. In this context, significant progress in CO2 utilization has been made in recent years, yielding a variety of products such as urea, carbonate, salicylic acid, and polyols [9\u201311]. Carbon Recycling International (CRI) recently began the first commercial demonstration plant of methanol synthesis by using CO2 as a row material [12].Among all products, Formic acid (FA) is an important industrial product that can be obtained by reducing CO2 via various catalytic routes, including thermochemical, photocatalytic, and electrochemical reduction [13\u201315]. FA is mainly used in textiles, cleaning, preservatives, hydrogen storage, agriculture, pharmaceutics, and food additives [16,17]. Currently, FA is synthesized in the industry by using two steps: the first step involves carbon monoxide reacting with methanol to form methyl formate, which is then converted into formic acid through acidification with H2SO4 or hydrolysis with water. The hydrolysis step requires an excess amount of water and increases the production cost of formic acid. Formic acid is also produced as a by-product during the liquid phase oxidation of hydrocarbons to acetic acid. Nonetheless, this is not a atom-economy as well as not used renewable carbon feedstocks [18].The number of homogeneous catalysts has been reported for CO2 hydrogenation to FA synthesis. For instance, Inoue et al. used triphenylphosphine (PPh3), containing Rh, Ru, and Ir complexes for catalytic hydrogenation of CO2 to formic acid with high yield [19]. Yoon and research groups worked on developing iridium-containing homogeneous complexes and organic polymer ligands as a catalyst for CO2 to formate synthesis, achieving the highest 40,000\u00a0h\n1 TOF with excellent selectivity [20,21]. Working on the same topic, Ertem et al. developed an amine-based iridium\u00a0containing catalyst for CO2 hydrogenation to formate synthesis using a very low reaction temperature (25\u00a0\u00b0C) and gas pressure (0.1\u00a0MPa) H2/CO2 with achieving 198\u00a0h\u22121 TOF. The same report shows the reverse reaction of formic acid to H2 and CO2 generation with the same catalyst at 60\u00a0\u00b0C with good TON and TOF 118000\u00a0h\u22121\n[22]. Also, a variety of transition metals, such as Pd, Ni, Cu, Fe, containing complexes with C-, N-, phosphorous ligand, N, N-chelated ligand, N-Heterocyclic carbine ligand (NHC), pincer complexes were also reported [23]. However, these catalysts have several drawbacks, including high energy requirements, high cost, catalyst separation from the reaction mixture, and other environmental issues [14].Keeping atom economy in mind, heterogeneous catalysts are the preferred choice for producing formic acid from renewable CO2. Thus, various mixed metal oxide (MMOs) catalysts such as Al2O3, MgO, CeO2, TiO2, ZnO, Cu2O, Ag2O, Co3O4, PbO, and SnOx have been used for the hydrogenation of CO2 to different chemicals [24\u201326]. Particularly, Mori and co-workers have used the impregnation method to synthesize PdAg nanoparticles supported on TiO2 for high-pressure hydrogenation of CO2 to formic acid, obtained >\u00a099% selectivity of FA, with 748\u00a0h\u22121 TON [27]. Also, PdAg supported on a hydrophilic N-doped polymer-silica was used for the same reaction at mild reaction conditions [28]. The Pd/ZnO catalyst demonstrated CO2 hydrogenation to FA under base-free conditions and found that the crystal plane of ZnO plays a vital role in the catalytic activity [29]. Pandey et al. synthesized Cu dispersed on a TiO2 catalyst using a typical co-precipitation method and used for hydrogenation of CO2, achieving the highest TON 6\u00a0h\u22121\n[30]. Furthermore, Chiang et al. investigated formic acid synthesis in a fixed bed reactor system using traditional CuZnO/Al2O3 and obtained 13.1% conversion of CO2 with 7.6% yield of formic acid [31].The literature reports indicate that the CO2 reduction reaction is still in its early stages, and a robust catalyst is required, which can be easily used in industrial production. Herein, we demonstrated that a lower percentage of Ir incorporation on Co3O4 and their catalytic application indirect hydrogenation of CO2 to HCOOH. The physicochemical properties of the synthesized composites were analyzed using analytical techniques to understand better the iridium role, morphology, and textural properties of the composites in the catalytic conversion of CO2 to formic acid.IrCl3.xH2O (99% Aldrich), Co(NO3)2.xH2O (99.5%, SRL India), Ammonia solution 25% (Finar, India). NaOH (96% SDFCL), N,N,N\u2032,N\u2032-tetramethylethane-1, 2-diamine (TEMDA >98%), KOH (85% SRL), Triethylamine (99.5% Loba Chemie), H2 and CO2 gas received from Raj sons Bhavnagar, India and used as received. A double distilled water prepared in the laboratory is used throughout the experiments.The different wt% of Ir on Co3O4 oxide were synthesized using co-precipitation, followed by the hydrothermal method. In a typical synthesis of 1\u00a0wt% Ir-Co3O4, 8\u00a0g of Co(NO3)2.6H2O was dissolved in 60\u00a0mL distilled water and stirred for 15\u00a0min at room temperature before adding 0.102\u00a0g of IrCl3.xH2O salt was added and stirred until complete dissolution of metal salt. Then 12.1\u00a0mL NaOH solution (6.3\u00a0M) was added dropwise as a co-precipitating agent and maintained the reaction pH between 9 and 10. The reaction mixture was then stirred at room temperature for 24\u00a0h. Then, the reaction mixture was transferred to a 100\u00a0mL Teflon-lined stainless-steel hydrothermal autoclave and placed in an oven at 120\u00a0\u00b0C for 5\u00a0h. Then the reactor was cooled to room temperature and the solid product was collected using simple filtration and washed several times with distilled water and methanol until the filtrate reached neutral pH. The obtained solid was dried in a hot air oven at 70\u00a0\u00b0C overnight before being calcined in a muffle furnace at 500\u00a0\u00b0C with 5\u00a0\u00b0C/min ramp rate for 5\u00a0h. The synthesized catalysts are donated as Ir-Co3O4-W, where W standards for wt% of Ir.The catalytic hydrogenation of CO2 into formic acid was carried out in a (300\u00a0mL Amar high-pressure reactor). The reactor was charged with 100\u00a0mL distilled water as a solvent, 0.2\u00a0g freshly prepared catalyst, and 10\u00a0mL\u00a0N,N,N\u2032,N\u2032-tetramethylethane-1, 2-diamine (TEMDA) was used as a base. Then the vessel was tightly closed and purged with N2 gas three times before the reactor was pressurized by CO2 and H2 gas (1:1) up to 62 bars. The reactor was heated using a heating mental under slow stirring. The reaction was started by increasing the agitator speed and continue for 6\u00a0h at 120\u00a0\u00b0C. When pressure ceasing happened, the reaction stopped, and the reactor was cooled to room temperature to ambient conditions. The catalyst was filtered out, and the product solution was analyzed using high-performance liquid chromatography (JASCO, CO-2060 Plus, Intelligent column thermostat, MD-2015 Multiwavelength Detector, PU-2089 Quaternary Gradient Pump). The product analysis was done using the sSupelcogel C-610\u00a0H column. The 0.1\u00a0N\u00a0H3PO4 as a mobile phase was passed through the column at a 0.6\u00a0mL/min flow rate. The samples were analyzed by selecting a 210\u00a0nm UV detector wavelength. The quantification of formic acid was carried out by plotting the standard calibration curve of formic acid by analyzing the formic acid solution of different known concentrations (2.5\u2013400\u00a0mmol) in 0.67 molar TEMDA solution). Then TON and TOF are calculated based on metal loading obtained from the ICP result.The different Ir wt% loaded on Co3O4 catalysts were prepared by co-precipitation, followed by the hydrothermal method (\nScheme 1). Iridium chloride and cobalt nitrates are used as Ir and Co precursors to form an Ir-Co3O4 composite using the appropriate precipitation agent (NaOH) during continuous stirring. The reaction was carried out in a basic aqueous solution and resulted in the formation of a hydrated complex of both metal salts. The hydrothermal treatment was used to convert this aqua complex to crystal oxide because the hydrothermal process generates autogenous pressure and produces monomer, followed by nucleation and crystal growth during the calcination process at higher temperature (\nScheme 2) [32].The powder X-ray diffraction method was used to determine the purity of the formed mixed metal oxide Ir-Co3O4 phase after calcination (\nFig. 1). The sharp diffraction peak at 2\u03b8=\u200919\u00b0, 31.3\u00b0, 37\u00b0, 39\u00b0, 44.9\u00b0, 48.3\u00b0, 55.7\u00b0, 59.5\u00b0, 65.4\u00b0, and 77\u00b0, which are indexed at corresponding diffraction plane (1 1 1), (2 2 0), (3 1 1), (2 2 2), (4 0 0), (3 3 1), (4 2 2), (5 1 1), (4 4 0), and (5 3 3) of cubic phase of Co3O4 and are good agreement with JCPDS No. 42\u20131467 [33]. The diffraction peaks of iridium oxide were not observed due to very low wt% loading of IrCl3.xH2O salt on cobalt precursor, the spinal structure of Co3O4 selected the major plane (3 1 1) to determine the crystal size 6\u2009nm and an average crystalline size of 1\u2009wt% Ir-Co3O4 system is 1.88\u2009nm determined by Debye Scherer\u2019s equation (See in ESI S2).The surface elemental composition and chemical state of synthesized 1\u2009wt% Ir-Co3O4 catalyst was studied by X-ray photoelectron spectroscopy (XPS), as shown in \nFig. 2. Fig. 2 (A) survey spectrum confirmed the presence of Ir 4f, Co 2p, and O 1s species in the synthesized catalyst. The deconvoluted XPS spectra of Co 2p could split into two distinct peaks located with spin orbits at Co 2p3/2 and Co 2p1/2 and binding energy at 779.7 and 195.1\u2009eV, respectively. The splitting characteristic of Co 2p3/2 and Co 2p1/2 indicate the presence of divalent and trivalent Co species in the cubic spinal structure of Co3O4. The fitting peak at 779.7\u2009eV is a spin-orbit doublet of Co 2p3/2 deconvoluted into two peaks at 779.7\u2009eV (FWHM 1.32) 781.3\u2009eV (FWHM 2.4) which are assigned to Co3+ 2p3/2 and Co2+ 2p3/2 configurations respectively. While the shoulder peak of Co 2p1/2 at a binding energy of 795.1\u2009eV could split into two distinct peaks at 795.1\u2009eV (FWHM 1.8) and 797.0\u2009eV (FWHM 2.5), these are attributed to the 2p1/2 spin-orbits of Co3+ and Co2+ respectively [34]. The typical two small shakeup satellite peaks of Co 2p3/2 are located at 790.0 (FWHM 2.1), and 804.5 (FWHM 2.5) eV are conformed to Co3+ and Co2+ species are present in Fig. 2 (B) spectra [35]. The Co2+ and Co3+ cations are found in Co3O4, with Co2+ being tetrahedrally coordinated ions and Co3+ being octahedrally coordinated cations [36]. Fig. 2 (C) shows that Ir 4\u2009f high-resolution spectra are resolved in single doublet components of Ir 4\u2009f7/2, indicating that Ir exists in two chemical states. The lower energy spin-orbit splitting Ir 4\u2009f7/2 could be deconvolute into two distinct peaks with binding energies of 60.5 (FWHM 1.80) and 62.2\u2009eV (FWHM 1.89), indicating that the iridium is present in the Ir+3 and Ir+4 chemical state. While the shoulder peak located with spin-orbit Ir 4\u2009f5/2 at the binding energy 65.1\u2009eV (FWHM 1.87) ascribed to the Ir+4 state present in the material. This is derived from the Ir+4 state in the form of IrO2 on the Co3O4\nsupporting material\n[37,38]. Fig. 2 (D) depicts the XPS spectra O\u00a01s of the Ir/Co3O4 system and split into two peaks where the lower binding energy at 530\u2009eV could correspond to the lattice oxygen O2-, and the higher binding at 531.8\u2009eV can be attributed to the adsorped oxygen [39]. The considerable difference between O2- affections with the Co ion in the cubic spinal Co3O4 system may be single O three coordinated with three Co3+ and two coordinated O bonded with one Co2+ and Co3+\n[40]. According to the results of the XPS study, the spinal structure of Co3O4 has more adsorption oxygen vacancies, as well as Ir+3 and Ir+4 species, and these results are in excellent agreement with surface area and pore size.N2 adsorption/desorption at liquid nitrogen temperature was used to determine the BET surface area and pore diameter of synthesized Ir-Co3O4 catalysts (See ESI Fig. S4). Catalysts were degassed at 300 for 4\u2009h before analysis. The surface area of pristine Co3O4 was found to be reduced from 16\u2009m2/g to 7.2\u2009m2/g when 0.5\u2009wt% iridium was added to Co3O4 but increased to 47.7\u2009m2/g when 2\u2009wt% iridium was added. In terms of pore size, the pristine Co3O4 has a pore diameter that was slightly reduced with the addition of Ir up to 1.25\u2009wt%. When 2\u2009wt% of Ir was incorporated into Co3O4, the pore diameter dramatically reduced to 388\u2009\u00c5; this indicates that the pores of the Co3O4 were occupied by iridium up to 1.25\u2009wt%. Further increasing the amount of iridium, the cluster of iridium formed on the surface of Co3O4. The results of BET surface area analysis and pore size distribution are summarised in \nTable 1.The FE-SEM images of simple Co3O4 and 1\u2009wt% of Ir-Co3O4 catalyst (\nFig. 3 A and B) show no distinct morphology. However, the elemental mapping displayed the uniform distribution of Ir contents on the Co3O4 surface (Fig. 3 C-F) with 1\u2009wt% of Ir content was confirmed by SEM EDX (See ESI Fig. S2). The presence of the Ir and Co on 1\u2009wt% Ir-Co3O4 were identified using the lattice fringes value shown in Fig. 3 (G) and (H). The lattice fringes value obtained through the SAED pattern 1 wt% Ir-Co3O4 catalysts were identical to the X-ray pattern as shown in (Fig. 1, JCPDS No. 42\u20131467). The composited solid crystalline sample reveals the various lattice fringes depicted in the selected different zones in Fig. 3 (G and H). The majority of separated lattice fringes are shown in Fig. 3 (G) and (H) 0.464, 0.24, 0.23, 0.28\u2009nm correspond to (1 1 1), (3 1 1), (2 2 2), and (2 0 0) lattice indices, respectively, and exactly match with the XRD result d\u2010spacing of spinal cubic phase Co3O4 (JCPDS No. 42\u20131467). The active metal Ir oxide fringes are displayed in Fig. 3 (G) 0.258\u2009nm correlate with (2 0 0) plane of Ir crystal and conform to the d\u2010spacing with (JCPDS 04\u2013009\u20138479). The SEAD pattern also confirmed the plane, or d\u2010spacing value of Ir-Co3O4, see Fig. 3 (I).The surface acidity of 1\u2009wt% Ir-Co3O4 oxide catalyst was quantified using temperature-programmed desorption (TPD) with NH3 and the results are shown in \nFig. 4. Further ammonia adsorption was carried out at the heating rate of 10\u2009\u00b0C/min by flowing 5% NH3/He stream (30\u2009mL/min) from temperature 200\u2013800\u2009\u00b0C. The TPD result of the Ir-Co3O4 oxide catalyst indicates a broad distribution of strong acidic sites in the catalyst. Typically, acidic strength is determined by NH3 adsorption above >\u2009400\u2009\u00b0C [41]. Fig. 4 depicts the profile of NH3 TPD adsorption on the Ir-Co3O4 oxide catalyst. NH3 desorption shows two distinct regions, with a peak around 450\u2013635\u2009\u00b0C indicating a strong acidic site [42,43]. The quantitative analysis of NH3-TPD desorption of lower loaded Ir-Co3O4 oxide catalyst is estimated to be 0.963\u2009mmol/g acidic sites are present in the material. The series of Ir metal loaded on the Co3O4 surface was confirmed by the Inductively Coupled Plasma Mass Spectrometry (ICP-MS). The quantitative Ir concentration in the composite is quite analogous to the theoretical value of Ir, hence used for further study and determining the TON and TOF based on this result.To explore the hydrogenation of CO2 (\nScheme 3), initially, the hydrogenation reaction operates in the absence of a catalyst and base, there was no CO2 conversion was observed. (\n\nTable 2, entry 1). Furthermore, in the presence of a base (TEMDA) without catalyst addition leads to no conversion of CO2. The results concluded that the hydrogenation of CO2 to formic acid required a catalyst to activate both CO2 and H2.The reaction was then carried out using only the supporting material Co3O4 as a base catalyst gave 7\u2009mmol of FA. When a series of Ir-Co3O4 catalysts were used, initially, 0.5\u2009wt% of Ir-Co3O4 catalyst gives 360.9\u2009mmol of formic acid (Table 2, entry 4). Followed by 1 and 1.25\u2009wt% Ir-Co3O4 catalyst gives excellent yield of formic acid 399 and 400\u2009mmol, respectively (Table 2, entries 5, 6). However, no significant difference was observed when the reaction was carried out with a 2\u2009wt% Ir-loaded catalyst (Table 2, entry 7).Furthermore, different types of bases were used with 1\u2009wt% Ir-Co3O4 catalyst for hydrogenation of CO2 to formic acid. Initially, the hydrogenation reaction of CO2 does not show any catalytic activity without the addition of base (\nTable 4, entry 1). The experiment concluded that base is necessary for the conversion of CO2 into other products. Further, the reactor was pressurized by an equal amount of CO2 and H2 (1:1) with 0.1 molar concentration of alkali bases like NaOH and KOH giving 5 and 73\u2009mmol of formic acid. Additionally, compared the different amine-containing bases for CO2 hydrogenation. When 1\u2009mole NH3 was used as a base 11\u2009mmol of formic acid was observed (Table 4, entry 4). Then, investigated the activity of 0.6 molar piperidine, which provided 19\u2009mmol yields of formic acid (Table 4, entry 5). The majority of the researchers used tertiary amine as a base for the synthesis of formic acid from CO2 because it efficiently coordinated with CO2 and converted to formic acid or formate. The 0.6 molar concentration of triethylamine yields 16\u2009mmol of formic acid (Table 4, entry 6). When use N,N,N\u2032,N\u2032-tetramethylethane-1,2-diamine (TEMDA) as a base, the yield of formic acid (399\u2009mmol) increased 25 times better than triethylamine (Table 4, entry 7).The higher catalytic activity of TEMDA, may be due to the presence of two active sites to co-ordination with CO2 and less hindered amine base. It can easily coordinate with CO2 molecules and form carbamate zwitterion intermediate species. The comparative results of various Ir based catalysts with Ir-Co3O4 are shown in Table 3 Most of the articles used trimethylamine and K2CO3 as a base, giving less formate yield than TMEDA. As a result, N,N,N\u2032,N\u2032-tetramethylethane-1,2-diamine base was chosen for further optimization study. The effect of temperature on the hydrogenation of CO2 to formic acid synthesis was investigated in the range of 80\u2013140\u2009\u00b0C (See \nFig. 5 (A). The formic acid yield suggests that at the lower reaction temperature, the rate of CO2 conversion was also slow; the highest conversion of CO2 was obtained at 120\u2009\u00b0C. When the reaction temperature was raised above 120\u2009\u00b0C, the yield of CO2 hydrogenation did not increase. The reaction conducted at 140\u2009\u00b0C, yield of formic acid was the same 397\u2009mmol.The highest catalytic activity was observed in 0.67\u2009M concentration of TEMDA; when decreasing the concentration of TEMDA to 0.33\u2009M, the formic acid yield suddenly decreased to 210\u2009mmol (\nTable 5, entry 1). Further, increasing the concretion of TEMDA, there are no dramatic changes observed in the yield of formic acid, and 0.67\u2009M TEMDA concluded for the further optimation studies (Table 5, entries 2\u20134).\nFig. 5 shows how the rate of CO2 hydrogenation to selectively formic acid synthesis correlates with time (B). The CO2 conversion profile shows that the initial CO2 conversion rate was very high in one hour, yielding 281\u2009mmol of formic acid and that the formic acid yield increased steadily up to 6\u2009h, yielding 384\u2009mmol. The CO2 to formic acid conversion rate was very slow after a 6\u201312\u2009h reaction. The time variation study reveals that 6\u2009h is enough to get the maximum yield of formic acid.Further, extend the optimization study of CO2 hydrogenation effect of gas ratio variation. To rule out the hydrogen source, CO2 hydrogenation was performed without H2 pressure and no conversion was obtained. This experiment confirms Ir-Co3O4 catalyst does not produce hydrogen from water (\nTable 6, entry 1). As a result, a 62\u2009bar CO2 pressure reaction also does not form formic acid or formate. When the reaction was carried out at a lower pressure, 7\u2009bar CO2 and H2 gas respectively obtained 213\u2009mmol yields of formic acid (Table 6, entry 2), when the pressure was increased to 21\u2009bar CO2 and 41\u2009bar\u2009H2 gas, 391\u2009mmol of formic acid was obtained (Table 6, entry 3). Further, reduce of the H2 gas pressure to 21\u2009bar and increasing the CO2 pressure to 41\u2009bar also decreases the formic acid yield to 320\u2009mmol (Table 6, entry 4). Then using a 31\u2009bar (1:1) ratio of CO2 and H2 gases for the further subsequent reaction was obtained as a higher yield of formic acid 398.5\u2009mmol (Table 6, entry 5). Further, increase the pressure of CO2 and H2 in equal amounts, but the yield of formic acid remains the same (398\u2009mmol) (Table 6, entry 6). According to the gas ratio variation study, the synthesis of formic acid from CO2 is a temperature and pressure-dependent reaction.After each reaction, the catalyst was successfully recovered and washed with methanol before being dried at 200\u2009\u00b0C for 2\u2009h. The catalyst was used for the recycling experiment, as shown in \nFig. 6. Initially, the first cycle catalyst yielded 403\u2009mmol of FA, but the second and third cycles yielded only 8 and 13\u2009mmol of formic acid yield decrease. Even after the fifth catalytic cycle, the 1\u2009wt% Ir-Co3O4 catalyst demonstrated excellent catalytic activity with 372\u2009mmol formic acid in 6\u2009h, indicating the structural and catalytic stability of the Ir-Co3O4 metal oxide catalyst. The post catalyst was characterized by XPS analysis to understand the chemical changes of Ir and Co (see S7). No changes were observed in the chemical state of Ir, but in the case of Co the formation of CoO was observed; this may be due to the decomposition of Co3O4. At the same time, the decomposition of Co3O4 may be the reason for the slightly loss in the yield of FA.The formate ion formation was confirmed using 1H and 13C NMR, FT-IR, and quantified by HPLC. The experiment was conducted on NMR 600\u2009MHz using D2O as an NMR solvent, which appeared peak at 4.8\u2009ppm. The standard formic acid 1H proton appeared at 8.13\u2009ppm, which acidic formic proton does not show due to hydrogen bonding with water and NMR solvent. Then the confirmation of the mixture of TEMDA and formic acid shows a 1H proton peak at 8.38\u2009ppm because the base receives the acidic proton from FA and form formate, which causes the change in proton value shift from 8.13 to 8.38, indicating that it is the formate ion. The subsequent reaction was carried out in D2O as a reaction solvent; in this case, the source of transferred hydrogen to reduce CO2 by gases molecular hydrogen transfer from the metal surface rather than the solvent is shown in \nFig. 7. (A). Fig. 7 (B) shows the FT-IR spectra and detects the reaction product formate ion instead of FA due to the experiment being performed in basic media, TEMDA used as a base. Initially, distilled water showed a broad band at the position 1645\u2009cm\u22121. Then only TEMDA showed a significantly low intense band as compared to other samples. Fig. 7 (B) depicts the FT-IR spectra of a standard mixture of TEMDA and FA, in which TEMDA reacts with FA to form tetramethyethylendamine formate, but formic acid did not completely consume with base, resulting in a spectrum that shows the formic acid band at (1214\u2009cm\u22121) as well as formate ion band 1351 and 1384\u2009cm\u22121. In the reaction mixture, formate ions were observed through their most intense band located at 1351, 1384 and 1590\u2009cm\u22121, respectively [50\u201352]. The standard formic acid stretching band is located at different positions 1216 and 1718\u2009cm\u22121 than reaction mixture spectra. The FT-IR study concludes the presence of formate ions in the reaction mixture.Based on the characterization results and the combination of iridium and cobalt oxide catalytic performance, herein proposed a plausible mechanism of the catalytic route for the hydrogenation of CO2 to the formic acid formation (\nScheme 4). The initial step would be for gaseous CO2 molecules to dissolve in water and convert into carbonic acid in the reaction medium. When the reaction temperature is raised to 120\u2009\u00b0C, the base TEMDA activates the CO2 molecule to form the carbamate zwitterion intermediate. As a Lewis acid, CO2 can easily co-ordinate with Lewis base (TEMDA) to form the carbamate intermediate. The formed carbamate zwitterion intermediate is very unstable and easily reacts with other species. Simultaneously, the active iridium nanoparticle participates in the activation of the H2 molecule on the catalyst surface and dissociates into the activated H species [53]. Particularly, Ir nanoparticles facilitated delocalized electron transfer to the vacant orbital of CO2 and promoted the CO2 hydrogenation.It can provide a more negative hydride, resulting in a more reactive nucleophilic attack on the electrophilic CO2 molecule [54]. The hydride (-H) is then transferred from the Ir interface side to the carbamate zwitterion, hydrogenated, and converted into a formate product, and upon acidification, it yields formic acid.In summary, the lower wt% of Ir loaded on traditional cobalt oxide demonstrated for the hydrogenation of CO2 to formate synthesis. The Ir/Co3O4 based mixed metal oxide catalyst was synthesized by typical co-precipitation followed by the hydrothermal method. The use of noncarbonated sources to directly hydrogenate CO2 for formic acid synthesis using TEMDA as a base over a spinal Ir-Co3O4 oxide catalyst. The synthesized cubic spinal Ir-Co3O4 catalysts are confirmed by various analytical tools. The Ir-Co3O4 catalyst selectively converted CO2 to formic acid via catalytic hydrogenation and obtained an excellent yield of (399\u2009mmol) and TON 1916\u2009h\u22121. While revealing excellent recyclability and producing the highest final formate concentration of 399\u2009mmol within 6\u2009h at milder reaction conditions, the formation of formic acid from CO2 was confirmed by NMR, FT-IR and quantified by HPLC. Its promising approach for direct CO2 hydrogenation and significant implications in the field of CO2 conversion chemistry are encouraged by the results of excellent catalytic performance, durability, simplicity, and high-pressure stability.\nBalasaheb D. Bankar: Conceptualization, Investigation, Writing \u2013 original draft, Data curation, Formal analysis. Krishnan Ravi: Data curation. Rajesh J. Tayade: Data curation. Ankush V. Biradar: Conceptualization.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.CSMCRI communication No. 209/2022. B. D. Bankar acknowledges to UGC government of India for the senior research fellowship. Dr. A. V. Biradar acknowledges MLP 0028, HCP 0009 CSIR India for the financial support. Also, Dr. P. S. Subramanian for discussion and encouragement. The analytical division provides the centralized instrumentation facility with all requisite instrumental analysis of CSIR- Central Salt and Marine Chemicals and Research Institute, Bhavnagar.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jcou.2022.102315.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  The utilization of abundantly available CO2 feedstock for the synthesis of high-value compounds and fuels is the main focus of catalysis research. Herein, we report the synthesis of different wt% Ir-Co3O4 oxide as a heterogeneous catalyst prepared by co-precipitation, followed by the hydrothermal method. The P-XRD analysis revealed the formation of a cubic phase with a high intense (3 1 1) plane with a 6\u00a0nm crystal size. The synthesized catalyst was used for direct hydrogenation of CO2 assisted by N,N,N',N' tetramethylenylenediamine. Under optimized reaction conditions, the 2.9\u00a0wt% Ir-Co3O4 catalyst demonstrated an excellent yield of formic acid (399\u00a0mmol). The formic acid yield in N,N,N',N; tetramethylenylenediamine is 25 times better than traditionally used trimethylamine. The outstanding performance of the Ir-Co3O4 catalyst was due to the stoichiometric amount of active Ir content uniformly distributed on Co3O4 support with a mixed Ir+3/4 oxidation state, which readily exchanges electrons between Co3O4 and Ir during the reaction. The catalyst was successfully recycled up to five times with negligible loss in the yield of formic acid.\n               "}
{"full_text": "Data will be made available on request.Climate change has emerged as one of the most pressing issues in politics and society as a whole. The emission of greenhouse gases, mostly CO2, has reached its highest level in human history and is a significant contributor to global warming [1,2]. A significant fraction of these emissions originate from the chemical industries which either depend directly on or utilize products from fossil fuels [3]. Moreover, demand for petrochemical products by the chemical industries is expected to increase in the coming years. However, some of these chemical intermediates can be produced by utilizing CO gas. For example, the hydrolysis of methyl formate to produce formic acid and the catalytic carbonylation of methanol to produce acetic acid are some of the vital chemical processes that utilize CO as a building block [4]. As a result, rather than using fossil fuels in these processes, green technologies such as solid oxide electrolysis cells (SOECs) could be utilized to produce CO through CO2 electrolysis. CO2 gases from different industrial outputs could be electrochemically reduced to CO thereby reducing the overall industrial carbon emission [5,6].Different competing green technologies are currently being investigated for CO production. Among the electrochemical CO2 reduction possibilities, three technologies easily stand out; low temperature electrolysis, molten carbonate electrolysis and the SOECs. Of the three technologies, the SOEC is the most advanced technology for CO2 electrolysis with years of operational hours [4]. In low temperature electrolysis, achieving high selectivity towards CO production is a non-trivial issue. Consequently, scarce catalysts such as IrO2 and noble cathodes (Au and Ag) must be utilized. On the other hand, in molten carbonate electrochemical cell, the rapid corrosion of the electrolysis cell container remains a major challenge [4,7,8]. The SOEC technology, however, presents higher performance efficiencies at an industrially relevant scale. Therefore, with respect to the different indices of comparison such as the operating efficiency, faradaic efficiency, cell voltage, and area-specific resistance, the SOEC outperforms the molten carbonate and the low temperature electrolysis modes [4,7,8].Despite the high efficiencies in SOECs, significant degradation, especially on the fuel electrode, has been observed. For example, considering the state-of-the-art Ni-YSZ fuel electrode materials, significant electrode degradation due to Ni migration and agglomeration has been reported [9\u201313]. Furthermore, noticeable degradation due to carbon deposition is observed when carbon-containing fuels are utilized [14\u201316]. Vanesa et al. [14] have investigated the formation of carbon on a Ni-YSZ electrode operating in fuel cell mode with a fuel mixture of 75% CO and 25% H2 at 1073\u00a0K. They reported a pronounced carbon deposition on the Ni-YSZ electrode, resulting in an increase in polarization resistance, decrease in porosity and deactivation of the electrochemical activity. Similarly, He et al. [15] have investigated the extent of carbon formation on a Ni-YSZ pellet after four hours of exposure to humidified methane fuel gas at a temperature range of 773\u20131073\u00a0K. They observed extensive carbon formation on the Ni-YSZ pellet. The deposited carbon was observed to dissolve into the bulk of the Ni particles leading to significant expansion of the Ni-YSZ pellet. Such expansion could lead to delamination and deactivation of the electrode in single cells. Yuefeng et al. [17] demonstrated that the Ni-YSZ fuel electrode is deactivated in pure CO2 electrolysis at 0.9\u00a0V and 700\u00a0\u00b0C. However, a better stable performance was observed at higher operating voltage of 1.3\u00a0V.These issues, i.e carbon deposition and electrode deactivation, could be minimized or even resolved by using electrode materials that have mixed ionic and electronic conducting (MIEC) properties as well as suppress carbon deposition. The MIEC properties ensure that the electrochemical reactions extend beyond the three-phase boundary. Therefore, for these reasons, attention has shifted to ceria containing electrodes [7,15,18]. Under a reducing atmosphere, doped ceria oxide has MIEC properties, exhibiting a mixed-valence of Ce3+ and Ce4+. As a consequence, the electrochemical reaction zone is extended from the three-phase boundary to the entire electrode surface. Furthermore, the electronic conductivity of ceria can partially compensate for Ni agglomeration and depletion thereby reducing the effect on performance\u00a0[7]. With regards to carbon deposition, numerous works in literature have shown that the mixing of fuel electrode cermet with ceria reduces the amount of carbon deposition on the electrode [15,18\u201320]. For instance, He et al. [15] compared the amount of carbon deposition on Ni-YSZ pellet, with and without a ceria catalyst and observed less carbon on the pellet with a ceria catalyst than without a ceria catalyst. In general, the presence of localized electrons and oxygen vacancies in ceria electrodes has been observed to play a non-trivial role in preventing carbon deposition during CO2 reduction [7,18\u201320]. In line with this, complete replacement of the YSZ oxide phase (in Ni-YSZ) with the GDC oxide phase (Ni-GDC) is being pursued.In this work, the high temperature CO2 electrolysis on solid oxide cells (SOC) consisting of a Ni-GDC fuel electrode is examined in detail. The electrochemical activity of the electrode is investigated by using electrochemical impedance spectroscopy (EIS) at different operating conditions. Impedance measurements were obtained at different compositions of CO2 and CO, as well as at different temperatures (750\u2013900\u00a0\u00b0C). Furthermore, a long-term stability test was performed to study the performance and durability of the electrode during operation. Finally, post-test analysis was carried out in order to understand the degradation behavior.For the electrochemical measurements, electrolyte-supported single cells were fabricated. The fuel electrode is made of commercial NiO-Ce0.9Gd0.1O0.95 (GDC) powder from Marion Technologies (NiO: GDC, 65:35\u00a0wt ratio), while the LSCF (La0.58Sr0.4Co0.2Fe0.8O3-\u03b4) oxygen electrode powder was self-synthesized using a modified Pecheni method [21]. To prepare the electrode paste, NiO-GDC powder was mixed in a binder solution comprising 6\u00a0wt% ethyl cellulose (binder) dissolved in \u03b1-terpineol (dispersant). The slurry was then mixed using a planetary vacuum mixer (THINKY Mixer ARV-310) and subsequently homogenized for about 30\u00a0min by roll milling. A similar procedure was used to create the LSCF oxygen electrode slurry. Dense 8YSZ electrolyte supports from Kerafol\u00ae (d=20\u00a0mm, thickness 250\u00a0\u00b5m) were used to create the button cells. A thin layer (4\u20135\u00a0\u00b5m) of GDC was screen printed (EKRA screen printing Technologies) on one side of 8YSZ substrates and sintered at 1350\u00a0\u00b0C for 1\u00a0h under air to form a barrier layer for the oxygen electrode. After that, the fuel electrodes (15\u201318\u00a0\u00b5m) were screen printed on the opposite side of the electrolyte. Five different sintering temperatures were considered: 1150, 1200, 1250, 1300, and 1350\u00a0\u00b0C for 2\u00a0h at a heating rate of 2\u00a0\u00b0C\u2027min\u22121. Based on the polarization resistance, 1200\u00a0\u00b0C for 2\u00a0h was chosen as an optimized sintering condition. The LSCF layer was screen printed on the GDC barrier layer side and subsequently sintered at 1080\u00a0\u00b0C for 3\u00a0h. Finally, a NiO layer screen printed on the fuel electrode side was used as a current collector. The single-cell configuration before reduction is represented by NiO-GDC/8YSZ/GDC/LSCF. Following the same procedure, NiO-YSZ electrode was also fabricated and sintered at 1350\u00a0\u00b0C for 4\u00a0h, which is the optimized sintering condition for this electrode.For the electrochemical measurement, a two-electrode (four-wire) NorEcs Probostat\u2122 set-up was used in the characterization of the single cells [22]. The cell was heated up to 900\u00a0\u00b0C (with 1\u00a0\u00b0C\u2027min\u22121), after which the nickel oxide cermet (NiO-GDC) was gradually reduced to nickel (Ni-GDC) as described by Foit et al. [23]. Following the reduction, IviumStat (Ivium Technologies) potentiostat/galvanostat devices were used to acquire the impedance spectra as well as the current density-voltage characteristics. The frequency range during the impedance measurement was varied from 110\u00a0kHz to 0.11\u00a0Hz with an AC amplitude of 50\u00a0mV and 21 frequencies per decade. Similarly, the current density-voltage (I-V) characteristics were obtained as previously described [23]. The quality of the impedance spectra was analyzed and validated through the Kramers Kronig transformation test [24]. Impedance measurements were taken at OCV under various temperature ranges (750\u2013900\u00a0\u00b0C) and CO2 partial pressures. Long-term stability tests of the button cells were carried out at 900\u00a0\u00b0C with a current density of \u2212\u00a00.5\u00a0A\u2027cm2 for 1070\u00a0h. Impedance spectra were analyzed with both the complex non-linear least-square (NLLS) method as well as the distribution of relaxation times (DRT) transformation. A commercially available NLLS-fit program (RelaxIS\u00ae software, RHD-Instruments) was utilized in the fitting procedure and the DRT transformations.The morphology of the cells was examined with Quanta FEG 650 (FEI\u00a9) scanning electron microscope.Single cells of NiO-GDC were fabricated and analyzed based on their sintering temperature. Five different sintering temperatures were considered; 1150, 1200, 1250, 1300 and 1350\u00a0\u00b0C for 2\u00a0h. However, a tape test was performed on the electrodes showed that the electrodes sintered at 1150\u00a0\u00b0C showed poor adhesion to the electrolyte, hence it was not considered for the rest of the measurement. \nFig. 1a-d shows the SEM images of the cell before the reduction process. The SEM images clearly show an increase in particle agglomeration with the increase in sintering temperature. The cell sintered at 1350\u00a0\u00b0C exhibits the most pronounced particle agglomeration while the 1200\u00a0\u00b0C sintered cell shows the least particle growth. Fig. 1e shows the impedance spectra of the cells obtained at 900\u00a0\u00b0C under OCV conditions. A decrease in polarization resistance (Rp) with decreasing sintering temperature is observed and the lowest Rp is observed for the cell sintered at 1200\u00a0\u00b0C. The result agrees with the microstructural observation from the SEM. An increase in particle agglomeration observed at higher sintering temperatures results in a decrease in the electrode surface area and thus, a decrease in the electrochemical reaction zone. Consequently, an increase in Rp is observed with increased particle agglomeration [25]. Optimized cells, sintered at 1200\u00a0\u00b0C were further used for the electrochemical characterization and the long-term degradation test.To characterize the cell performance, I-V characteristics as well as impedance measurements were compared to those of conventional Ni-YSZ cells. \nFig. 2a compares the I-V characteristics of electrolyte-supported single cells containing Ni-GDC and Ni-YSZ fuel electrodes, respectively. Both single cells were prepared using 8YSZ electrolyte support with an LSCF oxygen electrode and measured in the same test rig. It can be seen that the Ni-GDC electrode containing single cell exhibits a higher current density of \u2212\u20091.16\u2009A\u2009cm\u22122 compared to the Ni-YSZ cell (\u22120.63\u2009A\u2009cm\u22122) at 1.5\u2009V and 900\u2009\u00b0C. Fig. 2b shows the Nyquist plots obtained from the Ni-GDC cell in comparison to that of the Ni-YSZ cell. For the Ni-GDC cell, a lower Rp value of 0.23\u2009\u03a9.cm2 is observed compared to the 0.44\u2009\u03a9.cm2 observed for the Ni-YSZ cell at 900\u2009\u00b0C. The higher performance of the Ni-GDC could be attributed to the enhanced electrochemical properties of the GDC as a result of the MIEC property [26\u201329].Further I-V measurements were obtained for the Ni-GDC cell under varying operating temperatures. Fig. 2c shows the I-V characteristics of the cell as a function of operating temperature (750\u2013900\u2009\u00b0C). The current density increases with an increase in temperature. Such a trend is expected due to the enhancement of electrochemical kinetics at higher operating temperatures. A maximum current density of \u2212\u20091.16\u2009A\u2009cm\u22122 is observed at 1.5\u2009V and 900\u2009\u00b0C. The continuity of the I-V curves across the OCV indicates that the Ni-GDC fuel electrodes can function as reversible SOCs [13]. In most of the cell measurements, the observed open circuit voltage was within 10\u2009mV of the theoretical open circuit voltage according to the Nernst equation, which indicates sufficient cell sealing.To investigate thermally activated processes, impedance spectra were obtained and analyzed at different temperatures from 750\u00b0 to 900\u00b0C in both OCV conditions as well as under polarization. \nFig. 3a illustrates the Nyquist plots as a function of temperature at OCV. Two distinct arcs are easily identified in the impedance spectra; a low and a high frequency arc. While the low frequency arc is relatively unchanged with temperature variation, the high frequency arc shows a pronounced increase in magnitude with the decrease in temperature. This suggests that the high frequency electrochemical processes are thermally activated processes. Similar trend was also recorded under polarization (as shown in supplementary Fig. S1). In general, the decrease in temperature, from 900\u00b0 to 750\u00b0C, caused an increase in the real and imaginary contributions in the Nyquist diagram of the impedance spectra. Such an increase is attributed to the reduction of ionic conductivity and electrochemical reaction kinetics in the electrodes with decreased temperature. The ohmic resistance (Rs) of the cell is determined from the intercept with the real axis at the high frequency in the Nyquist plot. Consequently, the activation enthalpy of the ohmic resistance is calculated from the slope of the Arrhenius equation as represented in Eq. (1) and illustrated in Fig. 3b. The determined value of 61\u2009kJ/mol is consistent with the values of ionic conductivity of the 8YSZ electrolyte [30].\n\n(1)\n\n\nln\nR\n=\n\u2212\nln\n\n\n\u03c3\n\n\n0\n\n\n+\n\n\n\n\nE\n\n\nA\n\n\n\n\n\n\nR\n\n\ng\n\n\nT\n\n\n\n\n\n\nThe impedance spectra were analyzed with both the DRT transformation and the NLLS method. \nFig. 4a shows the DRT representation of the impedance spectra as a function of temperature. Five peaks (P1, P2, P3, P4 and P4a) are observed in the DRT plot within the measured frequency range of 0.11\u2009Hz to 110\u2009kHz. The impedance spectra, however, were modeled with an equivalent circuit consisting of four time constants in series to a resistor and an inductor (LR-RQ-RQ2-RQ3-Ws) as shown in Fig. 4b. Furthermore, a comparison was made between the simulation of the fit and measured data. The comparison shows good agreement between the DRT of the proposed ECM and the DRT of the measured data (Supplementary Fig. S3). Also, the error plot showed non-systematic distribution around the frequency axis, which indicates that the proposed ECM can effectively reproduce the obtained impedance data across the measured frequency range (Supplementary Fig. S3).The time constants RQ1, RQ2 and RQ3 correspond to the processes P1, P2 and P3 on the DRT plot respectively. While the P4 peak corresponds to the infinite length Warburg short element (Ws) with the P4a peak interpreted as the satellite peak of the Ws [31,32]. The DRT reveals a significant dependence of the high/mid frequency peaks on temperature variation. P1, P2 and P3 exhibit an increase in magnitude with the decrease in temperature. On the other hand, P4 is almost independent of temperature variation. The measurements indicate that the mid frequency process (P3) dominates the electrode process at the lower temperature of 750\u2009\u00b0C.The absolute values of the resistances were obtained from the NLLS fitting of the impedance spectra with the equivalent circuit model depicted in Fig. 4b. Fig. 4c illustrates the Arrhenius plot of the determined resistances. The resistances R1, R2 and R3 corresponding to processes P1, P2 and P3 respectively, exhibit a significant increase with a decrease in operating temperature, while Ws is relatively unchanged with the decrease in temperature. The trend shows good agreement with the observed DRT plot. R1 and R3 exhibit high activation energies of 109\u2009\u00b1\u200910\u2009kJ\u2009mol\u22121 and 99\u2009\u00b1\u20092\u2009kJ\u2009mol\u22121, respectively, while R2 shows an activation energy of 77\u2009\u00b1\u200910\u2009kJ\u2009mol\u22121.Considering the electrochemical processes occurring in the electrode, different electrode reaction steps exhibit different temperature dependencies and these dependencies may identify the possible electrode process. For instance, gas diffusion processes exhibit an almost independent temperature dependency while charge transfer processes and processes from the transfer of ionic species show strong temperature dependency and high activation energies [33\u201336].Measurements under different compositions of the CO2 fuel gas were performed to further investigate the fuel electrode processes. The CO2/CO ratio was systematically changed from 90/10\u201350/50. The impedance spectra as well as current-voltage characteristics were obtained as a function of CO2/CO ratio. \nFig. 5a illustrates the Nyquist plots obtained from the variation of CO2 content at OCV. An increase in the amount of CO2 in the fuel gas led to an increase in Rp. The mid and low frequency arcs exhibit a more pronounced dependence, increasing in magnitude with increasing CO2 content. This observation is contrary to what is expected when fuel gas is increased. In fact, in CO2 electrolysis mode at OCV conditions, there is lower electrochemical activity towards CO2 reduction than CO oxidation, hence an increase in Rp is observed with increasing CO2 content. Such observation could be attributed to preferential adsorption or higher activation energy of CO2 desorption on the active catalyst sites of the oxide phase and Ni metal [23,37]. This result is in agreement with a similar experiment by Foit et al. [23] on CO2 electrolysis in Ni-YSZ. They observed that the increase in CO2 content in the fuel gas composition led to an increase in ASR at OCV conditions. However, at higher current densities, increasing the CO2 gas compositions resulted in a decrease in ASR. They opined that at higher current densities, mass transport limitation dominates the overall reaction rate. Hence, decreasing the CO2 content resulted in a lesser amount of fuel gas for reaction leading to a decrease in the electrochemical reaction.The corresponding DRT plots of the impedance spectra are represented in Fig. 5b. Similar to the trend in the Nyquist plot, the low/mid frequency P4 and P3 peaks exhibit pronounced dependence on CO2 variation while P1 and P2 peaks are relatively constant. This suggests that the underlying contributing processes of P4 and P3 peaks are most likely fuel electrode processes while P1 and P2 could be oxygen electrode contributions. Furthermore, the P4 and P3 peaks are observed to exhibit the highest contribution to the electrochemical impedance. Fig. 5c illustrates the obtained resistance from the fitting of the impedance spectra using the equivalent circuit model. The trend is in agreement with the observation in the DRT plot, wherein Ws and R3 resistances show a significant increase with an increase in CO2 content. The significant CO2 content dependence of the low frequency P4 peak as well as the observed independent temperature behavior indicates that this process is most likely a diffusion process [34,35,38]. However, the GDC cermets have been reported to show a low-frequency peak resulting from the chemical capacitance caused by the variation in the oxygen nonstoichiometry of the GDC electrode [38,39]. Therefore, the low frequency P4 peak is attributed mostly to a gas diffusion process and possible contribution from the oxygen nonstoichiometry of the GDC electrode.Considering the frequency regime of the P3 process, reactions at the electrode/electrolyte interface can be ruled out since these processes typically exhibit high relaxation frequencies [33,40\u201342]. With a frequency range between 40\u2009Hz and 250\u2009Hz, this suggests a process towards the electrode sub-surface. Such mid frequency process could be attributed to gas-solid interaction such as adsorption, dissociation and desorption of the gas species [38]. Several authors [7,18,19] have attempted to suggest possible elementary mechanisms of CO2 reduction on ceria containing cermets. Chueh et al. [18] investigated the surface electrochemistry of CO2 reduction and CO oxidation on a ceria cermet. They opined that the electrochemical reduction of CO2 to CO occurs via two single-electron transfer steps, with carbonate ((CO3)2-) formation as an intermediate process. The formed carbonate further absorbs and saturates the electrode surfaces thereby reducing the overall kinetics of CO2 reduction. In general, the carbonate adsorption process is regarded as a major rate-determining step in CO2 reduction [7,18,19]. Therefore, following the variation of temperature and CO2 content, the P3 process has shown to exhibit the highest resistance and hence the rate-determining step in the CO2 electrochemical reduction. Therefore, this process could be inferred to be related to an adsorption process. The mid frequency P3 peak is therefore attributed to a possible surface electrode reaction process (adsorption/desorption) of the gas species in addition to a charge transfer process on the electrode surface.To investigate oxygen electrode processes, measurements under different partial pressures of oxygen (pO2) were performed from 0.1 to 1\u2009atm. The obtained impedance spectra were analyzed with both the DRT method and ECM fitting. \nFig. 6a and b represent the DRT transformation and the ECM fitting of the impedance spectra as a function of pO2 respectively. The P4 peak is independent of pO2, while P3, P2 and P1 show very slight changes with pO2 variation. From the DRT plot alone, it is arguable to attribute some of the peaks to the oxygen electrode process alone. However, this could indicate that the contribution of the oxygen electrode is minimal. The fitting results (Fig. 6b) of the impedance spectra with the equivalent circuit model showed that R1 (representing P1 peak) and R2 (P2 peak) resistances are mostly, contributions from oxygen electrode processes. However, following the inconsistency between the DRT representation and the NLLS fitting of the impedance, a further experiment was necessary to clarify the impedance contribution from the oxygen electrode. For this, impedance measurement was performed on a symmetrical half-cell containing LSCF electrodes (in two-electrode measurements). The obtained DRT was compared in Fig. 6a. The comparison shows that the P1 and P2 peaks are essentially oxygen electrode processes while P3 and P4 are mainly fuel electrode resistance contributions with slight contributions from the oxygen electrode. This is in agreement with the observed trend in the NLLS fitting. However, the focus of the current study is on the fuel electrode processes, hence the electrochemical processes of the oxygen electrode were inferred from the numerous literature on LSCF electrodes [40\u201343]. Overall, the obtained DRT representation of the LSCF impedance spectra (with their corresponding frequency range) for symmetrical half-cell is in good agreement with literature observation. [40,42]. Chen et [42] al. investigated the performance of LSCF symmetrical cells with different fabrication methods. From their analysis, they ascribed the low frequency peak between 1 and 10\u2009Hz to gas diffusion process, the mid frequency peak between 10 and 500\u2009Hz to surface exchange and ion diffusion process (which corresponds to the P3 and P2 in this report) and lastly, a high frequency peak (around 1000\u2009Hz) to charge transfer process across the interface. A similar result was also observed by Leonide et al. [40] in the impedance study of LSCF and LSF electrodes; a low frequency (0.3\u201310\u2009Hz) gas diffusion process, a mid frequency (2\u2013500\u2009Hz) oxygen surface exchange process followed by oxide diffusion in the bulk of the electrode and lastly a high frequency charge transfer process were observed. Therefore, following the high activation energy of R1 (as shown in Fig. 4c) as well as the high frequency range of the P1 peak (similar to ref [40,42]), there is no doubt that this is in good agreement and coincides with the charge transfer process of the oxygen electrode. \nTable 1 summarizes the possible electrochemical process contribution related to the individual polarization resistances.Long term stability test were performed to investigate the performance stability of the cells during long operating times. The measurement was performed at 900\u2009\u00b0C at a current density of \u2212\u20090.5\u2009A\u2027cm\u22122 for up to 1070\u2009h on two different cells with a fuel gas composition consisting of 80% CO2 and 20% CO. \nFig. 7a illustrates the degradation rate of the cell, represented as an increase in cell voltage as a function of time. The cell degradation was obtained by evaluating the slope of the curve. A degradation rate of 31\u2009mV\u2027kh\u22121 could be determined. In another cell, the evolution of the degradation mechanism was investigated during the long-term stability test by performing impedance measurements at OCV every 100\u2009h. Fig. 7b and c illustrate the evolution of the impedance spectra as well as the ohmic (Rs) and polarization (Rp) resistances respectively as a function of operation time. The Rs increased from 0.48 to 0.53\u2009\u03a9\u2027cm2 while the Rp increased from 0.28 to 0.35\u2009\u03a9\u2027cm2 after the degradation test. It is known that the loss of electrode contact surface with the current collector could also lead to an increase in the ohmic resistance and thus an increase in degradation rate. Such ohmic resistance increase due to loss of contact surface is usually accompanied by a proportional increase in the Rp. However, in our case, the increase in Rp is higher and thus disproportionate to the ohmic resistance increase, indicating that ohmic resistance is most likely, not due to contact loss. Fig. 7d illustrates the equivalent circuit analysis of the spectra which shows that the degradation behavior is dominated by the high/mid frequency processes of R1, R2 and R3 while Ws is relatively unaffected. This implies that both the oxygen electrode and the fuel electrode contributed to the degradation mechanism.Post-test analysis was performed on the measured cell to investigate the morphology and the microstructure of the electrodes. The microstructure of the long-term measured cell was compared to that of a freshly reduced cell. \nFig. 8a-c represents the fuel electrode of the reduced cell while Fig. 8d illustrates the microstructure of the corresponding oxygen electrode. Similarly, Fig. 8e-g shows the microstructure of the fuel electrode after long-term degradation test, while Fig. 8h represents the corresponding oxygen electrode. The secondary electron image of the electrode microstructure between Fig. 8a and e illustrates an increase in the Ni particle size after the degradation test. This indicates Ni particle agglomeration during CO2 electrolysis. The Ni agglomeration is further confirmed in the backscattered electron image of Fig. 8f-g when compared to Fig. 8b-c, where gray particles are Ni and white bright particles are GDC. In addition, Ni depletion and pore formation at the electrode/electrolyte interface are also visible after the long-term degradation test as compared to the reference cell (Fig. 8a,b and e,f). The observed microstructural changes would influence the measured impedance result in different ways. For example, the formation of a Ni-depleted layer as a result of Ni migration away from the electrolyte increases the electrolyte thickness thereby leading to an increase in ohmic resistance, as seen in Fig. 7b. Also, Ni agglomeration would effectively reduce the active surface area for electrochemical reactions, causing an increase in polarization resistance (Fig. 7c). Lastly, the continuous progression of these effects during cell operation would inevitably result in a decrease in cell performance over time due to an increase in area specific resistance. An extensive microstructural analysis of the electrodes, in relation to the observed electrochemical degradation processes, will be the objective of another paper.The impedance analysis also reveals that the high frequency processes, which are majorly oxygen electrode processes also contributed to the observed degradation. The secondary electron image comparison between Fig. 8d and h reveals no significant change in the LSCF microstructure. However, the mechanism of LSCF oxygen electrode degradation has been extensively studied [44\u201346]. One such mechanism is the formation of an insulating SrZrO3 layer due to the reaction between volatile SrO and the YSZ electrolyte [44,47]. To prevent this reaction, a GDC barrier layer between the electrolyte and the oxygen electrode is adopted. However, since the GDC barrier layer in Fig. 8d and h is not fully dense, this reaction cannot be entirely prevented. Monaco et al. [47] opined that the formation of an insulating SrZrO3 layer causes loss of Zr4+ in the 8YSZ electrolyte thereby reducing the ionic conductivity of the electrolyte and thus, leading to an increase in ohmic resistance of the cell. The increase in ohmic resistance entails a decrease in overall cell performance and thus, an increase in cell degradation [44\u201346].In this study, an electrolyte-supported single cell consisting of Ni-GDC fuel electrode and LSCF oxygen electrode was fabricated and analyzed under high temperature CO2 electrolysis conditions. Impedance measurements were carried out at different temperatures as well as under different CO2:CO fuel gas compositions and oxygen partial pressures. The obtained impedance spectra were evaluated with both DRT and NLLS fitting. Four time constants, representing four peaks in the DRT spectra, were used to fit the impedance spectra. The high frequency P1 peak corresponds to the oxygen electrode charge transfer process while the middle frequency processes of P2 and P3 consist of contribution from surface reaction processes from both the fuel and oxygen electrodes. Lastly, the low frequency process is assigned to the gas diffusion process in addition to the surface reaction in the fuel electrode. Long-term degradation analysis was performed at 900\u2009\u00b0C and a current density of \u2212\u20090.5\u2009A\u2027cm\u22122. The cell shows a low degradation rate of 31\u2009mV\u2027kh\u22121 during 1070\u2009h of operation. Furthermore, analysis of the degradation mechanism showed that the high/mid frequency processes contributed more to the degradation rate. Microstructural evaluation of the measured cell with SEM revealed Ni particle agglomeration, increase in electrode porosity and Ni migration away from the electrolyte.\nIfeanyichukwu D. Unachukwu: Methodology, Investigation, Formal analysis, Validation, Conceptualization, Data curation, Software, Visualization, Writing \u2013 original draft, Writing \u2013 review & editing. Vaibhav Vibhu: Methodology, Formal analysis, Validation, Conceptualization, Software, Supervision, Visualization, Writing \u2013 review & editing. Jan Uecker: Investigation, Formal analysis, Validation. Izaak C. Vinke: Methodology, Supervision, Validation, Project administration, Conceptualization, Resources, Software, Visualization. R\u00fcdiger-A. Eichel: Supervision, Project administration, Resources. L.G.J. (Bert) de Haart: Methodology, Supervision, Validation, Project administration, Conceptualization, Resources, Software, Visualization.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors gratefully acknowledge funding by the German Federal Ministry of Education and Research (BMBF) within the iNEW 2.0 Project: incubator sustainable and renewable value chains, under grant agreement number 03SF0627A.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jcou.2023.102423.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  The challenges of high degradation rate and significant carbon deposition, which are common with Ni-YSZ electrodes, have shifted attention to other electrode materials with enhanced performance in SOECs using carbon-containing fuels. In this study, the performance and electrochemical behavior of the Ni-GDC fuel electrode under CO2 electrolysis were investigated. The study was performed over a range of operating conditions, varying the operating temperature, the CO2 content of the fuel gas as well as the oxygen partial pressures in the oxygen electrode gas. Long-term stability test was performed up to 1070\u00a0h at 900\u00a0\u00b0C and a current density of \u2212\u00a00.5\u00a0A\u2027cm\u22122. The electrochemical impedance spectra obtained from the various measurement were evaluated with DRT as well as an equivalent circuit model consisting of 4 time-constant; (LR-RQ1-RQ2-RQ3-Ws). The low frequency Warburg (short) element (Ws) was attributed to gas diffusion and surface processes at the fuel electrode, the mid frequency processes of RQ2 and RQ3 are assigned to the combined contribution of fuel and oxygen electrode. The high frequency RQ1 was assigned to the charge transfer process at the oxygen electrode. A low degradation rate of 31\u00a0mV\u2027Kh\u22121 was observed during the long-term stability test. Furthermore, analysis of the degradation rate illustrates that significant contributions to the degradation were from the mid and high frequency processes, in addition to ohmic resistance. SEM analysis of the measured cell shows agglomeration of Ni particles, increase in electrode porosity as well as Ni migration away from the electrode/electrolyte interface.\n               "}
{"full_text": "Data will be made available on request.Metal\u2013organic frameworks (MOFs) prepared from nitrogen-rich ligands are extremely versatile materials; over the last years, a great number of MOF materials with carboxyl and/or pyridine ligands have been designed and built [1]. Now, there is an increasing interest in using azolates as linkers to obtain MOFs, because result in strong metal-nitrogen bonds which endow high chemical and thermal stabilities to the frameworks [2\u20134]. N-based ligands include pyrazole, imidazole, triazole, and tetrazole that contain nitrogen atoms with Lewis basic activity which can act as coordinating and interactive sites to build MOFs. However, the tetrazolate ligands might be the most promising candidates to obtain versatile MOFs, due to their possibility of donor N sites to coordinate by different modes to the metal [1,5]. Tetrazole has similar pKa than carboxylate, which shows that the conditions to prepare the corresponding MOFs are similar. Tetrazole-based MOFs (Tz-MOFs) are robust and stable materials with good porosity, adequate topology, and adsorption characteristics comparable to those from carboxylate-based MOFs. Compared to bidentate carboxylate ligands, tetrazolate derivatives have strong coordination capabilities and different coordination modes. After deprotonation (partial or total), tetrazoles turn into azolate anions, which change the basicity characteristics of these linkers and therefore the coordination to metals [6]. In view of the good properties as stability and porosity, tetrazole compounds have applications in gas adsorption [7\u201310], magnetism [11,12], sensing [13,14], potential applications in optics [15\u201317] and are used to obtain energetic materials [18\u201321]. Additionally, in medicinal chemistry tetrazole compounds can be substituted for the carboxylate functions [2]. In general, MOFs with carboxylate ligands have been widely explored as catalysts [22\u201326], while MOFs prepared from azolate linkers and low-valent metals (Co2+, Ni2+, Zn2+, etc) have been less studied [27\u201330]. Thus, it is worth to synthesize tetrazole derivatives and exploring their potential applications.One of the most important current problems is the large emission of CO2. To mitigate this problem there are different strategies, being the valorization of CO2 to value-added products one of the most promising [31\u201333]. MOFs have been widely studied as heterogeneous catalysts for the addition of CO2 to epoxides to yield cyclic carbonates because of the Lewis or basic sites on their structures, thus, many MOFs are attractive due to the possibility to generate unsaturated metal sites by removing the coordinated solvent molecules. In particular, Co-MOFs (quite earth abundant and low-cost metal) have been used as catalysts for the oxidation of olefins due to their redox capability and as Lewis acid catalysts [34].One-pot cascade processes are greener, simpler and more efficient than general step by step reactions since it is not necessary to isolate or purify the intermediate products. However, there are a limited number of proper catalysts designed to be effective with a wide scope of substrates. Thus, the direct synthesis of cyclic carbonates from alkenes and CO2 is attracting growing importance [35\u201337].Based on the literature data and fascinating qualities of these ligands, we are interested in building new MOFs of cobalt from two interesting tetrazole-based ligands 2,6-di(1H-tetrazol-5-yl)naphthalene (H2NDTz) and 2,6-bis(4-(1H-tetrazol-5-yl)phenyl)naphthalene (H2NDPhTz) (\nFig. 1). Thus, besides the synthesis and characterization of the new MOFs, and due to coordinative unsaturated cobalt sites generated after thermal activation, we have demonstrated their catalytic efficiency in the epoxidation of alkenes, cycloaddition of CO2 to epoxides and in the synthesis of cyclic carbonates from styrene, through a one-pot tandem epoxidation-cycloaddition reaction. This tandem reaction has been rarely studied with Co-MOFs catalysts, thus one example has been reported so far [38].Details for the synthetic procedure of H2NDTz and H2NDPhTz ligands can be found in the supplementary information.A solution of CoCl26\u2009H2O (22.5\u2009mg, 0.12\u2009mmol) in 0.5\u2009mL of water was added to a 2.0\u2009mL dimethylformamide (DMF) solution of H2NDTz (29.5\u2009mg, 0.12\u2009mmol) or H2NDPhTz (49.9\u2009mg, 0.12\u2009mmol). The 2.5\u2009mL solution was poured into a 4\u2009mL vial and heated at 90\u2009\u00b0C for 1 d. Block-shaped pale-brown crystals were obtained and were washed using DMF, water and methanol and then dried in air to afford a total of 38.0\u2009mg of a crystalline product in the case of Co-NDTz [Co3(NDTz)3(DMF)3(H2O)6] and 40\u2009mg in the case of Co-NDPhTz [Co3(NDPhTz)3(DMF)3(H2O)6. Crystal data of both MOFs is provided in CIF format, accessible through CCDC numbers: 2180927, for Co-NDTz, and 2180928 for Co-NDPhTz.Co-NDTzs and Co-NDPhTz were thermally activated before catalytic experiments by heating them at 110\u2009\u00b0C under vacuum for 10\u2009h.In a typical reaction, styrene (4.8\u2009mmol), t-butylhydroperoxide (TBHP, 5.5\u2009M in decane, 7.2\u2009mmol), activated Co-NDTz catalyst (0.2\u2009mol% based on Co) were added into a Supelco glass microreactor (5\u2009mL). The reaction mixture was heated at 50\u2009\u00b0C and stirred. Aliquots were taken at different times and reaction evolution was followed by GC-MS. After reaction was complete, the catalyst was separated by centrifugation, thoroughly washed, dried under vacuum and after a new activation, it was reused.Epoxide (0.020\u2009mol), biphenyl (internal standard (0.002\u2009mol); nBu4NBr (0.06\u2009mmol), and activated Co-NDTz catalyst (0.0114\u2009mmol based on Co) were added into a picoclave B\u00fcchi reactor (10\u2009mL). The reaction mixture was purged with CO2. Then, it was charged with 3\u2009bar of CO2 and heated at 50\u2009\u00b0C for 1\u201320\u2009h. The catalyst was recovered by centrifugation and activated before reuse in a new cycle. Reaction products were monitored by GC-MS.Styrene (4.8\u2009mmol), TBHP (5.5\u2009M in decane, 7.2\u2009mmol), nBu4NBr (0.018\u2009mmol), and thermally activated Co-NDTz catalyst (0.2\u2009mol% related to styrene) were introduced into a picoclave B\u00fcchi reactor (10\u2009mL). The reaction mixture was purged several times with CO2, charged with 3\u2009bar of CO2 and heated at 110\u2009\u00b0C overnight. The catalyst was recovered by centrifugation, washed with acetone and activated again at 110\u2009\u00b0C before the reuse in a new cycle. Reaction products were monitored by GC-MS.H2NDTz and H2NDPhTz (\nScheme 1) were prepared in two steps using 2-triflate,6-bromonaphthalene as common precursor. Thus, H2NDTz was synthesized according to a modified method to that reported [39,40]. It was obtained by reaction of brominated precursor with copper cyanide [41] and subsequent conversion of the cyano groups into the corresponding tetrazoles. H2NDPhTz was prepared by a Suzuki-Miyaura reaction between 2-triflate,6-bromonaphthalene and 4-cyanophenylboronic acid followed by reaction with sodium azide under the same conditions that the above ligand.Both ligands were obtained in good yields (>80 %) and the structures were confirmed by 1H NMR (Figs. S1 and S2) and elemental analyses (see supporting information).The synthesis of the cobalt-naphthalene tetrazole MOFs was carried out under solvothermal conditions. Co-NDTz was obtained by reaction of CoCl2.6\u2009H2O with H2NDTz in a mixture DMF/H2O: 1/1 at 70\u201390\u2009\u00b0C for 1\u2009day. However, when ligand H2NDPhTz was employed, the corresponding Co-NDPhTz is only obtained when a mixture DMF/H2O: 1/10 ratio was used at 90\u2009\u00b0C. Both materials have been fully characterized by elemental analysis, Fourier Transform Infrared Spectra (FT-IR), Thermogravimetric analysis (TGA), and powder X-ray diffraction (Fig. 1). FT-IR spectra show that the tetrazolate ligand is incorporated into the network with CN bands at \u223c1600, 800\u2009\u2212\u20091300\u2009cm\u22121; CO frequency due to DMF molecules appears at \u223c1650\u2009cm-1 (Fig. 1b).TGA under air atmosphere shows the decomposition patterns for both MOFs. Co-NDTz shows a degradation pattern in two steps, which occur at decomposition temperatures of around 300 and 380\u2009\u00b0C respectively. However, Co-NDPhTz shows a unique degradation step with an initial decomposition temperature of around 340\u2009\u00b0C. Besides, this thermogram shows a weight loss of around 100\u2009\u00b0C attributed to solvent molecules entrapped within the network. From the residue obtained by TGA, assuming the formation of Co2O3, it was determined a Co content of 13.6\u00a0% for Co-NDTz and 10.8\u00a0% for Co-NDPhTz. These values and the N content obtained by elemental analysis showed us that both materials have a ratio ligand/Co: 1/1.The XPS survey spectra of Co-tetrazole-MOFs are shown in Figs. S3 and S4, indicating the presence of Co, N, C, and O elements in the materials. The Co 2p traces of both materials exhibit bands at \u223c783.1, 783.5 and 798.3, 798.6\u2009eV with a difference of \u223c15.1\u2009eV corresponding to Co2+ species 2p3/2 and Co 2p1/2 respectively; satellite peaks are also observed at 788.5 and 804.9\u2009eV. XPS of C1s exhibited a predominant peak at 285.9, 286.4\u2009eV (C-C, CN units) and N1s spectra showed a band at 401.5, 402.1\u2009eV (CN) (Figs. S4 and S5) [42].The morphology of the new cobalt MOFs was examined by SEM (Fig. 1e). Co-NDTz showed elongated oval rough aggregates while Co-NDPhTz showed a plate-like morphology.It has not been possible to obtain single crystals of adequate size for the resolution of the structure by single-crystal X-ray diffraction, therefore, the structural elucidation was completed based on analysis of powder X-ray diffraction patterns (Fig. 1a) and computer modellization. Thus, the PXRD pattern of Co-NDTz was first indexed with a monoclinic primitive cell, with lattice parameters a =\u200913.00\u2009\u00c5, b =\u200914.95\u2009\u00c5, c =\u20096.48\u2009\u00c5, \u03b2\u2009=\u200992.16\u00b0 (Table S1). A Pawley refinement was successfully completed, further supporting the feasibility of this cell. The short value of the c parameter strongly suggests the formation of a MOF with a rod-shaped secondary building unit (SBU), similar to previously reported structures including related pyrazole [43,44] or tetrazole [45,46] based linkers. Electron density maps were generated by applying the charge flipping method to the integrated intensities from the PXRD pattern, and used to build a crystal model in the P21\n/c space group, consisting of inorganic SBUs running along the crystallographic c axis, formed by cobalt atoms coordinated to bridging tetrazole rings (\nFig. 2a). The model was geometrically optimized with force-field based energy minimization procedures coupled with Rietveld refinement cycles, using Biovia Materials Studio Software package [47]. In the refined structure, there is one crystallographically independent cobalt atom coordinated to four nitrogen atoms from the tetrazolate linkers, and to two oxygen atoms from additional water ligands that complete the octahedral environment of the metal centers. The tetrazolate rings are coordinating to the metal atoms through the nitrogen atoms at 2, and 3 positions, creating short bridges that extend the rod-shaped SBU, which are then connected through the organic linkers to produce a 3D framework (Fig. 2b). Based on this structure, an isoreticular crystal model was built up for the extended Co-NDPhTz MOF (Fig. 2c). The corresponding lattice parameters obtained after a Pawley refinement are a =\u200922.77\u2009\u00c5, b =\u200915.89\u2009\u00c5, c =\u20097.26\u2009\u00c5, \u03b2\u2009=\u200990.06\u00b0 and the geometry optimization of the structure was therefore completed. We noticed that the relative intensities of some peaks in the experimental pattern are lower than the calculated one, which is possibly due to preferred orientation effects, expected for the plate like morphology of the crystals (Fig. 1e), and the diffraction data acquisition in reflexion geometry (Figs. S5 and S6).Nitrogen gas adsorption isotherms were measured after the evacuation at 373\u2009K overnight to remove solvent molecules (under these conditions, PXRD pattern is maintained). The surface area calculated using Brunauer-Emmett-Teller (BET) method from the nitrogen gas adsorption data result in 43.0\u2009m2g\u22121 for Co-NDTz (Fig. S7) and 4.47\u2009m2g\u22121 for Co-NDPhTz. However, the CO2 sorption measured at 273\u2009K (Fig. 2d) reveals for case of Co-NDTz a Dubinin-Astakhov [48] CO2 specific surface area of 636\u2009m2g\u22121 and a CO2 uptake of 2.35\u2009mmol\u00b7g-1; whereas for phenyl extended MOF, Co-NDPhTz, the Dubinin-Astakhov CO2 specific surface area was 308\u2009m2.g-1 and the CO2 uptake of 1.31\u2009mmol\u00b7g-1; which indicates a lower CO2 accessibility to the tetrazole group in Co-NDPhTz. This result suggests a blockage or collapse of the structure which could be attributed to the length and freedom of rotation of the bonds that make up this linker.The different N2/CO2 adsorption was recently observed for a different type of cobalt MOF, which exhibited a BET surface area of only 6.8\u2009m\u00b2.g\u22121 but a CO2 uptake of 2.26\u2009mmol.g-1, attributed in this case to the interaction of the amide functional group of the framework with the polar CO2\n[49]. Most of the cobalt MOFs reported exhibit CO2 uptake between 1.0 and 3.0\u2009mmol.g-1\n[50,51]. In our case, it is the presence of tetrazole groups, along with the rigidity of the linker NDTz, that made Co-NDTz exhibit a high CO2 uptake capacity. It is known that strong dipole-dipole and acid-base interactions are present between protonated and deprotonated forms of tetrazole ring and CO2 carbon dioxide [52].MOFs have been evaluated as heterogeneous catalysts for different types of reactions being effective and selective due to their properties as porosity (high surface area or the important number of active sites [53]. It is also known that for a MOF to be used as a heterogeneous catalyst, it is necessary that there exist coordinative unsaturated sites (CUSs) [54], and considering that the Co-MOFs, herein reported, have free and coordinated solvent molecules, the Co-MOFs should be thermal activated at 110\u2009\u00b0C under reduced pressure for 4\u2009h before each reaction. The PXRD pattern shows that activated Tz-MOFs maintain the crystallinity (Fig. S8). Encouraged by the availability of both Lewis acidic and potentially redox-active Co sites in thermally activated Co-NDTzs, we have investigated its catalytic performance in the tandem synthesis of cyclic carbonates from olefins and carbon dioxide. Before carrying out the tandem reaction, a preliminary evaluation of the novel catalysts in the separated reactions was carried out. Both, the epoxidation of olefins and the coupling of CO2 with epoxides to obtain cyclic carbonates were evaluated individually.The possibilities of MOFs as oxidation catalysts have been reported by different authors [55], now we have evaluated Co-NDTzs in olefin epoxidation and the results are presented in \nTable 1. The epoxidation reaction was initially examined under solvent-free conditions using styrene as substrate and either hydrogen peroxide or oxygen (4\u2009bar) as oxidants, along with activated Co-NDTz (0.2\u00a0% mol), at 90\u2009\u00b0C. Under these conditions, no conversion was achieved. When a solution of TBHP in decane (5.5\u2009M), was employed as an oxidant, the styrene oxide was selectively obtained after 14\u2009h with excellent yield (>99\u00a0%) (entry 1). Control experiments revealed the synergic effect between the oxidant and the MOF, since only 4\u00a0% of styrene oxide was obtained in absence of catalyst (entry 2). Co-NDPhTz seems to be less active than Co-NDTz (entry 3) since only 14\u00a0% conversion was observed. To explore the versatility of the catalyst in the epoxidation reactions, an internal alkene as cyclooctene was also oxidized, however only 34\u00a0% (64\u00a0% at 110\u2009\u00b0C) of the epoxide is observed after 24\u2009h of reaction (entries 4\u20135). Some examples of the CG-chromatograms obtained in this reaction are collected in the ESI.The removal of the catalyst by filtration after 4\u2009h stopped the reaction, and the filtrate afforded nearly no additional conversion after stirring for another 8\u2009h (Fig. S8b). These observations suggest that the catalyst is a true heterogeneous catalyst. Solids could be isolated from the reaction suspension by simple filtration. Recovered Co-NDTz was reused five times and, no significant loss of catalytic performance was observed (Fig. S9a). The structural integrity was verified by PXRD (Fig. S8).The cycloaddition of CO2 to epoxides to obtain cyclic carbonates is the second reaction explored with these catalysts.The reaction conditions were first optimized using a Picoclave B\u00fcchi glass reactor of 10\u2009mL, Co-NDTz as catalysts and epichlorohydrin (ECH) as substrate (\nFig. 3), due to its reactivity and interest, because it can be produced from glycerine obtained from vegetable oil [56]. To optimize the reaction conditions, those reported so far for different Co-MOFs used as catalysts to promote this reaction, were tested [49,57\u201359]. Thus temperatures between 40 and 100\u2009\u00b0C, CO2 pressures between 1 and 8\u2009bar and different amounts of TBAB (from 2.5 to 10\u2009mol\u00a0%) were initially tested. From the above experiments, the most favorable conditions for this reaction with Co-NDTz as a catalyst were a temperature of 50\u2009\u00b0C and 3\u2009bar of CO2, but using a much lower amount of TBAB (0.3\u2009mol\u00a0%) (\nTable 2). Moreover, with an ECH:Co ratio of 1750:1 a good conversion of ECH was achieved in 90\u2009min (entry 1) and the corresponding cyclic carbonate was selectively obtained. It should be noted that the ECH:Co ratio used was also much higher than any of those reported so far for Co-carboxylate-MOFs, that do not exceed the ratio ECH:Co of 1000:1 [57\u201359].Using the same reactor, some other epoxides were used to evaluate the catalyst (Fig. 3). When a bulkier and less reactive epoxide such as styrene oxide (SO) was used (entry 2) a good conversion was obtained (65\u00a0%) taking into account the time employed (1.5\u2009h). Usually, a good conversion of this epoxide requires longer reaction times (9\u201312\u2009h) [57\u201359]. The same conversion was achieved with 1,2-epoxyhexane (HE) although it took 7\u2009h of reaction (entry 3).Finally, an internal epoxide such as cyclohexene oxide (CHO) (entry 3) was evaluated. After 20\u2009h of reaction, only 26\u00a0% of epoxide conversion was achieved. In order to check if the structure of the catalyst had an influence on this result, the reaction was carried out using Co-NDPhTz as catalysts (entry 5). In the same conditions, a slight increase in the conversion was obtained which indicated a great difficulty in opening this epoxide probably caused by the steric hindrance between the two rings as was observed in previous works [60,61].Two control experiments were carried out, using only TBAB to promote the reaction (entry 6) or using Co-NDTz in absence of TBAB (entry 7) in the same conditions as the above experiments and using ECH as substrate. In both cases, a lower conversion of ECH was observed after 90\u2009min of reaction, which confirms the cooperative effect between Co-NDTz and TBAB.A scale-up of the reaction was carried out using ECH as substrate with a reactor of higher capacity that allowed the use of 6 times more epoxide than in the previous studies. For this experiment, the temperature was also of 50\u2009\u00b0C, but a continuous pressure of CO2 at 3\u2009bar was maintained and a lower amount of catalysts was employed, ECH:Co ratio of 3500:1. In these conditions (entry 8) a 63\u00a0% of ECH conversion was attained which involved a TON of 2205. The catalyst was removed by centrifugation, washed with methanol three times, and then, it was dried overnight at 110\u2009\u00b0C under vacuum. The catalyst was used in a new reaction to investigate its recyclability. Thus, in a second run using fresh ECH and the same reaction conditions, a slightly lower ECH conversion was obtained although the corresponding cyclic carbonate was obtained as a unique product indicating a >\u200999\u00a0% of selectivity after the first recycling. Similar ECH conversions were obtained in the next runs (up to six runs) which confirms the good recyclability of this catalyst even using a big amount of substrate (Fig. S10a). CG-chromatograms of the fifth run are collected in the ESI.The catalyst was also removed by filtration after 30\u2009min of reaction and no additional conversion after stirring for another 90\u2009min was observed (Fig. S10b) confirming again the true heterogeneous character of this catalyst.The reaction was performed with the extended catalyst (Co-NDPhTz) (entry 9) obtaining also a good ECH conversion and excellent TON value.MOFs have been also evaluated as heterogeneous catalysts for cascade reactions by combining acid-base character, redox properties, and metal\u2013organic nodes [62]. Because activated Co-NDTz demonstrated its ability as an effective catalyst in the epoxidation of olefins as well as in the cycloaddition of CO2 to epoxides, we were encouraged to study the one-pot oxidative carboxylation of styrene with carbon dioxide to obtain cyclic carbonates. Thus, activated catalyst, styrene, TBAB (as co-catalyst necessary to open the ring epoxide), TBHP (as an oxidant), and CO2 were introduced in a Picoclave B\u00fcchi glass reactor, when the reaction was performed at 50\u2009\u00b0C, only the corresponding epoxide was obtained (\nTable 3, entry 1); if the reaction mixture was heated to 110\u2009\u00b0C, the epoxidation and the CO2 coupling could be completed with a single workup stage (Table 3, entry 2) The kinetic profile of this control reaction is shown in \nFig. 4. The reaction performed with Co-NDPhTz yields selectively the epoxide and only traces of carbonate were detected.It has been reported that in the plausible mechanism for this tandem reaction, tert-butanol (TBOH) derived from TBHP might coordinate to the metal at the apical positions and inhibit the carboxylation step [63]. This is in agreement with our experimental observation, since it is required to thermally activate the MOF before using it to afford CUS, otherwise no conversion is observed in any reaction. Thus, when the reaction temperature is increased to 110\u2009\u00b0C, the tert-butanol is thermally and in situ decoordinated, and the second step is enabled. To confirm this fact, the CO2 cycloaddition of epoxides has been performed in presence of tert-butanol under the standard conditions described in the above section (Table 2). It was observed that even after 5\u2009h at 50\u2009\u00b0C no conversion was observed, however, the same batch was subsequently heated to 110\u2009\u00b0C and the carbonate was obtained in 97\u00a0% yield after 10\u2009h Some examples of this CG-chromatograms are collected in the ESI.Based on the above mentioned, a possible mechanism for the tandem epoxidation-carboxylation reaction from styrene and CO2 is shown in \nScheme 2. More into details: firstly, the reaction of unsaturated Co(II) sites with TBHP to form a Co(III)-peroxy complex that restores the Co(II) sites by releasing a t-butoxy radical. This radical reacts with styrene to generate a tert-butoxperoxy derivate that yields the styrene oxide and t-butanol which is coordinated to the cobalt. After activation at 110\u2009\u00b0C, it is uncoordinated and the epoxide is activated by coordination to the cobalt active center. Then, the less hindered carbon of the epoxide is attacked by bromide from the TBAB, followed by the insertion of carbon dioxide into the Co-oxygen bond giving a cobalt-carbonate intermediate. The cyclic carbonate is obtained by a closure intramolecular of the cycle accompanied simultaneously by the regeneration of both Co-MOF and co-catalyst.The catalyst could be reused at least five times with moderate loss of activity (from 78\u00a0% to 70\u00a0% yield), although longer reaction times are needed (72\u2009h) to achieve the same conversion (Table 3, Fig. S11). PXRD patterns, SEM images and FT-IR of the Co-MOFs recovered from the tandem reaction, revealed that the structures were mainly maintained in both cases (Figs. S8, S12-S13).The tandem epoxidation-carboxylation reaction has been studied with different types of catalysts, a comparison with recently reported metal-based catalysts along with some recent reviews can be found in Table S3. Higher CO2 pressures are usually needed for this transformation (from 5 to 100 in one case), and little attention has been paid to the recyclability of the system. For example, similar conditions were employed by Zhang and coworkers [64], with Ce-based MOF, and Lin and coworkers, with a Mn/Hf-MOF [65], in both the catalytic system was recycled 3 times. Besides, metal loadings are much higher compared to our system which only requires 0.2\u2009mol\u00a0%.We have obtained two new porous tetrazole-based MOFs: Co-NDTz and Co-NDPhTz using 2,6-di(1H-tetrazol-5-yl)naphthalene (H2NDTz) and extended 2,6-bis(4-(1H-tetrazol-5-yl)phenyl)naphthalene (H2NPhDTz) as linkers under solvothermal conditions. Based on the structural analysis, we can say that they both are isoreticular MOFs. After thermal activation, both compounds have a great number of coordinative unsaturated and redox Co(II) sites and proved to be excellent catalysts for the selective epoxidation of alkenes and in the CO2 cycloaddition to epoxides. Also, we have proved that activated Co-NDTzs result effective for the one-pot tandem epoxidation-carboxylation of styrene with CO2. This work opens a new via to develop nitrogen-rich tetrazole-MOFs as efficient heterogeneous bifunctional catalysts.Antonio Valverde: synthesis, characterization, formal analysis, investigation, and writing the original draft. M. Carmen Borrallo-Aniceto: characterization and catalytic experiments. Urbano D\u00edaz: characterization and review. Eva M. Maya: catalytic experiments, formal analysis, investigation, and writing \u2013 review & editing Felipe G\u00e1ndara: characterization and review. F\u00e9lix S\u00e1nchez: review & editing. Marta Iglesias: conceptualization, funding acquisition and writing \u2013 review. All the authors discussed the results and commented on the manuscript.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Authors acknowledge to Grants PID\n2020-112590GB-C22 and PID\n2020-112590GB-C21 funded by MCIN/AEI/10.13039/501100011033. A.V.G. thanks for FPU17/03463. We acknowledge Dr. F\u00e1tima Esteban from ICMM X-Ray Diffraction Facility for assistance in PXRD data acquisition.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jcou.2022.102298.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  The development of efficient catalysts that include the advantages of homogeneous and heterogeneous catalysts is a challenge that can be achieved with metal-organic frameworks (MOFs) since they can incorporate different functionalities in their structure that make them promising catalysts for different processes. Herein, two new isoreticular nitrogen-rich naphthalene cobalt-MOFs, Co-NDTz and Co-NDPhTz, were successfully prepared under solvothermal conditions from the corresponding linkers (H2NDTz=2,6-naphthaleneditetrazole and H2NDPhTz=2,6-bis(4-(1H-tetrazol-5-yl)phenyl)naphthalene). These Co-tetrazole-based MOFs combine Lewis acid and redox functionalities and good CO2 adsorption and after being thermally activated resulted to be excellent efficient catalysts for the epoxidation of alkenes and CO2 cycloaddition to epoxides yielding cyclic carbonates, reaching turnover numbers up to 2500. Furthermore, these two reactions take place following a highly desired one-pot tandem process and a cyclic carbonate was obtained from styrene and CO2 under solvent-free conditions. In addition, the heterogeneous catalysts are easily recycled without noticeable loss of catalytic activity and without important structural deterioration.\n               "}
{"full_text": "\n\n\nData will be made available on request.\n\n\nData will be made available on request.The production of wastewater containing acetic acid is common in chemical industries such as petrochemicals and wood pulp mills [1,2], and acetic acid has always been a by-product in some conventional wastewater treatment processes of macromolecular organic matter. Zhong et\u00a0al. [3] reported several short-chain acids, such as acetic acid, as major by-products that formed during the ozonation degradation of humic acids, and Reisz et\u00a0al. [4] found that acetic acid was formed when O3 was used for 2-propanol oxidation, because the \u03b1-position methyl group coordinated with the carboxyl group is not easy to be further oxidized [5]. Acetic acid-contaminated water not only causes great harm if ingested by livestock and to crop irrigation but also adversely affects the respiratory system and sensory organs of human beings [6,7]. Therefore, the effective degradation of acetic acid is of great significance for the complete mineralization of organic compounds in wastewater.Several physical approaches, such as adsorption and membrane separation methods, have been found to exhibit excellence performance of acetic acid removal [8,9]. However, acetic acid was not completely destructed and the adsorption materials required frequent regeneration. Besides, traditional chemical method using alkali to neutralize wastewater usually leads to an enormous load of total dissolved solid in the treated effluent for biodegradation. It was reported that almost twice amount of NaOH or CaCO3 were required to neutralize acetic acid in the wastewater for further biodegradation [10]. Biodegradation would be a very slow process, which was scarcely possible under the practical circumstance [11]. Advanced oxidation processes (AOPs) have been widely used to remove organic matter from aqueous environments [12]. The reactive species, such as \u00b7OH and \u00b7O, can effectively break down organic matter into harmless products [13]. Sannino et\u00a0al. [14] found the heterogeneous photo-Fenton oxidation of acetic acid on LaFeO3, could reach 60% after 5 h oxidation. Cihano\u011flu et\u00a0al. [15] combined ultrasound with a catalyst to oxidize acetic acid wastewater, but the chemical oxygen demand (COD) degradation was only 25.5%. In addition to the drawbacks of incomplete degradation, the large scale of the removal unit and secondary contamination would restrain its practical application [16].By contrast, catalytic ozonation has been proven to be an effective method to produce reactive oxygen species and further degrade organic matter. Transitional metal oxides showed great potential for ozone decomposition and form more reactive oxygen species and hydroxyl radicals with higher redox potentials, thus further promoting the degradation efficiency of the target organic pollutants [17,18]. It has been found that nano-CeO2 greatly improved the oxidation efficiency of the H2O2/O3 system and promoted the degradation of acetic acid small molecules [19]. Peng et\u00a0al. showed that the O3 oxidation efficiency of succinic acid reached 100% in combination with a Ni/Al2O3 catalyst [20]. Nevertheless, little attention has been paid so far to the degradation of acetic acid by catalytic ozonation, and the development of the corresponding catalysts is of great significance.In this work, a range of metal oxides (MnO2, Co3O4, Fe3O4, and CeO2) were loaded on \u03b3-Al2O3 and their catalytic ozonation performance for acetic acid degradation were tested. The influencing factors (acetic acid concentration, O3 concentration, pH, and ozonation temperature) were also investigated. The physiochemical properties of the catalysts were characterized with X-ray diffraction (XRD), scanning electron microscope (SEM), and X-ray photoelectron spectroscopy (XPS). Tert-butanol (TBA) and p-benzoquinone (PBQ) were used to explore reactive species for acetic acid oxidation. The mechanism of acetic acid ozonation on MnO2/\u03b3-Al2O3 catalyst was further proposed based on the in situ diffuse reflectance Fourier transform infrared spectroscopy (in situ DRIFTS).The catalysts were prepared via wet impregnation method. The preparation method was as follows: Firstly, a certain amount of nitrates of manganese, cobalt, iron, and cerium was dissolved in the deionized water. Then the \u03b3-Al2O3 powder (as support, 2000 mesh, 99%) was dosed into the nitrates solution and vigorously stirred for 1 h. The impregnated powder was dried at 100 \u00b0C for 5 h and finally calcined in air at 500 \u00b0C for 3 h to obtain the catalysts.The crystalline phases of the catalysts were determined using X-ray diffraction (XRD, Rigaku Ultima IV powder diffractometer, Japan) with a Cu Ka radiation. The morphological properties of the catalysts were measured using scanning electron microscopy (SEM, Supra 55, Zeiss, Germany) with energy-dispersive X-ray spectroscopy (EDS). The surface elements on MnO2/\u03b3-Al2O3 were analysed using X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi+, Thermo, USA).The experimental set-up used for catalytic ozonation of acetic acid was shown in Fig.\u00a01\n. A pulse power supply (M10K-08, Suzhou Allftek, China) was used to drive the dielectric barrier discharge (DBD) reactor DBD to generate O3. The volume fraction of O3 was around 30.0\u201340.0 g Nm\u22123. The concentrations of acetic acid and oxidation products (CO and CO2) in the gases from the outlet of the ozonation reactor were on-line analysed using a gas chromatograph (GC) (GC2014, Shimadzu, Japan) equipped with two flame ion detectors (FIDs) and two columns: a 2 m Porapark-N column (Dalian Institute of Chemical, China) with a methanizer prior to the FID to analyse CO and CO2; and a capillary column (SH-Stabilwax-DA, Shimadzu, Japan) to analyse the acetic acid concentration in the water liquid solution. By-products from acetic acid oxidation were analysed using high-performance liquid chromatography (HPLC, Agilent 1260, USA) equipped with an Eclipse Plus C18 column and an ultraviolet (UV) detector. The mobile buffers of A (KH2PO4, 0.015 mol L\u22121) and B (Acetonitrile with 0.5 mL min\u22121) were used. The pH value of the acetic acid solution was modified using sodium hydroxide and measured using a pH meter (AZ 86505, Hengxin, China). The O3 oxidation mechanism of acetic acid was investigated by adding OH quencher tert-butanol (TBA, 362 mM) and the superoxide radical (\u00b7\n\n\nO\n\n2\n\u2212\n\n) quencher p-benzoquinone (PBQ, 1.1 mM) to the acetic acid solution.The degradation (X, %) was calculated using Eq.\u00a01.\n\n(1)\n\n\nX\n=\n\n\n\nC\n0\n\n\u2212\n\nC\nt\n\n\n\nC\n0\n\n\n\u00d7\n100\n%\n\n\n\nWhere C0\n is the initial concentration of acetic acid in g\u00b7L\u20131, and Ct\n is the concentration of acetic acid in g\u00b7L\u20131 at reaction time t (min).Mineralization (Yi+1\n) and energy efficiency (\u03b7i\n) (g\u00b7kWh\u20131) were defined as follows:\n\n(2)\n\n\n\nY\n\ni\n+\n1\n\n\n=\n\n\nm\ng\n\n\nm\n0\n\n\n\u00d7\n100\n%\n\n\n\n\n\n\n(3)\n\n\n\n\u03b7\ni\n\n=\n\n\nm\ng\n\n\nP\n\n\n\nt\ni\n\n60\n\n\n\n\n\u00d7\n100\n%\n\n\n\n\nThe power discharge (P, kW) was calculated by:\n\n(4)\n\n\nP\n\n=\n\n\n\nP\na\n\n\u00d7\nf\n\n1000\n\n\n\n\n\n\n\n(5)\n\n\n\nP\na\n\n=\n\n\u2211\ni\n\n\n(\n\n\n\nV\n\ni\n+\n1\n\n\n+\n\nV\ni\n\n\n2\n\n)\n\n\n(\n\n\n\nI\n\ni\n+\n1\n\n\n+\n\nI\ni\n\n\n2\n\n)\n\n\n(\n\nt\n\ni\n+\n1\n\n\n\u2212\n\nt\ni\n\n)\n\n\n\n\nWhere P\na (J) is the energy injection in a pulse discharge period, and f (Hz) is the pulse frequency. Vi and Vi+1 was the discharge voltage at ti and ti+1 (s), respectively.\nEq.\u00a06 was used to calculate the amount of CO2 and CO generated from acetic acid oxidation.\n\n(6)\n\n\n\nm\ng\n\n=\n\n\u2211\n\nt\n=\n0\n\n\nt\n=\nt\n\n\n\n(\n\n\n\n60.5\nF\n\n\n22.4\n\n\n\n\n\n\n(\n\n\n[\nCO\n]\n\n+\n\n[\nC\n\nO\n2\n\n]\n\n\n)\n\n\nt\n\ni\n+\n1\n\n\n\n\u2212\n\n\n(\n\n\n[\nCO\n]\n\n+\n\n[\nC\n\nO\n2\n\n]\n\n\n)\n\n\nt\ni\n\n\n\n2\n\n\n(\n\n\nt\n\ni\n+\n1\n\n\n\u2212\n\nt\ni\n\n\n)\n\n\n)\n\n\n\n\nWhere m0\n is the initial amount of acetic acid in g. mg\n is the amount of acetic acid in g and calculated from the average concentration of [CO2] and [CO] at ti+1\n and ti\n time (min) with the difference in oxidation time (ti+1\n\u2212ti\n). P is the discharge power in W, 60 is the conversion factor between hour and minute, and 1000 is the conversion factor between W and kW. F is the flow rate of the bubbling gas, 0.100 L min\u20131; 22.4 is the molar volume (L) of a gas at standard state and 60.5 is the molecular weight of acetic acid. The discharge power (P) was calculated from waveforms of the pulse voltage supplied from the pulse power supply, using the voltage probe (P6015A, Tektronix, USA) and current probe (CP8030H, Cybertek, China) and oscilloscope (MDO 3022, Tektronix, USA).Firstly, the degradation, mineralization, and energy efficiency of acetic acid wastewater catalytic ozonation over MnO2/\u03b3-Al2O3, Co3O4/\u03b3-Al2O3, Fe2O3/\u03b3-Al2O3, and CeO2/\u03b3-Al2O3 catalysts were compared, and the results were shown in Fig.\u00a02\n. It is seen that MnO2/\u03b3-Al2O3 achieved the best degradation among all the catalysts evaluated, and the degradation rapidly increased to 49.2% within 20 min and gradually to 88.4% at 300 min (Fig.\u00a02a). By contrast, the degradation of acetic acid was only 39.9% with O3 bubbling and almost zero with O2 bubbling within 300 min. Besides, the degradation of acetic acid in the presence of catalysts were all much higher than that of O3 bubbling alone, which should be related to the promotion of O3 decomposition in the presence of the catalysts. The mineralization of acetic acid (oxidation to CO and CO2) is important to reflect the catalytic performance, which was also the key issues for acetic acid wastewater treatment. The mineralization of acetic acid is linear to ozonation time, and the slopes of the mineralization of acetic acid using MnO2/\u03b3-Al2O3 were the largest (Fig.\u00a02b). The energy efficiency using MnO2/\u03b3-Al2O3 reached the highest level at 50 min, and the value was approximately 14.9 g\u00b7kWh\u22121 and also the highest in comparison with the other three catalysts (Fig.\u00a02c). This demonstrated that MnO2/\u03b3-Al2O3 was an excellent candidate for acetic acid wastewater treatment.Besides, the effects of MnO2 loading amount on the degradation, mineralization, and energy efficiency of acetic acid were also investigated. The highest degradation of acetic acid was 88.4% at 300 min of ozonation time over 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst (Fig.\u00a02d). By contrast, the degradation of acetic acid only reached 56.3%. The mineralization over 1.0 wt.% MnO2/\u03b3-Al2O3 was also the highest in general, which increased to a maximum level of 88.2% (Fig.\u00a02e), with the best average energy efficiency 14.9 g\u00b7kWh\u22121 (Fig.\u00a02f). Fig. S1 indicates that the CO2 was the primary product from acetic acid ozonation since the CO2 concentration was approximately 6.5\u201343 times as many as CO.After the optimal loading amount of manganese was determined, other primary factors influencing the catalytic performance, such as catalyst dosage, acetic acid concentration, O3 concentration, and ozonation temperature were further studied. The influence of the dosage of the 1.0wt.% MnO2/\u03b3-Al2O3 catalyst on acetic acid degradation is shown in Fig.\u00a03\n. It was found that the both degradation and mineralization were increased as the dosage of 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst increased. However, there was no distinct difference between 30 and 40 g/L. Therefore, 30 g/L 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst was regarded as the optimum dosage with an overall consideration of the cost and effectiveness. The highest energy efficiency was also achieved when 30 g/L of 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst was dosed, reaching around 15 g\u00b7kWh\u22121. By contrast, the mineralization of acid acetic only reached 43.2% in the absence of the catalyst, with an energy efficiency of 5.1 g\u00b7kWh\u22121 (Fig. S2).Besides, the effects of initial acetic acid concentration on energy efficiency at various ozonation time were studied. As shown in Fig.\u00a04\n, the highest and lowest energy efficiencies were achieved when the initial acetic acid concentrations were 1 g\u00b7L\u22121 and 0.5 g\u00b7L\u22121, respectively. Since the ozonation of acetic acid related with the adsorption of acetic acid and O3 on catalyst surface, the presence of the highest initial concentration of acetic acid is possibly due to the optimum adsorption of acetic acid and O3 on catalyst surface under the experimental condition. Fig.\u00a04b shows the mineralization of acetic acid at 300 min ozonation time as a function of initial acetic acid concentration. The mineralization reached 89%, 84%, 67%, and 50%, when the initial acetic acid concentration was 0.5, 1, 1.5, and 2 g\u00b7L\u22121, respectively. Fig.\u00a04c shows the energy efficiency of acetic acid at 300 min ozonation time as a function of initial acetic acid concentration. The energy efficiency reached the highest around 15 g/kWh when the initial acetic acid concentration was 1 g\u00b7L\u22121. The mineralization decreased with increasing the initial acetic acid concentration, however, the amount of acetic acid mineralized increased (Fig.\u00a04d). This finding implied that in order to get a large amount of mineralized acetic acid, a high initial acetic acid concentration is required. However, in order to get the highest energy efficiency, the ozonation should be carried out with optimized initial acetic acid concentration (1 g\u00b7L\u22121).In order to investigate the amount of O3 consumed for the catalytic ozonation of acetic acid, and the effects of O3 concentration on energy efficiency was also studied (Fig.\u00a05\n). The O3 concentration was concisely adjusted by the pulse frequency of the DBD reactor. The O3 concentrations were 20.7, 35.6, 47.0, and 55.0 g Nm\u22123 when the frequencies were 50, 100, 150, and 200 Hz, respectively. In general, the higher O3 concentration, the higher acetic acid degradation, mineralization, and energy efficiency. The maximum energy efficiency (25.5 g\u00b7kWh\u22121) at 20 min was achieved when the frequency was 50 Hz. However, the energy efficiency decreased with increasing frequency (Fig.\u00a05c). It was found that the O3 concentration drop was within 4.0\u20136.0 g/Nm3 after 140 min of ozonation time (Fig.\u00a05d), although the O3 concentration increased with the frequency. This indicated that the amount of O3 used for acetic acid was limited, and most of the O3 did not take part in acetic acid ozonation and flowed away from the ozonation reactor.The reaction temperature also affects the catalytic ozonation results of acetic acid wastewater, and Fig.\u00a06\n demonstrates the degradation, mineralization, and energy efficiency of acetic acid catalytic ozonation in the range of 25-70 \u00b0C. It is seen that the catalytic ozonation reaction over MnO2/\u03b3-Al2O3 catalyst was not very sensitive to reaction temperature. The degradation of acetic acid at 25 \u00b0C was much lower within the first 50 min, and this might be the weaker volatilization of acetic acid compared with higher temperatures (Fig.\u00a06a). The mineralization of acetic acid was little affected by reaction temperature, indicating that the volatilized acetic acid would be finally oxidized into CO2 (Fig.\u00a06b). In addition, the energy efficiency was also very close at various reaction temperature within 300 min (Fig.\u00a06c). The highest degradation, mineralization, and energy efficiencies were achieved, and they were 85.2%, 95.3%, and 15.8 g\u00b7kWh\u22121, respectively, at 25 \u00b0C. In addition, the degradation (X) data at different ozonation temperatures were fitted with the first-order reaction. It is seen that ln(1\u2212X) was linear to ozonation time and the linear coefficient R2 were between 0.96\u22120.99 (Fig. S3). This suggests that the acetic acid catalytic ozonation is exactly first-order reaction over 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst.The initial pH value of the acetic acid solution was adjusted with 0.1 M NaOH. Fig. S4 shows the energy efficiencies at different pH values as a function of ozonation time. When the pH was 3.42 (the original pH of acetic acid solution), the energy efficiency was at the maximum level of 14.9 g\u00b7kWh\u22121, higher than that when the pH value was 7.06 or 11.26. It has been reported that \u00b7OH radical has a higher oxidation ability than other types of reactive oxygen species (such as \n\n\nO\n\n2\n\u2212\n\n), especially under acidic conditions, the similar pH effect was reported by Sahni et\u00a0al. [21] during the degradation of polychlorinated biphenyls using liquid-phase discharge plasma. Therefore, \u00b7OH radicals might also play a major role in acetic acid ozonation, and more details would be discussed in next section.The crystalline structures of \u03b3-Al2O3 and MnO2/\u03b3-Al2O3 with 0.5 wt.%, 1.0 wt.%, and 10 wt.% MnO2 loadings were characterized, and the XRD patterns are shown in Fig.\u00a07a\n. The peaks ascribed to \u03b3-Al2O3 were clearly observed, indicating that the crystalline structure of the support was well reserved after the addition of MnO2. However, the crystallinity of \u03b3-Al2O3 greatly decreased with the increase of MnO2 loadings. It was noted that as the loading amount of MnO2 increased to 10 wt.%, new diffraction peaks at 28.68\u00b0, 37.32\u00b0, 42.82\u00b0, 56.65\u00b0, and 72.38\u00b0 ascribed to MnO2 phase (JCPDS PDF#24-0735) were observed, and they were indexed to (110), (101), (111), (211), and (112) planes of MnO2, respectively. By contrast, when the MnO2 loadings were 0.5 wt.% and 1.0 wt.%, no characteristic peaks were found, indicating MnO2 was amorphous or highly dispersed on the surface of \u03b3-Al2O3. The SEM images of 1.0 wt.% MnO2/\u03b3-Al2O3 showed that MnO2 crystallites were in the form of nanoparticles, and the Mn and O elements were uniformly dispersed (Fig.\u00a07b).The surface chemical states of the 1.0 wt.% MnO2/\u03b3-Al2O3 catalysts before and after acetic acid ozonation were characterized with XPS. As shown in Fig.\u00a07c, e and f, the Mn 2p3/2 spectra were deconvoluted into three peaks at 640.6, 641.7, and 643.0 eV, which were ascribed to the Mn2+, Mn3+, and Mn4+ species, respectively [22]. It is noted that Mn species on the surface of 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst mainly existed in the form of Mn4+, and the relative ratio of Mn4+ species accounted for 43.8% and 48.1% before and after acetic acid ozonation. Besides, the relative ratio of Mn4+ of 0.5 wt.% MnO2/\u03b3-Al2O3 and 10 wt.% MnO2/\u03b3-Al2O3 were calculated to be 35.4% and 36.5%, respectively, much lower than that of the 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst. A higher Mn4+ ratio for the manganese-based catalysts is typically strongly linked to a superior catalytic activity [23,24], and this also explained the best catalytic ozonation performance of the 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst. In addition, it was noted that the ratio of Mn2+ decreased from 27.5% to 16.2% after the catalytic ozonation process (Table\u00a01\n), while the ratio of Mn3+ and Mn4+ increased. This was possibly due to a portion of Mn2+ being oxidized by O3.The O 1s spectra before and after ozonation were deconvoluted into the peaks at 531.7\u2013531.8 eV that were assigned to surface chemisorbed oxygen (Oads) and at 530.8\u2013530.9 eV that were ascribed to lattice oxygen (Olatt) (Fig.\u00a07d) [25,26]. Due to the high reaction activity, surface chemisorbed oxygen played an important role in a series of organic substances oxidation reactions. The relative ratio of Oads/Ototal was 49.1%, 71.2%, and 32.7% for 0.5 wt.% MnO2/\u03b3-Al2O3, 1.0 wt.% MnO2/\u03b3-Al2O3, and 10 wt.% MnO2/\u03b3-Al2O3 catalyst, respectively. Obviously, the relative ratio of Oads species at the surface of the 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst was greatly higher than that of 0.5 and 10 wt.% MnO2/\u03b3-Al2O3, and it accounted for a much higher ratio than the Olatt species, corresponding to its superior catalytic ozonation performance. Besides, the relative ratio of Oads and Olatt species were not distinctly changed (Table\u00a01), suggesting the excellent stability of 1.0 wt.% MnO2/Al2O3 catalyst in the process of acetic acid wastewater catalytic ozonation.In situ DRIFTS was carried out to investigate the mechanism of acetic acid catalytic ozonation over MnO2/\u03b3-Al2O3 catalyst, and the results were shown in Fig.\u00a08\n. An acetic acid/water bubbler and O3 generator were used to feed the in situ cell with a gas mixture of acetic acid, water, and O3 and to simulate the gas-liquid-solid reaction on the MnO2/\u03b3-Al2O3 catalyst surface. The absorption in the \u03bd(OH) region (3200 cm\u20131) was attributed to the surface hydroxyl group, and the absorption peak at 3460 cm\u20131 was attributed to the hydroxyl of adsorbed water [27]. The absorption peak at 1345 cm\u20131 is caused by the \u03b4(CH3) of the CH3C=O group, while the absorption peak at 1640 cm\u20131 was assigned to the \u03b4(H2O) vibration. The absorption peak at 1563 cm\u20131 is attributed to \u03bd(COO) of acetic acid [28]. The absorption peak at 982 cm\u20131 is due to peroxide \n\n\nO\n\n2\n\n2\n\u2212\n\n\n on the metal oxide (M-\n\n\nO\n\n2\n\n2\n\u2212\n\n\n) [29]. As shown in Fig.\u00a08d, the peak height of the reactive species M2+-\n\n\nO\n\n2\n\u2212\n\n generated on the 1.0 wt.% MnO2/\u03b3-Al2O3 after 80 min was greater than that on the 0.5 and 10 wt.% MnO2/\u03b3-Al2O3. The M-\n\n\nO\n\n2\n\n2\n\u2212\n\n\n peak was ascribed to an oxygen molecule adsorbed on the surface of MnO2 that was associated with oxygen vacancies and would play an important role in catalytic oxidation reactions of acetic acid [30]. This explained why 1.0 wt.% MnO2/\u03b3-Al2O3 exhibited much better acetic acid ozonation performance than 0.5 and 10 wt.% MnO2/\u03b3-Al2O3. Besides, the COO peak height on the 0.5wt.% MnO2/\u03b3-Al2O3 increased with time, while that of the 1.0 wt.% MnO2/\u03b3-Al2O3 was relatively stable at a low level (Fig.\u00a08e). This suggested that the COO species accumulated on the 0.5 wt.% MnO2/\u03b3-Al2O3 surface and could not be further oxidized due to a lack of reactive oxygen species, leading to an undesirable catalytic ozonation performance. The peak height ratio of COO/\n\n\nO\n\n2\n\n2\n\u2212\n\n\n could be regarded as an indicator to reflect the ability of a catalyst to inhibit the accumulation of COO (acetic acid), and the results shown in Fig.\u00a08f demonstrated that 1.0 wt.% MnO2/\u03b3-Al2O3 had the lowest COO/\n\n\nO\n\n2\n\n2\n\u2212\n\n\n ratio. This implies that the best performance of the 1.0 wt.% MnO2/\u03b3-Al2O3 was enhanced by the surface \n\n\nO\n\n2\n\n2\n\u2212\n\n\n peroxide.To determine the main active species during acetic acid ozonation, TBA and PBQ radical quenchers were added to the acetic acid solution, and the results were shown in Fig.\u00a09\n. It was noted that when the hydroxyl radical (\u00b7OH) quencher TBA (362 mM) was added, the degradation of acetic acid decreased from 88.1% to 15.7%. When the superoxide radical (\n\n\nO\n\n2\n\u2212\n\n) quencher PBQ (1.1 mmol/L) was added, the degradation decreased from 88.1% to 63.2%. This finding suggests that the hydroxyl radical (\u00b7OH) and superoxide radical (\n\n\nO\n\n2\n\u2212\n\n) species contributed to acetic acid degradation, and the \u00b7OH species played a key role. This result was consistent with the results of the effects of pH values described above that a higher energy efficiency was obtained at a lower pH value. The by-products produced during the catalytic ozonation of acetic acid wastewater in the presence of MnO2/\u03b3-Al2O3 was analysed using the HPLC. At 120 min, oxalic acid was found to be the main by-product, and a small amount of formic acid was also detected (Fig.\u00a09c). This was similar to that of the experimental and theoretical research on the decomposition of acetic acid by pulsed DBD plasma reported by Matsui et\u00a0al. [31].The adsorption configuration of acetic acid on metal oxides was a bidentate state [32], and the breakings of the C-H and C-C bands were the primary steps in acetic acid oxidation that produce oxalic acid and formic acid [33]. By combining the results of this study with the researches concerning catalytic ozonation and acetic acid oxidation reported previously [32,34,35,36], the reaction steps and catalytic ozonation mechanism of acetic acid wastewater over 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst were proposed and illustrated in Fig.\u00a010\n. O3 was firstly adsorbed on the surface of MnO2/\u03b3-Al2O3 catalyst, and then decomposed into reactive oxygen species such as surface Mn-\n\n\nO\n\n2\n\u2212\n\n and Mn2+-\n\n\nO\n\n2\n\n2\n\u2212\n\n\n\nEq.\u00a07-(9) [37,38]. Besides, O3 reacts with adsorbed water to form surface hydroxyl group Eq.\u00a010 and (11) and \u00b7OH radicals Eq.\u00a010) [35], which also contributed to acetic acid oxidation (Eq.\u00a013 and (14). The \n\n\nO\n\n2\n\u2212\n\nspecies and \u00b7OH radicals played an important role in the oxidation of acetic acid mineralization into CO2 and H2O.\n\n(7)\nStep 1: Mn-O\u00a0+\u00a0O3\u00a0\u2192\u00a0Mn-O-O3\n\n\n\n\n\n(8)\n\n\n\nStep\n\n2\n:\n\n\nMn\n-\nO\n-\n\nO\n3\n\n\u2192\nMn\n-\n\n\nO\n\n2\n\u2212\n\n+\n\nO\n2\n\n\n\n\n\n\n\n(9)\n\n\n\nStep\n\n3\n:\n\n\n\n\nMn\n\n+\n\n-\n\n\nO\n\n2\n\u2212\n\n+\n\n\nMn\n\n+\n\n-\n\n\nO\n\n2\n\u2212\n\n\u2192\n\n\nMn\n\n\n2\n+\n\n\n-\n\n\nO\n\n2\n\n2\n\u2212\n\n\n+\n\nO\nsurf\n\n\n\n\n\n\n\n(10)\nStep 4: Mn-OH2\u00a0+\u00a0O3\u00a0\u2192\u00a0Mn-OH2-O3\n\n\n\n\n\n(11)\nStep 5: Mn-OH2O3\u00a0\u2192\u00a0Mn-OH\u00a0+\u00a0HO3\n\n\n\n\n\n(12)\nStep 6: HO3\u2219\u00a0\u2192\u00a0\u2219OH+O2\n\n\n\n\n\n(13)\nStep 7: Mn-OH\u00a0+\u00a0CH3COOH\u00a0\u2192\u00a0Mn-OH2\u00a0+\u00a0CH2COOH\n\n\n\n\n(14)\n\n\n\nStep\n\n8\n:\n\n\n\nCH\n3\n\nCOOH\n+\n\u00b7\nOH\n/\n\n\nO\n\n2\n\n2\n\u2212\n\n\n\u2192\nCOOH\n+\n\nCO\n2\n\n+\n\nH\n2\n\nO\n\n\n\n\nThe mineralization of acetic acid by catalytic ozonation was studied, the primary results are summarized as follows: Among the four metal oxides (MnO2, Fe2O3, Co3O4, and CeO2), MnO2 had the best catalytic performance for complete mineralization of acetic acid wastewater. The mineralization of acetic acid reached as high as 88.4% at 300 min with an average energy efficiency of approximately 14.9 g\u00b7kWh\u22121, when 30 g\u00b7L\u22121 of the 1.0 wt.% Mn/\u03b3-Al2O3 catalyst was used to treat 100 mL of acetic acid at a concentration of 1.0 g\u00b7L\u22121 at 25 \u00b0C. A 100 Hz pulse frequency of the DBD reactor was appropriate to produce enough inlet ozone concentration to maintain the mineralization and energy efficiency at a relatively high level. The catalytic ozonation of acetic acid over Mn/\u03b3-Al2O3 catalyst was not sensitive to the reaction temperature, and desirable mineralization of acetic acid could be achieved at an ambient temperature. The Mn/\u03b3-Al2O3 catalyst was efficient for ozone converting into reactive oxygen species such as \u00b7OH and \n\n\nO\n\n2\n\u2212\n\n in the solution and \n\n\nO\n\n2\n\n2\n\u2212\n\n\n on the catalyst surface, and they were essential for the complete mineralization of acetic acid, and Mn/\u03b3-Al2O3 catalyst was an excellent candidate for acetic acid wastewater treatment via catalytic ozonation for practical application.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\nErhao Gao: Investigation, Funding acquisition, Writing \u2013 review & editing. Ruiyun Meng: Data curation, Formal analysis. Qi Jin: Formal analysis. Shuiliang Yao: Conceptualization, Funding acquisition. Zuliang Wu: Formal analysis. Jing Li: Investigation. Erdeng Du: Formal analysis.This research was supported by the National Natural Science Foundation of China (No. 12075037), and Leading Innovative Talent Introduction and Cultivation Project of Changzhou City (CQ20210083).Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.chphi.2022.100149.\n\n\nImage, application 1\n\n\n\n\n\nImage, application 2\n\n\n\n", "descript": "\n                  The complete mineralization of acetic acid in wastewater through a biodegradation process is difficult due to the \u03b1-position methyl coordinated with the carboxyl group, and this work explored the oxidation performance of acetic acid by catalytic ozonation with metal oxides supported on \u03b3-Al2O3. It was found that MnO2/\u03b3-Al2O3 catalyst achieved superior mineralization performance to Co/Fe/CeOx supported on \u03b3-Al2O3 for acetic acid wastewater treatment. The effects of MnO2 loading, catalyst dosage, acetic acid concentration, O3 concentration, ozonation temperature, and initial pH value of the acetic acid solution were investigated. Typically, the mineralization of acetic acid over 1.0 wt.% MnO2/\u03b3-Al2O3 catalyst was as high as 88.4% after 300 min ozonation of 1.0 g\u00b7L\u20121 acetic acid at 25 \u00b0C with the highest energy efficiency around 15 g\u00b7kWh\u20121. By contrast, the mineralization of acetic acid could only reach 43.2% in the absence of the catalyst, with an energy efficiency of 5.1 g\u00b7kWh\u22121. Radical quenchers and indicated that \u00b7OH radical, \n                        \n                           \n                              O\n                           \n                           2\n                           \u2212\n                        \n                      species originated from ozone played an important role in the catalytic ozonation of acetic acid into CO2 and H2O. Besides, the catalytic ozonation mechanism of acetic acid over MnO2/\u03b3-Al2O3 was carefully proposed based on the in situ DRIFTS results.\n               "}
{"full_text": "Data will be made available on request.Ethylene glycol (EG) is an essential product and versatile feedstock for the manufacturing of chemicals, fuels and polymers and so on [1\u20133]. Nowadays, EG is largely dependent on petroleum-derived ethylene in commercial synthesis. Fortunately, a promising approach for the preparation of EG, namely the plant of coal to ethylene glycol (CTEG) has been recently developed rapidly in China (Fig. S1) [4,5]. The process of CTEG includes a crucial step, namely control hydrogenation of dimethyl oxalate (DMO) to EG [6], which also has two main steps including DMO preliminary hydrogenation to methyl glycolate (MG) and then conversion of it to EG (Scheme 1\n) [7\u201310]. Moreover, EG can dehydrate further to ethanol and the final product should meet a rigid ultraviolet transmittance to serve as raw materials for polyester. Currently, DMO production from syngas (CO\u00a0+\u00a0H2) is already commercialized using Pd-based catalyst while the catalyst for DMO hydrogenation still possesses a lot of problems [11], such as short lifetime, toxic promoter and rigorous reaction conditions [12,13].The toxic chromium has been used as the preferred promoter to improve the stability of Cu-based catalysts in industry, which is detrimental to the health [14]. Therefore, strategies to enhance the copper-based catalyst lifespan mainly comprise doping with other elements (Pd, Au, Ag, Ni, Co or B) [12,15\u201322], and modifying the support by using specific oxides (TiO2, La2O3, Al2O3) [23\u201326] or molecular sieve (SBA-15, HZSM) [27,28]. However, these strategies usually lead to a higher costs, complicating fabrication process and even elevating reaction temperature, thus it is not suitable for large-scale production [29].We have previously found that dextrin modified Cu-SiO2 catalysts could largely increase catalytic performance for DMO hydrogenation [30]. The results suggested that the copper species, particularly the copper phyllosilicate (CuPS) and CuO species would be obviously affected by the amount of dextrin. Dextrin is an organic polymer with large molecular weight. However, \u03b2-cyclodextrin (\u03b2-CD) is a water-soluble cyclic oligosaccharide with different structural property, especially its inner hydrophobic cavity and outer OH-groups, may bring variational catalytic performance (Scheme S1B) [31]. It has been reported that the \u03b2-CD and SiO2 are compatible to generate a common phase and thus form mesoporous silica materials [32,33]. Moreover, the binuclear copper (II) complex with \u03b2-CD (Scheme S1C) could be readily prepared as an efficient catalyst for conversion of arylboronic acids [31,34]. Besides, combination of Cu-\u03b2-CD complex and SiO2 has been rarely reported to the vapor-phase hydrogenation reactions. Although we have previously reported two works on the introduction of \u03b2-CD to Cu-SiO2 catalysts for DMO hydrogenation [35,36], their preparation method, silica sources, catalyst morphology, catalytic performance as well as the mechanisms of \u03b2-CD effect on the catalyst were much different. The novelty in this work is to dig out the role of polyhydroxy molecular template in tuning the Cu-SiO2 based catalysts for DMO hydrogenation reactions.At present work, four Cu-SiO2 catalysts were synthesized by the ammonia evaporation method assisted with polyhydroxy molecular templates (Scheme 1). To further investigate the template effect, we also prepared 0.2PEG-Cu-SiO2 catalyst with polyethylene glycol (PEG, Average MW\u00a0=\u00a010,000, Scheme S1A) as template. Experimental results showed that PEG was coated in the catalyst precursors while \u03b2-CD was washed away during catalyst preparation, which was different from the results of dextrin. However, \u03b2-CD modified Cu-SiO2 catalyst was demonstrated to exhibit higher activity, whereas the 0.2PEG-Cu-SiO2 catalyst adversely deteriorated the catalytic performance.All samples (theoretical Cu loading: 25\u00a0wt%) were obtained with ammonia evaporation method. In brief, 10.6\u00a0g of copper nitrate trihydrate and the preset mass fraction (0.2 and 0.5) of \u03b2-CD were mixed into 150\u00a0ml of DI H2O with stirring and ultrasonication. Then the received aqueous ammonia solution (30\u00a0ml) was introduced and stirred. After this step, 21\u00a0g of 40\u00a0wt% colloidal silica was introduced and stirred. In the following step, the above mixture was heated at 90\u00a0\u00b0C to evaporate the ammonia. Then the retained products were filtered and washed with DI H2O. In the end, the samples were dried (120\u00a0\u00b0C, 24\u00a0h) and calcined (450\u00a0\u00b0C, 4\u00a0h). The obtained samples were named as CD-Cu-SiO2.For comparison, the pure Cu-SiO2 and 0.2PEG-Cu-SiO2 samples were prepared similarly to CD-Cu-SiO2 catalyst only except that none adding any template or using PEG instead of \u03b2-CD, respectively. All samples were named as X-Cu-SiO2 catalysts, where X indicates the used templates (PEG, \u03b2-CD, or none).The catalytic activity test was conducted in a fixed-bed reactor, as depicted in the Supporting Information. All the indispensable data about the reference Cu/SiO2 catalyst based on our previous work has been acknowledged [30].The X-Cu-SiO2 catalysts were characterized by inductively coupled plasma optical emission spectrometer (ICP-OES), elemental analyzer (EA), N2 adsorption-desorption method, N2O chemisorption, thermogravimetric analysis (TGA), transmission electron microscope (TEM), X-ray diffraction (XRD), hydrogen temperature-programmed reduction (H2-TPR), hydrogen temperature-programmed desorption (H2-TPD), Fourier-transform infrared (FT-IR), and X-ray photoelectron spectroscopy (XPS). All the detailed characterization information is shown in the Supporting Information.To disclose the effect of different species and template's amount on the catalytic performance, the DMO hydrogenation over the X-Cu-SiO2 catalysts was examined. Fig. 1\n and Table S1 present the results of vapor-phase DMO hydrogenation of DMO as variation of reaction temperature. It shows that the DMO conversion over the 0.2CD-Cu-SiO2 sample was better than other three catalysts at the same reaction condition, especially during the low temperatures of 160 and 190\u00a0\u00b0C. With increasing the temperature from 160 to 190\u00a0\u00b0C, the selectivity of MG drastically decreased while EG increased much (from 28.7% to 97.4%) for 0.2CD-Cu-SiO2 (Fig. 1C). Remarkably, the DMO conversion (Conv.DMO) and EG selectivity (Selec.EG) at 190\u00a0\u00b0C reached maximum value of 99.9% and 97.4%, respectively. Further increasing the temperature, the Selec.EG slightly decreased with the byproducts including ethanol (EtOH), 1,2-butanediol (1,2-BDO) and 1,2-propanediol (1,2-PDO) increasing. Besides, it is clear that the 0.2PEG-Cu-SiO2 sample has extremely low Conv.DMO at low reaction temperature (lower than 220\u00a0\u00b0C). The optimized reaction temperatures (210\u2013270\u00a0\u00b0C) for the reference catalysts are much higher than 0.2CD-Cu-SiO2 (190\u00a0\u00b0C). We further optimized the other reaction parameters (such as time plots, space velocity, hydrogen pressure and the ratio of H2/DMO) for the optimal 0.2CD-Cu-SiO2 catalyst, and the results are shown in Fig. S2 and Fig. 2\n. Fig. S2 indicates that the influence of WLHSVDMO and reaction pressure is not obvious while the effect of H2/DMO is very obvious when it is lower than 50. All the reaction parameters demonstrate that the DMO hydrogenation is beneficial to procced under high reaction pressure and high H2/DMO ratio as well as suitable WLHSVDMO, which is consistent the literature report [2]. If the hydrogen pressure and H2/DMO ratio were too low and WLHSVDMO were too high, the DMO conversion and EG selectivity would be dropped down and generated much MG product.Furthermore, there is no denying that the stability of catalyst is crucial for industrial application and academic research. Fig. 2 shows that the 0.2CD-Cu-SiO2 was much more stable than Cu-SiO2, and it can maintain 99.9% DMO conversion with above 95% EG selectivity within 200\u00a0h of reaction time. However, the reactivity and lifespan of Cu-SiO2 sample was poor under identical reaction conditions, and the catalyst began to deactivate only after 25\u00a0h. Many efforts have been spent on such similar deactivations of Cu-SiO2 catalyst [14,19,24], and the improved reasons would be further discussed below (see Section 3.6).As we all know, the surface composition, surface areas and metal dispersion have significant effects on catalytic performance, thus the physicochemical properties of the X-Cu-SiO2 catalysts are summarized in Table 1\n. The actual Cu contents determined by ICP-OES for all the organic templates modified catalysts were slightly more than unmodified catalyst, but a little lower than the preset values of 25\u00a0wt%. This indicated that the Cu2+ ions weekly adsorbed on the support were lost during filtration and more copper species were coated by silica gel with organic template [19]. The elemental analysis of carbon for the dried catalyst precursors showed that carbon content was below than 0.3\u00a0wt% for both Cu-SiO2 and CD-Cu-SiO2 samples. Interestingly, the 0.2PEG-Cu-SiO2 precursor exhibited 4.17\u00a0wt% carbon loading, suggesting that \u03b2-CD was washed away while the PEG was successfully loaded in the catalyst precursors. Besides, Cu metallic surface area (SCu, 25.8\u00a0m2\u00a0g\u22121) and dispersion (22.5%) determined by dissociative N2O adsorption as well as specific surface area (SBET, 399.4\u00a0m2\u00a0g\u22121) were relatively increased for 0.2CD-Cu-SiO2 catalyst but dramatically decreased for the other organic additive modified catalysts. As mentioned above, the order of catalytic activities is similar to the variation trend of the surface areas. This confirms that SBET is essential for the vapor-phase DMO conversion, which is also proved by a number of reported papers [13,19,37].The N2 physisorption curves of the X-Cu-SiO2 samples are illustrated in Fig. 3\n. It suggests that all the catalysts exhibited Langmuir type IV isotherms with H1-type hysteresis loops. The shape of the hysteresis loops did not change a lot when adding the template, suggesting that introduction of \u03b2-CD or PEG would not affect the silica pore shape. However, their corresponding pore size distributions changed clearly, as illustrated in Fig. 3B. The pore sizes of the 0.2CD-Cu-SiO2 catalyst focus around 4.8\u00a0nm while the 0.2PEG-Cu-SiO2 catalyst focuses around 11.1\u00a0nm. Therefore, a suitable template is beneficial for the formation of higher BET surface area with ordered channel structures.\nFig. 4\n shows TGA plots of the samples under N2 atmosphere from 30 to 900\u00a0\u00b0C. The pure \u03b2-CD lost a weight loss of 13.4% assigned to the water molecule from 30 to 120\u00a0\u00b0C (Fig. 4A). The molecular backbones of \u03b2-CD and PEG then dramatically collapsed before 435\u00a0\u00b0C. As shown in Fig. 4B, the \u03b2-CD modified samples released a little more water molecule before 120\u00a0\u00b0C. However, the 0.2PEG-Cu-SiO2 catalyst showed an obvious degradation of the PEG from 120 to 435\u00a0\u00b0C. Finally, it can be observed that a slight weight loss (7.69%) for pure Cu-SiO2 catalyst, a higher weight loss (11.69% and 10.77%) for CD-Cu-SiO2 catalysts, and the most weight loss (17.53%) for 0.2PEG-Cu-SiO2 sammple during the whole heating process. The weight loss was attributed to the lost of water and copper species. For example, a small amount of CuPS during calcination would decompose to CuO [7,38].Many works reported the viewpoint that the copper species such as CuO particles, Cu-O-Si layer, and CuPS could be influenced by many factors like promoter [13,24], copper loading, [39] copper precursors, [40] and ammonia-evaporation temperature [7]. In particular, Chen et al. reported that more copper phyllosilicate would facilitate the DMO hydrogenation [7]. Thus, the copper species were determined by characterizations including XRD, FT-IR H2-TPR, H2-TPD and TEM.\nFig. 5\n presents the XRD results of the pure templates and the catalysts in which the feature of 2\u03b8 of \u223c22\u00b0 was ascribed to amorphous SiO2. As shown in Fig. 5B, the absence of characteristic diffraction peaks that appeared in the pure templates (Fig. 5A) indicated that the dried catalysts synthesized by organic compounds still lack long range ordering of the structure [33]. It is also because that the amounts of the used templates were low and the crystallinity may also would be affected by adding aqueous ammonia solution [28]. After calcination at 450\u00a0\u00b0C in air, the XRD patterns shows that some small peaks at 2\u03b8 of 31.0\u00b0, 34.8\u00b0 57.2\u00b0 and 63.3\u00b0 were from the CuPS phase (JCPDS 00\u2013003-0219) [41] exhibited in all the calcined catalyst precursors except that the 0.2PEG-Cu-SiO2 catalyst displayed strong diffraction peaks of CuO (JCPDS 05\u20130661) (Fig. 5C) [7,42]. After reduction in H2 atmosphere, the peaks at 2\u03b8\u00a0=\u00a043.3\u00b0, 50.4\u00b0, 74.1\u00b0 and 2\u03b8\u00a0=\u00a036.4\u00b0 are assigned to Cu (JCPDS 04\u20130836) and Cu2O (JCPDS 05\u20130667) species, respectively (Fig. 5D) [7]. This indicated that copper species of 0.2CD-Cu-SiO2 were uniformly dispersed in silica support while there were some large CuO particles in the calcined 0.2PEG-Cu-SiO2 catalyst.The FT-IR results are in accord with analysis from XRD pattern. Fig. 6A shows the FT-IR results of the pure PEG and \u03b2-CD in the 4000\u2013400\u00a0cm\u22121 wavenumber range. The peaks at 2930\u00a0cm\u22121 and 1033\u00a0cm\u22121 can be ascribed to the antisymmetric CH vibration of -CH2 and C-O-C vibrations, respectively. The broad band at \u223c3385\u00a0cm\u22121 is due to the hydroxy stretching vibration. It is interesting that we could not find the main FT-IR peaks of the pure organic template in the dried catalyst precursors from both XRD profiles and FT-IR spectra (Fig. 6B). The absorption bands at 1120, 800, and 470\u00a0cm\u22121 are attributed to the characteristic peaks of SiO2 [43]. Moreover, the presence of \u03b4\nOH vibration (673\u00a0cm\u22121) and \u03bd\nSiO shoulder peak (1030\u00a0cm\u22121) indicates the existence of CuPS [41,44], which were more stronger in 0.2CD-Cu-SiO2 sample but nearly disappeared in 0.2PEG-Cu-SiO2. On the contrary, the 0.2PEG-Cu-SiO2 sample exhibited a new band adsorption at 962\u00a0cm\u22121, indicating a new species of Cu-O-Si units [43]. Besides, it should also be mentioned that the peak of silica support (470\u00a0cm\u22121) would affect the vibrations of the CuO bond that appear at 460, 500, 575\u00a0cm\u22121 [45]. In short, the main copper species on the calcined 0.2PEG-Cu-SiO2 was large CuO particles and Cu-O-Si layer, while they were CuPS dominating on the other three catalyst precursors. After reduction in H2 atmosphere, all the samples had both copper and Cu2O species.The type and amount of copper species were further demonstrated by H2-TPR and H2-TPD profiles. Fig. 7A of the H2-TPR displayed a sharp reduction peak (190\u00a0\u00b0C) for pure Cu-SiO2 catalyst, which corresponds to comprehensive reduction of well dispersed CuO, CuPS, and Cu-O-Si units [7,38]. However, the H2-TPR peak shifts to a 194\u00a0\u00b0C and 201\u00a0\u00b0C with after using the template of \u03b2-CD, which maybe because enhanced chemical interaction between the metal species and SiO2 support [19,46]. A shoulder peak that appeared at 201\u00a0\u00b0C for 0.5CD-Cu-SiO2 may result from the reduction of larger particles of CuO [7]. In addition, the 0.2PEG-Cu-SiO2 exhibited a much higher reduction temperature (230\u00a0\u00b0C), which was assigned to the contribution of bulk CuO instead of well dispersed CuO [7]. Therefore, the H2-TPR results were in agreement with the XRD measurement.The H2 adsorption ability on the X-Cu-SiO2 catalysts were detected by the H2-TPD measurement. The catalysts were firstly activated at 350\u00a0\u00b0C, and showed a main peak centered at around 132\u00a0\u00b0C (Fig. 7B). For comparison, a H2-TPD test for the pure SiO2 displayed a weak H2 desorption peak in the 680\u00a0\u00b0C, indicating that strongly chemisorption H2 species on the SiO2 surface. [24] Therefore, the desorption peak located at 132\u00a0\u00b0C was assigned to chemisorption H2 on Cu species [47]. Besides, the peak at high temperature (737\u00a0\u00b0C) disappeared after adding temperate, suggesting that H2 was more easily released to take part in DMO hydrogenation reaction in the modified samples. The peak intensities of 0.5CD-Cu-SiO2 and 0.2PEG-Cu-SiO2 catalysts were clearly decreased. As identified by N2O chemisorption, the SCu and Cu dispersion for them were also much low even though their actual copper loadings were slightly >0.2CD-Cu-SiO2 sample. In a word, introduction of proper amount of \u03b2-CD into the Cu-SiO2 sample could obviously improve copper species dispersion and H2 activation.The morphology analysis based on TEM images are presented in Fig. 8\n and Fig. S3. Many rod-like copper phyllosilicate with a lamellar structure can be found in pure Cu-SiO2 catalyst (Fig. 8A). After adding a suitable amount of \u03b2-CD during catalyst preparation, the 0.2CD-Cu-SiO2 catalyst has spherical supporting silica structures (light gray) with more dispersed copper species distribution (dark gray, Fig. 8C). However, with more \u03b2-CD or PEG as template, the copper particles became much larger and even crystallized (Fig. 8B and D). After reduction, both pure Cu-SiO2 and 0.2CD-Cu-SiO2 samples showed similar particle sizes (\u223c 3.8\u00a0nm, Fig. 8E and F).The XPS was employed to analyze the surface metal species. The Cu 2p3/2 peaks at above 933.4\u00a0eV with the satellite peaks indicate that the chemical valence of Cu was +2 in the calcined samples (Fig. S4A). After deconvolution processing of the asymmetric Cu 2p3/2 envelope, two peaks at 934.2 and 936.4\u00a0eV are ascribed to dispersed CuO and CuPS, respectively [7,48,49]. Notably, it is important to emphasize that the proportion of CuPS is much more than CuO in the 0.2CD-Cu-SiO2 and inversely in the 0.2PEG-Cu-SiO2 catalyst. After reduction, the satellite peaks of Cu 2p disappeared (Fig. S4B). Fig. 9\n and Table 2\n show the Cu LMM XAES spectra and deconvolution results of the reduced catalysts to distinguish the surface Cu+ and Cu0 species. We have also noted that the Cu+/(Cu+ + Cu0) ratio (named as XCu+) increases (63.2% vs. 75.5%) with introduction of \u03b2-CD and decreases (63.2% vs. 49.8%) with using PEG as template. This result further proves that the interaction between the improved CuPS and the SiO2 carrier was increased significantly.Based on the above characterization results, it clearly indicated that the structure of \u03b2-CD modified Cu-SiO2 catalysts were different from dextrin coated Cu-SiO2 catalysts [30], which possessed smaller SBET, lower amount of CuPS and decreased metal dispersion. Due to the same preparation procedure, this phenomenon could be attributed to the different roles of polyhydroxy compounds in ammonia evaporation method. In our previous work, we have illustrated that dextrin could successfully coated copper nanoparticles during the catalyst preparation. However, this work has failed to generate Cu-\u03b2-CD complex, so that \u03b2-CD could be easily washed away during filtration. One reason was probably that the reported Cu-\u03b2-CD complex was formed in strong alkaline conditions like using NaOH instead of NH3\u00b7H2O [34,50]. Another reason was that the formation of Cu-\u03b2-CD complex was through a reversible reaction (Scheme 2\n, eq. 0). When the ammonia evaporation proceeding, the OH\u2212 concentration decreased so that even formed Cu-\u03b2-CD complex would disassemble again. Matsui et al. disclosed a ligand-exchanged reaction of Cu-\u03b2-CD complex with EDTA to give Cu(EDTA)2\u2212 and \u03b2-CD, suggesting that Cu-\u03b2-CD complex was not very stable [50].However, compared with PEG, the \u03b2-CD was not simply added and it should have been interacted with copper species, as evidenced by the more ordered pore structure and the improved copper phyllosilicate morphology from TEM images (Fig. 8B). This could be ascribed to the special structure of \u03b2-CD [33,51]. Because of the exterior hydrophilic surface of \u03b2-CD, they form a sphere-like structure with the connection of hydroxyls. Then the Cu2+ ions were cooperated with the nanosphere \u03b2-CD [52]. Thus 0.2CD-Cu-SiO2 catalyst possessed nanosphere copper particles instead of rod-like copper phyllosilicate. At a deeper level, copper phyllosilicate is composed of the alternate layers of SiO4 tetrahedra and discontinuous layers of CuO6 octahedra [17]. The synthesis of CuPS would be affected by a lot of factors, such as the pH of the precursor solution [41], the solution/silica contact time [41], CuO loading [53], and the ammonia evaporation temperature or time [7]. This work showed that the polyhydroxy compounds also have effect on its formation, as proved by TEM, XRD and FT-IR results. It is reported that there are three equilibrium reactions for the formation of CuPS (Scheme 2, eq. 1\u20133): [17,41,53] 1) The silica sol was dissolved to yield silicic acid (Si(OH)4); 2) The copper ammonia complex was hydrolyzed to Cu(OH)2(H2O)4 as ammonia evaporation; 3) The heterocondensation reaction of silica acid with the Cu(OH)2(H2O)4. Finally, CuPS monomer would polymerize and chemically interact with the SiO2 surface. We speculated that some Cu2+ ions were coated by dextrin or PEG, decreasing neutral metal complex. While \u03b2-CD probably inhibited the growing of copper phyllosilicate monomer to form rod-like structure owing to that \u03b2-CD forming a sphere-like structure. Therefore, polyhydroxy compounds play a key role in ammonia evaporation method, especially for copper phyllosilicate morphology.The topic about DMO hydrogenation has been published in numerous literatures (Table S2). Nevertheless, many reported Cu-based catalysts exhibited good catalytic performance for EG synthesis under conditions of lower WLHSVDMO (0.6 vs. 1.0\u00a0h\u22121), higher H2/DMO molar ratio (80 vs. 50), copper loading (30 vs. 17.78\u00a0wt%), P(H2) (3.0 vs. 2.0\u00a0MPa) or reaction temperature (200 vs. 190\u00a0\u00b0C). In conventional Cu-based catalysts for EG synthesis, if the reaction temperature were below 200\u00a0\u00b0C, the intermediary MG would be largely generated, accelerating catalyst deactivation [23]. This synthetic methodology employing low cost Cu-based catalysts with \u03b2-CD could become useful in industry given its good activity and stability at low temperature.The reasons for the excellent performance of 0.2CD-Cu-SiO2 catalyst are analyzed below. One is elevated Cu metallic dispersion and SBET, as well as more ordered channel structures. Another cause is that the improved copper phyllosilicate morphology, resulting in a higher XCu+ ratio (75.5%). Zhang et al. have reported on the optimizing in the catalytic property of Cu/10-SiO2, which was because of the suitable SBET, Dp and larger Cu dispersion [37]. Both the SBET (399.4\u00a0m2\u00a0g\u22121) and SCu (25.8\u00a0m2\u00a0g\u22121) were increased much for the 0.2CD-Cu-SiO2 catalyst so that its activity was enhanced a lot (Table 1), especially at low reaction temperature of 190\u00a0\u00b0C. Moreover, the H2-TPR and H2-TPD results obviously showed that the chemical interactions between the copper species and the silica carrier were increased and H2 activation became more easily than pure Cu-SiO2 (Fig. 7). Thus, the sintering of copper species could be slowed during the reaction. Moreover, a little carbon deposit on the catalyst surface plays significant roles in eliminating some very active species, which maybe favorable for over hydrogenation to ethanol and carbon deposition during reaction. Just as the role of catalyst pretreatment by DMO hydrogenation at high reaction temperature (450\u00a0\u00b0C for 24\u00a0h, the common reaction temperature is 180\u2013210\u00a0\u00b0C) to eliminate some very active species [19]. Therefore, the \u03b2-CD modified Cu-SiO2 catalyst presented better stability. When the mass fraction of \u03b2-CD was >0.2, the solvent was not enough to well dissolve it and finally the aggregation of copper nanoparticles happened (Fig. 7A and Fig. 8D). In addition, we also need to point out the reasons for the very poor activity of 0.2PEG-Cu-SiO2 sample. The large crystalline CuO bulk particles observed from the XRD, H2-TPR, and TEM images were the primary reason. As the carbon content was detected on the dried 0.2PEG-Cu-SiO2 precursor (Table 1), we propose that the copper species were coated by the PEG template instead of promoting the interaction between copper species and SiO2 carrier, resulting a poor SBET (258.5\u00a0m2\u00a0g\u22121) and SCu (9.3\u00a0m2\u00a0g\u22121) (Table 1). The lower copper surface area and larger copper nanoparticles resulted in a worse activity [54].Based on the XRD results, we know that all the copper oxides in the catalyst precursors were reduced to both Cu and Cu2O. The XPS spectra further demonstrated different Cu+/(Cu+ + Cu0) ratio among them. Much work so far has focused on the viewpoint that the synergic effect of Cu+ and Cu0 is effective in hydrogenation of DMO [7,19,39,55,56]. Generally, Cu+ sites bind and activate the ester and acyl groups while Cu0 sites absorb H2 molecules in ester hydrogenation [57]. The improved structure of 0.2CD-Cu-SiO2 was indicated to be capable of increasing the XCu+ (75.5%) than that on the pure Cu-SiO2 (63.2%). As the CuO species were easily reduced to Cu0, the XCu+ (49.8%) of 0.2PEG-Cu-SiO2 was decreased much [58]. Though the Cu+ species was also high for the 0.5CD-Cu-SiO2 sample (72.1%), it also displayed a poor activity. This is because that many other factors would also affect their catalytic results in addition to the XCu+ [38]. As shown in Table 1, although the actual Cu loading over 0.5CD-Cu-SiO2 was slightly higher than 0.2CD-Cu-SiO2, the Cu dispersion and SCu as well as SBET over 0.2CD-Cu-SiO2 was much higher than those over 0.5CD-Cu-SiO2. In addition, the Cu nanoparticles size was obviously increased from 3.7 to 6.8\u00a0nm over 0.5CD-Cu-SiO2 (Fig. 8). Therefore, the DMO hydrogenation on 0.2CD-Cu-SiO2 was more active than 0.5CD-Cu-SiO2.A green and facile route for the preparation of efficient Cu-SiO2 catalysts was demonstrated using low-cost \u03b2-CD as template. Unlike dextrin, the \u03b2-CD was not coated but successfully washed during catalyst preparation, leading to improving copper phyllosilicate morphology, increasing Cu dispersion and surface Cu+ species. Moreover, the modified structural properties of the 0.2CD-Cu-SiO2 could efficiently maintain the Cu dispersion without the deactivation. As a result, it possesses remarkable performance for DMO hydrogenation to ethylene glycol at 190\u00a0\u00b0C. However, the template of polyethylene glycol would obviously revise the distributions of copper species with formation of more Cu0 species and thus lower the DMO hydrogenation performance. Therefore, this synthetic strategy using \u03b2-CD for the modified Cu-based catalysts could become a suitable candidate in industry.\nRunping Ye: Conceptualization, Investigation, Writing \u2013 original draft. Chong Zhang: Software, Data curation, Formal analysis. Peng Zhang: Investigation, Data curation. Ling Lin: Conceptualization, Project administration, Writing \u2013 review & editing. Long Huang: Methodology, Investigation, Writing \u2013 review & editing. Yuanyuan Huang: Methodology, Investigation, Writing \u2013 review & editing. Tianyou Li: Methodology, Investigation, Writing \u2013 review & editing. Zhangfeng Zhou: Methodology, Investigation, Writing \u2013 review & editing. Rongbin Zhang: Methodology, Investigation, Writing \u2013 review & editing. Gang Feng: Writing \u2013 review & editing, Supervision, Funding acquisition. Yuan-Gen Yao: Writing \u2013 review & editing, Supervision, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by National Natural Science Foundation of China (Grants No. 22005296, 22102186, 21875096, and 21763018), National Key Research and Development Program of China (2017YFB0307301, 2017YFA0206802, 2018YFA0704500), and Strategic Priority Research Program of the Chinese Academy of Sciences (XDA21020800). The authors are also thankful for and Shiyanjia Lab (www.shiyanjia.com) on the XPS analysis.\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nThe following are the Supplementary data to this article. The introduction of plant of coal to ethylene glycol (Fig. S1), the influence of reaction parameters on the catalytic performance (Fig. S2), the details of catalyst characterizations, the details of catalytic performance test, the molecular structures of polyethylene glycol and \u03b2-cyclodextrin (Scheme S1), catalytic performance of DMO hydrogenation over X-Cu-SiO2 catalysts (Table S1), HRTEM images for X-Cu-SiO2 catalysts (Fig. S3), XPS spectra of X-Cu-SiO2 catalysts (Fig. S4), comparison of DMO hydrogenation over different CuSi catalysts (Table S2). Supplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106586.", "descript": "\n                  \u03b2-cyclodextrin (\u03b2-CD) was used to prepare remarkable and stable copper-based catalysts for hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG). The DMO conversion (95.6% vs. 99.9%), EG selectivity (53.6% vs. 97.4%) and lifetime (25\u00a0h vs. 200\u00a0h) were significantly increased on the 0.2CD-Cu-SiO2 sample compared with Cu-SiO2 at 190\u00a0\u00b0C. The prominent catalytic property was due to the significant roles of \u03b2-cyclodextrin to regulate surface dispersion of copper species along with their particle sizes, namely improved copper phyllosilicate morphology, increased BET surface area, improved Cu dispersion, and enhanced surface ratio of Cu+/(Cu+ + Cu0).\n               "}
{"full_text": "As the primary long-term drivers of climate change, the concentrations of the greenhouse gases CO2 and CH4 in the atmosphere are growing continuously, as reported by the Global Monitoring Division of the National Oceanic and Atmospheric Administration (NOAA)/Earth System Research Laboratory. Climate change makes it more attractive to reutilize CO2 and CH4 for the development of a carbon-neutral ecosystem. Great efforts are being directed at the chemical transformation of CO2 and CH4 into syngas [1\u20136], a mixture of CO and H2, that can be subsequently converted to fuels and chemicals using existing commercial processes [7\u201311]. However, this indirect route for CO2 and CH4 utilization is energy-intensive due to the high thermodynamic stability of the two molecules. Therefore, developing a single-step approach to convert CO2 and CH4 to fuels and chemicals under mild conditions, instead of syngas being the intermediary, is highly desirable.Non-thermal plasma (NTP) has been proved to be an effective way to directly convert CO2 and CH4 into high-value oxygenates such as alcohols, acids, ketones and aldehydes, under mild conditions through overcoming thermodynamic and kinetic limitations [12\u201314] and thus bypassing the process of syngas production. Earlier studies on converting CO2 and CH4 to oxygenates were carried out by Liu and co-workers using a dielectric barrier discharge (DBD) reactor. They found that the total selectivity of oxygenates, which included alcohols, acids, ketones, and aldehydes, changed with plasma parameters and reaction conditions, such as discharge gap, CH4/CO2 feed ratio, and starch as a dielectric layer. Pasting corn starch on the surface of the quartz tube reactor enhanced the formation of oxygenates, while inhibiting the generation of liquid hydrocarbons from CH4 and CO2\n[15,16]. Tosi et al. investigated the synthesis of acids from a CO2-CH4 DBD plasma, and discovered that the ratio of propanoic-butanoic acids to acetic acid increased with discharge duty cycle. In addition, Ni and Cu electrodes produced more carboxylic acids, particularly formic acid, than stainless steel. Furthermore, the possible mechanisms for acid generation were investigated by density functional theory (DFT) calculations [17\u201319]. Torsten et al. revealed the importance of O2 in the formation of methanol and formaldehyde during the plasma conversion of CO2 and CH4 in a DBD reactor using He as a diluent gas [20]. Krawczyk et al. reported that only methanol and ethanol were produced in the plasma-catalytic reaction of CO2 and CH4 with and without packing materials (i.e., Al2O3, Fe/Al2O3, NaY and NaZSM-5) in a DBD reactor, and the highest selectivity of alcohols (< 3.5%) was obtained in the non-packing system [21]. Rahmani et al. discovered that oxygenates produced in the plasma reforming of CO2 and CH4 contained 10 liquid fuels such as alcohols, ketones, and light organic acids. The formation of oxygenates decreased when the Ar addition was higher than 33%, and the total yield of oxygenates represented 2\u20134% of the total mass of the products [22]. Shirazi et al. performed DFT calculations to investigate the pathway in the plasma conversion of CO2 and CH4 to methanol on a crystalline Ni(111) surface [23], and they found that aldehydes were preferably formed in a CH4/CO2 DBD reactor in comparison with alcohols using a one-dimensional fluid model [24]. In addition, Lambert et al. reported formaldehyde generation with a selectivity of 11.4% in the plasma-catalytic conversion of CO2 and C2H6 over BaTiO3 supported vanadia/alumina catalysts [25]. Similar studies were also carried out by Chen and co-workers, who found that alcohols, aldehydes, and acids could be obtained with a maximum total oxygenate selectivity of 12%, which agreed with the results of chemical kinetic modeling [26].Very recently, significant progress has been made in plasma-driven CO2 and CH4 conversion to oxygenate, with a focus on catalyst investigations. Li et al. reported that packing Fe/SiO2 or Co/SiO2 catalyst into a DBD reactor greatly increased the formation of oxygenates to 40% selectivity, with methanol and acetic acid being the main products, whereas syngas and C2-C5 gaseous hydrocarbons were generated as the main products in the non-packing mode [27]. They also found that Fe/SiO2 promoted the generation of alcoholic products, while Co/SiO2 favored the formation of acids and long-chain products. Song et al. investigated the effect of catalysts\u2019 acidic properties on the total selectivity of oxygenates and carbon deposition in the plasma-catalytic conversion of biogas (a mixture of CO2 and CH4) [28], and the Pt/UZSM5 catalyst with a 100% ratio of weak acidic sites exhibited an oxygenates selectivity of up to 60%, including formaldehyde, methanol, ethanol, and acetone, implying that a high ratio of weak acidic sites benefits the formation of oxygenates. Shao and co-workers reported that packing Ni-based catalysts into the plasma region achieved around 30% selectivity of oxygenates, which is dependent on the microstructures, surface compositions, and reducibilities of the catalysts [29]. Our previous work achieved a liquid selectivity of 50\u201360% with acetic acid being the major liquid product through the development of a novel DBD reactor with a special water ground electrode [30].Despite great advances in the direct conversion of CO2 and CH4 to oxygenates using plasma technology, significant challenges remain. Firstly, the oxygenates that have been obtained to date consist of alcohols, acids, ketones, and aldehydes, and how to tailor the distribution of oxygenates is yet unclear. Second, current research on the catalyst design for this process is still in its infancy, and the correlation between catalytic active sites and specific oxygenate is still missing.Herein, a series of Cu-based catalysts were synthesized and tested in the conversion of CO2 and CH4 to oxygenates using a dielectric barrier discharge (DBD) plasma reactor at a reaction temperature of 60\u00a0\u00b0C and atmospheric pressure. The influence of the Cu-based catalysts, i.e., the electronic structure and acidity, on the formation of alcohols and acids was investigated. We use the support on which copper is anchored, i.e., Al(OH)3, Mg(OH)2, SiO2, TiO2 and HZSM-5 zeolite, to tune the valence state of copper species in the Cu-based catalysts, and use the preparation method of ion exchange and impregnation to regulate the valence state and Br\u00f8nsted acid sites of the Cu/HZSM-5 catalyst. The correlations between acids/alcohols and copper valence state are well established. Cu2+ was found to be more favorable for the formation of alcohols, whereas Cu+ contributes more to the production of acetic acid. Moreover, in addition to Cu+, Br\u00f8nsted acid sites of HZSM-5 can significantly improve the selectivity of acetic acid.The catalysts with 10\u00a0wt% Cu loading were synthesized using an incipient wetness impregnation method (IM). The HZSM-5 support was first calcined at 400 \u00b0C for 5\u00a0h to remove the impurities (e.g., adsorbed H2O). The precursor Cu(NO3)2\u00b73H2O was dissolved in deionized water, followed by the addition of HZSM-5 powder under stirring. After 5\u00a0h of aging at room temperature, the sample was dried at 120\u00a0\u00b0C overnight. Finally, the sample was calcined at 540\u00a0\u00b0C for 5\u00a0h in air, and the catalyst is noted as Cu/HZSM-5. Other catalysts, i.e., Cu/Al(OH)3, Cu/Mg(OH)2, Cu/SiO2 and Cu/TiO2, were synthesized using the same method.The Cu/HZSM-5 catalyst was also prepared using an ion exchange method (IE). The HZSM-5 zeolite was first calcined at 400\u00a0\u00b0C for 5\u00a0h to remove the impurities (e.g., adsorbed H2O), and the precursor Cu(NO3)2\u00b73H2O was dissolved in deionized water before adding HZSM-5 powder under stirring. The suspension was stirred at 80\u00a0\u00b0C for 2\u00a0h, before being filtered and washed with deionized water, and the process was repeated twice. Finally, the paste was dried at 120\u00a0\u00b0C for 12\u00a0h before being calcined in air at 540\u00a0\u00b0C for 3\u00a0h. The resulting sample is denoted as Cu/HZSM-5 (IE), while the Cu/HZSM-5 catalyst prepared by the incipient wetness impregnation method is denoted as Cu/HZSM-5 (IM). The as-prepared catalysts were used directly for plasma-catalytic conversion of CO2 and CH4 without any pretreatment.Conversion of CO2 and CH4 was carried out in a dielectric barrier discharge (DBD) catalytic reactor at low temperatures and atmospheric pressure, as illustrated in Fig. S1. The DBD reactor was made up of two coaxial glass cylinders with water circulating between the inner and outer cylinders, as well as two coaxial electrodes. The circulating water served as the ground electrode and was controlled at around 60\u00a0\u00b0C by a thermostat water bath. The inner high voltage electrode was a stainless-steel rod with an outer diameter (o.d.) of 2\u00a0mm, fitted along the axis of the inner glass tube (10\u00a0mm o.d. \u00d7 8\u00a0mm i.d.), which also functioned as the dielectric material. The discharge length was 40\u00a0mm, with a 3\u00a0mm discharge gap, and the catalyst (20\u201340 mesh) was packed into the entire discharge area. The temperature of catalyst bed was measured to be in the range of 60\u2013140\u00a0\u00b0C using a thermal infrared camera, as shown in Fig. S2. The DBD reactor was powered by an alternating current (AC) high voltage power source with a maximum peak voltage of 30\u00a0kV and a variable frequency range of 7\u201315\u00a0kHz. In this work, the flow rate of CO2 and CH4 was controlled by mass flow controllers with a CH4/CO2 ratio of 1:1 at a total flow rate of 40\u00a0mL/min. The discharge frequency was set at 9\u00a0kHz, and the reaction lasted 2\u00a0h. A four-channel digital oscilloscope (Tektronix, DPO 3012) with a high-voltage probe (Tektronix, P6015) and a current probe (Pearson 6585) was used to record the electrical signals. A cold trap consisted of anhydrous ethanol and liquid nitrogen was connected to the exit of the reactor to condense liquid products, which were quantified using a gas chromatograph (Shimadzu GC-2014\u00a0C) equipped with a flame ionized detector (FID) with an FFAP column. The gaseous products were analyzed using an online gas chromatograph (Shimadzu GC-2014\u00a0C) equipped with a thermal conductivity detector (TCD) and an FID. The change in gas volume before and after the reaction was monitored using a flowmeter.In this study, the conversion of CO2 and CH4, as well as the selectivity of main gaseous products (CO, C2-C4 hydrocarbons) and major liquid products (acetic acid and C1-C4 alcohols) are used as performance indicators. The specific calculation methods are shown as follows.\n\n\n\n\nX\n\nCO\n2\n\n\n\n%\n\n=\n\n\nmoles\n\nof\n\n\nCO\n2\n\n\nconverted\n\n\nmoles\n\nof\n\ninitial\n\n\nCO\n2\n\n\n\n\u00d7\n100\n%\n\n\n\n\n\n\n\n\n\nX\n\nCH\n4\n\n\n\n%\n\n=\n\n\nmoles\n\nof\n\n\nCH\n4\n\n\nconverted\n\n\nmoles\n\nof\n\ninitial\n\n\nCH\n4\n\n\n\n\u00d7\n100\n%\n\n\n\n\n\n\n\n\n\nS\nCO\n\n\n%\n\n=\n\n\nmoles\n\nof\n\nCO\n\nproduced\n\n\nmoles\n\nof\n\n\nCH\n4\n\n\nconverted\n+\nmoles\n\nof\n\n\nCO\n2\n\n\nconverted\n\n\n\u00d7\n100\n%\n\n\n\n\n\n\n\n\n\nS\n\n\nC\nx\n\n\nH\ny\n\n\n\n\n%\n\n=\n\n\nx\n\n\u00d7\n\nmoles\n\nof\n\n\nC\nx\n\n\nH\ny\n\n\nproduced\n\n\nmoles\n\nof\n\n\nCH\n4\n\n\nconverted\n+\nmoles\n\nof\n\n\nCO\n2\n\n\nconverted\n\n\n\u00d7\n100\n%\n\n\n\n\n\n\n\n\n\nS\noxygenates\n\n\n%\n\n=\n100\n%\n\u2212\n\n\n(\nS\n\nCO\n\n+\n\nS\n\n\nC\nx\n\n\nH\ny\n\n\n\n)\n\u2212\nca\n.\n10\n%\ncarbon\ndeposition\n\n\n\n\n\n\n\n\n\nS\n\n\nCH\n3\n\nCOOH\n\n\n(\n%\n)\n=\n\n\n2\n\n\u00d7\n\nmoles\n\nof\n\n\nCH\n3\n\nCOOH\n\nproduced\n\n\nmoles\n\nof\n\ntotal\n\noxygenates\n\nproduced\n\n\n\u00d7\n\nS\noxygenates\n\n\n\n\n\n\n\n\n\n\nS\n\n\nC\n\n1\n\u2212\n4\n\n\nOH\n\n\n(\n%\n)\n=\n\n\ncarbon\n\nnumber\n\n\u00d7\n\nmoles\n\nof\n\nalcohols\n\nproduced\n\n\nmoles\n\nof\n\ntotal\n\noxygenates\n\nproduced\n\n\n\u00d7\n\nS\noxygenates\n\n\n\n\n\nThe physicochemical properties of the as-prepared catalysts were analyzed using following techniques. The crystalline structure of the catalysts was determined by X-ray powder diffraction (XRD) using an X-ray diffractometer (Rigaku D-Max 2400) with Cu K\u03b1 radiation (\u03bb\u2009=\u20090.15406\u2009nm). The scanning range was from 5 to 80 (2\u03b8), with a step size of 0.02\u2009min\u22121 and a scanning speed of 10\u2009min\u22121. The valence state and chemical environment of copper species of the catalysts were examined by X-ray photoelectron spectroscopy (XPS) using a Thermo ESCALAB Xl+\u2009spectrometer. High-resolution transmission electron microscopy (HRTEM) images of the catalysts were recorded using a JEOL-JSM-2100\u2009F (Tecnai G2 F30 S-Twin) with an energy dispersive X-ray spectrometer (EDXS) at an accelerating voltage of 200\u2009kV. Scanning transmission electron microscopy-high angle annular dark field (STEM-HAADF) analysis was performed using the same apparatus as HRTEM. The redox properties of the catalysts were investigated by hydrogen temperature-programmed reduction (H2-TPR) using a chemisorption instrument (Quantachrome ChemBET 3000). TPR measurements were carried out in an H2/Ar mixture flow (10\u2009vol%\u2009H2, 120\u2009mL/min) from room temperature to 800\u2009\u00b0C at a heating rate of 10\u2009\u00b0C/min. A TCD detector measured the H2 concentration in the tail gas, which reflected the H2 consumption as a function of temperature. The acidity of the catalysts, i.e., the amount, strength, and type of acid sites, was studied by NH3 temperature-programmed desorption (NH3-TPD) and Pyridine adsorption FTIR (Py-FTIR). NH3-TPD was performed using the same system as for H2-TPR. The sample was purged with NH3 for 1\u2009h at 100\u2009\u00b0C, followed by purging with He to remove NH3 physically adsorbed on the catalyst. Finally, the sample was heated to 800\u2009\u00b0C under the flowing He and the composition of the effluent gas was measured by a TCD detector. Py-FTIR spectra in absorbance mode were collected accumulating 64 scans with 4\u2009cm\u22121 resolution on a Nicolet 6700 spectrometer equipped with an MCT cryodetector, cooled by liquid nitrogen. All samples were pressed into the form of self-supporting wafers and pretreated in situ for 4\u2009h under vacuum conditions (10\u22123 Pa) in a homemade quartz IR cell with CaF2 windows at 673\u2009K. Subsequently, pyridine vapor was exposed to the sample at room temperature for 30\u2009min and then the system was outgassed at 200\u2009\u00b0C for 30\u2009min, before being cooled to room temperature for measurement.The emission spectra of the CO2-CH4 plasma at different conditions were recorded using a Princeton Instruments ICCD spectrometer (SP2758) in the 200\u20131200\u2009nm range via an optical fiber facing the outside wall of the DBD reactor. The slit width and grating of the spectrometer were set to 20\u2009\u00b5m and 300\u2009g\u2009mm\u22121, respectively.Oxygenates, i.e., acetic acid (CH3COOH), alcohols (C1-C4OH), acetaldehyde (CH3CHO) and acetone (C3H6O), with acetic acid and alcohols being the major products were directly produced at mild conditions in this study, correspondingly denoted as R-COOH and R-OH. A small amount of acetaldehyde and acetone were also detected with selectivities less than 2% and 1%, respectively. Furthermore, no C5+ hydrocarbons were produced during this process. Therefore, our focus in this study is to get new insights into the formation of R-OH and R-COOH, as the formation of other liquid products is limited. \nFig. 1 shows the effect of Cu-based catalysts and the corresponding supports on the generation of R-OH and R-COOH. Clearly, there is no rule regarding the effect of supports on the selectivity of oxygenates when packing pure support material into the DBD reactor. The distribution of R-OH and R-COOH selectivity, however, changed regularly after anchoring copper on the corresponding support. That is, the selectivity of R-OH using the Cu-based catalysts decreased in the order of Cu/Al(OH)3 >\u2009Cu/Mg(OH)2 >\u2009plasma only >\u2009Cu/SiO2 >\u2009Cu/TiO2 >\u2009Cu/HZSM-5\u2009>\u2009HZSM-5, whereas the selectivity of R-COOH decreased in the opposite order. Therefore, Cu anchored to these materials can significantly tune the distribution of oxygenates. Moreover, when compared to the reaction using plasma only, the coupling of DBD with Al(OH)3, Cu/Al(OH)3 or Cu/Mg(OH)2 favored the production of R-OH, but inhibited the formation of R-COOH. This phenomenon was more pronounced for Cu/Al(OH)3. On the contrary, using Cu/HZSM-5 or HZSM-5 inhibited the formation of R-OH whilst enhancing the production of R-COOH. It is worth noting that packing Mg(OH)2 or TiO2 into the plasma area had a clear negative effect on the formation of oxygenates, but the generation of oxygenates was greatly enhanced by anchoring Cu.In addition, all of the cases gave a similar distribution of gaseous products, with CO and C2-C4 hydrocarbons being the most common gaseous products (Fig. S3). Moreover, Cu anchored on the supports slightly influenced the conversion of CO2 and CH4, as shown in Fig. 1(c) and (d). However, the CH4 conversion was almost twice the CO2 conversion in all cases, which agrees well with the results of the 0D chemical kinetic modeling of the CO2/CH4 plasma reaction [31]. A possible explanation for the lower CO2 conversion than CH4 is the regeneration of CO2 in plasma. Yao et al. reported that CO2 can be easily regenerated through the reaction CO +\u2009OH \u2192 CO2 +\u2009H in CO2/CH4 plasma [32], but Snoeckx et al. reported the charge transfer reaction, CO2\n+ + CH4 \u2192 CH4\n+ + CO2, contributes to 99% of the regeneration of CO2, which is considered to be the main reason for lower CO2 conversion than CH4\n[31].To understand the difference in the selectivity between R-OH and R-COOH over Cu-based catalysts in Fig. 1(a), a series of techniques, such as XRD, XPS, HRTEM, H2-TPR and NH3-TPD, were used to characterize the as-prepared catalysts. As inferred from XRD profiles in \nFig. 2(a), the as-prepared Cu-based catalysts show exclusively CuO phase with dominate characteristic peaks at 35.5\u00b0 and 38.8\u00b0 (JCPDS 48-1548), assigned to (11\u22121) and (111) plane of CuO, respectively [33].\nFig. 2(b) displays the reducibility of the as-prepared Cu-based catalysts changed when using different supports, reflecting the different interactions between copper species and support, i.e., Cu-O-M interfaces are formed (M is Al, Mg, Si, Ti and Al/Si). The Cu/Mg(OH)2 catalyst showed the highest reduction temperature, revealing that Mg(OH)2 has the strongest interaction with copper species. By contrast, Al(OH)3 and HZSM-5 exhibited weak interactions with copper species. Except for Cu/HZSM-5, other Cu-based catalysts showed a broad reduction peak, which can be assigned to the reduction of crystalline CuO since it follows a one-step reduction mechanism, i.e., CuO +\u2009H2 \u2192 Cu0 +\u2009H2O. Whereas the reduction of copper species on Cu/HZSM-5 is more complicated since it contains the reduction of both crystalline CuO and isolated Cu2+ species. In contrast to crystalline CuO, the reduction of isolated Cu2+ species is characterized by a two-step mechanism, i.e., Cu2+ +\u00a01/2\u2009H2 \u2192 Cu+ + H+ and Cu+ +\u00a01/2\u2009H2 \u2192 Cu0 +\u2009H+\n[34\u201336]. Thus, the results of H2-TPR reveal the crystalline CuO is the major copper species in as-prepared catalysts, which is consistent with the CuO crystalline phase detected in XRD patterns.Furthermore, the HRTEM images of the catalysts showed the lattice fringes of CuO (\nFig. 3), with d-spacings of 0.25\u2009nm and 0.23\u2009nm, respectively, obtained from interference fringes, representing CuO (11\u22121) and CuO (111) plane, respectively. Apart from the cases of Al(OH)3 and SiO2 supports, the interfaces between CuO and catalyst supports can be observed in their HRTEM images, which confirms the formation of Cu-O-M interfaces as inferred from H2-TPR profiles,. Because both Al(OH)3 and SiO2 are amorphous, it is difficult to observe the interface between CuO and Al(OH)3 or SiO2 through the lattice plane in their HRTEM images.In addition to crystalline CuO as identified above, Cu 2p3/2 XPS fitting curves shown in \nFig. 4(a) reveal that surface Cu species in all as-prepared catalysts exist in both Cu2+ and Cu+ forms [37,38]. However, the relative distribution of Cu2+ and Cu+ varied with the type of the support according to the deconvolution of the Cu 2p3/2 major peak, and the amount of Cu2+ species followed the descending order of Cu/Al(OH)3 >\u2009Cu/Mg(OH)2 >\u2009Cu/SiO2 >\u2009Cu/TiO2 >\u2009Cu/HZSM-5. Moreover, the amount of Cu2+ can also be estimated from the peak area of the Cu2+ satellite peak, since the Cu2+ species are featured by its large satellite peak, whereas satellite peaks are absent in the spectra of Cu+ or Cu0 species. As shown in Fig. 4(b), the peak area of the Cu2+ satellite peak, with binding energies between 940 and 945\u2009eV, decreased in the order of Cu/Al(OH)3 >\u2009Cu/Mg(OH)2 >\u2009Cu/SiO2 >\u2009Cu/TiO2 >\u2009Cu/HZSM-5, which agrees well with the changing tendency of Cu2+ amount obtained by deconvolution of Cu 2p3/2 main peak. Among these as-prepared catalysts, the copper supported on Al(OH)3 had the most surface Cu2+ and the least surface Cu+, whereas the copper supported on the HZSM-5 zeolite showed the complete opposite results.In addition, Fig. 4(a) shows that the binding energy of the Cu 2p3/2 varied with support, with varying magnitudes. When comparing Cu/Al(OH)3 to Cu/Mg(OH)2, a noticeable shift to lower binding energy of Cu2+ was observed, which can be further confirmed by the down-shift binding energy of the Cu2+ satellite peak. Espin\u00f3s et al. found that the binding energies of copper supported metal oxides are very sensitive to copper dispersion and the type of support on which copper is deposited, i.e., the nature of interactions between copper oxide and support [37]. In this study, all of the Cu-based catalysts had a comparable average particle size of CuO in the range of 22\u201325\u2009nm calculated by Scherrer\u2019s formula (Table S1). Thus, the variance in Cu 2p3/2 binding energy in this study was mainly caused by the varied chemical natures of the supports, as evidenced by the H2-TPR results (Fig. 2(b)). Obviously, the interaction of CuO with support varied depending on the type of support, with Mg(OH)2 having the strongest interaction with CuO, whereas Al(OH)3 showed the weakest interaction with CuO (Fig. 2(b)). This explains the huge difference in Cu 2p3/2 binding energy between Cu/Mg(OH)2 and Cu/Al(OH)3.It is worth noting that the correlations between the molar ratio of R-COOH/R-OH and the valence state of copper, indicated by the Cu+ percentage of the Cu+ and Cu2+, are well established, as shown in Fig. 4(c). The ratio of R-COOH/R-OH increased with the amount of Cu+, demonstrating an approximately linear relationship. This finding suggests that Cu2+ species are beneficial to the formation of R-OH, whereas Cu+ species are desirable to form R-COOH. Among these Cu-based catalysts, Cu/Al(OH)3 possessed the largest quantity of Cu2+. This explains why, in the case of plasma only, the distribution of oxygenates shifted significantly from a mixture of R-OH and R-COOH, towards almost exclusively R-OH when packing Cu/Al(OH)3 into the plasma zone, as shown in Fig. 1(a). By contrast, Cu/HZSM-5 with the highest Cu+ content demonstrated strong activity toward R-COOH formation. This finding suggests that the valence state of copper is one of the crucial factors to improve the distribution of oxygenated products in plasma-catalytic conversion of CO2 and CH4.The best catalysts for the formation of alcohols and acetic acid in this study were Cu/Al(OH)3 and Cu/HZSM-5, respectively. Thus, these two catalysts were analyzed by XPS after the reaction, as shown in Fig. S4. The Cu2+ content of the Cu/Al(OH)3 catalyst decreased after the reaction, but a visible peak of Cu2+ satellite peak emerged on Cu/HZSM-5, indicating a slight increase in Cu2+ content. Furthermore, there is no peak shift in Cu 2p2/3 spectra on Cu/HZSM-5 before and after the reaction.Interestingly, pure HZSM-5 support was found to be the most beneficial catalyst for the formation of acetic acid (Fig. 1(a)), but its activity toward acetic acid generation decreased after anchoring Cu, although Cu/HZSM-5 had the highest Cu+ amount. To be explicit, the Cu/HZSM-5 catalyst was also prepared using the ion exchange method (IE), and its catalytic performances are compared with those of Cu/HZSM-5 (IM) in \nFig. 5.Apparently, the method for the preparation of Cu/HZSM-5 had a significant influence on the selectivity of R-OH and R-COOH. The Cu/HZSM-5 (IE) catalyst inhibited the formation of R-OH, but greatly enhanced the generation of R-COOH, resulting in higher total selectivity of oxygenates. Based on the XRF results (Table S2), the copper efficiencies of the Cu/HZSM-5 catalysts prepared by IE and IM methods, expressed by the mole numbers of the produced oxygenate per mole copper, are given in Fig. 5. The Cu/HZSM-5 (IE) clearly showed a much higher copper efficiency than the Cu/HZSM-5 (IM), especially in terms of R-COOH production. In addition to higher R-COOH selectivity, the conversion of CO2 was also enhanced, as shown in Fig. 5(b), and the gap between the conversion of CO2 and CH4 was narrowed by using the Cu/HZSM-5 (IM) catalyst. This finding in turn supports the increased selectivity of R-COOH over Cu/HZSM-5 (IE), since the stoichiometric ratio of CH3COOH formation from CO2 and CH4 molecules is 1:1, which means that the selectivity of CH3COOH approaches the theoretically maximum value only when the conversion of CO2 is equal to that of CH4.To understand the substantial differences in the catalytic activity of the Cu/HZSM-5 catalysts prepared by IE and IM methods, their chemical properties were comparatively investigated. As shown in \nFig. 6(a), different from the Cu/HZSM-5 (IM) catalyst, no peak associated with copper species was observed in the XRD pattern of the Cu/HZSM-5 (IE) catalyst, indicating that copper species were highly dispersed on Cu/HZSM-5 (IE). This finding can be further confirmed by EDS element mapping images of Cu/HZSM-5(IE) in \nFig. 7. Clearly, the Cu mapping image showed a homogeneous distribution of copper species, and the location of the Cu element appearing is almost always accompanied by the Al element but not for the Si element. This exhibits the feature of ion-exchange zeolite, i.e., Cu ion exchanges with H sites located at the Si-(OH)-Al unit of HZSM-5. However, where Al appeared, in turn, Cu was not always there by comparing the Al mapping with the Cu mapping in Fig. 7. These findings suggest that H sites were partially substituted by Cu ions, which is consistent with the low loading of Cu obtained on the Cu/HZSM-5\u00a0(IE) catalyst (Table S2) and the high dispersion of Cu ions on HZSM-5. These results reveal that copper species anchored on HZSM-5 mainly exist in the form of isolated copper species when using the ion-exchange method for catalyst preparation, but in the form of crystalline CuO when using the impregnation method (Figs. 2 and 3).XPS was then used to identify the electron structure of this type of isolated copper species, as shown in Fig. 6(b). The isolated copper species mainly exist in the form of Cu+ on Cu/HZSM-5 (IE) catalyst, according to the lack of the characteristic satellite peak of Cu2+ in Cu 2p3/2 XPS spectra. That is, the ion-exchange method is more favorable for Cu+ formation than the IM method. A similar result was also reported by Wu et al. [39]. Combined with the performance of Cu/HZSM-5 (IE) in Fig. 5, the high amount of Cu+ on Cu/HZSM-5 (IE) is one reason for the high activity toward CH3COOH formation, but low activity to C1-C4OH generation, which is consistent with the conclusion reached above, highlighting again that Cu+ enhances the formation of R-COOH.Furthermore, the reducibility of the Cu/HZSM-5 catalysts using different preparation methods was comparatively investigated by H2-TPR analysis, as shown in Fig. 6(c). Cu/HZSM-5 (IE) displayed two small H2-consumption peaks due to its low Cu amount (Table S2), i.e., the low-temperature and the higher-temperature peaks in the range of approximately 200\u2013600\u2009\u00b0C, which can be correspondingly assigned to the reduction of isolated Cu2+ to Cu+ and Cu+ to Cu0 according to the literature [34\u201336]. On the one hand, the H2-TPR results reveal that in addition to the main species of Cu+ confirmed by XPS in Fig. 6(b), the isolated Cu2+ species, exist in the Cu-exchange HZSM-5 (IE) catalyst, as evidenced by the typical two-step reduction of isolated Cu2+, i.e., Cu2+ to Cu+ and Cu+ to Cu0, whereas crystalline CuO has a one-step reduction mechanism from Cu2+ to Cu0, as described in Fig. 2(b). On the other hand, Cu+ species were further confirmed to be the main isolated copper species, not Cu2+, since the area of the higher-temperature peak was much higher than that of the low-temperature peak in the H2-TPR profile of Cu/HZSM-5 (IE). Otherwise, the area of the low-temperature peak should be approximately equal to that of the high-temperature peak, if Cu2+ is the unique isolated copper species. Although the H2-TPR results agree with the results of the XPS analysis in Fig. 6(b), the H2-TPR and XPS analyses gave different amounts of Cu2+. H2-TPR showed a certain amount of Cu2+ existed in Cu/HZSM-5 (IE), whereas only a very small amount of Cu2+ was obtained by XPS. This finding is mainly caused by the feature of the XPS technique that only copper species located at a depth lower than 10\u2009nm can be detected.In contrast, an overlapping peak was observed in the H2 consumption profile of the Cu/HZSM-5 (IM) catalyst (Fig. 6(c)), which can be ascribed to the reduction of crystalline CuO particles, as supported by the XRD results (Fig. 6(a)) and HRTEM analysis (Fig. 3). The overlapped peak consists of four peaks (denoted as peak A-D) that were fitted using Gaussian deconvolution. According to the literature [40], isolated Cu2+ species may also exist in addition to the main reduction peak of crystalline CuO. Peak A, located at a low temperature, represents the reduction of isolated Cu2+ to Cu+; Peak B, located at a moderate temperature, being a major one, is assigned to the one-step reduction of CuO particles to Cu metal; Peak C, appeared at a high temperature, represents the reduction of isolated Cu+ to Cu0; Peak D may be related to the interaction of crystalline CuO and HZSM-5. In fact, the existence of isolated copper species on HZSM-5 is reasonable when using the IM method since the IM process is inevitably accompanied by ion-exchange with regard to HZSM-5. Therefore, these characterization results highlight that crystalline CuO is the dominant copper species on Cu/HZSM-5 (IM). On the Cu/HZSM-5 (IE) catalyst, however, only isolated copper species exist, i.e., isolated Cu+ and Cu2+ with Cu+ ions being the most abundant.In addition, the acidity of the Cu/HZSM-5 catalysts prepared by the IE and IM methods was determined using NH3-TPD and Py-FTIR, as shown in Fig. 6(d) and Fig. S5, using the pure HZSM-5 zeolite as a reference. The NH3-TPD profile of pure HZSM-5 displays two desorption peaks in the 150\u2013300\u2009\u00b0C and 300\u2013600\u2009\u00b0C ranges, representing abundant weak acid sites and a minor quantity of strong acid sites on HZSM-5, respectively (Fig. 6(d)). In addition, the Py-FTIR spectra of pure HZSM-5 shows that both Br\u00f8nsted acid (H) sites and Lewis acid sites coexist in HZSM-5, with respective infrared bands at around 1540 and 1455\u2009cm\u22121 probed by pyridine [41] (Fig. S5). When compared to HZSM-5 in Fig. 6(d), the peak area of the high-temperature desorption peak was significantly increased after anchoring Cu on HZSM-5 using the IM method, while the peak area of the low-temperature desorption peak was slightly lowered. This finding suggests that Cu anchored on HZSM-5 via the IM method increased the amount of strong acid sites while decreasing the number of weak acid sites. However, the corresponding Py-FTIR spectrum of Cu/HZSM-5 (IM) presented a significant drop in both H sites and Lewis acid sites (Fig. S5), with the H sites almost completely disappearing. The NH3-TPD and Py-FTIR studies of Cu/HZSM-5 (IM) reveal that the increased part of strong acid sites is mostly due to the addition of crystalline CuO (Fig. 6(a)) as strong Lewis acidic species. By contrast, the NH3-TPD profile and Py-FTIR spectrum of Cu/HZSM-5 (IE) were totally different, as shown in Fig. 6(d) and Fig. S5. The Cu-exchange HZSM-5 catalyst results in a decrease in both the acid amount and acid strength, as well as H sites of HZSM-5 due to the exchange of Cu ion with H sites in the Si-(OH)-Al unit of HZSM-5. It should be noted that a fraction of the H sites in HZSM-5 remained in the Cu/HZSM-5 (IE) catalyst, almost all of the H sites, however, disappeared in the Cu/HZSM-5 (IM) catalyst.Combing with the corresponding performance of pure HZSM-5, Cu/HZSM-5 (IM) and Cu/HZSM-5 (IE) as shown in Fig. 1 and Fig. 5, pure HZSM-5 possessed the highest amount of Br\u00f8nsted acid sites and exhibited greatest activity toward R-COOH formation. Compared to pure HZSM-5, the H sites in Cu/HZSM-5 (IM) disappeared, and the associated selectivity of R-COOH decreased as well, despite the fact that Cu+ in Cu/HZSM-5 (IM) had a beneficial effect on R-COOH formation (Fig. 4(c)). Furthermore, Cu/HZSM-5 (IE) with a portion of H sites displayed higher activity toward R-COOH than Cu/HZSM-5 (IM) without H sites. Interestingly, although Cu/HZSM-5 (IE) had fewer H sites than pure HZSM-5, Cu/HZSM-5 (IE) showed comparable R-COOH selectivity, which can be attributed to the synergy of Cu+-H site as reported in the literature [42]. These findings indicate that Br\u00f8nsted acid sites may be another important factor for tuning R-COOH selectivity in plasma-catalytic conversion of CO2 and CH4 in addition to the copper valence state. Song et al. found that a higher ratio of weak acid sites is beneficial for the generation of oxygenates, such as formaldehyde, methanol, ethanol, and acetone, in the plasma-catalytic conversion of CO2 and CH4, but the relationship between weak acid sites and the distribution of oxygenates was not demonstrated [28]. In this study, the roles of Br\u00f8nsted acid sites in enhancing the formation of R-COOH will be discussed as follows.Different from catalytic reactions, CO2 and CH4 molecules in the plasma-catalytic system are pre-activated by energetic electrons into active species in Fig. S6, such as excited CO2*\u2009, CH4*\u2009and CO*\u2009, CH3, CH2, CH, H and O radicals [30,43]. According to the simulation results reported by De Bie et al. [44], CH3 radicals are much more abundant than CH species in the CH4-CO2 DBD. In addition, OH could be produced via three main reactions, i.e., CH +\u2009O \u2192 C +\u2009OH, H2 +\u2009O \u2192 OH +\u2009H and CO2(v) +\u2009H \u2192 CO +\u2009OH, which has been confirmed in detail in our previous studies [30].In this study, the proportion of CH3OH in the R-OH products exceeded 60%, followed by ethanol. That is, the enhanced formation of Cu2+ species in the production of R-OH can also be considered to promote the formation of CH3OH. Generally, Cu2+ species are accepted as the active sites in the oxidation of CH4 to CH3OH using O2 or N2O as an oxidant [45,46]. Herein, the active O species (777.5\u2009nm, 3s5S0 \u2192 3p5P; and 844.7\u2009nm, 3s3S0 \u2192 3p3P) [47], presented in Fig. S6, are highly efficient for oxidation. Therefore, it is reasonable to hypothesize that Cu2+ species act as the active sites for CH4 oxidation to CH3OH with O species produced from CO2 splitting in this study. As shown in \nFig. 8, possible pathways are proposed for plasma-catalytic conversion of CO2 and CH4 to CH3OH.Methanol can be produced through the CH4 oxidation reaction. That is, Cu-CH3, as the important intermediate species, could be formed by CH4*\u2009dissociation or CH3 adsorption on copper sites. Subsequently, the resulting Cu-CH3 species react directly with gaseous OH to produce CH3OH in pathway \u2460. Alternatively, as shown in pathway \u2461, O radicals could insert into the Cu-C bond of Cu-CH3 to form Cu-OCH3 (methoxy) according to the literature [42] and followed by methoxy protonation with H radicals or nearby H adsorbed (Had) on the catalyst for the final synthesis of CH3OH. In addition, the recombination of gaseous CH3 and OH radicals can result in the formation of CH3OH [30].In addition, another way to generate methanol in this study is the CO2 hydrogenation reaction, i.e., the CO2 to CH3OH pathway, and the H atoms derive from CH4 splitting in plasma, which is supported by the existence of H radials (656.3\u2009nm, 3d2D \u2192 2p2P0) in the emission spectra of the CO2/CH4 plasma (Fig. S6) [48,49]. As shown in Fig. 8, CO\u2009(Angstrom bands at 451\u2009\u2212608\u2009nm), formed in the plasma gas-phase reactions (Fig. S6), can be directly adsorbed onto the copper center, followed by stepwise hydrogenation with H radicals or Had to form CH3OH with HCOad, H2COad\n, and H3COad being the adsorbed intermediates in pathway \u2462. On the other hand, gaseous CO2 *\u2009could be adsorbed onto a copper center to form Cu-CO2, and subsequently, the Cu-CO2 is hydrogenated with H radicals or Had step by step via formate pathways, resulting in the formation of H3COad intermediate in the pathway \u2463. Finally, H3COad reacts with H radicals or Had to produce CH3OH. In our previous study, we described in detail the above surface-reaction pathways that lead to the formation of CH3OH during plasma-catalyzed CO2 hydrogenation [50].\n\nFig. 9 shows four main pathways proposed for plasma-catalytic conversion of CO2 and CH4 to CH3COOH. Pathway \u2460, Br\u00f8nsted acid sites can enable CH3COOH generation directly through the CO2 protonation pathway even without Cu center in the catalyst, i.e., gaseous CO2*\u2009react with H sites in the Si-(OH)-Al unit of HZSM-5 to yield -COOH species [51], followed by direct C-C coupling with gaseous CH3 radicals to produce CH3COOH via the E-R process. In this way, the formation of CH3COOH is determined by the amount of Br\u00f8nsted acid sites, and this pathway is well supported by the experimental results, i.e., pure HZSM-5, which possessed the highest amount of Br\u00f8nsted acid sites, exhibited higher activity toward CH3COOH in the absence of a Cu center. Pathway \u2461, the resulting -COOH species can also couple with neighboring Cu-CH3 species, which is produced by CH4 dissociation on a Cu center [51\u201356] to form CH3COOH in the presence of a copper center. Pathway \u2462, gaseous CO2*\u2009is directly inserted into the bond of Cu-CH3 to form Cu-OOCCH3 (acetate species) [52\u201354] and subsequently, CH3COOH is formed via proton transfer from the adjacent Br\u00f8nsted acid site to the Cu-OOCCH3 species. Alternatively, in Pathway \u2463, gaseous H radicals react directly with Cu-OOCCH3 to form CH3COOH, which accounts for CH3COOH generation on Cu-based catalysts lacking or having fewer Br\u00f8nsted acid sites, such as Cu/SiO2, Cu/TiO2 and Cu/HZSM-5 (IM) in this study. In addition, surface -OH Br\u00f8nsted acid sites can be generated by CH4 splitting on a metal center (M) into M-CH3 and M-H, with the H from M-H eventually bonding to the adjacent O site of zeolite or the surface O of oxides to form -OH [52\u201354], which maintains the catalytic cycle of Br\u00f8nsted acid sites involved in Fig. 9.It is worth noting that both pathways \u2461 and \u2462 reveal that the formation of CH3COOH results from a synergy of metal and Br\u00f8nsted acid sites, which necessitates the development of a bi-functional catalyst. In this study, Cu-exchanged HZSM-5 serves as a bifunctional catalyst due to the co-existence of isolated Cu+ and Br\u00f8nsted acid sites, enabling CH3COOH production through pathways \u2461 and \u2462, as well as pathway \u2460 due to accessible Br\u00f8nsted acid sites. This explains why Cu-exchanged HZSM-5 showed fewer Br\u00f8nsted acid sites than pure HZSM-5 while still exhibiting strong CH3COOH selectivity. The above findings suggest that acid sites are another important factor in tuning the selectivity of acetic acid, and the metal-Br\u00f8nsted acid site synergy provides a potential strategy to enhance the formation of R-COOH. However, Br\u00f8nsted acid sites are not required for R-OH formation.Tosi et al. found that direct coupling of gaseous CH3 with COOH species results in CH3COOH production in addition to the formation of CH3COOH on the surface of the catalysts, and the COOH species were produced in the gas phase via CO2 protonation with gaseous H radicals [19]. Our previous study also confirms the formation pathway of CH3COOH by recombination of gaseous CH3 radicals with COOH species [30].Direct conversion of CO2 and CH4 to high-value oxygenates such as alcohols (R-OH), acids (R-COOH), acetaldehyde and acetone with R-OH and R-COOH being the major oxygenates, was successfully achieved over the Cu-based catalysts driven by plasma at 60\u2009\u00b0C and atmospheric pressure. For the first time, the correlations of oxygenate distribution versus copper valence state were experimentally confirmed. Cu2+ species exhibit superior activity towards the formation of R-OH products. Cu+ species, however, are critical for the generation of R-COOH. In addition to Cu+ species, Br\u00f8nsted acid (H) sites on the Cu-exchange HZSM-5 catalyst also promote the synthesis of R-COOH via the CO2 protonation route, as well as the synergy of isolated Cu+ and H sites. These findings, i.e., the correlation between valence state/Br\u00f8nsted acid and the catalyst activity, provide valuable insights into improving existing catalyst performance and designing cutting-edge highly selective catalysts for the oriented generation of oxygenates from plasma-catalytic CO2 and CH4 conversion.\nYuezhao Wang: Conceptualization, Validation, Formal analysis, Resources, Data curation, Writing \u2013 original draft, Writing \u2013 review & editing. Linhui Fan: Conceptualization, Validation, Formal analysis, Resources, Data curation, Writing \u2013 original draft. Hongli Xu: Validation, Formal analysis, Resources, Data curation. Xiaomin Du: Validation, Formal analysis, Resources. Haicheng Xiao: Supervision, Funding acquisition. Ji Qian: Resources, Data curation. Yimin Zhu: Resources, Supervision. Xin Tu: Conceptualization, Formal analysis, Writing \u2013 review & editing, Supervision, Funding acquisition. Li Wang: Conceptualization, Validation, Formal analysis, Resources, Data curation, Writing \u2013 original draft, Writing \u2013 review & editing, Supervision, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.We acknowledge financial support from the National Natural Science Foundation of China (No. 21908016), the PetroChina Innovation Foundation (No. 2019D-5007-0407) and the LiaoNing Revitalization Talents Program (No. XLYC1907008). X. Tu acknowledges the funding from the European Union\u2019s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No. 813393.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.apcatb.2022.121583.\n\n\n\nSupplementary material\n\n\n\n.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  Direct conversion of CO2 and CH4 into value-added oxygenates under mild conditions is highly desirable since it has great potential to deliver a sustainable low-carbon economy and a carbon-neutral ecosystem. However, tuning the distribution of oxygenates in this process remains a major challenge. Here, the electronic structure and acidic properties of copper-based catalysts were exploited as strategies to tune the distribution of oxygenates (alcohols and acids) in the plasma-catalytic conversion of CO2 and CH4 at a reaction temperature of 60\u00a0\u00b0C and atmospheric pressure. We use support, on which copper is anchored, to regulate the distribution of Cu2+ and Cu+ in the Cu-based catalysts. Comprehensive characterization of the catalysts together with the reaction performances reveals that Cu2+ species are favorable to the formation of alcohols, whereas Cu+ species are critical to enhancing acetic acid production. Furthermore, the Br\u00f8nsted acid sites of HZSM-5 significantly improved the selectivity of acetic acid, while the synergy of isolated Cu+ center and Br\u00f8nsted acid sites, developed via Cu-exchange HZSM-5, exhibits potential for acetic acid formation. Finally, possible pathways for the formation of alcohols and acetic acid have been discussed. This work provides new insights into the design of highly selective catalysts for tuning the distribution of alcohols and acids in the plasma-catalytic conversion of CO2 and CH4 to oxygenates.\n               "}
{"full_text": "No data was used for the research described in the article.Over the years, the world has experienced climatic change due to increasing atmospheric levels of greenhouse gases (GHGs), which are responsible for global warming [1]. CO2 has been identified as the dominant contributor to global warming, accounting for over 65\u00a0% of the total annual GHG emissions [2]. The global CO2 emissions from combustion of fossil fuels for energy production and industrial processes have been increasing tremendously [3]. Although in 2020, there was an observed slight decline in CO2 emissions of about 8\u00a0% compared to 2019. This reduction was attributed to a sharp temporary cutback in global atmospheric CO2 emissions due to the COVID-19 pandemic and the corresponding restrictions resulting into limitations on the use of petroleum products. But this decline in CO2 emissions did not result into a significant reduction in the CO2 atmospheric levels due to the prevailing residual amount of this gas and emissions from other sectors. Moreover, recent figures show that a 6\u00a0% increase in CO2 emissions was observed in 2021 compared to 2020, which was partly attributed to the high energy demand during economy recoveries after easing the COVID-19 restrictions [3\u20135].Regulations and strategies to control the increasing atmospheric CO2 levels have been proposed by governments and the private sector [6]. Among the strategies designed to curb the CO2 problem, are the technologies for carbon capture and storage (CCS). Under CCS, CO2 is trapped at point sources such as industries and thermal power plants, and then stored in geological formations or under ocean beds. However, the volatile nature of CO2 and the high energy costs involved in these processes make them uneconomical. In this respect, effort has been directed towards the development of efficient carbon capture and utilization (CCU) technologies. The CCU approach, transforms the captured CO2 into value added chemical products such as fuels, polymers, among others [7,8]. Considerable effort has been made to design catalytic processes for CO2 conversion because it is an inexpensive, readily available, non-toxic, and renewable carbon resource [9,10]. However, due to the thermodynamic stability of CO2, it is usually challenging to activate in typical reactions and its utilization as a raw material on a large scale is still lacking. Nonetheless, some industrial processes utilizing CO2 as a chemical feedstock have been reported including synthesis of urea, salicylic acid, inorganic carbonates, organic carbonates and polymers [11,12].Recently, CO2 utilization as a feedstock in industrial polymer production has attracted a lot of attention, due to the long term storage potential of this greenhouse gas in the copolymers [13]. Polymers that contain CO2 groups are commonly known as polycarbonates and the aliphatic ones like poly(propylene)carbonate)PPC can undergo biodegradation at the carbonate linkages producing water and CO2 as byproducts [13]. This is in sharp contrast to the common non-biodegradable petroleum derived polymers such as polyethylene (PE), polypropylene (PP), polyvinyl chloride (PVC), and polystyrene (PS) [14]. Currently, most polycarbonates are produced via an interaction between bisphenol-A and diphenyl carbonate or through a reaction of the sodium salt of bisphenol-A and phosgene to form poly (bisphenol-A carbonate) (BA-PC). However, this approach produces large amounts of phenolic by-products [15] not to mention the hazardous nature of phosgene [16]. An improved approach in the synthesis of BA-PC from Bisphenol A was reported by Asahi Kasei, where the phenol generated is recycled and the CO2 produced utilized as one of the starting materials, eliminating the need for use of phosgene [17,18].A sustainable approach to the preparation of polycarbonates was first demonstrated by Inoue and co-workers [19]. They reported an alternating copolymerization of propylene oxide (PO) with CO2 which resulted into an aliphatic poly(propylene)carbonate (PPC) as shown in Scheme 1\n\n. Since its discovery, there has been growing interest in large scale production of biodegradable CO2-based polycarbonates. The CO2-based aliphatic polycarbonate plastics have excellent properties such as biodegradability, high transparency, UV stability, thermoplasticity and high young\u2019s modulus [20]. This makes CO2-based polymers useful raw materials in the manufacture of barrier materials, plasticizers, plastic film products, medical plastics, reactive hot melt adhesives, binders, and many other products [13,21].In comparison to the current industrial processes for polycarbonate synthesis, the CO2 and epoxide copolymerization approach is advantageous because it lowers the amount of fossil fuel-based monomer raw materials incorporated in the polymer. When CO2 is used as the monomer, the production of PPC saves about 43\u00a0% by mass of petrochemical resources in comparison to poly(propylene oxide) (PPO) prepared by PO homopolymerization. Moreover, PPC formation usually follows chain growth mechanisms, which theoretically imply that the copolymerization of CO2 and PO has better control of molecular weight and selectivity, unlike BA-PC which follows step-growth mechanisms [22]. Moreover, the epoxide monomers involved in the process are not as toxic as phosgene that has been largely used in polycarbonate production processes [16].Nevertheless, the CO2-epoxide copolymerization process is still limited in its industrial application for polycarbonate synthesis, primarily due to CO2 thermodynamic stability. CO2 being in the most highly oxidized state of carbon, has a more negative standard Gibbs free energy (\u0394G\u00a0=\u00a0\u2212394.37 KJmol\u22121) in comparison to other C1 molecules. Thus, CO2 activation and utilization in chemical production requires a lot of energy [23]. The high energy barrier in CO2 activation can be overcome by either direct input of energy or via reaction of CO2 with chemically reactive species [24]. Due to the temperature-sensitive nature of the products of the CO2 based copolymerization, use of chemically reactive species in the presence of suitable catalysts is the preferred option [25,26]. Nonetheless, the advantages of copolymerization of CO2 with other monomers to form polycarbonates still outweigh the thermodynamic challenges involved in the process.Several copolymers have been synthesized from the reaction of CO2 and energy-rich molecules such as amines, epoxides, aziridines, episulfides and oxetanes [24,27]. However, the most widely studied copolymerization reaction is that of CO2 with PO to form PPC [28,29]. It should be noted that in the CO2 and PO co-polymerization process, the carbon dioxide molecule does not undergo reduction. The inert nature of CO2 is overcome by the energy (\u2248114 kjmol\u22121) released during the PO ring opening process [30]. The earlier work of the CO2/PO copolymerization successfully produced an alternating aliphatic copolymer of PPC with a high molecular weight [19]. Ever since, the copolymerization reaction triggered interest of several researchers to not only increase yields of the synthesized PPC but also to improve the selectivity towards CO2 insertion into the polymer chain. The PPC with high CO2 content possesses unique properties such as biodegradability and excellent oxygen barrier performance that has made it a potential raw material in substances used in food packaging, healthcare and automotive industries [13]. However, due to the amorphous nature of PPC and relatively low glass-transition temperature (Tg), its application in comparison to aromatic polymers like BPA-PC is still limited [21]. In this respect, more effort is still being sought to design efficient catalyst systems for the copolymerization of CO2 and epoxides to produce polycarbonates of desirable chemical and mechanical properties.In a typical CO2 and PO copolymerization reaction, there are two common side reactions that can occur. The first is the cyclization or back-biting reaction, and the second is the consecutive insertion of the epoxide monomer into the growing polymer chain. The chain back-biting reaction leads to the depolymerization of the polycarbonate to form the thermodynamically stable cyclic carbonate [31]. Consecutive insertion of the epoxide leads to the formation of ether bonds in the copolymer, which lowers the Tg of the polymer and compromises its material properties at room temperature [32].In the presence of a suitable catalyst, the CO2 and PO copolymerization reaction follows a coordination-insertion mechanism as illustrated in Scheme 2\n. The metal center of the catalyst is coordinated to the epoxide, followed by a nucleophilic attack which results in ring opening to form a metal alkoxide. PO being an aliphatic epoxide, its ring-opening is usually favored at the least hindered CO bond, and the CO2 molecule is then inserted into the metal\u2013oxygen bond resulting in the formation of a metal carbonate. Depending on the selectivity of the catalyst and reaction conditions, different pathways can be undertaken after this step. The metal carbonate can undergo cyclization or back-biting (path 1) to form the propylene carbonate by-product, regenerating the metal alkoxide which can propagate the reaction further [33]. On the other hand, propagation by multiple coordination and insertion of CO2/PO producing the polycarbonate chain (path 2) can also occur. This would result in the formation of a copolymer with alternating CO2/PO linkages. However, homopolymerization of epoxide can also occur and in this case, metal alkoxides successively attack the epoxide monomer instead of CO2 insertion (path 3) forming ether linkages in the copolymer [25].Moreover, chain transfer reactions can also take place when the copolymerization is conducted in the presence of protic compounds such as water (in the epoxide monomer), forming hydroxyl-terminated copolymers and metal alkoxides or hydroxide species [34]. The molecular weight (MW) of the formed copolymers is believed to be reliant on the amount of polymer formed in a living polymerization reaction. In this case, the degree of polymerization is dependent only on catalyst concentration which eventually results in a linear increase of polymer MW with monomer conversion [35]. However, experimental results generally show lower values of the MW for the polycarbonate than expected implying that MW is dependent on both chain transfer agents and catalyst concentrations [35,36].In PO and CO2 copolymerization, although PO ring-opening is favored at the methylene carbon\u2013oxygen (\u2013CH2-O-) bond due to lower steric hindrance, cleavage can also occur at the methine carbon\u2013oxygen (\u2013CH-O-) bond, giving rise to differences in the regioselectivity of the resultant polymers [37]. Spectroscopic studies of the copolymers show that the 13C NMR spectrum gives three signals in range of 154.0\u2013155.2\u00a0ppm typical of carbonyl region for polycarbonates. These signals are usually assigned to head-to-head (HH), head-to-tail (H-T), and tail-to-tail (T-T) linkages in the ratio of 1:2:1 respectively as illustrated in Scheme 3\n. The H-T linkages are formed by successive ring opening at the same carbon center, while the HH and T-T linkages are formed by sequential ring opening at the methine CO bond followed by ring opening at the methylene CO bond. Thus, the H-T linkages are the dominant linkages in the polycarbonate copolymers [37]. Additionally, different stereochemistries of the polycarbonates are also possible for the H-T linkages giving either isotactic (all the substituents have the same stereochemistry), or syndiotactic (alternating stereochemistries), or atactic (all substituents have random stereochemistry) polymer domains [13,25]. The catalyst systems utilized in the copolymerization reaction, exhibit differences in the regioselectivity of the formed polymer. For example, the catalyst developed by Quan et al., 2003, produced H-T linkages in higher concentrations of 70\u201377\u00a0% giving high Tg values of 37\u201342\u00a0\u00b0C, respectively [38]. Interestingly, some studies have shown that the regioregularity of polycarbonates is a key factor that influences material properties of the polymers [25].Ever since the successful copolymerization of CO2 and PO using an equimolar Et2Zn/H2O catalyst system [19], many catalyst systems of either heterogeneous or homogeneous nature [39], have been developed for the ring-opening copolymerization process (ROCOP). Several studies have investigated performance of homogeneous systems, owing to their high activities and selectivity for the CO2/PO copolymerization [40]. However, the use of these catalysts for chemical synthesis on an industrial scale is limited due to their complex synthesis, difficult separation of the catalysts from the product, and lack of recyclability [41]. In contrast, the ease of preparation, stability and recyclability have made heterogeneous catalyst systems attractive for large scale CO2/PO copolymerization. But the heterogeneous catalysts that have catalyzed this process with high polycarbonate selectivity, have often suffered low polymer yields while those with high activity usually exhibit low selectivity for PPC [25,26]. In this respect, significant research effort is still devoted to development of efficient heterogeneous catalysts for CO2 and PO copolymerization.In 1969, Inoue et al. utilized a ZnEt2/H2O (1:1\u00a0mol ratio) catalyst system for CO2 and PO copolymerization to form PPC. The reaction was carried out at 80\u00a0\u00b0C and 20\u201350\u00a0atm CO2 pressure for a period of 48\u00a0h, giving PPC with 88\u00a0% carbon content but in a low yield of only 4.2\u00a0g polymer/g catalyst [19]. When utilizing the ZnEt2/H2O catalyst system, the C2H5(ZnO)nH species produced from the hydrolysis of ZnEt2 was considered to be the active species responsible for the formation of the polycarbonate [42]. The DFT study done on the ZnEt2/H2O catalyst system by Pan et al.,2013 proposed a monometallic mechanism for the CO2/PO reaction. The low catalytic activity observed by this system was attributed to its monometallic mechanism unlike its homogeneous counterparts that usually follow a bimetallic mechanism [43]. Despite the poor activity of ZnEt2 /H2O catalyst system, it stimulated further research into similar systems utilizing different ZnEt2/active hydrogen-containing species. Studies have showed that compounds having two or more labile hydrogen atoms such as amines, alcohols or carboxylic acids coupled with ZnEt2 generated effective catalysts for the copolymerization of CO2 and PO [35,44]. On the other hand, ZnEt2/monoprotic sources of similar compounds only produced propylene carbonate [45].Kobayashi et al. studied the CO2 and PO copolymerization using ZnEt2 and diprotic phenolic compounds as catalysts in dioxane solvent at 35\u00a0\u00b0C and 30\u00a0atm [46]. Under optimal conditions, the best system was ZnEt2/resorcinol (1:1\u00a0mol ratio) which gave a higher copolymer yield of 3.5 g polymer/ g catalyst compared to the yield of 1.03\u00a0g/g catalyst obtained when using the Et2Zn/H2O (1:1) system under similar reaction conditions [46]. Using a similar approach, Kuran et al. investigated the CO2 and PO copolymerization catalyzed by ZnEt2 and di- or tri- protic compounds in 1,4-dioxane solvent at 35\u00a0\u00b0C and 60\u00a0atm of CO2. The ZnEt2/resorcinol catalyst was also among the most active systems giving a copolymer yield of 28\u00a0% with respect to PO used [47]. When utilizing Et2Zn/carboxylic acid systems, aromatic carboxylic acids displayed higher activities in the copolymerization reaction compared to their aliphatic counterparts [48]. The Et2Zn/primary amine (1:1) systems displayed comparable activities as the Et2Zn/H2O (1:1) systems while the Et2Zn/secondary amine systems hardly showed any activity in the copolymerization reaction [42]. Despite the interest and effort focused on the dialkylzinc based catalysts, these systems generally exhibited low turnover frequencies (TOFs) and give copolymers with very broad molecular weight distributions. These shortfalls could be partially attributed to the lack of optimal access to metal centers by monomers resulting into production of a wide range of polymers.The next major advancement in the CO2 and PO copolymerization reaction was reported by Soga et al. where the performance of metal salts of acetic acid [49], and air-stable salts of zinc hydroxide/dicarboxylic acid derivatives were investigated for this process [50]. Under solvent free conditions, at 80\u00a0\u00b0C and after a duration of 43\u00a0h the zinc acetate salt gave a copolymer yield of 2.3\u00a0g polymer/g catalyst with 100\u00a0% carbonate linkages (Fc), and weight average molecular weight (Mw) of 20,000\u00a0g/mol. Whereas the yield when using cobalt acetate was 0.89\u00a0g polymer/g catalyst (Fc\u00a0=\u00a0100\u00a0%, Mw\u00a0=\u00a025,000\u00a0g/mol) after 70\u00a0h [49]. The copolymerization reactions performed using Zn(OH)2/dicarboxylic acids as catalysts showed that aliphatic dicarboxylic acids had superior activity compared to their aromatic counter parts [50]. The Zn(OH)2/glutaric acid (Zn(OH)2/GA) catalyst system turned out to be most active, giving a copolymer yield of 69.1\u00a0g polymer/g Zn, with a number average molecular weight (Mn) of 12,000\u00a0g/mol. The zinc glutarate (ZnGA) catalyst formed from Zn(OH)2/GA catalyst system gave an activity higher than that of the ZnEt2/H2O catalyst (23.8\u00a0g copolymer/g Zn) when tested under the same experimental conditions [50]. The ease of synthesis, non-toxicity, economic viability and high activity of ZnGA, have made it an attractive class of catalysts for industrial production of polycarbonates [32]. Extensive studies have been undertaken to improve the structure and surface properties of ZnGA using numerous zinc sources, combined with different synthetic approaches.In 1995, Darensbourg et al. synthesized ZnGA catalysts by addition of glutaric acid to zinc oxide (ZnO) in toluene which upon heating gave a white crystalline solid [51]. In their study, supercritical CO2 was found to be a suitable substitute for ordinary organic solvents in the CO2/PO copolymerization reaction. Under optimal conditions, a polymer yield of 15.9\u00a0g polymer/g Zn (Fc\u00a0=\u00a091\u00a0%, Mn\u00a0=\u00a026,783\u00a0g/mol) was attained when methylene chloride was used as a co-solvent [51]. Later, Ree et al. [52] studied the effects of different zinc compounds and glutaric acid derivatives on the synthesis of ZnGA, and performance of the subsequent catalysts. When ZnO precursor was used in ZnGA catalyst synthesis, it gave the best catalytic activity compared to other zinc sources. Under optimized reaction conditions this ZnGA catalyst, produced a high molecular weight copolymer (Mw\u00a0=\u00a0343,000\u00a0g/mol, Mn\u00a0=\u00a0143,000\u00a0g/mol, Mw/Mn (PDI)\u00a0=\u00a02.4), with a yield of 64.0\u00a0g polymer/g catalyst [52]. Although, the resultant PPCs were also amorphous, and exhibited a Tg of 38\u00a0\u00b0C and decomposition temperature (Td) of 252\u00a0\u00b0C under a nitrogen atmosphere [52].Studies also showed that the crystallinity and morphology of the ZnGA affected its activity. In 2002, Wang et al. successfully synthesized ZnGA catalysts using ultrasonic stirring. These catalysts exhibited high crystallinity and small particle size. They gave a higher catalytic performance compared to the ZnGA prepared using mechanical stirring methods. In addition, the PPC polymer obtained exhibited slightly high Tg (39.4\u00a0\u00b0C) and Td (278\u00a0\u00b0C) values compared to those reported in literature [53]. The ultrasonic stirring route was also used by Meng et al.[54], who reported high crystalline ZnGA catalysts with particle sizes in the range 0.2\u20130.3\u00a0\u03bcm, that produced PPC with a high molecular weight and in a good yield of 160.4\u00a0g polymer/g catalyst. In 2004, Eberhardt et al. successfully synthesized solid zinc glutarate catalysts with controlled amounts of Zn-ethyl sulfinate initiating groups [55]. When zinc ethyl sulfinate groups were incorporated in diethylzinc based carboxylates, the ZnGA catalysts that were formed showed enhanced catalytic activity up to a factor of 16 in the CO2/PO copolymerization reaction compared to those obtained with ZnO-based glutarates [55]. However, this catalyst modification had limited potential for industrial application because of its procedural complexity and need for use of expensive precursors.Investigations into the catalytic mechanism for the ZnGA system showed that there was adsorption of CO2 and PO onto ZnGA, but PO tended to be more easily adsorbed and inserted into the zinc oxide bond of the catalyst as compared to CO2. Implying that the catalyst surface played a key role in the process and could be modified by PO adsorption. This further implied that the reactivity of ZnGA in the copolymerization process is initiated by PO rather than CO2\n[56]. Theoretical and experimental studies led to the proposal that a bimetallic catalytic pathway is followed for the CO2/PO copolymerization over the ZnGA system, involving successive insertions of CO2 and epoxide into Zn-alkoxide and Zn-carboxylate groups on the surface of the catalyst [57]. Further, the ideal separation between two adjacent Zn atoms which resulted in the optimal activation energy required for the copolymerization was suggested to be in the range 4.6\u20134.8\u00a0\u00c5, as shown in the crystal structure of the synthesized ZnGA (Fig. 1\n(a) [57]. In another study aimed at improving the frame work of ZnGA, a nanosized surface-etched ZnGA catalyst was prepared using a mild-HCl solution. This surface modified ZnGA exhibited increased productivity (\u223c83\u00a0%) in the copolymerization reaction with a turnover number (TON) of 132.1\u00a0g PPC/g catalyst superior to the standard-ZnGA [58]. Kim et al. 2004, also demonstrated that the catalytic activity of the ZnGA systems mainly originates from the outer surfaces of the Zn-dicarboxylates [59].Having found out that the activity of ZnGA is restricted to the outer surface of its particles, Padmanaban and Yoon also reported enhanced catalytic activity when ZnGA was treated with various metal chlorides to form surface-modified ZnGA-metal chloride catalysts [60]. The catalysts treated with iron (III) chloride (ZnGA-Fe) and zinc chloride (ZnGA-Zn) exhibited 25.6\u00a0% and 38.3\u00a0% increased performances respectively in comparison to the untreated ZnGA catalysts. Moreover, the surface-modified catalysts produced high-molecular-weight polymers [60].The catalytic activity of ZnGA in the copolymerization reaction was observed to depend mainly on the surface area of the catalyst, and in this respect, studies were also undertaken to utilize different supports to enhance the catalyst surface area. Catalyst supports that have been employed for this process include metal oxides (alumina, titanium oxide, magnesium oxide, zeolites), nonmetal oxides (silica, carbon), and polymers [61]. Zhu et al. 2002, reported the first PPC product in the copolymerization reaction catalyzed by supported ZnGA catalysts using a perfluorinated compound as the support [62]. The alternating PPC obtained under optimal reaction conditions was in very high yield (126\u00a0g polymer/g catalyst, Mw\u00a0=\u00a056,100\u00a0g/mol), with a high Tg (46.46\u00a0\u00b0C) and Td of 255.8\u00a0\u00b0C [62]. Meng et al. 2005, also reported a silica-supported ZnGA (SiO2/ZnGA) catalyst prepared by grinding ZnGA and SiO2 together in a planetary ball grinder under vacuum. The SiO2-ZnGA exhibited a high catalytic yield of 358.8\u00a0g polymer/g Zn [63].When ZnGA was dispersed onto the surface of acid-treated montmorillonite (MMT) in quinolone it gave a ZnGA\u2013MMT catalyst with smaller crystal sizes. Under optimal conditions, ZnGA\u2013MMT gave PPC with a high molecular weight and in a good yield (115.2\u00a0g polymer/ g ZnGA). The resultant PPC exhibited slightly high Tg (38\u00a0\u00b0C) and a Td\u00a0>\u00a0250\u00a0\u00b0C which was attributed to the presence of residual MMT in the copolymer [64]. The MCM-41 supported ZnGA catalyst when utilized for CO2 and PO copolymerization, exhibited enhanced catalytic activity (89.5\u00a0g polymer/g catalyst) producing PPC with a high molecular weight under optimal conditions [65]. Gao et al. 2015, also reported an increase in polymer yield while utilizing a silica supported ZnGA catalyst (from 194\u00a0g polymer/g Zn over unsupported ZnGA to 392\u00a0g polymer/g Zn over ZnGA/SiO2). Moreover, a high carbonate content (>97\u00a0%) with Mn of>10,000\u00a0g/mol for the alternating PPCs were observed in comparison to previous studies [66].Double metal cyanides (DMCs) are another type of catalysts that have been widely investigated in the CO2/PO copolymerization reaction. These catalysts, also known as Prussian blues are a class of molecular salts made up of crystalline metal cyanide frameworks [67,68]. Their structures have two different metal centers, where one metal coordinates via the carbon atom of the cyanide (CN\u2013) ligand and the other via the nitrogen atom. The general structural formula of DMCs is Tx[M(CN)\ny\n]z, where T metal ions include Zn(II), Co(II), Fe(II) or Ni(II), whereas for M, cations such as Co(III), Fe(III), Cr(III), Fe(II) or Ir(III) are frequently used [69,70]. Typical divalent transition metal hexacyanometallates (III) exhibit structures based upon the cubic T3[M(CN)6]2 framework as showed in Fig. 2\n, where [M(CN)6]2 ion complexes are linked via the octahedrally coordinated nitrogen bound T2+ ions [71].Several studies have demonstrated that the DMC catalysts are active in the CO2/PO copolymerization process. Kruper and Swart, 1995 demonstrated that the three dimensional Zn3[Fe(CN)6]2 based systems were mildly active for the random copolymerization of CO2/PO to produce PPC. The polymer product comprised of 2.8\u00a0g of PPC and about 0.53\u00a0g of polypropylene oxide (PPO). The total ratio of carbonate to the ether was 4.8 with 71\u00a0% PO conversion [73]. In 2004, Chen and co-workers, employed a Zn3[Co(CN)6]2 catalyst system (Fig. 2) for the CO2/PO copolymerization reaction. This system exhibited enhanced catalytic activity of over 1000\u00a0g copolymer/g of catalyst in comparison to its analog based on Zn3[Fe(CN)6]2\n[74]. Under optimal reaction conditions (110\u00a0\u00b0C, 3.8\u00a0MPa of CO2,10\u00a0h), a PPC yield of 2000\u00a0g/g Zn3[Co(CN)6]2 was obtained exhibiting a 0.24\u00a0M fraction of CO2.Investigations on the structures of the DMC catalysts informed the possible reaction mechanism for the copolymerization process, concluding that an epoxide ring opening occurs at the Zn-OH group of the DMC in the initiation step of the reaction [75\u201377]. In typical CO2/epoxide copolymerization reactions over DMC catalysts, Dharman et al. [77] and Stahl and Luinstra [76], proposed a multistep reaction mechanism with competing routes as shown in Scheme 4. After epoxide ring opening and coordination to the catalyst, carbonate linkages are formed through CO2 insertion in a step-by-step polymerization process. On the other hand, ether linkages are also formed through a Lewis base assisted nucleophilic attack of an hydroxyl group at the coordinated PO in chain growth process [76,77].In this respect, the copolymerization process using DMCs as catalysts usually give polymers with a mixture of carbonate and ether linkages, as well as cyclic carbonate byproducts [79]. Significant efforts have been made to improve the activity and yield of PPC in the copolymerization process over DMC catalysts. In the 2005, Kim et al. performed the copolymerization of CO2 and epoxides using DMC synthesized from a zinc salt and K3Co(CN)6 in the presence of tert-butanol and poly(tetramethylene ether glycol) (PTMEG) as the complexing reagent. The study observed a high reactivity of DMC catalysts in the copolymerization of PO with supercritical CO2 (sCO2), giving a yield of 343\u00a0g of polymer/g Zn in 2\u00a0h compared to the homogeneous diethylzinc-based catalyst [51,80]. The polymer yields in the absence of sCO2 were 507\u00a0g and 535\u00a0g of polymer/g Zn at 50\u00a0\u00b0C and 80\u00a0\u00b0C reaction temperatures, respectively in 24\u00a0h. The copolymer carbonate fraction (FC) at a lower temperature was higher (22\u00a0% at 50\u00a0\u00b0C) than that at a higher temperature (13\u00a0% at 80\u00a0\u00b0C). Further, the catalyst system showed higher activity (526\u00a0g polymer/g Zn) for alicyclic epoxides like cyclohexene oxide (CHO) in the CHO/CO2 copolymerization (PCO2\u00a0=\u00a0140 psi) at 80\u00a0\u00b0C after 4\u00a0h in comparison to aliphatic epoxides like PO [80]. In 2006, Robertson et al. reported a series of anhydrous DMC catalysts of the formula Co(H2O)2[M(CN)4].4H2O (M\u00a0=\u00a0Ni, Pd, Pt) for the random copolymerization of CO2/PO [81]. These DMC catalysts exhibit a two dimensional structure as shown in the X-ray crystal structure for the (Co(H2O)2[Pd(CN)4].4H2O) catalyst displayed in Fig. 3\n\n[81]. The reactions gave polymers with high number average molecular weights (Mn\u00a0=\u00a02.33\u20130.26\u00a0\u00d7\u00a0105 g/mol) and broad molecular weight distributions (MWDs\u00a0=\u00a05.8\u20132.3). The activity of these catalysts was significantly lower than that of Zn3[Co(CN)6]2 based systems, but they produced no cyclic propylene carbonate in the polymer [81].The Co[Ni(CN)4] catalyst showed better performance in the CO2/PO copolymerization (TOF\u00a0=\u00a01,860\u00a0mol PO (mol Co)\u22121h\u22121, at 130\u00a0\u00b0C and 54.4\u00a0atm CO2), but with low CO2 incorporation in the PPC polymer chain (FC\u00a0=\u00a020\u00a0%) [81].Studies have shown that modification of the coordination environment around the central metal (M) in the Znx[M(CN)y]z DMC catalyst, affects its catalytic activity in the CO2/PO copolymerization reaction. Zhang et al. reported that distortion of the octahedral coordination sphere by replacing one of the CN\u2013 ions with other anions, reduced efficiency in the copolymerization reaction [82]. In their study, low CO2 mole fractions (<0.36) in the polymer and high cyclic carbonate by-product yields (1.1\u201362.5\u00a0%) were observed. The low catalytic performance (<500\u00a0g polymer/g catalyst) of the resultant catalysts in comparison to the original DMC (1,466\u00a0g polymer/g catalyst) was attributed to the change in the electron-donating effect of the substituting groups. Thus, distorting the coordinative sphere around Zn affects the efficiency of the catalyst [82]. When Zn3[Co(CN)6]2 DMC catalyst was utilized in the absence of chain transfer agents, the catalytic activity in the CO2/PO copolymerization yielded 60,600\u00a0g polymer/g catalyst after 10\u00a0h [83]. However, the carbonate content in the obtained PPC ranged between 34\u00a0% and 49\u00a0%, although there was a decrease in the formation of the propylene carbonate by-product to below 1.0\u00a0%. Due to the dependence of PPC stability on the carbonate content, the Mn of the as-polymerized PPC with a carbonate content of 48\u00a0% reached 130,000\u00a0g/mol but decreased to 60,000\u00a0g/mol after 24\u00a0h of storage at 70\u00a0\u00b0C, and further dropped to 40,000\u00a0g/mol after 7\u00a0days [83].In another study, Varghese et al. reported a DMC catalyst where the potassium ion in the traditionally prepared DMC system was replaced with H+ from H3Co(CN)6) [84]. The hydrogen containing DMC catalyst exhibited enhanced catalytic performance in the CO2/PO copolymerization reaction giving a PPC yield of 1260\u00a0g polymer/g catalyst. The PPC displayed a carbonate fraction of 66\u00a0% and a polycarbonate selectivity of 97\u00a0%, which was a great improvement compared to the traditionally prepared DMC catalyst with a yield of 600\u00a0g polymer/g catalyst, carbonate fraction of 10\u00a0%, and PC selectivity of 92\u00a0% [84]. Furthermore, two active 2D- nanolamellar DMC catalysts (Zn\u2013Ni DMC and Co\u2013Ni DMC catalysts) synthesized via ball milling were reported by Guo et al. with an improved catalytic activity for the CO2/PO copolymerization reaction [85]. At optimal reaction conditions (60\u00a0\u00b0C, 24\u00a0h, 4\u00a0MPa of CO2), the Zn\u2013Ni DMC catalyst displayed a higher activity. The PPC obtained had a higher molecular weight (Mn\u00a0=\u00a010,344\u00a0g/mol, PDI\u00a0=\u00a01.45) containing 83.5\u00a0% carbonate linkages and higher content of CO2 (42.7\u00a0%), while the Co\u2013Ni DMC catalyst gave PPC (Mn\u00a0=\u00a08,478\u00a0g/mol, PDI\u00a0=\u00a01.44) with 73.7\u00a0% carbonate linkages, and lower CO2 content (35.9\u00a0%) [85]. Recently, Penche et al. 2021, reported a series of porous hexacyanometallate (III) complex catalysts for the ring-opening copolymerization of CO2-PO reaction. In their study, the catalysts displayed moderate activity giving copolymers exhibiting carbonate units in the range 16 to 33\u00a0%, coupled with fairly high molecular weights (Mw\u00a0=\u00a06,000\u201385,400 gmol\u22121) [86].Generally, DMCs have exhibited high catalytic activity in the CO2/PO copolymerization reaction. However, a major drawback of these systems is the low CO2 incorporation into the growing polymer chain causing high degree of ether linkages, due to significant homopolymerization of PO. In this respect, low percentages of carbonate linkages (20\u201340\u00a0%) in PPC are observed for the DMC systems. In addition, the attendant copolymerization reaction conditions are harsh, with temperatures in the range of 80 to 130\u00a0\u00b0C and pressures from 50 to 100\u00a0atm. Moreover, formation of the cyclic carbonate byproduct in significant yields is also reported for a number of catalysts [87\u201389].In an effort to further improve catalytic performance of DMC based catalysts in the CO2/PO copolymerization, Lu et al. 2013, designed multi-metal cyanides (MMCs) consisting of three different metals, which exhibited enhanced catalytic activity [90]. The catalysts were synthesized from different salts in varying ratios as shown in Table 1\n.The polycarbonates produced using the MMCs catalysts, exhibited slightly higher carbonate linkages compared to those formed with the DMC catalysts. MMC-2 exhibited higher activity compared to MMC-1 and MMC-3. However, under optimal conditions and at 70\u00a0\u00b0C, the PPC produced by MMC-2 displayed slightly lower carbonate content and molecular weight compared to PPC formed using MMC-1 and MMC-3 which was attributed to the electron atmosphere around the central metal [90].In 1991, success in the CO2/PO copolymerization reaction was achieved by Chen et al. using ternary rare-earth metal coordination catalysts [91]. The ternary rare-earth metal catalyst system composed of yttrium phosphonates, triisobutyl aluminium, and glycerin (Y(P204)3-Al(i-Bu)3-glycerin) showed the highest activity [91]. The copolymers formed had a high molecular weight, narrow molecular weight distributions, and high thermal stability. Furthermore, the copolymerization process gave high polymer yields within shorter reaction times compared to the previous organometallic catalytic systems. However, the copolymers were randomly arranged with low carbonate linkages (30\u201340\u00a0%) [91]. This motivated further studies that led to the design of composite catalytic systems as an effort to improve the carbonate content in the polymer [92].In order to address the selectivity and activity challenges suffered by different catalysts in the copolymerization reaction, synergetic effects of combined catalysts were explored. An efficient composite catalyst of ZnGA/DMC (Zn3[Co(CN)6]2) for the CO2/PO copolymerization that produced PPC with high molecular weight and in high yield was reported by Meng et al. [93]. This composite catalyst containing a small amount of DMC, exhibited a higher activity, selectivity, and shorter reaction time than that of the traditional ZnGA catalyst. An alternating PPC with a high CU (97.7\u00a0%) and in high yield (508\u00a0g polymer/g cat) was obtained under optimal reaction conditions of 24\u00a0h, at 70\u00a0\u00b0C and 50\u00a0bar CO2 pressure. The resultant PPC (Mw\u00a0=\u00a0380,000\u00a0g/mol) also showed good thermostability (Tg\u00a0=\u00a042.0\u00a0\u00b0C, Td (5%)\u00a0=\u00a0253.4\u00a0\u00b0C) [93].Using the same composite catalyst system but with a higher ratio of DMC (DMC/ZnGA; 10:1), An et al. [94] reported an increased activity for the CO2/PO copolymerization reaction in the presence of a PPG initiator compared to the findings of Meng et al. [93]. This increase in activity was attributed to the high concentration of DMC in the composite catalyst. However, this compromised selectivity towards the carbonate insertion in the polymer resulting in low carbonate unit contents (45.1\u00a0%) in the polymer. The ZnGA catalyst in the composite improved the carbonate unit content in the copolymer and also transferred the propagating chain with more alternating structures compared to the DMC catalyst. Thus, the polymer chain growth was attributed to the DMC catalyst, while the increased CO2 insertion into the growing polymer chain was attributed to the ZnGA catalyst. Under optimal conditions, oligo (propylene\u2010carbonates) with Mn of 1,228\u00a0g/mol and a high yield of 1,689\u00a0g/g cat were obtained [94]. Despite the cost-effective synthetic procedures of these carboxylate/DMC composites, the catalyst reaction conditions are still harsh for the thermodynamically unstable PPC products formed.In an earlier study, Tan and Hsu, 1997 combined a yttrium rare earth ternary (RET) coordination complex with the diethylzinc catalyst for the CO2/PO copolymerization reaction [92]. The yttrium carboxylate-dialkylzinc glycerol catalyst system gave an alternating PPC with a high carbonate content (95.6\u00a0%), in high yield and within a short reaction time [92]. The polycarbonates generated had alternating arrangements with high Mn (100,000\u00a0g/mol) in a 12\u00a0h reaction period, unlike the polycarbonates formed with the other RET catalysts without ZnEt2\n[91,92]. The resulting yield (4,200 (g/mol of Y)/h) was much higher than that obtained (2,451\u00a0g/mol of Y) for a 16\u00a0h run using only the RET catalyst system, reported by Chen et al. [91,92]. On the other hand, the yttrium carboxylate without the diethylzinc gave random copolymers with low carbonate units having inferior properties for industrial applications [92]. The RET/ZnEt2 composite offered a high activity in the copolymerization process due to the RET carboxylate and the enhanced generation of alternating PPC copolymers was attributed to the diethylzinc [92]. When this catalyst composite was combined with ZnCo-DMC as reported by Dong et al. in 2012, there was a reduction in the carbonate content in the PPC, though activity remained high [95]. The reduction in the selectivity for the CO2 was attributed to the DMC catalyst in the composite catalyst.Liu et al. 2003, developed a neodymium carboxylate-based complex composed of Nd(CCl3CO2)3, ZnEt2, and glycerine (Fig. 4\n), for the copolymerization process [96]. The catalyst system gave a high PPC yield with an activity as high as 6,875\u00a0g/mol of Nd in an 8\u00a0h run at 90\u00a0\u00b0C. The copolymers obtained had a Tg above 37\u00a0\u00b0C, though the metal residue in the polymer affected its stability [96].When yttrium-benzoate complexes (Y(RC6H4CO2)3 (R\u00a0=\u00a0H, OH, Me, or NO2) in combination with ZnEt2 and glycerol were tested for the copolymerization process, they gave alternating PPCs with up to 98.5\u00a0% carbonate linkages with productivities exceeding 100 (g/mol Zn)/h [37]. By varying the substituents at different positions of the aromatic ring in benzoate, the microstructure of PPC could be moderately adjusted and the head-to-tail linkage in polymer varied from 68.4 to 75.4\u00a0%. The polymers obtained by these catalysts had broad molecular weight distributions (Mw/Mn) which was attributed to the steric factor of the ligand in the yttrium complex, whereby substituents at the 2- and 4-positions are believed to affect the coordination or insertion of the monomer into the polymer chain. Although RET/ZnEt2 based composites have exhibited promising performance in the copolymerization process, the complex synthetic procedures of the RET coordination compounds limit their large-scale industrial application.Generally, the dialkylzinc based catalysts exhibited low catalytic activity in the CO2/PO copolymerization reaction, low turnover frequencies (TOFs) and gave copolymers with low and broad molecular weight distributions (Table 2\n, entries 1 and 2). These shortfalls were attributed to lack of optimal access to metal centers by monomers resulting into production of a wide range of polymers. On the other hand, the double metal cyanides exhibited high catalytic activity in the copolymerization process, gave higher polymer yields in shorter reaction times. However, higher reaction temperatures and pressures were required for these systems (Table 2, entries 3 and 4). Moreover, the polymer products were marred with low carbonate linkages and broad to narrow molecular weight distributions. When carboxylates were utilized in the CO2/PO copolymerization process, the polymer yield was low but with high carbonate content, high molecular weight, and narrow molecular weight distributions (Table 2, entry 5). Use of composites, that employ more than one catalyst for the copolymerization reaction, resulted into improved performance due to favorable synergistic effects (Table 2, entries 7\u20139). The PPC obtained were in high yields, having high molecular weights and high carbonate content within a short reaction time.Polypropylene carbonate (PPC) is one of the most common CO2-based copolymers whose production not only contributes to mitigation of climate change through CO2 utilization, but this copolymer has several applications due to its attractive properties. However, PPC large scale production is still limited by the chemical inertness of CO2 and the lack of suitable catalytic systems that can give high turnover numbers combined with attractive properties of the resultant polymer. Over the past few decades, researchers in both academia and industry have made effort to develop efficient homogeneous and heterogeneous catalytic systems for PPC synthesis. However, the performance of the heterogeneous systems so far developed do not favorably compete with their homogeneous analogues, despite the several positive attributes enjoyed by the former.Dialkylzinc based catalysts were the first active systems designed for the CO2/PO copolymerization process. However, these catalysts exhibited low turnover frequencies (TOFs) and gave copolymers with very broad molecular weight distributions. These setbacks were attributed to the multinuclear nature of the catalysts and the lack of optimal access to metal centers by monomers producing a wide range of polymers. Improved performance in the CO2/PO copolymerization process was observed with zinc glutarate based catalyst systems. The easy synthesis, non-toxicity and high activity of the ZnGA systems, made them attractive for industrial application in polycarbonate synthesis. Despite the high catalytic activity and selectivity for polycarbonates exhibited by heterogeneous ZnGA, most of their structures were not well-defined and their TOFs were not as competitive as those of the homogeneous catalysts. Thus, more effort is still needed to improve the structure and surface properties of the ZnGA based catalysts. Double metal cyanides (DMCs) are another class of catalysts that have exhibited promising catalytic activity in the CO2/PO copolymerization reaction. However, the major drawback of these systems is low CO2 incorporation into the growing polymer chain causing high ether linkages due to PO homopolymerization. Moreover, DMCs require high reaction temperatures which degrades the PPC resulting in formation of the cyclic byproducts which are thermodynamically more stable.Rare-earth metal catalysts in the presence of co-catalysts have also exhibited enhanced catalytic activity in the CO2/PO copolymerization process. The polymers formed exhibit high molecular weight, narrow molecular weight distributions, and high thermal stability. Generally, the stability and high catalytic activity of these systems have made them potential candidates for industrial application in PPC synthesis. However, the complex and costly synthetic procedures required for these catalysts hinders their use on a large scale. In attempt to improve both selectivity and activity in the CO2/PO copolymerization process, catalyst composites have been designed by combining different catalysts under appropriate proportions. Typical catalyst composites comprising of DMCs and ZnGA catalysts showed enhanced catalytic performance in the copolymerization process due to favorable synergistic effects. The increase in activity of the composites was attributed to the DMC in the composite catalyst. Although this also led to low selectivity towards carbonate insertion in the polymer resulting in low carbonate unit contents in the polymer. On the other hand, the enhanced CO2 insertion into the copolymer was attributed to the ZnGA catalyst.Despite the promising achievements made in synthesizing active catalysts for PPC formation, still to date, a balance between catalyst activity and selectivity for the desired polymers with good properties is still lacking. Generally, catalysts with a high activity give PPC with poor material properties while several catalysts with comparatively low activities give PPC with promising polymer properties. In this respect, further research should be sought to develop efficient heterogeneous catalyst systems that can produce PPC copolymers in excellent yields without compromising the desired polymer properties. The use of composite catalysts offers favorable synergistic effects of different catalyst in the CO2/PO copolymerization reaction, since enhanced selectivity and activity has been observed for these systems compared to the individual catalysts. In this respect, further investigations into catalyst composites is likely to offer more efficient catalysts for the copolymerization process.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors wish to express their sincere thanks for financial support from the International Programme in Chemical Sciences (IPICS), under the International Science Programme (ISP), Uppsala University, Sweden, and The German Academic Exchange Service (DAAD).", "descript": "\n                  Chemical conversion of carbon dioxide (CO2) into value-added products is an attractive industrial process because it offers several economic and environmental advantages. This review presents advances and challenges in the CO2 and propylene oxide (PO) co-polymerization using heterogeneous catalysts to form poly (propylene carbonate) (PPC), an environmentally friendly polymer with several applications. In the co-polymerization process, CO2 is employed as a green carbon source, an alternative to the toxic phosgene which has numerous negative environmental impacts. However, this route of polycarbonate production, is hindered by the chemical inertness of CO2, and to overcome this, various catalysts have been developed. A number of heterogeneous catalysts including carboxylates, double metal cyanides and composites, have achieved varying success in activating CO2 in the production of polycarbonates. The effect of different reaction conditions including pressure, temperature and solvent has been explored. The limitations faced by various heterogeneous catalysts and improvements made over the past decades have been highlighted. Mechanistic insights for the production of PPC from CO2 and PO have been presented and the differences in both the regioselectivity and stereochemistry of the resultant polymers discussed.\n               "}
{"full_text": "Biomass is considered as a potential renewable feedstock for production of value added chemicals such as fuels and fine chemicals. Biomass, having low nitrogen and sulphur content is easily renewable and excellent source of energy with low CO2 emission [1,2]. The production of industrially important bio-chemicals using sugars has recently become a priority for many countries. 5-Hydroxymethylfurfural (HMF), is an important chemical, which can be produced by single step acid catalysed conversion of sugars. The production of 5-HMF over heterogeneous catalysts has many advantages including minimal catalyst deactivation, low waste, and cost-effective purification of products [3,4]. Among various solid acid catalysts, nanocrystalline zeolites with high surface area are widely used in different refining and petrochemical industries as heterogeneous catalysts [5]. Zeolites are also effective for the enhancement of hydrocracking activity [6,7]. The synthesis of heteropoly acids loaded zeolite catalysts like 12-Molybdophosphoric acid (PMo), Ni and Ni\u2013PMo loaded HZSM-5 zeolites were well reported in the literature [8\u201310]. The conversion of sugars using Lewis acidic zeolites to produce useful chemicals is also studied [11]. Some researchers evaluated the conversion of C6 sugars like glucose into HMF using bi-functional catalyst systems [12,13]. A novel solid proton conducting material has been made by loading different weight percentages of heteropoly acids (HPA) onto Y-zeolite [14] (Scheme 1\n) [15].In the present work, a series of zeolite supported solid acid catalysts by loading different weight percentage of phosphomolybdic acid was synthesized. They were found efficient for conversion of C6 sugars into HMF, which is an important bio-based platform chemical.Natural zeolite (NZ) was repeatedly washed with double distilled water, dried, crushed, and then grinded in Agate mortar. These particles were then washed several times with water and dried in an oven at 100\u00a0\u00b0C for 1 h and then calcination was done at 400\u00a0\u00b0C for 4\u00a0h in muffle furnace. Thus obtained activated zeolite was grinded again in agate mortar to form smaller particle size.10\u00a0g NZ was loaded with 0.1\u00a0g, 0.3\u00a0g, 0.5\u00a0g, 0.7\u00a0g of phosphomolybdic acid (PMA) (for 1\u00a0wt%, 3\u00a0wt% 5\u00a0wt%, 7\u00a0wt% loading respectively) in 100\u00a0ml ethanol solution. The solution was stirred without heating in a 5 MLH magnetic stirrer at 350\u00a0rpm for 24\u00a0h. The prepared solution was aged for 48\u00a0h and then filtrated. The residue was dried at 100\u00a0\u00b0C for 4\u00a0h. The prepared PMA/NZ catalysts (PMA/NZ-1, PMA/NZ-3, PMA/NZ-5, PMA/NZ-7) were crushed and then stored in the desiccator.N2 adsorption-desorption was done by using Quantachrome (Model: Autosorb -Iq-Tpx) surface area analyzer. XRD patterns were recorded on X\u2019pert Pro-3 powder, using Ni-filter and Cu K\u03b1 radiation (E\u00a0=\u00a08047.8\u00a0eV, \u03bb\u00a0=\u00a01.5406A\u00b0) in 2\u03b8 range of 10\u00b0- 80\u00b0 at a scanning rate of 1\u00b0 /min. Morphology and surface topography was studied by FESEM (Tescan Model: MIRA-3 LMH). Thermo gravimetric analysis (TGA) of samples was carried out using TA InstrumentsSDT-Q-600, with a heating rate of 10\u00a0\u00b0C/min under nitrogen flow (50\u00a0cm3/min)..The reaction was carried out in a 100\u00a0ml stainless steel hydrothermal autoclave kept in silicon oil bath. To prepare the solution, 3\u00a0g. of reactants (glucose and fructose) were taken with 0.15, 0.2, 0.3 and 0.6\u00a0g. of PMA/NZ-1, PMA/NZ-3, PMA/NZ-5 and PMA/NZ-7 catalysts respectively mixed with 20\u00a0ml absolute ethanol, one at a time. The solution was poured into the autoclave and sealed pack the instrument. The whole assembly was then placed in oil bath and stirred at magnetic stirrer continuously at 140\u00a0\u00b0C for 4\u00a0h. A thermometer is kept in silicon oil bath to observe temperature time to time. Thus obtained solution was cooled at room temperature and kept out from the autoclave after 24\u00a0h and filtered using whatman filter paper.From the reaction it was found that among all the catalysts, PMA/NZ-3 showed maximum conversion percentage of glucose and fructose. So all the characterization studies were done for PMA/NZ-3 catalyst only. In Fig. 1\n, the N2 adsorption desorption isotherm of synthesized PMA/NZ-3 catalyst exhibited a BET surface area of 7.2\u00a0m2 g\u22121 and pore volume of 0.005\u00a0cm3\u00a0g\u22121 and the pore diameter was determined to be 16.664\u00a0\u00c5 using BJH method. The BET surface area of natural zeolite is 99.096\u00a0m2\u00a0g\u22121 and pore volume is 0.024\u00a0cm3\u00a0g\u22121.In order to understand the phase symmetry of the prepared catalyst, a systematic study on the XRD was undertaken. The XRD pattern as shown in Fig. 2\n expressess that the material dominantly contains a mordenite mineral phase ((Ca,Na2,K2) Al2Si10O24\u00b77H2O) with major characteristic peaks at 2\u03b8 values of 9.79\u00b0 and 26.27\u00b0. Other mineral phases such as quartz (SiO2) and hematite (Fe2O3), as well as other types of zeolites such as clinoptilolite ((Na,K,Ca)2-3Al3(Al,Si)2Si13O36\u00b712H2O) and heulandite ((Ca,Na)2-3Al3(Al,Si)2Si13O36\u00b712H2O) were also observed. Further, no peak corresponds to heteropoly acids were observed in the XRD structure of modified zeolite. These results imply that the heteropoly acid is well dispersed on the support surface as an amorphous or microcrystalline phase without altering the support phase.The prepared catalyst PMA/NZ-3 is subjected TGA analysis to evaluate the amount of physically adsorbed water removed from the catalyst surface. Two weight loss regions are obtained in the thermogram of the catalyst as shown in Fig. 3\n. The first weight loss occurred between 150 and 250\u00a0\u00b0C, which corresponds to the loss of water: while the other weight loss obtained at 450 \u2013 550\u00a0\u00b0C relates to decomposition of Keggin structure accompanied by water removal.In Fig. 4\n, SEM-EDAX analysis of PMA/NZ catalyst showed the presence of Si, Al, O, Mo and P. The presence of molybdenum and phosphorus in the synthesized catalyst confirmed the loading of phosphomolybdic acid on the NZ suface.The percentage yield of 5-HMF produced by conversion of glucose and fructose was estimated by UV\u2013Visible spectrophotometer. 5-HMF absorbs at 284\u00a0nm under UV\u2013VIS region. A set of standard solutions of HMF has been prepared for calibration. The results are summarized in Table 1\n.In this work zeolites loaded with phosphomolybdic acid has been synthesized and used in conversion of glucose into value added chemicals namely HMF. Characterization of the prepared material by XRD, SEM \u2013EDAX and TGA confirmed the loading of PMA on zeolite structure. It is demonstrated that stable catalysts are possible to prepare which retain their structure and exhibit good catalytic properties. The activity is however increased initially marked upon loading till 3\u00a0wt% with heteropoly acid molecules and it decreases on further loading. The effect of loading is very clearly seen at low heteropoly acid loading (3\u00a0wt%).\nSonal Gupta: Conceptualization, Methodology, Software, Visualization, Investigation, Data curation, Writing \u2013 original draft, Writing \u2013 review & editing. A.B. Gambhire: Software, Data curation. Renuka Jain: Conceptualization, Methodology, Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.We would like to thank Government College, Kota, Rajasthan for all guidance and CSIR, New Delhi for financial support.", "descript": "\n                  Renewable biomass as a sustainable feedstock has been enormously explored for the manufacturing of high value-added products such as biofuels, commodity chemicals, and new bio-based materials. The present paper describes the synthesis of a series of natural zeolite based catalysts by loading different weight percentage of phosphomolybdic acid. The efficiency of catalysts were checked by conversion of carbohydrate (glucose and fructose) to 5-HMF, which is an important industrial chemical used in the production of various value-added chemicals, materials and biofuels.\n               "}
{"full_text": "Data will be made available on request.Biodiesel is regarded as a promising transmit solution to alleviate the environmental issue of burning fossil fuels [1,2]. However, the intensive development of the biodiesel industry has resulted in a huge accumulation of glycerol, a by-product from the biodiesel production process, which has led to abundant glycerol waste and a dramatically price drop in glycerol market that is further struck the economic profit of producing biodiesel [3,4]. Thus, converting glycerol into value-added products has become a promising approach to solve the problems of glycerol waste and improve the economics of biodiesel processes [4,5]. Among the various value-added products, glycerol carbonate (GLC) with the versatile properties can provide a wide range of applications [5,6], for example using as a solvent, a cosmetic ingredient, a laundry detergent, a building eco-composite, or a chemical intermediate [5\u20137]. Therefore, the value-added conversion of GL to GLC has attracted wide attentions, and four main methods have been developed to convert GL into GLC: (1) carbonization with phosgene or carbon monoxide [6]; (2) direct reaction with CO2\n[8]; (3) glycolysis with urea [9]; and (4) transesterification with dimethyl carbonate (DMC) [7,10,11]. Comparing the pros and cons of each of these four methods \u2014 such as the toxicity of co-reactant, thermodynamic equilibrium limitation of the reaction, difficulties in by-product separation, and the reaction conditions \u2014 GL conversion with DMC (shown in Scheme 1\n) appears to be the most promising route to form GLC, as DMC has been considered as a green chemical and this conversion route can be conducted at relatively mild operational conditions.Converting GL with DMC to GLC requires the presence of catalysts, where homogenous catalysts, such as KOH, NaOH, and H2SO4, have been reported to achieve excellent catalytic performance, but this type of catalyst is difficult to separate from the reaction system [12]. Thus, heterogeneous catalysts have attracted more attention due to their efficient recyclability and good catalytic performance. Compared to acid catalysts, the presence of base catalysts can lead to relatively high yield and selectivity of glycerol carbonate, and also fast reaction rate in the transesterification of GL and DMC [12], which has been reported that the basic site of a catalyst is responsible for activating GL by cleaving its OH bonds [13,14]. Liu et al. [15] established a good correlation between catalytic activity and surface basicity for transition metal doped hydrotalcite catalysts. However, it is still debatable if tuning the amount and strength of catalyst basic sites can determine their catalytic performance in GL transesterification. For example, Hu et al. [16] reported that 15\u00a0wt%\u00a0K/CaO catalyst calcinated at 700\u00a0\u00b0C shows higher glycerol conversion (99\u00a0%) than CaO (92\u00a0%) in the transesterification of GL and DMC, but the amount of basic sites of 15\u00a0%\u00a0K/CaO-700 (30.37\u00a0mmol/g) is lower than that of CaO (33.93\u00a0mmol/g). MgO with a trapezoidal morphology has been synthesized and tested in glycerol transesterification, and it showed the highest glycerol conversion and GLC yield but with the lowest amount and weakest strength of basic sites compared to the MgO catalyst in a rod-like, spherical, flower-like, and nest-like structure [17]. Therefore, the surface basicity of a catalyst might not be the only factor that affects its catalytic performance in the glycerol transesterification.In addition, alkali and alkaline earth metals are high abundant in Earth\u2019s crust [18], in particular calcium, sodium, magnesium and potassium, and their unit price is relatively cheap which provides great potential to be applied in industry. So, many alkali and alkaline earth metal based and modified catalysts have been studied in transesterification of GL and DMC, for example, introducing lithium to ZnO, La2O3 and ZrO2 support catalyst have been found significantly enhanced the conversion of glycerol to GLC from barely converted to over 90\u00a0% [7,19,20]. Moreover, the ionic radius and valence state of the alkali and alkali metals have been reported playing important roles on the doping location and coordination sphere, which further influence their catalytic abilities. Sugiura et al. [21] studied the alkali metal ion substitution on a layered calcium phosphate compound (octacalcium phosphate), and revealed that the difference in ionic radius between alkali metal and calcium affects the location of alkali metal ions in the layered compound. Ferreira et al. [22] used diffuse reflectance UV\u2013vis spectra to analyse the coordination number changes of CeO2 after the addition of Ca and Mg, with Ca/CeO2 showing a lower coordination number (approximately 8) of Ce4+ ions. However, the effect of alkali and alkaline earth metals in improving the catalytic performance of metal oxides for glycerol carbonate production has not been fully understood. Additionally, no systematic investigation into the different combinations of the modifiers and base metal oxides currently exists, which has hindered the development of effective catalysts and efficient catalytic processes in glycerol value-added conversion.Therefore, this research work presents a systematically study of the promotional roles of alkali and alkaline earth metals on improving catalytic performance of La2O3 in transesterification of GL and DMC. The ionic radius and valence state for alkali and alkaline earth metals were found as dominant factors for improving the catalytic activity of La2O3, and other factors including molecular weight, surface composition, crystallinity, electron status, specific surface area and basicity of modified La2O3 samples were further elucidated.All chemicals used in this work are of analytical grade and without further purification. Lanthanum nitrate (La(NO3)3\u00b76H2O), ammonium carbonate, magnesium nitrate (Mg(NO3)2), barium nitrate (Ba(NO3)2), N,N-dimethylformamide (DMF, 99\u00a0%), glycerol (99\u00a0%), methanol (99\u00a0%) and tetraethylene glycol (99\u00a0%) were purchased from Alfa Aesar. Lithium nitrate (LiNO3), sodium nitrate (NaNO3), potassium nitrate (KNO3), calcium nitrate (Ca(NO3)2), strontium nitrate (Sr(NO3)2), dimethyl carbonate (99\u00a0%), 4- (hydroxymethyl)-1,3-dioxolan-2-one (90\u00a0%) were purchased from Thermo Fisher Scientific Inc. Glycidol (96\u00a0%) was purchased from Sigma-Aldrich.The support material, La2O3, was synthesised using a modified version of the precipitation method from Li et al. [19]. The preparation process is as follows, 0.03\u00a0mol lanthanum nitrate and 0.12\u00a0mol ammonium carbonate were each dissolved in 120\u00a0ml deionised (DI) water. The obtained ammonium carbonate solution was slowly added into lanthanum nitrate solution under mechanical stirring at room temperature and continuously stirred for 6\u00a0hrs. Then the white precipitate was separated in a centrifuge (SIGMA\u00ae 2-16P) and washed with DI water. Finally, the solid paste was dried at 110\u00a0\u00b0C for 24\u00a0hrs and calcined at 800\u00a0\u00b0C for 6\u00a0hrs in a Muffle furnace (Carbolite\u00ae ELF 11/14B) to obtain final La2O3 catalysts.La2O3 doped by 25\u00a0mol% alkali and alkaline earth metals were prepared by wet impregnation method. A certain amount of M(NO3)n (where M\u00a0=\u00a0Li, Na, K, Mg, Ca, Sr, Ba; n\u00a0=\u00a01,2) was dissolved in DI water and then 0.5\u00a0g La2O3 powder was added. The suspension was magnetically stirred for 12\u00a0hrs. The resulting slurry was evaporated at 80\u00a0\u00b0C in a water bath to remove excess water. The solid residue was dried at 110\u00a0\u00b0C for 10\u00a0hrs and then calcined at various temperatures (400\u00a0\u00b0C \u2212 800\u00a0\u00b0C) for 2\u00a0hrs.The samples were denoted as xM/La2O3T, where\u00a0\u00d7\u00a0represents the mass percentage or molecular percentage, M represents the alkali and alkaline earth metal, and T represents the calcination temperature. The default value for\u00a0\u00d7\u00a0and T are 25\u00a0mol% and 600\u00a0\u00b0C when the sample presented without \u2018x\u2019 and/or \u2018T\u2019.The crystal phases of pristine and modified La2O3 catalysts were characterised by powder X-ray diffraction (XRD) using a Bruker Phaser-D2 diffractometer with Cu K\u03b1 X-ray source. The scanning range (2\u03b8) was from 10\u00b0 to 90\u00b0, with a slit of 1\u00b0 at a scanning rate of 10\u00b0 min\u22121. The electron states for the samples were analysed via X-ray photoelectron spectroscopy (XPS), conducted on a Thermo ScientificTM K-AlphaTM+ spectrometer equipped with a monochromatic Al K\u03b1 X-ray source (1486.6\u00a0eV) operating at 100\u00a0W. All peaks have been calibrated with C1s peak where the standard binding energy (B.E.) is 284.8\u00a0eV for adventitious carbon source. The basicity of the samples was tested via CO2 temperature-programmed desorption (TPD). This was carried out by placing a 0.10\u00a0g sample into a U-shape reactor and pre-treating it under Helium flow (at 50 sccm) at 600\u00a0\u00b0C for 30 mins. CO2 was then introduced and adsorbed on the samples for 45 mins at room temperature. During the desorption process, the cell was heated up to 1000\u00a0\u00b0C with a ramping rate of 10\u00a0\u00b0C\u00a0min\u22121 under the Helium flow (at 50 sccm).The amount of alkali and alkaline earth metal doped on La2O3 was analysed by inductively coupled plasma optical emission spectrometry (ICP-OES) operated on a Varian Vista Pro instrument with axial view. The samples were digested in aqua regia and then diluted to a certain concentration before the measurements. The surface area and pore size distribution of the samples was determined via the N2 physical adsorption\u2013desorption experiments at 77\u00a0K in a chemisorption (& physisorption) gas sorption analyser (Quantachrome autosorb IQ). The samples were first degassed at 200\u00a0\u00b0C in vacuum for 2\u00a0hrs, and then their N2 isotherms were measured and analysed based on the Brunauer-Emmett-Teller (BET) equation theory. The morphology of the samples was determined by scanning electron microscope (SEM, JEOL JSM-6390A). Before measurements were taken, the sample was suspended in ethanol solution and dispersed in ultrasonic bath for 1\u00a0min. Then, the suspensions were added dropwise onto a copper tape for SEM analysis.The catalytic performance of La2O3 catalysts doped by alkali and alkaline earth metals were tested via glycerol (GL) conversion with dimethyl carbonate (DMC) in a stainless-steel reactor (Yanzheng\u00ae YZPR-100). The thermocouple was built in a stainless-steel blind tube inside the reactor for the temperature control. The reaction mixture was stirred with a magnetic stirrer during the reaction.The ratio 1:3 of GL and DMC was mostly claimed as optimal reactant ratio based on literature review [15,16,19,20,23\u201327], so in a typical experiment, 3.0\u00a0g of GL and 9.0\u00a0g of DMC were added into the reactor with 0.10\u00a0g of catalyst. After sealing the reactor, the mixture was continuously stirred and heated to a desired reaction temperature for a certain time. After this time was reached, the reactor was cooled down in an ice-water bath to stop the reaction. A certain amount of internal standard substance (ISTD), tetraethylene glycol, and DMF were added into the reaction mixture. Then the catalyst the catalyst was separated from the liquid phase in the centrifuge (SIGMA\u00ae 2-16P). The collected catalyst was retained and prepared for further recycling experiments. The obtained liquid phase was further analysed by gas chromatography (GC, Shimadzu GC-2010 plus), equipped with a flame ionization detector (FID) and a Stabilwax-MS (30\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm) column. The qualitative analysis of the reaction products was carried via Shimadzu gas chromatography-mass spectra (GC\u2013MS).In this section, the synthesis strategy of the wet impregnation method for doping alkali and alkaline earth metals on La2O3 catalysts is firstly illustrated in Section 3.1, then the catalyst characterisation obtained from the methods presented above for pristine La2O3 catalyst and La2O3 catalysts modified by 25\u00a0mol% alkali and alkaline earth metals are systematically discussed in Section 3.2. The catalytic performances of the pristine La2O3 and modified La2O3 catalysts in GL and DMC transesterification are then presented in Section 3.3, along with a discussion of the key characteristics which might affect the performance. The plausible mechanism was discussed based on modified La2O3 samples in Section 3.4. Finally, the optimal operating conditions for the transesterification of glycerol via doped La2O3 are shown in Section 3.5.The wet impregnation method was used for modifying La2O3 catalyst with alkali and alkaline earth metals. The principle behind this method is discussed based on synthesising Na/La2O3 catalyst which is shown in Scheme 2\n. The synthesis procedure includes three main steps: impregnation, drying, and calcination. During the first step of impregnation, the pre-synthesised La2O3 powder is uniformly dispended in the NaNO3 aqueous solution containing the Na+, NO3\n\u2013, H+ and OH\u2013 ions, and part of La2O3 phase could be transferred into La(OH)3 and La2O2CO3 phases during contact of water and CO2\n[28], and the reaction mechanisms are defined by Eq. (1) and Eq. (2) respectively. During the impregnation, Na+ ions might diffuse into the pores of La2O3 and be adsorbed onto the porous surface, and Na+ ions could also be adsorbed on the external surface of the support by forming the ion pair with its oxo/hydroxo-groups [29].During the drying procedure, the precursor of Na+ forms a homogenous distribution of the Nax(NO3)y(OH)z crystals on the surface of the La2O3 (or on La(OH)3 and La2O2CO3) crystal phases which is illustrated in Eq. (3),. As the solvent is removed during the drying process, it can result in a redistribution of the modified metal phase on the support material [30]. The dopants inside the pores which have a smaller size can more easily migrate out to the external surface and contribute to the formation of the Nax(NO3)y(OH)z crystals. In the calcination step, Nax(NO3)y(OH)z starts to decompose to NaNO3 at 100\u2013200\u00a0\u00b0C, then NaNO3 is converted into molten salt at around 308\u00a0\u00b0C [31], where the molten salt phase can increase the mobility of Na which can lead Na to enter the lattice of the support material. As described in Eq. (4), the molten NaNO3 salt firstly starts decomposing to NaNO2 at 380\u00a0\u00b0C [31], and then the formed NaNO2 salt further decomposes to Na2O at around 600\u00a0\u00b0C [32\u201334]. In the meantime, La(OH)3 and La2O2CO3 phases can start converting into La2O3 at 600\u00a0\u00b0C and complete the conversion at around 800\u00a0\u00b0C [35]. The formation of surface defects also occurs during the calcination step [35,36].\n\n(1)\n\n\n\n\n\n\nLa\n\n2\n\n\nO\n3\n\n\n\n\n\ns\n\n\n\n\n+\n\n\n\n\nNaNO\n\n3\n\n\n\n\n\na\nq\n\n\n\n\n+\n\nH\n2\n\nO\n\u2192\n\n\n\n\nLa\n\n2\n\n\nO\n3\n\n\n\n(\ns\n)\n\n\n+\n\n\n\n\nNa\n\n+\n\n\n\n(\na\nq\n)\n\n\n+\n\n\n\n\nNO\n\n\n3\n\n-\n\n\n\n(\na\nq\n)\n\n\n+\n\n\n\n\nH\n\n+\n\n\n\n(\na\nq\n)\n\n\n+\n\n\n\n\nOH\n\n-\n\n\n\n(\na\nq\n)\n\n\n\n\n\n\n\n\n(2)\n\n\n\n\nLa\n\n\n2\n\n\n\n\n\n\nO\n\n\n3\n\n\n\n\n(\ns\n)\n\n\n+\n\n\nH\n\n\n2\n\n\nO\n+\n\n\nCO\n\n\n2\n\n\n\u2192\n\n\nLa\n\n\n2\n\n\n\n\n\n\nO\n\n\n3\n\n\n\n\n(\ns\n)\n\n\n+\nLa\n\n\n\n\n(\nO\nH\n)\n\n\n3\n\n\n\n\n(\ns\n)\n\n\n+\n\n\nLa\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\nCO\n\n\n3\n\n\n\n\n(\ns\n)\n\n\n\n\n\n\n\n\n(3)\n\n\n\n\n\n\nLa\n\n2\n\n\nO\n3\n\n\n\n\n\ns\n\n\n\n\n+\n\n\n\n\nNa\n\n+\n\n\n\naq\n\n\n+\n\n\n\n\nNO\n\n\n3\n\n-\n\n\n\naq\n\n\n+\n\n\n\n\nH\n\n+\n\n\n\naq\n\n\n+\n\n\n\n\nOH\n\n-\n\n\n\naq\n\n\n\u2192\n\n\nNa\n\nx\n\n\n\n\n\n\n\nNO\n\n3\n\n\n\n\ny\n\n\n\n\n\nO\nH\n\n\n\nz\n\n\u2219\n\n\n\n\nLa\n\n2\n\n\nO\n3\n\n\n\n\n\ns\n\n\n\n\n\n\n\n\n\n\n(4)\n\n\n\n\nNa\n\nx\n\n\n\n(\n\n\nNO\n\n3\n\n)\n\ny\n\n\n\n(\nO\nH\n)\n\nz\n\n\u2192\nN\na\nN\n\nO\n3\n\n\u2192\nN\na\nN\n\nO\n2\n\n+\n\nO\n2\n\n\u2192\n\n\nNa\n\n2\n\nO\n+\nN\nO\n+\nN\n\nO\n2\n\n,\n\n\n\n\nNa\n\nx\n\n\n\n(\n\n\nNO\n\n3\n\n)\n\ny\n\n\n\n(\nO\nH\n)\n\nz\n\n\u2219\n\n\n\n\nLa\n\n2\n\n\nO\n3\n\n\n\n\n\ns\n\n\n\n\n\u2192\n\n(\nN\na\n)\n\n\n\n\n\nLa\n\n2\n\n\nO\n3\n\n\n\n\n\ns\n\n\n\n\n\n\n\n\nThe synthesis strategy for doping other alkali and alkaline earth metals \u2013 Li, K, Mg, Ca, Sr, and Ba \u2013 on La2O3 is similar to that for Na doping La2O3, but the decomposition temperatures of their nitrates to oxides are different to that of NaNO3 which are summarised in Table S1.The bulk and surface composition of La2O3 samples are analysed by ICP-OES and XPS, and the results are listed in Table S2. The bulk composition of each element tested by ICP-OES is consistent with the designed composition where the amount of alkali and alkaline earth metals are around 25\u00a0mol% of La. While the surface compositions of Li and Na are around 68\u00a0% which are three times higher than their overall composition, that of Mg is about 48\u00a0% and two times higher than its bulk composition, and the surface composition for the other metals are similar to their bulk results. This result suggests that the majority of Li, Na and Mg are doped on the La2O3 surface, while K, Ca, Sr and Ba could form another phase along the bulk of La2O3. As illustrated in the synthesis strategy in Section 3.1, during the drying process, the dopants with smaller ionic radii more easily migrate out of the inner pores and adsorb on the surface. Zhang et al. [37] also found that the surface composition of Li and Na with smaller ionic radii are higher than K, due to Li and Na are more easily to migrate on the ZnO surface than K.The crystal structure of prepared La2O3 samples and the doping location of alkali and alkaline earth metals were studied via XRD analysis, and the results are shown in Fig. 1\n. The XRD patterns of pristine La2O3 and the modified La2O3 samples followed the hexagonal structure lanthanum oxide phase (P-3\u00a0m1, JCPDS 83\u20131344), and the main diffraction peaks of La2O3 are observed at 2\u03b8\u00a0=\u00a026.1\u00b0, 29.1\u00b0, 29.9\u00b0, 39.5\u00b0, 46.0\u00b0 and 52.1\u00b0, corresponding to the (100), (002), (011), (012), (110) and (103) crystal planes, respectively. The blue stars in Fig. 1 represent for the La2O2CO3 phase (JCPDS 84\u20131963) which is inevitably formed when La2O3 sample is exposed to ambient atmospheric conditions [19,38]. In addition, no extra crystalline phases of alkali metal oxides were observed from their XRD patterns, therefore, the alkali metal could uniformly disperse on La2O3 surface [19,39,40]. Unlike surface doping, the bulk doping can influence the phase stability and crystal growth [40,41]. Diffraction peaks for alkaline earth metal doped La2O3 samples become much more weaken and broaden than that for alkali metal doped La2O3 samples, indicating the heavy doping of alkali earth metal inhibited the formation of La2O3 crystalline structure [41]. Nevertheless, the presence of CaO, SrCO3, and BaCO3 phase further confirmed that Ca, Sr, and Ba formed extra crystalline structure. These results suggest that most alkaline earth metals are incorporated in the bulk La2O3.X-ray photoelectron spectroscopy (XPS) analysis was used to clarify the electron environments for pristine and modified La2O3 samples. The XPS profiles for La 3d and O 1\u00a0s orbitals are shown in Fig. 2\n and the corresponding binding energies of these orbitals are listed in Table S3. Due to spin\u2013orbit coupling, the La 3d spectrum separated into two groups known as La 3d3/2 and La 3d5/2, respectively [19,42]. Each group can be further deconvoluted into one main peak (denoted as I), and two satellite peaks (denoted as II and III) [43,44]. The peaks of La 3d5/2 I for the pure La2O3 material are centred at 834.7\u00a0eV, and it is negatively shifted to between 834.2\u00a0eV and 834.4\u00a0eV for the La2O3 catalysts doped by alkali metals, while no peak shifting was observed for the samples doped by alkaline earth metals. This result suggests that alkali metals donate electrons to La, which makes the La in alkali metal doped La2O3 catalysts be able to donate more electrons to reactants [39].The peaks of O1s spectra are located at around 528.7\u00a0eV, 530.7\u00a0eV, and 531.4\u00a0eV, corresponding to the lattice O2\u2212 (OL), chemisorbed surface O\u2212 (OS) and weakly adsorbed OH\u2013 and CO3\n2\u2013 (OA) species, respectively [15,39,42,44]. The binding energy of O 1\u00a0s spectra are positively shifted towards the higher energy field for both La2O3 catalysts doped by alkali metals and alkaline earth metals, with the latter elements being shifted more. This result indicates that a larger amount of electrons transferred from O sites to alkaline earth metals than to alkali metals, which indicates that alkaline earth metals have a strong interaction with O in La2O3\n[45,46]. This is also consistent with the XRD result discussed in the previous section that an extra phase of alkaline earth metal oxide was formed on the La2O3 surface.The surface basicity of La2O3 catalysts was measured by CO2-TPD analysis and is shown in Fig. 3\n, and the corresponding densities of their basic sites are calculated and listed in Table S4. The CO2 desorption peak for the pure La2O3 catalyst is centred at 400\u00a0\u00b0C \u2212 500\u00a0\u00b0C, so the basic sites for pristine La2O3 sample can be denoted as strong basic site [19]. After doping alkali and alkaline earth metals on the La2O3 surface, the CO2 desorption peaks shifted to a higher temperature field compared to that of pristine La2O3 sample and are located at 600\u00a0\u00b0C \u2013 800\u00a0\u00b0C, in which the basic sites for the modified samples can be denoted as extra strong sites [19 47]. As a result, the basicity of the La2O3 sample became stronger with introducing alkali and alkaline earth metals. Additionally, alkali metal doped La2O3 samples contains 4.56\u20135.60\u00a0\u03bcmol/m2 basic sites, higher than alkaline earth metal doped ones with 0.26\u20135.34\u00a0\u03bcmol/m2 basic sites, which is consistent with the XPS results that the electrons around O in alkali metal doped La2O3 samples are more dense than those in alkaline earth metal doped La2O3 samples [47].The average crystallite sizes of La2O3 and modified La2O3 were calculated via the Debyee-Scherrer equation, and the results are listed in Table S2, where the average crystal size of pure La2O3 is 68\u00a0nm, which is the largest of all the samples. The average crystal sizes of the alkali metal doped La2O3 samples were the next largest at about 42.2\u00a0nm-57.9\u00a0nm and the smallest were the alkaline earth metal doped La2O3 samples (at about 18\u00a0nm-34.3\u00a0nm). As a result, adding alkali and alkaline earth metal can hinder the growth of La2O3 crystal [30]. This is consistent with their specific surface areas measured by N2 isotherms which are shown in Fig. 4\n a, and their specific surface areas, calculated by BET theory, are listed in Table S2. The surface areas of La2O3 samples doped by alkali metals are smaller than that of La2O3 samples doped by alkaline earth metals. The main pore size of all samples is distributed in the range of 2\u20134\u00a0nm, as shown in Fig. 4 b, while the pore width and total volume of modified La2O3 catalysts, especially in the range of 2\u20134\u00a0nm, became broader and higher than that of the pristine La2O3 catalysts. The surface morphologies of La2O3 and promoted La2O3 catalysts were individually determined by SEM analysis, and presented in Figure S1. The morphologies of the pure La2O3 sample and La2O3 samples doped by alkaline earth metals are similar and show a flake-like structure, while nanorod-like structures are shown in the La2O3 samples doped by alkali metals.In this section, the catalytic performance of pristine La2O3 catalyst and modified La2O3 catalysts by alkali and alkaline earth metals are presented in Section 3.3.1, followed by a thorough discussion on the potential factors of the dopants and catalyst characteristics in improving the catalytic performance of La2O3.The catalytic performance of the pristine La2O3 catalysts and the La2O3 catalysts modified by 25\u00a0mol% alkali and alkaline earth metals was examined in the transesterification of GL to GLC at 70\u00a0\u00b0C and 2\u00a0hrs and the results are presented in Fig. 5\n a. The pristine La2O3 catalyst has barely any conversion of GL into GLC at 70\u00a0\u00b0C after 2\u00a0hrs, while doping alkali and alkaline earth metals on La2O3 catalysts significantly improved the catalytic performance of La2O3 catalyst in GL and DMC conversion. La2O3 catalysts doped by Li, Na and K achieved 48\u00a0%, 85\u00a0% and 40\u00a0% GL conversion, respectively, and the GL conversions for La2O3 catalysts doped by Mg, Ca, Sr and Ba were 13\u00a0%, 41\u00a0%, 21\u00a0% and 23\u00a0%, respectively. The GLC yield follows the similar trend as the GL conversion for alkali and alkaline earth metal doped La2O3 catalysts. Among all the modified La2O3 catalysts, Na doped La2O3 catalyst shows the best catalytic performance, and Ca doped La2O3 catalyst led to relatively higher GL conversion and GLC yield than the La2O3 catalysts doped by other alkaline earth metals. In addition, the alkali metal doped La2O3 catalysts showed relatively better catalytic performance than alkaline earth metal doped La2O3 catalysts.To rule out the effect of molecular weight for doping metals, a fixed mass ratio, of alkali and alkaline earth metals were doped on La2O3 and tested in the GL transesterification, where the results are presented in Fig. 5b. The 3.5\u00a0wt% mass ratio of Na/La is the equivalent mass ratio to the 25\u00a0mol% of Na/La catalyst, and the equivalent molar ratios (eq. mol%) for other dopants are listed in Table 1\n. The catalytic performance of La2O3 based catalysts with the fixed mass ratio shows a similar trend as that for the catalysts with a fixed molecular ratio. As shown in Fig. 5 b, 3.5\u00a0wt% Na/La2O3 and 3.5\u00a0wt% Ca/La2O3 catalyst also achieved the highest glycerol conversion of 85\u00a0% and 37\u00a0%, respectively, among La2O3 catalysts doped by alkali metal and alkaline earth metal, and La2O3 catalysts doped by alkali metals showed better catalytic performance than the ones doped by alkaline earth metals. This result shows that the molecular weight of dopants is not the dominant factor that determines the ability of alkali and alkaline earth metals on improving catalytic performance of La2O3 catalyst.The catalytic performance of a catalyst is determined by its properties which can be tuned by the dopant added. So, in this section, the correlation between the catalyst properties and the catalytic performance are discussed, and the dominant effects of the dopants on improving the catalytic performance of La2O3 catalyst are revealed.Interestingly, the ionic radius of Na (1.02\u00a0\u00c5) and Ca (0.99\u00a0\u00c5), listed in Table 1, is similar to the cation radius of La (1.03\u00a0\u00c5), and Na/La2O3 and Ca/La2O3 catalyst have showed the best GL conversion and GLC yield among the La2O3 catalysts modified by alkali metals and alkaline earth metal, respectively, so the similarity in their radius might be the dominant factor in affecting the dopant interaction with support material. Thus, a correlation between the ionic radius and catalytic performance was established and shown in Fig. 6\n.For alkali metal promoted La2O3 samples, the ionic radius ratio of Na/La is 0.99 which indicates the radii of Na+ and La3+ are very similar, and the ionic radius ratio of Li/La and K/La are 0.74 and 1.34, indicating the radius of Li+ and K+ relatively smaller or larger than that of La3+, respectively. Interestingly, the GL conversion under La2O3 catalyst doped by Na, which has a similar ionic radius with La, is higher than that under La2O3 catalyst doped by metals with a dissimilar ionic radius with La, such as Li and K. This result indicates that the La2O3 catalyst doped by a dopant with a similar ionic radius to La can maximise the improvement of its catalytic performance in glycerol and DMC transesterification. The same trend was also found for the La2O3 catalysts doped by alkaline earth metals, with Ca having the most similar ionic radius to La, and having the highest GL conversion among the alkaline earth metals. Song et al. [7] and Kaur et al. [20] also claimed that ZnO and ZrO4 catalyst doped by lithium showed better catalytic performance than those catalyst doped by other alkali metals because Li has a similar ionic radius (0.76\u00a0\u00c5) to Zn (0.74\u00a0\u00c5) and Zr (0.72\u00a0\u00c5). Therefore, the similarity of ionic radius for the dopant to its support material can be considered as one of the determining factors in improving the catalytic performance of the catalyst. The potential reason might be when the ionic radius of the dopant is smaller or larger than the support material, it could cause a significant structure distortion around the dopant, which might affect the stability of the active sites, so when the ionic radius is similar to the support cation, the formed active sites is more stable which leads to the better catalytic performance. Moreover, when comparing the dopants in the same period, the catalysts doped by alkali metals are superior to the ones doped by the alkaline earth metal in GL and DMC reaction, indicating that the valence state of the dopant also affects the catalytic performance of modified La2O3 catalysts.The surface concentration of alkali and alkaline earth metals was tested via XPS, as stated in Section 3.2.1, and the surface concentration of Li, Na and Mg is higher than their overall concentration, where that of K, Ca, Sr, and Ba is similar to their overall concentration. It has been reported that surface doping might be better than bulk doping [40], but in this work, the surface concentration of the dopant did not show a proportional effect on the catalytic performance of modified La2O3 catalysts. For instance, the surface concentration of Li and Na is similar, but 25\u00a0mol% Na doped La2O3 catalyst achieved 85\u00a0% GL conversion, much better than 48\u00a0% GL conversion for 25\u00a0mol% Li doped La2O3 catalysts at the same reaction conditions. In addition, 25\u00a0mol%\u00a0Mg/La2O3 catalyst achieved less GL conversion than 25\u00a0mol% Ca/La2O3 catalyst, although the surface concentration of Mg as about two times higher than that of Ca. Thus, the surface concentration is not the dominant factor for La2O3 catalysts doped by alkali and alkaline earth metals showing different catalytic ability in GL conversion.The diffraction peaks for La2O3 catalysts doped by alkaline earth metals from XRD analysis as shown in Fig. 1 are much broader and weaker than those for the pristine La2O3 catalysts and the ones doped by alkali metals, which agrees with the previous study [49,50]. The broader and weaker peaks could result from the formation of a non-crystalline phase, the aggregation of particles [51], or an incomplete La2O3 crystal phase. As illustrated in the synthesis strategy, given in Section 3.2.2, La2O3 can easily transfer to La(OH)3 and La2O2CO3 phases when it comes into contact with water or carbon dioxide in the atmosphere, so during the drying and calcination process, the La(OH)3 and La2O2CO3 gains are decomposed into La2O3 and rebuild the crystalline structure. Castro et al. [41] reported that alkaline earth metals, especially Mg and Ca, could prevent host material reforming from the calcination process. So a higher calcination temperature 800\u00a0\u00b0C was used to synthesise the modified La2O3 catalysts, and their XRD patterns, as shown in Figure S2, indicate a better crystalline structure of hexagonal lanthanum oxide was formed. Conversely, the catalytic performance for the samples calcined at 600\u00a0\u00b0C is better than that for the samples calcined at 800\u00a0\u00b0C (as shown in Figure S3), despite the crystal structures being well formed at 800\u00a0\u00b0C. In addition, the catalyst calcined at 800\u00a0\u00b0C achieved a similar trend as the ones calcined at 600\u00a0\u00b0C in GL and DMC conversion, where Na/La2O3 and Ca/La2O3 showed the best catalytic performance among the La2O3 catalysts doped by alkali metals and alkaline earth metals, respectively, and the alkali metal doped La2O3 catalysts showed relatively higher GL conversion than alkaline earth metals doped La2O3 catalysts. Thus, these results further suggest that the crystallisation degree for La2O3 catalysts doped by alkaline earth metals is not responsible for improving the catalytic performance.The full analysis of electronic states of La2O3 catalysts modified by alkali and alkaline earth metals is given in Section 3.2.3, where alkali metals as dopants on La2O3 catalysts can donate their electrons to La but alkaline earth metals showed little effect on the electronic environment of La. This might be due to the alkali metals having a lower electronegativity than alkaline earth metals [52], meaning alkali metals are more likely to donate their electrons than alkaline earth metals. This phenomenon has also been reported on alkaline earth metals doped on ZrO2\n[51] and Ni/La2O3 catalysts [53]. Additionally, alkali metals affect the electron distributions around O sites, while alkaline earth metals as dopants on La2O3 catalysts more strongly affect the electron distributions around O sites, which could be due to the extra phase of alkaline earth metal oxides formed [51]. This result is consistent with their catalytic performance where alkali metal doped La2O3 catalysts showed relatively better catalytic performance of alkaline earth metal doped La2O3 catalysts, which further reveals that the valence states of alkali and alkaline earth metals is one of dominant factors. However, for dopants within the same group, this cannot explain the trend in their catalytic performance.The basic sites have been reported to be an important factor for a catalyst to achieve high GL conversion, so the correlation between basic site density and catalytic activity of alkali and alkaline earth metals doped La2O3 is presented in Fig. 7\n. The basic site density of La2O3 catalyst significantly increased after doping with alkali and alkaline earth metals as presented in Table S3, but the catalytic activities of La2O3 based catalysts are not proportional to the density of their surface basic sites. For instance, the basic site density of Na/La2O3 and Ca/La2O3 catalysts are very similar (5.60\u00a0\u03bcmol/m2 and 5.34\u00a0\u03bcmol/m2, respectively), but the GL conversion under Na/La2O3 catalysts is about 40\u00a0% higher than that under Ca/La2O3 catalyst. In addition, the basic site density of Li and K doped catalysts are also close to that of Na doped La2O3 catalyst, but Li and K modified La2O3 catalysts showed much lower GL conversion and GLC yield than Na/La2O3 catalyst. Therefore, these results imply that basic sites are important to the transesterification of GL and DMC, but it is not a determining factor.The plausible mechanism for the transesterification of GL and DMC on the La2O3 catalyst is proposed and shown in Fig. 8\n. The carbonyl group of DMC and the hydroxyl group of GL are activated on the La site and the O site, respectively [25]. Then the activated GL anion attacks the carbonyl carbon of activated DMC to form a 1-(o-methoxy-carbonyl)glycerol complex (denoted as intermediate 1) and one molar methanol. The intermediate 1 then further cyclise to GLC with another molar of methanol. The addition of alkali metals on La2O3 catalysts can form the M\u2212La centre that help activate DMC and cyclise intermediate 1, which leads to higher GL conversion and GLC yield (as shown in Fig. 5 a) than the pristine La2O3 catalyst. The extra phase formed by adding alkaline earth metals can provide more active sites and benefit the catalytic reaction of GL and DMC. The better catalytic performance of La2O3 catalyst promoted by alkali metal than by alkaline earth metals can illustrate that M\u2212La active centre could be more effective in reducing activation energy of the reaction than extra phase of alkaline earth metal oxide.As La2O3 catalyst doped by 25\u00a0mol% Na achieved the highest GL conversion and GLC yield, Na/La2O3 catalyst was chosen for further investigation on the effect of dopant/support metal molar ratio, calcination temperature, catalyst dose, reaction temperature, reaction time and reusability. Each variable is considered independently using a standard set of conditions, and the best optimal condition for each variable is carried forward to the next variable analysis.The impact of Na/La molar ratio for modified La2O3 (xNa/La2O3) was studied in the GL conversion to glycerol carbonate. The catalytic activity of xNa/La2O3 calcined at 600\u00a0\u00b0C was studied under the reaction temperature of 70\u00a0\u00b0C for 2\u00a0hrs and the results are shown in Fig. 9\n. The result indicates the GL conversion was significantly increased from 32\u00a0% to 85\u00a0% with increasing the amount of Na from 7\u00a0mol% to 25\u00a0mol% doping on La2O3 catalysts, and the GLC yield was increased from 25\u00a0% to 59\u00a0% respectively. With increasing the amount of Na further to 50\u00a0mol% of La, GL conversion slightly increased to 89\u00a0%, but the GLC yield dropped to 47\u00a0% in which the yield of glycidol (GLD) increased from the decarbonation reaction of GLC. Therefore, it can be inferred that the GL conversion and GLC yield can be boosted with increasing amount of Na to 25\u00a0mol% of La2O3 catalyst due to an increase in the number of active sites, but further increasing the amount of Na benefits the generation of the by-product. Thus, to balance the high GL conversion and GLC yield, 25\u00a0mol% Na/La ratio was chosen as the optimal doping ratio, which are used for further optimisation.To investigate the effect of calcination temperature on the catalyst activity, 25\u00a0mol% Na/La2O3 was calcined at various temperatures ranging from 400\u00a0\u00b0C to 800\u00a0\u00b0C, the obtained catalysts were tested in the GL and DMC conversion at 70\u00a0\u00b0C, and the reaction time was set for 2\u00a0h. The results are presented in Fig. 10\n a. With the increase of calcination temperature from 400\u00a0\u00b0C to 600\u00a0\u00b0C, the GL conversion significantly increased from 5\u00a0% to 85\u00a0%, but then dropped from 85\u00a0% to 70\u00a0% with increasing the calcination temperature further to 800\u00a0\u00b0C. The crystalline structure of 25\u00a0mol% Na/La2O3 catalysts calcined at 400\u00a0\u00b0C to 800\u00a0\u00b0C was tested via XRD, and the result is presented in Fig. 10 b. When the 25\u00a0mol% Na/La2O3 catalyst was calcined at 400\u00a0\u00b0C and 500\u00a0\u00b0C, the catalyst was mainly in the La2O2CO3 phase, which then gradually transferred to the La2O3 phase with increasing the calcination temperature, and La2O3 phase was the only phase presented at the Na/La2O3 catalyst calcined at 800\u00a0\u00b0C, which is consistent with the literature work [35]. In addition, the crystallinity of the catalysts increased with increasing the calcination temperature. These results indicate that La2O3CO3 phase is not active in the conversion of GL into GLC. Increasing calcination temperature not only benefits to forming La2O3, but also promotes the NaNO3 decomposition to Na2O leading to better interaction with La2O3 catalyst. Although the La2O3 was also completely formed at 700\u00a0\u00b0C and 800\u00a0\u00b0C, the GL conversion and GLC yield are decreased with GLD yield increased which might be due to the aggregation of particles where Li et al. [19] reported a similar phenomenon of the decrease catalytic performance of Li/La2O3 catalyst calcined at 700\u00a0\u00b0C in GL and DMC conversion compared to the catalyst calcined at 600\u00a0\u00b0C. Thus, 25\u00a0mol% Na/La2O3 catalysts calcined at 600\u00a0\u00b0C were used for the further optimisation of reaction parameters.The amount of 25\u00a0mol% Na/La2O3 catalysts calcined at 600\u00a0\u00b0C was investigated in the range of 0.01\u00a0g to 0.20\u00a0g for 70\u00a0\u00b0C and 2-hour GL and DMC reactions. As shown in Fig. 11\n a, with the amount of catalyst added from 0.01\u00a0g to 0.15\u00a0g, the production of GLC increases from 50\u00a0% to 60\u00a0% then drops to 40\u00a0% when further increasing the catalyst dose to 0.20\u00a0g, but the yield of by-product GLD also increased with the increase of catalyst dose. This result indicates that increasing the total number of active sites not only benefits the GL conversion to GLC, but also boosted the GLC decomposition to GLD. The catalyst performance peaks between 0.10\u00a0g and 0.15\u00a0g of catalyst, as the GLD yield is lower when 0.10\u00a0g 25\u00a0mol% Na/La2O3 catalyst was used, it seems to be more effective. Thus, 0.10\u00a0g of 25\u00a0mol% Na/La2O3 catalysts calcined at 600\u00a0\u00b0C was used for next optimisation of reaction temperature.The optimisation of the reaction temperature with 0.10\u00a0g of 25\u00a0mol% Na/La2O3 catalyst in GL and DMC conversion was carried out in the range of 50\u00a0\u00b0C \u2212 90\u00a0\u00b0C for 2\u00a0hrs, and the result is presented in Fig. 11 b. The GL conversion significantly increases from 33\u00a0% to 96\u00a0% at a temperature increase from 50\u00a0\u00b0C to 80\u00a0\u00b0C. With temperature increasing further to 90\u00a0\u00b0C, the amount of glycerol converted is steady, but the yield of glycerol carbonate reduced. The maximum production of glycerol carbonate was 59\u00a0% when the reaction temperature was 70\u00a0\u00b0C. The glycerol transesterification to GLC is an endothermic reaction, therefore GL conversion increases with the reaction temperature increase [54]. However, the higher reaction temperature can drive GLC decomposed to GLD. Therefore, 70\u00a0\u00b0C was selected as the optimal reaction temperature.In the transesterification of glycerol and dimethyl carbonate, there is one primary product of glycerol carbonate and one by-product of glycidol. As shown in Fig. 11 c, with an increase in the reaction time from 0.5 hr to 4\u00a0hrs, the conversion of glycerol improved from 66\u00a0% to 95\u00a0%, and the yield of glycerol carbonate enhanced slightly from 58\u00a0% to 62\u00a0%. When the reaction time is beyond 2\u00a0hrs, the selectivity of GLC decreased and the amount of by-product GLD increased. So the optimal reaction time was selected as two hours.To investigate the reusability of the catalyst, the spent 25\u00a0mol% Na/La2O3 catalyst was recycled and reused in a fresh glycerol reaction operating at the optimal conditions (70\u00a0\u00b0C and 2\u00a0hrs), and the results are shown in Fig. 11 d. The catalyst activity slightly deceased after two cycles. The GL conversion dropped from 85\u00a0% to 71\u00a0% after the second cycle, and further dropped to 61\u00a0% after the fourth cycle. To reveal the reason of the deactivation, the following experiments were designed, and the results indicated that the catalyst decay is due to surface wearing which some active components lose in the solution as a fine powder because very high stirring speed (1000\u00a0rpm) was applied for the reaction system to ensure the catalyst well dispersing in the immiscible GL and DMC environment. In group A, to test whether the decay is caused in the process of diluting reaction media or cleaning the residue reactants and products from the catalyst, the fresh Na/La2O3 catalyst was washed in ethanol (A1) and DMF (A2), respectively, at room temperature and dried completely at 110\u00a0\u00b0C in the oven before adding into the reactor. The glycerol conversion and GLC yield under pre-treated Na/La2O3 catalyst, shown in Table 2\n, are similar to that under fresh catalyst without treatment. Therefore, the solvents used in the regeneration process is not the reason causing the deactivation of the catalyst. In group B, the catalyst was separated from the reaction system via centrifuge after 1\u00a0hr reaction (B1), and after removing the catalyst particles the solution media continues the reaction for another 1\u00a0hr (B2). The results, listed in Table 2 for group B, show that the glycerol conversion and glycerol carbonate production increased after removing the catalyst particles from the reaction system, which revealed the some fine catalyst particles still remained in the solution and the deactivation of the catalyst likely resulted from wearing of active components from the catalyst surface. As the GL conversion continues after removing the La2O3 catalysts by centrifuge, one possible reason may be the leaching of Na ions from La2O3, and the leached Na promotes the reaction [7,19]. To clarify whether the deactivation is due to the leaching of Na ions from the doped La2O3, a certain amount of NaNO3 was added in GL and DMC, but no products were detected from gas chromatograph as shown in group C. So the reaction cannot be conducted with only the presence of Na+. As a result, the catalyst decay is not due to the leaching of Na+ into reaction system. Thus, the catalyst deactivation is mainly due to the surface wearing rather than the regeneration process or leaching.As the catalyst surface wearing is mainly caused by the friction between the magnetic stirrer and the bottom of the reactor, reducing the friction through increasing the volume of the solution and using mechanical agitator rather than a magnetic stirrer can be a potential solution to prevent the catalyst surface wearing. Moreover, using stronger support materials or coating the catalyst into stronger support materials, such as Al2O3 and ZrO2, or applying advanced synthesis methods, such as sol\u2013gel and precipitation, have been reported as effective way to prevent and reduce the catalyst decay from catalyst surface wearing [55].Alkali and alkaline earth metal doped on La2O3 catalysts were successfully synthesised via a wet impregnation method, and their catalytic activities were tested in glycerol conversion to glycerol carbonate. Na was found to be the best dopant among the alkali and alkali earth metals for improving La2O3 catalytic performance, and 25\u00a0mol% Na doped La2O3 catalyst achieved highest GL conversion (85\u00a0%) and GLC yield (60\u00a0%) at 70\u00a0\u00b0C in a 2-hour reaction. La2O3 catalysts doped by alkali metals achieved relatively higher GL conversion and GLC yield than that doped by alkaline earth metals.Ionic radius and valence state of dopants play significant roles in affecting the catalytic performance of La2O3 catalysts. In general, alkali metals were well dispersed on La2O3 surface, while alkaline earth metals were formed an extra phase and aggregated on La2O3 surface. The electron distribution around La on La2O3 surface was affected by doping alkali metal, while doping alkaline earth metal on La2O3 surface affected activity of O sites. The similarity of ionic radius between the dopant and La was found as one of the determining factors for improving La2O3 catalytic performance. The dopant with an ionic radius close to La led to a larger improvement in La2O3 catalytic activity than other dopants with a smaller or larger ionic radius, and the dopant with a lower valence state showed a better enhancement for La2O3 catalytic activity. The basic sites on La2O3 surface were found important to the transesterification of GL and DMC, but the basic site density on modified La2O3 surface was not a determining factor for the catalytic performance of La2O3 catalysts. The findings about the effect of radius and charge of alkali and alkaline earth metals on La2O3 catalytic activity are expected to help to understand the promotional role of the dopant for designing more efficient and lower cost catalysts for glycerol value-added conversion to glycerol carbonate.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Thanks to the financial support from the UK research council EPSRC (EP/V041665/1). The authors appreciate Mr Fergus Dingwall from School of Engineering in The University of Edinburgh for all his technical support; Dr. Laetitia Pichevin and Dr. Nicola Cayzer from School of Geoscience in The University of Edinburgh for their kind help on ICP-OES and SEM analysis; and Dr. Gary Nichol from School of Chemistry in The University of Edinburgh for providing XRD facility and valuable advice.Supplementary data to this article can be found online at https://doi.org/10.1016/j.cej.2023.141486.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  Alkali metals and alkaline earth metals have been used to promote the catalytic performance of metal oxides in transesterification of glycerol and dimethyl carbonate, however, the promotional roles of the dopants in influencing the catalytic performance of the metal oxides have not been fully investigated which hinder the development of low-cost and high-efficiency catalysts in transesterification of glycerol and dimethyl carbonate. This paper, for the first time, systematically studied the influence of ionic radius and valence state of dopants, surface concentration of dopants and the basicity of the catalysts on the catalytic performance of La2O3 in transesterification of glycerol and dimethyl carbonate. Our results suggest that the ionic radius and valence state of the dopants are the determining factors, while the basic site density is not a crucial factor, although the basicity of catalyst surface is important in activating glycerol and dimethyl carbonate. Among alkali and alkaline earth metals, 25\u00a0mol% Na/La2O3 catalyst achieved the highest glycerol conversion (85\u00a0%) and glycerol carbonate yield (60\u00a0%) in the 70\u00a0\u00b0C and 2-hour reaction. After the detailed investigation, a plausible mechanism of glycerol and dimethyl carbonate transesterification on Na/La2O3 catalyst has been proposed. This research could help understand the promotional role of alkali metals and alkaline earth metals and the results may guide future design of metal oxide catalysts.\n               "}
{"full_text": "Aqueous phase reformingAttenuated total reflection-infraredBinding energyCO-Temperature programmed desorptionCO2-Temperature programmed desorptionDensity functional theoryDiffuse reflectance infrared Fourier transform spectroscopyExtended X-ray absorption fine structureFourier-transform infrared spectroscopyH2-Temperature programmed reductionIonic exchangeInternational Energy AgencyInfraredIncipient wetness impregnationMulti walled carbon nanotubesNH3- Temperature programmed desorptionProton exchange membrane fuel cellsPolyvinylpyrrolidoneSolution combustion synthesisSol-gel in acidic conditionsSol-gel in basic conditionsSingle wall carbon nanotubesTransmission electron microscopyTetraethylorthosilicateTurn over frequencyTime on streamTemperature programmed oxidationUrea matrix combustion methodWater gas shiftWeight hourly space velocityX-ray absorption near edge structureX-ray absorption spectroscopyX-ray photoelectron spectroscopyX-ray diffractionThe use of fossil fuels are considered the leading cause of global warming. For this reason, alternative and sustainable sources of energy are explored nowadays to overcome this issue. The use of biomass to substitute fossil oil to produce materials and energy brought to the development of the biorefinery concept. According to the International Energy Agency (IEA) definition, it is intended as the integration of different processes for the sustainable production of goods starting from biomass [1]. One of the advantages of biorefinery is the possibility of targeting small, decentralized plants; as a matter of fact, several drawbacks have been raised for large-scale implementation. One reason is due to the difficulty to reach a complete and effective exploitation of the organic content of the starting biomass. For example, in the biodiesel industry, 1\u00a0kg of undesired crude glycerol is produced together with 10\u00a0kg of the desired product, generating a considerable amount of waste [2]. Similarly, pyrolysis and hydrothermal liquefaction aim at producing a biofuel (namely bio-oil or bio-crude) [3], but a significant fraction of the carbon present in the feed is lost in the aqueous stream [4].Aqueous phase reforming (APR) process was proposed by Dumesic and coworkers to valorize oxygenated molecules and obtain a gas mixture rich in hydrogen [5]. It derives that it can be applied to carbon-laden wastewaters present in bio- and conventional refineries to increase the conversion efficiency of the plant, reducing the amount of waste that should be treated, and obtaining at the same time a valuable product. Most of the literature refers to the use of model compounds, such as alcohols (methanol, ethanol), polyalcohols (ethylene glycol, glycerol, xylitol, sorbitol) and represent a pillar of the present review since it focuses mainly on the development of the catalyst. Fewer works have been dedicated to complex matrixes (glucose, xylose, woody biomass) [6,7]. Finally, limited researches described the performance of multicomponent mixtures, close to an industrial application (e.g. wastewater from the brewery industry, food industry) [8\u201314].Previous reviews on aqueous phase reforming focused both on the influence of the reaction conditions and the catalytic systems. Davda et\u00a0al. reported in the first review the fundamentals of their pioneering research [15]. Chen et\u00a0al. reported the different reaction mechanisms among the substrates [16], while Coronado et\u00a0al. summarized in their review a large number of issues, helping to compare various parameters [17]. Vaidya et\u00a0al. classified the available literature on the base of the starting substrate [18]. Finally, very recently, El Doukkali et\u00a0al. reviewed the research of effective catalysts for the valorization of glycerol through steam reforming, hydrogenolysis and, indeed, APR [19].Despite most of the works looked at the development of an effective catalyst, this subject has not been comprehensively reviewed so far in the APR field. For this reason, the present work aims to review the actual knowledge on the design of catalytic systems for the valorization of biomass-derived compounds via APR.Chapter 2 deals with a brief introduction to the process to show its possible applications and advantages. Thermodynamic and kinetic considerations are reported, with a particular focus to the latter, since its knowledge constitutes the basis for the rational design of the catalyst.The core of the work is based on the effect of different parameters on the performance of the process. The scientific contributions were extensively reviewed, starting from the pioneering works of the Dumesic's research groups up to the most recently published ones.Chapter 3 outlines the influence of the active site, both in monometallic and bimetallic systems, with specific attention to Pt-based bimetallic systems; moreover, structure sensitivity effects are discussed. Chapter 4 describes the impact of the support's choice on the APR performance. Chapter 5 deals with the effect of the preparation method as a key step for determining the properties of the catalyst and, in turn, the yield of the desired product. The open questions and challenges related to catalytic and technological subjects are discussed in Chapter 6.The research papers have been classified in the corresponding chapter according to the main aim of the study, although defining proper boundaries is not trivial. For example, bimetallic catalysts can alter both electronic properties (modifying the bond strength of reactants, intermediates and products) and structural properties (modifying the dispersion and, consequently, the number of available sites).Throughout the review, specific attention is put to the use of density functional theory (DFT), microkinetic modelling or in situ techniques to get information for the design of new catalysts. In fact, since the catalyst structure can be modified by the environment during the reaction itself, in situ characterizations can inspect connections between the catalyst structure and its performance. In conclusion, the primary outcomes are summarized and integrated to have some final messages for the design of an active, selective, and stable APR catalyst.The aqueous phase reforming is a process carried out at mild temperatures (220\u2013270\u00a0\u00b0C) and pressures (30\u201360\u00a0bar), in the presence of a catalyst [15]. In these conditions, the aqueous solution remains in the liquid phase, leading to an energetic advantage thanks to the avoided vaporization. This is one of the potential benefits of APR compared to the conventional steam reforming process, that is performed typically at about 800\u00a0\u00b0C. The reaction stoichiometry is reported in equation (1).\n\n(1)\n\n\n\nC\nn\n\n\nH\n\n2\ny\n\n\n\nO\nn\n\n\u2194\nn\nC\nO\n+\ny\n\nH\n2\n\n\n\n\n\nIn Fig.\u00a01\n, the influence of temperature on the Gibbs free energy for the reforming of alkanes and alcohols is reported. It can be observed that oxygenated hydrocarbons have a more favorable equilibrium, i.e., the hydrogen production can occur at lower temperatures than if obtained from alkanes.In the same temperature range, the water gas shift (WGS) reaction is favored as well (equation (2)).\n\n(2)\n\n\nC\nO\n+\n\nH\n2\n\nO\n\u2194\nC\n\nO\n2\n\n+\n\nH\n2\n\n\n\n\n\nThis occurrence allows to generate in one reactor a gas mixture where CO is present in negligible percentage, making the gas stream a compatible feed for low temperature proton exchange membrane fuel cells (PEMFCs); contrarily, in the steam reforming plant, a double configuration with high- and low-temperature water gas shift reactors is necessary to maximize the hydrogen yield.Combining equations (1) and (2) we can derive the conventional reaction stoichiometry for APR (equation (3))\n\n(3)\n\n\n\nC\nn\n\n\nH\n\n2\ny\n\n\n\nO\nn\n\n+\nn\n\nH\n2\n\nO\n\u2194\nn\nC\n\nO\n2\n\n+\n\n(\n\ny\n+\nn\n\n)\n\n\nH\n2\n\n\n\n\n\nIt is important to recall here that, despite the advantageous thermodynamics for hydrogen production, another reaction involving hydrogen consumption is more favored: the methanation reaction (equation (4)).\n\n(4)\n\n\nC\n\nO\n2\n\n+\n4\n\nH\n2\n\n\u2194\nC\n\nH\n4\n\n+\n2\n\nH\n2\n\nO\n\n\n\n\nThe APR mechanism is constituted by different steps, in which the interactions of the molecule with the active site, the support and their interface play a fundamental role to determine the product distribution. Knowing the possible steps involving the reacting molecule is essential to properly design the catalyst, tuning its structural, morphological, and textural properties to favor one pathway rather than the other. The possible reaction mechanism of glycerol APR is described in Fig.\u00a02\n. As will be reported in the next paragraphs, glycerol was chosen as model compound for several works, due to the importance of its valorization for the biodiesel value chain.The substrate can undergo dehydration (a) with the acid sites of the support, leading to hydroxyacetone, which can be subsequently hydrogenated (b) to propylene glycol. This route is undesired because hydrogen is consumed.If the molecule interacts with the metal active site, it is generally agreed that the dehydrogenation (c) of the molecule is the first step (glyceraldehyde intermediate). Afterwards, it can follow two routes. C\u2013O bond breaking (d) can occur, leading to the formation of alkanes. In this case, since the C\u2013H activation of alkanes is thermodynamically hindered at typical APR temperatures, their hydrogen content cannot be exploited. On the other hand, C\u2013C bond breaking (e) can lead to the formation of carbon monoxide. This intermediate, while adsorbed on the active site, may interact with water that, once catalytically activated (i.e., H2O \u2192 OH\u00a0+\u00a0H), produces hydrogen and carbon dioxide via WGS (f). Another detrimental route is that CO undergoes methanation (or Fischer-Tropsch reaction), consuming hydrogen (g).Taking into consideration the reported possible reaction pathways, some key points should be considered by the catalyst designer to maximize the hydrogen production:\n\n\u2022\nDehydrogenation, C\u2013C bond breaking, H2O activation, and water gas shift reaction should be favored;\n\n\n\u2022\nC\u2013O bond breaking, methanation/Fischer-Tropsch and dehydration should be avoided.\n\n\nDehydrogenation, C\u2013C bond breaking, H2O activation, and water gas shift reaction should be favored;C\u2013O bond breaking, methanation/Fischer-Tropsch and dehydration should be avoided.As graphically reported in Fig.\u00a03\n, preparing the catalyst involves three main choices regarding the preparation, the metal and the support. In the next chapters, the available literature will be presented with the aim of systematically pointing out how each of these choices can be beneficial or detrimental to the activation of each of the reported steps, i.e., favor C\u2013C bond breaking more than C\u2013O breaking, or activate water dissociation without worsening the support stability. APR is strongly sensitive to the reaction conditions, such as nature and concentration of the feed, reaction temperature and pressure, catalyst amount, reactor configuration. It derives that fair comparisons of results between different works are difficult and will be limited in this review.In the present chapter, the effect of the active metal is reported. At the beginning, a comparison between monometallic systems is performed, both using first-principles and experimental methods. Afterwards, the synergy of bimetallic catalysts is reported, with a higher attention to Pt-based and Ni-based formulations. Specific paragraphs are dedicated to systems considered particularly interesting both for the obtained results and for the extent of characterization (i.e., Pt\u2013Mo, Pt\u2013Co, Pt\u2013Ni, Pt\u2013Re, Ni\u2013Cu). Finally, the influence of the particle size is discussed.Platinum is the most studied metal for aqueous phase reforming since it combines high activity and moderate selectivity. These results have been primarily explained by several DFT investigations. Davda et\u00a0al. studied the C\u2013C and C\u2013O bond cleavages of ethanol by DFT on Pt [20]. Firstly, the authors reported that Pt\u2013C bonds are more stable than Pt\u2013O bonds examining the stability of different isomeric species. Secondly, they suggested that the C\u2013C bond cleavage should be faster than the C\u2013O bond cleavage because the energy for the transition state of the former is one order of magnitude lower than the latter (4 vs 42\u00a0kJ/mol) [21]. It follows that, recalling Fig.\u00a02, path e) is more favored than path d).In Fig.\u00a04\n the possible steps involved in ethanol reforming on Pt are depicted [22]. After being adsorbed on the active site, ethanol can follow two routes for the first dehydrogenation according to which H is abstracted (ethoxy CH3CH2O or 1-hydroxyethyl CH3CHOH); in the second step, this intermediate is dehydrogenated in acetaldehyde. Thereafter, the adsorbed acetaldehyde further dehydrogenates into acetyl (CH3CO), ketene (CH2CO) and ketenyl (CHCO) (please note that only the most plausible species were reported here among the possible pathways). It is just the ketenyl species that finally is subjected to C\u2013C bond breaking to CO and methylidyne (experimentally confirmed). This outcome has been compared in a proper work in which DFT, microkinetic investigation and experiments have been combined [23]. The authors observed that the level of hydrogenation of the intermediates influenced the C\u2013C or C\u2013O cleavage barriers because of geometric and electronic effects. In this sense, more dehydrogenated species facilitated C\u2013C bond breaking having the C\u2013C bond parallel to the surface. The microkinetic model also highlighted that the formation of 1-hydroxyethyl via \u03b1 C\u2013H scission (CH3CHOH) is predominant compared to CH3CH2O after first ethanol dehydrogenation (bold arrows in the corresponding figure). Further theoretical works on platinum with different substrates were reported in Refs. [24\u201326], attaining analogous conclusions.On ruthenium, the activation energy for C\u2013C bond cleavage is lower than on platinum, being 38\u00a0kJ/mol when the surface CH2CO species is formed (while it is 90.24\u00a0kJ/mol on Pt after CHCO), so it is more sensitive to the first dehydrogenation steps [27]. However, it is far more active towards methanation than Pt, decreasing the hydrogen selectivity [28].The decomposition of glycerol has been explored also on Pd, Rh, Cu, Ni, focusing on dehydrogenation, C\u2013C and C\u2013O cleavage [29]. Glycerol binds similarly on Pt, Pd, Rh, Cu, through the hydroxyl groups. The four possible mono-dehydrogenated species are reported, derived from C\u2013H or O\u2013H cleavage of the terminal or central carbon. Furthermore, the binding energies related to more dehydrogenated intermediates and C\u2013O cleavage are reported as well. In agreement with previous studies, it has been reported that C\u2013C breaking is favorable on Pt after several dehydrogenation steps, and it is more facile than C\u2013O breaking. Despite the simplification of the models, the results were coherent with the experimental outcome. The pathway is similar on Pd, i.e., dehydrogenation up to C3H5O3 is necessary and C\u2013C is more favorable, but the lowest C\u2013C scission transition state energy is higher than on Pt, suggesting that Pt is more active. Despite dehydrogenation is always necessary, the situation is different on Rh and Ni, where the energies between C\u2013C and C\u2013O are more comparable. Moreover, the transition states are quite low in terms of activation energy, explaining the experimentally known high activity of Ni despite its low selectivity. Finally, on Cu the transition states show a quite high activation energy, highlighting the low Cu activity. It is worthy to point out that the model was further implemented considering the presence of adsorbed CO. Higher coverage of CO increased the free energy for glycerol dehydrogenation, emphasizing the pivotal role of the water gas shift reaction in the process to reduce the concentration of adsorbed CO.Davda et\u00a0al. initially experimentally investigated Pt, Ni, Ru, Rh, Pd and Ir supported on silica for ethylene glycol APR [30]. In their kinetic study, the authors reported the promising ability of Pt to produce hydrogen, while Pd showed low activity and the highest selectivity (Fig.\u00a05\n). Looking at non-noble metals, Ni was comparable to Pd, but suffered from low selectivity (alkane production) and it was more prone to deactivation. The use of a cheap material, such as Ni, would increase the cost-effectiveness of the process. However, its low stability and H2 TOF (ten-fold lower than platinum) would impede its utilization in a larger scale context.It has been largely reported that bimetallic catalysts perform better than monometallic in many different fields, and APR is part of this family [31]. Before entering in the core of the paragraph, it is worthy to cite the first outcomes derived from the works carried out in the Dumesic's research group.Huber et\u00a0al. used a high-throughput system to study several catalyst formulations for ethylene glycol APR [32]. They found that Sn addition to Raney Ni catalyst greatly improved its performance. Sn was mainly present at Ni defect-sites and as an alloy (e.g. Ni3Sn), without affecting the Ni particle size. In this way, Sn hindered the CO methanation that, over Ni catalysts, preferentially occurs in the defects. This outcome was further explored in successive works. The addition of Sn, Au or Zn to Ni/alumina catalyst promoted the hydrogen selectivity, with Sn being the best promoter [33]. Being alumina not stable, Sn was added to Raney-Ni catalysts. With only 400:1 Ni:Sn atomic ratio, the methane production was halved, and eliminated at 18:1. The modified catalyst deactivated in the first 48\u00a0h by sintering, and afterwards by re-oxidation of the active site because of the interaction with water (no coke). Leaching of Ni was explained by formation of organometallic species after interaction of the catalyst with the feed. This phenomenon was reduced as well thanks to Sn addition. Finally, since Pt and Pd showed the best performance in monometallic systems, more than 130\u00a0Pt and Pd bimetallic catalysts based on these noble metals were screened using APR of ethylene glycol as probe reaction [34]. Four bimetallic catalysts were particularly interesting for their results: Pt\u2013Ni, Pt\u2013Co, Pt\u2013Fe and Pd\u2013Fe. The authors reported that alloying Pt with Ni, Co and Fe led to an electronic modification of Pt which caused the decrease in the binding energy (BE) of hydrogen and carbon monoxide and the increase in the dissociative adsorption energy of hydrogen. On the other hand, Fe addition in Pd-based catalyst promoted the WGS, which is considered the rate determining step on monometallic Pd.The next paragraphs report the outcome of several bimetallic systems, including the motivation that the authors proposed to explain the structure-activity relationship. The huge amount of information required a strong effort to rationalize and understand how to present it. They were not reported chronologically; rather, it was in the aim of the authors give a progressive insight into the phenomena, collecting different researches on similar systems to have common outputs. The paragraphs are divided according to the main actor of the system, e.g. platinum, nickel, etc., despite some of them could have been reported in more than one paragraph (i.e., the description of Pt\u2013Ni could have been stated in the Pt or Ni section). Finally, ad hoc sub-paragraphs were dedicated to the most interesting Pt-based (Pt\u2013Co, Pt\u2013Mo, Pt\u2013Ni, Pt\u2013Re) and Ni-based (Ni\u2013Cu) systems in order to go deeper with their comprehension. If not clearly stated, glycerol is considered as the feedstock.Godina et\u00a0al. investigated the APR of xylitol over mono and bimetallic catalysts [35]. Among the platinum-based bimetallic catalysts, Pt\u2013Re showed the highest conversion at each weight hourly space velocity (WHSV) investigated (Fig.\u00a06\n-A). This outcome was related to the higher water gas shift, promoted by Re. Low amount of erythritol was found in the liquid phase, suggesting that C\u2013O cleavage was favored over C\u2013C. Pt led mainly to propane, being active for C\u2013C, while Pt\u2013Re mostly led to butane and propane. Overall, because of the lower hydrogen selectivity of the bimetallic system than the monometallic one, the hydrogen turnover frequency of the former was lower than the latter (about 23 min\u22121 vs 26 min\u22121, at WHSV 9.5 h\u22121).Larimi et\u00a0al. investigated different promoters (Rh, Cr, Re, Pd, Ru, Ir) for Pt-based catalysts supported on \u03b3-alumina [36]. As reported in Fig.\u00a06-B, while the conversion was approximately constant (about 82%), the hydrogen selectivity was minimum for monometallic Pt (69.9%) and maximum for Pt\u2013Rh (89%). The variation of the lattice constant proved the formation of a solid solution. It was suggested that Rh helped the WGS, at the same time increasing the mobility and reactivity of surface oxygen atoms. The benefic addition of Rh was also reported in Ref.\u00a0[37], where the use of MgO-supported nano-sheet catalysts allowed stability during 100\u00a0h TOS (time on stream), overcoming the limitations reported by Ref.\u00a0[38].The addition of Rh can have different effects on different supports. When added to Pt/\u03b1-alumina, both the hydrogen yield and the stability increased, by preventing coke formation favoring methanation [39]. On the other hand, Pt/\u03b3-alumina's performance was negatively influenced by Rh, with a slight decrease of the hydrogen yield and catalyst life span (46.3\u00a0h vs. 51.6\u00a0h).Guo et\u00a0al. investigated the APR of glycerol on different Pt-based bimetallic catalysts [40]. The alloys were pre-synthesized and successively loaded on the support (\u03b3-Al2O3) to exclude the influence of the particle size. Among the promoters (Ni, Co, Fe, Cu), Fe showed the highest carbon conversion to gas and hydrogen yield, with 1:1 considered as optimal atomic ratio. Water gas shift tests showed consistent results with the APR tests. The coherent results between APR and WGS confirmed that the latter has a strong influence on the overall result. To further investigate this aspect, in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) was utilized. It was showed that formate desorption is not easy on Pt/alumina, while it was easier for the Pt\u2013Fe system. It was suggested that Fe promoted the water activation producing OH groups that can react with the proximate CO adsorbate on Pt. This is the reason why 1:1 atomic ratio led to the best results, because it is the one which presented the highest proximity between the two metals. Moreover, it favored the formate decomposition that causes otherwise the sites blocking. Further works on Pt\u2013Fe bimetallic systems were reported in Refs. [41,42]. Moreover, further DRIFTS spectra showed new CO adsorption sites on Fe in Ref.\u00a0[43].Dosso et\u00a0al. studied bimetallic Pt based catalysts promoted with Ni or Co for APR of polyols derived from glucose degradation [44]. Pt and Ni are completely miscible, and this occurrence affected the electronic properties of Pt reducing the binding of adsorbate. This is in agreement with the results reported in Ref.\u00a0[34]. X-ray photoelectron spectroscopy (XPS) showed that Pt was mainly in the oxidized state (Fig.\u00a07\n-A), likely because of the donation of electrons. The interaction of Pt with Co and Ni was suggested by the presence of a broad peak of reduction in the H2-TPR, as depicted in Fig.\u00a07-B. Pt\u2013Ni showed higher yield than Pt\u2013Co. No direct comparison with the monometallic could be performed as the authors used also different preparation techniques, namely urea matrix combustion method (UCM) and incipient wetness impregnation (IWI). Being Ni not active for water gas shift, a high selectivity to CO was observed (almost 20-fold the one of Pt\u2013Co). The authors finally highlighted the higher presence of coke on Pt\u2013Ni compared to Pt\u2013Co, not directly explained by its limited higher activity.Pendem et\u00a0al. studied the effect of potassium promotion on Pt/Hydrotalcite catalysts for APR of glycerol [45]. Despite potassium addition decreased the surface area, it contributed to the increase of the surface basicity (measured by CO2-TPD) and stabilizing the Pt precursor, increasing the final dispersion of the active site. Adding 2.8% of K increased the glycerol conversion from 27 to 88%. Further addition up to 16.9% K decreased the activity likely for the covering of Pt particles (67% conversion). However, increasing the loading up to 28% changed the morphology improving further the dispersion and allowing to reach 83% conversion. Moreover, the increase of basicity caused an increase in the H2 selectivity. Finally, the catalyst showed stable performance after four runs, without appreciable leaching of both Pt and K. The effect of a second metal in modifying the acid-base properties of the catalyst was also reported in Ref.\u00a0[46] for Mn. The results were comparable with [47] in the case of ethylene glycol APR, despite no leaching was observed in this case.Another way to improve the APR performance was reported adding Ru to Pt-based catalysts on carbon supports prepared by impregnation [48]. In the APR of wheat straw hydrolysate, Pt\u2013Ru/Multi walled carbon nanotubes (MWCNT) showed the highest activity and hydrogen selectivity. It was attributed to the electron donor effect of Ru that increased the Pt electron density, filling its d-band and reducing the strength of adsorption of organics. The proximity of Ru and Pt clusters was the key to enhance the catalysis. In fact, the same did not happen using a different support, activated carbon. In this case, the catalyst was constituted by unpaired Pt and Ru atoms, which consequently were unable to express their synergy.Despite not performed in the context of APR, it is interesting to cite here that the addition of Ru improved also the sulfur tolerance of Pt monometallic [49]. The authors exploited the sulfur spillover concept, due to the higher resistance of ruthenium, and the hydrogen spillover that can regenerate the catalyst. Extended X-ray absorption fine structure (EXAFS) determination of the coordination number showed the presence of an alloy at the three investigated ratios (Fig.\u00a08\n - step 1). Sulfur species may interact with platinum (step 2) and then move to Ru by spillover (step 3), or directly interact with the latter (step 4). The authors reported that sulfur poisoning caused dealloying; afterwards, hydrogen spillover (step 5) could allow the removal of sulfur. Indeed, the regenerated sample showed similar spectra than the fresh one, through a re-alloying mechanism. In this sense, the atomic balance between the two metals is fundamental to favor both the sulfur-trap capacity (given by Ru) and the spillover at proximity.The last example reported in the paragraph regards the use of Zn as promoter. Lei et\u00a0al. used atomic layer deposition to modify Pt/alumina catalysts by ZnO promotion [50]. A higher hydrogen production was obtained when Pt was deposited before than Zn. This is because it took advantage of the two interfaces Pt\u2013ZnO and Pt\u2013Al2O3. Thanks to X-ray absorption near edge structure (XANES) spectroscopy, the authors suggested that the significant charge transfer from Zn to Pt promoted the H2 selectivity. Moreover, the bimetallic system showed only limited sintering (the particle size increased from 0.9\u00a0nm to 1.2\u00a0nm), contrarily to the monometallic catalyst (from 1.0\u00a0nm to 2.4\u00a0nm).\nTable 1\n summarizes the effects of the cited promoters on the performance of APR when added to Pt-based catalysts. It can be observed that the positive impact is commonly associated to the promotion of water gas shift, which can be obtained by different paths (increase of surface basicity, water activation, easier intermediates desorption). One typical drawback is that the increase of the conversion is (partially) counterbalanced by the decrease of the hydrogen selectivity, since methanation is promoted as well. It is important to observe that in some cases different results were obtained, despite similar catalytic systems and analogous substrates were used, such in the case of Pt\u2013Rh on gamma alumina ([37,39]), or for Pt-Mn ([46,47]). On the other hand, similar results were obtained also with different substrates such as ethylene glycol or glycerol for Pt-Fe ([34,40]), or three different polyols in the case of Pt-Co ([44]).Dietrich et\u00a0al. studied a Pt\u2013Mo bimetallic catalyst supported on carbon prepared by IWI [51]. Transmission electron microscopy (TEM) showed that the bimetallic nanoparticles sintered during glycerol reforming, with the average size moving from 2\u00a0nm to 5.1\u00a0nm. X-ray absorption spectroscopy (XAS) showed that the Pt\u2013Mo nanoparticles are metallic Pt-rich with the presence of Mo in different states (metallic when close to Pt, oxide when in isolated clusters, as reported in Fig.\u00a09\n-A). Even more interesting, operando XAS was carried out to evaluate if the state of the catalyst changes during the reaction. It was reported that Pt remained in the metallic state and its coordination number increased, coherently with the results reported by TEM. Overall, the basic structure remained the same, with a Pt-rich core and Mo in the near surface region of the nanoparticle. This is coherent with DFT calculations which predicts an analogous core-shell composition when starting from metallic Pt and MoO precursors.The comparison with monometallics and the extensive study on the role of Mo using operando spectroscopy and DFT was performed in a following work [52]. The authors reported that Pt was the real active site, while Mo acted as a promoter modifying the electronic properties of Pt and its interaction with the adsorbed molecules. DFT calculations showed that Mo decreased the Pt\u2013CO binding energy; as a matter of fact, XANES showed that CO was the most abundant surface species in the case of Pt/C (about 0.6 fractional coverage), while it was under the detection limit for Pt\u2013Mo/C. At the same time, it reduced the kinetic barrier for dehydrogenation and even more for C\u2013O bond cleavage, leading globally to a reduction of the H2 selectivity (Fig.\u00a09-B). Further DFT investigations in agreement with the ones reported here can be found in Ref.\u00a0[53], where the role of Mo oxide is highlighted regarding the change of WGS mechanism and the bifunctionality of the catalytic system is suggested.The promotion of Co was studied on Pt supported on MWCNT by IWI [54]. Co raised by 4.6 times the glycerol activity normalized per Pt surface. The improvement was ascribed to the WGS promotion, so by removing the strong CO adsorbed on the Pt site facilitating the water activation. The hydrogen selectivity slightly increased compared to the monometallic Pt, in the whole range of glycerol conversion (Fig.\u00a010\n-A). Pt\u2013Co and Pt showed an increase of the particle size from 1\u00a0nm to 2\u00a0nm (more than one-week test), indicating a good stability of the catalytic system. Pt\u2013Co particles were found in three different configurations: Pt clusters (59%), Co core \u2013 Pt shell (30%) and Pt\u2013Co alloy (11%). Therefore, despite only 40% of the particles was in the bimetallic form, still the performance improved. The high presence of Pt on the surface, contrarily to the case of the Pt\u2013Mo where it was in the core, may explain the maintenance of high hydrogen selectivity. XANES outcome reported a downward shift in the d-band center compared to the monometallic Pt, which is related to a lower binding energy of CO.The same research group further evaluated the effect of Co addition to Pt-based catalysts performing operando X-ray absorption spectroscopy (XAS) investigation [55]. Adding Co increased the site time yield up to four-fold when the catalyst composition was 1Pt5Co (i.e., 1:5\u00a0Pt-to-Co molar ratio). Please note that monometallic Co showed negligible activity. The comparison between fresh and operando structures of the catalyst exhibited modification because of the hydrothermal environment, despite preserving the bimetallic configuration. The formation of an alloy was considered the reason for the improvement in the performance, rather than the nature of the particles (e.g. Pt shell/Co core). Being maintained the selectivity, the authors suggested that the role of Co was to exalt the Pt properties, at the same time reducing the CO binding energy and improving the WGS.Previously Wang et\u00a0al. investigated by XAS Pt\u2013Co nano-alloyed particle systems supported on single wall carbon nanotubes (SWCNT) [56]. XAS spectra showed a core (Co) and shell (Pt\u2013Co) structure, which greatly enhanced the activity without decreasing the selectivity. The electron transfer from Co to Pt may perturbate the latter, probably affecting the reaction path.Co promotion on Pt based catalysts was also studied on CeO2\u2013ZrO2 mixed oxides supports for ethylene glycol APR [57]. As reported in Fig.\u00a010-B, the optimal Co:Pt molar ratio in terms of APR activity and WGS (CO conversion) was found equal to 0.5. The authors suggested that the oxophilic properties of cobalt favored the WGS and penalized the formation of coke (likely hindering acetic acid intermediate formation [58]). Further increase of Co loading decreased the ethylene glycol conversion. The interaction between the two metals was investigated via TPR. Adding the promoter, the strong metal-support interaction Pt\u2013CeO2ZrO2 decreased. Other phenomena such as the formation of surface defects and higher dispersion of the bimetallic can contribute to the higher activity. As a matter of fact, the activity trend followed the results of the CO chemisorption. The promoter helped also to stabilize the platinum particles, which were less aggregated in the bimetallic catalyst compared than the monometallic.Platinum and nickel are likely the most studied active metals for APR, since they can be defined as representatives of two classes. Platinum is the most effective monometallic catalyst, but it suffers from a high cost; on the other hand, nickel is cheaper, and despite its performance are less enthusiastic than platinum, its use would increase the cost-effectiveness of the process. Ni/C exhibited much lower activity than Pt/C, particularly at low WHSV, and deactivated rapidly because of Ni leaching during APR of xylitol [35]. Even for Ni-based systems, as will be shown more extensively later, the addition of a promoter improves the performance. For example, it was cited that the presence of Sn in a Raney Ni catalyst led to a hydrogen production per unit volume equal to 350\u00a0\u03bcmol H2/cm3, comparable to the value obtained by Pt/Al2O3 (450\u00a0\u03bcmol H2/cm3) [32]. For this reason, several researchers investigated bimetallic systems which involved these two metals.He et\u00a0al. studied the optimal ratio between Pt and Ni on MWCNT prepared by IWI for glycerol APR [59]. Adding the second metal allowed to increase the dispersion of Pt. The formation of Pt\u2013Ni alloy modified the electronic properties of Pt, increasing the interaction with the support and thus the dispersion. Adding Ni also caused the highest increase of glycerol conversion (from 26% to 81%), while the carbon conversion approximately increased from 8% to 15% (Fig.\u00a011\n-A). The hydrogen yield increased from 1.83\u00a0mmol H2/g glycerol for the monometallic to a maximum 2.43\u00a0mmol H2/g glycerol, together with a strong increase of methane (approximately 3 folds higher). This is coherent with the known methanation activity of Ni. Looking at the effect of Ni loading, it was reported that 1:1 was the optimal atomic ratio, since excessive Ni may not form an alloy with Pt and rather cover it.Different Pt:Ni ratios (1% Pt and 3\u201318% Ni) on ceria doped alumina catalysts were also investigated in Ref.\u00a0[60]. Two different types of surface alloys, NiPt and Ni3Pt, were identified. Adding Ni up to 6% reduced the crystallite size causing the highest activity (96% conversion) and H2 selectivity (approximately 83%), as depicted in Fig.\u00a011-B. The shift in the Pt 4f7/2 XPS peaks highlighted the modification of the electronic environment by Ni addition. The synergy between the two metals is proved by the fact that physical mixture of the two catalysts did not perform similarly well. The improvement compared to monometallic Pt was referred to the improvement of WGS thanks to Ni addition, higher dispersion, and lower H2 and CO binding energy on Pt, making the sites more available. Looking at the catalyst stability, the authors showed that monometallic Ni severely deactivated (metal oxidation, carbon deposition and leaching), while 1Pt6Ni was stable on 85\u00a0h TOS (small particles aggregation but not carbon deposition was observed).Possible structural modification of bimetallic Pt\u2013Ni catalysts under APR conditions were followed thanks to in situ EXAFS analysis [61]. Fig.\u00a011-C shows the modification subjected by the Pt\u2013Ni clusters under APR. When the catalyst is reduced, the core is Ni-rich, while the shell is Pt-rich. However, under APR conditions, the Ni\u2013Pt particles restructures, with platinum diffusing to the core while Ni segregates to the surface. Importantly, this behavior was reversible prior subsequent re-reduction. The enhanced activity of the bimetallic catalyst compared to monometallic Pt was explained by this segregation. This is coherent with DFT studies which showed higher activity of Ni-terminated (i.e., Ni-rich surface) Pt\u2013Ni nanoparticles due to the increase of oxygen binding energy which boosts the initial dehydrogenation rate [62].Pt\u2013Re/C was initially studied for the gas-phase glycerol reforming, where it was noted that the H2 turn over frequency (TOF) increased by one order of magnitude compared to Pt/C [63]. In a simplified kinetic model, the authors observed that the CO adsorption equilibrium constant was 10 times lower for the bimetallic catalyst compared with the monometallic. Kunkes et\u00a0al. carried out microcalorimetric measurements of CO adsorption and CO-TPD studies [64]. Pt/C catalyst reported 115\u00a0kJ/mol as heat of CO adsorption, while for Pt\u2013Re/C catalyst it approached the one of pure Re (105\u00a0kJ/mol). Moreover, the partial oxidation of Re sites under the APR condition may form sites with lower binding energies. Avoiding that the sites are blocked by CO (or other intermediates) is of paramount importance because C\u2013C bond breaking mainly happens with transition states multiply bonded to Pt sites, therefore the latter must be preferably free [21]. Thanks to attenuated total reflection-infrared (ATR-IR) in situ analysis, it was reported that CO desorption from catalyst surface was more facile on the bimetallic catalyst compared to the monometallic [65]. ReOx can interact with CO adsorbed on Pt, facilitating its desorption (CO spillover), contributing in this way to the increase of activity. Another reason for the higher activity can be related to the increased C\u2013C and C\u2013O bond breaking capacity, with the latter to a major extent. It means that the Re addition decreased the hydrogen selectivity because it favored hydrogenation reactions of dehydrated intermediates. This is because Re increased the acidity (Re-OH, Br\u00f8nsted type) of the catalytic system.King et\u00a0al. tested Pt\u2013Re for the first time under APR conditions [66]. Re addition led to an increase of glycerol and H2 TOF. Re alone did not show any activity, so the effect was just as promoter. However, it decreased the H2 selectivity. As previously reported, the promotion was induced by multiple reasons: on one side, Pt\u2013Re alloy can have higher activity by modification of the electronic properties that can affect the CO adsorption strength. Moreover, Re can favor water gas shift by water activation, which is limited on platinum [67]; in this way, higher OH surface coverage may be obtained, leading to a higher formation of COOH, which is a key intermediate for the reaction. This outcome was proved by CO stripping voltammetry in Ref.\u00a0[68], where the authors observed that the Pt\u2013Re had a lower onset potential of CO oxidation than Pt one, and it was ascribed to a higher concentration of OH species (stronger binding of oxygen species which promoted water activation). Finally, the acidic properties could explain the decrease in H2 selectivity. KOH addition showed further increase in liquid products, so that it can be assumed that base addition can compensate the surface acidity provided by ReOH.The correlation between the surface properties of the bimetallic catalyst and the product distribution was further investigated in Ref.\u00a0[69]. XPS showed that Pt binding energy (BE) shifted, and it was higher with higher Re loading. The increase of the BE (electron deficiency) was due to the interaction with Re (possibly alloy formation). However, Re BE increased as well, ascribed to metal-support interaction. In order to simulate the interaction with water, the catalyst was treated with steam, leading to a change of oxidation in both platinum and rhenium. In particular, the presence of new Re-OH species increased the surface acidity, which in turn affected the product distribution favoring dehydration products via acid-catalyzed reactions. Interestingly, it was reported that the hydrogen selectivity increased with the conversion: this is due to the more difficult dehydration of smaller molecules compared with glycerol. Similar conclusions were reported in Ref.\u00a0[70] and in the case of xylitol [71].In conclusion, it is worthy to highlight some of the outcome reported for these characteristic bimetallic Pt-based catalysts. Table 1 resumes the effects of Mo, Co, Ni and Re addition to Pt, as reported in the last sub-paragraphs. Commonly the promoters induced an electronic perturbation of platinum, which decreased the interaction with the adsorbates and increased the availability of free active sites. Apart from Co, each of the promoters caused a strong increase of the conversion but associated with a decrease of the H2 selectivity. In fact, often the C\u2013O cleavage pathway is favored: for example, due to the enrichment of surface acidity in the case of Re, or due to methanation in the case of Ni. It is important to observe that tuning the reaction conditions, like adding a base, could overcome the reported drawbacks.In view of a possible industrial application of APR, the use of non-noble metals in the catalyst formulation may be a milestone for the success. For this reason, the use of transition metals-based catalysts, such as Ni, has been explored.Rahman investigated bimetallic Ni-based catalysts for APR of glycerol over multiwalled carbon nanotubes [72]. In the 1\u00a0Pt-xNi series, 1\u00a0Pt\u20133Ni was found the one with the highest activity (glycerol conversion higher than 99%) and hydrogen yield (Fig.\u00a012\n-A). Monometallic Ni was confirmed active towards methanation, reporting the highest yield for methane and the lowest glycerol conversion (44%). As depicted in Fig.\u00a012-B, Ni influenced the electronic state of Pt (shift of the binding energy). While Ni monometallic suffered from deactivation after 65\u00a0h TOS, Cu and Pt addition allowed to maintain a stable hydrogen production in 100\u00a0h long runs.The influence of several promoters (Mg, Cu, Zn, Sn, Mn) on the APR of ethylene glycol was recently studied over Ni\u2013Al hydrotalcites [73]. The screening reported that the conversion was not affected by the incorporated metal. On the other hand, the selectivity was improved in the case of Mg promoter. The authors explained this outcome because of the change in the electronegativity of Ni. Interacting with MgAlO, Ni became more electronegative, reducing the possibility that the feed may be dehydrated.Luo et\u00a0al. studied the Ni:Co ratio for APR of glycerol on alumina support [74]. 1:3 was found the optimal ratio to maximize the hydrogen yield, thanks to the tradeoff between higher selectivity thanks to Co addition and lower activity due to the decrease of superficial Ni. The synergy between the two metals, proved by H2-TPR, was confirmed by the fact that both monometallics had much lower hydrogen yield. Slight addition of Ce further increased the hydrogen yield promoting the formation of NiO instead of Ni aluminate, improving the dispersion, and decreasing methane selectivity. Moreover, it avoided Ni sintering and stabilized the alumina support. Nevertheless, coke remained one of the causes for catalyst deactivation.The effect of cobalt addition to a bimetallic Ni\u2013Fe catalyst was systematically studied thanks to Fourier-transform infrared spectroscopy (FTIR) characterization before and after APR of ethylene glycol [75]. The results are reported in Fig.\u00a013\n. Ni was found to be responsible for ethylene glycol activation (about 30% carbon conversion to gas for Ni alone vs 10% for Fe alone) thanks to the high activity of metallic Ni for C\u2013C bond breaking. The bimetallic Ni\u2013Fe further improved the conversion at 45%, also reaching higher hydrogen selectivity. Fe was present as Fe3O4 and was supposed to be involved in the WGS via a redox mechanism: Fe2+ was oxidized by water (producing hydrogen) to Fe3+; afterwards, the latter was reduced back by CO (producing carbon dioxide). The addition of Co strongly increased the conversion up to 95%, maintaining the hydrogen selectivity but at the same time cutting the alkane selectivity from 75% to less than 5%. It was suggested that Co increased the capacity to adsorb ethylene glycol at the same time decreasing the one for hydrogen. Finally, the authors highlighted the instability of the catalyst due to the re-oxidation of Fe and leaching of Ni.Further improvements of Ni catalyst were reported in Ref.\u00a0[76]. A Ni\u2013B amorphous alloy catalyst showed higher hydrogen production and stability than Raney Ni: the improvement was attributed to the boron oxides which, surrounding the hexagonal close-packed Ni, caused their stabilization and the higher resistance against sintering.Tuza et\u00a0al. studied Ni\u2013Cu catalysts supported on hydrotalcite-like compounds, with varying composition from 20% Ni monometallic to 20% Ni\u201320% Cu [77]. Ni monometallic showed the highest activity among the samples. The addition of copper increased the Ni dispersion and reducibility. The increase of hydrogen selectivity (250\u00a0\u00b0C) during 12\u00a0h TOS was ascribed to the lower particle size, since multiple Ni clusters are necessary for CO dissociation and methane production. At 270\u00a0\u00b0C a decrease of the hydrogen selectivity (defined by the authors as the molar fraction in the gas phase) with TOS was observed because of hydrogen-consuming reactions (acetol hydrogenation), with increase of CO2 selectivity.In a successive work, Manfro et\u00a0al. looked also at surface acidity properties [78]. They observed that bimetallic catalysts were more acidic than monometallic, explaining the higher acetol formation. The absence of re-oxidation contrarily to other case reported in literature was motivated by the stabilization of the support. The slight sintering that occurred did not cause deactivation in 6\u00a0h TOS.Ni\u2013Cu supported on MWCNT for glycerol APR was also run up to 110\u00a0h, in order to get information also on extended time on stream [79]. X-ray diffraction (XRD) spectra showed a shift in the Ni peak, indicating the higher interaction with Cu in the case of higher loading, with possibility of an alloy due to their complete miscibility. Moreover, the dispersion increased when Cu was added (further increase led to similar dispersion than the monometallic). At the same way, the 1Cu12Ni was the most stable against sintering. The catalytic tests showed superior performance of the bimetallic formulation, with 1Cu12Ni giving the highest hydrogen yield. It reported lower CO ascribed to the higher WGS activity; moreover, methanation activity of Ni decreased, attributed to the alloy formation and the reduction number of clusters with multiple Ni atoms which are necessary for CO hydrogenation (4-fold lower methane formation in the bimetallic catalyst compared to the monometallic one). Looking at the stability, 6Cu12Ni and 12Cu12Ni showed high sintering, while 1Cu12Ni did not, leading stable H2 yield along three consecutive 110\u00a0h TOS runs. Only minor sintering was observed, but no re-oxidation for Cu and Ni and leaching. WGS activity was assessed with a proper test.Park et\u00a0al. studied different Ni-based catalysts supported on LaAlO3 with different promoters (Cu, Co, Fe) and found that Ni\u2013Cu was the one with highest glycerol conversion and hydrogen selectivity [80]. It was the catalyst with the highest dispersion, while the monometallic Ni showed the worst dispersion. Furthermore, it showed less coke deposition (whisker type, less harmful because it grows on one side of the metal particle) than the monometallic (graphite type) and no sintering (contrarily to the monometallic). However, it was higher than the Ni\u2013Co formulation.\nTable 2\n summarizes the effects of the promoters on the performance of APR when added to Ni-based catalysts. Since the beginning of the APR research, Ni has been reported as one of the most interesting non-noble metals due to its high activity. The addition of the promoters like Sn, Mg, Co and Cu helped to increase the selectivity by different mechanisms (e.g. by blocking defect sites or decreasing the particle size). This outcome is a key improvement since the lack of selectivity due to the high tendency of methanation is a drawback of Ni-based catalysts. At the same way, most of the unsolved issues still regards the instability by re-oxidation, carbon deposition or leaching, which may prevent a long-term use.Apart from the catalytic systems previously reported, some others are worthy to be cited. In fact, despite they may be less frequently applied, their outcome may be interesting to stimulate further research. Therefore, in the following, formulations based on noble metals (Ir and Ru), bulk and metal-free catalysts will be reported.Liu et\u00a0al. tailored the properties of a Ir-ReOx/SiO2 catalyst with a noble metal (Ru, Pd, Pt) to favor either the APR of glycerol (C\u2013C breaking) or its hydrogenolysis (C\u2013O breaking) [81]. Pt showed higher conversion compared to Ru and Pd, but lower selectivity to acetol and propylene glycol (1,2-PrD), due to its higher selectivity towards APR. Being the particle size similar, it cannot be ascribed to structure-sensitivity features. Increasing the conversion led to an increase in the C3 products selectivity. This is coherent with previous literature since hydrogen selectivity decreased with the reaction time due to parasite consecutive hydrogenation reactions. The properties of the Pt\u2013Ir-Re system were investigated looking at the monometallic and combination of bimetallic (Fig.\u00a014\n). Ir provided the highest conversion among the monometallic and the highest selectivity to acetol. Pt\u2013Ir and Ir\u2013Re showed higher conversion than the monometallic, with the former providing more APR activity: this is suggested by the higher presence of hydrogen and propylene glycol, meaning that more hydrogen was produced. This outcome showed the synergy among Pt and Ir since the corresponding monometallic catalysts reported very low activity. The authors suggested that new species were formed, with XRD showing a shift in the Pt and Ir peaks indicating the formation of alloys. The physically mixed catalyst had similar conversion level to the Ir-ReOx/SiO2 but lower than the trimetallic, underlying the importance of the proximity between the components, which cannot be obtained through a simple mixing. Adding ReOx allowed to reduce the metal particle size. Finally, it was observed that low loading of Pt (and likely smaller particles) favored the C\u2013O hydrogenolysis.Espinosa-Moreno et\u00a0al. studied the APR of glycerol over Ir-based bimetallic catalysts with Cu and Ni [82]. The authors observed that Ir\u2013Cu led only to 0.9% of carbon conversion to gas, despite the high glycerol conversion (76.5% on La2O3 support). This result was not trivial from literature. In fact, Ir was initially suggested by Dumesic and coworkers as a plausible APR catalyst thanks to its high hydrogenolysis activity (referring to the work of Sinfelt [83]). The addition of Cu was not sufficient to overcome the WGS limitation and the possibility that CO poisoned it.The performance of Ru-based catalyst was improved as well thanks to the use of promoters, under multiple points of view. Its stability was improved thanks to the use of nitrogen, which avoided Ru sintering increasing the metal-support interaction, favoring at the same time the initial dispersion [84]. At the same time, higher glycerol conversion and hydrogen selectivity were obtained. Moreover, it exalted the basicity of the support. This feature helped to activate the water molecule, which is a necessary step for the WGS reaction: its facilitation allowed to improve the overall performance because the sites were less occupied by the CO.Novel unconventional types of catalysts can be finally reported, such as the use of cobalt aluminate spinel for the APR of glycerol [85] and metal-free catalysts [86]. Cobalt is not stable due to oxidation and leaching. However, the spinel formation with alumina may overcome these limitations. Thanks to a mutual protective effect, the spinel catalysts had surface area higher than cobalt oxide and alumina alone. Increasing the Co loading decreased the surface acidity (measured by NH3-TPD) because alumina, which is a Lewis acid, was substituted by cobalt oxide. Analogously, basic sites formation was favored by Co loading (measured by CO2-TPD). Reference tests with alumina and cobalt oxide alone gave respectively the highest and lowest selectivity for hydroxyacetone, which is acid-catalyzed; at the same time Co3O4 gave the highest H2 and CH4 selectivity, being methanation favored by basic sites. The spinel structure helped in reducing the methanation activity and the re-oxidation of Co (5% vs 10%), even in a reducing atmosphere. The results showed the predominance of C3 liquid products, highlighting the low activity for C\u2013C breaking of Co. The conversion decreased with time on stream also due to coke and sintering of large particles.Esteve-Adell et\u00a0al. used for the first time a metal-free catalyst to perform the APR of ethylene glycol, i.e. graphene, obtained by alginate pyrolysis [86]. In this context, the presence of defects is pivotal to activate reactions that commonly require metals to overcome kinetic limitations. Graphene allowed higher conversion compared to graphite-derived materials. One main point of concern was the low stability, that ended up in negligible activity at the third run. It was ascribed to the possibility that carboxylic acids by-products or hydrogen itself may react with the active sites (carbonyl groups and diketones). In particular, the author proposed that they consist of frustrated acid-base Lewis pairs, i.e., Lewis acid and bases close enough to interact with H2 but not enough to interact between each other, acting as dehydrogenation sites. IR spectra and thermogravimetric analysis showed the presence of adsorbate organics.Before moving on the third element chosen by the catalyst designer, the choice of the support, it is worthy to discuss one issue still related to the active metal category. It has been proven that aqueous phase reforming is a structure-sensitive reaction, i.e., its rate, normalized per exposed metal surface atom, changes with the particle size [87]. However, in the APR field, there is no apparent agreement if smaller of bigger particles are more benefic for hydrogen production. In the following, the main works analyzing this topic are reported. Despite in some cases TOF was not reported, valuable comments can be drawn.Lehnert and Claus reported for the first time that hydrogen selectivity increased with increasing particle size, without affecting the conversion (Fig.\u00a015\n-A) [88]. Increasing the particle size implies that the number of face atoms increases, while the number of edges and corner atoms decreases [89]. Therefore, it can be postulated that a greater extent of face atoms permits favorable adsorption of the oxygenates for the C\u2013C breaking.On the other hand, Wawrzetz et\u00a0al. reported different results. Despite the agreement on the fact that the conversion was slightly affected by the particle size, they reported higher hydrogen selectivity for smaller particles (Fig.\u00a015-B) [90]. This result was ascribed to the higher concentration of metal sites that hindered the dehydration pathway (i.e., the catalytic chemistry favored by the support). Having smaller particles means highly coordinatively unsaturated metal atoms and at the same time higher concentration of metal atoms at the support-metal boundary. Both aspects may affect the mechanism followed by the molecule and its fate (hydrogen or alkane production).The latter outcome was supported by the results reported in Ref.\u00a0[91], where small Pt particles favored C\u2013C breaking more than C\u2013O. Moreover, sintering happened during the first 60\u00a0min (from approximately 2\u00a0nm to less than 4\u00a0nm) and then it was constant up to 1440\u00a0min. Similar results were reported for Pt particles in Refs. [92\u201394], where quantum effects have been proposed to explain the change of TOF close to a critical diameter. Moreover, methanation is favored by bigger particles [36].Very recently, Vikla et\u00a0al. prepared different Pt/Sibunit catalysts not only to study Pt size effects, but also looking at the influence of its distribution (uniform or egg-shell) [95]. The hydrogen production rate, normalized per unit of surface Pt, linearly increased with the mean particle size (up to 10.7\u00a0nm), in accordance with [88]. Furthermore, an optimal size for the agglomerated particles was found (about 21\u00a0nm).A change in the activity was reported in Ref.\u00a0[96], where platinum nanoparticles with different particle size were synthesized for glycerol APR to obtain structure-activity relationship. Normalizing to the Pt surface area, it was showed that larger particles increased the conversion (the activity). Moreover, the product distribution was affected (the selectivity). It was reported that small particles (with higher concentration of edge sites \u2013 Pt(100)) favored the dehydrogenation, while larger particles (with higher concentration of facet sites \u2013 Pt(111)) favored the dehydration.A stability-related TOF was defined by Duarte et\u00a0al., who looked at the influence of Pt loading on alumina supported catalysts (0.3\u20132.77\u00a0wt%) by IWI [97]. The authors reported that the amount of coke decreased when the number of surface Pt increased; in other words, larger particles caused less coke deposition. Moreover, sintering was observed for all the catalysts, and that was proved to occur during the initial phase of the reaction, since no decay in the performance was observed. The same research group previously studied the effect of Pt loading on xylitol APR in Ref.\u00a0[98] and sorbitol [99].Higher H2 TOF for smaller particles were observed also for Ru-based catalysts [100]. The use of in-situ ATR allowed the authors to observe that on the catalyst with high loading (5\u00a0wt%) and large particles (4\u00a0nm) C\u2013O breaking bond was favored, leading to acetylide intermediates and methane as the final product. On the other hand, the catalyst with low loading (0.5\u00a0wt%) and small particles (1\u00a0nm) did not lead to any acetylide. The difference was attributed to the fact that acetylide needs two bonds with the metallic site, therefore it was more likely to happen on flat metallic surface. Analogous outcome can be found in Ref.\u00a0[101], where Ru catalysts were prepared at different loading to have particle size greater than 2\u00a0nm (representatives of high-coordination flat terraces) and smaller than 2\u00a0nm (representatives of low coordination step/edges). They observed that small Ru particles favored hydrogen selectivity, with CH4/H2 ratio less than one. At the same time, the C1/C2 products ratio is smaller, indicating less activity for breaking. This result is contrast with DFT studies, which foresee lower energy barriers on smaller particles rich in edges and steps. The discrepancy was ascribed to the possibility that more active smaller sites can be deactivated by CO blocking which is strongly adsorbed. To confirm this hypothesis, Fischer-Tropsch synthesis was carried out, which did not report alkane formation on small Ru particles because CO was not activated.Van Haasterecht et\u00a0al. studied the influence of Ni particle size on different supports (carbon nanofibers, alfa and gamma alumina, zirconia, SiC) for APR of ethylene glycol [102]. Narrower peaks in the XRD of spent catalysts, together with TEM images, confirmed that sintering phenomena occurred for each of the supported catalyst in this order: CNF\u00a0>\u00a0ZrO2\u00a0>\u00a0SiC > \u03b3-Al2O3\u00a0>\u00a0\u03b1-Al2O3. The difference was attributed to the different inter-particle distance (due to the different surface area) and the initial particle size. The authors reported that smaller particles grow faster and more than bigger particles; furthermore, they suggested that it was due to Ostwald ripening mechanism due to the high solubility of Ni and absence of the influence of the catalyst loading on the growth rate.In heterogeneous catalysis the function of the support is typically reported as a mean to increase the metal dispersion. However, its behavior may be active in determining the performance of the reaction. For example, the acid-base functionalities of the supports are well-known and exploited in many important industrial reactions, such as hydroisomerization. In the context of APR, the nature of the support has been reported as a mean for tuning the hydrogen production.The support can affect the quality of the hydrogenation sites [103] due to its electronegativity. In the case of APR, it has been reported that Pt/alumina is less hydrogenating because of the lower electronegativity of the support compared to Pt/amorphous silica alumina [104].Shabaker et\u00a0al. screened different supports for platinum-based catalysts [105]. Looking at the H2 TOF reported in Fig.\u00a016\n, the ranking at 225\u00a0\u00b0C was TiO2\u00a0>\u00a0black, carbon, Al2O3\u00a0>\u00a0SiO2\u2013Al2O3, ZrO2\u00a0>\u00a0CeO2, ZnO, SiO2, with Al2O3, ZrO2 and TiO2 showing the highest hydrogen selectivity. The reaction temperature significantly influenced the ranking, with Pt-black being the one with the second worst TOF at 210\u00a0\u00b0C. SiO2 and CeO2 were reported to dissolve under the hydrothermal conditions.As reported in the introduction, the support has a strong role in modifying the hydrogen selectivity because of its acid-base properties. Wen et\u00a0al. studied the effect of supports with different acidity with a Pt-based catalyst [106]. The authors showed that the scale of hydrogen yield was SAPO-11\u00a0<\u00a0AC\u00a0<\u00a0HUSY\u00a0<\u00a0SiO2\u00a0<\u00a0MgO\u00a0<\u00a0Al2O3, which is qualitatively coherent with the range of acidity (i.e., SAPO-11 and HUSY are zeolites, while MgO is a basic support). During 240\u00a0min TOS, each support was stable except for MgO and SAPO-11. Possible structure-sensitivity effects were excluded in Ref.\u00a0[38] by pre-synthesizing platinum colloids that were then loaded on the supports through different techniques. It was pointed out that the scale of basicity (measured via CO2-TPD) was analogous to the range of hydrogen yield, i.e., MgO\u00a0>\u00a0Al2O3\u00a0>\u00a0CeO2\u00a0>\u00a0TiO2\u00a0>\u00a0SiO2. The authors suggested that the support basicity polarized water, inducing its dissociation and facilitating the WGS step. Again, although MgO showed the best performance, it was not hydrothermally stable. Further screening of supports highlighted the importance of the basicity for Pt-based catalysts in Ref.\u00a0[107].Liu et\u00a0al. studied different supports for Pd-based catalysts in the APR of ethylene glycol [108]. Among NiO, Cr2O3, Al2O3, ZrO2, Fe2O3 and Fe3O4, the latter showed the best performance thanks to the promotion of the water gas shift, which is the rate determining step for Pd.Kim et\u00a0al. investigated the effect of support on Pt\u2013Re systems [109]. The authors reported that the activity increased in the order alumina\u00a0<\u00a0silica\u00a0<\u00a0activated carbon\u00a0<\u00a0CMK-3. The ordered mesoporous carbon was explained as the best thanks to the easier access of the active site for the reactants and escape of the products and higher dispersion. Similar outcome was obtained by CMK-9 in the case of Pt-Fe [41].Zirconia and boehmite supports were explored in the APR of hydroxyacetone [110]. The support alone (and zirconia notably) produced more coke than the platinum-supported catalyst (measured by CHN analysis and temperature programmed oxidation), indicating that the metal plays a role in preventing such deactivation via aldol condensation mechanism. Boehmite may have fewer coke thanks to its lower acidity. The in-situ ATR study allowed identifying the most present dimer between the two possible intermediates.It is interesting to observe also the influence that the support modification may have on the stability of the entire catalytic system. Stekrova et\u00a0al. studied the influence of different Ce, Zr and La oxide supports for the APR of methanol [111]. Nickel catalysts are often subjected to deactivation due to re-oxidation and sintering of the metal particles. ZrO2 and CeO2 thanks to their oxygen storage and mobility are useful supports for WGS. Furthermore, oxygen vacancies can be increased if doped with lower-valence metal, like La, also increasing the metal-support interaction and, consequently, the Ni dispersion. The authors showed that Ni did not re-oxidize during the reaction. However, atomic absorption spectroscopy showed the occurring of Ni leaching. Ni sintering also occurred during the reduction step. While it was the same for pure zirconia and mixed ceria-zirconia oxide, it was lower in the case of mixed Ce\u2013La, likely due to the strong interaction between Ni and La. Cerium carbonate was reported as the main cause for the deactivation of the WGS step, even if the activity was restored by thermal regeneration (300\u00a0\u00b0C, air flow). Looking at the performance of APR, they reported that the use of mixed oxides is beneficial for the activity. Surprisingly, pure CeO2 support showed the lowest WGS activity, despite its effectiveness in the gas-phase system. On the other hand, the highest WGS activity was reported by mixed 17Ce\u20135La\u2013Zr support, linked to its highest basicity.\nTable 3\n summarizes the main effects of the support on the performance of APR. The acid -base properties and the reducibility were mostly investigated since they are directly involved in the water gas shift step that, as reported throughout this work, is a key step to promote the hydrogen yield. Less attention has been put on the textural properties, despite it can be important to promote the selectivity. In fact, it can be favored not only playing with the nature of the active sites, but also avoiding that hydrogen, once produced, may contact other feed molecules/intermediated and hydrogenate them.Alumina is one of the most studied supports for APR. Different alumina supports for Pt catalysts were studied in the APR of ethylene glycol [113]. It was reported that the hydrogen yield decreased in the order alfa\u00a0>\u00a0delta\u00a0>\u00a0gamma thanks to the high dispersion obtained in the former. The authors reported that the absence of chlorine in the alfa sample (prepared with a higher temperature treatment compared with gamma) improved the dispersion because chloride facilitated the sintering. Moreover, it was suggested that Pt is more anchored on dry (e.g. alfa) alumina surface than on the hydroxylated one (gamma).Making an analogy with the active site, it has been reported that binary supports, i.e., mixed oxides, are able to increase the hydrogen production thanks to a synergy between the components. \u03b1-Al2O3 modified with CeO2 and ZrO2 improved by more than 50% the hydrogen yield compared to unmodified \u03b1-Al2O3 [91]. This is because CeO2 and ZrO2 participate in the water activation, therefore promoting the WGS reaction [114].Iriondo et\u00a0al. modified alumina-supported Ni-based catalysts with Mg, Ce, Zr or La for glycerol APR [115]. La addition caused the highest increase in glycerol conversion, followed by Ce, Zr and Mg which was worse than the un-modified alumina. The supports did not influence the selectivity. Neither affected the dispersion, as there was not a clear trend. The authors suggested that geometric effects were due to the presence of the promoters on the Ni surface, similarly to Sn on Ni. All the samples showed deactivation ascribed to re-oxidation of Ni, while sintering was not reported.The influence of two different supports (alumina and nickel aluminate) have been studied for the aqueous phase reforming of methanol [112]. Pt supported on NiAl2O4 showed higher dispersion than on alumina (80% vs. 70%). The use of the spinel increased the methanol conversion from 26.5% to 99.9% and the hydrogen yield from 23.3% to 95.7%. Please note that in this case, Ni was present in an oxide state, so it was not able to activate the methanation reaction. Being NiAl2O4 not able to convert methanol, a synergy between the active site and the support was supposed to explain the improvement of the performance, due to several reasons. In situ DRIFTS was used to detail the CO formation process for the first time, showing that it is achieved via dehydrogenation of methanol to methoxy and formaldehyde species, followed by the decomposition of the latter to CO. Interestingly, the pathway was different for the alumina-supported catalyst. Indeed, it had mediocre dehydrogenation activity (the formation of formaldehyde and CO was observed at a longer time) and WGS (peaks of CO were already present, while they were absent for the spinel-supported catalyst). In the spinel support, platinum was more reducible, therefore more active for dehydrogenation (first reason for the synergy); the motivation was ascribed to the oxygen vacancy present in NiAl2O4. Furthermore, they contributed to carry out the water gas shift activating water via a redox mechanism which is faster than the associative mechanism observed for alumina. Summarizing, as reported in Fig.\u00a017\n, the first step (dehydrogenation) is performed in the same way on both catalysts: playing Pt a vital role for dehydrogenation, its characteristics affected the performance, and they improved thanks to the fact that it was more reducible. In the second step, the efficacy of WGS was further prompted by the redox mechanism, rather than the associative one.Moreover, the catalyst was stable along 600\u00a0h of time on stream, with only 10% of loss in conversion (no information on the possible change of selectivity were reported).Ceria support showed higher dispersion, WGS activity and resistance to coke than alumina [116]. Furthermore it suppressed methanation [117] by poisoning the responsible active sites. This mechanism would be analogous with the one proposed for the Sn-modified Ni Raney [32]. The effect of CeO2 addition to alumina supports for Pt catalysts in APR of glycerol was studied for the first time in Ref.\u00a0[118]. 3% and 6% ceria doping increased the hydrogen and methane yield compared to unpromoted alumina, while 9% loading decreased the hydrogen yield. The catalysts showed similar hydrogen selectivity. The authors ascribed the improvements to the increase in the availability of metal surface area and reducibility of platinum precursor by adding ceria.Bastan et\u00a0al. looked at the APR of glycerol over Ni-based catalyst supported on mixed alumina/MgO oxide, searching for a structure-activity relationship for the different Al/Mg ratio [119]. As reported previously, Ni suffers from sintering. MgO can be used to stabilize Ni particles. Increasing the Mg content also increased the Ni dispersion, while decreasing its reducibility. The A2M1 support showed the highest conversion because of the higher presence of surface metallic Ni. The spent catalyst showed re-oxidation of Ni and no carbon deposition, while no information on sintering was reported. Despite of the re-oxidation, the study of the performance showed stable hydrogen yield along 24\u00a0h of time on stream.Pt/SiO2\u2013Al2O3 with different Si/Al molar ratio (range 0\u20131) was investigated to modify the surface acidity and influence the product distribution [120]. Maximum conversion and hydrogen production rate were reported at Si/Al ratio equals to 0.1. Increasing the ratio led to an increase of the methane selectivity and a decrease of ethane selectivity. The authors reported that increasing the Si loading led to an increase of the Br\u00f8nsted/Lewis acidity ratio (because Si\u2013OH species are present on the surface), as well as an increase of the weak acidic sites and a decrease of the strong (and total) acidic sites. This difference on the quality and quantity of acidic sites can alter the platinum dispersion as well (higher at higher Si/Al ratio). The authors correlated the conversion with the amount of strong acid sites and the hydrocarbon selectivity to the weak Br\u00f8nsted acid sites. However, contrarily to most of the works reported in literature, the authors assumed that higher alkane selectivity can be attributed to both C\u2013O and C\u2013C cleavage.Despite the promising use of alumina, its transition into boehmite is well known under APR conditions, as reported in Refs. [121,122]. The use of silica deposition has been explored to overcome this limitation [123]. During the structural modification into boehmite, the metal particles may be encapsulated or sinter due to the losing contact with the support. One strategy to prevent this occurrence is increasing the support hydrophobicity, preventing the Al centers attack. The authors used tetraethylorthosilicate (TEOS). The total acidity decreased (measured by NH3-TPD), due to the substantial decrease of Lewis acidity (measured by pyridine adsorption) and a small increase of Br\u00f8nsted acid sites. The silylation decreased the catalyst activity: the complete conversion was reached for pure alumina after 5\u00a0h, while the necessary time increased up to 7\u00a0h and 12\u00a0h with the increase of silylation time (0, 4, 8, 12\u00a0h). In Fig.\u00a018\n-A the modification of the hydrogen production rate is reported. The rise in acidity caused the increase in hydroxyacetone selectivity; the WGS remained effective (lack of CO). The boehmite formation was slowed down or prevented, while some minor sintering occurred. Globally, the support life-time increased from 12\u00a0h to 36\u00a0h, likely due to the removal of the protective layer.Similarly, Van Cleve et\u00a0al. used alkyl phosphonate coatings to improve alumina hydrothermal stability [124]. The authors investigated phosphonic acids with different tail lengths, from C1 to C18. The initial surface area decreased due to this pretreatment likely because of the blockage of smaller pores. While alumina became boehmite if untreated, the alkyl phosphonate slowed down the transition, and it was slower the longer the tail length (Fig.\u00a018-B). The reason may be ascribed to the hydrophobic properties or the density (the smaller the chain, the higher the loading) of phosphonate groups. The role of the tails was investigated after an oxidative treatment where the tails were removed. Since alumina remained stable, it was suggested that the key to the improvement was the interaction between the support and the head of the alkyl phosphonate. Nevertheless, the longer chain (C18) showed higher stability despite the lower P density. It may be since the high coverage of C1 led, on average, to lower coordination compared to C18, so that the support was less stabilized.The addition of cerium, yttrium and calcium was investigated to improve the redox capacity and basicity of zirconia-supported catalysts in methanol APR [125]. Base Ni/ZrO2 catalyst reported 48% of conversion and 40% hydrogen yield. Ce addition decreased the activity, in contrast with other works in the literature. Despite the improvement of Ni reducibility, Ce negatively influenced the quality of the surface basicity (decrease of weak basic sites, with the formation of intermediate strength ones). Therefore, it seemed that the basicity plays a more significant role than the redox properties. Ca and Y addition increased the total surface basicity equally, with the former acting mainly in the formation of weak basic centers, and the latter in the establishment of medium/strong basic centers. The Ca-doped catalyst showed better performance than the Y-doped one because strong basic sites can negatively affect the results by preventing the desorption of CO2. Finally, increasing the Ca loading from 4% to 14% worsened the activity, likely due to the decrease of Ni surface area (despite the higher amount of weak and intermediate basic sites).The addition of Ce is common on Zr support. We cite here its investigation on two examples, with Ni- and Pt-based catalysts. Bastan et\u00a0al. investigated the effect of the Ce\u2013Zr support composition for Ni-based catalysts prepared by co-precipitation for APR of glycerol [126]. XRD showed that the two oxides were present in a mixed form, not as separated clusters, and presented higher surface area than the single oxide support. Zr included in the Ce lattice lowered the crystallite size due to its lower ionic radius. The Ni reducibility changed in the mixed oxide, with the increasing of Zr loading leading to an increase of the reduction temperature peak, likely suggesting stronger metal-support interaction. Finally, the increase of Zr also allowed to increase the Ni dispersion. It is likely that this difference caused the increase of glycerol conversion for the mixed oxides, however the hydrogen selectivity decreased. The best catalyst (Ni/Ce0.3Zr0.7O2) was tested also for stability with 25\u00a0h TOS. No drop in the conversion was reported; moreover, commonly deactivation mechanisms such as Ni re-oxidation and sintering were excluded.The effect of cerium/zirconium molar ratio for Pt-based catalyst is studied in Ref.\u00a0[127]. The activity and hydrogen selectivity were maximized when the Ce/Ce\u00a0+\u00a0Zr molar ratio was 0.4. The lowest CO concentration in this case suggests also more effective WGS. The authors reported that this ratio led to a higher abundance of oxygen vacancy sites which, being correlated with the platinum dispersion, may be considered the cause for the higher performance of this formulation thanks to strong metal-support interactions. Raman analysis was carried out to measure the defect sites. Mixed oxides have larger lattice strain and higher oxygen vacancy concentration, with the one at 0.4 ratio having the highest. H2-TPR showed improved reducibility in the case of mixed oxides, ascribed to the higher oxygen mobility. Finally, the sample 0.4 showed also less sintering compared to Pt supported on only ceria or zirconia.Larimi et\u00a0al. investigated ternary Pt\u2013Ce\u2013Zr solid solutions prepared by controlled precipitation methods with different Ce/Zr ratio, looking at its influence on the oxidation state of Pt and its particle size (dispersion) [128]. The presence of a ternary solid solution was confirmed by XRD. When Ce/Zr ratio was equal to 1, the dispersion was the highest and it was the one with the highest hydrogen activity. Looking at the TOF, it was reported that face atoms (larger particles) are more active, while higher H2 selectivity was obtained with the smallest particles. XPS analysis showed the electronic interaction between Pt, Ce and Zr through the increase of Pt binding energies. Electron withdrawing from Pt should cause the decrease of the Pt\u2013CO binding, enhancing in this way the WGS. Furthermore, the catalyst was proven stable in the APR condition, without agglomeration of Pt.Finally, Harju and coworkers reported a mild effect of zirconia particle size on butanol conversion, while the product distribution and the stability was negatively affected using bigger particles due to the onset of internal mass transport limitations [129]. In the 250\u2013420\u00a0\u03bcm range, butyric acid selectivity increased, favoring the active metal (Rh) leaching and decreasing the hydrogen formation by favoring hydrogen-consuming reactions (e.g. hydrogenolysis).Several carbon supports (activated carbon, single and multi-walled carbon nanotubes, superdarco (i.e., methane and steam treated) carbon and graphene oxide) were studied for Pt-based materials [130]. Activated carbon reported the highest gas production among the supports while graphene oxide the highest hydrogen selectivity. Moreover, SWCNT performed better than MWCNT likely because of the larger pores that facilitate the reaction of large molecules present in the hydrolysate.Wang et\u00a0al. studied the influence of surface functional groups in MWCNT used as support for Pt based catalysts in ethylene glycol APR [131]. HNO3 functionalization increased the Pt dispersion but decreased the turnover frequency. The authors ascribed this outcome to the presence of oxygen containing groups, which can create a competitive adsorption between water and ethylene glycol, due to the increase in hydrophilicity. An annealing treatment which removed these groups allowed to restore the original TOF.Finally, it is worthy to cite the effect that the carbon support morphology can have on APR. Meryemoglu et\u00a0al. looked at three ranges of activated carbon: < 88\u00a0\u03bcm, 177-88\u00a0\u03bcm, 177\u2013250\u00a0\u03bcm [132]. Smaller particles had higher surface area and pore volume. Pt size was similar, so the difference can be ascribed only to the support. It was showed that activity increased with decreasing particle size, as well as for narrower size distribution. Kim et\u00a0al. investigated the importance of the configuration of the support looking at 3-D and 2-D ordered mesoporous carbon with hollow- and rod-type configuration [133]. The order of hydrogen production was different from the order of the metal dispersion, highlighting the importance of the structure of the support. It was noted that 3-d ordered mesoporous carbon (OMC) allowed higher resistance towards sintering for the Pt particles; furthermore, the mesoporosity of hollow-type framework configuration favored the transport of reactant and products to and from the active sites, respectively.The preparation of a catalyst involves different steps, and each of them can affect its final characteristics. First of all, the formation of the metal-support system is necessary. It may be carried out via impregnation, deposition, ion exchange, etc. During this stage, liquid/solid interfacial phenomena are important and can affect the behavior of the final catalyst. In the APR literature, most of the works deals with impregnation (wet or incipient wetness) and will not be reported in this paragraph, unless for the sake of comparison. Only limited information was reported on the procedures and methods themselves, since outside of the scope of the work. Then, the second stage involves (commonly) gas-solid reactions, firstly in the calcination (oxidation) step, and afterwards in the activation (hydrogenation, sulfidation) step. The modifications caused by this second stage strongly influence the structure of the catalyst as obtained after the first stage [134]. This is the reason why both of them are analyzed in the following.Alternative methods to the traditional impregnation can contribute to increase the reducibility of the metal, facilitate alloy formation or the interaction between two metals in case of bi-metallic systems [135,136]. However, most of the time they are developed to increase the dispersion, so that a high metallic surface area available for the reaction can boost the productivity. For example, it was reported that silica-supported platinum catalysts prepared via ionic exchange (IE) showed higher metal dispersion (60%) than the ones prepared via incipient wetness impregnation (20%) [91]. This result allowed to nearly double the hydrogen yield. Analogously, urea matrix combustion method (UCM) compared with IWI for the APR of different polyols allowed higher dispersion and, as a consequence, higher hydrogen yield [44].As far as the choice of the metal precursor is concerned, Lehnert and Claus evaluated the influence of different platinum precursors (amines, nitrates, sulfites) with IWI technique [88] They observed slight differences among the salts in terms of glycerol conversion and hydrogen selectivity, with tetrammine platinum (II)-nitrate giving the highest hydrogen rate of production.Lemus et\u00a0al. developed a method to synthesize stable Pt size-controlled nanoparticles [137]. The novelty consisted in synthesizing the metal nanoparticles in situ, in the presence of the support, therefore leading to the immobilization on it (contrarily to methods where the nanoparticles are first prepared and later deposited, ex situ). The authors used polyvinylpyrrolidone (PVP) and NaBH4 as capping and reducing agent, respectively. Bigger particles were obtained by ex situ method, as well as with a reference method without PVP. However, the addition of PVP only after the contact with activated carbon led to a higher dispersion, as no competition was present between the nanoparticles and the capping agent.Apart from the activity and selectivity, also the stability can be affected by the preparation method. El Doukkali et\u00a0al. compared the preparation of Pt and Ni catalysts on alumina via incipient wetness impregnation (IWI) and sol-gel method under basic conditions (SGB) [138]. Sol-gel created a material with higher surface area (375 vs. 140\u00a0m2/g) and pore volume (0.43 vs. 0.23\u00a0cm3/g). The basic agent leads to spherical cluster formation and a spongy material. The pore volume decreased in SGB after adding the metal, likely because they were incorporated in the pores network, while they were mainly in the outer surface in the case of IWI. SGB also increased the interaction of Ni with the support (leading to Ni aluminate) favoring the dispersion. The latter was confirmed by broader peaks in XRD. SGB method also stabilized Pt particles which otherwise sintered during the reduction step when prepared via IWI. The carbon conversion to gas increased with SGB catalysts. In a subsequent work, the authors added also the study of sol-gel preparation under acidic conditions (SGA) [139]. The differences in the preparation method are reported in Fig.\u00a019\n-A. SGA led to a material with fibrous and laminated morphology and high surface area, allowing an effective dispersion of Pt and Ni: the modification in the morphology facilitated adsorption of reactant and desorption of the products, whereas the higher dispersion increased the activity. The catalysts prepared by SGA had 50% higher surface area than IWI, but 50% lower than SGB. Looking at the performance, the activity was in the same trend (Fig.\u00a019-B). Finally, SGB led to catalysts more resistant to sintering compared to the ones prepared by SGA.The same research group reviewed the deactivation mechanisms for the catalysts subjected to APR in Ref.\u00a0[141]. No leaching was reported for Pt, Ni and PtNi catalysts both at 230 and 250\u00a0\u00b0C. The textural characterization showed that the decrease of surface area involved the materials prepared by sol-gel routes more than the ones by IWI. XRD showed higher transition to boehmite in the supports synthesized by sol-gel, likely due to their higher surface area that facilitates the incorporation of water molecule. Ni sintering occurred, and it was higher for those prepared by IWI. However, contrarily to literature, Pt did not sinter. XPS showed that monometallic Ni completely re-oxidized, while it was metallic only in the 11\u201325% in the bimetallic case (lower for the sol-gel because of the higher dispersion). Pt remained in the metallic form after the reaction. Finally, temperature programmed oxidation (TPO) showed the presence of carbonaceous species adsorbed on the catalysts.Roy et\u00a0al. compared sol-gel method and solution combustion synthesis (SCS) combined with wet impregnation method for Ni/CeO2 catalysts [142]. The first one showed higher carbon conversion to gas and hydrogen selectivity, ascribed to the growth of Ni nanoparticles prepared by SCS during the reaction. Analogously, the same research group looked at Ni/\u03b3-Al2O3 catalysts, revealing that in this case SCS samples outperformed the ones prepared by sol-gel in terms of activity (i.e., ethanol conversion), hydrogen selectivity and TOF [143,144]. The extensive structural and superficial characterization showed that it was not only due to the smaller particles obtained by SCS. In fact, SCS sintered in a lower extent, produced less coke and bulk spinel formation (promoting WGS) than SG.Other examples of better dispersion of the active phase in the case of sol-gel route compared to impregnation are reported in Ref.\u00a0[145].Following the preparation steps, two further phases are commonly involved, calcination and reduction.Callison and coworkers used a colloidal synthesis procedure for the preparation of platinum particles [96]. The synthesis procedure was optimized varying the concentration of NaBH4 (reducing agent), reduction time of Pt precursor and immobilization time on the support. Afterwards, the reduction temperature was varied from 25\u00a0\u00b0C to 90\u00a0\u00b0C to change the particle size. As expected, it was shown that it increased with the increase of reduction temperature.Morales-Marin et\u00a0al. used bulk nickel-aluminate catalysts reduced at different temperatures from 300 to 850\u00a0\u00b0C [146]. The authors observed that increasing the reduction temperature had a two-fold effect and the results of the catalytic tests are reported in Fig.\u00a020\n. On one side, the nickel dispersion increased because its migration from the spinel structure to the surface is favored. On the other hand, both surface acidity and (to a greater extent) basicity increased. Up to 450\u00a0\u00b0C, less than 5% conversion was reported due to the absence of the Ni0 active site. In this range, hydroxyacetone was the main product, deriving from dehydration reactions catalyzed by Lewis acid sites. As a matter of fact, this outcome is properly exploited in works where the aim is the hydrogenolysis of glycerol and the production of C3 products [147]. From 600 to 850\u00a0\u00b0C, the glycerol conversion and gas production gradually increased, with the maximum hydrogen yield at 850\u00a0\u00b0C. However, hydrogen selectivity partially reduced due to its consumption in parallel reactions. Nickel oxidation and sintering were identified as the leading cause for catalyst deactivation, while leaching and coke were excluded. Interestingly, sintering was observed mainly for the larger particles, in contrast with previous literature [102].El Doukkali et\u00a0al. evaluated the influence of calcination temperature (550\u2013750\u00a0\u00b0C) on the stability of alumina and Ni/Pt particles [148]. The Pt particle size was similar, independently from the calcination temperature. Moreover, in one case the active sites were incorporated during the SGB synthesis of alumina; in the second case, they were impregnated by IWI after sol-gel alumina preparation (SGI). Ni particles were bigger for SGI, but less sensitive to the calcination temperature, while sintering occurred with SGB, causing more deactivation. Bigger particles are more resistant to sintering (4-fold increase for SGB, two-fold for SGI) and re-oxidation, which are common deactivation causes for Ni. Increasing the calcination temperature decreased the activity but increased the stability.Irmak et\u00a0al. studied different preparation and reduction methods of platinum on activated carbon, alumina and titania by IWI [149]. Thermal treatments were carried out under hydrogen and nitrogen: in the latter case the activity improved thanks to the lower particle size. Afterwards, reducing the precursor chemically (using NaBH4) rather than thermally, further improved the catalyst activity, thanks to the fast reduction process when it is added dropwise to the solution.Novel techniques, even if not applied for APR yet, are worthy to be mentioned. Keshavarz et\u00a0al. used microemulsion systems to prepare Pt and Re based catalysts with controlled particle size for heptane reforming [150]. Two different microemulsions, neutral and acidic, were prepared. Re particle size was larger for neutral microemulsions, while Pt particle size was larger for acidic microemulsions, as well as for Pt\u2013Re. Interestingly, the nature of the microemulsions did not affect the final acidic properties of the supported catalyst.The influence of surfactants on the synthesis of platinum nanoparticles via microemulsion method [151] was studied as well. In this method, nanosized water droplets in which the metal salt is dissolved work as a nanosized reactor during the reduction while dispersed into a continuous oil phase. Four surfactants were used: Sodium 1,4-bis(2-ethylhexoxy)-1,4-dioxobutane-2-sulfonate (AOT), cetyltrimethylammoniumbromide (CTAB), poly(oxyethylene) sorbitanmonooleate (Tween80) and poly(ethylene glycol) p-(1,1,3,3-tetramethylbutane)-phenyl ether (TX-100). It was reported that ionic surfactants allowed producing smaller nanoparticles than non-ionic ones (AOT\u00a0<\u00a0CTAB\u00a0<\u00a0Tween80\u00a0<\u00a0T-X-100, i.e., anionic surfactant\u00a0<\u00a0cationic surfactant\u00a0<\u00a0nonionic surfactant) thanks to the influence of the different head group charge, as it affects the dynamic process of collision, nucleation and growth of the droplets. Other examples of microemulsion synthesis of nanoparticles can be found in Ref.\u00a0[152], where the authors prepared NiPt bimetallic nanoparticles for methane dry reforming.Finally, Roy et\u00a0al. reported the use of a radio-frequency plasma treatment to modify the surface of a Ni/alumina catalyst [153]. The plasma modification mainly influenced the metal-support interaction leading to a higher dispersion of the metal, leading to an increase of the catalytic activity.In Table 4\n the influence of preparation method reported in the chapter is summarized. Among the others, sol-gel methods seemed particularly interesting for their simplicity and possibility to increase the dispersion; as usual, the trade-off is reported being affected the stability of the catalyst, which is particularly sensitive when alumina support is used. Once the support-active metal system is formed, the conditions used in the calcination and reduction steps can still play a role in modifying the final catalyst properties due to its dynamic structure.In the present work we collected the efforts of several research groups whose aim is the design of an effective catalyst for APR. The scientific outcomes have been classified according to the effect of three main steps: choice of the preparation method, choice of the metal, choice of the support. It is important to observe that the reality is not as simple as reported, and the boundaries are much more flexible. For example, glycerol conversion rate decreases in the order TiO2\u00a0>\u00a0ZrO2\u00a0>\u00a0CeZrO2\u00a0>\u00a0CeO2 on Pt, but the ranking changes into TiO2\u00a0>\u00a0CeZrO2\u00a0>\u00a0ZrO2\u00a0>\u00a0CeO2 on Pt-Re [154]. It means that changing the metal, apparently also changed the effect of the support. Similarly, changing the support affected also the ranking between preparation methods, as reported in the works of Roy et\u00a0al., where CeO2 was more active prepared by sol-gel than SCS, but viceversa for Al2O3 [142\u2013144]. Notwithstanding, an ideal combination of these ingredients may be proposed at the end of the present work. The active elements should likely rely on bimetallic systems. Nowadays, platinum appears inevitable due to its peculiar characteristics. Despite of its cost, the key point refers to avoiding its deactivation, which seems an affordable task, at least with model compounds. From the available literature, it should be accompanied by a promoter to increase mainly its water gas shift activity. Among the others, Re and Fe seemed the most suitable. If the metals could explicate the WGS effectively, the support may just play the role of dispersive medium. For this reason, mesoporous carbon, thanks to its inert behavior with respect to the aqueous phase and controlled pore size, may be a suitable support. Finally, the coupling of in situ formation of nanoparticles and their activation by chemical reduction methods could be a preparation method able to guarantee high dispersion and stability of the active phase.Apart from the characteristics of the catalyst, the hydrogen yields strongly depend on the nature of the substrate, due to severe selectivity issues that arise with the increasing complexity of the molecule. In other words, we can expect high yield for small molecules (methanol, ethylene glycol, glycolic acid), while it will decrease for glycerol, xylitol or sorbitol. For example, we observed with glycolic acid 65% hydrogen yield with Pt/C while it was 38% for sorbitol, at iso-conversion conditions [11,12].However, looking also to an industrial application, the TOF values could be even more interesting. Lange reported that its value should be in the range 0.033\u201316.7 tons of product per ton of catalyst per min [155]. For Pt-based catalysts, for example, it could reach 3 min\u22121 [95,156] and this figure is promising for the future. However, it remains to be proofed also for more complex systems (see paragraph 6.2). For example, Pt\u2013Rh catalyst applied to the APR of pure glycerol showed 83.5% glycerol conversion and 89% hydrogen selectivity, while these values dropped respectively to 43.1% and 39% for crude glycerol [36].The future research should focus on each of the three cited topics to improve the performance of the catalytic system in the APR scenario. In the following, some points worthy of consideration are reported.In the field of the preparation method, innovative aspects such as the effect of the orientation of the active sites compared with unoriented Pt materials on graphene prepared by IWI, showed that the oriented material reported 2 order of magnitude higher catalytic activity expressed as TOF than unoriented ones [157]. Moreover, novel preparation techniques for bimetallic catalysts may be developed to handle the harsh conditions of the reduction treatment and reaction (for example, leading to stronger metal-support interactions) and stabilize the bimetallic clusters.Apart from experimental testing, the rational design of heterogeneous catalysts, thanks to the use of multiple tools such as DFT and micro-kinetic models, should benefit from further understanding of the reaction mechanism and lead to more effective catalysts. In this sense, the fact that APR occurs in liquid conditions is an obstacle, since the presence of a solvent that interacts with reagents, intermediates and products, modifies the energetic and reaction pathways [158]. The aqueous environment also affects the common knowledge in the reactivity of the catalytic system. It means that typical gas-phase WGS catalysts, such as Cu and CeO2, are not trivially effective catalysts also in the liquid phase.The use of a second metal to improve the activity, selectivity and stability of the primary active site has been proved to be commonly effective. However, despite the achievements, the complexity of the systems often requires further efforts. For example, it has been reported that the bimetallic structures modify under the reaction conditions: therefore, the development of in situ characterization techniques is necessary. Furthermore, new and cheap materials, such as tungsten, can be promising for future applications in the field [159].Much work needs to be done also on the support side, to clarify its role in the reactivity of the total system. For this reason, understanding phenomena such as charge transfer, spillover and perimeter activation may help in the design of new catalysts with tailored properties [160].Apart from the ones cited in this work, the catalyst will face new challenges when it goes towards the use of real wastewater streams. In fact, the complexity of the multi-components mixture may rise competitive adsorption phenomena [12]; moreover, the presence of inorganics or high molecular weight organics can lead to fast deactivation [161].Focusing on the latter, also the studies with model compounds, despite trying to assess the stability, are often referred to short runs, and so insufficient to probe the stability as demanded by chemical industry. Furthermore, studies on catalyst synthesis scalability and regeneration protocols should be developed.The literature cited herein applies the APR to simple model compounds, since its aim is the study and development of effective catalysts. However, the application of APR is devoted to the valorization of complex multicomponent mixtures, as it is the case of biorefinery wastewater streams. For this reason, in the last decade, the research started to investigate such feedstocks. Due to is versatility, APR could be applied to treat the water fractions derived from lignocellulosic biomass processing (e.g. not fermentable sugars post hydrolysis, aqueous phase from pyrolysis and hydrothermal liquefaction, etc.), aqueous effluents from food processing (e.g. breweries, cheese factories, etc.), crude glycerol from the biodiesel sector, and others [162]. Most of these works used simple catalytic systems (typically monometallic platinum catalysts), however they provide a range of hydrogen productivity into a more industrially relevant environment. For example, referring to the brewery wastewater, it was estimated that about 294\u00a0mL H2/g COD could be produced via APR, while anaerobic digestion could reach roughly half (150\u00a0mL H2/g COD) production [8]. Under the economic point of view, Larimi and coworkers showed that glycerol APR has lower production cost than glycerol steam reforming (3.55 vs 3.65 $/Kg), and this is competitive with other technologies which aim at a renewable hydrogen production (such as biomass gasification, dark fermentation, solar thermal electrolysis) [163,164]. Globally, the application of APR at industrial scale can be competitive if the cost of the feedstock is competitive as well. As a matter of fact, it can account for most of the production cost (e.g. up to 92% in the case of hydrogen from sorbitol syrup [165]).Aqueous phase reforming has been conceptualized as a strategic process for the valorization of biomass-derived compounds for hydrogen production. Since 2002, most of the literature focused on the pursuit of the optimal catalytic system that maximizes activity, selectivity and stability. Despite the efforts, the complexity of the reaction and the intercorrelation among the variables hindered, at the moment, the possibility to turn this process from the laboratory to the industrial scale. The aim of the present review was reporting, in a comprehensive way, the influence of several variables which can affect each of the three figures.Scope of the preparation method was mainly maximizing the dispersion to increase the number of available active sites. Alternative methods to the conventional impregnation techniques, such as ionic exchange, sol-gel and microemulsions reached this aim. Furthermore, they modified the electronic properties of the metal, for example via alloy formation or strong metal-support interaction, which in turn affected the reducibility, the tendency to CO binding or sintering.Theoretical investigation and first-principle methods, such as DFT, showed the intrinsic predisposition of metals to activate one or another pathway. Among the others, Pt showed higher tendency to C\u2013C cleavage than C\u2013O cleavage, maximizing the hydrogen production. The use of a promoter allowed to exalt or suppress some characteristic features of the monometallic catalytic form. Different promotion phenomena were reported. Ensemble (or geometric) effects were shown when Sn addition hindered the CO methanation on Ni-based systems; stabilizing effects have been attributed to Ru and Rh when protected Pt from coke deposition and sulfur poisoning, respectively; ligand (or electronic) effects were largely reported when the promoters decreased the interaction between carbon monoxide and Pt active site, favoring WGS (Re, Co, Fe, Mo). Overall, it has been widely documented that the second metal can tune the catalyst modifying the binding energy with reactants, intermediates or products, improving the reducibility of the first metal or its dispersion, changing the surface acid-base properties. Each of these modifications can have a different degree of importance, and it depends on the catalytic system as a whole. For example, it seemed that increasing the metal surface area is more important than the increase of (weak) basic properties, which in turn play a more important role than the metal reducibility. Trade-off is ubiquitous in the design of the optimal catalyst, as the example of the choice of the metal particles exemplifies. Despite results are not totally coherent, we can assume that the size of the particles mainly affects the selectivity, with the smaller ones favoring dehydrogenation and C\u2013C cleavage, while the bigger ones favoring dehydration and methanation; if the conversion is affected, this is higher for larger particles, which caused less coke deposition as well.Finally, the choice of the support mainly affected the dispersion thanks to its surface area and favored (or not) dehydration acid-catalyzed reactions. Basic character of the support was linked to the promotion of water gas shift and, in turn, to higher hydrogen production. The hydrothermal stability, such as the case of Al2O3 and MgO, is a severe issue that can be overcome properly modifying the surface composition and morphology.Despite several challenges remain to be tackled, we strongly believe that the developments in the field of catalysis through innovative preparation methods, rational design and in situ characterization techniques can pave the way to the synthesis of effective catalysts for aqueous phase reforming and sustainable hydrogen production.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The project leading to this research has received funding from the European Union\u2019s Horizon 2020 research and innovation program under grant agreement N 764675. Gianna Moscoso Thompson is gratefully acknowledged for the graphical abstract artwork.", "descript": "\n                  The aqueous phase reforming (APR) is a catalytic reaction able to produce hydrogen from oxygenated compounds. The catalytic system plays a pivotal role to permit high conversion of the substrate, high selectivity towards hydrogen, and stability in the view of an industrial application. These figures of merit depend on several strategies taken by the researchers to properly design the catalyst, like the preparation method, the choice of the active metal together with possible promoters, the type of the support and so on. The available literature reports several studies where these parameters are evaluated and discussed. In this review, they were critically examined with the aim of finding correlations between the properties of the catalyst and the activity, selectivity and stability for the APR of carbon-laden water fractions. Both theoretical and experimental works have been included in the literature survey. When available, studies with the use of in-situ techniques allowed to increase the understanding of the catalytic phenomena involved in the reaction. Great attention was also reported to recently published works, so that the review could present the most up-to-date developments in the field. The most important outcomes regarding each parameter have been highlighted; moreover, the synergy among each of them has been pointed out, together with the trade-off that the researcher has to deal with in the pursuit of the optimum catalyst.\n               "}
{"full_text": "\n\n\nData will be made available on request.\n\n\nData will be made available on request.Specific surface area (cm-1)Concentration of species (M)Catalyst layerDiffusion coefficient (cm s-1)Diffusion driving force (m-1)Voltage (V)Faraday constant (A s mol-1)Force term (kg m-2 s-2)Gas diffusion layerHeight (m)Identity vectorCurrent density (A m-2)Mass flux (kg m-2 s-1)Permeability (m2)Constant of inter-diffusion coefficient fitted by Fuller, Schettler and GittingsHenry law constant (kPa)Length (m)Molar mass (kg mol-1)PtIr loading (kg cm-2)Number of electron transfer in reactionPressure (Pa)The sum of consumed or produced species in reactionUniversal gas constant (J K-1 mol-1)Area (m2)Temperature (K)Velocity (m s-1)Diffusion volume (m3 mol-1)Width (m)Molar fractionTransfer coefficientThickness (m)PorosityOverpotential (V)Contact angle (\u00b0)Dynamic viscosity (Pa s)Density (kg m-3)Conductivity (S m-1), or interfacial tension (N m-1)Electric potential field (V)Mass fractionAnodeCathodeFlow channelSpecies indexSpecies indexLiquid phasemembraneribAmong the energy generation technologies from renewable sources to conventional combustion, fuel cells have attracted much attention due to their high efficiencies and low emissions. Polymer electrolyte membrane fuel cells (PEMFCs) are formed by polymer electrolyte membranes as proton conductors and Pt-based nanomaterials as anode and cathode catalysts. The combined advantages of low operating temperature, high power density, light weight and readiness for mass production, make PEMFC a promising candidate for various applications, such as vehicle propulsion, stationary electricity generation and portable power. However, hydrogen is the currently used fuel that is difficult to be stored, transported and distributed. As compared with hydrogen and other fuels (e.g., methanol and ethanol) for fuel cells, ammonia is a high energy density carbon-free fuel that is readily liquified under \u223c8 bar of pressure at ambient temperature [1].Direct ammonia fuel cells (DAFCs) utilize gaseous ammonia or its aqueous solution as a fuel which is oxidized to nitrogen and water within the anode catalyst layer (CL), coupled with a four-electron reduction of oxygen at the cathode. Comparatively, the ammonia oxidation reaction (AOR) kinetics is far inferior to the hydrogen oxidation reaction in PEMFCs. The AOR is considered to follow the Gerischer-Mauerer mechanism [2]. According to this mechanism, *NHx (x=2,1 and 0) adsorbates are created during the process of ammonia dehydrogenation. *NH and *NH2 adsorbates may dimerize and form N-N bonds and the products can be subsequently dehydrogenated to dinitrogen. However, nitrogen adatoms (*N) will not dimerize and are considered to form a surface poison and deactivate the Pt catalyst [3].The basic study on DAFCs with aqueous KOH as electrolyte and Pt as anode dates back to 1960s [4]. Afterwards, there were few reports on low-temperature DAFCs until Lan et\u00a0al. adopted an alkaline anion-exchange membrane as the electrolyte for DAFCs [5]. Cr-decorated Ni (CDN) and PtRu/C were used as the anode catalysts and MnO2 as the cathode catalyst. They found CDN showed better catalytic activity than PtRu/C. However, the peak power density of those fuel cells is usually below 10 mW cm-2, limited by the unsatisfactory stability of alkaline membranes at high temperature, the sluggish kinetics of anode ammonia electrooxidation, and deactivation of the Pt-based cathode catalyst by ammonia crossover from the anode [6]. Furthermore, Silva et\u00a0al. investigated the effect of different Pt: Au atomic ratios on catalytic activity in anode and found that PtAu/C 70: 30 presented better performance than Pt/C in DAFC while Au/C showed no activity [7]. A first-principles study by Herron et\u00a0al. predicted Pt is the most active monometallic catalyst followed by Ir and Cu, whereas other metals such as Au, Ag, Pd, Ni, Co, Rh, Ru, Os and Re have significantly lower activity [8]. Song et\u00a0al. studied the kinetics of Pt, Ir and PtIr alloy as electrocatalysts experimentally and found that Ir had a lower AOR onset potential, and lower peak current than Pt. Moreover, PtIr could combine the advantages of both metals and exhibit extraordinary catalytic performance among these metals [9]. Meanwhile, polymeric anion exchange membranes with high OH- conductivity, low ammonia permeability, mechanical robustness and chemical stability above 80 \u00baC under highly basic and chemically aggressive conditions are required to further improve the DAFC performance [10]. A series of poly(arylene piperidinium)s (PAPipQs) with the mentioned features have been available recently [11]. Based on a judicious choice of catalysts (PtIr anode and Fe-N-C cathode), PAPipQs-based membrane DAFCs demonstrated superior performance with a record peak power density of 135 mW cm-2 at 80 \u00baC [12]. Besides, Achrai et\u00a0al. reported a DAFC with a KOH-free anode feed. This DAFC used Pt1Ir10/C as anode and Ag-based dispersed electrocatalyst as cathode of which the peak power density hit record for this type of fuel cell and could reach 180 mW cm-2 at 120 \u00baC [13]. To further improve the cathode's catalytic performance, Hu et\u00a0al. synthesized Mn-Co spinel on three different carbon supports (BP2000, Vulcan XC-72R and multiwalled carbon nanotubes) which all showed good ammonia tolerance. Especially, compared with the Pt-C cathode, Mn-Co-BP2000 cathode paired with Pt-Ir anode could improve the peak power density of the cell to 128.2 mW cm-2 at 80 \u00baC under a 2 bar backpressure [14]. Jeerh et\u00a0al. studied the ORR activity of the LaCoO3-\u03b4 based perovskites and found that the co-doping of Cr and Fe elements can significantly improve the ORR performance of the catalysts. Moreover, the DAFC based on LaCr0.25Fe0.25Co0.25O3-\u03b4/C cathode achieved an open-circuit voltage (OCV) of 0.72 V and a maximum current density of \u223c320 mAcm-2, which showed almost the same performance as the DAFC with commercial Pt/C cathode [15]. Although the study of low-temperature DAFCs is still at an early stage, ammonia is one of the clean and inexpensive energy sources which makes DAFCs promising power devices for the carbon-neutral economy [16]. However, the challenges, such as minimizing the ammonia crossover, developing suitable anode and cathode catalysts, maintaining mechanical stability of the alkaline exchange membranes and achieving high power density at low temperature, still need to be dealt with [17]. To the best of our knowledge, there is no theoretical simulation of DAFCs to guide the optimization of the cell performance at the membrane electrode assembly (MEA) level.The simulation of DAFCs involves various complex processes, such as species transport, momentum transport, and electrochemical reaction. Especially, the liquid-phase ammonia fuel is oxidized to gas-phase nitrogen within the anode CL, where the phase transition is similar to that at the anode of direct methanol fuel cells (DMFCs). Until recently, the numerical model of DAFCs has not been presented yet, so this work is developed mostly on the base of DMFC models. In several early investigations into DMFCs, Baxter et\u00a0al. developed a one-dimensional single-phase model of DMFCs which considered kinetics of methanol oxidation and active specific surface area in the anode CL [18]. Then some other important factors, such as the methanol crossover [19], methanol concentration [20] and methanol mixed with air [21], began to be taken into account in other one-dimensional models. However, due to the limitation of the model dimension and the lack of considering two-phase flow in DMFCs, these models were not relatively accurate. In contrast, Wang et\u00a0al. developed a two-phase, multicomponent model for liquid-fed DMFCs which considered diffusion and convection of liquid and gas phases in backing layers and flow channels [22]. Ge and Liu focused on using the Tafel equation to describe electrochemical kinetics and the effects of two phases in the anode and cathode sides of a DMFC model [23]. Many researchers also showed an interest in the effects of crossover, operation conditions and intrinsic parameters on cell performance. Yang and Zhao developed a two-dimensional, two-phase mass transport model of DMFCs. In this model, the effects of porosity and anode flow rates on cell performance and methanol crossover were studied [24]. Liu and Wang developed a three-dimension isothermal mixture multiphase flow model to investigate the interplay between local current density and methanol crossover [25]. Biswas et\u00a0al. studied a DMFC anode model which could accurately predict the rate of methanol crossover affected by the geometric and material properties of anode layers [26]. Garc\u00eda-Salaberri et\u00a0al. developed a three-dimensional DMFC model which considered gas diffusion layer (GDL) compression and showed the optimal methanol concentration [27,28].These mentioned 2D/3D two-phase models accurately simulated multi-phase mixture properties and the effect of operating conditions or material properties on cell performance. The objective of this work is to develop a three-dimensional, two-phase multicomponent model for the anode CL based on the unique kinetics of DAFCs and the physical properties of involved species, and provide the insight into optimizing the anode CL structure by investigating the effect of its porosity, PtIr loading and thickness on cell performance at an MEA level. In our model, the Maxwell-Stefan model is applied to this DAFC anode in the porous region and flow channel, Darcy's law is used to describe the fluid dynamics in the porous region, and the Brinkman equation is applied in the flow channel [29]. The effects of CL material properties on cell performance and ammonia crossover are investigated.\nFig.\u00a01\n shows the schematic diagram of the DAFC anode, which consists of a flow channel, GDL, CL and an alkaline anion-exchange membrane. The GDL and CL are assumed to be isotropic porous regions. The ammonia solution flows along the flow channel and diffuses through the GDL to the CL where it will be oxidized to nitrogen according to the AOR (2NH3\u00a0+\u00a06 OH-\u00a0\u2192\u00a0N2\u00a0+\u00a06H2O\u00a0+\u00a06 e-). The nitrogen generated in the CL will aggregate and transport through the GDL to the flow channel. The ammonia liquid phase and nitrogen gas phase diffuse in opposite directions, which decelerate the mass transport of both phases and deteriorate the cell performance. All phases are assumed to be continuous. Nitrogen is assumed to be insoluble in liquid phase. The velocity of liquid-gas phase in the flow channel is assumed to be the same. The geometric parameters of the model are listed in Table\u00a01\n.Based on this model, several governing equations, including mass conservation equation, momentum conservation equation, and species conservation equation are used to describe the multiple processes such as mass transport and fluid dynamics.The Maxwell-Stefan model is used to describe the multicomponent flow. The model takes the collisions between different species, including ammonia, water and nitrogen into account and the corresponding equations are listed as Eqs. (1-3):\n\n(1)\n\n\n\u2207\n\u00b7\n\nj\ni\n\n+\n\u03c1\n\n(\nu\n\u00b7\n\u2207\n)\n\n\n\u03c9\ni\n\n=\n\nR\ni\n\n\n\n\n\n\n\n(2)\n\n\n\nj\ni\n\n=\n\u2212\n\u03c1\n\n\u03c9\ni\n\n\n\n\u03a3\n\nk\n\n\nD\n\ni\nk\n\n\ne\nf\nf\n\n\n\nd\nk\n\n\n\n\n\n\n\n(3)\n\n\n\nd\nk\n\n=\n\u2207\n\nx\nk\n\n+\n\n1\np\n\n\n[\n\n(\n\nx\nk\n\n\u2212\n\n\u03c9\nk\n\n)\n\n\u2207\np\n]\n\n\n\n\n\n\n\n(4)\n\n\n\nD\n\ni\nk\n\n\n=\nk\n\nT\n\n1.75\n\n\n\n1\n\np\n(\n\nv\ni\n\n\n1\n3\n\n\n\n+\n\nv\nk\n\n\n1\n3\n\n\n\n)\n\n\n\n\n(\n\n\n1\n\nM\ni\n\n\n+\n\n1\n\nM\nk\n\n\n\n)\n\n\n1\n2\n\n\n\n\n\n\n\n\n(5)\n\n\n\nD\n\ni\nk\n\n\ne\nf\nf\n\n\n=\n\nD\n\ni\nk\n\n\n\n\n[\n\n(\n1\n\u2212\ns\n)\n\n\u03b5\n\n\n]\n\n\n1.5\n\n\n\n\n\n\n\n\n(6)\n\n\n\nR\ni\n\n=\n\nM\ni\n\n\n\u2211\nm\n\n\nR\n\ni\n,\nm\n\n\n\u2212\n\n\u03c9\ni\n\n\n\u2211\ni\n\n\n\nM\ni\n\n\n\u2211\nm\n\n\nR\n\ni\n,\nm\n\n\n\n\n\n\n\n\n\n(7)\n\n\n\nR\ni\n\n=\n\n\n\n\u03bd\ni\n\n\ni\nv\n\n\n\nn\nF\n\n\n\n\n\n\n\n\n(8)\n\n\n\n\u03bd\n\n\nH\n2\n\nO\n\n\n=\n\u2212\n6\n,\n\n\n\u03bd\n\nN\n\nH\n3\n\n\n\n=\n2\n,\n\n\n\u03bd\n\nN\n2\n\n\n=\n\u2212\n1\n\n\n\nwhere ji\n is the flux of species i, \u03c9 and x refer to mass fraction and molar fraction separately and Dik\n is binary diffusivity between species i and k. The binary diffusivity is defined by Eq.\u00a0(4). However, in porous medium with two-phase flow, Dik\n should be corrected by porosity and saturation. Therefore, the effective diffusivity is given by Eq.\u00a0(5), as defined by Bruggeman correction. dk\n is the force exerted by species k. Ri\n in Eq.\u00a0(1) is coupled with electrochemical reaction in Eq.\u00a0(6) and given by Faraday's law in Eq.\u00a0(7). \u03bdi\n is specified by Eq.\u00a0(8). The Maxwell-Stefan model is applied for the entire flow in both flow channel and porous medium regions.Darcy's law is applied to both gas and liquid flows in porous medium. The law is given by Eqs. (9-13):\n\n(9)\n\n\n\nu\nl\n\n=\n\u2212\nK\n\n\nk\n\nr\nl\n\n\n\n\u03bc\nl\n\n\n\u2207\n\np\nl\n\n\n\n\n\n\n\n(10)\n\n\n\nu\ng\n\n=\n\u2212\nK\n\n\nk\n\nr\ng\n\n\n\n\u03bc\ng\n\n\n\u2207\n\np\ng\n\n\n\n\n\n\n\n(11)\n\n\n\u2207\n\u00b7\n\n(\n\n\u03c1\nl\n\n\nu\nl\n\n)\n\n=\n\nQ\nl\n\n,\n\n\u2207\n\u00b7\n\n(\n\n\u03c1\ng\n\n\nu\ng\n\n)\n\n=\n\nQ\ng\n\n\n\n\n\n\n\n(12)\n\n\n\nk\n\nr\nl\n\n\n=\n\ns\n3\n\n\n\n\n\n\n\n(13)\n\n\n\nk\n\nr\ng\n\n\n=\n\n\n(\n\n1\n\u2212\ns\n\n)\n\n3\n\n\n\n\nwhere K is the absolute permeability of the porous medium and kr\n is a function of saturation that represents the relative permeability of one phase. \u03bc and u denote the viscosity and velocity of gas or liquid phase separately. Q represents the sum of the consumed or produced species in the ammonia electrooxidation reaction.The two-phase property is defined by Eqs. (14-17):\n\n(14)\n\n\n\np\nc\n\n=\n\np\ng\n\n\u2212\n\np\nl\n\n=\n\u03c3\ncos\n\n\u03b8\nc\n\n\n\n(\n\n\n\u03b5\n\n/\nK\n\n)\n\n\n0.5\n\n\nJ\n\n(\ns\n)\n\n\n\n\n\n\n\n(15)\n\n\nJ\n\n(\ns\n)\n\n=\n1.417\ns\n\u2212\n2.120\n\ns\n2\n\n+\n1.263\n\ns\n3\n\n\n(\n\n90\n\u2218\n\n<\n\n\u03b8\nc\n\n<\n\n180\n\u2218\n\n)\n\n\n\n\n\n\n\n(16)\n\n\n\u03c1\n=\n\n\u03c1\nl\n\n\ns\nl\n\n+\n\n\u03c1\ng\n\n\ns\ng\n\n\n\n\n\n\n\n(17)\n\n\n\u03c1\nu\n=\n\n\u03c1\nl\n\n\ns\nl\n\n\nu\nl\n\n+\n\n\u03c1\ng\n\n\ns\ng\n\n\nu\ng\n\n\n\n\nwhere the relation of liquid and gas phase pressure is connected by capillary pressure pc\n in Eq.\u00a0(14). Furthermore, \u03c3 denotes interfacial tension, \u03b8c\n is the contact angle, J(s) lists the Leverette function. The average density and velocity of two phases defined by Eqs.\u00a0(16)-(17) are utilized in the Maxwell-Stefan model.The liquid saturation is mainly determined by the mass fraction of nitrogen while the vaporizing of both ammonia and water is also considered. The relation is shown by Eqs. (18-20):\n\n(18)\n\n\n\np\n\nN\n\nH\n3\n\n\n\n=\n\nk\nH\n\n\nx\n\nN\n\nH\n3\n\n,\nl\n\n\n\n\n\n\n\n\n(19)\n\n\n\n\u03c9\n\n\nN\n2\n\n,\ng\n\n\n=\n\n\n\nM\n\nN\n2\n\n\n\n(\np\n\u2212\n\np\n\n\nH\n2\n\nO\n,\ns\na\nt\n\n\n\u2212\n\np\n\nN\n\nH\n3\n\n\n\n)\n\n\n\n\n\u03c1\ng\n\nR\nT\n\n\n\n\n\n\n\n\n(20)\n\n\ns\n=\n\n\n\n\u03c1\ng\n\n\n(\n\n\u03c9\n\nN\n2\n\n\n\u2212\n\n\u03c9\n\n\nN\n2\n\n,\ng\n\n\n)\n\n\n\n\n\u03c1\nl\n\n\n(\n\n\u03c9\n\n\nN\n2\n\n,\nl\n\n\ns\na\nt\n\n\n\u2212\n\n\u03c9\n\nN\n2\n\n\n)\n\n+\n\n\u03c1\ng\n\n\n(\n\n\u03c9\n\nN\n2\n\n\n\u2212\n\n\u03c9\n\n\nN\n2\n\n,\ng\n\n\n)\n\n\n\n\n\n\nwhere kH\n represents the Henry's constant. \u03c9sat N2,l is saturated nitrogen mass fraction in liquid phase and is assumed to be 0.Electrochemical equations are used to describe the kinetics of the ammonia electrooxidation reaction in the anode. Since the anode reaction is much more sluggish than that in the cathode, the polarization only in the anode is taken into account for the potential loss of the whole cell. Butler-Volmer equation as shown in Eq.\u00a0(21) is chosen, which is coupled with the effect of mass transport.\n\n(21)\n\n\n\ni\n\nl\no\nc\n\n\n=\n\ni\n0\n\n\n[\n\nc\nR\n\ne\nx\np\n\n(\n\n\n\n\u03b1\na\n\nF\n\u03b7\n\n\nR\nT\n\n\n)\n\n\u2212\n\nc\nO\n\ne\nx\np\n\n(\n\n\n\u2212\n\n\u03b1\nc\n\nF\n\u03b7\n\n\nR\nT\n\n\n)\n\n]\n\n\n\n\n\n\n\n(22)\n\n\n\n\n\n\nc\nR\n\n=\n\n\n(\n\n\nc\n\nN\n\nH\n3\n\n\n\n/\n\nc\n\nN\n\nH\n3\n\n\n\nr\ne\nf\n\n\n\n)\n\n\u03b3\n\n\n\n\n\n\n\n\nc\n\nN\n\nH\n3\n\n\n\n>\n\nc\n\nN\n\nH\n3\n\n\n\nr\ne\nf\n\n\n,\n\u03b3\n=\n0\n;\n\nc\n\nN\n\nH\n3\n\n\n\n\u2264\n\nc\n\nN\n\nH\n3\n\n\n\nr\ne\nf\n\n\n,\n\u03b3\n=\n1\n\n\n\n\n\n\n\n\n\n(23)\n\n\n\ni\nv\n\n=\n\na\nv\n\n\ni\n\nl\no\nc\n\n\n\n\n\n\n\n\n(24)\n\n\ni\n=\n\n\n\u222b\n\n\u222b\n\n\u222b\n\n\ni\nv\n\nd\nv\n\n\n\n\nS\n\n\n\n\nwhere iloc\n is the electrode reaction current density, i0\n is the exchange current density. cR\n and cO\n are the mole concentration of reduction and oxidation separately, \u03b1a\n is the anode transfer coefficient, \u03b1c\n is the cathode transfer coefficient, av\n is the specific area, and iv\n is the current density. cR\n and cO\n are acquired by coupling with mass transport equations. The average current density in CL is defined by Eq.\u00a0(24).\nav\n is defined by Eq.\u00a0(25). mPtIr\n is the mass loading of PtIr, \u03b4CL\n the thickness of CL.\n\n(25)\n\n\n\na\nv\n\n=\n\n\n\nm\n\nP\nt\nI\nr\n\n\nE\nC\nS\nA\n\n\n\u03b4\n\nC\nL\n\n\n\n\n\n\n\n\n\u03b7 is defined by Eq.\u00a0(26).\n\n(26)\n\n\n\u03b7\n=\n\n\u0394\n\n\u03d5\n\u2212\n\nE\n\ne\nq\n\n\n\n\n\nwhere Eeq\n is the thermodynamic equilibrium potential.Besides, part of the polarization is caused by Ohmic resistance of each component in the anode and the solution. According to Ohm's law, the potential loss is given by Eq.\u00a0(27).\n\n(27)\n\n\n\ni\nk\n\n=\n\u2212\n\n\u03c3\nk\n\n\u2207\n\n\u03d5\nk\n\n\n\n\nwhere \u03c3 is the conductivity and k denotes the electrolyte or solid part of the anode.The Brinkman equation is used to describe the fluid dynamics in the flow channel, which is given by Eq.\u00a0(28) and Eq.\u00a0(29).\n\n(28)\n\n\n0\n=\n\u2207\n\u00b7\n\n[\n\u2212\np\nI\n+\n\u03bc\n\n(\n\u2207\nu\n+\n\n\n(\n\n\u2207\nu\n\n)\n\nT\n\n)\n\n]\n\n+\n\nF\nt\n\n\n\n\n\n\n\n(29)\n\n\n\u03c1\n\u2207\n\u00b7\n(\nu\n)\n=\n0\n\n\n\nwhere F\nt is the force term for the influence of gravity and other volume forces.In the membrane, not only hydroxide ions but water and ammonia can transfer to the other side. The ammonia crossover is the main cause of cathode voltage loss and catalyst deactivation. Based on the previous work of methanol crossover in DMFC models, the crossover mechanism can be explained by three parts: molecular diffusion, electroosmotic drag by proton and hydraulic transport. However, there is no proton transferring in the anion exchange membrane, so the effect of electroosmotic drag is neglected and the crossover is defined by Eqs.\u00a0(30)-(31):\n\n(30)\n\n\n\nN\n\nN\n\nH\n3\n\n\n\nx\no\nv\ne\nr\n\n\n=\n\u2212\n\nD\n\nN\n\nH\n3\n\n,\nm\ne\nm\n\n\n\u2207\n\nc\n\nN\n\nH\n3\n\n\n\n\u2212\n\n(\n\n\n\nK\n\nm\ne\nm\n\n\n\n\u03bc\nl\n\n\n\n\n\n\u0394\n\n\np\n\nl\n,\nc\n\u2212\na\n\n\n\n\n\u03b4\n\nm\ne\nm\n\n\n\n\n)\n\n\nc\n\nN\n\nH\n3\n\n\n\n\n\n\n\n\n\n(31)\n\n\n\nI\n\nx\no\nv\ne\nr\n\n\n=\n3\nF\n\nN\n\nN\n\nH\n3\n\n\n\nx\no\nv\ne\nr\n\n\n\n\n\n\nThe physical and chemical processes within the DAFC anode are computed by the several modules in COMSOL Multiphysics 5.4: (1) Second Current Distribution is applied in both the CL and the membrane. \u2018Porous Electrode\u2019 is applied to the CL. (2) Darcy's Law is applied for two phases in both the CL and GDL. The Brinkman equation is applied in the flow channel. (3) Transport of Concentrated Species and the Maxwell-Stefan model are applied in CL, diffusion layer and flow channel. Half of the DAFC anode model is built due to the symmetrical structure and the symmetry plane of the anode model is set as \u2018symmetry\u2019 in Comsol Multiphysics. Hexahedral mesh is applied for the entire model. A probe is set in the CL to integrate the current source and get the average current density.In COMSOL Multiphysics, two end planes of the flow channel are set as inlet and outlet separately. The potential of the GDL upper surface is set as the operating potential while the potential of the membrane bottom is set as zero. Other outside surfaces of the model are all set as \u2018no flux\u2019 and \u2018insulation\u2019, so no species (such as ammonia) or current can get through these surfaces.To verify the accuracy of this model, the calculated polarization curve of the model is compared with the experimental polarization curve of DAFCs. The procedure for assembling the DAFC and its testing were similar to our previous work [14]. In particular, the MEA was composed of carbon cloth (W0S1009, Ce Tech Co., Ltd) supported by 3.4 mgPGM cm-2 PtIr/C (40% PtIr on Vulcan XC-72R, Pt/Ir\u00a0=\u00a01:1, Premetek Co.), 15 \u00b5m alkaline polymer electrolyte membrane (AEM, Alkymer W-211415, EVE Institute of New Energy Technology) and carbon paper (28BC, SGL Carbon) supported by 1.7 mg cm-2 MnCo-BP2000 (home-made cathode catalyst [14]). The anode catalyst ink was prepared by mixing PtIr/C, Nafion solution (Dupont, 5wt%) and isopropanol under ultrasonication for 1 h in an ice-water bath. The preparation of cathode catalyst (MnCo-BP2000) ink is the same as the anode. Both the mass ratio of the anode and cathode catalyst to Nafion is 3:1. The AEM was immersed in 2.0 M KOH for 12 h to replace the Cl- anion in the AEM to OH-, and then was washed three times with deionized water. Finally, it was pressed between carbon cloth and carbon paper to make the MEA.The DAFC was tested (G20, Greenlight Innovation Corp. Canada) under the conditions of 1 M/3 M ammonia in 3 M KOH as anode fuel (5.0 mL min-1, controlled by a peristaltic pump), humidified O2 as cathode oxidant (200 mL min-1, backpressure: 20 kPa) and the cell temperature of 80 \u00baC. The physicochemical properties and operation conditions of this model are listed in Table\u00a02\n.To optimize the DAFC anode model, the effects of structure parameters of the anode CL (e.g., porosity, thickness and PtIr loading) on cell performance are investigated, which can guide the optimization of electrode to achieve high performance DAFCs.As shown in Fig.\u00a02\n(a), the calculated polarization curve of this model is compared with the experimental polarization curve of DAFC. The theoretical open-circuit voltage (OCV) of the DAFC is 1.17 V [5]. In contrast, the experimental value of the OCV is only 0.59 V with 1 M ammonia and 0.66 V with 3 M ammonia, possibly attributed to the high onset potential of ammonia electrooxidation and mixed potential in the cathode caused by ammonia crossover. The OCV is set equal to the experimental value because the cathode overpotential is not considered in this anode model. Both the calculated and experimental results demonstrate that the cell voltage decreases with increasing the current density due to the polarization. In the whole current density range, good agreement is obtained between the predicted polarization curve and experimental data. The electrochemical polarization and concentration polarization produced in the oxygen reduction process in the cathode which is not as severe as in the anode and neglected in this model might account for the slight deviation. Fig.\u00a02(b) presents the predicted rate of ammonia crossover as a function of current density. Because the ammonia concentration is the key factor which can cause crossover in Eq.\u00a0(30), the parasitic current calculated for 3 M ammonia (3 M model) is higher than that for 1 M ammonia (1 M model) when the cells were operated under the same current density. In addition, the parasitic current decreases with the increase of the cell current density since more ammonia is consumed in AOR. The 1 M model in Fig.\u00a02(a) shows severe concentration polarization at 390 mA cm-2, which means that most ammonia in the CL is consumed. This can be verified by the result in Fig.\u00a02(b) where the parasitic current drops to nearly 0 mA cm-2 at the same current density and only a small amount of ammonia penetrates through the membrane to the cathode.In the 3 M ammonia model with 0.3 V cell voltage, the distribution of ammonia concentration and gas saturation in the anode cross-section are shown respectively in Figs.\u00a03\n(a) and (b). Fig.\u00a03(a) shows that the ammonia concentration drops to 2 M in the flow channel region due to ammonia transferring to the GDL and the fluid density diminishing caused by nitrogen. Meanwhile, from diffusion layer to membrane, the concentration decreases sharply and reaches almost zero in the CL because most ammonia is consumed during the electrooxidation process. The gas saturation in Fig.\u00a03(b) reflects the nitrogen volume fraction in the anode. The result shows that gas saturation is higher in the CL and GDL than that within the flow channel, because nitrogen is generated in the CL. As shown by Eq.\u00a0(5), the diffusion rate is slower in porous media than that in channel. Nitrogen can easily aggregate in the CL and GDL and hinder the ammonia transport from flow channel to the CL, which explains why the ammonia concentration is much lower in the porous region than that in flow channel in Fig.\u00a03(a) even though ammonia is consumed in the CL by AOR. Nitrogen flows with the ammonia solution stream and accumulates, causing higher gas saturation in the downstream region.The polarization and ammonia distribution of the DAFC model (1 M) are investigated under different anode CL porosity \u03b5cl\n (in Table\u00a02). As shown in Fig.\u00a04\n(a), the porosity almost shows indiscernible effect on polarization when the current density is below 300 mA cm-2. In contrast, as the current density is higher than 300 mA cm-2, the concentration polarization becomes dominant and further turns more obvious with varying CL porosity. The current density decreases by 20 mA cm-2 at 0.1 V when the porosity increases from 0.4 to 0.5. Such difference can be further investigated by comparing the ammonia concentration distribution in the anode with the CL porosities of 0.4 and 0.5 when the cells are operated at 0.1 V, as illustrated in Fig.\u00a04(b) and Fig.\u00a04(c), respectively. The results indicate that the concentration polarization is severe and most ammonia is consumed to nearly zero in the CL. When focusing on the CL, a lower ammonia concentration is found in the case of porosity of 0.5, suggesting a slight decrease in cell performance. Obviously, higher porosity in the CL can enhance the diffusion of both ammonia and nitrogen. However, considering the low ammonia concentration in the CL, the porosity has a higher impact on nitrogen diffusion than ammonia diffusion since nitrogen has a larger diffusivity than ammonia in the feed solution. Therefore, nitrogen generated in the CL can easily diffuse to the GDL with increasing the porosity and in turn hinder the ammonia transport, which accounts for the decreased cell performance with increasing porosity within the high current density region.The polarization curves for various CL thickness (\u03b4cl\n) ranging from 30 to 60 \u00b5m is shown in Fig.\u00a05\n(a). In the mediate current density region where the ohmic polarization has a major influence, voltage loss increases when the CL thickness increases from 30 to 60 \u03bcm. This is because the resistance is proportional to the layer thickness. However, as the current density is above 300 mA cm-2, the cell shows better performance and less concentration polarization upon increasing the CL thickness reaches. This can also be verified by the distribution of ammonia concentration in the anode at 0.1 V, as shown in Fig.\u00a05(b) and Fig.\u00a05(c) with the CL thickness of 30 and 60 \u03bcm, respectively. The results show that the ammonia in the 30 \u03bcm CL is almost consumed while some ammonia remains in the 60 \u03bcm CL. Since there is no enhancement in AOR kinetics with the increase of the CL thickness, the larger volume of a thicker CL can accommodate more ammonia and mitigate the concentration polarization.In addition to the porosity and thickness of CL, the effect of PtIr loading in the CL on cell polarization is also investigated. Due to the sluggish AOR kinetics, high anode catalyst loading is usually required to achieve high-performance DAFCs [13,36]. The PtIr/C loading of 3.4 mgPGM cm-2 used in our experiment is included in the finite element models. As shown in Fig.\u00a06\n, the increase of PtIr loading leads to the improvement in DAFC performance. The improvement in the high current density region is smaller than that in the low current density region. For instance, the current density at 0.42 V is tripled, namely increasing from 22 mA cm-2 to 66 mA cm-2, when the PtIr/C loading increases from 1.4 to 5.4 mgPGM cm-2. In contrast, the current density at 0.1 V is moderately improved from 359 mA cm-2 for 1.4 mgPGM cm-2 to 405 mA cm-2 for 5.4 mgPGM cm-2. Such results indicate that high PtIr loading can reduce the activation polarization of the DAFCs due to the increase of active AOR sites in the CL, but has a less significant effect on reducing the concentration polarization, which is mainly caused by the insufficient supply of NH3 and the accumulation of nitrogen in the catalytic layer.In summary, a three-dimensional two-phase multicomponent model of a DAFC anode incorporated with electrochemistry, mass transport and fluid dynamics has been developed and validated by experimental results. The modeling results indicate that the rate of ammonia crossover decreases with the increase of current density. By manipulating the parameters in the model including porosity, thickness and PtIr loading of the CL, their effects on cell performance and the distribution of ammonia and nitrogen are elucidated. Increasing CL porosity can degrade the cell performance at high current density due to the blockage of ammonia transport by nitrogen gas. The increase of CL thickness increases ohmic polarization in the middle current density region and provides larger volume for ammonia transport and oxidation, which reduces the concentration polarization at high current density. The electrochemical performance of DAFCs can be effectively improved by increasing the active sites of AOR in CL with higher PtIr loadings.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was financially supported by the National Key Research and Development Project of China (No. 2019YFB1504502).", "descript": "\n                  Great progress has been made in recent years in the development of low-temperature direct ammonia fuel cells (DAFCs), motivated by the recognition that ammonia is a carbon-free hydrogen carrier with high energy density, low production cost, and ease in liquefaction at ambient temperature. However, the sluggish kinetics of ammonia electrooxidation and especially complicated mass transport in the anode catalyst layer hinder the further development of DAFCs. In this work, a three-dimensional two-phase multicomponent DAFC model considering the effect of ammonia crossover has been developed. Maxwell-Stefan model, Darcy's law and Brinkman equation are utilized to simulate the multicomponent fluid motion and transport. The predicted polarization curve simulates all experimental results well. The model shows the rate of ammonia crossover decreases with the increase of current density. Besides, the effects of the physicochemical property of the anode catalyst layer, including porosity, thickness and PtIr loading, on cell performance are investigated. The modeling results indicate that decreasing porosity and increasing thickness can slightly improve the electrochemical performance of DAFCs at high current density. Meanwhile, higher PtIr loading can effectively reduce voltage loss before approaching the limiting current density.\n               "}
{"full_text": "Data will be made available on request.Ethanol is one of the most versatile and available biomass-derived molecules, with an increasing industrial production in the last years [1]. It is an interesting building block molecule used in the manufacture of drugs, plastics, and other compounds, both by enzymatic and catalytic routes [2\u20134]. Nevertheless, its main use is as biofuel, providing sustainable energy with properties close to gasoline, but its poor lubricant properties can negatively influence the engine\u2019s durability. Mixtures of ethanol with higher alcohols (mainly branched ones) can overcome this problem. The Guerbet reaction is the most promising route to obtain these second-generation biofuels [5,6]. The complex mechanism of this reaction, involving dehydrogenation, condensation and hydrogenation steps, hinders its industrial application, requiring more research to increase the selectivity. Most of the previous works have been performed in the gas-phase, with temperatures relatively high (>300\u00baC), defining the 1-butanol as the target compound, and reaching maximum yields lower than 30% [7\u201314]. Due to the extended number of side reactions, complex mixtures are obtained, being difficult to purify because of the similar physiochemical properties of most of these products.It is expected that working in condensed phase, at high pressure and close to the ethanol critical point, the activity changes significantly and the product distribution control increases. However, there are only few studies based on the optimum conditions obtained in the gas-phase. Thus, Riitonen and co-workers reached selectivities towards 1-butanol up to 70%, working with different \u03b3-Al2O3 supported metal catalysts, at 250\u00baC with 100\u202fbar, with conversions between 10% and 30% [15\u201317]. Similar selectivities are reported with more complex configurations, using microwaves-assisted reaction and Ni/Al2O3 catalyst [18] or combining metal catalyst with homogeneous bases [19]. At the same temperature but at 176\u202fbar, a selectivity to butanol higher than 83% is reported by Ghaziaskar and Xu, using 8% Ni/\u03b3-Al2O3, with 35% of ethanol conversion [20]. With these bifunctional catalysts, the acid sites promote the condensation as well as the CO hydrogenations by the Meerwein-Poondorf-Verley (MPV) mechanism whereas the metal nanoparticles enhances the dehydrogenation and CC hydrogenation steps. The main route competes with undesired acid-catalyzed additions and different acetals and acetates are also obtained. The relevance of these side reactions increases with the molecular weight of the compounds, whereas the acid-catalyzed condensation decreases. Consequently, the Guerbet mechanism is limited to the first condensation (C4).The co-presence of basic sites is expected to improve the initial dehydrogenation. This approach, scarcely studied, could be adapted to enhance the condensation step, increasing the size of the products to six and eight carbon atoms. Thus, mixed oxides have been considered, reaching selectivities of 1-butanol close to 70%, but with low ethanol conversions (<5%) [21]. Miller and co-workers propose a nickel supported mixed oxide (Ni/La2O3-\u03b3-Al2O3), obtaining 41% of ethanol conversion with more than 70% of butanol [22,23].In the last years, the industry interest has shifted to these higher alcohols since they have a high value in the production of plasticizers, soaps, and fine chemicals, in addition to their properties as solvents and fuel additives [23,24]. However, the production of these heavy alcohols from ethanol has been poorly studied. Preliminary studies in gas-phase propose a sequential configuration (from ethanol to butanol and from butanol to 2-ethyl hexanol (2EH)) with different catalysts and reaction conditions [25\u201329]. This configuration is quite complex from the technical point of view, because of the low selectivity of the first stage. The limited literature in liquid phase (batch configuration) proposes a maximum of 32% of conversion with selectivities of 22% of hexanol and 60% of butanol, at 230\u00baC, using 7\u201310% Cu/MgAl, with a catalyst/reactant mass ratio close to 0.1 [6,30,31]. Despite these promising results, there is a lack of systematic study that allows identifying the relevant catalytic properties and the reaction conditions to improve the selectivity to higher alcohols (>C4) and their corresponding precursors.This work presents a comprehensive study of the liquid-phase ethanol self-condensation considering the production of heavy alcohols (C6-C8) as a one-step process. The activity of different catalysts (HAP, MgAl, MgZr, MgFe, MgCaAl) was studied analyzing the results as a function of their morphological and physic-chemical properties. These materials were chosen considering the previous literature for Guerbet gas-phase condensations, with well-recognized works highlighting the activity and selectivity of HAP [32\u201335] and different mixed oxides [14,21,31]. Despite the different structures of these materials, the same type of active sites (acid and base ones) as well as their hydrogenation capacity by the MPV mechanism are highlighted as the key parameters for the reaction, allowing the comparison. The reaction conditions (no solvent, low catalytic loading) were chosen to achieve a tight control of the activity to facilitate the identification of the catalytic behavior on each single step of the process. These results are analyzed to propose a mechanism, detecting similarities and differences with respect to the gas-phase configuration. Different metals were supported on the most promising materials, analyzing if the presence of nanoparticles with hydrogenation and dehydrogenation activity have a crucial role enhancing the production of the target compounds.A commercial hydroxyapatite (HAP) (Sigma Aldrich) is used in this work, whereas the different mixed oxides (MgZr, MgAl, MgFe and MgCaAl) were prepared in the lab. The details of each particular preparation method are included in the Supplementary Information.Bifunctional catalysts were prepared supporting Pt, Ni, Cu, Ru or Pd by the dry-impregnation method, using nitrate precursors. This was done by adjusting the volume of the metallic precursor solution (prepared to achieve the target 1\u202fwt% of metal loading) to the pore volume of the support, ensuring the total impregnation and a high dispersion. The impregnated catalyst was dried for 24\u202fh, calcined to 700\u202f\u00b0C and reduced with a H2 flow of 20\u202fmL\u00b7min\u22121, up to a temperature of 450\u202f\u00b0C (ramp 5\u202f\u00b0C\u00b7min\u22121), holding this temperature for 3\u202fh.The catalytic morphology was determined by N2 physisorption, using a Micromeritics ASAP 2020 instrument, applying the BET and BJH methods to calculate the surface area, the pore diameter and volume. The crystalline phases were analyzed by X-ray diffraction (PANalytical X\u2032Pert Pro), working with the Cu-K\u03b1 line (0.154\u202fnm) in the range 2\u03b8 =\u202f10 \u2013 120\u00b0. These analyses were done with fresh and spent materials to identify possible changes in the structure during the reaction.The acidity and basicity quantifications (fresh and spent catalysts) were performed by a programmed temperature desorption (TPD) in a Micromeritics AutoChem II 2920, following the desorption of the probe molecules (NH3 or CO2) by a Pfeiffer Vacuum-300 mass spectrometer. A previous cleaning step with He ensures the absence of physisorbed compounds. The saturation was done for 20\u202fmin with a 20\u202fmL\u00b7min\u22121 flow (2.5% NH3 in He or 99.5% of CO2). The desorption was monitoring from room temperature to 950\u00baC, with a slope of 5 \u00baC\u00b7min\u22121.The evolution of the catalytic surface with the reaction was analyzed by diffuse reflectance infrared spectra using a Thermo Electron Nicolet FTIR spectrometer equipped with a MCT/A detector. Spectra were recorded in the 4000\u20131200\u202fcm\u22121 range, with a resolution of 4\u202fcm\u22121, collecting 256 scans/spectrum. 20\u202fmg of catalyst were used in each experiment, being placed inside a high temperature cell. Catalytic measurements were conducted at 230\u00baC under inert atmosphere (N2 flow of 20\u202fmL\u00b7min\u22121) or under an atmosphere saturated in ethanol.The metal loading of the bifunctional catalysts was determined by Inductively Coupled Plasma Mass Spectrometer (ICP-MS) using a HP 7900 of Agilent. Approximately 50\u202fmg of the sample were inserted into a microwave-assisted Teflon bomb; adding HCl (2.25\u202fmL) and HNO3 (0.75\u202fmL) to dissolve the sample. The metal dispersion and particle size distribution was quantified by transmission electronic microscopy (TEM) using a MET JEOL 1011. Histograms and average particle size were calculated by analyzing 100 particles in each sample, using the software Confocal ImageJ.The ethanol condensation was performed in a 0.5\u202fL stirred batch autoclave reactor equipped with a PID temperature controller and a backpressure regulator (Autoclave Engineers EZE Seal). Firstly, 200\u202fmL of ethanol (EtOH) (VWR, 100%) and the catalyst (0.5 or 2\u202fg as a function of the experiment) was added to the reactor. The air was purged with N2, and condensation was carried out at 230\u202f\u00b0C, under 30\u202fbar of N2 (pressure at room temperature, increasing to 80\u202fbar at 230\u202f\u00b0C) under stirring (1000\u202frpm) for 8\u202fh. This temperature is the average of those reported in the bibliography (from 200 to 250\u00baC) [6,22,23,31], considering the limit conditions allowed by the reactor. Based on these points, the conditions were selected as an equilibrium that guarantees the liquid state of all the compounds involved, the conditions that could promote the reaction to heavier compounds (>C4) as well as the minimum severity of the reaction, in good agreement with the desired sustainable character of the process.A filter placed inside the sampling port prevents the catalytic extraction during the sampling. The evolution of the different compounds involved in the reaction was analyzed by gas chromatography (GC) in a Shimadzu GC-2010 equipped with a FID detector, using a 30\u202fm long CP-Sil 8 CB capillary column. Peak assignment was carried out by GC-MS in a Shimazdu GC/MS QP 2010 Plus Instrument, using a 30\u202fm long TRB-5MS capillary column. EtOH and the majority peaks calibrations were done using commercial samples, whereas minority products calibration was carried out using the relative carbon concept proposed by Scanlon and Willis [36]. The analytical conditions (detailed in the Supplementary Information, Table S1) were chosen to guarantee that the ethanol signal does not saturate the detector, even in the case of initial samples (the highest signal expected). Each reported experimental point corresponds to the average value of at least two analyses. The maximum standard deviation of the reported values is 6%. The results were analyzed in terms of conversion, selectivities, and carbon balance, according to Eqs. (1), (2), and (3), where \u201c\n\n\n\nn\n\n\ni\n\n\n\"\n\n corresponds to the number of carbons in each compound, and \u201d\n\n\nC\n\n\ni\n,\nt\n\n\n\u201d is the molar concentration of this compound at the time analyzed:\n\n(1)\n\n\nSelectivity\n:\n\n\u03c6\n\n\n\n%\n\n\n\n=\n\n\n\n\n\nn\n\n\ni\n\n\n\u00b7\n\n\nC\n\n\ni\n,\nt\n\n\n\n\n2\n\u00b7\n\n\n\n\n\n\n\nEtOH\n\n\n\n\n0\n\n\n\u2212\n\n\n\n\nEtOH\n\n\n\n\nt\n\n\n\n\n\n\n\n\u00b7\n100\n\n\n\n\n\n\n(2)\n\n\nConversion\n:\n\nx\n\n\n\n\n%\n\n\n\n=\n\n\n\n\n\n\nEtOH\n\n\n\n\n0\n\n\n\u2212\n\n\n\n\nEtOH\n\n\n\n\nt\n\n\n\n\n\n\n\n\nEtOH\n\n\n\n\n0\n\n\n\n\n\u00b7\n100\n\n\n\n\n\n\n\n(3)\n\n\nCarbon\n\nBalance\n:\n\nCB\n=\n\n\n\n\u2211\n\n\n\n\nn\n\n\ni\n\n\n\u00b7\n\n\nC\n\n\ni\n,\nt\n\n\n\n\n\n\n\n2\n\u00b7\n\n\n\nEtOH\n\n\n\n\n\n0\n\n\n\n\n\n\n\n\nAt reaction conditions, the presence of solid deposits is discarded (assumption corroborated by TPO analyses). The differences in carbon balance closures are then attributed to the formation of light compounds, not detected in the liquid phase. This hypothesis was corroborated by the analysis of the gas phase, being recovered using a sampling bag. The gases accumulated were analyzed by GC-MS. Only a qualitative analysis is possible, the quantification and temporal evolution evaluation being not possible due to the accumulative sampling required.The accuracy of GC-FID analyses in these conditions was probed analyzing ethanol signal and the standard deviation of 12 repetitions of the same sample, as indicated in the Supplementary Information (see Table S2). The values indicated are congruent with the theoretical conversion required to obtain the products detected. However, this second methodology is discarded with the aim to compare the relevance of permanent gases produced in the reactions, analyses that could be not possible if the conversion is calculated based on products.Initial experiments were focused on the comparison between bulk materials to identify the catalytic properties that maximize the selectivity to higher alcohols, mainly C6 and C8, or their corresponding condensed precursors (in absence of metal particles with hydrogenation activity). In all the cases, anhydrous ethanol was used as reactant, setting the temperature at 230\u00baC. The evolution of ethanol conversion with time is analyzed in \nFig. 1.The absence of an external solvent (high reactant initial concentration, 789\u2009g\u00b7L\u22121) as well as the low catalytic loading (2.5\u2009g/L) justifies the conversions obtained. Similar conversions are reported in most of the previous literature when working without external solvent, even with higher catalytic loading [14,17,21,23]. In fact, comparing the results in terms of mass of ethanol reacted per time and catalyst loading, the values obtained in this study (from 0.473 to 0.907\u2009g/g\u00b7h with MgZr and MgAl (3/1), respectively) are better than most of those drawn from studies in absence of reduced metals (0.710\u2009g/g\u00b7h [14,30], 0.647\u2009g/g\u00b7h [21]), with the exception of Perrone and co-workers, who obtained values of 1.326\u2009g/g\u00b7h with a partial substitution of Mg2+ and Al3+ by Cu2+ and La3+, respectively [31].Although higher conversions could be anticipated working with higher catalytic loading, reported conversions are more useful for gaining further understanding on reaction mechanism and the role of the different surface sites. MgZr shows the poorest activity, maximum conversion of 1.2% after 8\u2009h, suggesting that longer times could rise this value because of the increasing trend observed. Low conversions (1.3\u20131.5%) are also reached with MgAl (2/1) and MgCaAl, but in these cases, the values remain almost constant after the first 2 and 3\u2009h, respectively. On the other hand, conversions obtained with HAP, MgFe and MgAl (3/1) are significantly higher, with values from 2% to 2.3%. As in the previous case, three different trends are observed, with a constant conversion in less than 2\u2009h with MgFe, a continuous evolution observed with HAP, and an intermediate behavior of MgAl (3/1), with a fast increase during the first 2\u2009h and a second and slower increasing trend at longer times after an intermediate and flat step.\n\nTable 1 summarizes the product distributions after 8\u2009h reaction time, in terms of absolute and relative selectivity of main families of compounds. Considering the large number of products obtained, the absolute selectivities are lumped by families, according to their number of carbons. Thus, C2 only involves the acetaldehyde, whereas C4 corresponds to crotonaldehyde, butanal, and 1-butanol, with a similar distribution for C6 and C8 families. Although quantitatively analyzed, side products (as esters) are not included in these families, being considered together as \u201cundesired liquid by-products\u201d. Reported carbon balance closure is calculated comparing the initial ethanol loading and the concentration of ethanol and all the desired and undesired reaction product analyzed in the liquid phase after performing the reaction. Although this carbon balance closure is very high, as the ethanol conversions are low, selectivity to gas products can be important. Thus, apparent selectivities to gaseous by-products have been estimated from ethanol conversion and carbon balance, being these values also reported in Table 1.As anticipated, the liquid-phase configuration promotes a different distribution than the gas-phase one, with a decrease of C4s in favor of a higher number of heavy compounds (C6 and C8), mainly observed with MgAl (2/1) and MgCaAl. These compounds are the heaviest detected in this study, concluding that subsequent dehydrogenations and/or aldolizations require more severe conditions. In the gas-phase reactions, only some traces of C6s are detected using bimetallic modified mixed oxides [31,43].More than 26% of the total compounds correspond to C6 and C8 when using MgCaAl, the best catalyst promoting condensation. In the case of MgAl (2/1), this percentage decreases to one half of this value. In terms of functional groups, MgAl (2/1) shows a high hydrogenation activity, alcohols representing almost 80% of the total. On the other hand, with MgCaAl, alcohols only correspond to 14%, suggesting that condensation prevails over hydrogenation. In both cases, all the C6-C8 alcohols are obtained following the same ratio as the carbon families. Regarding the aldehydes, both materials produce hexanal, which hardly condensates with other ethanol molecule, limiting the production of C8s.C6 are the heaviest compounds detected with MgZr (7.6%). A similar distribution of aldehydes and alcohols is obtained. The high concentration of crotonaldehyde (almost 30%) with respect to butanal (almost negligible) suggests that the CO hydrogenation is easier than the CC one, most of the butanal being directly converted into butanol.HAP, MgFe, and MgAl (3/1) do not show relevant activity for heavy condensations, with less than 3% of C6 and C8 compounds. In the case of HAP, the results are congruent with a lack of activity promoting the condensation of heavy compounds, obtaining a sample enriched in C4 (>40%) with a good balance between aldehydes and alcohols (35% of butanol). On the contrary, MgFe and MgAl (3/1) demonstrate a poor condensation activity, with total selectivities lower than 10%, producing almost 90% of undesired gases (ethylene and diethyl ether).To sum up, these materials reveal differences not only in terms of conversion but also in the product distribution. The instability of some of the catalysts at reaction conditions could be a possible justification of these discrepancies. IPC results indicate the absence of metals in the reaction liqueur since Mg, Al, Ca, Fe or Zr are not detected. The catalytic leaching is then discarded as a deactivation cause. On the other hand, the comparison between crystallographic phases of fresh and spent materials (XRD diffractograms shown in Fig. S1 and Table S3, discussed below) reveals a good correspondence between peaks before and after the reaction, prevailing the amorphous structure of these materials, without observing the parent hydrotalcite structure (crystalline one). Thus, the physical and morphological stability of these materials is also corroborated. In this context, a relevant role of their catalytic properties is suggested, promoting different steps of the main mechanism, preventing undesired lateral reactions, and minimizing the adsorption of different compounds. This discussion requires the analysis of the morphological and physico-chemical properties of these materials, main data being summarized in \nTable 2.Morphological results discard any relevant role of the external surface area or mass transfer limitations. In general, materials with a fast initial conversion (MgFe, MgAl (3/1), MgAl (2/1)) exhibit high concentration of acid sites (mainly weak and medium ones), whereas materials with a lower acidity require longer times. These results demonstrate the relevance of acid sites adsorbing the ethanol molecule. This conclusion was verified by DRIFT spectroscopy, observing more pronounced bands with MgAl (3/1), the most acidic material. The identification of all the DRIFT bands, based on previous literature [37,38], as well as their discussion, is detailed in Fig. S2.MgAl (2/1), MgCaAl, HAP and, in a small extent, MgZr have the highest capacity to promote the condensation, in agreement of their high concentration of medium and strong basic sites, releasing water. Previous literature indicates that some oxides derived from a hydrotalcite-type precursor undergo rehydration in presence of water, modifying their surface chemistry by the reconversion of the O2- basic sites to OH- ones (Br\u00f8nsted sites responsible of aldol condensations) [39]. This hydration depends on their surface structure and can also occur with HAP [40]. This fact would be enough to increase its condensation capacity, not so relevant to observe differences in the crystallographic structure (surface phenomenon). This effect was corroborated by comparing the acidity and basicity of fresh and spent catalysts. Results obtained (shown in Figs. S3-S4 and Table S4) indicate the relative enrichment in medium-strength basic sites of MgZr and HAP, the materials that suffer reactivation, whereas the strength of acid sites remains almost invariable, resulting in stronger basic/acid pairs. According to the literature, these sites promote the dehydrogenation of ethanol via E1cb elimination mechanism [41], justifying the reactivation observed with these materials. This reactivation has not been reported in gas-phase reactions, suggesting that Br\u00f8nsted sites are not stable at high temperatures.A similar reactivation and increase in the basic/acid sites strength is observed with MgAl (3/1), but it shows low condensation capacity (5% of >C2). This fact is justified by the lack of correlation between basic/acid pairs, prevailing the activity only catalyzed by acidity. In this case, water is released by undesired acid additions (yielding 1,1-diethoxyetane, selectivities up to 8%) and ethanol dehydrations, producing ethylene and diethyl ether in large amount (>84%). These two compounds were identified in the analysis of the gas phase by GC-MS. In good agreement with their presence, the carbon balance closure with this material (98.1%) is the lowest one. In other materials, a clear decrease in the acidity without a relevant change in the strength distribution is observed (43%, 61%, and 62% with MgAl (2/1), MgFe, and MgCaAl, respectively). With these materials, the expected rehydration effect is shielded by the adsorption of unsaturated intermediates, severely blocking the acid sites and hindering further reaction progresses.The different profiles of the two MgAl mixed oxides deserves special attention, suggesting a strong influence of the preparation method and the Mg/Al ratio. The minor differences in terms of weak and medium-strength acidity discard a different dehydrogenation capacity. In fact, the conversions during the first 3\u2009h are very similar. XRD diffractograms (Fig. S1) illustrate relevant differences, consequences of the interaction between oxide phases during the preparation. In good agreement, a spinel (MgAl, JCPDS 00\u2013021\u20131152) is detected in MgAl (2/1), in addition to periclase (MgO, JCPDS 03\u2013065\u20130476), the only phase observed with MgAl (3/1), suggesting that Al is in an amorphous phase or in crystals too small to be identified with the resolution of this equipment. The different coordination state of Mg and Al on these crystalline phases (and the corresponding morphology of the acid sites) seems to play a key role in the interaction of reaction intermediates with the catalysts, producing opposite effects in the catalytic activity. DRIFT spectra (Fig. S2) visualize these differences. Thus, signals of ethoxide species (1460\u2009cm\u22121\n[42]) are significantly more evident in the case of MgAl (3/1) than in MgAl (2/1).The reaction products from the two materials with the lowest weak acidity (HAP and MgZr) are enriched in acetaldehyde, suggesting that these sites are involved in the condensation, directly by an acidic mechanism ore stabilizing the basic sites (basic/acid pairs). The high acidity justifies the poor results obtained with MgAl (3/1) and MgFe, with more than 84% of carbon as gaseous by-products. The sampling method for the analysis of these gases does not allow accurate and continuous quantification, but in both cases more than 90% corresponds to ethylene, with lower amounts of diethyl ether and other compounds in traces. These results contrast with those obtained with MgAl (2/1), material that prevents the formation of gaseous by-products. The high acidity of this material is well balanced with its basicity, suggesting the primacy of basic/acid pairs over isolated acid sites, promoting the main route of the Guerbet reaction.There is a good correspondence between the medium-strength basic/acid sites and the total selectivity to C6 and C8 compounds, indicating that these sites are the most relevant ones to promote condensations, as induced from the total C6-C8 selectivity of MgCaAl (25.8%) and MgAl (2/1) (12.7%). Even with this amount of C6 and C8, MgAl (2/1) produces the maximum amount of C4 (78.6%), followed by HAP (40.2%), being suggested as promising supports for next studies. These catalysts produce the maximum total alcohol selectivity, being enriched in butanol.Thus, the conversion is not always related with a high activity in the Guerbet reaction since also undesired ethanol dehydration and acid-catalyzed additions occur, obtaining light gases. MgFe and MgAl (3/1) produce selectivities to diethoxyethane at initial times higher than 9.7% and 10.4% (values that corresponds to relative weights of 39% and 24% of this compound in the products\u2019 mixture), respectively. Only in the case of MgAl (2/1), this conversion corresponds to desired products, observing 1-butanol since the first samples.To explain these results, a separate experiment with MgAl (2/1) and pure butanol as reactant was done. The evolution of the main intermediates is detailed in Fig. S5. Less than 1.7% of conversion is reached after 8\u2009h (99.8% carbon balance), with ethanol and butanal (16.3%) as the main reaction products and less than 1% selectivity for C6s and C8s (0.2% 3-hexen-1-ol, 0.4% of 1-hexanol, 0.2% of 2-etil-1-hexanol). Thus, the C6 adducts are suggested to be produced mainly by the condensation between crotonaldehyde and acetaldehyde, but not so easily from butanal (the selectivity of this compound is more than 20 times higher than when using ethanol). Thus, once crotonaldehyde is partially hydrogenated to butanal, the condensation capacity decreases. This hypothesis is congruent with the stability of the enol intermediate produced during the condensation in presence of CC double bonds. The quantification of acetaldehyde and ethanol suggests a partial reversible character of condensation not observed in gas phase or when ethanol is used as reactant since the reverse reaction is catalyzed by the same active sites than the direct one, the condensation prevailing in presence of aldehydes. According to this study, strong acid sites and basic/acid pairs are required to promote the butanol double dehydrogenation and condensation. Their low concentration as well as the adsorption of the C6 and C8 compounds with the subsequent active-sites blockage conditions the low second condensation ability and justifies the prevalence of hydrogenated C4 compounds.In all the cases, the hydrogenation capacity decreases with the size of the aldehydes. This is an anticipated result considering the absence of metal nanoparticles and the prevalence of the MPV route. According to this mechanism [8], the hydrogenation requires the co-adsorption of the aldehyde and an alcohol, on an acid site, obtaining a cyclic compound as the reaction intermediate. The stability of this intermediate decreases with the size, being the most unstable the one of six carbons (crotonaldehyde + ethanol). Thus, the ratio of butanol to the total C4 family of compounds reaches values higher than 99% with Mg/Al (2/1) whereas this percentage decreases to 45% when analyzing the C8 hydrogenation capacity.To establish a reaction mechanism, these results after 8\u2009h must be analyzed together with the evolution in time of the different intermediates. The most relevant data are shown in \nFig. 2, excluding MgFe and MgAl (3/1) because of their negligible condensation activity.Acetaldehyde is the first intermediate obtained with all the materials, with initial selectivities of 100% and a continuous decreasing trend with the time. This decreasing trend is more marked in those catalysts with higher condensation activity. According to these results, the ethanol dehydrogenation to obtain acetaldehyde is the starting point of the process, discarding the direct coupling between two ethanol molecules as a relevant step. Despite the lack of total agreement about the ethanol condensation in gas-phase, this mechanism prevails in the literature over the ethanol direct condensation [25,44].The typical evolution of a successive condensation C4-C6-C8 is observed with MgZr and, being not so marked, with HAP. With these two materials, C4s and C6s compounds reach a maximum in selectivity, slightly displaced in conversion in the case of C6s, according to their intermediate character. The high hydrogenation activity of MgAl (2/1) alters these curves, observing a high accumulation of C4s due to the stable character of butanol. The high condensation activity of crotonaldehyde is observed with MgCaAl, material with which C4s and C6s almost appear simultaneously, with a high proportion of the last ones, consuming most of the crotonaldehyde obtained in the first condensation.As to the evolution of each intermediate, a first hydrogenation of CC bonds is observed for the C4 family (\nFig. 3a), suggesting a consecutive production of butanal and butanol that is assumed to be extrapolated to heavier compounds (C6 and C8). The CO hydrogenation by the MPV mechanism as well as the subsequent condensations justify the low selectivity to these aldehydes, with a fast production of heavy compounds or alcohols once these intermediates are obtained. The slow but continuous hydrogenation activity of these materials (see Fig. 3b) suggests that longer times could enhance the selectivity of these target compounds with all the catalysts except MgFe and MgAl (3/1) because of their lack of condensation activity.According to these results, the basis of the mechanism in the liquid phase is based on the one in the gas phase but with some modifications. Different side reactions are observed (ethoxides not detected in gas phase and quite relevant in the liquid one). The softer conditions of this configuration and the absence of noble metals justify the slow rates of hydrogenations, increasing the opportunity to obtain heavier chemicals by the condensation of unsaturated compounds. The presence of isomers as well as partially hydrogenated derivatives rises the total number of compounds. Considering the experimental results, the scheme proposed for the liquid-phase ethanol condensation can be updated to a more complex one considering the different compounds detected with more than four carbon atoms, as shown in \nScheme 1. In this scheme, the different isomers that could be simultaneously obtained are indicated, as well as the fact that C8s are produced by the sequential addition of an acetaldehyde molecule to crotonaldehyde, the direct condensation of two crotonaldehyde molecules or the condensation involving butanal being discarded as relevant at these conditions.The temporal evolution of reaction products also supports the previously mentioned required trade-off between hydrogenation and condensation activity, the fraction of heavy compounds reaching a maximum with those materials highly selective to alcohols, whereas observing a slow but continuous increasing trend with MgAl (2/1) and HAP, as shown in \nFig. 4. These results indicate that higher selectivities can be obtained modifying the reaction conditions.According to the preliminary analysis, a higher catalytic loading is anticipated to have a positive influence on the product carbon-length if the condensation activity prevails over the hydrogenation one. To check this hypothesis, the results obtained with 2.5\u2009g/L are compared with those reached using 10\u2009g/L. Once the catalytic loading is increased, the conversion is 160% higher with MgAl (2/1), from 1.5% to 3.9% (with a continuous increasing trend during the 8\u2009h), whereas a lower increase from 2.1% to 2.9% is observed with HAP (flat conversion after 5\u2009h). However, the main differences are related to the selectivity distribution, compared in \nFig. 5.There is an increase in the selectivity of the C6s, in detriment of the C4s and, in the case of the HAP, of acetaldehyde too. The amount of C8 compounds is negligible in both cases, suggesting that stronger sites are required to promote this step. The increase in the condensation activity is in detriment to the hydrogenation one. This is congruent with the decreasing stability of MPV intermediates with the size of the intermediates, as discussed before. Thus, hydrogenated compounds decreased from 79.1% to 35% MgAl (2/1), whereas the initial 34.9% reached with 2.5\u2009g/L of HAP declines to 10.4% with 10\u2009g/L. In both cases, these reductions are proportional to the increases in C6s. Butanol is the main alcohol in both cases, 30.9% and 9% with MgAl (2/1) and HAP, with only 2.9% and 1.4% of hexanol, respectively. No significant differences in terms of side products were observed with any of these materials (selectivities from 6% to 11%), being more than 80% due to the 1,1-dietoxiethane, the side product obtained by the ethanol dimerization. On the contrary, the control over the reaction decreases, observing 19.2% (MgAl (2/1)) and 55.4% (HAP) of undetected gases.Results obtained with MgAl (2/1) (>71% of target compounds) are significantly better than those reached with HAP, suggesting a good alcohols production by a second hydrogenation step. The poor increase of activity observed with HAP suggests that the positive effect of rehydration has a limited impact, and the results are mainly conditioned by other aspect, whose negative role is more evident as the reaction advances. In this context, previous literature suggests that water can play a double role, with a negative influence if the interaction with the catalyst occurs via adsorption on the strong sites [43].The analysis of the influence of water content is of great interest, from the catalytic point of view and to evaluate the technical-economic viability of this process. As to the mechanism, it allows to identify the rate determining step. As to the technical approach, the use of aqueous ethanol is preferred in terms of costs since anhydrous ethanol is more expensive due to the required additional dehydration steps. Simple separation processes such as distillation allow reaching a maximum ethanol purity of 95% (v/v), limited by the minimum-boiling ethanol-water azeotrope, requiring expensive technologies (azeotropic distillation with benzene or cyclohexane, distillation combined with adsorption) to fully remove water from ethanol.The results obtained in absence of water (anhydrous ethanol used as reactant) were compared to those introducing 2.5% and 5% (v/v) of water. Main results after 8\u2009h are summarized in \nTable 3. Results in terms of ethanol conversion seem to be not very conclusive, with some materials for which the water promotes it (an increase in conversion of 44% observed with MgAl (3/1)), and materials with the opposite trend, more evident with HAP and MgFe (decreases of 52% and 60%, respectively). An intermediate situation is observed with MgAl (2/1), with almost constant conversion despite the water content.These results corroborate that water preferentially interacts with the catalysts via dissociative adsorption on the strong sites, producing the deactivation of the materials. This phenomenon has been previously reported by Miller and co-workers\n[45]. Thus, Lewis strong basic sites (O2-) are converted into weaker Br\u00f8nsted sites (OH-), modifying the catalytic activity of these materials [6,14,46]. At the same time, Lewis acid sites are blocked by the adsorption of hydroxyl anions. In fact, the activity decreases in those catalysts that have the highest concentration of strong basic sites. The increase in the activity observed with MgAl (3/1) is explained by the lowest dissociation due to the lowest strong basicity (prevailing the molecular adsorption), and the subsequent lower blockage of the acid sites that promote the dehydrogenation.The condensation capacity is also altered, enriching the final mixtures in acetaldehyde (more than 60% in all the cases), observing a total disappearance of C8 compounds and a very significant decrease in C6 and C4 condensed ones. A reduction in the condensation capacity is observed with all the materials. This result is congruent with the adsorption of water on the condensation active sites, preventing the advance of the reaction. Thus, in presence of water, the Guerbet reaction is limited by the condensation step and, even in those cases in which the ethanol dehydration is promoted, there is not a clear advance to the target compounds.Acid sites are also involved in the hydrogenation by MPV mechanism. In good agreement, their blockage explains the almost total absence of alcohols even when feeding only 2.5% of water. This situation affects to all the fractions, observing only traces of butanol (lower than 1% in all the cases), with a total disappearance of C6 and C8 alcohols, even in those cases when the corresponding condensated adducts are still produced in significant amount, such as in the case of both MgAl materials.Water also promotes the production of 1,1-diethoxyethane, except for HAP, with a constant selectivity of 5.6% with and without water. This acetal is obtained by the reaction between an alcohol and an aldehyde molecule, and it has been observed in the literature with selectivities close to 40% in presence of acid materials [14]. With these basic-acid materials, its selectivity reaches a maximum of 14.1% with MgAl (3/1) and 5% of water. In all the cases, this compound represents more than 90% of the total undesired products detected.To sum up, a negative influence of free water is demonstrated, in agreement with the conclusions obtained in gas-phase, even with those catalysts showing reactivation in presence of the small amount produced during the reaction. Thus, the typical percentage of water presents in an azeotropic ethanol inhibits the reaction. Considering the increase in costs, the economic viability requires an improvement in the selectivity towards the target alcohols. The hydrogenation via the MPV mechanism is not enough to promote it, suggesting the use of bifunctional catalysts to activate the hydrogen produced during the dehydrogenation.Improving the dehydrogenation, the excess of acetaldehyde is expected to promote the condensations. Among the transition metals, the dehydrogenation activity of Cu is highlighted in the literature [47]. On the other hand, alcohols are produced by hydrogenation steps. The presence of noble metal nanoparticles could have a positive effect activating the hydrogen produced during the dehydrogenation, enhancing the hydrogenation activity [5]. This section analyzes the activity of different bifunctional catalysts (Cu, Ru, Pd, Pt) using MgAl (2/1) as support. Although most of the C4 obtained with MgAl (2/1) is butanol, the presence of dehydrogenation active metals could enhance the enolization of crotonaldehyde or butanal prevailing over the total hydrogenation of these intermediates. Improving the hydrogenation, not only the alcohols selectivity but also the conversion is expected to increase reducing the relevance of adsorption processes. In all the cases, a theoretical 1\u2009wt% of metal loading is used, to limit the interference on the support properties and to guarantee the appropriate metal dispersion. Main results related to the catalytic characterization are summarized in \nTable 4\n.\nThe specific metal loadings measured by ICP-MS (>0.93) indicate a high similarity between materials, discarding any effect of this parameter in the discussion of their catalytic behavior. In the same way, TEM microscopy (Fig. S6) reveals the presence of metal particles in the range of 4\u20135\u2009nm, with a very similar dispersion of Pd and Pt (29\u201330%), being slightly lower in the case of Cu (24%), and a bit higher with Ru (38%). As expected, the presence of metal particles partially modifies the acidity and basicity of the support. Thus, all the materials show a decrease in the acidity (more relevant in the cases of weak and medium sites), as well as the corresponding decrease (with the exception of Ru/MgAl) in the basicity. The slight increase in the strong acidity of some materials (Cu and Ru) is explained by the strong acid character of metal ions, suggesting the coexistence of some cations on the surface, together with the metal particles visualized by TEM. To sum up, the characterization of these materials corroborates a partial alteration in the morphological and chemical properties of the support, justifying the need of working with low amounts of metal (1%) to minimize these effects and guarantee a correct analysis of these effects.\n\nFig. 6 compares the main results, in terms of ethanol conversion and product distribution. A clear improvement in the conversion is observed with Pd (3.4%) and mainly Cu/MgAl, reaching a final value of 6.8% (2.68\u2009g of EtOH converted per g of catalyst and hour). This value is higher than those reported in the literature for systems with similar metal loading, even working with catalyst/reactant mass ratios four times higher [14,21]. The closest value published (2.46\u2009g/g\u00b7h) corresponds to a catalyst involving Cu and Ni, two metals for dehydrogenation [48]. This supports the hypothesis of the high relevance of the first dehydrogenation on the global reaction.The conversion is proportional to the weak-strength basic/acid pairs (see \nFig. 7a) for all the materials expect for the Cu/MgAl, material with which the conversion is almost double than the expected one. This result indicates that the presence of reduced metals (Pd, Pt, Ru) is not relevant for conversion, only affecting to the product distribution, whereas the well-known dehydrogenation activity of Cu plays a key role in the reaction.All the bifunctional catalysts produce mixtures enriched in acetaldehyde. This suggests that the condensation activity is affected by the metals. However, C6s and C8s reach more relevance, representing 14.9% with the parent support but 25% with Pt. These data, together with the continuous increasing trend observed in their profiles (shown in Fig. S7-10), indicate that the dehydrogenation activity is faster than the consumption by condensation of the acetaldehyde and longer times are expected to produce an enrichment in the heavy fractions. Lateral reactions are also favored by bifunctional catalysts, obtaining a ratio between the sum of esters and ethers and the sum of all the compounds involved in the Guerbet route that increases from 0.06 of the original MgAl (2/1) to 0.18, 0.28, 0.15, and 0.23, with Cu, Ru, Pd, and Pt, respectively.The total selectivity to target compounds (alcohols) decreases from 79.1% (MgAl (2/1)) to 49.7%, 12.3%, 42.6%, and 42.7% (Cu, Ru, Pd, and Pt). These values are justified since the presence of heavier compounds hinders the hydrogenation in absence of reducing atmosphere, using the hydrogen removed during dehydrogenation that is not desorbed from the liquid ethanol. Thus, the MPV hydrogenation mechanism prevails, the total selectivity of alcohols being proportional to medium-strength acidity, as illustrated in Fig. 7b. However, interesting conclusions can be extracted by analyzing the distribution of these alcohols, shown in \nFig. 8.Butanol represents almost 100% of the alcohols observed with MgAl (98.7%), whereas the presence of metal nanoparticles produces a significant decrease in this percentage in favor of heavier fractions (in good agreement with the higher conversions). Thus, the alcohol distribution obtained with Cu/MgAl indicates 81.9% of butanol, with 15.7% of C6 alcohols and 2.4% of C8. The same analysis reports 88.5%, 8.3%, and 2.1% of butanol, C6, and C8 alcohols with Ru/MgAl; 77.6%, 18.3% and 4.1%, respectively, with Pd/MgAl; and 69.5%, 22.2%, and 8.3%, with Pt/MgAl. This last catalyst presents a very interesting hydrogenation capacity combined with the highest selectivity for unsaturated compounds. However, these results are not significantly different from those reached with Cu/MgAl, a catalyst that has a conversion 240% higher than Pt/MgAl. In global terms, and considering that Cu/MgAl produces the second higher alcohol total selectivity, this material is chosen as the optimum one among the tested in this screening.The activity of different catalysts in the ethanol liquid-phase condensation reveals a strong influence of acidity and basicity of the materials, global results being limited by the dehydrogenation activity. This limitation is more evident as the size of the molecule that must undergoes dehydrogenation increases, in such a way that dehydrogenation metal phases are required to promote the production of C6s and C8s and their subsequent alcohols. These compounds, not observed in gas-phase, are produced in the condensed one since the hydrogenation rate is significantly slower, promoting successive condensations.Some materials are reactivated by the rehydration of aluminum oxide phases with the water in situ produced. However, the competitive adsorption of water and ethanol on the acid sites produces a decrease in the dehydrogenation activity in presence of small percentages of free water (2.5, 5%), conditioning the complete evolution of the reaction.Cu is identified as the optimum metal observing a synergetic effect with the MgAl (2/1) support. 1% Cu/MgAl (2/1) allows a conversion almost five times higher than the one obtained with the parent material, producing almost 50% of alcohols with a selectivity distribution enriched in C6s and C8s (18.1%). These results involve a significant improvement in this field, supporting the liquid-phase production of heavy alcohols from ethanol with low catalytic loadings.\nLaura Faba: Methodology, Supervision, Data curation, Writing \u2212 original draft. Jennifer Cueto: Investigation, Data curation, Writing \u2212 original draft. M\u00aa \u00c1ngeles Portillo: Investigation, Resources. \u00c1ngel L. Villanueva Perales: Formal analysis, Writing \u2212 review & editing. Salvador Ord\u00f3\u00f1ez: Conceptualization, Supervision, Formal analysis, Writing \u2212 review & editing. Fernando Vidal: Supervision, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work has been carried out in the framework of the Project BIOC4+ (PY18-RE-0040) funded by Junta de Andaluc\u00eda and European Union (ERDF funds). Authors would like to acknowledge the technical support provided by Servicios Cient\u00edfico-T\u00e9cnicos de la Universidad de Oviedo.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.apcata.2022.118783.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  The production of higher alcohols (C4+) via ethanol liquid-phase condensation is studied in this work, screening catalysts with different acid/base properties, observing similarities but also relevant differences with respect to gas-phase reactions in the gas phase. The mechanistic analysis demonstrates the relevance of acidity, mainly to promote the dehydrogenation steps. In the same way, side reactions and hydrogenations have less relevance than in gas-phase, promoting the condensations and, subsequently, obtaining heavy compounds. The highest alcohol selectivity is reached with MgAl (2/1), with more than 79% of C4+ selectivity, but the activity of this material is conditioned by the low conversion obtained. The presence of water reduces the activity because of a competitive adsorption on the catalytic sites whereas the activity increases significantly when using bifunctional catalysts. The best results, obtained with 1% Cu/MgAl (2/1), allow rising the conversion up to more than 460% respect to the parent mixed oxide, with almost 44% of the alcohol mixture enriched in heavy compounds, mainly C6 and C8.\n               "}
{"full_text": "In the past decades, many environmental problems have emerged in rapid succession, and effective measures are urgently needed to deal with the situation. Among them, all kinds of sewage containing various refractory organic matter discharged from production and living cause more severe pollution of surface water and groundwater, which is more prominent in developing countries and regions [1\u20137]. These emerging pollutants (Eps), such as pharmaceuticals, personal care products, surfactants, endocrine-disrupting chemicals and sterols, are difficult to remove after entering the environmental water body, causing great harm to human production and life. So, it received wide attention and many related studies have been devoted to effectively eliminating such refractory organic pollutants. To date, the removal of Eps has been extensively studied, including adsorption [8], flocculation [9], centrifugation [10], coagulation [11], gravity separation [12], biodegradation treatment [13\u201315] and advanced oxidation processes (AOPs) [16\u201322]. Among them, AOPs can generate hydroxyl radicals (\u2022OH, E0\u00a0\u200b=\u00a0\u200b1.8\u20132.7\u00a0\u200bV), sulfate radicals (SO4\n\u2022\u2212, E0\u00a0\u200b=\u00a0\u200b2.5\u20133.1\u00a0\u200bV), superoxide radicals (O2\n\u2022\u2212, E0\u00a0\u200b=\u00a0\u200b0.94\u00a0\u200bV), singlet oxygen (1O2, E0\u00a0\u200b=\u00a0\u200b0.65\u00a0\u200bV) and other reactive oxygen species (ROS), which can effectively realize the oxidation and even mineralization of refractory organic pollutants [23\u201328].Different advanced oxidation technologies use specific oxidants. Common oxidants (such as hydrogen peroxide (H2O2), peroxymonosulfate (PMS), ozone (O3), and persulfate (PS)) can be activated in a variety of ways to generate ROS. Activation methods include ultrasonic activation [29], thermal decomposition activation [30], microwave activation [31], UV photocatalysis [32\u201334], alkaline activation [35] and transition-metal catalyst activation [36,37]. Among these activation methods, the catalyst can effectively catalyze the reaction by reducing the reaction energy barrier. Especially transition metal activation (metal ions or metal oxides) is the most common and efficient way [38\u201341]. Many studies have confirmed that many transition metal-based catalysts (Fe, Cu, Co, Mn, Fe3O4, Co3O4, etc.) can effectively catalyze these oxidants for water treatment. Compared with traditional metal oxides, the emerging research hotspot metal sulfide has attracted much attention. It has appeared in many research fields, such as lithium-ion batteries, supercapacitors, oxygen generation reactions, CO2 reduction [42\u201347], etc. At the same time, many works have been published on the application of metal sulfides in the field of AOPs. Related studies have found that metal sulfide has higher electric conductivity, electrochemical activity, catalytic activity and other excellent physical and chemical characteristics than the corresponding oxides, making metal sulfide an ideal candidate to replace metal oxides [48\u201350]. However, metal sulfide synthesis, characterization and application still lack systematic summaries. Meanwhile, the catalytic mechanism and rules of different metal sulfides in AOPs also need to be further elucidated to help the future research direction of metal sulfides in treating water pollution in AOPs.The current research mainly focuses on applying metal sulfide materials in AOPs to treat water pollution. The organizational structure of this review article is as follows. After briefly comparing the similarities and differences between metal oxides and metal sulfides, the first section introduces the synthesis and characterization of metal sulfides in detail. Subsequently, the second part summarizes the application of different types of metal sulfide in AOPs (including catalysis, light and electricity). The third part discusses the application of heterojunction nanocomposites made of metal sulfide combined with other materials in AOPs. Section four concludes the reaction mechanism of metal sulfide in different reaction systems through density functional theory (DFT) calculation, and the active root of metal sulfide is revealed. Finally, some constructive suggestions on the feasibility and industrialization development of metal sulfide materials as catalysts for wastewater treatment were put forward. The review will help the advanced oxidation technology of metal sulfide play a more critical role in water pollution control and guide the development direction of this technology in the future.Metal oxide plays a vital role in the field of catalysis. It is widely used as the primary catalyst, cocatalyst and carrier [51]. As the primary catalyst, metal oxide catalysts can be divided into transition metal oxide catalysts and main group metal oxide catalysts. Sulfide is the simplest inorganic sulfur anion, metal sulfide is a sulfur anion combined with metal or semi-metal positive ions to form M\nx\nS\ny\n compounds, and bimetallic sulfide is A1-x\nB\nx\nS\ny\n, where x and y are integers [52,53]. To date, hundreds of metallic sulfide materials have been discovered, many of which have simple structures and a high degree of symmetry. In addition, they have excellent physical and chemical properties. Metal sulfide can be formed by the reaction of sulfur with metal to form binary compounds or by the reaction of hydrogen sulfide (or hydrosulfuric acid) with metal oxides or hydroxides. Metal sulfide is generally colored solid and insoluble in water. Only alkali metal sulfide and (NH4)2S are readily soluble in water, and a few alkali earth metals are slightly soluble in water, such as CaS, SrS, BaS and others soluble in water. In analytical chemistry, various sulfides can be identified and separated according to their solubility differences in water and their corresponding characteristic colors. Alkali earth metal sulfide is a specific solubility, such as N2S, K2S, ZnS, MgS, FeS, MnS are soluble in dilute acid, while PbS, CdS, Sb2S3, SnS, Ag2S, CuS, HgS are insoluble in dilute acid. Recently, researchers have devoted a great deal of energy to studying transition metal sulfide catalysis in many catalytic fields. So far, dozens of metal sulfides and metal sulfide nanocomposites have been reported for various catalytic fields. Fig.\u00a01\n shows the key keywords of metal sulfide application in the environmental field in recent years based on the Web of Science database, which represents the focus direction and route of current research. In addition, it is an excellent photocatalyst because the conduction band of metal sulfide catalyst is composed of d and sp orbitals, and the valence band is composed of S 3p orbitals, which is much more negative than O 2p orbitals [54]. The narrow band gap provides an excellent response to the entire solar spectrum compared to the oxide materials [55].It is well acknowledged that natural metal sulfides (e.g., pyrite, mackinawite, chalcocite, chalcopyrite, molybdenite, etc.) were extensively distributed on the earth. However, in targeted applications and academic studies, scholars are more inclined to choose synthetic metal sulfides for decontaminating the aquatic environment due to their high purity, high reactivity, and good dispersion. Compared with natural ores, synthetic metal sulfide is more beneficial for researchers to process and analyze. Currently, various methods have been employed to synthesize metal sulfides, including hydrothermal and solvothermal, template, precipitation, thermal composition, electrochemical, etc., summarized in Table\u00a01\n. Hence, the following sections will describe the available synthesis routes in more details.Hydrothermal processes can be defined as any homogeneous or heterogeneous chemical reaction performed under sealed high temperature and pressure conditions using water or organic solvents, in which the reactants can dissolve, and the resulting products are insoluble [56]. The method shows extraordinary potential for the preparation of advanced materials (bulk single crystals, fine particles, and nanoparticles) since the as-synthesized samples possess the characteristics of high purity, excellent dispersibility, good uniformity, and proper crystal shape.Zhang et\u00a0al. [57] successfully fabricated the Co3S4 nanoparticles in the autoclave and kept the condition at 180\u00a0\u200b\u00b0C for 24\u00a0\u200bh. Similarly, the microwave hydrothermal method employing polyvinylpyrrolidone (PVP) as the surfactant was used to prepare doughnut-like CuS particles [58]. This strategy also applies to the preparation of other sulfides, such as CuFeS2, MoS2, NiCo2S4, and CuCo2S4 [45,59]. Meanwhile, the developments of multiple combination processing of catalysts, such as mechanochemical-hydrothermal, electrochemical-hydrothermal, sonar-hydrothermal, etc., have contributed significantly to the synthesis of many high-activity catalysts [60\u201362].The template method refers to the synthesis of nanoparticles by using mesoporous matrix materials as templates, which is the most widespread and successful strategy for developing advanced materials [63]. For example, Kim et\u00a0al. [64] recently prepared the bulk MoS2 using an anodic aluminum oxide template with a hole size of 80\u00a0\u200bnm and neck width of 10\u00a0\u200bnm. Tang et\u00a0al. [65] grew NiS2 on carbon cloth, which was used as an efficient 3D hydrogen evolution cathode in neutral solutions. Similarly, the soft template method is also deemed an effective strategy, which has been applied to fabricating various metal sulfides. Jiang et\u00a0al. [66] synthesized the CuS hollow spheres through a facile microemulsion template route at room temperature using copper naphthenate as the metal precursor and thioacetamide as the source of S2\u2212.However, hard and soft templates have many disadvantages, such as cumbersome preparation procedures, difficulty in eradicating the template, tedious preparation time and high cost, which limit the large-scale applications of samples to a certain extent [67]. Therefore, the self-assembly and template-free methods attracted much attention. For example, as shown in Fig.\u00a02\n(a), Tu et\u00a0al. [68] reported that the hierarchically ZnIn2S4 nanosheet-constructed microwire arrays could be received via the template-free strategy. The samples were constructed by vertical nanosheets with about 1\u20135\u00a0\u200b\u03bcm diameters and more extensive than 10\u00a0\u200b\u03bcm in average length, which preferentially exposed (006) facets. In addition, Yu and co-workers [69] demonstrated a facile biomolecule-assisted one-pot route toward fabricating novel CdS/MoS2/graphene hollow spheres, regarded as the high-efficiency and low-cost photocatalysts for hydrogen evolution, owing to the unique hollow-shaped structure and enhanced charge separation ability. At the same time, as a new kind of porous crystal material, metal\u2013organic\u00a0\u200bframeworks (MOFs) are also widely used as precursors for synthesizing metal sulfide. Wu et\u00a0al. [70] first synthesized the cobalt-containing MOFs precursor ZIF-67 by a simple agitation method and then successfully converted the MOFs precursor into a hollow amorphous CoS\nx\n hexagonal cage in situ by adding thioacetamide. The authors further utilized this strategy to synthesize bimetallic MOFs precursors by substituting sectional Co(II) for M(II) (M\u00a0\u200b=\u00a0\u200bMn, Ni, Cu, Zn) [71]. Hollow bimetallic sulfide polyhedral were prepared by hydrothermal sulfurization in TAA solution.The mixture of metal salts should be continuously heated under certain conditions during the thermal decomposition process to obtain the target products. Deng et\u00a0al. [72] prepared the CoS2/CC (commercial carbon cloth) nanoparticles by thermal composition method, in which the precursor Co3O4/CC and sulfur powder were placed in the backward position and upstream side of the porcelain boat, respectively. The temperature of the tube furnace was quickly elevated to 450\u00a0\u200b\u00b0C in 30\u00a0\u200bmin and kept for 120\u00a0\u200bmin in N2 atmosphere and the as-synthesized catalyst functioned as a 3D flexible electrode for water oxidation. Davar et\u00a0al. [73] also fabricated CuS nanoparticles via a facile and low-temperature thermal decomposition method. In addition, as shown in Fig.\u00a02(b), Vu and his colleagues [74] successfully prepared NiS by utilizing Ni(NO3)2 and thioacetamide as the precursors, and the materials were incorporated into an electrochemical sensor.The precipitation method has also been extensively used to synthesize solid catalysts. Both ion salt and precipitating agent are necessary for the precipitation, and the formed precipitate should be washed, dried and calcined under specific conditions. In the precipitation method, a series of factors will impact the properties of the precipitation products, such as the types of precipitating agent and their concentrations, precipitation temperature, precipitation pH, stirring speed and feeding order, so the operational conditions are essential to be optimized [75]. Kumar et\u00a0al. [76] reported a route for the preparation of CdS nanoparticles with the precipitation method using the equimolar (0.1\u00a0\u200bM, 20\u00a0\u200bmL) solutions of cadmium acetate (Cd(OAc)2) and Na2S as precursors. Rafiq et\u00a0al. [77] used 2-mercaptoethanol as a capping agent, homogenous solutions of 0.1\u00a0\u200bM Zn(NO3)\u00b74H2O and Na2S\u00b75H2O, to prepare ZnS quantum dots. They found that the synthesized ZnS quantum dots have excellent potential in dye degradation. The precipitation method attracted researchers due to some advantages, such as mild reaction conditions and simple reaction control.In this process, the powder particles are subjected to severe mechanical deformation by the collision with the ball in the stainless-steel container and constantly broken, cold welding and fracture, resulting in the solid-state reaction and mechanical chemical reaction in the powder blend. This low-cost, easily scalable mechanical-chemical route has prepared uniform 5\u00a0\u200bnm-sized CuS quantum dots (QDs) [78]. as shown in Fig.\u00a02(c), CuS quantum dots have a high surface area with dislocation and planar defects, such as twinning and stacking defects, which are conducive to electrocatalytic and photocatalytic performance. In addition, several nickel sulfide powders, such as Ni3S2, Ni7S6, Ni\nx\nS6 and Ni3S4, were prepared by ball milling with NiS powder as raw material [79]. Ambrosi et\u00a0al. [80] prepared a highly active MoS2 electrocatalyst using thin MoS2 slices. Ball milling improves the electrochemical and electrocatalytic properties of MoS2. In addition, CuCrS2 and NiCr2S4 were synthesized by mechanical alloying using ball milling technology [81], and the materials were sintered and treated to enhance electrical conductivity.The electrochemical method can electrodeposit the materials on the electrode surface by providing a constant voltage or current to the charged particles to induce directional movement of charged particles. The transfer of electrons usually accompanies the process of preparing materials, and the cost should be considered. Zhao et\u00a0al. [62] reported that the ternary mixed metal Ni\u2013Co\u2013Fe sulfides based on three-dimensional (3D) nickel foam (NiCoFeS/NF) could be fabricated by a facile electrodeposition-solvothermal process, and the metal sulfides were used for efficient electrocatalytic water oxidation in alkaline media. Wang et\u00a0al. [82] demonstrated a single-step potentiostatic method for the electrodeposition of Cu2S nanoparticles onto fluorine-doped tin oxide electrodes from CuCl2 and thiourea aqueous solution to develop counter electrodes for quantum-dot-sensitized solar cells.Microwave is a kind of electromagnetic wave, which, like high-frequency electromagnetic waves, is generated by the periodic change of electric and magnetic field energy in the electromagnetic oscillation circuit. Microwave radiation usually refers to an electromagnetic wave with a frequency of 300\u2013300,000\u00a0\u200bMHz and a wavelength of less than 1\u00a0\u200bm. According to their wavelength, microwaves can be divided into decimeter waves, centimeter waves and millimeter waves. Microwaves are generated by magnetrons. Microwave radiation is much weaker than infrared radiation and needs to be processed before it can be received using a receiver. Microwave heating drivers can only accept limited 2.45\u00a0\u200bGHz frequencies. Through microwave radiation heating, the microwave generates regular interactions with the material to convert electromagnetic energy into heat energy. Heat is generated from within the material, as opposed to traditional heating methods, which transport heat from the outside to the inside. This internal heating shortens reaction times and saves energy. Therefore, compared with traditional methods, microwave irradiation is faster and more efficient. Chen et\u00a0al. [83] prepared CdS, ZnS, CoS, PbS, CuS, Bi2S3, Sb2S3, Ag2S and other metal sulfides by microwave method. Glycol was used as the solvent. Finally, various metal sulfides were successfully prepared.The wet chemical method has also been applied to fabricate metal sulfides. Chen et\u00a0al. [84] fabricated the ZnS rods via the wet chemical method under reflux conditions. They used Zinc acetate and thioacetamide as raw materials, and thioglycolic acid acted as a capping agent. Similarly, microwave irradiation is employed to prepare metal sulfides, providing a homogeneous heating process for rapidly synthesizing nanocrystals with controllable size and shape [67]. CuS/Graphene composite was obtained under microwave irritation, Cu(NO3)2 and Na2S2O3 were used as the copper source and the sulfur source, respectively [85]. The cation exchange method may also be an excellent choice, and the technique was used by Zhu et\u00a0al. [86], who prepared (Ag, Cu)2S hollow spheres with a diameter of 700\u00a0\u200bnm to 1\u00a0\u200bm with the molar ratio of CuS to Ag+ 2:1\u00a0\u200bat room temperature.It is known that the different processes will result in various products with different shapes and catalytic performances. Table\u00a02\n displays the main differences among other preparation methods, all the preparation methods have certain limitations, and the structure of the catalyst highly depends on the preparation method, which is an essential factor that influences the catalytic activity of the established catalytic oxidation system.In general, even if the as-synthesized materials are the same, there will still have some differences in morphology, structure and catalytic activity due to different preparation conditions. Herein, we take CuS as an example to compare the differences in as-synthesized samples prepared under different conditions (seen in Table\u00a03\n). As expected, the preparation conditions significantly impact the morphological structure of the materials, which may further affect the catalytic performance and reusability of the materials. Therefore, it is essential to utilize effective characterization techniques to help us better understand the basic structure and morphology of the catalysts and further clarify the mechanisms of the catalyst-induced decontamination reaction.The surface morphology, specific functional groups, chemical compositions, thermal properties and crystal phase of as-synthesized samples will directly or indirectly influence the catalytic performance of catalysts. The frequently used characterization methods for diverse metal sulfides are summarized in Table\u00a04\n. X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), Brunauer-Emmett-Teller specific surface area (BET), Fourier transformed infrared spectroscopy (FTIR), photoluminescence (PL) spectra and thermogravimetric analysis (TGA) are the commonly used methods for identifying the chemical compositions, analyzing the phase structure, measuring the specific surface area and pore size, clarifying surface functional groups, determining the surface vacancies and transfer, as well as recombination of photogenerated electron-hole pairs and revealing the thermal properties, respectively. Thus, the following sections will focus on the six widely used characterization technologies to further understand the correlation between catalyst characteristics and catalytic performances.X-ray of specific energy is used as the excitation source to irradiate the surface of materials and excites photoelectrons, and then uses electron energy analyzer to detect photoelectrons according to different energy distributions. The composition and oxidation state can be obtained according to the deconvolution of the peak position and intensity of each element, and the relative content of elements can be determined. According to different studies, the peak positions of each bonding configuration can vary in a certain range, possibly owing to their various preparation environments [36,87\u201390]. From XPS analysis, the mechanism of catalyst activity and the reasons for its inactivation and poisoning can be acquired. However, XPS is not a destructive tool for the qualitative and quantitative study of the chemical state and composition of the elements, which exist on the thin surface of the samples (up to 5\u00a0\u200bnm) [67].XRD is one of the most powerful tools for accurately revealing the phase composition, grain size and crystal structure of catalysts. The diffraction pattern of the powder is usually obtained by Debye Scherrer and Guinier methods. The material composition can be qualitatively examined by comparing the position of the diffraction peak with the powder diffraction file. The grain size can be determined by fitting the peak widths based on the Debye-Scherrer equation, and the crystallinity can be received according to the diffraction peak shape and area. Besides, the Scherrer equation calculates the average size of the fresh and used catalysts. Thus, it is favorable to understand the reason for deactivation better. The phase structure of metal sulfides (single, double and ternary components) can be identified by XRD tests [91\u201395].It is well known that most heterogeneous catalysts are porous materials and their pore structure, size and pore volume highly depend on the catalyst preparation methods. The specific surface area represents the total surface area per gram of catalyst (m2/g), which is an essential parameter for evaluating catalytic properties. The gas adsorption Brunauer-Emmett-Teller (BET) methods are one of the most commonly used. It is a multi-molecular layer adsorption formula based on the classical statistical theory of BET. The specific surface area is related to particle size, shape, surface defects and pore structure, which affects the chemical and physical properties of the as-synthesized materials.The specific surface area is greatly affected by different preparation conditions. Generally, smaller particle sizes result in larger surface areas and enhance the surface adsorption ability and catalytic performance of as-prepared samples. Zhang et\u00a0al. [57] found that the hydrothermal-synthesized nanosheets possessed a higher surface area (62.8\u00a0\u200bm2/g) than that of the Co3S4 nanoparticles (32.1\u00a0\u200bm2/g), which may be favorable for the improved H2 evolution on CdS/Co3S4. Chen et\u00a0al. [84] revealed that the BET value of the ZnS increased as the thioglycolic acid content elevated. Furthermore, commercial mackinawite has been proven to degrade p-chloroaniline efficiently [87].Infrared spectroscopy is a common means to identify the molecular structure and determine the surface properties of catalysts. Quantitative analysis can also be carried out for individual components or mixtures of various components, especially for samples that are difficult to separate and cannot find significant characteristic peaks in the ultraviolet and visible regions. In addition, the structure of the unknown substance can be concluded according to the position and shape of the absorption peak in the spectrum. The intensity of the absorption peak can identify the content of each component for composites. Similarly, FTIR is beneficial for the investigation of catalytic mechanisms. For example, by ATR-FTIR analysis, Zhou et\u00a0al. [84] studied the role of sulfur conversion in sulfate radical generation of PMS/FeS2 system. Meanwhile, Zhu et\u00a0al. [96] found the graphene-supported hollow cobalt sulfide nanocrystals for PMS activation were highly efficient for bisphenol A elimination from the FTIR study.PL spectra have been extensively employed as a proven technique for studying the surface vacancies, transfer, and recombination of photogenerated electron-hole pairs of composites [97]. Ayodhya et\u00a0al. [98] revealed that the synergistic effect of CuS and rGO would significantly restrain the recombination of hole-electron pairs and thus result in an enhanced separation of a photogenerated carrier by PL spectra analysis. Besides, the different emission phenomena from the PL spectra may also ascribe to the different morphology, sizes, and crystalline. Likewise, by comparing the PL spectra of CdS and CdS/rGO nanocomposite, Sagadevan et\u00a0al. [99] found the immobilization of CdS nanopowders on the rGO sheets decreased the PL intensity, suggesting that the efficient charge separation process took place inside the composite matrix.Thermal analysis is defined as the technique for revealing the physical properties of a substance or its reaction product, which are measured as a function of temperature [100,101]. Differential scanning calorimetry (DSC), thermogravimetry (TG/DTG) and differential thermal analysis (DTA) are the most commonly used thermal analysis techniques, which have been successfully performed to investigate the thermal behaviors of heterogeneous catalysts, including the interaction between metal active components and carriers, the coordination state and distribution of the active metal ions, the deactivation mechanism of catalyst, the phase transitions, as well as the thermally induced chemical reactions and decompositions [43,50,102]. The thermal stability and structural changes of the prepared CuS and rGO-capped CuS composite were estimated through TGA in the range of temperature 30\u00b0C-1000\u00a0\u200b\u00b0C [98]. In addition, the thermal analysis method was successfully applied to choose the ideal calcination temperature of urchins-like cadmium zinc sulfide nanostructured particles [103].Electron microscopy technology can provide information about the interaction between electrons and target materials. Microscopic information about the catalysts can be obtained after data conversion, amplification, and other processing. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) are frequently used techniques for comparative analysis of morphological changes of fresh and used catalysts, and the energy dispersive spectroscopy (EDS) is used to explore the change of element content of catalyst [102,104\u2013110]. Besides, the atomic force microscopy (AFM) technique is regarded as a powerful tool that can also determine morphology (e.g., size, surface texture, and roughness) and electrical properties of as-prepared catalysts with high spatial resolution, thus becoming more and more widely utilized in material characterization.Moreover, Raman spectroscopy is a non-destructive method requiring no special sample preparation and nondestructive analysis and presenting analytical advantages such as speed, cost, sensitivity, repeatability and stability of analyses, which plays a vital role in the structural characterization of metal sulfides, mainly providing valuable information about phase composition, crystal symmetry and orientation, quality of the crystal and the total amount of the substance. Furthermore, compared with infrared absorption spectroscopy, the recording of Raman spectroscopy does not require special sample preparation and nondestructive analysis, the fiber optic probe brings the excitation laser into the sample and transmits the scattered light to the spectrometer, enabling remote detection of the Raman signal [111]. Xing et\u00a0al. [44] confirmed the reduction capability of the exposed Mo4+ by analyzing the Raman spectra of MoS2 and Fe3+.Temperature-programmed desorption (TPD) involves heating a sample at raising the temperature and followed by detecting the released gas using a suitable device [112]. Quantitative analysis of TPD is widely employed to clarify the desorption strength of heterogeneous catalysts, particularly for the supported and multi-component metal catalysts, by performing experiments at different heating rates and then accompanied by theoretical treatment processing. For example, O2-TPD was introduced to investigate the effect of substitution-induced oxygen-ion vacancies, and H2-TPD was utilized to study active metal surface area on supported particles [113]. In addition, electrochemical analysis is occasionally used due to the excellent conductivity of metal sulfides. Cyclic voltammetry scan and electrochemical impedance spectra (EIS), as well as Tafel plot, are the most commonly used methods [43,45,47,49,74,89,114]. Understanding the electrochemical corrosion characteristics and electron and charge transfer ability of samples is helpful. For example, the excellent electrical conductivity of iron sulfides has been unveiled by the CV cycle profile and the typical Nyquist plots [115].Chiu et\u00a0al. [116] investigated the heterogeneous interface's lattice, atomic structure and composition by using aberration-corrected scanning transmission electron microscopy (AC-STEM) at the interplanar junction of WS2/WSe2 and WSe2/MoS2. AC-STEM observations of heterojunction surfaces indicate that the in-plane heteroepitaxy growth is initiated by replacing sulfur atoms at the edge of the pre-grown transition metal dichalcogenides. Pan et\u00a0al. [117] used synchrotron X-ray absorption near edge spectroscopy (XANES) to elucidate the electronic structure changes of CuCo2S4 NSs at unique core/shell electrodes. Furthermore, the corresponding Fourier transform (FT) k3\u03c7(k) function of the extended X-ray absorption fine structure (EXAFS) spectroscopy further confirmed the successful vulcanization of CuCo2O4 NPs into CuCo2S4 nanosheets. Through atomic-resolution scanning tunneling microscope (STM) simulation images with energy wave function integration below 0\u00a0\u200beV and 1.5\u00a0\u200beV, Liu et\u00a0al. [118] observed that two simulation images of single-layer MoS2 with 2S, 3S, and 4S vacancy chains showed almost the same black dot characteristics around the defects because any atoms do not occupy the defect.In addition, in situ characterization has become an indispensable characterization method in catalytic reactions, which not only helps us to further explore the catalytic mechanism but also makes outstanding contributions to the further design of various effective catalysts. In order to reveal the synergistic effect of 1T and 2H phases in MoS2 in the photo Fenton oxidation process, Chen et\u00a0al. [119] detected the change of Mo chemical state under light irradiation by using solid EPR in operando. The EPR technique within operando has the potential to recognize the presence of Mo(III) instantaneously. By in-situ EPR, we demonstrate that the optimized 2H/1T heterojunction allows interfacial electron transport from the semiconductor 2H to the metal 1T phase and synchronizes partial reduction Mo(IV) to Mo(III) at the interface.Up to now, the catalytic mechanisms involved in metal sulfide-based AOPs have been exclusively revealed as the radical and non-radical pathways. Previous findings have suggested that metal sulfides could function as an effective activator for organic pollutants elimination in water and wastewater. According to the previous findings, there are two general mechanisms commonly proposed by metal sulfide-based AOPs. One is the improvement of the electron transfer efficiency resulting from the reductive S2\u2212 on the catalyst surface [45,46,87,120]; the one is the unsaturated S atoms on the surface of the metal sulfide can capture protons resulting in the formation of H2S and exposing metal active sites with reducing properties, thus accelerating the rate-limiting reaction step [44,91]. However, there are still some differences in various metal sulfide-based oxidation systems. As shown in Table\u00a05\n, some reported metal sulfide-AOPs are summarized. Therefore, a deep understanding of the catalytic performance and mechanisms of the target AOPs is beneficial to developing a more robust system for organic pollutants removal.In the following part, we will discuss the catalytic performance and the mechanism for mono-metal metal sulfides-based catalytic oxidation systems, such as Fe\nx\nS\ny\n, CuS, Co\nx\nS\ny\n, and MoS2 to reveal the superiority of the metal sulfides-based system.Pyrite (FeS2) is an abundant sulfide mineral containing structural Fe2+ and has been regarded as an electron donor for the removal of pollutants and an activator of either oxygen or peroxides. Feng et\u00a0al. [121] studied the catalytic oxidation of 1,4-dioxane in FeS2/PMS system. They found that nearly 100% degradation of 1,4-dioxane (50\u00a0\u200bmg/L) was obtained after 40\u00a0\u200bmin, and the removal efficiency was higher than the FeS2/PDS and FeS2/H2O2. They systematically studied the role of disulfides and activation sites and proposed the possible reaction mechanism of FeS2/PMS system (Fig.\u00a03\n(a)). The reaction mechanisms include the following aspects: 1) the direct activation of PMS by FeS2 can be ignored; 2) a synergistic effect was observed, which was attributed to the controlled activation of PMS by Fe2+ generated by the reaction of FeS2 with Fe3+; 3) electron transfer from S2\n2\u2212 on pyrite surfaces to Fe3+ on or near the surfaces of pyrite; 4) PMS reacted with the S2\n2\u2212 on the surface of pyrite to release Fe2+, which then activated PMS rapidly to generate radicals; 5) \u2022OH and SO4\n\u2022\u2212 were involved in the pyrite-PMS system and \u2022OH were the main active species. Thus, the combination of FeS2 and PMS could effectively remove target pollutants.FeS/PS system was also used to eliminate 2,4-dichlorophenoxyacetic acid (2,4-D) [122]. The authors found that 2,4-D could be efficiently removed and mineralized by FeS/PS system. The quenching experiments implied that both \u2022OH and SO4\n\u2022\u2212 were responsible for 2,4-D degradation, but \u2022OH was the dominant active species. The quenching tests and EPR analysis have proposed the possible mechanism of 2,4-D oxidation in FeS/PS. PS would decompose to yield SO4\n\u2022\u2212 by both Fe(II) and Fe2+, then the formed SO4\n\u2022\u2212 in the aqueous solution or on the surface of FeS could directly react with 2,4-D to generate \u2022OH via the oxidation of H2O (or \u2022OH). Subsequently, the hydrolysis of S2O8 would generate H2O2 and further lead to the Fenton reaction in the presence of Fe(II) (or Fe2+), then the generated SO4\n\u2022\u2212 and \u2022OH would result in efficient degradation of 2,4-D.The formation of Fe sulfides usually occurs in anoxic deposits, where sulfides are produced by the metabolic activity of sulfate-reducing bacteria [120]. Yuan et\u00a0al. [123] studied the degradation performance of p-chloroaniline (PCA) by persulfate activated with ferrous sulfide ore particles, FeS was proposed as an alternative electron donor, which could act as a continuous-releasing source of dissolved Fe2+ and Fe(II), leading to the formation of SO4\n\u2022\u2212 and then initiate a series of radical chain reactions. However, Fan and his co-authors [87] found that the FeS/PS system displayed excellent potential for PCA degradation and mineralization across an extensive initial pH range (3.0\u201311.0). The results were ascribed to the independent oxidations of Fe(II) and S(-II) and the regeneration of Fe(II) induced by S(-II) at the FeS surface, the clarified surface reaction catalytic mechanisms of FeS/PS system were different from Yuan et\u00a0al. [123], because \u2022OHfree, SO4\n\u2022\u2212\nfree diffusing from the FeS surface mainly contributed to PCA elimination.Pirgal\u0131o\u011flu et\u00a0al. [124] investigated the CuS/O3 system for treating aqueous single-dye solutions. They found that CuS increased the oxidation rates of ozonation side-products via a decrease in TOC values of the treated dye solution, and they discussed the catalytic ozonation kinetics and mechanism in detail in three parts: 1) the measurement of copper ions present in the liquid phase; 2) the homogeneous catalytic effect of liquid phase copper ions on ozonation efficiency; 3) the effect of CuS on the enhancement of ozone decomposition to generate more hydroxyl radicals. Nekoueia et\u00a0al. [125] demonstrated that the synthesized CuS hollow nanospheres@N-doped CNCs hybrid composites had the outstanding potential for enhanced adsorptive removal and catalytic oxidation of ciprofloxacin (CIP) when PMS was present, and the mechanism of PMS activation was as follows: firstly, after PMS was added into the catalytic system, H2O molecules were adsorbed on the Cu(II) part of the as-prepared catalyst to produce \u2261Cu(II)\u2013OH according to Eq. (1), and then Cu(II) reacted with HSO5\n\u2212 to produce \u2022OH. Furthermore, the \u2261Cu(II)\u2013OH reacted with HSO5\n\u2212 resulting in \u2261Cu(II)\u2212(OH)OSO3\n\u2212 generation and further decomposed to SO4\n\u2022\u2212 via Eqs. ((2)\u2212(3)).\n\n(1)\n\u2261Cu(II)\u2212 \u2013OH\u00a0\u200b+\u00a0\u200bHSO5\n\u2212\u2192 \u2261Cu(III)+ SO4\n2\u2212\u00a0\u200b+\u00a0\u200b\u2022OH\n\n\n\n\n(2)\n\u2261Cu(II)\u2212 \u2013OH\u00a0\u200b+\u00a0\u200bHSO5\n\u2212 \u2192 \u2261Cu(II)\u2212(OH)OSO3\n\u2212\u00a0\u200b+\u00a0\u200bOH\u2212\n\n\n\n\n\n(3)\n\u2261Cu(II)\u2212 (OH)OSO3\n\u2212 \u2192 \u2261Cu(III)\u00a0\u200b\u2212\u00a0\u200b\u2013OH\u00a0\u200b+\u00a0\u200bSO4\n\u2022\u2212\n\n\n\nSimilarly, the CuS mineral was selected as an activator to persulfate (PS) for eliminating atrazine (ATZ), and 91.6% ATZ removal was obtained in 40\u00a0\u200bmin under acidic pH [126]. They proposed the one-eight-lives method to evaluate the kinetic of CuS/PS system for ATZ degradation, which the following equation could conclude: \u2212d[ATZ]/dt = (2.985\u00a0\u200b\u00d7\u00a0\u200b10\u22124 mmol/(L\u00b7min)) [ATZ]0.023[PS]0\u00b776318[CuS]0.80801. Based on the control experiments, they found that the CuS/PS system was superior to other systems (Cu2+/PS and CuO/PS system). Furthermore, the possible reaction mechanism of CuS/PS system was proposed as follows according to the radical scavenger tests and EPR study. Specifically, PS was first absorbed on the surface of CuS. Then PS was activated, followed by the generation of main ROS (SO4\n\u2022\u2212 and \u2022OH) for ATZ removal. Simultaneously, the PS would induce the oxidation of CuS to produce Cu2+ and S(-II), S(-II) could not promote the degradation of ATZ, which was further oxidized to the different intermediate valence of sulfur species. Besides, the leached Cu2+ could also activate PS to produce SO4\n\u2022\u2212 and \u2022OH in the liquid phase, but the role of the homogeneous process was insignificant.It was reported that the fabricated graphene nanosheet-supported hollow cobalt sulfide nanocrystals (Co3S4@GN-X and CoS@GN-X, X in accordance to the GO weight) via a template method (zeolitic imidazolate frameworks) were able to activate PMS to degrade bisphenol A (BPA) [96]. The mechanisms for enhanced BPA degradation efficiency are shown in Fig.\u00a03(b). Firstly, PMS was activated by CoS to produce abundant SO4\n\u2022\u2212. Secondly, for the pristine CoS with poor removal efficiencies and electron transfer, the formed SO4\n\u2022\u2212 was inclined to accumulate and then diffuse out of the catalyst surface after reaching saturation, as well as be converted into \u2022OH by H2O/OH\u2212. For CoS@GN-60, the existence of graphene provided adsorption territories and high-speed electron flow for BPA elimination. In contrast to \u2022OH, SO4\n\u2022\u2212 was more likely to degrade BPA molecules through the charge-transfer mechanism. Therefore, once SO4\n\u2022\u2212 was generated, it would bound onto the catalyst by electrostatic force and then quickly capture electrons from the absorbed BPA across graphene since it functioned as the conductor. Thus, no excessive SO4\n\u2022\u2212 was released into the bulk solution to combine with DMPO when conducting EPR tests. Besides, the \u2022OH generated by the side reaction of SO4\n\u2022\u2212 could combine with H2O/OH\u2212 of the bulk solution, which restrained SO4\n\u2022\u2212 transfer. Consequently, the combination of sulfides and the suitable carrier is more conducive to the degradation of target contaminants. In summary, sulfides combined with the excellent performance of the carrier are more conducive to removing target pollutants. In addition, another study also found a key role in hollow structures. Wu et\u00a0al. [70] studied the effect of the hollow structure of the catalyst. A new hollow amorphous CoS\nx\n hexagonal cage catalyst was prepared by an aqueous solution-assisted solvothermal method, which activated PMS effectively to degrade antibiotics through advanced oxidation processes. The results show that the prepared hollow amorphous CoS\nx\n cage exhibits excellent tetracycline (TC) decomposition ability under PMS activation, which is far superior to the conventional Fenton reaction of solid Co3O4 and CoS. The excellent catalytic performance of PMS activated CoS\nx\n is due to the Co3+/Co2+ and S2\u2212/S2\n2\u2212 cycles and the existence of hollow structures.Ridruejo et\u00a0al. [127] demonstrated that the complete removal of acid tetracaine could be achieved by electro-oxidation and photoelectron-Fenton processes with a boron-doped diamond anode at 100\u00a0\u200bmA/cm2. The results implied the viability of the manufactured CoS2-based cathode was highly suitable for water treatment since the use of an air-diffusion cathode containing CoS2 nanoparticles could enhance the electro-generation of H2O2. Moreover, Yin et\u00a0al. [128] successfully established a strategy to prepare the quasi-single cobalt sites in the nanosized pores of SBA-15 (QS-CoS). Their findings suggested that the QS-CoS catalysts were highly efficient for PMS activation due to the confined space, and abundant silicon hydroxyl groups in the as-synthesized SBA-15 contributed to the generation of the resultant quasi-single cobalt sites in the form of Co\u2013O\u2013Si. Meanwhile, both SO4\n\u2022\u2212 and \u2022OH were produced in QS-CoS/PMS system, SO4\n\u2022\u2212 was the dominant one responsible for phenol degradation. The possible mechanisms could be described as follows according to Eqs. ((4)\u2212(7)).\n\n(4)\nCo2+\u2013O\u2013Si\u00a0\u200b+\u00a0\u200bHSO5\n\u2212 \u2192 Co3+\u2013O\u2013Si\u00a0\u200b+\u00a0\u200bSO4\n\u02d9\n\n\u2212\n\u00a0\u200b+\u00a0\u200bOH\u2212\n\n\n\n\n\n(5)\nCo3+\u2013O\u2013Si\u00a0\u200b+\u00a0\u200bHSO5\n\u2212 \u2192 Co2+\u2013O\u2013Si\u00a0\u200b+\u00a0\u200bSO5\n\u02d9\n\n\u2212\n\u00a0\u200b+\u00a0\u200bH+\n\n\n\n\n\n(6)\nSO4\n\u2022\u2212\n\u00a0\u200b+\u00a0\u200bH2O \u2192 SO4\n2\u2212\u00a0\u200b+\u00a0\u200b\u2022OH\u00a0\u200b+\u00a0\u200bH+\n\n\n\n\n\n(7)\nSO4\n\u2022\u2212\n\u00a0\u200b+\u00a0\u200b\u2022OH\u00a0\u200b+\u00a0\u200bC6H5OH \u2192 several steps \u2192 CO2 +H2O\u00a0\u200b+\u00a0\u200bSO4\n2\u2212\n\n\n\nXing and co-workers [44] revealed that commercial 2H-type MoS2 could function as an excellent co-catalyst to signi\ufb01cantly enhance the decomposition ef\ufb01ciency of H2O2 by 47.2% and significantly decrease the consumption of H2O2 (0.4\u00a0\u200bmmol/L) and Fe2+ (0.07\u00a0\u200bmmol/L) in AOPs . The mechanism of the catalytic reactions used MoS2 as the co-catalyst is shown in Fig.\u00a03(c). At \ufb01rst, the protons of the solution were captured by the unsaturated S atoms on the surface of MoS2 and led to the formation of H2S, which was clearly identified by EPR test, because the MoS2 powder showed a signal at g\u00a0\u200b=\u00a0\u200b2.0, implying the existence of sulfur vacancies. Subsequently, the surface Mo4+ was oxidized to Mo6+ as well as followed by the reduction of Fe3+ to Fe2+ (Eq. (8)), which significantly enhanced the original rate-limiting of Fe3+/Fe2+ conversion in conventional AOPs (Eq. (9)). In the Fenton reaction, Mo6+ was further reduced to Mo4+ with the assistance of H2O2 to ensure the catalytic cycling of MoS2 according to Eq. (10).\n\n(8)\nFe3+\u00a0\u200b+\u00a0\u200bMo4+ \u2192 Fe2+\u00a0\u200b+\u00a0\u200bMo6+\n\n\n\n\n\n(9)\nFe3+\u00a0\u200b+\u00a0\u200bH2O2 \u2192 Fe3+\u00a0\u200b+\u00a0\u200b\u2022O2H\u00a0\u200b+\u00a0\u200bH+\n\n\n\n\n\n(10)\nMo6+\u00a0\u200b+\u00a0\u200bH2O2 \u2192 Mo4+\u00a0\u200b+\u00a0\u200bH2O\u00a0\u200b+\u00a0\u200bO2\n\n\n\nFurthermore, they pointed out that the ef\ufb01ciency of the AOPs involving metal sul\ufb01de co-catalysts could be further improved by visible light illumination due to the light-induced sensitization of selected pollutants. Remarkably, the reaction rate constant increased by 18.5 times compared with conventional AOPs, when MoS2 and rhodamine B were used as co-catalyst and target pollutants, respectively. In addition, they studied the co-catalytic performance of all the chosen sul\ufb01des (MoS2, WS2, Cr2S3, CoS2, PbS and ZnS) and the ef\ufb01ciency of reduction of Fe3+ to Fe2+ was as the following order: WS2\u00a0\u200b>\u00a0\u200bCoS2\u00a0\u200b>\u00a0\u200bZnS\u00a0\u200b>\u00a0\u200bMoS2\u00a0\u200b>\u00a0\u200bPbS\u00a0\u200b>\u00a0\u200bCr2S3\u00a0\u200b>\u00a0\u200bconventional Fenton. Moreover, the MoS2 has excellent stability and reusability even after ten cycles, and the newly developed H2O2-based AOPs not only shows outstanding potential for TOC degradation (90%) but also significantly decreased the COD value of actual wastewater from 10,400\u00a0\u200bmg/L to 360\u00a0\u200bmg/L at a record-low dosage of H2O2 and Fe2+. The above results indicated that the co-catalytic effect of the metal sul\ufb01des was universal, and it would make significant progress in the practical application of transition metals involved AOPs for environmental remediation. Similar results were also observed by Dong et\u00a0al. [91]. They used the co-catalytic WS2 on the Fenton reaction to improve the decomposition of H2O2 for the reduction of Cr(VI) and remediation of Phenol synchronously, which further implies the general applicability of metal sulfides to the degradation of organic pollutants.Currently, metal sulfide catalysts studied in AOPs are usually nanomaterials, and it has been reported that the differences in the microstructure of metal sulfide greatly affect its catalytic performance. Zhu et\u00a0al. [129] investigated the degradation of aromatic organic compounds by 3D-MoS2 sponge loaded with MoS2 nanospheres and graphene oxide (GO) in AOPs. Exposure of Mo4+ active sites on 3D-MoS2 can significantly improve the concentration and stability of Fe2+ in AOPs, so that Fe3+/Fe2+ is in a stable dynamic cycle, thus effectively promoting the activation of H2O2/PMS. More importantly, the authors found that 2D-MoS2 solids in powder form are difficult to recover and reuse after degradation, which may increase costs and cause secondary pollution to the environment. In addition, this shortcoming may prevent the widespread application of MoS2 in industrialization. However, 3D materials have a larger specific surface area and hierarchical pore structures, which is more conducive to the adsorption of organic pollutants and electron transfer in the advanced oxidation process. Compared with traditional 2D-MoS2, 3D-MoS2 sponge is more conducive to industrial production, and the feasibility of its industrial application is confirmed by pilot experiments.CdS is regarded as the classic IIeVI semiconductor, whose direct band gap at room temperature is 2.43\u00a0\u200beV, which is regarded as the best semiconductor, and researchers have paid considerable attention to its potential application [130,131]. Yang and co-workers [132] developed an efficient visible-driven photo-Fenton system based on self-assembled CdS nanorods. The CdS/Fe2+ photo-Fenton system has highly degraded sulfamethazine (SMT) under visible light irradiation. The SMT of 20\u00a0\u200bmg/L is almost completely degraded within 90\u00a0\u200bmin, and the pH range is 4\u20138. The large number of photoelectrons produced in the system can reduce the dissolved O2 to H2O2 by direct two-electron reduction pathways and accelerate the conversion of Fe3+ to Fe2+.ZnS, as an economical and efficient environmental protection material, administered due to its under ultraviolet (UV) light, has good catalytic activity and high-efficiency theory, which can rapidly generate electron-hole pairs (e\u2212\u2212h+) under photoexcitation, so it has been widely researched [133,134]. However, it has a wide and direct band gap (3.80\u00a0\u200beV for wurtzite), a crucial factor inhibiting its visible light response. Tie et\u00a0al. [135] synthesized the novel nitrogen-doped ZnS microspheres by a simple one-step method. The prepared nitrogen-doped ZnS catalyst showed excellent photocatalytic activity for removing organic pollutants under natural sunlight and producing hydrogen by water cracking. The prepared catalyst showed good photodegradation performance against many organic pollutants such as methyl orange, methylene blue, Rhodamine B, ciprofloxacin and sulfa. Nitrogen doping improves the visible light absorption capacity and electron transfer efficiency of ZnS, thus improving the photocatalytic performance of the catalyst.Bismuth sulfide has attracted extensive attention recently due to its high photocatalytic activity in degrading organic pollutants. Among these bismuth sulfides, Bi2S3 has been extensively reported in the field of photocatalysis due to its unique optical properties and band gap (1.3\u00a0\u200beV) [136,137]. In recent years, the photocatalytic activity of Bi2S3-based NSs has attracted the attention of many researchers [138]. As shown in Fig.\u00a03(d), Gao et\u00a0al. [139] constructed an S-scheme heterostructure photocatalyst MoS2/Bi2S3/BiVO4 supported on 3D lignosulfonate modified poly(vinyl formal) sponges to enhance the synergic adsorption and photo-Fenton degradation of various fluoroquinolones in water. The synergistic adsorption of the carrier and the catalytic action of the photocatalyst significantly improve the applicable pH of the Fenton reaction (2.0\u20139.0) and reduce the dosage of Fe2+ to 0.014\u00a0\u200bmmol/L. Photogenerated electrons and the redox conversion of Mo(IV)/Mo(VI) and Bi(III)/Bi(V) accelerate the conversion of Fe3+ to Fe2+.Relative to the extensively investigated one-component metal sulfides, bimetallic sul\ufb01des have been demonstrated to be more remarkable owing to the excellent electrical conductivity and thermal stability as well as the synergistic effects of metal ions, which have been widely used as electrocatalysts for supercapacitor and OER [47,49,93,114]. However, only a few studies have been done on applying bimetallic sulfides in AOPs-based wastewater treatment. In the following part, we take sulfide carrollite (CuCo2S4) and chalcopyrite (CuFeS2) as examples to discuss the catalytic performance and the mechanisms in the catalytic oxidation system.As shown in Fig.\u00a04\n(a), The spinel sul\ufb01de carrollite (CuCo2S4) was introduced to activate PMS for the elimination of bisphenol S (BPS) in water [46]. It was verified that the catalytic activity and stability of hydrothermally fabricated CuCo2S4 during the oxidation process were superior to those of commonly used cobalt and copper oxides/sulfides. They also found that the neutral pH condition was most favorable for the degradation of BPS. Meanwhile, the synergistic surface redox couples of Cu(II)/Cu(I) and Co(III)/Co(II) played a vital role in the catalytic activation of PMS to produce SO4\n\u2022\u2212, and SO4\n\u2022\u2212 was proved to be the dominant oxidant species for the elimination of BPA by quenching experiments and EPR study.Furthermore, the high-resolution XPS and turnover frequency (TOF) tests proposed a possible process for forming oxidant species. They studied the changes of high-resolution XPS spectra of fresh and used CuCo2S4 in detail. The ratio of Co(III)/Co(II) increased from 0.88 to 1.07 after the reaction, suggesting that electrons transferred from cobalt to PMS in the activation process. However, the ratio of Cu(I)/Cu(II) decreased from 1.25 to 0.92, indicating that a part of Cu(I) has been oxidized to Cu(II). The appearance of Cu(I) and Co(III) in fresh CuCo2S4 implied that Cu could represent carrollite (I)Co2\n(III)(S4)(\u2212VII) [140], and the reactions involved in CuCo2S4/PMS can be described in the following equations (Eqs. (11)\u2212(15)), and the role of SO4\n\u2022\u2212 was excluded due to its much lower reducing potential (1.10\u00a0\u200bV). Eq. (13) was thermodynamically beneficial for (ECo(III)/Co(II)\u00a0\u200b=\u00a0\u200b1.84\u00a0\u200bV, ECu(II)/Cu(I)\u00a0\u200b=\u00a0\u200b0.15\u00a0\u200bV), proposing the synergistic effect between the two metal sites in CuCo2S4. Furthermore, TOF of as-prepared Co3O4, CuCo2O4 and CuCo2S4 were determined to be 1.31, 0.57 and 5.87\u00a0\u200b\u00d7\u00a0\u200b10\u22123\u00a0\u200bs\u22121, respectively. The TOF of CuCo2S4 was 5 and 10 times greater than Co3O4 and CuCo2O4, respectively. It was reported that binary cobalt sulfides have much lower optical band gap energy and much higher conductivity than the corresponding oxides [45,93]. More importantly, the replacement of oxygen with sulfur would produce a more \ufb02exible structure because the electronegativity of S2\u2212 was lower than O2\u2212.\n\n(11)\n\u2261Cu(I)\u00a0\u200b+\u00a0\u200bHSO5\n\u2212\u2192 \u2261Cu(II)\u00a0\u200b+\u00a0\u200bSO4\n\u2022\u2212\u00a0\u200b+\u00a0\u200bOH\u2212\n\n\n\n\n\n(12)\n\u2261Cu(II)\u00a0\u200b+\u00a0\u200bHSO5\n\u2212\u2192 \u2261Cu(I)\u00a0\u200b+\u00a0\u200bSO5\n\u2022\u2212\u00a0\u200b+\u00a0\u200bH+\n\n\n\n\n\n(13)\n\u2261Cu(I)\u00a0\u200b+\u00a0\u200b\u2261Co(III) \u2192 \u2261Cu(II)\u00a0\u200b+\u00a0\u200b\u2261Co(II)\n\n\n\n\n(14)\n\u2261Co(II)\u00a0\u200b+\u00a0\u200bHSO5\n\u2212\u2192 \u2261Co(III)+SO5\n\u2022\u2212\u00a0\u200b+\u00a0\u200bH+\n\n\n\n\n\n(15)\n\u2261Co(II)\u00a0\u200b+\u00a0\u200bHSO5\n\u2212 \u2192 \u2261Co(III)\u00a0\u200b+\u00a0\u200bSO4\n\u2022\u2212\u00a0\u200b+\u00a0\u200bOH\u2212\n\n\n\nFeCo2S4, as a bimetallic sulfide, not only has a strong synergistic effect between Co and Fe bimetallic sulfide but also has more metal active sites, which is conducive to the occurrence of various catalytic reactions [141,142]. In addition, FeCo2S4 also has excellent electrical conductivity, which makes the material in electrochemistry, photoelectric chemistry, energy storage and other fields of wide attention [143,144]. As demonstrated in Fig.\u00a04(b), Li et\u00a0al. [145] prepared FeCo2S4\u2013C3N4 nanomaterial as a novel multiphase catalyst for the activation and degradation of sulfamethoxazole (SMX) by PMS. 20\u00a0\u200bmg/L FeCo2S4\u2013C3N4 and 0.15\u00a0\u200bmM PMS can effectively degrade SMX (91.9%, 0.151\u00a0\u200bmin\u22121). Radical scavenger test and EPR analysis confirmed that the singlet oxygen (1O2) led nonradical pathway is the primary reaction mechanism of SMX degradation.Nie et\u00a0al. [45] verified that CuFeS2-PMS system displayed excellent activity for BPA removal compared with Cu2S, FeS2, CuFeO2, and Co3O4. It was found that BPA was almost eliminated (99.7%) and the TOC removal reached 75% within 20\u00a0\u200bmin. SO4\n\u2022\u2212 was confirmed to be the leading reactive species responsible for the BPA elimination by ESR and quenching tests. They pointed out that the as-synthesized CuFeS2 powders possessed rich active surface Cu+ and Fe2+, the sulfur species play a vital role in enhancing the reduction cycle of Cu2+/Cu+ and Fe3+/Fe2+. Thus, they proposed a possible mechanism for the CuFeS2-PMS system.As demonstrated in Fig.\u00a04(c), Cu+ and Fe2+ on CuFeS2 surface react with PMS to form SO4\n\u2022\u2212 (Eqs. (16)\u2013(17)). Based on the XPS spectra of the Fe 2p and Cu 2p, rich active Cu+ and Fe2+ species on fresh catalyst surface initiate PMS activation quickly, resulting in a quick degradation of BPA in the initial 1\u00a0\u200bmin in the CuFeS2-PMS system. Simultaneously, Cu+ and Fe2+ would be oxidized to Cu2+ and Fe3+, and SO4\n\u2022\u2212 reacted with H2O to form \u2022O2H (Eqs. (18)\u2013(19)). Subsequently, the surface Cu2+ and Fe3+ were reduced and recycled to Cu+ and Fe2+ (ECu\n2+\n/Cu\n+\u00a0\u200b=\u00a0\u200b0.17\u00a0\u200bV, EFe\n3+\n/Fe\n2+\u00a0\u200b=\u00a0\u200b0.77\u00a0\u200bV) via strongly reductive sulfur species such as S2\n2\u2212 (Eqs. (20)\u2013(23)). Furthermore, the regeneration of Fe2+ active sites was also carried out by the electron transfer from Cu+\u00a0\u200bto Fe3+ (Eq. (24)). The regenerated Cu+ and Fe2+ active sites on the CuFeS2 surface were capable of activating PMS again, leading to the continuous formation of reactive species. Their findings indicated that bimetallic metal sulfides have outstanding potential for oxidizing organic contaminants.\n\n(16)\nCu+\u00a0\u200b+\u00a0\u200bHSO5\n\u2212 \u2192 Cu2+\u00a0\u200b+\u00a0\u200bSO4\n\u2022\u2212\u00a0\u200b+\u00a0\u200bOH\u2212\n\n\n\n\n\n(17)\nFe2+\u00a0\u200b+\u00a0\u200bHSO5\n\u2212 \u2192 Fe3+\u00a0\u200b+\u00a0\u200bSO4\n\u2022\u2212\u00a0\u200b+\u00a0\u200bOH\u2212\n\n\n\n\n\n(18)\nSO4\n\u2022\u2212\u00a0\u200b+\u00a0\u200bOH\u2212 \u2192 SO4\n2\u2212\u00a0\u200b+\u00a0\u200b\u2022OH\n\n\n\n\n(19)\nSO4\n\u2022\u2212+ H2O \u2192 SO4\n2\u2212\u00a0\u200b+\u00a0\u200b\u2022OH\u00a0\u200b+\u00a0\u200bH+\n\n\n\n\n\n(20)\n2S2\u2212\u00a0\u200b+\u00a0\u200b2Cu2+(Fe3+) \u2192 2Cu+(Fe2+)\u00a0\u200b+\u00a0\u200bS2\n2\u2212\n\n\n\n\n\n(21)\n2S2\n2\u2212\u00a0\u200b+\u00a0\u200b2Cu2+(Fe3+) \u2192 2Cu+(Fe2+)\u00a0\u200b+\u00a0\u200bSn\n2\u2212\n\n\n\n\n\n(22)\nS\nn\n\n2\u2212\u00a0\u200b+\u00a0\u200b2Cu2+(Fe3+) \u2192 2Cu+(Fe2+)\u00a0\u200b+\u00a0\u200bS0\n\n\n\n\n\n(23)\nS0\u00a0\u200b+\u00a0\u200b2Cu2+(Fe3+) \u2192 2Cu+(Fe2+)\u00a0\u200b+\u00a0\u200bSO4\n2\u2212\n\n\n\n\n\n(24)\nCu+\u00a0\u200b+\u00a0\u200bFe3+ \u2192 Fe2+\u00a0\u200b+\u00a0\u200bCu2+\n\n\n\nBarhoumi et\u00a0al. [146] developed a novel electrochemical advanced oxidation process with heterogeneous catalyst chalcopyrite for the degradation and mineralization of tetracycline (TC). They revealed that the performance of the electro-Fenton (EF)/chalcopyrite process was superior to conventional EF, acquiring nearly complete mineralization of the TC solution under optimized conditions after 360\u00a0\u200bmin, which was attributed to the self-regulation of Fe2+ and Cu2+ content in the reaction medium, exerting a synergistic effect. Moreover, they noticed that oxalic and oxalic acids were more rapidly destroyed when using chalcopyrite, and the results could be accounted for the existence of Cu2+ ions, whose carboxylate complexes were more reactive towards \u2022OH, thus led to an enhancement in their oxidation rate.In recent years, as a terylene chalcogenide, ZnIn2S4 with a typical layered structure has attracted significant interest due to its excellent electrical and optical properties, intense visible light capturing ability and excellent chemical stability [147,148]. More importantly, ZnIn2S4 has a matching bandgap structure with semiconductor materials such as g-C3N4 and TiO2, which makes it possible to construct binary heterojunctions based on ZnIn2S4 to improve the charge separation and migration efficiency of photocatalytic systems [149\u2013151]. Shi et\u00a0al. [152] synthesized a series of carbon quantum dots (CQDs) in different proportions as ideal cocatalysts to enhance the photocatalytic activity of ZnIn2S4 microspheres in the visible range. Meanwhile, the optimal CQDs(5)/ZnIn2S4 hybrid photocatalyst showed excellent methyl orange (MO) degradation ability (10\u00a0\u200bmg/L and 100% degradation within 40\u00a0\u200bmin). The degradation rate was 2.34 times higher than that of ZnIn2S4 alone.Cu2FeSnS4 (CFTS) is one of the important ores in tin ore [153]. In addition, it has been proven to be an earth-abundant quaternary semiconductor and is an alternative material for solar energy conversion with excellent photovoltaic properties [154,155]. As a result, it is receiving increasing attention due to its potential applications in water decomposition, solar cells and pollutant degradation [156]. As demonstrated in Fig.\u00a04(d), Kong et\u00a0al. [157] prepared flower-like Cu2FeSnS4 nanomaterials for the activation of persulfate (PS) to degrade bisphenol A (BPA) in model industrial wastewater. The results showed that Cu2FeSnS4 catalyzed the decomposition of PS more effectively than mono-metal Cu/Fe/Sn sulfide and showed catalytic activity over a wide pH range. EPR spectroscopy, X-ray photoelectron spectroscopy and radical quenching experiments reveal the synergic effect of Cu, Fe and tin nutraceutical systems during PS activation. The intrinsic electron transfer between Cu, Fe and Sn, especially the Fe(II)\u2217 species formed on the surface of CFTS after Fe(II) is complexed by S, overcomes the inhibition of the M(n+1)+/Mn+ redox cycle.In view of the unsatisfactory performance of single-component catalysts, the combination of two or more catalytic materials to construct heterogeneous structures has become an effective strategy to improve catalyst activity. It can not only generate electron redistribution and achieve a synergistic effect at the interface by combining different components but also generate a new interface structure by changing the composition and crystal phase of the structure, to achieve efficient catalytic function. Metal sulfide-based heterojunction catalysts show outstanding performance in chemical reactions due to their beneficial interfacial properties. However, the rational design of heterogeneous catalysts with the required interface properties and charge transfer properties remains challenging. At present, many studies have reported the application of heterojunction catalysts constructed of metal sulfide and other excellent materials in the field of advanced oxidation.Guo et\u00a0al. [158] reasonably designed the hollow flower-shaped polyhedron \u03b1-Fe2O3/defective MoS2/Ag Z-scheme heterojunction using the polyhedron \u03b1-Fe2O3 as template by a one-pot hydrothermal deposition process. Hollow floral polyhedral heterojunction uses multiple light reflections in the hollow structure to achieve the enhanced photocatalytic activity. As a link between \u03b1-Fe2O3 and Ag nanoparticles, defective MoS2 provides a large number of active sites and broadens the light response region. Importantly, the photocatalytic Fenton degradation of 2,4-dichlorophenol and salicylic acid by the hollow flower-shaped polyhedron \u03b1-Fe2O3/Defective MoS2/Ag is higher than that of \u03b1-Fe2O3 and \u03b1-Fe2O3/defective MoS2.As demonstrated in Fig.\u00a04(e), Zhao et\u00a0al. [159] successfully prepared for the first time a novel p-n heterojunction photocatalyst with core-shell structure n-BiVO4@p-MoS2 by in situ hydrothermal method. The shell thickness of MoS2 can be easily adjusted by changing the concentration of MoS2 precursor in solution. For the photocatalytic reduction of Cr6+ and the photocatalytic oxidation of crystal violet, BiVO4@MoS2 sample showed excellent photocatalytic performance. Its enhanced photocatalytic activity was mainly due to the high specific surface area and strong adsorption capacity of the catalyst for pollutant molecules. The core-shell geometry also increases the contact area between BiVO4 and MoS2 and promotes the charge transfer at the BiVO4/MoS2 interface.Density functional theory (DFT) studies the electronic structure of multi-electron systems and is a quantum mechanical method [160,161]. The main goal of DFT is to use electron density as a fundamental quantity to replace the wave function. Density refers to the number density of electrons. A functional says that energy is a function of electron density, which in turn is a function of spatial coordinates [162]. A function of a function is a Functional. DFT is a method to study the electronic structure of multi-electron systems by electron density [163]. Specifically, in operation, DFT simplifies difficult problems involving electron-electron interaction into non-interaction problems through various approximations, and then puts all errors into a single term (XC potential), and then analyzes this error. The most basic function of DFT is to calculate the electron density distribution and ground state energy of a system [164]. The core of the concrete calculation method is Hohenberg-Kohn theorem and Kohn-Sham equation. DFT is of great significance, which makes it possible to calculate the electronic structure of larger systems. The first-principles calculation does not rely on other empirical parameters, and only requires five basic constants (particle mass, electric quantity, Planck constant, light speed and Boltzmann constant). Starting from the chemical composition and crystal structure of the material, various ground state properties of the material can be obtained by solving the Schrodinger equation, such as band structure, optical properties, mechanical properties, thermodynamic properties and magnetic properties. In recent years, after continuous development, density functional theory occupies a mainstream position in material simulation, with advantages of small error and high efficiency, and prominent advantages in the calculation of transition metal atomic system and phonon system [165\u2013168].DFT calculations are often combined with related experiments to supplement and extend the experimental discipline. The mechanism behind chemical reaction processes can be further explored by studying the structure of materials and small molecules (e.g., bond length, vibration). In recent years, much progress has been made in information about the surface binding positions of metal sulfide-based catalysts and oxidants by reliable DFT calculations. It mainly includes computational structural properties (3D, 2D, 1D, heterogeneous structure model, adsorption energy, binding energy), calculation of electronic properties (charge density, state density, band structure, orbital hybridization, charge projection, differential charge density, spin density), calculation of defect properties (formation energy, transition energy level, migration path and barrier) and calculation of adsorption and catalytic properties (adsorption configuration, free energy, molecular decomposition barrier).As shown in Fig.\u00a05\n(a), Fang et\u00a0al. [169] synthesized a new P-doped CdS nanorods@NiFe layered double hydroxides (LDH) Z-scheme photocatalyst (P\u2013CdS@NiFeLDH) by hydrothermal and calcination. By DFT calculation, it is found that phosphorus doping can form a new Fermi level, reduce the intermediate band gap, and thus extend the lifetime of photogenerated electron carriers. As a result, the degradation rate of bisphenol A by P\u2013CdS@NiFe-LDH photocatalyst reached 98% within 160\u00a0\u200bmin, and the photocatalytic degradation performance was significantly improved. Huang et\u00a0al. [170] demonstrated that monodisperse iron atoms are confined to MoS2 nanosheets with dual catalytic sites as highly active and stable catalysts to catalyze the efficient oxidation of aniline through the activation of PS. The DFT calculation further explained the high catalytic performance of Fe0\n\u00b7\n36Mo0\n\u00b7\n64S2 by the low oxidation state of Fe and the strong metal-carrier interaction in Fe\nx\nMo1-x\nS2. The Bader charge calculation results of Fe and Mo atoms in Fe0\n\u00b7\n36Mo0\n\u00b7\n64S2 showed that the perturbation of the electronic structure of MoS2 by Fe doping would trigger the catalytic activity of MoS2, and the charge transfer between the restricted Fe atom and the MoS2 carrier indicates that the metal-carrier interaction is vital. This interaction results in a rapid charge transfer between Fe and Mo during the catalytic process. As displayed in Fig.\u00a05(b), Guo et\u00a0al. [171] synthesized Co, S (pg-C3N4/Co3O4/CoS) using in situ template method and microwave vulcanization. DFT calculations show that electron migration from carbon nitride to CoS promotes the Co3+/Co2+ cycle. The calculation results of the projected density of state showed that the introduction of Co3O4 makes the Fermi level move towards g-C3N4 conduction band, and the formation of CoS on Co3O4 further makes the Fermi level enter the bottom of g-C3N4 conduction band. This means that g-C3N4 has better electrical conductivity, which is favorable for photogenic charge transfer. As shown in Fig.\u00a05(c), Ye et\u00a0al. [172] synthesized MIL-88\u00a0\u200bB(Fe)/Fe3S4 hybrid material with many unsaturated iron sites to treat the antibiotic trimethoprim by electro-Fenton processes. DFT calculations established the structure modeling, charge density and adsorption of H2O2. The results showed that the electron transfer from O to Fe promotes the transition from Fe(III) to Fe(II). Due to the presence of S atoms, more dense charge accumulation is concentrated near the Fe site, and the enhanced charge redistribution produces more electron-rich Fe active sites and accelerates the Fe(III)/Fe(II) redox cycle by increasing the rate of electron transfer. The adsorption energy of H2O2 on the catalyst showed that the adsorption of H2O2 on the catalyst was stable and conducive to the reaction.We have summarized the recently developed progress in synthesizing, characterization, and application of metal sulfides. Various methods have been extensively applied to prepare metal sulfides, including hydrothermal and solvothermal, template, thermal decomposition, and precipitation. The powerful characterization techniques (e.g., XPS, XRD, BET, FTIR, SEM, TGA, AC-STEM, XAFS and in situ characterization) are beneficial for us to understand the physicochemical properties of the catalyst better. The as-synthesized and natural metal sulfides can be introduced into various heterogeneous catalytic systems for wastewater treatment, and the performance and the mechanism of catalytic oxidation were discussed in detail. Although the mechanisms of catalytic oxidation were inconsistent to some degree, the following two conclusions were generally proposed by metal sulfide-based AOPs. One is the enhancement of the electron transfer efficiency caused by reductive S2\u2212 since S2\u2212 possesses lower electronegativity than O2\u2212; the other is that unsaturated S atoms on the surface of the metal sulfide can capture protons to form H2S and expose the metal active sites with reducing properties, thereby accelerating the rate-limiting reaction step. In addition, we discussed the theoretical calculation of metal sulfides, and elaborated the reaction mechanism of metal sulfides in different AOPs from the perspectives of modeling, adsorption energy calculation and free energy calculation.Despite the excellent potential of metal sulfide-based AOPs for wastewater treatment, there are still some issues that need to be taken into consideration.\n\n(1)\n\nPreparation of metal sulfides. The preparation methods of functionalized polysulfide still face specific difficulties, particularly compared to the one-component metal sulfides. Although many synthetic methods were employed for the fabrication of metal sulfides, the yield and efficiency of these methods are not satisfied yet, thus limiting large-scale practical applications.\n\n\n(2)\n\nRoles of sulfur species. It is verified that metal sulfides have higher catalytic activity than the corresponding oxides. However, previous work only focused on the surface S of the catalyst, while it is far from completion because there is almost no concern about sulfur in aqueous water. Therefore, it is essential to further inspect the water chemistry of sulfur speciation, particularly the transformation and destination of sulfur. It is beneficial better to understand the mechanisms of metal sulfide-based AOPs.\n\n\n(3)\n\nLeaching of toxic metal ions. Many studies demonstrated metal sulfide's superiority in improving catalytic oxidation activity. However, the leaching of toxic metal ions cannot be overlooked. To counter this drawback, we should design novel supported catalysts. The material with high stability and superior electrical conductivity is preferably selected as the supporter, such as metal oxide and carbon-based materials, and then the leached toxic metal ions can be immobilized by the supporters.\n\n\n(4)\n\nStability and reusability. The stability and reusability of as-prepared catalysts in a continuous operation need to be further investigated to be applied for practical wastewater treatment. It is essential to effectively prevent the loss of active components and avoid catalyst poisoning.\n\n\n(5)\n\nLaboratory research stage to engineering application stage. At present, most of the work in this research field adopts beaker experiments, which is quite different from the actual engineering application. Therefore, a great deal of effort is still needed to scale up these reaction processes to achieve practical applications of metal sulfide catalysts. It should be further explored in improving catalytic performance and reducing the use cost. Low-cost, abundant resources, stable and durable metal sulfide catalysts should be carefully selected.\n\n\n(6)\n\nIn-depth exploration of the reaction mechanism. At present, there are many catalytic systems based on metal sulfides. However, for different reaction systems, the depth and breadth of theoretical calculation are still very limited, especially the lack of in-depth analysis of the change in the electronic structural properties of catalysts. Therefore, further work should thoroughly elucidate the root causes of the superior catalytic properties of metal sulfides through in-depth DFT calculations.\n\n\n\nPreparation of metal sulfides. The preparation methods of functionalized polysulfide still face specific difficulties, particularly compared to the one-component metal sulfides. Although many synthetic methods were employed for the fabrication of metal sulfides, the yield and efficiency of these methods are not satisfied yet, thus limiting large-scale practical applications.\nRoles of sulfur species. It is verified that metal sulfides have higher catalytic activity than the corresponding oxides. However, previous work only focused on the surface S of the catalyst, while it is far from completion because there is almost no concern about sulfur in aqueous water. Therefore, it is essential to further inspect the water chemistry of sulfur speciation, particularly the transformation and destination of sulfur. It is beneficial better to understand the mechanisms of metal sulfide-based AOPs.\nLeaching of toxic metal ions. Many studies demonstrated metal sulfide's superiority in improving catalytic oxidation activity. However, the leaching of toxic metal ions cannot be overlooked. To counter this drawback, we should design novel supported catalysts. The material with high stability and superior electrical conductivity is preferably selected as the supporter, such as metal oxide and carbon-based materials, and then the leached toxic metal ions can be immobilized by the supporters.\nStability and reusability. The stability and reusability of as-prepared catalysts in a continuous operation need to be further investigated to be applied for practical wastewater treatment. It is essential to effectively prevent the loss of active components and avoid catalyst poisoning.\nLaboratory research stage to engineering application stage. At present, most of the work in this research field adopts beaker experiments, which is quite different from the actual engineering application. Therefore, a great deal of effort is still needed to scale up these reaction processes to achieve practical applications of metal sulfide catalysts. It should be further explored in improving catalytic performance and reducing the use cost. Low-cost, abundant resources, stable and durable metal sulfide catalysts should be carefully selected.\nIn-depth exploration of the reaction mechanism. At present, there are many catalytic systems based on metal sulfides. However, for different reaction systems, the depth and breadth of theoretical calculation are still very limited, especially the lack of in-depth analysis of the change in the electronic structural properties of catalysts. Therefore, further work should thoroughly elucidate the root causes of the superior catalytic properties of metal sulfides through in-depth DFT calculations.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to acknowledge the financial support from the National Natural Science Foundation of China (No. 52200105, 52070133) and Sichuan Program of Science and Technology (2023NSFSC0344, 2023JDZH0010).", "descript": "\n                  Due to its unique physical and chemical properties, metal sulfide has been proven to be a promising ideal candidate for metal oxide catalysts and has been widely used in many catalytic fields. In recent years, advanced oxidation processes (AOPs), especially those based on metal sulfides, have been recognized as one of the most effective techniques for controlling water pollution due to their superior catalytic performance and stability. However, there is a lack of systematic summary and elaboration of the reported works on metal sulfide catalysts. This work reviews the synthesis, characterization and application of metal sulfide in AOPs for water decontamination. In addition, we further summarized the catalytic oxidation mechanisms of different metal sulfide-based AOPs and combined them with density functional theory (DFT) calculation to clarify the active root of the catalytic reactions of various metal sulfides. Finally, the application of metal sulfide is prospected, including the challenges in large-scale preparation, sulfur hydrochemistry and metal ion leaching, and the stability and reusability of metal sulfide. This review will help guide the future development of metal sulfide and further develop efficient and stable metal sulfide-based AOPs to better deal with water pollution.\n               "}
{"full_text": "No data was used for the research described in the article.A novel, sustainable, clean, and efficient energy is required to meet the ever-increasing energy demand [3,4]. The idea of generating H2 from renewable or non-renewable sources without harming the environment is considered the most promising solution for decarbonizing large sectors of the global economy [5]. Besides, H2 qualifies to work as an energy carrier/storage, which can upgrade the renewable energy sector [6]. Unfortunately, H2 is unavailability in nature and is usually obtained from fossil fuels because it is more cost-effective [7]. Shifting toward producing clean H2 at affordable prices requires novel, eco-friend, inexpensive production methods [8]. Indeed, the generation of \u201cgreen\u201d H2 from a solar-driven water splitting process (WSP) represents the trendiest clean energy source, which can diminish our addiction to fossil fuels and mitigate the problem of anthropogenic climate disasters [9,10]. Immense efforts have been devoted to develop the \u201cideal\u201d photocatalyst material, which is still the main obstacle in this field. The majority of high-performance photocatalysts still need one of the following points or all of them, including (1) a noble co-catalyst, (2) a complex fabrication procedure, or (3) a \u201ccrutch\u201d to secure hot electrons like PV-PEC model [11,12].In 2009, graphitic carbon nitride (g-C3N4), a fangled metal-free organic conjugated polymeric semiconductive (n-type) was presented, for the first time, as an active photocatalyst material [13]. Referring to Table 1\n, the distinctive properties of g-C3N4 make it possessed research attention since then. However, virgin g-C3N4 can not harness long wavelength (400\u00a0\u2264\u00a0\u03bb\u00a0\u2264\u00a0700\u00a0nm) and suffer from high recombination kinetics between photoinduced charges (e\u2212/h+) [14]. Besides, the g-C3N4 bulk own low texture features, which also dwarf the photocatalytic performance [15]. More details are provided in Table 1 (cons). The aforementioned shortcomings can be alleviated by several means, such as structure re-engineering, doping method, framework defect engineering, substitution method, surface modification, nano heterojunction, etc. [12,16\u201319].Compared to non-nano counterparts, the fabrication of a nano-scale photocatalyst can improve the contact area with the reaction medium, enhance charge separation efficiency and modify the photoelectronic properties [37\u201340]. For instance, Xu and co-workers fabricated g-C3N4 hollow nanotubes (labeled as CNNTs) with the assistance of hydroxylamine sulfate. As expected, the CNNTs with distinguished texture-photoelectronic properties showed an impressive H2 evolution rate (HER) at 180\u00a0\u03bcmol/h with a ca. 19.5 folds improvement over bulk g-C3N4 [40].The essential role of a co-catalyst in photocatalyst material is to improve charge separation or/and increase active site abundance-accessibility. Therefore, presenting a cheap, abundant, eco-friendly co-catalyst with high performance is critical for potential applications. The NiO [41,42], NiS [43\u201345], MoS2 [46], Ag2O [47], Cu2O [48], Ni(OH)2 [49], Fe2O3 [50], and Co3O4 [50], belongs to the transition metal oxides/sulfides category. The NiO can be classified as a cheap, stable, and eco-friendly compound with good performance as a co-catalyst [41,42]. Historically, NiO was used for the first time as a co-catalyst in 1982 to improve the performance of SrTiO3 in water vapor decomposition [51]. Since 2012, nickel-based substances have been used as co-catalyst for g-C3N4, including, but not limited to, Ni(OH)2 [49], Ni [52,53], Ni2P [54], and NiSx [43\u201345]. The fall of the nobility: this report aims to proffer nickel oxide (NiO) as a cheap and efficient co-catalyst to offer an alternative to scarce extravagant noble metals (ex. Pt, Pd).To our knowledge, the NiO loaded over deficient, porous S-doped g-C3N4 nanofiber synthesis via the electrospinning process has not been reported yet. In light of this, it will be an exciting and valuable task to synthesize NiO nanoparticles from scratch and immobilize it over S-doped g-C3N4 nanofiber (x%NiO/S-g-C3N4-F) via an electrospinning process. This paper evaluated the performance of the H2 evolution rate from the water under simulated solar illumination (full spectrum) within a photoelectrochemical system (PEC, H type). In addition, this report scrutinized the impact of the NiO load amount (ca. 1\u20135\u00a0wt%) on the light absorption, charge separation, and photocatalytic activity of S-g-C3N4-F. Subsequently, the as-synthesized samples were characterized by various spectroscopic techniques and electron microscopy. Eventually, the results signalized that the inorganic-organic nanohybrid 2%NiO/S-g-C3N4-F photocatalyst could be a potential candidate for solar-to-fuel conversion applications.All the starting chemicals and solvents used to prepare or test the fabricated material are presented in Table 2\n. All chemicals were used as received without any further purification.A former study showed that NiO particles could be prepared via a facile calcination method (see Fig. 1\n) [1]. In detail, Nickel (II) nitrate hexahydrate (5.0\u00a0g) in the solid phase was put in a crucible with a lid (200\u00a0mL) and calcined at 650\u00a0\u00b0C (5\u00a0\u00b0C/min) under an ambient atmosphere for 3\u00a0h and left to cool naturally overnight. As-obtained bulk NiO was added to 80\u00a0mL of distilled water and grinding using two homogenizer models. The first round with ULTRA-TURRAX (IKA\u00ae T25) for 10\u00a0min at 8500 RPM. Next, the 2nd round with ULTRA-TURRAX (Model: T10B) with a smaller blade for another 5\u00a0min at 6500 RPM. The obtained slurry was heated over a hot plate (85\u201390\u00a0\u00b0C) to remove the added water. Eventually, the collected samples were kept in sealed glass bottles for further use or characterization.The synthesis procedure of S doped g-C3N4 bulk (termed S-g-C3N4 bulk) by thermal treatment of TU, whereby 7.6\u00a0g (ca. 0.10\u00a0mol) was added to a crucible with led (300\u00a0mL) and calcinated at 575\u00a0\u00b0C/2\u00a0h, under an ambient atmosphere, with 5\u00a0\u00b0C/min as a ramping rate. After that, the prepared sample is left overnight in the oven to cool naturally. A similar procedure for product collection, grinding, and drying was followed as in the preparation of NiO.The synthesis procedure of NiO/S-doped g-C3N4 nanofibers (1D) comprised solution preparation, whereby a specific amount of NiO (10\u201350\u00a0mg) dispersed into 78\u00a0mL of solvent (40\u00a0mL DMF\u00a0+\u00a06.0\u00a0mL HAc\u00a0+\u00a032\u00a0mL ethanol) by using a homogenizer (6000 RPM) for 5\u00a0min. After that, 5.0\u00a0g of Thiourea (TU, NH2CSNH2) was dissolved in the solution with stirring (300 RPM) at room temperature for an hour. Next, 7.5\u00a0g of polyvinylpyrrolidone (PVP, Mw 1,300,000) was added and mixed (600 RPM) at room temperature for another hour until it became homogeneous. Then, the prepared solution was gently stirred (150 RPM) overnight at room temperature to remove any remaining bubbles.The electrospinning machine operated at a closed system (see Fig. S1) with a temperature of around 28\u201332\u00a0\u00b0C under controlled humidity (HUM\u00a0\u223c\u00a030%). An irradiation lamp (75\u00a0W) and dehumidifier device worked continuously to keep the humidity at the set point in the closed chamber (see Fig. S2). To keep the consistency in the experiment, the dehumidifier device was not turned off overnight to maintain the humidity at 30% for the following day. The spinning precursor was loaded into a plastic syringe (25\u00a0mL), and the stainless-steel needle size was 19 gauge in all experiments, as shown in Fig. S1. Pictures of the electrospinning setup with other details are provided in the supplementary material.The flow rate of the syringe pump was 0.8\u00a0mL/h for all electrospinning processes. Besides, the distance between the positive electrode and the ground collector was 20\u00a0cm, and the voltage difference was kept at 20\u00a0kV (1\u00a0kV/cm). One radiation heating unit (75\u00a0W) was operated continuously and located 50\u00a0cm higher than the middle point between the needle head and collector. Afterward, collected \u201cgreen\u201d fiber was thermally treated at 575\u00a0\u00b0C/2\u00a0h to polymerize g-C3N4 and completely remove the PVP and extra organic components [2,55]. A former study proved that PVP nanofiber fabricated via the electrospinning method is almost decomposed at 500\u00a0\u00b0C, as the TGA analysis showed [55]. The ramping rate was 5\u00a0\u00b0C/min, and the prepared sample was left to cool naturally in the oven overnight. For comparison, S-doped g-C3N4 nanofiber (labeled as S-g-C3N4-F) was prepared following the same procedure of NiO/S-g-C3N4-F without adding the NiO.A test run showed that the yield of pure S-g-C3N4-F synthesized at 575\u00a0\u00b0C/2\u00a0h without NiO is about 22%. Based on that, the added amount of NiO co-catalyst (10\u201350\u00a0mg) represents about 1\u20135\u00a0wt% of the total weight. The corresponding samples were denoted as x%NiO/S-g-C3N4-F, where x\u00a0=\u00a01, 2, 3, 4, and 5\u00a0wt%. Collected samples were kept in sealed glass bottles for testing and characterization purpose. More details about the journey from chemicals to target compounds through intermediate phases under thermal treatment are graphically presented in Fig. 1.The working electrode for testing the performance was prepared as follows procedure: (i) 100\u00a0mg of prepared photocatalyst material dispersed within Nafion solution (5\u00a0mL, 0.5\u00a0wt%), after that, (ii) the obtained slurry was loaded over indium tin oxide (ITO) glass (30\u00a0\u00d7\u00a030\u00a0mm, 1850\u00a0\u00c5 thick, 10\u201315 \u03a9/in2) by slowly dripping, (iii) the as-prepared electrode was left at room temperature for 12\u00a0h, after that (iv) the as-prepared electrode was dried over a hot plate at ca. 100\u00a0\u00b0C for another 12\u00a0h.Different characterization techniques were used to identify the prepared material, quantify the structure features, investigate the photogenerated charges separation efficiency, and other purposes, as shown below in Table 3\n.The photocatalytic hydrogen production system was performed in a photoelectrochemical cell (PEC, H-type) with a circle window from quartz (dia. 24\u00a0mm) to transfer the incident light efficiently, as shown in Fig. 2\n. As prepared photocatalyst material loaded over ITO-coated glass was immersed in a methanol aqueous solution (20\u00a0vol%, 200\u00a0mL). After that, the photocatalytic system was exposed to lamp irradiation (Metal halide 400\u00a0W, full spectrum, 380\u00a0\u2264\u00a0\u03bb\u00a0\u2264\u00a0780\u00a0nm). The light intensity of 400\u00a0W Metal halide was quantified at about 90\u00a0mW/cm2 [57].The distance between the irradiation device and the photocatalytic fuel cell (PEC) was kept constant in all the experiments at 20\u00a0cm. Nafion 117 (dia. 20\u00a0mm) was used as a replaceable ion transport membrane. Besides, a solid square of a platinum electrode (20\u00a0\u00d7\u00a020\u00a0mm) was used for the hydrogen reduction reaction. A simplified sketch for the setup used to test photocatalysis material and an image of the actual setup are presented in Fig. 2 (a, b). The water displacement method was used to quantify the production rate of H2 [58\u201360]. The photocurrent response for prepared samples was also quantified without any bias potential via a digital multimeter with PC data logging. The photocatalytic fuel cell (PFC) was isolated in a closed box to eliminate any external light that may manipulate the results. The redox reactions for the water-splitting process are also presented in Fig. 2 (b).Powder X-ray diffraction patterns (XRD) were used to investigate the phase and composition of the as-prepared samples. Fig. 3\n (a) shows the diffraction peaks of pure NiO at 2\u03b8 of 37.26\u00b0, 43.26\u00b0, 62.81\u00b0, and 79.36\u00b0, which were respectively indexed as (111), (200), (220), and (222) plane of cubic NiO structure according to JCPDS file no. 47\u20131049. Besides, Fig. 3 (b, c) show the XRD patterns of S-g-C3N4 bulk and NiO/S-g-C3N4-F following the JCPDS file no. 87\u20131526, where the two peaks (2\u03b8) at 13.1 and 27.6\u00b0 ascribed from (100) and (002) planes [20,52]. Besides, the slight shifting of the main peak for g-C3N4 can be related to injecting the S in the polymeric backbone (see Fig. 3 (b, c)) [61]. As shown in Fig. 3 (c)), one small peak of NiO was observed at 43.06\u00b0 (plane 200) in the XRD of NiO/S-g-C3N4-F. The XRD of NiO/S-g-C3N4-F emphasized the presence of NiO, and a similar XRD pattern was captured in related reports (ex. Ni/g-C3N4 [53], NiO/g-C3N4 [62]). The average crystalline grain size (DXRD) for as-prepared samples was calculated from the full width at half-maximum (FWHM) of the XRD peaks using the Debye-Scherrer equation [63]. Accordingly, the average DXRD of NiO, S-g-C3N4 bulk, and S-g-C3N4-F were quantified as 25.1 and 3.64, and 0.58\u00a0nm, respectively. The DXRD of the NiO/S-g-C3N4-F series quantified around 0.60\u00a0nm (see Fig. 3 (c)), which indicates that the NiO did not impact the crystallinity, contrary to the structure. The DXRD error of the NiO/S-g-C3N4-F series was estimated at around 6.61% (ca.\u00a0\u00b1\u00a00.03966\u00a0nm) [64]. Fabricated S-g-C3N4-F destroys the order in the long-range without effect on the short-range order, which creates an amorphous structure [65]. Referring to XRD of S-g-C3N4-F and NiO/S-g-C3N4-F series, the peak of 13.1\u00b0 disappeared within XRD noise, and the main peak (002) became broader, which can be attributed to low crystallinity and effect of introducing the S as heteroatom into the framework [65\u201367]. To sum up, the XRD of NiO, S-g-C3N4-F, and NiO/S-g-C3N4-F confirmed the formation of the target materials without any impurities.\nFig. 4\n shows the morphologies of grinded NiO (a), 2%NiO/S-g-C3N4 \u2033green\u201d fiber (c), and calcinated 2%NiO/S-g-C3N4-F (g, h). The diameter of 2%NiO/S-g-C3N4 \u2033green\u201d fiber was analyzed via ImageJ software and quantified in a range of 131\u20131151\u00a0nm with an average of 540\u00a0nm over the whole tested area (Fig. 4 (d)). Besides, the EDX of 2%NiO/S-g-C3N4 \u2033green\u201d fiber confirmed the presence of Ni with good distribution over the host material surface (see Fig. 4 (e, f-yellow)). Based on the EDX, there is a high properly that the small nanoparticles that decorated the \u201cgreen\"/calcinated S-g-C3N4-F surface are NiO nanoparticles.As shown in Fig. 4 (g, h), the building unit for 2%NiO/S-g-C3N4 nanofiber consists of short nanofiber (ca. 40\u201360\u00a0nm) that coalesce and fusion to construct the like-fiber structure with high porosity. The average diameter for calcinated 2%NiO/S-g-C3N4-F was quantified following the same technique as 463\u00a0nm (see Fig. 4 (i)). Interestingly, the calcinated 2%NiO/S-g-C3N4-F was carried on the flower leaf tip without even bending (Fig. 4 (j), indicating that the structure is filled with porous. Relied on the BET result (section 4.3), the density for 2%NiO/S-g-C3N4-F was quantified as ca. 9.17\u00a0mg/cm3. After thermal treatment, the distribution of NiO over electrospun PVP transferred from 7.5\u00a0g to about 1.0\u00a0g of S-g-C3N4-F, which increased Ni intensity as shown in EDX element mapping (Fig. 4 (k)). Referring to Fig. 4 (k), the EDX spectra of NiO/S-g-C3N4-F confirmed the presence of S as one of the constituent elements, which can be considered evidence of successfully injecting the S into the g-C3N4 matrix. Moreover, the excellent distribution of C and N elements emphasized that the S-g-C3N4 represents the main building units. In addition, the apparent concentration for EDX of 2%NiO/S-g-C3N4-F showed an approximate ratio of 1:1 for Ni to O, which is also considered solid evidence of the formation of the NiO component. To sum up, the synergetic effect of fabrication S-g-C3N4-F with porous nanofiber structure and the magnetic properties of NiO can improve the incident light absorption efficiency, enhance the photoinduced charges separation and material recoverability.Improvement in the g-C3N4 texture features can enhance the charge separation efficiency and harvesting capability of the incident light [12]. The texture properties for S-g-C3N4 bulk and 2%NiO/S-g-C3N4-F were analyzed by nitrogen adsorption\u2013desorption experiment, and the result are presented in Fig. 5\n (a). The typical structure of S-g-C3N4 bulk as a stacking layer offers humble texture features as 3.78\u00a0m2/g and 0.026\u00a0cm3/g for BET surface area (SSA) and pore volume (Vp), which deteriorate photoreactivity [15]. In contrast, the NiO/S-g-C3N4-F structure consists of short nanofibers (ca. 40\u201360\u00a0nm) that coalesce and fusion to construct the porous nanofiber structure. Consequently, the well-developed porosity for 2%NiO/S-g-C3N4-F showed a substantial improvement compared to S-g-C3N4 bulk with 65.5\u00a0m2/g and 0.143\u00a0cm3/g as SBET and Vp, respectively. Besides, a graphical comparison between S-g-C3N4 bulk and 2%NiO/S-g-C3N4-F in SSA and Vp is shown in Fig. 5 (b).The BET hysteresis loop for 2%NiO/S-g-C3N4-F (Fig. 5 (c)) is located at 0.85\u00a0<\u00a0P/P0\u00a0<\u00a01.00 (green area) and takes the shape of type IV (BDDT classification), which emphasizes that the structure is filled with mesoporous (2\u221250\u00a0nm) [68,69]. In addition, Fig. 5 (d) shows the pore size distribution for 2%NiO/S-g-C3N4-F concentrated in the range of 1.5\u20135\u00a0nm. A similar H3 hysteresis loop was captured in a former study for TiO2/g-C3N4 nanofiber, which was related to the coalescence of nanoparticles to construct the fiber structure that created the slit-shaped pores [70]. Indeed, the enhancement in the texture features for NiO/S-g-C3N4-F enhances the active site availability and accessibility, as well as the light harness capability. Furthermore, the charges separation efficiency for NiO/S-g-C3N4-F expect to be improved as a synergetic effect of (i) the presence of carbon vacancies (Vc) [56,71,72], (ii) the magnetic properties of NiO [53,73], and (iii) the impact of nanofiber structure (diam. 463\u00a0nm) on decreasing the hot charges transmission distance [12].The radiative recombination process for photogenerated charges (e\u2212/h+) was studied by photoluminescence (PL) spectroscopy. The peak position for PL spectra of S-g-C3N4 bulk and NiO/S-g-C3N4-F centralized within the Vis-light zone (see Fig. 6\n (a)), which inform that the optical band gap (Eg) is less than 2.70\u00a0eV [74]. Compared to the S-g-C3N4 bulk, the PL peak for S-g-C3N4-F was not detected, meaning the luminous recombination probability was efficiently suppressed. In addition, the convexity intensity in the x%NiO/S-g-C3N4-F series decreased with increased NiO content from 1 to 2\u00a0wt%, as shown in Fig. 6 (b). However, the overloaded of NiO (3\u00a0\u2264\u00a0x\u00a0\u2264\u00a05\u00a0wt%) over the S-g-C3N4-F surface slightly facilitates the charges recombination rate, as shown in the overlap of PL spectra (Fig. 6 (b)).Grafted Sulfur as an electron-rich with smaller electronegativity (\u03c7\u00a0=\u00a02.58) in the g-C3N4 matrix narrowed the bandgap and downshifted the Fermi level due to occurring spin polarization in the material, which increased overlap in the orbitals of C, N, and S atoms and the \u03c0 states [66,75,76]. Moreover, using thiourea (TU) as a precursor source created defects within the polymeric backbones (see Fig. 1), which improved the excitons (e\u2212/h+) separation efficiency and expanded the light-responsive range [56,72]. The black color for g-C3N4-based material was captured in former reports (ex. Ni/g-C3N4 [53], metal oxide/g-C3N4 hybrids [50]). The black color for S-g-C3N4-F with or without NiO (Fig. 7\n (a)) can be related to the effect of changing the thermal treatment atmosphere during the decomposition of PVP (C6H9NO)n, which generated different gases (ex. N2, CO2, CO) [77]. Besides, both PL and XRD spectra support the supposition that S-g-C3N4-F based material owns an amorphous structure. A former study confirmed that the atomic arrangement impacted the band structure and charge recombination rate, which can justify the significant difference in the PL result between S-g-C3N4 bulk and S-g-C3N4-F [65]. A similar distinguished PL response was captured in a former study that fabricated amorphous g-C3N4 via thermal treatment (620\u00a0\u00b0C/2\u00a0h) under an Ar atmosphere. Based on the theory of band tails, formation of a proper disorder in the crystallinity of g-C3N4 narrowed the bandwidth (Eg) to 1.90\u00a0eV [65]. Based on that, the outstanding PL result for S-g-C3N4-F and NiO/S-g-C3N4-F series can be ascribed to the synergistic effect of the sulfur dopants and the defective-amorphous- nanofiber structure.\n\n(1)\n\n\nE\ng\n=\n1240\n/\n\n\u03bb\ng\n\n\n\n(\ne\nV\n)\n\n\n\n\n\nThe optical properties of S-g-C3N4 bulk, S-g-C3N4-F, and NiO/S-g-C3N4-F series were studied using UV\u2013Vis diffuse reflectance spectroscopy (DRS). The absorption edge (\u03bbg) is quantified from the intercept between the tangent of the absorption curve and the abscissa coordinate, as shown in Fig. 7 (c) [78]. The bandgap energy (Eg) for as-prepared samples was quantified according to equation (1) [78].The bandgap (Eg) for fabricated g-C3N4-based materials was quantified in the following order: S-g-C3N4-F (1.77\u00a0eV)\u00a0<\u00a01%NiO/S-g-C3N4-F (1.73\u00a0eV)\u00a0<\u00a02%NiO/S-g-C3N4-F (1.76\u00a0eV)\u00a0<\u00a03%NiO/S-g-C3N4-F (1.84\u00a0eV)\u00a0<\u00a04%NiO/S-g-C3N4-F (1.95\u00a0eV)\u00a0<\u00a05%NiO/S-g-C3N4-F (2.12\u00a0eV)\u00a0<\u00a0S-g-C3N4 bulk (2.61\u00a0eV). Fig. 7 (b) graphically compares the capability of S-g-C3N4 bulk and 2%NiO/S-g-C3N4-F in harnessing solar energy. Indeed, the overload of NiO over S-g-C3N4-F negatively impacts the photoelectronic character. Former studies (ex. MoS2/CdS [79], Ni2P/g-C3N4 [54], Ni(OH)2/a-Fe2O [80]) stated that co-catalyst could lower the light absorption capability of the main photocatalyst material through the light-blocking effect. Former studies quantified the Eg for NiO as 3.5\u00a0eV, which consider another plausible explanation for increased Eg for NiO/g-C3N4 nanocomposite with increased NiO content [62,81]. Based on that, the co-catalyst amount and distribution must be optimized to maximize the photocatalysis performance.Photocatalytic H2 evolution rate for prepared samples was measured under simulated solar irradiation (400\u00a0W Metal halide, 380\u00a0\u2264\u00a0\u03bb\u00a0\u2264\u00a0780\u00a0nm) using methanol (20\u00a0vol%) as a hole scavenger reagent. In the current study, no H2 production was detected before introducing the photocatalyst into the reaction medium or in the absence of the irradiation source. Based on that, the generated H2 was formed from photocatalytic reactions. Fig. 8\n (a, b) shows the photocatalytic H2-production performance (HER) and photocurrent value of different g-C3N4-based materials, including the S-g-C3N4 bulk, S-g-C3N4-F, and x%NiO/S-g-C3N4-F series. The photocurrent value was quantified on top of every hour for 60\u00a0s over 4 cycles for each sample. The photocatalytic HER was quantified in the following order: S-g-C3N4 bulk\u00a0<\u00a0S-g-C3N4-F < 5%NiO/S-g-C3N4-F < 4%NiO/S-g-C3N4-F < 1%NiO/S-g-C3N4-F < 3%NiO/S-g-C3N4-F < 2%NiO/S-g-C3N4-F.Based on the collected data, the S-g-C3N4 bulk exhibits a modest hydrogen production rate (HER) at 22.2 \u03bcmol/h with about 0.08\u00a0\u03bcA as photocurrent value (see Fig. 8 (a, b)). The humble performance of S-g-C3N4 bulk can be related to the bulk structure features with a high recombination rate between photogenerated charges, and the inability to harness visible light. In comparison, the HER of pure S-g-C3N4-F reached 63.55 \u03bcmol/h with a 2.86-fold improvement over S-g-C3N4 bulk, suggesting the pivotal role of morphology in enhancing photocatalytic performance. As can be seen, introducing a proper amount of NiO (x\u00a0=\u00a02\u00a0wt%) as a co-catalyst over S-g-C3N4-F increased active site availability and further suppressed the charge recombination rate (Fig. 6 (b)). Consequently, the HER of 2%NiO/S-g-C3N4-F enhanced to 107.04 \u03bcmol/h, representing almost double the pure S-g-C3N4-F performance (see Fig. 8 (a)).The maximum apparent quantum yield (AQY) of the 2%NiO/S-g-C3N4-F and S-g-C3N4 bulk was calculated at about 1.51% and 0.31% at wavelength 420\u00a0nm. Unfortunately, the overload of NiO over the host material body harms the light-harvesting capability and facilitates the charge (e\u2212/h+) recombination rate, as PL and DRS confirmed (Fig. 6 (b), Fig. 7 (c)). As a result, increased NiO content to 3\u20135\u00a0wt% over S-g-C3N4-F relapsed the HER with maintaining a better performance over virgin g-C3N4 in all cases present in this study (Fig. 8 (a)). A similar photocatalytic pattern was captured in documented reports for Ni/S-doped g-C3N4 [52] and NiO/g-C3N4 [62]. It is well-documented that the strong interaction between noble metals and host photocatalyst material facilitates charge flowability, which reinforces overall performance [12,54]. As well, creating a firm interface between the photocatalyst and the co-catalyst could afford the unobstructed charge-transfer channel [46,82]. Due to g-C3N4 tri-s-triazine (heptazine) ring structures, the transition metal ions can be readily absorbed on its surface [83]. Based on that, the significant difference in performance between S-g-C3N4-F and NiO/S-g-C3N4-F indicates the presence of an intimate interface between them.In addition, multiple runs of photocatalysis were carried out under identical conditions to test the stability of 2%NiO/S-g-C3N4-F in evolving H2 from the water dissociation process. After four consecutive cycles, the HER for 2%NiO/S-g-C3N4-F declined by only 9.3% from the initial performance (Fig. 8 (c)), indicating high stability [68,72]. The performance pattern for x%NiO/S-g-C3N4-F series and the significant enhancement for S-g-C3N4-F over S-g-C3N4 bulk is consistent with the PL and DRS results. Fig. 8 (d) displays the relationship between the HER rate and photocurrent value for fabricated g-C3N4. As expected, the HER for fabricated g-C3N4 and quantified photocurrent value showed harmonicity.A former study by Chaudhary et al. succeeded in loading g-C3N4 with Ni through thermal treatment with nickel carbonate (NiCO3) at 550\u00a0\u00b0C/2\u00a0h under an N2 atmosphere [73]. The fabricated Ni/g-C3N4 generated H2 at a rate of 40 \u03bcmol/h under natural \u201coutdoor\u201d sunlight. Moreover, the photocurrent value for Ni/g-C3N4 reached almost double the pure g-C3N4 value [53]. In this study, the fabricated noble-free 2%NiO/S-g-C3N4-F showed a competitive performance to many related reports (see Table 4\n). For example, the HER of 2%NiO/S-g-C3N4-F reached about 78.6% of Pt3%/S-doped g-C3N4 performance (136.0 \u03bcmol/h) [68]. The noble-free g-C3N4-based material that satisfies the thermodynamic requirements to split the water and form \u201cgreen\u201d H2 under the visible light spectrum has more potential to be used in potential applications [12]. To sum up, the distinguished texture properties of nanofiber morphology with the merit of adding NiO as a co-catalyst and the impact of injecting the S within the polymeric backbones consider the reasons behind the impressive photocatalytic performance of 2%NiO/S-g-C3N4-F.Nickel (Ni) and its derivatives (e.g. NiS, NiO) distinguish as cheap, abundant, and relatively effective co-catalyst [87]. Indeed, the magnetic properties of Ni can improve the separation efficiency of photogenerated carriers (e\u2212/h+) by rapidly capturing and transferring hot electrons [53,73]. Besides, the NiO as a co-catalyst can act as an active site to lower the H2 evolution overpotential and facilitate electron transfer (see Fig. 9\n (a)), which suppresses the recombination (bulk/surface) rate between generated charges (e\u2212/h+), as PL spectra confirmed. Indeed, NiO can also represent a reactive site by acting as the hydride-acceptor and proton-acceptor centers, as shown in Fig. 9 (b)) [44].Under thermal treatment, the sulfur vaper self-gas foaming impacts the texture properties of the prepared g-C3N4. Besides, inserting a non-metallic element (ex., S, B, O) into a polymeric bone does not create mechanical barriers or structural limitations. And most importantly, the integration of S as an electron-rich non-metallic element into polymer backbones increased the orbital overlap and the \u03c0 states as well as hybridized the S, C, and N orbital and created a sub-energy level below the conduction band (CB), as shown in Fig. 9 (c) [66,75]. Moreover, a former study stated that inserting S with smaller electronegativity (2.58) inside of N (3.04) altered the surface electronic properties by downshifting the CB (0.45\u00a0eV) and widened the VB by 0.12\u00a0eV, which positively impacted the mobility of hot h+ and enhanced redox reaction efficiency [76]. Based on that, the liberator electrons (e\u2212) are expected to transfer from the valance band (VB) to the sub-energy level before moving to the conduction band (CB). This two-step mechanism for hot electrons can somewhat suppress the recombination rate. In addition, the carbon vacancies (Vc) in the S-g-C3N4 matrix can impact the photogenerated carriers transmission (Fig. 9 (d)) [71]. Former studies showed that formation vacancies within the g-C3N4 framework could impact textural-photoelectric properties [88\u201390] and extend charges (e\u2212/h+) lifetime [89]. However, it usually requires a toxic, harmful treatment to create defects in the g-C3N4 matrix, like alkali-assisted (ex. KOH, NaOH, and Ba(OH)2) [88]. This study developed the carbon vacancies (Vc) in modified g-C3N4 as a sequence using the thiourea, a simple and efficient method (see Fig. 1) [56,72].The water-splitting process occurs in two steps: (1) two water molecules react with generated holes through an oxidation reaction, forming an oxygen molecule and four hydrogen protons (H+) (see Fig. 9 (a)), after that (2) two hot electrons (e\u2212) react with two hydrogen protons (H+) over a platinum plate to form H2 that appears in tiny bubbles. The excess generated holes (h+) are consumed by reacting with methanol as a sacrificial reagent.This study fabricated and tested the inorganic-organic nanohybrid NiO/S-g-C3N4 nanofiber in evolution of hydrogen from water. The fabrication method for NiO/S-g-C3N4-F consists of the electrospinning method with simple thermal treatment. The significant enhancement in HER for NiO/S-g-C3N4-F can indicate the existence of intimate interfacial between constituent material. In the current study, there was no H2 production before introducing the photocatalyst into the reaction medium or in the absence of the irradiation source. Based on that, the generated H2 was formed from photocatalytic reactions. Amongst all fabricated samples in the study, the S-g-C3N4-F decorated with NiO (2\u00a0wt%) demonstrates the highest H2 production rate and photocurrent value at 107.04 \u03bcmol/h and 1.01\u00a0\u03bcA, which surpassed the S-g-C3N4 bulk by 4.82 and 12.62-folds, respectively. Due to the magnetic character, the fabricated NiO/S-g-C3N4-F can be collected from reaction media by a simple method. Moreover, the fabricated 2%NiO/S-g-C3N4-F shows high stability, as the H2 formation rate slightly declined (<10%) after 4 consecutive cycling tests. The improvement in the photocatalytic performance of NiO/S-g-C3N4-F was ascribed to the following reasons: (i) acted the NiO as an electron-trapping center, (ii) the merit of fabrication of amorphous nanofiber structure, and (iii) the red-shift of the absorption edge due to modify the electronic structure by injecting the S in the polymeric bones. This work demonstrates that NiO is an efficient co-catalyst, which can pave the way for upgrading the efficiency of photocatalyst material without harming economic viability. Eventually, the fabricated 2%NiO/S-g-C3N4-F as an efficient, stable, and low-cost photocatalyst can be a candidate for photocatalytic applications, especially Vis-light-driven H2 evolution and pollutants photodegradation.Conception and design of study: Suleiman M. Abu-Sari, Wan Mohd Ashri Wan Daud, Muhamad Fazly Abdul Patah, Bee Chin Ang. acquisition of data: Suleiman M. Abu-Sari, Wan Mohd Ashri Wan Daud, Muhamad Fazly Abdul Patah, Bee Chin Ang. analysis and/or interpretation of data: Suleiman M. Abu-Sari, Wan Mohd Ashri Wan Daud, Muhamad Fazly Abdul Patah, Bee Chin Ang. Drafting the manuscript: Suleiman M. Abu-Sari, Wan Mohd Ashri Wan Daud, Muhamad Fazly Abdul Patah, Bee Chin Ang. revising the manuscript critically for important intellectual content: Suleiman M. Abu-Sari, Wan Mohd Ashri Wan Daud, Muhamad Fazly Abdul Patah, Bee Chin Ang. Approval of the version of the manuscript to be published (the names of all authors must be listed): Suleiman M. Abu-Sari, Wan Mohd Ashri Wan Daud, Muhamad Fazly Abdul Patah, Bee Chin Ang.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to acknowledge the Ministry of Higher Education Malaysia for the financial support through the Research Grant Scheme DP KPT (Project number: FRGS/1/2019/STG01/UM/02/5) and the University of Malaya through Impact-Oriented Interdisciplinary Research (Grant IIRG003B\u20132020IISS).The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.rineng.2023.100952.", "descript": "\n                  The energy dilemma is exacerbating. Artificial photosynthesis is a plausible blueprint for solar-to-fuel conversion applications to satisfy future energy demands. Developing a cheap, efficient photocatalyst material can be the watershed moment in this field. Herein, low-cost sulfur self-doped g-C3N4 nanofiber decorated by nickel oxide (denoted as x%NiO/S-g-C3N4-F) was obtained via electrospinning and one-step thermal treatment (575\u00a0\u00b0C). Outstandingly, the modified g-C3N4-based material could interact and harvest long wavelengths up to 706\u00a0nm. Moreover, the quantified specific surface area (SSA) for 2%NiO/S-g-C3N4-F is more than 17.3 folds larger than S self-doped g-C3N4 bulk (denoted as S-g-C3N4 bulk). As a result, the optimal photocatalytic property of 2%NiO/S-g-C3N4-F is almost five times as high as S-g-C3N4 bulk, achieving 107.04 \u03bcmol/h. The suggested photocatalysis mechanism was proposed and supported by the results. Significantly, loading a proper amount of NiO over modified S-g-C3N4 promoted performance as well as convenient recovery and reusability. According to the experimental and characterization results, the fabricated 2%NiO/S-g-C3N4-F consider a potential candidate for photocatalytic applications, especially Vis-light-driven H2 evolution.\n               "}
{"full_text": "\n\n\nData will be made available on request.\n\n\nData will be made available on request.The quest for green energy production and the search for environmentally friendly energy production is one of the top global priorities today. Along with the application and utilization of renewable energy sources, there is an urgent need to develop multifunctional energy storage and energy conversion technologies to balance the demands of supply/demand management of the produced clean and sustainable energy [1\u20133]. Supercapacitors are undoubtedly one of the most promising energy storage devices as they are safe to operate and demonstrate a potential for ultra-high power density storage, long cycle stability and fast response [4\u20136]. In supercapacitors, the physicochemical properties of electrode materials, such as morphology, microstructure and electrical conductivity govern their electrochemical performances [7\u20139]. For effective storage of electrical energy and its conversion to chemical energy, electrocatalytic water-splitting devices are considered to be a good option due to their high energy density and environmentally friendly processes [10\u201312]. Considering the requirements of successful implementation of electrolysis in practice, effective high performance electrode materials with capabilities for fast electron transfer and high electrochemical activity are solicited. In the past few decades, research on multifunctional energy storage and conversion materials has mainly focused on transition metal oxides, hydroxides and sulfides [13\u201316]. However, the relatively low electrical conductivity and poor long-term stability limit their wider applications. Therefore, it is highly desirable to design and develop multifunctional materials with enhanced activity, high charge storage capacity and long cycling stability for building next-generation supercapacitors and electrocatalysts [17].Transition metal phosphide, as a promising candidate material, has attracted much attention in recent years [18\u201321]. Due to the good thermal stability and electrical conductivity, they are beneficial both for energy storage and conversion applications [22\u201325]. Amongst numerous candidates, the phosphate nickel-cobalt bimetallic compound (P-(Ni, Co)) has a higher electrical conductivity, better electrochemical activity and charge storage capacity, rendering them suitable as a potential multifunctional electrode material for next-generation applications [26,27]. A few exploratory reports are available in the literature including the work of Chen et\u00a0al. who reported three-dimensional (3D) hierarchical NiCoP@CoS tree-like core-shell nanoarrays designed for the one-dimensional core to act as \u201chyperchannel\u201d for electron transfer and the two-dimensional vertical shell for fast ion diffusion; the composite exhibited battery-type electrochemical performance, including high specific capacitance, enhanced rate capability and good cycling stability [28]. Wang's group fabricated flexible NiCo2O4-P (phosphated nickel cobalt oxide) on carbon filaments, that exhibited improved oxygen evolution reaction (OER) and hydrogen evolution reaction (HER) properties argued to arise from the heteroatomic doping and improved active sites (due to nanoparticles), leading to an improvement of conductivity thus promoting faster electron transfer during water splitting [29]. Lee et\u00a0al. reported a branched spike-covered hollow structure of NiCo2Px with enhanced HER and OER performances in a wide range of pH [30]. Yan et\u00a0al. explored complex p-n heterojunction nanorod arrays Co-Pi/Co3O4/Ti:Fe2O3 (cobalt phosphate/cobalt oxide/titanium-doped ferric oxide) that showed improved photocurrent density and water oxidation kinetics; this was reported to occur due to the faster removal of photogenerated holes, thus increasing the charge injection and bulk separation efficiency, leading to a facile charge transfer and separation of electron-hole pairs [31]. P-(Ni, Co) bimetallic compound has a prominent redox reaction site due to the synergistic effect of nickel and cobalt catalysts and is considered to be excellent electrode material for supercapacitors as well as a good electrocatalyst. Substitution by metal ions (Xn+) with greater electronegativity and lower pKa of [X(H2O)m]\nn\n\n+ has been demonstrated to shift the formal redox potential of parent metal positively in transition metal complexes and (hydr-)oxides due to inductive effects leading to greater ORR/OER activity, attributed to optimized binding of the reaction intermediates on the surface in rate-limiting steps [32]. However, its further utilization is still limited due to low rate capability and poor cycling stability resulting normally due to the structural collapse during repeated fast redox processes.In general, substrates like nickel foam (NF) have been widely utilized due to their high electrical conductivity and strong mechanical flexibility [33\u201335]. Furthermore, electrode materials with large open frames, such as Prussian Blue (PB), have faster ion diffusion paths [36]. The PB crystal structure can be adjusted due to the characteristics of polynuclear metals that cause their internal metal ions to be replaced by alkali metal ions with different valence states [37]. Iron (Fe) atoms of Prussian blue can be replaced by metal ions like nickel (Ni), cobalt (Co), copper (Cu), zinc (Zn) and manganese (Mn), and the physicochemical characteristics such as morphology, crystals, and internal pore structure can be altered. Open framework structures like Co-Fe mixed oxides with retainable architectures have already been demonstrated to exhibit superior OER performance [38,39]. More importantly, Prussian blue analogues (PBA) can be synthesized by a very simple method leading to their wide application as multifunctional electrode materials for energy storage and water splitting [40\u201342]. Modification of PBAs as phosphates leads to further improvement in their electrochemical performances [43\u201345]. Therefore, the design starting from the growth of the first layer of 1D NiCo nanowires (NiCo NWs), followed by a layer of Prussian blue analogues with a third layer of phosphides on a 3D conductive NF substrate to form a 3D complex network seems to be a promising approach for enhancing the conductivity and structural stability. Thus, a 3D core/shell nanostructure of P(Ni, Co, Fe) on the NF substrate results in a self-supported binder-free multifunctional electrode with improved surface-active sites providing a transport path to more electrolyte ions thus effectively boosting electrochemical performances of the electrode.In this work, a hierarchical P(Ni, Co, Fe) nanoarrays electrode was obtained by phosphating NiCo NWs/NiCo-PBA core/shell nanoarray structures fabricated through an in situ self-sacrificial growth process. The NiCo NWs/NiCo-PBA core/shell nanoarrays were synthesized through a two-step process. The NiCo NWs core to boost the charge transport and improve conductivity was synthesized by hydrothermal method, whereas the NiCo-PBA shell that contributes to increasing the surface-active sites for redox reactions was obtained by a chemical bath method, utilizing the core materials as a self-sacrificial source for Ni and Co. This composite electrode exhibits superior specific capacitance, good rate capability and improved cycling stability. Furthermore, the hybrid structure electrode shows a promising application for OERs with excellent electrocatalytic activity. This work reveals a simple and effective solution for the fabrication and design of high-performance multifunctional electrode materials for supercapacitors and enhanced OER activities.NiCo NWs were prepared via a hydrothermal process (Scheme\u00a01). Prior to the synthesis process, nickel foam was cleaned with 1\u00a0M HCl, ethanol, and deionized (DI) water under ultrasonication for 15\u00a0min, respectively, and then dried at 60\u00a0\u00b0C. For the synthesis of NiCo NWs, 1\u00a0mmol Ni(NO3)2\u00b76H2O and 2\u00a0mmol Co(NO3)2\u00b76H2O were dispersed in 50\u00a0ml DI water at room temperature with continuous stirring. Then 4\u00a0mmol urea was added to the above solution under strong magnetic stirring for 30\u00a0min. The derived solution was then transferred into a sealed autoclave containing a piece of clean nickel foam (NF, 3\u00a0\u00d7\u00a05\u00a0cm2) and maintained at 120\u00a0\u00b0C for 6\u00a0h. After cooling down to room temperature, the obtained product was cleaned with absolute ethanol and DI water several times followed by overnight drying at 60\u00a0\u00b0C prior to further use.A piece of NF with NiCo NWs nanoarrays was put into 80\u00a0mL potassium ferricyanide (K3Fe(CN)6) (1\u00a0mM) solution, and then heated at 60\u00a0\u00b0C for 10\u00a0h in an atmospheric oven. After cooling down to room temperature in the ambient, the obtained product was washed and dried at 60\u00a0\u00b0C overnight.P(Ni, Co, Fe) nanoarrays were fabricated through a gentle phosphorization process. The mass ratio of NaH2PO2\u00b7H2O and NiCo NWs/NiCo-PBA nanoarrays was about 10:1. Typically, NaH2PO2\u00b7H2O powder was put at the upstream side and center of the tube furnace, while the sample was placed at the downstream side and about 2\u00a0cm away from the hypophosphite powder. Then, the furnace was heated to 350\u00a0\u00b0C at a heating rate of 5\u00a0\u00b0C\u00a0min\u22121 and maintained for 2\u00a0h under a nitrogen atmosphere to obtain P(Ni, Co, Fe) nanoarrays upon naturally cooling down the samples to room temperature.The morphologies of the materials were observed by scanning electron microscopy (SEM, ZEISS, Ultra 55) equipped with Energy-dispersive X-ray (EDX) spectroscopy. The crystal structure was evaluated by an X-ray diffractometer (XRD, PANalytical X\u2019 Pert PRO, Netherlands). Fourier transform infrared spectroscopy (FTIR, Nicolet iS10) was used to detect the functional groups on the prepared samples.All the electrochemical performances of the electrode were carried out in a three-electrode system by using an as-prepared electrode as the working electrode, platinum wire as the counter electrode, and saturated calomel electrode as a reference electrode (SCE) with 6\u00a0M KOH aqueous solution as the electrolyte. Cyclic voltammetry (CV), galvanostatic charge-discharge (GCD), electrochemical impedance spectroscopy (EIS) and long-term cycling measurements were performed using Gamry Interface 1000 electrochemical workstation.Commercial activated carbon (AC) was used to prepare the negative electrodes as follows: AC powder, carbon black and polyvinylidene fluoride in N-methyl-pyrrolidone were mixed in a ratio of 8:1:1 to obtain the electrode slurry. Then, the slurry was coated on Ni foam substrate and dried overnight at 120\u00a0\u00b0C in a vacuum oven. Asymmetric devices were assembled with P(Ni, Co, Fe) cathode and AC anode. According to the charge balance, the mass ratio of the two electrodes is around 20:1.The specific capacitances (Cs\n, F g\u22121) of the electrodes from CV and GCD tests were calculated according to the following equations:\n\n(1)\n\n\n\nC\ns\n\n=\n1\n/\n\n(\n\nm\n\u00b7\n\u03c5\n\u00b7\n\n\u0394\n\nV\n\n)\n\n\u00b7\n\n\u222b\n\n\nV\n0\n\n\nV\n\nI\n\n(\nV\n)\n\nd\nV\n\n\n\n\n\n\n(2)\n\n\n\nC\ns\n\u2032\n\n=\n\n(\n\nI\n\u00b7\n\n\u0394\n\nt\n\n)\n\n/\n\n(\n\nm\n\u00b7\n\n\u0394\n\nV\n\n)\n\n\n\n\n where m (g) is the mass of active electrode material, \u03bd (mV s\u22121) is the scan rate, \u0394V (V) is the potential window, \n\n\n\u222b\n\n\nV\n0\n\n\nV\n\nI\n\n(\nV\n)\n\nd\nV\n\n is the integrated area under the CV curve, I (A) is the current during the discharge process, and \u0394t (s) is the discharged time.The energy (E, Wh kg\u22121) and power (P, W kg\u22121) densities of the asymmetric supercapacitor were determined from Eqs.\u00a0(3) and (4), respectively:\n\n(3)\n\n\nE\n=\n1\n/\n3.6\n\u00b7\n1\n/\n2\n\u00b7\n\nC\nS\n\u2032\n\n\n\u0394\n\n\n\nV\n\n\n\n\n\u2032\n\n2\n\n\n\n\n\n\n\n\n(4)\n\n\nP\n=\n3600\n\u00b7\nE\n/\n\n\u0394\n\nt\n\n\n\n\n where Cs\u2019 (F g\u22121) is specific capacitance calculated from the GCD test, \u0394V\u2019 (V) is the operating voltage, and \u0394t (s) is the discharging time.For the determination of electrocatalytic performance, the working electrode was the catalyst material, the counter electrode was a platinum wire and the reference electrode used was mercury-mercury oxide (Hg/HgO) (1\u00a0M KOH). The electrolyte was an alkaline solution of 1\u00a0M KOH with a pH of 13.6. Prior to each test, the electrolyte was degassed to remove oxygen by bubbling nitrogen gas for 30\u00a0min. All the potential shown below were transformed to the reversible hydrogen electrode (RHE) by using the equation \n\n\nE\n\nR\nH\nE\n\n\n=\n\nE\n\nH\ng\n/\nH\ng\nO\n\n\n+\n0.098\n\nV\n+\n0.0591\n\n\u00d7\np\nH\n\n.Linear sweep voltammetry (LSV) was carried out in the window potential of 1\u223c1.7\u00a0V vs RHE at a scan rate of 5\u00a0mV\u00a0s\u22121. The overpotential (\u03b7\u00a0=\u00a0ERHE \u2013 1.23\u00a0V) was calculated at current densities of 100\u00a0mA\u00a0cm\u22122 and 300\u00a0mA\u00a0cm\u22122, respectively. Tafel slope (b) was calculated according to the Tafel equation \n\n\u03b7\n=\nb\n\nl\no\ng\n(\ni\n)\n+\na\n\n, where i represent the current density. EIS was measured at 1.6\u00a0V (vs RHE) with an amplitude of 5\u00a0mV and a frequency range between 100\u00a0kHz to 0.01\u00a0Hz.Electrochemical surface area (ECSA) was estimated by considering a double-layer capacitance \n\n(\n\nC\n\nD\nL\n\n\n)\n\n according to Eq.\u00a0(5) where \n\n\n\n\nC\n\n\n\u2033\n\n\n\nS\n\n is a general specific capacitance with a typical value reported for the alkaline electrolyte of 40\u00a0\u00b5F\u00a0cm\u22122\n[46\u201348]. To determine the double layer capacitance, a series of CV curves were measured between 0.4 \u223c 0.9\u00a0V at scan rates of 10, 20, 30, 50, 80 and 100\u00a0mV\u00a0s\u22121. Plotting half the charge current (at 0.52\u00a0V vs RHE) against the scan rate yields a straight line with the slope equaling the double-layer capacitance.\n\n(5)\n\n\nE\nC\nS\nA\n=\n\nC\n\nD\nL\n\n\n/\n\n\n\nC\n\n\u2033\n\n\n\nS\n\n\n\n\n\nChronopotentiometry was performed to evaluate the stability of the electrodes when they are under a current density of 100\u00a0mA\u00a0cm\u22122 for 240\u00a0min. LSV measurement was recorded after the stability tests and the overpotential was compared before and after the testing of the stability of all the electrodes.As shown in Scheme\u00a01, the hierarchical P(Ni, Co, Fe) nanoarray electrodes were obtained by phosphating NiCo NWs/NiCo-PBA core/shell nanoarrays structure fabricated through an in situ self-sacrificial growth process. The NiCo NWs/NiCo-PBA core/shell nanoarrays were synthesized through a two-step method. During the hydrothermal growth of NiCo NWs core, the Ni ions cooperate with Co ions to form the NiCo NWs. Then, the NiCo-PBA nanoparticles shell was formed in a chemical bath, with the former core serving as a self-sacrificial source of Ni and Co ions inducing the in-situ growth of the shell. Finally, a phosphating step induces an ultrathin layer of P(Ni, Co, Fe) on the surfaces of the first and second layers. This 3D hierarchical overlapped triple-layered composite can potentially possess more surface/interface, higher active sites and better electron transfer efficiency with a more effectively modulated electronic structure.The morphologies and microstructures of the synthesized materials were studied by scanning electron microscopy (SEM). As shown in Fig.\u00a01\na and b, quasi-vertical densely packed nanowires are uniformly anchored on the NF substrate, which forms interconnected and overlapping charge transfer pathways in the nanowires. Anisotropic NiCo NW arrays of 1\u20133\u00a0\u03bcm length and \u223c 30\u00a0nm width are formed. The morphologies of the as-prepared NiCo NW nanoarrays synthesized at 120\u00a0\u00b0C for different reaction times are shown in Fig. S1(\u2020ESI). The Morphology tends to change from nanoflakes to nanosheets, mixed nanosheets/nanowires to only nanowires, governed by the crystallization growth processes. Furthermore, the energy storage performance of the electrode fabricated for 6\u00a0h is obviously superior compared to the others reported here (Fig. S2, \u2020ESI); thus, the optimal preparation condition of 120\u00a0\u00b0C/6\u00a0h was chosen for further modifications in this work. The in situ induced second layer NiCo-PBA nanocubes have a size of \u223c100\u00a0nm encapsulated on each nanowire like Chinese candied haws (Fig.\u00a01c and d). Fig. S3 (\u2020ESI) shows the corresponding Energy-dispersive X-ray (EDX) mapping of the NiCo-PBA nanoarrays. The homogeneous distribution of the Ni, Co, Fe, C and N elements intuitively suggests that the core/shell nanostructure is composed of NiCo NW cores and PBA nanoparticle shells due to the nature of the synthesis of these materials.Furthermore, with the introduction of P, some of the interconnected nanowire and nanoparticle joints are welded by the conductive phosphate layer to form interconnected and overlapped clusters with the appearance of a rougher surface of the P(Ni, Co, Fe) nanoarrays. (Fig.\u00a01e and f). Furthermore, at higher magnification, the nanowire skeleton can be clearly observed, indicating the preservation of the nanowire structure. The conductive and interconnected phosphate layer can effectively increase the electron transfer capability of the composite. The flexible coating layer can also relieve the stress from volume changes, thus improving the energy storage and conversion performances that require cyclic charging and discharging sequences. The corresponding EDX mapping (Fig.\u00a01g-l) of the P(Ni, Co, Fe) nanoarrays shows the homogeneous distribution of the Ni, Co, Fe, C, N and P elements, while the presence of the Fe, C and N elements also indicates that the PBA nanoparticles are uniformly distributed. Therefore, it can be confirmed that this complex 3D hierarchical overlapped triple-layer composite may be composed of NiCo NWs core, NiCo-PBA nanoparticles second shell and the outermost phosphide layer as hypothesized.X-ray diffraction (XRD) was employed to study the crystal structure of the prepared materials. As shown in Fig.\u00a02\na and S4a (\u2020ESI), the three strong peaks located at 44.51\u00b0, 51.85\u00b0 and 76.37\u00b0 can be ascribed to the NF substrate (JCPDS card No. 04-0850) [49]. It can be seen that all the peak intensities seem to be very weak and the noise is high, making it difficult to distinguish the different phases except for the NF substrates. This is due to the strong signals from the NF substrates, the complex composition of the composite and the relatively low crystallinity. For clear comparison to eliminate the interference of the strong NF peak and confirm the crystal phase, the crystal structures of all kinds of powder materials by using the same synthesized methods without NF substrates were collected and shown in Fig. S4b (\u2020ESI), which is in good accordance with the results shown in Fig.\u00a02a. Combined with the analysis of all XRD results, it is known that the peaks located at 17.1, 26.5, 33.7, 39.5, 44.4, 47.1 and 62.1\u00b0 can be indexed to (0 2 0), (2 2 0), (3 0 0), (2 3 1), (0 5 0), (3 4 0) and (4 5 0) planes of NiCo carbonated hydroxides (JCPDS card No. 16-0164 for nickel carbonated hydroxide and JCPDS card No. 48-0083 for cobalt carbonate hydroxide, respectively), which also demonstrates the successful formation of NiCo compound on the NF substrate [50]. In addition, the successful growth of the shell materials NiCo-PBA nanoparticles on the NiCo NWs core by the in situ induction is also demonstrated. The extra peaks located at 17.2, 24.3, 34.6 and 38.6\u00b0 correspond to the (2 0 0), (2 2 0), (4 0 0) and (3 3 1) planes of PBAs (JCPDS card No. 46-0908 for Ni2Fe(CN)6\u00b70.5H20 and JCPDS card No. 46-0907 for Co3[Fe(CN)6]2\u00b710H20, respectively). After phosphating, the peaks located at 31.8, 41.1, 48.2 and 54.3\u00b0 can be indexed to planes of phosphates (JCPDS card No. 13-0213 for NiP2, JCPDS card No. 29-0497 for CoP and JCPDS card No. 34-0996 for FeP4, respectively).In order to understand the bonding in the complex structure of the P(Ni, Co, Fe) nanoarrays, Fourier transform infrared (FTIR) spectra were recorded from 750 to 4000\u00a0cm\u22121 (Fig.\u00a02b). The peaks located at 1379.8, 1067.4 and 827.8\u00a0cm\u22121 can be assigned to arise from symmetrical stretching vibration, stretching vibration and bending vibration of CO3\n2\u2212, respectively, confirming the existence of CO3\n2\u2212 [51,52]. The peaks centered at 3598.5, 3496.3 and 1484.4\u00a0cm\u22121 can be ascribed to OH stretching and bending vibrations in water molecules [50]. Moreover, the weak peaks located under 1000\u00a0cm\u22121 reveal the existence of metal-OH bonds [53]. These results are consistent with the XRD analysis, confirming the compositions of NiCo NWs. Compared to the FTIR results from NiCo NWs, the peaks centered around 2091.9\u00a0cm\u22121 is the most prominent corresponding to the Fe-CN-M band in PBAs [54]. After phosphating, all the peaks are weakened and are broader, which suggests the complex composition and low crystallinity of the composite, which again is in agreement with the results obtained from XRD studies. Thus we have demonstrated the successful formation of the complex 3D hierarchical overlapped triple-layer P(Ni, Co, Fe) nanoarrays, which may possess high surface/interface active sites, high electron transfer efficiency, stable structure, and thus can be used as multifunctional electrodes.In order to evaluate the electrochemical performances of the electrodes, a series of comparative studies were designed with the binder-free NiCo NWs, NiCo NWs/NiCo-PBA and P(Ni, Co, Fe) electrodes in 6\u00a0M KOH aqueous solutions in a standard three-electrode system. Fig.\u00a03\na presents the cyclic voltammetry (CV) curves of these electrodes with the potential range of 0 \u223c 0.6\u00a0V at 20\u00a0mV\u00a0s\u22121. Obviously, the P(Ni, Co, Fe) electrode has the largest peak current and integration area, indicating the highest capacitance of the complex 3D hierarchical overlapped triple-layer composite. In addition, compared to a single crystal phase, the positions of the redox peak pairs of the multi-phased materials increase but shift, showing more redox reaction sites, leading to a higher capacitance. Fig.\u00a03b describes the CV curves of the P(Ni, Co, Fe) electrode at different scan rates from 1 to 20\u00a0mV\u00a0s\u22121. An obvious couple of redox peaks reveal the typical electrochemical behaviors generated from Faradaic reactions related to Ni-O/Ni-O-OH and Co-O/Co-O-OH [55,56]. With the increase in the scan rates, a slight shift of redox peaks can be observed, which indicates good electrochemical reversibility.The galvanostatic charge-discharge (GCD) curves at 5\u00a0mA\u00a0cm\u22122 were measured as shown in Fig.\u00a03c in order to further confirm the practical usefulness of these electrode materials. All the GCD curves have approximately symmetric charge and discharge time, which also indicates good reversibility. Obviously, the P(Ni, Co, Fe) electrode has the longest discharge time, demonstrating the highest specific capacitance. Fig.\u00a03d and 3e show the GCD curves of the P(Ni, Co, Fe) electrode between 0 and 0.5\u00a0V at different current densities. The trend of the GCD curves follows same trend as results obtained in the CV measurements. GCD curves of NiCo NWs and NiCo NWs/NiCo-PBA electrodes are provided in Fig. S5 (\u2020ESI) for comparison. Obviously, the P(Ni, Co, Fe) electrode exhibits a much higher capacitance as presented in Fig.\u00a03f. The specific capacitance of the P(Ni, Co, Fe) electrode is calculated to be 1125.8\u00a0F\u00a0g\u22121 (3.7\u00a0F\u00a0cm\u22122) at a current density of 2\u00a0mA\u00a0cm\u22122, which is approximately twice that of the NiCo NWs (524.9\u00a0F\u00a0g\u22121, 1.8\u00a0F\u00a0cm\u22122), and about 4 times higher than that of the NiCo NWs/NiCo-PBA (275.6\u00a0F\u00a0g\u22121, 1.4\u00a0F\u00a0cm\u22122). Moreover, a high specific capacitance of 565.1\u00a0F\u00a0g\u22121 is maintained for the P(Ni, Co, Fe) electrode even at a high current density (20\u00a0mA\u00a0cm\u22122), which suggests an excellent rate capability. These superior electrochemical performances of the P(Ni, Co, Fe) electrode are obtained due to the complex 3D hierarchical overlapped triple-layer nanoarrays structure.\nFig.\u00a04\na shows the corresponding relationship between the log (scan rate) and log (peak current) calculated from CV measurements (the inset in Fig.\u00a04a, CV curves of the P(Ni, Co, Fe) electrode at scan rates from 0.2 to 2\u00a0mV\u00a0s\u22121) to explore the energy storage mechanism of the electrochemical performances following the relationship as given in the equation: i=a\u03bdb\n, where a and b are constants, and i is the current. The b value can reflect the energy storage mechanism. When it is close to 0.5, it means the diffusion is the main process; when it is approximately next to 1, it stands for the capacitive behavior [7,57]. As shown in Fig.\u00a04a, we obtain b values close to 1 and the good linear relationship both indicate that the as-prepared P(Ni, Co, Fe) electrode is not controlled by the ion diffusion, revealing the capacitive characteristics.Electrochemical impedance spectroscopy (EIS) was studied to determine the ion transport kinetics as shown in Fig.\u00a04b. The inset shows the equivalent series circuit, and the corresponding resistance of each electrode is shown in Table S1 (\u2020ESI). Fig.\u00a04b shows that the Nyquist plot of the P(Ni, Co, Fe) electrode has the smallest semicircle in the high-frequency region, indicating the smallest charge transfer resistance at the interface between the electrolyte and electrode, which may be caused by the improved activity of the phosphate layer and the effective charge transfer efficiency, compared to the other electrodes studied here [58]. The 45\u00b0 Warburg region at the middle frequency region is related to the diffusion of the electrolyte ions into the bulk of electrodes. Compared to the NiCo NWs electrode, the NiCo NWs/NiCo-PBA electrode shows no obvious increase, indicating that the diffusion resistance in the material is not significantly reduced due to the lower conductivity of the PBA nanoparticles [59]. Obviously, the P(Ni, Co, Fe) electrode presents the fastest migration and diffusion after phosphating due to the introduction of P and the complex 3D hierarchical nanostructures that enhance the electrical conductivity of the material. Of course, the highest slope of the P(Ni, Co, Fe) electrode at low frequencies indicates the best capacitive performance. As shown in Fig.\u00a04c, the cycling stability of the P(Ni, Co, Fe) electrode is explored to evaluate its potential for application in practical devices. The P(Ni, Co, Fe) electrode possesses much better cycling stability with capacitance retention of 97.1% after 5000 cycles at 50\u00a0mA\u00a0cm\u22122 and 89.9% after continuous 5000 cycles at 100\u00a0mA\u00a0cm\u22122, which reveals the enhanced structural stability of the 3D hierarchical nanoarrays, and thus effectively relieving the volume expansion during the repeated charge-discharge process. The inset in Fig.\u00a04c presents the GCD curves before and after long-term cycling, and the little-changed shapes indicate the fast rate capability and good stability. Moreover, the insert in Fig.\u00a04d shows the SEM image of the complex composite after long-term cycling tests, which demonstrates the integrity of the synthesized electrode materials.To further confirm the usefulness of these materials for practical applications, the as-prepared P(Ni, Co, Fe) electrode was used as the cathode to construct an asymmetric supercapacitor (ASC) with activated carbon cloth as anodes. In Fig. S6a-d (\u2020ESI) we have plotted the electrochemical performances of the AC anode in 6\u00a0M KOH. In Fig. S6e (\u2020ESI) the comparison CV curves of P(Ni, Co, Fe) and AC electrodes at 20\u00a0mV\u00a0s\u22121 were presented. Fig. S6f (\u2020ESI) presents the CV curves of the ASC device at 10\u00a0mV\u00a0s\u22121 under different applied voltages which allowed us to choose a working voltage of 1.6\u00a0V for further studies. In Fig.\u00a05\na the schematic diagram of the device structure is presented while in Fig.\u00a05b the CV curves of the ASC device at different scan rates are shown and in Fig.\u00a05c the GCD curves at different current densities are plotted. It can be immediately observed that the CV curves have similar shapes, indicating fast electron transport in the devices. The capacitance is about 42.1\u00a0F\u00a0g\u22121 at 0.1\u00a0A\u00a0g\u22121 (Fig.\u00a05d), which may be limited by the lower capacitance of the AC anode. As shown in Fig.\u00a05e, the capacitance retention is 93.9% after 5000 cycles at 5\u00a0A\u00a0g\u22121, demonstrating good cycling stability, which is further confirmed by the CV and EIS results before and after cycling (the insets in Fig.\u00a05e). After a long-term charging/discharging cycling, there seems to be a slight upward trend in capacity, probably because of the improved wettability and permeability of the electrolyte into the bulk of the electrode. Moreover, from the Ragone plots as shown in Fig.\u00a05f, the estimated energy density is 15.0\u00a0Wh\u00a0kg\u22121, and the maximum power density is 4000\u00a0W\u00a0kg\u22121. Two ASC devices were connected in series to successfully power different colored LED indicators after charging at 10\u00a0mV\u00a0s\u22121, showing the promise for practical applications (inset of Fig.\u00a05f).These electrodes were also evaluated for oxygen evolution reaction (OER) in alkaline media (1\u00a0M KOH aqueous solution). Linear sweep voltammetry (LSV) curves are shown in Fig.\u00a06\na. It is possible to notice two extra oxidation peaks at 1.30 and 1.35\u00a0V before the OER takes place, corresponding to the oxidation of Co2+/Co3+ and Ni2+/Ni3+, respectively. These transformations are inevitable, and other works described such reactions enhancing the performance of the electrodes. Wu et\u00a0al. found that the incorporation of trivalent nickel (Ni3+) was related to the improvement of OER activity and electrode stability [60]. Meanwhile, Menesez et\u00a0al. found that Co3+ acts as an active site enhancing the electrophilicity of adsorbed O facilitating the formation of OOH species [61]. LSV curves also indicate a significant improvement in the electrocatalytic activity of the P(Ni, Co, Fe) electrode reaching an overpotential value of only 252\u00a0mV at 100\u00a0mA\u00a0cm\u22122\nFig.\u00a06b), which presents a better performance compared to the classical ruthenium oxide (RuO2) electrocatalyst for OER that shows an overpotential of 270\u00a0mV at 50\u00a0mA\u00a0cm2\n[62]. Moreover, the Tafel slope (Fig.\u00a06c) shows that the overpotential needed to increase the production rate of oxygen is lower for P(Ni, Co, Fe) (68\u00a0mV\u00a0dec\u22121), followed by NiCo NWs/NiCo-PBA and NiCo NWs with values of 106 and 113\u00a0mV\u00a0dec\u22121, respectively. One of the main challenges of OER lies in its mechanism, which is a four-electron transfer through multi-step reaction pathways where several intermediates are formed. Researchers have proposed several paths explaining the mechanism of oxygen evolution [63]. Bockris and Otagawa studied the OER mechanism in perovskites, involving electron transfer steps and chemical steps (association and dissociation), establishing that the adsorption of the hydroxyl group on the electrode occurs as a first step followed by the rate-determining step (RDS) corresponding to the electrochemical desorption of OH forming hydrogen peroxide as an intermediate which decompose to yield oxygen [64]. Breaking the metal and oxygen group bond is involved in the RDS. The Tafel slope is closely related to the OER mechanism. From the literature, it is possible to find that when the Tafel slope is 120\u00a0mV\u00a0dec\u22121 the RDE corresponds to a single electron transfer reaction, but when this value is 60\u00a0mV\u00a0dec\u22121, RDE is the chemical reaction [65]. The most accepted OER mechanism in alkaline media is shown in Eqs. (6)-((10) where M corresponds to an active site and the place where the specie is adsorbed [66,67]:\n\n(6)\n\n\nM\n+\n\nO\n\n\nH\n\n\u2212\n\n\n\u2194\nM\n\u2212\nO\nH\n+\n\n\ne\n\n\u2212\n\n\n\n\n\n\n\n\n(7)\n\n\nM\n\u2212\nO\nH\n+\nO\n\n\nH\n\n\u2212\n\n\n\u2194\n\nM\n\u2212\nO\n+\n\nH\n2\n\nO\n+\n\n\ne\n\n\u2212\n\n\n\n\n\n\n\n(8)\n\n\nM\n\u2212\nO\n+\nM\n\u2212\nO\n\n\u2194\n2\nM\n+\n\n\nO\n2\n\n\n\n\n\n\n\n(9)\n\n\nM\n\u2212\nO\n+\nO\n\n\nH\n\n\u2212\n\n\n\u2194\nM\n\u2212\nO\nO\nH\n+\n\n\ne\n\n\u2212\n\n\n\n\n\n\n\n\n(10)\n\n\nM\n\u2212\nO\nO\nH\n+\nO\n\n\nH\n\n\u2212\n\n\n\u2194\n\nO\n2\n\n+\nM\n+\n\nH\n2\n\nO\n+\n\n\ne\n\n\u2212\n\n\n\n\n\nAs was mentioned before changes in the Tafel slope have been associated with the RDS during the OER process, and smaller values suggest that the RDS is at the end of the electron transfer reaction, being an indication of a good electrocatalyst [65]. Mary et\u00a0al. have associated a value of 40\u00a0mV\u00a0dec\u22121 for Ni films with an RDS corresponding to the formation of peroxide intermediates (M-OOH) [68]. Doyle et\u00a0al. indicated that a Tafel slope of 120\u00a0mV\u00a0dec\u22121 is most likely associated with the adsorption of OH\u2212 as a rate-limiting step for iron oxide electrodes, and values of 64\u00a0mV\u00a0dec\u22121 correspond to M-O as RDS for NiFe layered double hydroxide [69,70]. Therefore, in this work, the RDSs of NiCo NWs and NiCo NWs/NiCo-PBA are close to 120\u00a0mV\u00a0dec\u22121 indicating that the limiting step for both electrodes is the formation of M-OH. Meanwhile, the P(Ni, Co, Fe) electrode with a Tafel slope of 68\u00a0mV\u00a0dec\u22121 is closest to 64\u00a0mV\u00a0dec\u22121 leading to the intermediate formation of M-O as the rate-limiting step for the OER process thus explaining the obtained better catalytic activity. Tafel slope corresponds well with the increase of overpotential from 100 to 300\u00a0mA\u00a0cm\u22122 (Fig.\u00a06b) where P(Ni, Co, Fe) only need an increase of 31\u00a0mV, unlike for the other electrodes. This attractive electrode performance may be due to the higher interfacial charge transfer kinetics as was mentioned above (Fig.\u00a04b), and a better mass transfer due to the open structure of the electrode together with a high surface-active site density, which is comparable to that of most of high-end OER electrodes reported in the literature.The electrochemical surface area (ECSA) was evaluated by the electrochemical double-layer capacitance (CDL\n) calculated at 0.52\u00a0V from CV curves (Fig.\u00a06d and S7, \u2020ESI) at different scan rates. As shown in Fig.\u00a06e, the CDL\n of P(Ni, Co, Fe) is 49,160 \u00b5F, whereas that of the NiCo NWs/NiCo-PBA and NiCo NWs electrodes were 2010 and 688 \u00b5F, respectively. ECSA values calculated following Eq.\u00a0(5) are summarized in Table S2 (\u2020ESI). The P(Ni, Co, Fe) electrode shows a much higher specific area (83 m2 g\u22121) compared to the other electrodes fabricated in this study, indicating that the number of effective active sites for water oxidation increases after the phosphorization process leading to a better electrochemical oxygen evolution reaction activity. The electrocatalytic stability studied at 100\u00a0mA\u00a0cm\u22122 (Fig.\u00a06f) shows that the overpotential remains stable for 240\u00a0min under a continuous OER process for all the electrodes. A slight decrease in the working potential during the first 60\u00a0min for NiCo NWs and NiCo NWs/NiCo-PBA can be attributed to the activation of the catalyst with an increase in the active species participating in the OER, which is typically found in metal-hydroxide catalysts [71,72]. In contrast, the P(Ni, Co, Fe) electrode exhibits a lower and constant potential of 1.52\u00a0V during the chronopotentiometry measurement.During the chronopotentiometry test, the overpotential at 100\u00a0mA\u00a0cm\u22122 changes until an equilibrium is reached. LSV curves were measured before and after the chronopotentiometry test as shown in Fig.\u00a07\n\na. Overpotential increases by 41, 102 and 40\u00a0mV for NiCo NWs, NiCo NWs/NiCo-PBA and P(Ni, Co, Fe), respectively. The higher increase for NiCo NWs/NiCo-PBA is associated with the low stability of Prussian blue analogues in alkaline media [73,74]. Fig S8 (\u2020ESI) shows a small increase in P(Ni, Co, Fe) Tafel slope after the chronopotentiometry test, suggesting the nature of the material and its operation mechanism slightly changed. Overpotential increase is in agreement with what is found in the literature for CoMnP and NiCoP, where the deactivation of the materials is associated with the formation of oxides like metal oxides (MOx), phosphate (POx), or phosphides (PO) on the surface [75,76]. Nonetheless, the P(Ni, Co, Fe) electrode still shows a much lower overpotential than the other two electrodes, revealing its excellent OER electrocatalytic activity.To explain the charge transfer process during OER, EIS measurements were performed at a potential of 1.6\u00a0V vs RHE. Nyquist plots in Fig\u00a07b present a semicircle shape with a complete loop demonstrating that mass transport is not a limitation during the OER process. EIS data was modeled using the equivalent circuit inset in Fig.\u00a07b. This circuit includes a solution resistance (Rs), two parallel constant phase elements (CPEf and CPEct), one resistance associated with the substrate surface (Rf) and a charge transfer resistance (Rct) representing the resistance caused by the OER process on the catalyst interface. Table S3 (\u2020ESI) lists the values of the parameters for the fitting process. The inset figure in Fig\u00a07b shows the charge transfer resistance, which demonstrates that the Rct value (0.196 \u03a9) of P(Ni, Co, Fe) is much smaller than the values for NiCo NWs (1.257 \u03a9) and NiCo NWs/NiCo-PBA (0.975 \u03a9) suggesting a more efficient charge transport.This multifunctional electrode comprising of 3D hierarchical overlapped triple-layer nanoarrays structured P(Ni, Co, Fe), presents superior electrochemical performances including high specific capacitance, high rate capability, and low overpotential, that arise from: 1) the unique interconnected and overlapped nanowire structure anchored on the 3D NF substrate, the in situ induced PBA nanoparticles, and highly conductive phosphate layer provide multidimensional and multiscale synergistic effects, boosting the electron transfer, ion diffusion and catalytic reaction kinetics; 2) the multidimensional hierarchical structure with close contact interface/surface provides more edges and defects, thereby generating more effective surface area and active sites leading to higher electrocatalytic activity and the adsorption capacity of electrolyte ions; 3) the introduction of the phosphorous enhances the electrical conductivity of the material, reducing electron losses during charge transfer process, thus facilitating the redox reaction, improving the specific capacitance, and increasing the electrocatalytic activity of OER.A simple growth strategy to obtain a complex 3D hierarchical overlapped triple-layered nanoarray structured P(Ni, Co, Fe) electrode materials was developed by phosphating NiCo-NWs/NiCo-PBA core/shell nanoarrays structures fabricated through an in-situ self-sacrificial growing process. Owing to the unique multidimensional structure, more surface-active sites, improved conductivity, and enhanced catalytic reaction kinetics, the promising electrodes exhibit superior specific capacitance of 1125.8\u00a0F\u00a0g\u22121 (3.7\u00a0F\u00a0cm\u22122) at 2\u00a0mA\u00a0cm\u22122, good rate capability and improved cycling stability. The asymmetric device assembled with the hybrid electrode as cathode and activated carbon as anode show an energy density of 15\u00a0Wh kg\u22121 and a power density of 4000\u00a0W kg\u22121. Furthermore, the hybrid structure presents excellent oxygen evolution reaction performance with an overpotential of 252\u00a0mV at 100\u00a0mA\u00a0cm\u22122 and low Tafel slope of 68\u00a0mV\u00a0dec\u22121, and overall water splitting abilities with a low cell voltage of 1.52\u00a0V at 100\u00a0mA\u00a0cm\u22122 and stability for 4\u00a0h. The strategy presented in this work is simple, effective and novel for designing highly conductive multifunctional electrode materials for energy storage and electrocatalysis.\nXingyan Zhang: Conceptualization, Methodology, Formal analysis, Investigation, Visualization, Supervision, Writing \u2013 original draft. Mar\u00eda Isabel Alvarado-\u00c1vila: Formal analysis, Investigation, Visualization, Writing \u2013 original draft. You Liu: Software, Visualization. Dongkun Yu: Formal analysis. Fei Ye: Formal analysis. Joydeep Dutta: Resources, Supervision, Writing \u2013 original draft.The authors declare that they have no competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.X. Y. Zhang would like to thank Ragnar Holms Foundation for the Post-doctoral fellowship. M. I. Alvarado-\u00c1vila would like to thank the National Commission for Scientific and Technological Research, (CONICYT) for the Doctoral scholarship \u201cBeca Chile\u201d 2018-72190682. D. K. Yu would like to thank the China Scholarship Council (CSC) for the Doctor scholarship (202006360037). J. Dutta acknowledges the support of the MISTRA Terraclean project (Diary No. 2015/31) and Vinnova (Diary No. 2021-02313).Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.electacta.2022.141582.\n\n\nImage, application 1\n\n\n\n", "descript": "\n                  Highly efficient and environmentally friendly multifunctional electrode materials for application in supercapacitors to electrocatalysis are important for advances in the future of electrical energy storage and green hydrogen production. This work reports a simple growth strategy to obtain hierarchical P(Ni, Co, Fe) modified electrodes by phosphating a core/shell composite of nickel-cobalt (NiCo) Prussian blue analogues fabricated through an in situ self-sacrificial growth process. Due to the unique microstructure, abundant surface-active sites, and enhanced interfacial conductivity, the hybrid electrode exhibits specific capacitance as high as 1125.8\u00a0F\u00a0g\u00a0\u2212\n                     1 (3.7\u00a0F\u00a0cm\u22122) at 2\u00a0mA\u00a0cm\u22122, excellent rate capability and improved cycling stability (97.1% retention capacitance after 5000 cycles at 50\u00a0mA\u00a0cm\u22122 and 89.9% after continuous 5000 cycles at 100\u00a0mA\u00a0cm\u22122). Furthermore, the hybrid structure shows excellent oxygen evolution reaction performance with an overpotential of 252\u00a0mV at 100\u00a0mA\u00a0cm\u22122 and 283\u00a0mV at 300\u00a0mA\u00a0cm\u22122, with a low Tafel slope of 68\u00a0mV\u00a0dec\u22121, and overall water splitting abilities with a cell voltage of 1.55\u00a0V at 100\u00a0mA\u00a0cm\u22122. This work provides insights into the design of next-generation high-performance multifunctional electrode materials by controlling the surface/interface of multicomponent structures for enhancing their properties.\n               "}
{"full_text": "The renewability and sustainability issues of producing chemicals, materials and fuels from depleted fossil resources have greatly accelerated and expanded the development of renewable energies and resources [1\u20133]. As a highly-accessible and renewable natural carbon source, biomass is currently gaining traction as a feedstock for the manufacture of valuable chemicals, functionalized materials and high-energy-density biofuels [4\u20136]. The replacement of fossil-derived products with bio-based ones is of paramount importance for human society to develop sustainably. In this regard, the chemical processing of cellulose, the most plentiful component of biomass on the planet, is one of the most promising and appealing approach to produce value-added chemicals. Because the resultant products display comparative or ever superior characteristics in many ways to fossil-based counterparts [7\u20139]. It is well known that the hydrolysis of cellulose through chemical or biological approach results in the formation of D-glucose, which can be subsequently transformed into 5-(hydroxymethyl)furfural (HMF), a key biomass-derived platform molecule, via successive isomerization and dehydration process [10\u201312]. HMF contains aldehyde and alcohol group at its 2 and 5 positions, both of which are pretty strong reactivity in oxidizing and reducing environment [13,14]. Therefore, HMF as well as its derivate are currently utilized to produce a wide range of high-value bio-based chemicals, which are expected to (partially) replace voluminously consumed petroleum-based chemicals for the manufacture of fine chemicals, bio-fuels and plastics [15\u201317].Among various HMF valorization routes, selective oxidation of the aldehyde and/or alcohol group(s) of HMF can fabricate several valuable and intriguing furanic chemicals (Fig.\u00a01\n), such as 2,5-diformylfuran (DFF), 5-hydroxymethyl-2-furancarboxylic acid (HMFCA), formyl 2-furancarboxylic acid (FFCA), 2,5-furandicarboxylic acid (FDCA) and furan-2,5-dimethylcarboxylate (FDMC). These products have been extensively described as monomers for novel polymers, pharmaceutical intermediates, precursors for sustainable dyes and fungicides as well as platform chemicals for other value-added chemicals [18\u201324] (Fig.\u00a01). In particular, FDCA has a similar dicarboxylic acid structure to the fossil-derived terephthalic acid (TPA), thus it has been hailed as a renewable substitute to TPA for the manufacture of polyethylene terephthalate (PET)-analogous polymeric materials [25]. It is interesting to note that polyethylene furanoate (PEF), fabricated by the polymerization of FDCA and ethylene glycol, even discloses more attractive thermal properties concurrently with superior gases (O2, CO2, and H2O) barrier resistance with regard to PET, which enables it a promising packing material for food and beverage [26,27]. It has been estimated that an annual market volume for FDCA is approximately 50.5 million metric tons (MT) with an estimated value of $ 50.5 billion [28]. Recently, Avantium received a funding of 25 million Euros and plans to build an FDCA pilot plant with an annual output of 5000 tons from fructose in 2023 [29]. Novamont and Stora Enso also announced that they will invest 10 or 9 million Euros to build an FDCA pilot plant from fructose respectively [30,31].Recently, great efforts have been dedicated to developing new catalysts for the selective oxidation of HMF via thermo-catalytic, electro-catalytic or photo-catalytic approach [32,33]. There are several reviews documented the progresses in the selective oxidation of HMF [33\u201337]. In comparison with other methods, thermo-catalysis is the most widely investigated one because of its high catalytic efficiency, convenient operation and ample catalyst candidates. Especially, the heterogeneous catalyst holds prominent advantages over the homogeneous counterparts regarding the separation and reusability of the catalyst. It should be noticed that the selective oxidation of HMF to DFF, HMFCA, FFCA or FDCA over heterogeneous catalyst actually proceed via a similar reaction mechanism but with different oxidation degrees. In many cases, the same catalyst can be finely addressed to produce different oxidation products from HMF through manipulating the reaction parameters (reaction solvent, temperature, time and so on). For instance, a semicrystalline nanoporous multiblock copolymer matrix supported Au NPs (Au/sPSB) enabled the selective oxidation of HMF to DFF, HMFCA, FFCA, FDCA or FDMC with high yields respectively through judicious choice of the reaction condition and medium [38]. Thus, the design of efficient heterogeneous catalysts for the oxidation of HMF follows common principles in many cases no matter what the target product.In this review, we will present an overview of the established heterogeneous catalyst design strategy for the selective oxidation of HMF over transition metal (noble and cheap metal)-based catalyst as well as the metal-free catalyst. In contrast to the previously published reviews, we highlight the universal catalyst design strategy toward enhancing the catalytic performance of the catalyst for the oxidation of HMF. Particular attention will be focused on the following aspects: (1) the reaction mechanism of HMF oxidation over different catalysts; (2) the developed approaches aiming at boosting the catalytic activity and stability of the supported catalyst via manipulating the basic-acidic property, the redox ability, porosity, as well as the surface properties of the support; (3) enhancing the catalytic performance of the catalyst through introducing other constituents as the promoter to tailor the geometric as well as the electronic configurations of the active centers; (4) the reported catalytic systems for the one-pot oxidative esterification of HMF to FDMC.Noble metal-based catalysts, such as gold (Au), palladium (Pd), platinum (Pt), ruthenium (Ru) and silver (Ag) have been extensively studied for the selective oxidation of HMF for a long time because of their high activity for activating molecular oxygen (O2). Although these noble metal catalysts generally revealed high catalytic activity for HMF oxidation, their catalytic behaviors are quite diverse and strongly relate to the type of active metal species and the basicity of the reaction medium. For example, Ru-based catalysts were found to be more active for the conversion of HMF to DFF in the organic solvent than others [39,40]. A few effort has been devoted to investigating the mechanism of oxidation of HMF to DFF over Ru catalysts. Nie and co-workers well demonstrated that the oxidation of HMF to DFF over Ru/C in N,N-dimethylformamide (DMF) operated by the Langmuir\u2013Hinshelwood mechanism [40] (Scheme 2\n\n), in which HMF and O2 were firstly adsorbed and dissociated to corresponding alcoholate (R\u2013CH2O\u2217), hydrogen (H\u2217) and atomic oxygen (O\u2217) species over Ru/C surface, respectively. Then, a hydrogen of R\u2013CH2O\u2217 was abstracted by O\u2217 species, resulting in the generation of the adsorbed DFF (R\u2013CHO\u2217) and hydroxyl (OH\u2217) species. Subsequently, the resultant OH\u2217 species react with H\u2217 species to form the adsorbed water (H2O\u2217) species. Finally, R\u2013CHO\u2217 and H2O\u2217 quickly desorb from the Ru/C surface to finish the catalytic cycle. Kinetic isotope effects (KIE) were also investigated in this study and proved that the activation of C\u2013H bond is the kinetically-relevant step for the oxidation of HMF into DFF over Ru/C. A similar reaction process was also proposed by Sarmah and co-workers for the oxidation of HMF to DFF over Ru nanoparticle-supported H-beta catalyst (Ru/H-beta) [41].As for the consecutive oxidation of HMF to FDCA in aqueous solution with oxygen, two competitive reaction paths were reported according to the first step oxidation reaction, which can be dominantly proceeded either via oxidizing the aldehyde side chain of HMF to carboxyl group (Scheme 1, route 1) or converting its hydroxymethyl moiety to aldehyde group (Scheme 1, route 2). Then, the resultant intermediate (HMFCA or DFF) was oxidized to FDCA via FFCA. It has been widely observed that Au- and Pd-based catalysts offered higher catalytic activity for the oxidation of aldehyde group of HMF than that of hydroxymethyl side chain [33,34,42], thus the oxidation of HMF over Au- and Pd-based catalysts generally conducted via route 1 even under base-free conditions (Fig.\u00a02\n) [43\u201345]. In contrast, Ru catalysts favored the oxidation of hydroxymethyl moiety instead of the aldehyde group of HMF in the first oxidation step, therefore the oxidation of HMF to FDCA over Ru-based catalyst followed route 2 [46,47]. Moreover, the reaction path for the HMF oxidation also has been influenced by the basicity of the reaction medium. Pt and AuPd alloy catalysts preferred the oxidation of the aldehyde group of HMF than hydroxymethyl moiety in the first oxidation step (route 1) might because of the basic medium favoring hydration of the aldehyde to the gem-diol [48,49]. Whereas different reaction pathway (route 2) was observed for the oxidation of HMF over Pt and AuPd alloy catalysts under base-free condition [50\u201352].Even though the reaction path for the oxidation of HMF to FDCA over noble metal catalysts relies on the active metal species and the basicity of the reaction medium, the role of water or oxygen during the HMF oxidation process was demonstrated to be the same in different catalytic systems by isotope labeling technology. In 2012, Davis and co-workers detected 18O incorporated HMFCA and FDCA products when HMF oxidation reaction was performed over Au/TiO2 and Pt/C catalysts in H2\n18O under 16O2 pressure [53]. Whereas 18O atoms were not found to be incorporated into HMFCA and FDCA when analyzing the products of HMF oxidation under 18O2 pressure in H2\n16O. These results indicated that aqueous-phase oxidation of the aldehyde proceeds through a geminal diol generated by the reversible hydration reaction of aldehydes and water in a basic environment (Fig.\u00a02). Thus, water served as the direct oxygen source during the oxidation of the aldehyde side chain of HMF, DFF or FFCA to carboxyl group. Meanwhile, O2 was considered to play an indirect role during oxidation through scavenging the electrons deposited over the metal NPs to end the catalytic cycle. A similar reaction mechanism was also reported in many other works [46,54,55].To further elucidate the role of molecule oxygen during the HMF oxidation process, electron paramagnetic resonance (EPR) and density functional theory (DFT) calculations were utilized to detect and simulate the generated active oxygen species over noble metal NPs. Liu et\u00a0al. found that different active oxygen species was produced over different facets of Pt NPs, such as Pt (100) surface prefer to generate \u2027OH species whereas \u2027O2\n\u2212 species are largely formed on a Pt (111) surface, both of them could boost the dehydrogenation reaction by promoting the break of O\u2013H and C\u2013H bonds during HMF oxidation [48]. Lei and co-workers also revealed that oxygen could be hydrogenated to H2O2 over Pd (111) surface, which could remove the extra electrons in the HMF oxidation process and complete the catalytic cycle [56]. Thus, O2 participated in the reaction by forming active oxygen species, which are generally involved in the dehydrogenation process, or served as a scavenger to wipe out the deposited electrons over the metal NPs during the HMF oxidation process.However, there is a lack of studies that unravel the fundamental reasons for the different reaction paths for the oxidation of HMF over various noble metals, which is of great significance for the understanding of the reaction mechanisms. Investigating the adsorption behavior of the intermediates during the HMF oxidation process over different noble metal species may provide some useful information. Efforts are also supposed to be dedicated to gain more insights into the oxygen activation process over different noble metals.Stabilized metal nanoclusters (NCs) or nanoparticles (NPs) in the solution exhibit some peculiar advantages over the traditional supported metal clusters for the oxidation of alcohols because of their precisely controllable sizes and shapes as well as freely rotational and three-dimensional nature [57,58]. However, only a few works reported the selective oxidation of HMF over the dispersed noble metal catalytic systems. The first colloidal noble metal NPs for the HMF oxidation was reported by Siankevich et\u00a0al. [55] They prepared a series of polyvinyl pyrrolidone (PVP) protected Pt NPs (Pt-PVP-GLY) with different sizes (1.5\u20135\u00a0nm) for the oxidation of HMF in the base-free aqueous solution. The activity of the Pt NPs decreases with the increasing of the size of NPs and Pt NPs with the smallest size of 1.5\u00a0nm afforded the highest FDCA yield of 95%. Similarly, size-dependent effect on HMF oxidation activity was also found in PVP stabilized Pd nanoparticles (Pd NPs) [59].The morphology of the NPs in the solution phase can be easily manipulated by regulating the preparation conditions, thereby allowing one to better understand the structure\u2013reactivity relationship. Liu et\u00a0al. studied the effect of different exposed facets on the catalytic activity of Pt nanocrystals for the oxidation of HMF [48]. Interestingly, oxygen was activated to different active oxygen species over different facets. Specifically, Pt (100) surface prefers to generate \u2027OH species whereas \u2027O2\n\u2212 species are largely formed on a Pt (111) surface [48]. Density functional theory (DFT) calculations revealed that the energy barriers of O\u2013H and C\u2013H bond scission on \u2027OH-precovered Pt (100) surface are much lower than that on \u2027O2\n\u2212-precovered Pt (111) surface [48]. Therefore, Pt (100) surface displays a higher activity for the HMF oxidation than Pt (111). As for single-crystalline Pd nanocrystals, (111)-faceted nanooctahedra (Pd\u2013NOs) disclosed 2.6 times higher turnover frequency (TOF) than (100)-faceted nanocubes (Pd\u2013NCs) for the oxidation of HMF to FDCA [56]. Unlike the situation of Pt nanocrystals, DFT results indicated that oxygen tends to hydrogenate to H2O2 over Pd (111) surface, which can participate in the HMF oxidation process and complete the catalytic cycle [56]. In contrast, oxygen prefers to dissociate to the atom state over Pd (100) surface, which cannot serve as active oxygen species for the HMF oxidation reaction. These efforts provided some profound insights into the structure\u2013reactivity relationship of noble metal nano-catalysts for the oxidation of HMF. However, some issues deserved to be further studied in the colloidal noble metal NPs systems for the oxidation of HMF, such as the lack of in-depth understanding of the effect of different stabilizer, the relatively high catalyst loading and the requirement of base additive to reach a high FDCA yield in some cases.The immobilization of noble metal NPs on a support not only can improve their stability and manipulate their spatial distribution but also can enhance their activity through tuning the support properties and the metal-support interactions. The manipulation of the basic-acidic property, the redox ability, porosity, as well as surface property of the support can effectively boost the activity of the metal nanoparticles. One should be noted that these characteristics of the support do not separate from each other but actually act together to make the catalyst more active. However, to make the story clear and understandable, we have summarized the developed strategies for engineering each property separately as an individual section. Furthermore, the fabrication of supported bimetallic NPs is also an important strategy for tuning the geometric as well as the electronic configurations of NPs. The developed supported bimetal catalytic systems are also discussed in this review.The base/acid properties of the support are well demonstrated to play a critical role in the catalytic activity of the supported noble metal catalysts for the oxidation of HMF. In general, the reaction rate of the alcohol oxidation reactions can be enhanced in an alkaline environment or the presence of basic support. Zhu et\u00a0al. revealed that the strong basicity of Mg-Beta zeolites can promote the catalytic activity of Au/Mg-Beta catalyst for the HMF oxidation and reduce the dosage of additional base [60]. Because the basic environment could favor the formation of alkoxide intermediate as well as smooth the activation of C\u2013H in alcohol oxidation process [61,62]. Therefore, we begin our discussion with catalytic systems based on basic supports. As summarized in Table 1\n, typical basic support, including MgO, Mg(OH)2, hydrotalcite (HT) and MgAl2O4 etc. were widely employed for supporting Au, Ru, Pd and AuPd NPs for the oxidation of HMF. Interestingly, the employment of basic supports enabled the efficient oxidation of HMF to DFF or FDCA in the base-free condition in most cases (Table 1). Actually, the involvement of base additives during the HMF oxidation process raises many concerns regarding the product purification process and production costs. One should be noted that the activity of the catalyst strongly relates to the basicity of the basic support. Generally, strong basic supports (e.g. MgO, Mg(OH)2) afforded a relatively lower DFF or FDCA yield (selectivity) than that of supports with moderate basic sites (e.g. HT and MgAlO). Liu et\u00a0al. attributed the relatively lower activity of Ru/MgO than Ru/Mg2AlOx to the strong basic nature of the MgO, which promoted the undesirable degradation and polymerization by-reactions of the susceptive HMF [40]. However, this is probably not the only reason for the inferior performance of Ru/MgO. Antonyraj et\u00a0al. found that the surface area of Ru/MgO (73\u00a0m2\u00a0g\u22121) is much lower than Ru/MgAlO (200\u00a0m2\u00a0g\u22121) [63], and the larger Au particle size of Au/MgO (> 10\u00a0nm) than Au/HT (3.2\u00a0nm) was also observed [64], which are probably the other reasons for the inferior catalytic performance of Au/MgO for the conversion of HMF to FDCA.Even though high FDCA yields (95%\u2013100%) were accomplished over the basic support catalytic systems under base-free conditions, the generated acidic products (FDCA, FFCA and HMFCA) may react with the basic support, resulting in the etch of active component and support, which leads to the downgrade of the catalyst activity and stability. In particular, 26%\u201338%\u00a0mol/mol Mg of the support leached from Ru(OH)x/MgO and Ru(OH)x/HT catalyst after the reaction and the concentration of Mg2+ in the solution corresponded to the generated FDCA concentration, suggesting that the basic support may serve as a solid base to neutralize the produced FDCA and push the reaction forward [65]. Unexpectedly, HT supported Au [64], Pd [43] or Pt [50] catalysts showed similar FDCA productivity (5.7\u20136.9 molFDCA h\u22121 molmetal\n\u22121) and demonstrated to be relatively stable for the base-free oxidation of HMF to FDCA. An FDCA yield decrease of around 10% was observed after three or five runs with only a small amount of support leaching (around 3% Mg2+ etched after each run) for Au/HT [64]. That's to say, the transformation of HMF to FDCA can also proceed smoothly without stoichiometrical consumption of basic support, indicating that the basic support not just simply acted as solid base to react with FDCA to form FDCA salts. For example, Gupta et\u00a0al. found that the basicity of HT support accelerated the transformation of aldehydes of HMF or FFCA to hemiacetals intermediate as well as the generation of metal alkoxide species during HMF oxidation process [64]. Subsequently, Ardemani et\u00a0al. revealed that the moderate basic sites of HT support can react synergistically with the high fraction of surface gold to modulate the adsorption behavior of HMF and HMFCA intermediate, which afforded high activity for the oxidation of HMF under base-free condition (Fig.\u00a03\na) [66].After gaining some insights into the role of basic support during the base-free HMF oxidation process, several strategies were developed to prevent the leaching of the support and improving the life of the catalyst. The restriction of the interactions between the carboxylic acid products (HMFCA, FFCA and FDCA) and basic metal components of the support can significantly relieve the leaching of metal species. Gao et\u00a0al. reported that the introduction of La into AuPd/CaMgAl-layered double hydroxide (LDH) brings in new La3+-O2+ pairs on AuPd/LaCaMgAl-LDH catalyst, leading to the formation of abundant highly dispersed LaOx species over the support surface, which can enhance the stability of the catalyst by alleviating the strong interactions between acidic products (HMFCA, FFCA and FDCA) and metal species (Fig.\u00a03b) [67]. Therefore, the leaching of Mg and Ca species were significantly suppressed in AuPd/LaCaMgAl-LDH catalyst (0.8 and 0.3\u00a0wt%) in comparison with that of in AuPd/CaMgAl-LDH catalyst (4.5 and 2.6\u00a0wt%) [67]. In addition, the incorporation of an acid-resistant phase into the basic support is another approach to avoid the potential metal leaching. For instance, the synergy between hydrotalcite (HT) and activated carbon (AC) endows HT-AC composite with the characteristics of these two components and prevents the leaching of HT in acidic environment [68]. The dilution of Au/MgO with inert MgF2 phase also enabled Au/MgF2\u2013MgO to catalyze HMF oxidation under base-free condition with relatively high FDCA productivity of 15.8 molFDCA h\u22121 molAu\n\u22121 [45]. More importantly, the final pH value of the reaction solution was 3.8 whereas no Mg leaching was noticed, indicating the strong stability of Au/MgF2\u2013MgO catalyst in an acidic environment. To eliminate the support leaching issue at the source, many novel basic supports without basic metal oxides or hydroxides were developed in recent year, such as magnesium-doped carbon (C\u2013O\u2013Mg) [69] and DOWEX 50WX2-100 resin [70]. Especially, DOWEX 50WX2-100 resin supported Pt catalyst (Pt@Dowex-Na) provided high FDCA yield (99%) and excellent stability in neat water under continuous flow (FDCA yield remained 99% after 43\u00a0h time-on-stream and no Pt leaching was detected) [70]. Although only a few works focus on the oxidation of HMF in continuous mode, continuous-flow oxidation of HMF is important for the large-scale utilization of HMF.Amphoteric, acidic and inert supports, such as TiO2, ZrO2, Al2O3 and active carbon, etc., with better chemical stability than basic supports especially under acidic conditions, were also extensively investigated for supporting noble metal NPs. Generally, as summarized in Table 2, a certain number of base additives (e.g. NaOH, Mg(OH)2 or NaHCO3) are required to achieve high FDCA yields (62%\u201399%) over TiO2, ZrO2 and Al2O3 supported noble metal catalysts. When HMF oxidation reaction was performed over non-basic supports supported catalysts under base-free conditions, a very low FDCA yield was usually offered. For example, only 2%\u20135% FDCA yields were obtained from HMF over Au/TiO2, Pt/ZrO2 [49] or Au/ZrO2 [45] in the absence of base additives. Note that ring-opening products of HMF such as levulinic acid (LA) and formic acid (FA) were detected over TiO2, ZrO2 and Al2O3 supported Ru(OH)x under base-free conditions [71]. In general, the basic sites of the support are supposed to favor HMF oxidation reactions while their acidic sites may promote the undesirable HMF ring-opening by-reactions. Hence, base additives are essential for accelerating HMF oxidation reactions and suppressing HMF ring-opening reaction in the presence of TiO2, ZrO2 or Al2O3 supported noble metal catalysts. Moreover, TiO2, ZrO2 and Al2O3 supported Au and Pt catalysts revealed higher FDCA yields (95\u201399%) than Ru and Pd catalysts (62\u201386%) in alkaline solution [44,74\u201376]. Especially, Au/Al2O3 offered the highest FDCA productivity of 24.8 molFDCA h\u22121 molAu\n\u22121 among TiO2, ZrO2 and Al2O3 supported noble metal catalysts at a relatively low reaction temperature of 70\u00a0\u00b0C [77].Notably, many studies also pointed out that the Lewis acid site of the support (e.g. CeO2) can serve as the active site to adsorb the alcohol, then favor the formation of alkoxide intermediates during the alcohol oxidation process [78,79]. In addition, Odriozola and co-workers found a positive correlation between the Br\u00f8nsted acidity of CexZr1-xO2 (hydroxyl groups) and the corresponding FDCA yields of CexZr1-xO2 supported Au [80]. To be specific, CexZr1-xO2 support with higher acidity offered higher FDCA yields [80]. Recently, Ag NPs were found to be very active for the oxidation of HMF to HMFCA with unexpected selectivity [81\u201384]. In particular, the acidic ZrO2 supported Ag NPs offered the highest HMFCA yield of 92% among various supports with different acidic and basic properties (ZrO2, TiO2, CeO2 and MgO) [81]. Additionally, the modification of support with acidic groups (e.g. \u2013HSO3) enabled the direct transformation of carbohydrates to FDCA [85] or DFF [86] in a one-pot process. For example, Rathod and co-workers realized the one-pot two-step production of FDCA from fructose with an overall yield of 64% over a sulfonated glucose-derived carbon-supported Pd catalyst (Pd/CC) [85].Although the inert active carbon with stable physical properties renders them ideal candidate support for the oxidation of HMF under base-free conditions, the catalytic activity of active carbon-supported noble NPs was lower than that of metal oxide one [39,40,42,74] [87]. The introduction of heteroatom dopant (e.g. N and P) into the carbon matrix can significantly regulate the physicochemical property of the support, thereby improving the catalytic activity of the supported noble metal NPs. Nitrogen-doped carbon is a type of material with basic sites and stable under acidic conditions. The N dopant of the carbon support could strongly coordinate with noble metal species (e.g. Au [88], Pt [52], Ru [89] and Pd [90]) to enhance its dispersity, rendering the formation of ultrafine metal NPs with boosted activity and stability. For example, Han et\u00a0al. found that the activity of the nitrogen-containing carbon-supported Pt catalyst (Pt/C-EDA-4.1) was strongly related to the relative content of different types of nitrogen species, especially the concentration of basic pyridine-type nitrogen (N-6) [52]. Besides, it was reported that phosphorus can modify the 3\u00a0d electron density of noble metals and promote the activation of the substrate during the alcohol oxidation process [91,92]. Accordingly, highly porous nitrogen- and phosphorus-co-doped graphene sheets supported Pd catalyst (Pd/HPGSs) [93], P-decorated CNF supported Au\u2013Pt NPs (Au\u2013Pt/P-HHT-CNF) [94] and phosphorus-doped carbon-supported Ru catalyst (Ru/OMC-P0.56) [95] were reported to be more active for the oxidation of HMF than the catalyst without P-doping. However, in comparison with N-doped carbon supports, fewer efforts have been devoted to the study of the mechanism of P-doped carbon supported catalysts for the HMF oxidation, especially regarding the promotion effects of the P-doping for the oxidation of HMF.Reducible oxides are extensively employed for supporting noble metal NPs for the alcohol oxidation reactions because the oxygen vacancy (Ov) of the reducible oxides can effectively activate molecular oxygen, thereby working synergistically with noble metal NPs to boost the catalytic performance of the catalyst (Table 3). CeO2, Mn3O4, MnO2, NiO [100] and CoO are typical reducible oxides containing redox couples (e.g. Ce4+/Ce3+, Co3+/Co2+, Mn4+/Mn3+/Mn2+) and Ov. Casanova et\u00a0al. revealed that Au\u2013CeO2 [74] was very active for the oxidation of HMF to FDCA, offering an impressive FDCA productivity of 122.9 molFDCA h\u22121 molAu\n\u22121 (5\u00a0h, 130\u00a0\u00b0C, 10\u00a0bar Air), which is much higher than that of non-reducible oxides supported catalysts as reviewed in the 2.3.1 section. This work suggested that redox couples and Ov of the support were involved in the catalytic circles. Later, they further reported that the catalytic performance of the Au\u2013CeO2 for the oxidation of HMF can be improved by the reductive pretreatment of the catalyst, which could increase the amount of Ce3+ and oxygen vacancies of the catalyst, thus contributed to the O2 activation and dehydrogenation process during the HMF oxidation reactions [74].In general, the doping of heteroatom with different valences\u00a0in the reducible oxides can generate new surface defects to compensate for the charge imbalance, thus regulating the redox property of the support [101]. For instance, Miao et\u00a0al. obtained an enhanced FDCA yield of 99% over Ce0.9Bi0.1O2-\u03b4 supported Au (Au/Ce0.9Bi0.1O2-\u03b4) when comparing with Au/CeO2 (39%) under the same reaction conditions (2\u00a0h, 65\u00a0\u00b0C, 10\u00a0bar O2) [102]. It has been proved that introducing Bi3+ to substitute Ce4+ in the CeO2 lattice would lead to the improvement of the surface Ov concentration, thereby improving its oxygen activation capacity [102]. Moreover, the formed Bi\u2013O\u2013Ce linkages favored the hydrogen transfer process during the HMF oxidation reactions. A similar phenomenon was observed in Ce0.8Bi0.2O2-\u03b4 supported Pt NPs (Pt/Ce0.8Bi0.2O2-\u03b4) [103]. Interestingly, an FDCA productivity of 392 molFDCA h\u22121 molPt\n\u22121 was achieved over Pt/Ce0.8Bi0.2O2-\u03b4 at 23\u00a0\u00b0C (0.5\u00a0h, 10\u00a0bar O2), which is the highest value among ever reported noble-metal-based catalysts. According to the same strategy, Gao et\u00a0al. fabricated Mn\u2013Ce mixed oxides with tunable redox property for supporting Ru NPs (Ru/Mn6Ce1OY) [47]. As shown in the hydrogen temperature-programmed reduction (H2-TPR) and oxygen temperature-programmed desorption (O2-TPD) profiles of the Mn\u2013Ce oxides (Fig.\u00a04\n), both the reduction and oxygen desorption peaks shift toward a lower temperature with the increasing of Ce content, indicating the enhancement of its higher oxygen mobility. It has been well demonstrated that the strong metal-support interaction between Ru NPs and Mn\u2013Ce oxide afforded the catalyst an exceptional FDCA yield of > 99% under base-free conditions (15\u00a0h, 150\u00a0\u00b0C, 10\u00a0bar O2) and good catalytic stability against aggregation and oxidation of active Ru NPs [47].In addition to the heteroatom-doping method, regulating the morphology of the reducible oxides support is another way to modify the redox property of the support, which has been rarely studied for the catalysts in HMF oxidation. Li and co-workers compared the catalytic activity of Au/CeO2 with different CeO2 morphologies (rod, cube and octahedra) for the oxidation of HMF to FDCA, and found that the TOF value of Au/CeO2-rod (6.3 min\u22121) was 7\u201332 times higher than that of Au/CeO2-cube (0.9 min\u22121) and Au/CeO2-oct (0.2 min\u22121) [104]. The authors announced that CeO2-rod has the richest surface oxygen vacancies, which largely promoted the catalytic activity of Au/CeO2-rod by efficiently activating HMF and O2 [104]. Recently, Liao et\u00a0al. obtained more than twice higher FDCA productivity over MnO2 supported atomic Pd catalyst (Pd\u2013MnO2) than Pd NPs (PdNP\u2013MnO2) [105]. The experimental and DFT calculation results revealed that the incorporated atomic Pd species worked synergistically with MnO2 framework to create more Ov and to promote the mobility of the surface lattice oxygen (OL), which favored the activation of oxygen and enhanced the HMF adsorption capacity. This finding presented a new way to design robust catalyst with strong metal-support interaction by regulating the redox property of the support toward the efficient production of FDCA.In comparison with bulk support, the utilization of porous materials as the support for noble metal NPs holds unique advantages not only for precisely tailoring the NPs size and morphology but also facilitating the accessibility of active sites to substrate molecules. Therefore, controlling the porosity of the support is an effective strategy to improve the activity and stability of the supported NPs (Table 4).Zeolites, as extensively industrialized catalysts, have been widely studied as support in many biomass conversion processes because of their rigid porous structure and high specific surface area [109]. Cai and co-workers found that cage-type Y zeolite supported Au catalyst (Au/HY) offered the best catalytic performance for the oxidation of HMF, whereas low FDCA yields of 1%\u201315% were achieved over H-MOR and Na-ZSM-5 zeolites supported Au catalysts (Au/H-MOR and Au/Na-ZSM) [110]. The authors concluded that the activity of the catalyst is strongly related to the Au NPs size and smaller Au NPs provide better catalytic performance. Especially, the unique small super-cage of HY zeolite resulted in the formation of ultra-small and uniformly dispersed Au NPs inside the HY cage with a diameter of 1.0\u00a0nm. A special hydrothermal approach was also developed to direct incorporate Pt NPs into the crystals of Beta zeolite, resulting in the formation of highly active Pt@Beta zeolite catalyst for the oxidation of HMF to FDCA thanks to the enhanced interaction between Pt and silica species [111].Besides, mesoporous metal oxides were also developed as supports for improving the activity of the supported noble metal NPs. Lolli et\u00a0al. fabricated an ordered mesoporous CeO2 by using SBA-15 as a hard template for supporting Au NPs (Au/m-CeO2) for the oxidation of HMF to FDCA [112]. The small crystallites and high surface area of mesoporous CeO2 endowed the Au/m-CeO2 with better catalytic performance than commercial ceria-supported Au catalyst. Masoud and co-workers further emphasized the importance of the support pore structure for the stability of the Au NPs and the ordered mesoporous structure can minimize the growth of Au NPs during the HMF oxidation reactions [113]. Recently, Pichler et\u00a0al. adopted a surface-casting method to manufacture high-surface-area ZrO2 (ZrO2\nH-aero, 239\u00a0m2\u00a0g\u22121) [114]. The high surface area of the support endows the formation of ultra-fine Ru clusters (0.8\u20131\u00a0nm) in Ru/ZrO2\nH-aero catalyst, which displayed superior activity for the oxidation of HMF to FDCA under base-free aqueous solution with a high FDCA yield of 97% [114].Apart from the zeolites and traditional mesoporous metal oxides, many novel metal-free porous materials were also developed as the support for noble metal NPs for the efficient oxidation of HMF. In particular, the pore structure and specific surface areas of the porous organic polymers can be fine-tuned by simply applying different monomers. Various covalent triazine frameworks (CTF), a class of highly stable polymers, were prepared in molten ZnCl2 as the support for Ru NPs and the resulting Ru/CTF gave DFF yield of 73%, which is higher than that of Ru/C, Ru/\u03b3-Al2O3, Ru/hydrotalcite and Ru/MgO because of the high surface areas (2349 m2 g\u22121) and large pore volume (1.96\u00a0cm3\u00a0g\u22121) of CTF [115]. Furthermore, Ru/CTF also revealed a superior catalytic performance for the base-free oxidation of HMF to FDCA (78%) in water than Ru/C and Ru/\u03b3-Al2O3 [116]. In addition, mesoporous poly-melamine-formaldehyde (mPMF) [117], activated chitosan carbon (PVP-ACS-800) [118] and hierarchical porous nitrogen-doped carbon (NC2) [119] were also employed as porous carbon material to confine small noble metal NPs for the oxidation of HMF and offered desirable FDCA yields.As we discussed above, one of the important targets of increasing the porosity of the support is to achieve better metal dispersity and reduce the size of metal NPs. The unique pore dimension and channel structure of porous materials endow an ideal space for the incorporation of metal NPs with tailed sizes and morphologies. The size effect of the NPs has been reported in many cases and higher dispersion of the supported active phases generally affords a better reaction rate [120,121]. Meg\u00edas-Sayago et\u00a0al. investigated the size effect of the active carbon immobilized Au NPs in the range of 4\u201340\u00a0nm for the oxidation of HMF to FDCA [120]. The authors claimed that the increase of Au NPs sizes resulted in the decrease of the exposed Au (100)/Au (111) facet ratio and an exponential decay trend was observed between the product selectivity and Au (100)/Au (111) exposure ratio (Fig.\u00a05\n). Because the Au (100) face thermodynamically favored the oxygen reduction reaction and the smaller Au NPs exhibited higher Au (100) face concentration, which largely promoted the hydroxide ion concentration over the Au NPs surface, eventually boosted its catalytic activity for the oxidation of alcohol/aldehyde groups. This work provides new perspectives to gain insights into the size effect of the supported noble metal NPs, thereby may offer guidelines for developing efficient catalytic systems for the oxidation of HMF. However, rare reports focused on the interpretation of in-depth causes for the size effect of other supported noble metal catalysts (Pd, Pt, Ru and Ag).For the supported heterogeneous catalyst, the catalytic reactions generally occur over the catalyst surface, the surface nature of the support, such as the type and amount of the functional groups, hydrophilic and hydrophobic property as well as the coating property, thus plays a crucial role for the catalytic activity of the supported metal NPs. In this section, we will discuss the reported strategies for the modification of the support surface properties to enhance the catalytic performance of the loaded metal NPs for the oxidation of HMF, especially by introducing or altering the functional groups of the support and coating the support surface with another phase (Table 5).As discussed in section of 2.3.1, the introduction of basic functional groups (e.g. nitrogen-containing functional groups) on the carbon material support can increase the basicity of the support. In addition, nitrogen-containing functional groups are also frequently introduced onto the surface of the support as anchoring sites for soluble active metal species. For example, 3-aminopropyltriethoxysilane (APTES) functionalized silica-coated Fe3O4 NPs (Fe3O4@SiO2\u2013NH2), poly (4-vinylpyridine)-functionalized carbon-nanotube (PVP/CNT) and bi-imidazole groups grafted SBA-15 (SBA-Im) were reported as ligand containing supports to immobilize Ru3+ for the selective oxidation of HMF to DFF [123\u2013125]. The resultant supported Ru complex catalysts afforded high DFF yields of 86\u201394% in organic solvents [123\u2013125]. Moreover, cumbersome procedures for the modification of the pristine support with nitrogen-containing functional groups impede their large-scale application.In addition to nitrogen-containing functional groups, Wan and co-workers revealed that modifying the oxygen-containing functional groups over carbon nanotubes (CNTs) can deeply influence the catalytic activity of CNTs supported noble metal catalysts for the base-free oxidation of HMF into FDCA [50,126]. For instance, CNT pretreated by 30\u00a0wt% H2O2 generated a high concentration of phenol and carbonyl/quinone groups over its surface, offering a desirable FDCA yield of 94% under base-free conditions [126]. And the increasing amount of these oxygen-containing groups greatly enhanced the absorption ability of CNT toward HMF and reaction intermediates, consequently boosting the catalytic activity of the supported Au\u2013Pd NPs [126]. Interestingly, Chen et\u00a0al. found that the optimal inter-site distance of acidic sites (-HSO3 species) and oxidation sites (Ru species) in sulfonic acid groups decorated reduced graphene oxide-supported Ru NPs catalyst (Ru/S-rGO-2) was determined to be 12.5\u00a0\u00b1\u00a02.2\u00a0nm and \u201ccloser\u201d or \u201cfurther\u201d will promote the generation of undesirable by-products (such as humins and levulinic acid) in the one-pot conversion of fructose to DFF [86].The wettability of the catalyst support directly influences the adsorption capacity of the catalyst toward the substrates, thus wettability modulation of the support is an effective method to improve the performance of the supported metal NPs. However, a few works focus on studying the effect of the wettability of the catalyst on the oxidation of HMF. Wang and co-workers obtained an FDCA yield of 99% over hydrophilic mesoporous poly (ionic liquid) (MPIL) supported Au\u2013Pd alloy [127]. Whereas the hydrophobic material supported Au\u2013Pd alloy catalyst only provided a low FDCA yield of 36% under the same reaction conditions [127]. The authors revealed that hydrophilic support has a stronger affinity toward HMF than FDCA, which facilities the accessibility of HMF and leaving of the FDCA over the catalyst surface [127]. The establishment of facile and low-cost approaches to modify the wettability of the catalyst may be a promising method to enhance the activity of the catalyst for the oxidation of HMF.Iron-based magnetic materials are widely employed as catalyst support for both noble and non-noble metal NPs because of their low cost, facile separation and recovery features [128]. However, the self-interactions between magnetic nanoparticles make them easily aggregated, which deteriorates the catalytic performance of the catalysts. Surface encapsulation modification is a widely adopted strategy to avoid the aggregation of iron-based magnetic material and enhance their stability in basic and acidic environment [128]. Zhang and co-workers designed a series of hydroxyapatite (HAP), carbon or graphene oxide encapsulated \u03b3-Fe2O3 or Fe3O4 supported Ru, Pd, CoOx or Mn3O4 catalysts for the oxidation of HMF to DFF or FDCA. Among them, \u03b3-Fe2O3@HAP-Ru [129] and Fe3O4/Mn3O4 [130] provided high DFF yields of 82\u201389% under mild conditions. Pd/C@Fe3O4 [131], C\u2013Fe3O4\u2013Pd [132], \u03b3-Fe2O3@HAP-Pd (0) [133] and nano-Fe3O4-CoOx [134] gave satisfactory FDCA yield of 69%\u201393%. It was also reported that Fe3O4 in the Fe3O4 decorated reduced graphene oxide supported Pt NPs (Pt/Fe3O4/rGO) can promote the dispersion of the supported Pt NPs [135]. More importantly, Fe3O4 can interact with the supported Pt NPs to modulate the exterior electron state of Pt and Fe atoms, leading to a more active Pt electron state with high catalytic activity, thus afford a high FDCA yield of 99% even in a base-free condition. Moreover, the FDCA yield increased from 71% in Pt/CeO2 to 100% under base-free conditions after the introduction of a nitrogen-doped carbon coating over the CeO2 support (NC\u2013CeO2) [51]. Because the introduction of nitrogen-doped carbon coating leads to the generation of plentiful surface defects on the support and abundant electron-deficient metallic Pt species, as well as the increase of the basicity of the support [51].To circumvent the deactivation, instability and leaching issues of the supported single noble metal NPs, a second metal has been introduced as an additive to tailor the geometric as well as the electronic configurations of the active centers (Table 6).The combination of Au and Pd results in the formation of Au\u2013Pd bimetal NPs, which is the most investigated bimetal catalyst for HMF oxidation. For instance, Villa and co-workers obtained a moderate FDCA yield of 80% over Au/C under mild reaction conditions whereas the yield of FDCA decreased to 60% after five recycling tests because of the irreversible adsorption of intermediates or aggregation of Au NPs [136]. It is interesting to note that Pd or Pt-modified Au/C catalyst revealed enhanced activity and stability than pristine Au/C for HMF oxidation. Especially, Au8\u2013Pd2/AC catalyst provided a higher catalytic activity than Au8\u2013Pt2/AC, offering almost quantitative FDCA yield and an impressive FDCA productivity of 99 molFDCA h\u22121 mol\u22121\nnoble metal under mild reaction conditions. Importantly, an FDCA yield of 99% can be maintained over Au8\u2013Pd2/AC even after five runs, which proves that alloying Au and Pd overcomes the deactivation of Au/C for the oxidation of HMF. A similar promotion effect was also observed in Au\u2013Pd/CNT [126], zinc hydroxycarbonate (ZOC) supported Au\u2013Pd [137], Au6Pd1/TiO2 [138], Pd\u2013Au/HT [139], porous bowl-like nitrogen-doped carbon-supported AuPd bimetallic (AuPd/pBNxC) nanoreactors [140] and Au0.5Pd@Co3O4 catalyst [141]. In addition, Xia and co-workers observed that the formation of PdAu alloy NPs reallocates the electron density of the NPs as a result of the different electronegativity of Au (2.4) and Pd (2.2), leads to a reduction of NPs size and promotes the dispersion of Pd species [139]. The observed geometric effects and electronic effects via alloying Pd with Au greatly improve the catalyst activity of the PdAu alloy NPs for HMF oxidation, especially promotes the oxidation of HMFCA [139]. Liao and co-workers further confirmed that the introduction of Pd into Au NPs induces the electron transfer from Pd to the 5d orbitals of Au in a Au0.5Pd@Co3O4 catalyst, which promoted the generation of highly active oxidation sites (Au+ species), eventually leading to an enhanced catalytic performance of AuPd alloy NPs (Fig.\u00a06\n) [141].In addition to AuPd alloy NPs, the incorporation of budget non-precious metal into noble NPs not only can reduce the cost of the catalyst but also contributes to enhancing the catalytic activity of the NPs. TiO2 or CeO2-supported AuCu alloy NPs provided a better catalytic activity and stability than the monometallic sample, in which Cu species in Cu\u2013Au alloy NPs actually serves as a promoter or dispersing agent to elevate the dispersity of the active Au species [142\u2013144]. Gupta et\u00a0al. further compared the promotion effect of introducing a second non-noble metal (Co, Cu and Ni) for the base-free oxidation of HMF over Pd/Mg(OH)2 [145,146]. Among various prepared bimetal catalysts, the Ni0.9Pd0.1/Mg(OH)2 catalyst revealed a better catalytic performance than monometallic Pd as well as Co and Cu-derived bimetallic catalysts. The authors interpreted that the incorporation of Ni into Pd NPs endows the formation of electron-rich Pd sites because of the electron transport between the two components, which favors the absorption and activation of oxygen molecule, thus providing better catalytic performance toward HMF oxidation.Several cheap metals, including Pb, Bi, Ni and Sn, were also introduced as the synergetic component to enhance the catalytic activity of the supported Pt NPs for HMF oxidation. A high FDCA yield of 99% was achieved over Pb\u2013Pt/C at 25\u00a0\u00b0C in 1.25\u00a0mol L\u22121 NaOH aqueous solution [147]. The characterization of the recovered solid after the reaction revealed that the introduced homogenous Pb salts were deposited on the surface of the Pt/C during the reaction, resulting in the formation of bimetal Pb\u2013Pt/C catalyst, which can be directly reused without the addition of extra Pb salts in the next catalytic cycle [147]. Apart from Pb, Bi was also introduced as a promoter for enhancing the reactivity of Pt/C catalyst for the oxidation of HMF [148]. The promotion effect of Bi-doping was ascribed to the interaction between bismuth and the \u03c0 electrons of the furan ring. The oxophilicity of bismuth enhances the chemical adsorption of the geminal diol intermediate over the surface of Bi\u2013Pt/C catalyst, which facilitates the dehydrogenation process on active Pt sites. In addition, the presence of Bi promoter in Bi\u2013Pt/C catalyst can prevent the over-oxidation of the Pt NPs during the reaction, thereby improving the reusability of the catalyst. Recently, Shen et\u00a0al. found that combining Pt and Ni in Pt\u2013Ni/AC renders Pt NPs an improved CO absorption and oxidation ability, thereby offering the highest FDCA productivity of 25.3 molFDCA h\u22121 molnoble metal\n\u22121 under base-free conditions among noble metal catalysts [149]. Moreover, Sn-doping in 2-Pt1Sn1\u2013H2 catalyst can also modify the electronic structures of Pt NPs and results in the formation of abundant Pt (0) species with high catalytic activity and stability [150]. Obviously, the incorporation of secondary metal into noble metal NPs is an efficient method to improve their catalytic activity and stability. More efforts should be devoted to figuring out the genuine active sites of the complex bimetal catalysts as well as the intrinsic function of the dopants for enhancing the catalyst activity.Basic supports can readily react with the generated acidic products (FDCA, FFCA and HMFCA) during HMF oxidation process, resulting in the etch of active component and support, which leads to the downgrade of the catalyst activity and stability [65,71,151]. Several strategies were developed to prevent the leaching of basic support, including the constraint of the interactions between the acid products [67], the incorporation of an acid-resistant phase [45,68] and replacing the basic metal oxides or hydroxides with novel basic supports [69,70,73,90]. For other types of supported catalysts, the reasons for the catalyst deactivation mainly include the absorption of byproducts [44], the over-oxidation of the metal active sites [75,81] as well as the sintering of metal particles [81,96,98]. In addition, the leaching of N species was observed during the HMF oxidation process over Pt/C-EDA-4.1 catalyst and resulted in the catalyst deactivation [52]. Interestingly, Pd/HPGS catalyst exhibited high stability for HMF oxidation and can maintain its activity after 20 consecutive cycles thanks to the strong metal-support interactions [93]. Therefore, the modulation of metal-support interactions can be an effective strategy to enhance the stability of the supported noble metal catalysts. Accordingly, as presented in Table 3\n\n\n\n\n\n\n\n\n\n\n the reducible oxide-supported noble metal catalysts generally exhibited good catalytic stability and can be recycled several times (4\u20138) without activity deactivation. The existing robust metal-support interaction endowed the active noble metal species with high resistance to aggregation and oxidation during the oxidation process, thus revealing good recyclability [47,105]. Moreover, the encapsulation effect of the zeolite framework largely eliminates the leaching-induced deactivation of the catalyst, offering excellent catalyst stability [111]. The incorporation of secondary metal into noble metal NPs is another efficient method to improve the catalytic stability of the catalyst [136,138,142,148].As we summarized above, noble-metal-based catalysts exhibit excellent catalytic activity and recyclability toward HMF oxidation even in the absence of base promoter or under very mild reaction conditions. However, from an economic perspective, the implementation of these catalysts for the oxidation of HMF in a large-scale practical process certainly increases the production cost of the products because of the poor availability and exorbitant price of noble metals. Lately, the development of cost-effective non-precious metal-based catalytic systems for HMF oxidation is gaining momentum. A tremendous amount of non-precious metal oxides (e.g. V2O5, MnO2, CoOx and CeO2) were successfully applied for the efficient oxidation of HMF. It has been widely accepted that the oxidation of HMF over non-precious metal oxides operates by the Mars-van Krevelen (MvK) mechanism [152\u2013156]. Generally, the MvK mechanism includes several elementary steps, (1) alcohols were absorbed and activated on the surface of the oxide and gave rise to the generation of alkoxy species and protonated OL species via the O\u2013H bond dissociation; (2) the adsorbed alkoxide intermediates were subsequently converted into aldehydes and protonated OL species after deprotonation from the \u03b2-carbon by the vicinal OL, meanwhile, electron transfer from the substrates to the metal cations, leading to the reduction of the metal oxides; (3) the recombination of these formed protonated OL species leads to the formation of H2O and O2 as well as Ov; (4) the catalytic circle was finished after the reduced metal oxides were re-oxidized and oxygen vacancies were replenished by the dissociative chemisorption of aerial oxygen. Accordingly, this reaction mechanism comprises a metal oxide redox cycle, thus the reducibility and oxidizability of the catalyst play a key role in its catalytic activity for HMF oxidation [157,158]. The acid-base property of the catalyst was also reported could influence the dissociation of O\u2013H bond [159\u2013161].Kinetic parameters and reaction pathways are fundamental aspects to the interpretation of the catalytic reaction mechanisms. The selective oxidation of HMF to DFF involves the cleavage of \u03b2C\u2212H and O\u2013H bond, both of which was reported to be the rate-determining step during the HMF oxidation process over different catalysts. In a manganese oxide (OMS-2), KIE studies revealed that dissociating \u03b2C\u2212H bond of the adsorbed alkoxide intermediates is a kinetically-relevant step in the catalytic cycle [157]. A kinetic ratio of (kH/kD) 4.19 was observed for the competing oxidation of HMF and its deuterated counterpart at the methylene group (R-CD2OH) [157]. This study pointed out that the dissociative chemisorption of O2 over the catalyst surface is also a kinetically-relevant step during HMF oxidation. For the oxidation of HMF to DFF over a vanadium oxide nanobelt-arrayed mesoporous microsphere (VOx-ms) [163], HMF molecules with a deuterated hydroxyl group (R\u2013CH2OD) offered a unified kH/kD ration of 1.05 whereas R-CD2OH gave a kH/kD value of 2.21, which also suggested that \u03b2C\u2013H bond cleavage is the rate-determining step during HMF oxidation. However, L. Suib and co-workers revealed that the energy for the break of O\u2013H bond was higher than that of C\u2013H bond in a mesoporous manganese incorporated cobalt oxide (meso Mn-CoOx) by the DFT calculations, implying that O\u2013H bond dissociation is the rate-determining step during the HMF oxidation [156]. The DFT calculation results are in discordance with the previous KIE studies, thus the exact mechanism for the oxidation of HMF to DFF over non-precious metal oxides needs to be further studied.It is similar to the case of noble metal catalysts, HMF can be oxidized into FDCA over non-precious metal oxides proceeds via two reaction paths (Scheme 1. Route 1 and 2). The composition and structure of the catalyst, as well as the reaction medium, significantly affected the reaction route for HMF oxidation. Hara and co-workers investigated the rate-constants (k\n\n1\n-k\n\n5\n) for the oxidation of HMF to FDCA over activated MnO2 and found that the oxidation of HMF to DFF in the first step gave a k\n\n1\n value of 2.4\u00a0\u00d7\u00a010\u22123, which is almost ten times higher than that of conversion of HMF to HMFCA, indicative of that HMF was converted into FFCA mainly through DFF (route 2) [155]. Moreover, the k\n\n5\n value is obviously lower than that of k\n\n1\n-k\n\n4\n, suggesting that the oxidation of FFCA to FDCA is the rate-determining step during this catalytic circle. The same reaction path was also reported in many Mn-based catalytic systems [152,164,165]. However, the oxidation of HMF to FDCA over the (Fe, Co, Ni)-doped MnOx and holey 2 D Mn2O3 nanoflakes catalysts proceeded via route 2 (Fig.\u00a07\n) [162,166]. The discrepancy reaction pathways over Mn-based catalysts may be ascribed to the different adsorption behaviors of hydroxymethyl and aldehyde moiety of HMF over different catalysts, which often receives cursory attention during the reaction mechanism investigations. Furthermore, Zhang and co-workers calculated the activation energy (Ea) of each step for the conversion of HMF to FDCA over Fe0.6Zr0.4O2 catalyst in [Bmim]Cl [167]. The results revealed that the oxidation of FFCA to FDCA possessed the highest E\na value of 110.2\u00a0kJ\u00a0mol\u22121, which is higher than that oxidation of HMF to HMFCA (82.7\u00a0kJ\u00a0mol\u22121) and HMFCA to FFCA (86.4\u00a0kJ\u00a0mol\u22121). Similarly, the same research group also reported that the oxidation of FFCA to FDCA is the rate-determining step for the oxidation of HMF to FDCA in [Bmim]Cl over heteropoly acid catalyst [168]. It should be noted that the oxidation of FFCA to FDCA was observed to be the slowest step during the HMF oxidation process over almost all the non-precious metal oxides catalytic systems, but a reasonable interpretation has not been proposed yet.In section 2.3.2, various kinds of non-precious reducible oxides were employed as the support to enhance the catalytic activity of noble metal NPs for HMF oxidation. Actually, non-precious transition metal oxide, such as V2O5 and MnO2 itself contains redox couple (e.g. V5+/V4+ and Mn4+/Mn3+/Mn2+) and OL, which are capable of oxidizing alcohols alone (Table 7). Several inorganic vanadium-containing materials, such as V2O5 [169] and VOPO4\u00b72H2O [170,171], show high selectivity toward the oxidation of HMF to DFF in DMSO. To avoid the use of the solvent with a high boiling point (such as DMSO), Yan et\u00a0al. fabricated a (010)-faceted vanadium oxide nanobelt-arrayed mesoporous microsphere (VOx-ms) for the selective oxidation of HMF to DFF, in which an excellent DFF yield of 89% was obtained within 1\u00a0h in aqueous solution [163]. DFT calculations revealed the OL of VO sites prefers to absorb the O\u2013H group of HMF rather than \u2013CHO moiety, thus guaranteeing high DFF selectivity [163].In comparison with vanadium-based catalysts, manganese oxides not only can selective oxidation of HMF to DFF but also enables the efficient formation of FDCA in basic aqueous solution. MnO2 is a typical manganese oxide with diverse crystal structures, but pure MnO2 displays offered DFF yields as low as 2%\u201342% in organic solvents [156,160,172\u2013174]. Nitrogen-doped MnO2 (N\u2013MnO2) afforded quantitative DFF yield was achieved at 25\u00a0\u00b0C within 6\u00a0h [175]. The authors demonstrated that nitrogen doping slightly elongates the Mn\u2013O bonds and descends the Mn\u2013O coordination numbers of MnO2, which creates more surface defect sites and coordinatively unsaturated Mn sites, eventually enhance the catalytic performance of the catalyst. Several polycrystalline MnO2 catalysts were also developed for the oxidation of HMF to DFF [174,176,177]. In particular, manganese oxides containing MnCO3, \u025b-MnO2 and Mn2O3 provided DFF yield of 88% in ethanol, in which MnCO3 and \u025b-MnO2 were believed to work concertedly to enhance the catalytic performance [176]. It should be noted that the most adopted solvent (DMF or DMSO) in vanadium or manganese-based catalytic systems is ranked as hazardous or problematic solvent according to the CHEM21 selection guide for organic solvent while ethanol belongs to recommended solvent [178]. The development of catalysts that enable the effective oxidation of HMF to DFF in eco-friendly and low boiling point solvent is an important direction for the production of DFF in the future.Even though pure MnO2 has low catalytic activity toward the oxidation of HMF to DFF in organic solvents, Hara's group achieved an FDCA yield of 91% over commercially available activated MnO2, which is much higher than that obtained over other manganese oxides (Mn2O3, Mn3O4, MnO and MnOOH) [153]. Later, the effect of MnO2 crystal structure (\u03b1-, \u03b2-, \u03b3-, \u03b4-, \u03b5-, and \u03bb-MnO2) on the HMF oxidation has been investigated [155,179]. DFT calculations demonstrated that the Ov formation energies at the planar oxygen sites in MnO2 crystal structure are generally higher than those at the bent oxygen sites (Fig.\u00a08\n) [155]. Especially, the ratio of oxygen sites with relatively lower vacancy formation energies in \u03b2- and \u03bb-MnO2 is higher than that of in \u03b1- and \u03b3-MnO2, indicating the first two types are likely to be better candidates than later ones for the HMF oxidation [155]. Indeed, the succeeding experimental analysis shows that the activity of the manganese oxides for the oxidation of FFCA to FDCA decreases in the order of \u03b2-MnO2\u00a0>\u00a0\u03bb-MnO2\u00a0>\u00a0\u03b3-MnO2\u00a0\u2248\u00a0\u03b1-MnO2\u00a0>\u00a0\u03b4-MnO2\u00a0>\u00a0\u03b5-MnO2. What\u2019 more, increasing the surface area of \u03b2-MnO2 can further enhance its catalytic performance. Similarly, Bao and co-workers constructed holey 2 D Mn2O3 nanoflakes by the calcination of an Mn-based metal\u2013organic framework (MOF) precursor, which provided almost quantitative FDCA yield thanks to its abundant surface pores structure [162]. Although satisfactory FDCA or DFF yield can be obtained over pure manganese oxides, high catalyst loading as well as a long reaction time are required, meanwhile the productivity of\u00a0the product, especially for FDCA (0.08\u20130.3 mmolFDCA h\u22121 g\u22121\ncatalyst), is quilt low over these single V and Mn oxides.In addition to vanadium and manganese-based catalysts, CuO and Co3O4 revealed outstanding catalytic performance toward the oxidation of HMF to FDCA by using NaClO as oxidant, in which FDCA yields of 96%\u2013100% were obtained under mild reaction conditions (40\u00a0\u00b0C, 2\u00a0h) [180]. However, the involvement of large amounts of oxidant (NaClO/HMF molar ratio\u00a0=\u00a045) for HMF oxidation made this process cost-intensive. Very recently, Lin's group developed an innovative NiOx catalyst for the efficient oxidation of HMF to FDCA [181], where a high FDCA yield of 97% was achieved by using only 6 equiv. NaClO as oxidant at 25\u00a0\u00b0C in a short reaction time of 30\u00a0min with a remarkable FDCA productivity of 24.3 mmolFDCA h\u22121 gcatalyst\n\u22121.In recent years, cheap binary and ternary metal oxides have gained boosting interest for HMF oxidation because of their higher catalytic activity and better recyclability than single cheap metal oxides (Table 8). The binary and ternary metal oxides are composed of two or three kinds of cheap metal species, in most cases, only one of them is the active site for HMF oxidation and other constituents serve as the promoter to enhance the activity of the active component. Moreover, the introduced second metal species can also interact with the original active species to form new active sites.Both vanadium oxides and phosphates display good catalytic activity toward the selective oxidation of HMF to DFF. However, DFF yields lower than 90% were obtained over single vanadium oxides or phosphates in extended reaction time (> 10\u00a0h in most cases). It was reported that the introduction of Cu or Fe can enhance the mobility of OL of the catalyst [182,183]. For example, Hou and co-workers demonstrated that binary \u03b1-CuV2O6 nanobelt offered almost quantitative DFF yield within 3\u00a0h in DMSO [182]. The introduction of Cu results in the formation of unique Cu\u2013O\u2013V units in \u03b1-CuV2O6 with excellent OL mobility due to the bimetallic synergistic effects.Various cheap metal species such as K, Fe, Co, Ce, and Cu were introduced into manganese oxides as the second or third component for modifying their redox or acid\u2013base properties. The hydrothermal treatment of KMnO4 and MnSO4\u00b7H2O in acidic aqueous solution results in the generation of cryptomelane octahedral molecular sieves (K-OMS-2) with a composition of KMn8O16\u00b7nH2O, which revealed excellent catalytic activity for the selective oxidization of benzylic alcohols to aldehydes or ketones [184]. Inspired by this work, Fu's group successfully achieved a DFF yield of 99% from HMF over K-OMS-2 in DMSO whereas MnO2 only afforded a DFF yield as low as 4% [185]. Interestingly, the subsequent treatment of K-OMS-2 in 1\u00a0mol L\u22121 HNO3 solution produces an H\u2013K-OMS-2 catalyst, which offered better catalytic activity than K-OMS-2 for HMF oxidation because of the presence of Br\u00f8nsted acid sites over H\u2013K-OMS-2 catalyst [185]. Subsequently, Nie and Liu further compared the catalytic performance of manganese oxides with different morphologies for HMF oxidation (OMS-1, OMS-2, OMS-6, OMS-7, \u03b3-MnO2, amorphous MnO2 (AMO), birnessite-type MnO2 (Na-OL-1)), among which OMS-2 showed the best catalytic activity and offered a DFF yield of 97% in DMF within 1\u00a0h [157]. Note that the activities of the manganese oxides correlate well with their reducibility and oxidizability. The particular (2\u00a0\u00d7\u00a02) tunnel structure of OMS-2 endows its high reducibility and oxidizability, thus provides high catalytic activity for HMF oxidation. OMS-2 only contains the oxidation sites for HMF oxidation, Sarmah et\u00a0al. thus combined the H-Beta catalyst, which contains both of Br\u00f6nsted and Lewis acidic sites, and OMS-2 for the transformation of carbohydrates into DFF, where excellent DFF yields of 97%, 95%, 93% and 91% were achieved from fructose, sucrose, glucose, and starch in a one-pot two-step process respectively [186].The superior catalytic performance of K-OMS-2 catalyst originates from its particular (2\u00a0\u00d7\u00a02) tunnel structure formed in the presence of Mn and K in an acidic environment whereas the combination of manganese oxides with Fe, Co, Ce, La [187] and Cu [188] species boosts the activity of the catalyst mainly through modulating its redox or acid\u2013base properties. Neatu and co-workers reported that the co-existence of multiple phases (bixbyite-Mn2O3, MnO2 and hematite) in Mn0.75/Fe0.25 catalyst works synergistically to enhance its catalytic activity through strengthening its medium basic sites (O2\u2212) [161]. Recently, Lin's group further investigated the promotion effect of Fe2O3 for MnO2 catalyst for the oxidation of HMF, in which HMF conversion of 97% with DFF selectivity of 98% was achieved within 5\u00a0h [160]. It has been well demonstrated that the Fe2O3-doping greatly increased the content and activity of Mn4+-O2\u2212 acid-base pair in Mn6Fe1Ox, which was proved to be active sites for HMF oxidation.In addition to Fe-doping manganese oxides, Co has also been extensively introduced as a promoter to enhance the catalytic performance of the catalyst. Gui et\u00a0al. observed that the synergy between Mn and Co species induces the formation of the dominant CoMnO3 phase in Mn0.50-Co0.50-O and spinel CoMn2O4 hollow spheres catalyst, which promotes the mobility of OL and also enriches its content, thus increasing its catalytic activity [172,189]. Interestingly, Xu and co-workers obtained an FDCA yield of 95% (Co\u2013Mn-0.25, 120\u00a0\u00b0C, 1\u00a0MPa O2, 5\u00a0h) over Co\u2013Mn oxides prepared by the solid-state grinding method, which outperformed the pure MnOx [165]. The excellent catalytic performance of Co\u2013Mn-0.25 catalyst is ascribed to its high OL mobility and variable oxidation states of surface Mn species after the introduction of Co species. Recently, Lin's group further investigated the influence of the Ov concentration of the Mn\u2013Co oxides on its catalytic performance for the oxidation of HMF [190]. Experimental and theoretical calculation results well proved that increasing the Ov amount not only can boost the OL reactivity of Mn\u2013Co oxides by weakening the Mn\u2013O bond intensity but also promote the adsorption and activation of HMF and O2 over the catalyst [190].On the other hand, the incorporation of Mn into cobalt oxide can also improve the mobility of OL of the catalyst by Jahn-Teller (J-T) distortion (Fig.\u00a09\n) [156,191]. For example, L. Suib and co-workers designed a mesoporous manganese incorporated cobalt oxide (meso Mn-CoOx) for the oxidation of HMF, where 80% HMF conversion with DFF selectivity of 96% was obtained over 5% Mn-doped CoOx (meso 5% Mn-CoOx) [156]. Especially, the TOF value achieved over meso 5% Mn-CoOx catalyst is 300\u2013391 folds higher than that of meso-CoOx or -MnOx samples. The substitution of Co3+ in meso-CoOx with Mn3+ species elongates the Mn\u2013O bonds because of the J-T distortion, which improves the mobility of OL. It seems like that the key to enhance the catalytic performance of the metal oxides is to improve the mobility of OL; however, the deep reasons for this have not been clarified yet.Taking the excellent oxygen storage of CeO2 into consideration, Han and co-workers introduced CeO2 as a promoter to increase the OL transmission from Ce to Mn species to boost the activity of the catalyst for HMF oxidation [152]. H2-TPR and XPS results revealed that CeO2 in MnOx-CeO2 mixed oxide (MC-6) improves its oxygen mobility and surface Mn4+ and Ce3+ concentration. Mn4+ was assumed to be the primary active site for HMF oxidation over MnOx-CeO2 mixed oxide while the replenishment of OL was accomplished via transforming the OL of Ce species to MnO2 lattice. This mechanism indicates both Mn4+ and Ce3+ species are involved in the HMF oxidation process. Yu et\u00a0al. further demonstrated that the intrinsic active sites in (Fe, Co, Ni)-doped MnOx catalysts are M3+O(-Mn4+)2 clusters, which can provide more active OL and stronger OL regeneration ability than Mn4+O(-Mn4+)2 clusters in pure MnOx catalyst [166]. These works suggested that the introduced dopant metal species can also interact with the original active species to form new active sites.Most of the above-listed multitudinous Mn-based oxides emphasized the important role of excellent OL mobility for their superior catalytic performance toward HMF oxidation. Actually, the modulation of the acid\u2013base properties of the catalysts can also greatly improve their alcohol oxidation activity. For instance, Parvulescu et\u00a0al. put forward that the introduction of Cu species into Mn\u2013Al LDH greatly enhances its basicity, which weakens the intensity of the hydroxyl group of HMF, thus facilitating the oxidation of HMF into DFF under base-free aqueous solution [192]. Dibenedetto and co-workers also obtained 89% FFCA yield with impressive productivity of 8.3 molFFCA h\u22121 gcatalyst\n\u22121 over CuO\u00b7CeO2 mixed oxides in base-free aqueous solution [193]. Especially, the authors revealed that the balance between acidity or basicity of the catalyst plays a key role to achieve high FFCA selectivity. The same research group further revealed a volcano relationship between the ratio of strong basic and acid sites (nb/na) in Mg\u2013Ce binary mixed oxides and their DFF selectivity [159]. The highest DFF yield of 96% was obtained over MgO\u00b7CeO2 with nb/na ratio of 1.1 in base-free aqueous solution. A higher or lower nb/na ratio than 1.1 leads to the formation of over-oxidized products FFCA and undesirable ring-opening by-products. To further increase the oxidation activity of CuO\u00b7CeO2 for the production of FDCA from HMF, MnO2 was included in the catalyst as a strong oxidant [164]. Different from the supported noble metal catalysts, in which the basic support generally provides better catalytic performance; the balance of the basic and acid sites of the non-noble metal oxides is important for the catalytic performance of the catalyst.Mo-containing Keggin heteropolyacids (HPMo12O40, HPAs) have already been experimentally and theoretically demonstrated as efficient catalysts for the production of DFF and FDCA from HMF or carbohydrates [154,168,194]. Especially, the partial replacement of Mo centers in HPAs by V (MVP-HPAs) can further enhance its oxidation ability through increasing its redox capacity [168,195]. However, HPAs function homogeneously in the majority of solvents, thus suffers from separation and recovery issues in the catalysis process. Note that the substitution of H+ in HPAs by Cs+ can generate insoluble cesium salt of HPAs (such as, CsxH3-xPMo12 and CsMVP-HPA), which can be employed as heterogeneous catalyst for the oxidation HMF and comparable products yield were achieved as the homogeneous HPAs [195,196]. Besides, Xu and co-workers found that CeCu(OH)6Mo6O18 heterogeneous trimetal catalysts shown excellent catalytic activity for the oxidation of HMF into DFF in p-chlorotoluene [197].Another active Mo-derived catalyst for the oxidation of HMF is molybdenum oxide (MoOx), but the lack of sufficient acidity for fructose dehydration limits its application for the one-pot conversion of fructose to HMF. Therefore, Yang and co-workers developed a series of 3D flower-like micro/nano Ce\u2013Mo composite oxides with tunable redox and acidic properties, in which f-Ce9Mo1O\u03b4 catalyst offered an DFF yield of 74% from fructose [198]. However, the fructose dehydration reaction was performed under N2 atmosphere while the HMF oxidation occurred under oxygen atmosphere, thus switching the inert gas to oxygen is inevitable to guarantee high DFF yield in this case. Interestingly, Zhao et\u00a0al. accomplished the one-pot production of DFF from fructose by the integration of sulfonate zirconia and molybdenum oxide (MZS), where a satisfactory DFF yield of 74% was obtained from fructose over 10-MZS catalyst under oxygen flow [199]. An appropriate MoO3 content (10\u00a0wt%) in the catalyst is important to achieve a balance between its acidity and oxidability, thus promotes the fructose dehydration and HMF oxidation process while suppressing the direct fructose oxidation by-reactions under oxygen flow. Recently, Lei and co-workers revealed that the combination of zirconia and molybdenum oxides (ZrxMoyO\u03b4) also enables the conversion of fructose into DFF in a one-pot process with a DFF yield of 61% under static air within 4\u00a0h [200]. Moreover, These Mo-based catalysts generally provided good recyclability after removing the absorbed organic impurities by a calcination post-treatment even if the substrate is carbohydrates.Iron-based materials were extensively employed as magnetic support to facilitate the recovery of the catalyst or promoter to enhance the catalytic activity of Mn-oxides, which were previously regarded inactive for the oxidation of HMF. Interestingly, Li and co-workers designed a series Fe-based catalysts by controlling the structure and crystal facets of Fe-oxides, which revealed excellent catalytic activity for the production of DFF from HMF or fructose [201\u2013203]. In particular, the (111) crystal facet of octahedral Fe3O4 NPs was proved to be highly active for the selective oxidation of HMF to DFF because of the presence of negatively charged oxygen species in (111) crystal facet [202,203]. In addition to regulate the morphology of Fe oxides, the judicious selection of the reaction medium can also transform the inert Fe-based materials into active catalysts for HMF oxidation. Zhang and co-workers proposed a series of ionic liquid (IL)-promoted Fe-based catalytic systems for the base-free conversion of HMF or fructose to FDCA, among which FDCA yields of 38\u201361% was achieved from HMF or fructose over Ce0.5Fe0.15Zr0.35O2 [204] or Fe0.6Zr0.4O2 [167,205] in 1-butyl-3-methylimidazolium chloride ([Bmim]Cl) without base additive. In contrast, only negligible FDCA was detected in other common solvents (e.g. H2O, methanol and 1,4-dioxane) over the same catalyst. The solvent effect seems to dominate the catalytic activity of the catalysts, but the original catalytic activity of the catalysts in IL remains unclear. Moreover, an effective strategy for the separation of FDCA and reuse of IL is urgently needed to improve the competitiveness of the IL/metal oxides catalytic system.Many soluble non-noble metal salts and organic complexes are reported to exhibit excellent catalytic activity toward the oxidation of HMF and even outperform the heterogeneous ones in some cases [208]. However, some flaws of homogeneous catalysts, such as difficulties in the separation of catalyst from the final product as well as the recycling of the catalyst, severely hinder their practical application and commercialization. Heterogenization of the soluble active components on stable support is a workable and widely employed solution for the above problems. Especially, encapsulation and tethering through a covalent bond or electrostatic interaction are the major adopted strategies to immobilize the homogeneous non-noble catalyst for HMF oxidation (Table 9).Encapsulation of the catalyst complex within the small pores of the support not only can avoid the catalyst leaching but also can maximally imitate the homogeneously-catalyzed reaction process. In 2003, Ribeiro and co-workers designed a bifunctional acidic and redox catalyst by encapsulating cobalt acetylacetonate into sol\u2013gel silica (Co-gel), which afforded an overall FDCA yield of 71% from fructose via a one-pot process within only 1\u00a0h without base additive [209]. Lee and co-workers employed Cr-MIL-101-encapsulated phosphomolybdic acid (PMA-MIL-101) as a heterogeneous bifunctional catalyst for the production of DFF from HMF or fructose [210]. 2PMA-MIL-101 catalyst even offered a slightly higher DFF yield of 91% from HMF than that of homogeneous phosphomolybdic acid (88%). Due to the Mo species leaching (9%), the spent 2PMA-MIL-101 catalyst lost 10% DFF yield from fructose in the second run as compared to the fresh one. To avoid the leaching of metal active species, Wang et\u00a0al. fabricated a mesoporous silica nanofibers-encapsulated H5PMo10V2O40 catalyst (HPMoV/SiO2) through surfactant-directed pore formation and electrospinning methods, which can be reused at least 10 runs for the oxidation of HMF to DFF without catalyst deactivation and metal leaching [211]. Recently, Wang and co-workers also developed a stable polyoxometalate-based mesoporous poly (ionic liquid) catalyst (PMoV2@CP-5.5-400) through a partial carbonization process (Fig.\u00a010\n) [212].Comparing with the encapsulation approach, anchoring the soluble components in the support surface via a covalent bond or electrostatic interaction is a preferred methodology because of its better stability and general applicability. Many metal acetylacetonate complexes are known as active homogeneous catalysts for the oxidation of alcohols [213]. In recent years, various organic and inorganic supports, including organically functionalized SBA-15 mesoporous silica (SBA-Py-VO-2) [214], montmorillonite K-10 clay [215], hydroxyapatite encapsulated magnetic \u03b3-Fe2O3 [216], polyaniline [217], were employed to heterogenise the soluble acetylacetonate complexes for HMF oxidation, and high product yields (around 90%, DFF or HMFCA) could be obtained.Actually, N atoms in many N-containing groups and complexes have lone pair electrons, which exhibit a high affinity toward various metal cations by coordinating with their free orbit, thus the immobilization of metal cation on the N-containing materials produces stable heterogeneous catalyst [218\u2013223]. For example, Saha and co-woks prepared a stable porphyrin-based Fe(III)-porous organic polymer (FeIII\u2013POP-1), which provided excellent productivity of 20.8 molFDCA h\u22121 gcatalyst\n\u22121 by using air as oxygen source under base-free condition [219]. N-containing basic materials can also tightly bond homogeneous heteropolyacid (HPA) through the acid-base interaction. Therefore, polyaniline [224], amino-functionalized CeO2 nanofibers [225] and chitosan nanofibers [226] were employed to immobilize polyoxometalates for the production of DFF from HMF or carbohydrates with good catalyst stability. In particular, chitosan nanofibers supported POM (HPMoV/CS-f (25)) provided excellent DFF yield of 94%, 62% and 31% from HMF, fructose and glucose respectively as a result of its unique coexistence of redox capacity and acid\u2013base properties [226].Overall, immobilized catalytic systems not only addressed the separation and reusability issues of homogeneous catalysts but also present a promising opportunity to enhance the catalytic activity of the homogeneous counterparts through elaborately modulating the surroundings around the active sites. Nevertheless, the heterogenization methods usually involve a massive amount of organic solvents as well as lengthy and multi-step operations, more efforts should be made for the development of facile and low-cost approaches to immobilize the homogeneous catalyst.It's similar to the case of the supported noble metal catalyst, the catalytic activity of the supported non-noble metal oxides heavily depends on the properties of the support, such as the base-acid property, redox property as well as porosity of the support. Various strategies were developed to manipulate these properties of the support to boost the catalytic activity of the catalyst in many ways, like improving the dispersity of the active components as well as the support and metal oxides interactions (Table 10).It has been reported that increasing the Br\u00f8nsted acidity of the support can promote the catalytic activity of the catalyst for the oxidation of HMF to DFF [80,158,227]. Riisager and co-workers found that V2O5/H-beta catalyst contained the lowest Lewis acidity but highest Br\u00f8nsted acidity (derived from V2O5) comparing with H-ZSM-5, H\u2013Y, and H-mordenite supported V2O5 catalysts, thus offering the highest V2O5 dispersity and highest DFF yield [227]. This result is similar to Odriozola's conclusion, in which increasing Br\u00f8nsted acidity of the support was believed to promote the formation of the alkoxy intermediate during the HMF oxidation process [80]. The introduction of Br\u00f8nsted acidity in the support also enables the formation of bi-functional catalysts for the one-pot production of DFF from fructose. DFF yields of 63\u201378% were obtained from fructose over protonated graphitic carbon nitride supported vanadium catalyst (V\u2013g-C3N4(H+)) [228], protonated nitrogen-doped carbon-supported molybdenum trioxide (Mo-HNC) [229] and carbon sphere-supported molybdenum oxide (MoOx/CS) [230]. These works pointed out that the balance of the acid density and redox ability of the catalyst is important to achieve a high DFF yield in the one-pot process [228\u2013230].Basic supports are also widely employed to enhance the catalytic activity of the catalyst for the oxidation of HMF. N-doped carbon materials are extensively adopted as support for loading non-noble metal species because the strong interaction between nitrogen and metal species can produce stable metal NPs, sub-nano clusters or even single-atom-catalysts [231,232]. Nitrogen-doped graphene confined Cu NPs (Cu/NG) was reported to be active for the oxidation of HMF to DFF with the aid of 2,2,6,6-tetramethyl-piperidin-1-oxyl (TEMPO) [233] or FDCA by using TBHP as an oxidant [234]. Zhang et\u00a0al. further compared the catalytic activity of various non-noble metal nitrides (MNx/C-T, M\u00a0=\u00a0Fe, Co, Cu, Cr, and Ni; T represents the calcination temperature) for the oxidation of HMF, among which FeNx/C-900 offered the best catalytic performance with 97% DFF yield [235]. This research pointed out that Fe\u2013N4 species (an iron moiety coordinated with four nitrogen groups) of the catalyst is the main active site for HMF oxidation and its concentration as well as chemical circumstances are crucial for the catalytic activity of the catalyst. Subsequently, several N-doped carbon supported Co catalysts were developed for the oxidation of HMF to FDCA, such as CoNx/C-900 [236], CoOx-MC [237], Co\u2013Mn/N@C [238], nitrogen-doped carbon-supported single-atom cobalt catalyst (Co SAs/N@C) [239]. The presence of N species in the catalyst not only enhanced the basicity of the catalyst but also promoted the dispersity of the active components and even created new kinds of active sites by coordinating metal species.The employment of redox and porous materials as support generally boosts the catalytic activity of the non-noble metal oxides by means of promoting the dispersity of the active species and/or metal oxides\u2013support interactions. The spontaneous redox reaction between VO3\n\u2212 anions and the polyaniline-functionalized carbon nanotubes (CNTs) (VO2-PANI/CNT) afforded the formation of homogeneously dispersed VO2 NPs, which provided a high DFF yield of 96% from HMF [240]. The enhancement of the oxygen mobility of the support also greatly promoted the catalytic activity of the supported metal oxides for the oxidation of HMF [241,242]. For example, Fang et\u00a0al. recently revealed that the catalytic activity of Mn\u2013Co\u2013O supported Co3O4 NPs can be improved by enhancing the oxygen mobility of Mn\u2013Co\u2013O support [242].By using the strong interactions between the metal oxides and zeolite, Wang and co-workers designed copper-containing mordenite zeolite (V2O5@Cu-MOR) [244] and high-silica MOR supported vanadium oxide catalyst (10V2O5@MOR (60)) [245] for the production of DFF from HMF or fructose, in which uniform and stable isolated V species were generated inside the pore channel of the zeolite. To further improve the dispersity of the active species, Fang and co-workers developed a universal method to fabricate mesoporous KIT-6 encapsulated ultrafine high-loading metal-oxides NPs (Co3O4, CuO, Fe3O4 and NiO) through calcinating a self-assembled MOF precursor inside the silica mesopores [243]. The authors concluded that the confinement effects of the KIT-6 mesopores enabled the generation of uniformly dispersed ultrafine metal-oxides NPs within the mesopores (Fig.\u00a011\n), which thus provided high FDCA yields (80%\u201399%) and excellent productivities (40\u201350 molFDCA h\u22121 molmetal\n\u22121) under very mild base-free conditions.For the single non-noble metal catalysts, the absorption of organic impurities [160,164,172,189,193], the leaching of metal species during the oxidation process [228\u2013230], the variation of the surface area as well as the redox cycle of the metal cation [153,162] are the common causes for the deactivation of the catalyst during the recycling experiments. The strong interaction between the metal species and N component can protect the active sites of the catalyst from leaching issue and guarantee a good reusability of the catalyst [175,233,234,238]. Especially, N\u2013MnO2 catalyst can be reused at least 8 runs without decreasing its catalytic activity [175]. The established stable chemical bonds between the acetylacetonate complexes and support [215,217,218] as well as the robust interaction between polyoxoanion and N-containing groups [224\u2013226] also enables the catalyst a good stability. In comparison with single metal-based catalysts, the introduction of a second or third component can also increase its stability, which may be beneficial from the optimized redox or acid\u2013base properties [159,165,182,201,207]. The slight deactivation of some binary and ternary metal oxides could be contributed to the reversible transformation of the catalyst crystal structure [159]or the reduced redox properties [166].Taking into consideration the fact that metal-based catalysts generally suffer from the leakage issue of metal ions, high catalyst cost and environmental pollution problem, the development of metal-free catalysts can be an intriguing and promising approach to address these above drawbacks. In recent years, low-cost carbon-based materials bearing various heteroatom dopants or rich surface functional groups were developed as efficient catalysts for the oxidation of HMF (Table 11).In 2015, Watanabe and co-workers, for the first time, realized the conversion of HMF to DFF over a nitrogen-doped activated carbon catalyst (N-doped-AC-8), achieving a higher DFF selectivity (93%) than that of noble-metal catalysts (Ru/C and Pt/C, 0\u201314%) albeit at a relatively low HMF conversion (23%) [246]. Graphite-type nitrogen (Nc) species over the N-doped carbon surface are in charge of the selective oxidation of alcohols due to the good correlations between the conversion of alcohols and the amount of Nc species. Differentiating from the case of metal oxides, the oxidation of alcohol over N-doped carbon catalyst operates by a Langmuir\u2013Hinshelwood process. As shown in Fig.\u00a012\n, oxygen was adsorbed and activated on a carbon site adjacent to the Nc and/or on the nitrogen atom to produce oxygen radicals, then alcohol, which might adsorb on Nc site, was oxidized to aldehyde by the formed active oxygen species. In addition, Nc species were considered as the active sites for the conversion of HMF to FDCA in a zeolitic-imidazole framework (ZIF-8)-derived nitrogen-doped nanoporous carbon (NNC) catalyst [247], nitrogen-doped graphene (NG-800) [248] and bamboo sawdust-derived nitrogen-doped carbon (NC) material [249]. Especially, the obtained FDCA yields over NNC catalysts prepared at various calcination temperature corresponds well with the Nc species contents of the catalysts, which further confirmed the key role of Nc species for HMF oxidation [247]. Recently, Tao et\u00a0al. also verified that the graphitic N and pyridinic N species in N-doped graphene (NG) should be responsible for the activation of oxygen to generate active superoxide radicals (\u00b7O2\n\u2212) by combining DFT calculation and EPR experiment [250]. However, these above catalysts gradually deactivated with the increase of the recycling runs because of the decrease of the amount of Nc species during the oxidation process.To enhance the activity and stability of N-doped carbon catalysts, TEMPO [248] and HNO3 [251] were introduced as co-catalysts. Significantly increased DFF yields of 95\u2013100% were observed with the aid of co-catalyst whereas only a trace amount of DFF was detected without co-catalyst [248,251]. The presence of TEMPO or HNO3 may generate reactive oxygen species over the catalyst, thus improving the DFF yield. The recyclability of these catalysts has also been greatly increased [248,251]. Interestingly, Cao and co-workers found that the introduction of P into N-doped carbon can significantly enhance the selectivity of DFF because of the high concentration of P\u2013C and Nc species of the catalyst [252]. This work avoids the involvement of co-catalyst or second oxidant during the HMF oxidation process and inspires us to develop efficient carbon-based catalysts for HMF oxidation by a heteroatoms co-doping strategy.In addition to the heteroatom-doping strategy, the introduction of functional groups over the carbon surface can also create active sites for HMF oxidation [253]. Lv et\u00a0al. revealed the carboxylic acid groups in the graphene oxide (GO) can oxidize HMF to DFF, affording a DFF yield of 90% in a reaction time of 24\u00a0h [254]. The little loss of GO reactivity during the recycling tests for the oxidation of HMF is largely related to the partial reduction of oxygen-containing functional groups [254]. Glucose-derived carbocatalyst (CC\u2013SO3H\u2013NH2) comprising of both acidic groups (-SO3H and \u2013COOH) and basic groups (-NH2) was employed for the direct production of DFF from carbohydrates [255]. Accordingly, the basic sites of the CC-SO3H\u2013NH2 catalyst are responsible for the isomerization process while its acid sites work as active sites for the dehydration and selective oxidation process. Anyway, green and sustainable metal-free catalytic systems will play a more important role in the oxidation of HMF and other biomass-derived compounds if the catalyst activity and stability can be further improved.In recent years, great progress has been accomplished for the production of FDCA and high FDCA yields can be obtained from HMF over noble or non-noble catalytic systems. Nevertheless, FDCA is a solid powder with a high boiling point (around 420\u00a0\u00b0C at normal atmosphere) and poor solubility in common solvents, which made it difficult to purify FDCA by conventional crystallization and rectification method [256]. The poor solubility of FDCA in water and major industrial solvents seriously encumbers its production at a high concentration. These limitations of FDCA have shifted research interest toward its methyl ester derivative, that is, furan-2,5-dimethylcarboxylate (FDMC). Benefiting from the good solubility of FDMC in methanol, the oxidative esterification of HMF to FDMC in high substrate concentration (>\u00a020\u00a0wt%) with good yields (around 90%) has already implemented [257,258]. In addition, FDMC can be readily separated and purified from the reaction medium by sublimation as a result of its relatively low boiling point (around 140\u00a0\u00b0C at 10 Torr) [259]. More importantly, FDMC enables the production of PEF with high quality while a colored product of PEF is obtained in the case of FDCA due to the decomposition of FDCA during the polymerization process [260]. However, comparing with FDCA, a few studies focused on the production of FDMC from HMF. Only several heterogeneous catalysts, such as Au, PdCoBi and Co, were found to be active for this reaction (Table 12). It should be noted that a systematic review of this research area has not been published to date.In 2008, Taarning and co-workers, for the first time, accomplished the one-pot oxidative esterification of HMF to FDMC over Au/TiO2 catalyst with excellent yield (98%) and productivity (102.2 molFDMC h\u22121 molAu\n\u22121) in the presence of sodium methoxide [259]. Semicrystalline nanoporous multiblock copolymer matrix-supported Au NPs (Au/sPSB) also provides almost quantitative FDMC yield from HMF with the aid of a base promotor [38]. However, the involvement of base additives during HMF oxidation process makes these catalytic systems less attractive. Interestingly, CeO2 and ZrO2 supported Au catalysts can circumvent the use of alkali additives due to their appropriate acid\u2013base properties, which enables the base-free conversion of HMF to FDMC with almost quantitative yield under mild reaction conditions [261,262]. Kinetic curves for the oxidation\u2013esterification of HMF showed that the oxidation of alcohol to aldehyde over Au/CeO2 is the rate-limiting step and aldehyde can be rapidly converted into ester via hemiacetal intermediates [261]. However, Au/CeO2 catalyst was significantly deactivated in the second reuse because of the absorption of organic material and a calcination process (250\u00a0\u00b0C, air, 12\u00a0h) is required to regenerate the catalytic activity of the catalyst [261].To improve the activity and stability of Au NPs toward the oxidation esterification of HMF, Cu or Pd was introduced as the second component [263,264]. The \u03b3-Al2O3 supported Au-CuOx nanohybrids (Au\u2013Cu/\u03b3-Al2O3) afforded much higher FDMC yield (98%) than that of \u03b3-Al2O3 supported single Au catalyst (48%) [264]. The authors well demonstrated that the introduction of Cu induced a strong electron interaction between Au and Cu species, thus leading to the generation of abundant intensely interacted Au\u2013CuOx interfaces as active sites for HMF oxidation. And the negatively charged Au species in Au-CuOx hybrids also improved the oxygen activation capacity of the catalyst [264]. Nevertheless, the interaction between Au and Cu species became weaker after the reaction, resulting in the deactivation of the catalyst during the recycling experiments. Similar to the case in the oxidation of HMF to FDCA [139], the introduction of Pd greatly promoted the activity of AuPd\u2013Fe3O4 for the oxidation of alcohol group of HMF, rending FDMC yield of 92% at room temperature [263]. By contrast, a single Au\u2013Fe3O4 failed to oxidize the hydroxyethyl group of HMF to aldehyde, thus affording the formation of 5-hydroxymethylfuroic acid methyl ester (HMFE) with a yield of 92%. In addition, AuPd\u2013Fe3O4 sample revealed better catalyst stability than Au\u2013Fe3O4 with significantly decreased metal leaching in the reaction solution thanks to the synergistic electron transfer between Au and Pd [263].Besides Au-based catalysts, Stahl and co-workers surprisingly found that the combination of Pd, Bi and Te with specific formulations rendered highly effective catalysts for oxidation-esterification of primary alcohols [265]. Recently, Fu and co-workers further revealed that FDMC yield of 93% was achieved from HMF when employing Pd/C as the catalyst and Co(NO3)2 and Bi(NO3)3 as promoters with the aid of K2CO3 [256]. Interestingly, the prepared heterogeneous catalyst (PdCoBi/C) afforded a slightly higher FDMC yield of 96% than the case with homogeneous promoters. However, it remains ambiguous for the nature of the promotion effects of additives (Bi and Co) in this above case. The noble metal catalysts, especially Au-based catalysts, revealed high catalytic activity for the oxidation esterification of HMF even under mild and base-free conditions, but the stability of the catalyst remains barely satisfactory.Recently, N-doped carbon-supported Co catalysts were developed as unique non-noble metal catalysts for the efficient oxidative esterification of alcohols in methanol [266,267]. Accordingly, Fu and co-workers attempted to employ CoxOy-N@C catalysts for the oxidative esterification of HMF in the presence of K2CO3. They found that CoxOy-N@C catalyst showed high catalytic activity for the conversion of aldehyde group to ester but weak catalytic activity for the oxidation of alcohol group to aldehyde [268]. Therefore, K-OMS-2 [268], \u03b1-MnO2 [269] and Ru@C [270,271] were introduced as the co-catalyst to promote the oxidation of hydroxymethyl moiety of HMF and HMFM, and thus high FDMC yields of 95\u2013100% could be achieved over these multicomponent catalytic systems. During the oxidative esterification of HMF over N-doped carbon-supported Co catalyst, the pyridinic N species of the catalyst worked as Lewis base to abstract the acidic hydrogen of hemiacetal intermediate, forming pyridinic N+-H species. And the presence of Co3O4 species could facilitate the regeneration of pyridinic N by forming [OH\u2212]ad to react with N+-H species. Lin's group recently further verified that substrate, oxygen and methanol were adsorbed and activated on nitrogen-doped carbon shells (C\u2013N) of Co@C\u2013N catalyst while Co species behaved as electron donator to enhance electron density of C\u2013N material, improving its catalytic activity for oxidative esterification of HMF (Fig.\u00a013\n) [272].In comparison with N-doped carbon-supported cobalt oxides, N-doped carbon-supported Co NPs, such as hollow yolk\u2013shell Co@CN [258], porous cobalt/nitrogen co-doped carbons (Co@NCs) [273], hollow Co NPs embedded nitrogen-doped graphite (Co@CN) [274], successfully realized the base-free oxidative esterification of HMF, where FDMC yields of 89\u201395% could be achieved with a reaction time of 12\u201324\u00a0h. The enhanced catalytic performance of these catalysts may be related to the unique hollow structure, high specific surface area or optimized basic and acid sites of the catalyst [258,274]. In addition, Liu and co-workers surprisingly found that the catalytic activity of N-doped carbon-supported Co NPs (Co NPs-N@C) can be further boosted by reducing the size of Co NPs to single-atom Co species [275]. The isolated Co\u2013N3C species of Co SAs\u2013N@C catalyst served as the active sites during the alcohol oxidation process, which can significantly reduce the energies for O2 and alcohol activation as well as hemiacetal intermediate formation than in the case of Co NPs-N@C catalyst. It has been well documented that the oxygen was activated to \u00b7\n-O2 species by gaining electrons from the N-doped carbon-supported Co catalyst during the HMF oxidative esterification process [270,275]. Therefore, the enrichment of electron density of CoNx species may contribute to enhance the catalytic activity of Co catalysts. Indeed, metal NPs (Co or Cu NPs) were reported to be able to work synergistically with single CoNx sites by inducing the electron transfer from metal NPs to CoNx sites, which reduced the energy for O2 adsorption and favored the generation and release of active oxygen species, and thus promoting the oxidative esterification of HMF to FDMC [257,273,276]. Even though Co catalysts displayed satisfactory catalytic activity for the oxidative esterification of HMF, the Co catalysts generally gradually lose its catalytic activity during the recycling experiments because of the absorption of organic impurities, the aggregation and oxidation of Co NPs [257,274].It is beyond doubt that Co-based catalysts represent an important advance towards the efficient production of FDMC from HMF by using non-noble catalysts, but the Co-based catalysts were commonly prepared with expensive precursors (such as, 2-methylimidazole and dicyandiamide) and energy-consuming preparation process (\u2265 800\u00a0\u00b0C with an inert atmosphere), which certainty hinders their industrial application. Therefore, the development of novel and facile strategy for the preparation of low-cost, efficient and stable Co-based catalysts are very necessary. In addition, the reason for the exclusive catalytic performance of Co-based catalysts toward the oxidative esterification of alcohols remains ambiguous. Thus, a clearer understanding on the reaction mechanism of the oxidation esterification of HMF based on elaborate experiments and computational simulations is of crucial importance.In this review, we systematically summarized the proposed universal catalyst design strategy toward the efficient selective oxidation of HMF. Overall, great progress has been achieved for the selective oxidation of HMF, many novel heterogeneous catalytic systems were developed and exhibited excellent catalytic activity even under mild reaction conditions. Nonetheless, to implement the large-scale production of downstream value-added products by selective oxidation of HMF, there is definitely a long way to go regarding how to address the following issues:\n\n(1)\nOne of the challenges for commercializing the HMF oxidation processes is to design highly effective and stable but low-cost catalyst. The exorbitant prices as well as unsatisfactory stability of noble-metal catalysts severely hinder their practical application. Non-noble and metal-free catalysts seem to be economical alternatives for noble catalysts while their relatively low catalytic activity generally requires high catalyst loading to guarantee a high product yield and selectivity. On the other hand, non-noble metal oxides are less stable than noble metal NPs in an acidic environment. The fabrication of non-noble mixed-metal oxides (binary, ternary and even high-entropy metal oxides) and encapsulation of metal oxides with acid-resistant phase has emerged as workable approaches for designing highly durable nonprecious-metal catalysts. The catalytic activity of metal-free heteroatom (N, P, and S) doped carbon-based catalysts largely relies on the concentration of dopants as well as their coordination environment. Multiple heteroatoms co-doping strategy may herald the advent of a new avenue to fabricate highly effective and robust carbon-based catalysts for HMF oxidation. But, how to precisely regulate the number of dopants as well as their coordination environment should be considered.\n\n\n(2)\nMost previous works for the catalytic oxidation of HMF mainly focused on process optimization and scarce studies provide insightful information regarding the reaction mechanisms of HMF oxidation reactions. Convincing and evidential mechanism investigation is strongly encouraged to proceed through the combination of DFT calculations and advanced operando characterization techniques, which may be helpful to elucidate the relationship between the catalyst structure and activity.\n\n\n(3)\nEven though many recently developed novel catalytic systems afforded satisfactory catalytic performance for HMF oxidation, the catalyst preparation process often involved lengthy and multi-step operations as well as a massive amount of organic solvent, which is not suitable for large-scale industrial application. Therefore, in addition to excellent catalytic performance, the preparation method of the catalyst should be as simple as possible and inexpensive when designing a new catalyst for HMF oxidation.\n\n\n(4)\nBased on this review, the majority of the established catalytic systems only focused on the selective oxidation of pure HMF, whereas HMF currently has a similar price\u00a0to its oxidation products. Thus, it's of significance to develop multi-functional catalysts bearing both acid\u00a0sites and oxidation sites for the direct transformation of budget carbohydrates or even lignocelluloses into desirable products in a one-pot process. Especially, there\u00a0are only several catalytic systems that have been reported for the production of FDCA from carbohydrates in a one-pot process with low catalytic efficiency [85,168]. To accomplish this goal, the developed catalysts should have good tolerance to the by-products such as humins.\n\n\n(5)\nAnother limitation for the production of FDCA from HMF is the use of low substrate concentration solutions. The development of novel solvent systems with better FDCA solubility and providing better protection for the sensitive HMF molecule may be a solution for these issues. Moreover, the replacement of HMF with more stable HMF derivates can also realize the production of FDCA at high substrate concentration (10\u201320\u00a0wt%). Thus, the development of new catalysts, which can operate well at high substrate concentration solution and enable the efficient oxidation of HMF derivates, is an important research topic in the future.\n\n\n(6)\nThe industrial-scale production of bulk chemicals generally operates under flow continuous conditions, which can offer better productivity and lower production cost. Thus, to close the gap between the laboratory investigation and industrial application of HMF oxidation processes, more efforts should be devoted to evaluating the catalytic performance (such as reaction pathways, optimal reaction conditions and kinetic parameters) of the catalyst under continuous flow conditions by employing fixed bed reactors.\n\n\nOne of the challenges for commercializing the HMF oxidation processes is to design highly effective and stable but low-cost catalyst. The exorbitant prices as well as unsatisfactory stability of noble-metal catalysts severely hinder their practical application. Non-noble and metal-free catalysts seem to be economical alternatives for noble catalysts while their relatively low catalytic activity generally requires high catalyst loading to guarantee a high product yield and selectivity. On the other hand, non-noble metal oxides are less stable than noble metal NPs in an acidic environment. The fabrication of non-noble mixed-metal oxides (binary, ternary and even high-entropy metal oxides) and encapsulation of metal oxides with acid-resistant phase has emerged as workable approaches for designing highly durable nonprecious-metal catalysts. The catalytic activity of metal-free heteroatom (N, P, and S) doped carbon-based catalysts largely relies on the concentration of dopants as well as their coordination environment. Multiple heteroatoms co-doping strategy may herald the advent of a new avenue to fabricate highly effective and robust carbon-based catalysts for HMF oxidation. But, how to precisely regulate the number of dopants as well as their coordination environment should be considered.Most previous works for the catalytic oxidation of HMF mainly focused on process optimization and scarce studies provide insightful information regarding the reaction mechanisms of HMF oxidation reactions. Convincing and evidential mechanism investigation is strongly encouraged to proceed through the combination of DFT calculations and advanced operando characterization techniques, which may be helpful to elucidate the relationship between the catalyst structure and activity.Even though many recently developed novel catalytic systems afforded satisfactory catalytic performance for HMF oxidation, the catalyst preparation process often involved lengthy and multi-step operations as well as a massive amount of organic solvent, which is not suitable for large-scale industrial application. Therefore, in addition to excellent catalytic performance, the preparation method of the catalyst should be as simple as possible and inexpensive when designing a new catalyst for HMF oxidation.Based on this review, the majority of the established catalytic systems only focused on the selective oxidation of pure HMF, whereas HMF currently has a similar price\u00a0to its oxidation products. Thus, it's of significance to develop multi-functional catalysts bearing both acid\u00a0sites and oxidation sites for the direct transformation of budget carbohydrates or even lignocelluloses into desirable products in a one-pot process. Especially, there\u00a0are only several catalytic systems that have been reported for the production of FDCA from carbohydrates in a one-pot process with low catalytic efficiency [85,168]. To accomplish this goal, the developed catalysts should have good tolerance to the by-products such as humins.Another limitation for the production of FDCA from HMF is the use of low substrate concentration solutions. The development of novel solvent systems with better FDCA solubility and providing better protection for the sensitive HMF molecule may be a solution for these issues. Moreover, the replacement of HMF with more stable HMF derivates can also realize the production of FDCA at high substrate concentration (10\u201320\u00a0wt%). Thus, the development of new catalysts, which can operate well at high substrate concentration solution and enable the efficient oxidation of HMF derivates, is an important research topic in the future.The industrial-scale production of bulk chemicals generally operates under flow continuous conditions, which can offer better productivity and lower production cost. Thus, to close the gap between the laboratory investigation and industrial application of HMF oxidation processes, more efforts should be devoted to evaluating the catalytic performance (such as reaction pathways, optimal reaction conditions and kinetic parameters) of the catalyst under continuous flow conditions by employing fixed bed reactors.There are no conflicts to declare.We are grateful for funding supported by the National Natural Science Foundation of China (Grant Nos. 22078275; 21978246), the National Key Research and Development Program of China (Grant No. 2019YFB1503903), the Key Area Research and Development Program of Guangdong Province (Grant No. 2020B0101070001), the Fundamental Research Funds for the Central Universities (Grant No. 20720190014), PetroChina Innovation Foundation (2019D-5007-0413).", "descript": "\n                  The selective oxidation of 5-hydroxymethylfurfural (HMF), a versatile bio-based platform molecule, leads to the formation of several intriguing and useful downstream chemicals, such as 2,5-diformylfuran (DFF), 5-hydroxymethyl-2-furancarboxylic acid (HMFCA), formyl 2-furancarboxylic acid (FFCA), 2,5-furandicarboxylic acid (FDCA) and furan-2,5-dimethylcarboxylate (FDMC). These products have been extensively employed to fabricate novel polymers, pharmaceuticals, sustainable dyes and many other value-added fine chemicals. The heart of the developed HMF oxidation processes is always the catalyst. In this regard, this review comprehensively summarized the established heterogeneous catalyst design strategy for the selective oxidation of HMF via thermo-catalysis. Particular attention has been focused on the reaction mechanism of HMF oxidation over different catalysts as well as enhancing the catalytic performance of the catalyst through manipulating the properties of the support and fabricating of multi-component metal nano-particles and oxides. The current challenges and possible research directions for the catalytic oxidation of HMF in the future are also discussed.\n               "}
{"full_text": "Carbon cycles are different processes which transfer carbon in many ways from energy production to atmospheric pollution. Burning of fossil fuels in daily life for the industrial operations and/or energy supply to human society has liberated most common greenhouse gases (GHGs) such as carbon dioxide ( \n\n\n ) and methane ( \n\n\n )\u00a0[1,2]. These are major contributors to the greenhouse effect, by trapping radiation or heat in the atmosphere. Increase in GHGs emissions has triggered climate change and global warming\u00a0[3]. These emissions have had vital impact on our ecosystem, thereby initiating unaccountable anomalous weather conditions and natural calamities\u00a0[3]. Therefore, the necessity of mitigation of these greenhouse gases has gained interest and search for the methods to produce clean energy. Reforming reactions are interesting methodologies that not only mitigate, but also convert these dreadful GHGs into value added products\u00a0[4]. Among many reforming reactions, dry reforming of methane (DRM) has its own importance by converting two greenhouse gases into useful syngas ( \n\n\n ). The syngas ratio from DRM can be used as direct feed stock for F\u2013T (Fischer\u2013Tropsch) synthesis\u00a0[5]. Although DRM has many advantages, it is a great challenge to design an active and suitable catalyst for it. Moreover, it is extremely difficult to activate highly stable \n\n\n bond in \n\n\n and \n\n\n bond in \n\n\n at low temperatures. Further, at low temperatures this reaction has been accompanied by side reactions, such as (i) reverse water gas shift (RWGS: \n\n\n ), (ii) methane decomposition (MD: \n\n\n ) and (iii) Boudouard reaction (BR: \n\n\n )\u00a0[4,6]. So for the low-temperature DRM the catalyst should suppress the side reactions, be sustainable in moist environment as well as resistant to coking from individual reactions of methane and \n\n\n . Several advantages of low-temperature dry methane reforming are demonstrated by a number of researchers\u00a0[7\u20139]. Recently, Yabe et\u00a0al. reported about the electricity generation and hydrogen storage from renewable resources by using catalytic low-temperature dry reforming of methane\u00a0[10,11]. Noble metal and non-noble metal catalysts were extensively studied for DRM reaction\u00a0[12\u201315]. Although noble metal catalysts showed high activity, industrialization through these catalysts is unimaginable for economic reasons. So in this scenario non-noble metal catalysts were considered as alternative source for DRM. Among them, \n\n\n based catalysts showed better performance and promising activity compared to noble metal catalysts. However, \n\n\n catalysts are highly prone to coking. \n\n\n nanoparticles of around 5nm and below are known to effectively reduce coking\u00a0[16\u201318]. In order to enhance the resistance towards coking many supports and their modifications to improve oxygen storage capacity (OSC) have been taken into consideration. Cerium dioxide and ceria modified materials as supports have been receiving great attention and are extensively studied in catalysis from the past decades\u00a0[4]. The generation of oxygen vacancies and OSC in \n\n\n can be achieved by enhancing the \n\n\n redox cycle with simple alterations\u00a0[19]. Further, it is more advantageous to obtain that with shape controlled synthesis of ceria. The ceria nanocrystals with definite morphologies expose unique surface properties, for instance, ceria rods with prevailing {110}, {100} facets and cubes with {100} facets\u00a0[20]. Previous literature reports of both theoretical and experimental studies on ceria nanocrystals revealed that the surface energy of these dissimilar crystal facets varies with definite morphologies\u00a0[21,22]. The theoretical calculations carried out on nano ceria crystals revealed that the energy of formation of oxygen vacancy was surface sensitive and the stability of \n\n\n was closely related to the crystal facets. It was concluded that the rods with {110} and {100} facets have shown a more facile migration of lattice oxygen atoms from the bulk to the surface than cubes with {100} facets\u00a0[20,22,23]. Furthermore, combination of a suitable support and metal yields improved catalytic activity. The strong synergy between \n\n\n and \n\n\n might generate a hybrid nanostructure that enhances the interfacial metal-support interactions\u00a0[23]. These interactions were vital in improving the thermal stability as well as dispersion of active metal on support.A considerable number of publications show every year how significant advantages can be achieved by the application of microprocess technology in terms of product yield, purity and time required for chemical and biochemical conversions compared to equivalent bulk reactions\u00a0[24]. Microreactor technology demonstrated the advantages of microfluidic devices for a very efficient performance of chemical and biochemical processes under controlled and repeatable conditions. Recently, new concepts such as continuous processing, flow chemistry, high-throughput screening and process intensification have been established to open novel pathways in process design and engineering. Process intensification with miniaturization provides insights into the different scales on which process intensification can be applied and enables the development of novel and sustainable plants that offer dramatic process improvements over the existing state of the art in terms of plant size, waste production and other factors\u00a0[25]. As a result of the small size, the surface to volume ratio is much higher than in conventional reactors. This in turn affects other properties such as the flow regime and mass and heat transfer. Since high pressures and temperatures can be handled much easier on a very small scale, microreactors open up new process windows\u00a0[26]. There is no doubt that process intensification through microreactor applications clearly has the potential to revolutionize chemical and biochemical synthesis. Between-two-plates reactors have recently gained some attention in the field, as it has been shown that they enable seamless scale-up possibilities, good process control and present a simple to build, as well as simple to assemble solution\u00a0[27\u201330].Compared to homogeneous catalysis, heterogeneous catalysis is one of the key tools to increase the sustainability of the diversity of chemical syntheses by simplifying product processing and allowing easy separation and reuse of the catalyst. The incorporation of active, selective and stable solid catalysts in the form of immobilized coatings or micrometer-sized powders in microreactors offers additional advantages by providing highly specific surfaces for the catalysis of the normally demanding three-phase reactions (i.e.\u00a0gas\u2013liquid\u2013solid or liquid\u2013liquid\u2013solid reactions)\u00a0[31].Lattice Boltzmann (LB) methods have emerged in late 1980\u2019s as an alternative approach to solving lattice-gas automata (LGA)\u00a0[32]. They originally proved to be a computationally cheaper option of LGA solving of hydrodynamic problems, especially in low Reynolds number regimes. By the beginning of the new millennium the field has expanded to various applications such as: turbulence modeling, flow through complex geometries, heat and mass transfer, even reactive flows and others\u00a0[33]. Some studies have focused on LB modeling gas flow in specifically in microchannels\u00a0[34\u201337]. When considering gaseous reactive flows, combustion is an often studied topic\u00a0[38,39], although non-combustive model reactions have also been studied\u00a0[40]. Additionally, heterogeneous chemical reactions have been introduced in LB\u00a0[41,42], as well as heterogeneous catalytic reactions\u00a0[43\u201347]. The LB is showing great promise for use in modeling of packed- and fixed-bed reactors, among others, because of its easy implementation of complex geometries, therefore it is unsurprising that this has been a somewhat popular topic for its application\u00a0[30,41,43\u201345]. In spite of all the development in the LB field, a suitable yet simple boundary condition for a surface-reaction which would incorporate some sort of reaction kinetics is lacking.In this study, \n\n\n deposited on ceria rods ( \n\n\n ) was studied in detail in the process of low-temperature dry reforming of methane conducted in two different reactors systems. The catalyst was well characterized by different instrumental techniques. The study showed promising catalytic activity of both conventional fixed-bed and micro-channel reactors. Furthermore, a modified bounce-back approach to modeling heterogeneously catalyzed reactions within the scope of LB is presented. The results of the computations are compared with experiments in two geometries: a conventional fixed-bed reactor and a between-two-plates microchannel fixed-bed reactor.Synthesis of \n\n\n nanorods was performed according to the following protocol: 53.8g of \n\n\n (Sigma Aldrich) was dissolved in 140mL of ultrapure \n\n\n (Elga Purelab, model Option Q) yielding solution A. A separate solution containing 84mL of ultrapure \n\n\n and 4.9g of \n\n\n (Sigma Aldrich) was prepared (solution B). Solutions A and B were mixed and stirred for 30min on a magnetic stirrer (IKA, model C-MAG HS7). Afterwards, they were transferred to a Teflon\u00ae clad stainless steel autoclave, where they were hydrothermally aged for 24h at 100\u00b0C. After this time elapsed, the autoclave was quench cooled to room temperature and the suspension centrifuged. This was followed by 3-times washing and centrifuging with water and final washing and centrifuging with absolute alcohol. The yellow powder was freeze dried (Christ, model Alpha 1-2 LDplus) and calcined for 4h at 450\u00b0C in air using a muffle furnace (Nabertherm, model P330) and a heating ramp of 5\u00b0C min-1.Nickel was deposited by dissolving \n\n\n in 100mL of ultrapure water, followed by addition of appropriate mass of \n\n\n nanorods. The suspension was mixed on a magnetic stirrer and a dropwise addition of diluted ammonia aqueous solution was used to raise the pH value of the \n\n\n aqueous suspension, facilitating deposition of \n\n\n . The pH value of the suspension was raised slowly over the course of 2\u00a0h. After the pH value reached 9.5\u201310 the resulting suspension was centrifuged, dried overnight at 60\u00b0C in a laboratory drier and calcined for 4h at 450\u00b0C using a heating ramp of 5\u00b0C min-1. A nominal 1wt. \u00a0% nickel content was deposited. The nickel concentration in the solution after its deposition was analyzed by means of a photometric analysis (Spectroquant by Merck). For the 1wt. \u00a0% nickel sample, deposition efficiency of 99% was measured, meaning that the nominal loading corresponds to the actual nickel content.\n\n\n\n physisorption analysis (Micromeritics, model TriStar II 3020) was carried out at \u2212196\u00b0C on degassed catalyst samples to determine BET surface area, total pore volume and average pore diameter.\nIn-situ XRD measurements consisting of a series of reduction and oxidation steps of the catalyst were collected on the PANalytical Empyrean diffractometer having \n\n\n K\n\u03b1\n \u00a0radiation (\n\n\u03bb\n=\n1.54\n\n\u00c5\n\n). At first, the room temperature measurements were recorded in a step wise increment of 0.045\n\u2218\n with a count time of 0.5s and in \n\n2\n\u03b8\n\n range of 10 to 80\n\u2218\n. Soon after recording, the catalyst reduction was conducted in 5% \n\n\n balanced \n\n\n gas atmosphere at 300\u00b0C for 1h with a ramping of 5\u00b0C min-1; after completion of reduction, the measurements were recorded with similar procedure as mentioned above. Further, pure \n\n\n gas was used for purging in order to remove the \n\n\n atmosphere and then the compressed air was introduced into the chamber for 30min followed by cooling to 25\u00b0C . XRD diffractogram of the oxidized catalyst was again recorded and then the sample was purged with \n\n\n in order to create inert atmosphere. The second reduction was started from 25 to 500\u00b0C in \n\n\n flow with a ramping of 5\u00b0C min-1 for 1h and the XRD patterns were recorded at 500\u00b0C . \n\n\n gas was used to purge the chamber and then the second oxidation was carried out by using compressed air for 30min and then cooled down to 25\u00b0C with air atmosphere to record XRD diffractogram of the final oxidized sample.Micromeritics\u2019 AutoChem II 2920 apparatus equipped with a TCD detector was used to conduct the temperature programmed reduction experiments. 50mg of a catalyst was loaded in a U-shaped reactor and subjected to pretreatment in 5% \n\n\n at 300\u00b0C for 30min. Then followed purging with \n\n\n gas for 30min to remove physisorbed \n\n\n from the catalyst surface. After 15min of purging, the catalyst was cooled down to 10\u00b0C and reduction started with 5% \n\n\n gas mixture to 300\u00b0C with a heating ramp of 10\u00b0C min-1. After 1h, the gas flow was again shifted to 5% \n\n\n and kept for 30min. The reactor was then cooled down to 10\u00b0C in the same atmosphere. Then the re-reduction of the same sample was carried out from 10 to 500\u00b0C .\n\n\n\n pulse chemisorption measurements were also conducted on Micromeritics\u2019 AutoChem II 2920 instrument. The required amount of a catalyst was loaded in the U-shaped tube and reduced in 5% \n\n\n gas flow at 300\u00b0C for 1h and then the reactor was purged and cooled with \n\n\n . The consequent \n\n\n pulses were injected into the stream of \n\n\n until the saturation at 10\u00b0C . After that the reactor temperature was raised to 500\u00b0C for 1h. Then the same procedure was followed after performing reduction of the catalyst sample at 500\u00b0C .The DRM activity tests were performed at atmospheric pressure both in a conventional laboratory-scale fixed-bed reactor (differential reactor, I.D.: 9mm) made of quartz and a microchannel reactor made of stainless steel SS-316 plates separated by a graphite gasket (SGL Carbon, SIGRAFLEX\u00ae). From here onward the two systems will be referred to as \u201cconventional\u201d and \u201ctwo-plate\u201d, respectively. The two-plate channel dimensions were \n\n100\n\nmm\n\u00d7\n10\n\nmm\n\u00d7\n0.5\n\nmm\n\n. Schematic drawings of the two reactor systems are provided in a later section in Fig.\u00a02. In the conventional reactor, 50mg of a pelletized, crushed and sieved catalyst sample (average particle diameter: 0.375-0.500mm) was employed. For the two-plate reactor, 100mg of the catalyst sample was mixed with 1.5g of \n\n\n material (irregularly shaped particles, average particle diameter: 0.5mm) and carefully arranged in a single layer on the bottom plate over which the top plate was placed. The catalyst was reduced at 500\u00b0C for 1h in \n\n\n flow, then the reaction gas mixture composed of pure and undiluted \n\n\n and \n\n\n streams (\n\n\n\nCH\n\n\n4\n\n\n\u2215\n\n\nCO\n\n\n2\n\n\n=\n1\n\n:1) was fed into the reactor with three different total feed flow rates of 20, 40 and 60mLmin\u22121. The DRM reaction was carried out in the temperature range of 400\u2013500\u00b0C . The outlet gas stream was analyzed online by using a micro gas chromatograph (Agilent, model 490) equipped with Porapak Q and molecular sieves (MS5A) columns. The diameter of the catalytic bed in the conventional fixed-bed reactor was 9mm, which is sufficient in order to avoid wall effects (the utilized reactor diameter to the average particle diameter ratio is above 20).The model used consisted of two domains: a mesoscopic domain, which modeled the reaction and flow patterns in the reactor bed, using the lattice Boltzmann method; and a macroscopic domain which modeled the mass transport in the space leading up to the fixed-bed. The two domains are sketched in Fig.\u00a01.\n\n\nFor the purpose of numerical simulation of the reactive flow inside the domain 2 of the two reactors, the isothermal lattice Boltzmann (LB) \n\nD\n3\nQ\n19\n\n model was used\u00a0[48]. Although the reactions modeled are not isothermal processes, this assumption was made as the systems described are small in scale, and because of the large surface-to-volume ratio in such systems, no significant temperature gradients are expected. The central piece in the LB method is the LB equation: \n\n(1)\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\n\n\n\nx\n\n\u2192\n\n+\n\n\n\n\ne\n\n\u2192\n\n\n\ni\n\n\n\u0394\nt\n,\n\nt\n+\n\u0394\nt\n\n\n\u2212\n\n\nf\n\n\ni\n\n\ns\n\n\n\n\n\n\nx\n\n\u2192\n\n,\n\nt\n\n\n=\n\n\n\u03a9\n\n\ni\n\n\ns\n\n\n;\n\ns\n=\n\n\nCO\n\n\n2\n\n\n,\n\n\nCH\n\n\n4\n\n\n,\n\n\nH\n\n\n2\n\n\n,\nCO\n,\n\n\nH\n\n\n2\n\n\nO\n,\n\n\n\nwhere \n\n\nf\n\n\ni\n\n\ns\n\n\n is the distribution function of species \ns\n, pointing in direction \ni\n. \nf\n\u2019s are treated as the amount of substance at location \n\n\nx\n\n\u2192\n\n at time \nt\n with lattice velocity \n\n\n\n\ne\n\n\u2192\n\n\n\ni\n\n\n. The LHS of the equation represents the streaming of the \nf\n\u2019s in a time-step \n\n\u0394\nt\n\n and the RHS represents the collision with the collision operator \n\n\n\u03a9\n\n\ni\n\n\ns\n\n\n. The collision model of choice here is the two-relaxation-time collision (TRT)\u00a0[49]: \n\n(2)\n\n\n\n\n\u03a9\n\n\ni\n\n\ns\n\n\n=\n\u2212\n\n\n\n\nf\n\n\ni\n\n\ns\n+\n\n\n\u2212\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n+\n\n\n\n\n\n\ne\nq\n\n\n\n\n\n\n\u03c4\n\n\ns\n\n\n+\n\n\n\n\n\u2212\n\n\n\n\nf\n\n\ni\n\n\ns\n\u2212\n\n\n\u2212\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\u2212\n\n\n\n\n\n\ne\nq\n\n\n\n\n\n\n\u03c4\n\n\ns\n\n\n\u2212\n\n\n\n\n.\n\n\n\n\n\n\n\n\u03c4\n\n\ns\n\n\n+\n\n\n and \n\n\n\u03c4\n\n\ns\n\n\n\u2212\n\n\n are the symmetric and asymmetric relaxation times of the substance \ns\n. When modeling fluid flow, one controls the fluid\u2019s kinematic viscosity \n\u03bd\n through the symmetric relaxation time (\n\n\n\n\u03bd\n\n\ns\n\n\n=\n\n\n1\n\n\n3\n\n\n\n\n\n\n\u03c4\n\n\ns\n\n\n+\n\n\n\u2212\n\n\n1\n\n\n2\n\n\n\n\n\n)\u00a0[49]. In such applications \n\n\n\u03c4\n\n\ns\n\n\n\u2212\n\n\n remains a free parameter. Specifically for gas-flow applications in microchannels the latter can be manipulated to give the correct slip-flow at solid walls\u00a0[37]. However, when modeling advection\u2013diffusion systems with TRT, the asymmetric relaxation time is used to determine the species\u2019 molecular diffusion coefficient \nD\n\n (\n\n\n\nD\n\n\ns\n\n\n=\n\n\n1\n\n\n3\n\n\n\n\n\n\n\u03c4\n\n\ns\n\n\n\u2212\n\n\n\u2212\n\n\n1\n\n\n2\n\n\n\n\n\n)\u00a0[49]. In this case \n\n\n\u03c4\n\n\ns\n\n\n+\n\n\n becomes a free parameter. As present study is dealing with a flowing mixture of gases, the two cases are combined: \n\n\n\u03c4\n\n\ns\n\n\n+\n\n\n determines the species \ns\n\u2019s kinematic viscosity, while \n\n\n\u03c4\n\n\ns\n\n\n\u2212\n\n\n determines its molecular diffusivity coefficient. \n\n\nf\n\n\ni\n\n\ns\n+\n\n\n and \n\n\nf\n\n\ni\n\n\ns\n\u2212\n\n\n in Eq.\u00a0(2) represent the species \ns\n\u2019s symmetric and asymmetric link populations, respectively, which are computed as: \n\n\n\nf\n\n\ni\n\n\ns\n\u00b1\n\n\n=\n\n\n1\n\n\n2\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\u00b1\n\n\nf\n\n\n\n\ni\n\n\n\u0304\n\n\n\n\ns\n\n\n\n\n\n. \n\n\ni\n\n\n\u0304\n\n\n here represents the in-space opposite facing lattice direction (\n\n\n\n\n\ne\n\n\u2192\n\n\n\n\n\ni\n\n\n\u0304\n\n\n\n\n=\n\u2212\n\n\n\n\ne\n\n\u2192\n\n\n\ni\n\n\n\n, for further information on the \n\nD\n3\nQ\n19\n\n lattice see\u00a0[48]). Similarly \n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n+\n\n\n\n\n\n\ne\nq\n\n\n and \n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\u2212\n\n\n\n\n\n\ne\nq\n\n\n represent such equilibrium link populations and are computed as: \n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\u00b1\n\n\n\n\n\n\ne\nq\n\n\n=\n\n\n1\n\n\n2\n\n\n\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\n\n\n\ne\nq\n\n\n\u00b1\n\n\n\n\n\n\nf\n\n\n\n\ni\n\n\n\u0304\n\n\n\n\ns\n\n\n\n\n\n\ne\nq\n\n\n\n\n\n. \n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\n\n\n\ne\nq\n\n\n is the equilibrium distribution function of the substance \ns\n in \ni\nth direction. \n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\n\n\n\ne\nq\n\n\n depends on the local pre-collision density of \ns\n, \n\n\n\u03c1\n\n\ns\n\n\n\u22c6\n\n\n, which is proportional to \ns\n\u2019s partial pressure and it also depends on the local flow velocity \n\n\nu\n\n\u2192\n\n: \n\n(3)\n\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\n\n\n\ne\nq\n\n\n=\n\n\nw\n\n\ni\n\n\n\n\n\u03c1\n\n\ns\n\n\n\u22c6\n\n\n\n\n1\n+\n3\n\n\n\n\ne\n\n\u2192\n\n\n\ni\n\n\n\n\nu\n\n\u2192\n\n+\n\n\n9\n\n\n2\n\n\n\n\n\n\n\n\n\n\ne\n\n\u2192\n\n\n\ni\n\n\n\n\nu\n\n\u2192\n\n\n\n\n\n2\n\n\n\u2212\n\n\n3\n\n\n2\n\n\n\n\n\n\nu\n\n\u2192\n\n\n\n2\n\n\n\n\n,\n\n\n\nwith \n\n\nw\n\n\ni\n\n\n being the equilibrium weight in \ni\nth direction. \n\n\nu\n\n\u2192\n\n is calculated by first computing all post-collision densities of \ns\n, \n\n\n\u03c1\n\n\ns\n\n\n: \n\n(4)\n\n\n\n\n\u03c1\n\n\ns\n\n\n=\n\n\n\u2211\n\n\ni\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n,\n\n\n\nand then by computing the sum of mesoscopic momenta: \n\n(5)\n\n\n\n\nu\n\n\u2192\n\n=\n\n\n1\n\n\n\n\n\u2211\n\n\ns\n\n\n\n\nM\n\n\ns\n\n\n\n\n\u03c1\n\n\ns\n\n\n\n\n\n\n\u2211\n\n\ns\n\n\n\n\n\u2211\n\n\ni\n\n\n\n\nM\n\n\ns\n\n\n\n\n\n\ne\n\n\u2192\n\n\n\ni\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n.\n\n\n\n\nNote that here the \n\n\nf\n\n\ni\n\n\ns\n\n\n\u2019s are multiplied by the dimensionless molecular weights of \ns\n, \n\n\nM\n\n\ns\n\n\n. This ensures mass conservation in the system. \n\n\n\u03c1\n\n\ns\n\n\n\u22c6\n\n\n, the pre-collision density, is computed by also including a macroscopic reaction term, which accounts for the RWGS reaction. This is included in the model as follows\u00a0[41]: \n\n(6)\n\n\n\n\n\u03c1\n\n\ns\n\n\n\u22c6\n\n\n=\n\n\n\u03c1\n\n\ns\n\n\n+\n\n\n\n\n\n\n\n\u2212\n\n\nk\n\n\n+\n\n\n\n\n\u03c1\n\n\n\n\nCO\n\n\n2\n\n\n\n\n\n\n\u03c1\n\n\n\n\nH\n\n\n2\n\n\n\n\n+\n\n\nk\n\n\n\u2212\n\n\n\n\n\u03c1\n\n\nCO\n\n\n\n\n\u03c1\n\n\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\n\n\n\u03c4\n\n\ns\n\n\n+\n\n\n;\n\n\ns\n=\n\n\nCO\n\n\n2\n\n\n,\n\n\nH\n\n\n2\n\n\n\n\n\n\n\n\n\n\u2212\n\n\n\nk\n\n\n+\n\n\n\n\n\u03c1\n\n\n\n\nCO\n\n\n2\n\n\n\n\n\n\n\u03c1\n\n\n\n\nH\n\n\n2\n\n\n\n\n\u2212\n\n\nk\n\n\n\u2212\n\n\n\n\n\u03c1\n\n\nCO\n\n\n\n\n\u03c1\n\n\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\n\n\n\u03c4\n\n\ns\n\n\n+\n\n\n;\n\n\ns\n=\nCO\n,\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\n0\n;\n\n\ns\n=\n\n\nCH\n\n\n4\n\n\n\n\n\n\n\n.\n\n\n\n\n\n\n\nk\n\n\n+\n\n\n is a second order reaction constant for the RWGS reaction, while \n\n\nk\n\n\n\u2212\n\n\n is the constant of the reverse reaction. The equilibrium constant (\n\n\n\nk\n\n\ne\nq\n\n\n=\n\n\n\n\nk\n\n\n+\n\n\n\n\n\n\nk\n\n\n\u2212\n\n\n\n\n\n) was set to 5\u00a0[50]. Other side reactions were neglected here.The catalytic DRM reaction is carried out at the solid nodes, where the ordinary half-way bounce-back boundary condition with a slight modification is used to describe the walls. Slip flow at solid walls was not controlled and the effects of slip-flow on model accuracy were not studied in this work. The bounce-back can be carried out in the normal way, where at non-catalytic walls the incoming \n\n\nf\n\n\ni\n\n\ns\n\n\n\u2019s get translated into \n\n\nf\n\n\n\n\ni\n\n\n\u0304\n\n\n\n\ns\n\n\n\u2019s or at catalytic walls they get translated to a modified \n\n\n\n\nf\n\n\n\u02c6\n\n\n\n\ni\n\n\ns\n\n\n: \n\n(7)\n\n\n\n\nf\n\n\n\n\ni\n\n\n\u0304\n\n\n\n\ns\n\n\n=\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n;\n\n\nwall\n\n\n\n\n\n\n\n\nf\n\n\n\u02c6\n\n\n\n\ni\n\n\ns\n\n\n;\n\n\ncatalyst\n\n\n\n\n\n.\n\n\n\n\nAt solid nodes marked as the catalyst, all reactants\u2019 ( \n\n\n and \n\n\n ) \nf\n\u2019s get converted to products\u2019 ( \n\n\n and \n\n\n ) \nf\n\u2019s. However, each site has a limited amount of available catalytic sites, which is controlled through a free parameter \n\u03ba\n. \n\u03ba\n represents the amount of available sites and it does not discriminate against any molecule type nor does molecule size affect how much space a molecule occupies on the catalytic site, i.e.\u00a0all species have the same affinity towards the catalyst surface. Further parameters could be implemented into this model to control this, but such a model would require a molecular-level approach to determine these parameters. Because \n\u03ba\n limits how much substance can get to the catalytic surface, there are two cases of how this boundary is handled. In case where \n\n\n\n\u2211\n\n\ns\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n>\n\u03ba\n\n: \n\n(8)\n\n\n\n\n\n\nf\n\n\n\u02c6\n\n\n\n\ni\n\n\ns\n\n\n=\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\u2212\n\n\n\u03ba\n\n\n\u02c6\n\n\n;\n\n\ns\n=\n\n\nCH\n\n\n4\n\n\n,\n\n\nCO\n\n\n2\n\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n+\n2\n\n\n\u03ba\n\n\n\u02c6\n\n\n;\n\n\ns\n=\nCO\n,\n\n\nH\n\n\n2\n\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n;\n\n\ns\n=\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\n\n,\n\n\n\nwhere \n\n\n\n\u03ba\n\n\n\u02c6\n\n\n=\n\n\n\n\nf\n\n\n\u0304\n\n\n\n\n\n\n\u2211\n\n\ns\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\n\n\u03ba\n\n, with \n\n\nf\n\n\n\u0304\n\n\n being either \n\n\nf\n\n\ni\n\n\n\n\nCH\n\n\n4\n\n\n\n\n or \n\n\nf\n\n\ni\n\n\n\n\nCO\n\n\n2\n\n\n\n\n, whichever is smaller in value (limiting reagent). In case where \n\n\n\n\u2211\n\n\ns\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\u2264\n\u03ba\n\n: \n\n(9)\n\n\n\n\n\n\nf\n\n\n\u02c6\n\n\n\n\ni\n\n\ns\n\n\n=\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n\u2212\n\n\nf\n\n\n\u0304\n\n\n;\n\n\ns\n=\n\n\nCH\n\n\n4\n\n\n,\n\n\nCO\n\n\n2\n\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n+\n2\n\n\nf\n\n\n\u0304\n\n\n;\n\n\ns\n=\nCO\n,\n\n\nH\n\n\n2\n\n\n\n\n\n\n\n\nf\n\n\ni\n\n\ns\n\n\n;\n\n\ns\n=\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\n\n.\n\n\n\n\nThe factor 2 in both cases is added to keep the model in line with the stoichiometry of the DRM chemical equation. It was assumed that the reaction only takes place on the particle surface, so internal diffusion in particles was neglected.As the densities of \ns\n\u2019s (partial pressures) at the outlet were unknown, a constant density boundary condition could not be used to describe the outlet of the system. Instead a constant velocity boundary condition was used for the outlet. However, at the inlet the velocity was supposed to be set to a constant value (inlet velocity set in the experiments by the mass flow rate of gases), but more importantly the ratio of different species at the inlet needs to be set to 1:1:0:0:0 for \n\n\n , \n\n\n , \n\n\n , \n\n\n and \n\n\n , respectively. So a constant density boundary was used at the inlet. This could, however, not ensure that the velocity at the inlet would indeed match the inlet velocity set in the experiments. To address this issue a regulating algorithm was added to the inlet boundary condition, which ensured that the \n\n\n\u03c1\n\n\ns\n\n\n there were such, that the inlet velocity was correct. The boundary conditions used were the non-equilibrium boundaries by Guo et\u00a0al.\u00a0[51].Because the product gases are fast diffusing, a significant mass loss was initially observed at the inlet. This was due to products\u2019 partial pressures being kept at 0 at the inlet. This issue could be solved by making the simulation domain (domain 2) longer, but this would increase the computational costs. The solution was to split the domain in two, as described in a previous section. The added domain (domain 1) was used only to model the transport of products with a macroscopic model. This eliminated the mass loss due to back-diffusion. A simple 1D model was used for this: \n\n(10)\n\n\n\n\nD\n\n\ns\n\n\n\n\n\n\nd\n\n\n2\n\n\n\n\n\u03c1\n\n\ns\n\n\n\n\nd\n\n\nx\n\n\n2\n\n\n\n\n=\n\n\nu\n\n\nx\n\n\n\n\nd\n\n\n\u03c1\n\n\ns\n\n\n\n\nd\nx\n\n\n;\n\n0\n\u2264\nx\n\u2264\n\n\nL\n\n\n1\n\n\n,\n\n\n\nwhere \n\n\nu\n\n\nx\n\n\n is the average flow velocity in \nx\n-direction at domain 2 inlet and \n\n\nD\n\n\ns\n\n\n is the species\u2019 diffusivity. \n0\n is the inlet to domain 1 and \n\n\nL\n\n\n1\n\n\n is the junction of the two domains. Shooting method was used to solve Eq.\u00a0(10), by applying the following boundary conditions: \n\n\n(11)\n\n\n\n\n\u03c1\n\n\ns\n\n\n\n\n|\n\n\nx\n=\n0\n\n\n=\n0\n,\n\n\n\n\n(12)\n\n\n\n\n\n\n\n\n\u03c1\n\n\ns\n\n\n\n\n|\n\n\nx\n=\n\n\nL\n\n\n1\n\n\n\n\n\n\n\n\n\n\n1\n\n\n\n\n=\n\n\n\n\n\n\n\u03c1\n\n\ns\n\n\n\n\n\n\nL\n\n\n1\n\n\n\n\n\n\n\n\n\n\n2\n\n\n\n\n.\n\n\n\n\n\n\n\n\n\n\u03c1\n\n\ns\n\n\n\n\n\n\nL\n\n\n1\n\n\n\n\n\n is the species\u2019 density at the domain 2 inlet. This value was then modified by a combination of \n\n\n\n\n\n\nd\n\n\n\u03c1\n\n\ns\n\n\n\n\nd\nx\n\n\n\n\n|\n\n\nx\n=\n\n\nL\n\n\n1\n\n\n\n\n\n\n\n\n\n\n1\n\n\n\n\n (which was found by solving the model) and \n\n\n\n\n\n\nd\n\n\n\u03c1\n\n\ns\n\n\n\n\nd\nx\n\n\n\n\n|\n\n\nx\n=\n\n\nL\n\n\n1\n\n\n\n\n\n\n\n\n\n\n2\n\n\n\n\n.\n\n\nGiven the steady-state nature of this model, the dynamic-state solution was not necessarily physical, but the 1D model eventually \u201ccaught-up\u201d with the LB model. The fluid dynamics in the 1D model were not considered in this study and thus the pressure drop in the domain 1 was neglected and the bulk reaction was also neglected in this domain.The computational domains for the lattice Boltzmann model (domain 2) were created with Blender\u00a0[52]. Firstly four different shapes were designed by hand in Blender for both \u2014 catalytic and filler particles. Then the reactors\u2019 channels were designed. Next the particles were multiplied to the amount that was expected in the experiments by accounting for particles\u2019 average density and their average size, and sample masses. Then Blender\u2019s physics engine was used to randomly load these particles in the channels. The geometries were then exported and turned into lattices, which could be used in the LB computations. The staircase shape of the walls that was obtained through this process further randomized the particles\u2019 shapes.The computations used the model described above to compute (partial) pressure and velocity profiles inside the two-plate and conventional fixed-bed reactors. Total volumetric flow rates (\n\n\nV\n\n\n\u0307\n\n\n) studied in both systems were 20, 40 and 60mLmin\u22121. Because \n\n\nV\n\n\n\u0307\n\n\n\u2019s in the experiments were defined at room temperature, the inlet flow rate for the computations was estimated with the ideal gas law (\n\n\n\nV\n\n\n\u0307\n\n\n\u221d\nT\n\n) and mass conservation law (\n\n\n\n\n\nm\n\n\n\u0307\n\n\n\n\n1\n\n\n=\n\n\n\n\nm\n\n\n\u0307\n\n\n\n\n2\n\n\n\n). The viscosities of gasses as well as their molecular diffusivity coefficients were obtained from Engineering ToolBox web page\u00a0[53,54]. The values of the latter had to be extrapolated to get an estimate of the values at the studied conditions (500\u00b0C ).To first determine parameters \n\u03ba\n and \n\n\nk\n\n\n\u00b1\n\n\n, the computations for the conventional fixed-bed system were run, as this was due to its smaller size computationally less expensive. The two parameters were at first chosen as an arbitrary guess and then adjusted to capture the behavior exhibited in the experiments. The same values of both parameters were then used to simulate the process in the two-plate system without further tuning.Specific surface area (SSA) of the \n\n\n and \n\n\n samples are depicted in Table\u00a01. Catalysts showed SSA values of 98 and 93m2\ng\u22121 for \n\n\n and \n\n\n solids, respectively. The near equal values of SSA indicate that the \n\n\n deposition on \n\n\n does not show any considerable effect on surface area of \n\n\n . Also, the total pore volume and average pore diameter showed similar values.\n\n\n\nFig.\u00a03 shows the X-ray diffractograms of calcined \n\n\n and \n\n\n solids as well as of reduced and re-oxidized catalyst samples. The calcined \n\n\n sample showed high intense patterns at \n\n2\n\u03b8\n=\n28\n.\n55\n\n, 33.03, 47.48, 56.29 and 59.12\n\u2218\n. These diffraction lines are well corroborated with the crystal planes (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2) of cubic fluorite of \n\n\n as observed in JCPDS file 34-0394\u00a0[4]. In \n\n\n sample, all the high intense diffraction peaks are related to the \n\n\n fluorite and no \n\n\n peaks are observed. The addition of \n\n\n to \n\n\n showed a considerable shift in the \n\n\n diffraction peak from 28.55 to 28.68\n\u2218\n. This higher \n\n2\n\u03b8\n\n shift indicates incorporation of \n\n\n into \n\n\n lattice by the formation of \n\n\n solid solution\u00a0[55,56]. The reduction of \n\n\n catalyst at 300\u00b0C showed considerable decrease in the intensity of peaks, indicating that the reduction exhibited a significant effect on the crystallite size (Table\u00a02). The oxidation step conducted after reduction not only increases the peak intensity but also shows a gradual shift of the peak to its original position. This indicates that the oxidation of this material shows a reversible formation of \n\n\n solid solution. The re-reduction carried out at 500\u00b0C showed a little shift towards lower angle and also decrease in peak intensity indicating that the reduction process showed substantial changes in the \n\n\n lattice and crystallinity of the material. These shifts and intensities were clearly observed by a closer look (see Supplementary material, Fig. S1).Calculated lattice parameter values of \n\n\n (1 1 1) plane after considering the thermal expansion of \n\n\n and \n\n\n are presented in Supplementary material, Fig. S2. The lattice parameter value of \n\n\n was found to be equal to 5.4157\u00c5, further the addition of \n\n\n decreased the lattice parameter to a value of 5.3910\u00c5. As reported in literature, the decrease in lattice parameter can be referred as the \n\n\n incorporation into \n\n\n lattice in order to form a \n\n\n solid solution phase.Reduction of \n\n\n catalyst at 300\u00b0C showed a considerable increase in the lattice parameter value from 5.3910 to 5.4122\u00c5. According to literature, during reduction the transformation of \n\n\n from 4+ to 3+ resulted in an increase of ionic radii. So, the lattice expansion is mainly due to these high ionic radii species. Another possibility of lattice expansion could be due to the exsolution of \n\n\n . During reduction, oxygen atoms were stripped from the support lattice thus the electrons resulting from this reaction caused the decrease of support oxidation state, thereby expanding the lattice\u00a0[57]. Further oxidation of the reduced \n\n\n sample resulted in a similar lattice parameter value as the one belonging to the fresh catalyst, thus indicating that the reduction and oxidation cycles are reversible in \n\n\n . Re-reduction of oxidized catalyst at 500\u00b0C showed further enhancement in the lattice parameter value; the same behavior was also observed by other researchers\u00a0[58,59]. It should be noted that the re-oxidation step following the reduction at 500\u00b0C also exhibited the lattice parameter value similar to the one of the fresh catalyst sample, thus further confirming the reversible red-ox cycles in \n\n\n solid.\nFig.\u00a04 shows the calculated lattice parameter values for different crystalline plane of \n\n\n . After the addition of \n\n\n to \n\n\n , one can clearly observe a decrease in lattice parameter values as well as the same trend for all planes, thus indicating the uniform influence of \n\n\n on \n\n\n lattice. When the catalyst was reduced under \n\n\n till 300\u00b0C , some planes showed higher lattice parameter values and some planes exhibited lower values than the \n\n\n support. This observation supports the expansion of lattice that results into distortion of cubic \n\n\n . In-situ oxidation promotes the lattice parameter values to their original states, which is in agreement with the previously explained observation that the reconstruction of \n\n\n lattice is reversible. The reduction step conducted at 500\u00b0C caused an additional expansion of the ceria lattice due to the enhancement in extent of \n\n\n reduction. However, re-oxidation conducted after the second reduction cycle was able to reproduce lattice parameter values very similar to the ones determined for the fresh catalyst sample.\n\nFig.\u00a05 shows the results of TPR analysis in which the reduction behavior of \n\n\n catalyst was examined by undergoing different reduction cycles. The fresh catalyst was first reduced in 5% \n\n\n gas stream from RT to 300\u00b0C and then kept at this temperature for 1h. The obtained TPR profile (marked as TPR 300 in Fig.\u00a05) can be divided into four different reduction zones at 167, 237, 262 and 300\u00b0C , which were further denoted as \n\u03b1\n1, \n\u03b1\n2, \n\u03b2\n1 and \n\u03b2\n2 bands\u00a0[11,56,60,61]. \n\u03b1\n1 and \n\u03b1\n2 bands are related to the adsorbed oxygen and/or surface active oxygen species that can be easily reduced at lower temperatures. \n\u03b2\n1 band can be referred to the reduction of \n\n\n solid solution, while \n\u03b2\n2 reduction band could be attributed to the reduction of isolated \n\n\n and/or ceria. On the other hand, reduction of the re-oxidized sample at 500\u00b0C (TPR 500 profile) shows the same bands, however, \n\u03b2\n1 and \n\u03b2\n2 bands are shifted slightly to higher temperatures. This high-temperature shift may be attributed to the highly interacted \n\n\n in the \n\n\n lattice. As observed in the XRD examination, expansion of the lattice might increase the possibility of reduction of \n\n\n trapped in the ceria lattice. The total hydrogen consumption during the TPR 300 run was 15.0mLg\u22121, while in the case of re-reduced sample at 500\u00b0C this value was equal to 15.6mLg\u22121. The extent of ceria reduction is 23 and 24% for TPR 300 and TPR 500 runs, respectively. This is in agreement with the results illustrated in Fig.\u00a04 which indicate that the extent of \n\n\n reduction is enhanced at higher temperatures, thereby an increase in hydrogen consumption is noticed in the re-reduced sample.\nFig.\u00a06 shows the results of \n\n\n pulse chemisorption analysis for the \n\n\n catalyst undergoing different reduction cycles. The reported values of \n\n\n dispersion were found to be reproducible. The catalyst reduced at 300\u00b0C showed \n\n\n dispersion of 35%. Surprisingly, the catalyst reduced at 500\u00b0C showed higher dispersion (i.e.\u00a044%) compared to the solid reduced at 300\u00b0C . This can be explained by means of TPR and XRD studies, where high-temperature reduction showed expansion of \n\n\n lattice that can fetch \n\n\n -containing ensembles trapped inside the \n\n\n lattice. Participation of trapped (i.e.\u00a0re-surfaced) \n\n\n ensembles as an active phase might enhance the dispersion of \n\n\n after reduction at higher temperature. In addition to the above and as described during the explanation of results of XRD analysis, exsolution of \n\n\n from the solid solution is another possibility of lattice expansion. At high-temperature reduction, the increase in lattice expansion is due to the higher extent of \n\n\n exsolution from the \n\n\n lattice. Consequently, this increases the dispersion of \n\n\n after the reduction of the catalyst at 500\u00b0C . Further, after re-reducing the \n\n\n catalyst at 500\u00b0C (third cycle in Fig.\u00a06) a slight decrease of \n\n\n dispersion was observed that may be attributed to sintering. Mean particle size of \n\n\n ensembles calculated from the \n\n\n pulse chemisorption study is 1.6, 1.3 and 1.4nm (as it derives from TPR 300 cycle 1, TPR 500 cycle 2 and cycle 3 analyses, respectively).\nFig.\u00a07 shows SEM and TEM images of fresh \n\n\n and \n\n\n samples. The rod-shaped morphology of \n\n\n is observed from the figure. Further, the morphology has still existed even after \n\n\n deposition.Results of the catalytic activity test conducted in the two-plate reactor with varying reduction conditions of the \n\n\n catalyst prior to the run are shown in Fig.\u00a06. At \n\nT\n=\n500\n\n\u00b0\nC\n\n, the equilibrium conversions of \n\n\n and \n\n\n are 15 and 23%, respectively; the conversions reported in this study are below the equilibrium levels. The catalytic activity increased with the increase of reduction temperature from 300 to 500\u00b0C ; an increase of methane conversion from 5.6 to 10.5% was observed. A similar increment (from 8.5 to 14.9%) was measured in the case of \n\n\n conversion. We believe that the observed behavior could be explained by taking into account an identical trend regarding dispersion of \n\n\n in the \n\n\n catalyst pre-reduced before the DRM reaction at different temperatures (black curve in Fig.\u00a06).\nFig.\u00a08 shows the results of catalytic activity measurements conducted in both \u2014 two-plate reactor and conventional fixed-bed reactor varying the total gas flow rate from 20 to 60mLmin\u22121 with an increment of 20mLmin\u22121. The two reactor units showed promising activity: for instance, in the two-plate reactor conversions of \n\n\n and \n\n\n were found to be 10.5 and 15%, respectively, when utilizing the gas flow rate of 20mLmin\u22121. For the same gas flow rate, the conventional reactor showed conversions of \n\n\n and \n\n\n equal to 12.7 and 18.1%, respectively. Higher catalytic activity was observed in the conventional fixed-bed reactor compared to the two-plate reactor, although a double catalyst loading was used in the latter reactor system. The same trend was noticed for all the gas flow rates investigated. Furthermore, in the given range of experimental conditions enhanced production of water during the DRM reaction was observed in the two-plate reactor, whereas in the conventional reactor water formation was negligible. Water formed is mainly due to the RWGS reaction ( \n\n\n reacts with hydrogen and produces water and \n\n\n ). Higher conversion of \n\n\n over \n\n\n also indicates the co-existence of the RWGS reaction\u00a0[62]. A close observation on the conversion vs. time profiles presented in Fig.\u00a08 shows a noticeable difference in \n\n\n conversions as a function of time with an increase of gas flow rate. In the conventional fixed-bed reactor (Fig.\u00a08b), \n\n\n and \n\n\n conversions decrease in the same manner with the change of gas flow rates, whereas in the two-plate reactor the decrease of \n\n\n conversions is not in accordance with the temporal \n\n\n conversion profile. Low differences in \n\n\n conversions when increasing gas flow rate in the two-plate reactor are mainly due to the participation in side reaction (RWGS). One can therefore conclude that \n\n\n participation in RWGS to produce water is highly favorable in the two-plate reactor compared to the conventional fixed-bed reactor. Formation of water slowly decreased with the increase of gas flow rate, indicating that the decrease of contact time decreases the extent of side reactions (i.e.\u00a0RWGS reaction).\nIn the two-plate reactor, single-layered catalytic particles were placed at nearly 10cm length; on the other hand, in the conventional fixed-bed reactor the catalytic bed was of 9mm diameter and 3mm height. As such, the possibility of \n\n\n to react with \n\n\n is much higher along the catalytic bed in the two-plate reactor than in the short catalyst bed in the conventional fixed-bed reactor. Consequently, the two-plate reactor favored the progress of side RWGS reaction which is evidenced by (i) low decreases of \n\n\n conversions when increasing gas flow rate (compare \n\n\n conversion vs. time profiles for both reactor systems in Fig.\u00a08) and (ii) higher extent of water formation.The reaction rates calculated on the basis of \n\n\n and \n\n\n conversions are illustrated in Fig.\u00a09. In the two-plate reactor, the reaction rate (based on \n\n\n conversion) increased from 1.08 to 2.24mmolg\n\n\n\nNi\n\n\n\u22121\n\n\n\ns\u22121 with the increase of gas flow rate from 20 to 60mLmin\u22121. On the other hand, the reaction rate (based on conversion of \n\n\n ) increased to a value of 0.86mmolg\n\n\n\nNi\n\n\n\u22121\n\n\n\ns\u22121 when increasing the gas flow rate from 20 to 40mLmin\u22121; further increase of gas flow rate exhibited zero effect on the reaction rate (based on conversion of \n\n\n ) indicating that the DRM reaction was conducted in the kinetic regime in the case of gas flow rates of 40 and 60mLmin\u22121. Very similar trend was observed also in the case of the conventional fixed-bed reactor. Because \n\n\n participates in side reactions (one should note that \n\n\n takes part in RWGS reaction which is much faster than methane decomposition), conclusions about the DRM kinetics based on the conversion of this reactant are not appropriate.\n\n\nThe computations ran until a steady-state solution was reached, which varied from case to case. Final concentration profiles obtained are presented in Fig.\u00a010. The transition between the two domains appears smooth, however, computing the data derivatives reveals some distortion in the transition (data not shown). From Figs. 10d\u201310f it is evident, that the omission of the transport in domain 1, i.e.\u00a0only computing the transport through domain 2 and assuming the concentrations at the domain 2 inlet to be constant, would result in significant mass loss in the system. Domain 1 appears to be less necessary in the two-plate system, as can be assumed from Figs. 10a\u201310c. Although the range of flow rates was the same in both systems, the difference in back-diffusion appears because of the different cross-section geometries of the system, which result in different linear velocities, entering the system. Additionally, the catalyst was diluted in the two-plate system, which in turn meant that the concentration gradients of the products were possibly lower near the inlet and combined with faster convective transport, there was less potential for back-diffusion. The model results also show that about the same (non-negligible) amount of water vapor is produced in both systems.\nFig.\u00a011 is displaying the steady-state solutions of the velocity field as well as the partial concentration of \n\n\n along both channels. The velocity profile in Fig.\u00a011a shows that in the two-plate system the bulk of the gas flow traveled over the fixed bed. This happened due to the bed height being uneven and the bed particles not being distributed evenly through the depth of the channel, which is a consequence of the bed particles being poured into an open reactor, which is then closed by pressing the two plates together. In turn this could affect the two-plate reactor\u2019s performance. However, due to the advanced diffusive properties of gasses and reactor\u2019s small dimensions, the extent of these effects is likely minimal. Fig.\u00a011b is displaying the steady-state solution in the conventional reactor. Here the distribution of particles appears to be uniform and local spikes in flow velocity are observed due to gas being squeezed through narrow gaps in the bed.\n\n\nTables\u00a03 and 4 are comparing the results of computations with the experiments in both systems. There appear to be some differences in the reactants\u2019 conversions, where the model mostly underestimates especially the \n\n\n conversion and the \n\n\n conversion gets lower with increasing flow rate faster in the model than with the experiments. The model predicts the \n\n\n ratio well for lower flow rates, however, at 60mLmin\u22121 it deviates from the experiments in both systems.\n\nThe results of both \u2014 computations and experiments, are plotted together on one graph in Fig.\u00a012. Here the conversion is adjusted to the amount of catalyst in each system (conversion of the two-plate system was divided by 2, to account for twice the amount of the catalyst) and is plotted against the Reynolds number in the reactors (\n\nR\ne\n=\n\n\n\n\nD\n\n\np\n\n\n\n\n\n\nu\n\n\n\u0307\n\n\n\n\nx\n\n\n\n\n\u03f5\n\u03bd\n\n\n\n). Particle diameter (\n\n\nD\n\n\np\n\n\n), average linear velocity (\n\n\n\n\nu\n\n\n\u0307\n\n\n\n\nx\n\n\n) and bed porosity (\n\u03f5\n) used to calculate each \n\nR\ne\n\n were measured in the computations and assumed they were the same in experiments. This can give an insight into lower conversions of the two-plate system: as its cross-section dimensions are smaller the gases flow faster around the particles and the effective residence time of reacting species around catalytic sites is shorter than in the conventional system. Other differences may be setting the two systems apart however, as there is a noticeable \u201ckink\u201d in the pattern in the transition between the results of the two systems, especially with the \n\n\n conversions.\nFig.\u00a012 displays that the model successfully reproduced the reactant conversion trend in both systems. The model however struggled to correctly capture water vapor production shown in the experiments. In the experiments the conventional fixed-bed reactor produced significantly smaller amounts of water vapor than the two-plate system did. The model however predicts similar amounts of water vapor to be produced in both. From this it could be assumed that different residence times alone were probably not responsible for that.The synthesis of \n\n\n rods and their lattice modification with \n\n\n deposition was observed by means of XRD analysis. The lattice contraction of \n\n\n with deposition of \n\n\n illustrates that the incorporation of \n\n\n into the \n\n\n lattice results in the formation of \n\n\n solid solution. In-situ XRD examination conducted in the reductive atmosphere revealed that the expansion of \n\n\n lattice is enhanced with temperature. However, the lattice parameter values are quite similar to the untreated catalyst after oxidation, which evidences the reversible redox nature of \n\n\n lattice. TPR measurements showed an increase in consumption of \n\n\n content with temperature due to enhanced \n\n\n reduction. It was observed in low- and high-temperature TPR reduction cycles that trapping of \n\n\n ensembles in the \n\n\n lattice enhances dispersion of \n\n\n on the support.The promising catalytic activity in the low-temperature dry methane reforming is observed in both between-two-plates microchannel fixed-bed reactor and conventional fixed-bed reactor. The RWGS reaction is highly dominated in the former reactor system, which is identified with the enhanced amount of water formed during the reaction.The choice of TRT LB appears to have been appropriate, as the two collision parameters could be used to manipulate the gas viscosity as well as its diffusivity. Furthermore, the multiscale model significantly reduced the computational costs, by using the computationally more expensive LB method only in the more complex section of the domain. It performed well for finding steady-state solutions, but it needed a couple of iterations to start returning qualitatively smooth concentration profiles. The catalytic reaction model developed here was able to project the two systems\u2019 behavior, however with limited accuracy. With more computing power \n\u03ba\n and \n\n\nk\n\n\n\u00b1\n\n\n could be tuned more precisely, which could possibly improve the results. Additionally, there appear to be further differences between the two studied systems other than their geometry, as the model predicts similar \n\n\n vapor formation in both, whereas the experiments showed that the conventional fixed-bed reactor produced significantly lower amounts of water. This suggests that the residence time alone is probably not the reason for differing performances.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The financial support of the Slovenian Research Agency through PhD Grant MR-39080 (FS), Grants P2-0150, P2-0191 and projects J7-1816 and N2-0067 is acknowledged. The support through the H2020 project COMPETE, Slovenia\n (Grant No. 811040) is also acknowledged. The authors would also like to thank Dr. Janez Zava\u0161nik and Dr. Gregor \u017derjav for the TEM and SEM images, respectively. The graphite gasket was kindly provided by SGL Carbon.Supplementary material related to this article can be found online at https://doi.org/10.1016/j.cej.2020.127498.The following is the Supplementary material related to this article. \n\nMMC S1\n\n.\n\n\n\n\n", "descript": "\n                  Experimental and theoretical modeling on low-temperature dry methane reforming over Ni-containing CeO2 rods was studied. The catalyst was characterized by means of N2 physisorption, in-situ XRD, TPR and H2 chemisorption techniques. The characterization studies revealed the distortion of CeO2 flourite structure due to the Ni incorporation. Lattice expansion (due to reduction) and contraction (due to oxidation) suggest the reversible redox nature of CeO2. Ni\u2013O\u2013Ce solid solution formation was evidenced by both XRD and TPR studies. H2 chemisorption study revealed that the catalyst reduction temperature plays a significant role in Ni dispersion. The catalyst showed similar activity trends in two model geometries: a between-two-plates microchannel fixed-bed reactor and a conventional fixed-bed reactor. The activity tests were conducted in the kinetic regime, where conversions of CH4 were not influenced with the gas flow rate. A lattice Boltzmann model for mixed gas flow was developed along with a boundary condition for catalytic sites. The lattice Boltzmann model was used in a multiscale simulation of the studied reaction systems and produced data that qualitatively matched the experiments.\n               "}
{"full_text": "Triglycerides are long-chain organic compounds that are primarily present in different vegetable oils, animal fats. They are composed of carbon, hydrogen, and oxygen linkages, classified as glycerides (triglycerides, diglycerides, monoglycerides). In addition to glycerides, free fatty acids and phospholipids with metallic impurities are also present in the aforementioned feedstocks. These feedstocks produce biofuels via different catalytic and non-catalytic routes such as transesterification, hydrothermal conversions, or hydroprocessing route [1\u201316]. Newer ways for the production of sustainable aviation biofuels from biomass are being developed. Among different routes, the conversion of triglycerides to hydrocarbon fuel in single/multiple steps by hydroprocessing reactions is comfortable and well known to refinery engineers.Triglycerides are bulky in size, and for complete reaction, the molecules need to access the active sites on the catalyst surface through its porous structure. The catalyst activity, stability, life, and renewability are the crucial criteria for catalyst selection. Researchers have modified the conventional hydrocracking catalyst properties to achieve the desired yield and selectivity for hydroprocessing reactions. To complete triglyceride conversion to hydrocarbon fuel, changes in active metal composition and loading, different kinds of active metals, additives, etc., have been studied. Apart from conventional hydroprocessing catalyst supports, many other materials such as clays, carbon, oxides, and other mixed oxides derivatives have been tried and tested for hydroprocessing vegetable oil reactions.The encapsulated metal catalyst is gaining popularity as an advanced hydroprocessing catalyst with enhanced activity, chemoselectivity, and increased durability. Noble metal catalysts, for the hydroconversion of vegetable oil or oxygenated feeds, are not uncommon [1\u20133], but their use with feeds having high sulfur (>1000\u00a0PPM) is infrequent. Sulfur and carbon monoxides are known to poison noble metal catalysts [4,5]. Hydroprocessing of triglycerides and their co-processing with refinery streams have been studied and extensively published in the literature studies [1\u201316]. The hydroconversion parameters' effect has been evaluated for neat vegetable oil [17\u201325] or co-processing of vegetable oil with refinery streams [26\u201333]. Although noble metal catalysts have better hydrogenation than sulfided catalysts, they are more prone to deactivation due to CO or sulfur components present in the feed (typically in refinery feeds) [2\u20133]. One way to use noble metals in high sulfur or CO atmosphere is by encapsulating these metals inside some cage, e.g., zeolite cage, where bulkier components like CO, H2S, or other sulfur compounds may not enter the cage, while the smaller molecules, like H2, quickly enter the cage and adsorb on the metal [4].Choi et al. [8] have studied the effect of an encapsulated catalyst on the activity and selectivity for the oxidative dehydrogenation of methanol and iso-butanol. The prepared material, zeolite-like sodalite (SOD) material, which has a small pore size, having framework consisting of six-membered ring apertures of window size 3.6\u00a0\u00c5 and 5.2\u00a0\u00c5 and a cage diameter of 6\u20138\u00a0\u00c5, encapsulated Pt nanoparticles of size\u00a0<\u00a01\u00a0nm could fit inside their cages [8\u201310]. Choi et al. [8] reported the maximum size of the encapsulated Pt nanoparticles as 0.5\u00a0nm (5\u00a0\u00c5). These cages have dual benefits; they would protect the Pt clusters from sintering and prevent impurities like H2S and other 'S' containing compounds having a size more significant than that of the zeolite window to enter the cage [4,6\u20138]. Goel et al. [6] used a similar concept to encapsulate metal clusters inside small-pore zeolites. The prepared material activity was then tested for oxidative dehydrogenation of methanol and iso-butanol and hydrogenation of ethene and toluene. Wang et al. [5], using a similar concept by confining the Pd cluster inside silicalite for hydrogen production via formic acid decomposition. Studies by Choi et al. [4] and Juhwan et al. [7] provide a molecular-level understanding of encapsulation and hydrogen spillover mechanism. The authors suggested that surface hydroxyls, presumably Br\u00f8nsted acidity, play a crucial role in hydrogen spillover and activity enhancement. The hydroxyl groups' role in hydrogen spillover was confirmed experimentally by changing the hydroxyl concentration in the encapsulated zeolites; this resulted in a considerable change in the hydrogenation activity. Srinivas and Rao [36] were the first to observe the H2 spillover phenomenon on Pt supported on the carbon surface. It is expected that encapsulated noble metal may provide extra spillover hydrogen [34\u201336] to nearby NiS and MoS2 sites, increasing hydrogenation [36\u201337]. Availability of more active hydrogen on active MoS2 and Ni3S2 sites would enhance the catalyst's hydrogenation function, which would help reduce naphthenes and aromatics produced during the hydroconversion of jatropha oil [38].Sibi et al. [4] used encapsulated Pt inside sodalite (SOD) cage with NiMo/ZSM-5 for the hydroconversion reactions of Jatropha oil and studied the effect of the process parameters on the hydrocarbon product yield and isomerization activity. Product yield pattern and isomerization activity were reviewed on this catalyst. Aviation fuel has stringent cold flow specification (<-47\u00a0\u00b0C. Since the iso-paraffins are better in cold flow properties than normal-paraffins, it is important to track the isomerization activity for each catalyst. The yield of kerosene and iso/normal hydrocarbon ratio was compared with the conventional hydroprocessing catalyst. The reported maximum kerosene range hydrocarbon yield for NiMo/Pt@SOD@HZSM-5 was six times higher than NiMo/HZSM-5 catalyst for >99% of triglycerides conversion.SiO2-Al2O3, a more economical catalyst support [21] for vegetable oil processing, is being used to support this study. As per the aviation fuel specification, it is necessary to limit aromatics and naphthenes in aviation range hydrocarbon under a specific range [9\u201313]. A noble metal, being more active for hydrogenation reactions, strongly affects the product composition like paraffin, naphthenes, and aromatics in the aviation range of hydrocarbon during vegetable oil conversion [4]. The composition of the aviation range hydrocarbons affects the specification and strongly affects the catalyst life. Naphthenes and aromatics are well-known precursors for coke deposition [14]. Cyclic hydrocarbons are more prone to coke formation compared to normal paraffin [14\u201315]. Though aromatics and naphthenes contribute to the coke formation, their presence in the aviation turbine fuel (ATF) range is essential. They improve the freezing point, lubricity, and fuel compatibility with engine parts where sealants are used [16].American Society for Testing and Materials (ASTM) specifies certain limits for naphthenes and aromatics for renewable and non-renewable based aviation fuels. Aviation fuel, meeting the specifications as per ASTM standards (ASTM D1655 for petroleum origin and ASTM D7566 for alternative routes), is suitable for aircraft worldwide. ATF obtained via hydroprocessing of esters, vegetable oils, animal fats are covered in Annexure 2 of ASTM D7566 document [10]. ASTM D7566 specifies the limits of a maximum 0.5% percentage of aromatics and 15 % for naphthenes. To meet the ASTM D7566 specification, keeping the aromatics and naphthenes for the aviation range hydrocarbons within these limits is necessary.Hydroconversion of vegetable oil produces mixed hydrocarbon ranges, namely gasoline, aviation, and diesel range hydrocarbons [17\u201328]. Hydrocarbons obtained via hydroconversion of vegetable oil consist of naphthenes, paraffin, and aromatics in the final product [21\u201322]. A higher percentage of aromatics in hydro-processed lipids is not desirable in ATF [21\u201322,25]. Xing et al. [25] explained the mechanism for the production of aromatics during hydroconversion of esters. At lower temperatures, it was reported that fatty acids are first saturated, and then a hydrodeoxygenation reaction occurs at the metal sites to form long-chain hydrocarbons. They also compared the mechanism of saturated and unsaturated feed. Unlike saturated feeds, unsaturated feed directly undergoes deoxygenation at acid sites before saturation. The long-chain hydrocarbons then crack into gases at Br\u00f8nsted acid sites; and then undergo Diels\u2013Alder reactions on the Lewis acid sites to form aromatics hydrocarbons (AHCs). Lewis acidity supports aromatics formation in the hydroprocessed product; at the same time, Br\u00f8nsted acid is also responsible for hydroisomerization and hydrocracking reactions [25]. A trade-off between Lewis/Bronsted acid sites is required to maximize ATF range hydrocarbon and reduce aromatic content. It is expected that by increasing the hydrogenating function of the conventional hydrocracking catalysts, hydrocarbons with lower aromatics and cyclic, without compromising the product yield, could be produced.For better product (ATF) quality, a balance of acidic functionality and hydrogenation activity is an essential criterion in catalyst selection. Verma et al. [21,22] discussed the support effect on the product distribution for aviation range hydrocarbon using jatropha oil as feed. The catalyst was tailored by tuning the zeolitic acidity and porosity. Reported aviation range hydrocarbon yield was 40\u201345%, with high isomerization selectivity (iso/normal ratio) in the aviation range [2\u20136]. In another experiment, using Ni-Mo supported on high surface area semi-crystalline ZSM-5, very high aviation range hydrocarbon yield (77%) with iso/normal alkane ratio of 2.5 was reported. In addition to the catalyst acidity, hydrogenation strength also affects both qualities and the quantity of the aviation range of hydroprocessed products.The use of a noble metal catalyst along with a sulfided metal catalyst has been proposed in this work to enhance the hydrogenation activity of the catalyst for hydrogenation reaction. It is observed that high hydrogenation functionality reduces the aromatics and naphthenes present in the ATF range hydrocarbons to make the product better in quality (as per ASTMD-7566 specifications). In addition to limiting aromatics and naphthenes, high hydrogenating catalysts are also shown to have higher catalyst life due to reduced coke formation. With a desired Pt cluster size inside sodalite cage and optimized Pt@SOD to NiMo/SiO2-Al2O3 ratio, the desired product quality (reduced aromatics and naphthenes in aviation range hydrocarbons) could be achieved.Anhydrous sodium aluminate, NaAlO2 (Riedel-de-Ha\u00ebn, \u226599.95%); Sodium hydroxide, NaOH (Sigma Aldrich; \u226599.8%); Fumed silica, SiO2 (Sigma Aldrich 99.8%); Tetraammineplatinum(II) nitrate, [Pt (NH3)4](NO3)2 (Acros, 99%); tetraethyl orthosilicate, TEOS (Sigma Aldrich, 99.999 %), tetrabutylammonium hydroxide, Tetrabutylammonium hydroxide, TBAOH (Acros, 10% of TBAOH in water); Octadecyl dimethyl (3-(trimethoxy-silyl) propyl) ammonium chloride, (ODAC, 60 % diluted in methanol from Gelest Inc.), and deionized water were used during synthesis. Ammonium heptamolybdate and nickel nitrate precursors were obtained from Sigma Aldrich.Pt encapsulated sodalite was prepared by the hydrothermal crystallization method as described in the literature [6]. The gels' composition is as follows: 20 Na2O: 1.0 Al2O3: 1.5 SiO2: 160\u00b7H2O (mol/mol). NaAlO2\nand NaOH were dissolved in DI water (deionized water) and mixed with fumed SiO2. Metal precursor, [Pt (NH3)4](NO3)2, was dissolved in 10\u00a0ml water, and the prepared solution dropwisewas added to the gel at the rate of 0.08\u00a0ml/min. The gelobtained was then transferredinto a polypropylene container (125\u00a0ml), sealed, and vigorously stirred homogenized, the final mixture for 10\u00a0min. The gelformed was stirredin at 400\u00a0rpm and 100\u00a0\u00b0C for 7\u00a0h. An oil bath maintained the temperature. The solid product was then collected using a fritted funnel, and the filtrate was washed with DI water repeatedly till the pH was 7\u20138. The sample was then dried in ambient air overnight (14\u00a0h) at 100\u00a0\u00b0C. It was then heated in air at 100\u00a0ml/min at 350\u00a0\u00b0C for 3\u00a0h and treated in 9% H2/He (100\u00a0ml/min) at 650\u00a0\u00b0C for 2\u00a0h.To prepare Pt@SOD-NiMo-SiO2-Al2O3 catalyst, 3.0\u00a0g of SiO2-Al2O3\nsupportwas well mixed with 1.0\u00a0g ofPt@SOD support (Pt\u00a0=\u00a01.0%)by reducing the size in a mortar pestle. 4% NiO, 18% MoO3/SiO2-Al2O3\nwas synthesized by the incipient wetness impregnation method described in the literature [25]. After the metal impregnation, the resultant mixture was molded into pellets. The pellets of size 1\u20132\u00a0mm size were used for reactor loading.N2\nadsorption\u2013desorptionisotherms were used to examine the physical properties like surface area, pore-volume, and a pore radius of the catalysts at \u2212196\u00a0\u00b0C (BelsorbMax, BEL, Japan). The catalyst aciditywas measured by the ammoniaadsorption\u2013desorptiontechnique using an instrument equipped with a thermal conductivity detector, model Micrometrics 2900 instrument (USA). 0.25\u00a0g of the sample was saturated with NH3\nat 120\u00a0\u00b0C and flushed with helium at the rate of 100\u00a0ml/min to remove physically adsorbed NH3.The desorption of NH3\nwas carried outin the helium flow condition at the heating rate of 10\u00a0\u00b0C/min.XRD analysisto confirm the crystallinityof the sodalitematerialwasdone usinga Bruker D8 diffractometer (step size 0.002\u00b0 and scanning rate 1\u00b0/min) using Cu K\u03b1 radiation (40\u00a0kV and 40\u00a0mA).A field emission scanning electron microscope(FEI Quanta 200F) was used to obtain the surface images of the catalyst. ETD detector and lanthanum hexaboride (LaB6) doped in tungsten filament was used as an X-ray source, under high vacuum condition. Secondary electrons of acceleration voltage in the range of 10\u201330\u00a0kVwere used. Samples were prepared by spreading them on glued carbon tape. Energy-dispersive X-ray (EDX) coupled with SEM was used for the surface elemental analysis. TEM imageswere takenby the TECNAI electron microscope operated at 75\u00a0kV. Sample preparation was done by dispersing it in ethanol, and it was then sonicated for 15\u00a0min. The dried sample was then placed on a carbon-coated copper grid.Inductively coupled plasma (ICP) emission was used to quantify the Pt loading on the catalyst. Metal dispersions andtemperature-programmed reduction (TPR) and were determined by hydrogen chemisorption using Micromeritics 2720 equipment. The catalyst was reduced by heating to 650\u00a0\u00b0C (at 2\u00a0\u00b0C\u00a0min\u22121) in H2\n(99.999%) and held foronehour andthen evacuated for 1\u00a0h at 650\u00a0\u00b0Cto remove any chemisorbed hydrogen. Pt dispersion and total hydrogen uptake were measured by manual injection. The catalyst was washed in naphtha and dried at 100\u00a0\u00b0C before the elementalanalysis.Agilent 7890A, Refinery gas analyzer (RGA), equipped with 2- thermal conductivity detector (TCD) detectors, 1-FID, and seven columns (Column 1 HayeSep Q 80/100 mesh, Column 2 HayeSep Q 80/100 mesh, Column 3 Molsieve 5A 60/80 mesh, Column 4 HayeSep Q 80/100 mesh, Column 5 Molsieve 5A 60/80 mesh, Column 6 DB-1, Column 7 HP-PLOT Al2O3) and PCM: Electronic pneumatics control (EPC) module was used to analyze the gases.Varian 3800-GC withVt-5\u00a0ms column (30\u00a0m * 0.25\u00a0mm, 0.25\u00a0\u00b5m) was used to analyze hydrocarbon products formed during the reaction. The vegetable conversion was observed to confirm constant activity. The GC oven temperature program was: 35\u2013150\u00a0\u00b0C (rate of heating 3\u00a0\u00b0C/min; hold time: 5\u00a0min), 150\u2013300\u00a0\u00b0C (rate of heating 12\u00a0\u00b0C/min; hold time: 5\u00a0min), and 300\u2013320\u00a0\u00b0C (rate of heating 15\u00a0\u00b0C/min; hold time: 15\u00a0min). The experimentswere repeated,andstandard deviationswerealso plotted for each product component.The liquid hydrocarbon products selectivity was calculated on a relative percentage basis considering the entire range of hydrocarbon products formed as 100%. The liquid products analyzed on liquid GCwere reportedas relative percentages of lighter components (<C9), middle-range hydrocarbons (C9-C15), heavier range hydrocarbons (C16-C18), and oligomers (>C18). Complete conversion of jatropha oil hasbeen observedin all the experimental runs.Jatropha oil, which consists of main triglycerides, is used as the feedstock in all the experimental runs [8]. The vegetable oil is spiked with a 0.025% dimethyl-disulfide (DMDS) spiking agent to reduce leaching chances. The lipid composition of Jatropha curcas has been reported in the literature [38]. Approximately 90% of the jatropha oil consists of triglycerides. Other components are mono-glycerides, diglycerides, polar lipids, sterols, and sterol esters [38]. The scheme followed over the catalyst surface during hydroprocessing of vegetable oil (jatropha oil) is explained in Fig. 1\n. Liquid feed was mixed with hydrogen and fed to the reactor. Temperature, pressure, H2/Liquid-feed, and liquid hourly space velocity are major controlling parameters for the product pattern and composition. The gaseous product obtained consists of unreacted hydrogen along with H2S, CO, CO2. Condensable gases include water vapor, propane, and other hydrocarbons (<C6). The liquid products obtained could be grouped as lighter components (<C9), Mid-range (C9-C15 or kerosene range), Heavier range (C16-C18, Diesel range), and Oligomers (>C18 hydrocarbons).The prepared catalystswere loadedinside stainless-steel tubular fixed bed reactors (1.3\u00a0cm ID and 30\u00a0cm in length). Pt@SOD-NiMo--SiO2-Al2O3(4\u00a0g and 3\u00a0g) were loaded inside the reactor. Jatrophaoil was usedas feed. Ceramic beads were loaded at the top part and bottom part of the reactor. The hydrogen pressure inside the reactorwas controlledby a back pressure regulator (TESCOM). The inlet gas flowrate was maintainedby a mass flow controller (Brooks). The catalyst bedtemperatures were measuredby K-type thermocouples connected with a microprocessor-based temperature controller system.A high-pressure liquid pump (HPLC pump, Eldex made) was used to maintain the reactor's desired liquid flow rates. Reaction conditionswere variedover a wide range of temperaturesbetween 350 and 450\u00a0\u00b0C, the pressurebetween 60 and 100\u00a0bar, H2/feed ratio between 1500 and 3000 NL/L. Liquid hourly space velocity (LHSV) of 0.5\u20133\u00a0h\u22121 was maintained during the runs. All the necessary measuring componentswere calibratedbefore theexperiments.The reactor outlet stream was sent to the gas\u2013liquid separator. The gas\u2013liquid separator has a much higher surge volume than the inlet lines; hence due to sudden expansion, the gaseous fractions (containing hydrogen and lighter hydrocarbons with small quantities of H2S, CO, CO2) were separated from the liquid product. The gaseous products were sent to an alkali solution and then vented through a gas meter. Gas was collected at the outlet of the gasometer for analysis using gas chromatography (Refinery Gas Analyzer, Agilent India Pvt). The liquid product was obtained by draining it from thehigh-pressureseparator, and it was then analyzed using gas chromatography (Varian 3800-GC with Vt-5\u00a0ms column).Experiments were carried out to compare the activity of Pt encapsulated NiMo/SiO2-Al2O3 catalyst with NiMo/SiO2-Al2O3 catalyst. The reaction scheme followed is shown in Fig. 1. Hydrogen is taken in excess in all the experiments to increase hydrogen solubility in liquid Jatropha feed. Pt encapsulated sodalite was prepared by hydrothermal crystallization method as reported in the experimental section. The preparation of Pt@SOD has been discussed in detail by Sibi et al. [4]. Pt@SOD, prepared, was characterized for its physical characteristics. After the incorporation of Pt inside the sodalite cage, the powder obtained was intimately mixed with SiO2-Al2O3, and then NiO and MoO3 were impregnated on the support. The final catalyst was then sulfided and reduced for reactivity test using Jatropha as a liquid feed. Product characterization includes product distribution (lighter, middle distillates, heavier hydrocarbons, and oligomeric compounds. Hydrogen activity was differentiated based on the product's componential analysis (PNA) obtained on these two different catalysts.Periodic, all-electron DFT calculations were performed using the double numerical plus polarization function (DNP) basis set of DMol3 module of Material Studio 8 (Biovia, San Diego) [39]. Revised Perdew Burke Erzenhof (RPBE) [40], generalized gradient approximation (GGA) exchange\u2013correlation functional was used for all the calculations. Convergence criteria for structure geometry optimization were set to 0.05\u00a0eV/\u00c5, 0.005\u00a0\u00c5, and 0.0001\u00a0eV, respectively, with respect to atom displacement, force, and energy. The smearing value of 0.005 au was used to improve the SCF convergenceThe TS search method available in DMol3 was used to find the transition state (TS) for each reaction step, the linear synchronous transit/quadratic synchronous transit (LST/QST) method was used to calculate the TS search. Obtained geometries TS search method further iterated using the \u201cTS optimization\u201d using the DMol3 module to complete refined TS structures. The vibrational frequencies of the transition state (TS) were analyzed. TS confirmed all the vibrational frequencies to be real, except one, in the reaction coordinate direction. The spin-polarized setup is used for both geometry optimization and TS calculations. The adsorption energy of the intermediate species (Eads) was calculated from the following equations,\n\n\n\n\nE\n\nads\n\n\n=\n\nE\n\nmol\n+\nP\nt\n-\nS\nO\nD\n\n\n-\n\nE\n\nPt\n-\nS\nO\nD\n\n\n-\n\nE\n\nmol\n\n\n\n\n\n\nEmol+Pt-SOD, EPt-SOD, and Emol are the energies of the molecule adsorbed at Pt3 encapsulated inside SOD, Pt3 encapsulated inside SOD, and molecule in the gas phase respectively.The activation energies (Ea) for spillover of hydrogen was calculated by the difference in the energy of the transition (ETS) and initial states (EIS)\n\n\n\n\nE\na\n\n=\n\nE\n\nTS\n\n\n-\n\nE\n\nIS\n\n\n\n\n\n\nThe initial structures of the SOD zeolite, having the chemical formula Na6Al6Si6O24, were adapted from the material studio database (shown in Fig. 2\n(a)). Pt3 triangular cluster is encapsulated inside the SOD cage, as shown in Fig. 2(b). Considering the reported maximum size of the encapsulated Pt nanoparticles as 5\u00a0\u00c5 by Choi et al. [8], the Pt3 cluster was chosen, having a diameter of 4.2\u00a0\u00c5, for this study. The Pt3 triangular cluster is encapsulated inside the SOD cage, as shown in Fig. 2(b). The geometry optimized SOD encapsulated Pt3 (Pt3@SOD) is used to study the chemisorption of H2 and the H spillover from Pt3 cluster to the SOD zeolite.The prepared catalyst was characterized by SEM, TEM, XRD for physicochemical characteristics. Support crystallinity is an essential aspect during the encapsulation of noble metal. The crystalline nature of the material (sodalite) was confirmed by XRD, SEM, and TEM analysis, as shown in Figs. 3, 4\n\n. XRD analysis of Pt@SOD is shown in Fig. 3. Peaks at 2\u03b8 values of 140, 240, 320, and 370 correspond to sodalite's crystal structure (JCPDS 11-0401). SEM-EDX of the sodalite surface confirms the absence of Pt on the surface (Fig. S1); however, the Pt particles in the sodalite were confirmed by ICP analysis (Table S1).Since the noble metal is being used along with the Ni and Mo oxide catalyst and the sulfidation, reducing the Pt is crucial before the catalyst hydrogenation activity runs. TPR was performed at different temperatures to finalize the reduction temperature. Fig. S1 shows the TPR of Pt@SOD; the prominent peak observed was in the range of 450\u2013650\u00a0\u00b0C. This peak may be attributed to the reduced state of Pt(0) from Pt (II) [4]. Peaks observed beyond 800\u00a0\u00b0C, may be due to the collapse of sodalite, causing the inner Pt to get exposed and reduced at this higher temperature. The SEM image of Pt@SOD reduced at 900\u00a0\u00b0C (Fig. 5\n) confirms the collapsed sodalite structure at a higher temperature (900\u00a0\u00b0C). The thermal stability of the SOD cage was also studied. The reduced catalyst retains its crystallinity was confirmed by XRD analysis for Pt@SOD reduced at temperatures 350, 450, 550, and 650\u00a0\u00b0C. TPR shows that the catalyst crystalline nature is retained until 650\u00a0\u00b0C of Pt@SOD, as shown in Fig. S2.Study the effect of the temperature on product yield and quality; the reactor temperature was varied from 380 to 440\u00a0\u00b0C. Other parameters were kept constant for the catalyst, liquid hourly space velocity 1\u00a0h\u22121, H2/feed ratio as 2200 (vol/vol), and pressure at 100\u00a0bar. Since hydrodeoxygenation of triglycerides to hydrocarbon is spontaneous and highly exothermic, catalyst activity was compared in terms of lighter, midrange, and heavier hydrocarbons. The complete conversion has been observed for both the catalyst in the given range of temperature variation. The effect of temperatures on the yield% for lighter components (<C9) in Fig. 6\n (a), mid-range (C9-C15 or kerosene range) in Fig. 6(b), heavier range (C16-C18, diesel range) in Fig. 6(c), and oligomers (>C18) in Fig. 6(d) over NiMo/SiO2-Al2O3 and Pt@SOD-NiMo-SiO2-Al2O3 catalyst has been shown. Cracking was observed to be 6\u201317 times higher for NiMo/SiO2-Al2O3 than that of Pt@SOD-NiMo-SiO2-Al2O3 catalyst. Light hydrocarbon (<C9) increased rapidly from 5 to 50% with an increase in temperature, from 380 to 440\u00a0\u00b0C on NiMo/SiO2-Al2O3 catalyst, while the rise in same range hydrocarbons was only from 1 to 6% for similar temperature range on Pt@SOD-NiMo-SiO2-Al2O3 catalyst. The higher increase in lighter hydrocarbons over NiMo/SiO2-Al2O3 catalyst was due to the catalyst's high acidity and increased acidic activity. The reason for the lower acidity of Pt encapsulated NiMo/SiO2-Al2O3 catalyst is that for the same weight of catalyst bed, Pt@SOD-NiMo-SiO2-Al2O3 has 33% lower SiO2-Al2O3 group present per grams of a catalyst than NiMo/SiO2-Al2O3 catalyst.\nFig. 6(b) shows that mid-range hydrocarbons (C9-C15) increased slowly for both the catalysts; a maximum was observed around 420\u00a0\u00b0C over NiMo/SiO2-Al2O3. Conventional NiMo/SiO2-Al2O3 catalyst showed a 2\u20133-time higher yield when compared to Pt@SOD-NiMo-SiO2-Al2O3 at different temperatures. Lower yield over Pt@SOD-NiMo-SiO2-Al2O3 may be attributed to lower acidity per gram of total catalyst in the case of Pt encapsulated catalyst.In Fig. 6(c), the diesel range hydrocarbons (C16-C18) yield was observed to be higher for Pt@SOD-NiMo-SiO2-Al2O3 catalyst compared to NiMo/SiO2-Al2O3 in the temperature range of 380\u2013400\u00a0\u00b0C. Variation of diesel yield with temperature was much significant over NiMo/SiO2-Al2O3 due to a higher acidic function, which shows higher cracking ability at higher temperatures. Diesel yield obtained over Pt@SOD-NiMo-SiO2-Al2O3 was 1.4\u20135 times higher than that of NiMo/SiO2-Al2O3 catalyst.A maximum yield (11%) of oligomeric hydrocarbons (>C18) (Fig. 6 (b)) was observed over NiMo/SiO2-Al2O3 catalyst at 420\u00a0\u00b0C; it decreased to <4% at a higher temperature 440\u00a0\u00b0C on NiMo-SiO2-Al2O3 catalyst, while over Pt@SOD-NiMo-SiO2-Al2O3 catalyst maximum yield of oligomeric hydrocarbons (>C18) was observed at higher temperatures 420\u2013440\u00a0\u00b0C. The lower yield of oligomers (6\u20137%), were observed over a new catalyst (Pt@SOD-NiMo-SiO2-Al2O3), compared to conventional NiMo/SiO2-Al2O3 catalyst. Lower yield may be attributed to higher hydrogenation activity in the presence of the Pt, which provides active hydrogenation for the reaction [7]. In the case of Pt encapsulated catalyst, high hydrogenation function would remove most of the unsaturation and hence the precursors for oligomer formation.From the studies on sulfided metal oxide catalysts during the hydroconversion of triglycerides (jatropha oil) [8\u201313], it has been reported that the isomerization activity of hydrocracking catalysts increases with an increase in the acidity of the catalyst [6]. In the case of Pt@SOD-NiMo-SiO2-Al2O3, which has lower acidity, it is expected that the presence of strong hydrogenation sites would hydrogenate the triglyceride chains [9], which eventually decreases the overall isomerization selectivity. In the temperature range of 380\u2013440\u00a0\u00b0C, the isomerization activity was 2\u201330 times higher for NiMo/SiO2-Al2O3 than that of Pt@SOD-NiMo-SiO2-Al2O3 catalyst (Fig. 7\n). Maximum isomerization was observed at the lowest temperature for NiMo Pt@SOD-/SiO2-Al2O3 catalyst while it was highest at maximum temperature (440\u00a0\u00b0C) for NiMo/SiO2-Al2O3 catalyst. An optimum acidity is required for the required isomerization activity and better catalyst life [11]. Since isomers have better freezing characteristics than their linear counterpart; the catalyst acidity improves the isomerization of the hydroprocessed product, which improves the freezing point of mid-range hydrocarbons. Though literature shows that acidity also increases coke formation during hydrocracking reactions, optimum catalyst acidity is essential to check cracking ability and coking reactions [39]. Increased hydrogenation function limits naphthenes and aromatics, as evident in the subsequent results.\nFig. 8\n shows the influence of temperature on naphthenes and aromatics distribution along with H2 consumption for C9-C15 range hydrocarbons on NiMo/SiO2-Al2O3 catalyst. Aromatics concentration varied between 0.2 and 3% in the temperature range of 380\u2013440\u00a0\u00b0C. The minimum aromatics (0%) was observed at 400\u00a0\u00b0C for NiMo/SiO2-Al2O3 catalyst. Paraffins (Fig. S3), varying from 94 to 78% from 380 to 440\u00a0\u00b0C. Similar to the aromatics trend, a minimum was observed for naphthenes at 400\u00a0\u00b0C. Naphthene content increases more rapidly (3\u201319%) than aromatics in the range of 400\u2013440\u00a0\u00b0C. Poly-aromatics was less than <1% in the entire temperature range on the NiMo/SiO2-Al2O3 catalyst. Polyaromatics observed was much lower for Pt@SOD-NiMo-SiO2-Al2O3 catalyst due to high hydrogenating functionality present due to the noble metal-based catalyst (Fig. S4).\nFig. 9\n shows the influence of temperature on naphthenes and aromatics distribution for C9-C15 range hydrocarbons along with H2 consumption on Pt@SOD-NiMo-SiO2-Al2O3 catalyst. Mono-aromatics was <0.6% in the entire range of temperature. Low aromatics were expected, as noble metals are more active for hydrogenation than sulfided Ni and Mo. In the mid-range product, 5\u201315 times reduction in aromatics and 3\u201315% reduction in naphthenes in the entire operating range of temperature was observed for Pt@SOD-NiMo-SiO2-Al2O3 catalyst. Since noble metal catalysts are more hydrogenating, as expected, there were 1\u20132 times excess hydrogen consumption compared to conventional NiMo-SiO2-Al2O3 catalysts. Noble metal catalysts have a high hydrogenating function. Hence they adsorb more hydrogen compared to sulfide catalysts. In the case of the encapsulated Pt@SOD-NiMo-SiO2-Al2O3 catalyst, the adsorbed hydrogen is expected to be transferred to nearby sites via a spillover mechanism [7]. TPR data shown in Table 1\n shows that the hydrogen adsorbed on Pt@SOD-NiMo-SiO2-Al2O3 catalyst is more significant than combined chemisorbed hydrogen on Pt@SOD and NiMo-SiO2-Al2O3 catalyst. The only possible excess hydrogen adsorbed could be explained by the spillover phenomenon. Maximum naphthene yield was observed at 440\u00a0\u00b0C for NiMo/SiO2-Al2O3 catalyst. The product's higher paraffinic yield was observed (Fig. S3) when the triglyceride was hydro-processed over NiMo-SiO2-Al2O3 catalyst than Pt encapsulated catalyst. In both the cases' aromatics and naphthene content, a minimum was observed at 400\u00a0\u00b0C temperature due to the kinetic and thermodynamic limitation of hydrogenation function at lower and higher temperatures [40]. In the case of Pt@SOD-NiMo-SiO2-Al2O3 catalyst, the naphthene and aromatic content were within the range of ASTM limits for aviation hydrocarbons obtained by hydroconversion of vegetable oil (ASTM D7566) [11].Juhwan et al. [7], studies hydrogen spillover mechanism on the metal oxides, nature of spillover species, migration mechanism, and theoretical catalytic functions. They studied that surface hydroxyl groups, especially Br\u00f8nsted acid sites, play a crucial role in hydrogen migration via spillover from active metal to nearby support. The role of hydroxyl groups in hydrogen spillover was confirmed experimentally by changing the hydroxyl concentration in the encapsulated zeolites; this resulted in a considerable change in the hydrogenation activity. The activation energy for each step of migration of H2 via spillover has also been calculated by DFT [7]. The H-spillover's overall activation energy on the zeolite surface was reported to be <125\u00a0kJ/mol by Choi et al. [8].Pt promotional effect in hydrogenation reactions, over Pt@SOD-NiMo-SiO2-Al2O3 catalyst, the dissociative chemisorption of H2 molecule over Pt3 cluster encapsulated inside the SOD cage, and the spillover of H from the Pt cluster to the SOD zeolite Si-O-Al site are studied. The energies of the physisorbed molecular H2, dissociated chemisorbed H2, Transition state (TS) for H spillover, and the final geometry after the H spillover are given in Table 2\n. Molecular H2 is physisorbed at the Pt3@SOD through bonding with two Pt atoms, as shown in Fig. 2(a). The Pt-H bond distances, Pt(1)-H(1) and Pt(1)-H(2), are measured to be \u223c1.85\u00a0\u00c5. The internal H-H bond distance is measured as 0.74\u00a0\u00c5, indicating molecular adsorption of H2. The molecular H2 is then dissociated over the Pt3 cluster, where the H atoms are adsorbed to different Pt atoms, as shown in Fig. 10\n(b). The internal H-H bonds are entirely broken, as observed from the H(1)-H(2) bond distance of 1.62\u00a0\u00c5. The Pt(1)-H(1), Pt(3)-H(1), Pt(2)-H(2), and Pt(3)-H(2) bond lengths are calculated to be 1.68\u00a0\u00c5, 1.78\u00a0\u00c5, 1.66\u00a0\u00c5 and 1.82\u00a0\u00c5, respectively. The dissociative chemisorption of H2 is an exothermic process with reaction energy \u221269\u00a0kJ/mol. The H2 dissociative chemisorption reaction is highly exothermic in nature has been experimentally seen in the high H2 chemisorption value of the Pt incorporated catalyst in Pt/NaA and Pt/SiO2 catalysts [7].For the catalyst to be able to hydrogenate the bulky naphtha molecule, the spillover of chemisorbed H atoms from the Pt cluster to the zeolite and from zeolite to the NiMo catalyst is indeed essential. Juhwan et al. [7] have extensively studied the H spillover over the zeolite-A and found the process to be fast under the reaction condition with an overall activation energy \u223c95\u00a0kJ/mol in the presence of surface \u2013OH group. However, the spillover from the Pt cluster to the zeolite Si-O-Al backbone has not been studied. As the H is strongly chemisorbed at the Pt cluster, the spillover of H from Pt to the SOD backbone is an important step and was analyzed here using the DFT method, as shown in Fig. 10(b, b' and c). The transition state obtained for the H spillover reaction is shown in Fig. 10b', whereas the final state after the spillover with one H is transferred to the zeolite Si-O-Al site is shown in Fig. 10c, forming a \u2013OH group at the sodalite cage. Both the Pt(2)-H(2) and Pt(3)-H(3) bond distance increases from 1.66\u00a0\u00c5 and 1.82\u00a0\u00c5 in the initial state (dissociated chemisorbed H2, Fig. 10(b)) to 2.11\u00a0\u00c5 and 2.78\u00a0\u00c5 in the transition state (Fig. 10(b')). At the same time, the O(z)-H(2) bond distance decreases from 3.91\u00a0\u00c5 in the initial state (Fig. 10(b)) to 1.35\u00a0\u00c5 in the transition state (Fig. 10(b')), indicating the transfer of H atom from the Pt3 cluster to the O atom of the SOD framework. The H is completely transferred to the zeolite O atom in the final state, as shown in Fig. 10(c), with O(z)-H(2) bond length is measured as 1.06\u00a0\u00c5. The complete table for the important bond lengths is given in the SI, Table S2. The activation barrier for H spillover from the Pt3 cluster to the SOD Si-O-Al site is calculated to be 111.5\u00a0kJ/mol, whereas the reaction energy for the H spillover process is endothermic by 95.8\u00a0kJ/mol. DFT results obtained in this study indicate that the activation barrier for H spillover from the Pt3 cluster to the zeolite is lower compared to the H- spillover over the zeolite surface (<125\u00a0kJ/mol) as reported by Choi et al. [8]. This suggests that for the Pt@SOD-NiMo(S)-SiO2-Al2O3 catalyst, the H2 molecules will preferably dissociate over the metallic Pt3 cluster and then spillover first to the zeolite Si-O-Al site and then to the outer surface NiMo(S) active site, where it will hydrogenate the hydrocarbon fragments.In Fig. 11\n, the Pt@SOD-NiMo-SiO2-Al2O3 catalyst showed stable activity for >900\u00a0h. C9-C18 yield remained >85% throughout the run. The catalysts were active and showed stable activity. The yield of aviation range hydrocarbon was 2\u20134 times in the entire run for conventional NiMo/SiO2-Al2O3 catalyst compared to Pt@SOD-NiMo-SiO2-Al2O3 catalyst. Even though NiMo/SiO2-Al2O3 showed better activity for aviation range hydrocarbons, but due to the presence of higher aromatics and naphthenes in this range of hydrocarbon, there is a need for a secondary step to reduce the contents to the recommended level (ASTM D7566 sets the limit for aromatics <0.5% and naphthene as <15% in sustainable aviation fuel obtained from non-petroleum sources originating from hydroprocessing of esters and fatty acids). Pt@SOD-NiMo-SiO2-Al2O3 catalyst showed better product quality (<0.6% aromatics, <5% naphthenes) but lower product yield and isomerization activity. It is expected with further optimization of catalyst, specially Pt@SOD percentage in final catalyst and acidity, a better yield with desired product quality can be obtained for hydroconversion of triglycerides.The state-of-the-art reported catalyst Pt@SOD-NiMo(S)-ZSM-5 had limitations in terms of catalyst life (Table 3\n; stable activity only for 350\u00a0h)4. Faster deactivation is correlated to Oligomeric (>C18) products, which tend to adhere to catalyst surface and contribute to rapid deactivation. The undesirable coke precursors (>C18 oligomers) are nearly 3 times more on Pt@SOD-NiMo(S)-ZSM-5 than the current catalyst. A much lower deactivation rate as evidenced by much longer stable activity (660\u00a0h) makes this catalyst superior to all the catalysts reported to date for this reaction. Higher acidity of ZSM-5 (0.95\u00a0mmol\u22121) and mixed micro-mesoporosity (pore sizes: 0.6\u00a0nm, 13\u00a0nm) compared to the current catalyst supported on mesoporous SiO2-Al2O3 (acidity: 0.77\u00a0mmol\u00a0g\u22121; pore size: 8.6\u00a0nm), makes the former more susceptible to deactivation during hydrocracking reaction.Sustainable aviation fuel (mid-range product) with reduced aromatics and naphthene was achieved in the temperature range of 380\u2013440\u00a0\u00b0C with the encapsulated noble metal-based catalyst combined with NiMo/SiO2-Al2O3 hydroprocessing catalyst. Pt@SOD reduced the aromatics by 5\u201315 times and naphthenes by 3\u201315 times when compared with conventional NiMo/SiO2-Al2O3 catalyst. However, the isomerization activity was observed to be much lower with the Pt@SOD-NiMo-SiO2-Al2O3 catalyst. Although better product quality was obtained with Pt@SOD-NiMo-SiO2-Al2O3, the mid-range hydrocarbon yield was almost half of the yield obtained on NiMo/SiO2-Al2O3 catalyst. With the optimized, Pt@SOD loading and acidity, a trade-off between naphthenes and aromatics percentage in the aviation range vs. aviation range hydrocarbon yield and isomerization activity may be achieved. Theoretical DFT studies suggest that encapsulated Pt inside sodalite cage would provide extra spillover hydrogen to nearby sites for jatropha conversion. Noble metal along with sulfided catalyst was used in a high sulfur condition (>1000\u00a0PPM sulfur), and the activity was stable (>500\u00a0h) for the hydroconversion reaction using jatropha oil as liquid feed along with hydrogen as gaseous feed. An optimized ratio of encapsulated noble metal and acidity would improve the product quality without affecting the yield.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors thank the Council of Scientific & Industrial Research-Indian Institute of Petroleum Dehradun for research funding. The author also acknowledges staff members of the biofuel division and analytical division for help and support.", "descript": "\n                  Encapsulation of noble metals inside the zeolite cage has attracted much attention in the area of catalysis due to unique properties, i.e., shape selectivity, higher activity, and product yield. In the combined work, encapsulated Platinum (Pt) inside the sodalite cage combined with sulfided nickel and molybdenum supported on silica-alumina (NiMo(S)/SiO2-Al2O3) is used to improve the catalyst hydrogenation function by providing spillover hydrogen to the nearby active sulfided NiMo sites. Encapsulation ensures the protection of the noble metal sites from poisoning due to the sulfur-containing compounds. They have a higher hydrodynamic diameter than that of the sodalite window cage. Experimental results showed platinum encapsulated sulfided nickel and molybdenum supported on silica-alumina (Pt@SOD-NiMo(S)-SiO2-Al2O3) had a substantial effect on the product composition, yield, and hydrogen consumption compared with the conventional NiMo(S)/SiO2-Al2O3 catalyst for the hydroconversion reaction of jatropha oil to produce carbon\u2013neutral green fuels. Sustainable aviation fuel with lower aromatics and naphthene content (<1%) was produced over the bi-functional Pt encapsulated Ni-Mo(S)/SiO2-Al2O3 catalysts in the reaction temperature range of 380\u2013420\u00a0\u00b0C. Aromatics and naphthalene obtained in the aviation range product were as much as 15 times lower than the conventional NiMo(S)/SiO2-Al2O3 catalyst. The Pt@SOD-NiMo(S)-SiO2-Al2O3 catalyst showed stable activity for >600\u00a0h of the run, whereas state-of-the-art reported catalyst Pt@SOD-NiMo(S)-ZSM-5 had limitations in terms of catalyst life (stable activity only for 350\u00a0h).\n               "}
{"full_text": "Data will be made available on request.Despite the foreseen transformation that will take place in the next few years, the reduction in oil consumption will not reach 60\u00a0% until 2050, with an 80\u00a0% reduction when the energy model is zero emissions to the atmosphere [1]. Therefore, it is necessary to continue processing different highly complex crude oils for years, generating high volumes of complex wastewater that must be treated before discharge [2]. The presence of certain organic and inorganic compounds derived from the production processes can cause inhibition or even toxic effects on conventional biological treatment of centralised refinery wastewater treatment plants (RWWTPs) [3,4].Among the different residual aqueous streams produced in refineries, the spent caustic streams are considered a hazardous waste due to their complex features [5]. Spent caustic solutions are derived from multiple sources due to the scrubbing of cracked gas in ethylene crackers or Merox units processing of liquefied petroleum gas, gasoline, kerosene or natural gas [6]. Spent caustic streams are classified into three types depending on their composition and the process in which they are generated: sulfidic, cresylic, and naphthenic caustic streams. Usually, refineries do not separate each type of spent caustic, and they are mixed for a standard treatment rather than implementing specific solutions for each one [7].The spent caustic streams contain a high COD, sulfides that produce very strong odors, aromatic compounds such as phenols, amines, calcium carbonate, sulphates and a very high alkaline pH, which makes their handling and treatment extremely difficult due to corrosion and precipitation problems in industrial facilities [8\u201311]. The amount of sulphates in spent caustic wastewater is not a problem for aerobic biological treatment in RWWTPs [12]. However, in the case of sulfides, there is an internal limit established in some wastewater treatment plants (WWTPs) of 25\u00a0ppm to avoid corrosion of concrete and steel, safety issues for operators and odours, among other problems [12]. The presence of amines in spent caustic streams is also very important. One of the amines most widely used in crude oil refining processes is methyldiethanolamine (MDEA) [13,14]. MDEA is a tertiary alkanolamine used as absorbent in natural gas sweetening process for removal of hydrogen sulfide (H2S) and carbon dioxide (CO2) [14]. Although MDEA is more selective towards H2S, this amine is also used currently in CO2 capture processes. This leads to an increase in the arrival of MDEA to WWTPs of the industries in which it is used [15,16]. Regarding toxicity, contradictory reports can be found in the literature. Even though MDEA does not seem to be toxic to aerobic treatment systems [17], certain problems have been identified due to its inhibitory effect on biological processes such as nitrification [18].Different technologies have been studied for the treatment of spent caustic wastewaters, such as thermally activated persulphate [19], biological treatment based on biomass adaptation to simulated amine concentration [17], ozonation with microelectrolysis [20] and cyclic thermal oxidation processes [21]. In these systems, it must be noted that the highest concentration of MDEA studied was 3,760\u00a0ppm from a simulated wastewater working at an acidic pH between 1 and 5 [20]. This concentration is considered in the range of accidentally maximum values found in the inlet streams of biological treatments in RWWTPs.Wet air oxidation (WAO) has proven to be an effective process for the treatment of wastewater with high organic matter content, allowing the total or partial degradation of compounds that are toxic or refractory to biological processes [22\u201325]. The elimination of organic compounds identified in the wastewater can be practically total, with COD reductions of more than 70\u00a0% in most cases. The removal of MDEA in a real wastewater with higher concentration of the contaminant (in the range of g/L) has been studied in two previous works by a WAO process [26] and a catalytic process (known as catalytic wet air oxidation, CWAO) using a commercial activated carbon as catalyst [27]. In the case of CWAO, most works have been performed using synthetic wastewaters with phenols and derivatives as dominant model pollutants [28]. The operation conditions, type of catalyst and removal efficiencies of all these studies are summarized in Table 1_SM. The catalysts used in this process are normally classified as: noble metal catalysts, non-noble metal catalysts and metal-free carbon materials [29]. Synthesised metal-free carbon materials can promote catalytic wet air oxidation processes for wastewater treatment as an alternative to noble metals and rare earth oxide catalysts. The absence of metals in the catalyst based on carbonaceous materials avoids possible leaching and subsequent treatment needs compared to processes using metal-containing catalysts [28,30].Petroleum coke (petcoke) is a black-colored solid composed primarily of carbon, generated as a product of the coking process oil refineries or in other heavy hydrocarbon cracking processes at high temperatures and rotational speeds [31]. It contains limited amounts of sulfur, metals, and non-volatile inorganic compounds. The most extensive use of petcoke is as a source of energy or carbon in different industrial applications [32]. Due to the majority use of coke as fuel in industries traditionally considered as a source of pollutant gas emissions, global demand of petcoke is expected to decline and, therefore, finding novel applications is key [33]. To address this danger, potential applications are being studied in which this co-product can be used due to its special characteristics. In this study, the preparation of activated carbon materials in catalytic wet air oxidation (CWAO) for the treatment of wastewater of the own refinery fits into the circularity guidelines that must govern technological processes in the near future. The use of materials prepared from the own refinery petcoke as possible catalysts for the treatment of wastewater has not been already reported despite the large number of studies described in literature (Table 1_SM).In this work, it is proposed a different approach for the onsite treatment of the spent caustic streams containing MDEA in high concentration coming from the natural gas sweetening process, in which accidentally discharges of MDEA could reach concentrations up to 2.6\u00a0g/L. Currently, this highly MDEA-containing stream is mixed with wastewater streams coming from other operation units, diluting the pollutant concentrations upstream of RWWTP. However, mixing is often not sufficient to reduce the concentration to values that can be assumed by a conventional biological process in RWWTPs. The aim of the study is to increase the biodegradability of this wastewater using WAO and CWAO processes. The onsite treatment of highly MDEA-containing wastewater streams rather than at the end-of-pipe after dilution would be beneficial from the point of view of reducing the capital expenditures, as lower wastewater flows would require smaller equipment sizes. Additionally, the reuse of a low value-added material such as refinery petcoke as a catalyst in the CWAO process, which is likely to become a waste product owing to increased legislation on emissions, could reduce the operational costs due to the feasibility of using milder operating conditions of pressure and temperature in the oxidation process.Linde Gas Espa\u00f1a, S.A.U supplied air pressurised bottles for WAO and CWAO experiments. Potassium hydroxide and hydrochloric acid (37\u00a0% v/v) used for the petcoke activation, were purchased from Labkem and Sigma Aldrich, respectively.The spent caustic wastewater used in this study comes from an amine unit located in a petroleum refinery in Spain. Samples were collected and immediately stored at 4\u00a0\u00b0C to avoid variations in composition. The so-called fuel grade green petcoke comes from a coker unit of the same petroleum refinery. The petcoke sample was stored in a closed container before use.WAO and CWAO experiments were performed in a 500\u00a0mL capacity T316 stainless steel autoclave reactor, resistant to high pressures and temperatures: model 4575A manufactured by the Parr Instrument Company, USA. The reactor was equipped with an electrically heated jacket, a turbine stirrer, and a variable speed magnetic drive. The temperature using a thermocouple immersed in the liquid phase and the stirring rate were controlled using a Parr 4842 controller and the pressure by the gas inlet and a gas release valve located on the top of the reaction vessel. The liquid samples were taken through a dip tube immersed in the reaction mixture. A schematic diagram of the experimental setup is shown in Fig. 1\n.Typically, 250\u00a0mL of the spent caustic refinery wastewater is placed in the reactor. The pH of the refinery wastewater (9.6\u00a0\u00b1\u00a00.1) was not modified and adjusted prior or during the treatment. Initially, nitrogen gas flow was passed through the head space of reaction vessel to ensure inert conditions, and continuous stirring was fixed at 400\u00a0rpm. Then, the air was supplied up to 15\u00a0bar to maintain the wastewater in a liquid phase. Finally, the reactor was heated to the operating temperature and then the air pressure was increased to the selected value. The temperature and air pressure of WAO and CWAO experiments varied between 150 to 250\u00a0\u00b0C and 10 to 90\u00a0bar, respectively [5]. The same procedure is used for CWAO runs but adding 1 g/L of a carbon-based catalyst prepared from petroleum coke in the initial loading of the wastewater into the reaction vessel. The reproducibility and the accuracy of the performed WAO and CWAO runs were evaluated periodically between tests. Methyldiethanolamine (MDEA), Total Organic Carbon (TOC), Chemical Oxygen Demand (COD), sulfides (S2\u2212) and pH were periodically monitored for samples withdrawn along the reaction time for 60\u00a0min. Prior to analyses, the samples were filtered through a 0.7\u00a0\u03bcm glass fiber filter.TOC (Total Organic Carbon) was determined in a combustion/non-dispersive infrared gas analyser model TOC-V Shimadzu. The pH was monitored using a GLP-22 digital pH meter (HACH LANGE SPAIN, S.L.U). Chemical Oxygen Demand (COD), Total Solids (TS), and Volatile Total Solids (VTS) were measured following APHA-AWWA Standard Methods 5220.D, 2540.B and 2540.E respectively. Total Kjeldahl Nitrogen (TKN) was measured using a Vapodest 450 (Gerhardt, Analytical Systems) for the digestion of the samples, following APHA-AWWA Standard Method 4500-Norg C. Sulfides (S2\u2212), nitrates (NO3\n\u2212), and ammonium (NH4\n+) concentration were determined using a Smartchem 140 (AMS Alliance), following APHA-AWWA Standard Methods [34]. Methyldiethanolamine (MDEA) concentration was monitored by gas chromatography (GC) in a Varian 450-GC equipped with a column HP-PONA (High-resolution Performance column, 50\u00a0m\u00a0\u00d7\u00a00.20\u00a0mm, 0.50\u00a0\u03bcm, for the detection of Paraffins, Olefins, Naphthenes and Aromatics) and a Flame Ionization Detector (FID). The injection and detector temperatures were set at 300 and 250\u00a0\u00b0C, respectively. The oven temperature was maintained at 90\u00a0\u00b0C for 7\u00a0min, raised to 250\u00a0\u00b0C at a 50\u00a0\u00b0C/min rate, and finally held for 5\u00a0min. The degradation by-products from MDEA oxidation were analysed by direct aqueous-injection gas-chromatography coupled to a mass detector (320 GC\u2013MS) using a Bruker column Stalbiwax-MS (30\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm, 0.25\u00a0\u03bcm). This specific column for aqueous samples was initially maintained at 50\u00a0\u00b0C for 3\u00a0min, then heated to 180\u00a0\u00b0C at 12\u00a0\u00b0C/min, and finally maintained for 5\u00a0min at 250\u00a0\u00b0C (7\u00a0\u00b0C/min). The injector was held at 320\u00a0\u00b0C, and He (1\u00a0mL/min) was used as the carrier gas. Biodegradability tests were performed to evaluate the effect of refinery wastewater on acclimated and controlled biomass cultures according to the literature experimental procedure included in the supplementary information. For this purpose, measurements of oxygen consumption (OC) and oxygen uptake rates (OUR) in different pulses of treated wastewater and sodium acetate were made after months of acclimatisation to a sodium acetate-rich feed used as a readily biodegradable substrate of the biomass taken from a local wastewater treatment plant [35].\nTable 1\n shows the physicochemical characterisation of the spent caustic wastewater. The concentration of MDEA was about 2.5\u00a0g/L. The TOC and COD of the corresponding MDEA concentration are almost the 85\u00a0% of the total TOC and COD of the wastewater. In the case of nitrogen MDEA represents approximately 97\u00a0% of total in the water, mostly in the form of organic nitrogen (MDEA provides more than 98\u00a0% of the organic nitrogen). Thus, MDEA is the most abundant compound, with the presence of sulfides (750\u00a0ppm) also detected in significant concentrations. The high MDEA concentration is due to an unusual and extreme situation of accidental discharge from ethylene crackers or Merox units to the centralized refinery's WWTP. Additionally, the wastewater showed a low concentration of metals, a very alkaline pH due to the amine content and 36\u00a0g/L of TSS. These characteristics allow classify this process stream as a hazardous waste and make it difficult to handle and treat [8,11,36\u201339].The fuel-grade green petcoke (PC) collected from a coker unit had the typical characteristics (included in the supplementary information) of a refinery coke [32], with a low specific surface area (ca. 10\u00a0m2/g), a CHNS distribution of 81.7\u00a0%, 3.7\u00a0%, 1.4\u00a0%, and 5.3\u00a0%, respectively, and heavy metals content of approximately 1,700\u00a0ppm of vanadium and 400\u00a0ppm of nickel (Table 2_SM). Thermogravimetric analyses evidenced, a weight loss lower than 14\u00a0% at 1000\u00a0\u00b0C in an inert nitrogen atmosphere and almost negligible up to 500\u00a0\u00b0C under air atmosphere (Fig. 1_SM). Table 2\n displays physicochemical characterization of activated carbon materials prepared using different KOH:Petcoke ratios [30]. The chemical activation produced an increase in the specific surface area of the petcoke, due to an increase in the porosity of the material [40]. The increase of KOH:Petcoke ratio from 2 to 4 enhanced the BET specific surface area from 1,043 to 3,459\u00a0m2/g. These areas are similar or even higher than those reported for other petroleum cokes chemically activated under different conditions [30,41,42]. Nitrogen adsorption\u2013desorption isotherms (Fig. 2_SM) depict typical profiles of microporous materials according to the IUPAC. According to the adsorption isotherms, these materials fill the micropores in a continuous process at low relative pressures (P/P0\u00a0<\u00a00.015) [43]. According to the Howarth-Kawzoe model, the pore size distribution was centred in the microporous range with values of less than 2\u00a0nm (Fig. 3_SM). The total pore volume of the activated petcoke materials goes from 0.48\u00a0cm3/g to 1.63\u00a0cm3/g, pore volumes that are up to twice as large compared to other carbonaceous materials obtained by petroleum coke activation [30]. The CHNS elemental analysis showed a significant reduction of N and S contents after chemical activation. Likewise, the content of V and Ni also decreased due to acid washing (HCl) for the removal of KOH. The increase of KOH:PC ratio, calcination temperature, heating ramp and nitrogen flow rate showed a positive effect on the development of more microporous materials with higher total pore volume. A KOH:PC ratio of 4, using a calcination temperature of 790\u00a0\u00b0C (heating temperature of 20\u00a0\u00b0C/min and a nitrogen flow rate of 100\u00a0mL/min) provide the activated carbon material with maximum specific surface area and total pore volume (3,459\u00a0m2/g and 1.63\u00a0cm3/g, respectively) and pore size distribution centred at 1.8\u00a0nm. X-ray diffraction (Fig. 4_SM\n) shows the difference between the structures of the initial petcoke and synthesised carbonaceous materials consisting of graphite microcrystallites characteristic of activated carbon-type materials with a high specific surface area [42]. SEM-EDX analyses of the raw and activated petcoke materials displayed particles of a non-homogeneous morphology and size with a much more porous surface and higher O/C ratios as the KOH:PC increased by the effect of the activation agent (Fig. 5_SM).Wet air oxidation (WAO) has proven to be an effective process for the treatment of wastewater with high organic matter content, allowing the total or partial degradation of compounds that are toxic or refractory to biological processes [44]. The WAO process at different reaction temperatures and air pressure was studied for the concentrated MDEA wastewater stream. Additionally, solid catalysts synthesized from green fuel-grade petcoke in the form of microporous activated carbonaceous materials were also tested. Figs. 2 and 3\n\n show the removal of MDEA, COD, TOC, and sulfides of highly MDEA concentrated spent caustic wastewater for WAO and CWAO experiments, respectively. Preliminary blank experiment at 250\u00a0\u00b0C and 90\u00a0bar under inert nitrogen atmosphere showed no modification of the initial characteristics of the wastewater. This indicates a negligible thermal degradation and the need of dissolved oxygen in the wastewater for the oxidation of pollutants.Initially, the highly concentrated MDEA stream was treated at 150\u00a0\u00b0C under different air pressures ranging from 10 to 90\u00a0bar, which can dissolve oxygen amounts from 0.05 to 0.4\u00a0mL O2 per gram of water [45,46]. As it can be seen in Fig. 2, removal of sulfide reached 90\u00a0% in all WAO experiments. This fact is attributed to the high reactivity of sulfides as compared to the organic matter contained in the water [8]. This is especially noteworthy for the WAO experiments at the lowest temperature (150\u00a0\u00b0C), where sulfide removal was above 90\u00a0% even with low air operation pressure (10\u00a0bar). The increase of air pressure from 10 to 90\u00a0bar increased the MDEA elimination but had a slight effect on TOC and COD removals.Sulfides are dissolved in the wastewater due to the basicity of the waste stream (pH\u00a0=\u00a09.6). If the pH approaches neutrality (pH\u00a0=\u00a07), the sulfides can abandon solution as acid gas, which must be treated before emission into the atmosphere [8]. Sulfides may exist in three different forms, H2S, HS\u2212 and S2\u2212 depending on the pH of the medium. At a pH of 7, sulfides are present in the form of H2S gas and HS\u2212. If the pH decreases further, most of the sulfides would be in the form of H2S gas. On the contrary, when the pH increases the majority of the sulfides are in the form of dissolved hydrosulfide (HS\u2212) and sulfide (S2\u2212) [47]. The treatment of these pollutants has been studied by additions of iron salts for their precipitation [48], biological oxidation [49], neutralization combined with conventional or advanced oxidations such as the Fenton process [47], or by physicochemical separation systems such as electro-coagulation [50]. These processes generate secondary sludge effluents that require a further waste management, increasing treatment costs, consumption of chemicals, etc. Besides, in some cases, the efficiency of sulfide removal is not high enough to make the water amenable for conventional biological treatments. The WAO process overcomes these disadvantages, enabling sulfide removal under mild operating conditions without generating secondary waste effluents [51]. It must be pointed out that for WAO experiments at 150\u00a0\u00b0C and different air pressures, the pH did not decrease from the initial value of 9.6 to<8.5, a value at which most of the sulfides are dissolved in water in the form of S2\u2212\n[47]. Thus, the elimination of these sulphur compounds by stripping in the form of acid gas is discarded. The elimination is through the oxidation to sulphates (SO4\n2\u2212) by the oxygen dissolved in the reaction medium with a stoichiometric consumption of 2\u00a0g O2/g S2\u2212\n[50].Likewise, at 150\u00a0\u00b0C, the organic matter removal in terms of the total organic carbon (TOC) was 6\u00a0%, 9\u00a0% and 14\u00a0% for operating air pressures of 10, 50 and 90\u00a0bar, respectively. The reduction in terms of COD was significantly higher (18\u00a0%, 27\u00a0% and 28\u00a0%) due to the simultaneous removal of sulfides previously mentioned, wich have a contribution to COD. Despite the moderate removal of TOC and COD, the most abundant pollutant identified in the water, MDEA, was eliminated by 10\u00a0%, 35\u00a0% and 62\u00a0%. Therefore, at the temperature of 150\u00a0\u00b0C MDEA cannot be removed completely resulting in concentrations in the treated effluent that exceed 1,000\u00a0ppm even for 90\u00a0bar of air operation pressure, which is still too high for the downstream biological treatment, despite the dilution of this effluent by mixing with the rest of the wastewater streams coming from different refinery processes.Therefore, in order to enhance the performance and taking into account the effect of the increase of pressure and temperature on the overall economy of the WAO process due to the rise in energy consumption, the reaction temperature was raised to 200\u00a0\u00b0C. The increase of 50\u00a0\u00b0C produced an enhanced in the amount of oxygen dissolved in the water (ca. 1.5 times for the same working air pressures [45,46]). The increase of temperature hardly enhanced the TOC removal, maintaining similar efficiencies to those obtained at 150\u00a0\u00b0C, whereas the COD removal increased, with reductions of 24\u00a0%, 32\u00a0% and 38\u00a0% for the air pressures of 10, 50 and 90\u00a0bar, respectively. In the case of MDEA removal, more than 80\u00a0% and 90\u00a0% were achieved at air pressures of 50 and 90\u00a0bar, leading to final concentrations of 497 and 177\u00a0ppm, respectively. In these cases, the dilution of this stream with the rest of the wastewater streams of the refinery would allow reaching MDEA concentration at which very limited effects on the biological process of the wastewater treatment plant are expected. The WAO process is therefore effective for the treatment of wastewater with high MDEA concentration at typical operation conditions of temperature (200\u00a0\u00b0C) and air operation pressure (up to 90\u00a0bar). The main pollutants of the wastewater, MDEA and sulfides, are removed without the need to achieve a high removal of organic matter.At this point, the increase of temperature to 250\u00a0\u00b0C was also assessed for further reduction of organic matter. The effect of temperature on the amount of dissolved oxygen in the reaction medium is increased by 1.7 times compared to the same working pressures and 200\u00a0\u00b0C [45,46]. At 250\u00a0\u00b0C and 50 or 90\u00a0bar air pressure, the elimination of MDEA was almost complete (99\u00a0%) and over 95\u00a0% for sulfides. In this case, the mineralization of organic carbon increases up to 40\u00a0%, with COD reductions of more than 65\u00a0%. A decrease in final pH of the wastewater occurs from the natural initial value of 9.6 to 4.8 because of the oxidation process. This acidic character can cause corrosion problems in the facilities needed for treatment [52] even having been established as a worldwide problem in sewers due to corrosion caused by H2S in wastewater [53]. This decrease in pH is associated to the organic matter oxidation and conversion of sulfides to sulphates and sulfuric acid by the high pressure WAO process [54]. This acid partially consumes the alkalinity of the spent caustic from the wastewater. In addition, the organic matter contained in the water is oxidized to low molecular weight carboxylic acids that contribute to the acidity of the water. Among the distinct low molecular weight carboxylic acids, WAO process is characterized by promoting the formation of acetic acid as main oxidation by-product [44]. Acetic acid is quite difficult to further oxidize to carbon dioxide and water by WAO process, being necessary to use very extreme operating conditions or to work with catalysts that facilitate the process [55]. In this work, acetic acid concentrations higher than 400\u00a0ppm were detected in the wastewater after WAO treatment. This formation of carboxylic acids together with the oxidation of sulfides to sulphates, results in the consumption of alkalinity and consequent decrease in the pH of the water.Initially, operation conditions of 250\u00a0\u00b0C and 50\u00a0bar of air pressure were used for testing the catalytic performance of activated carbonaceous materials. Air pressure at 50\u00a0bar was used as the increase of air pressure above this value hardly improved the COD reduction and TOC mineralization for WAO experiments. In these conditions, the catalytic activity of commercial activated carbon (CWAOAC) with a specific surface area of 1,288\u00a0m2/g and activated carbon materials from petcoke (CWAOPC) with different specific surface areas (1,013\u00a0m2/g, 2,196\u00a0m2/g and 3,221\u00a0m2/g) was evaluated to compare its effect on oxidation yields (Fig. 3a and b). In general, non-significant differences were observed for the four carbonaceous materials and the non-catalytic WAO process at these operating conditions and 60\u00a0min of reaction for all the experiments. The commercial activated carbon achieves near-complete removal of MDEA and sulfides. A slight increase in the mineralization degree of the organic matter and reduction of COD was observed for the CWAOAC compared to the WAO. The carbonaceous material synthesized from petcoke (CWAOPC_2,196\u00a0m2/g) shows similar results to the commercial activated carbon, with amine and sulfide degradations of 99\u00a0% and 95\u00a0%, and TOC and COD reductions of 56\u00a0% and 76\u00a0%, respectively. These performances are maintained for the carbonaceous material with a lower specific surface area (CWAOPC_1,013\u00a0m2/g) and the TOC and COD removals were slightly improved using the carbonaceous material of the highest specific surface area (CWAOPC_3,221\u00a0m2/g) and higher O/C ratio according to the SEM-EDX analyses (Fig. 5_SM).In order to determine a higher influence of the catalyst compared to the non-catalytic WAO process, the air pressure and temperature were decreased to milder operating conditions. This study was carried out with the CWAOPC_2,196\u00a0m2/g carbonaceous material prepared from petcoke as the most available catalyst of the three CWAOPC materials and similar catalytic performance to commercial activated carbon (CWAOAC).The influence of temperature at constant air pressure of 50\u00a0bar and air pressure at constant temperature of 150\u00a0\u00b0C for the catalyst CWAOPC of 2,196\u00a0m2/g is shown in Fig. 3c and 3d, respectively. The performance of CWAO at 50\u00a0bar when the temperature decreased from 250\u00a0\u00b0C to 150\u00a0\u00b0C evidenced a more remarkable difference between the results of CWAO and WAO experiments. The elimination of MDEA exceeds 90\u00a0% for the two upper temperatures of 200\u00a0\u00b0C and 250\u00a0\u00b0C, reaching a MDEA removal of 77\u00a0% at 150\u00a0\u00b0C (only 35\u00a0% for the WAO treatment). At lowest temperature, the catalyst proves a significant effect, increasing the removal of MDEA, the main pollutant in the actual refinery wastewater stream, by more than twofold. Concerning the removal of sulfides and the reduction in TOC and COD, a slight improvement was also observed in comparison to the results of the WAO process.The effect of the air pressure for the CWAO in the range of 10 and 90\u00a0bar was also seen at the lowest temperature of 150\u00a0\u00b0C. At 10\u00a0bar, 61\u00a0% of the initial high concentration of MDEA in the wastewater (2,521\u00a0ppm) is removed, with the total oxidation of sulfides to sulphates. This is a 50\u00a0% improvement in amine degradation referred to the oxidation without catalyst. Increasing the pressure to the upper studied limit of 90\u00a0bar yields an MDEA removal of more than 80\u00a0% compared to 62\u00a0% obtained with the WAO process. Thus, the carbonaceous catalyst CWAOPC_2,196\u00a0m2/g improves the performance of the oxidation process. The functional groups of the catalyst promote the reduction of the oxygen in the medium, generating radicals of greater oxidizing power that can oxidize the compounds present in the wastewater [28]. The oxidation under these mild operation conditions with the catalyst results in a higher oxidation of the MDEA and organic compounds with a slight enhancement of COD and TOC removals.Therefore, the catalytic material with high porosity and specific surface area prepared from petcoke proves to be an effective catalyst for the WAO process, improving the yields achieved under mild operating conditions. The CWAO process, which achieves high removals of MDEA and sulfides from the water under mild operating conditions, will probably improve the biodegradability of the wastewater. This is also an important factor in terms of the performance of the centralized biological treatment system of the refinery's wastewater treatment plant. Despite this, the actual biodegradability of the oxidation products generated in the WAO and CWAO processes needs to be studied. This would allow establishing the real effect of the effluent generated in the process on the biological treatment system, as well as the optimal operating conditions at which the pretreatment of the stream with high MDEA should be carried out.Most of the studies of WAO for wastewater treatment are mainly focused on quantifying the degradation of the organic matter in terms of COD and TOC reductions. However, it is essential to identify oxidation by-products or intermediates from nitrogen-containing organic compounds to evaluate their potential toxicity and/or refractory behavior for subsequent biological treatment. The WAO of nitrogen-containing compounds can produce different products including ammonium, nitrate, nitrite, nitrous oxide and nitrogen gas depending on the pollutant and reaction conditions [21\u201323]. When the main nitrogen input is an amine-containing compound, ammonium is mainly produced as stable end-product at harsh oxidation conditions [51].\nFig. 4 shows the concentration of nitrogen-containing compounds in the wastewater such as the MDEA itself, other nitrogen organic products and inorganic nitrogen as ammonium and nitrates/nitrites, for the WAO and CWAO treated waters after 60\u00a0min under different operation conditions. The initial MDEA concentration (2,521\u00a0ppm) is about 300\u00a0ppm in terms of nitrogen content and contributes to ca. 95\u00a0% of the total nitrogen of the wastewater. WAO experiments at 150\u00a0\u00b0C evidenced a low formation of inorganic nitrogen compounds. MDEA and nitrogen-containing organic by-products were the most abundant compounds. The contribution of MDEA was up to ca. 50\u00a0% for the highest air pressure. The increase of temperature at 200\u00a0\u00b0C resulted in the elimination of MDEA of 27\u00a0%, 83\u00a0% and 95\u00a0% for the 10, 50 and 90\u00a0bar, respectively. The increase of air pressure led to 10\u00a0% of inorganic nitrogen such as NO3\n\u2212/NO2\n\u2212. At the highest temperature (250\u00a0\u00b0C), the increase of air pressure enhanced the inorganic nitrogen products up to 40\u00a0%, out of which 30\u00a0% corresponding to ammonium.The use of the carbonaceous material synthesized from petcoke as a catalyst (CWAOPC_2,196\u00a0m2/g) shows a higher contribution of non-identified organic nitrogen compounds at low temperature (150\u00a0\u00b0C) and different air pressures in comparison to the analogous WAO experiments, and the oxidation to NO3\n\u2212/NO2\n\u2212 and ammonium was hardly detected. As the temperature was increased up to 250\u00a0\u00b0C at 50\u00a0bar, the oxidation of the nitrogen organic compounds led to ammonium contents higher than those observed in WAO experiments at the analogous operation conditions, and low presence of NO3\n\u2212/NO2\n\u2212. Thus, the increase in temperature and air pressure using the carbonaceous activated carbon as catalyst for CWAO, decreased the MDEA contribution to the total nitrogen content promoting the generation of ammonium as main inorganic nitrogen by-product. Ammonium obtained as a product of MDEA oxidation has a certain refractory character to the WAO and CWAO processes, resulting in low elimination in the form of nitrogen gas [23] as also attested from the negligible decrease of total nitrogen. Nevertheless, ammonium can be consumed as a nutrient by aerobic biological treatment systems [56].As shown in Fig. 4\n, the contribution of non-identified organic nitrogen compounds stemming from the partial oxidation of MDEA (termed as \u201cother organic\u201d in the graph) to total nitrogen varies depending on the operation conditions of WAO and CWAO. The oxidation of MDEA follows a very complex mechanism with a large number of by-products whose formation has not been explained in many cases [13,20,22,23,57\u201359]. As other amines, MDEA is oxidized to organic acids and glycine as main by-products, but unspecified formyl-amides have also been reported as secondary by-products [57]. Formation of diethylamine (DEA), N-methylamine (MMA), or ethanolamine (MEA) may occur concurrently with a methyl group transfer from MDEA or by direct oxidation. These compounds may react with other compounds formed in the oxidation process of MDEA, leading to a chain of reactions that are difficult to define [57]. In the CWAO, the degradation of MDEA using metal-free carbon catalysts has been demonstrated to be strongly dependent on the chemical surface groups of the materials, which play a key role in the production of active oxidizing species. Thus, it is well recognized that the adsorption of O2 over the carboxyl groups of activated carbon surface produces its dissociation to form O2\n\u2212 species [60]. Then, hydroxyl radicals (OH) are generated by electron transfer of O2\n\u2212or attractingH+ of the carboxyl groups. In addition, the basic groups of the activated carbon surface have also an important role attracting small molecules of carboxylic acids, which can react with OH to generate CO2 and H2O. In the case of MDEA, the hydroxyl radicals can also attack the CN bond of the MDEA to generate other intermediate products which will be oxidized to carboxylic acids, ammonium, CO2 and H2O [27]. In this study, the analysis of the oxidation products of the treated wastewater under the different operating conditions has resulted in the detection of acetic acid, ammonium, and other by-products of the oxidation of MDEA such as dimethylamine, tetrazole-1,5-diamine, 2-amino-1-propanol, and alanine among others.Acetic acid appeared in all the oxidation reactions (WAO and CWAO), being its contribution over the total organic carbon more important as temperature and air pressure is increased or when catalyst was used (Fig. 6_SM of SI). The rest of the identified compounds varied without a clear trend. Fig. 5\n shows the potential oxidation reactions of MDEA based on the products identified for the different operating conditions of WAO and CWAO. In these reactions, water and carbon dioxide are also produced because of the mineralization of organic matter as deduced from the reduction of the TOC and COD of the wastewater.\n\n(1)\n2C5H13NO2\u00a0+\u00a08O2\u00a0\u2192\u00a05CO2\u00a0+\u00a05H2O\u00a0+\u00a02NH3\u00a0+\u00a02,5C2H4O2\n\n\n\n\n\n(2)\nC5H13NO2\u00a0+\u00a01,5O2\u00a0\u2192\u00a0CO2\u00a0+\u00a0H2O\u00a0+\u00a0C2H4O2\u00a0+\u00a0C2H7N\n\n\n\n\n(3)\n2C5H13NO2\u00a0+\u00a07O2\u00a0\u2192\u00a05CO2\u00a0+\u00a05H2O\u00a0+\u00a0NH3\u00a0+\u00a0C2H4O2\u00a0+\u00a0C3H9NO\n\n\n\n\n(4)\n7C5H13NO2\u00a0+\u00a046O2\u00a0\u2192\u00a032CO2\u00a0+\u00a040H2O\u00a0+\u00a0NH3\u00a0+\u00a0C2H4O2\u00a0+\u00a0CH4N6\n\n\n\n\n\n(5)\n2C5H13NO2\u00a0+\u00a08O2\u00a0\u2192\u00a05CO2\u00a0+\u00a06H2O\u00a0+\u00a0NH3\u00a0+\u00a0C2H4O2\u00a0+\u00a0C3H7NO2\n\n\n\nThe complex nature of the petrochemical wastewater and the variety of generated by-products makes necessary to analyze the actual biodegradability of the effluents after the WAO or CWAO treatment. The generation of acetic acid and ammonium as main products of the oxidation process suggests a potentially increased biodegradability of the effluent, which should be easily treated in the conventional biological treatment system of the refinery's water treatment plant [3,56].To evaluate the rapid biodegradability of the treated effluents obtained in different oxidation conditions of WAO and CWAO runs, respirometric tests were performed with an activated sludge culture acclimatised for months to sodium acetate as biodegradable substrate. The toxicity and inhibition effects of the treated effluents on the readily biodegradable substrate were also assessed using the classical respirometric bioassays [35], but no conclusive results could be obtained. The spent caustic wastewater, containing 2,521\u00a0ppm of MDEA, high COD, TOC and sulfides, evidenced a low biodegradability of ca. 4\u00a0% compared to the sodium acetate solution used as a readily biodegradable substrate.\nFig. 6\na depicts the compositional percentage of MDEA and acetic acid (both relative to TOC), and ammonium relative to Total Nitrogen (TN) in relation to the biodegradability of the treated effluents. Fig. 6b shows the results of biodegradability of treated wastewaters at 150, 200 and 250\u00a0\u00b0C using 50 and 90\u00a0bar air pressure for WAO and the same temperatures using 50\u00a0bar for CWAO (activated carbon material prepared from petcoke with specific surface area of 2,196\u00a0m2/g).The samples of WAO treatment showed a remarkable increase of biodegradability at 250\u00a0\u00b0C (ca. 50\u00a0%) as compared to the results at 200\u00a0\u00b0C (ca. 20\u00a0%), regardless the applied air pressure (50 or 90\u00a0bar). This fact is attributed to the absence of MDEA and higher contribution of acetic acid in the remaining TOC of the sample for 250\u00a0\u00b0C. Moreover, the percentage of ammonium respect to the TN was also much higher at 250\u00a0\u00b0C, which is decreasing the amount of other nitrogenated organic by-products, presumably with lower biodegradability. At 150\u00a0\u00b0C, the presence of MDEA in significant amounts and presence of less biodegradable nitrogenated organic compounds, leads to a dramatical decrease of biodegradability (<5 %). In contrast, the CWAO significantly increases the biodegradability at the three studied temperatures using 50\u00a0bar of air pressure. At 250\u00a0\u00b0C, the high content of acetic acid and ammonium are probably boosting the increase of biodegradability up to almost 70\u00a0%. At lower temperatures, 200 and 150\u00a0\u00b0C, the biodegradability decreased up to ca. 30\u00a0% and 25\u00a0%, respectively, due to remaining amounts of MDEA and lower contents of ammonium, but these values are much higher than those obtained in WAO at the same operation conditions.The fact that the biodegradability cannot reach values close to 100\u00a0% is attributed to the presence of nitrogenated organic compounds produced by the oxidation of MDEA, which may require longer degradation times than those established in the respirometric tests. It should be noted that these tests show the biodegradability of the effluents generated in short periods of time, for an activated sludge acclimatized to a model substrate [35]. The performance of an activated sludge process in refinery's wastewater treatment plant should be better, leading to a higher biodegradability of the effluents of the WAO and CWAO processes. These treatment plants have activated sludge systems acclimatized to less biodegradable compounds, with a higher concentration of active biomass than that used in these respirometric tests and operating at higher hydraulic residence times [3,4]. In addition to the improvement expected from operation at the refinery's wastewater treatment plant, biodegradability achieved by WAO and CWAO treatment of the spent caustic stream at the outlet of the unit, where the pollutant load is the highest, makes it viable for the effluent to reach the biological treatment system after dilution with other wastewater streams generated in the refinery. This would further reduce the pollutant load reaching the treatment system making treatment even easier.The WAO process is an effective treatment of the refinery spent caustic stream with high MDEA contamination from amine absorption units in the purification of refinery gas streams. A deep study of the operating conditions has been performed as MDEA is a tertiary amine that can cause significant problems in the centralized biological system of the refinery's wastewater treatment plant. The experiments were performed at the original pH of the wastewater despite its high alkalinity, which assure that sulfides are removed by oxidation to sulphates and using air instead of pure oxygen as source of oxidising agent. The increase of temperature has a more important effect than the air pressure on the performance of the treatment in terms of the reduction of MDEA, sulfides, COD and TOC. The WAO at 250\u00a0\u00b0C or even 200\u00a0\u00b0C at 50\u00a0bar of air pressure can be considered an effective technology for the on-site treatment of highly MDEA concentrated wastewater. Under these conditions, more biodegradable streams enriched in acetic acid and ammonium compounds were achieved, although their biodegradability could not overpass ca. the 50\u00a0%. On the other hand, microporous carbonaceous materials prepared from a refinery petcoke by chemical activation under different conditions were tested as catalysts in CWAO. These materials showed high specific surfaces areas ranging between 1,013 and 3,459\u00a0m2/g. A carbonaceous material prepared in this work with a specific surface area of 2,196\u00a0m2/g proved a better performance at milder operation conditions than WAO experiments using the same operation conditions. The MDEA removal was improved by more than twofold at 150\u00a0\u00b0C and 50\u00a0bar of air pressure. At 250\u00a0\u00b0C, the CWAO was able to achieve an increase of biodegradability up to 70\u00a0%. The biodegradability decreased up to ca. 30 and 25\u00a0% at 200\u00a0\u00b0C and 150\u00a0\u00b0C, respectively. But these values were still significantly higher than those obtained for WAO experiments in analogous operation conditions (20\u00a0% and\u00a0<\u00a05\u00a0%). It must be noted that these results of biodegradability were estimated according to respirometric tests, whereas biological treatment of the refinery's wastewater treatment plant would operate with a more acclimatized activated sludge and long hydraulic residence times, which can overcome potential problems caused by the presence of remaining MDEA and nitrogenated by-products. In this sense, a compromise between the operating conditions and the resultant biodegradability of the effluent for subsequent mixture with other refinery wastewaters is needed to reduce the operational expenditure of WAO and/or CWAO and determine the techno-economic feasibility of the industrial implementation of this technology as a pre-treatment step for the highly MDEA concentrated wastewater streams.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors thank the financial support of the Community of Madrid through the projects IND2018/AMB-9611 and S2018/EMT-4341 REMTAVARES-CM. Moreover, the authors are grateful to Repsol for providing wastewater samples.Supplementary data to this article can be found online at https://doi.org/10.1016/j.cej.2023.141692.The following are the Supplementary data to this article:\n\nSupplementary Data 1\n\n\n\n", "descript": "\n                  Different operating conditions of wet air oxidation and catalytic wet air oxidation have been studied for the treatment of highly concentrated methyldiethanolamine wastewater streams from amine units of acid gas recovery in petrol refineries. These units occasionally generate streams of high methyldiethanolamine content that require special actions to avoid undesirable impacts on the downstream biological process of the petrochemical wastewater treatment plant due to its inhibition effect. The wet air oxidation treatment achieved remarkable removals of methyldiethanolamine, sulfides, chemical oxygen demand and total organic carbon (99%, 95%, 65% and 38%, respectively). Likewise, activated petroleum coke materials from the own refinery plant were tested as catalysts in the process. These materials were prepared under different conditions (chemical activating agent and thermal carbonization process). The catalytic wet air oxidation treatment using an activated petroleum coke was able to remove the methyldiethanolamine at milder operation conditions keeping a similar performance in terms of wastewater treatment removals as compared to the non-catalytic experiments. This technology significantly increased the biodegradability of the treated effluents ranging from 25 to 70 % due to the formation of more biodegradable substrates (acetic acid and ammonium) for further biological treatment.\n               "}
{"full_text": "Fossil crude is the major raw material used in the petrochemicals industry and the production of transportation fuels. However, the depletion of fossil reserves and the increase in the world's energy demand makes it difficult for the petrochemicals and transportation fuels industry to meet global demand. The more these crude reserves are depleted, the more exploration and refining quality become difficult, which leads to an increase in the price of transportation fuels and petrochemicals (Hosukoglu\u00a0et\u00a0al., 2012; Petrequin,\u00a02012; Liu\u00a0et\u00a0al., 2011; Ong\u00a0and Bhatia,\u00a02010; Galvis\u00a0and de Jong,\u00a02013; Leckel,\u00a02009; Larson\u00a0et\u00a0al., 2010). The Fischer-Tropsch (FT) process is a technology that can produce synthetic transportation fuels and chemicals from biomass, coal and natural gas-derived synthesis gas or syngas (H2\u00a0+\u00a0CO). This technology converts the syngas into hydrocarbons and oxygenates free of sulphur (S) and nitrogen (N) over a wide range of iron (Fe)-, cobalt (Co)-, nickel (Ni)- and ruthenium (Ru)-based catalysts. It has been suggested that the following reactions occur during Fischer-Tropsch synthesis (FTS) (Schulz\u00a0et\u00a0al., 1994; Idem\u00a0et\u00a0al., 2000; Rodr\u00edguez\u00a0Vallejo and de Klerk,\u00a02013; Davis,\u00a02001; van\u00a0Steen and Schulz,\u00a01999; Ponec,\u00a01978):\n\n(1)\nAlkanes nCO\u00a0+\u00a0(2n+1)H2\u00a0\u2192\u00a0C\nn\nH2\n\nn\n\n+2\u00a0+\u00a0nH2O\n\n\n\n\n(2)\nAlkenes nCO\u00a0+\u00a02nH2\u00a0\u2192\u00a0C\nn\nH2\n\nn\n\u00a0+\u00a0nH2O\n\n\n\n\n(3)\nWater-gas shift (WGS) CO\u00a0+\u00a0H2O \u21c4 CO2\u00a0+\u00a0H2\n\n\n\nSome other side reactions that occur during FTS are (Nijs\u00a0and Jacobs,\u00a01980; Lebouvier\u00a0et\u00a0al., 2013; Kummer\u00a0and Emmett,\u00a01953; Lee\u00a0et\u00a0al., 2014):\n\n(4)\nAlcohols nCO\u00a0+\u00a02nH2\u00a0\u2192\u00a0H(-CH2-)\nn\nOH\u00a0+\u00a0(n-1)H2O\n\n\n\n\n(5)\nBoudouard reaction 2CO\u00a0\u2192\u00a0C\u00a0+\u00a0CO2\n\n\n\nThe use of Fe-based catalysts for FTS has been facilitated with chemical promoters such as the group IA elements and structural promoters such as ZnO, Al2O3, MgO and SiO2 (Jager\u00a0and Espinoza,\u00a01995; Raje\u00a0et\u00a0al., 1998; Bukur\u00a0et\u00a0al., 1990; Yang\u00a0et\u00a0al., 2005). This is due to their ability to undergo WGS reaction to make up the deficient H2 in syngas derived from coal. Fe-based catalysts have a low cost, flexible reaction conditions and high FTS activity, although somewhat less active than Co-based catalysts (Li\u00a0et\u00a0al., 2016; Lohitharn\u00a0and Goodwin,\u00a02008; Griboval-Constant\u00a0et\u00a0al., 2014).The promotion of Fe-based catalysts with group IA metals such as Na, K and Cs can lead to a shift in the product distribution and the production of higher molecular weight hydrocarbons in FTS (Uner,\u00a01998; Miller\u00a0and Moskovits,\u00a01988; Xiong\u00a0et\u00a0al., 2015). A typical Fe-based catalyst with a group IA metal as a chemical promoter can influence the dispersion and reduction behaviour of Fe oxides (Xiong\u00a0et\u00a0al., 2015). This could also lead to an increase in the CO consumption and dissociation rates, the CO2 production rate and the olefin to paraffin (O/P) ratio, and a drop in the overall FT reaction rate. Na and K have been observed to decrease the CH4 selectivity and exhibit a much higher WGS reaction (An\u00a0et\u00a0al., 2007; Dry\u00a0and Oosthuizen,\u00a01968). Na and K promotional effects are related to the Fe metal local electron density modification and blockage of the catalytic active sites (An\u00a0et\u00a0al., 2007; Dry\u00a0et\u00a0al., 1969). These can be explained in terms of the Na and K electropositivity, which causes transfer of the charge to the Fe metal leading to a decrease in the adsorption of H2 (Li\u00a0et\u00a0al., 2016; Li\u00a0et\u00a0al., 2015). Li et\u00a0al. (Li\u00a0et\u00a0al., 2016) reported the promotional effect of Na on Fe-based catalyst in FTS to restrain the reduction property of Fe oxides in the catalyst. And, the Na was also reported to facilitate carbonisation of the Fe-based catalyst, decrease the FTS activity, methane selectivity and paraffins, and increase the production of heavier hydrocarbons (C5+) and olefin. Xiong et\u00a0al. (Xiong\u00a0et\u00a0al., 2015) also prepared a carbon nanotube (CNTs)-supported Fe catalyst with different IA group alkali promoters, and found that Na increased the crystallite size of Fe oxides, while the surface area of the Fe/CNT catalyst decreased in the presence of Na. In addition, the CO conversion and long chain hydrocarbons (C5+) were also found to increase, and the presence of Na in the Fe-based catalyst slightly inhibited the reducibility of Fe oxides.The dependency of Na promotion on hydrocarbon production and CO conversion for Fe-based catalysts in FTS is still not well established in principle. Further research is needed to understand Na effects in Fe-based catalysts. The present study was carried out to investigate the effects of Na in Fe-based catalysts at different reaction temperatures (250 \u2013 310 \u00b0C) during FTS. For this reason, an Fe-based catalyst (Fe/Al2O3) and its Na-promoted form (FeNa/Al2O3) were prepared and characterized with various techniques to examine their FT activity at different reaction temperatures. The emphasis was on the Na effects in Fe-based catalysts in terms of the methane formation rate, conversions of H2 and CO, WGS reaction, and olefins and paraffins of the lower, middle and higher range hydrocarbons.The Fe/Al2O3 FTS catalyst used in this study was prepared using the solution of Fe(NO3)3\u00b79H2O (\u2265 98%, Sigma-Aldrich) as Fe precursor. This precursor was used for impregnation of the active Fe (15\u00a0wt.%) metal on 3\u00a0mm alumina (Al2O3) support from Sigma-Aldrich. The incipient-wetness impregnation (IWI) preparation procedure was used to prepare the Fe/Al2O3 catalyst sample. Briefly, 33.1\u00a0g of Fe(NO3)3\u00b79H2O (\u2265 98%, Sigma-Aldrich) was weighed into 50\u00a0mL deionised water (H2O), and stirred at 60\u00a0rpm and room temperature for 30\u00a0min to obtain the Fe precursor solution. The 3\u00a0mm alumina (Al2O3) support (30\u00a0g, used as received, without drying) from Sigma-Aldrich was then poured into the solution and continue stirring for 24\u00a0h (20\u2013100 \u00b0C heating value) to obtain a homogeneity sample. After the impregnation procedure, the catalyst sample was dried at 120 \u00b0C for 6\u00a0h, calcined in air at 500 \u00b0C for 3\u00a0h and ground and sieved to 15\u201360\u00a0\u00b5m particle size, in order to obtain the Fe/Al2O3 FTS catalyst.Simultaneously, and using the solutions of Fe(NO3)3\u00b79H2O (\u2265 98%, Sigma-Aldrich) and Na2O (80%, Sigma-Aldrich) as Fe and Na precursors, respectively, the same IWI preparation technique described above was used to prepare the Na-promoted Fe/Al2O3 FTS catalyst. These precursors were used for impregnation of the active Fe (15\u00a0wt.%) and Na (2\u00a0wt.%) metals on 3\u00a0mm alumina (Al2O3) support from Sigma-Aldrich. Briefly, 22.1\u00a0g of Fe(NO3)3\u00b79H2O (\u2265 98%, Sigma-Aldrich) and 0.7\u00a0g of Na2O (80%, Sigma-Aldrich) were weighed into 50\u00a0mL deionised water (H2O), and stirred at 60\u00a0rpm and room temperature for 30\u00a0min to obtain the Fe and Na precursor solutions. The 3\u00a0mm alumina (Al2O3) support (20\u00a0g, used as received, without drying) from Sigma-Aldrich was then poured into the solution and continue stirring for 24\u00a0h (20\u2013100 \u00b0C heating value) to obtain a homogeneity sample. However, the catalyst sample was dried at 120 \u00b0C for 12\u00a0h after the impregnation technique, calcined in air at 500 \u00b0C for 3\u00a0h and ground and sieved to 15\u201360\u00a0\u00b5m particle size, in order to obtain the FeNa/Al2O3 FTS catalyst.The FTS catalysts were characterized using different techniques. The physical adsorption technique of N2 physisorption was done for the catalysts Brunauer-Emmett-Teller (BET) surface area, pore volume and average pore size distribution using a Micromeritics ASAP 2460 instrument. Prior to this technique, each catalyst sample was degassed at a temperature of 120 \u00b0C for 3\u00a0h and the results of the analysis were summarized accordingly.The scanning electron microscopy (SEM) technique was performed on the FTS catalysts to determine their morphologies, using a Jeol JSM-7800F Field Emission Scanning Electron Microscope instrument with high resolution in the scanning electron mode. The accelerating voltage and working distance used during this technique were 5 \u2013 10\u00a0kV and 15\u00a0mm, respectively. Elemental composition analysis of the catalysts was carried out during SEM analysis using the Energy-dispersive X-ray (EDX) spectroscopy technique. Prior to the SEM and EDX analyses, a small amount of each catalyst sample was dropped on the SEM grid and the sample was coated with a thin carbon layer for the catalyst topographic evaluation. The images of the catalyst samples were viewed using a digital charge coupled device (CCD) camera equipped with the SEM instrument.The image characteristics of the prepared FTS catalysts were investigated using a Jeol JEM-2100F Field Emission Transmission Electron Microscope. The transmission electron microscopy (TEM) technique was carried out at an accelerating voltage of 200\u00a0kV and the microscope was equipped with a LaB6 source and CCD camera for electron emission and imaging, respectively. Before the TEM technique, each sample of the catalyst was suspended in ethanol and the suspension was then dropped on a copper grid coated with a thin carbon film for the assessment of the catalyst image characteristics. The measurements were performed at a beam spot range of 50 \u2013 100\u00a0nm in the transmission electron mode.The catalysts crystallinity was determined by the powder X-ray diffraction (XRD) technique using a PANayltical X'Pert Pro-powder diffractometer instrument. This instrument was fitted with an ID X'Celerator detector of PHD lower and upper level of 6.67\u00a0keV and 12.78\u00a0keV, respectively. The instrument was also fitted with a programmable divergence slit with a radiation length of 10\u00a0mm. The XRD technique was conducted using the 2\u03b8 range of 10 to 90\u00b0 with Cu K\u03b1 radiation of \u03bb\u00a0=\u00a00.15405\u00a0nm at 40\u00a0mA and 40\u00a0kV operation conditions. The diffractometer was configured with a PW3064 sample spinner of 1\u00a0second rotation time, and the scan step size and time were 0.0170\u00b0 2\u03b8 and 87\u00a0s, respectively.The hydrogen-temperature programmed reduction (H2-TPR) technique was conducted using a Micromeritics AutoChem II 2920 instrument, where each catalyst sample was put in a pre-heated tubular reactor (quartz) with a thermocouple for continuous measurement of the reduction temperature. Before the H2-TPR technique, each sample of the catalyst was degassed with a high purity argon (99.999%, Ar) at temperature of 150 \u00b0C for 1\u00a0hour, and after which, the reactor temperature was dropped down to 50 \u00b0C. The reduction was carried out with 10% H2/Ar gas mixture of a flow rate of 50 cm3(STP)/min and the reactor temperature was raised to 950 \u00b0C at a ramping rate of 10 \u00b0C/min against the thermal conductivity detector (TCD) signal.The prepared FTS catalysts were evaluated in a tubular fixed-bed reactor with an internal diameter (ID) of 10.2\u00a0mm and a tube length (TL) of 412.8\u00a0mm. Prior to the evaluation of the catalysts, 1.0\u00a0g of each prepared FTS catalyst was loaded into the isothermal area of the fixed-bed reactor. The upper and lower levels of the reactor were filled with glass beads, thereby placing the catalyst bed in the middle of the reactor. Each catalyst was reduced in-situ for 12\u00a0h using a syngas ratio of 0.9, at a temperature of 300 \u00b0C and pressure of 1\u00a0bar. The FT reaction was carried out using the same reduction syngas (H2/CO) ratio of 0.9, weight hourly space velocity (WHSV) of 3.7 SLph/gcat, four different reaction temperatures (T) of 250 \u00b0C, 270 \u00b0C, 290 \u00b0C and 310 \u00b0C, and a pressure (P) of 10 bar(g). The tail-gas was analysed using an INFICON Micro Gas Chromatography (GC) Fusion 4-Module System with different TCDs and columns. Module A, which contained TCD and a Rt-Molsieve 5A column was used to analyse H2, N2, CO and CH4. Module B, which contained TCD and a Rt-U-Bond column was used to analyse CO2, C2H4 and C2H6. Modules C and D contained Rt-Alumina and Rxi-1\u00a0ms columns, respectively, and their TCDs were used to analyse the other hydrocarbons in the tail-gas. The liquid, wax and H2O were collected at the cold and hot traps set in between the reactor and back pressure regulator for analysis using a flame ionization detector (FID), DB-5MS, ZB-1HT and packed columns, respectively. The catalyst activity was evaluated by CO conversion, H2 conversion, formation rate and hydrocarbon selectivity on carbon basis (CO2-free), as defined in the equations below:\n\n(6)\n\n\n\nCO\n\nconversion\n\n\n:\n\n\n\n\nX\n\nc\no\n\n\n\n\n\n(\n%\n)\n\n=\n100\n\nx\n\n\n\n\nM\n\nC\nO\n,\ni\nn\n\n\n\n\u2212\n\n\nM\n\nC\nO\n,\no\nu\nt\n\n\n\n\nM\n\nC\nO\n,\ni\nn\n\n\n\n\n\n\nwhere MCO,in\n and MCO,out\n are the molar flowrate (mol/h) of CO at inlet and CO at outlet, respectively.\n\n(7)\n\n\n\nH\n2\n\n\n\nconversion\n\n\n:\n\n\nX\n\nH\n2\n\n\n\n(\n%\n)\n\n=\n100\n\nx\n\n\n\n\nM\n\n\nH\n2\n\n,\ni\nn\n\n\n\n\u2212\n\n\nM\n\n\nH\n2\n\n,\no\nu\nt\n\n\n\n\nM\n\n\nH\n2\n\n,\ni\nn\n\n\n\n\n\n\n\n\n\n\nM\n\n\nH\n2\n\n,\ni\nn\n\n\n and \n\nM\n\n\nH\n2\n\n,\no\nu\nt\n\n\n are the molar flowrate (mol/h) of H2 at inlet and outlet, respectively.\n\n(8)\n\n\n\nFormation\n\nrate\n\n\n:\n\n\nr\n\nC\nn\n\n\n\n(\n\nmol\n/\n\ng\ncat\n\n.\nh\n\n)\n\n=\n\n\n\nM\n\n\nC\nn\n\n,\no\nu\nt\n\n\n\nW\n\nc\na\nt\n\n\n\n\n\n\n\n\n\n(9)\n\n\n\nCO\n2\n\n\n\nselectivity\n\n\n:\n\n\nS\n\nC\n\nO\n2\n\n\n\n\n(\n%\n)\n\n=\n100\n\nx\n\n\n\nM\n\nC\n\nO\n2\n\n,\no\nu\nt\n\n\n\n\nM\n\nC\nO\n,\ni\nn\n\n\n\n\u2212\n\n\nM\n\nC\nO\n,\no\nu\nt\n\n\n\n\n\n\n\n\n\n\n\n\nM\n\nC\n\nO\n2\n\n,\no\nu\nt\n\n\n is the molar flowrate (mol/h) of CO2 at outlet.\n\n(10)\n\n\n\nHydrocarbon\n\nselectivity\n\n\n:\n\n\nS\n\nC\nn\n\n\n\n(\n%\n)\n\n=\n100\n\nx\n\n\n\n\nM\n\n\nC\nn\n\n,\no\nu\nt\n\n\n\nx\n\nn\n\n\n\nM\n\nC\nO\n,\ni\nn\n\n\n\n\u2212\n\n\nM\n\nC\nO\n,\no\nu\nt\n\n\n\u2212\n\n\nM\n\nC\n\nO\n2\n\n,\no\nu\nt\n\n\n\n\n\n\n\n\n\n\n\nM\n\n\nC\nn\n\n,\no\nu\nt\n\n\n is the molar flowrate (mol/h) of hydrocarbon with carbon number, n at outlet. Wcat\n (g) is the weight of each catalyst during the FT reaction.The selectivity to C13+ hydrocarbons (including oxygenates) is defined as:\n\n(11)\n\n\n\nS\n\nC\n\n13\n+\n\n\n\n\n(\n%\n)\n\n=\n100\n\u2212\n\n\n\u2211\n\nn\n=\n1\n\n12\n\n\nS\n\nC\nn\n\n\n\n\n\n\n\n\n\n\n\n\nwhere\n\nn\n\n=\n1\n\nto\n12\n\n\n\n\nThe prepared surface area, pore volume and average pore size of the FTS catalysts (Fe/Al2O3 and FeNa/Al2O3) are summarized in Table\u00a01\n. The result of the physical adsorption technique of N2 physisorption show that the Na-promoted Fe/Al2O3 catalyst has a lower BET surface area compared to the unpromoted Fe/Al2O3 catalyst. Na promotion decreased the BET surface area and influenced the textural properties of the Fe/Al2O3 catalyst (Li\u00a0et\u00a0al., 2016; Xiong\u00a0et\u00a0al., 2015; X.\u00a0An\u00a0et\u00a0al., 2007). The pore volume and the average pore size of the catalysts show no significant differences in the N2 physisorption results. This indicates that no blockage of the catalyst pores was experienced during the loading of the Fe and Na metals on the Al2O3 support.The SEM images showing the morphologies of the Fe/Al2O3 FTS catalyst are shown in Fig.\u00a01\n(a) and (c), while that of the Na-promoted FTS catalyst (FeNa/Al2O3) are shown in Fig.\u00a01(b) and (d). The images show that both prepared catalysts formed sphere-like catalyst structures that can clog together to form bigger sphere-like catalyst particles. This is also clearly seen in the TEM images in Fig.\u00a02\n(a) and (c) for the Fe/Al2O3 catalyst sample and Fig.\u00a02(b) and (d) for the FeNa/Al2O3 catalyst sample. These images show the characteristic features of the prepared catalyst samples. The particle size distribution ranging from 1 to 5\u00a0nm was observed with an evenly and unevenly distribution of the Fe and Na active metals in the TEM images. There was also a slight increase in the particle size when the Fe/Al2O3 catalyst was promoted with Na. This corresponds to what was observed in the N2 physisorption results, as the Na-promoted catalyst showed a smaller BET surface area than the unpromoted catalyst. The TEM images also show that Na promotion inhibited the crystallinity of the Fe/Al2O3 catalyst (Fig.\u00a02(b)). This is also noticeable in the H2-TPR profile of the FeNa/Al2O3 catalyst sample, and it will be discussed later in this section. The particle size distribution can range from 30 to 120\u00a0nm when both the prepared catalysts clog together, as seen in the SEM and TEM images.The elemental analyses of EDX performed during the SEM techniques confirmed the presence of Fe and Na active metals in the prepared Fe/Al2O3 and FeNa/Al2O3 FTS catalysts, as summarized in Table\u00a01. The nominal amount of Na prepared relative to the 15 wt% of Fe was 2 wt% in FeNa/Al2O3 FTS catalyst. This nominal amount of Na was used to have a clear indication of Na in the Na-promoted catalyst. There were 15.4 and 11.1 wt% Fe contents detected in the Fe/Al2O3 and FeNa/Al2O3 catalyst samples, respectively. Na was not detected in the Fe/Al2O3 catalyst sample, while 1.6 wt% Na was detected in the FeNa/Al2O3 catalyst sample. This indicates successful loading of both Fe and Na FTS active metals onto the Al2O3 support.\nFig.\u00a03\n shows the powder XRD patterns of the prepared Fe/Al2O3 and FeNa/Al2O3 FTS catalysts. These patterns contain peaks of the crystallites present in the catalyst samples. It can be noted that the XRD patterns contain haematite (Fe2O3) crystalline peaks at 2\u03b8 values of 24.3, 33.3, 49.6 and 64.5\u00b0 for both the Fe/Al2O3 and FeNa/Al2O3 catalyst samples. The magnetite (Fe3O4) peaks can be found at 2\u03b8 values of 35.8, 54.4 and 63.0\u00b0, and these are the characteristics of the iron oxides phases, including the haematite peaks. The Fe2O3 and Fe3O4 peak characteristics agree with their reduction steps, which will be discussed later in this section. The XRD patterns also contain alumina (Al2O3) peaks at 2\u03b8 values of 41.1, 58.0, 67.2, 72.5 and 75.8\u00b0. Na2O and Na2O2 crystalline peaks can be seen at 2\u03b8 values of 27.9, 37.4 and 44.6\u00b0 for the FeNa/Al2O3 catalyst sample. Therefore, the presence of these crystalline peaks suggests the preparation and characteristics of the Fe/Al2O3 and FeNa/Al2O3 FTS catalysts.H2-TPR experiments were performed on the prepared Fe/Al2O3 and FeNa/Al2O3 FTS catalysts to investigate their reduction behaviour, as represented in Fig.\u00a04\n. The powder XRD patterns of the catalysts showed the characteristic crystalline peaks of Fe2O3 and Fe3O4. These Fe oxides are known to form active metallic Fe during appropriate reduction conditions. Fe2O3 proceeds to form Fe metal (Fe0) in two or three stages, and it could form iron oxide (FeO) for supported Fe based catalysts (Al-Dossary\u00a0and Fierro,\u00a02015; Jin\u00a0and Datye,\u00a02000; Qing\u00a0et\u00a0al., 2012). The H2-TPR profile of the Na-promoted catalyst (FeNa/Al2O3) shows that reduction occurred at a temperature of 295 \u00b0C (stage b). This could be attributed to the effects of Na in the catalyst, as the 295 \u00b0C temperature reduction step is not present in the H2-TPR profile of the Fe/Al2O3 catalyst. The Fe/Al2O3 and FeNa/Al2O3 FTS catalysts H2-TPR profile started at a reduction temperature of 215 and 230 \u00b0C (stage a), respectively. Stage c is a reduction step for FeNa/Al2O3 catalyst, at a temperature of 38 5 \u00b0C, which is higher than the reduction temperature of 360 \u00b0C for the Fe/Al2O3 catalyst at stage c. This indicates that Na promotion inhibited the reduction behaviour of the Fe/Al2O3 catalyst slightly, and the reduction step is thus assigned to reduction of Fe2O3 to Fe3O4 (Xiong\u00a0et\u00a0al., 2015; Qing\u00a0et\u00a0al., 2012; Li\u00a0et\u00a0al., 2013). Stage d occurs at a reduction temperature of 445 \u00b0C and is ascribed to reduction of Fe3O4 to FeO in the Fe/Al2O3 catalyst, due to strong interaction between the active Fe metal and Al2O3 support. This stage is not present in the H2-TPR profile of the FeNa/Al2O3 catalyst, as the Na helped to generate some amorphous phases, as seen in the TEM images. Therefore, Na reduced the interaction between the active Fe metal and the Al2O3 support and provided a better reduction step for FeO to Fe0, as seen at stage e (reduction temperature of 550 \u00b0C). Stage f (reduction temperature of 765 \u00b0C) represents the reduction step of Fe-aluminates present in the both catalyst systems.The catalytic performance of the prepared Fe/Al2O3 and FeNa/Al2O3 catalysts for FTS were evaluated in a tubular fixed-bed reactor. This was carried out at four different reaction temperatures \u2013 250 \u00b0C, 270 \u00b0C, 290 \u00b0C and 310 \u00b0C to investigate how the Na-promoted Fe/Al2O3 catalyst behaves at different temperatures compared to the unpromoted Fe/Al2O3 catalyst. Although, some researchers have reported the effects of Na on Fe-based catalysts, limited studies are available on its behaviour at different temperatures (Xiong\u00a0et\u00a0al., 2015; Li\u00a0et\u00a0al., 2014; H.M.T.\u00a0Galvis\u00a0et\u00a0al., 2013; H.M.T.\u00a0Galvis\u00a0et\u00a0al., 2013). The present study was conducted for each catalyst under the following conditions: a total reaction time on stream (TOS) of circa 360\u00a0h; a pressure (P) of 10 bar(g); H2/CO of 0.9; a weight hourly space velocity (WHSV) of 3.7 SLph/gcat. A summary of the FTS experiment results is provided in Table\u00a02\n and the TOS performance profiles at reaction temperature of 310 \u00b0C are represented in Fig.\u00a05\n. Fig.\u00a05 shows a rapid increase in the CO conversion, which stabilised within few hours of TOS. The CO conversion for Fe/Al2O3 catalyst reached a steady state after 20\u00a0h TOS, while the CO conversion for FeNa/Al2O3 catalyst reached a steady state after 80\u00a0h TOS. The later steady state behaviour of the FeNa/Al2O3 catalyst can be explained with respect to the Na inhibition of the reduction behaviour of Fe/Al2O3 catalyst as discussed earlier for the H2-TPR profiles.The CO and H2 conversions at the steady state were used to evaluate the FTS activity of the catalysts, as shown in Fig.\u00a06\n. The behaviour of the FeNa/Al2O3 catalyst did not show the same trend as the Fe/Al2O3 catalyst at different temperatures. The CO and H2 conversions of the Fe/Al2O3 catalyst increased when the reaction temperature was increased, while the CO and H2 conversions of the FeNa/Al2O3 catalyst gave a similar trend at different temperatures. Na as an alkali metal promoter in Fe based catalysts is known to increase CO conversion and decrease H2 conversion during FTS due to increase in the catalyst dissociative adsorption rate of CO and surface basicity, respectively (Li\u00a0et\u00a0al., 2016; Li\u00a0et\u00a0al., 2015; H.M.T.\u00a0Galvis\u00a0et\u00a0al., 2013; Ribeiro\u00a0et\u00a0al., 2010).Na improves the conversion of CO in the low to average conversion range and is ineffective or hinders CO conversion in the high conversion range (Li\u00a0et\u00a0al., 2016; Ribeiro\u00a0et\u00a0al., 2010). However, comparing the CO conversion of both catalysts in Fig.\u00a06 indicates that the presence of Na increased CO conversion at a reaction temperature of 290 \u00b0C and 310 \u00b0C. The Fe/Al2O3 catalyst and the Na-promoted Fe/Al2O3 catalyst (FeNa/Al2O3) gave a similar CO conversion rate at a reaction temperature of 270 \u00b0C, and Na inhibited CO conversion at a reaction temperature of 250 \u00b0C. The CO and H2 conversions of the FeNa/Al2O3 catalyst obtained at reaction temperature of 250 \u00b0C are more similar and the difference becomes more obvious as the reaction temperature increases. The reason for these phenomena could be competitive adsorption of dissociative CO and H2 on the Na-promoted catalyst at a reaction temperature of 250 \u00b0C and 270 \u00b0C (Li\u00a0et\u00a0al., 2014). This suggests that Na could be effective at improving CO conversion at certain higher reaction temperatures and hinders CO conversion at lower reaction temperatures. The H2 conversion results for the FeNa/Al2O3 catalyst were much lower than that of the H2 conversion of the Fe/Al2O3 catalyst at a reaction temperature of 250 \u00b0C and 270 \u00b0C, and closer at a reaction temperature of 290 \u00b0C and 310 \u00b0C. This also suggests that Na effects in Fe-based catalysts to increase CO conversion and decrease H2 conversion are dependent on the reaction temperature during FTS.\nFig.\u00a07\n shows a linear relationship between the CO2 selectivity and reaction temperature of the Fe/Al2O3 and FeNa/Al2O3 catalysts. The Na-promoted Fe/Al2O3 catalyst exhibited a much higher CO2 selectivity at a reaction temperature of 250 \u00b0C, 270 \u00b0C, 290 \u00b0C and 310 \u00b0C than the unpromoted Fe/Al2O3 catalyst did. This explains the FeNa/Al2O3 catalyst behaviour in terms of the hydrocarbons production rate, where the FT rate is lower than the FT rate of Fe/Al2O3 catalyst at these reaction temperatures, as also seen in Table\u00a02. The higher CO2 selectivity suggests an improved Boudouard (2CO\u00a0=\u00a0C\u00a0+\u00a0CO2) or WGS (CO\u00a0+\u00a0H2O\u00a0=\u00a0CO2\u00a0+\u00a0H2) reaction rate at different reaction temperatures, when the Fe/Al2O3 catalyst was promoted with Na. The reason for the increased WGS reaction at all reaction temperatures could not be ascertained, because there was a decrease in the CO adsorption rate at a reaction temperature of 250 \u00b0C, and an increase in the CO adsorption rate at a reaction temperature of 290 \u00b0C and 310 \u00b0C. However, Na promotion decreased the H2 conversion in the Fe/Al2O3 catalyst at all reaction temperatures (Fig.\u00a06). This indicates that H2O or oxygen adsorption could be increased by the presence of Na in the Fe/Al2O3 catalyst, which would lead to an increase in the WGS reaction at all reaction temperatures.The linear relationship between the formation rate of methane (C1) and the reaction temperature of the Fe/Al2O3 and FeNa/Al2O3 catalysts is shown in Fig.\u00a07. The C1 formation rate of the Na-promoted Fe/Al2O3 catalyst that is based on CO and H2 is seen following different trend from CO2 selectivity that is based on CO and H2O. Table\u00a02 also shows the C1 selectivity of the Fe/Al2O3 and FeNa/Al2O3 catalysts with the same trend of C1 formation rate of the Fe/Al2O3 and FeNa/Al2O3 catalysts at different reaction temperatures. The results seen in Fig.\u00a07 indicate that the C1 formation rate of the FeNa/Al2O3 catalyst at a reaction temperature of 290 \u00b0C and 310 \u00b0C was significantly higher than the C1 formation rate of the Fe/Al2O3 catalyst at the same reaction temperature. This was in the reverse order at a reaction temperature of 250 \u00b0C and 270 \u00b0C, as the C1 formation rate of the FeNa/Al2O3 catalyst was significantly lower than the C1 formation rate of the Fe/Al2O3 catalyst.Many studies have reported that Na is an effective group IA promoter that decreases the C1 formation rate (Dry\u00a0and Oosthuizen,\u00a01968; Galvis\u00a0et\u00a0al., 2013; Ribeiro\u00a0et\u00a0al., 2010). However, the FTS evaluation experiments for the Fe/Al2O3 and FeNa/Al2O3 catalysts show that the C1 formation rate of the Na-promoted Fe/Al2O3 catalyst decreased at certain lower reaction temperatures, but increased at certain higher reaction temperatures. The conclusion for these phenomena could be ascertained to competitive adsorption of dissociative CO and H2 on the Na-promoted catalyst at reaction temperatures of 250 \u00b0C and 270 \u00b0C (Li\u00a0et\u00a0al., 2014; Ribeiro\u00a0et\u00a0al., 2010). This indicates that Na effects in Fe-based catalysts to decrease C1 formation rate are dependent on the reaction temperature during FTS.The linear relationship between the selectivity of the lower hydrocarbons (C2\u00b0-C4\u00b0 and C2\np-C4\np) and the reaction temperature for the prepared Fe/Al2O3 and FeNa/Al2O3 catalysts can be seen in Fig.\u00a08\n. This can also be seen in Table\u00a02, as the overall C2-C4 selectivity increased with the increasing reaction temperature when the Fe/Al2O3 catalyst contained Na. This indicates improved C2-C4 selectivity for the Na-containing Fe/Al2O3 catalyst at all reaction temperatures during FTS.Many researchers have reported that Na is an effective alkali metal promoter that increases the selectivity of lower olefin (C2\u00b0-C4\u00b0); and decreases the selectivity of lower paraffin (C2\np-C4\np) (An\u00a0et\u00a0al., 2007; Dry\u00a0and Oosthuizen,\u00a01968; Ribeiro\u00a0et\u00a0al., 2010; Abbot\u00a0et\u00a0al., 1986; Lama\u00a0et\u00a0al., 2018). A slightly higher C2\np-C4\np selectivity was observed with the Na-promoted Fe/Al2O3 catalyst at a reaction temperature of 290 \u00b0C. The same trends can be seen in Fig.\u00a08, which shows that C2\u00b0-C4\u00b0 selectivity increased and C2\np-C4\np selectivity decreased at different reaction temperatures when the Fe/Al2O3 catalyst was promoted with Na. The reason for this behaviour is the improved surface basicity of the Na-promoted Fe/Al2O3 catalyst, as its inhibited H2 conversions were also observed at all reaction temperatures. This suggests that the effects of Na in Fe-based catalysts, in terms of increasing C2\u00b0-C4\u00b0 selectivity and decreasing C2\np-C4\np selectivity, are independent of the reaction temperature during FTS.\nTable\u00a02 shows that overall C5-C12 selectivity increased at all reaction temperatures (250\u00a0\u00b0C, 270\u00a0\u00b0C, 290\u00a0\u00b0C and 310\u00a0\u00b0C) when the Fe/Al2O3 catalyst was promoted with Na. This suggests that Na-containing Fe-based catalysts can be used to improve selectivity towards middle range hydrocarbons (C5-C12) at different reaction temperatures in FTS. The behaviour of middle olefins (C5\u00b0-C12\u00b0) and paraffins (C5\np-C12\np) at different reaction temperatures with the Fe/Al2O3 and FeNa/Al2O3 catalysts can be explained using Fig.\u00a09\n. It can be seen that the C5\u00b0-C12\u00b0 and C5\np-C12\np selectivity was higher with the FeNa/Al2O3 catalyst at all reaction temperatures compared to that of the Fe/Al2O3 catalyst. Na has been reported to facilitate the carbonisation of an Fe-based catalyst and in context, this can improve the selectivity towards C5+ hydrocarbons (Li\u00a0et\u00a0al., 2016; Xiong\u00a0et\u00a0al., 2015; Ribeiro\u00a0et\u00a0al., 2010). Na was also observed to hinder the reduction behaviour of the Fe/Al2O3 catalyst as earlier discussed. This is due to the increase in the strength of Fe-O bonds present in the Fe2O3 of the FeNa/Al2O3 catalyst, which then leads to an increase in the selectivity towards C5\u00b0-C12\u00b0 and C5\np-C12\np (Li\u00a0et\u00a0al., 2016). This indicates that the effects of Na in Fe-based catalysts, in terms of improving selectivity towards C5\u00b0-C12\u00b0 and C5\np-C12\np, are independent of the reaction temperature during FTS. The trend seen with C5\u00b0-C12\u00b0 and C5\np-C12\np selectivity of the FeNa/Al2O3 catalyst is not linear at all the reaction temperatures, as C5\u00b0-C12\u00b0 selectivity tends to be: constant at a reaction temperature of 270\u00a0\u00b0C, 290\u00a0\u00b0C and 310\u00a0\u00b0C; lower at a reaction temperature of 250\u00a0\u00b0C. The FeNa/Al2O3 catalyst C5\np-C12\np selectivity tends to be: constant at a reaction temperature of 290\u00a0\u00b0C and 310\u00a0\u00b0C; linear at a reaction temperature of 250\u00a0\u00b0C and 270\u00a0\u00b0C. The trend seen with a Fe/Al2O3 catalyst C5\u00b0-C12\u00b0 and C5\np-C12\np selectivity is linear at all the reaction temperatures. It can also be seen that the increment in the Na-promoted Fe/Al2O3 catalyst (FeNa/Al2O3) selectivity towards C5\u00b0-C12\u00b0 became more obvious as the reaction temperature was raised. This phenomenon could be related to different competitive adsorption behaviours of dissociative CO and H2 at different reaction temperatures in the Na-promoted Fe/Al2O3 catalyst during FTS. Therefore, the Na effects in Fe-based catalysts, in terms of raising the selectivity of C5\u00b0-C12\u00b0 are dependent on the reaction temperature during FTS.\nFig.\u00a010\n shows the selectivity of the higher hydrocarbons (C13+), including oxygenates (Oxy), as a function of the reaction temperature of the Fe/Al2O3 and FeNa/Al2O3 catalysts. There is a drop in product selectivity to C13+, including oxygenates for the Na-promoted Fe/Al2O3 catalyst (FeNa/Al2O3) at different reaction temperatures, when compared with the unpromoted catalyst (Fe/Al2O3). Although, the drop in C13+ (including oxygenates) product selectivity for the FeNa/Al2O3 catalyst becomes more obvious as the reaction temperature rises, it can still be seen to produce less C13+ (including oxygenates) at all reaction temperatures. The phenomenon indicates that Na hindered the product selectivity of C13+ (including oxygenates) during the FTS experiment. This could relate to an increase in product selectivity of lower and middle range hydrocarbons (C2-C4 and C5-C12) for the Na-containing catalyst (FeNa/Al2O3), at all reaction temperatures studied, when compared with the unpromoted catalyst (Fe/Al2O3). Therefore, Na can be used to reduce product selectivity towards C13+ (including oxygenates) in Fe-based catalysts. Na effect in Fe-based catalysts, in terms of reducing product selectivity towards C13+ (including oxygenates) is independent of the reaction temperature during FTS.The alumina-supported Fe catalyst (Fe/Al2O3) and its Na-promoted catalyst (FeNa/Al2O3) were prepared and characterized with various techniques to investigate their behaviour at different reaction temperatures (250 \u2013 310 \u00b0C) in FTS. It was found that there was a slight increase in particle size when the Fe/Al2O3 catalyst was promoted with Na. This corresponds with what was observed in the N2 physisorption results, as the Na-promoted catalyst (FeNa/Al2O3) gave a smaller BET surface area than that of the unpromoted catalyst (Fe/Al2O3). The TEM images showed that Na promotion inhibited the crystallinity of the Fe/Al2O3 catalyst. This was also noticed in the H2-TPR profile of the FeNa/Al2O3 catalyst sample. Na reduced the interaction between the active Fe metal and Al2O3 support to give a better reduction step from FeO to Fe0. The emphasis was on the Na effects in Fe-based catalysts, in terms of: C1 formation rate; H2 and CO conversion; WGS reaction; olefins and paraffins of the lower, middle and higher range hydrocarbons. The results show that the Na effects in Fe-based catalysts of increasing CO conversion and decreasing H2 conversion are dependent on the reaction temperature during FTS. The C1 formation rate of the Na-promoted Fe/Al2O3 catalyst was lower at certain lower reaction temperatures and higher at certain higher reaction temperatures. The conclusion drawn was that they were due to competitive adsorption of dissociative CO and H2 on the Na-promoted catalyst at reaction temperatures of 250 \u00b0C and 270 \u00b0C. The higher CO2 selectivity suggests an improved WGS reaction rate at different reaction temperatures, when the Fe/Al2O3 catalyst is promoted with Na. It was also found that the Na effects in Fe-based catalysts to raise C2\u00b0-C4\u00b0, C5\u00b0-C12\u00b0 and C5\np-C12\np selectivity and reduce C2\np-C4\np and C13+ (including oxygenates) selectivity are independent on the reaction temperature in FTS. Therefore, Na-containing Fe-based catalysts could be used to improve selectivity towards light olefins (C2\u00b0-C4\u00b0) and middle range hydrocarbons (C5-C12) at different reaction temperatures in FTS. The promotion of the alumina-supported Fe catalyst (Fe/Al2O3) and its Na-promoted catalyst (FeNa/Al2O3) with another group IA metal such as K and Cs is recommended for FTS at different reaction temperatures.The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.\nAliu A. Adeleke: Conceptualization, Methodology, Formal analysis, Investigation, Writing \u2013 original draft, Visualization. Muthu Kumaran Gnanamani: Investigation, Writing \u2013 review & editing, Supervision. Michela Martinelli: Formal analysis, Investigation. Burtron H. Davis: Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.We would like to acknowledge the financial support provided by the University of South Africa (UNISA) and National Research Foundation (NRF): UID 111348 and 113648.", "descript": "\n                  Na in Fe-based catalysts can be used to increase CO conversion and C2-C4 olefins and decrease the conversion of H2 and C1 selectivity, but its behaviour at different reaction temperatures is of importance in Fischer-Tropsch synthesis (FTS). The dependency of the C1 formation rate, the conversions of H2 and CO, the water-gas shift reaction, the olefins and paraffins of the C2-C4 and C5-C12 hydrocarbons, and C13+ hydrocarbons on the reaction temperature for prepared Fe/Al2O3 and FeNa/Al2O3 catalysts was evaluated in a tubular fixed-bed reactor. This was done to investigate the effects of Na in Fe-based catalyst at different reaction temperatures (250 \u2013 310 \u00b0C). The results show that the effects of Na in Fe-based catalysts to increase CO conversion and decrease H2 conversion are dependent on the reaction temperature in FTS. The Na-promoted Fe-based catalyst (FeNa/Al2O3) gave a lower C1 formation rate at certain lower reaction temperatures (250 \u00b0C and 270 \u00b0C) compared to the unpromoted Fe-based catalyst (Fe/Al2O3). The presence of Na in the Fe-based catalyst improved the C1 formation rate at certain higher reaction temperatures (290 \u00b0C and 310 \u00b0C). Na was found to hinder the selectivity towards C2-C4 paraffins and C13+ hydrocarbons, including the oxygenates, and improve the formation of C2-C4 olefins and C5-C12 hydrocarbons at different reaction temperatures.\n               "}
{"full_text": "Data will be made available on request.The rational design of crystalline organic-inorganic hybrids such as metal-organic frameworks (MOFs) is an important challenge in materials science [1\u20133]. In the crystal engineering of extended coordination networks, N,N\u2032-ditopic ligands such as pyrazine (pyz) and 4,4\u2032-bipyridine (4,4\u2032-bipy) are often employed to link inorganic or metal-organic chains or layers into 2D and 3D structures [4\u201316]. Regarding pyz-pillared 3D materials, the most common types are those derived from metal-carboxylate [8\u201315] or cyanoheterometallic sheets (Hofmann-type MOFs with the general formula M1[M2(CN)4]) [16\u201326]. These pillared-layer MOFs have been investigated for applications involving, amongst others, supercapacitors [8], magnetism [9,10], gas storage and separation [11\u201314,17\u201320], switchable spin-crossover materials [20\u201324], iodine capture [25], and rechargeable alkaline batteries [26]. Other pyz-bridged 2D/3D solids include the copper(I) rhenate hybrid CuReO4(pyz) [27], transition metal nitroprusside derivatives of the type Fe1\u2212xTx(pyz)[Fe(CN)5NO] (T\u00a0=\u00a0Co, Ni, Cu) [28], metal carbonyl MOFs with the formula fac-M(CO)3(pyz)3/2 (M = Cr, Mo, W) [29], and the molybdenum oxide hybrid compound [Mo2O6(pyz)] [30].The two pyz-bridged Mo-containing materials mentioned above are interesting since they represent two extremes in terms of the metal oxidation state, i.e., from 0 in fac-Mo(CO)3(pyz)3/2 to VI in [Mo2O6(pyz)]. MOFs with low-valent metal nodes are rare and the former tricarbonyl derivative constituted the first crystallographically characterized example of a Mo0-based material [29]. The structure consists of fac-M(CO)3(pyz)3/2 coordination layers that stack along the a-axis, leading to the formation of hexagonal pore channels of 5.5\u20137.5\u00a0\u00c5 diameter (\nFig. 1a). The void space inside the pores is occupied by a non-coordinated pyz molecule, resulting in the final formula fac-Mo(CO)3(pyz)3/2\u00b71/2pyz (1). This compound crystallized in the triclinic space group P-1. Syntheses of 1 contained a small amount of a second crystalline phase with the formula fac-Mo(CO)3(pyz)3/2 (2), identified as a dense cubic phase consisting of two interpenetrating coordination networks (Fig. 1b). The structure of [Mo2O6(pyz)] (3) consists of layers of corner-sharing MoO5 square pyramids connected through pyz groups into a 3D covalently bonded metal oxide-organic ligand framework (Fig. 1c), in a manner analogous to that found in [MoO3(4,4\u2032-bipy)0.5] [30,31].The molybdenum-pyrazine coordination compounds described above are potentially interesting as (pre)catalysts for the following reasons: (i) during the last two decades, a broad variety of molybdenum carbonyl complexes, including tricarbonyl complexes such as [Mo(CO)3(L)\nn\n] [L =\u20091,2,4-triazole (trz, n\u2009=\u20093) or tris(1-pyrazolyl)methane (tpm, n\u2009=\u20091)], have revealed good performance as catalyst precursors for oxidation reactions, especially the epoxidation of olefins [32\u201336]; (ii) in several cases, the direct use of the aforementioned heteroleptic complexes as pre-catalysts leads to the formation of catalytically active crystalline metal oxide-organic ligand hybrids [37]; (iii) interesting catalytic behaviors have been found for a wide variety of molybdenum oxide-organic hybrids, e.g., 1,2,4-triazole-based hybrids displayed reaction-induced self-separation behavior when applied as catalysts for oxidation reactions [38\u201340]. With these considerations in mind, the present investigation was undertaken, in which the molybdenum-pyrazine compounds 1\u20133 were examined as (pre)catalysts for the oxidation of sulfides and the epoxidation of olefins, including biorenewable terpene and unsaturated fatty acid methyl ester (FAME) substrates.The following chemicals, reagents and solvents were obtained from Sigma-Aldrich (unless otherwise indicated) and used as received: (for synthesis) molybdenum hexacarbonyl (Fluka), molybdenum trioxide (Fluka), pyrazine (>99 %), toluene (99.9 %), acetonitrile (99.9 %, Riedel-de Ha\u00ebn), diethyl ether (99.8 %, Riedel-de Ha\u00ebn); (for catalytic tests) cis-cyclooctene (95 %, Alfa Aesar), methyl oleate (99 %), methyl linoleate (95 %, Alfa Aesar), dl-limonene (>95 %, Merck), cyclododecene (mixture of isomers, 96 %), methyl phenyl sulfide (99 %), diphenyl sulfide (98 %), 5.5\u2009M tert-butyl hydroperoxide in decane (<4 % water), \u03b1,\u03b1,\u03b1-trifluorotoluene (\u226599 %), acetone (99.5 %, Riedel-de Ha\u00ebn), and the internal standards methyl decanoate (99 %), undecane (>99 %) and mesitylene (98 %).Mo(CO)6 (0.16\u2009g, 0.60\u2009mmol) and pyz (52 equiv.) were added to a PTFE-lined stainless-steel autoclave (40\u2009mL capacity) in a glove box under an argon atmosphere. The autoclave was sealed and heated in an oven to 150\u2009\u00b0C at a ramp rate of 0.5\u2009\u00b0C\u2009min\u22121. After heating at this temperature for 40\u2009h, the autoclave was cooled to room temperature over a period of 16\u201324\u2009h, and the resultant dark shiny solid product was transferred to a Schlenk tube and washed with acetonitrile (3\u2009\u00d7\u200920\u2009mL) to remove excess pyz and residual Mo(CO)6. Finally, the solid was vacuum-dried at room temperature for 2\u2009h. Yield: 115\u2009mg, 56 % (based on Mo). Anal. Calcd for C9H6MoN3O3\u00b7C2H2N (340.15): C, 38.84; H, 2.37; N, 16.47 %. Found: C, 38.60; H, 2.38; N, 16.81 %.A mixture of Mo(CO)6 (0.48\u2009g, 1.80\u2009mmol) and pyz (10 equiv.) was placed under vacuum (ca. 0.1 bar) for 10\u2009min in a Schlenk tube. Toluene (30\u2009mL) was then added and the mixture was refluxed under a nitrogen atmosphere for 3\u2009h. The resultant dark precipitate was isolated by centrifugation, washed with acetonitrile (3\u2009\u00d7\u200920\u2009mL), and vacuum-dried at room temperature for 2\u2009h. Yield: 0.43\u2009g, 80 % (based on Mo). Anal. Calcd for C9H6MoN3O3 (300.11): C, 36.02; H, 2.01; N, 14.00; Mo, 31.97 %. Found: C, 35.72; H, 2.23; N, 13.89; Mo, 30.5 %.A mixture of MoO3 (0.11\u2009g, 0.76\u2009mmol), pyrazine (60\u2009mg, 0.76\u2009mmol) and Milli-Q water (15\u2009mL) was heated under autogenous pressure and dynamic conditions (20\u2009rpm) for 3 days at 160\u2009\u00b0C in a 23\u2009mL Teflon-lined stainless-steel autoclave. The resultant yellow-orange solid was collected by filtration, washed with Milli-Q water (2\u2009\u00d7\u200910\u2009mL), acetone (2\u2009\u00d7\u200910\u2009mL) and diethyl ether (2\u2009\u00d7\u200910\u2009mL), and finally vacuum-dried at room temperature for 2\u2009h. Yield: 0.10\u2009g, 71 % (based on Mo). Anal. Calcd for C4H4Mo2N2O6 (367.98): C, 13.06; H, 1.10; N, 7.61 %. Found C, 13.40; H, 1.15; N, 7.80 %.Elemental analysis for C, H and N was performed using a Truspec 630\u2013200\u2013200 instrument. Inductively coupled plasma atomic emission spectroscopy (ICP-AES) analyses (for Mo) were performed at the Central Testing Laboratory, University of Aveiro, using a Horiba JobinYvon Activa M spectrometer (detection limit of ca. 0.02\u2009mg\u2009dm\u22123; an experimental range of error of 5 %). Powder X-ray diffraction (PXRD) patterns were collected on an Empyrean PANalytical diffractometer (Cu-K\u03b1 X-radiation, \u03bb\u2009=\u20091.54060\u2009\u00c5) in a Bragg-Brentano para-focusing optics configuration (45\u2009kV, 40\u2009mA) at ambient temperature, using a spinning flat plate sample holder. Samples were step-scanned in the range from 5\u00b0 to 70\u00b0 (2\u03b8) with steps of 0.026\u00b0. A PIXEL linear detector with an active area of 1.7462\u00b0 was used with a counting time of 99\u2009s per step. Scanning electron microscopy (SEM) images and elemental mappings (Mo) were obtained on a Hitachi SU-70 SEM microscope equipped with a Bruker Quantax 400 detector operating at 20\u2009kV. Scanning transmission electron microscopy (STEM) images were collected using a Hitachi HD2700 microscope equipped with a Bruker EDS detector. The samples were prepared by depositing a drop of a suspension of the solid sample in ethanol onto holey amorphous carbon-film-coated 400 mesh copper grids (Agar Scientific). Thermogravimetric analysis (TGA) was performed under air using a HITACHI STA 300 system with a heating rate of 5\u2009\u00b0C\u2009min\u22121. The textural properties were determined from N2 sorption isotherms at \u2212196\u2009\u00b0C, which were measured using a Quantachrome instrument (automated gas sorption data using Autosorb IQ2). The samples were pre-treated at 60\u2009\u00b0C for 8\u2009h, under vacuum (< 4\u2009\u00d7\u200910\u22123 bar). The specific surface area (S\nBET) was calculated using the Brunauer, Emmett and Teller equation, and the total pore volume (V\np) was based on the Gurvitch rule (for a relative pressure (p/p\n0) of at least 0.99). The external/mesoporous surface area (S\next) and micropore volume (V\nmicro) were calculated using the t-plot method. Attenuated total reflectance (ATR) FT-IR spectra were measured on a Bruker Tensor 27 spectrometer equipped with a Specac Golden Gate Mk II ATR accessory having a diamond top plate and KRS-5 focusing lenses (resolution 4\u2009cm\u22121, 256 scans). Diffuse reflectance (DR) UV-Vis spectra were recorded at room temperature in the range 190\u2013900\u2009nm using a JASCO V-780 spectrophotometer equipped with a JASCO ISV-469 integrating sphere, with Spectralon as reference material. The spectra were collected in the reflectance mode, with a bandwidth of 2\u2009nm, a scan speed of 200\u2009nm\u00a0min\u22121, and a data pitch of \u2248\u20090.5.The catalytic reactions were carried out using 10\u2009mL borosilicate batch reactors equipped with a Teflon valve for sampling and a Teflon-lined magnetic stirrer. Initially, catalyst (in an amount equivalent to 18\u2009\u03bcmol Mo), \u03b1,\u03b1,\u03b1-trifluorotoluene (TFT) (1\u2009mL) and substrate (1.8\u2009mmol) were added to the reactor, which was then immersed in a temperature-controlled oil bath at 35 or 70\u2009\u00b0C under stirring (1000\u2009rpm). After 10\u2009min, the preheated oxidant tert-butyl hydroperoxide (TBHP) (2.75\u2009mmol for cis-cyclooctene (Cy8) and cyclododecene (Cy12), and 4.00\u2009mmol for bio-based olefins and sulfides) was added to the reactor, and this moment was taken as the initial instant of the catalytic reaction. The use of H2O2 as oxidant led to the complete degradation of the solids, forming solutions (no solid-state catalyst), and thus TBHP was preferred. Concerning the choice of solvent, TFT is appealing since it is readily available, relatively inexpensive, noncoordinating, and environmentally benign, and possesses a high boiling point and strong capacity to dissolve a large range of organic compounds [41]. It has been explored as a solvent in various olefin epoxidation catalytic systems, frequently yielding positive results [42,43].The evolution of the reactions was monitored by analyzing freshly prepared samples by gas chromatography (GC), using a Varian 450 GC instrument equipped with a BR-5 capillary column (30\u2009m\u2009\u00d7\u20090.25\u2009mm\u2009\u00d7\u20090.25\u2009\u00b5m) and a FID detector. The quantification of reactants/products was based on calibrations; the internal standards used were undecane for the substrates Cy8, Cy12 and dl-limonene (Lim), methyl decanoate for methyl oleate (MeOle) and methyl linoleate (MeLin), and mesitylene for methyl phenyl sulfide (MPS) and diphenyl sulfide (DPS). The values of initial catalytic activity (mmol\u2009gcat\n\u22121 h\u22121) and initial turnover frequency (mol\u2009molMo\n\u22121 h\u22121) were calculated based on substrate conversion at 12\u2009min reaction.The catalyst stability was evaluated by reusing the recovered solids in consecutive batch runs, keeping constant the initial mass ratio of catalyst:Cy8:TBHP between runs (the solids are denoted i-runn, where i is the catalyst identification number and n is the number of the batch run from which the solid was isolated). After each run, the solids were separated from the reaction mixture by centrifugation (3500\u2009rpm), thoroughly washed with acetone, dried overnight under atmospheric conditions, and finally vacuum-dried (ca. 0.1\u2009bar) at 60\u2009\u00b0C for 1\u2009h. The obtained solids were characterized by ATR FT-IR and DR UV-Vis spectroscopies, PXRD and/or SEM/STEM. Hot filtration (or leaching) tests were performed to check if soluble active species were present in the liquid phase for the system 3/TBHP/TFT/Cy8, at 70\u2009\u00b0C, under similar conditions to those used for a typical batch run. Specifically, at 1\u2009h reaction (in the presence of solid), the hot solid-liquid biphasic mixture (catalyst/TBHP/TFT/Cy8, at 70\u2009\u00b0C) was filtered through a 0.2\u2009\u00b5m PTFE membrane filter, and the hot filtrate was transferred to a separate preheated (70\u2009\u00b0C) reactor, after which stirring was continued at 70\u2009\u00b0C and GC was used to monitor any further reaction. In addition to the hot filtration tests, ambient filtration tests were performed as follows: at 30\u2009min (for used 1 and 2, referred to as 1-run1 and 2-run1) or 1\u2009h (for 3) reaction in the presence of the solid catalyst, the biphasic solid-liquid mixture (catalyst/TBHP/TFT/Cy8, at 70\u2009\u00b0C) was subjected to centrifugation (5\u2009min at 6000\u2009rpm and then 5\u2009min at 9000\u2009rpm) and afterwards filtration (using a 0.2\u2009\u00b5m PTFE membrane filter) at ambient temperature; the filtrate was transferred to a separate preheated (70\u2009\u00b0C) reactor, after which stirring was continued at 70\u2009\u00b0C and GC was used to monitor any further reaction.The influence of the reaction conditions on catalytic performance was evaluated for 1 and 3 using different amounts of catalyst (initial Mo:Cy8 molar ratio of 0.01, 0.005 and 0.0025, maintaining TBHP:Cy8\u2009= 1.5), oxidant (initial TBHP:Cy8 molar ratio of 1.5 and 2.2, maintaining Mo:Cy8\u2009=\u20090.01) and different temperatures (55 or 70 \u00b0C, with initial TBHP:Cy8 molar ratio\u2009=\u20091.5 and Mo:Cy8\u2009=\u20090.01).The tricarbonyl-pyrazine-molybdenum(0) MOF fac-Mo(CO)3(pyz)3/2\u00b71/2pyz (1) was obtained as a dark shiny crystalline solid in 56\u00a0% yield after heating a mixture of Mo(CO)6 and an excess of pyz in a Teflon-lined stainless-steel autoclave at 150\u2009\u00b0C for 40\u2009h. The PXRD pattern of 1 (\nFig. 2b) is in agreement with that reported by Voigt et al. for the same phase obtained by heating the reagents in a sealed ampoule rather than an autoclave [29]. When Mo(CO)6 was reacted with a 10-fold excess of pyz in refluxing toluene, an 80 % yield of the cubic phase, fac-Mo(CO)3(pyz)3/2 (2), was obtained, which had previously only been isolated by mechanical separation of single crystals formed as a minor secondary phase in the synthesis of 1 by the ampoule method [29]. The PXRD pattern of 2 (Fig. 2f) matches well with the simulated pattern calculated for the crystal structure (Fig. 2e), which confirms the purity of the as-synthesized MOF. The pyrazine-pillared molybdenum(VI) oxide hybrid [Mo2O6(pyz)] (3) was synthesized in 69 % yield by the hydrothermal treatment of a mixture of MoO3, pyz and H2O in the mole ratio 1:1:1100 at 160\u2009\u00b0C for 3 days. Previously, Liang et al. obtained 3 through a similar procedure except that pyrazine-2-carboxylic acid was employed as the pyz source rather than pyz itself, with the former undergoing thermal decarboxylation under the hydrothermal synthesis conditions [30]. The phase purity of 3 was confirmed by PXRD (Figs. 2i and 2j).Crystal morphologies were assessed by SEM (\nFig. 3). MOF 1 shows an irregular rod-like morphology with particle sizes up to 30\u2009\u00b5m in length, with some showing a hexagonal cross-section. The particles of 2 generally present an irregular and pseudo-spherical aspect with sizes up to 5\u2009\u00b5m. For the hybrid 3, the high crystallinity revealed by PXRD is confirmed by the SEM images, which show rectangular slab- or block-like crystallites with sizes in the range 5\u201320\u2009\u00b5m.The thermal stability of the materials was evaluated by TGA (\nFig. 4). For comparison, the TGA curve of neat pyz is shown. Pyz sublimes at room temperature and ambient pressure, and hence the TGA curve shows an abrupt and complete mass loss between 25 and 70\u2009\u00b0C (differential thermogravimetric maximum (DTGmax) at 66\u2009\u00b0C). Although the MOF 1 contains free pyz molecules in the hexagonal pore channels, the TGA curve under air does not show a low-temperature (<70\u2009\u00b0C) weight loss step corresponding to the removal of these weakly bound moieties. Instead, 1 starts to decompose above 70\u2009\u00b0C, showing two main overlapping weight loss steps with DTGmax values of 102\u2009\u00b0C and 120\u2009\u00b0C. The first of these is probably due to partial decarbonylation combined with removal of the free pyz molecules, while the second step is attributed to the decomposition of the remaining pyz molecules in the framework. Cubic 2 displays a similar behavior with DTGmax values of 97\u2009\u00b0C and 118\u2009\u00b0C. The molybdenum(VI) oxide hybrid 3 displays high thermal stability, with no weight loss being registered up to 335\u2009\u00b0C. Decomposition of the pyrazine pillars then takes place abruptly between this temperature and 380\u2009\u00b0C (DTGmax 367\u2009\u00b0C). Considering that complete decomposition of 1-3 under air leads to MoO3, the weights of the residues at 400\u2009\u00b0C for 1 (41.7 %), 2 (49.6 %) and 3 (78.5 %) are in agreement with the calculated values of 42.3 %, 48.0 % and 78.2 %, respectively.Materials 1-3 possessed relatively low specific surface areas; S\nBET in the range 9\u201358\u2009m2 g\u22121 (\nTable 1). MOF 1 possessed higher S\nBET than 2 (58 and 9\u2009m2 g\u22121, respectively) and some microporosity (V\nmicro =\u20090.012\u2009cm3 g\u22121); 2 did not possess microporosity. These results are somewhat consistent with the dense (cubic phase) framework of 2. Hybrid 3 did not possess microporosity, but the total pore volume was relatively high (V\np =\u20090.44\u2009cm3 g\u22121 compared to < 0.09\u2009cm3 g\u22121 for the MOFs), which may correspond to inter/intraparticle mesopores. The three materials did not exhibit well-defined pore size distribution curves, suggesting that they do not possess ordered open pore systems.The pyz-bridged materials 1-3 were explored as epoxidation (pre)catalysts, firstly in the model reaction of cis-cyclooctene (Cy8) using tert-butyl hydroperoxide (TBHP) as oxidant. The three materials led to biphasic solid-liquid mixtures. The influence of the reaction conditions on the performance of 2 and 3 was studied (Fig. S1 in the Supplementary data). For 2 with a reaction temperature of 70\u2009\u00b0C, an increasing amount of MOF (initial TBHP/Cy8 molar ratio kept constant at 1.5) led to increasing initial Cy8 conversion, although the initial catalytic activity (mmol\u2009gcat\n\u22121 h\u22121) and initial turnover frequency (TOF, mol\u2009molMo\n\u22121 h\u22121) decreased, following the order (1991\u2009mmol\u2009gcat\n\u22121 h\u22121 and 684\u2009mol\u2009molMo\n\u22121 h\u22121 for Mo:Cy8\u2009=\u20090.0025) >\u2009(1292\u2009mmol\u2009gcat\n\u22121 h\u22121 and 444\u2009mol\u2009molMo\n\u22121 h\u22121 for Mo:Cy8\u2009= 0.005) >\u2009(962\u2009mmol\u2009gcat\n\u22121 h\u22121 and 330\u2009mol\u2009molMo\n\u22121 h\u22121 for Mo:Cy8\u2009= 0.01). For the organometallic MOF 2, the in situ oxidative decarbonylation (discussed below) of the Mo(0) sites into oxidized molybdenum sites is required prior to the catalytic reaction. An increasing amount of 2 represents a decreasing initial molar ratio of TBHP:Mo, which may negatively impact on the oxidative decarbonylation rate of 2 (with TBHP) and consequently on the overall rate of olefin epoxidation. Nevertheless, in the studied range of Mo:Cy8 ratios, 2 led to 92\u201395 % epoxide (Cy8Ep) yield at 2\u2009h and 97\u2013100 % yield at 4\u2009h (the epoxide was always the only reaction product, i.e., Cy8Ep selectivity was 100 %). For 3, the initial activity and initial TOF were greater than zero solely for the highest Mo:Cy8 ratio (74\u2009mmol\u2009gcat\n\u22121 h\u22121 and 14\u2009mol\u2009molMo\n\u22121 h\u22121, respectively), and Cy8Ep yield at 24\u2009h increased in the order 69 %, 88 % and 100 % for Mo:Cy8\u2009=\u20090.0025, 0.005 and 0.01, respectively. As opposed to 2, the oxidative decarbonylation requirement does not apply for 3 (which is already in the oxidized form); a higher amount of 2 corresponds to a higher amount of oxidized metal sites from the initial instant of the catalytic reaction, enhancing the epoxidation reaction rate.Control experiments performed without catalyst or without oxidant showed negligible conversion, indicating that the catalytic reaction required the simultaneous presence of the molybdenum catalyst and the oxidant. This is consistent with mechanistic studies reported in the literature for molybdenum(VI)-catalyzed epoxidations, being generally accepted that acid-base reactions between the metal center (acting as a Lewis acid) and the hydroperoxide oxidant (acting as a base) occur to give active oxidizing species responsible for transferring an oxygen atom to the olefin molecule; this involves the nucleophilic attack of the olefin on an electrophilic oxygen atom of the oxidizing species (heterolytic mechanism), and the concomitant formation of the epoxide and tert-butyl alcohol (co-product of TBHP conversion) [44\u201346].For both 2 and 3 an increase in the amount of oxidant from TBHP/Cy8\u2009=\u20091.5 to TBHP/Cy8\u2009=\u20092.2 increased the epoxidation reaction rate significantly (e.g., conversion at 1\u2009h increased from 87 % to 94 % for 2, and from 24 % to 31 % for 3) (Fig. S1 in the Supplementary data). For a Mo:Cy8:TBHP molar ratio of 1:100:152, a decrease in the reaction temperature from 70\u2009\u00b0C to 55\u2009\u00b0C led to a significantly slower epoxidation reaction with 2 (initial activity =\u2009367\u2009mmol\u2009gcat\n\u22121 h\u22121, TOF =\u2009126\u2009mol\u2009molMo\n\u22121 h\u22121).Further catalytic experiments were performed with a reaction temperature of 70\u2009\u00b0C and a Mo:Cy8:TBHP molar ratio of 1:100:152. Under these established conditions, 1 led to faster epoxidation than 2 (e.g., leading to a quantitative yield of the epoxide at 2\u2009h, \nFig. 5). The slightly higher reaction rate for 1 than for 2 may stem from the different crystal structures, such as the dense framework of 2 vs. the more open framework of 1, which may result in different amounts of accessible molybdenum sites. Control tests performed with (i) the free organic ligand, (ii) the synthesis metal precursor Mo(CO)6, and (iii) the physical mixture of the free organic ligand and Mo(CO)6 (in molar amounts of metal and ligand equivalent to those added together with 1 and 2) led to homogeneous mixtures, and Cy8 conversion at 1\u2009h was 3 %, 41 % and 28 %, respectively, compared to 96 % for 1 and 87 % for 2. Hence, the MOFs give very active species when compared with their individual organic and/or inorganic components.There are few examples of monometallic tricarbonylmolybdenum(0)-based compounds tested for catalytic epoxidation of olefins. To the best of our knowledge, for the model reaction of Cy8/TBHP, only five mononuclear complexes (and no MOFs or hybrids) were previously reported. \nTable 2 compares the catalytic results of the previously studied complexes to those for MOFs 1 and 2. Under similar reaction conditions, 1 led to superior catalytic results to [Mo(CO)3(1,2,4-trz)3] (trz = triazole) and comparable results to [Mo(CO)3(1,2,3-trz)3] [35]. The mononuclear complexes [Mo(CO)3(tpm*)] (tpm* = tris(3,5-dimethyl-1-pyrazolyl)methane) and [Mo(CO)3(tpm)] (tpm = tris(1-pyrazolyl)methane) led to 96 % and 99 % conversion at 6\u2009h and 2\u2009h, respectively, at 55\u2009\u00b0C. Depending on the tricarbonyl metal complex, different species were formed under the epoxidation reaction conditions. Polymeric species with the empirical formula [MoO3(L)] were formed from [Mo(CO)3(L)3] with L =\u20091,2,3-trz and 1,2,4-trz [35], and the hexamolybdate salt [{MoO2(tpm*)}2(\u00b52-O)][Mo6O19] was formed from [Mo(CO)3(tpm*)] [48]. The oxidizing species formed were stable in consecutive catalytic batch runs, somewhat in parallel to that verified for 1 and 2 (discussed below).As found for 2, the reaction mixture with 1 as (pre)catalyst was biphasic solid-liquid. The solids were recovered and reused in two consecutive batch runs, which led to similar kinetic curves for each type of material (runs 2 and 3 in Figs. 5A and 5B). There was a slight fall off in initial activity between runs 1 and 2 for both catalysts: from 1142\u2009mmol\u2009gcat\n\u22121 h\u22121 to 983\u2009mmol\u2009gcat\n\u22121 h\u22121 for 1, and from 962\u2009mmol\u2009gcat\n\u22121 h\u22121 to 820\u2009mmol\u2009gcat\n\u22121 h\u22121 for 2. For 2, the kinetic curves for runs 2 and 3 merged with that for the first run for reaction times longer than 30\u2009min, while for 1 complete conversion in runs 2 and 3 took slightly longer (4\u2009h) than in run 1 (2\u2009h).The used catalysts (from 1 and 2) were characterized by PXRD, SEM/STEM, ATR FT-IR and UV-Vis spectroscopies. PXRD showed that the recovered solids were X-ray amorphous (Fig. 2). Morphological changes occurred during the first catalytic run with 1 and 2, leading to solids with irregular particle sizes, which included aggregates of nanoparticles (Fig. S2 in the Supplementary Data). The ATR FT-IR spectra of the used catalysts suggest that chemically similar species were formed from 1 and 2 (1-run1 and 2-run1, \nFig. 6). The loss of the strong \u03bd(CO) bands near 1765 and 1890\u2009cm\u22121, coupled with the appearance of a band at about 940\u2009cm\u22121 (assigned to \u03bd(MoO)) and very broad bands in the spectral range 400\u2013800\u2009cm\u22121 (assigned to \u03bd(Mo\u2013O\u2013Mo)), shows that the precursor materials underwent in situ oxidative decarbonylation. Changes in the spectral range 1000\u20131500\u2009cm\u22121 (associated mainly with internal pyz modes) indicate alterations in the metal-organic ligand coordination modes. The similar spectra found for 1-run1 and 2-run1, and 1-run3 and 2-run3, correlate with the roughly coincident kinetic curves of the used solids (Fig. 5). Accordingly, 1-run1 and 2-run1 gave similar UV-Vis diffuse reflectance spectra consisting of a relatively narrow and intense peak at 250\u2009nm, with a shoulder on the high energy side at about 200\u2009nm, and overlapping shoulders on the low energy side with maxima at about 300 and 335\u2009nm (Fig. S4 in the Supplementary data). The absence of absorption peaks between 400 and 800\u2009nm indicates that practically all molybdenum centers are oxidized to MoVI and no Mo0 or other reduced forms are present (cf. the parent materials 1 and 2 (Fig. S4), which display a strong, very broad absorption band across the whole visible region, consistent with the black color of the solids). Hence, the absorption bands between 250 and 350\u2009nm are assigned to ligand to metal charge transfer transitions (O2\u2013 \u2192 Mo6+) involving oxygens in bridging (Mo\u2013O\u2013Mo) and terminal (MoO) positions [49,50]. The bands between 300 and 350\u2009nm are especially indicative of a molybdenum oxide substructure containing connected molybdenum(VI)-oxo species in which the Mo6+ centers are octahedrally coordinated. The absorption bands exhibited by 1-run1 and 2-run1 may also have contributions from the pyrazine ligand which, in its free form, displays a band at about 255\u2009nm, assigned to the \u03c0-\u03c0* electronic transition of the aromatic ring, and a broad band beyond 275\u2009nm (\u03bb\nmax 320\u2013340\u2009nm), assigned to the forbidden electronic transition n-\u03c0* [51].Hot filtration (leaching) tests with the recovered solids 1-run1 and 2-run1 led to roughly comparable results to the respective normal batch run in the presence of solid catalyst (Fig. S3 in the Supplementary data). These results may be partly due to the difficult separation of the nanoparticles, which may pass through the membrane filter with a 200\u2009nm pore size. As an alternative, ambient separation tests involving centrifugation and filtration operations were carried out for 1-run1 and 2-run1 (Fig. S3). The recovered solids were used instead of the original MOFs 1 and 2 because during run 1 the latter are converted to the respective oxidized compounds 1-run1 and 2-run1, and thus one cannot exclude the possibility of the reaction mixture of batch run 1 containing side products of the conversion of the original catalyst. The filtrates were left to react further at the catalytic reaction temperature, which led to smaller increments in Cy8 conversion between 0.2\u2009h and 4\u2009h reaction compared with those without solid catalyst separation (i.e., normal batch run 2): for 1-run1, 16 % increment in conversion compared to 33\u00a0% without solid catalyst separation; for 2-run1, 17 % increment in conversion compared to 44 % without solid catalyst separation. These results suggest that the solid catalysts contributed to the catalytic reactions. ICP-AES analyses of the respective filtrates indicated that the amount of molybdenum was 209\u2009mg\u2009dm\u22123 for 1 and 119\u2009mg\u2009dm\u22123 for 2, which corresponds to ca. 2 and 1\u2009mol%, respectively, of the initial amount of molybdenum added to the reactor.Catalyst reuse was also explored for the hybrid 3 (Fig. 5C). The kinetic curves are practically identical for the three consecutive batch runs, and epoxide selectivity remained 100\u00a0%. The initial catalytic activity was 74, 109 and 103\u2009mmol\u2009gcat\n\u22121 h\u22121 for runs 1, 2 and 3, respectively. Characterization of the catalyst recovered after each run by PXRD and ATR FT-IR spectroscopy indicated that the structural and chemical features of 3 were preserved (Fig. 2 and Fig. 6). Morphologically, the particle sizes seemed to decrease, which may have contributed to the slight increase in activity observed in consecutive batch runs (Fig. S2 in the Supplementary Data). A hot filtration test at 1\u2009h reaction for 3 led to a slower reaction without the solid, but conversion was nevertheless significant (50 % at 6\u2009h without solid, compared with 84 % in the presence of the solid catalyst, Fig. S3), suggesting that the system may have a homogeneous catalytic contribution and/or very small solid particles that could not be separated by the membrane filter. The ambient leaching test, which involved centrifugation and filtration operations, was carried out. The filtrate was left to react further at the catalytic reaction temperature, leading to 9 % increment in Cy8 conversion between 1\u2009h and 6\u2009h reaction, compared to 60\u00a0% without solid catalyst separation (Fig. S3), suggesting that the solid catalyst 3 contributed to the catalytic reaction. ICP-AES analyses of the respective filtrate indicated that the amount of molybdenum was 3\u2009mg\u2009dm\u22123, which corresponds to only ca. 0.03\u2009mol% of the initial amount of molybdenum added to the reactor.The principal structural motif of 3, comprising perovskite-like layers of corner-sharing {MoO5N} octahedra, is found in a few other organically modified molybdenum(VI) oxide hybrids, namely those containing pyridine [52], 4,4\u2032-bipy [31], 1,2,3-triazole (1,2,3-trz) [53], 1,2,4-triazole (1,2,4-trz) [31] and 4-(1,2,4-triazol-4-yl)benzoic acid (trPhCO\n\n2\n\nH) [54] building blocks. The last four of these have been tested as catalysts for the epoxidation of Cy8 with TBHP. Under the same reaction conditions (70\u2009\u00b0C, TFT cosolvent, 1\u2009mol% Mo, 1.5 equiv. TBHP), the hybrids 3 and [MoO3(1,2,3-trz)0.5] [35] led to identical results (83\u201384 %/100 % Cy8Ep yield at 6/24\u2009h), which were superior to those obtained with [MoO3(1,2,4-trz)0.5] (54\u00a0%/91\u00a0% Cy8Ep yield at 6/24\u2009h) (Table S1 in the Supplementary data). The hybrid [Mo4O12(trPhCO\n\n2\n\nH)2]\u00b70.5\u2009H2O was only tested at 55\u2009\u00b0C, giving 44 %/82 % epoxide yield at 6/24\u2009h [54], while [Mo2O6(4,4\u2032-bipy)] led to 99 % epoxide yield (99 % selectivity) after 2\u2009h in refluxing CHCl3 (using a high catalyst amount of 2.7\u2009mol% Mo) [55].\nTable S1 compares the performance of 3 for the model reaction of Cy8 with other polymeric hybrids of the type [Mo2O6(L)\nx\n] in which the Mo(VI) oxide subtopologies consist of 1D chains or ribbons. Under roughly similar reaction conditions, and based on Cy8 conversion at 6\u2009h, 3 performed better than [Mo2O6(Htrgly)]\u00b7H2O (Htrgly = 2-(4H-1,2,4-triazol-4-yl)acetic acid; 59\u00a0% conversion) [40], [Mo2O6(trpzH)(H2O)2] (trpzH = 4-(3,5-dimethyl-1H-pyrazol-4-yl)-1,2,4-triazole; 74 % conversion), [Mo2O6(m-trtzH)(H2O)2] (m-trtzH = 5-[3-(1,2,4-triazol-4-ylphenyl)]-1H-tetrazole; 21\u00a0% conversion), [Mo2O6(p-tr2Ph)]\u00b7H2O (p-tr2Ph = 1,4-phenylene-4,4\u2032-bis(1,2,4-triazole); 17 % conversion), and [Mo2O6(tr2ad)]\u00b7H2O (tr2ad = bis(1,2,4-triazol-4-yl)adamantane; 18\u00a0% conversion) [56], and worse than [Mo2O6(trethbz)2]\u00b7H2O (trethbz = (S)-4-(1-phenylpropyl)-1,2,4-triazole; 98 % conversion) [54]. Under different reaction conditions, higher conversion was reported for [Mo2O6(pent-pp)] (pent-pp = 2-(1-pentyl-3-pyrazolyl)pyridine; 98\u00a0% conversion at 6\u2009h, 55\u2009\u00b0C, Mo:Cy8:TBHP molar ratio =\u20091:113:172), but this material suffered loss of activity after run 1 and was chemically and structurally unstable [57]. Generally, the literature for the [Mo2O6(L)\nx\n] family of hybrids lacks detailed catalyst stability studies, making it difficult to establish fair comparisons with other catalysts.Catalysts 1-3 were further examined for the epoxidation of other olefins, including the bio-olefins dl-limonene, methyl oleate (MeOle) and methyl linoleate (MeLin), and the reaction scope was expanded to include sulfoxidation (Table S2 in the Supplementary data).\ndl-Limonene is the racemic mixture of d-limonene and l-limonene, which are the main compounds in citrus and pine needle essential oils, respectively. Upgrading of limonene by catalytic oxidation is a critically important pathway towards its valorization as a renewable platform chemical [58]. The epoxidation of dl-limonene gives mono- and diepoxides, namely 1,2-epoxy-p-menth-8-ene (LimEp) and 1,2:8,9-diepoxy-p-methane (LimDiEp) (\nScheme 1), which can undergo ring-opening reactions to give a wide range of oxygenated derivatives, from simple limonene-1,2-diols (LimDiol), which are useful precursors of bioactive compounds [59,60], to cyclic carbonates [61,62] and a broad variety of bio-based polymers [63,64], such as limonene-based non-isocyanate polyurethanes [61,62]. The reaction of dl-limonene was fast in the presence of 1 and 2, giving 100 % conversion at 2\u2009h and 4\u2009h, respectively (\nFig. 7). Mono and diepoxide products were mainly formed in a total yield of 67 % at 2\u2009h for 1 (LimEp/LimDiEp molar ratio of 3.2) and 66 % at 4\u2009h for 2 (LimEp/LimDiEp molar ratio of 5.7), and LimDiol was formed via LimEp epoxide ring-opening, in 27 % and 28 % yield, respectively. Without catalyst, conversion was only 6 % at 24\u2009h. To the best of our knowledge, the only tricarbonylmolybdenum(0)-based compound that has previously been reported as a (pre)catalyst for the epoxidation of limonene is [Mo(CO)3(1,1,1-tris(methylaminomethyl)ethane)], which led to 18 % conversion of (R)-(+)-limonene at 24\u2009h/55\u2009\u00b0C [47]. Although the reaction of dl-limonene was slower in the presence of 3, the combined epoxide selectivity was much higher, with LimEp and LimDiEp being obtained in 64 % and 36 % yield, respectively, at 24\u2009h (LimEp/LimDiEp molar ratio\u2009=\u20091.8) (\nFig. 8).Methyl oleate and methyl linoleate are the methyl esters of oleic acid and linoleic acid, which are the major components of most vegetable oils. The epoxides of fatty acid methyl esters (FAMEs) have widespread use as solvents, lubricants, PVC stabilizers and plasticizers, and as intermediates for the synthesis of biobased polyols for polyurethane production [65,66]. With MeOle as substrate, 1-3 led to 94 %/99 %, 88 %/100 % and 32 %/94 % conversion, respectively, at 4\u2009h/24\u2009h, and methyl 9,10-epoxyoctadecanoate (MeOleEp) was formed with 100\u00a0% selectivity (Fig. 7 and Fig. 8, Scheme 1). The reaction of the diene MeLin in the presence of 3 gave mainly the monoepoxide isomers and regioselectivity towards the epoxidation of the 9,10 and 12,13\u2009CC double bonds was similar. Specifically, methyl 9,10-epoxy-12-octadecenoate and methyl 12,13-epoxy-9-octadecenoate (MeLinEp) were formed in equimolar amounts with a total yield of 58 % at 69 % conversion, 24\u2009h (\nScheme 2). The total monoepoxide selectivity decreased slightly between 2\u2009h and 24\u2009h reaction (from 100 % at 12 % conversion to 85 % at 69 % conversion) owing to the formation of the diepoxide methyl 9,10\u201312,13-diepoxyoctadecanoate (10 % MeLinDiEp yield at 24\u2009h). For 1 and 2, the MeLinEp yields at conversions of 71\u201374 % (4\u2009h) and 85\u201391 % (24\u2009h) were 50 %/36 % for 1 and 55 %/49 % for 2 (Fig. 7). The decreasing yields of the monoepoxides (MeLinEp) were accompanied by increasing diepoxide yields (16 %/25 % for 1 and 16 %/31 % for 2 at 4\u2009h/24\u2009h) and increasing total yields of cyclization products (8 %/30 % for 1 and 0 %/5 % for 2 at 4\u2009h/24\u2009h). The cyclization products were methyl 10,13-epoxy-9,12-dihydroxyoctadecanoate and methyl 9,12-epoxy-10,13-dihydroxyoctadecanoate, which may be formed via intramolecular cyclization of the epoxydiol intermediates (Scheme 2). The latter may be relatively unstable due to the proximity of the diol and epoxy ring in the molecule (separated by a methylene group) and thus may not be present in measurable amounts in the reaction medium [67\u201372].For the two FAMEs, no reaction occurred without catalyst, at 24\u2009h. It is noteworthy that for 1 and 2 the slightly higher catalytic activity observed for 1 with Cy8 as substrate is maintained for the bio-derived olefins. In compliance with the results for Cy8, 1 and 2 display much higher catalytic activity than 3 in the reactions of the bio-derived olefins.To the best of our knowledge, only three hybrid materials of the type [Mo2O6(L)\nx\n] have been reported as catalysts for biomass-derived olefins, namely, [Mo2O6(2,2\u2032-bipy)] [71], [Mo2O6(trethbz)2]\u00b7H2O [54] and [Mo2O6(pent-pp)] [57] (Table S1 in the Supplementary data). Based on total epoxide selectivity at high conversions (>90 %) of MeOle and dl-limonene, 3 seems to exhibit the most promising catalytic performance. The MeLin reaction was somewhat slower for 3 than for the 1D hybrid [Mo2O6(trethbz)2]\u00b7H2O, which may be due to an interplay of several factors such as differences in the catalysts\u2032 structural dimensionality and solubility in the medium containing the olefin.Catalyst 3 was also studied for the epoxidation of cyclododecene (\nScheme 3). The reaction gave 100 % selectivity to 1,2-cyclododecane epoxide at 86 % conversion, 24\u2009h. Without catalyst, the reaction was sluggish (7 % conversion at 24\u2009h). Comparing the results for cis-cyclooctene and cyclododecene, the reaction is slower for the olefin with the larger ring size, suggesting that steric hindrance effects may be important.To assess the capacity of 1-3 to promote sulfoxidation, the catalytic oxidation of methyl phenyl sulfide (MPS) and diphenyl sulfide (DPS) was studied under mild conditions, at 35\u2009\u00b0C and atmospheric pressure. Without catalyst, MPS and DPS conversion was 66\u00a0% and 69\u00a0%, respectively, at 24\u2009h, and only the corresponding sulfoxide was formed. Catalysts 1-3 promoted the one-pot conversion of sulfide-sulfoxide-sulfone. Thus, 1 and 2 led to quantitative yields of the sulfones after 24\u2009h reaction, while the sulfone yields for 3 were 84 % from MPS and 95 % from DPS (\n\nSchemes 4 and 5). The molybdenum-catalyzed sulfoxidation of sulfides (or sulfoxides) may involve a proton transfer of the hydroperoxide oxidant to an oxido ligand (MoO) forming a Mo-OH group and an additional \u03b7\n1-hydroperoxido ligand (Mo-OOH). Interactions between the resultant oxidizing species and the sulfur-containing nucleophile (via an oxygen atom of Mo-OOH) leads to O-O bond rupture and formation of a SO bond, giving the sulfoxide (or sulfone) product, and tert-butanol [72].In this work, molybdenum-pyrazine coordination compounds were investigated for the first time in the catalytic epoxidation of olefins and sulfoxidation of sulfides. The organometallic network solids with the framework composition [Mo(CO)3(pyz)3/2] (1, 2) and the polymeric oxomolybdenum hybrid [Mo2O6(pyz)] (3) promoted the epoxidation of cis-cyclooctene, cyclododecene and bio-based olefins (fatty acid methyl esters and terpene substrates), and sulfoxidation reactions, giving products with interesting applications. In general, the reactions were highly selective to the epoxides (in the case of olefins) and sulfones (in the case of sulfides). Compounds 1-3 led to biphasic solid-liquid reaction mixtures. While the solid phase with 3 showed retention of the crystalline structure and could be recovered and reused without loss of activity, the reaction seemed to be partly homogeneously catalyzed by dissolved metal species. On the other hand, the tricarbonyl-pyrazine-molybdenum(0) compounds 1 and 2 were converted in situ to a polymeric oxomolybdenum catalyst which was more active than 3 and performed steadily in consecutive batch runs. The amorphous nature of the oxomolybdenum catalyst (derived from the MOFs) hampers its structural elucidation. The compound warrants, nevertheless, further study due to the convenience of 1 and 2 as catalyst precursors, especially the cubic phase 2 which can be readily prepared on a large scale.\nDiana M. Gomes: Methodology, Validation, Investigation, Writing - Original Draft. Andreia F. Silva: Methodology, Validation, Investigation, Writing - Original Draft. Ana C. Gomes: Validation, Investigation, Writing - Original Draft. Patr\u00edcia Neves: Project administration, Validation, Investigation, Writing - Original Draft. Anabela A. Valente: Resources, Conceptualization, Supervision, Writing - Review & Editing. Isabel S. Gon\u00e7alves: Resources, Conceptualization, Supervision, Writing - Review & Editing. Martyn Pillinger: Resources, Funding acquisition, Project administration, Visualization, Writing - Review & Editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was developed within the scope of the project CICECO-Aveiro Institute of Materials, UIDB/50011/2020, UIDP/50011/2020 & LA/P/0006/2020, financed by national funds through the FCT (Funda\u00e7\u00e3o para a Ci\u00eancia e a Tecnologia)/MCTES (Minist\u00e9rio da Ci\u00eancia, Tecnologia e Ensino Superior) (PIDDAC). We acknowledge support and funding provided within the CENTRO 2020 Regional Operational Program (project references CENTRO-01\u20130145-FEDER-028031 and PTDC/QUIQOR/28031/2017) and the COMPETE 2020 Operational Thematic Program for Competitiveness and Internationalization (POCI-01\u20130145-FEDER-030075), co-financed by national funds through the FCT/MEC (Minist\u00e9rio da Educa\u00e7\u00e3o e Ci\u00eancia) and the European Union through the European Regional Development Fund under the Portugal 2020 Partnership Agreement. D.M.G. (grant ref. 2021.04756.BD) acknowledges the FCT for a PhD grant (State Budget, European Social Fund (ESF) within the framework of PORTUGAL2020, namely through the Centro 2020 Regional Operational Program). A.C.G. thanks the FCT/MCTES for funding through the Individual Call to Scientific Employment Stimulus (CEECIND/02128/2017).Supplementary data associated with this article can be found in the online version at doi:10.1016/j.cattod.2023.114050.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  The horizons of epoxidation and sulfoxidation processes may be expanded by developing new, efficient, and versatile catalysts. In the present work, three pyrazine-bridged molybdenum(0/VI)-based coordination network solids have been investigated for the epoxidation of olefins and the oxidation of sulfides. The materials studied were the Mo0-based metal-organic framework (MOF) fac-Mo(CO)3(pyz)3/2\u00b71/2pyz (1) with a structure consisting of stacked fac-Mo(CO)3(pyz)3/2 coordination layers, the cubic phase fac-Mo(CO)3(pyz)3/2 (2) with a dense framework consisting of two interpenetrating coordination networks, and the molybdenum oxide-pyrazine hybrid material [Mo2O6(pyz)] (3) with a structure consisting of perovskite-like MoO3 layers pillared by pyz molecules. In the model reaction of cis-cyclooctene with tert-butyl hydroperoxide (TBHP) at 70\u00a0\u00b0C, quantitative yields of the epoxide were obtained within 2\u00a0h for 1, 4\u00a0h for 2, and 24\u00a0h for 3. Catalysts 1-3 were further examined for the epoxidation of other olefins, including the bio-olefins dl-limonene, methyl oleate and methyl linoleate, and the reaction scope was expanded to include the oxidation of sulfides. In the reactions of the bio-olefins, 3 was highly selective, giving only diepoxide and/or monoepoxide products. While the tricarbonyl-pyrazine-molybdenum(0) compounds displayed higher activity, by-products were obtained in the reactions of dl-limonene and methyl linoleate, namely limonene-1,2-diol and hydroxytetrahydrofuran cyclization products, respectively. Catalysts 1-3 displayed high activity for the selective oxidation of sulfides (methyl phenyl sulfide and diphenyl sulfide) to sulfones under mild conditions (35\u00a0\u00b0C).\n               "}
{"full_text": "The electrochemical oxygen evolution reaction (OER) is a key component of promising routes for clean energy, such as hydrogen production via water electrolysis, regenerative fuel cells, and electrochemical CO2 reduction (CO2RR) [1\u20136]. The OER, which involves four coupled electrons and protons, exhibits slow reaction kinetics and high overpotentials that limit the process efficiencies [7\u20139]. Thus, intensive efforts have been directed toward developing electrocatalysts that enhance the kinetics of the OER. Precious metals such as Ir and Ru are regarded as the best electrocatalysts for the OER in acidic media [10\u201313].Particularly, numerous studies have focused on Ru-based materials owing to Ru having a higher activity and lower price than those of Ir [14\u201317]. Electrochemically produced hydrous RuO\nx\n shows a remarkable OER catalytic activity because the abundant unsaturated Ru leads to the participation of large amounts of lattice oxygen in the reaction [18\u201320]. The oxygen participating through the lattice oxygen oxidation mechanism (LOM) significantly reduces the theoretical overpotential of the OER, leading to an enhanced catalytic activity [21,22]. However, hydrous RuO\nx\n dissolves to a significant extent in the acidic medium during the OER, resulting in a low stability [23\u201325]. To suppress the dissolution and increase the stability of Ru-based electrocatalysts, crystalline rutile RuO2 prepared by thermal treatment, which possesses stable lattice oxygen, has been proposed as an OER catalyst. Crystalline RuO2 shows a conventional adsorbate evolution mechanism (AEM) for the OER, which decreases the dissolution to provide a good stability, but leads to a lower catalytic activity than that of hydrous RuO\nx\n.A recently proposed strategy to improve the activity and stability simultaneously is to introduce another element that can significantly modify the electronic structure of the Ru oxide. Alloying with Ir is one of the typical approaches for this purpose [23,26\u201332]. Qiao\u2019s group reported a nanocrystalline Ru@IrO\nx\n catalyst with an increased valence state of the Ir in the shell and a decreased valence state of the Ru in the core [27]. This charge redistribution in this structure enhanced the catalytic activity and stability in acidic conditions. Recent studies have reported that other elements, such as Pt, Y, Cr, Te, and Ni, can modify the electronic structure of Ru oxide catalysts adequately [18,33\u201336]. Lin et al. synthesized a rutile Cr0.6Ru0.4O2 catalyst with a higher Ru oxidation state than that of RuO2 and a superior OER activity and stability [37]. Density functional theory (DFT) calculations demonstrated that this electronic structure reduces the energy barriers for the formation of *OOH. Huang\u2019s research group reported Ru-Ni oxide nanosheets with a downshifted d-band center electronic structure [38]. According to DFT calculations, a modified d-band electronic structure and transformed t2g and eg orbitals reduce the energy barriers for O2 formation. However, few studies have focused on the real behavior and role of the introduced elements in the reaction [39].Here, we investigated the mechanisms through which Ni introduced in Ru oxide enhances the catalytic activity for the OER. A nanosized Ru and Ni oxide (denoted as RuNiO\nx\n) electrode was manufactured via a simple dip coating process modified with a capping agent followed by a thermal treatment. The RuNiO\nx\n electrode exhibited a better OER performance than that of a RuO\nx\n electrode and was stable for 100\u00a0h. The electronic structure of the RuNiO\nx\n electrode is close to that of a hydrous RuO\nx\n electrode, which possesses oxygen vacancies. The results of in-situ/operando X-ray absorption near-edge structure spectroscopy (XANES), near-edge X-ray absorption fine structure spectroscopy (NEXAFS), and on-line inductively coupled plasma mass spectrometry (ICP-MS) analyses demonstrated that the Ni distorts the Ru oxide structure at anodic potential, increasing the amount of oxygen vacancies. This phenomenon increases the participation of lattice oxygen, leading to an enhanced OER catalytic activity of the RuNiO\nx\n electrode.All chemicals were of analytical grade and were used without further purification. A commercial Ti foam was purchased from Alantum Corporation. KOH, Ni(NO3)2\u00b76H2O, and polyvinylpyrrolidone (PVP) were purchased from Sigma-Aldrich. RuCl3 was obtained from Alfa Aesar. H2SO4 and ethanol were purchased from Samchun Chemical Co., ltd. (Korea). Deionized water was obtained using an Arium\u00ae Mini Ultrapure Water System (Sartorius AG).The RuNiO\nx\n electrode was prepared on a Ti foam via a dip coating procedure, which can be easily applied in industrial processes (Fig. S1). To increase the active surface area of the electrode, PVP was used as a surfactant. A piece of Ti foam (TF, 1\u00a0cm\u00a0\u00d7\u00a02\u00a0cm) was washed with ethanol and dried in nitrogen. Ni(NO3)2\u00b76H2O (5\u00a0mmol), RuCl3 (5\u00a0mmol), and 50\u00a0mg of PVP were dissolved in 9\u00a0mL of ethanol and 1\u00a0mL of water. The TF was submerged in the solution and shaken for 20\u00a0s. The TF was then dried in a convection oven at 70\u00a0\u00b0C. The coated TF was calcined in air at 450\u00a0\u00b0C for 1\u00a0h; the sample was denoted as RuNiO\nx\n. A NiO\nx\n electrode was prepared using an identical procedure to that of the RuNiO\nx\n electrode, except that the RuCl3 was replaced with Ni(NO3)2\u00b76H2O. A RuO\nx\n electrode was prepared with the same procedure, except that the Ni(NO3)2\u00b76H2O was replaced with RuCl3. Finally, a RuO\nx\n-L electrode was prepared via the same procedure, but without the PVP.The electrochemical tests were performed in a standard three-electrode system with a potentiostat (VSP, BioLogic) using an Ag/AgCl (3.5\u00a0M KCl) electrode and a graphite electrode as the reference and counter electrodes, respectively. All potentials were calibrated to the reversible hydrogen electrode (RHE) scale. Electrochemical impedance spectroscopy (EIS) measurements were conducted at 1.5\u00a0V vs. RHE in the frequency range from 1\u00a0Hz to 100\u00a0kHz.The morphologies of the prepared electrodes were studied via high-resolution scanning electron microscopy (HR-SEM, Regulus 8230, Hitachi). X-ray diffraction (XRD) measurements were conducted using an Empyrean diffractometer (Malvern Panalytical) equipped with a Cu K-alpha radiation source. X-ray photoelectron spectroscopy (XPS, PHI 5000 VersaProbe, ULVAC-PHI, Inc.) was used to investigate the chemical states in the electrodes. To analyze the electronic structures, synchrotron-based NEXAFS measurements were conducted at the 10D beamline of the Pohang Accelerator Laboratory (PAL, Pohang, South Korea). X-ray absorption spectroscopy (XAS) analyses were performed at the 1D beamline of the PAL.In situ/operando XAS analyses were conducted using a homemade electrochemical flow cell. The catalyst-coated TF electrodes were placed on a thin carbon-coated Kapton film surface and contacted with a 0.05\u00a0M H2SO4 electrolyte. The graphite and Ag/AgCl electrodes were installed on the back side of the flow cell. The set-up details were reported in a previous paper [40].On-line ICP-MS analyses were performed using the electrochemical flow cell. The dissolution was quantified by mixing 1\u00a0ppm Re in 1\u00a0M HNO3 as an internal standard. The electrode area was 0.1\u00a0cm2 for the on-line ICP-MS measurement. The set-up details were reported in a previous paper [40]. The working electrodes were the prepared Ru and Ni electrodes, and a 0.05\u00a0M H2SO4 solution was used as an electrolyte. The Ag/AgCl electrode and a Pt wire were used as the reference and counter electrodes, respectively.A membrane electrode assembly (MEA; active area: 10\u00a0cm2) was used as the electrochemical cell for the CO2RR tests. An Ag black (APS 20\u201340\u00a0nm, 99.9% in metal basis, Alfa Aesar) pasted gas diffusion layer electrode was used as the cathode, and the prepared RuNiO\nx\n and RuO\nx\n electrodes and a commercial IrO2 electrode were used as anodes. The membrane (Sustainion\u00ae X37-50 Grade RT, Dioxide Materials) was pretreated in a 1\u00a0M KOH solution for 48\u00a0h and washed with deionized water several times. The MEA was then placed in a homemade CO2 electrolyzer single cell with pin-type and serpentine flow channels on the anode and cathode sides, respectively. Humidified CO2 heated to 90\u00a0\u00b0C with a heating mantle was used as the reactant gas; the flow rate was 200\u00a0cm3 and the gas was fed to the cathode side. A 0.1\u00a0M KHCO3 solution was flowed through the anode side using a pump operating at 10\u00a0r min\u22121. The electrochemical single-cell tests were performed using a potentiostat (BioLogic, VSP, VMP3B-10), which can measure up to 10 A at room temperature. The outlet gas was analyzed via gas chromatography (GC, Agilent 7890A). The GC inlet was connected to the cathode through a water trap. Argon (99.999%) was used as the carrier gas. Hydrogen was detected using a thermal conductivity detector, and carbon compounds, such as CO and CH4, were detected using a flame ionization detector (FID). To enhance the detection of CO, a methanizer was installed before the FID. The partial current densities of the products were determined using the total measured current and the volume of each product calculated from the GC peaks.RuNiO\nx\n electrode was prepared on Ti foam through dip coating process, which is easily used in industrial process (Fig. S1). The Ti foam substrate was immersed and withdrawn in the Ru and Ni salt contained solution and then calcined in air at 450\u00a0\u00b0C for 1\u00a0h to make stable lattice oxygen. In order to increase catalytic active surface area of electrode, PVP as a surfactant was added in the solution for dip coating. Fig. 1\n(a) shows SEM images of the RuNiO\nx\n electrode. The particles on the electrode surface have diameters of approximately 10\u201320\u00a0nm and are uniformly distributed on the Ti foam substrate. Similar morphologies were observed for the RuO\nx\n and NiO\nx\n electrodes (Figs. S2\u2013S4). Without PVP in solution, the achieved Ru electrode (denoted as RuO\nx\n-L) had a larger grain size than those of the other electrodes, indicating that the added PVP successively decreases the grain size in the dip coating process (Fig. 1b and Fig. S5). Because of the thin RuNiO\nx\n layer, the XRD pattern of the RuNiO\nx\n electrode only exhibits a peak corresponding to the Ti foam substrate. Thus, we manufactured carbon-supported RuNiO\nx\n, RuO\nx\n, and NiO\nx\n (RuNiO\nx\n/C, NiO\nx\n/C, and NiO\nx\n/C, respectively) powders via similar synthesis processes to those of the electrodes, but using carbon black (Ketjenblack 300\u00a0J, Nouryon) instead of the Ti foam. As shown in Fig. S6, the XRD pattern of NiO\nx\n/C is consistent with that of NiO, and RuO\nx\n/C shows peaks corresponding to RuO2 and RuClO2. These results indicate that the crystalline oxides formed well during the thermal treatment. RuNiO\nx\n/C exhibited a multiphase crystalline structure, which matched with the NiO, RuO2, and RuClO2 peaks. The peaks of RuO\nx\n/C were shifted toward higher 2\u03b8 values with respect to those of RuO2 and RuClO2, suggesting that Ni was incorporated into the Ru oxide lattice.The electronic structure of manufactured electrodes surface was analyzed by X-ray photoelectron spectroscopy (XPS) and NEXAFS. The Ru 3d XPS peaks of the RuO\nx\n and RuO\nx\n-L electrodes located at 280.6\u00a0eV (Fig. 1c) were attributed to rutile RuO2, which is in good agreement with the XRD results [34]. The Ru 3d peak of the RuNiO\nx\n electrode was shifted toward higher binding energies by 0.2\u00a0eV with respect to those of the RuO\nx\n and RuO\nx\n-L electrodes; this is a similar phenomenon to that of hydrous RuO\nx\n and in agreement with previous research. This suggests that the incorporated Ni successfully modified the electronic structure of the Ru surface. The Ni 2p XPS spectrum in Fig. 1(d) shows a peak for the RuNiO\nx\n electrode at 855.3\u00a0eV, which corresponds to the Ni2+ oxidation state of NiO and is identical to that of the NiO\nx\n electrode. The NEXAFS spectrum of Ni also shows similar electronic structures for the NiO\nx\n and RuNiO\nx\n electrodes (Fig. S7). In the O 1s XPS spectrum (Fig. 1e), the strong peaks of the RuO\nx\n and NiO\nx\n electrodes at 529.3\u00a0eV were attributed to lattice oxygen. For the RuNiO\nx\n electrode, a peak at 530.8\u00a0eV attributed to oxygen vacancies was also observed [41]. This suggests that the RuNiO\nx\n electrode possesses a larger number of oxygen vacancies than those of the RuO\nx\n and NiO\nx\n electrodes [42]. To obtain detailed information about the t2g and eg orbitals, an O 1s NEXAFS study was conducted (Fig. S8). The RuNiO\nx\n electrode has a similar O 1s NEXAFS spectrum to those of the RuO\nx\n and NiO\nx\n electrodes, which indicates that the introduction of Ni had no significant effects on the t2g and eg orbitals. The electronic structure of the synthesized electrodes was further characterized via Ru and Ni K-edge XANES, which is a bulk-sensitive technique and is based on the electron transitions from the 1s orbital to unoccupied 4p orbitals. The Ru K-edge spectra (Fig. 1f) of the RuO\nx\n and RuNiO\nx\n electrodes show similar peak shapes, in contrast with the XPS results. This indicates that the RuO\nx\n and RuNiO\nx\n electrodes have similar Ru electronic structures to that of rutile RuO2, but the surface Ru in the RuNiO\nx\n electrode has an electronic structure similar to that of hydrous RuO\nx\n. The Ni K-edge spectra of the NiO\nx\n and RuNiO\nx\n electrodes also show identical peak shapes, which is in agreement with the XPS results (Fig. S9). This suggests that the bimetallic oxide of RuNiO\nx\n hardly affects the electronic structure of the entire electrode.Energy-dispersive X-ray spectroscopy (EDS) elemental mapping coupled to transmission electron microscopy (TEM) and SEM was used to determine the elemental distribution and composition of the RuNiO\nx\n electrode (Fig. 1g and Fig. S10). For convenience, the RuNiO\nx\n/C material used in the XRD analysis was used in the TEM analysis as well. Ru and almost Ni are homogenously distributed on the carbon support. According to the EDS elemental analysis of the RuNiO\nx\n electrode, the ratio of O to Cl is 96.7:3.3, indicating that most of the RuNiO\nx\n electrode has a RuO\nx\n structure.The electrocatalytic OER performances of the synthesized electrodes were evaluated in a standard three-electrode system. Fig. 2\n(a) shows the linear sweep voltammetry (LSV) results, which were obtained with a 0.05\u00a0M H2SO4 electrolyte saturated with O2. The RuNiO\nx\n electrode required an overpotential of 217\u00a0mV to achieve 10\u00a0mA\u00a0cm\u22122, a much lower value than those of the RuO\nx\n (248\u00a0mV) and RuO\nx\n-L electrodes (286\u00a0mV). Also, it is higher than that of RuNiO\nx\n-L (262.3\u00a0mV) as shown in Fig. S11(a). These results are in agreement with previous research and reveal that the introduction of Ni enhances the OER catalytic activity of Ru [38]. To determine an optimal Ru:Ni ratio, LSV curves were obtained for RuNiO\nx\n electrodes with different Ni ratios (Fig. S12). The optimal Ru:Ni ratio was determined to be Ru1Ni1. The electrocatalytic activity of the RuO\nx\n electrode is better than that of the RuO\nx\n-L electrode owing to the small particle size, which indicates that the addition of PVP during the dip coating process successfully enhanced the OER performance by increasing the electrochemical surface area. The NiO\nx\n electrode showed a deficient OER performance, confirming the poor intrinsic activity of the Ni oxide. The excellent OER performance of the RuNiO\nx\n electrode was further confirmed via the analysis of the Tafel plots, as shown in Fig. 2(b) and Fig. S11(b). The Tafel slope of the RuNiO\nx\n electrode (56\u00a0mV dec\u22121) is lower than those of the RuO\nx\n (70\u00a0mV dec\u22121), RuO\nx\n-L (79\u00a0mV dec\u22121), RuNiO\nx\n-L (75\u00a0mV dec\u22121) and NiO\nx\n (307\u00a0mV dec\u22121) electrodes. This indicates that the catalytic reaction kinetics were enhanced by the introduction of Ni. A list of the OER catalysts is compared in Table S1 to address the large improvement obtained in this work. Also, we carried out CV with multiple scan rates in the potential range of 0.8 to 0.9 VRHE to obtain the ECSA of the as-prepared catalysts (Fig. S13). The calculated double layer capacitance (C\ndl) can be used to represent the ECSA because the ECSA is directly proportional to C\ndl. It is noteworthy that the C\ndl value of RuNiO\nx\n (268.24 mF cm\u22122) is higher than that of RuO\nx\n (232.58 mF cm\u22122), RuNiO\nx\n-L (208.12 mF cm\u22122) and RuO\nx\n-L (134.81 mF cm\u22122). These results demonstrate that RuNiO\nx\n could have large active surface area results in higher OER activities. To evaluate the stability of the RuNiO\nx\n electrode, a chronopotentiometric experiment was conducted at 10\u00a0mA\u00a0cm\u22122. The RuNiO\nx\n electrode maintained the same performance for 100\u00a0h, confirming its good stability in acidic media (Fig. 2c).To investigate the effects of Ni introduction, comparison of XPS before and after OER was performed. In Ru 3d XPS spectrum, the positive shift of binding energy for Ru electrode after OER (A.O) (280.7\u00a0eV) can be ascribed to altered, close to hydrous RuO\nx\n (280.8\u00a0eV) (Fig. 2d). Meanwhile, RuNiO\nx\n maintains their electronic structure (280.8\u00a0eV). The XPS Ni 2p spectra (Fig. S14) show that the RuNiO\nx\n and NiO\nx\n electrodes maintain their electronic structures after the OER tests. The changes in the oxygen vacancies after the OER can be studied from the XPS O 1s spectra (Fig. 2e). For Ru-based OER catalysts, the dissolution of Ru has the effect on long-term durability. As the OER potential increases, the dissolution rate increase. Since the RuNiO\nx\n catalyst has greatly improved OER activity, the applied overpotential to the anode during the OER stability test was very small. Therefore, since the dissolution rate was very small, it is possible to maintain the performance even for a long-time operation. The surface of all manufactured electrode reveals increased peaks for oxygen vacancy and hydroxide, which is the representative property of hydrous RuO\nx\n and lattice oxygen participation. It is clear that relative peak intensity assigned oxygen vacancy and Ru XPS peak position of RuNiO\nx\n electrode after OER is higher than that of RuO\nx\n electrode. This would be simple reason of enhanced performance by Ni introduction. However, one thing to note is that the chemical state of RuO\nx\n electrode surface is also converted to that of hydrous-RuO\nx\n after OER. Thus, it would be a bit insufficient to explain the effects of Ni introduction.To confirm the applicability under neutral conditions, the performance of the RuNiO\nx\n electrode as an anode for the electrochemical CO2RR was evaluated. In neutral media, the trends are similar to those in acidic conditions, as show in Fig. 2(f) and Fig. S15. The RuNiO\nx\n electrode could reach 10\u00a0mA\u00a0cm\u22122 of current density at an overpotential of only 386\u00a0mV, which is a lower value than those of the RuO\nx\n (440\u00a0mV) and RuO\nx\n-L (480\u00a0mV) electrodes. The electrochemical CO2RR performance was evaluated using a homemade zero-gap CO2 electrolyzer and gaseous CO2 to accelerate the reduction reaction while minimizing the mass transfer resistance (Fig. 2g). An Ag electrode and a RuNiO\nx\n electrode were used as the cathode and anode, respectively, and 0.1\u00a0M KHCO3 was used as the electrolyte. For comparison, a commercial IrO2 electrode and the synthesized RuO\nx\n-L electrode were evaluated under same conditions. The cell voltage of RuNiO\nx\n electrode is lower than other electrodes, demonstrating high OER catalytic activity under CO2RR condition (Figs. S16 and S17). Furthermore, the RuNiO\nx\n electrode exhibited a CO selectivity higher than 90% for all values of the applied current density, which represents a similar activity with commercial IrO2 electrode. By contrast, for the RuO\nx\n-L electrode, as the applied current density increased, the CO selectivity decreased and the H2 selectivity increased. The high cell voltages of RuO\nx\n-L electrode accelerate the dissolution of Ru, leading Ru metal ion crossover to cathode. This can be attributed to the low CO partial current density and poor stability of the RuO\nx\n-L electrode under neutral conditions. In the case of RuNiO\nx\n-L, the CO selectivity maintains lower than that of RuNiO\nx\n as shown in Fig. S18.To gain further insights into the role of Ni in the enhancement of the OER activity, in-situ/operando XANES and on-line ICP-MS analyses were conducted in acidic conditions. The Ni K-edge XANES spectra of the NiO\nx\n and RuNiO\nx\n electrodes obtained at different applied potentials are shown in Fig. 3\n(a\u2013c). The XANES peaks of the NiO\nx\n electrode at potentials in the range of the open circuit voltage (OCV) to 1.63\u00a0V are similar, indicating that the Ni2+ oxidation state was maintained during the OER (Fig. 3a). However, the pre-edge shape, which represents the coordination geometry, changes with the applied potential [43]. The pre-edge peak of the NiO\nx\n electrode appeared at a potential of 1.43\u00a0V and shifted toward lower photon energies as the potential increased to 1.63\u00a0V. This suggests that the Ni oxide structure was distorted for applied potentials above 1.43\u00a0V. For the RuNiO\nx\n electrode, the oxidation state of Ni was also maintained when the applied potential changed (Fig. 3b). However, the pre-edge of the RuNiO\nx\n electrode shows significant differences with respect to that of the NiO\nx\n electrode. A pre-edge peak appeared at the OCV and shifted toward lower photon energies as the applied potential increased. The detailed comparison of the NiO\nx\n and RuNiO\nx\n electrodes at 1.23 and 1.63\u00a0V is shown in Fig. 3(c). Both electrodes retained an oxidation state of Ni which is consistent with Ni2+, but the pre-edge peak of the RuNiO\nx\n electrode appeared at a lower potential and was located at lower photon energies than that of the NiO\nx\n electrode. These results indicate that the RuNiO\nx\n electrode presented distortion at lower potentials and to a greater extent than the NiO\nx\n electrode.The changes in the Ru electronic structure during the OER were investigated via in-situ/operando Ru-K edge XANES at potentials in the range of the OCV to 1.63\u00a0V. The spectrum for the RuO\nx\n electrode maintains its features (which resemble those of RuO2) but shows a slight increase in the white-line intensity at potentials above 1.43\u00a0V (Fig. 3d). It has been reported that an increased white-line intensity indicates the presence of hydrous RuO\nx\n with a disordered structure and high amounts of structural water [44]. Thus, the results suggest that a transition of the RuO\nx\n electrode from rutile RuO2 to hydrous RuO\nx\n occurred at a potential of approximately 1.43\u00a0V. The Ru XANES spectrum of the RuNiO\nx\n electrode shows an increased white line intensity at the OCV and maintains its features as the potential increases up to 1.63\u00a0V (Fig. 3e). These tendencies are also confirmed at EXAFS spectra, as shown in the Fig. S19. The hydrous RuO\nx\n peak with 2.1\u20132.25\u00a0nm appeared at 1.43 and 1.63\u00a0V for RuO\nx\n and RuNiO\nx\n, respectively. Especially, the Ru\u2013O bonding length of RuNiO\nx\n decreased during structure distortion. These results demonstrate that the RuNiO\nx\n electrode has the electronic structure of hydrous RuO\nx\n at lower potentials than the RuO\nx\n electrode does (Fig. 3f). In summary, the distorted Ni and modified RuO\nx\n formed at the active sites of RuNiO\nx\n changed the initial electronic distribution and coordination environment of the RuNiO\nx\n, and hence could enhance its OER performance [45\u201347].In order to probe the structure transition, on-line ICP-MS study was performed with changing applied potential in acid condition. On-line ICP-MS is a powerful technique to measure real-time electrochemical dissolution which is caused by direct dissolution, structure transition and oxygen lattice participation for OER [48,49]. For the RuO\nx\n electrode, the dissolution began at about 1.3\u00a0V and increased in extent as the potential was increased (Fig. 3g and h). This indicates the formation of hydrous RuO\nx\n at potentials above 1.3\u00a0V and the occurrence of oxide-mediated dissolution, which is an indication of lattice oxygen participation in the OER. The dissolution peaks associated with lattice oxygen participation for the RuNiO\nx\n electrode at the highest potential are significantly larger than those for the RuO\nx\n electrode, indicating a large extent of lattice oxygen participation. Small dissolution peaks were observed for the RuNiO\nx\n electrode near 1.0\u00a0V, which correspond to the structure transition. Ni dissolution exhibited similar trends to those of Ru dissolution. The NiO\nx\n electrode has dissolution peaks only at potentials above 1.3\u00a0V, indicating that the transition and direct dissolution began above 1.3\u00a0V. For the RuNiO\nx\n electrode, after the first cycle, the dissolution peak at the highest potential was reduced, and other dissolution peaks appeared near 1.0\u00a0V. This reveals that the direct dissolution was inhibited, and the structural transition occurred at around 1.0\u00a0V. These results are consistent with the in-situ/operando XANES results.Based on the ex-situ characterization and in-situ/operando studies, we propose the role of Ni in the enhancement of the OER activity of Ru oxides as shown in Fig. 4\n. Before explaining the role of Ni, the behaviors of the RuO\nx\n and NiO\nx\n electrodes with different applied potentials should be clarified. The NiO\nx\n electrode retained its structure until the potential reached 1.23\u00a0V. Above 1.23\u00a0V, the structure was distorted, leading to asymmetry. The RuO\nx\n electrode also maintained its electronic structure until the potential reached 1.23\u00a0V, and above said potential, the surface of the rutile RuO2 was converted to hydrous RuO\nx\n, which enabled the participation of lattice oxygen in the OER. When Ni was introduced in the RuO2 structure, the Ru oxide lattice distance was slightly reduced due to the short Ni\u2013O distance and the mismatch of the RuO2 and NiO structures, which led to abundant oxygen vacancies. The structure of the RuNiO\nx\n electrode was already distorted at the OCV, and it was distorted further as the applied potential increased, generating abundant oxygen vacancies. It is proposed that oxygen vacancies lead to large structural distortion at low potentials. This phenomenon also enables the penetration of water and ions into the RuNiO\nx\n structure, modifying the electronic structure further and causing the transition to hydrous RuO\nx\n, which enhances the participation of lattice oxygen in the OER. According to theoretical studies, the distorted RuNiO\nx\n structure with an eg-dz\n2 misalignment can minimize the energy barrier for the OER [38], and the LOM provides a unique local configuration by modifying the metal\u2013oxygen hybridization [21,22]. These changes in the electronic structure and the LOM pathway significantly reduce the energy barriers for each intermediate in the OER. Thus, the introduced Ni not only modified the initial electronic structure of the Ru oxide (which was observed via XPS), but also created a large number of oxygen vacancies at low overpotentials by distorting the lattice oxygen structure. This phenomenon of the RuNiO\nx\n electrode can accelerate the conversion of the OER mechanism from AEM to LOM, enhancing the catalytic activity. On the basis of results, we summarized OER mechanisms of RuO\nx\n based on previously proposed mechanism in Fig. S20\n[22]. Fig. S20(a) showed the AEM of RuO\nx\n. The reaction can happen either on Ru site via an OOH intermediate. However, LOM differs by participation of activated lattice oxygen in the reaction as shown in Fig. S20(b). Water attacked the activated oxygen and removed as O2 from the surface leaving behind an oxygen vacancy.In conclusion, we unveiled the role of introduced Ni in Ru oxide which enhanced catalytic activity of OER via in-situ/operando technique. A nanosized RuNiO\nx\n electrode was simply fabricated via a modified dip coating process. The electrode demonstrated a remarkable OER performance (an overpotential of 210\u00a0mV for 10\u00a0mA\u00a0cm\u22122) and good stability in acid media (100\u00a0h). The pristine RuNiO\nx\n electrode possesses an electronic structure close to that of hydrous RuO\nx\n, with a larger number of oxygen vacancies than that of the prepared RuO\nx\n electrode. According to the in-situ/operando XANES and ICP analyses, the electronic structure of the RuO\nx\n electrode became similar to that of hydrous RuO\nx\n at potentials above 1.43\u00a0V. The introduction of Ni distorted the structure of the oxygen lattice, creating abundant oxygen vacancies and changing the electronic structure to that of hydrous RuO\nx\n at a low potential. This increased the lattice oxygen participation and changed the OER mechanism of the RuNiO\nx\n electrode from AEM to LOM at a lower potential than that of the RuO\nx\n electrode. These findings provide information on the behavior of bimetallic oxide materials during the reaction and constitute fundamental insights into the development of efficient electrocatalysts.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by institutional program grants from the Korea Institute of Science and Technology and Korea Institute of Energy Technology Evaluation and Planning (KETEP) granted financial resource from the Ministry of Trade, Industry & Energy, Republic of Korea (No. 20224C10300020) and \u201cCarbon to\u00a0X\u00a0Project\u201d (2020M3H7A1098229) through the National Research Foundation (NRF) funded by the Ministry of Science and ICT, Republic of Korea. This research was also supported by the National Research Council of Science & Technology (NST) grant by the Korean government (MSIT) (No. CAP21011-100) and National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (NRF-2021R1A2C2093467). We also acknowledge Advanced Analysis Center at KIST for the TEM and 1D XRS KIST-PAL beamline for measuring the hard X-ray absorption spectroscopy (XAS).Supplementary data to this article can be found online at https://doi.org/10.1016/j.jechem.2022.09.032.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  Introducing Ni in Ru oxide is a promising approach to enhance the catalytic activity for the oxygen evolution reaction (OER). However, the role of Ni (which has a poor intrinsic activity) is not fully understood. Here, a RuNiO\n                        x\n                      electrode fabricated via a modified dip coating method exhibited excellent OER performance in acidic media, and neutral media for CO2 reduction reaction. We combined in-situ/operando X-ray absorption near-edge structure and on-line inductively coupled plasma mass spectrometry studies to unveil the role of the Ni introduced in the Ru oxide. We propose that the Ni not only transforms the electronic structure of the Ru oxide, but also produces a large number of oxygen vacancies by distorting the oxygen lattice structure at low overpotentials, increasing the participation of lattice oxygen for OER. This work demonstrates the real behavior of bimetallic oxide materials under applied potentials and provides new insights into the development of efficient electrocatalysts.\n               "}
{"full_text": "Carbon nanostructures show promise in fuel cell applications as a catalyst support. Much research is still being conducted in order to enhance the catalytic activity and selectivity and reduce the cost of catalyst preparation [1]. There are currently three main synthesis methods used to produce graphitic nanofibres GNFs and/or carbon nanotubes CNTs. The methods are electric arc discharge (EAD) [2], laser ablation technique (LA) and chemical vapour deposition technique CVD [3,4]. Among the techniques stated, CVD is a common method used to produce GNFs and CNTs providing a good carbon yields and low amounts of unwanted soot.The porous nature of amorphous silica provides nucleation sites for catalyst precursor precipitation and hinders subsequent grain growth and sintering. This will lead to the formation of small catalyst particles for carbon growth. The small and uniform size of catalyst particles will lead to the formation of uniform size synthesised CNS [8]. The catalyst particle size is an important factor in carbon nanostructures growth where carbon filament growth rate increases with decreasing catalyst particle size [1,5]. It is reported that by using pores silica [6,7] and pores alumina [8] as a support for a metal catalyst, high quality of CNTs can be grown. It has also been reported that synthesising CNTs using a silica support is very attractive due to uniform pores size of silica, good and homogenous CNTs between 1 and 6\u00a0nm diameter were successfully synthesised at 1000\u00a0\u00b0C [9].Even though unsupported NiO catalyst precursor can give high deposition yields of 57\u00a0g\u00a0g-1\ncatalyst h\u22121 of deposited carbon, but the type of graphitic nanostructures formed is a mixture of platelet and herringbone with the diameter of the fibres ranging from 20 to 500\u00a0nm, the selectivity for the formation of graphitic structures and the fibres diameter is hard to control [4]. However, the unsupported Fe2O3 catalyst precursor, has much better specific selectivity for the graphitic structure formed, but the fibre diameters are range from 50 to 200\u00a0nm but the yield of deposited carbon is also very low, 2.25\u00a0g\u00a0g-1\ncatalyst h\u22121\n[3]. The hypothesis behind the current study is that a Ni, Fe bimetallic catalyst will have a high yield, high selectivity and if supported on a porous silica, a well-defined size distribution for the deposited carbon.The aim for this study is to investigate the effect of the CVD variables of reaction temperature and gas composition on the nanostructured carbon produced by using a supported NiFe2O4 spinel as a catalyst precursor. The controlled factors for the synthesis are temperature and different ratios of reaction gases, C2H4 and H2 which reported play a vital role in nanostructures formed [8,10].A high purity grade porous silica, 200\u2013500\u00a0\u00b5m particle size with a pore size of 60\u00a0\u00c5, (Sigma Aldrich) was impregnated with a nickel nitrate hexahydrate Ni(NO3)2\u00b76H2O (puriss grade, Sigma Aldrich), iron (III) nitrate nonahydrate Fe (NO3)3\u00b79H2O (ACS reagent, Sigma Aldrich) solution by incipient wetness technique [7,11]. These solutions were then added into weighed porous silica powder to produce a final 5\u00a0wt% Ni0.36Fe0.64 reduced metal catalyst in the silica support. The heating process carried out at 130\u00a0\u00b0C evaporated off the water and left behind iron-nickel nitrate impregnated in the pore sites of the silica.\n\n(1)\n\n\n2\nF\ne\n\n\n(\nN\n\nO\n3\n\n)\n\n3\n\n.\n9\n\nH\n2\n\n\nO\n\n\nl\n\n\n\n+\nN\ni\n\n\n(\nN\n\nO\n3\n\n)\n\n2\n\n.\n6\n\nH\n2\n\n\nO\n\n\nl\n\n\n\n\n\u2192\n\n\u0394\n130\n\u00b0\nC\n\n\nN\ni\nF\n\ne\n2\n\n\n\n(\nN\n\nO\n3\n\n)\n\n\n8\n\n\ns\n\n\n\n\n+\n15\n\nH\n2\n\n\nO\n\n\ng\n\n\n\n\n\n\n\nThe calcination at 400\u00a0\u00b0C for 4\u00a0h caused the iron-nickel nitrate to thermally decompose to iron-nickel oxide (also known as spinel structures catalyst precursor).\n\n(2)\n\n\nNiF\n\ne\n2\n\n\n\n(\nN\n\nO\n3\n\n)\n\n\n8\n\n\ns\n\n\n\n\n\n\u2192\n\n\u0394\n400\n\u00b0\nC\n\n\n\nN\ni\nF\n\ne\n2\n\n\nO\n\n4\n\n\ns\n\n\n\n\n+\n8\nN\n\nO\n\n2\n\n(\ng\n)\n\n\n\n+\n\n2\n\nO\n\n2\n\n\ng\n\n\n\n\n\n\n\n\nPrior to the reaction, the iron-nickel oxide is reduced to Ni0.36 Fe0.64 metal under the hydrogen flow.\n\n(3)\n\n\nNiF\n\ne\n2\n\n\nO\n\n4\n\n\ns\n\n\n\n\n+\n\n4\n\nH\n\n2\n\n\ng\n\n\n\n\n\n\u2192\n\n\u0394\n400\n\u00b0\nC\n\n\n\nN\ni\nF\n\ne\n\n2\n\n\ns\n\n\n\n\n+\n4\n\nH\n2\n\n\nO\n\n(\ng\n)\n\n\n\n\n\n\nThe catalyst produced analysed using XRD (Bruker D8 Advance) at every stage to ensure formation of NiFe2O4 spinel and Ni0.36 Fe0.64 metal.The growth of carbon nanostructures was carried out at atmospheric pressure in a fixed bed reactor (quartz-tube 5.0\u00a0cm diameter and 150\u00a0cm length) located in a vertical oven with C2H4 (Research Grade N3.2, BOC) as a carbon source. The feedstock gases H2 and C2H4 were controlled using mass flow controllers (MKS GE250A) providing a mixture of H2/C2H4 of 20/80, 50/50 and 80/20 maintaining a total flow of 100 sccm (standard cubic centimeters per minute). During the initial and the intermediate stages of CVD, the chamber is placed in a vacuum stated by pumping out the chamber using commercial vacuum pump. This is important to ensure the chamber is free from any residual gas that may affect the reaction. Prior to the reaction, the catalyst precursor was first reduced in 10% of H2 for 4\u00a0h. The reactions time were kept constant at 2\u00a0h. The temperature and gas composition were controlled as outlined in Fig. 1\n.All reactions were repeated three times to present reproduced data and minimise the calculation error. For the yield calculation, the percentage of the carbon deposition is defined as follows:\n\n(4)\n\n\nyield\n\n\n(\n%\n)\n\n=\n100\n\u00d7\n\n\n\n\nweight\n\n\no\nf\n\nc\na\nr\nb\no\nn\n\n\nd\ne\np\no\ns\ni\nt\ne\nd\n\no\nn\n\n\nt\nh\ne\n\n\nc\na\nt\na\nl\ny\ns\nt\n\n\nweight\n\no\nf\n\nc\na\nt\na\nl\ny\ns\nt\n\n\n\n\n\n\n\n\nThe output gases and the catalyst activity were monitored using mass spectrometer (MS) and analysed by MASsoft 7 software (Hiden analytical). Scanning electron microscopy (SEM) FEG\u00a0\u2212\u00a0ESEM Philips XL\u00a0\u2212\u00a030, and transmission electron microscopy (TEM) JEOL 2000fx, equipped with selected area electron diffraction (SAED) were used to characterise the GNFs and CNS produced.The samples produced from the CVD were then underwent acid treatments to remove the silica support and the catalyst precursor. First, the samples were treated in 20% concentration of hydrofluoric acid (HF) in three 30\u00a0min periods to dissolve the silica. It was then filtered and washed using excess distilled water before further purification in 12\u00a0M nitric acid, stirred for 18\u00a0h at room temperature before filtered and washed using excess distilled water. Finally, the samples were left to dry in a drying cupboard at 80\u00a0\u00b0C overnight.Impregnated silica with iron, nickel nitrates calcined at 400 \u00b0C for 4\u00a0h, showed a weak diffraction pattern due to NiFe2O4 at 2\u03b8 values of 37\u00b0 (311) and 63\u00b0 (440) as in Fig. 2\n. However due to the high relative intensity of silica support, these NiFe2O4 peaks were not clearly observed. It is also believed, for all the catalyst precursor preparations, that due to the slow drying process, these metal precursors formed clusters close to the entrance of the pores in the silica as reported by Bond et al.\n[12]. Then a heat treatment applied at 400\u00a0\u00b0C formed a cap of catalyst particle over a pores site of silica preventing the diffraction from the catalyst precursors in the pores.Prior to the CVD reaction, the catalyst precursor underwent a reduction in 10% H2 (in 90 sccm Ar) to form iron-nickel alloy. In the \u2018after H2 reduction\u2019 pattern, it was clearly shown that the NiFe2O4 phase had disappeared and a new pattern for iron-nickel alloy was observed at 2\u03b8 values of 45\u00b0 (111) and 52\u00b0 (200).\nFig. 3\n shows the XRD profiles for the samples at the different stages of the experiment. The \u2018after CVD\u2019 profile did not show the present of any carbon peak in the sample due to the small yield of CNS diffraction overshadowed by the silica diffraction. However, the XRD profile for \u2018after acid treatments\u2019 samples showed the high relative intensity of carbon peaks at 2\u03b8 values of 26\u00b0 (002) and 42\u00b0 (111). The very low intensity of the iron-nickel which is still observed due to the metal catalyst particles trapped inside the carbon nanostructures thus preventing dissolution by the acids. The catalyst particles trapped inside the CNS structures is a common scenario especially when growing MWCNTs similar to previous work proposed [13\u201316].\nFig. 4\n shows that for the silica supported NiFe2O4 catalyst precursor, the carbon deposition increased with the increase of the hydrogen content from 20% to 50%. The maximum yield obtained was around 762%\u2013898% mc/mcat observed at a H2 composition of 50%. However, further increase in the hydrogen content to 80% seemed to reduce the carbon deposition, due to the decreasing carbon source for deposition dropping the yield to 399%\u2013564% mc/mcat. The reduction in yield was found to be the case for all the reaction temperatures studied. Furthermore, the amount of carbon deposition at the temperature of 400\u00a0\u00b0C, 500\u00a0\u00b0C, 600\u00a0\u00b0C and 700\u00a0\u00b0C was within a similar range irrespective of the gas mixture.The MS plots for the silica supported NiFe2O4 during reduction under hydrogen (10% H2/90% Ar) are given in Fig. 5\n. The temperature was increased to 400 \u00b0C (taking ca. 40\u00a0min), upon reaching this temperature the hydrogen was introduced at a flow of 10 sccm and the Ar was reduced to 90 sccm making 10% H2/90% Ar. No hydrogen partial pressure detected after being introduced and an increase in H2O partial pressure suggested the onset of reduction of NiFe2O4. This can be seen from the production of water (H2O with m/z 18) an increased partial pressure of H2O after the introduction of hydrogen. The partial pressure of hydrogen was at a minimum level during the first 45\u00a0min of introduction before it started to increase. The decreased formation of H2O after two hours of the introduction of hydrogen suggested the entire oxide precursor was reduced to iron-oxide metal.\nFig. 6\n (a\u2013d) shows the MS plot for the samples synthesised using hydrogen reduced silica supported Ni0.36Fe0.64 catalyst during the CVD reaction under different reactant gas compositions. These reactions with different reaction temperature and gas composition were selected to show the catalytic activity of NiFe2O4 throughout carbon deposition analysis. Fig. 6 (a) shows minimal activity of catalyst and low partial pressure of CH4 and C2H6 which later lead to low carbon deposited during synthesis.\nFig. 6 (b) shows the MS plot for the reaction at 700 \u00b0C with 20% H2/80% C2H4. With the increase of the reaction time, an increase of CH4 and C2H6 were detected. It was found that the dehydrogenations of C2H4 to CH4 and hydrogenation to form C2H6 were maintained throughout the synthesis period. However, by increasing the hydrogen to 50%, the catalytic activity was observed to be increased, as shown in Fig. 6 (c). This can be seen from the steady production of the CH4 and C2H6 throughout the 2\u00a0h reaction and until the reaction was stopped. The increase in the hydrogen content assist the hydrogenation process, forming C2H6 and keeping the surface of the catalyst clean for further reactions to occur.The MS plot for the further increase of hydrogen content to 80% is shown in Fig. 6 (d). The dynamic activity of the catalyst can be seen from the maintained CH4 partial pressure and consumption of C2H4 throughout the reaction period, but there was a lower partial pressure of C2H6 and CH4 which lead to a low yield of CNS formation.From the mass spectrometry, the dehydrogenation of C2H4 was observed by the decrease of C2H4 partial pressure and the formation of H2 was found to occur immediately after H2 and C2H4 were introduced into the CVD furnace. This can be observed by the decrease in the partial pressure of C2H4 and the increase of partial pressure of H2 occurring immediately after the reactants gases were introduced.The result of C2H4 decomposition is the production of carbon atoms at the surface of the catalyst particles. The H2 molecules released during this dehydrogenation may reduce the oxide metal catalyst precursor or may interact with the free C2H4 to form higher density hydrocarbons as in agreement with previous experimental works [10,17]. The carbon atoms produced during the C2H4 dehydrogenation will then diffuse onto the metal catalyst surface. The C2H4 dehydrogenation is believed to produce a carbon atom at the surface of the metal catalyst.The diffusion of C atoms into the catalyst particles follows on from the dehydrogenation of C2H4. With the solubility of C atoms into the catalyst surface, it is expected that catalyst particles attain a highly mobile or quasi-liquid state at the reaction temperature for CNS growth. The formation of the metal solid (NixFey-C) then resulted in lowering the melting point of the catalyst particle, which also softened the catalyst particle [18]. The increase in the reaction temperature will continue softening the catalyst particles and will encourage the carbon solubility as compared to the initial solid state of catalyst particles. This can be seen from the higher CNS yield produced as the reaction temperature increased from 400\u00a0\u00b0C to 700\u00a0\u00b0C as shown in Fig. 4.After the formation of the quasi-liquid phase, the C atoms become free to diffuse through into the catalyst particles at the area with high carbon concentration. The carbon precipitation stage follows on after the C atom was found to be adsorbed and diffused into the catalyst particle. The precipitation and propagation of the carbon were found to be coinciding with the formation of hydrocarbon gases, e.g C2H6, C3H6 and C3H8\n[5,19]. The continuous precipitation and propagation of graphite is determined by the dehydrogenation and hydrogenation on the catalyst surface and the rate of carbon diffusion into the catalyst particles. The precipitation and propagation were also influenced by the reaction conditions: reaction temperature and reactant gas compositions. There was no further analysis such as in situ TEM to determine the exact precipitation and propagation rate or carbon formation on the catalyst. However, from the final carbon yield production during CVD process in Fig. 4, it is possible to study the pattern of optimum reaction conditions for carbon growth. It was found that for silica supported NiFe2O4, the reaction temperatures investigated 400 \u00b0C\u2013700 \u00b0C and the composition of 50% H2/50% C2H4 as evidenced by the high partial pressure of CH4 and C2H6 produced during the reaction which later gave the maximum carbon produced, hence the maximum rates of carbon precipitation occurred. the reaction of 50/50 ethene/hydrogen was observed to give the highest yield, which suggests that there is a balance between the diffusion of the carbon atom through the catalyst particle leading to the precipitation and propagation of MWCNTs as well as the flux of carbon onto the surface hydrogenation of the carbon onto the catalyst surface to keep the catalyst face clean, allowing sites to remain free for a prolonged reaction.From the MS data, the catalyst deactivation can be detected by an increase of the C2H4 and increased of H2 partial pressure in Fig. 6. The fast deactivation of the catalyst was observed for the reaction using silica supported NiFe2O4 catalyst precursor using 20% H2/80% C2H4, shown a lowest CNS yield production. Further investigation of the MS plot found that the dehydrogenation of C2H4 happened at a very low intensity of hydrogen applied in the reaction. This rapid catalyst deactivation resulted in formation of a graphite layer surrounding the catalyst particles which stopped any further reaction from occurring. The catalyst activity was high at the beginning of the introduction of reactant gases, but later on the reactivity dropped. The trend of the deactivation process found in this study agreed with many other studies that have reported this phenomenon of a reduction in conversion of C2H4 (dehydrogenation) to carbon and H2 or a slow propagation of CNS after a prolonged period [5,10,20,21].\nFig. 7\n (a) shows the micrograph of un-impregnated silica support. The CNS was successfully synthesised using silica supported NiFe2O4, Fig. 7(b) showing the growing CNS. More details in morphology of the nanostructures were represented in Fig. 7 (c, d). Given that the limitations of magnification in SEM analysis, details studied on the structures and morphology of the CNS were analysed using TEM. The structures present at the different reaction conditions are represented in the Fig. 8\n.The TEM micrographs in Fig. 8 show nanostructures of MWCNT carbon observed when synthesised at 400 \u00b0C to 700 \u00b0C. At 700 \u00b0C reaction temperature, MWCNTs with encapsulation, Fig. 8(a) were obtained at 20% hydrogen composition. By increasing the hydrogen content to 50% and 80%, only the MWCNTs were observed to be found, Fig. 8(b, c). At the lower reaction temperature, 600 \u00b0C, disordered CNS with encapsulation, Fig. 8(d) were observed, whereas disordered, Fig. 8(e) and disordered CNS with hollow herringbone GNFs, Fig. 8(f) were observed when the hydrogen content was increased to 50% and 80% respectively. For the CVD reaction at 500 \u00b0C, the encapsulated nanostructures with disordered CNS were obtained at 20% and 50% hydrogen content Fig. 8(g, h) and disordered nanostructures with encapsulated CNS found at 80% hydrogen content Fig. 8(i). However, at 400 \u00b0C, the encapsulated CNS was observed at all reactant gas compositions Fig. 8(j, k, l).The formations of MWCNTs were observed at the reaction temperature of 700 \u00b0C, regardless of the ratio of hydrogen/ethene composition. For better understanding and clarification, detailed analysis of the morphology and structures was undertaken for the reactions at 700 \u00b0C conditions. The range of diameters of the CNS observed at700 \u00b0C under different gas composition was shown in Fig. 9\n.The analysis of 100 fibres shows an average diameter of Davg\u00a0=\u00a031\u00a0\u00b1\u00a04\u00a0nm. The tubes diameters were distributed in a range of 25\u201345\u00a0nm where the majority of the tubes were distributed in the range of 30\u201335\u00a0nm as in Fig. 9(a). For the synthesis in a condition of 50%H2/50%C2H4 the average diameter of MWCNTs present in a sample was Davg\u00a0=\u00a030\u00a0\u00b1\u00a04\u00a0nm. The fibres were distributed over a diameter range of 25\u201345\u00a0nm where the maximum distribution was at 30\u201335\u00a0nm, Fig. 9(b). However, for the reaction in a reactant gas of 80%H2/20%C2H4. The average diameter of MWCNTs was calculated to be Davg\u00a0=\u00a032\u00a0\u00b1\u00a04\u00a0nm. The diameters were distributed over a range of 25\u201350\u00a0nm with the maximum at 35\u00a0nm, Fig. 9 (c).This uniformity in CNS diameters was due to the formation of uniform catalyst precursor particles confined within the silica pores. Most of the previous work done also illustrated the formation of a uniform diameter of CNS depending on the pores size of the support [6,11,22\u201324], while the other studies based on CVD synthesis using the unsupported catalyst showed the nanostructure diameters were in the range of 20\u2013250\u00a0nm [4,19,25,26]. In the work presented here the uniformity of the diameter of CNS is controlled by the size of the porous template in the silica particle itself\nFig. 10\n (a) shows the MWCNT structures present in a sample synthesised at 700 \u00b0C with 20%H2/80% C2H4. The presence of some encapsulated carbon is also shown in the figure. The insert in Fig. 10 (b) shows the SAED images with two arcs perpendicular with the structures, confirming this structure is MWCNT. Fig. 10 (c, d) shows the TEM images for the MWCNTS synthesised at a reaction temperature of 700 \u00b0C with reactant gas composition of 50% H2/50% C2H4 at a higher magnification with the insert of SAED pattern for the selected fibre. The TEM analysis of the structures of samples synthesised at 700 \u00b0C with reactant gas composition of 80% H2/20% C2H4 are shown in Fig. 10 (e, f). The structures present were confirmed to be MWCNTs, shown by images the higher magnification and the SAED pattern inserted in Fig. 10 (f).The predominant nanostructures and compositions present in the samples synthesised at 700 \u00b0C with a different H2/C2H4 composition observed on the selected area electron diffraction (SAED) pattern over 50 fibres randomly selected nanostructures are summarised in table 1\n. The dominant structures present were observed as MWCNTs, which were present at 20% H2 and 50% H2 (over 80% and 50% ethene respectively). While increasing the hydrogen percentage to 80%/20% ethene, the nanostructures present were still predominantly MWCNTs with a small presence of hollow herringbone GNFs.It was found that the addition of nickel into the iron matrix forming binary iron-nickel alloy found to favour the development of MWCNTs, which can be seen from the table 1.For all the reaction at synthesis temperature of 400 \u00b0C\u2013700 \u00b0C, the CNS production using silica supported NiFe2O4 catalyst produced a uniform diameter of CNS between 20 and 50\u00a0nm and the highest yield produced at the gas composition of 50%H2/50% C2H4. It was also shown the activity of the catalyst during the formation of CNS can be monitored using in situ mass spectrometry by monitoring the exhaust gas from the on-going synthesis. During the CNS precipitation and propagation, it was found that the highest activity of catalyst occurred at the gas composition of 50%H2/50% C2H4.It also shown using silica supported NiFe2O4 as a catalyst precursor, the synthesis at 700 \u00b0C will produce MWCNTs at all H2/C2H4 reactant gas compositions. While, the reaction at lower temperature produced disordered CNS and encapsulated carbon.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to acknowledge the financial support of the Ministry of Higher Education Malaysia (MOHE) under the Fundamental Research Grant Scheme FRGS/1/2015/TK05/UPM/02/3/030115-1704FR.", "descript": "\n                  The carbon nanostructures (CNS) were successfully grown during the chemical vapour deposition of ethene (C2H4) and hydrogen (H2) over a supported Ni0.362Fe0.64 catalyst. The temperature of the reaction was varied between 400\u00a0\u00b0C and 700\u00a0\u00b0C with different ratios of hydrogen and ethene (20/80, 50/50 and 80/20). The increase of the H2 in the reaction gas gives higher deposition yield of carbon where the maximum yield is observed at a mixture of 50/50 of H2 and C2H4 respectively. The results showed that the structures of the carbon formed by the decomposition of ethene were dependent on the reaction temperature and the gas ratio employed. Graphitic nanofibers (GNFs) and multiwall carbon nanotubes (MWCNTs) were produced when the temperature reached 700\u00a0\u00b0C, while at the lower temperature 600 \u00b0C, disordered CNS with encapsulation and some amorphous nanostructures tended to form.\n               "}
{"full_text": "The continuously growing demand for clean transportation fuels with the depletion of conventional petroleum reserves promotes an increase of poor and heavy oil upgrading. Therefore, it is significant to efficiently convert the residual oil into light fractions oil in the refining industry, in which the composition of residual oil is complex, containing paraffinic, naphthenic and aromatic hydrocarbons with high contents of sulfur (S), nitrogen (N), vanadium (V), nickel (Ni) and so on, but it is difficult and challenging to achieve the efficient conversion of residual oil. Among several residual oil conversion processes, slurry-phase hydrocracking of heavy oil technology is an alternative technology, which has been attracting great attention because of its ability to process various heavy oils and achieve a higher feedstocks conversion (Go et\u00a0al., 2018; Lim et\u00a0al., 2018). It is known that the catalyst plays a significant role on the performance in the slurry-phase hydrocracking process.Slurry-phase hydrocracking catalysts mainly include oil-soluble dispersed catalysts and solid powder dispersed catalysts (Watanabe et\u00a0al., 2002; Al-Attas et\u00a0al., 2019). Oil-soluble dispersed catalysts were obtained by introducing transitional metals (such as Mo, Ni, Co and Fe) into oil soluble precursors to form organometallic compounds (Kang et\u00a0al., 2020; Yang et\u00a0al., 2019; Li et\u00a0al., 2019). The active metal of catalyst can rapidly saturate the free radicals produced via \u03b2-scission reaction of C\u2013C by incorporating hydrogen into the cracked active hydrocarbons, which favor to limit the aromatic condensation reaction and over cracking reaction (Bellussi et\u00a0al., 2013; Nguyen et\u00a0al., 2016; Kim et\u00a0al., 2017). It has been found that oil-soluble Mo catalyst exhibited higher performance compared with other metals catalysts in the slurry-phase hydrocracking process (Kim et\u00a0al., 2018; Liu et\u00a0al., 2019). Moreover, the synergy effect of bimetals catalyst on the heavy oil conversion and the yield of light distillates oil was observed (Kim et\u00a0al., 2019; Nguyen et\u00a0al., 2015). Oil-solution catalysts possess great catalytic performance in the slurry-phase hydrocracking, but the high-cost during organometallic compounds preparation restrict their wide application in the petroleum refining industry.Solid powder dispersed catalysts, generally inorganic minerals, were widely employed in the early stage of the development of slurry-phase hydrocracking technology, in which inorganic minerals were considered as the hydrogenation active phase and coke carries in the slurry-phase hydrocracking process (Sanaie et\u00a0al., 2001; Matsumura et\u00a0al., 2005). Matsumura et\u00a0al., (2005) used the natural limonite as catalyst to upgrade Brazilian Marlim vacuum residue (VR), VR conversion was less than 80\u00a0wt%, C5-340 oC fraction yield was about 34\u00a0wt%, and the coke yield was higher than 5\u00a0wt%. Red mud containing a mixture of Fe, Al and Ti oxide was employed for the slurry-phase hydrocracking of VR, the result showed that VR conversion was about 65\u00a0wt%, and the yield of naphtha and diesel was lower than 34\u00a0wt%, as well as the active phase for the hydrocracking reaction was pyrrhotite (Fe(x-1)S\nx\n) derived from iron oxide (Nguyen-Huy et\u00a0al. 2012, 2013). Quitian et\u00a0al. (Quitian et\u00a0al., 2016) found that ore catalyst of molybdenite and hematite were able to inhibit the gas and coke formation caused by decomposition and condensation reactions, and promote the hydrogenation of the free radicals formed primarily via the thermal cracking of C\u2013C bond. Although the inorganic minerals as catalyst have an advantage over the low cost in the slurry-phase hydrocracking process, their inferior catalytic activity can\u2019t meet the demand for highly efficient conversion of heavy oils.In order to overcome the drawbacks of oil-solution dispersed catalysts and fine inorganic minerals catalysts, the transitional metals such as Mo, Co and Ni were supported on carbon, alumina, silica-alumina, even inorganic minerals to prepare the hydrocracking catalysts, which not only can provide more hydrogenation active sites, but also play a role of coke carries (Looi et\u00a0al., 2012; Park et\u00a0al., 2019; Viet et\u00a0al., 2012). MoS2-amorphous-silica-alumina (ASA) catalyst employed in the slurry-phase hydrocracking promoted the cracking reaction and changed favorably the product distribution (Sanchez et\u00a0al., 2018). Puron et\u00a0al., (2013) found that NiMo/Al2O3 catalyst exhibited higher asphaltenes conversion with lower coke deposition and a reduced gas yield at achieving similar VR conversion compared with NiMo/ASA catalyst in the slurry-phase hydrocracking of Maya VR, due to its larger pore lessening diffusion limitation of asphaltenes molecules. Sahu et\u00a0al., (2016) reported that Ni\u2013Mo supported on goethite catalyst showed VR conversion of 80\u00a0wt% with the low boiling point liquid products of about 70\u00a0wt%, and found the products distribution depending on the physical and chemical properties of the catalysts. In comparison with the chemical synthesis materials of ASA and Al2O3, the natural minerals not only have an advantage in very low cost, but also contains some metals such as Ti and Zn, especially Fe, which can transform to sulfided iron (Fe1-x\nS) acted as the hydrocracking active sites (Du et\u00a0al., 2018). Cortes et\u00a0al., (2019) employed the modified vermiculite as support to prepare the hydrocracking catalyst for Athabasca Bitumen, and observed that Fe in vermiculite favored to improve the catalyst activity for the Bitumen conversion. Our group has studied the catalyst supported on natural bauxite modified by acid-treatment and hydrothermal method for hydrocracking of coal tar, the results shows that the acid-treatment and hydrothermal modifications can enhance the catalyst performance in the hydrocracking process (Yue et\u00a0al. 2016, 2018). However, the catalyst presented lower conversion even though using relatively light oil of coal tar with boiling point higher than 510 oC fraction less than 40\u00a0wt% as feedstock, while the processes of acid-treatment and hydrothermal modification led to wastewater discharge and increasing energy consumption.In this study, natural rectorite after calcination was used as support to prepare Mo catalysts for the slurry-phase hydrocracking of VR, in which the effect of calcination modification on the natural rectorite properties and catalyst performance was investigated. The supports of calcined rectorite and catalysts were characterized by XRD, FTIR, Py-FTIR, H2-TPR and XPS, as well as the catalyst performance was evaluated in an autoclave reactor with VR as feedstock to principally examine VR conversion and the yield of naphtha and middle distillates in the hydrocracking process. This work is significant for the development of high-efficiency and low-cost catalysts for the slurry-phase hydrocracking.Vacuum residue (VR) used as feedstock for the slurry-phase hydrocracking was supplied by China Petrochemical Corporation, its properties are shown in Table\u00a01\n.Natural rectorite was supplied by Zhongxiang Rectorite Co., Ltd. (Hubei Province, P. R. China), its chemical analysis compositions are shown in Table\u00a02\n. The calcination modification of natural rectorite was conducted at 450, 500 and 600 oC, which are denoted as Rec-450, Rec-500 and Rec-600. Mo supported on calcined rectorite catalyst was prepared by using the incipient wetness impregnation method with aqueous solution of ammonium molybdate tetrahydrate ((NH4)6Mo7O24\u00b74H2O, Adamas, 98%). The wet catalyst after set at 25 oC for 12\u00a0h was dried at 120\u00a0\u00b0C for 10\u00a0h, and then calcined at 500\u00a0\u00b0C for 4\u00a0h. The content of MoO3 in catalyst is 5\u00a0wt%. The catalysts are designated as Rec-Mo, Rec-450-Mo, Rec-500-Mo and Rec-600-Mo.X-ray diffraction (XRD) analysis of the sample was performed on an Ultima IV diffractometer with Co K\u03b1 radiation at 40\u00a0kV and 40\u00a0mA, and the pattern was recorded in the 2\u03b8 range from 5 to 90\u00b0 with a step of 0.02\u00b0. Fourier transform infrared (FTIR) was employed to examine the framework structure of natural rectorite, which was carried out on a Nicolet iS 10 spectrometer with the sample diluted by KBr at the ratio of 1:100, and the FTIR spectra were recorded in the wavenumber range from 4000 to 400\u00a0cm\u22121 with scans of 32. Pyridine adsorbed FTIR (Py-FTIR) of sample was conducted after heated at 350\u00a0\u00b0C for 5\u00a0h under a vacuum of 1.3\u00a0\u00d7\u00a010\u22123\u00a0Pa, and the adsorption of pure pyridine vapor at 30\u00a0\u00b0C for 20\u00a0min was followed. Adsorbed pyridine was removed by evacuating at 200 and 350\u00a0\u00b0C. H2 temperature programmed reduction (H2-TPR) of natural rectorite and catalyst was carried out on an ASAP-2920 instrument with a thermal conductivity detector (TCD). The sample was pretreated at 300\u00a0\u00b0C for 30\u00a0min under Ar atmosphere, and then heated from 50 to 950\u00a0\u00b0C with a rate of 10\u00a0\u00b0C/min in the 10\u00a0vol% H2/Ar stream. X-ray photoelectron spectroscopy (XPS) analysis of catalyst was performed on a Thermo Scientific ESCALAB 250Xi instrument with a monochromatic Al K\u03b1 source, and C1s peak with a binding energy of 284.6\u00a0eV and Al 2p peak with a binding energy of 74.6\u00a0eV were used as the references to calibrate the binding energy scale of Mo species. XPS spectra were recorded by using a XPSPEAK41 software after a background subtraction, and Gaussian-Lorentzian function was used for the spectrum deconvolution. The relative concentrations of the species of MoS2, MoSxOy and Mo6+ oxide for each sulfided catalyst were determined through their corresponding peak area. For example, the relative MoS2 concentration was calculated as following:\n\n[MoS2](%)\u00a0=\u00a0AMoS2/(AMoS2+AMoSxOy\u00a0+\u00a0AMo\n6+)\n\nwhere AX represents the peak area of species X.The slurry-phase hydrocracking performances of catalysts were evaluated in a 300\u00a0mL stainless-steel autoclave reactor using VR as a feedstock. VR, catalyst and appropriate sulfur powder were loaded into the reactor with the catalyst content of 5\u00a0wt%. Prior to the hydrocracking reaction, the catalyst was presulfurized with sulfur powder at 250 oC for 35\u00a0min and 350 oC for 35\u00a0min. The hydrocracking of VR was carried out at 420 oC under an initial H2 pressure of 13\u00a0MPa with the H2 to oil ratio (v/v) of 1000 for 90\u00a0min with vigorous agitation. After the hydrocracking reaction, the autoclave reactor was rapidly cooled to ambient temperature. The mixture of product and catalyst was collected, and then centrifuged to separate the liquid product and solid residue. The solid residue was washed with toluene to obtain the toluene insoluble, which included catalyst and coke. The liquid product was divided into four fractions on the basis of the boiling point (BP) in a decompression distillation plant, which are naphtha (BP\u00a0<\u00a0180 oC), middle distillates (BP 180\u2013350 oC), vacuum gas oil (VGO, BP 350\u2013520 oC) and VR (BP\u00a0>\u00a0520 oC). VR conversion and the yields of gas, naphtha, middle distillates, VGO and coke were calculated as following:\n\nVR conversion (%)\u00a0= (mass of >520 oC fraction in feed\u00a0>\u00a0mass of >520 oC fraction in product) / (mass of >520 oC fraction in feed)\u00a0\u00d7\u00a0100%\n\n\n\n\nGas yield (wt%)\u00a0= (mass of gas in product) / (mass of the feed)\u00a0\u00d7\u00a0100%\n\n\n\n\nNaphtha yield (wt%)\u00a0= (mass of <180 oC fraction in product / mass of the feed)\u00a0\u00d7\u00a0100%\n\n\n\n\nMiddle distillates yield (wt%)\u00a0= (mass of 180\u2013350 oC fraction in product \u2013mass of 180\u2013350 oC fraction in feed) / (mass of the feed)\u00a0\u00d7\u00a0100%\n\n\n\n\nVGO yield (wt%)\u00a0= (mass of 350\u2013520 oC fraction in product \u2013 mass of 350\u2013520 oC fraction in feed) / (mass of the feed)\u00a0\u00d7\u00a0100%\n\n\n\n\nCoke yield (wt%)\u00a0= (mass of coke) / (mass of the feed)\u00a0\u00d7\u00a0100%\n\n\n\nFig.\u00a01\n shows the XRD patterns of raw rectorite and rectorites calcined at different temperatures. The peaks at about 2\u03b8\u00a0=\u00a07.0o and 19.8o were assigned to the characteristic diffraction peaks of rectorite crystalline structure (Zhang et\u00a0al., 2010; Bao et\u00a0al., 2019), while there were some impurities such as rutile TiO2 corresponding to 2\u03b8\u00a0=\u00a027.4o and hematite Fe2O3 corresponding to 2\u03b8\u00a0=\u00a033o and 35.5o in the raw nature rectorite (Nguyen-Huy et\u00a0al., 2012; Liu et\u00a0al., 2019), because the rectorite contained Fe2O3 of 8.2\u00a0wt% and TiO2 of 4.0\u00a0wt% on the basis of the chemical compositions analysis in Table\u00a02. The intensity of diffraction peaks of rectorite crystalline structure decreased after calcination, especially the peak at 2\u03b8\u00a0=\u00a07.0o, indicating that the crystalline structure of rectorite was destroyed during the calcination process, possibly due to the loss of interlay water under the high temperature calcining. No obvious change was observed for the diffraction peaks of nature rectorites calcined at different temperatures, indicating the calcination temperature higher than 450 oC had no remarkable effect on the rectorite crystalline structure.FT-IR spectra of nature rectorites calcined at different temperature are shown in Fig.\u00a02\n. The peaks at 3640 and 3430\u00a0cm\u22121 observed in IR spectra of rectorites were assigned to the bending vibration of hydrogen band of the hydroxyl stretching in SiOH and interlaminar water, respectively. The peak at 1040\u00a0cm\u22121 was associated with the in-plane Si\u2013O\u2013Si stretching vibration, and the peak at 705\u00a0cm\u22121 was ascribed to the bending vibration of Si\u2013O\u2013Al (Zheng et\u00a0al., 2013). The peak at 3640\u00a0cm\u22121 corresponding to the hydrogen band vibration of SiOH hydroxyl stretching disappeared after rectorite calcination because of the \u2013OH dropping from SiOH structure. Moreover, the intensity of peak at 3430\u00a0cm\u22121 dramatically decreased for the rectorite calcined at 600 oC, nearly disappeared, indicating that the interlaminar water fell off from the rectorite structure when heated at 600 oC. In addition, all IR peaks positions of rectories had no shift during the calcination process.The acid properties of the calcined rectorites were investigated by Py-FTIR. The Py-FTIR spectra of the calcined rectorites measured at 200 and 350 oC are shown in Fig.\u00a03\n. The band at 1450 and 1540\u00a0cm\u22121 are assigned to the Lewis (L) and Br\u00f8nsted (B) acid sites, respectively (Tan et\u00a0al., 2008; Wei et\u00a0al., 2019). No obvious peak was observed for raw rectorite at both of 200 and 350 oC in the Py-FTIR spectra. While there was slightly weak peak at 1450\u00a0cm\u22121 for the calcined rectorite measured at 200 oC. The detail amounts of acid sites at 200 and 350\u00a0\u00b0C for the calcined rectorites are summarized in Table\u00a03\n. It is found that the amount of acid sites of calcined rectorites had no distinct increase compared with that of raw rectorite, which were less than 10\u00a0\u03bcmol/g for both L and B acid sites, suggesting that the calcined modification did nearly not increase acid sites on the rectorite.\nFig.\u00a04\n displays H2-TPR profiles of nature rectorites calcined at various temperatures. The reduction peaks of H2-TPR profiles can reflect the hydrogen consumption. Raw rectorite exhibited two reduction peaks, which were at center of 495 and 770 oC, respectively. Nature rectorite contained Fe2O3 with the content of 8.2\u00a0wt% more than other metallic oxides, except for Al2O3 and SiO2, according to the chemical compositions analysis of rectorite. Thus, the peaks of rectorite were considered as the reduction peaks of iron oxides species. The reduction steps of iron oxides species in hydrogen usually follows as: from Fe2O3 to Fe3O4, and then from Fe3O4 to FeO, finally from FeO to Fe, as reported in the literatures (Mogorosi et\u00a0al., 2012; Cheng et\u00a0al., 2015). Hence, the peak at center of 495 oC was ascribed to the reduction of hematite (Fe2O3) to magnetite (Fe3O4), and the peak at center of 770 oC was attributed to formation of Fe though the reduction of FeO. The rectorite calcined at 450 oC also showed two reduction peaks, and the reduction peak positions were similar with that of raw rectorite, suggesting no new iron oxides species formation, but the area of peak at 495 oC was remarkably larger than that of raw rectorite, it may be that calcination made inert iron oxides into active iron phase. A peak at center of 615 oC appeared for the rectorite calcined at 500 oC, associating with the reduction of magnetite to FeO, and it became stronger accompanied with the peak at 490 oC becoming weaker for the rectorite calcined at 600 oC, indicating that some amount of hematite converted into magnetite in the rectorite when calcined at the temperature higher than 500 oC.H2-TPR profiles of Mo catalysts supported on rectorites calcined at different temperatures are shown in Fig.\u00a05\n. It is clear that there were two reduction peaks centered at the temperature of 560 and 780 oC for all Mo catalysts supported on rectorites, in which the peak at 560 oC was ascribed to the reduction of octahedral Mo6+ species to tetrahedral Mo4+ species, Fe2O3 species to Fe3O4 species and Fe3O4 species to FeO species, as well as the peak at 780 oC was attributed to the tetrahedral Mo4+ species to Mo and FeO species to metallic iron. In comparison with the catalyst supported on raw rectorite, the two reduction peaks positions had no obvious shift for the catalyst supported on calcined rectorites, however, remarkable change of peak area was observed for the catalyst supported on calcined rectorites, especially the low-temperature reduction peak. The low-temperature reduction peak area of catalysts supported on calcined rectorites decreased compared with that of catalyst supported on raw rectorite, and it changed in the order of Rec-Mo\u00a0>\u00a0Rec-500-Mo\u00a0>\u00a0Rec-450-Mo\u00a0\u2248\u00a0Rec-600-Mo, indicating the amount of octahedral Mo6+, Fe2O3 and Fe3O4 species on catalyst of rectorite calcined at 500 oC was slightly less than that on catalyst of raw rectorite, but more than those on catalyst of rectorite calcined at 450 and 600 oC, it may be that the change of hydroxyl on rectorite during the calcination had an effect on the Mo species phase formation on the catalysts surface.The chemical surface compositions of sulfided catalysts supported on rectorites were investigated by XPS. Fig.\u00a06\n displays Mo 3d XPS spectra and deconvolution results of sulfided catalysts supported on rectorites calcined at various temperatures. The Mo 3d XPS spectra include three doublets, the doublets with bending energy at 229.3 and 232.5\u00a0eV are related to Mo 3d5/2 and Mo 3d3/2 levels for the Mo4+ in MoS2 phase species, the doublets with bending energy at 230.5 and 233.8\u00a0eV are ascribed to Mo 3d5/2 and Mo 3d3/2 levels for the Mo5+ in MoSxOy oxysulfide species, and the doublets with bending energy at 232.7 and 235.9\u00a0eV are assigned to Mo 3d5/2 and Mo 3d3/2 levels for the Mo6+ in MoO3 oxide species (Nikulshin et\u00a0al., 2014; Pimerzin et\u00a0al., 2017; Zhou et\u00a0al., 2018). The deconvolution results of different catalysts based on XPS spectra are summarized in Table\u00a04\n. It is found that the sulfidation degree of Mo species (Mo4+ proportion) on the calcined rectorites catalysts was higher than that on raw rectorite catalyst, especially the catalyst on rectorite calcined at 500 oC, it may be ascribed that calcination modulated the surface hydroxyl on rectorite, further influenced the interaction of Mo species and rectorite. In general, the catalyst with high sulfidation degree presents high hydrogenation activity in the hydrocracking and/or hydrotreating process (Nikulshin et\u00a0al., 2014; Pimerzin et\u00a0al., 2017; Cui et\u00a0al., 2013; Liu et\u00a0al., 2020).The performance of catalyst supported on calcined rectorite was evaluated in the slurry-phase hydrocracking of VR at 420\u00a0\u00b0C under an initial H2 pressure of 13.0\u00a0MPa. Fig.\u00a07\n shows VR conversions of the different catalysts. The VR conversion of Rec-Mo catalyst was about 77.0\u00a0wt%, and no notable change of the conversion was observed for the catalysts supported on calcined rectorites, exceptionally Rec-450-Mo catalyst with VR conversion of 81.0\u00a0wt%. It has been reported that the slurry-phase hydrocracking over oil dispersed catalyst or catalyst with few acid sites occurred above 420 oC was considered as a thermal cracking reaction path accompanied with hydrocracking reaction path (Kim et\u00a0al., 2017; Matsumura et\u00a0al., 2005; Nguyen et\u00a0al., 2013). The thermal cracking reaction followed free radical mechanism, mainly depending on the reaction temperature, while hydrocracking reaction followed carbenium ion mechanism, which was principally affected by the acid sites of catalysts at the same reaction condition. There were few acid sites in rectorites examined by Py-FTIR, shown in Table\u00a03. Hence, it is concluded that the almost same VR conversions for all catalysts were attributed to the slurry-phase hydrocracking controlled by the thermal cracking reaction following free radical mechanism.The products distribution is significant for the conversion of heavy and poor feedstocks into light oil fractions in the refining industry, in which more amount of the valuable fractions of naphtha and middle distillates corresponding to the boiling point range of gasoil and diesel was expected to produce, but gas and coke as worthless product, especially coke causing negative effect on the catalyst and reactor, were as infamous fraction. The yields of naphtha and middle distillates are shown in Fig.\u00a07, and the products distribution of the catalysts supported on rectorites calcined at various temperatures are shown in Fig.\u00a08\n. The yields of naphtha and middle distillates for the various catalysts increased following as Rec-Mo (40.4\u00a0wt%)\u00a0<\u00a0Rec-600-Mo (52.7\u00a0wt%)\u00a0<\u00a0Rec-450-Mo (53.5\u00a0wt%)\u00a0<\u00a0Rec-500-Mo (61.7\u00a0wt%), indicating that the calcination of rectorite was beneficial for the enhancement of the yields of naphtha and middle distillates in the slurry-phase hydrocracking process.The detail yields of gas, naphtha, middle distillates, VGO, residues and coke of the various catalysts were distinctly different, as shown in Fig.\u00a08. Rec-Mo catalyst presented the naphtha yield of 14.9\u00a0wt% and middle distillates yield of 25.5\u00a0wt%, both of which obviously increased for the catalysts supported on calcined rectorites, especially Rec-500-Mo catalyst with the naphtha yield of 25.4\u00a0wt% and middle distillates yield of 36.3\u00a0wt%. Moreover, the gas yield of Rec-Mo catalyst was up to 28.2\u00a0wt%, distinctly higher than that of the catalysts supported on calcined rectorites, and the gas yield of Rec-500-Mo catalyst reduced to 7.9\u00a0wt%. In addition, there was no remarkable change on the coke yield for the various catalysts. It is concluded that the catalysts supported on calcined rectorite had better performance compared with Rec-Mo catalyst on the basis of product distribution, especially Rec-500-Mo catalyst, it is attributed that the higher hydrogenation activity of the catalyst restrained over cracking reaction of intermediate product to produce gas.In this study, the catalysts supported on natural rectorite were prepared, and the effect of calcination modification on the catalysts properties was examined. The catalyst performance was conducted in an autoclave reactor at 420 oC and an initial H2 pressure of 13\u00a0MPa. The reaction results show that the catalysts supported on calcined rectotire exhibited similar VR conversions with the catalyst supported on raw rectorite, it is ascribed that the thermal cracking reaction following free radical mechanism controlled the reaction process, because there were few acid sites on the catalyst surface. However, the yields of naphtha and middle distillates for the catalysts supported on calcined rectorite were obviously higher compared with that on raw rectorite, especially the yields of naphtha and middle distillates over Rec-500-Mo catalyst up to 61.7\u00a0wt%, indicating that the calcination of rectorite was beneficial for improving the yields of naphtha and middle distillates in the slurry-phase hydrocracking process, it is attributed that the higher sulfidation degree of molybdenum oxide species on catalyst promoted the hydrogenation reaction, thus inhibited the over-cracking reaction of intermediate product to produce gas. This study is significant for the development of high-efficient and low-cost catalyst for the slurry-phase hydrocracking of heavy and poor oil.The authors acknowledge National Key Research and Development program (2018YFA0209403) and National Natural Science Foundation of China (Youth) program (21908027) for financing this research.", "descript": "\n                  In order to develop high-efficiency and low-cost catalyst for the slurry-phase hydrocracking of vacuum residue (VR), the catalyst supported on natural rectorite was prepared, and the effect of calcination modification of rectorite on the catalyst properties and performance was investigated. The support of rectorite and catalyst were characterized by XRD, FTIR, Py-FTIR, H2-TPR and XPS to examine their structures and properties. The comparative reaction results show that VR conversions for the catalysts supported on calcined rectorite were similar with that on raw rectorite, possibly due to the VR cracking reaction controlled by the thermal cracking following free radical mechanism because of few acid sites observed on the catalysts surface. However, the yields of naphtha and middle distillates for the various catalysts were obviously different, and increased following as Rec-Mo (40.4\u00a0wt%)\u00a0<\u00a0Rec-600-Mo (52.7\u00a0wt%)\u00a0<\u00a0Rec-450-Mo (53.5\u00a0wt%)\u00a0<\u00a0Rec-500-Mo (61.7\u00a0wt%), indicating that the calcination of rectorite favored the enhancement of the yields of naphtha and middle distillates for the catalyst in the slurry-phase hydrocracking process, it is attributed that the higher sulfidation degree of molybdenum oxide species on the catalyst surface promoted hydrogenation reaction, thus restrained over-cracking reaction of intermediate product to produce gas.\n               "}
{"full_text": "Due to rising environmental concerns and challenges in satisfying future demand, the demand for renewable raw materials to replace petroleum oil-based products is significantly increasing [1]. It is one of the green synthesis choices that can help with long-term sustainability [2]. Vegetable oils (VOs), which currently account for the majority of renewable feedstocks used to make bio-based products, might be a viable alternative for the production of bio-based products [3]. Hence, the use of VOs as a potential feedstock for a variety of functional materials and their applications in a variety of sectors has recently gained greater attention. Long-chain fatty acid triglyceride esters are utilized as a trustworthy starting material for the production of a wide range of bio-based fuels and chemical products [4]. However, the degree of unsaturation present in the oil causes rancidity, stability, lubricity, compatibility, and chemical degradation problems which limit their use as petroleum oil alternatives [5\u20137]. As a result, the scientific community is paying more attention to the functionalization of VOs via epoxidation. Non-edible vegetable oils, such as castor oil [8], jatropha oil [9], Cynara cardunculus seed oil [1], cottonseed oil [10], Mahua oil [11], etc., have emerged as possible alternatives to address low-cost material demands without competing with food crops. Argemone mexicana oil is an ideal choice for making epoxidized oil, which is a renewable alternative for a variety of applications. This is because of their inherent biodegradability, availability, sustainability, non-toxicity, and ease of chemical modification of VO, as well as environmental concerns and limited supply of petroleum [12].Argemone mexicana oil (AMO) is a non-edible seed oil derived from Mexican prickly poppy seeds, an annual growing weed plant in the Papaveraceae family [13]. It is primarily composed of triglycerides of unsaturated long-chain fatty acids with high linoleic acid (36.6\u201361.4%) and oleic acid (18.5\u201340%), which are playing an important role in the epoxidation process [14]. Epoxidation is the most common VOs modification method for functionalizing ethylenic double bonds in VOs and converting them into the highly reactive epoxy group. They are adaptable building blocks for making bio-based products like plasticizers, lubricants, PVC stabilizers, and surface coating formulations [8,15,16]. This can also result in a variety of stable products with a highly reactive oxirane ring that aids in the investigation of a variety of chemical reactions [17]. Generally, the epoxidation of vegetable oils is performed with peroxy acid formed in situ since H2O2 has very low solubility in vegetable oils [11,17\u201319]. Acetic acid is more selective than formic acid, because the side reaction of peracetic acid decomposition is lower than performic acid, and the side reaction of ring opening is lower than with formic acid [20]. Consequently, epoxidized oil has increased viscosity, lubricity, oxidative stability, compatibility, and thus making the materials more susceptible to microbial degradation, all of which are crucial characteristics of epoxy products [3,12,21,22]. Various studies have been reported on the epoxidation of VOs with peroxy acids formed in situ utilizing homogenous catalysts, acidic ion exchange resins (AIER), or biocatalysts for epoxy formation with a wide range of applications, particularly for PVC plasticizers, and lubricants [10,15,21,23]. During epoxidation employing peroxy carboxylic acid as an oxidizer catalyzed by a homogenous catalyst, mainly sulphuric acid, produces commercially viable epoxidized oils [19]. [24] used peracetic acid to make epoxy canola oil employing sulphuric acid catalyst to convert ethylenic unsaturation to oxirane (81%) at 7\u00a0h. However, they can cause a variety of undesirable side reactions as well as corrosion problems [25,26]. Moreover, enzymes are also used as effective biocatalysts for the epoxidation of VOs in recent years [27]. used immobilized lipase as a biocatalyst for epoxidation of genetically modified high oleic acid soybean oil with or without free fatty acid and toluene, resulting in the epoxide conversion of 95% at 35\u00a0\u00b0C. The chemo-enzyme epoxidation of different VOs such as soybean oil [28], Sapindus Mukorossi seed oil [29], and Karanja oil [30] with other biocatalysts such as Novozym 435, have been reported using hydrogen peroxide as an oxidant. But, the chemo-enzyme epoxidation of VOs is strongly influenced by hydrogen peroxide concentration as well as high temperatures, resulting in enzyme deactivation [26]. The employment of heterogeneous catalysts, including AIER for epoxidation of VOs, would be more advantageous in terms of separation ease, reusability, eco-friendly, and cost effectiveness [31]. [32] have reported on Karanja oil (iodine value, 89\u00a0g/100g) epoxidation with peracetic acid catalyzed by Amberlite IR-120 catalyst. The researchers have studied the effects of stirring speed, molar ratio of hydrogen peroxide to ethylenic double bond in the oil, molar ratio of acetic acid to ethylenic double bond in the oil, temperature, and catalyst loading for the epoxidation process [24]. also described on epoxidation of canola oil with in situ generated peroxy acetic acid using Amberlist IR120H resin as a catalyst and obtained oxirane oxygen content of 90% at 7\u00a0h.Metal oxide solid acid catalysts are used in a variety of organic synthesis processes, including aromatic nitration, esterification, transesterification, and epoxidation [33]. Thus, heterogeneous metal oxide catalysts are preferred to circumvent the limits of the aforementioned catalysts [3]. There have been a few reports on the epoxidation of soybean oil, canola oil, podocarpus falcatus seed oil, methyl oleate, and sunflower oil using heterogonous metal oxides as a catalyst such as sulfonated \u2013ion exchange resins [32], sulfonated- SnO2 [3], solid sulfonated silica acid [34], Ti\u2013SiO2 [18], Alumina [35], tungsten [36], respectively [34]. used sulfonated silica solid acid catalyst for epoxidation of podocarpus falcatus seed oil with hydrogen peroxide as an oxidizer. The maximal ethylenic conversion to oxirane was reported to be 84.75% under optimal conditions ethylenic double bond to H2O2 molar ratio of (2.5:1), catalyst loading (5%), temperature (70\u00a0\u00b0C), and time (4\u00a0h). Sulfated tin (IV) oxide is also classified as a super solid acid because of its high surface acidity and is employed in the majority of acid-catalyzed processes like esterification and transesterification [19]. [37] reported that a sulfate-doped metal oxide surface can function as a solid acid and an oxidative catalyst [3]. also used sulfated tin (IV) oxide to epoxide unsaturation in canola oil using peroxyacetic acid produced in situ from H2O2 and acetic acid. They reported a maximum epoxide conversion of 100% at optimum epoxidation conditions [33]. studied solid acid sulfated \u2013 Zirconia for effective epoxidation of castor oil. Because certain sulfate-doped metal oxides such as SnO2, ZrO2, TiO2, and Al2O3 have both Lewis and Br\u00f8nsted acid sites derived from metal oxides and sulfates doped on the surface of metal oxides, they have been widely used as a solid acid in organic chemical modifications. Furthermore, as compared to metal oxides without sulfate, they produce super acid materials with high surface acidity and significantly larger surface areas [38]. As a result of their exceptional catalytic activity, they have attracted a lot of attention and are frequently utilized as a solid acid catalyst for a range of organic modifications. As far as the authors knowledge, sulfated tin (IV) oxide solid acid-catalyzed epoxidation of Argemone mexicana oil with peroxy acetic acid formed in situ has not been reported elsewhere.Various researchers have studied kinetic modelling strategies to estimate kinetic constants for the epoxidation of cottonseed oil by Prileschajew method [39]. have used a semi-batch reactor to study kinetic modelling for the epoxidation of cottonseed oil with performic acid by Prileschajew method. The results of their study showed that the reaction enthalpy of epoxidation and ring opening was \u2212230\u00a0kJ/mol and \u221290\u00a0kJ/mol, respectively with initial reaction conditions of 50\u201370\u00a0\u00b0C, an organic phase 30\u201340%, formic acid 0.02\u20130.05\u00a0mol/min and time 25\u201350\u00a0min [40]. used a kinetic model under adiabatic conditions to investigate the variables affecting the risk of thermal runaway for the epoxidation of cottonseed oil. It has been noted that adiabatic temperature rise and time to maximum rate were sensitive to the content of acetic acid and hydrogen peroxide [41]. estimated the kinetic constants for the epoxidation of cottonseed oil by peroxyacetic acid using a batch reactor. The authors developed a kinetic modelling technique to predict kinetic constants for the ring opening reaction involving water, acetic acid, and peracetic acid. They reported that ring opening by acetic and peracetic acids more quickly than water and hydrogen peroxide.Therefore, this study aimed to synthesize and characterize sulfated\u2013tin (IV) oxide solid acid as a heterogeneous catalyst for AMO epoxidation with peroxy acetic acid formed in situ. The influences of various AMO epoxidation parameters (viz. molar ratio of the ethylenic double bond in the AMO to H2O2, molar ratio of the ethylenic double bond in the AMO to acetic acid, catalyst concentration, and reaction temperature) were investigated. The physicochemical characteristics of AMO and its epoxidized oil (EAMO) were also examined. Moreover, a kinetic model for AMO epoxidation was analyzed to proceed to the acceptable degree of double bond conversion.Hydrogen peroxide (30%), glacial acetic acid (99.5%), ammonia solution (30%), iodine crystals, HBr solution (48%), sodium thiosulphate, anhydrous sodium sulfate, sulphuric acid (98%), ethyl acetate, Stannous chloride dihydrate (SnCl2.2H2O, 97%) and chloroform (99%) were purchased from Sigma-Aldrich (Germany). The other chemicals and reagents utilized in this experiment were analytical grade.Oil was extracted from Argemone mexicana seed (AMS), collected from Addis Ababa, Ethiopia, using the soxhlet method with chloroform as the solvent. The maximum oil yield was achieved at a temperature near the boiling point of the corresponding solvent [42]. After the complete extraction process, the extracted oil was separated from the solvent using a rotary evaporator and vacuum pump at 70\u00a0\u00b0C. Prior to the epoxidation process, the obtained oil was refined and stored at \u22124\u00a0\u00b0C for further use in the epoxidation process.In this study, sulfate group-doped tin (IV) oxide solid acid catalyst for the AMO epoxidation was prepared by applying the chemical co-precipitation method [3,43\u201345]. 75 g of SnCl2.2H2O (97%) and 1.5\u00a0L deionized water were mixed with a continuous stir followed by a dropwise addition of 30% of aqueous ammonia solution to maintain the desired pH of \u223c9.0. At room temperature, a white tin hydroxide (Sn (OH)4) powder gel was precipitated after 6\u00a0h. The resultant gel was filtered using Whatman's filter paper and rinsed with distilled water until a neutral solution. The gel was then oven-dried at 100\u00a0\u00b0C for overnight. Then, 20\u00a0g of the obtained gel powder was impregnated with 300\u00a0mL of 1\u00a0M sulphuric acid solution for 1\u00a0h. Further acid-treated tin hydroxide powder gel was oven-dried at 100\u00a0\u00b0C for 12\u00a0h. Both oven-dried tin hydroxide and its related acid-treated gel powder were calcined at 500\u00a0\u00b0C for 4\u00a0h. Finally, several surface characterizations were performed on the resulting pure tin (IV) oxide and sulfate group doped-tin (IV) oxide catalysts using FTIR, XRD, BET/BJH, DSC, TGA, SEM-EDX methods, thereby understanding catalyst activity.FTIR (Thermo fisher FTIR spectrometer-Nicolet iS50), at 4\u00a0cm\u22121 resolution with KBr as a background matrix in the range of 4000\u2013400\u00a0cm\u22121, were used to determine the formation of pure tin (IV) oxide and its sulfated solid acid catalyst. In FTIR analysis, 5\u00a0mg of both resultant catalysts and 95\u00a0mg of KBr crystal were thoroughly mixed, grounded into a fine powder, and then pelletized using a hydraulic press at 10 tons. Thus, the functional groups present on the catalyst surfaces were identified using this technique.XRD was also conducted on a diffractometer with Ni-filtered CuK\u03b1 radiation at \u03bb\u00a0=\u00a00.154\u00a0nm in the 2\u03b8 range of 10\u201380\u00b0 and thus the crystal structure and sizes of both catalysts were determined. Accordingly, the mean crystal sizes of the catalysts were calculated using the Debye Scherer equation as indicated in eq. (1).\n\n(1)\n\n\nD\n=\n\n\n0.9\n\u03bb\n\n\n\u03b2\n\ncos\n\n\u03b8\n\n\n\n\n\nwhere D \u2013denotes the average diameter of crystalline size (nm), \u03bb \u2013denotes the wavelength of CuK\u03b1 radiation at 0.154\u00a0nm, \u03b2 \u2013denotes full width at half maximum intensity (FWHM) in radian, and \u03b8 \u2013denotes for Bragg angle (\u0585).Brunauer - Emmett-Teller (BET) devices based on adsorption and desorption of N2 gas isotherms via Quantachrome Nova 2200e Surface Area analyzer (USA) was used to characterize surface areas of both resultant catalysts (pure tin (IV) oxide and it\u2019s sulfate doped - tin (IV) oxide solid acid catalyst). The specific surface areas of both catalysts were determined from a nitrogen adsorption study conducted at a low temperature (\u2212196.15\u00a0\u00b0C) using the high vacuum conventional volumetric glass system and were evacuated at 250\u00a0\u00b0C for 2\u00a0h before exposure to nitrogen gas at \u2212196.15\u00a0\u00b0C under reduced pressure (10\u22125tor). Besides, the micrometric Pore Size analyzer Barret \u2013Joyner Halenda (BJH) technique was used to evaluate the pore volume and average pore diameter of the obtained catalysts.Thermal gravimetric analysis (TGA) was also employed to study the thermal stability of the resultant catalysts under the temperature range of 25\u2013750\u00a0\u00b0C using TA instruments with SDT Q600 under nitrogen flow in which weight losses were evaluated. Then, further confirmation analysis was conducted using differential scanning calorimetry (DSC).All epoxidation processes were conducted in 500\u00a0mL three-necked round bottom flasks with a magnetic stirrer and placed in the hot plate's temperature-controlled water bath. One side of the flask was inserted with a thermometer and used to measure reaction temperature, while the other middle neck of the flask was fitted to a water-cooled reflux condenser. Primarily, the required amount of AMO was mixed with acetic acid and sulfated tin (IV) oxide solid acid catalyst at 30\u00a0\u00b0C at a stirring speed of 1000\u00a0rpm. Then, 30% of hydrogen peroxide solution was added dropwise to the reaction mixture in the first 30\u00a0min. The reaction time recording was begun once this oxidizer was completely added to the reaction mixture. The influence of reaction conditions, such as molar ratio of the ethylenic double bond in the oil to H2O2 and acetic acid, catalyst concentration, and the reaction temperature were studied. Upon completion of the epoxidation process, catalysts were removed by filtration. Before analysis, the reaction products were collected and ethyl acetate (50\u00a0mL) was used to separate the aqueous and organic oil layers periodically, then washed with both NaHCO3 (5%) solution and distilled water until pH \u223c7, and the trace amount of water and other impurities was absorbed with anhydrous Na2SO4. Finally, the resultant epoxy products were separated from ethyl acetate using a rotatory evaporator. The epoxy oxygen content was analyzed using AOCS Cd 9\u201357 methods in which 0.1\u00a0N hydrobromic solution in acetic acid (glacial) was used as a titrate (as per eq. (2).), and the ethylenic double bond conversion into an epoxy group was determined in terms of iodine value (IV). The amount of double bonds in the oil is closely related to the iodine value (IV), which is an indicator of the overall unsaturation in the AMO. As a result, IV was determined to find out the quantity of double bonds in the oil. Hence, the ethylenic double bond conversion was investigated in terms of the iodine (IV) value measurements, using the AOCS Cd 1\u201325 method according to eq. (3). Further epoxidized oil formation confirmation analyses were conducted using FTIR, 1H NMR, and 13C NMR methods.\n\n(2)\n\n\nO\nO\nC\n=\n\n\n0.1\nM\n\nx\n\n1.6\n\nx\n\n\n(\n\nB\n\u2212\nV\n\n)\n\n\n\nw\ne\ni\ng\nh\nt\n\no\nf\n\ns\na\nm\np\nl\ne\n\n\n(\ng\n)\n\n\n\n\n\n\nwhere, OOC stands for oxirane oxygen content (%), B stands for volume of hydrobromic acid solution used to titrate a blank, and V stands for volume of hydrobromic acid solution used to titrate the test sample.\n\n(3)\n\n\nI\nV\n=\n\n\nM\n\nx\n\n12.69\n\nx\n\n\n(\n\nB\n\u2212\nS\n\n)\n\n\n\nw\ne\ni\ng\nh\nt\n\no\nf\n\ns\na\nm\np\nl\ne\n\n\n(\ng\n)\n\n\n\n\n\n\nwhere IV stands for iodine value, M stands for molarity of sodium thiosulphate, B stands for mL of sodium thiosulphate used to titrate a blank, and S stands for the mL of sodium thiosulphate used to titrate the test sample. In addition, the ethylenic DB conversion in the AMO to the epoxy group was determined using eq. (4) [46].\n\n(4)\n\n\nE\nt\nh\ny\nl\ne\nn\ni\nc\n\nD\nB\n\nc\no\nn\nv\ne\nr\ns\ni\no\nn\n\n\n(\n%\n)\n\n=\n\n\nI\n\nV\n0\n\n\u2212\nI\nV\n\n\nI\nV\n\n\nx\n100\n\n\n\nwhere, IV denotes the iodine value of the AMO before epoxidation in g I2/100g of oil, and IV is the iodine value of the EAMO in g I2/100g.To better understand the influences of epoxidation reaction conditions on epoxide conversion, the different experimental trials were conducted following the one-variable-at-a-time method. All epoxidation experiments were done with a constant 50\u00a0mL of AMO and 1000\u00a0rpm mixing speed for 6\u00a0h. In this study, the effect of sulfated \u2013 tin (IV) oxide solid acid catalyst loading varied from 5 to 15% with corresponding to the weight of AMO on the ethylenic double bond conversion and epoxy oxygen ring content, while other reaction parameters such as the molar ratio of an ethylenic double bond in the AMO to 30% of H2O2 1:3, the ethylenic double bond in the AMO to acetic acid ratio 1:2, and reaction temperature 70\u00a0\u00b0C were taken from the literature [3]. The effects of the ratio of the ethylenic double bond in the AMO to hydrogen peroxide (30%) on the ethylenic double bond (DB) conversion and epoxy oxygen content were investigated by varying the range from 1:1 to 1:4\u00a0at fixed optimal catalyst loading value, the molar ratio of the ethylenic double bond the AMO to acetic acid 1:2, and reaction temperature (70\u00a0\u00b0C). The effect of the ratio of the ethylenic double bond in the AMO to acetic acid was varied from 0.5 to 2.5\u00a0at fixed other epoxidation parameters. The reaction temperature was also altered from 50 to 80\u00a0\u00b0C to investigate its impact on an ethylenic double bond conversion and thus epoxy oxygen content. The overall AMO epoxidation process is shown in Scheme 1\n.The physicochemical characteristics of the epoxidized product (EAMO) such as density, kinematic viscosity at 40\u00a0\u00b0C (mm2/s), kinematic viscosity at 100\u00a0\u00b0C (mm2/s), viscosity index, flash point (oC), epoxy oxygen ring content (%), and iodine values (g I2/100g of the AMO) were examined using established methods. Using a Rheometer (MCR 102, USA) instrument, the dynamic viscosity of AMO and its epoxidized oil (EAMO) were measured as a function of temperature ranging from 20 to 100\u00a0\u00b0C at a constant shear rate of 50 per second. Furthermore, the kinematic viscosity and viscosity index of both AMO and its epoxidized oil was determined using the determined dynamic viscosity value based on the ASTMD2270 standard table. The viscosity index (VI) of EAMO was calculated using eq. (5). The flash point of AMO and its epoxidized oil were also examined. According to a standard procedure, the iodine values (IV) of both samples were tested using eq. (3).\n\n(5)\n\n\nV\n\nI\nx\n\n=\n\n\n\n\u03b3\nA\n\n\u2212\n\n\u03b3\nx\n\n\n\n\n\u03b3\nA\n\n\u2212\n\n\u03b3\nB\n\n\n\nx\n100\n\na\nt\n\n40\n\n\no\n\nc\n\nx\n100\n\n\n\nWhere \n\nV\n\nI\nx\n\n\n denotes viscosity index of the AMO/EAMO, \n\n\n\u03b3\nx\n\n\n denotes kinematic viscosity of Epoxidized oil (EAMO) at 40\u00a0\u00b0C, \n\n\n\u03b3\nA\n\n\n and \n\n\n\u03b3\nB\n\n\n denotes kinematic viscosity of oil A and B at 40\u00a0\u00b0C are used as reference oil taken from ASTM-D2270-10 table for \n\n\n\u03b3\nx\n\n\n at 100\u00a0\u00b0C.FTIR spectra of obtained pure and sulfated tin (IV) oxide catalysts are presented in Fig. 1\na and b. The characteristic peaks at 600\u00a0cm\u22121 show the existence of O\u2013Sn\u2013O stretching. This revealed the complete conversion of tin (IV) hydroxide gel into SnO2 when calcined at 500\u00a0\u00b0C for 4\u00a0h. Similarly, bands at 1286, 1145, and 1018\u00a0cm\u22121 were observed in the spectra after sulfation of tin (IV) hydroxide gel with 1\u00a0M\u00a0H2SO4, indicating symmetric and asymmetric stretching frequencies of the sulfated group (SO) and confirming a bidentate chelation (linkage) mode between sulfate group and tin (IV) oxide. This further indicates the formation of sulfate doped \u2013 tin (IV) oxide as a solid acid catalyst after being calcined at 500\u00a0\u00b0C for 4\u00a0h. Because of the existence of sulfate doped on the surface of tin (IV) oxide catalyst, robust acidic properties can be recognized [38]. Thus, improving the properties of sulfate doped tin (IV) oxide catalyst is important for promoting the epoxidation of AMO with peroxyacetic acid generated in situ. Similar research works were reported by Refs. [43\u201345].The XRD spectra of attained pure tin (IV) oxide and sulfated tin (IV) oxide solid acid catalyst are illustrated in Fig. 2\na. The major absorption peaks at 26.63\u00b0, 33.90\u00b0, 37.99\u00b0, 51.81\u00b0, 54.80\u00b0, 61.91\u00b0, 64.78\u00b0, 65.99\u00b0, 71.75\u00b0 are related to diffraction from planes (110), (101), (200), (211), (220) of tin (IV) oxide particles. This shows the complete conversion of tin (IV) hydroxide gel into pure tin (IV) oxide by calcination at 500\u00a0\u00b0C for 4\u00a0h, and thus confirms the tetragonal crystal phase. Similarly, after sulfation of this gel with 1\u00a0M\u00a0H2SO4 treatment, XRD peaks at 26.63\u00b0, 33.93\u00b0, 37.98\u00b0, 51.84\u00b0, 54.76\u00b0, 61.91\u00b0, 64.82\u00b0, 65.98\u00b0, and 78.73\u00b0 are related to the above-mentioned diffraction planes. This indicates that the sulfate group doped on the catalyst's surface did not cause any crystalline changes. This is supported via the DCS plot in Fig. 2b, confirming that the prepared catalysts have a single phase. However, sulfation of tin (IV) oxide reduces the crystalline size of the obtained tin (IV) oxide thereby increasing the surface area of the catalyst which in turn enhances its catalytic activity. The mean crystalline sizes of pure tin (IV) oxide and tin (IV) oxide doped with the sulfate group were determined using the Debye Scherrer eq. (1) based on XRD peak width measurements. Accordingly, the calculated mean crystalline size of tin (IV) oxide and sulfated tin (IV) oxide solid acid catalyst was determined to be 35.62\u00a0nm and 16.64\u00a0nm, respectively. This smaller crystal size of the sulfate group-linked tin (IV) oxide catalyst is related to the addition of sulfate ions. This could be due to sulfate chelation on the catalyst surface, which prevents the tin (IV) oxide particles from coagulating during the calcination process at 500\u00a0\u00b0C for 4\u00a0h. As a result, the surface area of the sulfated tin (IV) oxide solid acid catalyst increased and thus improved its catalytic performance for the AMO epoxidation process. This demonstrates that sulphuric acid treatment has a significant impact on reducing crystalline size, increasing surface area, and so improves the catalytic activity of sulfated tin (IV) oxide for epoxidation. Comparable results were reported in the literature [3,45].Catalyst surface area is one of the critical parameters that have a significant influence on catalytic activity, and thus epoxide conversion. Brunauer \u2013 Emmett \u2013 Teller (BET) was employed to analyze the surface area of each prepared catalyst while Barret \u2013Joyner Halenda (BJH) method was used to characterize pore volume and average diameters. As indicated in Table 1\n the BET surface areas of Tin (IV) oxide and sulfate group linked tin (IV) oxide were 14.84\u00a0m2/g and 60.61\u00a0m2/g, respectively. The development of sulfate linkage with tin (IV) oxide gives its increased surface area for the sulfate group linked tin (IV) oxide solid acid catalyst. The deposition of sulfates on the surface of tin (IV) oxide increased its pore volume from 0.06 to 0.13\u00a0cm3/g and caused an increased in the average pore diameter from 10.97 to 11.20\u00a0nm, according to the BJH pore size distribution result. Sulfation of tin (IV) oxide enhanced pore volume and is anticipated to boost epoxide output [12,38].The thermal analyses of pure tin (IV) oxide and sulfated tin (IV) oxide after calcination at 500\u00a0\u00b0C are displayed in Fig. 3\na and b. The weight loss of both catalysts was determined in the temperature ranges of 25\u2013800\u00a0\u00b0C. Tin (IV) oxide was determined to be quite stable up to 800\u00a0\u00b0C, and the weight loss percentage change was found to be insignificant. Up to 600\u00a0\u00b0C, sulfated tin (IV) oxide was similar stability with very little weight loss, but after 600\u00a0\u00b0C, the weight loss increased due to the evolution of the sulfate group from the catalyst surface. Sulfated tin (IV) oxide loses weight when heated to 800\u00a0\u00b0C compared to pure tin (IV) oxide [3,38].The surface morphology of both tin (IV) oxide and sulfated tin (IV) oxide catalysts was characterized by SEM analysis. As indicated in Fig. 4\na and b, the surface morphology of both catalysts has no significant change after the impregnation of sulfate ions. Sulfation of the catalyst improved the catalytic oxidative activity of tin (IV) oxide surface as compared to non-sulfated metal oxide. Thus, sulfation is the key to boosting the conversion of Argemone Mexicana oil to its epoxidized oil [12,33,43].The effect of catalyst concentration on AMO epoxidation is shown in Fig. 5\na. In this study, the influence of sulfated tin (IV) oxide solid acid catalyst on the course of AMO epoxidation was evaluated. It was investigated by varying the amounts of the catalyst 5, 7.5, 10, and 15% of the corresponding weight of oil keeping the ratio of the ethylenic double bond in the oil to acetic acid and 30% H2O2 as 1:3:2 3. All epoxidation reactions were examined at a constant agitation speed of 1000\u00a0rpm at 70\u00a0\u00b0C. As illustrated in Fig. 5a, the ethylenic double bond (DB) conversion in the AMO to its epoxidized oil gradually increased with an increase in catalyst concentration up to 12.5% due to an increment in the active sites of the catalyst. The maximum double bond conversion of 95.05% and related epoxy oxygen content of 6.25 was achieved after 6\u00a0h. However, more upsurge in sulfated tin (IV) oxide catalyst concentration resulted in considerably the same conversion or less. This might be an increased rate of oxygen ring cleavage beyond the maximum value of the catalyst external surface active sites content during epoxidation [8]. In the current investigation, a sulfated tin (IV) oxide solid acid catalyst concentration of 12.5% was shown to be the best value for AMO epoxidation. The results of the study showed that pure tin (IV) oxide calcined at 500\u00a0\u00b0C has no substantial double bond conversion of AMO into an epoxy group under these experimental conditions [3].\nFig. 5b shows the effect of hydrogen peroxide (30%) on the course of AMO epoxidation at a catalyst concentration of 12.5%, temperature of 70\u00a0\u00b0C, and the molar ratio of the ethylenic double bond in the AMO to acetic acid 1:2. Hydrogen peroxide has a significant influence on in situ epoxidation [47]. Thus, the ratio of an ethylenic double bond in oil to H2O2 (30%) was varied at 1:1, 1:1.5, 1:2, 1:2.5, 1:3, 1:3.5, and 1:4 to study its influence on the in situ epoxidation of ethylenic double bond conversion in the AMO. As shown in Fig. 5b the rate of ethylenic double bond conversion increased with increasing the molar ratio of the ethylenic double bond in the AMO to H2O2 (30%). The molar ratio of the ethylenic double bond in AMO to hydrogen peroxide of 1:2.5 resulted in the maximum ethylenic double bond conversion of 95.5% with the highest epoxy oxygen content of 6.25. As the ethylenic double bond in the AMO to hydrogen peroxide (30%) beyond 1:2.5, the ethylenic double bond conversion declined. This is due to an excess supply of 30% of H2O2 can cause an upsurge in the degradation rate of oxirane oxygen content [25].The effect of the ratio of the ethylenic double bond in the AMO to acetic acid (varied at 1:0.5, 1:1, 1:1.5, 1:2, 1:2.5, 1:3) on the in situ epoxidation is shown in Fig. 5c. Though carboxylic acid such as acetic acid acts as a good oxygen carrier during the AMO epoxidation, it is the main contributor to the degradation of the oxirane oxygen ring formed [1,21]. It was revealed that the epoxide conversion rate improved with an increase in acetic acid concentration, but further addition resulted in oxygen ring degradation. Thus, the molar ratio of the ethylenic double bond in the AMO to acetic acid was found to be 1:1.5. Under this condition, the maximum ethylenic double bond conversion of 95% and the epoxy oxygen content of 6.25 were obtained after 6\u00a0h. Beyond this acetic acid value, the epoxy oxygen content was decreased due to epoxy oxygen ring cleavage owing to the higher acetic acid content.\nFig. 5d shows the effect of epoxidation temperature (varied at 313, 328, 343, and 358 k) on the development of in situ epoxidation, while other epoxidation parameters (viz. molar ratio of the ethylenic double bond in the AMO to 30% H2O2 of 1:2.5, molar ratio of the ethylenic double bond in the AMO to acetic acid of 1:1.5, and catalyst concentration 12.5%) were kept constant. As shown in Fig. 5d, an increase in temperature up to 358 k increased the rate of the ethylenic double bond conversion. However, after 6\u00a0h, the rate of epoxy formation was found to be slightly constant for 343\u00a0K and 358\u00a0K.\nFig. 6\n reveals the reusability of sulfate group-doped tin (IV) oxide solid acid on AMO epoxidation was evaluated for four consecutive epoxidation processes under optimized experimental conditions. After the epoxidation reaction, the catalyst was separated and carefully washed, and then refluxed with ethyl acetate to remove reaction products that had formed on the catalyst's surface. Then it was dried overnight in the oven at 100\u00a0\u00b0C. The in situ epoxidation was carried out at optimal reaction parameters of molar ratio of the ethylenic double bond in oil: H2O2: acetic acid 1:2.5:1.5, employing regenerated sulfate chelated tin (IV) oxide solid acid catalyst 12.5% at 343\u00a0K for 6\u00a0h. The results of the study showed that repeated washing with ethyl acetate leads to the separation difficulty of oil remnants from the pores of the catalyst which poisons catalyst active sites. This causes a gradual loss of catalytic activity after three repeated cycles, and thus lowers ethylenic double bond conversion. It is revealed that oil interaction with catalyst active sites was limited as the number of repeated cycles increased thereby resulting in the lower double bond conversion of AMO.To explore the reaction mechanism of the AMO epoxidation process catalyzed by sulfated tin (IV) oxide solid acid, experimental runs were conducted at 313, 328, 343, and 358\u00a0K to analyze the kinetics and thermodynamics of in situ epoxidation of AMO. The process of epoxidation of AMO is essentially a heterogeneous reaction. To explain the heterogeneous catalytic epoxidation process, the Langmuir \u2013 Hinshelwood \u2013 Hougen \u2013 Watson (LHHW) kinetic expression was suggested. Thus, the reaction that takes place on the catalyst's active sites is principally regulated by three reaction steps: (1) the adsorption of reactants, (2) the surface reaction between adsorbed reactants on the active sites of the catalyst, and (3) desorption of products. However, the in situ epoxidation principally depends on two key reaction steps (i.e. peroxyacetic acid formation and epoxidation steps) since desorption of the product on the catalyst surface is assumed to be weak. In the presence of sulfate group doped tin (IV) oxide solid acid catalyst, epoxidation of AMO with peroxy acetic acid produced in situ from H2O2 and CH3COOH is as in eq. (6):\n\n(6)\n\n\nI\n.\n\n\nH\n2\n\n\nO\n2\n\n\n(\nl\n)\n\n+\nC\n\nH\n3\n\nC\nO\nO\nH\n\n(\nl\n)\n\n\n\n\u21cc\n\nK\n1\n\n\n\nC\n\nH\n3\n\nC\nO\nO\nO\nH\n\n(\nl\n)\n\n+\n\nH\n2\n\nO\n\n(\nl\n)\n\n\n\n\n\nAssuming that the adsorption of AMO and peracetic acid on catalyst active sites is a mild reaction. Then, the surface reaction between AMO and PAA indicates the formation of epoxidized Argemone mexicana oil (EAMO) and is written as in eq. (7):\n\n(7)\n\n\nI\nI\n.\n\nA\nM\nO\n+\nP\nA\nA\n+\nC\n\n\n\u21cc\n\nK\n2\n\n\n\nE\nA\nM\nO\n\u2212\nC\n\n\n\n\nThe desorption of EAMO is given by eq. (8):\n\n(8)\n\n\nI\nI\nI\n.\n\nE\nA\nM\nO\n\u2212\nC\n\n\u21cc\n\nK\n3\n\n\nE\nA\nM\nO\n+\nC\n\n\n\nwhere AMO is Argemone mexicana oil, EAMO is epoxidized argemone mexicana oil, C is sulfated tin (IV) oxide solid acid catalyst, and PAA is peracetic acid.The kinetic analysis for the epoxidation of AMO was conducted on the basis of the subsequent assumptions.\n\n(I)\nThe rate-controlling step (the slowest step) is considered to be the epoxidation AMO (eq. (7)) whereas peracetic acid formation (eq. (6)) is a rapid and simultaneous step which does not considered a rate-controlling step. Thus, the overall rate equation of epoxidation of AMO considering the epoxidation surface reaction on active sites of catalyst is a rate-determining step is shown in eq. (9):\n\n\n\n\n(9)\n\n\n\u2212\n\nr\n\nA\nM\nO\n\n\n=\n\n\n\u2212\nd\n\nC\n\nA\nM\nO\n\n\n\n\nd\nt\n\n\n=\n\nk\n\u2032\n\n\nC\n\nA\nM\nO\n\n\n\nC\n\nP\nA\nA\n\n\n\n\n\n\n\n\n(II)\nSince an excess peracetic acid was used for epoxidation, t can be assumed that peracetic acid content is constant during the epoxidation reaction (i. e. CPAA,o\u00a0=\u00a0CPAA). Moreover, the epoxidation of AMO is assumed to be a pseudo-first-order reaction. Thus, the rate equation is written as in eq. (10):\n\n\n\n\n(10)\n\n\n\n\n\u2212\nd\n\nC\n\nA\nM\nO\n\n\n\n\nd\nt\n\n\n=\nk\n\nC\n\nA\nM\nO\n\n\n\n\n\nwhere, k\u00a0=\u00a0k\u2019 CPAA\nThe rate-controlling step (the slowest step) is considered to be the epoxidation AMO (eq. (7)) whereas peracetic acid formation (eq. (6)) is a rapid and simultaneous step which does not considered a rate-controlling step. Thus, the overall rate equation of epoxidation of AMO considering the epoxidation surface reaction on active sites of catalyst is a rate-determining step is shown in eq. (9):Since an excess peracetic acid was used for epoxidation, t can be assumed that peracetic acid content is constant during the epoxidation reaction (i. e. CPAA,o\u00a0=\u00a0CPAA). Moreover, the epoxidation of AMO is assumed to be a pseudo-first-order reaction. Thus, the rate equation is written as in eq. (10):Thus, eq. (10) in terms of fractional conversion (XAMO) of oil can be rewritten as in eq. (11):\n\n(11)\n\n\n\n\n\u2212\nd\n\nC\n\nA\nM\nO\n,\no\n\n\n\n(\n\n1\n\u2212\n\nX\n\nA\nM\nO\n\n\n\n)\n\n\n\nd\nt\n\n\n=\nk\n\nC\n\nA\nM\nO\n,\no\n\n\n\n(\n\n1\n\u2212\n\nX\n\nA\nM\nO\n\n\n\n)\n\n\n\n\nSince, \n\n\nC\n\nA\nM\nO\n\n\n=\n\nC\n\nA\nM\nO\n\n\n\n(\n\n1\n\u2212\n\nX\n\nA\nM\nO\n\n\n\n)\n\n\n\nWhere, CAMO, o denotes the initial concentration of ethylenic double bond in the AMO.\n\n(12)\n\n\n\n\n\u2212\nd\n\nX\n\nA\nM\nO\n\n\n\n\nd\nt\n\n\n=\nk\n\n(\n\n1\n\u2212\n\nX\n\nA\nM\nO\n\n\n\n)\n\n\n\n\n\nThe rate equation can be stated as follows after integrating eq. (12) at XAMO\u00a0=\u00a00, t\u00a0=\u00a00 and X\u00a0=\u00a0XAMO at t\u00a0=\u00a0t.\n\n(13)\n\n\n\n\u222b\no\n\nX\n\nA\nM\nO\n\n\n\n\n\n\u2212\nd\n\nX\n\nA\nM\nO\n\n\n\n\n1\n\u2212\n\nX\n\nA\nM\nO\n\n\n\n\n=\nk\n\n\u222b\n0\nt\n\nd\nt\n\n\n\n\nThen, after integration of eq. (13), the final rate equation can be written as in eq. (14):\n\n(14)\n\n\n\u2212\nln\n\n(\n\n1\n\u2212\n\nX\n\nA\nM\nO\n\n\n\n)\n\n=\nk\nt\n\n\n\n\nAs a result, the experimental data were fitted with linear regression, and the epoxidation rate constants (K) at various temperatures were determined using the slope of \n\n\u2013\nln\n\n(\n\n1\n\n\u2013\n\nX\n\nA\nM\nO\n\n\n\n)\n\n\n vs. time plot and tabulated in Table 2\n. As shown in Table 2, the rate constant (K) values increased with the corresponding reaction temperature increment, revealing that the reaction was pseudo-first-order with respect to AMO.Arrhenius equation (\n\nk\n=\n\nA\n\nE\na\n/\nR\nT\n\n\n\n) was utilized to compute the activation energy of AMO epoxidation using the slope of \u2013lnk vs. (1/T, k\u22121) plot, and is presented in Fig. 7\n. Therefore, according to the Arrhenius equation plot, the resultant activation energy obtained was 47.03\u00a0kJ/mol. This value confirmed that the chemical reaction utilizing sulfate -doped tin (IV) oxide solid acid catalyst was kinetically controlled.The main thermodynamic parameters analyzed under the present study (viz. Gibb\u2019s free energy (\u0394G), enthalpy (\u0394H), and entropy (\u0394S)) can be determined using eqs. (15)\u2013(15)\u2013(17)(15)\u2013(17):\n\n(15)\n\n\n\u0394\nH\n=\n\nE\na\n\n\u2212\nR\nT\n\n\n\n\n\n\n(16)\n\n\nK\n=\n\n\nR\nT\n\n\nN\nh\n\n\n\ne\n\n\n\u0394\nS\n\nR\n\n\ne\n\n\n\n\n\n\n(17)\n\n\n\u0394\nG\n=\n\u0394\nH\n\u2212\nT\n\u0394\nS\n\n\n\n\nUsing eqs. (15)\u2013(15)\u2013(17)(15)\u2013(17), thermodynamic parameters of epoxidation of AMO using sulfate -doped tin (IV) oxide solid acid catalyst were found to be \u0394H\u00a0=\u00a044.18\u00a0kJ/mol, \u0394S\u00a0=\u00a0\u2212137.91 Jmol\u22121k\u22121, and \u0394G\u00a0=\u00a091.12\u00a0kJ/mol and tabulated in Table 3\n. The \u0394H value determined was the enthalpy of activation for the epoxidation process. The positive values of \u0394H indicate that the energy input (heat) from an external source is required to raise the energy level and transform the reactants to their transition states. Thus, the positive value of enthalpy of activation reveals that the epoxidation process is endothermic in nature. Similar results were reported in various literature [48\u201352]. The negative value of entropy revealed that the epoxide product is more stable as compared to the AMO. The positive value of Gibb\u2019s free energy also showed that the epoxidation of AMO is a non-spontaneous process which is also confirmed by the positive value of enthalpy. The present study is in reasonable agreement with the prior reports, which revealed comparable observations of activation energy of 44.85\u00a0kJ/mol for epoxidation of soybean oil and 44.65\u00a0kJ/mol for palm oleic acid [17,19].\nFig. 8\na and b shows the FTIR results of AMO and its epoxidized oil (EAMO). It was revealed that the removal of ethylenic double bonds in the AMO and the formation of its epoxidized product (EAMO) by their absorption peaks. The bending and stretching vibration of the ethylenic double bond (=C\u2013H of unsaturated fatty acids) in the AMO is visible in bands at 3008\u00a0cm\u22121 and 721\u00a0cm\u22121. However, removing these bands from AMO revealed that the oil has been completely converted to its epoxidized form (EAMO). This was further corroborated by the presence of a new band at 825\u00a0cm\u22121, which was not observed in the AMO, showing that an epoxy oxygen ring (C\u2013O \u2013 C) has been formed in the epoxidized oil (EAMO). This matches the appearance of an epoxy oxygen ring in the 785\u2013880\u00a0cm\u22121 absorption peak range [17]. The absence of a broad proton signal of the \u2013OH group in epoxidized oil indicates that no major side reactions occurred during in situ epoxidations of AMO with sulfated tin (IV) oxide soil acid catalyst.Nuclear magnetic resonance spectroscopy (NMR) was used to better understand the synthesis of epoxidized oil (EAMO). This also signifies the ablation of the ethylenic double bond in the AMO and the appearance of an epoxy oxygen ring in the final epoxy product during in situ technique. In Fig. 9\na 1H NMR spectra shows the presence of the ethylenic double bond (-C = C-) in the AMO at a chemical shift of 5.3\u00a0ppm. However, these ethylenic double bonds in this oil have been vanished in the epoxidized product (EAMO) as shown in Fig. 9b. Furthermore, the existence of new oxirane oxygen ring bands in the epoxy product at 2.7\u20133.3\u00a0ppm and 1.5\u20131.87\u00a0ppm, confirming the conversion of ethylenic double in the AMO to EAMO by in situ epoxidation using sulfate group doped tin (IV) oxide solid acid catalyst.Moreso, the removal of the ethylenic double bond in the AMO at 130\u00a0ppm (Fig. 10\na) and the appearance of a new peak at 53\u201358\u00a0ppm (Fig. 10b) of 13C NMR spectra of EAMO also indicated that the conversion of the ethylenic double bond in the AMO into EAMO.Argemone mexicana oil (AMO) is remarkable a renewable resource for the epoxidation with peroxy acetic acid formed in situ using a heterogeneous solid acid catalyst. Thus, the physicochemical characteristics of AMO were examined as shown in Table 4\n. Moreover, the fatty acids content of AMO were 24.92% oleic acid (C18:1), 59.43% of Linolenic acid (C18:2), and 15.65% saturated fatty acids and contained an iodine value (IV) of 118.21\u00a0g I2 per 100g of oil. AMO epoxide (EAMO) could be used to synthesize valuable goods such as plasticizers, lubricants, polymers, stabilizers, and others [3]. Thus, the physicochemical characteristics of epoxidized AMO were also examined and tabulated in Table 4. The amount of epoxide generated is dependent on the number of double bonds present in the oil, which is defined by the iodine value. The unsaturation of the raw material increased as the iodine value rises. When compared to AMO, the iodine value decreased from 118.21 to 5.62\u00a0g I2/100g of AMO. This shows the conversion of unsaturation present in the AMO into its epoxidized form. The obtained flash point of EAMO was 280\u00a0\u00b0C. Thus, the epoxidized version of AMO can be utilized as a plasticizer in polymeric materials and as a high-temperature diesel fuel additive as a lubricant because it contains ether and ester functionality which enhances its compatibility [3]. The oxidative stability of AMO and EAMO was determined according to A Metrohm AG Rancimat model 892 (Herisau/Switzerland). It was used to assess the oxidative induction time (OIT) in accordance with AOCS Official Method Cd 12b-92, AOCS 1992. Thus, the oxidative induction time of the extracted AMO and its epoxidized oil (EAMO) was found to be 2.13\u00a0h and 68.41\u00a0h, respectively (Table 4 and Supplementary Fig. S1).The viscosity of epoxidized products such as lubricants and plasticizers is critical to their lubricity. The viscosity of fluids decreases as temperature rises, and a measure called the viscosity index was employed to enumerate this trend. The greater the viscosity index value, the less the viscosity of the substance changes with temperature [3]. This work investigated the kinematic viscosities of AMO and its epoxidized oil (EAMO) at a temperature ranging from 20 to 100\u00a0\u00b0C (Fig. 11\n). It was revealed that the viscosity of both AMO and its epoxidized oil (EAMO) decreased with rises in the temperature. The kinematic viscosity of AMO was 31.05\u00a0mm2/s at 40\u00a0\u00b0C and 5.56\u00a0mm2/s at 100\u00a0\u00b0C, and its epoxidized form (EAMO) had a value increased to 131.5\u00a0mm2/s at 40\u00a0\u00b0C and 7.26\u00a0at 100\u00a0\u00b0C (Table 4). The reason behind the increment of kinematic viscosity of EAMO was due to the ethylenic double bond in the AMO was removed through epoxidation. The kinematic viscosities of the present study were within the range of vegetable oil-based epoxy products such as lubricants (5\u2013225\u00a0mm2/s at 40\u00a0\u00b0C and 2\u201320\u00a0mm2/s at 100\u00a0\u00b0C) [53]. The viscosity index of AMO decreased from 163.3 to 136.8 due to the disappearance of the double bond in the AMO. As a result, the viscosity and viscosity index of the EAMO falls within the given ranges, meeting the ISO VG 100 grade viscosity for industrial applications [53].\nFig. 11 depicts the trend of dynamic viscosity as a function of temperature for AMO and epoxidized AMO (EAMO). It was revealed from Fig. 11 that as the temperature increased the dynamic viscosity of both AMO and EAMO decreased. It was also shown in Fig. 11 that the dynamic viscosity of EAMO is much greater than AMO from 20 to 60\u00a0\u00b0C. The reason was due to the conversion of the double bond in the AMO to epoxide product, EAMO during in situ epoxidation.\nTable 5\n [7,18,24,54\u201357], depicts the comparison of literature with other heterogeneous metal oxide catalytic system for epoxidation of vegetable oils. The ethylenic double bond conversion of 89.7%, 75% and 96% were obtained from epoxidation of soybean oil using Ti\u2013SiO2 catalyst at reaction condition of 90\u00a0\u00b0C and 54\u00a0h [18], Alumina catalyst at reaction condition of 80\u00a0\u00b0C and 10\u00a0h [54] and HY zeolite catalyst at reaction condition of 70\u00a0\u00b0C and 3\u00a0h [56], respectively. Similarly, The ethylenic double bond conversion of 90% was obtained from Cardanol oil and Jatropha oil using Amberlite IR 120\u00a0at the catalyst loading of 20\u201322\u00a0wt%, and reaction condition of 65\u00a0\u00b0C, and 7\u00a0h. These results were in comparable to the present study as shown in Table 5.Sulfated tin (IV) oxide solid acid catalyst was successfully synthesized and characterized in this study. Sulfated tin (IV) oxide solid acid was an effective catalyst for the epoxidation of AMO with peroxyacetic acid formed in situ. The maximum ethylenic double bond conversion of 95.5% with an epoxy oxygen content of 6.25 was obtained at the molar ratio of the ethylenic double bond in the AMO: H2O2, acetic acid was 1:2.5, 1:1.5, catalyst concentration 12.5% and reaction temperature at 343 k for 6\u00a0h. Epoxy group formation was confirmed using FT-IR, 1H, and 13C NMR spectroscopy. The physicochemical characteristics of EAMO indicate improved viscosity and oxidative stability, which leads to high lubricity when compared to its precursor, AMO. The catalyst and the AMO epoxide product were potential sources for PVC bioplasticizers synthesis.1) Fekadu Ashine, Subramanian Balakrishnan, Zebene Kiflie, and Belachew Zegale Tizazu conceived and designed the experiments.2) Fekadu Ashine performed the experiments.3) Fekadu Ashine, Subramanian Balakrishnan, Zebene Kiflie, and Belachew Zegale Tizazu analyzed and interpreted the data.4) Fekadu Ashine, Subramanian Balakrishnan, Zebene Kiflie, and Belachew Zegale Tizazu contributed reagents, materials, analysis tools or data.5) Fekadu Ashine, Subramanian Balakrishnan, Zebene Kiflie, and Belachew Zegale Tizazu wrote the paper.No funding was received to assist with the preparation of this manuscript.We do not have any conflict of interest.All authors mutually agreed that the manuscript to be submitted to the Heliyon Journal and the work has not been published/ submitted or is being submitted to another journal.The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.We request you to kindly consider our manuscript for possible publication in your esteemed journal.The authors would like to thank Addis Ababa Science and Technology University and Chemical and Construction Inputs Industry Development Institute to allow experimental set-up work and analytical instruments for characterization.The following is the supplementary data related to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data related to this article can be found at https://doi.org/10.1016/j.heliyon.2023.e12817.", "descript": "\n                  In this study, sulfated tin (IV) oxide solid acid catalyst was prepared for the epoxidation of Argemone mexicana oil (AMO) with peroxyacetic acid formed in-situ. The catalyst was synthesized using the chemical co-precipitation method and characterized. The effects of various epoxidation parameters on ethylenic double bond conversion (%) and oxygen ring content were analyzed. The maximum ethylenic double bond conversion of 95.5% and epoxy oxygen content of 6.25 was found at the molar ratio of AMO to 30% of H2O2\u00a0=\u00a01:2.5, molar ratio of AMO to acetic acid\u00a0=\u00a01:1.5, catalyst concentration\u00a0=\u00a012.5%, and reaction temperature\u00a0=\u00a070\u00a0\u00b0C at reaction time\u00a0=\u00a06\u00a0h. The kinetic and thermodynamic features of the epoxidation of AMO were also analyzed with appropriate models. The results of the kinetic study of the epoxidation reaction followed pseudo first order with the activation energy\u00a0=\u00a00.47.03\u00a0kJ/mol. Moreover, the thermodynamic constants of epoxidation of AMO were found as \u0394H\u00a0=\u00a044.18\u00a0kJ/mol, \u0394S\u00a0=\u00a0\u2212137.91 Jmol\u22121k\u22121) and \u0394G\u00a0=\u00a091.12\u00a0kJ/mol. The epoxidized product of AMO was further analyzed using FTIR, 1H NMR, and 13C NMR. The results of these analyses confirmed the successful conversion of the ethylenic double bond in the AMO to EAMO.\n               "}
{"full_text": "With the rapid development of the global economy, industrialization and urbanization have caused the demand for fossil fuels to expand rapidly. Environmental warming and its associated environmental problems have caused widespread concern around the world; greenhouse gases in the atmosphere are the main cause of global warming, and carbon dioxide is the greenhouses gas with the highest emissions [1]. Carbon dioxide is mainly derived from the burning of fossil fuel, automobile exhaust, animal and plant respiration, and corpse decay. Carbon dioxide emissions are high, but its utilization rate is low. At present, it seems to be meaningful to convert carbon dioxide into different chemicals [2]. Therefore, new catalysts need to be developed, and in-depth research on methods to functionalize carbon dioxide must be performed. To solve this problem, an increasing number of scientists have become committed to developing various technologies to capture and fix carbon dioxide [3].Metal-organic frameworks (MOFs) are a class of porous crystal materials formed by the self-assembly of metal ions or metal clusters and organic ligands and are promising materials with unique properties. MOFs have a large specific surface area, porous structure, and multichemical composition and are easy to be functionalized. MOFs have been applied to gas adsorption, storage and separation; sensors; electrochemistry; and catalysis.Combined with the environmental problems caused by carbon dioxide, at present, many works have reported that metal organic frameworks act as catalysts to fix carbon dioxide [4\u20137]. Wang W et\u00a0al., synthesized a metal-organic framework containing zinc metal and the H4tmpe ligand with a stable structure to fix carbon dioxide [4]. Beyzavi MH et\u00a0al., reported a new Hf-based metal\u2013organic framework (HfNU-1000) incorporating Hf6 clusters [5], which demonstrated a high catalytic efficiency for the activation of epoxides and facilitated the quantitative chemical fixation of CO2 into five-membered cyclic carbonates under ambient conditions, rendering this material an excellent catalyst. Wang S et\u00a0al., reported a cobalt-containing zeolitic imidazolate framework (Co-ZIF-9) that served as a robust MOF cocatalyst to reduce CO2 by cooperating with a ruthenium-based photosensitizer [6]. Li P et\u00a0al., successfully constructed a highly porous MOFs via solvothermal assembly of a clicked oct carboxylate ligand and Cu (II) ions, incorporating both exposed metal sites and nitrogen-rich triazole groups [7]. The high-efficiency and size-dependent selectivity toward small epoxides for catalytic CO2 cycloaddition make this MOF a promising heterogeneous catalyst for carbon fixation.The materials discovery process involves several stages, including synthesis, testing and characterization, etc [8]. These steps are usually carried out sequentially, and therefore, only a few materials can be synthesized, tested or measured at a time. The process of discovering and developing new materials currently entails considerable effort, time and expense. For MOFs, the possible combinations of the numerous building blocks under different topological symmetries are almost infinite [9]. Therefore, the construction of metal-organic frameworks is diverse, and people can only construct these frameworks via experience and guesswork. To explore the high-performance MOF materials, a theoretical prediction method with high efficiency and accuracy should be developed and applied.Machine learning technology has achieved great success in finance, medical, biology, and informatics [10\u201312]. Today, with the development of artificial intelligence, machine learning (ML) is an emerging research paradigm that will revolutionize material discovery [13]. It has been indicated that machine learning can be applied to chemical discovery [14\u201316]. Machine learning, as the name implies, uses the most primitive learning method of humans (regular learning) to give a machine the ability to process data, train the algorithm using training data, and test the accuracy of the algorithm using test data. ML makes effective decisions through the experience generated by historical data [17]. This material discovery method deviates from the traditional method and can screen large-scale materials with excellent performance at one time. ML is more efficient and faster than traditional methods.In our work, based on the reported experimental results, the characteristics and performance of MOFs for fixing carbon dioxide into cyclic carbonate were extracted to establish a data set, which was further applied to train and test five ML algorithms, including SVM, KNN, DT, SGD, and NN, to setup classifiers. The tested ML algorithms were extended to classify 1311 hypothetical MOFs via screening for high catalytic performance; the characteristics of these MOFs were finally extracted, as shown in Scheme 1\n [18\u201320]. This work applies machine learning methods to small data sets to achieve large-scale screening of catalysts in the hope that these methods can play a guiding role in future experimental work, thereby accelerating the discovery of new materials and saving manpower, material resources, and financial resources.The metal ions/clusters and organic ligands in a MOF determine its structure and properties. Here, the topological structure of a MOF is completely abandoned, and the types of metal atoms and organic ligands are selected as the features of the MOF. The MOF\u2019s characteristics and carbon dioxide conversion rates were collected from approximately one hundred published paper. Data set information is available in the Supplementary Information B. The machine learning algorithm learns a classifier to predict the catalytic properties of MOFs based on their structural features. TOF value is used as an indicator for catalyst performance evaluation, and the calculation method of TOF value is described in Supplementary Information A. The reported reaction temperatures in experiments range from room temperature to 140\u00a0\u00b0C, which could be directly used in ML. Here, the TOF values have been transformed to the same temperature according to Arrhenius equation, and the specific implementation method is presented in Supplementary Information A.S1. The median of revised TOF value is taken as the classification limit. The target value (revised TOF value) is divided into two categories, a value above the median which classified a good category, marked as 1, and a value below the median is classified a bad category, marked as 0. After processing, the target value distribution of the machine learning data set samples is average.To predict the carbon dioxide conversion rate of MOFs based on their structural properties, supervised learning was applied. In supervised learning, the computer learns from labelled historical examples for which the outcomes are known to make predictions on future data for which the outcomes are unknown. Based on python and the scikit-learn package[21]\n], we used five machine learning methods, including SVM, KNN, DT, SGD method and NN, and more information is available in the Supporting Information A.S2. The specific workflow is shown in Fig.\u00a01\n.The performance of each ML algorithm was evaluated by calculating the precision, recall and F1 score. As shown in Scheme 2\n, the precision was calculated with \n\n\n\nT\nP\n\n\n(\nT\nP\n+\nF\nP\n)\n\n\n\n, where TP is the number of true positives and FP is the number of false positives. Precision measures how many of the samples that were predicted to be positive are true positive examples.Recall is\n\n\n\nT\nP\n\n\n(\nT\nP\n+\nF\nN\n)\n\n\n\n , where TP is the number of true positives and FN is the number of false negatives. Recall is the ability of the classifier to find all positive samples. Recall measures how many of the positive samples are predicted to be positive.The F1 score, also known as the balanced F-score or F-measure, can be interpreted as a weighted average of precision and recall, and it can be calculated with the following equation:\n\n(1)\n\nF\n1\n\ns\nc\no\nr\ne\n=\n\n\n2\n\u2217\n\n(\nprecision\u00a0\u2217\u00a0recall\n)\n\n\n\n(\nprecision\u00a0\n+\n\u00a0recall\n)\n\n\n\n\n\nthe F1 score reaches its best value at 1 and worst value at 0, and the relative contributions of precision and recall to the F1 score are equal.The data set was derived from the experimental literature in which MOFs are used as catalysts to immobilize carbon dioxide as a cyclic carbonate [1,5,7,22\u201381]. Useful information was extracted from the literature, such as the metals, organic linkers, MOFs, reactants in the CO2 fixation reaction, and TOF values. There are five reactants reported in the experimental results, including propylene oxide (PO), epichlorohydrin (ECH), butylene oxide (BO), styrene oxide (SO) and epibromohydrin (EBP), which have similar structures, as shown in Fig.\u00a02\n. The database in this study contains 106 data entries (the total MOF number). For each MOF, the number of metal species, organic ligands and reactant types are 2, 2 and 1, respectively, and every MOF was described with 85 structural characteristics composed of 23 metals, 57 organic ligands, and 5 reactants. Each structural property can be encoded as a binary parameter (also called a variable or feature), 0 or 1, indicating its presence or absence, respectively, in a specific MOF [9]. TOF values is taken as the target. The data set information is available in the Supplementary Information B.The machine learning algorithm learns a classifier (also known as a \u201cmodel\u201d) to predict the catalytic properties of MOFs based on their structural features. The prediction result of the classifier is 0 or 1, which represents materials with a poor prediction performance or excellent performance, respectively. It is hoped that the classifier model can be generalized to predict the approximate performance of a novel species and the values of its structural parameters. If the predicted performance of the new material does not reach the desired level, then there is no point in synthesizing the material, and vice versa. Therefore, an accurate prediction model can guide the synthesis and experimentation of new materials.Before training the models, the data set was randomly divided into 80% for training and 20% for testing. Five machine learning methods, including SVM, KNN, DT, SGD, and NN, were trained by adjusting the hyperparameters. The final models selected were built with the following configuration. The SVM classification with the libsvm implementation method from scikit-learn was used (svm.SVC). The learning of the hyperplane in the SVM algorithm used radial basis function (RBF) kernel functions for the decision function. Our implementation of SVC finds the best parameters; penalty parameter C of the error term is 100, and the gamma value of the kernel coefficient of RBF is 0.01. The K Neighbors Classifier class, with the number of neighbors being 23 and the algorithm parameter being auto, attempts to decide the most appropriate algorithm based on the values passed to the fit method. The Decision Tree Classifier class with a maximum depth of tree of 8 and balanced classes weights was used to build the model. The SGD classifier algorithm possesses hyper-parameter values for loss of hinge and penalty (aka, the regularization term) of l2. The NN model implements a multilayer perceptron (MLP) algorithm that trains using backpropagation. The numbers of hidden nodes in the two hidden layers were set equal to 5 and 5. After training the machine learning models, the next step is to evaluate the models.Five machine learning methods were used to train the model and then test the test set. The model performance for predicting catalysis was evaluated through calculation of the accuracy, precision, recall, and F1 score. The accuracy scores of the five models on the training set and the test set are shown in Fig.\u00a03\n(a). SVC, NN and SGD have a strong learning ability for the training data set, and KNN and DT perform are worse than the former three methods. Regarding testing capability, SGD has a strong testing ability with an accuracy of 86.4%, that of SVC and NN are both 81.8%, and KNN and DT is lower than 80%. For SVC, SGD and NN, the prediction results are consistent with the experimental results at least 80% of the tested materials.In some cases, accuracy is not the most comprehensive tool for evaluating models. Indicators such as precision and recall are better for measuring machine learning model performance than accuracy under certain circumstances. Next, to further evaluate the model, precision, recall and the F1 score were used to evaluate the classification performance, and the results are shown in Fig.\u00a03(b). By comparison, we find that SVC, SGD, NN, models have the highest precision of more than 92%, which means that among the catalysts with an excellent predicted performance, over 92% of the MOFs are truly excellent. SVC, SGD, NN models all have the recall of 82%, and among them, SGD has the highest recall rate of 0.864, which means that 86.4% of the excellent MOFs verified by experiments are also predicted with SGD.The F1 score is the harmonic average of precision and recall, and commonly it\u2019s ideal when some model\u2019s F1 score is higher than 80%. Through the analysis of the prediction and testing results, the F1 score of SVC, SGD, NN methods can reach more than 80%, and KNN and DT are lower than that value. Through the reliability analysis for the five classifiers, SVM, SGD, and NN are all get the high cores in accuracy, precession, recall and F1 evaluation. Here, to improve the reliability of the classification, three trained classifiers from SVM, SGD, and NN were combined to improve the reliability, which could be considered as an ensemble learning method.We combined 1311 hypothetical MOFs using 23 metals and 57 organic ligands, which is the component of the reported MOF for fixing CO2, and further applied the trained classifiers to screen out MOFs with high catalytic activity. The five organic reactants in the previous experimental work were used to predict the performance of metal organic framework materials. Finally, 6555 (1311\u22175) samples were available for classification.The classified results for the CO2 fixation performance are presented, which contains five layers, and each layer corresponds to one organic reactant and contains 1311 points, which correspond to 1311 MOFs. If a MOF\u2019s performance in CO2 fixation is predicted as \u201cgood\u201d by one of the three models at the same time, the MOF is regarded as a highly active material and is specified as a red point. For the other prediction results (none \u201cgood\u201d from the three methods), the MOF is regarded as a bad material and is marked as a blue point. Prediction results can be found in the Supplementary Information A.S3.All of the five organic reactants have an epoxyethane group, and their structures are similar to each other. Here, the versatility of a MOF\u2019s catalytic performance among the organic reactants is investigated using the voting method and is presented in Fig.\u00a04\n. If a MOF displays \u201cgood\u201d activity on more than three reactants, the MOF\u2019s performance is specified as \u201cexcellent\u201d and is marked as a red point. If a MOF displays \u201cbad\u201d activity on all five of the reactants, it is specified as \u201cpoor\u201d and is marked as a blue point. If a MOF\u2019s activity is only \u201cgood\u201d for 1\u20133 reactants, it is specified as \u201cmoderate\u201d and is marked as a green point. In Fig.\u00a04, the abbreviated form of the ligand is used. The full name of the ligand is in the Supplementary Information A.S3.The predicted \u201cexcellent\u201d MOFs are what we ultimately need. As shown in Fig.\u00a05\n, certain metals could be combined with most organic ligands to form high-performance MOFs, and some organic ligands also combine with most metals to form excellent MOFs. Here, the excellent ratios are evaluated for the metal and ligand, which are calculated with the following formula:\n\n(2)\n\nExcellent\u00a0Ratio\n=\n\n\nExcellent\u00a0ligands\u00a0\n\n(\nmetals\u00a0\n)\n\n+\nModerate\u00a0ligands\n\n(\nmetals\n)\n\n\u2217\n50\n%\n\n\nTotal\u00a0ligands\u00a0\n\n(\nmetals\n)\n\n\n\n\n\n\n\nThe excellent ratios for metals are presented in Fig.\u00a05. 11 metals, Y, Zr, Ni, Cu, Li, Na, K, Rb, W, V, and Mn have shown slightly better performance than other metals. In the above metals, Li, Na, K, Rb and W, are all derived from the multi-metal MOFs predicted through the machine learning, and the present results could not ensure their single-metal MOFs have good catalytic performances. Here Mn, V, Cu, Ni, Zr, and Y are the most likely candidates.As shown in Fig.\u00a06\n, MA, BDC-NHx(Me)3-x(I-), NH2-BPY, NH2-BDC, bpH2, tactmb, tdcbpp, TBPP, TCPP, DABCO, TATAB, AIP, BTC, H3L, TCPE, NDC,BTB, compared with other ligands, display outstanding performance. In the ligands, MA, BDC-NHx(Me)3-x(I-), NH2-BPY, NH2-BDC, bpH2, DABCO, TATAB, AIP, BTC, BTB, NDC, are in the dual-ligand MOFs, and the present results also could not ensure their performance when they applied as ligand singly. TBPP and TCPE in the training data set are only one data, this may lead to the overestimated activity of them through the ML. Then for single ligands, tactmb, tdcbpp, TCPP, H3L are recommended as the most ideal candidates, and their structures are shown in Fig.\u00a07\n.The predicted six metals and four organic ligands could be combined to form 24 high-performance MOFs, as shown in Fig.\u00a08\n. Six MOFs (specified as stars) have been synthesized and reported [33,34,82\u201385], and their crystal structures are shown in Fig.\u00a08.In the six excellent metals, V, Mn, Cu and Ni, are 3d metal, and Y and Zr are the first two 4d metals. All of them have multiple valence, similar atomic radius and electronegativity. Due to that the metal ion is always the center of the catalytic activity, the 3d and the beginning of the 4d metal could be considered.For the MOFs reported as catalysts that can fix carbon dioxide into cyclic carbonate in previous experimental works, their structural characteristics and catalytic activities are collected into a data set, which is applied to train classifiers with five ML algorithms, and three classifiers are combined to predicted 1311 novel MOF structures via ensemble learning. The results show that the ML model could predict a MOF\u2019s catalytic performance according to its structural feature. The best metals, Mn, V, Cu, Ni, Zr, and Y, and best ligands, tactmb, tdcbpp, TCPP, H3L, are discovered. The six metals and four ligands could be combined into 24 MOFs that possess strong potential for being catalysts for carbon dioxide fixation.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This study was supported by the National Natural Science Foundation of China (21676004). Authors declare that there are no conflicts of interests.The following is the supplementary data related to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data related to this article can be found at https://doi.org/10.1016/j.jmat.2021.02.005.", "descript": "\n                  The process of discovering and developing new materials currently requires considerable effort, time, and expense. Machine learning (ML) algorithms can potentially provide quick and accurate methods for screening new materials. In the present work, the features of the metal organic frameworks (MOFs) as a catalyst for fixing carbon dioxide into cyclic carbonate were extracted to build a data set, which were collected from the experimental results of approximately 100 published papers. Classifiers were trained with the data set with various ML algorithms, including support vector machine (SVM), K-nearest neighbor classification (KNN), decision trees (DT), stochastic gradient descent (SGD), and neural networks (NN), to predict the catalytic performance. The ML models were trained on 80% of the data set and then tested on the remaining 20% to predict the carbon dioxide fixation ability. The trained ML model was extended to explore 1311 hypothetical MOFs, and some structures displayed a strong catalytic ability. Finally, the six best metal ions (Mn, V, Cu, Ni, Zr and Y) and four best ligands (tactmb, tdcbpp, TCPP, H3L) were determined. These six metals and four ligands could be combined into 24 MOFs, which are strongly potential catalysts for carbon dioxide fixation. Using machine learning methods can speed up the screening of materials, and this methodology is promising for application not only to MOFs as catalysts but also in many other materials science projects.\n               "}
{"full_text": "\n\n\nNo data was used for the research described in the article.\n\n\nNo data was used for the research described in the article.As a result of increasing public attention to the major environmental risks posed by the widespread use of fossil fuel sources, a deal was reached to transition to an energy system that relies on cleanliness, safety, and recycled materials [1\u20132]. Hydrogen energy is an attractive future energy carrier since it can be created easily and without producing any greenhouse effect [3\u20136]. Despite being the most prevalent element in the universe's elements, in nature, hydrogen isn't found in its purest form. In this context, the production of low-cost, environmentally friendly hydrogen from renewable energy sources is becoming increasingly important. Water is a valuable source of hydrogen, and water electrolysis is the most common method of obtaining hydrogen from it [7\u20138].The HER has already been examined in both acidic [9\u201311] and alkaline [12\u201313] solutions, as well as its dependence on pH and temperature effects. Several electrode materials have been used to investigate HER, including Hg, Rh, Pt, Au, Sn, Cu and Ag [14\u201317]. It is widely available that increasing any material's surface area can be calculated in general to improve its cathodic efficiency through a synergistic combination of multiple parts engaged in it [18\u201319]. Moreover, the electrocatalytic productivity of any substance is mostly determined by the metal surface's binding energy with adsorbed hydrogen [20\u201321]. Based on the preceding, catalysts based on noble metals have already been used as effective HER electrocatalysts [22\u201323].The first electrodeposition of rhodium plating was demonstrated by Marino in 1912 [24]. According to Cinamon, sulphate and phosphate baths are used for rhodium coating [25]. Rh is very corrosion resistant due to its association with the platinum group of metals. In our research work, we achieved coating using PC and DC techniques. The main reason for selecting PC over traditional DC plating is to avoid the continuous entry of current into the bath, which leads to burning deposits or uneven coating. This led to the PC becoming popular in recent times. Pulse coating also leads to fine grain size deposition by varying the current TON and TOFF formulas. For more than a decade, PC coatings have been used in industries like aerospace, shipyards, and auto manufacturing [26\u201327]. Catalytic metals, especially Cu, Zn, and Al mixed oxides, are frequently used in chemical engineering and pollution control [28\u201329]. The presence of catalytically active transition metal species (e.g., Cu, Co, Fe, Ni, V, Rh) allows for easy separation of the end products [30\u201331].In this paper, Rh has been coated on SS304 by electrodeposition. Pt is always superior. We tried with Rh, to know its ability and get an idea of Rh's comparison to Pt [22]. The prepared Rh specimens were characterised by SEM, EDX, AFM and XRD. HER performance was accessed by LSV, Tafel and Chronoamperometry techniques.The rhodium bath was optimised to coat the rhodium on an SS304 substrate. Even though rhodium is cost-effective, the availability of Rh on the steel surface is very low, and hence the cost of Rh in the study can be controlled. The amount of Rh deposited on the substrate is 0.5 \u00b5gcm\u22122. The bath's composition and working requirements are presented in Table\u00a01\n. The rhodium sulphate solution (Rh2(SO4)3) was supplied by Arora Matthey Limited, Kolkata. Using a pH metre (HI2020 edge pH meter, HANNA, USA), the bath pH was kept at 1.3 by adding H2SO4 as needed. All depositions were performed at a temperature of 45\u00b0C. Electrodeposition was performed on a specified surface (1.76 cm2) of the polished (abrasive sandpaper of many grades ranging from 80 to 1800 scale) SS304 substrate (15 mm in diameter and 1 mm in thickness). The anode was made of insoluble platinised titanium (Ti anode fabricators private limited, Chennai, India) [32]. During plating, the electrodes, anode and cathode, were placed at a distance of 3 cm apart. The experiment was carried out in a 100 mL capacity beaker (rhodium solution and H2SO4) designed electrochemical cell and a power source (Agilent N6705A DC Power Analyser, USA). All depositions were recorded and completed in 20 minutes under continuous conditions. The coatings were rinsed and dried with distilled water. Various instrumental tools were used to examine the deposited coatings for surface morphology and compositional information.The surface morphology of the rhodium coating was analysed using SEM ((model: FESEM Carl ZEISS), interfaced with EDS (model: Oxford Nanoanalysis 250). The surface morphologies of coated materials have been described using atomic force microscopy investigations (model: Nanosurf\u00ae EasyScan 2 AFM & STM) to confirm the evidence of other research methods. X-ray diffraction (Riakgu Mini Flexell Desktop Diffractometer with Cu-Ka (l\u00a0=\u00a01.5406 \u00c5).) at 40 kV and 40 mA, scanning from 10o\u00a0to 100o\u00a0of 2\u03b8 was used to identify the phases and crystal structure of the coated samples.In the cell, electrodeposited rhodium electrodes were exposed to cathodic and anodic polarisation to determine the amount of hydrogen produced during the study. The rhodium coatings were electrodeposited on the working electrode and a platinum electrode as a counter electrode. As a reference electrode, Ag/AgCl was used. Sinsil International Private Limited in Bengaluru, India, supplied all of the electrodes. Using the CompactStat.h10800 workstation, Ivium Technologies, The Netherlands, the electrochemical behaviour of the coatings was characterised using CompactStat.h10800. The glass setup is equipped with marked micro-burettes for measuring the quantity of H2 released throughout electrolysis. The distance between the electrodes in our setup is 5 cm, near the cathode and anode exit holes are provided. It is easier for hydrogen and oxygen to go out from the chamber rather than mixing and was achieved by utilising a modified glass tubular cell, as illustrated in Fig.\u00a01\n.As illustrated in Fig.2\n, SEM is used to study the surface morphology of both PC (with different duty cycles) and DC source coatings, as illustrated in Fig.\u00a02. The optimised current density in our work was 4.1 A/dm2. One of the key factors that determines the rate of HER is current density. In comparison to all the PCs and DC sources (Fig.\u00a02 a), PC 75% samples reveal more homogeneous and smaller granules, which produce a smoother coating surface. The Rh was irregular in scale on the surface in Fig.\u00a02 b, 2c, and 2d because there are variations in its spread, the AFM confirmed this impression (Fig.4). The degree of uniformity decreased from PC 75% to DC. The duty cycle percentage has a considerable impact on the morphology of the surface of Rh coatings. Strong adherence and brightness were obtained for all Rh alloy coating sources, although a 75% duty cycle rendered one appropriate for HER activity [33].\nFig.\u00a02 depicts the EDX spectrum of Rh metal ions incorporated into the Rh bath solution coating. Furthermore, the weight % of Rh is presented in Table\u00a02\n and demonstrates that the Rh concentration in PC coatings is less than in DC coatings, despite the same deposition conditions. The experimental data from Table\u00a02 indicates that the Rh content of the coating declines in the bath from DC to PC coating, and the data matches concerning SEM images. (Fig.2). Type SS304 is a grade of austenitic steel with the following chemical composition by weight percentage; C 0.08, Mn 2.00, P 0.042, S 0.032, Si 0.72, Cr 18-20, Ni 8-12, Ni 0.10 and Fe 67-71.This change in the weight percentage of the coating causes a change in the surface roughness, as a result of this, it enhances the electrocatalytic activity [34].The AFM is a strong tool for characterising coating roughness in terms of average smoothness, which is answerable for improved electrocatalytic action. As a result, as shown in Figs.\u00a04(a) and 3\n\n(b), a 3D AFM image of DC duty cycle and PC 75% coatings are captured. In comparison to the PC 75% source, there were considerable alterations of the surface roughness in the DC coating. The excess surface area of the active rhodium on SS304 is caused by spines and corrugations on the DC electrode. The average roughness of the PC technique was 15.9 nm, which is lower than the DC coating of 42.0 nm, indicating that the Rh is deposited relatively consistently and even. This level of uniformity in the PC method increased the electronic charge density during LSV, which contributed to the high HER activity.\nFig.\u00a05\n depicts XRD patterns for Rh deposited on SS304 at different coating sources in our study. Peaks may be seen at 43.2o, 50.9o and 75.3o respectively. The Rh (111), (200) and (220) planes of the cubic Rh crystal can be indexed by three diffraction peaks Crystallographic search match software and powder diffraction files were used to analyse the peaks of the XRD pattern (PDF no. 1-1213). The grain sizes can be determined using Debye\u2013Scherrer Equation [30] as given below\n\n\n\nD\n=\n\n\nK\n\u03bb\n\n\n\u03b2\ncos\n\u03b8\n\n\n\n\n\n\nWhere \u03bb is the X-ray wavelength, \u03b8 is the Bragg angle, and \u03b2 is the FWHM of the diffraction peak. The average grain size of the coatings is 7nm, 10nm, 12nm, and 14nm for DC, PCs, 25%, 50%, and 75% duty cycles respectively.As illustrated in Fig.\u00a06\n, the Rh electrocatalytic activity was initially investigated for H2 evolution in H2SO4 media. All of the samples exhibit a favourable hydrogen evolution process. At duty cycle 75%, Rh coated by the PC technique has a relatively low overpotential for hydrogen evolution. For HER, the overpotential of a Rh catalyst deposited by a 75% duty cycle sample is comparable to that of pure Pt. These findings support the sample's superior performance (75% duty cycle). During the hydrogen evolution reaction, the 25% and 50% samples show more overpotential and less current. It was predicted that Rh would be particularly active for HER. Indeed, the catalyst Rh demonstrated a catalytic start at virtually zero overpotential, and catalytic current rapidly increased in the sample attained at 75% duty cycle. Further cathodic sweeping revealed H2 bubble development and discharge from the surface are both very active.Tafel plots, as illustrated in Fig.\u00a07\n, were used to assess the Rh electrocatalytic activities. Tafel slopes reveal the nature of the HER process. The Volmer Heyrovsky or Volmer-Tafel mechanistic pathways are used to express the kinetics factor of electrocatalytic HER. Table\u00a03\n shows the Tafel slopes of 40.7, 61.3, and 74.1 mV/dec for coated Rh deposition with PC, respectively utilising 75 %, 50 % and 25 % duty cycles. The 75 % duty cycle coated Rh catalysts had a much lower Tafel slope, indicating increased electrocatalytic activity, which is indeed more than the prepared DC samples (69.9 mV/dec). The moderate Volmer-Tafel reaction mechanism has a Tafel slope of 40.7 mV/dec is hydrogen atom desorption and that hydrogen atom desorption is the rate-determining step [35\u201336].The chronopotentiometry (CP) of both PC (75%, 50% and 25% respectively) and DC coatings was investigated, as well as their electrocatalytic stability. A persistent current is administered in-between the two electrodes in this method by monitoring the voltage of one of the electrodes as a function of time regarding the substance of the reference electrode. The CP experiment was carried out at a steady current of -0.35 mAcm\u22122 for 3 hrs. The electrocatalytic behaviour of the coatings was evaluated using this technique by monitoring the sum of hydrogen freed for an early 180 seconds. The amount of hydrogen liberated is recorded and registered in Fig.\u00a09. When compared to DC coatings, PC coatings emit a greater amount of H2 gas. This demonstrated that the preferred electrode material for HER is PC coating. Fig.\u00a08\n\n depicts the chronopotentiograms of PC 75%, PC 50%, PC 25%, and DC coatings. At first, the graph shows a substantial drop in potential as a function of time for both coatings. This is because, at the start of the electrolysis, the reduction of hydrogen ions and the evolution of hydrogen gas occur at a faster pace due to the rapid supply of current [37\u201338]. After a few minutes, there was little fluctuation in the potential with time, indicating the development of equilibrium. This shows that the release of hydrogen occurs efficiently on the electrode's surface and measurement of hydrogen release is confirmed by the fitted graduated burette in the three electrode arrangements glass tube as shown in the Fig.\u00a01.The coating of rhodium on SS304 by pulse and direct technique was successfully done by the PC and DC method. The coating grain size was reduced and it was achieved by the PC technique at PC 75% is inferred by SEM. EDX confirmed the presence of Rh in the base metal. The roughness of the surface is highlighted by AFM analysis, which is supported by surface morphology outcomes. The SEM results were XRD verified. LSV demonstrates that the created coating has less overpotential and provides greater current. The lesser Tafel slopes demonstrate the efficacy of the catalysts and validate the HER mechanism explained by the Volmer-Tafel. Values of chronopotentiometry approve the complete consequences by providing a higher hydrogen collecting volume through electrolysis. The present research has the calibre to deliver its importance and to be commercialised for industrial use (Fig.\u00a03).The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Dr. Praveen B.M reports financial support was provided by Board of Research in Nuclear Sciences. Dr Praveen B.M reports a relationship with BRNS that includes: employment and funding grants. Dr. Praveen BM has patent pending to Licensee. Dr. Praveen B.M employee of Srinivas University, College of Engineering and Technology, MangaluruWith the approval of Project No. 37(2)/14/18/2018-BRNS, dated 11/07/2018, instrumentation and financial funding from the Department of Atomic Energy (DAE), Board of Research in Nuclear Sciences (BRNS), Mumbai, Government of India and the authors also wish to acknowledge the management of M.S. Ramaiah College of Arts, Science and Commerce, Bengaluru for the constant support and encouragement through MSRCASC seed money funding granted in the year 2022. Srinivas University, Institute of Engineering & Technology, Srinivas Nagar, Mangaluru, Karnataka, has provided laboratory support", "descript": "\n                  The theory and kinetics of the hydrogen evolution reaction (HER) on electrodeposited rhodium in acidic media (0.5 M H2SO4 solution) were looked into. An electrodeposition approach using direct current (DC) and pulse current (PC) was used to deposit rhodium on a stainless steel 304 (SS304) substrate. Several parameters, including rhodium concentrations, current densities, temperature, pH, and coating duration, were used to optimise the rhodium bath. Scanning electron microscopy (SEM), X-ray diffraction (XRD), and energy-dispersive X-ray (EDX) analyses were used to assess the change in surface shape and chemical composition. The best coating was demonstrated at PC 75% duty cycle with an optimised current density of 4.0 A/dm2, which was better than the remaining PC cycles and DC source coating, indicating the most productive activity for hydrogen production. The activity of Rh catalyst coatings resembled that of pure platinum metal. Cyclic voltammetry (CV), chronopotentiometry (CP), and potentiodynamic polarisation techniques were studied to determine the HER. The results obtained from the PC technique with a 75% duty cycle give more HER performance.\n               "}
{"full_text": "Amine compounds are indispensable intermediates in fine and bulk chemical industries for the synthesis of polymers, surfactants, pharmaceuticals, and agrochemicals [1\u20133]. Currently, a large quantity of industrial relevant amines, such as aliphatic amines, aromatic amines, and aminoalcohols, are manufactured from non-renewable fossil resources via several functionalization steps, which suffer from low energy efficiency and environmental pollution [2]. Catalytic reductive amination (RA) of carbonyl compounds is a well-known class of reaction that is widely used for the clean synthesis of various amines with water as the main by-product [4,5]. Due to the increasing concerns on depletion of fossil recourses and rising CO2 concentration in the atmosphere, the synthesis of valuable amines from biomass-derived compounds has been gaining increasing attention [1,2,6\u20138].The RA of biomass-derived carbonyl substances including aldehydes and ketones proposes a potential alternative for the high-efficiency synthesis of renewable amines in mild conditions of temperatures \u2264 120\u00a0\u00b0C and pressure \u2264 4\u00a0MPa [2,9]. For instance, Zhang et\u00a0al. [7] reported a novel bifunctional Ru/ZrO2 catalyst with co-existence of Ru0 and acidic RuO2 species, realized the efficient conversion of lignocellulose-derived glycolaldehyde into useful ethanolamine at 75\u00a0\u00b0C and 2\u00a0MPa H2. Hara and coworkers [10] reported that Nb2O5 supported Ru catalysts could effectively catalyze RA of various carbonyl compounds at 90\u00a0\u00b0C. Using activated carbon supported Pd nanoparticles, Iborra et al. [11] attained N-substituted-5-(hydroxymethyl)-2-furfuryl amines (yield up to 100%) in the RA of 5-hydroxymethylfurfural at 100\u00a0\u00b0C and 0.3\u00a0MPa H2. Besides noble metals, base metals, such as nickel (Ni/Al2O3 [5], Ni@SiO2 [12], Ni/MC [13]), cobalt (MOF-Co [14], Co/N\u2013C [15]) and copper (Cu/ZrO2 [16]), also presented high performance for the RA of different carbonyl compounds. Despite these achievements, it is still of significant importance to obtain efficient and stable base metal catalysts for the green production of valuable amines from biomass and its derived carbonyl substances via RA reaction under mild conditions.5-Amino-1-pentanol (5-AP) is a valuable amino alcohol that is extensively required as a pharmaceutical intermediate for the manufacture of therapy for cancer and inflammation, especially the alkaloid manzamine with high medicinal value [17]. Recently, our group reported a novel and environmentally friendly process to sustainably synthesize 5-AP by the RA of bio-furfural derived 2-hydroxytetrahydropyran (2-HTHP) with ammonia [18\u201320] (Scheme 1\n). The in situ ring cleavage tautomerization of 2-HTHP to reactive 5-hydroxypentanal (5-HP) contributes to the high efficiency of the clean synthesis of 5-AP. Note that based on the relatively higher hydrogenation activity (\u223c80 times) of the 2-HTHP intermediate as compared with the direct hydrogenolysis of tetrahydrofurfuryl alcohol. Huber and coworkers [21] developed a multi-step process for the synthesis of useful 1,5-pentanediol (1,5-PD) with high tech-economy from bio-fufural derived tetrahydrofurfuryl alcohol. Traditional metal oxides especially for ZrO2 supported Ni catalysts [18] and the hydrotalcite-derived Ni\u2013Mg3AlO\nx\n catalysts [19] exhibited high activity with 90.8%\u201393% 5-AP yield in the RA of 2-HTHP at 80\u00a0\u00b0C and 2\u00a0MPa H2. Nevertheless, the stability of these monometallic Ni catalysts was still unsatisfactory due to the sintering and surface oxidation of activity metal particles, even though the stability of Ni\u2013Mg3AlO\nx\n catalyst has been improved a lot (losing 18% activity after 120\u00a0h time running) by virtue of the unique layered double hydroxides (LDHs) structure [22,23]. Additionally, despite Ru-based catalysts being much cheaper than Rh- or Ir-based catalysts for the synthesis of 1,5-PD, their stability in the direct hydrogenation of 2-HTHP to 1,5-PD was still unsatisfactory, especially for oxide supported Ru catalysts, losing more than 50% of activity in less than 24\u00a0h [21]. Clearly, the stability of the metal catalysts played a critical role in the further application of the bio-fufural derived useful 5-AP and 1,5-PD.Taken that bimetallic catalysts generally present improved catalytic activity, selectivity, and stability in comparison to their monometallic counterparts [24\u201326] and cobalt based catalysts have also been applied to synthesize amines through RA reaction into consideration [14,15,27], NiCo/Al2O3 bimetallic catalysts with LDHs precursor structure were constructed in this work and investigated not only for the RA of 2-HTHP to synthesize valuable 5-AP, but also the direct hydrogenation of 2-HTHP into 1,5-PD. To our knowledge, there were few studies about NiCo bimetal catalysts in RA reactions [28]. The NiCo/Al2O3 bimetallic catalysts were fabricated by a simple co-precipitation method and much attention has been paid to the structure-activity relationship of the NiCo/Al2O3 nanocatalysts for the RA of 2-HTHP to 5-AP owing to the complex reaction network [18,19] and the long-term stability of the catalysts. The incorporation of Co into Ni/Al2O3 catalysts was discovered to greatly enhance the catalytic stability, not only in the RA of 2-HTHP to 5-AP, but also in the direct hydrogenation to 1,5-PD, probably associating with the formation of Ni\u2013Co nanoalloy which inhibited the sintering and surface oxidation of active metal particles. To further elucidate the reaction mechanism and the difference in activity between active metals, DFT (density functional theory) calculations were also performed on the RA of 2-HTHP, in particular on activation of 2-HTHP in the presence of H2, NH3, and Co or Ni.Ru/Al2O3 (5.0\u00a0wt%), Pt/Al2O3 (10.0\u00a0wt%) and Pd/Al2O3 (10.0\u00a0wt%), 1,5-pentanediol (98%), and dihydropyran (99%) were purchased on Alfa Aesar. 1,2-Pentanediol (98%) and 5-amino-1-pentanol (95%) were obtained from Aladdin Chemical Reagent Co. LTD. Aqueous solution of 2-HTHP (\u223c21.8\u00a0wt%) was prepared through the method of an autocatalytic hydration of dihydropyran presented by Huber et\u00a0al. [29]. Dihydropyran and deionized water with a mass ratio 1:4 were added into a 2\u00a0L autoclave and performed under 2.0\u00a0MPa N2, 100\u00a0\u00b0C for 1\u00a0h. The concentration of 2-HTHP was determined by gas chromatograph using 1,2-pentanediol as the internal standard.The Ni\nx\nCo\ny\nAl-LDH hydrotalcite precursors were synthesized via co-precipitation method under pH\u223c10 with various chemical component [(Ni2+\u00a0\u200b+\u00a0\u200bCo2+)/Al3+\u00a0=\u00a02/1, Ni2+:Co2+\u00a0=\u00a01:0, 5:1, 2:1, 1:1, 1:2, 1:5, 0:1]. Briefly, a mixed aqueous solution of Ni(NO3)2\u00b76H2O, Co(NO3)2\u00b76H2O, and Al(NO3)3\u00b79H2O with 0.5\u00a0M total metal concentration, and a mixture of alkali with Na2CO3 (1\u00a0M) and NaOH (5\u00a0M) were poured simultaneously into a three-necked flask while being vigorously stirred. Next, the forming suspension was aged under 80\u00a0\u00b0C maintain 24\u00a0h. And the obtained sediments through filtration and abstersion with deionized water to pH blow 7. After drying at 110\u00a0\u00b0C 12\u00a0h to obtain Ni\nx\nCo\ny\nAl-LDH and then calcined under 700\u00a0\u00b0C for 3\u00a0h to get Ni\nx\nCo\ny\nAl-LDO (LDO: layered double oxide). The Ni\nx\nCo\ny\nAl-LDO samples were activated under 650\u00a0\u00b0C keeping 3\u00a0h with flowing H2 to gain Ni\nx\nCo\ny\n/Al2O3 catalysts. Finally, the samples were transferred into a glovebox filled with Ar gas under sealed quartz tube after cooling to room temperature.X-ray diffraction (XRD) were performed via Rigaku D/MAX-2400 diffractometer in reflection mode with copper K\u03b1 radiation source (\u03bb\u00a0=\u00a00.15406\u00a0nm). The mean diameter of nanoparticles was calculated by Scherrer equation [30]. The in situ XRD characterization was got during programmed temperature to 750\u00a0\u00b0C with H2 (99.999%) at a flow rate of 80\u00a0mL/min with a rate of 5\u00a0\u00b0C/min and the patterns were recorded from 300 to 750\u00a0\u00b0C interval 50\u00a0\u00b0C. X-ray photoelectron spectra (XPS) measurements were tested with ESCALAB250xi spectrometer using an Al K\u03b1 source (h\u03bd\u00a0=\u00a01486.6\u00a0eV). The C 1s at 284.6\u00a0eV served as the reference for validating all binding energies. The catalysts by prereduction were passivated under 1% O2/N2 stream at 30\u00a0\u00b0C maintained 2\u00a0h before taking the XRD and XPS tests.The textural characteristics of the samples containing the BET surface area and average pore diameter were determined through nitrogen adsorption isotherms using a Micromeritics Tristar 3020 under the temperature of liquid nitrogen. The samples were disposed of with N2 flow at 300\u00a0\u00b0C for 4\u00a0h before the test. The scanning electron microscopic (SEM) experiments were accomplished using SU8020 electron microscope at 1\u00a0kV. Transmission electron microscopic (TEM) patterns were characterized by a JEM2010 electron microscope. The transmission electron microscope of FEI Talos 200x (USA) was employed to carry out high angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and energy-dispersive X-ray spectroscopy (EDX) mapping investigations.The H2 temperature-programmed reduction (H2-TPR) was characterized by a Quantachrome Automated Chemisorption Analyzer. Before test, the obtained calcined samples (contain \u223c10\u00a0mg Ni) were pretreated in helium gas stream at 200\u00a0\u00b0C for 30\u00a0min. The samples were cooled to 35\u00a0\u00b0C, reduced under a 10% H2/Ar stream rate of 30\u00a0mL/min, then heated to 900\u00a0\u00b0C at a heating ramp of 10\u00a0\u00b0C/min. The equally quantity of sample after calcined was initially pre-reduced at 650\u00a0\u00b0C for 2\u00a0h with a 10% H2\u2013N2 flow (30\u00a0mL/min) for the sake of determination the degree of reduction of the reduced catalysts for reaction. The acquired reduced sample underwent a second H2-TPR following a temperature increase to 900\u00a0\u00b0C after cooling to below 50\u00a0\u00b0C, and the TCD captured the signal changes. The reduction degree was then calculated with the formula below [31]:\n\n\n\nR\ne\nd\nu\nc\nt\ni\no\nn\n\nd\ne\ng\nr\ne\ne\n\n\n(\n%\n)\n\n=\n\n\nT\nP\nR\n\np\ne\na\nk\n\na\nr\ne\na\n\no\nf\n\nc\na\nl\nc\ni\nn\ne\nd\n\ns\na\nm\np\nl\ne\n\u2212\nT\nP\nR\n\np\ne\na\nk\n\na\nr\ne\na\n\no\nf\n\nr\ne\nd\nu\nc\ne\nd\n\ns\na\nm\np\nl\ne\n\n\nT\nP\nR\n\np\ne\na\nk\n\na\nr\ne\na\n\no\nf\n\nc\na\nl\nc\ni\nn\ne\nd\n\ns\na\nm\np\nl\ne\n\n\n\n\n\n\nA Huasi DAS-7200 automatic chemical adsorption equipment was used to conduct measurements of NH3 temperature programmed desorption (NH3-TPD). Firstly, the 200\u00a0mg of calcined samples underwent a 3-h pretreatment at 650\u00a0\u00b0C under 5% H2\u2013Ar (30\u00a0mL/min). After being cooled to 100\u00a0\u00b0C, the reduced samples were subjected to a 10% NH3\u2013N2 stream for 1\u00a0h (30\u00a0mL/min). The sample was subsequently heated up to 600\u00a0\u00b0C with a 10\u00a0\u00b0C/min rate under He flow, and TCD was used to record the signal of NH3 desorption.NMR spectra of 5-AP were captured at 25\u00a0\u00b0C on a Bruker AV\u2162400. Chemical shift values for 1H and 13C are given as \u03b4 values (ppm) with respect to the deuterated solvent and coupling constants (J) in Hz.A stainless steel autoclave reactor (100\u00a0mL) was used to perform the 2-HTHP RA reaction, with the stirring speed set to 800 r/min. The samples from calcination were activated for 3\u00a0h at 650\u00a0\u00b0C under pure H2 (80\u00a0mL/min). A typical experiment involved adding 15\u00a0g 2-HTHP (21.8\u00a0wt%) aqueous solution to the reactor first, followed by adding the reduced catalyst (0.1\u00a0g) and 15\u00a0g of aqueous ammonia (25\u00a0wt%) to the autoclave operated in the glove box fulling of Ar. After replacing with H2 3 times and pressuring to 2.0\u00a0MPa, then preheated the reactor to the needed temperature, which remained constant throughout the reaction process.On a tubular fixed-bed reactor, the stability of the screened catalyst was tested at 80\u00a0\u00b0C and 2\u00a0MPa H2. Loading with the calcined Ni\nx\nCo\ny\nO-LDO sample (2.0\u00a0g) in the center section of the reactor, which had mesh sizes ranging from 40 to 60. Both sides were then covered in quartz powder (20\u201340 meshes). Before the run, the calcined sample was activated with a reduction program to 650\u00a0\u00b0C kept for 3\u00a0h in pure H2 with an 80\u00a0mL/min flow rate. When the reactor cooled to the target reaction temperature, and pressurized to 2\u00a0MPa with H2, at a weight ratio of 1:1, the mixed solution was infused into the reactor with a rate of 5\u00a0g/h containing 2-HTHP aqueous solution and ammonia. The reaction solution was gathered every 5\u201310\u00a0h in the gas-liquid separator.The reactant products were evaluated via a GC-MS system (Agilent 7890A/5975C). The identified products includes 5-amino-1-pentanol (5-AP), 5-imino-1-pentanol (5-IP), 1,5-pentanediol (1,5-PD), THP-oxypentanimine (THPOPI), 5-[(5-hydroxypentyl)imino]-1-pentanol (5-HPIP), and di-1-pentanolamine (DPA). For quantitative analysis, 1,2-pentanediol (1,2-PD) was used as the internal standard. The 2-HTHP conversion and product selectivity were demonstrated as reported previously [19,20].Geometries of molecules and intermediates involved in reactions depicted in Chart 1\n were adequately optimized on basis of DFT, adopting the hybrid functional M06-2X [32], with the addition of triple-\u03b6 6\u2013311++G\u2217\u2217 basis for light elements (H, C, N, and O) and the LANL2DZ basis\u00a0\u200b+\u00a0\u200bECP for metals. When required, more isomers were considered. The effects of solvent were included utilizing the implicit solvation model SMD [33]. M06-2X functional and SMD solvation models are acknowledged as accurate for reaction energy prediction. To determine whether the stationary points were actually minima (no negative frequencies), harmonic approximation vibrational frequency calculations were performed on optimized geometries or 1st order saddle points (a negative frequency) for stable molecules or transition states, respectively. Thermochemical values were computed under T\u00a0=\u00a0298.15\u00a0K and p\u00a0=\u00a01.00\u00a0atm as well as the harmonic frequencies and IR values.Relaxed scans of the potential energy hypersurface about the ring-opening dihedral angle \u03b4 (C(\u03b1), C(\u03b2), C(\u03b1\u2019), O) of 2-HTHP were performed (\u00b15-degree 36 steps). All computations including all atomic species used the integration grid for the electronic density's 974 angular points and 250 radial shells. Double electron integrals and their derivatives were calibrated with an accuracy of 10\u221212 a.u. The Self-Consistent Field (SCF) methodology employed was the Bacskay-designed quadratically convergent approach, which is considered slower but more accurate than conventional SCF with DIIS deduction [34]. Set the root-mean-square (RMS) variation of the density matrix convergence criterion as 10\u221210 and the maximum variation of the convergence criterion of the density matrix is set as 10\u22128. Convergence requirements for geometry majorizations were established as follows: 2\u00a0\u00d7\u00a010\u22126 a.u. for maximum force, 1\u00a0\u00d7\u00a010\u22126 a.u. for RMS force, 6\u00a0\u00d7\u00a010\u22126 a.u. for maximum displacement, and 4\u00a0\u00d7\u00a010\u22126 a.u. for RMS displacement. The GAUSSIAN G16.C01 package was used to execute all of the calculations.\nTable\u00a01\n displays the textural properties of Ni\nx\nCo\ny\n/Al2O3 samples with diverse Ni/Co ratios. The concentrations of Ni and Co in Ni\nx\nCo\ny\n/Al2O3 catalysts measured by XRF were close to their nominal loadings. The BET surface area decreased generally from 124.3\u00a0m2/g for Ni/Al2O3 without Co to 41.6\u00a0m2/g for monometallic Co/Al2O3 sample with the decrease of Ni/Co mole ratios in Ni\nx\nCo\ny\n/Al2O3 samples. In addition, the average pore diameter of the bimetallic Ni\nx\nCo\ny\n/Al2O3 catalysts was located in the mesopore range of 18.4\u201327.6\u00a0nm (Table\u00a01, Fig.\u00a0S1).\nFig.\u00a01\n displays the XRD profiles of Ni\nx\nCo\ny\nAl-LDH, Ni\nx\nCo\ny\nAl-LDO, and Ni\nx\nCo\ny\n/Al2O3 samples with different Ni/Co molar ratios. The diffraction peaks at 2\u03b8\u00a0=\u00a023.5\u00b0, 35.2\u00b0, 39.6\u00b0, 47.2\u00b0, 61.3\u00b0, and 62.6\u00b0 corresponding to (006), (009), (015), (018), (110), and (113) planes for typical LDHs structure were revealed in all Ni\nx\nCo\ny\nAl-LDH samples (Fig.\u00a01a), respectively [35\u201337]. No Ni or Co phase was detected as a separate crystalline phase, indicating the successful incorporation of the metals into the LDHs structure. In addition, the characteristic reflections gradually broadened as the Ni content increased, which may associate with the intercalation of more Ni chelates in the hydrotalcite layers [36,38].\nFig.\u00a01b shows the XRD patterns of the hydrotalcite-derived Ni\nx\nCo\ny\nAl-LDO samples after calcination at 700\u00a0\u00b0C. The reflection peaks of the LDHs structure disappeared completely and were replaced with the presence of the characteristic diffraction peaks of NiO or Co3O4 or their mixture. For monometallic CoAl-LDO (Fig.\u00a01b, A), the main diffraction peaks at 2\u03b8\u00a0=\u00a031.2\u00b0, 36.8\u00b0, 44.8\u00b0, 55.6\u00b0, 59.3\u00b0, and 65.2\u00b0 were observed, assigning to the lattice planes of Co3O4 spinel with (220), (311), (400), (422), (511), and (440), respectively (JCPDS 78\u20131970). With the decrease of cobalt amount and the increase of Ni loadings in the Ni\nx\nCo\ny\nAl-LDO samples, the spiculate of the diffraction peaks assigning to Co3O4 spinel weakened gradually, meanwhile diffraction peaks at 2\u03b8\u00a0=\u00a037.3\u00b0, 43.3\u00b0, and 62.9\u00b0 assigning to the lattice planes of NiO (111), (200), and (220) (JCPDS 71\u20131179) appeared and intensified (Fig.\u00a01b, B\u2212E). No diffraction peaks belonging to Al2O3 support were observed in all samples, demonstrating that Al2O3 existed in an amorphous form.After activating under 650\u00a0\u00b0C with H2, the metal oxides diffraction peaks disappeared, meanwhile three distinctive diffraction peaks at around 2\u03b8\u00a0=\u00a044.4\u00b0, 51.7\u00b0, and 76.0\u00b0 were found, assigned to the crystal planes of face-centered cubic (fcc) nickel or cobalt (Fig.\u00a01c) [39]. Detailed characterization presented that the characteristic diffraction peak of the (111) plane for fcc Ni and Co of monometallic Ni/Al2O3 and Co/Al2O3 catalysts centered at 44.5\u00b0 and 44.3\u00b0 (Fig.\u00a01d), respectively. Worth mentioning, in bimetallic catalysts, the (111) plane in this region shifted toward the centre, probably attributing to the incorporation of the Co atoms into the Ni lattice in bimetallic catalysts, that is the generation of Ni\u2013Co alloy phase during high temperature treatment in H2 [40,41]. The mean metal crystallite sizes of the Ni\nx\nCo\ny\n/Al2O3 catalysts as obtained from the Scherrer equation slightly decreased from around 7.9\u00a0nm for monometallic Ni/Al2O3 to 7.0\u00a0nm for Ni1Co1/Al2O3 and then gradually increased to 11.8\u00a0nm for monometallic Co/Al2O3 with increasing Co content (Table\u00a01). The slightly lower crystallite size for the bimetallic Ni1Co1/Al2O3 probably originated from the peak broadening caused by the heterogeneity of the bimetallic composition.\nFig.\u00a02\n displays the detailed structural variation of the representative Ni\nx\nCo\ny\n/Al2O3 catalysts during the in situ XRD reduction. For monometallic NiAl-LDO (Fig.\u00a02a), the characteristic peaks of nickel oxide gradually reduced with increasing reduction temperature, simultaneously the characteristic peaks of metal Ni0 vanished at 2\u03b8\u00a0=\u00a037.3\u00b0, 43.3\u00b0, and 62.9\u00b0 after the reduction temperature elevated up to 650\u00a0\u00b0C, demonstrating the reduction of NiO to Ni0. As for monometallic CoAl-LDO (Fig.\u00a02c), the diffraction peaks of Co3O4 presented to 31.2\u00b0, 36.8\u00b0, 44.8\u00b0, 55.6\u00b0, 59.3\u00b0, and 65.2\u00b0 became weaker and eventually disappeared with increasing temperature, meanwhile, the diffraction peak of CoO appeared at 42.3\u00b0 between 450\u00a0\u00b0C and 550\u00a0\u00b0C and subsequently transformed to metal Co0 at a temperature above 600\u00a0\u00b0C. These results indicate the consequent reduction of Co3O4 to CoO and further to Co0, which is consistent with previous research on supported Co catalysts [42,43]. For the bimetallic Ni2Co1Al-LDO sample, the diffraction peaks of NiO or Co3O4 generally disappeared with increasing temperature to 550\u00a0\u00b0C and above, accompanied by the appearance of diffraction peaks at 43.8\u00b0, 50.9\u00b0, and 75.5\u00b0, which may relate with the formation of Ni\u2013Co alloy phase as mentioned above (Fig.\u00a02b). The temperature for the obvious presence of metallic phase in the Ni2Co1Al-LDO sample is at least 50\u00a0\u00b0C lower than that of monometallic NiAl-LDO and CoAl-LDO, showing that the reducibility of the bimetallic oxide sample is greatly improved.\nFig.\u00a03\n displays the SEM morphology of Ni\nx\nCo\ny\nAl-LDO and reduced Ni/Al2O3 catalyst. A three-dimensional flower-like layer structure was observed in all calcined samples, indicating that the layered microstructure of hydrotalcite was preserved even after calcination at 700\u00a0\u00b0C. Note that the stratified structure of the NiAl-LDO was mainly preserved during activation under 650\u00a0\u00b0C with H2 (Fig.\u00a03d), which is consistent with previous findings [19]. Such unique structure features would provide a high exposure of active sites for the catalytic reactions.TEM images of reduced samples were taken to reflect the microstructure of the monometallic and bimetallic Ni\nx\nCo\ny\n/Al2O3 (Fig.\u00a04\na\u2013c). High dispersion of uniform metal Ni was clearly described on Ni/Al2O3 catalyst, while the distributions of metal particles became poorer with introducing Co. The average sizes of the metal nanoparticles gradually increased from 7.1\u00a0nm for Ni/Al2O3 to 11.4\u00a0nm for Co/Al2O3 with increasing Co loading, agreeing well with the XRD results (Table\u00a01). For the Ni2Co1/Al2O3 catalyst, the HRTEM images showed lattice spacings of approximately 0.202\u00a0nm for Ni\u2013Co particles (Figs. 4d and S2). Comparing with the standard lattice distance of Ni(111) (0.204\u00a0nm) and Co(111) (0.205\u00a0nm), the slightly smaller NiCo(111) lattice spacing may associate with the formation of lattice-strained configuration and strong interaction between bimetallic atoms [44]. Furthermore, the high-angle annular dark-field (HAADF)-STEM was also taken for Ni2Co1/Al2O3 (Fig.\u00a04e), which displayed the high dispersion of bimetallic NiCo nanoparticles. The elemental mapping demonstrated homogeneous distributions of Ni, Co, Al, and O elements (Fig.\u00a04f\u2013i). Meanwhile, the line-scan spectra of elements Ni and Co (Fig.\u00a04j) illuminated similar Gaussian distributions along with a single particle, confirming the formation of Ni\u2013Co alloy nanostructure [45]. There have been claimed that the generation of bimetallic alloy particles is helpful for inhibiting the sintering of active metal nanoparticles [24]. Thus, the generation of Ni\u2013Co alloy nanoparticles in our case may contribute to the greatly improved reaction stability of the bimetallic catalysts.H2-TPR experiments were performed to explore the reductive behavior and the interaction of metal oxides with the support of Ni\nx\nCo\ny\nAl-LDO catalyst precursors (Fig.\u00a05\n). Monometallic NiAl-LDO showed a couple of H2 consumption peaks with a minor one at around 460\u00a0\u00b0C and a major one at about 737\u00a0\u00b0C. The minor H2 consumption peak might be associated with the reduction of greatly distributed Ni oxides on the surface, while the major peak might be ascribed to the reduction of Ni2+ species with a strong metal-support association [19]. For the CoAl-LDO sample, two distinct reduction peaks with one at 475\u00a0\u00b0C and another broad one centered at about 725\u00a0\u00b0C were observed, which might be associated with the transition of Co3O4 to CoO and then transferring to Co [42,46]. Such a reduction process is in consistent with the in situ XRD study (Fig.\u00a02c). There are two reduction areas for calcined bimetallic Ni\nx\nCo\ny\nAl-LDO within the scope of 400\u2013500\u00a0\u00b0C and 600\u2013700\u00a0\u00b0C (Fig.\u00a05B-F), which are likely attributable to the reduction both of Co3O4 and NiO. Both reduction peaks shift generally towards lower temperatures compared with monometallic samples, indicating that the incorporation of Co species into bimetallic Ni\nx\nCo\ny\nAl-LDO samples enhanced the reduction of bimetallic samples, which is also reflexed by the calculated reduction degree shown in Table\u00a01. The above results illustrate the possible existence of Ni\u2013Co strong interaction in bimetallic Ni\nx\nCo\ny\nAl-LDO samples, which facilitates the reduction of metal oxide as a consequence [44].According to the reports that the surface acidic site of the catalyst could facilitate the adsorption and activation of CO group and promote the generation of imine by the condensation of CO with NH3 [7,47]. Meanwhile, the production of the iminium ion may facilitate imine intermediate hydrogenation on account of the abundance of acid sites [48,49]. The surface acidity of the Ni\nx\nCo\ny\n/Al2O3 catalysts was characterized by NH3-TPD. As shown in Fig.\u00a06\n, three distinct NH3 desorption regions were observed at 150\u2013250\u00a0\u00b0C, 250\u2013400\u00a0\u00b0C, and 400\u2013600\u00a0\u00b0C, which might be associated with the acidic sites of weak (WA), medium-strength (MSA), and strong (SA), respectively [7,50]. The concentration of acidic sites especially for medium-strength and strong of Ni\nx\nCo\ny\n/Al2O3 catalysts gradually decreased with the increase of cobalt content (Fig.\u00a06 and Table\u00a0S1), which may be result of the increasing amount of CoO\nx\n species with a rather low acidity in the catalysts. A previous study by Li and coworkers [44] also found that supported Co catalysts presented lower acidity than those Ni catalysts.\nTable\u00a02\n presents the catalytic performances of a series of LDHs-derived monometallic and bimetallic NiCo/Al2O3 catalysts in the conversion of 2-HTHP by RA to 5-AP at the condition of 60\u00a0\u00b0C and 2.0\u00a0MPa H2. A high conversion of 2-HTHP (91%) was detected even without a catalyst, on account of the high reactivity of 2-HTHP in the presence of ammonia [19]. The main by-products, 5-IP and THPOPI, were found with selectivity of 43% and 29%, respectively (entry 1). The condensation of 5-IP intermediate with 2-HTHP is most likely to generate THPOPI. Similar results were achieved with Al2O3 support as a catalyst (entry 2). The Ni\nx\nCo\ny\n/Al2O3 catalysts derived from hydrotalcite with different chemical compositions exhibited different catalytic activities for the RA of 2-HTHP (Table\u00a02, entries 3\u201310). For monometallic Ni/Al2O3 catalyst with 72% Ni loading, a high selectivity of 90% to the target product 5-AP was achieved at 100% conversion of 2-HTHP, and then the selectivity of 5-AP further increased to 93% under an optimum temperature of 80\u00a0\u00b0C (Table\u00a02, entry 4 and Fig.\u00a0S3). With the increase of cobalt from a Ni/Co ratio of 5:1 to 0:1, not only the conversion of 2-HTHP dropped to 96%, but also the selectivity of 5-AP decreased monotonously from 87% on Ni5Co1/Al2O3 to 9% over Co/Al2O3, and the imine product 5-IP increased gradually from 0 to 38%. The selectivity to the secondary imine 5-HPIP, generated from the condensation of the target product 5-AP with the 5-HP intermediate, exhibited a trend of first increasing and then decreasing. Meanwhile, a higher amount of DPA, a secondary amine by-product, was detected with a selectivity of 9% over cobalt-riched Ni1Co5/Al2O3 and Co/Al2O3 with 67% and 78.6% Co loadings (Table 1), respectively. These results indicate that cobalt-riched catalysts not only presented lower activity for the conversion of 2-HTHP, but also much inferior in the hydrogenation of imine intermediates to the target amine product as compared with nickel-riched catalysts. The catalytic results of conventionally impregnated monometallic Ni and Co catalysts further supported the above findings (entries 10, 11). Ni/Al2O3-IM showed medium 5-AP selectivity (33%) at full 2-HTHP conversion while Co/Al2O3-IM presented not only lower 2-HTHP conversion (94%) but also rather low 5-AP selectivity (\u223c0%). Moreover, commercial Al2O3 supported noble metals of Pd, Ru, and Pt (all with 10% metal loading), as well as Raney Ni, were also investigated for the catalytic RA of 2-HTHP (entries 13\u201316). These catalysts demonstrated rather low catalytic performances for the synthesis of 5-AP, with only the Raney Ni catalyst exhibiting slightly higher 5-AP selectivity of 54% at 94% 2-HTHP conversion.Clearly, the above results indicate that the LDHs-derived Ni-based nanocatalysts, particularly for those with high Ni/Co ratios (i.e.\u00a0\u2265\u00a05:1), exhibited outstanding catalytic activity for the RA of 2-HTHP. Since the influence of the crystallite sizes and reduction degree of the Ni\nx\nCo\ny\n/Al2O3 catalysts, especially the catalysts with Ni/Co ratio greater than 1:5 showed lower catalytic activity (Table\u00a01), the monotonous decrease of activity for the generation of 5-AP of the Ni\nx\nCo\ny\n/Al2O3 catalysts (Table\u00a02, entries 3, 5\u201310) would be associated with the nature of Ni and Co in the hydrogenation of imine intermediates, and the surface acidity of the catalysts. It has been claimed that the rate-controlling step in RA of carbonyl compounds is the hydrogenation of imines [15], and the hydrogenation of 5-IP to 5-AP is deduced to be the rate-controlling step for this reaction [19]. Therefore, the advantageous catalytic activity of the Ni-riched Ni\nx\nCo\ny\n/Al2O3 catalysts might primarily account for their higher activity in imine intermediates hydrogenation as compared with the Co-riched catalysts. Larger metal particle size and lower degree of reduction on the Co-riched catalysts (i.e. Ni:Co ratio \u22641:5, Table\u00a01), which mean lower amounts of active surface sites, would also partially result in their inferior hydrogenation activity for the imine intermediates. The experiments on the direct hydrogenation of 2-HTHP to synthesize 1,5-PD at similar conditions as the RA of 2-HTHP except for adding deionized water instead of ammonia solution showed the same trend as the RA of 2-HTHP, that is both the conversion of 2-HTHP and 1,5-PD selectivity decreased with increasing Co content (Table\u00a0S2). Such finding further supported the inferior hydrogenation activity of the Co component at the low reaction temperature of 60\u00a0\u00b0C, although several reports showed Co-based catalysts presented high RA activity at temperatures above 100\u00a0\u00b0C [14,15,26]. Despite 2-HTHP could facilely react with NH3 through reactive 5-HP intermediate to form imine intermediates of 5-IP and THPOPI even without the presence of a catalyst (Table\u00a02, entries 1, 2), the presence of larger amounts of surface acidic sites of the Ni-rich catalysts (Fig.\u00a06) may promote the formation of imine intermediates and their further hydrogenation as reported previously [7,47\u201349], which would to some extent contributes to the higher RA activity of these catalysts.It is well known that stability is an important fact for a heterogeneous catalyst. Nonetheless, metal catalysts, especially monometallic catalysts, often suffer from sintering, surface oxidation of active sites, and/or loss of active components in the RA reaction in the presence of ammonia [5,8,47]. Although the stability of the Ni\u2013Mg3AlO\nx\n catalysts with hydrotalcite precursor structure prepared by co-precipitation improved a lot as compared with the Ni/ZrO2 catalyst synthesized by a conventional method of impregnation in the RA of 2-HTHP, the stability of the former is still unsatisfactory, deactivation obvious after 90\u00a0h running [19]. Herein, the stability of the Ni2Co1/Al2O3 bimetallic catalyst derived from hydrotalcite precursor was first evaluated in the RA of 2-HTHP to see whether the incorporation of Co could enhance the stability of the catalyst or not. To our delight, the catalyst maintained high stability during 180\u00a0h running (Fig.\u00a07\na). The conversion of 2-HTHP retained under 100% and the selectivity of target product 5-AP just slightly declined from around 90% to close to 82% during the 180\u00a0h running. The deactivation rate based on the decrease of 5-AP yield was calculated to be 9% after 180\u00a0h time on-stream, which is obviously lower than that of Ni\u2013Mg3AlO\nx\n catalysts (18% after 120\u00a0h reaction) and Ni/ZrO2 (\u223c19% after 90\u00a0h reaction) reported previously [18,19]. Then, we compared the stability of the NiCo/Al2O3 bimetallic catalysts with the monometallic Ni/Al2O3 at a five times higher feeding rate of WHSV\u00a0=\u00a02.5 h\u22121 (Fig.\u00a07b). The yield of 5-AP for Ni/Al2O3 noticeably decreased from 72% to 39% after 60\u00a0h reaction. In contrast, the yield of 5-AP for the bimetallic Ni5Co1/Al2O3 and Ni2Co1/Al2O3 just slightly decreased from 70% to 54% and from 62% to 52%, respectively. The deactivation rate for the monometallic Ni catalyst reached 46%, which is \u223c2.2 and 2.9 times higher than that of the latter bimetallic catalysts, revealing the outstanding stability of the NiCo bimetallic catalysts. Note that the Ni2Co1/Al2O3 bimetallic catalyst also presented good stability in the reaction of 2-HTHP direct hydrogenation to synthesize useful 1,5-PD. No appreciable decrease in the yield of 1,5-PD was displayed after 180\u00a0h of time on-stream (Fig.\u00a07c). Previous studies by Huber and co-workers [21] showed that the oxide supported Ru catalyst, i.e. Ru/TiO2, suffered from rapid deactivation in the hydrogenation of 2-HTHP to synthesize 1,5-PD, losing more than 50% of activity in less than 24\u00a0h.Clearly, the above findings reveal the high stability of the NiCo bimetallic catalyst with LDH precursor structure, and the incorporation of Co into the Ni/Al2O3 catalysts could remarkably increase the stability of the bimetallic catalyst, although the incorporation of Co to some extent lowered the selectivity to the product of 5-AP (Table\u00a02), which is probably due to the low hydrogenation activity of Co at the low reaction temperature of 60\u00a0\u00b0C. As discussed above, the strong interaction between the metal nickel and cobalt species assisted the formation of a highly dispersed alloy structure, which would eventually contribute to the increased stability of the bimetallic catalyst as compared with the monometallic Ni/Al2O3 catalyst [44,51]. It should be remarked here that the Ni\u2013Al2O3 nanocatalysts with similar Ni loadings (\u223c50\u00a0wt%) prepared by a similar co-precipitation method reported previously by our group presented substantially good stability in the RA of not only 2-HTHP (biomass-derived aldehyde) [20], but also 5-diethylamino-2-pentanone (biomass-derived ketone) [52] without appreciable deactivation during 150 and 200\u00a0h running, respectively. The somewhat better stability of these Ni\u2013Al2O3 nanocatalysts as compared with the Ni\nx\nCo\ny\n/Al2O3 catalysts presented in this work is probably due to that the former catalysts prepared with lower metal ions concentration (0.1\u00a0M vs. 0.5\u00a0M in this work) and thus exhibited higher metal dispersion (smaller Ni particle sizes, i.e. \u223c5.4\u00a0nm vs. 7\u20138\u00a0nm in this work) [52]. Taken together, despite the stability of the NiCo bimetallic catalysts with LDH precursor structure needs further improvement, the incorporation of Co exactly promoted the stability of the bimetallic catalysts, and may shed light on designing novel bimetallic catalysts for not only RA reaction but also several other reactions, such as hydrogenation and oxidation.To further investigate the improvement in stability with the incorporation of Co, the used Ni2Co1/Al2O3 and Ni/Al2O3 catalysts after 60\u00a0h at a higher feeding rate were characterized by XRD, TEM, and XPS. As shown in Fig.\u00a08\na and b, apparent diffraction peaks of Ni0 and Ni\u2013Co alloy were observed at around 2\u03b8\u00a0=\u00a044.4\u00b0, 51.7\u00b0, and 76.0\u00b0, and no additional peaks appeared after the RA reaction. The Ni\u2013Co alloy characteristic diffraction peaks were well maintained and no separate phases of Ni0 and Co0 could be seen (Fig.\u00a08b). The Ni crystallite size of Ni/Al2O3 catalysts significantly grew from 7.9\u00a0nm to 14.6\u00a0nm after the reaction, while the Ni\u2013Co alloy particles just slightly grew from 7.8\u00a0nm to 10.1\u00a0nm under similar reaction conditions, showing that the incorporation of Co could retard the sintering of the active metals. TEM characterizations supported the findings by XRD that a more obvious growth of metal particles on the used Ni/Al2O3 than on the used Ni2Co1/Al2O3 (Fig.\u00a08c and d). XRD and TEM characterization of the Ni2Co1/Al2O3 catalysts after 180\u00a0h time-on-stream RA of 2-HTHP or direct hydrogenation of 2-HTHP at WHSV of 0.5 h\u22121 also confirmed the good stability of the bimetallic catalyst (Figs.\u00a0S5 and S6). XPS characterization was also presented to reveal the variation of the valence of the surface Ni species before and after the catalysts were used. As displayed in Fig.\u00a0S4, a large amount of Ni0 species on monometallic Ni/Al2O3 catalyst was surface oxidized to Ni2+ after utilization, as the Ni0/(Ni0 + Ni2+) ratio decreased obviously from 42.6% to 10.9%, while the ratio for Ni2Co1/Al2O3 slightly dropped from 31.3% to 24.0%. It has been reported that zero-valent metal is the active site for the hydrogenation of imine intermediate to form amine [15,52], the decline in the Ni0/(Ni0\u00a0\u200b+\u00a0\u200bNi2+) ratio means the decrease in hydrogenation sites. Obviously, the above findings by XPS indicate that the incorporation of Co could retard the surface oxidation of the more active Ni0 species, which may associate with its higher reducibility (Figs.\u00a02 and 5) and eventually contributes to the high stability of the NiCo bimetallic catalyst during the reaction. The inevitably dissolved oxygen in the reactant caused the variation of the valence on the surfaced active metal during the long-term time-on-steam reaction. Taking the generation of Ni\u2013Co alloy by incorporating Co into consideration, the Ni\u2013Co alloy species in the bimetallic catalyst could on one hand inhibit the sintering of the metal particles, on the other hand, retard the surface oxidation of active metals during the reaction, and thus remarkably improved of stability of the bimetallic catalysts.Previously, in our proposed reaction pathway for the RA of 2-HTHP, 2-HTHP is first equilibrated with its ring-cleavaged tautomer of 5-HP, which is quickly condensed with NH3 to obtain the 5-IP imine intermediate, followed by the hydrogenation to the 5-AP in the presence of adequate H2 [18,19]. DFT calculations were performed in this work to provide a computational model for potential reaction routes. Reactant 2-HTHP has two possible conformers (namely, A and B), resting with the position of the hydroxyl group (axial or equatorial, respectively). Conformer A is less stable than conformer B (\u0394H\u00a0=\u00a00.8\u00a0kJ/mol and \u0394G\u00a0=\u00a00.9\u00a0kJ/mol); these two conformers may convert one to the other by means of tautomerization and rotation about \u03c3 bond (vide postea). Either conformer A or B admits further rotamers, deciding by the orientation of the OH group about the C\u2013O bond; the energy differences between the stable rotamers are negligible, as the rotational energy barriers. This is found also for other species involved in the reaction paths; in general, the spacial orientation of OH and NH groups origins rotamers, with small energy differences among them. The reactions involving 2-HTHP have been computed for conformers A and B separately; the other reactions are common for the two conformers (and hence the computed values are the same). Reactant 2-HTHP may undergo three primary reactions: (i) a monomolecular reaction of ring opening to give 5-HP (reaction 1 in Chart 1); (ii) hydrogenation to give 1,5-PD (reaction 3); and (iii) addition of ammonia with elimination of water to give 5-IP (reaction 7). It may also react with products of reactions 3 and 7 to give by-products by addition and condensation (reactions 5 and 9, respectively); for these reactions activation energies were not computed. Reaction 1 may hypothetically proceed via two different ways, viz. radical ring opening (1r) or tautomerization (1t). Reaction 1r is very energetically unfavored (Table\u00a03\n), whereas reaction 1t has a lower value of \u0394H and \u0394G, even without metal presence. Moreover, reaction 1r implies two subsequent steps, viz. the radical breakage of the O\u2013C bond and the subsequent intramolecular hydrogen shift (whose values of \u0394H\n\u2021 and \u0394G\n\u2021 are depicted in brackets). The values of \u0394H\n\u2021 and \u0394G\n\u2021 of reaction 1t are much lower than those of reaction 1r; the presence of metal further reduces these values. The computed values of \u0394G for reaction 1t are compatible with the experimental values of equilibrium composition at the room temperature reported in the literature [28]. The Boltzmann populations calculated at T\u00a0=\u00a0298.15\u00a0K and p\u00a0=\u00a01.00\u00a0atm using the calculated thermochemical quantities are x(1-HTHP): x(5-hydroxypentanal)\u00a0=\u00a095.4%:4.6%. We can therefore conclude that reaction 1 proceeds via tautomerization and not via ring opening. Once the tautomerization of 2-HTHP has occurred, 5-hydroxypentanal may react either with H2 (reaction 2) or NH3 (reaction 6); reactions 2 and 6 are competitive. However, reaction 6 has a higher rate than reaction 2\u00a0\u200bat a low reaction temperature. Note that the calculations for reactions 1 and 7 were also repeated by modeling the metal surface with a (2 frozen\u00a0\u200b+\u00a0\u200b2 relaxed)-layer slab composed of 6\u00a0\u200b\u00d7\u00a0\u200b6 metal atoms; a few possible anchorings (top, bridge, hollow) and molecule's orientations were tested but only those giving the lowest energy differences were not rejected. As expected, the numerical values are quantitatively different from those obtained with the above (more simplified) model; however, the overall trend and general conclusions that can be dragged from these results remain the same.For reaction 3, \u0394H\n\u2021 and \u0394G\n\u2021 are not available, since no transition state was identifiable. The target product 5-AP may be obtained either via the addition of ammonia to 1,5-PD and the release of water (reaction 4) or hydrogenation of 5-IP (reaction 8). The activation energy of reaction 4 is rather high, which means the dehydrogenation amination of 1,5-PD is very hindered at low temperatures. Reaction 8 has a larger activation energy than reaction 6; this supports the experimental evidence that reaction 8 is the rate-controlling step for the synthesis of 5-AP [19]. The reason why without a catalyst the reaction of the addition of ammonia is more favored than hydrogenation of 2-HTHP may be the reaction medium include a significant amount of unreacted ammonia, whereas H2 not only needs to get dissolved in the solution, but also needs to be activated over metal sites. The by-product 5-HPIP may not form via two subsequent reactions of 10 and 11, but more likely by means of the reaction between 5-AP and the reaction intermediate 5-HP [19]. Finally, the presence of metal always reduces the activation energies, confirming its role as a catalyst; in particular, Ni is more active than Co in each reaction, supporting the rather low activity of Co catalyst in the RA of 2-HTHP (Table\u00a02).A series of Al2O3-supported monometallic and bimetallic Ni\u2013Co nanocatalysts with diverse Ni/Co ratios derived from hydrotalcite precursors were synthesized through co-precipitation approach and used for the catalytic RA and direct hydrogenation of biofurfural-derived 2-HTHP to synthesis useful 5-AP and 1,5-PD, respectively. Both yields towards the target 5-AP and 1,5-PD products decreased with increasing incorporation of Co. However, the introduction of Co improved the reducibility of the NiCo/Al2O3 bimetallic catalysts and promoted the reaction stability of the bimetallic catalysts in both reactions with over 180\u00a0h time-on stream. Detailed characterization of the catalysts before and after the reaction indicated that the incorporation of Co resulted in the formation of NiCo alloy, which helps to inhibit the agglomeration of the metal particles and hinder the surface oxidation of the more reactive Ni0 species as compared with Co0. DFT-based modeling of the reaction mechanisms confirmed the reaction pathway proposed that 5-AP is formed via the hydrogenation of 5-IP intermediate, and also supported the higher reactivity of Ni in the RA of 2-HTHP to synthesize 5-AP as compared with Co. The important findings in this work would shed light on the development of more reactive and stable nanometal catalysts in the RA or hydrogenation of carbonyl compounds from not only biomass but also fossil resources to produce useful chemicals.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was financially supported by the National Natural Science Foundation of China (21872155, 22102198, and 22272187), the Strategic Pilot Science and Technology Project of the Chinese Academy of Sciences (XDA21010700), and the CAS \"Light of West China\" Program. The authors also acknowledge the helpful discussion for Prof. George W. Huber at the University of Wisconsin-Madison.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.gce.2023.01.003.", "descript": "\n                  Al2O3-supported monometallic Ni, Co, and bimetallic Ni\u2013Co nanocatalysts originated from layered double hydroxide precursors were synthesized by co-precipitation method, and used for the synthesis of useful 5-amino-1-pentanol (5-AP) and 1,5-pentanediol (1,5-PD) by reductive amination (RA) or direct hydrogenation of biofurfural-derived 2-hydroxytetrahydropyran (2-HTHP), respectively. In both reactions, the yield of the target products decreased monotonously with the increasing amounts of Co in the NiCo/Al2O3 catalysts, owing probably to the replacement of highly reactive Ni by Co component with inferior hydrogenation activity at the low reaction temperature of 60\u00a0\u00b0C. However, the incorporation of Co could improve the reducibility of the NiCo/Al2O3 bimetallic catalysts and promote the reaction stability of the catalysts, especially for Ni2Co1/Al2O3, in both reactions with over 180\u00a0h time-on-stream. Characterization of the catalysts before and after the reaction showed that the incorporating Co could inhibit the sintering of metal particles and hinder the surface oxidation of the more reactive Ni0 species, thanks to the formation of Ni\u2013Co alloy in the bimetallic catalysts. DFT-based modeling of the reaction mechanisms is also performed, supporting the reaction pathway proposed previously and also the much higher activity of Ni in the RA of 2-HTHP as compared with Co.\n               "}
{"full_text": "Alkynyl carboxylic acids are an important class of compounds owing to their existence in a variety of biologically active natural products [1]. The direct carboxylation of terminal alkynes with CO2 is a particularly useful synthetic method since CO2 is inexpensive, easily-available, non-toxic and thus can be considered as an ideal C1 subunit in chemical synthesis [2,3]. To date, a number of catalyst systems for terminal alkyne carboxylations have been reported [4\u20139], and these reactions are conducted most frequently with noble metal Ag catalysts. Though several homogenous copper catalytic systems were found in the carboxylation reaction [10\u201313], further application was seriously impeded by the recyclability and reusability of the catalysts as well as the metal contamination of the products. To solve these problems, a handful of Cu-based heterogeneous catalysts have been developed for the alkyne carboxylation with CO2 in recent years. For instance, He and co-workers showed CuBr supported on activated carbon as recyclable catalyst for the carboxylation reaction [14]. Kim's group recently incorporated CuCl2 within polyvinyl imidazolium tri-cationic ionic liquid and found that such a specie exhibited an enhanced activity in catalyzing the coupling of alkyne and CO2 [15].Coordination polymers (CPs) and their unique subclass of metal-organic frameworks (MOFs) are an emerging class of functional organic-inorganic hybrid materials, which can be directly adopted as heterogeneous catalysts or as catalyst supports/precursors for many important organic transformations [16\u201320]. Recently, a few MOFs, such as MIL-101 [6], ZIF-8 [21], and UiO-66 [22], have been used as supports for the preparation MOF-supported Ag nanoparticle catalysts for the coupling of alkynes with CO2. In contrast, reports on the catalytic CO2 fixation with only MOFs as catalysts are much more scarce [23,24]. The copper(II) MOF-catalyzed carboxylation of terminal alkynes, first reported by Zhao [23] and later developed by Verpoort [24], demonstrated the feasibility of the utilization of copper(II) MOFs as efficient heterogeneous catalysts. Very recently, we reported a mixed-ligand [Cu4(\u03bc3-OH)2]-cluster-based CP catalyst for carboxylation of terminal alkynes with CO2 [25]. Further development of new Cu(II)-based MOF catalysts for these types of reactions is, therefore, highly desirable.According to the Irving-Williams stability series [26], late 3d transition metal ions (soft Lewis acids, such as Co(II), Ni(II), Cu(II) and Zn(II)), being coordinated by N-heterocyclic ligands (soft Lewis bases, such as imidazole-, triazole-, and pyridine-based compounds), tend to favor the formation of stable complexes. In our previous work, the flexible fluorinated bis(1,2,4-triazle) ligand 1,4-bis(1,2,4-triazole-1-ylmethyl)-2,3,5,6-tetrafluorobenzene (Fbtx) has been proved to be a good candidate to assemble with Cu(II) and Co(II), affording stable MOF catalysts for several diverse transformations [27\u201329]. On the other hand, developing a fast and efficient method for the preparation of MOF-based catalysts is of great significance. In this aspect, the use of microwave approach is much more appealing than the conventional solvothermal method owing to the apparent advantages such as shorter reaction time, fast kinetics of crystal nucleation and growth, higher yields, and better reproducibility [30]. Very recently, we have also reported the application of microwave assistance in the rapid production of a Cu(II)-based MOF catalysts [25]. In this contribution, we report the microwave-assisted synthesis, structural characterization, and catalytic properties of a new Cu(II)-based MOF [Cu(Fbtx)2Br2]\nn\n (denoted as CZU-7). Single-crystal X-ray diffraction analysis revealed that CZU-7 features a two-dimensional coordination framework with the 44-sql topology. Furthermore, the crystalline CZU-7 was shown to be an efficient heterogeneous catalyst for the direct carboxylation of terminal alkynes with CO2 under mild conditions.A mixture containing Fbtx (31.3\u00a0mg, 0.1\u00a0mmol), CuBr2 (22.3\u00a0mg, 0.1\u00a0mmol) and distilled water (6\u00a0mL) was placed in a 100-mL Teflon-lined container. The container was sealed and then irradiated at 100\u00a0\u00b0C in microwave oven with power 300\u00a0W for 100\u00a0min. After cooling down to room temperature, blue block crystals of CZU-7 were collected, washed with distilled water, and dried at room temperature (Yield: 79.3% based on Fbtx). Anal. calcd for C24H16Br2CuF8N12: C 34.00, H 1.90, N 19.82%; found: C 34.52, H 1.91, N 19.75%. Selected IR peaks (KBr pellet, cm\u22121): 3168 (w), 3126 (m), 3088 (m), 3033 (w), 2971 (w), 2946 (w), 1526 (s), 1493 (s), 1438 (w), 1398 (m), 1358 (w), 1329 (w), 1289 (s), 1274 (s), 1224 (w), 1196 (m), 1135 (s), 1044 (s), 1032 (s), 999 (m), 987 (m), 916 (w), 892 (m), 862 (m), 793 (m), 752 (s), 670 (s), 637 (w), 570 (w).In a typical experiment, a 50-mL stainless steel autoclave, equipped with a magnetic stirring bar, was charged with CZU-7 (33.9\u00a0mg, 1\u00a0mol%), Cs2CO3 (1.96\u00a0g, 6\u00a0mmol), 1-ethynylbenzene (0.41\u00a0g, 4.0\u00a0mmol) and DMF (20\u00a0mL). Once added, CO2 (0.3\u00a0MPa) was introduced into the reaction mixture under stirring at 100\u00a0\u00b0C for 16\u00a0h. After the reaction, the mixture was cooled to room temperature and monitored by high-performance liquid chromatography (HPLC, Shimadzu LC-10\u00a0CE).For a long time, developing a fast and efficient method for the preparation of CP-based materials is of great significance. In this work, CZU-7 was synthesized by microwave-assisted hydrothermal reaction at 100\u00a0\u00b0C, which led to a significant yield (79.3%) with 100\u00a0min of reaction time. In contrast, reaction of CuBr2 with Fbtx under hydrothermal conditions at 120\u00a0\u00b0C for 3\u00a0days gave rise to blue single crystals of CZU-7 in 58.2% yield. The phase purity of the samples of CZU-7, prepared by microwave-assisted method and hydrothermal method, was confirmed by similarities between the as-synthesized and simulated PXRD patterns (Fig. S2). Scanning electron microscopy (SEM) analysis was performed to determine the morphology and size of CZU-7 prepared using the microwave method. Fig. S3 shows that the CZU-7 material is mainly consisted of microparticles formed from packing of nanosheets with the thickness of about 120\u00a0nm. CZU-7 is insoluble in water and in most of common organic solvents, such as ethanol, acetonitrile, dichloromethane, DMF, and DMSO.X-ray structural analysis revealed that CZU-7 crystallizes in the centrosymmetric triclinic space group P\n\n\n1\n\u00af\n\n. The asymmetric unit contains half of a Cu(II) ion, two halves of crystallographically unique Fbtx ligands, and one terminal-coordinated bromide ion. As shown in Fig. 1a, the Cu(II) center displays a distorted octahedral geometry, surrounded by four nitrogen atoms (N1, N1#1, N4 and N4#1) from four Fbtx ligands with the CuN bond distances of 2.003(7) and 2.017(8) \u00c5 and two bromide atoms (Br1 and Br1#1) from two bromide ions with the CuBr bond distance of 3.110(7) \u00c5. Each Fbtx ligand adopts the anti-conformation to avoid the steric hindrance and connects adjacent Cu(II) ions to form a 2-D window-shaped layer with 44-sql topology (Fig. 1b). The adjacent layers are stacked in a staggered pattern along the b-axis, which are further interconnected via interlayer \u03c0\u00b7\u00b7\u00b7\u03c0 interactions between the tetrafluorinated benzene rings with a centroid-centroid separation of 3.41\u00a0\u00c5 (Fig. 1c). The 3D lattice of CZU-7 has no solvent-accessible area, as calculated by the PLATON package [31].\nThe catalytic activity of CZU-7 for carboxylation of terminal alkynes with CO2 was investigated with 1-ethynylbenzene using Cs2CO3 in DMF at 100\u00a0\u00b0C over 16\u00a0h. The catalytic activity of CZU-7 for carboxylation of terminal alkynes with CO2 was investigated with 1-ethynylbenzene using Cs2CO3 in DMF at 100\u00a0\u00b0C over 16\u00a0h. The conversion increased with the catalyst loading (1\u00a0mol% of Cu) and reached 87% (Table 1\n, entries 1\u20134), and the reaction proceeded with 100% selectivity in favor of phenylpropiolic acid. The catalytic efficiency was slightly lower than that of our previous reported Cu-MOF system [25], and was comparable to those reported for Cu(IN)-MOF [24]. The reaction using the CZU-7 catalyst obtained from microwave method resulted in a slightly higher conversion than that obtained from hydrothermal method (entry 5), which may be attributed to the morphology of nanostructures. Lower temperatures (40 to 80\u00a0\u00b0C) reduced the conversion, while a higher temperature (120\u00a0\u00b0C) did not increase the conversion (Table 1, entries 6\u20139). We also found that a temperature of 100\u00a0\u00b0C and a reaction time of 16\u00a0h were required (Fig. S4) The screening of different solvents (Table 1, entries 1 and 10\u201315) revealed that DMF was found to be the most effective for the carboxylation reaction to give a 87% conversion (Table 1, entry 4), which may be due to the fact that DMF is a better solvent for both CO2 (DMF is a weak base) and Cs2CO3 than the other organic solvents [32].Variation of Cs2CO3 to the other bases like Na2CO3 and K2CO3 under otherwise identical conditions led to the decrease of the conversion into 17\u201329% (Table 1, entries 16 and 17). Surprisingly, if we added 3\u00a0mmol CsCl to the reaction system of K2CO3, the conversion elevated from 29% to 41% (Table 1, entry 18), thus revealing that Cs+ could promote this reaction. It was observed that CsF and CsOAc were inactive to this reaction (Table 1, entries 19 and 20). The results indicated that Cs2CO3 was the suitable base for deprotonation of terminal alkynes [11,33]. The impact of CO2 pressure on the reaction was also investigated (Fig. S5). The carboxylation reaction took place smoothly under a CO2 pressure of 0.3 Mpa. To gain insight into the catalytic mechanism, several control experiments were conducted. In the absence of CZU-7, the reaction of 1-ethynylbenzene with CO2 resulted in negligible conversion (Table 1, entry 5). The introduction of ligand Fbtx did not promote the reaction at all (Table 1, entry 21). Copper salts (CuBr2, Cu(NO3)2 and CuSO4) and copper oxides (Cu2O and CuO) were found to catalyze the carboxylate reaction with moderate conversion (Table 1, entries 22\u201326), which implied that the introduction of metal Lewis acidic catalysts is favorable.To extend the substrate scope, our further examination focused on the carboxylation of several representative terminal alkyne substrates (Table 2\n). Under the optimized reaction conditions of CZU-7 (1\u00a0mol% of Cu), 6\u00a0mmol Cs2CO3, 20\u00a0mL DMF, 100\u00a0\u00b0C, and 0.3 Mpa CO2, the corresponding carboxylation products were obtained in moderate to good yields (65\u201384%) when aromatic alkynes with either an electron-donating group (Table 2, entries 2\u20135) or electron-withdrawing groups (Table 2, entries 6\u201311) were used. Even for an alkyne with a heteroaromatic ring group (Table 2, entry 12), a yield of 72% was achieved. The aliphatic alkyne 1-hexyne successfully underwent the carboxylation reaction to give the corresponding product (Table 2, entry 13).Thermogravimetric analysis (TGA) measurement indicated that CZU-7 can maintain their structural integrity until the collapse of the coordination frameworks beginning at ca. 245\u00a0\u00b0C (Fig. S6). The temperature-dependent PXRD patterns (Fig. S7) showed that its structure was retained after the crystalline samples were heated at 220\u00a0\u00b0C in air for 6\u00a0h. To check the stability of CZU-7, we carried out leaching and recycling experiments. When the filtrate of the reaction mixture after catalysis was used instead of CZU-7 under identical conditions, the conversion was negligible (Fig. S8). Inductively coupled plasma-mass spectrometry (ICP-MS) analysis showed that no copper species leached into the supernate. These results demonstrate the heterogeneous nature of the catalyst. To determine the catalyst recyclability, the carboxylation of 1-ethynylbenzene with CO2 catalyzed by CZU-7 was performed under the same conditions as described above except for the use of the recovered catalyst. As shown in Fig. 2\n, the recovered catalyst could maintain its good activity after five consecutive cycles. PXRD patterns (Fig. S9) and FT-IR spectra (Fig. S10) of the recovered samples are almost identical to those of the as-synthesized one, preserving the structural integrity of the material during catalysis. To determine the oxidation state of copper before and after reactions, XPS experiments of CZU-7 have been carried out (Fig. S11). The binding energy of Cu2p3/2 at 933.41 and Cu2p1/2 at 953.24\u00a0eV for the sample is assigned to the characteristics of the Cu\n\u03b4+ (\u03b4\u00a0<\u00a02) species [27]. The observation of the Cu 2p peaks demonstrated no change in the oxidization state of copper, further implying the CZU-7 catalyst is stable during the reaction processes. Moreover, the XPSPEAK fitting results of copper species of CZU-7 revealed that simulated peaks for various Cu forms were fit to the measured peak areas. It was observed that the ground complex catalyst with a smaller particle size (100\u2013300 mesh) could accelerate the rate of the carboxylation. Because the surface area of the catalyst is very small, this reaction should be catalyzed by the active sites on the surface of the catalyst.On the basis of the above results and the previously suggested reaction mechanisms involving the Cu(II)-based complexes in alkyne carboxylation reactions [15,24,34], we propose a plausible mechanism shown in Scheme 1\n. First, terminal alkyne coordinates to the copper(II) center of the Cu(II) CP, enhancing the acidity of the alkyne CH bond. Then, deprotonation reaction of terminal alkyne by Cs2CO3 affords the Cu(II) acetylide intermediate A. Meanwhile, the insertion of CO2 into sp-hybridized carbon\u2013copper bond leads to the formation of the Cu(II) propiolate intermediate B, which subsequently reacts with another terminal alkyne and Cs2CO3 releasing cesium propiolate, simultaneously regenerating Cu(II) acetylide intermediate A. At the end of the reaction, the acidification of cesium propiolate affords the propiolic acid product.In summary, a new stable Cu(II)-based MOF CZU-7 with the 44-sql topological structure was synthesized via a facile and fast microwave heating method. The as-synthesized CZU-7 catalyst could serve as an efficient and recyclable catalyst for the direct carboxylation of terminal alkyne with CO2. This work highlights the feasibility of using Cu(II)-based MOF materials as heterogeneous catalysts for CO2 conversion.\nZhong-Hua Sun: Investigation, Data curation, Writing \u2013 original draft. Xin-Yan Wang: Methodology, Validation. Kun-Lin Huang: Software, Visualization. Ming-Yang He: Visualization, Resources. Sheng-Chun Chen: Investigation, Writing \u2013 review & editing, Conceptualization, Funding acquisition.We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled, \u201cHeterogeneous catalytic carboxylation of terminal alkynes with CO2 over a copper(II)-based metal-organic framework catalyst\u201d.We gratefully acknowledge financial support by the National Natural Science Foundation of China (21676030), and Jiangsu Province Prospective Industry-University-Research Cooperative Research Program of China (NO. BY2016029-08).\n\n\n\nSupplementary material 1\n\nImage 2\n\n\n\n\n\n\nSupplementary material 2\n\nImage 3\n\n\n\n\n\n\nSupplementary material 3\n\nImage 4\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106472.", "descript": "\n                  Developing efficient, inexpensive and robust heterogeneous catalysts for CO2 conversion is greatly important. In this study, a new stable copper(II)-based metal-organic framework [Cu(Fbtx)2Br2]\n                        n\n                      (denoted as CZU-7, CCDC-2165700, Fbtx\u00a0=\u00a01,4-bis(1,2,4-triazole-1-ylmethyl)-2,3,5,6-tetrafluorobenzene) was synthesized via a facile and fast microwave heating method. CZU-7 exhibits a two-dimensional layer structure with the 44-sql topology. This material was demonstrated to be an efficient heterogeneous catalyst for the direct carboxylation of 1-ethynylbenzene with CO2, and various propiolic acids were synthesized in moderate to good yields under optimized reaction conditions. Moreover, the catalyst could be easily recovered and reused five times without significant loss of catalytic activity.\n               "}
{"full_text": "Data will be made available on request.Rising environmental issues and diminishing resources caused by increase in the usage of fossil fuels encourage researchers to find alternative sources of energy which is renewable, sustainable, and eco-friendly [1\u20134]. Biomass has significant potential as an alternative energy source because of its high energy density and minimal environmental impact; however, raw bio-oils derived from biomass contain a mixture of organic compounds with fatty acids as one of the main components, resulting in large amount of oxygen (30\u201340\u00a0wt%). This results in high viscosity, low vapor pressure, high corrosiveness, and low stability of these bio-oils which limit their direct application as transportation fuels [5\u20138]. Thus, improvement of the quality of bio-oils are indispensable to satisfy the standard requirements of transportation fuels.The fatty acids in the bio-oils can be converted into liquid hydrocarbons by several methods, such as decarboxylation (DCO) [9\u201311], hydrodecarbonylation (HDC) [12,13], and hydrodeoxygenation (HDO) [14,15]. The ultimate aim of these methods is to remove the oxygen content in the bio-oils. In DCO, the fatty acid is converted by removing the carboxyl group resulting in CO2 as the side product, while in HDC, the fatty acid is initially converted into aldehyde before CC bond cleavage occurs to produce CO as the side product. The difference between these methods lies in the reaction environment. Under inert conditions, decarboxylation mainly takes place over decarbonylation. On the other hand, decarbonylation dominantly occurs over decarboxylation in H2 environment [16,17]. Both reactions generally requires high temperature because of the stability of the CC bond. Due to the removal of CO2 or CO in these methods, the main product is a hydrocarbon with a less carbon atom, which is undesirable in terms of atom economy. Meanwhile, HDO converts fatty acids by using H2 and remove O by forming water in the process, involving series of hydrogenation and dehydration reactions. Hence, compared to DCO and HDC, HDO generally requires higher H2 pressure to occur, depending on the type of feed [10,18]. Nevertheless, HDO is more favorable than the other methods because it maintains the carbon efficiency and produces more environmentally friendly byproduct in the form of water [19\u201321].The HDO of fatty acids is challenging due to low reactivity of the carboxylic group [22]; therefore, efforts have been made using various catalysts to improve the reactivity. Noble metal catalysts (Pt, Pd, Ru, Rh) are generally used because of their high deoxygenation activity even at low temperatures and hydrogen pressures [23]. Meanwhile, non-noble metal catalysts such as Ni and Co have also been explored [24\u201327]. Although their deoxygenation activity is not as high as the noble metal catalysts at similar loading level, they are more affordable [28]. Nevertheless, both catalysts are more likely to remove oxygen through DCO or HDC mechanism because of their great affinity towards \u03b71 (C)-acyl configuration [29\u201332]. For example, Ni/ZSM-5 was used in the conversion of palmitic acid, and favored the formation of pentadecane over hexadecane, highlighting the preference of DCO or HDC mechanisms over HDO [33]. Therefore, modifications of the catalysts have been conducted to alter the deoxygenation pathway. For example, Phan et al. [34] modified Pt/UiO-66 using amine groups for deoxygenation of stearic acid to n-alkanes. The defective-UiO-66 improved the HDO activity significantly due to the increased number of accessible Pt sites, acidity, and basicity of the catalysts. Moreover, defects induced by the amine groups allowed H-transfer for HDO because of enhanced H2 adsorption on the support. Kon et al. [35] loaded MoOx on Pt/TiO2 to modify interaction between Pt and the Lewis acid sites formed by MoOx/TiO2, which led to a synergistic effect on lauric acid HDO to dodecane with full conversion and 86\u00a0% dodecane yield at 180\u00a0\u00b0C. For non-noble metals, Mo has been widely combined with Ni or Co in the traditional HDO catalysts (NiMo/Al2O3 or CoMo/Al2O3) in reduced or sulfided form to modify the carboxylic acid-catalyst interaction [36,37]. Other promoters, such as WOx\n[38], ReOx\n[39,40], and Nb2O5\n[41,42] have also been used to increase Lewis acid sites for activation of the CO group of fatty acids, thereby preventing HDC/DCO reaction.HDO of fatty acids over modified noble metal catalysts generally results in the formation of alcohols [40,43]. To obtain hydrocarbons as the final product, alcohols must be dehydrated over acidic supports, such as Al2O3\n[44,45] and zeolites [46\u201348]. The dehydration of alcohols leads to formation of alkenes, which can be hydrogenated rapidly over the noble metal sites. This means that the HDO of fatty acids to hydrocarbons over noble-metal based catalysts requires a promoter to control the activity of noble metals and an acidic support to convert alcohols into the final product. These reactions can be seen as a cascade reaction.Ruthenium, one of the noble metals, has an excellent hydrogenation activity similar to Pt and Pd [22,49]; thus, the modification of Ru with promoters is important to regulate the activity of Ru. Addition of Sn [2,50], B [51], or Mo [52] could improve the HDO activity due to the changed electron density of Ru resulting from its interaction with the promoter. For Sn, the interaction of Ru and Sn can result in two different active sites, Ru3Sn7 and Ru-SnOx, which depend on the support used. Both active sites can catalyze the reduction of fatty acids to alcohols, although the mechanism is slightly different. The Ru3Sn7 alloy catalyzes reduction of fatty acids to alcohols through the changed electron density of both Ru and Sn, resulting in higher hydrogenation activity of Ru and higher CO adsorption capability of Sn [50]. On the other hand, the Ru-SnOx reduces fatty acids to alcohols through CO activation on Lewis acid sites formed by SnOx and hydrogenation on Ru [53,54]. Comparing both active sites, the reduction activity is higher over Ru3Sn7 alloy than Ru-SnOx because of lesser side reactions. On Ru-SnOx, the presence of isolated Ru may cause side reactions of HDC or DCO of fatty acids, while the presence of isolated SnOx may catalyze esterification reaction [53]. Therefore, Ru3Sn7 is more preferred active site for the HDO of fatty acids.As aforementioned, in order to achieve full HDO of fatty acids to hydrocarbons, specific sites for dehydration must also exist on the catalyst. Using Al2O3 as support for RuSn catalysts may result in high dehydration activity but formation of the Ru3Sn7 alloy phase may be hindered because of the strong metal-support interaction between Sn and Al2O3 as previously reported [55\u201357]. Thus, designing supports with low metal-support interaction to enable the formation of Ru3Sn7 alloy without suppressing their dehydration activity is highly desirable, and is the main focus of this study.Herein, we report for the first time the HDO of octanoic acid to C8 hydrocarbons with high yields over RuSn catalysts with Ru3Sn7 as the main active site on SiO2-doped Al2O3 (SiAl) support. The deposition of SiO2 enables the formation of Ru3Sn7 on Al2O3 support, minimizing side reactions. The effect of temperature, H2 partial pressure, contact time, and SiO2 loading were investigated to understand the important factors and derive plausible mechanism for the reaction.The commercial supports used in this study were SiO2 (hydrophilic fumed silica, CAB-O-SIL M\u22125) obtained from Cabot Co. Ltd. (USA) and \u03b3-Al2O3 from Alfa Aesar Co. Ltd. (USA). The chemicals used in the catalyst synthesis were RuCl3\u00b7xH2O (99.9\u00a0%), SnCl4\u00b75H2O, (98.0\u00a0%), ethanol (>99.5\u00a0%), ammonium nitrate (NH4NO3, 99.999\u00a0%), polyvinylpyrrolidone (PVP, K-30), sodium borohydride (NaBH4, 99\u00a0%), all purchased from Sigma Aldrich Co. Ltd. (USA), tetraethylorthosilicate (TEOS, 99\u00a0%), obtained from Alfa Aesar Co. Ltd. (USA), and hydrochloric acid (HCl, 37\u00a0%), obtained from Samchun Chemicals Co. Ltd. (South Korea). The chemical used in the reactivity studies was octanoic acid (>98\u00a0%), obtained from Sigma Aldrich Co. Ltd. (USA). The standards used for quantitative analysis were octanal (99\u00a0%), dioctyl ether (99\u00a0%), 1-octene (99\u00a0%), trans-2-octene (97\u00a0%), all obtained from Sigma Aldrich Co. Ltd. (USA), trans-3-octene (97\u00a0%), from Alfa Aesar Co. Ltd. (USA), and octanol (99\u00a0%), from Acros Organics Co. Ltd. (USA). The solvent used for gas chromatography (GC) was acetone, purchased from Daejung Chemicals Co. Ltd. (South Korea).The SiAl supports were synthesized via a liquid-phase deposition method as reported elsewhere [58]. Briefly, a solution containing TEOS and ethanol (0.693\u00a0g of TEOS corresponds to 200\u00a0mg of SiO2) was mixed with 1\u00a0g of Al2O3. Hydrolysis of TEOS was conducted by adding 0.13\u00a0g NH4NO3 to the solution with continuous stirring at 40\u00a0\u00b0C for 168\u00a0h. The sample was then collected by filtration and dried at 100\u00a0\u00b0C for 24\u00a0h. The sample was calcined at 500\u00a0\u00b0C for 4\u00a0h with a ramping rate of 1\u00a0\u00b0C/min. The synthesized supports were denoted SiAl(x:y), where\u00a0x\u00a0and y are the nominal loadings in wt% of SiO2 and Al2O3, respectively. For example, SiAl(1:9) indicates that the nominal loading of SiO2 in the support is 10\u00a0wt%, while that of Al2O3 is 90\u00a0wt%.The RuSn catalysts were synthesized by a sol-immobilization method following a reported procedure [59]. For 10\u00a0g of support, a certain amount of RuCl3\u00b7xH2O and SnCl4\u00b75H2O were mixed in 200\u00a0mL of DI water to obtain a Ru loading of 1.4\u00a0wt% and Sn/Ru molar ratio of 2 calculated based on previous study [60]. Then, 3.6\u00a0g of PVP was added to the solution, which was then cooled to near 0\u00a0\u00b0C in an ice bath under continuous stirring. An aqueous solution of 0.1\u00a0M NaBH4 (NaBH4/metal mol ratio\u00a0=\u00a05) was added rapidly to the mixture. After 30\u00a0min, 10\u00a0g of finely ground support (SiO2, Al2O3 and SiAl(x:y)) was then added to the solution and 0.1\u00a0M HCl was added to decrease the pH of the solution to 1\u20133. The solution was stirred overnight at room temperature. The solid suspension in the solution was separated by centrifugation at 3000\u00a0rpm for 10\u00a0min, and the recovered solid was washed with distilled water five times to remove unreacted precursors, and dried then for 12\u00a0h at 100\u00a0\u00b0C. The catalysts were then pelletized and activated in the reactor at 460\u00a0\u00b0C for 4\u00a0h in H2 atmosphere with a ramping rate of 5\u00a0\u00b0C/min prior to reaction.The surface area of the catalysts and supports were acquired using an ASAP 2420 apparatus (Micromeritics, USA). The samples were first degassed for 30\u00a0min at 90\u00a0\u00b0C, followed by heating at a constant temperature of 300\u00a0\u00b0C for 4\u00a0h. The bulk phase of the catalysts was analyzed by powder X-ray diffraction (PXRD) using a Rigaku Ultima IV X-ray diffractometer (Japan) with a Cu K\u03b1 radiation beam (\u03bb\u00a0=\u00a00.15418\u00a0nm) operated at 45\u00a0kV and 200\u00a0mA. The measurements were performed within a scanning range of 10 to 80\u00b0 at a scan rate of 2\u00b0/min and step size of 0.02\u00b0.X-ray absorption spectroscopy (XAS) was carried out to analyze the local atomic structure of the catalysts at Ru K-edge (22.117\u00a0keV). The analysis was performed at the 7D beamline located at the Pohang Accelerator Laboratory (PLS-II, 3\u00a0GeV, South Korea). The spectra of the samples were recorded at room temperature in transmission mode. Ru foil and commercial RuO were used as standards for the calibration. The obtained data were analyzed using the Athena and Artemis programs. For the extended X-ray absorption fine structure (EXAFS) analysis, the Fourier-transforms of the background-subtracted data were obtained using a Hanning window in the k range of 3\u201312 A with a dk of 1 A\u22121.Temperature-programmed desorption (TPD) measurements were conducted to analyze the acidity of the catalysts and the supports using a Micromeritics AutoChem II 2920 instrument (USA) equipped with a thermal conductivity detector (TCD). The catalysts were reduced at 460\u00a0\u00b0C for 4\u00a0h in H2/Ar. After reduction, the temperature was reduced to 100\u00a0\u00b0C under pure He, and then 5\u00a0%NH3/He was introduced to the sample for 1\u00a0h. The NH3 desorption was then carried out at elevated temperatures from 100 to 800\u00a0\u00b0C with a ramping rate of 5\u00b0/min under He flow.The acid sites of the supports were also characterized by pyridine adsorbed infrared spectroscopy (IR) measurements using a JASCO FT/IR-4200 spectrometer equipped with a triglycerine sulfate (TGS) detector. The samples were pressed into self-supporting wafers with a diameter of 2\u00a0cm (ca. 30\u00a0mg) which was placed at the center of an infrared flow cell equipped with KBr windows maintained at 25\u00a0\u00b0C. The samples were pretreated at 460\u00a0\u00b0C in H2 for 1\u00a0h and cooled to room temperature under vacuum. Spectra were recorded in absorbance mode (128 scans, 4\u00a0cm\u22121 resolution) in the region of 4000\u20131000\u00a0cm\u22121. Subsequently, the samples were exposed to 0.5\u00a0kPa of flowing pyridine vapor for 15\u00a0min and evacuated at 150\u00a0\u00b0C for 15\u00a0min. The spectra were taken at room temperature after evacuation. The amounts of Br\u00f8nsted acid sites (BAS) and Lewis acid sites (LAS) were determined from the area of the peaks at ca. 1540 and 1445\u00a0cm\u22121, respectively. The molar adsorption coefficient for the band at 1540\u00a0cm\u22121 was 1.67\u00a0cm\u00a0\u03bcmol\u22121 and that for band at 1445\u00a0cm\u22121 was 2.2\u00a0cm\u00a0\u03bcmol\u22121\n[61].The Si, Al, and O surface chemical states of the synthesized SiAl supports were obtained by X-ray photoelectron spectroscopy (XPS) analysis with an Axis Supra equipment (Kratos Analytical Ltd., UK) equipped with a monochromatic Al K\u03b1 source (1486.7\u00a0eV). The Si and Al core-level spectra were recorded in the 2p region, while the O core-level spectra were recorded in the 1\u00a0s region. Before analysis, the catalysts were reduced at 460\u00a0\u00b0C for 4\u00a0h in H2/Ar and sealed in a glass vial.Particle size distributions in the catalysts were obtained by Field emission transmission electron microscopy (FE-TEM) analysis using a Tecnai F30 S-Twin (Thermo Fisher Scientific, USA) operated at 300\u00a0kV. The catalysts were ground into fine powder and reduced under H2/Ar at 460\u00a0\u00b0C for 4\u00a0h before analysis.The synthesized catalysts were tested for gas phase HDO of octanoic acid. The reaction was conducted in a continuous down flow fixed-bed reactor (SUS 316) with a reactor diameter of 0.75\u00a0cm and length of 40\u00a0cm. The reactor was equipped with heating jacket, in which the temperature can be controlled to 700\u00a0\u00b0C. The reactor was also equipped with a pre-heater and a line-heater that were set at 250\u00a0\u00b0C to ensure vaporization of the octanoic acid before entering the reactor. The H2 gas was mixed with the feed in the pre-heater zone. The flowrate of the gas was controlled using a mass flow controller (MFC, Brooks Instrument 5850E Series (USA)). The reactor was also equipped with a chiller operated at \u22125\u00a0\u00b0C to condense reaction products and separate them from the H2 gas.In a typical reaction test, 1\u00a0g of catalyst was placed inside the reactor on top of quartz wool. The catalyst volume was set at 3\u00a0mL by adding inert glass beads. The catalyst was then activated inside the reactor at 460\u00a0\u00b0C for 4\u00a0h in H2 with a ramping rate of 5\u00a0\u00b0C/min. The reactor was then cooled to the desired reaction temperature and pressurized with H2 to 20\u00a0atm using a back-pressure regulator located at the end of the reactor system. The H2 and vaporized octanoic acid feed were introduced to the reactor at a H2/octanoic acid molar ratio of 70.8. The weight hourly space velocity (WHSV) was set at 1\u00a0h\u22121, calculated using the following equation:\n\n(1)\n\n\nW\nH\nS\nV\n\n\n\n\nh\n\n-\n1\n\n\n\n\n=\n\n\noctanoic\n\nacid\n\nmass\n\nf\nlowrate\n(\n\ng\nh\n\n)\n\n\nAmount\n\nof\n\ncatalyst\n(\ng\n)\n\n\n\n\n\n\nThe effect of H2 partial pressure and contact time were investigated at 350\u00a0\u00b0C and 20\u00a0atm. For the study of the effect of H2 partial pressure, the total flowrate of gas was fixed at 180\u00a0mL/min while the gas composition was varied using inert N2 as balance. To check the effect of contact time, the catalyst amount was varied while the H2/octanoic acid feed molar ratio was kept at 70.8. The WHSV was varied from 1 to 40\u00a0h\u22121 which corresponds to contact time from 0.025 to 1\u00a0h, calculated according to the following equation:\n\n(2)\n\n\nC\no\nn\nt\na\nc\nt\n\nt\ni\nm\ne\n\n\n\nh\n\n\n=\n\n\nAmount\n\nof\n\ncatalyst\n(\ng\n)\n\n\noctanoic\n\nacid\n\nmass\n\nf\nlowrate\n(\n\ng\nh\n\n)\n\n\n\n\n\n\nThe reaction products were collected every hour and analyzed using an off-line gas chromatograph (GC, Younglin Chromass 6500 GC, South Korea) equipped with a flame ionization detector (FID), an autosampler (Younglin YL3000A, South Korea), and a DB-624 column (30\u00a0m\u00a0\u00d7\u00a00.32\u00a0mm\u00a0\u00d7\u00a01.8\u00a0\u03bcm, Agilent Technologies, USA). The measurements were taken after steady state conditions were achieved, and all the presented data points were obtained by averaging three measurements after stabilization. In all cases, the mass balance reached 98\u00b12\u00a0%. The reactants and products were quantified by their column retention time in comparison with standards. Details of the analysis procedure are given in the Supplementary Information (SI).Catalytic activity was expressed by conversion of octanoic acid (Eq. (3)) and selectivity to the products (Eq. (4)), defined as follows:\n\n(3)\n\n\nC\no\nn\nv\ne\nr\ns\ni\no\nn\n\n\n\n\n%\n\n\n\n=\n\n\n\n1\n-\n\n\nn\n\nOA\n,\n1\n\n\n\nn\n\nOA\n,\n0\n\n\n\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n(4)\n\n\nP\nr\no\nd\nu\nc\nt\n\ns\ne\nl\ne\nc\nt\ni\nv\ni\nt\ny\n\n\n\n\n%\n\n\n\n=\n\n\n\n\nv\ni\n\n\nn\ni\n\n\n\n\u2211\n\nv\ni\n\n\nn\ni\n\n\n\n\u00d7\n100\n\n\n\n\nIn these equations, \u03c5 is the carbon number of species i (8 for 1-octanol, octanal, 1-octene, and i-octene; 16 for dioctyl ether and octyl octanoate), \u03b7i,1 is the number of moles of the product and \u03b7i,0 is the initial number of moles of the compounds.The physicochemical properties of the catalysts and the supports are summarized in Table 1\n. The BET surface area of the supports ranged from 185 to 251\u00a0m2/g and did not change significantly after deposition of the metals, indicating no substantial modification of textural properties of the supports.The bulk and crystal structures of fresh and spent catalysts were identified by PXRD shown in Fig. 1\na and Fig. S1, respectively. For RuSn/SiO2, a broad feature can be observed around 2\u03b8 of 20\u00b0, corresponding to the amorphous SiO2 support. Meanwhile, for the other catalysts, peaks observed at 2\u03b8 of 39.7\u00b0, 46.0\u00b0, and 66.7\u00b0 (depicted as diamond symbol) correspond to \u03b3-Al2O3. In addition to the support peaks, discernible features can be observed at 2\u03b8 of 30.1\u00b0, 32.9\u00b0, 35.8\u00b0, 40.9\u00b0, and 43.1\u00b0, which match well with the (310), (222), (321), (411), and (420) planes of Ru3Sn7 alloy with cubic crystal structure (PDF No. 26\u2013504). This indicates that the crystalline phase of Ru species in the catalysts was mainly RuSn alloy phase. However, this phase was not observed in RuSn/Al2O3, possibly due to strong interaction of Sn with Al2O3, which stabilized Sn in an oxidation state close to\u00a0+2 and hampered the reduction of Sn and interaction with Ru [55]. The crystallite sizes of the Ru3Sn7 alloy were calculated from the peak broadening of the (411) plane by the Scherrer equation and the results are reported in Table 1. The crystallite size of the fresh catalysts increased in the order, RuSn/SiAl (2:8) (3.5\u00a0nm)\u00a0<\u00a0RuSn/SiAl (3:7) (4.2\u00a0nm)\u00a0<\u00a0RuSn/SiAl (1:9) (5.5\u00a0nm)\u00a0<\u00a0RuSn/SiO2 (9.4\u00a0nm). The crystallite sizes of the spent catalysts increased (Table 1 and Fig. S1) which indicate possible agglomeration of metal sites after reaction.The local atomic structure of the catalysts was examined using X-ray absorption spectroscopy (XAS) at the Ru K-edge as shown in Fig. 1b and c. Fig. 1b shows the normalized Ru K-edge XANES spectra of the catalysts and references with absorption edge energy, E0, shown as vertical lines. Depending on the charge distribution of the observed atom, the E0 might shift to higher or lower energy because of changing chemical environment. The XANES spectra of RuSn/Al2O3 show a similar E0 energy to that of the Ru foil at 22.116\u00a0eV, indicating that the charge distribution in RuSn/Al2O3 is similar to the Ru foil. This suggests that there is almost no interaction between Ru and Sn in RuSn/Al2O3. In contrast, the XANES spectra of RuSn/SiO2 and RuSn/SiAl (3:7) show that the E0 shifted to lower energies (E0\u00a0=\u00a022.110\u00a0keV). This means that the presence of Sn in these catalysts modified the local chemical environment of Ru by changing the electron density of Ru through electron transfer from Sn to Ru. The direction of electron transfer is consistent with Pauling electronegativity differences, where Ru has larger Pauling electronegativity (2.2) than Sn (1.96). Unlike RuSn/Al2O3, the interaction of Ru-Sn in these catalysts gives clear evidence for alloy formation.The Fourier-transform EXAFS spectra (Fig. 1c) show the interaction of Ru atom with other atoms based on radial distances. Peaks corresponding to Ru-O and Ru-O-Ru which are located at radial distances of 1.59 and 2.73\u00a0\u00c5 were not observed in all the catalysts, suggesting that all catalysts have Ru in the reduced form. For RuSn/Al2O3, an intense peak was observed at a radial distance of 2.34\u00a0\u00c5, corresponding to the Ru-Ru bond length. On the other hand, for RuSn/SiO2 a peak located at a higher radial distance of 2.53\u00a0\u00c5 was observed, attributed to the Ru-Sn bond length [62,63]. Interestingly, the RuSn/SiAl (3:7) shows a similar peak location as RuSn/SiO2, indicating that Ru and Sn interacts strongly in these catalysts, further confirming the result obtained by XRD. Thus, the results provide clear evidence that Ru-Sn interaction exists in both RuSn/SiO2 and RuSn/SiAl (3:7), while only Ru0 phase is present in RuSn/Al2O3.The strength of acid sites on the supports and catalysts were determined by NH3-TPD. Desorption profiles are shown in Fig. 1d-1 for supports and Fig. 1d-2 for the corresponding supported RuSn catalysts. Generally, peaks below 200\u00a0\u00b0C, between 200 and 400\u00a0\u00b0C, and above 400\u00a0\u00b0C are considered weak, medium, and strong acid sites, respectively. The Al2O3 support showed three desorption peaks located at 166, 295, and 406\u00a0\u00b0C (Fig. 1d-1), which indicates that Al2O3 has three different acid sites that correspond to weak, medium and strong sites [64,65]. The strength of Lewis acidity depends on the coordination of the Al and O atoms, in which the lower coordination number brings stronger acidity [66,67]. With the addition of 30\u00a0% SiO2, the peaks that was initially located at 295 and 406\u00a0\u00b0C shifted to higher temperatures of 313 and 697\u00a0\u00b0C, meaning that strong acid sites was introduced with deposition of SiO2. The increased acid strength could be attributed to the formation of Br\u00f8nsted acid sites through the interaction of SiO2 and Al2O3 in the form of isolated silanols (SiOH) anchored on the surface of Al2O3 or the bulk transformation of Al2O3 to SiO2-Al2O3 as reported in several literatures [68\u201370].When Ru and Sn were added to the supports, different NH3 desorption profiles can be observed (Fig. 1d-2). On Al2O3, the medium and strong acid strengths increased significantly as indicated by the shifting of the peaks at 166 to 170, and 406 to 538\u00a0\u00b0C. This indicates that the addition of metals could increase the acidity as well. This could be achieved by forming new acid sites such as SnOx due to its strong interaction with Al2O3\n[71]. For RuSn/SiAl (3:7) catalyst, the addition of metals slightly shifted the medium acid sites from 313 to 297\u00a0\u00b0C while two strong acid peaks appeared at 604 and 684\u00a0\u00b0C. The shifting of medium acid sites could be attributed to the deposition of metal on the support, while the strong acid sites became enhanced caused by the formation of SnOx similar to the RuSn/Al2O3.Further verification of the acid sites on the supports was carried out by pyridine FTIR as shown in Fig. S2 for pyridine adsorption and Fig. 1e for the respective OH region. The pyridine adsorption spectra of SiO2 (Fig. S2) showed no bending vibrations, suggesting that SiO2 has negligible acidity. For Al2O3, bending vibrations were observed at 1450 and 1493\u00a0cm\u22121, corresponding to pyridine adsorbed on Lewis acid sites and on both of Lewis and Br\u00f8nsted acid sites, respectively [72,73]. The bending vibration at 1540\u20131550\u00a0cm\u22121 that is characteristic for the Br\u00f8nsted acid sites was not observed, indicating that acidity of Al2O3 is dominated by Lewis acid sites. With the addition of SiO2, the adsorption band intensity at 1450\u00a0cm\u22121 decreased significantly. This indicates that SiO2 blocked the Lewis acid sites on Al2O3. Although there was no adsorption band of pyridine related to Br\u00f8nsted acid sites (1540\u00a0cm\u22121), the adsorption band at 1493\u00a0cm\u22121 could be an indicator of the presence of Br\u00f8nsted acid sites by comparison of the absorbance intensity at that wavelength between the supports. With the addition of 10\u00a0% SiO2, the absorbance intensity at 1493\u00a0cm\u22121 was similar to that for the pure Al2O3; however, the intensity at 1450\u00a0cm\u22121 for the former was slightly lower, suggesting that the comparable intensity at 1493\u00a0cm\u22121 was caused by the presence of Br\u00f8nsted acid sites. A further increase in the SiO2 loading decreased the absorbance intensity at 1493\u00a0cm\u22121, which is due to decrease in Lewis acid sites. Comparing the absorbance intensity between SiAl (2:8) and (3:7), the absorbance was higher for SiAl (3:7) at 1493\u00a0cm\u22121 while the opposite trend was observed at 1450\u00a0cm\u22121, suggesting that the decreasing Lewis acid sites is counterbalanced by the newly formed Br\u00f8nsted acid sites. In addition, the vibration band at around 1550\u00a0cm\u22121 that belongs to Br\u00f8nsted acid sites started to appear on the SiAl (3:7) support. Correlating with the NH3-TPD results (Fig. 1d) reveals that the appearance of stronger acid sites for SiAl support was caused by the appearance of the Br\u00f8nsted acid sites.As aforementioned, the presence of Br\u00f8nsted acid sites on SiAl supports was due to the formation of silanol species when SiO2 was deposited on Al2O3; however, there are two possible forms of silanol species on Al2O3. The first one is by the bulk transformation of Al2O3, in which Si atoms replaced Al atoms in the lattice to form Si-OH-Al, and the other one is by the formation of silanol species anchored on the top of Al2O3. To determine which of the two silanol species exists in the SiAl supports, the FTIR spectra of the evacuated supports in the OH stretching region were recorded and shown in Fig. 1e. The SiO2 showed one vibration band at 3745\u00a0cm\u22121 attributed to Si-OH species. For Al2O3, the bands at 3768, 3728, 3680, and 3590\u00a0cm\u22121 were assigned to the vibration of terminal OH over one tetrahedrally-coordinated Al ion, terminal OH over an octahedrally-coordinated Al ion, bridging OH, and triply-bridging OH groups, respectively [74,75]. As SiO2 loading increased, a new vibration band at 3738\u00a0cm\u22121 appeared. This vibration band is assigned to the isolated silanols over Al2O3 support [70]. The lower frequency of this band compared to the one in parent SiO2 (3745\u00a0cm\u22121) indicates that the silanols have greater OH bond ionicity, resulting in stronger Br\u00f8nsted acid sites [69]. This is consistent with the earlier interpretation that the increased acid strength observed by NH3-TPD is related to the presence of isolated silanols over Al2O3. Additionally, the peak at 3610\u00a0cm\u22121 assigned to OH groups of Si-OH-Al was not observed in all the SiO2-containing supports. This suggests that the formation of bulk Si-OH-Al species did not occur, but rather the deposited SiO2 was formed on top of the Al2O3 support, clarifying that only surface modification of Al2O3 occurred with SiO2 deposition.The surface states of the SiAl supports at Si 2p and Al 2p regions were determined by XPS and reported in Fig. 1f and Fig. S3. The adventitious C 1s carbon peak was used as the reference peak at 284.8\u00a0eV. The Si 2p spectra of SiO2 showed a very distinct peak at 103.7\u00a0eV, a typical peak of the Si4+ of SiO2\n[76,77]. The addition of SiO2 on Al2O3 resulted in a shifted spectra to lower binding energy as compared to the pure SiO2 and the shifts became even larger as SiO2 loading increased, indicating a decrease in the Si oxidation state from\u00a0+4 to between 0<\u00a0x\u00a0<4. The reduction of Si oxidation state in SiAl supports means that Si is interacting with other species in the form of Si-O-M, which is typically observed in aluminosilicates, where M is not Si but most likely Al [69]. The increase in the SiO2 loading led to increase in the number of SiO2 interacting with Al2O3 species, resulting in larger shifts in the binding energy. Meanwhile, the Al 2p spectra shows a similar peak location for all samples at 74.2\u00a0eV (Fig. S3). However, peak deconvolution reveals two distinct contributions at 75.2\u00a0eV that belongs to AlOH species and at 74.2\u00a0eV that corresponds to Al-O species [77\u201379]. The peaks at 75.2\u00a0eV did not change significantly with increasing loading of SiO2, while that at 74.2\u00a0eV shifted to lower binding energy with increasing SiO2 loading. This indicates that the deposition of SiO2 modified Al-O interaction, and gives additional evidence for interaction between Si, O and Al species. The interaction of Si with Al possibly induced the formation of terminal silanols which are anchored on Al2O3 and are likely the main reason for the increased acidity in SiAl supports, as confirmed previously by NH3-TPD and pyridine FTIR results.The structural images of the catalysts were captured by FE-TEM and are shown in Fig. 2\n and Fig. S4. It can be observed that metal particles were distributed evenly in all catalysts. In all the catalysts, except RuSn/Al2O3, the lattice spacing of the metal particles (identified by yellow lines) was calculated to be 0.22\u00a0nm, which corresponds well with the lattice spacing of the Ru3Sn7 alloy (411) plane. On the other hand, the lattice spacing for RuSn/Al2O3 was calculated to be 0.21\u00a0nm, matching well with the lattice spacing of the Ru0 (111) plane. These results provide further evidence that RuSn alloys were not formed on RuSn/Al2O3 in accordance with the XRD result. Fig. 2 also shows the size distributions of the catalysts and the average particle size determined from 100 particles for each sample observed in the FE-TEM images. It can be seen that the average metal particle size decreased with the presence of SiO2; furthermore, the metal particles became more narrowly distributed. This may be related to strong interactions between Sn and Ru in the alloys which minimized agglomeration of the Ru species [56,80]. Thus, it can be inferred that the presence of SiO2 on Al2O3 modified the interaction between the metals and the supports which was beneficial to the formation of the RuSn alloy phase in the SiAl support compared to the pure Al2O3 support.The reactivity of the catalysts in the HDO of octanoic acid was tested at a total pressure of 20\u00a0atm, H2/feed molar ratio of 70.6, temperature of 300\u2013400\u00a0\u00b0C, and WHSV of 1\u00a0h\u22121. Fig. 3\n and Table S1 show the conversion and product selectivity as a function of temperature over all the catalysts. The conversion of octanoic acid exceeded 95\u00a0% even at the lowest temperature (300\u00a0\u00b0C) on all the catalysts except on RuSn/Al2O3, in which the conversion reached 81\u00a0% at highest temperature (400\u00a0\u00b0C). This indicates that all the catalysts except RuSn/Al2O3 possessed a very high activity in the applied reaction conditions. This is likely due to the existence of the Ru3Sn7 alloy phase that can selectively adsorb CO over the Sn sites [2,50]. Contrastingly, the lower activity of RuSn/Al2O3 is attributed to the presence of a different active phase of Ru-SnOx as indicated by the characterization results. Hence, it is clearly proven that the Ru3Sn7 alloy is a key active phase for hydrogenation of CO bond.\nFig. 3 confirms that using different supports resulted in significant changes in the product distribution. For RuSn/SiO2, the main product was octanol at all temperatures, followed by octanal and octyl octanoate. At 300\u00a0\u00b0C, the selectivity to octanol was 74.6\u00a0%, before gradually decreasing to 38.7\u00a0% with the increase in temperature due to the enhanced formation of lighter C7 products (n-heptane and i-heptenes). For RuSn/SiAl (1:9), C8 hydrocarbons consisting of paraffin (n-octane) and olefins (1-octene, 2-octene, 3-octene, and other i-octenes) were the main products at all temperature regions. When the SiO2 loading increased to 30\u00a0%, the selectivity to n-octane also rose significantly and became the main product because of the different acidity possessed by the supports as will be explained further. Meanwhile, the RuSn/Al2O3 produced octyl octanoate and octanal as main C8 products while n-heptane and i-heptenes were also produced significantly as C7 products. Increase in temperature improved the selectivity to the C7 products to 46.6\u00a0%. It should be noted that the reaction over RuSn/Al2O3 was conducted from 350\u00a0\u00b0C because of severe clogging of the reactor at lower reaction temperature possibly due to the extensive formation of octyl octanoate.The products shown in Fig. 3 can be distinguished into hydrocarbons (n-octane, i-octenes, n-heptane, and i-heptenes) and oxygenates (octanal, octanol, dioctyl ether, and octyl octanoate). Fig. 4\n shows the distribution of these products and the ratios between them at 350\u00a0\u00b0C to assess the activity differences of the catalysts. As can be seen in Fig. 3 and Fig. 4(a), the selectivity to C7 products was the highest on RuSn/Al2O3 with a total selectivity of 29\u00a0%, indicating that RuSn/Al2O3 showed a decent HDC/DCO activity due to the aforementioned active phase of Ru-SnOx. This implies that isolated Ru species exist and is responsible for the HDC/DCO activity due to their oxophilic property to adsorb acid on the surface.[81\u201383] In contrast, the other catalysts showed minimum production of C7 compounds, meaning that the Ru3Sn7 alloy phase minimizes HDC/DCO activity and preserves the carbon number of the products. The selectivity to C8 hydrocarbons (Fig. 4a) can be divided into two categories: favoring n-octane or i-octenes. The selectivity to both products differed significantly depending on the supports used. The RuSn/SiAl (1:9) catalyst favored the production of i-octenes with selectivity of 64\u00a0%, while the RuSn/SiAl (3:7) catalyst favored the formation of n-octane with selectivity of 82.7\u00a0%. In contrast, the proportion of C8 hydrocarbons on RuSn/SiO2 and RuSn/Al2O3 was very low.In terms of the oxygenates (Fig. 4b), RuSn/SiO2 favored the production of octanol, octanal, and octyl octanoate with selectivities of 66.0, 13.7, and 13.8\u00a0%, respectively. With the addition of SiO2 on Al2O3 support, the production of these oxygenates decreased, meaning that octanol was immediately converted into i-octenes or n-octane. On RuSn/Al2O3, formation of octanal and octyl octanoate could be clearly observed, suggesting a lower activity and propensity for side reactions such as esterification. It has been discussed previously that the acidity provided by SnOx and Al2O3 could increase esterification activity [2,84].\nFig. 4(c) shows the product selectivity ratio between n-octane and i-octenes, as well as the hydrocarbons/oxygenates ratio at 350\u00a0\u00b0C. The n-octane/i-octenes ratio increased significantly with increasing SiO2 loading from 0.27 on RuSn/SiO2 to 5.27 on RuSn/SiAl (3:7). The n-octane/i-octenes ratio on RuSn/Al2O3 was zero, indicating no formation of n-octane. Meanwhile, the hydrocarbons/oxygenates ratio also increased significantly from 0.07 on RuSn/SiO2 to 299.83 on RuSn/SiAl (3:7), before decreasing to 0.55 on RuSn/Al2O3. The significant increase in n-octane/i-octenes ratio with increasing SiO2 loading suggests that hydrogenation of i-octenes occurred extensively, while the sharp increase in hydrocarbons/oxygenates ratio on RuSn/SiAl (3:7) indicates high octanol dehydration activity. These results suggest that the different acidity of the supports played a very important role in determining the final product distribution, while the presence of RuSn active phase ensures the high initial activity of octanoic acid conversion.To understand more clearly about the effect of temperature on product distribution over all the catalysts, C7 and C8 hydrocarbons yield as well as their ratios are summarized in Fig. 5\n. It can be observed that the overall C7 hydrocarbons yield increased with increasing temperature on all catalysts as a result of high energy input to break CC bonds in the reactant (Fig. 5a). The C7 hydrocarbons yield was much higher on the RuSn/Al2O3 and RuSn/SiO2 catalysts. For the C8 hydrocarbons (Fig. 5b), the increase in temperature from 300 to 400\u00a0\u00b0C also increased the yield on all the catalysts because of endothermic nature of dehydration. The highest yield was obtained on RuSn/SiAl (2:8) and (3:7) with comparable results, followed by RuSn/SiAl (1:9), RuSn/Al2O3 and RuSn/SiO2. It could be seen that the C8 hydrocarbons yield followed a curved trend with maximum around 350\u00a0\u00b0C for the RuSn/SiAl catalysts, and decreased slightly at higher temperature because of favorable formation of C7 hydrocarbons. This is more clearly visualized in Fig. 5c which compares the C7/C8 hydrocarbons ratio at different temperatures. At all temperatures, the C7/C8 hydrocarbons ratio was the highest on RuSn/Al2O3, followed by on RuSn/SiO2, decreasing with increasing temperature. The C7/C8 hydrocarbons ratio on the RuSn/SiAl catalysts was similar and increased with temperature.The distinctive behavior observed between the RuSn catalysts supported on pure Al2O3 and SiO2 and those dispersed on the SiAl supports hints at likely participation of the support. The results show that RuSn/Al2O3 prefers to produce C7 than C8 hydrocarbons, although it should be noted that C8 hydrocarbons were still formed mainly as 1-octene at 350\u00a0\u00b0C and a mixture of i-octenes at 400\u00a0\u00b0C (Fig. 3). The formation of i-octenes indicates that octanol was dehydrated over Al2O3 during high temperature reactions because of its endothermicity and high dehydration activity possessed by Al2O3\n[17,85]. The dehydration seemingly occurred by unimolecular dehydration mechanism that follows an E2 mechanism, in which an octanol is initially adsorbed on Al2O3 surface, resulting in deprotonation of C\u03b2-H bond and cleavage of C\u03b1-O bond [44,86]. However, the proportion of octenes was minimal in this case due to severe formation of side products, i.e., C7 hydrocarbons and octyl octanoate, caused by the low hydrogenation activity possessed by the active sites. For RuSn/SiO2, the results show preferential formation of C8 oxygenates. In spite of this tendency, C8 hydrocarbons were still observed at high temperature as indicated by the decrease in the C7/C8 ratio. Further inspection of the C8 products shows that the C8 hydrocarbons consist of n-octane and i-octenes mixture (Fig. 4c) caused by the direct deoxygenation of octanol over the metal sites.When SiO2 is deposited on Al2O3, the production of C8 hydrocarbons was favored instead of oxygenates, highlighting the high dehydration activity over the supports which increased with SiO2 loading. As observed in NH3-TPD (Fig. 1d) and pyridine FTIR (Fig. S2) results, the deposition of SiO2 resulted in the formation of Br\u00f8nsted acid sites in the form of terminal silanols that are highly active for the dehydration of alcohols according to the literature [87,88]. Thus, although the Lewis acid sites decreased with increasing SiO2 loading, the increasing amount of Br\u00f8nsted acid sites still resulted in high octanol dehydration activity because of the different dehydration activity possessed by Lewis and Br\u00f8nsted acid sites [88]. The dehydration over Br\u00f8nsted acid sites follows alcohol adsorption on the support and proton transfer from the silanol groups to the OH groups of alcohol, eliminating water in the process and forming alkoxy group as an intermediate. The alkoxy group then stabilizes by forming double bond, resulting in the formation of \u03b1-olefin [89].According to Fig. 3, the C8 hydrocarbons consisted of n-octane, 1-octene, 2-octene, 3-octene, and other i-octenes. Since the dehydration of octanol results in the formation of 1-octene, the formation of 2-octene, 3-octene, and other i-octenes suggests that isomerization occurred during the reaction. In general, isomerization can occur on either Lewis acid sites or Br\u00f8nsted acid sites. On Lewis acid sites, the isomerization occurs through hydration-dehydration mechanism, in which the formed 1-octene rehydrates to form 2-octanol and dehydrates again to form 2-octene [90]. On Br\u00f8nsted acid sites, the isomerization occurs through the migration of double bonds achieved by olefin protonation over the Br\u00f8nsted acid sites [91,92]. Regardless of the isomerization method, the final major product obtained on SiAl supports was n-octane, which comes from the over-hydrogenation of octenes. The increasing amount of n-octane with higher SiO2 loading suggests that Br\u00f8nsted acid sites highly affects the over-hydrogenation activity by strongly adsorbing octene molecules, resulting in higher retention time for over-hydrogenation to occur on the metal sites.The HDO performance of RuSn/SiAl catalyst was then compared to other reported catalysts as shown in Table S2. The comparison was conducted by calculating the hydrocarbons productivity, which is the production rate of hydrocarbons obtained through HDO mechanism without any carbon loss. In the case of octanoic acid, the HDO products would be the n-octane and i-octenes. As can be observed in Table S2, the RuSn/SiAl (3:7) showed better HDO performance compared to the other catalysts in the cited works, although the different reaction conditions should also be considered.The stability of the catalyst was tested over the course of 60\u00a0h using RuSn/SiAl (3:7) as the catalyst. The stability test was carried out at 350\u00a0\u00b0C, WHSV of 1\u00a0h\u22121, and pressure of 20\u00a0bar, and H2/feed of 70.8. It should be noted that the catalyst amount used in this test was 1/4 of the one used in Fig. 2, which explains the different product distribution. The long-term test result is shown in Fig. S5. Minimal change in the catalyst activity with no significant drop in the performance was observed during the reaction, suggesting the high stability of the catalyst although the metal sites marginally agglomerated during the reaction, as evidenced by the XRD of spent catalysts (Fig. S1), surface area result (Table S4), and TEM of the spent catalyst (Fig. S6). Generally, the agglomeration of metal sites results in the deterioration of catalyst activity, but this is not the case in this study. This could possibly mean that formation of alloy was enhanced during the reaction, resulting in the growing of the alloy metal sites since high H2 pressure and temperature with long reaction time were applied. A similar result was obtained by Karim et al. [93], where the formation of PdZn alloy was observed after reaction even without any pretreatment at 250\u00a0\u00b0C. It was postulated that the reduction of ZnO in the presence of Pd and H2 facilitated the alloying process during reaction. In addition, time-resolved XRD for the reduction of CuO observed by Rodriguez et al. [94] showed that the crystallinity of Cu0 depended on the reduction time of CuO, in which longer reduction time resulted in increased crystallinity. Oezaslan et al. [95] also observed similar result where Pt-Cu alloy crystallinity was enhanced after reducing at high temperature and prolonged time caused by metal insertion and particle growth. According to Luo et al. [50], the formation of Ru3Sn7 alloy started at temperature as low as 350\u00a0\u00b0C, so enhanced formation of Ru3Sn7 alloy and alloy particle growth during the reaction could be the reason for the high catalytic stability.In order to understand the effect of H2 on the hydrogenation activity of octanoic acid over RuSn/SiAl catalysts, the H2 partial pressure was varied between 3.9 and 19.7\u00a0atm with N2 as balance at 20\u00a0atm. Fig. 6\n shows the octanoic acid conversion and product selectivity as a function of H2 partial pressure. The octanoic acid conversion was insensitive to changing H2 partial pressure, hovering around 70\u00a0%. At low H2 partial pressure, the major product was octyl octanoate, followed by octanal, octanol, and i-octenes, with some minor products such as dioctyl ether and n-octane. As the H2 partial pressure increased, the selectivity to octyl octanoate increased, followed by decreasing octanol selectivity. The selectivity to i-octenes remained similar at all H2 partial pressure, meaning that the decreasing octanol selectivity at higher H2 partial pressure was due to the esterification of octanoic acid and octanol to form octyl octanoate, which most likely occurred on the acidic sites of the support. This also means that octanoic acid seemed to preferably adsorb on the support, while H2 molecules adsorbed on the Ru3Sn7 surface. At this condition, saturation of the Ru3Sn7 sites by H2 molecules occurred so further conversion of octyl octanoate either by hydrogenolysis or hydrogenation was suppressed, resulting in higher amount of octyl octanoate. Nevertheless, the conversion remained similar because of the bi-functionality of the catalyst that could convert octanoic acid by both hydrogenation and esterification.To understand the reaction pathway for the conversion of octanoic acid over RuSn/SiAl, a contact time study was carried out over RuSn/SiAl (1:9) at 350\u00a0\u00b0C, H2/Feed ratio of 70.8, and pressure of 20\u00a0atm. Fig. 7\n, Table S3, and Fig. S7 show the contact time analysis results in which the products are categorized into hydrocarbons (Fig. 7a) and oxygenates (Fig. 7b).The increase in the contact time raised the conversion value from 70 to 100\u00a0% due to prolonged contact of reactants with catalyst. The major hydrocarbon products shown in Fig. 8\na were 1-octene, n-octane, and i-octenes; the minor products were n-heptane and i-heptenes with selectivities less than 1\u00a0%. For 1-octene, the selectivity followed a positive curvilinear trend in which it initially increased from 1.1 to 20.2\u00a0% at contact time of 0.025\u20130.25\u00a0h and then decreased afterwards to 6.4\u00a0%, suggesting that 1-octene is an intermediate product. For i-octenes, a similar trend to that of 1-octene was observed; however, the inflection point was located at higher contact times, suggesting that i-octenes was an intermediate product formed afterwards. It should be noted that significant increase in the selectivity to i-octenes were observed at a contact time of 0.25\u20130.5\u00a0h, which coincided with decrease in the selectivity to 1-octene, suggesting that i-octenes was produced from 1-octene through isomerization. The selectivity to n-octane increased linearly with contact time (0 to 30.6\u00a0%), indicating that n-octane is a final product from hydrogenation of i-octenes and 1-octene.For oxygenates, the selectivity to octanal initially increased until a contact time of 0.04\u00a0h and decreased subsequently. This trend is similar to the trend of secondary or intermediate products, as shown previously for i-octenes or 1-octene; however, octanal, which should be a primary product in this reaction via the hydrogenation of octanoic acid should exhibit a selectivity trend which follows a continuous downward trajectory. This unexpected trend suggests that the increase in selectivity at short contact time is most likely due to the dehydrogenation of octanol from the equilibrium state of octanal and octanol. However, at longer contact time, the equilibrium shifts to the formation of octanol due to longer interaction and increasing number of reactants on the catalyst surface. In case of octanol and octyl octanoate, both shows similar trend in which the selectivity decreased continuously with increasing contact time. This suggests that octanol is converted into either octanal, octyl octanoate, or dioctyl ether at short contact time and possibly to 1-octene at longer contact time. For dioctyl ether, the selectivity profile followed a curvilinear trend like octanal, meaning that it is an intermediate product that is formed from octanol at short contact time and is converted by subsequent hydrogenation to octanol and 1-octene at longer contact time.Based on the observations above, the following reaction network can be proposed as shown in Fig. 8. First, octanoic acid can be transformed by two different reaction pathways: either by forming octanal (black arrow) through hydrogenation or by forming i-heptenes through DCO (dotted red arrow). With the presence of the Ru3Sn7 alloy phase in the RuSn/SiAl catalysts, DCO activity is suppressed significantly leading to preferential conversion of octanoic acid to octanal. The formed octanal can also undergo two different reaction routes to form either octanol (black arrow) by further hydrogenation or i-heptenes by HDC (dotted red arrow). Again, on Ru3Sn7 alloy such i-heptenes formation is very limited, so octanal is favorably converted into octanol. When octanol is formed, it can be converted by four different reactions: full dehydration to form 1-octene (black arrow), partial dehydration to form dioctyl ether (dotted orange arrow), esterification with octanoic acid to form octyl octanoate (dotted blue arrow), and reverse reaction to form octanal by dehydrogenation (black arrow). The tendency for these reactions depends on the reaction temperature, contact time, and acidity of the support. As can be observed in Fig. 3, the high temperature favors full dehydration over partial dehydration. Under isothermal conditions, short contact time favors reverse reaction to form octanal and octyl octanoate as major products with dioctyl ether as minor product. On the other hand, long contact time favors full dehydration of octanol to 1-octene. Based on the support acidity, the absence of acidity such as in SiO2 inhibits the dehydration process; hence, the formation of octyl octanoate dominates on RuSn/SiO2. On the other hand, the presence of Lewis acid sites in Al2O3 promotes the dehydration of octanol to 1-octene.Both reactions of partial dehydration and esterification are reversible reactions. If dioctyl ether is formed, it can be converted into octanol and 1-octene by decomposition (dotted orange arrow). If octyl octanoate is formed, it can be further converted back into octanol by hydrogenation (dotted blue arrow) or into octanoic acid and octanol by hydrolysis (dotted black arrow). Meanwhile, 1-octene formed from dehydration of octanol can undergo either isomerization to form 2-octene, 3-octene, and other isomers (dotted yellow arrow), or direct hydrogenation to form n-octane (black arrow). Similar to 1-octene, i-octenes can also be converted into n-octane through hydrogenation (black arrow). The presence of medium and strong acid sites enables the isomerization of 1-octene and hydrogenation of octenes by providing H-transfer to the metal site. Overall, the HDO of octanoic acid to n-octane over the RuSn/SiAl catalysts involves sequential reaction steps that occur on two different main sites. The metallic sites of Ru3Sn7 enable the conversion of octanoic acid to octanol, in which with close proximity to the support catalyzes the dehydration of octanol to form octenes. Further hydrogenation by the metallic sites results in the formation of n-octane as the final product.The conversion of octanoic acid was carried out over RuSn/SiAl catalysts with different SiO2 loadings. Characterization of the catalysts suggests that the addition of SiO2 on Al2O3 modified the surface of Al2O3 which was beneficial to the formation of Ru3Sn7 alloy phase. This alloy phase was crucial to the high conversion of octanoic acid to octanol, and the diminished formation of C7 hydrocarbons through HDC/DCO mechanism. The addition of SiO2 on Al2O3 was also beneficial for modulating the acidity of Al2O3, in which the modified acidity enhanced the dehydration of octanol and hydrogenation of octenes. The stronger acidity observed in the RuSn/SiAl catalysts was attributed to the formation of terminal silanols, which induced strong Br\u00f8nsted acid property on the support. It was revealed that n-octane was formed in significant amounts as SiO2 loading increased due to a longer interaction between octenes and support, facilitating the over-hydrogenation. The synergetic relationship between RuSn metal alloy and SiAl support resulted in high activity for octanoic acid transformation to C8 hydrocarbons with preserved carbon number.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This study was supported by the National Research Foundation of Korea (NRF) (2020M1A2A2080851) and the Institutional Research Program of KRICT (KK2311-10).Supplementary data to this article can be found online at https://doi.org/10.1016/j.cej.2023.141912.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  The ever-increasing demand for substitute energy sources encourages the usage of biofuels as transportation fuel; however, the unstable properties and low calorific value of biofuels caused by high oxygen content limits their direct application. In this study, we report the upgrading of a model biofuel in octanoic acid by hydrodeoxygenation (HDO) using RuSn supported on SiO2-doped Al2O3 (SiAl). The presence of SiO2 on Al2O3 was crucial for the formation of Ru3Sn7 alloy phase because SiO2 modified the metal-support interaction between RuSn and Al2O3. The alloy phase conferred significant hydrogenation activity to convert octanoic acid to octanol and minimized carbon loss by reducing decarbonylation activity. The deposition of SiO2 also resulted in improved acidity, which increased with increasing SiO2 loading from 10 to 30\u00a0%, due to the formation of isolated terminal silanols (Si-OH). The presence of the silanol groups was important for dehydration of octanol to octane, since the silanols retained octenes longer on the support, facilitating over-hydrogenation. The cooperative metal-support activity on RuSn/SiAl enables high and selective hydrogenation and dehydration activity for upgrading fatty acids in biofuel to hydrocarbons.\n               "}
{"full_text": "The root causes of a series of frontier hot issues, such as global climate anomalies, energy supply shortages, and ecological deterioration are the cliff-like decline of energy reserves and insufficient sustainable development [1]. As a clean energy source without supply constraints, The development of clean energy source has become a new trend towards global climate anomalies, energy supply shortages, and ecological deterioration. Solar energy converted by photoelectric devices can meet the urgent needs of a large amount of electronic equipments for human activities. Inspired by the mechanism of \u201cphotosynthesis\u201d, biomimetic constructed dye-sensitized solar cells (DSSCs) relied on their cost-effective and environmental-friendly characteristics have posted great potential on photoelectric conversion application. Generally, the DSSCs was assembled with a unique sandwich structure, matching with four components: photoanode coated on conductive glass (TiO2, ZnO, WO3, SnO2, Nb2O5), dye sensitizer, redox electrolyte, and counter electrodes (CEs) [2,3]. Noteworthily, the CEs component concerning on the electrical conductivity, electrocatalysis, and ion diffusion levels plays a non-negligible role in determining photovoltaic performance [4]. Platinum (Pt) as one of the most typical CEs owing to its high price, low abundance, and detrimental effects on the environment, limits large-scale industrial production [5,6]. Hence, it has become an urgent task to explore low-cost, harmless, and stable non-platinum CEs for sustainable photoelectric conversion.Currently, carbon materials featuring high porosity and good chemical inertness to corrosive electrolytes provide a good alternative to Pt [7\u20139]. [10,11]. Nevertheless, the disadvantage of the carbon materials is that it limites by electrical conductivity, which has a effect on the efficient transfer of electrons between and within the carbon particles, thus leading to higher diffusion resistance of I3\n\u2212 ions. Moreover, the complete framework obtained from such conventional carbon materials tend to show carbon wholeness at the macro scale of millimeter to centimeter scale, resulting in low bonding strength, poor dispersion, and poor contact between the large particles and the FTO interface.Moreover, considering the higher photovoltaic level, the improvement of the electrocatalytic performance of pure carbon materials still needs to be explored and optimized. Current research focuses on the construction of hetero-structured systems by arranging different materials (selenides, metal alloys, sulfides) [12\u201316], or by chemically doping heteroatoms (nitrogen, sulfur, phosphorus, boron) [17\u201319] with carbon defect sites. There are still key issues that need to be addressed, such as the non-renewability and the technical difficulty of the former, the reproducibility and precise control of the position of the doping defects and network pore structure of the latter, and the poor contact between the conductive media. A general design-principle goes beyond these existing methods of structuring carbon materials to achieve precise control of defect doping in structured-carbon. It can create more active sites for the absorption and transfer of I3\n\u2212 and H\n+ ions, thus achieving the improvement of comprehensive performance of electrical conductivity, catalysis, and electron transport. The precise control of the structural properties of carbon materials, such as particle size, morphology, and porosity, is an exploratory solution to the existing bottleneck.Herein, based on the principle of anisotropic integrated design, this work is expected to generate new physical properties and electronic activities by constructing unique structural arrangements and optimizing the composition ratio, and to generate more active sites by reducing agglomeration and synergistic effects. Specifically, pitaya peel-derived carbon structures were used as the matrix of in situ self-generated N-doped CNTs-coated Ni nanoparticles embedded in N-doped 3D network (Ni@NCNTs/PC-X), realizing the construction of self-generated Ni-N-C hybrid sites in 3D network structures as CEs material. More interestingly, in addition to providing anchor points to stabilize Ni atoms and adjust the electronic structure of surrounding atoms, melamine can also participate in the formation of N-doped porous carbon carriers and N-doped self-grown CNTs, improving the mass transfer catalysis process of Ni@NCNTs/PC-X. The three dimensional network characteristics of Ni@NCNTs/PC-X active sites were revealed by structural characterization, morphology and chemical composition analysis. Based on its good electrical conductivity and electrocatalytic electrochemical oxidation\u2013reduction performance, its photovoltaic performance as a CEs material was initially explored, and its potential application in DSSC was verified. It is expected to promote the application of biomass-based carbon materials in photovoltaic electrodes.Melamine (C3H6N6), nickel acetate tetrahydrate (Ni(CH3COO)2\u00b74H2O), acetic acid (C2H4O2) were provided by Aladdin Chemical Reagent Co., Ltd (Shanghai, China). Polymer spacers (Surlyn, 60\u00a0\u03bcm thickness), FTO glass slides with a size of 2.5\u00a0\u00d7\u00a02\u00a0cm, resistivity of \u223c7 ohm/sq were bought from Yingkou Opivite New Energy Technology Co., Ltd (Liaoning, China).First, the rinsed red pitaya peels were stripped of their folds and roasted at 70\u00a0\u00b0C for 24\u00a0h. The obtained 1\u00a0g of pitaya peels were calcined at 800\u00a0\u00b0C under N2 for 2\u00a0h. The calcined product was stirred in 2\u00a0g/20\u00a0ml KOH solution for 6\u00a0h, dried and then annealed at 700\u00a0\u00b0C under N2 for 2\u00a0h. The product (PC) was acid-washed with 1\u00a0M HCl for 2\u00a0h, washed repeatedly with distilled water and ethanol until neutral, and then dried at 80\u2103. 1\u00a0g of PC, 1\u00a0g of Ni(CH3COO)2\u00b74H2O, and 1\u00a0g of melamine (2\u00a0g, 4\u00a0g, and 8\u00a0g) were added to 30\u00a0ml of 37% CH3COOH solution and stirred until gel-like. Afterward, the gel was annealed at 700\u00a0\u00b0C at N2 for 2\u00a0h to obtain self-grown N-doped CNTs-coated Ni nanoparticles on N-doped porous carbon, named Ni@NCNTs/PC-X (X\u00a0=\u00a01, 2, 4 and 8), where X indicates the mass ratio of added melamine to PC.Preparation of counter electrodes (CEs): 0.2\u00a0g of the prepared CEs material, 10 \u03bcL of polyethylene glycol, and 20 \u03bcL of Triton X-100 were added into 1\u00a0ml of ultra-pure water, and mixed well by sonication. The doctor blade technique was adopted for the preparation of CEs with clean FTO as the substrate. The homogeneous membrane formed by adding 20\u00a0mM H2PtCl6\u20226H2O to the surface of FTO was sintered at 450\u00a0\u00b0C for 30\u00a0min to obtain Pt as CEs.Preparation of photoanodes: FTO conductive glass was washed twice with conductive glass cleaning solution, distilled water, and ethanol in sequence, and dried in an oven. The P25/Ni-2 sol-gel (detailed in Supplementary Information: Fig.S2, Fig.S3, Fig.S4) was scraped to a certain thickness on clean FTO through a filmmaker and sintered at 450\u00a0\u00b0C for 30\u00a0min. In order to obtain dense calcined electrode layers, it is necessary to immerse the above electrodes in 40\u00a0mM TiCl4 solution at 70\u00a0\u00b0C for 30\u00a0min and to undergo secondary sintering at 450\u00a0\u00b0C for 30\u00a0min. Finally, dye-sensitized photoanode was obtained by immersing the prepared electrode into 0.5\u00a0mM N719 and avoiding light at 25\u00a0\u00b0C for 18\u00a0h.The assembly of DSSCs: Photoanode (P25/Ni-2, P25) and CEs (Ni@NCNTs/PC-X, Pt) were assembled together with a 60\u00a0\u03bcm Surlyn film to obtain DSSCs with an active area of 0.25 cm2. Then, an electrolyte consisting of 0.025\u00a0M I2/0.1\u00a0M guanidine thiocyanate/0.5\u00a0M 4\u2011tert\u2011butyl pyridine/0.6\u00a0M 1-methyl-3-propylimidazolium iodide in a mixture of acetonitrile/valeronitrile (85:15 v%) was injected into the CEs cell through the reserved hole. The reserved hole was sealed to prevent the electrolyte from volatilizing.Samples were characterized and measured by X-ray diffractometer (Rigaku D/MAX 2500\u00a0V, Japan), Four-point probe meter detector (RTS-8, 4Probes Tech Ltd. China), Field emission scanning electron microscopy (FE-SEM, Zeiss Sigma 300, Germany), X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific K-Alpha Plus), Fourier transform infrared spectroscopy (FTIR, Thermo Fisher Scientific, USA), Raman spectroscopy (inVia Reflex, Renishaw, UK), A Brunauer-Emmett-Teller (BET, JW-BK400, Jingwei Gaobo Co., Ltd. China), an electrochemical workstation (CS350, Wuhan Corrtest Instruments Corp., Ltd, China), and a solar simulator with a digital source meter (SOFN INSTRUMENTS CO., Ltd, China) (see supplementary information for details).Based on anisotropic integrated design principles, we wish to propose a system of self-grown N-doped CNTs-supported Ni nanoparticles on N-doped porous carbon. The design with abundant active sites provides suitable energy for binding triiodide ions, where N-doped CNTs sites can shorten the diffusion distance of triiodide ions, and N-doped porous carbon matrix provides fast electron transport pathways and acts as physical barriers in process catalysis reactions. Previous literatures have demonstrated that composite metal-based catalysts generally exhibit enhanced catalysis action than single metal-based catalysts because of the synergistic action of multiple elements. Here, we are particularly concerned with the construction of unique Ni-intercalated composite co-doped N-doped CNTs by Ni metal particles on active N-doped PC (Ni@NCNTs/PC-X). Using PC, Ni(CH3COO)2\u00b74H2O, and melamine as precursors, a simple high-temperature in-situ pyrolysis, doping, self-growth, and self-assembly method were used. The nitrogenous hybridization of melamine promoted the formation of N-doped ordered porous carbon and provided the required spatial arrangement. Melamine provided both C and N sources. The carbothermal reduction of Ni(CH3COO)2\u00b74H2O precursor enhanced the graphitization effect of Ni-doped porous carbon. It was partially inserted into the carbon wall of N-doped porous carbon as a Ni-doped catalyst for N-doped self-growing CNTs. The catalytic graphitization and nitridation eventually lead to Ni intercalation and N-doped CNTs (Ni@NCNTs). The synergistic effect between the components is expected to enhance the electrocatalytic performance of this hybrid by the Ni metal component and N-doped CNTs.To achieve this goal, Ni@NCNTs/PC-X composites were designed and constructed in a simple and efficient manner, as shown in Fig.\u00a01\n. To further study the effect of experimental parameters on the structure, a series of samples (Ni@NCNTs/PC-X composites) with different melamine contents (1\u00a0g, 2\u00a0g, 4\u00a0g, and 8\u00a0g) were designed and prepared based on PC, Ni(CH3COO)2 4H2O and melamine. Ni@NCNTs was in situ immobilized on N-doped PC by Ni-N-C hybridization and uniformly distributed in the non-agglomerated N-doped PC particles to form an interconnected conductive framework. Therefore, a well-networked Ni@NCNTs/PC composite was creatively constructed by in-situ growth of Ni@NCNTs among N-doped PC particles at 700\u00a0\u00b0C. All of these self-grown N-doped CNTs embedded nickel nanoparticles supported by N-doped porous carbons are discussed in the morphology and composition analyses. To investigate the effect of N and Ni co-doping on PC-modified CEs, electrochemical results confirmed that co-doped Ni@NCNTs/PC-X exhibited better performance than unmodified PC. This detail is discussed in the chapters on photovoltaics, electrocatalysis, and electrochemical properties.The crystallization properties of in-situ self-grown N-doped CNTs-coated Ni nanoparticle composites (Ni@NCNTs/PC-X) of pitaya peel-porous carbon were investigated by XRD patterns and Raman spectroscopy (Fig.\u00a02\na and b). As far as XRD patterns, in addition to the characteristic diffraction peaks of amorphous carbon, Ni@NCNTs/PC-X has broad and weak characteristic diffraction peaks at 2\u03b8=25.8\u00b0 and 41.5\u00b0 They represent the typical diffraction peaks in graphitic carbon (002) and (100) planes, indicating the coexistence of amorphous graphitic carbon. And more, there are distinct peaks at 2\u03b8=44.2\u00b0 (111), 51.5\u00b0 (200), and 76.2\u00b0 (220), originating from Ni doping. PC has a broad diffraction peak at 25.8\u00b0 After the introduction of melamine, the peaks of the Ni@NCNTs/PC-1 to Ni@NCNTs/PC-8 composites gradually narrowed and enhanced at 25.8\u00b0, indicating that the graphitization effect was enhanced by Ni catalysis and melamine nitridation. Therefore, Ni@NCNTs/PC-X shows typical characteristics of Ni-containing graphitic carbon compared to the typical amorphous carbon of PC. Furthermore, in Fig.\u00a02b, both PC and Ni@NCNTs/PC-X have two distinct peaks at \u223c1348 cm\u22121 and \u223c1588 cm\u22121, which belong to the D and G peaks of graphite [20]. It cannot be ignored that the ID/IG values of the Ni@NCNTs/PC-X are all lower than that of PC (1.08). This indicates that N-doped CNTs grow on the surface of porous carbon, and the carbon in the composite material has a higher structural order. In addition, the melamine content has an important effect on the ID/IG value of Ni@NCNTs/PC-X, that is, the ID/IG value of Ni@NCNTs/PC-X composites decreased from 1.07 (Ni@NCNTs/PC-1) to 0.99 (Ni@NCNTs/PC-8). The lower ID/IG value of Ni@NCNTs/PC-X composites indicated the increased graphitization, which agreed with the XRD analysis.Ni@NCNTs/PC-X was successfully constructed by this unique Ni-embedded graphical-carbon of self-grown CNTs on the N-doped-carbon surface combined with the abundant defective sites, which has good electrical conductivity, excellent electrolyte diffusivity, and electrocatalytic properties. Ni@NCNTs/PC-X (Ni@NCNTs/PC-1 of 129.1 S cm\u22121, Ni@NCNTs/PC-2 of 165.3 S cm\u22121, Ni@NCNTs/PC-4 of 215.7 S cm\u22121, Ni@NCNTs/PC-8 of 226.8 S cm\u22121) has a higher degree of graphitization, a more regular and ordered structure, and a higher electrical conductivity than pure amorphous PC with defects (18.3 S cm\u22121). The increased electrical conductivity benefits from the unique ordered graphite-carbon micro-nano structure through the hybrid by the Ni metal component and N-doped CNTs supported by N-doped PC. These key performances were confirmed by the above-mentioned XRD and Raman analyses, as well as subsequent in-depth analyses of quantitative indicators of elemental composition and chemical state, microstructure and pore structure.In particular, the elemental types-chemical configurations of PC and Ni@NCNTs/PC-4 were further explored by high-resolution XPS spectroscopy. The full XPS spectrum of Ni@NCNTs/PC-4 (Fig.\u00a02c) shows the existence of N and Ni peaks in Ni@NCNTs/PC-4, which indicates that Ni ions and N are doped into PC. In Fig.\u00a02d, the four peaks of high-resolution C 1\u00a0s spectrum fitted at 284.5\u00a0eV, 284.9\u00a0eV, 286.2\u00a0eV, and 288.9\u00a0eV are CC/C\u00a0=\u00a0C, CN, C\u00a0=\u00a0N/CO, and C\u00a0=\u00a0O, respectively [21,22]. Observations of CN and C\u00a0=\u00a0N bonds provide strong support for the arguments for N-doped porous carbon and N-doped CNTs. The high-resolution spectrum of N 1\u00a0s at 398.0\u00a0eV, 400.1\u00a0eV and 403.5\u00a0eV represent pyridine N, pyrrole N and graphite N with relative contents of 24.07%, 47.28% and 28.65%, respectively. For Ni@NCNTs/PC-X, the higher ratios of pyridine N and pyrrole N formed in the structure of Ni@NCNTs/PC-4 (24.07%, 47.28%) have better electrocatalytic effects compared with Ni@NCNTs/PC-1 (18.31%, 41.45%), Ni@NCNTs/PC-2 (20.36%, 43.57%), Ni@NCNTs/PC-8 (22.13%, 45.06%) [23]. The predominant pyridine-N with lone pair electrons may be beneficial for electrocatalysis because the electron-donating effect is beneficial for reducing the electronic work function of carbon [23]. Besides, the characteristic metallic Ni in the Ni 2p spectrum was found at 855.3\u00a0eV, confirming the thermal reduction of nickel acetate (Ni2+) to Ni as catalysts in carbon under a nitrogen atmosphere [24]. Also, the presence of trace amounts of Ni2+ may be caused by the oxidation of air left over from the heating process.To further confirm the existence and quantitative analysis of the microscopic pore structures of PC and Ni@NCNTs/PC-X composites, N2\n adsorption-desorption experiments were measured and analyzed. As shown in Fig.\u00a03\na-b, all samples exhibited typical IV isotherms and obvious hysteresis loop characteristics when the relative pressure (P/P0) was 0.4\u20131, confirming the existence of mesoporous features. As shown in the inset of Fig.\u00a03b, the pore sizes of the composites are concentrated between 2 and 3\u00a0nm with the average pore size of 2.5\u00a0nm, which belongs to the mesoporous size. The specific surface area (SBET) of pitaya peels-based PC is 1101.9 m2/g, and the large SBET is beneficial to the charge transport and the growth of CNTs on its surface. The SBET of the Ni@NCNTs/PC-X composites is mainly controlled by the porous PC and CNTs grown on it. With the growth of CNTs, the SBET decreased from 1041.3 m2/g (Ni@NCNTs/PC-1) to 884.1 m2/g (Ni@NCNTs/PC-2), 602.9 m2/g (Ni@NCNTs/PC-4) and 513.7 m2/g (Ni@NCNTs/PC-8). With the increase of melamine content, the growth of CNTs on the surface of porous PC increased and the SBET decreased, leading to the reduction of electrolyte catalyzed by the composite, which may be unfavorable. In Ni@NCNTs/PC-X composites, more electrocatalytic sites can be provided for I3\n\u2212 to promote electron transport and reduce charge transport resistance by adjusting the amount of melamine, appropriately using porous carbons with high SBET and high mesoporosity, and maintaining a high catalytic activity area.The morphology of the Ni@NCNTs/PC-X corresponds to its SBET, and the increase of the number of carbon nanobundles may have a negative impact on its SBET contribution (Fig.S1), which is not beneficial to provide more electrocatalytic active sites. The morphological characterization of Ni@NCNTs/PC-X composites (Fig.S1) shows that the morphology of the composites can be adjusted to obtain the best electrocatalytic performance by reasonably controlling the mass ratio of melamine to porous carbon. After activation by KOH etching, pores of different sizes appeared on the surface of the porous carbon material, which can accelerate the diffusion of electrolytes in the carbon-based electrode, as shown in Fig.\u00a03c-h. However, small-scale dispersion and non-uniform morphology lead to poor electrical contact performance between materials with high interfacial resistance and high potential barrier, which can adversely affect the charge conduction between carbon matrices. Therefore, it is important to construct an intermediate bridge that can shorten the charge transfer. Interestingly, when a certain mass of melamine was added to the porous carbon material, the high-temperature carbonized precursor and Ni-catalyzed graphitization pyrolysis induced the successful self-growth of carbon nanotube bundles of different lengths on the porous carbon surface. The use of steric hindrance during the preparation of Ni@NCNTs can avoid the increase in the size of PC due to aggregation, ensure a smaller size effect, and also play a positive role in shortening mesoporous channels. Therefore, the protruding carbon nanotube bundles effectively shorten the transport distance of charge in the carbon material, thereby reducing the resistance of charge transport. Furthermore, a three-dimensional conductive interpenetrating network of Ni@NCNTs/PC-X is creatively designed by embedding Ni@NCNTs uniformly into the internal structure and particle gaps of PC. In addition, the EDS mapping shows that C, N, and Ni elements are well homogeneous in Ni@NCNTs/PC-4 (Fig.\u00a03i).The fabrication of DSSCs assembled with P25/Ni-2 as photoanode and Ni@NCNTs/PC-X as CEs is represented in Fig.\u00a04\na. To investigate the electrocatalytic performance of pitaya peel-derived carbon-based composites as CEs, we assembled DSSCs with P25/Ni-2 as photoanode and Ni@NCNTs/PC-X as CEs (Fig.\u00a04b and Table\u00a01\n). The PCE of DSSCs based on the activated pitaya peel-derived PC as CEs was 2.47%, which may be attributed to the high SBET of PC for good catalytic reduction performance for I3\n\u2212. It was also found that the PCE values of the DSSCs based on Ni@NCNTs/PC-X as CEs were correlated with the ratio of melamine to PC. When the amount of melamine was 0.5\u00a0g, 1\u00a0g, and 2\u00a0g (X\u00a0=\u00a01, 2, and 4), the PCE value gradually increased. With the addition of melamine increased to 4\u00a0g (X\u00a0=\u00a08), the PCE value decreased. Using Ni@NCNTs/PC-4 as CE, the optimal PCE is 5.13%, Voc is 0.69\u00a0V, Jsc is 13.27\u00a0mA/cm2, and FF is 0.56. Compared with P25/Ni-2-Pt, the Voc of P25/Ni-2-Ni@NCNTs/PC-4 decreased slightly, which may be related to the uneven surface of CE. The different PCE values are related to the specific surface area, and the unique micro-nano structure of Ni@NCNTs/PC-X electrode materials. Furthermore, the self-grown CNTs on the surface of the PC effectively shortened the electron transport path, resulting in the formation of a three-dimensional network structure of the dispersed PC. It shows that N doping level and microscopic pore structure had a synergistic effect on photovoltaic performance. The photovoltaic performance of DSSCs assembled with P25/Ni-2 as photoanode and Ni@NCNTs/PC-X as CEs exhibited a small statistical standard deviation, indicating that Ni@NCNTs/PC-X as CEs has good photovoltaic reproducibility.The CV curves of Ni@NCNTs/PC-X (Fig.\u00a04c and d) were used to evaluate the electrocatalytic properties and the kinetics of the I3\n\u2212/I\n\u2212redox reaction of CEs. For CV curves, the position and size of the two pairs of redox peaks are often used to evaluate the electrocatalytic performance of the electrode. The first redox peak is usually defined as Eq.\u00a0(1) and the other as Eq.\u00a0(2)\n[12]:\n\n(1)\n\n\n\n\n\nI\n\n3\n\u2212\n\n+\n2\n\n\ne\n\n\u2212\n\n=\n3\n\n\nI\n\n\u2212\n\n\n\n\n\n\n\n(2)\n\n\n3\n\nI\n2\n\n+\n2\n\n\ne\n\n\u2212\n\n=\n2\n\nI\n3\n\u2212\n\n\n\n\n\nIn general, the smaller the potential difference (Ep) between the first redox peaks, the faster the kinetics of the redox reaction, which means a better reduction of I3\n\u2212\n[25]. It can be seen from Table\u00a01, Ep values in descending order are PC (1110\u00a0mV) > Ni@NCNTs/PC-2 (668\u00a0mV) > Ni@NCNTs/PC-1 (636\u00a0mV) > Ni@NCNTs/PC-8 (620\u00a0mV) > Pt (588\u00a0mV) > Ni@NCNTs/PC-4 (474\u00a0mV). Apparently, Ni@NCNTs/PC-4 CE has the best catalytic activity compared with other CEs, which may be due to the best synergistic effect of PC with its self-grown CNTs leading to its good morphology and structure. The CV curve of Ni@NCNTs/PC-4 as CE (Fig.\u00a04e) shows that the diffusion rate of the I3\n\u2212/I\n\u2212ion pairs on the electrode surface increases with the increase of scanning speed, and the Ox-1/Red-1 peak moves in the direction of positive and negative potentials, respectively. Also, the linear relationship between the peak current density and the square root of the scanning rate (shown in the inset) is consistent with Langmuir isotherm theory, implying that diffusion is the controlling mechanism of the redox reaction of I3\n\u2212/I\n\u2212, and there is no interaction between the surface interface of Ni@NCNTs/PC-4 and the electrolyte [26].The Rs values of different CEs based on pitaya peel-derived carbon ranged from 19 to 23 \u03a9, which was close to 17.18 \u03a9 for CE based on Pt, shown in Nyquist plots (Fig.\u00a04f). The size of Rs has little effect on PCE value, so the key is the impact of Rct on PCE. Due to the lack of sufficient active sites, the activated pitaya peel-derived carbon (PC) has a larger Rct value of 12.53 \u03a9. The Rct values of Ni@NCNTs/PC-X as CEs are all lower than that of PC, indicating that the growth of CNTs could effectively reduce the Rct value [27]. The Rct value of Ni@NCNTs/PC-X CEs decreases from 10.44 \u03a9 (Ni@NCNTs/PC-1) to 5.30 \u03a9 (Ni@NCNTs/PC-2) and 5.21 \u03a9 (Ni@NCNTs/PC-4), which is close to Pt's 4.91 \u03a9. Furthermore, the Rct value of Ni@NCNTs/PC-8 increased to 6.26 \u03a9 when the melamine level increased to 4\u00a0g (X\u00a0=\u00a08). The variation of Rct value of Ni@NCNTs/PC-X CEs is consistent with the J-V curve. The reason for the increased Rct value of Ni@NCNTs/PC-8 may be that the number of CNTs growing on PC surface increased, SBET and Rct decreased due to the increase of melamine content. The lower Rct values of Ni@NCNTs/PC-4 imply a higher electrocatalytic performance for I3\n\u2212, which is more favorable for electron transport.Meanwhile, the exchange current density (J0) of Ni@NCNTs/PC-4 (0.0013\u00a0mA/cm2) is higher than those of PC (0.0007\u00a0mA/cm2), Ni@NCNTs/PC-1 (0.0008\u00a0mA/cm2), Ni@NCNTs/PC-2 (0.0011\u00a0mA/cm2) and Ni@NCNTs/PC-8 (0.001\u00a0mA/cm2) (Fig.\u00a04g). Ni@NCNTs/PC-4 has the lowest Rct and the highest J0, indicating that it has a good electrocatalytic effect on the reduction of I3\n\u2212\n[28]. Ni@NCNTs/PC-4 composites exhibit higher Jlim (limiting diffusion current density) values than other CEs materials. According to Eq.\u00a0(3), it can be confirmed that the surface between the I3\n\u2212 ions in the electrolyte and the Ni@NCNTs/PC-4 has a large diffusion coefficient (D).\n\n(3)\n\n\n\n\nJ\n\nl\ni\nm\n\n\n=\n2\nn\ne\nD\nC\n\nN\nA\n\n/\nl\n\n\n\n\nThe strong diffusion rate of the as-prepared composite CEs is mainly attributed to its large specific surface area and suitable carbon nanotube growth on its surface, which is of great significance to shorten the diffusion distance of I3\n\u2212 ions, reduce the charge transfer resistance, and ultimately improve the electrocatalytic performance.The present results were also compared with those of composites prepared by similar methods using carbon from other biomass sources or commercial carbon materials in Table\u00a02\n. The PCE value (5.13%) of the Ni@NCNTs/PC-4-based CEs assembly was higher than previously reported biomass carbon-based CEs (PCE=1.09\u20134.98%). This also confirms the contribution of self-generated Ni-N-C hybrid sites in 3D network structures as CEs for improving the photoelectric conversion efficiency.The CV curves overlapped well after continuous scanning without significant changes, as shown in Fig.\u00a05\n(a, c, e, g). Furthermore, no exfoliation of the electrode material from the surface of the FTO substrate was exhibited (insets of Fig.\u00a05), which corroborates the excellent electrochemical stability of Ni@NCNTs/PC-X as CEs in electrolytes. Fig.\u00a05(b, d, f, h) shows the corresponding peak current densities of oxidation and reduction under different scanning times. The increase in the number of scans did not lead to a large change in the curve shape and peak current density of CEs, which also suggested that the CEs prepared by Ni@NCNTs/PC-X composites had good stability in the redox reaction of I3\n\u2212/I\n\u2212. Notably, all CEs exhibited remarkable repeatability in 50 consecutive scan curves, and the as-prepared Ni@NCNTs/PC-X CEs exhibited high electrochemical stability.\nFig.\u00a06\n takes the Ni@NCNTs/PC-4 composite with the best electrocatalytic performance as an example to explore its electrochemical properties and cycling stability. All CV curves (5\u2013150\u00a0mV/s, Fig.\u00a06a) show good shape retention and good capacitance values. At a high scanning rate of 150\u00a0mV/s, the shape of the curve can also be well maintained, showing good electrochemical stability and reversibility. The GCD measurements of the composites were simultaneously performed from 1 A/g to 10 A/g. From Fig.\u00a06b, it can be seen that the potential changes linearly with time, and the shape is a typical symmetrical triangle, which also indicates that it is contributed by the Ni@NCNTs/PC-4 electric double layer capacitance. There is almost no voltage drop in the GCD curve, suggesting that the electrode has the advantages of low internal resistance, fast charging and discharging, and good rate performance. The specific capacitances (Fig.\u00a06c) of Ni@NCNTs/PC-4 are 91.0, 90.9, 87.1, 85.1, 83.9, 82.0, 79.7, and 77.5 F/g, respectively. To test the cycling stability of the Ni@NCNTs/PC-4 composite, as shown in Fig.\u00a06d, the capacitance was characterized by 10,000 consecutive charge-discharge cycles at 6 A/g between \u22121 and 0\u00a0V. The results show that Ni@NCNTs/PC-4 electrode has a high capacitance retention rate of 98.91% after 10,000 cycles, which can be clearly confirmed by the last 10 charge-discharge cycles (inset of Fig.\u00a06d). The electrochemical test and analysis of Ni@NCNTs/PC-4 as an electrode material further confirmed its good electrocatalytic performance and charge-discharge stability, and proved its practicability as a photovoltaic electrode from multiple perspectives.The purpose of this study is to solve the problems of conductivity, catalytic performance, dispersion, contact resistance and ion diffusion of traditional biomass carbonaceous materials as electrodes in DSSCs. To this end, a Ni-N-C hybrid 3D ionized sites-network CE was construed in the N-doped 3D network by in-situ self-grown N-doped CNTs and Ni nanoparticles embedded using the carbon structure derived from the pitaya peel as the matrix. The mass ratio of melamine to 3D structured-carbon increased from X\u00a0=\u00a01 to X\u00a0=\u00a08, and the number of in-situ self-grown CNTs on the surface of 3D structured-carbon increased gradually, while the surface area (SBET) decreased gradually. The optimal SBET of Ni@NCNTs/PC-4 exhibits excellent electron transfer ability and good stability. In addition to providing anchor points to stabilize Ni atoms, melamine can also participate in the formation of N-doped porous carbon supports and N-doped self-grown CNTs, improving the mass transfer of Ni@NCNTs/PC-4 during catalysis. Structural characterization, micromorphological and chemical composition analyses revealed that Ni@NCNTs/PC-4 had abundant active sites. The DSSCs assembled with Ni@NCNTs/PC-4 CEs exhibited good photovoltaic performance with a PCE value of 5.13%\u00b10.05, which was higher than that of the DSSCs with PC (2.47%\u00b10.05) and close to that with Pt (5.60%\u00b10.10). This novel Ni@NCNTs/PC-x expands the choices of electrodes for designing DSSCs, especially for the purpose of using renewable biomass carbon sources for massive production.This work was supported by the National Natural Science Foundation of China (31960293).\nGenhui Teng: Writing \u2013 review & editing. Baorui Liu: Conceptualization. Zhe Kang: Investigation. Yanhui Xie: Methodology. Dongying Hu: Supervision, Writing \u2013 review & editing. Dawei Zhao: Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.recm.2022.11.001.\n\n\nImage, application 1\n\n\n\n", "descript": "\n                  The technical bottleneck of carbon materials as counter electrodes (CEs) lies in their limited electrical conductivity, extended ion diffusion paths, poor dispersion, and high contact resistance. Problem-oriented in-situ self-grown N-doped CNTs-coated Ni nanoparticles based on N-doped carbonaceous structures derived from pitaya peel (PC) are adopted to construct Ni-N-C hybrid 3D ionized network sites (Ni@NCNTs/PC-4) as CEs. Structural characterization, micromorphological and chemical composition analyses revealed the 3D network structure of Ni@NCNTs/PC-4 with abundant active sites. They effectively shorten the diffusion distance of I3\n                     \u2212 ions with a smaller charge transfer resistance (5.21 \u03a9) than that of PC (12.53 \u03a9). DSSCs based on Ni@NCNTs/PC-4 display good optoelectronic properties, in which the short-circuit current density (Jsc) is 13.27\u00a0mA/cm2, higher than those of Pt (11.66\u00a0mA/cm2) and PC (6.99\u00a0mA/cm2). The PCE value (5.13%) of DSSCs based on Ni@NCNTs/PC-4 is also higher than that of DSSCs based on PC (2.47%). Overall, this work provides a preliminary research and new ideas for further in-depth study of biomass-derived 3D structured-carbons that contribute to key electrodes in DSSCs.\n               "}
{"full_text": "Data will be made available on request.The excessive consumption of single-use plastics in the age of consumerism has generated a staggering amount of global plastic production. The short service life of these highly durable plastics creates enormous pressure on municipal waste management systems. Experts estimate an increase in global plastic waste production from 240 Mt/y in 2016\u2013430\u00a0Mt/y in 2040 in a business-as-usual scenario [1]. This data translates to the outflow of \u223c1.71 and \u223c0.75 billion metric tonnes of plastic waste into the aquatic and terrestrial environment by 2020\u20132040 [1]. Sufficient evidence has demonstrated the possible trophic transfer of the fragmented plastic waste through the aquatic food web, causing an increased risk of toxicity to humans as one of the top predators [2]. Scientists have also pointed out that plastics may hinder the carbon sequestration ability of phytoplankton and zooplankton, hence impeding the role of oceans as the most significant carbon sink on Earth [3]. Therefore, immediate actions are needed to handle the increasing amount of plastic waste, especially the unrecyclable ones. While the gradual bans of single-use plastics could serve as a temporary solution to the crisis, a paradigm shift from a linear to a circular economy of plastics is regarded as a sustainable solution in the long term without compromising the societal benefits of plastics [4,5].At present, there are nearly 28 technology providers around the world that have developed or are currently developing thermal-chemical recycling technologies to promote the circular economy of plastics [6]. Pyrolysis is one of the key technologies used by these providers, where the plastic waste is degraded in an inert environment at high temperatures (400\u2013600\u00a0\u00b0C) to produce value-added products, including liquid fuels, chemical feedstock, and carbon nanomaterials [7,8]. Given its strategic importance, plastic waste pyrolysis is an essential component in the SuSChem Strategic Innovation and Research Agenda which requires alignment of all actors in the innovation ecosystems [9]. At present, the technological readiness level of plastic pyrolysis stands at 6\u20137 [10]. Despite the significant progress in plastic pyrolysis development, several important issues limit the potential of large-scale plastic waste pyrolysis. As plastic pyrolysis is an endothermic reaction, the significant energy cost in the scaled-up process reduces the cost-competitiveness of the pyrolysis oil against the fossil-based oil. Most research teams utilized resistive heating in plastic pyrolysis, which is associated with low heating and cooling rates, and therefore low energy efficiency. This limitation is responsible for the high energy consumption and thus aforementioned low economic viability of large-scale plastic waste recycling. A potential solution to this challenge lies in replacing resistive heating with induction heating, which is a non-contact technique involving the induction of eddy current on the surface of a ferromagnetic metal placed in an alternating magnetic field. The eddy current generates the Joule heating effect, which leads to rapid heat generation of the metal. As most plastic pyrolysis reactors are made of stainless steel (which is a good susceptor in an alternating magnetic field), the application of induction heating in plastic pyrolysis could be a critical and innovative solution to achieve higher energy efficiency and, therefore, higher economic feasibility for large-scale application.Several research teams have investigated induction heating in biomass waste pyrolysis. Tsai et al. [11] reported the pyrolysis of biomass wastes to bio-oils in a stainless-steel reactor via induction heating. Muley et al. [12] also developed a two-stage reactor for the pyrolysis of pinewood sawdust with induction heating applied to biomass pyrolysis and catalytic upgrading stages. Compared to resistive heating, the adoption of induction heating in the catalysis stage lowered the degree of coke deposition in catalyst particles. This observation was related to a more efficient heat transfer from the stainless-steel reactor surface (susceptor) to the catalyst bed during the induction heating process. The pyrolysis of electronic waste through induction heating (with graphite crucible as susceptor) also led to a more significant weight reduction (by 7 %) than pyrolysis through resistive heating [13]. Compared with biomass pyrolysis, scientific investigations on the roles of induction heating in plastic pyrolysis are limited in the literature. Nakanoh et al. [14] were the first team who mentioned the feasibility of the process. However, no information was provided on the plastic pyrolysis product yields and compositions. Zeaiter [15] reported on the rapid decomposition of high-density polyethylene waste via induction heating during pyrolysis, resulting in insufficient contact between the pyrolysis intermediates and catalyst particles. This phenomenon ultimately led to high wax/liquid and solid yields but low gas yields. To date, there is an insufficient understanding of the feasibility of plastic pyrolysis via induction heating. To fill this knowledge gap, an exploratory study was performed to investigate the thermal and catalytic pyrolysis of neat plastics in a self-fabricated stainless-steel (SS316) fixed bed reactor using induction heating. Specifically, the effects of catalyst properties on the product yields and compositions from the pyrolysis of low-density polyethylene (LDPE), high-density polyethylene (HDPE), and polypropylene (PP) pellets in a nitrogen environment were investigated. This is the first study that reports the behaviors of thermal and catalytic pyrolysis of different polymers via induction heating of the reactor wall, which will provide valuable insights into novel strategies for plastic waste valorization.The LDPE, HDPE, and PP pellets (particle size: 3\u20135\u00a0mm) were supplied by a Spanish energy and petrochemical company and were used as received. For catalytic pyrolysis, ZSM-5 catalyst (CBV 2314, Zeolyst International, SiO2/Al2O3 =\u00a023) and spent fluid catalytic cracking (FCC) catalyst obtained from a Spanish-based international energy company were used. Before the experiments, the catalysts were calcined at 550\u00a0\u00b0C in the air for 3\u00a0h to remove all the adhered impurities (including moisture) [16]. The calcination step also converts the ZSM-5 zeolite from ammonium form to hydrogen form [17] and removed the coke formed on the FCC catalyst.The thermal degradation behavior of the plastics was analysed using a thermogravimetric analyser (TGA, STAR system, Mettler Toledo). The samples were purged with pure nitrogen gas (20\u00a0mL/min) and then heated at 10\u00a0\u00b0C/min in the temperature range of 25\u2013800\u00a0\u00b0C [18]. The raw TGA data were differentiated to obtain the derivative (DTG) curves for the samples. The molecular weight distribution (MWD), molecular weight averages (Mw), and polydispersity indexes (PI) of the plastic pellets and waxes from thermal pyrolysis were determined using gel-permeation chromatography (GPC) coupled with infrared detector (GPC-IR6) (Polymer Char, Spain) [19]. The samples were prepared by dissolving approximately 10\u00a0mg of plastics in 1\u00a0mL of 1,2-dichlorobenzene at 150\u00a0\u00b0C, followed by in-line filtration. The Mw and PI values referred to the monodisperse PS standards [19].The density and strength of the Br\u00f8nsted acidic sites on the catalysts were characterized based on the thermogravimetric measurement of temperature-programmed decomposition of n-propylamine (NPA) in a DSC-TGA thermal analyser (Model: SDT Q600, TA Instruments) [20]. The sample was heated at 600\u00a0\u00b0C for 30\u00a0min in nitrogen flow (30\u00a0mL/min) to remove all absorbed impurities and then saturated with NPA at 150\u00a0\u00b0C. After that, the sample was heated in nitrogen flow (30\u00a0mL/min) at 150\u2013700\u00a0\u00b0C with a heating rate of 10\u00a0\u00b0C/min. The textural properties of the catalysts were characterized based on the nitrogen adsorption-desorption isotherm produced by the Micromeritics Gemini 2360 instrument at 196\u00a0\u00b0C [21]. Before the analysis, the samples were outgassed at 350\u00a0\u00b0C for 4\u00a0h. The surface area of the samples was computed using the Brunauer-Emmett-Teller's (BET) equation from the nitrogen adsorption curve in the region 0.05\u00a0\u2264\u00a0P/Po \u2264\u00a00.3. The pore size distributions of HZSM-5 and FCC catalysts were computed using the Barrett-Joyner-Halenda (BJH) method from the desorption isotherms. The XRD diffractograms were recorded on a Panalytical X\u2032Pert Pro (The Netherlands) using Cu K\u03b1 radiation generated at 45\u00a0kV and 40\u00a0mA. A scanning range from 5\u00b0 to 100\u00b0 was used at a speed of 0.03\u00b0/s.A fixed bed reactor system (\nFig. 1) was used for plastic pyrolysis through induction heating. The reactor was an AISI 316 stainless steel tube (length, L: 10.0\u00a0cm, outer diameter, OD: 2.2\u00a0cm). The plastic pellets and the catalysts were placed in the middle section of the reactor between the layers of quartz wool. The reactor also acted as a susceptor in an alternating magnetic field produced by a 3-turn copper coil (L: 5.0\u00a0cm, inner diameter, ID.: 4.5\u00a0cm) connected to an induction heater (Easyheat system, Ambrell, UK) with an output power of 1.2\u00a0kW. The current frequency was fixed at 315\u00a0kHz. A Superwool Plus fiber (13\u00a0mm) layer was placed in between the reactor and copper coil to reduce heat losses from the reactor wall to the surrounding. The reactor wall temperature was measured using a pyrometer (CTM-3CF75H1-C3, Micro-Epsilon) via a small opening in the insulation layer.The top part of the reactor was welded to an inlet AISI 316 tube (L: 5.0\u2009cm, OD: 2.5\u2009cm), while the bottom part was welded to an outlet AISI 316 tube (L: 5.0\u2009cm, OD: 1.8\u2009cm). A stainless-steel cold trap connected to the bottom reactor outlet was maintained at 2\u2009\u00b0C to collect the condensable reaction products. The non-condensable gases were collected in a water column placed after the cold trap. The total volume of the gas product evolved during plastic pyrolysis was determined based on the volume of water displaced from the water column into a measuring cylinder.In this exploratory study, 1.00\u2009g of plastic pellets (LDPE, HDPE, or PP), together with 0.20\u2009g of catalyst (in the case of catalytic pyrolysis), were placed on 0.20\u2009g of a quartz wool layer in the reactor. As a basis to estimate the effect of the catalyst, a series of blank experiments were carried out, hereafter termed \"thermal pyrolysis\" in this paper. Before the pyrolysis process, the reactor system was purged with nitrogen gas (120\u2009mL/min) for 10\u2009min to prevent the oxidation/combustion of the plastics during pyrolysis. Next, the nitrogen gas flow was turned off. The induction heater was turned on for 30\u2009min to allow reactor heating. After 30\u2009min, the induction heater was turned off, and the reactor was allowed to cool to room temperature. If the wax product was formed in the cold trap, the wax was collected and weighed. If a liquid product was formed, it was extracted with 3\u2009mL of dichloromethane before being weighed and analysed. All the pyrolysis experiments were performed in duplicates. The average values and per cent errors of product yields are presented in Section 3. In general, the experimental errors fall within 7%.According to the temperature measurement results (Section S1, Supplementary materials), induction heating successfully raised the reactor temperatures to 500\u2013700\u2009\u2103 within 10\u2009min. The heating rate of the system is estimated to be 50\u201385\u2009\u2103/min depending on the induction heater power. By definition, the system/process could be classified as fast pyrolysis, which has a typical heating rate of 10\u2013200\u2009\u2103/min [22]. Based on the temperature measurement results, the plastic pyrolysis experiments occurred at 650\u2009\u00b0C to allow full plastic conversion.As discussed in Section 2.4 and Section S1 (Supplementary materials), the heating power of the induction heater was selected to ensure the complete conversion of the plastic pellets in all experiments. Consequently, no plastic remained in the reactor after pyrolysis reactions. The liquid yield was calculated based on the following equation:\n\n(1)\n\n\nLiquid yield\n,\n\n\nx\n\n\nl\n\n\n\n\n\n\n%\n\n\n\n=\n\n\n\n\nm\n\n\nl\n\n\n\n\n\n\nm\n\n\np\n\n\n\n\n\u00d7\n100\n%\n\n\n\nwhere mL (g) is the liquid product mass, and mp (g) is the mass of the plastic pellet used. Similarly, the wax yield was calculated based on the mass of the wax product collected. The amount of coke formed on the catalyst was characterized using the temperature-programmed oxidation (TPO) technique described in the literature [23]. The spent catalyst (10\u2009mg) was heated to 800\u2009\u00b0C using a thermo-gravimetric analyser (TGA/DSC STAR system, Mettler) at a heating rate of 15\u2009\u00b0C/min, followed by a hold for 10\u2009min at 800\u2009\u00b0C, with an airflow rate of 100\u2009mL/min. The analysis result was used to calculate the total mass of coke formed on the catalyst. The coke yield was then calculated using the following equation:\n\n(2)\n\n\nCoke yield\n,\n\n\nx\n\n\nc\n\n\n(\n%\n)\n=\n\n\n\n\nm\n\n\nc\n\n\n\n\n\n\nm\n\n\np\n\n\n\n\n\u00d7\n100\n%\n\n\n\nwhere mc (g) is the total mass of coke formed.The light hydrocarbon compounds in the gas products were analysed using the Shimadzu 2010 GC equipped with an FID column (Equity-1 column, Supelco, L: 50\u2009m, ID: 0.53\u2009mm, stationary phase: poly(dimethylsiloxane), film thickness: 3\u2009\u00b5m). Helium gas (31\u2009mL/min) was used as carrier gas. The heating program used was: 50\u2009\u00b0C for 6\u2009min, then increased to 238\u2009\u00b0C at 12.5\u2009\u00b0C/min, and then held for 5\u2009min. The split ratio was 2.0. Both the injector temperature and detector temperature were maintained at 250\u2009\u00b0C. The abundance of hydrocarbon compounds in the gas products was reported according to the carbon numbers [24,25]. The chemical composition of the liquid products was analysed using the Thermo Scientific\u2122 Q Exactive\u2122 GC Orbitrap\u2122 GC-MS/MS System with a method adapted from the literature [26]. The GC-MS/MS system was equipped with the Thermo Scientific TG-5SILMS capillary column (internal diameter: 0.25\u2009mm, length: 30\u2009m, film thickness: 0.25\u2009\u00b5m, PN 26096\u20131420). Helium (1.2\u2009mL/min) was used as a carrier gas. The injector was operated at 280\u2009\u00b0C, with a split ratio of 25:1. The oven was initially maintained at 40\u2009\u00b0C for 2\u2009min, then heated up to 320\u2009\u00b0C at a ramp rate of 30\u2009\u00b0C/min, and finally held at 320\u2009\u00b0C for 15\u2009min. The MS was equipped with an electron ionization source (ionization voltage\u2009=\u200970\u2009eV, m/z\u2009=\u200950\u2013600). The ion source and transfer line were operated at 280\u2009\u00b0C and 150\u2009\u00b0C, respectively. The peaks in the total ion chromatogram were identified using Xcalibur Qual Browser software (Xcalibur version 4.2.47) by mass spectra searching the National Institute of Standards and Technology (NIST) Mass Spectral Search Program for the NIST/EPA/NIH EI and NIST Tandem Mass Spectral Library Version 2.3. The C7-C40 alkanes standard (1000\u2009\u00b5g/mL) purchased from Sigma Aldrich was used to facilitate hydrocarbon compound identification in the liquid products.Information on the thermal behavior of the plastics is essential to determine suitable pyrolysis process conditions. \nFig. 2a shows the TGA curves of LDPE, HDPE, and PP. In general, all the samples exhibited a one-step mass-loss process, which is a typical degradation behavior observed in other clean plastic samples [27]. The samples contain negligible moisture evidenced by the absence of a peak at \u223c100\u2009\u00b0C in the TGA plots. Complete degradation of all plastics occurred within the range of 435\u2013500\u2009\u00b0C. The decomposition peak temperatures of the plastics are observed in the following order: PP (458\u2009\u00b0C) <\u2009LDPE (471\u2009\u00b0C) <\u2009HDPE (482\u2009\u00b0C) (Fig. 2b). Although PP, LDPE and HDPE consist of long hydrocarbon chains, the presence of methyl groups in the PP polymer chain lowers its thermal stability when compared to PE samples [28]. The lower decomposition temperature of LDPE compared to HDPE is related to the higher degree of branching in the former polymer [29].GPC is another critical analysis that provides information on the MWD of the polymer samples. The MWD of the plastic samples is shown in Fig. 2c, while the numerical values of mass average (Mw), number average (Mn), and Z average (Mz) molecular masses of the samples are provided in \nTable 1. All the samples display single modal peaks, which is similar to the observations made by C\u00e1ceres et al. [30] and Zhang et al. [31]. LDPE and HDPE exhibit broader peaks, indicating wider MWD [32]. This observation is accompanied by high PI values (13.4 and 16.8 respectively, Table 1). In comparison, PP produced smaller and higher peaks, with PI values of 6.62. The Mn of the samples is arranged in the following order: LDPE (12600) <HDPE (13300) <\u2009PP (32443). These values are very different from those measured in other works [33], as neat plastics properties are often adjusted by the manufacturers for different industrial applications.The Br\u00f8nsted acidic sites in zeolitic catalysts play an important role as proton donors during the catalytic cracking of hydrocarbons [34], including plastic decomposition products [35], into smaller hydrocarbons. The Br\u00f8nsted acidic sites in zeolites originate from the protons that balance the negative charges due to the substitution of Si by Al atoms in the zeolitic framework. The abundance of Br\u00f8nsted acidic sites typically increases with aluminum content, which is represented by the Si/Al ratio of the zeolitic catalysts. When a catalyst saturated with NPA is heated, the Br\u00f8nsted acidic sites catalyze NPA decomposition into propene and ammonia at a stoichiometric ratio via the Hoffmann elimination reaction [36], enabling estimation of the Br\u00f8nsted acidic sites density on the catalyst. The strengths of the acidic sites can be estimated based on the product evolution peak temperature(s) [36]. Xian et al. [34] demonstrated different accessibility of the Br\u00f8nsted acidic sites in a zeolite sample, related to their locations in the crystal grains.As shown in \nFig. 3a, HZSM-5 exhibits three peaks at different temperatures. The peaks observed at 400\u2009\u00b0C and 460\u2009\u00b0C are attributed to the evolution of products from NPA decomposition on Br\u00f8nsted acidic sites. The peak temperatures observed in this study closely match the HZSM-5 samples used by Xian et al. [34] and Xian et al. [36] (410\u2013420\u2009\u00b0C and 470\u2013490\u2009\u00b0C respectively), as well as the HZSM-5 sample prepared by Caeiro et al. [37] (415\u2009\u00b0C and 485\u2009\u00b0C respectively). The peak at 400\u2009\u00b0C denotes the presence of stronger Br\u00f8nsted acidic sites that can catalyze NPA decomposition at a lower temperature. In contrast, the peak at 460\u2009\u00b0C represents weaker Br\u00f8nsted acidic sites that can catalyze NPA decomposition at higher temperatures. A broad peak at 280\u2009\u00b0C exists due to the desorption of physisorbed NPA from weak acidic sites, which was also observed by Xian et al. [36]. Previous works by other teams indicate that desorption peaks below 300\u2009\u00b0C represent acidic sites that are too weak to decompose NPA, and therefore unlikely to catalyze hydrocarbon cracking [38]. The integration of peak areas (excluding the peak at 280\u2009\u00b0C) revealed a Br\u00f8nsted acidic sites density of 1082.3\u2009\u00b5mol/gcat for HZSM-5 (in agreement with data reported by Losch et al. [39]), with 72.4 % contribution from the strong Br\u00f8nsted acidic sites (\nTable 2). When compared to HZSM-5, the FCC catalyst exhibits only two peaks with very low intensity. Similar to HZSM-5, the peak displayed at 212\u2009\u00b0C is attributed to the desorption of physisorbed NPA, while the peak at 416\u2009\u00b0C is related to the desorption of NPA decomposition products. No additional peak is observed. The quantification of the area under peak (416\u2009\u00b0C) indicates a Br\u00f8nsted acidic sites density of 100.5\u2009\u00b5mol/gcat, contributed solely by weak Br\u00f8nsted acidic sites. In addition to the Y zeolite, the fresh FCC catalyst also consists of binder and filler materials that improve the mechanical performance of the catalyst in the FCC process, while reducing the density of Br\u00f8nsted acidic sites per unit catalyst mass. Furthermore, the spent FCC used in this study could be poisoned by transition metal contaminants, especially vanadium and nickel, during FCC operations [40,41], rendering a great number of active sites ineffective for cracking plastic pyrolysis intermediates in this study.The textural property of a catalyst is another significant factor that influences product selectivity during the catalytic cracking of plastics [35]. The nitrogen adsorption-desorption isotherms and pore size distribution curves of HZSM-5 and FCC catalysts are shown in Fig. 3b,c. A combination of type II and type IV(a) isotherms is observed for both samples according to the updated classification standards of adsorption-desorption isotherms by IUPAC [42]. At low relative pressure (0\u2009\u2264 P/Po \u2264 0.02), the number of nitrogen molecules adsorbed increased quickly due to micropore filling and single-layer adsorption in the mesopores, leading to the convex shape of the curves. Further increase of relative pressure leads to multilayer adsorption. At P/Po \u2248\u20090.44, condensation of the adsorbate molecules occurs in the mesopores (capillary condensation). Due to the presence of macropores in the samples, adsorption saturation is not observed. The desorption curves of both samples do not coincide with their respective adsorption curves at P/Po \u2265\u20090.44, and different hysteresis loops are observed. The hysteresis loop for HZSM-5 is the H4 type, indicating a mixture of microporosity and mesoporosity dominated by narrow crack pores [43]. In contrast, an H3-type hysteresis loop is observed for the FCC catalyst indicating plate slit, crack, and wedge structures [44]. HZSM-5 possess a higher micropore area and mesopore areas (Table 2), thus a higher BET surface area (410.6\u2009m2/g) than those of FCC catalyst (147.5\u2009m2/g). Fig. 3c shows that HZSM-5 has a peak distribution below \u223c70 \u00c5, with a single peak of 38\u2009\u00c5 (3.8\u2009nm), leading to a large micropore volume (0.124\u2009cm3/g) and small mesopore volume (0.082\u2009cm3/g). The observed pore distribution characteristic is typical for HZSM-5 zeolite [45]. FCC catalyst also produces a single peak with similar intensity to HZSM-5. Fig. 3b indicates a significant proportion of HZSM-5 zeolite in such a mixture. Nevertheless, the FCC catalyst has a wider pore distribution, up to 150 \u00c5 (15.0\u2009nm), leading to a larger mesopore volume (0.131\u2009cm3/g). This is accompanied by a smaller micropore volume (0.042\u2009cm3/g). The average pore diameter based on the BJH desorption curve is 4.80\u2009nm for HZSM-5, and 11.90\u2009nm for the FCC catalyst.The crystallinity and average crystallite size of the H-ZSM5 and the FCC catalyst is examined using the XRD technique. The H-ZSM5 sample (\nFig. 4) depicts sharp peaks at 2\u03b8\u2009=\u20098.03\u00b0, 8.97\u00b0, 13.24\u00b0, 13.99\u00b0, 16.06\u00b0, 23.11\u00b0, 24.01\u00b0, and 30.24\u00b0, which are related to (101), (111), (102), (112), (022), (051), (313) and (062) planes in the X-ray diffraction pattern of the MFI topology, indicated by the Powder Diffraction File (PDF) 01\u2013079\u20131638 [46]. Similar diffraction patterns are also observed in the hierarchical ZSM-5 zeolite synthesized by Jesudoss et al. [47] and meta-promoted ZSM-5 zeolite characterized by [48]. When compared with HZSM-5 zeolite, FCC catalyst displays peaks with lower intensities. This is possibly attributed to the partial destruction of the FCC catalyst crystallinity after repeated usage in the FCC process. The diffraction lines exhibited by the FCC catalyst match the Y-type zeolite (PDF number: 01\u2013077\u20131549), while the small peaks observed at 45.8\u00b0 and 67.0\u00b0 are attributed to the characteristic peaks of Al2O3\n[49]. The average crystallite sizes of the HZSM-5 zeolite and the FCC catalyst were estimated to be 35.85\u2009nm and 18.75\u2009nm using the Debye Scherrer equation:\n\n(3)\n\n\nL\n=\n\n\n0.89\n\n\u03bb\n\n\n\u03b2\ncos\n\n\u03b8\n\n\n\n\n\n\nwhere L is the average crystallite size (nm), \u03bb is the X-ray wavelength (0.154\u2009\u00c5), \u03b8 is the Bragg diffraction angle (in radian) and \u03b2 is the full width at half maximum (FWHM) (in radian).The product yields for the thermal and catalytic pyrolysis of HDPE, LDPE, and PP are provided in \nFig. 5. Thermal pyrolysis of plastics produced high wax yield (72.4\u201373.9\u2009wt%). Two types of waxes were formed in each experimental run (regardless of the plastic type): white wax in the cold trap and yellow/green wax in the tubing that connected the reactor to the cold trap (Fig. S2, Supplementary materials). This observation signifies the following: when thermal pyrolysis occurs, the plastic vapor fills up the reactor space rapidly. Most of the vapor exited the reactor (due to the volume expansion) and condensed in the cold trap as white wax. As no nitrogen gas flows through the reactor in the experiments, the residence time of the plastic vapor in the reactor was affected by the rate of plastic conversion. The plastic vapor that remained in the reactor was further cracked to form lighter molecules, which then condensed in the tubing as yellow/green wax. This observation is also supported by the production of non-condensable gases up to a reaction time of 30\u2009min, even after the complete plastic conversion. The formation of wax products from plastics indicates the occurrence of random C-C bond scission along the polymer chain, resulting in fragments with high molecular weights. This is a typical observation for plastic pyrolysis in a continuous system with short residence time, for example, a fluidized bed reactor [50]. A longer residence time promotes extensive cracking of the intermediates, producing more gas products and fewer waxes [51]. All the wax samples displayed a single peak with narrower MWD (\nFig. 6) compared to the plastic pellets samples in Fig. 2c. This is accompanied by low PI values (1.32\u20131.46). The Mw values of the wax samples are arranged in the following order: LDPE (400) <\u2009PP (600) <\u2009HDPE (900) (\nTable 3). Assuming these waxes consisted of -CH2- groups [33], then these waxes contain chains with average carbon numbers of 28 (LDPE), 42 (PP) and 64 (HDPE). These data show that thermal pyrolysis within a short period is sufficient to reduce the molecular weight of the polymers to 1.5\u20134.5 % of their original molecular weights. When compared to the plastic pyrolysis waxes produced by Arabiourrutia et al. [33], the waxes obtained in this study have smaller Mw and PI values. The observed difference could be ascribed to the higher heating rates used in this study.The presence of catalysts in plastic pyrolysis reduced the wax yield while increasing the gas and liquid yields significantly. When HZSM-5 zeolite was used, all the plastics were converted to liquid products (24.0\u201327.2\u2009wt%) and gas products (345.0\u2013448.0\u2009cm3/g). A small number of cokes were also formed on the used zeolites, which is discussed further in Section 3.4. Despite the inferior textural properties (compared to HZSM-5) and absence of strong Br\u00f8nsted acidic sites, the FCC catalyst demonstrated catalytic activity in plastic pyrolysis within a short reaction time. Nevertheless, the catalyst was unable to convert all the heavy hydrocarbons into smaller molecules, and the wax product was mixed with the liquid product in the cold trap. These observations show that catalysts can effectively reduce the activation energy of polymer cracking, which was also observed by Zhou et al. [51] in microwave-assisted pyrolysis of plastic waste. Interestingly, plastic-type seems to influence the distribution of products during catalytic pyrolysis over FCC catalyst. LDPE produced soft wax and gas product without liquid products, while HDPE resulted in a mixture of soft wax and liquid products, together with the gas product. On the other hand, all the wax products from PP pyrolysis were converted into gas and liquid products. A different trend was observed in catalytic pyrolysis over HZSM-5, where the LDPE and HDPE showed closely similar product distribution compared to PP. Such differences could be related to the relatively weaker catalytic performance of the FCC catalyst (compared to HZSM-5), which influenced the product formation from HDPE and LDPE in different ways. More detailed investigations are necessary to fully understand the role of Br\u00f8nsted acidic sites in the catalytic pyrolysis of plastics assisted by induction heating.The application of induction heating in plastic pyrolysis using different reactor setups has been attempted by several research teams. Whajah et al. [52] reported LDPE pyrolysis over Fe3O4 mixed with Ni- or Pt-based catalysts (1:1 mass ratio for plastic to the Fe3O4/catalyst mixture) in a glass batch reactor. The Fe3O4 particles were heated to 350\u2009\u00baC, and the pyrolysis reactions lasted for 120\u2009min with 100\u2009mL/min N2 flow. Due to the small amount of heat generated by the metallic susceptor in the alternating magnetic field, a longer reaction time was needed to achieve 94 % plastic conversion, with gas yields of 15\u201380\u2009wt%, liquid yields of 2\u201343\u2009wt%, and coke yields of 1.3\u20132.4\u2009wt%. In contrast, Hassani [53] performed LDPE pyrolysis over HZSM-5 zeolite via the induction heating of a semi-batch reactor wall to 400\u2013500\u2009\u00baC, with a catalyst loading of 9\u201325\u2009wt%. The author reported gas yields of 60\u201375\u2009wt%, liquid yields of 20\u201355\u2009wt%, and coke yields of 2\u20134\u2009wt%. Similar to the aforementioned works, this study demonstrated that catalytic pyrolysis of plastics produced high gas yields and comparatively low liquid yields. Nevertheless, induction heating of the reactor wall in this study generated a large amount of heat in a short time and ensured a total plastic conversion. The short pyrolysis time could also reduce the secondary reactions, leading to lower coke yields in this study compared to Whajah et al. [52] and Hassani [53].To obtain further insights on the effects of induction heating on plastic pyrolysis behavior, the evolution rates of gas products during thermal and catalytic pyrolysis of HDPE, LDPE, and PP were recorded (\nFig. 7). As plastic pyrolysis typically leads to the generation of light hydrocarbon gas, Fig. 7 indicates the plastic decomposition rate during pyrolysis powered by induction heating. In all experiments, rapid gas evolution was observed between 3 and 5\u2009min (corresponding to the temperature of 450\u2013550\u2009\u00b0C). This observation is in line with the TGA data where plastic decomposition was observed at 450\u2013510\u2009\u00b0C (see Fig. 2a). In all experiments, the generation of gas products slowed down after the temperature reached 600\u2009\u00b0C after 7\u2009min. After 10\u2009min, complete plastic conversion into volatile products occurred. The experiments were performed without inert gas flow to enable precise observation of the product gas evolution rates in different experimental runs. During the reactions, some of the gas products that remained in the reactor underwent further cracking to produce lighter hydrocarbons leading to a further increase in the product gas volume. The secondary cracking effects could be reduced when a sweeping gas is provided to remove the pyrolysis intermediates from the reactor during plastic pyrolysis.In thermal pyrolysis (Fig. 7a), only a small amount of gas products (95\u2013132\u2009mL/g plastic) were produced, accompanied by wax formation, regardless of the plastic type (Fig. 6). Similar observations were made by Zeaiter [15], indicating the limited degree of polymer chain cracking during thermal pyrolysis. Several research teams reported that plastic pyrolysis with high heating rates (fast pyrolysis) produced more gas products, while pyrolysis with low heating rates (slow pyrolysis) produced more liquid products [54,55]. Fig. 7a also reveals higher product gas evolution (representing more extensive cracking) of PP compared to HDPE and LDPE, as PP (with a branched-chain structure and tertiary carbon) is more susceptible to C-C bond cleavage. Thus, PP can be cracked more easily than PE (linear structure with limited branching) to produce permanent gases, which is also supported by the TGA result (Fig. 2). Similar observations were also made for PP pyrolysis by resistive heating [56] and microwave heating [51].The presence of catalysts significantly increased the product gas evolutions to different degrees (Fig. 7b, c). In the presence of HZSM-5, the gas production volumes increased more than trifolds (345\u2013448\u2009mL/g plastic) when compared to thermal pyrolysis. This is related to the high catalytic property of HZSM-5, which is extensively reported in plastic pyrolysis research. The microporous structure and high acidic properties of HZSM-5, as mentioned in Section 3.2, promoted a high degree of plastic cracking, leading to the formation of liquid products and more C1-C4 products. In comparison, catalytic pyrolysis over FCC catalyst produced a lower amount of gas products (164\u2009\u2212192\u2009mL/g plastics). This is attributed to the less severe plastic cracking due to the lower acidity of the catalyst (Section 3.2).\nFig. 7b, c also reveals the lower degree of gas product evolution during catalytic cracking of PP compared to HDPE and LDPE, which is in contrast with the observation in Fig. 7a. Such observation is related to the limited contact between the PP molecules (with branched-chain structure) with the catalyst surface, leading to lower cracking and thus gas product volume. The steric hindrance effect in catalytic pyrolysis of PP was also observed by Li et al. [57] and Palza et al. [58].\n\nFig. 8 summarizes the composition of gas products from thermal and catalytic pyrolysis (in terms of carbon number), with the details of the gas components provided in Table S1 (Supplementary materials). The gas products from thermal pyrolysis are rich in C2 (12.63\u201318.27 %) and C3 (67.39\u201372.19 %) compounds. The presence of catalysts increased the abundance of C3 and C4 compounds to higher extents, at the expense of C1 and C2 compounds. The possible explanation for this observation is the reduced residence time of the plastic vapor in the reactor during catalytic pyrolysis. Due to the abundance of C3 compounds (propane and propene) and C4 compounds (butane and butene), the average molecular weight of the gas products was 41.87\u201346.67 (\nTable 4). The presence of the catalysts increased the rates of plastic decomposition and vapor formation. As the vapors exited the reactor rapidly, there was insufficient time for extensive cracking of hydrocarbon molecules into smaller hydrocarbon compounds.The GC/MS chromatograms (Figs. S2-S7) reveal a significant number of hydrocarbons in the liquid products. A tall peak that corresponds to the C6 compound in all chromatograms could be attributed to the hexane used for cleaning of autosampler of the GC instrument. Therefore, the percentage of C6 compounds in the liquid products was not analysed to prevent data misinterpretations. The presence of doublets and triplet peaks in the chromatograms signifies the presence of alkadienes, alkenes, and alkanes with similar carbon numbers in the liquid samples, which is a typical characteristic of plastic pyrolysis products [18,51]. Detailed information on the liquid product compositions is provided in Table S2 (Supplementary materials). To simplify the data analysis, the abundance of alkanes, alkenes, and aromatic compounds in the liquid products is illustrated in \nFig. 9, whereas the details of the compounds in the liquid products are provided in Table S1. A careful examination of the chromatograms shows that catalytic pyrolysis of HDPE and LDPE possessed similar product distributions, while the liquid products from PP pyrolysis showed different product characteristics. This is clear evidence of the influence of plastic structures on the product formation route during pyrolysis.The type of catalyst used in plastic pyrolysis also has an apparent influence on the liquid product compositions. Liquid products from catalytic pyrolysis of HDPE and LDPE over HZSM-5 contained hydrocarbons with carbon numbers in C7-C15, represented by the tall peaks in the chromatograms (Figs. S3\u20134). HZSM-5 produced a high amount of aromatic compounds in the liquid products from the three plastics pyrolyzed. This observation is related to the high abundance of benzene, toluene, ethylbenzene, xylene and naphthalenes. HDPE pyrolysis produces an exceptionally high percentage of aromatic compounds (93.0 %) at the expense of alkenes (5.5 %) and alkanes (1.49 %). A similar trend is observed for LDPE pyrolysis, which produced 74.8 % aromatic compounds, 18.2 % alkenes, and 7.0 % alkanes. The considerably high proportion of naphthalene in the liquid products is related to the extensive cyclization of pyrolysis intermediates catalyzed by HZSM-5 [59]. A slightly different trend is observed for PP pyrolysis, which produced considerably more alkenes (38.3 %) and alkanes (16.9 %) and fewer aromatic compounds (44.9 %) (Fig. S5). Past investigations show that the highly acidic and microporous structure of HZSM-5 zeolite promotes \u03b2-scissions of C-C bonds along the polymer chain to produce carbonium ions and olefinic compounds. The carbonium ions can undergo further \u03b2-scissions to produce smaller oligomers, followed by isomerization, oligomerization, cyclization and aromatization of the oligomers, forming a wide range of hydrocarbon compounds [60]. The bulky structures of PP molecules hinder effective contact of the molecules with the catalyst particles, leading the decreased cyclization and aromatization. Due to the high contents of aromatics (mainly C7-C8 compounds), the plastic pyrolysis liquid products over HZSM-5 possess low average molecular weight (AMM) (9.73\u201311.14).The chromatograms representing catalytic pyrolysis over FCC catalyst (Figs. S6-S8) show different trends to those over HZSM-5. All the liquid products contain C5-C35 hydrocarbons, represented by peaks with lower heights. Such observations indicate a wide product distribution. As FCC catalysts possess lower surface area and weaker acidity than those of HZSM-5, the catalyst was incapable of producing small hydrocarbons with narrow product distribution. Catalytic pyrolysis of plastics over FCC catalyst produced liquid products that contained more alkenes (38.3\u201367.6 %) and alkanes (16.9\u201325.1 %), with a small number of aromatic compounds (6.7\u20139.5 %). Catalytic pyrolysis of HDPE and LDPE produced a slightly higher amount of C19-C40 alkanes (13.4\u201315.6 % compared to PP, 7.9 %), while catalytic pyrolysis of PP produced a slightly higher amount of C9-C19 compounds (11.0 % compared to HDPE and LDPE, 8.5\u20139.7 %). Therefore, the AMM of the liquid products over the FCC catalyst are higher (21.74\u201322.49) than those over HZSM-5. No apparent difference was observed in terms of the abundance of aromatic compounds according to carbon number. These observations show that the strength and density of Br\u00f8nsted acidic sites play a more critical role than the BET surface area in the aromatization of plastic pyrolysis intermediates. The effects of plastic-type exerted minor influences on the liquid product distribution in this case.The temperature-programmed oxidation of the catalysts used in plastic pyrolysis provides qualitative and quantitative information on the amount and types of cokes formed on the catalysts. \nFig. 10 indicate different coking behaviors on the spent catalysts. For the HZSM-5 zeolite, 2\u20133 peaks were observed (in addition to the peak at \u223c100\u2009\u00b0C which signifies moisture presence). This observation indicates the formation of different types of cokes that (i) can be decomposed at lower temperatures (150\u2013330\u2009\u00b0C), and (ii) can be decomposed at higher temperatures (330\u2013760\u2009\u00b0C). In general, cokes are formed via the aromatization of hydrocarbon oligomers on Br\u00f8nsted acidic sites. Investigations on plastic pyrolysis [61] and alkene aromatization [62] demonstrated that cokes could exist in the spaces between the catalyst particles or on the catalyst surface (termed soft coke), or inside the catalyst micropores, where the Br\u00f8nsted acidic sites are located (termed hard coke). Soft coke consists mostly of oligomers and can be removed via heating in an inert environment. In contrast, hard coke is a polyaromatic hydrocarbon formed via hydrogen transfer, isomerization, cyclization, and aromatization of plastic cracking products on the Br\u00f8nsted acidic sites [63]. Due to the more complex structure of the hard coke, higher temperatures are required for its complete oxidation.For the HZSM-5 zeolites, the coke yields are 2.36 %, 2.14 %, and 2.18 % for HDPE, LDPE, and PP, respectively, concerning the mass of plastic samples pyrolyzed. For the FCC catalyst, coke decomposition produced peaks at similar positions to those observed on ZSM zeolites. Strong peaks at \u223c640\u2009\u00b0C were observed in all the FCC catalyst samples, irrespective of the plastic type. In contrast, small peaks were observed at 220\u2013240\u2009\u00b0C on the FCC catalyst used for HDPE and LDPE pyrolysis. The peak is almost non-existent for the FCC catalyst used for PP pyrolysis. The strong peak observed at \u223c640\u2009\u00b0C is attributed to the coke formed on the weak acidic sites of the FCC catalyst. The coke yields are 1.32 %, 1.38 % and 1.70 % for HDPE, LDPE, and PP, respectively. The characterization results in Section 3.2 indicate the almost non-existence of the strong acidic sites in the regenerated FCC catalyst (Table 2). The formation of coke, despite in small amount, shows that the sites played a non-negligible role during plastic pyrolysis. The difference in the coke yields observed in HZSM-5 zeolite and FCC catalysts are attributed to the higher density and strengths of Br\u00f8nsted acidic sites in the former catalyst. The high acidity of HZSM-5 is known to aromatize plastic volatiles, which leads to coke formation. The rapid deactivation of HZSM-5 in plastic pyrolysis remains an issue that requires innovative solutions [51]. A more detailed study is required to unravel the possible influence of plastic-type on the catalyst coking behavior.To illustrate the advantage of induction heating in plastic pyrolysis, the electrical energy consumed by the induction heater during plastic pyrolysis was measured using a C.A 8436 Qualistar+\u2009Power Quality analyser (Chauvin Arnoux). Measurements under different conditions show that the electrical energy consumption was independent of the masses of plastics and catalysts during pyrolysis (as heat energy was supplied in excess to the reaction system) and was only related to the induction heater power and reactor dimensions. The electrical energy consumed during 10\u2009min of plastic pyrolysis was 228\u2009kJ (\u224863.4\u2009Wh) (Fig. S9a, Supplementary materials). This value is significantly lower than those recorded by Wu et al. [64] for biomass pyrolysis, namely 665.2\u2009kJ at 550\u2009\u00b0C (with the presence of metal hollow balls) and 819.9\u2009kJ (without the presence of metal hollow balls) at 600\u2009\u00b0C. The observed difference could be attributed to the smaller size of the reactor heated in this study (reactor diameter: 2.2\u2009cm, length of the reactor to be inductively heated: 5.0\u2009cm) when compared to the reactor used by Wu et al. [64] (reactor diameter: 5.6\u2009cm, length of the reactor to be inductively heated: not specified).As plastic/biomass pyrolysis using induction heating is rarely reported in scientific literature, data on the electrical energy consumption of the process is scarce. To provide a direct comparison, electrical energy consumption during plastic pyrolysis in a fixed bed reactor using an electric furnace (based on resistive heating) was also measured using the power quality analyser. In contrast to the reactor heating profile depicted in Fig. S1, a long heating time (25\u2009min) is required to raise the reactor temperature to \u223c650\u2009\u00b0C, followed by plastic pyrolysis for 10\u2009min. These steps consumed 1225.8\u2009kJ (340.5\u2009Wh) and 270\u2009kJ (75.1\u2009Wh) respectively, resulting in total energy consumption of 1535\u2009kJ (426.4\u2009Wh) (Fig. S9b, Supplementary materials). The measurement values indicate an 85 % saving in electrical energy consumption when plastic pyrolysis is performed via induction heating over resistive heating. These values indicate the potential energy saving for plastic pyrolysis via induction heating. A similar observation was also made by Okan et al. [65], where sewage sludge pyrolysis driven by induction heating (4.62\u2009MJ/kg) consumed less energy than resistive heating (6.65\u2009MJ/kg) under similar reaction conditions. A more accurate comparison of energy consumption in plastic pyrolysis using induction heating and resistive heating will be necessary using the same reactor which can be inductively and resistively heated.In this study, the reactor was heated to a high temperature at a reaction time sufficient to ensure complete plastic conversion. Therefore, the heat energy supplied to the system was in excess compared to the theoretical energy needed to depolymerize the plastic pellets. This statement implies that the electrical energy consumption per unit mass of reactant would decrease at a higher reactant loading, as suggested by Okan et al. [65]. The energy efficiency of the plastic pyrolysis via induction heating can be improved when the potential effects of reaction parameters (frequency of the alternating magnetic field generated by the induction heater, current used for induction heater, the magnetic permeability of the reactor material etc.) which affects reactor temperature during induction heating are fully understood. Statistical process optimization is another strategy to increase the energy efficiency of the plastic pyrolysis process.In literature, different reactor designs have been proposed for plastic pyrolysis to maximize conversion and liquid/gas yields, while minimizing coke formation on catalysts [66,67]. Plastic pyrolysis in batch and semi-batch reactors exhibit satisfactory process performance at the laboratory scale and provide scientists with ample information on the process performance. Nevertheless, the development of reactor designs that allow continuous operation mode is essential to enable process scale-up. In addition, significant energy consumption and long heating time are recognized as a few of the technical challenges in plastic pyrolysis. Microwave-assisted pyrolysis is proven to be an excellent strategy to overcome the aforementioned challenges, as the heat energy can be rapidly produced from metallic susceptors inside the reactor vessel [68], which minimizes the heating/cooling time and energy losses to the surroundings. In addition, the heat transfer limitation can be minimized, as the susceptors are mixed with the plastic feed. Due to the incompatibility with the stainless-steel reactor, microwave-assisted pyrolysis is typically performed in a specialized reaction system, which contains a reactor vessel made of microwave-transparent materials (quartz/ceramic). The limited mechanical strengths of these materials restrain the prospects of microwave-assisted pyrolysis in the plastic recycling sector, especially when considering the possible pressure fluctuations due to the volatile gases released during the process. This study presents the possibility to conduct plastic pyrolysis in a stainless-steel fixed-bed reactor in an alternating magnetic field. The reactor was heated up to 650\u2009\u00b0C within 10\u2009min, which was sufficient for complete plastic conversion. Due to the short residence time of plastic pyrolysis intermediates in the reactor heating zone, thermal pyrolysis produced mainly wax products with a small amount of hydrocarbon gas. The addition of zeolitic catalysts significantly improved the yields of gas and liquid products at the expense of the wax product. The product yields and compositions, as well as the catalyst deactivation behaviors, can be explained with references to past investigations on plastic pyrolysis using resistive heating and microwave heating. This exploratory study demonstrates the feasibility of plastic pyrolysis via induction heating of a reaction vessel, which is comparable to microwave heating in terms of heating/cooling time. The use of a stainless-steel reactor allows possible pressure variation during the process, hence increasing the safety margin when considering the process upscaling potential. The avoidance of metallic susceptors in the plastic feed could lead to a more convenient treatment of solid residue in the reactor after plastic pyrolysis. These findings also demonstrate the high flexibility of the inductively heated plastic pyrolysis system in plastic waste valorization into various products, depending on the desired end-uses and market value of these products. It is believed that the application of induction heating can be extended to various existing reactor designs (auger reactor [69], conical spouted bed reactor [70], vertical falling film reactor [71], among others) and process designs (pyrolysis, gasification [72], hydrothermal processing [73] among others), which is an interesting option to improve energy efficiency while retaining the unique advantages related to these reactor/process designs. Nevertheless, more fundamental studies are needed for a comprehensive understanding and control of inductively heated plastic pyrolysis. The lack of data on the yields of H2, as well as C5 and C6 compounds in plastic pyrolysis products in this study, is a significant shortcoming that will be tackled in future studies.The study examined the feasibility of induction heating in the thermal and catalytic pyrolysis of HDPE, LDPE, and PP plastics using two HZSM-5 (MFI topology) and FCC (FAU topology) catalysts, which has not been explored in the literature. In all the experimental runs, the plastic conversion started within 3\u20135\u2009min of reaction time, and complete plastic conversion was observed within 10\u2009min. This observation corresponded to the rapid reactor temperature increase upon induction heating. Thermal pyrolysis produced a significant amount of wax (72.4\u201373.9\u2009wt%) and a small amount of non-condensable gas products (rich in C3 followed by C2 compounds). The presence of the catalysts significantly enhanced the gas (70.6\u201373.9\u2009wt% for HZSM-5 and 62.4\u201375.2\u2009wt% for FCC catalyst) and liquid yield (24.0\u201327.2\u2009wt% for HZSM-5 and 0\u201335.9\u2009wt% for FCC catalyst), at the expense of wax yield (0\u2009wt% for HZSM-5 and 0\u2009\u221225.4\u2009wt% for FCC catalyst). The gas products for the catalytic pyrolysis were rich in C3 followed by C4 compounds, regardless of the plastic and catalyst properties. On the other hand, a significant variation in terms of liquid product distribution is observed for catalytic pyrolysis. HZSM-5 zeolite produced liquid products with high aromatics content, especially C7-C10 aromatics (representing toluene, ethylbenzene, xylene, naphthalene and several alkylbenzene isomers). The FCC catalyst produced a high number of liquid products, which were richer in alkenes and alkanes in the C9-C40 range. The outlined differences are attributed to the different textural properties and acidity profiles of the catalysts, which led to different reaction pathways during plastic pyrolysis. These observations validate the ability of the catalysts in product distribution modification despite the short residence time. Polymer types also influenced product yields and distribution, in agreement with past investigations in plastic pyrolysis. Coke analysis revealed the formation of cokes (2.14\u20132.38\u2009wt% for HZSM-5 and 1.32 \u2013 1.70\u2009wt% for FCC catalyst) of different strengths, corresponding to the different acidic sites on the catalysts. Based on the similar decomposition peaks of these cokes in HZSM-5 and FCC catalysts, it is hypothesized that the cokes were formed on the strong and weak acidic sites of these zeolitic catalysts.To conclude, this exploratory study demonstrates the feasibility of new pyrolysis powered by induction heating in the production of different value-added products from plastics. The adoption of induction heating could be an interesting strategy that reduces the time needed for plastic waste pyrolysis, especially at a larger scale. In addition, the prospects of applying induction heating on stainless-steel reactors also provide an interesting scaling-up strategy that allows possible pressure variation. As induction heating is also known to have higher energy conversion efficiency than resistive heating, future studies will be dedicated to the quantification of energy consumption in plastic pyrolysis assisted with induction heating.\nSyie Luing Wong: Investigation, Methodology, Writing \u2013 original draft. Sabino Armenise: Conceptualization, Validation. Bemgba Bevan Nyakuma: Writing \u2013 reviewing and editing. Anna Bogush: Formal analysis, Data curation. Sam Towers: Formal analysis, Data curation. Chia Hau Lee: Software, Resources. Keng Yinn Wong: Software, Resources. Ting Hun Lee: Software. Evgeny Rebrov: methodology, Supervision. Marta Mu\u00f1oz: Supervision, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Syie Luing Wong and Sabino Armenise have received support from the European Union\u2019s Horizon 2020 research and innovation program under the Marie Sk\u0142odowska-Curie grant Agreement No. 754382, GOT ENERGY TALENT. The content of this publication does not reflect the official opinion of the European Union. Responsibility for the information and views expressed in this paper lies entirely with the authors. Marta Mu\u00f1oz also gratefully acknowledges the financial support from the Comunidad de Madrid (S2018/NMT-4411) for a project titled \u201cAdditive Manufacturing: from Material to Application.\u201d Syie Luing Wong wishes to thank John Pillier, Li SiRui, Joe Gregory, and Deema Kunda for all the guidance and assistance provided during his research associated with the School of Engineering, University of Warwick, UK. Syie Luing Wong and Sabino Armenise also acknowledge the contributions of Carlos Prieto and the FCC team from CEPSA to this study in terms of materials, technical analysis, and intellectual discussions.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jaap.2022.105793.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  Plastic pollution compromises the environment and human well-being, and a global transition to a circular economy of plastics is vital to address this challenge. Pyrolysis is a key technology for the end-of-life recycling of plastics, although high energy consumption limits the economic feasibility of the process. Various research has shown that the application of induction heating in biomass pyrolysis reduces energy consumption when compared to conventional heating. Nevertheless, the potential of induction heating in plastic pyrolysis is rarely explored. This paper presents an exploratory study on the thermal and catalytic pyrolysis of high-density polyethylene, low-density polyethylene, and polypropylene in a fixed bed reactor through induction heating. An MFI-type HZSM-5 zeolite (SiO2/Al2O3 =\u00a023) and an FAU-type spent fluid catalytic cracking (FCC) catalyst with distinctive Br\u00f8nsted acidity and textural properties were used. A complete conversion of the plastic feedstocks was achieved within 10\u00a0min, even without a catalyst. Thermal pyrolysis produced wax (72.4\u201373.9\u00a0wt%) and gas products, indicating a limited degree of polymer cracking. Catalytic pyrolysis over HZSM-5 and FCC catalyst significantly improved polymer cracking, leading to higher gas (up to 75.2\u00a0wt%) and liquid product (up to 35.9\u00a0wt%) yields at the expense of wax yield (up to 25.4\u00a0wt%). In general, the gas products were rich in C3 and C4 compounds. The liquid product composition was highly dependent on the catalyst properties, for example, the HZSM-5 produced high aromatics, while the FCC catalyst produced high alkenes in the liquid products. The catalyst acidity and textural properties played an essential role in plastic pyrolysis within the short reaction time. This study demonstrated the feasibility of a fast, energy-efficient, and versatile plastic valorization technology based on the application of induction heating, where the plastic feed can be converted into wax, gas, and liquid products depending on the end-use applications.\n               "}
{"full_text": "Hydrogen peroxide (H2O2), as a green and environment-friendly oxidant, is widely used in chemical, biological, environmental, and other industries. Currently, H2O2 production is mainly based on the anthraquinone autoxidation (AO) process. However, the AO process has complicated processes, high production costs, and an unfriendly environment, and the safety risks of high-concentration hydrogen peroxide during transportation, storage, and use cannot be ignored (Fukuzumi et al., 2021). The direct synthesis of H2O2 from H2 and O2 has broad application prospects because of its environmental-friendly, economic efficiency, ready-to-use, and only water as a by-product (Huynh et al., 2021; Lewis and Hutchings, 2018).Although H2O2 directly synthesizing has excellent potential, the formation of H2O2 and H2O from H2 and O2, the hydrogenation and decomposition of H2O2 are both thermodynamically favorable (Han et al., 2021), which makes the preparation of high-performance hydrogen peroxide catalysts more difficult. Many experimental and theoretical studies (Han et al., 2021; Agarwal et al., 2021; Tian et al., 2017; Hang and Chung, 2020; Lari et al., 2017; Lyu et al., 2019; Priyadarshini et al., 2021) were conducted to obtain high-performance hydrogen peroxide catalysts. Palladium is the practical active catalyst for H2O2 directly synthesizing. The Langmuir-Hinshelwood reaction mechanism and proton-electron transfer reaction mechanism are the two main reaction mechanisms for directly synthesizing hydrogen peroxide (Han et al., 2021). These two mechanisms agree for oxygen adsorption but diverge for hydrogen adsorption and dissociation. According to the Langmuir-Hinshelwood reaction mechanism, hydrogen is dissociated on the surface of Pd to form H* (* indicates adsorption on the catalyst surface), and then H* and O2* to form \u2013OOH* intermediate. In comparison, the proton-electron transfer reaction mechanism involves hydrogen dissociated on the surface of Pd to form protic hydrogen (H+) and electrons. Electrons are transferred to O2* by the catalyst, and then the proton H+ reacts with O2* to form \u2013OOH* intermediate. The DFT study concluded that both reaction mechanisms are inverse thermodynamically feasible (Agarwal et al., 2021). Although the mechanism of hydrogen peroxide generation is still somewhat controversial, the role of palladium as an active catalytic center is recognized; therefore, the modulation of palladium can directly affect the synthesis of hydrogen peroxide and ultimately obtain a high-performance catalyst. Changing the catalyst's synthesis process and heat treatment conditions could change palladium's geometry and electronic structure (such as particle size, crystal plane, and Pd0/Pd2+ ratio), which are conducive to improving H2O2 selectivity and production rates (Tian et al., 2017; Hang and Chung, 2020; Lari et al., 2017; Lyu et al., 2019; Priyadarshini et al., 2021; Han et al., 2017). Alloying Pd with various metals (e.g., Au, Ag, Zn, Sn, Co, Ni) (Li et al., 2018; Ricciardulli et al., 2021; Doronkin et al., 2020; Gu et al., 2016; Wilson et al., 2018; Wilson et al., 2018; Kanungo et al., 2019; Nazeri et al., 2021; Tian et al., 2017; Kazici et al., 2017; Zhang et al., 2021; Lee and Chung, 2020; Maity and Eswaramoorthy, 2016; Zhang et al., 2018) enhance H2O2 selectivity by changing the geometric and electronic structure of palladium.Support modification is another important strategy to improve the catalytic performance of Pd catalysts. The surface properties of the support can adjust H2O2 productivity and H2O2 selectivity by affecting the morphology, dispersion, electronic state and alloying degree of the Pd active metal. Support regulation of oxide (SiO2, Al2O3, TiO2, etc.) and carbon materials (Lewis and Hutchings, 2018; Vu et al., 2021; Liang et al., 2020; Edwards et al., 2014; Edwards et al., 2009; Piccinini et al., 2012; Villa et al., 2016; Hu et al., 2014; Gudarzi et al., 2015; Gudarzi et al., 2015; Thuy Vu et al., 607 (2020).; Garc\u00eda et al., 2015) (activated carbon, ordered mesoporous carbon, CNT, etc.) are common in H2O2 directly synthesizing. Hutchings (Edwards et al., 2009; Ntainjua et al., 2008) and co-workers found that porous carbon supports were the best supports for Pd and Au-Pd catalysts because of the superior catalytic activity and lowest rate of H2O2 hydrogenation and decomposition side reactions. In the follow-up research works, the team further found that nitric acid pretreatment of carbon supports for Au-Pd catalysts could switch off the sequential H2O2 hydrogenation and decomposition side reactions compared with TiO2, ZrO2, CeO2 and other oxide supports, making porous carbon become a significant support for the direct synthesis of H2O2 from H2 and O2.\nN-doped is an effective modification method for porous carbon, which could change the surface properties and pore structure and provide plentiful chemically active sites (Wei et al., 2018). Perathoner (Abate et al., 2010) and co-workers found that the Pd-based catalysts supported on N-doped CNTs showed higher catalytic activity and turnover, and the improvement of this activity may be related to the electronic effect of N sites in the support. Rosa Arrigo (Arrigo et al., 2014; Arrigo et al., 2016) and co-workers believed that the N functional groups on the surface of carbon support could form a robust Pd-N bond with Pd, which could not only inhibit the agglomeration of Pd nanoparticles but also provide electrons to Pd, thus improving the stability and activity of Pd catalyst. Ji (Ji et al., 2021) et al. found that an appropriate concentration of N functional groups (2.72 at.%) could maintain the electron-deficient state of Pd, which was conducive to improving H2O2 productivity.\nN-doped porous carbon was concerned with potential support for the direct synthesis of H2O2 from H2 and O2. However, the current experimental studies cannot well explain the effect of N functional groups on the synthesis of H2O2. To further explore the influence of N functional groups on H2O2 directly synthesizing, the method of theoretical calculation relies on the following: (i) constructing Pd/C heterojunctions with different nitrogen contents; (ii) investigating the effect of N-doping on the formation energy and electron transfer of Pd/C heterojunctions; (iii) investigating the effect of N-doping on O2 and H2O2 adsorption, and OO bond dissociation. Theoretical calculations showed that N-doping could reduce the agglomeration degree of Pd nanoparticles and facilitate electron transfer from carbon support to Pd nanoparticles, which is conducive to the generation of H2O2. However, N-doping reduced the adsorption energy of O2 and H2O2, the active dissociation energy barrier of O2, which is detrimental to the desorption of H2O2 and the inactive dissociation of O2, reducing the selectivity of H2O2. In order to reduce the adverse effects of N-doping, N-doped porous carbons with different pore structures were constructed to accelerate the mass transfer and diffusion performance of Pd/C catalyst. The experimental results achieved the expected goal. Compared to other catalysts, the Pd/NPCs-PSS with a highly developed multistage pore structure showed excellent catalytic performance for the direct synthesis of H2O2 from H2 and O2, the H2O2 productivity and selectivity reached 328.4 molH2O2\u00b7kgcat\n-1\u00b7h\u22121 and 71.9\u00a0%, respectively.The periodic density functional theory (DFT) calculations of the present work were performed by the CP2K package (Kuhne et al., 2020) with the spin-polarized gradient corrected functional of Perdew Burke and Ernzerhof (PBE) (Perdew et al., 1996). The Goedecker\u2013Teter\u2013Hutter (GTH) pseudopotentials, DZVP-MOLOPT-GTH basis sets were utilized to describe the core electrons and valence electrons (VandeVondele and Hutter, 2007; Goedecker et al., 1996) (4s24p64d10 for Pd, 2s22p4 for O, 2s22p2 for C, 2s22p3 for N), respectively. 400Ry was set to plane-wave energy cutoff (Lippert and Parrinello, 2010). The electron and force convergence criteria were 1\u00a0\u00d7\u00a010\u22126 a.u. and 6\u00a0\u00d7\u00a010\u22124 Hartree Bohr, respectively. The transition state of reaction paths was performed using the climbing imagen nudged-elastic-band method (CI-NEB) (Henkelman et al., 2000), including 6 replicas, and the maximum force was 1\u00a0\u00d7\u00a010-3 atomic units. The transition states had been confirmed by vibrational analysis. The 3\u00a0\u00d7\u00a03\u00a0\u00d7\u00a01 Monkhorst\u2013Pack mesh k-points were used for work function calculation. The heterojunction of Pd/graphite consisted of 4 layers of graphite and 2 layers of Pd(111) crystal plane. We use VASPKIT (Wang et al., 2021) software 804 functions to automatically generate heterojunction with optimized p(1\u00a0\u00d7\u00a01) graphite and Pd(111)-p(1\u00a0\u00d7\u00a01). The mismatch tolerance of heterojunction is less than 1\u00a0%. The interlayer space is set to 2.5\u00a0\u00c5. The vacuum space was set to 17\u00a0\u00c5 along the Z-axis. The final periodic box of Pd/graphite is 9.9\u00a0\u00d7\u00a09.9\u00a0\u00d7\u00a039.5\u00a0\u00c5. The generated heterojunction model needs to be further optimized before being used. The detailed position information of Pd/graphite heteroatoms is shown in the Supporting Information (Table S5). We replaced carbon atoms with nitrogen atoms in Pd/graphite to build the nitrogen-doped graphite. The adsorption energy (Eads) of O2 over the Pd/graphite (Pd/N-graphite, Pd/2N-graphite) was calculated as below:\n\n\n\n\n\nE\n\n\na\nd\ns\n\n\n=\n\n\nE\n\n\nO\n2\n/\ns\nl\na\nb\n\n\n-\n\n\nE\n\n\ns\nl\na\nb\n\n\n-\n\n\nE\n\n\nO\n2\n\n\n\n\n\n\nWhere EO2/slab is the total energy of the Pd/graphite (Pd/N-graphite, Pd/2N-graphite) slab with the adsorbed atom(O2); Eslab is the total energy of Pd/graphite (Pd/N-graphite, Pd/2N-graphite) slab, and EO2 is the energy of the O2 in a box of 9.9\u00a0\u00c5 \u00d7 9.9\u00a0\u00c5 \u00d7 39.5\u00a0\u00c5.The interface formation energy (EF) of the Pd/graphite (Pd/N-graphite, Pd/2N-graphite) was calculated as below:\n\n\n\n\n\nE\n\n\nF\n\n\n=\n\n\nE\n\n\nt\no\nt\na\nl\n\n\n-\n\n\nE\n\n\ns\nu\np\np\no\nr\nt\n\n\n-\n\n\nE\n\n\nP\nd\n(\n111\n)\n\n\n\n\n\n\nWhere Etotal is the total energy of the heterostructure (Pd/graphite, Pd/N-graphite or Pd/2N-graphite). Esupport is the energy of graphite, N-graphite or 2\u00a0N-graphite, EPd(111) is the energy of Pd(111).The work functions (\u03a6) of the graphite, N-graphite, 2\u00a0N-graphite and Pd(111) surfaces according to the following equation:\n\n\n\n\u03a6\n=\n\n\nE\n\n\nv\na\nc\n\n\n-\n\n\nE\n\n\nF\n\n\n\n\n\n\nWhere EF is the fermi energy, and Evac is the electrostatic potential of the vacuum level.The activation barriers and reaction energy were calculated as below:\n\n\n\nEa\n=\n\n\nE\n\n\nT\nS\n\n\n-\n\n\nE\n\n\nI\nS\n\n\n\n\n\n\n\n\n\n\n\u0394\nE\n=\n\n\nE\n\n\nF\nS\n\n\n-\n\n\nE\n\n\nI\nS\n\n\n\n\n\n\nEIS, EFS and ETS were the energy of the initial state, final state and transition state.We used the VASPKIT code (Wang et al., 2021), Multiwfn, (Lu and Chen, 2012) VESTA (Momma and Izumi, 2011) and VMD (Humphrey et al., 1996) for post-processing of the CP2K calculated data.Palladium chloride (PdCl2, 59\u201360\u00a0%), were purchased from Shanghai Aladdin Biochemical Technology Co., ltd; Sodium polyacrylate (MW\u00a0=\u00a0450\u2013700\u00a0W, PAANa), Melamine (C3H6N6, 99\u00a0%), potassium bicarbonate (KHCO3, 99.5\u00a0%), Sodium polystyrene sulfonate (MW\u224870000, PSS), and Diallyldimethylammonium chloride (PDDA, 60\u00a0%) were purchased from Shanghai Macklin Biochemical Co., ltd; glucose (C6H12O6, 98\u00a0%), sodium borohydride (NaBH4, 98\u00a0%), ammonium hydroxide (NH3\u00b7H2O, 25\u201328\u00a0%) and hydrochloric acid (HCl, 36.0\u201338.0\u00a0%) were purchased from Sinopharm Chemical Reagent Co., ltd; ethanol (C2H5OH, 99.7\u00a0%) and methanol (CH3OH, 99.5\u00a0%) were purchased from Tianjin Fuyu Fine Chemical Co., ltd. All reagents were used as received without further purification. Deionized water was used throughout the experiment process.The porous carbon (PCs, NPCs, NPCs-PAANa, and NPCs-PSS; the experimental details for porous carbon are given in the Supporting Information) supported Pd catalysts were synthesized by an absorber-reduction method (Fig. 1\n). In a typical procedure, the porous carbon (100\u00a0mg, NPCs-PSS) was added to deionized water (50\u00a0ml) on a round flask; after ultrasonication for 30\u00a0min, a solution of PDDA (0.5\u00a0ml) was added to the mixture and stirred at 35\u2103 for 24\u00a0h. The modified porous carbon was obtained by centrifugation (10000\u00a0rpm/min). The modified porous carbon was added to deionized water (50\u00a0ml) in a round flask; after ultrasonication for 10\u00a0min, a solution of H2PdCl4 (0.5\u00a0ml, 0.075\u00a0M) was added to the mixture and stirred at 35\u2103 for 24\u00a0h. After impregnation, a fresh sodium borohydride solution (1\u00a0ml, 10\u00a0% wt) was added to the mixture, and the mixture was stirred at 35\u2103 for 5\u00a0h to reduce Pd entirely. The solid was obtained by centrifugation (10000\u00a0rpm/min) and dried in air at 80\u2103 overnight. The dried solid was placed in a tube furnace and treated in the air at 250\u2103 for 2\u00a0h, followed by H2 reduction at 250\u2103 for 2.5\u00a0h, The sold powder obtained after reduction was named Pd/NPCs-PSS. The experimental details for other catalysts are given in the Supporting Information. Elemental Analyzers, Fourier Transform Infrared Spectrometer (FTIR), Specific surface area and porosity analyzer (ASAP 2460), X-ray diffraction (XRD), Inductively coupled plasma-optical emission spectroscopy (ICP-OES), Raman Spectrometer, scanning electron microscopy (SEM), transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) were used to investigate the structure and surface properties of porous carbon and catalyst. The details of the instruments are provided in the Supporting Information.The first-principles calculations based on density functional theory (DFT) were used to investigate the reaction energetics and atomic interactions within the PdNC metal\u00a0\u2212\u00a0carbon heterointerface. The calculations were based on the Pd(111) crystal plane. Compared with other crystal planes, the Pd(111) crystal plane is considered the most favorable for the generation of H2O2 (Tian et al., 2013). The carbon supports were simplified as graphite, and the doped N atoms exist in graphitic nitrogen. The Pd/C heterojunctions comprised 4 layers of graphite and 2 layers of Pd(111) crystal plane; furthermore, C atoms on the surface of graphite adjacent to the Pd(111) crystal plane were replaced with N atoms to simulate the PdNC heterojunction. The specific model structures are shown in Fig. S6. The thermodynamic stability of heterojunctions can be described by the interface formation energy (EF). The interface formation energy (EF) of Pd/graphite, Pd/N-graphite, and Pd/2N-graphite were\u00a0\u2212\u00a00.19\u00a0eV, \u22120.81\u00a0eV and\u00a0\u2212\u00a00.97\u00a0eV (Table S3), respectively. The negative formation energy indicated that the Pd/graphite, Pd/N-graphite, and Pd/2N-graphite could form a stable interface; the Pd/N-graphite and Pd/2N-graphite with lower formation energy were conducive to the formation of small Pd nanoparticles, the small Pd nanoparticles were benefited hydrogen peroxide generation (Tian et al., 2017).The work function of the surface is a crucial parameter to studying charge transfer at the heterojunction interface. The work functions of the graphite, N-graphite, 2\u00a0N-graphite and Pd(111) surfaces were 4.53\u00a0eV, 3.72\u00a0eV, 3.68\u00a0eV and 5.07\u00a0eV (Fig. S7), respectively. The work function of graphite and Pd(111) surfaces agree with the experimental values of 4.6\u00a0eV and 5.3\u00a0eV (Oshima and Nagashima, 1997), respectively. As shown in Fig. S6, the fermi energy of Pd(111) surface was lower than graphite, N-graphite and 2\u00a0N-graphite; when those supports form a heterojunction structure with Pd(111) surface, electrons will flow from supports to Pd(111) surface until there have achieved equilibrium in the fermi energies. The interactions and electronic structure between Pd(111) surface and supports were investigated by calculating differential charge density. Compared with Pd/graphite, the electron density of C atoms on the surface of Pd/N-graphite and Pd/2N-graphite disappears, and the electrons are concentrated around Pd and N atoms (Fig. 2\n). It can be inferred that the graphitic nitrogen in the carbon support promotes the release of electrons from the support to Pd(111) layer, thereby exhibiting an electron transfer effect.Bader charge analysis was performed on the Pd/graphite, Pd/N-graphite and Pd/2N-graphite heterojunction models to quantify charge transfer at the interfaces. Bader change analysis results show that the amount of charge transfer from graphite, N-graphite and 2\u00a0N-graphite to the Pd(111) surface were 0.02 e, 0.49 e and 0.48 e, respectively. The plane-average electron difference and the electron displacement curve (Fig. S8) give the same conclusion. The transfer of electrons from N-graphite (2\u00a0N-graphite) to Pd(111) was conducive to maintaining a high concentration of Pd0 sites on the surface of Pd(111). Pd0 active site could promote the dissociation of H2 and the activation of O2, which was beneficial to hydrogen peroxide generation. However, excess Pd0 active sites will accelerate the dissociation of the OO bond in O2 and H2O2, resulting in decreased H2O2 selectivity.In the reaction mechanism of the direct synthesis of H2O2 from H2 and O2, the adsorption and activation of molecular oxygen on the surface of Pd were significant steps (Nugraha et al., 2017). The adsorption energy of O2 on the surface of Pd/2N-graphite, Pd/N-graphite and Pd/graphite were\u00a0\u2212\u00a00.86\u00a0eV, \u22120.83\u00a0eV and\u00a0\u2212\u00a00.82\u00a0eV (Table S3), respectively; indicating that the Pd supported on the N-doped carbon surface is beneficial to the adsorption of O2. The OO bond length of O2 after adsorption ranged from 1.38 to 1.39\u00a0\u00c5 (Table S3) and was higher than in the gas phase. According to the differential charge density figure (Fig. S9), we can indicate that the electrons of the adsorbed O2 molecules are transferred from Pd atoms and OO bond to the oxygen atoms, and weakens the OO bond. The d-Projected density of states (PDOS) of Pd (Fig. S10) were calculated to investigate the influence of N-doped on the d-band center of Pd. It is found that with the increase of nitrogen content, the d-band center of Pd was closer to the Fermi level (Fig. S10); according to the d-band center theory, the closer the d-band center of the metal catalyst to the Fermi level, the higher the metal surface activity and the stronger the adsorption of reactants. As the d-band center of Pd approaches the Fermi level, the adsorption energy of O2 and H2O2 decreases from\u00a0\u2212\u00a00.82\u00a0eV to\u00a0\u2212\u00a00.86\u00a0eV and\u00a0\u2212\u00a00.38\u00a0eV to\u00a0\u2212\u00a00.40\u00a0eV (Fig. 3\na-b), respectively, the lower adsorption energy was unfavorable for the desorption of O2 and H2O2. As shown in Fig. 4\n, the activation barriers (Ea) of O2 dissociation on Pd(111) surface decreased from 0.53\u00a0eV (Pd/graphite) to 0.49\u00a0eV (Pd/N-graphite) and 0.48\u00a0eV (Pd/2N-graphite). As mentioned above, N doping could reduce oxygen adsorption and dissociation energy (Ea) on Pd(111) surface, making oxygen more prone to active dissociation. In the reaction mechanism of hydrogen\u2013oxygen synthesis of hydrogen peroxide, the active dissociation of O2 is not conducive to the generation of hydrogen peroxide. The reactive dissociation of O* reacts with H* to produce the by-product H2O, resulting in the loss of intermediates, which should be avoided as much as possible in the direct hydrogen\u2013oxygen synthesis of hydrogen peroxide.The theoretical calculations showed that N doping could reduce the formation energy of Pd/graphite heterojunction; promote the transfer of electrons from carbon supports to Pd particles, which are beneficial to the generation of Pd nanoparticles with a high Pd0/Pd2+ ratio and small particle size. Smaller Pd nanoparticles and a higher proportion of Pd0 have a favorable effect on H2O2 production. However, N doping reduces the adsorption energy of O2 and H2O2, and the dissociation activation barriers of O2*, which is not conducive to improving H2O2 selectivity.In order to suppress the adverse effect of N doping on H2O2 synthesis, a series of Pd/C catalysts were fabricated by adjusting the carbon support pore structure. Firstly, N-doped porous carbon with high specific surface area and hierarchical pore structure were prepared by one-step multiple activation synthesis techniques. The different precursors produced by the hydrothermal treatment of glucose have undergone a one-step multiple activation process, in which potassium bicarbonate was the pore-forming agent, and melamine was the pore-forming agent and source of nitrogen, and simultaneously realized the pore-forming activation and nitrogen doping. Then, the Pd nanoparticles were loaded on porous carbon supports through an absorber-reduction strategy. Finally, the stable Pd/C catalyst was obtained by the further oxidation\u2013reduction heat treatment process.The morphology and size distribution of porous carbon are shown in Fig. 5\n. By adjusting the hydrothermal treatment process, the activated hydrothermal carbon has different morphologies and structures. The PCs and NPCs were composed of agglomerated carbon microspheres with an average particle size of 310\u00a0nm (Fig. 5a-b, Fig. S1a-b). The NPCs-PAANa prepared by adding trace PAANa dispersants were composed of mono-dispersed porous carbon spheres with an average particle size of 437\u00a0nm (Fig. 5c, Fig. S1c) (Gong et al., 2014). The NPCs-PSS prepared by adding trace PSS dispersant were formed by carbon spheres with particle sizes less than 100\u00a0nm and a rough surface(Fig. 5d, Fig. S1d) (Gong et al., 2014). All of the activated hydrothermal carbon surfaces have obvious irregular porous structures.The pore structure of porous carbons were analyzed by nitrogen adsorption\u2013desorption isotherm at 77\u00a0K (Fig. 6\na-b). The detailed data are displayed in Table 1\n. According to the IUPAC classification, the isothermal adsorption\u2013desorption curve of PCs showed type \u2160, indicating that it was dominated by microporous (Guo et al., 2020). The isothermal adsorption\u2013desorption curves of NPCs and NPCs-PAANA were type \u2163, and the nitrogen adsorption capacity increased sharply at low relative pressure P/P0\u00a0less than\u00a00.02 of the N2 sorption isotherms (Fig. 6a), indicating the presence of microporous, the hysteresis loop at relative pressure P/P0 of 0.4\u20130.8 can be classified as H2 hysteresis, indicating the presence of mesoporous (Rio et al., 2020). Similar to NPCs-PAANa and NPCs, NPCs-PSS also has H2 type hysteresis loop, indicating that NPCs-PSS also has a mesoporous structure. However, when the relative pressure P/P0 of 0.8\u20131.0, NPCs-PSS still has a surge in adsorption capacity, indicating stacking channels in NPCs-PSS (Gong et al., 2014), which was consistent with the SEM image (Fig. 5d). The specific surface area of porous carbons were calculated by Brunaure-Emmett-Teller (BET) model (Table 1). The specific surface area of PCs, NPCs NPCs-PAANa and NPCs-PSS were 1770\u00a0m2/g, 3421\u00a0m2/g, 3165\u00a0m2/g and 3173\u00a0m2/g, respectively. The difference in specific surface area between PCs and NPCs indicates that the volatile gas generated by high-temperature decomposition of melamine will etch hydrothermal carbon aromatic carbon skeleton, forming more pore structures and larger specific surface area (Li et al., 2019; Xu et al., 2020). The pore volumes of porous carbon were calculated by Barrett-Joyner-Halenda (BJH) model. The total pore volumes and the Vmic/Vtotal ratio of PCs, NPCs, NPCs-PAANa and NPCs-PSS were 0.99\u00a0cm3/g, 2.16\u00a0cm3/g, 2.44\u00a0cm3/g and 3.43\u00a0cm3/g, 0.79, 0.36, 0.31 and 0.36, respectively. The Vmic/Vtotal ratio of porous carbon indicates that the NPCs, NPCs-PAANA and NPCS-PSS have microporous and mesoporous structures. Fig. 6b shows that all of the porous carbons have micropores and mesoporous structures; micropores dominate the pore structure of PCs. In contrast, NPCs and NPCs-PAANa are dominated by mesoporous with a pore size of 2\u20135\u00a0nm. Besides the similar microporous and mesoporous structures as NPCs and NPCs-PAANa, NPC-PSS has stacked pores with pore diameters of 10\u2013100\u00a0nm. The rich mesoporous structure of porous carbon facilitates the mass transfer between catalysts and reactants (Yook et al., 2016).BET analysis showed that the adjusted glucose hydrothermal process combined with one-step multi-activation strategy successfully prepared the high specific surface area and high pore capacity carbon supports with micropore-mesoporous and micropore-mesoporous- macroporous multistage pore structures.The compositions of porous carbons were analyzed by elemental analyzer and XPS. XPS was used to detect the surface composition of the sample. The elemental analyzer was used to determine the contents of C, H, and N in the bulk phase, and the O content was calculated by the difference method. As shown in Table 2\n, the nitrogen contents of PCs, NPCs, NPCs-PAANa and NPCs-PSS obtained by elemental analyzer were 0.00\u00a0%, 0.85\u00a0%, 0.74\u00a0% and 0.84\u00a0% wt%, respectively, consistent with XPS determination, indicating that nitrogen atoms were successfully doped in the surface of porous carbon (Li et al., 2019). The N content of porous carbon obtained by XPS was consistent with the N-graphite in the theoretical calculation model.The samples were subjected to X-ray diffraction (XRD) to analyze the characteristics of carbon crystal structure. As shown in Fig. S12, The diffraction peak of PCs at 2\u03b8 values of 44\u00b0 and 24\u00b0 can be assigned as the (100) reflection of graphite and (002) reflection of amorphous graphite, respectively (Cai et al., 2020). The intensity of (002) peaks of the NPCs, NPCs-PAANa and NPCs-PSS were lower than the PCs, indicating that the N-doped damage the crystal structure of graphite and reduces the graphitization degree. The same conclusion can be obtained from the Raman spectra. As shown in Fig. 7\n, there are two prominent peaks at\u00a0\u223c\u00a01343\u00a0cm\u22121 and\u00a0\u223c\u00a01590\u00a0cm\u22121 for these four kinds of porous carbons, which belong to the D and G bands. The D band comes from the crystallization defect in the carbon material, which corresponds to the sp3 carbon atoms (Wei et al., 2018; Deng et al., 2015); the G band originates from the in-plane vibrations of the SP2 carbon atoms (Wei et al., 2018; Deng et al., 2015). The intensity ratio of the D band to the G band (ID/IG) was used to compare the degree of defects and graphitization in carbon materials (Wei et al., 2018; Deng et al., 2015). The ID/IG of NPCs, NPCs-PAANa and NPCs-PSS were higher than PCs, suggesting a more defective structure with lower graphitization. N-doped seems to promote the formation of defects in the carbon lattice and reduces the degree of graphitization. The carbon lattice defect sites induced by nitrogen-doped were active due to a reduction in symmetry at or near the crystalline edges (Huang et al., 2021; Zhu et al., 2013).XPS was used to investigate porous carbon's chemical states of C, N and O elements. According to the C 1\u00a0s spectrum of the porous carbons (Fig. S2), three different functional groups such as CC/CC (284.8\u00a0eV), CO (286.3\u00a0eV) and O-CO (288.7\u00a0eV) were observed. A satellite peak corresponding to the \u03c0-\u03c0* transition was detected near 290.9\u00a0eV, indicating the existence of the graphene-like microstructure on the porous carbon surface (Huang et al., 2020; Shi et al., 2017; Nguyen et al., 2020). The O 1\u00a0s spectrum of the porous carbons (Fig. S3) had three different functional groups such as CO (531.7\u00a0eV), OCO (533.1\u00a0eV) and O-CO (534.4\u00a0eV) (Wu et al., 2021; Cai et al., 2020). The N 1\u00a0s spectrum can be divided into four functional groups with different binding energies (Fig. 8\n); the peak near the binding energy of 398.0\u00a0eV, 400.1\u00a0eV, 401.4\u00a0eV and 403.6\u00a0eV are classified into Pyridinic N, Pyrrolic N, Graphitic N and Oxidized N (Arrigo et al., 2016; Wu et al., 2021; Mao et al., 2020; Mao et al., 2019; Jia et al., 2020; Zhang et al., 2020), respectively. According to Table S2, the Graphitic N was the dominant N functional group of NPCs, NPCs-PAANA and NPCS-PSS, and the contents were 40.20\u00a0%, 52.04\u00a0% and 54.43\u00a0%, respectively, which was consistent with the theoretical calculation model.The Fourier transform infrared (FTIR) spectroscopy show that PCs, NPCs, NPCs-PAANa and NPCs-PSS have similar FT-IR spectra (Fig. S13). The broad peaks around 3000\u00a0\u223c\u00a03600\u00a0cm\u22121 could be assigned to the intermolecular hydrogen bonds of carbon and aromatic OH asymmetry stretching vibration. A small peak around 2926\u00a0cm\u22121 could be assigned to the stretching vibrations of \u2013CH3 and \u2013CH2 groups in aliphatic or cycloalkanes (Sheng et al., 2019). A weaker peak at 1387\u00a0cm\u22121 belongs to the stretching vibration of the \u2013CH3 group (Cai et al., 2020; Cuong et al., 2020). The characteristic peaks at 1580\u00a0cm\u22121 and 1178\u00a0cm\u22121 belonged to vibrational stretching of CC and CO in aromatic rings (Chang et al., 2020). No prominent nitrogen-containing groups were found in the FT-IR spectra, which may be related to the low nitrogen content (Gong et al., 2014).As shown in Fig. 6c-d, Pd/C catalysts' pore structure, size distribution, specific surface area, and pore volume are consistent with their carbon supports (Table 2), indicating that the Pd/C catalyst prepared by the absorber-reduction method could inherit the structural properties of the porous carbon support. The specific surface area and pore volume of Pd/PCs, Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS catalysts were 1517 m2g\u22121, 2915 m2g\u22121, 2742 m2g\u22121 and 2839 m2g\u22121, and 0.85 cm3g\u22121, 1.81 cm3g\u22121and 3.06cm3g\u22121, respectively. The higher specific surface area and pore volume of Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS catalysts are beneficial for exposing a high density of active sites and ensuring fast mass transfer.\nFig. 9\n displays the XRD patterns of Pd/PCs, Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS in the range of 10-90\u00b0 (2\u03b8). The Pd/PCs have five different peaks; the peaks at 39.92\u00b0, 46.43\u00b0, 67.76\u00b0, 81.64\u00b0 and 86.12\u00b0 can be indexed to the (111), (200), (220), (311) and (222) planes of face-centered cubic (fcc) Pd (JCPDS# 87\u20130643) (Li et al., 2018), respectively. Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS have only observed the diffraction peak of the Pd (111) plane and are consistent with the theoretical calculation model. Compared with the Pd (111) peak in Pd/PCs, the intense peak around 40\u00b0 in Pd/NPCs downshifts by 0.3\u00b0, and the full-width half-maximum (FWHM) increases by 0.9\u00b0 (Fig. S14). The same situation also exists in Pd/NPCs-PAANa and Pd/NPCs-PSS, indicating Pd/NPCs, Pd/ NPCS-PAANa and Pd/ NPCS-PSS may have slight lattice stretching. The average size of Pd nanoparticles of Pd/C catalysts were calculated by the Scherrer formula. The calculated Pd nanoparticles size of Pd/PCs, Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS were 14.2\u00a0nm, 5.1\u00a0nm, 5.5\u00a0nm and 3.3\u00a0nm, respectively.\nFig. 10\na-d presents TEM images of Pd/PCs, Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS, and it can be observed that Pd nanoparticles are evenly dispersed on the surface of porous carbon. The average particle sizes of Pd nanoparticles on Pd/PCs, Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS were 14.9\u00a0nm, 4.6\u00a0nm, 5.2\u00a0nm and 4.2\u00a0nm, respectively, agree well with the XRD results. It is worth noting that the average particle size of Pd nanoparticles was strongly correlated with the concentration of N functional groups of the porous carbon (Fig. S5). The average particle size of Pd nanoparticles in nitrogen-containing porous carbons (NPCs, NPCs-PAANa and NPCs-PSS) was smaller than N-free porous carbons (PCs) and negatively correlated with the nitrogen concentration, which is consistent with the theoretical calculation results. Theoretical calculation shows that the formation energy of Pd/C heterojunction was significantly reduced with nitrogen doping and further reduced with increased nitrogen content. The lower formation energy is conducive to anchoring Pd nanoparticles on the surface of the porous carbon support, inhibiting the agglomeration of Pd nanoparticles during redox heat treatment, and facilitating the formation of small particle-size Pd nanoparticles. The smaller Pd particle size could increase hydrogen peroxide production (Tian et al., 2017). The lattice spacing of Pd(111) in Pd/NPCs (0.231\u00a0nm), Pd/NPCs-PAANa (0.233\u00a0nm) and Pd/NPCs-PSS (0.231\u00a0nm) were slightly larger than that of Pd/PCs (0.230\u00a0nm) (Fig. 10e-h), agree well with the XRD results. Energy-dispersive X-ray spectroscopy (EDS) elemental mapping Fig. 10i-k showed that the C, N and O atoms were uniformly distributed on the NPCs-PSS support. It is beneficial for the dispersion of the Pd nanoparticles (Fig. 10\nl).The electronic state of Pd and N atoms was further researched by Pd 3d and N 1\u00a0s X-ray photoelectron spectroscopy (XPS). Fig. 11\n shows that the complex Pd 3d was deconvoluted into four different peaks with the binding energies of Pd0 (\u223c341\u00a0eV, \u223c335\u00a0eV) and Pd2+ (\u223c343\u00a0eV, \u223c337\u00a0eV) (Liang et al., 2020; Thuy Vu et al., 607 (2020).). These results indicated that the Pd2+ exited in all catalysts, even though the catalysts were heat treatment in the hydrogen atmosphere. It was worth noting that the binding energies of Pd0 (Table S2) decreased in the order of Pd/PCs (335.83\u00a0eV)\u00a0>\u00a0Pd/NPCs (335.80\u00a0eV)\u00a0=\u00a0Pd/NPCs-PAANa (335.80\u00a0eV)\u00a0>\u00a0Pd/NPCs-PSS(335.72\u00a0eV), suggesting that N doping can facilitate electron flow from the porous carbon support to the supported Pd nanoparticles (Li et al., 2019), theoretical calculations can infer the same result. Benefiting from electron transfer, the Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS have a higher Pd0/Pd2+ ratio than the Pd/PCs (Table S2). The nitrogen species of catalysts were the same as their supports, but the content was reduced (Table S1), possibly because Pd nanoparticles cover part of the N sites.The above evidence confirms that nitrogen doping can effectively inhibit the agglomeration of Pd nanoparticles and promote the Pd nanoparticles to obtain electrons from the N-dope porous carbon support, which is consistent with the theoretical calculation results.\nFig. 12\n shows the change of catalytic performance with the time of different Pd/C catalysts for the direct synthesis of hydrogen peroxide. During the reaction period (3.0\u00a0h), the concentration of hydrogen peroxide in the solution increased (Fig. 12a), indicating that all the catalysts maintained high activity throughout the reaction. Compared with the Pd/PCs, the Pd/NPCs supported by N-doped porous carbon have higher hydrogen peroxide productivity (Fig. 12b) and hydrogen conversion rate (Fig. 12c) but lower hydrogen peroxide selectivity (Fig. 12d); the experimental result agrees well with the conclusion of theoretical calculations. According to Fig. 12, there are two points worth noting: (i) at the beginning of the reaction (0.5\u00a0h), the hydrogen conversion rate was positively correlated with the Pd0/Pd2+ ratio (Table S2), and the Pd/NPCs-PAANa had the highest hydrogen conversion rate. An hour later, the hydrogen conversion rate of Pd/NPCs-PSS exceeded the Pd/NPCs-PAANa (Fig. 12c); (ii) the catalysts (Pd/NPCs, Pd/NPCs-PAANa, and Pd/NPCs-PSS) showed observed differences in H2O2 selectivity. The Pd0/Pd2+ ratio of Pd/NPCs-PSS fall in between Pd/NPCs and Pd/NPCs-PAANa; however, the Pd/NPCs-PSS had the highest hydrogen peroxide selectivity and productivity (Fig. 12d);The size of Pd nanoparticles is one of the factors affecting the direct synthesis of hydrogen peroxide by hydrogen and oxygen. Han's (Tian et al., 2017) research showed that H2O2 selectivity is strongly correlated with particle size when the particle size of Pd nanoparticles is 1.4\u20132.5\u00a0nm, and the effect of particle size on H2O2 selectivity is slight when the particle size is 2.5\u201330\u00a0nm. According to the Fig. 10, the average particle sizes of Pd nanoparticles on Pd/PCs, Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS were 14.9\u00a0nm. 4.6\u00a0nm, 5.2\u00a0nm and 4.2\u00a0nm, respectively, especially for Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS with average particle size between 4.2\u00a0nm and 5.2\u00a0nm. Therefore, the effect of particle size on the catalytic properties of Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS can be ignored in this study. The Pd0/Pd2+ ratio of Pd nanoparticles is another important factor affecting the direct synthesis of hydrogen peroxide (Edwards et al., 2012). Pd0 is the main active site of H2 dissociation, Pd2+ could inhibit the dissociation of H2O2 and improve the H2O2 selectivity. The Pd0/Pd2+ ratios of Pd/PCs, Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS were 1.53, 2.07, 3.33 and 2.35 (Table S2), respectively. A higher Pd0/Pd2+ ratio indicated a higher H2 conversion and a lower H2O2 selectivity. At the initial stage (0.5\u00a0h), Pd/NPCs-PAANa showed the highest H2 conversion, while H2O2 selectively decreased in the order of Pd/PCs\u00a0>\u00a0Pd/NPCs\u00a0>\u00a0Pd/NPCs - PAANa. Although the difference of Pd0/Pd2+ ratio could partly explain point (i) and point (ii), it does not explain the abnormality at point (i) and point (ii). We believe that there are other factors affecting the catalytic performance of hydrogen peroxide. Earlier research suggested that the mesoporous structure of porous carbon could alter the H2O2 selectivity by affecting the mass transfer within the catalyst (Yook et al., 2016; Park et al., 2014; Fellinger et al., 2012). Compared with Pd/PCs, the Pd0 content of Pd/NPCs was increased 7\u00a0%, and the H2 conversion was increased 17\u00a0%. Compared with Pd/NPCs, the Pd0 content of Pd/NPCs-PAANa was increased 9.5\u00a0%, and the H2 conversion only increased 3.1\u00a0%. These results imply that compared with the microporous structure, the mesoporous structure is more conducive to the diffusion of H2 in the catalyst and improved H2 conversion, which is consistent with Choi (Yook et al., 2016). After an hour of reaction, the H2 conversion of Pd/NPCs, Pd/NPCs-PAANa and Pd/NPCs-PSS were 63.0\u00a0%, 66.1\u00a0% and 66.4\u00a0%, respectively, and with a decrease of 7.9\u00a0%, 7.9\u00a0% and 5.8\u00a0%, respectively. The slow decline rate of Pd/NPCs-PSS may be related to the particular pore structure. Compared with Pd/NPCs-PAANa, the larger average pore size and the higher mesoporous ratio of Pd/NPCs-PSS are more conducive to H2 mass transfer in the catalyst and improve H2 conversion.According to the mechanism of direct synthesis of hydrogen peroxide (Han et al., 2021), OO bond dissociation is the key that affects the selectivity of H2O2. A lower Pd0 content is beneficial for improving the selectivity of hydrogen peroxide. At the initial stage (0.5\u00a0h), H2O2 selectivity was negatively correlated with Pd0 content, except Pd/NPCs-PSS. The H2O2 selectivity of Pd/NPCs-PSS was up to 78.8\u00a0%, indicating that other factors may affect the H2O2 selectivity. The solution does not contain hydrogen peroxide at the beginning of the reaction, the dissociation of O2 has an important effect on the hydrogen peroxide selectivity. Xu and co-worker studied (Tian et al., 2020) that the selectivity of hydrogen peroxide is facilitated when O co-adsorbed on the surface of Pd. The Pd/PCs, Pd/NPCs, and Pd/NPCs-PAANa have similar pore structures and average pore diameters, resulting in no significant difference in the diffusion of O2 within these catalysts. The Pd/NPCs-PSS has a special macropore-mesoporous structure and average pore diameters up to 4.32\u00a0nm, which could promote the mass transfer of O2 in the catalyst. The good mass transfer of O2 leads to the faster co-adsorption equilibrium of O on the Pd surface, which improves hydrogen peroxide selectivity. As the reaction progress, the concentration of hydrogen peroxide in the solution gradually increases (Fig. 12a), and the dissociation of hydrogen peroxide becomes an essential factor affecting the catalytic performance. Kakimoto's research (Park et al., 2014) showed that the porous carbon's mesoporous structure could improve the catalyst layer's mass transfer and reduce the contact time of H2O2 inside the catalyst, thereby reducing the decomposition of H2O2. In addition to a large number of 2\u20135\u00a0nm mesopores, the Pd/NPCs-PSS also had a large number of 10\u2013100\u00a0nm mesopores-macropores. The mesopores-macropores structure enhanced the mass transfer of H2O2, reduced the residence time of H2O2 in the catalyst, and improved the selectivity of H2O2.In addition to the above, two other points are worth noting here, (i). the H2O2 selectivity of Pd/NPCs-PAANa exceeded the Pd/NPCs after reaction an hour, and the difference tended to expand (Fig. 12d); (ii) the difference in H2O2 selectivity between Pd/NPCs-PAANa and Pd/NPCs-PSS gradually decreased as the reaction progressed (Fig. 12d). After fully investigating the effect of Pd nanoparticle size, the surface electronic state of Pd and the pore structure of the catalyst support, these two points seem to imply that the dispersion of the catalyst support also has the potential to affect the selectivity of hydrogen peroxide, which worth to research in the future.\nTable S4 compares the catalytic performance of Pd/NPCs-PSS catalysts with other carbon-supported Pd-based catalysts. At ambient pressure, the hydrogen peroxide productivity and selectivity of Pd/NPCs-PSS were up to 328.4 molH2O2 kgcat\n-1h\u22121 and 71.9\u00a0%., respectively. The excellent catalytic performance of Pd/NPCs-PSS indicates that N-doped porous carbon with a macropore-mesoporous-microporous structure is an excellent material for directly synthesizing hydrogen peroxide from hydrogen and oxygen.A series of experiments were performed to study the influence of Pd/C catalysts on the hydrogenation and decomposition of hydrogen peroxide. According to Fig. 13\n, the hydrogenation rates of Pd/C catalysts are higher than the decomposition rates, which indicates that hydrogenation was the primary side reaction (Thuy Vu et al., 607 (2020).; Tian et al., 2020). It is worth noting that Pd/NPCs-PSS had the lowest dissociation and hydrogenation rates than Pd/NPCs and Pd/NPCs-PAANa; although Pd/NPCs-PSS has the smallest Pd particle size (Fig. 10d), the difference in metal particle size distribution too small to fully explain the substantial difference in catalytic results. Therefore, we believe that the pore structure of the catalyst support (Fig. 6 b) also influences hydrogen peroxide's hydrogenation and dissociation performance. The unique macropore-mesoporous-micropore structure of the Pd/NPCs-PSS catalyst reduced hydrogen peroxide's dissociation and hydrogenation rates by increasing the diffusion rates of hydrogen and hydrogen peroxide in the catalyst.Reusability is an important property of the catalyst. In-cycle tests were performed on Pd/NPCs-PSS to examine the stability and reusability of the catalyst. After being reused two times, the H2 conversion, H2O2 selectivity and productivity decreased (Fig. S15). ICP, XRD and XPS were used to analyze the Pd invasion, geometry and electronic morphology of Pd/NPCs-PSS after rescue. The ICP analysis of Pd/NPCs-PSS before and after circulation implied that the leaching of Pd during the catalyst recycling could be negligible (Table S2). the XPS analysis of Pd/NPCs-PSS after circulation implied that the catalyst recycling could reduce the Pd0 content (Fig. S16). The XRD analysis of Pd/NPCs-PSS shows that the catalyst recycling changed Pd's morphology and particle size. The crystal plane, unsuitable for H2O2 production, can be observed in the XRD pattern (Fig. S17), such as the (100) crystal plane. The average particle size of Pd calculated by the Scherrer formula was 16.2\u00a0nm implying that Pd agglomerated in catalyst cycling.In summary, guided by theoretical calculations, we prepared a series of Pd/C catalysts comprising highly dispersed Pd nanoparticles deposited onto hierarchically porous nitrogen-doped carbon material to efficiently synthesize hydrogen peroxide from hydrogen and oxygen. DFT results indicated that the N doping reduces the formation energy of Pd/C heterojunction, promoting the transfer of electrons from carbon support to Pd nanoparticles, forming Pd nanoparticles with small particle size and high Pd0/Pd2+ ratio, and beneficial to improving the hydrogen peroxide productivity. However, N doping shifts the d-band center of Pd toward the Fermi level, and lowers the active dissociation energy barrier of O2, which reduces hydrogen peroxide selectivity. The experimental results showed that adjusting the pore structure of the N-doped porous carbon supports could reduce the negative effect of N doping for H2O2 selectivity. Compared with other catalysts, the special pore structure of Pd/NPCs-PSS catalyst improved the mass transfer rate of H2, O2 and H2O2 in the catalyst, which was the key to inhibiting the negative effects of N doping. At ambient pressure, the hydrogen peroxide productivity and selectivity of Pd/NPCs-PSS were up to 328.4 molH2O2\u00b7kgcat\n-1\u00b7h\u22121 and 71.9\u00a0%, respectively. This study provides a possible solution to design high-performance Pd/C catalysts to directly synthesize hydrogen peroxide from hydrogen and oxygen at atmospheric pressure.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Thanks for the computing support of the State Key Laboratory of Public Big Data, Guizhou University. This work was financially supported by National Natural Science Foundation of China, China (Grant No.22068009), Natural Science Basic Research Program of Guizhou Province, China (Grant No.ZK[2022]088), Cultivation Project of Guizhou University, China (Grant No. [2020]38). the Open Project of Guizhou University Laboratory and Equipment Departments, China (No.SYSKF2022-045) and College Students innovation and entrepreneurship training program of Guizhou Institute of Technology, China (No.S202014440090)Supplementary data to this article can be found online at https://doi.org/10.1016/j.arabjc.2022.104452.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  Nitrogen-doped porous carbon is potential support for directly synthesizing H2O2 from H2 and O2. Here, density functional theory (DFT) was used to study the effect of N-doped porous carbon on H2O2 directly synthesized. The theoretical calculation results showed that N-doped improved H2O2 productivity and H2 conversion by increasing the dispersion of Pd nanoparticles and the Pd0/Pd2+ ratio. However, N-doped decreased H2O2 selectivity by reducing oxygen's dissociation energies. The experimental results showed that adjusting the pore structure of N-doped porous carbon could improve the adverse effects of N-doping for H2O2 selectivity. The H2O2 productivity and selectivity of Pd/C catalyst with a macropore-mesoporous-microporous hierarchical porous structure were up to 328.4 molH2O2\u00b7kgcat\n                     -1\u00b7h\u22121 and 71.9\u00a0%, respectively, at ambient pressure. The macropore structure enhances the transfer and diffusion performance of the catalyst and effectively inhibits the effect of N-doping on OO bond dissociation, which improves H2O2 productivity and selectivity. This research provides a possible solution for designing a high-performance Pd/C catalyst to directly synthesize H2O2 from H2 and O2 at ambient pressure.\n               "}
{"full_text": "A major challenge of our century is the substitution of fossil energy carriers by renewable resources. For chemical industry, the independence from crude oil and coal as carbon source is highly desirable. An alternative, ubiquitous, nontoxic, and sustainable carbon source is CO2, but the stable nature of the molecule makes catalytic activation essential [1]. Processes which enable activation and chemical conversion of CO2 require high energy input, desirably provided by renewable sources (e.g. wind, solar, geothermal, etc.). In conjunction with renewable energy, CO2 has the potential to generate a closed carbon cycle, mitigating CO2 emissions and the related issue of global warming [2]. Large-scale industrial processes with the currently greatest economic potential for CO2 utilization are the production of hydrocarbon fuels, methanol, or \u2013 for fine chemical production \u2013 urea and salicylic acid [3].Reverse water gas shift reaction (rWGS) is among the most promising technologies to convert CO2 into synthetic fuels or CO as their precursor (other examples include direct hydrogenation of CO2 and methane dry reforming \u2013 MDR) [4]. For MDR, the high (400\u202f\u00b0C\u2013600\u202f\u00b0C) operating temperatures and the associated problems of catalyst sintering and coke formation are major drawbacks [5]. Additionally, catalysts are reported to be very sensitive to sulphur impurities, which cause catalyst deactivation [6]. While direct hydrogenation of CO2 is seen as very promising for methanol synthesis on an industrial scale, as it is thermodynamically more favourable than rWGS [7], the indirect route via rWGS and CO is reported to give a 20 % higher methanol yield compared to direct hydrogenation [4]. Furthermore, it has been suggested that rWGS plays a major role in selective methanation of CO2, and it occurs in Fischer-Tropsch reactors operated with high CO2 feeds [8]. In summary, rWGS is a very promising reaction for activation and utilization of CO2, which is the motivation for the present study.The rWGS reaction is an equilibrium-limited reaction (Eq. (1)) and due to its endothermic nature CO formation is favoured at high temperatures. At lower temperatures, the equilibrium favours the water-gas shift reaction (i.e. reverse of Eq. (1)). Moreover, at lower temperatures methanation is a well-known side reaction [4].\n\n(1)\n\n\nH\n2\n\n+\nC\n\nO\n2\n\n\u2009\n\u21cc\n\u2009\nC\nO\n+\n\nH\n2\n\nO\n\u2009\n\n\u0394\nr\n\nH\n\n\u00b0\n298\n\n=\n42.1\n\u2009\nk\nJ\n\u2009\n\n\nm\no\nl\n\n\n-\n1\n\n\n\n\n\nIn recent years, a lot of work was dedicated to the improvement of rWGS catalysts or the design of novel materials. The most widely studied materials are supported catalysts based on copper, platinum, or rhodium [9,10]. Copper has the advantage of lower operating temperatures and suppression of methanation. A comparison of supported platinum catalysts and supported iron and copper reveals improved CO2 conversion rates; however, the selectivity with respect to CO is diminished. [4]. Furthermore, nickel and cobalt based as well as bimetallic catalysts [11,12] were studied. Notably, there is the need for cheap, abundant catalyst materials, as the high cost of noble metals constitutes the major constraint for their large scale application [13]. For instance, Wang et al. showed that the combination of oxygen vacancies and finely dispersed Ni was the reason for the high catalytic activity of their Ni/CeO2 catalyst [11]. Pastor-P\u00e9rez et al. demonstrated high CO2 conversion levels exclusively to CO under various reaction parameters for doped FeCu catalysts [14]. Also, perovskite based rWGS catalysts were tested for their performance as demonstrated by Kim et al. for barium zirconate-based materials [15] or by Daza et al. for cobalt-based perovskites [16]. Additionally, various promotion elements were investigated, e.g. the enhancement of the catalytic activity with potassium, as studied by Chwen et al. [17]Among the wide variety of catalysts reviewed in literature, iron-based catalysts have shown the greatest potential, due to their thermal stability and high oxygen mobility [18], while remaining a feasible option in terms of production costs. Ko et al. performed DFT calculations on CO2 dissociation, finding preferred CO2 to CO dissociation on Fe-containing bimetallic particles [19], thus, highlighting the potential of iron-based catalysts even further. A further advantage of application of reducible oxides in rWGS is that their enhanced oxygen mobility prevents coking [20].Perovskite type oxides with the general formula ABO3 consist of two cations of different sizes (A is larger than B). The wide range of possible structures (many A and B cations can be combined) in conjunction with the option to introduce catalytically active or promoting elements via doping facilitate systematic catalyst development [21]. Moreover, many different types of perovskites have been the subject of extensive research and their properties have been studied thoroughly, particularly because of their wide-spread application in many areas such as solid-state electrochemistry, fuel cell technology, and catalysis [22\u201324]. Among the desired properties of perovskites are their excellent thermal stability (especially important in solid oxide fuel cells due to the high operation temperatures between 600\u202f\u00b0C and 900\u202f\u00b0C) and their resilience towards catalyst poisons at higher temperatures, including the possibility of catalyst regeneration via redox cycling.A further outstanding ability of some perovskites is nanoparticle exsolution [25]. Upon reductive treatment (e.g. in H2) or under reaction conditions (in sufficiently reducing reaction environments) the perovskite is partially reduced. Consequently, reducible lattice cations move to the surface, exsolve there, and form metallic nanoparticles. Catalytically highly active and easily reducible dopants are exsolved preferentially [21]. This process enables in-situ growth of active catalysts [23] and, in comparison to traditional deposition techniques, more finely and more highly dispersed catalyst nanoparticles are formed [25,26]. Moreover, this process is more economical (both with respect to time and cost), as no expensive precursors or complicated \u2018deposition\u2019 procedures are needed [27]. In addition, different studies have found that nanoparticles emerged via exsolution show enhanced sintering stability due to \u201canchoring\u201d, even at high reaction temperatures [25,28,29]. Hence, perovskites have the potential to resolve a major problem of many rWGS catalysts, i.e. rapid deactivation caused by severe aggregation of metal particles at high temperatures [13]. In fact, perosvkite catalysts are promising candidates to reduce similar problems in many CO2 utilization reactions [30,31]. Moreover, Tsounis et al. could show that tailoring perovskites enables selectivity tuning and can also suppress competing side reactions [32].Intensive research on nanoparticle exsolution and its mechanisms has been done already by the solid-state electrochemistry community [25,29,33]. Whether monometallic or bimetallic nanoparticles are formed during dopant and lattice cation reduction mainly depends on two factors: temperature and reductive power of the gas environment [34,35]. These properties open the possibility of catalyst engineering and tailoring materials for respective reactions.The advantages outlined above were the motivation to synthesize not-yet intensively studied perovskite-based rWGS catalyst materials and to study their catalytic performance in the temperature range of 300\u202f\u00b0C\u2013700\u202f\u00b0C and with different doping compositions including in situ XRD. The perovskite catalysts studied were La0.9Ca0.1FeO3-\n\n\u03b4\n, La0.6Ca0.4FeO3-\n\n\u03b4\n, Nd0.9Ca0.1FeO3-\n\n\u03b4\n, Nd0.6Ca0.4FeO3-\n\n\u03b4\n, Nd0.6Ca0.4Fe0.9Ni0.1O3-\n\u03b4\n, and Nd0.6Ca0.4Fe0.9Co0.1O3-\n\n\u03b4\n. The listed materials were selected due to current research on highly active rWGS materials and to demonstrate a possible design approach for finely tuned catalysts.As described in previous work [21], the Pechini synthesis was used to prepare the samples. The following starting materials were mixed in the appropriate stoichiometric ratios: La(CH3COO)3\u00b71.5H2O (99.9 %, Alfa Aesar), Nd2O3 (99.9 %, Strategic Elements), CaCO3 (99.95 %, Sigma-Aldrich), Fe (99.5 %, Sigma-Aldrich), Co(NO3)3\u00b76H2O (99.999 %, Sigma-Aldrich), and Ni(NO3)3\u00b76H2O (98 %, Alfa Aesar). Solutions (either in H2O or in HNO3 (65 %, Merck) \u2013 both doubly distilled) of the needed amounts were produced and mixed. A 20 % excess of citric acid (99.9998 % trace metal pure, Fluka) was added to the resulting mixtures to trigger complex formation. The solvents were evaporated off, self-ignition of the remaining gel was induced by heating, and the formed powders were calcined at 800\u202f\u00b0C for 3\u202fh. The Ni doped perovskite Nd0.6Ca0.4Fe0.9Ni0.1O3-\n\n\u03b4\n was additionally calcined a second time at 1000\u202f\u00b0C, in order to try to achieve phase purity. The calcined products were homogenised via grinding and the powder samples were used for characterization [using Brunauer-Emmet-Teller (BET) analysis, scanning electron microscopy (SEM), and in-situ X-ray diffraction (XRD)] as well as catalytic experiments. Purchased La0.6Sr0.4FeO3-\n\n\u03b4\n (LSF, Sigma Aldrich) was catalytically tested with the same setup as a benchmark material. A Fe2O3/Cr2O3 WGS catalyst (HiFUEL\u2122 W210, ThermoFischer Scientific) served as further reference. This commercially available catalyst can also serve as a tentative indicator of economic profitability: using the above-mentioned synthesis route and starting materials, production of the presented materials would cost about three times as much as the industrial catalyst. However, this is a very crude comparison, as no optimization or customization for industrial applications were done for the presented materials.A PANalytical X'Pert Pro diffractometer in Bragg\u2013Brentano geometry (with separated Cu K\u03b11,2 radiation) and an X\u2019Celerator linear detector was used to perform XRD measurements. The in-situ experiments were carried out in an XRK 900 chamber (Anton Paar), providing a gas flow environment at ambient pressure. After a 30\u202fmin oxidative pre-treatment at 600\u202f\u00b0C, the samples were cooled to room temperature and the reaction atmosphere of H2 and CO2 in a 1:1 ratio, using Ar as balancing gas, was switched on. Flows of 20\u202fmL min\u22121 for each of the reaction gases and 50\u202fmL min\u22121 of Ar were used for the experiments with Nd0.6Ca0.4FeO3-\n\n\u03b4\n, while the respective flows were 7.5\u202fmL min\u22121 and 15\u202fmL min\u22121 for the experiments with Nd0.6Ca0.4Fe0.9Co0.1O3-\n\n\u03b4\n and LSF. The reaction was conducted at various temperatures and at each temperature step an in-situ XRD measurement was taken. To ensure equilibrium, each temperature was held for a period of 10\u202fmin prior to recording of the XRD pattern (\u223c30\u202fmin per measurement), resulting in holding every temperature for about 40\u202fmin. Data analysis and reflex assignment were performed with the HighScore Plus software (PANalytical) and the PDF-4\u202f+\u202f2019 database (ICDD - International Centre for Diffraction Data) [36].SEM images were recorded with secondary electrons on a Quanta 250 FEGSEM (FEI Company) microscope with an Octane Elite X-ray detector (EDAX Inc). An acceleration voltage of 5\u202fkV was used for satisfactory surface-sensitivity.To assess the performance of the investigated samples, trial rWGS reactions were conducted in a tubular flow reactor at ambient pressure (the setup was already described in [37]). Continuous sampling of the gas atmosphere was carried out online (with a measurement every 2\u20133\u202fmin) using a Micro\u2013Gas Chromatograph (Micro-GC, Fusion 3000A, Inficon). A carrier gas flow of 6\u202fmL min\u22121 (Ar) was used, while for both reactive gases (CO2 and H2) a flow of 3\u202fmL min-1 each was set, leading to an overall flow of 12\u202fmL min\u22121 (the gases were purchased from Messer Group GmbH). To assess the effect of the reactor on the catalytic activity, a blank test without catalyst was conducted. The found value \u2013 which was \u223c0.5 Mol% for CO at 600\u202f\u00b0C \u2013 was subtracted from all respective measurements for baseline correction of the measured conversion. The respective amounts of catalyst powder (20\u221275\u202fmg) were chosen such, that the CO generation remained below the thermodynamic limit \u2013 which was tested by repeat measurements with reduced amounts. The chosen flow and catalyst masses resulted in Weight Hourly Space Velocities (WHSV) around 30\u202fL g-1\u202fh-1 (the exact values are given in Table 1\n in Section 3.2). To make sure all test reactions start from the same state (fully oxidized perovskite), the samples were oxidized at 600\u202f\u00b0C for 30\u202fmin in an oxygen atmosphere of 1\u202fbar and a flow of 10\u202fmL min-1 O2.Comparisons of catalytic performance reported in literature tend to be not straightforward: aside from a meaningful indicator regarding the performance, reaction conditions (temperature, pressure, composition of the reaction environment\u2026) need to be given as well. Ref. [13], for instance, offers a nice overview of different catalysts used for rWGS for their given operating conditions. A common measure used by the catalytic community when comparing performances of catalysts are turn over frequencies (TOFs). However, in-depth knowledge about active sites (both nature and number) is necessary to feasibly obtain TOF values. This approach is hindered by the fact that perovskites are highly flexible (i.e. reaction parameter dependent) materials: For instance, the concentration of oxygen vacancies varies strongly depending on temperature. Moreover, the type of surface in contact with the reaction environment might vary (depending on dopants and termination). Also, possible (metal) nanoparticle exsolution influences the number of active sites, and consequently the activity, as well. Therefore, giving a TOF value is not straightforwardly possible. Instead, specific activities (activity per surface area) in mol m\u22122 s-1 were determined.In order to calculate this specific activity, the specific surface areas \n\na\nS\n\n (in m2\u202fg\u22121) of the materials (see Table 1 in Section 3.2.) were measured according to the BET method. Relevant isotherms of the degassed samples (4\u202fh at 300\u202f\u00b0C under vacuum) were obtained at \u2212196\u202f\u00b0C for fitting with a Micrometrics ASAP 2020 system. A specific activity \n\nr\n\nC\nO\n\n\n in mol m-2 s-1 [measuring how many moles of product (CO) were formed per m\u00b2 surface per s] was then derived according to Eq. (2), where the CO formation (mole fraction \n\nx\n\nC\nO\n\n\n in the product stream) and the total gas flow \n\nn\n\u02d9\n\n in mol s-1 were normalized to the catalyst surface area \n\na\n\nC\nA\nT\n\n\n. The total area \n\na\n\nC\nA\nT\n\n\n was obtained from the above-mentioned specific activity \n\na\ns\n\n and the used mass of catalyst \n\nm\n\nC\nA\nT\n\n\n.\n\n(2)\n\n\nr\n\nC\nO\n\n\n=\n\n\n\nn\n\u02d9\n\n\u22c5\n\nx\n\nC\nO\n\n\n\n\n\na\n\nC\nA\nT\n\n\n\n\n=\n\n\n\nn\n\u02d9\n\n\u22c5\n\nx\n\nC\nO\n\n\n\n\n\na\nS\n\n\u22c5\n\nm\n\nC\nA\nT\n\n\n\n\n\n\n\nSix different perovskite powders were manufactured for the current study. The investigated materials were chosen based on the following considerations:\n(i) Ferrite type perovskites have been selected as starting point, since Fe has been proven to be catalytically active in rWGS reactions [18,38], and, therefore, provides an already active host lattice\n(ii) Two of the investigated materials were B-site doped with 10 % Co or Ni, respectively. This means that in addition to an already catalytically active host lattice (see (i) above), further enhancements by reducible and catalytically active B-site dopants can be expected [4,39]. The reason for this expected enhancing effect of doping lies in the capability of the dopants of choice to \u2013 under the proper conditions (reductive atmosphere, high reaction temperatures) \u2013 diffuse to the surface and, by exsolution, form nanoparticles (as could be shown in previous studies [21,40]). Ni and Co are often used in combination with CeO2 and/or Al2O3 as support materials for rWGS catalysts as e.g. shown by work of Wang et al. [39].\n(iii) Nd and La were selected as A-cations, as both (as well as most rare earth elements) reportedly positively impact CO2 utilization reactions in general [41]. Usually, the latter is a very wide-spread A-site element, however, additional materials with Nd instead of La were studied as well. This was done to evade potential problems arising from coinciding peaks of lattice elements and dopants (especially Ni) in future X-ray photoelectron spectroscopy (XPS) investigations.\n(iv) Ca was used as A-site dopant in order to enhance electron and oxide ion conductivities (acceptor doping increases defect concentrations of electron holes and oxygen vacancies). A-site doping allows additional fine-tuning of the catalyst properties: both stability of the crystal as well as exsolution features can be affected [42]. For the doped perovskites here, 10 % and 40 % of Ca doping were chosen. Previous results [43] revealed effects of A-site dopant concentration on the electronic structure (more Ca leads to more partially oxidized Fe4+) and an increase of perovskite stability, thus leading to higher exsolution onset temperatures. For the B-site doped materials enhanced lattice stability was found as well, leading to the assumption that mainly dopants exsolve (given proper reducing conditions), while Fe ions remain in the lattice [21].In short, the perovskite materials La0.9Ca0.1FeO3-\n\n\u03b4\n, La0.6Ca0.4FeO3-\n\n\u03b4\n, Nd0.9Ca0.1FeO3-\n\n\u03b4\n, Nd0.6Ca0.4FeO3-\n\n\u03b4\n, Nd0.6Ca0.4Fe0.9Ni0.1O3-\n\n\u03b4\n, and Nd0.6Ca0.4Fe0.9Co0.1O3-\n\n\u03b4\n, were selected for their content of catalytically active constituents, stability (both thermal and chemical), and exsolution capabilities. For additional comparison, commercial La0.6Sr0.4FeO3-\n\n\u03b4\n (LSF, Sigma Aldrich) was used as reference. Characterization of the produced powders (structure and exsolution capabilities) was performed with XRD and SEM (details have been published in [43]). The XRD measurements revealed that all materials could be prepared successfully and that for all perovskites but the Ni-doped only the perovskite phase was present. They have similar distorted perovskite structures with an orthorhombic lattice, consistent with reports for La(1\u2212x)CaxFeO3 perovskites [44]. In case of Nd0.6Ca0.4Fe0.9Ni0.1O3-\n\n\u03b4\n, minimal amounts of NiO were found (cf. Fig. S2), meaning that Ni was not fully integrated. No additional phases were found for the Co-doped material, indicating complete incorporation. The commercial LSF is, due to the larger Sr cations compared to Ca, differently distorted with a rhombohedral lattice.To obtain the surface area of all synthesized catalyst materials, BET analyses carried out on the freshly prepared powders yielded surface areas ranging from 1.13 m\u00b2 g\u22121 to 5.07 m\u00b2 g\u22121 (see Table 1 in Section 3.2).To ensure the same starting conditions in all experiments and comparability of the results, all samples were pre-treated oxidatively (30\u202fmin in pure O2 at 600\u202f\u00b0C). In a second preparation step, the temperature of samples was reduced to 300\u202f\u00b0C, and only then, the gas atmosphere was changed to the reaction mixture. For all experiments the CO2 and H2 ratio was set to 1:1. Care was taken with respect to the used mass of catalyst, so that the reaction proceeds away from thermodynamic equilibrium (i.e. at 600\u202f\u00b0C and 700\u202f\u00b0C the limit is 40 % and 45 %, respectively). This has to be taken into account, since conversion rates around the equilibrium lead to pronounced back reaction contributions (WGS instead of rWGS) [45]. After switching to the reaction mixture, 60-minutes measurements were conducted between 300\u202f\u00b0C and 700\u202f\u00b0C. The temperature was increased in 100\u202f\u00b0C steps.\nFig. 1\n exemplarily displays the results for rWGS on the B-site undoped perovskite La0.6Ca0.4FeO3-\n\n\u03b4\n. Switching on the reaction environment at low temperatures (300\u202f\u00b0C) did not lead to detectable activity. Only after the reaction temperature was raised to 400\u202f\u00b0C, the onset of CO formation was observed. Similar onset temperatures where noted by Daza et al. on a related perovskite (LaFeO3) with a CO formation onset temperature of 450\u202f\u00b0C and similarly high CO selectivity [16]. When raising the reaction temperature to 500\u202f\u00b0C and 600\u202f\u00b0C, the CO formation increased significantly (\u223c15 % conversion at 600\u202f\u00b0C). It is worth mentioning that parallel consumption of the educts CO2 and H2 is also nicely visible in Fig. 1. This is especially important to note, since the amount of formed water could unfortunately not be quantified with desired accuracy by the used Micro-GC, despite water being visible in the chromatogram. All comparisons presented in this study use the reactivity values found at 600\u202f\u00b0C, since educt conversion at 700\u202f\u00b0C is too close to the thermodynamic equilibrium.The same procedure was used to assess the rWGS performance of all synthesized perovskites as well as of commercial LSF. In Fig. 2\n a comparative summary of all results is shown. Area specific activities (in mol m\u22122 s-1) were used for direct comparisons of the CO formation rate. To get the necessary values for such a comparison, the catalytic activity was related to the active surface areas, see Section 2.3 for details. In Table 1, all average specific activities at 600\u202f\u00b0C are given.The lowest activities were found for the La-based B-site undoped samples (La0.9Ca0.1FeO3-\n\n\u03b4\nand La0.6Ca0.4FeO3-\n\n\u03b4\n, purple curves in Fig. 2), which are comparable to the results of LSF (Fig. 2, black curve). Increasing the A-site dopant concentration had only a minor effect on the activity. The specific activity at 600\u202f\u00b0C was \n\n\n5.7\n\u00d7\n\u2009\n10\n\n\n-\n6\n\n\n\u2009\nm\no\nl\n\u2009\n\nm\n\n-\n2\n\n\n\u2009\n\ns\n\n-\n1\n\n\n and \n5.9\n\u00d7\n\n10\n\n-\n6\n\n\n\u2009\nm\no\nl\n\u2009\n\nm\n\n-\n2\n\n\n\u2009\n\ns\n\n-\n1\n\n\n, respectively (for LSF it was \n4.8\n\u00d7\n\n10\n\n-\n6\n\n\n\u2009\nm\no\nl\n\u2009\n\nm\n\n-\n2\n\n\n\u2009\n\ns\n\n-\n1\n\n\n). A change of A-site dopant (going from La to Nd) increased the activity (Nd0.9Ca0.1FeO3-\n\n\u03b4\n and Nd0.6Ca0.4FeO3-\n\n\u03b4\n, orange curves in Fig. 2). Both materials showed a CO formation onset when raising the temperature to 400\u202f\u00b0C. With every temperature step a further increase of activity was observable. The material with the lower Ca-doping (10 %) exhibited higher CO formation rates than the perovskite with higher Ca content at 600\u202f\u00b0C and 700\u202f\u00b0C. The specific activities at 600\u202f\u00b0C were \n11.3\n\u00d7\n\n10\n\n-\n6\n\n\n\u2009\nm\no\nl\n\u2009\n\nm\n\n-\n2\n\n\n\u2009\n\ns\n\n-\n1\n\n\n and \n6.6\n\u00d7\n\n10\n\n-\n6\n\n\n\u2009\nm\no\nl\n\u2009\n\nm\n\n-\n2\n\n\n\u2009\n\ns\n\n-\n1\n\n\n for Nd0.9Ca0.1FeO3-\n\n\u03b4\n and Nd0.6Ca0.4FeO3-\n\n\u03b4\n\n, respectively.A comparison of activities found for the Ni-doped sample Nd0.6Ca0.4Fe0.9Ni0.1O3-\n\n\u03b4\n in Fig. 2 (blue curve) and the undoped Nd0.6Ca0.4FeO3-\n\n\u03b4\n indicates that doping positively affects CO formation at elevated temperatures. Furthermore, it could be observed that at 500\u202f\u00b0C CO formation increased initially (unlike for the undoped materials, where the activity during each step was either constant or showed an initial drop). At 600\u202f\u00b0C, the specific activity for the Ni doped perovskite was \n18.0\n\u00d7\n\n10\n\n-\n6\n\n\n\u2009\nm\no\nl\n\u2009\n\nm\n\n-\n2\n\n\n\u2009\n\ns\n\n-\n1\n\n\n. The material doped with Co exhibited similar activation phenomena (Fig. 2, green curve). Already at 400\u202f\u00b0C, a slight increase of the CO formation could be observed in the isothermal regime. At 500\u202f\u00b0C, this effect was even stronger. When comparing all activities at all used temperatures, the largest value (\n27.2\n\u00d7\n\n10\n\n-\n6\n\n\n\u2009\nm\no\nl\n\u2009\n\nm\n\n-\n2\n\n\n\u2009\n\ns\n\n-\n1\n\n\n at 600\u202f\u00b0C) was found for the Co-doped perovskite.Besides comparing the novel perovskites to LSF, benchmarks of the catalytic performance against an industrial catalyst were conducted as well. Hence, the rWGS reaction was performed with the same parameters on the commercial HiFuel\u2122 high temperature WGS catalyst. It is composed primary of iron oxide with 7 % chromium oxide to enhance sinter stability, which makes the catalyst ideal for comparison to the iron-based perovskites of this study. At 600\u202f\u00b0C, the obtained specific activity was \n\u223c\n3\n\u00d7\n\n10\n\n-\n7\n\n\n\u2009\nm\no\nl\n\u2009\n\nm\n\n-\n2\n\n\n\u2009\n\ns\n\n-\n1\n\n\n, which is one order of magnitude lower than for the investigated undoped perovskites. But this result has to be evaluated very critically, as very strong sintering and consequently reduction of active surface area of the commercial catalyst was observed at high reaction temperatures. Consequently, the comparability of the specific activity of this reference material is questionable.To summarize, we found that exchanging La with Nd increases the catalytic activity, and doping the perovskite B-site with Ni or Co enhanced it even further. The Co-doped catalyst exhibited the best performance, highlighting that Co-doping is highly beneficial to rWGS activity. The activity enhancing effects of the addition of Co were reported for other materials, e.g. metal-carbides, as well [46].There are two conceivable reasons for the high activities observed in the Ni- and Co-doped perovskites: (i) exsolution of nanoparticles (which is a well-documented phenomenon in perovskites [25]) and (ii) reducibility of the materials.In literature and in previous work done by the authors, it was shown that nanoparticle exsolution can greatly enhance the catalytic activity of perovskites [34,47]. Moreover, the ongoing process of nanoparticle formation could be an explanation for the increasing activity over time at constant temperature observed for the Ni (500\u202f\u00b0C) and Co (400\u202f\u00b0C and 500\u202f\u00b0C) doped catalysts. Exsolution and the observed structural and morphological changes of our catalysts will be treated in depth in Section 3.3.With respect to reducibility, it should be noted that compared to the undoped perovskites, Nd0.6Ca0.4Fe0.9Ni0.1O3-\n\n\u03b4\n and Nd0.6Ca0.4Fe0.9Co0.1O3-\n\n\u03b4\n are more easily reduced in the reaction environment, as was demonstrated in a chemical looping experiment for Co-doping (see supporting info, Fig. S1). Furthermore, it has been shown by various groups that the rWGS reaction of catalysts containing an easily reducible oxide (e.g. as support) follows a surface redox mechanism (see for example the Pt/ceria or iron oxide based systems in Refs. [48,49]).Both aspects will be discussed more thoroughly in Section 3.5., giving a more detailed mechanistic insight.In addition, the occurrence of any side reactions was checked for all tested materials. Methane, for example, is a very well know side product, as the Sabatier reactions are the main side reactions of rWGS [4]. Here, however, no side products could be detected in the gas chromatograms indicating a high CO-selectivity: A key factor for a high selectivity towards CO \u2013 and an important property of perovskites \u2013 is the reducibility of the oxide support material, the oxygen ion mobility, and its capability for vacancy formation. For instance, for rWGS on the perovskites BaZr0.8Y0.16Zn0.04O3 and La0.75Sr0.25FeO3 nearly 100 % CO selectivity have been reported [4,15]. Also, O vacancies have been reported to be crucial for the catalytic activity of the Pd/CeO2/Al2O3 system, as they can be filled with O from CO2 [50]. One possible further reason for the high selectivity of the tested systems could be the size and distribution of the exsolved nanoparticles as discussed below. Similarly, Lu et al. observed that low Ni loadings (< 3 %) with well dispersed nanoparticles are highly beneficial to the selectivity towards CO [51]. They reported 100 % CO selectivity in the temperature range from 400\u202f\u00b0C to 750\u202f\u00b0C.To directly follow the structural changes of the novel perovskites during catalytic reactions and to get insights into the active phase of the different perovskites, in-situ XRD measurements were performed in the reaction environment (i.e. at 1\u202fbar in a flow cell, 1:1 ratio of CO2 and H2) on selected perovskites. Resulting XRD diffractograms for the undoped Nd0.6Ca0.4FeO3-\u03b4 are shown in Fig. 3\n.At low rWGS reaction temperatures (below 400\u202f\u00b0C), only the reflexes corresponding to the perovskite host lattice were visible. Importantly, the perovskite was stable and no decomposition of the material was observed at all reaction temperatures. This is crucial for possible industrial applications, where catalyst regeneration by oxidation/reduction cycles can be realized on stable materials [52]. Although rWGS reaction conditions are reducing, no formation of any metallic Fe-phase could be observed by in-situ XRD over the whole temperature range of the experiment. The chemical potential of the gas phase was not sufficient for the formation of metallic nanoparticles on the surface. However, between 500\u202f\u00b0C and 600\u202f\u00b0C weak signals (2\u03b8 of 30.0\u00b0, 35.3\u00b0 and 62.0\u00b0) were evolving, which could be assigned to the occurrence of Fe3O4. This phase transformed to FeO at around 600\u202f\u00b0C, indicated by the disappearance of the Fe3O4 signals and emerging signals at 2\u03b8 of 36.0\u00b0, 41.8\u00b0, and 60.5\u00b0. This transition agrees with the Fe-O phase diagram [53], which shows a transition around 570\u202f\u00b0C. It was previously reported that iron-oxide is an active phase for catalysing rWGS [4], and in the current experiments this phase is forming under reaction conditions.Additionally, at a temperature of 400\u202f\u00b0C a small signal evolved at 2\u03b8 of 29.4\u00b0, resulting from formation of CaCO3 on the perovskite surface. Similarly, above 500\u202f\u00b0C trace amounts of a graphite phase could be observed. Formation of carbonates under reaction conditions is a well-known phenomenon for A-site doped perovskites, as reported in literature [54]. The amount of CaCO3 increased at higher reaction temperatures, simultaneously with the formation of the iron oxide species. This might be due to the additional driving force of establishing stoichiometric balance in the perovskite structure after the exsolution of B-site cations. Interestingly, both iron-oxide and CaCO3 signals diminish at the highest temperature of 700\u202f\u00b0C.Figs. S3\u2013S6 in the supporting info show diffractograms of all the B-site undoped samples, La0.9Ca0.1FeO3-\u03b4, La0.6Ca0.4FeO3-\u03b4, Nd0.9Ca0.1FeO3-\u03b4, and Nd0.6Ca0.4FeO3-\u03b4, obtained by ex-situ XRD after rWGS reactions (samples from the activity measurements with the last temperature at 700\u202f\u00b0C). In agreement with the in-situ XRD results, for Nd0.6Ca0.4FeO3-\u03b4 both CaCO3 and an iron-oxide phase were found. The latter is Fe3O4, probably because FeO present at the highest temperatures (see in-situ XRD experiment) underwent a phase transition back to Fe3O4 when cooling down.For Nd0.9Ca0.1FeO3-\u03b4, Fe3O4 was also visible, as well as trace amounts of CaCO3. In contrast, for the samples with La no iron-oxide species could be observed. Furthermore, for La0.9Ca0.1FeO3-\u03b4 no new phases were found at all after rWGS, while for La0.6Ca0.4FeO3-\u03b4 formation of CaCO3 was observed. Two trends follow from these observations. Firstly, the exchange of La with Nd enables the formation of the iron oxide phases under rWGS conditions. Secondly, more Ca-doping on the perovskite A-site leads to stronger CaCO3 formation during reaction, as would be expected.Another observation arising when comparing the diffractograms before and after rWGS is the slight shift of all reflex positions towards smaller diffraction angles. This effect is strongest for La0.6Ca0.4FeO3-\u03b4 and is due to expansion of the unit cell in reducing conditions as the result of oxygen vacancies being formed and the partial reduction of the Fe (from Fe4+ to Fe3+ and from Fe3+ to Fe2+) [55].Formation of CaCO3 was also visible in SEM images recorded after rWGS reactions on La0.6Ca0.4FeO3-\u03b4 and Nd0.6Ca0.4FeO3-\u03b4, Figs. S8/S10. After reaction, larger crystals with regular shape (often triangular) and sizes between 200\u202fnm and 400\u202fnm could be observed. These crystals were assigned to CaCO3 (this was supported by EDX measurements performed on the B-site doped materials, see below). In case of lower Ca-doping, La0.9Ca0.1FeO3-\u03b4 and Nd0.9Ca0.1FeO3-\u03b4, no CaCO3 can be seen in the SEM images (Figs. S9/S11). For B-site undoped materials, no formation of nanoparticles was observed on the perovskite surface by SEM, although formation of FeO could be observed for Nd0.6Ca0.4FeO3-\u03b4 by in-situ XRD and Fe3O4 could be found for both perovskites with Nd in the ex-situ XRD patterns. Conceivably, iron oxide occurred either as exsolved nanoparticles that were below the detection limit of SEM, or decomposition occurred without apparent changes in morphology. Generally, the perovskites preserved their overall surface structure (except for CaCO3 formation), highlighting their excellent thermal stability.The absence of metallic surface iron species on the B-site undoped Nd0.6Ca0.4FeO3-\u03b4 could also be confirmed by in situ NAP-XPS data (cf. Fig. S14A).In-situ XRD measurements on the reference material LSF (Fig. S7) showed a very similar behaviour as for Nd0.6Ca0.4FeO3-\u03b4. At low reaction temperatures, only the perovskite phase was observable. At 650\u202f\u00b0C, the formation of FeO could be observed, as indicated by the weak signal at 2\u03b8 of 41.7\u00b0. Moreover, trace amounts of SrCO3 started to appear at 2\u03b8 of 25.0\u00b0, analogous to the formation of CaCO3 in the Ca doped materials. Unlike for the Ca doped samples, SEM could reveal the formation of small nanoparticles (20\u221230\u202fnm) on the perovskite surface for LSF, Fig. S12.To investigate the influence of exsolution on the rWGS activity, the experiment on Nd0.6Ca0.4FeO3-\u03b4 was repeated with an additional pre-treatment step in H2/H2O (32:1) at 700\u202f\u00b0C (60\u202fmin). The bottom XRD pattern in Fig. 4\n indicates successful Fe nanoparticle formation by reductive treatment (2\u03b8 of 44.6\u00b0 and 62.9\u00b0, corresponding to a metallic Fe-phase), which has already been shown in earlier work [21].Upon stepwise increases of the rWGS reaction temperature, the metallic Fe signal started to decrease between 450\u202f\u00b0C and 500\u202f\u00b0C. At the same time, new diffraction lines appeared at 2\u03b8 of 30.0\u00b0, 35.3\u00b0, 56.6\u00b0, and 62.2\u00b0, which could be assigned to Fe3O4. Under rWGS reaction conditions, the initially metallic Fe nanoparticles where oxidized to Fe3O4 in the temperature range of 450\u202f\u00b0C\u2013550\u202f\u00b0C. Between 600\u202f\u00b0C\u2013650\u202f\u00b0C, a further change of the observed phases occurred. The Fe3O4 nanoparticles where reduced to FeO (2\u03b8 of 35.9\u00b0, 41.7\u00b0, and 60.4\u00b0). Additionally, starting from 500\u202f\u00b0C, CaCO3 and graphite were formed. At the highest reaction temperature (700\u202f\u00b0C), the carbonate phase diminished and only FeO and graphite were left on the catalyst surface.In combination with the activity data of the catalytic measurements (cf. Section 3.2), it can be concluded that the formed Fe3O4 and FeO is correlated to the high rWGS activity. While the specific activities of all the B-site undoped materials were similar at 500\u202f\u00b0C, starting from 600\u202f\u00b0C, the Nd-based perovskites exhibited higher catalytic activities compared to the respective perovskites with La. The temperature region between 500\u202f\u00b0C and 600\u202f\u00b0C is exactly where the iron-oxide phases started to evolve in the in-situ XRD experiment with Nd0.6Ca0.4FeO3-\u03b4. Furthermore, an iron-oxide phase only occurred in the ex-situ XRD experiments with the Nd-based perovskites. This phase is formed under reaction conditions, both from the perovskite without reductive pre-treatment and from already exsolved Fe particles. The unwanted formation of CaCO3 observed for the materials with a higher amount of Ca-doping could explain why the activity of Nd0.6Ca0.4FeO3-\u03b4 was lower compared to Nd0.9Ca0.1FeO3-\u03b4. These results demonstrate how rich the surface chemistry of perovskites can be, and that the surface is responding dynamically to changes of the chemical potential of the reaction environment, as well as the strong effect these changes have on the rWGS activity.In-situ XRD results for the Co-doped perovskite Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4 are displayed in Fig. 5\n. Here, the catalyst was exposed to the rWGS reaction environment without any reductive pre-treatment (only initial oxidation was performed for a defined surface state). Above 525\u202f\u00b0C, the formation of a FeO and/or CoO phase could be observed at 2\u03b8 of 36.2\u00b0, 40.0\u00b0, and 60.7\u00b0. Unfortunately, it was not possible to find a clear assignment to either FeO or CoO (or a mixed Fe-Co-oxide \u2013 FexCo1-xO) due to the overlap of the two signals in XRD and the limited resolution of the diffractometer. However, from previous work with EDX mapping, it is known that due to easy reducibility Co is preferentially exsolved at lower temperatures [21]. At 550\u202f\u00b0C, additional formation of a metallic bcc phase could be observed at 2\u03b8 of 44.8\u00b0 and 65.0\u00b0. Again, no clear assignment was possible with XRD due to the signal overlap of Fe and Co, however the fact that no metallic phase occurred in the experiment without Co-doping (cf. Fig. 3), the higher diffraction angle compared to the metallic phase in the experiment with B-site undoped Nd0.6Ca0.4FeO3-\u03b4 and pre-treatment (cf. Fig. 4), and the preferential exsolution of Co suggest a predominance of Co in this phase. Both the B-site metal oxide and metallic phases are more pronounced than in the case of no Co doping. These results show that doping with Co enhances the exsolution process and also enables the reduction of the exsolved elements to a metallic state within the reaction atmosphere, which is preferred for catalysis. These findings where supported by in situ NAP-XPS data as well (cf. Fig. S14B), with the evolvement of metallic Co during rWGS reaction. Already at 500\u202f\u00b0C, a small amount of CaCO3 could be observed at 2\u03b8 of 29.3\u00b0, which got larger simultaneously with the formation of the B-site cation containing phases. Importantly, the perovskite host lattice was stable over the whole temperature range.\nFigs. 6 and 7\n\n summarize the results from SEM and EDX measurements for Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4. After the rWGS reaction, larger, smooth crystals (sizes around 250\u202fnm) on the surface could be observed. They could be identified as formed CaCO3 on the surface, which was supported by the enrichment of Ca and C seen in the EDX spectrum at the position of one of these crystallites (Fig. 7, spectrum C). Furthermore, the formation of finely dispersed nanoparticles with sizes between 20\u202fnm and 50\u202fnm was visible. The EDX spectrum at the position of one particle (Fig. 7, spectrum B) reveals a larger Co L signal (as a shoulder of the Fe L peak) compared to a position in between particles (Fig. 7, spectrum A). This supports the theory that Co has been preferentially exsolved before Fe, and that the nanoparticles can be supposed to be mainly composed of Co. It should be pointed out that the spatial and depth resolution of the EDX analysis is limited, meaning that the obtained spectrum includes signals from the surrounding of the particle. Therefore, an unequivocal determination of the particle composition would require additional TEM studies.The in-situ XRD results suggest that during reaction the nanoparticles were either primarily metallic or already oxidized. Due to the exposure to air (and hence oxidation), when transferring the samples to the SEM, it is no longer possible to distinguish between the two cases. Therefore, it is not clear which of the B-site metal oxide or metallic phases observed with XRD, or even both, correspond to the nanoparticles seen in the SEM images. Astonishingly, these nanoparticles were stable without sintering effects even at high reaction temperatures up to 700\u202f\u00b0C. The reason is that nanoparticles formed by exsolution are anchored to the perovskite surface, as has been shown by Neagu et al. [33] in an in-situ TEM study. The stable nature and the prevention of sintering of these formed nanoparticles make perovskite catalysts extremely valuable for industrial applications.As shown in the catalytic data (cf. Section 3.2), the Co-doped perovskite exhibited the highest rWGS activity. A possible reason could be the co-existence of nanoparticles containing metallic Co and the FeO/CoO oxide phase on the perovskite surface (in vicinity to oxygen vacancies of the host lattice), both enhancing hydrogen dissociation and redox activity (further details see Sec. 3.5 below). This is supported by the fact that the activity measurements of the Co-doped catalyst showed the largest increase between 500\u202f\u00b0C and 600\u202f\u00b0C, the temperature region where these phases started to appear in the in-situ XRD experiment. For Sr-doped lanthanum cobaltite perovskites (La1-xSrxCoO3-\u03b4), Daza et al. reported the importance of metallic cobalt for the conversion of CO2 to CO as well [56].For the Ni doped catalyst Nd0.6Ca0.4Fe0.9Ni0.1O3-\u03b4, a diffractogram was obtained ex-situ after rWGS reaction (Fig. S2 in the supporting information) using the sample from the activity measurement (last temperature 700\u202f\u00b0C). Here, the NiO impurity observed in the pristine sample could not be found anymore. Instead, a metallic Ni-phase (fcc) was present, indicated by the reflexes at 2\u03b8 of 43.9\u00b0 and 51.2\u00b0. This suggests a reduction of the NiO phase to metallic Ni during rWGS, probably around 500\u202f\u00b0C, which is the temperature where increasing activity at constant parameters was observed in the catalytic experiments. The formation of additional metallic Ni (by exsolution) cannot be confirmed conclusively. Also, alloying of Ni with Fe within the fcc phase might be possible by exsolution. Further experiments would be necessary to determine the exact behaviour. Besides the Ni containing phases, formation of CaCO3, as well as trace amounts of Fe3O4 could be observed. This agrees with the results observed for the B-site undoped Nd0.6Ca0.4FeO3-\u03b4. Compared to the latter, there was more CaCO3 and less Fe3O4. This can be explained by the changed driving forces for segregation, evoked by the formation of the Ni phase. As this leads to a B-site sub-stoichiometry in the remaining perovskite, the A-site Ca segregation is enhanced, while B-site Fe segregation is reduced.The larger CaCO3 crystals found on the surface of La0.6Ca0.4FeO3-\u03b4, Nd0.6Ca0.4FeO3-\u03b4, and Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4 were also observed in the SEM images of Nd0.6Ca0.4Fe0.9Ni0.1O3-\u03b4 (Fig. S13). Here, they were even bigger with sizes between 400\u202fnm and 700\u202fnm. This larger size matches the result of an enhanced Ca segregation obtained from the XRD measurements. Furthermore, the larger size of the crystals allowed for an EDX mapping to confirm their chemical nature as CaCO3. At high magnification, also very small nanoparticles (< 15\u202fnm) were visible on the surface. These probably consist of Ni, in agreement with the metallic Ni-phase observed with XRD. The formation of Ni nanoparticles can explain the higher catalytic activity of the Ni-doped catalyst compared to the undoped ones, which was observed in the catalytic measurements (cf. Section 3.2).To further investigate the promoting effect of the formed nanoparticles, and to answer if the exsolved nanoparticles are really enhancing the catalytic activity, additional experiments were conducted. These additional activity tests for Nd0.9Ca0.1FeO3-\u03b4 and Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4 were performed with linear temperature ramps.For the undoped Nd0.9Ca0.1FeO3-\u03b4, it was found that exsolution of metallic Fe particles does not enhance the catalytic activity (for more details see the supporting info Section \u201cExsolution Enhanced Catalytic Reaction\u201d and Fig. S15).For the B-doped Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4, the reaction temperature was increased from 300\u202f\u00b0C to 570\u202f\u00b0C (heating rate =1\u202f\u00b0C min\u22121). After reaching 570\u202f\u00b0C, the reaction temperature was reduced with the same rate (to 300\u202f\u00b0C) (see also Fig. S16). For this special experiment 570\u202f\u00b0C was chosen as highest temperature, because at this temperature exsolution is already possible, but CaCO3 formation is still minor (see in-situ XRD results, Fig. 5). From the in-situ XRD results, in this temperature range the exsolved particles are expected to be mainly cobalt-oxide.Since the cobalt-oxide-nanoparticles only form upon the first heating in reaction atmosphere (they are formed in-situ at high reaction temperatures where cation mobility is sufficient), heat-up and cool-down behaviour are expected to be different in case the particles affect the activity of the catalyst. If the formed nanoparticles would have no influence on catalytic reactivity, the CO formation rate should be equal for both up and down ramping. Indeed, a hysteresis-like behaviour was found as can be seen in Fig. 8\n, with a maximum increase in formed CO by 0.9\u202fmol% at 490\u202f\u00b0C. This is a clear indication that for the cooling ramp a higher catalytic reactivity was observed than for the heating ramp, which can be interpreted as an evidence for a catalytic effect of the exsolved nanoparticles.To show that this method is genuinely suitable of determining differences in catalytic activity, a second identical experiment was conducted directly after the first heat-up/cool-down cycle (i.e. without changing the catalyst, and without any intermediate treatment). The catalytic activity during both the second heating and cooling phases followed the cooling ramp of the first cycle exactly (see Fig. 8, dashed curves). This confirms that nanoparticles, evolving upon the very first heat-up, were improving the rWGS reactivity over the whole temperature range of all following heat-up or cool-down ramps, indicating reversible and stable catalytic behaviour of the nanoparticle decorated perovskite catalyst. Consequently, with this experiment strong structural deactivation phenomena in the investigated temperature range (up to 570\u202f\u00b0C, prior to the increased formation of CaCO3) could be ruled out as well.In heterogeneous catalysis on reducible oxides, commonly lattice oxygen is found in the reaction product. This effect was explained by P. Mars and D. W. van Krevelen by suggesting a regenerative redox mechanism that consists of two steps [57]: In the first step, the catalyst is reduced by one of the educts, which is oxidized by taking up an oxygen atom from the catalyst and thus creates a vacancy in the catalyst\u2019s oxygen sub-lattice. In the following second step, the catalyst is regenerated by reaction with the second educt, which donates an oxygen atom to the catalyst and hence becomes reduced.The rWGS reaction (Eq. (1)) on oxide catalysts usually also proceeds via a Mars-van Krevelen (MvK)-type redox mechanism. There, in the first step, hydrogen reacts with lattice oxygen of the catalyst forming the first product water, which desorbs leaving an oxygen vacancy behind. The two electrons formed in this reaction step are also transferred to the catalyst oxide ensuring charge neutrality. In Eq. (3) the half reaction of H2 oxidation proceeding on a reducible oxide is written in Kr\u00f6ger-Vink notation with \n\nO\nO\n\u00d7\n\n, \n\nv\nO\n\u2219\u2219\n\n, and \ne\n'\n denoting regular lattice oxygen (relatively neutral), oxygen vacancy (relatively two-fold positive), and electron (relatively negative), respectively.\n\n(3)\n\n\nH\n2\n\n+\n\nO\nO\n\u00d7\n\n\u2192\n\nH\n2\n\nO\n+\n\nv\nO\n\u2219\u2219\n\n+\n2\ne\n'\n\n\n\nIn the second step, these electrons are consumed by reduction of carbon dioxide, which annihilates an oxygen vacancy by donating an oxygen atom to the catalyst, thus forming the second product carbon monoxide.\n\n(4)\n\n\n\nC\nO\n\n2\n\n+\n\nv\nO\n\u2219\u2219\n\n+\n2\n\ne\n'\n\n\u2192\n\nO\nO\n\u00d7\n\n+\nC\nO\n\n\n\nBoth half reactions are coupled by the electron transfer via the solid catalyst substrate and thus for the case of steady state conditions, the following relationship for the reaction rate densities \nr\n holds:\n\n(5)\n\n\nr\n\n\nH\n2\n\n-\no\nx\n\n\n=\n\nr\n\n\n\nC\nO\n\n2\n\n-\nr\ne\nd\n\n\n=\n\nr\n\nr\nW\nG\nS\n\n\n\n\n\nTherein, the subscripts \u201cH2-ox\u201d and \u201cCO2-red\u201d refer to Eqs. (3) and (4), respectively. Both are equal to the net rate of the rWGS reaction rrWGS, which is compared in a surface area-normalised form (thus called specific activity) in Fig. 2.As one can see in this figure, upon changing the composition of the investigated perovskite-type catalysts, the reaction rate is improved. In the following, we thus propose a relation between the observed catalyst activities in Fig. 2 and the corresponding catalyst composition based on the characteristics of an assumed MvK-type redox mechanism on mixed conducting perovskites. To do so, it is helpful to first summarize previous results on similar materials:(i) CO2 reduction proceeds on the perovskite surface and is only little affected by exsolution of metallic particles [40]. The availability of a sufficiently high concentration of electrons in the electro-catalyst increases efficiency drastically. Thus, the CO2 reduction rate is higher on more easily reducible oxides [40].(ii) Exsolved metallic particles can enhance the H2 oxidation rate significantly [34,47]. The improvement of H2 oxidation rate occurs by spillover of adsorbed hydrogen species from the metal to the oxide \u2013 i.e. the bare oxide surface suffers from a depletion of reactive species, hence here H2 activation is limiting the net reaction rate, while exsolved metallic particles help sustaining a significant coverage with an active hydrogen species [58].(iii) Surface enrichment of Sr causes performance degradation of perovskite-type electro-catalysts, but enrichment of La shows only minor effects [59,60].With this in mind, let us now look at the catalyst performance data in Fig. 2. Doping LaFeO3 with calcium instead of strontium does not visibly change the catalytic activity of the investigated perovskite. This is in agreement with previous results, since both Ca and Sr do not show any redox activity and their effect on defect chemistry (e.g. electron concentration) is only due to their charge.Changing La against Nd caused an improvement in the specific activity by up to a factor of two (compare the purple and orange curves in Fig. 3). This result is difficult to be unambiguously explained in quantitative terms using the data available so far. However, two potential qualitative explanations shall be briefly discussed here. First, the different redox activity of La and Nd may be the reason for the observed behaviour. While La in the perovskite bulk has only a minor (if any) redox activity, there are indications that surface La actively participates in redox reactions [61]. Assuming a slightly easier reducibility of surface Nd compared to surface La may explain the increased rWGS activity of the Nd-based catalysts via improvement of the CO2 reduction reaction (Eq. (4)). Second, the iron oxide phases observed on Nd0.6Ca0.4FeO3-\u03b4 and Nd0.9Ca0.1FeO3-\u03b4 may contribute to an increased CO2 reduction rate. Formation of these phases might as well be simply a consequence of the high redox activity of the Nd-containing surface, without a significant enhancing effect of the iron oxide phases on the CO2 reduction. The exact relations are still to be investigated further. However, any role of FeO and Fe3O4 for an improved H2 oxidation activity can be definitely ruled out as demonstrated in Ref. [58].A further noteworthy increase of the catalyst performance was achieved by introduction of the reducible elements Co and Ni, which under reaction conditions both caused decoration of the oxide catalysts by metallic precipitates (see Figs. 2 and 6). Hence \u2013 by considering the above-mentioned previous results of H2 oxidation on exsolution-decorated electro-catalysts \u2013 we suggest the associated specific activity improvement to be mainly caused by an enhancement of the H2 oxidation rate (Eq. (3)). Some additional activity enhancement may originate from the easier reducibility and the corresponding higher oxygen vacancy concentration especially of the cobalt doped material.A summary of both interpretations is sketched in Fig. 9\n, which shows the job sharing of oxide surface and metal exsolution, catalysing CO2 reduction and H2 oxidation, respectively. The connection of both half-reactions is achieved by electrons and oxygen vacancies flowing via the mixed conducting perovskite-type catalyst.To summarize, the high reactivity of Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4 (and to some extent Nd0.6Ca0.4Fe0.9Ni0.1O3-\u03b4) can be explained by the synergistic effect of the perovskite host lattice with its capability for high oxygen vacancy concentration, responsible for effective CO2 activation, the formed CoO (and/or FeO) on the surface, which is contributing to the oxygen chemistry, and the exsolved metal nanoparticles, which are enhancing the H2 adsorption and dissociation ability of the catalyst surface.To assess the deactivation behaviour of the investigated perovskites, rWGS reactions at constant high temperatures (600\u202f\u00b0C) were conducted. Here, results for Nd0.6Ca0.4FeO3-\u03b4 and Co-doped Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4 are highlighted in Fig. 10\n. Results for the other materials are displayed in the supporting info (Figs. S18\u2013S20). To ensure the same well-defined starting point for all experiments, all samples were subjected to an oxidizing pre-treatment at 600\u202f\u00b0C for 30\u202fmin. Following a cooling down period (to 300\u202f\u00b0C in O2), the reaction gas atmosphere was switched on. A sharp increase of reaction temperature to 600\u202f\u00b0C in order to find possible pronounced activation or deactivation effects during reaction onset did not reveal any such effects.For Nd0.6Ca0.4FeO3-\u03b4 (Fig. 10), a slow deactivation over time was observed which could be seen in the slightly decreasing CO signal. This slow deactivation is very likely attributed to formation of CaCO3 crystallites on the surface, as observed in SEM images and XRD after reaction (Figs. S10 and S4). The CaCO3 is formed in two stages: Ca-segregation to the surface and its reaction with CO2 from the reaction atmosphere. This phenomenon of carbonate formation is a well-known issue for perovskites with alkaline earth metals as A-site cations, as shown in literature e.g. for BaCo0.4Fe0.4Zr0.2O3-\n\n\u03b4\n [54]. The formed CaCO3 crystallites are covering part of the perovskite surface, blocking active sites. The slow deactivation is a result of the proceeding growth of CaCO3. Possible strategies for reducing this deactivation process are either a reduction of the Ca-content, the use of an A-site sub-stoichiometric material or a change of the A-site composition to elements that are less prone to segregation or carbonate formation [62]. For the tested perovskites with lower A-site Ca doping, an already reduced tendency for segregation and CaCO3 formation was observed, as shown by XRD and SEM data (Figs. S3/S9 and S5/S11).Nd0.9Ca0.1FeO3-\u03b4, La0.9Ca0.1FeO3-\u03b4, and La0.6Ca0.4FeO3-\u03b4 showed a similar behaviour with respect to catalytic activity with slow deactivation over time (Figs. S18\u2013S20). For the Co-doped Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4, the same behaviour was observed as well, see Fig. 10 (inset). In in-situ XRD data (Fig. 5), the formation of CaCO3 was observed above a reaction temperature of 500\u202f\u00b0C. With increasing reaction temperature, the corresponding reflex grew more intense, indicating the growth of the CaCO3 crystallites.Six rare earth based perovskite-type oxides were investigated with respect to their rWGS performance, as well as their corresponding surface structure and morphology (using XRD and SEM). The highest catalytic activity was achieved when formation of metallic nanoparticles by exsolution occurred during reaction, as could be shown for the Ni- and Co-doped materials Nd0.6Ca0.4Fe0.9Ni0.1O3-\u03b4 and Nd0.6Ca0.4Fe0.9Co0.1O3-\u03b4. We propose that this is due to a Mars-van-Krevelen-type mechanism of the rWGS reaction on these materials with a highly beneficial job-sharing ability of the metal-particle-decorated catalyst surfaces. On the one hand, dissociative H2 adsorption takes place on the metal particles. By spillover, the active hydrogen species can subsequently increase the reduction rate (and thus the oxygen vacancy generation rate) in the perovskite backbone compared to the undecorated perovskite. This backbone, on the other hand, is responsible for CO2 activation. It shows a pronounced defect chemistry, including oxygen vacancies, and can reduce CO2, accepting one of its O atoms to re-fill a vacancy.The choice of both A-site and B-site compositions of the perovskite material play an important role for the resulting behaviour and performance. For the design of an optimal catalyst material, several aspects must be considered:\n\n(i)\nThe used rare earth metal on the A-site can \u2013 via its aptitude to redox behaviour \u2013 influence the reducibility of the perovskite backbone and its ability for oxygen vacancy formation. This, in turn, has an effect on the CO2 reduction rate, but also on the exsolution process (this was also seen in previous work regarding exsolution [43]). In this work, we have shown that exchanging La with the presumably more redox-active Nd increased the catalytic activity for rWGS.\n\n\n(ii)\nDoping of the A-site with an alkaline earth metal is important for the perovskite defect chemistry, affecting electronic and ionic conductivity, oxygen vacancy concentration, and thus CO2 reduction rate and exsolution behaviour. However, alkaline earth metal doping comes with the disadvantage of segregation effects and carbonate formation during rWGS. For our Ca-doped materials, we observed CaCO3 covering the surface, thus blocking active sites and resulting in catalyst deactivation over time. The carbonate formation increased with a higher Ca content and more pronounced B-site metal exsolution.\n\n\n(iii)\nUsing an easily reducible element (such as Co or Ni) as B-site dopant in combination with a less reducible main component (in our case Fe) not only facilitates formation of metal particles on the surface, but also enables a preferential exsolution of the dopant element (found in this study both for Ni- and Co-doping). At the same time, the less reducible main component ensures that a stable perovskite backbone is retained, because it is not completely reduced under the same conditions. The metal particles strongly increase the catalytic activity for rWGS, as was observed for both the Ni- and the Co-doped material. This stable backbone can provide a good anchoring of exsolved nanoparticles, thus preventing sintering and ensuring a high amount of gas/metal/oxide three-phase boundaries. For our Co-doped material, no sintering of the metal nanoparticles could be observed, even up to 700\u202f\u00b0C.\n\n\n(iv)\nThe choice of the B-site dopant is crucial. Cobalt is known to be highly active for rWGS, and consequently the Co-doped perovskite showed the best rWGS performance. In contrast, Fe particles exsolved by reductive pre-treatment, did not enhance the catalytic activity.\n\n\nThe used rare earth metal on the A-site can \u2013 via its aptitude to redox behaviour \u2013 influence the reducibility of the perovskite backbone and its ability for oxygen vacancy formation. This, in turn, has an effect on the CO2 reduction rate, but also on the exsolution process (this was also seen in previous work regarding exsolution [43]). In this work, we have shown that exchanging La with the presumably more redox-active Nd increased the catalytic activity for rWGS.Doping of the A-site with an alkaline earth metal is important for the perovskite defect chemistry, affecting electronic and ionic conductivity, oxygen vacancy concentration, and thus CO2 reduction rate and exsolution behaviour. However, alkaline earth metal doping comes with the disadvantage of segregation effects and carbonate formation during rWGS. For our Ca-doped materials, we observed CaCO3 covering the surface, thus blocking active sites and resulting in catalyst deactivation over time. The carbonate formation increased with a higher Ca content and more pronounced B-site metal exsolution.Using an easily reducible element (such as Co or Ni) as B-site dopant in combination with a less reducible main component (in our case Fe) not only facilitates formation of metal particles on the surface, but also enables a preferential exsolution of the dopant element (found in this study both for Ni- and Co-doping). At the same time, the less reducible main component ensures that a stable perovskite backbone is retained, because it is not completely reduced under the same conditions. The metal particles strongly increase the catalytic activity for rWGS, as was observed for both the Ni- and the Co-doped material. This stable backbone can provide a good anchoring of exsolved nanoparticles, thus preventing sintering and ensuring a high amount of gas/metal/oxide three-phase boundaries. For our Co-doped material, no sintering of the metal nanoparticles could be observed, even up to 700\u202f\u00b0C.The choice of the B-site dopant is crucial. Cobalt is known to be highly active for rWGS, and consequently the Co-doped perovskite showed the best rWGS performance. In contrast, Fe particles exsolved by reductive pre-treatment, did not enhance the catalytic activity.These considerations can be used to tune the perovskite catalyst composition, achieving a material exhibiting excellent catalyst properties \u2013 exsolution of catalytically highly active metal nanoparticles with good sintering resistance and optimal size, distribution, and composition from a stable and reducible perovskite backbone, while showing only minimal carbonate formation. Not only changing the used elements and their ratio, but also introducing a sub-stoichiometry to either of the perovskite sites is conceivable, giving an even wider range of possibilities for optimization. Thus, an optimal rWGS performance tuned to the actual process parameters \u2013 with a high catalytic activity and stable operation \u2013 can be realized.Our results show that this material class is ideal to meet current challenges for industrial scale rWGS. Moreover, rWGS belongs to the most promising reactions for future CO2 conversion and utilization systems and belongs to the closest to implementation.This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement n\u00b0 755744 / ERC - Starting Grant TUCAS).\nL. Lindenthal: Conceptualization, Investigation, Formal analysis, Validation, Writing - original draft, Writing - review & editing. J. Popovic: Investigation, Formal analysis. R. Rameshan: Data curation, Investigation, Formal analysis. J. Huber: Investigation, Formal analysis. F. Schrenk: Investigation, Formal analysis. T. Ruh: Data curation, Validation, Writing - review & editing. A. Nenning: Data curation, Validation. S. L\u00f6ffler: Investigation, Formal analysis. A.K. Opitz: Supervision, Validation, Writing - review & editing. C. Rameshan: Conceptualization, Funding acquisition, Project administration, Supervision, Writing - original draft, Writing - review & editing.The authors report no declarations of interest.The X-ray measurements were carried out within the X-Ray Center of TU Wien. The authors acknowledge TU Wien Bibliothek for financial support through its Open Access Funding Programme.Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.apcatb.2021.120183.The following is Supplementary data to this article:\n\n\n\n\n", "descript": "\n                  Reverse Water-Gas Shift (rWGS) is among the reactions with the highest readiness level for technological implementation of CO2 utilization as an abundant and renewable carbon source, and its transformation for instance into synthetic fuels. Hence, great efforts are made in terms of further development and comprehension of novel catalyst materials. To achieve excellent catalytic performance, catalytically active (nano)particles that are evenly distributed on (and ideally embedded in) an active support are crucial.\n                  An extremely versatile material class that exhibits the desired properties are perovskite-type oxides due to the fact that they can easily be doped with highly active elements. Upon controlled reduction or during reaction, these dopants leave the perovskite lattice and diffuse through the material to form nanoparticles at the surface (by exsolution) where they can greatly enhance the activity.\n                  Here, six perovskites were studied and their exsolution capabilities as well as rWGS performance were explored. Nanoparticle exsolution significantly enhanced the rWGS activity, with the catalytic activity being in the order Nd0.6Ca0.4Fe0.9Co0.1O3-\n                     \n                        \u03b4\n                      > Nd0.6Ca0.4Fe0.9Ni0.1O3-\n                     \n                        \u03b4\n                      > Nd0.9Ca0.1FeO3-\n                     \n                        \u03b4\n                      > Nd0.6Ca0.4FeO3-\n                     \n                        \u03b4\n                      > La0.6Ca0.4FeO3-\n                     \n                        \u03b4\n                      > La0.9Ca0.1FeO3-\n                     \n                        \u03b4\n                      > La0.6Sr0.4FeO3-\n                     \n                        \u03b4\n                     (benchmark). Moreover, it could be shown that nanoparticles formed due to exsolution are stable at high reaction temperatures. In this paper, the flexibility of the investigated perovskite materials is demonstrated, on the one hand facilitating a material design approach enabling control over size and composition of exsolved nanoparticles. On the other hand, the studied perovskites offer a tuneable host lattice providing oxygen vacancies for efficient CO2 adsorption, activation, and resulting interface boundaries with the ability to enhance the catalytic activity.\n               "}
{"full_text": "Influenced by the continuously growing ecological effects stemming from anthropogenic climate change, carbon dioxide (CO2) conversion and utilization have been receiving increasing interest from the scientific community.\n1\n\n,\n\n2\n Among all the different CO2 utilization processes, electrocatalytic CO2 reduction reaction (CO2RR) is regarded as an attractive strategy that can not only help to reduce the atmospheric CO2 level and subsequent global warming effect but also relieve mankind\u2019s dependence on fossil fuels for sustainability.\n3\n\n,\n\n4\n In the past decade, substantial efforts have been dedicated to increase the multi-carbon (C2+) selectivity for CO2RR, in particular, using high-alkalinity electrolytes (e.g., 0.1\u201310\u00a0M KOH) to lower overpotentials and promote the yields of target products.\n5\n However, the direct synthesis of C2+ products from CO2RR in alkaline electrolyte suffers from the inevitable side reaction of carbonate formation at the catalytic interface, which results not only in the depletion of both CO2 and OH\u2212 but also carbonate precipitation on the electrode, which degrades the catalytic performance.\n6\n\n,\n\n7\n\nThe production of hydrocarbons and oxygenates in CO2RR has been generally proposed as the initial conversion of the CO2 reactant into \u2217CO intermediates, followed by further reduction of those adsorbed \u2217CO intermediates.\n8\n Thus, a two-step tandem approach in which CO2 is initially reduced to carbon monoxide (CO) and subsequently into C2+ products has been emerging as an alternative strategy.\n6\n\n,\n\n9\n Owing to the advance of solid-oxide electrolysis cell (SOEC) technology,\n6\n high-rate CO2RR to CO has been developed for commercial applications, making this two-step cascade design feasible. On one hand, CO electroreduction reaction (CORR) is a carbonate-formation-free platform that can be operated in high-alkalinity conditions and waive the problem of carbonate formation.\n9\n On the other hand, as CO is widely accepted as the key intermediate for C\u2212C coupling,\n10\n the direct use of CO feed for electrolysis can significantly increase the local CO concentration at the catalyst-electrolyte interface, which inherently favors the generation of C2+ products.\n7\n\n,\n\n11\n Furthermore, as CO\u2217 is widely accepted as a key intermediate in CO2RR, CO2RR is believed to share common reaction pathways with CORR. Hence, understanding the CORR mechanism can also benefit the mechanistic exploration of CO2RR. The catalyst design strategies of CORR may also be applicable for CO2RR.To date, Cu is generally known as the only monometallic element that can convert CO2/CO into C2+ hydrocarbons and oxygenates with appreciable activities and selectivities.\n10\n\n,\n\n12\n\n,\n\n13\n Substantial research efforts have been devoted on Cu-containing catalyst surfaces to gain fundamental understandings of CORR. A variety of strategies, such as alloying,\n14\n facet tailoring,\n15\n morphology control,\n16\n\n,\n\n17\n and oxygen engineering,\n18\n have been pursued on Cu-based catalysts for CORR, targeting steering product selectivity and boosting catalytic activity. Some specific features in CORR, especially compared with its kindred CO2RR, have been discovered. In this review, we will provide an overview of current CORR advances in the mechanistic understanding, representative design strategies for the Cu-based catalysts (Table\u00a01\n), and potential influential factors such as electrolyte and CO coverage. Furthermore, the rational design of CORR electrolysis reactors is also critical for the promotion of CORR performances. Finally, we will summarize the existing challenges that should be overcome to push this technology to the next level, and propose several perspectives for future opportunities.Although Cu shows appreciable selectivity and activity for catalyzing CORR toward C2+ products, the possible products have been reported with a wide distribution. Typically, C2+ products, including ethylene (C2H4), ethanol (EtOH), acetate, and n-propanol, can be obtained simultaneously,\n28\n\n,\n\n29\n but none of these products can overwhelmingly dominate the product spectrum. From a mechanistic perspective, the broad C2+ product distribution is likely to be caused by the existence of multiple bifurcations in the reaction pathways.\n30\n\n,\n\n31\n Thus, understanding the mechanism of CORR is of great significance, which can allow the provision of a roadmap for the catalyst design with optimal reaction activities and selectivities. As CO electroreduction is a highly complicated process composed of diversified reaction pathways along with multiple electron-and-proton transfer steps, computational investigations have been employed as major toolkits in providing theoretical insights for the formation of C2+ products.\n30\n\n,\n\n31\n In addition, spectroscopic studies to identify reaction intermediates have also been widely used to aid the mechanism exploration.In 2013, Koper and co-workers proposed a C\u2212C coupling mechanism through proton and electron transfer as follows: (1) CO adsorption; (2) coupling of \u2217CO with CO(g) via electron transfer to form \u2217C2O2\n\u2212; (3) protonation of \u2217C2O2\n\u2212 to form \u2217OCCOH (Figure\u00a01A).\n31\n The characteristic of this mechanism shows that it can be performed at a relatively low overpotential, and its initial potential was calculated to be \u22120.4\u00a0V (versus reversible hydrogen electrode [RHE]). Using in situ Fourier transform infrared (FT-IR) spectroscopy, the authors further provided experimental evidence for the formation of a hydrogenated CO dimer (\u2217OCCOH) at low overpotentials during CO reduction on Cu(100) electrodes in LiOH solution (Figure\u00a01B).\n32\n However, the \u2217OCCOH intermediate could not be observed on Cu(111) under identical conditions, suggesting that CO dimerization is a structure-sensitive process, in agreement with previous experimental and computational observations.\n35\n\n,\n\n36\n Similarly, N\u00f8rskov and co-workers predicted that OCCHO\u2217 should be the most thermodynamically favorable state derived from OCCO\u2217.\n33\n Their work proposed that the energy barrier for \u2217CO dimerization and surface hydrogenation of OCCO\u2217 to form OCCHO\u2217 on Cu(100) are dramatically lower than those on Cu(111) (Figure\u00a01C),\n33\n indicating that Cu(100) is more active than Cu(111).Garza et\u00a0al. presented the free energy change with voltage for the reaction \u2217CO\u00a0+ CO \u2192 \u2217COCO on Cu(100) and Cu(111).\n34\n For Cu(100), only at high overpotentials, \u2217CHO was predicted to be more stable than \u2217COCO, while on the Cu(111) surface, \u2217COCO was highly unstable compared with its components at all the calculated potentials. Thus, on Cu(100) at low overpotentials, \u2217CO dimerizes and reduces to \u2217COCHO, leading to C2+ products. On Cu(100) at high overpotentials and on Cu(111) at all potentials, \u2217CO reduces to \u2217CHO, which can lead to both CH4 and C2+ products (Figure\u00a01D).\n34\n The results were consistent with experiments that C2H4 and CH4 share a common intermediate on Cu(111) and at high potentials on Cu(100), while the C2H4 formation proceeds on a distinctive path at low overpotentials on Cu(100).\n35\n\n,\n\n37\n\n,\n\n38\n The calculations by Goodpaster et\u00a0al. showed that the kinetic barrier for CO dimerization on Cu(100) increases with an increasing negative voltage and vice versa for the \u2217CHO formation,\n39\n supporting the mechanism proposed by Garza et\u00a0al.\n34\n and Ou et\u00a0al.\n40\n\nThere is also a possibility of reducing \u2217CO to \u2217COH, rather than \u2217CHO. Different studies have investigated this question, with different computational techniques employed.\n37\n\n,\n\n39\n\n,\n\n41\u201343\n In gas-phase computational hydrogen electrode (CHE) calculations, Nie et\u00a0al. found that \u2217COH was favored,\n43\n while the opposite trend was observed in explicit solvent simulations.\n41\n Akhade et\u00a0al. found that the presence of adsorbed K+ ions increases the selectivity for reducing \u2217CO to \u2217CHO by stabilizing these species and destabilizing the transition state to form \u2217COH.\n42\n Garza et\u00a0al. proposed that the stability of \u2217COH depends on the adsorption site: on a hole site of a Cu(100) surface, \u2217COH is close in energy to \u2217CHO.\n34\n In contrast, on a Cu(100) site (the preferred site of CO top adsorption),\n44\n the energy of \u2217COH is higher than \u2217CHO.\n34\n Thus, the authors concluded that \u2217COCHO is much more stable than \u2217COCOH, especially at high potentials.However, a recent study by Cheng et\u00a0al. suggested that the C\u2212C coupling is not the rate-determining step in CORR.\n45\n The authors investigated the CO adsorption isotherms on Cu in a broad pH range. Combining with the electrokinetic data, the authors demonstrated that the reaction orders of adsorbed CO at p\nCO <0.4 and >0.6 atm are first and zeroth, respectively, for C2+ products on three Cu catalysts. Thus, the authors proposed that the hydrogenation of CO with adsorbed water is the rate-determining step in CORR, and the site competition between CO and water leads to the observed transition of the CO reaction order.The pathways toward C2H4 and EtOH seem to proceed via several common intermediates, as C2H4 and EtOH are known to have similar onset potentials, suggesting that they share a common potential-limiting step.\n46\u201348\n In addition, experimental observations have shown that the Faradaic efficiencies (FEs) of these two products often shift similarly when different alkaline cations are used in the electrolyte.\n49\n Calle-Vallejo et\u00a0al. presented that the bifurcation step between C2H4 and EtOH is the conversion of \u2217CH2CHO intermediate.\n31\n Deoxidation of \u2217CH2CHO leads to the C2H4 pathway, while hydrogenation of the intermediate to \u2217CH3CHO is the penultimate step of the EtOH formation.\n31\n Garza et\u00a0al. proposed that the bifurcation step between C2H4 and EtOH on Cu(100) surface is the conversion of \u2217COCHO intermediate (Figure\u00a02B).\n34\n The formation of \u2217COCHOH intermediate leads to the C2H4 pathway, of which the activation energy is 0.49 eV at 1.0\u00a0V versus RHE. By comparison, the formation of glyoxal (OHC\u2212CHO) leads to the EtOH pathway, with the activation energy of 0.58 eV at the same electrode.\n34\n These findings may also explain the origin of a more favorable C2H4 production on Cu surfaces.The role of water has also been explored extensively in the formation of C2H4 and EtOH. Xiao et\u00a0al. used grand canonical quantum mechanics to predict the detailed mechanisms on Cu(111) and suggested that the surface water may directly donate a proton to the OH group of the carbon-containing intermediate and favor the dehydration reaction (the activation energy is 0.86 eV), thus benefiting the formation of C2H4.\n30\n In contrast, the dehydration reaction is kinetically blocked by surface water (the activation energy is 1.13 eV), thus hindering the formation of EtOH (Figure\u00a02C).\n30\n To test the possibility that oxygen in the product might arise from water rather than from CO, Lum et\u00a0al. conducted electroreduction of C16O in H2\n18O electrolyte on oriented Cu surfaces and found that 60%\u201370% of EtOH contained 18O, which was originated from the solvent.\n51\n The authors further extended the previous all-solvent density functional theory (DFT) meta-dynamics calculations to consider the possibility of incorporating water,\n50\n\n,\n\n52\n and found a new mechanism involving water molecules in a concerted reaction with the \u2217C\u2212CH intermediate to form \u2217CH\u2212CH(18OH), subsequently leading to (18O)EtOH (Figure\u00a02D).\n50\n By comparison, the conversion of \u2217C\u2212CH intermediate into \u2217C\u2212CH2 leads to the formation of C2H4, which competes with the formation of EtOH.Acetate is not a major CO2RR product on Cu-based electrodes, and the production of acetate with considerable selectivity at high current densities has only recently been demonstrated in CORR under highly alkaline conditions.\n14\n\n,\n\n15\n Thus, the mechanistic understanding of the acetate formation from CORR/CO2RR is relatively limited compared with other C2+ products. Birdja et\u00a0al. proposed that acetate and EtOH seem to be formed from the Cannizzaro disproportionation of acetaldehyde.\n53\n However, in experiments, the molar production of acetate observed often greatly exceeds that of EtOH in CORR, suggesting that the Cannizarro disproportionation is not a dominant pathway.\n15\n\n,\n\n54\n\nTo gain more insights about the formation of acetate, the isotopic labeling experiments using C18O were conducted to verify the reaction pathways toward acetate. For instance, Lum et\u00a0al. performed C16O reduction in H2\n18O and found that one O atom in acetate originates from CO and the other O is from electrolyte (Figure\u00a03A).\n50\n Coincidently, Jouny et\u00a0al. performed isotopic labeled C18O reduction on Cu in flow cells and gained a similar result that only one O of acetate was labeled, supporting that the O from electrolyte was also involved in the formation of acetate.\n50\n\n,\n\n54\n Based on the results of isotopic labeling experiments, the authors proposed that the observed acetate consisted of one O originating from CO and one O originating from an OH\u2212 anion reacting with intermediate species.To further understand the acetate formation pathway on Cu surfaces, Luc et\u00a0al. performed DFT calculations and proposed a pathway toward acetate (Figure\u00a03B).\n15\n This pathway involves water incorporation into ethenone (CH2\u2013CO), a ketene species, to form acetic acid (CH3\u2013COOH). In addition, Li et\u00a0al. suggested a same mechanism on oxide-derived Cu that the formation of acetate probably arises from attacking of OH\u2212 on a surface-bound ketene or other carbonyl-containing intermediate after C\u2013C bond formation.\n18\n This contention is supported by the observation of substantially increased acetate formation when increasing the KOH concentration from 0.1 to 1 M.In the follow-up work, Jouny et\u00a0al. demonstrated that C\u2013N bonds can be formed through co-electrolysis of CO and NH3 with considerable acetamide selectivity (Figure\u00a03C).\n55\n The results verified the mechanism for the acetate pathway since ketene or ketene-like intermediate is a key driving force behind acylation of nucleophilic co-reactants. This ketene-involved mechanism may also explain why more CH3COO\u2212 is produced by CORR than CO2RR. The basic principle is that the higher local pH under CORR condition provides more OH\u2212 for the formation of CH3COO\u2212 than that under CO2RR condition.The elementary steps through which the C3 products are formed have not been well studied yet. Ren et\u00a0al. presented that the formation of n-propanol is initialed from the intermolecular C\u2013C coupling between CO and C2 intermediates,\n56\n followed by proton/electron transfer to form propionaldehyde (CH3CH2CHO).\n49\n\n,\n\n57\n Propionaldehyde is then reduced on the Cu sites to n-propanol. Similarly, Xiao et\u00a0al.\n30\n and Zheng et\u00a0al.\n10\n also reported that C3 formation can proceed via C\u2013C coupling between C2 and C1 intermediates. Hence, most of the related reports use this mechanism to interpret the formation of n-propanol.\n19\n\n,\n\n22\n\n,\n\n26\n\n,\n\n27\n\n,\n\n58\n\n,\n\n59\n The coupling mechanism between active C1 and C2 species can be classified into two different modes as CO\u2217\u2013CH2CHO\u2217\n51\n\n,\n\n56\n\n,\n\n57\n and CO\u2217\u2013OCCO\u2217,\n19\n\n,\n\n26\n\n,\n\n27\n respectively. Among them, the coupling of CO\u2217 and OCCO\u2217 is mostly used in CORR, as CO species are abundant on the surface (Figure\u00a03D).\n26\n Another mechanism of the n-propanol formation was also reported by using CO reaction with acetaldehyde.\n60\n On Cu surfaces, the inadequate stabilization of \u2217C2 intermediates leads to desorption rather than further intermolecular reduction with \u2217CO for C3 generation.\n58\n To ensure the production of C3 at high production rates, C2 intermediates must be formed and stabilized on the catalyst surface and thus be available to be coupled with adsorbed CO.\n27\n\nSince the hydrogenation processes are critical reaction steps for the conversion of CO to hydrocarbons and oxygenates, the electrolyte pH may also be essential for CORR. Experimentally, highly alkaline electrolytes have been shown to improve the formation rates of C2+ products in CORR on Cu-based surfaces,\n61\n while the CH4 formation is favored by a low-pH environment. Xiao et\u00a0al. predicted the atomic mechanisms underlying electrochemical reduction of CO,\n52\n finding that (1) at acidic pH, the C2+ pathways are kinetically blocked and the CH4 pathway proceeds through \u2217CO \u2192 \u2217COH \u2192 \u2217CHOH\u2192\u2217CH2 \u2192 \u2217CH3 \u2192 CH4; (2) at neutral pH, the C1 and C2+ pathways share the common \u2217COH intermediate, where the branch to C\u2212C coupling is realized by a CO\u2212COH pathway; and (3) at high pH, the C\u2212C coupling through adsorbed CO dimerization dominates, suppressing the C1 pathway by kinetics, thereby boosting selectivity for multi-carbon products. Liu et\u00a0al. found a \u223c0.36\u00a0V shift in overpotential for C2+ products at pH 7\u201313, which translates to over three orders of magnitude enhancement in C2+ activity (Figures\u00a04A and 4B).\n62\n This finding was ascribed to the decreased activation barrier of \u2217OCCO protonation with increasing pHs (Figure\u00a04C).\n62\n\nHowever, some reports have indicated that the formation mechanisms of C2H4 and EtOH are not affected by pH. Hori et\u00a0al. presented that the partial current densities of C2H4 and EtOH as a function of potential exhibited good correlations regardless of the pH value (Figure\u00a04D),\n57\n indicating that the formation of C2H4 and EtOH are likely independent of the electrolyte pH.\n57\n In this case, pH can affect the formation of C\u2212C bond by changing the electroreduction activity toward CH4.\n30\n\n,\n\n57\n\n,\n\n62\n For the same CH4 partial current density, an increase in the local electrode pH shifts the potential to a more negative position on both the RHE and the standard hydrogen electrode (SHE) scales. The suppression of CH4 pathway can subsequently lead to enhanced selectivities toward C2H4 and EtOH.\n49\n\n,\n\n57\n Li et\u00a0al. varied the concentrations of Na+ and OH\u2212 at the same absolute electrode potential, and demonstrated that higher concentrations of cations (Na+), rather than OH\u2212, exert the main promotional effect on the production of C2+ products.\n63\n The promotional effect of OH\u2212 determined at the same potential on the RHE scale is likely caused by larger overpotentials at higher electrolyte pH. The authors also suggested that highly alkaline electrolytes can be beneficial in improving the CORR performance at the device level due to engineering considerations, as the ion-exchange membrane and oxygen evolution reaction (OER) catalysts favor alkaline over neutral conditions. In addition, higher pH environment may reduce the required voltage in full cell operations, because the equilibrium potential of OER can shift to less positive values but the C\u2212C coupling chemistry remains unaffected.\n63\n\nFor the acetate pathway, as it is proposed to be formed through attacking of OH\u2212 on a ketene-like intermediate, high alkaline electrolyte may favor the acetate production.\n18\n\n,\n\n54\n Correspondingly, the enhanced acetate selectivity was observed at higher KOH concentrations (Figure\u00a04E).\n15\n\nAs an important part of electrolyte, the nature of cations in the electrolyte may influence the activity and selectivity of Cu for CORR. Hori et\u00a0al. reported that alkaline cations affect the CORR selectivity on polycrystalline Cu,\n49\n suggesting that larger cations favor the formation of C2+ species such as C2H4, EtOH, and n-propanol. Cation effects were explained by Hori et\u00a0al. in terms of the potential variation in the outer Helmholtz plane, which originates from a difference in the hydration number of different cations.\n49\n Larger cations are less hydrated and expected to adsorb more easily on the cathode surface, shifting the potential to more positive values, thereby steering the selectivity toward C2H4 instead of CH4. Such experimental observations were confirmed by Kyriacou et\u00a0al.\n64\n Later, P\u00e9rez-Gallent et\u00a0al. proposed potential-dependent and structure-sensitive cation effects for CORR on Cu electrodes,\n65\n and found the presence of larger cations may stabilize the hydrogenated dimer intermediate (OCCOH) at low overpotentials (Figure\u00a05A),\n65\n thereby promoting the formation of C2H4, especially on Cu(100).By systematically varying the concentration of Na+ and OH\u2212 at the same absolute electrode potential, Li et\u00a0al. found that the chelation of Na+ leads to a drastic decrease in the formation rates of C2+ products (Figures\u00a05B\u20135D).\n63\n One possibility was proposed by Bell and co-workers that the existence of alkali cations at the outer Helmholtz plane can offer field-assisted stabilization for intermediates with dipole moments, such as \u2217CO, \u2217OCCO, and \u2217OCCHO.\n66\n\n,\n\n67\n The modified local electric field at higher cation concentrations can lead to higher densities of \u201chot spots\u201d that favor the formation of C2+ products. Another possibility is that interfacial water structure is modified at higher concentrations of cations,\n68\n\n,\n\n69\n which can facilitate the C\u2212C coupling pathway by better solvating the transition state complex.In addition to cations, the concentration or species of anions may influence the CO adsorption, thus influencing the CORR process. For example, Sebastian-Pascual et\u00a0al. investigated the interfacial properties of Cu(111) and Cu(100) in phosphate buffer solutions in the presence of CO.\n70\n Combining ab initio molecular simulations with voltammetry experiments, they found that CO adsorbs on the surface in the potential region close to the desorption of phosphate species anions from the electrolyte. The predominant adsorbed species is HPO4\u2217 on Cu(100), while it is PO4\u2217 on Cu(111). Due to the lower binding energy of PO4\u2217 on Cu(111), the adsorption of CO on Cu(111) takes place at less negative potentials than on Cu(100). Ovalle et\u00a0al. systematically explored the adsorption and desorption of CO on polycrystalline Cu electrodes in the presence of specifically and nonspecifically adsorbing anions at different concentrations.\n71\n They found that, at an electrolyte concentration of 10\u00a0mM, the adsorption and desorption of COatop are virtually independent of the identity of the anions. In contrast, at an electrolyte concentration of 1 M, the COatop coverage is significantly affected by the electrolyte anions, and the saturation coverages of COatop are lower compared with those in 10\u00a0mM electrolytes. The magnitude and mechanism of the modulation depend on the identity of anions. Weakly or nonspecifically adsorbing anions (SO4\n2\u2212, ClO4\n\u2212) limit the COatop saturation coverage by blocking a fraction of CO adsorption sites. Chloride ions, which specifically adsorb on Cu electrodes, can lower the CO coverage by modulating the CO adsorption energy.As the adsorption of intermediates is affected by the surface coverage of CO due to adsorbate-adsorbate interactions,\n72\n tuning the coverage of adsorbed \u2217CO may influence the binding of surface intermediate and subsequently the product selectivities. Sargent and co-workers investigated the influence of CO coverage for the formation of ethylene and oxygenates, and found that lower CO coverages stabilize the ethylene-relevant intermediates, whereas higher CO coverages favor the oxygenate formation (Figure\u00a06A).\n61\n Noskov and co-workers presented a DFT study on the effect of CO coverage for the CO\u2212CO coupling energy on Cu, and suggested that the CO dimerization barrier becomes lower as the \u2217CO coverage increases for all facets (Figure\u00a06B).\n73\n Zheng and co-workers investigated the influence of CO coverage for the formation of acetate, and found that higher CO coverage on the surface favors the C\u2212C coupling as well as the stabilization of ethenone intermediate, which is a key intermediate for acetate formation, thus promoting the formation of acetate.\n14\n\nGenerally, alloying is an effective strategy to tune the electronic structure and thus modulate the intrinsic adsorption property of intermediates.\n74\n For example, Wang et\u00a0al. demonstrated the improvement of CORR selectivity to n-propanol by an Ag-doped Cu catalyst.\n19\n As a result of strain and ligand effects, the Ag doping in Cu leads to two classes of neighboring Cu atoms with distinct electronic structures. A pair of adjacent Cu atoms with different electronic structures can act as an C\u2212C coupling active site to promote both C1\u2212C1 coupling and C1\u2212C2 coupling, thus resulting in the formation of C3 products (Figures\u00a07A and 7B).\n19\n These findings are analogous to the enhanced coupling effect of Cu0 and Cu+, proposed by Xiao et\u00a0al.\n75\n Li et\u00a0al. demonstrated a Pd-doped Cu catalyst to tune the adsorption of hydrogen at the Cu surface and thereby promote the formation of alcohols.\n20\n The introduction of Pd provides optimal H-binding for alcohol production at neighboring Cu sites, hydrogenating post-C\u2212C coupling of reaction intermediates along the alcohol pathway. The synthesized Pd-doped Cu catalyst achieved a Faradaic efficiency of 40% toward alcohols and a partial current density of 277 mA cm\u22122 from CORR, which was a 2-fold increase in the alcohol-to-ethylene ratio compared with the bare-Cu catalyst (Figure\u00a07C).\n20\n Wang et\u00a0al. demonstrated that planar CuAg electrodes can reduce CO to acetaldehyde with over 50% FE and over 90% selectivity at a modest electrode potential.\n21\n The FE to acetaldehyde was further enhanced to 70% by increasing the roughness factor of the CuAg electrode. The authors indicated that Ag ad-atoms on Cu weaken the binding energy of the reduced acetaldehyde intermediate and inhibit its further reduction to EtOH, suggesting that the improved selectivity to acetaldehyde is due to the electronic effect from the Ag incorporation.\n21\n\nIntroducing dopants with stronger CO\u2217 adsorption than Cu may excessively strengthen the surface CO\u2217 and promote H\u2217 adsorption concurrently,\n74\n\n,\n\n76\n while it is also critical to suppress H2 evolution. To enrich the surface \u2217CO coverage and inhibit\u00a0the competing hydrogen evolution reaction (HER) simultaneously, Zheng and\u00a0co-workers developed an atomically-ordered CuPd intermetallic compound catalyst composed of a high density of Cu-Pd pairs that feature as catalytic sites, enabling FEacetate of \u223c70% and an acetate partial current density of 425 mA cm\u22122.\n14\n The combination of Cu and Pd can enrich surface \u2217CO coverage, stabilize ethenone as a key acetate-path intermediate, and inhibit HER due to the absence of Pd clusters,\u00a0thus substantially promoting the acetate formation. In contrast to the ordered\u00a0CuPd catalyst, aggregated Pd atoms hinder \u2217CO reduction as they bind those\u00a0carbonic species too strongly, leading to the formation of H2 (Figures\u00a07D\u20137G).\n14\n\nIn addition to binary alloys, ternary alloys were also used to modulate the intrinsic adsorption property of intermediates during CORR. For example, Wang et\u00a0al. presented a silver-ruthenium co-doped copper (Ag\u2013Ru\u2013Cu) catalyst, which shows high selectivity, production rate, and stability for n-propanol electrosynthesis.\n22\n The co-doping of Ag and Ru in Cu induces CO adsorption near the C1\u2013C1 and C1\u2013C2 coupling sites, and thus results in higher \u2217CO coverage on the surface than Ag\u2013Cu or Cu, which can promote multiple C\u2013C coupling. Additionally, the adsorption energy of key C2 intermediate for C1\u2013C2 coupling on Ag-Ru\u2013Cu is higher than that on Ag\u2013Cu or Cu; this may reduce the desorption of C2 intermediates, thus increasing the residence of C2 intermediates necessary for C3 generation (Figure\u00a07H).\n22\n\nIn addition to regulating the adsorption of intermediates through electronic effects, alloying can also affect the catalytic performance by changing the catalyst geometry. For instance, Guan et\u00a0al. reported a series of Cu-Au alloys with different compositions as CORR catalysts, which enabled substantially different product distributions of CH4 and C2H4 (Figures\u00a07I and 7J).\n23\n Compared with pure Cu catalyst with a good C2H4 selectivity, the introduction of Au causes steric hindrance and increases the distance between two adsorbed CO intermediates at neighboring Cu sites, thus enhancing the C1 pathway by generating \u2217CHO intermediate and leading to the formation of CH4.Single-atomic catalysts have been becoming a research hot spot due to their unique electronic structures of metal atoms coordinated with nonmetal atoms, and a variety of metal single-atomic catalysts have been demonstrated with excellent CORR performances.\n24\n\n,\n\n25\n For example, Bao et\u00a0al. developed a Cu single-atomic catalyst anchored to Ti3C2Tx nanosheets for CORR.\n24\n The atomically dispersed Cu\u2013O3 sites favor C\u2013C coupling to generate the key \u2217CO\u2013CHO species, and then induce the decreased free energy barrier of the potential-determining step, leading to a high selectivity for C2+ products (Figures\u00a08A\u20138D).\n24\n\nIn addition to the single-atomic catalysts, dual-atomic catalysts are also effective to produce C2+ products, in which metal atom pairs may contribute synergistically to favor the coupling of two CO molecules. For instance, Li et\u00a0al. developed a dual metal atomic catalyst with uniform distributions of two adjacent Cu-Cu or Cu-Ni atoms anchored on nitrogen-doped carbon frameworks, featuring distinctive catalytic sites for CORR.\n25\n The dual Cu atomic sites facilitate electroreduction of two CO molecules and subsequent carbon coupling toward ethylene and acetate. The replacement of one of the dual Cu atoms with Ni results in too strong CO adsorption, and thus only one single Cu atom functions as the catalytic site for the C1 reduction pathway (Figures\u00a08E\u20138G).\n25\n Hence, the rational design of new single-atomic or dual-atomic catalysts is a potential means to tune the product selectivities of CORR.In 1995, Hori et\u00a0al. firstly reported a facet-dependent selectivity of CO electroreduction.\n77\n C2H4 was favorably produced on Cu(100), and CH4 was predominantly yielded on Cu(111). The (110) electrode shows an intermediate product selectivity between (100) and (111). Thus, the synthesis of Cu-based catalysts with high Cu(100) exposure serves as a guideline for designing C2H4-selective catalysts. Using online electrochemical mass spectrometry (OLEMS), Schouten et\u00a0al. monitored the reduction of CO on Cu(111) and Cu(100) and suggested that C2H4 can be formed via two different pathways.\n37\n On Cu(100) (Figure\u00a09A), CO is reduced to only C2H4 and not CH4 at relatively low overpotentials, presumably through the formation of a surface-adsorbed CO dimer. On both Cu(100) and Cu(111) (Figure\u00a09B), at higher overpotentials, CO is reduced to C2H4 and CH4 simultaneously, suggesting a shared intermediate.\n37\n Later, those authors further compared the formation of C2H4 on Cu(100) and Cu(111) at different pH values to explore the two pathways for C2H4.\n79\n It was found that, although the formation of C2H4 on Cu(111) is clearly pH dependent, it is pH independent on Cu(100). The observed pH independence for the formation of C2H4 on Cu(100) supports the formation of the CO dimer on this crystal facet.Furthermore, the selectivity of acetate can be enhanced by decreasing the surface\u00a0exposed Cu(100), as the ethenone-involved acetate pathway has no facet preference. \n14\n For instance, Kang and co-workers reported a Cu nanosheet catalyst with two-dimensional triangular-shaped morphology, which selectively exposed (111) facets.\n15\n The catalyst showed an enhanced acetate selectivity of \u223c48% by suppressing the formation of C2H4 and EtOH (Figures\u00a09C and 9D).\n15\n As a result of the high exposure of Cu(111), an increase in the CH4 selectivity was also observed on these Cu nanosheets at potentials more negative than \u22120.7\u00a0V versus RHE.In addition to the facet types, the grain boundaries can also influence the CORR catalytic performances. Kanan and co-workers showed that the CO reduction activity is directly correlated with the density of grain boundaries in Cu nanoparticles.\n78\n The authors prepared Cu nanoparticles with different average grain boundary densities, quantified by transmission electron microscopy, and found that the specific activity for CO reduction to EtOH and acetate was linearly proportional to the fraction of Cu nanoparticle surfaces composed of grain boundary surface terminations, suggesting that grain boundaries alter surface properties of the catalyst to lower the reaction barrier (Figures\u00a09E and 9F).\n78\n They also suggested that grain boundaries may create surfaces with strong CO binding sites and these surfaces are responsible for the catalytic activity. The properties of grain boundaries can also be used to guide the catalyst design to promote the CORR performance. As the Cu(111) facet is C1-selective, whereas Cu(100) is C2-selective,\n46\n\n,\n\n80\n\n,\n\n81\n to promote the formation of n-propanol, Pang et\u00a0al. designed a highly fragmented Cu catalyst composed of a mixture of Cu(111) and Cu(100) facets, thereby creating additional opportunities to couple C1 and C2 intermediates.\n26\n The highly fragmented Cu catalyst presented an n-propanol selectivity of 20%, and a reaction rate that corresponds to a partial current density of 8.5 mA cm\u22122 for n-propanol.Among numerous catalyst design strategies, regulation of the catalyst morphology is also a way to affect the CORR performance. For instance, Zhuang et\u00a0al. developed a method of synthesizing open Cu nanocavity structures with a tunable geometry via the electroreduction of Cu2O cavities formed during acidic etching.\n27\n The cavity morphology can promote C2\u2212C1 coupling inside a reactive nanocavity via the nanoconfinement of C2 intermediates, thus enhancing C3 formation (Figures\u00a010A and 10B).\n27\n Wang et\u00a0al. developed a hierarchical Cu nanoflower electrodes to increase the roughness factor of the catalyst, which showed nearly 100% CORR selectivity for liquid oxygenated products at \u2012 0.23\u00a0V versus RHE, although the corresponding partial current densities remained relatively low (Figures\u00a010C\u201310E).\n17\n The authors demonstrated that the increased roughness factor of the electrode improves the selectivity for C2+ oxygenates by not only enabling operation at low overpotentials but also by suppressing competing HER.\n17\n\nOxide-derived Cu (OD-Cu) has often been reported to present higher CORR selectivities toward C2+ products than polycrystalline Cu.\n18\n By using in situ attenuated total reflectance surface-enhanced infrared absorption spectroscopy (ATR-SEIRAS), Lee et\u00a0al. showed that CO binds more strongly on the surface of copper oxides than pure metallic Cu.\n82\n As a result, CO can be densely populated on the surface and subsequently favor C\u2212C coupling toward C2+ products.\n82\n Kanan and co-workers presented that nanocrystalline Cu prepared from Cu2O (i.e., OD-Cu) produces EtOH, acetate and n-propanol with a total FE up to 57% at modest overpotentials.\n18\n By comparison, Cu nanoparticles with an average crystallite size similar to that of OD-Cu produce nearly exclusive H2 under identical conditions. The authors demonstrated that the enhanced CORR activity on OD-Cu electrodes relative to Cu nanoparticle electrodes is consistent with the presence of highly active sites on their grain boundary surfaces,\n18\n suggesting that engineering the grain boundaries by altering the oxide reduction method of nanocrystalline materials can improve activity and selectivity toward long-chain products.Because of the low solubility of CO (\u223c1\u00a0mM) in H2O, the geometric current density achieved in solution-phase CORR is on the order of 1 mA cm\u22122.\n13\n Gas diffusion electrodes (GDEs) have been designed to overcome the solution-phase mass-transport limits by creating gas-solid-liquid triple-phase boundaries, comprising CO, Cu-based catalysts, and electrolytes. The use of GDEs can enhance the CORR current densities by at least one order of magnitude. For example, using GDEs, Jouny et\u00a0al. presented a high CORR performance with a well-controlled electrode-electrolyte interface, with a total current density of up to 1 A cm\u22122,\n54\n suggesting attractive potentials of developing CO electrocatalysis in those systems. From a techno-economic perspective, commercially relevant productivity requires a current density of >200 mA cm\u22122.\n83\n\n,\n\n84\n Hence, the utilization of GDE-based systems is necessary for the CORR research.To increase the single-pass conversion rates and enlarge product concentrations, various design strategies have been explored for GDE-involved reactors. Kanan and co-workers investigated CO electrolysis with GDEs supplied by inter-digitated flow fields in electrochemical cells with different ion transport properties.\n13\n Using a cell with the GDE directly contacting with a Nafion membrane, the authors demonstrated >100 mA cm\u22122 partial current density for CO reduction to C2+ products, and direct production of 1.1\u00a0M acetate at a cell potential of 2.4\u00a0V for over 24\u00a0h (Figures\u00a011A\u201311C).\n13\n Wang and co-workers reported continuous generation of high-purity (up to 96%) acetic acid solutions via CORR on Cu nanocube catalysts in a porous solid electrolyte reactor.\n85\n Different from conventional liquid electrolyte, the porous solid electrolyte layer can efficiently conduct ions and does not introduce impurity ions into the generated liquid products. The porous solid electrolyte reactor can enable the continuous generation of electrolyte-free acetic acid solutions, with a current density of 1 A\u00b7cm\u22122 and a high acetic acid purity up to \u223c96% (Figures\u00a011D\u201311G).\n85\n\nRenewable-energy-powered electrocatalysis using Cu-based catalysts has been demonstrated as an attractive approach for converting CO2 or CO into multi-electron-transferred chemicals. Compared with direct conversion of CO2, a two-step cascade approach, in which CO2 is initially reduced to CO and subsequently into C2+ products, has been emerging as a carbonate-formation-free platform that can be operated under high-alkalinity conditions. In the past few years, significant progress has been achieved in CORR, including mechanistic studies, catalyst design, and system development. With the exploration of the reaction and further understanding of mechanism, it is promising to design Cu-based catalysts to selectively produce different C2+ products. The utilizations of GDE-based electrolyzers have exhibited a practical application prospect for CORR in chemical industry. Despite the achievements thus far, several key challenges that preclude the large-scale deployment remain to be solved. Below, we suggest some perspectives to push this technology into the next level.First, it is still difficult to know the comprehensive dynamic process and reaction mechanism of CORR. To understand the reaction mechanisms and further optimize catalyst performances, it is crucial to conduct operando measurements of reaction intermediates and catalytic products under experimental conditions, especially at high current densities. Thus, the development of techniques that can identify these reaction intermediates under in situ and operando conditions are highly desired. Second, computational investigation is a powerful tool to accelerate mechanism understanding. However, it is challenging for the present computational models to precisely describe comprehensive electrolyte-catalyst interfaces. Hence, development of new computational models that can accurately simulate solvated cations, protons, and hydroxide ions at the reaction interface is crucial to the CORR mechanism development. Third, to meet the needs of commercially relevant performances, the stability of CORR should be further promoted by rational designs applied to both catalysts and reaction systems. Finally, to reduce subsequent separation cost, it is critical to optimize the catalyst selectivity for increasing the purity of single C2+ products, which also requires substantial development in the design of reactors.The authors thank the following funding agencies for supporting this work: the National Key Research and Development Program of China (2018YFA0209401, 2017YFA0206901), the National Natural Science Foundation of China (22025502, 21975051), the Science and Technology Commission of Shanghai Municipality (21DZ1206800), and the Shanghai Municipal Education Commission (2019-01-07-00-07-E00045).G.Z. conceived the article\u2019s structure and supervised this work. G.Z., Y.J., and A.G. collected the literature, wrote the original manuscript, and participated in discussions and revisions of the manuscript.The authors declare no competing interests.We support inclusive, diverse, and equitable conduct of research.", "descript": "\n                  Electrochemical reduction of carbon monoxide has recently emerged as a potential approach for obtaining high-value products. Recent studies have shown that carbon monoxide can be electrochemically reduced to C2+ at high reaction rates, high selectivity, and inherently improved stability compared with carbon dioxide, highlighting the attractive potential of the electrocatalytic carbon monoxide reduction reaction (CORR). This review introduces recent progress in CORR mechanistic understanding and representative strategies for the design of copper-based electrocatalysts, such as alloying and doping, single-atom catalysts, crystal facet and morphology design, and oxide-derived copper. Several critical factors that influence the CORR activity and selectivity are also discussed, including pH, electrolytes, and the design of electrolysis reactors. Finally, the challenges and perspectives for further development in this field are summarized.\n               "}
{"full_text": "Atmospheric methane (CH4) from various sources, including biogases and natural gas leaks, is a significant greenhouse gas (GHG) with greenhouse effect much severer than carbon dioxide (CO2) [1\u20133]. The conversion of biomethane to hydrogen (H2) and useful solid carbon materials can substantially mitigate GHG emission by the carbon-negative effect and meanwhile supply renewable H2 to economic development. However, methane as the most stable hydrocarbon requires extremely high temperature (~1200\u00a0\u00b0C) for noncatalytic thermal decomposition, which produces H2 and low-value carbon black by\n\n(1)\n\n\nCH\n4\n\n\ng\n\n\u2194\nC\n\ns\n\n+\n2\n\nH\n2\n\n\ng\n\n,\n\u2206\n\nH\nr\no\n\n\u2248\n75\n\nkJ\n/\nmol\n\n\n\nIn the past years, continued research efforts have been made on catalytic decomposition of methane (CDM) aiming to reduce reaction temperature, enhance conversion rates, and control the morphology of carbon products. In the literature, metallic nanoparticle catalysts are commonly used for CDM at drastically lowered temperatures to form multiwalled carbon nanotubes (MWCNTs) and carbon nanofibers (CNFs). However, CDM is an endothermic and volume-increasing reaction that is still often performed at high temperatures (>700\u00a0\u00b0C) and far below atmospheric pressure to kinetically and thermodynamically favor the CH4 conversion (\u03c7\n\nCH4) [4\u201312].In general, carbon deposition and CNT growth from hydrocarbons on metallic catalysts involve multiple steps, including (i) metal surface-catalyzed dehydrogenation and carbonization requiring an activation energy (E\n\na, s\n), (ii) activated carbon (C) dissolution and diffusion through the metal phase, and (iii) dissociation of metal\u2011carbon bonds to release graphitic carbons such as graphene layers, CNT, and CNF [6,11,13]. The kinetic and thermodynamic behaviors of hydrocarbon decomposition and carbon formation are thus determined by the E\n\na, s\n, the C-metal binding energy (\u2206H\n\nf, MC\n\n\no\n), diffusion activation energy (E\n\na, d\n), and carbon solubility (S\n\nCM\n) in the metals. Metals with very large negative \u2206H\n\nf, MC\n\n\no\n values cause transport limitation by the excessive C-metal binding strength while metals with positive \u2206H\n\nf, MC\n\n\no\n lead to reaction limitations due to hindered dissolution and removal of the carbon products from catalyst surface. Thus, metals exhibiting moderate \u2206H\n\nf, MC\n\n\no\n and \u2206E\n\na, s\n and adequate S\n\nCM\n, such as Fe, Ni, and Co, are well-suited for chemical vapor deposition (CVD) of hydrocarbons [14\u201317]. Also, the \u2206E\n\na, s\n of hydrocarbon dissociation is greater for terrace sites than for step edge sites that results in preferential formation and dispense of graphene layers at the step edges leading to CNT growth on nanoparticles [13,18\u201321].Non-metallic catalysts, especially metal oxides, could also significantly lower the temperature for CDM but often produce carbon nanomaterials in more disordered structures [4,22]. The mechanisms of CNT formation on oxides are still not fully understood but the insolubility of carbon in solid oxides could cause transport limitation to hinder the formation of long CNTs and CNFs. The CNT formation on various oxides, e.g., SiO2, ZrO2, TiO2, Fe2O3, and Al2O3, have been investigated by in situ spectroscopic characterizations [20,23\u201325]. It is generally believed that carbonaceous radicals first form by chemisorption and dehydrogenation on the oxide surface and then migrate via surface diffusion to the catalyst tip where supersaturation is achieved to form and dispense CNTs [25\u201327].In recent studies, mixtures containing carbon monoxide (CO) and CH4 have been demonstrated for CVD conversion to MWCNTs at dramatically lowered temperatures (~ 500\u00a0\u00b0C) on in situ generated Fe metallic nanoparticles and partially reduced ferrite-type nanoparticle catalysts [22,28]. Such a mixed feed is of practical interest because of the potential to couple with the syngas production from CO2 dry reforming of CH4, i.e., CO\n2\u00a0+\u00a0CH\n4\u00a0\u2194\u00a02H\n2\u00a0+\u00a02CO [29]. The CVD conversion of the CO/CH4 mixtures apparently involve simultaneous endothermic CDM and the following highly exothermic CO-disproportionation, i.e., the Boudouard reaction,\n\n(2)\n\n2\nCO\n\ng\n\n\u2194\nC\n\ns\n\n+\n\nCO\n2\n\n\ng\n\n,\n\u2206\n\nH\nr\no\n\n\u2248\n\u2212\n172\n\nkJ\n/\nmol\n\n\n\nPure CO is a precursor for growing high-purity CNTs on iron-based nanoparticles, which also involve CO surface dissociation, formation of metal carbide, and subsurface carbon transport [30,31]. The metal or oxide catalyzed CVD of pure gases were found to exhibit overall first order for CH4 and second order kinetics for CO at low conversions [32,33]. However, the overall rate of carbon formation by reaction of CH4/CO mixtures on the Fe metal nanoparticles was found to exhibit a linear dependency on the square root of the product of reactant concentrations (i.e., [CH4]0.5\u00b7[CO]0.5) [4]. These indicate that synergistic effects between CH4 and CO have caused the mixture reaction kinetics to drastically deviate from those of the pure gas reactions on the catalysts.Here, a Cr-doped ferrite (FeCr) nanocrystalline catalyst is demonstrate for CDM to H2 and MWCNTs with cofed CO at a low temperature of 500\u00a0\u00b0C. The nanocrystalline FeCr catalyst is a well-known H2S-resistant catalyst for water-gas shift (WGS) reaction with intrinsic stability in H2O and CO2 atmospheres [34,35]. These basic properties of the FeCr catalyst are highly desirable for conversions of biomethane, which commonly contain H2S impurity. In our recent studies, the FeCr nanoparticles were found to induce MWCNT formation while suppressing methanation during WGS reaction at 500\u00a0\u00b0C under high pressure and H2-lean conditions [22]. The FeCr-catalyzed MWCNT formation in WGS conditions appeared to be associated with the byproduct CH4 and methanation intermediates. The involvements of both CH4 and CO were further evidenced by the rapid conversion of a CH4(20%)/CO(80%) dry mixture to H2 and MWCNTs at conditions corresponding to the WGS reaction. The present work focuses on experimental investigation on the synergistic effects of CH4 and CO on the FeCr-catalyzed CVD conversion of CH4/CO mixtures at a low temperature of 500\u00a0\u00b0C and atmospheric reaction pressure.The nanocrystalline Cr-doped iron oxide (FeCr) catalyst with an atomic ratio of Fe:Cr\u00a0=\u00a010:1 was synthesized by the coprecipitation method from a mixed salt solution as explained in our earlier work [36]. The starting solution for coprecipitation contained iron nitrate and chromium nitrate precursors with Fe:Cr atomic ratio of 10:1. An aqueous ammonia solution was added to the mixed salt solution under stirring to induce solid precipitation until completion at pH around 9. The precipitated solids were recovered by filtration and subsequently dried at 80\u00a0\u00b0C for 12\u00a0h. The dried particles were then calcined at 500\u00a0\u00b0C in air for 3\u00a0h using heating and cooling rates of 5\u00a0\u00b0C/min. The calcined solid was ground intensively and mesh-sieved to remove the large agglomerates. The microscopic images and elemental analysis results, which are presented in later sections, showed that the actual primary grain size and Fe:Cr atomic ratio of the as-synthesized FeCr oxide were 15\u201325\u00a0nm and\u00a0~\u00a010, respectively.The CVD reaction was carried out in a packed-bed reactor (PBR) made of a fused silica-coated alumina tube with outer and inner diameters of 0.57\u00a0cm and 0.39\u00a0cm, respectively. The tube had a total length of 10\u00a0cm with each end mounted with a\u00a0~\u00a01\u00a0cm-thick multilayered glass cloth (Fig. 1\n). In most cases, 0.1\u00a0g of catalyst particles was distributed on loosely packed quartz wool along the tube reactor. After the PBR was installed in a programmable furnace with a 1.5-m long preheating coil. The catalyst was activated by partially reducing the hematite (Fe2O3) to magnetite (Fe3O4) in a processing gas flow containing H2, CO, CO2, and H2O, which had a reductant to oxidant ratio of R/O\u00a0=\u00a01.4, where R/O = (p\n\nCO\n\u00a0+\u00a0p\n\nH2)/(p\n\nCO2\u00a0+\u00a0p\n\nH2O\n) [34,36]. The activation process was performed at 400\u00a0\u00b0C for 4\u00a0h using heating and cooling rates of 5\u00a0\u00b0C/min to achieve partial reduction of Fe3+ to Fe2+ without formation of metallic Fe and crystallite growth [22,35]. The catalyst bed was then purged with pure N2 gas at a flowrate of 20\u00a0cm3 (STP)/min for at least 3\u00a0h while the temperature was increased to 500\u00a0\u00b0C and stabilized before switching to a reactant gas flow.The entire reaction system was similar to that described in our previous reports [22,37]. It included mass flow controllers for regulating the feed flowrate and composition and an online GC\u2013MS (GC, Agilent GC 6890\u00a0N and MS Agilent 5975B) for analyzing product gas compositions. The GC was equipped with a thermal conductivity detector (TCD). The exiting gas flowrate was also monitored by a film flowmeter at a 10-min interval for verification. A digital pressure gauge was used at the reactor entrance while the PBR exit remained at atmospheric pressure. The actual temperature fluctuation of the furnace hosting the PBR was \u00b13\u00a0\u00b0C over the reaction durations. The efficient heat exchange through the small-diameter ceramic tube was able to stabilize the gas temperature at the PBR exit, which varied between 495 and 510\u00a0\u00b0C depending on the feed gas composition.The gas reactions were performed in the PBR immediately after the catalyst activation process. The reaction temperature and exiting pressure were 500\u00a0\u00b0C and ambient pressure (1.013\u00a0bar), respectively. The feed pressure at the PBR entrance was typically around 1.10\u00a0bar, which is slightly higher than the ambient exiting pressure to drive the gas flow. The reaction experiments were conducted for feed gases with CH4 mole fraction varying from 100% (i.e., pure CH4, 99.99%) to 0% (i.e., pure CO, 99.9%). The total feed flowrates were kept at around 16\u00a0cm3 (STP)/min for mixtures. In most cases, the reaction was terminated when the feed entrance pressure exceeded 1.5\u00a0bar under the fixed gas flowrate because of the increased flow resistance from blocking effect of the growing carbon deposits. After terminating the reaction, the reactor was cooled down to room temperature in a continued N2 flow. The solid products, i.e., the carbon-deposited catalysts, were immediately retrieved and sealed in glass vials to minimize possible oxidation by air before various ex situ characterizations.Pure CO and CH4 are common precursors for CNT production on metallic iron-based catalysts usually at temperatures much higher than 500\u00a0\u00b0C [38]. The low-temperature FeCr-catalyzed reactions were tested for pure CH4 and CO, respectively, with and without the pretreatment with the other. These experiments were performed to study the effects of pretreatment in one pure gas on the catalyst's activity in subsequent reaction with the other.Experiments were first carried out for CH4 and CO single gas reaction without pretreatment in the other. The feed flowrates were kept at ~5\u00a0m3 (STP)/min for pure gas reactions. The volume increasing CDM reaction is thermodynamically favored at low pressures. Thus, the reaction was tested for CH4 partial pressure varying from 0.05\u00a0bar to 1.013\u00a0bar in balancing N2, i.e., molar fraction of CH4 in the CH4/N2 mixture increasing incrementally from 5% to 100% at atmospheric pressure. The endothermic CDM is expected to be kinetically and thermodynamically limited at the low temperature of 500\u00a0\u00b0C. Hence, pure CH4 reaction was tested for 20\u00a0h and 126\u00a0h, respectively, to ensure observation of the final carbon morphology. The CO-disproportionation reaction is exothermic and volume-decreasing and therefore, is thermodynamically favored at low-temperature and high pressure. The reaction of the pure CO was conducted at atmospheric pressure (1.013\u00a0bar) and 500\u00a0\u00b0C for 20\u00a0h.To investigate the effects of catalyst pretreatment in one gas on the subsequent reaction with the other, experiments were conducted by switching feeds of pure CH4 and CO, respectively. All reactions were performed at 500\u00a0\u00b0C and 1.013\u00a0bar. For studying the effects of CO-pretreatment on the subsequent reaction with CH4, the pure CO flow was first fed a flowrate of 10\u00a0cm3 (STP)/min for 30\u00a0min; then the catalyst bed was purged by pure N2 gas at 10\u00a0cm3 (STP)/min for 1\u00a0h; and finally, the pure CH4 gas was fed at a flowrate of 10\u00a0cm3 (STP)/min for 21\u00a0h. For studying the effects of CH4-pretreatment on the subsequent reaction with pure CO, the pure CH4 was first fed at a flowrate of 10\u00a0cm3 (STP)/min for 30\u00a0min; the catalyst bed was then purged by pure N2 gas at 10\u00a0cm3 (STP)/min for 1\u00a0h; and finally, the pure CO gas was fed at a flowrate of 10\u00a0cm3 (STP)/min for 21\u00a0h. The flowrates and compositions of the exiting gas were monitored during the entire processes. The final solid products were retrieved and characterized by microscopic examinations and chemical analyses.The gas products of the gas reactions were monitored by the online GC, and the solid products including carbon materials and reacted catalyst particles were retrieved and examined by various ex situ characterizations. The crystalline phases of the catalysts were identified by X-ray diffraction (XRD) using a PANalytical X'Pert Pro diffractometer with Cu K\u03b1 radiation (\u03bb\u00a0=\u00a01.5406\u00a0\u00c5) and, in some cases, confirmed by transmission electron microscopy (TEM) and electron diffraction (ED). The morphology and dimensions of the solid materials were observed by SEM and TEM and their elemental compositions were determined by the energy dispersive X-ray spectroscopy (EDS) technique. The SEM-EDS analyses were performed by a FEI Scios DualBeam microscope equipped with Ametek Octane Super EDAX. The catalysts at different processing stages were analyzed by X-ray photoelectron spectroscopy (XPS) to study the metal oxidation state variations with reaction conditions. The XPS experiments were performed on a Thermo VG Scientific spectrometer using Al-K\u03b1 (1486.7\u00a0eV) radiation as the excitation source at room temperature. The pressure of the catalyst chamber was maintained below 10\u22128\u00a0Pa to avoid a large amount of noise in the spectra from contaminants. The obtained binding energies were adjusted by referencing the spectra to the carbon (C 1\u00a0s) peak at 284.6\u00a0eV. The carbon products were also characterized by Raman shift spectroscopy for assessment of the graphitic phase and crystallinity.The gas and solid products of the FeCr-catalyzed conversion of CH4/CO mixtures had consistent types of gas components and carbon morphology for varied CH4 mole fraction in the feeds (y\n\nf, CH4). The gas products contained CH4, CO, CO2, and H2, which were expected from the two primary reactions (1) and (2). Although H2O was not discernable by the TCD-equipped GC, the following secondary reactions likely coexisted [39]:\n\n(3)\n\n\nCO\n2\n\n\ng\n\n+\n\nH\n2\n\n\ng\n\n\u2194\nCO\n\ng\n\n+\n\nH\n2\n\nO\n\ng\n\n,\n\u2206\n\nG\nr\no\n\n\u2248\n+\n13\n\nkJ\n/\nmol\n\n\n\n\n\n(4)\n\n\nCO\n2\n\n\ng\n\n+\n\nCH\n4\n\n\ng\n\n\u2194\n2\nCO\n\ng\n\n+\n2\n\nH\n2\n\n\ng\n\n,\n\u2206\n\nG\nr\no\n\n\u2248\n+\n40\n\nkJ\n/\nmol\n\n\n\n\n\n(5)\n\nCO\n\ng\n\n+\n\nH\n2\n\n\ng\n\n\u2194\nC\n\ns\n\n+\n\nH\n2\n\nO\n\ng\n\n,\n\u2206\n\nG\nr\no\n\n\u2248\n\u2212\n18\n\nkJ\n/\nmol\n\n\nwhere the \u2206G\n\nr\n\n\no\n values are for standard pressure of 1\u00a0bar at 500\u00a0\u00b0C. These secondary reactions cause interdependences between the CH4 and CO reactions because of the participation of H2 and CO2 from their primary reactions (1) and (2). However, the CO2 reduction by H2 (eq. 3) and CO2-reforming of CH4 (eq. 4) are expected to be insignificant because they are thermodynamically unfavorable under such low temperature with low pressures of H2 and CO2. The involvement of CH4 steam reforming is not considered in this case because of the minimal amount of H2O from the very minor reactions (3) and (5) under low \u03c7\n\nCH4.The FeCr catalyst was first examined by SEM (Fig. S1), TEM, and EDS, which confirmed its primary particles size range of 10\u201325\u00a0nm and an overall Fe:Cr atomic ratio of ~10.6 (\u00b15%) (Fig. 2a). The XPS examination results (Fig. 2b) verified the partial reduction of Fe3+ to Fe2+ and Cr6+ to Cr3+ was achieved by activation in the processing gas (R/O\u00a0=\u00a01.4) without forming Fe0 or Cr0 [22,35]. In Fig. 2(c), the XRD pattern of the activated catalyst represents a combination of the Fe2O3 (hematite) and Fe3O4 (magnetite) spectra. There were no appreciable peaks to suggest any segregated phases of chromium trioxide, iron chromate, or chromium ferrite in the FeCr catalyst [40].The activated FeCr nanoparticle catalyst was able to actively catalyzed CH4 conversion to H2 and carbon nanomaterials in the entire range of y\n\nf, CH4. However, the morphology of the carbon products and conversions of CH4 (\u03c7\n\nCH4) and CO (\u03c7\n\nCO\n), as defined by Eq. (6), were found to depend on the y\n\nf, CH4 and differ markedly between the pure gases and mixtures.\n\n(6)\n\n\n\u03c7\ni\n\n=\n1\n\u2212\n\n\nmoles of\n\ni\n\nin product\n\n\nmoles of\n\ni\n\nin feed\n\n\n\n\n\ni\n=\n\nCH\n4\n\n\nCO\n\n\n\n\nThe results of SEM and Raman tests for the catalysts after reactions with pure CH4 and pure CO are presented in Fig. 2 (d-f). The carbon products from CH4 (Fig. 2e) and CO (Fig. 2f) were both of graphitic phases and primarily in particulate and rod-shape structures. However, for the same 20-h reaction duration, the amount of carbon from the pure CH4 reaction (Fig. 2d) was far less than that from the CO reaction (Fig. 2f). The carbon product from CH4 reaction developed into large particles with some small fiber-like structures formed after an extended reaction time of 126\u00a0h (Fig. 2e) that led to significant increase of the G-band Raman peak intensity. The fiber-like structures were not seen in the products from CO reaction under the same conditions.The mixed feed of CH4 and CO dramatically changed the morphology of the carbon product as compared to the CH4 and CO single gas reactions. As shown by the SEM images in Fig. 3\n, the reactions of CH4/CO mixtures resulted in formation of MWCNTs, and the content of MWCNT in the products appeared to depend on the y\n\nf, CH4, or CO content (y\n\nf, CO\n=1\u00a0\u2212\u00a0y\n\nf, CH4) in the feed. The Raman spectra in Fig. 3 revealed that the crystallinity of the graphitic carbon deposits was also affected by the feed composition. The addition of 5\u00a0mol% CO in the feed (i.e., y\n\nf, CH4\u00a0=\u00a095%) resulted in the formation of MWCNTs as the main carbon product with a small amount of graphitic carbon particles (GCPs). Increasing the y\n\nf, CO\n (i.e., decreasing y\n\nf, CH4) in the feed generally led to higher crystallinity for the carbon products as evidenced by the decreasing ratio of D/G peak intensities (I\n\nD\n\n/I\n\nG\n). The carbon products were predominantly of MWCNT structure for y\n\nf, CH4 varying from 10% to 90% but had significant amounts of GCP for y\n\nf, CH4 of 95% and 5%.The TEM images in Fig. 4\n unveiled distinct microstructures of the GCPs and MWCNTs in the products of the CH4/CO mixture reactions. In Fig. 4(a), the TEM image on the left shows the coexistence of long MWCNTs and GCP clusters. Compared to the activated FeCr (Fig. 2b), the XPS spectrum exhibited a new peak of metallic iron (Fe0) at around 708\u00a0eV in the Fe 2p3/2 envelope (Fig. 4a, right) [22]. Fig. 4(b) shows a HR-TEM image of the catalyst particle at a MWCNT tip, which suggested the formation of metallic Fe in the FeCr catalyst when reacting with the mixtures. The selected area ED pattern (Fig. S2a) was also characteristic of superimposed patterns of nanoscale ferrite, metallic Fe, and graphene [42\u201344]. In contrast, the TEM (Fig. 4c) and ED pattern of the catalyst/GCP cluster (Fig. S2b) reflected randomly gathered nanoscale ferrite structures [45]. The EDS results showed slightly higher Fe:Cr atomic ratio (~13.1) in the sample of MWCNT and GCP product (Fig. S2c) than the ratio of ~10.6 before reacting with the CH4/CO mixture. This might be caused by Cr6+/Cr3+ segregation from the ferrite solid solution during the formation of Fe0 cluster. The segregated chromium oxides could remain in the exposed oxides to be washed away by the acidic cleaning solution.The microstructures of the MWCNT in Fig. 4(b) and GCP in Fig. 4(c) remained unchanged in the entire range of y\n\nf, CH4 but the relative amounts of MWCNTs and GCPs apparently varied with y\n\nf, CH4 according to the SEM observations in Fig. 3. The MWCNTs were the principal component in products from feeds with y\n\nf, CH4 between 10% and 90% while GCPs were the main component in products from feeds with y\n\nf, CH4 of 95% and 5%. The TEM images in Fig. 4(b) from a representative area show that the diameters of the MWCNTs were in a range of 15\u201335\u00a0nm with unfirm wall thicknesses of 5.0\u00a0\u00b1\u00a01.0\u00a0nm. The MWCNT diameters were dependent on the sizes of the FeCr nanoparticles which ranged between 10 and 25\u00a0nm. The individual graphene layers were longer and more ordered in the MWCNTs (Fig. 4b) than in the GCP structures (Fig. 4c).The average \u03c7\n\nCH4 and \u03c7\n\nCO\n were determined for varied y\n\nf, CH4 over the first hour and second hour of reaction, respectively, and the results are presented in Fig. 5\n. The control and stabilization of the vary low CO and CH4 flowrates needed for y\n\nCH4 above 95% and below 5%, respectively, were challenging. Thus, mixture reactions were performed for y\n\nf, CH4 ranging from 5% to 95%. The experimentally measured conversions had large deviations, typically \u00b120%, mainly due to gas flowrate fluctuations caused by compaction, partial blockage, and movement of the catalyst bed under fast accumulation of carbons. Nevertheless, as shown in Fig. 5, the general trend of \u03c7\n\nCH4 changing with y\n\nf, CH4 could be observed with reasonable confidence while the dependence of \u03c7\n\nCO\n on y\n\nf, CH4 was less obvious.For CH4 and CO single gas reactions, the average \u03c7\n\nCH4 (~2.1%) was much smaller than the \u03c7\n\nCO\n (~25.0%) over the 2-h reaction time (Fig. 5a and b). This large difference of conversion can be explained by the fact that CH4 decomposition is endothermic with an equilibrium conversion (\u03c7\n\ne, CH4) of 16.8% while the Boudouard reaction of CO is highly exothermic with equilibrium conversion (\u03c7\n\ne, CO\n) of 92.1% at 500\u00a0\u00b0C and atmospheric pressure (Fig. S3). The reaction rate of the CH4 (r\n\nCH4) was also much lower than that of the CO (r\n\nCO\n) with the average r\n\nCH4 and r\n\nCO\n over the first hour of reactions estimated to be 0.83 and 15.4\u00a0mol/g-cat\u00b7h (Fig. 5c), respectively.For reactions of CH4/CO mixtures, besides the morphological change of carbon products from GCPs to MWCNTs, the \u03c7\n\nCH4 and r\n\nCH4 tended to increase substantially with y\n\nf, CH4 up to y\n\nf, CH4=95% (Fig. 5a). The addition of small amounts of CO also resulted in drastically greater \u03c7\n\nCH4 as compared to that of the CH4 in the inert N2 with the same y\n\nf, CH4. For example, the feed with y\n\nf, CH4 = 95% (i.e., with 5% CO) had an average \u03c7\n\nCH4 ~12%, which was more than five times the \u03c7\n\nCH4 for pure CH4 (\u03c7\n\nCH4 ~2.1%,). The \u03c7\n\nCH4 (<2.0%) was independent of the y\n\nf, CH4 for CH4 in N2 but was strongly dependent on y\n\nf, CH4 for CH4 in CO (Fig. 5a). The CO-promoted enhancement of \u03c7\n\nCH4 weakened as y\n\nf, CH4 decreased and eventually diminished at y\n\nf, CH4\u00a0<\u00a010% (i.e., y\n\nf, CO\n\u00a0>\u00a090%). The dependencies of \u03c7\n\nCH4 and \u03c7\n\nCO\n on the y\n\nf, CH4 suggest that r\n\nCH4 and r\n\nCO\n were significantly affected by the mixture feed composition. As can be seen from Fig. 5(c) that the average r\n\nCH4 and r\n\nCO\n over the first hour reactions were substantially enhanced by adding small amounts of CO and CH4, respectively. Such a kinetic enhancement for each component declined with increasing the content of the other. Also, for both CH4 and CO, the conversions in the first hour were generally higher than in the second hour of reaction that were expected by the decrease of catalyst surface accessibility due to the growing carbon deposits on the catalyst surface.The above observed effects of CH4/CO feed composition on the gas conversions (\u03c7\n\ni\n), reaction rates, and produced carbon morphology indicate strong synergistic effects between the CH4 and CO reactions on the FeCr catalyst. These synergistic effects may include thermal and chemical states and interactions of produced gases at the catalyst surface. Firstly, the highly exothermic reaction of CO could increase the catalyst surface temperature to enhance the endothermic CDM both thermodynamically and kinetically. Meanwhile, the CO-facilitated CH4-decomposition could thermodynamically benefit CO conversion by timely consuming the heat generated at the catalytic surface. These mutual benefits of reaction heat effects likely played significant roles in enhancing the \u03c7\n\nCH4 and r\n\nCH4. Secondly, the coexistence of CH4 and CO with the resultant H2 may help to further reduce the ferrite surface (Fig. 4b) that consequently altered the chemisorbed intermediates and the morphology of the carbon deposits.To study the mutual influences between CH4 and CO reactions on the FrCr catalyst surface, the activated FeCr catalyst was pretreated for 30\u00a0min in pure CH4 or CO before reacting with the other. The CH4-pretreated catalyst produced predominantly MWCNTs with a few GCPs when subsequently reacting with the pure CO (Fig. 6a and b). However, the CO-pretreated catalyst generated primarily GCPs with no appreciable MWCNTs in the subsequent reaction with pure CH4 (Fig. 6c and d).The \u03c7\n\nCH4 on the CO-pretreated catalyst was <0.5% (Fig. 6d), which was drastically lower than that on the fresh catalysts shown in Figs. 6(b) and 5(a). The \u03c7\n\nCO\n on the CH4-pretreated catalyst presented in Fig. 6(b) was also significantly lower than those on the fresh catalyst for pure CO (Figs. 6d and 5b) or mixtures of low y\n\nf, CH4 but comparable to \u03c7\n\nCO\n for mixtures of high y\n\nf, CH4 (i.e., low y\n\nf, CO\n) (Fig. 5b). These decreases of \u03c7\n\nCH4 and \u03c7\n\nCO\n can be attributed to the reduced catalytic site accessibility by carbon deposits from the CH4 and CO pretreatments and the lack of beneficial heat effects existing in mixture reactions. The CO-pretreated catalyst suffered more severe activity loss for the subsequent CH4 reaction apparently because of the thicker carbon deposits from the fast CO reaction.The chemical and morphological properties of the pretreated catalyst surface play key roles in determining the structures of the carbon products during the subsequent reactions with different gases. The SEM and EDS examinations revealed similar textures for the catalyst particles after the 30-min pretreatment in CH4 (Fig. 7a) and CO (Fig. 7b), respectively. However, The EDS results in Fig. 7(a) and (b) show that the CO-pretreated catalyst had a substantially greater amount of carbon deposits (C/Fe\u00a0~\u00a01.8) than the CH4-pretreated catalyst (C/Fe\u00a0~\u00a00.8). This difference was expected because the endothermic CH4 reaction (eq. 1) was severely limited both kinetically and thermodynamically at 500\u00a0\u00b0C (Fig. 5). The thicker carbon deposits could cause greater blockage of the catalytic sites that led to more drastic reduction of \u03c7\n\nCH4 on the CO-pretreated catalyst as compared to the decrease of \u03c7\n\nCO\n on the CH4-pretreated catalyst.The XPS spectra for catalysts after the pretreatments by CO and CH4 are presented in Fig. 7(c-f). The Fe 2p spectra of the catalysts in Fig. 7(c) show two distinct envelopes at 710.2\u2013711.0\u00a0eV and 723.8\u2013724.3\u00a0eV, which are assigned to the Fe 2p3/2 and Fe 2p1/2 sub-levels, respectively. The deconvoluted Fe 2p XPS spectra exhibited two main peaks and a satellite peak for each sub-level, which are related to the contributions from both Fe2+ and Fe3+ ions. The peaks at 709.9\u2013710.6\u00a0eV and 723.3\u2013723.9\u00a0eV are assigned to Fe2+ states, while the peaks at 712.5\u2013712.8\u00a0eV and 725.5\u2013726.8\u00a0eV are assigned to Fe3+ ions. Furthermore, the satellite peaks at 718.2\u2013719.2\u00a0eV and 732.0\u2013732.7\u00a0eV are attributed to both Fe2+ and Fe3+ ions. The spectrum suggests that Fe2+ and Fe3+ ions coexisted on the surface of the Fe/Cr catalyst after the pretreatments in CO and CH4. The CO-pretreated sample showed a blue shift in the binding energy of Fe 2p as compared to the CH4-pretreated sample. Such a shift may be attributed to the higher population of Fe3+ ion on the surface of CO-pretreated sample, which is confirmed by the higher Fe3+/Fe2+ ratio found on the CO-pretreated catalyst as compared to the CH4-pretreated sample (Table 1\n). This confirms that CH4 is a stronger reductant than CO under the current reaction conditions. The XPS spectrum-based estimate of chemical composition was achieved by the area integration protocol described in our previous work [46]. The Fe 2p spectra are accompanied by the broadness of satellite peaks that may have caused the relatively large uncertainty in chemical quantification [47].In the FeCr solid oxide solutions of low Cr contents, the Fe3+ cations are substituted by the doped Cr ions in the ferrite lattice, which is evidenced by the XRD results in Fig. 2(c). The Cr dopant is known to promote the redox cycles of Fe\n3+\u00a0\u2194\u00a0Fe\n2+ that in turn enhances the catalytic activity of ferrite [46,48]. The Cr 2p XPS spectra of the samples in Fig. 7(d) show the Cr 2p3/2 and Cr 2p1/2 sub-bands. Each sub-level can be divided into two peaks in which the peaks at ~576\u00a0eV and\u00a0~\u00a0586\u00a0eV are ascribed to the Cr3+ states and the peaks at ~578\u00a0eV and\u00a0~\u00a0587\u00a0eV correspond to the Cr6+ states. The Fe 2p binding energy of CO-pretreated sample was higher than that of the CH4-pretreated sample, which could be caused by the presence of more Cr6+ ion on the surface of CO-pretreated sample. This observation is consistent with the fact that surface Cr6+/Cr3+ ratio of the CO-pretreated catalyst was greater than that of the CH4-pretreated catalyst (Table 1). Like the larger Fe3+/Fe2+ ratio in the CO-pretreated sample, the greater Cr6+/Cr3+ ratio of the CO-pretreated catalyst is also attributed to the relatively weaker reducing power of CO than CH4 under the specific conditions.The single-phase hematite structure of the preactivated fresh FeCr catalyst (Fig. 2c) suggests that the following isomorphous substitution for Fe3+ by Cr6+ occurs to create Fe3+ vacancies (V\n\nFe(III)\u2034),\n\n(7)\n\nCr\n\nO\n3\n\n\n\u2192\n\n\nFe\n2\n\n\nO\n3\n\n\n\n3\n\nO\nO\nX\n\n+\nCr\n\n\nVI\n\n\nFe\n\nIII\n\n\n\n\u2022\n\u2022\n\u2022\n\n\n+\n\nV\n\nFe\n\nIII\n\n\n\u2034\n\n,\n\u2206\n\nG\nT\n\no\n,\n\u2217\n\n\n\n\n\nThe XPS findings of the Fe and Cr oxidation states after activation in the processing gas confirm the transition of hematite to magnetite by partial Fe\n\n3+\n\u00a0\u2192\u00a0Fe\n\n2+ reduction, e.g., \n\n1\n2\n\n\nFe\n2\n\n\nO\n3\n\n+\n\n1\n2\n\n\nH\n2\n\n\u2192\nFeO\n+\n\n1\n2\n\n\nH\n2\n\nO\n,\n\u2206\n\nG\nT\n\no\n,\n\u2217\n\n\n, that may be facilitated by the extinction of V\n\nFe(III)\u2034 via reduction of Cr6+ (Cr(VI)) to Cr3+ (Cr(III)),\n\n(8)\n\n\n3\n2\n\n\nO\nO\nX\n\n+\nCr\n\n\nVI\n\n\nFe\n\nIII\n\n\n\n\u2022\n\u2022\n\u2022\n\n\n+\n\nV\n\nFe\n\nIII\n\n\n\u2034\n\n+\n\n3\n2\n\n\nH\n2\n\n\u2192\nCr\n\n\nIII\n\n\nFe\n\nIII\n\n\nX\n\n+\n\n3\n2\n\n\nH\n2\n\nO\n,\n\u2206\n\nG\nT\n\no\n,\n+\n\n\n\n\n\nAlthough the CO and CH4 can also reduce the above oxides at sufficiently high temperatures, H2 is expected to be the most effective reductant at the low temperature. Thus, the amount of H2 and the content of Cr dopant are critical to the extent of FeCr catalyst reduction and the following equilibrium concentrations of FeO ([Fe2+]) and Cr(III)\nFe(III)\n\nX\n ([Cr3+]) are defined by thermodynamics principles,\n\n(9)\n\n\n\nFe\n\n2\n+\n\n\n\n=\n\n\n\n\n\n\nFe\n\n3\n+\n\n\n\n\u2219\n\np\n\nH\n2\n\n\n\n\np\n\nH\n2\nO\n\n\n\n\n\n1\n/\n2\n\n\n\u2219\nexp\n\n\n\n\u2212\n\u2206\n\nG\nT\n\no\n,\n\u2217\n\n\n\nRT\n\n\n\n\n\n\n\n(10)\n\n\n\nCr\n\n3\n+\n\n\n\n=\n\n\n\n\n\nCr\n\n\nVI\n\n\nFe\n\nIII\n\n\n\n\u2022\n\u2022\n\u2022\n\n\n\n\n2\n\n\u2219\n\n\n\nO\nO\nX\n\n\n\n3\n/\n2\n\n\n\u2219\n\np\n\nH\n2\n\n\n3\n/\n2\n\n\n\n\np\n\nH\n2\nO\n\n\n3\n/\n2\n\n\n\n\u2219\nexp\n\n\n\n\u2212\n\u2206\n\nG\nT\n\no\n,\n+\n\n\n\nRT\n\n\n\n\n\nThe presence of water vapor on surface may limit the reductions of metal ions and prevent the formation of metallic Fe0, which was confirmed in our previous work on MWCNT formation in H2-lean water gas shift reaction environments containing CH4 and CO [22].For the current FeCr catalyst, metallic Fe was not observed after being pretreated with the pure CH4 and CO but was found after reacting with the CH4/CO mixture (Fig. 4a and b). This may be caused by the significant H2 generated from the CO-promoted CDM, which could induce further reduction of Fe2+ to Fe0 in reaction of the CH4/CO mixtures. The formation of the metallic Fe could also facilitate the growth of MWCNTs with higher crystallinity and uniformity.The gas environment-dependent oxidation state variations of Fe and Cr on the FeCr catalysts dictate the types of surface intermediates, such as chemisorbed\u2011oxygen and carbonaceous species, that consequently affect the \u03c7\n\nCH4 and \u03c7\n\nCO\n and carbon morphology. The deconvoluted C 1\u00a0s spectra in Fig. 7(e) show three peaks at ~284\u00a0eV, ~286\u00a0eV, and\u00a0~\u00a0288\u00a0eV, which are related to the C\u2013C/C\u2013H, C\u2013O/C=O, and carboxylic groups, respectively. The relative quantity of C\u2013O/C=O and carboxylic species was lower on the surface of CH4-pretreated catalyst than on the CO-pretreated catalyst (Table 1).The O 1\u00a0s spectra of the samples in Fig. 7(f) can be fitted into two peaks in which the lower and higher binding energy bands are assigned to the lattice oxygen species (OI) and chemisorbed oxygen species (OII), respectively. Moreover, the CH4-pretreated sample exhibited higher OI/(OI\u00a0+\u00a0OII) ratio as compared to the CO-pretreated sample (Table 1). These indicate a decrease of chemisorbed oxygen species on the CH4-pretreated catalyst surface. This is consistent with the fact that the CH4-pretreated catalyst possessed less C\u2013O/C=O and carboxylic species but had greater amounts of CC and CH species.The variations of metal oxidation states and the types of chemisorbed oxygen and carbonaceous intermediates on the catalyst surface determine the gas conversions and carbon morphology in reactions with CH4 and CO. Although carbon coated metallic catalysts, such as Fe and Ni nanoparticles, were found to actively catalyze the MWCNT formation because metal-carbide can form at relatively low temperatures, carbon deposits from CH4 and CO were found to behave differently [49,50]. Literature studies have found that, during CO reaction on the Ni/SiO2 catalyst, some chemisorbed oxygen atoms (OII) could penetrate through the surface to restructure the metal surface that enables subsurface migration of carbon [49]. It was also found that CO carbonization was not determined by the dissociation of CO but by the reaction of adsorbed oxygen with CO via a Langmuir-Hinshelwood mechanism. However, the carbon deposits from CDM were found to remain on the surface due to the inadequate surface OII that hindered the carbon transport and growth.The pretreatment by CH4, a stronger reductant than CO, could cause Fe3+ reduction to greater extents, even possible generation of surface Fe0 atomic clusters [51]. Thus, although the CH4-pretreated catalyst lacked surface OII to sustain fast CH4 decomposition, it could catalyze the subsequent reaction with CO and promote the formation of MWCNT (Fig. 6a) because CO could create a larger OII population (Table 1). On the contrary, the CO-pretreatment was unable to induce metallic Fe at the low pressure and temperature. Consequently, the CO-pretreated catalyst was unable to support the subsurface carbon transport for dispense of MWCNT when continuing to reaction with CO and inefficient in promoting the reaction with pure CH4 due to severe surface carbon blockage (Fig. 6d).In the mixture reaction, CO reaction increases the OII and catalyst surface temperature that enhance the \u03c7\n\nCH4 to generate more H2 (Fig. 5a). The increased H2 presence on the catalyst surface may further reduce the surface Fe2+, i.e., FeO\u00a0+\u00a0H\n2\u00a0\u2192\u00a0Fe\n0\u22efH\n2\nO, to form metal Fe clusters on surface (Fig. 4a and b). The O 1\u00a0s spectra for the sample reacted in CH4/CO mixtures (Fig. 4a) exhibit a peak at 533.6\u2013533.8\u00a0eV, which was absent in the spectra for samples from reactions with pure CH4 and CO. This peak can be ascribed to oxygen in surface groups (OIII) such as chemisorbed carboxyl and/or hydroxyl species.The surface hydroxyl groups (OH\n\u2212)\nS\n may be resulted from reaction with H2, which spontaneously dissociates on Fe clusters to readily react with O and C [31], e.g., 2O\n\nO\n\n\nX\n(O\n\nI\n)\u00a0+\u00a02H\u00a0\u2194\u00a0V\n\nO\n\n\u2022\u2022\u00a0+\u00a02(OH\n\u2212)\nS\n when reaction occur at the metal/oxide boundaries. The resultant V\n\nO\n\n\u2022\u2022 could facilitate the OI/OII transitions that promotes the formation of surface intermediates such as \u201cM\u22efO\n\ns\n\u00a0\u2212\u00a0R\u201d where M\u00a0=\u00a0Fe\n2+/0 and R\u00a0=\u00a0CO\n\nm (0\u2264m\u22641), CH\n\nx (0\u2264x\u22643), or H, depending on the gas composition and temperature. The ab initio and density function theory computational studies have shown lowered reaction barriers for both CO and CH4 dissociations on Fe nanostructures with exothermic path to surface bound elemental carbon (Cs). Such carbon atom formation on nanostructured Fe is key to the initial CNT formation that involves the well-known cap lift-off process in early stage [6,11,31]. The evidence of cap formation was also observed on some premature MWCNT/catalyst particles from mixture reaction but not in the samples of pure CO reaction, where only thick graphitic layers and rods were found (Fig. S4).The magnetite phase nanocrystalline FeCr (10:1) solid oxide solution could effectively catalyze CDM at 500\u00a0\u00b0C, which is much lower than the temperatures needed for CDM on the conventional metal nanoparticle catalysts. The \u03c7\n\nCH4 increase dramatically and the morphology of carbon products altered when a small amount of CO was cofed. For example, a feed with 5% CO (i.e., y\n\nf, CH4 = 95%) achieved an average \u03c7\n\nCH4 that was over five times the conversion of pure CH4 and obtained predominantly MWCNT product instead of GCPs formed from pure CH4 or CO. The causes of the CO-promoted catalytic activity enhancement for CDM have been found to lie in the synergistic effects of CH4 and CO reactions on the chemical states of the catalyst surface and mutually beneficial reaction heat effects. The strongly exothermic Boudouard reaction increases the local temperature at the catalytic sites to thermodynamically and kinetically benefit the endothermic CH4 decomposition. The enhanced CDM generates more hydrogen at the catalyst surface that induces further reduction of Fe2+ to form Fe0 metal clusters in the catalyst surface. These metallic Fe clusters along with the CO-originated surface oxygens (OII) facilitate the transfer of C and H out of the active surface sites to sustain fast conversions of both CH4 and CO. It can be also inferred by the microscopic observations and XPS results that the surface metallic Fe clusters are imperative to MWCNT formation by enabling subsurface C transfer for continued MWCNT growth. The doped Cr participate in the gas-induced solid state defect reactions that facilitate the redox cycles of Fe\n3+/Fe\n2+ and transfer of OII, which are vital to maintaining the surface catalytic activities. With its well-known sulfur resistance and low-cost production, nanocrystalline FeCr catalysts is potentially useful for conversion of biogas to produce clean H2 and MWCNTs under desirably mild conditions.\nCRediT authorship contribution statement:\n\nXinhui Sun: Investigation, Data Curation, Writing (draft). Devaiah Damma: Characterizations, Data Curation. Zishu Cao: Characterizations. NOE T. Alvarez: Validation, Supervision. Vesselin Shanov: Resources (Lab), Supervision. Antonios Arvanitis: Resources (Reactor), validation. Panagiotis G. Smirniotis: Resources (Catalysis Lab), Supervision. Junhang Dong: Conceptualization, Funding acquisition, Resources, Writing \u2013 drafting, reviewing, and editing.\nDeclaration of Competing Interest.\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.None.This research was financially supported by the Development Service Agency of Ohio through the Ohio Coal Research and Development program (Grant no. OOECDO-D-17-13) and the U.S. Department of Energy/Office of Science (Grant no. DE-SC0018853). The alumina tubes were provided by Media and Processes Technology Inc., Pittsburgh, USA.\n\n\n\nSupplemmentary material\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106475.", "descript": "\n                  Catalytic decomposition of methane (CDM) to H2 and multiwalled carbon nanotubes (MWCNTs) was achieved by a nanocrystalline Cr-doped ferrite (FeCr) catalyst at 500\u00a0\u00b0C and atmospheric pressure with minor cofed CO. The exothermic Boudouard reaction increased the temperature and H2 from CDM at catalyst surface that induced Fe2+ reduction to Fe0. The Fe0 clusters along with the CO-originated surface oxygens enabled transfer of C and H to sustain the surface CDM and CO reactions. The metallic Fe-enabled C transfer led to the formation of MWCNTs. The Cr6+/3+ dopants facilitated the Fe redox cycles and maintained surface oxygens for high catalytic activity.\n               "}
{"full_text": "Electrocatalytic CO2 reduction reaction (CO2RR) is a promising way to convert CO2 into valuable carbon-containing fuels or chemicals at room temperature and atmospheric pressure [1\u20136]. The development of high-efficiency catalysts for CO2 reduction reaction has become a topical issue. According to the previous studies, metal-based catalysts (Pb, Pt, Au, Ag, Cu, etc.) exhibited good catalytic activity for CO2RR [7\u201311]. Among these metals, Cu is the only catalyst that has the ability to realize C\u2013C coupling reaction to generate C2+ products, due to its unique adsorption characteristics of these intermediates [12]. Therefore, Cu-based catalysts become the star materials in electrocatalytic CO2 reduction field.CO2 is an extremely stable linear molecule with two CO bonds length of 116.3 pm and bond energy of 750\u00a0kJ\u00a0mol\u22121, which determine that CO2 activation is a key step limiting the catalytic activity of Cu-based catalyst [13,14]. In addition, Cu has the moderate adsorption properties of intermediates, resulting in a complex variety of hydrocarbon products. So the low CO2RR activity and C2+ products selectivity of Cu-based catalysts is still worth investigating [15\u201317].In order to improve the activity and selectivity of Cu-based catalysts [18], various strategies were proposed such as regulating morphology structure [19,20], defect [21], and surface/interface engineering [22]. As we know, the CO2RR require multiple steps and involve various intermediates. Generally, the binding strengths of involved intermediates are linearly correlated. For example, both \u2217COOH and \u2217CO rely on C\u2013metal binding, thus the adsorption strengths of \u2217COOH and \u2217CO on the metal catalyst surface are essentially correlated. This relationship determines that the binding energy of individual intermediate at Cu site cannot be regulated independently. If we want to improve the activation process of CO2, we need to adjust the stronger adsorption capacity for \u2217COOH, which will incidentally bring stronger \u2217CO adsorption capacity, thus affecting products selectivity. Cu-based bimetallic catalysts offer the possibility to break this scaling relationship [23,24]. Two kinds of metal component can provide more possible adsorption sites for intermediates, and can more fully realize the regulation of the binding energy of intermediates at different sites, thus improving the activity of CO2RR and optimize the selectivity of products.Since Watanabe et\u00a0al. reported the Cu-based bimetallic catalysts in 1991, a variety of bimetallic Cu-based catalysts have been synthesized and used for CO2 reduction, such as Cu\u2013Au [25], Cu\u2013Ag [26], Cu\u2013Pb [27], and Cu\u2013Pt [28]. It has been fully demonstrated that Cu-based metallic catalyst can significantly promote CO2 activation. For example, due to the stronger CO2 adsorption ability of Pd than Cu, the reaction kinetics of \u2217COOH generation on the Pd surface is faster than on the Cu surface. Therefore, combining Pd with Cu is an effective approach to promote CO2 reduction. In addition, density functional theory (DFT) calculations show that the stepped CuPd(111) surface has a stronger CO2 adsorption and activation capacity than the Cu(111) surface [29]. Experimental results also demonstrated the Cu\u2013Pd alloy can effectively promote the activation of CO2 and inhibit the formation of hydrocarbons on the catalyst surface, resulting in much higher selectivity of CO product [30]. In addition, Cu-based bimetallic catalysts can also promote the C\u2013C coupling reactions. For example, Hoang et\u00a0al. synthesized a nanoporous Cu\u2013Ag alloy by additive-controlled electrodeposition [31]. The catalyst exhibited a record CO2RR performance with the Faradic efficiency (FE) for C2H4 and C2H5OH reaching 60% and 25%, respectively. The nanoporous surface of Cu\u2013Ag caused low-coordination metal atoms steps and edges, which are favored to facilitate the C\u2013C coupling process. Therefore, Cu-based bimetallic catalysts show an attractive perspective in CO2RR.Although the study about Cu-based bimetallic catalysts is likely to be a future priority, what is the key factor that affects CO2 reduction performance on Cu-based bimetallic catalysts? What is the role of the different metal elements? How can we design Cu-based bimetallic catalysts more rationally? These problems still greatly puzzled researchers. In this review, we begin with a brief background on the CO2RR process and the basic principles that determine the selectivity of Cu-based bimetallic catalysts. Then, we summarize the research progress of Cu-based bimetallic catalysts in CO2 reduction reaction (Scheme 1\n). In short, we first discuss the advantages of the morphology of Cu-based bimetallic catalysts, including dendritic, nanowires, polyhedron, and core\u2013shell structure. Subsequently, we focus on the local electric field effect induced by the Cu-based bimetallic nanoneedle structure. Then, the interface engineering between bimetallic and the series of special phenomena that occurred on the bimetallic interface (including interfacial effects, interfacial atomic arrangements, interfaces strain) are discussed. Next, some commonly studied effects in bimetallic systems, such as the electronic effect and tandem effect are also analyzed. For a deeper understanding, some critical examples that combine experimental and computational studies in the CO2 reduction reaction are given. Finally, a perspective on the research and design of Cu-based bimetallic catalysts is proposed, wishing to better understand the CO2 reduction process on Cu-based bimetallic catalysts and to provide some insights for future studies.With the continuous development of instrument science and the electrocatalytic CO2 reduction reaction, great progress has been made in the CO2 reduction mechanism [32\u201336]. In general, CO2 reduction reaction involve the multi proton/electron transfer process. The first step is the adsorption/activation of CO2 molecule. Due to the extremely stable property of CO2, the adsorption/activation ability of active sites on CO2 has a great influence on CO2 reduction activity [13]. In most cases, the high energy barrier of CO2 adsorption/activation makes it a rate-determining step (RDS) during CO2 reduction process [37].After CO2 adsorption and activation, various intermediates (such as \u2217CO2, \u2217COOH, \u2217CO, and \u2217CH2) will form with the transfer of protons (H+) and electrons. Those intermediates play a critical role in the process to form hydrocarbon products. For instance, \u2217CO2 can obtain an H+ to form two possible intermediates: \u2217OCHO or \u2217COOH, which is determined by the adsorption configuration of CO2. The \u2217COOH, as the main intermediate for the generation of CO product, but it is also an important intermediate to form HCOOH (depending on the binding energy of C and O at the active sites) (Fig.\u00a01\n). As a result, there is often a competitive relationship between different products. One critical intermediate to form C1 products is \u2217CO. For example, the \u2217CHO and \u2217COH can form by hydrogenation process of \u2217CO, then the \u2217CHO can undergo 3 pairs of proton/electron transfer process to form CH3OH, while the \u2217COH can form CH4 by 5 pairs of proton/electron transfer process (Fig.\u00a01).Besides, \u2217CO is also an important intermediate to generate C2 products. Distinguished from the C1 pathway, the formation of C2 products needs the C\u2013C coupling process. At present, there are three main pathways of C\u2013C coupling: (1) Two \u2217CO direct coupling to form \u2217COCO; (2) \u2217CO coupling with \u2217COH to form \u2217COCOH; (3) \u2217CO coupling with \u2217CHO to form \u2217COCHO. Moreover, there is also the possibility of coupling between two \u2217COH intermediates or two \u2217CHO intermediates to form \u2217COHCOH or \u2217CHOCHO, respectively. And another possible pathway was reported by Janik and his colleagues [38]. The \u2217CO intermediate hydrogenation to form \u2217COH or \u2217CHO, subsequently the \u2217CHO or \u2217COH undergo hydrogenation and dehydration process to form \u2217CH2 intermediate, then two \u2217CH2 direct coupling can generate C2H4 product. All in all, the C\u2013C coupling process is a critical and complex step of CO2 reduction. The coverage of \u2217CO has a great impact on the C\u2013C coupling process. Huang et\u00a0al. investigated the influence of \u2217CO coverage on the Cu(100) facet and found that with the increase of \u2217CO coverage, the energy barrier of \u2217CO\u2013\u2217CO coupling decreasing, which can be attributed to the weakening of the Cu\u2013\u2217CO bonds [39]. In summary, how to design catalysts for promoting CO2 adsorption/activation and C\u2013C coupling, and how to tune the adsorption/desorption properties of critical intermediates are research hotspots.For the rational design of Cu-based bimetallic catalysts, the role of the second metal should be considered. One important point is that the second metal acts as the direct reaction active site. The Cu-based bimetallic catalysts require the rational utilization of the inherent properties of both metal components, and the incorporation of these properties into the bimetallic systems. For example, in the pathway of HCOOH generation, the O atom in the \u2217OCHO intermediate rather than the C atom will adsorb on the surface catalyst active sites. For higher selectivity of HCOOH product, the stronger O affinity of metal sites is required. Previous works have demonstrated that metal electrocatalysts, which has high O affinity and low H (such as Sn, Bi, In, and Pb) affinity exhibit excellent selectivity for HCOOH.Due to the difference in intrinsic catalytic properties of catalytic sites, the main product on the different metal sites will be various. Therefore, the products distribution of the bimetallic catalysts can be adjusted. In bimetallic systems, combining Cu with another metal that produces different intermediates to activate the tandem effect is an effective strategy to improve the C2+ products catalytic performance. Most commonly, combining Cu with Au, Ag, Zn, and other \u2217CO-producing metals results in an enrichment of \u2217CO on the catalyst surface, the \u2217CO will be spillover to the adjacent Cu sites and be further coupled to form C2+ products.Moreover, the introduction of secondary metals may not act as active sites, it can assist in changing the electronic structure of Cu, such as the coordination environment and electronic orbitals of Cu. In this case, the d-band electrons of Cu can pair with the s-band or p-band electrons of the adsorbate, thus facilitating the adsorption of reactants and the formation of intermediates. The introduction of secondary metals may also only change the geometry structure of the bimetallic catalysts, which is manifested by the change of morphology, the creation of interfaces, and the adjustment of atomic arrangements. For example, in porous bimetallic structures, \u2217CO intermediate desorption is suppressed due to the domain-limiting effect, which leads to a locally high concentration environment of \u2217CO on the catalyst, thus enhancing the selectivity and activity of the C2 products.It is generally believed that the high coverage of adsorbed C1 intermediates on the Cu surface is the key to triggering the interaction of adjacent intermediates for further C\u2013C coupling, the method of assembling secondary metals with different C1 intermediate production capacities in Cu is the key to the goal. This requires full consideration of how the CO2 reduction process fits and cooperates with the two components to maximize the reduction efficiency of CO2 on the catalyst surface. In addition, the interaction between the second metal and the host metal (Cu) should also be considered. We consider and summarize these principles in detail, and present them in the content that follows.Research on Cu-based bimetallic alloys dates back to 1972, at that time, two different papers but discussed the same topic\u2014the change of properties caused by alloying of Ni and Cu were published in the same journal. The one discussed this change from the point of ensemble size, however, the other proposed a view of d-character [40\u201342]. Although the two papers have different views, they all made convincing explanations and ended up with the same conclusion\u2014alloying the two metals increases the activity of the catalytic reaction [40]. Compared with the single metal catalyst, the performance improvement of bimetallic catalysts is mainly attributed to the multiple factors among structure, interface, and various synergistic effects resulting from the coupling of two metals. For the rational design of Cu-based bimetallic catalysts and understand the promotion mechanism between Cu-based bimetallic catalysts on CO2 reduction. We have sorted out the experimental progress and related derivative studies on those factors and discussed the influence factors in detail in the following parts.Morphology can change the local microstructure of nanocatalysts and determine their catalytic performance [43]. The synthesis of Cu-based bimetallic nanocrystals with different morphology, such as nanodendrite, polyhedral, porous hollow, and core\u2013shell structure, has been widely studied [44]. Each of these structures has its unique advantages (Fig.\u00a02\n). In this chapter, we have briefly classified and discussed the morphology effect of Cu-based bimetallic catalysts for CO2 reduction reaction.Dendrite structures exhibit the advantages of high roughness and high surface area, which is beneficial for CO2 adsorption and electronic transfer. Keerthiga et\u00a0al. proposed that the dendritic structure can alter the crystal orientation and crystallinity of the metal deposits, thus facilitating the catalytic reaction [45]. Koh et\u00a0al. found the dendrite structure has high index facets and low-coordination sites, which are beneficial to the stabilization of reaction intermediates [46]. Based on these studies, the high-efficiency CO2 reduction reaction with the multi-electron/proton transfer process can readily occur on the dendritic structure.Roth et\u00a0al. investigated the effect of Cu\u2013Ag dendrite structure on the distribution and selectivity of CO2 reduction products [47]. Compared with Cu foams, the presence of Ag during electrodeposition significantly changed the size and shape of dendrites in the pore walls of Cu\u2013Ag foams, resulting in increased surface area and roughness. The high surface area enables Cu\u2013Ag dendrite catalysts to have higher CO2 reduction current density, which represents higher catalytic activity (Fig.\u00a03\n). In addition, the study indicated that the production of hydrocarbons depends on the asymmetry of the structure, and the rough surface is conducive to the production of higher hydrocarbons. Thus, dendrite structure was considered to be more favor the production of hydrocarbon. Ingole et\u00a0al. demonstrated the dendritic structures are more favor for ethylene than other structures. This enhancement is mainly attribute to the nano-hot spots, diffusion control of ionic species and pore size which more favorable for CO adsorption [48]. Similarly, Klingan et\u00a0al. demonstrated that needle-like structures at the dendrites edges can lead to an increase in local pH, promoting C2H4 production while suppressing CH4 yield [49]. Reller et\u00a0al. further demonstrated this enhancement is indeed related to the needle-like structure, because the selectivity to ethylene decreases significantly with the coarsening of the edge needle-like structure [50].In addition, due to the higher roughness, and higher density of stepped sites, the dendritic structure can promote CO2 reduction performance by suppressing the hydrogen evolution reaction (HER) [51]. For example, Hoffman et\u00a0al. prepared Cu\u2013In bimetallic catalysts with a dendrite structure [52]. Due to the exposed stepped sites and variety of surface facets, the dendrite structure exhibited an optimized performance of formate (FE\u00a0=\u00a062%) at \u22121\u00a0V by inhibition of HER.One-dimensional materials such as bimetallic nanowires [53], nanorods [54], and nanotubes are considered to be one of the most common materials for electrocatalysis as its multiple advantages [55]. Firstly, one-dimensional structures have better conductivity because the charges can efficiently transfer along with the axial orientation. Secondly, the uniform structure can promote the adsorption of reactants, and improve the stability of the catalyst and directional mass transfer. Jang et\u00a0al. prepared a one-dimensional CuIn nanowire electrocatalyst by anodization and electrodeposition (Fig.\u00a04\na and b) [53]. The CuIn catalyst exhibited performance with over 68% selectivity of CO (Fig.\u00a04c). They proposed that the CuIn nanowires provide a larger surface area and further promote the charge transfer. And alloy formation changes the properties of local active sites. The In atoms replace the Cu atoms and suppress the H2 generation. In addition, due to the oxygen-deficient indium and oxygen of indium hydroxide have slightly positive and negative charges. By delivering electrons and protons, CO2 would be adsorbed and stabilized on the polarized surface as depicted in Fig.\u00a04d.In recent years, one-dimensional nanowire arrays have received extensive attention. The abundant space between adjacent nanowire structures, not only facilitates the mass transport of the reactants at the solution/catalyst boundary, but also promotes the timely release of the bubbles generated by the reaction in the solution, avoiding the inactivation of active sites. For example, He et\u00a0al. prepared a Cu\nx\nAu\ny\n nanowire arrays (NWAs) for efficient catalysis CO2 reduction. The authors proposed that the nanowire array structure significantly limits the diffusion of OH\u2212 and HCO3\n\u2212, which results in high local pH thus facilitating the coupling of CO intermediates. In addition, the nanowire arrays structure can also decrease the diffusion rate of CO generated from the nanowire surface to the bulk solution. Results in a higher CO concentration in the nanowire structure, and increase in the \u2217CO/\u2217H ratio for further reduction on the nanowire, it is beneficial to generate EtOH at low overpotentials [56].Polyhedron of different shapes have also been widely explored because their expose different crystal planes and more active sites. Yang et\u00a0al. prepared bimetallic Cu\u2013Pd catalysts with different morphologies and compositions, which exhibited different CO selectivity (Fig.\u00a05\na\u2013c) [57]. In particular, spherical Cu\u2013Pd nanoalloys (Cu\u2013Pd\u2013S) have the highest CO conversion faradaic efficiency (93%), while the H2 evolution reaction is dominant on the concave cubic (Cu\u2013Pd\u2013C) and dendritic Cu\u2013Pd (Cu\u2013Pd-D) nanoalloys (Fig.\u00a05d and 5e). This morphology-dependent selectivity is attributed to the equilibrium of rate-determining steps in the CO2 reduction reaction. The severe HER rates of the concave cubic and dendritic Cu\u2013Pd nanoalloys can be mainly attributed to their unique structures providing more active edge/corner sites for water dissociation.Due to the suitable distance of the bimetallic domains in the polyhedral structure, the transformation of intermediate can be effectively facilitated and the reduction performance of CO2 can greatly improve. Recently, Ma et\u00a0al. realized the confined growth of Cu with (100) facets on Ag nanocubes to form Ag\u2013Cu Janus nanostructure [58]. Compared with Cu nanocubes, the bimetallic Janus structure exhibits very good C2 products selectivity. The exposed Cu(100) facets can lower the energy barrier for C\u2013C coupling and improve the selectivity for C2H4. In addition, the Cu\u2013Ag polyhedral structure also showed higher FE for C2H4 and C2 products than the Ag\u00a0+\u00a0Cu mixture. Due to the long distance between the Ag and Cu domains in the Ag\u00a0+\u00a0Cu mixture, it is difficult for the CO spillover from Ag to Cu. In the Ag\u2013Cu polyhedral structure, this suitable distance between Cu and Ag is beneficial to the diffusion of the intermediate product \u2217CO and reduces the energy barrier of C\u2013C coupling, resulting in the high selectivity of C2 products in Ag\u2013Cu polyhedral.The core\u2013shell structure has the advantages of a larger specific surface area, short diffusion paths, prominent hierarchical structure, and fast charge transfer rate. Therefore, the core\u2013shell structure is attractive as electrocatalysts, as the core is the main active component with specific functions, while the shell acts as a protective layer to enhance the performance of the core material or generate new properties. Huang et\u00a0al. prepared a Cu@Ag core\u2013shell nanoparticles with tunable shell thickness. The Cu@Ag-2 catalyst with the optimal Ag shell thickness exhibits the highest C2 FE (67.6%). The analysis results show that the Cu@Ag-2 catalyst with appropriate Ag shell thickness can effectively generate and enhance the local CO concentration on the Cu core, and the unique core-shell structure provides more active sites and faster charge transfer. Thus promoting the conversion of CO2 to C2 [59].The core-shell structure can regulate the concentration of intermediates through a confinement effect and thus facilitating C\u2013C coupling. Zhong et\u00a0al. prepared Ag@Cu core-shell catalysts, achieved the tuning of the pore size in the porous Cu shell (Fig.\u00a06\na\u2013c) [60]. In situ experiments and finite element theoretical simulations demonstrate that the Ag@Cu core-shell catalysts with an average pore size of 4.9\u00a0nm induce the highest local \u2217CO concentration due to the confinement effect and therefore promote C\u2013C coupling. Thus, exhibited high Faraday efficiency of the C2+ products reached 73.7% under constant current reaction conditions with a total constant current density of 300\u00a0mA\u00a0cm\u22122 (Fig.\u00a06d and e).Moreover, a heterogeneous nanostructure of Cu-rich cores embedded in an In(OH)3 shell-like matrix was reported [61]. The derivative catalysts show high electrocatalytic performance at moderate overpotential. By comparing the properties before and after the In(OH)3 shell dissolved, they demonstrated that the presence of In(OH)3 is critical to the high selectivity of CO, and thus confirm the heterogeneous nanostructure is essential for improved activity and stability of CO2RR [62].By adjusting the shape and structure of bimetallic catalysts, the activity and selectivity of CO2 reduction can be effectively regulated. According to the previous work, we can conclude that controllable shape and structure can be achieved by various methods such as co\u2013reduction reaction [63], galvanic replacement reaction [64], solvothermal method [65], and seed-mediated method [66]. However, many challenges still need to be solved for nano bimetallic catalysts with controllable structures. For example, since Cu can be combined with a variety of metals to form bimetallic catalysts, strategies of generic and shape controllable synthesis methods suitable for different bimetallic systems are necessary. In addition, during the CO2 reduction process, the sharp dendrite structure is easy to be etched, the nanoparticles are easy to agglomerate, and the nanowires will fracture, thus resulting in performance degradation. Therefore, how to maintain the complete structure of the catalyst during the reaction is very critical. At present, many morphology protection methods have been developed such as: organic coating, construction of second metal protection layer and formation of stable oxide surface. However, the dynamic process of dissolution and reconstruction of electrode morphology should also be the research focus in the future. For example, Larraz\u00e1bal et\u00a0al. found the equilibration process may result in the creation of more active sites or generate metal pathways by bypass the less conductive In(OH)3 shell [61]. Weng et\u00a0al. proposed the Pd atoms pre-deposited on the Cu surface cause sustained morphological and compositional reorganization, which sustained refreshing the catalyst surface and thereby keeping the catalytic performance of CO2 reduction to hydrocarbons [67].As a special morphology, a sharp needle structure with a high curvature is always accompanied by a strong electric field in its vicinity [68]. Dong et\u00a0al. demonstrated that the strength of the electric field effect is related to the sharpness of the needle by comparing a series of well-defined morphological Cu\u2013Sn bimetallic foils, rods, and cones [69]. The COMSOL finite element analysis shows that the local electric field strength gradually increases as the top diameter (D\ntop) of the Cu cone decreases, while the electric field strength tends to decrease as the bottom (D\nbottom) radius increases (Fig.\u00a07\n). It is concluded that a high curvature can induce strong electric field effects and promote CO2 reduction. This is consistent with the experimental result that the Cu\u2013Sn electrode with the highest curvature has the best CO faradaic efficiency (FE) of 82.7%.Generally, the low solubility of CO2 in the electrolyte makes it difficult for CO2 molecules to adsorb on the electrode surface. Therefore, it is proposed that an electric field can promote the adsorption/activation of CO2. For example, N\u00f8rskov et\u00a0al. used theoretical calculations to prove the local electric field can significantly alter the free energy of the CO2 reduction to CO on Ag metals [70]. In addition, some studies have demonstrated that alkali metal cations can overcome the limitation of low CO2 solubility through non-covalent interactions with adsorbed CO2. Liu et\u00a0al. proposed a strategy to improve CO2 adsorption capacity by aggregating alkali metal cations concentration induced by nanoneedle electric field (Fig.\u00a08\na and b). Specifically, the local electric field can increase the K+ concentration on the electrode surface by a factor of 20, and as the K+ concentration increases, CO2 rapidly stabilizes on the sharp electrode [71].Similarly, due to the transfer difficulty of electrons at grain boundaries (GBs), the electrons tend to accumulate at grain boundaries. Zhong et\u00a0al. designed an Au\u2013Cu bimetallic nanochain aerogel (NC\u2013Au\u2013Cu) containing rich grain boundaries. They found the role of GBs is electron channels to promote the accumulation of charges. The charge accumulation induces a high local electric field for K+ concentration [72]. High K+ concentrations favor the concentration of CO2 near the electrode surface and reduce the formation energy of the rate-determining step of \u2217CO2 to \u2217COOH intermediate. Lee et\u00a0al. found the strong electric field is concentrated at the tip of the dendrite [73], the enrichment of K+ inhibits hydrogen generation by preventing H+ from approaching the surface while promoting CO2 adsorption/activation.Furthermore, electric field effects can promote CO2RR by facilitating the C\u2013C coupling process (Fig.\u00a08c and d) [74]. N\u00f8rskov et\u00a0al. reported the theoretical evidence of a CO dimerization mechanism on Cu(111) and Cu(100). The combination of field and solvation effects considerably reduces the thermodynamic and kinetic energy requirements for the formation of CO dimers [75]. This prediction has been proved by experimental results. Zhou et\u00a0al. reported that the nano-tip arrays with vertical alignment induce stronger electric fields, and that strong electric fields lead to more K+ enrichment, resulting in stronger \u2217CO adsorption and a lower energy barrier for C\u2013C coupling (Fig.\u00a08e) [76].The adsorbent interacts with the metal surface can obtain an adsorbate-surface dipole. The interaction of these dipoles with the electric field can have a profound effect on the catalytic reaction [77]. Resasco et\u00a0al. reported that surface intermediates with larger dipole moments (e.g., \u2217CO2, \u2217CO and \u2217OCCO) can be stabilized by electrostatic fields in the Outer Helmholtz Plane (OHP) [78]. This work provides a theoretical perspective on how the electric field affects the adsorption behavior of intermediates.Although the concentrated electric field at the tip can enhance CO2 adsorption/activation and promote C\u2013C coupling, the low density of the needle tip limits the effective area. It is worth further exploring how to achieve a more pointed tip and a larger effective active region. Safaei et\u00a0al. reported a hierarchical multi-layered structure to increase the density of the active tip [79]. The performance of CO2RR can be improved by multi nano-tips on porous carbon fibers. In detail, they prepared the new layered structure by partially covering the nanoneedles with thiol and then surface secondary gold electrodeposition. The new catalyst exhibited a stronger local electric field and three times higher current density than a single needle tip. Periodic array structures can also further enhance electric field effects and improve catalytic performance. The ordered nano-tip array structure had a higher local electric field, achieving almost double the selectivity of the C2 products compared to the disordered nano-tip array [76].Although electric field effects do kinetically facilitate the reduction of CO2. And it is undeniable that the electric field effect is indeed a promising strategy for modulating catalytic performance, changing the adsorption energy on the catalyst surface, without changing the catalyst material or structure. However, other side-reactions (e.g., HER) can be equally enhanced by electric field effects, and the competition between side-reactions and CO2 reduction reaction dictates that the tip size is not necessarily the smaller the better [80]. Furthermore, it remains to be investigated whether electric field effects are still present in industrial high current applications and how useful they can be. Beside, although it has been proved that the electric field effect is stronger with higher curvature structure, there are some problems need to be investigated. For example, whether this correlation can be quantified? Whether there has the best value of curvature? Therefore, the construction and utilization of the electric field effect must establish the reasonable model of curvature and electric field strength.Conventional nanostructure and morphology control techniques do not always work in complicated CO2 reduction pathways. To improve the selectivity and activity of Cu-based catalysts, rationally modifying the interface of the bimetallic catalysts may be possible to break the conventional linear scaling relationship and modulate the binding strength of the target intermediate [81\u201384]. Interface engineering can be designed to contain multiple sites and optimize the chemical environment around the active sites in Cu-based bimetallic catalysts [85]. Experimental and theoretical studies have shown that various synergistic effects of interface engineering maybe present between the different metal elements, resulting in superior catalytic performance. In this chapter, we classify and discuss interface engineering into three parts: (1) interface effect, which highlights the role of intrinsic properties of bimetallic interfaces (such as interface type and size) in CO2 reduction reaction, (2) interface atomic arrangement, which focuses on the influence of changing the mixing patterns of Cu and M atom, and (3) interface strain, which focuses on the changes caused by mixing of two metals with different lattice parameters.The coupling of two metals generally results in the creation of a bimetallic heterogeneous interface, which plays an important role in the CO2 reduction reaction. The bimetallic heterogeneous interface can promote CO2RR through various roles, including the adjustment of intermediate adsorption, electronic/reactant transfer, generation of more active sites, and avoiding catalyst poisoning [85\u201387]. For example, Zhang et\u00a0al. found the interaction of the bimetallic interface plays an important role in adjusting the selectivity and activity of CO2RR [88]. Remarkably, the reaction rate was greatly influenced by the interface activity. The Cu\u2013Au interface can effectively promote the dissociation of H2O to H\u2217, while preferentially enhancing CO2RR and suppressing HER. In addition, they found the adsorb energy of CO is larger on the Cu\u2013Au interface compared with the pure Cu surface, which indicated that the CO is easier to be reduced at the Cu\u2013Au interface. The \u2217CHO and \u2217CO has stronger adsorption on the Cu\u2013Au interface, which is beneficial to promoting C\u2013C coupling and formation of C2 products.CO2 reduction is always dissatisfied on Cu-based catalysts due to the weak CO2 activation ability of Cu [89]. The adsorption and activation of CO2 can be effectively promoted by regulating the bimetallic interface. Ye et\u00a0al. conducted a comprehensive understanding of how CO2 and H2O interact with the Cu\u2013Ag interface to promote CO2 adsorption and activation. The synergy effect of Ag and Cu was the key to tune the CO2 (H2O)\u2013Cu\u2013Ag interactions and change the activation process of CO2 [90]. Compared with pure Ag and Cu, the adsorbates (H2O or CO2) on the AgCu interface exhibit different geometric and electronic structures. The observed interface restriction process of Ag and Cu significantly promotes the CO2 adsorption/activation process.Bimetallic heterointerfaces can also effectively promote C\u2013C coupling to generate C2 products. Huang et\u00a0al. fabricated Ag\u2013Cu heterointerfaces with different Cu domain sizes (Ag1\u2013Cu0.4, Ag1\u2013Cu1.1, and Ag1\u2013Cu3.2) (Fig.\u00a09\na\u2013f) [26], the Ag1\u2013Cu1.1 catalyst with optimal content obtained FE of C2H4 was 3.4-fold compared with pure Cu (Fig.\u00a09g). The adjustable interface of Ag-Cu is related to the FE of C2H4. The size of the Cu can adjust the interface and control the partial Cu oxidation state, then achieve the best \u2217CO\u2013\u2217CO for C2H4 generation (Fig.\u00a09h). Similarly, Wang et\u00a0al. constructed a bimetallic Ag\u2013Cu catalyst with a sharp interface to promote the reduction of CO2 to C2H4. A high 42% FE of C2H4 can be obtained, which is about two-fold than pure Cu catalyst [91]. In addition to increased C2H4 production, ethanol performance also can be improved by bimetallic heterointerfaces. Yeo et\u00a0al. obtained a high yield of C2H5OH by adding Ag to Cu2O nanowires and concluded that the improvement of C2H5OH is attributed to the amount of CO released from Ag sites on the Cu-Ag interface.The interface of bimetallic catalysts can also optimize the intermediate adsorption performance to regulate products selectivity. Peng et\u00a0al. prepared a Cu\u2013Bi bimetallic electrode by using a simple one-step co-deposition method [92]. The Cu\u2013Bi bimetallic catalysts achieved an amazing FE of formate (94.37%) and high current density (27.85\u00a0mA\u00a0cm\u22122) at \u22120.91 V vs. RHE. Due to the introduction of Cu atoms into the Bi electrode, the intrinsic and electronic structures of Bi electrodes will be tuned, and thus promote the CO2RR process transfer on Cu\u2013Bi surface to stabilizing intermediate \u2217OCHO, which is the important intermediate to generate formate. The surface of pure Bi favors the formation of COOH\u2217 intermediates, while generating unwanted CO product.By coupling Cu with secondary metals, interface effect have been proven to effectively enhance CO2 adsorption/activation, facilitate C\u2013C coupling, and optimize the adsorption property of intermediates in the CO2 reduction process. Many previous works give us a lot of inspiration on how to understand and use interface effects. Nevertheless, many problems should be studied in the future. For instance, current research on the effect of the interface may pay more attention to the selectivity or activity of the CO2 reduction reaction. Follow-up works need to focus on how to achieve a larger interface between the two metal domains and keep their large interface stable during the electroreduction process. The catalyst design should consider how to engineer advanced nanoheterostructures and generate abundant interfaces to maximize interface effects.In addition to the regulation of the bimetallic interface, the rational design of the interface atomic arrangement of the bimetallic catalysts is also an effective strategy to improve the catalytic CO2 reduction performance. Nishimura et\u00a0al. investigated the effect of metal atomic modifications of Cu on electrocatalytic CO2 reduction [93]. The results showed that atomic-level bimetallic effects on Cu-based electrocatalysts are typically mediated through the inhibition of undesired side reactions (e.g., the HER). The precise regulation of the relative position and arrangement of the introduced M metal atoms is important [93].In addition, the different interface atomic mixing modes for the two metals have a great influence on the CO2 reduction performance. For example, Ma et\u00a0al. prepared a series of Cu\u2013Pd catalysts (Cuat: Ptat\u00a0=\u00a01: 1) with three mixed atomic arrangements of ordered, disordered and phase separated modes [94]. Interestingly, the ordered sample had the best C1 FE (>80%), while the phase-separated structure exhibited the best C2 products selectivity. This result is attributed to the different arrangement of atoms and active sites on the surface of these catalysts. The ordered Cu\u2013Pd catalysts have more alternating Cu\u2013Pd sites than disordered Cu\u2013Pd catalysts. With the CO adsorbed on the Cu atoms facilitating the formation of CHO intermediates and the O atoms adsorbed on the Pd atoms stabilizing the CHO adsorption and thus further facilitating the production of C1. In contrast, the phase-separated structure has more adjacent Cu atoms, which exhibit favorable molecular distances and less steric hindrance. It is favorable for C\u2013C coupling.The intermetallic ordered structures are generally more thermodynamically stable than their disordered counterparts. Moreover, the intermetallic ordered structures provide a unique electronic structure and coordination environment [95]. For example, Kim et\u00a0al. found that the conversion of Au\u2013Cu (Auat: Cuat\u00a0=\u00a01:1) nanoparticles from a disordered arrangement to an ordered arrangement enables catalytic selectivity in CO2 reduction reaction (Fig.\u00a010\na) [96]. As shown in Fig.\u00a010b, the most ordered o-AuCu mainly produces CO, while the disordered d-AuCu mainly produces H2. The ordered conversion can form a three-atom-thick Au capping layer, which is responsible for the enhanced CO2 reduction catalytic performance. The DFT calculation compared the thermodynamic limit potentials UL (CO2) (limiting potential for CO2 reduction) and UL (H2) (limiting potential for H2 release) of the three model systems. The UL (CO2) values of the three models indicate that o-AuCu has the best activity to reduce CO2 to CO (Fig.\u00a010c\u2013e). The smaller value of UL (CO2) \u2013 UL (H2) means a higher selectivity for CO2 reduction. As shown in Fig.\u00a010e, o-AuCu has better selectivity than d-AuCu.The phase separated structure with moderate atomic spacing and low steric hindrance, can provide a large number of active sites for optimizing the binding of intermediates [97]. For example, Wang et\u00a0al. prepared a phase-separated Cu3Sn/Cu6Sn5 catalysts by electrochemical deposition of Sn on Cu foams. The catalyst exhibited 82% Faraday efficiency of HCOOH at \u22121.0\u00a0V (vs. RHE) and excellent CO2 reduction stability (about 42\u00a0h) [98]. While the main products of both Cu3Sn and Cu6Sn5 are H2. DFT calculations showed that the adsorption energy of \u2217HCOO is more negative than \u2217COOH on Cu3Sn/Cu6Sn5, indicating that the formate is more likely to be formed on phase-separated Cu3Sn/Cu6Sn5. In addition, the Gibbs free energy of HER on Cu3Sn/Cu6Sn5 is more positive than on Cu3Sn and Cu6Sn5, suggesting that the phase separation Cu3Sn/Cu6Sn5 catalyst can exhibit an inhibition HER effect compared to Cu3Sn and Cu6Sn5.Introducing the metal vacancies on the catalyst surface can regulate surface atomic arrangement for the better electronic structure of adjacent atoms and reaction energy barrier of intermediates. Zhuang et\u00a0al. reported a Cu2S catalyst with surface Cu vacancies. The core S atoms and Cu vacancies on the shell can efficiently convert C2 alkene products to alcohol products [99]. The ratio of ethanol to ethylene greatly increased. DFT calculations show that the catalyst with Cu2S core and Cu vacancies increases the energy barrier for CO2 reduction to the ethylene pathway, while it does not affect the ethanol pathway (Fig.\u00a011\n).Zhu et\u00a0al. also reported an Au\u2013Cu alloy catalyst with surface enriched vacancies [100]. In particular, the de-alloyed Au3Cu alloy catalyst (De-Au3Cu) with enriched Cu vacancies exhibited the highest CO Faraday efficiency of 94.3% at \u22120.43 V vs. RHE (Fig.\u00a012\na). The DFT calculations (Fig.\u00a012b\u2013d) indicated the free energy for adsorption/activation of CO2 (\u2217CO2 to \u2217COOH) on the De-Au3Cu catalyst is lower than Au3Cu and Au(100). CO2 is more likely to be adsorbed on vacancies rather than on metal sites.During the CO2 reduction reaction, atomic arrangement engineering is a promising tool for catalyst design. By adjusting the surface atomic arrangement, changing the atomic spacing and coordination environment between adjacent Cu atoms and secondary metal M atoms, we can effectively change the adsorption mode and configuration of adsorbates on the catalyst surface. The adjusted adsorption strength of intermediates and optimized reaction pathway result in better catalytic activity and higher selectivity of CO2 reduction reaction. However, the implementation of precise controls bimetallic structure (such as the proportion and atomic arrangement pattern of bimetallic atoms) still is a great challenge. Atomic-scale material synthesis methods such as atomic layer deposition techniques should be used and studied.Different lattice parameters of two different metal surfaces can lead to strain effects [101,102]. It is proven that the strain effects can effectively influence the performance of CO2 reduction [103]. The strain effect is usually induced by lattice mismatch and lattice dislocations. In addition, the structure transformations can also result in the strain effect. For example, Zhang et\u00a0al. reported a FePt/Pt NPs catalyst for high CO2 reduction performance. The high catalytic activity is attributed to the Pt strain. This strain can be changed by structure transition from the fcc (face centered cubic) structure to the fct (face centered tetragonal) structure [104]. Generally, there are two types of strain effects: tensile strain and compressive strain, we focus on the different effects of these two types of strain.The strain effect can enhance the performance of CO2 reduction. For example, due to the smaller lattice parameter of Cu, reducing the number of copper layers in Au@Cu core@shell can increase the tensile strain of the shell, and result in higher selectivity of C2H4 product [105]. Similarly, Reske et\u00a0al. investigated the effect of Cu cover layers of different atomic thicknesses on Pt [106]. As the thickness of the Cu cover increases, the lattice mismatch between Cu and Pt raises the interatomic distance of Cu and produces a tensile strain effect, thus leading to the suppressive H2 formation. Those studies provided new insight into strain-tuned CO2 reduction performance. This means it is a good approach that controlling the strain effect by varying the thickness of the Cu cover layer on the interface.Strain effects can also enhance the selectivity of CO2RR by inhibiting competing HER. Clark et\u00a0al. reported Cu\u2013Ag bimetallic catalysts with excellent performance to produce multi-carbon oxygenate by electrochemical reduction of CO2 [107]. This excellent performance was attributed to the selective suppression of HER induced by the compressive strain of a Cu\u2013Ag surface alloy. Specifically, compressive strain causes a valence band modifier, which leads the valence band structure of Cu to move to deeper levels. The weaker adsorption energy of H results in inhibited HER activity in the Cu phase. Similarly, Du et\u00a0al. obtained different elastic strain states of 32\u00a0nm and 5\u00a0nm Cu films and investigated their effect on CO2RR performance (Fig.\u00a013\na) [108]. In their study, tensile strain in 32\u00a0nm overlay facilitates CH4 production and increases the CH4 Faraday efficiency of CO2RR from 65.02% to 76.48%. In contrast, compressive strain significantly reduces its CH4 selectivity. The compressive strain in the 5\u00a0nm overlay reduces the H adsorption energy and thus suppresses the HER (Fig.\u00a013b and c).In bimetallic systems, the presence of strain effects has provided new regulatory strategies for designing catalysts with excellent performance. However, the lack of specific modulation methods for strain effects, the lack of characterization tools, and the absence of a uniform representation of the degree of strain have limited the application of this strategy. Therefore, in future research, in addition to investigating the effect of different strain types on the performance of CO2RR, it is important to further investigate what level of strain can have a significant effect on the performance of CO2RR.Reasonable utilize interface engineering is an important means to improve CO2 reduction performance. Thus, it is important to develop advanced interface research methods, such as in-situ ultrafast temporal resolution and spatial resolution spectroscopic techniques, which can directly observe the transient structure and interface transform during CO2 reduction, provides the possibility to reveal the dynamic evolution of the real active site. Obtain the most comprehensive information on the interfacial reaction process and reaction mechanism, and then summarize the regularity of interface engineering.Regulation of structural properties of the catalysts is the key strategy to alter the catalytic performance, as discussed in the previous section. Other approaches, such as the electronic effect and tandem effect, have been employed to improve the performance of Cu-based bimetallic catalysts for CO2RR. Due to the different electronic configurations, coupling two types of metallic atoms can cause charge redispersion, and leads to the electronic effect. Therefore, the reduction behavior of CO2 on the catalyst is altered and exhibits high selectivity of desired products. Moreover, by rational design the active sites between Cu and the second metal, the bimetallic catalysts can drive the tandem effect to exhibit more deeply reduced multi-carbon products.The electronic effect has a unique ability to facilitate the reaction rate and the activity of CO2RR. From a thermodynamic point of view, due to the different electronic configurations of the host metal and the guest metal, the introduction of the guest metal can change the electronic structure of the host metal, thus changing the binding energy of intermediates [109,110]. In addition, N\u00f8rskov et\u00a0al. reported that the electronic effects can result in the up and down shift of the d-band center, thus determining the reaction activity or binding energies of the intermediates and reactants [109,111]. Similarly, An et\u00a0al. demonstrated the performance of AuCu3@Au catalyst is related to the binding strength of \u2217COOH intermediate, influenced by the surface electronic structure (d-band center energy). The density of states indicates that the introduction of Cu causes the d-band center of AuCu3@Au to move toward the Fermi level, resulting in stranger adsorption of \u2217COOH on the catalyst surface [112].In Cu-based bimetallic catalysts, Liu et\u00a0al. prepared a homogeneous planar film to investigate the electronic effects on the CO2 reduction performance of Au\u2013Cu bimetallic films (Fig.\u00a014\na\u2013d) [97]. The result indicated that as the Au content increases, the d-band center deviates from the Fermi level. The oxygen binding strength is significantly weakened while the desorption energy of \u2217CO decreases, so favorable the release of \u2217CO and suppressed the formation of \u2217OCHO, thus increasing the CO yield and limiting the formation of HCOOH. Kim et\u00a0al. demonstrated that the increased number of electrons in the s-band and the upward shift of the d-band center position of Au can influence the bonding to CO and result in the activity of electrochemical reduction of CO2 to CO [113].Due to the different binding abilities of intermediates, electronic effects can lead to a change in reaction pathways. Zu et\u00a0al. proposed that the electronic effect promotes the formation pathway of HCOOH, while inhibiting the CO and H2 production [114]. As we known, the pathway to produce HCOOH is through the C atom of the \u2217COOH intermediate bound to the catalyst surface, but meanwhile will produce unnecessary CO and competition HER, thus the selectivity of HCOOH is always unsatisfactory. However, after the doping of Bi atoms in Cu nanocrystals, the electronic effect can result in a high HCOOH selectivity by another pathway to produce \u2217OCHO intermediate through the O atoms bound to the catalyst surface. This pathway is more favorable for HCOOH production because the \u2217OCHO intermediate almost does not produce CO.The electronic effects originate from the changes in electronic structure, it also can promote CO2 activation and the C\u2013C coupling process. Zhang et\u00a0al. investigated the changes in CO2 reduction properties by doping a series of transition metals on Cu(100) by the DFT method [115]. The result indicated the order of catalytic activity is Zn\u00a0>\u00a0In\u00a0>\u00a0Cd\u00a0>\u00a0Ni\u00a0>\u00a0Sn\u00a0>\u00a0Fe\u00a0>\u00a0Pd\u00a0>\u00a0Co. Due to the doping of Zn, not only alters the electronic around Cu, but also changes the atomic arrangement of the active sites. The active sites on the catalyst surface become electronegative, which are favorable for CO2 activation and lower the energy barrier of C\u2013C coupling.Overall, the electronic effects not only effectively promote the activation of CO2 [116,117], but also optimize the adsorption energy of intermediates [97,113,118], and even promote C\u2013C coupling reactions [115]. However, the electronic effects often co-exist with other effects (such as strain effects), so it is very important to distinguish those effects. The electronic effect refers to the change of the chemical environment around the A atoms caused by the specific chemical properties of the B atoms. The interface strain effect originates from the lattice distortion due to the different atomic radii of the two atoms. In general, the electron effect is a short-range effect, and the strain effect is a long-range effect. The electronic effects require the surface atoms of the catalyst to be located on the surface or in the first or second subsurface layer. However, it is reported the strain effect has no such limitation and only strain effects can influence the reactions of more than multi atomic layers. For example, Strasse et\u00a0al. found the strain effects could change the distance of surface Pt atoms, even though the other atoms are buried deeper than 1\u00a0nm [119]. Maark et\u00a0al. demonstrated as the number of Cu overlayers increases (from 1 to 3 layers), electronic effects are filtered out, ultimately there are only the strain effects that exist in the Cu overburden [120]. Although it is difficult to distinguish those two effects by the experimental method, DFT calculations can easily examine them individually. For example, by constructing a Cu-M bimetal with the same lattice constant as Cu, it is possible to exclude strain effects and study electronic effects alone.Electrochemical CO2RR is a multiple electron/proton transfer process and requires multiple steps and intermediates. Tandem catalysts offer the possibility of breaking the linear scaling relationship and improving catalytic performance by coupling multi-step reactions [121]. The Tandem catalytic process refers to the intermediates (such as \u2217CO) that are selectively produced on metal A sites, in the subsequent CO2RR process, transferred to the metal B sites, and further reduced to other products on metal B sites [122].In Cu-based bimetallic tandem catalysts, \u2217CO is the most common intermediate, which can be generated from some metals such as Au, Ag, and Zn. The transfer of \u2217CO to the Cu can enhance the possibility of C2 products generation. Therefore, improving the availability of \u2217CO intermediates is the key to the subsequent deep reduction reactions. Zhang et\u00a0al. investigated bimetallic tandem catalysis of Cu-modified Ag and Au, and gave the experimental evidence of \u2217CO spillover by ATR-SEIRAS [123]. The free energy barriers for CO spillover on both Au and Ag surfaces are small, and the \u2217CO adsorption on surface Cu sites is more stable than on Ag or Au sites, demonstrating the thermodynamic and kinetic feasibility of CO spillover from Ag or Au to Cu surfaces. It is beneficial to C\u2013C coupling and improves the C2 products selectivity.In addition, high coverage of \u2217CO allows for more CO spillover and is thermodynamically favorable for CO2RR to multi-carbon products [124\u2013126]. Gao et\u00a0al. operated the Raman spectroscopy revealed the formation and transfer of \u2217CO on Ag/Cu catalysts, \u2217CO spillover efficiency of approximately 95% from Ag to Cu, which significantly contributed to the selectivity of the C2 products [127]. The higher \u2217CO coverage enables easier \u2217CO adsorption of Cu, as well as greater opportunity for C\u2013C coupling.Ren et\u00a0al. propose a two-site mechanism and \u2217CO insertion mechanism to explain the whole tandem reaction process [128]. In the Cu\u2013Zn catalysts, CO2 is firstly being reduced to \u2217CO on Cu or Zn (1\u00a0\u2192\u00a02 in Fig.\u00a015\na). Secondly, \u2217CO can be further reduced to \u2217CHO or \u2217CH\nx\n (x\u00a0=\u00a01\u20133) on the Cu site (2\u00a0\u2192\u00a03 in Fig.\u00a015a). Thirdly, \u2217CO will be desorbed on Zn, and the desorbed \u2217CO will diffuse to the Cu site (2\u00a0\u2192\u00a03 in Fig.\u00a015a). Then the spilled-over \u2217CO can insert into the bond between the Cu surface and \u2217CH2, to form \u2217COCH2 (3\u00a0\u2192\u00a04 in Fig.\u00a015a). Further reduced to CH3CHO (4\u00a0\u2192\u00a05 in Fig.\u00a015a) and finally to form C2H5OH (5\u00a0\u2192\u00a06 in Fig.\u00a015a). Similarly, Lee et\u00a0al. further demonstrated that during the formation of C2H5OH, the sample which enriched Cu\u2013Ag biphasic boundary can facilitate the insertion of \u2217CO into the \u2217CH2 and Cu, and result in the enhancement of the C2H5OH pathway (Fig.\u00a015b) [129]. The homogeneous mixing Cu\u2013Ag biphasic boundary can provide a sufficient interface for enhancing tandem catalytic performance.The Tandem effect can effectively improve the selectivity of C2 products, such as C2H5OH [128,130,131] and C2H4[112]. However, many factors affect the tandem process. Firstly, it depends on the space management of the \u2217CO transfer. Zhang et\u00a0al. showed in a tandem catalytic process, the appropriate spatial management of \u2217CO transport can enhance the yield of C2 products. The migration of \u2217CO between the ZnO and Cu sites does not require the two active sites to be adjacent [132]. Lee et\u00a0al. also found the spacing of Ag and Cu should be abundant and suitably, so the insertion process of CO can become beneficial [129].In addition, Buonsanti et\u00a0al. reported the facet-dependent selectivity for tandem catalytic. The results show that due to the tandem effect of Cuoh ((111) facets with high CH4 selectivity) and Cucub ((100) facets with high C2H4 selectivity), both Cuoh-Ag and Cucub-Ag catalysts can inhibit CH4 and H2, while promoting the selectivity of C2H5OH. Cuoh-Ag has a better selectivity of C2H5OH than Cucub-Ag. Due to the enriched CO generated by Ag, the key step of the C2H5OH pathway is \u2217CH\nx\n\u2013\u2217CO coupling process. However, the active sites which can promote this step are only distributed at the edges and corners of the Cucub-Ag catalyst, this means that the Cucub-Ag cannot efficiently reduce CO2 to C2H5OH [130]. Except for this facet-dependent tandem catalysis performance, they reported a size-dependent tandem catalysis on CO2 reduction selectivity [131]. In the Cucub-Ag catalyst, with the enriched \u2217CO intermediate, Cucub with a smaller size has higher selectivity of C2H5OH due to its larger edge-to-faces ratio.The tandem effect shows great potential for improving the selectivity of valuable C2 products. In this process, \u2217CO is the main intermediate, so its spillover processes are studied in detail. The efficiency of \u2217CO spillover can effectively promote the tandem effect. The high \u2217CO coverage provides more available CO, which is beneficial to C\u2013C coupling and generation of C2 products. However, it is worth noting that the key evidence of \u2217CO spillover is usually obtained from Operando Raman Spectroscopy. More accurate and convenient characterizations should be developed in the future. Based on advanced in situ characterization techniques, future studies on tandem catalysts should focus on the efficient utilization of \u2217CO intermediates and their effects on CO2 reduction reaction. For example, it has been demonstrated that under high \u2217CO coverage, \u2217CO with top adsorption mode is more favorable for the reduction of CO2 to C2 products. In addition, the confinement effect is an effective strategy to enrich the \u2217CO intermediate, which deserves further study.By summarizing the CO2 reduction process, it can be concluded that: (1) CO2 adsorption/activation is often the rate-determining step that determined the activity of CO2RR. (2) The C\u2013C coupling process is the key to the generation of the more valuable C2 products. (3) The different adsorption abilities between two metals and intermediates can lead to different reduction products. Cu-based bimetallic catalysts offer new strategies to break the traditional linear scale relationship and improve the activity and selectivity of CO2RR. We summarize recent advances in the reduction of CO2 over Cu-based bimetallic catalysts. The role that second metal act as a reaction site is considered in detail. Such as the strain effect can change the d-band center of Cu, and thus alter the adsorption behavior of the intermediates. The tandem effect enables the coupling of multiple reaction steps to achieve sequential successive catalysis. In addition, the role of the second metal that only change the morphology and electronic structure is also briefly analyzed (morphology effect and electronic effect). For CO2 reduction reaction, provides a comprehensive comprehension of the design of Cu-base bimetallic catalysts in conjunction with the theoretical understanding. However, the reduction of CO2 to valuable additional products at low cost and high efficiency still is a big challenge. It is required to develop more efficient catalysts and a more instructive understanding of the mechanisms of CO2 reduction. Therefore, we briefly discuss the perspective for future studies in these areas.\n\n(1)\nThe atmospheric CO2 concentration is extremely low compared to the ideal experimental conditions (saturated CO2 concentration in the electrolyte). It remains to be seen whether the prepared catalysts are still effective under atmosphere circumstances. Therefore, it is important to achieve efficient conversion of CO2 at low concentrations to meet the carbon reduction target.\n\n\n(2)\nThe electrocatalytic reduce CO2 to multiple carbon products in liquid electrolytes faces two challenges: the low activity of the catalyst in non-alkaline electrolytes, and the formation of carbonates in alkaline electrolytes. Therefore, it is necessary to develop a strategy for designing efficient reduce CO2 in liquid electrolytes. In addition, the electrolyte currently used for CO2 reduction is mainly KHCO3, because it has been suggested beneficial to the dissolution of CO2 and thus promotes the reaction. However, the current research on how the electrolyte properties (e.g., concentration, species, pH) affect the CO2 reduction performance is still very limited. Therefore, more attention needs to be paid to the effect of electrolytes in the future.\n\n\n(3)\nIt is well known that the adsorption configuration often influences catalytic performance. However, there are a few studies about whether the reactant molecules/intermediate species affect the catalyst interface property. To maintain the stability of the catalyst. Research in this area should be considered seriously.\n\n\n(4)\nAt present, the reduction products on Cu-based bimetallic catalysts are mostly C2 products. Multi-carbon products, such as C3 and C4 products are still few and have very low selectivity. In the complex process of CO2 reduction to multi-carbon products, it is particularly important to design catalysts with C3 and C4 products. Using the multi-metal system to complete the reactions of each step and in turn, to form a tandem catalytic process to obtain products with high added-value is a possible strategy. Therefore, in the future should focus on the design of Cu-based ternary (or even more than ternary) catalysts.\n\n\nThe atmospheric CO2 concentration is extremely low compared to the ideal experimental conditions (saturated CO2 concentration in the electrolyte). It remains to be seen whether the prepared catalysts are still effective under atmosphere circumstances. Therefore, it is important to achieve efficient conversion of CO2 at low concentrations to meet the carbon reduction target.The electrocatalytic reduce CO2 to multiple carbon products in liquid electrolytes faces two challenges: the low activity of the catalyst in non-alkaline electrolytes, and the formation of carbonates in alkaline electrolytes. Therefore, it is necessary to develop a strategy for designing efficient reduce CO2 in liquid electrolytes. In addition, the electrolyte currently used for CO2 reduction is mainly KHCO3, because it has been suggested beneficial to the dissolution of CO2 and thus promotes the reaction. However, the current research on how the electrolyte properties (e.g., concentration, species, pH) affect the CO2 reduction performance is still very limited. Therefore, more attention needs to be paid to the effect of electrolytes in the future.It is well known that the adsorption configuration often influences catalytic performance. However, there are a few studies about whether the reactant molecules/intermediate species affect the catalyst interface property. To maintain the stability of the catalyst. Research in this area should be considered seriously.At present, the reduction products on Cu-based bimetallic catalysts are mostly C2 products. Multi-carbon products, such as C3 and C4 products are still few and have very low selectivity. In the complex process of CO2 reduction to multi-carbon products, it is particularly important to design catalysts with C3 and C4 products. Using the multi-metal system to complete the reactions of each step and in turn, to form a tandem catalytic process to obtain products with high added-value is a possible strategy. Therefore, in the future should focus on the design of Cu-based ternary (or even more than ternary) catalysts.The authors declare no conflicts of interest.The authors acknowledge the financial support from the International Science and Technology Cooperation Program (Grant No. 2018YFE0203400 and 2017YFE0127800), the National Natural Science Foundation of China (Grant No. 22002189, 21872174, and U1932148), Hunan Provincial Natural Science Foundation (2020JJ2041, 2020JJ5691) and Key R&D Program of Hunan Province (2020WK2002). We are grateful for resources from the High Performance Computing Center of Central South University.", "descript": "\n                  Electrocatalytic CO2 reduction reaction (CO2RR) is one of the effective means to realize CO2 resource utilization. Among the high-efficiency metal-based catalysts, Cu is a star material profiting from its ability for CO2 reduction into valuable hydrocarbon products. However, due to the difficulty in activating CO2 and regulating intermediate adsorption/desorption properties, the CO2RR activity and selectivity of Cu-based catalysts still cannot meet the requirements of industrial applications. The design of Cu-based bimetallic catalysts is a potential strategy because the introduction of the second metal can well promote the activation of CO2 and break the linear scaling relationship in intermediate adsorption/desorption. In this review, the synergistic enhancements of Cu-based bimetals on CO2 activation and intermediate adsorption/desorption are analyzed in detail, including the advantages caused by the morphology of Cu-based bimetallic catalysts, the local electric field effect induced by the special nanoneedle structure, the interface engineering (strain effect, atomic arrangement, interface regulation), and other particular effects (electronic effect and tandem effect). Finally, the challenges and perspectives on the development of Cu-based bimetallic catalysts for CO2 reduction are proposed.\n               "}
{"full_text": "The authors declare that the data supporting the findings of this study are available within the article and the supplemental information. The data and results supporting the present study are available from the lead contact upon request.Single-atom catalysts (SACs), comprised of isolated metal atoms anchored on supports, have attracted great attention in recent years because of their high efficiency and excellent selectivity toward various reactions.\n1\u20136\n When the metals on the supports are downsized to the minimum, SACs reach superhigh atom-utilization efficiency (close to 100%).\n7\u201310\n Moreover, the isolated metal atoms usually deliver different electronic states than that of the metal nanoparticles, endowing excellent catalytic performance due to the strong metal-support interaction and distinctive coordinated environments.\n11\u201313\n The unique, well-defined, and uniform active sites of SACs also provide an ideal platform for studying the electrocatalytic mechanism at the atomic level.In the last decade, rational design and fabrication of SACs have been wildly developed, and some inspiring results (including synthesis methods and electrochemical performance) have been reported.\n14\u201317\n There novel synthesis methods include the incipient wet chemical method,\n18\n\n,\n\n19\n pyrolysis method,\n20\n\n,\n\n21\n physical deposition method,\n22\n\n,\n\n23\n and atom trapping method among others.\n24\n\n,\n\n25\n For instance, the wet chemical method has been demonstrated to be a general and effective strategy for the synthesis of SACs and is achieving some inspirational results. However, owing to the limitations of the mechanism, the synthesis of SACs with large-scale production and high metal loading remains a great challenge for the wet chemical method.\n26\n\n,\n\n27\n The facile pyrolysis method has been reported to show some advantages in the realization of high metal loading of SACs, but the loading could not be easily controlled owing to the complex and unknown high-temperature chemical reaction pathway.\n20\n\n,\n\n28\n In a word, significant progress has been achieved on the preparation strategies of SACs, while some issues still need to be addressed. Firstly, the metals tend to leach off in the operating conditions due to the high free energy of single-metal atoms, so strong metal interaction or complicated defect strategies are usually needed to stabilize the single-metal atoms.\n29\n\n,\n\n30\n Secondly, sophisticated experimental design is needed for many synthesis methods owing to the disparate coordination environment between various metal atoms and defect types.\n31\u201334\n Thirdly, the SACs derived from the bulk metal salt or compounds still need to be achieved at high temperatures, and the processes are generally time consuming.\n35\u201337\n Moreover, the single-batch production ability is usually at the milligram level, which greatly limits its practical application.\n38\n\n,\n\n39\n Therefore, developing a general, high-efficient, large-scale, and tunable method without high temperature to synthesize SACs is still urgently needed.Herein, we demonstrate a general and high-efficiency \u201cplasma bombing\u201d strategy to directly transform commercially available metal salts into SACs on a large scale. Notably, the proposed strategy can be readily extended to various transition metals, benefiting the synthesis and practical applications of the SACs. As a typical case, the prepared SAC-Fe/NC catalyst displays a remarkable performance and durability both for rotation disk electrode (RDE) and the assembled zinc-air batteries (ZABs). Furthermore, density functional theory (DFT) calculations provide deep insights into the remarkable oxygen reduction reaction (ORR) mechanism that the adsorbent-induced spin-crossover effect is involved and the superb ORR performance is mainly contributed by the pyridinic-type Fe-N4 moieties.The synthesis process of the \u201cplasma bombing\u201d strategy to prepare SACs is shown in Figure\u00a01A (more detailed synthesis information is shown in the experimental procedures section). Briefly, the bulk metal salts are excited and stripped to form the flow of the single-metal atoms under the nitrogen \u201cplasma bombing\u201d treatment, and simultaneously the single-metal atoms are trapped and anchored by the defective and nitrogenous sites on the carbon supports to obtain the SACs.The SAC-Fe/NC is characterized as a model to verify the successful preparation of the \u201cplasma bombing\u201d strategy to fabricate SACs. Transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) images of the SAC-Fe/NC display highly disordered carbon structures, and no obvious metal singles (e.g., metal nanoparticles and/or metal clusters) are found on the carbon supports (Figures\u00a01B and S1; supplemental experimental procedures). In addition, the X-ray diffraction (XRD) patterns of pure NC and SAC-Fe/NC are shown in Figure\u00a0S2: the location and densities of the peaks of SAC-Fe/NC are almost the same as pure NC, indicating no peaks of Fe crystals are detected. Moreover, the Fe signals can be clearly observed from the XPS survey and high-resolution Fe 2p XPS spectrum (Figures\u00a0S3A and S3B), confirming the existence of the Fe species on the carbon supports. Furthermore, the highly dispersed and homogeneously distributed Fe sites can be observed in the aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (AC HAADF-STEM) image (Figure\u00a01C). The HAADF-STEM EDS mapping images show that the Fe and N elements are uniformly dispersed over the whole carbon supports (Figure\u00a01D). The metal content of the Fe in SAC-Fe/NC is as high as 8.5 wt % (Table\u00a0S1), determined by the inductively coupled plasma optically emitting spectrometer measurement (ICP-OES).X-ray absorption near edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) are conducted to determine the fine electronic structure of the Fe species on SAC-Fe/NC. Figure\u00a01E exhibits the Fe K-edge XANES of SAC-Fe/NC and standard materials (e.g., Fe foil and Fe2O3). The Fe K-edge absorption energy of SAC-Fe/NC is between the Fe2O3 and Fe foil, indicating the Fe atoms exhibit a positive valence state below\u00a0+3. A dominate peak at approximately 1.44\u00a0\u00c5 is detected in the EXAFS data of SAC-Fe/NC, corresponding to the first Fe-N coordination shell (Figure\u00a01F). A Fe-Fe bond at approximately 2.2\u00a0\u00c5 for Fe foil is not observed in SAC-Fe/NC, confirming the Fe species exist as signal isolated atoms on the supports (Figure\u00a01F). Figure\u00a01G displays the best-fitting result of the EXAFS data of SAC-Fe/NC compared with the experiment data (Fe species are coordinated with four pyridinic-type N atoms), revealing that the Fe atoms are bonded by four pyridinic-type N atoms (Fe-(pd)-N4 coordination structure). The atomic configuration of SAC-Fe/NC is further investigated by the Fe K-edge wavelet transform (WT)-EXAFS, due to the powerful resolution in both k and R spaces of the WT. Different from the WT signals of Fe foil and Fe2O3, we do not find the Fe-Fe coordination in SAC-Fe/NC, yet a bright and clear Fe-N signal is observed, which further identifies the isolated feature of Fe species in SAC-Fe/NC (Figures\u00a01H and 1J).The specific surface area and pore structure of the carbon supports is characterized by the Brunauer-Emmett-Teller. As shown in Figure\u00a0S4, the specific surface area and\u00a0pore volume of SAC-Fe/NC are evaluated to be as high as 624 m2 g\u22121 and 0.95\u00a0cm3 g\u22121, respectively. In addition, the N2 adsorption shows a typical type I behavior of SAC-Fe/NC, indicating the existing small pore size of the carbon supports. The pore size of the SAC-Fe/NC is further estimated to be 1.2\u00a0nm by calculating the adsorption branch via nonlinear DFT (Figure\u00a0S4B). The carbon structure of the prepared catalyst and pure NC are also characterized by Raman spectroscopy. There are two obvious broads obtained for both SAC-Fe/NC and pure NC, which belong to the D band and G band, respectively (Figure\u00a0S5). The ID/IG values of the SAC-Fe/NC and pure NC are calculated as 1.15 and 1.09, respectively. The higher the ID/IG values are, the more structural defects. The structural defects of the SAC-Fe/NC are generated by the \u201cplasma bombing\u201d treatment, and more defects are beneficial to capturing the dissociated Fe atoms to form the Fe SACs.As a new strategy for fabricating SACs, one of the most important aspects is robust and general. The proposed \u201cplasma bombing\u201d strategy can be readily extended to the preparation of other SACs with the similar synthesizing procedures of SAC-Fe/NC by only varying the metal precursors (detailed information is shown in the experimental procedures section). The atomically dispersed atoms (Mn, Ni) are confirmed by the AC HAADF-STEM images (Figure\u00a0S6), and XRD patterns as well as HAADF-STEM EDS mapping images of the prepared samples provide more solid evidence for the successful synthesis of various SACs (Figures\u00a0S7 and S8; Table\u00a0S1). To demonstrate the superiority of the proposed \u201cplasma bombing\u201d strategy to fabricate SACs, recently reported SAC preparation methods are summarized in Table\u00a0S2. The proposed \u201cplasma bombing\u201d strategy outperforms most of the previously reported SAC synthesis approaches, in terms of synthesis time, metal loading, universality, and operational difficulty. In short, the proposed \u201cplasma bombing\u201d strategy is a general, novel, and high-efficiency way to synthesize SACs.The ORR performance of the prepared catalysts is measured in O2-saturated 0.1\u00a0M KOH solution, and Pt/C is also evaluated for comparison. The SAC-Fe/NC exhibits a high half-wave potential (E1/2) of 0.920\u00a0V and kinetic current density (jk) of 9.890 mA cm\u22122 (Figures\u00a02A and 2B and Table\u00a0S3), much higher than those of Pt/C (0.860 V, 1.755 mA cm\u22122) and pure NC (0.768 V, 0.165 mA cm\u22122). Compared with the Pt/C (72\u00a0mV dec\u22121) and pure NC (114\u00a0mV dec\u22121), the SAC-Fe/NC exhibits the smallest Tafel slope of 67\u00a0mV dec\u22121, confirming much faster kinetics of the ORR process (Figure\u00a02C). Furthermore, a low peroxide yield of less than 4% over the potential range of 0\u20131.0\u00a0V is obtained for the SAC-Fe/NC, indicating a dominant four-electron ORR process (Figure\u00a0S9). Besides the high activity, there is almost no activity decay in terms of the E1/2 and jk after accelerated durability test (ADT), indicating superb durability of the prepared SAC-Fe/NC (Figure\u00a02D and Table\u00a0S4).To probe the potential commercial application of the SAC-Fe/NC, it is applied as the air-cathode electrode for the homemade ZABs. As shown in Figure\u00a02E, the peak power density of SAC-Fe/NC-based ZAB is 263 mW cm\u22122, which is about 1.7-fold higher than the Pt/C-based ZAB (151 mW cm\u22122). What\u2019s more, the specific capacity of the SAC-Fe/NC-based ZAB reached 803 mAh g\u22121, much better compared with Pt/C-based ZAB (701 mAh g\u22121) (Figure\u00a02F). The rate-discharge curves over various discharge current densities of the SAC-Fe/NC-based ZAB are shown in Figure\u00a02G, and a stable plateau is displayed at each discharge current density. Moreover, the SAC-Fe/NC-based ZAB can stably work for over 104\u00a0h without an obvious decrease in the discharge voltage, which also reveals that the SAC-Fe/NC-based ZAB possessed excellent discharge stability (Figure\u00a0S10). Additionally, the unit output of the SAC-Fe/NC is 1.16\u00a0g (Figure\u00a0S11), and considering its efficient preparation process (about 4\u00a0h for a single batch), the proposed \u201cplasma bombing\u201d method exhibits great potential for large-scale production. The performance comparison of SAC-Fe/NC catalyst and the recently published non-noble metal-based SACs in both RDE and ZAB level are summarized in Figures\u00a02H and 2I and Tables\u00a0S5 and S6. It can be seen that SAC-Fe/NC demonstrates outstanding performance among these catalysts, further verifying the highly effective and durable quality of SAC-Fe/NC catalyst synthesized by the \u201cplasma bombing\u201d strategy. The excellent ORR performance of the prepared SAC-Fe/NC catalysts could owe to the unique and uniform active sites, as well as the strong metal-support interaction and distinctive coordinated environments. Moreover, to further verify the high efficiency and practicality of the proposed \u201cplasma bombing\u201d strategy, the ORR activity (including RDE and ZAB level) of the SACs prepared by different synthetic methods is also summarized in Table\u00a0S2. The ORR performance of prepared SAC-Fe/NC is found to outperform most of the reported catalysts (Table\u00a0S2). It provides solid evidence to confirm the practicality of the proposed \u201cplasma bombing\u201d strategy.First principal calculations were carried out to provide theoretical insights into the structure and electronic structure evolution of Fe-N4 sites during the oxygen reduction reaction process. Previous investigations on the spin-crossover (SC) ORR mechanism show that the spin states in Fe cations have a visible impact on the ORR performance of Fe-based SACs.\n40\n To determine the ground state of pyridinic-type (denoted as Fe-pd-N4) and pyrrolic-type FeN4 (denoted as Fe-po-N4), various spin states, including low-spin (LS), intermediate-spin (MS), and high-spin (HS) states, are considered. As summarized in Tables\u00a0S7 and S8, the type of nitrogen atoms influences the ground state of Fe SAC. Fe-pd-N4 and Fe-po-N4 prefer MS and HS state as the ground state, respectively. The distribution of spin-charge density indicates that the Fe cations make a major contribution to the magnetic moments; see Figure\u00a03A and Tables\u00a0S7 and S8. It is worth noting that there may be a limited number of transferring charges from Fe cations in FeN4 moieties to O\u2217 anions; the O\u2217 anions with partially filled p orbitals exhibit relatively large local magnetic moments of 0.40 and 0.45\u00a0\u03bcB in Fe-pd-N4 and Fe-po-N4 systems, respectively. Moreover, their ground states are vulnerable to adsorbed intermediates. The OOH\u2217, O\u2217, and OH\u2217 configurations for Fe-pd-N4 exhibit LS, MS, and HS ground state, respectively, while those for Fe-po-N4 are in HS, MS, and HS in the ground state, respectively, revealing an adsorbent-induced spin-crossover (AISC) ORR process in Fe SACs.Based on the computational hydrogen electrode (CHE) methodology, four ORR pathways, including LS, MS, HS, and AISC pathways, are calculated to evaluate the catalytic performance of Fe-pd-N4 and Fe-po-N4. The energy-favoring AISC pathway plays a major contribution to the ORR performance, while the contributions of LS, MS, and HS pathways on catalytic activity are relatively small. The rate-determining step of the LS pathway in Fe-pd-N4 is the formation of OH\u2217, while that of the others is the desorption of OH\u2217. As displayed in Figures\u00a03B\u20133E, the LS and MS pathways exhibit a smaller overpotential than the HS pathway in both Fe-pd-N4 and Fe-po-N4. Moreover, the overpotential of LS, MS, and SC pathways in Fe-pd-N4 is estimated to be 0.57, 0.48, and 0.73 V, smaller than those of LS (0.69 V), MS (0.70 V), and SC (0.90 V) pathways in Fe-po-N4, respectively, indicating that the ORR activity of pyridinic-type Fe-pd-N4 is better than that of pyrrolic-type catalysts.In conclusion, a general \u201cplasma bombing\u201d strategy to transfer bulk metal salts into SACs is developed and validated, which is confirmed to be facile, effective, tunable, and capable of large-scale production. Driven by the energy of the plasma bombing, the surface metal salts are evaporated to generate single-metal atoms, which are trapped and anchored by the defective and nitrogenous sites of the carbon support, forming the isolated SAC-M/NC catalysts (M\u00a0= Fe, Mn, Ni, and so on). Impressively, the prepared SAC-Fe/NC exhibits both high activity and durability for ORR, and it also achieves high performance for the assembled ZABs. More importantly, DFT calculations provide theoretical insights into the AISC ORR mechanism in FeN4-based SACs and demonstrate that the Fe-pd-N4 sites of the SAC-Fe/NC play a major role in the high ORR performance. The findings open new paths for the rational design of SACs from non-noble bulk metal salts to isolated single-metal atoms, which possess great opportunities for the practical application of SACs in ZABs and beyond.Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Xinlong Tian (tianxl@hainanu.edu.cn).This study did not generate new unique materials.In a normal procedure, the FeCl2 (50\u00a0mg) and powder of NC (100\u00a0mg) were separately placed on the porcelain boat. The porcelain boat was placed in the plasma-enhanced chemical vapor deposition (PECVD), and then started on the tube furnace of the PECVD with the following parameters: the temperature was 400\u00b0C, holding time was 60\u00a0min, and under the N2 flowing. When the temperature reached 400\u00b0C, we turned on the PECVD with the following parameters: radio frequency power was 500 W, the processing time was 40\u00a0min, the tube pressure was 50 Pa, and under the N2 flowing. After the temperature of the instrument dropped to room temperature, the SAC-Fe/NC was obtained.TEM, HRTEM, and HAADF-STEM EDX images were performed by using a Thermo Scientific Talos F200X G2 operated at 200 keV. AC HAADF-STEM image was obtained on an aberration-corrected FEI Titan G2 60\u2013300 field-emission TEM (FEI,\u00a0USA), operated at 300 keV (\u03b1max\u00a0= \u223c100 mrad). XRD was conducted on HAOYUAN powder diffractometer (DX-2700BH). XPS data were collected using a Thermo Scientific NEXSA X-ray photoelectron spectrometer with a monochromatized Al Ka X-ray source (1,486.6 eV). Raman spectra were collected using a Horiba Scientific with 532-nm laser excitation. The actual loading of the metals was determined by using ICP-OES (Aglient 5110). The X-ray absorption spectra including XANES and EXAFS of the sample at Fe K-edge were measured at the beamline 1W1B of the Beijing Synchrotron Radiation Facility (BSRF, Beijing).All ORR performance tests were conducted on a Gamry 1010E electrocatalytic station with a three-electrode electrochemical cell, in which Hg/HgO was applied as the reference electrode, and a graphite rod was used as the counter electrode. A glass carbon electrode (GCE, diameter 5\u00a0mm) was applied as the working electrode. The catalysis ink was prepared via ultrasonically 5-mg catalysts dispersed into 1\u00a0mL ethanol including 30\u00a0\u03bcL Nafion (5 wt %). After that, we dropped 7\u00a0\u03bcL of the catalysis ink onto the GCE to obtain the working electrode. The LSV curves were recorded at a rotating speed of 1,600 revolutions per minute (rpm) with a scan rate of 5\u00a0mV s\u22121 in O2-saturated 0.1\u00a0M KOH. The ADT method was applied to test the durability, with 10,000 cycles of potential cycling from 0.6 to 1.0\u00a0V at 100\u00a0mV s\u22121, and then we recorded the LSV curves.For the rotating ring-disk electrode tests, the Pt ring electrode was biased at 1.2\u00a0V versus reversible hydrogen electrode (RHE), leading to the electro-oxidation of H2O2, which occurred during the ORR process. The H2O2 yield and n per oxygen molecule were calculated by the following equations:\n\n(Equation\u00a01)\n\n\n%\n\nH\n2\n\n\nO\n2\n\n=\n200\n\n\n\nI\nR\n\n/\nN\n\n\n\n\nI\nD\n\n+\n\nI\nR\n\n\n/\nN\n\n\n\n\n\n\n\n\n(Equation\u00a02)\n\n\nn\n=\n4\n\n\nI\nD\n\n\n\nI\nD\n\n+\n\n\nI\nR\n\n/\nN\n\n\n\n\n\n\nwhere ID and IR are the disk and ring currents, respectively. N is the ring current collection efficiency (37%).All potentials in this work are quoted with respect to an RHE.In the homemade ZAB, the polished zinc plate was used as the anode electrode, catalyst loaded on carbon paper (catalysts loading: 1\u00a0mg cm\u22122) was applied as the air-cathode electrode, and 6\u00a0M KOH solution was used as electrolyte. The polarization curves of the assembled ZAB were recorded via a Gamry 1010E electrochemical workstation. The discharge polarization curve was carried out on a LANHE (CT2001A) battery testing system.Spin-polarized DFT calculations were performed using the Perdew-Burke-Ernzerhof (PBE)\n41\n functional and the projector augmented wave (PAW)\n42\n\n,\n\n43\n potential as implemented in the Vienna Ab Initio Simulation Package (VASP).\n44\n\n,\n\n45\n The Ueff\u00a0= U \u2013 J\u00a0= 3 eV\n46\n was applied to Fe\u2019SD orbitals. An energy cutoff of 500 eV and a convergence criterion of 10\u22125 eV for self-consistent calculations was adopted. All structures were fully relaxed until the total force on each atom was less than 0.05 eV/\u00c5. The solvent effect was included by using the implicit solvation model as implemented in the VASPsol code.\n47\n\n,\n\n48\n The van der Waals interaction was corrected based on the DFT-D3 scheme.\n49\n The thickness of the vacuum layer was larger than 15\u00a0\u00c5. A 6\u00a0\u00d7\u00a06\u00a0\u00d7\u00a01 graphene supercell was used to model the Fe-pd-N4. The Fe-po-N4 model was derived from the pyrrole-type FeN4 model.\n30\n A \u0393-centered k-point with a resolution less than 0.03\u00a0\u00d7\u00a02\u03c0/\u00c5 was used. VASPKIT code\n50\n and VESTA software\n51\n were used for calculation pre-processing and post-processing.The CHE model\n52\n was used in our calculations. The Gibbs free energy of molecules and ORR-related absorbates was calculated by G\u00a0= EDFT\u00a0+ ZPE \u2013 TS, where EDFT, ZPE, and S were the DFT energy, zero-point energy, and entropy, respectively, and temperature T was adopted as 298.15 K. The ORR involves four four-electron pathways on the active sites. The theoretical overpotential at equilibrium potential was determined according to \u03b7\u00a0= 1.23 \u2013 |\u25b5Gmax/e\u2212|, where \u25b5Gmax was the maximum free energy change of adjacent electronic steps.This work was supported by the Hainan Province Science and Technology Special Fund (ZDYF2020037, 2020207), the National Natural Science Foundation of China (21805104, 22109034, 22109035, 52164028, 62105083), the Start-up Research Foundation of Hainan University (KYQD(ZR)- 20008, 20082, 20083, 20084, 21065, 21124, 21125).Conceptualization, P.R. and X.-L.T.; methodology, P.R., D.-X.W., and J.-M.L.; investigation, P.R., D.-X.W., J.L., P.-L.D., Y.-J.S., and X.-L.T.; writing - original draft, P.R. and X.-L.T.; writing - review & editing, P.R., D.-X.W., and X.-L.T.; funding acquisition, D.-X.W., J.L., J.-M.L., P.-L.D., Y.-J.S., and X.-L.T.; supervision, X.-L.T.The authors declare no competing interests.Supplemental information can be found online at https://doi.org/10.1016/j.xcrp.2022.100880.\n\n\nDocument S1. Figures\u00a0S1\u2013S11, Tables\u00a0S1\u2013S8, and supplemental experimental procedures\n\n\n\n\n\nDocument S2. Article plus supplemental information\n\n\n\n", "descript": "\n                  Single-atom catalysts (SACs) have attracted tremendous attention owing to their unique intrinsic properties, while the facile synthesis, especially the direct transformation of metal salts into single-metal atoms, is still a huge challenge. Here we report a practical approach to access the large-scale synthesis of SACs, in which the metal salts are excited and stripped as mobile single-metal atoms flow under the \u201cplasma bombing\u201d treatment. Simultaneously, they are trapped and anchored by the defective and nitrogenous sites of the supports. More importantly, the synthesis approach is quite general, and it can be extended to the fabrication of various SACs. As an illustration, the prepared SAC-Fe/NC delivers a remarkable oxygen reduction reaction (ORR) performance. In combination with the theoretical analysis, an adsorbent-induced spin-crossover ORR mechanism is proposed, and the Fe species coordination with four pyridinic-type N atoms is demonstrated as the main contributor to the high performance of the prepared SAC-Fe/NC.\n               "}
{"full_text": "Non-edible lignocellulose is the most abundant, cheapest and fastest growing sustainable biomass resource, composed of three primary biopolymers: cellulose (a polymer of glucose), hemicellulose (a polymer mainly of pentoses) and lignin (a highly cross-linked polymer of substituted phenols) [1]. In order to produce value-added bio-products which could displace petroleum feedstocks, lignocellulose must first be transformed into simpler and more easily processed platform chemicals. This approach, similar to that used in conventional petroleum refineries, would allow the simultaneous production of biofuels and biochemicals in an integrated facility, a biorefinery [2].In 2004, the U.S. Department of Energy (DOE) [3] released a report, later revised by Bozell et al. [4], identifying the top value-added platform chemicals in a future biorefinery. HMF was identified as one of the most appealing and promising building block molecules. This furan derivative can be produced from agricultural waste and forest residue such as polysaccharides (i.e. cellulose and hemicellulose) by acid-catalysed hydrolysis to C6 monosaccharides, followed by dehydration [5]. In contrast to most petrochemical products, HMF is an oxygen-rich, functionalized compound. Its conversion to value-added chemicals usually involves several chemical transformations (e.g. hydrogenation, dehydration, hydrogenolysis, oxidation, etc.) which are promoted by multifunctional heterogeneous catalysts [6,7]. The development of related catalytic heterogeneous processes has become highly topical to produce valuable bioproducts such as: tetrahydrofuran 2,5-diyldimethanol (THFDM) [8\u201310], 2,5-dimethylfuran (DMF) [11], 2,5-furandicarboxylic acid (FDCA) [12,13], C6 linear alcohols [14,15], 3-hydroxymethylcyclopentanone (HCPN) and 3-hydroxymethylcyclopentanol (HCPL) [16,17] (Fig. 1\n).One much less studied reaction is the conversion of HMF into the linear diketone derivatives 1-hydroxyhexane-2,5-dione (HHD) and 2,5-hexanedione (HXD). Although a large-scale synthetic route to HHD is currently not available, the presence of a hydroxymethyl functionality offers opportunities for the synthesis of valuable chemicals, as recently highlighted [18\u201320]. HXD is employed as a solvent and as an intermediate for the synthesis of polymers, amines and surfactants [21,22]. HHD can be produced from HMF in water under H2 pressure via the metal-catalysed selective hydrogenation of the carbonyl group to furan-2,5-diyldimethanol (FDM), followed by the acid-catalysed ring-opening of the FDM unsaturated ring [23]. Notably, this pathway does not involve the formation of a more stable saturated tetrahydrofuran ring, whose ring-opening requires harsh reaction conditions (i.e. T\u202f>\u202f140\u202f\u00b0C; P\u202f>\u202f60\u202fbar) [24]. HXD can also be produced from FDM via hydrogenolysis to DMF, followed by hydrolytic ring-opening of the latter or by HHD via scission of the hydroxyl group (Fig. 2\n) [22,25].The formation of HHD from HMF was firstly reported in 1991 by Schiavo et al. using Pd/C in an aqueous solution of oxalic acid (pH\u202f=\u202f2) at 70\u202fbar H2 and 140\u202f\u00b0C [23]. In 2009, Luijkx et al. reported a similar process using HCl [26]. In 2014, Liu et al. developed two binary catalytic systems using Pd/C either in CO2/H2O (forming carbonic acid) [27], or in THF with co-added Amberlyst-15 [28], affording in both cases 77% yield of HHD (see Table SI1). Recently we reported that HHD is formed as a low yield intermediate (<7%) in the conversion of HMF to HCPN and HCPL over M-Al2O3 catalysts in H2O (M\u202f=\u202fCo, Ni, Cu). HHD was rapidly converted to HCPN via an aldol condensation reaction, catalysed by basic sites, followed by hydrogenation [17].The requirement of hydrogenating metal phases and acidic sites for the production of HHD prompted us to investigate the deposition of various transition metals over zeolite supports to prepare easily tuneable bifunctional catalysts. Among the most investigated zeolites, Beta (with BEA topology) exhibits excellent properties to the aimed transformation due to its high hydrothermal stability, large specific surface area (>600\u202fm2\u202fg\u22121), 3D large-pore channel system (5.5\u20137.6\u202f\u00c5) and dual Lewis/Br\u00f8nsted acidity [29]. Beta zeolite-based catalysts have been used for the conversion of furfural into levulinic acid [30,31] and for the hydrodeoxygenation of furoins into alkanes [32]. However, to the best of our knowledge, there are no studies reporting the formation of linear diketone derivatives using Beta zeolite-supported catalysts.Herein, we present the catalytic production of the diketone derivatives HHD and HXD from HMF by zeolite-supported transition metals in H2O. A series of transition metal-loaded (M) Beta zeolites were prepared (M\u202f=\u202fCo, Ni, Cu, Ru, Pd), characterised and tested in a batch stirred reactor under H2 pressure with Pd showing the highest catalytic activity. Consequently, the effects of Pd particle size, zeolite Si/Al ratio and reaction conditions were investigated, and catalyst stability and recyclability were evaluated. This work establishes Beta zeolite-supported Pd catalysts as promising candidates for the upgrading of HMF into valuable biomass-derived linear diketone derivatives by demonstrating for the first time the conversion of HMF to HHD and HXD in water by a solid state, bifunctional (no acid co-added) catalyst.Metal-loaded zeolites were prepared by incipient wetness impregnation (IWI) of commercially available Beta (Si/Al\u202f=\u202f12.5) and ZSM-5 (Si/Al\u202f=\u202f11.5) zeolites purchased from Zeolyst Int. The corresponding aqueous solutions of the metal precursors were prepared using: PdCl2 (Sigma-Aldrich), Pd(NO3)2\u00b72H2O (Aldrich), RuCl3 (Aldrich), NiCl2\u00b76H2O (Aldrich), CuCl2\u00b72H2O (Aldrich) and CoCl2\u00b76H2O (Fluka). Prior to impregnation, the parent NH4-zeolites were calcined at 550\u202f\u00b0C (heating rate of 2\u202f\u00b0C\u202fmin\u22121) in static air for 5\u202fh, producing the respective H-zeolites. The deposition of the metal was carried out by adding dropwise the aqueous solution of the precursor to the zeolite support at room temperature (3\u202fwt% for the Pd and Ru samples; 10\u202fwt% for the Cu, Ni and Co samples). After impregnation, the catalysts were dried in a rotary evaporator at 65\u202f\u00b0C under vacuum for 1\u202fh. Subsequently, the dried samples were calcined in air at 500\u202f\u00b0C for 5\u202fh (heating rate of 2\u202f\u00b0C\u202fmin\u22121). The reduction treatment was performed under pure H2 flow (100\u202fcm3\u202fmin\u22121) for 5\u202fh at 200\u202f\u00b0C (for the Pd and Ru samples), 300\u202f\u00b0C (Cu sample) and 500\u202f\u00b0C (Ni and Co samples) with a heating rate of 2\u202f\u00b0C\u202fmin\u22121. Finally, the catalysts were passivated under a flow of 1% v/v O2/N2 (100\u202fcm3\u202fmin\u22121) for 2\u202fh at room temperature.Pd-loaded Beta zeolite was also prepared by deposition-coprecipitation (DP-CP) using the urea-based method developed by Geus et al. [33]. First, 2\u202fg of the calcined Beta zeolite were placed in a 250\u202fml round-bottom flask. Then, an aqueous solution (100\u202fml) containing PdCl2 (0.005\u202fM) and urea (1.2\u202fM, Sigma) was added dropwise with constant stirring (550\u202frpm) at room temperature. The suspension (pH\u202f=\u202f4\u20134.5) was heated to 95\u202f\u00b0C to initiate urea hydrolysis. After 3\u202fh, the pH of the suspension remained stable at pH\u202f\u2248\u202f7.5. The solution was cooled to room temperature and the precipitate was collected by filtration, washed with deionized water, dried at 110\u202f\u00b0C overnight and subsequently calcined in air at 500\u202f\u00b0C for 5\u202fh (heating rate of 2\u202f\u00b0C\u202fmin\u22121). The calcined sample (catalyst precursor) was then reduced under pure H2 flow (100\u202fcm3\u202fmin\u22121) at 200\u202f\u00b0C for 5\u202fh (heating rate of 2\u202f\u00b0C\u202fmin\u22121). Finally, the reduced catalyst was passivated under a flow of 1% v/v O2/N2 (100\u202fcm3\u202fmin\u22121) for 2\u202fh at room temperature. Hereafter, the sample prepared by the DP-CP urea method will be referred as Pd(u)/Beta.Partial dealumination of the calcined Beta zeolite (Si/Al\u202f=\u202f12.5) was carried out by acid treatment using HNO3 aqueous solutions of different concentration (0.1, 0.5, 2 and 5\u202fM) at room temperature for 1\u202fh (20\u202fmL\u202fg\u22121 zeolite). After filtration and washing with deionized water, the materials were dried overnight (110\u202f\u00b0C) and calcined in static air at 500\u202f\u00b0C for 5\u202fh (heating rate of 2\u202f\u00b0C\u202fmin\u22121). Afterwards, the obtained dealuminated zeolites were impregnated with Pd following the DP-CP urea method described above. Hereafter, the four dealuminated and impregnated samples will be abbreviated as Pd(u)/Beta-dAlx (x\u202f=\u202f1\u20134), where x\u202f=\u202f4 refers to the sample showing the highest degree of dealumination (higher Si/Al ratio).The prepared catalysts were characterised by powder X-Ray diffraction (PXRD) on a Panalytical X\u2019Pert Pro diffractometer with Co K\u03b11 radiation (\u03bb\u202f=\u202f1.7890\u202f\u00c5) in the 2\u03b8 angle range 10\u221280\u00b0 (scanning speed of 0.023\u00b0\u202fs\u22121). Metal content of the catalysts was determined by inductively coupled plasma - optical emission spectroscopy (ICP-OES) using an Agilent 5110 SVDV instrument. The samples were digested in a strong acidic medium (10\u202fml HCl and 20\u202fml HNO3) and then diluted with water (1:10 v/v). Textural properties were evaluated through N2 adsorption-desorption isotherms at 77\u202fK, using a Micromeritics TRISTAR II instrument. Prior to the measurement, the samples were outgassed under vacuum at 120\u202f\u00b0C for 20\u202fh. The BET equation was used for specific surface area calculation, whereas pore volume was determined by the BJH method.Acidity of the catalysts was determined by temperature programmed desorption of ammonia (NH3-TPD) in a Quantachrome ChemBET 3000 unit. Firstly, the samples were outgassed under a He stream (100\u202fcm3\u202fmin\u22121) heating at 10\u202f\u00b0C\u202fmin\u22121 up to 350\u202f\u00b0C. Afterwards, the samples were cooled to 150\u202f\u00b0C and saturated under an ammonia stream (100\u202fcm3\u202fmin\u22121) for 10\u202fmin. Subsequently, the physically adsorbed ammonia was removed by flowing helium (100\u202fcm3\u202fmin\u22121) for 30\u202fmin at 150\u202f\u00b0C. Finally, the chemically adsorbed ammonia was desorbed by heating to 650\u202f\u00b0C with a rate of 10\u202f\u00b0C\u202fmin\u22121 under He flow (100\u202fcm3\u202fmin\u22121). Ammonia concentration was monitored continuously using a thermal conductivity detector (TCD).Thermogravimetric analysis (TGA) was carried out on a Q600 TA Instrument; ca. 5\u202fmg of sample were loaded into an alumina microcrucible and heated to 800\u202f\u00b0C at 10\u202f\u00b0C\u202fmin\u22121 under a flow of air (100\u202fcm3\u202fmin\u22121). Elemental analysis (C and H content) of the used catalysts was carried out on a Thermo EA1112 Flash CHNS Analyser. TEM images were obtained with a JEOL 2100 transmission electron microscope operating at 200\u202fkV. The samples were dispersed in acetone, stirred in an ultrasonic bath and deposited on a carbon-coated Cu grid. SEM imaging and energy-dispersive X-ray (EDX) spectroscopy were run on a Hitachi S-4800 Field-Emission scanning electron microscope.Solid state 27Al NMR experiments were performed on a 9.4\u202fT Bruker DSX 400\u202fMHz spectrometer using a Bruker Triple Resonance 4\u202fmm HXY (in double resonance mode) probe under Magic Angle Spinning (MAS) at a rotational rate of 10\u202fkHz. One-dimensional MAS NMR spectra were recorded using a rotor-synchronized (1 period) Hahn echo sequence with a radio frequency pulse of 50\u202fkHz (\u03c0/2 pulse of 1.7\u202f\u03bcs duration) and a quantitative recycle delay of 1\u202fs. Whilst the quantitative interpretation of 27Al MAS NMR data has to be performed with caution due to non-uniform excitation of sites with different magnitudes of the quadrupolar coupling constants [34], the similar values observed for tetrahedral and octahedral sites (i.e. 1\u20132\u202fMHz) allow for an estimation of their ratio [35,36].The performance of the catalysts was studied in high pressure 100\u202fml batch stirred reactors (Parr Instrument Co.) A glass liner was loaded with 45\u202fml of an aqueous solution of HMF (0.04\u202fM) and 0.06\u202fg of catalyst and placed into the stainless-steel reactor. After sealing the vessel, the reactor was flushed three times with N2 and heated to the required reaction temperature (80\u2013155\u202f\u00b0C). Once the targeted temperature was reached, the vessel was pressurised with H2 to the respective value (5\u201360\u202fbar of H2) and stirring was set to 600\u202frpm. After the end of the reaction (typically 6\u202fh), the identity and distribution of the products were determined by the combination of 1H and 13C NMR spectroscopy (Bruker AVANCE III HD spectrometer), GC-MS (Agilent 6890\u202fN GC with a 5973 MSD detector) and GC (Agilent 7890A GC with an FID). GC and GC-MS were equipped with a DB-WAXetr capillary column (60\u202fm, 0.25\u202fmm i.d., 0.25\u202f\u03bcm). Standard reference compounds used: HMF (Sigma), FDM (Manchester Organics), THFDM (Ambinter) and HXD (Sigma-Aldrich). Details regarding calculations of conversion, yield and selectivity are provided in the Supporting Information (SI).PXRD patterns (2\u03b8\u202f=\u202f10\u201380\u00b0) of the Beta-supported metal catalysts after reduction show the characteristic peaks of the corresponding metallic phase (Fig. SI1). No crystalline phases of the metal oxides precursors were observed, confirming their complete reduction under the H2 treatment. Well-defined reflections associated with the zeolitic structures (Beta or ZSM-5) were identified [37], verifying that crystallinity of the zeolitic support was preserved after impregnation. The composition of the prepared catalysts was determined by ICP-OES (Table 1\n), showing metal contents close to the corresponding nominal values (Pd, Ru\u202f=\u202f2.7\u20132.9\u202fwt%; Ni, Cu, Co\u202f=\u202f8.7\u20139.3\u202fwt%).The comparison of the catalytic performance of several zeolite-supported metal catalysts, prepared by IWI, in the conversion of HMF is presented in Table 1 (110\u202f\u00b0C, 20\u202fbar H2). Temperature was set at 110\u202f\u00b0C in order to minimize the extent of oligomerisation reactions which are favoured by the presence of acidic sites [24].A preliminary control reaction with non-impregnated Beta zeolite (entry 1) showed negligible HMF conversion (4%) to FDM (1% yield), verifying that a reduced metal phase is essential for the conversion of HMF to the targeted diketone derivatives. An additional control experiment using the more reactive FDM intermediate as the substrate over Beta zeolite (entry 2) resulted in 10% FDM conversion with a concomitant colour change of the reaction mixture from pale to dark yellow. However, no products were detected by GC and carbon mass balance (Cmb) was only 90%. The decrease in Cmb suggests that the highly reactive unsaturated intermediates formed by the hydrolytic ring-opening of FDM, such as 1-hydroxyhex-3-ene-2,5-dione (HHED) [27,28], may lead to heavier ill-defined products, such as humins [38,39]. This undesired oligomerisation reaction always takes place in parallel with productive FDM conversion. It should be noted that decarboxylation of HHED to levulinic and formic acid [40] was not observed due to the highly reducing conditions employed.Among the screened active metal phases, the Pd/Beta catalyst (entry 3) afforded the highest HMF conversion (80%) and selectivity to HHD (56%), whereas HXD was also detected as a minor product (8% selectivity). However, Cmb was only 77%, consistent with the formation of undetectable oligomers. The Ru/Beta catalyst (entry 4) showed lower HMF conversion (41%) and HHD selectivity (44%). The non-noble metal based catalysts (i.e. Ni, Cu and Co, entries 5\u20137) showed even lower HMF conversion (15\u201324%) despite having considerably higher metal loadings. Moreover, selectivity to HHD was rather poor (13\u201325%), whereas HXD was not detected. It should be noted that the higher Cmb observed for the Ru, Ni, Cu and Co supported catalysts (84\u201398%) is a direct consequence of the lower HMF conversion. The superior catalytic activity of Pd relative to other transition metals has also been demonstrated for the hydrogenation of HMF to THFDM in water, using Pd/C carbon [41] or Pd@MIL-101(Al)-NH2 MOF [42].The effect of the zeolitic support was also explored by using ZSM-5 (Si/Al\u202f=\u202f11.5) instead of Beta (Si/Al\u202f=\u202f12.5) and preparing Pd/ZSM-5 (entry 8). The latter also showed high HMF conversion (74%) and HHD selectivity (41%), albeit slightly lower than Pd/Beta. Notably, Pd/ZSM-5 afforded the highest selectivity to HXD (14%) which can be attributed to the higher concentration of Br\u00f8nsted acid sites in ZSM-5 [43] and the different structural frameworks (MFI in ZSM-5 vs. BEA in Beta) which affect the shape selectivity by either mass transfer or transition state effects [44,45]. Overall, both zeolite-supported Pd catalysts gave the highest HMF conversion and HHD selectivity but also showed the lowest Cmb due to the formation of heavier undetectable oligomers [27,28].In order to clarify the role of water in the reaction mechanism, an isotopic labelling experiment was performed using D2O as the solvent under the same reaction conditions (110\u202f\u00b0C, 20\u202fbar H2). GC-MS revealed the formation of [D3]-HHD and [D4]-HXD as the main products, as well as traces of [D4]-HCPN. FDM was also detected but it was not deuterated (Fig. SI2). Specifically, higher m/z values were observed in the mass spectra of the products when D2O was employed as the solvent instead of H2O: m/z\u202f=\u202f118 ([D4]-HCPN), 133 ([D3]-HHD) and 118 ([D4]-HXD) compared to m/z\u202f=\u202f114 (HCPN), 130 (HHD) and 114 (HXD). This in turn suggests that two D2O molecules participate in the catalytic mechanism, specifically in the ring-opening of FDM via consecutive hydration-dehydration steps (Fig. 3\n), as originally proposed by Horvat et al. [40]. Importantly, none of the above compounds is formed in water-free reaction mixtures [7]. Therefore, H2O not only serves as an environmentally benign solvent but is also necessary for FDM ring-opening [46].The effect of the Pd particles size in the Beta zeolite-supported catalysts was also investigated. In addition to the catalyst prepared by IWI and PdCl2 (Pd/Beta), two more catalysts were prepared by either (i) IWI and Pd(NO3)2 as the precursor (Pd(n)/Beta) or (ii) DP-CP with urea and PdCl2 (Pd(u)/Beta). Samples were then calcined and reduced as before. The different preparation methods led to different morphologies of the supported Pd nanoparticles (NPs), as deduced by TEM imaging (Fig. 4\n and SI3) and PXRD (Fig. SI4).The Pd/Beta catalyst (Fig. 4a) resulted in intermediate Pd NPs (average diameter of 5.2\u202f\u00b1\u202f3.4\u202fnm). Employment of Pd(NO3)2 as the precursor (Fig. 4b) led to much larger Pd NPs with a significantly less uniform particle size distribution (average diameter of 16.2\u202f\u00b1\u202f10.4\u202fnm). The higher dispersion of Pd catalysts with PdCl2 as the precursor has been ascribed to the formation of complex PdxOyClz species on alumina/aluminosilicate surfaces [47,48]. The DP-CP method (Fig. 4c) resulted in the smallest Pd NPs and the most uniform size distribution (average diameter of 3.5\u202f\u00b1\u202f1.5\u202fnm). The smaller and more uniform particle size observed for Pd(u)/Beta can be associated with the slow and homogenous generation of hydroxide ions through the hydrolysis of urea at 95\u202f\u00b0C which hinders the uneven precipitation of PdII species due to a sudden, local increase of pH [49,50].\nTable 2\n shows the conversion of HMF and the obtained product distribution for the three Beta zeolite-supported Pd catalysts after 6\u202fh under 20\u202fbar of H2 at 110\u202f\u00b0C. All the catalysts have similar Pd contents varying between 2.6 and 2.8\u202fwt% (based on ICP-OES). A direct correlation between higher HMF conversion and smaller Pd particle size was identified. Thus, the Pd(n)/Beta catalyst (dM\u202f=\u202f16.2\u202f\u00b1\u202f10.4\u202fnm) showed the lowest HMF conversion (56%). Moreover, the lower hydrogenation activity of Pd(n)/Beta resulted in the lowest selectivity to HHD (39%) due to a higher degree of oligomerisation of the unsaturated intermediates formed via FDM ring-opening (Fig. 3). On the other hand, the Pd(u)/Beta catalyst (dM\u202f=\u202f3.5\u202f\u00b1\u202f1.5\u202fnm) afforded almost complete HMF conversion (96%) and the highest selectivity to HHD (56%).The N2 adsorption-desorption isotherm of the Pd(u)/Beta catalyst was similar to the parent Beta zeolite (Fig. SI5), presenting features of Type I isotherms. The resulting textural properties showed a slight decrease in specific surface area and pore volume after the incorporation of Pd (621\u202fm2\u202fg\u22121 and 0.297\u202fcm3\u202fg\u22121\nvs. 574\u202fm2\u202fg\u22121 and 0.286\u202fcm3\u202fg\u22121) due to partial blockage of the zeolite pores by Pd NPs, characteristic of a highly dispersed metal phase [51]. Thus, the enhanced hydrogenation activity with the decrease of Pd particle size can be attributed to the corresponding higher metal dispersion (higher metal active surface) [52] and to the higher uniformity in Pd particle size which favours the adsorption and hydrogenation of furanic compounds [41,42]. Additionally, well dispersed and uniform Pd NPs increase catalytic lifetime by hindering leaching and sintering of particles [53]. In this sense, SEM-EDX analysis of the Pd(u)/beta catalyst (Fig. SI6) confirmed the absence of residual chlorine (from the PdCl2 precursor) which is known to increase metal atom mobility and cause sintering.The product distribution obtained over the Pd(u)/Beta catalyst (Table 2) revealed that at 96% HMF conversion no product exceeded 1% selectivity apart from the targeted diketone derivatives HHD and HXD. Notably, full conversion was achieved in 24\u202fh and only two well-defined peaks corresponding to HXD and HHD were observed in the respective GC chromatogram of the product mixture (Fig. SI7). However, Cmb was found lower than 80% for all runs. Taken together, these observations suggest that an undetectable by GC fraction of products is produced via an acid-catalysed oligomerisation of unsaturated intermediates, formed via FDM ring-opening [24,27]. The formation of oligomers could be potentially suppressed by co-addition of organic solvents [54]. Alternatively, techniques such as biphasic reactive extraction, adsorbent-based separation or reactive distillation could be applied to separate the oligomer fraction from the targeted compounds [55,56].The time evolution of HMF conversion, product distribution and Cmb over the Pd(u)/Beta catalyst are depicted in Fig. 5\n. HMF was swiftly consumed, reaching 87% conversion in 120\u202fmin and 96% in 360\u202fmin. FDM was detected at early reaction times, with a maximum yield of 4% at 20\u202fmin but was fully consumed in 240\u202fmin. HHD and HXD yields rapidly increased during the first 120\u202fmin (47% and 7%, respectively) but did not significantly change afterwards, achieving final values of 54% (HHD) and 9% (HXD). Notably, exposing a mixture of HHD and HXD to a fresh batch of Pd(u)/Beta with or without H2 (T\u202f=\u202f110\u202f\u00b0C, 3\u202fh) showed no interconversion between HHD and HXD. A selectivity vs. conversion plot (Fig. SI8) is consistent with HHD and HXD being formed via two separate pathways (Fig. 2). Previous works have shown that hydrogenolysis of FDM to DMF and hydrolysis of the latter can lead to HXD [21,25]. However, we did not detect any trace of DMF, indicating that HXD is formed through a currently unidentified mechanism.The observed time profile supports the proposed reaction mechanism for the conversion of HMF into HHD (Fig. 3) which begins with the hydrogenation of the HMF carbonyl group to form FDM, followed by the hydrolytic furan ring-opening and hydrogenation to form HHD [10,17,20,28]. The intermediate nature of FDM was confirmed by a separate experiment using a lower amount of catalyst (40\u202fmg instead of 60\u202fmg); FDM yield reached a maximum of 23% before gradually decreasing as HHD was being formed (Fig. SI9). The time evolution of the Cmb showed a pronounced decrease within 120\u202fmin (from 100% to 73%) but remained practically constant afterwards (71% in 360\u202fmin). This is consistent with the formation of undetected heavier oligomers, catalysed by the zeolite acid sites.Since the catalytic properties of the Pd(u)/Beta catalysts are directly related to the acidity of the zeolite framework, it was anticipated that the removal of Al atoms would affect catalyst activity and selectivity [57]. In order to understand the influence of the Si/Al ratio on the two competitive acid-catalysed reaction pathways, i.e. FDM ring-opening and oligomerisation of unsaturated intermediates, four Beta zeolite-supported Pd catalysts with different Si/Al ratio were prepared (Pd(u)/Beta-dAlx, x \u202f=\u202f1\u20134) via dealumination (acid treatment) and subsequent Pd impregnation (DP-CP with PdCl2). The parent Beta zeolite (Si/Al\u202f=\u202f12.5) was partially dealuminated by using gradually more concentrated HNO3 aqueous solutions of 0.1, 0.5, 2 and 5\u202fM which led to increasingly higher Si/Al atomic ratios in the range of 17.8\u201334.5 (Table 3\n).Crystallinity and microporosity of the zeolitic support were preserved after the dealumination treatment based on the respective PXRD patterns and N2 isotherm profiles (Fig. SI10). Removal of Al resulted in a progressive decrease of acid sites according to NH3-TPD measurements (Fig. SI11 and Table SI2), due to removal of: (i) extra-framework aluminium, associated to Lewis acidity and (ii) tetrahedrally coordinated aluminium, associated to Br\u00f8nsted acidity [58,59]. Important differences were observed in the catalytic performance of the Pd(u)/Beta-dAlx catalysts (Table 3). A non-negligible increase of Cmb was observed over the dealuminated supports (from 71% for Pd(u)/Beta to 79\u201382% for Pd(u)/Beta-dAlx). However, a gradual decrease in HMF conversion and selectivity to HHD was also observed as the Si/Al ratio was increased due to the lower extent of the hydrolytic FDM ring-opening, catalysed by Br\u00f8nsted acid sites.The effect of reaction temperature (80\u2013155\u202f\u00b0C) and H2 pressure (5\u201360\u202fbar) on the production of diketone derivatives from HMF over the Pd(u)/Beta catalyst was also investigated (Table 4\n). Comparison with the original run (entry 1) revealed that lowering the temperature below 110\u202f\u00b0C resulted in lower HMF conversion (entries 2\u20133). Raising the temperature to 125\u202f\u00b0C or 140\u202f\u00b0C (entries 4\u20135) restored HMF conversion (\u226595%). Notably, HCPN was also detected as a minor product (4\u20135% selectivity) as the temperature increased due to promotion of FDM ring-rearrangement, resulting in a slight improvement of Cmb (from 71% at 110\u202f\u00b0C to 78% at 140\u202f\u00b0C). However, the combined selectivity of targeted HHD and HXD was practically not affected (65\u202f\u00b1\u202f1%), although a marginal shift towards HXD was observed (56% HHD and 9% HXD at 110\u202f\u00b0C vs. 53% HHD and 13% HXD at 140\u202f\u00b0C). Further increasing the temperature to 155\u202f\u00b0C (entry 6) led to a significant decrease of HHD selectivity, mainly due to oligomerisation (Cmb\u202f=\u202f65%).Having established that T\u202f=\u202f110\u202f\u00b0C is the optimal temperature for the production of the targeted compounds, the effect of H2 pressure was explored. Running the reaction under 40\u202fbar of H2 afforded again 96% HMF conversion, albeit with a slightly higher Cmb (77%, entry 9). Increasing the H2 pressure to 60\u202fbar resulted in 98% HMF conversion with 66% HHD selectivity in 6\u202fh and 100% conversion with 68% HHD selectivity in 24\u202fh (entries 10\u201311). Cmb increased to 82%, consistent with a faster hydrogenation rate of the unsaturated intermediates (vs. oligomerisation) to form HHD.The stability of the Pd(u)/Beta catalyst was examined by testing the catalytic activity of the supernatant after physically separating the catalyst from the reaction mixture. HMF conversion did not increase any further (\u224850% conversion at 110\u202f\u00b0C and 20\u202fbar H2) and product distribution did not change once the catalyst was filtered off after 40\u202fmin (Fig. SI12). Likewise, Pd concentration in the supernatant was less than 0.2\u202fppm (<0.5% of the total Pd content), according to ICP-OES. Both results verify that Pd does not leach into the solution phase and confirm the heterogeneous nature of the catalytic system.The reusability of the Pd(u)/Beta catalyst was investigated by evaluating its catalytic activity upon consecutive runs (110\u202f\u00b0C, 20\u202fbar H2). The used catalyst was recovered after each run by filtration at room temperature, washed with deionized water and dried at 25\u202f\u00b0C overnight. Fig. 6\na depicts the TGA curves of the fresh and the used Pd(u)/Beta catalysts, showing a noticeable increase in the total weight loss for the used catalyst (28% vs. 7%). Furthermore, the elemental microanalysis of the used catalyst (Fig. 6a inset) revealed a significant carbon content (9.32% weight), consistent with deposition of organic compounds on the catalyst\u2019s surface during turnover. This result compensates to a certain extent (5\u20136%) for the lower Cmb observed. The PXRD pattern of the used Pd(u)/Beta catalyst (Fig. 6b) showed the expected reflections of the Pd metallic phase, indicating that Pd remains reduced after turnover. However, TEM images of the used catalyst (Fig. 6c and SI13) revealed an increase in Pd particle size (dM\u202f=\u202f7.6\u202f\u00b1\u202f2.3\u202fnm, Fig. 6d) compared to the fresh catalyst (dM\u202f=\u202f3.5\u202f\u00b1\u202f1.5\u202fnm, Fig. 4c), indicative of aggregation and formation of larger Pd particles [60].Recycling tests were conducted after recovering the Pd(u)/Beta catalyst (Fig. 7\na). HMF conversion gradually decreased from 96% (1st run) to 10% (4th run) due to deposition of organic compounds and aggregation of Pd NPs (Fig. 6 and Fig. SI13). In order to restore the catalytic activity, the used Pd(u)/Beta catalyst (after 4 consecutive runs) was subjected to a regular regeneration treatment: calcination (air/500\u202f\u00b0C/5h) and reduction (H2/200\u202f\u00b0C/5\u202fh). Catalytic activity was partially restored (5th run), affording 90% HMF conversion to HHD (40% yield) and HXD (4% yield). A non-negligible amount of FDM (7% yield) and HCPN (12% yield) was also detected in the product mixture.In order to investigate the observed change in selectivity, a separate set of experiments was conducted during which the catalyst was regenerated at the end of each run (Fig. 7b). HMF conversion gradually decreased from 96% (1st run) to 56% (3rd run) with a concomitant increase in FDM yield (6% in 3rd run), as also observed for the dealuminated samples (Table 3). Measurement of the 27Al MAS NMR spectra of the fresh and the regenerated Pd(u)/Beta catalyst (Fig. SI14) revealed a decrease in the fraction of tetrahedrally coordinated Al after turnover and regeneration. This in turn suggests a lower number of Br\u00f8nsted acid sites [61] which promote FDM ring-opening. Moreover, PXRD verified an increase in Pd particle size after regeneration (Fig. SI15), consistent with a lower hydrogenation activity. Therefore, the observed differences in activity and selectivity during recycling can be ascribed to aggregation of Pd NPs and partial loss of Br\u00f8nsted acidity [61\u201365].Diketone derivatives such as HHD and HXD were produced from HMF over a bifunctional Beta zeolite-supported Pd catalyst in water under relatively mild reaction conditions. The DP-CP method afforded the most active catalyst, compared to IWI, due to smaller and more uniformly dispersed Pd particles among the zeolitic support. Complete conversion of HMF was achieved at 110\u202f\u00b0C and 60\u202fbar of H2 with 68% selectivity to HHD. Leaching of Pd was not observed and catalytic activity could be partially restored after a simple regeneration step. Selectivity to HHD is mainly limited by the formation of heavier ill-defined oligomers. The key distinguishing feature of this study is the synergic effect of the zeolite acid sites and the highly active hydrogenating Pd metallic phase which promotes the hydrolytic ring-opening and subsequent hydrogenation of the FDM intermediate without necessitating co-addition of an acid or use of an organic solvent.This work was supported by the Engineering and Physical Sciences Research Council (EPSRC), UK (EP/K014749).Supplementary data to this article can be found online at https://doi.org/10.1016/j.jcat.2019.04.038.The following are the Supplementary data to this article:\n\nSupplementary Data 1\n\n\n\n", "descript": "\n                  Conversion of 5-hydroxymethylfurfural (HMF) in water to the linear diketone derivatives 1-hydroxyhexane-2,5-dione (HHD) and 2,5-hexanedione (HXD) was investigated over a series of Beta zeolite-supported transition metal catalysts (Co, Ni, Cu, Ru, Pd). Their catalytic performance was tested in a batch stirred reactor (T\u202f=\u202f110\u202f\u00b0C, PH2\u202f=\u202f20\u202fbar) with Pd showing the highest activity and selectivity to HHD and HXD. The effects of Pd particle size, zeolite Si/Al ratio and reaction conditions (T\u202f=\u202f80\u2013155\u202f\u00b0C, PH2\u202f=\u202f5\u201360\u202fbar) were also investigated. The incorporation of Pd into Beta zeolite by the deposition-coprecipitation method produced the most efficient catalyst, affording complete HMF conversion (T\u202f=\u202f110\u202f\u00b0C, PH2\u202f=\u202f60\u202fbar) predominantly to HHD (68% selectivity) and HXD (8% selectivity). The combination of a bifunctional acid/redox solid catalyst and water enhances the hydrolytic ring-opening and subsequent hydrogenation of the furan ring. Catalytic activity can be partially restored by a simple regeneration treatment. This work establishes a catalytic route to produce valuable diketone derivatives from renewable furanic platform sources in water.\n               "}
{"full_text": "\n\n\nNo data was used for the research described in the article.\n\n\nNo data was used for the research described in the article.With the development and progress of industry, air pollution is becoming more and more serious. The various kinds of pollution accumulated in the atmosphere cause harm to plants, river soils, building materials and human health that cannot be ignored [1]. Therefore, a series of measures such as emission reduction, atmospheric governance and resource utilization are being actively launched [2]. An end to pollution control (especially denitrification of flue gas) is urgent, in the terminal treatment of atmospheric nitrogen oxides technology, the most direct and efficient method is to purify pollutants by selective catalytic reduction with ammonia (NH3-SCR), during which, as the core of catalytic reaction, the preparation and configuration of catalysts have become the focus of the field of atmospheric catalysis [3]. At present in denitration management, common catalysts include vanadium, manganese, cerium, titanium and molecular sieve materials. MnOx has rich variable valence states, large specific surface area, high chemical adsorption oxygen content on the surface, such catalyst has rich surface active sites, high crystallinity and rich structural morphology [4,5]. Therefore, scholars gradually began to study MnOx catalysts to improve the NH3-SCR reaction activity at low temperatures. Qiu et\u00a0al. [6]. prepared ordered mesoporous material MnCo2O4 by nano-casting method with good low temperature SCR activity and N2 selectivity, and showed high anti-poisoning performance with the NOx conversion rate maintained at about 80% after 10 h reaction. The NOx conversion rate was maintained at about 80% after 10 h reaction. Chen et\u00a0al. [7]. prepared a novel MnCoOx sphere catalyst, which not only exhibited high low-temperature activity for the NH3-SCR of NOx but also significantly enhanced SO2 resistance. Over MnCoOx sphere catalyst the formation of MnSO4 is significantly inhibited in the presence of SO2, so that the reaction over the SO2-poisoned Mn(5)Co(5)Ox catalyst can still proceed by the LH mechanism and maintain its high catalytic performance.However, the conventional single catalyst is easily deactivated by the erosion of H2O and SO2 in the flue gas [8\u201310]. The catalyst will be inactivated for a long time, so it has been affected in the industrial development [11,12]. Various metals (Cu, Co, Cr, Ni, Fe, Sn, Mg) were doped into MnOx catalyst, and the modification of Co enhanced the performance and tolerance of SCR to SO2 [13,14]. Co for crossing one of the metallic elements, CoOx has unique oxygenation and reduction properties good co-catalyst, which can enhance the catalytic performance in the reaction high activity and high selectivity. Hu et\u00a0al. [10] synthesized Co3O4 and manganese-doped Co3O4 nanoparticles by co-precipitation method and used them as catalysts for NH3 selective catalytic reduction of NO (NH3-SCR). The NH3-SCR activity of Mn0.05Co0.95Ox catalyst was greatly enhanced by the addition of Mn oxide. The results show that the incorporation of manganese can provide more acidic sites on the catalyst, and the binary nitrate species from NOx adsorption are significantly activated on the surface of Mn0.05Co0.95Ox catalyst. Xu et\u00a0al. [15] designed and fabricated monolithic porous MnCoxOy nanocubes on a titanium grid, as a denitrification catalyst for NH3-SCR. Characterization results show that the surface of the titanium grid was uniformly coated with cubic arrays, which prevented the migration and aggregation of metal oxides and allowed the components to act synergistically during the catalytic reaction. In addition, due to its robust structure and morphology, the catalyst can maintain a high NOx conversion while exhibiting excellent catalytic cycle stability and good hydrogen resistance. Liu [16]. prepared a series of Ba- and Co-doped MnOx catalysts by citric acid complexation. The effects of Ba and Co doping on the performance of MnOx low-temperature NH3-SCR were investigated. The experimental results showed that the addition of Co then promoted the catalytic performance of manganese oxide. When Ba and Co were co-doped, the performance of the catalyst was significantly improved. 3BaMnCoOx showed the most excellent catalytic performance, with the catalytic activity above 99% when the reaction temperature was higher than 180\u00a0\u00b0C.But the single structure catalyst is still not resistant to the poisoning caused by SO2 and H2O. Core-shell nanomaterials have cavity and composite properties, the unique shell structure can effectively reduce the contact probability of active substances with H2O or/and SO2, slow down catalyst poisoning, prolong the service life of catalysts [17\u201320].Up to now, many scholars have prepared core-shell catalysts with different combinations of active components and shell materials [21,22]. The most typical core-shell catalyst is the one with Mn, Ce, Co and other metal oxides as the core and TiO2 as the shell and its SCR reaction at low temperature was studied, the results show that, as opposed to a single active component, core-shell structure catalyst on the basis of no reduction in activity [23,24]. It significantly improved the resistance to H2O and SO2, this is because the shell effectively inhibits the formation of sulfate species on the surface. Therefore, core-shell catalyst has good SO2 resistance [25\u201328].In this work, a novel Co(3-x)MnxO4@TiO2 core-shell catalyst was prepared by two-step method. The sol-gel method can easily achieve molecular level mixing, and the dynamic coating method can uniformly distribute the material on the surface of the coated material. Therefore, the catalysts prepared by the above two methods have better hybrid type, stability and uniformity at the molecular level. Firstly, MnOx were prepared by sol-gel method, then Co(3-x)MnxO4@TiO2 catalyst was prepared by dynamic coating method, which performed typical core-shell structure. In this paper, surface response method was used to explore the optimum preparation conditions, and on this basis to study the activity and resistance of the catalyst [29,30]. Then, the effect of SO2 in flue gas on SCR activity of core-shell catalyst was studied, and the mechanism and reason of improving SO2 resistance of core-shell catalyst with special morphology were analyzed. Through the analysis of above conclusions, the best proportion of core-shell catalyst was obtained to improve the activity and resistance.Co(3-x)MnxO4 was prepared by sol-gel method. Firstly, CO(NO3)3\u00b76H2O and Mn(AC)2\u00b74H2O were weighed to dissolve in A beaker filled with deionized water, and then C6H12O6 was weighed to dissolve in A beaker filled with deionized water (C6H12O6: Mn+ CO =0.8). Add the solution in A beaker to the solution in B beaker under 60\u00a0Hz ultrasound, continue to stir, then aged at 70\u00a0\u00b0C for 3\u00a0h to form wine red transparent sol, black porous powdered forebody was obtained after 12\u00a0h drying at 100\u00a0\u00b0C, finally, the samples were obtained by calcinating in air. The catalysts were denoted as Co(3-x)MnxO4, where x refers to 0.5, 1, 1.5, 2 and 2.5, thus the Mn:Co molar ratio was set as 1:5, 1:2, 1:1, 2:1 and 5:1 [31].After the preparation of the core material, the dynamic coating method was used to coat the shell material. Co(3-x)MnxO4 catalyst prepared in the above experiments was weighed and added with 100\u00a0ml anhydrous ethanol and a concentrated ammonia water solution (28wt%, 15\u00a0mol/L), by ultrasonication for 30\u00a0min. Then, different molar amounts of C16H36O4Ti were added dropwise and the reactions were allowed to proceed for 24\u00a0h at 45\u00a0\u00b0C under continuous mechanical stirring. The resultant products were separated, collected and washed with deionized water and ethanol for several times. Finally, the obtained powders were dried in an electric oven at 100\u00a0\u00b0C for 12\u00a0h and calcined at 500\u00a0\u00b0C for 2\u00a0h. The catalysts were denoted as Co(3-x)MnxO4@TiO2(y), where y refers to the (Mn+Co):Ti molar ratio of 1:5, 1:2, 1:1, 2:1 and 5:1, respectively [31].The Response Surface Methodology (RSM) is also known as Response Surface Design Methodology. The central composite design (CCD) is an experimental design developed based on partial experimental design and 2-level full factorial experimental design approach. The addition of a point to the 2-level experimental design corresponds to an additional level of thus, the nonlinear relationship between response values and factors can be investigated. Suitable for 2\u223c6 influencing factors. The number of experiments is usually between 14 and 90. In this paper, the 3 factors and 5 levels require 20 groups of experiments. Combining the results of scholars' studies and the literature, the response surface method was used to investigate the effects of core-shell ratio, ammonia addition and calcination temperature on the catalyst performance. The CCD experimental design was carried out using Design expert software and the design parameters are shown in Table\u00a01\n. The results of the CCD 3 factor 5 level experimental design using Design expert software and the experimental results are shown in Table\u00a02\n.The morphological characteristics of all prepared catalysts were characterized by transmission electron microscopy (TEM), high-resolution TEM (HR-TEM), and element mapping using a FEI TECNAI G2 F20 instrument operated at 200kV. The N2 adsorption and desorption isotherms were obtained by an analyzer (ASAP 2020, Micromeritics, USA) pre-treated at 300\u00a0\u00b0C for 2h to characterize the pore volume, average pore size, and specific surface area of the nano-composites. X-ray diffraction (XRD) patterns were obtained in a system (D8 forward, Bruker, Germany), with Cu Ka radiation scanning 2\u03b8 from 10 to 90\u00b0, 6\u00b0/min. Raman spectroscopy of the catalyst was performed at room temperature with the Fisher Scientific DXR2 at 532\u00a0nm for an exposure time of 50\u00a0s.X -ray photoelectron spectroscopy (XPS) spectra were obtained on the Thermo Feather Science Al K\u03b1 250XI spectrometer, calibrated against the C1s peak (binding energy at 284.6\u00a0eV) of the surface contamination, and the peak differential simulation was further analyzed using XPS peak software.\nIn-situ DRIFT experiments were performed on the spectrometer (IS50, Thermo, USA), which was equipped with an MCT/ A detector cooled by liquid nitrogen and an in-situ drift reaction unit with a zinc selenide window. Before each experiment, the samples were pre-treated at 400\u00a0\u00b0C with N2 at 100\u00a0mL/min for 1\u00a0h. The background spectrum is collected in flowing N2 and automatically subtracted from the sample spectrum. Record 650\u20134000\u00a0cm\u22121 by collecting 100 times with a resolution of 4\u00a0cm\u22121.The SCR activity of NO by NH3 was performed in a fixed-bed reactor with an inner diameter of 7 mm. The temperature was raised from 75 to 275\u00a0\u00b0C. The typical reactant gas consisted of 500\u00a0ppm NO, 500\u00a0ppm NH3, 100\u00a0ppm SO2 (if added), 10 vol% H2O (if added), 5 vol% O2 and balance N2 with a gas hourly space velocity (GHSV) of 24000 h\u20131. Water vapor (10 vol%) was generated by passing N2 through a heated bottle containing deionized water. The NO, O2, NH3 and SO2 were mixed in the mixing tube through the gas flow-meter, and after passing the sample, it enters the Fourier gas infrared analyzer for detection. The NOx (NO and NO2) conversion and N2 selectivity were calculated as follows:\n\n(1)\n\n\n\nNOx\n\nconversion\n\n\n(\n%\n)\n\n=\n\n\n\nC\n\nNOx\n\nin\n\n\u2212\n\nC\n\nNOx\n\nout\n\n\n\nC\n\nNOx\n\nin\n\n\n\u00d7\n100\n\n\n\n\n\n\n(2)\n\n\n\nN\n2\n\n\nselectivity\n\n(\n%\n)\n\n=\n1\n\u2212\n\n\n2\n\u00d7\n\nN\n2\n\n\nO\nout\n\n\n\nNO\n\nx\nin\n\n+\nN\n\nH\n\n3\nin\n\n\n\u2212\nNO\n\nx\nout\n\n\u2212\nN\n\nH\n\n3\nout\n\n\n\n\n\u00d7\n100\n\n\n\n\nTwenty groups of Co(3-x)MnxO4 core materials were prepared under different conditions, and the optimum preparation conditions were investigated by activity detection. The optimum preparation conditions were selected by surface analysis: Mn:Co=2:1, calcination temperature was 450\u00a0\u00b0C, reaction temperature was 175\u00a0\u00b0C. On the basis of kernel optimization, the second order surface response was carried out. The coating of TiO2 shell was investigated onto the optimized active component material via the design of three factors (core-shell ratio, ammonia water content and calcination temperature) with five levels, thus, twenty groups of samples were prepared under different combination conditions. The catalytic activity of these catalysts was tested, and then the surface analysis diagram was prepared according to the test results, furthermore, the best coating conditions was selected to make the optimal CoMn2O4@TiO2 core-shell catalyst.The experimental results of 20 tests with 5-level of 3 factors in CCD design (as listed in Table\u00a01) are shown in Table\u00a02. The response surface analysis is not an examination of indicators, but rather multiple indicators, analyzing the interaction between factors and the most important factors affecting the experiment. It is the center point as the core of the surface to build a 3D surface map, so there are repeated experiments that are randomly generated, the purpose is to verify the accuracy and precision of the experimental results. The above results were imported into the software and analyzed to establish an objective function model of NOx conversion (Y), respectively. It should be pointed out that in the regression equation analysis of variance, the corresponding P values of X1X2, X1X3 and X2X3 in Y model was 0.6825, 0.5846 and 0.8682, which was far greater than 0.0001, respectively, so that they were deleted to optimize the regression equation, as presented in Eq.\u00a0(3), respectively. The quality of a design of experiments is measured by the variability of the estimate of its coefficients and consequently by the variability of the response estimate.It should be noted that the model P value of NOx conversion rate is less than 0.0001, which shows that the NOx conversion rate model is significant. The P value of missing degree is 0.5169, greater than 0.05, it indicates that the degree of absence is not significant. More importantly, the R\n2 variance of the regression equation of NOx conversion is 0.9257, close to 1, it shows that the model established by RSM has a good fitting degree. As you can see from Table\u00a02, the central point of the CCD design, a catalyst of 500\u00a0\u00b0C calcined with 1.5 active component (core-shell ratio 1:1) and 0.5 ml ammonia water (ammonia water to active component ratio 4.17) was prepared, six tests were performed at 225\u00a0\u00b0C (run 1, 2, 3, 13, 18, 20), and the SCR activity (94.9\u00b13.6%) was higher than 90%. It shows that the experimental method is feasible and the experimental data is repeatable.When the core-shell ratio is 1:5 (Run-19), the NOx conversion rate of Run-19 catalyst is 86.98%, which is lower than that of Run-3 (central point). It indicates that the proportion of active ingredients is too low to play an effective catalytic role. The excess of active component (Run11, core-shell ratio 5:1) also leads to pore clogging and surface accumulation, the effective active site of the catalyst could not be completely dispersed, which cannot show catalytic effect effectively. The SCR activity of Run-4 was compared, high temperature calcination is not good for catalytic reaction (run-5, Run-15, Run-16, and Run-17, also proved this). This may be due to the sintering of active ingredients or the reduction of specific surface area. As one can see from the activity tests on Run-7, less content of ammonia water is not good for catalytic reaction (Run-8, Run-15, Run-16, also proved this). This may be due to that too little amount of ammonia water as a binder may lead to inadequate core-shell bonding, the instability of core-shell structure also results in a decreased activity. Verified by multiple groups of experiments, it can be found that the experimental results designed by CCD have certain repeatability. Therefore, the above model equations can better predict the experimental values, the theoretical basis was laid for RSM optimization of preparation parameters and reaction conditions.\n\n(3)\n\n\n\n\n\nY\n=\n92.98\n+\n1.33\n\nx\n1\n\n+\n1.31\n\nx\n2\n\n\u2212\n7.61\n\nx\n3\n\n\u2212\n1.64\n\nx\n1\n\n\nx\n2\n\n\u2212\n2.20\n\nx\n1\n\n\nx\n3\n\n\n\n\n\n\n\n\u2212\n0.66\n\nx\n2\n\n\nx\n3\n\n\u2212\n5.63\n\nx\n1\n2\n\n\u2212\n3.89\n\nx\n2\n2\n\n\u2212\n7.08\n\nx\n3\n2\n\n\n\n\n\n\n\n\n\nFig.\u00a01\n shows the 3D response surface diagram of core-shell ratio (a), calcination temperature (b) and ammonia water addition amount (c) to NOx conversion rate, respectively. Fig.\u00a01(a) shows a good 3D hillside shape that the NOx conversion rate increases first and then decreases with the increase of core-shell ratio, this indicates that there is an optimal load value, which can be obtained through model optimization analysis [28]. The reason for this rule may be that too low core-shell ratio results in too few active components, it is difficult to meet the demand of gas molecules, while the excessive core-shell ratio leads to the accumulation and blockage of TiO2 shell materials that prevents gas molecules from contacting the pore [29,30]. According to the analysis in Fig.\u00a01(b), when the calcination temperature is in the range of 400\u223c480\u00a0\u00b0C, it has little effect on NOx conversion of catalyst that can reach more than 85%, and the NOx conversion rate is on the rise. In the temperature range of 480\u223c600\u00a0\u00b0C, the NOx conversion rate decreases gradually. Many research results show that high temperature may lead to the catalyst sintering, that is, the grain size of active component becomes larger, specific surface area reduction, the catalyst has thermal deactivation [31\u201333]. As the calcination temperature was gradually increased and the calcination time lengthened, the sintering of catalyst will be aggravated accordingly. With the gradual increase of temperature and the extension of calcination time, the sintering process with crystallization changes will become more serious. Low calcination temperature may lead to incomplete calcination of catalyst [34]. Combined with Fig.\u00a01(a), Fig.\u00a01(c) showed that the core-shell catalyst activity increased first and then decreased with the addition of ammonia water, but to a lesser extent, the activity was above 85%. Therefore, the amount of ammonia water is not the main factor affecting the activity, optimal range of ammonia water content is 0.3\u223c0.7\u00a0mL (ratio of ammonia water to active component: 2.5\u223c5.8). The reason is that ammonia water acts as an adhesive to facilitate the hydrolysis of C16H36O4Ti, it will not affect the change of active component [35]. However, too little amount will lead to insufficient core-shell bonding, unstable micro-porous structure and easy collapse. Too much amount will affect the distribution of acid sites and reduce the activity of catalyst to a certain extent.The amount of ammonia water is fixed as 0.5, the NOx conversion rate can reach 93.8% within a 5% reduction in maximum activity of 4.7%. The scaling range is selected by the transversal method shown in Fig.\u00a02\n, when the amount of active component is 1.1\u223c2.2, nitrogen oxide conversion can achieve a maximum activity reduction of 5%. To facilitate calculation and preparation, the approximate value of active component is 1\u223c2 (core-shell ratio is 1:2\u223c2:1). Taking the boundary value as the limiting condition, two groups of catalysts were prepared, the optimal core-shell ratio was selected by activity and also resistance tests. The purpose of this experiment is to select the core-shell ratio samples with the best nitrogen selectivity and sulfur tolerance under the premise of ensuring a certain high activity.The NOx conversion results of catalysts with different proportions at different temperatures are presented in Fig.\u00a03\n(a). TiO2 alone has almost no activity below 225\u00a0\u00b0C, so when comparing catalysts with different core-shell ratios, the effect of activity brought by TiO2 is not considered. In the range of core-shell ratio of from 1:2 to 2:1, four ratios were selected for the full-temperature window activity detection at 75\u223c225\u00a0\u00b0C. It can be analyzed from Fig.\u00a03(a), with the increase of reaction temperature, the SCR activity of catalysts with different proportions also increased, except for the single active component CoMn2O4, activity of other catalysts reached the highest and tended to be stable at 225\u00a0\u00b0C. The activity of CoMn2O4@TiO2 (1:1) and CoMn2O4@TiO2 (1:2) catalysts was lower than that of CoMn2O4. This is because that, with the same mass of catalyst, a decrease in activity was due to a decrease in the proportion of active components [36]. Compared with CoMn2O4, SCR activity of CoMn2O4@TiO2 (1.5:1) and CoMn2O4@TiO2 (2:1) catalysts was improved to some extent, respectively. It is more obvious below 150\u00a0\u00b0C, which indicates that the two core-shell catalysts have better activity at low temperature.As shown in Fig.\u00a03(b), the N2 selectivity of the two catalysts decreases with the increase of temperature, which may be due to the oxidation reaction of NO and NH3 in the process of NH3-SCR reaction, which coexists with O2 to produce NO2, N2O and other by-products, leading to the decrease of selectivity [37]. These side effects are more likely to occur at high temperatures, and the storage rate of NH3 increases with the increase of temperature. At the same time, a part of NO is oxidized to NO2, which is conducive to the occurrence of \"fast-SCR\" reaction. However, excessive NO2 in the system will also lead to preferential reaction with NH3, thus the N2 selectivity is further reduced [38]. As can be seen N2 selectivity of CoMn2O4@TiO2 is always higher than that of CoMn2O4. The results shows that the Ti element coated in shell obviously inhibited the formation of N2O, thus improved the selectivity to N2. Compared to other catalysts of different proportions, CoMn2O4@TiO2 (2:1) always had the highest N2 selectivity to above 80% in full temperature window test, which can reach more than 90% when low temperature range below 175\u00a0\u00b0C. Therefore, it can be concluded that CoMn2O4@TiO2(2:1) core-shell catalyst has good stability, high N2 selectivity and high SCR activity.Sulfur resistance of catalysts with different core-shell ratios in the presence of 100ppm SO2 is shown in Fig.\u00a04\n. Under the condition of 100\u00a0ppm SO2 for 20\u00a0h at 225\u00a0\u00b0C, CoMn2O4 catalyst activity decreased significantly to about 40%, SO2 is connected for 20 h, which could not be recovered by removing SO2. It shows that the catalyst is completely inactivated in the presence of SO2 for a long time and the result of this decrease in activity is irreversible. In contrast, the activity of CoMn2O4@TiO2 core-shell catalyst was higher than CoMn2O4 sample after SO2 treatment for 20\u00a0h. Importantly, the activity of CoMn2O4@TiO2 could be restored to close to the initial value after SO2 removal. This indicates that CoMn2O4@TiO2 core-shell catalyst has better SO2 resistance and toxicity reversibility than CoMn2O4. This may be due to that the coating of TiO2 shell inhibits the formation of sulfate on the surface of active component, Ti(SO4)O in CoMn2O4@TiO2 protects CoMn2O4 from SO2 passivation [39]. The reason may be reaction between SO2 and TiO2 on the outer surface of the catalyst generates TiSO4O, thus trapping the SO2 in the flue gas outside the catalyst. In addition, SO2 might react with NH3 in the gas to form (NH4)2SO4 There may be the following reactions are TiO2\u00a0+\u00a0SO2\u00a0\u2192\u00a0TiSO4O and 2SO2\u00a0+\u00a0O2\u00a0+\u00a02NH3\u00a0+\u00a02H2O\u00a0=\u00a02(NH4)2SO4 (correspond to Eqs.\u00a0(13) and (14) in 3.4).After SO2 was injected at 225\u00a0\u00b0C for 4\u00a0h, the activity of all CoMn2O4@TiO2 core-shell catalysts decreased to a certain extent. The activity of CoMn2O4@TiO2 (2:1) and CoMn2O4@TiO2 (1.5:1) was still about 80%. Higher than CoMn2O4@TiO2 (1:1) and CoMn2O4@TiO2 (1:2). After SO2 infiltration for 20\u00a0h, CoMn2O4@TiO2 (2:1) can still maintain nearly 80% activity, but the catalytic activity of the other three core-shell ratio catalysts further decreases to less than 60%. The activity of CoMn2O4@TiO2 (2:1) catalyst did not decrease significantly that still be maintained at more than 75%, indicating that this catalyst has better SO2 resistance with the core-shell ratio of 2:1 that ensures sufficient active sites on the active components. It also ensures that the uniform coating of TiO2 shell inhibit the formation of shell ammonium sulfate on the core surface, thus it has good resistance to SO2 poisoning [40]. Based on the above research, 2:1 core-shell ratio was selected as the preparation ratio of CoMn2O4@TiO2 core-shell catalyst.It has been reported in the literature that H2O and SO2 in exhaust gas can induce catalyst inactivation [41]. Fig.\u00a05\n shows the influence of H2O and SO2 on the NOx conversion of CoMn2O4@TiO2 (2:1) and CoMn2O4 catalysts at 225\u00a0\u00b0C. In the presence of 10 vol% H2O, the catalytic activity of the two catalysts were both decreased. The catalytic activity of CoMn2O4 catalyst decreases to 45% after 20 h and only recovered to 80% when the water was removed from the flue gas, however, the catalytic activity of CoMn2O4@TiO2 (2:1) remained stable at more than 80% after 20h H2O was introduced, and quickly recovered to the original value after the removal of H2O. The results show that CoMn2O4@TiO2 (2:1) catalyst has good water resistance The stability of catalyst activity may be due to the competitive adsorption of reactants at the active sites on the catalyst surface [42]. When 100ppm SO2 was introduced into the reaction system, the NOx conversion rates for CoMn2O4@TiO2 (2:1) and CoMn2O4 ranged from 100% to 75% and 30%, respectively. After SO2 was stopped, the NOx conversion rate of CoMn2O4@TiO2 (2:1) gradually increases to 100% within 4h and is relatively stable. However, the NOx conversion rate of CoMn2O4 can only be maintained at about 40%. When 10 vol.% H2O and 100ppm SO2 were introduced into the system together, the NOx conversion of the two catalysts showed a decreasing trend. After 24\u00a0h, the NOx conversion rate of CoMn2O4@TiO2 (2:1) remains around 78%, while NOx conversion rate of CoMn2O4 can only remain around 58%. After removing the two medium agents, the NOx conversion rate of CoMn2O4@TiO2 (2:1) still recovered to 98% but CoMn2O4 catalyst only recovered to 83%. It shows that CoMn2O4@TiO2 (2:1) has better water and sulfur resistance than CoMn2O4. Considering the consistency of active components, the difference in toxicity resistance of the two catalysts must come from titanium dioxide shell protection in the core-shell structure.The micro-structure and morphology of CoMn2O4 and CoMn2O4@TiO2 (2:1) catalysts were studied by TEM and HRTEM. As shown in Fig.\u00a06\n(a) and (b), CoMn2O4 catalyst is in the form of scales, and CoMn2O4@TiO2 catalyst is in the form of cluster centered on a nuclear sphere, with an average diameter of about 200\u00a0nm and a shell thickness of about 40\u00a0nm. HRTRM images of CoMn2O4@TiO2 catalyst in Fig.\u00a06(d) show high crystallinity with spacing distances between lattice planes of 0.165 and 0.234\u00a0nm, corresponding to manganese-cobalt spinel crystal plane and anatase titanium dioxide crystal plane, respectively [43]. The lattice stripes of spinel oxides in the nuclei indicate the presence of crystalline Mn-Co oxides in the catalysts. As can be seen in Fig.\u00a06(e, h), the addition of Ti will cause a certain degree of aggregation of CoMn2O4@TiO2 catalyst. To understand the element distribution of CoMn2O4@TiO2 catalyst, element mapping was used to determine the elements, as shown in Fig.\u00a06(f, g, i, j). The element mapping profiles of manganese and cobalt spinel phases are significantly smaller than those of oxygen and titanium. O and Ti cover the surface of Mn and Co elements, indicating that manganese and cobalt species are covered by titanium dioxide shells. Therefore, CoMn2O4@TiO2 core-shell catalyst coated with titanium dioxide shell is formed.To identify the phase composition of CoMn2O4 and CoMn2O4@TiO2 catalysts, XRD analysis was performed in Fig.\u00a07\n(a), diffraction peaks of the two catalysts are similar, and both have obvious CoMn2O4 crystal phase peaks, while CoMn2O4 @TiO2 catalyst also has distinct TiO2 crystal phase peaks. The diffraction peaks of CoMn2O4@TiO2 catalyst are consistent with the typical diffraction peaks of single TiO2 nanocrystals and CoMn2O4 spinel phase [44]. The peak values at 14.2\u00b0, 29.4\u00b0, 32.96\u00b0, 36.5\u00b0, 47.2\u00b0, 60.8\u00b0 and 65.3\u00b0 may be CoMn2O4 phase peaks of different crystal phases [45]. The latter peaks at 25.4\u00b0, 38.6\u00b0, 48.8\u00b0 and 75.5\u00b0 correspond to the crystal signal diffraction peaks of TiO2 in the shell of CoMn2O4@TiO2 at different exposed crystal planes [46]. CoMn2O4@TiO2 (2:1) shows the peak of anatase TiO2 and CoMn2O4 spinel phase, indicating that both anatase TiO2 and CoMn2O4 have good crystallinity. In the Figure, CoMn2O4 peak strength of CoMn2O4@TiO2 (2:1) catalyst is reduced, which may be caused by the good dispersion of anatase TiO2 on the catalyst surface. XRD results showed that TiO2 crystals were likely to be uniformly coated on CoMn2O4 catalyst, which was further confirmed by Raman spectra.More information about the crystal structure was obtained by Raman spectroscopy measurements as shown in Fig.\u00a07(b). Some studies have shown that the vibration characteristics of single manganese-cobalt spinel show low Raman activity, which only appears in the 302, 380, 652 and 853\u00a0cm\u22121 region [47]. The peaks of CoMn2O4@TiO2 are located at 167, 396, 508, and 639\u00a0cm\u22121 respectively, which are typical diffraction peaks of anatase titanium dioxide [48,49]. This result is consistent with the XRD analysis in Fig.\u00a07(a), and also coincides perfectly with the diffraction peak of pure TiO2 crystalline phase in Fig.\u00a07(b). The diffraction peaks of CoMn2O4@TiO2 are almost identical to those of pure TiO2 and pure CoMn2O4 spinel phases, which indicates that CoMn2O4@TiO2 is made from the latter two configurations, and some of the peaks of CoMn2O4@TiO2 are higher than those of TiO2, which may be because the CoMn2O4 diffraction peaks are similar to those of TiO2, so the peaks overlap and show more obvious. It is worth noting that the CoMn2O4 nucleon structure in CoMn2O4@TiO2 shows a low intensity band in the Raman spectrum, and it is difficult to observe any characteristic peaks of metal oxides. These facts proved that no obvious Mn and Co phase was observed due to the cover of TiO2 on the surface of CoMn2O4 cluster centered on a nuclear sphere, and the interaction between metal oxides may be greater. The strong interaction between titanium dioxide and active metal oxides contributes to the activity of NH3-SCR. This result is in good agreement with subsequent XPS results. Therefore, XPS and Raman results support the presence of CoMn2O4 as an active component in CoMn2O4@TiO2 (2:1), which is critical for excellent SCR performance.\nFig.\u00a08\n shows the FT-IR spectra of the two catalysts and the presence of a certain phase can be determined by detecting some specific functional groups. For CoMn2O4 catalyst, it can be seen from the Figure that there is an obvious peak near the wave number of 500\u00a0cm\u22121. After checking, it is highly consistent with the position of Mn-O bond and Co-O bond in tetrahedral gap, and the peak at 632\u00a0cm\u22121 is consistent with the position of Mn-O bond and Co-O bond in octahedral gap. The formation of the bond is also good, and the CoMn2O4 prepared by the sol-gel method has a spinel structure and good crystal structure. The diffraction peak positions of CoMn2O4@TiO2 catalyst and CoMn2O4 catalyst are similar, but there are obvious vibration bands at 1121\u00a0cm\u22121 and 1405\u00a0cm\u22121, which is due to the relatively strong electronic positioning trend of anatase TiO2 and the vibration peak. This indicates the existence of TiO2 crystal phase, but the peak position is relatively weak, because TiO2 accounts for less in the core-shell catalyst, and the main peak is dominated by CoMn2O4. However, the vibration peaks of Mn-O bond and Co-O bond in CoMn2O4@TiO2 catalyst are relatively weak, which is also because the TiO2 coating on the catalyst reduces the vibration frequency of manganese oxide, resulting in reduced light transmittance.The chemical elements and oxidation states on the surface of the catalyst were determined by X-ray photoelectron spectroscopy (XPS). XPS spectra of Mn, Co, O and Ti are shown in Fig.\u00a09\n. Fig.\u00a09(a) and Table\u00a03\n shows the Mn2p XPS spectra of CoMn2O4@TiO2 catalyst and CoMn2O4 catalyst, the Mn2p spectra of the two catalysts are typical peaks centered on 641.7eV and 653.6\u00a0eV, which are assigned to Mn2p3/2 and Mn2p1/2, respectively. Through peak fitting, Mn2p3/2 spectrum can evolve into Mn2+ peak (640.8\u00a0eV) Mn3+ peak (642.0\u00a0eV) and Mn4+ peak (644.0\u00a0eV), while Mn2p1/2 spectrum evolves into Mn4+ peak (653.44\u00a0eV) [50]. As shown in Fig.\u00a09(a), the Mn content of CoMn2O4 is much higher than that of CoMn2O4@TiO2, which is attributed to the core-shell structure of CoMn2O4@TiO2, resulting in a decrease in the proportion of Mn content. It is well known that Mn in different valence states will affect the electron transfer and REDOX ability of the catalyst. Manganese ion is more active in high oxidation state, because Mn4+ ion can promote the oxidation of NO into NO2\n[51]. We all know that Mn in different states can influence the electron transfer and REDOX capacity of catalyst. Mn3+ and Mn4+ is considered to be the active site for adsorption and activation, along with redox couples of Mn3+\u2194 Mn4+. Higher ratios of Mn3+ and Mn4+ are favorable for Mn dissolution species switch between various valence states and promote Catalytic cycle and performance. The (Mn3++Mn4+)/ (Mn2++Mn3++Mn4+) ratio of CoMn2O4@TiO2 catalyst is 77.7%, which is higher than that of CoMn2O4 catalyst 67.4%. Therefore, it is beneficial to the transformation ability of Mn species between different valence states and promotes the catalytic cycle and performance. Therefore, it is more suitable for low-temperature SCR, which may be the reason for the higher reactivity of CoMn2O4@TiO2. The results show that CoMn2O4@TiO2 catalyst can rapidly oxidize the reduced Mn2+ ions to Mn3+ and Mn4+ ions during SCR reaction. This can improve the REDOX capacity of CoMn2O4@TiO2 catalyst, which is conducive to the improvement of catalytic performance.In addition, XPS spectra of Co2p on CoMn2O4 and CoMn2O4@TiO2 catalysts are shown in Fig.\u00a09(b) and Table\u00a03, which are 780.5\u00a0eV Co3+ and 782.2\u00a0eV Co2+, respectively. The other spin orbital component (Co2p1/2) is at 796.2\u00a0eV and 797.8\u00a0eV, corresponding to Co3+ and Co2+ configuration, respectively [52]. In addition, peaks centered at 786.6eV and 802eV can be assigned to satellite peaks of Co2p [53]. As can be seen from the Figure, the Co3+/ (Co2++Co3+) of CoMn2O4@TiO2 is 53.5% higher than that of CoMn2O4 (51.8%), indicating that the average valence state of Co ions on CoMn2O4@TiO2 catalyst is higher. Due to the stronger REDOX capacity of Co3+, Co2+ is far less active than Co3+. Therefore, more Co3+ species are conducive to the improvement of REDOX performance, which also determines that CoMn2O4@TiO2 has a better catalytic activity.As shown in Fig.\u00a09(c) and Table\u00a03, the O1s spectra of both catalysts can be divided into two main peaks. A high binding energy peak (531.2eV) corresponds to surface adsorbed oxygen (hydroxyl or oxygen ion, O\u03b1) or defective oxide, while the second low binding energy peak (529.7eV) corresponds to lattice oxygen (O2\u2212, O\u03b2) [54,55]. The O\u03b1/(O\u03b1+O\u03b2) ratio of CoMn2O4@TiO2 catalyst decreased from 31.9% to 26.1% compared with CoMn2O4 catalyst. It can be clearly seen that the binding energies of O\u03b1 and O\u03b2 at CoMn2O4@TiO2 shift to higher values compared with CoMn2O4 catalyst, and higher binding energies indicate more stable covalent bonds. In addition, the higher concentration of O\u03b1 species was conducive to the occurrence of NH3-SCR reaction [56]. The concentration of O\u03b1 species in CoMn2O4 is higher than that in CoMn2O4@TiO2, however, a single O\u03b1 concentration does not determine a higher activity of CoMn2O4 catalyst, but only leads to a catalyst with the same high activity at the optimum reaction temperature as CoMn2O4@TiO2 above 200\u00a0\u00b0C, which is consistent with the activity experimental results.For Ti2p XPS spectra in Fig.\u00a09(d) and Table\u00a03, the Ti2p spectra of CoMn2O4@TiO2 catalyst can be divided into Ti2p3/2 and Ti2p1/2 titanium dioxide forms respectively. These bands were observed at 457.9 eV and 463.5 eV, where \u25b3BE= gap was 5.6 eV, characteristic of Ti3+ species. The bands at 459.8 eV and 464.5 eV were Ti4+ species [57,58]. For CoMn2O4@TiO2 catalyst, Ti4+ is still the main phase, and there is a certain amount of Ti3+ on the surface of the catalyst. This is because the existence of oxygen vacancies in titanium dioxide can lead to the corresponding charge balance, part of Ti4+ in titanium dioxide receives electrons from oxygen and becomes Ti3+ (Ti4++ E\u2212\u2194 Ti3+) [59,60].According to XPS analysis, the REDOX capacity of core-shell catalyst CoMn2O4@TiO2 is slightly higher than that of CoMn2O4 catalyst, which explains the increased selectivity of N2 to CoMn2O4@TiO2. In the SCR reaction of CoMn2O4@TiO2 catalyst, there is electron transfer between Mn, Co and Ti, that is Mn4++Co2+\u2194 Mn3++Co3+, Mn3++Ti4+\u2194Mn4++Ti3+(corresponding to Eqs.\u00a0(11) and (12) in 3.4). The strong electron interaction plays an important role in NH3-SCR process under dry and wet conditions.The adsorption of NH3 on the catalyst surface plays a key role in the SCR reaction, so the temperature-programmed desorption of NH3 (NH3-TPD) was used to investigate the number and strength of acid sites on the catalyst surface. The NH3-TPD curves of CoMn2O4 and CoMn2O4@TiO2 catalysts are shown in Fig.\u00a010\n. The CoMn2O4 catalyst shows three major desorption peaks centered at 111, 259 and 602\u00a0\u00b0C, respectively. While the area peaks of CoMn2O4@TiO2 catalyst were at 104, 309 and 736\u00a0\u00b0C. The NH3-TPD curves for both catalysts can be divided into the following three regions of 50-200\u00a0\u00b0C, 200-400\u00a0\u00b0C and 400-800\u00a0\u00b0C, which can be attributed to NH3 adsorption at weak, medium and strong acid sites, with strong peaks in the high temperature interval probably due to N2 desorption [41]. The corresponding peak areas of the various peaks are denoted as Px (x\u00a0=\u00a01, 2, 3). The position of the medium-intensity acid peak of CoMn2O4@TiO2 was significantly shifted to a higher temperature range compared with that of CoMn2O4, and this result indicated that the introduction of TiO2 was beneficial to improve the catalytic activity at high temperatures [61]. The strong acid peak sites of CoMn2O4@TiO2 catalyst are much higher than those of CoMn2O4 catalyst, indicating that it has more strong acid sites and stronger acidity. For CoMn2O4, the peak at 259\u00a0\u00b0C corresponds to NH4\n+ desorption from the Br\u00f6nsted acid site, while the peaks at 602\u00a0\u00b0C correspond to NH3 desorption from the Lewis acid site [62]. In contrast, in the desorption curve of CoMn2O4@TiO2, the broad weak peak at 309\u00a0\u00b0C corresponds to the desorption of NH3 from the Br\u00f6nsted acid site, while the sharp peak at 736\u00a0\u00b0C is from the desorption of the Lewis acid site. It is well known that Lewis acid sites play an important influence in SCR systems, so CoMn2O4@TiO2 has a higher SCR activity compared to CoMn2O4.The percentage of peak area for each acid site was calculated for both catalysts in Fig.\u00a010, and the number of surface acid sites was listed in Table\u00a04\n based on the total acid amount. The amount of surface acids follows the order of CoMn2O4@TiO2 > CoMn2O4. It can be clearly seen that the CoMn2O4@TiO2 sample presents stronger acid sites (0.54 \u03bcmol/g) compared to the CoMn2O4 sample (0.43 \u03bcmol/g). This indicates that the TiO2 shell leads to a catalyst with more acid sites and stronger acidity.The redox performance of catalysts plays an important role in the NH3-SCR reaction, so the reduction of CoMn2O4 and CoMn2O4@TiO2 catalysts was tested using programmed temperature rise reduction (H2-TPR). As shown in Fig.\u00a011\n, the H2-TPR spectrum of CoMn2O4 catalyst can be divided into 2 peaks in the range of 50\u2013800\u00a0\u00b0C. The first peak at around 390\u00a0\u00b0C could be the combination of MnO2\u2192Mn2O3\u2192Mn3O4 and Co3+\u2192Co2+\n[19]. The second peak centered at 552\u00a0\u00b0C can be explained by the coexistence of Mn3O4\u2192MnO and Co2+\u2192Co0\n[19]. Unlike CoMn2O4, another peak exists for the CoMn2O4@TiO2 catalyst, observed at around 208\u00a0\u00b0C can be attributed to the reduction of surface oxygen species [57]. No obvious electron transfer peak position of Ti was observed in the spectrum, which may be due to several reasons. Firstly, the participation ratio of Ti is too small, much lower than that of (Mn+Co), so the peak position is very low and cannot be visualized. The second reason is that TiO2 is uniformly distributed on the surface with high stability, and it is difficult to observe the valence change of Ti without obvious transition. The third reason may be due to the metal-oxygen bonding between Ti and Mn and Co, so the observed peak positions overlap, which may also be the reason why the peak area of CoMn2O4@TiO2 catalyst is significantly larger than the peak area of CoMn2O4.The presence of lattice defects on CoMn2O4@TiO2 catalysts is related to the appearance of oxygen vacancies caused by crystal structure defects. In contrast, the two reduction peaks of the CoMn2O4@TiO2 catalyst were similar to those of the CoMn2O4 catalyst at 50\u2013800\u00a0\u00b0C. However, after Ti doping, the position of the reduction peak gradually shifted to a lower temperature, indicating that the catalyst has both better redox performance and higher catalytic activity at low temperatures. Both peaks show an increase in the area of the reduced peaks belonging to Mn3O4 and MnO. This indicates that the oxygen adsorbed on the surface is reduced and Mn4+ is more easily reduced to the lower valence state, which corresponds to the XPS results.The peak areas and total hydrogen consumption of each reduction peak are given in Table\u00a05\n. Based on the H2 consumption and reduction temperature of the catalysts, it is concluded that the ease of reduction of the catalysts is in the order of CoMn2O4@TiO2>CoMn2O4. The number of redox sites increases after doping with TiO2 shell layer because H2 consumption increases [63]. Meanwhile, the good interaction of oxides such as Mn, Co and Ti are beneficial to improve the NH3-SCR activity.\nIn-situ DRIFT spectra of adsorbed ammonia species and NO+O2 over time on CoMn2O4 and CoMn2O4@TiO2 catalysts at 225\u00a0\u00b0C are shown in Fig.\u00a012\n. As can be seen from Fig.\u00a012(a), after NH3 pre-treatment for 30min, the surface of CoMn2O4 catalyst is covered by different kinds of adsorbed ammonia species. The bands at 1242, 1340, 1450 and 1540\u00a0cm\u22121 were attributed to the coordination ammonia binding to the Lewis acid site, and the bands at 1607 and 1680\u00a0cm\u22121 were attributed to NH4\n+ at the Br\u00f8nsted acid site [64]. After the introduction of NO+O2 mixture gas, the number of ammonia species decreased to a certain extent, and then began to increase with time, indicating that the adsorbed ammonia reacts rapidly with gaseous NO/NO2. After adsorption to the catalyst, NH3 existed in the form of -NH2 and H+. After adsorption saturation of NH3, it further reacted with NO to generate N2 and H2O (corresponding to Eqs. (4) and (5) in 3.4). As shown in Fig.\u00a010(a), bands of ammonia species coincide with several NOx bands, including bridging nitrate (1242\u00a0cm\u22121), nitrite species (1340\u00a0cm\u22121), monotone nitrate (1450 and 1540\u00a0cm\u22121), and nitrogen dioxide (1607\u00a0cm\u22121) [65,66]. In Fig.\u00a012(b), CoMn2O4@TiO2 catalyst also shows Lewis acid sites (1242,1340,1450 and 1540\u00a0cm\u22121) and Br\u00f8nsted acid sites (1607 and 1680\u00a0cm\u22121) [67,68]. When NO+O2 is added, it can be found that the change trend of CoMn2O4@TiO2 is similar to CoMn2O4. Bridging nitrate bands (1242\u00a0cm\u22121), nitrite species (1340\u00a0cm\u22121), monotone nitrate bands (1450 and 1540\u00a0cm\u22121) and nitrogen dioxide (1607\u00a0cm\u22121) were also presented [69]. The peak intensity of ammonia and NOx species decreased first and then increased with the increase of time. The results showed that ammonia species had strong adsorption and activation capacity for the two catalysts. The results showed that compared with CoMn2O4 catalyst, the reaction adsorption peak of CoMn2O4@TiO2 catalyst was more obvious, and the catalyst had more active intermediates, which was beneficial to SCR reaction. From in situ infrared results, we conclude that both Lewis and Br\u00f8nsted acid sites are involved in SCR reactions of both catalysts [68]. It is suggested that Eley-Rideal (E-R) mechanism exists, and the adsorbed ammonia reacts with gaseous NO/NO2 to form N2 (or N2O) and H2O [70].On the other hand, the role of adsorbed NOx in CoMn2O4 and CoMn2O4@TiO2 catalysts was studied by in-situ drift spectroscopy under the same reaction conditions. The reaction drift spectra of pre-adsorbed NOx and ammonia are shown in Fig.\u00a012(c) and (d), respectively. The basic reaction principle of this method is the reaction between NH3 and NO in the gas. Firstly, NO oxidizes with oxygen to form NO2, and NH3 gets an H+ to form NH4\n+, and then NH4\n+ reacts with NO2 to form the intermediate NH4NO2, the last, intermediate can further decompose N2 and H2O (correspond to Eqs. (6)\u2013(8) in 3.4). At the same time, the above-mentioned two reaction pathways are accompanied by side reactions. Mn+ reacts with -NH2 produced in the previous process to generate -NH and H+, and then -NH reacts with NO in the flue gas to generate N2O (corresponding to Eqs. (9) and (10) in 3.4). Bridging nitrate bands (1242\u00a0cm\u22121), nitrite species (1340\u00a0cm\u22121), monotone nitrate bands (1450 and 1540\u00a0cm\u22121) and nitrogen dioxide (1607\u00a0cm\u22121) appeared when NO+O2 mixture was introduced for 30\u00a0min [71]. The band intensity of different types of NOx increases with time, and then decreases gradually with the addition of NH3. NOx species were quickly covered by the characteristic peak of NH3, and the adsorbed NOx species reacted with NH3. to form N2. When NH3 was added, the peak position remained constant, but the intensity of nitrate species decreased with time. These results indicate that the adsorbed NOx can also participate in the NH3-SCR reactions of the two catalysts. The peak intensities of different nitrates were as follows: nitrogen dioxide > monotone nitrate > bridged nitrate. Combined with reactivity, nitrogen dioxide and monodentate nitrate were the main active species. Nitrogen dioxide has high reactivity at low temperature and is easy to react with adsorbed ammonia to form N2 and water [72]. The diffusion of nitrogen dioxide can be coupled with NO oxidation and \u201cfast SCR\u201d chemistry to promote NH3-SCR reaction. More obvious types of NOx with adsorption peaks can be observed on the surface of CoMn2O4@TiO2. Therefore, the reaction of NH3-SCR to CoMn2O4 and CoMn2O4@TiO2 catalyst mainly follows the typical E-R mechanism, while the L-H mechanism exists but does not play an important role.Reaction between reactants:\n\n(4)\n\n\nN\n\nH\n3\n\n\u2192\n\u2212\nN\n\nH\n2\n\n+\n\n\n\nH\n\n+\n\n\n\n\n\n\n\n(5)\n\n\n\u2212\nN\n\nH\n2\n\n+\n\nN\nO\n\u2192\n\nN\n2\n\n+\n\n\nH\n2\n\nO\n\n\n\n\n\n\n(6)\n\n\nN\n\nH\n3\n\n\n(\ng\n)\n\n\n+\n\n\n\nH\n\n+\n\n\n\u2194\n\nN\n\nH\n4\n+\n\n\n(\n\na\nd\ns\n\n)\n\n\n\n\n\n\n\n(7)\n\n\nN\nO\n\n\n(\ng\n)\n\n+\n\n1\n/\n2\n\n\nO\n2\n\n\n(\ng\n)\n\n\u2192\nN\n\nO\n2\n\n\n(\ng\n)\n\n\n\n\n\n\n\n(8)\n\n\n\n\ne\n\n\u2212\n\n+\n\nN\n\nH\n4\n+\n\n+\n\nN\n\nO\n2\n\n\u2192\nN\n\nH\n4\n\nN\n\nO\n2\n\n\u2192\n\nN\n2\n\n+\n\n\nH\n2\n\nO\n\n\n\n\nElectron transfer between different elements of the catalyst:\n\n(9)\n\n\n\n\nM\n\n\nn\n+\n\n\n\n+\n\n\u2212\n\nN\n\nH\n2\n\n\n\u2192\n\n\n\nM\n\n\n\n(\n\nn\n\u2212\n1\n\n)\n\n+\n\n\n\n+\n\n\u2212\n\nN\nH\n\n+\n\n\n\nH\n\n+\n\n\n\n(\n\n\n\nM\n\n\nn\n+\n\n\n=\n\nM\n\n\nn\n\n\n3\n+\n/\n4\n+\n\n\n,\n\nC\n\n\no\n\n\n3\n+\n\n\n\n)\n\n\n\n\n\n\n\n(10)\n\n\n\u2212\n\nN\nH\n\n+\n\nN\nO\n\n\n(\n\na\nd\ns\n\n)\n\n\u2192\n\nN\n2\n\nO\n\n+\n\n\n\nH\n\n+\n\n\n\n\n\n\n\n(11)\n\n\nM\n\n\nn\n\n\n4\n+\n\n\n+\n\nC\n\n\no\n\n\n2\n\n+\n\n\n\u2194\n\nM\n\n\nn\n\n\n3\n\n+\n\n\n+\n\nC\n\n\no\n\n\n3\n\n+\n\n\n\n\n\n\n\n\n(12)\n\n\nM\n\n\nn\n\n\n3\n\n+\n\n\n+\n\nT\n\n\ni\n\n\n4\n\n+\n\n\n\u2194\n\nM\n\n\nn\n\n\n4\n\n+\n\n\n+\n\nT\n\n\ni\n\n\n3\n\n+\n\n\n\n\n\n\nBlocking effect of the TiO2 shell layer on H2O and SO2:\n\n(13)\n\n\nT\ni\n\nO\n2\n\n\n+\n\nS\n\nO\n2\n\n+\n\nO\n2\n\n\u2192\nT\ni\nS\n\nO\n4\n\nO\n\n\n\n\n\n\n\n(14)\n\n\n2\nS\n\nO\n2\n\n+\n\n\nO\n2\n\n+\n\n2\nN\n\nH\n3\n\n+\n\n2\n\nH\n2\n\nO\n\n=\n\n2\n\n\n(\n\nN\n\nH\n4\n\n\n)\n\n2\n\nS\n\nO\n4\n\n\n\n\n\nThe above reaction mechanism is divided into three main parts, the first part is reaction between reactants, NH3 and NO reaction. After adsorption to the catalyst, NH3 existed in the form of -NH2 and H+. After adsorption saturation of NH3, it further reacted with NO to generate N2 and H2O. In addition to that, NO oxidizes with oxygen to form NO2, and NH3 gets an H+ to form NH4\n+, and then NH4\n+ reacts with NO2 to form the intermediate NH4NO2, the last, intermediate can further decompose N2 and H2O (corresponding to Eqs. (4) to (8) in 3.4). The second reaction mechanism partly involves electron transfer between different elements of the catalyst. Mn+(Mn+\u00a0=\u00a0Mn3+/4+, Co3+) reacts with -NH2 produced in the previous process to generate -NH and H+, and then -NH reacts with NO in the flue gas to generate N2O. In the SCR reaction of CoMn2O4@TiO2 catalyst, there is electron transfer between Mn, Co and Ti, that is Mn4++Co2+\u2194 Mn3++Co3+, Mn3++Ti4+\u2194Mn4++Ti3+(corresponding to Eqs. (9) to (12) in 3.4). The third reaction mechanism is partly the blocking effect of the TiO2 shell layer on H2O and SO2. The reason for the high resistance of CoMn2O4@TiO2 may be due to reaction between SO2 and TiO2 on the outer surface of the catalyst generates TiSO4O, thus trapping the SO2 in the flue gas outside the catalyst. In addition, SO2 might react with NH3 in the gas to form (NH4)2SO4 There may be the following reactions are TiO2\u00a0+\u00a0SO2\u00a0\u2192\u00a0TiSO4O and 2SO2\u00a0+\u00a0O2\u00a0+\u00a02NH3\u00a0+\u00a02H2O\u00a0=\u00a02(NH4)2SO4 (correspond to Eqs. (13) and (14) in 3.4).Based on the above results, the possible reaction pathway of CoMn2O4@TiO2 was proposed. Although the Langmuir-Hinshelwood (L-H) mechanism exists, it is relatively weak. The Eley-Rideal (E-R) mechanism plays a dominant role in the NH3-SCR reaction, and the electron cycle interacts with CoMn2O4@TiO2 catalyst.It is speculated that the reaction between NH4\n+ and NO2 will eventually decompose the intermediate into N2. Meanwhile, part of the -NH2 material reacts with gaseous NO to form N2 and water, while the rest is further oxidized to -NH and then reacts with gaseous NO to form N2O. The REDOX cycle is accomplished by electron transfer, transferring reactive oxygen species in SCR reactions. According to Fig.\u00a013\n, there is an electron transfer between Mn, Co and Ti. Mn4+ gains electrons to become Mn3+, Co2+ loses electrons to become Co3+, Mn3+ loses electrons to become Mn4+ and Ti4+ gains low electrons to become Ti3+.According to the analysis of catalytic activity and XPS, the catalytic activity below 125\u00a0\u00b0C at CoMn2O4@TiO2 is low, probably due to the small number of O\u03b1 species. With increasing temperature (>125\u00b0C), the catalytic activity of CoMn2O4@TiO2 is higher. This is due to stronger Lewis acid sites and higher REDOX capacity. In addition, CoMn2O4@TiO2 showed higher SO2 and enhanced water tolerance. The synthesis of core-shell structure enhanced the surface acidity and REDOX property (Mn4++Co2+\u2194Mn3++ Co3+, Mn3++Ti4+\u2194Mn4++ Ti3+) through the coating of TiO2. The reaction of NH3-SCR on CoMn2O4 and CoMn2O4@TiO2 catalysts mainly follows the typical E-R mechanism, while the L-H mechanism is weak. Considering the similar relationship between the two samples, the reason why CoMn2O4@TiO2 has higher SO2 resistance than CoMn2O4 may be its unique core-shell structure. The uniformly distributed titanium dioxide shell of CoMn2O4 core minimizes SO2 poisoning to surface active sites, thus improving the high stability and SO2 tolerance [73\u201375].In summary, a new type of CoMn2O4@TiO2 core-shell cluster centered on a nuclear sphere catalyst was successfully prepared by sol-gel method and external dynamic coating method. The optimal preparation conditions were selected by the 3D diagram obtained by the two-stage curved surface response experiment (Mn: Co= 2:1), ammonia water and active component addition ratio of 4.17, calcination temperature of 45\u00a0\u00b0C\u00a0\u00b0C\u00a0\u00b0C, reaction temperature of 225\u00a0\u00b0C). Then, under the premise of ensuring a certain high activity, the catalyst with a core-shell ratio of 2:1 was selected through SO2 tolerance test, which had high NH3-SCR catalytic activity and SO2 resistance. The catalyst prepared by this method has high activity, excellent N2 selectivity and high stability, achieving more than 95% NO conversion in the 170-225\u00a0\u00b0C range. It also enhances the resistance to SO2 and H2O in the low-temperature NH3SCR reaction, In the presence of 100\u00a0ppm SO2 for 20\u00a0h, the catalytic activity can still maintain close to 80%, and after the removal of SO2, the activity can basically recover to the initial level. The reason is uniform distribution of TiO2 shell in CoMn2O4 nucleus reduces the exposure of surface active sites to SO2 and the formation of by-products such as TiSO4O, which provides high stability and improves SO2 tolerance. TEM results showed that CoMn2O4@TiO2 catalyst had an obvious cored shell structure, and TiO2 was uniformly coated on the surface of CoMn2O4, XRD and Raman results show that the catalyst has obvious CoMn2O4 crystal phase and TiO2 crystal phase, and FT-IR results show that CoMn2O4@TiO2 catalyst has higher oxidation reducibility, XPS results show that synthesis of core-shell structure enhanced the surface acidity and REDOX property (Mn4++Co2+\u2194Mn3++ Co3+, Mn3++Ti4+\u2194Mn3++ Ti3+) through the coating of TiO2. NH3-TPD shows that the TiO2 shell leads to a catalyst with more acid sites and stronger acidity. H2-TPR shows the ease of reduction of the catalysts is in the order of CoMn2O4@TiO2 > CoMn2O4, the number of redox sites increases after doping with TiO2 shell layer although H2 consumption decreases. In situ DRIFT results show that reaction of NH3-SCR on CoMn2O4 and CoMn2O4@TiO2 catalysts mainly follows the typical E-R mechanism. On the whole, this study provides a new method to solve the SO2 poisoning problem of low temperature SCR catalyst.\nZhiyong Qi: Put forward the experimental content and design the experimental scheme. Perform experiments and write papers. Fengyu Gao: Guide the experiment process, help to revise the experiment paper, and propose the modification plan. Songjin Ko: Auxiliary operation experiment, auxiliary modification of the paper. Xiaolong Tang: Guide the experiment process, help to revise the experiment paper, and propose the modification plan. Honghong Yi: Guide the experiment process, help to revise the experiment paper, and propose the modification plan. Hengheng Liu: Auxiliary operation experiment, auxiliary modification of the paper. Ning Luo: Auxiliary operation experiment, auxiliary modification of the paper. Ying Du: Auxiliary operation experiment, auxiliary modification of the paperThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was financially supported by National Natural Science Foundation of China (U20A20130, 21806009) and Fundamental Research Funds for the Central Universities (06500152).", "descript": "\n                  In this study, the Co(3-x)MnxO4@TiO2 core-shell catalysts prepared by sol-gel and external kinetic coating method were investigated for DeNO\n                        x\n                      by response surface methodology (RSM) using 3-factor with 5-level experiments. Compared with CoMn2O4 sample, the CoMn2O4@TiO2 catalyst was optimized to obtain almost 100% NO\n                        x\n                      conversion at 125\u00b0C to 225\u00b0C that also maintained above 80% SCR activity with 100\u00a0ppm SO2 and 10 vol.% H2O in the testing time within 24 h. The XRD and Raman show a clear distinction of that CoMn2O4 spinel phase was well-dispersed on anatase TiO2 over CoMn2O4@TiO2 catalyst, which exhibited a core-shell structure with obvious distribution boundary of CoMn2O4 core coated by TiO2 shell confirmed by TEM images. This intact catalyst presented complete TiO2 coating structure due to the detected bonds Mn-O, Co-O and particular Ti-O via FT-IR analysis. NH3-TPD and H2-TPR profiles indicated that the CoMn2O4@TiO2 catalyst owned more acid sites, stronger acidity and enhanced redox capacities benefiting from its core-shell structure after TiO2 coating. According to XPS analysis, higher content of (Mn3++Mn4+)/(Mn2++Mn3++Mn4+) (77.7%), Co3+/ (Co2+\u00a0+\u00a0Co3+) (53.5%) and surface oxygen (26.1%) were in favour of valence electron interaction (Mn4++Co2+\u2194Mn3++ Co3+, Mn3++Ti4+\u2194Mn4++ Ti3+). The NH3-SCR reaction pathways over CoMn2O4 and CoMn2O4@TiO2 catalysts were compared and proposed through DRIFTS experiments, which mainly follows the typical Eley-Rideal (E-R) mechanism, while the Langmuir-Hinshelwood (L-H) mechanism is weak at 225\u00a0\u00b0C. This study opens up a new avenue for designing efficient and environment-friendly NH3-SCR catalysts and looks promising for practical application.\n               "}
{"full_text": "In 2020, based on U.S energy consumption, the energy consumption summed up to 93 quadrillion Btu. Fossil fuels (petroleum, natural gas, and coal) accounted for 79% of total. Therefore, in order to meet up with the present and the future energy challenges, green diesel is the holy grail of future energy sustainability. Also, as nations round the planet pledge to deal with emission emanated from the use of fossil fuel, here comes an idiom which is becoming new and new from nation leaders: the green hydrogen (biodiesel, propane, solar or hydro, wind, biogas, e.t.c.).Biodiesel also known as green diesel is a renewable and sustainable energy fuel obtained from transesterification of vegetable oils, animal fats, and biomass algae employing ethanol or methanol as organic solvent in the presence of salt of sodium or potassium catalyst [1,2]. With the problems of conventional fuel crisis, toxic effects, environmental problems, high mortality rate recorded against the utilization of conventional diesel, biodiesel as come to stay as the only present and future replacement for conational diesel.However, there are major difficulties to overcome before the green diesel vision can be realized. The major drawbacks against its recognition are the cost of production, low desire output, and high viscous nature. One way to reduce cost of production was the used of wastes materials as feedstocks [3\u20135]; more vivid way to tackle the problem of high viscous nature and improve the product yield is to increase the methanol-oil molar ratio (Adepoju et al., 2022) [6]. According to stoichiometric ratio, 1:3 is required for reaction to reach completion, but the yield of biodiesel always low with high waste product (glycerol). Twice ratio (1:6) of methanol-oil molar ratio and application of process modelling and optimization software have been identify as a way to tackle the problem [7\u201310].\nAsimina triloba also known as Carica papaya (CP) is a tropical fruit mainly cultivated in tropical climates (Africa, Asian and South America), the fruit contained approximately 15\u201320% seeds always discarded as a waste after fruit consumption. It has been reported that the dried seed contains 14.1\u201335% oil content depending on the species of the CP. This oil among all other vegetable oils has been reportedly to be non-edible (acid value\u00a0=\u00a03.60 mg KOH/g oil) with high degree of unsaturation [11,12]. Hence, its suitability as waste feedstock for green diesel synthesis in the presence of suitable derived base catalyst.Replacing potassium or sodium hydroxide with catalysts derived from agricultural, industrial, and domestic biomass wastes have not only served as feedstock in green diesel synthesis, but also solve the problem of wastes disposal that constituted to the environmental health challenges of plant, animal, and aquatic life [13,14]. Meanwhile, derived catalysts are categories into three categories namely; heterogeneous catalyst, heterogenized-homogeneous catalysis, and biocatalysts. Among these catalysts, heterogeneous catalyst is a solid catalyst of calcium-potassium based compound (Ca\u2013K). It application in synthesis of green diesel (transesterification process) is due to its superior properties such as ease of recoverability and reused, non-toxic, and of low cost. Heterogenized-homogeneous catalysis and Biocatalysts have major shortcomings such as slight reaction conditions, high selectivity and efficiency, in-ability to convert a cellular catalyst into a bioprocess, recoverability problem, in ability to sustainability in harsh environmental conditions, instability in aqueous media, cofactor dependability, reaction allergic, and inhibition inactivation [11,15]. It's not surprising why [16] synthesized CaO from natural waste materials and applied it to conversion of waste vegetable oils blends with hydrotreated kerosene for biodiesel production, while [17] testified the used of cocoa pod husk as heterogeneous catalyst for biodiesel production, while [18] adopted lard oil for the production of biodiesel via heterogeneous catalyst. The work of [19] further emulated the use of heterogeneous catalyst for the conversion of waste cooking oil-Calophyllum inophyllum to biodiesel, but the works reported by Refs. [20,21] utilized the green heterogeneous base catalysts as an effective bio-based for the conversion of oil to biodiesel. These showed that replacement of homogeneous catalyst with derived heterogeneous catalyst has come to stay, and wood ash powder could also serve as biobase material for biodiesel synthesis.Wood ash is the inorganic and organic residual powder remain after the wood combustion or unbleached wood fibre in fireplace, bonfire or in an industrial power plant. Generally, along with oxygen, the major components of wood ash are calcium (Ca), potassium (K), magnesium (Mg), silicon (Si), and phosphorous (P) [22,23]. Due to high Ca\u2013K base present in the residual powder, this can be used for Ca\u2013K base for green diesel synthesis with proper process modeling and optimization of process variables condition [24]. No wonder, the used of wood ash as catalyst for biodiesel production was reported by Ref. [25]. [26] utilisized wood ash for biodiesel production, while wood ash biocatalyst as a green catalyst and its application for synthesis of benzochromene derivate was reported by Ref. [27]. In another study [28], employed calcined wood ash for synthesize of biodiesel, while [29] review the application of wood ash as catalyst in various oil-biodiesel synthesis. These reports showed that catalyst can truly be obtained from the wood ash due to high percentage of calcium and potassium oxide present in its composition.Modeling and optimization of process condition for green diesel production have been reportedly proved to increase the product yield and reduce the production cost making green diesel acceptable as a replacement for fossils fuel [30,31]. Displayed in Table 1\n are the past report on the use of different software for process modeling and optimization of synthesized of biodiesel from difference feedstock (oil). It was observed that no report either past or present have modeled or optimized process conditions of biodiesel synthesis from Asimina triloba oil employing I-Optimal design/Integrated Variance. Therefore, this study obtained the optimum biodiesel yields using I-Optimal design in three-level-three-variables process conditions (reaction time, catalyst amount, and ethanol-oil ratio). Catalyst employed was derived from residual wood ash. Economic appraisal of the biodiesel synthesized was also elucidated, and the biodiesel qualities were examined by comparing with biodiesel recommended standard.Matured Asimina triloba seed was obtained from market fruits seller, in Otuoke, Bayelsa, Nigeria. The seed was washed with distil water, sun-dried for two days to semi-dried the seed, oven dried to constant weight at 100 oC for 24 h in a Genlab laboratory oven (fan assisted circulation, high accuracy Pt 100A duplex sensors<0.6 oC, temperature range-ambient +5\u2013100 oC, 8 stages profile control). The dried was separated from the chaff by winnowing, and the cleaned seed was milled into powder using a 4-liter Laboratory blender (single phase, one speed universal motor with maximum ambient temperature rating 40 oC, speed 15, 600 rpm-20,000\u00a0rpm, with working capacity 180\u20131900 mL, dimension 6.5 in), the milled seed powder was then extracted with organic solvent to produce oil. Residual wood ash was obtained from bakery located at Yenegoa, Bayelsa State, Nigeria; the ash was separated from unwanted impurities by using a stainless steel 60 mesh 250 \u03bcm. The fine residual ash wood powder (RAWP) was kept in a clean container for further processing.Chemicals used in this work were obtained from Sigma Aldrich Nigeria Limited, and were of standard graded.Extraction of oil from the powder was carried out using a continuous process method known as solvent extraction. 100 g of the powder was loaded in a muslin bag and then put in extraction chambered, the round bottom of Soxhlet extractor (500 ml capacity) was filled with 250 ml of solvent (n-hexane), and the mounted with condenser and set on a four-phase heating mantle at temperature between 68 and 70 oC for until the powder is free of oil. The oil with solvent phase in the flask was transferred to an evaporator reactor (heated at 80 oC) to allow oil made free of solvent by recycled. The recycled solvent was re-used, and the oil was filtered using funnel and filter paper (250 mm size) to remove any particles associated during extraction. The percentage yield of the oil was obtained using Eqn. (1):\n\n(1)\n\n\n\u03b2\n\n\n(\n\nw\n/\nw\n\n)\n\n%\n=\n\n\n\nm\n\no\ni\nl\n\n\n\nm\n\nc\np\ns\np\n\n\n\n\nX\n\n100\n\n\n\nWhere \n\n\u03b2\n\n= oil yield, \n\n\n\nm\n\no\ni\nl\n\n\n\nm\n\nc\np\ns\np\n\n\n\n\n = ratio of mass o oil yield to that of mass of Asimina triloba seed powder.The coated values were taken from the duplicate samples.The properties of oil such as density, moisture content, viscosity, saponification value, iodine value, acid value, and peroxide value via AOAC, 1997 [42] standard procedures. Cetane number, higher heating value and API gravity are computed using Eqns. (2)\u2013(4):Cetane number - [43].\n\n(2)\n\n\nC\ne\nt\na\nn\ne\n\nN\no\n=\n46.3\n+\n\n\n5458\n\nS\nV\n\n\n\u2212\n\n0.225\n\nI\nV\n\n\n\n\nHigher Heating Value (HHV) - [43].\n\n(3)\n\n\n\nH\nH\nV\n\n\n(\nM\nJ\n/\nk\ng\n)\n\n=\n49.43\n\u2212\n\n[\n0.041\n\n(\n\nS\nV\n\n)\n\n+\n0.015\n\n(\n\nI\nV\n\n)\n\n]\n\n\n\n\n\nAmerican Petroleum Institute (API) - [44].\n\n(4)\n\n\nA\nP\nI\n=\n\n\n141.5\n\nS\np\ne\nc\ni\nf\ni\nc\n\ng\nr\na\nv\ni\nt\ny\n\n@\n\n\n15\no\n\nc\n\n\n\u2212\n\n131.5\n\n\n\n\nThe RAWP characterization was performed using SEM (model 6300F by JEOL solution for innovation, USA) to study the study the surface topography and composition of the catalyst. The XRF (model NEX CG, made by Rugaku, Austin, USA) stimulated over K\u03ac and Cu radiation source, and to confirm the elemental analysis of the samples and the quantitative structure of the samples. FTIR (model 3116465, made in Japan) was used to examine the key functional group and confirm the existence of distinctive absorption bands of the elements in the catalyst powder. The BET isothermal adsorption method (QUANTACHROME, 1KE) and Hammett indicator was used to institute the coefficient of determination along with the pore size, pore diameter, and surface area. The procedures were as follows:Before BET analysis, Initially, the samples were preheated at 150 \u00b0C for 45 min underneath helium flow to eliminate some adsorbed loaded entities on the catalyst surfaces, then 5% CO2 was ran over the sample with helium at flow rate 25 ml/min for 40 min. The basic strength of calcined powder and the mixed powder surfaces were assessed by CO2-temperature-programmed desorption (Temperature programmed desorption (TPD) [made in BEL132 Cat, Japan]).The elemental compositions of the samples were checked by (AXS Bruker wavelength) dispersive X-ray Fluorescence (XRF) spectrophotometer with an Rh source and tube of 2.2 kW power. The specific surface area was disclosed by using Brunauer\u2013Emmett\u2013 Teller (BET) method via N2 adsorption/desorption isotherm analysis (Surface area & pore size analysis [Belsorb III, Japan]) of the catalyst was carried out on volumetric adsorption analyser at 196\u00b0C.FTIR spectrometers (Agilent Technologies Model Cary 122\u00a0630 FT-IR spectrometer) with spectral range from 400 to 4000 cm\u22121 rely on the same basic principle as NDIR analyzers, i.e., the fact that many gases absorb IR radiation at species-specific frequencies. However, FTIR spectroscopy is a disperse method, which means that measurements are performed over a broad spectrum instead of a narrow band of frequencies. This test method is performed by directing an x-ray beam at calcined samples and calcined mixed sample and measuring the scattered intensity as a function of the outgoing direction. The beam is separated, and then scattered when FTIR spectrometer directs beams of IR at the sample and measures how much of the beam and at which frequencies the sample absorbs the infrared light. The spectra were evaluated and identify based on the reference database.Scanning electron microscopy (SEM) is used to examine the morphology of the powder before consolidation. The procedure is straightforward. Two-sided carbon tape is fixed to an SEM sample stub, and the UHMWPE powder is sprinkled onto the surface. A light gold, platinum, was applied (100 \u00c5), and the samples were examined in an SEM chamber. The flakes are 50\u2013100 \u03bcm in diameter. The surface morphology was disclosed by field emission scanning electron microscopy (FE-SEM, QUANTA FEG 250).Based on acid value of CP oil, since the Acid value of the oil is greater than the recommended standard (acid value >3.0 mg KOH/g oil) for one step reaction biodiesel production, two steps production processes was adopted as follows:This stage requires the reduction in acid value of the oil. The esterification process involves the use of acid to esterify the oil to acceptable acid value (<3.0 mg KOH/g oil). The procedure reported in our work [45], with little modifications. In this case, chloric acid (HClO3) was used as an acid because of its high acid strength and ability to accelerate reaction condition to earlier completion than HCl and H2SO4. Three necked batch reactor placed on hot magnetic plate fitted with a condenser and magnetic stirrer. 100 ml of oil was measured, and preheated in the reactor on the hot plate with temperature control. 0.15 ml of HClO3 was mixed with 50 ml of methanol in a separate flask and then transferred to the preheated oil in a batch reactor. The reaction time between 30 and 40 min was sufficient to complete the reaction. The resultant product was transfer to a separating funnel for phase separation. The methanol-oil phases was washed with distil water to remove methanol. The washed esterified oil was then dried over anhydrous Na2SO4 to remove washed water, and the Na2SO4 was recovered by filtration. The acid value of esterified oil was determined using official standard method of AOAC, 1990, and the minimum acid value esterified Asimina triloba oil (EASO) was determined via response surface design experiment before transesterification (second stage).In this case, biodiesel was produced using the esterified oil (EASO) with the lowest acid value. The procedures for transesterification of the EASO to biodiesel was as reported in our recent work [46], with little modification. Ethanol was used as a solvent for the synthesis of biodiesel owing to its less chemical toxic and energy carrier ability than methanol.The reaction was carried out in a three-necked batch reactor flask. 80 ml of the cooled EASO was heated at 40 oC for 60 min for oil pretreatment, 2.5 (wt. %) RAWP was dissolved in 16 ml E-OH (ethanol) in a 250 ml flask. The mixture was transferred to preheated oil in the reactor, and the reaction temperature was monitored at 40 oC for further 40 min until reaction completion. The un-dissolved catalyst was separated from the products by decantation, the ethanol-biodiesel-glycerol phases were separated from biodiesel via separating funnel through gravity settling. The catalyst in the biodiesel was recovered by washing with heated ethanolic sodium bicarbonate, and then filtered. The wet biodiesel was dried with an inorganic drying agent (anhydrous Na2SO4). The pure biodiesel was obtained by filtration, and the yield of biodiesel was computed determined. The procedures were carried out in duplicate based on three factors considered for experimental design via response surface design. The produced biodiesel properties were determined using the [42], and the properties were compared with biodiesel recommended standard.Biodiesel production was experimentally designed in two stages with esterification to reduce the oil acid value to minimum level, and transesterification to convert the reduced oil acid value to biodiesel. In the first stage, a total of nine experimental runs were design and carried out without repetition using hybrid design. In the second stage, an I-Optimal design was adopted which generated more experimental runs than box benhken.For esterification experimental design, three factors were considered; HClO3 conc.: K1, reaction time: K2, and Oil/M\u00a0\u2212\u00a0OH ratio: K3, along with three levels to established the minimum acid value of the oil. A total of nine (9) experimental runs were designed and carried out based on varied factors design level Table 2\n depicts the range of value considered and the variables level.Process transesterification experimental design was carried out using statistical software. Since I-optimal designs provide lower average prediction variance across region of experimentation. It optimality is desirable for response surface methods (RSM) where prediction is important. The algorithm picks points that minimize the integral of the prediction variance across the design space. Hence, the design expert 13.1.4.0 with build time 2632 m.s was adopted for this process by considering three variables in three levels; catalyst amount; X1, reaction time: X2, and E-OH/oil molar ratio: X3, respectively. A total of twenty (20) experimental runs were generated and were experimentally carried out. Table 2 depicts the range of value considered and the variables level.In esterification, the results of experimental values in 9 runs were used as a based to determine the esterified oil with low acid value. Each run was carried out at different acid concentration, reaction time and methanol/oil molar ratio as displayed in Table 2. The minimum acid value was validated in triplicate and the average mean value was used for transesterification stage.Transesterification analysis involved the step by steps analysis with a view to optimize biodiesel production. Regression coefficient, test of significant, hypothesis test, multiple regression, and analysis of variance were used to test the variables significant, coefficient of determination, and interactive effects. The contour and 3-dimensional plots were used to establish the response values and operating conditions variables. The model equation that related the response (biodiesel yield) with the variables interaction is expressed with polynomial regression model of the Kth -order as presented in Eqn. (5).\n\n(5)\n\n\nY\n=\nA\n+\n\nB\n1\n\nX\n+\n\nB\n2\n\n\nX\n2\n\n+\n\nB\n3\n\n\nX\n3\n\n+\n\u2026\n+\n\nB\nK\n\n\nX\nK\n\n\n\n\n\nComparative analysis of the hybrid design and Kapla-Meier Estimator (KME) statistical software was evaluated based on the root mean square error (RMSE), coefficient of determination (R-square), adjusted (\n\n\nR\n\na\nd\nj\n.\n\n2\n\n)\n\n, and predicted (\n\n\nR\n\np\nr\ne\nd\n.\n\n2\n\n\n) expressed in Eqns. (6)\u2013(8).\n\n(6)\n\n\nR\nM\nS\nE\n=\n\n\n\n\u2211\n\n(\n\n\n\u03b4\n\ni\n,\n\nc\na\nl\n\n\n\u2212\n\n\u03b4\n\ni\n,\ne\nx\np\n\n\n\n)\n\n\nN\n\n\n\n\n\n\n\n\n(7)\n\n\n\nR\n\na\nd\nj\n.\n\n2\n\n=\n1\n\u2212\n\n\n\u2211\n\n\ni\n=\n1\n\nn\n\n\n\n(\n\n\n\n\u03b4\n\ni\n,\n\nc\na\nl\n\n\n\u2212\n\n\u03b4\n\ni\n,\ne\nx\np\n\n\n\n\n\n\u03b4\n\na\nv\ng\n,\n\ne\nx\np\n\n\n\u2212\n\n\u03b4\n\ni\n,\ne\nx\np\n\n\n\n\n)\n\n2\n\n\n\n\n\n\n\n(8)\n\n\n\nR\n\np\nr\ne\nd\n.\n\n2\n\n=\n1\n\u2212\n\n{\n\n\n\n(\n\n1\n\u2212\n\nR\n\np\nr\ne\nd\n.\n\n2\n\n\n)\n\n\n(\n\nN\n\u2212\n1\n\n)\n\n\n\nN\n\u2212\n\n(\n\nN\nI\n+\n1\n\n)\n\n\n\n}\n\n\n\n\nWith \n\n\n\u03b4\n\ni\n,\n\nc\na\nl\n\n\n\n calculated acid value, \n\n\n\u03b4\n\ni\n,\ne\nx\np\n\n\n\n experimental acid value, is the number of experimental runs, \n\n\n\u03b4\n\na\nv\ng\n,\n\ne\nx\np\n\n\n\n average experimental acid value, NI is the number of variables.\nTable 3\n displayed the properties of Asimina triloba oil extracted via solvent extraction as compared with the earlier reported on the same oil. It was observed that the properties were well within the earlier reported by other researchers. The little disparity in the results could be due to how ripe the Asimina triloba was before consumption (unripe Asimina triloba has high acidity which helps remove the bacterial that caused urinary tract infections), the varieties of the fruits, and the growth region. Naturally, Asimina triloba has about seven (7) varieties includes: sunflower, Taylor, Taytwo, Mary Foos Johnson, Mitchel, Davis, and Rebecoas Gold. However, the high acid value of the oil obtained in this study proved that the matured Asimina triloba seed come from unripe fruits, hence, the reason for its potential as feedstock for biodiesel synthesis in a two steps reaction.Since the acid value of oil (3.80 mg KOH/g oil) was found to be higher than the recommended oil (<3.00 mg KOH/g oil) standard for biodiesel production, the oil need to be esterify using a strong acid. To avoid assumption based on trial and error approach, a total of nine (9) runs were design and were carried out based on constraint variables with acid value as the output value. Table 4\n present the results obtained for the experimental runs with acid value recorded as the output of the esterified oil.From the table, the experimental runs with lowest acid value was run 6 with acid value of 1.20 (mg KOH/g oil), having an FFA\u00a0=\u00a00.60. This run with the HClO3 conc.\u00a0=\u00a00.25 (% vol.); CH3OH/Oil ratio\u00a0=\u00a05 (vol/vol); and reaction temperature\u00a0=\u00a060 (oC) was used as the condition for esterify oil transesterification for biodiesel production.The result of BET analysis carried out based on data reduction acquisition employing the DA method with nitrogen as adsorbate for sample weight of 0.12 g, revealed the plot of pore volume against pore diameter in Fig. 1\n. It was observed that the highest pore volume was recorded at 0.321 (cc/g) at corresponding value of 2.820 nm pore diameter. It was noticed that higher pore diameter produced low volume which explained while the reaction process was faster to synthesis biodiesel. Further analysis by BJ adsorption method based on liquid density with data reduction parameter evaluation showed a plot of cumulative pore volume, surface area against the pore diameter (Fig. 2\n). The overall maximum values that described the catalyst potentiality for biodiesel production were found at surface area of 441.368 m2/g, pore volume of 0.213\u00a0cc/g, and pore diameter of 2.136 nm. The pore size distribution data based on pore with, cumulative pore volume and surface area, dv/d, and ds/d, are as presented in supplementary file (Sup. 1)In order to determine the compositions of the elements present in the catalyst sample, XRS-FP analysis was carried out using Gaussian method and the results was as presented in Table 5\n. It was observed that the catalyst consist of various elements, but the major elements are CaO with concentration of 42.516 (wt. %), K2O with concentration of 12.163 (wt. %), and SiO2 with concentration of 23.942 (wt. %). Other elements are present in small quantities, but also aids in biodiesel synthesis. The high CaO indicated that the residual ash has potential to be used as alkaline base for biodiesel production, and the presence of K2O supported the strength of basicity of the catalyst. However, the presence of SiO2 showed that the residual wood ash possesses some acidic strength but SiO2 also help the production of biodiesel due to it weak base nature. The raw results as obtained from the analysis are presented in supplementary file (Sup. 2).The results of SEM analysis carried out on residual wood ash used as catalyst at magnification of 500x and 1000x are presented in Fig. 3\n(a\u2013b). Observed from the figures indicated a cohesive jointed-like crack shape with permeable surfaces. The structures showed an ice-like surface mineral particle accountable for the slick flora of the element present. The soapy like look could be attributed to the presence of high concentration of SiO2 found in the catalyst. However, the whiteness, brightness and opacity look could be due to TiO2 present. The fusion adherent effects and the brilliant glossy glaze found in the residual wood ash via XRS-FP Analysis Report could be attributed to the presence of CaO and K2O, which make it as a potential alkaline base catalyst for the synthesis of biodiesel.\nFig. 4\n represented the results obtained for the FTIR analysis of catalyst sample used for biodiesel synthesis. Various functional elements are found at different wavelengths and angular phases. Normally, the mid-IR spectrum is divided into four regions: the single bond region (2500-4000 cm\u22121), the triple bond region (2000-2500 cm\u22121), the double bond region (1500-2000 cm\u22121), and the fingerprint region (600-1500 cm\u22121) (Nandiyanto and Ragadhita, 2019). The functional groups present can be identified as follows (Coates, 2000):At fingerprint region, the following functional group can be found (i) aliphatic organohalogen compound such as C\u2013F, C\u2013Cl, C\u2013I, and C\u2013Br. (ii) the Alcohol, OH out-of-plane bend, (iii). Phenol, C\u2013O stretch, (iv) the primary, secondary, and tertiary alcohol, C\u2013O stretch, (v) the primary or secondary, OH in-plane bend, (vi) phenol or tertiary alcohol, OH bend, (vii) the peroxide, C\u2013O\u2013O-stretch, (viii) the Epoxy, oxirane rings, and Aromatic ethers, aryl-O stretch, (viii) the Alkyl-substituted ether, and Cyclic ethers with large rings, C\u2013O stretch, (ix) the primary, secondary, and tertiary, both amine and aromatic CN stretch, (x) the carboxylate salt, the P\u2013O\u2013C, aromatic and aliphatic phosphates, the carbonate ion, sulfate, nitrate, phosphate, silicate e.t.c can be found.At the double bond region, there exists the following functional group: (i) the nitrogen-oxy compounds, the open-chain imino \u2013C = N-, open chain azo \u2013N = N- (ii) the carbonyl compound such as: ketones, carboxylic acid, aldehydes, Ester, Amide, Acid halide, Aryl carbonate (iii) the primary and secondary amine\u00a0>N\u2013H bend, (iv) Aromatic ring (aryl) such as CC\u2013C aromatic ring stretch (v) the olefinic (alkene) such as Alkenyl CC stretch, aryl substituted CC, conjugated CC.At the triple bond region (2000-2500 cm\u22121), the following functional groups can be found: (i) Acetylenic (alkyne) such as C \n\n\u2261\nC\n\n terminal and medial alkyne, (ii) the transition metal carbonyl, (iii) the Ester carbonyl (iv) the Nitrogen multiple and cumulated double bond compound such as Thiocyanate (-SCN), Isocynate (-NCO asym. stretch), Cyanate (-OCN and C\u2013OCN stratch), aromatic and aliphatic cyanide, (iv) the ether and oxy compound of methoxy, C\u2013H stretch (CH3\u2013O-).At the single bond region (2500-4000 cm\u22121), there exist the functional groups such as (i) Alkyne C\u2013H stretch, (ii) Olefinic (alkene) such as terminal (vinyl) C\u2013H stretch, the pendant (vinylidene) C\u2013H stretch, medial, cis-or trans-C-H stretch, (iii) Saturated aliphatic (alkene/alkyl) such as methyl C\u2013H asym./asym stretch, methylene C\u2013H asym./sym stretch, methyne C\u2013H stretch, methoxy, methyl ether O\u2013CH3, C\u2013H stretch, methylamino, N\u2013CH3, C\u2013H stretch (iv) The Acetylenic (alkyne) such as alkyne C\u2013H stretch (v) the alcohol and hydroxyl compound such as hydroxyl group, H-bonded OH stretch, normal polymeric OH stretch, Dimeric OH stretch, internally bonded and non -bonded hydroxyl group, OH stretch, primary, secondary, tertiary alcohol, OH stretch, phenols, OH stretch, (vi) Ether and oxy compound of methoxy, C\u2013H stretch (CH3\u2013O-) (vii) Primary and secondary amino of aliphatic and aromatic primary amine NHH stretch, aliphatic and aromatic secondary amine\u00a0>N\u2013H stretch, heterocyclic amine\u00a0>N\u2013H stretch, and imino compounds =N\u2013H stretch (vii) The thiols of S\u2013H stretch, (viii) the common inorganic ions such as ammonium ion.The wavelength peak found in this study are within the aforementioned regions, therefore, it can be concluded that the analysis of the calcinated wood ash as catalyst for biodiesel synthesis was viable.Further analysis via quantitative analysis on the residual wood ash used as catalyst based on phase data view showed the plot of intensity against the angular phase diagram. The plot showed a zig-zag plot with graphite as the major compound containing carbon as the major element. Other compound such as Quartz, Adamite, Gahnite, Zincite, and Willemite are present with figure of merit based on weight fraction via the analysis (Fig. 5\n.). The presences of quartz identify in the catalyst represent the catalytic current which helped in oscillatory vibration of the phases. The graphite identify in the catalyst responsible for high tension lubrication responsible for low viscous product. The presence of adamite in the catalyst helped in formation of biodiesel colour, usually adamite always yellow in colour which responsible for light yellowish biodiesel produced in this work. Other compounds also aids in biodiesel formation during production.\nTable 6\n presented the I-optimal twenty (20) designs of constraint variable runs, the experimental yield, the residual value (errors), and the leverages. It was observed that the maximum biodiesel yield was obtained at run 3 (98.77% vo./vol.), while the minimum yield was obtained at run 17 with a yield of 84.00% (vo./vol.). This proved that the conversion of Asimina triloba seed oil to biodiesel via two steps reaction processes was successful, and the catalyst derived from residual wood ash is suitable for the synthesis. The experimental values and the predicted value were displayed graphically in Fig. 6\n(a), which shows an absolute perfect straight line with no intercept, passing through the origin with all points lie in the line. This also shows that the experimental results via.average value recorded helped in optimal design prediction. The graph of residual plotted against the runs showed a sinusoidal waveform with amplitude of \u00b12.1, indicated that the experimental runs design were design with both ends equity (Fig. 6(b)).Based on Test of significant, Table 7\n represented the full results obtained using I-optimal design, it was noted that all the variables were significant at p-value< 0.005 including the model value. This indicated that the coefficients of determination were found close to 100%. The value of the.Predicted R2 of 99.07% is in reasonable agreement with the Adjusted R2 of 99.78%; i.e. the difference is less than 2%. The model equation based on the coefficient estimate which represents the expected change in response per unit change in factor value when all remaining factors are held constant are presented in Eqns. (9) and (10):In terms of coded value\n\n(9)\n\n\n\nExp\n.\n\u00a0Yield\u00a0\n\n(\n%\nv\no\nl\n.\n/\nv\no\nl\n.\n)\n\n=\n+\n89.84\n+\n5.83\n\nX\n1\n\n+\n2.00\n\nX\n2\n\n\u2212\n0.4321\n\nX\n3\n\n+\n1.36\n\nX\n1\n\n\nX\n2\n\n\n\n+\n0.2902\n\nX\n1\n\n\nX\n\n3\n\n\n\n+\n0.3966\n\nX\n2\n\n\nX\n3\n\n+\n1.91\n\nX\n1\n2\n\n\u2212\n2.71\n\nX\n2\n2\n\n+\n0.4578\n\nX\n3\n2\n\n\n\n\n\n\nIn terms of actual value\n\n(10)\n\n\n\nExp\n.\n\u00a0Yield\u00a0\n\n(\n%\nv\no\nl\n.\n/\nv\no\nl\n.\n)\n\n=\n+\n89.83572\n\u2212\n2.32016\n\nX\n1\n\n+\n50.66581\n\nX\n2\n\n\u2212\n9.75584\n\nX\n3\n\n\n\n+\n0.271285\n\nX\n1\n\n\nX\n2\n\n+\n0.02901\n\nX\n1\n\n\nX\n\n3\n\n\n\n+\n0.793211\n\nX\n2\n\n\nX\n3\n\n+\n0.019149\n\nX\n1\n2\n\n\u2212\n10.83093\n\nX\n2\n2\n\n+\n0.47773\n\nX\n3\n2\n\n\n\n\n\n\nThe equation in terms of coded and actual factors can be used to make predictions about the response for given levels of each factor. By default, the high levels of the factors are coded as +1 and the low levels are coded as \u22121 for coded and in actual, the levels should be specified in the original units for each factor. The coded equation is useful for identifying the relative impact of the factors by comparing the factor coefficients whereas the actual equation should not be used to determine the relative impact of each factor because the coefficients are scaled to accommodate the units of each factor and the intercept is not at the center of the design space.Based on each factor significant contribution towards the response (biodiesel) with p-value highly significant, the significant of each variable on the response are displayed in Fig. 7\n(a\u2013c). It was noted that higher reaction time, closely high catalyst amount, and high ethanol to oil molar ratio favoured the response.Further on variable effects on the response, the model quadratic response effects on the response are presented in 3-dimensional contour form displayed in Fig. 8\n(a\u2013c). The statistical analysis by confirmation and prediction, predicted a yield of 98.93% (vol/vol) at the following variables conditions; X1\u00a0=\u00a059.976 (min), X2\u00a0=\u00a03.204 (% wt.); X3\u00a0=\u00a06.979 (vol/vol). This value was validated via three experimental runs; an average value of 98.87% (vol/vol) was obtained. These also proved that I-optimal designs provide lower average prediction variance across the region of experimentation in this study, which is desirable for response surface methods (RSM) where prediction and optimum validation are important. The algorithm picks points that minimize the integral of the prediction variance across the design space.To examine the qualities of biodiesel produced from Asimina triloba seed oil using residual wood ash as catalyst, the physicochemical properties of the diesel was examined and the results were compared with biodiesel recommended standard- ASTM D6751 [49] and EN 14214 [50]. Observations from the table (Table 8\n) indicated that the properties of the Asimina triloba biodiesel produced are well within the biodiesel recommended standards. Nearly all the properties of oil have reduced in values as the conversion through two steps processes have proved to be effects ways for synthesis of biodiesel with an active catalyst from residual waste ash (see Table 9).To ascertain the content and composition of esters present in the biodiesel, the characteristics of biodiesel were determined by GC-MS. Table 9\n displayed the presence of methyl ester compound present in the biodiesel sample, the retention time, and the molecular weight, while the peaks are attached in supplementary file. Observation from the table indicated that the conversion of Asimina triloba oil to biodiesel in two step approached using wood ash as precursor for the synthesis of CaO\u2013K2O \u2013SiO2 base catalyst was efficacious carried out.The methyl ester compounds found in the biodiesel formed are methyl tridecanoate, tetradecanoic, methyl myristate, methyl palmitoleate, methyl palmitate, methyl ester, methyl elaidate, methyl stearate and methyl elaidate. These results are in agreement with the previous study by Ref. [50], that the GC-MS fatty acid composition of Carica papaya seed oil contained oleic acid, palmitic acid, lauric acid, stearic acid, hexadecenoic acid, linoleic acid, and myristic acid.\nTable 9\n\n indicated the results of syntheses of FAME using developed catalyst from biomass sources, the percentage yields and the percentage minerals content. Observation from the table showed that the biodiesel conversion yield was high with respect to the concentration of CaO found within the source. The results further showed that most reports derived only CaO from the biomass source which account for low and high yield of biodiesel reported in various work, but the analysis of the wood ash used in this study account for the presence of other compound such as SiO2 and K2O which support the strength of the CaO catalyst during production also found in wood ash. As earlier said SiO2 and K2O are also base and aids in formation of biodiesel. Also, the calcined temperatures reported by researchers were noticed to be around 105\u2013900 oC, these temperature were less than the thermal temperature treatment of 1000 oC reportedly used in this study. Also, the process optimization used for experimental design and validation produced the optimum yield, whereas, most of the biodiesel yield reported by the authors were not validated by any optimization software like I-optimal design adopted in this study; the yields recorded were maximum experimental yields. Therefore, the CaO\u2013 SiO2\u2013K2O based catalyst derived from wood ash in this study proved to be suitable Bioresource catalyst for conversion of Asimina triloba oil to biodiesel.To examine the strength of catalyst developed from residual wood ash, the used catalyst after production was recycled, and the impurities in the catalyst was removed by washing primary alcohol, centrifuge for 30 min at 3500\u00a0rpm, and then filtered. The residual catalyst was oven dried to constant weight and then cools at room temperature for 40 min. The obtained purified catalyst was reused at the maximum biodiesel yield and variables condition (reaction time of 50 min, catalyst amount of 5.50 (%wt.), and 7:1 (vol/vol) EtOH/OMR).The results obtained at vary reusability test were presented in Fig. 9\n using Microsoft excel plot. Catalyst reusability test showed the biodiesel yield approaches \u223c100% at first-three cycles (1\u20133 cycles). This could be attributed to freshly active pore site, pore diameter and large surface area of the catalyst. At 4th to 6th cycles, there was no significant reduction in biodiesel yield <10% reduction. This reduction can be due to the adherent of impurities at the surface of the catalyst during reaction and formation of by-product. At 7th cycle, there exists greater reduction in biodiesel conversion to 90.20% owing to formation of glycerol and ethanol covering the active catalyst sites, hereby rendering the activities of catalyst minima. The 8th-10th cycles the conversion yield reduced greatly (87-80%), this proved that the catalyst interface is completely less active as results of high impurities and the need to introduce new freshly made catalyst for biodiesel conversion cost effectiveness. Hence, catalyst reusability test was altered at 7th cycles. This phenomenon exhibited by wood ash catalyst can be attributed to nonstop intermediate products of mono/diglyceride formed during reaction which blocked the catalyst hovels, and the development of water-oxygen reaction that takes place at the catalyst superficial, hereby reduces the catalyst sensitivity.Meanwhile, the biodiesel conversion yield from wood ash catalyst after 7th cycle may be low, but the yields are still better than most homogeneous catalyst reportedly used for biodiesel production [12,40,57,58], but for the purpose of cost estimation and acceptability of biodiesel as replacement for conventional diesel, the conversion yield must be targeted to optimum value.As earlier reported, the main factor affecting the full application of biodiesel nationwide is its cost of production, hence, the need for its economic appraisal. Generally, the materials used for the production of biodiesel in this study are wastes obtained freely from nearby locations. The key main factors (Table 9) to be considered in cost estimation in developing new solid base catalysts are wastes assemblies; refinement/planning and categorizations [14,32]\nBasis: 50 L of biodiesel.The cost of production was estimated in Nigeria Naira (N) by Eqn. (5):\n\n\n(11)\n\n\n\u2229\n\n(\nN\n)\n\n=\n\u03b1\n+\n\u03b2\n+\n\u03b3\n+\n\u03bc\n+\n\u03b5\n+\n\u03b8\n\n\n\n\nTherefore, \n\n\n\u2229\n\n(\nN\n)\n\n\n is N 700.00, which is the cost of producing 50 litre of biodiesel using wastes materials. This is equivalent to $1.68 as at 16th June 2022 (N 1\u00a0=\u00a0$0.0024).Comparing this result with the previously reported works ([59] reported $46.34/kg [32], reported $3.01/kg for 10 L), it can be observed that this is the best result reported so far on cost of biodiesel production ($1.68/kg for 50 L of biodiesel). This proved that replacing conventional diesel with biodiesel/green diesel is highly economical, environmentally friendly and always available [32].In this work, biodiesel was synthesized from Asimina triloba waste seed oil. The oil was obtained from solvent extraction of seeds. The acid value of the oil was high, and was esterified in the first stage by strong acid, HClO3 in ten experimental runs. The esterified oil was converted to biodiesel in second stage using a residual wood ash as catalyst. Catalyst analysis and characterization showed that the produced catalyst has high basic strength with potassium and calcium as major elements. Catalyst reusability test showed that the residual wood ash catalyst can be recycled and used in seven cycles. Statistical optimization via I-Optimal design established the optimum biodiesel yield of 98.87% (vol/vol) at stable constraint variables. Economic appraisal showed that the wastes adopted in this study proved to be the best materials for biodiesel production. The produced biodiesel have some fuel properties when compared with biodiesel standard.No research funds available for this work.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Authors acknowledge the effort of Lillian for her input in this work. The efforts of technical staff of Spectral Laboratory Services of Engineering and Science Analyses, No. 14 Forte Oil Station Polytechnic Road, Tudunwada Kaduna South, Kaduna, Nigeria are highly appreciated.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.cscee.2022.100252.", "descript": "\n                  This study critically examined the visibility of Asimina triloba oil for synthesis of biodiesel in the presence of ethanolic CaO\u2013K2O \u2013SiO2 base catalyst developed from residual wood ash powder. The oil was extracted using continuous extraction method. The physicochemical properties of the oil was determined for its production routes Two steps reaction process were adopted; first to lower the acid value of Asimina triloba oil by considering three factors namely; HClO3 conc., M-OH/Oil molar ratio, and reaction time. The second step convert the esterified oil to biodiesel by considering three variables namely: reaction time: X1, catalyst amount: X2, and E-OH/oil molar ratio: X3 as variables constraint using I-Optimal design by. Catalyst obtained from (RWA) was characterized using SEM, FTIR, XRF-FT, BET isothermal adsorption, and qualitative analysis. Catalyst reusability, and economical appraisal of the biodiesel synthesized were also examined. The results showed that Asimina triloba seed is rich in oil with 40.34%, and the oil is highly unsaturated with 88.90%, having high acid value of 3.80 mg KOH/g oil. Based on first step, the acid value of the oil was reduced to 1.20 mg KOH/g oil which was used as esterified oil in the second stage. A maximum experimental biodiesel yield of 98.73% (vol/vol) was obtained, but I-optimal design predicted 98.93% (vol/vol) yield at the following variables conditions; X1\u00a0=\u00a059.976 (min), X2\u00a0=\u00a03.204 (% wt.); X3\u00a0=\u00a06.979 (vol/vol) which was validated as 98.87% (vol/vol). Catalyst characterization showed that the RWA contained high amount of CaO of 42.516 (% wt.), K2O of 12.168 (% wt.), and SiO2 of 23.942 (% wt.) which serve as base catalyst. Catalyst reusability showed that the catalyst RWA degradation effects start at 8th cycles, and the economic appraisal showed the production of biodiesel using Asimina triloba oil is cost effective fuel properties in line with biofuel standard.\n               "}
{"full_text": "As a clean and efficient energy source, hydrogen energy is pursued after by scientists [1]. The use of solar energy to conduct photoelectrochemical (PEC) catalysis is a highly prospect way to produce renewable energy [2]. Thus, many metal-based compounds are exploited for the realization of potent reaction efficiency [3\u20135]. Among them, BiVO4 can be regarded as a typical photoanode material for PEC water splitting and has attracted substantial attentions due to its satisfactory light absorption and high theoretical photocurrent density [6\u20138]. However, the recombination of photogenerated electrons and holes on the electrode surface and the slow reaction kinetics often make the photocurrent density of BiVO4 lower than its theoretical value, thereafter constricting its practical application in photoelectrochemical water splitting [9\u201311].In order to solve the drawbacks of BiVO4, many studies raised up a variety of effective strategies to promote charge separation and suppression of charge recombination [12,13]. In addition to element doping and morphology control to restrain charge recombination [14\u201317], proper co-catalyst loaded can not only augment the light absorption of electrode materials, but also enhance the kinetics of water oxidation reaction [18,19]. For instances, Kim et\u00a0al. successfully prepared a NiOOH/FeOOH/BiVO4 electrode and obtained a photocurrent density of 4.5\u00a0\u200bmA/cm2 with Na2SO3 hole scavenger [20]. Based on this, Zhang et\u00a0al. prepared an ultra-thin \u03b2-FeOOH nano-layer rich in oxygen vacancies. The obtained FeOOH/BiVO4 photoanode showed excellent PEC performance for water oxidation, with 4.3\u00a0\u200bmA/cm2 photocurrent obtained in Na2SO4 solution (At 1.23\u00a0\u200bV vs RHE, AM 1.5\u00a0\u200bG illumination) [21]. Rong et\u00a0al. took the lead in synthesizing NiFeOOH on carbon nanotubes and confirmed that this structure can expose more active sites. The synergistic effect of carbon nanotubes and NiFeOOH also endowed the composite with excellent stability [22]. Recently, Fang et\u00a0al. constructed dual co-catalyst systems using NiFeOOH and Co\u2212Pi to form a heterojunction with BiVO4. The core\u2212shell structure of BiVO4/NiFeOOH/Co\u2212Pi electrode displayed excellent photocurrent density (2.03\u00a0\u200bmA/cm2 at 1.23\u00a0\u200bV vs RHE) [23]. And Co\u2212Pi has been commonly regarded as one of the representatives of high-efficiency cobalt-based catalysts, and it can form a stable and effective catalytic system with a variety of materials [24,25]. However, the photocurrent density of BiVO4/NiFeOOH/Co\u2212Pi is still lower than the theoretical value of BiVO4. To this end, finding a more effective and easily synthesized photoanode remains a prerequisite in PEC water splitting. We turned our attention to other cobalt-based salts, hoping to find a more suitable co-catalyst. After trying to load cobalt silicate on NiFeOOH/BiVO4 (Co\u2212Sil/NiFeOOH/BiVO4) [26], we noted Co\u2212Ci co-catalyst can increase the carrier concentration in CO2 reduction and perovskite solar cells [27\u201329]. Moreover, Yadav and AmalenduChandra studied the dynamic characteristics of carbonate ion in aqueous solution by using the dispersion-corrected density functional theory [30]. The results showed that there was a strong hydrogen bond between carbonate and metal ions, which made the structure more tightened. Therefore, cobalt\u2212carbonate (Co\u2212Ci) co-catalyst is considered one of the candidates for high-efficiency cobalt-based catalysts.Herein, inspired by the above works, we introduced cobalt-based co-catalysts into NiFeOOH/BiVO4, and obtained respectively Co\u2212Ci/NiFeOOH/BiVO4, Co\u2212Pi/NiFeOOH/BiVO4 and Co\u2212Sil/NiFeOOH/BiVO4 electrodes. Three composite electrodes all show higher PEC performance than BiVO4. However, it is obvious that the Co\u2212Ci/NiFeOOH/BiVO4 photoanode has achieved the most excellent performance. The photocurrent density of 4.1\u00a0\u200bmA/cm2 is obtained in 0.5\u00a0\u200bmol/L Na2SO4 solution, and both the injection and separation efficiency have also been enhanced. As a co-catalyst, NiFeOOH plays a role of hole transport layer, making the separated holes more easily captured by Co\u2212Ci. It can also be used as a passivation layer on the surface of BiVO4, because NiFeOOH makes the charge recombination lower at the interface. The promotional effect of Co\u2212Ci is due to rich oxygen vacancies in the thin Co\u2212Ci layer that improve the PEC performance of the photoanode. Compared with Co\u2212Pi and Co\u2212Sil, it also can better match the NiFeOOH/BiVO4 interface. Interestingly, we note that among the three different co-catalysts, carbonate ion has a smaller ionic radius than phosphate and silicate ion, suggesting that the charge around Co\u2212Ci may be more intensive and can accelerate the charge transfer at the interface. The more compact structure of the carbonate allows Co\u2212Ci to better match the NiFeOOH/BiVO4 interface and stimulate the performance of the electrode. As an unpopular co-catalyst in the field of photoelectrochemical water splitting, Co\u2212Ci has brought unexpected performance to BiVO4 photoanodes, and also provides potential application prospects for other photoanodes. In the follow-up work, perfecting the Co\u2212Ci reaction mechanism is an urgent task, and various carbon-based catalysts should be also received continuous attention.Na2SiO3\u00b79H2O and KHCO3 were purchased from Tianjin Kaixin Chemical Co., Ltd. Na2HPO4 was purchased from Shanghai Zhongqin Chemical Reagent Co., Ltd. FeCl2\u00b74H2O, NaOH, Na2SO3, KI, Ni(NO3)2\u00b76H2O, Co(NO3)2\u00b76H2O, Bi(NO3)2\u00b75H2O, Na2SO4 and Na2H2PO4\u00b72H2O were purchased from Sinopharm Chemical Reagent Co., Ltd. p-benzoquinone and vanadium acetylacetonate VO(acac)2 were purchased from Shanghai Aladdin Biochemical Co., Ltd. Fluorine-doped tin oxide (FTO) conductive glass was purchased from Zhuhai Kaiwei Photoelectric Technology Co., Ltd. Hexamethylenetetramine (C6H12N4) was purchased from Yantai Shuangshuang Chemical Co., Ltd. Dimethyl sulfoxide was purchased from Tianjin Damao Chemical Preparation Plant. All chemicals are analytically pure, and deionized (DI) water was used for all experiments.Using the BiVO4 (please refer to the support information for specific steps) prepared by electrodeposition of the classic three-electrode system [20], BiVO4/NiFeOOH was obtained by a simple solution impregnation method. Two aqueous solutions of 2\u00a0\u200bmmol/L Ni(NO3)2\u00b76H2O and FeCl2\u00b74H2O were prepared, and 5\u00a0\u200bmmol/L hexamethylenetetramine was dissolved in the above solution. Then, two aqueous solutions were mixed in a culture flask at a volume ratio of 1:2 and BiVO4 was placed in the culture flask obliquely. Finally, an equal volume of NaOH (20\u00a0\u200bmmol/L) was poured into the bottle and it was stood for 10\u00a0\u200bmin [23].For the Co\u2212Ci co-catalyst, a simple and efficient PED method was adopted, using 0.3\u00a0\u200bmmol/L Co(NO3)2\u00b76H2O as a cobalt source and dissolving it in 0.1\u00a0\u200bmol/L potassium bicarbonate [29]. The constant-potential electrodeposition method was adopted, the starting voltage is \u22120.1\u00a0\u200bV, and the deposition time is 360\u00a0\u200bs (the light source is a xenon lamp that simulates sunlight AM 1.5\u00a0\u200bG 100\u00a0\u200bmW/cm2). Co\u2212Sil or Co\u2212Pi co-catalysts also followed the similar steps, using 0.3\u00a0\u200bmmol/L Co(NO3)2\u00b76H2O as the cobalt source to dissolve in NaSiO3 solution and pH\u00a0\u200b7 phosphoric acid buffer, respectively. Co\u2212Sil was prepared by PED for 15\u00a0\u200bs, while Co\u2212Pi was deposited by electrodeposition for 300\u00a0\u200bs.All the instruments used in the characterization can be seen in the supplementary information.In the photoelectrochemical performance test of all photoanodes, we used CHI 660D and CHI 760E electrochemical workstation (Shanghai Chenhua Instrument Co., Ltd.) to build a three-electrode system . The prepared photoanode was the working electrode (WE), Pt electrode was the counter electrode [31], and Ag/AgCl (3.5\u00a0\u200bmol/L KCl) electrode was the reference electrode (RE). The electrolyte used in this work was 0.5\u00a0\u200bmol/L Na2SO4 (pH=7.35) aqueous solution. The light source was a xenon lamp produced by Beijing Zhongjiao Jinyuan Technology Co., Ltd. combined with an AM 1.5\u00a0\u200bG filter, and the light power density on the surface of each composite material was equal (the calibration was 100\u00a0\u200bmW/cm2). Gas chromatograph was used to understand the water splitting performance of the composite electrode. In addition, the incident photocurrent conversion efficiency (IPCE) was achieved with a xenon lamp (PLS-SXE300C) equipped with a monochromator (71SWS, Beijing NBeT Technology Co., Ltd.). In all photoelectric tests, the working electrode was backlit.In the first place, BiOI was obtained on FTO conductive glass by using the classical three-electrode system, and BiVO4 was prepared according to the previous reports. We can see clearly in the scanning electron microscope (SEM) that the thin slice BiOI with voids and the interlaced worm-like BiVO4 were obtained by calcining (Fig.\u00a01\n a and b). As shown in Fig.\u00a01c, after loading NiFeOOH on BiVO4, the smooth white particles adhered on the tail side of BiVO4 became rough and made the structure more tightened [32,33]. Fig.\u00a0S1 (a, b, c) show the SEM images of Co\u2212Pi/NiFeOOH/BiVO4, Co\u2212Sil/NiFeOOH/BiVO4, Co\u2212Ci/NiFeOOH/BiVO4, respectively. Furthermore, there is no obvious difference among the above electrodes. Fig.\u00a01d shows a flower-like nanocluster structure with a radius of about 1\u00a0\u200b\u03bcm in Co\u2212Ci/NiFeOOH/BiVO4, which can also be seen in NiFeOOH/BiVO4, Co\u2212Pi/NiFeOOH/BiVO4 and Co\u2212Sil/NiFeOOH/BiVO4. From the SEM\u2212EDS element images of Co\u2212Ci/NiFeOOH/BiVO4, it can be clearly seen that Bi, V, O, C, Fe, Ni and Co are uniformly distributed, which proves the successful preparation of this composite electrode (Fig.\u00a02\na). In addition, the SEM\u2212EDS element images of NiFeOOH/BiVO4, Co\u2212Pi/NiFeOOH/BiVO4, and Co\u2212Sil/NiFeOOH/BiVO4 can all be confirmed in Fig.\u00a0S2, Fig.\u00a0S3, Fig.\u00a0S4. The internal structure of Co\u2212Ci/NiFeOOH/BiVO4 was further examined with a high-resolution transmission microscope (HR-TEM). 0.312\u00a0\u200bnm and 0.342\u00a0\u200bnm correspond to the BiVO4 (\u2212121) and NiFeOOH (120) crystal faces, respectively (Fig.\u00a02 b and c) [23]. The amorphous layer at about 4\u00a0\u200bnm was observed from the region contained in the dashed line profile, and combined with X-ray diffraction (XRD) analysis, it can be considered that Co\u2212Ci is amorphous. XRD is usually used to analyze the crystal structure of the material. In Fig.\u00a0S5a, it is obviously observed that the peaks at 2\u03b8=27\u00b0 and 33.9\u00b0 correspond to (110) and (101) planes of SnO2 (JCPDS No. 46\u20131088) in FTO conductive glass, respectively. The 2\u03b8 values of 18.2\u00b0, 18.6\u00b0, and 28.8\u00b0 are attributed to (110), (011) and (\u2212121) planes of monoclinic scheelite BiVO4 (JCPDS: 14\u20130688). Other peaks also basically have crystal planes corresponding to the above two substances. Although the diffraction peaks of \u03b1-NiOOH (JCPDS No. 06\u20130075) and \u03b1-FeOOH (JCPDS No. 29\u20130713) are blocked by SnO2 and BiVO4, but the related signals can still be found in NiFeOOH/BiVO4, Co\u2212Pi/NiFeOOH/BiVO4, Co\u2212Sil/NiFeOOH/BiVO4 and Co\u2212Ci/NiFeOOH/BiVO4. Due to amorphous properties of the cocatalyst such as Co\u2212Ci, no obvious characteristic peaks were found in three different cocatalysts of Co\u2212Ci, Co\u2212Sil and Co\u2212Pi, which was consistent with the results of TEM analysis and previous reports [26,34]. In the XRD spectrum, only the characteristic diffraction peak at 2\u03b8=28.8\u00b0 of BiVO4 was observed to slightly move to the high diffraction angle after loading the co-catalyst, suggesting that the crystal structure of BiVO4 was not destroyed in the system composed of NiFeOOH and Co\u2212Ci (Co\u2212Sil or Co\u2212Pi) double-layer co-catalysts.X-ray photoelectron spectroscopy (XPS) can be used to prove the composition and chemical valence of elements, and can further prove the successful preparation of Co\u2212Ci/NiFeOOH/BiVO4. Fig.\u00a0S6 (a-c) shows the XPS whole patterns of BiVO4, NiFeOOH/BiVO4 and Co\u2212Ci/NiFeOOH/BiVO4, and all the elements are marked. The element composition is consistent with the SEM\u2212EDS results. Fig.\u00a03\n shows the detailed high-resolution XPS spectra, clearly showing element composition and the changes of chemical bond state. Bi 4f7/2 and Bi 4f5/2 in BiVO4 are located at 158.5\u00a0\u200beV and 163.9\u00a0\u200beV, respectively. After co-catalyst was loaded, the two peaks shifted slightly to the low angle direction (Fig.\u00a03a). Similarly, the peak positions of V 2p3/2 and V 2p1/2 also slightly changed (Fig.\u00a03b). As for the O 1s peak, in BiVO4, it can be well fitted for two types of oxygen. One is 529.3 eV O2\u2212, which is caused by the combination of oxygen atoms and metals. The other is 531.9\u00a0\u200beV adsorbed oxygen (H\u2014O\u2014H) related to hydroxyl groups in adsorbed water molecules (Fig.\u00a03c) [35]. In NiFeOOH/BiVO4 and Co\u2212Ci/NiFeOOH/BiVO4, O 1s can be clearly identified by three peaks, and the newly existing signal is at about 530.7\u00a0\u200beV. This peak is related to oxygen vacancies, confirming the co-catalyst is helpful to produce many oxygen vacancies, which can better promote water oxidation [21]. Fig.\u00a03d is fitted with two peaks at 855.8\u00a0\u200beV and 872.9\u00a0\u200beV, and is accompanied by a satellite peak. These two peaks correspond to Ni 2p3/2 and Ni 2p1/2, respectively, proving that Ni exists in the form of +2. The Fe 2p spectrum shows that Fe exists in the form of +3. The peaks for NiFeOOH/BiVO4 at 711.3 eV and 724.8 eV correspond to Fe 2p3/2 and Fe 2p1/2, and 716.3 eV and 730.8 eV are two satellite peaks (Fig.\u00a03e). By comparing the peak areas of Ni 2p and Fe 2p (\u22481:2), the existence of NiFeOOH is proven [23]. Finally, the two characteristic peaks representing Co 2p3/2 and Co 2p1/2 are located at 779.2\u00a0\u200beV and 796.2\u00a0\u200beV, respectively (Fig.\u00a03f). Co 2p3/2 can be fitted to two peaks of 779\u00a0\u200beV and 782.7\u00a0\u200beV, which are attributed to Co3+ and Co2+, respectively, indicating Co3+ coexists with Co2+ in Co\u2212Ci/NiFeOOH/BiVO4. Combined with C 1S in Fig.\u00a0S6d, the preparation of Co\u2212Ci is successful [27,36]. In addition, UV\u2013Vis diffuse reflectance is used to compare the absorbance and absorption boundary of different materials to infer the optical properties of the test sample. As shown in Fig.\u00a04\na, the absorbance of NiFeOOH is slightly enhanced compared to BiVO4, and the absorption edge is almost unchanged. After loading three different co-catalysts, Co\u2212Ci, Co\u2212Sil, and Co\u2212Pi, the three dual-co-catalyst-loaded electrodes have better absorbance, indicating that the light absorption capacity of the composite electrode has been enhanced. Compared with BiVO4 and NiFeOOH/BiVO4, the absorption edges of the three composite electrodes have obvious red shift, which proves that the synergistic effect of double layer co-catalyst can broaden the light absorption region and better capture and utilize light. The light harvesting efficiency (LHE) in Fig.\u00a0S7a reveals that the three photoanodes of Co\u2212Ci/NiFeOOH/BiVO4, Co\u2212Sil/NiFeOOH/BiVO4, and Co\u2212Pi/NiFeOOH/BiVO4 absorb more than 80% of the light. In Fig.\u00a04b, we determine that the band gap of BiVO4 is about 2.41\u00a0\u200beV by extending the slope of Tauc curve to X axis, which is basically consistent with the known band gap of BiVO4. As the basic characteristics of semiconductors, the band gaps of different catalysts generally do not change. In the subsequent Mott\u2212Schottky figure, the flat band potential (V\nfb) of the composite electrode is known by linear fitting, which can better explain the excellent catalytic performance of Co\u2212Ci in water oxidation [37].Photoanode oxygen evolution reaction (OER) has the problem of slow reaction kinetics. The reaction kinetics can be accelerated by increasing oxygen vacancies and preparing ultrathin co-catalysts [38,39]. Electrochemical impedance spectroscopy (EIS) was used to clarify the interface charge transfer resistance of Co\u2212Ci/NiFeOOH/BiVO4, Co\u2212Sil/NiFeOOH/BiVO4, Co\u2212Pi/NiFeOOH/BiVO4, NiFeOOH/BiVO4 and BiVO4. It can be observed in Fig.\u00a05\n (a and b) that the semicircle diameter of Co\u2212Ci/NiFeOOH/BiVO4 photoanode is small, indicating its optimal charge transfer ability. In the equivalent circuit fitting, the intercept between the curve and the X axis represents the equivalent series resistance [10], and the semicircle reflects the interfacial charge transfer resistance (R\nct) between the electrode and electrolyte [40]. Co\u2212Ci/NiFeOOH/BiVO4 obtains 2.4845\u00a0\u200b\u03a9 [10] and 108.9\u00a0\u200b\u03a9 (R\nct), which are much smaller than those of other electrodes (Table. S1). It is proven that Co\u2212Ci can effectively improve the transfer rate of photo-induced carrier compared with Co\u2212Pi and Co\u2212Sil, since the radius of Co\u2212Ci is smaller, and thus can match the NiFeOOH/BiVO4 interface more compatibly and provide the shorter transmission channel. Mott\u2212Schottky (M\u2212S) curve is an important mean to analyze the characteristics of semiconductor materials (Fig.\u00a05c and d). Under the applied voltage of 0\u20130.8\u00a0\u200bV, the slopes of all capacitance\u2212voltage curves are positive, indicating that the five photoanodes are n-type semiconductors in a certain range [41]. By Mot\u2212Schottky formula: \n\n1\n\nC\n2\n\n\n=\n\n\n2\n\n\ne\n0\n\n\n\u03b5\ne\n\n\nN\nd\n\n\n\n\n\n\nv\n\u2212\n\nv\nf\n\n\u2212\n\n\nk\nT\n\n\ne\n0\n\n\n\n\n, we can know that the donor carrier concentration (N\nd) of Co\u2212Ci/NiFeOOH/BiVO4 is 7 times higher than that of BiVO4, and it is also superior to Co\u2212Pi/NiFeOOH/BiVO4 and Co\u2212Sil/NiFeOOH/BiVO4 (Table. S2) [16,42,43]. The increase of the carrier concentration leads to the increase of the conductivity of Co\u2212Ci/NiFeOOH/BiVO4, which in turn increases the photocurrent density, and this is mutually corroborated with the claim that Co\u2212Ci in EIS analysis increases the transfer rate of photo-induced carrier.The PEC performances of BiVO4, NiFeOOH/BiVO4, Co\u2212Pi/NiFeOOH/BiVO4, Co\u2212Sil/NiFeOOH/BiVO4 and Co\u2212Ci/NiFeOOH/BiVO4 were explored. As shown in Fig.\u00a06\na, Co\u2212Ci/NiFeOOH/BiVO4 achieves a photocurrent density of 4.1\u00a0\u200bmA/cm2, which is 530% higher than that of BiVO4, and about 1.9 times higher than that of a single co-catalyst loaded NiFeOOH/BiVO4. At the same time, the initial potential of Co\u2212Ci/NiFeOOH/BiVO4 is reduced to about 0.3\u00a0\u200bV (vs RHE) compared to the 0.41\u00a0\u200bV (vs RHE) of the other four electrodes. In addition, both Co\u2212Pi/NiFeOOH/BiVO4 and Co\u2212Sil/NiFeOOH/BiVO4 showed better water oxidation performance than NiFeOOH/BiVO4, proving that the loaded double-layer cocatalyst was a feasible means to improve the performance of PEC [44]. This is due to the fact that the synergistic effect formed by the double-layer co-catalyst broadens the light absorption range (Fig.\u00a04a), allowing more photo-generated carriers to be generated. Under the condition of dark reaction, the photoanode loaded with co-catalyst also showed lower initial potential, demonstrating that the presence of co-catalyst improved the water oxidation kinetics of the electrode (Fig.\u00a0S7b). Fig.\u00a06b shows the transient photocurrent curve under chopped light (light-dark interval of 5\u00a0\u200bs). All electrodes show excellent photo-response. The Co\u2212Ci/NiFeOOH/BiVO4 electrode shows the highest photocurrent density in the different voltages, and the photocurrent density is close to 4.1\u00a0\u200bmA/cm2 at 1.23 V (vs RHE), which is consistent with the linear sweep voltammetry (Fig.\u00a06a). 1 mol/L Na2SO3 was used as the hole scavenger in 0.5\u00a0\u200bmol/L Na2SO4 electrolyte solution to explore the charge injection efficiency and charge separation efficiency of the electrode. The charge injection efficiency of BiVO4 is only 15%, which is due to the waste of most of its holes caused by charge recombination. After loading double-layer co-catalysts, the injection efficiencies of Co\u2212Pi/NiFeOOH/BiVO4, Co\u2212Sil/NiFeOOH/BiVO4 and Co\u2212Ci/NiFeOOH/BiVO4 are improved to certain extents. Among them, Co\u2212Ci/NiFeOOH/BiVO4 reached up to 68%, which was much higher than that of other electrodes. In this case, more holes could participate in the reaction (Fig.\u00a06c). It is also confirmed that the good matching between NiFeOOH/BiVO4 and Co\u2212Ci co-catalysts made the electrode present the most favorable reaction interface. The change of charge separation efficiency can be clearly observed in Fig.\u00a06d. Similar to the injection efficiency, the separation efficiency of photo-induced carrier of Co\u2212Ci/NiFeOOH/BiVO4 was also improved, indicating that its charge separation ability was significantly improved. Combined with Co\u2212Ci in EIS, the charge migration rate was greatly improved, and it is concluded that Co\u2212Ci/NiFeOOH/BiVO4 can make the surface reaction kinetics faster.In order to better understand the characteristics of electrode materials, a series of efficiency calculations were carried out. The applied bias photon-to-current efficiency (ABPE) reflects the conversion ability of solar energy to chemical energy of the photoanode (Fig.\u00a07\na). The maximum efficiency of Co\u2212Ci/NiFeOOH/BiVO4 reaches 0.96% at 0.82\u00a0\u200bV (vs RHE), which is much greater than 0.19% of BiVO4 under the same voltage. Thanks to this efficient photoelectric conversion, Co\u2212Ci/NiFeOOH/BiVO4 stands out in the same type of electrodes. The incident photon-to-current efficiency (IPCE) is an important parameter for evaluating electrode materials. In Fig.\u00a07b, it is found that all electrodes are active in the wavelength ranging from 350\u00a0\u200bnm to 520\u00a0\u200bnm, which is highly consistent with UV\u2013Vis diffuse reflectance (Fig.\u00a04a). The IPCE value of all the electrodes almost reaches the peak at 415\u00a0\u200bnm. Co\u2212Ci/NiFeOOH/BiVO4 has 54% conversion efficiency, which means that more than half of the photogenerated electrons are captured. In addition, the light harvesting efficiency LHE and IPCE were used to study the absorbed photon-to-current efficiency (APCE). Similar to IPCE, Co\u2212Ci/NiFeOOH/BiVO4 has the highest conversion efficiency (65%) at 415\u00a0\u200bnm (Fig.\u00a07c). It can be seen from IPCE and APCE that the photon capture and utilization of the electrode modified by the dual-layer co-catalyst are greatly stronger than those of pure bismuth vanadate and BiVO4 modified by a single co-catalyst, so Co\u2212Ci co-catalyst is undoubtedly a very promising one. Meanwhile, electron paramagnetic resonance (EPR) indicates that there are a large number of oxygen vacancies in the composite electrode (Fig.\u00a07d) [17,45]. This is precisely because the rich oxygen vacancies in Co\u2212Ci make the PEC performance of Co\u2212Ci/NiFeOOH/BiVO4 improved, which is mutually confirmed by the O 1s peak in XPS (Fig.\u00a03c).Gas chromatograph was used to understand the water splitting performance of the electrode, and quantitative analysis of oxygen and hydrogen was executed by GC integration every 0.5\u00a0\u200bh for 3\u00a0\u200bh. Generally, H2 and O2 are produced at the same time almost following the ratio of 2:1. In our comparison tests, BiVO4 produces 53\u00a0\u200b\u03bcmol of H2 in 3\u00a0\u200bh, while Co\u2212Ci/NiFeOOH/BiVO4 produces 170\u00a0\u200b\u03bcmol of H2. The Faraday efficiency is obtained by dividing the actual amount of H2 produced by the theoretical amount of H2 [46]. The Faraday efficiency of Co\u2212Ci/NiFeOOH/BiVO4 reaches 96%, indicating that almost all the photo-generated charges are used to split water (Fig.\u00a08\na). At the same time, in order to analyze the influence of Co\u2212Ci, Co\u2212Sil and Co\u2212Pi on the stability of NiFeOOH/BiVO4 photoelectrocatalytic process, and reasonably evaluate whether the composite electrode has application value, the stability of the electrode was tested under AM 1.5\u00a0\u200bG illumination (100 mW/cm2) and 1.23\u00a0\u200bV vs RHE voltage for 3\u00a0\u200bh (Fig.\u00a0S8a). The photocurrent density of all electrodes showed a downward trend at the beginning due to the electron\u2212hole recombination, but Co\u2212Ci/NiFeOOH/BiVO4 still maintained the highest photocurrent density within 3\u00a0\u200bh. In addition, Figure\u00a0S8b shows the SEM image of Co\u2212Ci/NiFeOOH/BiVO4 after a 3 h stability test. It is observed that although the worm-like basic shape of the original porous electrode surface is not destroyed, the gap between BiVO4 becomes relatively close, and the electrode surface becomes rough. Moreover, Fig.S5b provides the XRD pattern of Co\u2212Ci/NiFeOOH/BiVO4 after 3 h stability test, and it is found that the crystal structure of the electrode has not changed, proving that the electrode has a certain degree of durability, which is very important for the practical application of photoelectrochemical water splitting.According to the analyses above, a possible mechanism for the superior performance of Co\u2212Ci/NiFeOOH/BiVO4 to the same type of photoanodes is proposed (Fig.\u00a08b). On the one hand, NiFeOOH is introduced to form the carrier channel to inhibit the rapid recombination of charges. On the other hand, the O2\u2212 strength in Co\u2212Ci/NiFeOOH/BiVO4 is enhanced, which may be due to the fact that Co\u2212Ci has stronger metal binding ability and smaller molecular radius, so it can be well matched with NiFeOOH/BiVO4. The oxygen vacancies generated by Co\u2212Ci as a co-catalyst (hole transport layer) can better promote water oxidation, which is conducive to the transmission of photogenerated electrons to the FTO substrate and the generation of hydrogen at the cathode through the external circuit. In Table\u00a0S3, our work is compared with the recent photocurrent density of BiVO4 photoanode under AM 1.5\u00a0\u200bG (100\u00a0\u200bmW\u00a0\u200bcm\u22122) illumination.In this work, we refer to the Co\u2212Ci co-catalyst used in CO2 reduction and perovskite solar cells, and apply it in the field of photoelectrochemical water splitting through the method of PED. The PEC performance of the related electrodes was compared. All tests showed that this amorphous Co\u2212Ci achieved more outstanding performance than Co\u2212Pi (or Co\u2212Sil) on NiFeOOH/BiVO4. Under AM 1.5\u00a0\u200bG irradiation and pH\u00a0\u200b7.35 Na2SO4 solution, the Co\u2212Ci/NiFeOOH/BiVO4 photoanode exhibits a photocurrent density of 4.1 mA cm\u22122, which is more than 500% that of BiVO4 photoanode and is more than twice that of the traditional Co-Pi/NiFeOOH/BiVO4 electrode. The enhanced PEC activity is attributed to the fact that the Co\u2212Ci co-catalyst not only can bring abundant oxygen vacancies, but also can better match NiFeOOH/BiVO4. Co\u2212Pi/NiFeOOH/BiVO4 broadens the light absorption range, increases the index level carrier concentration and reduces the charge transfer resistance (R\nct), resulting in higher photocurrent density, more stable durability and up to 0.97% of the photoelectric conversion efficiency (ABPE). Furthermore, the low cost and simple synthesis methods of Co\u2212Ci provide strong support for its practical application and are expected to be applied to other candidate electrodes.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was financially supported by the National Natural Science Foundation of China (Nos. 52173277 and 21808189), the Key Science and Technology Foundation of Gansu Province, China (No. 20YF3GA021), and the Natural Science Foundation of Gansu Province, China (No. 20JR5RA523).The following is the Supplementary data to this article:\n\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.apmate.2021.11.010.", "descript": "\n                  The cobalt\u2212carbonate (CoCi)/NiFeOOH double-layer co-catalyst was prepared on bismuth vanadate (BiVO4). Compared with the same type of electrode (Co\u2212Pi/NiFeOOH/BiVO4 and Co\u2212Sil/NiFeOOH/BiVO4), the photoelectrochemical (PEC) performance of the composite electrode presents the most excellent performance. The Co\u2212Ci/NiFeOOH/BiVO4 electrode prepared by photoelectric deposition (PED) achieves a photocurrent density of 4.1\u00a0\u200bmA/cm2 at 1.23\u00a0\u200bV vs RHE, and the applied bias photon-current efficiency (ABPE) is up to 0.95%. In addition, with the help of the equivalent circuit fitting in electrochemical impedance spectroscopy (EIS), the charge transfer resistance (R\n                     ct) of Co\u2212Ci/NiFeOOH/BiVO4 is only 108.9\u00a0\u200b\u03a9, which is 16% that of BiVO4. The enhanced PEC performance of Co\u2212Ci/NiFeOOH/BiVO4 in the double-layer cocatalyst system is attributed to the outstanding advantages of Co\u2212Ci cocatalyst in oxygen vacancy defects, superior to other cobalt-based catalysts in promoting charge transfer and improving the kinetics of water oxidation. This makes Co\u2212Ci co-catalyst become one of the favorable competitors in the field of photoelectric catalysis.\n               "}
{"full_text": "\n\n\nData will be made available on request.\n\n\nData will be made available on request.There is an increasing urgency to identify alternative energy sources to fossil fuels in order to meet the need to supply sustainable, clean energy as well as reduce greenhouse emissions to mitigate rising global temperatures, extreme and fluctuating weather patterns, and the negative impact on the earth's ecosystem [1\u20134]. To this end, hydrogen is an attractive energy carrier as a source of clean efficient power in stationary, portable and transport applications [5] as it has a high energy density (142 MJ\u00b7kg\u22121\nvs 54 MJ\u00b7kg\u22121 for natural gas) as well as the potential to be generated in high purity from water splitting where the only by-product is oxygen [6\u201312]. However, hydrogen is a flammable gas which forms potentially explosive environments and, as such, there are significant safety concerns over its storage and transportation; moreover, compression and liquefaction of hydrogen are energy intensive processes. The use of hydrogen storage materials is one of the most promising solutions as they are stable and safe to handle and would allow for the generation of hydrogen on site [6,13-25]. To this end, sodium borohydride has appropriate credentials for use as a storage material as it has a high stability and a high hydrogen content (10.8 wt%) and is nontoxic, inexpensive and water soluble (Eq. 1) [6,13b,c,j,k, 25-32].\n\n(1)\n\n\nNaB\n\n\nH\n\n4\n\n+\n4\n\n\nH\n\n2\n\n\nO\n\n\u2192\nNaB\n\n\n(\n\nOH\n\n)\n\n4\n\n+\n4\n\n\nH\n\n2\n\n\n\n\n\nAs the thermal decomposition of NaBH4 requires temperatures in excess of 400\u00b0C and its hydrolysis in water is slow, considerable effort has been dedicated to developing cost-effective catalysts that can achieve the rapid and controllable release of hydrogen that will be required for this technology to become commercially viable. While homogeneous catalysts have been shown to facilitate the solvolysis of hydrogen-rich boron compounds [33\u201338], noble metal nanoparticles (NPs) have recently attracted considerable attention as the hydrogen generation rate can be controlled through their size, morphology and environment and the catalyst can be recovered and reused in much the same manner as a conventional heterogeneous catalyst [39\u201343]. While the high activity obtained with small nanoparticles is due to their high surface area to volume ratio and the large number of active sites, they are unstable with respect to aggregation to less reactive species which limits their practical applications [44\u201345], for example, integration into hydrogen-based fuel cells for use in vehicles and portable electronic devices [46\u201348]. One potential solution to overcome aggregation under conditions of catalysis has been to stabilise the nanoparticles by encapsulation into a support such as porous carbon structures [49\u201360], zeolites [61\u201365], mesoporous silicas [66\u201368], porous organic polymers [69\u201370], metal organic frameworks [71\u201377] and, most recently, dendrimers [78\u201380]. Additional benefits of this strategy include control of the growth and morphology due to the confinement [81\u201387], modification of their properties through surface-support interactions [88\u201393] and incorporation of functionality to affect synergy, for instance, bimetallic nanoparticles [94\u201397]. At present, the most efficient supported NP catalyst for the hydrolysis of sodium borohydride is based on RuNPs confined in zeolite-Y; this system gave a turnover frequency of 550 molH2.molRu\n\u22121.min\u22121\n[64].Ionic liquids have also been used for the stabilization of nanoparticles [98\u2013101]; however, the weak electrostatic interactions involved do not always provide sufficient stabilisation to prevent aggregation under the conditions of catalysis [102\u2013103]. One possible approach to improve nanoparticle stability has been to introduce a heteroatom donor such as a phosphine, amine, nitrile, ether, or thiol that can supplement this weak stabilization by forming a covalent interaction to the nanoparticle surface [104]. This approach has proven successful with significant improvements in catalyst stability and performance; for example, palladium nanoparticles stabilised by a phosphine-functionalised imidazolium-based ionic liquid are markedly more efficient hydrogenation catalysts than their unmodified counterparts [105\u2013109] while RuNPs stabilised by a phosphine-functionalised ionic liquid exhibited a solvent dependent chemoselectivity for the hydrogenation of aromatic ketones as reactions performed in ionic liquid were highly selective for reduction of the carbonyl group whereas the use of water as the solvent resulted in hydrogenation of both the carbonyl and the arene. Moreover, the phosphine was shown to exert a marked influence on catalyst efficiency as the corresponding phosphine-free RuNP catalyst was markedly less selective in both solvents [110\u2013111]. However, even though this strategy has been shown to improve catalyst performance, functional ionic liquids are prohibitively expensive as a bulk solvent, leaching contaminates the product and recovery, and purification of the ionic liquid can be difficult, which has limited their implementation.These issues have been addressed by grafting ionic liquids onto supports such as mesoporous silica, polymers, and MOFs on the basis that the resulting material would stabilise the nanoparticles in much the same manner as an ionic liquid, while the covalent attachment would prevent leaching of the ionic liquid, facilitate separation and recovery of the catalyst, and reduce the amount of ionic liquid, as the catalyst would be confined within the support [112\u2013117]. Polymers are particularly attractive supports as their modular construction would enable the hydrophilicity, ionic microenvironment, charge density and redox properties to be modified in a rational manner, additional functionality to be introduced and the composition and stoichiometry of the metal precursors to be defined to facilitate access to synergistic bi- and trimetallic nanoparticles. We have recently been exploring this approach and developed heteroatom donor-decorated polymer-immobilised ionic liquids, reasoning that the heteroatom donor could influence the size, size distribution and morphology of the nanoparticles as well as modify their surface electronic structure and, thereby, modulate their efficacy as catalysts. In this regard, there have been an increasing number of reports of the beneficial effect of ligands on the performance of heterogeneous nano-catalysts, which have been attributed to steric, electronic and solubility factors [118]. Our early studies showed that palladium nanoparticles immobilized on a polyethylene glycol-modified phosphine-modified PIIL is a remarkably efficient catalyst for aqueous phase Suzuki-Miyaura cross-couplings [119], the chemoselective hydrogenation of \u03b1,\u03b2-unsaturated ketones, nitriles and esters, [120] and the hydrogenation of nitroarenes [121]. Moreover, gold nanoparticles stabilized by a phosphine oxide-modified polymer immobilised ionic liquid catalyses the highly selective reduction of nitroarenes to afford N-arylhydroxylamines and azoxyarenes [122] and the corresponding ruthenium nanoparticles catalyse the aqueous phase hydrogenation of aryl and heteroaryl ketones and levulinic acid with remarkable efficacy and selectivity [123].While support-grafted ionic liquids have been used to stabilise catalysts for a wide range of transformations, there appear to be only two reports of their use to support nanoparticle catalysts for the hydrolytic evolution of hydrogen from hydrogen-rich boron derivatives, which is somewhat surprising as polymer immobilised ionic liquids are functional and tuneable supports for molecular and nanoparticle catalysts. An imidazolium-based organic polymer has recently been used to prepare highly dispersed ultrafine AuPd alloy NPs for the hydrolytic release of hydrogen from ammonia borane which outperformed both its monometallic counterparts [124] and we have recently reported that phosphine decorated polymer immobilized ionic liquid stabilized PtNPs are highly efficient catalysts for the hydrolytic generation of hydrogen from NaBH4\n[125]. This study has now been extended to investigate the efficacy of phosphine oxide and amine-decorated polymer immobilised ionic liquid stabilised RuNPs as catalysts for the hydrolysis of NaBH4 on the basis that the heteroatom donor could disrupt the key hydrogen-bonded surface-coordinated ensemble between the acidic hydrogen of water and the hydridic hydrogen of borohydride and thereby influence catalyst performance. Herein, we report the results of a comparative study to explore the influence of polymer composition on catalyst performance and reveal that that RuNPs stabilised by an amino-modified polyionic liquid outperform their phosphine oxide-decorated and unmodified counterparts. Kinetic studies in combination with deuterium isotope effects have been used to probe the mechanism and a tandem hydrogenation of 1,1-diphenylethene with hydrogen generated from the catalytic hydrolysis of NaBH4 in D2O gave a mixture of isotopologues resulting from reversible \u03b2-hydride elimination/re-insertion at a surface Ru-D competing with reductive elimination.All reagents were purchased from commercial suppliers and used without further purification, RuCl3.3H2O 99.9% (PGM basis) was purchased form Alfa Aesar (47182) and polymers 1a-f were prepared as previously described and their purity confirmed by 1H and 13C{1H} NMR spectroscopy and elemental analysis. Ethanol was distilled over iodine activated magnesium with a magnesium loading of 5.0 g L\u22121 and diethyl ether from Na/K alloy under an atmosphere of nitrogen.To a round bottom flask charged with 1a (4.0 g, 6.5 mmol) and ethanol (100 mL) was added a solution of RuCl3\u00b73H2O (1.3 g, 6.5 mmol) in ethanol (20 mL). The resulting mixture was stirred vigorously for 5 h at room temperature after which time a solution of NaBH4 (2.0 g, 52.0 mmol) in water (10 mL) was added dropwise and the suspension stirred for an additional 18 h before concentrating to dryness under vacuo. The crude black solid was triturated with cold acetone (2\u00a0\u00d7\u00a0100 mL) then washed with water (100 mL) followed by ethanol (2\u00a0\u00d7\u00a040 mL) to afford a black solid that was recovered from the washings via centrifugation followed by filtration through a frit. The final product was rinsed with ether until a fine black powder was obtained which was dried under vacuum to afford 2a in 87% yield (4.06 g). ICP-OES data: 5.85 wt% ruthenium and a ruthenium loading of 0.58 mmol\u2219g\u22121.Catalyst 2b was prepared from 1b (1.0 g, 0.83 mmol), RuCl3\u00b73H2O (0.17 g, 0.83 mmol) and NaBH4 (0.25 g, 6.64 mmol) in ethanol (25 mL) as described above to afford a fine black powder in 50% yield (0.54 g). ICP-OES data: 7.02 wt% ruthenium and a ruthenium loading of 0.70 mmol\u2219g\u22121.Catalyst 2c was prepared from 1c (5.0 g, 6.25 mmol), RuCl3\u00b73H2O (1.30 g, 6.25 mmol) and NaBH4 (1.89 g, 50 mmol) in ethanol (100 mL) as described above to afford a fine black powder in 53% yield (2.82 g). ICP-OES data: 7.24 wt% ruthenium and a ruthenium loading of 0.72 mmol\u2219g\u22121.Catalyst 2d was prepared from 1d (4.0 g, 2.68 mmol), RuCl3\u00b73H2O (0.46 g, 2.68 mmol) and NaBH4 (0.81 g, 21.4 mmol) in ethanol (100 mL) as described above to afford a fine black powder in 78% yield (3.32 g). ICP-OES data: 1.83 wt% ruthenium and a ruthenium loading of 0.18 mmol\u2219g\u22121.Catalyst 2e was prepared from 1e (5.0 g, 7.75 mmol), RuCl3\u00b73H2O (1.60 g, 7.75 mmol) and NaBH4 (2.34 g, 62 mmol) in ethanol (100 mL) as described above to afford a fine black powder in 67% yield (3.88 g). ICP-OES data: 3.43 wt% ruthenium and a ruthenium loading of 0.34 mmol\u2219g\u22121.Catalyst 2f was prepared from 1c (4.0 g, 5.11 mmol), RuCl3\u00b73H2O (1.06 g, 5.11 mmol) and NaBH4 (1.54 g, 40.9 mmol) in ethanol (100 mL) as described above to afford a fine black powder in 79% yield (3.45 g). ICP-OES data: ICP-OES data: 6.97 wt% ruthenium and a ruthenium loading of 0.69 mmol g\u22121.Comparative catalytic hydrolysis reactions were conducted in water at the appropriate temperature in a thermostated 50 mL round bottom flask. In a typical experiment, a flask charged with a stir bar, catalyst 2a-f (0.2 mol%) and NaBH4 (0.021 g, 0.57 mmol) and fitted with a gas outlet and connected to the top of an inverted water-filled burette designed to monitor the progress of the reaction by measuring the volume of water displaced with time. The flask was stabilised at 303 K and the reaction was initiated by adding water (2 mL) and the system was immediately sealed by replacing the gas outlet; the tap to the water filled burette was then opened, the time zero volume recorded, and the water displacement monitored. The optimum activity for each catalyst was determined by varying the catalyst loadings between 0.08 and 0.32 mol% at 303 K and measuring the hydrogen produced as a function of time. Kinetic studies were also conducted according to the protocol described above using the following catalyst loadings: 0.26 mol% 2a, 0.32 mol% 2b, 0.45 mol% 2c, 0.11 mol% 2d, 0.16 mol% 2e and 0.32 mol% 2f for a range of temperatures (294 K, 298 K, 303 K, 308 K and 313 K) and the corresponding activation energies (Ea) were determined from an Arrhenius plot of the initial rate against 1/T.The reaction order in catalyst was determined by performing the hydrolysis reactions at 298 K with NaBH4 (0.28 M, 0.021 g) in water (2 mL) and varying the catalyst concentration from 0.14 mol% to 0.69 mol% for 2a, 0.16 mol% to 0.63 mol% for 2b, 0.23 mol% to 1.1 mol% for 2c, 0.058 mol% to 0.28 mol% for 2d, 0.12 mol% to 0.27 mol% for 2e and 0.25 mol% to 0.64 mol% for 2f. The reaction order in sodium borohydride concentration was investigated by performing reactions at 298 K in water (200 mL) using 0.026 mmol of catalysts 2a (0.0448 g), 2e (0.0764 g) and 2f (0.0376 g) and varying the amount of sodium borohydride between 6.6 \u03bcmole and 185 \u03bcmole (i.e. [NaBH4]0\u00a0=\u00a00.035, 0.07, 0.13, 0.26, 0.39, 0.53, 0.65, 0.78, 0.9 mM), such that the catalyst:NaBH4 mole ratios ranged from 4:1 and 1:6. The effect of sodium borohydride concentration on the initial rate of hydrolysis at high concentrations of sodium borohydride, i.e. under the conditions of catalysis, was also determined using 2e (0.0026 g, 0.884 \u03bcmol) to catalyze the hydrolysis of NaBH4 solutions (2 mL) with varying concentrations of sodium borohydride ranging from 0.55 mmol to 2.2 mmol ([NaBH4]0\u00a0=\u00a00.28, 0.56, 0.83, 1.1 M).The effect of the concentration of NaOH on catalyst efficacy was explored by conducting catalytic hydrolysis reactions at 303 K in 2 mL of alkaline 0.28 M NaBH4 (0.021 g) across a range of sodium hydroxide concentrations (i.e. [NaOH]\u00a0=\u00a00.035, 0.07, 0.14, 0.28, 5.0, 10, 50, 100 mM) catalyzed by 0.26 mol% 2a (0.0025 g) and monitoring the gas evolution.Recycle studies were performed at 303 K as described above using 2 mol% 2a (0.0193 g, 0.0114 mmol) and 2e (0.0335 g, 0.0114 mmol) to catalyze the hydrolysis of sodium borohydride (0.021 g, 0.57 mmol) in water (20 mL). The progress of the reaction was monitored as described above and when the hydrolysis was complete an additional portion of fresh sodium borohydride (0.021 g, 0.57 mmol) was added, and the procedure repeated. After the 5th run samples of the catalysts were isolated and analysed by TEM.A borate-buffered solution was prepared by dissolving Na2B4O7\u00b710H2O (9.53 g, 25 mmol) and NaCl (4.39 g, 75 mmol) in distilled water (900 mL) in a volumetric flask. When the borate was completely dissolved the pH of the solution was adjusted to 7.2 by gradual addition of boric acid (20.99 g, 0.34 mol); the solution was then made up to one liter. Recycle studies were conducted by adding NaBH4 (0.021 g, 0.57 mmol) to a flask containing 1 mol% 2e (0.0165 g, 0.0056 mmol) and 20 mL of the aqueous borate buffer solution. The flask was maintained at 303 K and the progress of the reaction was monitored as described above. When the hydrolysis was complete an additional portion of fresh sodium borohydride (0.021 g, 0.57 mmol) was added, and the procedure repeated for comparison with the recycle study described above in the absence of buffer.Hot filtration studies were conducted at 303 K following the protocol described above using either 0.2 mol% 2a (0.0019 g) or 0.16 mol% 2e (0.0026 g) to catalyze the hydrolysis of sodium borohydride (0.021 g, 0.57 mmol) in water (2 mL). The progress of the reaction was monitored as a function of time and the mixture filtered through a 0.45 \u03bcm syringe filter when the conversion reached ca. 50% (10 min for 2a and 7.75 min for 2e), after which the burette assembly was reconnected, and the gas evolution monitored for a further 30 min. In a complementary procedure, a hydrolysis reaction that had reached completion was filtered through a 0.45 \u03bcm syringe filter and an additional portion of NaBH4 (0.021 g, 0.57 mmol) added to the filtrate and the gas evolution monitored.A flask was charged with 2 mol% catalyst (2a 0.0186 g; 2e, 0.0335 g), water (20 mL) and sodium metaborate (0.0765 g, 0.57 mmol) and the resulting mixture stirred at 303 K for the predetermined time (t\u00a0=\u00a00 min, 20 min, 40 min, 60 min) to investigate whether the pre-stirring time influences catalyst efficacy. After pre-stirring for the allocated time, the reaction was initiated by addition of the NaBH4 (0.021 g, 0.57 mmol) and the rate of hydrogen evolution quantified by measuring the volume of water displaced with time.Tandem hydrogenations were performed using two Schlenk flasks connected through tubing. One of the flasks was charged with a stir bar, either NaBH4 (0.042 g, 1.11 mmol) or NaBD4 (0.046 g, 1.11 mol) and 0.26 mol% 2e (0.0025 g) and the hydrolysis started by addition of either D2O (2 mL) or H2O (2 mL). The reaction flask was immediately stoppered, isolated from the second flask by closing the stopcock and stirred for 70 min. The second Schlenk flask was charged with 1,1-diphenylethene (0.180 g, 1.00 mmol), 0.5 mol% Pd/C and either CH3OH (2 mL) or d4-methanol (2 mL). After 70 min the second flask was evacuated briefly before opening the connector to the hydrolysis flask. The reaction was allowed to stir at 303 K for 18 h before the solvent was removed and the residue analyzed by 13C{1H} NMR spectroscopy and GC-MS to establish the composition and quantify the distribution of isotopologues.The polymers required for this study were prepared via radical polymerisation of the corresponding imidazolium-based ionic liquid monomer, either styrene, (4-vinylphenyl)methanamine or diphenyl(4-vinylphenyl)phosphine oxide and the corresponding imidazolium-based ionic liquid cross-linker in the ratio x\u00a0=\u00a01.84, y\u00a0=\u00a01.0 and z\u00a0=\u00a00.16, as previously described [119\u2013123]. Catalysts 2a-f were prepared by the wet impregnation of the polymer support with ruthenium trichloride to afford precursors with a 1:1 ratio of ruthenium to neutral monomer, followed by in-situ reduction of the ruthenium with NaBH4; to afford the product as a fine black powder in high yield; the synthesis and composition of the polymers and the catalysts is shown in Fig.\u00a01\n. The composition and purity of polymers 1a-f was determined using a combination of solution and solid state 13C{H} and 31P{H} NMR spectroscopy and elemental analysis while the loaded RuNP catalysts were characterised by solid state 13C{H} and 31P{H} NMR spectroscopy, infra-red (IR) spectroscopy, high resolution transmission electron microscopy (HRTEM), SEM, X-ray photoelectron spectroscopy (XPS) and inductively coupled plasma-optical emission spectroscopy (ICP-OES) (See Fig.\u00a02\n and the supporting information for full details). The ruthenium loadings in 2a-f were determined to be 0.18\u22120.75 mmolg\u22121 using ICP-OES.The solid state 13C{1H} NMR spectra of 1a-f and 2a-f each contain resonances from \u03b4 121 to 149 ppm, which correspond to the aromatic carbon atoms of polystyrene and the carbon atoms of the imidazolium ring, as well as signals between \u03b4 10 and 51 ppm which belong to the methylene carbon atoms of the polystyrene backbone and the methyl group attached to the imidazolium ring. Additional signals at \u03b4 71 and 59 ppm for 2b, 2d and 2f belong to the carbon atoms of the polyethylene glycol (PEG) chain and the terminal OMe, respectively, and a signal at \u03b4 49 ppm for 2e and 2f is associated with the CH2NH2. The surface of the RuNP catalysts was characterised by X-ray photoelectron spectroscopy by analysing the Ru 3p region as the C 1s and Ru 3d region overlapped. For catalyst 2a, stabilised by unmodified imidazolium-based polymer, a Ru 3p3/2 peak at 463.19 eV was assigned to RuO2, and satellite features were fitted at 465.97 eV (Table S2 and Fig.\u00a02a). The presence of RuO2 species is most likely due to surface oxidation of the pre-formed metallic Ru nanoparticles. The corresponding Ru 3p3/2 peak for catalysts containing the phosphine oxide (2c and 2d) or amine (2e and 2f) was shifted to lower binding energy (462.56 and 461.37 eV for 2c and 2d, respectively and 462.83 and 462.89 eV for 2e and 2f, respectively) compared to the Ru 3p3/2 binding energy of 463.19 eV for catalyst 2a (Table S2 and Fig.\u00a02b-f). A shift to lower binding energy may be indicative of electron transfer from the heteroatom of the phosphine oxide or amine to the RuNPs. Catalyst 2d containing O=PPh2 and PEG heteroatom donors gave the largest shift (-1.72 eV) in binding energy of the Ru 3p3/2 peak (461.37 eV for 2d) relative to 2a. TEM micrographs of 2a-f revealed that the ruthenium nanoparticles were ultrafine and near monodisperse with average diameters between 1.6 and 2.8 nm; representative micrographs and the corresponding distribution histograms based on the sizing of >100 particles for 2a-f are shown in Fig.\u00a02. SEM images revealed that the catalyst materials were far more granular than their polymeric counterparts, which appeared largely smooth.The hydrolysis of sodium borohydride was identified to investigate the efficacy of catalyst 2a-f on the basis that PEG-modified \u2018click\u2019-dendrimer stabilised noble and bimetallic metal nanoparticles catalyse this reaction with promising initial TOFs and as such would provide a formative benchmark for comparative evaluation. Preliminary catalyst testing was conducted using recent literature protocols as a lead [75,78]; reactions were initially performed at 303 K using 0.2 mol% of 2a-f to catalyse the hydrolysis of a 0.28 M solution of sodium borohydride (Fig.\u00a03\na, b). The reaction was monitored by quantifying the amount of hydrogen liberated as a function of time using water displacement from an inverted burette assembly and all data were corrected by subtracting the background hydrogen generated over the same time under identical conditions. Hydrogen evolution started immediately with no induction period which is consistent with the metallic state of the ruthenium. Under these conditions, RuNP@NH2-PIILS (2e) gave the highest initial TOF of 135 moleH2.molRu\n\u22121.min\u22121 and reached 92% conversion after 20 min, whereas its PEGylated counterpart RuNP@NH2-PEGPIILS (2f) was less active with a slightly lower TOF of 117 moleH2.molRu\n\u22121.min\u22121. Removal of the amino-group from either of these systems resulted in a reduction in the activity with RuNP@PIILS (2a) and RuNP@PEGPIILS (2b) giving initial TOFs of 121 moleH2.molRu\n\u22121.min\u22121 and 89 moleH2.molcat\u22121.min\u22121, respectively. In contrast, under the same conditions, catalysts 2c and 2d, stabilised by phosphine oxide-decorated polymer, were both less active than their respective amino-modified analogues with initial TOFs of 70 moleH2.molRu\n\u22121.min\u22121 and 103 moleH2.molRu\n\u22121.min\u22121, respectively (Fig.\u00a03b). For comparison, 0.16 mol% Ru/C (5 wt%) catalysed this hydrolysis under the same conditions but only reached 57% conversion after 25 min with a TOF of 69 moleH2.molRu\n\u22121.min\u22121. The initial TOF for 2e improved to 177 moleH2.molRu\n\u22121.min\u22121 when the reaction was performed in dilute solution (10 mL) with a reduced catalyst loading of 0.08 mol%. A series of baseline hydrolysis reactions conducted by substituting catalysts 2a-f with their corresponding polymers 1a-f confirmed that the RuNPs were essential for catalysis as the gas evolution did not exceed the background reaction under the same conditions.As there is no clear correlation between the efficacy of catalysts 2a-f and the nanoparticle size, further studies will be conducted to explore the surface electron density of the RuNPs as a function of the support and to investigate whether the amine influences the hydrogen bonded surface ensemble responsible for substrate activation or improves the dispersibility of the catalyst in the reaction mixture and thereby access to the active site. To this end, amine-modified supports have previously been reported to improve the performance of nanoparticle catalysts compared with the corresponding unmodified catalyst. For example, ruthenium nanoparticles stabilised within the pores of amine-modified MIL-53 (MIL-53(Al)-NH2) is a significantly more active catalyst for the dehydrogenation of amine-borane than its unmodified counterpart, MIL(Al)-53; this was attributed to the formation and stabilization of ultra-small RuNPs [76]. There are also numerous additional reports of the beneficial effect on catalyst performance of incorporating an amine onto the surface of a support. For instance, a marked improvement in the activity and selectivity of platinum nanowires for the partial hydrogenation of nitroarenes to N-phenylhydroxylamine [126\u2013127], an enhancement in the activity of RuNPs for the hydrogenation of levulinic acid to \u03b3-valerolactone [128], an improvement in activity for the transfer hydrogenation of nitroarenes catalysed by RuNP confined in an amine-modified porous organic polymer [129], an increase in activity for the PtNP-catalysed hydrogenation of quinoline [130], improvements in activity and selectivity for the Pt/Co and PdNP catalysed semi-hydrogenation of alkynes [131\u2013133], and highly selective reduction of the carbonyl in cinnamaldehyde with MOF-confined Pt nanoclusters [134].Although a comparison of the efficacy of 2a-f with literature reports of other supported ruthenium nanoparticles should be treated with caution because of the vastly disparate experimental conditions and protocols employed to collect data, the initial TOF of 177 moleH2.molRu\n\u22121.min\u22121 is higher than that of 80 moleH2.molcat\u22121.min\u22121 obtained with PEGylated click dendrimer-stabilised RuNPs [78], and 105 moleH2.molRu\n\u22121.min\u22121 with ruthenium electrodeposited on nickel foam [135] and a marked improvement on 67 moleH2.molRu\n\u22121.min\u22121 obtained in 5% wt NaOH with RuNPs nanoclusters stabilised by confinement in the framework of Zeolite-Y [64], 25 moleH2.molRu\n\u22121.min\u22121 for RuNP@ZIF-67 [77] and 35 moleH2.molRu\n\u22121.min\u22121 for carbon-supported bimetallic RuCo nanoparticles [136]; but lower than that of 550 moleH2.molRu\n\u22121.min\u22121 obtained with RuNPs stabilised in Zeolite-Y [64] and 505 moleH2.molRu\n\u22121.min\u22121 with nanoporous ruthenium prepared by chemical dealloying RuAl [137]; to the best of our knowledge these latter systems are the most active ruthenium-based catalysts for this hydrolysis.As the highest TOF was obtained with 2e, a thorough study of the reaction kinetics together with deuterium isotope effects, recycle experiments and a tandem reaction using the liberated hydrogen for the tandem hydrogenation of 1,1-diphenylethene with deuterium labelling was undertaken, details of which are discussed herein; for comparison, full details of the corresponding experiments with catalysts 2a-d and 2f are provided in the supporting information and discussed in context where appropriate. There have been numerous reports of an enhancement in activity for the metal nanoparticle catalysed hydrolytic evolution of hydrogen from sodium borohydride and amine borane in the presence of added base. For example, Astruc has reported a marked increase in the initial TOF for the hydrolysis of NaBH4 catalysed by click dendrimer-supported RuNPs from 80 moleH2.molRu\n\u22121.min\u22121 to 186 moleH2.molRu\n\u22121.min\u22121 in the presence of 0.2 M NaOH; an increase in TOF was also observed for a host of other catalysts including Rh, Au, Pd, Co, Ni, Fe and Co nanoparticles with the exception of PtNPs which experienced a strong negative effect [78]. Significant enhancements in TOF were also obtained for the hydrolysis of hydrogen-rich boron compounds with MNP@ZIF-8 (M\u00a0=\u00a0Ni, Co), NiPtNP@ZIF-8 and CoPtNP@dendrimer nanocatalysts in the presence of NaOH [72,73,75,80]. This enhancement has been attributed to coordination of the hydroxide to the nanoparticle surface which increases the electron density and facilitates activation of the O-H bond; in contrast, Pt is an electron-rich metal and highly reactive towards oxidative addition and as such the hydroxide ions occupy surface active sites and prevent substrate coordination. Such a large enhancement in activity for a dendrimer-stabilised RuNP-based catalyst prompted us to study the efficiency of 2a for the catalytic hydrolysis of NaBH4 as a function of the concentration of sodium hydroxide; reactions were conducted using 0.26 mol% of 2a to catalyse the hydrolysis of alkaline solutions of 0.28 M NaBH4 with sodium hydroxide concentrations ranging between 0.035 mM to 100 mM (Fig.\u00a04\n). There was no apparent variation in the initial TOF at low concentrations of NaOH (< 0.035 mM) while the TOFs decreased gradually at concentrations above 0.07 mM; this decrease became more dramatic when the sodium hydroxide concentration reached 5 mM and the initial TOF eventually dropped from 136 moleH2.molRu\n\u22121.min\u22121 in the absence of sodium hydroxide to 39 moleH2.molRu\n\u22121.min\u22121 in a 100 mM NaOH solution of NaBH4. To this end, there have been several reports of a decrease in the hydrogen generation activity with increasing NaOH concentration (1-10 wt% NaOH) for the ruthenium-catalysed hydrolysis of NaBH4\n[138\u2013142], which were attributed to strong interactions between the hydroxide ions and water decreasing the available free water needed for the hydrolysis of NaBH4\n[138]. However, it is interesting to note that high concentrations of NaOH have been shown to enhance the hydrogen generation activity for the non-noble metal catalysed hydrolysis of NaBH4, i.e. these systems tolerate high concentrations of hydroxide and coordination of the OH\u2212 to the surface does not appear to prevent substate binding [143\u2013148]. As the decrease in hydrogen generation rate for 2a at a NaOH concentration as low as 0.001 wt% (0.28 mM) is unlikely to be due to a reduction in the activity of water, as described by Amendola, the high rate obtained in the absence of NaOH may reflect the intrinsic activity of ruthenium to facilitate oxidative addition as a late transition metal while the reduction in activity in the presence of even a minor amount of sodium hydroxide (NaOH:catalyst between 0.05:1 and 0.4:1) may be attributed to the hydroxyphilic nature of ruthenium with the hydroxide ions occupying surface active sites and preventing substrate coordination and activation, as described above; even at these concentrations there would be sufficient OH\u2212 ions to populate the surface of the nanoparticle and disrupt the strongly hydrogen bonded NaBH4\u2014H2O ensemble involved in the rate limiting O-H bond activation step (vide infra).Kinetic studies were subsequently undertaken to determine the temperature dependence of the rate and obtain activation parameters for the hydrolytic release of hydrogen from NaBH4 for a comparison with related systems reported in the literature. A set of reactions were conducted to monitor the hydrolysis of a 0.28 M solution of NaBH4 as a function of time to determine the initial rates across a range of temperatures from 294 K to 313 K. The apparent activation energies (Ea) for the hydrolysis catalysed by 2a-f, determined from an Arrhenius plot of lnk against 1/T (lnk\u00a0=\u00a0lnA - Ea/RT) using the initial rates calculated from the linear slope of the graph, ranged from 38.9 kJ mol\u22121 to 51.8 kJ mol\u22121 (Fig.\u00a05\na-b and Fig. S1 in the supporting information). These values lie within the range reported for the hydrolysis of NaBH4 with other RuNP catalysts including 35 kJ mol\u22121 for RuNPs stabilised in the framework of Zeolite-Y [64], 41 kJ mol\u22121 for water-dispersible, acetate-stabilized RuNPs [149], 36 kJ mol\u22121 for RuNPs confined in ZIF-67 [77], 47 kJ mol\u22121 for RuNPs immobilised by the anion exchange resin IRA-400 [150] and 41.8 kJ mol\u22121 for ruthenium immobilised on Al2O3 pellets [151], but slightly lower than 61.1 kJ mol\u22121 for RuNPs supported on amine-modified graphite [139], 56.0 kJ mol\u22121 for RuNP@IRA-400 [138], 58.2 kJ mol\u22121 for Ru(acac)3\n[152] and 66.9 kJ mol\u22121 for ruthenium supported on carbon [153]. There does not appear to be a correlation between the activation energies and the initial rates which may be attributed to variations in the number of active sites or their availability as this determines the pre-exponential factor (A) [76, 154].The hydrogen release was next investigated as a function of the concentration of 2e across a range of catalyst loadings from 0.12 mol% to 0.28 mol% in 0.28 M NaBH4 (Fig.\u00a06\na) and the logarithmic plot of the initial hydrogen generation rate versus catalyst concentration gave a straight line with a slope of 1.04 (Fig.\u00a06b), indicating that the hydrolysis of NaBH4 is first order with respect to the catalyst. Similarly, the corresponding slopes for the logarithmic plots obtained with catalyst 2a-d and 2f varied between 0.70 and 1.04, which are all consistent with first order kinetics; full details are presented in Fig. S2 in the supporting information. This data is also consistent with recent reports of noble metal nanoparticle-catalysed hydrogen generation from hydrogen-rich boron derivatives including a slope of 0.73 for RuNPs confined in Zeolite-Y [64], 0.94 for RuNPs stabilized by polyvinylpyrrolidinone [155], 1.06 for Ru(acac)3\n[152], 1.17 for porphyrin-stabilised RuNPs [156], 0.85 for PtCoNP@dendrimer [78], and 0.82 for Ni2Pt@ZIF-8 [73]. The variation in the rate of hydrolysis of NaBH4 as a function of the substrate concentration was also investigated using catalyst 2e. As the order of reaction with respect to NaBH4 has been reported to depend on the amount of NaBH4 in solution (i.e. the NaBH4:catalyst ratio), changing from 1 to 0 as the concentration of NaBH4 increases [145], kinetic data was obtained by conducting a series of reactions with 0.026 mmol of catalyst 2e and varying the initial concentration of NaBH4 from 0.066 mM to 0.52 mM as these amounts correspond to catalyst:hydride ratios between 2:1 and 1:4 (Fig.\u00a07\n). Such low catalyst/hydride mole ratios were used to avoid the BH4\u2212induced dynamic saturation of the active sites on the catalyst surface which would give zero order kinetics; under these conditions the surface is not completely covered by NaBH4 and there are active sites. The slope of 1.02 obtained from the logarithm plot of hydrogen generation rate versus concentration of NaBH4 confirms that the hydrolysis is first order in substate, which undergoes rate limiting diffusion on the catalyst surface. Under the same conditions, slopes of 1.02 and 1.01 were also obtained with catalysts 2a and 2f, respectively, which are both consistent with first order kinetics; see Fig. S3 in the electronic supporting information. First order kinetics with respect to NaBH4 have previously been reported for ruthenium on carbon [142], palladium on carbon [157] and Pd and Pt dispersed on functionalised surfaces of carbon nanotubes [158] when reactions were conducted at low concentrations of NaBH4. A similar study conducted with catalyst 2e at much higher catalyst/hydride mole ratios between 1:625 and 1:2500 gave a slope of 0.26 which is indicative of zero order kinetics due to saturation of the active sites on the catalyst surface during the reaction (Fig. S4 in the supporting information), as described by Patel [145]. A slope of 0.17 was also obtained using catalysts 2d which is also consistent with zero order kinetics; similar kinetics have previously been described for ruthenium nanoclusters [159], Ru supported on IRA 400 [150] and ruthenium on carbon [153].The kinetic isotope effect (KIE) is a valuable tool for elucidating information about the rate limiting step (RLS) of a reaction that has been routinely used to probe the catalytic hydrogen generation from borohydride and amine borane (AB) [160,72,79,80]. While the reaction kinetics are complicated and the mechanism still not fully understood [42] it is clear that both NaBH4 and ammonia-borane are hydride donors and provide one of the two hydrogen atoms of the derived hydrogen gas while water provides the other in the form of a proton [41,43] and that the rate determining step involves activation of one of the O-H bonds of water, as measured by the large primary KIE obtained when the hydrolysis is performed in D2O instead of H2O [78,79,80,83,161,162]. Activation of an O-H bond has been proposed to occur via oxidative addition involving a hydrogen-bonded ensemble between a surface-coordinated borohydride and a water proton; the hydrogen could then be liberated either via reductive elimination between a borohydride-derived NP-H and the water-derived NP-H (Fig.\u00a08\n, pathway a-c) or a concerted \u03c3-bond metathesis-like process between a surface coordinated [BH4]\u2212 and a water-derived NP-H (Fig.\u00a08, pathway d-e), which may be facilitated by hydroxide. Alternatively, the protonic and hydridic hydrogen atoms may be transferred to the nanoparticle surface by oxidative addition of both the O-H and B-H bonds, respectively, to afford a dihydride that would generate hydrogen and BH3-OH via reductive elimination (Fig.\u00a08, pathway f-g), as proposed by Astruc for the CoNP@ZIF-8 catalysed hydrolysis of NaBH4\n[75]. While the pathways described in Fig.\u00a08 are all initiated by oxidative addition of the O-H bond of water via a hydrogen-bonded ensemble involving a surface-coordinated borohydride, Jagirdar [163] and Ma [164] have suggested that activation of the O-H bond and generation of H2 could occur via a hydrogen-bonding interaction between a surface adsorbed water and a surface hydride generated via rapid hydride transfer from NaBH4 to the NP surface.The role of H2O in the hydrolysis of NaBH4 catalysed by 2e was explored by conducting the reaction in D2O and monitoring the hydrogen evolution as a function of time to determine the KIE. Reactions were conducted under the conditions of catalysis i.e. 0.16 mol% of 2e was used to catalyse the hydrolysis of 2 mL of a 0.28 M solution of NaBH4 at 30\u00b0C. A comparison of the efficacy of 2e as a catalyst for the hydrolysis of NaBH4 in H2O and D2O revealed that the reaction was more rapid in H2O than in D2O with a primary kinetic isotope effect (k\nH/k\nD) of 2.31 (Fig.\u00a09\na); similar values of k\nH/k\nD were obtained with catalysts 2a (k\nH/k\nD\u00a0=\u00a01.76) and 2d (k\nH/k\nD\u00a0=\u00a01.53) and the corresponding data is presented in Fig. S5a-b in the supporting information. This value is comparable to the solvent isotope effect of 2.25 obtained by Astruc for the gold-ruthenium nanoalloy catalysed visible light-accelerated hydrolytic dehydrogenation of NaBH4 and amine-borane [165] as well as 1.8 determined in a detailed kinetic analysis of the platinum-catalysed hydrolysis of NaBH4 in alkaline media [162], 2.3 for dendrimer-stabilised RhNPs [79], 2.4 for PtCo@dendrimer [80] and 2.49 for NiNP@ZIF-8 [72] and supports a mechanism with rate limiting cleavage of an O-H bond of water in a surface-coordinated hydrogen-bonded ensemble of the type described above and shown in Fig.\u00a08. The same comparison of initial rates between reactions conducted in H2O and D2O under stoichiometric conditions using 26 \u03bcmol of 2e for the catalytic hydrolysis of 200 mL of a 0.13 mM solution of NaBH4 at 30\u00b0C (catalyst:NaBH4 ratio of 1:1) gave a primary kinetic isotope effect of 1.7 (Fig. S6d in the supporting information), which is also consistent with rate limiting oxidative addition of water. However, this KIE does not distinguish between a rate limiting step in which a surface coordinated NaBH4\u2014\u2014HOH ensemble activates an O-H bond towards oxidative addition through a hydrogen-bonding interaction to afford a water-derived metal hydride and a surface-coordinated borohydride, such as that shown in Fig.\u00a08 pathway a, and concerted activation of both the B-H and O-H bonds in a similar hydrogen-bonded ensemble; the latter process would most likely occur via oxidative addition of the O-H bond and rapid hydride transfer from the borohydride (Fig.\u00a08 pathway a-c) rather than oxidative addition of both the O-H and B-H bonds (Fig.\u00a08, pathway f-g) as borohydrides are extremely potent transfer reagents. For the same reason, a subsequent \u03c3-bond metathesis involving the surface-coordinated borohydride and the water-derived RuNP hydride would also be unlikely (Fig.\u00a08, pathway e).Thus, the mechanism was further probed by comparing the rates of hydrolysis of NaBD4 and NaBH4 catalysed by 2e at 30\u00b0C. Analysis of the initial rates obtained for the hydrolysis of 200 mL of a 0.13 mM solution of NaBH4 and NaBD4 catalysed by 26 \u03bcmol of 2e, i.e., a substrate/catalyst ratio of 1, gave a primary kinetic isotope effect (k\nH/k\nD) of 2.72 (Fig.\u00a09b). Reassuringly, comparable values were also obtained with catalysts 2a (k\nH/k\nD\u00a0=\u00a02.25) and 2d (k\nH/k\nD\u00a0=\u00a02.37), full details of which are presented in Fig. S6a-b in the supporting information. These values are comparable to that of 2.2 obtained for the visible light-accelerated H2 evolution from NaBH4 catalysed by a gold-ruthenium nanoalloy; which, together with a KIE of 2.5 obtained for the hydrolysis of NaBH4 in D2O, was taken to indicate that both the O-H and B-H bonds were activated by the ruthenium atoms in the rate limiting step, most likely via concerted oxidative addition-hydride transfer, involving the surface-coordinated hydrogen-bonded [BH3H\u2212]\u2014\u2013H-OH ensemble, rather than oxidative addition of both the O-H and B-H bonds [75,165]. Interestingly though, comparison of the rates obtained under the conditions of catalysis using 2e to catalyse the hydrolysis of 2 mL of 0.28 M solutions of NaBH4 and NaBD4 at 30\u00b0C gave a KIE of 0.65 (Fig.\u00a09c); similar values were also obtained with catalysts 2a (k\nH/k\nD\u00a0=\u00a00.87) and 2d (k\nH/k\nD\u00a0=\u00a00.85), full details of which are provided in Fig. S5d-f in the supporting information. These are inverse kinetic isotope effects and would be consistent with a surface-coordinated borohydride activating an O-H bond of water in the hydrogen-bonded ensemble prior to hydride transfer.The hydrogen liberated from the catalytic hydrolysis of NaBH4 was used for the hydrogenation of 1,1-diphenylethene with various labelling experiments to determine the fate of the liberated hydrogen. In the first of these, the tandem reaction was conducted using 0.26 mol% 2a to generate hydrogen from a 0.28 M solution of NaBH4 in D2O at 30\u00b0C in a sealed tube; after 70 min the connector was opened to the second flask which contained 1,1-diphenylethene and 0.5 mol% Pd/C in d4-methanol and the resulting mixture was stirred for 18 h. Interestingly, analysis of the crude mixture by 1H, 2H and 13C NMR spectroscopy and mass spectrometry revealed that a mixture of all eight isotopologues of 1,1-diphenylethane had been generated (Scheme\u00a01\n). Analysis of the methine region (\u03b4 44.5 ppm) of the 13C{1H} NMR spectrum was used to identify and assign each of the isotopologues, which appear as a set of four singlets at \u03b4\u00a044.88, 44.81, 44.73, and 44.66 ppm corresponding to I, II, III, and IV, respectively, while V, VI, VII and VIII appear as a set of four 1:1:1 triplets at \u03b4 44.46, 44.39, 44.31 and 44.24 ppm, respectively, resulting from a J\nCD of 19.5 Hz due to the deuterium atom attached to the methine carbon; the methyl group of these isotopologues has either zero, one, two, or three deuterium atoms. The experimental spectrum of the reaction mixture and the summed simulated spectrum of each isotopologue are shown in Fig.\u00a010\n (see Fig. S71 in the supporting information for full details of the simulated spectrum for each isotopologue). The summed simulated spectrum is remarkably similar to the experimental spectrum, which supports the assignment of the isotopologues and their relative proportions and confirms that the coupling constants, chemical shifts and line intensities and widths have been correctly determined. On the basis that the hydrogen generated from the hydrolysis of NaBH4 in D2O should result from a water-derived proton and a borohydride-derived hydride, the deuterium incorporation for all isotopologues II-VIII should be one. To this end, the total deuterium incorporation of 1.3 is slightly higher than expected and could be due to H/D exchange either with the d4-MeOH on the Pd/C during the hydrogenation or from the generation of a mixture of HD and D2 by exchange at the NP surface after O-D bond activation. A complementary experiment using hydrogen liberated from NaBH4/H2O for the hydrogenation of 1,1-diphenylethene in d4-methanol gave a total deuterium incorporation of 0.3, which confirms that H/D exchange occurs on the surface of the Pd/C; moreover, this deuterium incorporation corresponds to the excess of 0.3 above the total deuterium incorporation of one that was expected when the hydrogenation was performed in d4-MeOH with hydrogen generated from NaBH4/D2O. The hydrogenation was also performed in toluene with hydrogen generated from NaBH4 in D2O to investigate exchange at the NP surface. Under these conditions, the total deuterium incorporation of 0.93 was close to one, indicating that H/D exchange at the NP surface is slow; a total deuterium incorporation of 1.76 was also obtained when the hydrogenation was performed in toluene using hydrogen generated from NaBD4 in D2O, which is reassuringly close to the predicted value of two. Finally, the generation of minor amount of isotopologues containing -CHD2 and -CD3 (III, IV, VII and VIII) from each of these deuterium labelling experiments is consistent with H/D scrambling via facile reversible \u03b2-hydride elimination from a surface M-CPh2CH2D species, reinsertion of the resulting Ph2C=CHD into a surface M-D followed by reductive elimination from (D)HPd-CPh2CH3-nDn (n\u00a0=\u00a02, 3); full details of the relative proportions of each isotopologue obtained from these labelling studies are summarised in the supporting information. A higher than stoichiometric incorporation of deuterium recently reported for the hydrogenation of styrene using 'HD' generated from the hydrolysis of tetrahydroxydiboron with D2O using quantum dot stabilised PtNPs was also attributed to facile reversible alkene insertion-extrusion involving metal-hydride/deuteride species [166].Recycle studies were conducted with 2 mol% loading of 2e to investigate its activity profile during reuse and thereby its stability and longevity and potential for use in a scale-up system. The practical issues associated with separating and recovering a small amount of catalyst by filtration without loss of material after each run meant that it was not possible to perform a conventional recycle experiment. As such, a reuse experiment was undertaken by monitoring the hydrolysis until gas evolution was complete, the aqueous reaction mixture was then charged with a further portion of NaBH4 and the gas evolution monitored; this sequence was repeated to map the catalyst efficacy against reaction time and reuse number. While the comparative conversions and TOFs shown in Fig.\u00a011\na, b were obtained during the first 2 min of the hydrolysis to enable a meaningful comparison between runs, complete conversions were obtained for each run within 4 min. The resulting gas evolution-time profile and corresponding conversion-cycle number profile in Fig.\u00a011a, b shows a minor but gradual drop in conversion across five reuses, from 89% after 2 min in the first run to 78% after the same time in the 5th run. The drop in catalyst activity in successive runs, defined as the percentage reduction in the initial TOF, shows that 2e retains 71% of its activity across five reuses (Fig.\u00a011b, red); this is comparable to recycle studies reported for other noble metal nanoparticle catalysts including; RuNPs immobilised in ZIF-67 [77], PtCoNPs supported on carbon nanospheres [167], ruthenium nanoparticles immobilised within the pores of amine-functionalised MIL-53 [76], ruthenium supported on graphite [139], RuCo nanoclusters incorporated in PEDOT/PSS polymer [168], RuNP stabilized by polyvinylpyrrolidone, zeolite-confined RuNPs [64], click dendrimer-stabilized PtCo, Rh and Pt nanoparticles and gold-transition metal nanoalloys [72,73,78,79,80,165] and Ru-RuO2/C [141].Sneddon et\u00a0al. previously reported that the use of a borate buffered solution for the rhodium-catalysed release of hydrogen from ammonia triborane extended the catalyst lifetime such that Rh/Al2O3 showed little change in the hydrogen release rate over 11 cycles [169]. Following this lead, a preliminary comparative recycle hydrolysis conducted in freshly prepared aqueous borate buffer (pH maintained between 7.2 and 8) containing 0.28 M NaBH4 and 1 mol% 2e resulted in a marked increase in activity as evidenced by the initial TOF of 133 moleH2.molRu\n\u22121.min\u22121 obtained for the first run compared with 95 moleH2.molRu\n\u22121.min\u22121 for the corresponding reaction in water. The initial TOF increased to 146 moleH2.molRu\n\u22121.min\u22121 in the second run but then decreased gradually in subsequent cycles to 109 moleH2.molRu\n\u22121.min\u22121 in the final run (Fig.\u00a011d); even though this represents a 26% reduction in activity over the 5 cycles, it remains higher than the TOFs obtained in water under the same conditions. Interestingly, the data in Fig.\u00a011c, d also shows that the conversion-time profile changes quite dramatically in successive cycles such that the conversion increases from 54% after 5 min in the first run to 80% at the same time interval in the final run; in contrast, for reactions conducted in the absence of buffer, conversions decreased gradually in successive runs (Fig.\u00a011b). A hydrolysis catalysed by 1 mol% 2e was also conducted in 0.34 M boric acid to provide a benchmark as the borate buffer solution was prepared with this concentration of boric acid and, under otherwise identical conditions, the initial TOF of 66 moleH2.molRu\n\u22121.min\u22121 was significantly lower than that obtained in the aqueous borate buffer solution (See Fig. S7 in the supporting information). Further studies are currently underway to identify an optimum buffer for this reaction and to develop an understanding of the changes in the conversion-time profile in consecutive runs as well as the origin of the enhancement in activity obtained when the catalysis is conducted in aqueous buffer.ICP-OES analysis of the aqueous reaction mixture recovered after the fifth run revealed that the ruthenium content was below the detection limit, suggesting that the reduction in activity was unlikely to be due to leaching of the ruthenium to generate a homogeneous species that was less active. Hot filtration studies were also conducted to explore whether soluble ruthenium species might be responsible for the gas evolution. Following a typical protocol, a hydrolysis reaction catalysed by 2 mol% 2e was filtered through a 45-micron syringe filter at ca. 50% conversion. The hydrogen liberated from the filtrate was monitored and corresponded to the background hydrolysis in the absence of catalyst (Fig.\u00a012\n, blue line), indicating that the active species had been removed in the filtration i.e. it is heterogeneous, and that leaching does not generate active soluble ruthenium species. In a complementary hot filtration study a catalytic hydrolysis that reached completion was filtered through a syringe filter (0.45 \u03bcm) and a fresh portion of NaBH4 added to the filtrate. The hydrogen liberated also corresponded to the uncatalyzed hydrolysis providing further support that the active species is heterogeneous (Fig.\u00a012, orange line). TEM analysis of the catalyst isolated after the fifth run revealed that the ruthenium nanoparticles remained essentially monodisperse with a mean diameter of 1.8 \u00b1 0.5 nm compared with 1.8 \u00b1 0.6 nm for the freshly prepared catalyst (Fig.\u00a012b) which suggests that agglomeration is not responsible for the drop in conversion with increasing use.There have been several reports that the sodium metaborate tetrahydrate by-product generated during the hydrolysis of NaBH4 deactivates the catalyst by adsorption on the surface [67,71,76,80,135,170-172], although Wie has demonstrated that the activity of deactivated Ru on nickel foam catalyst can be partially replenished by washing the catalyst with deionised water and completely replenished by washing with HCl to remove the NaBO2\n[135]. As such, a series of poisoning studies were undertaken to examine the influence of the by-product on catalyst performance; this involved pre-stirring an aqueous suspension of 2e with 100 equivalents of sodium metaborate prior to addition of NaBH4 and monitoring the progress of the reaction as a function of the pre-stirring time. A 11B NMR spectrum of a typical reaction solution confirmed that the tetrahydroxyborate anion B(OH)4 was the sole by-product as the spectrum contained a single sharp resonance at \u03b4\u00a02.2 ppm [162,173]; no other species such as partially hydrolysed intermediates were detected. A comparison of the hydrogen evolution in the absence of NaBO2 against the corresponding reaction with added NaBO2 as a function of the pre-stirring time (Fig.\u00a013\na,b) confirms that the addition of metaborate passivates the catalyst. The conversions obtained after a reaction time of 2 min and the corresponding initial TOFs as a function of pre-stirring time reveal that the passivation is instantaneous as the TOF drops from 84 moleH2.molRu\n\u22121.min\u22121 in the absence of NaBO2 to 80 moleH2.molRu\n\u22121.min\u22121 immediately after the addition of the NaBO2 with no pre-stirring (time\u00a0=\u00a00 min); the TOFs continue to drop gradually to 57 moleH2.molRu\n\u22121.min\u22121 as the pre-stirring time was increased to 60 min.Finally, the formation of NaBO2 can also be monitored by measuring the pH of the reaction solution as a function of time for the catalytic hydrolysis of a 0.028 M solution of NaBH4 using 2 mol% of 2e. Fig.\u00a014\n shows that the pH of the reaction solution clearly maps to the conversion with a gradual increase from pH 8.3 at time\u00a0=\u00a00 min, recorded immediately after addition of the NaBH4, to pH\u00a0=\u00a011.1 after ca. 2.5 min when the gas evolution had finished; for comparison a 0.028 M solution of NaBO2 in the absence of catalyst or NaBH4 has a pH of 11.30, which correlates with the pH of a hydrolysis reaction at high conversion.Ruthenium nanoparticles stabilized by polymer immobilized ionic liquids catalyze the hydrolytic evolution of hydrogen from sodium borohydride; catalyst stabilized by an amino-modified imidazolium-based polymer was the most active with an initial TOF of 171 moleH2.molRu\n\u22121.min\u22121, this is among the highest to be reported for a RuNP-based system. Kinetic studies revealed that the reaction was first order in catalyst as well as sodium borohydride at low hydride/catalyst mole ratios but zero order with respect to NaBH4 concentration with high hydride/catalyst mole ratios. The apparent activation energies of 38.9 kJ mol\u22121 to 51.8 kJ mol\u22121 are in the region commonly reported for the platinum group metal catalyzed hydrolysis of hydrogen rich boron derivatives; the apparent activation energy of 38.9 kJ mol\u22121 for RuNP@NH2PIILS is lower than each of the other catalysts tested and consistent with its higher initial TOF. A kinetic isotope effect (kH\n/kD\n) of 2.3 obtained for reactions conducted in H2O and D2O and a kH\n/kD\n of 2.72 for reactions conducted with NaBH4 and NaBD4 at a low catalyst/hydride mole ratio indicate that both the O-H and B-H bonds are activated by the ruthenium atoms in the rate limiting step, most likely via a concerted oxidative addition-hydride transfer involving the surface-coordinated hydrogen-bonded [BH3H-]\u2014\u2013H-OH ensemble rather than oxidative addition of both the O-H and B-H bonds. Interestingly though, the kH\n/kD\n of 0.67 obtained from comparing the initial rates of hydrolysis for NaBH4 and NaBD4 under conditions of catalysis, i.e. at a high catalyst/hydride mole ratio, is an inverse KIE which would be consistent with a surface-coordinated borohydride activating an O-H bond of water in the hydrogen-bonded ensemble prior to rapid hydride transfer. Reuse experiments showed that RuNP@NH2-PIILS retains 79% of its activity over 5 runs and poisoning studies conducted by adding NaBO2 to a catalytic reaction suggest that the reduction in activity is most likely due to passivation of the catalyst by absorption of the metaborate by-product on the nanoparticle surface. A tandem hydrogenation of 1,1-diphenylethene in d4-MeOH with hydrogen generated from the catalytic hydrolysis of NaBH4 in D2O gave a mixture of all eight possible isotopologues with a total deuterium incorporation greater than one while the use of toluene for the hydrogenation using NaBH4/D2O gave a total deuterium incorporation close to one. This is consistent with slow H/D exchange at the NP surface and fast H/D exchange on the surface of the Pd/C coupled with H/D scrambling via facile reversible beta hydride elimination-reinsertion during the hydrogenation. This programme is currently exploring the use of PIIL supported bimetallic nanoparticles with varying proportions of noble and earth abundant metals to establish how the composition of the NP influences catalyst performance with the aim of identifying an optimum synergism that will be suitable for use as a hydrogen generation system for portable applications of proton exchange membrane fuel cells (PEMFC). In addition, PIILs are an ideal support to investigate how polymer properties such as charge density, the number and type of heteroatom donor and functionality, porosity and hydrophilicity influences the size, morphology, and efficacy of the nanoparticles as well as to tailor catalyst-support interactions to enhance efficacy. Ultimately, this catalyst technology will be extended to include the hydrogen evolution reaction to develop stable, durable, highly active cost-effective catalysts for use in AEM based electrolysers and fuel cells.The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Reece Patterson, Anthony Griffiths reports financial support was provided by Engineering and Physical Sciences Research Council.R.P. gratefully acknowledges the Engineering and Physical Sciences Centre for Doctoral Training in Renewable Energy Northeast Universities (\u2018ReNU\u2019)\nEP/S023836/1 for a studentship and A.A. thanks Taibah University, Saudi Arabia for a Scholarship. We also thank (Dr Tracey Davey) for the SEM images (Faculty of Medical Sciences, Newcastle University) and Zabeada Aslam and the Leeds electron microscopy and spectroscopy centre (LEMAS) at the University of Leeds for TEM analysis. This research was funded through a studentship (Anthony Griffiths) awarded by the Engineering and Physical Sciences Centre for Doctoral Training in Molecules to Product (EP/SO22473/1). The authors greatly acknowledge their support of this work. This article is dedicated to the memory of Professor Stephen A. Westcott (Canada Research Chair holder in the Department of Chemistry & Biochemistry, Mount Allison University, Canada) who recently passed away; a fantastic scientist, a great ambassador for chemistry teaching and research in Canada and across the globe, a generous, genuine and kind human being but most of all the best of friends.Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.mcat.2022.112476.\n\n\nImage, application 1\n\n\n\n", "descript": "\n                  Ruthenium nanoparticles stabilised by polymer immobilized ionic liquids catalyse the hydrolytic release of hydrogen from sodium borohydride. The composition of the polymer influences performance and ruthenium nanoparticles stabilised by an amine-decorated imidazolium-based polymer immobilised ionic liquid (RuNP@NH2-PIILS) was the most efficient with a maximum initial turnover frequency (TOF) of 177 moleH2.molRu\n                     \u22121.min\u22121, obtained at 30\u00b0C with a catalyst loading of 0.08 mol%; markedly higher than that of 69 molH2.molRu\n                     \u22121.min\u22121 obtained with 5 wt% Ru/C and one of the highest to be reported for a RuNP catalyst. The apparent activation energy (Ea) of 38.9 kJ mol\u22121 for the hydrolysis of NaBH4 catalysed by RuNP@NH2-PIILS is lower than that for the other polymer immobilized ionic liquid stabilised RuNPs, which is consistent with its efficacy. Comparison of the initial rates of hydrolysis in H2O and D2O catalysed by RuNP@NH2-PIILS gave a primary kinetic isotope effect (k\n                     H/k\n                     D) of 2.3 which supports a mechanism involving rate limiting oxidative addition of one of the O-H bonds in a strongly hydrogen-bonded surface-coordinated [BH3H\u2212]\u2014-H2O ensemble. The involvement of a surface-coordinated borohydride is further supported by an inverse kinetic isotope effect of 0.65 obtained from a comparison of the initial rates for the hydrolysis of NaBH4 and NaBD4 under the conditions of catalysis i.e., at a high hydride/catalyst mole ratio. Interestingly though, when the comparison of the initial rates of hydrolysis of NaBH4 and NaBD4 was conducted in dilute solution with a hydride/catalyst mole ratio of 1 a kinetic isotope effect (k\n                     H/k\n                     D) of 2.72 was obtained; this would be more consistent with concerted activation of both an O-H and B-H bond in the rate limiting step, possibly via a concerted oxidative addition-hydride transfer in the surface-coordinated hydrogen-bonded ensemble. Catalyst stability and reuse studies showed that RuNP@NH2-PIILS retained 71% of its activity over five runs; the gradual drop in the initial TOF with run number appears to be due to passivation of the catalyst by the sodium borate by-product as well as an increase in viscosity of the reaction mixture rather than leaching of the catalyst.\n               "}
{"full_text": "Direct photocatalysis using plasmonic metal (gold, silver, copper, or aluminum) nanoparticles (NPs) under visible-light irradiation, also called plasmonic catalysis, has drawn significant interest in the last decade.\n1\u201313\n When plasmonic metal NPs are illuminated, they can efficiently absorb visible light due to the excitation of localized surface plasmon resonance (LSPR), where the conduction electrons of the NPs collectively oscillate with the electromagnetic (EM) field of incident light.\n14\n The plasmon-resonant excitation generates energetic hot electrons that can trigger chemical reactions of a reactant adsorbed on the plasmonic NPs, under mild reaction conditions.\n1\u20137,12,13\n However, only a limited number of the transformations can be directly catalyzed by NPs of plasmonic metals, especially when compared to the range of transformations accessible with transition-metal complexes in homogeneous catalysis systems. For specific reactions, photocatalysts of alloy NPs of a plasmonic and a transition metal of inherent catalytic activity for specific reactions have been developed to expand the application of the plasmonic photocatalysis to a wider range of selective organic synthesis reactions.\n15\n\nTransition-metal complexes are widely used for the homogeneous catalytic synthesis of many important organic compounds.\n16\u201319\n Combining the optical function of plasmonic metal NPs with the inherent bond-forming or bond-breaking ability of transition-metal complexes may enable the transition-metal complexes to catalyze reactions with the light energy harvested by plasmonic metal NPs under mild conditions. With traditional catalysis reactions, intense thermal heating is often required to bridge reaction activation barriers and achieve sufficient catalytic efficiency. When plasmonic metal NPs are introduced as antennas to direct energy to the active reaction site, there may be no need for such intensive heating. This difference in the way catalysis is expected to prove beneficial for selective synthesis. In a combined system, the plasmonic NP surfaces do not act as catalytically active sites themselves; instead, the NPs facilitate chemical transformations via the transfer of energy and light-generated hot electrons to reactive transition-metal complex sites, as schematically illustrated in Figure\u00a01\n. The adsorption and activation of reactants at the complex sites of the new catalysts are different from those on metal NP surface.It is well known that the LSPR light absorption of plasmonic NPs can generate EM near-fields, with intensities much higher than that of the incident radiation.\n4,6\u201311\n The plasmon field enhancement can significantly change the light-matter interaction in excitonic systems.\n14,20\n For instance, the field intensity in a narrow junction between Ag NPs has been predicted to be 106 times that of the incident light alone.\n4\n These narrow junctions between closely spaced NPs are the so-called hot spots.\n21,22\n Whether the high intensity of EM fields at the hot spots can be utilized for catalysis by surface-bound active sites has not been investigated. We consider that the strong EM fields at hot spots may radically change the interaction between reactant and transition-metal complexes when the complexes are immobilized within the hot spots. A strong interaction may direct the energy of the incident light to the reaction site. Also, since the number density of plasmonic NPs is larger at hot spots than other regions in a sample, the number of hot electrons at hot spots is also predicted to be higher. These properties may facilitate a chemical transformation and thus improve the catalytic efficiency of the metal complexes.Electron transfer may contribute to the catalysis process by changing the oxidation state of the transition metal complexes, a key mechanism that can promote catalysis of many reactions.\n23\n In a combined system, the transfer of energy and hot electrons is determined by energy alignment: the energy of photoexcited electrons have to be sufficiently high to be injected into the metal complex sites (via a medium) and change the metal oxidation state. The light harvested by plasmonic NP antennas promotes conduction electrons to energy levels above the Fermi level,\n4,12,14\n and energy distribution of hot electrons depends on the energy of the incident photons and the light absorption mechanism of the plasmonic metal NPs.In addition to the intraband electron excitation in the plasmonic NPs due to the LSPR effect, interband excitation of d \u2192 sp transitions by visible light can generate electron-hole (e\u2013-h+) pairs in Au NPs, which can participate in a chemical reaction of adsorbate on the NP surface.\n24\n The interband excitation in NPs of non-plasmonic metals can also facilitate chemical reactions.\n25\n It is of great interest to know whether the interband excitation can promote catalytic performance of the metal complexes at a distance. Small Au NPs possess only weak LSPR absorption and the light absorption of the NPs exposed to short wavelengths (<450\u00a0nm) induces predominantly interband excitation.\n12\n This property of small Au NPs can therefore be used to investigate whether interband excitation can promote the catalysis even when the NP surfaces are not catalytically active sites themselves. Hence, the mechanisms of the systems with the plasmon-antenna-promoted catalysts will be different from those of the plasmonic metal NP catalysts and homogeneous metal complex catalysts. These features are of great interest from a fundamental research perspective.To verify the efficacy of the proposed strategy of exploiting the antenna-effect of plasmonic metal NPs and to develop protocols for transition-metal-catalyzed reactions under mild conditions, we designed a structure as illustrated in Figure\u00a01. In this structure, Ni2+ complexes and Ag or Au NPs are immobilized on \u03b3-Al2O3 nanofibers. Immobilizing Ni2+ complexes to the \u03b3-Al2O3 supports maintains the complexes at fixed locations relative to the metal NPs, allowing us to investigate the influence of the high intensity of EM near-fields and transfer of hot electrons from the metal NPs to Ni2+ ions. Such a structure can achieve stable photocatalytic performance and make the catalysts recyclable.Ni2+ complexes have been extensively used as homogeneous catalysts.\n26\u201329\n Nickel is a common, inexpensive transition metal and has been used to catalyze reductive cleavage of C\u2013O bonds.\n30\u201335\n This reaction was chosen as a model reaction because it is the essential reaction step for the production of high-value aromatic chemicals from biopolymer lignin.\n36\n Aryl ether C\u2013O bonds are relatively unreactive; for example, the dissociation energy of the aryl ether C\u2013O bonds in the \u03b1-O-4 linkage is 218\u00a0kJ mol\u22121.\n31\n So catalysts usually function at high temperatures and hydrogen pressures (>120oC and added hydrogen pressure),\n37\n inevitably yielding saturated hydrocarbons. To avoid undesired hydrogenolysis of aromatic rings and to achieve selective cleavage of the aryl ether C\u2013O bonds, the reaction should ideally be conducted under mild reaction conditions (low temperature and pressure). Hence, the catalysts must be highly active and selective to the C\u2013O bond. Cleavage of C\u2013O bonds in aryl ethers with a photocatalyst [Ir(ppy)2(dtbbpy)]PF6 at room temperature was reported recently.\n38\n Although these types of organometallic homogeneous photocatalysts are efficient, they are generally expensive and their use requires costly recycling processes.In the present study, photocatalysts with immobilized Ni2+ complexes and plasmonic metal NPs were applied to the hydrogenolysis of benzyl phenyl ether that has an \u03b1-O-4 linkage.\n39\n We demonstrate that by varying the metal NP loading on the alumina support, the number of the plasmonic hot spots could be substantially altered. The hot electrons generated by both interband excitation of small Au NPs and intraband excitation (via LSPR effect) of Ag NPs can promote the catalysis. A significant observation is evidence for a light-induced, enhanced chemisorption of the reactant molecules at the catalytic active Ni2+ sites. This is beneficial to the catalytic performance of the Ni2+ sites. We also find that transfer of the hot electrons from the plasmonic metal NPs to the Ni2+ complexes via a \u201cbridge\u201d of the aromatic ring of the reactant is essential to the performance of the photocatalysts.The photocatalysts were prepared following the procedures shown in Figure\u00a0S1. The prefix number of the sample name indicates the plasmonic metal content in weight percentage (wt %), and ASN (alumina-slilane-NH2) is the support of Al2O3 nanofibers grafted with a silane containing an amino group. The metal content in the catalysts was measured by inductively coupled plasma optical emission spectrometry (ICP-OES), confirming the contents of Ni and plasmonic metal (Table S1). For example, the 2.5Au-ASN-Ni2+ catalyst contains 2.5 wt % of Au NPs, and Ni2+ ions are immobilized on the ASN support.The \u03b3-Al2O3 nanofibers are \u223c5\u00a0nm thick and 100\u00a0nm long, which were sintered to form a highly porous framework of randomly oriented fibers.\n40\n The cage-like nanofiber configuration (see Figure\u00a01) could confine the NPs formed within the structure and allow reactant molecules to readily diffuse to the Ni2+ complex reaction sites in the vicinity of the metal NPs through inter-fiber voids. Plasmonic NPs (Ag or Au) and Ni2+ were immobilized on the alumina fibers in the procedure schematically illustrated in Figure\u00a01. The transmission electron microscopy (TEM) images indicate that the metal NPs were dispersed throughout the \u03b3-Al2O3 fiber support (Figures 2A and 2D). The supported Au NPs were relatively small, most of them smaller than 5\u00a0nm (Figure\u00a02B and inset of Figure\u00a02A), the mean particle size is 2\u00a0nm. The particle size distributions of the Ag NPs (Figure\u00a02E and inset of Figure\u00a02D) were broader, most of the Ag NPs were smaller than 15\u00a0nm and the mean particle size is about 9\u00a0nm. Line scan analysis (energy dispersion X-ray spectroscopy [EDX]) of plasmonic metal and Ni compositional fluctuations in Figures 2C and 2F indicate that the Ni2+ ions were immobilized on the bare support between the metal NPs, instead of accumulating on the surface of the plasmonic metal NPs.The other structural and chemical properties of the catalyst were characterized by nitrogen adsorption, Fourier transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS) and are shown in Table S2 and Figures S2\u2013S6. The results verify that the photocatalysts possess the assembled structure as we designed.Diffuse reflectance ultraviolet-visible (DR UV-vis) spectra of the samples are shown in Figures 2G and 2H. \u03b3-Al2O3 fibers and grafted \u03b3-Al2O3 fibers (ASN) have no obvious absorption in the visible-light region (wavelength > 400\u00a0nm), while two broad bands peaked at wavelengths of 378 and 652\u00a0nm are clearly observed for Al2O3-silane-NH2-Ni2+ (ASN-Ni2+). These two bands can be assigned to the 3A2g\n(F) \u2192 3\nT\n1g\n(P) and 3A2g\n(F) \u2192 3\nT\n1g\n(F) transitions for distorted octahedral Ni2+ complexes.\n41,42\n These absorption bands in 2.5Au-ASN-Ni2+ and 4.3Ag-ASN-Ni2+ samples overlap with the strong absorption of the supported metal NPs in the samples. Both 4.3Ag-ASN-Ni2+ and 4.3Ag-ASN samples have absorption bands centered at 405\u00a0nm, which are the LSPR absorption of Ag NPs.\n8\n The 2.5Au-ASN-Ni2+ catalyst shows a broad light absorption and the absorption intensity decreases when the wavelengths are longer than 550\u00a0nm. No sharp characteristic LSPR peak of Au NPs at \u223c530\u00a0nm\n43\n is observed. This could be attributed to that the small Au NPs (with a mean size of 2\u00a0nm). The interaction of Au NPs with the amino group of the grafted silane also changes the electron density of the small Au NPs and their light absorption due to LSPR effect\n44,45\n more significantly than that of the relatively larger Ag NPs (with mean size of 9\u00a0nm).\nFigure\u00a03\n shows the catalytic performance for the C\u2013O bond cleavage of benzyl phenyl ether (one of the \u03b1-O-4 model compounds) by the photocatalysts prepared with different amounts of plasmonic metal NPs (parameters given in the caption). Both the 2.5Au-ASN-Ni2+ and 4.3Ag-ASN-Ni2+ catalysts gave high conversion of the ether (98% and 96%, respectively) under visible-light irradiation of 0.96\u00a0W cm\u22122, while no reaction was observed in the dark. The selectivity toward production of phenol was around 50%, indicating that a homolytic C\u2013O bond scission of benzyl phenyl ether occurred. There are no products of undesired hydrogenolysis of aromatic rings; the high selectivity is because the moderate conditions of the photocatalytic process induce over-reduction.The catalytic activities of a sample prepared with Ni2+ complexes but without the metal NPs, a sample with the metal NPs but no Ni2+ ions yielded very low conversion (6% or nil). This highlights a significant synergistic effect occurred when the Ni2+ complexes were immobilized in the presence of plasmonic metal NPs: when the light absorption by the metal NPs combines with the inherent catalytic capability of Ni2+ complexes, this results in superior photocatalytic activity of the composite. It is also evident that conversion of the reactant increased with increased metal NP loading of the catalyst. The synergistic effect becomes significant when the metal NP content is\u00a0high.Simply the mixture of Ni(NO3)2 and 4.3Ag-ASN exhibited negligible activity. Hence, the possibility that Ni2+ species in liquid phase are active for the catalytic reaction is excluded. The mixture of Ni(NO3)2 and 2.5Au-ASN exhibited a low conversion of about 20%. It appears that the gold NPs can interact with Ni2+ species in liquid while Ag NPs do not. The detailed reason is unknown. In those control experiments, the overall contents of Ni2+ ions and silver in the mixture of Ni(NO3)2 and 4.3Ag-ASN are the same as those in the 4.3Au-ASN-Ni2+. This is also the case for the Au sample. The significant differences in catalytic performance between the mixtures and our designed catalysts suggest that the interaction between metal NP-ASN and immobilize Ni2+ ions has substantial impact on the overall photocatalytic process, the reaction is a complete heterogeneous catalysis process for Ag-ASN-Ni2+ catalysts and proceeds mainly by heterogeneous catalysis process for Au-ASN-Ni2+ catalysts. Plasmonic NPs considerably enhance the catalytic activity of Ni2+ complexes only when they are immobilized in the close proximity to the NPs.It was reported that photothermal effect of plasmonic NPs could elevate the temperature of the NPs, which may play a critical role in the catalysis.\n46\n To understand if local heating will contribute to the Ni2+ catalysis, we conducted the photo-reaction using the ASN-Ni2+ catalyst at elevated temperatures (80\u00b0C \u2013120\u00b0C); only 2% of conversion increase of C\u2013O bond cleavage was observed (see Table S3), which means that a 40\u00b0C increase in reaction temperature could not enhance the catalytic activity of the Ni2+complexes in our reaction system. Therefore, photothermal effect caused by irradiation could not drive the reaction in the present study. The hot electrons from the illuminated metal NPs may play the critical role in activating the aryl ether C\u2013O bonds as is the case for some other plasmonic catalysts reported in the literature.\n1\u201313\n But the Au and Ag NPs are not the active catalytic sites in the present study, as evidenced by the experimental data (see Figure\u00a03). Hence, these catalysts may be expected to operate differently from the well-known mechanism where light-generated hot electrons promote reactions.\n2\u20138\n\nWe compared the catalytic performance enhancement of Ni2+ when the number density of metal NPs are different by using catalysts with different metal NP contents and maintaining the metal NP content in the reaction systems and other experimental conditions identical. For instance, the reaction vessel was loaded with 66\u00a0mg of 1.3Ag-ASN-Ni2+ and 20\u00a0mg of 4.3Ag-ASN-Ni2+. The quantities of the two catalysts were calculated from the Ag content (Table S1). Figure\u00a04\nA shows that the percent conversion of reactant in the reaction driven by 4.3Ag-ASN-Ni2+ is significantly higher than by 1.3Ag-ASN-Ni2+ although the content of immobilized Ni2+ ions in the latter is 3.8 times of that in the former (derived from data in Table S1). The turnover number (TON) with respect to the Ni2+ sites in 4.3Ag-ASN-Ni2+ was\u00a0calculated from the conversion, being >18 times higher than that in 1.3Ag-ASN-Ni2+. Given that the Ag NP content in the two systems is the same, the overall Ag NP content in the catalytic system cannot be the reason for the large difference in the catalytic performance.The major difference between the two photocatalytic systems shown in Figure\u00a04A is that the Ag NPs are more closely packed in 4.3Ag-ASN-Ni2+. The junctions between closely spaced NPs can generate \u201chot spots\u201d where the strong EM field coupling of the closely spaced NPs generates huge EM field enhancement.\n4,6\n In a unit volume of\u00a0catalyst, the Ag NPs in the 4.3Ag-ASN-Ni2+ is more than three times the Ag NPs in the 1.3Ag-ASN-Ni2+ catalyst, and therefore, there are a greater number of nano-junctions between Ag NPs. The results in Figure\u00a04A can be interpreted as a manifestation of the fact that the photocatalytic activity heavily depends on the number of Ag NPs\u2019 nano-junctions in the photocatalysts. An environment of high number density of Ag NPs can significantly promote catalytic performance of the immobilized Ni2+ complexes around the particles.The effect of catalyst mass on photocatalytic activity was examined, and the results are provided in Table S4. The conversion increased as more catalysts were used. But the conversion did not increase linearly with catalyst mass. The deviation from the linearity may be attributed to the screen effect.\nFigure\u00a0S7 shows EDS mapping results of the 4.3Ag-ASN-Ni2+ photocatalyst, which reveals the existence of closely spaced NPs, consequently supporting the role played by plasmonic hot spots.\n21,22\n Analysis of the results (Figure\u00a0S7) indicates the distribution of plasmonic hot spots regions overlapped with concentrations of surface-bound Ni2+ complexes co-located in the plasmonic hot-spot regions.A similar phenomenon can be found in the xAu-ASN-Ni2+ photocatalytic systems as seen in Figure\u00a04B. However, the effect of high number density of Au NPs on the catalytic performance is weaker than that of Ag NPs. The performance improvement caused by increasing the number density of small Au NPs (from 0.7 wt % Au to 8.4 wt % Au; see TEM images in Figure\u00a0S8) is less than that caused by increasing the number density of Ag NPs (from 1.3 wt % Ag to 4.3 wt % Ag; see TEM images in Figure\u00a0S9).We performed simulations of the near-field enhancement of single Au NP (2\u00a0nm) and dimer Au NPs (2\u00a0nm) with 1\u00a0nm gap between them (Figure\u00a0S10) by using an electrostatic eigenvalue method; the details of the simulation methodology are from Davis and G\u00f3mez.\n47\n Under 400\u00a0nm excitation, the intensity of EM fields at the surface of an isolated Au NP is \u223c10 times larger than the field intensity of the incoming photon flux, while the enhancement of EM field occurring in the middle of the dimer is \u223c12 times. In contrast, the EM field localization upon LSPR excitation of Ag NPs is significantly higher (Figure\u00a0S11). The intensity of EM fields at the surface of an Ag NP (9\u00a0nm) is \u223c300 times larger than the field intensity of the incoming light, while the enhancement of EM field occurring at the hot spot of the dimer is \u223c8,000 times. The hot spot regions will have high concentrations of energetic electrons (because the high number density of the plasmonic NPs) and interaction of the EM near-field with reactant in these regions is likely to be very strong. Thus, these sites are considered most important for plasmonic photocatalysis.We note that our wet-chemistry synthesis approach results in distributions of sizes of the plasmonic metal particles and a range of inter-particle distances in the catalysts as previously reported.\n1,2\n Uniform metal particle size and inter-particle distance is not a prerequisite for formation of plasmonic hot spots.\n46\n The 4.3Ag-ASN-Ni2+ catalyst has many more hot spots, compared to 1.3Ag-ASN-Ni2+ as shown in Figure\u00a0S9. The intense field enhancement at the hot spots between the metal NPs clearly influences the catalytic performance of Ni2+ complexes under mild reaction conditions, and the results of Figure\u00a03 represent a statistical average which is dominated by the response augmented by plasmonic hot spots.The existence of hot spots has also been experimentally observed. Highly intense EM near-fields affect the properties of molecules within them;\n48\n for example, the intensity of signals in surface-enhanced Raman scattering (SERS) of the molecules.\n48\u201350\n SERS predominantly originates from EM near-field enhancement, and its intensity is approximately proportional to the square of the EM field intensity. Thus, the intensity of SERS active peaks of reactant is expected to be commensurate with intensity of the enhanced EM near-field around Ag NPs experienced by the reactant molecules. The Raman spectra of benzyl phenyl ether adsorbed on the catalysts with different Ag NP contents were compared (seen in Figure\u00a05\nA).The results show that when the aryl ether molecules were adsorbed on the catalysts with a high content of Ag NPs (4.3Ag-ASN-Ni2+), peaks at 1,153\u00a0cm\u22121 ascribed to the aryl ether C\u2013O bond vibration become much more intense, even when compared to their concentrated counterpart (pure benzyl phenyl ether). No SERS signals were detected with the catalyst samples prepared with a lower Ag NP content (1.3Ag-ASN-Ni2+ and 1.3Ag-ASN) or with small Au NPs. Given that the photocatalysts have a similar specific surface area (Table S2), the adsorption of the aryl ether molecules on the alumina support has no obvious effect on the Raman signal intensity. The SERS results can be attributed to three possibilities: (1) there is a higher concentration of the aryl ether adsorbed on the samples with a higher Ag NP content; (2) the EM near-field is much more intense (there are more hot spots) in the photocatalysts with increased number density of Ag NPs, and this significantly enhances the Raman signal of adsorbed molecules; or (3) the pronounced Raman signals are due to synergistic effect of the two reasons above. The near-field enhancement of the sample with small Au NPs (the 2.5Au-ASN-Ni2+) is relatively weak as indicated by the simulation results and the fact that there is no LSPR absorption peak at about 520\u00a0nm wavelength. This is consistent with the result that no enhanced Raman signal is observed.The reactant adsorption is a key step prior to surface reaction in heterogeneous catalysis, the impact of light irradiation on the adsorption has not been extensively investigated to date. To explore whether the light irradiation can affect reactant adsorption, we conducted adsorption experiments of benzyl phenyl ether on different catalysts in the dark and under light irradiation with different light intensities. The results are summarized in Figure\u00a05B. The adsorption amount is given by the percentage of the initial concentration of the reactant before light irradiation, which was adsorbed by the catalyst under irradiation. The adsorption capacity is calculated based on the Ni amount collected in Table S1. In the experiment, the cleavage reaction did not proceed as KOH was not added. No chemicals other than those added were detected. The adsorption capacity (AC) was calculated as\n\n\n\nAC\n\n(\n%\n)\n\n\u00a0\n=\n\u00a00\n.\n01\n\u00a0\u00d7\u00a0\n\n(\n\n\n\n\nC\n0\n\n\u2212\nC\n\n\n\nC\n0\n\n\nn\n\nNi\n\n\n\n\n\n)\n\n\u00a0\u00d7\u00a0\n100\n,\n\n\n\nwhere 0.01 is the total molar amount of benzyl phenyl ether, C0 and C are the concentrations of the ether before and after adsorption experiment, respectively, and nNi is moles of nickel (calculated by the data shown in Table S1). The adsorption properties of ASN-Ni2+, 1.3Ag-ASN-Ni2+, and 4.3Ag-ASN-Ni2+ catalysts are compared in Figure\u00a05B.The initial concentration used for the adsorption experiment (0.01 M) was much lower than that used in the reaction. The amount of Ni2+ (the adsorption site) in these samples was 0.011\u20130.013\u00a0mmol, and as discussed below, only the adsorption on a fraction of the Ni2+ sites, which are in the intense EM fields, was influenced by light irradiation. When a high initial concentration of reactant was added, it was difficult to accurately detect the concentration changes cause by light-induced adsorption.It is also worth noting that only trace amounts of benzyl phenyl ether were adsorbed on the illuminated 4.3Ag-ASN catalyst, the sample without Ni2+ but with high Ag NP content. This implies that the EM near-fields alone are not leading to increased reactant adsorption. One needs to consider the interaction between the ether and Ni2+ complex sites of the catalysts too. We measured FTIR spectra of the catalysts after the ether adsorption experiments (Figure\u00a0S12). For the ASN-Ni2+ samples that adsorbed the ether in the dark or under light irradiation of intensity of 1.11\u00a0W cm\u22122, the position of the peak located at 1,151\u00a0cm\u22121 (which is assigned to the C\u2013O bond stretch mode of the ether) remains unchanged (Figures S12A and S12D).However, for both catalysts containing Ag NPs and the surface-bound Ni2+ complexes (1.3Ag-ASN-Ni2+ and 4.3Ag-ASN-Ni2+), this band is blue-shifted to a higher wavenumber under light-irradiation conditions. This indicates an enhanced interaction can occur between the ether and Ni2+ sites when hot spots are activated (Figures S12E and S12F). Also, the intensity of the band at 1,350\u00a0cm\u22121 (the vibration mode of NO3\n\u2212 that coordinates with Ni2+ sites) decreased markedly after the sample was illuminated (Figures S12B and S12C). This can be explained if NO3\n\u2212 is displaced by the ether during adsorption under visible-light irradiation. The light-induced adsorption of the ether is irreversible since it remained adsorbed after the light was switched off. The specific adsorption sites (Ni2+ sites), the coordination of the adsorbate to the sites, the requirement of activation energy (light harvested by the plasmonic metal NPs) for the adsorption, and irreversible adsorption process suggested chemisorption of the ether at Ni2+ sites.The adsorption of the ether increases significantly with the light intensity increases, on the photocatalyst 4.3Ag-ASN-Ni2+ (Figure\u00a05B). The adsorption of the ether on 1.3Ag-ASN-Ni2+ catalyst (with much less hot spots) is considerably lower, regardless of the light intensities used. These data demonstrated that the ether adsorption depends on the number density of Ag NPs and intensity of the incident light but not on the overall number of Ni2+ complexes in a sample. The higher light intensity and the Ag NPs density determine the stronger EM near-fields generated close to the active sites Ni2+. 1.3Ag-ASN-Ni2+ sample contains slightly more Ni2+ complexes than the 4.3Ag-ASN-Ni2+ sample (Table S1). It follows that the chemisorption only appears at the Ni2+ sites subject to the EM fields of high intensity. In 1.3Ag-ASN-Ni2+, a small number of sites with the sufficient EM field intensity for chemisorption as the sample has small number of hot spots. Increasing the incident light intensity amplifies the intensity of the EM near-fields, and as the number of such sites increases, so does the chemisorption. Such an increase is insignificant for 1.3Ag-ASN-Ni2+ compared with that for 4.3Ag-ASN-Ni2+. To the best of our knowledge, there have been no reports on light-enhanced chemisorption of molecules with irradiation of continuous-wave visible-light of moderate intensity (\u223c1\u00a0W cm\u22122). Very recently, we reported that a plasmonic alloy system can selectively concentrate the reactant molecules to catalyst surface and thus increase the reaction rate by several orders under light.\n51\n It is proven that light irradiation generates an optical plasmon force that can add to van der Waals force and selectively attract reactant molecules to the active sites. In the plasmonic antenna system in the present study, similar optical plasmon force contributes to concentrate the ether to catalyst surface and thus enhances the chemisorption of ether on Ni2+ sites. This provides essential knowledge on how visible-light irradiation significantly promotes the catalytic performance of metal cations by plasmonic metal.Following the chemisorption, the dependence of photocatalytic performance on the intensity of the light source is also investigated in Figure\u00a06\n. The temperature of the reaction mixture was carefully maintained at 90\u00b0C or 80\u00b0C\u00a0\u00b1 2\u00b0C to guarantee that thermal effects could be discounted. It is clear that the higher the light intensity, the greater is the contribution of light irradiation to driving the reaction. When the light intensity is 0.96\u00a0W cm\u22122, over 90% of the conversion is due to light irradiation for the C\u2013O bond cleavage of benzyl phenyl ether. Clearly, light irradiation is the principle driving force of the reaction. Higher light intensity generates more hot electrons in metal NPs and also generate stronger EM field to increase the reactant chemisorption; thus, irradiation-induced enhancement of chemisorption and reactant conversion are correlated.Correlating the light intensity dependence of the chemisorption and the reactant conversion with observation that the catalytic performance is not regulated by the overall Ni2+ content in the photocatalyst system (Figure\u00a04A), we infer that the C\u2013O bond cleavage predominantly takes place on the Ni2+ sites within the hot spots, where the ether molecules chemically adsorbed. The adsorbed ether molecules will experience an intense oscillating EM field, which could facilitate the activation of the molecules for the cleavage reaction. The hot spots can capture both light energy and reactant molecules, have higher concentration of hot electrons and thus, are an ideal location for the catalysis.Besides, as can be seen in Figure\u00a06, increasing the light intensity to higher than 0.76\u00a0W cm\u22122 resulted in a dramatic increase in the conversion of the C\u2013O bond cleavage reaction catalyzed by 2.5Au-ASN-Ni2+ (or 4.3Ag-ASN-Ni2+). There exists an energy threshold to activation of benzyl phenyl ether molecule. It is between 0.59 and 0.76\u00a0W cm\u22122 for 2.5Au-ASN-Ni2+ catalyst at 90\u00b0C and between 0.76 and 0.96\u00a0W cm\u22122 for 4.3Ag-ASN-Ni2+ at 80\u00b0C. The intensity threshold is a common feature in photocatalysis systems mediated by plasmonic NPs and influenced by various factors.\n52\n\nThe catalytic performance variation with the wavelength reveals the energy alignment of the combined system. The action spectrum (Figures 7A and 7B) shows the dependence of the irradiation wavelength on the conversion efficiency of the photocatalytic reactions. The reaction rates were converted to the apparent quantum yields (AQYs), which were calculated as\n\n\n\nAQY\u00a0\n\n(\n%\n)\n\n=\n\n\n\nY\n\nlight\n\n\n\u2212\n\nY\n\ndark\n\n\n\nn\n\n\u00d7\n100\n,\n\n\n\nwhere the Ylight and Ydark are the amount of reactant being converted under light irradiation and dark conditions, respectively;\n15\n n is the number of incident photons. The number of reactant being cleaved in the dark, Ydark, was subtracted from that observed when the system was irradiated, Ylight, to clearly illustrate the contribution of light irradiation to the overall conversion. The AQY is a wavelength-dependent quantity that is given by the ratio of the number of molecules produced to the number of incident photons. By definition, it is a quantity that is independent of the intensity of the incident radiation, and its trend provides physical insight on the mechanism, accounting for the conversion of photonic energy into chemical potential energy.For the Ag-Ni2+ photocatalytic system, the highest conversion rate was achieved when illuminating with a 400\u00a0nm peak wavelength (Figure\u00a07B). According to the UV-vis spectra of the photocatalyst in Figure\u00a02H, the most intense LSPR absorption of Ag NPs in this photocatalyst occurs at 405\u00a0nm, well matched with the LED wavelength with which the highest cleavage reaction rate conversion was observed. Also, the AQY spectrum generally matches the absorption of photocatalyst (Figure\u00a07B). Given the temperature of the reaction mixture and the irradiation energy of each wavelength were carefully maintained identical, the significant AQY variation with irradiation wavelength corroborates that the reaction was mainly driven by light. The results evidence that the enhancement of the catalytic performance is caused by the LSPR absorption of Ag NPs. The LSPR absorption in the range between 400 and 600\u00a0nm can generate hot electrons with sufficient energy to reduce Ni2+ in the complexes.For the 2.5Au-ASN-Ni2+ photocatalyst, the AQYs show a mismatch with the catalyst light absorption: the highest conversion rate was achieved when illuminating with a 400\u00a0nm peak wavelength (single-color LED source with band width\u00a0\u00b1 10\u00a0nm), while there is no absorption peak at this wavelength (Figure\u00a07A). Hence, light absorption is not the sole factor determining the catalytic reaction over the small Au NPs. The high AQY at 400\u00a0nm suggests that there is an energy alignment governing the electron transfer: only hot electrons of sufficient energy can trigger the reaction that takes place at Ni2+ complex sites.According to the diagram shown in Figure\u00a07C, photons with 397\u00a0nm wavelength (\u223c3.12 eV) are able to excite d electrons of Au NPs to the energy level for Ni2+ reduction. Photons with longer wavelengths have insufficient energy to reduce Ni2+ ions. On the other hand, a much lower reaction rate is observed under irradiation with wavelength\u00a0< 380\u00a0nm (Figure\u00a07A). It has been reported that visible-light illumination can make the reduction potential of Ni2+ complex shift negatively.\n53\n We also found that the UV irradiation shifted Ni2+/Ni0 reduction potential by 0.2 eV increasing the difficulty in reduction of Ni2+ complexes. Another possible reason is that light with wavelength around 380\u00a0nm causes the 3A2g(F) \u2192 3T1g(P) transition for the distorted octahedral Ni2+ complex (which exhibits light absorption peaked at 378\u00a0nm). A fraction of light can be absorbed for this transition; accordingly, fewer hot electrons are generated. Consequently, the reduction of Ni2+ complex, which is critical for the C\u2013O bond cleavage as discussed later, proceeds at a slow rate. Therefore, there is a narrow energy window at about 3.12 eV, in which photons generate hot electrons by interband electron excitation in the Au NPs to reduce Ni2+ ions driving the reaction. The energy alignment in Figures 7C and 7D may not be strict as the alignment is affected by other factors such as temperature, pH, and solvent. However, it provides a good approximation of the alignment. As shown in Figures 7C and 7D, isopropanol acts as a sacrificial electron donor and is oxidized into acetone on the metal particles,\n2,15\n accomplishing the electron-hole separation in high efficiency.In contrast, the energy level of d band in silver is much lower than that in gold\n54\n so that the hot electrons generated by interband transition in Ag NPs do not have sufficient energy to reduce Ni2+ ions directly (Figure\u00a07D).The catalysts with Au NPs or Ag NPs have similar architecture but distinct differences in LSPR absorption characteristics. The weak LSPR absorption of the small Au NPs (seen in Figure\u00a02G) can be expected to generate a much weaker EM near-field. The EM field enhancement at the hot spots of small Au NP dimer is also predicted to be weak (Figure\u00a0S10). Hence, the LSPR absorption is possibly too weak to be the main driving force for promoting the reaction. Light-excited interband transition (from d band to sp band) in the small Au NPs should be main mechanism for generating hot electrons that induce the catalytic reaction.Interestingly, light-induced adsorption phenomenon occurred in 2.5Au-ASN-Ni2+ photocatalytic system as well (Figure\u00a0S13). It is known that metal NPs produce large field gradients in a wide wavelength range (not only LSPR wavelength)\n55\n as the oscillating EM field of incident light always changes the charge distribution of metal NPs. Also, electrical charges tend to accumulate in sharp edges and high curvature surface of metal particles (lightening rod effect). The higher charge density at surface of the smaller particles results in sharper gradient around the particles. The two effects are applicable to small Au NPs, leading to the increasing chemisorption under light irradiation.The above observation implies that the reaction is operated via the transfer of energy and light-excited hot electrons from metal NPs to Ni2+ active sites, modulating the oxidation state of Ni2+ in the process. In many reactions catalyzed by the transition-metal complexes, the change in the oxidation state of the transition metal enables the bond-breaking or bond-forming processes (oxidative addition and reductive elimination).\n23\n\nThe electron paramagnetic resonance (EPR) spectra of the photocatalysts shown in Figure\u00a08\n support our inference. The EPR spectra of the photocatalysts (4.3Ag-ASN-Ni2+ and 2.5Au-ASN-Ni2+) with the reactant benzyl phenyl ether and solvent isopropanol (IPA) after purple light irradiation (400\u00a0nm of wavelength) exhibits a six-line signal (traces b and d in Figure\u00a08), which is similar to that observed from the ASN-Ni2+ sample after reduction in H2 atmosphere at 200\u00b0C (trace e in Figure\u00a08). The g-factor (\u223c2.01) is close to the reported Ni NPs with a characteristic of paramagnetic behavior.\n56\n So a fraction of the nickel in this H2 reduced sample exists in the Ni0 state. The paramagnetic signals are not observed in the spectra of other samples, even the system of the photocatalysts in IPA solvent after light irradiation (without the ether, see Figure\u00a0S14). The results indicate that in the presence of the ether, a Ni2+ \u2192 Ni(0) transformation takes place in the photocatalyst under irradiation of 400\u00a0nm. The hot electrons generated on the illuminated metal NPs migrate to the Ni2+ complex and reduce Ni2+ ion to Ni0. However, the results shown in trace d in Figure\u00a0S14 that direct electron transfer from Ag NPs to the Ni2+ complex is not evidenced by the EPR spectrum of the photocatalyst dispersed in IPA solvent in the absence of \u03b1-O-4 reactant under irradiation of 400\u00a0nm wavelength. It seems that the ether molecules are involved in the reduction of Ni2+ to Ni0. Surprisingly, we observed the Ni2+ \u2192 Ni0 reduction after illumination, when the photocatalyst was dispersed in benzotrifluoride (BTF) or toluene (traces f-i in Figure\u00a08) in the absence of the ether, but this phenomenon was not observed when BTF was replaced with N,N-dimethyl-formamide (DMF) (traces j and k in Figure\u00a08). It has been reported that for the C-C stretching modes of unsaturated hydrocarbons, temporary electron transfer from plasmonic metal NPs into the normally unoccupied anti-bonding \u03c0\u2217 orbitals can happen in a SERS system.\n57\n Therefore, the aromatic ring in the molecules of the aryl ether, BTF, and toluene could serve as a \u201cmolecular bridge\u201d for the transfer of the photo-induced hot electrons from the plasmonic metal NPs to Ni2+ sites (inset in Figure\u00a08). The hot electrons have sufficient energy to migrate to the unoccupied anti-bonding \u03c0\u2217 orbitals of the aromatic ring.For the Au-Ni2+ system, the ether molecules are also necessary for the charge transfer process to produce Ni0 state species (comparing the results of trace d in Figure\u00a08 and trace a in Figure\u00a0S14). So generation of hot electrons on Au NPs surface is a prerequisite but not sufficient condition. The transfer of the hot electrons plays a critical role in the Ni2+ reduction.No products were detected when toluene or BTF was used as the solvent (Table S5), suggesting that the IPA is necessary in the reaction as the hydrogen source for the reductive cleavage. The reducing agent IPA adsorbed on the surface of Ag NPs or Au NPs releases H atoms at the surface and is itself oxidized as a sacrificial electron donor in the presence of KOH and acetone was detected.\n2\n Using IPA as reduction agent avoids high pressure H2 that has a safety risk. The aryl ether molecules adsorbed at Ni2+ complex sites, and molecules with aromatic rings could assist transfer of the light-generated hot electrons from the illuminated NPs to Ni2+ sites. The reduced Ni0 state is chemically unstable in the reaction environment because the reductive potential of Ni0/NiII (\u22121.2\u00a0V versus SCE)\n28\n is more negative than that of benzyl phenyl ether (\u22120.62\u00a0V versus SCE, seen in Figure\u00a0S15). Therefore, the reduced Ni0 sites could catalyze the cleavage of the C\u2013O bonds in the aryl ether compound. Given the dependence of the chemisorption on the EM near-field intensity and the required molecular bridge for the hot electron transfer, the reaction should take place predominately at the Ni2+ complex sites which are simultaneously close to the metal NPs and subject to intense EM fields, i.e., within plasmonic hot spots.The efficiency of electron transfer by benzyl phenyl ether bridge should be also correlated to the ratio of NP to Ni2+. The mean density of the Ag NPs and Au NPs are estimated and summarized in the Table S6 and correlated to mean densities of APTMS and Ni complexes. The distributions of supported NPs and Ni2+ complexes based on the mean densities are presented as below. The densities are averaged data, summarized from TEM images in Figures 2, S8, and S9. The Ag NPs density in 4.3Ag-ASN-Ni2+ sample is 8 times of Ag NPs density in 1.3Ag-ASN-Ni2+ sample. The density of the Ag and Au NPs determines the ratio of NP:Ni2+. The higher the NP:Ni2+ ratio, the higher the efficiency of electron transfer by benzyl phenyl ether bridge. Thus, the electron transfer efficiency in 4.3Ag-ASN-Ni2+ is much higher than in 1.3Ag-ASN-Ni2+ because there are 36 Ni2+ corresponds to per Ag NP in 4.3Ag-ASN-Ni2+ and 508 Ni2+ corresponds to per Ag NP in 1.3Ag-ASN-Ni2+. Similarly, the Au NP:Ni2+ ratio in 8.4Au-ASN-Ni2+ is much higher (1:10) than the other two catalysts, 1:17 and 1:47. The electron transfer by benzyl phenyl ether bridge in 8.4Au-ASN-Ni2+ and 2.5Au-ASN-Ni2+ catalysts is more efficient than that in 0.7Au-ASN-Ni2+ catalyst.On the basis of these results, we propose a tentative reaction pathway for the cleavage, the main features are depicted in Figure\u00a09\n. The aryl ether molecules first diffuse within the framework of the photocatalyst and physically adsorb onto the catalyst surface (the adsorption observed in the dark; Figure 5B). Visible-light irradiation induces chemisorption of the ether at Ni2+ complex sites (i-ii), the ether molecules coordinate to Ni2+ ions replacing the NO3\u2212 ions. Then photo-generated hot electrons transfer to Ni2+ sites (via the unoccupied anti-bonding \u03c0\u2217 orbitals of the aromatic ring of the ether) and reduce the ions to Ni0 species (iii). The reduced Ni0 species can catalyze the C\u2013O bond cleavage, as described by literature reports,\n37,58\n yielding the final products (v). Here, IPA is oxidized on the plasmonic metal NPs (iv), similar to the mechanism reported.\n2\n The IPA oxidation releases hydrogen and electrons, which are needed in hydrogenolysis. The synergistic effect between the optical antenna function of plasmonic metal NPs and the catalytic ability of Ni2+ sites dominates the transformation of the aryl ether. The synergistic effect is accredited to light-induced chemisorption of reactant molecules and the generation and function of the hot electrons as shown in Figure\u00a09.The stability and recyclability of the catalysts investigated. The silane was grafted onto the Al2O3 support in toluene at 110\u00b0C, so that the Al\u2013O\u2013Si bond is stable at the reaction temperatures (80\u00b0C and 90\u00b0C). We compared the contents of Si in 4.3Ag-ASN-Ni2+ and reused 4.3Ag-ASN-Ni2+ photocatalysts measured by the EDX (Figure\u00a0S16), they are similar (2.2\u20132.4 wt %). The recycle stability of the photocatalyst is illustrated in Figure\u00a0S16. The 4.3Ag-ASN-Ni2+ catalyst was reused for 3 cycles of C\u2013O bond cleavage of benzyl phenyl ether under light irradiation. The product conversion moderately decreased during recycling, and the leaching of Ni2+ ions during the reaction caused the decrease in Ni content of the used catalyst. The soluble Ni2+ ions were not active for the hydrolysis at 80\u00b0C. The decrease could also be due to the partial loss of Ag NPs on the external surface of aggregates of randomly oriented fibers and leaching of nickel, thus lowering the chemisorption of reactant amount and formation possibility of Ni0 species, as seen in the EDS mapping result displayed in the figure.In summary, with the designed photocatalyst structure we have demonstrated that Au NPs or Ag NPs can act as optical antennas to absorb visible-light and promote the catalytic performance of Ni2+ complex immobilized in the hot spots. The new plasmonic-antenna-promoted catalysts with 2.5 wt % of small Au NPs (\u223c 2\u00a0nm) or 4.3 wt % of Ag NPs (\u223c 9\u00a0nm) exhibited superior catalytic activity for the cleavage of relative stable C\u2013O bonds in benzyl phenyl ether under visible-light irradiation and mild reaction conditions without reduction of the aromatic rings. The results signify a new mode to activate chemical reactions by combining the advantages of plasmonic metal NPs and the chemical bond formation ability of transition-metal complexes, which is complementary to known plasmonic catalysis and transition-metal complex catalysis.We have provided strong evidence suggesting that the catalytic performance is enhanced by the synergy of the following two effects: hot-electron transfer to the catalytically active Ni2+ complex sites and light-enhanced chemisorption of reactant species. Both of these effects are greatly augmented by the strong EM near-field localization attainable with plasmonic hot spots. The energy alignment is commonly a key issue for the hot-electron transfer in the new photocatalysts. We believe these results will stimulate further research into the creation of novel plasmonic-antenna-promoted catalysts that exploit the high chemical reactivity of transition-metal complexes.The chemicals were purchased from commercial suppliers and used as provided: benzyl phenyl ether (Sigma-Aldrich, >98%), isopropanol (Sigma-Aldrich, >99.5%, anhydrous), toluene (Fisher, >99.99%, GC assay), N,N-dimethyl formamide (Sigma-Aldrich, >99.8%, anhydrous), \u03b1,\u03b1,\u03b1-trifluorotoluene (Sigma-Aldrich, >99%, anhydrous), acetic acid (Ajax Finechem, >99.7%), nitric acid (Ajax Finechem, 68%\u201370%), C12-14H25-29O(CH2CH2O)5H surfactant (Sigma-Aldrich), (3-aminopropyl)trimethoxysilane (Sigma-Aldrich, >97%), potassium hydroxide (Sigma-Aldrich, >99.99%), phenol (Ajax Finechem, AR), silver nitrate (Merck, AR), gold chloride trihydrate (Sigma-Aldrich, >99.9%), nickel(II) nitrate hexahydrate (Scharlau, >98%), sodium borohydride (Sigma-Aldrich, >98%), sodium aluminate (Sigma-Aldrich, anhydrous), H2 (Supagas, >99.999%), and Ar (Supagas, >99.99%)The photocatalysts were prepared following the illustration shown in Figure\u00a0S1. \u03b3-Al2O3 nanofibers were prepared following a procedure reported previously\n59\n and used as support of the catalysts. Boehmite (with a chemical formula of AlOOH) nanofibers were prepared from NaAlO2 and then converted to \u03b3-Al2O3 fibers by calcination at 450oC for 5 h. The details of the fiber preparation are provided in the Supplemental Information. Amino groups were then grafted on 3.0\u00a0g of \u03b3-Al2O3 fibers by refluxing in 50\u00a0mL of toluene solution of 9.7\u00a0mmol of (3-aminopropyl)trimethoxysilane (APTMS) for 40 h. The solid sample was collected by washing and filtrating with H2O and ethanol, and finally dried at 60oC (Al2O3-silane-NH2 is abbreviated to ASN).Plasmonic Ag or Au NPs were prepared on ASN support by an impregnation-reduction procedure. For example, in the synthesis of 1.3 wt % of Ag NPs on ASN support, 1.0\u00a0g of the ASN support was dispersed into 133\u00a0mL of deionized water under vigorous stirring with a magnetic stirrer for 20\u00a0min. 27.8\u00a0mL of 0.01\u00a0M AgNO3 aqueous solution was then added to the suspension and stirred for further 20\u00a0min. Next, 74\u00a0mL of NaBH4 (0.038 M) aqueous solution was added dropwise to the suspension over 30\u00a0min under continuous stirring. The suspension was aged overnight, and then the solid was separated by washing with water and dried at 60\u00b0C under vacuum. The obtained sample was labeled as 1.3Ag-ASN.The Ni2+ ions were introduced via complexation with the free amine groups of the silane grafted on the \u03b3-Al2O3 fibers. The procedure for introducing Ni2+ ions to Ag-ASN samples is as follows: 0.5\u00a0g of obtained sample with supported Ag NPs (xAg-ASN, where x denotes the weight percentage of silver in the catalysts) was mixed with 30\u00a0mL of Ni(NO3)2 aqueous solution (0.017 M) by a shaker for 24\u00a0h at room temperature. Then the solid was washed with water 3 times before drying at 60oC under vacuum. The catalysts obtained were labeled as xAg-ASN-Ni2+. This procedure immobilizes Ni2+ complexes on the sample surfaces, including the narrow gaps between metal NPs, where hot spots are likely to be generated. The \u03b3-Al2O3 nanofibers are \u223c5\u00a0nm thick and 100\u00a0nm long, which were sintered to form a highly porous framework of randomly oriented fibers.\n40\n The fibers possessed a large specific surface area of 260 m2\u00b7g\u22121, as seen in Table S2 and Figure\u00a0S2, where most of the surface area was available for grafting of the silane possessing an NH2 group to form Ni2+ complexes and immobilizing plasmonic NPs. Moreover, the cage-like structure (see Figure\u00a01) could confine the NPs formed within the structure and allow reactant molecules to readily diffuse to the Ni2+ complex reaction sites in the vicinity of the metal NPs through inter-fiber voids.The photocatalytic reaction was conducted in a light reaction chamber. A 10\u00a0mL Pyrex glass tube was used as the reaction container. After adding the reactants and catalyst, the tube was filled with argon and sealed with a rubber septum cap. Then the tube was placed above a magnetic stirrer with stirring, and illuminated under a halogen lamp (Philips Industries: 500W, wavelength in the range of 400\u2013750\u00a0nm). An air conditioner was set to the light reaction chamber to control the reaction temperature. Reactions were also conducted in the dark at the same temperature for comparison. The reaction temperature in the dark was maintained the same as the reaction under irradiation. All the reactions in the dark were conducted using an oil bath placed on a magnetic stirrer. The tube was wrapped with aluminum foil to avoid exposure of the reaction to light. After the reaction, the mixture was collected and filtered through a Millipore filter (pore size 0.45\u00a0\u03bcm) to remove the solid photocatalyst. The products were analyzed by an Agilent 6890 gas chromatography (GC) with HP-5 column. An Agilent HP5973 mass spectrometer was used to identify the product. All the products concentrations were calibrated with an external standard method.XRD patterns of the samples were recorded on a Philips PANalytical X\u2019Pert PRO diffractometer using CuK\u03b1 radiation (\u03bb=1.5418\u00a0\u00c5). The working power was 40 kV and 40 mA. The diffraction data were collected from 10\u00b0 to 80\u00b0 with a resolution of 0.01\u00b0 (2\u03b8). FT-IR measurements were conducted on PerkinElmer Spectrum.\n2\n The samples were prepared in KBr pellets and stabilized under controlled relative humidity before acquiring the spectrum. Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectra were collected by a Nicolet-5700 spectrometer in the wavenumber range between 4,000 and 620\u00a0cm\u22121 at resolution 4\u00a0cm\u22121. The particle size and morphology of the catalyst samples was characterized with a JEOL2100 transmission electron microscope, equipped with a Gatan Orius SC1000 CCD camera. Nitrogen physisorption isotherms were measured at \u2212196\u00b0C on the Tristar II 3020. Prior to each measurement, the sample was degassed at 120\u00b0C for 24\u00a0h under vacuum. The specific surface areas of the samples were calculated by using the Brauner-Emmet-Teller (BET) method and the nitrogen adsorption data in a relative pressure (P/Po) range between 0.05 and 0.2. Varian Cary 5000 spectrometer was used to collect the data for the diffuse reflectance UV-visible (DR-UV-vis) spectra of the samples. EPR spectra were recorded with a Burker EPR ELEXSYS 500 spectrometer operating at a frequency of 9.5 GHz in the X-band mode. Metal content in catalysts were obtained by inductively coupled plasma optical emission spectrometry (ICP-OES) carried out on a R4 PerkinElmer ICP-OES 8300DV instrument. Prior to analysis, the powder samples were dissolved in HNO3 (70%) and diluted by deionized water. Electrochemical measurements were conducted with Bio-Logic SAS potentiostat model VSP. All Raman spectra were taken with a 532\u00a0nm excitation laser, 10\u00a0s exposure, one-time accumulation, and a Raman imaging setup based on a Renishaw Invia Raman microscope was used.We acknowledge financial support from the Australian Research Council (DP150102110). The electron microscopy work was performed through a user project supported by the Central Analytical Research Facility (CARF), Queensland University of Technology.P.F.H. performed all the experiments. H.Y.Z. and S.S. supervised the project. E.W. and Q.X. provided valuable discussion and revised the manuscript. D.G. provided simulation of EM field. The manuscript was written through contributions of all authors.The authors declare no competing interests.Supplemental Information can be found online at https://doi.org/10.1016/j.chempr.2019.07.022.\n\n\nDocument S1. Figures S1\u2013S16 and Tables S1\u2013S6\n\n\n\n\n\nDocument S2. Article plus Supplemental Information\n\n\n\n", "descript": "\n                  Plasmonic catalysis has drawn significant interest recently, as the catalysis can be driven by visible light. Here, we show a new tactic to apply low-flux visible-light irradiation on plasmonic metal nanoparticles (NPs) to initiate catalysis with surface-bound transition-metal complexes under mild conditions. Ni2+ complexes (as catalytic reaction sites) and Au or Ag NPs were immobilized on \u03b3-Al2O3 nanofibers to produce plasmonic-antenna-promoted catalysts. The light irradiation on Au or Ag NPs enhanced photocatalytic activity of the Ni2+ complexes for reductive cleavage of C\u2013O bond by 18-fold or 17-fold, respectively. The intense electromagnetic near-fields of the plasmonic metal NPs significantly increased the chemisorption of the reactant to the Ni2+ active sites. The light-excited hot electrons transfer via a molecular bridge of the aromatic ring of the reactants. The light-enhanced chemisorption plays a key role in this photocatalyst\u2019s structure that comprises a plasmonic antenna and catalytically active metal complex sites.\n               "}
{"full_text": "Levulinic acid (LA) is regarded as a representing chemical derived from lignocellulosic biomass[1\u20133]. GVL has been identified as a platform molecule and a promising option for producing value-added chemicals, which can be efficiently obtained through the hydrogenation of LA[4\u20139]. Besides, GVL can also be used as solvents and fuel additives due to their excellent thermal stability and energy storage ability[10,11].The catalytic reduction of LA to GVL has been reported with various supported metal and bimetal catalysts[12\u201318]. Ru metal-based catalysts were especially intrinsically in this process owning to their high activity to activate carbonyl groups[8,12,19,20]. Yan et al[21] reported 96% of the selectivity of GVL with 63% LA conversion at 160\u00a0\u00b0C, 4.5\u00a0MPa H2 for 6\u00a0h. Almeida et al [22]developed a TiO2-SiO2 hybrid support loaded Ru catalyst. High LA conversion (96%) can be obtained at 130\u00a0\u00b0C, 3.0 Mpa H2 for 6\u00a0h with 5.0 wt% Ru-TS catalyst. In recent reports, a second metal was used in this reaction to reduce the use of noble metal and improve the reaction activity. A 0.5 wt % Ru-5 wt % Ni/MMT bimetal catalyst was developed and 93% conversion of LA can be achieved at 220\u00a0\u00b0C for 4\u00a0h with the pressure of 2.5 Mpa[23]. For the Ru-Sn/C catalyst, the Sn addition can lead to the formation of bimetal Rn-Sn species and stabilized the catalyst. Complete LA conversion at 180\u00a0\u00b0C with 3.5 Mpa H2 was achieved[24].Zeolites have been applied as a promising catalyst for the conversion of biomass owing to their high thermal stability, complex porous structure, and rich acid sites[25]. Benefiting from these unique structures, it could be an interesting attempt to fabricate the acid-metal bifunctional catalysts. However, diffusional restriction lowers the catalytic performance in both the solvent and gaseous-based liquid phase reaction for biomass upgrading. In this work, we consider using an MCM-49 zeolite as a carrier to prepare highly efficient bimetallic zeolite catalysts. The active center of MCM-49 zeolite was claimed to be located at the 12 MR cups (0.71\u00d70.71\u00d70.91\u00a0nm) on the external surface of zeolites[26]. Thus, the more easily accessed sites make this zeolite potential support for biomass conversion which often involves diffusion problems.Herein, Ru-Mn bimetal supported on the MCM-49 zeolite catalyst was designed, converting LA to GVL. It is shown that the Mn promoted Ru/MCM-49 catalysts are efficient catalysts for the hydrogenation of LA into GVL. A large amount of accessible acid sites in layered MCM-49 zeolites is favorable for catalyzing the cleavage of C-O bonds, a key step in the hydrogenation of LA to GVL. The effect of Mn addition on the yield of GVL was investigated. This work contributes to an in-depth understanding of the complicated structure-performance relationship of bimetallic Ru-based zeolite catalysts for the LA hydrogenation reaction.MCM-49 zeolite was obtained by SINOPEC (RIPP). Levulinic acid, \u03b3-valerolactone, ruthenium trichloride hydrate (RuCl3\u00b7xH2O, 99.98%), and Mn(CH3COO)2\u00b74H2O (\u226599%) were obtained from Sigma-Aldrich. Ru-Mn/MCM-49 catalysts with different weight percent of Mn were prepared by the two-step incipient wetness impregnation. The MCM-49 zeolite was dried at 120\u00a0\u00b0C for 6\u00a0h before impregnation. In a typical procedure, Mn(CH3COO)2\u00b74H2O was dissolved in 4\u00a0mL of water, followed by the addition of a calculated amount of MCM-49 support with stirring. The precursor was dried at 60\u00a0\u00b0C overnight and then calcined at 550\u00a0\u00b0C for 3\u00a0h to obtain the Mn/MCM-49 samples. After that, Ru (2.0 wt%) was loaded onto the Mn/MCM-49 catalyst by incipient wetness impregnation. After impregnation and aged overnight, samples were dried at 80\u00a0\u00b0C and were further calcined at 400\u00a0\u00b0C for 2\u00a0h. The obtained samples were denoted as Ru-MCM-49, Ru-Mn(x)/MCM-49, where\u00a0\u00d7\u00a0represents different weight percent (0.2, 0.5, 0.7 and 1.0 wt%) of Mn.X-ray diffraction patterns were measured on a Rigaku X-ray diffractometer with Cu K\u03b1 radiation and a tube current of 35\u00a0mA. FTIR spectra of zeolites were recorded on the Perkin Elmer's System 2000 IR spectrometer (4000\u2013400\u00a0cm\u22121). Element contents were obtained by ICP-AES on an ICP 6000 SERIES instrument.The N2 adsorption-desorption isotherms were recorded at 77\u00a0K using a Micromeritics ASAP 2004 surface area analyzer. The microporous volume, surface area, and pore size distribution were calculated by the DFT method. SEM analysis was performed on a Hitachi 4800 with an accelerating voltage of 20\u00a0kV. TEM analysis was carried out on a JEOL JEM-2100FMII microscope and operated at an accelerating voltage of 200\u00a0kV. XPS measurements were carried out on Phi510 X-ray photoelectron spectrometer with Mg K\u03b1 radiation operating at 250\u00a0W.The thermogravimetric study of the samples was carried out on the German NETZSCH STA 449 F3 thermal analyzer, with a heating rate of 10\u00a0\u00b0C/min (air flow: 50\u00a0mL/min). NH3-TPD analysis was performed on a Micromeritics-2920 instrument. The sample was treated at 300\u00a0\u00b0C for 2\u00a0h in He gas. Then, the sample was purged with NH3 in He (30%) at 60\u00a0\u00b0C for 1\u00a0h. TPD was measured in the range of 100\u2013750\u00a0\u00b0C. Pyridine-adsorbed FTIR spectrum was acquired on a Nicolet Magna-IR 560. A self-supporting wafer with 30\u00a0mg of zeolite was loaded on an in-site cell equipped with a CaF2 window. Before pyridine adsorption, the powder zeolite was evacuated at 400\u00a0\u00b0C under vacuum (P\u00a0<\u00a010\u22123 Pa) for 3\u00a0h. After equilibration, the sample was evacuated at 200\u00a0\u00b0C and 350\u00a0\u00b0C, and data were recorded accordingly.The H2-TPR analysis was carried out on the Autochem-2920 instrument. For one experiment, the sample of 200\u00a0mg was pre-treated with N2 for 2\u00a0h. The signals were collected by rising the temperature from 50 to 500\u00a0\u00b0C with a rate of 5\u00a0\u00b0C/min using H2/He mixture (5% H2). The CO sorption experiment was conducted on a Bruker Vertex 70 FTIR system equipped with a Harrick diffuse reflectory accessory. The sample was reduced in-situ at 400\u00a0\u00b0C for 2\u00a0h. The background data was acquired at 160\u00a0\u00b0C after purging with He gas. CO saturation for the catalyst was conducted with CO/He mixture (5/45\u00a0mL/min) at 160\u00a0\u00b0C synchronized with the adsorption time. The CO \u201ccut off\u201d experiment was performed using He as the balanced gas (50\u00a0mL/min).The in-situ XAFS data were taken at 8-ID (ISS) in the National Synchrotron Light Source II (NSLS II), Brookhaven National Laboratory (BNL). The in-situ measurements were done using a Clausen cell where around 1\u00a0mg sample was loaded in a 1.0\u00a0mm (OD) and 0.9\u00a0mm (ID) quartz capillary.Aqueous-phase LA hydrogenation was conducted in a 50\u00a0mL batch reactor under the conditions of 160\u00a0\u00b0C, 2.5\u00a0MPa H2. The as-prepared catalysts were reduced at 10% H2/Ar flow at 400\u00a0\u00b0C for 2\u00a0h prior to the reaction. Subsequently, 10\u00a0mL LA aqueous solution (50\u00a0g/L) and 50\u00a0mg catalyst were added to the reactor, which was then purged with N2 three times. The reactor was then sealed with H2 and heated up to the desired reaction temperature. Reactions were conducted at 160\u00a0\u00b0C for 180\u00a0min under a constant stirring speed of 800\u00a0rpm. The collected liquid products were analyzed by Agilent GC-7820 with FID detector.Powder XRD patterns for the reduced catalysts are shown in Fig. 1\n. Typical diffraction peaks of MCM-49 can be observed for all the samples. No obvious Ru2O3 phase was found, indicating the highly dispersed Ru species. The crystalline structure of MCM-49 zeolite was also confirmed by the FTIR spectra (Fig. S1). There was no significant morphology change demonstrated by SEM images for the catalyst with different Mn concentrations which can be seen in Fig. S2. The N2 physisorption profiles for samples are shown in Fig. 2\na. The uptake at the P/P\n0\u00a0=\u00a00.5\u20130.9 indicates the presence of mesopores in the catalysts. Pore size distribution (Fig. 2b) calculated by DFT method suggests that all the samples present mesopores at 1.9\u00a0nm and 2.7\u00a0nm. Table 1\n gives the porous parameters of samples. Compared with the MCM-49 zeolite, both the microporous area and the external surface area decreased after metal loading. At high Mn loading (\u02c30.7 wt%), the external surface area of the catalyst significantly decreased while the microporous area was kept unchanged. It can be speculated that the micropores of zeolite can be easily blocked by metal sites located at the micropores and mesopores. In this situation, the excess amount of Mn tends to partly render the diffusion of reacting molecules in the reaction.The physical chemistry parameters of the typical catalysts are collected in Table 2\n. The Mn modified Ru catalysts exhibit nearly the same Ru loading about 2.0 wt%. It was observed that the Ru concentration at the surface of the catalyst is much smaller than that at the bulk, which suggests that Ru particles prefer to locate at the micropores of the MCM-49 zeolite. CO chemisorption was applied to estimate the diameter and dispersion state of Ru particles. It was found that the size of Ru particles decreased with the increase of Mn content. High Ru dispersion was obtained by Mn doping. Furthermore, the EDS-mapping images of Ru-Mn(0.7)/MCM-49 in Fig. 3\n further revealed the highly dispersed RuO nanoparticles. Consequently, the results illustrated that highly dispersed Ru species occurred in Ru-Mn(0.7)/MCM-49 through the presence of the Mn element.\nFig. 4\n gives the activities of the tested catalysts for the LA hydrogenation. In previous reports, some kinds of derivatives, such as 1,4-pentanediol, MTHF and pentanoic acid were observed in the LA hydrogenation reaction[27\u201330]. In this report, GVL was found to be the main product and targeting intermediate 4-hydroxypentanoic acid (4-HPA) was detected. No other products such as methyl-tetrahydrofuran and pentanoic acid formed by hydrogenation products of GVL were observed. When Ru/MCM-49 was used as the catalyst, the LA conversion was relatively low (64%) with the GVL selectivity of 87%. Mn modified samples exhibited enhanced hydrogenation activity. However, LA conversion and GVL selectivity were substantially boosted over Ru-Mn/MCM-49 catalysts as evidenced by Fig. 4. With increasing the Mn content, the LA conversion continuedly increased. When 0.7 wt% Mn was loaded onto Ru/MCM-49 catalyst, the LA conversion was improved significantly to 98%, and high selectivity (100%) to GVL. The high activity of the Ru-Mn/MCM-49 catalyst in this reaction suggests that the catalytic activity is determined by the Mn species. The comparison results in Table 3\n from the literature demonstrate the high performance of the Ru-Mn/MCM-49 catalyst for the LA hydrogenation in this study.As shown in Fig. S5, the yield of GVL decreased from 98% in the first run to 63% in the fourth run at the reusability test. However, this stability was much better than the Ru/C catalyst in the literature[31,32]. Further, the TGA results (Fig. S6) showed that the carbon deposit was the important reason for the loss of activity. The mass loss between 200\u00a0\u00b0C and 400\u00a0\u00b0C was the adsorption of small amounts of residual organic species or carbon deposits of surface metal potentials on the catalyst. In the TGA profile, a significant weight loss in the range of 200\u00a0\u00b0C\u00a0\u2013\u00a0400\u00a0\u00b0C can be seen, indicating that carbon deposition mainly occurs on the surface of the catalyst. The carbon deposits covered the active metal sites on the surface of the catalyst so that the activity of the reaction was reduced.The acid sites of the catalysts were evaluated by NH3-TPD and Py-FTIR studies. NH3-TPD plots of samples, shown in Fig. S3, exhibit that all the catalysts present similar acid amounts and acid strength. The desorption peaks at the temperature range (220\u2013330\u00a0\u00b0C) are ascribed to the weak acid sites while NH3 desorption peaks at higher temperatures (380\u2013460\u00a0\u00b0C) are assigned to strong acid sites. For the impregnated zeolites catalysts, the acid properties of the catalyst were mainly determined by supports[33]. In some reports, a slight change in the acid sites can be observed for the Ru exchanged zeolites[34,35]. In this study, due to the low Ru metal loading, MCM-49 zeolite contributed to the main acid sites, including the weak and strong acid sites, and all the samples showed similar amounts of acid sites. The Py-FTIR spectra of all samples were carried out to estimate the amounts of Br\u00f8nsted acid sites and Lewis acid sites. As shown in Fig. S4, the band at 1540\u00a0cm\u22121 from the bonding of Br\u00f8nsted acid sites appeared on the sample Ru/MCM-49. The quantitive results are listed in Table 4\n. For the various Ru-Mn/MCM-49 catalysts, the amounts of the Br\u00f8nsted acid sites of catalysts decreased with the Mn promotion. The band at 1445\u00a0cm\u22121 and 1605\u00a0cm\u22121 can be assigned to Lewis acid sites. The L/B ratio was also calculated on the basis of the two kinds of acid sites. It can be found that the amount of Lewis acid sites is rather stable for all the catalysts, but the L/B ratio increases significantly with Mn modification. To identify the intrinsic activity of the catalyst, the turnover frequency (TOF) was calculated and was listed in Table 5\n. The Ru-Mn(0.7)/MCM-49 catalyst exhibited the highest TOF value of 1529\u00a0h\u22121, which was 1.6 times higher than that of the Ru/MCM-49 catalyst.The reaction pathway for the LA hydrogenation was shown in scheme 1\n. The hydrogenation of levulinic acid into 4-hydroxypentanoic acid (4-HPA) followed by the dehydration of 4-HPA into GVL are the two main steps of the reaction. The first step of the reaction is the hydrogenation reaction which requires metal sites and for the dehydration reaction (second step) the active component is acid sites[14,36\u201339]. The relatively high TOF among the test catalysts indicated that a high L/B ratio clearly had a positive effect on the hydrogenation reaction. Because the Lewis sites preferentially interact with the CO bonds of LA, the L/B ratio is closely related to this reaction. However, an increase in Br\u00f8nsted acid acidity (Ru/MCM-49) led to the significant generation of the intermediate product 4-HPA, with a concomitant decrease in the GVL selectivity as shown in Table 5. As a result, the Lewis acid sites originated from the zeolite, enabling the activation of LA and the followed dehydration reaction, improving the efficiency of this reaction.In Fig. 5\n, H2-TPR profiles are present for the catalysts. Main reduction peaks at 110\u2013120\u00a0\u00b0C were observed for all the catalysts. The reduction peak lower than 150\u00a0\u00b0C can be ascribed to the reduction of highly dispersed Ru particles located at the external surface of zeolite[40]. Obviously, the Ru existed as RuOx in the MCM-49 zeolite support, which can be reduced at relatively low temperatures. The value of the reduction temperature for the ruthenium species was aligned with the results reported for Ru/zeolite and Ru/SiO2\n[41,42]. The Mn loading can efficiently improve the reduction temperatures of the catalysts. The highest reduction temperature was obtained for the sample Ru-Mn(0.7)/MCM-49, indicating the presence of strong metal-support interaction in the sample.To further elucidate the dynamic changes in the local coordination environment, in-situ XAFS characterizations were performed to investigate the chemical state and local structure of Ru in the reduced catalysts. As should in Fig. 6\n(a), both reduced samples and Ru foil show very similar XANES structures, indicating the samples have been fully reduced to metallic Ru. Moreover, the Fig. 6(b) EXAFS data give a similar conclusion that both reduced samples have average local structures very similar to metallic Ru. These results demonstrate that the phase of Ru can be fully reduced as observed in the XANES results.XPS characterizations were performed to investigate the chemical state of the catalysts. Fig. 7\n shows the major C, O, Mn, Ru, Si and Al peaks, the Ru 3p spectra, and the Ru 3d spectra. Ru 3d (Fig. 7b) was not suitable to determine the chemical state of Ru species which was easily obscured by carbon C 1\u00a0s peak[43\u201345]. As shown in Fig. 7c, the binding energy of 461.8\u00a0eV designed at Ru 3p3/2 evidenced the metallic state of Ru(0) for Ru-Mn/MCM-49. After doping Mn, the binding energy of Ru 3p3/2 shifted to a low value, and it becomes 461.3\u00a0eV for Ru-Mn(1.0)/MCM-49 catalyst. The change of peak value indicates that electronic interaction excites between Ru and Mn species. With the increase of Mn content, the enhancement of electron density of metallic Ru can be inferred because lower binding energy of Ru 3p was observed. The shift of binding energy for Ru species suggests the catalysts exhibit higher H2 dissociation ability after loading Mn[46].\nIn-situ CO-Drifts were further employed to reflect the sensitives of environment change and the electronic state of the adsorption site by the CO vibration frequency. The spectra of CO adsorption on the Ru/MCM-49 and Ru-Mn(0.7)/MCM-49 catalysts were obtained in Fig. 8\n. Low adsorption band at 2050\u00a0cm\u22121, high adsorption bands around 2115\u00a0cm\u22121, and high adsorption bands around 2171\u00a0cm\u22121 are observed. The band at 2171\u00a0cm\u22121 can be designed as the CO molecular in the gas phase rather than the multicarbonyl species on Run+ as reported in previous report[47], which is also demonstrated by the CO\u2013 cut-off experiment as shown in Fig. S5. The band at 2056\u00a0cm\u22121 was assigned to the Ru particles. The peak at higher wavenumber (2115\u00a0cm\u22121) falls in the range of high dispersed Ru metal sites. In this work, two different kinds of dispersed Ru sites were confirmed. It is clear that the addition of Mn changes the state of metallic sites on the Ru-Mn/MCM-49 catalyst. The relative percent of Ru particles in Ru/MCM-49 was higher than that in Ru-Mn/MCM-49, which was in agrees with the H2-TPR results. Then a high GVL selectivity was observed for the sample Ru-Mn/MCM-49. It has been reported by Novod\u00e1rszki et al[48] that the LA conversion was closely related to the adsorption state of the surface intermediate and the GVL selectivity can be enhanced by the efficient intermediate adsorption and the high surface H coverage. However, too strong H coverage may decrease the GVL selectivity by further hydrogenation. In the LA hydrogenation reaction, hydrogen was adsorbed on the surface of Ru through the formation of hydrogen bonds between hydrogen and Ru. Hence, an increase of the high dispersed Ru species was beneficial to the dissociation and activation of H2 on the catalyst surface; thus, the reaction rates for the hydrogenation of LA can be enhanced. Hence, the results of in-situ CO-DRIFTS confirm that Ru species exists as high dispersed metallic species in the sample Ru-Mn/MCM-49, in agreement with the above H2-TPR and EXAFs observations.Bifunctional Ru/MCM-49 catalyst modified by Mn species was developed for the hydrogenation of LA to GVL. XPS result revealed the presence of Mn-Ru interaction with the Mn modification. The shift of position for metallic Ru 3p to lower energy suggested the catalyst exhibited higher H2 dissociation ability by Mn addition. It was also found that the amount of Br\u00f8nsted acid sites for the sample Ru-Mn/MCM-49 was significantly decreased, which could be beneficial for the formation and stabilization of GVL. Doping of Mn to Ru-MCM-49 remarkedly improved the activity, and the activity of the catalyst was related closely to the Mn concentration. Ru-Mn(0.7)/MCM-49 catalyst exhibited the highest activity, with the TOF value of 1529\u00a0h\u22121. Under the reaction conditions, 98% LA conversion and 100% GVL selectivity were achieved. The electron density of metallic Ru can be enhanced by the addition of Mn which improves the dissociation ability of Ru sites in the reaction as revealed by XPS and CO-DRIFTS. The stable adsorption and activation of LA and H2 can efficiently proceed on the metallic Ru sites facilitating the hydrogenation reaction.\nWenlin Li: Conceptualization, Funding acquisition, Supervision. Feng Li: Investigation. Xin Ning: Investigation. Kaixi Deng: Investigation, Methodology. Junwen Chen: Investigation, Supervision. Jiajun Zheng: Data curation. Ruifeng Li: Resources.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Financial support from the State Key Laboratory of Catalytic Materials and Reaction Engineering (RIPP, SINOPEC, No. 18-ZC0607-0007) is gratefully acknowledged. This research used resources 8-ID (ISS) of the National Synchrotron Light Source II, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Brookhaven National Laboratory under Contract No. DE-SC0012704.Supplementary data to this article can be found online at https://doi.org/10.1016/j.crcon.2022.05.003.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  Selective hydrogenation of Levulinic acid (LA) to \u03b3-Valerolactone (GVL) is an important reaction to produce high value-added chemicals and fuels but remains a big challenge. Herein we reported a Ru/zeolite catalyst with Mn promotion, which exhibited excellent catalytic performance (yield: 98%) towards LA to GVL. The intrinsic activity (TOF) also increased obviously with the Mn addition. The particle size of Ru gradually decreased with the increase of Mn loading and a strong interaction between Ru and support was observed for the Ru-Mn/MCM-49 catalyst. The addition of Mn not only offered a good dispersion of Ru species on MCM-49, but also increased the L/B ratio of the catalyst, thereby contributing to the high GVL selectivity. High dispersed Ru sites were the intrinsic active sites of the catalyst verified by the in-situ experimental studies. The dissociation of the reactants was significantly enhanced, resulting in higher catalytic activity.\n               "}
{"full_text": "H2 is considered an ideal clean energy carrier to replace traditional fossil fuels.\n1\u20134\n Electrocatalytic H2 evolution reaction (HER) from renewable energy by water splitting is a promising technology for efficient H2 production.\n5\u20138\n Various Earth-abundant metal-based catalysts with considerable HER activities have been broadly reported to replace precious platinum group metal (PGM) electrocatalysts.\n9\u201314\n Unfortunately, a large gap still exists between non-precious metals and PGM catalysts toward HER, owing to the lack of efficient active sites. The occurrence of too weak (e.g., for Cu and Zn) or too strong (e.g., for Co, Ni, and W) H adsorption results in a lower reaction rate through affecting the initial H adsorption and/or the ultimate molecular H2 desorption from the catalyst surface.\n15\n In particular, metallic Ni and its alloys are widely regarded as the most promising non-precious-metal electrocatalysts alternative to PGM for industrial-scale water electrolysis; yet, the strong H adsorption on Ni has severely affected its durability.\n16\n\nThe Fe triad elements (i.e., Fe, Co, and Ni) share similar chemical and physical characteristics, are abundant on Earth, and have been popularly used as catalysts for many hydrogenation applications.\n17\n\n,\n\n18\n For example, metallic Co is the most widely used catalyst in Fischer-Tropsch synthesis (FTS) for converting H2 and carbon monoxide from coal or natural gas into long-chain hydrocarbons.\n19\n\n,\n\n20\n However, metallic Co and Fe typically show poor activity toward HER, although Co- and Fe-based chalcogenides,\n21\n nitrides,\n22\n carbides,\n23\n and phosphides,\n24\n and their interaction with heteroatom-doped carbon and macrocyclic Co complexes\n25\n\n,\n\n26\n have been reported as having considerable activity for HER. Co can exist in two crystallographic structures: the hexagonal close packed (hcp) and the face-centered cubic (fcc) phase. Many reports have revealed that the hcp Co is more active than fcc Co, due to the intrinsic stacking faults that offer a greater number of active sites.\n27\n\n,\n\n28\n Nevertheless, the large size of nanoparticles (NPs) with limited surface area and strong H binding over Co greatly slow down the desorption of H to evolve H2 and limit the HER kinetics. To achieve enhanced HER catalytic performances on Co, downsizing the catalysts to fine NPs and creating disordered structure to engender more exposed active sites, along with optimized electronic structures for H binding, are critical.Here, we show a hcp metallic Co catalyst modulated by vanadium oxide (VOx) clusters (denoted as Co(VOx)) with extraordinary HER activity as well as an atomic-level understanding of the activity origins. The VOx-modulated metallic Co catalyst ((Co(VOx)-y, y is the V5+/Co2+ precursor molar ratios) is prepared by a facile electrodeposition process. Due to the strong interaction between Co and VOx, the crystallographic, electronic structure, and coordination environment can be rationally regulated at atomic scale by adjusting the doping levels during the electrodeposition process. The VOx doping plays an important role for achieving a refined Co NPs and highly disordered Co structure. The optimal Co(VOx)-3% catalyst exhibits extraordinarily low overpotentials and high activity and stability toward HER, which are distinctly different from the poor HER activity and durability of metallic Co. X-ray absorption spectroscopy (XAS), X-ray crystallography, operando Raman spectroscopy, and density functional theory (DFT) calculations are applied to study the roles of VOx clusters, uncovering the highly disordered structures and partial electron transfer from Co to VOx, which dramatically decreases the H adsorption on V-Co(001) to achieve enhanced HER.The Co(VOx)-y (y\u00a0= 0%, 1%, 3%, and 6%) catalysts with different doping levels are electrodeposited directly on Ni foam (NF) substrate, and all of the electrodes after electrodeposition show a black color. The X-ray diffraction (XRD) pattern of Co and Co(VOx) samples electrodeposited on carbon fiber paper to remove the effect of diffraction signals from NF were collected by using the Empyrean PANalytical diffractometer in the grazing incidence measurement mode. As shown in Figure\u00a01\nD, for pure Co, the characteristic peaks located at 41.5\u00b0, 44.5\u00b0, and 47.3\u00b0 are indexed to the (100), (001), and (101) planes of hcp metallic Co (joint committee on powder diffraction standards [JCPDS]: 05-0727),\n20\n respectively. VOx doping has a remarkable influence on the crystal structure of Co. With the increment of VOx doping, the two characteristic peaks at 41.5\u00b0 and 47.3\u00b0 almost disappear, while the (001) peak broadens and decreases sharply, indicating the formation of a poorly crystallized Co structure and refined NPs, as illustrated in Figure\u00a01A.To identify the structural transformation of Co by VOx doping, the electrodeposition behavior of the Co(VOx)-y electrodes was analyzed. The representative Co(VOx)-3% electrode is discussed in the following sections, unless stated otherwise, as it shows the highest HER activity. Figure\u00a0S1 shows that pure Co was electrodeposited at a more positive potential than that for pure VOx deposition from the NH4VO3 precursor. However, the deposition of VOx can significantly influence the Co crystallization, as evidenced by a negative shift of the electrodeposition potential for Co(VOx)-3%. This negative shift in potential indicates that VOx regulates the nucleation and crystallization process of metallic Co. Because the formed VOx are ultrafine NPs in nature (see scanning electron microscopy [SEM] information below), thus, the crystallization structure and size of Co can be well regulated by VOx.The morphology of the obtained catalysts was first characterized by SEM. Figure\u00a0S2 shows the SEM image of uniformly and densely distributed Co NPs on NF. At high magnifications, it is revealed that the size of NPs is in the scale of tens of nanometers. For the electrodeposited VOx, the NF substrate is uniformly covered with the composite constituted by a large number of NPs with sizes of merely several nanometers (Figure\u00a0S3). Figure\u00a0S4 depicts the different magnified SEM morphology for Co(VOx)-3% with respect to the individually deposited Co or VOx. In the presence of vanadium, the well-ordered structure of Co NPs was transformed into an irregular nanocluster structure composed of fine NPs with a size of few nanometers, indicating the crystalline refinement effect of VOx on Co. Transmission electron microscopy (TEM) was used to learn more about the structure of Co(VOx) composites. As seen in Figures 1B and S5, the TEM morphology of pure Co shows large NPs with sizes ranging \u223c30\u201350\u00a0nm. The clear lattice patterns and diffraction rings from the selected area electron diffraction (SAED; inset in Figure\u00a01B) both indicate the well-crystallized Co structure. The TEM images of the Co(VOx)-3% sample reveal the uniformly distributed fine NPs structure with a size of \u223c5\u201310\u00a0nm (Figure\u00a01C). The high-resolution TEM (HRTEM) image in Figure\u00a01C shows a lattice fringe of 0.205\u00a0nm, which is consistent with the (001) plane of Co and corresponds well with the SAED pattern (Figure\u00a01C, inset). However, the poorly defined rings in SAED demonstrate the poor crystalline phases and a disordered Co structure. The elemental distribution was confirmed by TEM-energy dispersive spectroscopy (EDS) mapping, in which all the elements are uniformly distributed throughout the whole sample (Figures 1E\u20131H). In addition, the HRTEM morphologies of Co(VOx)-1% and Co(VOx)-6% samples (Figures S6 and S7) depict Co NPs with sizes between \u223c15\u201320\u00a0nm and \u223c5\u201310\u00a0nm with good and poor lattice fringe for Co (001) plane, respectively, demonstrating that both the morphology and crystal structure of Co NPs can be well tuned through VOx modulation.X-ray photoelectron spectroscopy (XPS) was used to investigate the valence states of the as-prepared Co and Co(VOx) samples. As seen from Figure\u00a0S8, in the Co2p region, the two main peaks for Co2p3/2 and Co2p1/2 are located at 781.2 eV and 797.3 eV, accompanied by two shake-up satellite peaks (786.6 eV and 803.1 eV), demonstrating an oxidized Co surface. In addition, an enhanced oxidation degree of Co after the increment of the doped VOx causes a slightly negative energy shift in Co2p configuration.\n29\n The high-resolution XPS of the V2p region shows characteristic peaks at 517.1 eV and 516.4 eV assigned to V5+ and V4+ species,\n30\n and a slightly negative shift of V2p was observed with increasing VOx doping. The energy shift on Co2p and V2p reveals a direct interaction between Co and VOx and a charge transfer from Co to VOx.\n31\n The O1s XPS spectra show a main peak at \u223c531.1 eV, corresponding to O in VOx and the oxidized Co-O in the composite. It is noted that the metallic Co peaks were not observed in the XPS spectra because of the surface oxidized layers.The HER performance of the as-prepared Co(VOx) catalysts with different VOx concentrations (0%\u20136%) along with the commercial 20 wt % Pt on carbon black (Pt/C) electrodes are evaluated in 1\u00a0M KOH without ohmic potential drop (iR) correction (Figure\u00a02\nA). Linear sweep voltammetry (LSV) of all the prepared Co(VOx) electrodes show significantly enhanced HER activity compared to the bare Co electrode. In particular, the Co(VOx)-3% electrode shows the best HER activity, suggesting the electrodeposited VOx with optimal doping plays an important role in the activity enhancement. Specifically, to deliver a current density of \u2212100 mA cm\u2212\n2, the Co(VOx)-3% electrode requires an overpotential of merely 178\u00a0mV, which is much smaller than 344\u00a0mV and 275\u00a0mV required for Co and Co(VOx)-1%, respectively, and 203\u00a0mV for Co(VOx)-6%. The obtained performance of Co(VOx)-3% is comparable or outperforms recently reported state-of-the-art noble-metal-free HER catalysts, although it still shows a larger overpotential than Pt/C catalyst (93\u00a0mV) (Table S1). The corresponding Tafel slopes were derived from the LSV curves to investigate the reaction kinetics. As shown in Figure\u00a02B, the Co(VOx)-3% and Co(VOx)-6% electrodes have similar Tafel slopes of 40 and 36\u00a0mV dec\u22121, respectively, which are much smaller than those of Co(VOx)-1% (60\u00a0mV dec\u22121) and Co (125\u00a0mV dec\u22121), suggesting that HER follows the Volmer-Heyrovsky mechanism at the Co(VOx) electrodes. Also, a Tafel slope of 23\u00a0mV dec\u22121 was obtained for the Pt/C electrode. The significantly decreased Tafel slope of Co(VOx)-3% and Co(VOx)-6% indicate fast intrinsic kinetics of HER due to the formation of the fine and disordered Co structure (see below).The long-term electrocatalytic durability is a pivotal parameter for a HER electrode. Figure\u00a02C shows that the Co(VOx)-3% catalyst retains a stable HER activity for over 50 h. In contrast, the bare Co electrode shows a noticeable activity decay during the continuous HER operation. Additionally, the almost overlapped LSV curves obtained on the Co(VOx)-3% electrode before and after stability demonstrate the robustness of the Co(VOx)-3% electrode (Figure\u00a02D). The stable activity of the catalyst was further studied by the post-HER characterizations. As shown in Figure\u00a0S9, the almost unchanged morphology of the fine NPs and the retained uniform distribution of the constituting elementals both indicate the rigid structure of the Co(VOx)-3% catalyst, and the HRTEM also reveals the existence of the lattice fringe for metallic Co.To understand the origin of the enhanced HER activity on Co(VOx) catalysts, the electrochemical active surface area (ECSA) was first determined by measuring the double-layer capacitance (C\ndl) of the electrodes derived from cyclic voltammetry (CV) curves in a non-Faraday region with different sweep rates (Figure\u00a0S10).\n32\n The C\ndl values obtained on Co(VOx) electrodes are all larger than that on pure Co (12 mF), with the Co(VOx)-3% electrode being the largest (30 mF), followed by Co(VOx)-6% (24 mF) and Co(VOx)-1% (17 mF) (Figure\u00a02E). These results demonstrate that the fine NPs and disordered structure provide more active sites for HER. Moreover, the intrinsic catalytic activity of each active site was evaluated by normalizing current against ECSA. Figure\u00a0S11 shows that the Co(VOx)-3% electrode still exhibits a higher HER activity than Co and Co(VOx)-1%, demonstrating the profound role of VOx in enhancing the intrinsic activity of each active site. It should be noted that the ECSA normalized HER activity of Co(VOx)-3% and Co(VOx)-6% electrodes almost overlapped, suggesting that the more exposed active sites are contributing in the improved HER on Co(VOx)-3%, and this observation is consistent with the projected DFT simulations (see below). In addition, the charge transfer process of the prepared samples was investigated by electrochemical impedance spectroscopy (EIS), and an equivalent resistor-capacitor circuit model (R\ns, resistor; R\nct, charge transfer resistance; C, capacitance) was used to fit the impedance spectra. As shown in Figure\u00a02F, the EIS spectra reveal a significantly smaller R\nct for the Co(VOx)-3% electrode (12\u00a0\u03a9) than that of Co(VOx)-6% (22\u00a0\u03a9), Co(VOx)-1% (81\u00a0\u03a9), and Co (107\u00a0\u03a9), verifying the fast electron transfer kinetics of HER on Co(VOx)-3%. Furthermore, the Faradaic efficiency close to unity is obtained on the Co(VOx)-3% electrode by measuring the generated H2 gas by using gas chromatography (Figure\u00a0S12).To obtain further information on the elemental oxidation states and in particular to determine the influence of VOx on the atomic structure of Co and eventually the HER performance, the Co K-edge X-ray absorption near edge structure (XANES) was studied in detail. As seen from Figure\u00a03\nA, the XANES spectra of Co and Co(VOx)-1% contain similar pre-edge characteristic features, referring to an oxidation state close to metallic Co for Co in these catalysts. With increasing the VOx amount, both the positively shifted pre-edge peak and higher white line intensity demonstrate a higher oxidation degree of Co (Figure\u00a03B). In other words, the Co(VOx)-3% and Co(VOx)-6% catalysts are oxidized and disordered more at higher VOx doping levels. Figure\u00a03C shows the resulting Fourier-transformed extended X-ray absorption fine structure (FT-EXAFS) for all prepared samples. There are fewer scattering contributions from the Co-O bond for Co and Co(VOx)-1%, but the other three nearest-neighbor (NN; 2nd\u20134th) shells show strong intensity, which indicates the formation of metallic Co with highly ordered atomic structure.\n33\n It should be noted that from XPS spectra, we observed that these two samples are oxidized, but the XAS spectra reveal a metallic phase structure. This is because XPS is a surface-sensitive technique, whereas XAS is designed for bulk-averaged atom-atom correlation information. With the increment in VOx doping, the significantly reduced intensity amplitude for Co(VOx)-3% and Co(VOx)-6% demonstrates the existence of an abundant disordered structure. Furthermore, the increased 1st NN shell intensity reveals the higher oxidation degree, whereas the decreased 2nd\u20134th NN shell intensity indicates a much less efficient atomic packing and a greater structural disorder of metallic Co due to the regulation effect of VOx, which corresponds well to the XRD and XPS results.\nEx situ experiments yield valuable information on the chemical nature of Co(VOx); yet, the origin of catalytic activity in these NPs remains largely unknown. To address this issue, operando Raman spectroscopic experiments were performed to detect structural changes during HER conditions. To conduct operando experiments, the catalyst was held at the applying potential for 10\u00a0min before acquiring the spectrum on Co(VOx)-3% and Co electrodes. As shown in Figure\u00a03D, several intense peaks are observed at 518\u00a0cm\u22121, 679\u00a0cm\u22121, and 806\u00a0cm\u22121 on Co(VOx)-3% under open circuit potential (OCP). The detected bands at 518\u00a0cm\u22121 and 679\u00a0cm\u22121 are assigned to Co oxide, which correspond to F1\n2g and A1g vibrations, respectively.\n34\n The Raman peak at 806\u00a0cm\u22121 is attributed to O-V-O stretching vibrations. As the potential was increased to more negative values, e.g., \u22121.1 V, the vibration band peak at 679\u00a0cm\u22121 almost disappeared, whereas the characteristic peak at 518\u00a0cm\u22121 weakened, which is attributed to the reduction of the Co oxide layer. The intensity of the VOx peak at 806\u00a0cm\u22121 was also decreased, which is mainly due to the dissolution of loosely bonded or physically adsorbed VOx in KOH. It should be noted that the Co-O-V features cannot be totally removed, as the peak at 806\u00a0cm\u22121 remains with increasing the overpotential. This is further evidenced by the inductively coupled plasma optical emission spectroscopy (ICP-OES) analysis (V/Co\u00a0= 2.7% for fresh sample versus V/Co\u00a0= 1.4% after stability) as well as the V/Co atomic ratio of 1.2% from TEM-elemental spectra (Figure\u00a0S9F) after long-term stability test. In addition, the operando Raman spectra of the pure Co electrode shows two well-defined Raman peaks at 523\u00a0cm\u22121 and 691\u00a0cm\u22121 for Co oxide under the OCP condition (Figure\u00a03E). However, the peak intensities are much lower than those of the Co(VOx)-3% NPs, and a similar trend for the reduction of the Co oxides can be found when higher negative potentials are applied.From the above results, it can be concluded that under HER conditions, Co oxides are first reduced to expose Co-O-V active sites to electrolyte for H2 evolution. This is further proven by applying a negative potential at different LSV cycles on the Co(VOx)-3% electrode, for which a profound Co oxide reduction peak was observed during the first 3 cycles (Figure\u00a03F), demonstrating the fast reduction of Co oxide for the exposure of Co-O-V active sites toward HER. It is noted that the developed electrodeposition approach can be easily extended to other conductive substrates, such as carbon fiber paper and copper foam. As seen from Figure\u00a0S13, similar performances were achieved as those on the NF substrate, where the Co(VOx) electrode always shows a significantly enhanced HER activity compared with the electrodeposited bare Co electrode, demonstrating this is a general synthetic approach to fabricate active Co(VOx) electrodes on conductive substrates.Understanding the doping effect of the oxidized V (VO4 cluster) on the HER activity on Co(001) catalysts as the determined structure of the catalysts was followed by a series of DFT calculations to elucidate the HER on Co(001) and V-doped Co(001) (V-Co(001)) catalysts under different VO4 coverage levels (\u03b8\u00a0= 0.25, 0.50, and 0.75). The structural model of bare Co(001) was constructed as a 4\u00a0\u00d7 4 periodic supercell (Figure\u00a0S14). We considered the key reaction steps in alkaline HER, including the water dissociation reaction and the adsorption/combination of reaction H intermediates (H\u2217). Figure\u00a0S15 shows the calculated reaction energy diagram of water dissociation on Co(001) and V-Co(001) (\u03b8\u00a0= 0.25). The energy barriers for water dissociation are similar on these two surfaces (0.90 eV and 1.01 eV on Co(001) and V-Co(001) (\u03b8\u00a0= 0.25), respectively), which are quite smaller than those for other active catalysts for alkaline HER, such as Ni2P/NiTe2.\n7\n These results indicate that the water can be efficiently dissociated on both Co(001) and V-Co(001) surfaces with similar energy barriers.Then, we considered all possible adsorption sites for H\u2217 closed to the VO4 cluster (Figure\u00a0S16) and all possible stable structures of V-Co(001) with \u03b8 \u2265 0.50 (Figures S17 and S19). Figure\u00a04\nD shows the calculated free energy diagram for HER on bare Co(001), V-Co(001) (\u03b8\u00a0= 0.25-0.75), and V-Co(001) (\u03b8\u00a0= 0.50) with the highest HER activities, which are shown in Figures 4A, 4B, S18, and S20. For the bare Co(001), the \n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n is highly negative (\u22120.32 eV), which indicates a strong interaction between H\u2217 and Co(001),\n35\n manifesting poor HER reaction kinetics. Introducing a low coverage of the VO4 cluster on Co(001), namely V-Co(001) (\u03b8\u00a0= 0.25), significantly increased the value of \n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n to \u22120.25 eV, suggesting an enhanced HER activity compared to bare Co(001). The deformation electronic density calculation (Figures 4C and S21) and Bader analysis show that the VO4 cluster has led to an increase in electronic charge density on VO4 and a loss of electron charge density on the surrounding six Co atoms, and there are 0.23 electrons transferred from each Co to VO4 based on the Bader analysis. Fewer electrons localized on Co sites closest to VO4 results in the weak H adsorption on V-Co(001) and thus the enhanced HER activity. Moreover, we found that a further increase of the coverage of VO4 clusters on the Co(001) surface can further increase the values of \n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n. For example, the \n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n for V-Co(001) (\u03b8\u00a0= 0.50) and V-Co(001) (\u03b8\u00a0= 0.75) are \u22120.12 eV and 0.14 eV, respectively. These results indicate that the HER activities of V-Co(001) with a high VO4 coverage (\u03b8 \u2265 0.50) are much higher than those with low VO4 coverage (\u03b8 \u2264 0.25). In addition, considering that the number of active sites for V-Co(001) (\u03b8\u00a0= 0.75) (Figure\u00a0S20) are fewer than those of V-Co(001) (\u03b8\u00a0= 0.50) (Figure\u00a0S18), we conclude the HER activity follows the order of V-Co(001) (\u03b8\u00a0= 0.50) > V-Co(001) (\u03b8\u00a0= 0.75) > V-Co(001) (\u03b8\u00a0= 0.25) > Co(001), which is consistent with experimental results. It has to be noted that the binding of reaction intermediates on low-coordinated\u00a0metal atoms is stronger than that on high-coordinated metal sites.\n36\n The small particle size of Co can significantly increase the proportion of low-coordinated metal atoms, such as step atoms, kink atoms, and vacancies, which can increase the binding strength of H\u2217. Hence, the absolute value of adsorption energy of H\u2217 should be decreased as Co particle size increases. In other words, the downsized Co NPs should have a stronger H\u2217 adsorption and decrease HER activity. Although the downsized Co particle itself has a strong H\u2217 adsorption energy, for our developed Co(VOx) catalyst, the VOx modification can significantly decrease the H\u2217 adsorption energy compared with that of bare Co, thus leading to the excellent HER performances.In summary, a Co(VOx) catalyst was developed with a refined NP size, disordered structure, and a significantly enhanced alkaline H2 evolution activity. The best Co(VOx) electrode can only be achieved at appropriate VOx doping levels to offer suitable H binding as well as abundant active sites during HER. This study provides a promising approach for achieving notable HER performance with metallic Co, which is known to be a poor catalyst for HER, and sheds light on the crucial role of atomic structure modification through a facile VOx engineering strategy. The development of highly active metallic Co-based catalysts for HER would find applications in the water electrolysis industry and also potentially provide a multifunctional catalyst for the FTS industry for which metallic Co is currently most widely used for the synthesis of valuable long-chain hydrocarbons from H2 and carbon monoxide. The proposed approach can also be potentially extended to other non-active metallic Fe triad materials for developing efficient low-cost electrocatalysts for H2 production.Further information and requests for resources should be directed to the Lead Contact, Prof. Chuan Zhao (chuan.zhao@unsw.edu.au).This study did not generate new unique materials.The authors declare that data supporting the findings of this study are available within the article and the Supplemental Experimental Procedures. All other data are available from the lead contact upon reasonable request.The Co(VOx) electrodes were prepared by a direct electrodeposition approach in a\u00a0three-electrode system with the NF as the working electrode, graphite plate as the counter electrode, and a double junction saturated calomel electrode (SCE) as the reference electrode. The NF was cleaned in a 5\u00a0M HCl solution for 10\u00a0min and rinsed Milli-Q water before use. The electrodeposition was carried out in an electrolyte\u00a0consisting of 0.45\u00a0M CoCl2\u00b76H2O, 4.5 to 27\u00a0mM (V/Co\u00a0= 0%\u20136%) NH4VO3, and 0.35\u00a0M H3BO3, by dissolving the chemicals in 50\u00a0mL Milli-Q water under sonication. The electrodeposition was performed on a CHI 760D electrochemical workstation at \u20131.9\u00a0V (versus SCE) for 600 s. The Co(VOx) mass loading on NF was 2.3\u00a0\u00b1 0.4\u00a0mg cm\u22122. The preparation of other control samples can be found in detail in the Supplemental Experimental Procedures.All the electrochemical measurements were carried out on a CHI 760D electrochemical workstation by using the prepared Co(VOx) and Co electrodes on NF as the working electrode and graphite plate and double junction SCE (saturated KCl) as the counter and reference electrode, respectively. All potentials measured were calibrated to the reversible H2 electrode (RHE) through the following equation: E\nRHE\u00a0= E\nSCE\u00a0+ 0.241\u00a0V\u00a0+ 0.059 pH. The linear sweep voltammetry (LSV) polarization curves were recorded at a scan rate of 3\u00a0mV s\u22121 in 1\u00a0M KOH without iR compensation. Chronoamperometric measurements were obtained under the same experimental setup. EIS spectra were measured at 150 mV overpotential in the frequency range of 0.1\u2013100000\u00a0Hz with an amplitude of 10\u00a0mV in 1\u00a0M KOH electrolyte.All of the spin-polarized DFT calculations were performed using the vienna Ab initio simulation package (VASP) program,\n37\u201339\n which uses a plane-wave basis set and a projector augmented wave method (PAW) for the treatment of core electrons. The Perdew, Burke, and Ernzerhof exchange-correlation functional within a generalized gradient approximation (GGA-PBE)\n40\n was used in our calculations. For the expansion of wave functions over the plane-wave basis set, a converged cutoff was set to 450 eV. Spin-polarization effect and dipole correction were considered in all cases.The structural model of bare Co(001) was constructed as a 4\u00a0\u00d7 4 periodic supercell (Figure\u00a0S14A), which contains four atomic layers with the bottom two layers fixed in their respective bulk positions and all the other atoms fully relaxed. Experimental observations show that V-dopant atoms are oxidized and adsorb on the Co(001) surface. In order to simulate V-doped Co(001) catalysts, VO4 clusters are added to the Co(001) surface to model V-doped Co(001) catalysts with different VO4 coverages (V-Co(001) (\u03b8\u00a0= 0.25, 0.50 and 0.75)). For example, Figure\u00a0S14B shows the simulation model of V-Co(001) (\u03b8\u00a0= 0.25). The vacuum space was set to larger than 18\u00a0\u00c5 in the z direction to avoid interactions between periodic images. In geometry optimizations, all the structures were relaxed up to the residual atomic forces smaller than 0.005 eV/\u00c5, and the total energy was converged to 10\u22124 eV. The Brillouin zone was sampled using 3\u00a0\u00d7 3\u00a0\u00d7 1 \u0393-centered mesh. The deformation electronic density of the V-Co(001) was defined as \n\n\u0394\u03c1\n\n(\nr\n)\n\n=\n\u00a0\u03c1\n\n\n(\nr\n)\n\n\nV\n\u2212\nCo\n\n(\n001\n)\n\n\n\n\u2212\n\u03c1\n\n\n(\nr\n)\n\n\nCo\n\n(\n001\n)\n\n\n\n\u2212\n\u03c1\n\n\n(\nr\n)\n\n\nVO\n4\n\n\n\n, where \n\n\u03c1\n\n\n(\nr\n)\n\n\nV\n\u2212\nCo\n\n(\n001\n)\n\n\n\n\n represent the charge density of the V-Co(001) system, and \n\n\u03c1\n\n\n(\nr\n)\n\n\nCo\n\n(\n001\n)\n\n\n\n\n and \n\n\u03c1\n\n\n(\nr\n)\n\n\nVO\n4\n\n\n\n represent the charge density of the bare Co(001) and the VO4 cluster at the same coordinates as those in the V-Co(001) system, respectively.The overall HER mechanism was evaluated with a three-state diagram consisting of an initial H+ state, an intermediate H\u2217 state, and 1/2 H2 as the final product. The free energy of H\u2217\n\n\n\n(\n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n)\n\n\n is proven to be a key descriptor to characterize the HER activity of the electrocatalyst. A electrocatalyst with a positive value leads to low kinetics of adsorption of H, whereas a catalyst with a negative value leads to low kinetics of release of H2 molecules.\n35\n The optimum value of \n\n\n|\n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n|\n\n\n should be zero; for instance, this value for the well-known highly efficient Pt catalyst is near zero as \n\n\n|\n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n|\n\n\u2248\n0.09\n\u00a0eV\n\n.\n5\n The \n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n is calculated as\n35\n\n\n\n(Equation\u00a01)\n\n\n\u0394\n\nG\n\nH\n\u2217\n\n\n=\n\u0394\n\nE\n\nH\n\u2217\n\n\n+\n\u0394\n\nE\nZPE\n\n\u2212\nT\n\u0394\n\nS\nH\n\n\n\n\nwhere \n\n\u0394\n\nE\n\nH\n\u2217\n\n\n\n is the binding energy of adsorbed H and \n\n\u0394\n\nE\nZPE\n\n\n and \n\n\u0394\n\nS\nH\n\n\n are the difference in zero point energy (ZPE) and entropy between the adsorbed H and H2 in the gas phase, respectively. As the contribution from the vibrational entropy of H in the adsorbed state is negligibly small, the entropy of H adsorption is \n\n\u0394\n\nS\nH\n\n\u2248\n\u2212\n\n1\n/\n2\n\n\nS\n\nH\n2\n\n\n\n, where \n\n\nS\n\nH\n2\n\n\n\n is the entropy of H2 in the gas phase at the standard conditions. Therefore, the \n\n\u0394\n\nG\n\nH\n\u2217\n\n\n\n value for the Co(001) surface should be \n\n\u0394\n\nE\nH\n\n+\n0\n.\n24\n\u00a0eV\n\n.\n35\n\nMore details of the characterization methods followed, XAS data collection, and operando Raman measurements are provided in Supplemental Experimental Procedures.All physical characterizations were carried out at the Mark Wainwright Analytical Centre (MWAC) and Electron Microscope Unit at the University of New South Wales (UNSW). XAS spectra were recorded on the multiple wiggler XAS beamline 12 ID at the Australian Synchrotron (AS1/XAS/15778). Thanks to Dr. Hangjuan Ren for the schematic drawing. Thanks to Dr. Rosalie Hocking at Swinburne University of Technology and Dr. Bernt Johannessen at Australian Synchrotron for the help of XAS data collection. Thanks to Mr. Kamran Dastafkan for proofreading the paper. This research was undertaken with the assistance of resources provided by the Pawsey and the National Computational Infrastructure (NCI) facilities at the Australian National University, which were allocated through the National Computational Merit Allocation Scheme and supported by the Australian Government and the Australian Research Council grant (LE190100021). C.Z. is grateful for the award of a Future Fellow from Australian Research Council (FT170100224).C.Z. and Y.L. designed the experiments. Y.L. undertook electrochemical experiments; performed the XRD, TEM, and XAS; and interpreted data. X.T. and S.C.S. undertook the DFT calculation. W.Y. collected the Faradic efficiency data, X.B. helped to collect the XAS data, Z.S. performed XPS, and T.Z. collected SEM. C.Z., Y.L., X.T., and S.C.S. co-wrote the paper. All authors discussed the results and commented on the manuscript.The authors declare no competing interests.Supplemental Information can be found online at https://doi.org/10.1016/j.xcrp.2020.100275.\n\n\nDocument S1. Figures S1\u2013S21 and Table S1\n\n\n\n\n\nDocument S2. Article plus Supplemental Information\n\n\n\n", "descript": "\n                  Promoting active and stable H2 evolution reaction (HER) on metallic Fe triad materials is important yet challenging. Here, we report a metallic Co catalyst modulated by vanadium oxide (VOx) clusters (denoted as Co(VOx)) for active alkaline HER activity. Systematic X-ray absorption spectroscopy (XAS) and X-ray crystallography studies verify that VOx clusters endow Co with a highly disordered lattice and downsized particle size. The best Co(VOx) electrode is achieved with an optimal doping level of 3%, which delivers \u2212100 mA cm\u22122 at an overpotential merely of 178\u00a0mV, in contrast to 344\u00a0mV and continuous activity decay on pure Co. The lower or higher doping level is unable either to regulate the atomic structure and reduce the H binding or to provide fewer active sites. Density functional theory (DFT) calculations reveal that VOx enables efficient electron transfer from Co to VOx, thus decreasing the H-adsorption on V-Co(001) for enhanced HER.\n               "}
{"full_text": "Data will be made available on request.Worldwide, there is a great desire to develop technologies for the efficient capture and conversion of carbon dioxide to fuels and chemicals, such as methanol [1\u20133]. Methanol is a carrier of carbon and hydrogen [4,5] and, as an energy carrier, it can be used directly as a fuel in direct methanol fuel cells (DMFCs) and internal combustion engines (ICEs) [6,7]. Additionally, methanol can be used in the production of high value-added chemicals (e.g. formaldehyde, methyl tert-butyl ether and acetic acid) and as a feedstock to produce hydrocarbons (such as alkanes, olefins or aromatics) and inherently fuels [6\u20138].Traditionally, Cu/ZnO-based catalysts have been employed in the industrial production of methanol from the syngas stream (CO/CO2/H2) generated in the steam reforming of natural gas [9\u201311]. The mechanism of methanol formation from CO/CO2 has been under debate for decades [12,13]. Since CO is the predominant carbon-containing molecule in syngas [2,12] and Cu is an outstanding metal for CO2 reduction to CO (through reverse water-gas shift, CO2 + H2 \u21cc CO + H2O), CO has been considered the primary carbon source in methanol production for decades (CO + 2\u00a0H2 \u21cc CH3OH) [10,12]. It was in the late 1980\u00a0s when the use of 14C-labeled isotopes provided evidence to suggest CO2 as the main carbon source in the methanol production (CO2 + 3\u00a0H2 \u21cc CH3OH + H2O) from CO2/CO/H2 mixtures [14]. Since then, many studies have been driven in the same direction [10,12,15\u201317], and currently, CO2 is perceived as the major reactant under industrial conditions. However, as Grabow and Mavrikakis suggested based on density functional theory (DFT) calculations and microkinetic modeling [12], under typical methanol production conditions, both CO and CO2 hydrogenation routes can coexist.In addition to the active discussion about the carbon source, the knowledge gained regarding the nature of the active sites in the CO2-to-methanol (CTM) process has grown exponentially in recent decades [18\u201320]. CO2 is a stable molecule and greatest difficulties in achieving great methanol selectivity in CO2 hydrogenation are related to kinetic limitations [21]. Therefore, a molecular understanding of the key aspects that govern the activity and selectivity of a catalyst is crucial. Overall, over copper-based catalysts, the CO2 hydrogenation to methanol reaction has been described as a structure-sensitive reaction in which not all of the surface atoms have the same role and activity [15,18,22]. For copper-zinc oxide (Cu-ZnO) binary systems, copper is responsible for the adsorption, dissociation and spillover of atomic hydrogen (H*) [23], while zinc oxide enhances the dispersion of Cu nanoparticles and facilitates the adsorption of CO2\n[24]. The Cu-ZnO interface and surroundings have been described as the most active sites responsible for the activity wherein the intermediate species (e.g. carbonates and formates) are further hydrogenated to methanol [11,25\u201327]. Lately, ZrO2-based catalysts are emerging as active [18,28,29] and cost-effective solutions for the efficient synthesis of methanol [30] and a few experimental studies have explored the synergistic interactions of Cu/ZnO/ZrO2 catalysts and the active interplay toward methanol production [10,18,31,32]. Alone on ZrO2, both Cu [28,33] and ZnO [34,35] can also display some activity in the hydrogenation of intermediate species to methanol. Also, the ZrO2/Cu inverse configuration has shown excellent properties for an efficient methanol synthesis from CO2\n[36,37].In recent years, more investigations have described not only the synergistic effect of binary Cu/ZnO catalysts but also the effects of the locations of both the active metal (Cu) and promoter (ZnO) on methanol production from CO2\n[19,38,39]. The formation of ZnO particles or reduced Zn on the Cu surface have been found to improve the Cu dispersion and eventually, the accessible Cu surface area [11,19,25,39]. Experimental and computational studies on CO2 hydrogenation to methanol have pointed out that the formation of ZnO aggregates on top of Cu particles promoted methanol production which may be related to the increase in the number of active ZnO-Cu pairs [19,38]. Palomino et al. [38] experimentally demonstrated that ZnO added on top of Cu (100) and (111) surfaces yields a superior methanol production compared to the inverse copper-added-on-top-of-zinc oxide catalyst. Moreover, the highest production of methanol was observed at a relatively low surface coverage (\u03b8Zn) of 0.15\u20130.20\u00a0monolayer (ML) and similar values of \u03b8Zn \u2248\u00a00.20\u00a0ML were reported by Nakamura and coworkers for ZnO over polycrystalline copper [40], and Kattel and collaborators over Cu(111) substrates [19]. By a combination of experimental, and DFT calculations and modeling based on thermodynamics, Kuld and coworkers [41] found the highest methanol turnover frequency (TOF) at a surface coverage of \u03b8Zn \u2248\u00a00.47\u00a0ML, with the TOF being greater when using larger Cu particles.The above examples highlight the potential of and interest in synthesizing and testing zinc-on-top-of-copper catalysts with ZnO surface coverages of approximately \u03b8Zn \u2248\u00a00.1\u20130.2\u00a0ML. An atomic-scale synthesis technique, such as atomic layer deposition (ALD), is an efficient technique to reach this range of surface coverage. The ALD technique is based on the sequential use of self-terminating gas\u2013solid reactions and can offer accurate atomic level control of the deposited metal concentrations [42,43]. To modify metal oxide interfaces, single atoms can be uniformly distributed on high surface area supports by ALD [43\u201345]. The first studies in atomic-scale synthesis (by ALD) toward CTM reaction have already been reported (e.g. ZrO2-ALD on Cu/SiO2\n[46] and Ni-ALD on Cu nanoparticles on\u00a0\u03b3-Al2O3\n[47]), bringing out the benefits of this technique to enhance the catalyst activity, selectivity and stability. In a recent publication, Saedy and coworkers [48] applied preferential chemical vapor deposition (PCVD) and incipient wetness impregnation in the synthesis of ZnO/Cu/Al2O3 catalysts for the CTM reaction. Zinc oxide introduced by the PCVD method on a prereduced copper phase resulted in a more active and selective catalyst compared to the impregnated catalyst [48]. By means of various diffraction, spectroscopic and microscopy characterization techniques, the investigators demonstrated a more efficient production of active/selective ZnO/Cu interfaces compared to the more traditional impregnation method [48]. Furthermore, the authors suggested that the inverse ZnO/Cu interface may result in a more active system than the conventional Cu/ZnO interface [48]. In a recent research [49], ZnO was added by ALD (323\u00a0K, diethylzinc as a precursor) on copper hydroxide nanowires and the size of ZnO was tuned from isolated species to nanoparticles by increasing the number of ZnO cycles from 1 to 20. The maximum methanol production rate was found after 3 cycles.In this work, we focused on studying the catalytic performance of diverse copper-zinc oxide on zirconia catalysts for the hydrogenation of carbon dioxide to methanol. Atomic layer deposition (ALD) and incipient wetness impregnation were applied for the incorporation of Zn and Cu into the catalyst, respectively. Zn was deposited by ALD in one ALD cycle, which in practice should correspond to atomically dispersed ZnO species covering 10\u201320% of the surface (0.1\u20130.2\u00a0ML). By alternating the order in which Cu and Zn were attached to the catalyst, we created different metal-oxide configurations. A combination of various characterization techniques, such as diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), hydrogen temperature programmed reduction (H2-TPR), carbon dioxide temperature programmed desorption (CO2-TPD), X-ray photoelectron and absorption spectroscopy (XPS and XAS) and scanning transmission electron microscopy-energy-dispersive X-ray spectroscopy (STEM-EDS), DFT calculations and catalytic tests, led us to identify the most active configuration. We expect that the findings of this research shall enhance the understanding of the elemental features in the zinc oxide/copper/zirconia system and help to consider the location of zinc oxide (promoter) and copper (metal) as a crucial parameter to produce active sites for the efficient hydrogenation of carbon dioxide to methanol.In total, five samples were synthesized: ZnO/ZrO2, Cu/ZrO2, ZnO/Cu/ZrO2, Cu/ZnO/ZrO2 and ZnO/Cu/ZnO/ZrO2. The self-made samples were prepared following two different methods, depending on which metal was incorporated into the porous structure of the support (monoclinic zirconia, ZrO2). Thus, Cu was added by incipient wetness impregnation (IWI), while Zn was added by atomic layer deposition (ALD). Monoclinic zirconia, provided by Saint-Gobain NorPro as cylindrical pellets (length 5\u00a0mm, diameter 3\u00a0mm) was used as a support material (surface area of 70\u00a0m2 g\u22121). Prior to its utilization, ZrO2 was crushed and sieved to a particle size of 250\u2013420\u00a0\u00b5m and calcined in a muffle furnace (Nabertherm P330) in ambient air at 873\u00a0K for 5\u00a0h (10\u00a0K\u00a0min\u22121) to remove possible surface impurities. Cu nitrate trihydrate, Cu(NO3)20.3\u00a0H2O (CAS: 10031\u201343\u20133, Sigma Aldrich, 99\u2013104% purity) and Zn acetylacetonate, Zn(C5H7O2)20.3\u00a0H2O (Zn(acac)2, CAS: 14024\u201363\u20136, Volatec) were used as copper and zinc precursors, respectively. The targeted areal number density for Cu and Zn was 2 atoms/nm2 (Zn/Cu atomic ratio of one). The Cu and Zn loadings in wt% measured by ICP\u2013OES are shown in \nTable 1, and a scheme that shows the sequence of samples prepared is shown in \nFig. 1.For the impregnation of Cu by IWI, either on ZrO2 or ZnO/ZrO2, the corresponding amount of Cu precursor was dissolved in the exact amount of deionized water needed to fill the pore volume of the support. The water uptake capacity of the support (\u2248 0.3\u2009mL\u2009g\u22121) was experimentally estimated by adding deionized water drop by drop to a known amount of dried support (393\u2009K, 24\u2009h). Approximately 3\u20135 drops of the Cu nitrate solution were added at a time to an Erlenmeyer flask containing the dried support. Next, the partially wet support was first gently mixed, and then the flask was shaken for 2\u20133\u2009min to ensure an even distribution of the solution. After adding the final drops of the solution, the slightly damp material was aged for 5\u2009h at room temperature and dried overnight at 393\u2009K in an oven under static air. Finally, the dried material was calcined at 673\u2009K for 2\u2009h (5\u2009K\u2009min\u22121) in a tube furnace with a constant flow of 100\u2009mL\u2009min\u22121 of synthetic air (AGA 99.999% purity, 20% O2, 80% N2).The deposition of Zn on ZrO2 by ALD started by treating the calcined support in a flow-type fixed bed F-120 ALD reactor (ASM Microchemistry) at 523\u2009K for 10\u2009h to remove moisture before the actual ALD process. Then, the solid zinc acetylacetonate reactant was vaporized at 393\u2009K in flowing nitrogen and reacted to the pretreated ZrO2 support by one cycle of the ALD process for 3\u2009h at 473\u2009K and a pressure of ca. 3 mbar. Reactant-originated acetylacetonate ligands were removed by oxidative treatment in a tube furnace in synthetic air flow (100\u2009mL\u2009min\u22121) at 773\u2009K for 2\u2009h (5\u2009K\u2009min\u22121). The same procedure was followed when Zn was deposited on ZrO2, Cu/ZrO2 and Cu/Zn/ZrO2. \nFig. 2 shows a conceptual scheme with the configurations of the various copper-zinc-zirconia catalysts that were synthesized and tested in this research.The metal content of the catalysts was determined by inductively coupled plasma atomic emission spectroscopy (ICP\u2013OES). Samples (ca 0.100\u2009mg) were weighed in Teflon vessels, and a mixture of nitric acid (HNO3, 65%, 2.5\u2009mL) and hydrochloric acid (HCl, 37%, 7.5\u2009mL) was added. Vessels were closed and placed in a microwave oven (Milestene, Ethos) and heated (1\u2009h, 200\u2009\u00b0C). After cooling, the samples were diluted with MQ-grade water, and the Cu- and Zn-contents were determined with an F-AAS instrument (Varian 220\u2009F) using an air-acetylene burner.The surface area and cumulative pore volume of zirconia were obtained by nitrogen physisorption isotherm (liquid nitrogen, 77\u2009K) in a Thermo Scientific Surfer equipment. The support sample was weighted to a quartz glass burette (ca. 200\u2009mg) and degassed at 573\u2009K for 3\u2009h. Specific surface area was calculated from the isotherm according to the Brunauer-Emmett-Teller (BET) method [50]. The cumulative pore volume was calculated based on the Barrett-Joyner-Halenda (BJH) method [51].The phase/crystallinity of the support was studied by X-ray diffraction that was carried out on a ground sample in a PANanalytical X\u2018\u0301Pert Pro MPD Alpha 1 device equipped with Cu K\u03b11 radiation (45\u2009kV and 40\u2009mA). The X-ray scanning range was from 10\u00b0 to 100\u00b0 (2\u03b8) with a step size of 0.0131\u00b0 and a time per step of 51\u2009s. The results are shown in the supporting information (Fig. S1). The characteristic monoclinic phase (JCPDS 37\u20131484) was identified with main reflections at 24.5\u00b0, 28.3\u00b0, 31.5\u00b0, 34.2\u00b0, 35.4\u00b0, 40.8\u00b0, 49.3\u00b0, 50.2\u00b0, 54.1\u00b0 and 55.5\u00b0.The reducibility of the metal oxides was studied by hydrogen temperature-programmed reduction. The experiments were performed using an Altamira AMI-200 characterization system with a thermal conductivity detector (TCD) connected to an OmniStarTM mass spectrometer (MS) produced by Pfeiffer Vacuum. A total of 150\u2009mg of sample was placed in a U-shaped quartz reactor and treated in constant He flow (AGA 99.999% purity) at 473\u2009K for 60\u2009min and cooled back to 303\u2009K. The sample was then heated from 303 to 873\u2009K or 1173\u2009K in 2% H2/Ar (AGA 99.999% purity) with a heating ramp of 5\u2009K\u2009min\u22121. The TPR measurement of ZrO2 and ZnO/ZrO2 was performed up to 1173\u2009K, while 873\u2009K was used as the maximum temperature for the Cu-containing samples. The TPR results are qualitative (given as arbitrary units) and the areas under the peaks cannot be compared with each other. The total gas flow was set at 50\u2009mL\u2009min\u22121 (STP conditions) during the whole measurement.Similarly, carbon dioxide desorption was studied by CO2 temperature-programmed desorption. The experiments were performed using an Altamira AMI-200 characterization system with a TCD connected to an OmniStarTM MS produced by Pfeiffer Vacuum. A total of 150\u2009mg of sample was placed in a U-shaped quartz reactor, treated in He flow at 473\u2009K for 60\u2009min, and cooled back to 303\u2009K. Then, the sample was activated by heating the solid from 303 to 623\u2009K in 2% H2/Ar with a heating ramp of 10\u2009K\u2009min\u22121 and 60\u2009min hold time. The sample was then cooled to 323\u2009K in He flow and maintained at that temperature for 30\u2009min. Thereafter, the reduced sample was exposed to a constant flow of 10% CO2/He (AGA 99.999% purity) for 30\u2009min at 323\u2009K and flushed for 60\u2009min at the same temperature in He flow to remove the physisorbed CO2. Finally, the sample was heated from 323 to 1073\u2009K in He flow to desorb the chemisorbed CO2. The possible desorbed products were continuously monitored during the experiment by mass spectrometry. The amount of CO2 desorbed for each measurement was quantified by using calcium carbonate (CaCO3) as an internal standard. About 4\u2009mg of CaCO3 was mixed with the sample and decomposed during the desorption step into CO2 (g) and CaO (s) in the range of 800\u2013950\u2009K. The total gas flow was set at 50\u2009mL\u2009min\u22121 (STP conditions) for all measurements.The electronic structure of the samples was studied by high-energy-resolution fluorescence-detected X-ray absorption spectroscopy (HERFD-XAS). The experiments were performed at the ID20 beamline of the European Synchrotron Radiation Facility (ESRF, Grenoble, France) [52,53]. The beam was monochromated by a combination of a Si(111) premonochromator and a Si(311) channel-cut monochromator. The spectrometer was a von Hamos spectrometer based on three Si(333) crystal analyzers.X-ray photoelectron spectroscopy (XPS) was used to study the surface composition of the reduced samples (623\u2009K for 30\u2009min, 2% H2/Ar, transfer to XPS was done through air). The measurements were performed with a Kratos AXIS Ultra DLD X-ray photoelectron spectrometer using a monochromated AlK\u03b1 X-ray source (1486.7\u2009eV) run at 100\u2009W. A pass energy of 80\u2009eV and a step size of 1.0\u2009eV were used for the survey spectra, while a pass energy of 20\u2009eV and a step size of 0.1\u2009eV were used for the high-resolution spectra. Photoelectrons were collected at a 90\u00b0 take-off angle under ultra-high vacuum conditions, with a base pressure typically below 1\u00d710\u22129 Torr. The diameter of the beam spot from the X-ray was 1\u2009mm, and the area of analysis for these measurements was 300\u2009\u00b5m x 700\u2009\u00b5m. Both survey and high-resolution spectra were collected from three different spots on each sample surface in order to check for homogeneity and surface charge effects. All spectra were charge-corrected relative to the position of C-C bonding of adventitious carbon at 284.8\u2009eV.Scanning transmission electron microscopy high-angle annular dark-field (STEM-HAADF) images were acquired for the prereduced samples (623\u2009K, 60\u2009min, 50\u2009mL\u2009min\u22121 of 2% H2/Ar, STP conditions) by a JEOL JEM-2200FS double aberration corrected, high-resolution microscope, operated at 200\u2009kV acceleration voltage. The chemical elemental mapping analysis was conducted with an X-ray energy-dispersive spectroscopy (EDS) detector. The samples were drop-casted using acetone onto a gold grid coated with an ultrathin holey carbon film.The evolution of the surface species during the cyclic CO2-H2 adsorption was measured by diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) carried out in a Nicolet Nexus FTIR spectrometer with a high temperature/pressure Spectra-Tech reactor chamber equipped with a dome and ZnSe windows. The gas flow leaving the chamber was monitored with an OmniStar GSD 301 spectrometer by Pfeiffer Vacuum. The scans were collected from 4000 to 600\u2009cm\u22121 at a scan resolution of 4\u2009cm\u22121. Prior to the experiment, a background spectrum was acquired under room conditions using an aluminum mirror in constant Ar flow. Approximately 20\u2009mg of crushed sample powder was placed in the sample holder, heated from room temperature to 673\u2009K in a constant flow of 10% O2/N2/Ar (synthetic air: AGA 99.999% purity, 20% O2, 80% N2; Ar: AGA 99.9999% purity) and kept there for 60\u2009min to remove possible surface impurities. The temperature was then decreased to 623\u2009K in the same atmosphere, and the gas was switched to 10% H2/Ar (H2: AGA 99.999% purity; Ar: AGA 99.9999% purity) and kept for 60\u2009min to activate the metal oxides. The catalyst was then cooled to the desired reaction temperature, either 450, 500 or 550\u2009K, and flushed with Ar for 30\u2009min. The cyclic adsorption of CO2-H2 consisted of three consecutive full cycles of the following sequence: i) CO2 adsorption for 12\u2009min under flowing 10% CO2/He (AGA 99.999% purity), ii) switching to Ar flow for 12\u2009min, iii) switching to H2 adsorption for 12\u2009min under flowing 10% H2/Ar and, iv) switching to Ar flow for 12\u2009min. During each step within each cycle, a spectrum (100 scans, approximately 2\u2009min) was recorded at 0, 5 and 10\u2009min to monitor the surface of the catalyst with respect to time on stream. The total gas flow was set at 50\u2009mL\u2009min\u22121 (STP conditions) during the whole experiment. A summary of the experimental conditions (timing, gases, flow rates, and temperatures) used for DRIFTS experiments is depicted in \nFig. 3.To assist infrared band identification, density functional theory (DFT) calculations were performed using the BEEF\u2013vdW exchange\u2013correlation functional [54] as implemented in the GPAW [55] software. The monoclinic zirconia was described by a two-layer-thick slab model, built from a 3\u2009\u00d7\u20092\u2009m-ZrO2(111) supercell with periodic boundary conditions used in the lateral directions. The final cell measurements were 20.67\u2009\u00d7\u200914.79\u2009\u00d7\u200924.0 \u00c5 with angles of 90\u00b0/90\u00b0/116.5\u00b0. Cu(111) and stepped Cu(110) surfaces were modeled as three-layer periodic slabs, where the bottom layers were kept fixed in their bulk geometry in unit cells of 4\u2009\u00d7\u20094 and 3\u2009\u00d7\u20094, respectively. The core electrons of all elements were described by projector-augmented wave (PAW) [56] setups in the frozen-core approximation. A real\u2013space grid basis was used with a maximum grid spacing of 0.2 \u00c5. Periodic boundary conditions were used in two directions and the reciprocal space was sampled at the \u0393 point. A Hubbard U correction [57] of 2.0\u2009eV was applied to the d\u2013orbitals of the zirconium atoms. The atomic structures were optimized using the Fast Inertial Relaxation Engine (FIRE) algorithm as implemented in the Atomic Simulation Environment (ASE) [58,59] package until the maximum residual force was below 0.005\u2009eV \u00c5\u22121. Vibrational frequencies for modes involving adsorbate and binding catalyst surface atoms were determined using the Frederiksen method [60]\n. The frequencies of the combined modes were obtained by adding up the frequencies of their individual contributions.The catalytic performance was evaluated in a high-pressure continuous-flow fixed-bed equipped with a stainless-steel tube reactor with a mesh placed in the midsection of the reactor. One gram of calcined catalyst, sieved to 0.25\u20130.42\u2009mm, was loaded in the reactor. Prior to the catalytic reaction, the catalyst was activated by in situ reduction at 623\u2009K for 60\u2009min with a constant flow of 10% H2/N2 (v/v; H2: AGA 99.999% purity, N2: AGA 99.999% purity). Activity tests were conducted at 450, 500 and 550\u2009K with a total pressure of 3.0\u2009MPa and a gas hourly space velocity (GHSV) of 7500\u2009h\u22121 (STP conditions: 273.15\u2009K and 1\u2009bar). The reaction mixture was composed of H2/CO2/N2 (\u2053 71/23/6, v/v/v; H2: AGA 99.999% purity; CO2, AGA: 99.995%; N2: AGA 99.999% purity). The volumetric flows were 6.3\u2009L\u2009h\u22121 of H2, 2\u2009L\u2009h\u22121 of CO2 and 0.6\u2009L\u2009h\u22121 of N2 (STP conditions) and the volume of catalyst used per experiment was approximately 1.2 10\u22123 L. For each catalyst, the reaction temperature was increased in steps of 50\u2009K. Initially, the reactor temperature was stabilized at 450\u2009K and kept there for 90\u2009min. Next, the temperature was increased to 500\u2009K (10\u2009K\u2009min\u22121), stabilized, and kept there for another 90\u2009min. Finally, the procedure was repeated for the highest temperature of 550\u2009K. Thus, the data depicted in this manuscript were collected after 90\u2009min under each operating condition. The unreacted gases and reaction products were continuously monitored in an Agilent 490 Micro Gas Chromatograph (microGC) fitted with a thermal conductivity detector (TCD) and equipped with two columns: a) MS-5 molecular sieve for permanent gases H2, N2, CH4 and CO and, b) PoraPLOT U for CO2, CH3OH and H2O. The CO2 conversion (X\nCO2, Eq. 1) and product selectivity (Si, Eq. 2) were calculated by internal normalization standard with N2. The CO2 conversion, the selectivity of CH3OH, CO and CH4 and the space-time yield of CH3OH (STY\nCH3OH, mmol h\u22121 gcat\n\u22121, Eq. 3) were calculated according to the following formulas:\n\n(1)\n\n\n\n\nX\n\n\n\n\nCO\n\n\n2\n\n\n\n\n(\n%\n)\n=\n\n\n\nF\n\nCO\n2\n,\nin\n\n\n-\n\nF\n\nCO\n2\n,\nout\n\n\n\n\nF\n\nCO\n2\n,\nin\n\n\n\n*\n100\n%\n\n\n\n\n\n\n(2)\n\n\n\nS\ni\n\n(\n%\n)\n=\n\n\nF\n\ni\n,\nout\n\n\n\n\nF\n\nCH\n3\nOH\n,\nout\n\n\n+\n\nF\n\nCO\n,\nout\n\n\n+\n\nF\n\nCH\n4\n,\nout\n\n\n\n\n*\n100\n%\n\n\n\n\n\n\n(3)\n\n\n\n\n\n\n\n\n\n\n\nSTY\n\nC\n\nH\n3\n\nOH\n\n\n(\nmmol\n\n\n\nh\n\n\n\u2212\n1\n\n\n\n\n\ng\ncat\n\n\n\n\u2212\n1\n\n\n)\n=\n\n\n\n\nF\n\nCO\n2\n,\nin\n\n\n*\n\nX\n\nCO\n2\n\n\n*\n\nS\n\nC\n\nH\n3\n\nOH\n\n\n\n\nm\ncat\n\n\n\n\n\n\n\n\n\nwhere F\nCO2 and F\n\ni\n are the molar flow rates of CO2 or products (CH3OH, CO and CH4) and m\ncat is the mass of catalyst in grams.The list of samples (support and self-made catalysts), the corresponding sample code used throughout the manuscript, the Cu and Zn metal loadings measured by ICP\u2013OES and the Cu and Zn areal number density, the amount of CO2 desorbed in the CO2-TPD experiments and the number of reduction peaks (for Cu-species) and temperature at the maximum height of the peak determined by H2-TPR are listed in Table 1.The STEM-HAADF images and EDS mapping of elements for the support and self-made samples are depicted in \nFig. 4. Based on the results, the elements mapped for each sample seem to be evenly distributed throughout the samples and no high concentration spots were detected in any of the samples. The estimation of the Cu and Zn particle size distributions was impossible due to the poor contrast between Cu, Zn and Zr.Also, the percentage of a monolayer of ZnO (% ML) was estimated as a ratio between the areal number density of ZnO in each sample (Zn atoms per nm2, calculated from the Zn metal loadings measured by ICP\u2013OES) and the average areal number density of Zn in one bulk ZnO monolayer obtained through equation reported elsewhere [42] (12.0\u2009nm\u22122, from ZnO density = 5.61\u2009g\u2009cm\u22123). According to these calculations, one Zn-ALD cycle (Zn/Zr, Cu/Zn/Zr and Zn/Cu/Zr samples) yielded approximately 15% of a bulk ZnO ML equivalent.The electronic structure (oxidation states) of copper and zinc was investigated by high-energy-resolution fluorescence-detected X-ray absorption spectroscopy (HERFD-XAS). The Cu and Zn absorbance at the Cu and Zn K-edges are depicted in \nFig. 5 for the various catalysts and references (Cu foil, Cu2O, CuO, Zn foil, ZnO and Zn(acac)2). Along with the samples prepared and reported in Table 1, two more samples were prepared and analyzed: Zn/Cu/Zr (5\u2009wt% Cu) and Zn/Cu/Zr (15\u2009wt% Cu). These samples were synthesized to observe a possible effect of the Cu loading on the electronic structure of Cu and Zn. All the samples analyzed by HERFD-XAS were studied after the calcination treatment detailed under Section 2.1 (673\u2009K or 773\u2009K when Cu or Zn were added last, respectively), except a Zn/Zr sample that was also studied before the calcination step, i.e., after the Zn-ALD cycle.Analyzing the Cu K-edge HERFD-XAS and first derivative (relative to incident energy) spectra (Fig. 5a-b) of the Zn/Cu/Zr samples with various Cu loadings, the spectra of the samples with 5\u2009wt% and 15\u2009wt% Cu were similar to that of CuO, which indicated a clearer presence of bulk CuO when increasing the Cu content. The characteristic edge transition for CuO can be observed at ca. 8.984\u2009keV [32]. The rest of the samples (Zn/Cu/Zr with 1.1\u2009wt% Cu, Cu/Zn/Zr and Cu/Zr) showed the absence of the first main shoulder that caused an apparent shift of the first peak of the derivative spectrum toward higher energies (Fig. 5b). The shoulder in CuO stems from 1\u2009s to 4p transitions with a charge transfer from the ligand (1s13d10 _L_ final state, with _L_ denoting a ligand hole) [61,62]. This ligand effect decreases from CuO to the catalysts and it is typically absent in tetrahedral Cu(II) complexes [63]. The local coordination in bulk CuO is nearly square planar while the difference in the catalysts indicates a different local coordination in the surface-dominated Cu species. Thus, the pre-edge intensity (at 8.979\u2009eV) decreases from CuO to the catalysts, indicating more likely an octahedral coordination in our samples, where the pre-edge is dipole-forbidden. However, the absence of the characteristic Cu1+ 1\u2009s-4p transition feature at ca. 8.98\u2009keV [63] indicates the absence of Cu2O in any of the samples. Thus, it can be concluded that the Cu valence state seemed to be Cu2+ for all the catalysts.For the Zn K-edge (Fig. 5c-d), the HERFD-XAS spectra and 1st derivative were very similar for all the samples, which suggests that the chemical environment for Zn is not significantly influenced either by the order in which Cu and Zn were added to the catalyst or by the Cu loading. Comparing the HERFD-XAS spectra of the samples with the ZnO (wurtzite) reference, there were significant differences; however, the presence of the Zn2+ oxidation state can be assumed. The peak at 9.679\u2009eV present on ZnO reference, assigned to multiple scattering on atomic neighbors beyond the first shell [64], was not visible in any of the samples (except for a small peak observed on the Cu/Zn/Zr sample) which evidenced the presence of atomically dispersed ZnO species. It was computationally [64] and experimentally [49] demonstrated that ZnO XANES features are developed while increasing the number of ZnO atomic shells and, consequently, the ZnO cluster size. To the best of our understanding, it was expected that Zn was atomically dispersed as lone ZnO units with an oxidation state of Zn2+.To complement the bulk structural-chemical information obtained by HERFD-XAS, the surface composition was analyzed by XPS. The samples were reduced ex situ and momentarily exposed to the atmosphere during the preparation before the analysis.The elemental composition and the relative amount of the diverse zinc components are included in \nTable 2. The survey spectra of the samples in the range of 0\u20131200\u2009eV are included in Fig. S2 and the high-resolution XPS spectra of the Zn 2p and Cu 2p regions are depicted in \nFig. 6. As expected, all samples exhibited zirconium and oxygen, the relative amount of which decreased when adding other components (Table 2). The surface concentration of copper and zinc content measured by XPS varied between 2\u20134 and 3\u20137\u2009at%, respectively, for the samples. The Zn/Cu atomic ratio was close to two for the samples with one ALD cycle of Zn and Cu by impregnation. Compared to the bulk atomic ratio of about one (see Table 1), the observed Zn/Cu ratio is consistent with zinc located on the surface and copper somewhat clustered. Addition of zinc on Cu/Zr decreased the surface concentration of copper, as expected (from 3.0 to 1.9\u2009at% for Zn/Cu/Zr). According to XPS, the Cu/Zn/Zr sample had a higher surface concentration of both copper and zinc than the inverse Zn/Cu/Zr. While the reason for this observation is not fully clear, we speculate that it may have to do with the details of impregnation (zinc oxide is amphoteric [65] and the impregnation solution was acidic; part of the zinc may have dissolved and migrated to the outer surface during drying). Addition of zinc on Cu/Zn/Zr again decreased the surface concentration of copper (from 4.0 to 2.8\u2009at% on Zn/Cu/Zn/Zr).Taking a closer look at the high resolution XPS results, the Zn 2p3/2 region (Fig. 6a) could be deconvoluted to three different species of Zn2+. These correspond to ZnO-like species (1021.5\u2009eV) and Zn mixed state denoted a (1023.1\u2009eV) and Zn mixed state denoted b (1024.4\u2009eV). The presence of metallic zinc can be discarded due to the absence of a peak at a slightly lower binding energy (at \u2053 1021\u2009eV [66]). The highest fraction of zinc in ZnO-like species was in the Zn/Cu/Zr sample (over 85%). The Cu 2p3/2 region (Fig. 6b) showed the presence of Cu metal (Cu0) and Cu2+ in all Cu-containing samples. Although the samples had been reduced before the XPS measurements, sample transfer through air to XPS had evidently been sufficient to oxidize part of the Cu0 to Cu2+.The reducibility of CuO was studied by H2-TPR (hydrogen temperature-programmed reduction). The results are shown in \nFig. 7. The reduction profiles for Zr and Zn/Zr samples are also shown in Fig. 7 as a reference. The Zr sample showed a very small and broad reduction peak with a maximum at 860\u2009K that can be related to the formation of surface oxygen vacancies on the support [67,68]. On the Zn/Zr sample, a single reduction peak was found at 815\u2009K, which can be associated with the partial reduction of the support and/or with the partial reduction of ZnO to metallic Zn [69,70].The Cu-containing samples showed various reduction peaks in the range of 360\u2013510\u2009K, depending on the order in which Cu and Zn were incorporated into the support. Reduction of CuO in the Cu/Zr sample generated up to three reduction peaks: two overlapping peaks with maxima at 405 and 418\u2009K and a third peak with a maximum at 477\u2009K. The double reduction peak at relatively low temperatures (frequently named the \u03b1 and \u03b2 peaks [71\u201374]) has been previously reported in samples with relatively low CuO loadings of approximately 5\u2009wt%. In accordance with these authors, well-dispersed Cu2+ species are first reduced to Cu+ (\u03b1-peak) and subsequently to Cu0 (\u03b2-peak). The third peak at higher temperature can be assigned to the reduction of more poorly dispersed bulk-like CuO particles or CuO particles with a stronger metal-support interaction [72,73,75].The deposition of Zn on top of Cu/Zr (Zn/Cu/Zr sample) led to a partial improvement in the reducibility of CuO. The previously reported \u03b1 and \u03b2 peaks in the Cu/Zr sample merged into a single reduction peak at 418\u2009K, while the reduction at 477\u2009K in Cu/Zr shifted to a lower temperature. This latter observation suggests that the addition of Zn after Cu improved the dispersion of bulk CuO particles [19,76]. Surprisingly, when zinc was added by ALD before copper impregnation (Cu/Zn/Zr sample), higher reduction temperatures than in the Zn/Cu/Zr sample were observed with two reduction peaks at ca. 448 and 500\u2009K, the latter presumably generated by the reduction of larger CuO particles. The positive effect of zinc addition after copper impregnation on the reducibility of bulk CuO is also observable in the Zn/Cu/Zn/Zr sample with a shift of the peak maximum from 500 to 475\u2009K.The CO2-TPD (carbon dioxide temperature-programmed desorption) profiles of pure Zr and the self-made Cu-Zn on Zr samples are displayed in \nFig. 8. Up to four different desorption peaks or desorption domains can be identified. The first two peaks (at ca. 380 and 445\u2009K) can be assigned to weakly basic sites with different gas\u2013solid interactions, while peaks at approximately 565 and 700\u2009K can be attributed to moderate and strong basic sites, respectively [77,78]. Accordingly, the desorption profile of pure Zr showed the presence of a significant amount of desorbed CO2 in the range of 340\u2013600\u2009K with a main desorption peak at 380\u2009K (most likely bicarbonate species on surface -OH groups) and a less significant and broader peak at 445\u2009K. The addition of Cu and/or Zn resulted in the appearance of new desorption peaks at higher temperatures while the main peak at approximately 380\u2009K became broader and less intense.The amount of desorbed CO2 (Table 1) followed the order (from higher to lower): Zn/Cu/Zr >\u2009Zn/Cu/Zn/Zr >\u2009Zn/Zr >\u2009Zr >\u2009Cu/Zn/Zr >\u2009Cu/Zr ranging from 115 to 63\u2009\u00b5mol of CO2 per gram of catalyst. In general, ZnO played an important role for the adsorption of CO2, especially when it was introduced after copper. The deposition of ZnO on Cu/Zr sample (Zn/Cu/Zr sample) increased the total amount of adsorbed CO2 from 63 to 115 \u03bcmol CO2 gcat\n\u22121 (\u223c80% higher) while the deposition of ZnO on Cu/Zn/Zr (Zn/Cu/Zn/Zr sample) increased the adsorbed CO2 from 85 to 109 \u03bcmol CO2 gcat\n\u22121 (\u223c30% higher). This trend shows the benefits of the zinc-after-copper pair in promoting the adsorption capacity of the catalyst and the advantages of adding the Zn atoms after Cu. When the results were expressed as molecules of CO2 per Zn atom (Table 1), Zn/Cu/Zr was still a superior catalyst (0.67 molecules of CO2 per Zn atom); however, the Zn/Cu/Zn/Zr sample performed worse, turning out to be the catalyst with the lowest capacity among the Zn-containing catalysts (0.3 molecules of CO2 per Zn atom). These results showed that Zn has an essential role in the adsorption of CO2 but the amount of CO2 and the Zn loading are not linearly correlated, the Zn/Cu/Zr being the preferred configuration to maximize the amount of adsorbed CO2.To understand the CO2 and H2 adsorption of Cu and Zn on ZrO2, the surface species were monitored during three consecutive adsorption cycles of CO2 and H2 by in situ DRIFTS. To follow the catalyst surface with the experiment, representative spectra of the first and third cycles of adsorption are depicted in \nFig. 9 in the fingerprint region (1700\u20131200\u2009cm\u22121). The results are shown for the studied catalysts at three temperatures (450, 500 and 550\u2009K). The selected spectra in Fig. 9 show the surface species after either 5 or 10\u2009min of a certain gas flow (CO2, H2 or Ar). In the supplementary material (Figs. S3-S7), monitoring with respect to time on stream (first five minutes of exposure to CO2 and H2) of the surface species during the first cycle of CO2 and H2 adsorption is displayed. An example of the evolution of several MS signals during the DRIFTS experiment carried out at 500\u2009K with the Zn/Cu/Zr sample is displayed in Fig. S8.The cyclic adsorption on the Zr sample (Fig. 9a) was considered to serve as a reference for the CO2 and H2 adsorption capabilities of the support material at various temperatures. After 10\u2009min of CO2 flow, the surface showed the presence of bicarbonate species, HCO3\n- at ca. 1625, 1427 and 1220\u2009cm\u22121 and bidentate carbonates CO3\n2- at ca. 1560\u20131530, and 1330\u2009cm\u22121, with the band positions exhibiting good agreement with the literature [79\u201382]. According to the literature, the presence of terminal -OH groups is required for the formation of bicarbonate species, while carbonates (either monodentate, bidentate or polydentate) require the presence of coordinately unsaturated (c.u.s.) Zr4+ and O2- sites [79,80,82]. Experimentally, the highest intensity for bicarbonate species was observed at the lowest temperature (450\u2009K), and the intensity decreased with increasing temperature, especially from 450 to 500\u2009K. In contrast, the intensity of bidentate carbonate species was not significantly affected by the temperature of the experiment. The spectrum after one minute of CO2 flow (Fig. S3) did not differ from the spectrum after 10\u2009min, which indicated a rapid saturation of the surface with carbonate and bicarbonate species. After switching the gas flow from CO2 to Ar, bicarbonate species vanished almost completely during 10\u2009min of Ar purge. At 450\u2009K, vibrational bands at ca. 1564 and 1330\u2009cm\u22121 corresponding to bidentate carbonate species remained slightly visible, which suggested a stronger adsorption of these species to ZrO2 compared to bicarbonate species. A similar observation was made by Kouva and coworkers [79] when studying the adsorption of CO2 on ZrO2 by DRIFTS in the range of 373\u2013673\u2009K. Both the first and third cycles of adsorption yielded comparable spectra with no apparent accumulation of any species throughout the experiment.The incorporation of Zn atoms onto ZrO2 by ALD (Zn/Zr sample, Fig. 9b) clearly modified the CO2 adsorption capacity of ZrO2. After 10\u2009min of CO2 exposure (either during the first or third cycle of adsorption), a crowded carbonate region (1600\u20131300\u2009cm\u22121) was observed. The vibrational band at 1220\u2009cm\u22121 during the CO2 flow indicated the presence of bicarbonate species, most likely on unsaturated Zr or Zn sites [83,84]. After 10\u2009min of Ar purge after the first cycle of CO2 adsorption, the spectra still displayed a busy carbonate region, indicating that Zn or the Zn-ZrO2 interface can store CO2-related species (especially carbonates at ca. 1535\u2009cm\u22121) with a greater adsorption strength than ZrO2. The switch from Ar to H2 flow led to the formation of new species such as formates (*HCOO-) located at ca. 2966, 2873 (Fig. S4), 1575, 1379 and 1365\u2009cm\u22121 (Fig. 9b), while the intensity of carbonate species (at ca. 1535\u2009cm\u22121) decreased in parallel. The formation of formate species was already visible after 1\u2009min of H2 flow (Fig. S4), which indicated a speedy hydrogenation of carbonates to formates on a Zn/Zr sample. In addition, more formates were clearly detected with increasing experimental temperature and number of adsorption cycles. This indicated: i) the ability of the Zn or Zn-ZrO2 interface to accumulate carbonates during the CO2 flow and to further convert them to formates, and ii) the high stability of formates on the Zn/Zr sample even at high temperature since they did not disappear or further react between cycles.The computational and experimentally observed infrared vibrational frequencies for formate species are displayed in \nTable 3. The frequencies for formate species were computed on different model systems such as m-ZrO2 (111), ZnO/m-ZrO2 (111), Cu (111), and Cu (110). The model surfaces used in DFT calculations are displayed in Fig. S9. Additionally, the computed IR frequencies for possible intermediates, such as formic acid *HCOOH, carboxyl *COOH and dioxymethylene *H2COO, which are intermediate species in methanol formation [19,85], are given in Figs. S10-S13. According to the DFT calculations, formate species produced four different bands on ZnO/ZrO2 in the 3000\u20131200\u2009cm\u22121 region that correspond to different functional groups and types of vibration. The bands were located at 1326, 1358, 1535 and 2961\u2009cm\u22121, and they correspond to symmetric stretch \u03c5s(O-C-O), bending \u03b4(C-H), asymmetric stretch \u03c5as(O-C-O) and stretching \u03c5(C-H), respectively. The experimental band at ca. 2970\u2009cm\u22121 can be attributed to the combination of C-H bending and asymmetric O-C-O stretching modes [86].The first cycle of CO2 adsorption over the Cu/Zr sample (Fig. 9c) led to a less crowded fingerprint region compared to Zn/Zr sample. As for the Zr sample, the main species detected were bicarbonates (at ca. 1625, 1430 and 1220\u2009cm\u22121) and bidentate carbonates (at ca. 1557 and 1330\u2009cm\u22121): the bicarbonates were rapidly removed after 10\u2009min of Ar purge. Additionally, according to the band observed at 1535\u2009cm\u22121, carbonates were present on the catalyst surface. Similar to the Zn/Zr sample (Fig. 9b), formates were formed during exposure to H2; however, the position of the band of the O-C-O asymmetric vibration differed (1575\u2009cm\u22121 on the Zn/Zr sample and 1569\u2009cm\u22121 on the Cu/Zr sample). A lower IR frequency for formate species on Cu (111 and 110) compared to that on a Zn/Zr sample was also predicted by DFT calculations (Cu(111) vs. ZnO/m-ZrO2, Table 3). After three consecutive cycles of CO2 and H2 adsorption, formates reached the highest intensity at 500\u2009K, followed by 550 and 450\u2009K. During the 3rd cycle of H2 flow on the Cu/Zr sample, the intensity of formate species slightly decreased from 5 to 10\u2009min on stream, which evidenced the strong adsorption of formate species on the Cu/Zr sample. Herein, by using a combined experimental and computational (DFT calculations) approach, Larmier and coworkers [85] found that on a Cu/ZrO2 catalyst, formate species exhibit a low Gibbs free energy, which dictates their strong adsorption on the catalyst surface.With Zn/Cu/Zr (Fig. 9d) and Cu/Zn/Zr (Fig. 9e) samples, the detected species and the discussed trends with respect to temperature and CO2 and H2 cycles displayed a combination of the results observed based on Zn/Zr and Cu/Zr samples. Thus, mainly due to the presence of Zn in both Zn/Cu/Zr and Cu/Zn/Zr samples, the fingerprint region was highly occupied by CO2-related species after exposing the catalysts to CO2. Formate was the main species detected after completing the cycles of adsorption of CO2 and H2, and no further hydrogenated species were observed (such as formic acid, *HCOOH, dioxymethylene, *H2COO or methoxy, *H3CO).Based on the DRIFTS results discussed above, carbonates were present on all samples, and formate was the prevalent hydrogenated species in the selected experimental conditions. To compare the ability of each catalyst to hydrogenate carbonates into formates, a \u201cformates to carbonates intensity ratio\u201d was calculated and followed throughout the first to third cycles of adsorption for the various temperatures (Fig. 9f). Bands at 1535\u20131533 and 1575\u20131569\u2009cm\u22121 were chosen as representative frequencies for carbonates and formates, respectively. Comparing the evolution of the ratio by temperature, clear trends could be observed at 500 and 550\u2009K, while at 450\u2009K, the ratio remained roughly constant at ca. 1 throughout cycles, which indicated a poor reduction of carbonates to formates at this temperature. At 500 and 550\u2009K, the ratio was approximately 1.5 and 1.3, respectively, which pointed out the positive effect of temperature in the transformation of carbonates to formates. In all cases, the increase of the ratio originated from the concurrent decrease in the intensity of carbonate and the increase in the intensity of formate. At 500\u2009K, the Zn/Zr sample yielded the highest ratios at the end of each cycle which indicated the crucial role of the ZnO-ZrO2 interface to form formate from carbonate. Especially outstanding was the increase of the ratio for the Zn/Cu/Zr sample at 500 and 550\u2009K when hydrogen was introduced into the system (increase highlighted with solid black in lines in Fig. 9f). This behavior revealed the good ability of the Zn/Cu/Zr configuration in the hydrogenation of carbonate to formate. In addition, the Zn/Cu/Zr was the unique sample that exhibited a rising ratio during H2 flow under any condition. This indicated that the Zn/Cu/Zr configuration was able to remain active in the transformation of carbonate to formate during the 10\u2009min of H2 flow, presumably due to a greater ability to keep CO2 adsorbed, as observed in Fig. 8.The catalytic performance of Cu-Zn on zirconia catalysts is illustrated in \nFig. 10 for the three selected reaction temperatures: 450, 500 and 550\u2009K. Fig. 10a shows the CO2 conversion, Fig. 10b-c show the production of CH3OH expressed as the space-time yield in mmolCH3OH gcat\n\u22121 h\u22121 and in mmolCH3OH gCu\n\u22121 h\u22121, respectively, and Fig. 10d shows the product selectivity for CH3OH, CO and CH4. Among the three reaction temperatures tested, all the catalysts showed activity at 500 and 550\u2009K, while there was no measurable activity at 450\u2009K. Greater CO2 conversion values were achieved for all the samples with increasing reaction temperature. The CO2 conversion values attained with both Cu- and Zn-containing samples, i.e., Zn/Cu/Zn/Zr, Zn/Cu/Zr and Cu/Zn/Zr, were higher than those achieved with Zn/Zr, Cu/Zr and Zr samples, which evidenced the importance of the Cu-Zn interactions in the CO2 hydrogenation reaction in accordance with the literature [18,19,87]. Interestingly, the addition of Zn on top of Cu (Zn/Cu/Zr and Zn/Cu/Zn/Zr samples) promoted CO2 conversion, and the most remarkable improvement occurred at 550\u2009K, with conversion values of approximately 9% for Zn/Cu/ZrO2 and 6.5% for Zn/Cu/Zn/ZrO2. Under the same operating conditions, the other samples (Cu/Zn/Zr, Cu/Zr and Zn/Zr) yielded CO2 conversion values between 2% and 4%.Regarding the product selectivity (Fig. 10d), temperature had a remarkable effect on the product distribution. At 500\u2009K, the prevailing product was methanol with all catalyst combinations (except on the Zr sample, which only produced CO). Thus, the highest CH3OH selectivity was achieved at 500\u2009K with Zn/Cu/Zn/Zr sample (close to 80%) followed by Zn/Cu/Zr (71%) and Cu/Zn/Zr (68%) samples. The selectivity toward CO increased significantly when increasing the temperature from 500 to 550\u2009K, according to the endothermicity of the reverse-WGS reaction (\u0394H\u00b0298k = + 41\u2009kJ\u2009mol\u22121) [88]. Advantageously, the Zn/Cu/Zr sample did not produce any methane under any of the conditions, in contrast to the Zn/Cu/Zn/Zr and Cu/Zn/Zr samples that produced methane at both 500 and 550\u2009K with selectivities around 3\u20136%.In terms of methanol production (Fig. 10b-c), the most efficient catalysts were Zn/Cu/Zn/Zr and Zn/Cu/Zr, with similar production rates at 500\u2009K (\u2053 1.9\u2009mmolCH3OH gcat\n\u22121 h\u22121) and 550\u2009K (\u2053 3.8\u2009mmolCH3OH gcat\n\u22121 h\u22121). For Cu/Zn/Zr sample, the methanol production was lower (1.3 and 2.9\u2009mmolCH3OH gcat\n\u22121 h\u22121 at 500 and 550\u2009K, respectively). Moreover, Zn/Cu/Zr and Zn/Cu/Zn/Zr samples showed a space-time yield of methanol approximately three times higher than that achieved with Cu/Zr and Zn/Zr samples. When the methanol production was referred to in terms of grams of copper (Fig. 10c), the Zn/Cu/Zr sample produced 165 and 340\u2009mmolCH3OH gCu\n\u22121 h\u22121 at 500 and 550\u2009K (similar rates were attained with the Zn/Cu/Zn/Zr sample), while the Cu/Zn/Zr sample produced 107 and 241\u2009mmolCH3OH gCu\n\u22121 h\u22121 under the same operating conditions. Therefore, the activity results highlight the potential interest of adding Zn by atomic layer deposition after Cu to promote methanol production catalysis.The presence of zinc oxide overlayers on top of copper particles in industrial-type Cu/ZnO/Al2O3 catalysts during the hydrogenation of carbon oxides to methanol has been demonstrated in diverse studies during the last decade [11,89,90]. Schott et al. [91] reported the unique properties of ZnO layers on the surface of copper particles due to the partial reduction of ZnO to a less strongly oxidized Zn\u03b4+ state under the reducing conditions of methanol synthesis. Similarly, Behrens and collaborators [11] demonstrated that the Cu-ZnO synergy lies in their strong metal support interaction leading to partial coverage of the copper surface with ZnOx under reducing (activation) conditions. Thus, ZnO nanoparticles dispersed on top of copper are special entities that could show particular physico-chemical properties not observed for bulk oxides [38,89,92].As previously described in the introduction there seems to be agreement among scholars in setting the optimal coverage of copper particles by ZnOx for a higher catalytic activity at relatively low values of 15\u201320% monolayer (ML) [19,38,40]. Nevertheless, higher coverages (\u03b8Zn = 0.47) have also been reported as optimal values [41]. Although this difference is not fully understood, all of these studies agree that larger coverages of ZnO may negatively affect the catalyst activity in terms of methanol production [19,38,40,41]. Furthermore, the computational studies carried out by Kuld and coworkers [41] predicted a greater TOF of methanol for ZnO particles smaller than 7\u2009nm, sizes that can be easily accomplished by ALD. Based on our results, we observed that zinc added by one ALD cycle yielded \u205315% ML of ZnO, while we speculate that Cu formed larger structures from nanoparticles to small clusters. Based on this remark, the addition of copper by impregnation after zinc ALD might have disabled some of the ZnO species and prevented their assistance in any further reaction. This speculation is supported by the similar methanol production rates observed for Zn/Cu/Zr and Zn/Cu/Zn/Zr samples.To place these arguments and calculations on a more solid basis, we have compared the results included in this manuscript with the results reported in the literature for earlier studies (\nTable 4). In our work, with the Zn/Cu/Zr sample, we report rates of 165 and 340\u2009mmolCH3OH gCu\n\u22121h\u22121 at 500 and 550\u2009K, respectively, while 169 and 321\u2009mmolCH3OH gCu\n\u22121h\u22121 were produced with the Zn/Cu/Zn/Zr sample under the same temperature conditions. Recently, Saedy et al. [48] accomplished a methanol productivity of \u2053 56\u2009mmolCH3OH gCu\n\u22121h\u22121 at 523\u2009K and 3.0\u2009MPa with a ZnO/Cu/Al2O3 catalyst (Zn added by PCVD) and a Zn/Cu atomic ratio of 0.5. Higher Zn/Cu atomic ratios (up to 1.64) did not significantly affect the production of methanol. It is also important to highlight that we achieved a similar areal number density and % of ML of ZnO for the Zn/Cu/Zr sample to those reported by Saedy et al. (\u2053 1.5 Zn atoms/nm2, \u2053 13% ML). In a similar study, Gao et al.\n[27] synthesized ALD ZnO-coated Cu/SiO2 catalysts with various exposure times and ALD cycles and tested them in the hydrogenation of carbon dioxide. Among the prepared catalysts, the most active catalyst (exposure time of 30\u2009s to the Zn precursor and one ALD cycle) was more selective toward carbon monoxide (96\u2009mmolCO gCu\n\u22121h\u22121) than toward methanol (10.6\u2009mmolCH3OH gCu\n\u22121h\u22121) at 523\u2009K and 4.0\u2009MPa. However, for that particular ZnO/Cu/SiO2 catalyst, the areal number density was significantly lower than the value that we report in our manuscript (\u2053 0.6 Zn atoms/nm2, \u2053 5% ML).In general, the data included in Table 4 highlight the good performance of the catalysts prepared and tested in the present work for carbon dioxide hydrogenation to methanol. It is worth mentioning that this level of activity was achieved with relatively low Cu and Zn metal loadings (\u20531\u20132\u2009wt%) which led to considerably high methanol production rates when they were expressed per gram of copper. This range of metal contents seemed to be relatively efficient to achieve a good and even distribution of both copper and zinc and to produce a significant amount of active ZnO-Cu sites. As a future challenge, it would be worth investigating whether it is viable to obtain similar methanol production rates per gram of copper with higher metal contents. In this context, atomic layer deposition (ALD) can be an outstanding synthesis method for the scale up of catalysts toward higher metal loadings.Tuning the interaction of zinc and copper must be considered an important parameter to control the catalytic performance toward methanol formation. In this work, by alternating the order in which metal (copper) and promoter (zinc) were added to the catalyst, a series of catalysts with various metal-promoter-support configurations were synthesized. The order in which the zinc promoter was introduced onto the catalyst by atomic layer deposition (ALD) compared to the active copper metal by impregnation affected the catalytic activity. Zinc ALD after copper impregnation (zinc-on-copper) yielded higher CO2 conversion and methanol production rates than copper-on-zinc, although the overall copper and zinc loadings were similar. Advantageously, unlike the other catalysts, the zinc-on-copper zirconia catalyst (Zn/Cu/Zr sample) did not produce any methane with high methanol production rates under the tested operating conditions.Infrared studies of cyclic adsorption of CO2 and H2 revealed that zinc ALD on impregnated copper accumulated carbonates and bicarbonates (CO3\n2-, HCO3\n-) exceptionally well during the carbon dioxide feed and transformed them into formate species (*HCOO) during the hydrogen feed. Together with the higher CO2 conversion and methanol production rate achieved with the Zn/Cu/Zr sample, this suggests that the catalytic activity to some extent relies on the ability of the catalyst to transform carbonates to formates. The DFT calculations accurately predicted the band position for formate species on different model surfaces (i.e, ZnO/ZrO2, Cu(111) and Cu(110)) compared to the experimental bands observed by DRIFTS. The formate pathway was the favored mechanistic route of carbon dioxide hydrogenation to methanol under the selected experimental conditions. In addition, the CO2 temperature programmed desorption analyses showed the great capacity of the zinc-copper on zirconia catalyst for the adsorption of CO2. When evaluating the molecules of CO2 adsorbed per atom of zinc, the zinc-on-copper configuration adsorbed more CO2 molecules than the copper-on-zinc configuration (0.67 versus 0.45 molecules CO2 per zinc atom). Additionally, according to TPR studies, the zinc deposited after copper impregnation improved the homogeneity of copper oxide species and the reducibility of the bulk CuO.Overall, this work provides insight into the significance of the zinc oxide/copper/zirconia interactions for selective hydrogenation of carbon dioxide to methanol and highlights the potential of atomic layer deposition (ALD) in the synthesis of atomically dispersed metal catalysts for an efficient methanol synthesis.\nAitor Arandia: Conceptualization, Investigation, Methodology, Formal analysis, Validation, Visualization, Supervision, Writing \u2013 original draft, Writing \u2013 review & editing. Jihong Yim: Investigation, Methodology, Formal analysis, Validation, Writing \u2013 review & editing. Hassaan Warraich: Investigation, Formal analysis, Methodology. Emilia Lepp\u00e4kangas: Investigation, Methodology. Ren\u00e9 Bes: Investigation, Formal analysis, Writing \u2013 review & editing. Aku Lempelto: Investigation, Software, Formal analysis, Visualization, Writing \u2013 review & editing. Lars Gell: Investigation, Software, Formal analysis, Visualization, Writing \u2013 review & editing. Hua Jiang: Investigation. Kristoffer Meinander: Investigation, Formal analysis, Visualization, Writing \u2013 review & editing. Tiia Viinikainen: Methodology, Writing \u2013 review & editing. Karoliina Honkala: Conceptualization, Writing \u2013 review & editing, Funding acquisition. Simo Huotari: Investigation, Formal analysis, Visualization, Writing \u2013 review & editing. Riikka L. Puurunen: Conceptualization, Methodology, Resources, Writing \u2013 review & editing, Supervision, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The work at Aalto University has been financially supported by the Academy of Finland (COOLCAT consortium, decision no. 329977 and 329978; ALDI consortium, decision no. 331082). This work made use of Aalto University Bioeconomy, OtaNano and RawMatters infrastructure. Hannu Revitzer (Aalto University) is thanked for the ICP-OES analysis, Aalto workshop people (especially Seppo J\u00e4\u00e4skel\u00e4inen) for working on the reactor modifications. The DFT calculations were made possible by computational resources provided by the CSC \u2014 IT Center for Science, Espoo, Finland (https://www.csc.fi/en/) and computer capacity from the Finnish Grid and Cloud Infrastructure (urn:nbn:fi:research-infras-2016072533). The University of Helsinki acknowledges support from Academy of Finland (project 295696) as well as ESRF for beamtime and Blanka Detlefs and Christoph Sahle for expert support. Preliminary XANES measurements were performed using the Helsinki Center for X-ray Spectroscopy Hel-XAS instrument under the proposal number 2021\u20130011.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.apcatb.2022.122046.\n\n\n\nSupplementary material.\n\n\n\n.", "descript": "\n                  The development of active catalysts for carbon dioxide (CO2) hydrogenation to methanol is intimately related to the creation of effective metal-oxide interfaces. In this work, we investigated how the order of addition of copper and zinc on zirconia influences the catalytic properties, the catalytic activity and selectivity toward methanol. Regarding the carbon dioxide conversion and methanol production, the catalysts on which the promoter (zinc) was atomically deposited after copper impregnation (i.e., ZnO/Cu/ZrO2 and ZnO/Cu/ZnO/ZrO2) were superior catalysts compared to the reverse copper-after-zinc catalyst (Cu/ZnO/ZrO2). Temperature-programmed experiments and in situ diffuse reflectance infrared Fourier transform-spectroscopy (DRIFTS) experiments allowed us to elucidate the benefits of the zinc-after-copper pair to store CO2 as carbonate species and further convert them into formate species, key intermediates in the formation of methanol. This research provides insights into the potential of atomic layer deposition in the development of tailored heterogeneous catalysts for efficient CO2 valorization to methanol.\n               "}
{"full_text": "Metal-organic framework, denoted as MOF, is defined by the International Union of Pure and Applied Chemistry (IUPAC) as \u201ca coordination network with organic ligands containing potential voids\u201d [1]. Since the early 1990s, after the first scientific reports on the development of a new class of porous materials, there has been strong interest in this topic. Almost 30 years of intense research has led to numerous potential applications of MOFs in a wide variety of fields including gas adsorption, separation, catalysis, photocatalysis and bio-sensing. Intensive studies on MOF applications have also included their application in fuel cells and supercapacitors [2\u20139]. Several synthesis routes of metal-organic networks have been developed over the years. The most utilised are conventional solvothermal and non-solvothermal, microwave-assisted and mechanochemical methods [2,4,10]. Numerous scientific papers report on both solvothermal and non-solvothermal syntheses of MOFs, giving the exact synthesis procedures, and the changes of MOFs\u2019 parameters by the modification of synthesis conditions can be found in the literature. Several MOFs have been synthesised using non-solvothermal methods which require the selection of metal precursors, organic linkers and solvents, as well as the appropriate synthesis temperature. The remarkable success of MOFs in a wide range of applications has pushed scientists to use MOF materials as precursors to obtain catalytic materials with unprecedented properties. However, despite the fact that the recent development in synthesis of metal organic frameworks pushes the limits of the chemical and mechanical resistance of those materials, they are used in a wide range of industrial applications based on catalysis. The next milestone in the application of metal organic frameworks in industry may be not only further improvements in the chemical and mechanical endurance of those materials, but also their structuring into monolith-like, short channel structures membranes or arranged structures which guarantees high heat and mass transport properties. Since the remarkable success in development of structured catalysts in industry-based heterogeneous catalysts including gas exhaust abatement in the automotive sector and stationary source abatement, water gas shift, combustion and NOx abatement [11], the structuring of MOFs into structured catalysts seems to be a natural step forward in their evolution.Several works have recently been published describing the ways of the preparation of structured materials based on metal organic frameworks [12\u201318]. In the work written by Chen et al. [18], various attempts to produce composite HKUST/Fe3O4 materials in different bodies like pellets, films and foams are described. The authors have developed a method of shaping of composite HKUST/Fe3O4 materials by using carboxymethylcellulose as a binder. By using freeze-drying or gel-induced surface hardening, various foam-like or thin films with high porosity properties have been developed. A complementary method for the preparation of MOF-based foams is described in the work published by Garai et al. [19], where the shaping of metal organic frameworks by transferring them into areogel or xerogel and further solvent removal was proposed. However, despite the versatility of proposed method, the use of foams derived by the aerogel and xerogel method is limited, due to a high fragility of derived structures. In the deposition of metal organic frameworks on the metallic surfaces, much attention has been paid to the preparation of electrodes for lithium-ion batteries [20]. The deposition of metal organic frameworks based on zeolite-imidazole frameworks was performed by annealing treatment. The porous zinc-cobalt oxide porous plates prepared in this way revealed remarkable, high reversible properties as anode materials and considerable lithium storage capacities.Despite the fact that the metal organic framework materials demonstrate great catalytic properties in many catalytic reactions including catalytic oxidation [21\u201325], selective catalytic reduction [26], alkylation, transesterification [10], water gas shift and conversion of methane to fuels, their heat and mass transfer properties may be successfully tuned up by either their direct shaping into structured catalysts or their deposition on existing carriers. Although several works describing the use of three-dimensional printing of metal organic frameworks to monoliths have recently been published [17], literature reports describing deposition of MOFs on supported carriers are scarce.Structural reactors owe their significant success mainly to their wide use in the automotive and energy industries, where the ceramic or metal monoliths are commonly in use for oxidation and selective catalytic reduction reactions [27]. The catalytic oxidation of hydrocarbons is one of the most important reactions for the conversion of hydrocarbons to obtain valuable products. Over the numerous catalytic reactions, the oxidation of cyclic hydrocarbons such as cyclohexane or cyclohexene results in the formation of value-added products that can be further used in fine chemical synthesis. The exemplary oxidation of cyclohexene with H2O2 may be used as an alternative method for the synthesis of adipic acid, which is further used in production of Nylon-66 [23]. Additionally, the oxidation of cyclohexene may also result in the formation of epoxides and unsaturated ketones and alcohols which are valuable products in organic syntheses and the fragrance industry. Recently, the catalytic oxidation of cyclohexenene to the mixture of oxygen-containing products has been reported for SBA-15 [28], core shell-structures [24] and MIL-101 [21] or modified Ni-MOF-74 catalyst [29]. Although literature reports provide information on the successful use of metal organic frameworks on cyclohexene catalytic oxidation instead of conventional mesoporous catalysts, a common feature of the work is the use of powder catalysts which practically eliminates their wider application. The main reason for that is the necessity of additional mixture/catalyst filtration to receive products instead of simple structured catalyst removal from the batch reactor.In this work, we present an optimised method for the preparation of composite metal organic frameworks for structured catalysts based on metallic plates, woven gauzes and metallic foams as catalysts for aerobic oxidation of cyclohexene. The choice of those types of structures is not accidental, as they are used as catalyst supports: metal monoliths for oxidation and reduction reactions, meshes for oxidation/separation processes and foams for oxidation reactions. The prepared structured catalysts with deposited thin metal organic frameworks have revealed considerable surface areas and remarkable, good adhesion parameters. The catalytic activity tests have proven that the composite metal organic framework catalysts may be successfully used in aerobic oxidation of cyclohexene to produce value-added fine chemicals.All chemicals used in this study were reagent grade and are commercially available. They include nickel acetate tetrahydrate, cobalt acetate tetrahydrate, zinc acetate dihydrate, nickel nitrate hexahydrate, cobalt nitrate hexahydrate, zinc nitrate hexahydrate, 2,5-dihydroxyterephthalic acid (DHTP), all from Sigma-Aldrich, and methylene chloride, n-hexane, N,N-dimethylformamide (DMF), n-propanol, from Chempur Poland.The synthesis protocol used in this study consisted of three steps: support pre-treatment, in situ MOF deposition and material activation. Structured supports used in this study were FeCrAl plate (GoodFellow, 0.3\u00a0mm thick Fe 72.8%, Cr 22%, Al 5%, Y 0.1%, Zr 0.1%), steel woven gauzes (17.5 mesh/in., wire diameter 0.1\u00a0mm; Fe 73%, Cr 20%, Al 5%) and NiCr foams (Recemat BV; 27\u201333 ppi, estimated average pore diameter 0.6\u00a0mm, Ni 60\u201380%, Cr 15\u201340%, Fe 0.5%, Cu 0.1\u20130.3%).Prior to the deposition of MOF on to the structured carriers, the structures were cut into small pieces \u2013 FeCrAl plates 1\u00a0cm\u00a0\u00d7\u00a01\u00a0cm, FeCrAl gauze 1\u00a0cm\u00a0\u00d7\u00a01\u00a0cm, NiCr foams 1\u00a0cm\u00a0\u00d7\u00a01\u00a0cm \u2013 and subsequently cleaned in an ultrasound bath using acetone, n-propanol and distilled water to remove impurities. Subsequently, FeCrAl plates and wire gauzes were calcined at 1100\u00a0\u00b0C in a ventilated oven for 24\u00a0h to obtain a thin alumina layer. This procedure of FeCrAl alloy treatment was previously reported as enhancing further adhesion between alloy and deposited material [30].In the second step, the M(M\u00a0=\u00a0Zn; Ni; Co)--MOF-74 layers were deposited in situ by modifying the solvothermal method for powder synthesis recently reported in the literature [31,32]. The detailed synthesis conditions are summarised in Table 1\n.The first layer deposition of Zn-MOF-74 was performed from Solution I by using zinc acetate as a metal precursor. After dissolution of the appropriate amounts (see Table 1) of metal salt and 2,5-dihydroxyterephtalic acid (DHTP) in N,N-dimethylformamide DMF, the metal salt solution was added to the DHTP solution dropwise to avoid precipitation. The resulting solution was then transferred to Teflon liners with structured carriers previously suspended on scaffolding. The as prepared stainless-steel bombs with Teflon vessels were tightly capped and placed in oven at 100\u00a0\u00b0C for 20\u00a0h. The resulting structured carriers with deposited MOF layers and non-deposited MOF crystals were washed using the sequence proposed elsewhere [33]: methyl chloride three times, and n-hexane three times. The resulting materials were then dried at room temperature and activated in a vacuum drier at 180\u00a0\u00b0C for 6\u00a0h. The double and triple deposition of Zn-MOF-74 was performed by changing synthesis solution I to synthesis solution II with zinc nitrate as a metal precursor.The general procedure for deposition of Co-MOF-74 and Ni-MOF-74 was performed as for deposition of Zn-MOF-74, with the difference that the appropriate metal nitrate (Co or Ni) was used as a metal precursor in all three-layer deposition steps.The crystallinity of prepared materials was determined by XRD analyses using an X'Pert Pro MPD (PANalytical) diffractometer with CuK\u03b1 radiation at 30\u00a0mA and 40\u00a0kV. The diffraction patterns were collected in the range of 5\u201380\u00b0 2\u03b8 with a 0.033\u00b0 step for 12\u00a0min. The determination of crystallinity M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 layers deposited on FeCrAl plates was determined by means of Grazing Incidence X-Ray Diffraction analysis (GIXRD). Analyses were performed only for M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 layers deposited on FeCrAl plates due to the GIXRD method limitations. The GIXRD analyses were performed in 5\u201375\u00b0 2\u03b8 range with a 0.033\u00b0 and constant omega angle 1\u00b0.The morphology of prepared structured catalysts was determined by using a Nova Nano SEM 300 FEI Company scanning electron microscope for high-quality magnification imaging. To enhance the visibility of the structure of and the distribution of the Me-MOF-74 layers on structured carriers, the obtained materials were pseudo-coloured using Fiji software. The exact colours of LUT's were determined of an activated MOF samples by using AvaSpec-ULS3648 High-resolution spectrometer equipped with a high-temperature reflection probe (FCR-7UV400-2-ME-HTX, 7\u00a0\u00d7\u00a0400\u00a0\u03bcm fibres, Avantes BV) and a Mikropack DH-2000-BAL Deuterium-Tungsten Halogen Light Source working in the 200\u20131000\u00a0nm spectral range. The exact colour of the prepared material was determined by AvaSoft 8 software with colour measurements extension (Avantes BV). The determined colours were presented using HEX and RGB values (Table 1).Kr and N2 sorption experiments were performed on ASAP 2020 (Micromeritics) for structured supports, powder samples and MOF layers deposited on FeCr plates and NiCr foams, respectively. Prior to analyses, the samples were outgassed at 250\u00a0\u00b0C for 12\u00a0h. The BET specific surface areas were calculated for p/p0 in the range of 0.06\u20130.2 and for Kr adsorption and p/p0\u00a0=\u00a00.06\u20130.2 for N2 adsorption experiments.The Me-MOF-74 layers deposited on FeCrAl plates were examined by X-ray Photoelectron Spectroscopy with an ESCA Prevac spectrometer equipped with a hemispheric XPS analyser of charged particles and AES analysers (VG Scienta R3000) and Mg/Al anticathodes. The sample charging effect was corrected using C 1s band at 248.8\u00a0eV.The prepared Me-MOF-74 samples were characterised by FTIR spectroscopy using two modes: ATR FTIR for non-deposited MOF crystals that were collected after in situ MOF deposition, and by in situ DRIFT for composite Me-MOF-74 samples deposited on FeCrAl plates. The ATR-FTIR studies were carried out using a Bruker Vertex 70v spectrometer equipped with Bruker Platinum ATR (diamond crystal), by averaging 128 scans in the range of 4000\u2013400\u00a0cm\u22121 with a 4\u00a0cm\u22121 resolution. The in situ DRIFT spectra were collected by using a Thermo Nicolet iS 10 equipped with MCT detector and Praying Mantis High Temperature Reaction Chamber with ZnSe windows (Harrick). The in situ experiments were performed on dehydrated at 110\u00a0\u00b0C for 1\u00a0h in He flow (AirProducts) catalysts samples. To avoid the presence of water and oxygen, the He line was equipped with an Agilent moisture/oxygen trap. The spectra were collected by averaging 128 scans with 4\u00a0cm\u22121 resolution and BaSO4 as a background.The FTIR sorption experiments by using CO (Linde) and CD3CN as probe molecules were performed by using a NICOLET iS 10 spectrometer. The spectra were taken in the 4000-650\u00a0cm\u22121 range with 4\u00a0cm\u22121 resolution by averaging 128 scans. Prior to the spectroscopic measurements, the obtained Me-MOF-74 crystals were pressed into the self-supporting wafers and activated under vacuum at 270\u00a0\u00b0C with 5\u00a0\u00b0C/min temperature ramp. The qualitative determination of the nature of the active sites in prepared MOF-74 samples was determined by low temperature (\u2212100\u00a0\u00b0C) carbon monoxide (Linde) and room temperature CD3CN (Sigma Aldrich) chemisorption. Prior to the chemisorption of probe molecules, the adsorbed gases were distilled by freeze and thaw cycles to remove impurities. The resulting spectra were presented as a substructured spectra after each portion of adsorbed probe molecule and activated sample as a background.To determine the nature and the chemical distribution of deposited metal organic frameworks on structured carriers, the \u03bcRaman mapping analyses were performed by using high resolution confocal Raman microscope - Witec Alpha 300 M+ equipped with three ZEISS lenses (x10, x50, x100), two diffraction gratings 600 and 1800, and two 633\u00a0nm and 488\u00a0nm with power of approximately 50 and 75\u00a0mW, respectively. The \u03bcRaman spectra were taken for FeCrAl plates due to the optical microscope limitations.The effectiveness and stability of the prepared structured metal organic framework materials was determined in two ways. The effectiveness of MOF-74 in situ deposition was determined by weighing the washed and activated composite materials before and after layering. The mechanical stability test was performed by ultrasound irradiation methods proposed recently in literature for structured catalysts [34\u201336]. In brief, the washed and activated structured catalysts were immersed in polypropylene jars filled with n-propanol and irradiated in a 40\u00a0kHz ultrasound bath (Ultrasonix proclean 0.7\u00a0M, 60\u00a0W). The weight loss was determined after 15\u00a0min of ultrasonic irradiation.Catalytic activity during the aerobic oxidation of cyclohexene was measured under atmospheric and 10\u00a0bar O2 pressure for powder samples and MOF deposited on NiCr foams as representative for structured catalysts. The aerobic oxidation of cyclohexene was measured under atmospheric conditions and were performed in glass reactor vessel equipped with a reflux condenser. In a typical experiment, the 50\u00a0mg of catalyst (for MOF/NiCr foams 50\u00a0mg of catalyst refers to the 50\u00a0mg of MOF deposited on NiCr foam) and 10\u00a0cm3 of cyclohexene were placed in the reactor and heated to 80\u00a0\u00b0C for 4\u00a0h under oxygen flow. The oxygen flow (Oxygen 5.0, Linde Gas) was controlled by Bronkhorst mass flowmeters and set to 20\u00a0ml/min. Prior to the reaction, the glass reactor was purged with molecular oxygen for 15\u00a0min with 20\u00a0ml/min flow. The experiments under 10\u00a0bar O2 pressure were performed in a Buchi Miniclave Stainless Steel reactor. The catalytic experiment procedure was similar to experiments at atmospheric pressure. The O2 pressure was set to 10\u00a0bar by using a Buchi manometer at the reactor vessel. Prior to the catalytic experiments, the pressure reactor was purged with molecular oxygen for 15\u00a0min.The catalytic reaction products were analysed by the method described in ref. [21], using a gas chromatograph (Thermo Scientific A Trace 1310) coupled with a single quadrupole mass spectrometer (ISQ) equipped with an RXi-5MS capillary column (Restek, USA, 30\u00a0m, 0.25\u00a0mm ID, 0.25\u00a0mm film thickness.). Prior to analysis, the reacting mixtures were thoroughly cooled down in an ice bath to avoid CH evaporation, and approx. 10\u00a0mg of PPh3 was added to reduce cyclohexenyl hydroperoxide to 2-cyclohexen-1-ol and avoid further mixture oxidation.The migration of metal (Zn, Ni, Co) from prepared MOF samples to the reaction mixture during the catalytic reaction was determined by atomic absorption spectrometry using a Thermo Scientific ICE3000 series AAS spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). To determine the metal content in post-reaction mixtures, the external standard method was used. The results were processed using Solaar 2.01 software. All standards and reagents were of trace analysis grade.The synthesis of metal organic frameworks may be performed in various conditions by using metal precursors and organic linkers, of which metal nitrates and acetates are commonly used [2]. Since the choice of the starting reagents for synthesis of MOF in powder form may influence the crystal size and the synthesis time, the application of the in situ crystal deposition over the metallic structures should consider crystal-surface interactions [37]. It followed from this that acetates and nitrates were natural choices due to their acidic properties in a liquid solution. The choice of the acetates and nitrates is dictated by their dual role as metal precursors and acidic environment generators. The acidic environment is favourable and commonly used in structured reactor preparation in metallic support pre-treatment [30]. It was previously reported that the use of an acidic environment induces the formation of thin alumina layer on FeCrAlloy material, which increases further adhesion of the deposited layer [38]. Another problem related to the nature of the precursor is that, while acetates can be used for synthesis of various MOF, their use for MOF-74 synthesis is limited for the preparation of Zn-MOF-74 though conventional synthesis and Ni- and Co-MOF-74 through dry-gel synthesis [39]. Based on available literature reports, we used zinc acetate as a starting point mixture in the optimisation of in situ synthesis. To monitor the acidity of the synthesis solutions, we performed measurements of pH before and after in situ solvothermal synthesis (Table 2\n). The acetate solutions\u2019 pH values before the synthesis are very close to neutral point, whereas nitrate-based precursor solutions are strongly acidic (pH\u00a0\u2248\u00a02.7). Despite the fact that the in situ synthesis of Zn-MOF-74 resulted in well crystallised MOF-74, as already been postulated in the literature [39,40], the amount of MOF-74 deposited on structured carriers was considerably low. Hence, for the double and triple synthesis of MOF layers on metallic supports, we used metal-nitrates as metal precursors. However, it has to be pointed out that the use of the metal nitrate as an MOF metal precursor at the first layer deposition did not result in either deposition of the MOF layer at the metallic carrier or formation of the Zn-MOF-74 crystals on the bottom of the reaction vessel. To confirm the crystallinity and the purity of obtained materials, PXRD for non-deposited powder MOF-74 (Fig. 1\n, left column) and GIXRD for MOF-74 deposited on FeCrAl plates (Fig. 1, right column) were performed. In all prepared materials, as well as for the non-deposited crystal phase and thin layer deposited on metallic carriers, the presence of Zn-MOF-74 (JCPDS 00-062-1198), Ni-MOF-74 (JCPDS 00-62-1029) and Co-MOF-74 (JCPDS 00-063-1147) structures without impurities [39,41,42] was confirmed. The use of GIXRD analysis allowed high quality diffraction patterns on MOF layers deposited on FeCrAl plates to be obtained. Despite the fact that the GIXRD measurement was performed at a low angle, we could still observe reflections at 25.6, 35.1, 37.8, 43.5, 52.6 (024) and 57.6\u00b0, which are characteristic of \u03b1-Al2O3 [43] (JCPD 04-005-4503) from FeCrAl support. The \u03b1-Al2O3 is the result of the FeCrAl support calcination at 1100\u00a0\u00b0C which enhances the adhesion of deposited MOF layers. The detailed phase analysis was previously reported in our previous paper [44] and also in GIXRD profile analysis in supporting information (Figs. S1-S2). It may be seen that the intensity of characteristic \u03b1-Al2O3 reflections decreases in the Co-MOF-74 >Ni-MOF-74> Zn-MOF-74 order, which may suggest that the thickness of metal organic framework layers in prepared structured catalysts increases. It is also worth mentioning that, in all considered materials, we observed that the crystallisation of MOF material over the metallic support was strongly influenced by the number of metallic supports placed in the Teflon liners for in situ deposition. Once the total amount of metallic supports exceeded 1\u00a0g per synthesis, we did not observe the metal organic framework crystals either in reacting vessels or deposited on the structured carriers.To determine the structure and the purity of the MOF layers deposited on FeCrAl plates, the XPS analysis of triple deposited MOF-74 layers on FeCrAl plates was performed. The results of the XPS analyses are presented in Fig. 2\n. The survey spectra of the triple deposited MOF-74 layers deposited on FeCrAl plates (black lines) and calcined FeCrAl plates are presented in Fig. 2 A, D, G. It may be seen that the survey spectra of Zn-MOF-74, Ni-MOF-74 and Co-MOF-74 do not reveal any lines originating from calcined FeCrAl plates (cf. red lines) and only signals from Me(Zn, Ni, Co) 2p, O1s and C1s may be observed. Since the alumina is mainly present at the calcined FeCrAl plate surface due to the migration of alumina at 1100\u00a0\u00b0C calcination, we used the signal at 75\u00a0eV originating from Al 2p [45] as an internal marker to determine the purity deposited MOF-74 layers. The zoomed area for 75\u00a0eV region for Me(Zn, Ni, Co)-MOF-74 catalysts are presented in Fig. 2 B, E, H. It may be seen that, for all considered cases, the Al 2p line does not occur at the XPS spectra of Me (Zn, Ni, Co)-MOF-74 catalysts. The XPS spectra for Zn 2p, Ni 2p and Co 2p for Me (Zn, Ni, Co)-MOF-74 are presented in Fig. 2 C, F, I. The Zn-MOF-74/FeCrAl catalyst reveal two main peaks at 1022.2 and 1045.3\u00a0eV (Fig. 2 C) that may be attributed to Zn 2p3/2 and Zn 2p1/2 [46]. For the Ni-MOF-74/FeCrAl catalyst two main group bands were detected with the peaks at 855.9 and 873.6\u00a0eV and associating satellite peaks at 860.7 and 879.4\u00a0eV, which may be attributed to Ni 2p3/2 and Ni 2p1/2 [47], respectively. At the XPS spectrum of Co-MOF-74/FeCrAl, catalyst peaks at 781.9 and 797.8\u00a0eV and associating satellite peaks at 785.8 and 802.6\u00a0eV are observed. These may be attributed to Co 2p3/2 and Co 2p1/2 [48], respectively.The effectiveness of the in situ MOF deposition over structured supports was determined gravimetrically after each deposition. The results are presented in Fig. 3\n A. The effectiveness of the MOF deposition on the structured carriers was presented as a mass increase per geometrical surface area of metallic support. Such deposition results are commonly used for the comparison of coating loading in structured reactors engineering [27,49]. The lowest MOF loading was observed for the layers deposited on FeCrAl plates. For this support, the individual deposition of Zn- and Co-MOF-74 layers never exceeds 0.32\u00a0mg/cm2 (maximum value achieved for Zn-MOF-74 after double deposition). The maximum mass increase after triple deposition was achieved for Co-MOF-74, and was equal to 0.669\u00a0mg/cm2. The deposition of MOF layers of on FeCrAl wire gauze results in considerable MOF mass increase on metallic support. In general, the MOF loading on wire gauze increases on average by a factor of two, with some minor derogations for Co-MOF-74 at single deposition where this value increases almost four-fold, and for Zn-MOF-74 at triple loading, where the mass increase is almost one order of magnitude higher than for the FeCrAl plate. When considering the total mass increase on the FeCrAl wire gauze in comparison with the FeCrAl plate, the mass loading factor increases in a arrange 2.9-fold for Zn-MOF-74, two-fold for Co-MOF-74 and up to 2.1 times for Ni-MOF-74 (cf. Table S1). The highest metal organic metal loading by in situ deposition was achieved for NiCr foam. Analysis of the obtained MOF loading values (Table S1) reveals that the maximum MOF loading was achieved after triple deposition of Co-MOF-74. Considerable high values were achieved for double deposition of Zn-MOF-74. It must be emphasised that the total mass increase forms the following order Co-MOF-74>Zn-MOF-74>Ni-MOF-74, which is similar to MOF loading on the FeCrAl plate and wire gauze. It must be also pointed out that the Ni-MOF-74 indicated the worst adhesion properties on all considered metallic carriers.The morphology of deposited coatings on structured supports was determined using two methods: digital photography and SEM microscopy. The results of digital photography imaging are presented in supplementary materials in Figs. S3\u2013S5 for Me (Zn, Ni, Co)-MOF-74 layers deposited on FeCrAl plates, FeCrAl wire gauzes and NiCr foams, respectively. In the case of Zn-MOF-74, the single deposition on each structured support is barely seen in digital pictures. Considerable changes in layer deposition on each structured support may be observed after double and triple deposition (Figures S3-S5, B and C). For Ni and Co-MOF-74 layers, the single deposition of MOF material may be observed. To determine in detail the morphology of prepared structured catalysts, SEM analysis was performed. To enhance the visibility of SEM images, pseudo colouring by using defined RGB colours determined by UV\u2013Vis spectroscopy was performed. The SEM images are presented in Fig. 4\n for three structured carriers, and in Fig. 5\n for triple deposited MOF layer on NiCr foams with 2000x magnification. Since the whole matrix contained 27 images per single SEM magnification, the results for each deposition for M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 are presented in supplementary materials in Figs. S6\u2013S14. The deposition of Zn-MOF-74 on structured carriers is presented in Figs. S6-S8. It can be seen that, after single deposition, surfaces of all three structured carriers at the lowest magnification (200x) do not show any substantial changes in carrier morphology. This changes upon increasing magnification from 2000x up to 5000x. The surface seems to be coated with a thin layer of MOF with visible small crystals of irregular shape. This phenomenon changes after double in situ coating (Fig. S7). In this case, even a quick look at the catalyst's surface at low-magnification images reveals the complete coverage of the structured carrier. The crystals began to grow in more regular shape, similar to hexagonal rods. The shape of the Zn-MOF-74 structures is more evident for wire gauze and foam structures. The MOF-74 growth on structures is evident, and good adhesion may be observed. The higher magnifications also reveal smaller crystals found on larger ones (Figure S6-S8 E-F). The triple deposition reveals full surface coverage in all three structured carriers. The MOF crystals reveal full developed shapes. Detailed analysis of SEM images allows the thickness of the Zn-MOF-74 layers to be determined, which in that case is equal to 40\u00a0\u03bcm. The important feature of Zn-MOF-74 layers is depicted in Fig. 4 A1, B1, C1 as well as in Figs. S6\u2013S8 G-I, where, for the foam carrier, the MOF crystals are perpendicularly oriented to the foam surface, in contrast to the FeCrAl plates and wire gauzes, where the stochastic orientation prevails.The SEM images for Ni-MOF-74 are presented in Fig. 4 A2, B2, C2 for triple deposition and in Figs. S9\u2013S11 for single, double and triple deposition. It may be seen that the crystal morphology is far different from that of Zn-MOF-74 crystals. The surfaces of all three structured carriers are covered with spherical crystals with an average diameter of 10\u00a0\u03bcm. However, it must be emphasised that the crystals form a thin layer which is more visible after double and triple coating of wire gauze and foam carriers. One can observe that surface coverage is uniform after double deposition on structured carriers. After triple deposition, the carriers\u2019 surfaces reveal point-crystal growth (Fig. 4 B2 and C2). The thickness of the Ni-MOF-74 layers was equal to the average MOF particle diameter, i.e. 10\u00a0\u03bcm. The average thickness after triple coating was approx. 30\u00a0\u03bcm (cf. Fig. S11 G).The Co-MOF-74 morphology is presented in Fig. 4 A3, B3, C3 and Figs. S12\u2013S14. The crystal morphology exhibits more regular hexagonal shape in comparison with Zn-MOF-74. It can be seen that complete carrier coverage is achieved after single deposition in all considered carriers (Figures S12 A-I). It must be emphasised that, for single deposited Co-MOF-74 on NiCr foam, there is different morphology in comparison with Co-MOF-74 deposited on the FeCrAl plate and wire gauze. The foam surface seems like it was treated by some kind of MOF primer and forms the incubation-like centres for further crystal growth. The morphology of the Co-MOF-74 crystals is similar for all kinds of metal supports after single deposition (Figures S12 A-I). In all considered structured carriers, the hexagonal crystal is perpendicularly oriented to the metallic carriers. The triple deposition of Co-MOF-74, however, causes crystal aggregation, and local crystal spots can be observed especially for the FeCrAl wire gauze and NiCr foam. However, the presence of the local crystal hypertrophies is not evident as in the case of Zn- and Ni-MOF-74 layers. It must also be pointed out that the thickness of the Co-MOF-74 layers is lower than for Ni-MOF-74 and is equal to 20\u00a0\u03bcm (average single crystal size). Due to the growth of the MOF crystals perpendicular to the support surface, the crystal tends to fill the free space between crystals rather than to overgrow already grown crystals.The results of the krypton and nitrogen adsorption on bare structured carriers, MOF powders and MOF deposited on metallic supports are summarised in Table 3\n. The krypton adsorption on structured supports revealed that structured carriers are non-porous solids (Table 3\nA). The measured SBET for the FeCrAl plate, wire gauzes and NiCr foams were equal to 0.027, 0.012 and 0.039\u00a0m2/g, respectively. The nitrogen adsorption on powder samples (Table 3 B), collected using the in situ solvothermal method, revealed that the specific surface SBET areas of prepared samples were approx. 1000\u00a0m2/g for all prepared powder MOF-74 samples, which corresponds well with the results presented in the literature [39,42]. Since for the characterisation of metallic structured catalysts with deposited porous metal organic framework layers there is no proposed methodology for the presentation of the SBET results, the data presentation was two-fold. To compare the specific surface of the M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 layer over representative FeCrAl support, the SBET was referred to the mass of MOF-74 deposited on the metallic carrier. This value was determined gravimetrically after M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74/FeCrAl plate activation. However, to compare the values of the specific surface between the supported catalysts, the SBET was referred to the total mass of the structured catalyst. When analysis of SBET for the FeCrAl plate referred to the deposited MOF layer (Table 3 C), it may be seen that the values for SBET are lower than the calculated specific surface areas for powder samples, and are equal to 331.6\u00a0m2/g for Zn-MOF-74, 823.5\u00a0m2/g for Ni-MOF-74 and 716.7 for m2/g for Co-MOF-74. It may be observed that a considerable decrease was observed for Zn-MOF-74, where the value of specific surface area was approx. 700\u00a0m2/g lower than for its powder counterpart. The difference between the calculated SBET values may be two-fold. The successful in situ synthesis of Zn-MOF-74 over metallic structures was achieved by the optimised triple synthesis, where the primer layer on Zn-MOF-74 was prepared from the zinc acetate solution, whereas double and triple deposition was synthesised by using a nitrate solution as zinc precursor. For Ni- and Co-MOF-74 catalysts, the observed SBET decrease was lower and equal to approx. 200\u00a0m2/g and 300\u00a0m2/g. In this case, however, the Ni- and Co-MOF-74 the triple deposition may cause crystal overgrowth which may influence the overall SBET value. Additionally, the multiple layer deposition may also influence the availability of micro and mesopores for adsorbed molecule. Analysis of the SBET values referred to the total mass of the structured catalyst (Table 3 D; mass of the metallic carrier\u00a0+\u00a0mass of the deposited layer) leads to the general conclusion that the amount of the deposited metallic organic frameworks on the structured support increases in the following order: FeCrAl wire gauze\u00a0>\u00a0FeCrAl plate\u00a0>\u00a0NiCr foam, which is different than the gravimetrical measurements from Table S1 and Fig. 3. However, it must be emphasised that the values determined by the gravimetrical method were performed after structure catalyst washing after in situ deposition and are not impacted by the high temperature UHV activation of catalysts samples in the sorption meter. Analysis of the literature data on TGA analysis of the metal organic frameworks leads to the conclusion that, at approx. 300\u00a0\u00b0C, M (M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 is equal to 30\u00a0wt %. of the initial mass [39,40]. In this study, the activation of MOF prior to the N2 sorption was performed under 250\u00a0\u00b0C to ensure effective activation. Since the metal supports used in this study are non-porous solids, we can estimate the mass of the catalyst deposited on the surface of the structured supports by formula previously proposed in the literature [50]:\n\n(1)\n\n\nm\n\n\nM\nO\nF\n\nd\ne\np\no\ns\ni\nt\ne\nd\n\no\nn\n\n\nt\nh\ne\n\ns\nu\np\np\no\nr\nt\n\n\n\n=\n\n\n\nS\n\nB\nE\nT\n,\n\nM\nO\nF\n\n\nd\ne\np\no\ns\ni\nt\ne\nd\n\no\nn\n\nt\nh\ne\n\ns\nu\np\np\no\nr\nt\n\n\n\nS\n\nB\nE\nT\n,\n\nM\nO\nF\n\np\no\nw\nd\ne\nr\n\n\n\n\n\u00b7\n\n\nm\n\nM\nO\nF\n\np\no\nw\nd\ne\nr\n\n\n\u00b7\n1000\n,\n\n\n\nm\ng\n\n\n\n\nwhere: mMOF deposited on the support is the approximated mass of the deposited MOF layer on structured support, SBET, MOF deposited on the support is the specific surface area of the structured reactor (structured support with MOF layer), MOF powder is the mass of the powder used to calculate SBET equal to 1\u00a0g. The calculated values of MOF mass deposited on different structured supports lead to the conclusion that the MOF-74 layers are favourably deposited on NiCr foams and FeCrAl plates. However, to fully characterise the effectiveness of the in situ layering, the type of the MOF-74 by metal should be considered. It may be seen that the lowest calculated MOF masses were obtained for Zn-MOF-74. Despite the fact that, in the case of Zn-MOF-74, XRD analyses revealed a characteristic pattern for MOF-74 crystals at the metallic support, deep analysis of the SEM pictures for individual depositions shows that the well-defined crystals are formed after triple deposition (Fig. 5 and Figs. S6-S8). The first two layers should therefore be defined as intermediate MOF-layers or primer MOF-layers. The decrease in calculated MOF referred to the total mass of the structured catalysts using N2 sorption is related to the low contribution of well-crystallised MOF on the overall mass of the deposited layer. The opposite situation can be observed for Ni- and Co-MOF-74 layers deposited on structured carriers. Here, the total mass of deposited MOF calculated from N2 sorption gives two-order of magnitude higher values of deposited MOF when comparing to Zn-MOF-74. The SEM results clearly show the growth of well-defined crystals on structured supports after single deposition (Figs. S9-S11).To determine the molecular nature of the prepared structured MOF catalysts, IR and Raman analyses were performed. The detailed IR analysis of prepared samples using ATR, DRIFT and transmission IR can be found in supporting information (Fig. S15). The characterisation of the active centres in prepared materials was performed by the sorption of two probe molecules: carbon monoxide and CD3CN. Both probe molecules are commonly used to study acidic and basic properties of heterogeneous catalysts. The results of CO and CD3CN adsorption are presented in Fig. 6\n and Fig. 7\n, respectively. The low temperature of carbon monoxide adsorption on M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 gives the rise of the main band at 2160-2180\u00a0cm\u22121, which corresponds to Me2+-CO adducts formed in the prepared metal organic framework catalysts. It has been previously reported in the literature [51,52] that the values of the main CO adsorption bands for M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 decreases in the following order: Ni (2180\u00a0cm\u22121)\u00a0>\u00a0Zn (2173\u00a0cm\u22121)\u00a0>\u00a0Co (2162\u00a0cm\u22121). The high C\u2013O stretching frequencies are derivative of the smallest size and the highest polarisation of Ni2+ ion for the Ni-MOF-74 sample (Fig. 6 B) [51,52]. It must be emphasised that, upon increase of partial pressure of carbon monoxide, the minor bands at 2150-2100\u00a0cm\u22121 and 2200-2250\u00a0cm\u22121 can be observed and may be attributed to some combination overtones of \u03bd(CO). It was also observed that, at high CO coverages, for Zn-MOF-74 and Ni-MOF-74 an additional band at around 2135\u00a0cm\u22121 is formed, which was previously assigned to liquified CO in the MOF pores [53].The results of CD3CN probe molecule adsorption on M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 catalysts are presented in Fig. 7. The adsorption of CD3CN probe molecules shows rise of a sharp and intensive band at 2110\u00a0cm\u22121, which is characteristic of deuterated \u03bd(CD3) vibrations, and two intense bands at 2237 and 2290\u00a0cm\u22121, which may be attributed to physiosorbed CD3CN and coordinated CN species to Lewis acid sites, respectively [53,54]. The acidic properties of various MOF materials by using CD3CN as a probe molecule has recently been reported for MIL-140C (Zr), MIL-140D (Zr) [55], MIL-100 (Al, Fe, Cr) [54]. It must be pointed out that the values of \u03bd(CD3) and \u03bd(CN) vibrations are similar to those reported for MIL 140C, D and MIL-100 metal organic frameworks, which may lead to the conclusion that they possess similar acid strength.The complementary experiments of molecular properties of prepared samples were performed by \u03bcRaman spectroscopy. The results of \u03bcRaman analysis were presented as a \u03bcRaman maps (Figs. 8 and 9\n\n), for two reasons. The \u03bcRaman mapping allowed us to show the distribution of the MOF over the metallic carrier. Comparison of the \u03bcRaman maps leads to the conclusion that the most uniform distribution was achieved for Ni and Co-MOF-74 samples (Fig. 8 C and D). Indeed, the \u03bcRaman maps also exhibit local layer overlapping (brighter spots at \u03bcRaman maps), which is in good agreement with SEM images for samples after MOF triple deposition. However, it must also be pointed out that the determination of the surface homogeneity using only \u03bcRaman maps must be carried out with a high degree of caution, since \u03bcRaman maps for the homogeneous calcined FeCrAl plate also reveal some local increase in Raman intensity. The corresponding Raman spectra (Fig. 9) exhibit the structure of prepared composite samples. The Raman spectrum of the calcined FeCrAl plate (Fig. 9 A) reveals bands at 418, 630 and 750\u00a0cm\u22121, which may be attributed to \u03b1-Al2O3 of hexagonal symmetry (band at 418\u00a0cm\u22121) [56], \u03b1-Fe2O3 (band 630) and \u03b3-Fe2O3 [57]. The \u03bcRaman of the FeCrAl plate may be treated as a marker. Since the depth of the sample penetration is relatively high for Raman scattering, the presence or absence of a marker band may be useful in determining the surface thickness. In our previous work, we reported that the use of various characterisation techniques such as XPS, \u03bcRaman and EDX allows the determination of the in-depth distribution of the active phase [58]. Here, we can observe that, for the M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 composite catalysts deposited on metallic support, there was no signal originating from the metallic support. The Raman spectra of M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 reveal two main band group regions: to 820\u00a0cm\u22121 and 1200-1700\u00a0cm\u22121. The 1200-1700\u00a0cm\u22121 reveals bands at 1275, 1412, 1501, 1560 and 1619\u00a0cm\u22121, which may be attributed to \u03bd(C\u2013O) from deprotonated hydroxyls, symmetric \u03bd(COO\u2212) and stretching and deformation vibrations of benzene rings [41], respectively. The bands at lower frequencies, at approx. 820 and 560\u00a0cm\u22121, may originate from benzene ring bending and deformation vibrations, respectively [41,51]. The additional bands, at approx. 413\u00a0cm\u22121, can be due to \u03bd(Me\u2013O) vibrations [51]. Comparison of the Raman maps for M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 structured catalysts and the FeCrAl plate lead to the conclusion that the metallic carrier is uniformly covered with the MOF layer. Similar observations can be observed from the analysis of XPS results (cf. Fig. 2 A).The adherence of the deposited on metallic support M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 layers was evaluated by using an ultrasound bath mechanical resistance test. This type of examination is frequently used for layer adherence testing in structured catalyst characterisation [37,59,60]. The results of the MOF layer adherence performance for various structured supports are presented in Fig. 3 B. The results are presented as a percentage of mass loss during ultrasonic irradiation treatment. The best adherence properties were observed for NiCr foams. After the ultrasonic irradiation test for Zn-MOF-74, almost 50% of the deposited material remained at the support surface. This value was slightly lower for the Ni and Co-MOF-74 layer, with 40% and 35% of the material deposited over a metallic foam. The metal organic framework layers deposited on FeCrAl wire gauzes indicated lower adherence to the structured support. In the case of Zn-MOF-74, almost all of the deposited material was removed from the structured support, whereas, for Ni and Co-MOF-74, 10% and 20% of the deposited material remained on the support. Comparison of the layer adherence to the support carrier after mechanical resistance testing for FeCrAl plates and NiCr foams leads to the conclusion that the stability of the deposited MOF material is derivative either of the available geometrical area and its shape or of the total volume of the support which is sonochemically treated. During the mechanical stability experiment, the structures were stochastically placed in an ultrasonic bath. Their natural arrangement in the bath left one of the sides less subjected to ultrasounds. What is more, comparison of the support structure morphology for wire gauzes and foams may lead to the conclusion that intensity of ultrasound waves can be gradually screened by the bone-like structure of NiCr foam. It must be emphasised that the literature reports on the deposition of metal organic frameworks on metallic supports is rather scarce, which makes comparison of the obtained results with other literature reports impossible. Since the metal organic frameworks are mainly formed into the desired shapes, such as pellets, foams or monoliths with the addition of a binder [17\u201319], or as required in the case of their use as the electrodes [13], the influence of the other kinds of forces of the prepared materials has been considered.The catalytic activity of prepared M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 powders and Me (Zn, Ni, Co)-MOF-74 deposited on NiCr foams was measured in the aerobic oxidation of cyclohexene. The results are summarised in Table 4. It must be emphasised that bare metallic supports revealed no activity in the aerobic activation oxidation of cyclohexene. The result of catalytic activity is expressed as a function of total conversion of cyclohexene and individual selectivity to the main products: cyclohexene oxide, 2-cyclohexen-1-ol, 2-cyclohexen-1-one and trans-cyclohexane-1,2-diol. It may be seen that the activity of all prepared powder catalysts exceeds 50% conversion. The activity of prepared powder samples was: 66.5% for Zn-MOF-74, 59.0% for NiMOF-74 and 52.3% for Co-MOF-74 catalysts. Analysis of the selectivity for prepared samples shows that, for Zn- and Ni-MOF-74 catalysts, the oxidation reaction proceeds mainly to 2-cyclohexen-1-ol and 2-cyclohexen-1-one. In the case of the Zn-MOF-74 catalyst, the selectivity to cyclohexene-1-ol and 2-cyclohexen-1-one was 65.4% and 13.9%, whereas for Ni-MOF-74 it was 74.3% and 13.3% respectively. The selectivity for the cyclohexane oxide was 12.5 and 8.7% for Zn-MOF-74 and Ni-MOF-74, respectively. However, when analysing the oxidation reaction results for Co-MOF-74, it may be seen that the cobalt oxide favours the epoxidation reaction, with cyclohexane oxide as the main product with almost 19% selectivity, whereas the contributions of the 2-cyclohexen-1-ol, 2-cyclohexen-1-one and the other products were lower. Moreover, among the products, trans-cyclohexane-1,2-diol was not detected. Additionally, the contribution of the side products reached 30%. Although in the literature [29,61] we can find some results on cyclohexene catalytic oxidation over Me-MOF-74 catalysts, comparison of the obtained results is impossible due to different synthesis procedures for MOF-based materials and their different physicochemical properties. For example, Ruano et al. [61] synthesised the catalysts from metal acetate solutions (Zn, Co, Ni, Mn and Cu)-MOF-4 with another nanocrystalline structure. Furthermore, the morphology of prepared MOFs in Ref. [61] was far from that of our materials. The SBET values presented in Refs. [61] were 948, 693 and 514\u00a0m2/g for Zn-, Co- and Ni-MOF-74, respectively. These SBET results are considerably lower than the SBET values presented in this work. The next difference between our work and [61] lies in the fact that, during the catalytic activity tests, Ruano et al. [61] used H2O2 or tetr-buty hydroperoxide (TBHP) as an oxidising agent together with atmospheric oxygen. Indeed, both oxidising agents can be used to either initialise radical reaction (TBHP) or oxidise cyclohexene, but the oxidising effect is supposed to be higher than in the case of molecular oxygen. Despite this fact, the authors presented cyclohexene conversion reaching 71.5% for Co-MOF-74, 40% for Ni-MOF-74 and 5% for Zn-MOF-74, and analysis of the reaction product was performed by gas chromatography equipped with flame ionisation detector. In relation to the work written by Sun et al. [29], the preparation results were different from the preparation conditions presented in this study.When analysing the oxidation results under 10\u00a0bar O2 pressure, a general increase of the activity for Ni- and Co-MOF-74 samples can be observed. The conversion of cyclohexene for Ni-MOF-74 increases up to 81.7%, whereas for Co-MOF-74 the conversion is equal to 67.9%. The individual selectivity for the oxidation products changes for Ni-MOF-74 at 10\u00a0bar O2, with considerable increase to 2-cyclohexen-1-one, cyclohexane-1,2-diol and other products. In the case of Co-MOF-74, with the reaction at elevated O2 pressure, the selectivity of oxidation products remains at the same level, with a slight increase of selectivity to cyclohexane-1,2-diol. For Zn-MOF-74 powder catalysts, we could see no considerable changes in either conversion or selectivity. Catalytic activity was also determined for MOF catalysts deposited in situ on NiCr foams. Through analysis of the results of the catalytic activity under 10\u00a0bar O2 pressure over structured M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 deposited on NiCr foams, a general decrease in conversion of cyclohexene can be observed. It can be seen that, in all considered MOFs deposited on NiCr foams, the conversion of cyclohexene decreased by a factor of two. The reason of this phenomenon can be explained by the decrease of the effectiveness factor of the catalyst in cyclohexene oxidation. Despite the fact that in all catalytic experiments the same catalyst amount was used (50.0\u00a0mg), it must be pointed out that, in the case of powder catalysts, the availability of the active sites is higher due to the wide distribution of catalysts in the reacting mixture. The comparison of the SEM results in Figs. S6-S14 for both supported catalysts and powder MOF-based materials shows that the size of the individual grains varies from 5 to 10\u00a0\u03bcm, whereas the thickness of the deposited layer is as high as 40\u00a0\u03bcm. The considerable thickness of the MOF layer on the support may lead to a considerable decrease in the catalytic activity of prepared materials according to the Thiele modulus. However, the calculation of the Thiele modulus and effectiveness factor calculations exceeds the scope of this article, indicating future directions for the application of structured reactors with deposited MOFs.The characterisation of MOF materials in the catalytic oxidation of cyclohexene should consider also a factor related with the migration of a metal from MOF structure to the reaction solution. The results of the metal content in post-reaction mixtures are presented in Table 4. Analysis of the obtained results leads to the conclusion that, in the case of Zn-MOF-74 and Ni-MOF-74, the metal content in the post-reaction mixture was below the detection limit. Only small amounts of zinc ions were detected in the post-reaction mixture (0.12\u00a0mM). Noticeable amounts of metal in the post-reaction mixture were observed for Co-MOF-74. The amount to detected cobalt was approx. 3\u00a0mmol for the Co-MOF-74 powder sample for the oxidation reaction under atmospheric and 10\u00a0bar O2 pressure. However, for MOF deposited on NiCr, the value of detected Co was one order of magnitude lower, and was equal to 0.39. The decrease of cobalt migration to the reaction mixture may be related with the generally lower activity of the Co-MOF-74/NiCr catalyst and the good adhesion of the MOF to the NiCr foam surface. It must be emphasised that, in the case of MOF catalysts deposited on NiCr foams, the catalysts were placed in the reaction vessel and simply removed after the reaction, whereas cobalt catalysts in powder form required additional filtration to separate the reacting mixture and powder catalyst. The lack of additional filtration of the post-reaction mixture and catalyst in the case of MOF deposited on NiCr may be a fundamental step towards the wider application of MOF materials as heterogeneous catalysts.The aim of this paper was to obtain and characterise thin metal organic framework layers on various metallic structured supports by using spectroscopic and microscopic methods, and to determine their potential application in the catalytic oxidation of cyclohexene. The in situ deposition of metallic organic framework thin layers consists of three steps, including support pre-treatment, in situ solvothermal deposition and MOF-layer activation to remove residual solvents from synthesis protocol. The prepared structured carriers with deposited MOF-74 layers were characterised with various characterisation techniques to determine the surface morphology and their molecular structure. The in situ deposition of metal organic frameworks was the most effective for Zn- and Co-MOF-74 on NiCr foams, giving the approx. 4\u00a0mg/cm2 mass increase after triple coating. We have indicated that there is no difference in molecular structure between in situ deposited and non-deposited crystalline phase of metal organic frameworks. The high mechanical resistance of prepared M(M\u00a0=\u00a0Zn; Ni; Co)-MOF-74 layers on NiCr foams and FeCrAl plates was confirmed by the ultrasonic irradiation performance.The activity of prepared MOF catalysts both in powder form and MOF deposited on NiCr foams was measured in the catalytic oxidation of cyclohexene. The prepared catalysts revealed high activity in the studied reaction, with the conversion exceeding 50% for powder catalysts under both atmospheric and elevated pressures. The catalysts deposited on NiCr foams revealed twice lower conversion in comparison with their powder counterparts. However, the use of structured catalysts did not require their additional filtration from the reaction mixture, which makes them favourable for further testing as heterogeneous catalysts in the organic reagents oxidation.We believe that the in situ deposition of metal organic frameworks from Me2(dobdc) group, proposed in this study, will lead to the substantial development of MOF materials and their further application in heterogeneous catalysis as structured reactors.\nP.J. Jod\u0142owski: Formal analysis, Investigation, Data curation, Writing - review & editing. G. Kurowski: Formal analysis, Investigation. K. Dymek: Formal analysis, Investigation. R.J. J\u0119drzejczyk: Formal analysis, Investigation. P. Jele\u0144: Formal analysis, Investigation. \u0141. Kuterasi\u0144ski: Formal analysis, Investigation. A. Gancarczyk: Formal analysis, Investigation. A. W\u0119grzynowicz: Formal analysis, Investigation. T. Sawoszczuk: Formal analysis, Investigation. M. Sitarz: Formal analysis.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to acknowledge dr Jakub Marchewka (Faculty of Materials Science and Ceramics, AGH University of Science and Technology) for digital photography of prepared structured catalysts and also Maciej Bik (Faculty of Materials Science and Ceramics, AGH University of Science and Technology) for GIXRD profile fitting and phase assignment.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.micromeso.2020.110249.", "descript": "\n                  The aim of this study was to obtain and characterise thin metal organic frameworks layers supported on various metallic structured carriers such as FeCrAl plates and woven gauzes and NiCr foams. The thin layers of the metal organic frameworks were fabricated by in situ solvothermal deposition, optimised by the selection of metal precursor and the layering/washing order. The parameters of the resulting metal organic framework coatings were characterised in terms of layer thickness in correlation with the fold overlap, morphology, chemical properties and mechanical resistance to ultrasonic irradiation. Several techniques were used to characterise metal-organic framework layers, including in situ FTIR, \u03bcRaman mapping, XRD, low temperature sorption of liquid nitrogen, and SEM. The results of structural analysis of prepared structured catalysts revealed that the surfaces of the structured carriers are uniformly covered with Me-MOF-74 thin layers. The mechanical stability tests showed that the metallic foams possessed high mechanical resistance and may be considered as a structured support for heterogeneous catalysts.\n               "}
{"full_text": "Data will be made available on request.The progress in industrialization and urbanization have created different environmental pollution issues worldwide. Regardless of much developments, industrial units are one of serious threats to the environment. Continuous production waste and their discharging into water resources is extremely terrible due to the mutagenic and carcinogenic nature of the pollutants [1\u20134]. Behind this water pollution, the textile, paper, leather, plastic, pharmaceutical, food processing and cosmetics units are principal offenders. The discharged water from such industries contains both organic and inorganic particulates that are malignant to human beings and aquatic life [5\u20138]. So, for healthy environment of all the living organisms, the removal of these wastes from water is necessary. In this regard, several strategies have been employed, i.e., ozonation, electrocoagulation, electrochemical destruction, photo-Fenton degradation, membrane filtration, coagulation, ion exchange and adsorption [3,9\u201313]. Among these, the photocatalytic process is proved to be highly efficiency and if the catalyst is active under visible light, then this process also viable economically. Various catalysts have been studied that showed promising efficiency for the removal of toxic pollutants, i.e., composites, metallic oxides (Mn3O4, CoO, NiO, CuO, ZnO, CeO2, SnO2) doped materials (Sn doped titania, N-doped Zirconia, NiFe2O4 and ZnO heterostructures and BaFe12O19) have been applied as photocatalysts and their response was promising [3,14\u201319]. Amongst these materials, ferrites have been investigated extensively owing to their ease of preparation, high stability as well as magnetically recoverable nature. There are six types of hexaferrite materials (M\u2212type, U-type, W-type and X-Z-types). M\u2212type hexaferrite possess hexagonal crystal lattice with P63/mmc space group in which 11 symmetry sites are present with 64 ions per unit cell. Due to high magnetization, chemical stability and large microwave magnetic loss, these hexaferrites are highly explored for different applications. Ferromagnetic oxides include M\u2212type hexaferrites (MFe12O19), where M may be divalent cation, i.e., Ba and Sr. The SrFe12O19 (SFO) belongs to metallic oxide-type hard magnetic compounds and has been used usually in many technological applications [4,20\u201322]. This compound with magneto plumbite structure exhibits electromagnetic behavior owing to its higher magnetic parameters, low conductivity loss and high permeability. It has been extensively employed in microwave absorbers, permanent magnetic designs, magnetic recording media (MRM), high-frequency electromagnetic (HFEM) devices, sensors and EM (electromagnetic) protecting devices [23]. The SFO being n-type semiconductor has been utilized as magnetic component in the fabrication of metal oxide photocatalysts [3,16,17,24]. Owing to their magnetic nature, these composite photocatalysts can be recollected from reaction mixture via simple and economic method by using a magnet. Xie et al. [25] fabricated magnetic Bi2O3/SrFe12O19 heterojunction and observed the enhanced photocatalytic performance of SrFe12O19 due to the heterojunction of p-n type. These hexagonal ferrites have been studied occasionally in heterogeneous catalysis for photodegradation of dyes. Different methods have been used to synthesize SFOs, i.e., co-precipitation, sol\u2013gel, hydrothermal and microemulsion method etc [4,16,23]. Among these synthesis strategies, the micro-emulsion approach offers various advantages versus other synthesis techniques, i.e., to control the particle shape, size, surface area, morphology and homogeneity. It is also a facile, fast and eco-benign versus others [3].Based on aforementioned facts, microemulsion is facile and low temperature operating strategy which imparts good homogeneity and produces crystallites in nano range. Herein, we have fabricated pure SrFe12O19 (SFO) and Ni2+ doped SrNixFe12-xO19 (x\u00a0=\u00a00.05\u20130.25) NPs named as SFNO (1\u20135, Ni content\u00a0=\u00a00.0, 0.05, 0.10, 0.15, 0.20, 0.25 are named as SFO, SFNO1, SFNO2, FNO3, SFNO4 and SFNO5, respectively) via facile microemulsion route. Hence, the fabricated materials were assessed for the removal of CV dye under the irradiation of visible light. Furthermore, the effect of Ni substitution on ferroelectric and dielectric behavior of SFO material was also investigated.The iron nitrate nonahydrate (\u226599\u00a0%), nickel nitrate hexahydrate (\u226599\u00a0%), strontium nitrate (99.99\u00a0%), cetyltrimethyl ammonium bromide (CTAB), ammonia (25\u00a0%) and CV dye (C25H30N3Cl) were precured from Sigma Aldrich and were utilized as received. Distilled water was utilized for the washing purposes and solutions preparation.All the metal nitrate solutions were mixed according to required compositions (stoichiometric amounts) followed by magnetic stirring and heated up to 60\u00a0\u00b0C. To minimize the aggregation of particles and to well control the crystallites size, the aqueous CTAB in equal volume was poured to all the mixture. Drop wise pouring of NH3 solution resulted into precipitation of metallic hydroxides at 10\u201311 pH. The precipitates were neutralized up to pH 7 by washing with water. To minimize moisture contents, the precipitates was subjected to drying in an electric oven at 90\u00a0\u00b0C for 8\u00a0h, then these were sintered at 850\u00a0\u00b0C where metallic hydroxides were transformed into desired mixed metal oxide (hexaferrite) lattice. Lastly the powdered material was transformed into pallets for dielectric and ferroelectric characterizations (Fig. 1\n). Total six samples were prepared at using\u00a0\u00d7\u00a0(Ni content\u00a0=\u00a00.0, 0.05, 0.10, 0.15, 0.20, 0.25), which were named as SFO, SFNO1, SFNO2, FNO3, SFNO4 and SFNO5, where\u00a0\u00d7\u00a0concentration is 0.0, 0.05, 0.10, 0.15, 0.20, 0.25, respectively.The crystalline phase of the prepared composites was studied by Philips\u00a0X\u00a0pert PRO (3040/60 model) X-ray diffractometer (Cu K-alpha radiation and \u03bb\u00a0=\u00a01.5406 A\u00b0) in 2-theta range of 20-60\u00b0. Characteristic functional groups were identified by FTIR analysis (Nicolet FTIR interferometer (model-8400S) in 400\u20134000\u00a0cm\u22121 range. Raman spectra were traced in 0\u2013800\u00a0cm\u22121 range through spectrometer (model 6400 triple JobinYvon-Atago/Bussan). At room temperature, the ferroelectric loops were studied by P-E (M/s Radiant Instrument USA) tracer. Wayne Kerr 6500B Impedance analyzer with DC bias current from 0 to 100\u00a0mA was used to analyze dielectric response in frequency range 20\u00a0Hz to120 MHz using silver coating pellets of 6\u00a0mm thickness at 300\u00a0K. The PL spectra were acquired by fluorescence spectrophotometer (SHIMATZU-RF 5301PC). UV\u2013Visible diffuse reflectance absorption spectra for all composites were measured via double beam spectrophotometer (Carry, Agilent).A 10\u00a0mg/L solution of CV dye was prepared in distilled water and 500\u00a0mL of dye was taken out and 10\u00a0mg of catalyst was added. Prepared mixture was stirred for 10\u00a0min to stabilize the reaction mixture. To acquire equilibrium, the mixture (catalyst and dye) was shaken for 20\u00a0min in dark followed by illuminating to visible light (150\u00a0W Xe lamp) for various time intervals. Analyte was withdrawn for every 10\u00a0min, centrifuged and absorbance (588\u00a0nm) was recorded. The percentage CV dye removal was assessed using Eq. 1.\n\n(1)\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nD\ne\ng\nr\na\nd\na\nt\ni\no\nn\n\n\n(\n%\n\n)\n\n=\n1\n-\n\n\nA\nt\n\n\nA\no\n\n\n\u00d7\n100\n\n\n\nwhere, At and Ao shows the absorbance of dye solution at specific time interval\u2019t\u2019 and at zero time.The XRD patterns of pure SFO and SNFO (1\u20135) composites in 2-theta range of 20\u00b0-60\u00b0 are shown in Fig. 2\na. Intense diffraction peaks were evolved at 2 theta, i.e., 20.27\u00b0, 23.15\u00b0, 30.32\u00b0, 32.35\u00b0, 35.22\u00b0, 40.00\u00b0, 49.93\u00b0, 54.94\u00b0 and 56.22\u00b0, which corresponds to reflection planes of (103), (006), (110), (007), (108), (205), (209), (217) and (2011), respectively and were well in line with standard (JCPDS # 00\u2013033-1340). Absence of any secondary peak confirmed the formation of single phase hexaferrite lattice. On increasing the dopant amount, the position of peaks shifted towards the lower diffraction angle, where the peak intensity was also declined (Fig. 2b). The observed trend might be endorsed to distortion in crystalline lattice, which may originate due to dissimilarity in cationic radius of host Fe3+ (64\u00a0pm) and Ni2+ cations (69\u00a0pm) [3]. The mean crystalline size (D) was determined using Eq. 2.\n\n(2)\n\n\nD\n\n=\n\n\n0.94\n\u03bb\n\n\n\u03b2\nc\no\ns\n\u03b8\n\n\n\n\n\nwhere, \u03bb and \u03b2 represent the wavelength used of X-rays used and full width at half maximum whereas \u03b8 is Bragg\u2019s angle. Cell volume (Vcell) was calculated from lattice constants \u2018a\u2019 and \u2018c\u2019 as shown in Eq. 3. The other structural parameters, i.e., theoretical density (\u03c1X-ray) and bulk density (\u03c1bulk) values were determined using Eqs. 4\u20135.\n\n(3)\n\n\n\nV\n\ncell\n\n\n=\n\n\na\n2\n\nc\n\n\n\n\n\n0.866\n\n\n\n\n\n\n\n\n\n\n(4)\n\n\n\u03c1\nx\n-\nr\na\ny\n\n=\n\n\n2\nM\n\n\n\nN\nA\n\n\nV\n\ncell\n\n\n\n\n\n\n\n\n\n\n(5)\n\n\n\u03c1\nb\nu\nl\nk\n\n=\n\nm\n\n\u03c0\n\n\nr\n\n2\n\nh\n\n\n\n\n\n\nWhere, M and NA denotes the molar mass and Avogadro\u2019s number, while m and r are the mass and radius and h is the height of pallet correspondingly.The average crystalline size was found to be in 34\u201350\u00a0nm range. Lattice constant \u2018a\u2019 remained constant, while \u2018c\u2019 was decreased by increasing the concentration of Ni because the ionic radius of dopant (Ni) is larger than the host (Fe) cations, which leads to expansion in Vcell\n[26]. The c/a ratio was less than 3.98, which indicated that the hexagonal structure was formed with P63/mmc space group [27]. The X-ray density decreased from 5.08 to 4.93 to g/cm3, whereas bulk density was increased from 2.127 to 3.421\u00a0g/cm3 with increase in Ni dopant (Table 1\n). The \u03c1X-ray was found to be greater than \u03c1bulk, which probably is an indication of certain amount of pores in the lattice, highly beneficial for to enhance the photocatalytic and optical properties of the fabricated material [3].\nFig. 3\na depicts the FTIR spectra of bare SFO and SFNO (1\u20135) in 4000\u2013400\u00a0cm\u22121 range. The two Fe\u2013O stretching vibrational bands at 447 and 593 (cm\u22121) at tetra and octahedral sites were appeared in 400\u2013800\u00a0cm\u22121 range. The octahedral band appears in 400\u2013500\u00a0cm\u22121 range, while the tetrahedral band originates in 500\u2013700\u00a0cm\u22121 range [28]. Considerable change in band position takes place, which might be ascribed to the variation in Fe+3\u2013O-2 and Ni+2\u2013O-2 bond lengths in doped lattice. Possibly the variation in magnetic dipole moment and ionic replacement, the band intensity decreases with increase in of Ni substitution [29]. The Fe-O bending vibrational band was evolved at 1129\u00a0cm\u22121, which was in agreement with already reported literature [27]. The \u2013OH anti-symmetric stretching band was appeared at 1384\u00a0cm\u22121. The deformational vibration and stretching vibration correspond to hydroxyl group that was detected at 1633\u00a0cm\u22121\n[30]. Band at 2337\u00a0cm\u22121 represents the deformation of water molecule. The band appeared near 3400\u00a0cm\u22121 might be associated with hydroxyl group due to presence of moisture contents [16].In order to explore the characteristic vibrational modes of fabricated materials, the Raman analysis was employed in 0\u2013800\u00a0cm\u22121 range (Fig. 3b). There are 64 atoms in the unit cell of SrFe12O19 where only 42 modes consisting of 11A1g\u00a0+\u00a014E1g\u00a0+\u00a017E2g were found as Raman active. Raman bands associated to octahedral site vibrations confirmed that the composites had magneto plumbite structure of strontium hexaferrite lattice. According to literature data used by Kreisel et al. [31], the peaks observed in Raman patterns have been ascribed to different vibrational bands. Bands appeared at 173 and 329 (cm\u22121) correspond to E1g symmetry of whole hexaferrite block and E2g vibrations [32,33]. Mode observed at 410\u00a0cm\u22121 can be associated to A1g vibrations at octahedral (12\u00a0k) sites [34]. The strong E2g band detected at 329, 529, 606 (cm\u22121) might be ascribed to Fe-O bond of FeO6 octahedral vibration [35]. The E2g band evolved at 606\u00a0cm\u22121 can be linked to stretching vibrations of Fe-O at 4f2 octahedral sites of hexaferrite lattice [36]. The band appeared at about 677\u00a0cm\u22121 might be present owing to vibrations of bipyramidal (2b) sites of A1g symmetry. For doped materials, a slight shifting in Raman peaks towards lower wavenumber was observed with increase in Ni doping content that is possibly due to difference in ionic masses of host Fe (55.85 amu) and Ni (58.71 amu) dopant [3].Ferroelectric (P-E) loops of SFO and SFNO (1\u20135) composites were measured at 300\u00a0K under \u221210\u201310\u00a0mV/cm electric field (Fig. 4\na). On applying the field, the\u00a0+\u00a0ve side of field showed increasing polarization, while \u2013ve side revealed a decrease in electric polarization. None of the prepared material attained saturation polarization (Psat), maximum polarization (Pmax). On increasing the Ni content, the Pmax, coercivity (Er) and remnant polarization (Pr) were increased. Maximum polarization and remnant polarization were increased from 4.06 to 10.98 mC/m2 and 4.0 to 10.6 mC/m2, respectively for highly doped material (Fig. 4b). The observed trend in substituted SFNOs might be due to that the one FeO6 octahedron within sub-unit of SrFe12O19 unit cell at 3 crystallographic sites, i.e., 4f2, 2a and 12\u00a0k wherever Fe is sited normally at octahedron center. When external field is applied the Fe cations demonstrate off-center shifting within the hexaferrite lattice, which probably prompt the electric polarization. Also, the enhanced polarization in substituted materials might be endorsed to Ni doping at Fe site where the cationic radius of Ni is slightly bigger than Fe, which may cause a variation in Fe-O and Ni-O bond length [3]. Pristine SFO indicates slightly less conductive behavior that seems to be decrease on increasing the Ni doping in crystallite lattice. Remnant polarization is less than Pmax for all the materials that is in well agreement with reported studies [3,20,21]. Such ferroelectric behavior of Ni doped SrFe12O19 indicates their potential utilization in various electronic devices [3].Field dependency of dielectric constant (\u03ad) for all the materials was analyzed in 20\u00a0Hz-1.5\u00a0GHz at 300\u00a0K. Dielectric constant was measured from capacitance values using Eq. 6.\n\n(6)\n\n\n\n\u03b5\n\u0307\n\n=\n\n\nCD\n\n\n\u03b5\n\noA\n\n\n\n\n\n\n\nWhere, C and D denote the capacitance and pallet thickness while \u03b5o and A are permittivity constant of free space and cross-sectional area of the pallet. For all the samples, at lower frequency, the dielectric constant was higher showing dispersion in this regime whereas in high frequency region, it remains almost constant with their lowest value (Fig. 5\na). The observed behavior might be ascertained to electronic polarization (induced due to fluctuation in oxidation states of metallic cations) as well as space charge type polarization. At high frequency, the constant and lowest \u03b5\u2032 value is possibly due to incapability of dipoles of dielectric material to follow the rapid changing AC field that starts to trail behind the field thereby resulting of drop in \u03ad. The dependence of frequency on \u03b5\u2032 might be described via Koop\u2019s polarization theory established on Maxwell and Wagner model of double layer [17]. According to this, a dielectric material contains bilayer structure where first layer having conducting grains of high conductivity is separated from second layer of less conductive grain boundaries. When electric field is applied, the electrons via hopping mechanism get arrived at boundaries where they accumulate due to high resistivity of grain boundaries. This charge accumulation at grain boundaries enhances the interfacial type polarization that intern rise the \u03b5\u2032 in low field. The electronic flow through grain boundaries inhibits at high frequency that results in drop of dielectric constant [3]. Fig. 5b illustrates the dielectric constant versus doping content at specific frequencies for pure and doped SFNOs. The SFNO5 shows highest \u03b5\u2032 than undoped SFO, which is possibly owing to reduction of B-site Fe cations that decreases the hooping rate among Fe+2 and Fe+3 cations thus increasing the \u03ad value [16,37].\nFig. 6\na demonstrates the frequency dependent dielectric loss (\u03b5\u201d) of pure SFO and Ni doped strontium hexaferrite different compositions (1\u20135). The dielectric loss shows similar trend to the dielectric constant that is it is high at low frequency. As the frequency was increased, it starts to drop and outside specific frequency, it responds rather independently from applied frequency. Such behavior can be associated to Maxwell\u2013Wagner polarization of interfacial type, which occurs in heterogeneous dielectric medium [3]. Fig. 6c demonstrates the variation of tan\u03b4 of hexaferrites versus applied field at room temperature. At about 100\u00a0kHz, the tan\u03b4 dropped down swiftly and later, it remains almost constant. This initial decline in loss tan might be described by Koop\u2019s phenomenological model. The dielectric losses in hexaferrite materials are normally represented in resistivity terms that is the dielectric materials with lower resistivity values display high dielectric loss and vice versa. Tangent loss showed inverse relation with Ni doping in SrFe12O19, which probably is due to low electron hopping rate among Fe2+ and Fe3+ state responsible for high conduction [16]. In ferrites, polaron hopping perhaps the basic reason for dielectric loss when frequency is high and electron hopping was the reason for dielectric loss at low frequency. The low dielectric losses in these materials are beneficial for their utilization in different high frequency applications [17].The impact of frequency on AC resistivity of doped and undoped Sr hexaferrite samples is shown in Fig. 7\na. Like dielectric constant, AC resistivity shows dispersion at low field, however, on increasing the frequency, resistivity of all material was dropped and at highest frequency, it was lowest with constant value. The basic reason for variation in resistivity might be that on octahedral sites of structure, the hopping rate of charge carriers takes place among Fe+2 and Fe+3 states. Hopping of charge carriers increases by the external field when the frequency is high. Electrical conduction is improved by the hopping of charge carriers as it leads to decrease in resistivity [3].In region of low frequency, the AC conductivity of all the compositions did not change and was low values, but gradually enhanced on increasing the applied field. In structure of ferrites, the conductivity is generally associated to octahedral sites where the hopping of electrons takes place among Fe+3 and Fe+2cations. At higher frequency region, it started to increase with frequency and attained its maximum value at highest AC field. Perhaps the degree of crystallinity, crystallite size and temperature of a material may affect the hopping phenomenon between localized states. AC conductivity was enhanced probably due to accessibility of hopping between electrical charges when the frequency was high. In higher frequency region, the free charges are able to move as they have sufficient energy and thus lead to enhance the conductivity process [17].To investigate the charge separation ability and recombination of photo-electron-hole pairs, the PL analysis was performed in 400\u2013800\u00a0nm range. At specific excitation wavelength, a strong emission peak was detected at 470\u00a0nm (Fig. 8\n). For pristine SFO, the sharp PL peak intensity indicates its fast recombination rate of charge carriers, while for doped compositions, the PL intensity dropped quickly showing their low recombination rate. In case of substituted materials, the low recombining rate increases the life time of the light induced carriers, which intern improves the optical efficiency of photocatalytic material [17]. The better separation ability of doped materials might be endorsed to O2 vacancies and creation of extra energy levels within the conduction band and valence band possibly owing to intrinsic defects on increasing the doping amount in hexaferrite lattice [38].The UV\u2013vis analysis of fabricated materials were studied to analyze the optical behavior and bandgap (Eg), which is highly useful to investigate the photocatalytic application. Fig. 9\na shows the UV\u2013vis diffuse reflectance spectra of pure SFO and SFNO(1\u20135) composites in 200\u2013800\u00a0nm range. Direct optical bandgap for all the composites was estimated by employing the Tauc\u2019s model (Eq. 7) by plotting h\u03c5 vs \u03b1h\u03c52.\n\n(7)\n\n\n\n\n\n\u03b1\nh\n\u03c5\n\n\n\n\n=\n\nk\n\n\n\n\n\nh\n\u03c5\n-\n\nE\ng\n\n\n\n\n\n\n1\n/\n2\n\n\n\n\n\n\nWhere, \u03b1, k and \u03c5 are absorption coefficient, Boltzmann constant and photon frequency of visible light used. Bandgap was estimated on extrapolation of linear part of plot at \u03b1\u00a0=\u00a00 (Fig. 9b). The bandgap for undoped strontium hexaferrite was 2.31\u00a0eV, while for doped composites, a substantial decline in Eg was observed. Bandgap energy was tuned from 2.11\u00a0eV (SFNO1) to 1.66\u00a0eV (for SFNO5) (Fig. 9b). The observed decline in bandgap for substituted materials might be explained on the basis of specific impurity levels that may appear within the forbidden energy band on increasing the Ni dopant in lattice, which accelerates the intensification of donor level overhead the prior valence band, while acceptor level under prior conductance band [39]. Other factors such as structural strain, variation in average crystallite size or surface area/volume might have definite consequence on optical bandgap [16,17].The catalytic capability of pure SFO and highly doped SFNO5 material was appraised for CV dye removal under visible light illumination. Fig. 10\n(a-b) demonstrates the visible light absorption curves of CV dye for SFO and SFNO5 materials recorded at specific time intervals. Pristine SFO showed a degradation of 55\u00a0%, whereas highly doped SFNO5 catalyst demonstrated superior efficiency of 91\u00a0% removal of CV in 90\u00a0min under the exposure of visible light irradiation. This better photocatalytic competence of SFNO5 material for degradation of CV dye might be endorsed to structural strain created on doping by metallic cations with different cationic size and charge, as it is most probable that a definite level of oxygen or cationic vacancies may be generated within the lattice to balance the electrical charge neutrality. Actually, the Ni doping tuned not only the Eg of substituted catalyst (SFNO5), but also minimize the recombination of e--h+ pair via acting as efficient trappers for photoactive charge carriers [3]. In comparison to previous studies (Table 2\n), it is concluded that the SrNixFe12-xO19 showed promising efficiency for the degradation of CV dye, which could have potential applications as a photocatalyst under solar light irradiation.For photo-degradation reaction (PDR) of CV dye, the apparent rate constant (kapp) was calculated from Langmuir\u2013Hinshelwood relation as shown in Eq. 8.\n\n(8)\n\n\n-\nl\nn\n\n\nA\nt\n\n\nA\n0\n\n\n=\nk\nt\n\n\n\n\nThe kapp was determined from slope of plot, i.e., -lnAt/Ao versus irradiation time (t). The linear fitting of obtained line with Adj R2\u00a0>\u00a00.95 confirmed that the degradation pathway of dye followed pseudo first order kinetics. The kapp values for PDR of CV dye over undoped SFO and SFNO5 catalysts under visible light irradiation were determined as 8.02\u00a0\u00d7\u00a010-3 min\u22121 and 2.337\u00a0\u00d7\u00a010-2/min, respectively (Fig. 11\na-b).In order to recognize the mechanism of charge transferring and to analyze the key radicals involved during the photo-degradation of CV dye over substituted SFNO5 catalyst, scavenging experiments were conducted. The AgNO3 and EDTA and 2-propanol (TP) were taken as trapping agents being scavengers of electron (e-), h+ and hydroxyl (\u2022OH) radicals, respectively. Without using scavenging agent, the percent removal of dye was 91\u00a0%. A radical drop in decolorization was noticed with TP, which showed the most active specie, hydroxyl (\u2022OH) radicals involved in the CV dye removal. A significant decrease in degradation was noted also with AgNO3 that indicated that h+ were the active agents that played imperative role in removal of CV dye over SFNO5 catalyst. Though, no substantial change in degradation was observed with EDTA, indicating no direct impact of e- towards the photo-gradation of dye. For above-mentioned scavengers, the percent degradation of CV dye was dropped to 22, 33 and 43 (%) showing the rate constant of 0.0009, 0.0017 and 0.0029 (min\u22121), separately. The experimental results showed that hydroxyl radicals were key active species which played the major part in decolorization of CV dye. The h+ also influenced the degradation of CV dye significantly, whereas e- showed slight contribution in degradation over SFNO5 (Fig. 12\na). The order of scavenging effect was found as, \u2022OH\u00a0>\u00a0h+ > e-.Degradation of dye effluents on photocatalysts are based on creation of electron-hole pair within the semiconducting material when it is irradiated to the light, which operates the oxidation- reduction processes of adsorbed species on the surface of photocatalyst [17]. These photo-induced e- are excited from the valance band (VB) and entered into conduction band leaving behind the h+ in VB. This e-/h+ pair on the surface of SFNO5 probably plays its crucial role for oxidation/reduction reactions involved during the degradation of CV dye (Fig. 13\n). The O2 (dissolved in water) in dye medium captures the photo-excited electrons, where super oxide (O2\n\u2022-) radical is produced. On the other hand, H2O interact with h+ and resultantly, generates OH\u2022 radical [16]. The OH\u2022 degrade the CV dye by oxidative process and convert them into CO2, H2O and inorganic ions. An overall CV dye degradation mechanism is presented in Eqs. 9\u201319.\n\n(9)\n\n\nSrN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n+\nh\nv\n\u2192\n\n\n\n\n\nS\nr\nN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n\n\n\n\ne\n\nCB\n\n-\n\n\n\n\n\n+\n\n\n\n\n\nh\n\nVB\n\n+\n\n\n\n\n\n\n\n\n\n\n(10)\n\n\nSrN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n\n\n\n\nh\n\nVB\n\n+\n\n\n\n\n\n\u2192\n\n\nS\nr\nN\n\ni\nx\n\n\nF\n\n12\n-\nx\n\n\n\nO\n19\n\n+\n\n\nH\n+\n\n+\nO\nH\n\n\n\n\n\n\n(11)\n\n\nSrN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n\n\n\n\nh\n\nVB\n\n+\n\n\n\n\n\n+\n\u2192\n\n\n\n\nS\nr\nN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n\n\n\n\nh\n\nVB\n\n+\n\n\n\n\n\n+\n\nO\n\nH\n.\n\n\n\n\n\n\n\n(12)\n\n\nSrN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n\n\n\n\ne\n\nCB\n\n-\n\n\n\n\n\n+\n\u2192\n\n\n\n\n\nS\nr\nN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n\n\n\n\ne\n\nCB\n\n-\n\n\n\n\n\n+\n\n\nO\n2\n\n\u2219\n-\n\n\n\n\n\n\n\n\n(13)\n\n\n\nO\n2\n\n\u2219\n-\n\n\n+\n\n\nH\n+\n\n\n\n\n\nH\n\nO\n2\n\u2219\n\n\n\n\n\n\n\n(14)\n\n\nH\n\nO\n2\n\u2219\n\n+\n\nH\n\nO\n2\n\n\u2192\n\n\n\n\n\n\nH\n2\n\n\nO\n2\n\n+\n\n\nO\n2\n\n\n\n\n\n\n\n(15)\n\n\nSrN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n\n\n\n\ne\n\nCB\n\n-\n\n\n\n\n\n+\n\n\nH\n2\n\n\nO\n2\n\n\n\n\u2192\n\n\n\n\n\nS\nr\nN\n\ni\nx\n\nF\n\ne\n\n12\n-\nx\n\n\n\nO\n19\n\n+\n\nO\n\nH\n\u2219\n\n+\n\nO\n\nH\n-\n\n\n\n\n\n\n\n(16)\n\n\n\nH\n2\n\n\nO\n2\n\n+\n\n\nO\n2\n\n\u2219\n-\n\n\n\u2192\n\n\n\n\nO\n\nH\n\u2219\n\n+\n\nO\n\nH\n-\n\n\n\n+\n\nO\n2\n\n\n\n\n\n\n\n(17)\n\n\n\nH\n2\n\n\nO\n2\n\n+\n\nh\nv\n\u2192\n\n\n\n\n\n2\nO\n\nH\n\u2219\n\n\n\n\n\n\n\n(19)\n\n\nDye\n\n+\n\nO\n\nH\n\u2219\n\n\u2192\n\n\n\n\nI\nn\nt\ne\nr\nm\ne\nd\ni\na\nt\ne\ns\n\n\n\n\n\n\n(19)\n\n\nIntermediates\n\u2192\n\n\nC\n\nO\n2\n\n+\n\n\nH\n2\n\nO\n\n+\n\ni\nn\no\nr\ng\na\nn\ni\nc\n\ni\no\nn\ns\n\n\n\n\nThe Ni doped SrNixFe12-xO19 (for\u00a0\u00d7\u00a0\u00a0=\u00a00.0\u20130.25) were fabricated through simple micro-emulsion method and influence of substitution on the dielectric, optical and photo-catalytic behavior was analyzed. Ferroelectric loops got widened on doping, which enhanced the maximum polarization, coercivity and remnant polarization. Dielectric parameters showed dispersion in low frequency region, but remained almost independent at higher frequencies. Bandgap was reduced from 2.31\u00a0eV (pure SFO) to 1.66\u00a0eV for highly doped SFNO5 composition. The substituted SFNO5 material showed much better photocatalytic efficiency regarding the removal of CV dye versus undoped SFO under visible light irradiation. Effect of various scavenging agent to assess the key active species for photodegradation of dye was studied, where OH\u2022 radicals were observed as major species involved in degradation process. The Ni doping affected the optical, conductivity and dielectric properties significantly of doped SrNixFe12-xO19. The photo-catalytic efficiency revealed a potential application for the photodegradation of dye in wastewater under visible light irradiation, which will be highly economical versus UV light based catalytic process.This research was funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project (Grant No. PNURSP2022R124), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.\nZunaira Irshad: Investigation, Writing \u2013 original draft. Ismat Bibi: Conceptualization, Supervision. Aamir Ghafoor: Methodology. Farzana Majid: Formal analysis. Shagufta Kamal: Project administration. Safa Ezzine: Validation. Zainab M. Elqahtani: Resources, Data curation, Funding acquisition. Norah Alwadai: Software, Writing \u2013 review & editing. Noureddine El Messaoudi: . Munawar Iqbal: Visualization, Writing \u2013 original draft, Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors express their gratitude to Princess Nourah bint Abdulrahman University Researchers Supporting Project (Grant No. PNURSP2022R124), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. The authors extend their appreciation to the Deanship of Scientific Research at King Khalid University, Saudi Arabia for funding this work through the Research Groups Program under grant number R.G.P.1: 255/43.", "descript": "\n                  A series of Ni doped SrNixFe12-xO19 (x\u00a0=\u00a00.0, 0.05, 0.10, 0.15, 0.20, 0.25) hexaferrites were fabricated through facile micro-emulsion approach and doping content effect was investigated based on structural, dielectric, optical and photocatalytic properties. The prepared materials were characterized via X-ray diffraction (XRD), Raman, Fourier transformed infrared (FTIR), photoluminescence (PL) and UV\u2013vis techniques. The Ni doped SrFe12O19 structure was hexagonal having space group P63/mmc with mean crystallite size in 34\u201350\u00a0nm range. Ferroelectric polarization, coercivity and Remnant polarizations were increased with Ni doping. Dielectric analysis showed the higher values of dielectric parameters with dispersion in low frequency region, which were independent of at high frequency. Optical bandgap was reduced on substitution, which was in 2.31\u20131.66\u00a0eV range, which was in good association with decline in PL analysis. The PCA (photocatalytic activity) of pure SrFe12O19 (SFO) and doped SFNO5 (x\u00a0=\u00a00.25) material was assessed for removal of crystal violet (CV) dye under visible light illumination. The SFNO5 showed much better PCA and 91\u00a0% dye was removed in 90\u00a0min with a rate constant of 2.337\u00a0\u00d7\u00a010-2 min\u22121 versus pristine SFO (only 55\u00a0% with rate constant 8.02\u00a0\u00d7\u00a010-3 min\u22121). Results revealed that tuned bandgap and enhanced AC conductivity of doped materials make it crucial candidates for optoelectronic and SOFCs (Solid oxide fuel cell) applications. Owing to the excellent PCA and magnetically separable nature, SrNixFe12-xO19 has potential for the dyes removal from the effluents under solar light irradiation, which could be more economically versus UV based photocatalytic process.\n               "}
{"full_text": "Data will be made available on request.In recent years, there has been a concerted effort to decrease the reliance of modern consumer products on the fossil fuel industry and the need for a sharp decline in their use to keep the temperature increase below 1.5\u00a0\u00b0C (Welsby et al., 2021). One of the main products from the petroleum industry are chemical feedstocks, which represents around 10\u00a0% of the global petroleum industry in terms of production volume, but a significantly higher financial value due to the increased price of petrochemicals compared to fuels. In particular, BTX are used to manufacture common plastics like poly ethylene (PE), polystyrene (PS), polyurethanes (PUL), and nylon (NY) among other chemicals (\u201cBiogreen Energy Syngas,\u201d 2020). In consequence, there is growing interest in replacing fossil derived with bio derived feedstocks, although it has been suggested that despite biomass is an \u201cindispensable\u201d resource for the circular economy, combusting it to electricity only is inefficient and the path to chemicals should be pursued (Hamer, 2020). Accordingly, lignin as renewable source of aromatics is an important element in achieving the \u2018net-zero emission by mid-century\u2019 target outlined by The Paris Agreement (Griffin et al., 2018). Lignin possesses a phenolic polymer structure comprised of p-hydroxy-phenyl, guaiacol, and syringyl groups, and hence has good potential to provide renewable aromatic compounds for future use as feedstocks to the chemical manufacturing industry. Aromatic production could be an economical way of kick-starting large-scale lignocellulosic pyrolysis as a means to reduce dependency on oil derived chemicals and help promote renewable biomass as a staple industry in the global economy.Fast pyrolysis, which is the rapid heating of biomass in absence of oxygen for a short residence time followed by rapid quenching of the produced vapours can be used to produce either fuels or aromatics, with appropriate variations in process conditions, which includes the use of catalysts (Liu et al., 2021a; Farooq et al., 2022; Hendry et al., 2020). On an economic level, biomass pyrolysis has been assessed for fuel production. A report published in 2015 by the US Department of Energy shows an economic projection of how advanced bio-fuels can be made cost competitive vs fossil fuel derived transportation fuels (Dutta et al., 2015). The conclusion of this report was that biomass pyrolysis can compete with fossil fuel derived fuels based on a total yield of combined gasoline and diesel equivalent fuels of 78 gallons / US ton dry of biomass and a fuel price of around $3.50 /gallon final product. The report outlines that the total product yield should be minimum 25\u00a0wt% of dry feedstock for the whole process, including ex-situ HDO upgrading of pyrolysis vapours.Recent studies show that pyrolysis under hydrogen atmosphere operating at a rapid heating rate namely hydropyrolysis (HyPy) is a promising technology for converting biomass into liquid fuels (e.g., bio-oil and C4\u00a0+\u00a0hydrocarbons), since the addition of H2 enhances the H/C ratio of the bio-oil, reduces the O/C ratio, facilitates hydrotreating reactions that involve CC coupling, hydrocracking, alkylation, decarboxylation, decarbonylation, hydrogenation, HDO, and recombination (Oh et al., 2021; Stummann et al., 2021). In particular, HDO of oxygenated aromatics (e.g., phenolic compounds) occurs via hydrogenation of the aromatic ring followed by deoxygenation (Liu et al., 2021b). The second advantage is that the H\u2022 radicals \u2018stabilise\u2019 the highly reactive intermediates species (e.g. ketone and aldehyde) and prevent them from condensing into coke, which reduces the carbon recovery and lead to higher operational costs (Resende, 2016). Presence of hydrogen also enhances demethylation of methoxy groups in phenolic products and favours dehydration vs C loss pathways (\u2013CO, \u2013CO2). For example, non-methoxy phenolics reached 19.68C% at 30\u00a0bar and 500\u00a0\u00b0C, where monocyclic and polycyclic aromatic hydrocarbons as well as condensable aliphatic hydrocarbons were observed from lignin hydropyrolysis (Wang et al., 2022). Temperature plays an important role in hydropyrolysis, where HDO takes place at increasing temperatures to with efficient elimination of furanic and phenolic oxygen-containing compounds at T\u00a0>\u00a0500\u00a0\u00b0C (Wang et al., 2013). However, Tian et al (2021) showed that when hydropyrolysis temperature is increased from 700 to 800\u00a0\u00b0C and higher, the hydropyrolysis oil (from pine sawdust) decreased from 35\u00a0wt% to 25\u00a0wt% at expenses of gas phase (from 50 to 70\u00a0wt%) (Tian et al., 2021).The majority of studies show that hydrogen atmosphere (either at pressure or ambient) with no catalyst can be sufficient to provide the desired bio-oil composition and vaporisation before the HDO step (Zheng et al., 2017). This pathway demands a good understanding of the HyPy products so that a suitable HDO catalyst can be selected. For BTX production, the pyrolysis vapour should consist of simple phenols and the HDO catalyst should selectively produce BTX molecules by removing the oxygen content (Jan et al., 2015). Some studies show HDO catalysts further cracking the bio-oil phenols into small olefins or alkane gas molecules, which have lower value than aromatic BTX compounds (Marker et al., 2013). Additional reduction of BTX molecules can promote char build-up and catalyst deactivation, which is undesirable.There are two main catalyst types used in the hydropyrolysis of lignocellulosic biomass, the pyrolysis catalyst and the HDO catalyst, where the most well-known conversion processes include both of them, such as the integrated hydropyrolysis and hydroconversion (IH2\u00ae)) and a two-step biofuel process (H2Bioil) (Venkatesan et al., 2020). The pyrolysis catalyst is primarily used to increase the hydrogenation of the biomass molecules and promote the breaking of the weakest CC bonds in the structure which in turn, breaks down the lignin and cellulose polymer chains into oligomers, dimers and monomers (Sirous-Rezaei and Park, 2020). The most efficient catalysts for this are either powered metal oxide catalysts on a fluid bed reactor or acid site zeolite-based structures. Zeolites are relatively expensive to prepare and have several well-documented shortcomings when used in biomass pyrolysis, namely char build-up, fast deactivation and high regeneration energy, suggesting they may not be suitable for large scale processes, although they promote aromatics production (Stummann et al., 2018; Zhu et al., 2022; Jindal et al., 2022).In the literature the main success in HDO catalysts for BTX production have been in metal oxides on a porous, surface-active supports under high H2 pressure (Resende, 2016; Stummann et al., 2021; Venkatesan et al., 2020; Wang et al., 2013). Metal oxide reducing powders appear to be a better catalyst due to their higher reactivity, which does not depend on molecule mobility rates through zeolite pores, as well as regeneration being easier and faster than that of zeolites (Jan et al., 2015). Palladium is a precious metal in the same chemical group as Nickel and Platinum and is used widely in industry as a hydrogenation catalyst, usually for the reduction of CC or CO bonds in organic chemistry. Jan et al published a study in 2015 that showed very high yields for BTX from lignin samples using in-situ palladium doped HZSM-5 zeolite (Jan et al., 2015). The reported yield was 40\u00a0wt% aromatics at 600\u00a0\u00b0C and 17\u00a0bar of H2. However, the catalyst was used with a catalyst to biomass ratio of 10:1 or 20:1, with the 1:1 resulting in only 4\u00a0wt% aromatics produced, with C6-C10 aromatic and polyaromatic molecules counting in that total. BTX consisting of C6 and C7 aromatics made up approximately 18\u00a0wt% with 20:1 catalyst to biomass ratio. The high catalyst to biomass ratio would cause problems when scaling up these experiments as the assumed reason for such high catalyst ratio is the high tendency for palladium and zeolites to quickly form char that leads to rapid deactivation. Stummann et al. (2018) studied the hydropyrolysis of beech wood at 26\u00a0bar and 450\u00a0\u00b0C in presence of \u2018bog-iron\u2019 as a hydropyrolysis catalyst and nickel-molybdenum alloy supported on aluminium oxide as the HDO catalyst. This study showed 24.7\u00a0wt% of condensed organic and C4\u00a0+\u00a0molecules, which can be used as gasoline additive (Stummann et al., 2020). This is important as one of the reaction paths to aromatics from hydropyrolysis is through C4\u00a0+\u00a0intermediate species (Norinaga et al., 2014).\nSirous-Rezaei and Park (2020) studied the HyPy/HDO of kraft lignin using HY as in-situ catalyst and Fe/HBeta, FeReOx/MCM-41, Fe/ZrO2 and FeReOx/ZrO2 as ex-situ HDO catalyst, under a continuous flow of H2 at 1 atmosphere pressure (Sirous-Rezaei and Park, 2020). While FeReOx/ZrO2 resulted the most efficient HDO catalyst (7.1\u00a0wt% aromatic hydrocarbons of which 4.8\u00a0wt% BTX), all FeReOx/MCM-41, Fe/ZrO2 and FeReOx/ZrO2 led to significantly lower yield of coke compared to a zeolite-supported catalyst like Fe/HBeta. Similar aromatics (7.6 wt.%/5.4\u00a0wt% BTX) were obtained using palladium rhenium oxide (Sirous-Rezaei and Park, 2020). The lower cost ferrous rhenium catalyst material and low reaction pressure is significant because it represents a relevant advantage if this process is scaled up. The products from the iron-based catalyst also gave 11.4\u00a0wt% oxygenates of which 8.7\u00a0wt% was phenol. These compounds are usually the primary targets for HDO catalysts suggesting phenols deoxygenation can be improved. The same research group showed the efficiency of ferrous rhenium oxide on zirconium oxide support in the upgrading and deoxygenating of mono-phenols into BTX (Starting material \u2013 BTX wt%: Guaiacol \u2013 21.6\u00a0wt%, m-Cresol \u2013 61.7\u00a0wt%, Anisol \u2013 48.3\u00a0wt%) (Sirous-Rezaei et al., 2018). However, the starting materials for these studies were pure lignin/phenols without any additional species that could interfere with the catalyst action (e.g. cellulose fraction). An early research show that Etek (Etanol Teknik) lignocellulosic filtration residue from acid straw hydrolysis resulted in 34\u00a0% gas, 61\u00a0% liquid and 5\u00a0% coke (down from 19.6\u00a0% in N2) when underwent hydropyrolysis at 800\u00a0\u00b0C with both lignin and cellulose derived compounds accounted in the liquid product (Windt et al., 2009).Overall, the literature shows that integrated Hy-Py/HDO is the most effective pathway for deoxygenating pyrolysis bio-oil and also that bifunctional zirconia supported metal catalysts such as FeZrO2 can turn lignin into valuable aromatics under mild hydropyrolysis conditions. However, selectivity on BTX is a unresolve challenge. Therefore, this study wants to investigate in-house synthetised ZrO2 supported Ce, Na Fe, PdFe catalysts in the integrated Hy-Py/HDO of Etek lignin waste under ambient H2 pressure at 600\u00a0\u00b0C using a pyroprobe reactor coupled with a gas chromatograph/mass spectrometer (Py-GC/MS), with the aim to (i) maximise BTX selectivity on oil product, (ii) deoxygenation of the Hy-Py bio-oil and (iii) compare the effect of hydrogen and the selected catalysts on the products distribution with those obtained in presence of nitrogen.Etek (Etanolteknik AB, SE) lignin is an industrial filtration residue remaining after a two stages weak acid hydrolysis of soft wood and consists of 41\u00a0% holocellulose and 59\u00a0% lignin (Nowakowski et al., 2010). The Etek lignin contained 51\u00a0wt% C, 5.7\u00a0wt% H, 1.6\u00a0wt%\u00a0N and 37.7\u00a0wt% O (by difference), while the proximate analysis gave 76.6\u00a0wt% volatiles, 19.4\u00a0wt% fixed carbon and 4\u00a0wt% ash. Na2CO3 (99.6\u00a0% purity), CeO2 and ZrO2 (99.0\u00a0% purity), ZrOCl2\u00b78H2O (99.5\u00a0% purity) and PdCl2 and FeN3O9x9H2O used for the synthesis of the catalysts were all purchased from Sigma-Aldrich.Na/ZrO2 (1.5_1), CeNa/ZrO2 (1_1_1), CeNa/ZrO2 (2_1_1) were synthetized mixing by pestle & mortar Na2CO3, CeO2 and ZrO2 and then calcining the mixture at 900\u00a0\u00b0C for 4\u00a0h in air. ZrO2 support for Fe/ZrO2 and PdFe/ZrO2 was instead prepared by dropwise addition of ZrOCl2\u00b78H2O to water (100\u00a0mL). After addition, the suspension was aged 20\u00a0h at 90\u00a0\u00b0C, dried in an oven at 110\u00a0\u00b0C for 15\u00a0h and subsequently calcined in flowing air (30\u00a0mL/min) at 500\u00a0\u00b0C for 3\u00a0h. 5\u00a0%Fe and 1\u00a0% Pd on ZrO2 were then prepared by incipient wet impregnation method with PdCl2 and FeN3O9x9H2O as precursors. The dried resulting salts were then calcined in air at 500\u00a0\u00b0C for 3\u00a0h to achieve good dispersion (Hendry et al., 2020).A -micro-pyrolysis reactor coupled with a gas chromatography/mass spectrometer (Py-GC/MS) was used for the experimental campaign. The system consisted of a CDS 5250 (Chemical Data Systems, USA) pyroprobe hyphenated to a GC\u2013MS Trace DSQ II (Thermo Scientific, USA). Fig. 1\n schematises the reactor setup, consisting in an autosampler quartz tube (1.9\u00a0mm diameter) where one mg of the Etek lignin was sandwiched between 2 layers of 5\u00a0mg of selected catalyst and the two materials were kept in place and separated by layers of quartz wool.The reaction tests were performed at 600\u00a0\u00b0C (heating rate of 10\u00a0\u00b0C/ms) in presence of hydrogen (100\u00a0mL/min) with final temperature maintained for 30\u00a0s. Although the true temperature in the microprobe is likely 75-100\u00b0 C lower than the set point (Thangalazhy-Gopakumar et al., 2011). The temperature of 600\u00a0\u00b0C was selected since literature indicates that this temperature was required to maximise recovery of phenolics from Etek lignin pyrolysis (Hendry et al., 2020). The volatiles generated in the pyroprobe were collected in a tar trap for separating the non-condensable gases and further analyse them by GC\u2013MS. The tar trap consisted in a 1/8\u2033 Tenax kept at 40\u00a0\u00b0C. The condensed species were then desorbed from the tar trap at 300\u00a0\u00b0C for 3\u00a0min and sent to a GC\u2013MS for analysis by a transfer line heated at 350\u00a0\u00b0C. The bio-oil was injected in the GC at 280\u00a0\u00b0C, using 25\u00a0mL/min of helium, as a carrier, and the split ratio of 80:1. The column used was an Agilent: HP-5MS, 19091S-433; length, 30\u00a0m; internal diameter, 250\u00a0\u03bcm; film thickness, 0.25\u00a0\u03bcm The GC oven temperature started at 40\u00a0\u00b0C for 2\u00a0min, heated up to 320\u00a0\u00b0C at 12\u00a0\u00b0C/min and kept at temperature for 15\u00a0min. The raw area percentages were recalculated by excluding the GC\u2013MS peaks after 20\u00a0min (see supplementary material) and normalising to 100\u00a0%. This was done because those peaks belonged to contaminants leached out from the GC column at high temperature. The peaks were finally identified by the NIST library. The Total Sum Normalization (TSN) method for normalising GC\u2013MS data, which consists in dividing the area of each peak in each chromatogram by the total sum of all peaks within that chromatogram, was used in this work. Sum normalised data were multiplied by 100 and expressed in terms of their percent contribution (area%) to the total area. To evaluate the catalysts\u2019 ability on steering the HyPy/HDO reactions to produce BTX in the oil product from the starting lignin fraction in Etek lignin, the term selectivity was used. The product was studied in term of functional groups using Excel. Non-aromatic molecules were classified in the following order: cycloalkene, cycloalkane, epoxide ring, esters, ketones\u00a0+\u00a0aldehydes and olefins as summarised in Table 1\n including a brief description of the main uses. Alcohol groups was a very common feature so were ignored in terms of classifications. Hence there are molecules in the cycloalkene class that are also ketones and have ester groups on branches off the cycloalkene ring. 2-Dodecenal is an example of an olefin chain with an aldehyde group but was classified in the ketone\u00a0+\u00a0aldehyde group and not the olefin group as per the order above. The BTX content based on starting Etek lignin weight was estimated considering (i) the bio-oil wt% resulting from Etek lignin hydropyrolysis in absence of catalyst (61.2\u00a0wt% based on dry lignin) under similar conditions (ambient pressure H2, 800\u00a0\u00b0C) (Windt et al., 2009); (ii) a fraction of bio-oil detectable by GC\u2013MS (\u223c16.4\u201340\u00a0%) (Windt et al., 2009); (iii) the fraction of HDO oil (23\u00a0wt%) obtained from lignocellulosic material under similar conditions (450\u00a0\u00b0C, H2 at ambient pressure and 5\u00a0%Pd catalyst\u00a0+\u00a0NiMo/Al2O3) (Gholizadeh et al., 2016) and (iv) the fraction of BTX from the GC\u2013MS analysis (63\u00a0%) assuming this is proportional to the actual weight content of BTX.Proximate and ultimate analysis of the Etek lignin were obtained by using an Exeter CE-440 Elemental Analyser and a Mettler Toledo TGA2 thermogravimetric analysis, respectively. The catalysts characterisation included XPS spectra that were acquired using a PHI Quantera II Scanning XPS Microprobe -instrument. Samples were sputter-cleaned using argon ions prior to analysis. Calibration of spectra was done using the C 1\u00a0s peak (284.8\u00a0eV) for adventitious carbon. SEM-EDX analysis was done using a Zeiss Leo 1530 microscope equipped with a FEG and operated at a voltage of 2.5\u00a0kV and equipped with an Oxford Instruments X-Max silicon drift detector (SDD) on non-coated samples. TEM analyses were performed using a Titan Themis 200 scanning/transmission electron microscope (S/TEM) equipped with an X-FEG Schottky field emission gun and a Super-X high sensitivity windowless EDX detector complemented by a Gatan Enfinium EELS Detector. The catalyst samples for TEM were ultrasonically dispersed in ethanol and then deposited on carbon-coated copper grids using capillary and dried in air for 30\u00a0min. XRD was instead run using a Bruker Nonius X8-Apex2 CCD equipped with an Oxford Cryosystems Cryostream (typically operating at 100\u00a0K), and an X-ray source with a Cu anode working at 40\u00a0kV and 40\u00a0mA, and an energy-dispersive one-dimensional detector. The Fe amount on the Fe/ZrO2 catalyst was determined by atomic absorption spectroscopy (AAS) using a PerkinElmer Analyst 100, while surface analysis was carried out using a Tristal II Plus Micromeritics analyzer. H2-TPR was carried out using a lab-made instrument (Bagnato et al., 2020). The catalyst was first dried at 110\u00a0\u00b0C for 18\u00a0h. and then heated at 10\u00a0\u00b0C/min from 25\u00a0\u00b0C to 900\u00a0\u00b0C in a 5\u00a0% H2/He flow (40\u00a0mL/min). The effluent gases were analysed by a TCD detector. Surface acidity was analysed by NH3-TPD using an AutoChem II system (Micromeritic, USA).\nTable 2\n shows the HyPy oils\u2019 chemical composition from the GC\u2013MS analysis together to the biomass component each group is originated from. The HyPy oil product using CeNa/ZrO2 (1_1_1) (see supplementary material) contained 107 distinct compounds. This shows the low selectivity of the catalyst, with Ketones\u00a0+\u00a0Aldehydes (11.04\u00a0%), and Guaiacols (22.25\u00a0%) being the most abundant functionalities. The effect of hydrogen vs nitrogen during the Etek lignin pyrolysis was compared (see supplementary material) using the data produced in a previous work (Hendry et al., 2020). The high HyPy temperature used (600\u00a0\u00b0C) was necessary to maximise lignin break-down into monomers, which would then be converted into hydrocarbon via hydrogenation, decarbonylation, deoxygenation pathway generating fuel grade hydrocarbons (Hendry et al., 2020; Jindal et al., 2022). Clearly, the distribution of the bio-oil compounds shifted to lower molecular weight compounds, such as butanone and cyclopentadiene, although the most abundant compounds in N2 atmosphere (e.g. phenol; phenol-2-ethoxy; 3-creosol; levoglucosan) remained qualitatively unchanged when H2 was used. Very limited benzene was detected suggesting that both Ce and Na are not good in hydrogenating and deoxygenating the substituted aromatic rings. The GC\u2013MS analyser from the hydropyrolysis reactions in presence of the same catalyst but with two moles of Ce instead of one [CeNa/ZrO2 (2_1_1)] identified 46 distinct compounds in the oil product. This shows an increased peak area of the catalyst over CeNa/ZrO2 (1_1_1), with the 20 most abundant molecules making up 87.73\u00a0% of the total bio-oil, most of which presented a lower retention time compared to CeNa/ZrO2 (1_1_1) and larger content in anhydrosugars derived from cellulose, suggesting that Ce is more effective in breaking down cellulose into small oligomers/monomers including some cycloalkenes and cycloalkanes. Presence of Ce/ZrO2 (1_1) resulted in 81 distinct compounds, mostly Oxygenated sugars (34.7\u00a0%), Guaiacols (23.5\u00a0%) and Esters (7.66\u00a0%), indicating its propension in converting cellulose-derived components than CeNa/ZrO2 (2_1_1). The GC\u2013MS analysis in presence of Na/ZrO (1.5_1) identified 60 distinct compounds with Ketones\u00a0+\u00a0Aldehydes (23\u00a0%) and Guaiacols (42.3\u00a0%) being the most abundant. Basic sites of the catalyst can favour reduction of acids and deoxygenation via ketonization and aldol condensation reactions (Stefanidis et al., 2016). In comparison, when Na/ZrO3 (1.5_1) was pyrolysed in presence of N2, the GC\u2013MS resulted in 47 distinct compounds mainly Guaiacols (62.4\u00a0%) (Hendry et al., 2020). This suggests that the catalyst had good reactivity to both the cellulose and lignin components of the biomass in presence of H2. Olefin intermediates quickly reform into larger molecules, eventually leading to polyaromatic hydrocarbons and char (Norinaga et al., 2014). The presence of H2 gas stabilises the olefin intermediates and stops further reactions that produce coke on the catalysts\u2019 surface. This also explains why the tests in presence of H2 showed much higher cellulose derived compounds.The GC\u2013MS analysis with 1\u00a0wt% Pd and 5\u00a0wt% Fe on ZrO2 identified 62 distinct compounds, which are much closer to the desired products (i.e. deoxygenated hydrocarbons) than Na/ZrO2 and the other two Na/Ce catalysts. Ketones\u00a0+\u00a0Ketone Alcohols, Cycloalkenes, Guaiacols, and BTX were the main functionalities with total area from each being 12.06\u00a0%, 18.67\u00a0%, 7.88\u00a0% and 16.31\u00a0%, respectively. The presence of polyaromatic hydrocarbons (5.74\u00a0% - highest in any of the catalytic tests) shows that olefins created from cellulose reduction are contributing to the end products and indicate that species are over-reduced and forming char deposits on the catalytic acid sites, limiting further catalytic activity (Stefanidis et al., 2016). The higher area% of cycloalkenes and lower yield of Ketones\u00a0+\u00a0Aldehyde shows that PdFe/ZrO2 is more active and more reducing than Na/ZrO2. Likewise, the 16.3 area% of BTX and low relative area% in Alkylphenols and Guaiacols shows the reducing factor is equally applied to the lignin component as well as the cellulose component.The HyPy bio-oil obtained using Fe/ZrO2 showed the best characteristics among all the employed catalysts. The GC\u2013MS analyser identified only 11 distinct compounds in the product flow from the hydropyrolysis- test, which was duplicated for consistency. Ketones (acetone), Cycloalkenes and BTX were the most abundant functionalities with total area from each being respectively: 15.3\u00a0%, 13.5\u00a0%, and 66.8\u00a0% (35.4\u00a0% benzene, 25.6\u00a0% toluene and 5.8\u00a0% xylene). This shows a very high selectivity of the catalyst compared with the other catalysts. FeZrO2 also resulted in a carbon distribution of 85.5\u00a0% C5-C10 (10.5\u00a0% C5, 41.1\u00a0% C6, 27.6\u00a0% C7 and 6.3\u00a0% C8, 3.3\u00a0% C10), which was similar to the 84\u00a0% aromatic hydrocarbons (C6-C11) obtained by the Hy-Py/HDO of pine sawdust at 500\u00a0\u00b0C and 20\u00a0bar over a hydrocracking catalyst (20:1 cat:biomass wt ratio) bed maintained at 500\u00a0\u00b0C (Venkatesan et al., 2020). They also showed that at 1\u00a0bar, \u223c63\u00a0% aromatics (34\u00a0% BTX of which 12.8\u00a0% B., 13.8\u00a0% T. and 5.8\u00a0% X.) and significant amount of Poly Aromatic Hydrocarbons (PAH) (\u223c16\u00a0%) were produced at 500\u00a0\u00b0C, which confirm the higher selectivity of FeZrO2 on BTX under the studied conditions. Similarly, Agarwal et al. (2017) produced alkylphenolics (17\u00a0wt% on lignin) and aromatics (4\u00a0wt% on lignin) for a total of 34\u00a0wt% lignin oil, from the hydrotreating of Kraft lignin at 450\u00a0\u00b0C, 100\u00a0bar H2, 4 hrs using Fe rich limonite catalyst (Agarwal et al., 2017). To compare the BTX yield with those in literature, the weight % of the BTX (based on starting Etek lignin) was estimated (1.5\u20133.6\u00a0wt%) resulting in between those obtained using FeZrO2 (1.3\u00a0wt%) and PdReOx/ZrO2 (5.4\u00a0wt%) under lower temperature (350\u00a0\u00b0C), where oxygenated phenolics were the predominant oil components (Sirous-Rezaei and Park, 2020). However, proper BTX quantification would be required.to confirm this. Moreover, the deoxygenation power resulted the best with only about 5.5\u00a0wt% O in the HyPy oil compared to 15.5\u00a0wt% for FePd/ZrO2, based on GC\u2013MS HyPy oil composition and larger for the other catalysts, which confirms the selective recovery of C ad H in the HyPy oil.This study proposes a reaction pathway leading to BTX considering the already proposed mechanisms for the cracking and reforming of biomass. To evaluate if the main products distribution from the HyPy of Etek lignin was dictated by the thermodynamical favourability, the main reactions for three representative model compounds (phenol and 2-methoxy-phenol for lignin and glucose for cellulose) were simulated using HSC Chemistry 5.11, Outokumpu between 300 and 800\u00a0\u00b0C, at ambient pressure. For phenol and 2-methoxy-phenol the thermodynamic viability of hydrodeoxygenation to benzene (R1), cyclohexane (R4), toluene (R6) and methyl-cyclohexane (R7), the hydrogenation to cyclohexadiene (R2), cyclohexanol (R3), methyl-cyclohexanol (R8) and the methanation reactions (R5 and R9) were evaluated. For the cellulose fraction of Etek lignin instead, the HDO of glucose to 1,3-cyclopentadiene (R10), acetone (R11) and 3-cyclopentene 1,2-diol (R12) were considered as listed below:\n\n(1)\nR1: C6H6O\u00a0+\u00a0H2\u00a0=\u00a0C6H6\u00a0+\u00a0H2O\n\n\n\n\n(2)\nR2: C6H6O\u00a0+\u00a02H2\u00a0=\u00a0C6H8\u00a0+\u00a0H2O\n\n\n\n\n(3)\nR3: C6H6O\u00a0+\u00a03H2\u00a0=\u00a0C6H12O\n\n\n\n\n(4)\nR4: C6H6O\u00a0+\u00a04H2\u00a0=\u00a0C6H12\u00a0+\u00a0H2O\n\n\n\n\n(5)\nR5: C6H6O\u00a0+\u00a010H2\u00a0=\u00a06CH4\u00a0+\u00a0H2O\n\n\n\n\n(6)\nR6: C7H8O\u00a0+\u00a0H2\u00a0=\u00a0C7H8\u00a0+\u00a0H2O\n\n\n\n\n(7)\nR7: C7H8O\u00a0+\u00a04H2\u00a0=\u00a0C7H14\u00a0+\u00a0H2O\n\n\n\n\n(8)\nR8: C7H8O\u00a0+\u00a03H2\u00a0=\u00a0C7H14O\n\n\n\n\n(9)\nR9: C7H8O\u00a0+\u00a011H2\u00a0=\u00a07CH4\u00a0+\u00a0H2O\n\n\n\n\n(10)\nR10: C6H12O6\u00a0+\u00a03.6H2\u00a0=\u00a01.2C5H6\u00a0+\u00a06H2O\n\n\n\n\n(11)\nR11: C6H12O6\u00a0+\u00a04H2\u00a0=\u00a02C3H6O\u00a0+\u00a04H2O\n\n\n\n\n(12)\nR12: C6H12O6\u00a0+\u00a04.8H2\u00a0=\u00a01.2C5H8O2\u00a0+\u00a03.6H2O\n\n\n\u0394G, \u0394H, \u0394S and the equilibrium constant (Kc) were calculated as function of temperature and the resulting log Kc for the above reactions are shown in Fig. 2\n, where values of Kc larger than zero indicate spontaneity and larger the number greater the thermodynamic viability. As can be seen for the two phenolic species, in general lower the temperature more thermodynamically favourable are the reactions R1 to R10 (due to exothermicity of all these reactions) and only the methanation and HDO to benzene and toluene are thermodynamically favoured at temperature between 500 and 600\u00a0\u00b0C as used in this work, with the first being more favourable. This is in agreement with the experimental findings with the exception of cyclohexadiene (R2), where only benzene and toluene were detected. Typically, the phenolic ring is hydrogenated to form cyclohexanol or cyclohexanone at low temperature (as supported by the Kc in Fig. 2) and this explains the absence of these products in this work. The HDO reactions of glucose (R10-12) are instead endothermic and therefore favoured at higher temperature as shown in Fig. 2, where it can be seen that they are thermodynamically favourable under the studied temperature. If acetone (R11) is common to both Fe/ZrO2 and PdFe/ZrO2, the HDO to 1,3-cyclopentadiene occurs only in presence of Fe/ZrO2 suggesting that Fe is more prone to deoxygenating cellulose. Due to the reducing nature of the H2 atmosphere, it is expected that cellulose is reduced to non-condensable gases as shown by V.K. Venkatakrishnan et al (2014), with the remaining contribution from cellulose being reduced to simple chain molecules like Ketones, Aldehydes, Olefins as well Cyclopentenes (Venkatakrishnan et al., 2014). Recently, Li et al (2022) showed that cellulose hydropyrolysis under 25\u00a0bar hydrogen and 500\u00a0\u00b0C mainly produced C5-C7 ketones with carbon yield of 27.2\u00a0% (Li, Miao, et al., 2022). In particular, C5-C7 chain ketones and cyclic ketones were generated by furfural through HDO and hydrocracking reactions. The olefin-radical pathway for cellulose to BTX conversion is not prevalent in H2 atmosphere because the presence of H2 will protonate and stabilise the radicals and prohibit further aromatisation reactions from cellulose derivatives. However, it is assumed that cyclopentadiene is formed via this route.Various reaction pathways have been previously proposed to explain the mechanism for lignin hydropyrolysis and bio-oil vapour HDO upgrade into BTX (Jan et al., 2015; Liu et al., 2019; Sirous-Rezaei et al., 2018; Sirous-Rezaei and Park, 2020). It involves the hydrogenation of the lignin structure during pyrolysis into phenolic monomers, including guaiacols, which then are reduced to BTX. At high temperature as used in this work, phenol deoxygenation occurs either via decarbonylation or dehydration of the alcohol group (RC), resulting in mono-aromatic hydrocarbons (MAHs). Over acid sites, these MAHs may undergo alkylation, cyclization, minor rates of hydrogenation to form cycloalkanes and polymerization (RD) to form naphthalenes and poly-aromatic hydrocarbons (Gamliel et al., 2018). Based on previous literature, it is expected that coking reactions took place at 600\u00a0\u00b0C (Gholizadeh et al., 2016).The GC\u2013MS analysis clearly indicates that FeZrO2 can selectively produce BTX from the HyPy/HDO of Etek lignin under the studied conditions. Therefore, the properties of this catalyst were examined by different techniques to elucidate the reasons beyond it and correlate them to the proposed reactions mechanismss. AAS confirmed that the Fe content in the FeZrO2 catalyst was 4.98\u00a0wt%, while the surface analyses resulted in a BET surface of 65\u00a0m2/g, a pore volume of 0.26\u00a0cm3/g and a BJH average pore size of 20\u201330\u00a0nm (see supplementary material). The relatively large volume of mesopores represent an important factor for the promotion of hydrotreating reactions, since it favours the diffusion of large molecules (e.g. guaiacols) into the pores where they can then react in the active sites. Guaiacol HDO in presence of Pd/meso-ZSM-5 catalyst exhibited superior guaiacol conversion and product distribution when compared with Pd supported on conventional microporous ZSM-5, due to the improved diffusion and accessibility of active sites inside meso-ZSM-5 (Wang et al., 2020). X-ray powder diffraction pattern of Fe-doped ZrO2 catalyst (see supplementary material) appears as a very well crystalline mixture of different zirconium oxide phases. ZrO2 monoclinic patters typical at ambient temperature (2-\u03b8 of 28.2\u00b0, 31.5\u00b0, 38.5\u00b0, 50.1\u00b0 and 59.8\u00b0) are visible, although additional XRD peaks (30.4\u00b0, 31.4\u00b0 and 35.4\u00b0) suggest the presence of other ZrO2 phases. ZrO2 has been shown to be a good support for hydropyrolysis reactions due to its thermal stability and surface acidity (Li, Su, et al., 2022), XPS spectra of the FeZrO2 catalyst were taken (see supplementary material). The Fe 2p binding energy (BE) bands indicate presence of Fe3O4 (711 and 715\u00a0eV) and Fe2O3 (725\u00a0eV) (Bagnato et al., 2020), while the Zr3d bands indicate presence of ZrO2 (182.5\u00a0eV and 183.9\u00a0eV) and ZrOx suboxide (180.1, 182.5\u00a0eV). Finally, the O1s spectra show the oxygen corresponding to ZrO2 (531 and 532.8\u00a0eV), ZrOx (530.6\u00a0eV) and Fe species (528.9\u00a0eV). SEM-EDX (see supplementary material), which was used to establish the overall size of the particles and the content in iron, suggests particles in the 10\u201350\u00a0nm size with presence of some larger agglomerates and confirm that the weight content of iron close to 5\u00a0wt% as per synthesis. TEM (see supplementary material) confirms that the FeZrO2 particles are of 20\u201330\u00a0nm size, but do not allow for a clear distinction between Zr and Fe due to poor contrast difference, although the darker area of\u00a0\u223c\u00a05\u00a0nm can be assigned to Fe. It is well known that a good dispersion of metals on the support surface is related to good catalytic activity (Rhodes et al., 2005). Therefore, to evaluate the dispersion of Fe on the ZrO2 support, TEM-EDX (see supplementary material) was analysed. The TEM-EDX show a good dispersion of the Fe on the ZrO2 support and confirmed that the Fe nanoparticles are well dispersed in the rage of 5\u00a0nm, which can be linked to the catalyst performance. In this context, the low selectivity of PdFe vs Fe could be linked to the poor dispersion of Pd as shown by TEM-EDX (see supplementary material). The acidity, basicity and H2 reducibility of FeZrO2 were also analysed. The H2-TPR (see supplementary material) showed an H2 uptake of 280\u00a0\u03bcmol/g with reduction to Fe0 occurring between 450 and 600\u00a0\u00b0C with peak at 550\u00a0\u00b0C, which confirm that reported in a previous work (Bagnato et al., 2020; Hendry et al., 2020; Spreitzer & Schenk, 2019). NH3-TPD resulted (see supplementary material) in an acidity of about 260\u00a0\u00b5mol NH3/g, which consisted of weak and mild acid sites denoted by desorption of NH3 from 100 to 400\u00a0\u00b0C, while FePdZrO2 had considerably lower acidity as shown in previous work (Bagnato et al., 2020). These findings support the fact that under the studied conditions PdFeZrO2 performance was not good as the mono metallic FeZrO2 catalyst and suggest that the selectivity and deoxygenation capability of FeZrO2 is due to the synergic activity of weak and mild acid sites in bonding the carbonyl group and the capacity of Fe of adsorb H2 at high temperature and hydrogenating the aromatic ring. It was recently shown that aromatic rings are repulsed by oxyphilic Fe/ZrO2, which instead adsorb and hydrogenate the carbonyl group followed by deoxygenation in the ZrO2 acid sites (Yung et al., 2019; Li, Miao, et al., 2022)). Moreover, Kumar et al. (2021) studied the influence of the support ZrO2, for Co based catalyst at different reaction temperature, 300\u2013600\u00a0\u00b0C, at 1\u00a0bar for the hydropyrolysis of prot lignin (Kumar et al., 2021). ZrO2 resulted in 29.8\u00a0% H-type phenolics at 600\u00a0\u00b0C due to enhanced demethoxylation. In addition, the Fe performance can be ascribed to a better distribution of active Fe inside the support mesopores and the ability of monoaromatics to enter the mesopores to reach the closely located Fe and ZrO2 active sites. A similar effect was observed in the hydrotreating of guaiacol with Ni-ZrO2 on mesosilica, where mesopores facilitated confinement of Ni and ZrO2 nanoparticles, which proximity was considered necessary for achieving a high selectivity into deoxygenated products HDO (Lopez et al, 2020).Therefore, the remarkable catalytic activity and selectivity of Fe/ZrO2 was linked to the iron oxophilicity, the strong reduction potential of zero-valent iron, the good dispersion of Fe nanoparticles on the support and the presence of mesopores and Lewis acid sites on the zirconium oxide support, which increased the interactions between lignin derived phenolic molecules and the catalyst surface. Fe/ZrO2 showed also good potential to also reduce cellulose-derived molecules to acetone and cyclopentadiene. The above suggest that Fe/ZrO2 can be used for hydrodeoxygenation of biomass-derived oxygenates under the tested conditions, which is crucial to achieve a cost-effective scale-up of the process.The hydropyrolysis and hydrodeoxygenation of Etek lignin was studied at 600\u00a0\u00b0C in presence of zirconia supported metal catalysts under ambient pressure. Fe/ZrO2 led to considerably enhanced HDO of lignin-derived phenolics to aromatic hydrocarbons in comparison with the other catalysts. The Fe/ZrO2 mild-strong reduction sites provided- good reducing activity, selectively hydro-deoxygenating carbonyl groups in phenolic and guaiacol compounds to benzene (35.4\u00a0%), toluene (25.6\u00a0%) and xylene (5.8\u00a0%), as the main products with a carbon distribution of 85.5\u00a0% in C5-C10 hydrocarbons.\nWilliam Lonchay: Investigation, Formal analysis, Visualization, Validation, Writing \u2013 original draft. Giuseppe Bagnato: Investigation, Validation, Writing \u2013 review & editing. Aimaro Sanna: Conceptualization, Methodology, Supervision, Validation, Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by Innovate UK/KTN (KTP project no. 10013135). We kindly acknowledge Joe Perkins-Hall and Karen Sam, CDS-Analytical for Py-GC support and Aaron Naden, Department of Chemistry, University of St. Andrews for TEM-EDX analysis.Supplementary data to this article can be found online at https://doi.org/10.1016/j.biortech.2022.127727.The following are the Supplementary data to this article:\n\nSupplementary Data 1\n\n\n\n", "descript": "\n                  The use of lignin to produce Benzene, Toluene and Xylene (BTX) is a promising pathway to strength the economic case, over the production of advanced bio-fuels alone. In this work, Ce, Na, Pd and Fe supported on zirconium oxide were evaluated for the ex-situ hydropyrolysis (HyPy)/hydrodeoxygenation (HDO) of Etek lignin under mild conditions (600\u00a0\u00b0C, 1 atmosphere) towards the production of BTX. Fe/ZrO2 was able to selectively produce BTX (67 area%) and cycloalkenes (13.5 area%) and strongly deoxygenate the HyPy oil to about 5\u00a0wt% oxygen content, resulting in an oil with a carbon distribution of 85.5\u00a0% in C5-C10 hydrocarbons. The high selectivity of Fe/ZrO2 was related to the iron oxophilicity, the strong reduction potential of zero-valent iron, the good dispersion of Fe nanoparticles on the support and the presence of mesopores and acid sites, which enhanced the interactions between the reacting species and the catalyst surface.\n               "}
{"full_text": "No data was used for the research described in the article.The development of catalytic converters for neutralization of exhaust gases has a history of almost 50\u00a0years and is constantly accompanied by an increasing stringency of emission standards for hazardous atmospheric pollutants CO, NOx and CHx [1]. To meet these standards at today's level, significant improvements in the performance of three-way catalysts (TWCs) are required. One of the main challenges in this regard is to solve the problem of a cold-start when the catalytic converter and exhaust gases have a temperature below the ignition temperature. The maximum emission of hazardous pollutants is observed during the first 30\u00a0s from engine start [2,3]. One approach to solving this problem is to localize the catalyst in the closest position to the engine, i.e. close coupled position. This reduces the time to reach ignition temperature to 10\u201315\u00a0s [1]. However, during engine operation the temperature in this zone varies in a wide range and reaches 1000\u00a0\u00b0C and more. This poses new challenges for development of catalysts with increased thermal stability while maintaining high activity at low temperatures. [4,5]. The use of natural gas, containing mostly methane, as a clean fuel for natural gas fueled vehicles (NGVs) and for power generation in gas turbines is becoming increasingly common as an alternative to conventional diesel or gasoline fuel [6\u20138]. Over the past 10\u00a0years, the number of NGVs worldwide has more than doubled [6]. The increased use of NGVs is related to lower emissions of pollutants, such as carbon monoxide and volatile organic compounds, compared to conventional engines [9,10]. However, the residual unburned methane in exhaust gases poses a great potential threat to the environment because of its strong greenhouse effect, which is \u223c30 times greater than that of CO2. Therefore, an effective aftertreatment system is required to reduce the fraction of residual methane [11]. Complete catalytic oxidation of methane has become the main technology for purifying the exhaust gases of NGVs [12]. This process is complicated due to high strength of the CH bond in the methane molecule, low temperature of exhaust gases, not exceeding 500\u00a0\u00b0C, and high concentration of water in the exhaust gases, poisoning the catalyst [11,12]. Under these conditions, Pd-based catalysts proved to be the most active [13]. To prevent the decomposition of active PdO into less active metallic Pd and to minimize the high-temperature formation of NOx and SOx, it is necessary to lower the methane oxidation temperature. [7,14]. However, the exhaust temperature may rise to 800\u2013850\u00a0\u00b0C during fast driving. [7]. To maintain high low-temperature activity under these conditions, it is necessary to increase the thermal stability of both the support and the highly dispersed oxidized forms of palladium, which exhibit the greatest activity in methane oxidation.Ceria is an indispensable component of modern TWC catalysts. First of all, ceria serves as an OSC component to minimize the drop of conversion when the composition of the reaction mixture fluctuates around the stoichiometric air/fuel ratio [1]. An important feature of CeO2 is the accumulation and release of oxygen under redox conditions, which are realized during the operation of TWC catalysts. This is achieved due to the Ce4+-Ce3+ redox couple, which has the ability to change from Ce4+ (CeO2) under oxidizing conditions to Ce3+ (Ce2O3) under reducing conditions and vice versa [15]. However, the thermal stability of cerium oxide is not sufficient, so the doping of CeO2 with transition metal ions is used to increase thermal stability and preserve OSC. The most studied is the modification of CeO2 by Zr4+ ions in a certain concentration (ZrO2\u00a0\u2264\u00a033\u00a0mol%), which does not change the fluorite structure of Ce1-xZrxO2 solid solution [16], but increases the oxygen capacity and thermal stability [17,18].The use of tin for CeO2 doping has been studied in a number of works and has shown to provide a significant increase of OSC in series of CexSn1-xO2-\u03b4 oxides with low amount of tin (x\u00a0\u2265\u00a00.8) [19\u201322]. Doping with tin was also found to prevent the growth of mixed oxide crystallites leading to increase of the specific surface area [19,23,24]. The enhancement of OSC as a result of modification by tin is explained by an increase in the concentration of defects in the mixed oxide lattice. This leads to the formation of additional oxygen vacancies associated with the redox couple Sn4+/Sn2+ (along with the Ce4+/Ce3+ pair) [19,25,26] promoting increased oxygen mobility. The authors of [19] also observed a lower reduction temperature of mixed oxides, which was attributed to the lower oxygen binding energy in Ce-O-Sn than in Ce-O-Ce or Sn-O-Sn.The doping of SnO2 with cerium also leads to increase in the thermal stability of the SnO2-based catalysts. The SnCe binary oxides are characterized by a larger specific surface area and a smaller particle size than undoped SnO2 [27], which determines the dispersion of deposited Pd and thus affects the activity [28]. Generally, the activity of Pd/SnO2 catalysts in CO oxidation reaction is significantly lower than the activity of Pd/CeO2 and Pd/Ce-Sn-O catalysts. The onset of the CO oxidation on Pd/SnO2 is observed at temperatures above 100\u2013170\u00a0\u00b0C [28,29], whereas in the case of Ce-based and Ce-Sn-based catalysts the CO oxidation starts at room temperature and even below. At the same time, Sn-rich catalysts (SnO2 phase) have a larger specific surface area compared to Ce-rich catalysts (CeO2 phase) [30]. A lower reduction temperature was observed for Sn-rich catalysts, which the authors attribute to the higher oxygen mobility in Sn-rich catalysts due to the coexistence of surface and bulk oxygen vacancies. In contrast to Sn-rich catalysts, in Ce-rich catalysts oxygen vacancies exist predominantly on the surface [30]. For Ce-doped tin dioxide nanoparticles an increase in thermal stability and resistance to sintering at high temperature is observed. In this case the thermal stability is explained by the processes of cerium segregation on the surface of tin oxide particles with the formation of a Ce-rich surface layer responsible for particle growth inhibition [31]. It should be noted that Ce-doped SnO2 in Sn-rich catalysts shows higher activity in methane oxidation than Sn-doped CeO2 in Ce-rich catalysts [30]. On the other hand, the introduction of small amounts of Sn into CeO2 lattice increases the activity of the catalysts in CO oxidation [19].Pd and Pt being widely known for their high activity in oxidation of CO and hydrocarbons, are used together with cerium oxide in TWC catalysts [1,32\u201336]. The deposition of Pd on cerium oxide promoted with transition metal ions can increase the thermal stability and preserve the active component in a highly dispersed state due to the strong metal-support interaction. Increased thermal stability and catalytic performance in CO oxidation for Pd-Ce-Sn-O co-oxides have been reported in [22,29,37\u201339]. A study of Ce-rich and Sn-rich Pd/Sn1-xCexO2 catalysts showed that the Ce-rich Pd/Sn0.2Ce0.8O2 sample has a higher CO oxidation activity than the Sn-rich Pd/Sn0.8Ce0.2O2 and non-promoted Pd/CeO2 and Pd/SnO2 catalysts [37]. Compared to the non-promoted SnO2 and CeO2 oxides, the binary CeSn oxides have smaller crystallite sizes and larger surfaces. The authors concluded that the active forms of oxygen in Ce-rich and Sn-rich supports are lattice and chemisorbed oxygen, respectively. However, to accurately establish the nature of oxygen and the mechanisms of the catalytic reaction the additional studies are required, including investigation of catalysts with different compositions of mixed oxides.Pd/CeO2-SnO2 catalysts with varying Ce/Sn ratio showed high activity in methane oxidation [40]. The highest activity is achieved by Pd deposition on Sn0.9Ce0.1O2 support calcined at 1100\u00a0\u00b0C. However, in this work the catalyst was calcined only at 450\u00a0\u00b0C that doesn't give an opportunity to judge about preservation of activity under operating conditions at high temperature. Therefore, the study of Pd/CeO2-SnO2 catalysts with varying Ce/Sn ratio calcined at high temperatures is promising for advancing the development of thermostable methane oxidation catalysts.In our previous work [29] the Pd/Ce-Sn-O catalysts synthesized by the counter precipitation method with an equimolar Ce/Sn ratio were studied. It was shown that the calcination of samples at temperatures of 800\u20131000\u00a0\u00b0C leads to an increase in catalytic activity in CO oxidation reaction at low temperatures below 150\u00a0\u00b0C. It was shown that upon calcination above 600\u00a0\u00b0C a nanoheterophase catalyst structure was formed, which is a catalytically active PdxCe1-xO2-\u03b4 dispersed phase on the surface of SnO2 nanoparticles. The formation of such structure in Pd/Ce0.5Sn0.5O2 catalyst provided high thermal stability up to 1000\u00a0\u00b0C and preserved active PdOx clusters stabilized on the surface of PdxCe1-xO2-\u03b4 solid solution. Despite the relationship between increase in activity and formation of nanoheterophase structure of the catalysts detected during thermal activation, it remains unclear whether the increase in reaction rate compared to unmodified catalysts is a consequence of the specific spatial structure of the catalyst, or whether modification of the active center of the catalysts occurs.This work presents a detailed study of the local structure of active centers of Pd/Ce(Sn)O2 catalysts using XRD, TEM, XPS, Raman methods and is aimed at establishing the relation between the catalyst structure and catalytic properties. For this purpose we studied the effect of Ce/Sn ratio in Pd/CeO2-SnO2 catalysts calcined at high temperature of 800, 900 and 1000\u00a0\u00b0C on physico-chemical properties and activity in CO and methane oxidation.The Pd/CeSn samples with varying Ce/Sn ratio were synthesized by counter precipitation method used in our previous work [29]. The initial (NH4)2[Ce(NO3)6] was synthesized from Ce(NO3)3\u00b76H2O (JSC Reaktiv, analytical grade) according to Ushakov et al. [41]. [Pd(H2O)2(NO3)2] was obtained from metallic Pd (Krastsvetmet, 99.9%) in compliance with Khranenko et al. [42]. Na2[Sn(OH)6] was prepared from SnCl4\u00b75H2O (Reachem) by dissolving in a 10% excess of NaOH with subsequent crystallization. Purity of the products was verified by XRD.(NH4)2[Ce(NO3)6] and [Pd(H2O)2(NO3)2] were dissolved in water or in an aqueous solution of HNO3 to obtain an acid solution (30\u00a0ml). Na2[Sn(OH)6] was dissolved in water or an aqueous solution of NaOH (50\u00a0ml). The amounts of NaOH and HNO3 were calculated for the complete hydrolysis of all components. The amount of [Pd(H2O)2(NO3)2] was calculated so as to equalize the molar ratio Pd/(Ce\u00a0+\u00a0Sn) to the molar ratio Pd/Ce in a reference sample 1wt%Pd/CeO2 (0.0163) synthesized by the procedure described in [29]. The amounts of the reagents for the synthesis of 4\u00a0g of catalyst and Pd loading (wt%) are given in Table 1\n.The resulting solutions were rapidly mixed with each other which led to the neutralization and precipitation of hydroxides of all the metals. The reaction mixture was then cooled to \u221215\u00a0\u00b0C and held for a day at this temperature. The precipitate was isolated on a filter at 2\u00a0\u00b0C and repeatedly washed with cold water and then with acetone. The obtained precipitates were dried in an oven at 80\u00a0\u00b0C for 4\u00a0h. After that, the catalyst samples were placed in the preheated furnace and calcined for 2\u00a0h in air at 450, 600, 800, 900, and 1000\u00a0\u00b0C.The synthesized catalysts are denoted as Pd/CeSnX-T, where T is the calcination temperature, X describes the atomic ratio of the support components and is calculated from atomic fraction of Sn (X\nSn) and Ce (X\nCe) in the catalyst by the following eq. X\u00a0=\u00a0X\nSn / (X\nSn\u00a0+\u00a0X\nCe)\u00b7100.In all cases, atomic absorption spectrometry (AAS) did not detect palladium in mother liquors and rinsing waters. CeO2 and SnO2 are chemically inert compounds and therefore it is hard to prepare solutions for analysis. To determine the ratios of elements we dissolved the hydroxide precipitate, which was obtained after the mutual neutralization, in 6\u00a0M hydrochloric acid. The results of ICP-MS analysis (with Quadrupole ICP-MS iCAP spectrometer) confirmed the ratio of the metals to be close to initial loadings. The atomic ratio Pd:Ce is 0.017:1 for Pd/Ce reference catalyst (1.04\u00a0wt% in terms of \u201cPd\u00a0+\u00a0CeO2\u201d stoichiometry) and Pd:(Ce\u00a0+\u00a0Sn) is 0.0158:1, 0.0163:1, 0.0163:1 and 0.0163:1 for Pd/CeSn75, Pd/CeSn50, Pd/CeSn25and Pd/CeSn15 series, correspondingly.An X-ray diffraction study (XRD) was carried out on a Shimadzu XRD\u20137000 diffractometer (Cu K\n\n\u03b1\n radiation, Ni filter on the reflected beam). Diffraction patterns were recorded in a stepwise mode with the accumulation time necessary for recording reflections of the phases that are present in small amounts. The refinement of lattice constants and quantitative phase analysis were performed by Rietveld full-profile analysis using PowderCell2.4 software [43]. The crystallite sizes (D) were calculated on the base of Scherrer Eq. [44] after exclusion of the instrumental contribution. For the fluorite phase, the DFluorite value was calculated from reflex 220 (2\u0398\u00a0=\u00a047.5\u00b0); for the rutile phase, the DRutile value was taken as the average of that calculated from reflexes 200 (2\u0398\u00a0=\u00a038.0\u00b0) and 111 (2\u0398\u00a0=\u00a038.9\u00b0). Taking into account the results of work [45] the independence of the integral broadening method on the size and shape distribution of the samples was applied for calculations. The procedure of deconvolution and fitting of the XRD lines was performed using the WinFit 1.2.1 software [46].The catalysts microstructure was studied by transmission electron microscopy (TEM) using a Thermo Fisher Scientific Themis Z double Cs-corrected electron microscope operated at an accelerating voltage of 200\u00a0kV. The spectrum imaging data were obtained using a Super-X G2 EDX detector and a HAADF detector for image registration in scanning (STEM) mode. Crystal lattices on the obtained (S)TEM images were analyzed using the Fourier method. The samples were dispersed by ultrasound in ethanol and deposited on standard copper grids covered with a holey carbon film.The X-ray photoelectron spectra were obtained using an Ultra Axis DLD (Kratos Analytical, UK) and ES-300 (Kratos Analytical, UK) photoelectron spectrometers. The analyzers were calibrated relatively Au 4f\n7/2 and Cu 2p\n3/2 lines with standard binding energies 84.0\u00a0eV and 932.7\u00a0eV, respectively [47]. Mg K\n\n\u03b1\n source with h\u03bd\u00a0=\u00a01253.6\u00a0eV was used as a primary radiation. Spectral data acquisition was carried out in the regime of constant pass energy of the analyzer. The decreased X-ray source power 65\u00a0W was used to prevent ceria photoreduction during spectra recording. The calibration of experimental spectra was performed using the U'''-component of the Ce 3d line, the binding energy of which was taken equal to 916.7\u00a0eV [48]. The spectra processing was carried out using the XPSCalc software developed at the Boreskov Institute of Catalysis SB RAS, approved for films and powder catalytic systems [49\u201351]. The spectra decomposition into individual components was described by the Gaussian-Lorentzian distribution with subtraction of the background of inelastically scattered electrons by Shirley model. The quantitative composition was calculated from areas of the lines taking into account the atomic sensitivity factors [47].Raman spectra were obtained using an InVia (Renishaw, UK) confocal Raman dispersion spectrometer equipped with a Leica microscope with a 50\u00d7 objective. Excitation was performed with continuous lasers: a semiconductor laser with 785\u00a0nm and 100\u00a0mW and a solid-state Nd:YAG laser, second harmonic, 532\u00a0nm, 100\u00a0mW. To prevent sample heating, only 10% of the maximum laser intensity was employed together with 50% defocusing mode and signal accumulation time up to 100\u00a0s. The Raman spectra were measured in the 100\u20133200\u00a0cm\u22121 range with a spectral resolution of 2\u00a0cm\u22121 and 1\u00a0cm\u22121 upon excitation at 532 and 785\u00a0nm, respectively. Data are reported for the range of 100\u20131000\u00a0cm\u22121, in which important changes were observed in the spectra.An automated setup with a flow reactor [50] was used for investigation of catalytic properties in temperature-programmed reaction with CO (TPR-CO), in temperature-programmed CO oxidation reaction (TPR-CO\u00a0+\u00a0O2) and in temperature-programmed CH4 oxidation reaction (TPR-CH4\u00a0+\u00a0O2). The concentrations of CO (m/z\u00a0=\u00a012, 28), CO2 (m/z\u00a0=\u00a012, 28, 44), O2 (m/z\u00a0=\u00a032), H2 (m/z\u00a0=\u00a02) and CH4 (m/z\u00a0=\u00a013) were measured with aid of the quadrupole mass spectrometer RGA 200 (SRS). Neon was used as the reference inert standard for precise calculation of the component concentrations in the reaction mixtures. Digital mass-flow controllers for individual gases and mass-spectrometer operate at room temperature.In the TPR-CO experiments the reaction mixture containing 1.0\u00a0vol% CO, 0.5\u00a0vol% Ne and helium the balance was introduced at a flow rate of 100\u00a0cm3/min to the catalyst sample (0.2\u00a0g) preliminary cooled in the reactor to \u221220 (\u221240) \u00b0C. As the steady-state concentrations of CO and CO2 were established, the sample was heated from \u221220 (\u221240) to 450\u00a0\u00b0C at 10\u00a0\u00b0C/min heating rate. The concentrations of CO, CO2, O2, H2 and H2O were measured during the reaction. Before each TPR-CO experiment, the catalysts were pre-treated by 20% O2/He gas mixture at 450\u00a0\u00b0C during 2\u00a0h with subsequent cooling in this mixture and with subsequent purging with helium.The activity in TPR-CO\u00a0+\u00a0O\n\n2\n and TPR-CH\n\n4\n\u00a0+\u00a0O\n\n2\n was measured using the following experimental conditions. The catalysts grain size was 0.14\u20130.25\u00a0mm. The catalysts weight and catalysts volume were 0.2\u20130.3\u00a0g and 0.25\u00a0cm3 for TPR-CO\u00a0+\u00a0O2 accordingly. The reaction mixture velocity and GHSV were 1000\u00a0cm3/min and 240,000\u00a0h\u22121, accordingly. A 0.05\u00a0g of catalyst was diluted by quartz to 0.25\u00a0cm3 for TPR-CH4\u00a0+\u00a0O2. For CH4 oxidation the reaction mixture and GHSV were 100\u00a0cm3/min and 120\u00a0L\u00a0g\u22121\u00a0h\u22121, accordingly. For CO oxidation initial gas composition contains 0.2\u00a0vol% CO, 1.0\u00a0vol% O2, 0.5\u00a0vol% Ne, and helium the balance. For CH4 oxidation initial gas composition contains 0.1\u00a0vol% CH4, 1.0\u00a0vol% O2, 0.5\u00a0vol% Ne, 0 (10.0) vol% H2O and helium the balance. 10\u00a0vol% of steam (H2O) were introduced by flowing the helium through a temperature-controlled saturator. The catalysts were cooled to \u221220\u00a0\u00b0C if the experiments were carried out in a CO oxidation reaction. If experiments were carried out in CH4 oxidation, the initial temperature was 50\u00a0\u00b0C.The experiments were carried out using repeated cycles of the catalysts heating/cooling in the reaction mixture in the temperature range from \u221210 (+50) \u00b0C to 450 (600) \u00b0C at constant heating rate 10\u00a0\u00b0C/min for CO (CH4) oxidation. Changes in the concentrations of CO (CH4), O2 and CO2 during the reaction were monitored by measuring at a frequency of 0.34\u00a0Hz.Rates of the CO\u00a0+\u00a0O2 or CH4\u00a0+\u00a0O2 reactions were calculated using the initial sections of TPR\u00a0\u2212\u00a0CO\u00a0+\u00a0O2 (TPR-CH4\u00a0+\u00a0O2) curves in the region of conversions not higher than 20% under differential reactor operation, where diffusion limitations and changes in the reaction mixture composition over the catalyst bed can be neglected (Fig. S1). As it is reported in previous work [29] the internal and external diffusion effects and heat transfer effect could be neglected during the kinetic experiments. The specific reaction rate was determined per catalyst surface using the formula W(molecules/cm2\u00a0\u00d7\u00a0s)\u00a0=\u00a0C\n\n0\n\u00a0\u00d7\u00a0X\u00a0\u00d7\u00a0V\nRM / m\u00a0\u00d7\u00a0S\n\nsp\n, where C\n\n0\n is the initial concentration of CO (or CH4) (molecules/cm3), X is the CO conversion, V\nRM is the space velocity of the reaction mixture (cm3/s), m is the catalyst weight (g), and S\n\nsp\n\n\u2013 specific surface area (cm2/g). The weight reaction rate was determined per 1\u00a0g of Pd using the formula W(molecules/g(Pd)\u00a0\u00d7\u00a0s)\u00a0=\u00a0C\n\n0\n\u00a0\u00d7\u00a0X\u00a0\u00d7\u00a0V\nRM / m, where C\n\n0\n is the initial concentration of CO (or CH4) (molecules/cm3), X is the CO conversion, V\nRM is the space velocity of the reaction mixture (cm3/s), and m is the weight of Pd (g).The catalytic properties of Pd/CeSn samples with various content of tin were studied in the CO oxidation reaction. The previously discovered tendency towards an increase in the catalytic activity of Sn-modified catalysts with calcination temperature extends to the entire studied range of Ce/Sn ratio. Comparison of the weight rates (per 1\u00a0g Pd) of CO oxidation for catalysts calcined in the temperature range of 450\u20131000\u00a0\u00b0C (Fig. S2) showed that the activity of the catalysts increases significantly with an increase in the calcination temperature up to 800\u00a0\u00b0C and above. In this regard, the main emphasis in this work is placed on a detailed study of catalysts calcined at temperatures of 800, 900, and 1000\u00a0\u00b0C.In our previous work [29] Pd/CeO2-SnO2 catalysts with eqimolar Ce/Sn ratio were investigated. In the present work, the synthesis of a catalyst with a ratio Ce/Sn\u00a0=\u00a01 was reproduced. Fig. S3 shows the dependences of CO conversion for two different samples of 1%Pd/CeSn50\u2013800. From the presented data, we can see that the dependences of CO conversion are almost identical, which indicates a good degree of reproducibility of the synthesized samples.\nFigure 1\n shows the dependences of T\n50 (temperature of 50% CO conversion) and the weight rates (per 1\u00a0g Pd) of CO oxidation at 25\u00a0\u00b0C on the catalyst calcination temperature. The temperature dependences of CO conversion and the Arrhenius dependences of the CO oxidation rate are shown in Fig. S4. The presented data shows a dome-shaped dependence of the T\n50 values on the relative tin content. The minimum T\n50 value is attained in the case of Pd/CeSn50 and Pd/CeSn25 calcined at 800\u2013900\u00a0\u00b0C, while the maximum values are observed for the Pd/CeSn15 and Pd/CeSn75 catalysts. The dome-shaped dependences are also observed for the rate of CO oxidation on Pd/CeSnX with rate maxima at X\u00a0=\u00a050 (T\ncalc\u00a0=\u00a0800\u20131000 \u043eC) and X\u00a0=\u00a025 (T\ncalc\u00a0=\u00a0800\u2013900 \u043eC). The activity of the Pd/CeSn25 catalyst is similar to that of the Pd/CeSn50 catalyst after calcination at 800\u2013900\u00a0\u00b0C. After calcination at 1000\u00a0\u00b0C the activity of Pd/CeSn25 and Pd/CeSn75 catalysts becomes significantly lower than that of Pd/CeSn50. Thus, the activity of the catalysts strongly depends on the amount of tin oxide introduced and the activation temperature. At the same time, the observed dome-shaped dependence on the composition may be related both to the implementation of different catalyst morphologies and to the formation of new active centers on the surface. It can be assumed that the nature of the active oxygen in these centers differs significantly, determining its reactivity.The temperature-programmed reaction with CO was used to determine the reactivity of oxygen in the catalysts. The shape of the CO consumption curves versus temperature and the temperature profile of the CO2 evolution are related to the presence of oxygen with different bonding energies in the catalyst. Fig. S5 shows a typical TPR-CO experiment showing the time dependence of CO consumption, evolution of CO2, O2, H2, H2O, and temperature as well. A description of the experiment is provided in the Supplementary Material. Based on previous work [29,51\u201353], the CO2 release at temperatures above 200\u00a0\u00b0C is associated with the interaction of CO with bulk oxygen of CeO2 and SnO2 particles or particles of their solid solutions based on fluorite and rutile phases. The presence of Pd in these phases leads to a decrease in the temperature of interaction of oxygen with CO due to activation of oxygen in these phases and its diffusion to the palladium active sites (PdOx, PdO) on the catalyst surface [54]. A sharp peak of CO2 evolution in the temperature range of 180\u2013200\u00a0\u00b0C is associated with the interaction of CO with oxygen of PdO particles [29,51\u201353]. The evolution of CO2 in low-temperature region of \u221240\u2013100 \u00b0\u0421 takes place due to interaction of CO with oxygen of PdOx(s)/Pd-Ce-O(s) clusters.As it can be seen from the TPR-CO data in Fig. 2\n, the main release of CO2 occurs in the range of 150\u2013450\u00a0\u00b0C. According to Table S1 and Fig. S6, the total amount of reactive oxygen varies significantly depending on the tin content in the samples. The data presented in Table S1 and Fig. S6 show that calcination at 800\u00a0\u00b0C produces the maximum amount of reactive oxygen in all samples studied. However, considering all calcination temperatures, the amount of reactive oxygen passes through a maximum depending on the concentration of tin in the samples. The calcination at 800\u00a0\u00b0C yields the maximum amount of oxygen for the equimolar Ce/Sn ratio, while for higher calcination temperatures of 900\u00a0\u00b0C and 1000\u00a0\u00b0C the maximum amount of oxygen is observed for the Pd/CeSn25 sample.TPR-CO curves of Pd/CeSn25\u20131000, Pd/CeSn75\u2013900 and Pd/CeSn75\u20131000 show peaks at about 190\u00a0\u00b0C, which are related to the reduction of PdO nanoparticles [51,52]. The amount of PdO calculated from the area of these peaks is 14, 37 and 49% of the total Pd content in Pd/CeSn25\u20131000, Pd/CeSn75\u2013900 and Pd/CeSn75\u20131000 catalysts, respectively. It should be noted that PdO particles are not formed during the calcination of the samples at 800\u00a0\u00b0C, although the above data indicate the highest amount of active oxygen in the catalysts calcined at 800\u00a0\u00b0C. The formation of PdO particles in the Pd/CeSn50 catalysts is also not observed. It can be assumed that in those samples where the PdO phase is not formed, the maximum interaction of palladium with structures of cerium and tin oxides occurs.The most noticeable differences in TPR-CO curves are observed in the low temperature region, as shown in Fig. 2 d-f. The release of CO2 at the lowest temperatures is an extremely important factor, since it largely determines the activity of the catalyst. Fig. 2 d-f shows that the onset of CO2 evolution depends on both the tin content and the calcination temperature. For Pd/CeSn15 and Pd/CeSn25 catalysts the onset of CO2 evolution shifts towards higher temperatures from \u221220 to 20\u00a0\u00b0C as the calcination temperature increases from 800 to 1000\u00a0\u00b0C. For the Pd/CeSn50\u2013800 catalyst, CO2 evolution starts at the temperature of 0\u00a0\u00b0C which remains practically unchanged with increasing calcination temperature of the catalyst up to 1000\u00a0\u00b0C, thus demonstrating high thermal stability. For the Pd/CeSn75 catalyst the characteristic feature is the ability to release CO2 at very low temperatures around \u221240\u00a0\u00b0C if the catalyst is calcined at 800 and 900\u00a0\u00b0C. These data allow us to conclude that the Pd/CeSn75 catalyst has the most weakly bound and reactive oxygen among the catalysts calcined at 800\u2013900\u00a0\u00b0C, but the amount of oxygen of this type is quite low. After calcination at 1000\u00a0\u00b0C, the CeSn75\u20131000 catalyst loses weakly bound oxygen, whereas the Pd/CeSn50\u20131000 catalyst has the highest amount of weakly bound oxygen in this case. The presented data show that the Pd/CeSn50 catalyst has a compromise property of retaining the highest amounts of weakly bound oxygen at high calcination temperatures up to 1000\u00a0\u00b0C.Fig. S7 shows the temperature dependence of CH4 conversion and the Arrhenius dependence of CH4 oxidation rate. Based on these data, the values of T\n10 and weight rate of CH4 oxidation (per 1\u00a0g Pd) at 450\u00a0\u00b0C were calculated as a function of relative tin content X for Pd/CeSnX catalysts (X\u00a0=\u00a015, 25, 50, 75) calcined at 800, 900 and 1000\u00a0\u00b0C. These data presented in Fig. 3\n show that the highest T\n10 values are observed for the Pd/CeSn15 catalyst with lowest tin content. An increase in the tin content leads to a decrease in the T\n10 value. The lowest T\n10 values and, accordingly, the highest reaction rate are achieved for the Pd/CeSn75 catalyst. Being calcined at 1000\u00a0\u00b0C the Pd/CeSn75 catalyst is characterized by a significantly higher reaction rate compared to the Pd/CeSn50\u20131000 and Pd/CeSn25\u20131000 catalysts (4 and 10 times higher, respectively). Thus, the catalytic testing of the studied catalysts in the reaction of methane oxidation shows that the activity of the Sn-rich catalysts is significantly higher than the activity of the Ce-rich ones.The effect of water vapor on the catalytic activity was studied on the Pd/CeSn75 catalyst, which shows the highest activity in the oxidation of methane in the series of Pd/CeSnX catalysts. As shown in Fig. 3 \u0441, the \u042210 value for the Pd/CeSn75\u2013800 catalyst increases from 380 to 505\u00a0\u00b0C when 10% H2O is added to the reaction mixture. At the same time, the reaction rate decreases by about an order of magnitude (Fig. 3 d).\nFigure 4\n shows XRD patterns obtained for Pd/CeSn catalysts of various compositions calcined at 450, 800, 900, and 1000\u00a0\u00b0C. Table S2 presents the structural parameters and phase content calculated from the XRD data. Due to the low palladium content in the samples, the diffraction patterns do not comprise any reflections that could be attributed to known palladium-containing phases.Pd/CeSnX catalysts calcined at 450\u00a0\u00b0C are characterized by a significant broadening of diffraction reflections due to the small crystallite size, the presence of crystalline structure defects in the formed phases and local fluctuations of phase composition. The poor resolution of the reflections does not allow a correct quantitative phase analysis to compare the amounts of coexisting phases with the structure of fluorite and rutile. Moreover, for Ce-rich samples (Pd/CeSn15\u2013450 and Pd/CeSn25\u2013450) it is not possible to confidently state from the XRD data whether a phase with rutile structure is present or not. Diffraction data for catalysts calcined at 450\u00a0\u00b0C also do not allow the crystal lattice parameters to be determined with sufficient accuracy. Nevertheless, it can be confidently stated that the lattice parameter of fluorite phase in all samples a\nFluorite\u00a0=\u00a05.26\u20135.40\u00a0\u00c5 differs significantly from the lattice parameter of pure cerium oxide (a(CeO2)\u00a0=\u00a05.411\u00a0\u00c5, ICDD PDF-2 card #34\u2013394). This indicates the incorporation of tin ions into the cerium oxide lattice with the formation of a Ce1-ySnyO2\n(or Pd\n\nx\n\nCe\n\n1-x-y\n\nSn\n\ny\n\nO\n\n2-\u03b4\n\n) solid solution [55].\nAn increase in the calcination temperature of catalysts to 800\u00a0\u00b0C and higher leads to an increase in the crystallite size due to coarsening. This is accompanied by narrowing of the reflections in the diffraction patterns. Fig. 4 and Table S2 show that the Pd/CeSn catalysts calcined at 800\u00a0\u00b0C represent two-phase mixtures consisting of a phase with fluorite structure and a phase with rutile structure. A slight shift of the lattice parameters in the observed phases (Table S2) relative to the lattice parameters of pure cerium oxide (a\u00a0=\u00a05.411\u00a0\u00c5, ICDD PDF-2 card #34\u2013394) and tin oxide (a\u00a0=\u00a0b\u00a0=\u00a04.738\u00a0\u00c5, c\u00a0=\u00a03.187\u00a0\u00c5, ICDD PDF-2 card #41\u20131445) indicates the incorporation of tin and cerium atoms into the lattice of the CeO2 and SnO2 phases to form primary solid solutions based on fluorite (Ce1-xSnxO2-\u03b4) and rutile (Sn1-yCeyO2-\u03b4). Quantitative phase analysis shows that the phase content of the samples correlates well with the nominal composition of the samples. It should be noted that the samples with a higher tin content comprise crystallites of a smaller average size compared to Ce-rich samples (Table S2), and this decrease in particle size is relevant to both the fluorite and rutile phases. Thus, the fluorite phase in the Pd/CeSn25\u2013800 catalyst has a crystallite size of 6.8\u00a0nm, while in the Pd/CeSn75\u2013800 sample it is 4.5\u00a0nm, and the rutile phase in these two samples has crystallite sizes of 9\u00a0nm and 4.3\u00a0nm, respectively. The diffraction patterns of the samples calcined at higher temperatures (900 and 1000\u00a0\u00b0C) are characterized by narrow, well-resolved reflections (Fig. 4, Table S2). At the same time, the tendency for the crystallites in the observed phases of fluorite and rutile to be smaller in Sn-rich samples is retained.HRTEM images of catalysts calcined at 450\u00a0\u00b0C show crystallites of mixed CexSn1-xO2-\u03b4 oxides of various compositions (Fig. 5\n). The crystallite size in all samples varies from 1 to 4\u00a0nm; in the Pd/CeSn50\u2013450 sample individual largest crystallites with a size of 4\u20137\u00a0nm are observed.As can be seen from the Fast Fourier Transform (FFT) image shown in Fig. 5 a, all interplanar distances observed in the crystallites of the Pd/CeSn25\u2013450 catalyst correspond to the reflections of the fluorite phase. Also, there are areas in the sample containing poorly crystallized, amorphous species. Apparently, these species are the nuclei of the rutile phase. Thereby the rutile phase in the sample has not yet been formed, which explains the absence of rutile reflections on XRD patterns (Fig. 4). Similar amorphous regions are observed in the Pd/CeSn50\u2013450 sample, however, crystallites of the rutile phase are present \u2013 the characteristic reflections appear in FFT patterns, corresponding to distances in rutile phase (Fig. 5 b). The additional ring in the selected area electron diffraction (SAED) pattern between the (220) and (311) rings of CeO2 is clearly visible, the position of this ring corresponds to the distance of 1.76\u00a0\u00c5 and thus can be attributed to the (211) ring of the SnO2 rutile phase. HRTEM images of the Pd/CeSn75\u2013450 catalyst clearly show crystallites with the interplanar distances corresponding to the fluorite and rutile phases. The intensity of the rutile phase reflections in the FFT patterns (Fig. 5 c) exceeds the intensity of the fluorite ones. Also, HRTEM images show the presence of amorphous, poorly crystallized particles.According to EDX analysis (Fig. 5 d, f), Pd/CeSn25\u2013450 and Pd/CeSn75\u2013450 catalysts show quite homogeneous distribution of Sn and Ce, i.e. particles of CexSn1-xO2-\u03b4 oxides with various composition are strongly mixed in the volume of the samples. On the contrary, the Pd/CeSn50\u2013450 sample is notable for a \u201ctwo-colour\u201d EDX pattern, where tin-rich and cerium-rich regions are clearly distinguished (Fig. 5 e). Quantitative analysis gives atomic ratios Ce:Sn\u00a0\u2248\u00a01.7 in cerium-rich regions and Ce:Sn\u00a0\u2248\u00a00.3 in tin-rich regions.Palladium in all samples is in a highly dispersed state and is not detected as a separate phase. A low-intensity signal from palladium can be seen in the EDX spectra, quantified from 0.4 to 1.5\u00a0wt% in different sections of the samples, with higher amounts of palladium being detected in the Ce-rich regions.TEM study of Pd/CeSn catalysts showed that calcination at 800\u00a0\u00b0C leads to formation of polycrystalline agglomerates consisting of two types of oxide nanocrystallites (Fig. 6\n). According to EDX data regions enriched in either tin or cerium can be distinguished. Analysis of interplanar spacings allows us to attribute the Sn-rich crystallites to rutile phase and the Ce-rich ones to fluorite phase, in agreement with the XRD results. The crystallites of rutile and fluorite phases are spatially mixed in the bulk of the sample agglomerates, thus forming a nanoheterophase structure over the studied range of tin content. The crystallite size of both phases has a non-linear dependence on the tin content in the catalyst. The highest dispersion is observed for the Pd/CeSn75\u2013800 catalyst, where the size of the fluorite crystallites varies within 3\u20136\u00a0nm, and the size of the rutile crystallites is about 5\u201310\u00a0nm. The crystallite size of the fluorite phase in the other two samples is slightly higher, 5\u201310\u00a0nm. The size of rutile nanocrystallites varies in the range of 5\u201315\u00a0nm in Pd/CeSn25\u2013800 and reaches up to 20\u00a0nm in the Pd/CeSn50\u2013800 catalyst (Fig. S8). Larger Sn-rich crystallites are surrounded by smaller Ce-rich crystallites, thus forming composites with intercrystalline mesoporosity. Only for the PdCeSn75 catalyst no preferential localization of cerium on the surface is observed, most likely due to too high tin content in the sample.All the catalysts calcined at 800\u00a0\u00b0C comprise palladium in a highly dispersed state, no large palladium-containing structures are detected in TEM images. EDX analysis shows rather uniform distribution of palladium in the studied regions without agglomeration (Fig. S8).The catalysts calcined at high temperatures of 900\u00a0\u00b0C and 1000\u00a0\u00b0C were investigated by TEM. The calcination of the Pd/CeSn25 catalyst at 900\u00a0\u00b0C and higher leads to a noticeable sintering and enlargement of a part of the fluorite phase crystallites. As it can be seen from Fig. 6, Ce-rich particles of 50\u2013100\u00a0nm in size appear in the Pd/CeSn25\u2013900. These particles are characterized by developed intracrystalline microporosity. Such pores do not form in smaller fluorite particles.Earlier studies of Pd/CeSn50\u2013900 and Pd/CeSn50\u20131000 catalysts [29] showed formation of heterogeneous structures, represented by the SnO2-based particles covered with CeO2-based particles. A higher amount of tin in these catalysts was shown to prevent cerium oxide sintering, the crystallite size of the fluorite phase does not exceed 20\u00a0nm.The particle size of the fluorite phase in Sn-rich Pd/CeSn75\u2013900 catalyst is even smaller \u2013 about 5\u20138\u00a0nm, the particles of the rutile phase are enlarged to 5\u201320\u00a0nm (Fig. 6). Thus, the crystallite size of both phases in Sn-rich catalysts remains the smallest in the whole series of catalysts.All catalysts are characterized by the formation of intercrystalline mesopores after thermal treatment. The most developed porous structure is achieved with a small spread in the size of the support crystallites, which is typical for the samples with the maximum tin content \u2013 Pd/CeSn75\u2013900 (Fig. 6) and Pd/CeSn75\u20131000.TEM data for the Pd/CeSn75\u2013900 catalyst (Fig. 7\n) show that palladium leaves the support lattice to form PdO nanoparticles with a size of about 5\u201310\u00a0nm. These particles are densely surrounded by the CexSn1-xO2-\u03b4 nanoparticles, forming structures of the core@shell type. As can be seen from Fig. 7, PdO particles are most closely in contact with Ce-rich nanoparticles, and the Sn-rich particles are located on the surface of these agglomerates. Thus, these particles can be considered as having a core@shell1@shell2 structure, where the core is PdO, the shell1 consists of the fluorite CexSn1-xO2-\u03b4 (x\u00a0>\u00a00.6) nanoparticles, and the shell2 is formed by the rutile SnyCe1-yO2-\u03b4 (y\u00a0>\u00a00.6) ones.The calcination of the Pd/CeSn75 catalyst at 1000\u00a0\u00b0C leads to further segregation of palladium from the support lattice and enlargement of the Pd-containing particles. The formed particles are clearly visible in the TEM and STEM images, their size is about 100\u00a0nm (Fig. 8\n). It should be noted that rather low oxygen signal is detected in the EDX spectra acquired from the central area of particles (Fig. 8 e). At the same time the Pd signal in center is noticeably higher, and an inner border in Pd map is distinguishable. This assumes the core@shell particle structure with metal palladium core covered by PdO shell. These core@shell Pd@PdO particles are covered with nanoparticles of cerium\u2011tin oxides, which size varies in the range of 5\u201315\u00a0nm. In this case, due to the large size of palladium particles, the Ce-rich and Sn-rich oxide nanoparticles are mixed randomly without particular localization on the surface of certain phases, which was observed for the Pd/CeSn75\u2013900 sample.The nanoscale palladium phase was detected only in Sn-rich catalysts after high-temperature treatment (900\u20131000\u00a0\u00b0C). In all other studied samples palladium is in a highly dispersed state.HRTEM investigation (Fig. 9\n) reveals the presence of small species on the support surface, which can be attributed to PdOx clusters [53]. It should be noted that the observed clusters were located on the surface of fluorite particles. The size of the clusters does not exceed 1\u00a0nm. Under the influence of the electron beam, the clusters detected on extended facets of ceria exhibit dynamic behavior and move intensively across the surface until they reach the surface defects. We associate this dynamic process with the partial reduction of clusters under the influence of the electron beam, which decreases their stability on extended ceria surfaces. The Pd-containing clusters are stabilized on the defects of the support, such as surface steps and intercrystalline boundaries. Analysis of the interplanar distances in these clusters shows the correspondence to metallic palladium. More often, epitaxy between the lattice of such clusters and the fluorite crystals is observed, that is typical for Pd/CeO2 systems [56\u201358]. Nevertheless, many of the clusters do not exhibit the periodic contrast due to the small size and resulting structural rearrangement. In [59] similar amorphous Pd particles stabilized on ceria surface were shown to be highly active and stable towards CO oxidation.A high-resolution EDX mapping (Fig. 10\n) for the Pd/CeSn50\u2013800 sample revealed that not only palladium but also tin signals are recorded intensely in the area where the cluster is located. Meanwhile, the cluster is located on the surface of Ce-rich nanoparticle. Apparently, the calcination of the catalyst results in the segregation of both palladium and tin from the lattice of mixed Ce(Pd,Sn)O2 oxides with the formation of mixed palladium\u2011tin clusters. Most likely, such clusters are initially oxidized under atmospheric conditions, but get easily reduced during TEM investigation under the action of the electron beam [51,60].Despite the formation of a nanosized palladium phase in Sn-rich catalysts, part of palladium remains in a highly dispersed state after calcination at 900\u20131000\u00a0\u00b0C. The HRTEM and HAADF-STEM images presented in Fig. 11\n show a region of a Pd/CeSn75\u2013900 sample. It can be seen that clusters <1\u00a0nm in size are located on the surface of the CexSn1-xO2-\u03b4 particles. Using high-resolution EDX mapping, we confirm the localization of the clusters on the surface of Ce-rich crystallites. Also, individual clusters are located at the grain boundaries between Ce-rich and Sn-rich crystallites thus having an interface with the rutile phase too. Again, in the presented EDX-mapping data the Sn signal is clearly visible in the region containing the clusters. Thus, the earlier conclusion about the formation of PdSnxOy clusters in the samples during the segregation of Pd and Sn from the fluorite solid solution is confirmed.High-temperature treatment at 1000\u00a0\u00b0C is accompanied by further segregation of doping elements from the crystal lattices of the fluorite and rutile phases. As a result, HRTEM images (Fig. 12\n) show numerous clusters and amorphous species on the surface of the support particles. These highly dispersed species undergo constant movement under the influence of the electron beam followed by slight agglomeration near the surface defects. Due to the reducing effect of the electron beam some TEM images of the clusters show interplanar spacings of about 0.23\u00a0nm, which corresponds to the (111) spacing in metallic palladium.Thus, all catalysts retain the highly dispersed active state of palladium even after calcination at 1000\u00a0\u00b0C. Only in Sn-rich catalysts a part of palladium undergoes sintering with the formation of PdO nanoparticles stabilized by the CexSn1-xO2-\u03b4 support in the form of core@shell structures.\nFigure 13\n shows the Pd 3d spectra obtained for all catalyst series depending on the composition and calcination temperature. All spectra in Fig. 13 are normalized to the total integral intensity of Ce 3d\u00a0+\u00a0Sn 3d lines. The Pd 3d spectra were decomposed using three individual doublets with E\nb(Pd 3d\n5/2)\u00a0=\u00a0337.8\u00a0eV, 336.2\u00a0eV and 336.6\u2013337.2\u00a0eV. The provided spectra show that for the Pd/CeSn25 and Pd/CeSn50 catalyst series palladium is predominantly in one state, characterized by E\nb\u00a0=\u00a0337.8\u00a0eV (Fig. 13 a, b). According to the literature [4,35,52,53,61] this state is interpreted as Pd2+ ions incorporated into the lattice of cerium oxide particles. An additional state with E\nb(Pd 3d\n5/2)\u00a0=\u00a0336.0\u2013336.2\u00a0eV is observed in Pd/CeSn25\u2013800 (Fig. 13 a, curve 2), Pd/CeSn50\u2013900 and Pd/CeSn50\u20131000 (Fig. 13 a, curves 3,4), which can be reliably referred to PdOx surface clusters [53,60]. In the case of Pd/CeSn25 catalyst calcination at higher temperatures of 900 and 1000\u00a0\u00b0C leads to a shift of this doublet by 0.5\u00a0eV towards higher binding energies to the value E\nb(Pd 3d\n5/2)\u00a0=\u00a0336.6\u00a0eV, which is associated with the formation of PdO particles [62]. In the series of catalysts with high tin content (Fig. 13 c) the intensity of the palladium line is significantly lower, indicating either a larger size of palladium particles or their encapsulation by the support. Calcination of the Pd/CeSn50 from 800 to 1000\u00a0\u00b0C leads to a shift of the Pd 3d\n5/2 line from 337.5 to 336.7\u00a0eV, which implies the transition from isolated Pd2+ ions in the oxide matrix to nanosized PdO particles.Quantitative data on the atomic ratio of Pd/(Ce\u00a0+\u00a0Sn) depending on the catalyst calcination temperature are shown in Fig. S9. Thus, the XPS data show that the Pd/(Ce\u00a0+\u00a0Sn) ratio decreases with increasing tin content in the catalysts, and this effect is pronounced with an increase in the calcination temperature. After calcination of the Pd/CeSn25 catalyst at 600 \u00b0\u0421, the ratio Pd/(Ce\u00a0+\u00a0Sn) equals 0.016, which is the typical value for isotropic distribution of palladium over the catalyst, while for the Pd/CeSn50\u2013600 and Pd/CeSn75\u2013600 catalysts the palladium concentration in the surface layers is significantly lower. The calcination procedure affects this Pd/(Ce\u00a0+\u00a0Sn) ratio differently. For catalysts with a high cerium content (Pd/CeSn50 and Pd/CeSn25), calcination at 800\u20131000\u00a0\u00b0C leads to an increase in the Pd/(Ce\u00a0+\u00a0Sn) ratio compared to this value for catalysts calcined at 600\u00a0\u00b0C. This suggests a partial segregation of palladium ions to the surface and/or dispersion of palladium particles and their participation in the formation of active centers. This effect was previously observed in catalysts with equimolar content of cerium and tin in the Pd/CeSn50 catalyst [29]. It should be noted that the highest Pd/(Ce\u00a0+\u00a0Sn) ratio was obtained for the Pd/CeSn25\u2013800 catalyst, which exhibits high specific activity (Fig. 1 b). In the case of the Pd/CeSn75 catalyst with low cerium content no segregation of palladium into the surface layers of the catalyst is observed, only sintering or encapsulation of palladium occurs in accordance with the interpretation of the shifts of the Pd 3d line.Analysis of the Ce 3d line shows that the proportion of Ce3+ in the samples increases with increasing tin content (Fig. S10). In the case of the Pd/CeSn75\u2013800 catalyst an abnormally high Ce3+ content of about 40\u201345% of the total cerium content is observed. Such high values of Ce3+ concentrations are unattainable within the crystal lattice of CeO2 fluorite and may indicate that part of cerium ions, namely Ce3+, is outside the ordered phase of CeO2 particles. These cerium ions are either located in the amorphous layers on the surface of cerium oxide particles or they dope the lattice of tin oxide.\nFigure 14\n shows the Raman spectra of the catalysts obtained with laser excitation wavelengths of 532\u00a0nm and 785\u00a0nm. It was established that the spectra of the catalysts calcined at 450 and 600\u00a0\u00b0C are almost identical. Therefore, Fig. 14 shows the Raman spectra of the catalysts starting from the calcination temperature of 600\u00a0\u00b0C.For Pd/CeSn samples calcined at 600\u00a0\u00b0C, a strongly broadened Raman peak is observed in the region of 460\u00a0cm\u22121, corresponding to the well-known F2g vibrational mode of the crystalline cubic fluorite type structure of ceria, originated by the tension vibrations of the oxygen atoms that surround the cerium atoms [63]. As the calcination temperature increases, this peak narrows, with the strongest narrowing observed for the Pd/CeSn50 sample. For the Pd/CeSn75\u2013450 and Pd/CeSn75\u2013600 samples with low cerium content the F2g vibrational mode is almost not manifested, indicating that the ordered structure is formed during high-temperature annealing. Also, for the samples calcined in the temperature region of 600\u2013800\u00a0\u00b0C, a general rise in the Raman spectrum is observed in the vibration region below 750\u00a0cm\u22121, which indicates significant lattice distortions of the fluorite structure of cerium oxide. One can also observe a weak peak in the region of \u223c835\u00a0cm\u22121 (at ex\u00a0=\u00a0532\u00a0nm), which can be attributed to stretching vibration of peroxide species formed on CeO2 surface according to literature data and our recent studies [64,65]. This peak does not appear during further annealing.Tin oxide with a tetragonal rutile structure has a complex structure of vibrational states \u2013 18 oscillations in the first Brillouin zone [66]. Of these, the symmetrical A2u and Eu modes are IR active, whereas the remaining A1g, B1g, B2g, and Eg modes are Raman active. It should be noted that the Raman spectrum of SnO2 is well manifested only for well-crystallized and rather large particles (tens of nm) or bulk samples [66]. In our case we see intense SnO2 Raman peaks only for the Pd/CeSn75 sample calcined at 900 and 1000\u00a0\u00b0C - this is an intense narrow band A1g vibration at \u223c635\u00a0cm\u22121, excited by 532\u00a0nm radiation. For the other samples with lower tin content we observe only a weak A1g band at 635\u00a0cm\u22121.Of great interest is the broad band in the region of \u223c234\u00a0cm\u22121, which is formed in all samples during calcination at T\u00a0\u2265\u00a0800\u00a0\u00b0C and is observed when using radiation of 785\u00a0nm, while the radiation of 532\u00a0nm does not excite this type of vibration.It is known that vibration of IR active Eu mode of SnO2 lie in this region, which usually does not show up in Raman spectra. However, these vibrations become active in Raman spectra when the structure of SnO2 is distorted, for example, as a result of a lack of oxygen. Thus, in [67] the bands of vacancy-related Raman modes at 234, 573, and 618\u00a0cm\u22121 related to Eu, B1u, and A1g modes were observed. Therefore, the band in the Raman spectrum at 234\u00a0cm\u22121 can be attributed to the Eu vibration of the distorted rutile SnO2 structure; the less intense peaks in the region from 250 to 650\u00a0cm\u22121 also refer to different SnO2 modes (Eu, Eg, B1u, and A1g) in the distorted structure.The incorporation of Pd into the matrix is confirmed by Raman spectra (a broad shoulder in the region of 500\u2013700\u00a0cm\u22121 with excitation at 532\u00a0nm for samples calcined at low temperatures - up to 800\u00a0\u00b0C) [60].Thus, the Raman spectroscopy data indicate a high degree of lattice distortion of the rutile and fluorite phases, which is consistent with the XPS data and confirms the structural data obtained by XRD and TEM methods.The study of the catalytic properties of Pd/CeSn catalysts for the CO oxidation reaction allowed us to establish two general patterns. The first one describes the catalytic activity as a dome-shaped dependence on the Ce/Sn ratio. The highest activity in CO oxidation is observed for catalysts with cerium content from 75 to 50%. In agreement with [37] this range extends for Ce-rich catalysts with 80% cerium content. However, the activity decreases noticeably already at 85% Ce content in accordance with our data. The second pattern is related to the effect of calcination on the catalytic activity. The effect of thermal activation is observed as the calcination temperature increases from 450 (600)\u00a0\u00b0C to 1000\u00a0\u00b0C, and the maximum activity is observed at T\u00a0=\u00a0800\u2013900\u00a0\u00b0C. The effect of thermal activation is manifested to some extent for all the compositions studied, but the greatest increase in activity is observed for the equimolar Ce/Sn ratio. It is noteworthy that the amount of reactive oxygen passes through a maximum depending on the concentration of tin in the samples. The calcination at 800\u00a0\u00b0C yields the maximum amount of oxygen for the equimolar Ce/Sn ratio, while for higher calcination temperatures of 900\u00a0\u00b0C and 1000\u00a0\u00b0C the maximum amount of oxygen is observed for the Ce-rich sample. Such high amount of reactive lattice oxygen is explained by formation of additional oxygen vacancies associated with the redox couple Sn4+/Sn2+ (along with the Ce4+/Ce3+ pair) [19,25,26] promoting increased oxygen mobility.The activity of catalysts in the methane oxidation reaction increases with increasing Sn content, and the ignition temperature of the reaction decreases significantly in accordance with the work [40].The study of the catalytic properties of Pd/CeSn catalysts shows a strong mutual influence and synergy of Ce and Sn components, which are responsible for the formation of the support morphology and structure.The addition of tin to Pd/CeO2 system leads to a significant thermostabilizing effect, which involves the suppression of sintering processes due to the formation of a nanoheterophase structure consisting of crystallites of the tin-doped fluorite phase and the cerium-doped rutile phase. Phase inhomogeneity sterically hinders the sintering of crystallites. Mutual doping of the fluorite and rutile phases with Sn and Ce ions leads to a change in the lattice parameters and a strong distortion of the corresponding crystal lattices. In work [31] the increased thermal stability of the CeO2-SnO2 system was explained by the cerium segregation on the surface of tin oxide particles with the formation of a Ce-rich surface layer. Due to the distortion of the crystal lattice, a large number of defects are formed, which, in turn, serve as centers for stabilization of palladium in a highly dispersed state with formation of various microstructures.Even after calcination at high temperatures up to 1000\u00a0\u00b0C high dispersity of rutile and fluorite phases is retained due to the spatial mixing of particles forming a nanoheterophase structure. The mutual arrangement of the particles of two phases is determined by their content and the size of the crystallites. If the Ce:Sn atomic ratio is \u22651, there is a tendency for fluorite phase nanoparticles to localize on the surface of larger Sn-rich particles, forming composites with intercrystalline mesoporosity. It is this microstructure that is favorable for catalysts that are highly active in low-temperature CO oxidation.The formation of isolated Pd2+ ions with their preferential localization throughout the volume of CeO2-based fluorite phase particles is characteristic when using catalysts with higher cerium content. This type of interaction between palladium and cerium oxide can be characterized as the strongest from a chemical point of view and is associated with the relatively small size of CeO2 particles and their defectiveness. In this case, palladium is in the most dispersed (atomic) state and can penetrate into the volume of the fluorite particles, which in this case are highly defective due to the introduction of Sn4+ ions into their lattice and the formation of a CexSn1-xO2-\u03b4 solid solution. Calcination at temperatures of 800\u20131000\u00a0\u00b0C stimulates the transfer of palladium atoms to the upper layers of CexSn1-xO2-\u03b4 particles. Thus, the number of active catalytic centers on the surface significantly increases, leading to an overall increase in the activity of catalysts per weight content of palladium. The calcination of Pd/CeSn catalysts leads not only to the mentioned transfer of Pd atoms, a similar process also occurs with the Sn ones doping the fluorite phase, resulting in formation of active PdxSnyOz clusters (oxidized due to the contact with the atmosphere and with the fluorite particles). The Sn atoms are mostly localized closer to the surface of the fluorite particle on which the cluster is located. Most likely, the segregation of Sn occurs as an interface layer of SnOx, on which palladium is localized in the form of PdOx. Some of the Sn atoms are assumed to be incorporated directly into the active palladium particle. In summary, the data obtained indicate the formation of a new type of catalytically active centers in which PdOx clusters are modified by tin during calcination at T\u00a0=\u00a0800\u2013900\u00a0\u00b0C. This effect of modification of PdOx active centers by tin atoms during catalyst calcination should affect the reactivity of oxygen. Indeed, the onset of the CO2 evolution upon interaction with CO can occur even at temperatures below 0\u00a0\u00b0C for the catalysts calcined at 800\u00a0\u00b0C and higher. The catalysts calcined at a low temperature of 450\u00a0\u00b0C show the onset of interaction with CO only at 50\u00a0\u00b0C and above (Fig. S11). Note that within the Mars-Van Krevelen mechanism, the most reactive oxygen forms determine the ignition of the reaction and the low-temperature activity of the catalysts. Thus, the thermal activation of catalysts with high cerium content is associated with the reorganization of the morphology and structure of catalysts with the formation of active centers based on PdOx clusters with their modification by tin atoms.The calcination of catalysts with high tin content results in such a reorganization of the morphology and structure of the support that leads to encapsulation of palladium in the form of PdO particles inside a shell of nanoparticles of fluorite and rutile phases. Previously, the formation of catalysts with the PdO@CeO2 structure was studied in a number of articles [68\u201370]. These works demonstrated high activity of such catalysts in the reaction of low-temperature oxidation of CH4. Catalytic data obtained in the reaction of the CH4 oxidation for the Pd/CeSn system show that the Sn-rich sample has significantly higher activity than the samples with lower tin content. Comparing the structural and catalytic data, we can conclude that the formation of PdO particles during calcination of the Sn-rich sample at 800\u20131000\u00a0\u00b0C is responsible for the catalyst activity with respect to methane oxidation. A characteristic property of the Sn-rich sample is the preservation of activity regardless of the calcination temperature in the range of 800\u20131000\u00a0\u00b0C, which is apparently a consequence of the formation of core@shell PdO@(CeO2\u00a0+\u00a0SnO2) structures.A study of the effect of water vapor on the activity of catalysts showed inhibition of methane oxidation, which was demonstrated previously in a number of works [10,12,71]. According to [72,73], an increase in methane conversion can be achieved by increasing the Pd content. In turn, that increase in the conversion is expected to lead to a shift of the methane conversion curve to the low-temperature region, which is the goal of creating catalysts for the efficient oxidation of methane at low temperatures [7]. A detailed study of the effect of water vapor on catalysts with increased Pd content is planned for further research.In this work we carried out the detailed study of the structural and catalytic features of the thermostable 1%Pd/CeO2-SnO2 catalysts. It was demonstrated that by varying the Ce/Sn ratio and calcination temperature we can regulate the structure of the Pd-centers active in the reactions of CO and CH4 oxidation.The Pd/CeO2-SnO2 catalysts were synthesized by counter precipitation varying the Ce/Sn ratio from 15/85 to 75/25 and the calcination temperature in a wide range of 450\u20131000\u00a0\u00b0C. To establish the structural features of the catalysts and their catalytic activity, the catalysts were studied by a complex of structural (XRD, HRTEM), spectral (XPS, Raman), and catalytic (TPR-CO\u00a0+\u00a0O2, TPR-CH4\u00a0+\u00a0O2, TPR-CO) methods. The catalysts, due to the synergism between components, were shown to possess high thermal stability after calcination up to 1000\u00a0\u00b0C with preservation of low-temperature (T\u00a0<\u00a0100\u00a0\u00b0C) activity in the CO oxidation reaction at a Ce/Sn ratio from 3/1 to 1/3. In the case of the catalyst with the highest tin content (Ce/Sn\u00a0=\u00a01/3), high-temperature treatment leads to a decrease in the activity in the CO oxidation, but significantly increases the activity in the CH4 oxidation reaction.It was found that the effect of thermal activation for Pd/CeSn catalysts is associated with a change in the state and structure of active centers involving palladium. The initial samples calcined at 450\u2013600\u00a0\u00b0C comprise palladium in ionic form (Pd2+) located in the lattice of the support nanoparticles, especially the fluorite ones. Annealing at 800\u00a0\u00b0C and higher is accompanied by the segregation of doping ions from the support structure with the formation of PdOx clusters on the surface of fluorite phase. The active PdOx clusters were found to be modified by Sn atoms, which also segregate from the fluorite lattice during calcination. The preservation of active cluster forms in the highly dispersed state during high-temperature treatment is due to their stabilization on the defects of the support surface. Due to parallel precipitation of tin and cerium oxides during the synthesis of catalysts, the two support phases are mutually doped with cerium and tin, respectively, which leads to distortion of crystal lattices and defectiveness of nanoparticles. Efficient spatial mixing of defective nanoparticles of the rutile and fluorite phases ensures the resistance of the support to sintering due to the formation of a nanoheterophase structure.In the Pd/CeSn75 catalyst with high tin content, the thermal activation is also associated with the reorganization of the support morphology, but it also includes the formation of a new phase based on palladium oxide. The segregation of Pd2+ ions in this case leads to the formation of both the tin-modified PdOx clusters and the PdO nanoparticles occluded inside the agglomerates of the support particles. The formed core@shell structure PdO@(CeO2\u00a0+\u00a0SnO2) shows high thermal stability and provides catalyst activity in methane oxidation reaction.\nOlga A. Stonkus: Investigation, Visualization, Writing - original draft. Andrey V. Zadesenets: Investigation. Elena M. Slavinskaya: Investigation, Writing - review & editing. Andrey I. Stadnichenko: Investigation, Visualization. Valery A. Svetlichnyi: Investigation, Visualization. Yury V. Shubin: Investigation, Visualization. Sergey V. Korenev: Methodology, Project administration. Andrei I. Boronin: Conceptualization, Project administration, Writing - review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the Ministry of Science and Higher Education of the Russian Federation within the governmental order for Boreskov Institute of Catalysis (project \u0410\u0410\u0410\u0410-\u041021-121011390053-4) and for Nikolaev Institute of Inorganic Chemistry (project #121031700315-2). The TEM studies were carried out using facilities of the shared research center \u201cNational center of investigation of catalysts\u201d at Boreskov Institute of Catalysis.\n\n\nSupplementary material\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2022.106554.", "descript": "\n                  1%Pd/CeO2-SnO2 catalysts with varying Ce/Sn ratio were synthesized by counter-precipitation followed by calcination in a wide temperature range. The catalysts with Ce/Sn\u00a0<\u00a03/1 possess high thermal stability after calcination up to 1000\u00a0\u00b0C while maintaining low-temperature activity in CO oxidation. The PdOx clusters serving as active centers in CO oxidation are modified by Sn upon calcination. High tin content (Ce/Sn\u00a0=\u00a01/3) provides the activity of the catalysts in CH4 oxidation due to stabilization of PdO nanoparticles in the form of core@shell PdO@(CeO2\u00a0+\u00a0SnO2) structures. Formation of the nanoheterophase structure upon calcination plays a key role in the stabilization of Pd-active centers of different types.\n               "}
{"full_text": "The catalysed formation of linear or isomerised aliphatic and aromatic hydrocarbons from the polymerisation of syngas (H2/CO mixture) is known as Fischer-Tropsch synthesis (FTS) [1]. FTS typically takes place under two environments: at high temperatures (330\u2013350\u00a0\u00b0C) on an iron catalyst, and at low temperatures (180\u2013250\u00a0\u00b0C) on a cobalt catalyst [2]. Whilst cobalt is the more costly metal, it shows higher catalytic activity and selectivity towards the formation of longer chain hydrocarbons than its iron congener [2].Many details of the FTS mechanism are still uncertain due to limitations around studying FTS in situ with an applied heterogenous catalyst. Consequently, there are several aspects of the mechanism that remain intensely debated. For example, activation of CO, a theorised step in FTS initiation, has been explained using different mechanistic pathways, such as associative or dissociative adsorption and with or without hydrogen assistance [3]. Another area of debate in the FTS mechanism is the process of chain growth, as spectroscopic characterisation of absorbates and reaction intermediates face many practical limitations [4\u20136]. Namely, chain growth intermediates (CxHy) are present in low concentration under steady state conditions alongside significantly higher surface coverages of CO, H, and products. Such compexity makes it difficult to determine which species are active in the chain growth mechansim and further to understand what the active sites are for this aspect of the reaction.A number of groups have used model approaches to study both CO activation and intermediate CxHy adsorption and chain growth on Co(0001) single crystal surfaces [7\u20139]. Regarding chain growth mechanisms, unsaturated absorbates such as ethylene and propene were shown to either dissociatively chemisorb to Had and the corresponding alkyne, or, in the presence of co-adsorbed CO, be hydrogenated to alkylidyne species [9]. When sufficiently heated, these alkylidyne species are shown to dimerise [8]. Whilst this shows that the presence of co-adsorbed CO impacts the reaction pathway of hydrocarbon polymerisation to form more reactive alkylidyne intermediates, spectroscopic characterisation of these intermediates under more realistic conditions, specifically at pressure, remains a challenge. In addition to this \u201cpressure gap\u201d, there is a significant difference between single crystal and polycrystalline applied catalytic materials, known as the \u201cmaterials gap\u201d [10].Inelastic neutron scattering (INS) and quasi-elastic neutron scattering (QENS) are vibrational and diffusional spectroscopic techniques that could provide insight into catalytic processes [11,12]. As neutrons have no electrical charge, they do not have the same limitations found in electromagnetic vibrational spectroscopy. All vibrational modes should theoretically be visible with INS/QENS, although the large incoherent scattering cross section of 1H results in these techniques being uniquely sensitive to hydrogen nuclei. QENS can measure the diffusional dynamics and mobility of hydrogen and CXHY species on the a catalyst surface, providing a more in-depth insight into the FTS mechanism. Several INS studies by Lennon and Parker have been performed on conventional supported and unsupported Fe FTS catalysts [13\u201315]. Recently, they have reported their first paper on Co surfaces using supported and unsupported Co catalysts [16]. These studies have focoused on the nature of hydrocarbon and coke species post FT catalysis, in addition to seminal work on the nature of adsorbed hydrogen on these surfaces. To date, no specific neutron scattering studies have been carried out that follow the FTS chain growth mechanism from model CxHy intermediates.As with all studies of specific adsorbates on supported metal catalysts, differentiation of species bound to metal or support is challenging. Skeletal metals, commonly known by their commercial name of Raney metals or as sponge metals, are nanoporous catalysts with a high surface area and limted or no support structure. The absence of adsorbate-support interactions as well as the resulting high thermal conductivity make skeletal metals potential candidates for studying CXHY adsorption and bridging the \u201cpressure\u201d and \u201cmaterials gap\u201d. Most neutron spectroscopy used with skeletal metals have been INS studies performed with skeletal nickel [17\u201321], with only a handful of studies using skeletal cobalt [16,22,23]. This includeds a recent report by Lennon and Parker on H2 adsorption on skeletal nickel and cobalt [16].Here we report preliminary experiments investigating skeletal cobalt as an adsorbate for an ethylene adsorption study using INS and QENS. The adsorbed ethylene is used as a model for CxHy FTS chain growth intermediates, as previously pioneered through the UHV single crystal work of Weststrate and co-workers [8,9]. The experiments intend to investigate the viability of using neutron scattering and skeletal cobalt catalysts, in first emulating surface science findings and then in the future proceding to perform experiments at elevated pressures. These neutron techniques were supported by temperature programmed reduction (TPR) experiments and X-ray photoelection spectroscopy (XPS) studies. Limitations in the use of skeletal cobalt as a bridge to UHV single crystal studies are discussed.500\u00a0g of reagent grade, >98% Sigma-Aldrich NaOH pellets were dissolved in 2\u00a0L distilled water at 40\u00a0\u00b0C. 100\u00a0g of Goodfellow\u00a9 33\u00a0wt% Co:66\u00a0wt% Al alloy catalyst precursor was slowly added to the base solution and stirred [Diffraction patterns of the Goodfellow\u00a9 alloy precursor sealed in cellulose tape were recorded using monochromated Cok\u03b11 radiation using a Bruker D8 Discover diffractometer operating in transmission mode over the 2\u03b8 range 10\u2013110\u00b0, with a 0.00750404\u00b0 step)] over a period of 13\u00a0h. Phase analysis was performed using the ICDD PDF database.The reaction proceeded for 1\u00a0h at 40\u00a0\u00b0C. The resultant catalyst was washed with distilled water until neutral (pH indicator paper was used). This reaction was repeated two more times and washed as before ensuring that the product remained submerged in water throughout. Due to the large mass of catalyst present, drying on a Schlenk line was impractical. As such, ~ 50\u00a0g of wet catalyst slurry was decanted into an Inconel 718 can until full, then sealed with a copper gasket. The gas inlet and outlet lines on the can were then connected to hydrogen and helium inlet gas lines and placed into a furnace. The catalyst was then dried under 270 SCCM He at 130\u00a0\u00b0C for 16\u00a0h, after which hydrogen was blended into the gas mixture at 130 SCCM at 280\u00a0\u00b0C for 3\u00a0h. The hydrogen supply was then turned off and the sample was held at 280\u00a0\u00b0C for 30\u00a0min under He flow.Ethylene gas lines were connected to the Inconel 718 canister as it was lowered into the neutron beam pathway. The sample was cooled to <10\u00a0K and an INS spectrum was recorded for 12\u00bd\u00a0hours. A buffer volume of 0.0175\u00a0mol (868.4\u00a0mbar in 499.76\u00a0cm3 at 298\u00a0K) of pure ethylene gas (SIP analytical N5.0 grade) was introduced to the catalyst sample at 200\u00a0K for 10\u00a0min. Sample was then cooled back down to <10\u00a0K and another INS spectrum was recorded for 13\u00a0h 20\u00a0min. The canister was then heated up to room temperature in order for any condensed ethylene gas to evaporate and adsorb onto the catalyst. Sample was held at room temperature for 10\u00a0min. Sample was then cooled back down to <10\u00a0K and a final INS spectrum was recorded for 16\u00bd\u00a0hours. Data were recorded on the TOSCA spectrometer and the MAPS spectrometer at the ISIS neutron and Muon Source.Catalyst slurry was prepared as described in Section 2.1 and decanted into an Inconel 718 can and dried, reduced, and washed as previously described. The catalyst was then decanted into a cylindrical aluminium canister inside an Ar glovebox and sealed with indium wire.The aluminium canister was connected to a pure ethylene gas line and placed into the neutron beam. A base temperature measurement was recorded at <10\u00a0K before heating to 300\u00a0K with data collected at intermediate temperatures. The sample was then dosed with 0.0202\u00a0mol (1000.0\u00a0mbar in 499.76\u00a0cm3 at 298\u00a0K) pure ethylene gas (N5.0) and allowed to equilibrate for 10\u00a0min before the measurement was repeated. Residual gas was removed by evacuation at room temperature before dosing with a gas mixture of 50\u00a0mbar pure (N5.0) CO and 850\u00a0mbar pure ethylene gas in 499.76\u00a0cm3 at 298\u00a0K. After equilibration for 10\u00a0min, the sample was cooled for measurement. Data were recorded on the IRIS spectrometer at the ISIS neutron and Muon Source using the pyrolytic graphite analyser.Surface area analysis required pre-passivation of sample to allow exposure in air, no other procedure required sample passivation. 5000\u00a0ppm O2 in He gas mixture was flowed into a Schlenk flask containing 100\u00a0mg of catalyst before TPR/TPD treatment at 30\u00a0mL\u00a0min\u22121 after a needle was inserted through the Suba-seal to prevent building up of internal pressure. Sample was left to oxidise for 30\u00a0min. Passivated samples were degassed at 150\u00a0\u00b0C for 24\u00a0h. BET surface area and porosity was measured using N2 substrate gas at 77.3\u00a0K (liquid N2) with a Quantachrome Quadrasorb and a Micromeritics ASAP 2020 Plus.Temperature programmed reductions were performed on the as synthesised material and after ethylene dosing using an Altamira AMI-300 Lite. Approximately 75\u00a0mg (exact measurements recorded for each experiment) of catalyst was loaded in-between quartz wool and reduced under 5% H2/Ar flowing at 30 SCCM with a 5\u00a0\u00b0C\u00a0min\u22121 ramp rate from room temperature to 300\u00a0\u00b0C. Measurement of hydrogen consumption was determined by a thermal conductivity detector (TCD) and calibrated by pulsing argon through a 574\u00a0\u03bcL sample loop. Samples were then cooled to room temperature and dosed with an excess of ethylene. Multiple TPR cycles without exposure to air were also performed on dosed and un-dosed samples to probe ethylene deposition.XPS analysis was performed using a Kratos Axis SUPRA XPS fitted with a monochromated Al k\u03b1 X-ray source (1486.7\u00a0eV), a spherical sector analyser and 3 multichannel resistive plate, 128 channel delay line detectors. Samples were transferred from glovebox to the instrument using the vacuum transfer arm in order to avoid exposure to the air. All data were recorded at 150\u00a0W and a spot size of 700\u00a0\u00d7\u00a0300\u00a0\u03bcm. Survey scans were recorded at a pass energy of 160\u00a0eV, and high-resolution scans recorded at a pass energy of 20\u00a0eV. Electronic charge neutralization was achieved using a magnetic immersion lens. Filament current\u00a0=\u00a00.27A, charge balance\u00a0=\u00a03.3V, filament bias\u00a0=\u00a03.8V. All sample data were recorded at a pressure below 10\u22128\u00a0Torr and a room temperature of 294\u00a0K. Data were analysed using CasaXPS v2.3.20PR1.0 and the spectra were calibrated with C1s peak at 284.8\u00a0eV.Following the synthetic procedure, skeletal cobalt was prepared by the caustic dealumination of the cobalt-aluminium alloy. XRD analysis (Fig. 1\n) demonstrated that the material comprised of a mixture of small FCC and HCP Co metal crystallites, in keeping with numerous other observations of nanoparticular support Co and skeletal cobalt catalysts [24]. The synthesised material was found to have a surface area of 23 m2g\u22121, with a pore volume of 0.07\u00a0cm3\u00a0g\u22121 and average pore diameter of 3.8\u00a0nm, in keeping with previously synthesised materials [25].Based on the recent findings from Davidson et al. that the surface of skeletal cobalt is dominated by hydroxylated and oxidic species, a temperature programmed reduction of the catalyst was performed (Fig. 2\n). The profile of the TPR shows has three features centred at 140\u00a0\u00b0C, 160\u00a0\u00b0C and 180\u00a0\u00b0C and finally a series of residual reduction features up to 300\u00a0\u00b0C. As anticipated, the volume of hydrogen consumed is significantly lower, at 0.96\u00a0mmol\u00a0g\u22121, than that seen for bulk reduction of Co3O4 or CoO to Co (16.6 and 13.3\u00a0mmol\u00a0g\u22121 respectively) and is indicative of a passivated surface oxide/hydroxylated layer on the catalyst. A temperature of 280\u00a0\u00b0C was therefore used to reduce the catalyst for ethylene dosing experiments. Notably, this reduction procedure resulted in the surface area of the material dropping to 12\u00a0m2\u00a0g\u22121. Given that a key criterion of this model material is high Co surface area, to maximise CxHy concentration and make INS/QENS experiments viable, this degree of sintering is concerning. Further studies to remove surface hydroxyl species with milder reduction treatments are required as part of ongoing work.After reduction of surface hydroxyl species, the material was transferred to TOSCA and ethylene dosed at \u221273\u00a0\u00b0C (200\u00a0K) and then again at 30\u00a0\u00b0C. The spectra collected from TOSCA after ethylene dosing and background subtraction are shown in Fig. 3a. It is noted that the observed signal in both spectra was relatively low and indicates that sintering of the sample limited the concentration of H containing species. Both samples show peaks at 988 and 1102\u00a0cm\u22121, which in reference to skeletal Ni [26] suggests hydrogen bound in the threefold-site (i.e. Co3H). If the C2H4 was partially dehydrogenated to adsorbed acetylene and H2, then features should be present c. 300\u20131200\u00a0cm\u22121 dominated by the symmetric and asymmetric CH bending modes. The absence of these bands supports the explanation that complete dehydrogenation to form adsorbed hydrogen and coke-type species might be occurring. The 85\u00a0cm\u22121 band could represent external vibrational modes of associatively adsorbed ethylene that have broadened and shifted due to interaction with the metal surface [27]. The presence of both adsorbed hydrocarbon and completely dehydrogenated species suggests that different sites on the cobalt surface exist with very different reactivity. Alternatively, the 85\u00a0cm\u22121 mode could be interpreted as another hydrogen adsorption mode; vide infra.After the second dose at 30\u00a0\u00b0C, a richer spectrum was obtained with a strong, sharp peak at 300\u00a0cm\u22121, assigned to a CH3 torsion from a metal-bound methyl species [28]. This is higher in energy than a typical C-CH3 group and makes the presence of an ethylidyne species unlikely [29]. Although, specific surface coverage and adsorption onto a metal surface could shift features, making definitive exclusion of ethylidyne not possible. DFT studies, as part of potential future work, to provide simulated spectra present an opportunity to resolve this. The assignment (of a methyl group) was further confirmed from the higher energy transfer data recorded on MAPS (Fig. 3b) that shows a peak at 2940\u00a0cm\u22121 indicating a CH stretch from a saturated hydrocarbon [30]. It should also be noted that there is no apparent signal around 3400\u00a0cm\u22121 that would indicate the presence of hydroxyl groups. Regarding the other modes seen on TOSCA, the 85\u00a0cm\u22121 mode seen at lower temperature dosing was retained, while the hydrogen on the threefold Co site appears slightly decreased, and the intensity of the broad band between 550 and 900\u00a0cm\u22121 increases. Due to its low effective mass physisorbed ethylene would show molecular recoil in this energy range which results in no well-defined peaks [27] and represents a transfer of momentum from the neutron to the sample without excitation of vibrational modes. CH modes from hydrocarbon fragments would be expected in this region, but other interpretations can also be considered.A recent INS study of hydrogen adsorption on Ni and Co catalysts by Davidson et al. [16] has tentatively suggested the formation of bound molecular hydrogen on cobalt that resemble Kubas complexes [31]. In this model the dihydrogen molecule is chemisorbed to the metal through an \u03b72 \u03c3-bond. This increases the HH bond length, causing the H2 rotational band to shift to 46\u00a0cm\u22121 from the 118\u00a0cm\u22121 of pure hydrogen. If this is also the origin of the 85\u00a0cm\u22121 band on our catalyst, it suggests that Co can support a stable intermediate in the dissociation to the hydride species that may have a strong influence on catalytic reaction mechanisms. Furthermore, it suggests that this interaction strength may easily be followed by the frequency of this mode and can be directly linked to the length of the HH bond, in this case calculated at 0.96\u00a0\u00c5. Bound molecular hydrogen would lead to increased structure in the obtained spectra [32] which was not reported in the study by Davidson et al. However, comparison with H2 complexes in solution [26] suggest that the symmetric HMH stretch and the rock, wag and torsion modes would fall in the range of 950\u2013400\u00a0cm\u22121, which could explain at least some of the intensity of the broad feature in the observed spectra. Unfortunately, this cannot be taken in any way as a confirmation of the chemisorbed molecular hydrogen. Similar features where they appear in [16] are assigned to bound M-H species on different adsorption sites, which is an equally valid interpretation for our data. Furthermore, the presence of overlapping hydrocarbon modes in this region further complicates the assignment.The spectra thus show that mixed-mode adsorption is occurring on the catalyst at lower temperature. This does not contradict the possibility of a Kubas-type chemisorbed molecular hydrogen and metal hydride, although the more conventional assignment would be for associatively and dissociatively chemisorbed ethylene. The second dose at higher temperature shows that the adsorbates have reacted, with strong evidence for a methyl species appearing at 300\u00a0cm\u22121 in the spectrum with other features that may be assigned to complexed molecular hydrogen and/or bound saturated hydrocarbon fragments. Interestingly, Weststrate and co-workers on ethylene adsorption on single crystal (0001) Co surfaces under UHV, suggests that at <\u221275\u00a0\u00b0C ethylene should associatively absorb [8,9]. While at 30\u00a0\u00b0C ethylene would dehydrogenate into adsorbed hydrogen and acetylene on the skeletal cobalt. Observations, from INS on the skeletal cobalt, of dissociatively chemisorbed species at low temperature, and the clear presence of a methyl group at higher temperature, demonstrates a notable deviation from the surface chemistry seen by Weststrate.Given the dramatic differences in reaction conditions and catalytic surface, this discrepancy is perhaps unsurprising, with several possible factors being possibly responsible. The higher reactivity of certain sites on the polycrystalline surface could reduce the temperature, from the 125\u00a0\u00b0C seen by Weststrate, required for the dehydrogenation of acetylene to form surface coke [8]. The role of the diverse surfaces, steps and kinks on the Co mediated Fischer-Tropsch mechanism is unresolved and remains an area of debate. Theoretical studies have shown that the relative stability of C1Hx vs C2Hx is dependent on the surface termination, with close packed surfaces favouring C2Hx and higher index surfaces C1Hx [33]. The results invoking a concept of an ensemble of active sites where CO activation occurs at high index sites and chain growth on close packed surface. The propensity of higher index planes towards CO activation but also CC breaking being reflected in the significant dehydrogenation activity and probable CC cleavage (re. evidence of methyl groups) observed in this study using the polycrystalline skeletal catalyst vs a close packed Co (0001) surface. Madey and co-workers showed that the INS spectrum of ethylene adsorbed on skeletal Ni changed dramatically on heating to room temperature due to dehydrogenation, although the specific species could not be identified [17]. Interestingly, they note that this dehydrogenation was far more significant than observed on analogous EELS studies on a Ni(111) surface and postulated that steps and defects on the skeletal catalyst reduced the activation energy of dehydrogenation. To further clarify the presence of multiple dehydrogenated species, a TPR was performed on a skeletal cobalt that had been dosed with ethylene at 30\u00a0\u00b0C (Fig. 2). Two significant features were observed between 200 and 250\u00a0\u00b0C with a small 3rd feature centred at 300\u00a0\u00b0C. Unfortunately, given that a TCD was used instead of mass spectrometry, it was not possible to identify the desorbed hydrogenated species. However, the results confirm that multiple dehydrogenated carbon species were formed during ethylene adsorption. Further, it is anticipated that ambient pressure hydrogenation of coke species would require higher temperatures than observed.Alternatively, the potential hydroxylation/oxidation of the surface through contamination which was noted to be highly facile by Davidson et al. could have reduced surface coverage and altered the reactivity of adsorbed ethylene species. Water and surface hydroxyl species, present under high conversion in Fischer-Tropsch influence Co structure and oxidation state and have been implicated in a possible CO activation pathway [34]. However, little evidence of hydroxyl groups can be found from INS of the catalysts after reduction; vide supra. Further, TPR analysis post ethylene dosing showed no reduction features in the temperature range associated with original hydroxylated sample. Yet, analysis of the reduced sample by XPS appears to contradict these findings. As indicated from the Co 2p spectra XPS (Fig. 4\n) shows that the surface of the sample was completely oxidised to Co2+, with a multiplet peak at 780.34\u00a0eV and satellite structure indicative of Co(OH)2 [35]. The O1s spectra further indicate surface metal oxide and hydroxide with peaks at 529.94 and 532.06\u00a0eV respectively. However, a note of caution is required, as Davidson et al. highlighted the potential for rapid surface hydroxylation on sample transfer, even within a glovebox, as done for the present XPS analysis. A reference TPR experiment where the skeletal cobalt was reduced, then cooled to room temperature and stored under a static environment for 60\u00a0min, before being exposed to a second TPR step is illuminating. As shown in Fig. 2, this procedure results in a reduction feature at broadly the same temperature as that seen for the as synthesised material, indicating that simple storing under a static inert environment provided sufficient oxygen/water ingress to facilitate surface hydroxylation. Consequently, XPS is shown to be sensitive to the effect of handling the sample, even when carried out within a glovebox. It can therefore be concluded that experiments must be conducted in situ and that the surface of sample is probably free of hydroxyl species within such experiments (INS/QENS/TPR).Lastly, the diffusional properties of adsorbed species were investigated by QENS. Fig. 5\n shows the peak intensity integrated between \u221217.5 and 17.5 \u03bceV and normalised to this value at the lowest temperature, representing the total elastic scattering from the sample as temperature varies. Very little change is seen in the undosed cobalt; however, the dosed material show significant decreases around 75\u00a0K. This indicates the onset of motion that may be considered a phase change. Pure ethylene has a melting point of 104\u00a0K, so the depression of this melting point demonstrates there is an interaction between the adsorbate and the cobalt catalyst. Full fitting of the Q-resolved data at 300\u00a0K to obtain diffusion parameters was attempted, but the diffusion was too rapid for the instrument to measure, which is suggestive of rapid rotational motions of the methyl group confirmed by INS. Interestingly, repeating the experiment with co-dosed CO increased the temperature when elastic intensity decreased and implies that CO is inhibiting the motion of surface bound CxHy species. This result suggests that future INS experiments should be performed with co-dosing of CO to identify possible changes in adsorbate speciation.A skeletal cobalt catalyst was synthesised and used in a preliminary INS/QENS study of adsorbed ethylene. We show that the material requires substantial reduction prior to use in adsorption studies. While samples were stable under a reducing atmosphere, any transfer of the sample, even within a glovebox, was sufficient to result in complete surface hydroxylation. In addition, this prerequisite reduction step can easily result in significant sintering of the skeletal cobalt. Recorded INS spectra had relatively weak and broad signals possibly associated with a small uptake of ethylene, due to sintering during reduction. It should also be noted that these data were the result of a subtraction to remove the strong background signal from both the Co sample and the Inconel sample holder which is unfortunate and could lead to distortion of the spectra. However, without adequate sample containment the sample will oxidise, and this treatment is therefore unavoidable. INS showed that the adsorbed hydrogen and CxHy environment is significantly more complex on a skeletal cobalt catalyst than analogous single crystal UHV studies, with a range of CH and adsorbed hydrogen species being observed on the former. This is hypothesised as being due to the presence of highly reactive defects and high index surfaces on the skeletal cobalt catalyst. Dissociated hydrogen adsorbed in a 3-fold site was observed in addition to a possible Kubas-type chemisorbed molecular hydrogen. CH fragments could also possibly be observed, with dosing at room temperature resulting in definitive methyl species. Distinction between M-CH3 or M-C-CH3 cannot be made from the literature available and requires further investigation. QENS studies confirmed that ethylene reacted with the skeletal cobalt surface, although diffusional parameters were too rapid to be measured. The co-dosing of ethylene with CO inhibited the motion of the adsorbed ethylene or dehydrogenated species.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors wish to thank Prof. Stewart Parker for discussions about neutron scattering and performing the MAPS measurement. The ISIS Neutron and Muon source is thanked for their grants of beam time for the neutron experiments (doi: https://doi.org/10.5286/ISIS.E.RB2010089, doi: https://doi.org/10.5286/ISIS.E.RB2010090 and doi: https://doi.org/10.5286/ISIS.E.RB2190135-1). The X-ray photoelectron (XPS) data collection was performed at the EPSRC National Facility for XPS (\u201cHarwellXPS\u201d), operated by Cardiff University and UCL, under Contract No. PR16195. UK Catalysis Hub is kindly thanked for resources and support provided via our membership of the UK Catalysis Hub Consortium and funded by EPSRC grant: EP/R026939/1, EP/R026815/1, EP/R026645/1, EP/R027129/1 or EP/M013219/1 (biocatalysis). Edward Jones acknowledges funding for this work through his studentship received from EPSRC Sustainable Hydrogen CDT (EP/S023909/1).", "descript": "\n                  To bridge the materials gap of single crystal work on the Fischer-Tropsch chain-growth mechanism, ethylene adsorption on a model skeletal cobalt catalyst was studied. Speciation and mobility of surface species were characterised using inelastic (INS) and quasi-elastic (QENS) neutron scattering. INS spectra demonstrated that highly reactive sites facilitated ethylene dehydrogenation at lower temperature than in single crystal studies. Adsorbed hydrogen was assigned to Co3H and potentially a Kubas species. After adsorption at 30\u00a0\u00b0C, methyl groups were identified. CO co-adsorption was shown to modify the dynamics of the adsorbed species. Further analysis demonstrated the sensitivity of skeletal cobalt to surface hydroxylation.\n               "}
{"full_text": "Over the past decades, the global environment has been rapidly deteriorated due to the increasing use of fossil fuels [1]. In the current age, it is highly anticipated to develop green energy devices. Rechargeable metal-air batteries are the environmental-friendly energy conversion devices with high performance and practical feasibility [2\u20134]. Among them, the rechargeable zinc-air battery (RZAB) is particularly attractive [5], with several advantages such as intrinsic safety, high specific capacity, and high theoretical energy (1086 Wh kg\u22121, including O2) [6,7]. The air\u2013cathode of RZAB needs a bi-functional catalyst having activities for both using O2 (when discharging) and producing O2 (when charging) efficiently, that is, oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) [8]. Commercially available catalysts are usually platinum on carbon (Pt/C) for ORR and iridium dioxide or ruthenium dioxide for OER, which all involve noble metals and exhibit mono-functional catalytic performance [4,5]. Such disadvantages limit the wide use of RZAB.As alternatives to noble metallic elements, first-row transition metal oxides and their derivatives are considered as good ORR/OER catalysts [9\u201311]. However, simple metal oxides often exhibit relatively low electrical conductivity [12] and small specific surface area (<400\u00a0m2 g\u22121) [13]. Hence metal oxides should be integrated with conductive materials for the improved performance. It is known that conductive carbon black (CB) is commercially mature activated carbon having high electrical conductivity and plentiful pores, which provide large surface areas for chemical modifications [12,14,15]. In this sense, the combination of conductive carbon with metal oxides can enhance the ORR/OER electrochemical performance [16].In general terms, nevertheless, the degree of combination of metal oxides and carbon is relatively low, leading to the easily leaching of metal oxides from carbon and the resultant unstable performance. That being so, it is reasonable to look for a binder that helps to anchor metal oxides to the carbon surface, and such binder can be polymers. In this scheme, monomers are first anchored within a porous carbon matrix followed by the polymerization of them, forming the polymer-grafted carbon, and then heteroatom groups in polymer chains are able to act as ligand donors to bind metallic ions in a coordination mode, making possible the subsequent formation of metal-oxide active sites anchored on polymer-grafted carbon.Unfortunately, a substantial number of polymers are prepared using radical initiators that include organic/inorganic peroxide and azo compounds, which are either dangerously explosive or toxic to human health and the environment [17]. On the other hand, acoustic cavitation generated by ultrasonic irradiation can produce numerous micro-bubbles and cause violent collisions of particles, thereby leading to high temperatures locally (>5000\u00a0K) and a large number of free radicals [18]. It has been demonstrated that free radicals produced by ultrasound are able to induce the polymerization of monomers [19\u201321]. Moreover, the ultrasonic environments are favourable for the formation of metal oxides without using extra chemical reagents or apparatus [22,23]. From the green chemistry perspective, ultrasonication is an environmentally friendly and highly energy efficient method for the preparation of polymers and metal oxides.In this work we introduce the convenient route to the effective combination of earth-abundant tri-metallic oxide and conductive CB using an amide-type polymer induced by ultrasonic cavitation. The monomer N-isopropyl acrylamide (NIPAm) undergoes the polymerization using ultrasonication instead of any radical initiator. The polymer chains formed are grafted on the surface of CB, and the amide groups in the polymer can bind three types of metal ions (i.e., Mn2+, Ni2+ and Fe2+) through the efficient coordination, resulting in the production of metal oxide by the second round of ultrasonic irradiation. Thanks to the amide-type polymer as a binder, the Mn-Ni-Fe tri-metallic oxide anchored on polymer-grafted CB has both the appropriate amount of metal-oxide active sites and a sufficient number of hierarchical pores, leading to the enhanced ORR/OER electrocatalytic performance compared to its ultrasonication-free counterpart and commercial catalysts, which proves to be suited as a redox bi-functional catalyst in the RZAB. Accordingly, this work provides appealing insight into the effective combination of the two inherently incompatible parts for the construction of composite materials.N-isopropyl acrylamide (NIPAm, 98%, Adamas-beta), ferrous sulfate heptahydrate (FeSO4\u00b77H2O, Adamas-beta), manganese (II) chloride (MnCl2, 99%, Adamas-beta), nickel (II) chloride hexahydrate (NiCl2\u00b76H2O, 99%, Adamas-beta), isopropyl alcohol (C3H8O, 99.7%, Sinopharm Chemical), Ketjenblack\u00ae CB (EC300J conductive carbon black, Japan Lion), Nafion\u00ae perfluorinated resin solution (5\u00a0wt%, Sigma-Aldrich), \u03b1-Al2O3 with diameter of 50\u00a0mm (99%, Tianjin Aida), platinum on carbon (Pt/C, 10\u00a0wt%, Sigma-Aldrich), ruthenium oxide (RuO2, 99.95%, Adamas-beta), potassium hydroxide (KOH, >90%, General reagent), zinc acetate (Zn(ac)2, 99.5%, Adamas-beta), high-purity nitrogen and oxygen gases (99.999%, Xuzhou Special Gases), and ultrapure water (18.2 M\u03a9 cm, Sartorius arium).Field-emission scanning electron microscopy (FESEM) images were acquired using Quanta FEG 250. Transmission electron microscopy (TEM) as well as high-resolution TEM (HRTEM) images were obtained on FEI Tecnai TF30. Nitrogen adsorption\u2013desorption isotherms were obtained by Quantachrome Autosorb iQ. Pore width distributions were gained based on the quenched solid density functional theory (DFT) model. X-ray diffraction (XRD) patterns were acquired by PANalytical X'pert3 Powder using the Cu K\u03b1 radiation (\u03bb\u00a0=\u00a01.54178\u00a0\u00c5). X-ray photoelectron spectroscopy (XPS) measurements were conducted on Thermo Fisher ESCALAB 250Xi using the Al K\u03b1 radiation (1486.6\u00a0eV) and binding energies were calibrated according to the C 1\u00a0s peak (284.8\u00a0eV). An ultrasonic exfoliation instrument with six-sided distributed ultrasonic transducers (Scientz-CHF-5B, Ningbo Scienz) was used to intensify the dispersion of CB. An ultrasonic instrument equipped with a microwave generator (Scientz-IIDM, Ningbo Scienz) was employed to synthesize catalysts.From the beginning, 0.5\u00a0g of CB was dispersed in 100\u00a0mL of ethanol in a 250\u00a0mL sealed beaker and ultrasonically treated using the ultrasonic exfoliation instrument for half an hour (20\u00a0kHz, 576\u00a0W). Then, 1.5\u00a0g of NIPAm was dissolved in 100\u00a0mL of ultrapure water, which was added to the black suspension of the dispersed CB. After that, the mixture was degassed by high-purity N2 using a capillary glass-made gas inlet for 10\u00a0min at a flow rate of 50 sccm. Further, the degassed mixture was transferred to a three-neck flask and placed in the ultrasound/microwave instrument. The ultrasonic horn was inserted 5\u00a0cm below the liquid level through the centre neck of the three-neck flask. As for the other two necks, one was linked to a reflux condensation maintained at 0\u00a0\u00b0C, and the other was sealed using a polytetrafluoroethylene (PTFE) cap. The mixture was then treated under a 500\u00a0W ultrasound for half an hour with the assistance of microwave (200\u00a0W) at the initial 500\u00a0s for rapid warming. Subsequently, 10\u00a0mL of 0.05\u00a0M NiCl2, 10\u00a0mL of 0.05\u00a0M FeSO4 and 10\u00a0mL of 0.05\u00a0M MnCl2 solutions were added to the mixture, followed by the second round of ultrasonic irradiation for another half an hour (20\u00a0kHz, 500\u00a0W). Finally, the ultrasonically treated mixture was centrifuged and dried at 80\u00a0\u00b0C overnight. The obtained material was denoted U-MnNiFe@NCB. To investigate the ultrasonic effect, a sample was prepared using stirring (at 80\u00a0\u00b0C) to replace the ultrasonic treatments mentioned above, leaving the other conditions unchanged. This control sample was denoted MnNiFe@NCB.Electrochemical tests were conducted using an electrochemical workstation (Iviumstat.h, Ivium Technologies) at 25\u00a0\u00b0C. In a three-electrode configuration, a rotating disc electrode with glassy carbon surface (GC-RDE) connected with a rotator (AFMSRCE, Pine Research) was employed to load a catalyst, acting as a working electrode. Additionally, a Ag/AgCl electrode (3.5\u00a0M KCl) connected to the main cell through a Luggin capillary was employed as a reference electrode, and a platinum foil was used as a counter electrode. All the potentials measured in this paper were converted to the reversible hydrogen electrode (RHE) based on the following equation: E(RHE)\u00a0=\u00a0E(Ag/AgCl)\u00a0+\u00a00.2046\u00a0+\u00a00.059\u2219pH. The catalyst ink was made by mixing 3.0\u00a0mg of the catalyst, 70\u00a0\u03bcL of isopropyl alcohol, 170\u00a0\u03bcL of ultrapure water as well as 10\u00a0\u03bcL of 5\u00a0wt% Nafion\u00ae perfluorinated resin solution together under ultrasonication for half an hour to form a homogenized mixture. Meanwhile, the GC-RDE (6\u00a0mm in diameter for GC) was polished with \u03b1-Al2O3 and washed thoroughly with ultrapure water and ethanol. Afterwards, 8\u00a0\u03bcL of the catalyst ink was transferred onto the polished surface of GC-RDE using a pipette and dried naturally. The catalyst loading was 0.49\u00a0mg\u00a0cm\u22122. Regarding the inks of the commercial catalysts, 1\u00a0mg of 10\u00a0wt% Pt/C or 2\u00a0mg of RuO2 was mixed with 10\u00a0\u03bcL of Nafion\u00ae perfluorinated resin solution, 170\u00a0\u03bcL of isopropyl alcohol and 70\u00a0\u03bcL of ultrapure water under ultrasonic irradiation for half an hour. The loadings of Pt/C or RuO2 were 0.16\u00a0mg\u00a0cm\u22122 and 0.32\u00a0mg\u00a0cm\u22122, respectively.Regarding ORR experiments, the gas (O2 or N2) with a flow rate of 80 sccm was first fed into the electrolyte (0.1\u00a0M KOH) for half an hour and then through the headspace above the electrolyte solution during measurements. Cyclic voltammetry (CV) measurements were undertaken to activate catalysts with a scan rate of 50\u00a0mV\u00a0s\u22121 for 50 cycles prior to electrocatalytic measurements. Linear sweep voltammetry (LSV) experiments at different rotating rates (400\u20132025\u00a0rpm) with a scan rate of 10\u00a0mV\u00a0s\u22121 were conducted. As for LSV curves, the currents obtained in the N2-saturated solution were subtracted from those in the O2-saturated solution in order to eliminate strong capacitative currents from porous carbon. The Koutecky-Levich (K-L) equation shown below was employed to obtain the electron-transfer number [24]:\n\n(1)\n\n\n\n1\nj\n\n=\n\n1\n\nj\nk\n\n\n+\n\n1\n\nB\n\n\n\u03c9\n\n\n1\n/\n2\n\n\n\n\n\n\n\nwhere \u03c9 is rotating rate (rpm), jk\n is kinetic current density (mA cm\u22122), j is total current density (mA cm\u22122), and B is the K-L slope, which can be expressed as [24,25]:\n\n(2)\n\n\nB\n=\n0.2\n\n\nnFC\n\n0\n\n\nD\n\n0\n\n\n2\n/\n3\n\n\n\n\nv\n\n\n-\n1\n/\n6\n\n\n\n\n\nwhere F is the Faraday constant (96485C mol\u22121), n is the electron-transfer number, C\n0 is the saturated O2 concentration (1.2\u00a0\u00d7\u00a010\u20133 mol L\u20131), D\n0 is the diffusion coefficient of O2 (1.9\u00a0\u00d7\u00a010\u20135 cm2 s\u22121), and v is the kinematic viscosity in 0.1\u00a0M KOH (0.01\u00a0cm2 s\u22121). For the OER measurements, the LSV curves were recorded with a scan rate of 10\u00a0mV\u00a0s\u22121 at a rotating rate of 1,600\u00a0rpm after the electrolyte (0.1\u00a0M KOH) was saturated with N2 with a flow rate of 80 sccm. The iR-drop compensation was made to minimize the ohmic potential loss of electrodes due to the solution resistance.Custom-made RZABs comprising air-cathodes and zinc-plate anodes were assembled, each with three poly(methyl methacrylate) (PMMA) plates. The air\u2013cathode was made by pressing catalyst-loaded carbon paper, air-breathable water-proof layer (placed in the middle), and nickel foam at 10\u00a0MPa. The catalyst ink for the RZAB measurements was prepared by mixing 3\u00a0mg of U-MnNiFe@NCB, 70\u00a0\u03bcL of isopropyl alcohol, 140\u00a0\u03bcL of ultrapure water and 10\u00a0\u03bcL of 5\u00a0wt% Nafion\u00ae perfluorinated resin solution together under ultrasonication for half an hour. As for the Pt/C-RuO2 catalyst, 1\u00a0mg of 10\u00a0wt% Pt/C and 2\u00a0mg of RuO2 were mixed in 70\u00a0\u03bcL of ultrapure water, 170\u00a0\u03bcL of isopropyl alcohol and 10\u00a0\u03bcL of 5\u00a0wt% Nafion\u00ae perfluorinated resin solution under ultrasonication for half an hour. For either of these catalysts, the catalyst area on the electrode is 1.77\u00a0cm2, and the mass loading of catalyst is 1.69\u00a0mg\u00a0cm\u22122 in each battery. The electrolyte comprising 0.2\u00a0M Zn(ac)2 and 6.0\u00a0M KOH was circulated by a peristaltic pump. The specific capacity was calculated by the following equation [26]:\n\n(3)\n\n\n\nC\ns\n\n=\n\n\nI\n\u0394\nt\n\n\nm\n\nZn\n\n\n\n\n\n\nwhere I is current (mA), C\ns is specific capacity (mAh g\u22121), \u0394t is time from the beginning until zinc was totally consumed (s), mZn\n is the consumed zinc mass (g).The pristine CB undergoes the two-round ultrasonic treatments, being converted to the polymer/metal oxide-modified material U-MnNiFe@NCB, as shown in Fig. 1\na. The FESEM images of CB and U-MnNiFe@NCB at different magnifications (Fig. 1b-g and Fig. S1) show that the two samples have the similar morphology of nanosized carbon granules, suggesting that the ultrasonic modification does not change the basic structure of the CB. Free radicals generated by the ultrasonic cavitation are able to induce the polymerization of NIPAm, resulting in the formation of PNIPAm that is grafted onto the carbon granules. The grafted polymer chains contain the structural units terminated by amide groups, which can act as good donor groups bonded to metal ions. To determine whether PNIPAm was successfully grafted and its content, U-MnNiFe@NCB, MnNiFe@NCB and CB were vacuum heated at 250\u00a0\u00b0C for 6\u00a0h with a ramp rate of 10\u00a0\u00b0C\u00a0min\u22121 separately. As shown in Table S1, the grafting proportion of U-MnNiFe@NCB is ca. 23.0%, revealing that its surface is indeed modified with sufficient quantities of PNIPAm by the ultrasonication. By comparison, the grafting proportion of MnNiFe@NCB is only 6.7% (Table S1), showing the low content of volatile organic components in the material without undergoing the ultrasonication. This difference in the grafting proportions can be explained by the fact that the polymer PNIPAm with the network of carbon skeletons formed using the ultrasound favour the adherence to the surface of porous carbon compared to the monomers.The TEM images of U-MnNiFe@NCB at different magnifications (Fig. 2\na and b) also exhibit its typically granular morphology, consistent with the FESEM observations. Further, the HRTEM image of U-MnNiFe@NCB (Fig. 2c) shows its overwhelmingly amorphous feature, together with a few lattice fringes. A magnified view of the lattice fringes by the inverse fast Fourier transform (IFFT) of the circled region is shown in the inset of Fig. 2c. The d-spacing is 0.34\u00a0nm, which can be ascribed to the (002) surface of graphitized carbon [27]. Also, another two types of lattice fringes are found in the HRTEM image of U-MnNiFe@NCB (Fig. 2d). As magnified in the insets of Fig. 2d, the d-spacings of these lattice fringes include 0.21\u00a0nm and 0.30\u00a0nm, which should be ascribed to the metal oxide formed by the ultrasonication. Fig. 2e\u2013k exhibit the high-angle annular dark-field (HAADF)-TEM and energy-dispersive X-ray spectroscopy (EDS) elemental mapping images of U-MnNiFe@NCB. The three metallic elements Fe, Mn and Ni coexist with O, indicating the formation of the tri-metallic oxide on the CB.The XRD patterns of U-MnNiFe@NCB, MnNiFe@NCB and CB are exhibited in Fig. 3\n. The widened bands centred at 24.0\u00b0 are seen from all the three patterns, corresponding to the (002) plane of graphitized carbon [28] and consistent with the HRTEM observation (Fig. 2c). The widened characteristics of diffraction peaks also confirm the three materials possess the amorphous carbon matrices. By comparison with MnNiFe@NCB and CB, U-MnNiFe@NCB has the two additional small yet sharp peaks centred at 30.3\u00b0 and 43.4\u00b0 observable from its XRD pattern, which can be ascribed to the (220) and (400) planes of Mn-doped NiFe2O4, respectively [29,30]. The d-spacings of the (220) and (400) planes of Mn-doped NiFe2O4 are 0.30\u00a0nm and 0.21\u00a0nm, respectively [29,30], which are consistent with the HRTEM data (Fig. 2d). Thus, it can be inferred that the metal ions Ni2+, Fe2+ and Mn2+ attached to the polymer-grafted carbon granules by the amide groups are converted to Mn-doped NiFe2O4 under the ultrasonication.The elemental compositions of Mn-doped NiFe2O4 anchored on the polymer-grafted CB, along with the control samples MnNiFe@NCB and CB, are evaluated using the XPS technique. The wide-scan XPS spectra of the three materials are shown in Fig. S2a and the corresponding elemental information is summarized in Table S2. The pristine CB contains only two elements C and O, whereas U-MnNiFe@NCB possesses the largest quantities of N, O, Mn, Ni and Fe among the three materials, and MnNiFe@NCB prepared without using the ultrasonication has the intermediate amounts of these elements, indicating that the grafted polymer chains produced by the ultrasonic cavitation, acting as adhesives, allow the CB to bind with the metallic elements effectively. Without the ultrasonication or any radical initiator, the polymerization of NIPAm could not be triggered spontaneously. Because the polymer PNIPAm formed in U-MnNiFe@NCB binds more tightly to the carbon surface than its monomer NIPAm as mentioned above, MnNiFe@NCB contains the lower amounts of potential N-/O-donor groups anchored on the surface of CB and thus the reduced number of the coordinated metal ions. For the non-metallic elements, the narrow-scan C 1\u00a0s spectrum of U-MnNiFe@NCB (Fig. 4\na) can be split into the four subpeaks C=C (284.7\u00a0eV) [31,32], C\u2013C/C\u2013O/C\u2013N (285.4\u00a0eV) [31], C=O (286.4\u00a0eV) [31] and O=C\u2013O (289.2\u00a0eV) [32]. Likewise, the narrow-scan N 1\u00a0s spectrum of U-MnNiFe@NCB (Fig. 4b) can be split into the three peaks N\u2013Q (quaternary nitrogen, 401.5\u00a0eV) [33], HNC=O (400.2\u00a0eV) [34] and metal-N (399.2\u00a0eV) [35], and its O 1\u00a0s counterpart (Fig. 4c) can also be decomposed into the three subpeaks O\u2013C=O (533.5\u00a0eV) [31], N\u2013C=O (532.0\u00a0eV) [36] and O2\u2013 (530.5\u00a0eV) [37]. By contrast, the C 1\u00a0s spectrum of CB that lacks nitrogen or metallic elements (Fig. S2b) can be split into the three subpeaks C=C (284.8\u00a0eV) [31], C\u2013C/C\u2013O (285.4\u00a0eV) [31] and O=C\u2013O (289.2\u00a0eV) [32], whereas its O 1\u00a0s counterpart (Fig. S2c) can also be split into the two peaks O\u2013C=O (533.5\u00a0eV) and C\u2013OH (532.3\u00a0eV) [31]. Indeed, the chemical bonding of the non-metallic elements is different between U-MnNiFe@NCB and CB, because of the existence of the grafted polymer and metal oxide in the former. Evidently, the existence of the metal-N bonding on the surface of U-MnNiFe@NCB shows that the metal ions are coordinated to the amide groups of the grafted PNIPAm. Additionally, the existence of the lattice oxygen O2\u2013 indicates the formation of transition metal oxide [37,38]. In other words, the metal ions bind to the N and O donor atoms of the amide groups that are part of the grafted polymer in the coordination mode, forming the metal oxide under the ultrasonication. Regarding the metallic elements, on the other hand, the high-resolution Mn 2p, Fe 2p and Ni 2p XPS spectra of U-MnNiFe@NCB (Fig. 4d\u2013f) show the typical characteristics of 2p3/2 Fe3+ (711.7\u00a0eV), 2p1/2 Fe3+ (724.9\u00a0eV), 2p3/2 Mn2+ (642.4\u00a0eV) and 2p3/2 Ni2+ (855.8\u00a0eV) [39,40]. The surface contents of Mn, Ni and Fe in U-MnNiFe@NCB are 0.28 at%, 0.23 at% and 0.91 at%, respectively, as summarized in Table S2; that is, the atomic ratio of Mn to Ni to Fe is 1:0.82:3.25. Thus, the chemical formula of Mn-doped NiFe2O4 anchored on the polymer-grafted CB synthesized by the ultrasonication can be written as Mn0.6Ni0.5Fe2O4, the formation of which should involve the following steps under the ultrasonic irradiation [41]:\n\n(4)\n\n\n\n\nH\n\n\n2\n\n\nO\n\n\n\u2192\n\n\n)\n)\n)\n\n\nH\n\u00b7\n+\nO\nH\n\u00b7\n\n\n\n\n\n\n(5)\n\n\n2\nH\nO\n\u00b7\n\n\n\u2192\n\n\n)\n)\n)\n\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\n\n\n(6)\n\n\n\n\n2\nF\ne\n\n\n2\n+\n\n\n+\n\nH\n2\n\n\nO\n2\n\n\n\u2192\n\n)\n)\n)\n\n\n\n\n2\nF\ne\n\n\n3\n+\n\n\n+\n2\n\n\nOH\n\n-\n\n\n\n\n\n\n\n(7)\n\n\n\n\nFe\n\n\n2\n+\n\n\n+\n\n\n2\nF\ne\n\n\n3\n+\n\n\n+\n8\n\n\nOH\n\n-\n\n\n\u2192\n\n)\n)\n)\n\n\n\n\nFe\n\n3\n\n\nO\n4\n\n+\n\n\n4\nH\n\n2\n\nO\n\n\n\n\n\n\n(8)\n\n\n\n\nNi\n\n\n2\n+\n\n\n+\n\n\n2\nH\n\n2\n\n\nO\n2\n\n+\n2\n\n\nOH\n\n-\n\n\n\u2192\n\n)\n)\n)\n\n\nN\ni\nO\n+\n\n\n3\nH\n\n2\n\nO\n+\n\nO\n2\n\n\n\n\n\n\n\n(9)\n\n\n3\n\n\nMn\n\n\n2\n+\n\n\n+\n\nH\n2\n\n\nO\n2\n\n+\n6\n\n\nOH\n\n-\n\n\n\u2192\n\n)\n)\n)\n\n\n\n\nMn\n\n3\n\n\nO\n4\n\n+\n\n\n4\nH\n\n2\n\nO\n\n\n\n\n\n\n(10)\n\n\n\n\n4\nF\ne\n\n3\n\n\nO\n4\n\n+\n3\nN\ni\nO\n+\n1.2\n\n\nMn\n\n3\n\n\nO\n4\n\n\n\u2192\n\n)\n)\n)\n\n\n\n\n6\nM\nn\n\n\n0.6\n\n\n\n\nNi\n\n\n0.5\n\n\n\n\nFe\n\n2\n\n\nO\n4\n\n\n\n\n\nIt can be seen that ultrasound plays important roles in the generation of the tri-metallic oxide: the ultrasonic cavitation induces the formation of free radicals that can produce peroxide, which reacts with the metal ions in a series of steps until the formation of Mn0.6Ni0.5Fe2O4. Without the ultrasound, metal oxide could hardly be obtained under the given conditions. This can also explain why the diffraction peaks of Mn0.6Ni0.5Fe2O4 or any other metal oxide are absent in the XRD pattern of MnNiFe@NCB obtained without employing the ultrasound (Fig. 3). Accordingly, the ultrasonication helps to induce not only the polymerization of NIPAm but also the formation of the tri-metallic oxide Mn0.6Ni0.5Fe2O4, and the grafted polymer chains act as binders to anchor the active sites (i.e. the metal oxide) to the porous electrode (i.e. the CB).In addition to active sites, porous properties have significant effects on the catalytic performance, and in this case the CB provides the porous domains for mass transport. N2 adsorption\u2013desorption isothermal experiments were conducted to investigate pore structures of the three materials. The isotherms of U-MnNiFe@NCB, MnNiFe@NCB and CB (Fig. 5\na) can be classified as Type I/II [42,43]. Specifically, the largely increased adsorption volumes at p/p0 \u2248 0 indicate the existence of micropores [44,45], the hysteresis loops at p/p0\u00a0=\u00a00.45\u20130.95 are caused by mesopores [46,47], and the large uptakes at p/p0\u00a0>\u00a00.95 imply the presence of macropores [48]. Both U-MnNiFe@NCB and MnNiFe@NCB are characteristic of hierarchical pores originating from the CB. The surface areas of U-MnNiFe@NCB, MnNiFe@NCB and CB are 414\u00a0m2 g\u22121, 615\u00a0m2 g\u22121, 725\u00a0m2 g\u22121, respectively. The change of surface areas reflects the degree of pore blockage by the grafted polymer chains, that is, the more the polymer on the surface, the higher the degree of pore blockage and the smaller the surface area, which is consistent with the corresponding grafting proportions of the materials mentioned above (Table S1). Moreover, the pore width distributions of U-MnNiFe@NCB, MnNiFe@NCB and CB based on the DFT method are shown in Fig. 5b. Overall, the three curves look similar to each other, confirming the pore structures of U-MnNiFe@NCB and MnNiFe@NCB stem from the characteristics of the CB and that no major changes are made to the pore width distributions after the surface modifications either with or without the ultrasonication. On the other hand, there is a noticeable difference between U-MnNiFe@NCB and CB in the micropore domains. U-MnNiFe@NCB has the two lower peaks centred at 0.8\u00a0nm and 1.4\u00a0nm than the CB, whereas their mesopore counterparts are comparatively close to each other, indicating that the ultrasonic modification has the stronger effect on micropores. This can also be quantitatively demonstrated according to the data from Table S3. The CB has the surface area and pore volume of micropores roughly three times greater than U-MnNiFe@NCB, while the differences between CB and U-MnNiFe@NCB in surface areas and pore volumes of mesopores are very little. It can then be inferred that the grafted polymer tends to \u2018bury\u2019 the micropore sites. In other words, the ultrasonic modification does not lead to a significant loss of mesopores that are responsible for mass transport, whereas micropores act as effective anchor points for the ultrasonically-induced polymer chains that capture the metal-oxide active sites, therefore achieving the fine balance between active sites and mass transport in this material design.Next, the electrochemical measurements were conducted to evaluate the catalytic activities towards ORR and OER for U-MnNiFe@NCB, MnNiFe@NCB and CB. Regarding the ORR performance, Fig. 6\na shows the LSVs of the three samples alongside commercial 10\u00a0wt% Pt/C. U-MnNiFe@NCB achieves an acceptable level of ORR catalytic activity, having the onset potential (Eonset\nORR) of 0.92\u00a0V, the half-wave potential (E1/2\nORR) of 0.73\u00a0V and the limiting current density of 5\u00a0mA\u00a0cm\u22122 at 1,600\u00a0rpm, superior to MnNiFe@NCB, CB and 10\u00a0wt% Pt/C. The inferior ORR performance of MnNiFe@NCB and CB can be ascribed to the lack of metal-oxide active sites in these two materials. The ORR electron-transfer numbers of U-MnNiFe@NCB, MnNiFe@NCB, CB and 10\u00a0wt% Pt/C are 3.7, 3.2, 2.3 and 3.6, respectively, according to the corresponding K-L plots (Fig. 6b and Fig. S3), showing the ORR by U-MnNiFe@NCB is predominantly the 4-electron process that is energetically favourable for electrochemical devices. Moreover, the OER catalytic performance of U-MnNiFe@NCB is compared with that of MnNiFe@NCB, CB and RuO2. The LSVs of the four materials concerning their OER behaviours are exhibited in Fig. 6c. U-MnNiFe@NCB shows the onset potential (Eonset\nOER) of 1.64\u00a0V at 10\u00a0mA\u00a0cm\u22122, more negative than MnNiFe@NCB (1.79\u00a0V), CB (1.89\u00a0V) and RuO2 (1.67\u00a0V), which indicates the better OER performance of U-MnNiFe@NCB than the other three samples [49]. This can also be confirmed by its minimal overpotential requirement for OER. The OER overpotential of U-MnNiFe@NCB (at 10\u00a0mA\u00a0cm\u22122) is 410\u00a0mV, the smallest value among the samples (MnNiFe@NCB: 560\u00a0mV; CB: 660\u00a0mV; RuO2: 440\u00a0mV). Furthermore, considering the ORR/OER activities holistically (Fig. 6d), U-MnNiFe@NCB exhibits the superior bi-functional performance to the commercial Pt/C and RuO2. The \u0394E gap (Eonset\nOER\u2013E1/2\nORR) can reflect the ORR/OER performance of electrocatalysts: the wider the \u0394E gap, the less efficient the bi-functional catalyst [50]. As shown in Fig. 6d, the \u0394E gap of U-MnNiFe@NCB is the narrowest among the catalysts studied, showing its reasonable performance towards ORR/OER. Clearly, the major reason why U-MnNiFe@NCB exhibits the much improved bi-functional ORR/OER performance compared to MnNiFe@NCB is the much larger amount of metal-oxide active sites (produced by the ultrasonication) anchored on the surface of CB by the grafted polymer chains (also produced by the ultrasonication) for U-MnNiFe@NCB compared to its non-ultrasonically treated counterpart.Custom-made RZABs were designed to evaluate the bi-functional ORR/OER electrocatalytic performance of U-MnNiFe@NCB in real electrochemical devices. Fig. 7\na shows the schematic illustration of the two zinc-air batteries connected in series, each made of two electrodes\u2014a zinc-plate anode and an air\u2013cathode. The electrolyte that contains 0.2\u00a0M Zn(ac)2 and 6.0\u00a0M KOH flows in each battery by an external peristaltic pump. The air\u2013cathode is composed of nickel foam, air-breathable water-proof layer and carbon paper loaded with a catalyst. Fig. 7b presents the open-circuit voltage (OCV) performance of the two RZABs using U-MnNiFe@NCB and the mixture of 10\u00a0wt% Pt/C and RuO2 as the cathode catalysts (abbreviated as the U-MnNiFe@NCB battery and the Pt/C-RuO2 battery respectively). It can be seen that the U-MnNiFe@NCB battery exhibits an OCV of 1.47\u00a0V, the voltage being slightly larger than the Pt/C-RuO2 battery (1.43\u00a0V). Additionally, Fig. 7c shows the charging/discharging polarization curves of the U-MnNiFe@NCB and Pt/C-RuO2 batteries. There is a larger voltage gap between the charging and discharging data of the Pt/C-RuO2 battery compared to its U-MnNiFe@NCB counterpart, indicating the enhanced performance of the latter [3]. As exhibited in Fig. 7d, the power density of the U-MnNiFe@NCB battery achieves the maximum (96 mW cm\u22122), the value being much larger than for the Pt/C-RuO2 battery (45 mW cm\u22122). Furthermore, the specific capacity of the U-MnNiFe@NCB battery was measured to be 746.6 mAh gZn\n\u20131 (Fig. S4), which is larger compared to the Pt/C-RuO2 battery (630.2 mAh gZn\n\u20131). Indeed, U-MnNiFe@NCB proves to be the fine bi-functional electrocatalyst for RZABs. Finally, the stability test of the U-MnNiFe@NCB battery was undertaken at a current density of 5\u00a0mA\u00a0cm\u22122 with repeatedly 5-min charging and 5-min discharging for 160 cycles (96,000\u00a0s in total). The result in Fig. 7e shows that the voltage gap between charging and discharging undergoes almost no obvious change (from 0.73\u00a0V at the beginning to 0.76\u00a0V in the end), revealing the good stability of U-MnNiFe@NCB. Fig. 7f shows the demonstration of a white LED light powered by the two U-MnNiFe@NCB batteries connected in series. Though still in the infancy, these gadgets are instructive as the non-noble metallic oxides produced ultrasonically have the comparable bi-functional catalytic activities to the noble metal counterparts.To summarize, this work demonstrates the important roles of ultrasonication in the fabrication of metal oxide/porous carbon composite materials in which polymer chains act as binders. Reactive free radicals generated by ultrasonic cavitation not only help to polymerize monomers without using dangerous radical initiators but also favour the formation of the Mn-Ni-Fe tri-metallic oxide that is otherwise impossible to obtain without ultrasonication. Our investigations have also shown that ultrasonically induced polymers are inclined to anchor within micropore domains instead of mesopores, which would not block mesopore pathways. This is particularly important, because the increase in the number of active sites (metal oxide) does not come into conflict with the enhancement in mass transport (relevant to mesopores) in this material design. With numerous active sites and fast mass transport, U-MnNiFe@NCB possesses the superior bi-functional ORR/OER electrocatalytic performance to its ultrasonication-free counterpart and commercial catalysts. Moreover, two in-series connected RAZBs using U-MnNiFe@NCB as air-cathodes are capable of powering small electronics, demonstrating its potential to work well in a real device. As a consequence, this work provides useful frameworks of ideas for understanding multiple effects of ultrasonication on the assembly of composite materials that comprise inherently incompatible components.\nBolin Jin: Data curation, Formal analysis, Investigation, Validation, Visualization, Writing \u2013 original draft, Writing \u2013 review & editing. Peiyao Bai: Formal analysis, Investigation, Validation, Visualization, Writing \u2013 original draft. Qiang Ru: Formal analysis, Investigation, Validation. Weiqi Liu: Formal analysis, Validation. Huifen Wang: Formal analysis, Visualization. Lang Xu: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing \u2013 original draft, Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Natural Science Foundation of China (51702358), the Natural Science Foundation of Jiangsu Province (BK20170281) and the Fundamental Research Funds for the Central Universities (2019ZDPY02). L.X. holds the Jiangsu Specially-Appointed Professorship.Supplementary data to this article can be found online at https://doi.org/10.1016/j.ultsonch.2021.105846.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  As a promising electrochemical energy device, a rechargeable zinc-air battery (RZAB) requires cost-effective cathode catalysts for oxygen reduction reaction (ORR) and oxygen evolution reaction (OER). Some earth-abundant transition metal oxides have certain levels of bi-functional ORR/OER catalytic activities yet low electronic conductivities. The addition of high-electronic-conductivity material such as carbon black could result in another problem because there is low compatibility between metal oxide and carbon. In this work, polymer chains are ultrasonically prepared to act as binders to anchor metal-oxide active sites to porous domains of carbon black. The monomer N-isopropyl acrylamide is polymerized under ultrasonication instead of using conventional radical initiators which are dangerous and harmful. Reactive free radicals produced by ultrasonic irradiation can also help to form the Mn-Ni-Fe tri-metallic oxide. Thus, aided by the amide-type polymer as an adhesive, the tri-metallic oxide anchored on polymer-grafted carbon black prepared by ultrasonication possess a large number of metal-oxide active sites and hierarchical pores, contributing substantially to the enhanced ORR/OER electrocatalytic performance in the RZABs. Accordingly, this work provides interesting insight into the effective combination of inherently incompatible components for the fabrication of composite materials from an ultrasonic standpoint.\n               "}
{"full_text": "Shape-selective synthesis of ceria nanoparticles is of scientific and technological importance. Unique shape, size and surface defect-dependent ceria properties play a key role for various promising industrial applications including catalysis, electrochemistry, biology and optics [1\u20133]. Several strategies, such as hydrothermal [4\u20138], microwave-assisted hydrothermal [9], solvothermal [10] and micro-emulsion methods have been developed to synthesize CeO2 materials with well-defined cube, rod, polyhedron, nano-sheet morphologies [11]. Among all these strategies, hydrothermal synthesis is the most popular and convenient technique for preparing nanostructured CeO2 through varying precursor concentration, ratio of template agents, pH, hydrothermal temperature and aging time [8,12\u201314].In this context, cubic CeO2 materials often display remarkable structure-catalysis properties due to tunable oxygen vacancy (V\u00f6) on (100) facets. Up to date, many reports so far have also synthesized cubic CeO2 nanoparticles using this strategy [4\u20137,15,16]. However, high concentration of alkalis (6\u00a0mol/L) and prolonged reaction time (24\u201348\u00a0h) make this strategy very environmental and economical unfriendly. In addition, the size of resultant ceria materials often range from 20 to 60\u00a0nm, which show limited amounts of surface defects thus hindering further improvement of catalytic properties. Therefore, manufacture of finely dispersed ceria clusters with high surface defective contents are needed for advanced catalytic processes.Catalytic dehydrogenation of bio-derived oxygenates has been widely known as one of the key reactions for sustainable production of fuels and chemicals [17\u201319]. Catalytic activation of C\u2013H bond is known as the key reaction step for a variety of catalytic applications such as dehydrogenation, hydrogenolysis, dehydration and amination [20,21]. In particular, catalytic C\u2013H bond cleavage of bio-polyols leads to formation of various value-added carboxylic acids, which are essential building blocks for bio-degradable plastics and a variety of multifunctional polymers [22\u201325]. Typically, TOF values measured for conversion of bio-polyols such as glycerol, xylitol and sorbitol are in the range of 50\u20131500\u00a0h\u22121\u00a0at 160\u2013200\u00a0\u00b0C over monometallic Pt, Pd, Ru, Co, Ni, Cu and bimetallic NiCo, CuPd, PtCo, PtSn, AuPt catalysts [19,20,22,26\u201329]. As it is generally accepted that the cleavage of C\u2013H is a metal-activated reaction, considerable research efforts have been primarily focused on metal composition and morphology of metal particles for tunable activity and selectivity [15,22,27,30]. However, the morphological and electronic features (e.g., shape, size and surface defects) of heterogeneous supports, are yet to be detailed investigated for dehydrogenation of bio-polyols in aqueous phase. According to current literatures, CeO2 (100) and (110) facets show superior performances for facile C\u2013H activation of bio-polyols compared with other facets [31]. Large particle size (>20\u00a0nm) of ceria with low content of V\u00f6 are, however, major bottlenecking issues prevention further catalyst development for biomass conversion. The critical role of small size reflects on exposed surface of ceria materials. As already mentioned, previous studies have been primarily focused on synthesizing ceria size of >20\u00a0nm. Controllable fabrication of small ceria still remains a grand challenge in this area. Therefore, developing reliable synthetic approaches for small sized ceria materials is important for rational design of highly active catalysts for dehydrogenation as well as other energy applications.In the field of renewable H2, dehydrogenation of glycerol and polyols to green H2 co-producing valuable carboxylic acids, is known as an emerging technology for future bio-refineries (Scheme 1\n). Herein, we report a facile \u201clactic acid (LA) assisted hydrothermal method\u201d for synthesizing cubic CeO2 of 6\u00a0nm in size (CeO2-LA) with enhanced metal-support catalysis for dehydrogenation of glycerol and other bio-polyols. The structure and surface physicochemical properties analysis based on XRD, TEM, BET, Raman and XPS characterizations demonstrated that, the formation of finely sized cubic CeO2 leads to enhanced surface area and abundant surface V\u00f6 defects. The growing mechanism of finely dispersed clusters is ascribed to the synergism that, hydroxyl and carboxylic groups in LA molecule induce morphological confinement. The proposed PtCo/CeO2-LA catalysts with abundant surface V\u00f6 defects exhibit remarkable catalytic activity and good selectivity, leading a record high TOF value of 29,241\u00a0h\u22121\u00a0at 200\u00a0\u00b0C in dehydrogenation of bio-derived polyols. The influence of shape-, size-, V\u00f6-dependent catalytic properties on C\u2013H bond cleavage has been studied to reveal the underlying reasons for such performance enhancement. In addition, the investigated nano ceria materials display remarkable durability in hydrothermal conditions. The proposed synthetic strategy to tailor the V\u00f6 defect content can be potentially applied in rational design of other metal oxide catalysts for energy and environmental applications.Chloroplatinic acid hexahydrate (H2PtCl6\u00b76H2O), cobalt nitrate hexahydrate (Co(NO3)2\u00b76H2O), cerium nitrate hexahydrate (Ce(NO3)3\u00b76H2O), NaOH, Na2CO3, lactic acid, isopropanol, propanoic acid, glyceric acid, glycolic acid, acetic acid, formic acid, propylene glycol, ethylene glycol, ethyl alcohol, methyl alcohol, glycerol, xylitol, sorbitol, arabitol and mannitol were purchased from Sinopharm Chemical Reagent Co., Ltd.\nPtCo/CeO\n\n2\n\n-LA were prepared by LA-assisted hydrothermal method. A typical procedure is presented in Fig.\u00a0S1. Solution A (3.03\u00a0g of Ce(NO3)3\u00b76H2O dissolved in 25\u00a0mL deionized (DI) water) and solution B (0.8\u00a0mol/L NaOH and 0.25\u00a0mol/L Na2CO3) were simultaneously added dropwise to 250\u00a0mL beaker with 50\u00a0mL DI water, under continuous stirring at ambient temperature. The pH value of slurry in beaker was kept at 10.0\u201310.5 throughout the synthetic process. After stirring for 30\u00a0min at ambient temperature, solution C (0.06\u00a0g\u00a0H2PtCl6\u00b76H2O and 0.03\u00a0g Co(NO3)2\u00b76H2O dissolved in 10\u00a0mL DI water) and solution B were dripped into the beaker. After stirring another 30\u00a0min, the slurry was reduced with 0.1\u00a0g NaBH4 dissolved in 50\u00a0mL DI water (slowly added dropwise). Then the slurry was transferred to a 100\u00a0mL autoclave (2\u00a0g LA and 0.86\u00a0g NaOH were pre-added) and crystallized for scheduled time (1\u00a0h, 2\u00a0h, 3\u00a0h, 6\u00a0h, 10\u00a0h, 18\u00a0h and 24\u00a0h) at 200\u00a0\u00b0C. Furthermore, replacing to LA, isopropanol or propanoic acid were also selected as assistant to study growth kinetic of CeO2-LA. Finally, the slurry was filtered and washed three times with DI water. The solid sample was dried under 70\u00a0\u00b0C to obtained the final catalysts. The Pt and Co loadings of 1.36\u00a0wt% and 0.47\u00a0wt%, respectively, were determined by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES).\nPtCo/CeO\n\n2\n were synthesized by deposition\u2013precipitation method. Cubic CeO2 materials were first prepared by conventional hydrothermal method [32]. In a typical procedure, 1.74\u00a0g of Ce(NO3)3\u00b76H2O and 19.2\u00a0g of NaOH were dissolved in 80\u00a0mL DI water with stirring at ambient temperature. The resulting slurry was transferred to a 100\u00a0mL autoclave and crystallized for 24\u00a0h at 180\u00a0\u00b0C. Finally, the slurry was filtered, and washed three to five times with DI water. The solid sample was dried under 70\u00a0\u00b0C to obtained the final cubic CeO2. Supported PtCo/CeO2 catalyst was prepared by deposition\u2013precipitation synthesis. In a typical procedure, solution A (0.06\u00a0g of H2PtCl6\u00b76H2O and 0.03\u00a0g of Co(NO3)2\u00b76H2O dissolved in 50\u00a0mL DI water) and solution B (0.25\u00a0mol/L Na2CO3 aqueous solution) were simultaneously added dropwise to 500\u00a0mL beaker with 100\u00a0mL DI water and 1\u00a0g powder of cubic CeO2 under continuous stirring at ambient temperature. The pH value of slurry in beaker was kept at \n\n>\n\n 9 throughout the synthetic process. After stirring for 30\u00a0min at ambient temperature, the slurry was reduced with 0.1\u00a0g NaBH4 dissolved in 50\u00a0mL DI water (slowly added dropwise). After stirring another 12\u00a0h under ambient temperature, the slurry was filtered and washed three times with DI water. The solid sample was dried under 70\u00a0\u00b0C to obtained the final catalysts. The Pt and Co loadings of 1.30\u00a0wt% and 0.50\u00a0wt%, respectively, were also determined by ICP-OES.The morphologies and crystal structure of the PtCo/CeO2 and PtCo/CeO2-LA catalysts were measured by transmission electron microscope (TEM, JEOL JSM-2100F), high resolution TEM (HR-TEM) and X-ray diffraction (XRD, X'pert PRO MPD diffractometer instrument using Cu-K\u03b1 radiation with a scanning angle (2\u03b8) of 10\u00b0 - 80\u00b0, operated at 40\u00a0KV and 40\u00a0mA). The BET surface areas, pore volume and pore size were calculated according to N2 adsorption isotherms. Composition and valence states of surface metals were collected by X-ray photoelectron spectroscopy (XPS) measurements. Surface oxygen vacancy defect was measured by Raman spectra (LabRAM HR Evolution (HORIBA JobinYvon)) with the 514\u00a0nm laser lines. The Pt and Co loadings were determined by ICP-OES (VARIAN 720-ES, America Varian technologies).Evaluation of catalysts performance was carried out in a 30\u00a0mL autoclave. In a typical experiment, 0.05\u00a0g catalyst, 0.6\u00a0g NaOH and 15\u00a0mL glycerol aqueous solution (1.0\u00a0mol/L) were charged into the Parr reactor, following sealed and purged thrice with pure N2 (>1\u00a0MPa). Then, the reactor was heated to 200\u00a0\u00b0C and held at continue stirring in 1000\u00a0rpm under 1\u00a0MPa N2 pressure (N2 was used as an inert gas which does not affect experimental results. See Table S7). It is also important to mention that, N2 pressure was charged only for lab safety purpose. Thus 1\u00a0MPa was chosen for the benefit of conducting experiments. The activity of used catalysts was measured by conducting experiments with conversion <25%, roughly 1.0\u00a0h of batch time using 0.01\u00a0g of solid catalysts.After a designed reaction time, the liquid phase products were collected for quantitative analysis using a Shinadze HPLC LC-20AT system equipped with Phenomenex chromatographic column (Rezex ROA-Organic Acid H+ (8%), 300 \u00d7 7.7\u00a0mm) and refractive index (RID-10A) detectors. A typical operating condition were carried out at 60\u00a0\u00b0C with 0.005\u00a0mol/L H2SO4 aqueous solution as the mobile phase flowing at 0.8\u00a0mL/min. The definition of turn over frequency (TOF), conversion and selectivity is similar as previously described [22]. From the concentration values, conversion (X), selectivity (S) and turnover frequency (TOF based on Pt atom and calculated by initial reaction rate) were calculated as defined below. The yield was defined based on X and S (carbon based).\n\n\n\nConversion\n=\n\n\n\nCarbon\n\nm\no\nl\ne\ns\n,\nr\ne\na\nc\nt\na\nn\nt\n\n\ni\nn\ni\nt\ni\na\nl\n\n\n\u2212\n\nCarbon\n\nm\no\nl\ne\ns\n,\nr\ne\na\nc\nt\na\nn\nt\n\n\nf\ni\nn\na\nl\n\n\n\n\nCarbon\n\nm\no\nl\ne\ns\n,\nr\ne\na\nc\nt\na\nn\nt\n\n\ni\nn\ni\nt\ni\na\nl\n\n\n\n\u00d7\n100\n%\n\n\n\n\n\n\n\n\nSelectivity\n=\n\n\nCarbon\n\nm\no\nl\ne\ns\n,\n\np\nr\no\nd\nu\nc\nt\ns\n\n\nf\ni\nn\na\nl\n\n\n\n\nCarbon\n\nm\no\nl\ne\ns\n,\n\nr\ne\na\nc\nt\na\nn\nt\n\n\ni\nn\ni\nt\ni\na\nl\n\n\n\u2212\n\nCarbon\n\nm\no\nl\ne\ns\n,\n\nr\ne\na\nc\nt\na\nn\nt\n\n\nf\ni\nn\na\nl\n\n\n\n\n\u00d7\n100\n%\n\n\n\n\nTOF\u00a0=\u00a0\n\n\n\n\u25b5\n\nN\n\nr\ne\na\nc\nt\na\nn\nt\n\n\n\n\n\nN\n\nm\no\nl\ne\n,\n\nP\nt\n\n\n\u00d7\nR\ne\na\nc\nt\ni\no\nn\nt\ni\nm\ne\n\n\n\n (mole of loading Pt were determined by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES)).\n\n\n\nYield\n=\nConversion\n\u00d7\nSelectivity\n\n\n\n\n\nFig.\u00a01\n(a) and S1 depict the proposed two-step preparation of cubic CeO2-LA materials. Different from conventional one-step synthesis methods without using LA, Ce3+ was first precipitated at ambient conditions, the slurry of which was then further hydrothermally treated in a stainless steel reactor at 200\u00a0\u00b0C. In this process, LA was selected as soft template agent to control anisotropic growth of CeO2 nuclei. As shown in Fig.\u00a01(b), the TEM image presents that clearly cubic CeO2 crystal was formed via our LA-assisted hydrothermal method. Furthermore, the HRTEM image and fast Fourier transform (FFT) analysis of as-obtained cubic CeO2-LA (Fig.\u00a01(c)) show that two kinds of clear lattice fringe directions ascribed to (200) and (220) were observed, suggesting that the (100) planes are exposed on the surface of CeO2 nanoparticles [16,32]. The XRD pattern (Fig.\u00a01(d)) further reveals the pure cubic phase (fluorite structure, JCPDS 34-0394, space group Fm-3m) of the cubic CeO2-LA materials [16,32]. Notably, the particle size statistics (inset in Fig.\u00a01(b)) show that the average size of cubic CeO2-LA nanomaterial is 5.8\u00a0\u00b1\u00a00.9\u00a0nm, which is much smaller than cubic CeO2 synthesized by conventional hydrothermal methods (28.5\u00a0\u00b1\u00a024.2\u00a0nm, Fig.\u00a0S2). Obviously, in XRD pattern, the broadening of reflections of cubic CeO2-LA sample implied the significantly smaller cluster sizes (detailed calculation presented in Table S1) in comparison with conventional approaches (Table 1\n), which agrees well with the results of the particle size statistics based on TEM image. Incidentally, its surface areas is also increased significantly (see Table 1).The promising results on unique textural properties such as cubic crystal, small particle size and enhanced surface area, motivated us to inspect the plausible growth mechanism of CeO2-LA clusters. In general, kinetically controlled crystal growth in solution follows two major pathways: oriented attachment and Ostwald ripening (dissolution and precipitation) [33\u201335]. The conventional hydrothermal synthesis strategy of cubic CeO2 nanocrystals is known to follow oriented attachment pathways through initial nucleation of CeO2, subsequent effective collision and phase agglomeration of nuclei in concentrated OH\u2212 medium, which eventually results in large cluster size [35]. Interestingly, the collision and fusing of CeO2 nuclei is not significant in the presence of LA. As shown in Fig.\u00a02\n, instantaneous dissolution and recrystallization (Ostwald ripening) [36] of irregular nanoclusters to cubic CeO2 particles were observed in aqueous LA medium. Continued dissolution and recrystallization of ceria and cerium carbonates nuclei with extended hydrothermal time from 0 to 10\u00a0h, results in formulation of <6\u00a0nm sized cubic CeO2 particles. Notably, different from previous works [4\u20138,15], it is found that prolonged hydrothermal time from 3\u00a0h to 10\u00a0h only leads to slight enhancement of cluster size from 5.1\u00a0nm to 5.8\u00a0nm (Fig.\u00a03\n(a)), suggesting the collision and fusing of CeO2 nuclei are significantly inhibited in the proposed LA-assisted synthesis. Clearly, the Oswald ripening processes are manipulated by selectively promoting surface reconstruction rather than simple oriented attachment of cerium carbonates and CeO2 species.In particular, as shown in Figs. 2 and 3(b), when synthesis time is less than 3\u00a0h, the majority of the observed samples were nano-spheres, and composed of mixed cerium carbonates (PDF#44\u20130617) and CeO2 species. However, when the hydrothermal treatment time prolonged to 3\u00a0h and further to 6\u00a0h, the uniform cubic CeO2 nanomaterials were generated, with pure CeO2 diffraction peaks retained in XRD patterns. Li and coworkers [37] reported that the transformation of cerium carbonates phases go through the formation of amorphous cerium hydroxide and subsequent crystallization to form CeO2. Slowly dissolution of cerium carbonates in this work is beneficial for the recrystallization process leading to cubic CeO2 nanomaterials with smaller particle size.The existence of LA in aqueous medium clearly plays a key role in determining the fine dispersion of ceria clusters. To gain more insights for the formation mechanism, we performed further control experiments to understand the contribution of hydroxyl and carboxylic groups in LA molecule in formulating cubic ceria clusters during hydrothermal synthesis (see Fig.\u00a04\n). Therefore, we selected isopropanol (IP) and propionic acid (PA) as soft templates and synthesized CeO2 nanomaterials under similar conditions to establish the possible mechanism for the formulation of (200) and (220) facets.The critical role of hydroxyl and carboxylic groups on the formation of cubic ceria with <6\u00a0nm in size has not been well-documented in literature. According to TEM and XRD characterization in Fig.\u00a04, it is clear that, carboxylic groups quickly enable the formation of cerium carbonates phases. Under hydrothermal condition, the existing hydroxyl groups eventually facilitate the formation of cubic clusters with <6\u00a0nm in size.One should also be aware of the critical role of concentration of LA, IP and PA. Obviously, we have conducted synthesis under a condition where cubic morphologies are thermodynamically and kinetically favorable. Thus the size of ceria can be well controlled.As can be seen from Fig.\u00a04(a), mixed cubes, polyhedron and nano-spheres of CeO2 nanomaterials are coexisting when IP was used as the template. The average particle size of CeO2-IP (9.2\u00a0nm) is significantly larger than cubic CeO2-LA clusters (5.8\u00a0nm). Such observation indicates the CeO2-IP crystals in solution grow in an anisotropic fashion. When PA was used (Fig.\u00a04(b)), the obtained CeO2-PA clusters exhibit similar cubic shape and particle size to CeO2-LA clusters (Fig.\u00a04(c)). Taking into account the result of control experiments with time-dependent particle size data of CeO2-LA, it implies that the carboxyl functional groups play a critical role in manipulating morphological rearrangement through Ostwald ripening processes during hydrothermal treatment. Furthermore, the carboxyl functional groups are also found to stabilize the CeO2 nuclei, thus kinetics of self-assembly of cubic CeO2 clusters was significantly inhibited, by the repulsive forces through bonding the carboxyl functional groups in the soft template [38]. Teo and colleagues [35] also reported that, at a low pH, the growth of CeO2 particle favors the Ostwald ripening process through instantaneous dissolution and recrystallization of CeO2 nuclei, inducing the formation of small and individual particles. Hence, the dominant carboxyl mediated Ostwald ripening process results in the formation of cubic CeO2 with smaller size.Another interesting finding is that the pure CeO2 phase was generated within CeO2-IP clusters, while the as-prepared CeO2-PA clusters are mainly composed of both CeO2 and cerium carbonates phases (Fig.\u00a04(d)), suggesting that the hydroxyl groups control the reconstruction of Ce2(CO3)3 nuclei. This phenomenon was also found in time-dependent XRD patterns of CeO2-LA presented in Fig.\u00a03(b). As the hydrothermal treatment time prolong, the cubic CeO2 finally formed with dissolved cerium carbonates phases, indicating that the dissolution rate of cerium carbonates tuned by hydroxyl groups play a greatly role in determining the eventual morphology of cubic CeO2. Furthermore, the hydroxyl groups have reduction ability, which is favorable for the formation of surface V\u00f6 defects through reducing Ce4+ to Ce3+ [39,40]. Hence, detailed study of the surface features of CeO2-LA sample were conducted in the following sections.In summary, the experimental results shown above confirm that the formation of cubic shape and fine dispersion of CeO2-LA sample is induced by the strong synergism of hydroxyl and carboxyl functional groups during structural evolution (see Scheme 2\n). The carboxyl groups play a critical role in manipulating morphological rearrangement, thus the cubic shape and particle size can be well controlled. Hydroxyl groups determines the surface reduction rates of cerium carbonates phases. As a result, uniform CeO2-LA clusters with smaller particle size and abundant surface V\u00f6 defects can be achieved in the two-step preparation method proposed in this work.N2 adsorption/desorption results of cubic CeO2-LA and control samples are listed in Table 1. It is found that the surface areas and pore volume of CeO2-LA are 101.3\u00a0m2/g and 0.2\u00a0cm3/g, respectively, which are 5-fold and 10-fold higher compared with the control sample prepared in the absence of LA (20.6\u00a0m2/g and 0.02\u00a0cm3/g). The significantly enhanced surface area and pore volume often indicates more catalytically sites and defects are accessible to reactant molecules. Especially the V\u00f6 defects, created by removal of surface oxygen species simultaneously reduction of Ce4+ to Ce3+ species (Scheme 3\n), are proposed to reactive hot spot on CeO2 surface for many redox reactions catalyzed by metal oxides [3]. Therefore, we prepared CeO2-LA immobilized bimetallic PtCo catalysts and evaluated the performances for dehydrogenation of polyols as the model reaction. Prior to activity tests, metal valences and surface V\u00f6 defects of as-obtained bimetallic PtCo/CeO2-LA catalysts were characterized using XPS and Raman to gain insight into the unique surface and structural features.XPS spectra were collected to elucidate the chemical state of surface Ce and O species, which can reveal the content of surface defects [40]. The Ce 3d spectra (Fig.\u00a05\n(a)) of CeO2-LA was resolved into eight groups, suggesting the coexisting of Ce3+ and Ce4+ species in cerium oxides [6]. The peak positions and their attribution are summarized in Table S2 and Table S3. The 3d5/2 peak labeled V1 (885.1\u00a0eV) and 3d3/2 peak labeled U1 (903.3) are ascribed to Ce3+. In addition, the peaks labeled V0 (882.5\u00a0eV), V2 (888.4\u00a0eV), V3 (898.4\u00a0eV), U0 (900.8\u00a0eV), U2 (907.4\u00a0eV) and U3 (916.8\u00a0eV) are ascribed to Ce4+ species. To our best knowledge, it is generally accepted that the proportion of Ce3+ could reveal the concentration of V\u00f6 on CeO2 surface [3]. Following methodologies reported in the previous literature [6,32,40], the surface Ce3+ content was calculated based on the peaks areas of eight groups. The results show that the proportion of Ce3+ (23.7%) in cubic CeO2-LA sample is higher than CeO2 material synthesized in the absence of LA (19.0%), suggesting more abundant V\u00f6 defects on the surface of CeO2-LA.\nFig.\u00a05(b) presents the O 1s spectra of samples for further determining the chemical state of surface Ce species. The O 1s region is resolved into three groups, including the peak at 531.8\u00a0eV attributed to surface hydroxide and adsorbed H2O, the peaks at 530.2\u00a0eV and 529.2\u00a0eV assigned to Ce2O3 and CeO2, respectively (Table S4 and Table S5) [9,41]. These results further demonstrate the coexistence of Ce3+ and Ce4+ species in ceria samples. The peak area of lattice oxygen assigned to Ce2O3 in CeO2-LA sample is increased compared to the CeO2 material synthesized in the absence of LA, which means higher proportion of Ce3+ in CeO2-LA sample. This observation clearly shows good agreement for the Ce3+ and O 1s analysis, suggesting abundant V\u00f6 defects available on the surface of CeO2-LA.To further investigate the V\u00f6 defects content on CeO2-LA surface, the Raman spectra were collected and results are shown in Fig.\u00a05(c). The Raman spectra curves further confirmed different O storage capacity for cubic CeO2 synthesized by different method. The sharp characteristic peak at \u223c456\u00a0cm\u22121 is assigned to the Raman active optical-phonon F2g vibration mode of CeO2 with a fluorite structure [42]. In addition, the other low intense peak at \u223c573\u00a0cm\u22121 is detected on high energy sides of F2g peak, which is responsible for the V\u00f6 [43]. In general, the ratio of peak areas for the bands at \u223c573\u00a0cm\u22121 and \u223c456\u00a0cm\u22121 is defined as the concentration of V\u00f6 [44]. Detailed inspection on the two samples reveal that V\u00f6 concentration is significantly increased on CeO2-LA (27.2%), compared to the CeO2 material (3.3%) synthesized in the absence of LA. This result is good consistent with the observation from XPS analysis. Moreover, the relation between V\u00f6 concentration and hydrothermal reaction time was also investigated in detail (Figs. S3 and S4). It is clearly observed that V\u00f6 concentration increases with extended hydrothermal time, possibly due to the removal of more O species in the reductive atmosphere (-OH from LA and H2 released from NaBH4).More detailed catalyst characterization such as TEM-EDX and Aberration corrected STEM were performed to reveal the metal distribution of Pt and Co metals in the case of the PtCo/CeO2-LA and PtCo/CeO2 catalysts. As shown in Fig.\u00a06\n, the STEM-EDX mapping clearly shows that the Pt and Co elements have a good dispersion on PtCo/CeO2-LA catalyst compared with PtCo/CeO2 catalyst, which also confirms the presence of Pt and Co content in the catalysts. This information confirms the more introduction of V\u00f6 defects in the CeO2 surface leads to a better dispersion of the Pt and Co metals, which is consistent with other reports [45,46]. From aberration corrected STEM of PtCo/CeO2-LA catalysts in Fig.\u00a06(c) and Fig.\u00a0S6, it is shows that cubic morphology of CeO2 has been maintained very well after immobilizing Pt and Co content. Moreover, from the HR-TEM images in Fig.\u00a0S5, it is found that the lattice fringes corresponding to Pt (111) and Pt (200) planes are present along with Co (100) and Co (101) planes on PtCo/CeO2 and PtCo/CeO2-LA catalysts, indicating the presence of Pt\u2013Co surface on both catalysts.More importantly, detailed inspection on Fig.\u00a06 (c) further reveals the well dispersion of Pt species on cubic ceria. Taking into account the EDX mapping in Fig.\u00a06 (b), it is clear that, both Pt and Co elements are well dispersed on catalyst surface.Metal\u2013metal interfacial strong interaction are believed to be most essential to modulate cascade C\u2013H, C\u2013O and CO bond cleavage and formation in efficient conversion of biomass resources, due to tunable electronic structures at the interface [47]. Such, XPS of the PtCo/CeO2 and PtCo/CeO2-LA catalysts were further investigated to reveal the metal\u2013metal coupling effect. In Fig.\u00a07\n, the electron binding energy of Pt 4f5/2 and Pt 4f7/2 for both catalysts are 74.5\u00a0eV and 71.1\u00a0eV, respectively. The consistent electron binding energy of Pt 4f for the PtCo/CeO2 and PtCo/CeO2-LA catalysts indicates similar Pt\u2013Co and Pt-support coupling effect on both catalysts, which is further confirmed by the similar electronic structures of Co 2p for both samples. Hence, we can conclude that the more introduction of V\u00f6 defects in the CeO2 surface improve more sites for the growing of Pt, leading to a better dispersion, however, it cannot enhance Pt\u2013Co and Pt\u2013Ce electron coupling effect.As already mentioned in the above sections, catalytic activation of C\u2013H bond is known as the key reaction step for a variety of reactions such as dehydrogenation, hydrogenolysis, dehydration and amination [20,21]. In particular, catalytic C\u2013H bond cleavage of bio-polyols leads to formation of various value-added carboxylic acids, which are essential building blocks for bio-degradable plastics as well as other multifunctional polymers [22\u201325]. The catalytic performances of the PtCo/CeO2-LA and PtCo/CeO2 are investigated to understand possible size- and V\u00f6-dependent properties during conversion of bio-polyols to LA under N2 pressure. Glycerol, a byproduct of biodiesel production and chemical syntheses of perfumes, fragrances and pharmaceuticals from vegetable oil and animal fat, is selected as model compound. As shown in Fig.\u00a08\n(a) and Fig.\u00a09\n (detailed comparison see Table S6), the PtCo/CeO2-LA catalyst displays a record high activity (TOF: 29,241 h\u22121) for C\u2013H cleavage of glycerol (rate-determining step, see Scheme 4\n) [22]. In contrast, PtCo/CeO2 catalyst prepared in the absence of LA only shows a TOF value of 8079 h\u22121 under identical condition. Control experiments with uncalcined PtCo/CeO2 and PtCo/CeO2-LA catalysts were conducted to confirm that, negligible conversion of glycerol was found over those catalysts (X: 0.2\u20133%). Obviously, calcination is necessary to generate intrinsically active sites for glycerol conversion. In addition, the remaining LA on solid catalysts surface has been removed to exposed active sites. While the electron structure and surface morphology of Pt species show negligible differences for the two samples (Figs. 6 and 7), LA-assisted formation of abundant surface and V\u00f6 defects are the key contributing factor for such dramatic enhancement in activity.The plausible mechanism for activity enhancement for C\u2013H cleavage of glycerol (dehydrogenation) is discussed in this section. It is necessary to mention that, Pt-catalyzed dehydrogenation is the key step for glycerol conversion to LA [17,19,20,22]. As already shown in Scheme 4, one mole of glycerol should generate equivalent amount of H2, thus high glycerol conversion indicates significant generation of H2 in gaseous phase. Therefore, it is not surprising that H2 is the main gas phase product during experiments. Clearly, V\u00f6 defects are critical for dehydrogenation reactions. Among all critical parameters that have been extensively studied in literature, V\u00f6 is believed to be the key for surface redox properties of metal oxides [3,31,48\u201351]. This is because V\u00f6 defects can bind reactants and intermediates more strongly in their activation [3]. In this work, more V\u00f6 defects on the surface of PtCo/CeO2-LA catalysts can promote chemical adsorption of OH\u2212, which intrinsically facilitates glycerol dehydrogenation to form glyceraldehyde as key intermediate [22]. Hence, the record high activity was obtained over PtCo/CeO2-LA catalyst. We further investigated the effect of V\u00f6 concentration on glycerol conversion under N2 pressure. It is found in Fig.\u00a08(b) that the conversion of glycerol displays optimal values with V\u00f6 content, suggesting that the synergism between V\u00f6 and PtCo sites is critical for C\u2013H bond cleavage of glycerol. It is generally accepted that metal catalyzed C\u2013H bond cleavage is often restrained by poor activity for nucleophilic attacking activity of metal centers. The presence of V\u00f6 intrinsically facilitate electronic transfer from Ce3+ to metallic centers [22]. As a result, the \u03b1-bond activation mode is selectively promoted on the surface PtCo clusters. This result further confirms that adjusting chemical adsorption of reactants and OH\u2212 through tailoring V\u00f6 content is an effective approach for tunable dehydrogenation reactivity of bio-polyols.The catalytic reusability of the PtCo/CeO2-LA catalyst for glycerol were also tested at 1\u00a0MPa N2 pressure and 200\u00a0\u00b0C (Influence of pressure on conversion is shown in Table S7). As shown in Table 2\n, the TOF of PtCo/CeO2-LA catalyst show slight decreases under the recycle reaction. Thus, the PtCo/CeO2-LA catalyst is stable under the reaction conditions.It was previously found that, actually lowering the operating pressure favored conversion of glycerol to LA (Table S7). This is possibly because that, this reaction involves generation of one mole H2 and LA from one mole of glycerol molecules. Therefore, it is not surprising that, lowering N2 pressure slightly enhances glycerol conversion.TON values were also measured in Table 3\n. It is found that, TON value for PtCo/CeO2 catalyst was 20,197\u00a0at 200\u00a0\u00b0C. In contrast, PtCo/CeO2-LA catalyst shows a much higher TON value of 81,874 under the same condition. In addition, the regenerated PtCo/CeO2-LA catalyst also shows 79,899. It is clear that, the LA assisted PtCo/CeO2 catalyst displays superior activity and durability for glycerol conversion to LA and co-products.The catalytic performance of the PtCo/CeO2-LA catalyst for conversion of other sugar-derived polyols (xylitol, sorbitol, arabitol and mannitol) were also studied at 1\u00a0MPa N2 pressure and 220\u00a0\u00b0C. It is observed that synergistic C\u2013H bond cleavage was also achieved on the surface of the proposed catalyst, resulting in good selectivity of LA as the main product. As shown in Table 4\n, the selected long-chain bio-polyols are high efficient converted to C3 and C2 products with a high conversion (>90%), suggesting promising performances of PtCo/CeO2-LA catalyst for selective C\u2013C bond cleavage. And we observed that the C6 bio-polyols have a higher conversion but lower selectivity for LA, compared with C5 bio-polyols.As already mentioned in the above sections, More V\u00f6 defects on the surface of PtCo/CeO2-LA catalysts can intrinsically facilitate C\u2013H bond cleavage of bio-polyols. However, notably, the Ce3+, formed due to loss of considerable amounts of oxygen from its lattice, are readily oxidized to Ce4+ during storing at air atmosphere and quenching of partial V\u00f6 defects [52]. The loss of a large of number of V\u00f6 defects will inevitably reduce the catalytic activity. Hence, the performance of the PtCo/CeO2-LA catalyst with different storage time was evaluated, which can further understand the effect of surface V\u00f6 defects on the C\u2013H bond cleavage of bio-polyols. As shown in Fig.\u00a010\n, the fresh PtCo/CeO2-LA catalyst obviously displays higher activity with conversion of 39.9%, compared with aged PtCo/CeO2-LA catalyst in air at prolong time with conversion of 32.5% (1d), 30.2% (2d) and 25.4% (4d), while the selectivity of LA displays a negligible reduction. Pt 4f spectra of the fresh and aged PtCo/CeO2-LA catalyst in Fig.\u00a011\n(a) shows consistent electron binding energy, indicating similar Pt\u2013Co and Pt-support coupling effect on both simples. However, the weakened characteristic peak of Ce3+ on aged PtCo/CeO2-LA catalyst compared with fresh PtCo/CeO2-LA catalyst (Fig.\u00a011(b)) suggests the oxidation of Ce3+ to Ce4+ and quenching of V\u00f6 defects. Moreover, the Raman spectra also reveals the reduction of V\u00f6 defects (fresh: 12.1%, aged 1 day: 11.4%) on the surface of PtCo/CeO2-LA catalyst after an aging at air. This information indicates that the quenching of V\u00f6 defects decreased the activity of glycerol conversion, further confirming the great role of V\u00f6 defects in dehydrogenation of bio-polyols.Quenching of V\u00f6 defects on the surface of PtCo/CeO2-LA catalyst reduces its C\u2013H bond activity during conversion of bio-polyols, hence, it is urgent to develop a simple and efficient strategy for the regeneration of V\u00f6 defects. Several strategies, such as high temperature hydrogenation, ion doping, high-energy particle bombardment, atmosphere deoxygenation, mechanization and chemical reaction methods have been developed to form V\u00f6 defects in the past decades [53\u201357]. Among all these strategies, thermal reduction method is the most popular and convenient technique for forming V\u00f6 defects through varying reduction temperature and reducing atmosphere. However, this method usually requires high energy and seriously destroys the intrinsic structure of materials, hence it's not the best strategy for the regeneration of the PtCo/CeO2-LA catalyst.In this content, we developed a simple method for the regeneration of V\u00f6 defects on surface of the PtCo/CeO2-LA catalyst. Hernia radiation is known to create a large quantity of free radicals on the surface of metal oxides [58\u201360]. The formation of free radicals can decompose the unstable oxidative species on the surface of ceria to regenerate Ce3+. As shown in Fig.\u00a012\n(a), the PtCo/CeO2-LA catalyst was irradiated using hernia lamp, and it is found that the V\u00f6 defects content increased significantly with the extension of irradiation time (Fig.\u00a012(b)). Then, the regenerated catalysts with different irradiation time was evaluated at same reaction condition to reveal its reaction performance. In Fig.\u00a012(c), it is clearly that the activity of the PtCo/CeO2-LA catalyst was gradually recovered after regeneration of V\u00f6 defects with hernia lamp irradiation. When the irradiation time is 10\u00a0h (V\u00f6 defects content: 16.5%), the conversion of glycerol reached 42.3%, which even exceeds the conversion of the fresh PtCo/CeO2-LA catalyst (39.9%). However, the conversion began to decrease with the prolong irradiation time, because excessive V\u00f6 defects content (18.1%) will promote the enrichment of OH\u2212 on the surface of catalyst and inhibiting the transformation of glycerol [47].In summary, we have developed an environmental friendly and efficient strategy to synthesize cubic CeO2 supported PtCo bimetallic catalysts via LA-assisted hydrothermal method. The carboxyl and hydroxyl groups in LA synergistic promote phase transformation in hydrothermal process to the formation of cubic CeO2 with smaller particle size and enhanced surface area. Meanwhile, reductive atmosphere plays a key role in tailoring surface V\u00f6 content. Smaller particle size, enhanced surface area and abundant V\u00f6 defects contribute to improved dehydrogenation activity (TOF: 29,241\u00a0\u00b1\u00a01202 h\u22121) at 200\u00a0\u00b0C, during aqueous phase conversion of glycerol to LA. The PtCo/CeO2-LA catalyst also displays significant potential in conversion of long-chain bio-polyols. We also found that quenching of V\u00f6 defects on the surface of PtCo/CeO2-LA catalyst at air atmosphere reduces its C\u2013H bond activity during conversion of bio-polyols. In this content, a simple strategy to regenerate the quenched V\u00f6 defects and activity of the PtCo/CeO2-LA catalyst is also developed by irradiating of the deactivated catalyst hernia lamp. The methodology on shape-selective synthesis of metal oxides could be potentially utilized in other materials synthesis.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors want to thank financial supports National Natural Science Foundation of China (22078365, 21706290), Natural Science Foundation of Shandong Province (ZR2017MB004), Innovative Research Funding from Qingdao City, Shandong Province (17-1-1-80-jch), \u201cFundamental Research Funds for the Central Universities\u201d and \u201cthe Development Fund of State Key Laboratory of Heavy Oil Processing\u201d (17CX02017A, 20CX02204A) and Postgraduate Innovation Project (YCX2021057) from China University of Petroleum.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.gee.2022.08.008.", "descript": "\n                  Dehydrogenation is considered as one of the most important industrial applications for renewable energy. Cubic ceria-based catalysts are known to display promising dehydrogenation performances in this area. Large particle size (>20\u00a0nm) and less surface defects, however, hinder further application of ceria materials. Herein, an alternative strategy involving lactic acid (LA) assisted hydrothermal method was developed to synthesize active, selective and durable cubic ceria of <6\u00a0nm for dehydrogenation reactions. Detailed studies of growth mechanism revealed that, the carboxyl and hydroxyl groups in LA molecule synergistically manipulate the morphological evolution of ceria precursors. Carboxyl groups determine the cubic shape and particle size, while hydroxyl groups promote compositional transformation of ceria precursors into CeO2 phase. Moreover, enhanced oxygen vacancies (V\u00f6) on the surface of CeO2 were obtained owing to continuous removal of O species under reductive atmosphere. Cubic CeO2 catalysts synthesized by the LA-assisted method, immobilized with bimetallic PtCo clusters, exhibit a record high activity (TOF: 29,241 h\u22121) and V\u00f6-dependent synergism for dehydrogenation of bio-derived polyols at 200\u00a0\u00b0C. We also found that quenching of V\u00f6 defects at air atmosphere causes activity loss of PtCo/CeO2 catalysts. To regenerate V\u00f6 defects, a simple strategy was developed by irradiating deactivated catalyst using hernia lamp. The outcome of this work will provide new insights into manufacturing durable catalyst materials for aqueous phase dehydrogenation applications.\n               "}
{"full_text": "Data will be made available on request.As a prototypical model probe reaction, catalytic oxidation of carbon monoxide (CO) has attracted much research attention for many years [1]. With the development of proton exchange membrane fuel cell (PEMFC) technology, removal of CO from hydrogen raw materials by preferential oxidation has gradually become a research focus [2]. With expanding awareness of environmental conservation, the removal of CO from air has become a pertinent topic. CO could not only bind to human's hemoglobin, but also affect climate atmospheric chemistry and the ozone layer [3\u20135]. The catalytic oxidation of CO is one of the most effective strategies to mitigate the issues above. Typically, noble metal-based catalysts (Pt, Pd, Rh and Au etc.) are utilized for CO oxidation. However, large-scale applications of noble metal-based catalysts are restricted by their limited reserves and high cost, making it practically urgent to develop high-efficiency and low-cost catalysts for CO oxidation based on earth abundant elements [6].Recently, transition metal oxides (such as Co3O4, Fe2O3, Mn3O4, CuO and NiO etc.) have been employed as high active catalysts for CO oxidation [7,8]. Among the multifarious candidates, copper-based catalysts attracted much attention owing to their economic viability, environment-friendliness, and high catalytic activity [9]. For instance, CuO-CeO2 catalyst developed by Liu et al. could deliver a high CO conversion of 96% at 86\u00a0\u00b0C [10]. Regarding the progress achieved in copper-based catalysts and the features of active metal species, it has been established that only well dispersed copper species exhibited satisfactory catalytic activity, while that of bulk Cu species are negligible [11]. Consequently, it has aroused much attention to exploit highly dispersed copper-based catalysts to further improve its catalytic performance.In general, supports surfaces are utilized to load highly dispersed and active catalytic sites in a homogeneously distributed fashion, thereby maximizing utilization efficiency, catalyst stability, while minimizing aggregation of the active species and thus costs. In particular, metal-support interactions can modify the geometric and electronic structure of the interfacial active sites, which can further enhance catalytic activity and stability [12,13]. Recently, Papadopoulos et al. synthesized atomically dispersed copper-ceria catalysts using surfactant-assisted hydrothermal method and the catalytic activity of the catalyzers were greatly improved due to highly dispersed active centers and strong copper-ceria interaction [14].CuO/ZnO catalysts have been utilized in various thermochemical reactions, e.g., E.L. Reddy et al. reported that the surface area of Cu metal in catalyst significantly affects the catalytic activity of water-gas shift (WGS) reaction, when 30\u00a0wt% of CuO and 70\u00a0wt% of ZnO were utilized [15]. Yupeng Xie et al. found that CuO could be reduced to Cu (under reducing conditions) in CuO/ZnO catalysts in supercritical water gasification of lignin at 500 and 600\u00a0\u00b0C. The reduction may release active oxygen species to improve the decomposition and gasification of biomass. Additionally, metal Cu could also serve as the catalyst to accelerate the WGS reaction and carbon conversion [16]. C. Mateos-Pedrero et al. reported that in CuO/ZnO catalysts the copper dispersion and surface area increased with increasing the surface area of the ZnO support, resulting in better catalytic activity. In methanol steam reforming reaction, the selectivity increases with increasing the polarity ratio, which might be due to the presence of more selective Cu\u2013ZnO sites at the Cu\u2013ZnO polar interface of CuO/ZnO catalyst [17].Furthermore, the catalytic CO oxidation performance have been demonstrated to depend greatly on the sort of exposed crystal plane of oxide support [18,19]. Among the various metal oxides, ZnO support is more attractive owing to its low cost, nontoxicity, ease of preparation and its morphology can be well controlled to preferentially expose certain crystal facets. Therefore, ZnO has been utilized as supports in various catalytic reactions [20]. For instance, the selective hydrogenation of CO2 to CH3OH production was examined using Cu loaded on ZnO with various morphologies [19]. It was found that the platelike ZnO support based copper catalyst exhibited the highest selectivity towards methanol synthesis. To the best of our knowledge, though the morphology effect of ZnO carrier towards many reactions have been investigated [18,19], the influence of morphology engineering of CuO loaded ZnO supports on its catalytic performance for CO oxidation has been seldom reported.In this work, ZnO with different morphologies including nanorods (NRs) with varied aspect ratios and nanodisks (NDs) were synthesized, which served as support for CuO nanoparticles loading via deposition-precipitation method. The morphology engineering of CuO/ZnO catalysts for CO oxidation was systemically investigated. The ZnO support with different morphologies and the as-prepared catalysts derived from them were characterized by nitrogen adsorption, transmission electron microscopy (TEM), scanning electron microscopy (SEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopic (XPS) and inductively coupled plasma mass spectrometry (ICP-MS). The catalytic performance of the CuO/ZnO catalysts for CO oxidation was evaluated via fixed bed reactor. And a mechanism was proposed to highlight the importance of morphology of ZnO in CuO/ZnO catalysts for CO oxidation.ZnO NRs were prepared according to the reported procedure after modification [21,22]. \u223c 10.5\u00a0g of Zn(NO3)2\u00b76H2O and 28\u00a0g of NaOH were dissolved into 420\u00a0mL and 325\u00a0mL ethanol, respectively. And then the two solutions were mixed, followed by adding 231\u00a0mL ethylenediamine (EDA); and the mixture was kept at 120\u00a0\u00b0C for 12\u00a0h. Afterwards, the products were centrifugated and washed using deionized water and ethyl alcohol several times and calcined at 300\u00a0\u00b0C for 3\u00a0h in air flow. The ZnO NRs with different aspect ratios were also prepared with similar procedure except the reactions were carried out at 120\u00a0\u00b0C for 24\u00a0h. The products were named as ZnO-NR1 and ZnO-NR2, respectively.The ZnO NDs were prepared according to the reports with some modifications [19,23]. \u223c 12\u00a0g of Zn(OAc)2\u00b72H2O and 7.68\u00a0g of hexamine were continuously dissolved into 50\u00a0mL deionized water to obtain solution, respectively. And then the two solutions were mixed and reacted at 100\u00a0\u00b0C for 2\u00a0h. Afterwards, the as-prepared products were centrifugated and washed with deionized water and ethyl alcohol for at least 3 times and calcined at 300\u00a0\u00b0C for 3\u00a0h in air flow. And the products were denoted as ZnO-NDs.In order to load CuO onto the ZnO support via deposition precipitation method [24,25], \u223c1\u00a0g of ZnO (including commercial Zinc oxide (ZnO-C), ZnO-NR1, ZnO-NR2 and ZnO-NDs) were dispersed into 40\u00a0mL deionized water under ultrasonic vibration condition. Then 10\u00a0mL deionized water containing 0.16\u00a0g of Cu(OAc)2\u00b7H2O were put into the solution containing ZnO and the pH values of the mixture were regulated to 8\u20138.5 using 0.1\u00a0mol/L K2CO3. And then the above-mentioned mixture was transferred from beaker to a Teflon-lined stainless-steel autoclave and heated at 80\u00a0\u00b0C for 12\u00a0h. Afterwards, the obtained products were centrifugated and washed with deionized water and ethyl alcohol for at least 3 times, which were transferred to an oven and kept at 80\u00a0\u00b0C for 12\u00a0h. The as-prepared catalysts are denoted as CuO/ZnO-C, CuO/ZnO-NR1, CuO/ZnO-NR2 and CuO/ZnO-NDs, respectively.TEM and SEM images were collected using JEM-2100F field emission microscopy and Nova Nano SEM 450, respectively. XRD patterns were measured using a Lab XRD-7000\u00a0s powder X-ray diffractometer with Cu K\u03b1 (\u03bb\u00a0=\u00a00.154\u00a0nm) radiation. XPS measurements were carried out on an ESCALAB250Xi instrument. The charging effect was corrected by adjusting the binding energy of C1s to 284.6\u00a0eV. For XPS surface elemental compositional analysis/atomic fractions calculations, the relative ratio of atoms was calculated according to Eq. (1):\n\n(1)\n\n\n\nn\ni\n\n\nn\nj\n\n\n=\n\n\nI\ni\n\n\nI\nj\n\n\n\u00d7\n\n\n\u03c3\nj\n\n\n\u03c3\ni\n\n\n\u00d7\n\n\n\nE\n\nk\nj\n\n\n0.5\n\n\n\nE\n\nk\ni\n\n\n0.5\n\n\n\n\n(n: Number of atoms, I: Intensity (Integral peak area), E\nk: Kinetic energy, \u03c3: Photoionization cross section) [26\u201328]. N2 adsorption-desorption curves were collected using Autosorb-iQ-C instrument (Quantachrome) at 77\u00a0K. To determine the loading amount of Cu species on ZnO supports, ICP-MS were measured using a 7900 (Agilent) instrument.The CO catalytic oxidation performance was evaluated in a fixed-bed reactor using \u223c60\u00a0mg of screened catalysts (40\u201360 mesh) in a gas mixture (0.6% CO, 16.8% O2 and 82.6% N2) at a flow rate of 50\u00a0mL\u00a0min\u22121. Before each measurement, the CuO/ZnO catalysts were pretreated under air at 250\u00a0\u00b0C for 1\u00a0h. The effluent from the reactor was analyzed by an online gas chromatograph (Techcomp-GC7900) equipped with a flame ionization detector with Ni reformer. For kinetic studies, the reaction rates were measured at 398.15\u00a0K, while the conversion of CO was kept below or within 20% to assure differential reactor conditions. The conversion of CO was obtained according to the Eq. (2):\n\n(2)\n\nCO\n\nconversion\n=\n\n\nA\n\n\nCO\n2\n\n\n\n\n\nA\n\n\nCO\n2\n\n\n+\nA\n\nCO\n\n\n\n\n\n\nwhere A(CO2) and A(CO) are the concentrations of carbon dioxide and carbon monoxide, respectively.The morphological characteristics of the ZnO supports were examined using SEM. As illustrated in Fig. S1, the ZnO-NR1 and ZnO-NR2 show rod-like morphology, with average length of 160 and 750\u00a0nm, respectively. And the rod morphology can be observed more clearly from the TEM images (Fig. S2, Fig. S5(b)). ZnO-NDs exhibit regular hexagonal disk-like morphology with a diameter of 1\u20132\u00a0\u03bcm and a thickness of 0.5\u20131\u00a0\u03bcm. In comparison, the commercial ZnO support exhibits plate/flake morphology with rough surface and the thickness of the flake is 30\u201360\u00a0nm. Further loading copper species does not alter the morphology of the ZnO supports (Fig. 1\n).The presence of copper species can be inferred from the elemental mapping images of CuO/ZnO-NDs, the copper species are uniformly distributed on ZnO nanodisks (002) polar facet (Fig. 2\n). Moreover, the copper species in the as-prepared catalysts were characterized using ICP-MS measurements, and the copper contents are found within the range of 4.2\u20135.2\u00a0wt% (Table S1), which agrees well with the nominal copper contents during catalysts preparation.Fig. S3 illustrated the XRD patterns of the powder ZnO supports before loading CuO nanoparticles. All the diffraction characteristic peaks could be attributed to the wurtzite ZnO, and the peaks at 31.77\u00b0, 34.42\u00b0, 36.25\u00b0, 47.54\u00b0, 56.60\u00b0, 62.86\u00b0, 66.38\u00b0, 67.96\u00b0 and 69.10\u00b0, respectively correspond to the (100), (002), (101), (102), (110), (103), (200), (112) and (201) planes, respectively of ZnO (JCPDS: no. 36\u20131451). It is obvious that the ZnO supports exhibit different intensity of (002) plane owing to the different morphology. As inferred from XRD patterns, the intensity ratio of (002) to (100) plane (I(002)/I(100)) for ZnO-NR2 is 0.66 due to its highest length to diameter ratio in the samples with nanorod morphology and it possesses the most exposed (100) plane. With shortening of the nanorod length, the intensity of (002) plane increased gradually, and the I(002)/I(100) for ZnO-NR1 is 0.810, which is even larger than that of commercial ZnO carrier. ZnO-NDs exhibit the largest I(002)/I(100) value of 1.186, due to the most exposed (002) planes induced by the nanodisk morphology. After loading copper species onto the powder ZnO supports and catalyst palletization, the XRD pattern of CuO/ZnO still shows the typical peaks of ZnO as well as a rather weak peak for CuO(111) [JCPDS: 45\u20130937] at 38.73o, showing well dispersed small-sized CuO (Fig. 3\n), which is well in agreement with the TEM results.Fig. S4(a, b) shows the representative TEM image of the as-prepared CuO/ZnO catalysts, showing CuO nanoparticles dispersed onto ZnO commercial support as reference. As can be seen in the particle sizes derived from the TEM image (Fig. S4(c)), the mean size of CuO is \u223c3.1\u00a0\u00b1\u00a00.9\u00a0nm.N2 adsorption-desorption isotherms were collected to investigate the specific surface areas of the ZnO supports and CuO/ZnO catalysts. As demonstrated in Fig. 4(a), all the ZnO samples with different morphologies exhibit similar adsorption isotherms with negligible N2 uptake at low relative pressure and a rapid increase above relative pressure of 0.9, suggesting that the pores are mainly accumulated pores from the nanoparticles of ZnO. The Brunauer-Emmett-Teller specific surface areas are 50, 35, 3.7 and 43 m2g\u22121 for ZnO-NR1, ZnO-NR2, ZnO-NDs and commercial ZnO, respectively. It is demonstrated that the ZnO-NR1 exhibits the largest specific surface area due to its small size, and ZnO-NDs exhibit the smallest specific surface areas owing to their relatively big size. The specific surface area of the CuO/ZnO catalysts likewise exhibits the same trend as that of ZnO, and the specific surface area decreases a little after CuO loading as shown in Fig. 4(b).To investigate the catalytic performance of the CuO/ZnO catalysts, CO oxidation was measured in the normal mixture gas, and the effect of temperature on the CO conversion is illustrated in Fig. 5(a). All the prepared CuO/ZnO catalysts exhibit higher catalytic activity than CuO/ZnO-C in terms of much lower conversion temperature, and the catalytic activity exhibits a trend of CuO/ZnO-NDs\u00a0>\u00a0CuO/ZnO-NR1\u00a0>\u00a0CuO/ZnO-NR2\u00a0>\u00a0CuO/ZnO-C. For comparison, the values for 50% CO conversion (T50) are tabulated in Table S1, which reflect the same trend as that above mentioned. From another viewpoint, at a certain temperature of 150\u00a0\u00b0C, the CO conversion for CuO/ZnO-NDs is 52%, while the values are 33%, 25% and 25% for CuO/ZnO-NR1, CuO/ZnO-NR2 and CuO/ZnO-C, respectively. As can be analyzed, though the specific surface area of CuO/ZnO-NDs is much lower than that of CuO/ZnO-C derived from commercial ZnO support, its catalytic activity is more than 2 folds higher to that for CuO/ZnO-C at 150\u00a0\u00b0C. The specific reaction rates for all catalysts are provided in Table S1, which also reflect the same trend. The better catalytic performance of CuO/ZnO-NDs could partially be ascribed to its higher fraction of (002) polar plane (as inferred from XRD patterns and ZnO disk morphology from SEM and TEM images). And the ICP and XPS measurements (discussed below) show that ZnO-NDs attract and retain greater copper contents or CuO phase on the surface. From CuO/ZnO-NR1 to CuO/ZnO-NR2, both the specific surface area and (002) polar plane fractions lowered with increasing the length of ZnO nanorods, so that decreased catalytic performance was observed. In contrast, ZnO-NDs have the highest (002) plane fraction and thus generally show higher catalytic CO oxidation activity.Besides the catalytic activity, the stability of a catalyst is also an important parameter that determines its potential reuse and operational/catalyst replacement costs. As shown in Fig. 5(b), after six consecutive catalytic reaction cycles, the conversion of CO over CuO/ZnO-NDs remained \u223c50% at 170\u00a0\u00b0C without apparent reduction of activity performance, implying its good cyclic stability.XPS measurements were carried out prior to reaction tests and after CO oxidation utilizing spent catalysts to investigate the relative surface composition, oxidation states for constituent elements and interaction between copper species and ZnO supports for the CuO/ZnO catalysts. Fig. 6(a) and Fig. 6(g) show the Cu2p spectra for ZnO NR2 and NDs presenting pre-mortem catalyst surface analysis before CO oxidation reaction, respectively. The Cu 2p3/2 peak is centered at 933.6 and 933.7\u00a0eV (with Spin-orbit splitting (S.O.\u00b7S) value at 19.9\u00a0eV) for ZnO NR2 and NDs, respectively, which is attributed to CuO/Cu2+ species. The related prominent signature CuO satellite peaks can also be seen [28]. Fig. 6(d) and Fig. 6(j) presents the Cu2p spectra for the same ZnO NR2 and NDs revealing post-mortem analysis of the catalyst surface after reaction tests, respectively. The Cu 2p3/2 (CuO) peak is centered at 933.13 and 933.63\u00a0eV for ZnO NR2 and NDs, respectively. Fig. 6(b) and Fig. 6(h) show the Zn2p spectra for ZnO NR2 and NDs before CO oxidation reaction, respectively. The corresponding binding energies of Zn 2p3/2 are centered at 1021.4 and 1021.7\u00a0eV (with S.O.\u00b7S value of 23.0\u00a0eV), which is attributed to Zn2+ of stochiometric ZnO. Fig. 6(e) and Fig. 6(k) show the Zn2p spectra for ZnO NR2 and NDs after reaction tests, respectively. The corresponding binding energies of Zn 2p3/2 (of ZnO) are centered at 1021.63 and 1022.1\u00a0eV, respectively [29]. Fig. 6(c) and 6(i) show the O1s spectra of the ZnO NR2 and NDs prior to the reaction tests, respectively. The O 1\u00a0s spectra can be deconvoluted into three peaks centered at ca. 530.0, 531.5 and 533.3\u00a0eV, corresponding to the lattice oxygen of ZnO/CuO, and adsorbed hydroxyl (OH) /surface oxygen for CuO/ZnO NR2 and CuO/ZnO-NDs catalysts, respectively. Fig. 6(f) and 6(l) show the O1s spectra of the ZnO NR2 and NDs after CO oxidation reaction, respectively. The above-mentioned O 1\u00a0s spectra corresponding three peaks are centered at ca. 530.25, 530.15, 531.5, 531.45, 532.5 and 533.45\u00a0eV for NR2 and NDs, respectively [30].For the XPS spectra, prior to CO oxidation, compared to CuO/ZnO-NR2, the binding energies of CuO/ZnO-NDs for Zn2p and Cu2p shift positively, demonstrating a stronger interaction between CuO and ZnO-NDs. It is suggested that a p-n heterojunction may be formed at the interface of CuO and ZnO, resulting in the shift of binding energy as observed from XPS measurement [31]. Herein, the polar plane of ZnO could induce stronger interaction between ZnO and CuO, and correspondingly the binding energy shifts positively.From XPS compositional analysis, we also calculated associated surface atomic fractions of both the catalysts prior to the reaction and after the CO oxidation tests. And from O1s peak areas analysis post deconvolution, we computed the ratio of the surface oxygen/lattice oxygen (Table S2). From the surface atomic fraction analysis of the corresponding constituent elements, compared with the amount of CuO/Cu2+ species on ZnO NR2, the contents of CuO/Cu2+ species is higher on ZnO-NDs catalysts surface before reaction tests, as can be seen in Cu2p spectra (Fig. 6(a, g)) for CuO-NDs is less noisy and indicated by Cu/O and Cu/Zn ratios (as well as ICP results). However, there is an observable reduction in the fractions of CuO/Cu2+ species after reaction tests for both catalysts, as demonstrated by Cu2p XPS peaks becoming weak in Fig. 6(d) and Fig. 6(j) and evident from variation in Cu/O and Cu/Zn ratios. Similarly for the samples before test, we also observed that the exposed Zn2+ of ZnO is higher on the surfaces of CuO/ZnO-NR2 than CuO/ZnO-NDs, probably due to ZnO surface coverage by higher surface Cu2+ species (as discussed above) and higher adsorbed active oxygen species (Fig. 6(c, i)) in CuO/ZnO-NDs catalysts and indicated by Zn/O ratios. The higher copper contents on ZnO-NDs are also verified by ICP results (Table S1), as well as qualitatively verified by strong signal of Cu2p XPS as discussed (Fig. 6(a, g)). Because the same nominal amount of copper precursor and precipitation protocol was strictly followed, the greater Cu content retention on ZnO disk polar facets (Fig. 2) may well be regarded as a function of the greater fractions of the polar surfaces of ZnO-NDs for which the copper interaction is stronger than non-polar facets of ZnO-NRs [19] which is an interesting new insight in this work. After the CO oxidation tests, XPS compositional analysis reveals that there is a reduction in the amount of Zn2+ (ZnO), as indicated by Zn/O and Zn/Cu ratios. For the samples before reaction, from O1s peak analysis, we see that the percentage of surface/adsorbed Oxygen is much higher on CuO/ZnO-NDs as can be seen by proportion of the deconvoluted peak areas, and which was verified by increase in corresponding peak area ratios of OH/O-(Zn, Cu) (Table S2).The abundance and availability of activated surface oxygen species on ZnO-NDs is an attributing factor for enhanced catalytic activity, in addition to the relatively higher Cu2+ species [23]. From O1s peak area and compositional analysis after the reaction tests, we observe an increase in surface oxygen species for CuO/ZnO-NR2 and CuO/ZnO-NDs catalysts. O1s peak area analysis reveals an enhancement in the percentage of surface/adsorbed oxygen after the tests, which is much higher for CuO/ZnO-NDs verified by the proportion of the deconvoluted peak areas in Fig. 6(f) and Fig. 6(l), and which was verified by increase in corresponding peak area ratios of OH/O-(Zn, Cu) (Table S2) for the two catalysts. Thus, Our XPS study clearly identifies CuO/Cu2+ species from Cu2p peak analysis. The Cu2+ species surface atomic fractions can be compared easily between the samples before the reaction separately and after the reaction separately, in turn, as the catalyst samples conditions for the two sets are same. Additionally, the weakening of Cu2p peak due to active oxygen on ZnO NDs is indicative of the greater active oxygen species available for catalyzing CO oxidation, as can simply be qualitatively inferred from the O1s peak analysis as well. Since, we have pretreated our catalysts under air before CO oxidation at 250\u00a0\u00b0C for 1\u00a0h, there is a single CuO phase to the exclusion of any other oxidation state present on the sample. Thus, Cu contents reflected in ICP can be considered for quantitative CuO analysis. Thus, ICP clearly identifies higher Cu contents on NDs polar facets than on NR2 nonpolar facets, indicating stronger interaction. In summary, the enhancement in adsorbed active surface oxygen species, together with higher CuO/Cu2+ species (and thus higher interfacial sites) for CuO/ZnO-NDs catalysts, results in corresponding increase in catalytic turnover.In addition to the XPS analysis after CO oxidation, spent catalysts were evaluated after CO oxidation reaction utilizing XRD, SEM and TEM to assess support morphology and ZnO, CuO stability. Fig. S5 shows the representative TEM images of the typical CuO hemispherical shaped nanoparticles anchored onto ZnO lattice. Fig. S5(a) highlights the typical situation for CuO/ZnO-ND in spent catalysts, while Fig. S5(b, c) shows the interface in our typical CuO/ZnO-NRs catalyst, before and after reaction tests respectively.Fig. S6 shows the XRD patterns of the spent catalysts after CO oxidation, which reveals that the ZnO and CuO peaks remain essentially unchanged, with CuO still existing as weak as before the tests, illustrating the absence of significant Cu growth. Fig. S7 shows the SEM images of the spent catalysts after CO oxidation tests, which revealed the final morphology of the zinc oxide support. No morphology change was observed on the ZnO-NDs support which exhibited the highest thermal/shape stability, followed by ZnO-NR2, ZnO-NR1 and ZnO-C. Fig. S8 shows the corresponding elemental mapping images of the spent CuO/ZnO-NDs catalysts, which exhibits the ZnO wurtzite crystal hexagonal morphology and presence of Zn, O and Cu elements after CO oxidation tests. The ZnO disk morphology is intact after the CO oxidation tests showing its thermal stability at high temperatures and the uniform distribution of copper is maintained over the entire nanodisk top of the polar (002) facet with only negligible growth. Additionally, Fig. S9 shows the elemental mapping images of CuO/ZnO-C catalysts identifying distribution of Zn, O and Cu elements before and after CO oxidation tests. Even after the reaction tests, the Cu species show uniform distribution analogous to that before the reaction, elucidating the absence of drastic growth.The catalytic activity of copper-based catalysts is always influenced by oxidation states of copper, generated on the surfaces when usually exposed to air or under O2-rich reaction conditions. In nanoparticle catalysts comprising three valence states of copper, Cu2O phase is reported as more active than either metallic Cu or CuO [32\u201334].Li et al. [20,23] investigated ZnO shape effects in ZnO disks and rods. Based on diffraction intensity ratio of (002) polar plane to (100) nonpolar plane (I(002)/I(100)), ZnO disks were found to have a higher value. From SEM imaging and shape geometrical interpretation, it was concluded that the proportion of polar planes (with preferably more oxygen vacancies) is higher for disks than rods. Such oxygen vacancies on polar planes are found to be more crucial for catalytic activity than surface area, which agree with our findings as well. Mclaren, et al. [35] also reported that hexagonal nanodisks (with higher fractions of the more active polar ZnO surfaces) were found to have more than five times higher activity in the photocatalytic decomposition of methylene blue than nanorods. Very recently, Lyu et al., [36] synthesized Cu\nx\nO/ZnO catalyst which agreed well with our observations that the active sites of the catalysts are exposed Cu\nx\nO particles and the heterojunction at the interface. It was shown that due to the lower work function of the ZnO, electrons transfer from ZnO to CuO, thus generating higher active Cu species. Such electron rich interface is beneficial for the activation and dissociation of oxygen molecules and hence CO oxidation reaction. As regards mechanism, Luo et al. [37] reported a surface activation process on an atomic scale for copper species during CO oxidation utilizing environmental TEM. Surface roughening and low-coordinated Cu atoms formation upon CO exposure, and quasi-crystalline CuO\nx\n phase formation upon O2 exposure were observed. CO molecules formed strong Cu-CO bonds with Cu atoms on the surface, and then intercalated into the lattice (2\u20133 atomic layers) with the aid of O2, which induced lattice expansion. It was concluded that, on active CuO\nx\n surface, the reaction proceeds as follows: first O2 molecules dissociate due to the undercoordinated Cu atoms, then this lattice O atom combines with adsorbed CO and CO2 is produced. Sarkodie et al. reported promotional effects of Cu\nx\nO on the activity of Cu/ZnO catalyst towards efficient CO oxidation [38]. The fraction of active CuO\nx\n phase directly correlated with the catalytic activity and more Cu+ species resulted in enhanced CO adsorption ability. CO is chemisorbed on Cu+ surface, forming Cu+-CO, and then CO drifted to the interface between copper oxide and zinc oxide, where CO combined with activated O atoms from oxygen vacancies in the support lattice and CO2 is formed.In our case, as stated, since all the catalysts were pretreated in air at 250\u00a0\u00b0C for 1\u00a0h before each test, Cu species should be in oxidation state of CuO. Indeed, XRD, XPS and TEM results clearly confirmed that it was in CuO state, hence CuO worked as active centers that catalyzed the CO oxidation. We observed that the Cu\nx\nO/ZnO derived from ZnO-NDs showed greater CuO contents for the same copper-deposition-precipitation route as confirmed by ICP and XPS. This highlights that NDs attracts greater CuO fractions on its polar facets and this correlates well with the enhanced catalytic activity. Additionally, more active adsorbed oxygen species on the catalyst surface was confirmed by XPS spectra, as compared to the other catalysts.Based on our own findings and the literature, a reaction mechanism of CO oxidation was proposed, as described in Fig. 7\n. CO2 is generated via two possible pathways. The first pathway involves CO molecules from gas phase bonding to the Cu atoms in CuO nanoparticle surface, forming Cu2+-CO complex. Then the O2 molecules from the gas phase dissociate near the lattice O vacancy sites on CuO. Finally, the adsorbed CO molecules combine with the lattice O from CuO, which is replenished by molecular O2, to form CO2. In the second reaction pathway, the reaction proceeds at the interfacial sites between CuO and ZnO. The CO molecules either adsorbed on top surfaces of CuO drift to the interface between CuO and ZnO or adsorbed on the CuO interfacial perimeter sites, which react with the lattice O atoms on ZnO surfaces, thus forming CO2 molecules. Consumption of lattice O atoms creates O vacancies, which are filled by dissociating O2 molecules from gas phase. Due to the strong interaction between ZnO polar plane and CuO, the adsorption and oxidation of CO may be energetically more favorable in CuO/ZnO catalysts derived from NDs, resulting in better performance compared to other catalysts. The gas phase CO may adsorb on CuO nanoparticles, while the lattice Oxygen may be provided by CuO surface sites or ZnO at the interface between CuO and ZnO.In conclusion, we prepared a series of CuO/ZnO catalysts via altering the morphology of ZnO support and their catalytic activity towards CO oxidation was investigated. The CuO/ZnO-NDs with the most exposed (002) polar plane fraction derived from ZnO nanodisk exhibits more than 2 folds higher catalytic activity at 150\u00a0\u00b0C compared with Cu/ZnO-C using commercial ZnO as carrier. The resulting rationally tailored CuO/ZnO catalysts with nanodisk morphology were able to achieve complete CO conversion at 218\u00a0\u00b0C, which was much lower than that of Cu/ZnO-C (246\u00a0\u00b0C). The catalytic activity results indicated that the polar plane of ZnO played an important role in boosting CO oxidation via greater CuO retention and more adsorbed active oxygen, despite having lower specific surface area. This strategy can be applied to tailor ZnO based catalysts for industrially attractive applications such as CO2 hydrogenation reactions or be extended to other complex oxide-based catalysts.Shengnan Lu: Investigation, Writing-review & Editing.Houhong Song: Investigation, Writing-review & Editing.Yonghou Xiao: Resources, Conceptualization, Supervision, Project administration, Writing -review & Editing.Kamran Qadir: Resources, Conceptualization, Writing-review & Editing.Yanqiang Li: Writing -review & Editing.Yushan Li: Writing-review & Editing, Validation, Analysis.Gaohong He: Writing-review & Editing, Supervision.There are no conflicts of interest to declare.This research was supported by the Project of Central Government for Local Science and Technology Development of China (2022JH6/100100050), National Natural Science Foundation of China (21776028) and Liaoning Key Laboratory of Chemical Additive Synthesis and Separation (ZJKF2001).\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.colcom.2023.100698.", "descript": "\n                  Catalytic oxidation of CO is an effective route to eliminate the pollution of automobile exhaust. Tuning the fraction of exposed polar and non-polar crystal facets in ZnO provides an effective strategy to tailor metal-oxide interfacial active sites. Herein, the morphology of ZnO support was tuned to alter the fraction of exposed (002) polar/(100), (101) non-polar facets. ZnO nanodisks and nanorods were prepared, followed by deposition of \u223c3.1\u00a0\u00b1\u00a00.9\u00a0nm CuO nanoparticles. The impact of the ZnO morphology upon CuO deposition and catalytic activity towards CO oxidation were investigated. The CuO/ZnO derived from ZnO nanodisks (NDs) with the most exposed (002) polar plane exhibited more than 2 folds higher catalytic activity at 150\u00a0\u00b0C than that over CuO/ZnO with commercial ZnO support (CuO/ZnO-C), highlighting the morphology influence of ZnO. For its enhanced catalytic activity and cyclic stability, the CuO/ZnO-NDs could be a promising catalyst for CO oxidation.\n               "}
{"full_text": "Data will be made available on request.Nowadays, a large number of environmental problems caused by excessive CO2 emissions from anthropogenic sources occur frequently, such as melting glaciers, rising sea levels, drastic climate changes, etc [1]. Therefore, it is essential to develop an effective approach to mitigate carbon emissions. Considering that more than 60 % of the current energy demand comes from fossil fuel-based power plants, capture and sequestration (CCS) as well as capture and utilization (CCU) become indispensable for the mitigation of carbon emissions in the short term [2]. Among the two alternatives, CCU is the most promising solution and can be closely integrated with current industrial processes such as power plants and cement manufacturing [3]. However, the energy consumption in the separation, enrichment and transport of CO2 significantly increases the total cost of the CCU process.Recently, integrated CO2 capture and utilization (ICCU) technology has been proposed, which reduces the cost of the overall process by eliminating CO2 transport and storage [4]. This technology uses dual function materials (DFMs) and is gaining increasing interest by combining CO2 adsorption and utilization in a single reaction unit [5\u20137]. ICCU is a process of two periods or steps in which the feeding is varied in each one. In the first period, the DFM is saturated with CO2 by feeding an exhaust gas stream with a diluted concentration of CO2 (4\u201315 %). Then, in the second period, a reducing agent is introduced and the conversion of the adsorbed CO2 and the regeneration of the adsorption sites take place.Depending on the reducing agent, the composition of the DFM and the operating conditions different process can take place: dry reforming of methane (ICCU-DRM) [8,9], reverse water gas shift (ICCU-RWGS) [10,11] or methanation (ICCU-methanation) [12\u201314]. In addition, ICCU technology can also be an effective solution for storing surplus electrical energy in chemicals. For that, the energy surplus from renewable sources would be used to produce H2 from the hydrolysis of H2O and then reacted with the captured CO2. Therefore, ICCU can address the problem of intrinsic intermittency of renewable sources [15]. Among the different options, the ICCU-methanation technology is the most promising and its operation would approach a carbon neutral cycle. The first work of ICCU-methanation was published in 2015 [16], and since then the publications on CO2 adsorption and CH4 hydrogenation cycles are growing exponentially [17].DFMs are a combination of a compound based on Na [18,19], Ca [16,20], Mg [13,21] or K [13,22], as adsorbent, and on Ru [18,23], Ni [22,24] or Rh [25] as catalytic phase. Both phases are usually supported on a large surface area support. Specifically, \u03b3-Al2O3 is proposed as the best support [26]. In a previous work, we studied the effect of nickel loading in DFMs based on Na or Ca [20]. The highest activity was presented by DFM based on Na as adsorbent with 10 % Ni. Next, the addition of different promoters was studied. We conclude that the yield of CO2 adsorption and hydrogenation to CH4 in DFMs Ni-Na/Al2O3 is favored by the presence of a promoter [24]. Specifically, the sample promoted with Ru presents the best results. The presence of Ru creates synergistic aspects with Ni. It improves the reducibility of nickel and increases the ability of DFM to chemisorb H2.In the development of adsorbent materials with high capacity, reversibility and stability, Al-Mamoori et al. [27] studied Na and K doping of Ca-based adsorbents. The authors concluded that the addition of K and Na improved the performance of CaO since these materials had high CO2 adsorption capacity, fast kinetics and good stability above of 300\u00a0\u00b0C. On the other hand, Lee et al. [28] performed a comparative study of the adsorption and regeneration kinetics of Na2CO3 doped and conventional CaO adsorbents. In this study, the authors concluded that the addition of sodium carbonate to the calcium adsorbent can improve the cyclic stability of CO2 adsorption with fast kinetics. Recently, we have studied, for the first time in the scientific literature, the joint presence of Na and Ca in the same DFM [29]. DFMs are based on 4 % Ru as catalytic phase and 16 % adsorbent phase with different Na2CO3/CaO ratios. It has been possible to modulate the basicity and improve the Ru dispersion by varying the ratio and improve the CH4 production.In this work, the possibility of improving the activity of Ni-based DFMs is analyzed. For that, DFM 10Ni-16Na/Al2O3 is selected as reference. In the first place, the influence on the physicochemical properties and on the activity in the cyclic process of the joint presence of Na and Ca is analyzed. It is analyzed if it is possible to modulate the basicity and promote the CH4 production as in the DFMs based on Ru. Secondly, the influence of the incorporation of a small amount (1 %) of Ru to the reference DFM is studied. Note that the price per unit of mass of Ru is around 935 times higher compared to Ni (September 2022). The synergies that take place between both metals are studied. Third, the joint presence of Na and Ca and the simultaneous Ru incorporation is analyzed. For this, all the DFMs are extensively characterized and their activity is analyzed in CO2 adsorption and CH4 hydrogenation cycles. Finally, it is studied how hydrothermal aging in the presence of O2 modifies the physicochemical properties and the activity of the DFMs studied.Four DFMs were prepared by wetness impregnation. Firstly, appropriated amounts of Ca(NO3)2\u00b74H2O (Merck) and/or Na2CO3 (Riedel de-Ha\u00ebn) were impregnated over \u03b3-Al2O3 (Saint Gobain). The impregnated powder was dried at 120\u00a0\u00b0C overnight and then calcined at 550\u00a0\u00b0C for 4\u00a0h (1\u00a0\u00b0C\u00a0min\u22121). Afterwards, Ru(NO)(NO3)2 (Sigma Aldrich) and/or Ni(NO3)2\u00b76H2O (Fluka) were impregnated over the previous calcined powder. After drying at 120\u00a0\u00b0C, the DFMs were stabilized by calcining again at 550\u00a0\u00b0C for 4\u00a0h (1\u00a0\u00b0C\u00a0min\u22121). \nTable 1 lists the complete formulation of prepared DFMs and their nomenclature used in this work. The 10 %Ni/Al2O3 sample was also prepared as a reference.The calcined DFMs were placed in their granulated form (0.3\u20130.5\u00a0mm) in a quartz tube reactor and were heating from RT to 500\u00a0\u00b0C at 10\u00a0\u00b0C\u00a0min\u22121 during 1\u00a0h under 10 % H2/Ar (50\u00a0cm3 min\u22121).For studies of hydrothermal aging in the presence of oxygen, the DFMs were placed in their granulated form (0.3\u20130.5\u00a0mm) in a quartz tube reactor placed into a horizontal furnace. The DFMs were aged under 5 % H2O and 5 % O2 in Ar for 3\u00a0h, at a total flow rate of 550\u00a0cm3 min\u22121 at 550\u00a0\u00b0C.The specific surface, pore diameter and pore volume were determined from the N2 adsorption-desorption analysis. The reduced DFMs were pre-purged with nitrogen for 10\u00a0h at 300\u00a0\u00b0C using SmartPrep degas system (Micromeritics). Then the analysis were carried out at the nitrogen boiling temperature (\u2212196\u00a0\u00b0C) using an automated gas adsorption analyser (TriStar II, Micromeritics).X-ray diffraction spectra were obtained in a Philips PW1710 diffractometer. The samples were finely ground and were subjected to Cu K\u03b1 radiation in a continuous scan mode from 5\u00b0 to 70\u00b0 2\u03b8 with 0.02\u00b0 2\u03b8 per second sampling interval.Metallic dispersion was determined using the H2 chemisorption method in a Micromeritics ASAP 2020 equipment. Prior to the experiments, the samples (0.2\u00a0g) were reduced with pure H2 for 2\u00a0h at 500\u00a0\u00b0C. After that, the samples were degasified at the same temperature for 90\u00a0min. Finally, H2 was dossed for obtaining the adsorption isotherm at 35\u00a0\u00b0C.The morphology of the samples was analysed by transmission electron microscopy (TEM) in a JEM-1400 Plus instrument using a voltage of 100\u00a0kV. The reduced DFMs were dispersed in distilled water ultrasonically, and the solutions were then dropped on copper grids coated with lacey carbon film. In addition, High Angle Annular Dark Field (STEM-HAADF) images were obtained in a CETCOR Cs-probe-corrected Scanning Transmission Electron Microscopy microscope (ThermoFisher Scientific STEM, formerly FEI Titan3) operating at 300\u00a0kV and coupled with a HAADF detector (Fischione). The instrument had a normal field emission gun (Shottky emitter) equipped with a SuperTwin lens and a CCD camera. The samples were mixed with ethanol solvent and dropped onto a holey amorphous carbon film supported on a copper grid. In order to obtain spatially resolved elemental chemical analysis of the samples, the TEM apparatus was also equipped with an EDAX detector to carry out X-ray Energy Dispersive Spectroscopy (EDS) experiments. A 2k x 2k Ultrascan CCD camera (Gatan) was positioned before the filter for TEM imaging (energy resolution of 0.7\u00a0eV). The acquisition time for the analysis was 50\u00a0ms per spectrum and the used energy dispersion was 0.2\u00a0eV pixel\u22121.H2-TPD experiments were performed in a Micromeritics AutoChem II equipment. The samples (0.1\u00a0g) were pretreated at 500\u00a0\u00b0C (10\u00a0\u00b0C\u00a0min\u22121, 1\u00a0h) under 5 % H2/Ar (50\u00a0cm3 min\u22121) and then cooled down to 40\u00a0\u00b0C. After that, a 50\u00a0cm3 min\u22121 stream of pure hydrogen was fed long enough for complete saturation (60\u00a0min). Subsequently, DFMs were flushed out with Ar for 60\u00a0min in order to remove physisorbed H2. Finally, the desorption was conducted increasing temperature up to 750\u00a0\u00b0C (10\u00a0\u00b0C\u00a0min\u22121). The hydrogen desorbed was continuously monitored with a TCD detector.CO2-TPD experiments were carried out in a Micromeritics AutoChem II equipment. The samples (0.1\u00a0g) were pretreated at 500\u00a0\u00b0C (10\u00a0\u00b0C\u00a0min\u22121, 1\u00a0h) under 5 % H2/Ar (50\u00a0cm3 min\u22121) and then cooled down to 50\u00a0\u00b0C. Then, the samples were exposed to a gas stream composed of 5 % CO2/He (50\u00a0cm3 min\u22121) for 1\u00a0h at 50\u00a0\u00b0C to saturate the catalyst with CO2. Subsequently, the samples were exposed to He (50\u00a0cm3 min\u22121) for 90\u00a0min to remove the physically adsorbed CO2 and, finally, they were heated from RT to 1000\u00a0\u00b0C at 10\u00a0\u00b0C\u00a0min\u22121. The CO2 released was measured by mass spectrometry (HIDEN ANALYTICAL HPR-20 EGA).TPSR experiments were carried out in a quartz tube reactor placed in a horizontal furnace. The DFMs (0.3\u00a0g) were pretreated under 5 % H2/Ar at 500\u00a0\u00b0C (10\u00a0\u00b0C\u00a0min\u22121, 1\u00a0h), and then, the sample was cooled down to 50\u00a0\u00b0C. Subsequently, the DFMs were exposed to a gas stream composed of 25 % CO2/Ar with a flowrate of 700\u00a0cm3 min\u22121 for 20\u00a0min at 50\u00a0\u00b0C to saturate the catalyst with CO2. Previous to the experiments, the DFMs were purged in Ar (700\u00a0cm3 min\u22121) for 90\u00a0min to remove the physically adsorbed CO2. Afterwards, the samples were heated from 50\u00b0 to 600\u00a0\u00b0C at 10\u00a0\u00b0C\u00a0min\u22121 in a 5 % H2/Ar mixture with a flowrate of 700\u00a0cm3 min\u22121. The MultiGas 2030 FT-IR analyzer was used to quantify the formation of products during the reduction in the reactor effluent gas.H2-TPR experiments were carried out in a Micromeritics AutoChem II equipment. The samples (0.1\u00a0g) was loaded in a quartz tube reactor and pretreated at 350\u00a0\u00b0C for 15\u00a0min under 5 % O2/He (30\u00a0cm3 min\u22121) and then cooled down to 35\u00a0\u00b0C. The reducing gas flow was 30\u00a0cm3 min\u22121 of 5 % H2/Ar and the temperature was increased from 35\u00b0 to 1000\u00b0C with a heating rate of 10\u00a0\u00b0C\u00a0min\u22121. The water formed during reduction was trapped using a cold trap and the hydrogen consumption was continuously monitored with a TCD detector.The catalytic activity in the cyclic operation of CO2 adsorption and hydrogenation to CH4 of the DFMs was evaluated in a vertical tubular stainless steel reactor. The reactor was loaded with 1\u00a0g of DFM whose particle size was between 0.3 and 0.5\u00a0mm. Prior to the cycles, the DFM were reduced with a stream composed of 10 % H2/Ar. First, the temperature is increasing from RT to 500\u00a0\u00b0C and then is maintaining for 60\u00a0min. Once the DFMs were reduced, the temperature is stabilized at 280\u00a0\u00b0C in Ar and cycles of CO2 adsorption and hydrogenation to CH4 were started. The reaction temperature was varied between 280 and 520\u00a0\u00b0C, with intervals of 40\u00a0\u00b0C. At each temperature several isothermal cycles have been carried out until obtaining three consecutive quasi-identical cycles (cycle-to-cycle steady state). During the adsorption period, a stream with 10 % CO2 (Ar balance) was fed for 1\u00a0min, followed by a purge with Ar for 2\u00a0min to remove weakly adsorbed CO2. Next, during the hydrogenation period, a stream consisting of 10 % H2/Ar was fed for 2\u00a0min, followed by an Ar purge for 1\u00a0min before starting the adsorption period again. The total flow rate in the whole experiment was 1200\u00a0cm3 min\u22121, which corresponds to a space velocity of 45000\u00a0h\u22121. The flue gas composition was continuously measured using the MultiGas 2030 FT-IR analyzer for quantitative analysis of CO2, CH4, CO and H2O.The amount of CO2 stored was calculated from Eq. (1). For that, the amount that leaves the reactor was subtracted from the amount fed. To determine the amount of CO2 fed, the stream from the feed system was led directly to the analyser. This profile corresponds to the actual CO2 input that was fed to the reactor.\n\n(1)\n\n\nstored\n\n\n\nCO\n\n\n2\n\n\n\n(\n\n\n\n\u03bc\nmol\n\ng\n\n\n\u2212\n1\n\n\n\n)\n\n=\n\n1\n\nW\n\n\n\n\n\u222b\n\n\n0\n\n\nt\n\n\n\n[\n\n\n\nF\n\n\n\n\nCO\n\n\n2\n\n\n\n\nin\n\n\n\n(\nt\n)\n\n\u2212\n\n\nF\n\n\n\n\nCO\n\n\n2\n\n\n\n\nout\n\n\n\n(\nt\n)\n\n\n]\n\n\nd\nt\n\n\n\n\nThe CH4, H2O and CO productions were calculated from the following expressions:\n\n(2)\n\n\n\n\nY\n\n\n\n\nCH\n\n\n4\n\n\n\n\n\n(\n\n\n\n\u03bc\nmol\n\ng\n\n\n\u2212\n1\n\n\n\n)\n\n=\n\n1\nW\n\n\n\u222b\n0\nt\n\n\n\n\nF\n\n\n\n\nCH\n\n\n4\n\n\n\n\nout\n\n\n\n(\nt\n)\n\nd\nt\n\n\n\n\n\n\n\n(3)\n\n\n\n\nY\n\n\nCO\n\n\n\n(\n\n\n\n\u03bc\nmol\n\ng\n\n\n\u2212\n1\n\n\n\n)\n\n=\n\n1\nW\n\n\n\u222b\n0\nt\n\n\n\n\nF\n\n\nCO\n\n\nout\n\n\n\n(\nt\n)\n\nd\nt\n\n\n\n\n\n\n\n(4)\n\n\n\n\nY\n\n\n\n\nH\n\n\n2\n\n\nO\n\n\n\n(\n\n\n\n\u03bc\nmol\n\ng\n\n\n\u2212\n1\n\n\n\n)\n\n=\n\n1\nW\n\n\n\u222b\n0\nt\n\n\n\n\nF\n\n\n\n\nH\n\n\n2\n\n\nO\n\n\nout\n\n\n\n(\nt\n)\n\nd\nt\n\n\n\n\n\nCH4 selectivity is determined by relating the CH4 and CO productions since they were the only carbon based products that were detected:\n\n(5)\n\n\n\n\nS\n\n\nCH4\n\n\n\n(\n%\n)\n\n=\n\n\n\n\nY\n\n\n\n\nCH\n\n\n4\n\n\n\n\n\n\n\n\nY\n\n\n\n\nCH\n\n\n4\n\n\n\n\n+\n\n\nY\n\n\nCO\n\n\n\n\n\u00d7\n100\n\n\n\n\nThe carbon balance check was carried out from the following expression:\n\n(6)\n\n\n\n\ns\n\n\nCB\n\n\n\n(\n%\n)\n\n=\n\n(\n\n\n\n\n\nY\n\n\n\n\nCH\n\n\n4\n\n\n\n\n+\n\n\nY\n\n\nCO\n\n\n\n\ns\nt\no\nr\ne\nd\n\nC\n\n\nO\n\n\n2\n\n\n\n\n\u2212\n1\n\n)\n\n\u00d7\n100\n\n\n\n\n\n\nTable 2 lists the textural properties of the alumina and the four DFMs synthesized after the reduction protocol. The alumina used has a specific surface area of 197\u2009m2 g\u22121 and the specific surface values of the DFMs are between 101 and 119\u2009m2 g\u22121. Al2O3 has a mean pore size of 122\u2009\u00c5 and a pore volume of 0.62\u2009cm3 g\u22121. On the other hand, the DFMs present larger mean pore size values (147\u2013158\u2009\u00c5) and smaller pore volume values (039\u20130.46\u2009cm3 g\u22121). For that, the decrease in the specific surface area is attributed to a lower proportion of alumina in the formulation and to a partial blocking of the pores caused by the incorporation of the adsorbent and the metal. Comparing the DFMs with each other, it is observed that the joint presence of Na and Ca decreases the specific surface to a greater extent (101\u2013103 vs. 113\u2013119\u2009m2 g\u22121). Note that the total adsorbent content is the same in all DFMs (16 %). The joint presence of Na and Ca reduces the pore volume to a greater extent (0.39\u20130.40 vs. 0.43\u20130.46\u2009cm3 g\u22121). This fact indicates that the impregnation of two adsorbent phases blocks a greater number of pores.\n\nFig. 1 collects the diffraction spectra of the alumina used as support and of the reduced DFMs. The spectrum of alumina shows broad peaks of low intensity, characteristic of an amorphous solid. This background is also seen in all the DFMs spectra, in which only two additional peaks appear at 44.6 and 51.8\u00b0 2\u03b8 belonging to metallic nickel. In contrast, no peaks are identified for Na and Ca phases, which indicates that they are highly dispersed and/or they present poor crystallinity. Furthermore, no peaks are detected belonging to Ru-based phases in 1Ru10Ni-Na and 1Ru10Ni-NaCa formulations, indicating a high dispersion.\n\nTable 3 collects the metallic dispersion values estimated by H2 chemisorption of the DFMs. Disparate values (2.2\u20135.9 %) are obtained depending on the composition of the DFMs. The Ru-promoted DFMs show a noticeably higher dispersion (5.3\u20135.9 %) compared to the bare Ni DFMs (2.2\u20132.4 %). Taking into account that ruthenium is also reduced and both compounds can chemisorb H2 with a stoichiometric H/Me=\u20091/1 (Me=Ni or Ru) [30], these results denote that Ru promotes metallic phases dispersion.In order to analyze if the increase in dispersion is only due to the presence of a higher metallic content or if a synergistic effect is also being created between both metals, the chemisorbed H atoms for each DFM in the H2 chemisorption experiments are also determined and collected in Table 3. The bimetallic DFM 1Ru10Ni-Na presents a dispersion of 5.9 %. This dispersion corresponds to an H chemisorption capacity of Ru and Ni of 106.3 \u03bcmol g\u22121. In a previously prepared monometallic Ru-DFM under similar conditions (4Ru-Na, [29]) a dispersion of 19.6 % has been obtained, which corresponded to an H chemisorption capacity of 77.6 \u03bcmol g\u22121. Considering that 4Ru-Na DFM was composed of 4 % Ru and in the 1Ru10Ni-Na DFM the ruthenium content is of 1 %, we estimate that the ruthenium in 1Ru10Ni-Na DFM would be capable of chemisorbing 19.4 \u03bcmol g\u22121. Therefore, the rest of H chemisorbed in the bimetallic sample (86.9 \u03bcmol g\u22121) should correspond to nickel phase. This H chemisorption capacity is significantly higher to that observed for monometallic Ni-based sample (10Ni-Na = 37.5 \u03bcmol g\u22121). Therefore, the presence of Ru seems to promote the H chemisorption capacity of nickel.Similar results are obtained for DFMs with the joint presence of Na and Ca. The 10Ni-NaCa monometallic DFM has an H chemisorption capacity of 40.9 \u03bcmol g\u22121, while the contribution of nickel in the bimetallic DFM is 70.9 \u03bcmol g\u22121. The contribution of ruthenium (24.6 \u03bcmol g\u22121), in 1Ru10Ni-NaCa is estimated from the 4Ru-NaCa DFM [29]. Based on these results, the synergistic effect between nickel and ruthenium is confirmed, which leads to a larger exposed metallic nickel surface with respect to monometallic 10Ni-NaCa and 10Ni-Na samples. At this point, some authors have indicated that due to the higher melting point of ruthenium (2334\u2009\u00b0C) compared to nickel (1455\u2009\u00b0C), Ru can act as a shell with Ni in the nucleus, remaining protected it from sintering during the calcination stage [31,32]. On the other hand, the presence of Na or Na/Ca together does not seem to have a significant influence on the metallic dispersion.\n\nFig. 2 collects the TEM micrographs of the DFMs. Based on the assignment made in previous works [33] for a similar formulation (Ni/Al2O3 catalysts), where the authors assigned the dark areas in the micrographs to the Ni metal particles, while the gray areas to the \u03b3-Al2O3 support, we assign the darker spherical areas in the Ni-only DFMs (10Ni-Na and 10Ni-NaCa) to Ni particles. In fact, this assignment is in line with their higher atomic number (28) with respect to the adsorbents (Na = 11 or Ca = 20) and the support Al =\u200913).Regarding to the bimetallic DFMs (1Ru10Ni-Na and 1Ru10Ni-NaCa), the dark spherical particles are assigned to particles of Ni, Ru, or alloys of both in line with the observed by Li et al. [31] for the reduced Ni-Ru nanoparticles supported on SiO2 by TEM images. Similar assignment was made by Zhou et al. [34] in their study of bimetallic Ru\u2013M/TiO2 (M=Fe, Ni, Cu, Co) nanocomposite catalysts, where Ru-Fe, Ru-Ni and Ru-Co nanoparticles were observed in darker contrast. In fact, this assignment was further verified by EDX in their study. Thus, these results suggest the conformation of bimetallic Ni-Ru particles in the bimetallic DFMs (1Ru10Ni-Na and 1Ru10Ni-NaCa).Further evidences of this aspect as well as of the disposition of the metals and the adsorbent phases on the support have been obtained by HAADF images and EDX maps. \nFig. 3 shows the EDX maps obtained for each compound (Ru, Ni, Na, Ca and Al) in the DFM with the more complex composition (1Ru10Ni-NaCa). As can be observed, both adsorbent phases (Na and Ca) are homogenously disposed, covering the entire surface of the support (Al). In contrast, Ru and Ni coexist and are preferentially placed in the central area of the image. In fact, smaller Ru nanoparticles seem to be deposited on large Ni particles. Thus, these results confirm the existence of bimetallic Ni-Ru particles, as previously suggested by TEM images (Fig. 2).An average size of metallic or bimetallic particles is determined by counting at least 100 particles (Fig. 2) and the values are shown in Table 2. Very close values are obtained for the four DFMs (9\u201311\u2009nm), although the lowest values are presented by the DFMs with ruthenium in its formulation. At this point, a greater difference in metallic dispersions (estimated by H2 chemisorption) compared to particle sizes is observed. It is suggested that some of the particles observed in the TEM images of the DFMs are not completely reduced.In order to clarify this discrepancy, the hydrogen adsorption capacity by H2-TPD is also determined. \nFig. 4 shows the H2-TPD profiles of alumina and the synthesized DFMs. It can be seen that all the profiles exhibit two bands, before and after 450\u2009\u00b0C. The bands below 450\u2009\u00b0C are attributed to H2 chemisorbed on metallic particles (type I). On the other hand, the bands at higher temperatures are associated with the H2 desorption from the subsurface layers of the alumina or with the spillover phenomenon (type II) [35,36]. The alumina profile does not show variation below 350\u2009\u00b0C. At higher temperatures, an intense band appears, which is related to the dehydroxylation of alumina. On the other hand, the bands of the DFMs at low temperatures can be divided into several peaks. Ewald et al. [37], in their study of Ni/Al2O3 catalysts, assigned the main peak to hydrogen chemisorbed on the Ni surface. On the other hand, the peaks located at lower temperatures were attributed to hydrogen adsorbed on the corners of large Ni particles or on better dispersed particles. In DFMs, the position of the main peak shifts towards lower temperatures and its intensity increases with the addition of Ru. This fact shows a greater number of exposed metal atoms. The amount of H2 desorbed below 450\u2009\u00b0C is quantified and shown in Table 3. It can clearly be seen that in the DFMs promoted with Ru the amount desorbed is more than double compared to the DFMs only with nickel. Note that the ability of DFMs to supply dissociated hydrogen increases markedly with the addition of Ru. The increase in the number of exposed metal atoms has also been reported by other authors who incorporated Ru [38], Cr [39] or Fe [40] to Ni/Al2O3 catalysts.Based on H2 chemisorption (Table 3), TEM micrographs (Fig. 2) and H2-TPDs experiments (Fig. 4), it can be concluded that the Ru incorporation in the DFMs notably increases the dispersion and, consequently, the metallic surface area exposed. On the other hand, the presence of Na or the joint presence of Na and Ca does not significantly modify the disposition of the metals (Ni and Ru).The basicity of the samples is evaluated by CO2 temperature programmed desorption. \nFig. 5 shows the evolution of the CO2 signal measured by a mass spectrometer (m/e=44) as a function of temperature during the CO2-TPD experiments of alumina, the reference sample only with nickel (10Ni) and the DFMs. Depending on the desorption temperature, weak, medium and strong sites are distinguished. Weak basic sites are unstable and easily decompose below 200\u2009\u00b0C. Medium basic sites decompose between 200 and 600\u2009\u00b0C while strong basic sites are highly stable and decompose at temperatures above 600\u2009\u00b0C.Alumina exhibits a small desorption peak at low temperature. This peak is assigned to the decomposition of bicarbonates resulting from the interaction between CO2 and the surface hydroxyl groups of the alumina [41,42]. The incorporation of 10 % nickel (10Ni) does not modify the profile, obtaining also only a small desorption peak at low temperature. Porta et al. [13] did not observe a modification of the desorption profile with the incorporation of Ru over alumina support.The incorporation of Na, or the joint incorporation of Na and Ca, notably modifies the desorption profiles. With the incorporation of the adsorbent, the amount of CO2 desorbed increases remarkably. Porta et al. [14] studied by infrared spectroscopy the nature of CO2 adsorbed on DFMs based on potassium and barium as adsorbents and on ruthenium as metal. The authors only observed bands belonging to adsorbed carbonates on the adsorbent phases and did not observe surface species typical of CO2 adsorption on alumina. Based on these results, they were able to conclude that the adsorbent phases were covering the surface of the alumina.In order to compare the evolution of the different basic sites concentration depending on sample formulation, \nTable 4 shows the weak, medium, strong and total basicity for the reference samples and the DFMs. Alumina and the reference sample with nickel (10Ni) only show weak basicity. Also, as mentioned above, the amount desorbed is much lower compared to DFMs. The four synthesized DFMs present the three types of basicity. The addition of 1 % ruthenium does not notably modify the distribution of the basic sites. However, the joint presence of Na and Ca promotes medium basicity and strong basicity, and therefore total basicity (471\u2013446 vs. 372\u2013387 \u03bcmol g\u22121). This aspect is of special interest since in the cyclic process of CO2 adsorption and hydrogenation to CH4, only the CO2 that remains adsorbed during the storage period is the one that can be converted into methane in the hydrogenation period.The reactivity of CO2 adsorbed on DFMs is studied by temperature programmed surface reaction (TPSR) experiments. First, the sample is reduced at a controlled temperature up to 500\u2009\u00b0C with an isothermal step of one hour, to simulate the state of the DFMs in reaction. Next, the DFMs are saturated with CO2 and, finally, a stream with H2 is passed while the temperature is increased with a ramp of 10\u2009\u00b0C\u2009min\u22121 up to 600\u2009\u00b0C. During the experiment, the reactor outlet gases are continuously measured by an FTIR analyzer.\n\nFig. 6 shows the evolution of CH4 production with temperature for the four synthesized DFMs. In general, a main peak with a shoulder that extends up to 600\u2009\u00b0C is observed. The main peak is assigned to the hydrogenation of the CO2 adsorbed on the basic sites with weak strength, while the shoulder is assigned to the adsorbed on the sites with medium and strong strength basic sites. The DFMs without ruthenium in their composition (10Ni-Na and 10Ni-NaCa) show the start of CH4 production around 300\u2009\u00b0C. On the other hand, the incorporation of Ru leads to the start of production at a lower temperature (250\u2009\u00b0C). The amount of CH4 produced in each experiment is obtained by integrating the profiles and the values are shown in \nTable 5. The presence of ruthenium in the DFM, in addition to shifting production to a lower temperature, increases the amount of CH4 produced. On the other hand, the presence of Na or the joint presence of Na and Ca does not significantly modify the TPSR profiles.If the values of total basicity obtained (Table 4) are compared with the production of CH4 in the TPSR (Table 5), in all cases lower values are obtained for the TPSR experiments. In fact, methane/basicity ratios of 0.30\u20130.47 are obtained. Note that part of the CO2 is released at low temperatures when DFMs are not active for CO2 methanation. The highest ratios (0.43\u20130.47) are presented by the DFMs with ruthenium in their composition. Consequently, the DFMs 1Ru10Ni-Na and 1Ru10Ni-NaCa show the start of production at lower temperatures (Fig. 6), a greater amount of CH4 produced (Table 5) and the highest methane/basicity ratios. The best results are attributed to the synergies between Ni and Ru and to the higher intrinsic activity in CO2 methanation of ruthenium compared to nickel [43], which displaces the start of methanation at lower temperatures.The reducibility of the prepared samples is determined by H2 temperature programmed reduction (H2-TPR). \nFig. 7 shows the evolution of H2 consumption during the H2-TPR experiments of the DFMs prepared together with the reference sample without adsorbent (10Ni). The H2 consumption profile for the 10Ni reference sample (Fig. 7) can be deconvolved into three main contributions. According to the reported in the literature [44], the reducible Ni2+ species are generally classified as: \u03b1-NiO, \u03b2-NiO and \u03b3-NiO. The \u03b1-NiO species are related to free and easily reducible NiO species. \u03b2-NiO species are reduced at intermediate temperatures and are related to Ni2+ species that are not fully integrated into the nickel aluminate spinel structure. Finally, the reduction of the \u03b3-NiO species occurs at higher temperatures, related to the Ni that integrates into the nickel aluminate structure. These species are reduced to metallic nickel according to the following reactions: NiO+H2\u21c6Ni+H2O and NiAl2O4 +H2\u21c6Ni+Al2O3 +H2O, with Ni/H2 =\u20091 stoichiometry. In fact, the experimental H2/Ni ratio obtained is 1.01. This fact indicates that nickel is found as Ni2+, either as NiO, NiAl2O4 or a mixture of both.With the incorporation of the adsorbent, the profiles vary significantly. H2 consumption increases and shifts to lower temperatures. For the DFMs with only nickel, 10Ni-Na and 10Ni-NaCa, the H2/Ni ratio is 2.5 and 2.1, respectively. On the other hand, in the DFMs with ruthenium in their formulation, this is also reduced according to the RuO2 +\u20092\u2009H2\u21c6Ru+\u20092H2O reaction. Therefore, Ru is reduced with a H2/Ru ratio of 2. In both DFMs there is 1 % Ru, which consumes 99 \u03bcmol g\u22121 of H2 in its reduction. If the H2 consumption relative to ruthenium is subtracted from the total, the H2/Ni ratio results in 2.7 and 2.2, for the DFMs 1Ru10Ni-Na and 1Ru10Ni NaCa, respectively. At this point, an additional phenomenon is evident that contributes to the increase in hydrogen consumption observed for the four DFMs. Fig. S1 shows the evolution of the concentration of NO, NH3 and CH4 measured by the FTIR analyzer during the pretreatment H2-TPR of the TPSR experiment. The formation of NO is related to the decomposition of residual nitrates belonging to the adsorbent and metal precursors that have not been completely decomposed during the calcination step. The start of NH3 formation is detected at higher temperatures. NH3 formation requires the noble metal in its metallic state, so NH3 formation can be used as an indirect way of determining the temperature at which nickel begins to reduce. On the other hand, the formation of CH4 is attributed to the hydrogenation of the CO2 adsorbed on the samples, due to exposure to the environment before the experiment.With this additional information, the hydrogen consumption profiles of the DFMs can be correctly interpreted (Fig. 7). H2 consumption is related to the reduction of metals, residual nitrates and carbonates. On the other hand, the displacement of consumption at lower temperatures is assigned to a lower interaction between nickel and alumina due to the presence of the adsorbent. Keep in mind that in the DFM synthesis process, the adsorbent is incorporated in the first impregnation and the metals in the second.The H2 consumption profiles are different depending on the composition of the DFMs (Fig. 7). The DFMs 10Ni-Na and 10Ni-NaCa show the consumption start at 345 and 375\u2009\u00b0C. In this peak, the reduction of the \u03b1-NiO species occurs, as well as the hydrogenation of the nitrates and carbonates. The peak centered at 600\u2009\u00b0C is assigned to the reduction of \u03b2-NiO species and the peak centered at 770\u2009\u00b0C to the reduction of \u03b3-NiO species. On the other hand, with the addition of 1 % Ru (1Ru10Ni-Na and 1Ru10Ni-NaCa) the start of consumption shifts to 250 and 280\u2009\u00b0C and a peak with a larger area is obtained. In Fig. S1 it can be seen how in the DFMs with Ru in their composition, the production of NH3 and CH4 shifts towards lower temperatures and the amounts produced increase. On the other hand, the reduction shoulder of the \u03b2-NiO species disappears. However, the peak centered at 770\u2009\u00b0C assigned to the reduction of the \u03b3-NiO species is detected. It is evident that Ru promotes Ni reduction at lower temperatures. This fact is of great importance since Ni and Ru in their metallic state are the active sites for the hydrogenation of CO2 and, therefore, the hydrogenation activity would be expected to increase with the amount of reducible nickel.The presence of \u03b3-NiO species is detected in all four DFMs. The H2 consumption of these species is approximately 250 \u03bcmol of H2 g\u22121, therefore 16 % of nickel is in the form of spinel and is not reduced during pretreatment. On the other hand, part of the \u03b2-NiO species in the DFMs without ruthenium are not reduced either. At this point, the percentage of nickel reduced in the DFMs 1Ru10Ni-Na and 1Ru10Ni-NaCa is 84 %, while in the DFMs 10Ni-Na and 10Ni-NaCa this percentage decreases to 70 %. This difference explains the similar particle sizes obtained in the TEM images. It is confirmed that some of the particles observed in the TEM images of the DFMs are not reduced and therefore are not capable of chemisorbing H2. The calcination-reduction temperature selected in this study (550\u2013500\u2009\u00b0C), has been optimized in a previous work [24]. Finally, based on H2-TPRs (Fig. 7 and Fig. S1), it can also be concluded that the reduction pretreatment, carried out before the activity tests, completely eliminates the residual nitrates and carbonates.The catalytic activity is evaluated in cycles of CO2 adsorption and hydrogenation to CH4. Each cycle has a total duration of six minutes. First, it starts the adsorption period by introducing a stream of 10 % CO2/Ar for one minute, followed by a two-minute purge. The hydrogenation period is then started by introducing a stream of 10 % H2/Ar for two minutes and an additional one-minute purge takes place before starting the next cycle. \nFig. 8 shows the evolution of the concentration of CO2, CH4, H2O and CO in a complete cycle at 400\u2009\u00b0C for the four DFMs. The detailed description of the temporal evolution of reactants and products, as well as the mechanism, is detailed in previous works [18,20,29], as well as its modeling and simulation [45,46]. \nTable 6 collects the reactions suggested in each period. In the adsorption period, CO2 remains adsorbed forming carbonates. CO2 can be adsorbed on oxide (Eq. 7 and Eq. 8) or hydroxide sites (Eq. 9 and Eq. 10). On the other hand, in the hydrogenation period, the carbonates are decomposed by the presence of hydrogen (Eq. 11 and Eq. 12), the desorbed CO2 is hydrogenated to CH4 (Eq. 13) and part of the produced water remains adsorbed forming hydroxides (Eq. 14 and Eq. 15).If the evolutions of the reactants and products are compared (Fig. 8), depending on the DFM formulation, different concentrations are observed. For a more detailed interpretation, the quantities of CH4 and CO produced (Eq. 2 and Eq. 3, respectively), at each operating temperature, are determined and plotted in \nFig. 9. In all cases, the carbon balance closes with an error less than \u00b1\u20095 % (Eq. 6) and the H2O/CH4 ratio is very close to 2, in agreement with the stoichiometric of the Sabatier reaction (Eq. 13).CH4 production (Fig. 9a) shows a maximum around 400\u2009\u00b0C. At low operating temperatures, the ability to extract the stored CO2 is limited (Fig. 5 and Table 4), while at high temperatures the storage capacity is limited during the adsorption period due to the destabilization of carbonates. The DFM 10Ni-Na presents the lowest CH4 production in the temperature range studied, even so, it reaches 172 \u03bcmol g\u22121 at 400\u2009\u00b0C.The joint presence of Na and Ca (10Ni-NaCa) promotes the CH4 production at medium-high temperatures (360\u2013520\u2009\u00b0C). In the CO2-TPD experiments (Fig. 5 and Table 4) it has been observed that the presence of both adsorbents promotes total basicity since it increases the sites with medium and strong strength. At this point, this DFM is capable of retaining a greater amount of CO2 in the storage period, which is reflected in a greater CH4 production in the hydrogenation period. Specifically, the DFM 10Ni-NaCa produces 202 \u03bcmol g\u22121 at 400\u2009\u00b0C.The incorporation of 1 % Ru (1Ru10Ni-Na) to the DFM also improves CH4 production. On this occasion, it is promoted in the entire temperature range studied (280\u2013520\u2009\u00b0C), although especially at low-intermediate temperatures (280\u2013440\u2009\u00b0C). The DFM 1Ru10Ni-Na reaches to produce 250 \u03bcmol g\u22121 at 400\u2009\u00b0C. The production enhancement is assigned to a synergistic effect between Ru and Ni. As has been observed by H2-TPR (Fig. 7), Ru promotes the reduction of Ni in a more extent at lower temperature. In addition, the presence of Ru increases markedly the dispersion, thus increasing the number of active sites for CO2 hydrogenation. Liu et al. also observed an improvement in activity after the incorporation of Ru in a nickel-based catalyst for the continuous methanation of CO2\n[38]. The authors also highlighted the synergistic effects between Ni and Ru bimetallic catalysts, as well as their enhanced H2 and CO2 chemisorption capabilities.Finally, the DFM 1Ru10Ni-NaCa presents the highest CH4 production in the entire range of temperature studied. The promotion of low-temperature production is mainly attributed to the presence of Ru and the synergistic effect between Ru and Ni as above described. On the other hand, the promotion at high temperature is assigned to the modulation of basicity by the joint presence of Na and Ca. At this point, in the DFM 1Ru10Ni-NaCa, the contact between carbonates and metallic sites is promoted to a greater extent and consequently CO2 adsorption and hydrogenation to CH4 are also improved. Specifically, the DFM 1Ru10Ni-NaCa produces 298 \u03bcmol g\u22121 at 440\u2009\u00b0C.\nFig. 9b shows the CO production for the four formulations of DFMs. An upward trend is observed with temperature since its increase favors the RWGS (Eq. 16) in contrast to the methanation reaction (Eq. 13). \nTable 7 shows the selectivity (Eq. 5) of the DFMs operating at 400\u2009\u00b0C together with the CH4 production. DFM 10Ni-Na has a selectivity of 87.8 %. The incorporation of Ru increases the selectivity to 96.4 % and the joint presence of Na and Ca to 94.1 %. The improvement due to the presence of Ru is attributed to the fact that the greater number of available metallic sites favors the total hydrogenation of CO2 to CH4. On the other hand, the addition of Ca to the catalysts for CO2 methanation improves the selectivity to CH4 by strengthening CO2 chemisorption, while the addition of Na favors the formation of CO [47,48]. In addition, it has recently been concluded that the joint presence of both adsorbents further limits CO production by favoring contact between carbonates and metal sites [29]. Finally, the DFM 1Ru10Ni-NaCa presents the highest selectivity value (98.4 %). It is suggested that the two positive effects are added up.\n\n(16)\nCO2+H2\u21c6CO+H2O\n\n\nIn order to analyze the resistance of DFMs to the presence of O2 and H2O vapor, the effect of hydrothermal aging in the presence of oxygen is studied. For this, the DFMs were placed in a tubular quartz reactor placed in a horizontal furnace and a stream with 5 % H2O steam and 5 % O2 was fed for 3\u2009h at 550\u2009\u00b0C. This strategy is commonly used for NSR or SCR catalysts for NOx removal in diesel vehicle engines [49\u201351]. In this way, the state of the catalyst at the end of the vehicle's life can be simulated on a laboratory scale.\n\nFig. 10 shows the CH4 production at 400\u2009\u00b0C from the DFMs before (1st column) and after (2st column) the hydrothermal aging in the presence of O2. The aging process limits CH4 production from all DFMs. Specifically, production is reduced between 37 % and 47 %. Table S1 shows the textural properties and metal dispersion values of the aged DFMs. The aging process causes a reduction in specific surface area and pore volume and an increase in pore size. Continued exposure of DFMs to temperature in the presence of O2 and H2O causes sintering of the catalytic phase and agglomeration of the adsorbent, causing irreversible blockage of the smallest pores. Burger et al. [52] observed a progressive decrease in the specific surface area of NiAl2Ox and NiFeAl2Ox catalysts as the operation time with continuous CO2 and H2 feeding increased. The authors attributed the decrease to Ni particle growth and sintering of the mixed oxide phase. De-La-Torre et al. [49] also observed a reduction in specific surface area after hydrothermal aging for Pt-Ba/Al2O3 and Pt-Ce-Ba/Al2O3 NSR catalysts. The authors attributed the decrease to the formation of barium aluminate and blockage of the alumina pores by platinum and cerium. On the other hand, the aging process also causes a drastic reduction in the metallic dispersion of the DFMs (Table S1). The dispersion values of the aged DFMs are reduced between 53 % and 62 %. Again, the DFMs with ruthenium in their formulation (1Ru10Ni-Na and 1Ru10Ni-NaCa) show the highest values, 2.1 % and 2.8 %, respectively.If the CH4 productions of the aged DFMs (Fig. 10) are compared, they show a trend similar to that observed in Fig. 9. The lowest production is shown by the DFM 10Ni-Na (108 \u03bcmol g\u22121) with a selectivity of 88 %. Then the DFM 10Ni-NaCa produces 128 \u03bcmol g\u22121 with a selectivity of 96 %, followed by the DFM 1Ru10Ni-Na which produces 131 \u03bcmol g\u22121 with a selectivity of 97 %. It is confirmed that the promotion with Ru or the joint presence of Na and Ca also promote the production of CH4 after aging. Finally, DFM 1Ru10Ni-NaCa shows the highest production (155 \u03bcmol g\u22121) with the highest selectivity (98 %). Consequently, the contact between carbonates with metal sites and consequently the adsorption of CO2 and hydrogenation to CH4 is still promoted to a greater extent.The feasibility of boosting the CH4 production of the DFM 10Ni-16Na/Al2O3 through the joint presence of Na/Ca and the Ru incorporation has been studied. For that, four DFMs have been prepared by sequential wetness impregnation in which their formulation has been varied. The Ru incorporation notably increases the metallic dispersion. Synergistic aspects are created between Ni and Ru that limit nickel sintering and induce its reduction at lower temperatures. The reducibility is increased from 70 % (10Ni-Na and 10Ni-NaCa) to 84 % (1Ru10Ni-Na and 1Ru10Ni-NaCa). Furthermore, in TPSR experiments, CH4 production increases and starts at lower temperatures, confirming the synergies observed by other techniques. In short, the DFMs with Ru in their composition have a greater number of exposed metal atoms, so they have a greater capacity to supply dissociated hydrogen. On the other hand, the joint presence of Na and Ca does not seem to influence the disposition of the metallic phases.The four synthesized DFMs present three types of basicity (weak, medium and strong). On this occasion, the addition of 1 % ruthenium does not significantly modify the distribution of basic sites. However, the joint presence of Na and Ca promotes medium and strong basicity, and therefore total basicity. Note that all the DFMs present the same adsorbent loading (16 %), therefore the joint presence of Na and Ca improves the adsorption capacity, especially at medium-high temperatures.In the operation in cycles of CO2 adsorption and hydrogenation to CH4, once the cycle-to-cycle steady state is reached, the process is cyclic and repetitive. For all DFMs and at all operating temperatures (280\u2013520\u2009\u00b0C) the error with which the carbon balance is closed is less than \u00b1\u20095 % and the H2O/CH4 ratio is very close to 2, according to the stoichiometric of the Sabatier\u00b4s reaction. CH4 production shows a maximum around 400\u2009\u00b0C. At lower temperatures the extraction capacity of the adsorbed CO2 is limited, while at higher temperatures the storage capacity is limited. The lowest CH4 production is presented by the reference DFM 10Ni-Na, even so, it produces 172 \u03bcmol g\u22121 at 400\u2009\u00b0C. The joint presence of Na and Ca (10Ni-NaCa) improves the production at intermediate-high temperatures by boosting the adsorption capacity of the DFM, specifically it produces 202 \u03bcmol g\u22121 at 400\u2009\u00b0C. On the other hand, the Ru incorporation boosts the production in the entire range of temperatures studied, although especially at low-intermediate temperatures. The improvement is assigned to the synergistic aspects between Ni and Ru and to the higher number of exposed metal atoms. The DFM 1Ru10Ni-Na produces 250 \u03bcmol g\u22121 at 400\u2009\u00b0C. Finally, the joint presence of Na and Ca and the simultaneous Ru incorporation (1Ru10Ni-NaCa) presents the highest CH4 production in the entire range of temperatures studied. It is concluded that both effects are added. Mainly, at low temperatures, the presence of Ru boost the production, while at high temperatures, the greatest adsorption capacity is due to the joint presence of Na and Ca. Specifically, it produces 298 \u03bcmol g\u22121 at 440\u2009\u00b0C. In addition, it is the DFM with the highest selectivity to CH4 and the one that also presents the highest activity and selectivity after hydrothermal aging in the presence of O2.Alejandro Bermejo-L\u00f3pez: Validation, Methodology, Investigation, Writing \u2013 original draft. Be\u00f1at Pereda-Ayo: Conceptualization, Methodology, Visualization, Writing \u2013 review & editing. Jon A. Onrubia-Calvo: Methodology, Visualization, Writing \u2013 review & editing. Jos\u00e9 A. Gonz\u00e1lez- Marcos: Methodology, Data curation, Supervision, Funding acquisition. Juan R. Gonz\u00e1lez-Velasco: Conceptualization, Supervision, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Support for this study was provided by Project PID2019\u2013105960RB-C21 by MCIN/AEI /10.13039/501100011033 and the Basque Government (Project IT1509-2022). The authors thank for technical and human support provided by SGIker (UPV/EHU Advanced Research Facilities/ ERDF, EU).Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jece.2023.109401.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  ICCU-methanation is a promising technology that would approach a carbon neutral cycle. In this paper, the feasibility of boosting the activity of the dual function material (DFM) 10Ni-16Na/Al2O3 through the joint presence of Na/Ca and the Ru incorporation is studied. Four DFMs are prepared by sequential wetness impregnation and are extensively characterized by N2 adsorption/desorption, XRD, H2 chemisorption, TEM, STEM-HAADF and temperature-programmed techniques (H2-TPD, CO2-TPD, TPSR, and H2-TPR). The catalytic behaviour of DFMs in the cyclic process of CO2 adsorption and hydrogenation to CH4 is evaluated. The joint presence of Na/Ca improves CH4 production at intermediate-high temperatures by boosting the CO2 adsorption capacity. On the other hand, the Ru incorporation promotes CH4 production at low-intermediate temperatures by presenting synergistic aspects with nickel that lead to a greater number of exposed metal atoms. The Ru incorporation increases the metallic dispersion and the Ni reduction. Finally, the joint presence of Na/Ca and the simultaneous Ru incorporation presents the best activity results. It is concluded that both positive effects are added. Specifically, the DFM 1Ru10Ni-NaCa produces 298 \u03bcmol g\u22121 at 440\u00a0\u00b0C with a CH4 selectivity of 98.4 %. Furthermore, it is also the most active and selective DFM after hydrothermal aging in the presence of O2.\n               "}
{"full_text": "Biomass gasification is considered one of the most efficient routes to convert biomass feedstock into gaseous fuel through a partial oxidation process at high temperatures [1,2]. However, one of the main shortcomings of biomass gasification lies in the presence of tars in the product stream, which leads to fouling/clogging and corrosion of downstream equipment [3\u20135]. Hence, in order to minimize the amount of tar and improve the syngas composition, its catalytic conversion is one of most promising routes [6,7]. This process involves the oxidation of the tar components using steam to produce a syngas richer in H2 and, furthermore, the presence of the catalyst allows a more effective tar removal at lower temperatures than those in the non-catalytic tar conversion [8].The tar is as a complex mixture of condensable hydrocarbons, ranging from single-ring to five-ring aromatic compounds along with other oxygen-containing hydrocarbons and complex polycyclic aromatic hydrocarbons (PAH) [9]. Tar model compounds have been widely used in order to ascertain the catalyst performance and determine suitable operating conditions. In this work, toluene was selected as a tar model compound because it is a stable aromatic structure, especially at relatively low temperatures, apart from being one of the major tar species in the biomass gasification [10\u201314].Many parallel and consecutive reactions involve tar conversion, with the product distribution being the result of their competition. The main products obtained are hydrogen, carbon monoxide and carbon dioxide, and the major reactions occurring in the process are as follows:Toluene steam reforming\n\n(1)\nC7H8+7 H2O\u21927 CO+11 H2\n\n\n\nToluene steam dealkylation\n\n(2)\nC7H8+H2O\u2192C6H6+CO+2 H2\n\n\n\nWater gas shift (WGS):\n\n(3)\nCO+H2O\u2194CO2+H2\n\n\n\nThermal cracking\n\n(4)\nC7H8\u2192x CnHm+z H2\n\n\n\n\n\n(5)\nC6H6\u2192x CnHm+z H2\n\n\n\nHydrodealkylation\n\n(6)\nC7H8+H2\u2192C6H6+CH4\n\n\n\nToluene dry reforming\n\n(7)\nC7H8+7 CO2\u219214 CO+4 H2\n\n\n\nBoudouard reaction:\n\n(8)\nC+CO2\u21942 CO\n\n\nMethanation\n\n(9)\nCO+3 H2\u2194CH4+H2O\n\n\nCoke heterogeneous gasification:\n\n(10)\nC+H2O\u2192CO+H2\n\n\n\nAmongst these reactions, the most important ones are steam reforming (Eq. (1)) and water gas shift (WGS) (Eq. (3)).The selection of temperature and catalyst determines the extent of these reactions and the selectivity towards the different products [7,15\u201318].Tar catalytic conversion methods are classified as in-situ (or primary) and post-gasification (secondary) ones [14,19]. In the former, the tar reduction occurs during the gasification stage, with the catalyst being located in the gasifier itself. In the secondary approach, the gas produced in the gasifier is treated downstream in a secondary reactor where the catalyst is placed. Regardless the strategy followed, essential aspects conditioning the gasification process are those involving the reactor configuration, operating conditions and type of catalyst [20]. Fluidized bed reactors are one of the most developed technologies for biomass gasification, which require an appropriate catalyst in terms of activity and stability in order to reduce the tar content to 2\u00a0 g m\u22123 and avoid the need of a more expensive secondary catalytic reactor downstream [21,22].A wide range of materials with significant activity for cracking and reforming of heavy aromatic compounds have been investigated as primary catalysts [12,14]. Natural minerals, such as olivine and calcined dolomite, have been widely used in the steam gasification in fluidized beds, as they are active for tar removal, apart from being inexpensive and abundant [23]. Acid catalysts, such as alumina or zeolites, have also been used (prior to and after metal impregnation) as catalysts for tar abatement [24]. Nevertheless, the performance of all these primary catalysts can be greatly improved by metal phase addition [19,25\u201328]. Thus, support features, such as mechanical (resistance to attrition), physico\u2013chemical (surface area, porosity, acidity, composition and density) and catalytic ones (activity / selectivity and stability) play a relevant role in the metal-support interactions, as well as in the reforming reaction mechanism itself [29].From a catalytic point of view, nickel is known to be the most interesting metal phase for reforming applications [30]. Ni-based catalysts have been widely applied in the steam reforming of biomass tars due to both their high activity for breaking CC and OH bonds and performance in terms of H2 production [7,12,13,31\u201335]. However, their main drawbacks are related to their rapid deactivation, mechanical fragility and high cost compared to natural minerals or alumina [36]. Currently, use of iron as an active phase is gaining more attention for tar reduction due to its lower cost, abundance and lower toxicity compared to nickel [37]. Iron is known to be an active species for aromatic hydrocarbon destruction (breakage of CC and CH bonds), as well as for the WGS reaction. In fact, it has been proven effective for the aforementioned reactions in different oxidation states [24,38\u201340]. Therefore, iron impregnation of olivine, dolomite or Al2O3 seems to be interesting for synthetizing in-bed primary catalysts for gas\u2013solid contact reactors, such as fluidized or spouted beds, from both economic and environmental perspectives. Nevertheless, the lower activity of the iron species with respect to the Ni ones requires higher amounts of dopant, generally in the 10\u201330\u00a0wt% range [41].Accordingly, the aim of this work is to analyse the performance of olivine, dolomite and \u03b3 \u2013Al2O3 as primary catalysts, as well as the effect the impregnation of each catalyst with 10\u00a0wt% Fe has on the elimination of toluene, which has been selected as the model compound of biomass gasification tar. Furthermore, a detailed characterization of the fresh and deactivated Fe-doped catalysts has been carried out in order to determine the main deactivation mechanisms in this process. The results obtained will provide essential information for the selection of optimal primary catalysts for biomass gasification in the bench-scale unit equipped with an improved spouted bed reactor developed by our research group [21,42,43]. Furthermore, the results obtained may also be extrapolated to industrial gasification reactors, which are mainly fluidized beds. This study addresses multiple aspects that have not been jointly approached in the literature, as are catalysts preparation and characterization, influence of temperature, catalyst performance at zero time on stream, stability of Fe-loaded catalysts and the main deactivation causes.Six catalysts have been tested in the toluene steam reforming process. Three of them (olivine, dolomite and \u03b3 \u2013alumina) are primary catalysts, whereas the other three are those obtained by impregnating the aforementioned primary catalysts with Fe, i.e., Fe/olivine, Fe/dolomite and Fe/Al2O3. Besides, runs with silica sand were carried out for comparison purposes. Minerals Sibelco supplied the olivine and dolomite, and Alfa Aesar the \u03b3-Al2O3. These three primary catalysts provided satisfactory results in a previous study of biomass gasification in a fountain confined conical spouted bed reactor (CSBR), as they allowed reducing tar formation, as well improving the yield and composition of the syngas [21,42,43]. The catalyst particles were sieved in order to retain the fractions within the ranges of 90\u2013150\u00a0\u03bcm for olivine, 150\u2013250\u00a0\u03bcm for dolomite and 250\u2013400\u00a0\u03bcm for \u03b3 - Al2O3, which allow attaining similar fluidization regimes with these materials of different densities. Prior to use, dolomite was calcined at 900\u00a0\u00b0C for 4\u00a0h in a muffle oven in order to complete the decarboxylation of calcium and magnesium carbonates.The Fe loaded catalysts were prepared by wet impregnation of the supports with an aqueous solution of Fe(NO3)3\u00b79H2O (Panreac AppliChem, 98\u00a0%). The amount of saline precursor added was that corresponding to the desired final catalyst composition. The concentration of Fe was fixed at 10\u00a0wt% in order to compare the catalytic activity and selectivity of the three catalysts for same amount of metal loaded. Subsequent to the impregnation process, the prepared catalysts were dried at 100\u00a0\u00b0C for 24\u00a0h and then calcined at 1000\u00a0\u00b0C for 4\u00a0h. Given that the catalytic activity of iron species generally increases with their reduction state (Fe2O3\u00a0<\u00a0Fe3O4\u00a0<\u00a0FeO\u00a0<\u00a0Fe(0)) [26], these iron-impregnated catalysts were used once they had been subjected to an ex situ reduction process at 850\u00a0\u00b0C for 4\u00a0h under 10\u00a0vol% H2 stream, which ensured full reduction of ferric oxides into their metallic phase. The particle sizes of the Fe loaded catalysts were the same as those of their primary counterparts.The physical properties of the catalysts were determined by N2 adsorption\u2013desorption in a Micromeritics ASAP 2010 instrument. Based on the information of these isotherms, the catalysts features, such as those involving specific surface area and porous structure (average pore size and pore volume), were calculated by the Brunauer\u2013Emmett\u2013Teller (BET) method. Prior to the analysis, and in order to remove any impurity, the samples were degassed at 150\u00a0\u00b0C until a pressure below 2\u00b710\u22123 mmHg was reached. The chemical composition (wt%) of each catalyst was measured by X-ray fluorescence (XRF). More detailed information about the XRF methodology can be found elsewhere [43].The temperature-programmed reduction (TPR) of the catalysts was carried out in an AutoChem II 2920 Micromeritics, which allowed determining the catalyst reduction temperature before using it. This method consists in exposing the solid to a reducing gas flow of 10\u00a0vol% H2/Ar, while temperature is increased with a constant heating rate of 5\u00a0\u00b0C\u00a0min\u22121 from ambient one to 900\u00a0\u00b0C. The reduction temperature of each catalyst was ascertained by monitoring the H2 consumed.The crystalline structure of the fresh and deactivated catalysts was analyzed using X-ray powder diffraction (XRD) patterns. A Bruker D8 Advance diffractometer with Cu K\u03b11 radiation was used to conduct XRD. The detailed procedure followed is described elsewhere [44]. The metal crystallite size was calculated by using the Scherrer formula. Metal dispersion was calculated from metal crystallite size using the equation D (%)\u00a0=\u00a097.1/d (nm) and assuming that the size of Fe atom is the same as that of Ni atom, as reported elsewhere [45,46].The values of total acidity of the catalysts have been obtained by monitoring the differential adsorption of NH3 at 150\u00a0\u00b0C using simultaneously calorimetry and thermogravimetry in a Setaram TG-DSC 111 equipment.The amount of coke deposited on the used catalysts was determined by temperature-programmed oxidation (TPO) in a thermobalance (TGA Q5000TA Thermo Scientific). This TGA is connected on-line to a Blazer Instruments Mass Spectrometer (Thermostar) and the procedure followed to determine the coke deposited on each sample is as follows: (i) signal stabilization with He stream at 100\u00a0\u00b0C for 30\u00a0min, and (ii) a ramp of 5\u00a0\u00b0C\u00a0min\u22121 to 800\u00a0\u00b0C in a stream of O2 diluted in He, with this temperature remaining constant for 30\u00a0min in order to ensure full coke combustion.The experiments of toluene conversion with the different catalysts were performed in an Inconel fluidized bed reactor (300\u00a0mm in length and 10\u00a0mm in internal diameter), as shown in Fig. 1\n. The reactor is located within a radiant oven, which provides the heat for operating up to 900\u00a0\u00b0C. The temperature was measured and recorded by means of two K-type thermocouples, with one being located inside the reactor, approximately in the middle zone of the bed, and the other one close to the wall of the electric oven.The water for generating the steam and toluene were introduced by means of a high-pressure pump (ASI 521) and a syringe pump (PHD 4400), respectively. Their pumping flowrates were maintained constant in all the runs, with the values being 0.24\u00a0mL\u00a0min\u22121 for water and 0.06\u00a0mL\u00a0min\u22121 for toluene, which correspond to a steam/toluene ratio (S/T) of 4 and a molar steam/carbon (S/C) ratio of 3.35. Prior to feeding into the reactor, these two compounds were pumped separately into an evaporation system at 350\u00a0\u00b0C, which ensures their full vaporization. This plant is also provided with a nitrogen mass flow meter (Brooks SLA5800) that allows feeding up to 1 L min\u22121. In fact, a nitrogen flow rate of 300\u00a0mL\u00a0min\u22121 was used as fluidizing agent during the heating process prior to the reaction.The gaseous stream leaving the reactor was passed through a heater, whose temperature was kept at 300\u00a0\u00b0C in order to prevent the condensation of the products before entering the on-line analysis system. Then, the volatile stream circulated through a condensation device consisting of two coalescence filters, which ensured total condensation and retention of the non-reacted steam and toluene, as well as toluene derived products.This study deals with the effect of operating conditions in a catalytic process for tar elimination process, i.e., reforming temperature (in the 800\u2013900\u00a0\u00b0C range), catalysts type (olivine, dolomite and alumina, as well as their counterparts with Fe impregnation) and catalyst stability. Olivine was chosen to analyse the effect of temperature, whereas 850\u00a0\u00b0C was established as the suitable operating temperature to study the influence of catalyst type and time on stream. The effect of reaction time was studied for the Fe loaded catalysts in the 5\u2013115\u00a0min range in order to assess the evolution of catalyst activity and stability.Given that the density of the primary catalysts differs significantly (3300\u00a0kg\u00a0m\u22123 for olivine, 1666\u00a0kg\u00a0m\u22123 for alumina and 1275\u00a0kg\u00a0m\u22123 for dolomite), and in order to operate under the same hydrodynamic conditions in the fluidized bed reactor, the same bed volume was used in all experiments. Accordingly, as mentioned above, suitable particle sizes were chosen. Thus, 3.8\u00a0cm3 of the corresponding catalyst (or sand in case the experiment was carried out without catalyst) were placed in the bed in all cases, corresponding to a gas hourly space velocity (GHSV) of 820\u00a0h\u22121. Experiments at zero time on stream were repeated at least 3 times to ensure reproducibility of the results and the carbon mass balance closure was above 95\u00a0% in all runs.The analysis of the volatile stream leaving the reactor was conducted on-line by means of a GC (Agilent 7890) provided with a flame ionization detector (FID). The sample was injected into the GC prior to condensation by means of a line maintained at 280\u00a0\u00b0C in order to avoid the condensation of heavy tar compounds. The analysis of the non-condensable gases (after separating the tars from the gaseous stream in the condensation system) was carried out by means of a micro GC (Agilent 4900). The three independent modules with different columns (molecular sieve, porapak and plot alumina) allowed identifying and quantifying the gaseous products previously calibrated. This analysis methodology allowed a detailed quantification of the entire product stream.The conversion and product yields were taken as reaction indices to monitor and assess process performance. The carbon conversion of toluene was defined as the moles of carbon in the gaseous product stream divided by the moles of carbon in the toluene feed (Eq. (11)). Note that the moles of CO, CO2 and C1-C4 hydrocarbons formed (corresponding to the total amount of carbon moles in the gas) have been determined from the micro-GC analysis, whereas the moles of carbon in the feed were calculated based on the total amount of toluene introduced into the reactor (total volume of toluene injected in the run).\n\n(11)\n\n\n\n\nC\n\n\nconversion\n\n\n\n\n%\n\n\n=\n\n\nmoles\n\nof\n\ncarbon\n\nin\n\nthe\n\nproduct\n\ngas\n\n\nmoles\n\nof\n\ncarbon\n\nin\n\nthe\n\nfeed\n\n\n\u00b7\n100\n\n\n\n\nThe product yields were calculated as the ratio between the grams of each product (H2, CO, CO2 and CH4) in the gaseous stream and the grams of the model compound in the feed:\n\n(12)\n\n\nY\ni\ne\nl\nd\n\n\n\nw\nt\n%\n\n\n\n=\n\n\ng\n\nof\n\nthe\n\ncompound\n\nin\n\nthe\n\nproduct\n\ngas\n\n\ng\n\nof\n\nmodel\n\ncompound\n\nin\n\nthe\n\nfeed\n\n\n\u00b7\n100\n\n\n\n\nMoreover, H2 potential was also determined as the ratio between the concentration of H2 in the effluent gas and the maximum allowed by stoichiometry:\n\n(13)\n\n\n\n\nH\n\n\n2\n\n\n\np\no\nt\ne\nn\nt\ni\na\nl\n\n\n%\n\n\n=\n\n\nmoles\n\nof\n\n\n\nH\n\n\n2\n\n\n\ni\nn\n\nt\nh\ne\n\np\nr\no\nd\nu\nc\nt\n\ng\na\ns\n\n\nmaximum\n\nmoles\n\nof\n\n\n\nH\n\n\n2\n\n\n\na\nl\nl\no\nw\ne\nd\n\nb\ny\n\ns\nt\no\ni\nc\nh\ni\no\nm\ne\nt\nr\ny\n\n\n\n\n\n\nThe maximum number of H2 moles allowed by stoichiometry was calculated by considering toluene reforming reaction and that of WGS. Thus, H2 potential is defined based on the maximum number of H2 moles obtained when toluene is fully reformed to CO2 and H2.\nTable 1\n shows the physical properties (specific surface area, pore volume and average pore diameter) and chemical composition of the primary catalysts and those impregnated with Fe. As observed, olivine has the lowest specific surface area (1.92\u00a0m2 g\u22121) and pore volume (0.002\u00a0cm3 g\u22121) due to its non-porous structure. After impregnation with Fe(NO3)3\u00b79H2O solution, the specific surface area and the pore volume of dolomite and Al2O3 decreased mainly due to metal deposition, as it blocks some of the micropores of the catalysts. According to Kumar et al. [47], the presence of iron on alumina accelerates the shrinkage of alumina and transforms the alumina from gamma into other phases, which decreases the surface area because Fe2O3 particles act as heterogeneous nucleation sites for \u03b1-Al2O3 particles at high temperature. Nevertheless, the opposite trend was observed in the Fe/olivine, i.e., the specific surface area increased due to the deposition of Fe on the external surface area, and the pore volume and average pore size became larger, which suggests the collapse of the inter-pore structure of olivine. Note that the same trend has been observed for metal impregnation on supports with low porosity surfaces [22,48,49]. Apart from the impregnation process, the high calcination temperature also contributes to reducing the BET surface area and porosity of the Fe/ Al2O3 catalyst, although to a lesser extent. In a previous study [50], the same Al2O3 used in this study was calcined with air at 1000\u00a0\u00b0C during 5\u00a0h and its BET surface area and pore volume reduced to 87\u00a0m2 g\u22121 and 0.38\u00a0cm3 g\u22121, respectively.Dolomite is a calcium magnesium carbonate, i.e., CaMg(CO3)2, and therefore carbonates are decomposed into CaO and MgO in the calcination, which are the main constituents in the calcined dolomite, as shown in Table 1. Moreover, the XRF revealed that there is a high content of Fe in the Fe/olivine. In fact, the content of Fe in the raw olivine was of around 5.3\u00a0wt% and after impregnation, the Fe amount in the catalyst increased significantly to 17\u00a0wt%, which confirmed that the metal content was close to that corresponding to the impregnation (\u223c10\u00a0wt%) plus that in the original olivine. In the other two catalysts, namely Fe/Al2O3 and Fe/dolomite, the initial Fe content was negligible and after the impregnation increased up to 9.9 and 9.3\u00a0wt%, respectively. Thus, the Fe content of the three studied catalysts is consistent with the targeted metal loading of 10\u00a0wt%.\nTable 2\n shows the metal dispersion of each catalyst which was estimated based on the metal crystallite size obtained by XRD analysis (by applying Debye-Scherrer equation). As observed, the highest metal dispersion is attained for Fe/Al2O3 (2.6\u00a0%), whereas the poorest value is for dolomite (0.5\u00a0%). This result confirms that the physical structure of the support plays an essential role in the dispersion of the metal phase; that is, the support with the highest surface area as that of Al2O3 leads to the highest metal dispersion.The XRD patterns of the primary catalysts and Fe reduced catalysts are shown in Fig. 2\na and 2b, respectively. As observed, the three Fe doped catalysts show an intense peak of the metal iron phase at 2\u03b8\u00a0=\u00a044\u00b0 and two smaller ones at 2\u03b8\u00a0=\u00a065\u00b0 and 82\u00b0. Note that iron oxide phases were not detected in these catalysts, which is evidence of their full reduction. In both olivine (Fig. 2a) and Fe/olivine (Fig. 2b), the main crystalline phases observed are those corresponding to olivine (Mg1.81Fe0.19\u00b7(SiO4)) and enstatite (MgSiO3). Further diffractogram of unreduced Fe/olivine catalyst can be found elsewhere [22]. Regarding Fe/dolomite (Fig. 2b), apart from the metal iron phase, those of Ca(OH)2, CaO and MgO were also observed, with all of them being derived from the calcination of calcium magnesium carbonate, which is the main mineral species in the dolomite [51]. These last three phases (Ca(OH)2, CaO and MgO) were also observed in the XRD diffractogram of calcined dolomite (Fig. 2a). These alkaline earth oxides (CaO and MgO) containing Lewis basic sites may promote adsorption and migration of H2O and OH groups on the catalyst surface, and therefore promote carbon gasification and reduce carbon deposition [52]. The Ca(OH)2 diffraction peaks are evidence that CaO (a highly hygroscopic compound) absorbed humidity from the ambient and formed Ca(OH)2. In the Fe/Al2O3 catalyst, typical diffraction peaks corresponding to the Al2O3 support were detected, as well as hercynite (FeAl2O4), whose diffraction lines are located at 2\u03b8\u00a0=\u00a031\u00b0, 36\u00b0, 51\u00b0, 59\u00b0 and 64\u00b0. The high calcination temperature used (1000\u00a0\u00b0C) allowed the formation of hercynite spinel (FeAl2O4), which occurs at temperatures above 600\u00a0\u00b0C by the interaction between Fe species (Fe0, FeO and Fe3O4) and Al2O3, following the reaction mechanism reported in the literature [53,54]. Moreover, comparing the XRD patterns of Al2O3 before and after impregnation and calcination stages, there is a phase change from \u03b3-Al2O3 to a more stable one, which is probably the most stable one (\u03b1-phase) due to the high temperature of calcination used (1000\u00a0\u00b0C). The peaks assigned to Al2O3 in the Fe loaded catalyst in Fig. 2b are clear and sharp, which is evidence of its high crystallization degree, whereas the peaks assigned to Al2O3 in Fig. 2a are broad and low, thereby suggesting an amorphous structure with a small crystallization degree. Note that the same diffraction peaks than those observed for Al2O3 crystalline phases in Fig. 2a and b have been reported in the literature and correspond to \u03b3-Al2O3 and \u03b1-Al2O3, respectively [55,56]. Therefore, phase transformation is the consequence of the thermal degradation of the support, which affects adversely the physical properties of the catalyst by reducing catalyst surface area, thereby reducing catalyst activity. Several authors have called this process support sintering [35,57].The temperature programmed reduction (TPR) profiles of calcined Fe/olivine, Fe/dolomite and Fe/Al2O3 catalysts are shown in Fig. 3\n. Given that metal iron is expected to be the active phase for breaking CC and CH bonds [24,58], the reducibility of the catalysts is of great relevance. According to the literature [26,59] the reduction of Fe2O3 generally proceeds in two steps, as are: the reduction of Fe2O3 to Fe3O4 in the 350\u2013500\u00a0\u00b0C range and the reduction of Fe3O4 to metal Fe in the 500\u2013900\u00a0\u00b0C range. However, according to certain studies, the intermediate FeO is formed in the reduction from Fe3O4 to Fe0 [60,61]. These two regions associated with two or three reduction steps from Fe2O3 are observed in the three reduced catalysts, although differences in the interactions between the iron and the supports shifted the location of the peaks. In the TPR profile of the Fe/olivine, a broad reduction zone between 350 and 700\u00a0\u00b0C is observed with 3 peaks. The first two (at 470 and 530\u00a0\u00b0C) are associated with the reduction of Fe2O3 and Fe3O4/FeO, respectively, whereas the latter peak above 600\u00a0\u00b0C is due to the Fe atoms that migrated into the olivine support to form a very stable MgFe2O4 spinel phase [62]. Peaks at 380\u00a0\u00b0C and 500\u00a0\u00b0C appear in the Fe/dolomite, which are characteristic of iron species reduction, but there is also a broad peak at 750\u00a0\u00b0C, which corresponds to the reduction of Fe3+ from the calcium iron oxide (srebrodolskite, Ca2Fe2O5) to Fe, as was suggested by Zamboni et al. [63,64]. These authors observed the formation of this phase when iron nitrate was used in the wet impregnation of dolomite. In this study, no evidences of Ca2Fe2O5 are observed in the XRD diffractogram (Fig. 2), probably due to its low crystallinity. In the Fe/Al2O3 catalyst, apart from the two peaks identified at 380 and 580\u00a0\u00b0C, which are associated with the reduction of iron species (Fe2O3, Fe3O4 and FeO) , a third broad reduction zone appears between 700 and 900\u00a0\u00b0C, which is attributed to the reduction of iron aluminates (FeAl2O4), also identified in the XRD spectra [65]. Different authors suggested that the presence of alumina stabilizes Fe2O3 phase and the reduction goes through the formation of FeAl2O4 spinel, whose reduction occurs above 700\u00a0\u00b0C [66,67].The influence of temperature on toluene abatement on olivine catalysts is displayed in Fig. 4\n. Fig. 4a shows the evolution of carbon conversion and H2 potential. As observed, temperature has a great influence on carbon conversion and H2 potential, since their values increase from 3.6 and 2.6\u00a0% at 800\u00a0\u00b0C to 46.0 and 23.6\u00a0% at 900\u00a0\u00b0C, respectively. This increase in both parameters is attributed to the endothermic nature of the toluene reforming reactions, as well as of those involving decomposition and dehydrogenation, as all of them are promoted at high temperatures [68]. The same trend of carbon conversion and H2 potential with temperature on olivine catalysts was observed by other authors in the tar steam reforming [58,69].The yields of the compounds in the product stream is displayed in Fig. 4b. An increase in temperature leads to higher yields in gaseous compounds (including benzene) due to the promotion of both reforming and cracking reactions, with the highest yields being those of CO and CO2 at 900\u00a0\u00b0C (49.8 and 26.8\u00a0wt%, respectively). The yield of CH4 increases with temperature, but it is lower than 2.2\u00a0wt% at the three temperatures studied. It should be noted that the yield of C\n2-C4 hydrocarbons is hardly noticeable (below 0.01\u00a0wt%), and has not therefore been included in Fig. 4b. CH4 is mainly formed from dealkylation of the methyl group in the toluene structure and, to a minor extent, from the methanation of CO [8]. However, steam reforming of CH4 prevails over these reactions, since the content of CH4 in the products is very low [12]. The presence of an undesired compound (benzene) is due to incomplete decomposition of toluene [70], which is confirmed in Fig. 4b, where benzene yield increases from 0.6\u00a0wt% at 800\u00a0\u00b0C to 12.6\u00a0wt% at 900\u00a0\u00b0C at the expense of a decrease in toluene yield. Several reactions, such as steam dealkylation (Eq. 2), thermal cracking (Eqs. 4\u20135) or hydrodealkylation of toluene (Eq. 6) lead to the formation of benzene (all of them enhanced at high temperatures) [11,15,17,71]. However, the small amount of CH4 in the product stream is evidence that hydrodealkylation reaction (Eq. 6) is not significant [72]. It should be noted that the benzene produced from the aforementioned reactions can undergo reforming reactions to produce further CO and H2, although these reactions are limited due to benzene stability [11,12]. The yield of polycyclic aromatic hydrocarbons (PAHs, referred to the compounds heavier than toluene) also increases with temperature due to the promotion of condensation reactions of lighter tars. However, the low yield of these PAHs (below 1.2\u00a0wt% in the whole range of temperatures studied) is evidence that the extent of these reactions is almost negligible, probably due to the presence of steam [21,73]. Swierczynski et al. [3] also observed a yield of around 6\u00a0wt% of benzene and 14\u00a0wt% of polyaromatics in the product stream of toluene steam reforming at 850\u00a0\u00b0C when they used olivine as primary catalyst.\nFig. 4c displays the gas composition in the 800\u2013900\u00a0\u00b0C range. It can be observed that the effect of temperature on the gas composition is not very pronounce above 850\u00a0\u00b0C, i.e., the concentration hardly changes above this temperature. Between 800 and 850\u00a0\u00b0C, certain trends are observed when temperature is increased, as are: a slight decrease in H2 and CO2 concentrations (from 69.1 to 66.2\u00a0vol% and from 8.1 to 6.8\u00a0vol%, respectively) and an increase in that of CO (from 21.8 to 25.5\u00a0vol%). This result is explained by the promotion of the reverse WGS reaction due to its exothermic nature. The same trend with temperature was observed in other studies of catalytic reforming of tar model compounds, with this effect being attributing to the exothermic nature of the WGS reaction [68,74].In order to study the performance of primary catalysts, toluene conversion on olivine, dolomite and alumina was monitored at 850\u00a0\u00b0C and the results obtained are displayed in Fig. 5\n. The effect of thermal cracking was ascertained by comparing the results of carbon conversion (Fig. 5a), product yields (Fig. 5b) and concentration of gaseous compounds (Fig. 5c) obtained with the catalysts and those obtained with inert sand. As observed, the presence of any catalyst improves the overall efficiency of the process by increasing carbon conversion and the yields of gaseous compounds, especially those of H2, CO and CO2, as well as reducing that of toluene. This improvement over the results obtained with inert sand is associated with the promotion of steam reforming (Eq. 1), cracking (Eqs. 4\u20135) and WGS reactions (Eq. 3). The presence of primary catalysts also promotes steam dealkylation (Eq. 2) and thermal cracking (Eq. 4) reactions, since the concentration of benzene in the product stream is higher than that obtained with sand.Comparing the efficiency of the primary catalysts (Fig. 5a), Al2O3 leads to the highest conversion (58.4\u00a0%) followed by dolomite (39.1\u00a0%). However, the H2 potential with both catalysts is similar (28.5\u00a0% for Al2O3 and 28.9\u00a0% for dolomite). This latter result can be explained by the lower activity of Al2O3 and the higher of dolomite in the WGS. Thus, the higher activity of dolomite in the WGS reaction is related to CaO and MgO basic sites, with activity being higher as the Ca/Mg ratio is increased [75,76]. Furthermore, the presence of CaO and MgO also explains the higher yield of benzene at the expense of lowering that of toluene [77,78]. Moreover, olivine has the smallest influence on the toluene steam reforming, since it provided the lowest carbon conversion and H2 potential values. In this case, although the presence of Fe promotes reforming reactions, the low BET surface area (1.91\u00a0m2 g\u22121) and pore volume (0.002\u00a0g cm\u22123) are the factors leading to the low efficiency of this catalyst in the toluene elimination process. Studies reported in the literature confirm that dolomite and Al2O3 were more active than olivine for reducing the amount of tar derived from biomass gasification, as the extent of the WGS reaction is enhanced with dolomite [43,79].\nFig. 6\n compares the parameters involving toluene conversion (carbon conversion and H2 potential (a), product yields in the outlet stream (b) and the concentration of gaseous compounds (c)) for the Fe loaded catalysts. Fig. 6a reveals that Fe incorporation into the primary catalysts leads to higher carbon conversion and H2 potential than those on the primary catalysts in all cases, Fig. 5a, which is evidence of their higher catalytic activity for toluene reforming. Thus, on the one hand, it is well stablished that metal iron is active for CC and CH bond breakdown, which enhances hydrocarbon reforming and cracking reactions [58,80]. On the other, the addition of Fe promotes the WGS reaction because the adsorption of water molecules on the catalyst active sites is favoured, thus leading to higher H2 yields [81]. This improvement is especially remarkable with olivine, whose carbon conversion and H2 potential increases from 18 and 10.5\u00a0% to 73 and 31.9\u00a0%, respectively.As occurred with primary catalysts, that of Fe/Al2O3 provided the best results in terms of carbon conversion (87.6\u00a0%) and H2 potential (38\u00a0%) (Fig. 6a). However, the trends were reversed for Fe/olivine and Fe/dolomite after Fe incorporation, attaining higher carbon conversion in the former. This result is closely related to the change in the surface area of the catalysts caused by the impregnation, which definitely affects metal dispersion. As observed in Table 1, the surface area increased in the olivine when Fe was introduced, whereas it significantly decreased in the dolomite (from 17.42 to 3.55\u00a0m2 g\u22121). Furthermore, the results in Table 2 confirm the better dispersion of Fe on the olivine than on the Fe/dolomite, which suggests that the active sites are more accessible for the reactants in the former, as all the iron is located on the catalyst surface. This implies a higher catalytic activity of Fe/olivine, which explains the better results of carbon conversion on this catalyst than on Fe/dolomite, whose metal dispersion is the poorest.Comparing the product yields shown in Fig. 6b, the highest yields of CO and CO2 (mainly derived from the reforming and WGS reactions, respectively) and benzene (a cracking product) are obtained on the Fe/Al2O3 catalyst, whereas that of toluene is the lowest (below half of those obtained on Fe/olivine or Fe/dolomite). Note that Fe acts as the active phase for the reforming and WGS reactions, whereas the alumina support provides the acidity required for cracking reactions, i.e., the combination of both provides Fe/Al2O3 catalyst with the highest activity for these reactions. Moreover, a comparison of Fig. 6b with Fig. 5b shows that the yield of benzene increases greatly when Fe is added to the primary catalysts. It seems that the presence of Fe mainly catalyzed the conversion of toluene to benzene. Some studies suggested that temperatures higher than 800\u00a0\u00b0C increase the hydrodealkylation activity for the steam reforming of toluene on iron-based materials [72,82], whereas other researches concluded that the activity of iron-based materials leads to the decomposition of large tar compounds into small fragments of carbon species, which subsequently form benzene [83]. Therefore, it can be concluded that the higher benzene content is a combined effect of cracking and hydrodealkylation of toluene molecules on Fe active sites. The higher CH4 yields observed on Fe loaded catalyst than on primary catalysts also confirms this hypothesis.\nFig. 6c displays the gas composition obtained with the three Fe-impregnated catalysts. A comparison of these results with those for primary catalysts (Fig. 5c) shows the relevance of metal iron in the WGS reaction (Eq. 3), since the concentration of CO2 greatly increased in all the cases, whereas that of CO reduced. This is consistent with previous studies in the literature, in which a high activity of Fe is reported in the WGS reaction [38,40,84]. Analysing Fe loaded catalysts, Fe/Al2O3 led to the lowest concentration of CH4 and CO2 and the highest of H2 and CO, which is evidence of a high extent of steam and dry reforming of hydrocarbons (Eq. 1 and 7). According to Adnan et al. [85,86], this fact is attributed to the basic sites of Fe/Al2O3 catalysts, which promote endothermic CO2 reforming of hydrocarbons.The differences observed among these Fe-impregnated catalysts are the consequence of various factors. As previously stated, one of the most influential factor is related to the metal dispersion on the catalyst support, which plays a key role in the initial catalyst activity. A suitable metal-support interaction enhances the migration of metal crystallites, thereby obtaining a better dispersion of Fe on the support [26]. Furthermore, the physical structure of the support greatly influences the dispersion of the metal phase, as shown in Table 2, in which the highest Fe dispersion was obtained for Al2O3 (the support with the highest BET surface area and pore volume). The results in Fig. 6 confirm that the better surface properties of the Al2O3 support promote the dispersion of the active phase, and therefore lead to higher catalyst activity. Besides, the higher dispersion of Fe on olivine also explains the higher carbon conversion than on Fe/dolomite.Another factor is related to the activity of the support for cracking and/or reforming reactions, which is directly linked to its acidity [30,57]. Thus, the porous structure of olivine and dolomite barely have micro or mesopores, whereas alumina has a more developed porous structure, as shown in Table 1. Adnan et al. [85] suggested that toluene conversion reactivity is dominated by strong acid sites in the catalyst, which are directly attached to the surface of the catalyst. Thus, a higher surface area of the catalyst increases the number of strong sites available to contact with toluene, thereby leading to a higher acidity of the catalysts, and consequently to a higher conversion of toluene, as is the case of Fe/Al2O3, which has the highest acidity (Table 1) of the three Fe loaded catalysts [87]. Besides, Adnan et al. [88] stated that a higher content of Fe in the catalyst also promotes catalyst acidity, and therefore toluene conversion. Comparing the acidity of primary and Fe doped catalysts (Table 1), the presence of Fe increases the acidity of Fe/olivine and Fe/dolomite catalysts from 2.4 to 8.8 and from 8.7 to 10.5\u00a0\u00b5mol NH3 gcat\n\u22121, respectively, which explains the higher toluene cracking capability of Fe doped ones. Regarding the acidity value of Fe/Al2O3 (11.4\u00a0\u00b5mol NH3 gcat\n\n\u22121\n), it is much lower than that of the raw \u03b3-Al2O3. Indeed, as previously stated, the reduction in BET surface area caused by the calcination and impregnation stages leads to the blockage of some pores and reduces the number of acid sites available, thus reducing the total acidity of the catalyst. However, comparing Fig. 5b and 6b, benzene yield is higher when Fe/Al2O3 is used than when the primary Al2O3 is used, which suggests that the cracking activity of Fe/Al2O3 is higher. This result is explained by the combination of two issues. On the one hand, as was previously stated, the better performance of Fe for reforming and WGS reactions leads to higher H2 partial pressures in the reaction environment, thus promoting hydrodealkylation reactions (Eq. 6) which lead to higher benzene contents. On the other hand, the real acidity of \u03b3-Al2O3 under reaction conditions is much lower than that given in Table 1, as the high temperatures used in this study (850\u00a0\u00b0C) and the presence of steam accelerate the collapse of the porous structure and the transformation of \u03b3-alumina into other more stable phases, as stated elsewhere [47]. Thus, the blockage of pores and the transformation of \u03b3-phase into other ones (\u03b4, \u03b8 or \u03b1) reduces the number of acid sites available, and therefore its cracking activity.Other important issue involving catalytic activity is the reduction state of the iron species, with activity being higher as Fe species are further reduced (metal Fe is the most active phase). Thus, the XRD patterns in the three fresh catalysts reveal the presence of metal Fe, whereas the presence of other species with different reduction states, such as Fe2O3, Fe3O4 or FeO, was not initially observed (Fig. 2b). This is an evidence that the difference in the catalytic activities of Fe impregnated catalysts is mostly attributed to the interactions between the metal iron and the supports, as well as their physical structure. Thus, the better properties of Al2O3 (it acts as a textural promoter preventing the fast sintering of the iron metal, as well as stabilizing active sites on its surface) lead to better dispersion of the Fe oxide phase, and therefore better performance for toluene steam reforming and WGS reaction [38].Claude et al. [26] analysed the effect of Fe doped olivine and alumina catalysts and revealed that Fe/Al2O3 provided also higher toluene conversion than the Fe/olivine catalyst when temperature was 850\u00a0\u00b0C. Nevertheless, the olivine catalyst was the one of better performance at 750\u00a0\u00b0C. It should be noted that these authors reduced both catalysts in situ in the reactor with an inlet reactant gas mixture containing 31.5\u00a0%vol. H2, which simulates a feed containing a fraction of the reforming outlet stream. Therefore, the differences at this temperature can be explained by the presence of iron species with different reduction states (active for steam reforming) in the olivine, whereas Fe was only present as hercynit in the case of alumina.\nFig. 7\n displays the evolution of carbon conversion and H2 potential with time on stream in the toluene steam reforming at 850\u00a0\u00b0C on the three catalyst tested. As observed, Fe/Al2O3 provided the highest stability in terms of carbon conversion (Fig. 7a) and H2 potential (Fig. 7b), since it allowed operating for the longest period with the highest conversion (85.9\u00a0% after 35\u00a0min on stream). Fe/dolomite and Fe/olivine catalysts provided a rather stable activity for the first 15\u00a0min, but the deactivation rate increased greatly subsequent to this time, and therefore toluene conversion and H2 potential decreased rapidly. The decrease in these parameters in the range from 15 to 25\u00a0min is more pronounce on the Fe/dolomite catalyst (35.1\u00a0% and 18.0\u00a0%, respectively, at the end of 25\u00a0min on stream). Subsequent to this time, the Fe/olivine catalyst underwent more severe decrease to 31.0\u00a0% and 15.8\u00a0%, respectively, after 45\u00a0min on stream. Overall, the conversion level and H2 yield decreased gradually with reaction time when either catalyst was used, reaching similar steady values of around 30\u00a0% and 15\u00a0%, respectively, which is evidence of the deactivation underwent by the catalysts.The evolution of component yields in the product stream with reaction time is shown in Fig. 8\n for the three Fe doped catalysts. As observed, the yields of toluene and benzene follow opposite trends in the three catalysts. Once the activity for toluene reforming and cracking is low due to catalysts deactivation, the yield of toluene increases, whereas that of benzene decreases. Given the higher activity and stability on the Fe/Al2O3 catalyst for a longer time, benzene yield is higher and remains at around 45\u00a0wt% for a longer period than on Fe/olivine and Fe/dolomite. However, Fe/Al2O3 catalyst deactivation is more pronounced, attaining yields of H2 and CO lower than 10\u00a0wt%. Thus, Fe/dolomite provided lowest yields of toluene and highest of benzene when it was deactivated (65.2 and 14.6\u00a0wt%, respectively, for 65\u00a0min on stream), which means it is more active for toluene cracking than the other ones subsequent to this time. As previously stated, dolomite is well-known as an active catalyst for tar conversion when it is in the calcined state, i.e., CaO and MgO state, and therefore these species are still active phases for toluene cracking subsequent to the mentioned time [89].Moreover, as time on stream increased, H2 and CO2 yield decreased for the three catalysts, which is evidence of the lower extension of reforming and WGS reactions when the catalysts undergo deactivation. This reduction is more pronounced in the Fe/Al2O3 catalyst, in which the H2 yield decreased steadily from 17.5 to 7.3\u00a0wt% and that of CO2 from 105.0 to 36.0\u00a0wt% for 115\u00a0min on stream. These results and those corresponding to benzene and toluene yields confirm that, although the Fe/Al2O3 catalyst was able to maintain its reforming/cracking capacity for longer time, the deactivation was more severe.Regarding the CO yield, it was almost constant for Fe/Al2O3, whereas it decreased slightly for Fe/olivine and Fe/dolomite (from 13.7 to 6.9\u00a0wt% on both catalysts), although the latter allows operation for longer time on stream until reaching this final yield. This trend is a consequence of a balance between the attenuation of reforming (Eqs. 1 and 7) and WGS (Eq. 3) reactions and CO formation by mainly decarbonylation (cracking) [90,91].Regarding CH4 yields, they decreased slightly as reaction proceeded, attaining a value of around 1.25\u00a0wt% in the three catalysts. These results are evidence of the attenuation of the hydrodealkylation reaction (Eq. 6) when deactivation proceeded, although CH4 may still be formed by the cracking of the hydrocarbons in the reaction environment or by methanation (Eq. 9). A similar explanation holds for the slight increase in the yield of heavier hydrocarbons with time on stream. In this case, the attenuation of hydrocarbon reforming reactions by catalyst deactivation enables hydrocarbon rearrangement reactions, such as polymerization and/or cycloaddition, which lead to higher molecular weight species than toluene [21,92].The faster deactivation of olivine and dolomite catalysts shown in Figs. 7 and 8\n suggests that the role of metal-support interactions, as well as the structural characteristics of the supports, in the metal dispersion may be relevant in the catalyst deactivation mechanism [37]. In fact, the poorer metal dispersion on dolomite and olivine catalysts, and therefore the lower amount of Fe active sites available for reforming and WGS reactions, enhances catalyst deactivation, either by sintering, coke deposition or iron phase change (reduction in metal active sites by the oxidation of iron species). As reported by other researchers, iron is more active for tar cracking/reforming when it is in the metal state than oxidized, but the oxidizing nature of steam at high temperatures promotes the oxidation of Fe metal sites [37,83]. Furthermore, given the lower dispersion of Fe on Fe/olivine and Fe/dolomite catalysts, most of it will be deposited on the catalyst surface, which leads to faster coke deposition, and therefore faster deactivation [24]. In fact, the deactivation mechanism by coking for toluene steam reforming is well established in the literature on Fe based catalysts [58,88]. The deactivation mechanism for the three catalysts will be further discussed in the next section.Prevention and attenuation of catalyst deactivation is essential for improving the viability of this catalytic process at larger scale. Therefore, a detailed characterization of the deactivated catalysts was carried out in order to understand the main causes of catalysts activity decay. Based on the results obtained in this study and others reported in the literature, the main factors causing the deactivation of Fe impregnated catalysts are coke deposition and active phase oxidation. Nevertheless, sintering or iron loss by attrition may also be relevant [22,25,26,62].\nTable 3\n shows the textural properties of deactivated Fe/olivine, Fe/dolomite and Fe/Al2O3 catalysts once they were used for 115\u00a0min on stream. Comparing these properties with those displayed in Table 1 for fresh catalysts, the BET surface areas remained almost constant for Fe/olivine and Fe/dolomite catalysts, whereas the Fe/Al2O3 underwent a reduction from 12.48 in the fresh one to 7.63\u00a0m2 g\u22121 in the deactivated one. Pore volume and pore diameter of the Fe/olivine catalyst decreased from 0.017\u00a0cm3 g\u22121 and 234\u00a0\u00c5 to 0.010\u00a0cm3 g\u22121 and 217\u00a0\u00c5, respectively, thus revealing a partial blockage of the catalyst pores, but not their complete clogging. Moreover, the pore volume also decreased in the deactivated Fe/Al2O3, but the pore diameter increased from 206 to 285\u00a0\u00c5, which suggests that the smallest pores are fully blocked by coke deposition. Similarly, the pore diameter increased in the spent Fe/dolomite, which reveals blockage or partial obstruction of the smallest pores.\nTable 3 also displays the chemical composition of the deactivated catalysts. A comparison of these values with those of the fresh ones (Table 1) allows concluding that there is not significant iron loss by attrition phenomena. Thus, the iron oxide content decreased slightly in the deactivated catalysts, i.e., 3.15\u00a0wt% in the Fe/olivine, 0.49\u00a0wt% in the Fe/Al2O3 and 1.48\u00a0wt% in the Fe/dolomite. The absence of attrition phenomena was also checked by sieving the deactivated catalysts particles. Thus, their size is approximately the same as the fresh ones, i.e., 90\u2013150\u00a0\u03bcm for olivine, 150\u2013250\u00a0\u03bcm for dolomite and 250\u2013400\u00a0\u03bcm for alumina. The higher loss of Fe in the Fe/olivine and Fe/dolomite catalysts is related to the weaker interaction between the Fe and the support, as the metal species is mainly located on the surface [26]. Overall, these results of XRF analysis reveal that catalyst deactivation is not caused by the loss of metal phase.\nFig. 9 shows the XRD patterns of the spent catalyst, which allow assessing the influence of the active phase oxidation in the deactivation process. As observed, the diffraction lines attributed to metal iron (2\u03b8\u00a0=\u00a045\u00b0 and 2\u03b8\u00a0=\u00a065\u00b0) are only present in the Fe/dolomite sample, although their intensity is greatly reduced compared to those of the fresh one (Fig. 2). The XRD profiles for neither Fe/olivine nor Fe/Al2O3 contain these lines, which explains the slightly higher activity of Fe/dolomite for toluene reforming and WGS at longer times on stream. Nevertheless, lines for other iron phases have been detected in the three catalysts, with some of them being absent in the diffractograms for the fresh ones. The X-ray pattern of the deactivated Fe/Al2O3 only shows the presence of iron strongly incorporated into alumina (hercynite, FeAl2O4). A new phase of calcium iron oxide (Ca2Fe2O5) appears in the Fe/dolomite, which is formed due to the interaction between the metal iron and calcium oxide in the presence of steam [63]. The presence of Ca2Fe2O5 (reducible at high temperatures) promoted the redox reaction of Fe3+ to Fe0 due to its great oxygen-carrying capacity, which explains the presence of metal iron in the Fe/dolomite after the steam reforming process [78,93]. Given that MgO-iron oxide is detected, it can be assumed that iron only reacts with CaO, which was also concluded by Di Felice et al. [94]. Note that the spent Fe/dolomite catalyst shows certain decrease in the intensity of MgO and CaO peaks with respect to the fresh one (Fig. 2b), as well as the presence of the calcite phase (not observed in fresh catalyst), probably due to a very limited carbonation at 850\u00a0\u00b0C in the gasifier. These results confirm that CaO and MgO are still the most important active phases for toluene cracking on the spent catalyst. Concerning the Fe/olivine XRD profile, apart from the Mg1.81Fe0.19(SiO4) olivine phase previously detected in the fresh catalyst, new Fe3O4 lines appear at 2\u03b8\u00a0=\u00a018\u00b0, 21\u00b0, 30\u00b0, 54\u00b0 and 57\u00b0. These new iron phases, together with the absence or sharp reduction in Fe0 lines, are evidence of a loss of active phase by metal iron oxidation under the reaction conditions used on the three catalysts analysed. However, Fe3O4 phase in the Fe/olivine and Ca2Fe2O5 in the Fe/dolomite are still active for the reforming and WGS reactions, which explains their slightly higher carbon conversion than Fe/Al2O3 when they underwent deactivation (Fig. 7a).Another cause of catalyst deactivation with time on stream is the coke deposited on these catalysts. Therefore, spent catalysts were subjected to temperature programme oxidation (TPO) in order to assess the amount and nature of the coke deposited. This coke blocks the access of reactants to the metal sites or encapsulates the Fe particles, thereby deactivating the crystallite. The TPO analyses revealed that the highest amount of coke deposits were formed on the Fe/Al2O3 catalyst, followed by Fe/olivine and Fe/dolomite, with values being 4.10, 2.36 and 1.17\u00a0wt%, respectively. It is well stablished in the literature that the rate and extent of coke formation increases by increasing the acid strength of the catalyst [57,95]. Thus, the higher coke deposition on the Fe/Al2O3 catalyst is explained by the higher acidity of alumina than olivine and dolomite. Moreover, the lowest coke formation rate in the Fe/dolomite catalyst is explained by two facts: (i) the presence of Ca2Fe2O5 phase (oxygen carrier) improves oxygen mobility on the catalyst surface, and therefore leads to faster carbon removal by oxidation, and (ii) the presence of CaO and MgO in the dolomite favours steam-carbon reactions, thus hindering polymerization reactions that promote coke development [89]. However, it should be noted that the catalytic activity of CaO for steam reforming decreases dramatically when the carbonate is formed [96]. Zamboni et al. [63] suggested that the oxygen vacancies in the Ca2Fe2O5 structure favour the reduction of water and, furthermore, Ca2Fe2O5 rearranges by releasing oxygen, which oxidizes carbon species to CO2. Note that the CO2 yield (45.5\u00a0wt%) is the highest on the Fe/dolomite, even though the H2 yield is similar in the product stream once catalysts have been deactivated (Fig. 8).\nFig. 10\n displays the TPO profiles of the spent catalysts, which allow determining the nature and the possible location of the coke within the catalyst. Apart from the mentioned differences in the amount of coke deposited, Fig. 10 shows that the nature of the coke varies depending on the type of support. A prevailing peak is observed at 450\u00a0\u00b0C for Fe/olivine and Fe/Al2O3, which is related to the combustion of the amorphous coke (hydrogenated composition) deposited on metal particles. Two small peaks are also observed at 600\u00a0\u00b0C for these catalysts, which are related to a more structured coke located on the catalyst support, even though its content is almost negligible due to the low intensity of both peaks. Another oxidation peak was detected at around 270\u00a0\u00b0C for the Fe/olivine catalyst, which is related to a less structured coke or heavy hydrocarbon deposits [37]. Virginie et al. [24] also detected 3 peaks in the TPO of a 10\u00a0wt% Fe/olivine deactivated catalyst. The first one at around 360\u00a0\u00b0C, which is attributed to surface carbon oxidation, the second one at around 500\u00a0\u00b0C, which is due to the oxidation of iron carbide, and the third one at around 610\u00a0\u00b0C assigned to filamentous carbon oxidation. The Fe/dolomite catalyst seemed to be more effective than the other two catalysts for preventing coke formation and, furthermore, its coke burns at lower temperatures (the main peak below 400\u00a0\u00b0C). This is explained by the presence of Ca2Fe2O5 phase, which increases oxygen mobility on the catalysts surface, thus favouring coke gasification and inhibiting its growth and evolution towards a more structured coke [93].In view of these results, it may be concluded that the active phase oxidation is the main deactivation cause, but the coke deposited on the Fe active sites also causes their blockage, and therefore contributes to the catalysts deactivation. The deactivation of Fe/dolomite and Fe/olivine catalysts is faster because the iron is mainly located on the external surface, and therefore coking reactions encapsulate more easily the metal particles. Indeed, the surface area of olivine and dolomite supports is limited, and therefore Fe dispersion is more restrictive than in the alumina support. The latter undergoes more severe deactivation by coke deposition, but this occurs for longer reaction periods.This study approaches tar elimination by feeding toluene as a representative tar compound and shows that Fe incorporation into olivine, dolomite and alumina increases the activity and selectivity towards hydrocarbon reforming and WGS reactions. The results are evidence that an increase in temperature to 900\u00a0\u00b0C leads to an increase in carbon conversion and H2 potential due to the enhancement of toluene reforming and cracking reactions. Concerning the efficiency of the primary catalysts, alumina provides the highest carbon conversion followed by dolomite, with their H2 potential being similar. In fact, the higher acidity of alumina promotes catalytic cracking reactions, thus leading to higher carbon conversions, even though its activity for WGS reaction is more limited. Dolomite is the one of highest activity for WGS reactions, which is related to CaO and MgO basic sites obtained after calcination. In addition, the presence of these species also improves tar decomposition, which in turn increases benzene yield at the expense of decreasing that of toluene.Regarding Fe loaded catalysts, Fe/Al2O3 provides the best performance in terms of carbon conversion and H2 potential. In fact, the higher porosity and BET surface area of alumina compared to those of olivine and dolomite improves the dispersion of Fe, which acts as the active phase for reforming and WGS reactions, whereas the alumina support provides the acidity required for cracking reactions. Furthermore, Fe/Al2O3 is the most stable catalyst and allows operating for longer periods with higher conversion values, whereas Fe/olivine, and especially Fe/dolomite, undergo faster deactivation, as evidenced by the sharper decrease in the reaction indices.The analyses of spent catalysts show that the main deactivation cause is the active phase oxidation followed by coke deposition on Fe active sites during the toluene conversion process. The XRD patterns show new iron oxidized phases and the absence of Fe0 lines (except for Fe/dolomite, with their intensity being significantly lower than those of the fresh one), whereas TPO analyses reveal a higher coke deposition for Fe/Al2O3 (4.4\u00a0wt%), which fully blocks the smallest pores of the catalyst. The coke deposited in all the spent catalysts has an amorphous nature and blocks the access of reactants to the metal sites, thereby deactivating the catalysts.\nMaria Cortazar: Investigation, Data curation, Methodology. Jon Alvarez: Writing \u2013 original draft, Writing \u2013 review & editing, Formal analysis, Visualization. Leire Olazar: Investigation, Data curation. Laura Santamaria: Validation, Data curation. Gartzen Lopez: Conceptualization, Methodology, Supervision. Heidi Isabel Villaf\u00e1n-Vidales: Methodology, Validation. Asier Asueta: Supervision. Martin Olazar: Supervision, Conceptualization, Writing \u2013 review & editing, Project administration.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was carried out with the financial support of the grants RTI2018-098283-J-I00 and PID2019\u2212107357RB-I00 funded by MCIN/AEI/ 10.13039/501100011033 and by \u201cERDF A way of making Europe\u201d and the grants IT1218\u221219 and KK-2020/00107 funded by the Basque Government. Moreover, this project has received funding from the European Union\u2019s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 823745.", "descript": "\n                  The performance of olivine, dolomite and \u03b3-alumina primary catalysts was evaluated in the continuous tar elimination process in which toluene was selected as the biomass gasification tar model compound. Iron was incorporated into these catalysts in order to improve their catalytic activity. All the experiments were performed in a continuous flow fluidized bed micro-reactor, with a steam/toluene ratio of 4 and a space velocity (GHSV) of 820\u00a0h\u22121, which corresponds to a catalyst amount of 3.8\u00a0cm3. The effect of temperature was studied using olivine in the 800\u2013900\u00a0\u00b0C range, which allowed concluding that 850\u00a0\u00b0C was the best temperature for tar removal. The fresh and deactivated catalysts were characterized by N2 adsorption\u2013desorption, X-ray fluorescence (XRF), X-ray diffraction (XRD) and temperature-programmed oxidation (TPO). Tar conversion efficiency was assessed by means of carbon conversion, H2 yield (based on the maximum allowed by stoichiometry), gas composition and product yields, with Fe/Al2O3 leading to the highest conversion (87.6\u00a0%) and H2 yield (38\u00a0%). Likewise, Fe/Al2O3 also provided the highest stability, as it allowed operating for long periods with high conversion values (85.9\u00a0% after 35\u00a0min on stream), although it underwent severe deactivation. The analysis of the spent catalysts revealed that deactivation occurred mainly by coke deposition on the catalyst surface and iron phase oxidation, with Fe/olivine and Fe/dolomite leading to the faster deactivation due to their poorer metal dispersion related to their reduced surface area. The TPO profiles showed that the coke deposited on the three catalysts was amorphous with a very small contribution of highly structured carbon.\n               "}
{"full_text": "Fuel cells are electrochemical devices that efficiently convert the chemical energy of fuel and an oxidant into electricity. Specifically, SOFCs have advantages such as high efficiency for power generation and fuel flexibility, without the requirement for precious metal catalysts due to their high operating temperature (e.g. from 600 to 900\u00a0\u00b0C) [1\u20136]. Stationary combined heat-and-power fuel cell systems have already been commercialized. For example, the number of residential fuel cell systems (commonly called \u201cENE-FARM\u201d) installed in Japan has already exceeded 400,000 as of August 2021 [7]. The polymer electrolyte fuel cell class of ENE-FARM was launched in 2009, followed by the SOFC class in 2011, and the percentage of the SOFC class is gradually increasing largely due to its higher efficiency [8]. In addition, a 250\u00a0kW class SOFC-micro gas turbine hybrid system has been commercialized by Mitsubishi-Hitachi Power Systems (now Mitsubishi Heavy Industries), with a 1\u00a0MW system in the planning stage [9]. Meanwhile, Ceres Power in UK has set a target of producing SOFC systems with a cumulative capacity of several hundred MW, having signed commercial contracts with Robert Bosch in Germany, and Doosan in Korea [10]. Moreover, Bloom Energy in the U.S. sold a cumulative 80.9\u00a0MW of SOFC system capacity by 2018, representing a leading effort in the commercialization of SOFC systems [11]. However, research and development into SOFCs that operate, e.g., at relatively lower temperatures are still in progress due to the limited availability of materials which can maintain their performance for long periods of time at elevated temperature [12,13].Nickel cermets are widely used as SOFC anodes, and are comprised of Ni and an electrolyte component such as yttria-stabilized zirconia (YSZ) [1\u20136,14]. However, the Ni-based electronic conducting network in these anodes can be disrupted by repeated cycling of redox reactions, associated with the coarsening and aggregation of Ni particles [15]. Such redox processes occur when the fuel supply is interrupted upon system startup and shutdown, resulting in the deterioration of the overall efficiency of SOFC systems [16\u201326]. The development of SOFC anodes with high redox cycle durability has another important practical benefit, eliminating the need for an inert gas supply to prevent Ni oxidation during system shutdown, significantly simplifying system design and control.Therefore, it is desirable to develop SOFC anodes with high durability and resilience against redox cycling for the more widespread use of SOFCs. To solve this issue, the use of more redox stable materials such as strontium titanate (SrTiO3) may be helpful [27\u201330], whilst it can be doped with a higher valence cation at the A site to improve its electronic conductivity [31\u201335]. Strontium titanate doped with La3+ at the A site (i.e, Sr0.9La0.1TiO3, commonly known as LST) exhibits high electronic conductivity under SOFC operating conditions [31]. Furthermore, LST has a comparable thermal expansion coefficient to that of zirconia-based electrolyte materials [32]. Futamura et\u00a0al. demonstrated that, impregnation of porous anodes with fine metallic catalyst nanoparticles such as Ni and Rh, high durability against redox cycling could be achieved when supplied with 3% humidified hydrogen fuel, without sacrificing electrochemical performance [30]. However, when hydrogen-related fuel gas is supplied to SOFC systems, the water vapor concentration is reported to increase near the fuel gas outlet, and thus the Ni anode catalysts tend to oxidize [21], one of the major degradation mechanisms of SOFC anode materials. The electrochemical performance of impregnated anodes at high fuel utilization (i.e. at a higher water vapor concentration in the fuel feed) is slightly lower than that of conventional anodes [29,30]. For successful commercialization, both high performance and durability must be achieved simultaneously. The stability of anodes at high fuel utilization, if improved, enables higher power generation efficiency, because a larger fraction of the fuel can be utilized for power generation.As an alternative approach, novel SOFC anodes are being developed using Ni alloyed with transition metal elements rather than using pure Ni. In general, metallic Ni-alloys can exhibit higher electronic conductivity compared to ceramics such as donor-doped SrTiO3. When such alloy-based cermet anodes are exposed to an oxidizing atmosphere, the formation of a surface oxide layer suppresses further oxidation within the alloy particles, and it is possible that this could improve the durability against redox cycling in SOFCs. However, it is well known that doped elements in alloys generally act as scattering sites detrimental to electronic conductivity [36]. Therefore, it is important that Ni-alloys should exhibit both sufficient electronic conductivity and catalytic activity for efficient SOFC operation, even under a hydrogen-based fuel supply. Araki et\u00a0al. already reported the synthesis of various Ni alloy anodes using a spray drying technique, and investigated the effect of alloying on SOFC efficiency and H2S poisoning [37]. Meanwhile, Ishibashi et\u00a0al. prepared NiO-containing complex oxide powders and various Ni alloy anodes by ammonia co-precipitation, and reported electrochemical performance and redox cycling durability, finding that cobalt is a promising element for the formation of surface oxide layers for enhanced durability against redox cycling [38].Although the durability may be improved using alloying, cobalt is classed as a critical raw material, i.e. an element of high economic importance, but limited supply [39]. The price of Co will continue to increase due to the rise in demand for lithium-ion batteries, especially in battery electric vehicles (BEVs) [39\u201341]. As such, the overall amount of cobalt in SOFCs should be minimized where possible, so that the redox durability can be enhanced without risking supply chains or increasing the cost. The source of cobalt and the related geopolitical situation should also be considered.The objective of this study is to develop highly durable Ni\u2013Co alloy cermet anodes, whilst keeping the Co content as low as possible. Gadolinium-doped ceria (GDC) is selected as the electrolyte component in the anodes, and the effect of systematically adjusting the Ni:Co ratio on SOFC performance and microstructure will be investigated. Long-term tests will also be conducted at high fuel utilization (i.e. using a highly-humidified hydrogen supply).A stability diagram was computationally obtained using the software FactSage (Version 7.4, Thermfact Ltd., Canada) in order to examine the thermochemically stable phases of Ni\u2013Co alloys in reducing and oxidizing atmospheres at 800\u00a0\u00b0C (a typical SOFC operating temperature). In this study, we used the Phase Diagram module of FactSage to derive the Ni\u2013Co and NiO\u2013CoO phase diagrams and the Ni/NiO and Co/CoO phase boundaries.NiO and CoO composite oxide powders were prepared by ammonia co-precipitation, as described in Fig.\u00a01\n. Ni(NO3)2\u22c56H2O and Co(NO3)2\u22c56H2O (Kishida Chemical Co., Ltd., Japan) were used as the precursors for co-precipitation. Briefly, ammonia solution was added dropwise to an aqueous solution containing Ni and Co ions (in various ratios) to simultaneously precipitate these complex hydroxides [38]. The precipitates were then filtered, dried at 100\u00a0\u00b0C for 10\u00a0h in air, and calcined at 1000\u00a0\u00b0C for 2\u00a0h in air. The resulting composite oxide powders were mixed with GDC (Ce0.9Gd0.1O2, Rhodia, ULSA grade, USA) in a mass ratio of 48.1:51.9, resulting in a volume ratio of 50:50 for pure NiO (slightly changing when CoO was also incorporated). Ni\u2013Co-GDC cermet anodes were prepared with a variety of molar ratios, Ni:Co = (100-x):x, where x\u00a0=\u00a00, 5, 10, 20, and 30, herein referred to as Ni(100-x)Cox-GDC. Lanthanum strontium manganite (LSM, (La0.8Sr0.2)0.98MnO3, Praxair, USA) and scandium-stabilized zirconia (ScSZ, Daiichi Kigenso Kagaku Kogyo Co., Ltd., Japan) were used in the cathode. At the cathode side, a 1:1 LSM:ScSZ mass ratio was used close to the electrolyte, and 100% LSM powder was used close to the current collector [15,38].\nFig.\u00a02\n shows a schematic of the cell structure. In this study, the classical electrolyte-supported structure rather than electrode-supported or metal-supported structures is selected, in order to extract the anode-side voltage losses using a reference electrode [38]. The solid electrolyte is comprised of a plate of ScSZ (10\u00a0mol% Sc2O3, 1\u00a0mol% CeO2, 89\u00a0mol% ZrO2) with a diameter of 20\u00a0mm and a thickness of 200\u00a0\u03bcm. The anode layer was screen-printed onto the ScSZ electrolyte followed by heat-treatment at 1300\u00a0\u00b0C for 3\u00a0h in air. The cathode layer was then screen-printed onto the opposite side of the electrolyte followed by heat-treatment at 1200\u00a0\u00b0C for 5\u00a0h in air. The anode and cathode layers were both approximately 40\u00a0\u03bcm in thickness after heat-treatment. To separate the anode and cathode overvoltages, a reference electrode with an area of ca. 4\u00a0mm2 size was placed 2\u00a0mm away from the cathode, using Pt paste. A Pt mesh was attached to the surface of each electrode as a current collector, set on the electrodes after screen printing, and then each electrode was heat-treated. The electrode area was 8\u00a0\u00d7\u00a08\u00a0mm2 (0.64\u00a0cm2).\nFig.\u00a03\n shows the configuration of the electrochemical experimental setup. Cell performance tests were conducted using automated fuel cell evaluation systems (AutoSOFC, TOYO Corporation, Japan). Before measuring the cell polarization curves, the cell was sealed using a Pyrex glass ring and increasing the temperature to 1000\u00a0\u00b0C at 200\u00a0\u00b0C\u00a0h\u22121. Then, the cell temperature was maintained at 1000\u00a0\u00b0C, and 3% humidified hydrogen was supplied at 100\u00a0mL\u00a0min\u22121 for 1\u00a0h to reduce the metal oxide in the anode to the metallic state. The selected reduction temperature (1000\u00a0\u00b0C) is sufficient to reduce NiO to metallic Ni in the thin porous anode layers of the electrolyte-supported cells [15,21,29,30,38], whereas a higher temperature may be required for e.g. anode-supported and metal-supported cells with thicker anode layers. The current-voltage (I\u2013V) characteristics were then measured at 800\u00a0\u00b0C with 3% humidified hydrogen fuel supplied to the anode. The anode voltage (potential) was measured relative to the reference electrode on the cathode side. As the anode voltage includes both ohmic and non-ohmic overvoltages, these overvoltages (i.e., anode-side ohmic loss and anodic overvoltage) were separated via the current interrupt method [15,38]. Each electrochemical measurement was repeated three times, and the average values and their standard deviation are given to check their uncertainty.Redox cycling tests were conducted at 800\u00a0\u00b0C to evaluate the durability of the anodes, as described in Fig.\u00a04\n [15]. First, the cell was operated for 1\u00a0h at a current density of 0.2\u00a0A\u00a0cm\u22122 with 3% humidified hydrogen fuel supply (i); then, the fuel supply was interrupted for 1.5\u00a0h (ii); and finally, the supply of 3% humidified hydrogen was restarted (iii). These three steps are regarded as one cycle, and 50 cycles were applied in total. The anode voltage, non-ohmic anodic overvoltage, and anode-side ohmic loss were measured during these cycling tests.The I\u2013V characteristics and durability were also evaluated during continuous power generation up to 1000\u00a0h under 80% humidified hydrogen supply. The I\u2013V characteristics were measured under the same conditions as the cell performance test section described above. The anode voltage, anodic overvoltage, and anode-side ohmic loss were measured at a current density of 0.2\u00a0A\u00a0cm\u22122 and 80% humidified hydrogen over a course of 1000\u00a0h.Scanning electron microscopy (SEM) with energy-dispersive X-ray spectroscopy (EDS) was performed to evaluate the distribution of elements in the whole of anodes using dual-beam SEM (Versa 3D, FEI, USA). High-resolution observation and elemental analysis on the Ni(100-x)Cox-GDC anodes were carried out by means of scanning transmission electron microscopy (STEM, JEM-ARM200F, JEOL, Ltd., Japan).Phase diagrams of Ni\u2013Co and NiO\u2013CoO systems were calculated to estimate the stable phases in both reducing and oxidizing atmospheres. The resulting phase diagrams are shown in Fig.\u00a05\n, where the triangular symbols specify the positions at 800\u00a0\u00b0C for the following compositions: Ni-GDC (black); Ni95Co5-GDC (red); Ni90Co10-GDC (green); Ni80Co20-GDC (blue); and Ni70Co30-GDC (pink).\nFig.\u00a05(a) predicts that all five compositions will form Ni\u2013Co alloys with fcc structure in reducing atmosphere. Fig.\u00a05(b) predicts that all five compositions will form a solid solution with a single-phase NiO crystal structure in oxidizing atmosphere. Therefore, in the reducing atmosphere of the SOFC anode, the material is expected to be a true alloy, with Co dissolved in the Ni lattice. Fig.\u00a06\n shows a stability diagram calculated after Ishibashi et\u00a0al. [38], in which the coexisting equilibrium oxygen partial pressures of the metallic and oxide states of Ni and Co are plotted as a function of temperature. In this figure, the equilibrium boundary between Co and CoO is located at a lower oxygen partial pressure compared to Ni/NiO. This indicates that Co is slightly more easily oxidized compared to Ni. As such, when the fuel supply is interrupted in SOFC anodes, Co will be oxidized first, forming an oxide surface layer on the Ni\u2013Co alloy. Evidence of the formation of a Co-rich layer on surface of the Ni\u2013Co alloy has already been observed via higher-resolution TEM [38]. Therefore, based on the calculated thermochemical stability of Ni\u2013Co alloys in reducing and oxidizing atmospheres, it is expected that the redox cycle durability could be improved by alloying Ni with Co through the formation of a protective surface oxide layer. Further analysis is of scientific and technological interest, such as in-situ microstructural observation of redox processes at Ni\u2013Co alloy surfaces.\nFig.\u00a05 also confirms that a solid solution will be formed at elevated temperature in both the Ni\u2013Co and NiO\u2013CoO systems. Whilst the sintering temperature of the anodes was around 1300\u00a0\u00b0C, which is relatively low compared to typical anode-supported cells, these phase diagrams predict the formation of complete solid solutions even at higher sintering temperatures above 1300\u00a0\u00b0C.\nFig.\u00a07\n shows representative plots of (a) the anode voltage (anode potential) relative to the reference electrode, (b) the anodic overvoltage, and (c) the anode-side ohmic losses for each anode material, with error bars indicating their standard deviation, measured at 800\u00a0\u00b0C. Fig.\u00a07(a) reveals that the pure Ni-GDC anode exhibits the highest I\u2013V performance. As shown in Fig.\u00a07(b) and (c), the non-ohmic anodic overvoltage and the ohmic losses of anodes for which x\u00a0=\u00a010, 20, and 30 are higher than those of the pure Ni-GDC anode. These results suggest that Ni is more active than Co, and that alloying Ni with Co leads to a decrease in electronic conductivity. The dependence of overvoltage on the Co concentration above 10% will be a matter for future studies.In contrast, the anode for which x\u00a0=\u00a05 exhibits similar I\u2013V characteristics as the pure Ni-GDC anode, as shown in Fig.\u00a07(a). This result indicates that the decrease in power generation performance due to the addition of Co is negligible when the content of Co is sufficiently low. This result suggests that, by optimizing the ratio of Ni and Co, it is possible to prepare alloy-based cermet anodes with high redox cycle durability while maintaining power generation performance.Following the protocol outlined in Fig.\u00a04, the durability against redox cycling was measured three times for each of the five anode compositions. Typical results for the change in (a) anode voltage; (b) anodic overvoltage; and (c) anode-side ohmic loss during the test, measured at 800\u00a0\u00b0C, are shown in Fig.\u00a08\n. The percentage of anode voltage drop from the first cycle to the 50th cycle is shown in Fig.\u00a09\n, with error bars based on the standard deviation calculated from the results of three cells for each composition.In Fig.\u00a08, the observed changes during the early stages of the tests could be caused by e.g. the thermal history of the cells during the sintering processes. Regarding durability, Fig.\u00a08(a) reveals that the anode voltage for pure Ni-GDC significantly drops over 50 redox cycles, finally reaching around 0.7\u00a0V. In contrast, all the Ni\u2013Co alloy anodes display a much smaller drop in anode voltage, dropping to ca. 0.87\u00a0V after 50 redox cycles. Fig.\u00a09 plots the percentage change in anode voltage after 50 cycles, and a clear trend of decreasing degradation with increased Co content emerges. The largest difference is observed between the pure Ni-GDC anode and the Ni95Co5-GDC anode, indicating that the addition of even a small amount (5%) of Co can considerably improves the durability against redox cycling.\nFig.\u00a08(b) and (c) reveal that the addition of Co suppresses increases in both the non-ohmic anodic overvoltage and the ohmic losses during redox cycling. This is attributed to the preferential formation of a Co-based surface oxide layer on the Ni\u2013Co alloy, helping to suppress redox-induced aggregation of Ni particles, and thus preventing a decrease in electrode reaction area and maintaining electron-conducting network.Based on the fact that the durability of SOFC against redox cycling can be significantly improved even with the addition of just 5\u00a0mol% Co, herein the Ni95Co5-GDC anode is selected as a suitable anode composition for combining durability with minimal content of critical raw materials. Therefore, in the next section, the performance and durability of SOFCs with pure Ni-GDC and Ni95Co5-GDC anodes with 80% humidified hydrogen are evaluated.The electrochemical performance and long-term durability of the pure Ni-GDC and Ni95Co5-GDC anodes are herein investigated under more severe operating conditions, namely at higher water vapor partial pressure, simulating higher fuel utilization. The I\u2013V characteristics and durability up to 1000\u00a0h were measured at 800\u00a0\u00b0C using 80% humidified hydrogen supply. Fig.\u00a010\n(a) compares the initial anode voltage, whilst (b) and (c) compare the initial anodic overvoltage, and the initial anode-side ohmic loss, respectively. Meanwhile, Fig.\u00a011\n shows (a) the anode voltage, (b) the anodic overvoltage, and (c) the anode-side ohmic loss throughout the 1000-h durability test.\nFig.\u00a010 confirms that the Ni-GDC and Ni95Co5-GDC anodes exhibit almost the same I\u2013V characteristics under 80%-humidified hydrogen supply, with the anodic overvoltage and the anode-side ohmic loss also being very similar. This confirms that the addition of a low concentration of Co has negligible or slightly positive effect on the I\u2013V performance even at high fuel utilization, simulating the conditions at the downstream of the fuel supply in fuel cell systems, where redox-related degradation tends to occur. The reactions responsible for the slight improvement in electrochemical performance via alloying Ni with Co under these highly-humidified conditions are of scientific interest, where e.g. dealloying, segregation of secondary phases, formation of nano-composites, and/or their co-catalyst effects could occur.The anode voltage measured over the 1000-h durability test at 800\u00a0\u00b0C is shown in Fig.\u00a011(a). As the anode voltage in (a) remained stable during 1000\u00a0h, certain noisy fluctuation in the raw data shown in (b) and (c) may not influence the overall durability. Fig.\u00a011(a) confirms that the addition of Co slightly improves the long-term durability under highly humidified hydrogen supply. Fig.\u00a011(b) and (c) confirm that this improvement in the long-term durability is mainly associated with a decrease in non-ohmic anodic overvoltage, which is suppressed by the addition of Co.The non-ohmic overvoltages were averaged, and their standard deviations derived in every 100\u00a0h throughout the test, giving 80.2\u00a0\u00b1\u00a06.1\u00a0mV for the pure Ni-GDC anode, and 44.7\u00a0\u00b1\u00a02.0\u00a0mV for the Ni95Co5-GDC anode. Matsui et\u00a0al. previously reported that the Ni surface can be oxidized under highly humidified hydrogen supply, reducing the extent of the triple-phase boundary [42], increasing the non-ohmic overvoltage. As such, it is concluded that the addition of Co does not accelerate this process, and therefore can be used in SOFC anodes even in highly-humidified fuel streams.To elucidate the reasons for changes in the performance and durability, the distribution and chemical state of Ni and Co in the Ni\u2013Co alloy anodes were measured before the durability test. Fig.\u00a012\n shows representative elemental distribution maps of the Ni95Co5-GDC and Ni70Co30-GDC anodes, analyzed by SEM-EDS. As shown in Fig.\u00a012, in both samples (a) and (b), the location of Ni and Co within the anode layers is exactly the same. This analysis confirms that the elemental distributions of Ni and Co are almost identical in both anodes. The molar ratios of Ni and Co obtained by point analysis compiled in Table 1\n are very close to the nominal ratios targeted in the co-precipitation synthesis step, namely Ni:Co\u00a0=\u00a095:5, and Ni:Co\u00a0=\u00a070:30, confirming the validity of the phase diagram shown in Fig.\u00a05(a).\nFig.\u00a013\n shows STEM-EDS elemental distribution maps of the Ni-GDC anode before and after the redox cycling test, for Ni, O and Ce. Before the test, the elemental O distribution largely overlaps with Ce, but not with Ni, suggesting that pure Ni mainly exists as a metal. In contrast, after the durability test, oxygen is observed to additionally be present at the surface of the Ni particles, confirming that nickel particles are oxidized during the durability tests.Meanwhile, Fig.\u00a014\n shows elemental distribution maps for the Ni95Co5-GDC anode before and after the redox cycling test, in this case also including the Co distribution. Before the test, the distributions of Ni and Co closely overlap, and the distributions of O and Ce overlap. However, the O distribution does not overlap with either Ni or Co. This confirms that Ni and Co co-exist as an alloy, as expected from the phase diagram in Fig.\u00a05 (a). In contrast, after the durability test, oxygen can additionally be found uniformly on the surface of the Ni\u2013Co alloy particles. However, in this case, Co also appears to be more concentrated near the surface of the alloy particles.Quantitative evaluation of the different ratios of these elements near the surface of the alloy particles was performed using point analysis. Fig.\u00a015\n(a) shows a STEM image of the Ni95Co5-GDC anode after the redox cycling test, and Fig.\u00a015(b) shows a higher-magnification image of the highlighted area in Fig.\u00a015(a). The ratios of Ni, Co, and O in the Ni95Co5-GDC anode were then compared by point analysis at positions 1, 2, and 3, with 1 being closest to the surface, and 3 being furthest away. The particle on the right side of position 1 is GDC. The results of this quantitative analysis are shown in Table 2\n, confirming that the ratio of O is higher nearer the surface of the alloy particle. Meanwhile, the Ni:Co ratios were 62.6:37.4\u00a0at position 1; 91.4:8.6\u00a0at position 2; and 92.8:7.2\u00a0at position 3, confirming that the Co ratio also increases near the surface. These results clearly indicate that a Co-rich oxide film is formed at the surface of the alloy particles during the redox cycling durability test.These results confirm that a Ni-based oxide layer is formed on the Ni-GDC anode, while a Co-containing oxide layer is formed on the Ni\u2013Co alloy anode during the redox cycling test. As shown in the stability diagram in Fig.\u00a06, Co is more stable as an oxide than Ni. When NiO is reduced to metallic Ni, aggregation is promoted by the formation of Ni fine particles due to volume shrinkage, as verified by Matsuda et\u00a0al. [43]. Therefore, when the fuel supply is restarted, the dense oxide layer formed on the Ni\u2013Co alloy anode surface is more stable and more difficult to be reduced compared to the Ni-GDC anode, preventing aggregation during the reduction from NiO to metallic Ni.The EDS images of the Ni-GDC and Ni95Co5-GDC anodes after 1000-h power generation test under 80% humidified hydrogen supply are shown in Fig.\u00a016\n. In this case, the Ni-GDC anode images reveal thick oxide layers on the surface of the Ni particles. In contrast, images of the Ni95Co5-GDC anode reveal that both O and Co are present at the surface of Ni\u2013Co alloy particles. These micrographs indicate that a Co-containing dense oxide layer is formed at the surface of the Ni\u2013Co alloy particle even under high water vapor pressure. As mentioned in the previous section, the Ni95Co5-GDC anode suppressed the increase in non-ohmic anodic overvoltage during the 1000-h test under 80% humidified hydrogen supply, compared to the Ni-GDC anode. This can now be directly attributed to the difference in the surface compositions of these two anodes.Because Ni acts as an excellent catalyst in SOFC anodes, a decrease in the Ni ratio on the particle surface is expected to decrease the electrocatalytic activity. However, in this case the electrode activity is actually maintained in Ni\u2013Co alloys even when the Ni ratio on the electrode surface decreases. In addition, the alloying of the surface may suppress the aggregation of Ni\u2013Co particles. More detailed mechanisms to explain how the high electrocatalytic activity is retained in alloys is of scientific interest in future studies, and further investigation is needed. This will be performed via e.g. quantitative particle size analysis, impedance measurements, and distribution of relaxation time (DRT) analysis.When these alternative anode materials are applied to practical SOFCs, it will also be essential to evaluate the effects of using impurity-containing fuels [44\u201351]. Sulfur is a typical SOFC fuel impurity causing cell voltage drop [44\u201346]. Chlorine is another typical trace impurity in e.g. tap water [44,45,47], while phosphorous causes serious degradation even at ppb levels [44,45]. Siloxane is a contaminant in e.g. digester gas causing silica formation [44,45]. Trace impurities in biofuels affect the performance and durability of SOFCs [48\u201351]. As such, it will be important to evaluate impurity poisoning in Ni\u2013Co alloy anodes in future studies, associated with various chemical reactions with Ni and/or Co.Ni\u2013Co alloy cermet anodes for SOFCs were prepared and compared to pure-Ni cermet anodes. The initial I\u2013V performance of the Ni\u2013Co alloy anodes was comparable to that of the Ni-based anode. The redox cycling durability was considerably improved by alloying, even at low concentrations of 5\u00a0mol%. A 1000-h durability test was performed using highly humidified hydrogen supply, confirming that alloying can suppress increases in non-ohmic anodic overvoltage without compromising the I\u2013V performance, even at high fuel utilization. Electron microscopy revealed that for pure Ni-based anodes, a nickel oxide layer is formed on the surface of the Ni particles, while in the case of alloy anodes, a Co-containing oxide layer is formed during operation. Since Co is a more stable oxide compared to Ni, this Co-containing oxide layer suppresses aggregation of Ni-based particles under redox cycling, or at high water vapor partial pressure. As such, Ni\u2013Co alloy can be regarded as a robust cermet anode material for SOFCs, realizing high electrochemical performance, redox cycling durability, and long-term durability even under a high water vapor atmosphere, and with a Co content as low as 5\u00a0mol%. As such redox-stability is more critical for anode-supported cells than for electrolyte-supported cells, it is of technological interest to apply the redox-stable alloy anode material to anode-supported SOFCs and related electrochemical cells.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.A part of this study was supported by \u201cResearch and Development Program for Promoting Innovative Clean Energy Technologies Through International Collaboration\u201d of the New Energy and Industrial Technology Development Organization (NEDO) (Project No. JPNP20005). Collaborative support by Prof. H. L. Tuller, Prof. B. Yildiz, and Prof. J. L. M. Rupp at Massachusetts Institute of Technology (MIT) is gratefully acknowledged.\n\nAbbreviations\nDRT\n\ndistribution of relaxation time\n\nEDS\n\nenergy-dispersive X-ray spectroscopy\n\nGDC\n\ngadolinium-doped ceria\n\nLSM\n\nlanthanum strontium manganite\n\nLST\n\nlanthanum-doped strontium titanate\n\nScSZ\n\nscandium-stabilized zirconia\n\nSEM\n\nscanning electron microscopy\n\nSOFC\n\nsolid oxide fuel cell\n\nSTEM\n\nscanning transmission electron microscopy\n\n\ndistribution of relaxation timeenergy-dispersive X-ray spectroscopygadolinium-doped cerialanthanum strontium manganitelanthanum-doped strontium titanatescandium-stabilized zirconiascanning electron microscopysolid oxide fuel cellscanning transmission electron microscopy", "descript": "\n                  Ni alloys are examined as redox-resistant alternatives to pure Ni for solid oxide fuel cell (SOFC) anodes. Among the various candidate alloys, Ni\u2013Co alloys are selected due to their thermochemical stability in the SOFC anode environment. Ni\u2013Co alloy cermet anodes are prepared by ammonia co-precipitation, and their electrochemical performance and microstructure are evaluated. Ni\u2013Co alloy anodes exhibit high durability against redox cycling, whilst the current-voltage characteristics are comparable to those of pure Ni cermet anodes. Microstructural observation reveals that cobalt-rich oxide layers on the outer surface of the Ni\u2013Co alloy particles protect against further oxidation within the Ni alloy. In long-term durability tests using highly humidified hydrogen gas, the use of a Ni\u2013Co cermet with Gd-doped CeO2 suppresses degradation of the power generation performance. It is concluded that Ni\u2013Co alloy cermet anodes are highly attractive for the development of robust SOFCs.\n               "}
{"full_text": "Nowadays, the increasing energy demand and environment pollution have stimulated the research for the utilization of clean energy, in which hydrogen is considered as one of the most potential energy carrier [1]. For hydrogen economy, the greatest challenge lies in the hydrogen storage. Compared with gas-state and liquid-state hydrogen storage, solid-state hydrogen storage materials with higher energy density are safer [2]. Mg is not only used as a structural material [3,4] or biomedical material [5], but is also regarded as a particularly promising candidate for hydrogen storage, due to its low density, abundant resource and high theoretical hydrogen storage capacity [6,7]. Nevertheless, its thermodynamic stability, sluggish reaction kinetics and inherent low thermal conductivity impede the process of practical applications [8\u201311]. In the past decades, many solutions have been developed to overcome these problems, such as nanosizing [12,13], catalyzing [14,15], alloying [16], etc.Mechanical milling is widely used to disperse different catalysts. Particularly, transition metals (Nb, V, Ti, Co, etc.) [17,18], their oxide [19,20] and non-metal element [21,22], are usually dispersed into MgH2/Mg system. Among the transition metals, Ni-based compounds exhibit excellent catalytic effect on the hydrogen absorption/desorption of MgH2\n[23\u201325]. When the size is reduced to nanoscale, the catalytic effect will be further enhanced. Liu et al. [23] reported that Mg-Ni nano-composite, prepared by a wet chemical method, could absorb 85% of its maximum hydrogen capacity within 45\u00a0s at 125\u00a0\u00b0C. Furthermore, Ni nanofibers with a diameter of \u223c50\u00a0nm and porous structure could enhance the desorption properties of MgH2\n[26]. For example, the onset temperature (143\u00a0\u00b0C) and peak temperature (244\u00a0\u00b0C) of dehydrogenation are much lower for Mg with 4% Ni nanofibers addition, than that mixed with 4% Ni powders (300\u00a0\u00b0C for onset temperature and 340\u00a0\u00b0C for peak temperature). In addition, many carbon materials, including graphite [27], single-walled carbon nanotubes (SWNTs) [28], graphene nanosheets (GNs) [29] and activated carbon [30], have been studied to improve the hydrogen storage properties of MgH2. Moreover, recent theoretical calculation has shown that the combination of nanosizing and carbon materials leads to further enhanced hydrogen storage properties [31]. Huang et al. [32] revealed that carbon played an important role on inhibiting the aggregation of the catalysts. All in all, among the carbon materials, graphene, with unique 2D nanostructure, excellent electronic and thermal conductivity, and high chemical stability, is an ideal supporter to host disperse nanoparticles (NPs), which exhibits great catalytic effect in the hydrogen storage areas [33,34]. The combination of the transition metal and carbon seems to be more efficient on the improvement of hydrogen storage [35,36]. Liu et al. [37] synthesized porous Ni@rGO with GO and NiCl2\u00b76H2O as raw materials. Ni NPs loading on the rGO has better catalytic effect on the desorption kinetics of MgH2 than that of Ni or rGO alone. Zhang et al. [38] reported that Ni decorated graphene nanoplate (Ni/Gn), which was prepared with Gn and Ni(NO3)2\u00b76H2O, enhanced the sorption properties of MgH2. The sample of Mg@Ni8Gn2 absorbed 6.28 wt.% H2 in 100 s at 100\u00a0\u00b0C. To maximize the effect of graphene on preventing the growth and aggregation, a simple solvothermal method was adopted to prepare ordered structure of MgH2 NPs with good dispersion on graphene by Xia et al. [39]. The hydrogen capacity of the composite did not decay after 100 cycles, which is due to the stable structure.To sum up, carbon supported Ni NPs shows a positive effect on enhancing the hydrogen storage properties of MgH2. However, it is still a challenge to prepare uniformly dispersed Ni NPs on the surface of graphene, and the evolution of Ni during the hydrogenation/dehydrogenation cycles is not clear. In this work, Ni@rGO was successfully synthesized through annealing the Ni(OH)2@rGO with aqueous solution of GO and Ni(NO3)2\u00b76H2O as raw materials, and then it was doped into MgH2 by mechanical milling to improve the hydrogen absorption/desorption properties. The effects of milling time and Ni@rGO with different Ni loading amount have been investigated in detail. The evolution of Ni during the hydrogenation/dehydrogenation cycles has been clarified.Typically, 25\u00a0mL aqueous solution of GO (8\u00a0mg/mL, Shenzhen Matterene Technology Company) was diluted to 1\u00a0mg/mL, and then 248\u00a0mg Ni(NO3)2\u00b76H2O (AR) was added to the solution by ultrasonic vibration for 1\u00a0h. Freshly prepared aqueous solution of NaBH4 (5\u00a0mL, 1\u00a0M) was added to prepare composite precursor. After stirring the mixture solution at room temperature for 2\u00a0h, the products were collected by centrifuge and washed with deionized water and ethanol several times. Then, it was dried in a blowing dry oven at 100\u00a0\u00b0C for 12\u00a0h. Afterwards, the products was processed with experienced heat treatment at 500\u00a0\u00b0C for 2\u00a0h in a tube under Ar/H2. The mass ratio of Ni and GO was 2:8, 4:6 and 6:4, and the corresponding samples were denoted as Ni2@rGO8, Ni4@rGO6, and Ni6@rGO4.Magnesium hydride was prepared by reactive ball milling method according to our previous work [40]. Firstly, magnesium (purity > 99.9%, 400\u00a0mesh) was mechanically milled under Hydrogen atmosphere with an initial pressure of \u223c 1 MPa, followed by a long-period hydrogenation at 350\u00a0\u00b0C.The as-synthesized Ni2@rGO8 catalysts and MgH2 powders mixed in a fixed mass ratio of 10:90. The milling was then performed under the hydrogen pressure of 1\u00a0MPa, with the rotational speed of 450\u00a0rpm, and the ball-to-powder mass ratio of 40:1. The mixtures were milled for different periods (2\u00a0h, 5\u00a0h, 10\u00a0h, and 20\u00a0h). The corresponding samples were denoted as Mg-Ni2@rGO8-2h, Mg-Ni2@rGO8-5h, Mg-Ni2@rGO8-10h, and Mg-Ni2@rGO8-20h. The corresponding samples with different Ni@rGO addition were denoted as MH-Ni2@rGO8, MH-Ni4@rGO6 and MH-Ni6@rGO4.The loading amount of the as-prepared Ni@rGO was quantified by differential scanning calorimetry combined with thermogravimetry (TG-DSC, STA449 F3, Netzsch). The structure and morphology of the synthesized materials were characterized by powder X-ray diffraction (XRD, D8 Advance, Bruker), scanning electron microscopy (SEM, Sigma 500, Zeiss) and transmission electron microscopy (TEM, F200, FEOL), respectively.The decomposition performance of the composite was measured by DSC. Hydrogen storage properties were determined by using a homemade HPSA-auto apparatus [41], in which the modified Benedict-Webb-Rubin (MBWR) EOS was applied to calculate the compression factor of H2 gas. To reduce the error of the system, two high-accuracy pressure transducers (Keller, \u00b10.05% FS) and three thermocouple of Pt 100 were used to monitor the pressure and temperature. Absorption-desorption measurements were performed at various temperature with an initial pressure of 3\u00a0MPa for hydriding and 0.0004\u00a0MPa for dehydriding, respectively. Before ab/desorption measurements, the samples performed dehydrogenation and rehydrogenation at 340\u00a0\u00b0C. All materials handling was performed in a glove-box filled with purified argon (99.999%), in which water vapour and oxygen levels were below 1\u00a0ppm by a recycling purification system to prevent samples from hydroxide formation and/or oxidation.The XRD patterns of the as-prepared Ni2@rGO8, Ni4@rGO6, and Ni6@rGO4 are shown in Fig.\u00a01\n. The peak at 26\u00b0 is contributed by rGO (002). The characteristic diffraction peaks at 44.5\u00b0 (111) and 51.8\u00b0 (200) are attributed to face-centered crystalline nickel. Firstly, GO was reduced by the addition of NaBH4, and Ni(OH)2 was formed from the Ni2+ in the solution and anchored on the rGO (Fig. S1). After annealing in H2/Ar flow gases at 500\u00a0\u00b0C, the Ni(OH)2 was further reduced to Ni. Based on Scherrer equation, the average grain size of Ni in Ni2@rGO8, Ni4@rGO6, and Ni6@rGO4 were estimated to be 8.5\u00a0nm, 13.1\u00a0nm, and 15.4\u00a0nm, respectively.As-prepared catalysts of Ni2@rGO8, Ni4@rGO6, and Ni6@rGO4 were also characterized by SEM. As shown in Fig.\u00a02\n, it can be seen that the Ni NPs are uniformly anchored on the surface of rGO for all samples. The average particle diameter is around 11.5\u00a0nm for Ni2@rGO8, 13.5\u00a0nm for Ni4@rGO6, and 16.5\u00a0nm for Ni6@rGO4, suggesting that the particle size increases with the increasing of Ni loading amount. It is worthy of noticing that, after an exposure in air for a week, Ni NPs doped on the rGO were not oxidized, suggesting that there is a strong interaction between Ni NPs and rGO.The catalyst of Ni4@rGO6 was further characterized by TEM (Fig.\u00a03\n). TEM images show that Ni NPs are uniformly spread over the rGO without aggregation even if the Ni loading amount is up to \u223c66\u00a0wt.%. The diameter of Ni NPs is about 15\u00a0nm. Such superior confinement should be attributed to rGO, which can serve as a support for in-situ formation of Ni NPs and prevent the aggregation and growth during the following heat treatment [39]. In addition, the high-resolution TEM (HRTEM) shown in Fig.\u00a03b demonstrates lattice fringes with interplannar distance of 0.205\u00a0nm, 0.177\u00a0nm and 0.221\u00a0nm, which corresponds to the lattice planes of Ni (111), Ni (200) and graphitic carbon (100), respectively. Besides Ni and C, NiO (111) with interplannar distance of 0.241\u00a0nm was also observed. The present of NiO may be due to the oxidization of Ni during the TEM sample preparation. It is observed that the rGO appears in the surrounding of Ni and leaves the Ni facets with high activity exposed [42], which can enhance the \u201csynergistic effect\u201d.To verify Ni loading amount of Ni@rGO, TG measurement was adopted. The as-prepared samples were heated to 900\u00a0\u00b0C at a rate of 5\u00a0\u00b0C\u00a0min\u22121 under high pure air atmosphere. As a result, the rGO would be oxidized to carbon dioxide and released. Ni would be oxidized to NiO and remained as a residual with the increasing temperature. The Ni loading amount in Ni2@rGO8 was calculated to be 45\u00a0wt.% (Fig. S2), 66\u00a0wt.% for Ni4@rGO6, and 77\u00a0wt.% for Ni6@rGO4. It is worth noting that the Ni loading amounts are higher than the designed values, which is due to the loss of oxygen-containing functional group, i.e., carboxylic, hydroxyl, and carbonyl during the reduction process.After the preparation of the catalysts of Ni2@rGO8, catalyzed MgH2 was then made by reactive ball milling method, which is to mill the catalysts and MgH2 together under a hydrogen pressure of 1 MPa. In order to research the effect of milling time on the hydrogen storage properties, various milling time of 2\u00a0h, 5\u00a0h, 10\u00a0h, and 20\u00a0h were applied, and the XRD patterns of the MH-Ni2@rGO8 with different milling time are shown in Fig.\u00a04\n. The peaks of all samples mainly correspond to MgH2. However, the peaks of Ni phase were not found in the XRD patterns of all samples. In addition, with the increase of milling time, the diffraction peaks become wider and weaker, indicating that the crystallite size decreases. The grain size of MgH2 was estimated by Scherrer equation, which shows that the grain size decreases in an order of 16.0\u00a0nm, 13.8\u00a0nm, 11.9\u00a0nm, 9.4\u00a0nm with the increase of milling time. Decreased grain size of MgH2 would provide more diffusion channel of hydrogen, leading to an improvement of hydrogen sorption kinetics [29].The morphology of as-milled MH-Ni2@rGO8 for different time was further observed by SEM. As shown in Fig. S3, big particles (> 2\u00a0\u03bcm) can be clearly observed in the sample of MH-Ni2@rGO8-2h. While the milling time is increased to 5\u00a0h and 10\u00a0h, the particle size is noticeably decreased. However, when further prolonging the milling time to 20\u00a0h, the particle size becomes much larger, and flaky particles appeared. It is noted that the variation trend of particle size observed in SEM images is different from that of grain size calculated by XRD results, which indicates that the grain size could be reduced by increasing the milling time. However, powder particles may be much bigger due to the welding and aggregation during the long milling process.The dehydrogenation properties of MH-Ni2@rGO8 with different milling time were investigated by DSC at a constant heating rate (5\u00a0\u00b0C\u00a0min\u22121). As shown in Fig.\u00a05\n, both the onset temperature (Tonset) and peak temperature (Tpeak) of dehydrogenation for all MH-Ni2@rGO8 samples shift to low temperature compared with pure MgH2. For the samples of Ni2@rGO8 catalyzed MgH2, The Tpeak of hydrogen desorption is increased in an order of Mg-Ni2@rGO8-5h (280\u00a0\u00b0C), Mg-Ni2@rGO8-10h (290\u00a0\u00b0C), Mg-Ni2@rGO8-2h (295\u00a0\u00b0C), and Mg-Ni2@rGO8-20h (306\u00a0\u00b0C). It is obvious that 5 h is the best milling time for hydrogen desorption in our experimental condition. The desorption temperature reduces \u223c 98\u00a0\u00b0C compared to pure MgH2. The hydrogen desorption result is in accordance with the particle size of ball milled samples. When the milling time is short, the particles are not fully ground and still big; when the milling time is too long, the particles will be welded due to the strong mechanical energy input. The reduced particle size leads to the decreasing of diffusion distances for H atom and enhances the hydrogen absorption/desorption kinetics [39]. As shown in Fig. S3, the particle size of MH-Ni2@rGO8-2h (about 5\u00a0\u03bcm) is bigger than that of MH-Ni2@rGO8-5h (about 1\u00a0\u03bcm). While the milling time increases to 20\u00a0h, large particles (> 2\u00a0\u03bcm) can be easily observed. Therefore, with the increase of milling time, the Tpeak of hydrogen desorption decreases firstly and then increases.The hydrogen absorption/desorption kinetics of composite MH-Ni2@rGO8 ball milled for various times was then studied by volumetric method. The initial hydrogen pressure during absorption and desorption is 3\u00a0MPa and 0.0004\u00a0MPa, respectively. The hydrogenation curves at 200\u00a0\u00b0C are plotted in Fig.\u00a06\na, and the dehydrogenation curves at 300\u00a0\u00b0C are plotted in Fig.\u00a06b. The results show that the composite of MH-Ni2@rGO8-5h has the fast hydriding/dehydriding rate. It can absorb 5.6\u00a0wt.% H2 within 60\u00a0s at 200\u00a0\u00b0C and 4.8\u00a0wt.% H2 within 30\u00a0min at 100\u00a0\u00b0C (inset of Fig.\u00a06a). For the desorption properties, MH-Ni@rGO-5h can desorb 6.2\u00a0wt.% H2 in 20 min at 300\u00a0\u00b0C and 6\u00a0wt.% H2 in 60\u00a0min even at 280\u00a0\u00b0C (inset of Fig.\u00a06b). When the temperature reaches to 300\u00a0\u00b0C, all the Ni@rGO-containing samples reach 95% of their maximum dehydriding capacity in 20 min, indicating that faster hydrogen desorption kinetics are obtained at higher temperature. The hydrogen desorption behavior measured by volumetric method is the same as that by DSC. Fig.\u00a06 shows that the increase of the ball milling time would lead to a negative influence on the hydrogen storage kinetics when the milling time is longer than 5\u00a0h. The composite of MH-Ni2@rGO8-20h can only absorb 3.9\u00a0wt.% H2 within 30\u00a0min at 100\u00a0\u00b0C.Above results showed that the composite of MH-Ni@rGO with a milling time of 5\u00a0h had the best hydrogen storage properties. Thus, the milling time of 5\u00a0h was used for further study about the catalytic effect of Ni@rGO with different Ni loading amount on the hydrogen storage performance of MgH2. Fig.\u00a07\n presents the DSC profiles of the as-milled composite of MH-Ni2@rGO8, MH-Ni4@rGO6 and MH-Ni6@rGO4. The MH-Ni4@rGO6 displays the lowest desorption peak temperature (Tpeak\u00a0=\u00a0259\u00a0\u00b0C) in comparison of MH-Ni2@rGO8 and MH-Ni6@rGO4. Furthermore, it is worth to note that the onset temperature of MH-Ni4@rGO6 (Tonset\u00a0=\u00a0190\u00a0\u00b0C) is much lower than that of MH-Ni2@rGO8 (Tonset\u00a0=\u00a0240\u00a0\u00b0C) and MH-Ni6@rGO4 (Tonset\u00a0=\u00a0270\u00a0\u00b0C). In addition, MH-Ni6@rGO4 shows two desorption peaks at higher temperature, one is at 285\u00a0\u00b0C and the other is at 320\u00a0\u00b0C, which indicates that the excessive Ni NPs may aggregate on the rGO and weaken the interaction between Mg and rGO. Obviously, too much Ni loading would degrade the catalytic effect.\nFig.\u00a08\n gives the hydrogen absorption and desorption kinetics curves of MH-Ni2@rGO8, MH-Ni4@rGO6, and MH-Ni6@rGO4. The MH-Ni4@rGO6 can absorb 3.7\u00a0wt.% H2 at 100\u00a0\u00b0C in 10 min, higher than that of MH-Ni6@rGO4 (3.0\u00a0wt.%) and MH-Ni2@rGO8 (2.7 wt.%). The dehydrogenation kinetics is further improved in the sample of MH-Ni4@rGO6. At 300\u00a0\u00b0C, the MH-Ni4@rGO6 can desorb 6.1\u00a0wt.% H2in15\u00a0min, while the MH-Ni2@rGO8 needs 20\u00a0min to reach the same capacity. Moreover, the MH-Ni6@rGO4 can only desorb 5.7\u00a0wt.% H2, which is due to the higher Tpeak (320\u00a0\u00b0C).To facilitate comparison, representative hydrogen absorption/desorption data for MgH2 system (with metal/carbon catalysts, prepared by ball milling method) are summarized in Table 1\n\n[29,32,37,43,44,45]. Obviously, the Ni4@rGO6 shows outstanding catalytic efficiency in enhancing the ab/dehydrogenation kinetics of MgH2. In comparison with other reported systems, the sample of MH-Ni4@rGO6 is competitive in the hydrogen absorption at low temperature (100\u00a0\u00b0C) and hydrogen desorption at 300\u00a0\u00b0C.The composite of MH-Ni4@rGO6 exhibits excellent hydrogen absorption and desorption properties. The PCT measurements of hydrogen absorption and desorption for the sample were performed at 300\u00a0\u00b0C, 320\u00a0\u00b0C, and 340\u00a0\u00b0C in a hydrogen pressure range from 0.05 to 2\u00a0MPa. Particularly, to ensure equilibrium, each absorption/desorption stage lasts at least 2\u00a0h. As shown in Fig.\u00a09\na, the plateau pressure was measured as 0.205, 0.320, and 0.499\u00a0MPa for absorption and 0.162, 0.270, 0.420\u00a0MPa for desorption at 300, 320, 340\u00a0\u00b0C, respectively. The corresponding van't Hoff plots (Eq. 1) for both hydrogen absorption and desorption are shown in Fig.\u00a09b.\n\n(1)\n\n\nln\nP\n=\n\n1\nT\n\n\n(\n\n\n\u2212\n\n\u0394\n\nH\n\nR\n\n)\n\n+\nC\n\n\n\nWhere P is the H2 pressure, T is the temperature, \u0394H is the enthalpy, R is the gas constant (8.3145\u00a0J\u00a0mol\u2212\n1\u00a0K\u22121), and C is a constant. The C value equals to \u0394S/R, in which \u0394S refers to entropy.According to the fitting result, the hydride formation and decomposition reaction enthalpy value (\u0394H) are calculated to be \u221264.8 and 69.6\u00a0kJ\u00a0mol\u22121, respectively. The values are lower than the theoretical values of MgH2\n[38]. The results indicate that the addition of Ni@rGO can destabilize the MgH2, which might be a reason for lower onset dehydrogenation temperature shown in the DSC curves. Therefore, it can be concluded that the Ni@rGO can not only improve the absorption/desorption kinetics, but also change the thermodynamics.A number of kinetic models for the gas-solid reaction were adopted to analyze the evolution of kinetics, such as Johnson-Mehl-Avrami-Kolmogorov (JMAK) model [46,47], Chou model [48,49], etc. The classical JMAK model can well describe the hydrogenation and dehydrogenation of nucleation-growth-impingement mode. Thus, the improved kinetics of hydrogenation and dehydrogenation of MH-Ni4@rGO6 was further determined by the JMAK model (see Eq. 2) through fitting the absorption and desorption curves of MH-Ni4@rGO6 and the activation energy (Ea\n) can be calculated according to the Arrhenius equation (Eq. 3).\n\n(2)\n\n\nln\n\n[\n\n\u2212\nln\n(\n1\n\u2212\n\u03b1\n)\n\n]\n\n=\nn\nln\nt\n+\nn\nln\nk\n\n\n\n\n\n\n(3)\n\n\nk\n=\nA\nexp\n\n(\n\n\nE\na\n\n\nR\nT\n\n\n)\n\n\n\n\nWhere k is the reaction rate constant, n is the Avrami exponent of the reaction order, \u03b1 is the fraction transformed at time t, and A is the temperature-independent coefficient. The reacted fraction of 0.2\u00a0<\u00a0\u03b1\u00a0<\u00a00.8 was used in this study. Fig.\u00a010\n shows the JMAK plots for the absorption of the MH-Ni4@rGO6 at the temperature of 100, 150, 200\u00a0\u00b0C, and desorption at 280, 300, 320\u00a0\u00b0C. Generally, the rate-limiting process, growth dimensionality and nucleation behavior of the hydrides can affect the reaction order n. The n values of hydrogenation (0.91, 1.03 and 1.09) are close to 1 (Fig.\u00a010a), indicating that the hydriding reaction of the sample follows a diffusion-controlled mechanism [50,51]. There are numerous growth and nucleation scenarios consistent with a value of n\u00a0=\u00a01, including nucleation and growth along one-dimensional (1D) dislocation lines and thickening of cylinders, needles and plates. Jeon [52] pointed out that hydrogen atoms rapidly nucleate and accumulate along the defects and form a metal hydride layer in one dimension, followed by subsequent growth and thickening from the metallic core. Similarly, the n values are in close proximity to 1.5 for decomposition of MH-Ni4@rGO6 at 300 and 320\u00a0\u00b0C (Fig.\u00a010b). Thus, the phase transformation from MgH2 to Mg in this case exhibits a zero nucleation rate, which is consistent with previous report [28]. Besides, the Ea\n of MH-Ni4@rGO6 is calculated to be 47.6\u00a0\u00b1\u00a03.4\u00a0kJ\u00a0mol\u22121 for hydrogenation (Fig.\u00a010c) and 117.8\u00a0\u00b1\u00a03.4\u00a0kJ\u00a0mol\u22121 for dehydrogenation (Fig.\u00a010d). The value for hydrogenation is lower than that of MgH2-5\u00a0wt.% GNs (Ea\n of hydrogenation: 78.4\u00a0kJ\u00a0mol\u22121) [53], and the dehydrogenation value is also lower than pure MgH2 (157\u00a0kJ\u00a0mol\u22121) [37].Above results show that Ni4@rGO6 is an excellent catalyst to improve the hydrogen storage properties of MgH2. Generally, the uniformly distributed ultrafine Ni NPs could be beneficial to the decomposition of H2 and recombination of atomic hydrogen during hydrogen absorption/desorption [54]. Moreover, the graphene can also provide more nucleation sites for the alloy or hydride and hydrogen diffusion channel [39], exhibiting a \u201csynergistic effect\u201d with Ni. To gain a further insight into the catalytic mechanism, XRD measurement was carried on the samples of MH-Ni4@rGO6 at following five states: as-milled MH-Ni4@rGO6, after first dehydrogenation, after first rehydrogenation and after 8th de/rehydrogenation. As shown in Fig.\u00a011\n, XRD pattern of as-milled MH-Ni4@rGO6 sample exhibits diffraction peaks corresponding to MgH2 and Ni (Fig.\u00a011a), indicating that Ni has not reacted with Mg during ball milling process. After first dehydrogenation, only two phases can be detected, Mg and Mg2Ni (Fig.\u00a011b), implying that Ni has already reacted with Mg and transformed to Mg2Ni. During the next rehydrogenation and dehydrogenation cycles, except for the main phase transformation of Mg/MgH2, the phase transformation of Mg2Ni and Mg2NiH4 occurred (Fig. 11c\u2013e).It is well acknowledged that the XRD investigation can not give the information of trace amount of sample or amorphous sample. Thus, the evolution of Ni during the dehydrogenation process of MH-Ni4@rGO6 was also detected by TEM. As shown in Fig.\u00a012\na, the size of dehydrogenated MH-Ni4@rGO6 particles is around 200\u2013600\u00a0nm, and some small catalyst particles are anchored on the surface of matrix. Further HRTEM analysis (Fig. 12c\u2013d) shows that MH-Ni4@rGO6 sample was fully dehydrogenated to Mg and Mg2Ni with a crystallize size of smaller than 10\u00a0nm. Moreover, Ni NPs can't be found in the HRTEM images, implying that the Ni NPs were totally reacted with Mg and yielded Mg2Ni during the hydrogenation of MH-Ni4@rGO6, which agrees with the result of XRD (Fig.\u00a011). Moreover, it is worthy of note that, a metastable alloy of Mg6Ni was also identified by HRTEM, and evidenced by selected area electron diffraction (SAED) pattern (Fig.\u00a012b). It has been reported that Mg6Ni alloy can be formed in a Mg83Ni17 alloy due to the solute accumulation in solidification, and decomposes into Mg and Mg2Ni with a low velocity at the temperature of 300-350\u00a0\u00b0C [45]. In this study, the as-milled MH-Ni4@rGO6 sample was composed of MgH2 and Ni rather than Mg-Ni alloys (Fig.\u00a011). During the following dehydrogenation process, Mg6Ni and Mg2Ni alloys might be formed due to the entering of Ni atoms into the lattice of Mg. At high temperature, Mg6Ni alloy is not stable and transforms into Mg and Mg2Ni again. In addition, as shown in Fig.\u00a012d, Mg2Ni and rGO grains were clearly dispread in the surrounding of Mg, which indicates that the co-catalyst of Mg2Ni and graphene nanosheet may have a \u201csynergetic effect\u201d on the hydrogen storage properties of Mg.Hydrogen desorption kinetics of as-milled MH-Ni4@rGO6 and rehydrogenated MH-Ni4@rGO6 are presented in Fig.\u00a013\n. It can be clearly seen that the hydrogen desorption rate of rehydrogenated samples are faster than that of as-milled sample. According to the above XRD and HRTEM results, the only difference between as-milled and rehydrogenated sample is chemical surrounding of Ni. We believed that the change of dehydrogenation kinetics of MgH2 in the sample of MH-Ni4@rGO6 is due to the change of Ni: elementary Ni for as-milled sample and Mg2NiH4 for rehydrogenated samples. The result indicates that the in-situ formed Mg2Ni/Mg2NiH4 may have better catalytic effect than Ni.Interestingly, we found that Ni NPs in the sample of MH-Ni2@rGO8 seems to be more stable than Ni NPs in the sample of MH-Ni4@rGO6. As shown in Fig. S4, although Ni is not detectable in the as-milled MH-Ni2@rGO8, the diffraction peaks of Ni still present in the rehydrogenated samples even after 8 cycles. And Mg2Ni/ Mg2NiH4 is not appeared in all MH-Ni2@rGO8 samples, implying no reaction between Ni and Mg. For the sample of MH-Ni2@rGO8, the hydrogen desorption kinetics of as-milled sample is almost the same as that of rehydrogenated sample (Fig. S5). The results further proved that the in-situ formed Mg2Ni has more effective catalysis than Ni. It is well acknowledged that, Mg2NiH4 is easier to release hydrogen compared with MgH2. Therefore, the in-situ formed and uniformly dispersed Mg2NiH4 on the surface of rGO can serve as a \u201chydrogen pump\u201d to enhance the dehydrogenation kinetics [18,41]. The rGO could provide more active \u201ccatalytic sites\u201d and H \u201cdiffusion channels\u201d to reduce the dehydrogenation temperature and enhance the dehydrogenation kinetics [36], leading to a \u201csynergetic effect\u201d with Mg2NiH4. According to the results, it is assumed that there is a strong interaction between Ni NPs and rGO. When the amount of Ni is low, the binding may be strong enough to prevent the reaction between Ni and Mg/MgH2. However, if the amount of Ni is high, the interaction between Ni NPs and rGO will be weakened, therefore, the Ni NPs can react with Mg/MgH2 more easily. It is worth noting that too much Ni loading will weaken the interaction between Mg and rGO and degrade the catalytic effect.\n\n\n(1)\nNi@rGO with different loading amounts was synthesized by wet chemical method, and the average crystallites size of Ni for Ni2@rGO8, Ni4@rGO6, Ni6@rGO4 were calculated to be 8.5\u00a0nm, 13.1\u00a0nm, and 15.4\u00a0nm, respectively.\n\n\n(2)\nThe MH-Ni4@rGO6 composite absorbs 5\u00a0wt.% hydrogen in 20\u00a0min at 100\u00a0\u00b0C. And the composite shows enhanced dehydrogenation rate: it can release 6.1\u00a0wt.% hydrogen within 15\u00a0min at 300\u00a0\u00b0C. The activation energy for the rehydrogenation of MH-Ni4@rGO6 is 47.6\u00a0\u00b1\u00a03.4\u00a0kJ\u00a0mol\u22121. Hydride formation and decomposition reaction enthalpy (\u0394H) are determined to be \u221264.8 and 69.6\u00a0kJ\u00a0mol\u22121, respectively, indicating a little thermodynamic change for the composite.\n\n\n(3)\nWe found that the in-situ formed Mg2Ni/ Mg2NiH4 exhibits better catalytic effect than Ni. Ni couldn't react with Mg due to the strong interaction between rGO and Ni NPs when the loading amount of Ni is low.\n\n\nNi@rGO with different loading amounts was synthesized by wet chemical method, and the average crystallites size of Ni for Ni2@rGO8, Ni4@rGO6, Ni6@rGO4 were calculated to be 8.5\u00a0nm, 13.1\u00a0nm, and 15.4\u00a0nm, respectively.The MH-Ni4@rGO6 composite absorbs 5\u00a0wt.% hydrogen in 20\u00a0min at 100\u00a0\u00b0C. And the composite shows enhanced dehydrogenation rate: it can release 6.1\u00a0wt.% hydrogen within 15\u00a0min at 300\u00a0\u00b0C. The activation energy for the rehydrogenation of MH-Ni4@rGO6 is 47.6\u00a0\u00b1\u00a03.4\u00a0kJ\u00a0mol\u22121. Hydride formation and decomposition reaction enthalpy (\u0394H) are determined to be \u221264.8 and 69.6\u00a0kJ\u00a0mol\u22121, respectively, indicating a little thermodynamic change for the composite.We found that the in-situ formed Mg2Ni/ Mg2NiH4 exhibits better catalytic effect than Ni. Ni couldn't react with Mg due to the strong interaction between rGO and Ni NPs when the loading amount of Ni is low.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Natural Science Foundation of China (Grant No. 51671118), the research grant (No. 16520721800 and No. 19ZR1418400) from Science and Technology Commission of Shanghai Municipality. The authors gratefully acknowledge support for materials analysis and research from Instrumental Analysis and Research Center of Shanghai University.Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jma.2019.06.006.\n\n\nImage, application 1\n\n\n\n", "descript": "\n                  Uniform-dispersed Ni nanoparticles (NPs) anchored on reduced graphene oxide (Ni@rGO) catalyzed MgH2 (MH-Ni@rGO) has been fabricated by mechanical milling. The effects of milling time and Ni loading amount on the hydrogen storage properties of MgH2 have been investigated. The initial hydrogen desorption temperature of MgH2 catalyzed by 10\u00a0wt.% Ni4@rGO6 for milling 5\u00a0h is significantly decreased from 251\u00a0\u00b0C to 190\u00a0\u00b0C. The composite can absorb 5.0\u00a0wt.% hydrogen in 20\u00a0min at 100\u00a0\u00b0C, while it can desorb 6.1\u00a0wt.% within 15\u00a0min at 300\u00a0\u00b0C. Through the investigation of the phase transformation and dehydrogenation kinetics during hydrogen ab/desorption cycles, we found that the in-situ formed Mg2Ni/Mg2NiH4 exhibited better catalytic effect than Ni. When Ni loading amount is 45\u00a0wt.%, the rGO in Ni@rGO catalysts can prevent the reaction of Ni and Mg due to the strong interaction between rGO and Ni NPs.\n               "}
{"full_text": "Propylene is one of most important petrochemical products, as it could build huge variety of chemical commodities. In traditional petrochemical technology network, cracking technology undertake most supply of propylene. Nowadays, more than three-fourths propylene is produced as a by-product by naphtha steam cracking and fluid catalytic cracking (FCC) (Bhasin et\u00a0al., 2001; Hu et\u00a0al., 2019; James et\u00a0al., 2016; Nawaz, 2015; Wang et\u00a0al., 2020; Al-Douri et\u00a0al., 2017). However, the growth requirement of propylene cannot be satisfied. Facing challenges from supply and demand side, it is urgent to improve novel propylene generation technology. In recent years, a lot of attempts have been adopted to give traditionally cracking technology more flexibility and increase yield of propylene (Alabdullah et\u00a0al., 2020). Coupling catalytic and thermal cracking process has been got attention, which deepen the cracking degree of reactant and widen feeding to wide-range oils. A novel cracking technology called DCC was developed by Sinopec Research Institute of Petroleum Processing. Base on FCC technology, this technology further takes higher temperature and partial pressure of steam, which deepen cracking degree and thus produce more propylene (Akah and Al-Ghrami, 2015). However, the selectivity of propylene in these technologies is still poor and oil-feeding would cause serious energy-consumption (Blay et\u00a0al., 2017). Some novel technologies called on-purpose propylene production technology has been attractive alternatives to traditional cracking process, such as oxidative coupling of methane, methanol to propylene, Fischer-Tropsch synthesis and direct dehydrogenation of propane (PDH). Compared with the traditional process, they have highly selectivity and don't need oil-feeding. (Hu et\u00a0al., 2019). Recently, PDH due to cheap and widely available raw material (propane) receives much attention, this transformation from gas fuel (propane) to chemical product (propylene) also indicated huge economic profits. Thus, PDH is widely considered to the most promising propylene production technology in the future.For improving industrial sufficiency of PDH, developing eco-friendly, cheap and active catalysts play the core role. At present, Pt and Cr-based catalysts are successfully industrialized, and various kinds of (Ga, V, Zn, Zr, Ni, Co, Fe, Al, Sn, Mo and other noble metals) the catalysts have been studied and reported widely (Hu et\u00a0al., 2019; Nawaz, 2015; Sattler et\u00a0al., 2014). In order to provide a scientific understanding on these numerous catalysts, researches had made a lot of excellent reviews. Sattler et\u00a0al., 2014 classified common PDH catalysts by element and make a very detailed review on their reaction mechanism, promoter and support. Hu et\u00a0al., 2019 further widened types of catalyst and classified them into three categories: metal, metal oxide and carbon catalysts by an intuition of overall structure of materials. Chen et\u00a0al. (2020) subdivided the existing metal and oxide catalysts based on the active site structure. However, the structure-performance relationship is rarely summarized and explained in basic principles of chemistry. For the oxide catalysts, most reviews are classified by elements (Hu et\u00a0al., 2019; Nawaz, 2015; Sattler et\u00a0al., 2014). In addition, the role of adsorbed gas and coke product on activity is also lacked summary. Therefore, the structure-performance relationship in propane dehydrogenation still needs to be outlined systematically.In this review, we summarize theoretical and experimental researches achievement of catalysts in recent years. Reaction mechanism and nature of active sites are discussed in detail assisted by researches on well-defined model catalysts and density functional theory (DFT) calculations. Furthermore, we outline the effect of supports and promoter on structural and electronic properties of catalysts as well as their effect on catalytic performance. Coke formation, as the main culprit causing deactivation, its deposition mechanism and structure-activity relationship is highlighted. In the end, the influence of adsorption of co-feeding gas is introduced to emphasize importance of process intensification. We hope this paper would contribute rationally design and optimization PDH catalyst in the future.For understanding PDH processes intuitively, Fig.\u00a01\n displays a simplified process of main and side reactions in PDH reaction. Formation of equimolar proplyene from propane is the main reaction of PDH. Side reactions include cracking, hydrogenolysis and coke deposition. Some other side reaction such as oligomerization, cyclization also happened but occupy a small part. Cracking includes a group reaction of cleavage hydrocarbons to two smaller alkene or alkane. Their pathway in PDH includes two mechanisms: thermal cracking and catalytic cracking. Coke deposition refers to formation various carbon-rich hydrocarbons or macromolecule solid carbon though deep dehydrogenation. Due to its gas-solid two-phase reaction, thus it is not limited by reaction equilibrium in PDH. Both catalytic cracking and coke deposition are directly catalyzed through active sites or Lewis/Bronsted acidic sites. Specially, hydrogenolysis reaction initiate on the surface of metal, in which alkane would be converted into two smaller alkanes assisted by hydrogen (Sattler et\u00a0al., 2014).In the process of catalytic PDH, hydrogen atoms in propane would play a protecting role and prevent activation of C\u2013C bonds. DFT calculations also proved breaking the C\u2013C bond of alkane when the C\u2013H bond is abundant are difficult, fracture energy of C\u2013C bonds and C\u2013H bonds would be lower with taking away hydrogen atoms(Yang et\u00a0al. 2010, 2012; Hu\u0161 et\u00a0al., 2020). Therefore, propylene is widely thought as the key intermediates of side product of side reaction rather than propane, and it has a series reaction like relationship between side reaction and main reaction. (Yang et\u00a0al., 2010; Nyk\u00e4nen and Honkala, 2011; Valc\u00e1rcel et\u00a0al., 2002; Yang et\u00a0al., 2010). The barriers of propylene desorption and dehydrogenation is often used as criteria to evaluate selectivity (Sun et\u00a0al., 2018).The main reaction of PDH is strongly endothermic and increases in total moles of gas, thus lower partial pressure benefits the appropriate conversion. The high temperature is also necessary for breaking the limits of thermal balance. To maximize propylene generation, temperature needed in PDH reaction is about 550\u2013700\u00a0\u00b0C (Yang et\u00a0al., 2012). It is noted that some other methods are used to break through the limit of thermodynamics, such as oxidative dehydrogenation (Atanga et\u00a0al., 2018) (the addition of CO2, O2 or other oxides into reaction gas to increase the conversion by removing H2), adsorbed membrane reactor(James et\u00a0al., 2016; Kim et\u00a0al., 2016) (remove H2 by membrane exchange) and partial alkane combustion, however, they all have their own limitations and have not been used in industrial propane dehydrogenation. (Kong et\u00a0al., 2020).In summary, for PDH, fracture of C\u2013H bonds is the focused question. Catalyzing cleaver of C\u2013H bonds is an important organic reaction. For heterogeneous catalytic cleavage of C\u2013H bond, its mechanism is often considered as oxidation mechanism or electrophilic mechanism and could be catalyzed by oxide or metal catalysts (Lillehaug et\u00a0al., 2004). For alkane dehydrogenation, after a long screen in the past, PtSn- and Cr-based/Al2O3 catalysts have been applied in industrial application(Nawaz, 2015). Other catalysts, such as Ga (Yu et\u00a0al., 2020; Im et\u00a0al., 2016), V (Wu et\u00a0al., 2017), Zn (Camacho-Bunquin et\u00a0al., 2017; Schweitzer et\u00a0al., 2014), Zr (Otroshchenko et\u00a0al., 2017), Sn (Wang et\u00a0al., 2016), Fe (Tan et\u00a0al., 2016) based-catalysts, are also widely studied and may be competitors for applications in the future. However, they all have their own limits in stability, selectivity and conversion. Deep understanding structure-activity relationship of metal and oxide catalysts is very important to improve catalytic performance. During reaction condition, the formation of coke deposition and the adsorption of gaseous species also have important influence on the structure of the catalyst.VIIIB metals have good performance in catalyzing dehydrogenation reaction. Among them, Pt has the best catalytic performance (Sattler et\u00a0al., 2014). Pt-based catalysts have excellent one-way stability, high selectivity and undoubtedly first-class activity, which provide great convenience to product separation and long-time continuous reaction. The industrialized PtSnK/Al2O3 catalysts have stability up to several hours and high selectivity, which has been used to provide high purity propylene for polypropylene industry (Sattler et\u00a0al., 2014). The whole process is powered by a semi-continuous interstage heater and regenerated dynamically in moving bed for maintaining a uniform propylene outlet flow rate. Another famous Pt-based catalyst developed by Uhde co-feed steam and thus further improve anti-coke stability (Nawas et\u00a0al., 2015). However, high-level price still limits widely utilization of Pt-based catalysts in PDH. Pt has almost 100 times higher price than elements used in oxide catalysts. In general, loading of Pt is need to be controlled lower than 1\u00a0wt% to maintain its economic benefits. Recently, the latest platinum-based catalyst of UOP even has only a 0.3w%Pt loading (Ji et\u00a0al., 2020). Cost-efficient has been the key challenges for Pt-based catalysts. At the same time, pure Pt-based catalysts always have low selectivity and suffered by coking and sintering, it is necessary to improve the supports, promoters and synthesis methods (Liu and Corma, 2018) (Liu et\u00a0al., 2019). Except Pt, other VIIIB metal catalysts due to their lowly intrinsic selectivity or poor activity are less used in PDH. However, some of them have higher C\u2013H cleaver activity and lower price than Pt, this gives them great potential in future (He et\u00a0al., 2018).For metal catalysts, the whole PDH reaction could be catalyzed by metal cluster from multiple atoms to even single metal atom theoretically (Yang et\u00a0al., 2011). A whole adsorption diagrams during dehydrogenation and deep dehydrogenation are shown in Fig.\u00a02\na. Reaction mechanism of PDH in metallic catalysts is widely recognized as a reverse Horiuti-Polanyi (HP) mechanism, which could be divided into four steps(Sattler et\u00a0al., 2014):Propane would firstly adsorb on surface of metal thought a weak physisorption drove by van der Waals force. (Lian et\u00a0al., 2018; Yang et\u00a0al., 2010). Then, two adjacent C\u2013H bonds on propane are successively dissociated by two adjacent metallic atoms. Finally, propylene and H2 desorbed from metal surface. Two C\u2013H cleaver steps always have the highest energy barrier in this mechanism, thus PDH is also widely considered as a surface reaction-controlled process follows Langmuir\u2212Hinshelwood kinetics.While main reaction could occur on single metallic atom, undesired reaction step such as deep dehydrogenation and cleaver of C\u2013C bonds in PDH need more adjacent active sites. (Yang et\u00a0al., 2011). Hydrocarbons adsorbed on metal surface would more like retain their alkane structure(Yang et\u00a0al., 2010). For example, propylene adsorb on Pt (111) face though di-\u03c3 bonds, these bonds would still keep their SP3 hybridization structure. While the hydrogen atoms continuously dehydrogenate from propane, in order to maintain original coordination morphology of propane, more consecutive sites are favored(Lian et\u00a0al., 2018). Pure metal catalysts due to their multiple sites always would suffer severe side reaction.(Pham et\u00a0al., 2016). Generally, 3-follow active sites are considered to be the smallest site that could catalyze the side reactions effectively(Zhao et\u00a0al., 2015) Purdy et\u00a0al. (2020) synthesized a variety of Pd-based catalysts alloyed with different promoting metals including Zn, Ga, In, Fe and Mn. As showed in Fig.\u00a02b, Pb atoms in PbZn and PbIn alloy nanoparticle catalysts only exist as the isolated Pb sites and thus have the highest selectivity. Pb3Fe and Pb3Mn alloy have relatively lower selectivity because of the presence of three-fold Pb sites. For Pb3Ga alloy, although 3-fold Pb sites exist, the positive triangle structure of threefold Pb ensemble is distorted by alloying, so it has higher selectivity than Pb3Fe and Pb3Mn. This result reveals the importance of site continuity for selectivity.The coordination environment of metal atoms also plays important role. Defects in end of atomic arrangement, such as the edges or corners, have low coordination number and high electron deficiency. These unsaturated coordination metal atoms strongly adsorb reactants and is more preferred to dehydrogenation and deep dehydrogenation. Size or structure of nanoparticles would affect amounts of defect directly. Zhu et\u00a0al. (2015) synthesized a serious Pt/Mg(Al)O catalysts with different particle size form 1\u00a0nm\u201310\u00a0nm for PDH. They found sample with medium particle size shows the highest propylene formation rate. When particle size increases, propylene selectivity increase but conversion rate of propane decreased, which result a volcano curve relationship between particle size and propylene formation rate showed in Fig.\u00a02c. Author think this structural-sensitivity-like phenomenon is attributed to the surface of small platinum particles have more unsaturated step atoms which not only have a higher dehydrogenation activity but also stronger C\u2013C cleavage tendency than terrace atoms (Nykanen et\u00a0al., 2013).In sum, the continuity and coordination environment of metal active sites play important role in catalytic performance. Nature of active sites in metal catalysts could be regulated by the interaction of other components such as promoters and supports. Addition of second metal could alloy with active metal, resulting metal-metal bonds could adjust catalytic performance. Support provides adsorption sites and surface area available for dispersing metal particles. In some cases, interfacial area between support and metal also plays complex electronic and geometric interaction. Next two parts, we would outline these structure-activity relationships in detail.Pure metal catalysts always have low selectivity and suffered by serious coking and sintering. Alloyed with second promoting metal could effectively modify catalytic performance from ensemble and electronic effect. Sn is the most used promoter in Pt-based catalyst and famous as industrial promoter used in commercial PtSnK/Al2O3 catalysts. It has been widely reported that the addition of Sn could restrain side reaction and improve catalytic conversion (Pham et\u00a0al., 2016; Wang et\u00a0al., 2019). In-situ characterization techniques have identified transformation of chemical state of Sn from oxide state to metallic state and Sn atoms move into platinum lattice forming Pt\u2013Sn alloy during the PDH process (Deng et\u00a0al., 2014; Iglesias-Juez et\u00a0al., 2010; Kaylor and Davis, 2018). However, microscopic interaction between Sn and Pt in nanoparticle is still illusive due to complexity of real alloying metal particles. Ensemble effect and electronic effect has been used to explain the role of Sn. Electronic effect thinks Sn atoms would play its role as an electronic donor and increase the electronic density of 5d band of platinum atoms, enrichment of this electronic state benefit rapid desorption of proplyene due to repel force between electron-rich \u03c0 bonds in propylene and metal surface (Deng et\u00a0al., 2018; Long et\u00a0al., 2016). Ensemble effect think non-reactive Sn atoms would dilute large platinum ensembles into small clusters by formation of alloy and thus increased the selectivity of the reaction due to structural sensitivity of side reaction in PDH (Zhu et\u00a0al., 2014; Zhu et\u00a0al., 2017).For explaining role of Sn, the researchers have made a lot of effort in modeling ideal surfaces and DFT calculation. Yang et\u00a0al. (2012) performed a DFT research of dehydrogenation of propane on Pt, Pt2Sn1 and Pt3Sn1 (111) face for explaining the ensemble and electronic effects of Sn in various surface of Pt\u2013Sn alloy. Authors found that the Sn atom causes a downshift of d-band center of adjacent platinum atoms, which results in the weaker bond between carbon atoms of hydrocarbon and Pt atoms. However, downshift of d-band center also improves energy barrier of dehydrogenation step, thus lead to decrease in conversion. Results of calculation also indicate Pt3Sn bulk is the most suitable alloying surface. Another explanation think Sn atoms would preferentially cover and deactivate Pt atoms in edge, corner which are electron deficient and are apt to catalyze side reaction (Virnovskaia et\u00a0al., 2007). Nyk\u00e4nen and Honkala, 2013 compared the PDH performance on step sites of pure Pt (211) and Pt3Sn (211) face in which Sn atoms located on the step edge of platinum based on DFT. In Fig.\u00a03\na, as the Pt (211) step sites with coordination unsaturation would adsorb propylene tightly, after decorated by Sn, Pt3Sn (211) edge would have a obviously weaker bond with propylene. Sn atom located at edge significantly inhibited deep dehydrogenation and weaken propylene adsorption, and improve selectivity. For understanding catalytic process in experiment, well-defined Pt\u2013Sn alloy surface for experimental verification are also ongoing to provide a better understanding. Zhu et\u00a0al. (2014) synthesize a serious of Sn surface-enriched Pt\u2013Sn nanoparticles though a surface organic chemistry (SOMC) method and support them on MgAl2O4. As shown in Fig.\u00a03b, with the increase of Sn usage, the surface of nanoparticles gradually appears an enrichment of Sn and form spherical shell-like structure, while the particle size remains unchanged. With the process of Sn surface enrichment, the selectivity and conversion gradually increase. It is directly proved that isolation effect of Sn has an important effect on the PtSn catalysts.Although the electronic and ensemble effect have successfully explained many experimental phenomena in PtSn catalysts, these studies are only limited to the surface of PtSn nanoparticles. Under really experimental conditions, the interaction between Sn and Pt is far beyond surface interaction, the difference between the inner and outer layers in alloy could still cause large difference in reaction performance (Wu et\u00a0al., 2018). Ye et\u00a0al. (2020) synthesized a series of PtSn@Pt/SiO2 catalysts with a well-defined Pt1Sn1 surface. With the increase of the number of subsurface Pt\u2013Sn coordination bonds, the selectivity almost remains unchanged while the TOR continue increasing. Therefore, author considered that the isolated effect by surface Sn atoms mainly affect selectivity, while the TOR is more related to electronic effect of subsurface Sn. The structure-performance relationship is shown in Fig.\u00a03d. It can be seen that the surface and internal structure of Pt\u2013Sn catalyst have important effects on the catalytic performance.In addition, Sn element has many other effects on Pt-based catalysts. Pham et\u00a0al. (2016) studied the structural changes of Sn of PtSn/\u03b3-Al2O3 in regeneration-reaction process and found interaction between Sn and Al2O3 would play an important role. After the process of oxidative regeneration, Sn would de-alloy from Pt\u2013Sn nanoparticles and anchor on the Al2O3 support as Sn atoms or clusters. These adsorbed Sn sites provide nucleuses for recreation Pt\u2013Sn nanoparticles. Therefore, Pt\u2013Sn catalysts would recover their own high dispersion after redox process used Al2O3 as support, while the size of pure-Pt catalysts will continue increasing after several regenerations. Although the traditional idea is that Sn can only improve the activity of catalyst in the form of alloy, recent studies have shown Sn in oxide state could also improve the catalytic performance and like a strong metal-support interaction (SMSI). Deng et\u00a0al. (2018) reported a SMSI effect between Pt and SnOx in Pt\u2013Sn/SiO2 catalysts. Traditionally, Sn is pretreated by H2 to form an alloy structure though ensemble and electronic effect, author found Sn in Pt\u2013SnO2/SiO2 pretreating by O2 and N2 would play similar role, even all Sn species are only existed as SnO2. A more detailed structure modeling and comparison of catalytic performance is shown in Fig.\u00a03c. Compared with the Pt catalyst without Sn, the electronic state of Pt with SnOx was effectively increased after pretreatment of nitrogen, resulting in the improvement of selectivity and conversion. Identify promoting role of Sn still needs to further researches.Other metals, such as Ga (Bauer et\u00a0al., 2019), Zn (Rochlitz et\u00a0al., 2020) and In (Xia et\u00a0al., 2016), Cu (Ren et\u00a0al., 2018; Han et\u00a0al., 2014), V (Purdy et\u00a0al., 2020), Co (Cesar et\u00a0al., 2019), Mn (Wu et\u00a0al., 2018; Fan et\u00a0al., 2020) and Fe(Cai et\u00a0al., 2018), are also widely used in Pt-M catalysts (M stands for promoting metal) and their role of promoters are always explained like Sn. Nakaya et\u00a0al. (2020) synthesized a novel PtGa-Pb/SiO2 ternary alloy catalyst with isolated Pt atom structure. In the surface of PtGa nanoparticles, Pd atoms would selectively locate on threefold Pt sites driving by thermodynamic effect and only keep isolated Pt active sites. PtGa-Pb SAAC has higher stability and selectivity than the sample without Pb. X Sun et\u00a0al. (2018) synthesized a novel kind of PtCu Single-atom alloy catalyst (SAAC). By addition of Cu, Pt ensembles were dispersed to isolated single-atoms. PtCu SAA has a similar TOF but higher selectivity than pure Pt catalyst. DFT calculation shows that the Pt atom dehydrogenation activity on Pt\u2013Cu alloy is almost unchanged, but the propylene desorption ability increases significantly which prevent deep dehydrogenation and other side effects. By precisely controlling solid state transformation of Mn atom into Pt nanoparticles, Wu et\u00a0al. (2018) synthesize and identified Pt@Pt3Mn core-shell nanoparticle catalysts and pure Pt3Mn nanoparticle catalysts. The selectivity of Pt@Pt3Mn is effectively lower than that of Pt3Mn alloy. DFT calculation indicated Mn atoms in the subsurface would reduce the surface adsorption of propylene, thus inhibit side reaction. Interaction between these various promoters and platinum still needs more extensive study.Except Pt, other VIIIB metals could also be modified by second promoting metals. Ni, Co, Fe metals have higher dehydrogenation activity than Pt, but prone to generate cracking products or coke than producing propylene. (Chen et\u00a0al., 2020; Saelee et\u00a0al., 2018). Some other noble metals, such as Pd, Rh and Ru, have also been used in PDH because of similar electronic and geometric structures to platinum (Ma et\u00a0al., 2020; Purdy et\u00a0al., 2020; Natarajan et\u00a0al., 2020). Addition of appropriate promoters has potential to improve these essential shortcomings. He et\u00a0al. (2018) synthesized a Ni\u2013Ga nanoparticle catalyst supported by Al2O3 (70% delta, 30% gamma phase). This alloying NiGa catalyst had a high initial selectivity of 94% and long-term stability, while pure Ni has a poor selectivity approached to 0 and bad stability due to serious coking deposition. For studying the effect of Ga on Ni deeply, researchers further synthesized a serious of catalysts with different Ni:Ga surface ratio. Higher surface gallium content is main reason of the high selectivity in NiGa/Al2O3 catalyst. Raman et\u00a0al. (2019) studied performance of RhGa/Al2O3 catalyst in PDH. Addition of Ga would form solid intermetallic phases with Rh and improve catalytic performance. Continuing to increase amount of gallium would result in gradual formation of Rh\u2013Ga liquid metal solutions and a sharp rise in selectivity and activity will be observed while alloy is all liquid state. DFT showed present a synergistic effect between Ga and Rh, activation of propane is happened on single-atom Rh and has formed proplyene would diffuse and desorb from Rh to the Ga, and it may be the reason of high activity of isolated Rh atom.For metal-based catalyst, the role of the support is dispersing metal particles and avoiding the irreversible sintering at the high temperature of PDH. Support with high surface area is obviously benefit to disperse and stabilize uniform and ultra-fine metal particles. In addition, appropriate pore diameter which is matched to particle size would show higher resistance to sintering which is also called confinement effect. Because of the above property, aluminum and silicon-based supports with higher specific surface area and adjustable pore structure are widely used in metal-catalyzed PDH process, and ordered mesoporous materials and microporous molecular sieves have unique advantages in stabilizing nanoparticles though confinement effect. In terms of support-metal interaction, supports used in PDH are usually electrically inert and hardly reduction for avoiding undesired side reaction or structural collapse, coordination unsaturated defect on these nonreducible oxide support would provide strongly local-unsaturated adsorbed sites to anchor these metal particles (Ji et\u00a0al., 2020; Kwak et\u00a0al., 2009). Thus, adjusting the surface defected sites are important in improving metal particles dispersion and stability.Among aluminum-based materials, \u03b3-Al2O3 is the most common commercial support, which has low price, high thermal stability and strong resistance to abrasion (Sattler et\u00a0al., 2014). In the surface of \u03b3-Al2O3, coordinatively unsaturated pentahedral coordination Al3+ could anchor metallic component and prevent metallic nanoparticles aggregation (Gong and Zhao, 2019; Yu et\u00a0al., 2020; Kwak et\u00a0al., 2009). The amounts of unsaturated coordinated sites are closely related to the preparation method and morphology. Shi et\u00a0al. (2015) synthesized a novel PtSn/Al2O3 sheet catalyst for PDH. Rich-defected Al2O3 sheet would attribute to more unsaturated pentahedral Al3+ sites than commercial \u03b3-Al2O3, and thus effectively stabilize ultra-small raft-like Pt\u2013Sn clusters. PtSn/Al2O3 sheet catalyst has an extraordinary selectivity up to 99% and only suffered trace deactivation happened during 24h reaction test. Sheet-like structure also has a positive impact on mass transfer, and benefit better activity and selectivity at high space velocity. Gong and Zhao, 2019 synthesized a novel peony-like alumina nanosheet (Al2O3-MG) with richer pentahedral Al(\u2162) by glocuse-assisted hydrothermal method and supported PtSn catalyst on it. The strong interaction between pentahedral Al(\u2162) sites and PtSn nanoparticles improved anti-sintering ability, thus, PtSn/Al2O3-MG catalyst has an unexcepted stability and conversion.Although \u03b3-Al2O3 is good for achieving high dispersion of metal particles, however, acid sites on Al2O3 surface would catalyze undesired coke deposition and cracking reaction. Introducing basic promoters to adjust surface acidity is common solution. Addition of K, Na et\u00a0al. alkali metal would partially cover the strong acidic sites and result in the reduction of side reactions (Sattler et\u00a0al., 2014a). Other basic oxide such as Mg, Zn, Ca and some rare earth elements (Vu et\u00a0al., 2016; Im et\u00a0al. 2016) are also widely used for similar purpose. In addition to adjusting acidity, formation of MgAl2O4, ZnAl2O4 Spinel phase after addition of Mg, Zn would provide additional anchoring effect on metal particles by epitaxial metal-oxide interfacial caused by similar structure with Pt (111) face (Belskaya et\u00a0al., 2016). Ren et\u00a0al. (2018) systematically studied the promoting role of IB metals in Pt-M/MgAl2O4 and effect of MgAl-spinel on propane dehydrogenation. Compared with \u03b3-Al2O3, although the original size of particle in both supports are very similar, MgAl2O4 evidently provide a stronger interaction to Pt nanoparticles and showed a better reaction-regeneration performance in multiple regeneration. After the introduction of IB metal promoters (Cu, Ag, Au), selectivity and conversion of all Pt-M catalysts have improvement, among them Cu has the most preferred promoting effect. This may due to worse affinity with Pt of Ag and Au, and thus result in formation of amorphous alloy, while Cu would form stable Pt\u2013Cu intermetallic alloy with high stability and better dispersion. Some low melting-point oxides are also used in propane dehydrogenation to partially cover the acidic sites on the surface of alumina. Aly et\u00a0al. (2020) found that introducing B species into Pt/Al2O3 catalysts can a reduction of coke formation and side reaction. DFT calculation indicated formation of finely dispersed amorphous B2O3 on alumina, which covered stronger acid sites on Al2O3 and reduce unwanted side reaction. The improvement of alumina-based support still needs further more exploration and research.Except for aluminum-based materials, pure silica materials, especially pure silica zeolites, have been widely studied in the field of PDH because of their low acidity, high specific surface area and adjustable, homogeneous and ordered pore structure. However, inert property of silica material also leads weak adsorption capacity. At preparation process, due to the lack of strong electronic attraction, metal precursors could be often evenly distributed on the surface of silicon oxide (Fan et\u00a0al., 2020). At the reaction and regeneration process, due to the lack of strong adsorption sites, oxidized metal particles would suffer intensive sintering (Kaylor and Davis, 2018). Therefore, the research on silicon materials mainly focuses on improving the interaction between support and metal in both preparation and reaction-regeneration process.Creating the adsorbed sites on SiO2 materials by metal doping is a very common method to produce adsorption sites for metallic nanoparticle. Fan et\u00a0al. (2020) used MnOx modified mesoporous silicon nanoparticle as a support to disperse platinum. The modified of MnOx provide strong electron adsorption to Pt precursors thus result better dispersion of Pt. Moreover, metallic Mn produced from reduction also play promoting role though alloyed with Pt. DFT calculation shows that the formation of PtMn alloy not only promote the proplyene desorption, but also can keep a good activity of initial dehydrogenation. For zeolite support, although presence of aluminum atom provides additional anchored effect to metal particle, caused strong acidity would lead to serious coke deposition. Many studies have tried to exchange the framework structure of silica-alumina zeolite with other metals like Zn (Zhang et\u00a0al., 2015), Sn (Li et\u00a0al., 2017), Ti (Li et\u00a0al., 2017), Fe (Waku et\u00a0al., 2003) to obtain a relatively low acidity but keep stronger interaction. Zhang et\u00a0al. (2015) synthesized a Zn-ZSM-5 zeolite though use Zn precursor instead of Al precursor. Compared with the traditional Al-ZSM-5, the Zn-ZSM-5 has lower acidity which reduce coke deposition. At the same time, Zn also provide a strong interaction than Al-ZSM-5. Formation of Pt\u2013Zn nanoparticles also improve selectivity and conversion.Except for the introduction of heteroatoms, surface organometallic chemistry method (SOMC) also provides an optional method to directly anchoring stable metal nanoparticles by pure silicon material. Although silicon oxide materials due to its electric neutrality could not provide strong adsorption capacity for metal precursors in water through the electric attraction, hydroxyl groups on the surface of SiO2 can be connected with metallic organic precursors by SOMC and form highly dispersed even isolated metal sites (Xu et\u00a0al., 2019). Searles et\u00a0al. (2018) prepared a novel PtGa/SiO2 catalyst via grafting Pt and Ga precursor onto the surface of silica gel and form Ga and Pt single-sites, after reduction, they obtained homogeneous and ultrasmall Pt\u2013Ga alloying nanoparticles. This catalyst had amazing catalytic performance which high conversion (31.9%), selectivity (99%) by only usage of only 0.001g catalyst in a very high space velocity with ultrahigh stability and regeneration ability. Two different kinds of Ga species were observed in PtGa/SiO2: metallic Ga and remained single Ga3+ on SiO2. Metallic Ga formed bimetallic particles with Pt thus the selectivity and stability is improved. In addition, isolated Ga3+ on the surface of SiO2 is related to formation of strong Lewis acid sites, which may enable to the nucleation and stabilization of ultra-small Pt\u2013Ga nanoparticles. Another PtZn/SiO2 catalyst which has similar synthesis method was also prepared by Rochlitz et\u00a0al. (2020) PtZn/SiO2 catalyst has also a high conversion, selectivity, stability like mentioned-above PtGa/SiO2 catalyst. The formation of Pt\u2013Zn alloy is considered as an important reason to the high selectivity of PtZn/SiO2.Through dealumination of Si\u2013Al sieves, the silanol nest formed after dealumination on sieve has strong adsorption capacity than common silica material, these sites could effectively disperse metal precursor sites at the initial stage of preparation. Xu et\u00a0al. (2019a,b) further introduced this SOMC method into the synthesis of PtSn catalyst supported by dealuminated beta zeolite. Though directional interaction between organic functional group, isolated Sn atoms would localize in the framework of beta zeolite and prefer to form smaller PtSn clusters with Pt. PDH test showed that the Pt/Sn2.00-Beta catalyst had the highest conversion (50%) and selectivity (99%). It is worth noting that similar stable Pt clusters have been obtained by the same method on Y zeolite, which indicates that this method may be a general sintering inhibition method to Silica alumina zeolites. Ryoo et\u00a0al. (2020) co-impregnated Pt(NH3)4NO3 and nitrate ions of rare earth in the de-Ga molecular sieve. Silanol nest formed by dealumination would stabilize rare earth ions exist in the form of single atom and thus easy to be reduced. After reduction treatment at 700\u00a0\u00b0C in H2, rare earth metal-platinum alloy catalyst was obtained. This catalyst has high selectivity, conversion and shocking stability up to several days.Encapsulation of Pt-M clusters into micropores of zeolite by in-situ method or post-synthesis could also synthesize ultrasmall Pt alloyed cluster and effectively inhibit sintering due to confined effect. Wang et\u00a0al. (2020) synthesize a novel PtZn@S-1 catalyst via hydrothermal method. XPS found some of the Zn species exist as Zn2+ and located into lattice of S-1 zeolite, and others form Pt\u2013Zn alloys with Pt species. Ultrasmall Pt\u2013Zn alloys particles with high catalytic performance are encapsulated inside S-1 zeolite though in-situ synthesis. The confinement effect of S-1 channel effectively prevents the sintering of Pt\u2013Zn particles. Liu et\u00a0al. (2020) synthesized a novel K\u2013PtSn@MFI catalyst. Through XAS, TEM, author proved sub-nano Pt clusters(0.6\u00a0nm) are confined in the sinusoidal 10R channels of MFI. Interaction between Pt and Sn can be effectively controlled by adjusting the reduction method. Increasing the reduction temperature or time could can help Sn to enter into Pt clusters and provide promoting effect, thus increasing selectivity and slowing down deactivation rate. Under a condition similar industrial process, the K\u2013PtSn@MFI catalyst exhibited a selectivity of up to 97% and an initial conversion of 20% with only 3% deactivation in 70h.The traditional supports such as Al2O3 and SiO2 are electric inertia generally interact with metals by local defect sites. Although these defects have some anchoring effects, but they still failed to provide more effective electronic interaction with Pt. Therefore, some supports which provide strong metal-support interaction (SMSI) have received the much attention from researchers. The SMSI effect provides additional electronic interaction and modify interface structure between support and metal, thus it is expected to an interesting way to change catalyst performance. (Liu et\u00a0al., 2020). Jiang et\u00a0al. (2014) studied the promoting role of TiO2 on Pt/Al2O3 catalyst. After doping Ti into Al2O3, both selectivity and conversion have obvious improvement, despite Pt/TiO2\u2013Al2O3 and Pt/Al2O3 has similar particle size of about 2\u00a0nm. This phenomenon may be due to SMSI between TiO2 and Pt would increase electronic density of Pt atoms. Electron-rich Pt promote desorption of propylene and prevent side reaction. A high electronic density also weak attachment of coke precursor, so promote migration of coke from metallic nanoparticles to support. Liu et\u00a0al. (2017) synthesized a novel Pt/ND@G (graphene shell in nano diamond) catalyst which has high selectivity (90%) and shows slight deactivation in 100h. The electron transfer from defective graphene to platinum nanoparticles improved anti-sintering ability of Pt/ND@G. In addition, this electron transfer also increases electronic density of Pt surface and inhibit coke deposition and side reaction effectively.In regeneration process of PDH, treating catalysts with high temperature air would oxidize part Pt (0) to volatile Pt2+, Pt4+ species, and finally lead agglomerate (Kaylor and Davis, 2018; Pham et\u00a0al., 2016; Xu et\u00a0al., 2018; Uemura et\u00a0al., 2011). A strong interaction between Pt species and support under air condition is also important to inhibit sintering in regeneration process. CeO2 is well known for its special trapping platinum atom ability in high temperature. Xiong et\u00a0al. (2017) synthesis a novel PtSn/CeO2 catalyst. The strong interaction between Pt and CeO2 effectively minimized sintering of small Pt\u2013Sn clusters. Interestingly, during the oxide regeneration process, Platinum clusters supported on CeO2 would be dispersed into single-atoms. In reduction period, the dispersed Pt atoms will be self-reassembled to Pt\u2013Sn cluster and restore original dispersion, while in Pt/Al2O3 Pt particle would aggregate violently after regeneration. This self-assembly process is shown in Fig.\u00a04\n. Xu et\u00a0al. (2018) added \u03b2-Ga2O3 with different shapes to Pt/Al2O3 and studied their influence on propane dehydrogenation. PDH results suggest Ga2O3 can stabilize the particle size of platinum in air at high temperature. This may be because Pt particles could be captured by Ga2O3, which prevents Pt from being oxidized into PtOx large particle. Recent catalysts are shown in Table\u00a01\n.Oxide catalysts in PDH could be divided into supported catalysts and bulk catalysts. Compared with Pt, oxide catalysts have their own advantage such as significantly lower price and convenient regeneration process, while Pt catalysts need using dangerous Cl2 to redisperse large particles (Sattler et\u00a0al., 2014; Liu et\u00a0al., 2016; Zangeneh et\u00a0al., 2015). However, most oxide catalysts are suffered by serious coking deactivation and have much shorter one-way life than Pt catalysts, thus frequent regeneration is necessary (Nawaz et\u00a0al. 2015). In industry Catofin technology, CrOx-K2O/Al2O3 catalysts are regenerated every 10\u00a0min, several fixed bed reactors work in turn to keep a constant propylene outflow rate (Sattler et\u00a0al., 2014). Thus for oxide catalysts, stability during regeneration process always needs to pay more attention than one-way stability. Solid reaction or leaching in regeneration process are culprits of deactivation in regeneration process. For example, in industrial CrOx/Al2O3 catalysts, migration of Cr atoms into framework of Al2O3 would form inactive spinel phase and thus lead to irreversible deactivation. The inactivation process can be reduced by selecting supports and promoters properly. In addition, the intrinsic activity of the oxide catalyst is often tens of times lower than metal-based catalysts, which makes the oxide catalyst often need a high loading to achieve appropriate conversion.Catalytic mechanism of oxide catalysts in PDH is basis on metal-organic chemistry, unsaturated metal-oxide pairs are considered to be the source of active sites. After reduction activation process, Metal oxides would lose a part of oxygen atoms and produce unsaturated metal-oxygen sites, these unsaturated M-O pairs would tend to coordinate with carbon atoms and hydrogen atoms on propane and thus reduce activation energy of C\u2013H cleavage (Chen et\u00a0al., 2020; Estes et\u00a0al., 2016; Schweitzer et\u00a0al., 2014; Conley et\u00a0al., 2015). In Cr-based catalysts, coordination unsaturated Cr3+ and Cr2+ reduced from Cr6+ and Cr5+ have been widely confirmed as active sites by a series of in-situ characterization (Nawaz, 2015; Santhoshkumar et\u00a0al., 2009; Puurunen et\u00a0al., 2001; Gao et\u00a0al., 2019). Similar to Cr-based system, coordination unsaturated V3+ and V4+ reduced from V5+ are widely thought as active sites in V-based catalysts (Sokolov et\u00a0al., 2012; Langeslay et\u00a0al., 2018; Bai et\u00a0al., 2016; Zhao et\u00a0al., 2018; Liu et\u00a0al., 2016; Xie et\u00a0al., 2020; Rodemerck et\u00a0al., 2017). In some cases, the valence of oxides after reduction would not change such as Ga, Zn, but their coordination conditions are changed to unsaturation due to loss of oxygen atom (Sattler et\u00a0al., 2014). For Ga-based catalysts, tetrahedral (IV) coordination Ga3+ reduced from octahedrally (VI) coordinated Ga3+ are the main active sites. (Sattler et\u00a0al., 2014; Schreiber et\u00a0al., 2018\nChoi et\u00a0al., 2017; Kim et\u00a0al., 2017; Szeto et\u00a0al., 2018; Zheng et\u00a0al., 2005). Although in most oxide catalysts, increasing the reduction degree of the active site could greatly improve its activity due to increase of unsaturation, however, deep reduction is not necessarily beneficial to the oxide catalysts (Sun et\u00a0al. 2014, 2015; Dai et\u00a0al., 2020). For example, the active site of Co-based catalyst is considered to be unsaturated Co2+O. If CoOx is reduced to metallic Co nanoparticle, this will result in a sharp decrease in selectivity and stability (Dai et\u00a0al., 2020). In Zn-based catalysts, Zn metal formed by excessive reduction is inactive and easy to be sublimated and lost from the support due to its low melting point, resulting in irreversible deactivation.Reverse H\u2013P mechanism is also widely used to describe mechanism of oxide catalysts, but there is a little different in adsorption sites: while physical adsorption completed, propane would be dissociated to alkyl and hydrogen atom, alkyl adsorb on coordination unsaturated metal atoms, while hydrogen atoms adsorb on oxygen atom near these metal atoms. After that, the alkyl would undergo a \u03b2-hydrogen transfer and desorbed hydrogen would be adsorbed on metal ions, while propylene adsorbs on the top of metal ions by \u03c0-bond(Hu\u0161 et\u00a0al., 2020). Then, propylene and hydrogen are desorbed successively(Liu et\u00a0al., 2016; Estes et\u00a0al., 2016). This process has been showed on Fig.\u00a05\nb. The whole reaction is also considered as a surface reaction control mechanism and dehydrogenation step is rate-limited step (Xie et\u00a0al., 2020; Liu et\u00a0al., 2016). Although traditional model of oxide catalysts only includes metal-oxide pair M-O and follow HP mechanism, under different chemical conditions, active site structure and reaction mechanism may be different. (Olsbye et\u00a0al., 2005; Zhao et\u00a0al., 2018; Dixit et\u00a0al., 2018). DFT calculation is of great significance in predicting and interpreting influencing mechanism and detailed information of active sites in oxide catalysts. Zhao et\u00a0al. (2018) found that pretreating VOx/Al2O3 by hydrogen could converse VO bonds on VOx to V\u2013OH gradually, while pretreating with C3H8 would directly lead to the fracture of VO. Active sites with V\u2013OH structure have lower activity but slower deactivation (Fig.\u00a05a). With the increase of treating time of H2, both of the initial activity and deactivation rate would decrease. This also shows the complexity of active sites on surface of oxide. Although the active sites of most oxides are considered to be metal-oxygen pairs, Dixit et\u00a0al. (2018) made a systematic study in sites with different coordination environment and present of hydroxyl of PDH mechanism on (110) surface of Al2O3 by DFT. They also found a volcano relation between binding energy of dissociated H2 and TOF of PDH. Desorption of H2 in highest active sites is more suitable to be the rate-limiting step rather than traditional breaking of C\u2013H. AlIII\u2212OIII sites of hydroxylated Al2O3 (110) which has compromising Lewis acid/base ratio have highest active and follow a concerted mechanism. In addition to alkyl mechanism, mechanism participated with free radicals and carbocation have also been proved to have lower energy barrier in some cases (Mansoor et\u00a0al., 2018; Lillehaug et\u00a0al., 2004). Schreiber et\u00a0al. (2018) found traditional HP mechanism may not be suitable in Ga/H-ZSM-5 catalysts. They indicated while first step of dehydrogenation is still the common heterolytically dissociation process, the second step of dehydrogenation should happen with an intermediate state of carbenium rather than direct cleaver of C\u2013H in alkyl. These mechanisms are listed in Fig.\u00a05b.In sum, the active sites of the oxides are formed by metal-oxygen pairs after reduction, these sites are often coordination unsaturated. Which oxygen is reduced (M-O-M or M-O-S et\u00a0al., M is the active metal atom and S is the support), how many oxygen atoms are reduced, and whether forming adsorbed species such as hydroxyl decide the final micro-structure of active sites. In order to modify the reducibility of oxide catalysts, commonly used methods include regulating support and promoter. In addition, supports and promoters also affect the dispersion of active sites in oxide catalysts.Support plays a key role in oxide catalysts for PDH reaction. First and most obviously, support provide landing surfaces for dispersing active species and effect structure of active sites (Bai et\u00a0al., 2016). As the intrinsic activity of most oxide catalysts is much lower than that of Pt, a support with larger specific surface area is needed to increase the loading of oxide catalyst in order to achieve higher conversion. However, as increasing loading, polymerized structure of oxide would change from isolated sites to oligomer and finally crystals (Wu et\u00a0al., 2017) (Fig.\u00a06\n). The degree of aggregation not only determines amount of exposed active sites, but also plays decisive role in reduction tendency of oxygen atom and active structure during PDH process (Ruiz Puigdollers et\u00a0al., 2017; Liu et\u00a0al., 2016; Zhao et\u00a0al., 2019). A large number of studies in oxide catalysts focus on studying relationship between polymerization state and PDH catalytic performance. Isolated sites are generally considered to have special advantages. They not only have the maximum dispersion theoretically, but also prevent coke precursors contact each other to from graphitized coke (Rodemerck et\u00a0al., 2017; Liu et\u00a0al., 2020; Dai et\u00a0al., 2020; Hu et\u00a0al., 2015; Hu et\u00a0al. 2015, 2015). Supports with high specific surface area have obviously advantages in keeping them working as isolated sites. Rodemerck et\u00a0al. (2017) found that in the surface of VOx/MCM-41, isolated V sites have the highest activity and coke resistance. Although isolated V species have higher acidity than oligomerized counterparts, the coke precursors are hard to contact with each other and form coke, thus they would have both higher activity and stability in PDH. Synthesis method also plays an important role in dispersing oxide to isolated sites. Traditional impregnation method often leads to the uneven distribution of oxide species, some novel synthesis method such as hydrothermal one-pot method (Dai et\u00a0al., 2020), sol gel method (Hu et\u00a0al., 2018) and precursor modification has unique advantages. Zhao et\u00a0al. (2019) using Zn-based metal organic framework as precursor and synthesized highly dispersed ZnO@CN/S-1 catalyst. ZnO nanoparticles are highly dispersed and encapsulated into CN material. CN layers effectively hindered the loss of Zn in the high temperature due to physical wrapping effect. In addition, the electron interaction between CN materials and ZnO also promoted the desorption of propylene. This ZnO catalyst has excellent conversion, selectivity and stability.Evolution surface structure of oxide with the change of loading would also be largely determined by type of support, although some inert supports have large specific surface area, they could not disperse the active components well. Santhoshkumar et\u00a0al. (2009) found Cr is more likely to exist as isolated Cr (VI) species under low loading on the SBA-15. These isolated sites have highly activity and stability. However, with the increase of the loading amount, \u03b1-Cr2O3 gradually formed from isolated Cr species and result in a decrease of activity. On the contrary, although in the low loading Cr exists as oligomers on the Al2O3 with relatively low activity, \u03b1-Cr2O3 oxide does not appear even in a high loading. Therefore, CrOx/Al2O3 have higher conversion than CrOx/SBA-15 in high loading. Rossi et\u00a0al. (1992) studied the catalytic activity and texture properties of CrOx species on SiO2, Al2O3 and ZrO2. The result of PDH show CrOx supported on ZrO2 has a much higher activity due to the highly dispersed chromium species, this may be due ZrO2 has higher ability to disperse chromium catalyst as isolated sites. Supports also could affect coordination environment of active metals though M-O-S bonds (M is the active metal center, S is the support), and result in influence on the reducibility of supported oxide. Xie et\u00a0al. (2020) found ZrO2 can effectively improve the TOF of VOx than the sample using Al2O3 as support. The VOx/ZrO2 catalysts shows more loss of coordination oxygen atoms than VOx/Al2O3, and activity of lower coordinated V\u2013O pair on ZrO2 shows six times higher than Al2O3. Author also measured reducibility of V\u2013O\u2013V dimer on Al2O3 and ZrO2 though a DFT calculation, result showed V\u2013O\u2013S and V\u2013O\u2013V bond in VOx/ZrO2 are weaker than VOx/Al2O3.In the isolated active sites, metal-oxygen active pairs are closely connected with the support though M-O-S bonds, their catalytic performance are more greatly affected by the support. Organic precursor directed synthesis make M-O-S interaction in well-defined oxide catalysts possible. Szeto et\u00a0al. (2018) synthesized well-defined isolated Ga catalysts supported respectively on Al2O3 and SiO2 by SOMC method, which Ga species exist as isolated single-atoms on Al2O3 and double-atom on SiO2. Ga1/Al2O3 appears to show more active and selective catalyst than Ga2/SiO2. This may be due to Ga-O-Al sites have higher C\u2013H cleaver activity than Ga-O-Si, thus proved the importance of support in oxide catalysts. This strong interaction between oxide and support decide isolated sites sometimes may not be the best active site, for example, strong support-oxide interaction would cause isolated oxide sites difficult to be reduced into coordination unsaturation state, and thus resulting in low activity. Liu et\u00a0al. (2016) found isolated VOx has lower TOF than V2O5 crystalline in VOx/Al2O3. The lower reducibility of V\u2013O\u2013Al bond than V\u2013O\u2013V bond may be the reason of their poor activity, which result in most vanadium species are only reduced to V4+ instead of highly active V3+. The M-O-S bonds sometimes lead to the formation of B-acid sites which would catalyzes side reactions (Castro-Fern\u00e1ndez et\u00a0al., 2021).. Therefore, compatibility of isolated oxide catalysts should be carefully considered according to the types of supports.In some bulk catalysts such as ZrO2 (Zhang et\u00a0al., 2018), TiO2 (Li et\u00a0al., 2020), Al2O3(Wang et\u00a0al., 2020), Ga2O3 (Zheng et\u00a0al., 2005), particle size, crystal phase and degree of surface defects would play a similar role instead of dispersion. For Zr-based catalysts, oxygen vacancies on zirconia surface are widely considered as the source of activity. (Otroshchenko et\u00a0al., 2015). Zhang et\u00a0al. (2018) found that ZrO2 monoclinic crystal with smaller particle size shows higher intrinsic activity than bigger one. The effect is more obvious when the particle size is below 10\u00a0nm. This is due to more amount of surface defects such as corner, edge exist on the small ZrO2 crystal and results a higher oxygen vacancy concentration. At the same time, coking selectivity would decrease with the decrease of ZrO2 crystal size, this may be due to small nanoparticle have more dispersed active sites. Zhang et\u00a0al. (2019a,b) also studied the effect of crystal phase of ZrO2 on catalytic performance. Decreasing of particle size would increase activity of ZrO2 in either monoclinic phase or tetragonal phase. However, although they have similar activation energy, monoclinic ZrO2s have a higher conversion than tetragonal ZrO2. This phenomenon can be explained by that lattice oxygen in monoclinic ZrO2 has a better mobility than that on tetragonal ZrO2. Zheng et\u00a0al. (2005) tested catalytic performance of various Ga oxide with different morphology. Catalytic activity tests found \u03b2-Ga2O3 has the highest specific activity of PDH in different polymorphs of Ga2O3. NMR and NH3 adsorption indicated that \u03b2-Ga2O3 has higher Lewis acid sites and tetra Ga3+ density than other polymorphs of Ga2O3, this shows that the Lewis acid sites is closely related to the tetra coordinated Ga3+ and these Lewis acid sites are related with activity strongly.For oxide catalysts, adding appropriate promoters could effectively change the geometry and electronic structure of the active sites, thus affecting the performance of the catalysts (Sattler et\u00a0al., 2014). Interaction between promoters and active oxide could be classed to two interaction: oxide-oxide and metal-oxide interaction. This interaction depends on the existing form of promoter. For oxide-oxide interaction, heteroatom would effectively change redox properties or affects the acidity and basicity of oxide surface, thus effectively effect catalytic performance (Ruiz Puigdollers et\u00a0al., 2017). However, due to complexity of oxide catalysts, systematic researches on predicting what kind of element would in oxide catalysts is still difficult. From the experience, oxyphilic or basic elements often lead to lower activity but higher stability, while the dopants with acid elements (most of them could be used as independent active components) are conducive to the improvement of dehydrogenation activity (Liu et\u00a0al., 2020; Li et\u00a0al., 2016; Zhang et\u00a0al., 2019). Oxide-oxide interaction could also affect the aggregation morphology of active oxides though M-O-M bonds and thus affecting the dispersion of active sites. For metal-oxide interaction, metals especially noble metals could activate hydrogen molecules into high energy hydrogen atoms, so they could introduce hydrogen spillover or reverse spillover effect to adjust the reducibility of oxide or part in hydrogen evolution step in PDH (Otroshchenko et\u00a0al., 2015; Sattler et\u00a0al., 2014). The metal-oxide interface could also affect geometry and electronic structure of the active sites in some case (Liu et\u00a0al., 2016).Doping metal element could be achieved by simple co-impregnation or sequential impregnation method and effectively effect catalytic performance, thus they have been widely studied. Alkali metals such as Na and K are the most common metal promoters used in PDH. They could poison strong acid sites on surface of catalysts, which are considered to prone catalyzing side reactions such as coking. In addition, the addition of K is also conducive to improve dispersion of the oxide species on the support (Sattler et\u00a0al., 2014). Other elements, such as Ce (Zhang et\u00a0al., 2019), Zn (Liu et\u00a0al., 2020), Ni (Li et\u00a0al., 2016), have also been used in Cr-based catalysts as promoters. Zhang et\u00a0al. (2019) studied the role of Ce in CrOx/Al2O3 catalyst. The interaction between Ce and Cr results in more oxygen vacancies and after modification of Ce. Addition of Ce effectively reduced the inactive Cr6+ species and the dispersion of CrOx species, thus stability, selectivity and conversion are effectively improved. Wu et\u00a0al. (2017) explained in detail the structure-reaction relationship of Mg promoter in VOx/Al2O3. The addition of Mg would disperse V2O5 crystals (which is mainly contributors of coke) into oligomeric or isolated VOx structure (Fig.\u00a07\nc). This effect not only effectively decrease quantity of coke, but also results in a better conversion due to higher dispersion. However, excessive addition of Mg oxide would form MgO aggregation and cover active VOx, which lead to decline of activity. Precious metals are used as promoters. Liu et\u00a0al. (2016) prepared a novel ZnO/Al2O3 catalyst doped by trace amount (0.1\u00a0wt%) of platinum for PDH. Comparing to traditional ZnO catalyst, the addition of trace amount platinum greatly improves the reduction resistance of zinc oxide and reduces the formation of unstable metallic Zn. In addition, Pt improves the desorption of H2 and C\u2013H activation on the surface of Zn oxide. Further characterization indicate that Pt clusters are highly dispersed and covered by ZnO. Authors indicate these Pt clusters would increase the Lewis acidity of zinc though electronic interaction and promotes the desorption of adsorbed hydrogen atoms, this effect could reduce overreduction of unstable metallic Zn and increase ZnO activity (Fig.\u00a07b). Sn supported by SiO2 catalyst recently has high selectivity and activity, but at high temperature, the reduced Sn will aggregate and lead to the decrease of activity (Wang et\u00a0al., 2016; Wang et\u00a0al., 2020; Liu et\u00a0al., 2020). Wang et\u00a0al. (2016) found that after adding a small amount (0.05\u00a0wt%) of Pd into Sn/SiO2 would effectively increase stability of catalysts, the life of Sn catalyst was prolonged by more than two times. Authors found introduction of Pd could decrease aggregate degree of Sn species and reduce the loss of tin at high temperature, thus improve activity and stability.Except for metal elements, some non-metallic oxides such as SOx, POx can effectively improve the negative electron properties of oxide. Sun et\u00a0al. (2014) found that the introduction of SO4\n2\u2212 into Fe2O3 could effectively improve the yield of propylene and stability. This better catalytic performance may be due to the presence of SO4\n2\u2212 also lead to electron deficiency of Fe atoms, which makes absorb negatively charged second carbon of propane easier and provide a higher activity. In addition, promoting sulfur could also stabilize Fe chemical state and the inhibiting formation of harmful species, such as FeCx or FeS (Sun et\u00a0al. 2014, 2015). Tan et\u00a0al. (2016) synthesized a serious phosphorus containing Fe/Al2O3 catalysts. The addition of phosphorus plays an important role in creating more active sites, while the Fe2O3 without phosphorus will coked severely and have a low selectivity. After adding phosphorus, the catalyst with best performance has a conversion about 14% and a selectivity of 80%. In addition, longer duration experiment indicated that the conversion and selectivity of catalyst can be almost unchanged in 24\u00a0h. Gu et\u00a0al. (2020) modified vanadium oxide surface supported by Al2O3 with phosphorus by gas phase reduction method. After proper modification, the activity of the catalyst decreased slightly, but the stability increased greatly. This may be due to P dispersing VOx polymerization to isolated V sites, which have lower activity but higher stability. The addition of phosphorus also reduces the acidity of the oxide, thus improving the overall stability of the catalyst.While further improving second metal loading, dopant would increase in diffusion probability and react with the original oxide to bimetallic compound structure or solid solution. Formation of stable bimetallic oxides often leads to decreasing number of active sites and poor catalytic performance. But sometimes the formation of new structures is also conducive to the formation of better active sites than supported oxide. Chen et\u00a0al. (2008) synthesized a kind of Ga2O3\u2013Al2O3 solid solution catalyst with a spinel structure. Compared with the traditional Ga2O3/Al2O3 catalyst, the formation of solid solution can effectively disperse and stabilize the gallium oxide species. Although Ga2O3\u2013Al2O3 solid solution has a slight decreasing in the initial conversion rate compare to Ga2O3/Al2O3, it has more stable catalytic performance, this enhancement effect is shown in Fig.\u00a07d. Otroshchenko et\u00a0al. (2017) studied the promoting role of Cr in CrZrOx bimetallic bulk oxide. The addition of Cr promotes the lattice oxygen removal by activate H2 molecule to active hydrogen atom and thus improve coordinatively unsaturated Zr4+ sites formation, which has high dehydrogenation activity. Therefore, CrZrOx has higher activity than pure ZrO2 catalysts and even industrial CrOx/Al2O3 catalysts. In addition, lattice CrOx species are also responsible to reduce the strength and amount of strong acid sites on ZrO2, which reduce side reaction caused by Lewis acid sites. Compared with the supported CrOx/LaZrOx catalysts, the bimetallic CrZrOx oxide catalyst has better regeneration stability and activity at the same acidity number.Metals especially noble metals could also play a role of promoter by introducing hydrogen spillover effect, this effect could recombine adsorbed hydrogen atoms or adjust the reducibility of oxide catalysts. Sattler et\u00a0al. (2014) synthesized a Pt\u2013Ga2O3/Al2O3 with trace platinum content of only 0.001w%. Although both platinum and gallium oxide are considered to have catalytic activity in some previous studies, this catalyst has better conversion and selectivity than one of them exist alone. It is assumed that promoting role of Pt is accelerating recombination of the hydrogen atoms to H2 which is reverse reaction of hydrogen spillover. The stability and size of Pt particles would greatly affect the hydrogen spillover effect. Han et\u00a0al. (2019) studied the promoting effect of Cr in a series of supported CrZrOx catalysts. They found Cr atoms would activate H2 molecule to active hydrogen atom and promote coordinatively unsaturated Zr4+ sites formation, which has high dehydrogenation activity. This synergy effect could effectively reduce the usage of toxic Cr and expensive Zr but keep the activity unchanged.Promoter has a similar effect to improve catalytic properties of bulk oxide. Otroshchenko et\u00a0al. (2017) studied the influence of doping Li, Ca, Mg, Sm, La in MZrOx catalysts (M stands for dopant). Doping of ZrO2 with Ca or Li would result a low activity. On the contrary, Addition of La, Sm, Y would effectively enhance propylene formation rate of ZrO2. Creation of more surface defects on ZrO2 after adding heteroatoms may be the cause of these activity changes. Noble metals such as Pt, Rh, Ir and some hydrogen-activated metals such as Cu could provide hydrogen spillover effect, which improve reducibility and amounts of active sites (Otroshchenko et\u00a0al., 2015). Zhang et\u00a0al. (2020) carried out a more systematic study interaction between ZrO2 and Rh nanoparticles in Rh/ZrO2 catalyst. Experiment and calculation results indicated the addition of Rh increase the oxygen vacancy number on the surface of ZrO2 which is widely considered to main active sites, thus effectively improve conversion. However, excessive addition of Rh would lead to a decrease in activity, this may be due to too many oxygen vacancies would restrain the desorption of C3H6 and hindered PDH reaction (Fig.\u00a07a). Recent oxide catalysts are shown in Table\u00a02\n.Coke is the general term of deep dehydrogenation alkyls or graphitized carbon deposition (Sattler et\u00a0al., 2014). In PDH, coke deposition path mainly includes four steps: deep dehydrogenation, C\u2013C bonds breaking, formation of aromatic hydrocarbon and graphitization (Hu\u0161 et\u00a0al., 2020; Lian et\u00a0al., 2018; Zhao et\u00a0al., 2015). But actually, its detail mechanism and key intermediates is still ambiguous until now. Researchers have proposed some possible processes in deep dehydrogenation process basis on DFT. Valc\u00e1rcel et\u00a0al. (2006) studied stability of a variety of intermediates of deep dehydrogenation on Pt (111) include 1-propenyl, propylidyne, propenylidene, and propyne. Among these intermediates, propylidyne has the lowest energy and preferentially adsorb on hollow of three platinum atoms. This work pointed out the most stable hydrocarbon intermediate in PDH on Pt (111) surface. In another DFT calculation of Yang et\u00a0al. (2010), they indicated with the process of dehydrogenation the energy barrier of C\u2013C scissor decreases continuously. Propyne was found to be the most likely starting point for C\u2013C scissor to coke deposits. Many other studies have obtained similar conclusion (Saerens et\u00a0al., 2017). In addition to coking directly caused by active sites, acidic sites also catalyze the coking process, Lewis or Bronsted acid sites would catalyze coke formation like a catalyzed aromatic hydrocarbon process and follow an acid-catalyzed carbocation mechanism (Sattler et\u00a0al., 2014).Although the mechanism of coke deposition is still uncertain, C1 and C2 species are generally considered as main precursor of coke deposition rather than C3. These species have more lone electrons and would attract each other and form aromatic rings through surface-mediated mechanism. (Larsson et\u00a0al., 1996; Lian et\u00a0al., 2018; Saerens et\u00a0al., 2017). Jackson et\u00a0al. (1997) proved polycyclic aromatics formed on Pt/Al2O3 in PDH are more likely derived from C1 species rather than C3 species used a mathematical derivation. After detecting coke by mass spectrometry, they found main component of coke are pyrene and methyl pyrene and they cannot be divided by three. In addition, isotopic labeling experiments also proved C3 species would be divided into C1 in the process of coke deposition. Up to now, most of the DFT experiments use the polymerization of C1 or C2 to aromatic ring as the main carbon deposition process. (Lian et\u00a0al., 2018; Saerens et\u00a0al., 2017). After the formation of the first aromatic ring, the aromatic ring precursor expands continuously to polycyclic aromatic hydrocarbons through a Diels\u2212Alder mechanism, finally form highly graphitized coke.For either metal catalysts or oxide catalysts, coke deposition are the main culprit of deactivation. In industrial process, Pt-based catalysts need to be regenerated every 7\u20138\u00a0h. For Cr-based catalyst, this regeneration would be more frequent and one-way reaction time is only about 10\u00a0min(Sattler et\u00a0al., 2014). Although sometimes coke also show some advantages in Cr-based catalysts, such as they could provide additional energy in burning regeneration step. But for platinum-based catalysts, heat from combusting coke would lead to serious sintering even more serious than reaction step (Kaylor and Davis, 2018). In order to understanding effect of coke deposition, researches on coke deposition in PDH could be described from three levels: macro-levels, meso-levels and micro-levels (Ye et\u00a0al., 2019). In view of macro-level, coke will influent mass and heat transfer process, its influence is closely related to type of reactor and the reaction technology. Limitation of research focus of this paper, we would ignore detailed discussion for this influence. Recently, understanding and inhibiting coke deposition at a meso- and micro-level has attracted a lot of attention.In view of meso-level, coke product would narrow and finally block pore channels, causes diffusion resistance and leads to deactivation (Ye et\u00a0al., 2019). Proper pore structure could improve carbon capacity and thus reduce block effect. In addition, pore structure could play an important role in diffusion of propylene, and thus shorten the contact time between propylene and active sites to reduce coke deposition. Ye et\u00a0al. (2019) built a pore network model of PtSn/Al2O3 and use it to simulated its coke deposition process. According to the simulation results, increasing of pore connectivity and volume-averaged pore radius did not obviously influence the coke formation rate, but increased the maximum coke deposition. Decrease of pore size distribution would not affect the rate of coke deposition, but significantly increase maximum coke capacity. In addition to pore structure, the pore radius can't make a great impact on carbon capacity but is proportional to the rate of coke deposition. Ye et\u00a0al. also simulated the in-situ change of pore structure in deactivation process. Due to obvious diffusion limitation, in the first stage, propylene is difficult to diffuse from the inside to outside and thus coke form mainly in the nearly center part of catalyst, catalysts would suffer a rapid deactivation step. In the second stage, the pores in the center are almost blocked, so coke mainly exists in the outer region of the particles and the rate of coke deposition slows down.How to control the appropriate pore structure to minimize coke deposition has attracted extensive attention of researchers. Straight uniform pore is widely considered to be more favorable for propane dehydrogenation. Accumulation of particles could form some natural pores, but these inter-crystalline pores usually don't possess uniform size and have many structural defects such as curved and closed structures, which are not conducive to mass transfer. In industrial, shaping catalysts with high pressure is needed to homogenize size of channel of supports, this usually requires an ultra-high pressure of more than 100\u00a0MPa. Therefore, some monolith materials such as ordered mesoporous oxide, molecular sieve which has uniform and natural pore structure have attracted largely attention. Creation of ordered multistage pores also plays an important role in promoting mass transfer. Li et\u00a0al., 2017 prepared a series of PtSn/TS-1 catalysts with different particle sizes by hydrothermal synthesis. Although TS-1 is a microporous molecule sieve, with the decrease of particle size, mesopores gradually appear on surface of TS-1 and the hierarchically porous structure is formed. Compared with TS-1 with large particles, TS-1 with small particles has apparent advantages in conversion, selectivity and stability. The diffusion of products and reactants may be the key to explain this phenomenon. Calculation of Weisz\u2013Prater criterion indicated the PtSn/TS-1 of large particles is seriously affected by internal diffusion but small one has a low resistance to inter-crystalline diffusion. The hierarchically pore structure doesn't only significantly accelerate the diffusion of propylene, but also accelerating the diffusion of propane and avoiding the internal diffusion control. Liu et\u00a0al. (2020) embedded tin into dendritic mesoporous SiO2 nanoparticles for PDH. Dendritic mesoporous SiO2 nanoparticles have mesoporous structure and radial 3D pore with highly connectivity. Changing the ratio of template could regulate the pore structure with different pore size and connectivity, an appropriate pore structure could improve the reaction rate and selectivity.Dynamic radius of propane (4.3\u00a0\u00c5) and propylene are both relatively small, which reduces researchers' attention to the influence of pore structure. Because PDH is one-step reaction, the role of diffusion effect in PDH is often ignored. However, due main and side reactions of propane dehydrogenation is series reaction, the pore structure plays an important role in the selectivity of propane dehydrogenation. At present, most researches on the process of coke deposition focus on the microscopic reaction mechanism of coke. Next, we will discuss the micro-process of coke deposition in detail.At a micro scale, coke and coke precursors are rich in lone electrons or electron-rich \u03c0 bond, active sites would adsorb them strongly and thus result in deactivation. Therefore, those highly unsaturated coordination sites or electrophilic acidic sites are easier to be poisoned by coke. For example, step, edge sites on Pt nanoparticle and over reduced oxide sites are more prone to tightly adsorb C atom and suffered by deactivation (Zhu et\u00a0al., 2015). Coke would first generate and cover these sites tightly, and thus lead to decrease of conversion. Unsaturated sites are also the main sites which are prone to catalyze cracking and other side reactions, thus coke deposition process is always accompanied by the increase of catalyst selectivity (Gorriz et\u00a0al., 1992; Larsson et\u00a0al., 1996). However, even for terrace of Pt, coke deposition has also been proved to improve selectivity. Lian et\u00a0al. (2018) explained effect like double-edged sword of coke through a DFT-based kinetic Monte Carlo simulation on Pt (111) in PDH. At the beginning, although conversion in Pt-based catalysts is very high, but the main reaction is deep dehydrogenation rather than formation of propylene. As carbon deposits cover the surface, Consumption rate of propane decreases rapidly, and accompanied by a sharp increase in selectivity. Snapshots indicated the formation of coke covered and separated the Pt ensembles like alloying promoters such as Sn, and thus increase selectivity.Reducing coke deposition from both structural and electronic aspects has been widely used in PDH process. Isolated active sites are beneficial to reduce the contact between coke precursors and is helpful to reduce structural-sensitive coke deposition reactions. For metal catalysts, alloying metals as promoter could separate active metal ensembles and deep dehydrogenation species are difficult to form stable transition state on these small ensembles. Isolated catalysts such as SAACs have lower tendency to produce coke (Sun et\u00a0al., 2018). Electronic effect also plays an important role in inhibiting coke deposition. The promoting metal could transfer the electron to active metal as an electron donor, the electron-rich active sites would exclude the same electron rich \u03c0 orbitals of propylene and avoid strong adsorption or deep dehydrogenation (Wang et\u00a0al., 2018). Yang et\u00a0al. (2012) found Pt alloyed with Sn could effectively broaden d-bandwidth and lead to a downshift of d-center. This downshift of the center of d-band would lead to a higher dehydrogenation barrier, thus Sn slower carbon deposition rate though an electron effect. It should be noticed SMSI can also transfer electron to active sites similar to promoters and thus reduce the deep dehydrogenation reaction (Jiang et\u00a0al., 2014). In the oxide, the highly isolated active sites are also proved to be beneficial to reduce the formation of carbon deposition, because hydrocarbon precursors adsorbed on these sites are hard to combine with each other and form aromatic rings (Zhao et\u00a0al., 2019). In addition, the addition of basic promoters could modify affinity of the active site to electron, thus inhibiting coke deposition.In PDH, precursor of coke and formed coke deposition attached to active sites loosely would be constantly moving on active surface and even move to support surface derived by high temperature, which is also called a self-cleaning effect. This effect is confirmed in two different peaks of temperature programmed oxidation (TPO) experiments of spent catalysts, which showed coke may move into surface of support and further are dehydrogenated to more graphitized coke with higher combusted temperature (Jiang et\u00a0al., 2014; Redekop et\u00a0al., 2016; Wang et\u00a0al., 2018). Some researchers also attribute two kinds of combustion peak to some coke would migrate to far distance from platinum, which could lead oxygen overflow effect of platinum need to a higher energy barrier (Larsson et\u00a0al., 1996; Redekop et\u00a0al., 2016). Rich electron density would weaken interaction between coke precursor and active sites thus improve this self-cleaning effect (Iglesias-Juez et\u00a0al., 2010).Although the amount of coke deposited on Pt\u2013Sn catalysts are even more than on pure Pt catalysts in similar condition, Pt\u2013Sn catalyst still has higher stability. At the same time, the movement of coke is greatly affected by size of platinum particles. Peng et\u00a0al. (2012) observed by a high-resolution TEM, binding graphene layers will slough off from platinum particles to support due to the strain of structure, which is closely dependent on the particle size. Large Pt particles are more likely to form a carbon layer which coated on the surface of platinum particles. Smaller Pt particles would cause a greater tension on the carbon layer and thus coke would be form of carbon nanotubes or sheets, which are easier to desorb from Pt surface. The movement of the carbon layer will also cause the deformation of platinum nanoparticles due to its strong adsorption and result in the change of reaction performance (Wu et\u00a0al., 2016). Due to the complex structure of carbon materials, how to correctly understand the role of graphite layer coke still needs further study.In sum, strategies of anti-coke deposition could be started from both mesoscopic and micro perspectives. On the meso-scale, coking deactivation are mainly caused by channel block, increasing pore size could effectively reduce this process. In addition, proper pore structure is also important for reducing coke reaction. Straight, uniform and ordered channels are benefit to reduce the residence time of propylene inside catalyst, thus reducing coke deposition. On the micro-scale, deactivation of catalysts mainly comes from the covering of active sites by coke deposition. Isolated active sites are beneficial to inhibit the structure sensitive coke deposition process. From the view of electronic effect, the active sites rich in electrons can weaken the adsorption of coke precursors on the catalyst surface, thus inhibiting the deactivation of coke deposition. These structural and electronic effects could be achieved by adjusting promoters and supports reasonably.Co-feeding gas in PDH get more and more attention. Traditional feeding gas in PDH is pure propane, sometimes, in order to reduce coke deposition, a mixture of propane and hydrogen is also used in some cases (Saerens et\u00a0al., 2017). In industrial process, hydrogen would promote reverse reaction of deep dehydrogenation and thus inhibit coke formation, but it also inhibits propane dehydrogenation at the same time. Steam is another widely used co-feeding gas, it could eliminate coke though a water-gas conversion process and reduce the partial pressure of reaction gas. Some mild oxides such as N2O, CO2 (Xie et\u00a0al., 2019) are also used in propane dehydrogenation to improve conversion and reduce coke deposition. But strictly speaking, these are the category of oxidative dehydrogenation after addition these gases. Besides participating in the reaction, these gases also have important influence on the active site structure on the catalyst. How to understand their effect on the reaction is very important to optimizing technology.Hydrogen is the most important co-feed gas of PDH process in serious industrial Technology. In general, the role of H2 is considered to promote the deep dehydrogenation and dehydrogenation reverse reaction, and thus reduce the conversion and improve the stability. It should be noted that although hydrogen could participate in the reverse reaction of deep dehydrogenation, it could not reduce has graphitized carbon deposits, which also proves the irreversibility of graphitized coke deposits (Larsson et\u00a0al. 1996;\nRedekop et\u00a0al., 2016). DFT calculations and experiments also has proved hydrogen adsorbed on the surface would also promote the desorption of propylene and increase the energy barrier of dehydrogenation, thus inhibit further dehydrogenation of propylene to coke deposition. Sattler et\u00a0al. (2013) found addition of H2 not only reduces the total amount of coke deposition, but also effect the structure of coke deposit. Though the analysis of Raman spectrum and TGA, researcher observed the coke deposit has a smaller size and a more graphitized structure after adding hydrogen. This finding also verifies DFT's conclusion. In fact, sometimes the addition of hydrogen will lead to the improvement of both stability and conversion, this anti-common phenomenon is considered to be related to adsorption effect. Saerens et\u00a0al. (2017) tested influence mechanisms of hydrogen in Pt-based catalysts on the activity of PDH though a microreactor simulation. In addition to increasing the dehydrogenation barrier, highly partial pressure of hydrogen in feedback gas would not only reduce deep dehydrogenation species but also free active sites. Thus, appropriate addition of H2 would create more free sites on the surface and increase in activity. On the oxide, the addition of hydrogen could also change the properties of active sites, resulting in the change of reaction performance (Zhao et\u00a0al., 2018). Therefore for different types of catalysts, hydrogen atoms could not only participate in surface reactions, but also change the structure of active components, the role of hydrogen should be carefully considered due to different hydrogen absorption ability.Steam is a good choose of feeding gas in industrial process which can reduce the partial pressure of propane, promote the reaction equilibrium moving forward and at the same time can be used as a good heat conducting material. Though a reaction similar to water gas conversion, steam can reduce the rate of coke deposition. However, in high temperature, co-feeding steam will accumulate in breaking of structure of supporters, such as Al2O3, zeolite, basic metals are commonly used in improving stability of these materials. In some oxide catalysts, such as zirconia, the addition of steam would cover the active sites and completely inactivate the catalyst (Otroshchenko et\u00a0al., 2017). In some systems, the chemical reaction of water vapor with active component also plays a positive role. Shan et\u00a0al. (2015) found the addition of steam would change alloying structure of Pt\u2013Sn catalyst. After proper steam pretreatment time, the size of platinum particles will decrease and thus improve activity. Promotion of steam would oxidize the Sn(0) component to SnOx, resulting in the partly dealloying of Pt\u2013Sn catalysts and formation of highly active alloying components Pt3Sn, thus decrease dehydrogenation barrier. As a cheap additive, steam has high research value.Today, uncertain volatility of oil price, decreasing crude oil resource, energy consumption and carbon emission are driving traditional petrochemical industry to reexamine the utilization efficiency of carbon atoms. Utilization of propane is a very challenging but promising part in petrochemistry, while it traditionally is used as gas fuel with a lot of CO2 emission. In order to utilize these resources in green way, PDH technology need pay more attention. From the perspective of atomic economy, PDH is a fuel-chemical conversion process with high carbon utilization, in which propylene is its only product. In PDH, hydrogen as a by-product with almost zero pollution, which is an important low-carbon fuel and reaction gas. Traditional hydrogen comes from fossil fuels and is accompanied by a large amount of CO2 emissions. Undoubtedly, widening propane dehydrogenation process will provide a \u201c1\u00a0+\u00a01 > 2\u201d effect in either environment friendliness and economic benefits. However, traditional industry catalysts have their intrinsic shortcoming such as fast coke deposition and high toxicity of Cr-based catalysts and expensive price of Pt-based catalysts. In order to further improve this potential technology, structure-activity relationship of catalysts should be paid a lot of attention.For metal-based catalysts, it is proved that PDH takes place on adjacent metal atoms from one to multiple. The activity and selectivity of propane dehydrogenation are closely related to coordination environment and continuity of metal atoms. Catalytic performance of metal active sites can be effectively adjusted by forming alloy with second metal, electronic effect and ensemble effect often be used to reveal promoting role of alloy. Traditional support in metal-based catalysts is Al2O3 or SiO2. Highly dispersed alloy particles could be achieved by precisely adjusting the defect site structure or surface geometry of them. Recently, some semiconductor material such as TiO2, CeO2 were also used as supports due to their electronic effect. We think future research in metal catalysts should focus on the structure of alloys, defect sites engineering of inert supports and interaction between alloy nanoparticles and \u201cactive\u201d support. For alloying nanoparticles, different surface alloying structure such as substitutional solid-solution alloys, interstitial solid-solution alloys, intermetallic alloy and amorphous alloy with different alloying elements should be studied separately to detail the role of alloying promoter. The sub-structure and internal structure of nanoparticle should also be paid enough attention due to their important electronic interaction with surface atoms. In the aspect of support, the commonly used supports such as alumina and silica have high specific surface area and adjustable hole structure but still need to be further improved their interaction between metal and support. Their morphology and defect sites should be precisely modified to improve dispersion of supported nanoparticle. Considering that most of the traditional supports are electrically inert, those \u201cactive\u201d supports which interact with metal nanoparticle though electronic effect provide new possibilities for the further development of metal catalysts. How to make better utilization of their interaction and improve the specific surface area is more important.For oxide-based catalysts, the active sites of oxides are often considered to be metal-oxygen pairs with unsaturated coordination. Reduction plays an important role in determining the micro-structure of active sites include coordination of unsaturated metallic cation, hydroxyl et\u00a0al. Catalytic mechanisms on oxides catalysts are varied, which are closely related to the properties of metals and oxygen. Oxide could exist as different degree of polymerization on the support, such as isolated sites, polymerized sites or crystalline on the support. Due to the different bonding between the metal center and the support, their catalytic performance is also significantly different. Isolated sites are widely considered as the most favorable active sites for oxides because of their high dispersion and low coke deposition tendency, but considering their strong interaction with support, catalytic performance of isolated sites still needs to be carefully deliberated according to the coordination structure. Oxide catalyst could be further modified by adding metal or oxide promoters. Though metal-oxide interaction or oxide-oxide interaction, promoter would change the degree of polymerization or the electronic structure of active sites. Based on the above analysis, we outline that future development direction of oxide catalysts should focus on deeply understanding the mechanism and interface effect of oxide-oxide and metal-oxide. At present, most of the catalytic mechanism of oxide catalysts come from mechanism of homogeneous C\u2013H activation process catalyzed by metal complexes. However, the mechanism of heterogeneous catalysis is still unclear because detecting evolution of active sites in solid-gas two-phase is much more difficult than single-phase solution. Considering that there are many kinds of oxide catalysts, structure of active sites and reaction mechanism in each of them should be carefully analyzed individually. Metal and oxide promoters play an important role in improving the performance of oxide catalysts. Although many studies have found that a variety of element have positive effects in oxide catalysts, due to complex structure of interface, their promoting mechanism is still unclear. Adjusting their interaction through interface engineering is very promising to not only obtain higher activity but also establish structure-activity relationship between promoter and oxide. Especial the noble metal-oxide interaction, some recent reports even found that noble metals as promoters in oxide can achieve higher catalytic performance than noble metals as active components, regulatory interaction between noble metal and oxide though control the interface need more attention for replacing expensive noble metal catalysts with cheap oxides.Except for structure-activity relationship dominated by chemical effect, pore structure and mass transfer should be also emphasized. Reasonable design of highly interconnected, uniform pore with appropriate size plays an important role in further improving catalyst performance, especially anti-coking performance. The influence of pore structure should be paid more attention in the synthesis of supports. Influence of co-feeding gases on structure of active sites is also highly concerned in this review. We think future research in co-feeding gas should focus on some alkane-associated gas or gas easy to be separated in order to further improvement of dehydrogenation process.This work was supported by the National Natural Science Foundation of China (21872163, 21972166), National Engineering Laboratory for Mobile Source Emission Control Technology (NELMS2017A05), Beijing Natural Science Foundation (2202045, 2182060), PetroChina Innovation Foundation (2018D-5007-0505).", "descript": "\n                  Dehydrogenation of propane (PDH) technology is one of the most promising on-purpose technologies to solve supply-demand unbalance of propylene. The industrial catalysts for PDH, such as Pt- and Cr-based catalysts, still have their own limitation in expensive price and security issues. Thus, a deep understanding into the structure-performance relationship of the catalysts during PDH reaction is necessary to achieve innovation in advanced high-efficient catalysts. In this review, we focused on discussion of structure-performance relationship of catalysts in PDH. Based on analysis of reaction mechanism and nature of active sites, we detailed interaction mechanism between structure of active sites and catalytic performance in metal catalysts and oxide catalysts. The relationship between coke deposition, co-feeding gas, catalytic activity and nanostructure of the catalysts are also highlighted. With these discussions on the relationship between structure and performances, we try to provide the insights into microstructure of active sites in PDH and the rational guidance for future design and development of PDH catalysts.\n               "}
{"full_text": "With the acceleration of industrialization, water pollution caused by the uncontrolled discharge of a large number of wastes and hazardous pollution into water has become a critical issue worldwide [1\u20133]. Various advanced oxidation processes (AOPs), including photocatalysis, electrocatalysis, Fenton-like oxidation, persulfate oxidation, and others have been proved to be sustainable solutions to toxic chemical pollutants [4\u20136]. Catalysts, especially heterogeneous catalysts, play a central role in AOPs for environmental remediation because of their easy recyclability and separation [4,7]. Heterogeneous catalysts usually contain transition metals or metal oxides and supporting materials [7\u20139]. However, the immobilization of metal or metal oxides decreases the dispersion of active sites and enhances the diffusion resistance on the surface of a solid catalyst, limiting its catalytic efficiency [9,10]. Hence, developing cost-effective advanced catalysts with high atom utilization efficiency and perfect stability is critical for environmental catalysis.Recently, single-atom catalysts (SACs) have become ideal catalysts with excellent activity and stability in variety of catalytic reactions [11\u201313]. Due to the almost 100% atom utilization efficiency of metal atoms, unique electronic features and ultralow metal loading, SACs have been used extensively in environmental catalysis [13\u201315]. In 2017, single-atom Ag was modified on mesoporous graphitic carbon nitride and used to activate peroxymonosulfate (PMS) to degrade bisphenol A (BPA) under light conditions [16]. Subsequently, abundant novel SACs with distinctive catalytic properties have been developed for environmental remediation [17\u201320].Although various significative review articles over SACs were published in succession [4,5,21,22], comprehensive classification of AOPs over SACs for environment remediation was not well summarized and discussed. In this review, we summarize the technical classification of AOPs over SACs in environment application and discuss the roles of SACs in AOPs for water treatments. As such, this review classified the water treatment process according to catalytic technology, including photocatalytic degradation technology, electrocatalytic degradation technology, Fenton-like reaction, persulfate oxidation reaction and coupling reaction (Fig.\u00a01\n). We believe that SACs are interesting and promising materials for the degradation of water pollutant. Finally, the challenges and research tendency of SACs applied to water remediation are discussed.As advanced catalysts, SACs are a good choice for remediating organic pollution compared to traditional catalysts. The advantages of SACs are as follows (Fig.\u00a02\n): (1) the preparation route is relatively simple such as direct impregnation and pyrolysis; (2) dramatically reduced metal consumption due to the ultra-low metal loading and almost 100% atom utilization; (3) remarkable catalytic activity attributed to abundant atomically dispersed active sites and the unique electronic structures; (4) perfect stability originating from the special covalent bond between metal and non-metallic element; (5) the catalytic mechanisms and active sites were liable to identified. In general, SACs have been proved as excellent candidates and provided a powerful solution to achieve desirable AOPs catalysts with high activity and stability for environmental remediation (Table\u00a01\n).To clarify the advantages of SACs, we compare the reaction pathways on SACs and metal nanoparticles in AOPs. In the process of degrading organic pollutants, metal nanoparticles mainly play the role of trapping and transferring electrons, which can effectively improve the catalytic performance of the catalyst. For SACs, thanks to the formation of abundant chemical bonds with strong bond energies between single metal atom and non-metallic atom, numerous charges accumulate and redistribute on the surface of SACs, which not only significantly increase polarized active sites, but also promote the generation of reactive species (such as SO4\n\u2022-, O2\n\u2022-, \u2022OH and 1O2).Photocatalysis is a kind of technology that can oxidize organic contaminants absorbed on the surface of catalyst under light irradiation. It is an efficient, safe and environmentally friendly technology to deal with environmental pollution [23,24]. The key to achieve the perfect performance of photocatalyst is to ensure the efficient separation of photo-generated-carrier on the surface of photocatalytic material. To promote photocatalytic activities, it is an ideal strategy to decorate metal nanoparticles on the surface of conventional semiconductor photocatalyst [23,25]. For a specific photocatalytic reaction, SACs can affect the three key steps of photocatalytic reaction, including light absorption, charge separation and transfer, and surface catalytic reactions [26], as shown in Fig.\u00a03\nA. The introduction of single metal atoms can not only effectively improve the light capture ability of photocatalytic material, but also enhance the efficiency of photo-generated carrier separation [24,27]. Moreover, atomically dispersed single atoms endow abundant active sites to SACs due to metal-support interaction.In recent years, different types of single atoms, including noble metal (such as Au, Ag, Pt) and non-noble metal (Cu, Co, Ni, Ba, Zn, Bi, Mo and so on) have been loaded on the surface of various semiconductor nanomaterials in the form of atomic dispersion (Fig.\u00a03B), which are widely used in photocatalytic degradation of contaminations in water [23,24,26]. As a classic semiconductor catalyst, TiO2 has been widely used to prepare SACs due to its non-toxicity, low cost, reliable stability and suitable optical properties [28,29]. For example, Xu et\u00a0al. [30] reported on the nanopores TiO2 film loaded atomic Pt and applied to the photocatalytic degradation of trace ethenzamide. They found that the nanopores of TiO2 film could accelerate the eddy diffusion and the abundant Pt active sites could promote the molecular diffusion of low concentration pollutants at the same time. More important, the single Pt atom dispersed on TiO2 film could act as the electron capture center to reduce the recombination efficiency of photo-generated carriers and prolong the lifetime of photogenic holes, which enhance the degradation efficiency of low concentration of ethenzamide under vacuum ultraviolet (VUV) and ultraviolet (UV) illumination. In addition, ideal photoactive materials can not only absorb sun light, but also provide binding pockets to stabilize individual atom. In this respect, thanks to the abundant triazine rings and N sites, g-C3N4 could anchor the metal atoms to separate and transfer charge carriers, which could enhance the photocatalytic activity [4,31\u201333]. For instance, Xin et\u00a0al. [32] successfully synthesized single atom Ag anchored g-C3N4, which significantly enhanced the degradation of rhodamine B (RhB) and tetracycline (TC). They confirmed that the catalytic activity of the single-atom photocatalyst was modulated by the doping amount of monatomic Ag due to the ideal concentration of Ag atoms could cause a higher position of the conduction band, which endued the catalyst with stronger reduction performance. Yang et\u00a0al. [33] introduced single Co atom into polymeric carbon nitride (pCN) via a facile in situ growth strategy. They found strong covalent bands of Co-O bond and Co-N bond could be formed between Co single atom and pCN, which could efficiently expand the absorption of visible light and accelerate the separation and transfer of photo-generate carrier and promote the photocatalytic degradation efficiency of oxytetracycline.Electrocatalytic oxidation process, which is the most direct means to convert the input electrical energy to contaminant oxidation through direct electrode reactions or radical generation, has led to growing interests in degrading persistent organic pollutants [34]. The construction of efficient electrode materials is one of the major challenges to improve the electrocatalytic degradation of pollutants. Recently, unique SACs have been employed to produce electrode coating materials with high electron transfer rates and stability [35,36]. On the one hand, the abundant monatomic active centers in the cathode can form intermediate products with strong oxidation capacity from O2 and H2O, which is beneficial to promote the oxidative degradation of organic pollutants (Fig.\u00a04\nA). On the other hand, the non-metallic active sites in SACs can serve as effective binding sites for immobilizing metal atoms to augment the constitutive redox activity of the metal sites [37\u201339]. Generally, carbon materials are used as supports for anchoring metal single atom due to its perfect conductivity and numerous versatile covalent linkages (Fig.\u00a04B). For example, Pan et\u00a0al. [39] prepared a conductive membrane consisting of single Co atom and N atom co-doped graphene (NG-Co) for electrochemical degradation of MB and RhB. The single Co site served as major active site, while the pyrrolic N groups graphene acted as the key binding sites to immobilize the Co-active site on graphene, which improved the electrochemical catalytic performance of NG-Co. Zhao et\u00a0al. [40] designed an N-doped porous carbon electrode loaded with a bimetallic FeCu single-atom (FeCuSA-NPC) to degrade chlorinated pollutants (CPs). The synergistic effect of single-atom Fe and Cu is benefit to the rapid transfer of electrons and the formation of abundant \u2022OH in FeCuSA-NPC, leading to excellent mass activity of CP pollutants removal.As a \u201cgreen\u201d degradation technology, Fenton or Fenton-like reactions play an important role in the decomposition of refractory contaminants via producing strong reactive \u2022OH from H2O2 in aqueous media [41,42]. Traditional Fenton reaction mainly relies on Fe2+/Fe3+ catalytic decomposition of H2O2 to generate a large number of active species [43,44]. However, the inherent disadvantage of easy aggregation of iron activity sites still exists. To maximize the dispersion of the active sites and achieve higher atom utilization efficiency, single-atom iron was anchored on the surface of supported catalyst in Fenton-like reaction [45\u201347] (Fig.\u00a05\nA and B). Yin et\u00a0al. [46] anchored single-atom Fe onto nanopore SBA-15 (SAFe-SBA) by one step of calcination. The extended X-ray absorption fine structure (EXAFS) analysis confirmed the formation of four Fe-O bonds in SAFe-SBA. Moreover, SAFe-SBA catalyst was used to catalyze H2O2 for p-hydroxybenzoic acid and phenol degradation. The results showed that the degradation rate of SAFe-SBA catalyst was significantly enhanced compared with the aggregated iron sites (AGFe-SBA). The superior catalytic activity of SAFe-SBA was attributed to the single atomic dispersion, which benefited to the exposure of the maximum Fe active sites for H2O2 decomposition to induce more \u2022OH.Generally, the narrow working pH range (generally 2.0\u20133.0) of Fe-based catalysts hinders the large-scale utilization of Fenton-like reaction. Thus, single atoms Cu and Co are used in a wide pH range with H2O2 as the oxidant [48\u201350]. For instance, Wu et\u00a0al. [48] first loaded single Cu atoms decorated on N-doped graphene (Cu-SA/NGO) with atomically dispersed CuN4 moieties via an innovative pyrolysis after freeze drying method. The Cu-SA/NGO catalyst displayed excellent activity and stability in the degradation of various organic contaminants at neutral pH. The authors proposed that Cu-SA/NGO could provide more CuN4 sites for H2O2 activation to produce \u2022OH active species with lower energy barriers under acidic and neutral conditions. This provides a foundation for the application of single atom Cu-based catalysts in the degradation of organic pollutants. Xu et\u00a0al. [49] incorporated single Cu atoms in graphitic carbon nitride (Cu-C3N4) to activate H2O2 to generate \u2022OH at neutral pH. Then the Cu-C3N4 was rationally designed for Fenton filter to oxidize the organic contaminants in the wastewater treatment system (Fig.\u00a05C\u2013E).Recently, thanks to the generated high oxidizing species of free radicals (SO4\n\u2022- and \u2022OH), the persulfate oxidation technology (POT) has obvious advantages in effectively degrading organic wastewater. Many works reported that persulfate such as peroxymonosulfate (PMS) and peroxydisulfate (PDS) can degrade pollutants directly [51,52]. In 2018, Li et\u00a0al. [53] first anchored single-Co-atom on porous N-doped graphene to degrade BPA via the activation of PMS, SACs were began to apply into POT for organic contaminants degradation (Fig.\u00a06\n).PMS (HSO5\n-) has an asymmetric structure and long superoxide O-O bond (1.326\u00a0\u200b\u00c5). Thanks to the higher oxidation potential and long half-life of SO4\n\u2022-, PMS-based catalytic system combined with SACs has been emerged as a promising method for environmental remediation. There are various catalytic mechanisms of SACs in PMS system, mainly including free radicals (SO4\n\u2022-, O2\n\u2022-, \u2022OH) and reaction intermediate ROS (1O2 and high-valence metal oxidation species) (Fig.\u00a06A). At the same time, metal species can regulate the electronic properties of the surrounding non-metal elements, forming metal-non-metal active centers, which not only facilitates PMS activation, but also accelerates the formation of intermediate ROS [54]. Recently, various metal-based SACs have been used as PMS activators to degrade organic pollution [54,55]. In particular, Fe-based and Co-based SACs are most widely studied [57\u201359]. For instance, Duan et\u00a0al. [56] loaded single Fe atoms onto g-C3N4 (SAFe-CN) for PMS activation to degrade o-phenylphenol (OPP). The SAFe-CN coated water treatment filter remained 100% of OPP removal after 100\u00a0\u200bh of operation. More importantly, the author found that the dominated mechanism was not the non-radical pathway but electron transfer, which improved the utilization efficiency of PMS. Zhao et\u00a0al. [58] successfully anchored atomically Co in B-doped-CN network (BCN/CoN[2+2]) adopted structural constraint engineering strategy, and this catalyst can effectively activate PMS to degrade TCL with 80% removal in 5\u00a0\u200bmin. The experiment test and density functional theory calculations verified that the BCN/CoN[2+2] could drive the complete switching of PMS to 1O2 owing to the existence of the active site Co. Wang et\u00a0al. [59] anchored singly Fe on graphite carbon nitride (Fe-SA/PHCNS). They reported that the formation of \u2261FeN6=O species could activate the core catalytic site of \u2261Fe-N6, which endowed 97.2% average PMS utilization rate and less affected by environment factors. Moreover, a novel filter composed of Fe-SA/PHCNS and carbon felt displayed high efficiency filtration and oxidation performance for ACE degradation (Fig.\u00a07\n).In addition, other active transition metals (Cu, Mn) have been investigated in PMS [60\u201362]. Chen et\u00a0al. [61] embedded single-atom copper in reduced graphene oxide (SA-Cu/rGO) to activate PMS for the degradation of various antibiotics. They reported that the synergistic interaction between rGO matrix and single-atom Cu active sites could promote the formation of reactive species on the catalyst surface and in-situ decomposition of contaminants, improving the catalytic performance of the catalyst. Yang et\u00a0al. [62] successfully prepared SA Mn-N4 catalyst, which could highly activate PMS to oxidize BPA.PDS, another parent persulfate of SO4\n\u2022-, has a symmetric structure (-O3S-O-O-SO3-) and a short peroxide O-O bond (1.322\u00a0\u200b\u00c5), which makes PDS more stable. Compared with PMS, PDS is cheaper and lower toxicity. Moreover, PDS has no significant effect on pH value of water. Therefore, PDS has received widespread applied for refractory organic pollutants [63\u201365]. Huang et\u00a0al. [63] doped single-atom Fe in the plane of 2D MoS2 nanosheets. The catalyst displayed efficient catalytic activity and stability for propranolol degradation via activation of PDS, which attributed to the catalytic sites and the interaction between Fe and Mo. Du et\u00a0al. [65] anchored single iron atom on nitrogen-doped carbon (SAFe-N-C) for activating PDS to degrade chloramphenicol (CAP), the removal rate of SAFe-N-C outstandingly enhanced in degradation of CAP compared to N-C (from 15.3% to 93.1%). Instead of generating free radicals, Fe-N\nx\n and graphitic N active sites were oxidized by PDS to produce 1O2 on the catalyst, which immediately oxidized CAP into small molecules.Considering the problems encountered by using one catalytic technology alone in the environmental remediation process, coupling with different catalytic technologies can realize multifunctional co-catalytic strategies, which are extensively applied to degrade pollutants [3,4]. Commonly used coupling technologies include photo-Fenton, electro-Fenton, photo-PS, ozonation-PMS, piezo-PMS, and others [66\u201374]. Compared with the single catalytic technique, the coupled catalytic technology has perfect catalytic performance and unique reaction mechanism (Fig.\u00a08\n).For example, in order to overcome the high consumption of H2O2 in Fenton-like reaction, combination photocatalysis and electricity with Fenton-like reaction have been adopted, which can facilitate electron transfer in catalytic process [66,67]. Su et\u00a0al. [66] designed a defect engineered single Fe atom catalyst (Fe1-Nv/CN). They reported that nitrogen vacancies in Fe1-Nv/CN could act as electron trap sites, which can actuate the photoelectrons to concentrate on single Fe atoms for \u2022OH production under light irradiation, which could restrain the single-atom catalyst per se from quenching \u2022OH and promote the conversion efficiency of H2O2. As a result, the Fe1-Nv/CN catalyst displays a higher ciprofloxacin degradation activity under light irradiation. Liu et\u00a0al. [69] supported single-atomic-site Cu on carbon nitride (CN) by a pyrolyzing coordinated polymer strategy, and compared the degradation performance of tetracycline (TC) in SAS-Cu1.0/PS/Light and other similar systems. The SAS-Cu1.0/PS/Light system displayed outstanding catalytic performance for degradation of TC compared to other systems, indicating that light, catalyst and PS play a synergistic role in the degradation progress, which could generate abundant active species (O2\n\u2022-, SO4\n\u2022-, \u2022OH, 1O2). Generally, the reaction between O3 and H2O2 can spontaneously generate \u2022OH, but the reaction rate constant is extremely limited under acidic conditions. In this regard, Guo et\u00a0al. [74] designed a novel catalyst by anchoring single Mn atoms on g-C3N4, which effectively generated \u2022OH in situ through a new path in acid solution (pH=3) in the peroxone process, thus showing excellent activity and stability for oxidation of organic pollutants. Lan et\u00a0al. [13] immobilized Fe single atom on the surface of piezoelectric MoS2 nanosheet and then coupled the piezoelectric activation of PMS in Fenton-like reaction to degrade various contaminants. The introduction of Fe atomic sites could not only enhance the piezoelectric polarization and the piezoelectric charge separation of MoS2, but also improve the activation of PMS during the piezoelectric catalysis.Over the past decade, the introduction of SACs in environmental remediation has been considered as an effective strategy to degrade various organic pollutants due to their high intrinsic activity. In this review, the technical classification (including photocatalysis, electrocatalysis, Fenton-like, PS and coupling reactions) of SACs in environment application is summarized. Concurrently, the reactive mechanism of SACs in different catalytic reactions is discussed. Generally, SACs exhibit perfect catalytic properties due to the synergistic effects between metal single atom and the supports materials. Owing to the optimized mono-atom dispersion properties, the unique electronic properties and the formation of covalent bands between metallic and nonmetallic atom, SACs display unique advantages such as maximizing the utilization efficiency of metal atoms, excellent stability, ultra-low metal loading, as well as perfect catalytic activities in environment remediation. What's more, the development of advanced identification methods for SACs has made it possible to study mechanism at atomic level, which endows deep understanding of the basis of catalytic engineering.Undoubtedly, SACs show excellent catalytic performance in different catalytic technology for environment remediation, but the problems of stability and selectivity of SACs still need be overcome and addressed in the future (Fig.\u00a09\n).\n\n1.\nDuring the process of water treatment, the reaction process is carried out in a suspension, which makes it difficult to separate and recover SACs from the suspension. Therefore, how to immobilize the SACs in water treatment system is a challenge. Some effective strategies have been applied to improve the recovery and reuse ability of SACs in water treatment applications. For example, SACs can be coated on the surface of carbon materials (carbon felt, carbon cloth, functional cotton fiber, and so on) with a porous structure, and can be used as filters in water treatment processes. Moreover, SACs can be wrapped in magnetic materials to solve the problem of separation and recovery of catalysts in water treatment processes.\n\n\n2.\nTo avoid aggregation, the loading amount of metal atoms in SACs is generally minimal, which leads to unsatisfactory catalytic efficiency in practical application. Therefore, how to balance the catalytic activity and the limited single metal loading of SACs is of great significance.\n\n\n3.\nAlthough the single metal atoms can be fixed on the surface of supports through strong covalent bands, they are still easy to fall off during catalysis, leading to inactivation of the active site. Hence, it is necessary to solve the problem of the structural stability of SACs.\n\n\n4.\nDuring catalysis process, abundant highly reactive oxygen species will indiscriminately attack both the contaminants and the substrate material, which threatening the stability of SACs. Hence, how to improve the selectivity of SACs is an urgent problem to be solved.\n\n\nDuring the process of water treatment, the reaction process is carried out in a suspension, which makes it difficult to separate and recover SACs from the suspension. Therefore, how to immobilize the SACs in water treatment system is a challenge. Some effective strategies have been applied to improve the recovery and reuse ability of SACs in water treatment applications. For example, SACs can be coated on the surface of carbon materials (carbon felt, carbon cloth, functional cotton fiber, and so on) with a porous structure, and can be used as filters in water treatment processes. Moreover, SACs can be wrapped in magnetic materials to solve the problem of separation and recovery of catalysts in water treatment processes.To avoid aggregation, the loading amount of metal atoms in SACs is generally minimal, which leads to unsatisfactory catalytic efficiency in practical application. Therefore, how to balance the catalytic activity and the limited single metal loading of SACs is of great significance.Although the single metal atoms can be fixed on the surface of supports through strong covalent bands, they are still easy to fall off during catalysis, leading to inactivation of the active site. Hence, it is necessary to solve the problem of the structural stability of SACs.During catalysis process, abundant highly reactive oxygen species will indiscriminately attack both the contaminants and the substrate material, which threatening the stability of SACs. Hence, how to improve the selectivity of SACs is an urgent problem to be solved.In a word, there are still many challenges and opportunities to explore the feasibility of SACs in water treatment field in the future. This paper reviews the research progress on catalyst design, catalytic mechanism, different catalytic techniques and environmental applications over SACs, and makes efforts for its future industrial applications.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was funded by the Guangdong Basic and Applied Basic Research Foundation (Nos. 2020B1515020038), Science and Technology Program of Guangzhou (No. 202201020545) and Pearl River Talent Recruitment Program of Guangdong Province (2019QN01L148).", "descript": "\n                  Single-atom catalysts (SACs), consisting of metal single atoms and supporting materials, have shown remarkable potential due to their ultrahigh catalytic performances, maximum atomic utilization and environmental friendliness. More recently, SACs have become ideal catalyst materials and have been extensively applied in water treatment. This review summarizes the classification of advanced oxidation processes (AOPs, e.g., photocatalysis, electrocatalysis, Fenton-like reactions, persulfate oxidation and multi-technology coupling reaction) for the degradation of organic pollutants in water on SACs. The corresponding mechanisms for the removal of organic pollutants over SACs in the above technologies are also discussed. Distinguished from traditional nanoparticles and nanoclusters, the unique electronic properties of single metal atoms and the formation of covalent bands between metallic and nonmetallic atom promote the rapid generation of reactive oxygen species (SO4\n                     \u2022-, O2\n                     \u2022-, \u2022OH and 1O2), which endow SACs with excellent removal efficiency of organic pollutants. Finally, the opportunities and challenges of SACs applied in practical water treatment are proposed.\n               "}
{"full_text": "Data will be made available on request.An aqueous-phase bio-cured oil separated from the pyrolysis oil of woody biomass contains up to 60% of the carbon in the original biomass [1]. The aqueous fraction is composed of various oxygenated organic compounds, mainly small carbonyl compounds such as ketones, aldehydes, and organic acids [2,3]. Recently, researchers have begun to valorise carbon contained in aqueous-phase crude oil such as small olefins and aromatics with HZSM-5 via catalytic conversion [1].In line with the growing demand for bio-jet fuel [4], studies to prepare fuel precursors through C\u2013C bonding reactions (aldol condensation) of small oxygenates in the aqueous fraction are also being conducted [5,6]. These medium-chain fuel precursors are transformed to liquid biofuel (n-alkanes) through a hydro-deoxygenation process [7]. The medium-chain alkanes produced in this manner are suitable for the carbon range of jet fuel, and further conversion into branched alkanes through isomerization is required to obtain high quality bio-jet fuel (Fig. A1\n). For reference, since n-alkanes obtained from oil-based biomass are mainly long-chain hydrocarbons (C16 and C18), hydro-upgrading (hydro-cracking and isomerization) is performed in the last step to obtain high-yield and high-quality jet fuel. The catalyst used for hydro-upgrading is a bi-functional catalyst (Pt/HY, Pt\u2013Mg/HY, Pt\u2013Pd/HB, Pt/SAPO-11, etc.) wherein active metal and acid sites coexist and the cracking and isomerization reaction of paraffin occur simultaneously according to the reaction mechanism [8\u201310].However, unlike the existing hydro-upgrading catalysts suitable for long chain alkanes, more careful catalyst design is required for the isomerization of medium-chain hydrocarbons (C8, C13, etc) obtained through the C\u2013C bonding reactions of small oxygenates. This is because, during the isomerization reaction on the bi-functional catalyst, a cracking reaction of the medium-chain alkane is accompanied, thereby lowering the yield of jet-fuel. Therefore, in order to obtain a high-yield jet fuel while preserving the carbon number in medium-chain alkanes synthesized from small oxygenates, a catalyst in which isomerization is dominant rather than cracking of the medium-chain alkanes is required. In particular, it is important to find a composition with excellent catalytic performance at low temperatures, since hydro-cracking easily occurs at high temperatures due to endothermic reactions [11].Many studies have been conducted to improve the yield of isomers by changing the catalyst properties in the isomerization reaction of medium-chain alkanes [11,12]. As a result of performing the isomerization reaction of n-dodecane with Pt/SAPO-11 prepared under different synthesis condition for SAPO-11 (particle sizes of 65\u00a0nm to 4.5 \u339b), it is found that both suitable acidity and suitable particle size of SAPO-11 for shorter diffusion path are closely related to the yield of isomers [12]. In the isomerization reaction of n-octane, Ni\u2013Cu/SAPO-11 was prepared to suppress hydrogenolysis of n-alkane [11]. Hydrogenolysis was inhibited by reducing the active ensemble size by diluting the active metal (Ni) with inactive Cu, resulting in an isomer yield of about 63% at 340\u00a0\u00b0C. Low-temperature isomerization of n-hexadecane was performed using Pt\u2013Pd/HB (Si/Al\u00a0=\u00a025) with different Pt and Pd ratios. Bi-metallic catalysts showed an enhanced metal dispersion (92%) relative to mono-metallic catalysts (54%) and the conversion was also increased along with the metal dispersion (65%\u201377%) [13].A small amount (0.5\u00a0wt%) of rare earth metals (Ce, La, and Re) were loaded on Pt/ZSM-22 to suppress the Pt sintering during reduction treatment since the Pt dispersion is important in the low-temperature isomerization of alkane [14]. Ce or La oxides helped to protect the nano-sized Pt metal and induced an electron-deficiency state of Pt. However, the Ce-modified Pt/ZSM-22 showed better isomerization performance than the La-modified catalyst due to the strong interaction of a Pt\u2013O\u2013Ce bond. Rare earth metal (Ce, La, Nd, and Yb) - exchanged Pt/HB catalysts were also prepared [15,16]. The Ce-exchanged HB (0.2\u20130.8\u00a0wt% loading) exhibited higher conversion and selectivity for isomerized products than the parent HB catalyst due to the reducibility of Pt species facilitated by the ion-exchanged Ce [16]. On the other hand, in the case of La-exchanged Pt/HB, only a small amount of loading (0.3\u00a0wt%) had a slight positive effect on the isomerization performance due to the formation of new Lewis acid sites [15]. As such, in the bi-functional catalytic system, the catalytic performance for the isomerization reaction of n-paraffin is greatly affected by the pore size and acid site strength and density [12], the residence time of the olefinic intermediate in the pores [17,18], the metal dispersion (distance between metals) [11,13], and the metal/acid balance [17,19], as well as the reaction conditions, and studies are being conducted to increase the degree of isomerization at relatively lower temperatures. Nevertheless, it is still necessary to explore catalyst compositions that can maximize the isomer yield while minimizing the cracking reaction, in order to achieve high-yield and high-quality fuel through isomerization of medium-chain alkanes.In this work, a series of Pt\u2013La/HB (Beta zeolite, Si/Al\u00a0=\u00a038) catalysts were prepared for low-temperature isomerization of n-dodecane as a model compound of medium-chain alkane. In general, a Pt catalyst impregnated on an acidic support with a low Si/Al ration is more active but less selective for the isomerization reaction. However, beta zeolite with a low Si/Al ratio (B38) was selected to increase the conversion of n-alkane at relatively low temperatures and La was added on Pt/B38. The textural properties, metal dispersion, and acid properties of the La-loaded catalysts were characterized using various techniques and the accessible metal/acid ratio and an average acid step between two metal sites were discussed to highlight the importance of balancing the accessible metal and acid sites related to conversion and isomer yield.CP814C (B38) as a powder type beta zeolite was purchased from Zeolyst, and the zeolite was calcined before use at 500 \u05a0C for 1\u00a0h in air. As Pt and La precursors, chloroplatinic acid solution (Sigma-Aldrich, 5\u00a0wt%) and lanthanum nitrate hexahydrate (Samchun Chem., >98%) were used. The catalyst was prepared by the wetness impregnation method using a mixed aqueous solution of Pt and La precursors, followed by calcination at 450 \u05a0C for 4\u00a0h in air. To determine the effect of La addition, the loading amount of Pt was fixed at 1\u00a0wt%, and the prepared catalysts were denoted as follows: Pt\u2013La'x\u2019/B38, where x is the tentative La content.To determine the textural properties of the prepared catalyst, a Brunauer-Emmett-Teller (BET) analysis was conducted by using 3Flex (Micromeritics Co., LTD.). The sample was degassed at 200 \u05a0C overnight, and was cooled to \u2212196 \u05a0C for N2 adsorption. The structural property was estimated by an X-ray diffraction analysis (XRD, SmartLab 9kW/Rigaku Co. Ltd.). To observe the acidity of the catalyst's surface, NH3-temperature programmed desorption (NH3-TPD) and Fourier transform infrared spectroscopy (FTIR) were utilized by using a BELCAT II catalyst analyzer (BELCAT) and a Nicolet iS50+ (Thermo Scientific), respectively. In the case of NH3-TPD, NH3 was adsorbed on the sample at 50 \u05a0C, and then desorbed with increasing temperature to 800 \u05a0C. Meanwhile, pyridine, employed as a probe molecule in the FTIR analysis, was adsorbed at 100 \u05a0C, and then desorbed. In the desorption profiles of pyridine at below 300 \u05a0C, two bands overlap around 1445\u20131444\u00a0cm\u22121, corresponding to pyridine interacting with the hydroxyl group and pyridine bonded to a relatively stronger Lewis site, respectively. Thus, we chose the pyridine desorption profile of 300 \u05a0C to confirm the strong Lewis site because the pyridine bonded to the hydroxyl group disappears below 300 \u05a0C. A H2-temperature programmed reduction (H2-TPR) experiment was carried out to evaluate the temperature at which hydrogen consumption for Pt oxides occurred using an AutoChem II (Micromeritics). The contents of Pt in the prepared catalysts were determined by using inductively coupled plasma \u2013 optical emission spectroscopy (ICP-OES, OPTIMA7300DV/PerkinElmer), and a CO-pulse chemisorption (BELCAT II catalyst analyzer/BELCAT) was used for measuring the Pt dispersion and size. Visual images of the catalysts were obtained by using a field emission-scanning transmission electron microscope (FE-S/TEM, HF5000/Hitachi Co. Ltd.) equipped with an energy dispersive spectrometer (EDS). Before the TEM analysis, the catalysts were reduced at 350 \u05a0C for 1\u00a0h with H2.Low-temperature hydro-isomerization was conducted with low temperature and atmospheric pressure in a continuous fixed bed reaction system. The prepared catalyst (400\u00a0mg) was inserted into the tubular reactor (quartz, I.D. 5\u00a0mm), and then the reactor was mounted on a furnace. Hydrogen (Special Gas Co. Ltd., 99.999%) was injected with a flow of 100\u00a0ml/min into the reaction system. The reactor was heated to 350 \u05a0C for the catalyst's reduction, and then the hydro-isomerization was progressed at 160\u2013340 \u05a0C. The temperature deviation was maintained below 1%. As a model compound, n-dodecane (Sigma-Aldrich, >99%) was employed and pumped into the reaction system with 5.85 h\u22121 weight hourly space velocity (WHSV) by an HPLC pump (Optos 1HM, Eldex). The liquid sample was collected by using a cold trap, and the trap was cooled by ice to condense the products derived from the reactor.The collected sample was analyzed by a gas chromatograph (GC, Simadzu Co. Ltd./GC-2010 plus) equipped with a frame ionization detector (FID) and mass spectroscopy (MS, Simadzu Co. Ltd./GCMS-QP2010 SE). Both instruments used the same RTX-5 column (30\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm x 0.25\u00a0\u03bcm) and oven program (held to 50 \u05a0C for 1\u00a0min, heated to 220 \u05a0C (10 \u05a0C/min and held for 1\u00a0min), and further heated to 280 \u05a0C (15 \u05a0C/min and held for 3\u00a0min)). To quantitatively measure unconverted n-dodecane and isomerized products (iso-C12), n-pentadecane (Aldrich, >99%) was used as an internal standard, and the response factor of isomerized products was assumed to be the same with n-dodecane. The terminologies were defined by the following equations:\n\n\n\nC\no\nn\nv\ne\nr\ns\ni\no\nn\n\n\n(\n%\n)\n\n=\n\n\nt\nh\ne\n\na\nm\no\nu\nn\nt\n\no\nf\n\nc\no\nn\nv\ne\nr\nt\ne\nd\n\nn\n\u2212\nd\no\nd\ne\nc\na\nn\ne\n\n\n(\ng\n)\n\n\n\nt\nh\ne\n\na\nm\no\nu\nn\nt\n\no\nf\n\ni\nn\nj\ne\nc\nt\ne\nd\n\nn\n\u2212\nd\no\nd\ne\nc\na\nn\ne\n\n\n(\ng\n)\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n\n\ni\ns\no\n\u2212\n\n\nH\nC\n\n12\n\n\ny\ni\ne\nl\nd\n\n\n(\n%\n)\n\n=\n\n\nt\nh\ne\n\na\nm\no\nu\nn\nt\n\no\nf\n\np\nr\no\nd\nu\nc\ne\nd\n\ni\ns\no\n\u2212\n\nC\n12\n\n\n\n(\ng\n)\n\n\n\nt\nh\ne\n\na\nm\no\nu\nn\nt\n\no\nf\n\ni\nn\nj\ne\nc\nt\ne\nd\n\nn\n\u2212\n\nC\n12\n\n\n\n(\ng\n)\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n\n\ni\ns\no\n\u2212\n\n\nH\nC\n\n12\n\n\ns\ne\nl\ne\nc\nt\ni\nv\ni\nt\ny\n\n\n(\n%\n)\n\n=\n\n\nt\nh\ne\n\na\nm\no\nu\nn\nt\n\no\nf\n\np\nr\no\nd\nu\nc\ne\nd\n\ni\ns\no\n\u2212\n\nC\n12\n\n\n\n(\ng\n)\n\n\n\nt\nh\ne\n\na\nm\no\nu\nn\nt\n\no\nf\n\nc\no\nn\nv\ne\nr\nt\ne\nd\n\nn\n\u2212\n\nC\n12\n\n\n\n(\ng\n)\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n\n\n\n\ni\ns\no\n\u2212\n\n\nH\nC\n\n12\n\n\n\nn\n\u2212\n\n\nH\nC\n\n12\n\n\n\n\nr\na\nt\ni\no\n\n\n(\n\u2212\n)\n\n=\n\n\ni\ns\no\n\u2212\n\n\nH\nC\n\n12\n\n\ni\nn\n\np\nr\no\nd\nu\nc\nt\n\n\n(\ng\n)\n\n\n\nn\n\u2212\nH\n\nC\n12\n\n\ni\nn\n\np\nr\no\nd\nu\nc\nt\n\n\n(\ng\n)\n\n\n\n,\na\nn\nd\n\n\n\n\n\n\n\n\n\n\nn\nP\nt\n\n\nn\nB\nA\n\n\n\n(\n\u2212\n)\n\n=\n\n\nt\nh\ne\n\na\nm\no\nu\nn\nt\n\no\nf\n\ne\nx\np\no\ns\ne\nd\n\nP\nt\n\nm\ne\nt\na\nl\n\n(\n\n\nm\nm\no\nl\n\ng\n\n)\n\n\n\n\nt\nh\ne\n\na\nm\no\nu\nn\nt\n\no\nf\n\nm\ne\nd\ni\nu\nm\n\u2212\ns\nt\nr\ne\nn\ng\nt\nh\n\nB\nr\n\u00f8\nn\ns\nt\ne\nd\n\na\nc\ni\nd\n\ns\ni\nt\ne\ns\n\n(\n\n\nm\nm\no\nl\n\ng\n\n)\n\n\n\n\n\n\n\nThe textural properties and Pt dispersion of the prepared catalysts are summarized in Table 1\n. The specific surface area and total pore volume decreased with the impregnation of Pt and La on the bare B38 support. In particular, an increase in the amount of impregnated La had a greater effect on the reduction of micropore volume than the meso/macropore volume. As a result, the mean pore diameter increased from 2.15\u00a0nm to 2.29\u00a0nm. This reduction in micropores and increase in mean pore diameter are advantageous in terms of reducing the cracking reaction of olefinic intermediates, because retention of the intermediate in micropores induces cracking reactions at acid sites. The Pt dispersion of the Pt/B38 catalyst was 41.7%, whereas the Pt dispersion increased to more than 60% when impregnated with La on Pt/B38.The XRD patterns of the prepared catalyst are shown in Fig. A2\n. Compared with the XRD pattern of B38, the intensity of the characteristic peak of B38 decreased as the loading amount of La on the catalyst increased. Also, no characteristic peaks related to La were found even when 12\u00a0wt% of La was impregnated on the B38. This indicates that an amorphous lanthanum oxide (LaOx) was formed, which can be confirmed from the TEM images presented later. FE-STEM/TEM and EDS images of the prepared catalysts (Pt/B38, Pt\u2013La2/B38, and Pt\u2013La10/B38) are shown in Fig. 1 and Fig. A3. In the case of Pt/B38 (Fig. 1(a) and Fig. A3(a)), large Pt particles (approximately, 26\u201336\u00a0nm) are exposed to the external surface of the support, and small size Pt particles are distributed between them. Small Pt appears to be present within the pores rather than on the surface of the catalyst particles. The low Pt dispersion confirmed by the CO pulse method (Table 1) appears to be due to the large Pt particles exposed to the outside. As shown in Fig. 1(b) and Fig. A3(b), when Pt and a small amount of La were impregnated on B38, large and small Pt particles were also mixed. However, it is observed that, unlike Pt/B38, the size of Pt exposed to the outside was drastically reduced (approximately, 6\u201314\u00a0nm). As observed in the EDS mapping of the Pt\u2013La2/B38, Pt is distributed evenly on the support with La. It appears that co-impregnated La improves the dispersion of Pt, as shown in Table 1 (Pt dispersion). When 10\u00a0wt% of La was co-impregnated with Pt on B38 (Pt\u2013La10/B38), all large-sized Pt particles disappeared and nano-sized Pt particles (\u223capproximately, 3.3\u00a0nm) were uniformly distributed (Fig. 1(c) and Fig. A3(c)). It thus can be seen that as the amount of co-impregnated La increases, the uniformity and dispersion of nano-sized Pt particles appear to increase. However, unlike the STEM/TEM images, the Pt dispersion (63.7\u201368.4%) measured by the CO pulse (Table 1) was similar regardless of the amount of La loading for the Pt\u2013La5\u223c12/B38 catalysts. This appears to be due to the Pt size gradually decreasing when the amount of La loading exceeds a certain amount (about 5\u00a0wt%), whereas as the catalyst surface is covered with lanthanum, the number of externally exposed Pt (accessible Pt) decreases little by little. However, they still show higher dispersions (more than about 63.7%) than Pt/B38.As shown in the STEM image (Figs. 1(c-3)), amorphous LaOx was observed, which is consistent with the XRD result where La-related peaks did not appear (Fig. A2). Nevertheless, as shown in both EDS mapping (Figs. 1(c-4)) and line-EDS (Fig. A3(c)), La and Pt are well distributed throughout.H2-TPR was employed to further confirm the reduction in size of Pt particles by La addition to Pt/B38 (Fig. A4). For Pt/B38, the reduction peaks centered at 80\u00a0\u00b0C and 355\u00a0\u00b0C were attributed to the reduction of PtOx particles loaded on the external surface and dispersed in the internal pores of B38, respectively [19]. The reduction temperature for PtOx loaded on the external surface of B38 was gradually shifted to higher temperatures (above 200\u00a0\u00b0C) as the loading amount of La increases. This indicates that the size of PtOx particles loaded on the external surface of the support becomes smaller, which is accordant with the results of the FE-S/TEM analysis. However, excessive loading of La (12\u00a0wt%) reduced the amount of Pt species exposed on the surface, and thus the hydrogen consumption was relatively reduced. La thus appears to play a role in preventing agglomeration of co-impregnated Pt and facilitating uniform nano-size dispersion.\nTable 2\n summarizes acid properties of the prepared catalysts, based on the results of NH3-TPD and pyridine-FTIR (Fig. A5). When 2\u00a0wt% of La was loaded on Pt/B38, the total acid sites increased. However, as the amount of La was further increased, the amount of acid sites decreased slightly because the amorphous lanthanum oxides covered the acid sites. (Fig. 1 and A3). However, they still have more acid sites than B38 or Pt/B38. In addition, the Br\u00f8nsted acid sites related to the skeleton isomerization of the reactant also decreased according to the La loading, while Lewis acid sites were generated due to the formation of amorphous LaOx. Considering the ratio of accessible Pt and Br\u00f8nsted acid sites (nPt/nBA), Pt\u2013La2\u223c10/B38 had a higher nPt/nBA value (0.117\u20130.132) than that of Pt/B38 (0.076), but there was not a significant difference among the values. It is noteworthy that although nPt/nBA shows similar values, the size of accessible Pt becomes smaller as the amount of loaded La increases (Fig. 1, Fig. A3, and Fig. A4).The desired isomerization reaction entails dehydrogenation of n-alkane on Pt site, skeleton rearrangement of olefinic intermediate on active acid sites, and hydrogenation of iso-olefin on Pt site occurring continuously and smoothly, while preserving the number of carbons in the reactant. Fig. 2 shows the conversion and the selectivity of iso-dodecane (a), the yield of iso-dodecane (b), and the distribution of product (c and d) for B38, Pt/B38, Pt\u2013La10/B38, and La10/B38. B83 and La10/B38 were inactive for isomerization of n-dodecane at reaction temperature below 280\u00a0\u00b0C, but n-dodecane started to be converted above 300\u00a0\u00b0C where a severe cracking reaction was predominant, and thus the liquid product was hardly recovered. In the case of Pt/B38, n-dodecane stared to be converted at 180\u00a0\u00b0C and showed 100% conversion at 240\u00a0\u00b0C. However, most cracked hydrocarbons with less than five carbons and gases were generated in the high conversion section (Fig. 2(c)), and as the conversion increased according to the reaction temperature, the selectivity of iso-C12 fell inversely. As a result, the maximum yield (22.9%) of the desired iso-C12 was obtained at 200\u00a0\u00b0C with a conversion of about 30% (Fig. 2(b)). For Pt\u2013La10/B38, the conversion was delayed by about 20\u00a0\u00b0C compared to that of Pt/B38, but the selectivity of iso-C12 was maintained high in the range of 200\u2013250\u00a0\u00b0C. The selectivity of iso-C12 at 260\u00a0\u00b0C, which showed 100% conversion, decreased sharply, and the maximum iso-C12 yield (59.2\u201356.2%) was thus obtained in the range of 240\u2013250\u00a0\u00b0C (Fig. 2(b)). As shown in Fig. 2(c), distributions (mono-branched and multi-branched isomers and cracked hydrocarbons) in the product varied according to the conversion of n-dodecane over the bi-functional catalyst. n-dodecane was mainly transformed into cracked hydrocarbons over B38 acid catalyst. In the case of Pt/B38, isomers were mainly produced at low conversion (less than about 30%), but as the conversion increased, cracked products became the dominant species due to their secondary transformation. This appears to be due to poor hydrogenation resulting from low nPt/nBA (0.076). This trend was similar when La was co-impregnated with Pt on B38, but the conversion at which the secondary transformation began to appear in Pt\u2013La10/B38 was delayed by about two times (at about 60% conversion). For comparison, hydro-isomerization of n-dodecane was performed using a Pt/SAPO-11 which is one of the well-known catalysts for hydro-isomerization (not shown here). The conversion of 64.9% and the isomer yield of 52.1% were obtained at 300\u00a0\u00b0C. Thus, Pt\u2013La10/B38 shows better catalytic performance even at 250\u00a0\u00b0C than Pt/SAPO-11 that have been studied recently.The ratio of cracked product and isomer (C/I ratio) and the ratio of multi- and mono-branched isomers (multi/mono ratio) are provided in Fig. 2(d). Both C/I and multi/mono ratios of Pt\u2013La10/B38 were lower than those of Pt/B38. This means that skeleton isomerization and hydrogenation of n-dodecane are balanced on the Pt\u2013La/B38 catalyst. In conclusion, as shown in Table 2, it can be seen that the high nPt/nBA (0.117) of Pt\u2013La10/B38 is suitable for producing iso-C12 under the present reaction conditions, where the carbon number of the feed (n-dodecane) is preserved while minimizing the cracking reaction in the range of low reaction temperature.The average number of active acid sites that one n-dodecane molecule encountered during the catalytic reaction, nas, was estimated based on the initial product distribution to observe the characteristics of the diffusion path of olefinic intermediates between two Pt metal sites and the results are summarized in Table 3\n [18]. B38 was converted at high temperature (above 280\u00a0\u00b0C in Fig. 2(a)) because there was no metal active site for dehydro/hydrogenation, and mostly cracked products were produced, traveling approximately 3.8 acid steps. When Pt was impregnated on B38, isomers were mainly generated in the initial reaction, and the number of active acid sites involved in the transformation was greatly reduced. In the case of Pt\u2013La10/38, the nas value was close to 1.1. This is considered a result of the reduced the distance between nano-sized Pt particles, as can be seen from TEM images and Pt dispersion. This is consistent with the results of previous studies [18]. However, it is difficult to find a significant difference between nas values estimated from the initial reaction of two catalysts (Pt/B38 and Pt\u2013La10/B38). This is because the possibility that inactive acid sites exist under the operating conditions (180 and 200\u00a0\u00b0C) cannot be excluded. This is confirmed by observing the change of nas values according to the reaction temperature, shown in Fig. A6. As the conversion of n-dodecane increased, the nas values of two bifunctional catalysts increased. This indicates that the increase of the reaction temperature causes more acid sites to become active, and also increases the diffusivity of the olefinic intermediates. However, it is clear that Pt\u2013La10/B38 still shows lower nas values than Pt/B38 at similar conversions, and the gap between the two values increases as the temperature (conversion) increases. This means that in the Pt\u2013La10/B38 catalyst with well dispersed nano-sized Pt particles, even if the reaction temperature rises, the average number of active acid sites that the intermediates encounters while traveling through the surface of the catalysts is small due to the short distance between Pt particles.\nFig. 3 shows the results of hydro-isomerization of n-dodecane according to the La content impregnated on the Pt/B38 at 250\u00a0\u00b0C. Although the conversion was close to 100% on Pt/B38 at 250\u00a0\u00b0C, the recovered iso-C12 yield was very low due to the predominant cracking reaction (Fig. A7). As the amount of La loading increased, the conversion of dodecane decreased, but the selectivity and yield of the desired iso-C12 tended to increase. The maximum yield of iso-C12 was obtained in the 10% La impregnated catalyst.The reason for the change in catalytic properties (conversion, selectivity and yield) in the Pt\u2013La series catalysts even with similar nPt/nBA values is (1) an increase of uniformity of nano-sized Pt and (2) the smaller distance between two Pt particles caused by La loading, resulting in predominant hydrogenation of the olefinic intermediates. That is, the optimal loading of La reduce the distance between Pt particles and enhances the intimacy between the uniformly distributed nano-sized Pt and the active acid sites. A further increase of the La loading (more than 12\u00a0wt%) resulted in lowered catalytic performance, likely due to the coverage of the Br\u00f8nsted acid and Pt sites by LaOx with Lewis acid sites, as discussed in the results of FE-S/TEM and H2-TPR, consistent with a previous study showing that the conversion is related to the externally exposed Pt over the catalyst [19]. For the Pt\u2013La2\u223c10/B38 catalysts, the ratio of iso-C12/n-C12 was greater than 2.1. Note that the ratio decreases as the content of La increases because uncovered n-dodecane remains.\nFig. 4\n\n\n shows the time-on-stream stability of Pt\u2013La10/B38 in hydro-isomerization of n-dodecane. Initial conversion and isomer yield were 86% and 57%, respectively, but the isomer yield gradually decreased (46%) until 4\u00a0h of time-on-stream. After shutting down the reaction system, the catalyst was in-situ cleaned with acetone without any other regeneration process. In the subsequent reaction, the conversion and isomer yield were maintained within 75\u201380% and 55\u201360%, respectively, which means that the catalyst can be regenerated only by washing with acetone.Based on the above results, the hydro-isomerization reaction over the bifunctional catalysts is schematically drawn in Fig. 5. The mechanism of hydro-isomerization on the bi-functional catalyst has been well documented in several studies [13]. In general, the n-alkane is dehydrogenated to the alkene intermediates on the Pt active site and the intermediates are protonated and isomerized on the Br\u00f8nsted acid sites, and then hydrogenated on the nearby Pt active site to form a desired iso-alkane. Meanwhile, the low nPt/nBA value and the diffusion limitation of the intermediates in the microporous channels of acid supports increases the contact opportunity and the contact time with the acid sites, resulting in further cracking to obtain unwanted cracking products. That is, a proper arrangement of metal and acid sites is very important in bi-functional catalysts, as well as the textural structure of the catalysts. For Pt\u2013La10/B38, the size of Pt was very small and uniform. Such uniformly distributed nano-sized Pt particles (\u223capproximately, 3.3\u00a0nm) give the reactant many opportunities for dehydro/hydrogenation reactions, resulting in the production of many olefinic intermediates at the same time. The intermediates undergo skeletal rearrangement at an adjacent acid site and then hydrogenate at a nearby Pt site to form isomers. In addition, the La loading reduces the volume of micropores, thereby suppressing the cracking reaction that occurs from the diffusion limitation of the intermediates.A series of Pt\u2013La/B38 catalysts were prepared for low-temperature isomerization of n-dodecane and the effect of La addition on the catalytic performance (conversion, selectivity and yield) was investigated. Based on the results of CO pulse chemisorption, FE-S/TEM images, and a H2-TPR analysis, La (10\u00a0wt%) co-impregnation in Pt/B38 resulted in an increase of the Pt dispersion (41.7%\u201368.4%). Moreover, the size of Pt loaded on the external surface of B38 was significantly reduced from approximately 26\u201336\u00a0nm\u2013\u223c3.3\u00a0nm and the uniformity in Pt size was notably enhanced. La thus appears to play a role in preventing agglomeration of co-loaded Pt and facilitating uniform nano-size Pt dispersion. Pt/B38 showed good activity at relatively low reaction temperature (100% conversion at 220\u00a0\u00b0C), but most cracked hydrocarbons with less than five carbons and gases were generated in the high conversion section, resulting in a maximum yield, 22.9%, for the desired iso-C12. Meanwhile, for Pt\u2013La10/B38, the conversion was delayed by about 20\u00a0\u00b0C compared to that of Pt/B38, but the selectivity of iso-C12 was maintained high in a range of 200\u2013250\u00a0\u00b0C, resulting in the maximum iso-C12 yield (59.2\u201356.2%). Both the C/I and multi/mono ratios of Pt\u2013La10/B38 were lower than those of Pt/B38, indicating that skeleton isomerization and hydrogenation of n-dodecane are balanced on the Pt\u2013La/B38 catalyst. In conclusion, the desirable arrangement of active sites with higher nPt/nBA (0.1777) and lower nas (1.11) caused by La loading in the Pt\u2013La/B38 catalyst enhances the catalytic performance for low-temperature isomerization of n-dodecane.A National Research Foundation of Korea grant funded by the Korea government.\nIl-Ho Choi: Writing \u2013 review & editing, Writing \u2013 original draft, Methodology, Investigation, Formal analysis, Conceptualization. Hye-Jin Lee: Methodology, Formal analysis. Kyung-Ran Hwang: Writing \u2013 review & editing, Supervision, Project administration, Investigation, Conceptualization.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF2020M1A2A2079802).The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.micromeso.2022.112294.", "descript": "\n                  A series of Pt\u2013La/B38 catalysts were prepared for low-temperature isomerization of n-dodecane and the effects of La addition on the textural properties, metal dispersion, acid properties, and catalytic performance were investigated. La co-impregnated with Pt on B38 significantly reduced the Pt size and notably enhanced the uniformity in Pt size. A higher ratio of accessible Pt and medium-strength Br\u00f8nsted acid sites and a shorter distance between two Pt particles of Pt\u2013La10/B38 resulted in the maximum iso-C12 yield (59.2\u201356.2%) in the range of 200\u2013250\u00a0\u00b0C, due to the reasonable arrangement of active sites caused by La loading.\n               "}
{"full_text": "Data will be made available on request.Wastewater generation in urban cities is increasing steadily, and more than 48\u00a0% of global wastewater is discharged without treatment [1]. Many technologies, including aerobic and anaerobic treatment systems, have been reported for wastewater treatment. However, these conventional technologies have high energy requirements and are not self-reliant due to the need for an energy supply from fossil fuel sources [2]. Therefore, developing new water treatment technologies with integrated, efficient energy systems and sustainable practices will provide a positive approach toward a circular economy [3].One technology with the potential to significantly reduce the energy requirements in wastewater treatment plants is the microbial fuel cell (MFC). This bio-electrochemical system has gained widespread attention as a sustainable energy generation and wastewater treatment technology. In this system, the electroactive microorganisms are able to oxidize the organic substrates in wastewater under aerobic conditions, thereby releasing electrons and protons. The electrons are then transferred to the cathode via an external circuit, while the protons diffuse through a proton exchange membrane (PEM) [4]. The MFCs have demonstrated advantages such as low pressure and temperature operation and a minimal negative environmental footprint. However, their practical application has been limited by their low power generation, mainly due to the limited transfer of electrons from the microorganisms to the anode, and the high cost of their component materials [5]. Moreover, domestic wastewater contains many complex molecules that are difficult for the microorganisms to break down and oxidize [6]. Consequently, many studies have used synthetic, lab-produced wastewater. However, these synthetic wastewaters do not represent the actual characteristics of real wastewater and often require the addition of selective electroactive microorganisms cultured elsewhere. Therefore, to overcome the challenges of MFCs and make them compete with the conventional technologies for wastewater treatment and energy generation, new materials and designs must be developed to enhance the MFC performance, and these must be tested under operation using real wastewater for practical applications.The anode material of the MFC is a critical component in determining the power generation and enhancing the MFC performance due to its direct contact with microorganisms and its significant influence on the efficiency of the electron transfer process. Generally, the anode material should have a large surface area, electrical conductivity, good hydrophilicity, chemical stability, and biocompatibility [7]. Although various carbonaceous materials, including carbon cloth, felt, fiber, foam, reticulated vitreous carbon (RVC), and graphite, have been reported as MFC anode materials [8], they are limited by their high hydrophobicity, low surface area, and minimal porosity [9]. Consequently, many researchers have attempted to improve the properties of carbonaceous electrode materials by using various methodologies such as chemical, physical, and thermal treatment or coating with nanostructured materials [10\u201312]. In particular, the nanostructured coating of the anode is a simple and efficient technique. It provides a large surface area, porosity, stability, and selectivity for better interaction between the electrode and microorganisms, thus resulting in enhanced biofilm adhesion and acceleration of the electron transfer process. Metal and metal oxides such as Ni, Fe, Co, Au, Pd, Co2O3, MnO2, and Fe2O3 have been investigated extensively as anode catalysts for MFCs [13]. These materials possess remarkable properties, such as good electrical conductivity and electrocatalytic activity, for improving the performance of MFC [14,15]. However, their poor corrosion resistance and inadequate bacterial adhesion hinder their large-scale application in MFC [16]. It is noteworthy that the biocompatibility of metal and metal oxides as anode modification materials in MFC does not have a unified conclusion in the literature. Some metal and metal oxides, such as Fe, Fe2O3, TiO2, MnO2, FeO2 are reported to have good biocompatibility [12], while others, such as Cu, Ag, and Au, have limited biocompatibility with microbes [17]. Some authors hybridize metal and metal oxides with graphene-based materials to modify their properties for MFC applications [13].In this regard, tungsten nitride (W2N) and tungsten carbide (WC) have attracted considerable attention as specific transition metal groups due to their electrochemical stability, corrosion resistance, and platinum-like properties [18,19]. For example, previous work by the present authors investigated the use of WC on graphene oxide as an efficient anode catalyst for the MFC [20]. The results showed an enhancement in the generated power and current density due to increased surface wettability, surface area, and electric conductivity. Similarly, titanium nitride (TiN) nanoarray was in situ grown on carbon cloth (CC) for the anode in MFCs. The growth of TiN on the anode facilitated the enrichment of a large amount of Geobacter soli on the anode surface. The efficient growth of a large number of electroactive microorganisms was due to the metallic conductivity of TiN and its strong affinity towards microorganisms [21]. Towards the same aim, other researchers have investigated carbonaceous materials coated with three-dimensional macroporous structures, such as carbon foam, sponge, and carbon nanotubes, as high-performance anodes [22,23]. However, even though 3D microporous structures possess many advantages, their large-scale applications are hindered by the low interaction between the microorganisms and the anode surface, the complex fabrication procedures, and the high cost of these materials [15].MXenes, a class of two-dimensional inorganic materials with layers (a few atoms thick) consisting of transition metal carbides, nitrides, or carbonitrides, was first reported by researchers from Drexel University in 2011 [24]. MXenes have unique properties, such as good stability, high electrical conductivity, and rich surface functionalization, including a layered morphology for accelerated charge transfer and biocompatibility [25]. Consequently, the applications of MXenes in electrocatalysis [26], ion batteries, supercapacitors, water treatment, electronic devices, fuel cells, and bio-electrochemical systems have expanded rapidly [27,28]. Many studies have shown that MXenes are promising materials in electron transfer applications [29\u201331]. For example, due to their exceptional conductivity, they have been investigated for cathodic and anodic reactions in energy storage devices. This is due to the interlayer structure and oxygen termination groups that promote efficient electron transport [32,33].The first synthesized MXene was titanium carbide (Ti3C2\nTx\n), after which the family grew to include zirconium carbide (Zr3C2\nTx\n), molybdenum carbide (Mo2CTx\n), titanium nitride (Ti2N), niobium carbide (Nb2C), tantalum carbide (Ta4C3) [34]. Nevertheless, Ti3C2\nTx\n remains the most widely investigated MXene due to its more accessible synthetic pathways. This material is biocompatible, can bind strongly to carbonaceous substrates, and increases both the electron transfer rate and catalytic ability of the substrate [35,36]. However, despite the many outstanding properties of the reported MXenes, and the presence of rich surface functional groups, the fabrication of MXene-based composites via a simple and efficient process remains challenging due to the need for extreme synthesis conditions, such as high temperatures above 300\u00a0\u00b0C, which sometimes lead to damage of the crystal structure and the loss of its advantageous features [37,38]. In 2018 Ti3C2\nTx\n MXene was first reported as an anode for improved MFC [29]. This work showed an increase in power generation attributed to lower internal resistance and the promotion of microbial nanowires (pili), which increase the conductivity of the biofilm.Similarly, an MFC with a nickel ferrite/MXene on a carbon felt anode exhibited higher power density [31] due to more efficient electron transport and increased biological activity. MXene was also combined with a metal\u2013organic framework (MOF) named ZIF-67 to form a Ti3C2-ZIF-67 hybrid anode catalyst to improve the performance of MFCs. Due to the large specific surface area of MOFs, large pore size, biocompatibility, and particle size combined with the better hydrophilic surface and interlayer morphology of MXene, the power generation, and bacterial colonization were improved significantly [39]. The anolytes used in these works are synthetic wastewaters or solutions that do not represent the complexity of the wastewaters and lead to a different population of microorganisms, which then require that they are cultured separately.Herein, a W2N-Ti3C2\nTx\n composite catalyst (where Tx\n\u00a0=\u00a0O, F, and OH) is synthesized, characterized, and tested as the anode in an MFC to produce electricity during wastewater treatment. The aim is to improve the extracellular electron transfer and microorganism adhesion by depositing the as-synthesized material on a standard carbon cloth anode. The anolyte solution is real domestic wastewater without any additional inoculated microorganisms. The crystallinity, structural morphology, wettability, and biocompatibility of the composite catalyst are investigated via X-ray diffraction (XRD), scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HR-TEM), and atomic force microscopy (AFM). The composite catalysts coated on carbon cloth are tested in situ using a single-chamber air\u2013cathode microbial fuel cell. The present work highlights the great potential of such composite catalysts to enhance the performance of electrochemical systems.The Ti3C2\nTx\n MXene (where Tx\n represents surface functional groups such as \u2013O and \u2013F, OH) was produced by exfoliating the titanium aluminum carbide MAX phase precursor (Ti3AlC2) via the hydrofluoric acid (HF) etching approach shown schematically in Fig. S1a of the Supplementary Information\n[25,40]. In brief, the Ti3AlC2 MAX phase (1\u00a0g; \u226590\u00a0%, \u2264100\u00a0\u00b5m, Sigma-Aldrich) was added gently to a hydrofluoric acid solution (10\u00a0ml; 40\u00a0wt%, SungYoung Chemical Limited) with magnetic stirring. The solution was mixed for 6\u00a0h to complete the etching process. Following the reaction, the residual solids were separated by centrifugation and washed with deionized water until the pH of the supernatant was\u00a0\u223c\u00a07. The wet sample was then filtered under vacuum and dried at 80\u00a0\u00b0C overnight to obtain the MXene powder.The W2N was synthesized via the urea glass route, as shown schematically in Fig. S1b [20,41]. First, the metal precursor tungsten chloride (WCl6; 1\u00a0g, 99.9\u00a0% based on trace metals, Acros Organics) was dissolved in ethanol (2.54\u00a0ml). The solution was mixed for 30\u00a0min to form a stable brown W-orthoester solution. Then, urea (1\u00a0g) was added to the solution and mixed until a viscous solution was obtained. Finally, the solution was dried overnight at room temperature, followed by thermal treatment at 800\u00a0\u00b0C (with a ramping rate of 5\u00a0\u00b0C\u00a0min\u22121) under a flow of nitrogen (N2) for 4\u00a0h to obtain the desired product as a silvery black powder.To synthesize the W2N-MXene composite catalyst, 100\u00a0mg each of W2N and MXene were separately dispersed in ethanol (10\u00a0ml each) to form solutions A and B, respectively. Each solution was then sonicated for 1\u00a0h at room temperature, as shown schematically in Fig. S1c. Solution A was then gently added to solution B with stirring. The stirring was continued for 20\u00a0min until the ethanol had evaporated. The wet sample was then dried at 70\u00a0\u00b0C for 90 min, followed by heat treatment at 300\u00a0\u00b0C for 4\u00a0h under a flow of N2 at 5\u00a0\u00b0C\u00a0min\u22121, to yield a dark powder.The as-synthesized catalyst was coated onto carbon cloth via a simple ink-dropping approach, as shown schematically in Fig. S1c. First, the catalyst ink was prepared by adding the catalyst powder (10\u00a0mg) to a mixture containing 5\u00a0wt% Nafion binder (15\u00a0\u00b5l; Sigma-Aldrich), absolute ethanol (500\u00a0\u00b5l), and deionized water (500\u00a0\u00b5l). The mixture was then shaken several times, followed by sonication for 30\u00a0min to form a well-dispersed catalyst ink. Then, using a micropipette, the ink was repeatedly deposited onto the anolyte-facing side of the carbon cloth (Electro Chem. Inc., USA) until the entire surface was covered with catalyst ink. Finally, the sample was dried at room temperature to obtain the W2N-Ti3C2\nTx\n modified carbon cloth anode.The crystal structures of the as-synthesized catalysts were investigated under Cu K\u03b1 radiation using a Bruker D88 advanced diffractometer operated at 40\u00a0kV and 40\u00a0mA, in the angle range of 5\u00b0 to 90\u00b0 with a step size of 0.05. The morphologies of the as-synthesized composite catalysts were examined via field emission scanning electron microscopy (FE-SEM) with energy-dispersive X-ray spectroscopy (EDX; FEI TENEO VS) at an acceleration voltage of 20\u00a0kV and a working distance of 10\u00a0mm. High-resolution transmission electron microscopy (HR-TEM) was performed on an FEI-Titan ST electron microscope operated at 300\u00a0kV. After collecting the HR-TEM images, the same microscope was used to collect high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images and to characterize the elemental distributions on the as-synthesized catalysts. The thicknesses of the as-synthesized MXenes were determined by atomic force microscopy (AFM) using a dimension icon SPM scanner (Bruker) with an RTESPA-150 probe. Water contact angle analysis was performed at room temperature using a CAM 200 goniometer (KSV Instruments, USA). The formation of biofilm on the anode surface was investigated by the bio-SEM images using a Zeiss Merlin Gemini II microscope equipped with an in-lens critical point dryer (CPD300, Leica) and Everhart-Thornley detector (ETD).The ex-situ investigation of the prepared anodes was performed in a three-electrode cell comprised of the bio-anode, a Pt wire, and an Ag/AgCl electrode as working, counter, and reference electrodes, respectively, using a potentiostat cyclic voltammetry (CV) (Gamry Reference 600, United States of America) at a scan rate of 25\u00a0mV\u00a0s\u22121 from\u00a0+\u00a00.8\u00a0V to \u22120.8\u00a0V. Electrochemical impedance spectroscopy (EIS) was used to analyze the electron transfer behavior; the measurements were carried out using the same aforementioned electrochemical setup. The frequency range used for the analysis was 0.01\u00a0Hz to 100\u00a0kHz with an amplitude of 5\u00a0mV.As shown schematically in Fig. S2, the single chamber air\u2013cathode MFC reactor consisted of highly corrosion-resistant stainless-steel plate current collectors maintained at a constant separation of 2\u00a0cm, along with an anode composed of plain carbon cloth (CC), the Ti3C2\nTx\n/CC, or the W2N-Ti3C2\nTx\n/CC, and a 2.4\u00a0cm\u00a0\u00d7\u00a02.4\u00a0cm cathode consisting of platinum deposited onto carbon paper (Pt/CP), separated by a Nafion 115 membrane (Membrane International Inc. NJ, USA). All experiments were performed at room temperature and atmospheric pressure. In each case, the MFC was charged with domestic wastewater obtained from the KAUST wastewater treatment plant in Thuwal, Saudi Arabia. The anode chamber had a volume of 80\u00a0ml, which contained 60\u00a0ml of wastewater and 20\u00a0ml of glucose solution. Before each experiment, the anolyte chamber was purged with N2 to remove any dissolved oxygen and maintain anaerobic conditions. In contrast, free atmospheric oxygen was used as an electron acceptor in the cathodic compartment. The open circuit voltage (OCV) and electrode potentials were measured using a potentiostat and an Ag/AgCl reference electrode (HA-151B), respectively. All conducted measurements were recorded over time using a GL240-Graphtec data logger. After achieving a stable OCV, the circuit was closed using an external resistance, and the polarization curve was obtained via linear sweep voltammetry at a scan rate of 1\u00a0mV\u00a0s\u22121. The obtained current was used to calculate the power (P\u00a0=\u00a0IV). The current and power densities were obtained by normalizing the current and power to the anode geometric surface area (5.76\u00a0cm2), as given by Eq. (1) and Eq. (2), respectively. A current stability test was performed for 12\u00a0h at a constant cell voltage of 200\u00a0mV to evaluate the performances of the MFCs based on the as-synthesized catalysts.\n\n(1)\n\n\nC\nu\nr\nr\ne\nn\nt\nD\ne\nn\ns\ni\nt\ny\n\n\n\nm\nA\n\n\n\nm\n\n\n-\n2\n\n\n\n\n\n=\n\n\n\n\nI\n\nA\n\nan\n\n\n\n\n\n\n\n\n\n\n\n\n(2)\n\n\nP\no\nw\ne\nr\nD\ne\nn\ns\ni\nt\ny\n\n\n\nm\nW\n\n\nm\n\n\n-\n2\n\n\n\n\n\n=\n\n\n\n\n\nIV\n\n\nA\n\nan\n\n\n\n\n\n\n\n\n\nwhere I is the current generated, Aan\n is the anode geometric surface area, and V is the voltage of the MFC.The wastewater treatment efficiency was evaluated by measuring the COD before and after each MFC test. As given in Eq. (3), the coulombic efficiency (CE) was assessed based on the ratio of the recovered charge to the theoretical value, assuming that all substrate removal was converted entirely to electricity. Where 8 is the constant used for COD based on the ratio of the molecular weight of oxygen (O2) and the number of electrons exchanged per mole of O2 (b\u00a0=\u00a04), I is the current, F is Faraday\u2019s constant, Van is the volume of liquid in the anode compartment, tb is the time of MFC operation and \n\n\u0394\nC\nO\nD\n\n is the difference between the COD of the influent and effluent. The COD removal percentage was evaluated using the ratio of the difference between the influent and effluent CODs to the influent COD, as given by Eq. (4).\n\n(3)\n\n\nC\nE\n\n\n\n%\n\n\n\n=\n\n\n\n\n\n8\n\n\n\u222b\n\nt\nb\n\n\n0\n\nI\nd\nt\n\n\nF\n\nV\n\nan\n\n\n\u0394\nC\nO\nD\n\n\n\n\n\n\u2217\n100\n%\n\n\n\n\n\n\n(4)\n\n\nC\nO\nD\nr\ne\nm\no\nv\na\nl\n\n\n\n%\n\n\n\n=\n\n\n\n\n\n\n\nCOD\n\n\ninfluent\n\n\n-\n\n\nCOD\n\n\neffluent\n\n\n\n\n\nCOD\n\n\ninfluent\n\n\n\n\n\n\n\u2217\n100\n%\n\n\n\n\nThe Ti3C2\nTx\n material is structurally defined by the parent Ti3AlC2 MAX phase, which contains ionic-solid/covalent bonds between the Ti and C atoms, along with weaker metallic bonds between Ti and Al atoms. During the reaction, the etchant attacks the weak Ti\u2013Al bonds to selectively etch and remove the Al layer, while failing to break the Ti\u2013C bonds due to the high stability of the metal carbide in the acidic environment [42]. This is confirmed by the XRD pattern of the as-prepared Ti3C2\nTx\n and the Ti3AlC2 precursor in Fig. 1a, where the Ti3C2\nTx\n (red line) exhibits two significant peaks at 2\u03b8\u00a0=\u00a08.9\u00b0 and 18.2\u00b0 due, respectively, to the {002} and {004} crystal planes of the MXene [25] (see magnified peaks inseted in Fig. 1a). In contrast, the characteristic peak at 39\u00b0 due to the Al {104} crystal plane observed in the precursor (green line) is no longer visible after the etching process (red line). The peak at 41.5\u00b0 in the XRD pattern of the MAX phase precursor is attributed to the presence of TiC as an impurity. Meanwhile, the XRD pattern of the W2N-Ti3C2\nTx\n (black line, Fig. 1b) exhibits the peaks at 8.9\u00b0 and 18.2\u00b0 due to the Ti3C2\nTx\n, along with additional peaks at 38.35, 43.16, 61.97, 73.3, and 75.87\u00b0 due, respectively, to the {111}, {200}, {220}, {311}, and {222} crystal planes of W2N [43]. The latter peaks are observed in the XRD pattern of the pristine W2N (blue line, Fig. 1\nb) along with peaks attributed to tungsten nitride (WN) and tungsten (W) due to traces of tungsten mononitride along with metallic tungsten obtained during the synthesis process.The structures of the as-prepared Ti3C2\nTx\n and W2N-Ti3C2\nTx\n can be further elucidated by using Bragg's diffraction equation to calculate the d-spacing and c-lattice parameters. The d-spacing of an MXene indicates the complete interlayer spacing between single MXene sheets, while the c-lattice parameter encompasses two consecutive MXene sheets along with two complete interlayer spacings [44]. In the present work, the d-spacing and c-lattice parameters of the Ti3C2\nTx\n are calculated to be 0.985\u00a0nm and 1.970\u00a0nm, respectively. They are higher than those calculated for the Ti3AlC2 MAX phase (i.e., 0.923\u00a0nm and 1.846\u00a0nm, respectively). The difference of 0.062\u00a0nm between the d-spacings of the Ti3C2\nTx\n and the Ti3AlC2 MAX phase corresponds to the average atomic radii of oxygen (O) and fluorine (F) atoms. It is due to the intercalation of the Ti3C2\nTx\n MXene with these atoms during the HF-based preparation. Meanwhile, the calculated d-spacing and c-lattice parameters of the W2N-Ti3C2\nTx\n are 0.964\u00a0nm and 1.928\u00a0nm, respectively, which are lower than those of the MXene due to the deposition of W2N onto the surface and between the Ti3C2\nTx\n layers.The structure, morphology, and compositions of the as-prepared Ti3C2\nTx\n and W2N-Ti3C2\nTx\n are revealed by the SEM images in Fig. 2\n. The Ti3AlC2 exhibits a compact 3D morphology (Fig. 2a). In contrast, the Ti3C2\nTx\n exhibits an accordion-like structure with visible interlayer spacing (Fig. 2b). This accordion-like morphology is typical of MXenes that are synthesized using HF with a concentration above 30\u00a0%. The appearance of an accordion-like structure is due to the exothermic nature of the HF reaction with Al, which releases H2 gas [40]. The Ti3AlC2 structure is composed of stacked Ti3C2 layers separated by Al atoms. When Ti3AlC2 powder is immersed in an HF solution with high concentration, an exothermic reaction occurs between the Al atoms and HF and releases gas bubbles, presumed to be H2. This reaction removes the Al atoms between the layers, resulting in the exfoliation of Ti3C2 layers due to the loss of metallic bonding holding them together with the Al atoms. Because the experiments were conducted in an aqueous environment rich in fluorine ions, oxygen, hydroxyl, and fluorine are the most probable termination groups on the surface [25,45]. The visible interlayer spacings confirm the removal of the Al layers in agreement with the XRD results mentioned above. The HAADF-STEM images of Ti3C2\nTx\n and the W2N-Ti3C2\nTx\n composite in Figs. S3 and S4 further confirm the structure and reveal the presence of \u2013O and \u2013F groups on the Ti3C2\nTx\n MXene surface.Further, the SEM image of the W2N-Ti3C2\nTx\n in Fig. 2c demonstrates that the MXene structure is maintained in the composite and reveals that the W2N is deposited in the interlayer spacings and on the surface of the MXene. As well as preventing the 2D MXene layers from restacking and promoting the electron transfer process.The SEM\u00a0+\u00a0EDX elemental mappings of the MAX phase precursor and the Ti3C2\nTx\n MXene are provided in Fig. S5. These reveal that, after the etching process, Al atoms are no longer present, while C and Ti atoms are well distributed. These results confirm the presence of\u2013O and \u2013F functionalities on the MXene surface, in agreement with the calculated c-lattice parameters and the results in Fig. S3. Finally, a comparison of these results with the SEM\u00a0+\u00a0EDX mappings of the W2N-Ti3C2\nTx\n in Fig. 2d indicates that the addition of W2N does not cause any significant rearrangement in the Ti and C distributions in the catalyst. The numerous functional groups on the surfaces of the as-synthesized Ti3C2\nTx\n and W2N-Ti3C2\nTx\n are expected to play a significant role in their hydrophilic and adhesion properties. The quantitative elemental distributions of the MAX phase, the MXene, and the W2N-NPs-MXene obtained from the EDX mappings are summarized in Table S1.The 2D structure and interlayer spacing of the as-prepared Ti3C2\nTx\n are further revealed by the low- and high-resolution TEM images in Fig. 3\n. Here, the d-spacing is seen to be 0.98\u00a0nm (Fig. 3b), which agrees with that calculated from the XRD data. Further, the AFM results in Fig. 4\na and c indicate that the average thickness of the Ti3C2\nTx\n multilayer sheet, which consists of three-four stacked Ti3C2\nTx\n sheets, is 2.9\u20133.4\u00a0nm. Meanwhile, the average thickness of the multilayer sheet in the W2N-Ti3C2\nTx\n composite is 3.9\u00a0nm (Fig. 4b and d). This is attributed to the deposition of W2N between the MXene multilayers and on the surface and the exfoliation of some of these stacked sheets.The hydrophilicity of the plain CC and the carbon cloth coated with Ti3C2\nTx\n (Ti3C2\nTx\n/CC) and W2N-Ti3C2\nTx\n (W2N-Ti3C2\nTx\n/CC) are demonstrated by the water contact angle (WCA) results in Fig. 5\n. The WCA is seen to decrease significantly from 144\u00b0 for the pristine CC to 70\u00b0 for the Ti3C2\nTx\n/CC. The hydrophobic surface of the pristine CC is mainly due to the lack of oxygen-rich groups and can hinder the attachment of microorganisms. By contrast, the surface of the Ti3C2\nTx\n/CC is hydrophilic due to the presence of \u2013O and \u2013F groups. However, the WCA of the W2N-Ti3C2\nTx\n/CC is seen to increase slightly relative to that of the Ti3C2\nTx\n/CC (i.e., from 70\u00b0 to 77\u00b0) due to the removal of some \u2013O and \u2013F groups during the synthesis of the composite [46]. Even though the addition of W2N exhibited a slight increase in WCA from 70\u00b0 to 77\u00b0, the surface characteristics are still in the hydrophilic zone, whereas decreasing WCA from 144\u00b0 in the case of CC anode to 77\u00b0 for W2N-Ti3C2\nTx\n/CC is a significant improvement in the CC anode wettability [47]. Therefore, improving the hydrophilicity of the CC by coating it with the as-fabricated W2N-Ti3C2\nTx\n composite catalyst is sufficient for promoting microorganism attachment and biofilm growth on the anode surface.The ex-situ electrochemical behavior of CC, Ti3C2\nTx\n/CC, and W2N-Ti3C2\nTx\n/CC was investigated in a three-electrode cell. Fig. S6a shows the CV profiles of the plain CC, Ti3C2\nTx\n/CC, and W2N- Ti3C2\nTx\n/CC anodes. The results reveal that the W2N-Ti3C2\nTx\n/CC anode exhibited the highest current density (6.2\u00a0A\u00a0m\u22122), followed by Ti3C2\nTx\n/CC (3.3\u00a0A\u00a0m\u22122) and finally CC (0.3\u00a0A\u00a0m\u22122). This finding indicates that the electron transfer process was improved by the deposition of Ti3C2\nTx\n, and W2N- Ti3C2\nTx\n catalysts layers on the surface of CC. The presence of the catalyst layer enhanced the bio-catalytic activity, surface wettability, and electrical conductivity. Because of these improvements, the electron transfer resistance was reduced, thereby increasing the current density. EIS analysis was used to analyze the resistance to electron mobility of CC, Ti3C2\nTx\n/CC, and W2N-Ti3C2\nTx\n/CC in the three-electrode cell using real wastewater and 1\u00a0M glucose as the electrolyte solution. The obtained results further explain why the coated CC (Ti3C2\nTx\n/CC, and W2N-Ti3C2\nTx\n/CC) achieved higher current density compared with the CC. As shown in the Nyquist plot (Fig. S6b), plain CC exhibited the highest resistance to electron mobility. This is due to the high interfacial resistance (710\u00a0\u03a9) between the wastewater and the hydrophobic surface (WCA of 144\u00b0) of CC [48,49]. On the other hand, the Ti3C2\nTx\n/CC, and W2N- Ti3C2\nTx\n/CC anodes achieved an interfacial resistance of 350 and 200\u00a0\u03a9, respectively. This is attributed to the significant improvement in the anode wettability after coating the catalyst layer. The interfacial resistance achieved from the EIS analysis exhibits the same order as the internal resistance calculated from the slope of the polarization curve of the MFC (Fig. 8c.).As detailed in the Experimental section, the OCV and anode potentials (Ean) were measured for MFCs equipped with the CC, Ti3C2\nTx\n/CC, and W2N-Ti3C2\nTx\n/CC anodes, and the results are presented in Fig. 6\n. Here, each cell exhibits a decrease in anode potential (Fig. 6a) and an increase in OCV (Fig. 6b) with time. This is attributed to the accumulation of electrons generated by microorganisms during the accumulation stage. The bacteria degrade the organic substrate (glucose) in the wastewater solution, release Adenosine triphosphate (ATP) and Nicotinamide adenine dinucleotide-hydrogen (NADH), and convert the latter to Nicotinamide adenine dinucleotide (NAD+) to complete the metabolic cycle [21,50]. The released NAD+ accumulates within the growing bacteria until a steady state is reached, and a biofilm layer is formed with the maximum quantity of released electrons on the anode surface. Notably, the highest OCV (789\u00a0mV) and lowest anode potential (\u20130.118\u00a0V) are obtained for the W2N-Ti3C2\nTx\n, followed by Ti3C2\nTx\n/CC (785\u00a0mV and \u22120.214\u00a0V), and the plain CC (710\u00a0mV and \u20130.125\u00a0V). These results demonstrate the enhanced attachment of the microorganisms and biofilm formation (Fig. S7 shows an image of the used anode with biofilm attachment.) on the surface of the modified anodes due to the enhanced hydrophilicity and affinity of these anodes towards the microorganisms. In addition, it is worth noting that the W2N-Ti3C2\nTx\n /CC takes longer to reach the stable state relative to both the Ti3C2\nTx\n/CC and the plain CC, thereby indicating the continuous growth of a larger community of electroactive microorganisms on the former electrode surface.The process of biofilm formation on the surfaces of the plain CC, the Ti3C2\nTx\n/CC, and the W2N-Ti3C2\nTx\n/CC anode is further elucidated by the bio-scanning electron microscope images obtained after testing the MFCs (Fig. 7\n). Here, the plain CC anode (Fig. 7a and b) exhibits the lowest colonization of microorganisms compared to those of the Ti3C2\nTx\n/CC (Fig. 7c and d) and W2N-Ti3C2\nTx\n /CC (Fig. 7e) anodes, while the two modified anodes exhibit high colonization of many types of microorganism. This is due, in turn, to the hydrophobic nature of the plain CC, which is unfavorable for bacterial attachment, and the enhanced hydrophilicity and biocompatibility of the modified anodes. Moreover, the synergy between the W2N and Ti3C2\nTx\n in the composite catalyst significantly enhances the conductivity, increases the available surface area for microorganism attachment, and enhances the growth of electroactive microorganisms. This, in turn, promotes the electron transfer rate between the bacteria and the anode surface. Consequently, the W2N-Ti3C2\nTx\n/CC exhibits an extensive biofilm with structures that resemble microbial nanowires or pili (Fig. 7e), which are known to play a role in direct electron transfer between the electroactive bacteria and the electrode surface. Notably, the W2N-Ti3C2\nTx\n exhibits the maximum microorganism attachment among the three anodes, along with a diverse bacterial community, which can significantly influence the time required to achieve a high and stable OCV during the electron accumulation stage of MFC operation [51].The current densities, power densities, polarization curves, internal resistances, COD removal, and CE values of the MFCs containing the plain CC, the Ti3C2\nTx\n/CC, and the W2N-Ti3C2\nTx\n/CC anodes are presented in Fig. 8\n. The W2N-Ti3C2\nTx\n is seen to provide the maximum current density of 2.3\u00a0A\u00a0m\u22122 at 0.2\u00a0V, followed by the Ti3C2\nTx\n/CC, with a current density of 1.2\u00a0A\u00a0m\u22122, and the plain CC, with a current density of 0.5\u00a0A\u00a0m\u22122 (Fig. 8a). This corresponds to improvements of 360\u00a0% and 91\u00a0%, respectively, for the modified anodes relative to the plain CC anode and is attributed to the high electrical conductivity of Ti3C2\nTx\n (15000\u00a0S\u00a0cm\u22121), which helps to promote the extracellular transfer of electrons from the microorganisms to the electrode surface, thereby significantly reducing the electron transfer resistance of the MFC. Notably, this result correlates with the observation mentioned above on the SEM images of the anode obtained after MFC operation (Fig. 7). Meanwhile, the power densities of the MFCs can be evaluated from the plots of cell voltage vs current density in Fig. 8b. The results indicate that the MFC operated with the W2N-Ti3C2\nTx\n/CC anode exhibits the highest power density of 548\u00a0mW\u00a0m\u22122, which is respectively 52\u00a0% and 84\u00a0% higher than that obtained with the Ti3C2\nTx\n/CC (263\u00a0mW\u00a0m\u22122) and with the plain CC (88\u00a0mW\u00a0m\u22122). Because the power generated by the MFC is directly correlated with the voltage and current generated during operation, the main factors that improve the current generation can be the same factors that directly promote power generation [8]. Therefore, the enhanced power density of the modified anode is mainly due to the high electrical conductivity and hydrophilicity of the Ti3C2\nTx\n, which improve the charge transfer and microorganism attachment and decrease the electron transfer resistance. Further, the slopes of the polarization curves in Fig. 8b can be used to estimate the internal resistance, which can provide a comprehensive understanding of the activation losses, ohmic losses, and bacterial metabolic losses in the MFC [52]. The as-calculated internal resistances are presented in Fig. 8c, where the MFC operating with the W2N-Ti3C2\nTx\n/CC anode exhibits a significantly reduced internal resistance of 498\u00a0\u03a9, compared to 984\u00a0\u03a9 for the Ti3C2\nTx\n/CC, and 1687\u00a0\u03a9 for the plain CC anode. Further, the enhancements in power generation and current density obtained in the present study are compared with those reported previously in Table S2, where the as-fabricated catalyst is highly competitive.Finally, the wastewater treatment capability of each MFC is indicated by the COD of its influent and effluent in Fig. 8d. Here, the MFC based on the W2N-Ti3C2\nTx\n/CC anode exhibits the highest COD removal of 68\u00a0%, followed by the Ti3C2\nTx\n/CC anode (60\u00a0%), and the plain CC anode (40\u00a0%). Moreover, Fig. 8d also presents the CE of the tested MFCs, which were calculated according to the relationship between the removal of organic substrate and the generated current of the MFCs. Thus, the W2N-Ti3C2\nTx\n/CC anode exhibits the highest CE of 81\u00a0%, followed by the Ti3C2\nTx\n/CC anode (46\u00a0%) and the plain CC anode (31\u00a0%). The enhanced COD removal capacity and CE of the W2N-Ti3C2\nTx\n/CC anode are attributed to the more rapid growth of electroactive microorganisms relative to non-electroactive microorganisms on the surface of this anode, which leads to an increase in the number of electrons generated, an enhancement in the electron transfer process, and an increase in the induced current density. Moreover, the synergy between the W2N and Ti3C2\nTx\n significantly improves the bio-electrochemical kinetics at the anode by reducing the internal resistance and increasing the generated power and current densities [53].The large-scale utilization of MFCs significantly depends on the anode efficiency to interact with the microorganism for improved biofilm growth and stability to achieve optimum electron transfer for power generation. Challenges such as limited electron transfer, poor microbial colonization, stability, and high internal resistance are directly affected by the property of the material used as an anode in MFC. Therefore, emphasis must be placed on the design of cost-effective hybrid catalysts with the ideal chemical properties, surface area, wettability, biocompatibility, and electrocatalytic activity to drive better electroactive microorganism growth and biofilm stability for achieving high current and power density in MFC. Moreover, it is vital to emphasize that any commercial treatment approach must assure stability and reusability [54].Our work introduces a cost-effective synthesis approach for a W2N-Ti3C2\nTx\n composite catalyst, which enhances the power generation of an MFC operating with domestic wastewater if used as an anode. We must indicate that we used domestic wastewater without any additional external microorganism inoculation or pretreatment.The W2N-Ti3C2\nTx\n composite catalyst was synthesized, characterized, and then used as an anode to enhance the power generation of an MFC operating with domestic wastewater. The introduction of the composite catalyst on the surface of CC benefited the attachment of microorganisms. It promoted the significant growth of biofilm with structures resembling microbial nanowires (pili) that can enable direct electron transfer. Further, the high hydrophilicity of the Ti3C2\nTx\n MXene and the excellent affinity of the W2N towards the electroactive microorganisms combined to promote an enhanced interaction between the biofilm and the anode surface, compared to the Ti3C2\nTx\n MXene alone.These combined features enabled extracellular electron transfer from the electroactive microorganisms to the electrode surface. As a result, W2N-Ti3C2\nTx\n/CC anode generated 4.6 and 6.2 times higher current and power densities, respectively, than the plain CC anode. Furthermore, the synergy between the W2N and the Ti3C2\nTx\n improved the COD removal capacity and CE by 1.8 and 2.6 times, respectively, relative to the plain CC.In brief, the present study has introduced a simple, practical approach to synthesizing highly active, electrically conductive, hydrophilic, and cost-effective composite anode catalysts for use in high performance MFCs.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.We greatly acknowledge the funding provided by the King Abdullah University of Science and Technology (KAUST), BAS/1/1403. We also acknowledge the KAUST Electron Microscope and Surface Science Core Labs for helping with the SEM-EDX, TEM, and AFM analysis.Supplementary data to this article can be found online at https://doi.org/10.1016/j.cej.2023.141821.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  Microbial fuel cells (MFCs) have enormous potential to treat wastewater and reduce the energy demands of wastewater treatment plants while generating electricity using active microorganisms as biocatalysts. However, the practical application of MFCs is limited by the low power density produced, mainly due to poor anode performance. A tungsten nitride (W2N)-MXene composite catalyst is introduced to modify the anode surface for use in microbial fuel cells during domestic wastewater treatment. The aim is to improve the wettability, electrical conductivity, electron transfer efficiency, and microorganism attachment capability of the anode and ultimately increase the overall performance of the microbial fuel cell to produce electricity during wastewater treatment. In detail, a hydrofluoric acid etching approach is used to synthesize the Ti3C2\n                     Tx\n                      MXene, the urea glass technique is used to prepare the W2N particles, and an adequate mixing and heat treatment approach is used to produce the W2N-Ti3C2\n                     Tx\n                      composite catalyst. The W2N-Ti3C2\n                     Tx\n                      composite on carbon cloth anode provides one of the best performances recorded for MXene in this type of fuel cells and using real domestic wastewater: with a 523\u00a0% increase in the power density (548\u00a0mW\u00a0m\u22122), an 83\u00a0% decrease in the chemical oxygen demand (COD), and a 161\u00a0% increase in the electron transfer efficiency compared to those of the plain carbon cloth. We demonstrate that this outstanding performance is due to the improvements in hydrophilicity and microorganism attachment, particularly nanowires (or pili) which promote electron transfer. The present work offers an exciting avenue toward the process scale-up and optimization of single-chamber microbial fuel cells.\n               "}
{"full_text": "Nitrogen-containing heterocycles are ubiquitous scaffolds of many natural products and are being widely used in the field of pharmaceuticals [1\u20134]. Among them, quinazolinone derivatives are of utmost interest for a wide cross-section of chemists due to their remarkable bioactivities (Fig. 1\n) [5\u20138]. They exhibit a range of biological activities including cardiovascular, adipogenesis inhibitor, kinesin spindle protein inhibitor, antiinfective, anticancer, anticonvulsant, etc [9\u201314]. In addition, the use of quinazolinone derivatives in agriculture is still rare [15\u201318]. Therefore, there is still a hotspot in developing new means for synthesizing quinazolinone derivatives and application [19\u201324].Over the years, CH bond functionalization is considered as powerful and shortest approach for construction of carbon\u2013carbon and carbon-heteroatom bonds, which have potential application in the synthesis of biologically relevant molecules [25\u201327]. The azaarenes are among an important class of organic molecules whose synthesis have been described based on the concept of CH functionalization [28\u201333]. 2-Alkyl azaarenes such as 2-methyl pyridine, 2-methyl quinolone and 2-methyl quinazolinone have been successfully functionalized via addition to a variety of carbonyl compounds, such as aldehydes, isatins, \u03b1-oxoesters, etc [34\u201340]. Because ketones have an additional electron-donating group, which not only reduces their partial positive charge on the carbonyl carbon but also contributes to steric hindrance, they usually have lower reactivity than aldehydes. These factors make it challenging for ketones to go through nucleophilic addition compared to aldehydes. As a strategy to overcome this problem, an electron-withdrawing CF3 moiety adjacent to the reactive carbonyl carbon site is effective in activating the ketonic carbon. Shaikh reported a Yb(OTf)3-catalyzed benzylic CH bond functionalization of alkyl azaarene with a-trifluoromethylated carbonyl compounds [41]. In 2015, Teo described the InCl3-catalyzed functionalization of the CH bond of \u03b1-alkylazaarene with trifluoromethyl ketones [42]. Although the reactions proceed successfully, expensive rare metal catalysts and high reaction temperature were necessary. Based on this precedent, Teo observed that the reaction can also be catalyzed by FeCl3 catalyst (Fig. 2\n) [43]. In 2022, we developed a novel reaction for the synthesis of azaarene-equipped CF3-tertiary alcohols through addition of azaarenes to CF3-ketones under metal-free conditions [44].The literature survey indicated that most reported reactions are limited to using metal catalysts. Therefore, considering the potential of the greener approach in organic synthesis, the development of new strategies via improving the atom economy of chemical processes and simple and efficient coupling under solvent- and catalyst-free approach for the construction of quinazolinone derivatives are a challenging task and are in great demand.Coumarin is an important heterocyclic skeleton frequently found in numerous natural products, pharmaceutical molecules, fluorescent probes, and materials [45\u201350]. The combination of some privileged structures, a benzopyrone ring, a trifluoromethyl moiety and a quinazolinone ring, for the synthesis of quaternary carbon organic molecules could be of significant importance, especially for new drugs and materials.In continuation of our study on the application of 3-(trifluoroacetyl)coumarins [51\u201356], our aim was to develop a direct, metal free, environmentally benign route to access the quinazolinone derivatives along with avoidance of the aforementioned drawbacks. Herein, we report an efficient, eco-friendly direct approach to access the quinazolinone derivatives from 2-methyl quinazolinones with 3-(trifluoroacetyl)coumarins under catalyst- and solvent-free conditions (Scheme 1\n).All chemicals, except 2-(trifluoromethyl)-2-hydroxy-2H-chromene 1 and 2-methyl quinazolinone 2, which were synthesized according reported procedure, were purchased from commercial sources and used without further purification. 1H NMR and 13C NMR spectra were obtained using a Bruker DPX-400 spectrometer in CDCl3 or DMSO\u2011d\n6 solution with TMS as an internal standard. HR-MS(APCI) spectra were performed using a Waters Q-Tof MicroTM instrument, and X-rays were measured at 293\u00a0K on a Rigaku RAXIAS-IV type diffractometer.To a mixture of 2-(trifluoromethyl)-2-hydroxy-2H-chromene (0.4\u00a0mmol) and aryl ketone (0.4\u00a0mmol) in acetic acid (4\u00a0mL) was added NH4OAc (0.8\u00a0mmol) and the resulting mixture was heated under reflux. After completion of the reaction, the mixture was concentrated under vacuum to yield the crude product, which was further purified by column chromatography.White solid, mp: 213.2\u2013214.7 \u2103. 1H NMR (400\u00a0MHz, CDCl3) \u03b4 10.13 (s, 1H), 8.41 (s,1H), 8.23 (d, J\u00a0=\u00a07.9\u00a0Hz, 1H), 7.71 (m, 2H), 7.57 (m, J\u00a0=\u00a07.5, 4.2\u00a0Hz, 3H), 7.45 (t, J\u00a0=\u00a07.5\u00a0Hz, 1H), 7.32 (m, J\u00a0=\u00a012.4, 5.4\u00a0Hz, 2H), 4.03 (d, J\u00a0=\u00a015.3\u00a0Hz, 1H), 3.58 (d, J\u00a0=\u00a015.3\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, CDCl3) \u03b4 161.3, 160.7, 153.7, 152.1, 147.4, 146.51, 134.8, 133.2, 128.8, 127.3, 126.7, 125.1, 124.3, 122.8, 121.4, 118.3, 116.6, 76.2, 75.9, 36.9.19F NMR (376\u00a0MHz, CDCl3) \u03b4 \u221280.95. HRMS (APCI): m/z calcd for C20H14F3N2O4 [M\u00a0+\u00a0H]+ 403.0906, found 403.0923.White solid, mp: 257.2\u2013259.8\u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.08 (s, 1H), 8.43 (s, 1H), 7.86 (d, J\u00a0=\u00a07.6\u00a0Hz, 1H), 7.78 (d, J\u00a0=\u00a08.0\u00a0Hz, 1H), 7.66 (t, J\u00a0=\u00a07.8\u00a0Hz, 1H), 7.40 (m, 3H), 7.20 (d, J\u00a0=\u00a08.1\u00a0Hz, 1H), 4.34 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H), 3.37 (s, 1H), 2.51 (s, 3H), 2.20 (s, 4H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 162.2, 159.0, 153.8, 151.3, 146.8, 143.2, 132.9, 129.4, 128.4, 126.9, 125.6, 125.2, 124.0, 123.0, 119.2, 116.1, 75.4, 75.2, 74.9, 74.6, 36.4, 20.8, 12.6. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.52. HRMS (APCI): m/z calcd for C22H18F3N2O4 [M\u00a0+\u00a0H]+ 431.1219, found 431.1190.White solid, mp: 219.6\u2013221.0 \u2103. 1H NMR (400\u00a0MHz, CDCl3) \u03b4 10.35 (s, 1H), 8.47 (s, 1H), 8.10 (d, J\u00a0=\u00a07.8\u00a0Hz, 1H), 7.82 (s, 1H), 7.57 (m, 3H), 7.32 (m, J\u00a0=\u00a010.4, 4.9\u00a0Hz, 3H), 4.16 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.61 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 2.50 (s, 3H). 13C NMR (101\u00a0MHz, CDCl3) \u03b4 161.9, 160.3, 153.8, 151.3, 146.4, 145.9, 135.5, 134.6, 133.1, 128.7, 126.8, 125.0, 124.5, 122.4, 121.2, 118.4, 116.6, 76.1, 75.8, 36.4, 17.7. 19F NMR (376\u00a0MHz, CDCl3) \u03b4 \u221280.62. HRMS (APCI): m/z calcd for C21H16F3N2O4 [M\u00a0+\u00a0H]+ 417.1062, found 417.1083.White solid, mp: 207.8\u2013209.9 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.13 (s, 1H), 8.41 (s, 1H), 7.84 (m, 2H), 7.66 (t, J\u00a0=\u00a07.4\u00a0Hz, 1H), 7.46 (d, J\u00a0=\u00a08.3\u00a0Hz, 3H), 7.38 (d, J\u00a0=\u00a07.5\u00a0Hz, 1H), 7.01 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 4.21 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.39 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 2.36 (s, 3H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.7, 158.9, 153.7, 152.2, 146.1, 143.6, 136.5, 136.1, 133.0, 129.5, 126.7, 125.5, 125.2, 123.9, 121.1, 118.8, 116.2, 75.8, 75.5, 75.2, 74.9, 36.1, 21.1. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.12. HRMS (APCI): m/z calcd for C21H16F3N2O4 [M\u00a0+\u00a0H]+ 417.1062, found 417.1029.Yellow solid, mp: 244.3\u2013246.6 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.73 (s, 1H), 8.73 (d, J\u00a0=\u00a02.6\u00a0Hz, 1H), 8.40 (m, 3H), 7.86 (d, J\u00a0=\u00a07.6\u00a0Hz, 2H), 7.67 (d, J\u00a0=\u00a07.4\u00a0Hz, 2H), 7.48 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 7.39 (s, 2H), 7.25 (d, J\u00a0=\u00a09.0\u00a0Hz, 2H), 4.36 (d, J\u00a0=\u00a015.3\u00a0Hz, 0H), 3.45 (s, 0H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.1, 159.0, 156.9, 153.7, 152.4, 145.1, 143.7, 133.1, 129.5, 129.0, 128.6, 126.7, 125.2, 125.0, 123.8, 122.3, 121.4, 118.8, 116.3, 75.7, 75.4, 75.2, 74.9, 36.8. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.31. HRMS (APCI): m/z calcd for C20H13F3N3O6 [M\u00a0+\u00a0H]+ 448.0756, found 448.0732.White solid, mp: 210.4\u2013212.7 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.57 (s, 1H), 8.41 (s, 1H), 8.24 (s, 1H), 7.90 (d, J\u00a0=\u00a08.5\u00a0Hz, 1H), 7.83 (d, J\u00a0=\u00a07.8\u00a0Hz, 1H), 7.63 (t, J\u00a0=\u00a07.8\u00a0Hz, 1H), 7.38 (m, 3H), 7.25 (d, J\u00a0=\u00a08.5\u00a0Hz, 1H), 4.33 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.43 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.1, 159.0, 155.7, 153.7, 150.7, 143.6, 133.0, 130.9, 129.5, 128.4, 126.9, 126.7, 125.5, 125.1, 123.6, 122.8, 121.5, 118.8, 116.2, 75.7, 75.5, 75.2, 74.9, 36.6. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221261.06, \u221278.31. HRMS (APCI): m/z calcd for C21H13F6N2O4[M\u00a0+\u00a0H]+ 471.0780, found 471.0797.Yellow solid, mp: 204.3\u2013205.9 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.49 (s, 1H), 8.31 (s, 1H), 7.79 (d, J\u00a0=\u00a06.4\u00a0Hz, 1H), 7.59 (d, J\u00a0=\u00a07.2\u00a0Hz, 1H), 7.53 (dt, J\u00a0=\u00a08.3, 2.9\u00a0Hz, 2H), 7.40 (d, J\u00a0=\u00a08.2\u00a0Hz, 1H), 7.33 (d, J\u00a0=\u00a07.6\u00a0Hz, 1H), 7.27 (s, 1H), 4.31 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.34 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.4, 159.2, 158.1, 158.1, 158.0, 158.0, 155.6, 155.4, 153.9, 152.9, 143.0, 135.0, 134.9, 132.7, 129.3, 126.7, 125.5, 125.0, 123.9, 123.8, 123.7, 119.0, 116.1, 110.4, 110.1, 110.1, 109.9, 106.9, 106.9, 106.7, 106.7, 75.7, 75.4, 75.1, 74.8, 36.3. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.34, \u2212110.70, \u2212121.00. HRMS (APCI): m/z calcd for C20H12F5N2O4 [M\u00a0+\u00a0H]+ 439.0717, found 439.0701.White solid, mp: 199.2\u2013200.2 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.25 (s, 1H), 8.40 (s, 1H), 7.85 (d, J\u00a0=\u00a06.8\u00a0Hz, 1H), 7.65 (d, J\u00a0=\u00a07.3\u00a0Hz, 1H), 7.60 (m, J\u00a0=\u00a08.1, 2.5\u00a0Hz, 1H), 7.47 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 7.38 (t, J\u00a0=\u00a07.5\u00a0Hz, 2H), 7.14 (m, J\u00a0=\u00a010.9, 8.2\u00a0Hz, 1H), 6.86 (d, J\u00a0=\u00a08.2\u00a0Hz, 1H), 4.24 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.37 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 162.2, 159.6, 159.0, 154.1, 153.7, 150.4, 143.5, 135.5, 133.0, 129.5, 126.7, 125.2, 123.9, 122.9, 118.8, 116.2, 113.3, 113.1, 110.9, 75.7, 75.4, 75.1, 74.8, 36.1, 19.0. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.26, \u2212111.33. HRMS (APCI): m/z calcd for C20H13F4N2O4 [M\u00a0+\u00a0H]+ 421.0811, found 421.0801.White solid, mp: 250.0\u2013251.2 \u2103. 1H NMR (400\u00a0MHz, DMSO) \u03b4 12.66 (s, 1H), 8.39 (s, 1H), 7.93 (d, J\u00a0=\u00a02.4\u00a0Hz, 1H), 7.88 (d, J\u00a0=\u00a02.4\u00a0Hz, 1H), 7.84 (d, J\u00a0=\u00a07.7\u00a0Hz, 1H), 7.63 (d, J\u00a0=\u00a07.1\u00a0Hz, 1H), 7.43 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 7.39\u20137.30 (m, J\u00a0=\u00a018.3, 10.2\u00a0Hz, 2H), 4.43 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H), 3.38 (s, 1H). 13C NMR (101\u00a0MHz, DMSO) \u03b4 160.5, 159.2, 154.3, 154.0, 144.1, 143.2, 134.3, 132.7, 131.8, 130.6, 129.3, 126.8, 125.6, 125.0, 124.4, 123.9, 119.4, 116.0, 75.6, 75.3, 75.0, 74.8. 19F NMR (376\u00a0MHz, DMSO) \u03b4 \u221278.52. HRMS (APCI): m/z calcd for C20H12Cl2F3N2O4 [M\u00a0+\u00a0H]+ 471.0126, found 471.0106.White solid, mp: 220.7\u2013222.8 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.39 (s, 1H), 8.40 (s, 1H), 7.93 (d, J\u00a0=\u00a02.5\u00a0Hz, 1H), 7.82 (m, J\u00a0=\u00a07.8, 1.4\u00a0Hz, 1H), 7.62 (m, J\u00a0=\u00a08.0, 5.6\u00a0Hz, 2H), 7.43 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 7.33 (d, J\u00a0=\u00a07.4\u00a0Hz, 1H), 7.09 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 4.37 (s, 1H), 4.27 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.45 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.8, 158.9, 153.6, 146.8, 143.6, 134.9, 132.9, 131.1, 129.5, 129.0, 126.7, 125.1, 123.8, 122.6, 118.8, 116.2, 75.7, 75.4, 75.1, 74.8, 36.3, 18.9. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.25. HRMS (APCI): m/z calcd for C20H13ClF3N2O4 [M\u00a0+\u00a0H]+ 437.0516, found 437.0530.White solid, mp: 222.2\u2013224.2 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.23 (s, 1H), 8.39 (s, 1H), 7.84 (m, J\u00a0=\u00a07.7, 1.3\u00a0Hz, 1H), 7.65 (t, J\u00a0=\u00a07.1\u00a0Hz, 1H), 7.52 (s, 1H), 7.46 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 7.37 (m, 3H), 6.97 (d, J\u00a0=\u00a08.1\u00a0Hz, 1H), 4.22 (d, J\u00a0=\u00a015.3\u00a0Hz, 1H), 3.36 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.0, 159.0, 153.8, 150.7, 143.5, 134.6, 132.9, 129.5, 129.1, 126.7, 126.4, 125.2, 123.9, 118.8, 118.4, 116.2, 75.7, 75.4, 75.1, 74.8, 36.1. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.24. HRMS (APCI): m/z calcd for C20H13ClF3N2O4 [M\u00a0+\u00a0H]+ 437.0516, found 437.0505.Yellow solid, mp: 225.0\u2013226.0 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.39 (s, 1H), 8.40 (s, 1H), 8.08 (d, J\u00a0=\u00a02.3\u00a0Hz, 1H), 7.82 (d, J\u00a0=\u00a07.7\u00a0Hz, 1H), 7.74 (m, J\u00a0=\u00a08.7, 2.4\u00a0Hz, 1H), 7.63 (m, J\u00a0=\u00a011.4, 4.2\u00a0Hz, 1H), 7.44 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 7.35 (m, J\u00a0=\u00a08.9, 5.3\u00a0Hz, 2H), 7.02 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 4.26 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.40 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.7, 159.0, 153.7, 147.2, 143.6, 137.7, 133.0, 129.5, 129.2, 128.3, 126.7, 125.1, 123.9, 123.0, 119.2, 118.8, 116.2, 75.4, 75.3, 75.3, 74.8, 36.4. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.25. HRMS (APCI): m/z calcd for C20H13BrF3N2O4 [M\u00a0+\u00a0H]+ 481.0011, found 481.0027.Yellow solid, mp: 205.1\u2013206.4 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.21 (s, 1H), 8.39 (s, 1H), 8.12 (d, J\u00a0=\u00a02.4\u00a0Hz, 1H), 8.02 (d, J\u00a0=\u00a09.0\u00a0Hz, 1H), 7.78 (d, J\u00a0=\u00a02.4\u00a0Hz, 1H), 7.62 (s, 1H), 7.40 (m, 3H), 7.07 (d, J\u00a0=\u00a08.1\u00a0Hz, 1H), 4.23 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H), 3.41 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.8, 158.50, 152.8, 148.1, 142.2, 135.3, 134.9, 131.5, 126.8, 126.2, 123.8, 121.4, 120.8, 118.5, 116.8, 75.7, 75.4, 75.1, 74.8, 36.2. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.14. HRMS (APCI): m/z calcd for C20H13BrF3N2O4 [M\u00a0+\u00a0H]+ 481.0011, found 480.9995.White solid, mp: 218.4\u2013219.6 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.19 (s, 1H), 8.43 (s, 1H), 8.17 (d, J\u00a0=\u00a02.3\u00a0Hz, 1H), 7.88 (d, J\u00a0=\u00a07.4\u00a0Hz, 1H), 7.81 (m, J\u00a0=\u00a08.8, 2.4\u00a0Hz, 1H), 7.43 (m, J\u00a0=\u00a014.7, 5.8\u00a0Hz, 3H), 7.28 (t, J\u00a0=\u00a07.6\u00a0Hz, 1H), 4.35 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H), 3.38 (s, 1H), 1.85 (s, 3H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 162.1, 158.5, 152.8, 151.5, 147.0, 142.0, 135.2, 134.8, 131.4, 127.0, 126.8, 126.2, 123.9, 121.3, 121.0, 118.5, 116.8, 75.56, 75.3, 75.0, 74.7, 55.4, 36.4, 16.9. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.43. HRMS (APCI): m/z calcd for C21H15BrF3N2O4 [M\u00a0+\u00a0H]+ 495.0167, found 495.0154.White solid, mp: 200.9\u2013201.6 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.11 (s, 1H), 8.39 (s, 1H), 8.13 (d, J\u00a0=\u00a02.3\u00a0Hz, 1H), 7.79 (m, 2H), 7.50 (s, 1H), 7.44 (t, J\u00a0=\u00a06.5\u00a0Hz, 2H), 6.99 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 4.20 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.38 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H), 2.35 (s, 3H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.7, 158.5, 152.7, 152.0, 146.1, 142.3, 136.5, 136.1, 135.3, 131.5, 129.5, 126.7, 125.5, 123.8, 121.1, 120.7, 118.5, 116.7, 75.7, 75.4, 75.1, 74.8, 36.1, 21.1. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.07. HRMS (APCI): m/z calcd for C21H15BrF3N2O4 [M\u00a0+\u00a0H]+ 495.0167, found 495.0143.White solid, mp: 187.2\u2013189.4 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.41 (s, 1H), 8.40 (s, 1H), 8.13 (m, J\u00a0=\u00a016.5, 2.2\u00a0Hz, 2H), 7.81 (m, J\u00a0=\u00a08.3, 5.9, 2.3\u00a0Hz, 2H), 7.44 (t, J\u00a0=\u00a09.6\u00a0Hz, 2H), 7.02 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 4.26 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H), 3.40 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.8, 158.5, 153.7, 152.7, 147.1, 142.4, 137.7, 135.3, 131.5, 129.2, 128.3, 126.6, 123.8 (s, 2H), 123.0, 120.7, 119.1, 118.5, 116.8, 75.7, 75.4, 75.1, 74.8, 56.5, 36.4. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.19. HRMS (APCI): m/z calcd for C20H12Br2F3N2O4 [M\u00a0+\u00a0H]+ 558.9116, found 558.9077.Yellow solid, mp: 222.4\u2013224.9 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.22 (s, 1H), 8.41 (s, 1H), 8.19 (s, 2H), 8.04 (d, J\u00a0=\u00a06.9\u00a0Hz, 1H), 7.66 (t, J\u00a0=\u00a07.7\u00a0Hz, 1H), 7.58 (s, 1H), 7.42 (t, J\u00a0=\u00a07.5\u00a0Hz, 1H), 7.07 (d, J\u00a0=\u00a08.1\u00a0Hz, 1H), 4.21 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H), 3.41 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.7, 157.9, 152.8, 149.7, 148.1, 142.1, 137.3, 134.9, 131.3, 127.51, 126.8, 126.2, 123.7, 121.7, 121.4, 116.9, 110.2, 75.7, 75.4, 75.1, 74.8, 66.8, 36.1. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.11 (s, 1H). HRMS (APCI): m/z calcd for C20H12Br2F3N2O4 [M\u00a0+\u00a0H]+ 558.9116, found 558.9084.White solid, mp: 221.4\u2013222.8 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.09 (s, 1H), 8.45 (s, 1H), 8.20 (d, J\u00a0=\u00a09.4\u00a0Hz, 2H), 7.79 (d, J\u00a0=\u00a08.0\u00a0Hz, 1H), 7.49 (s, 1H), 7.21 (d, J\u00a0=\u00a08.0\u00a0Hz, 1H), 4.33 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H), 3.39 (s, 1H), 2.23 (s, 3H), 1.81 (s, 3H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 162.1, 157.8, 151.1, 149.7, 146.7, 143.4, 142.0, 137.2, 132.7, 131.1, 128.4, 127.8, 126.7, 123.9, 123.0, 121.9, 119.3, 116.9, 110.2, 75.4, 75.2, 74.9, 74.6, 36.4, 20.8, 12.6. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.45. HRMS (APCI): m/z calcd for C22H16Br2F3N2O4 [M\u00a0+\u00a0H]+ 586.9429, found 586.9394.White solid, mp: 161.9\u2013163.9 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.18 (s, 1H), 8.42 (s, 1H), 8.20 (m, J\u00a0=\u00a06.6, 2.2\u00a0Hz, 2H), 7.88 (d, J\u00a0=\u00a07.2\u00a0Hz, 1H), 7.48 (d, J\u00a0=\u00a06.6\u00a0Hz, 2H), 7.28 (t, J\u00a0=\u00a07.6\u00a0Hz, 1H), 4.33 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H), 3.36 (d, J\u00a0=\u00a015.8\u00a0Hz, 1H), 1.84 (s, 3H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 162.1, 157.8, 151.4, 149.8, 146.9, 141.9, 137.2, 135.1, 134.7, 131.1, 127.8, 126.7, 126.3, 123.8, 121.9, 121.4, 116.9, 110.2, 75.5, 75.3, 75.0, 74.7, 55.4, 36.3, 16.8. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.39. HRMS (APCI): m/z calcd for C21H14Br2F3N2O4 [M\u00a0+\u00a0H]+ 572.9272, found 572.9242.Yellow solid, mp: 215.5\u2013217.1 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.12 (s, 1H), 8.40 (s, 1H), 8.18 (s, 2H), 7.83 (s, 1H), 7.59 (s, 1H), 7.48 (m, J\u00a0=\u00a08.3, 1.9\u00a0Hz, 1H), 6.99 (d, J\u00a0=\u00a08.3\u00a0Hz, 1H), 4.17 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H), 3.40 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H), 2.37 (s, 3H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.7, 157.8, 151.9, 149.7, 146.0, 142.1, 137.3, 136.6, 136.2, 131.3, 127.5, 126.7, 125.52, 123.7, 121.7, 121.2, 116.9, 110.2, 75.7, 75.4, 75.1, 74.8, 66.8, 36.0, 21.12. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.04. HRMS (APCI): m/z calcd for C21H14Br2F3N2O4 [M\u00a0+\u00a0H]+ 572.9272, found 572.9244.Yellow solid, mp: 217.3\u2013218.9 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.76 (s, 1H), 8.74 (d, J\u00a0=\u00a02.6\u00a0Hz, 1H), 8.42 (t, J\u00a0=\u00a04.5\u00a0Hz, 2H), 8.19 (s, 2H), 7.51 (s, 1H), 7.26 (d, J\u00a0=\u00a09.0\u00a0Hz, 1H), 4.33 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H), 3.45 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.1, 157.9, 156.6, 152.3, 149.7, 145.2, 142.4, 137.4, 131.3, 128.9, 127.2, 126.5, 123.7, 122.3, 121.6, 117.0 (s, 11H), 110.3, 75.3, 75.0, 74.7, 36.6. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.24. HRMS (APCI): m/z calcd for C20H11Br2F3N3O6 [M\u00a0+\u00a0H]+ 603.8967, found 603.8925.White solid, mp: 204.5\u2013206.5 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.55 (s, 1H), 8.34 (s, 1H), 8.18 (s, 2H), 7.62 (m, 2H), 7.43 (s, 1H), 4.33 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.37 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.4, 158.0, 155.5, 152.6, 149.8, 141.4, 137.1, 134.9, 131.1, 128.0, 126.6, 123.7, 121.8, 116.7, 110.5, 110.2, 106.9, 75.5, 75.1, 74.8, 36.3. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.23, \u2212110.56, \u2212121.79. HRMS (APCI): m/z calcd for C20H10Br2F5N2O4 [M\u00a0+\u00a0H]+ 594.8927, found 594.8881.White solid, mp: 232.8\u2013234.4 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.66 (s, 1H), 8.36 (s, 1H), 8.18 (s, 2H), 7.94\u20137.92 (m, J\u00a0=\u00a06.4, 2.4\u00a0Hz, 2H), 7.42 (s, 1H), 4.39 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.37 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.5, 158.0, 154.0 (s, 4H), 149.9, 144.0, 141.8, 137.0, 134.4, 131.7, 131.1, 130.7, 127.8, 126.5, 124.5, 123.8, 122.2, 121.3, 116.8, 110.2, 75.4, 75.1, 36.5. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.35. HRMS (APCI): m/z calcd for C20H10Br2Cl2F3N2O4 [M\u00a0+\u00a0H]+ 626.8336, found 626.8307.White solid, mp: 107.3\u2013108.6 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.43 (s, 1H), 8.42 (s, 1H), 8.18 (m, J\u00a0=\u00a07.4, 2.2\u00a0Hz, 2H), 8.12 (d, J\u00a0=\u00a02.3\u00a0Hz, 1H), 7.81 (m, J\u00a0=\u00a08.7, 2.3\u00a0Hz, 1H), 7.51 (s, 1H), 7.03 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 4.24 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H), 3.41 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.7, 157.8, 153.5, 149.7, 147.7, 147.1, 144.7, 142.4, 141.1, 139.1, 137.5, 136.2, 135.1, 134.1, 133.0, 132.2, 131.3, 130.4, 129.4, 128.3, 128.1, 75.0, 74.7, 36.23. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.18. HRMS (APCI): m/z calcd for C20H11Br3F3N2O4 [M\u00a0+\u00a0H]+ 636.8221, found 636.8177.Yellow solid, mp: 212.5\u2013214.0 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 11.79 (s, 1H), 9.06 (s, 1H), 8.57 (d, J\u00a0=\u00a07.4\u00a0Hz, 1H), 8.23 (s, 2H), 8.18 (m, J\u00a0=\u00a011.1, 4.3\u00a0Hz, 1H), 7.98 (d, J\u00a0=\u00a08.1\u00a0Hz, 1H), 7.91 (t, J\u00a0=\u00a07.6\u00a0Hz, 1H), 7.81 (m, J\u00a0=\u00a06.1, 3.2\u00a0Hz, 1H), 7.74 (m, J\u00a0=\u00a07.0, 3.7\u00a0Hz, 2H), 4.75 (d, J\u00a0=\u00a015.9\u00a0Hz, 1H), 4.38 (s, 3H), 4.13 (d, J\u00a0=\u00a015.9\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.8, 158.7, 153.1, 148.1, 146.6, 143.9, 143.0, 134.8, 129.6, 126.8, 126.1, 125.3, 125.1, 123.9, 121.4, 121.0, 120.5, 119.4, 115.0, 75.7, 75.4, 75.1, 74.8, 36.1, 19.0. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.28 (s, 1H). HRMS (APCI): m/z calcd for C21H16F3N2O5 [M\u00a0+\u00a0H]+ 433.1011, found 433.1028.White solid, mp: 223.6\u2013225.7 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.21 (s, 1H), 8.42 (s, 1H), 7.89 (d, J\u00a0=\u00a07.9\u00a0Hz, 1H), 7.46 (d, J\u00a0=\u00a07.2\u00a0Hz, 1H), 7.41 (d, J\u00a0=\u00a07.1\u00a0Hz, 1H), 7.31 (m, 4H), 4.38 (d, J\u00a0=\u00a015.6\u00a0Hz, 1H), 3.93 (s, 3H), 3.39 (d, J\u00a0=\u00a09.8\u00a0Hz, 1H), 1.87 (s, 3H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 162.1, 158.7, 151.7, 146.9, 146.6, 143.5, 143.1, 134.9, 126.9, 126.2, 125.8, 125.1, 123.9, 123.7, 121.3, 120.4, 119.7, 115.0, 75.6, 75.3, 75.0, 74.7, 36.4, 16.9. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.50. HRMS (APCI): m/z calcd for C22H18F3N2O5 [M\u00a0+\u00a0H]+ 447.1168, found 447.1182.White solid, mp: 204.2\u2013206.3 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.59 (s, 1H), 8.41 (s, 1H), 8.29 (s, 1H), 7.97 (m, J\u00a0=\u00a08.6, 1.7\u00a0Hz, 1H), 7.35 (m, J\u00a0=\u00a016.7, 12.3, 3.6\u00a0Hz, 6H), 4.33 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.94 (s, 3H), 3.44 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 161.2, 158.7, 155.7, 150.7, 146.6, 144.0, 143.1, 131.0, 128.4, 127.0, 126.7, 125.5, 125.2, 123.8, 123.6, 122.8, 121.5, 120.6, 119.4, 115.1, 75.5, 75.3, 75.1, 74.8, 56.6, 36.6. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221260.99, \u221278.38. HRMS (APCI): m/z calcd for C22H15F6N2O5 [M\u00a0+\u00a0H]+ 501.0885, found 501.0868.White solid, mp: 89.7\u201391.5 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.53 (s, 1H), 8.33 (s, 1H), 7.65\u20137.55 (m, J\u00a0=\u00a022.7, 9.0\u00a0Hz, 2H), 7.33 (dt, J\u00a0=\u00a012.1, 7.6\u00a0Hz, 4H), 4.34 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.93 (s, 3H), 3.36 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.5, 159.2, 154.3, 154.0, 144.1, 143.2, 134.3, 132.7, 131.8, 130.6, 129.3, 126.8, 125.6, 125.0, 124.4, 123.8, 119.4, 116.0, 75.6, 75.3, 75.0, 74.8, 66.8, 36.4. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.39, \u2212110.70, \u2212120.93. HRMS (APCI): m/z calcd for C21H14F5N2O5 [M\u00a0+\u00a0H]+ 469.0823, found 469.0790.White solid, mp: 193.2\u2013195.7 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.24 (s, 1H), 8.38 (s, 1H), 7.59 (m, 1H), 7.33 (m, 4H), 7.13 (t, J\u00a0=\u00a09.5\u00a0Hz, 1H), 6.87 (d, J\u00a0=\u00a08.1\u00a0Hz, 1H), 4.23 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.92 (s, 3H), 3.41 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 162.16, 159.6, 159.2, 158.7, 154.2, 150.3, 146.6, 143.8, 143.0, 135.5, 126.7, 125.2, 123.9, 122.9, 120.6, 119.4, 115.1, 113.3, 113.1, 110.9, 36.1. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.33, \u2212111.33. HRMS (APCI): m/z calcd for C21H15F4N2O5 [M\u00a0+\u00a0H]+ 451.0917, found 451.0899.White solid, mp: 186.5\u2013188.6 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.67 (s, 1H), 8.36 (s, 1H), 7.94\u20137.88 (m, J\u00a0=\u00a017.6, 2.4\u00a0Hz, 2H), 7.39 \u2013 7.36 (m, J\u00a0=\u00a07.2, 1.9\u00a0Hz, 1H), 7.34 \u2013 7.28 (m, 3H), 4.43 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H), 3.92 (s, 3H), 3.37 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.5, 159.0, 154.3, 146.6, 144.0, 143.4, 134.3, 131.8, 130.6, 126.7, 125.7, 125.0, 124.4, 123.8, 120.4, 119.9, 114.8, 75.6, 75.3, 75.0, 74.7, 56.6, 36.4. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.56. HRMS (APCI): m/z calcd for C21H14Cl2F3N2O5 [M\u00a0+\u00a0H]+ 501.0232, found 501.0211.White solid, mp: 148.3\u2013150.3 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.45 (s, 1H), 8.43 (s, 1H), 7.98 (d, J\u00a0=\u00a02.5\u00a0Hz, 1H), 7.68 (m, J\u00a0=\u00a08.7, 2.5\u00a0Hz, 1H), 7.41 (m, J\u00a0=\u00a06.9, 2.1\u00a0Hz, 2H), 7.33 (t, J\u00a0=\u00a04.9\u00a0Hz, 2H), 7.16 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 4.31 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.95 (s, 3H), 3.44 (d, J\u00a0=\u00a015.7\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.8, 158.7, 153.7, 146.9, 146.6, 144.0, 143.0, 135.0, 131.1, 129.1, 126.7, 125.1, 123.9, 122.6, 120.5, 119.4, 115.0, 75.7, 75.4, 75.1, 74.8, 56.5, 36.3. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.34 (s, 1H). HRMS (APCI): m/z calcd for C21H15ClF3N2O5 [M\u00a0+\u00a0H]+ 467.0622, found 467.0598.Yellow solid, 232.7\u2013234.9 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.24 (s, 1H), 8.37 (s, 1H), 7.53 (t, J\u00a0=\u00a08.0\u00a0Hz, 1H), 7.38 (m, J\u00a0=\u00a010.1, 4.7\u00a0Hz, 2H), 7.32 (m, J\u00a0=\u00a013.1, 6.5\u00a0Hz, 3H), 6.99 (d, J\u00a0=\u00a08.1\u00a0Hz, 1H), 4.21 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.92 (s, 3H), 3.37 (s, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 159.9, 158.7, 153.9, 150.7, 146.6, 143.8, 143.1, 134.7, 132.8, 129.2, 126.7, 126.4, 125.2, 123.9, 120.6, 119.4, 118.4, 115.1, 75.6, 75.4, 75.1, 74.8, 56.6, 36.0. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.32. HRMS (APCI): m/z calcd for C21H15ClF3N2O5 [M\u00a0+\u00a0H]+ 467.0622, found 467.0605.White solid, mp: 173.6\u2013174.7 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.42 (s, 1H), 8.40 (s, 1H), 8.12 (d, J\u00a0=\u00a02.3\u00a0Hz, 1H), 7.80 (m, J\u00a0=\u00a08.7, 2.4\u00a0Hz, 1H), 7.36 (m, 4H), 7.06 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 4.27 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.94 (s, 3H), 3.40 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.7, 158.7, 153.8, 147.2, 146.6, 143.9, 143.0, 137.8, 129.3, 128.3, 126.7, 125.2, 123.9, 123.0, 120.6, 119.4, 119.2, 115.1, 75.5, 75.1, 74.8, 56.6, 55.4, 36.3. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.33. HRMS (APCI): m/z calcd for C21H15BrF3N2O5 [M\u00a0+\u00a0H]+ 511.0166, found 511.0092.Yellow solid, mp: 170.7\u2013171.5 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 11.74 (s, 1H), 9.01 (s, 1H), 8.57 (m, J\u00a0=\u00a07.9, 1.2\u00a0Hz, 1H), 8.18 (d, J\u00a0=\u00a08.7\u00a0Hz, 3H), 7.99 (d, J\u00a0=\u00a08.0\u00a0Hz, 1H), 7.91 (t, J\u00a0=\u00a07.1\u00a0Hz, 1H), 7.40 (m, J\u00a0=\u00a08.7, 2.4\u00a0Hz, 1H), 7.32 (d, J\u00a0=\u00a02.4\u00a0Hz, 1H), 4.72 (d, J\u00a0=\u00a015.8\u00a0Hz, 1H), 4.36 (s, 3H), 4.09 (d, J\u00a0=\u00a015.8\u00a0Hz, 1H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 164.3, 161.2, 159.7, 156.4, 154.8, 147.8, 146.0, 134.9, 130.6, 126.9, 126.5, 123.9, 122.0, 120.4, 113.4, 112.7, 100.3, 76.1, 75.8, 56.0, 35.8. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221280.16. HRMS (APCI): m/z calcd for C21H16F3N2O5 [M\u00a0+\u00a0H]+ 433.1011, found 433.1022.White solid, mp: 212.7\u2013214.7 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 11.72 (s, 1H), 9.03 (s, 1H), 8.41 (d, J\u00a0=\u00a07.9\u00a0Hz, 1H), 8.22 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 8.09 (s, 1H), 8.02 (d, J\u00a0=\u00a07.2\u00a0Hz, 1H), 7.79 (d, J\u00a0=\u00a07.5\u00a0Hz, 1H), 7.42 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 7.34 (s, 1H), 4.81 (d, J\u00a0=\u00a015.9\u00a0Hz, 1H), 4.37 (s, 3H), 4.08 (s, 1H), 2.80 (s, 3H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 163.3, 162.1, 159.3, 155.7, 151.8, 146.9, 143.4, 135.0, 130.4, 127.0, 126.2, 124.1, 123.8, 121.6, 121.3, 113.1, 112.7, 100.4, 75.6, 75.3, 75.0, 74.7, 36.3, 19.0, 17.0. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.64. HRMS (APCI): m/z calcd for C22H18F3N2O5 [M\u00a0+\u00a0H]+ 447.1168, found 447.1176.White solid, mp: 210.2\u2013210.9 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.19 (s, 1H), 8.32 (s, 1H), 7.60 (m, 3H), 7.47 (m, 2H), 7.36 (d, J\u00a0=\u00a08.4\u00a0Hz, 2H), 7.14 (d, J\u00a0=\u00a02.6\u00a0Hz, 2H), 6.88 (d, J\u00a0=\u00a08.2\u00a0Hz, 1H), 4.24 (d, J\u00a0=\u00a015.4\u00a0Hz, 1H), 3.37 (s, 2H), 2.36 (s, 4H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 162.2, 159.5, 159.1, 154.1, 151.9, 150.4, 143.4, 135.5, 134.4, 133.8, 129.1, 128.4, 126.7, 125.1, 123.9, 122.9, 119.0, 118.6, 117.3, 115.9, 115.2, 113.8, 75.1, 36.1, 20.7. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.30, \u2212111.35. HRMS (APCI): m/z calcd for C21H15F4N2O4 [M\u00a0+\u00a0H]+ 435.0968, found 435.0947.White solid, mp: 217.9\u2013219.5 \u2103. 1H NMR (400\u00a0MHz, DMSO\u2011d6\n) \u03b4 12.38 (s, 1H), 8.32 (s, 1H), 7.95 (d, J\u00a0=\u00a02.4\u00a0Hz, 1H), 7.65 (m, J\u00a0=\u00a08.7, 2.5\u00a0Hz, 1H), 7.60 (s, 1H), 7.45 (d, J\u00a0=\u00a08.5\u00a0Hz, 1H), 7.33 (d, J\u00a0=\u00a08.4\u00a0Hz, 2H), 7.09 (d, J\u00a0=\u00a08.7\u00a0Hz, 1H), 4.26 (d, J\u00a0=\u00a015.5\u00a0Hz, 1H), 3.37 (d, J\u00a0=\u00a015.7\u00a0Hz, 2H), 2.34 (s, 4H). 13C NMR (101\u00a0MHz, DMSO\u2011d6\n) \u03b4 160.8, 159.1, 153.6, 151.9, 146.9, 143.5, 135.0, 134.4, 133.8, 131.1, 129.3, 126.7, 125.1, 123.9, 122.6, 118.6, 116.0, 75.7, 75.4, 75.1, 74.8, 36.3, 20.7. 19F NMR (376\u00a0MHz, DMSO\u2011d6\n) \u03b4 \u221278.29. HRMS (APCI): m/z calcd for C21H15ClF3N2O4[M\u00a0+\u00a0H]+ 451.0672, found 451.0649.Green metrics calculations for compound 3aa.\n\n1.\n% Yield = (0.362\u00a0\u00d7\u00a0100)/0.402\u00a0=\u00a090\u00a0%\n\n\n2.\nAtom Economy = (402\u00a0\u00d7\u00a0100)/(242\u00a0+\u00a0160)\u00a0=\u00a0100\u00a0%\n\n\n3.\nCarbon Efficiency\u00a0=\u00a0[(1\u00a0\u00d7\u00a020)\u00a0\u00d7\u00a0100]/[(1\u00a0\u00d7\u00a011)+(1\u00a0\u00d7\u00a09)]\u00a0=\u00a0100\u00a0%\n\n\n4.\nRME = (0.362\u00a0\u00d7\u00a0100)/(0.242\u00a0+\u00a00.160)\u00a0=\u00a090\u00a0%\n\n\n5.\nMass intensity = (0.242\u00a0+\u00a00.160)/(0.362)\u00a0=\u00a01.11\n\n\n6.\nE-Factor = (1.11\u20131)\u00a0=\u00a00.11\n\n\n% Yield = (0.362\u00a0\u00d7\u00a0100)/0.402\u00a0=\u00a090\u00a0%Atom Economy = (402\u00a0\u00d7\u00a0100)/(242\u00a0+\u00a0160)\u00a0=\u00a0100\u00a0%Carbon Efficiency\u00a0=\u00a0[(1\u00a0\u00d7\u00a020)\u00a0\u00d7\u00a0100]/[(1\u00a0\u00d7\u00a011)+(1\u00a0\u00d7\u00a09)]\u00a0=\u00a0100\u00a0%RME = (0.362\u00a0\u00d7\u00a0100)/(0.242\u00a0+\u00a00.160)\u00a0=\u00a090\u00a0%Mass intensity = (0.242\u00a0+\u00a00.160)/(0.362)\u00a0=\u00a01.11E-Factor = (1.11\u20131)\u00a0=\u00a00.11Antifungal assays were performed against Fusarium oxysporum (F. oxysporum), Fusarium graminearum (F. graminearum), Phytophthora parasitica var ni-cotianae (P. nicotianae), Fusarium moniliforme (F. moniliforme), and Rhizoctonia solani Kuhn (R. solani) in vitro by the plate growth rate method. The synthesized compounds were dissolved in 2\u00a0% DMSO to yield a 10\u00a0mg/mL stock solution. Then, each solution was added to sterile potato dextrose agar (PDA) to give final concentrations of 0.1\u00a0mg/mL. After the mixture was chilled, the mycelium of the fungi was transferred to the test plate and incubated at 26\u00a0\u00b0C. When the mycelium of the fungi reached the edges of the control plate (without sample), the inhibitory index was calculated as follows: In-hibitory index (%) = (Db-Da)/(Db-Dc)\u00a0\u00d7\u00a0100\u00a0%, where Da is the colony diameter of the growth zone in the test plate, Db is the colony diameter of the growth zone in the con-trol plate, and Dc is the diameter of the mycelial disc. The median effective concentra-tion (EC50) of each compound with a significant fungicidal activity was further evalu-ated in three independent experiments. The statistical analyses were performed using SPSS software (IBM SPSS Statistic 26).Inevitably, with our ongoing interests in the development of simple and mild methodologies for organic syntheses, we began our studies by employing 3-(trifluoroacetyl)coumarin 1a and 2-methyl quinazolinone 2a as model substrates for the CH functionalization reaction, and the results are summarized in Table 1\n. An initial trial under a mild condition consisting of 20\u00a0mol% Na2CO3 in 1,4-dioxane at 90\u2103 gave only 20\u00a0% yield of 3aa was isolated after 24\u00a0h (entry 1). Subsequently, a comprehensive screening of reaction catalysts (acids, bases as well as amino acids) was carried out (entries 2\u20136). However, acids or bases used in the reaction did not give promising results and the isolated yields of desired product 3aa were poor (<50\u00a0%, entries 1\u20134). Moving toward greener catalysts, that is amino acid, did not give satisfactory results (entries 5\u20136). We further continued our screening under catalyst- and solvent-free conditions, surprisingly, we obtained the product 3aa in higher amounts (67\u00a0%) and short reaction time (2\u00a0h) compared to other previous conditions (entry 7). Therefore, we further performed a set of the reaction under solvent- and catalyst-free conditions at elevated temperatures (entries 8\u201312). To our delight, on only increasing the reaction temperature, the better yield of desired product 3aa was obtained (90\u00a0%, entry 9). Further increasing or decreasing the reaction time diminishes the yield of 3aa (entries 11\u201312). Therefore, the reaction of 1a with 2a at 120\u00a0\u00b0C under solvent- and catalyst-free conditions for 2\u00a0h was the optimized eco-friendly reaction condition for the synthesis of quinazolinone derivative 3aa (entry 9).Next, we started exploring the scope of the reaction using various substituted 2-methyl quinazolinones and 3-(trifluoroacetyl)coumarins under the optimized reaction conditions (Table 2\n). Generally, the reactions of quinazolinones with electron-withdrawing and -donating groups with 3-(trifluoroacetyl)coumarin 1a all proceeded smoothly to give corresponding products 3aa-3al (entries 1\u201312). In the case of disubstituted quinazolinones, the reaction led to decreased yields (entries 2, 7 and 9). To further expand the scope of the methodology, we subsequently investigated the reactions with various 3-(trifluoroacetyl)coumarins and quinazolinones (entries 13\u201337). In all cases, the reactions ran efficiently to give the desired products in moderate to high yields. Meanwhile, the structure of 3di was also determined by analogy on the basis of X-ray (CCDC 2225572).Additionally, various green chemistry parameters were investigated and demonstrated by the green assessment diagram for the synthesis of quinazolinone derivatives (Fig. 3\n). Our developed methodology has remarkable advantages in terms of several green parameters, for example, atom economy of the synthesis of 3aa was reached 100\u00a0%; carbon efficiency of the reaction was 100\u00a0%; reaction mass efficiency was observed up to 90\u00a0%; and most importantly the E-factor was reduced to 0.11.The in vitro antifungal activity of target compounds against F. oxysporum, F. graminearum, P. nicotianae, F. moniliforme and R. solani is summarized in Fig. 4\n. Triadimefon was used as positive controls at a concentration of 10\u00a0\u03bcg/mL. In general, most of the title compounds exhibited a certain degree of fungicidal activity at a concentration of 500\u00a0\u03bcg/mL. Overall, most of the desired quinazolinone derivatives showed fungicidal activities against the abovementioned five fungi. Gratifyingly, the antifungal activity against R. solani was obviously better compared to other four fungi, with a moderate to good inhibitor rate. Particularly, compounds 3bl, 3ce, 3\u00a0cl and 3ea displayed good (>93\u00a0%) in vitro fungicidal activity. Among them, compound 3\u00a0cl was the most potent and had the EC50 values of 10.6\u00a0\u03bcg/mL.In summary, we have described an environmentally benign, straightforward access for the synthesis of quinazolinone derivatives in the absence of catalyst and solvent in excellent yields. The advantages of the developed methodology are experimental simplicity, easy work-up, and excellent yields of products. This eco-friendly methodology also eliminates toxic metal catalyst-related environmental hazards and pollutions along with increase in atomic economy. Furthermore, the preliminary in vitro antifungal activity revealed that most of the synthesized compounds displayed promising fungicidal activities. Among them, compound 3\u00a0cl exhibited 95\u00a0% fungicidal activity against R. solani, with an EC50 value of 10.6\u00a0\u03bcg/mL.\nXiaodan Chang: Conceptualization, Investigation, Data curation. Liangxin Fan: Resources, Formal analysis. Lijun Shi: Data curation. Zhenliang Pan: Formal analysis. Guoyu Yang: Validation, Visualization. Cuilian Xu: Resources, Writing \u2013 review & editing. Lulu Wu: Conceptualization, Methodology, Investigation, Writing \u2013 original draft. Caixia Wang: Resources.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by Natural Science Foundation of Henan Province (22230420459, 222300420456) and Science and Technology Innovation Fund of Henan Agricultural University (KJCX2020A19).Supplementary data to this article can be found online at https://doi.org/10.1016/j.jscs.2023.101621.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  An environmentally benign highly atom-economic protocol for the construction of the CC bond has been developed under catalyst- and solvent-free conditions. This protocol involves the efficient coupling of 2-methyl quinazolinones with 3-(trifluoroacetyl)coumarins for the access of quinazolinone derivatives in excellent yields (up to 90\u00a0%). The crystal structure of compound 3di was investigated by X-ray diffraction analysis. The biological activities, such as in vitro antifungal activity of the quinazolinone derivatives against Fusarium graminearum, Fusarium moniliforme, Fusarium oxysporum, Phytophthora parasitica var nicotianae, and Rhizoctonia solani Kuhn, were investigated. The bioassay results indicated that most of the target products exhibited promising fungicidal activities, and compound 3\u00a0cl exhibited 95\u00a0% fungicidal activity against R. solani, with an EC50 value of 10.6\u00a0\u03bcg/mL.\n               "}
{"full_text": "Air pollution has always been a hot issue of public concern; substantial works have been applied to air pollution control for decades [1]. Volatile organic compounds (VOCs) have extensive sources, mainly including industrial and traffic exhaust emissions. The discharged pollutants such as benzene, toluene, formaldehyde, xylene and so on are toxic [2,3]. Secondary reactions and more complex pollutants are caused because of the diversity of VOCs [4]. It is an important factor that intensifies the atmospheric compound pollution, which directly affects human health and the environment [5]. In the last few decades, VOCs control is mainly based on end treatment, including single technologies such as adsorption [6\u201310], membrane separation [11], plasma [12], biodegradation [13], thermal catalysis [14\u201318], photocatalysis [19\u201321] by various environmental functional materials [22\u201325]. However, the increasing demands of VOCs control are difficult to be satisfied by single traditional technologies. The coupling and collaborative processing between several technologies have shown new vitality, such as adsorption-catalysis [26,27] adsorption-plasma [28] and plasma-catalysis [29], adsorption-plasma catalysis [30].Photothermal catalysis is a new coupling technology in recent years and has been rapidly developed, which can simultaneously solve the high-energy consumption of thermal catalysis [31,32] and the low efficiency of photocatalysis [33,34]. Photothermal catalysis is not a simple superposition of thermal catalysis and photocatalysis. It can make use of solar energy and thermal energy simultaneously and further improve catalytic efficiency through synergistic effect [35]. At present, several research methods of photothermal catalysis mainly include as follows:(1) Light driven thermal effect, without additional heat source; (2) To introduce additional heat source into the photocatalytic system; (3) To introduce light irradiation in the thermal catalytic system. The application of photothermal catalysis in the environmental field is limited, accounting for only about 2.1% (Fig.\u00a01\n). In recent years, photothermal synergy has been widely studied in VOCs oxidation (Fig.\u00a02\n) [36\u201349], carbon dioxide (CO2) reduction [50,51], hydrogen production [34,52], CO2 methanation [53,54] and methanol reaction systems [55,56]. Some progress has been made in catalyst design and the removal of complex pollutants. The reported photothermal catalysts can effectively degrade VOCs under different light irradiation conditions, and have excellent catalytic stability, such as MnOx [40,57], Co3O4 [58\u201360], TiO2 [12,61], perovskite [42,62] and noble metal composites [44,63], etc.Recently, the relevant review on photothermal catalysis mainly focuses on the design and application of nano-catalysts [64,65]. The mechanism of light-driven thermal catalysis is also briefly summarized by An et\u00a0al. [66]. However, the research on the mechanism of photothermal oxidation of VOCs, also the standardized comparison based on the performance over different photothermal catalytic systems are still rare. It is difficult to make a systemic comparison of the catalytic performances of various photothermal catalytic reactors because of the differences in performing parameters. On this basis, this paper focuses on the research progress of photothermal catalytic performance and the VOCs elimination mechanisms in batch and continuous systems, as well as the fabrication of photothermal catalytic materials (Fig.\u00a03\n). Considering the basic research and practical application, this review provides an opportunity to compare the efficiency of each method on a parallel basis. In addition, the application of theoretical calculation in photothermal catalysis and related mechanisms of VOC destruction are summarized and discussed. Finally, the existing problems are pointed out to guide photothermal catalysis technology in the future. This review will be devoted to point out a potential direction in the field of photothermal catalysis and make the related research work in the environmental field more successfully.In general, photothermal catalysts mainly include two types according to the current researches. One is metal oxide with both photo responsive properties and thermal catalytic activity, such a single oxide can be excited by both synchronously by light and heat. Another one is composite catalyst composed by noble metal particles (or transition metal oxides) and semiconductor oxide carrier. Among them, noble metal particles (or transition metal oxides) are dedicated to the thermal catalytic activity, and semiconductor oxide carriers play an excellent light response performance. Each part of the composite is activated by light and heat, respectively, and the synergistic effect is produced through material coupling. The light required for these catalysts comes from external light sources such as Xe lamp and Hg lamp used in the laboratory. The heat comes from the photothermal conversion caused by light irradiation or an external heat source. To achieve the effective performance of photothermal catalysis, the optical properties, photothermal conversion efficiency, also thermal catalytic activity of the catalyst should be considered [19,67,68].The photocatalytic efficiency is limited due to the low utilization rate of solar energy and the fast recombination of current carriers by traditional semiconductors with a wide band gap. To make full use of the solar energy, the catalyst should be designed to achieve broadband absorption between 300 and 2500\u00a0nm to maximize energy capture from solar light and activate the reaction. Different strategies have been developed to improve the light absorption capacity of catalysts, including the use of strong light absorption carriers (such as silica nanowire arrays [69,70] and MXene [71]) and the construction of plasma superstructures [72]. In addition, electron\u2013hole pairs are generated when the semiconductor absorbs photons with energy equal to or above the band gap. Eventually, these photo-generated carriers can migrate to the semiconductor surface and be transferred to the adsorbed molecules, thereby reduction or oxidation are initiated [19,73]. Therefore, to enhance solar spectral absorption or reduce the band gap is conducive to obtain a smaller carrier recombination rate, thus improving the photocatalytic efficiency.Photothermal conversion refers to the process of concentrating solar irradiation energy through reflection, absorption, or other ways and converting it into high temperature. When materials surface is exposed to electromagnetic irradiation, part of the photon energy is converted into other energy, such as thermal, electricity, chemical, and biomass energy [74\u201378]. To realize the efficient thermal utilization of solar energy, photothermal conversion materials are the basic medium required [79]. In general, materials with energy level transitions equivalent to that of photon energy are often used for photothermal conversions, such as metals, metal oxides, metal sulfides, and semiconductors [80\u201386]. In addition, another effective way to improve the photothermal conversion efficiency is to reduce the heat dissipation [87,88]. The photothermal conversion efficiency is an important index. In general, the ratio of the thermal energy converted by a material to the incident light energy is defined as the photothermal conversion efficiency (\u03b7) which can be calculated by Equation (1) [89].\n\n(1)\n\u03b7\u00a0=\u00a0Ethermal /Ephototons \u00d7 100%\n\n\n\nE\n\nphototons\n represents the energy of the incident photon, and E\n\nthermal\n represents the thermal energy converted after the catalyst absorbs the simulated light irradiation. Detailed calculation can be carried out according to the following formula [90]:\n\n(2)\n\n\u03b7 = Cp, i \u00d7 mi \u00d7 (T - Tsurr) / (A \u00d7 q \u00d7 t) \u00d7 100%\n\n\n\nThe Cp, i represent the specific heat capacity, which can be measured by a Heat Conduction Analysis Meter. The mi is the mass of the catalyst. The T and Tsurr are the catalyst temperature and ambient temperature, respectively. The A, q and t are the surface area of the catalyst, the power density of light source, and the irradiated time, respectively. The \u03b7 value depends on the absorbance of light and the surface area of the catalyst. Higher platform temperatures can be obtained through strong light absorption and surface area.Thermal oxidation is the preferred method for industrialization because of its economic and environmental advantages. The reaction is affected by many factors, such as low reaction temperature, competitive reaction, small amounts of active sites, brief residence time and catalyst deactivation. The presence of these factors leads to the formation of abundant intermediate products in the reaction system. During the catalytic process, the stubborn carbonaceous intermediate products are deposited on the surface of the catalyst, which hinders the contact between the pollutant molecules and the active sites. According to the photothermal conversion properties of the catalyst and the incident light intensity, the introduction of light in the thermal catalytic system has a positive effect on the temperature rise of catalyst surface. Catalysts with good thermal catalytic performance can directly utilize the heat energy generated by light, thus further reducing energy consumption, and having a positive impact in promoting the reaction. At present, the thermal catalytic activity of materials can be effectively enhanced through loading, doping modification [91,92], and defect engineering.Based on the description in Section 2.1, plenty of catalytic materials with photothermal synergistic effects have been designed and applied. Some reported photothermal catalysts for VOCs degradation and their properties are summarized in Table 1\n [36\u201339,42\u201349,58\u201363,93\u2013114].Transition metal oxides, which are usually polycrystalline and polyvalent, exhibit significant activity in the removal of gaseous pollutants [115\u2013119]. At present, the photothermal catalysis of VOCs by metal oxides has received a lot of attention, including Co3O4 [58,59], MnOx [102], TiO2 [57], etc.\nCo\n\n3\n\nO\n\n4\n\n. As an excellent catalytic material, Co3O4 has excellent light capture and electron-mediated properties [120,121]. As described in Section 2.1, a better photothermal catalysis performance can be improved by enhancing light absorption and heat conversion. The light absorption properties of catalysts could be modulated by shape and structure regulation or metallic element doping. Chen et\u00a0al. found that structure management and secondary metal doping techniques were effective to improve the solar light utilization efficiency of the catalyst [58]. NiOx/Co3O4 composites derived from Ni-doping ZIF-67 were prepared by the impregnation method. The full spectral absorption of NiOx/Co3O4 composites was significantly enhanced after Ni doping due to metal-to-metal charge transfer (MMCT). The MMCT process occurs when two metal ions with different valence states are bridged together. The difference in charge causes an electron transition and the two metal centers are reduced and oxidized, respectively. The transition oxo-bridged and the all-inorganic heterobinuclear units can extend light absorption from the ultraviolet region to the visible region [122]. In the meanwhile, the hollow structure of NiOx/Co3O4 allowed multiple reflections of light in the internal cavity, thus effective contact between photons and matter was promoted. Wang et\u00a0al. prepared ultrathin mesoporous Co3O4 nanosheets on stainless steel mesh (SS-Co3O4) with high photothermal properties by electrochemical deposition method [59]. SS-Co3O4 exhibited strong absorption throughout the entire solar spectrum. The temperature of SS-Co3O4 rapidly raised from room temperature to 110\u00a0\u00b0C and finally stabilized at 175\u00a0\u00b0C, which revealed its efficient photothermal conversion. The combined properties of SS-Co3O4 ultrathin with two-dimensional shape, foam porous structure and metal substrate resulted in the maximum photothermal conversion, more surface catalytic active sites, lattice oxygen species and mobility.In addition to improving the light absorption and heat conversion of Co3O4, it is also possible to improve the thermal catalytic activity at low temperatures. Reactions are more likely occurred when the atoms on the catalyst surface are in an unsaturated coordination state [123]. When Co3O4 forms a bimetallic catalyst with other elements, the electronic structure of the cations on the surface can be adjusted by weakening the Co-O bond [60]. Jin et\u00a0al. reported a quenching method that hot Co3O4 nanosheets were poured into copper nitrate solution and Cu2+ modified Co3O4 was obtained. Different from hydrothermal synthesis, Cu2+ only modified the surface layer of Co3O4 after quenching and Cu2+ was easier to replace Co2+than Co3+, thus more active sites were generated [60].\nMnO\n\nx\n\n. MnOx has good thermal catalytic activity at low temperatures [124\u2013126], most of the photothermal modification of MnOx focuses on light absorption and photothermal conversion. Yang et\u00a0al. synthesized a new hollow pellet composed of tightly packed nanosheets (R-MnO2-HS) [96]. R-MnO2-HS could effectively convert solar energy into thermal energy under full spectrum, visible and infrared light irradiation, the sample temperature raised from room temperature to 226, 220, 211\u00a0\u00b0C, respectively. Besides, there was almost no photocatalytic performance when the reactor temperature is kept at ambient temperature under full spectral irradiation, so the catalytic performance is mainly due to light-driven thermal catalysis. Therefore, the catalytic performance of R-MnO2-HS is mainly attributed to the light-driven thermal catalysis. The researchers also improved the light absorption and photothermal conversion of MnOx by combining multiple materials. Carbonaceous materials, especially graphene, display high capability of photo-absorption form the ultraviolet to near-infrared region [127\u2013129], and then light energy can be converted into thermal energy through non-radiative decay [130]. On this basis, multivariate manganese oxide catalysts have been reported widely. Dong et\u00a0al. prepared 2D graphene oxide, MnOx and polymerized carbon nitride nanosheets as building blocks through a filtration method (Fig.\u00a04\na). The superior photothermal effect of graphene caused the temperature of the film to rise to 85\u00a0\u00b0C in 15\u00a0min, which then initiated the thermal catalytic reaction of manganese oxide. Furthermore, the 2D/2D/2D assembly of these nanosheets in the membrane facilitated the transfer of energy and charge carriers between the various nanosheets. The increased temperature would activate the lattice oxygen and adsorbed oxygen molecules of MnOx, and promoted the transfer of photogenerated electrons and holes to the surface of the carbon nanotubes and the subsequent surface reactions [48]. Except for traditional manganese oxides, manganese potassium ore octahedral molecular sieve (OMS-2) has been widely used due to the unique characteristics of porosity, mixed-valence state, and easy release of lattice oxygen. Li et\u00a0al. [39,93,94] synthesized a series of OMS-2 composites doped with metal elements (Mg, Fe, Ce). The OMS-2 catalyst substituted by nanometer Mg (or Fe, Ce) ion had strong absorption in the whole solar spectrum region, it could effectively convert the absorbed solar energy into thermal energy. The catalyst surface temperature was higher than VOC ignition temperature, which promoted the catalytic reaction.\nTiO\n\n2\n. TiO2 is one of the most promising photocatalysts due to its excellent activity, good stability, non-toxicity, and relatively low price. However, TiO2 has a wide band gap and only responds to ultraviolet light. Moreover, the surface of TiO2 is easily occupied by toxic by-products and deactivated in the process of photocatalysis. The effect of reaction temperature on the photocatalytic oxidation of refractory carcinogenic benzene by anatase TiO2 nanosheets with (001) crystal faces were studied by Li et\u00a0al. [99]. The TiO2 nanosheets were coated on the surface of the Hg lamp, and the efficient photothermal catalytic oxidation of benzene was realized by using the heating effect of the UV and infrared light, without an extra heater. At the same time, it is an effective way to construct photothermal catalytic materials by combining other metal oxides with TiO2. MnOx/TiO2 [98], Co3O4/TiO2 [61], TiO2/CeO2 [103], CeMnxOy/TiO2 [101] and nano-TiO2-supported amorphous manganese oxide [104] were prepared through hydrothermal redox reaction in the presence of TiO2 (P25). These nanocomposites could effectively convert the absorbed solar energy into thermal energy, the surface temperature would be significantly higher than the light-off temperature of benzene oxidation [101]. In addition, the deposition of carbonaceous intermediates on TiO2 surface was inhibited by photothermal catalysis, thus the catalytic activity and durability were significantly improved [61,98].\nPolymetallic oxides. Perovskites can provide a stable framework for elemental doping and electron band structure regulation because of the stable structure, abundant metal ions and anions [62,131]. In the meantime, perovskites show a good prospect in visible light-driven photocatalysis because of their good band-edge potential and changed band structure. Chen et\u00a0al. [42] synthesized ABO3 type perovskite (A\u00a0=\u00a0La, Ce, Sm; B = Cr, Mn, Fe, Co, Ni), which was successfully applied to photothermal catalytic degradation of VOCs under visible light for the first time. All the synthesized ABO3 type perovskites showed excellent UV and visible light absorption, especially for LaMnO3 and LaNiO3. Besides, the band gap could be changed with different metal ions since B ion was located on the 3d orbit to form the conduction band, thereby the photo response, redox activity and other properties were accordingly affected. The band gaps of these perovskites followed the order: LaCrO3 (3.1\u00a0eV)\u00a0>\u00a0LaCoO3 (2.9\u00a0eV)\u00a0>\u00a0LaMnO3 (2.5\u00a0eV)\u00a0>\u00a0LaNiO3 (2.4\u00a0eV)\u00a0>\u00a0LaFeO3 (2.1\u00a0eV). The narrower band gap allowed for more efficient transfer of photogenerated electrons and separation of electron\u2013hole pairs [132]. Even under the premise of making full use of visible light, the utilization rate of the full solar spectrum is still low. The energy in the infrared region is wasted mainly by heat dissipation. Making full use of the energy in the infrared region becomes a hot issue that must be considered. Xu et\u00a0al. [47] found that CeO2/LaMnO3 composites had the characteristics of wide wavelength absorption (800\u20131800\u00a0nm), which could be used as a highly active photothermal response catalyst for the decomposition of VOCs under infrared irradiation. Under the infrared irradiation intensity of 280\u00a0mW/cm2, the maximum photothermal conversion efficiency of CeO2/LaMnO3 was 15.2%.The crystal structure of spinel contains tetrahedron and octahedron coordination sites, the accommodated different metal cations in the structure result in more reactive oxygen species. Chen et\u00a0al. synthesized ACo2O4 (A\u00a0=\u00a0Ni, Cu, Fe, Mn) spinel by co-precipitation method. The MMCT effect between Co ion and A ion was affected by the d\u2013d indirect transition band gap of A ion, which led to the difference in light absorption capacity. The strong light absorption and high solar thermal effect of NiCo2O4 in the whole solar spectrum (200\u20132500\u00a0nm) provided enough heat energy for the catalytic degradation of toluene [106].Compared with the high energy consumption in the catalytic process of metal oxides, noble metals are widely used in the thermal catalysis and photocatalysis of VOC due to their excellent activity [45,46,133\u2013135]. As noble metal particles tend to aggregate at high temperature, semiconductor metal oxides are often used as carriers [136]. The noble metal\u2013semiconductor structure is beneficial to broaden the spectral response region and the separation of photogenerated carriers. Schottky potential barriers between the noble metals and semiconductors can be generated due to the deposition of noble metals, thus promoting the separation of photogenerated electron\u2013hole pairs. Cai et\u00a0al. [63] reported that Pt nanoparticles were well dispersed on porous \u03b3-Al2O3 (Fig.\u00a04b (i)). Pt nanoparticles were used as a light absorbent and catalytic active site. As shown in Fig.\u00a04b (ii), Pt/\u03b3-Al2O3 showed obvious strong absorption in the wavelength range of 200\u20132500\u00a0nm, while pure \u03b3-Al2O3 has no obvious light absorption. The extraordinary optical absorption further confirmed that the thermal-electronic and photothermal effects of Pt nanoparticles could efficiently drive the catalytic reaction. On the other hand, the temperature of the catalyst's surface was monitored during the irradiation process (Fig.\u00a04b (iii). The temperature of catalyst raised to the platform temperature under irradiation in 15\u00a0min without an additional heater due to the excellent photothermal conversion. There was a positive correlation between temperature and Ag loading. The highest surface temperature was obtained over 2.81 Pt/\u03b3-Al2O3 (169\u00a0\u00b0C) [63]. The combination of noble metals with other transition metals or carbonaceous materials can also effectively improve light absorption and photothermal conversion. Jia et\u00a0al. reported that the photothermal conversion efficiency of hybrid nanomaterial Pt-rGO-TiO2 reached 14.1% under the infrared irradiation intensity of 116\u00a0mW/cm2 due to the effective photothermal conversion, increased light adsorption and well-dispersed Pt nanoparticles [45]. The light absorption, photothermal conversion are promoted mainly due to the localized surface plasmon resonance (LSPR), the strong metal-support interaction (SMSI) [45,114,137,138].As shown in Fig.\u00a05\na, the LSPR effect on the surface of the noble metal is enhanced under light irradiation. The energy of the hot electrons generated by the electrons' collective oscillation is much higher than that of the electrons in a thermodynamic equilibrium state [63,137]. In essence, hot electrons are the products of surface plasmon oscillations in nanostructures that pump electrons from a lower energy level to a higher energy level through Landau relaxation [139,140]. The LSPR effect of noble metal occurs under light, and the generation of hot electrons is conducive to the rapid heating of the system and the activation of semiconductor lattice oxygen. The range of LSPR induction can be broadened by manipulating the plasma nanostructure [41]. Mao et\u00a0al. [44] reported that Pt nanoparticles were limited to mesoporous micron-sized CeO2 successfully. A new hot electron-induced photoactivation process was proposed. The intense surface plasma absorption of Pt nanoparticles was beneficial to the catalytic activity and regional heating effect. Pt nanoparticles could absorb visible light photons and generate hot electrons, while the visible light photons could not be absorbed by CeO2 due to the large band gap. Besides, the LSPR thermal electron conversion to chemical reaction energy is inefficient due to the rapid relaxation of the carriers. It is valuable to design catalysts that are highly responsive to visible light irradiation and highly efficient in converting photons into chemical energy. Zou et\u00a0al. found that Pd-Ce catalyst was synthesized by a liquid-phase reduction method assisted with cetyltrimethylammonium bromide (CTAB), and more active interfaces were produced. The transfer of hot electrons from Pd particles to CeO2 because of surface plasmon resonance promoted the dissociation of adsorbed oxygen. The maximum light utilization rate of Ce/Pd catalyst for toluene oxidation and CO oxidation reached 0.42% and 1%, respectively, which was attributed to the effective Ce/Pd interface [41].The importance of SMSI in reducible oxides supported noble metal nanoparticles has been recognized in many thermal catalytic systems. SMSI promotes the activity of the loaded noble metals mainly by affecting the formation of negatively charged noble metals nanoparticles and chemisorbed oxygen. The noble metal nanoparticles were stabilized by geometric modification and electronic modification [108,141]. The strong interaction between the noble metal and the carrier interface can be enhanced by doping and modification of the carrier, different synthesis methods and the formation of composite catalysts, which is of great significance for improving the catalytic performance of photothermal synergistic catalytic materials. Yang et\u00a0al. reported that the SMSI in strontium titanate (STO) supported Pt nanoparticles (NPs) significantly promoted the photothermal oxidation of toluene under visible light. Chemisorbed oxygen and negatively charged Pt NPs were formed due to the SMSI effect, the oxygen activation and the surface plasma resonance effect of Pt NPs were promoted in the meanwhile. The above phenomenon promoted visible light absorption, photoionizing separation, and generation of reactive oxygen species [108].In recent years, defect chemistry and engineering techniques of metal oxides have attracted extensive attention because of their important role in regulating catalytic performance (Fig.\u00a05b) [142,143]. The existence of defects promotes the adsorption and activation of substrate molecules, thus accelerating reaction thermodynamics and kinetics [144,145]. In photocatalysis, electrons can be trapped by the defects and release thermal energy to the surface, which improves charge utilization and speeds up the surface reaction rate. In order to pursue more excellent catalytic performance, defect engineering has been widely studied in the past decades [146]. Li et\u00a0al. reported for the first time that porous Co3O4 nanorods (Co3O4-MNR) with a large amount of Co2+ vacancy defects greatly enhanced the photothermal catalytic activity of Co3O4 (Fig.\u00a05c) [95].It is found that the photocatalytic/catalytic performance of oxides significantly depends on the type and concentration of defects [142,143,146\u2013149]. Oxygen vacancy (OV) is the most utilized defect. OVs can effectively separate the photoexcited electrons and holes, thus improving the photocatalytic performance [150]. It is a natural defect of oxides and can be easily incorporated into the lattice of oxides by plasma treatment technology [151], acid treatment [152] and annealing in an anoxic atmosphere [153]. Zhang et\u00a0al. proposed that a solution plasma processing technique could be applied to process pre-synthesized TiO2 containing a large amount of OVs. Hydrogen dopants were added to the TiO2 lattice to produce defects, and the original visible absorption of colored TiO2 was still retained [12]. Huang et\u00a0al. demonstrated a simple and effective route to introduce OVs into (001) face of BiOI nanosheets by modification with low concentration nitric acid for the first time. The vacancy increased the maximum valence band of the BiOI nanosheet and ensured that more charge carriers were converted to \u00b7O2\u2212 in photothermal catalysis. In addition, the defective BiOI nanosheets could absorb more visible light than the original BiOI nanosheets, which was important for improving photothermal catalytic performance [111]. Carbon deposition in the process of catalytic oxidation of VOCs is the key factor that restricted the stability of catalysts. An et\u00a0al. introduced OVs into CeO2 through redox and steam treatment (ARCeO2), which showed high coke resistance performance. The abundant OVs in ARCeO2 enhanced light absorption, improved charge separation, and increased the generation of reactive oxygen species, thus improving its photothermal catalytic performance [49]. Although OVs can promote the catalytic reaction, however excessive OVs will act as the complex center and reduce the charge transfer to the catalyst surface [111]. Quantitative analysis and characterization of OVs have always been a difficulty in related studies. Some characterizations have been employed to measure oxygen vacancies, such as electron spin resonance (ESR) [154,155], x-ray photoelectron spectroscopy (XPS) [156,157], situ UV-Raman [158], etc. These studies further reveal that rational surface defects engineering is an extremely effective and advanced route in the promotion of photothermal catalysis.Moreover, the heterojunction was constructed for photothermal catalysis of VOCs because of the unique charge separation and transfer behavior, remarkable light response, as well as strong redox ability [114,159]. Z-scheme Ag3PO4/Ag/SrTiO3 Heterojunction [159], a new type of WO3/Ag/GdCrO3 named as Type B heterojunctions [50] and band gap-broken Ag3PO4/GdCrO3 heterojunction [51] were constructed by Rui et\u00a0al. Fig.\u00a06\na clearly revealed the charge transfer in WO3/Ag/GdCrO3 heterojunction. The thermal electrons and holes generated under photothermal process broke through the energy barrier of WO3/Ag/GdCrO3, and then combined with the photogenerated electron\u2013hole pairs generated by the semiconductor. Thus, the spatial separation of photoinduced charge was realized, and the strong light absorption and redox ability were retained [50]. In order to further clearly observe the effect of photothermal synergy on charge migration in heterojunction, photoinduced charge migration of Ag3PO4/GdCrO3 and Ag3PO4/Ag/GdCrO3 under different conditions was conducted by Rui et\u00a0al. (Fig.\u00a06b). The Ag3PO4 and GdCrO3 produced electron\u2013hole pairs in the irradiation of visible light, and hot electron\u2013hole pairs also induced by LSPR of Ag NPs. At the same time, the local electromagnetic fields induced by Ag LSPR effect could accelerate the electron generated by Ag3PO4 transfer to the Ag3PO4-Ag interface, which was beneficial to the production and separation of the light induced charge in Ag3PO4. When the temperature increased, the Ag NP enhanced LSPR effect can promote the formation of hot holes and improve the scattering rate of hot electrons. The photoinduced charge separation of GdCrO3 and Ag3PO4 was enhanced through interfacial charge transfer by plasma Ag NP as a bridge under photothermal conditions [51]. Therefore, the construction of heterojunction has great research potential and application value in the field of photothermal catalysis.Besides studying the oxidation mechanism of VOCs, the catalytic reactor also plays an important role in the combustion. In this review, the VOC photothermal catalytic system is divided into the batch system (Fig.\u00a07\na) and the continuous system (Fig.\u00a07b). On the one hand, this classification is intended to facilitate the comparison of the performance of photothermal catalysts, since the operation reactor has a significant effect on the performance. On the other hand, batch reaction systems can usually simulate the variation of pollutants in indoor or confined spaces, while continuous reaction systems can simulate the variation of pollutants in outdoor and industrial emissions.The equipment and operation of a batch photothermal catalytic reactor are relatively simple, which is commonly applied in laboratory research (Fig.\u00a07a). The system is easy to adapt to different operating conditions and pollutants. Also, it is suitable for the degradation with small air flow and multiple VOC varieties. Basically, batch photothermal catalysis occurs in a closed reactor and an external light source irradiates the inside of the reactor through a quartz window of a certain size. Meanwhile, the reaction temperature is maintained through insulation equipment; Sometimes additional heating equipment is used to provide the corresponding thermal energy. Some of reactants are fed into the system and interact with each other. However, it is difficult to achieve continuous degradation in batch system and the reaction time is long. In this review. The photothermal catalysis, photocatalysis and thermal catalysis performance and experimental parameters of various VOCs in the batch system are summarized in Table 2\n [36\u201340,44,46,93\u2013104,110,113,160]. Obviously, photothermal catalysis has a higher degradation rate and shorter reaction time than thermal catalysis and photocatalysis. Aromatic hydrocarbons with benzene rings are stable and require high temperatures to destroy. Most of the researches on photothermal catalysis VOCs focuses on benzene series. Moreover, the light response range of the catalyst is continuously expanded to obtain better catalytic activity and solar energy utilization.Traditional photocatalysis is carried out under the irradiation of ultraviolet light. Li et\u00a0al. coated the surface of the UV lamp with TiO2 and Pt/TiO2. The ultraviolet irradiation and nonradiative heat energy emitted from the UV lamp were fully utilized, and the efficient photothermal catalytic oxidation of benzene (\u223c70\u00a0ppm) was realized (Table 2) [36]. In addition to the light driven thermal effect under UV irradiation, the combined effect of UV light and external thermal sources on the catalytic reaction had been further investigated. Wang et\u00a0al. successfully improved the mineralization of gaseous benzene over TiO2 by adjusting the temperature from room temperature to 280\u00a0\u00b0C under UV irradiation. The high-temperature desorption could accelerate the mineralization rate of inactive sites. However, at high temperatures, the reduction of \u2219OH on the surface weakened the oxidation of benzene, while heat promoted the oxidation rate [110].The solar spectrum only contains about 5% ultraviolet (UV). Expanding the absorption and utilization of visible and infrared light is extremely urgent. Therefore, some works have been conducted to improve infrared and visible responses. Li et\u00a0al. reported that efficient photothermal catalytic oxidation of benzene (\u223c250\u00a0ppm) was realized by coating TiO2 on the surface of the Hg lamp. The thermal effect of infrared light was further employed without an extra heater. Under the same irradiation conditions, the photothermal catalytic activity of TiO2 was increased by 42.3 times compared with that at room temperature and showed excellent durability [99]. Ji et\u00a0al. commercialized SrTiO3 by constructing fluorine ions and Ag nanoparticles (Ag/F-STO). Under visible light irradiation at 90\u00a0\u00b0C, Ag/F-STO could degrade 95% of benzene, toluene, and xylene (800\u00a0ppm) through photothermal catalysis after 6\u00a0h with a high degradation rate constant [46]. Kong et\u00a0al. prepared bifunctional 0.1\u00a0wt % Pt/SrTiO3-x, achieved 100% mineralization against toluene (500\u00a0ppm) under visible light irradiation and mild thermal energy input (\u2264150\u00a0\u00b0C) after 1\u00a0h. The introduction of photocatalysis reduced the activation energy in the conventional thermal catalysis process, the generation of \u2219O2\u2212 and \u2219OH was also reduced by activating oxygen and consuming water [113].In order to make full use of solar energy in practical application, the research of full spectrum is gradually coming into view. The Pt/CeO2 nanocomposite prepared by Mao et\u00a0al. showed high photothermal catalytic activity under the full solar spectrum, visible light-infrared, or infrared light irradiation. Pt nanoparticles had strong surface plasma absorption in the whole solar spectrum region. The surface temperature of the catalyst was stabilized to 173\u00a0\u00b0C, and the toluene (\u223c4900\u00a0ppm) was completely degraded after 25\u00a0min of irradiation [44]. The mesoporous Co3O4 nanorods prepared by Lan et\u00a0al., which contained a large amount of Co2+ vacancy defects, showed high photothermal catalytic activity for benzene oxidation under UV-Vis-IR. Benzene (\u223c4900\u00a0ppm) was completely oxidized to CO2 after 40\u00a0min of irradiation [95]. The performance of a single metal oxide is limited, so the composite catalysts have become the research trend of many researchers [61,98]. The MnOx/TiO2 [98], Co3O4/TiO2 [61] nanocomposites were prepared and showed good catalytic activity and durability for benzene oxidation under light irradiation. Xie et\u00a0al. found that OMS-2/SnO2 nanocomposites showed good photothermal catalytic activity and stability of benzene under UV-Vis-IR irradiation. The surface temperature of the catalyst was stabilized to 260\u00a0\u00b0C, and the toluene conversion rate reached about 88% after 35\u00a0min of irradiation. The initial CO2 yield of OMS-2/SnO2 with photothermal catalytic degradation of benzene was 83.3\u00a0\u03bcmol\u00a0min\u22121g\u22121, which was 37.2 times higher than that of pure SnO2 [100].In addition to aromatic hydrocarbons, photothermal catalysis of other VOCs in batch systems has also been reported, such as aldehyde and alkanes. The MnOx-CeO2 mixed oxide prepared by Jiang et\u00a0al. had strong light responsiveness and low-temperature reducibility, a good catalytic synergistic effect was founded in formaldehyde emission reduction. The degradation rate of formaldehyde (250\u00a0ppm) reached 90.4% after 3\u00a0h of infrared irradiation at 75\u00a0\u00b0C. In addition, the depleted material after a long period of dark reaction showed encouraging self-repair under in-situ light [160]. Kang et\u00a0al. catalyzed propane over Pt/TiO2-WO3 catalyst. After the introduction of UV-vis light, the reaction temperature required for 70% conversion of C3H8 was reduced to a record-breaking 90\u00a0\u00b0C from 324\u00a0\u00b0C [138].In the batch reaction system, if the operation time is long enough and there is no side reaction or catalyst deactivation, the conversion rate per unit volume of the batch reaction system is higher. From the view of experimental research, the batch system is easy to operate and control, it can provide unique insights for the analysis of catalytic mechanism, kinetic parameters, and overall performance. However, the batch reaction system is not widely used in large-scale industries because of its high operating cost and low versatility. At present, the catalysts studied in batch VOC treatment system mainly focus on transition metal oxide catalysts, but there is still some space for improvement compared with noble metal catalysts, which needs further research and practical application promotion.On an industrial scale, most chemical processes employ continuous systems due to their excellent heat and mass transfer performance. In addition, the continuous working mode uses a higher energy light source to improve performance. For example, at higher energy levels, the light source can handle more intake volume per unit of time than other operation modes. After all, continuous systems are considered easier to scale up to the practical application. A typical continuous photothermal catalytic system is shown in Fig.\u00a07b.There are few studies on the photothermal degradation of VOCs by continuous systems. The continuous system is mainly based on toluene. In order to improve the traditional photocatalytic mode and expand the application of solar energy, the main irradiation light sources are concentrated in infrared, visible and full solar spectrum. Due to the short contact time between gas flow and the catalysts in the continuous system, the VOC concentrations in the current experimental study are mostly concentrated in the range of 50\u2013200\u00a0ppm. Some relevant research advances are summarized in Table 3\n [43,45,47,49,58,62,63,105\u2013107].In the study with infrared light as the source, Li et\u00a0al. found the surface temperature of CeO2/LaMnO3 composites raised rapidly to 275\u00a0\u00b0C due to excellent photothermal conversion. The toluene (200\u00a0ppm) conversion rate was 89% and the CO2 production rate was 87% under the infrared irradiation intensity of 280\u00a0mW/cm2. Such enhanced photothermal catalytic activity was mainly attributed to the synergistic effect of ultra-broadband, strong light absorption, efficient photothermal conversion, good low-temperature reducibility, and high lattice oxygen mobility, which was caused by the strong interaction between LaMnO3 and CeO2 [47]. Li's group also reported that the significant toluene conversion rate of Pt-rGO-TiO2 was 95% (150\u00a0\u00b0C), and the CO2 yield was 72% under the infrared irradiation intensity of 116\u00a0mW/cm2 [45].In the studies of visible light irradiation, An et\u00a0al. proposed that CeO2 (ARCeO2) with OVs performed high photothermal catalytic performance of typical VOCs (50\u00a0ppm) including styrene (T90\u00a0=\u00a0226\u00a0\u00b0C), n-hexane (T90\u00a0=\u00a0459\u00a0\u00b0C) and cyclohexane (T90\u00a0=\u00a0563\u00a0\u00b0C). In addition, rich OVs and weak acid ARCeO2, together with the synergistic effect of photothermal catalysis, promoted the oxidation of intermediates, thus enhanced the coke resistance of high photothermal catalysis. The photothermal catalytic activity of ARCeO2 did not decrease significantly at 200\u00a0\u00b0C for 25\u00a0h. Its excellent photothermal catalytic stability could be attributed to fewer intermediates and limited coke accumulation on ARCeO2 [49]. Chen et\u00a0al. studied the removal of toluene (220\u00a0ppm) on ACo2O4 (A\u00a0=\u00a0Ni, Cu, Fe, Mn) spinel by collecting inexhaustible solar energy to provide thermal energy. The order of photothermal catalytic performance of ACo2O4 was as follows: NiO2O4\u00a0>\u00a0CuCo2O4\u00a0>\u00a0FeCo2O4\u00a0>\u00a0MnCo2O4. NiCo2O4 showed the highest photothermal catalytic activity at 214\u00a0\u00b0C (toluene conversion rate was 93%, CO2 production rate was 80%) and good toluene oxidation stability (at least 20\u00a0h). The excellent photocatalytic performance of NiCo2O4 was mainly due to its strong visible light absorption. It was found that visible light irradiation could enhance the mobility of reactive oxygen species, and thus significantly improved the photothermal catalytic activity of NiCo2O4 [106].Similarly, the full utilization of full-spectrum light is of great significance for the development of continuous photothermal catalytic systems. Chen et\u00a0al. reported that NiOx/Co3O4 composite exhibited high photothermal catalytic activity (about 95% conversion and 80% mineralization at 295\u00a0\u00b0C) in toluene oxidation (210\u00a0ppm, gas space velocity per hour\u00a0=\u00a032,000\u00a0mL\u00a0g\u22121\u00a0h\u22121) under simulated sunlight [58]. Cai et\u00a0al. evaluated the toluene degradation performance of Pt/\u03b3-Al2O3 under full solar spectrum irradiation. Pt/\u03b3-Al2O3 displayed tunable optical properties and outstanding photothermal conversion due to the plasma photothermal effect of Pt NPs. 1.81\u00a0wt% Pt/\u03b3-Al2O3 showed highly efficient catalytic activity with 87% toluene conversion and 84% CO2 yield at 165\u00a0\u00b0C with solar irradiation intensity of 320\u00a0mW/cm2, as well as a decent stable continuous operation for 30\u00a0h [63].Up to now, there are few studies on VOCs degradation by continuous photothermal system, and there is a lack of theoretical support from basic experimental data. Furthermore, the evaluation of catalyst performance mainly focused on single contaminant component. The coexistence should be expanded to include mixtures such as multicomponent VOCs, NOx, and water vapor, to better simulate the actual industrial processing conditions. In the continuous system, noble metal catalyst shows more excellent catalytic performance, but the current catalyst test conditions are low concentration, low space speed conditions, and the industrial conditions are quite different. Therefore, the study of noble metal catalyst in continuous VOC treatment system should focused on improving its stability and durability under high space velocity and high concentration conditions.At present, some reports have proposed the routes and mechanisms of photothermal catalytic VOCs with heterogeneous catalysts based on experimental and theoretical researches. Although abundant studies have been conducted on different types of VOCs and catalysts, the comprehensive oxidation mechanism of VOCs involved in the degradation process is rarely summarized in detail. Therefore, the mechanisms of photothermal catalytic VOCs were systematically summarized in this paper.Light-driven thermal catalysis based on photothermal effect is the most common route in photothermal catalysis. The catalyst converts the absorbed light into thermal energy to heat up the surface of the catalyst. It can provide enough energy to reach the combustion temperature of VOCs on the surface of the catalyst without the need for an auxiliary heat source [107]. The photothermal effect of catalysts mainly caused by surface plasmon resonance, non-radiative relaxation in semiconductor and molecules thermal vibration [161]. Light can be converted to heat by one or several photothermal conversion routes depending on the properties of the material. Fig.\u00a08\n provide a diagram of light-driven thermal catalysis. The catalyst surface temperature increases due to the photothermal conversion when thermal catalysis is driven by solar light. VOC oxidation reaction occurs when the temperature reaches the light-off temperature. The appropriate catalyst can efficiently absorb light and release it as heat energy, resulting in a temperature rise to activate the oxidation process. Localized heating effects caused by strong surface plasma absorption of noble metal nanoparticles have also been reported [44]. Noble metal nanoparticles rapidly heat up after absorbing visible light, the heat energy generated allows the reactants to cross the barrier needed for the reaction. The light intensity around noble metal nanoparticles and the interface of nanoparticles are stronger after the surface plasma absorbs light. The photons produced by the surface plasmon effect can be converted to heat energy, the area around the noble metal particles will be heated to a higher temperature. Zou et\u00a0al. [41] reported the photothermal catalysis of toluene by the composite catalyst Pd/CeO2. It was suggested that the enhanced catalytic activity was mainly attributed to the strong surface plasmonic resonance of Pd hot electrons. Under light irradiation, the hot electrons were transferred to CeO2 rapidly through the Pd-CeO2 interface; The service life of hot electrons was prolonged and the system heated up rapidly; In the meanwhile, the surface adsorption of oxygen and the activity of lattice oxygen were improved, which were favorable to the catalytic activity.The solar light-driven thermal catalysis reaction system has the following advantages:(1) No additional heat input is required and high energy efficiency; (2) Due to the photothermal effect, the surface temperature of the catalyst rises instantly, and the heat is mostly concentrated on the surface of the catalyst; (3) In some cases, the photothermal catalysis can effectively inhibit the deactivation of the catalyst and increase the selectivity of the desired product.Based on the photochemical effects, catalysts are excited by light irradiation to produce hot carriers (such as electrons and holes), which can participate in the reactions [162]. The energy of hot carriers generated by photochemical effects is much than that produced by thermal excitation. At the meanwhile, the hot carriers can be transferred to unoccupied molecular orbitals of adsorbate molecules, thus resulting photochemical transformation. In addition to the advantages of promoting electron transfer, light also inhibits the carrier recombination. Jiang et\u00a0al. found that abundant oxygen vacancies existed in the lattice of Ce1-xBixO2-\u03b4 catalyst with the introduction of Bi3+. The introduction of Bi3+ ions would expand the response to visible light and enhance the redox activity at low temperatures. At the same time, the vacancy greatly promoted the migration of oxygen ions, which further promoted the low-temperature integration. The migration of oxygen ions captured by holes effectively inhibited the recombination of charge carriers under UV-Vis and the thermal activation under infrared light. The high temperature caused by infrared would further enhance the coupled ion and electron conductivity [38].It is interesting to note that the photochemical effect has a positive effect on reducing carbon deposition [49], enhancing water resistance [113] and catalyst self-remediation in the reaction process [160]. Jiang et\u00a0al. proposed a self-healing mechanism based on photo-remediation during photothermal process (Fig.\u00a09\na). The reduction of high valent metal ions and the toxicity of intermediate products would deactivate the catalyst after a long period of thermal catalysis. However, photo-remediation could partially repair depleted catalysts: (1) Photo oxidized species mainly recover reduced metal ions; (2) Harmful reaction intermediates could be removed from the closed active site; (3) The photoinduced holes transformed the inert \u2013OH into strongly oxidizing \u2219OH, which further reduced the VOCs and released the active sites [160].In fact, it is difficult to completely distinguish the photothermal and photochemical catalytic pathways, which are intertwined in the photothermal process. The photothermal catalytic system can dissipate the absorbed photon energy into heat energy under incident light irradiation, which can promote the transfer of charge carriers and improve catalytic activity. At the meanwhile, hot carriers generated by light irradiation can participate in catalytic reactions. Thus, the reaction mechanism of heterogeneous photothermal catalysis is usually the photothermal and photochemical effects.Under photothermal conditions, irradiation can promote the activation of lattice oxygen, and thus improve its thermal catalytic performance. Ji et\u00a0al. proposed that in the photothermal degradation of toluene on Ag/F-SrTiO3, the high-energy electrons on Ag nanoparticles induced by visible light could be transferred to the adsorbed oxygen to assist the activation of O2 (Fig.\u00a09b). The transfer of activated oxygen species (e.g. O2\n\u2212) from noble metal nanoparticles to the carrier had a positive effect on the oxidation reaction. On the other hand, the electron transfer could be accelerated by increasing the reaction temperature, thus the photogenerated electron\u2013hole separation rate could be promoted. The heat in the photothermal catalysis process was beneficial to the production, transfer, and oxidation capacity of reactive oxygen species, which was essential for improving the photothermal catalysis rate and the deep oxidation of accumulated surface intermediates. Furthermore, the thermal effect induced by the surface plasmon resonance effect of Ag nanoparticles could further promote the surface reaction of Ag nanoparticles, which was helpful to the photocatalytic activity [46]. Another photothermal synergistic pathway is proposed based on the Mars-van Krevelen (MvK) theory. The MvK mechanism [163] refers to the reaction between the reactant and the lattice oxygen of the catalyst. The reaction diagram is shown in Fig.\u00a09c. In this reaction, the reactants are oxidized and the catalyst is reduced, and oxygen vacancies are produced. The OV is further replaced by adsorbed oxygen, which causes the catalyst to be re-oxidized. Li et\u00a0al. have pointed out that the photothermal synergistic catalytic degradation mechanism of benzene, toluene and acetone conforms to the MvK mechanism [89].In general, the mechanism of photothermal synergism is mainly reflected in the following aspects: (1) The absorption of solar light by the catalyst accelerates the rate of charge separation and the ability of reactive oxygen generation; (2) Under simulated solar irradiation, the generation of reactive oxygen is beneficial to speed up the MvK redox cycle and reduce the activation energy; (3) Improving the photocatalytic efficiency through thermal catalysis can promote the deep oxidation of coke and alleviate the negative effect of carbon deposition on photocatalysis [105]. The improvement of catalytic activity after irradiation is not a simple effect of photoinduced heating, but a more complex photothermal synergistic effect and the specific mechanism still needs to be further explored.Theoretical calculations based on density functional theory (DFT) are widely applied to study and predict catalytic reactions due to great advances in computational methods and related software [164]. Computational chemistry has become a powerful tool to explore degradation processes and mechanisms. The cooperation between experimental research and theoretical calculation can provide a deeper insight into the mechanism of catalysis [165,166]. At present, DFT calculation is often used to calculate the formation energy of OV (Evo) and the adsorption energy (Eads) of gas molecules on the surface of the catalyst.OV formation energy refers to the energy required for an oxygen atom removed from the lattice oxygen of metal oxide to form a defect. As mentioned above, the catalytic oxidation of metal oxides is generally believed to be carried out through the MvK mechanism. The organic molecules adsorbed on the surface of the catalyst are oxidized by lattice oxygen, and the oxygen vacancies are subsequently supplemented by gas-phase O2. The lattice oxygen activity of the catalyst under sunlight irradiation can be further studied by DFT theory calculation. Hou et\u00a0al. considered the effect of the substitution of K+ by Ce4+ in the tunnel on the OV through DFT calculation and calculated the Evo in OMS-2 supercell. For pure OMS-2 (Fig.\u00a010\na (i)), the Evo was 2.32\u00a0eV. When four K+ ions were replaced by one Ce4+ ion, the Evo increased to 3.52\u00a0eV (Fig.\u00a010a (ii)). The theoretical results showed that only the substitution of K+ ions by Ce4+ ions in the OMS-2 channel was not conducive to oxygen vacancy formation, thus reducing the catalytic activity of OMS-2. When four K+ ions were replaced by two Ce4+ ions to form a Mn4+ ion vacancy to maintain the charge balance (Fig.\u00a010a (iii)), the Evo in the lattice near the Mn ion vacancy decreased to 2.23\u00a0eV, suggesting the high lattice oxygen activity. The lattice oxygen activity and catalytic activity of OMS-2 could be significantly improved by designing the unique Ce ion substituted with Mn vacancy nanostructure [167]. The results also provided a reference for other catalysts.Effective adsorption is a necessary condition for the rapid decomposition of VOCs. DFT is also used in the photothermal catalysis process to calculate the adsorption energy of various gas molecules in the reaction process on the catalyst surface. Borjigin et\u00a0al. studied the adsorption of benzene on the surface of CeO2 and Ag3PO4 during the photothermal degradation over Ag/Ag3PO4/CeO2. The optimized structure and adsorption energy obtained are shown in Fig.\u00a010b. For the CeO2 (111) model, a shorter O-C bond length (3.737\u00a0\u00c5) and higher adsorption energy (\u22120.63\u00a0eV) were obtained between the benzene ring and CeO2 (111) (Fig.\u00a010b (i)). The electron transfer of C6H6 to the CeO2 surface was also confirmed by the charge density differences transection maps (Fig.\u00a010b (ii, iii)). In addition, the calculation results of the Ag3PO4 (100) model showed that a longer O-C bond (3.853\u00a0\u00c5) and smaller adsorption energy (\u22120.31\u00a0eV) was formed between Ag3PO4(100) and benzene (Fig.\u00a010b (iv)), and the electron transfer is not clear (Fig.\u00a010b (v, vi)). According to theoretical adsorption analysis, the adsorption capacity of the CeO2 (111) surface was better than that of the Ag3PO4 (100) surface. Therefore, it could be considered that the CeO2 component in the composite promoted the decomposition of VOCs through strong adsorption, so the pollutants were enriched to the catalyst surface [114].Photothermal catalysis aims to convert harmful compounds into harmless substances. However, there are still many challenges in reaction mechanisms and practical application.\n\n(1)\n\nReaction mechanism. In the current photothermal catalysis mechanisms, the solar light-driven thermal catalysis is essentially based on thermal catalysis, and light irradiation just plays the role of providing thermal energy. The reaction paths photocatalysis and thermal catalysis are different, it is important to define their differences in the photothermal synergistic process. In the current research on photothermal synergistic catalysis of VOCs, the performance\u2013temperature curves under light and dark conditions are measured to judge whether light or heat participate in the reaction. At present, the surface temperature of catalyst is measured by thermocouple. The selectivity and yield of catalysis products can be observed and compared at the same temperature to determine whether light plays a heating role in the reaction. However, the accuracy of surface temperature measurements for catalysts has been debated in recent years. As a macroscopic measurement method, thermocouple has limited application in the nanometer scale and cannot accurately reflect the localized temperature of nanoparticles. More accurate temperature measurements such as scanning thermal microscopy (SThM) [168], non-contact measurement contact technology based on Raman spectroscopy/in situ infrared spectroscopy [169,170] need more exploration. Therefore, the relationship between light and heat in photothermal catalysis and the synergistic mechanism needs to be further explored. More experiments and theoretical studies need to be carried out, such as product yield and selectivity under different temperatures and different light intensities, activation energy calculation, photon utilization rate calculation, and other core experimental data analysis.\n\n\n(2)\n\nCatalyst deactivation. In catalytic reactions, catalyst deactivation can be affected by the change of catalyst performance, intermediate products, or byproduct poisoning. For example, in the catalytic degradation process of benzene, the coke will deposit on the surface of the catalyst, resulting in the deactivation of the catalyst. The deactivation of the catalyst results in a significant decrease in the overall removal efficiency and the generation of intermediate by-products. In addition, the operating cost of the system will increase significantly when the catalyst is frequently regenerated or replaced. It is a feasible option to improve the performance of the catalyst since regeneration is difficult and uneconomic.\n\n\n(3)\n\nLack of performance data under actual working conditions. Most of the research on photothermal catalysis of VOCs has been conducted in the laboratory. VOCs concentrations are usually in the range of 50\u20132000\u00a0ppm (Tables 2 and 3). Although the highly stable performance of photothermal catalysts has been reported, further evaluation under practical conditions is necessary. At present, the data of photothermal catalytic VOCs are scanty and usually limited to conditions without moisture, so there is a certain disparity between this assessment and the actual working conditions. Moreover, the evaluations of performance mainly focused on individual components, and the coexistence atmosphere should be expanded to include multiple VOCs (especially containing chlorine VOC, sulfur VOC, etc), NOx and other mixtures, so the industrial conditions can be better simulated.\n\n\n(4)\n\nAccuracy of theoretical calculations. In the process of theoretical calculation, active catalytic species or intermediates have been difficult to be studied by experimental methods. However, there are still many challenges in studying the mechanism of catalytic transformation, such as the limitations of the accuracy of the computational methods. Furthermore, the large number of intermediate products involved in the catalytic process increases the complexity of theoretical calculation. In addition, the complexity of commonly used organometallic complexes and ligands, as well as the solvent and additive effects in the process, pose additional challenges.\n\n\n\nReaction mechanism. In the current photothermal catalysis mechanisms, the solar light-driven thermal catalysis is essentially based on thermal catalysis, and light irradiation just plays the role of providing thermal energy. The reaction paths photocatalysis and thermal catalysis are different, it is important to define their differences in the photothermal synergistic process. In the current research on photothermal synergistic catalysis of VOCs, the performance\u2013temperature curves under light and dark conditions are measured to judge whether light or heat participate in the reaction. At present, the surface temperature of catalyst is measured by thermocouple. The selectivity and yield of catalysis products can be observed and compared at the same temperature to determine whether light plays a heating role in the reaction. However, the accuracy of surface temperature measurements for catalysts has been debated in recent years. As a macroscopic measurement method, thermocouple has limited application in the nanometer scale and cannot accurately reflect the localized temperature of nanoparticles. More accurate temperature measurements such as scanning thermal microscopy (SThM) [168], non-contact measurement contact technology based on Raman spectroscopy/in situ infrared spectroscopy [169,170] need more exploration. Therefore, the relationship between light and heat in photothermal catalysis and the synergistic mechanism needs to be further explored. More experiments and theoretical studies need to be carried out, such as product yield and selectivity under different temperatures and different light intensities, activation energy calculation, photon utilization rate calculation, and other core experimental data analysis.\nCatalyst deactivation. In catalytic reactions, catalyst deactivation can be affected by the change of catalyst performance, intermediate products, or byproduct poisoning. For example, in the catalytic degradation process of benzene, the coke will deposit on the surface of the catalyst, resulting in the deactivation of the catalyst. The deactivation of the catalyst results in a significant decrease in the overall removal efficiency and the generation of intermediate by-products. In addition, the operating cost of the system will increase significantly when the catalyst is frequently regenerated or replaced. It is a feasible option to improve the performance of the catalyst since regeneration is difficult and uneconomic.\nLack of performance data under actual working conditions. Most of the research on photothermal catalysis of VOCs has been conducted in the laboratory. VOCs concentrations are usually in the range of 50\u20132000\u00a0ppm (Tables 2 and 3). Although the highly stable performance of photothermal catalysts has been reported, further evaluation under practical conditions is necessary. At present, the data of photothermal catalytic VOCs are scanty and usually limited to conditions without moisture, so there is a certain disparity between this assessment and the actual working conditions. Moreover, the evaluations of performance mainly focused on individual components, and the coexistence atmosphere should be expanded to include multiple VOCs (especially containing chlorine VOC, sulfur VOC, etc), NOx and other mixtures, so the industrial conditions can be better simulated.\nAccuracy of theoretical calculations. In the process of theoretical calculation, active catalytic species or intermediates have been difficult to be studied by experimental methods. However, there are still many challenges in studying the mechanism of catalytic transformation, such as the limitations of the accuracy of the computational methods. Furthermore, the large number of intermediate products involved in the catalytic process increases the complexity of theoretical calculation. In addition, the complexity of commonly used organometallic complexes and ligands, as well as the solvent and additive effects in the process, pose additional challenges.Photothermal catalysis is a promising and sustainable technology, which can significantly improve the catalytic activity and regulate the reaction pathway through the synergistic effect of photochemical and thermochemical process. Photothermal catalytic VOCs shows many advantages in solving the problems of high-energy consumption of traditional thermal catalytic oxidation technology and low efficiency of photocatalytic VOCs purification technology. This paper reviews the research progress of photothermal catalytic technology, including catalyst design, the performance over different systems and the main photothermal catalytic mechanisms. Meanwhile, the challenges and development prospect of VOCs photothermal elimination technology is illustrated. There are still some scientific and technical issues needs be solved, which mainly include the effective design of catalysts, the deep exploration of essential mechanisms of photothermal synergistic effect, the synergistic control of multiple pollutants, etc. It is hoped that this review can improve deeply understanding of photothermal catalytic VOC technology and provide some reference for practical application.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work is sponsored financially by the National Natural Science Foundation of China (No.21906104 and No.12175145), the Shanghai Rising-Star Program (21QA1406600).", "descript": "\n                  Photothermal catalysis realizes the synergistic effect of solar energy and thermochemistry, which also has the potential to improve the reaction rate and optimize the selectivity. In this review, the research progress of photothermal catalytic removal of volatile organic compounds (VOCs) by nano-catalysts in recent years is systematically reviewed. First, the fundamentals of photothermal catalysis and the fabrication of catalysts are described, and the design strategy of optimizing photothermal catalysis performance is proposed. Second, the performance for VOC degradation with photothermal catalysis is evaluated and compared for the batch and continuous systems. Particularly, the catalytic mechanism of VOC oxidation is systematically introduced based on experimental and theoretical study. Finally, the future limitations and challenges have been discussed, and potential research directions and priorities are highlighted. A broad view of recent photothermal catalyst fabrication, applications, challenges, and prospects can be systemically provided by this review.\n               "}
{"full_text": "\n1. Introduction\nRice husk is a lignocellulosic biomass consisting of cellulose, hemicellulose, and lignin [1]. Cellulose can be converted into energy sources in the form of platform chemicals such as 5-hydroxymethylfurfural (5-HMF) and levulinic acid (LA) [2]. Cellulose conversion to 5-HMF and LA has been widely carried out using homogeneous catalysts, such as sulfuric acid and hydrochloric acid [3,4], as well as heterogeneous catalysts such as Nafion, SAC-13 [4], Dowex, cellulase-mimetic solid acid catalysts [5], and ZSM-5 zeolites in media based on water [6,7]. Glucose conversion to 5-HMF and LA carried out using MOFs [8,9] have also been reported. However, not many articles discuss the conversion of cellobiose to LA, in example [10] or 5-HMF. From all heterogeneous catalysts, hierarchical ZSM-5 zeolite modified with Mn3O4 species has drawn much attention, due to its unique properties of large surface area, distinctive structure, and reactivity [6,7,11]. Since the hierarchical ZSM-5 catalyst possesses the hierarchical pore system i.e micro- and mesoporosity [12,13], it has high thermal stability and well-spread active sites as well as fast mass transport [14]. As we can see from the previous study on Table 1\n, Chen et al. [11], manage to obtained 1.17% HMF, Krisnandi et al. [6], obtained an LA % yield of 15.83%, and Pratama et al. [7], obtained an LA % yield of 39.75% with conventional heating method.However, the typical conversion reaction of biomass-based cellulose to 5-HMF and LA [6,7] usually is time-consuming although it has used a catalyst and medium heat. As mentioned above, the conversion of delignified rice husks into LA using hierarchical Mn3O4/ZSM-5 had an optimal yield of 39.75% at a reaction time of 8 h [7]. In another study, the conversion of cellulose to LA obtained optimal yield (91%) at 6 hours of reaction (using Ni-HMETS-10 catalyst, 180 \u00b0C) [15] and optimal yield (55%) at 3 hours of reaction (using C4(Mim)2(2HSO4)(H2SO4)2 catalyst, 100 \u00b0C) [16]. Also, most of the time, it is difficult to reuse the catalyst since it is difficult to be separated from the very thick product after 8h reaction and most of the catalysts were decomposed [6].Therefore, it has drawn our interest to use a microwave-assisted method in performing this conversion reaction to shorten the reaction time and to avoid severe damage to the catalyst. Many researchers are starting to carry out catalytic conversion using microwaves instead of conventional methods because the heat transfer process in conventional heating is relatively slow and less effective [11]. The microwave-assisted organic synthesis (MAOS) method is currently widely used because it gives a fast, easy, and clean pathway route [17]. Other advantages of MAOS include low energy, short time, high yield, solvent-free, and recyclability of catalyst. Reactions such as Aldol, Claisen, and Michael may benefit from using microwave assistance because of the high temperatures they require [17]. Muley et al. [18], stated that microwaves give a positive impact on heterogeneous catalysts properties. With microwaves, the polar reactant molecules have a faster rotation rate and enter the transition state more quickly than with conventional heating systems. Rapid microwave heating and higher temperatures can increase product selectivity, as in the case of micro-plasma formation at metal sites, microwaves lead to increased product selectivity and rapid activation of reactants such as methane [18].Ren et al. [19], carried out a microwave-assisted synthesis in SO3H-functionalized ionic liquids (SFILs) as an efficient catalyst for the direct conversion of cellulose to LA using laboratory microwaves. The yield of LA rose up with the reaction time increasing from 5 to 30 min and decreasing afterward. As a result, a 5 min at 160 oC reaction produces 15.7% LA (Table. 1) and the most optimal LA (55%) was obtained by reacting for 30 min with the same temperature. In addition, the microwave-assisted catalytic conversion of cellulose to 5-HMF in ionic liquid also showed a higher 5-HMF yield compared to when using the oil-bath method [20]. From those results, it can be suggested that microwave heating has several advantages including short reaction time, a high percentage of conversion, product yield, and selectivity.Microwave irradiation assistance also applies to green chemistry because the time and energy are lower. This is in line with the UN SDG goal number 7 which is access to affordable, reliable, sustainable, modern, and clean energy for all [14]. However, there are challenges to scaling up the reaction-assisted microwave, such as the reduced intensity (attenuation) of electromagnetic waves at small penetration depths of the most absorbent solvents [21]. To overcome this limitation, the combination of microwave irradiation with ultrasound waves can help produce an increased depth of penetration beyond intensifying mixing and mass transfer [21,22].Biomass conversion using catalysts can also play an important role in supporting the UN's SDGs. Biomass conversion can produce carbon-neutral and even carbon-negative energy [23\u201325]. This is in line with UN SDG number 12, ensuring sustainable consumption and production patterns [26].The aim of this comparative study is to carry out the biomass conversion to LA and 5-HMF using a household microwave as a heating method and compare it to the reaction using the conventional yet established thermal heating method using an oil bath. The catalyst used was hierarchical ZSM-5 zeolite, impregnated with Mn3O4 which has been the chosen catalyst and used to great extent in our previous work [6,7], and the composition in the reaction mixture also followed a similar reported procedure. To determine the optimal performance of the microwave-assisted catalytic conversion reaction of biomass to LA, microwave power and reaction time were varied. The results then are compared to those obtained using conventional heating methods. Furthermore, the catalyst was reused for three cycles to confirm its reusability.The substrate materials in this conversion are delignified cellulose, from lignocellulosic biomass of rice husks, cellobiose, and glucose. Recently, there has been much interest in using cellobiose as a cellulose model compound [27,28]. Cellobiose is the most thermodynamically stable subunit of crystalline cellulose [29]. Cellobiose is a bridge between monosaccharides and cellulose, which contains two D-glucose units linked by a \u03b2-(1,4)-glycosidic bond [29]. Therefore, cellobiose has a conversion pathway similar yet shorter to cellulose, so it is very interesting to use it to develop efficient and effective catalytic systems of lignocellulosic biomass.\n2. Materials and methods\n\n1.1. Materials\nDelignified cellulose (24.07% cellulose, 21.21% lignin, and 33.33% holocellulose) was prepared from local Indonesian rice husks following the previous method [7], and the result is available in SI. 1, Table S1. The chemicals used in this experiment were all of the analytical grades: phosphoric acid (89.0 wt %), hydrogen peroxide (30.0 wt %), sulfuric acid (96.0 wt %), ethanol (95.0 wt %), sodium hydroxide (99.0 wt %), manganese chloride (99.0 wt%), and glucose were obtained from Merck (Darmstadt, Germany). Cellobiose was obtained from TCI (Tokyo, Japan) and used directly without purification. Meanwhile, tetraethyl orthosilicate (TEOS, 98.0%), sodium aluminate (99.0%), tetrapropylammonium hydroxide (TPAOH, 1.0 M), and polyacrylamide-co-diallyl dimethylammonium chloride (PDDAM, 10 wt %) were obtained from Sigma Aldrich (St. MO, USA). All chemicals were used without further purification or treatment.\n1.2. Preparation of Hierarchical Mn3O4/ZSM-5\nHierarchical ZSM-5 zeolite synthesis was carried out following the procedure reported by Krisnandi et al. [6]. The mixture was prepared following the molar ratio of previous studies with a composition of 0.29 g NaAlO2, 27.15 g 98% tetraethyl orthosilicate, 25.94 g TPAOH 40%, 111.83 mL H2O, and 1.0 g PDDAM. All mixtures were put into a teflon-coated stainless steel autoclave with volume c.a.250 ml. The autoclave was then put into the oven at 170 oC for 144 h and after that, the precipitate was filtered to obtain ZSM-5 hierarchical white powder.The hierarchical Mn3O4/ZSM-5 catalyst was prepared using the wet impregnation method by spraying Mn(II) from 2 mL of 0.214 M MnCl2.4H2O solution onto 1.0 g of the hierarchical ZSM-5 to obtain 2 wt % of Mn. The mixture was stirred to form a paste, dried, and then calcined at 550 \u00b0C for 8 h to obtain hierarchical Mn3O4/ZSM-5.\n1.3. Characterization of the catalysts\nThe as-prepared hierarchical ZSM-5 and Mn3O4/ZSM-5 were characterized using several techniques. The Powder X-ray diffraction (XRD) patterns were investigated on PANanalytical: X\u2019Pert Pro 2318 diffractometer using Cu-K\u03b1 radiation (\u03bb =1.54184 \u00c5) as the incident beam with 2\u03b8 ranging from 5 to 50\u00b0. The Si/Al ratios were obtained by X-ray Fluorescence in XRD Orbis EDAX, 100 kV, 40 mA, and 100 scans. Analysis of functional groups on zeolites was carried out using an Alpha-Bruker FTIR spectrometer, by measuring the KBr-pellet of the sample (20:1) with 128 scans at 4000-400 range of wavenumber. Surface area analysis was performed in Quantachrom-Evo Surface Area and Pore Analyzer instrumentation, in which the pore size distribution was determined using Barrett-Joyner-Halenda (BJH) desorption curve and Horvath Kawazoe (H-K) plot methods. The morphology and mesopores information of the catalysts was analyzed on Jeol JIB-4610F Scanning Electron Microscopy with Energy Dispersive Spectroscopy (SEM-EDS), while the Mn content was seen using the Atomic Absorption Spectrometer (AAS) Shimadzu AA-625-0.\n1.4. Conversion of biomass to LA\n\nConventional heating method\n[6,7]\n. The conversion of the substrate was carried out by adding 0.5 g substrate (delignified rice husk, cellobiose, or glucose), 0.05 g of hierarchical Mn3O4/ZSM-5 catalyst, 10 ml H3PO4 (40%, v/v), and 2 drops of H2O2 (30%, v/v) into a 20 mL-vial glass which then was immersed into oil-bath at 130 \u00b0C. After 2, 4, 6, and 8 h reaction time, the sample was immediately cooled in an ice bath to quench the reaction, then the filtrate was separated from the solid for product analysis.\nMicrowave-assisted method. For conversions using a household microwave (Samsung ME731K), the same mixture was placed into a 50-mL porcelain crucible and heated using 300 W, 450 W, and 600 W of power for 60 s. Prior to the reaction, the microwave heat was calibrated (the data is available in the SI.2, Table S2). Then, further reactions were performed at 30 s and 180 s reaction times, with the power that gave the best results. As a control, a reaction without a catalyst was also carried out.To evaluate the reusability and stability of the Mn3O4/ZSM-5 catalyst, the conversion of glucose as the model was carried out by adding 0.5 g glucose, 0.05 g of the catalyst, 10 ml H3PO4 (40%, v/v), and 2 drops of H2O2 (30%, v/v) into a 50-mL porcelain crucible using a microwave-assisted method and heated using 600 W of power for 180 s. The used Mn3O4/ZSM-5 catalyst was collected by filtration, washed thoroughly with deionized water to attain neutral pH, and dried at 75 \u00b0C for 30 minutes. The dried catalyst sample was then calcined at 550 \u00b0C for 3 hours. The recycled catalyst was reused according to the conversion experiment mentioned above until three times cycles. The products were analyzed using HPLC to determine the yield of LA. The used catalyst was characterized with FTIR, XRD, and SEM-EDS after four times use.Product analysis from the conversion reaction was carried out using High-Performance Liquid Chromatography (HPLC) at Welizer LC200 PG Instrument, and Nuclear Magnetic Resonance (NMR) at Agilent 500 MHz NMR spectrometer with DD2 console system.\nHPLC measurements were carried out to see the results of conversion products. Analysis was performed using HPLC Welizer LC200 Quarantine Isocratic Pump with UV detector. The column used is ODS with 0.1% HClO4 as eluent. The final eluent rate was adjusted to 1.0 mL/min at 40\u02daC. The wavelength used is 285 nm to identify compounds such as LA and 5-HMF with a retention time of 15 min. The conversion and yield of products were calculated using the equation (1) and (2). Details on the calculation to determine % conversion and % yield are available in the SI.3.\n\n(1)\n\n\n%\n\nc\no\nn\nv\ne\nr\ns\ni\no\nn\n\n=\n\n\n\ni\nn\ni\nt\ni\na\nl\ns\nu\nb\ns\nt\nr\na\nt\ne\nm\na\ns\ns\n\n\ng\n\n\n-\nr\ne\ns\ni\nd\nu\na\nl\nm\na\ns\ns\n\n\ng\n\n\n\n\ni\nn\ni\nt\ni\na\nl\ns\nu\nb\ns\nt\nr\na\nt\ne\nm\na\ns\ns\n\n\ng\n\n\n\n\nx\n\n100\n%\n\n\n\n\n\n\n(2)\n\n\n%\n\ny\ni\ne\nl\nd\n\nL\nA\n\no\nr\n\n5\n-\nH\nM\nF\n=\n\n\nt\no\nt\na\nl\nc\no\nn\nc\ne\nn\nt\nr\na\nt\ni\no\nn\n\n\n\n\nmg\n\nL\n\n\n\nL\nA\no\nr\n5\n-\nH\nM\nF\nX\ns\no\nl\nu\nt\ni\no\nn\nv\no\nl\nu\nm\ne\n\n(\nL\n)\n\n\n\nsubstratemass\n(\nm\ng\n)\n\n\nx\n\n100\n%\n\n\n\n\n\nNMR analysis. Prior to the NMR test, the organic phase of the product was extracted from the aqueous phase using 8 mL of ethyl acetate [30]. The separated organic phase was then added with MgSO4 to remove the remaining water [30]. Furthermore, the ethyl acetate was evaporated from the organic phase using a rotary evaporator at a temperature of 60 oC for 10 minutes. The result is a brown viscous liquid, this liquid is then analyzed using 1H-NMR and 13C-NMR. NMR characterization used Agilent 500 MHz NMR for 1H and 125 MHz for 13C with DD2 console system and D2O solvent.\n3. Result and discussion\nIn this work, a hierarchical ZSM-5 zeolite was synthesized and then modified with Mn3O4 2.14 wt% to reconstruct the best catalyst that was used in the conversion of delignified cellulose to LA [7]. For brevity, the characterization results from FT-IR analysis are available in SI. 4, Fig S2, while others are discussed below.\nXRD. Fig. 1\n shows the characteristic powder XRD pattern of the hierarchical ZSM-5 catalyst that is similar to the previous results [13] and having high crystallinity [6]. In addition, the existence of Mn3O4 in the ZSM-5 catalysts was shown by a peak in 30\u201335o and JCPDS: 24-0734 [31]. The structure of the catalyst did not significantly alter after being added with Mn3O4.\nSEM. Fig. 1 also shows that the as-synthesized hierarchical ZSM-5 forms a coffin-like shape, a typical MFI morphology [6], and after the impregnation process the shape is unchanged, which indicates that the impregnation does not damage the structure.\nSAA. Analysis of the surface area and pore size of the catalyst was presented in Fig. 2\n. The adsorption-desorption isotherm curves (Fig. 2 (a)) contain hysteresis loops in the range 0.4-0.9 P/Po, indicating the presence of mesoporosity, although the size pore distribution, calculated from BJH desorption curve (Fig. 2 (b)) and H-K plot (Fig. 2 (c)) show that the predominant pores are micropores that have an average pore diameter of 1.84 nm. However, it was also seen in Fig. 2 (b), that there was a mesopore distribution present at an average of 2.7 nm. Therefore, this as-synthesized ZSM-5 is labeled a hierarchical ZSM-5 because there are two types of pores that are micro (d < 2 nm) and meso (2 nm < d < 50 nm) in one structure [32]. Table 2\n summarizes the results of SAA of both hierarchical ZSM-5 and Mn3O4/ZSM-5.Conversion of biomass to LA has been thoroughly studied [6,7,11]. This conversion reaction uses phosphoric acid solvent (H3PO4 40% (v/v)) which serves to break the inter- and intramolecular hydrogen bonds belonging to cellulose [11]. In addition, 30% H2O2 solvent is also added that will react with the Mn2+ and Mn3+ from Mn3O4/ZSM-5 through a Fenton-like reaction as shown by equations (3) and (4) as reported by Pratama et al. [7].Mn2+ + H2O2 \u2192 Mn3+ + HO\u2022 + OH\u2212(3)Mn3+ + H2O2 \u2192 Mn2+ + HO2\u2022 + H+(4)The reaction in equation (3) will produce hydroxyl radicals to break the \u03b2-(1,4)-glycosidic bonds of cellulose and cellobiose so that they can be degraded into glucose [11]. The isomerized glucose will then be converted into LA and by-products such as formic acid with an intermediate compound in the form of 5-HMF [33]. General reaction scheme of the conversion of the cellulosic compounds to 5-HMF and Levulinic acid is depicted in Figure 3\n, while the detailed proposed mechanism available in SI.5.\nConventional heated reaction. Reaction conversion was carried out with the conventional method, equipped with an oil bath as a heating source. The results are displayed in Fig. 4\n, the current experiment was tested using delignified cellulose [6,7], glucose, and cellobiose as the substrate, and the results were obtained from measurements using HPLC and gave the same trend as shown in Fig. 4. From the graph we can see that for all substrates the longer the reaction time, the percentage of conversion tends to be higher. There was a slight decrease during the 8-h reaction although it was not significant because of the decrease in catalyst activity after the reaction lasted for 8 h. This result is consistent with the results reported in the previous studies [6,7], that using hierarchical Mn3O4/ZSM-5 as catalysts, it took quite some time to reach a high yield of LA. During the reaction time, the yield of 5-HMF is decreased simultaneously, following the increase of the yield of LA, because 5-HMF is an intermediate product that will be converted to LA (illustration in Fig. 3).\nMicrowave-assisted reaction. With the purpose of the reaction time, biomass conversion was carried out by using a household microwave. First, to find out the most optimal power required, the power used was varied at 300, 450, and 600 W, and the conversion was carried out for 60 s, using the three substrates, and the results are shown in Fig 5\n. When reaction control using a reaction mixture without Mn3O4/ZSM-5 catalyst was carried out, thick black charcoal was produced so it is difficult to identify and quantify (for brevity the results are not shown). This confirms that the reaction using the microwave also requires a heterogeneous catalyst. Fig. 5 (a) shows an increasing trend of % conversion along with the increment of power used. The best percentage yield is obtained when 600 W of power is applied. Therefore, in the next experiments, the power used was set to 600 W. Second, the reaction time was set to 30 s and 180 s (Fig. 5 (b)). It can be seen that the conversion continues to increase with a longer reaction time. However, the conversion for 240 and 300 s resulted in a charred solution so that the LA yield could not be determined. This is because household microwaves have limitations, there is no temperature regulation as well as stirring.The highest yield for LA was obtained when using a 600 W microwave for 180 s. Glucose gave the best percentage yield of 9.57%, followed by cellobiose at 6.12% and delignified cellulose at 4.33%. This shows that the longer the reaction time used, the more LA is produced. Interestingly, the percentage yield of the 5-HMF intermediate was rather low given by the three substrates (for brevity, the results are available in SI.3, Table S6). Glucose has the highest percentage yield because it has a shorter pathway for conversion to 5-HMF and LA. Meanwhile, delignified cellulose and cellobiose must face the \u03b2-(1,4)-glycosidic bond-breaking step first, as shown in Fig 3. In addition, delignified cellulose still contains some retained lignin which may inhibit the conversion process so delignified cellulose has the lowest percentage yield of LA, compared to cellobiose and glucose.Further look at the results, it shows that the percentage yield of LA from biomass conversion using the microwave-assisted reaction in 180 s, 600 W (Fig 5.(b)) was comparable with those obtained by using the conventional heating method at 130 oC for 4 h. This shows that the microwave method has the potential to convert in a short time so it is more efficient to use compared to conventional methods. The HPLC chromatograms of the microwave-assisted reaction in 180 s and the conventional heating at 130 oC for 4 h were compared (available in SI.6, Fig. S6), and it can be seen that, besides LA and 5-HMF, there are peaks that correspond to by-products such as formic acid [34]. However, the intensity of the peak from the microwave method is weaker compared to that of conventional methods. Furthermore, the 1H and 13C NMR measurements on the isolated LA also show that the product obtained with the microwave assisted method has higher purity than that of the conventional heating method. The analysis by NMR (SI 6., Fig. S7, S8) showed the presence of by-products in the form of acetic acid and formic acid [35,36] . The intensity of by-product converted using a microwave was lower than the conversion using conventional methods.As depicted in Fig. 6\n (a), the glucose conversion was still maintained at more than 45% after 4 times of use (3 cycles). Furthermore, the weight loss of the catalyst is illustrated in Fig. 4 (a), which was around thirty percent at the first use until after the fourth use. In addition, for the first cycle, the yield of LA as a product was 10.57% and decreased gradually to 7.00% after the 2rd cycle then tends to stay on the 3th cycle. The results confirmed that the catalyst had good reusability. The above analysis results show that Mn3O4/ZSM-5 has sufficient stability in the conversion of glucose to LA.To find out the properties of the reused catalyst, some characterization was carried out, i.e FTIR, powder XRD and SEM-EDS. As shown in Fig. 6 (b), the FTIR spectra of Mn3O4/ZSM-5 before and after use,. the intensity of the absorption band at 3500 cm-1, which is the absorption band of the silanol group (Si-OH) on the zeolite surface, was increased, after the catalyst was used and calcined several times. Furthermore, the spectra at 700-1250 cm-1 that was attributed to Si-O and Al-O vibrations of the aluminosilicate framework [38] remain. This suggests that the reuse of Mn3O4/ZSM-5 catalyst did not cause any damage to the zeolite framework. This was also supported by the PXRD pattern of Mn3O4/ZSM-5 (Fig. 6 (c)) after four times of use. The catalysts still had sharp PXRD diffraction peaks, and no new peaks appeared in contrast to the fresh catalysts, with the characteristics of the Mn peak remaining at the same intensity and position with no shifting compared to the fresh catalyst. It confirmed the structural stability of Mn3O4/ZSM-5 under catalytic reaction in the microwave-assisted method. The SEM images of the used catalysts in Fig. 6 (d) shows the after used catalysts which is similar to the Mn3O4/ZSM-5 catalyst before reaction. Furthermore, the Si/Al ratio of the used catalyst decreases to 28.4 and the Mn content decreases to 1.6%. This provided evidence that Mn (II) was removed from the Mn3O4/ZSM-5 structure during the reaction. This procedure suggests that the interaction of solid Mn3O4 as the active site in ZSM-5 and aqueous Mn2+ ions in the solution with H2O2 to create HO\u2022 radicals in a Fenton-like system took place in conversion using Mn3O4/ZSM-5 zeolites as a catalyst. This indicates that the reusability test of the catalyst in the microwave system does not damage the structure, however the active species of the catalyst such as Mn content and the Si/Al ratio on zeolite surface were calculated decrease because it has been used 4 times cycles reaction. The last cycle reaction shows results that are comparable with the 4 h reaction using the conventional heating methods in our previous work [6] identified that the EDX characterization for after used catalyst was (Si 12.28%, Al 0.11%, Mn 0.69%).This catalyst reusability test under the microwave-assisted system shows its excellence as confirmed by the overall conversion results, and its selectivity to LA (proof by HPLC and NMR measurement). In microwave synthesis, it has also been observed that the relatively lengthy heating and times result in an equivalent reaction temperature than attained for oil bath or conventional heating [37]. In addition, the stability of the catalyst under optimal conditions in this system is remarkable compared to that of the catalysts used in conventional heating systems as reported in our previous studies [6], in which the catalyst has undergone structural changes since the reaction started at 100 oC at 0 h, and removal of the active sites, Mn3O4, since 3 h reaction time.\n4. Conclusion\nThe hierarchical Mn3O4/ZSM-5 catalyst was successfully synthesized and used in the conversion of biomass to LA. It is confirmed that the conventional heating method at 130 oC for 8 h reaction time gives the highest conversion and LA yield. On the other hand, the reaction with a 600 W microwave for 180 s shows results that are comparable with the 4 h reaction using the conventional heating methods, with better purity of LA. In addition, the used catalyst after reactivation could be applied for 3 cycles of reaction without losing too much of its activity. This shows that the conversion using a microwave has the potential to be explored in the future to achieve cleaner reaction conditions in a very short reaction time.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was funded by BOPTN Research Fund number NKB-322/UN2.RST/HKP.05.00/2021 from the Ministry of Research and Technology Indonesia.", "descript": "\n                  Conversion of delignified cellulose from rice husk biomass, and model compounds of cellobiose and glucose to levulinic acid (LA) over hierarchical Mn3O4/ZSM-5 catalyst was carried out using a household microwave method, and then compared to the established conventional thermos-reaction method. The hierarchical ZSM-5 was prepared using a double template method, aiming for micro and mesoporous systems developed in the structure. The as-prepared ZSM-5 were modified with Mn3O4 through incipient wetness impregnation with Mn2+ solution followed by calcination at 550 oC. The catalysts were characterized using various techniques such as powder XRD, SEM, BET, AAS, and FT-IR which indicated the hierarchical structure of MFI zeolite (Si/Al of 30-34) with Mn loading of 2.14 wt%. The conversion products were analyzed using HPLC, 1H-NMR, and 13C-NMR instruments. The microwave-assisted reaction using 600 W for 180 s using delignified cellulose, cellobiose, and glucose gave conversion of 37.27%, 46.35%, and 54.29%, respectively which is close to the conversion given by the conventional reaction carried out at 130 oC for 4 h (36.75%, 55.62%, and 60.9%, respectively). Interestingly, the LA yield from the microwave-assisted reaction (4.33 %, 6.12 %, and 9.57 %) is higher than the yield from the conventional reaction, which only produced 5.2%, 4.88%, and 6.93% respectively. The microwave-assisted method is also shown to give less by-products compared to the thermochemical reaction. Therefore, it could be considered an alternative method for converting cellulose to LA.\n               "}
{"full_text": "With the increasing energy consumption and the challenge of the global climate problem, it is urgent to develop novel and efficient energy conversion and storage technologies [1\u20133]. In recent years, a lot of electrochemical technologies, which convert sustainable electric energy into chemical energy, have been developed. The typical electrochemical systems include water splitting [4\u20136], carbon dioxide reduction (CO2R) [7,8], nitrate reduction reaction [9,10], fuel cells [11], and metal-air batteries [12,13]. Oxygen evolution reaction (OER) is a key half-reaction in these technologies [14,15]. However, OER is a complex four-proton-coupled electron transfer process, and a large overpotential is required due to its sluggish kinetics [16\u201319]. Therefore, it is of great significance to develop an efficient catalyst with high durability and catalytic activity to reduce the overpotential of OER. Currently, commercial OER catalysts are Ir/Ru-based compounds, but the scarcity and high cost limit their industrial applications [4,20].Earth-abundant transition metals and their compounds show great potential as OER catalysts due to their decent performance and low cost [21\u201325]. Among them, Ni-based catalysts have attracted extensive attention [26,27]. However, these catalysts are still far from commercial applications due to their unsatisfied activity and poor stability. High-valent Ni compounds show much better OER performance, which are considered to be the active phase [28\u201332]. Recently, doping strategies have been used to promote the formation of high-valent Ni sites [30,33\u201338]. High-valence metallic dopants, such as Mo and V, do not exhibit satisfactory OER activity but can induce electronic effect to improve the catalytic activity of Ni-based catalysts [39]. Besides, Fe dopants have been widely used as a critical role in enhancing the OER activity of Ni-based catalysts [40]. Therefore, the doping strategy is a promising way to develop highly efficient OER catalysts. However, due to the complexity of the chemical structure, designing catalysts with multiple metallic dopants is still a great challenge.Based on Fe doped Ni catalysts (Fe\u2013Ni nanoparticles), we introduce multiple metallic dopants using a simple oil phase strategy in this study. The Fe\u2013Ni nanoparticles, with the other 3 dopants of Mn, Mo, and V, show an overpotential of 220\u00a0\u200bmV at 10\u00a0\u200bmA\u00a0\u200bcm\u22122 and a 200\u00a0\u200bh long-term electrochemical stability. Experimental results show that Mo and V dopants introduce high-valence Ni sites and promote the formation of NiOOH at a lower potential to enhance their OER performance. Meanwhile, Mn dopants increase the electrochemical active surface area (ECSA) and exposed active sites, further boosting the OER activity.Iron (III) acetylacetonate (Fe(acac)3, 97%) and (1-Hexadecyl) trimethylammonium chloride (CTAC, 96%) were purchased from Shanghai Aladdin Reagent Co., Ltd. Nafion solution (5\u00a0\u200bwt%) was supplied by Sigma-Aldrich. Molybdenum hexacarbonyl (Mo(CO)6, 98%), manganese(II) acetylacetonate (Mn(acac)3, 97%), vanadium(IV) oxy acetylacetonate (VO(acac)2, 95%), oleylamine (OAm, >70%), molybdenyl acetylacetonate (MoO2(acac)2, 95%), ruthenium oxide (RuO2, 99.95%) and glucose were bought from Heowns Biochem Technologies, LLC, Tianjin. Nickel (II) acetylacetonate hydrate (Ni(acac)2\u00b72H2O, 95%) was supplied by Meryer (Shanghai) Chemical Technology Co., Ltd. Other reagents were purchased from Tianjin Yuanli Chemical Co., Ltd.CTAC (0.5\u00a0\u200bg) was added into oleylamine (50\u00a0\u200bmL) in a 100\u00a0\u200bmL three-necked flask. After sonication for 30\u00a0\u200bmin, Fe(acac)3 (10\u00a0\u200bmg), Ni(acac)2\u00b72H2O (6.4\u00a0\u200bmg), Mn(acac)3 (8.9\u00a0\u200bmg), MoO2(acac)2 (8.8\u00a0\u200bmg), VO(acac)2 (6.5\u00a0\u200bmg), glucose (0.6\u00a0\u200bg), and Mo(CO)6 (165\u00a0\u200bmg) were added into the three-necked flask. And the mixture was sonicated for 30\u00a0\u200bmin and heated to 60\u00a0\u200b\u00b0C to obtain a homogeneous solution. The solution was heated to 220\u00a0\u200b\u00b0C and then kept for 2\u00a0\u200bh under magnetic stirring. The products were collected by centrifugation, then washed three times with an ethanol/cyclohexane mixture, and twice with 0.5\u00a0\u200bmol/L acetic acids (ethanol solution). Finally, the products were dried in the vacuum oven overnight. The other multiple metal doped nickel nanoparticles were synthesized by the same method.The catalyst ink was prepared by mixing 10\u00a0\u200bmg prepared products, 5\u00a0\u200bmg carbon black, 500\u00a0\u200b\u03bcL isopropanol, 500\u00a0\u200b\u03bcL Milli-Q ultrapure water, and 20\u00a0\u200b\u03bcL 5% Nafion solution for at least 30\u00a0\u200bmin. The homogeneous ink was dropped on a nickel foam and dried at room temperature to form a 0.25\u00a0\u200bcm2 effective catalytic area. The catalyst loading was about 8\u00a0\u200bmg/cm2.The morphologies and elemental distribution of the prepared samples were characterized by field emission scanning electron microscopy (SEM, Hitachi S-4800) and transmission electron microscopy (TEM, JEM-2100F) equipped with an energy-dispersive spectroscopy (EDS) detector. The crystalline structure of the catalysts was investigated by the powder X-ray diffractometer (XRD, Bruker D8) with Cu K\u03b1 radiation. X-ray photoelectron spectrometer (XPS, Thermo ESCALAB 250XI) was used to analyze the surface elemental composition and chemical states.All electrochemical tests were carried out at room temperature with a standard three-electrode system using an electrochemical workstation (Autolab PGSTAT302N). The nickel foam with catalyst, Ag/AgCl electrode with saturated KCl as the filling solution, and platinum foil were used as the working electrode, reference electrode, and counter electrode, respectively. The electrolyte was 1\u00a0\u200bM KOH solution (pH\u00a0\u200b=\u00a0\u200b14). All applied potentials against the Ag/AgCl reference were calibrated to the reversible hydrogen electrode (RHE) through the equation:\n\n\n\n\nE\n\nR\nH\nE\n\n\n=\n\nE\n\nA\ng\n/\nA\ng\nC\nl\n\n\n+\n0.197\n+\n0.0591\n\u00d7\np\nH\n\n\n\n\nLinear sweep voltammetry (LSV) was performed in the potential range from 1.0 to 1.8\u00a0\u200bV (vs RHE) at a scan rate of 1\u00a0\u200bmV\u00a0\u200bs\u22121 to evaluate the OER activity. Before LSV measurement, 20 cycles of cyclic voltammetry (CV) scans were carried out in the potential range from 1 to 1.8\u00a0\u200bV (vs RHE) at a sweep rate of 20\u00a0\u200bmV\u00a0\u200bs\u22121. Tafel slopes were derived from the linear region of the LSV polarization curves. To correct the i-R drop and analyze the electrode kinetics during the reaction, electrochemical impedance spectroscopy (EIS) measurement was performed in frequencies ranging from 0.1\u00a0\u200bHz to 100\u00a0\u200bkHz with an AC amplitude of 5\u00a0\u200bmV. The electrochemical double-layer capacitance (C\ndl) was carried out to determine ECSA. CV scans were carried out in a non-Faradic potential region at various scan rates ranging from 20 to 120\u00a0\u200bmV\u00a0\u200bs\u22121 at a potential window of 0.92\u20131.02\u00a0\u200bV (vs RHE). The long-term durability of the catalysts was evaluated by a chronopotentiometry experiment at 1\u00a0\u200bA\u00a0\u200bcm\u22122. The cyclic stability was conducted by 2000 cycles of CV measurements in the potential range from 1.0 to 1.8\u00a0\u200bV (vs RHE). All the potentials were corrected by 85% iR correction.The FeMnMoV\u2013Ni catalysts were synthesized via a simple oil phase strategy, as shown in Fig.\u00a01\na. Fe(acac)3, Ni(acac)2, Mn(acac)2, MoO2(acac)2, and VO(acac)3 were dissolved into oleylamine. At 220\u00a0\u200b\u00b0C, all metal precursors were reduced. Various catalysts, such as FeMoV\u2013Ni, FeMnV\u2013Ni, FeMnMo\u2013Ni, and Fe\u2013Ni, were synthesized by the same method. The XRD patterns of synthesized catalysts are shown in Fig.\u00a01b. The peaks at 44.5\u00b0, 51.9\u00b0, and 66.5\u00b0 are related to Ni (111), (200), and (220) planes (PDF#89\u20137128) [41,42]. The high-resolution transmission electron microscopy (HRTEM) image shows lattice fringes with an interplanar distance of 0.208\u00a0\u200bnm related to the Ni (111) plane. The selected area electron diffraction (SAED) pattern (Fig.\u00a0S1) of the catalyst displays (111), (002), (022), (222), and (133) diffraction rings of Ni, which suggests that the main phase of the catalyst is Ni phase. As shown in Fig.\u00a01c and Fig.\u00a0S2, TEM images and SEM images illustrate that the obtained products are irregular particles with an average size of \u223c105\u00a0\u200bnm. The EDS elemental mapping images (Fig.\u00a01e, S3) show that Ni, Fe, Mn, Mo, and V are uniformly distributed in the products, which indicates that the catalyst was multiple metal doped nickel nanoparticles.The OER performance of multiple metal doped nickel nanoparticles was evaluated by CV and LSV measurements on a typical three-electrode configuration in 1\u00a0\u200bM KOH solution. The LSV polarization curves for all samples are shown in Fig.\u00a02\na, and Fig.\u00a02b displays the overpotentials of all catalysts. The FeMnMoV\u2013Ni catalyst exhibits the best OER performance with an overpotential of 220\u00a0\u200bmV at 10\u00a0\u200bmA\u00a0\u200bcm\u22122 current density (\u03b7\n10\u00a0\u200b=\u00a0\u200b220\u00a0\u200bmV), which is superior to that of FeMoV\u2013Ni (\u03b7\n10\u00a0\u200b=\u00a0\u200b240\u00a0\u200bmV), RuO2 (\u03b7\n10\u00a0\u200b=\u00a0\u200b251\u00a0\u200bmV), FeMnV\u2013Ni (\u03b7\n10\u00a0\u200b=\u00a0\u200b287\u00a0\u200bmV), FeMnMo\u2013Ni (\u03b7\n10\u00a0\u200b=\u00a0\u200b295\u00a0\u200bmV), and FeNi (\u03b7\n10\u00a0\u200b=\u00a0\u200b363\u00a0\u200bmV). Meanwhile, the Tafel slope (Fig.\u00a02c) of FeMnMoV\u2013Ni is 50.9 mV/dec, which is also lower than that of FeMoV\u2013Ni (59.2 mV/dec), RuO2 (83.2 mV/dec), FeMnV\u2013Ni (87.6 mV/dec), FeMnMo\u2013Ni (88.1 mV/dec), and Fe\u2013Ni (106.6 mV/dec). These results indicate that the FeMnMoV\u2013Ni exhibits faster kinetics and excellent catalytic activity for OER. EIS measurements were performed to evaluate the electrode kinetics for all catalysts during the OER process [43], and the results are shown in Fig.\u00a02d. The semicircles in the EIS spectra represent the charge-transfer resistance (R\nct) at the electrode/electrolyte interfaces. The R\nct of FeMnMoV\u2013Ni (4.7\u00a0\u200b\u03a9) is lower than that of the other samples, demonstrating a rapid charge transfer during OER. Moreover, ECSA was assessed to evaluate the intrinsic catalytic activities of all catalysts. As shown in Fig.\u00a0S4, the C\ndl value of the FeMnMoV\u2013Ni catalyst is 8.8\u00a0\u200bmF\u00a0\u200bcm\u22122, which is the highest among all samples. The results indicate that Mn, Mo, and V dopants increase the ECSA and then promote OER performance. In addition, the LSV curves normalized by ECSA are presented in Fig.\u00a0S5, which represent the intrinsic OER activity of catalysts. It shows that FeMnMoV\u2013Ni and FeMoV\u2013Ni exhibit similar current densities, suggesting that Mn dopants do not promote the intrinsic OER activity. And FeMnV\u2013Ni and FeMnMo\u2013Ni also show similar standardized current density at the same overpotential. Compared with FeMnV\u2013Ni and FeMnMo\u2013Ni, FeMnMoV\u2013Ni exhibits a lower overpotential at the normalized current density of 0.2\u00a0\u200bmA\u00a0\u200bcm\u22122, indicating that the Mo and V dopants increase the intrinsic OER activity. According to the ECSA results, Mn dopants only increase the ECSA, while Mo and V increased both the ECSA as well as the intrinsic OER activity.High stability is important for practical applications. Fig.\u00a02e exhibits the LSV polarization curves before and after 2000 cycles of CV measurements with a scan rate of 50\u00a0\u200bmV\u00a0\u200bs\u22121. There is no obvious difference between these two curves, indicating the superior durability of the FeMnMoV\u2013Ni catalyst. Moreover, multiple current steps for the chronopotentiometry test are used to evaluate the stability of the catalyst. From Fig.\u00a0S6, the current density increased from 5 to 50\u00a0\u200bmA\u00a0\u200bcm\u22122 with a step of 5\u00a0\u200bmA\u00a0\u200bcm\u22122 per 500\u00a0\u200bs. Furthermore, a chronopotentiometry measurement was performed to evaluate electrocatalytic stability. As shown in Fig.\u00a02f, a long-term test was employed at 1\u00a0\u200bA\u00a0\u200bcm\u22122 over 250\u00a0\u200bh, and there is no significant increase in voltage, suggesting excellent electrocatalytic stability for the FeMnMoV\u2013Ni catalyst. Table\u00a01\n presents a comparison with previous reported transition metal doped catalysts, demonstrating the remarkable OER performance of the FeMnMoV\u2013Ni catalyst.We performed in situ Raman to identify the active site in the catalyst and confirm the promotion of different elemental dopants for the formation of active species. As shown in Fig.\u00a03\n, the in situ Raman spectra were recorded at applied potentials ranging from 1.1 V to 1.6\u00a0\u200bV vs RHE. At low applied potential, there are no obvious Raman peaks. With the increase of applied potential, two peaks located at \u223c476\u00a0\u200bcm\u22121 and 555\u00a0\u200bcm\u22121, which are attributed to the Ni\u2013O bond vibration in NiOOH [48,50,51], appeared and strengthened. Raman spectra indicate that the Ni phase transformed into NiOOH during the OER process. Noted that no other MOOH peaks appear, NiOOH is expected to be the real active site. The characteristic peaks of NiOOH first appear in the FeMnMoV\u2013Ni and FeMoV\u2013Ni at 1.3\u00a0\u200bV vs RHE and further strengthen with the increasing potential. The peaks of the FeMnMoV\u2013Ni catalyst are significantly stronger than that of the FeMoV\u2013Ni, indicating that Mn dopants facilitate the formation of NiOOH species. In comparison, the peaks of NiOOH appear until the potential reaches 1.4\u00a0\u200bV and 1.5\u00a0\u200bV in FeMnMo\u2013Ni and FeMnV\u2013Ni samples, respectively, indicating that the Mo dopants and V dopants promote the formation of NiOOH. EDS results show that all elements still distribute homogeneously in the sample after the OER performance test (Fig.\u00a0S7 and Fig.\u00a0S8).To understand the effect of the synergistic interaction of different elements of the catalyst on the electronic structure, the XPS tests were performed on the different catalysts after the OER performance test, as shown in Fig.\u00a04\n. The survey XPS spectra shown in Fig.\u00a0S9 display the presence of Ni, Fe, Mn, Mo, V, and O elements, which are consistent with the EDS results (Fig.\u00a0S7 and Fig.\u00a0S8). Fig.\u00a04a shows the high-resolution XPS spectra of Ni 2p for the different catalysts after OER. Compared with FeMnV\u2013Ni and FeMnMo\u2013Ni, the ratio of Ni3+/Ni2+ in FeMnMoV\u2013Ni is much higher (Table\u00a0S1). In comparison to FeMnMo\u2013Ni and FeMnV\u2013Ni, the ratio of Ni3+/Ni2+ in FeMnMoV\u2013Ni increased significantly, suggesting that the Mo dopants and V dopants could promote the formation of Ni3+ sites.\nFig.\u00a04b and Table\u00a0S2 show the XPS results of Fe 2p. The Fe 2p peak of the samples can be fitted to four peaks [52,53]. The XPS spectra of Mn, Mo, and V are shown in Fig.\u00a04c\u2013e and Tables\u00a0S3\u2013S5. The peaks at 642.1 and 653.3\u00a0\u200beV belong to Mn4+, and the peaks at 232.8\u00a0\u200beV and 235.9\u00a0\u200beV are attributed to Mo6+ in different samples [54]. After the OER test, V is divided into two valence states, V5+ and V4+ [55,56]. The chemical states of Fe, Mn, Mo, and V are similar in different samples, indicating that there is no obvious interaction between Mo, V, Mn, and Fe dopants. The O 1s maps are shown in Fig.\u00a04f, and the peaks located at \u223c536\u00a0\u200beV, \u223c533\u00a0\u200beV, and \u223c531\u00a0\u200beV are attributed to adsorbed H2O, M\u00a0\u200b\u2212\u00a0\u200bOH, and M\u00a0\u200b\u2212\u00a0\u200bO, respectively [57]. It is noteworthy that the M\u00a0\u200b\u2212\u00a0\u200bO signal related to NiOOH increases visibly in FeMnMoV\u2013Ni and FeMoV\u2013Ni catalysts, revealing that Mo and V dopants favor the formation of NiOOH [31,57,58].Here, we designed multiple metal doped nickel nanoparticles via a simple oil phase strategy. The nanoparticles with an average size of \u223c105\u00a0\u200bnm show exceptional OER performance and excellent stability. The FeMnMoV\u2013Ni catalysts exhibit superior OER activity with a low overpotential of 220\u00a0\u200bmV at of 10\u00a0\u200bmA\u00a0\u200bcm\u22122, a Tafel slope of 50.9\u00a0\u200bmV dec\u22121, and a 200\u00a0\u200bh stability at 1\u00a0\u200bA\u00a0\u200bcm\u22122. Mn, Mo and V dopants expose more active sites to boost the OER. In situ Raman analysis indicate that multiple metallic dopants facilitate the formation of NiOOH. Furthermore, XPS analysis show that the Mo and V dopants facilitate the formation of high-valence active sites. This work provides a facile strategy to develop multiple metal doped catalysts and opens a new window for catalyst design and various applications.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Natural Science Foundation of China (NSFC No. 51771132 and 52204320).The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.pnsc.2023.03.002.", "descript": "\n                  Developing efficient oxygen evolution reaction (OER) electrocatalysts is of great importance for sustainable energy conversion and storage. Ni-based catalysts have shown great potential as OER electrocatalysts, but their performance still needs to be improved. Herein, we report the multiple metal doped nickel nanoparticles synthesized via a simple oil phase strategy as efficient OER catalysts. The FeMnMoV\u2013Ni exhibits superior OER performance with an overpotential of 220\u00a0\u200bmV at 10\u00a0\u200bmA\u00a0\u200bcm\u22122 and a long-term stability of 250\u00a0\u200bh in 1\u00a0\u200bM KOH solution. In situ Raman analysis shows that the NiOOH site works as the active center and multiple metallic dopants facilitate the formation of NiOOH. Mo and V dopants promote the formation of high-valence state of Ni sites, and Mn dopants increase the electrochemical active surface area and expose more active sites. This work provides a novel strategy for catalyst design, which is critical for developing multiple metal doped catalysts.\n               "}
{"full_text": "In the past 170 years, human activities have been responsible for almost all of the rapid increase in CO2 emissions, which has been linked to global climate change and ocean acidification (Figure\u00a01\nA).\n1\n\n,\n\n2\n Strategies such as CO2 capture and catalytic conversion have been proposed to mitigate CO2 emissions. Among the latter strategy, thermocatalytic CO2 hydrogenation\n3\n\n,\n\n4\n provides a wide variety of products, such as CO, methane, methanol, olefins, ethers, and aromatics, whereas the CO2 mitigation efficiency depends on the source of H2; electrocatalytic CO2 reduction using water-derived protons is the subject of many recent studies due to its promise in the production of value-added chemicals,\n5\u20137\n such as syngas (CO and H2), formate, alkanes, olefins, and alcohols, although it is still far from competing with conventional processes.\n8\n\nOver the past 2 decades, the revolution in shale gas\u2014which contains both carbon and hydrogen resources\u2014has transformed the world's energy landscape.\n15\n After methane, ethane comprises the second-largest (3%\u201312%) hydrocarbon shale gas fraction, and the discovery of large global reserves of shale gas (ca. 16.1 quadrillion cubic feet) has boosted its supply significantly.\n16\n The ethane surplus far exceeds its demand in conventional petrochemical processes despite its low price, especially in the United States (Figure\u00a01B).\n10\n\n,\n\n17\n Recently, the strategy of simultaneously upgrading this underutilized ethane from shale gas with the greenhouse gas CO2 has introduced opportunities for synthesizing value-added gaseous (syngas\n18\u201321\n and ethylene\n18\u201320\n\n,\n\n22\n\n,\n\n23\n) and liquid (aromatics\n24\n\n,\n\n25\n and C3 oxygenates\n26\n) chemicals. As shown in Figure\u00a01C, these products are among the most essential feedstocks for the petrochemical, agricultural, cosmetic, and pharmaceutical industries all over the world. Global market reports forecast an increasing market value for these chemicals, primarily driven by the rapidly growing demand for synthetic chemicals, fertilizers, cosmetics, and pharmaceuticals.The key to the simultaneous upgrading of CO2 and ethane (SU-CO2Et) involves cleavage of the C=O bond in CO2 along with competitive cleavage of the C\u2013C (dry re-forming) and C\u2013H (oxidative dehydrogenation) bonds in ethane.\n18\n\n,\n\n19\n\n,\n\n23\n Such requirements pose challenges for catalyst materials engineering to develop not only controllable catalyst selectivity and activity for desired pathways but also high resistance to deactivation at elevated temperatures (e.g., due to metal sintering and coke deposition). However, catalysts that demonstrate these properties generally consist of precious platinum-group metals. Therefore, it is crucial to develop advanced catalysts using environmentally friendly and earth-abundant non-precious materials with improved activity and stability while lowering the utilization of expensive and scarce precious metals.\n27\n\nBimetallic materials typically present electronic and chemical properties that are distinct from those of the corresponding parent metals due to ligand and strain effects,\n28\n providing opportunities for designing new catalytic materials with enhanced catalytic performance.\n29\n The terminology \u201cbimetallic-derived catalyst\u201d used herein is a more general concept that refers to catalysts containing two metallic elements supported on a substrate, even though one or both of them may be oxidized under catalytic conditions. Thus, the bimetallic-derived catalysts include not only conventional bimetallic structures, such as core-shell, segregated, ordered, or random homogeneous alloys, but also metal oxide (MO)/metal and MO/MO interfacial structures.Bimetallic-derived catalysts have garnered considerable academic and commercial interest since the early 1960s, and Sinfelt introduced the term \u201cbimetallic clusters\u201d to refer to highly dispersed bimetallic entities present on the surface of a support.\n30\n An understanding of structure-function relationships is required to guide the design of bimetallic-derived catalysts with desired catalytic performance. These relationships can reveal key factors or descriptors\n31\n\n,\n\n32\n that control the reaction pathways as well as the binding and activation energies of critical reaction intermediates. Density functional theory (DFT) calculations in combination with machine learning have been shown to help the discovery of promising bimetallic-derived catalysts from millions of candidates.\n33\u201336\n However, the identification of catalyst structure-function relationships for SU-CO2Et is commonly restricted by the structural complexity of the supported powder catalysts (e.g., reduction/oxidation properties, alloying/dealloying, compositional ordering, site isolation, interfacial interactions, and other dynamic restructuring) and further complicated by the effects of temperature, pressure, and reactant composition. Thus, the atomically precise synthesis of well-defined and uniform bimetallic structures, combined with in situ bulk- and surface-sensitive techniques, is crucial to unravel structure-function relationships and to guide the design of bimetallic-derived catalysts for SU-CO2Et.Given the promising properties and applications of bimetallic-derived catalysts, this review focuses on precious and non-precious bimetallic materials for SU-CO2Et into value-added products. Drawing from the recent literature, we illustrate how to identify, control, and synthesize distinct types of active sites, including selecting different bimetallic compositions and utilizing a variety of MO supports. First, we highlight the importance of thermodynamics and kinetics in understanding bimetallic materials synthesis, and we provide an overview of the state-of-the-art synthesis methods and in situ/ex situ characterization techniques for bimetallic-derived catalysts. We then focus on how to apply distinct types of active sites to upgrade CO2 and ethane into select value-added chemicals (syngas and ethylene), and we illustrate how to gain insight into the structure-function relationships by using combined experimental and theoretical studies. The review concludes by highlighting challenges and opportunities in order to stimulate future fundamental investigations and potential commercial applications of bimetallic-derived catalysts for upgrading abundant shale gas and CO2.In the following discussions, we establish benchmarks for the investigation of bimetallic-derived catalysts in a bottom-up manner. Even though some of the general principles and state-of-the-art synthesis methods have not yet been applied to the synthesis of catalysts for SU-CO2Et, they provide guidance on the synthesis of bimetallic-derived catalysts with desired structures.Understanding the thermodynamics and kinetics of nucleation, growth, and phase formation in bimetallic-derived catalysts provides fundamental guidance on targeting synthesis for tailored applications in catalytic reactions. According to the second law of thermodynamics, a negative value of the change in Gibbs (or Helmholtz) free energy is necessary for a process to be spontaneous. The chemical equilibrium condition (dG\u00a0= 0) for materials synthesis at a specific temperature sets the thermodynamically stable structure of bimetallic-derived catalysts. Despite the importance of thermodynamics in circumscribing and predicting final structures, it is frequently ignored in favor of a trial-and-error or empirical materials synthesis process.Although the growth stage of a nanoparticle (NP) can be regulated via both thermodynamic and kinetic controls, the process of homogeneous nuclei formation can be understood thermodynamically by looking at the total Gibbs free energy change (\u0394G). The \u0394G of homogeneous nucleation (Figure\u00a02\nA) is defined as the sum of two opposite contributions, i.e., a bulk term (free-energy lowering) and a surface term (free-energy increasing) as shown in 1:\n37\n\n\n\n(Equation\u00a01)\n\n\n\u0394\nG\n=\n4\n\u03c0\n\nr\n2\n\n\u03b3\n+\n\n4\n3\n\n\u03c0\n\nr\n3\n\n\n(\n\n\n\u2212\n\nk\nB\n\nT\n\nln\n\n(\nS\n)\n\n\n\u03bd\n\n)\n\n,\n\n\n\nwhere r is the radius of a spherical nucleus, \u03b3 is the free energy increase per unit surface area of the nucleus, \u03bd is the molar volume of the nucleus, k\n\nB\n is the Boltzmann constant, T is the temperature, and S is the supersaturation of solution. As shown in Figure\u00a02A, a critical radius (r\u2217) can be obtained when d\u0394G/dr\u00a0= 0, which gives a critical free energy:\n37\n\n\n\n(Equation\u00a02)\n\n\n\u0394\n\nG\nHomo\n\u2217\n\n=\n\n\n16\n\u03c0\n\n\u03b3\n3\n\n\n\u03bd\n2\n\n\n\n3\n\n\n(\n\n\nk\nB\n\nT\n\nln\n\nS\n\n)\n\n2\n\n\n\n.\n\n\n\n\nThe r\u2217 refers to the minimum size in a metastable equilibrium at which nuclei can continue to grow spontaneously without being redissolved. Accordingly, the \u0394G\u2217Homo essentially represents the energy barrier that must be overcome for continuous growth to occur. Thus, the nucleation rate (J) can be written as follows:\n37\n\n\n\n(Equation\u00a03)\n\n\nJ\n=\nA\n\u22c5\nexp\n\n(\n\n\u2212\n\n\n\u0394\n\nG\nHomo\n\u2217\n\n\n\n\nk\nB\n\nT\n\n\n\n)\n\n=\nA\n\u22c5\nexp\n\n[\n\n\u2212\n\n\n16\n\u03c0\n\n\u03b3\n3\n\n\n\u03bd\n2\n\n\n\n3\n\n\nk\nB\n\n3\n\n\nT\n3\n\n\n\n(\n\nln\n\nS\n\n)\n\n2\n\n\n\n\n]\n\n.\n\n\n\n\nIf the nucleation occurs in the presence of heterogeneous structures such as seeds or supports (Figure\u00a02A), then the energy barrier \u0394G\u2217 for this heterogeneous nucleation would be decreased to:\n38\n\n\n\n(Equation\u00a04)\n\n\n\u0394\n\nG\nHetero\n\u2217\n\n=\n\n(\n\n\n2\n\u2212\n3\n\ncos\n\n\u03b8\n+\n\n\ncos\n\n\n3\n\n\n\u03b8\n\n4\n\n)\n\n\u0394\n\nG\nHomo\n\u2217\n\n,\n\n\n\nwhere \u03b8 is the contact angle and correlates with the characteristics of the nucleus and support (tentatively ignoring the charge effect on the surface).\nEquations 2, 3, and 4 provide a basic outline of how to regulate cluster formation within solutions by carefully controlling the key factors: synthesis temperature, precursor concentration, surface free energy, and contact angle (for heterogeneous nucleation). As illustrated by Figure\u00a02B, in contrast to facile nuclei formation at room temperature, the nucleation of reduced metal atoms at low temperatures (e.g., \u221260\u00b0C) can be inhibited owing to the retarded kinetics. Therefore, under controlled temperature, metal atoms can be highly (and even atomically) dispersed on supports.\n39\n\n,\n\n45\n\n,\n\n46\n\nEquations 2, 3, and 4 also indicate that both homogeneous and heterogeneous nucleation will occur simultaneously as the precursor(s) concentration exceeds the threshold of homogeneous nucleation, which in turn leads to non-uniform catalyst sizes, shapes, and morphologies. Thus, a straightforward way to achieve high and uniform dispersion on the support is to keep the precursor(s) concentration between the homogeneous and the heterogeneous thresholds. As illustrated in Figure\u00a02C, the homogeneous nucleation can be minimized by precisely controlling the injection rate in order to maintain the upper concentration limit (C\nup) below the critical supersaturation for the homogeneous nucleation.\n38\n\n,\n\n47\n Surface free energy \u03b3 also affects the preferred growth of crystallites or facets, and it can be regulated with different capping agents (e.g., inorganic ions, organic polymers, and biomolecules) or surface stabilizers.\n48\n Moreover, the selective chemisorption of the capping agent on a specific type of facet can act as a physical barrier, imposing additional kinetic control over the deposition of metal atoms on this facet.In practice, the surface of supports (such as those in Figure\u00a02D) is typically positively or negatively charged, which dictates the electrostatic interaction of the precursor cations or anions with the support. A thermodynamic model accounting for the support surface charge effect was developed by Yu and colleagues to elucidate the cluster evolution on supports.\n49\n As shown in Equation\u00a05, the total free energy of the supported cluster can be expressed as follows:\n\n(Equation\u00a05)\n\n\n\u0394\nG\n=\n\u0394\n\nH\n\nN\nP\n\n\n+\n\u0394\n\nH\n\n\nN\nP\n\n/\nsup\n\n\n\u2212\nT\n\u0394\n\nS\nM\n\n,\n\n\n\nwhere \u0394H\n\nNP\n and \u0394H\n\nNP/sup are the enthalpy changes due to the metal dispersion and the creation of the metal-support interface, respectively, and \u0394S\n\nM\n is the entropy change due to the mixing of metal clusters and their adsorption on the surface. Among these parameters, \u0394H\n\nNP/sup is the most important item to describe the metal-support interface and can be further represented by a key parameter, \u03b3\n\nss\n, which denotes the interfacial energy with the support. As indicated in Equation\u00a06,\n49\n\n\u03b3\n\nss\n reflects the dependence of interfacial energy on the electrostatic interactions of NPs with supports, which is in turn related to the point of zero charge (PZC) of the support and the pH of solution:\n\n(Equation\u00a06)\n\n\n\n\u03b3\n\ns\ns\n\n\n=\n\n\u03b3\n\ns\ns\n\n0\n\n+\n\n\u03b3\n\ns\ns\n\n\ne\nl\ne\nc\nt\nr\no\ns\nt\na\nt\ni\nc\n\n\n=\n\n\u03b3\n\ns\ns\n\n0\n\n+\n\n\u03b3\n\ns\ns\n\n0\n\n\u03bb\n\n(\n\n\np\nH\n\u2212\nP\nZ\nC\n\n\nP\nZ\nC\n\n\n)\n\n=\n\n\u03b3\n\ns\ns\n\n0\n\n\n(\n\n1\n+\n\u03bb\n\u0394\nz\n\n)\n\n.\n\n\n\n\nAs illustrated in Figure\u00a02E, the surface of supports can be positively (protonated) or negatively (deprotonated) charged by carefully tuning the pH of solution to below or above the PZC value, respectively. As a result, the oppositely charged metal precursor ions can be anchored onto the support in a highly dispersed manner via the strong electrostatic adsorption (SEA). The PZC values of the most commonly used oxide supports are tabulated in Figure\u00a02E.Likewise, the thermodynamically stable cluster size (r) for given precursors, supports (PZC), and pH values of a solution can also be estimated by taking d\u0394G/dr\u00a0= 0 to obtain the following correlation:\n49\n\n\n\n(Equation\u00a07)\n\n\n\nr\n3\n\n+\n3\n\n\n(\n\n\n\u03b3\n\ns\ns\n\n0\n\n\u2212\n\u03b2\nR\nT\n\n)\n\n\n\n\u03b3\n\ns\ns\n\n0\n\n\u03c0\n\u03bb\n\u0394\nz\n\n\nr\n+\n2\n\n\n4\n\u03b1\n\u03b3\n\u03a9\n\n\n\n\u03b3\n\ns\ns\n\n0\n\n\u03c0\n\u03bb\n\u0394\nz\n\n\n=\n0\n,\n\n\n\nwhere \u03b2 is a dimensionless factor that depends on the ratio of the surface concentration of clusters to that of adsorption sites, \u03a9 denotes the atomic volume of the bulk metal, and \u03bb is the independent proportionality constant of the cluster size. This relation can be used to predict the final stable NP cluster size for the given conditions.As illustrated in Figure\u00a03\n, multiple configurations (e.g., core-shell, segregated, ordered or random alloys, and MO/metal and M(O)/support interfaces) might be generated in bimetallic-derived catalysts. Many phase diagrams for binary alloys have been constructed by experimentally determining the equilibrated alloys of the constituents or by calculating the theoretical minimum free energy of the relevant phases. The phase diagram provides a straightforward way to identify the thermodynamically favorable phases that exist under equilibrium at a given temperature, pressure, and composition, which is crucial for designing and predicting the properties of bimetallic-derived catalysts. Subject to minimizing the total free energy, the most probable structures of the bimetallic NPs are core-shell, segregated, ordered (intermetallic compound), or random alloy. The alloying ability of bimetallic NPs can be approximated by the corresponding enthalpy of formation as follows:\n50\n\n\n\n(Equation\u00a08)\n\n\n\nH\np\nf\n\n=\n\nH\nb\nf\n\n\n(\n\n1\n\u2212\n\u03b1\n\nn\n\n-\n1\n/\n3\n\n\n\n)\n\n\u2212\n\u03b1\n\nn\n\n-\n1\n/\n3\n\n\n\n[\n\n\nx\nA\n\n\nE\nb\nA\n\n\n(\n\n1\n\u2212\n\n\nx\nA\n\n\n-\n1\n/\n3\n\n\n\n)\n\n+\n\nx\nB\n\n\nE\nb\nB\n\n\n(\n\n1\n\u2212\n\n\nx\nB\n\n\n-\n1\n/\n3\n\n\n\n)\n\n\n]\n\n,\n\n\n\nwhere \n\n\nH\np\nf\n\n\n and \n\n\nH\nb\nf\n\n\n are the enthalpies of formation of alloy NPs and bulk alloys at 0 K, respectively; \u03b1 is a shape factor; x\n\nA\n and x\n\nB\n are the atomic fractions of the elements A and B, respectively; n denotes the total number of atoms in the NP, correlating with the particle size; and E\n\nb\n refers to the cohesive energy of NPs. As indicated in Equation\u00a08, the alloy formation enthalpy is a monotonically decreasing function of n or NP size, and thereby there should be a critical n or NP size below which \n\n\nH\np\nf\n\n\n would become negative and lead to spontaneous alloying. Thus, reducing the size of NPs should promote the alloying between heterometals, especially those with poor miscibility in the bulk phase.It is noted that Equation\u00a08 cannot provide details about the thermodynamic surface composition. In practice, surface segregation of alloys is a common phenomenon in bimetallic-derived catalysts mainly due to the difference in surface free energy between the two constituents. A Langmuir-McLean model was proposed to approximate the surface fractions of the components (x\n\nsurf\n) under the equilibrium of segregation:\n56\n\n,\n\n57\n\n\n\n(Equation\u00a09)\n\n\n\n\nx\nA\n\ns\nu\nr\nf\n\n\n\nx\nB\n\ns\nu\nr\nf\n\n\n\n=\n\n\nx\nA\n\nb\nu\nl\nk\n\n\n\nx\nB\n\nb\nu\nl\nk\n\n\n\nexp\n\n(\n\n\n\u0394\n\nH\n\ns\ne\ng\n\n\n\n\nR\nT\n\n\n)\n\n=\n\n\nx\nA\n\nb\nu\nl\nk\n\n\n\nx\nB\n\nb\nu\nl\nk\n\n\n\nexp\n\n(\n\n\n\u03b1\n\n(\n\n\n\u03b3\nB\n\n\u2212\n\n\u03b3\nA\n\n\n)\n\n+\n\u0394\n\nH\n\ns\ne\ng\n\n\nm\ni\nx\ni\nn\ng\n\n\n+\n\u0394\n\nH\n\ns\ne\ng\n\n\ns\nt\nr\na\ni\nn\n\n\n\n\nR\nT\n\n\n)\n\n,\n\n\n\nwhere \n\n\u0394\n\nH\n\ns\ne\ng\n\n\n\n is the driving force of segregation and is composed of the contributions from the difference in surface free energy (\u03b3\n\nB\n \u2212 \u03b3\n\nA\n), the heat of mixing (associated with the relative strength of A\u2013A, B\u2013B, and A\u2013B bonds), and the change in lattice strain energy (owing to the mismatch in atomic size). As a consequence, the final surface configuration depends on the net result from the following contributions: (1) the surface energy of the two metals, with the metal showing lower surface energy tending to move to the surface of the NP and forming a shell; (2) the relative bond strength (cohesive energy) between the two different metal atoms compared with that of the same parent metals, where weaker heterometallic bonds favor segregation; (3) the relative atomic sizes, with the smaller atoms tending to occupy the core; and (4) the synthesis or pre-treatment temperature, which is critical to overcome the diffusion barrier of segregation.\n56\n\n,\n\n58\n\n,\n\n59\n Based on a principal component analysis, Wang and colleagues\n53\n revealed that the cohesive energy and atomic size were the primary independent factors controlling surface segregation, as indicated by the color matrix in Figure\u00a03B; furthermore, a strong correlation was obtained between the segregation energy and the core-shell structure preference. This provides a \"design map\" to predict or tune the surface configurations of bimetallic NPs.In addition to the above parameters, the adsorption of species (e.g., CO, O, H, OH, NO, and H2S) during synthesis or reaction also plays a significant role in dictating the surface composition of bimetallic-derived catalysts. The driving force of surface segregation between two metals can be modified due to their different binding strengths to a given adsorbate.\n57\n Using DFT calculations, Menning and co-workers reported an effective way to approximate the adsorbate-induced segregation of subsurface metal atoms to the surface in bimetallic systems.\n54\n Within a wide range of bimetallic systems, as shown in Figure\u00a03B, the driving force of segregation was predicted to be proportional to the difference in occupied d-band center (\u0394\u03b5d) between the subsurface and the surface configurations, but with different slopes for various adsorbates. Thus, adsorbate-induced segregation offers an additional way to vary the surface composition of bimetallic NPs and to explain the dynamic restructuring behavior that occurs during catalytic reactions.In addition, the catalyst support is more than just an inert carrier used to disperse metals on a large-area surface. Instead, the support typically interacts with NPs via chemical bonding and creates metal-support interfaces. These interactions may have a profound effect on the geometric and electronic properties of NPs and in turn on the catalytic performance. Specifically, as shown in Figure\u00a03C, the metal-support interactions could lead to the following effects: (1) charge transfer leading to electron redistribution and even variation of chemical state at the interfacial region; (2) change in NP morphology resulting from the strong interaction of certain facets of NPs with the support; (3) modification of surface composition with one metal segregating from the alloy NP toward the surface or into the lattice of the support due to preferential interactions; (4) metal ions in the support (e.g., oxides of Si, Al, Ga, In, Ti, V, and Nb) being reduced and incorporated into metal NPs to generate new alloy structures; and (5) strong metal-support interactions (SMSIs) resulting in a partial or complete encapsulation of the NPs by the MO layers, typically under reductive conditions. As will be demonstrated in the catalytic reactions of CO2 and ethane, these metal-support interactions and the resultant interfacial sites modify the adsorption of reactants, intermediates, and products, leading to unique catalytic properties. More details about the metal-support interactions can be found in previous reviews.\n55\n\n,\n\n56\n\nThe synthesis of supported bimetallic-derived catalysts involves complicated precursor-precursor/solute interactions in solution and metal-metal/support interactions in solid that vary under different temperatures, pressures, and atmospheric compositions. The controlled synthesis of supported bimetallic-derived catalysts with well-defined, uniform, and tunable structures remains challenging. A comprehensive discussion of synthesis methods of bimetallic NPs has been reviewed previously.\n59\u201362\n The following discussion briefly describes the conventional synthesis protocols and focuses on the state-of-the-art methods for the synthesis of supported bimetallic-derived catalysts.Co-impregnation, sequential-impregnation, and deposition-precipitation methods are the most common approaches used in the large-scale synthesis of supported bimetallic-derived catalysts. Although these methods are simple and economical, they usually lead to NPs with broad size distributions and compositional heterogeneity, which prevents rigorous correlation of the catalytic performance (i.e., activity, selectivity, and stability) with a well-defined structure. Colloidal methods may improve the uniformity but often require structure-stabilizing/directing ligands to cap catalytic surfaces. In practice, it is difficult to completely remove all the ligands, which would reduce the number of exposed metal sites available for catalysis. The colloidal metallic NPs mainly are used as model materials to compare the properties of the alloy with those of its monometallic parent NPs. Metal infiltration is a solvent-free method in which metal precursor salts, such as nitrate hydrate, are physically mixed with the support and subsequently heated above the melting temperatures of the metal salts. However, a non-uniform structure such as a core-shell pattern may be formed if the two metal precursors have very different decomposition temperatures. Thermal decomposition of bimetallic organometallic complexes provides an alternative way to form ultrafine bimetallic NPs. However, its application is limited by the high cost of organic precursors.Atomic layer deposition (ALD) is an efficient chemical vapor deposition method for the preparation of supported bimetallic NPs and MO-decorated NPs.\n52\n The self-limiting deposition reaction between the precursor vapors and the surface allows for an atomically homogeneous distribution, even on highly porous and heterogeneous supports. As illustrated in Figure\u00a04\nA, Lu and colleagues reported a strategy to selectively grow the secondary metal only on the primary metal surface while avoiding growth on the support. By judicious selection of the co-reactant, deposition temperature, and pulsing sequence, monometallic NP formation was prevented successfully.\n63\n\n,\n\n64\n In addition to secondary metals, MOs (e.g., Al2O3 and FeOx) also can be precisely deposited on primary metal NPs such as Pd and Pt with controllable thickness and deposition position.\n65\u201367\n ALD of MOs is also a powerful tool to manipulate the highly and lowly coordinated sites (HCSs and LCSs) of the NPs without changing the particle size. This strategy has been successfully applied to disentangle the size-dependent geometric and electronic effects of Pd/Al2O3 catalysts by selectively blocking the HCSs and LCSs of Pd NPs with the ALD of FeOx and Al2O3, respectively.\n68\n As such, ALD provides opportunities to explore directly the structure-function relationships on MO/metal interfaces in supported catalysts by synthesizing model interface structures.Adsorption or deposition of metal precursor ions typically relies on the electrostatic interactions, van der Waals interactions, and/or polar bonds between the precursor and the support. As illustrated in Figure\u00a02E, SEAs occur between the oppositely charged metal precursor ions and the support. Since the initial study by Brunelle in 1978,\n74\n the SEA method has drawn extensive attention for the preparation of uniform and ultrasmall supported mono/bimetallic NPs. Wong and colleagues\n75\n reported a simple and generalizable simultaneous SEA method using a common silica support with a variety of precious and non-precious metal (Pt, Pd, Co, Ni, and Cu) ammine precursor pairs. The obtained bimetallic NPs were ultrasmall and homogeneously alloyed with a narrow metal size distribution (0.9\u20131.4\u00a0nm). Ding and co-workers\n69\n recently developed a sequential SEA protocol with good size control of supported alloy NPs (1\u20133\u00a0nm) by regulating the sequential uptake of target cations (e.g., [Pd(NH3)4]2+) and anions (e.g., [PtCl4]2\u2212) from heterometallic double complex salts onto silica (Figure\u00a04B). Owing to its facile operation, precise synthesis, and low cost, the SEA method is promising for large-scale synthesis applications.The concept of reactive metal-support interaction (RMSI) is related to the chemical interaction between a metal and the support that induces the formation of bimetallic structures.\n55\n\n,\n\n76\n Xu and co-workers\n77\n reported an RMSI method to synthesize sub-2-nm bimetallic NPs (PtCo, RhCo, and IrCo) on mesoporous sulfur-doped carbon supports via the strong chemical interaction between metals and the sulfur atoms that are doped in the carbon substrate. Perovskite oxides (POs) with the nominal chemical formula ABO3 are versatile templates for the RMSI synthesis of supported bimetallic-derived catalysts due to their compositional flexibility, structural stability, and relatively low cost.\n78\n By substituting transition metals (M and M\u2032) into the B site of a PO with a moderate A-site deficiency (A1-xB1-y-zMyM\u2032zO3-\u03b4), uniform bimetallic NPs (e.g., PtNi and CoNi) can be achieved on the PO support via thermal reduction, during which in situ exsolution of metal NPs would occur.\n79\n\n,\n\n80\n In recent years, MXenes\u2014two-dimensional transition metal carbide materials with well-defined structures and widely tunable compositions\u2014have been explored as supports for bimetallic-derived catalysts. Li and colleagues\n70\n\n,\n\n76\n successfully applied the RMSI method between Pt NPs and Ti3C2Tx and Nb2CTx (T\u00a0= F, O, or OH) MXenes to synthesize MXene-supported well-defined Pt3Ti (6.0\u00a0\u00b1 3.2\u00a0nm) and Pt3Nb (2.6\u00a0\u00b1 0.7\u00a0nm) NPs (Figure\u00a04C). The RMSI method offers an opportunity to obtain supported bimetallic-derived catalysts with tunable chemical and structural properties.Conventional synthesis methods exhibit limited capability to produce homogeneously alloyed bimetallic NPs since they are typically constrained by the thermodynamic immiscibility of the constituents. Certain bimetallic-derived catalysts may possess promising unique electronic and geometric properties despite their inherent immiscibility, such as Cu-X (X\u00a0= Ag, Ni, Sn) NPs. To overcome the thermodynamic hindrance in bimetallic systems, Yang and colleagues\n71\n recently proposed a non-equilibrium synthesis strategy using a pulse thermal shock method, as illustrated in Figure\u00a04D. For each synthesis, metal precursor solutions were mixed well and dispersed on the substrates before being charged by the Joule heating pulses induced by current pulses (0.2 s). The metal precursors (e.g., metal nitrates) were instantly decomposed during the thermal shock, and then the resultant metals were mixed and kinetically trapped into bimetallic NPs following rapid quenching. This type of synthesis strategy sets an example for achieving bimetallic-derived catalysts using metals characterized by thermodynamic immiscibility.In recent years, the development of the concept of \u201csingle-atomic-site catalysis\u201d has spurred the emergence of single-atom alloy (SAA) and hetero-pair metal dimer catalysts, which expand the family of supported bimetallic-derived catalysts. SAAs contain isolated metal atoms on the matrix of another metal and can be synthesized using the galvanic replacement (GR) method, which enables the production of bimetallic and hollow NPs displaying ultrathin walls and even atomic dispersion on a primary metal substrate.\n81\n\n,\n\n82\n The bimetallic NPs obtained can also be deposited over solid supports by complementary wet-impregnation techniques, opening up the possibility of achieving supported SAA catalysts. As shown in Figure\u00a04E, the GR method was used with an aqueous suspension of Cu/metal/MO and a solution of Pt precursor to synthesize a mixed metal (MgAl) oxide-supported PtCu SAA catalyst.\n72\n The synthesis of hetero-pair metal dimers remains challenging, and the state-of-the-art examples are predominantly limited to C(N)-supported systems that use metal-organic framework (MOF) materials as templates.\n83\n As illustrated in Figure\u00a04F, a diatomic metal-nitrogen catalyst (Ni/Fe-N-C) was synthesized by an ion-exchange strategy based on the pyrolysis of a Zn/Ni/Fe MOF; in this method, Ni ions were exchanged with Zn nodes in the framework of an Fe-doped ZIF-8, and afterward residual Zn nodes were evaporated by a thermal treatment at 1,000\u00b0C.\n73\n\nThe characterization of supported bimetallic-derived catalysts is complicated by their wide range of sizes (1\u201310\u00a0nm), shapes (polyhedral, spherical, nanorod, tripod), morphologies (octahedral, tetrahedral, cubic), compositional ordering (core-shell, segregated, random, intermetallic), metal-support interfacial interactions, and reaction-driven restructuring. Although surface science techniques using well-defined crystal surfaces have yielded some structure-function relationships, the extension of such conclusions to practical supported catalysts is hindered by the \u201cpressure gap\u201d and \u201cmaterials gap.\u201d\n29\n Advancements in in situ/operando techniques have enabled atomic-level and time-resolved insights into the diverse geometric and electronic structures of supported bimetallic-derived catalysts under reaction conditions.\n84\n\n,\n\n85\n This section highlights some of the most common and powerful characterization techniques spanning different length scales from atoms to hundreds of nanometers (Figure\u00a05\nA) in order to demonstrate how to achieve comprehensive characterization of supported bimetallic-derived catalysts.Time-resolved in situ X-ray diffraction (XRD) is a prominent technique used to track the structural evolution of the long-range order of bimetallic-derived catalysts under reaction conditions. A significant amount of structural information is contained in the position, intensity, and shape of XRD peaks.\n95\n Phase identification\u2014the most common application\u2014can be achieved simply by comparing the sample diffraction peaks with those of known standards in a database.\n95\n By using the empirical Vegard's law or Zen's law, XRD can be used to estimate the lattice parameter/volume or composition of alloy NPs within the same class of unit cells.\n96\n\n,\n\n97\n Solving the crystal structure via Rietveld refinement of full XRD patterns provides more information about the lattice structure, crystallite size, particle composition, and atomic arrangement. However, XRD can be applied only to systems with long-range periodic atomic arrangement since it counts the Bragg diffraction signals. It is typically difficult to characterize diffraction peaks for particle sizes smaller than 2\u00a0nm due to the low crystallinity or significant disordering, limiting the applicability for NP characterization.In the past decade, advances in synchrotron source and detector technology have enabled a time-resolved total scattering analysis technique, pair distribution function analysis (PDF), to probe supported NP catalysts.\n98\n\n,\n\n99\n As illustrated in Figure\u00a05B, PDF counts the diffuse scattering in addition to Bragg diffraction and thus should be able to depict the probability of finding pairs of atoms separated by a distance r within a local atomic arrangement.\n100\n After a real-space Rietveld refinement, one can also obtain structural information such as the local bonding, lattice strain, and crystalline core size. It should be noted that in order to isolate the PDF information that is associated with metal NPs, differential PDFs should be used for supported catalysts for which contributions from the support can be subtracted from the raw PDF pattern.\n99\n Advantageously, the structural insights are not limited to the immediate coordination environment, enabling PDF to bridge the very important blind range (5\u201330\u00a0\u00c5) that exists among the other X-ray techniques, as shown in Figure\u00a05A.\n86\n\n\nIn situ X-ray absorption spectroscopy (XAS) is a powerful tool for providing information on the electronic and geometric properties of bimetallic-derived catalysts. As illustrated in Figure\u00a05B, XAS measurements consist of the X-ray absorption near edge structure (XANES) and the extended X-ray absorption fine structure (EXAFS) signals. In the XANES region, the appearance of a K or L pre-edge feature, if present, primarily arises from the dipole transition from the 1s or 2p state to the hybridized orbitals of nd \u2212 (n\u00a0+ 1)p.\n101\n The variation in the edge position or the area intensity of the \u201cwhite line\u201d offers information regarding the oxidation state of the element of interest. Within a similar ligand family, the edge position can be used to approximate the chemical state based on the formal oxidation state of standard reference compounds. Moreover, the linear combination fitting of XANES spectra combined with principal-component analysis offers a way to quantify and track the fraction of different compositions in heterogeneous samples under various reduction or reaction conditions.\n102\n The \u0394XANES technique (Figure\u00a05B) utilizes XANES signals relative to a reference sample (such as a substrate that is free of adsorbates). With the help of full multiple scattering ab initio calculations, the difference between these signals can reveal the signature of the adsorbates and the binding sites.\n88\n\nThe oscillatory part in the extended region results from the modulation by the scattering of the outgoing electron off the absorber atom itself (AXAFS) and the backscattering off the neighboring atoms (EXAFS).\n88\n\n,\n\n103\n Fitting the Fourier transformed EXAFS signal (FT-EXAFS, see Figure\u00a05B) reveals structural parameters such as the bond identity, coordination number, bond length, and mean-square disorder deviation (\u03c32). Additional correlation constraints on these parameters in bimetallic systems offer a way of validating the presence of metal-metal and metal-support bonding as well as estimating the particle size, shape, and configuration.\n104\n As a complement, the wavelet transformed EXAFS (WT-EXAFS, see Figure\u00a05B) is applicable to differentiate the equidistant heavier (e.g., transition metals) and lighter (e.g., C, N, O) backscattering atoms from the central atom, due to its better resolution of wave number dependence of the scattering.\n105\n\n,\n\n106\n AXAFS (Figure\u00a05B), which is employed less frequently than EXAFS, contains information about the interatomic potential (bonding electrons and/or oxidation state) and provides insight into the electronic interactions of the support with small- to medium-sized metallic and oxidic clusters (d\u00a0< 2\u00a0nm).\n88\n\n,\n\n103\n The AXAFS signal can be isolated by properly subtracting the double-electron excitation and EXAFS contributions. In recent years, time-resolved dynamic structural and mechanistic changes have been revealed simultaneously using state-of-the-art operando techniques, which couple quick EXAFS or energy dispersive XAS with XRD, three-dimensional tomography, vibrational (IR, UV-Vis, and Raman) spectroscopy, and online gas analysis (micro-gas chromatography and residual gas analysis/mass spectrometry) techniques.\n107\n The increasing application of such cutting-edge combinations will shed light on attaining a more convincing understanding of structure-mechanism-function relationships for bimetallic-derived catalysts.X-ray photoelectron spectroscopy (XPS)\u2014based on the photoelectric effect\u2014is a surface-sensitive technique for qualitatively or semiquantitatively analyzing the elemental and chemical compositions within the near-surface region.\n108\n Conventional XPS requires high-vacuum conditions due to the significant scattering of photoelectrons by gas molecules on the way to the detector, limiting its application under practical reaction conditions. Remarkable improvements in X-ray sources and differential pumping have enabled the application of XPS for unraveling electronic structures and surface compositions under elevated temperature and reaction atmosphere (up to a few tens of torrs).\n109\n For supported bimetallic-derived catalysts, metal-metal or metal-support electron transfers can be elucidated by examining the core-level shift of the metal relative to control samples (monometallic catalysts or inert-material-supported catalysts) or tabulated values. Moreover, a careful examination of the core level of C1s (Figure\u00a05C) and O1s provides information regarding reaction intermediates, such as the types of carbonaceous species (e.g., CO2\n\u03b4\u2212, CO, carbonates, formates, carbides, coke) and oxygen features (e.g., lattice oxygen, oxygen vacancy, OH).\n89\n Notably, the synchrotron light source offers an opportunity to obtain a depth profile, such as for core-shell configurations over bimetallic-derived catalysts, by varying the incident photon energy.\n108\n Despite the excellent capabilities of ambient-pressure XPS (AP-XPS), the signal-to-noise ratio may be significantly reduced for insulated materials and highly diluted catalytic systems.Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) accomplishes useful surface spectroscopic measurements. By exploiting the characteristic shifts in vibrational frequencies of isotopes, this vibrational spectroscopic technique enables operando analysis of surface intermediates. DRIFTS also can be used with specific probe molecules in order to determine the surface termination and binding sites. For instance, the bonding of a CO molecule to a transition metal atom (Figure\u00a05C) involves the donation and back-donation of electrons between the \u03c3-\u03c0\u2217 orbitals of CO and the d-orbitals of transition metals. Consequently, characteristic CO stretching features are observed on different metal surfaces and binding sites. DRIFTS measurements following CO adsorption often can be used to determine the surface termination of a bimetallic system by comparing the vibrational frequencies to those of the corresponding monometallic surfaces. However, this strategy is restricted in cases where carbonyl complex formation, CO decomposition, CO-induced segregation, and weak chemisorption (e.g., on Cu and Ag) may occur. In these cases, other probe molecules should be identified and used in the DRIFTS measurements.Compared with the above-mentioned techniques, scanning transmission electron microscopy (STEM) provides an intuitive view at an atomic scale for insights into the structural properties of supported bimetallic-derived catalysts. STEM is a process wherein pre-specimen lenses focus the beam into a small probe that is scanned in a raster pattern across the sample. A variety of signals can be emitted from the sample due to excitation by the high-energy electron probe. High-angle annular dark-field (HAADF) images (Figure\u00a05D) are obtained by collecting the electrons scattered to high angles with an annular detector, where the signal features a Z-contrast (Z\u03b1, \u03b1\u00a0= 1\u20132) dependence.\n110\n HAADF-STEM provides structural information on supported bimetallic-derived catalysts regarding the size (NPs, clusters, and even single atoms), morphology, lattice fringe, epitaxial growth, and metal-support interfaces. However, the image contrast may not be sufficient to differentiate the Z-similar elements or lighter elements on a heavy support, e.g., Fe and Ni on a CeO2 support. In such cases, HAADF-STEM combined with simultaneous spectroscopic imaging, such as energy-dispersive spectroscopy (EDS) or electron energy loss spectroscopy (EELS) (Figure\u00a05D), can be exploited to elucidate the spatial distribution of elements. The characteristic X-ray emissions from each element following beam irradiation can be collected by an energy-dispersive detector, allowing the analysis of elemental composition and distribution on supported bimetallic-derived catalysts. The characteristic losses in the energy of the transmitted electrons are related to the chemical state and electronic structure of the excited elements. Compared with EDS, EELS provides improved signal, spatial resolution (down to 1\u00a0nm), and sensitivity to the lower-atomic-number elements, such as C, N, and O. For heavier elements (e.g., Pt and Pd), EDS is typically preferred despite its lower spatial resolution. Improved spatial (sub-0.1\u00a0nm) and energy resolution (sub-0.1 eV) of electron microscopy (Figure\u00a05D) has been achieved with the advent of spherical and chromatic aberration correctors. Nowadays, atomic-lattice-resolution environmental (S)TEM has led to unprecedented insights into the dynamic restructuring that occurs under reaction conditions, such as metal sintering, phase transformation, oxidation-reduction processes, layered MO encapsulation, and MO interfacial behavior.\n111\n It should be noted that results from STEM imaging of particularly small areas may not be representative of the entire sample. Thus, STEM should be used in conjunction with other techniques such as those detailed above.In parallel with the aforementioned characterization techniques, macroscopic reactor studies constitute an important method to deduce microscopic structural and mechanistic information. Flow-reactor experiments (Figure\u00a05E) have been applied widely to evaluate the potential synergistic properties of bimetallic-derived catalysts, where the catalytic performance (activity, selectivity, and stability) of the bimetallic-derived catalysts can be compared with that of the sum of the corresponding monometallic-derived catalysts as well as of their physical mixtures. To obtain measurements of the intrinsic catalytic reaction kinetics, two typical approaches are used to eliminate the heat and mass transport limitations inside the catalyst particles and the reactor: (1) enhancing convection by increasing the flow rate at a constant space velocity, and improving diffusion by reducing particle size, and (2) mitigating temperature and concentration gradients using intraparticle and interparticle dilution of active sites with inert materials such as SiO2 powders and acid-purified quartz granules, respectively.\n112\n Multiple criteria (e.g., Kubota, Mears, and Kubota-Yamanaka criteria) can also be adopted to further diagnose the potential transport limitations within the intraparticle, interparticle, and reactor.\n20\n In addition, for steady-state measurements of intrinsic kinetic information such as activation barriers, activation entropy changes, and reaction orders, the overall conversions typically should be kept below 15% (even <5%) so as to enable the application of the differential plug-flow reactor model as well as to alleviate temperature and concentration gradients. It remains challenging to accurately measure activation barriers over catalysts with fast deactivation. In this scenario, a stepwise cooling-heating method might be used to approximate the activation barrier by means of averaging techniques.\n113\n The reaction mechanism can be deduced by examining the kinetic response to the partial pressure of reactants and/or products.In addition to structure characterizations and catalytic performance evaluations, the elucidation of structure-function relationships requires exploration of the surface species adsorbed on the catalyst under reaction conditions. However, even though certain chemisorbed species may be observed by the aforementioned spectroscopic approaches during a reaction, their identification does not necessarily confirm that the overall reaction proceeds via these surface intermediates, since they could be spectators, merely occupying and even poisoning the active sites. In contrast, isotopic labeling provides a more unambiguous means of studying the reaction mechanism. The deuterium kinetic isotope effect is based on a zero-point energy difference that occurs during deuterium substitution in a reaction. A primary isotope effect is observed when the substitution involves the dissociating bond (e.g., C\u2013D in alkanes/alkenes\n114\n\n,\n\n115\n and D\u2013D in dihydrogen)\n116\n in the rate-determining or kinetically relevant step(s). When the substitution involves neighboring (\u0251) or remote (\u03b2) bonds relative to the reaction center (e.g., \u2013CD3 substitution in methanol relative to the hydroxyl group), a secondary isotope effect affecting the reaction rate may occur due to the variations in electronic interactions, such as by hybridization or hyperconjugation.\n117\n\nAs illustrated in Figure\u00a05E, steady-state isotopic-transient kinetic analysis (SSITKA) and temporal analysis of products (TAP) are among the most powerful transient techniques to obtain kinetic and mechanistic information with isotopic labeling. SSITKA detects the response of reactor effluent species to a step switch of isotopic labels, and the response pattern provides a variety of in situ kinetic information regarding the abundance of surface intermediates, mean surface-residence time, reactivity, adsorbate surface coverage, and distribution of site heterogeneity.\n118\n TAP, introduced by Gleaves and colleagues\n119\n during the 1980s, is another transient pulse response technique operated in the well-defined Knudsen diffusion regime with sub-millisecond time resolution. The transient response pattern at a given temperature is a sensitive function of gas-solid interactions, and thereby it offers a unique opportunity to differentiate among the reaction sequences, probe the lifetime of surface intermediates, and isolate the role of gas diffusion inside porous materials such as zeolites.\n94\n\n,\n\n120\n\n,\n\n121\n Despite its excellent performance in kinetic and mechanistic studies, the applications of SSITKA and TAP are still limited by expensive facilities and isotopes as well as by complicated mathematical modeling. However, coupling these methods with other in situ spectroscopic techniques should expand insights into reaction intermediates and mechanisms over bimetallic-derived catalysts.The previous sections provided a survey of the general principles underlying materials synthesis, the state-of-the-art synthesis methods, and the essential characterization techniques for supported bimetallic-derived catalysts. In the current section, several representative bimetallic-derived catalysts with precious and non-precious metals are adopted to illustrate SU-CO2Et into different value-added products. In particular, we demonstrate how the combined in situ and ex situ techniques, coupled with DFT calculations, are utilized to identify distinct types of active sites and establish structure-function relationships.Dry re-forming of ethane (DRE) with CO2\u2014involving the scission of both the C\u2013H and the C\u2013C bonds of ethane\u2014is a promising approach to produce syngas, an important industrial feedstock for methanol synthesis and Fischer-Tropsch reactions. Ni-based catalysts are widely used for dry re-forming, although rapid deactivation owing to severe coke deposition and metal sintering restricts its applications. Extensive studies have been performed to improve the catalytic activity and stability by controlling bimetallic formation, metal dispersion (or size), oxygen storage capacity, support reducibility, and support acidity and basicity.\n122\n\n,\n\n123\n\nYan and co-workers\n21\n synthesized CeO2-supported PtNi bimetallic-derived catalysts for the DRE reaction with CO2 to produce syngas. It is noted that both the conversion (Figure\u00a06\nA) and the turnover frequency (TOF, not shown here) of reactants indicated that the PtNi/CeO2 catalyst was more active than the sum of the corresponding parent catalysts, revealing the synergistic effect from the formation of the Pt-Ni bimetallic bond. The in situ XRD experiments (Figure\u00a06B) revealed the phase evolution of the metals and the CeO2 support with temperature, showing that a PtNi alloy phase (2\u03b8\u00a0= 8.5\u00b0) was formed above 533 K. Simultaneously, a refinement of the CeO2 lattice constant suggested enhanced reducibility of CeO2 (Ce4+ \u2192 Ce3+) in the presence of the PtNi alloy above 533 K, which in turn promoted its CO2 activation properties via more oxygen vacancies. The XANES spectra of the Pt L3-edge and Ni K-edge of the PtNi/CeO2 catalyst revealed that Pt and Ni remained metallic under reaction conditions; and the EXAFS fittings of both edges (Figure\u00a06D) demonstrated the formation of Pt-Ni intermetallic bonds. The TEM images of the reduced and spent PtNi/CeO2 catalysts (Figure\u00a06E) indicated negligible metal agglomeration (2.3 and 2.5\u00a0nm, respectively) after the DRE reaction. To understand the active sites, it was crucial to determine the surface termination of the PtNi alloy configurations. As shown in Figure\u00a06C, comparison of the DRIFTS of CO adsorption over the bimetallic-derived catalyst and the corresponding monometallic catalysts illustrated that the CO vibrational features on PtNi/CeO2 resembled those on Pt/CeO2, indicating a Pt-terminated PtNi bimetallic surface. Therefore, DFT calculations were performed for the C\u2013C bond cleavage of ethane over a Pt-terminated-PtNi-Pt(111) model as well as over Pt(111), Ni(111), and mixed-PtNi-Pt(111) models (Figure\u00a06F). The DFT-calculated binding strengths of all the O- and C-bound intermediates over the Pt-terminated-PtNi-Pt(111) surface generally were weaker than those over the Pt(111), Ni(111), and mixed-PtNi-Pt(111) surfaces. Moreover, comparison of the energy changes along the re-forming of ethane (C\u2013C bond cleavage) revealed that the re-forming reaction on the Pt-terminated-PtNi-Pt(111) surface was the most energetically favorable, as illustrated in Figure\u00a06G. The combined results from characterization and DFT calculations provided the important insights that the Pt-rich PtNi bimetallic surface structure weakened the binding of surface oxygenate/carbon species and reduced the activation barrier for C\u2212C bond scission, leading to enhanced DRE activity as well as stability to produce syngas. This approach demonstrates how robust characterization of active sites can empower computational modeling to reveal structure-function relationships; in turn, these results establish structural targets for the synthesis of catalysts with improved properties.The oxidative dehydrogenation of ethane (ODHE) by CO2 leads to the formation of valuable ethylene with simultaneous consumption of CO2. However, compared with the C\u2013H bond (C2H5\u2013H, 415\u00a0kJ mol\u22121) of ethane, cleavage of the C\u2013C bond (CH3\u2013CH3, 368\u00a0kJ mol\u22121) is thermodynamically favored, thereby imposing a challenge on promoting the product selectivity toward ethylene, which requires breaking the C\u2013H bond while retaining the formed C=C bond. Recent advances in modified Cr/Zr-based catalysts,\n124\u2013126\n metal carbides,\n18\n\n,\n\n127\n MOs,\n128\n\n,\n\n129\n and especially bimetallic-derived catalysts\n19\n\n,\n\n23\n\n,\n\n125\n\n,\n\n128\n\n,\n\n130\n have demonstrated the feasibility of enhancing ethylene selectivity. However, few fundamental strategies are available to guide the development of ODHE catalysts due to their structural complexity under reaction conditions. Recent studies have suggested that the electronic and geometric properties of oxygen species may play a critical role in the selective cleavage of the C\u2013C/C\u2013H bonds of ethane.\n23\n\n,\n\n127\u2013129\n In addition, different metal-oxide or oxide-oxide interfacial configurations may tune the selective cleavage of C\u2013H/C\u2013C bonds by controlling the electronic and geometric features of oxygen species.Bimetallic Ni-Al mixed oxides derived from layered double hydroxides (LDHs) have been applied to ODHE.\n131\n Compared with the NiO/Al2O3 catalyst, Ni-Al mixed oxides are more selective to ethylene (selectivity 70%) due to the partial isolation of the electrophilic oxygen species. Recently, Zhou et\u00a0al.\n128\n reported a sulfate-modified Ni-Al mixed oxide catalyst and highlighted the role of the sulfate ion modifier in enhancing the ethylene selectivity by regulating the proportion of adjacent and isolated oxygen species. As illustrated in Figure\u00a07\nA, the Ni-Al mixed oxide catalysts were synthesized using Ni-Al hydrotalcite precursors interacting with sodium dodecyl sulfate (SDS) via a co-precipitation method. As the SDS amount increased from 0 to 0.05 mol, the layer thickness of the derived LDHs decreased from \u223c7.1 to \u223c2.0\u00a0nm (Figure\u00a07B). The XRD patterns in Figure\u00a07C indicated that the diffraction peaks on the Ni-Al mixed oxide catalysts were shifted slightly relative to those of pure NiO due to the doping effect of Al cations. The FT-EXAFS spectra in Figure\u00a07D demonstrated a significant reduction in the peak intensity associated with Ni-Ni bonding (in NiO) on the NiAl-S1 catalyst. The XPS results indicated a dominant coordination of sulfate with Ni atoms and the co-existence of Ni2+ and Ni3+, and the surface atomic ratio of Ni3+/Ni2+ (Figure\u00a07E) increased with the doping amount of S. Accordingly, the DRIFT spectra of linearly adsorbed CO on the partial positively charged Ni sites exhibited a blue shift with increasing sulfate concentration. The sulfate modifier can also impose a steric effect on the oxygen species, as indicated by a probe reaction of oxygen species coordination from N2O decomposition. The linear correlations in Figure\u00a07E suggested that the sulfate modifier interaction with the Ni3+ sites decreased the adjacent electrophilic oxygen sites and in turn increased the ethylene selectivity. Figure\u00a07F showed that NiAl-S1\u2014possessing the largest proportion of isolated oxygen species (\u223c100%)\u2014exhibited the highest ethylene selectivity (\u223c100%), confirming the importance of controlling the nature of interfacial sites and oxygen species for enhancing ethylene selectivity (Figures 7G and 7H). Similar effects were also observed recently in the ODHE by CO2 over CeO2-supported PdFe bimetallic-derived catalysts, in which the electron-enriched oxygen in the FeOx/Pd interface enhances the selective scission of C\u2013H bond to yield ethylene.\n23\n\nBimetallic-derived catalysts of given metal pairs typically show the capability of promoting reactions of CO2 and ethane toward only one pathway, i.e., either DRE or ODHE. Recent studies by Yan and colleagues\n19\n reported that changing the atomic ratio of Fe and Ni in NixFey/CeO2 bimetallic-derived catalysts enabled tunable selectivity toward either DRE or ODHE.As shown in Figure\u00a08\nA, introducing a small amount of Fe (Ni3Fe1/CeO2) promoted the DRE pathway (99% CO selectivity), while an increased amount of Fe (Ni1Fe3/CeO2) significantly shifted toward the ODHE pathway (78% C2H4 selectivity). Thus, distinct active sites selective to the DRE or ODHE reaction must exist on the NiFe/CeO2 catalysts, which can be regulated accordingly by changing the atomic Ni/Fe ratios. In situ XRD experiments (Figure\u00a08B) revealed slightly larger lattice constants for Ni1Fe3/CeO2 (3.575\u00a0\u00c5) and Ni3Fe1/CeO2 (3.570\u00a0\u00c5) compared with fcc metallic Ni (3.558\u00a0\u00c5), suggesting the formation of a Ni-rich NiFe alloy with the fcc structure on both catalysts. The XANES results showed that Ni in both catalysts remained metallic, while the Fe species in both catalysts exhibited similar near-edge features between Fe(II)O and Fe(II, III)3O4. The EXAFS fitting results (Figure\u00a08C) demonstrated a slight increase in the Ni-Ni(Fe) bond length over both catalysts relative to Ni3/CeO2, in accordance with the formation of a Ni-rich NiFe alloy as suggested by XRD. The Fe-Ce bonding at 3.6\u20133.7\u00a0\u00c5 was observed on both catalysts, demonstrating the strong interaction between Fe and the CeO2 support. Combined with the small coordination numbers of the Fe-O bond, it was suggested that the oxidized Fe species tended to form thin layers on CeO2. A STEM-EELS line-scan analysis (Figure\u00a08D) of Fe species indeed revealed the presence of a thin layer of FeOx on CeO2 particles in Ni1Fe3/CeO2, while such a layer was absent in Ni3Fe1/CeO2.DFT calculations (Figure\u00a08E) showed that, compared with the Ni(111), Ni-terminated-Ni3Fe(111), and bulk-terminated-Ni3Fe(111) models, the FeOx/Ni(111) interface promoted the stability of \u2217CH3CH2 and therefore the dehydrogenation pathway, while hindering the formation of \u2217CH3CH2O and its subsequent dehydrogenation and C\u2013C bond cleavage reactions via \u2217CH3CO. In addition, the DFT-calculated activation barrier of the ODHE pathway (oxygen-assisted C\u2013H bond cleavage) was lower than that of DRE (C\u2013C bond scission) on FeOx/Ni(111), consistent with the experimental results, which identified FeOx/Ni interfaces as the most likely active sites to promote selective C\u2013H bond cleavage in ethane. This work offers a better understanding of the structure-function relationships between interfacial active sites and catalytic performance; furthermore, it highlights the feasibility of tuning the product selectivity of the CO2-ethane reaction by using this understanding to control the active sites of versatile and non-precious NiFe bimetallic-derived catalysts.In addition to the metal effects, support effects or metal-support interfacial interactions also play a significant role in the activation of CO2 and ethane. Specifically, the oxide supports not only act as carriers to disperse catalytically active sites and hence improve conversion, but also exert enormous influence on the electronic and geometric properties of active sites through metal-support interactions, including the modification of oxygen storage/release/transfer, reaction pathways, deactivation resistance, and catalytic performance.\n20\n\n,\n\n125\n\n,\n\n130\n\n,\n\n132\n\nXie and co-workers\n20\n investigated PtNi bimetallic-derived catalysts on reducible (CeO2 and TiO2) and irreducible (\u03b3-Al2O3 and SiO2) oxide supports. The results of initial activity (both conversion and TOF, not shown here) indicated that PtNi catalysts supported on reducible oxides (CeO2 and TiO2) were generally more active than those supported on irreducible oxides (SiO2 and \u03b3-Al2O3). However, PtNi/TiO2 and PtNi/\u03b3-Al2O3 exhibited a rapid decay in the activity. PtNi/CeO2 and PtNi/SiO2 remained stable, with the former showing the highest CO yield (Figure\u00a09\nA) at the pseudo-steady state. Thermogravimetric analysis (TGA) indicated negligible coke deposition on the spent PtNi/TiO2 catalyst, while TEM imaging revealed an encapsulation layer blocking the metal active sites, due to the SMSI effects. On the spent PtNi/\u03b3-Al2O3 catalyst, the TEM (Figure\u00a09B), TGA (Figure\u00a09C), and Raman spectroscopy (Figure\u00a09D) results illustrated that the metal sites underwent noticeable agglomeration from 2.9 to 7.5\u00a0nm, and that the metal ensembles were encapsulated by graphic carbon. PtNi/CeO2 showed large amounts of carbonaceous species on the spent sample, but since they were identified primarily as disordered/amorphous in morphology, they could be removed readily as active intermediates during the re-forming reaction. Flow-reactor experiments revealed different kinetic behaviors for the DRE reaction on the irreducible SiO2- and reducible CeO2-supported PtNi catalysts.Pulse-reactor experiments using CO2 pulses indicated that the SiO2 support acted as a spectator, while the metal sites could weakly activate CO2. In situ DRIFTS studies validated that CO2 activation on PtNi/SiO2 proceeded not only with direct decomposition (CO2\u00a0+ 2\u2217 \u2192 CO\u2217\u00a0+ O\u2217) on metal sites but also with an H-assisted pathway to formate species (CO2\u00a0+ H\u2217 \u2192 formates). In contrast, the pulse-reactor experiments and the ceria lattice parameter derived from the Rietveld refinement of in situ XRD patterns (Figure\u00a09E) revealed that the reducibility of CeO2 was considerably enhanced in the presence of the PtNi alloy. The partially reduced CeO2 could provide additional sites (oxygen vacancies) on the support or on metal-support interfaces for the activation of CO2 (primarily CO2\u00a0+ 2# \u2192 CO#\u00a0+ O#), and the formed active oxygen species subsequently promoted ethane conversion. As illustrated in Figure\u00a09F, the effects of the reducibility of oxide supports should play a significant role in the catalytic performance, reaction kinetics, and reaction mechanisms during the DRE reaction. Similar effects were observed for the ODHE reaction over NiFe bimetallic-derived catalysts supported on reducible (CeO2) and irreducible (SiO2) oxides.\n132\n\nCatalytic conversion of CO2 and ethane into value-added gaseous products (syngas and ethylene) via the SU-CO2Et strategy represents a versatile strategy to mitigate CO2 emissions while utilizing underutilized fractions of abundant shale gas reserves. As summarized in this review, bimetallic-derived materials show promising catalytic activity, product selectivity, and stability for SU-CO2Et, as demonstrated through combined experimental and theoretical efforts in synthesis, reactor studies, and characterization. This section discusses some of the challenges and opportunities to further improve bimetallic-derived catalysts for the reactions of CO2 and ethane.The key to selectively upgrading ethane and CO2 into gaseous and liquid products is to finely control the extent of DRE and ODHE contributions. Based on the aforementioned studies, an illustration is provided in Figure\u00a010\n by considering the most\u00a0relevant key intermediates and reaction pathways [DRE, ODHE, RWGS (reverse water-gas shift), and hydrogenolysis] for different products (C2H4, H2, and CO) and by-products (CH4, H2O, formates, carboxylates, and bi-/carbonates). As implied in Figure\u00a010, designing a catalyst with targeted properties poses significant challenges since it requires the improvement of at least five important reaction steps: (1) CO2 activation, (2) oxygen transfer, (3) activation of C\u2013H/C\u2013C/C=C bonds, (4) formation of the C\u2013O bond, and (5) desorption of C2H4 and CO. Significantly, the failure to balance these five processes would lead to poor activity, selectivity, and/or stability. Thus, delicate design strategies, such as exploring the various synthesis methods described under Synthesis Methods of Supported Bimetallic-Derived Catalysts, need to be considered in order to achieve the desired active sites.For DRE catalysts (e.g., Ni-based catalysts) the efficient non-selective rupture of C\u2013H/C\u2013C/C=C bonds can be achieved easily, while the subsequent oxidation of the derived carbonaceous species is usually more difficult. Thus, these processes may be balanced by (1) enhancing carbon gasification as CO by weakening the binding of carbon and oxygen via alloying or modification with a precious metal with a relatively lower d-band center (e.g., Pd or Pt for Ni); (2) alleviating the rapid non-selective cleavage by hindering the highly mobile electrophilic oxygen species by decreasing large metal NPs into smaller entities; (3) enriching the availability of surface oxygen by introducing an oxophilic element (e.g., Fe, Co, or Mo for Ni) and controlling the CO2/ethane partial pressure ratios to achieve the desired oxygen coverage; or (4) promoting CO2 activation and mobility of resultant oxygen from the bulk to the surface or interface by doping a lower-valence metal into the oxide support bulk lattice, forming a solid solution (e.g., Ce-Zr).\n133\n\n,\n\n134\n\nIn contrast, the ODHE reaction is more likely to be limited by the C\u2013H bond activation (activity) and the desorption of ethylene (selectivity) than by the availability of oxygen from CO2 activation, since the reaction usually exhibits a zero-order kinetic response to the partial pressure of CO2. In this case, catalytic performance for ODHE may be improved by (1) choosing primary metals with considerable C\u2013H bond activation, less activity to hydrogenolysis (e.g., Pd),\n135\n and mild interaction with\u00a0ethylene; (2) weakening the electrophilic nature of oxygen by introducing a secondary metal with stronger M\u2013O bonding\n136\n to form new MO/metal interfaces (e.g., FeOx/Ni, FeOx/Pd, and SnOx/Pt) on the metal ensembles; (3) isolating adjacent electrophilic oxygen adatoms or selectively blocking the highly active sites (such as steps and kinks) for C\u2013C/C=C bond breaking using strongly binding modifiers (e.g., C, S, SiO2, and AlOx); (4) alleviating the supply of oxygen to prevent further oxidation by moderately suppressing the formation of oxygen vacancies on the support and retarding oxygen transfer via robust interfaces (e.g., FeOx/CeO2 interface); (5) maximizing the desired interfacial structures by reducing the size of the active phase or forming core-shell structures; (6) isolating active metal atoms (e.g., Ni or Pd) by forming well-defined intermetallic compounds\n137\n with less active metals (e.g., Zn, Ga, In, and Sn) to promote ethylene desorption by relatively downshifting the d-band center; or (7) breaking the trade-off relationship between activity and selectivity with SAA catalysts, anchoring single atoms of an active metal (e.g., Pd) onto a less active metal host (e.g., Cu, Ag, and Au).The synthesis of bimetallic-derived catalysts with the above-mentioned active-site configurations requires careful and precise synthesis strategies. For example, to obtain uniform and ultrasmall supported random alloy or intermetallic compound NPs, SEA with heterometallic double complex salts, RMSI with appropriate metal and support combinations, or ALD with appropriate gaseous precursors can be used. In addition, well-defined MO/metal interfacial structures can also be synthesized using the ALD method. To obtain the SAA catalysts, GR is among the most popular methods, involving an electroredox process between the sacrificial host metal NP template and the metal ions in solution. In addition, for bimetallic structures that cannot be synthesized by conventional thermal methods due to bulk thermodynamic instability, the pulse thermal shock method can be used to synthesize the desired alloy structures.Improving the SU-CO2Et processes using supported bimetallic-derived catalysts requires an understanding of a wide range of structural and chemical features, such as sizes, shapes, morphologies, compositional ordering, surface terminations, metal-support interfacial interactions, reaction-induced reconstructions, chemical states, surface intermediates, and reaction mechanisms. Thus, a definitive structural and mechanistic determination, especially under reaction conditions, is the premise to establish a reliable structure-function relationship. For a well-defined catalyst with a single type of active sites, the bulk electronic and geometric structural information of the bimetallic system can be elucidated relatively easily using combined characterization results from XANES, FT/WT-EXAFS, XRD-PDF, and HAADF/EDS/EELS-(E)STEM under reaction conditions. In contrast, the identification of the surface or interfacial features (compositional, structural, and chemical) in powder bimetallic-derived catalysts remains challenging owing to the interference by the bulk signals, the severe scattering of the signal on the way to the detector, and the short lifetime of the active intermediates. Intuitively, decreasing the size of metal ensembles would amplify the surface or interfacial contributions, enabling XANES, EXAFS, and AXAFS to probe the surface chemical state, interfacial bond formation, and interfacial electronic interaction, respectively. Moreover, it should be feasible to derive the chemical state and the signature of the adsorbate and its binding site using the \u0394XANES technique coupled with ab initio calculations.The surface properties of the bimetallic-derived catalysts can be revealed directly by\u00a0surface-sensitive analytical techniques, such as low-energy ion scattering (LEIS), DRIFTS, XPS, and electron microscopy. LEIS (Figure\u00a05C) is an extremely surface-sensitive technique used to quantify the composition of the topmost surface, although it is restricted to ex situ studies under ultrahigh vacuum conditions.\n90\n DRIFTS used with appropriate probe molecules such as CO is a useful approach to identify the composition of a bimetallic surface (overlayer, random/ordering alloy, or atom-isolated configurations, etc.). In addition, vibrational spectroscopies such as DRIFTS and Raman are useful techniques to identify key intermediates (e.g., carbonyl, formates, carboxylates, bi-/carbonates, and ethoxy) during some relatively low-temperature processes.\n5\u20137\n\n,\n\n138\n However, special care has to be taken when applying these techniques to the high-temperature reaction of SU-CO2Et, since the real intermediates might be too transient and weakly binding to be detected, while the observed species might be spectators. As a complementary technique, the SSITKA and TAP techniques coupled with isotope switching (e.g., 13CO2, C18O2, 13C2H6, or C2D6) and kinetic analysis should be employed to explore the key intermediates and reaction pathways. The advancements of differential pumping and microreactors have enabled the AP-XPS technique to identify the near-surface structural compositions and reaction intermediates over bimetallic-derived catalysts at elevated temperatures and in the presence of CO2 and light alkanes. Structural properties (e.g., size, shape, facets, morphology, defects, strain, alloying, and segregation) generally have been measured by ex situ (S)TEM. In recent years, both layered encapsulation- and RMSI-induced interfaces have attracted extensive attention due to their superior performance during reactions, while their properties are typically sensitive to distinct temperatures and chemical atmospheres. Emerging in situ (or semi-in situ) (S)TEM techniques\n139\u2013141\n with atomic resolution have demonstrated the feasibility of directly tracking surface encapsulations by MO clusters or layers (e.g., FeOx/Ni [or Pd]) and metal-support dynamic interactions (e.g., PtNi/CeO2) in the presence of CO2 and ethane at reaction temperatures.Equally as important as the metallic components, it is noted that the nature of the oxygen species is critical in determining the bond cleavage of ethane and its derivatives. Certain types of electrophilic oxygen species may exhibit unpaired electrons, which are usually too active and transient to be tracked by the above-mentioned techniques. In this situation, electron spin resonance spectroscopy is an appropriate approach to explore the evolution of these transient oxygen species, due to its extreme sensitivity to paramagnetic features.\n142\n\n,\n\n143\n At a steady state, the type and chemical state of oxygen species under reaction conditions can be elucidated by performing AP-XPS measurements of the O 1s feature. To gain a deeper understanding of the variations in oxygen reactivity due to electronic, geometric, and kinetic modifications at a specific orbital level, the total density of state (DOS) is usually calculated and projected to the surface atoms of interest (e.g., O and neighboring metals) using DFT calculations. Experimentally, resonant inelastic X-ray scattering (RIXS) spectroscopy\u2014essentially containing XAS and X-ray emission spectroscopy (XES)\u2014makes it feasible to probe the electron orbitals. The XAS and XES spectra provide information on the unoccupied and occupied electronic states, respectively; these results enable projection of the electronic structure onto the excited atom and can be compared directly with the linear combination of atomic orbitals approach used in DFT calculations. Therefore, RIXS should be a powerful tool to investigate intramolecular bonds and adsorption of surface species during the reactions of CO2 and ethane on bimetallic-derived catalysts. In the soft X-ray region (such as around the O and C K-edge), XAS and XES are dominated by dipole transitions, together with the symmetry character of the core level; thus, they can provide important information regarding the chemical bonding in an oxygen (2p)-projected manner.\n144\n\n,\n\n145\n RIXS also has been applied successfully for metal edges in the hard X-ray region to map out the metal (d)-projected DOS.\n146\n\n,\n\n147\n By combining the structural and chemical information elucidated by the RIXS- and DFT-derived site-projected DOS\u2014together with information obtained from the other techniques mentioned above\u2014it would be possible to determine the precise nature of the active sites responsible for the various pathways in the reaction of CO2 with ethane.As shown in Figure\u00a010, the CO2-assisted selective cleavage of the C\u2013H and/or C\u2013C bonds of ethane produces a variety of gaseous chemicals, including CO, H2, C2H4, CH4, and H2O. Among these, C2H4, CO, and H2 are important industrial feedstocks, which further can be upgraded into liquid products, such as via the aromatization reaction of C2H4 to aromatics\n24\n or the hydroformylation reaction of C2H4, CO, and H2 to C3 oxygenates.\n26\n These gas-phase heterogeneous reactions produce higher-value liquid products (e.g., aromatics and oxygenates) and reduce the costs associated with the separation of products from the gaseous reaction stream or from conventional liquid-phase homogeneous reactions (e.g., hydroformylation). However, the aromatization and hydroformylation reactions require additional types of catalytic active sites. In practical applications, it would be desirable to combine the CO2-ethane reactions (DRE and/or ODHE) with further upgrading (aromatization or hydroformylation), although this would result in additional challenges for the design of bimetallic-derived catalysts containing multifunctional active sites for different reactions. This prospect of achieving SU-CO2Et to higher-value products requires further advances in the synthesis and characterization of promising bimetallic-derived catalysts.This work was supported by the US Department of Energy, Office of Basic Energy Sciences, Catalysis Science Program, under contract DE-SC0012704. L.R.W. acknowledges the US National Science Foundation Graduate Research Fellowship Program grant DGE 16-44869.Conceptualization, Z.X. and J.G.C.; Writing \u2013 Original Draft, Z.X.; Writing \u2013 Review & Editing, Z.X., L.R.W., and J.G.C.; Supervision, Z.X. and J.G.C.", "descript": "\n                  Bimetallic-derived catalysts have played a pivotal role in many industrially important catalytic processes. In recent years, simultaneously upgrading the greenhouse gas CO2 and underutilized ethane (SU-CO2Et) has represented a promising route for producing value-added chemicals, such as synthesis gas and olefin. This review focuses on the synthesis and characterization of precious and non-precious bimetallic-derived catalysts and their application in SU-CO2Et. We discuss general principles and state-of-the-art strategies for catalyst synthesis, as well as in situ techniques for structural characterization and mechanistic insights. We then illustrate how to design and apply versatile bimetallic-derived catalysts for SU-CO2Et and to establish structure-function relationships by using combined experimental and theoretical approaches. We conclude the review by highlighting challenges and opportunities in active-site design and synthesis, in situ structural and mechanistic identification, and development of multifunctional bimetallic-derived catalysts for further upgrading abundant CO2 and shale gas into valuable aromatics and oxygenates.\n               "}
{"full_text": "According to statistics reported by International Energy Agency (IEA) and the American Energy Information Administration (EIA), the global energy consumption increases by 50 quadrillion Btu every year [1]. From another perspective, the intensive dependence on fossil fuels for electricity generation has absolutely altered our ecosystem [2\u20138]. Hence, research has been oriented to offer clean, abundant, and renewable sources for energy that suffice the global consumption and maintain a safe environment.In this context, fuel cells (FCs) appeared promising as advanced electrochemical energy converters of featured environmental flexibility, and high efficiency at low operating temperatures [9\u201312]. Technically, FCs convert directly and efficiently (up to 60%) the chemical energy into electricity with a tremendous (up to 90%) reduction in major pollutants [13]. The polymer electrolyte membrane FCs (PEMFCs) that belongs to the low-temperature FCs that utilize solid polymeric electrolyte for the ionic conduction has attracted a special attention due to their robustness, high power density, and low operational temperature. Of these low-temperature PEMFCs, the direct methanol FCs (DMFCs) and hydrogen FCs (HFCs) offered, recently, a comprehensive interest for electric vehicles and portable electronics [6,14,15]. However, their fast commercialization has been restricted by the huge cost of H2 containers, the prospective risks during H2 transportation and the toxicity and crossover of methanol [2,5,16]. Hence, the DFAFCs appeared more attractive owing to their higher theoretical open circuit potential (1.45\u00a0V compared to 1.23\u00a0V of HFCs and 1.21\u00a0V of DMFCs), lower toxicity, non-flammability and lower crossover through the Nafion membrane [17\u201320].Till now, the most common anodic catalysts for DFAFCs are mainly based on noble Pt and Pd metals [11,21\u201324]. However, Pt revealed a higher stability than Pd due to its higher dissolution resistance in harsh reaction conditions [25,26]. However, unfortunately, Pt can easily be poisoned by some reaction intermediates, such as carbon monoxide (CO) [27\u201329]. Several investigations reported that FAOR proceeds at Pt surfaces with a double-pathway mechanism [2,16,30\u201333]. The direct pathway which involves the direct oxidation of FA to CO2 (Eq. (1)) and the indirect pathway which involves the non-faradaic adsorption of poisoning CO at the Pt surface (Eq. (2)) followed by its subsequent oxidation at a high overvoltage (Eq. (3)).\n\no\nDirect dehydrogenation pathway:\n\n\n\n\n(1)\nHCOOH\u00a0\u2192\u00a0CO2\u00a0+\u00a02H+\u00a0+\u00a02e\u2212\n\n\n\n\n\no\nIndirect dehydration pathway:\n\n\n\n\n(2)\nHCOOH\u00a0\u2192\u00a0COads\u00a0+\u00a0H2O\n\n\n\n\n(3)\nCOads\u00a0+\u00a0H2O\u00a0\u2192\u00a0CO2\u00a0+\u00a02H+\u00a0+\u00a02e\u2212\n\n\n\nDirect dehydrogenation pathway:Indirect dehydration pathway:Considering the huge cost of Pt catalysts, it became crucial to amend it with a cheaper modifier to increase its utilization efficiency and catalytic performance [34,35]. In fact, previous modifications of Pt with metals (e.g. Au [32], Pd [36], Bi [37] and/or metal oxides (e.g., oxides of Ni [30], Mn [2], Cu [38], and Fe [39])) could greatly sustain better performance toward FAOR with enhanced structural and/or electronic properties.Although the difficultness in controlling size, shape and distribution of the catalyst component [20], electrodeposition represents a facile, fast and economic technique for assembling metal and metal oxide modified catalysts; ensuring a controlled production of a smooth surface with strong bonding with the substrate and offering opportunities for alloys and composite coatings with high hardness [40].In this study, a binary catalyst composed of Pt and Cu was fabricated onto the GC (a typical substrate for the deposition of nanoparticles that can be used for a simple investigation) and proved competent for FAOR. The catalyst was synthesized by the \u201csimultaneous co-electrodeposition\u201d protocol that ensured a convenient homogeneity of the catalytic constituents that were added in minute loadings (relatively to other procedures as the layer-by-layer approach) [2,32]. The molar Pt4+/Cu2+ ratio of the electrolyte during the catalyst's deposition was optimized to attain the highest catalytic activity toward FAOR. Furthermore, the catalyst\u2019s morphology, surface composition, and molecular structure were inspected to address the remarkable enhancement of these catalysts toward FAOR.Copper (II) sulfate pentahydrate (CuSO4\u00b75H2O, 99%), sodium hydroxide-pellets (NaOH), sodium sulfate anhydrous (Na2SO4), and FA (HCOOH, 98%) were purchased from Alfa Aesar while dihydrogen hexachloroplatinate (IV) hydrate (Premion\u00ae, H2PtCl6\u00b76H2O, 99.9%, metals basis) and sulfuric acid (AR, H2SO4, 98%) were purchased from Sigma Aldrich. All chemicals were of high purity and were used as received without further purification. A three-electrode electrochemical cell was used for the catalyst's preparation and electrochemical and catalytic inspections. A cleaned (by mechanical polishing with aqueous slurries of successively finer alumina powder (down to 0.06\u00a0mm) followed by a thorough washing with second distilled water) pristine and modified glassy carbon (GC) electrode (5.0\u00a0mm diameter) of a geometric area of ca. 0.196\u00a0cm2 was used as the working electrode, a spiral Pt wire was used as the counter electrode and an Ag/AgCl/NaCl (3\u00a0M) electrode was used as the reference electrode. All potentials were measured relative to this Ag/AgCl/NaCl (3\u00a0M) reference electrode.The \u201csimultaneous co-electrodeposition\u201d technique was employed to prepare PtxCuy catalysts onto the GC electrode surface with several molar ratios (starting from 1:0 till 1:4) [2,32]. The electrolyte of electrodeposition was 0.1\u00a0M Na2SO4 aqueous solution containing 2.0\u00a0mM H2PtCl6\u00b76H2O and 2.0\u00a0mM CuSO4\u00b75H2O For all catalysts, the electrodeposition of Pt and Cu onto the GC electrode surface was carried out potentiostatically at \u22120.2\u00a0V permitting the passage of only 9.4 mC.For a simple recognition of the electrodes' preparation, an abbreviation of PtxCuy was assigned to recognize the molar ratio of Pt4+ to Cu2+ in the deposition electrolyte, where x and y referred to the molar ratios of Pt4+ and Cu2+ ions, respectively. For example, the catalyst denoted as Pt1Cu4 correspond to a mole ratio of 1:4 for Pt4+ to Cu2+ ions in the deposition electrolyte.All electrochemical experiments were tested at room temperature (ca. 25\u00a0\u00b1\u00a01\u00a0\u00b0C) in aqueous solutions using a Bio-Logic SAS Potentiostat (model SP-150) operated with EC-Lab software. The electrocatalytic performance of the PtxCuy catalysts toward FAOR was inspected in aqueous solutions containing 0.3\u00a0M FA (pH \u223c3.5). The pH was adjusted by a dilute solution of NaOH. Current densities were always calculated on the basis of real Pt surface areas of the working electrodes (as Fig. S1 shows) employing a reference value of 420 \u03bcC cm\u22122\n[41].The morphology and elemental composition of PtxCuy catalysts were evaluated using a field-emission scanning electron microscope (FE-SEM, Quattro S, Thermo Fisher Scientific USA) whose accelerating voltage extended from 200\u00a0V to 30\u00a0kV with a magnification range from 6 to 2500000x that equipped with an energy dispersive X-ray spectrometer (EDS, AMETEK USA Element Detector). The crystallographic information of PtxCuy catalysts was obtained using a high resolution X-ray diffractometer (XRD-PANalytical X\u2019Pert Pro powder) that operated with a Cu anode, wavelength 0.154\u00a0nm, maximum 2.2\u00a0kW, and 60\u00a0kV. The inductively coupled plasma mass spectrometry, ICP-MS, (8800 ICP-MS, Agilent Technologies) was employed to assess the dissolution (loss) of Pt and Cu from the catalysts after stability measurements.\nFig. 1\na (Pt1Cu0 catalyst) displays the characteristic performance of a poly-Pt electrode in an acidic medium. This demonstrated the oxidation of Pt which extended over a potential range between ca. 0.6 and 1.2\u00a0V and coupled with the subsequent PtO reduction at ca. 0.46\u00a0V. Furthermore, the peaks that appeared in the potential range between 0 and \u22120.2\u00a0V were assigned to the hydrogen adsorption/desorption (Hads/des) at the Pt surface [25]. For Fig. 1b\u2013e, the current intensities of the Hads/des and PtO reduction peaks were gradually decreased with a parallel decrease in the intensity of the PtO reduction peak. This resulted from the distribution of the deposition charge between Pt and Cu which further indicated the successful deposition of Cu. As a result of Cu deposition, a new redox couple corresponding to the Cu oxidation (at ca. 0.45\u00a0V) and its subsequent reduction (at ca. 0.25\u00a0V) was developed [42\u201344]. With the further increase of Cu2+ in the solution (Fig. 1d and 1e), the current intensities of the Hads/des and PtO reduction peaks continued decreasing significantly concurrently with an observable increase in the current intensity of the Cu oxidation peak which appeared split [45,46] in two peaks; at 0.05\u00a0V and 0.45\u00a0V (almost similar to those obtained at the Pt0Cu1 catalyst (Fig. S2A)) to infer about a possible phase transformation. The calculated Pt surface area for all proposed catalysts was additionally calculated, based on the procedures in Fig. S1, as appeared in Table S1.The investigation is directed to elucidate the surface morphology, elemental composition, and molecular structure of the proposed PtxCuy catalysts. Fig. 2\n displays FE-SEM micrographs of the Pt1Cu0, Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts. As obviously seen in Fig. 2A, the deposition of Pt (Pt1Cu0 catalyst) occurred in spherical shape (ca. 110\u00a0nm in average diameter) with intensive aggregations (\u223c500\u00a0nm in diameter each). This morphology was retained for Pt with the deposition of Cu in starfish and/or intersected laminar structures (almost similar to the Cu morphology for the Pt0Cu1 catalyst appeared in Fig. S2B) in the Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts (Fig. 2B-E).Moreover, the EDS spectra of the Pt1Cu0, Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts (Fig. 3\na-e, respectively) provided a confirmation for the deposition of catalytic ingredients in the different catalysts where all peaks of C, O, Pt, and Cu appeared in their expected positions [47\u201349]. The elemental mapping of the Pt1Cu0 and Pt1Cu4 catalysts demonstrated the homogeneous distribution of C, O, Pt, and Cu elements in the proposed catalysts (Figs. S3A and B, respectively).Furthermore, the crystal structures of the different PtxCuy catalysts were examined by XRD (Fig. 4\na-e) where several diffraction peaks were identified for all catalysts at ca. 25\u00b0, 43\u00b0 and 79\u00b0 corresponding, respectively, to the (002), (100), and (110) planes of the hexagonal carbon structure (JCPDS card No. 075-1621) [50]. Also, the diffraction peaks identified at ca. 38.6\u00b0, 44.6\u00b0, 65.4\u00b0, and 78.7\u00b0 for all catalysts were assigned, respectively, to the (111), (200), (220), and (311) planes of the face-centered cubic (fcc) Pt lattice (JCPDS card No. 96-101-1112) [32,51]. The Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts showed three diffraction peaks at 2\u03b8 of ca. 43.1\u00b0, 50.2\u00b0, and 73.7\u00b0, consistent with those observed for the Pt0Cu1 catalyst (Fig. S2C), which were assigned to the (111), (200), and (220) facets of metallic Cu (JCPDS card No. 96-901-3016) [52\u201354]. The very small positive shift (ca. 0.05\u00b0) in the Pt diffractions in the Cu-modified catalysts might account for the composition change of the catalyst surface and/or Pt-Cu alloying [55].\nFig. 5\n (for easy comparison, all subfigures in Fig. 5 were grouped in a single figure - Fig. S4) represented the influence of Pt4+ and Cu2+ relative molar ratio in the deposition electrolyte on the catalytic activity of the proposed PtxCuy catalysts toward FAOR. First, it worth to point the inactivity of the unmodified GC electrode [56] and Cu [57] toward FAOR (see Fig. S2D). A blank test was carried out in FA-free solution (FAFS) that has the same pH (3.5) with the same measuring conditions to confirm our interpretation (will be mentioned later in text) for the peaks associated with FAOR. Fig. S5 confirmed that in FAFS, no peaks were detected at the same potentials like in case of the solution contacting FA. In Fig. 5A (Pt1Cu0 catalyst), two oxidation peaks were observed in the forward (anodic-going) scan at 0.34\u00a0V and 0.8\u00a0V. The first peak (at ca. 0.34\u00a0V) was assigned to the direct (preferred, because of its lower anodic overpotential that turns the output voltage of DFAFCs higher) oxidation of FA to CO2 (Eq. (1)). The current density of this peak will be abbreviated as I\np\nd. The second peak (ca. 0.8\u00a0V) was assigned to the oxidation of the COads to CO2 (Eq. (3)) after the hydroxylation of the Pt surface at a potential of ca. 0.7\u00a0V. The current density of this peak will be abbreviated as I\np\nind\n[58]. In reality, the core challenge of assigning Pt-based catalyst for FAOR is related to the adsorption of CO (COads) which occurs spontaneously from the non-faradaic dissociation of FA at open circuit potentials (Eq. (2)). This deactivates the Pt surface and prompts a potential poisoning for a significant number of Pt active sites, which, in turns, impede the direct \u201cpreferred\u201d pathway of FAOR. Balandin proposed the \u201cMultiplet theory\u201d that investigated the simultaneous adsorption of reacting species to a group of active atoms of a given catalyst [59]. He proposed a correspondence between the geometry of active centers and the energies of forming and breaking chemical bonds of the adsorbate/adsorbent clusters. According to this theory, the adsorption of CO on the Pt surface requires the presence of three neighboring Pt active sites with a certain geometry. If this contiguity was disturbed, the Pt\u2013CO bonding will not form; as will be elaborated below. In the backward (cathodic-going) scan, the Pt surface became clean (free of poisoning COads) after the oxidation of poisoning COads and that boosted FAOR as shown from the high peak current density of the backward scan (I\np\nb).For the Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts, critical changes appeared and influenced the relative ratios of I\np\nd, I\np\nind, and I\np\nb which denote important catalytic information. Generally, the I\np\nd/I\np\nind ratio evaluates the enhancement in the catalytic activity in the favorable direct oxidation pathway. On the other hand, the I\np\nd\n/I\np\nb ratio estimates the catalytic tolerance of the catalyst for poisoning CO species. These changes in current densities can associate a further change in the onset potential of the direct FAOR, E\nonset (measured at a constant current density of ca. 0.2\u00a0mA\u00a0cm\u22122), that reflects the capability of the catalyst to overcome unnecessary overpotentials (particularly of charge transfer) that normally detracts the voltage output of the cell. The I\np\nd/I\np\nind ratio of the Pt1Cu0 catalyst was ca. 0.65 (see Table 1\n), which is low to permit the movement for DFAFCs into a real commercialization. The increase of this I\np\nd/I\np\nind ratio of the catalyst is highly important for a commercial purpose. Fascinatingly, the CVs in Fig. 5B-E and the catalytic data in Table 1 demonstrated the importance of adding Cu to the catalysts and elaborated the influence of varying the (Pt4+ and Cu2+) molar ratios in the deposition electrolyte on the catalytic activity toward FAOR. Interestingly, both of I\np\nd/I\np\nind and I\np\nd\n/I\np\nb increased with the increase in the Cu2+ molar ratio. This reflected the critical role of Cu to direct FAOR in the direct pathway and to mitigate the CO poisoning. In addition, with the increase in Cu2+ molar ratio, a regular negative shift in E\nonset was observed. The best catalytic data was obtained for the Pt1Cu4 catalyst whose I\np\nd/I\np\nind was 3.58 (i.e., 6-times as that of the Pt1Cu0 catalyst). Its I\np\nd\n/I\np\nb was also the highest (0.73, i.e., 4-times as that (0.18) of the Pt1Cu0 catalyst). The negative shift in E\nonset of this Pt1Cu4 catalyst was as well the largest (ca. 336\u00a0mV). Fig. 6\n represents graphically these catalytic data. Also, several other parameters such as the potentials at I\np\nd, I\np\nind, and I\np\nb (E\np\nd, E\np\nind, and E\np\nb, respectively) were monitored and tabulated in Table S2. The potential changes associated with FAOR may relate to the surface composition change of the proposed catalysts.It is important to mention that the catalytic activities of the Pt1Cu5 and Pt1Cu6 (see Figs. S6A and B and data in Table S3) catalysts toward FAOR were lower than that of the Pt1Cu4, presumably due to the too low loading of the active (Pt) component in the catalysts. Hence, the Pt1Cu4 represented the best catalyst for FAOR among all the inspected catalysts in this investigation. Interestingly, this activity surpassed many of the reported activities for FAOR in literature (see Table S4 in the supplementary data file).Another important measurement besides the catalyst\u2019s activity is related to the catalyst\u2019s stability. Herein, the catalyst\u2019s modification with Cu was proposed not only to promote the catalytic activity but also to enhance the stability of the catalyst, which quickly deteriorates during continuous electrolysis. The stability of the entire set of our proposed catalysts were assessed by chronoamperometric measurements (see Fig. 7\n) for 3600\u00a0s at a constant potential of 0.2\u00a0V. Fig. 7 (a-e, respectively) displays the current transients (i-t curves) of the Pt1Cu0, Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts in aqueous solution of 0.3\u00a0M FA (pH \u223c3.5) at 0.2\u00a0V. As obviously seen in Fig. 7a and the attached inset, the current density of the Pt1Cu0 catalyst decayed rapidly due to the accumulation of poisoning CO on the Pt surface. This decay diminished largely for the Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts (see Fig. 7b-e). Interestingly, the highest stability was also recorded for the Pt1Cu4 catalyst (20% loss in the catalytic activity compared to 35% for the Pt1Cu0 catalyst, see Fig. 7a and e). This represented an additional merit for Cu in boosting the catalytic tolerance of the PtxCuy catalysts against CO poisoning during FAOR.The ICP-MS was employed to assess the loss in the catalytic ingredients (Pt & Cu) after the electrochemical stability inspection. As expected, a loss in Cu was observed, which was expected for Cu at high potentials. This loss in Cu increased with the molar ratio of Cu2+ ions in the deposition electrolyte (Table 2\n). On contrary, the loss of the active and precious material (Pt) in the Pt1Cu0, Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts was relatively minor and got decreased with the Cu2+ molar ratio (Table 2). This reinforced the role of Cu in boosting the durability of the PtxCuy catalysts and in ranking the Pt1Cu4 catalyst the best for FAOR.Electrochemical impedance spectroscopy (EIS) was used to interpret the catalysis of FAOR on the PtxCuy catalysts. The charge transfer resistance (R\nct) of the proposed PtxCuy catalysts was correlated to their catalytic performance toward FAOR [13,60,61]. Fig. 8\n (a-e, respectively) represents the Nyquist plots for all catalysts in aqueous solution of 0.3\u00a0M FA (pH \u223c3.5) at a potential of 0.2\u00a0V in the frequency range between 10\u00a0mHz and 100\u00a0kHz. The data fitting was carried out using the EC-Lab software and the equivalent circuit of this system was displayed in the inset of Fig. 8. Over there, R\ns and C\ndl referred to the solution resistance and double layer capacitance, respectively, of the electrochemical system. Analysis of R\nct for the whole set of catalysts grouped them in two categories; one of a lower and another of a higher R\nct than that (0.21 k\u03a9) of the unmodified Pt1Cu0 catalyst (Fig. 8a). The Pt1Cu1 and Pt1Cu2 catalysts recorded 0.14 and 0.15 k\u03a9, respectively, for R\nct as obviously seen from their smaller semicircle diameters (Fig. 8b and c). Such a decrease in R\nct inferred the existence of an electronic element in the catalytic enhancement [5,13]. This electronic enhancement could possibly result from the Pt-Cu alloying that might affect the Pt\u2013FA, Pt\u2013CO2 and/or Pt\u2013CO bonding or perhaps from the participation of Cu with its higher electrical conductivity than Pt [62,63] in the reaction mechanism of FAOR in the way facilitating the kinetics of charge transfer. This might associate a structural influence that could synergistically boost the catalytic enhancement. Surprisingly, the Pt1Cu3 and Pt1Cu4 catalysts owned higher (0.27 and 0.43 k\u03a9, respectively) R\nct than that of the Pt1Cu0 catalyst with larger semicircle diameters (Fig. 8d and e). This came consistent with diminishing the active Pt surface and the appearance of redox pair for copper (recall the splitting of the Cu peak that appeared only in Fig. 1d and e) that probably deactivated Pt electronically toward FAOR. This electronic deactivation was not equivalent to the geometrical (structural, third body) influence that Cu added to Pt which boosted synergistically the catalytic activity of the Pt1Cu3 and Pt1Cu4 catalysts toward FAOR. Table 3\n summarizes the electrochemical data (R\ns and R\nct) obtained from Fig. 8.To precisely confirm this claim, CO was allowed to be adsorbed at open circuit potential for 10\u00a0min and then stripped oxidatively in CO-free electrolyte containing 0.5\u00a0M Na2SO4 (pH \u223c3.5) at the Pt1Cu0, Pt1Cu1, Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts (Fig. 9\na-e, respectively). The Pt1Cu0 catalyst (Fig. 9a) showed (in the anodic scan) a zero current (Pt surface was blocked) up to ca. 0.73\u00a0V where CO started to desorb [64,65]. The charge (Q\nCO) consumed in the CO stripping is proportional to the poisoning level of COads and the onset potential of CO desorption (Eonset/\n\nCO) assess the minimum energy required for this desorption, which also accounts for the electronic properties of the Pt surface. Fortunately, the data of Fig. 9 agreed with the hypotheses of Fig. 8 in suggesting prevailing the electronic element in the catalytic enhancement of the Pt1Cu1 and Pt1Cu2 catalysts. This was obvious in the increased negative shift of their Eonset/\n\nCO (Fig. 9b and c). However, the amount of Cu in the Pt1Cu1 catalyst was not sufficient to provide an overall (electronic and geometric) enhancement for CO adsorption. The behavior of the Pt1Cu2 catalyst was much better in terms of Q\nCO and Eonset/\n\nCO. Interestingly, regardless the approximate agreement in their Eonset/\n\nCO, the Pt1Cu3 and Pt1Cu4 catalysts retained much lower Q\nCO than that of the Pt1Cu0 catalyst which highlighted the geometrical influence in retarding the adsorption of poisoning CO at the Pt surface. Table 4\n summarizes the data obtained from Fig. 9 which confirmed ca. 25, 60 and 75% improvement in the CO tolerance on the Pt surface of the Pt1Cu2, Pt1Cu3 and Pt1Cu4 catalysts, respectively. This confirmed the structural (third body) enhancement for the Pt1Cu2, Pt1Cu3, and Pt1Cu4 catalysts. Besides, Cu could also facilitate the oxidative removal of CO at lower potentials (electronic impact) at the Pt1Cu1 and Pt1Cu2 catalysts as respectively shown from the \u22120.17 and \u22120.22\u00a0V shift E\n\n\nonset/\n\n\nCO. Lastly, it is important to mention that although the Pt1Cu4 catalyst did not show any electronic enhancement, it acquired the highest activity and stability toward FAOR that originated solely from its structural (third body) enhancement effect.A PtxCuy binary catalyst was endorsed for efficient FAOR. The molar ratio of Pt4+ and Cu2+ ions in the deposition bath influenced, to a high degree, the catalytic performance and the enhancement mechanism toward FAOR. The Pt1Cu4 catalyst retained the highest catalytic activity (with up to ca. 6 times increase in the I\np\nd/I\np\nind index, 4 times increase in the I\np\nd\n/I\np\nb index and \u2212336\u00a0mV shift in E\nonset) of FAOR. This associated critical improvement in the catalytic stability that appeared in maintaining the highest current density and lowest current decay during prolonged electrolysis at 0.2\u00a0V, comparing to all other inspected catalysts. Based on the EIS and the CO stripping measurements, the catalytic enhancement of the Pt1Cu4 catalyst arose principally from a structural (third body) effect.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Supplementary data to this article can be found online at https://doi.org/10.1016/j.jscs.2022.101437.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  A propitious binary catalyst composed of Pt and Cu which were electrodeposited simultaneously onto a glassy carbon (GC) substrate was recommended for the formic acid (FA) electro-oxidation reaction (FAOR); the principal anodic reaction in the direct FA fuel cells (DFAFCs). The simultaneous co-electrodeposition of Pt and Cu in the catalyst provided an opportunity to tune the geometric functionality of the catalyst to resist the adsorption of poisoning CO at the Pt surface that represented the major impediment for DFAFCs marketing. The catalytic activity of the catalyst toward FAOR was significantly influenced by the (Pt4+/Cu2+) molar ratio of the electrolyte during electrodeposition, which also affected the surface coverage of Pt and Cu in the catalyst. Interestingly, with a molar (Pt4+/Cu2+) ratio of (1:4), the catalyst sustained superior (3.58 compared to 0.65 obtained at the pristine Pt/GC catalyst) activity for FAOR, concurrently with up to four-times (0.73 compared to 0.18 obtained at the pristine Pt/GC catalyst) improvement in the catalytic tolerance against CO poisoning. This associated, surprisingly, a negative shift of ca. 336\u00a0mV in the onset potential of FAOR, in an indication for the competitiveness of the catalyst to minimize superfluous polarizations in DFAFCs. Furthermore, it offered a better (ended up with 20% loss in the activity) stability for continuous (1\u00a0h) electrolysis than pristine Pt/GC catalyst (the loss reached 35%). The impedance and CO stripping measurements together excluded the electronic element but confirmed the geometrical influence in the catalytic enhancement.\n               "}
{"full_text": "\n\n\n\n\n\n\n\nSubject\nChemical Engineering: Catalysis\n\n\nSpecific subject area\nDevelopment of catalyst for green transformation of waste feedstock's to valuable chemicals\n\n\nType of data\nFigures\n\n\nHow the data were acquired\nXRD (Bruker D8 Discover), TGA (SDT Q600, T.A. Instruments), NH3\u2013TPD (Micromeritics Autochem-2920), Pyridine diffuse reflectance infrared Fourier transform spectroscopy (DRIFT, Thermofisher Scientific NICOLET iS50 FTIR spectrometer), XPS (PHI5000 Version Probe III (ULVAC-PHI)), HRSEM (Hitachi S-4800) and HRTEM (JEM-2100 Plus, JEOL, Japan). Raw profiles and images were collected for the biochar, activated biochar and metal loaded activated biochar.\n\n\nData format\nRaw and processed data\n\n\nDescription of data collection\nX-ray diffraction profiles were collected in 2\u03b8 range of 10\u201390\u00b0, TGA in N2 ambience at a heating rate of 5\u00a0\u00b0C min\u22121, acidity of the catalysts was measured using NH3\u2013TPD up to 800\u00a0\u00b0C using a thermal conductivity detector (TCD).Pyridine-DRIFT spectrum was recorded at 240\u00a0\u00b0C. EDS images were obtained using HRTEM JEM-2100 Plus instrument operated at 200 kV.\n\n\nData source location\nIndian Institute of Technology Madras, Chennai, India.\n\n\nData accessibility\nRepository name: Mendeley DataData identification number (doi): http://dx.doi.org/10.17632/v38nbb9rtp.3\nDirect URL to data: http://dx.doi.org/10.17632/v38nbb9rtp.3\n\n\n\nRelated research article\nAuthors\u2019 names: L. Gurrala, M. M. Kumar, A. Yerrayya, P. Kandasamy, P. Casta\u00f1o, T. Raja, G. Pilloni, C. Paek, R. VinuTitle: Unraveling the reaction mechanism of selective C9 monomeric phenols formation from lignin using Pd-Al2O3-activated biochar catalystJournal: Bioresource Technology,\nhttp://dx.doi.org/10.1016/j.biortech.2021.126204\n\n\n\n\n\n\n\n\n\n\u2022\nThe biochar obtained from biomass pyrolysis has a significant potential to be used as a catalyst support. Furthermore, the development of composite catalysts with Lewis acidic metal oxide supported on renewable carbon is vital for the production of chemicals and fuel molecules via hydrogenolysis and hydrodeoxygenation.\n\n\n\u2022\nThe detailed characterization methodology of biochar-derived catalyst (Pd-Al/ABC) is essential for researchers working in the field of green chemistry, renewable energy, and catalysis for the production of fine chemicals.\n\n\n\u2022\nThe present data can be used to gain fundamental insights on the properties of metal loaded activated biochar catalyst. The structural understanding can be used to probe the elementary reactions occurring on the catalyst active sites at different reaction conditions, and to develop structure-activity relationships.\n\n\nThe biochar obtained from biomass pyrolysis has a significant potential to be used as a catalyst support. Furthermore, the development of composite catalysts with Lewis acidic metal oxide supported on renewable carbon is vital for the production of chemicals and fuel molecules via hydrogenolysis and hydrodeoxygenation.The detailed characterization methodology of biochar-derived catalyst (Pd-Al/ABC) is essential for researchers working in the field of green chemistry, renewable energy, and catalysis for the production of fine chemicals.The present data can be used to gain fundamental insights on the properties of metal loaded activated biochar catalyst. The structural understanding can be used to probe the elementary reactions occurring on the catalyst active sites at different reaction conditions, and to develop structure-activity relationships.The chemically activated biochar (ABC) was prepared from biochar, which was obtained as a by-product from co-pyrolysis of biomass and waste plastic [2\u20134]. Al and Pd metals were loaded on ABC sequentially using wetness impregnation method [5]. The synthesized catalyst was analyzed to gain morphological and structural information. The raw and processed data of 2\u03b8 vs intensity values are reported in Mendeley data [6]. Fig.\u00a01\n shows the powder XRD profiles of the untreated biochar, ABC and metal loaded ABC catalysts. The XRD profile of untreated biochar shows crystalline diffractions due to the inorganic constituents in it, while after chemical activation most of the diffractions are absent for ABC. Only one sharp crystalline diffraction was observed at 2\u03b8 26.6\u00b0. For Pd loaded catalyst, diffraction at 2\u03b8 39.9, 46.4, 68.0 and 81.8\u00b0 were observed. The TG mass loss and differential mass loss profiles of ABC and Pd loaded ABC in presence of nitrogen are shown in Fig.\u00a02\n, and the associated data is presented in Mendeley data [6]. The major mass loss regimes were observed at 300-450\u00a0\u00b0C and 500-700\u00a0\u00b0C in the derivative mass loss profiles of ABC, 2Pd/ABC and 2Pd-Al/ABC. The mass loss at 850\u00a0\u00b0C follows the trend: 2Pd/ABC (19.2%) \u223c 2Pd-Al/ABC (19.1%) > ABC (15.2%). Fig.\u00a03\na shows the XPS survey spectrum of reduced 2Pd-5Al/ABC catalyst. The carbon 1s peak at 284.5 eV (CI) was taken as the reference [7]. The deconvoluted C1s spectrum shows different carbons corresponding to binding energies (BE) \u223c285.0, \u223c286.5, \u223c287.6 and \u223c289.1 eV, and these are labelled as C1, C2, C3 and C4, respectively. The relative area % of each of these deconvoluted peaks corresponding to different carbons are CI (10.8%), C1 (78.1%), C2 (5.4%), C3 (2.6%), C4 (3.1%). Similarly, the XPS of Pd3d was further deconvoluted to 5/2 and 3/2 of Pd0 and Pd2+, respectively. The BE values for Pd3d are 341.2, 342.9, 335.4, and 336.0 eV. The XPS data of binding energy vs intensity for the survey spectrum, C 1s and Pd 3d are available in Mendeley data [6].\nFig.\u00a04\n shows the HRTEM images of the ABC, Pd/ABC, and 2Pd-5Al/ABC catalysts at different magnification. The lattice fringes corresponding to Pd and Al2O3 are also displayed. Fig.\u00a05\n depicts the STEM image of 2Pd-5Al/ABC and elemental mapping images of C, Al, O, and Pd. The EDS spectrum shows the elemental composition of the loaded Al and Pd metals, which is \u223c5% and \u223c2%, respectively. Fig.\u00a06\n shows the TCD signal of NH3 evolved from temperature programed desorption from ABC and metal loaded ABC. The actual data of temperature vs TCD signal for the different catalysts are presented in Mendeley Data [6]. Three desorption peaks in the range of 100\u2013250\u00a0\u00b0C (D1), 250-500/550\u00a0\u00b0C (D2) and >500/550\u00a0\u00b0C (D3) are observed. TCD signal of ABC support without adsorbing NH3 (dashed line) using NH3-TPD shows desorption peak above 500\u00a0\u00b0C. The pyridine-DRIFT analysis of ABC and Pd-loaded catalysts are shown in Fig.\u00a07. A broad peak was observed at 1440\u00a0cm\u22121 in all pyridine-DRIFT profiles. The processed FTIR data of wavenumber vs transmittance are given in Mendeley data [6].\nA Bruker D8 Discover diffractometer was used to collect powder X-ray diffraction (XRD) patterns. An integrated multi-mode EIGER2 detector with Ni filter and CuK\u03b1 (\u03bb\u00a0=\u00a01.5406 \u00c5) operated at 40 kV, 30 mA was used. XRD profiles were recorded at a scan speed of 0.3\u00b0 s\u22121 with a step size 0.02\u00b0. The 2\u03b8 range was 10\u201390\u00b0. The mass loss profiles of the catalysts were recorded using T.A. Instruments SDT Q600 thermogravimetric analyzer equipped with internal air cooling unit and horizontal sample holder. The measurement was carried out in N2 ambience with 100\u00a0mL min\u22121 flow rate. Typically, 5\u00a0\u00b1\u00a00.4\u00a0mg of the catalyst sample was taken in an alumina cup, and heated up to 850\u00a0\u00b0C at a heating rate of 5\u00a0\u00b0C\u00a0min\u22121. An empty sample cup was loaded in the reference pan. The mass loss was continuously monitored, and the derivative mass loss was calculated to understand the regimes of major decomposition and the temperature corresponding to maximum mass loss rate.A Micromeritics Autochem-2920 instrument was used to determine the acidity of the catalysts using ammonia temperature programmed desorption (NH3\u2013TPD) method. The typical steps involved in the TPD measurement include: (a) activation of the catalyst at 300\u00a0\u00b0C in He ambience (40\u00a0mL min\u22121) for 60\u00a0min, (b) decrease in temperature to 50\u00a0\u00b0C, (c) adsorption of ammonia gas of 10% concentration in He (30\u00a0mL min\u22121) for 30\u00a0min at 50\u00a0\u00b0C, (d) evacuation of the physisorbed ammonia by flushing with He (30\u00a0mL min\u22121) for 60 min at 100\u00a0\u00b0C, and (e) monitoring the release of chemisorbed ammonia by increasing the temperature to 750\u00a0\u00b0C at 10\u00a0\u00b0C min\u22121 in continuous He flow (40\u00a0mL min\u22121). The concentration of the desorbed ammonia was measured using a thermal conductivity detector (TCD).A Thermofisher Scientific NICOLET iS50 FTIR spectrometer was used to conduct pyridine diffuse reflectance infrared Fourier transform (DRIFT) study to assess the type of acid sites in the catalyst. The catalyst sample was taken in a high vacuum cell provided by Harrick Scientific Products, and assembled with the spectrometer. The collection of a DRIFT spectrum involved the following steps: (a) initial activation of the sample in N2 ambience (20\u00a0mL min\u22121) at 300\u00a0\u00b0C followed by cooling it to 100\u00a0\u00b0C, (b) recording a baseline spectrum before pyridine adsorption at 100\u00a0\u00b0C, (c) injection of 30\u00a0\u00b5L pyridine while cooling down the sample cell from 100 to 50\u00a0\u00b0C, (d) evacuating the physisorbed pyridine by increasing the temperature to 100\u00a0\u00b0C and maintaining it for 60 min, and (e) raising the temperature further to 240\u00a0\u00b0C, and recording the DRIFT spectrum after 15\u00a0min at 240\u00a0\u00b0C. The final pyridine adsorption spectra was obtained by subtracting the baseline spectrum obtained at 100\u00a0\u00b0C from that obtained at 240\u00a0\u00b0C.High-resolution scanning electron microscopy (HRSEM) images of the catalysts were obtained using Hitachi S-4800 HRSEM, which was operated in the voltage range of 0.5\u201330\u00a0kV and current of 10\u00a0\u00b5A. A JEM-2100 Plus instrument (JEOL, Japan) operated at 200\u00a0kV was used to record high-resolution transmission electron microscopy (HRTEM) images of the catalysts. Typically, \u223c1\u00a0mg of sample was finely dispersed in isopropyl alcohol of 30\u00a0mL using ultrasonication. This highly dispersed sample was drop casted on a carbon-coated copper grid. Finally, isopropyl alcohol was dried at room temperature for 48\u00a0h. X-ray photoelectron spectroscopy (XPS) measurements were performed using PHI5000 Version Probe III (ULVAC-PHI). X-ray core level spectra were recorded using Al K\u03b1 radiation (h\u028b\u00a0=\u00a01486.6\u00a0eV). The carbon 1s peak at 284.5\u00a0eV was taken as the reference to determine the binding energy values of various elements in the catalyst.Not applicable.\nLakshmiprasad Gurrala: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Visualization, Writing \u2013 original draft, Writing \u2013 review & editing; M. Midhun Kumar: Methodology, Validation, Formal analysis; Changyub Paek: Methodology, Visualization, Funding; R. Vinu: Conceptualization, Methodology, Formal analysis, Visualization, Writing \u2013 original draft, Writing \u2013 review & editing, Resources, Supervision, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The corresponding author thanks ExxonMobil Research and Engineering (EMRE), U.S.A., for the research funding to IIT Madras. The National Center for Combustion Research and Development at IIT Madras is funded by Department of Science and Technology, Govt. of India.", "descript": "\n                  This article presents experimental data on the techniques used for the characterization of Pd-Al2O3 supported on activated biochar (2Pd-5Al/ABC) catalyst. The reported data is collected as a part of the research on the 2Pd-5Al/ABC catalyst used for lignin hydrogenolysis [1]. The data on X-ray powder diffraction, ammonia-temperature programmed desorption, pyridine diffuse reflectance infrared Fourier transform spectroscopy, and high-resolution scanning electron microscopy of various catalysts are valuable to study the changes in surface morphology and acidity upon metal loading. The data from thermogravimetric analysis, X-ray photoelectron spectroscopy, and scanning electron microscopy-energy dispersive X-ray spectroscopy are also provided to understand the thermal stability, ionic state of various metals and elemental composition of the catalyst, respectively. The data provided can be used for developing novel catalysts from renewable biochar, and the characterization of noble metal-metal oxide loaded catalysts can aid researchers to design composite catalytic materials for various applications.\n               "}
{"full_text": "Access to safe and clean drinking water has become a major source of contention and concern across the world. People in most parts of the world are having difficulty getting sufficient drinkable water. Water contamination has become more and more of a concern. Water pollution is a serious issue in developing countries with the majority of water in rivers and streams polluted [1]. Contamination and lack of drinkable water caused by microbiological and chemical pollution are major issues in rural areas all over the world. According to the WHO, the majority of people of developing nations suffer from health complications linked to microbiologically or chemically polluting potable water. The primary issue is generally the microbiological purity of drinkable water. The term \u201cgroundwater\u201d refers to an important supply of drinking water. Water pollution has resulted in the spread of waterborne diseases [2,3].Water pollution due to chemicals is either caused by organic or inorganic compounds. The presence of organic compounds is the primary cause of major drinking water supply issues. Synthetic chemical compounds such as dyes and pigments, insecticides, herbicides, nitroarenes and chlorinated compounds such as trihalomethans and their derivatives are examples of organic compounds [4]. Water could be polluted due to soluble organic compounds or the presence of too many inorganic ions such as sulfates, nitrates, mercury, cadmium, arsenic, lead, and other heavy metal ions. The mixture of mono-aromatic and di-aromatic benzene based hydrocarbon compounds is the common organic contaminant of natural waters [5].Nitrophenols and dyes are considered as water contaminants, which have high-quality biological and chemical stability and elevated toxicity [6]. Consequently, numerous techniques were developed for their oxidative and/or reduction. For instance, for the oxidative degradation, Fenton reactions are generally used for nitrophenols and dyes using hydrogen peroxide [7,8]. Currently, the catalytic reduction of nitrophenols and dyes in aqueous media using borohydride as a source of reducing agent is considered as a green and alternative facile method to resolve such ecological issue and gains increasing consideration. Moreover, the product of 4-nitrophenol reduction (aminophenol) is a fundamental intermediate in the synthesis of plasticizers, drugs, organic dyes, and hair dyeing agents [7,9].Aromatic amino group compounds and their precursors are important chemical raw materials that are used in the production of medicines and other fine products [10]. China produces 200,000 tons of 4-aminophenol per year with an annual demand increase of 8%. The pharmaceutical sector consumes roughly 85% of 4-aminophenol, which is used to synthesize benorilate (4-acetamidophenyl-2-acetoxybenzoate), clofibrate, and paracetamol (4-acetamidophenol). Benorilate and paracetamol, in particular, are potential antipyretic and analgesic replacements that do not have the same side effects as aspirin and phenacetin. As a result, they are an important component of non-steroidal anti-inflammatory medications all over the world. The catalytic reduction of nitro derivatives is the commercial approach for their manufacture. The most important easy way is the 4-nitrophenol reduction, which is for the most part accomplished by metal catalyst and reducing agent [11].Besides, 4-NP also acts as a pollutant, so it is a squeezing requirement for developing compelling and eco-friendly way to eliminate 4-NP from wastewater [12]. Conventional strategies include adsorption [13], photocatalytic and microbial degradation; [14], the electro-coagulation [15], electro-Fenton process [16], catalytic reduction by metal nanoparticles [17], and electrochemical treatment. However, all the above systems have their disservices. Among them, the metal nano-catalyst easily assists 4-nitrophenol reduction into 4-aminophenol because of its effective, practical and green eco-friendly properties [18]. Additionally, the reduction of nitro-aromatic to amines, such as 4-NP to 4-aminophenol, is a crucial step in synthetic organic and medicinal chemistry and the manufacture of numerous industrially vital chemicals [19].Metallic nanoparticles supported catalysts [20\u201322] have effectively left their promise in various fields of technological and scientific research. Presently, the world of catalysis including metallic nanoparticles has been revolutionized and is anticipating an oceanic change. Bimetallic nanomaterials are investigated widely in terms of their finer electronic, optical and catalytic properties over monometallic partners, because of the synergistic impacts between the two individual metals, alongside the across the extensive applications in sensors [23,24], plasmonics [25], catalysis [26,27], biomedical, and drug delivery [28]. Besides, the characteristics of the bimetals are firmly determined by the shape, size, types, and structure of the nanoparticles. Recently graphene oxide supported porous PtAu alloyed nanoflowers nanocomposites has been used for the catalytic reduction of 4-NP to 4-aminophenol (4-AP) by sodium borohydride [29]. Similarly, branched Pt@Ag core\u2013shell nanoparticles showed improved catalytic activity for the reduction of 4-nitrophenol [30]. Considerable attention has been given to zerovalent metal nanoparticles (like Au, Ni, Pd, Ag and Cu) due to their high stability and distinctive localized surface Plasmon resonance (SPR). Applications of noble metal nanoparticles in catalysis, optoelectronics, medicine, sensors and optics are known [31,32]. However, the remediation of toxic organic pollutants and their elimination from the aqueous environment is increasingly demanded day by day.Although smaller nanoparticles have better catalytic effectiveness in general, they agglomerate quickly due to their elevated surface energy owing to a reduction in catalytic activity. To address this problem, the reasonable way is to separate/isolate the metallic nanoparticles from one another by a synthetic substrate or physical obstruction, so they can't be in contact with one another straightforwardly. In the colloidal framework, metallic nanoparticles (NPs) are stabilized by steric stabilizers or electrostatic forces; noble metallic NPs can likewise be bolstered on chemically active oxides (for example titania, or ceria) [19,33].Here, in this work, the well-synthesized Ag and Cu-NPs on taro plant powder (Ag-Cu/TP) were fabricated. The synthesized nano-catalysts were used effectively in the catalytic reduction and hydrogenation of nitrophenols and organic dyes pollutants. Also, the study of the recyclability and stability of the prepared nano-catalysts were carried out for MO reduction.Taro-rhizomes were collected at a local market, rinsed in distilled water, then dried, and processed into powder by using a blender. Precursor's of silver nitrate and copper sulfate, as well as reducing agent (sodium borohydride) were from Merck Chemicals. Acid red 27 or Amaranth (AM), 2-nitrophenol (2-NP), 2,4,6-trinitrophenol (TNP), and 4-nitrophenol (4-NP), 4-nitroaniline (4-NA),\u00a0congo red (CR), methyl red (MR), methyl orange (MO), and Rhodamine B (RhB) were from BDH Chemicals, England.All of the compounds/chemicals purchased were analytical pure grade and were utilized directly without any additional purification.Bimetallic taro supported Ag-Cu nanocatalyst was synthesized by a simple adsorption method. The catalyst was prepared by using the following two-steps.Firstly, distilled water washed air dried 1\u00a0g powder was put in a 0.2\u00a0M solution of AgNO3 and CuSO4 mixture (50\u00a0mL each solution). Then, the powder mixture and salt solution were left for about 7\u00a0h at constant stirring in order to saturate the sites of adsorption of plant powder. Powders were removed through filtration and to remove the free ions, they were rinsed three times with distilled water (DW) and then filtered again and dried.In the second step, in order to reduce the metal ions into corresponding nanoparticles, 100\u00a0mg of dried treated powders was added to 50\u00a0mL of the reducing agent (sodium borohydride) solution (0.5\u00a0M). The color changed from brown to dark black after the treatment with NaBH4 confirming the reduction process and nanoparticles formation. To convert all ions into respective nanoparticles (Cu and Ag), the powders were left in the borohydride solution for about 1\u00a0h. The nano-catalyst (Ag-Cu/TP) after washing with DW was dried and stored for the next step.The prepared taro powder supported bimetallic silver-copper nano-catalysts (Ag-Cu/TP) were tested in eight different reactions. The different reactions include catalytic hydrogenation/reduction of nitrophenols (4-NP and 2-NP) into aminophenol, in picric acid reduction, and azo dyes MR, Acid Red 27, MO, CR and methylxanthene family RhB dye.For the nitroarenes hydrogenation, 1\u00a0mM aqueous solution of nitroarenes (2-NP, 4-NP and TNP) and 0.5\u00a0M solution of NaBH4 were prepared in a volumetric flask. In reaction, 3\u00a0mL of nitroarenes (NP or picric acid) was put in a reaction vessel (a quartz cuvette was used for analysis as a reaction vessel) to carry out its UV\u2013Vis spectra. After that, 0.5\u00a0mL of the reducing agent was added to study the variation in their UV\u2013Vis spectra. Usually, sodium borohydride addition increases the solution pH and forms nitrophenolate ions, due to which the variations in spectra occur. After NaBH4 addition, the catalyst was added (10\u00a0mg) into the reaction mixture. The decrease in \u03bb\nmax started once Ag-Cu/TP catalyst was added, which indicates the reduction of nitroarenes. The variation in \u03bb\nmax was noted and recorded (for TNP at 393\u00a0nm; for 4-NP at 401\u00a0nm, and for 2-NP at 418\u00a0nm) [34].The concentration of dyes employed for dye reduction was 0.05\u00a0mM. To reduce the dye, three mL of the chosen solution of dye was placed in a 3\u00a0mL UV cuvette and 0.5\u00a0mL reducing agent solution was added to record the spectra. After NaBH4 addition, the catalyst was added (10\u00a0mg) into the cuvette. The variation in \u03bb\nmax position and intensity at \u03bb\nmax after Ag-Cu/TP inclusion was recorded (i.e., for CR at 493\u00a0nm, for AM at 521\u00a0nm, for MO at 466\u00a0nm, for MR at 428\u00a0nm and for RhB at 555\u00a0nm).The reduction of the mixture of nitroarene was carried out by combining 1.5\u00a0mL of each reactant solution and recording the spectra. Then, 0.5\u00a0mL of reducing agent was added, and recorded its spectra. The variances in both intensity and \u03bb\nmax were recorded and monitored on a regular basis after the addition of the catalyst (10\u00a0mg).In the reduction reaction of MO, the reusability or recyclability of the Ag-Cu/TP catalyst was investigated. To investigate reusability, the Ag-Cu/TP was rinsed with distilled water three times and utilized for the next reaction at the same time.Stability study of Ag-Cu/TP catalyst was carried out at different intervals of time. The stability study was carried out on the reductive hydrogenation of 4-NP into 4-AP for up to 98\u00a0h.A chromatography glass column was developed for the dye reduction by packing with Ag-Cu/TP catalyst. The catalyst was mixed with a fixed amount of sand and then used for the reduction of dyes. For column packing, 2\u00a0g of Ag-Cu/TP was homogeneously mixed with 20\u00a0g of sea sand and then the column was packed through wet method, with cotton supports and sand at the bottom of the prepared column. The sea sand used was washed with 1\u00a0M HCl for 30\u00a0min, and then washed with 1\u00a0M NaOH, followed by washing four times with distilled water. Then, the washed sand was dried in a furnace at 400\u00a0\u00b0C for 4\u00a0h. The column top was filled for 1\u00a0cm with sea sand. The packed column was washed twice with distilled water before the reaction. Then, freshly prepared solutions of dyes mixed with sodium borohydride were fed continuously throughout the column. UV\u2013Vis spectra of the mixture eluted were carried out for each reaction using UV\u2013Vis spectrophotometer. The filled column without catalyst was used as a negative control to observe the adsorption capability of pure coarse sand.Different analysis techniques were used to characterize the Ag-Cu/TP catalyst, including SEM, FTIR, TEM, EDX, and XRD.The structural characterization of synthesized nanoparticles was investigated by using SEM. The synthesized Ag-Cu/TP was structurally characterized by using an SEM (JSM-5910, Joel) (Japan). Similarly, the elemental analysis was also carried out by using a JEOL-JSM-5910, (Japan), EDX system coupled with the SEM. For the TEM images, JEOL JEM-2100 electron microscope at 200\u00a0kV was used.Powder XRD was conducted by using a CuK source on a JEOL JDX-9C XRD analytical diffractometer (Japan). The Scherrer equation as shown below was used to figure out the average size of the catalyst crystallite.\n\n(i)\n\n\nD\n=\n\n\nK\n\u03bb\n\n/\n\n\u03b2\n\n1\n/\n2\n\n\n\ncos\n\n\u03b8\n\n\n\nIn the above equation, k represents a constant which has unity, \u03bb is the CuK\u03b1 source wavelength and equals to 1.5418\u00a0\u00c5, \u03b21/2 represents the full width at half maximum on 2\u03b8 scale, and \u03b8 is the Bragg's angle.A UV\u2013visible Shimadzu-1800 spectrophotometer was used\u00a0to conduct catalytic reduction tests on nitrophenols and dyes.The visual representation of the taro supported Ag-Cu/TP catalyst is presented in Fig.\u00a01\n. Metal ions treated powders look dark brown. The ions were absorbed by the taro powder because of the electrostatic contacts between positive metal ions and hydroxyl groups of phytochemicals present in the plant part. Besides, metal ions were absorbed by the pores of the plant powder. Because the \u2013OH of plant polyphenols reduced silver ions easily, the treated taro sample was dark brown in color [35]. To get rid of the ions that are weakly bound, which will cause leaching problems in the reduction process, the powder after metal ion treatment was washed three to four times with DW and then dried for further use at room temperature.The color of the sample immediately turned to dark black when treated with NaBH4 solution (Fig.\u00a01), indicating the reduction of metal ions to their nanoparticles and the formation of a catalyst. Metal ion reduction to their corresponding (Ag0 and Cu0) nanoparticles is shown in the given chemical reactions.\n\n\n(ii)\n\n\n2\nA\n\ng\n+\n\n+\n2\nB\n\nH\n\n4\n\u00af\n\n\n+\n6\n\nH\n2\n\nO\n\u2192\n2\nA\n\ng\n0\n\n+\n7\n\nH\n2\n\n+\n2\nB\n\n\n(\n\nO\nH\n\n)\n\n3\n\n\n\n\n\n\n\n(iii)\n\n\n2\nC\n\nu\n\n2\n+\n\n\n+\n4\nB\n\nH\n\n4\n\u00af\n\n\n+\n12\n\nH\n2\n\nO\n\u2192\n2\nC\n\nu\n0\n\n+\n14\n\nH\n2\n\n+\n4\nB\n\n\n(\n\nO\nH\n\n)\n\n3\n\n\n\n\n\nActive catalyst was prepared by treatment with NaBH4 as a reducing agent.Since catalytic activities of nanomaterials are primarily dependent on their crystal structure, shape, order, size, stability, and reusability. Materials with scattered and ordered nanoparticles frameworks have gained significant interest due to their fascinating catalytic and biological properties.The shape and surface morphology taro supported catalyst were evaluated using TEM and SEM analyses. For Ag-Cu/TP catalyst, both SEM and TEM were used for the surface morphology. As shown in Fig.\u00a02\n, the Ag-Cu/TP nano-catalyst was poly-dispersed size regularly distribution of nanoparticles on plant surfaces. The excellent catalytic effectiveness of Ag-Cu/TP was because of the uniform distribution of Ag and Cu nanoparticles to the powder surface. In addition to monodispersed particles, certain nanostructure clusters were found in Ag-Cu/TP SEM pictures, which are due to the accumulation of high metal ion concentration. Generally, the electronic and optical properties of nanomaterials/catalysts as reported previously in the literature are greatly dependent on the size, shape and nanoparticle distribution [36].As described in the catalytic reduction of nitrophenols and dyes, the excellent catalytic performance of the prepared catalyst is due to the uniform distribution of nanoparticles on the support surface.TEM images of Ag-Cu/TP showed the polydispersed and spherical-shaped smaller size nanoparticle formation on the surface of support taro. The average calculated size of the Ag-Cu/TP nanoparticles from TEM images was 2\u201333\u00a0nm [mean size\u00a0\u2248\u00a020\u00a0\u00b1\u00a013\u00a0nm] as presented in Fig.\u00a03\n.An EDX investigation was carried out for the elemental characterization of the Ag-Cu/TP catalyst. EDX is a technique which determines the composition of elements in a nanoparticle sample by detecting specific elements in the framework of a catalyst and quantifies their amount also. For the EDX pattern, the crystalline nature of the sample was clarified.EDX investigation of Ag-Cu/TP catalyst showed three distinct peaks at 0.5, 1.0, and 3.0\u00a0keV as presented in Fig.\u00a02(d). In the EDX spectrum, the peak at the 1\u00a0keV confirms the presence of copper [37], while the peak at 3\u00a0keV represents silver [38] in the Ag-Cu/TP catalyst. The weight proportion of the particular elements in the Ag-Cu/TP catalyst were 32.3% carbon, 34.28% oxygen, 13.17% copper, and 20.25% silver, as illustrated in Fig.\u00a02(d).Crystalline nature and the average size of prepared catalyst were evaluated by powder XRD pattern. The pattern of the powder XRD of Ag-Cu/TP catalyst is presented in Fig.\u00a04\n. The XRD of the prepared Ag-Cu/TP catalyst demonstrated five different diffraction peaks at a 2\u03b8 position of 38.58\u00b0, 43.67\u00b0, 50.67\u00b0, 64.69\u00b0 and 75.8\u00b0 as depicting in Fig.\u00a04. Peaks at 38.58, 64.69, and 75.8 correlate to Brag peaks [111], [220], and [311], respectively, which coincide with face centered cubic (fcc) structural features of supported Ag nano-catalyst [39,40]. The CuNPs in Ag-Cu/GP were identified at 43.67 and 50.67, correspond to Bragg's diffraction (111) and (200), respectively, and correlate to Cu nanoparticles' fcc structure [37].After the catalytic hydrogenation/reduction of 4-NP, an XRD study of Ag-Cu/TP was also performed as presented in Fig.\u00a04 (XRD pattern after catalysis). In the XRD pattern of Ag-Cu/TP, the most intense peak appears at 35.42\u00b0 which is related to the Bragg's index of (\u2212111) of CuO nanoparticles. This confirmed the pure monoclinic phase of CuO formation [41,42]. The intensities of XRD peaks of Ag and Cu nanoparticles decreased after catalysis, which showed the formation of oxides after five cycle reduction of MO dye. Some unassigned (\u2217) peaks were also observed, which might be because of the biomolecules crystallization.As demonstrated in Fig.\u00a05\n, the functional groups and chemical nature of the produced Ag\u2013Cu/TP catalyst were determined using the FT\u2013IR spectra. FT\u2013IR analysis of the pure taro powder was also carried out. Absorption bands at 3320, 2917, 3320, 2917, 16510, 1470, 1187 and 1015\u00a0cm\u22121 have been seen in the FT\u2013IR spectrum of pure taro powder (Fig.\u00a05), which correspond to hydroxyl (O\u2013H) stretching vibration, asymmetrical stretching of methylene (C\u2013H), the primary amine stretching vibration (N\u2013H), asymmetrical stretching of ether (C\u2013O\u2013C) group, bending of \u2013OH, C\u2013H and C\u2013OH side group vibrations [43\u201345].In the spectrum of Ag\u2013Cu/TP catalyst, the typical taro peaks were observed with a slight shift along with unique peak at 608\u00a0cm\u22121. The new peak observed at 608\u00a0cm\u22121 corresponds to the metal\u2013oxygen (M\u2212O) bond and confirms that Cu\u2013Ag nanoparticles were synthesized on the taro support [26]. Furthermore, when the pure taro powder and the supported Ag\u2013Cu/TP catalyst were examined, the bands of carboxyl, hydroxyl, and amine in the support catalyst appeared broad and slightly shifted in the catalyst, demonstrating the interaction between the metallic component and the taro powder. These findings reveal that Cu\u2013Ag nanoparticles were successfully produced over the taro powder surface.Animals, birds, plants, and marine life are all at risk from nitroaromatic molecules because of their toxicity. Due to its challenging decomposition at such trace levels, 4-NP poses harm to the ecosystem even at lower quantities. Catalytic reduction/hydrogenation is the best approach to transfer 4-nitrophenol into the beneficial product 4-aminophenol, which is then used in the manufacturing of antipyretic and analgesic medicines including acetaminophenol, acetanilide, and phenacetin [46]. As a result, reducing 4-NP is important in order to fulfill the needs for 4-aminophenol.The reducing agent sodium borohydride has been reported in researches to be unable to reduce 4-NP completely. Because borohydride ions function as powerful reducing species in the aqueous phase (E for H3BO3/BH4 = \u22121.33 V), 4-NP reduction by NaBH4 is thermodynamically favored (E for 4-NP/4-AP is \u22120.76 V). However, resulting in significant difference and kinetic restriction between the borohydride and nitrophenolate ions, the reduction is very slower kinetically, and the chance of this reaction decreases. To surpass the kinetic barrier, nanomaterials catalyzed/promote the hydrogenation process of 4-NP by allowing electron transport from the donor (borohydride ions) to the acceptor (i.e., nitrophenolate ions) [47]. Catalytic reduction of 4-NP to 4-AP followed a pseudo-first-order kinetic. When both borohydride and nitrophenolate ions are adsorbed on the surface of the active site of the catalyst, then electron transfer starts from negative borohydride to nitrophenolate ions. As a result, Ag and Cu nanoparticles boost up the reduction reaction by reducing the energy of activation and acting as a catalyst. In the presence of NaBH4 reducing agent, the transformation of NP to AP is a six-electron transfer process, as indicated in Scheme 1\n. As a result, the 4-NP reduction could not be achieved effectively without the use of a catalyst, as demonstrated experimentally by utilizing sodium borohydride alone in the previous study [35]. Accordingly, the reductive conversion of 4-NP required the use of an active catalyst. The chemical pathway for the transformation of NP to AP by sodium borohydride by the use of Ag-Cu/TP catalyst is shown in Scheme 1.The red curve in the spectrum shows that the solution of 4-NP has a significant absorption maxima (\u03bb\nmax) at 316 nm in the ultraviolet spectrum, as shown in Fig. 6\n(a). With the addition of 0.5 mL sodium borohydride, the sharp peak of 4-NP shifted from 316 to 401 nm. The formation of phenolate ions from 4-nitrophenol caused the peak position of 4-NP to shift since NaBH4 raised the solution pH, causing nitrophenolate ions to form. Due to the existence of nitrophenolate, the peak intensity and position at 401 nm remained static even after 20 min without the use of a catalyst. After catalyst addition, however, the height of the peak at 401 nm began to drop. The new peak in the reaction appeared at 302 nm (Fig. 6(a)). The synthesis of 4-AP was confirmed due to the drop in peak intensity at 401 nm and a rise in the intensity at 302 nm [48,49].The same Ag-Cu/TP catalyst was also utilized in the hydrogenation of 2-NP to 2-AP. The UV\u2013Vis spectrum of a dilute 2-NP have two prominent peaks at 356 nm and 273 nm, as shown by the red curve (Fig. 6(b)). With the introduction of NaBH4 to the sample vial containing 3 mL of 2-NP, a change in the peak position of 2-NP was noticed from 356 nm with a rising intensity to 418 nm (Fig. 6(b)). Until the catalyst was added, both intensity and position at 273 and 418 nm remained constant. Once Ag-Cu/TP catalyst was introduced, the peak at 418 nm began to decline. During this time, the peak in the UV area shifted from 273 to 294 nm. Thus, the reduction of 2-NP and generation of 2-aminophenolate ions are confirmed by the emergence of a second peak at 294 nm and a drop in the height of the main peak at 419 nm [50,51].From the peak of the UV\u2013Vis spectrum, the percent reduction of the NP reaction was measured using equation (iv)\n\n\n(iv)\n\n\nPercent Reduction\n=\n100\n\u2212\n\n(\n\n\nA\n\nt\n\n\n\n\u00d7\n100\n/\n\nA\n0\n\n\n)\n\n\n\n\nwhere A\n\n0\n signifies the absorbance maximum at time zero, and A\n\nt\n denotes the absorbance at a specific interval of time of reading.4-NP and 2-NP reduction processes by NaBH4 follow pseudo first order kinetic reaction. For the computation of the apparent rate constant (k\n\napp\n), the graph of reduction time (in minutes) versus ln (A\n\nt\n/A\n\n0\n) was used (Fig. 6(d)).The equation for the pseudo-first order is given below\n\n(v)\n\nr\n=\nl\nn\n\nC\nt\n\n/\n\nC\n0\n\n=\n\nd\nc\n\n/\n\nd\nt\n\n=\n\u2212\n\n\nK\n\na\np\np\n\n\n\nt\n\n\nwhere the reduction rate is represented by r; time of reaction is represented by t; while concentration by c; C\n0 represents initial concentration at time t = 0; C\n\nt\n stands for the concentration at specific interval t.Since here in this case, both reactants (4-NP and 2-NP) have a bright color and provide a maximum peak in the visual region, the reaction rate can be presented in comparative absorption intensity. Thus, the pseudo-first order in terms of adsorption can be expressed as:\n\n(vi)\n\nr\n=\nl\nn\n\nC\nt\n\n/\n\nC\n0\n\n=\nl\nn\n\n\nA\nt\n\n/\n\nA\n0\n\n=\n\u2212\n\n\nK\n\na\np\np\n\n\n\nt\n\n\n\nThe kinetic data for 4-NP and 2-NP and the graph between time versus ln(A\n\nt\n/A\n\n0\n) are presented in Fig. 6(d), reduced by NaBH4 and catalyzed by Ag-Cu/TP. The Ag-Cu/TP catalyst showed outstanding performance and 99.4% of 4-NP was reduced just in 5 min, while 97 percent reduction took 17 min with 2-NP (Fig. 6(c)). The calculated rate constant for Ag-Cu/TP at room temperature was 1.01 \u00d7 10\u22123s\u22121 for 2-NP reaction and 5.24 \u00d7 10\u22123s\u22121 for 4-NP as presented in Table 1\n. Impressively, the catalytic effectiveness of the prepared catalyst (Ag-Cu/TP) showed higher catalytic performance than recently reported bimetallic platinum-rhodium alloyed nano-multipods, where the reduction efficiency of 4-NP reaches to 94.02% in 10 min [52].The number of moles of nitrophenols (reactant) used in a known specific reaction per mol of catalyst (Ag-Cu/TP) per unit of time is called the Turnover frequency (TOF) as expressed by Eq. (vii) [53]. The computed TOF for the reaction of nitrophenols (4-NP and 2-NP) is revealed in Table 1.\n\n(vii)\n\n\nTOF\n=\n\n\nNo of mole of Reactant\n\n(\nNP\n)\n\n\nNo of mole of Catalyst\n\n\u00d7\n\n\nY\ni\ne\nl\nd\n\n\nT\ni\nm\ne\n\n\n\n\n\n\nThe higher TOF may be due to the dispersion of homogeneous and uniform silver and copper nanoparticles on taro support. The excellent catalytic performance of our catalyst (Ag-Cu/TP) was due to the uniform dispersion of metallic nanoparticles on the surface of support.Picric acid or TNP is a highly explosive and unstable chemical. Picric acid, which is ignited by flames, sparks or even heat, is commonly employed in military explosives and also has a rich history of nonmilitary and military applications. When subjected to friction or shock, heat or flame dried-out picric acid could burst, and it should be considered as an explosive material. Toxicity of picric acid is mostly acute; nevertheless, any long-term consequences, such as mutagenicity, have been described. Picric acid is a cutaneous sensitizer and a potent eye irritant in animals [60]. Nitro-substances like TNP have a very deficient -electron network resulting in greater xenobiotic characteristics due to their peculiar electron-withdrawing nature. Biodegradation is difficult for nitro-aromatic and other nitrogen-containing molecules. Because of the low electron density of the nitro group, electrophilic assault in the benzene ring of nitroarenes has become more difficult, that is the first phase in the reduction/degradation. As a result, di-nitroaromatic and tri-nitroaromatic molecules, such as picric acid, undergo the first reductive process.For the reductive hydrogenation of TNP, the same Ag-Cu/TP catalyst was utilized. TNP's aqueous solution has a yellowish color. The main absorption peak for TNP is illustrated at 357 nm as revealed in Fig. 7\n(a). After adding NaBH4 solution, the peak location was shifted from 357 to 393 nm, with increase in intensity. It was observed that with the addition of NaBH4, the color turned from light yellowish to dark yellow. The presence of phenolate ions in the reaction cuvette is evidenced by a change in peak \u03bb\nmax from 357 to 393 nm after adding NaBH4. The height of a peak \u03bb\nmax at 393 nm for TNP stayed unchanged for an hour without catalyst. This demonstrates that the reducing agent NaBH4 could not reduce TNP on its own. As a result, complete TNP reduction without a specific catalyst is incredibly challenging. The peak \u03bb\nmax at 393 nm began to decline after adding a catalyst to the cuvette. However, around 234 nm, a new peak appeared. With time, the unique peak that appeared at 234 nm grew in height. The catalytic reduction of TNP caused nitrophenolate ions to drop in intensity at 393 nm and aminophenolate ions to increase in intensity at 234 nm (Fig. 7(a)). The synthesis of aromatic amine (aminophenolate) is responsible for novel peak formation in the UV region at 234 nm [61]. TNP reduction was assisted by the Ag-Cu/TP, which occured at its surface by allowing the transfer of electrons from borohydride to the phenolate ions to cross the barrier of kinetic. TNP reduction follows a pseudo first order with regard to the substrate since it is independent of NaBH4 amount. A graph of ln (A\n\nt\n/A\n\n0\n) versus time of reduction was used to determine the rate constant (in second) for the pseudo-first-order kinetic of TNP reduction using Eq. (vi).Ag-Cu/TP catalyst demonstrated excellent activity and reduced 92.5% of TNP in 7 min as presented in Fig. 7(c). The k\n\napp\n calculated for TNP by Ag-Cu/TP from the slope was 5.62 \u00d7 10\u22123 s\u22121 (Fig. 7(d)). Ag-Cu/TP has a greater k\n\napp\n than our latest published Ag@CAF, which has a k\n\napp\n of 8.97 10\u22124\ns\u22121 for TNP [62].Dyes are currently commonly used and synthesized in a wide range of industries, including textiles, cosmetics, paper and pulp, printing, and so on. However, it was reported that about 10\u201315% of dye chemicals are lost into the aquatic environment, posing a threat to organisms as well as public health. Organic synthetic dyes are resistant to sunlight, base and acid, and because of the benzene ring in their molecules they have the potential to be genotoxic as well as detrimental to marine life. It produces an ecological imbalance by accumulation in water [63]. As a result, dyes must be removed and eliminated from the aquatic environment.The traditionally Amaranthus plant was used to produce AM dye. In recent years, it has been chemically manufactured as a trisodium salt on an industrial scale. Food Red 9 and Acid Red 27 are common names for the chemically manufactured AM dye. The carcinogenic impact of AM dye on albino rats has been studied in the literature. AM dye at 10 times the Acceptable Daily Intake causes lethal abnormalities, and levels of 47\u00a0mg\u00a0kg\u22121\u00a0by bodyweight of AM dye can affect liver function [64].The aqueous AM dyes produced strong single peak at 521\u00a0nm and smaller peaks at 330, 282, and 242\u00a0nm in the UV and visible regions of the UV\u2013Vis spectrum, as illustrated by the red curve (Fig.\u00a07(b)). After adding the reducing agent sodium borohydride, no change was observed in peak position and the peak remained at 521\u00a0nm. Even in the absence of a catalyst, the height and position of the \u03bb\nmax at 521\u00a0nm remained unchanged. However, after the catalyst was added to the cuvette, the intensity of the peak began to fall steadily. The prominent \u03bb\nmax at 521\u00a0nm vanished after the catalyst was introduced, while simultaneously a new peak formed at 267\u00a0nm (Fig.\u00a07(b)). The beginning of AM reduction was confirmed by the gradual peak decline and vanishing at 521\u00a0nm, and novel peak formation at 267\u00a0nm. Since the reduction of AM is independent of NaHB4, so the reaction is pseudo-first order. Ag-Cu/TP catalyst reduced 76.54% AM dye in 9\u00a0min as shown (Fig.\u00a07(c)). The k\n\napp\n calculated for AM by Ag-Cu/TP was 9.87\u00a0\u00d7\u00a010\u22124\u00a0s\u22121 (Fig.\u00a07(d)). Hence, Ag-Cu/TP displayed good action for AM reduction.The disintegration of MO dye is of major interest because it poses sever environmental and health concerns. MO could be reduced to smaller organic molecules by NaBH4 as a reducing agent. However, the reduction of MO by NaBH4 is very slow and time-consuming. Without a catalyst, the reduction ofMO by the reducing agent sodium borohydride is kinetically unfavorable but feasible thermodynamically. Noble metal nanoparticles like Au and Ag due to their distinctive SPR character and increased specific area are now widely used as efficient catalysts for the reductive degradation of MO dye.Firstly, 3\u00a0mL of aqueous dye solution was taken in a cuvette. After that, 0.5\u00a0mL of aqueous NaBH4 was added and their spectra in the 200\u2013800\u00a0nm range was then measured. Maximum absorption was observed at 466 and 280\u00a0nm for MO solution. Even after 20\u00a0min of NaBH4 addition, the absorbance at 466\u00a0nm remained unchanged. However, after adding Ag-Cu/TP, the absorbance at both 466 and 280\u00a0nm began to change (Fig.\u00a08\n(a)). With increased intensity, the peak at 466\u00a0nm consistently reduced, whereas the \u03bb\nmax at 280\u00a0nm shifted and reached 252\u00a0nm to some extent (Fig.\u00a08(a)). MO reduction was related to the steady decline and eventual vanishing of peak intensity at 466\u00a0nm. The primary amino group molecule (hydrazine derivative) formed during MO reduction resulted in the new peak appearance at 252\u00a0nm [65,66]. NaBH4 degrades MO at the azo (\u2013N=N\u2013) site by forming smaller amine derivative molecules in the presence of the prepared effective catalyst. The absorption bands of the \u2013NH2 group of the resultant products are attributed to the peak at 252\u00a0nm that was generated during the MO reaction. The possible process given in Scheme 2\n is most likely to be used in the reductive breakdown of MO on the Ag-Cu/TP catalyst. MO reduction requires electron transport from the source borohydride ions to the recipient dye molecule. Both the MO molecules and borohydride ions are initially adsorbed on the catalytic surface (Ag and Cu nanocatalysts). In this case, metal NPs serve as electron relays enabling electron transport between nucleophilic borohydride ions and electrophilic MO molecules [67]. Previous studies have found that metallic NPs in water accelerate the passage of electrons from borohydride ions to the dye resulting in formation of BO3\n3\u2212 ions [68]. These electron transfers might continue indefinitely resulting in further dye molecule degradation [69].The high reactivity of the prepared catalyst could be due to the high surface area and irregular morphology of the nano-catalyst (as shown in SEM images), which facilitate charge transfer and allow the depletion mechanism to overcome kinetic barriers. The reduction of MO by the reducing agent NaHB4 follows a pseudo-first-order kinetic.Eq. (iv) was used to calculate the percent degradation of MO using Ag-Cu/TP catalyst. Here in this report, as presented in Fig.\u00a08(c), the prepared catalyst (Ag-Cu/TP) showed 96% MO reduction in just 9\u00a0min. The calculated pseudo-first-order reaction rate constant for this reaction from the slope was 2.91\u00a0\u00d7\u00a010\u22123s\u22121 for Ag-Cu/TP at room temperature (Fig.\u00a08(d)).MR is used as an indicator marker in microbiology to detect bacteria that produces stable acids from glucose using mixed acid fermentation processes. Besides that, it's been discovered to be a great booster for the sonochemical disintegration of polychlorinated. When inhaled or ingested, MR produces sensitivity in the hair, skin, and eyes, as well as intestinal inflammation. Because of the breakdown of its molecule into toxic (carcinogenic and mutagenic) aromatic amines, the presence of MR in the aquatic system causes aesthetic issues that have a detrimental impact on public health [70]. MR produces visual and skin sensitivity if it is inhaled or ingested as well as digestive tract irritation.Our prepared catalyst was also used in the catalytic degradation of MR. Fig.\u00a08(b) shows the absorption maxima of aqueous MR solution at 428\u00a0nm. With the introduction of NaBH4, the peak height at 428\u00a0nm remained constant. With Ag-Cu/TP addition, the intensity of the absorption peak declined steadily. However, as seen in Fig.\u00a08(b), the additional two peaks at 307 and 246\u00a0nm appeared with time. The beginning of MR reduction at the azo region by producing smaller aromatic amines resulted in a drop in \u03bb\nmax at 428\u00a0nm and the raising of two separate peaks at 307 and 246\u00a0nm.Here, Ag-Cu/TP showed 97% MR reduction in just 9\u00a0min as presented in Fig.\u00a08(c). The pseudo-first order reaction rate constant was calculated through Eq. (vi). The rate constant of MR at normal temperature evaluated from the slope was 4.126\u00a0\u00d7\u00a010\u22123s\u22121 for Ag-Cu/TP. The plausible mechanism for reductive degradation of MR by Ag-Cu/TP is presented in Scheme 2.CR containing wastewater is brightly colored; its release into the aquatic environment inhibits photosynthesis by reducing light penetration, and is also cytotoxic to a wide range of aquatic creatures. As a result, the elimination of CR from aquatic environments is a major concern.Two major peaks were observed at 493\u00a0nm and 346\u00a0nm in an aqueous solution of CR. Without any active catalyst, the absorption peaks remained unchanged. After Ag-Cu/TP addition, the intensities at both points (493 and 346\u00a0nm) slowly decreased with time. After the catalyst was added, a novel peak at 250\u00a0nm appeared, as seen in Fig.\u00a09\n(a). CR reduction was initiated by the fall in intensities at both positions of 493 and 346\u00a0nm as well as the rise of a novel peak maxima at 250\u00a0nm. With the addition of a catalyst (Ag-Cu/TP), sodium borohydride degraded the CR molecules by producing amines of the aromatic nature with a small molecular weight. As a result, the peak at 250\u00a0nm that increased in height with time was due to the reducing of CR molecules [71]. This demonstrates that the presence of the prepared catalyst caused the deterioration of CR by NaBH4.An Eq. (iv) was used to calculate the percent reduction. Ag-Cu/TP degraded 89 percent of CR molecules in solution in 8\u00a0min, as shown in Fig.\u00a09(c). Since NaBH4 reduction of CR is a pseudo-first order [37], therefore, the rate of reaction of CR was 2.73\u00a0\u00d7\u00a010\u22123s\u22121 for Ag-Cu/TP catalyst (Fig.\u00a09(d)).In this study, the degradation rate of CR dye was higher than the earlier reported reduction of CR using filter paper coated Chitosan loaded CuNPs, which took 13\u00a0min for 0.012\u00a0mM CR reduction [47]. The reductive degradation of CR on silver and copper-decorated catalyst is probably presented by the plausible mechanism presented in Scheme 2.Rhodamine B is a fabric dye that belongs to the methyl-xanthine family. RhB is commonly employed in cell luminescence as a coloring agent; a tracer reagent in water quality studies, and a color tracer in the coloring of wool, cotton, leather, colored glass, jute, silk, and sprays. RhB is employed as a fluorescent sensing and colorimetric reagent in possible metal ions diagnosis due to some advantageous chemical characteristics [72,73]. If inhaled, RhB has harmful consequences for human health. Carcinogenicity, acute and neurological toxicity, development and reproduction toxicity have all been reported in humans and animals in the laboratory. This dye has the potential to cause tissue melanoma. Aside from these, it causes eye and skin burning, nasal itching, vomiting, throat incinerating, and chest pain. Some countries around the globe have banned and prohibited its usage in food goods due to its numerous negative health impacts. However, because of its low cost, great stability, and color, RhB is still being used illegally in food and other areas [59,74].\nFig.\u00a09(b), shows the UV\u2013Vis reduction spectrums of RhB by using Ag-Cu/TP catalyst\u00a0and reducing agent NaBH4. The same experimental technique was employed to reduce RhB as it was for other dyes. In the visible region, an aqueous RhB solution showed a maximum peak centered at 555\u00a0nm. The sharp absorption peak stayed stable at 555\u00a0nm for a long period without an appropriate catalyst. It has also been reported that without an appropriate catalyst, NaBH4 could not effectively degrade/reduce RhB [75]. However, after 10\u00a0mg of Ag-Cu/TP catalyst addition, maximum absorbance at 555\u00a0nm was obtained to steadily decline with time. The reduction dye may be seen in the gradual drop of the major peak of RhB, and a novel peak appearance at 242\u00a0nm. As stated previously, RhB de-ethylation is caused by the utilization of a NaBH4 with an active catalyst. As a result, the decline in \u03bb\nmax at 555\u00a0nm is because of RhB reduction [76]. Ag-Cu/TP reduced RhB dye by 71.2 percent in 7\u00a0min, as shown in (Fig.\u00a09(c)). This reduction by sodium borohydride follows a pseudo-first reaction model [75]. The kinetic data and graph of time (t) versus -ln (A\n\nt\n\n/A\n\n0\n) of RhB reduction is shown in Fig.\u00a09(d). As a result, the rate constant for RhB was 2.05\u00a0\u00d7\u00a010\u22123s\u22121. The plausible reduction mechanism of RhB dye of Ag-Cu/TP catalyst is presented in Scheme 2.For further study, a particular substratewas used to study the catalytic deterioration and reduction processes of the catalyst. The catalytic degradation/hydrogenation of the catalyst was tested in this work against a solution of nitroarenes.As seen in Fig.\u00a010\n(a), the mixture of TNP and 2-NP had the highest absorption at 359\u00a0nm. The \u03bb\nmax was shifted to 393\u00a0nm after NaBH4 addition, as shown in Fig.\u00a010(a). The production of phenolate ions caused the peak to move from 359 to 393\u00a0nm. With catalyst introduction, the peak intensity at 393\u00a0nm began to drop. The conversion of nitrophenols to corresponding aminophenols was demonstrated by the progressive drop in \u03bb\nmax at 393\u00a0nm and the generation of two additional peaks at 308 and 235\u00a0nm. The complete reduction of TNP and 2-NP mixture took 6\u00a0min using Ag-Cu/TP catalyst.The mixture of TNP and 4-NA showed a \u03bb\nmax at 358\u00a0nm (Fig.\u00a010(b)). The peak maxima was shifted from 358 nm to 392\u00a0nm after NaBH4 addition, as presented in Fig.\u00a010(b). The change in peak was because of phenolate ions formation in a solution. Once the catalyst was added, the peak height at 392\u00a0nm began to drop. The transformation of nitrophenols to aminophenols was demonstrated by the progressive drop in \u03bb\nmax at 392\u00a0nm and the development of a novel peak at 299\u00a0nm. The complete reduction of TNP and 4-NA mixture took 9\u00a0min using Ag-Cu/TP catalyst.As demonstrated in Fig.\u00a010(c), the mixture of 4-NP and 4-NA exhibited three different absorptions at 215, 281 and 404\u00a0nm. After adding NaBH4, the position of the peaks was unchanged, but the strength of the maximum (\u03bb\nmax) at 404\u00a0nm rose due to the generation of phenolate ions, as seen in Fig.\u00a010(c). The catalytic reduction/hydrogenation was demonstrated by a drop in both positions at 215 and 404\u00a0nm and the transfer in peak position from 281 to 295\u00a0nm with catalyst addition. Using Ag-Cu/TP, the full reduction of mixtures of 4-NP and 4-NA took 9\u00a0min.As demonstrated in Fig.\u00a010(d), the combination of 4-NP, 4-NA, and 2-NP exhibited three strong absorbance peaks at 403, 280, and 240\u00a0nm. After adding a reducing agent, the location of the peaks was unchanged, but the strength of the \u03bb\nmax at 403\u00a0nm rose due to phenolate ion production (Fig.\u00a010(d)). Once the catalyst was added, the absorption maxima at 403 and 240\u00a0nm dropped, while the third peak shifted from 280 to 293\u00a0nm. This fall in \u03bb\nmax at 403 and 240mn indicates catalytic hydrogenation of nitroarenes. Using Ag-Cu/TP, the full reduction of the three nitroarenes combination took just 8\u00a0min.As a result of the above findings, it was concluded that the Ag-Cu/TP catalyst is not only efficient in degradation of a particular signal pollutant, but also in the remediation of a combination of different pollutant molecules.When designing a wastewater treatment system, the cost of the process must be considered. A large proportion of the under-created and underdeveloped nation's enterprises could not pay the costs of usual sewage treatment plants since they are excessive to develop and supervise. Moreover, attention must be given to its operating costs. Many companies in under-developed and third world countries couldn't even afford to build and operate modern sewage treatment plants. Therefore, it is important to develop a low cost treatment method.Here in the current study, a novel column reactor for dyes remediation was developed as shown in Fig.\u00a011\n(a). A glass chromatographic column (Fig.\u00a011(b)) was filled with Ag-Cu/TP catalyst through the wet method by mixing it with the appropriate amount. The freshly prepared solutions of dye mixed with sodium borohydride were constantly passed through the developed column. Eluted solutions from the column were collected to carry out their UV\u2013Vis spectra for each reaction. The filled column without catalyst was used as a negative control to observe the adsorption capability of pure coarse sand.Reaction optimization for the development of columns was carried out for the reduction of MO dye. In order to study the catalyst amount influence on reduction using column through batch study, three different types of reactions were carried out. In the first experiment, 0.5\u00a0g of Ag-Cu/TP catalyst was mixed with 10\u00a0g of sand and then used for the column development. In the second experiment, 1\u00a0g of Ag-Cu/TP catalyst was mixed with 10\u00a0g of sand. In the third experiment, 2\u00a0g of Ag-Cu/TP catalyst was mixed with 10\u00a0g of sand and used for the column development. The three developed columns contain different amounts of catalyst. These columns were used for selective reduction of MO as a model pollutant. The flow rate of dye solutions passing the column was 2\u00a0mL/min and 5\u00a0mL/min.From the batch experiment, it was observed from the first experiment that 0.5\u00a0g of Ag-Cu/TP catalyst was able to reduce 37.7% of 0.05\u00a0mM of MO dye with a flow rate of 2\u00a0mL/min; however, its reduction efficiency declined to 22.6% with increasing the flow rate to 5\u00a0mL/min (as presented in Table 2\n). Results from the second experiment showed that 1\u00a0g of catalyst was able to reduce 67.4% of MO dye from aqueous solution with a flow rate of 2\u00a0mL/min, but its efficiency lowered to 49.2% by increasing the flow rate to 5\u00a0mL/min. Similarly, the column of 2\u00a0g of catalyst per 10\u00a0g of sand showed 72.8% with a flow rate of 2\u00a0mL/min, and 55.5% of MO reduction with the flow rate of 5\u00a0mL/min. From the results given in Table 2, it was concluded that 1\u00a0g of catalyst mixed with 10\u00a0g of sand showed the optimal catalyst amount as compared to 0.5 and 2\u00a0g of catalyst in 10\u00a0g of sand.The results of the catalytic reduction of azo dyes by column method are presented in Table 3\n. It was observed that the Ag-Cu/TP catalyst showed a good activity and reduced 68.7% and 68.0% of the CR and RhB dyes, respectively, with a flow rate of 2\u00a0mL/min. Similarly, it reduced 64.2% and 67.4% of MR and MO dyes, respectively, with a flow rate of 2\u00a0mL/min. Also, Ag-Cu/TP catalyst reduced 46.9% and 52.3% of the CR and RhB dyes, respectively, and 49.2% of MO and 46.4% of MR with a flow rate of 5\u00a0mL/min. The column was also used for the reduction of mixtures of dye solution, and it showed an excellent activity by reducing 69.0% of MO and CR solution. Besides catalyst, pure taro powder and sand were used to check their adsorption capabilities.The reusability and recycling of the effective catalyst are a major concern in today's catalysis field. Most catalysts are only used for a first or second cycle before being deactivated. Aside from catalytic efficiency, photocorrosion, durability, recyclability, and stability are also key factors to be considered when evaluating catalysts' performance because they can drastically reduce the method's cost. Catalysts in the type of colloidal nanoparticles have been successfully described in the literature as they exhibit photocatalytic activity; however, the issue is their recovering from reaction medium and their reusability for subsequent use. As a result, stable and completely separable catalysts are critical for the recyclability and reusability processes.The recyclability of synthesized Ag-Cu/TP catalyst was investigated using MO and 4-NP in this study. Unlike other pollutants, MO and 4-NP are easily reduced by sodium borohydride and Ag-Cu/TP. As a result, MO and 4-NP were chosen for a more detailed analysis of recyclability. For a 100 percent reduction of both compounds, the recyclability of the Ag-Cu/TP was evaluated for consecutive five cycles. The recyclability test of Ag-Cu/TP was tested after rinsing it three to four times in double DW when it was used for the reduction and utilized again in the reduction reaction at the same time. The disappearance and formation of peaks were observed periodically. Fig.12\n shows the recyclability study and the degradation of MO by using Ag-Cu/TP for five cycles. It was shown that Ag-Cu/TP took 9\u00a0min for the first and second use for MO reduction, while it took 10 and 11 min for 3rd and 4th uses, respectively. However, it took 14\u00a0min for 5th use of Ag-Cu/TP for MO degradation/reduction.However, in the 4-NP reaction, it took 5 and 6 min for the first and second uses, repectively For the 3rd and 4th uses, it took 8 and 9\u00a0min, respectively. While for the 5th use of 4-NP reduction, it took 12\u00a0min using Ag-Cu/TP (Fig.\u00a012(b)). The stability of Ag-Cu/TP, on the other hand, was assessed after 1, 18, and 36\u00a0h after preparation of catalyst. As demonstrated in Fig.\u00a012(b), the first usage took 5\u00a0min and 30\u00a0s, whereas the complete reduction required 7\u00a0min after 18\u00a0h. After 36\u00a0h of synthesis, however, the complete reduction of 4-NP took only 8\u00a0min (Fig.\u00a013\n).Because of the loss of nano-material during the recycling and washing process, the catalyst reduction performance in the continuous cycle reduced slightly. In fact, during the reduction phase, the reaction product might potentially block the active sites of the catalyst. However, after five cycles of turnover rates, Ag-Cu/TP was found to maintain strong catalytic activity. The reusability study indicated that the catalyst possesses excellent activity and stability and can be utilized many times in different reactions. The exceptional efficacy of the catalyst during the reusability study could be attributed to uniform dispersion of silver and copper nanoparticles on the support.Bimetallic Ag and Cu-based support catalyst was successfully synthesized by taro-rhizome through a low-cost method. SEM, EDX, and TEM pictures showed that Ag and Cu NPs were successfully synthesized on the plant powder surface. Powder XRD confirmed the crystalline nature and the fcc structure of synthesized catalyst.The synthesized Ag-Cu/TP nanomaterials have an outstanding catalytic activity for nitroarenes and dyes. The synthesized catalyst remained stable for several days, with exciting efficiency for 48\u00a0h, and could be simply recycled and utilized five times for 100 percent reduction of 4-NP and MO dye with very small changes in its catalytic activity. The catalyst that had not been treated with NaBH4 remained stable for a year without any change in catalytic properties. It was found that the catalyst has the best catalytic reduction activity not just for a particular substrate, but also for the reduction of a mixture of nitrophenols. As a result, this catalyst promises to evaluate the overall safety of diverse water-based systems by reducing nitrophenols\u00a0and organic dyes using the continuous column reactor. Because of their low cost and environmentally benign plant support, these catalysts might be produced in large quantities due to their superior metal ion absorption, high surface distribution, extraordinary recyclability, and stability.Not applicable.Not applicable.The authors confirm that the data supporting the findings of this study are available within the article.The authors extend their appreciation to the Deputyship for Research & Innovation, Ministry of Education in Saudi Arabia for funding this research work through the project number (IFPRC-061-247-2020) and King Abdulaziz University, DSR, Jeddah, Saudi Arabia.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors extend their appreciation to the Deputyship for Research & Innovation, Ministry of Education in Saudi Arabia for funding this research work through the project number (IFPRC-061-247-2020) and King Abdulaziz University, DSR, Jeddah, Saudi Arabia.", "descript": "\n                  Every day, the risk of water pollution grows. According to the World Health Organization (WHO), approximately half of the population of poor countries suffers from health problems linked to chemically or microbially contaminating drinkable water. Here in the present study, powder of taro-rhizome was used for preparation of supported bimetallic catalyst (Ag-Cu/TP). Characterization techniques used for analysis were XRD, TEM, SEM, FTIR and EDS. The prepared catalyst was then used in the catalytic reduction of chemical pollutants, including 4-nitrophenol (4-NP), 2,4,6-trinitrophenol (TNP), 2-nitrophenol (2-NP), methyl orange (MO), Acid Red 27 or amaranth (AM), Congo red (CR), Rhodamine B (RhB) and methyl red (MR). Turnover frequency was calculated for nitrophenol reductions, which were 0.98min\u22121 for 2-NP and 3.45min\u22121 for 4-NP. The Ag-Cu/TP catalyst\u00a0shows outstanding performance and 99.4% of 4-NP was reduced just in 5\u00a0min with pseudo-first \n                     order constant of 5.24\u00a0\u00d7\u00a010\u22123s\u22121, while 97% reduction of 2-NP took 17\u00a0min with the pseudo-first \n                     order constant of 1.01\u00a0\u00d7\u00a010\u22123s\u22121. The prepared Ag-Cu/TP catalyst was packed into a glass column and organic dyes containing water were filtered through it to test the cleaning effectiveness of water from dyes. Excellent dye reduction was achieved. The prepared column reactor works excellent in the remediation of dyes and reduces about 68.7% CR. In addition, stability and reusability study showed exceptional reusability up to 5 times with very small loss of activity and was stable for up to 10 days.\n               "}
{"full_text": "Polyhydroquinolines are an imperative cluster of 1, 4-dihydropyridine (1,4-DHP) nucleus containing heterocycles, that have fascinated much consideration because of their potent pharmacological and biological activities [1\u20136]. Polyhydroquinolines are also distinguished as calcium channel antagonists and platelet aggregation inhibitors. Indeed, a few 1,4-dihydropyridine derived commercial drugs constitute a major class of ligands for L-type Ca2+ channel (LTCC) blockers and are extensively used in the treatment for the therapy of the cardiovascular and Alzheimer's diseases [7\u20139]. Owing to the above noted most valuable therapeutic applications of the Polyhydroquinolines, synthetic researcher had developed number of methods for the synthesis of this valuable framework using several metal and halogen containing catalytic systems [10\u201330].In addition, heterocycles containing pyrimidine moieties occupy an outstanding position in medicinal chemistry due to its abundant therapeutic applications and that pyrimidine containing commercial drugs encloses the anticancer and antiviral activities [31,32]. Recently, many dihydropyrimidinone-functionalized scaffolds have been discovered antihypertensive and calcium channel modulating activities [33\u201336]. Certain developed protocols for the synthesis of dihydropyrimidinone compounds have been reported in the last decades by employing highly reactive enaminones as a precursor using different acid catalysts [37\u201339]. Most of the above reported methodologies for synthesis of polyhydroquinoline and dihydropyrimidinone compounds have been suffered with use of halogen containing acid catalysts, organic solvents and drastic conditions which are not environment friendly. Therefore, there is extreme necessity to develop a sustainable method to construct these compounds.Nowadays, ionic liquids have emerged as an excellent material due to their unique properties and have been extensively used in the industries as a catalysts as well as solvents [40]. The properties of ionic liquids such as low vapor pressure, wide liquid range, good conductivity, thermal, mechanical, and chemical stability, and modification of chemical structures of their ions offer a new improved technology [41,42]. Ionic liquids have already been found valuable in several fields of chemistry and it can be a replacement for volatile organic compounds in organic processes [43,44]. High efficiency is the most significant feature of green catalysis, and >90% processes worldwide depend on the fast-growing field of catalysis [45,46]. The ionic liquids catalysts are extremely important as they are environment friendly and leads to minimize the waste and cater sustainability [47,48].Currently, acidic ionic liquids (AILs) used as ecologically important reaction medium for the organic synthesis and has established extensive attention in organic transformations as an efficient catalyst [49\u201355]. Number of reviews discloses the importance of acidic ionic liquids encompassing a green catalyst as well as their application as a solvent in biodiesel production and industrial applications [56\u201359].It is well recognized point that the sustainability of chemical processes prominently depends on selection of renewable raw materials, use of benign solvents, low generation of hazardous waste, and reduction of energy consumption during manufacturing [60\u201362].Keeping in mind sustainable aspects and to introduce a new green protocol in terms of efficient reusable catalyst, herein in this endeavor we had demonstrated the synthesis of novel acidic ionic liquid [CEMIM][MSA]. The novel acidic ionic liquid ([CEMIM][MSA]) is synthesized for the first time and confirmed by various spectroscopic techniques such as IR, 1H NMR, 13C NMR, DEPT-135, Mass, HSQC, TGA and Elemental analysis. The acidic ionic liquid [CEMIM][MSA] was first time implemented as a catalyst for the green synthesis of polyhydroquinoline and 6-unsubstituted dihydropyrimidinone derivatives in eco-friendly media. Furthermore, we had established the sustainable, cost-effective, work up less and purification free process to isolate pure polyhydroquinolines and 6-unsubstituted dihydropyrimidinones with high yield in the very short span of time. The acidic ionic liquid catalyst [CEMIM][MSA] was recycled and reused five time without much loss in its catalytic activity. The plausible mechanism for the synthesis of polyhydroquinoline and 6-unsubstituted dihydropyrimidinone derivatives over [CEMIM][MSA] is discussed in detail. Molecular structure and hydrogen bonding study of polyhydroquinolines and 6-unsubstituted dihydropyrimidinone were carried out using a single crystal X-ray analysis.The main objective of our current research work is to design, the route of synthesis of acidic ionic liquid 3-(2-Carboxyethyl) -1-methyl-imidazole mesylate [CEMIM][MSA] using economically feasible raw materials. The synthesis is outlined in Scheme 1\n\n.\n1 equivalent of N-methyl imidazole and 1 equivalent of 3-chloro propionic acid compounds were treated at neat condition to obtain halogenated ionic liquid 3-(2- carboxyethyl) -1-methyl-imidazole chloride [CEMIM][Cl] with 96% yield. Further, the [CEMIM][Cl] treated in stoichiometric amount with methane sulfonic acid at ambient temperature in ethanol. The halide free acidic ionic liquid 3-(2-carboxyethyl)-1-methyl-imidazole mesylate [CEMIM][MSA] is obtained with 98% yield.The formation of acidic ionic liquid [CEMIM][MSA] was confirmed by spectral characterization techniques including Fourier transform infrared (FTIR), Nuclear magnetic resonance (1H NMR, 13C NMR, DEPT-135 and HSQC), Mass spectrometry, Elemental analysis and the spectral data is included in ESI (Figs. S1-S7).\nThe thermal stability of acidic ionic liquid [CEMIM] [MSA] was determined by thermogravimetric analysis (TG-DSC and TG-DTG) under N2 atmosphere at 25 to 1000\u00a0\u00b0C with the heating rate of 10\u00a0\u00b0C /min. The TG-DSC analysis plot is depicted in Fig. S8, and it showed two-stage decomposition. There is very slight weight loss occurred at 130\u00a0\u00b0C which correspond to the evaporation of water and volatile impurities. However, 31% weight loss observed at 284\u00a0\u00b0C reflects that the fragmentation of acidic ionic liquid [CEMIM] [MSA] has started followed by the main weight loss at 405\u00a0\u00b0C. Thus, complete decomposition of acidic ionic liquid take place at 405\u00a0\u00b0C. Moreover, DSC analysis results revealed an exothermal shift at temperatures 271\u00a0\u00b0C and 400\u00a0\u00b0C.These results pointed out that the acidic ionic liquid [CEMIM] [MSA] is a binary system catalyst and stable at high temperature reaction as well (TG-DSC and TG-DTG Figs. S8-S10 respectively included in the ESI).After confirmation of the structure of an ideal ionic liquid, the catalytic efficiency and applicability of acidic ionic liquid [CEMIM][MSA] as a catalyst was evaluated for the multicomponent synthesis of polyhydroquinolines and 6-unsubstitured dihydropyrimidinones.To optimize reaction condition for the synthesis of polyhydroquinolines (\nScheme 2\n\n), we have chosen the dimedone (1\u00a0mmol), 4-Cl benzaldehyde (1\u00a0mmol), ethyl acetoacetate (1\u00a0mmol), and ammonium acetate (1.5\u00a0mmol) as the reactants with 5\u00a0ml solvent volume for the model reaction.The initial experiments were carried in ecological solvents such as water and ethanol with 10\u00a0mol% [CEMIM][MSA] catalyst at ambient temperature. The reaction results, low yield of product and the time taken to complete the reaction is also quite long (Table 1\n, Entries 1,2). To evaluate activity of catalyst at higher temperature, same reactions were screened in Water, EtOH, it results in improvement in the yield of product (up to 85%) with reducing the reaction time (Table 1, Entries 3,4). Next step was testing of different ionic liquid catalysts were by changing catalyst loading concentration in EtOH medium at reflux temperature. It results in the poor to satisfactory yield of product (Table 1, Entries 5\u20138). However, model reaction performed at 100\u00a0\u00b0C in different ionic liquids as a reaction medium had achieved maximum yield (80%) of isolated polyhydroquinoline for ionic liquid [CEMIM][MSA] (Table 1, Entries 9\u201312). Also, the reaction progress was inspected under solvent free condition at higher temperature with 40\u00a0mol% [CEMIM][MSA] catalyst and it results in the satisfactory yield of product (Table 1, Entry 13). Inspired by the results of Water and EtOH at higher temperature with 20\u00a0mol% [CEMIM][MSA] catalyst, we have worked an experiment with model reactants in H2O: EtOH (1:1) solvent at 75\u00a0\u00b0C by charging 20\u00a0mol% [CEMIM][MSA] catalyst and astonishingly, the reaction completed in very less time and the yield of desired product obtained was 93% (Table 1, Entry 14). Then, the counter ion effect of synthesized ionic liquids [CEMIM][Cl] & [CEMIM][MSA] on the model reaction was evaluated on basis of experimental results, tabled in Table 1. The reaction performed by 20\u00a0mol% [CEMIM][Cl] in ethanol was observed 70% yield of product (Table 1, Entry 5), whereas the equal amount of [CEMIM][MSA] used in ethanol and H2O: EtOH solvents results, 80% and 93% yields of the product respectively as well as the reaction completed in short time (Table 1, Entries 4 &14). From the above results we can conclude that, the catalytic activity of ionic liquid gets enhanced when the chloride anion exchanged by mesylate anion. The ionic liquid [CEMIM][MSA] is more acidic in nature and thereby it perceived high catalytic performance compared to [CEMIM][Cl].Further the effect of solvents were investigated on the model reaction by employing 20\u00a0mol% [CEMIM][MSA] catalyst at various temperatures and results are depicted in Table 2\n. The reaction doesn't progress in non-polar solvent like toluene, o-xylene, and cyclohexane even when it was stirred for longer period (Table 2, Entries 1\u20133). Even by using the chlorinated and etheral solvents such as MDC, CHCl3, di isopropyl ether, methyl tert-butyl ether, the trace amount of product was obtained (Table 2, Entries 4\u20137). Then, the slightly polar aprotic solvents acetone and ethyl acetate, yielded up to 40% product while the low boiling polar aprotic solvents THF and acetonitrile were afforded 65% and 75% of polyhydroquinoline (Table 2, Entries 8\u201311). On using polar protic solvents such as water, MeOH, EtOH, IPA, H2O: EtOH (1:1), trifluoroethanol and ethylene glycol it had enhanced good to excellent yield of product (Table 2, Entries 12\u201318). However, satisfactory yield of product 10a was obtained by employing reaction at 100\u00a0\u00b0C for 1\u20131.5\u00a0h in high boiling polar aprotic solvents like DMSO, DMF, NMP and N, N-dimethyl acetamide (Table 2, Entries 19\u201322). In addition, trace amount of product was observed at 80\u00a0\u00b0C in neat condition after 1.5\u00a0h (Table 2, Entry 23). It is important to note that, the catalytic efficiency of catalyst enhanced in H2O: EtOH (1:1) polar protic medium at 75\u00a0\u00b0C and reaction had marched successfully to yield, 93% of product. The general investigation of solvent effect on product yield is tabled in (Table 2). It was observed that the ecofriendly solvent system H2O: EtOH (1:1) had high yield of product formation attributed owing to solubility of all reactant and intermediate product at optimized temperature as well as the low solubility of product at isolation temperature in reaction media.The loading of catalyst amount is a crucial factor in terms of reaction efficiency. To optimize catalyst loading, a set of experiment for model reaction at 75\u00a0\u00b0C in 5\u00a0ml of H2O: EtOH (1:1) with [CEMIM][MSA] catalyst (10 to 25\u00a0mol%) are displayed in Fig. 1\n. The results reflects that the yield of polyhydroquinoline 10a was maximum by loading 20\u00a0mol% of [CEMIM][MSA] catalyst. Further excess in the catalyst mol% did not affect the yield and time of reaction.Moreover, the effect of temperature for the synthesis of targeted molecule was scrutinized on the model reaction with 20\u00a0mol% of [CEMIM][MSA] catalyst and 5\u00a0ml of H2O: EtOH (1:1) solvent as shown in Fig. 2\n\n. The results signifies that the reaction performed at 25\u00a0\u00b0C and 50\u00a0\u00b0C was perceived satisfactory yield of product. Further, the yield of polyhydroquinoline 10a was raised up to 93% in 15\u00a0min when reaction implemented at 75\u00a0\u00b0C. On increasing the temperature up to 80\u00a0\u00b0C there was neither improvement in the yield of product 10a nor reduction in the time for the reaction to go to completion.Having the suitable condition in hand, the adaptability of the methodology was examined by changing different substituents in the reaction. The reaction of substituted aryl aldehydes was performed with dimedone, ethyl acetoacetate and ammonium acetate in H2O: EtOH (1;1) medium at 75\u00a0\u00b0C. The reaction results in 90\u201393% yield of polyhydroquinolines (10a-10v) in 15\u201330\u00a0min (Table 3\n, entries 1\u201322).Normally it is observed that the reactions of aromatic aldehydes containing electron withdrawing group such as Cl, Br, F, NO2 at different positions demonstrated high yield of the products than the reactions of aldehydes containing electron donating group such as \u2013CH3, -OH, -N(CH3)2, -OCH3, -OC2H5. The synthesized polyhydroquinolines (10a-10v) were crystalline compounds. These compounds were characterized based on their melting points, spectroscopic analysis (IR, 1H NMR, 13C NMR, DEPT-135 and LC-Mass) and were matched with those of literature authentic sample [25,63\u201368].Furthermore, to validate the formation of synthesized polyhydroquinolines the single crystal X- ray diffraction analysis of compounds 10j, 10v were carried out and their molecular structures were confirmed (Figs. 3 & 4\n\n\n). The compound 10j crystallized in monoclinic crystal system with the space group P 21/n, whereas the compound 10v crystallized in orthorhombic system with space group P bca. Based on the Single crystal XRD analysis results, the conformations of fused cyclohexenone and quinoline ring (10j and 10v), were identified as chair and boat conformations. The crystal data and structure refinement details for compounds 10j and 10v is tabled in Table S1.\nThe intra-molecular hydrogen bonding details for compounds 10j and 10v are tabled in Tables S2 & S3. It discloses the H bonding interaction between O-atom of carbonyl and H-atom of amino for 10j with a distance 2.11\u00a0\u00c5 and an angle of 172.8\u00b0. Whereas H bonding interaction between O-atom of carbonyl and H-atom of amino for 10v with a distance 2.06\u00a0\u00c5 and an angle of 174\u00b0. Then, the intermolecular packing interaction of molecules (10j and 10v) are displayed in Figs. S11 & S12 (Tables S2-S3 & Figs. S10-S11 are included in the ESI).Purity of compound is the most important entity in chemical reaction and to highlight this prominent point, we have checked the HPLC purity of typical compounds (10a and 10v). The purity of the synthesized polyhydroquinolines (without purification) was found to be >99.0%, which makes the process highly cost-effective.A plausible mechanism studies of acidic ionic liquid [CEMIM][MSA] employed as a catalyst for multicomponent transformation of polyhydroquinoline (10) under optimized condition is accredited in \nScheme 3\n\n\n. The acidic ionic liquid [CEMIM][MSA] contains the imidazolium cation moiety and carboxylic acid group as active catalytic centers. Initially, the aryl aldehyde (7) will be activated by imidazolium cation moiety and carboxylic acid group. Afterward, the mesylate anion abstract proton from active methylene compounds (6, 8) to generate carbanion which will rapidly react with activated aryl aldehyde to afford the Knovenagel intermediate (C, D) via elimination of water molecule and regeneration of ionic liquid. On the other side, the imidazolium cation moiety and carboxylic acid group activated the active methylene compounds (6, 8) to convert their enol form, which will further react with ammonium acetate (9) to give an enaminone compounds (A, B). Further, the mesylate anion abstract the proton from enaminone (A, B), the generated anion will react with intermediate adduct (C, D) to afford polyhydroquinolines (10) with elimination of water molecule and regeneration of catalyst. The hydrogen bonding effect of acidic ionic liquid [CEMIM][MSA] with aryl aldehydes (7) and active methylene compounds (6, 8) in polar media enhance their electrophilic character. Therefore, the acidic ionic liquid [CEMIM][MSA] could increase the rate of construction of polyhydroquinolines (10). The projected mechanism explain the formation Knovenagel compounds by the Knovenagel reaction of aryl aldehydes with active methylene compounds. It is followed by Michael addition reaction, which occurs between Knovenagel compounds and enaminone, followed by intramolecular cyclization and dehydration to polyhydroquinolines.\nThe versatility along with catalytic effectiveness of this catalyst for the three component one pot synthesis of dihydropyrimidinones (\nScheme 4\n\n) was explored. The reaction was carried out using 4-chlorobenzaldehyde, enaminones, and urea with 1:1:1.2\u00a0M ratio as the model substrates using 5\u00a0ml of solvent.Initially, the catalytic potency of acidic ionic liquid [CEMIM][MSA] as a catalyst was evaluated in toluene, methanol, and water solvents with minimum charging of catalyst quantity for the synthesis of dihydropyrimidinone 14a. However, the results were not satisfactory in terms of reaction time and the yield of product (Table 4\n, Entries 1\u20133). Consequently, the model reaction was obtained satisfactory to excellent yield of dihydropyrimidinone 14a, when it was performed with ionic liquid [CEMIM][MSA] and [CEMIM][Cl] as a catalyst in ethanol, isopropanol, and H2O: EtOH (1:1) solvents along with neat condition by fluctuating the temperature and catalyst concentration parameter (Table 4, Entries 4\u201312). However, the ionic liquids [CEMIM][MSA] and [CEMIM][Cl] were scrutinized as solvent and reaction placed at 120\u00a0\u00b0C for 180\u00a0min it results in the yield of product 55% and 40% respectively (Table 4, Entries 13\u201314). Next step is to study, the impact of other acid catalysts such as HCOOH, H3PO4, L-Proline, PTSA, Amberlyst-15 and NH2SO3H, using protic and aprotic solvents at boiling temperatures and for longer duration. It afforded poor yield of isolated dihydropyrimidinone 14a (Table 4, Entries 15\u201320). From the results of Table 4\n, we can conclude that 25\u00a0mol% [CEMIM][MSA] catalyst in H2O: EtOH (1:1) solvent at the temperature of 75\u00a0\u00b0C is found to be the superior optimum condition in relative to all other completed experiments with respect to yield and time.For exploring the robustness of the reaction using this typical condition, we further synthesized a series of 6-unsubstituted dihydropyrimidinones by treating various aryl aldehydes, different enaminones and urea (\nTable 5\n\n). The survey of the reveled protocol specified that the reaction of urea with aromatic aldehydes and various enaminones gave excellent yields of dihydropyrimidinones. The synthesized 6-unsubstituted dihydropyrimidinones were confirmed by spectroscopic techniques and physical results were compared with those reported in the literature [37\u201339].One of the illustrative derivative such as 4-(4-Methoxyphenyl)-5-phenylmethanone-yl - 3, 4 -dihydropyrimidin-2(1H)-one (14c) was thoroughly explored by spectroscopic data. The FTIR spectrum of compound 14c showed absorption peaks for NH stretching at 3336 and 3204\u00a0cm\u22121 while the >C=O stretching vibrations absorption peaks assigned at 1691 and 1654\u00a0cm\u22121. 1H NMR analysis revealed that the OCH3 proton observed singlet at \u03b4 3.73\u00a0ppm, whereas the characteristic methine proton signal appeared doublet at \u03b4 5.39\u00a0ppm. The aromatic proton signals were observed in the range of at \u03b4 6.89\u20137.28\u00a0ppm while two NH protons showed downfield signals at \u03b4 7.78 and \u03b4 9.25\u00a0ppm respectively. Next, the 13C NMR spectra was assigned the signals at \u03b4 39.72\u00a0ppm and \u03b4 55.54\u00a0ppm for OCH3 and methine carbon. The carbonyl C signals were observed at \u03b4 159.07 and \u03b4 192.05\u00a0ppm while aromatic C signals appeared in the range at \u03b4 113.31\u2013151.72\u00a0ppm. The GC\u2013MS analysis disclosed the molecular ion mass (m/z) at 308.20, which is like calculated mass. All the above interpreted spectroscopic results defined the formation of desired dihydropyrimidinone derivative 14c (All the spectra are included in the ESI as entry 14c).Moreover, in supporting to the spectroscopic results, the crystal structure of compound 14c was determined in detail by single crystal X-ray diffraction analysis. The compound 14c crystallized in monoclinic crystal system with P 21/c space group and the unit cell dimensions are; a\u00a0=\u00a09.4018 (6) \u00c5, \u03b1\u00a0=\u00a090\u00b0, b\u00a0=\u00a08.1698 (5) \u00c5, \u03b2\u00a0=\u00a0101.575 (3)\u00b0 and c\u00a0=\u00a020.6085 (13) \u00c5, \u03b3\u00a0=\u00a090\u00b0. ORTEP of 14c with confirmation of pyrimidine ring is as shown in Fig. 5\n and the crystal structure refinement details is expressed in Table S4.\nThe geometry of intramolecular hydrogen bonding for compound 14c is represented in Table S5. The proton attached nitrogen N1 forms an H bonds with the Oxygen O2 (>C=O) such that the N1\u2026.H1A\u2026.O2 distance is 2.789 \u01fa while that of the N2\u2026.H2A \u2026.O1 (>C=O) distance is 2.958 \u01fa. Also, the packing interactions of molecules (14c) along different axes is outlined as in Fig. S13 It is significant to remark that this is a sustainable process to prepare ten novel compounds (Table 5, Entries 4\u201313) not listed in the literature derived from dihydropyrimidinones while the four compounds (Table 5, entries 1\u20133 & 14) are known derivatives. (Tables S4-S5 & Fig. S13 are included in the ESI).A proposed mechanism for the developed transformation is projected in \nScheme 5\n\n\n. Initially the catalytic active centers of acidic ionic liquid [CEMIM][MSA] will activate the aryl aldehyde 11 and enaminone 12. Then urea 13 directly reacted with activated aryl aldehyde 11 to form the N-acyliminium intermediate E. Subsequently, the activated enol tautomer 12A of enaminone 12 will reacted with N-acyliminium intermediate E to produce an intermediate G via adduct F. The intermediate G underwent intramolecular cyclization to form hexahydropyrimidine compound H, which on further removal of dimethyl amine as by product to afford the dihydropyrimidinone 14.Atom economy is an essential aspect of green chemistry, and the efficiency of reactions were established by using the percent atom economy. The percent atom economy for the reaction of polyhydroquinolines (10a-10v) and dihydropyrimidinones (14a-14n) is calculated by using the formula, % Atom economy\u00a0=\u00a0(MW of product / \u2211 MW of reactants) x 100 and the results implied that, both the reaction have high percent atom economy (Fig. 6\n\n) which generated minimum amount of waste product. Then, the percent carbon economy exposes the percentage carbon in the reactants that remain in the ideal product and was evaluated for the polyhydroquinoline 10a and dihydropyrimidinone 14a by applying the formula,% Carbon economy\u00a0=\u00a0(Number of carbon in product / \u2211 number of carbon in reactants) x 100.The compounds 10a and 14a had shown 91.3% and 89.5% carbon economy which persisted maximum number of carbons.Reusability and recovery of the acidic ionic liquid catalyst [CEMIM][MSA] was examined for the synthesis of polyhydroquinoline (10a) and dihydropyrimidinone (14a) under the optimized conditions (\nFig. 7\n\n). After completion of the reaction, the catalyst was recovered from the filtrate and applied for repeated reactions for the same model reaction under the same reaction conditions. Based on recyclable results, it was observed that the [CEMIM][MSA] exhibited as highly efficient and recoverable for a five consecutive times with insignificant loss in its catalytic efficacy. The physicochemical transformation of recovered catalyst after fifth runs was inspected by employing NMR and Mass analysis (Figs. S16-S20). There is no remarkable transformation in the structure of recovered catalyst as referred by NMR (1H NMR, 13C NMR, and DEPT-135) and Mass in comparison with fresh catalyst. These results suggest that the catalyst was greatly stable up to fifth recycle and suitably proficient for its reuse under typical reaction conditions. (Figs. S16-S20 are included in the ESI).To explore the merits of the existing catalyst in association with other reported catalyst in literature for similar reaction. We have tabulated the performance of acidic ionic liquid [CEMIM][MSA] amongst other catalysts for the synthesis of polyhydroquinolines (10a-10v) and dihydropyrimidinones (14-14n) are revealed in Table 6\n\n. As it is marked from the results, acidic ionic liquid [CEMIM][MSA] can accomplished as highly proficient catalyst with respect to reaction times and yield of products. Furthermore, to highlight the noteworthy advantages of the developed method, we have matched our result with literature reported methods (\nTable 6\n). As described in Table 6\n, the developed strategies are superior as compared with defined approaches in terms of i) use of benign media and halogen free catalysts, ii) avoidance of drastic reaction condition, iii) workup less safe process, iv) easy recovery of catalysts, v) evasion of column and recrystallization methods to purify product, vi) short reaction time, and vii) excellent yields of product.All the chemical reactions were performed in round bottom single neck glass flask. The chemicals and solvents were purchased from Loba Chemicals, Sigma Aldrich and S. D. fine chemicals and used as received without further purification. All the melting points were measured by a Labstar melting point apparatus and were uncorrected. All the reactions were monitored by the thin layer chromatography (TLC) using silica gel coated plates. Infrared spectra were documented on a Perkin- Elmer, FTIR-1600 and Bruker, 3000 hyperion microscope with vertex 80 spectrophotometer in KBr and expressed in cm\u22121. Analysis of 1H NMR, 13C NMR, DEPT-135and HSQC spectra were determined on a Bruker Avance (300, and 400\u00a0MHz) spectrometer as D2O, DMSO\u2011d\n6 solutions, using tetramethyl silane (TMS) as the internal standard. Chemical shifts (\u03b4) were expressed in ppm. Mass spectra were recorded with a Waters Q-TOF micromass spectrometer by electrospray ionization (ESI). Thermogravimetric analysis (TGA) was performed on a Netzsch Sta 449 F3 Jupiter analyzer under a nitrogen atmosphere. Purity of compounds were analyzed by the HPLC on an Agilent 1200 system.X-ray diffraction data for compounds 10j, 10v and 14c were collected at T\u00a0=\u00a0297 (2) K on a Bruker APEXII CCD diffractometer with graphite monochromated Mo K\u03b1 (\u03bb\u00a0=\u00a00.71073 A\u00b0) radiation. Table S1 and Table S4 data are exhibited the unit cell parameters and other crystallographic details. Table S2, Table S3 and Table S5 data represented the hydrogen bonding details. The structures were solved using the direct methods of program SHELXS 2018 and refined anisotropically by full-matrix least-square on F2 with the program SHELXL 2018.Acidic ionic liquid [CEMIM][MSA] was synthesized based on the approach termed in preceding studies with some modifications [69\u201372]. In a 100\u00a0ml single neck flask, charged 1- methyl imidazole (1) (20\u00a0mmol) and 2- chloro propionic acid (2) (20\u00a0mmol) at 25\u201330\u00a0\u00b0C. The reaction mixture was agitated at 45\u00a0\u00b0C till completion of reaction (viscous mass). After that reaction mixture was washed and decanted with 3\u00a0\u00d7\u00a050 ml ethyl acetate at 25\u201330\u00a0\u00b0C to remove unreacted raw materials. Then the oily mass degassed under vacuum at 70\u201375\u00a0\u00b0C to give an intermediate ionic liquid [CEMIM][Cl] (3) as viscous white semisolid in 96.0% yield. Next, in same reaction flask [CEMIM][Cl] (3) (20\u00a0mmol) and 40\u00a0ml ethanol were charged, and reaction mixture stirred well for 15\u00a0min at 25\u201330\u00a0\u00b0C to observed homogeneous solution. To the intermediate ionic liquid solution, methane sulfonic acid (4) (21\u00a0mmol) was slowly added with vigorous stirring at 25\u201330\u00a0\u00b0C and mixture was stirred further 4\u20135\u00a0h on the same condition to carry anion metathesis reaction. Furthermore, the reaction mixture was distilled under vacuum at 50\u201355\u00a0\u00b0C to obtain viscous oil. The viscous oil washed and decanted with 3\u00a0\u00d7\u00a050 ml ethyl acetate to remove traces of acid. Finally, the oily residue degassed well under vacuum at 70\u201375\u00a0\u00b0C, afforded viscous a light green colored desired acidic ionic liquid [CEMIM][MSA] (5) with 98.0% yield.Charged dimedone 6 (1\u00a0mmol), aldehyde 7 (1\u00a0mmol), ethyl acetoacetate 8 (1\u00a0mmol), ammonium acetate 9 (1.5\u00a0mmol) and H2O: EtOH (1:1, 5\u00a0ml) in a flask. Charged acidic ionic liquid [CEMIM][MSA] (20\u00a0mol%) catalyst and the reaction mixture was agitated at 75\u00a0\u00b0C till completion of reaction. After completion of reaction, the mixture was cooled to 0\u20135\u00a0\u00b0C and stirred for 1\u00a0h. The precipitated solid was filtered and washed with 2\u00a0\u00d7\u00a02.5\u00a0ml cooled aqueous ethanol (1:1) to get a pure product in excellent yield.A three-component mixture of aldehyde 11 (1\u00a0mmol), enaminones 12 (1\u00a0mmol) and urea 13 (1.2\u00a0mmol) was treated with [CEMIM][MSA] (25\u00a0mol%) in a flask at 75\u00a0\u00b0C in H2O: EtOH (1:1, 5\u00a0ml) After completion of reaction, the crude mixture was cooled to 25\u201330\u00a0\u00b0C and stirred for 30\u00a0min. The precipitated solid was filtered out and washed with 2\u00a0\u00d7\u00a02.5\u00a0ml aqueous ethanol (1:1), which results in excellent yield of pure dihydropyrimidinones 14.The recovery of the catalyst [CEMIM][MSA] was evaluated for polyhydroquinoline synthesis on the model reaction dimedone (1\u00a0mmol), 4-Cl benzaldehyde (1\u00a0mmol), ethyl acetoacetate (1\u00a0mmol), ammonium acetate (1.5\u00a0mmol) and 20\u00a0mol% of [CEMIM][MSA] in aq. media at 75\u00a0\u00b0C. After filtration of product, the filtrate was concentrated on rotary evaporator to give crude residue containing ionic liquid and organic impurities. The residue was washed with ethyl acetate (3\u00a0\u00d7\u00a015\u00a0ml) to remove organic impurities. After that the catalyst was dried at 70\u00a0\u00b0C under reduced pressure and reused in consequent reaction for next run.White semisolid; Yield: 96.0%;\n1H NMR (500\u00a0MHz, D2O): \u03b4 2.93\u20132.95(t, 2H, CO-CH2), 3.82 (s, 3H, N-CH3), 4.40\u20134.43 (t, 2H, N-CH2), 7.36\u20137.38(m, 1H, Ar- H), 7.45\u20137.48(m, 1H, Ar- H), 8.71 (s, 1H, Ar- H);ESI-MS (+ve ion) (m/z): 155.09(M+);Anal. Calcd. For C7H11N2O2Cl (190.627): C, 44.10; H, 5.82; N, 14.70 Found: C, 44.07; H, 5.78; N, 14.76.Light green viscous oil; Yield: 98.0%;IR (KBr): 3483, 3110, 1725, 1574, 1413, 1190, 1051\u00a0cm\u22121;\n1H NMR (400\u00a0MHz, D2O): \u03b4 2.66 (s, 3H, S-CH3), 2.85\u20132.88 (t, 2H, CO-CH2), 3.74 (s, 3H, N-CH3), 4.32\u20134.35 (t, 2H, N-CH2), 7.28\u20137.30(d, 1H, Ar- H), 7.37\u20137.38(d, 1H, Ar- H), 8.63 (s, 1H, Ar- H);\n13C NMR (100\u00a0MHz, D2O): \u03b4 33.84, 35.54, 38.30, 44.70, 122.18, 123.43, 136.44, 174.16;DEPT-135 (100\u00a0MHz, D2O): \u03b4 33.83, 35.57, 38.29, 44.70, 122.18, 123.43, 136.43;ESI-MS (+ve ion) (m/z): 155.08(M+);ESI-MS (\u2212 ve ion) (m/z): 96.06(M+);Anal. Calcd. For C8H14N2O5S (250.272): C, 38.39; H, 5.64; N, 11.19. Found: C, 38.34; H, 5.61; N, 11.25.Light green viscous oil; Yield: 94.0%;\n1H NMR (400\u00a0MHz, D2O): \u03b4 2.66 (s, 3H, S-CH3), 2.83\u20132.87 (t, 2H, CO-CH2), 3.75 (s, 3H, N-CH3), 4.30\u20134.34 (t, 2H, N-CH2), 7.27\u20137.28(d, 1H, Ar- H), 7.34\u20137.35(d, 1H, Ar- H), 8.61 (s, 1H, Ar- H);\n13C NMR (100\u00a0MHz, D2O): \u03b4 33.75, 35.58, 38.28, 44.67, 122.18, 123.44, 136.42, 174.07;DEPT-135 (100\u00a0MHz, D2O): \u03b4 33.76, 35.50, 38.29, 44.61, 122.17, 123.42, 136.41;ESI-MS (+ve ion) (m/z): 155.09(M+);ESI-MS (\u2212 ve ion) (m/z): 96.06(M+);Light yellow solid; mp: 249\u2013250\u00a0\u00b0C; Yield: 94.0%;IR (KBr): 3274, 3205, 3077, 29,591,706, 1648, 1604, 1491, 1382, 1279, 1214, 1071\u00a0cm\u22121;\n1H NMR(400\u00a0MHz, DMSO\u2011d\n6): \u03b4 0.81 (s, 3H, CH3), 0.98 (s, 3H, CH3), 1.08\u20131.12 (t, 3H, CH3), 1.92\u20132.11 (m, 2H, CH2), 2.20 (s, 3H,CH3), 2.24\u20132.33 (s, 2H,CH2), 3.90\u20133.95 (q, 2H,CH2), 4.81 (s, 1H, CH), 7.06\u20137.13 (m, 4H, ArH), 8.83 (s, 1H, NH);\n13C NMR (100\u00a0MHz, DMSO\u2011d\n6): \u03b414.45, 18.66, 26.89, 29.52, 32.49, 36.04, 50.60, 59.52, 103.86, 110.16, 127.86, 129.66, 130.73, 145.65, 146.82, 150.24, 167.17, 195.12;DEPT-135 (100\u00a0MHz, DMSO\u2011d\n6): \u03b4 14.43, 18.64, 26.86, 29.50, 36.02, 39.85, 50.56, 59.55, 127.84, 127.87, 129.65;ESI-MS (m/z): 374.15 (M\u00a0+\u00a0H)+;HPLC purity: 99.09%;Yellow solid; mp: 256\u2013258\u00a0\u00b0C; Yield:91.0%;IR (KBr): 3227, 3206, 3077, 2958, 1701, 1648, 1495, 1380, 1280, 1216, 1168, 1072, 1030\u00a0cm\u22121;\n1H NMR(400\u00a0MHz, DMSO\u2011d\n6): \u03b4 0.84 (s, 3H, CH3), 0.99 (s, 3H, CH3), 1.12\u20131.14 (t, 3H, CH3), 1.92\u20132.06 (m, 2H, CH2), 2.20 (s, 3H,CH3), 2.10\u20132.33 (m, 2H,CH2), 3.63 (s, 3H, CH3), 3.90\u20133.96 (q, 2H,CH2), 4.76 (s, 1H, CH), 6.59\u20136.63 (m, 2H, ArH), 7.03\u20137.05 (m, 2H, ArH), 8.73 (s, 1H, NH);\n13C NMR (100\u00a0MHz, DMSO\u2011d\n6): \u03b4 14.49, 18.68, 27.03, 27.94, 29.64, 31.51, 32.49, 35.42, 50.79, 55.08, 59.33, 104.67, 110.81, 113.17, 128.86, 128.96, 140.44, 144.81, 149.76, 157.56, 167.41, 194.95;DEPT-135 (100\u00a0MHz, DMSO\u2011d\n6): \u03b4 14.48, 18.69, 27.03, 29.63, 35.42, 40.04, 50.79, 55.07, 59.32, 113.15, 128.86;ESI-MS (m/z): 392.18 (M\u00a0+\u00a0Na)+;Light yellow solid; mp:216\u2013218\u00a0\u00b0C; Yield:90.0%;IR (KBr): 3399, 3294, 3207, 3096, 2951, 1658, 1589, 1510, 1311, 1272, 1168, 1031\u00a0cm\u22121;\n1H NMR(400\u00a0MHz, DMSO\u2011d\n6): \u03b4 0.87 (s, 3H, CH3), 1.00 (s, 3H, CH3), 1.13\u20131.17 (t, 3H, CH3), 1.94\u20132.13 (m, 2H, CH2), 2.23 (s, 3H,CH3), 2.26\u20132.37 (m, 2H,CH2), 3.67 (s, 3H, OCH3), 3.96\u20133.98 (m, 2H,CH2), 4.73 (s, 1H, CH), 6.48\u20136.55 (m, 2H, ArH), 6.69 (s, 1H, ArH), 8.47 (s, 1H, OH), 8.84 (s, 1H, NH);\n13C NMR (100\u00a0MHz, DMSO\u2011d\n6): \u03b4 14.60, 18.62, 26.81, 29.68, 32.50, 35.54, 50.70, 55.78, 59.46, 104.72, 110.65, 112.26, 115.21, 119.97, 139.51, 144.72, 144.77, 147.07, 15.03, 167.56, 195.22;DEPT-135 (100\u00a0MHz, DMSO\u2011d\n6): \u03b4 14.60, 18.63, 26.81, 29.68, 53.54, 39.91, 50.70, 55.77, 59.46, 112.25, 115.21, 119.97;ESI-MS (m/z): 408.18 (M\u00a0+\u00a0Na)+;Light yellow solid; mp: 230\u2013232\u00a0\u00b0C; Yield: 92.0%;IR (KBr): 3280, 3212, 3077, 2966, 2937, 1708, 1645, 1605, 1527, 1489, 1381, 1278, 1211, 1151, 1108, 1072, 1030\u00a0cm\u22121;\n1H NMR(400\u00a0MHz, DMSO\u2011d\n6): \u03b4 0.83 (s, 3H, CH3), 0.99 (s, 3H, CH3), 1.09\u20131.13 (t, 3H, CH3), 1.93\u20132.13 (m, 2H, CH2), 2.29 (s, 3H,CH3), 2.40 (s, 3H, CH3), 2.25\u20132.49 (m, 2H,CH2), 3.91\u20133.96 (m, 2H,CH2), 4.87 (s, 1H, CH), 7.18\u20137.20 (d, 1H, ArH), 7.34\u20137.36 (m, 1H, ArH), 7.70\u20137.71 (d, 1H, ArH), 9.00 (s, 1H, NH);\n13C NMR (100\u00a0MHz, DMSO\u2011d\n6): \u03b4 14.38, 18.70, 19.83, 26.84, 29.52, 32.54, 36.25, 50.49, 59.66, 103.36, 109.75, 123.61, 130.70, 132.64, 132.95, 146.19, 147.52, 148.60, 150.64, 166.94, 195.07;DEPT-135 (100\u00a0MHz, DMSO\u2011d\n6): \u03b4 14.38, 18.70, 19.83, 26.84, 29.52, 36.25, 39.81, 50.49, 59.66, 123.62, 123.65, 132.95;ESI-MS (m/z): 399.20 (M\u00a0+\u00a0H)+;HPLC purity: 99.03%;Yellow solid; mp: 230\u2013232\u00a0\u00b0C; Yield: 91.0%;IR (KBr): 3336, 3204, 3083, 2897, 1691, 1654, 1609, 1510, 1435, 1370, 1325, 1250, 1215, 1157, 1028\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 3.73(s, 3H, OCH3), 5.39\u20135.40 (d, 1H, CH), 6.89\u20136.92 (d, 2H,Ar-H), 6.99\u20137.01 (d, 1H, CH), 7.25\u20137.28 (d,1H, ArH), 7.45\u20137.53 (m, 6H, ArH), 7.78 (s,1H,NH), 9.25\u20139.27 (d,1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 39.72, 55.54, 113.31, 114.30, 128.08, 128.44, 128.80, 136.76, 139.11, 141.86, 151.72, 159.07, 192.05;GC\u2013MS (m/z): 308.20 (M)+;Yellow solid; mp: 284\u2013285\u00a0\u00b0C; Yield: 93.0%;IR (KBr): 3313, 3211, 3084, 2911, 1697, 1653, 1611, 1499, 1376, 1325, 1253, 1216, 1162\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 5.44 (s, 1H, CH), 7.06\u20137.08 (d, 1H, CH), 7.31\u20137.51(m, 9H, ArH), 7.92 (s, 1H, NH), 9.44 (s, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 53.49, 112.24, 125.53, 126.88, 127.91, 128.49, 128.86, 131.04, 131.51, 133.51, 138.78, 142.68, 146.81, 151.55, 192.05;ESI-MS (m/z): 313.1 (M\u00a0+\u00a0H)+;Yellow solid; mp: 292\u2013295\u00a0\u00b0C; Yield: 92.0%;IR (KBr): 3289, 3074, 2896, 1698, 1657, 1614, 1447, 1378, 1328, 1224, 1163, 1134\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 5.49 (s, 1H, CH), 7.10\u20137.22 (m, 4H, CH & ArH), 7.49 (s, 6H, ArH), 7.95 (s, 1H, NH), 9.43 (s, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 53.42, 112.35, 113.54, 113.82, 114.61, 114.88, 122.86, 122.90, 128.50, 128.87, 131.04, 131.14, 131.51, 138.83, 142.64, 147.18, 147.26, 151.64, 161.05, 164.28, 192.11;ESI-MS (m/z): 297.1 (M\u00a0+\u00a0H)+;Yellow solid; mp: 282\u2013284\u00a0\u00b0C; Yield: 94.0%;IR (KBr): 3268, 2964, 1682, 1592, 1571, 1508, 1445, 1371, 1326, 1245, 1200, 1151, 1088\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 5.45\u20135.46 (d, 1H, CH), 7.03\u20137.14 (m, 3H, CH & ArH), 7.35\u20137.50 (m, 6H, ArH), 7.86\u20137.87 (d, 1H, ArH), 8.21 (s, 1H, NH), 9.38 (s, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 53.23, 112.97, 115.70, 128.40, 128.72, 138.90, 140.65, 140.69, 142.15, 151.64, 160.27, 163.50, 192.04;ESI-MS (m/z): 297.1 (M\u00a0+\u00a0H)+;Yellow solid; mp: 258\u2013260\u00a0\u00b0C; Yield: 90.0%;IR (KBr): 3327, 3206, 3087, 2923, 1687, 1652, 1620, 1439, 1369, 1320, 1250, 1210, 1180, 1157, 1078\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 2.17 (s, 3H, CH3), 2.20 (s, 3H, CH3), 5.36 (s, 1H, CH), 7.01\u20137.09 (m, 4H, CH & ArH), 7.47\u20137.50 (d, 5H, ArH), 7.78 (s, 1H, NH), 9.28\u20139.29 (d, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 19.45, 19.98, 53.52, 113.19, 124.22, 128.03, 128.45, 128.85, 130.06, 131.42, 135.82, 136.55, 139.00, 141.95, 157.75, 192.12;ESI-MS (m/z): 307.1 (M\u00a0+\u00a0H)+;White solid; mp: 278\u2013279\u00a0\u00b0C; Yield: 92.0%;IR (KBr): 3273, 3142, 2924, 1703, 1651, 1612, 1575, 1446, 1369, 1329, 1263, 1247, 1199, 1125, 1074\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 5.54 (s, 1H, CH), 7.00\u20137.11 (d, 1H, CH), 7.47\u20137.65 (m, 9H, ArH), 7.99 (s, 1H, NH), 9.50 (s, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 53.63, 112.06, 123.63, 124.78, 128.48, 128.88, 129.32, 129.75, 130.31, 131.02, 131.52, 138.79, 142.88, 145.74, 151.43, 192.04;ESI-MS (m/z): 347.1 (M\u00a0+\u00a0H)+;White solid; mp: 257\u2013258\u00a0\u00b0C; Yield: 92.0%;IR (KBr): 3286, 2923, 1702, 1673, 1650, 1597, 1440, 1369, 1324, 1256, 1200, 1178, 1089\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 2.34 (s, 3H, CH3), 5.43\u20135.44 (d, 1H, CH), 7.08 (s, 1H, CH), 7.25\u20137.42 (m, 8H, ArH), 7.92 (s, 1H, NH), 9.42 (s, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 21.43, 53.53, 112.23, 125.54, 126.88, 127.89, 128.69, 129.40, 131.05, 133.48, 136.03, 141.61, 142.18, 146.88, 151.60, 191.80;ESI-MS (m/z): 327.1 (M\u00a0+\u00a0H)+;Light yellow solid; mp: 270\u2013272\u00a0\u00b0C; Yield: 91.0%;IR (KBr): 3316, 3087, 2900, 1702, 1655, 1615, 1602, 1487, 1444, 1373, 1323, 1225, 1183, 1160\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 2.34 (s, 3H, CH3), 5.48\u20135.49 (d, 1H, CH), 7.02\u20137.11 (m, 3H, CH & ArH), 7.17\u20137.39 (m, 6H, ArH), 7.89 (s, 1H, NH), 9.38\u20139.39 (d, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 21.45, 53.47, 112.51, 113.51, 113.79, 114.40, 114.68, 122.70, 122.73, 128.58, 129.24, 130.69, 130.80, 136.06, 141.45, 141.89, 147.19, 147.28, 151.78, 161.03, 164.27, 191.82;ESI-MS (m/z): 311.1 (M\u00a0+\u00a0H)+;Light yellow solid; mp: 291\u2013293\u00a0\u00b0C; Yield: 92.0%;IR (KBr): 3327, 3199, 3096, 2916, 1700, 1658, 1617, 1602, 1507, 1438, 1372, 1322, 1225, 1182, 1157\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 2.33 (s, 3H, CH3), 5.45\u20135.46 (d, 1H, CH), 7.03\u20137.14 (d, 1H, CH), 7.09\u20137.11 (m, 2H, ArH), 7.22\u20137.24 (d, 2H, ArH), 7.35\u20137.39 (m, 4H, ArH), 7.85 (s, 1H, NH), 9.34\u20139.36 (d, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 21.43, 53.29, 112.91, 115.41, 115.70, 128.59, 128.80, 128.91, 129.26, 136.13, 140.71, 140.74, 141.45, 141.64, 151.73, 160.26, 163.49, 191.83;ESI-MS (m/z): 311.1 (M\u00a0+\u00a0H)+;Yellow solid; mp: 300\u2013302\u00a0\u00b0C; Yield: 90.0%;IR (KBr): 3334, 3208, 3089, 2933, 1691, 1649, 1615, 1438, 1368, 1320, 1252, 1209, 1179\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 2.17 (s, 3H, CH3), 2.19 (s, 3H, CH3), 2.34 (s, 3H, CH3), 5.35 (s, 1H, CH), 7.00\u20137.08 (m, 4H, CH & ArH), 7.24\u20137.27 (d, 2H, ArH), 7.37\u20137.40 (d, 2H, ArH), 7.78 (s, 1H, NH), 9.27 (s, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 19.47, 20.01, 21.43, 53.56, 113.18, 124.23, 128.04, 128.64, 129.36, 130.03, 135.77, 136.24, 136.51, 141.41, 141.47, 142.03, 151.79, 191.85;ESI-MS (m/z): 321.2 (M\u00a0+\u00a0H)+;White solid; mp: 262\u2013264\u00a0\u00b0C; Yield: 92.0%;IR (KBr): 3292, 2923, 1702, 1667, 1651, 1614, 1601, 1567, 1453, 1370, 1328, 1263, 1248, 1201, 1180, 1123, 1099, 1072\u00a0cm\u22121;\n1H NMR (300\u00a0MHz, DMSO\u2011d\n6): \u03b4 2.29 (s, 3H, CH3), 5.63 (s, 1H, CH), 7.04 (s, 1H, CH), 7.10\u20137.12 (d, 2H, ArH), 7.20 (s, 1H, ArH), 7.28\u20137.30 (d, 2H, ArH), 7.42\u20137.44 (d, 3H, ArH), 7.60 (s, 1H, NH), 9.19 (s, 1H, NH);\n13C NMR (75\u00a0MHz, DMSO\u2011d\n6): \u03b4 21.45, 53.88, 112.60, 123.59, 124.27, 129.36, 130.56, 135.85, 141.44, 141.67, 145.42, 151.89, 192.04;ESI-MS (m/z): 361.1 (M\u00a0+\u00a0H)+;In summary, we have successfully established a sustainable strategy for the synthesis of polyhydroquinoline and 6-unsubstituted dihydropyrimidinone derivatives using novel acidic ionic liquid [CEMIM][MSA] as recoverable and reusable catalyst in a single operation. The remarkable features of explored protocol are high catalytic efficacy of acidic ionic liquid [CEMIM][MSA] catalyst, minimum catalyst loading, green reaction profile, step and atom economic methods, shorter time period of reactions, economically feasible starting materials, workup less procedures, low generation of waste materials and purification free isolation of pure products with excellent yield.\nPriyanka Patil: Formal analysis, Methodology, Investigation. Suresh Kadam: Supervision, Software, Project administration. Dayanand Patil: Resources, Data curation, Validation. Paresh More: Conceptualization, Writing \u2013 review & editing, Writing \u2013 original draft, Supervision.None.The authors thank the SARTHI of Maharashtra for the funding provided through CSMNRF-2021/2021-22/896. The authors are also grateful to Solapur University, Punjab University, IIT Bombay and IIT Madras for spectral measurements. Further authors are thankful to the Principal and the Management of the V. G. Vaze College Autonomous for the laboratory facilities.\n\n\n\nSupplementary material\n\nImage 87\n\n\n\nCrystallographic data for the structures in this paper have been deposited with the Cambridge Crystallographic Data Centre as Supplementary Publication No. CCDC 2107726, 2,107,727, 2,107,728. Copies of the data can be obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, UK, (www.ccdc.cam.ac.uk/data_request/cif or e-mail: deposit@ccdc.cam.ac.Uk).", "descript": "\n                  Carboxylic acid functionalized imidazolium based novel acidic ionic liquid [CEMIM][MSA] was synthesized using economically feasible raw materials under very mild condition. The cost effective and sustainable synthesis of polyhydroquinoline and 6-unsustituted dihydropyrimidinone derivatives in green media was first time carried out very effectively using [CEMIM][MSA] catalyst. Acidic ionic liquid catalyst [CEMIM][MSA] with carboxylic acid group was easily separated and reused up to five cycles without much loss in its catalytic activity and stability. The introduced catalyst had showed a better catalytic performance in aqueous medium to obtain described compounds with excellent yields in shorter time. Purification free isolation of pure polyhydroquinolines and 6-unsustituted dihydropyrimidinones derivative makes the revealed protocol attractive and ecological.\n               "}
{"full_text": "", "descript": "\n                  The Ni\u2013MoO2 heterostructure was synthesized in suit on porous bulk NiMo alloy by a facile powder metallurgy and hydrothermal method. The results of field emission scanning electron microscopy (SEM), field emission transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) reveal that the as-prepared electrode possesses the heterostructure and a layer of Ni(OH)2 nanosheets is formed on the surface of Ni\u2013MoO2 electrode simultaneously after hydrothermal treatment, which provides abundant interface and much active sites, as well as much active specific surface area. The results of hydrogen evolution reaction indicate that the Ni\u2013MoO2 heterostructure electrode exhibits excellent catalytic performance, requiring only 41 mV overpotential to reach the current density of 10 mA/cm2. It also possesses a small Tafel slope of 52.7 mV/dec and long-term stability of electrolysis in alkaline medium.\n               "}
{"full_text": "Electrochemistry plays an essential role in the revolution from the current fossil-fuel-dominated energy market to a green, affordable, and sustainable energy economy because of its great potential to convert and store electricity from renewable energy sources [1]. For instance, hydrogen generation by electricity provided from renewable resources in conjunction with fuel cells and rechargeable batteries represent promising technologies for our future energy transformation. Catalysis is at the core of the processes taking place in these electrochemical devices. Their large-scale application is yet limited by low catalytic efficiency and high costs of electrode materials [2]. To address these problems, computational researchers have spent considerable efforts on the fundamental understanding of reaction mechanisms because a profound understanding of the elementary processes on the atomic scale may ultimately aid the rational design of multi-functional catalysts for the application in practice.Nanostructured catalysts have attracted tremendous interests for various electrochemical processes, such as water splitting [3,4], nitrogen fixation [5], carbon dioxide reduction [6], or rechargeable batteries [7,8], due to their large specific surface area, flexible structure, and unique electronic properties. To maximize these advantages, different nanostructures have been developed and designed, including zero-dimensional (0D) nanocages [9\u201311] and nanodots [12\u201314], one-dimensional (1D) nanotubes [15] and nanowires [7], two-dimensional (2D) nanosheets [16], three-dimensional (3D) nanocups [17], or mixed combinations thereof [18,19]. The precise control of structure and composition of catalysts can greatly boost their catalytic performances. Heterostructure is a typical nanostructure, which couples two or more low-dimensional nanostructures through van der Waals interactions [20]. This kind of architecture has attracted great interest in the catalysis community owing to their extraordinary chemical and physical properties [21]. Well-defined composites can work synergically at the interface to effectively tune the binding strength of reaction intermediates, and this may lead to an improved catalytic performance [22,23]. Another widely developed nanocomposite refers to the core-shell structure, which also has shown excellent advantages in many energy conversion applications [24\u201326]. This type of structure consists of two different constituent phases, and one of them serves as the core while the other acts as the shell layer to cover the core. A smartly constructed core-shell structure can lead to specific features at the interface. For example, Raza et\u00a0al. [27] reported a ceria-based nanocomposite in which ceria forms the core and a salt acts as the shell for the application as a low-temperature solid-oxide fuel cell. This functional nanocomposite exhibited extremely fast ion transport at around 300\u00a0\u00b0C with ionic conductivities exceeding 0.1\u00a0S/cm at the interfaces. Apart from the combination of different composites, nanostructured catalysts with diverse morphologies have reported to reveal outstanding performances in electrochemical devices. Liu et\u00a0al. [28] proposed a tetragonal VO2 hollow nanosphere as a cathode material for aqueous zinc-ion batteries. These hollow nanospheres not only facilitate Zn ion migration\u00a0but can also enhance the stability and conductivity of the electrode. Moreover, a newly developed nanocatalyst with flower-shaped structure has been reported [29]. The high surface-to-volume ratio and enhanced surface sensitivity and stability allows wide applications in the area of nanotechnology [30]. Another example refers to the work by Li et\u00a0al., who synthesized a trimetallic carbon nanoflower electrocatalyst with adjustable Co2+/Fe2+/Ni2+ ratio within a metal\u2013organic framework. The synergistic electronic effects between the different components result in an excellent overall water-splitting activity [31].The abovementioned nanostructured catalysts share the common advantage of high surface area, which can provide abundant active sites, and this may result in a larger current density for the electrocatalytic process at a fixed electrode potential. According to the Butler\u2013Volmer equation (cf. Eq. (1)) [32,33], an \u2018ideal\u2019 nanostructured catalyst should reveal large number of active sites, \u0393act, but at the same time also a small free energy of activation, G\nrds\n#. In Eqs. (1) and (2), j, j\n0 (cf. Fig.\u00a02), and \u03b7 indicate current density, exchange current density, and applied overpotential, respectively. Otherwise, e, k\nB, T, (\u03b3\u00a0+\u00a0rrds\n\u03b1\nrds), h, and z denote the elementary charge, Boltzmann's constant, absolute temperature in K, the number of electrons transferred before the transition state with highest free energy, Planck's constant, and the overall number of electrons transferred in the reaction, respectively. The catalytic properties of an electrode materials are mainly captured by the terms \u0393act and G\nrds\n# (vide supra)\u00a0and to a practically negligible extent also in the term (\u03b3\u00a0+\u00a0rrds\n\u03b1\nrds).\n\n(1)\n\n\nj\n\n(\n\u03b7\n)\n\n=\n\nj\n0\n\n\n{\n\nexp\n\n(\n\n\n\n(\n\n\u03b3\n+\n\nr\n\nr\nd\ns\n\n\n\n\u03b1\n\nr\nd\ns\n\n\n\n)\n\ne\n\u03b7\n\n\n\nk\nB\n\nT\n\n\n)\n\n\u2212\nexp\n\n(\n\n\n\u2212\n\n(\n\nz\n\u2212\n\u03b3\n\u2212\n\nr\n\nr\nd\ns\n\n\n\n\u03b1\n\nr\nd\ns\n\n\n\n)\n\ne\n\u03b7\n\n\n\nk\nB\n\nT\n\n\n)\n\n\n}\n\n\n\n\n\n\n\n(2)\n\n\n\nj\n0\n\n=\n\n\n\nk\nB\n\nT\nz\ne\n\n\u0413\n\na\nc\nt\n\n\n\nh\n\n\nexp\n\n(\n\n\n\u2212\n\nG\n\nr\nd\ns\n\n#\n\n\n\n\nk\nB\n\nT\n\n\n)\n\n\n\n\n\nDespite various types of nanostructured catalysts have been developed and synthesized, it is still a formidable task to precisely describe their structure\u2013performance relationship under reaction conditions, as well as to steer \u0393act and G\nrds\n# for the rational design of highly efficient electrocatalysts.Currently, theory-based catalyst discovery takes an increasing impact on the development of electrode materials for experimental investigations [34]. Electronic structure calculations in the density functional theory (DFT) approximation combined with thermodynamic analyses of binding energies have been well accepted as a rational tool to suggest electrode compositions to experimentalists. However, there are still many gaps between theory and experiment, on the one hand referring to the understanding of reaction mechanisms\u00a0and on the other hand referring to the accurate prediction of catalytic activity [35,36]. In this perspective, we briefly discuss pitfalls and challenges of theoretical modeling for the description of nanostructured catalysts with application in energy conversion and storage, and we point out potential solutions to fill these gaps.DFT is currently the most used approach to study electrochemical interfaces [37,38]. The first and probably the most important step to accurately predict the properties of electrocatalysts is to build a model that is as close as possible to the actual structure encountered in experiments. Given that DFT can cope with a few hundred atoms only, the theoretically constructed models are yet too simple to reflect the real structure under operando conditions. This finding is even more pronounced when considering that typical nanoparticles (NPs) in experiments contain several thousand atoms, which is clearly beyond the scope of DFT calculations. Recalling that the computational costs increase exponentially as soon as the number of atoms rises, so far there is a complex trade-off between theoretical models and the real structure in experiments, especially when calculating electronic properties such as band structures.Calle-Vallejo et\u00a0al. [39] theoretically investigated finite-size effects of adsorption energies on Pt NPs\u00a0with different sizes. The authors reported that the adsorption energies of intermediates in the oxygen reduction reaction (ORR) could differ by about 0.5\u00a0eV for different NP sizes and extended surface as depicted in Fig.\u00a01\na [40]. Such large deviations, however, can lead to unreliable and even completely wrong guidelines to devise new catalysts. To address this issue, Calle-Vallejo et\u00a0al. developed a \u2018coordination-activity plot\u2019 method to predict trends of adsorption energies and resolve the geometric structure of optimum active sites. The effects of first-nearest neighbors and second-nearest neighbors are considered both by applying different weights, summarized in the concept of generalized coordination numbers (\n\n\n\nC\nN\n\n\u00af\n\n\n). As shown in Fig.\u00a01b, the volcano map is constructed by compiling the coordination numbers and potential-determining steps of the reaction. On the left side of the volcano, low coordination is accompanied by too strong bonding whereas high coordination results in too weak bonding at the right volcano leg. The optimum binding strength is located around \n\n\n\nC\nN\n\n\u00af\n\n=\n8\n\n, and this coordination outperforms the Pt(111) surface (\n\n\n\nC\nN\n\n\u00af\n\n=\n7.5\n\n). In summary, the concept of coordination provides a path in heterogeneous catalysis for the design of optimum surface sites at affordable computational costs.The concept of generalized coordination numbers can also be used to describe the relationship between catalytic activity and various morphologies of NPs [41]. Fig.\u00a01c indicates four different types of shapes, including convex, concave, frame, and cross-shaped NPs. The most active site on these NPs are identified and compared with the flat Pt(111) surface. It turns out that the convex sites are less active than the concave sites for the ORR due to their smaller coordination number, culminating in a volcano trend that qualitatively explains the different catalytic performance of Pt nanostructures (cf. Fig.\u00a01d).Electrochemical experiments on the laboratory scale take place at 25\u00a0\u00b0C in an aqueous electrolyte with a well-defined pH value and a well-defined electrode potential. Quite in contrast, DFT calculations are commonly performed at 0\u00a0K under vacuum conditions, and the effects of pH and potential are often neglected in the computations. However, changes in pH and potential can greatly influence the structure and composition of electrocatalysts.Operando conditions in experiments are challenging for ab initio theory, particularly since the elementary reaction steps may take place on several surface facets simultaneously [42]. It is a common approximation in theory to determine the energetically favored surface facet in terms of surface energy and to build a slab model for this surface termination [43]. A potential opportunity to consider the presence of several surface facets refers to the Wulff construction [44]. While relying on surface energy calculations, this method is regarded as an effective tool to understand and predict the shape of NPs and its accuracy has been further improved recently [45,46]. In this context, we would like to point out the work of Opalka et\u00a0al. [47] as a prototypical example. While the authors made use of this advanced approach, even the prediction of NP shapes by the Wulff construction is accompanied with pitfalls, as outlined in the following.Opalka et\u00a0al. [47] systematically investigated the shape transformation of IrO2 NPs under application of an electrode potential when moving toward the potential regime of oxygen evolution reaction (OER). Typically, NP models are constructed at U\u00a0=\u00a00\u00a0V vs. SHE (standard hydrogen electrode). Opalka et\u00a0al., however, reported that the equilibrium shape of a IrO2 NP is strongly dependent on the applied electrode potential. While for U\u00a0=\u00a00\u00a0V vs. SHE the IrO2 NP exhibits two apical facets, namely the (100) and (110) terminations, at potentials exceeding U\u00a0=\u00a00.9\u00a0V vs. SHE, the (111) facet becomes thermodynamically stable. When increasing the potential further to U\u00a0=\u00a01.3\u00a0V vs. SHE, the entire NP forms (111) facets only. The reason why the preferred surface facet changes at different potentials is related to the binding strengths of the intermediate species on the electrode surface. The (111) facet of IrO2 NP is more favorable than the (110) facet at U\u00a0=\u00a01.3\u00a0V vs. SHE because the (111) facet binds surface oxygen, \u2217O, stronger than the (110) facet, and therefore, the surface energy of the (111) facet excels that of the (110) facet (cf. Fig.\u00a02). Consequently, the OER over IrO2 NPs commences from the (111) facet rather than on the (110) facet under experimental conditions. In contrast, in ab initio theory the reaction mechanism is commonly studied for the (110) facet [48\u201350], which can result in erroneous conclusions on mechanism and activity when comparing the outcome to experimental studies of nanosized systems.The above finding causes severe efforts for theoreticians to cover the following factors into the analysis of nanosized systems: a) a variety of surface facets need to be considered in the calculations, and even facets that are not stable at U\u00a0=\u00a00\u00a0V vs. SHE have to be taken into account because they can possibly be stable for elevated electrode potentials; b) for all these surface facets, a sufficient number of representative surface structures with different adsorbates (e.g.\u00a0OH, O, and OOH when moving toward anodic potential conditions) need to be calculated. Therefore, the resulting parameter space is enormous, and thus the corresponding computations can easily surmount the level of feasibility.In the following, we want to point out a methodical aspect in the calculation of surface structures with applications in energy conversion and storage. In most of the cases, the entire slab model is treated as a charge neutral system by using the concept of computational hydrogen electrode (CHE), developed by N\u00f8rskov and co-workers [51]. The CHE method regards that a proton-electron pair will transfer from the charge-neutral system into the electrode reservoir in each elementary step, but the description of charged systems or calculations under applied bias are not possible by this method. Consequently, scientists have made considerable efforts to move from a constant-charge description, as encountered with the CHE, to a constant-potential formalism, also denoted as grand canonical approach [52]. For instance, Gao et\u00a0al. [53] proposed a \u2018fixed potential\u2019 method to overcome this problem, in which the electrode potential (Fermi energy) is fixed while the total number of electrons can vary at the atomic level. They found that different electrode potentials will greatly influence the local electronic structures and binding environments, and this may alter catalytic activity in agreement with experimental studies [54]. Given that the binding energies obtained by the constant-charge and the fixed-potential methods can differ up to 0.50\u00a0eV [54], it appears that the conventional CHE approach is outdated and too inaccurate.Another important aspect that the CHE formulism does not address refers to the electric double layer (EDL) at the interface, which plays a vital role in heterogeneous catalysis. Modeling the EDL is a quite complex and challenging task given that not only the effect of solvent and electrolyte solutions but also the electrode potential needs to be considered [55]. Currently, ab initio molecular dynamics (AIMD) simulations are widely used to simulate the structures and dielectric properties of the EDL. Cheng et\u00a0al. used this approach to investigate electrified metal/water interfaces, such as the Pt(111)/water and Ag(111)/water interfaces, to unravel structure and capacitive behavior [56\u201358]. However, by the inclusion of ions at the interface the computational costs for AIMD increase, and accuracy may decrease given that this method is unable to consider long-range electrostatic interaction of charged electrolytes and electrodes [59]. In recent works, Zhang and coworkers connected the AIMD approach with machine learning (ML) models to speed up the convergence of the polarization, P, which lowers computational costs for the explicit modeling of electrified interfaces [60\u201363]. These efforts contribute to move the theoretical description of electrified interfaces closer to the experimental situation under operando conditions.Despite the critical discussion on the CHE method because of its simplicity, it is noteworthy to refer to the work of H\u00f6rmann et\u00a0al. [64]. The authors reported that the CHE approach is a first-order approximation to a fully grand canonical approach, and the results obtained by the CHE method are in qualitative agreement to a grand canonical ensemble, except for adsorbates with large dipole moments such as halides. This pinpoints that the CHE approach is yet valid for most electrocatalytic systems including the hydrogen and oxygen electrocatalysis, which are of high importance to energy storage and conversion. In summary, the community does not have a unified view on the application of computationally cost-effective charge-neutral methods (such as the CHE approach) and computationally more demanding grand canonical schemes, and there is a trade-off between the application of DFT to tackle the energetics of catalytic processes or AIMD to describe the electrochemical interface accurately [59\u201362].Another challenge refers to the description of the aqueous electrolyte in DFT calculations. While in experiments, the catalytic processes take place at the solid/liquid interface, which is prone to alter under operational conditions, it is still a common consensus to apply gas-phase DFT calculations to approximate the complex solid/liquid junction. One reason for this is that there is no unified method available of how to treat the solvent in ab initio electrochemistry. The solvent can be modeled by adding one or two solvent layers over the catalyst surface or by embedding a solvent environment on the whole reaction system, and these two different approaches refer to an explicit or implicit description, respectively. Apparently, the explicit solvent method provides a more accurate description of the reaction system at the expense of higher computational costs [65,66]. On the other hand, the dynamic nature of the solvent when modeled explicitly is not accounted for by DFT, and this may lead to artifacts because formally, averaging over all possible water configurations need to be carried out. The latter can only be achieved by computationally demanding AIMD simulations [67].Due to their lower computational costs, implicit solvent models are very popular in the computational electrochemistry community. A recent review by Ringe et\u00a0al. summarizes the application of implicit solvation models for the incorporation of solvent effects with applications in catalysis [38]. As shown in Fig.\u00a03\n, the solid\u2013liquid interface (SLI) can be simplified from a fully explicit quantum mechanical description to a fully implicit model by parametrization of each coarse-grained level. The implicit solvation approach accounts for the surrounding liquid electrolyte on the level of a continuous polarizable medium. This method can also mimic polarization of the electrode's electronic density under applied potential conditions and the concomitant capacitive charging of the entire double layer. Therefore, the implicit solvation scheme allows electronic structure calculations at the SLI at moderate computational costs compared with the accurate treatment of the solvent by AIMD. Despite its computational efficiency, the implicit solvation approach also reveals drawbacks that should be pointed out. These concerns address the extensive parametrization needed and the issue of transferability from one to another system, thus causing erroneous results in the worst case.An intimate interplay of experiment and theory may spur our understanding of electrochemical processes [68], and this is key for the development of catalysts with applications in energy storage and conversion. Synergy effects between theory and experiment can greatly accelerate the time-consuming process of catalyst exploration. Yet, to take the full advantage of theoretical approaches, the apparent size gap between theory and experiment needs to be carefully addressed.One key point is the complex trade-off between accuracy and computational costs. As discussed in the above, DFT is the method of choice for the investigation of catalytic processes at electrified solid/liquid interfaces, but it reaches its limit as soon as a few hundred atoms are part of the calculations. To address systems of larger size, the density functional tight binding method (DFTB) method is widely used. This approach is two to three orders of magnitude faster than DFT [69,70] due to relying on a semi-empirical approach that can handle thousands of atoms [71,72]. While the DFTB method may be a solution to model real NPs encountered in experiments, and thus, to close the gap between experiment and theory, it lacks accuracy compared with DFT, recalling that DFT can also only describe trends qualitatively but barely quantitatively [73].To our opinion, one promising opportunity to approach the gap between theory and experiment relating to nanocatalysts is the incorporation of ML\u00a0techniques (cf. Scheme 1\n). ML is a branch of artificial intelligence. While experiments and theory both can easily generate tons of data, the intrinsic correlations between data points often remain unknown. This is where ML comes into play because this method of data analysis is able to identify patterns, mine valuable information from data sets, and can train models, thus allowing to speed up research activities as discussed in the following.While relying on the DFT level to maintain reasonable accuracy, ML may contribute to expand the number of calculations conducted with the caveat that theoretical models need to be sufficiently trained by reference data. More precisely, a thorough assessment of the solvent contribution by the training of implicit solvation models (cf. Fig.\u00a03), the training of canonical or grand canonical approaches to assess their impact on the energetics of adsorbate species, and the extension of calculations to higher index surface facets for an accurate NP shape prediction under applied bias (cf. Fig.\u00a02) are of importance and relevance for nanocatalysts. Activity trends can then be derived by concepts such as the generalized coordination number (cf. Fig.\u00a01) or advanced activity descriptors that go beyond the prototypical thermodynamic reasoning [36,74,75]. Taking all these subtleties into account will be beneficial in that theoretical predictions and models are closer to the real systems in experiments [76\u201378].Nanocatalysts are auspicious for various energy storage and conversion applications and have already led to some achievements by merging computational and experimental methodologies. In this perspective, we have focused on the current gaps between computational modeling and real experimental conditions, and we have identified three main pitfalls for theoreticians. This comprises discrepancies between theoretical models and real structures, deviations between computational and experimental conditions, and the simplicity of computational protocols in terms of the electrode potential and solvation. If realistic models are studied under conditions that refer to the experimental ones by accounting for applied bias and solvation in an electrochemical environment, with the currently available computational resources we do not have a chance to tackle all the outlined pitfalls. To our opinion, machine learning (ML)\u00a0techniques are encountered as a gamer changer in this regard because they exhibit great potential to overcome the associated gaps between theoretical and experimental studies. Yet, one should pay attention to that fact that ML strongly depends on the reference data derived from density functional theory(DFT)\u00a0calculations and ab initio molecular dynamics(AIMD)\u00a0simulations. Therefore, success of this approach is only guaranteed for reliable and sufficiently large data sets. To foster further breakthroughs to fill the gap between theory and experiment, we suggest that the development of new methods for the modeling of electrochemical systems by DFT or AIMD should go hand in hand with the ML approach.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.KSE acknowledges funding by the Ministry of Culture and Science of the Federal State of North Rhine-Westphalia (NRW Return Grant). KSE is associated with the CRC/TRR247: \u201cHeterogeneous Oxidation Catalysis in the Liquid Phase\u201d (Project number 388390466-TRR 247), the RESOLV Cluster of Excellence, funded by the Deutsche Forschungsgemeinschaft under Germany's Excellence Strategy \u2013 EXC 2033\u2013390677874 \u2013 RESOLV, and the Center for Nanointegration (CENIDE). This article is based on the work from COST Action 18234, supported by COST (European Cooperation in Science and Technology).", "descript": "\n                  Computational approaches to describe catalysts under electrochemical conditions are steadily increasing. Yet, particularly the theoretical description of nanostructured catalysts, which have the advantage of a high surface area or unique electronic properties through refined synthetic protocols, is still hampered by the occurrence of pitfalls that need to be circumvented. In this perspective, we aim to introduce the reader to common pitfalls in the modeling of nanostructured catalysts with applications in energy conversion and storage, and we discuss the application of machine learning techniques as a potential solution to overcome the associated gaps.\n               "}
{"full_text": "No data was used for the research described in the article.CN and CC coupling reaction has gained immense interest recently as they find importance in synthesis of fine chemicals, natural products, pharmaceuticals, agrochemicals and functional materials [1]. In drug discovery, CN and CC coupling reactions are generally palladium catalyzed. As these heterocyclic products are found in various biologically active compounds & natural products there is a need to develop a facile method for CN bond forming reaction. Traditionally, palladium catalyzed CN coupling is more efficient & used widely [2\u20135], There are some limitations of this method such as high cost of the catalyst & availability of palladium. Use of palladium catalyst in the synthesis of API\u2019s can lead to contamination of drug substances. Hence, there is a need to find alternative catalytic method for CN coupling. Thus, use of non-noble transition metals such as copper [6\u20139], nickel [10\u201311], iron [12], cobalt [13] have found importance in such type of reactions.[14\u201324]. Yousefi et al. [21] reported Fe3O4@PVA/CuCl catalyst for CN coupling giving 90\u00a0% yield of the product in DMF at 110\u00a0\u00b0C whereas Sardarian et al. [22] reported Fe3O4@SiO2/Schiff base-Cu(II) NPs catalyst for CN cross coupling reaction for CN coupling reaction which gave 86\u00a0% yield at 100\u00a0\u00b0C in DMF. Mallick et al. [19] reported Cu-gCN catalyst in toluene at 100\u00a0\u00b0C giving 80\u00a0% yield of CN cross coupled product. Rout et al. [6] reported CuMoO4 nano-catalyst for CN coupling in DMSO at 90\u00a0\u00b0C giving 82\u00a0% yield of the product. Bazgir et al. [15] reported Cu@Cu2O nanoparticles on reduced graphene oxide and their catalytic activities in N-arylation of N-heterocycles giving 90\u00a0% yield in DMSO at 110\u00a0\u00b0C. Addition of various ligands increases the rate of the reaction but the disadvantage of using ligands is increase in additional cost for CN coupling reaction [25\u201327]. However, there are various disadvantages of these methods such as use of stoichiometric amount of metal reagents & use of expensive ligands which restricts its synthetic utility. Although reports are available for CN coupling reaction, it is still important to develop a ligand free, sustainable, economical & environmentally friendly method. Use of carbon-based heterogeneous catalysts is advantageous over homogenous catalysts such as ease of preparation, ability to incorporate various functionalities on the surface of the catalyst, separation, reusability, less cost, low toxicity etc.There are also reports for synthesis of various metal nanoparticles for number of applications. Metal nanoparticles are prepared from various sources such as flower extract, marine debris, fungus, vegetable waste, different species collected from sea coast, marine algae, plant extract, leaf extract, fruit extract & different types of biomass. Metals such as Cu, Fe, Ag, Pd, Au, Ni, Zr, etc were then loaded on it to form metal nanoparticles of various shapes & sizes. These nanoparticles have been used for various biomedical applications, which include antibacterial, antioxidant, free radical scavenging, antifungal, anticancer, larvicidal activity and showed biocompatibility properties. Other applications of metal nanoparticles include photocatalytic degradation of azo dyes & degradation of tetracycline and ibuprofen molecules [28\u201344].In continuation of our work, in the area of heterogeneous catalysis [45,46], we reported a ligand free approach towards CN coupling. Cobalt is found to be stable, non-toxic metal and can be considered as an effective alternative to palladium for CN coupling. There are very few reports available in literature for CN coupling using cobalt based heterogeneous catalysts, hence it\u2019s important to develop cobalt-based heterogeneous catalyst and study its utility for CN coupling reaction [47,48].In this work, we synthesized cobalt immobilized carbon-based nano-catalyst Co@CC by carbonization of glucose, its functionalization followed by immobilization of cobalt. The catalyst has been characterized using various characterization techniques. Cobalt based catalysts Co@CC was evaluated for CN coupling of several amines & aryl halides. Co@CC11, Co@CC12, Co@CC13 & Co@CC14 catalyst were synthesized by varying ratio of CC and CoCl2 ratio from 1:1 to 1:4 and were denoted as Co@CC11, Co@CC12, Co@CC13 & Co@CC14 respectively.Carbonaceous catalyst has gained lot of importance now-a-days as they can be prepared easily. Due to high density of oxygen functional groups present on surface of the catalyst, it is easy to immobilize metal on it. Carbon-based catalysts are highly stable, reusable, economical and eco-friendly. In this work, we synthesized cobalt immobilized carbonaceous catalyst and studied its catalytic activity for CN coupling of aryl halides & amines. Cobalt based carbonaceous solid acid catalyst was prepared using readily available cheaper source glucose. Glucose was heated with p-TSA to form carbonaceous material containing sulphonic acid (-SO3H), phenolic (\u2013OH) and carboxyl acid (\u2013COOH) functional groups. It was further treated with 3-aminopropyltrimethylsilane to introduce -OSi-CH2CH2CH2NH2 groups followed by treating with CoCl2 for immobilization of cobalt. The Co@CC12 catalyst was characterized with FT-IR (Spectrum 400), P-XRD (Rigaku MiniFlex 600) Power: 100\u2013240 VAC 1\u0278 15A 50/60\u00a0Hz, EDAX (Bruker), SEM (Quanta scanning electron microscope), HR-TEM (Jeol JEM F200), 13CP-MASS (Jeol ECX 400), XPS (Thermo K \u03b1) & BET (Quantachrome) techniques.As seen from FT-IR (Fig. 1\na), CO bonding peak at 1030\u00a0cm\u22121 and OS\u00a0=\u00a0O stretching vibration peak at 1157\u00a0cm\u22121 was observed. CH bending vibration peak for alkyl groups was observed at 1438\u00a0cm\u22121 & CC stretching vibration peak was observed at 1630\u00a0cm\u22121 for aromatic carbons. CO peak at 1703\u00a0cm\u22121 was observed for acid functionality. Strong peaks were observed for \u2013NH groups at 1582\u00a0cm\u22121 & for \u2013OH groups at 3205\u00a0cm\u22121. P-XRD (Fig. 1b) displayed broad peak at 2\u03b8\u00a0=\u00a015\u201330\u2070 for amorphous carbon with random orientation of aromatic carbon sheets and peaks at 2\u03b8 of 64.9\u2070, 58.9\u2070, 51.3\u2070, 39.7\u2070, and 29.0\u2070 were ascribed to the cobalt species. EDAX analysis (Fig. 1c) showed elemental composition of C, O, N, Si, S and Co to be 59.5\u00a0%, 23.9\u00a0%, 8.1\u00a0%, 3.3\u00a0%, 0.1\u00a0% and 5.1\u00a0% respectively. SEM analysis (Fig. 1d) showed heavy crumpling features of cobalt based nano-catalyst. HR-TEM analysis (Fig. 1e) showed well dispersed nanoparticles of particle size of 25\u201350\u00a0nm. 13CP/MAS NMR (Fig. 1f) showed signals for N-propyl groups from 11\u00a0ppm to 61\u00a0ppm. Broad signals for polycyclic aromatic carbon atoms, \u2013OH and \u2013COOH groups were seen at 129, 151, 172 & 209\u00a0ppm respectively.As seen from Fig. 2\n, HR-TEM elemental mapping data showed deposition of cobalt over carbon based support.XPS analysis (Fig. 3\n) showed binding energy peaks at 533\u00a0eV & 285\u00a0eV for O and C for Co@CC12. Presence of N and incorporation of 3-amino propyl trimethoxy group was confirmed from binding energy peak at 400\u00a0eV. Presence of Cl, S & Si was confirmed from peaks at 201, 167.5 and 103\u00a0eV respectively. Cobalt showed two spin orbital doublets at 797.1\u00a0eV in Co2p1/2 and 781.5\u00a0eV in Co2p3/2 confirming presence of Co3+ and Co2+ species. ICP analysis showed 3.92\u00a0% of cobalt in the catalyst. Surface area from BET analysis was found to be 15.5\u00a0m2/g.Co@CC12 was studied for CN coupling of indole & bromo-benzene using potassium hydroxide as a base. Various solvents like sulfolane, dimethylsulfoxide, water, tetrahydrofuran, \u03b3-valerolactone, toluene, acetonitrile & dimethylformamide were studied for the reaction (Table 1\n). The reaction was found to proceed only in sulfonated solvents sulfolane and dimethylsulfoxide giving 64\u00a0% and 35\u00a0% yield respectively. The reaction did not proceed in remaining solvents. Sulfolane gave higher yield of cross-coupled product. Sulfolane being thermally stable can be operated within a wide range of reaction conditions and reaction temperatures. The reaction when carried out under homogenous conditions using CoCl2 as a catalyst in sulfolane as a solvent showed no formation of cross coupled product.We also carried out reaction using CC catalyst (without cobalt) and it was found that only traces of product were observed during this reaction. We then studied the effect of different bases and their concentration for CN coupling of indole & bromo-benzene in sulfolane as a solvent using 20\u00a0wt% of Co@CC12 catalyst at 120\u00a0\u00b0C. We tried bases such as potassium tert-butoxide (KO\nt\nBu), potassium triphosphate (K3PO4), lithium hexamethyldisilazane (LiHMDS) and also studied various concentrations of base KOH (Table 2\n).It was found that the reaction proceeded in KO\nt\nBu base giving 41\u00a0% yield whereas the reaction did not proceed at-all in bases K3PO4 and LiHMDS. KOH was studied at different concentrations of 1\u00a0mmol to 4\u00a0mmol and it was observed that 4\u00a0mmol of KOH was found to give highest yield of cross-coupled product. After optimizing the solvent and concentration of base we then focused on optimizing various catalysts and catalyst concentration. The Co@CC12 catalyst was screened for various catalyst concentrations i.e. 10, 20, 30 & 40\u00a0wt% for reaction of indole & bromo-benzene using KOH as a base in sulfolane as a solvent (Table 3\n). The results showed that as the catalyst concentration was increased, yield of product also increased from 59\u00a0% to 74\u00a0% (Table 3, entries 1\u20133). At catalyst concentration of 40\u00a0wt% using Co@CC12 catalyst the reaction did not proceed at all at rt, whereas the yield of product was increased from 67\u00a0% to 91\u00a0% as the temperature was raised from 80 to 150\u00a0\u00b0C in 35\u00a0h (Table 3, entries 5\u20137). The starting was completely consumed at 150\u00a0\u00b0C in 35\u00a0h to give 91\u00a0% yield. The Co@CC11, Co@CC13, Co@CC14 catalysts were screened using optimized reaction conditions used for Co@CC12 catalyst.It was observed that at 40\u00a0wt% concentration of Co@CC11 catalyst using KOH in sulfolane at 150\u00a0\u00b0C the yield was 65\u00a0% after 35\u00a0h (Table 3, entry 8). The yield of product using Co@CC13 & Co@CC14 catalysts was similar to that of Co@CC12 catalyst (Table 3, entries 9\u201310). Hence Co@CC12 catalyst gave maximum yield of CN cross coupled product at 150\u00a0\u00b0C after 35\u00a0h.Various different indoles and aryl halides were reacted using Co@CC12 catalyst to form respective CN cross coupled product (Table 4\n). Indole was reacted with various aryl halides such as chloro-benzene, bromo-benzene, iodo-benzene, 4-ethyl-iodo-benzene, 4-isopropyl-iodo-benzene. It was observed that indole on reaction with chloro-benzene gave 65\u00a0% cross coupled product in 35\u00a0h (Table 4, entry 1), whereas in case of reaction of indole with bromo-benzene and iodo-benzene, starting was completely consumed in nearly 35\u00a0h giving 91\u00a0% yield (Table 4, entries 2 & 3). Indole on reaction with 4-ethyl-iodo-benzene & 4-isopropyl-iodo-benzene after 35\u00a0h gave 55 & 81\u00a0% yield respectively (Table 4, entries 4 & 5). Next, we studied reaction of 3-methyl-indole with aryl halides such as bromo-benzene, 4-ethyl-iodo-benzene & 4-isopropyl-iodo-benzene (Table 4, entries 6\u20138). It was observed that in case of 3-methyl-indole and bromo-benzene, starting was completely consumed in 24\u00a0h giving 85\u00a0% yield of the product, whereas in case of 4-ethyl-iodo-benzene & 4-isopropyl-iodo-benzene after 35\u00a0h, 70 & 75\u00a0% yield of cross coupled product was observed. Reaction of 6-bromo-indole with bromo-benzene gave 88\u00a0% yield whereas the yield in case of 4-ethyl-iodo-benzene & 4-isopropyl-iodo-benzene (Table 4, 9-11) was less. 2-methyl-indole, 8-methyl-indole and 2-phenyl-indole were also reacted with bromo-benzene, but the reaction was found to be slow giving 69, 66 & 65\u00a0% yield of the product (Table 4, entries 12, 14, 15) whereas reaction of 2-methyl-indole and 2-phenyl-indole with 4-ethyl-iodo-benzene gave 62 & 60\u00a0% yield respectively (Table 4, entries 13 & 16).The products were purified and characterized using NMR and HRMS. It was observed that electronic properties of the precursors did not have any impact on the yield of the products. The catalyst was easily prepared from readily available natural source glucose and showed good catalytic activity for CN coupling reaction. The catalytic activity of Co@CC12 was compared with other metal based catalysts reported in literature for CN coupling (Table 5\n). Co@CC12 catalyst showed good yield with respect to the reported ones.The catalytic process is considered efficient and economical only when a catalyst can be recycled and reused. Co@CC12 catalyst was evaluated for its recyclability. After first use, the catalyst was filtered, washed and dried for further recycles. We studied it\u2019 s reusability up-to five cycles for cross coupling of indole and bromo-benzene and it was observed to give consistent yield of the CN cross coupled product even after 5 recycles (Fig. 4\n). FT-IR, P-XRD analysis of the recycled catalyst is provided in the ESI (Fig 58, 59). In FT-IR, we observed peaks for reused catalyst for functionalities such as CO, -SO3H, \u2013OH & \u2013NH groups whereas P-XRD also displayed broad peaks in the range of 2\u03b8\u00a0=\u00a015-30\u2070 for amorphous carbon with random orientation of aromatic carbon sheets and peaks for cobalt species. There were no significant changes observed in fresh and recovered Co@CC12 catalyst. ICP analysis of reused catalyst showed only 0.04\u00a0% of cobalt has been leached.In conclusion, we developed a cobalt based Co@CC12 nano-catalyst and studied its catalytic activity for CN coupling reaction. It was found to be highly efficient for CN cross coupling of various amines and aryl halides giving yields in the range of 60\u201391\u00a0% in sulfolane as a solvent at 150\u00a0\u00b0C. The catalyst was studied for 16 different CN cross coupling reactions using various amines & aryl halides. The catalyst was easy to recycle and reused up to five cycles with consistency in yield of the products. The cobalt based Co@CC12 catalyst is reusable, economical and environment friendly compared to much expensive palladium and toxic copper-based catalyst and the method developed for CN coupling reaction is noble metal-free & ligand-free.All commercially available reagents were used without further purification unless otherwise stated. Column chromatography was performed on silica gel (60\u2013120 mesh). 1H and 13C NMR spectra were recorded on a Bruker spectrometer using CDCl3 as a solvent and TMS as an internal standard. NMR data are reported as follows: chemical shift, multiplicity (s\u00a0=\u00a0singlet, bs\u00a0=\u00a0broad singlet, d\u00a0=\u00a0doublet, dd\u00a0=\u00a0doublet of doublet, t\u00a0=\u00a0triplet, td\u00a0=\u00a0triplet of doublet, m\u00a0=\u00a0multiplet), coupling constants (Hz), and integration. High-resolution mass spectra (HRMS) were recorded using electron spray ionization (ESI) with a time-off-flight mass analyzer.Glucose (5\u00a0g, 27.7\u00a0mmol) and p-toluene sulfonic acid (10\u00a0g, 58.1\u00a0mmol) were taken in a clean and dry round bottom flask and stirred at 180 \u2103 for 24\u00a0h under nitrogen. After 24\u00a0h, a black solid mass was obtained which was then washed with millipore water and then with ethanol to obtain a black powder. This black powder was then dried in the oven at 100 \u2103 for 5\u00a0h. After drying to this black powder was added 20\u00a0mL of ethanol and 3-amino propyl triethoxy silane (APTES) (6\u00a0mL), sonicated for 30\u00a0min and further refluxed at 80 \u2103 for 8\u00a0h. The solid was collected by filtration, washed with water, followed by ethanol, and dried in the oven at 100 \u2103 for 5\u00a0h. This black powder was denoted as CC. After drying, to this black powder CC (0.5\u00a0g) was added 10\u00a0mL ethanol, CoCl2\u00b7H2O (0.5\u00a0g) and further refluxed 80 \u2103 for 12\u00a0h. The solid was collected by filtration, washed with water, followed by ethanol, and dried in the oven at 100 \u2103 for 5\u20136\u00a0h. This black powder was denoted as Co/CC11 catalyst.Black powder (0.5\u00a0g) (CC) was added 10\u00a0mL ethanol, CoCl2\u00b7H2O (1\u00a0g) and further refluxed 80 \u2103 for 12\u00a0h. The solid was collected by filtration, washed with water, followed by ethanol, and dried in the oven at 100 \u2103 for 5\u20136\u00a0h. This black powder was denoted as Co/CC12 catalyst.Black powder (0.5\u00a0g) (CC) was added 10\u00a0mL ethanol, CoCl2\u00b7H2O (1.5\u00a0g) and further refluxed 80 \u2103 for 12\u00a0h. The solid was collected by filtration, washed with water, followed by ethanol, and dried in the oven at 100 \u2103 for 5\u20136\u00a0h. This black powder was denoted as Co/CC13 catalyst.Black powder (0.5\u00a0g) (CC) was added 10\u00a0mL ethanol, CoCl2\u00b7H2O (2\u00a0g) and further refluxed 80 \u2103 for 12\u00a0h. The solid was collected by filtration, washed with water, followed by ethanol, and dried in the oven at 100 \u2103 for 5\u20136\u00a0h. This black powder was denoted as Co/CC14 catalyst.Aryl amine (1\u00a0mmol) & aryl halides (1.5\u00a0mmol) were taken in a round bottom flask along with solvent sulfolane. Catalyst Co@CC12 (40\u00a0wt%) & base KOH (4\u00a0mmol) was then added to it and the reaction mixture was heated to desired temperature. After completion of the reaction from TLC, water was added to the reaction and the reaction mixture was extracted in ethyl acetate and the organic layer was evaporated on rota-vac & the crude product was purified using column chromatography.\nShubham R. Bankar: Methodology, Investigation. Swapnali P. Kirdant: Methodology, Data curation. Vrushali H. Jadhav: Supervision, Conceptualization, Funding acquisition.The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:Vrushali Jadhav reports financial support was provided by National Chemical Laboratory CSIR. Jadhav Vrushali reports a relationship with National Chemical Laboratory CSIR that includes: employment.The remaining authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.S.R.B thanks UGC-New Delhi 276/(CSIR-UGC NET DEC.2018) & S.P.K. thanks CSIR-New Delhi 31/011(1151)2020-EMR-I for providing research fellowship. Dr. V. H. Jadhav thanks CSIR-NCL for providing the start-up fund (MLP036926) & all the facilities.Supplementary data to this article can be found online at https://doi.org/10.1016/j.rechem.2022.100682.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  CN cross coupling reaction is very important in synthesis of pharmaceuticals, natural products, agrochemicals, fine chemicals and functional materials. Traditionally, palladium or copper metals are used for CN coupling reaction. As palladium is expensive, we developed cobalt immobilized carbon-based nano-catalyst Co@CC for CN coupling. In this work, we synthesized non-noble metal-based Co@CC nano-catalyst by carbonization of glucose, it\u2019s functionalization followed by immobilization of cobalt on the surface of the catalyst. The catalyst was well characterized. The CN cross coupling reaction of various aryl halides & amines using Co@CC nano-catalyst was optimized for solvent, reaction temperature & catalyst concentration conditions. The catalyst showed high catalytic activity for CN coupling of various aryl halides & amines to form aryl amines in good to excellent yield up to 91\u00a0% in sulfolane as a solvent at 150\u00a0\u00b0C. The catalyst showed recyclability up to 5 times. The method developed for CN coupling reaction was noble metal free, ligand free, recyclable, sustainable, economical & environmentally friendly.\n               "}
{"full_text": "Cellulose, the most abundant polymer in nature, has vital importance for modern industry [1]. Cellulose is biocompatible, biodegradable, bioadhesive, and nontoxic properties using in many applications such as food, cosmetics, detergents, textile, and pharmaceutical areas [2]. On the other hand, chitosan deriving from chitin is the second abundant polysaccharide in nature [3]. It should be noted that chitosan has remarkable antimicrobial properties extracted from shrimp and other crustaceans. It is a natural biocompatible polymer that has been extensively investigated in pharmaceutical research and the. Recently, composite cellulose/chitosan was prepared by chemical modification including chemical grafting or physical treatments [4,5]. In recent years, natural composites including transition metal oxide, have a broad range of applications in gas sensors, supercapacitors, lithium-ion batteries, solar cells, electrochromic coatings, composite anodes for fuel cells, dye-sensitized photocathodes, as a contrasting agent for magnetic resonance.The pharmaceutical compounds have played essential roles in the human life [6]. Drugs are widely used not only to treat and prevent disease in humans and animals but also to improve the growth rate of agricultural products [7,8]. In addition to all the benefits of medicinal compounds, they are able to seriously pollute the environment, because their residues in the aquatic media have the potential to create resistance to environmental bacteria, making diseases difficult to treat and a great threat to public health [9,10]. Ciprofloxacin is a widely used antibiotic owing to its good performance in treating diseases [11]. Ciprofloxacin is released inactive forms due to partial metabolism and high structural stability after consumption [9]. The presence of antibiotics in the water cycle has caused serious concern because of their destructive impacts on human health and permanent damage to aquatic ecosystems [12]. They can disrupt the natural life cycle of native microbes. Therefore, many efforts have been made to find efficient methods that can ultimately convert ciprofloxacin into small biodegradable molecules or directly decompose it into carbon dioxide and water [10].Photocatalysis processes are considered as one of the efficient environmentally friendly methods to overcome the above-mentioned problems of antibiotics [13]. Photocatalytic degradation has the advantages such as energy saving, simplicity, mild reaction conditions and cost-effectiveness in the removal of pharmaceutical contaminants [14]. It has opened up a landscape in order to protect the environment through the design of novel high-performance photocatalysts for the removal of the toxic organic pollutants [15]. The electrons and holes are generated in the presence of photocatalyst and light source that can activate H2O and O2 to produce reactive oxygen species. These active species degrade antibiotics into substances with lower or even non-toxic biotoxicity [16]. Until now, prominent researches have been performed on the fabrication of various nanostructures with high photocatalytic properties [17]. However, there is still a need to design highly efficient photocatalysts with significant photoinduced charge transfer capabilities and a special electron-hole separation system simultaneously [18].Nickel oxide (a p-type semiconductor with antiferromagnetic behavior) and nickel (conducting ferromagnetic metal) possess unique electrical, optical properties, low cost, and high stability and can be used in catalysis, rechargeable batteries, and fuel cells [14]. They have impressive electrical conductivity, high electrochromic efficiency, high electro-activity, high effective surface area, cheap and straightforward synthesis procedures, and a wide modulation range [19]. They have been widely used in photocatalysis, supercapacitors, lithium-ion batteries, dye sensitizer in solar cells, transparent electrodes, biosensors, etc.[16,20]One of the major problems in photocatalytic processes is high recombination rate of photo generated electron-hole pairs which reduces the amount of reactive species for the degradation of organic compounds. The supporting the photocatalysts onto the suitable supports is an efficient solution for decreasing recombination of electron\u2013hole pairs [21].Here, in continuing our works, micro-crystalline aldehyde cellulose and nano chitosan were extracted from wastes of barely and shrimp wastes, respectively [22,23]. Cellulose and chitosan can attach covalently via condensation reaction and also Ni/NiO nanocomposite was synthesized using Calotropis procera. Finally, by adding Ni/NiO to cellulose/chitosan, nano-biocomposite Chit-Cell@Ni/NiO was designed. The biosynthesized nanocomposite has been utilized for the photocatalytic degradation of ciprofloxacin under sunlight. This study provides a green cost-effective strategy for degradation of ciprofloxacin without the use of excipients such as peroximonosulfate, H2O2, and NaBH4. Parameters affecting the degradation efficiency such as ciprofloxacin concentration, catalyst dosage, and pH value have been investigated and optimized. The mechanism and pathways of ciprofloxacin degradation were proposed through identification of intermediates.Extraction of nano chitosan by chemical method [24]:10\u00a0g samples of raw shrimp shell waste, after washing and drying, was added to sodium hydroxide (2.0\u00a0M) in the ratio of 1:16 (w/v) and stirred for 2\u00a0day\u00a0at room temperature (pH must be 11\u201313). Then, the samples were washed with water to obtain pH\u00a0=\u00a06.5\u20138 after filtering. The sample was dried at 80\u00a0\u00b0C for 16\u00a0h.The sample from the first step was added to HCl (1.0\u00a0M) in the ratio of 1:16 (w/v) and stirred at pH\u00a0=\u00a01\u20132.5\u00a0at 25\u00a0\u00b0C. After 24\u00a0h and filtering, the obtained sample was washed for having pH\u00a0=\u00a06.5\u20138.0. Chitin was obtained after drying at 75\u00a0\u00b0C for 15\u00a0h.The obtained chitin was added to 50% NaOH in the ratio (1:10) (w/v) at room temperature. After 48\u00a0h, chitosan was obtained after filtering and washing to gain pH\u00a0=\u00a06.5\u20138.0.Chitosan was added to 1\u00a0M of NaOH (1% (w/v)) to have pH\u00a0=\u00a07, the precipitate was then washed several times with water and centrifuged at 5000\u00a0rpm for 30\u00a0min and dried. The obtained sample was dissolved in 2% acetic acid with 1:15 (w/v) and refluxed at 80\u00a0\u00b0C within 30\u00a0min. Then, chitosan nanoparticles were achieved after washing several times.4\u00a0g of C. procera wood powder was added to 200\u00a0mL of distilled water at 90\u00a0\u00b0C for 10\u00a0min. A KOH solution (0.01\u00a0M) was then added drop-wise (drop rate 1\u00a0mL\u00a0min\u20131) at room temperature to reach the reaction pH to 9.170\u00a0mL of the obtained extraction and 20\u00a0mL of NiSO4 solution (0.05\u00a0M) were added to 130\u00a0mL of distilled water. The mixture of reaction was heated to 80\u00a0\u00b0C for 2.5\u00a0h. The resulting Ni/NiO nanoparticles was separated by centrifuge, washed with water, and dried in oven under vacuum. The as-synthesized sample was heated by the furnace at 500\u00a0\u00b0C for 4\u00a0h.10\u00a0g of barely wastes was treated with 1\u00a0g of NaClO2 salt in 50\u00a0mL of water at pH value 4.0\u00a0at 75\u00a0\u00b0C for 2\u00a0h, then filtered and the residue was washed with distilled water and ethanol (95%), and dried in an oven at 50\u00a0\u00b0C for 13\u00a0h. Then, dried residues were extracted with KOH 10% at room temperature for 8\u00a0h. After filtration, residue was washed until them neutral, and then washed with ethanol (95%). Finally, the samples were dried in an oven at 50\u00a0\u00b0C for 20\u00a0h (Tajik et\u00a0al. [23]).Sodium metaperiodate (0.27\u00a0g) was added to 0.5\u00a0g of microcrystalline cellulose suspended in 60\u00a0mL of distilled water. The mixture was stirred at 50\u00a0\u00b0C in the dark for 9\u00a0h. After this, the remaining NaIO4 was decomposed by adding of glycerol. Finally, the product was washed with distilled water and dried at room temperature for 24\u00a0h [22].The 0.03\u00a0g of aldehyde cellulose and 6\u00a0g of nano chitosan were dispersed in water for 8\u00a0h at 80\u00a0\u00b0C. Then, 0.5\u00a0g of Ni/NiO nanocomposite was added and refluxed for 3\u00a0h at 80\u00a0\u00b0C.The degradation tests of ciprofloxacin were carried out under direct sunlight on sunny days with radiation intensity of 260\u2013280 Klux from 11 p.m. to 2 p.m. The Chit-Cell@Ni/NiO nano-biocomposite was dispersed in 50\u00a0mL of the aqueous solution of ciprofloxacin in darkness conditions for 30\u00a0min. Subsequently, they were exposed to direct sunlight and stirred for 20\u00a0min. The initial and final concentration of ciprofloxacin was measured by a UV\u2013Vis spectrophotometer (JENWAY) with the maximum absorbance intensity at 275\u00a0nm [25] using the corresponding calibration plot. The efficiency of degradation of ciprofloxacin was calculated using the following equation:\n\n\n\nDegradation\u200aefficiency\n\n\n(\n%\n)\n\n=\n\n\n\n\nC\n0\n\n\u2212\n\n\nC\nt\n\n\n\nC\n0\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n\nC\n0\n\n\n\n and \n\n\nC\nt\n\n\n (mg/L) are initial and final concentrations of ciprofloxacin, respectively.First, cellulose aldehyde and chitosan were extracted from barely and shrimp wastes and attached together through condensation reaction between amine group of chitosan and C=O group of cellulose and the formation of Schiff base ligand. Then green synthesis of Ni/NiO was performed from C. procera as a reductant and stabilizing agent. Finally, the Chit-Cell@Ni/NiO nano-biocomposite was gained by adding Cellulose/chitosan to Ni/NiO.XRD of X-ray diffraction studies of Chit-Cell@Ni/NiO nano-biocomposite exhibit some peaks at 2\u03b8 = 9.8 \u00b0and 2\u03b8 = 20\u00b0 relating to chitosan (Fig. 1a) [26]. The two well-defined crystalline peaks observed around 2\u03b8\u00a0=\u00a022\u00b0 and 35\u00b0 were typical of cellulose indicating the success of present of cellulose [27]. Amount of Ni and NiO in the sample is low and this matter confirming by EDAX (Fig.\u00a0S1). The Williamson-Hall (W\u2013H) method was used to estimate the crystallite size of the as-synthesized Ni/NiO nanocomposite. In the following W\u2013H equation, \u03bb is the Cu-k\u03b1 radiation wavelength, d the crystallite size, \u03b2 the peak with at the half maximum, \u03b8 the brag angle, the terms \u03b7 shows the strain in the crystallites while the term d shows the size of the crystallites [28]. The constant k (k\u00a0=\u00a01/d) commonly is about 1 and varies in the range from 0.8 to 1.39. In the case of \u03b7\u00a0=\u00a00 the W\u2013H equation reduces to the Scherer equation. The crystallite size varies as 1/cos\u00a0\u03b8 and stain varies as tan\u00a0\u03b8 from the peak width. By construction of the plot of (\u03b2 cos\u00a0\u03b8) versus (\u03b7 sin\u00a0\u03b8) a straight line will be got with slope \u03b7 and intercept (k\u03bb/d) [29]. The intercept of plot was used to estimate the crystallite size, about 45\u00a0nm. This plot was drowning\n\n(line equation: \u03b2 cos \u03b8=(k\u03bb/d) (\u0572 Sin \u03b8)\n\n\nThe structure changes of nano chitosan and Chit-Cell@Ni/NiO nano-biocomposite were considered by FT-IR (Fig. 1b). In nano chitosan spectrum exhibited the peaks at 3434, 2994\u00a0cm\u22121 were assigned the OH and CH, respectively [24]. 1655, 1312, and 1023\u00a0cm\u22121 corresponded to acetylated amino groups, C=N stretching, and C\u2013O\u2013C stretching, respectively. In Chit-Cell@Ni/NiO nano-biocomposite spectrum, along with the peaks relating to nano chitosan, other peaks have confirmed the presence of cellulose and Ni/NiO nanoparticles. Typical absorption peaks of cellulose around 3448 and 2912\u00a0cm\u22121 are visible due to, respectively, \u2013OH and \u2013CH stretching vibrations (Fig.\u00a0S2). Stretching vibration mode of pure NiO nanocrystals has reported in the regions 430\u2013490\u00a0cm\u22121 and 600\u2013700\u00a0cm\u22121 which showed blue shift with respect to bulk NiO due to quantum size effect (Fig.\u00a0S1),[30 ].EDAX of nanochitosan shows C, N, and O in its structure (Fig.\u00a02\na). The presence of Ni, C, O, and N in Chit-Cell@Ni/NiO nano-biocomposite was confirmed its successful synthesis (Fig.\u00a02b). Also. tables illustrate the elemental analysis of Chit-Cell@Ni/NiO nano-biocomposite and 52.33, 5.59, 24.97, and 17.12 of C, N, O, and Ni were obtained.TGA, DTG, and DTA were used to evaluate the extracted nano chitosan and Chit-Cell@Ni/NiO nano-biocomposite thermal behaviors. TGA graph of nano chitosan is also depicted in Fig.\u00a03\na, which demonstrates two degradation steps. The evaporation of adsorbed water was shown at the first step with a 5.8% weight loss at 50\u2013100\u00a0\u00b0C. It seems that the nature of chitosan is hydrophilic [31]. The second step involved the 63.49% weight loss at 200\u2013410\u00a0\u00b0C attributing the cleavage of glycosidic linkages via dehydration and the decomposition of chitosan polysaccharides. TGA of Chit-Cell@Ni/NiO nano-biocomposite shows three steps. First, one was related to moisture evaporation (6.3% loss weight), and the second one corresponded to thermal degradation of cellulose, and chitosan (59.86% loss weight) and the third step was attributed to Ni/NiO nanocomposite synthesizing from natural plants (12.74% loss weight). It is important to note that the amount of the hydrogen-bonded water molecule in chitosan and the coordination biopolymers is of the order: nano Chitosan\u00a0<\u00a0Chit-Cell@Ni/NiO nano-biocomposite. Fig.\u00a03b and c were the DGA and DTA of nano chitosan and Chit-Cell@Ni/NiO nano-biocomposite confirming the two and three-step thermal behavior of nano chitosan and Chit-Cell@Ni/NiO nano-biocomposite thermal behaviors, respectively. It reveals that with increased amounts of Ni/NiO into the chitosan matrix lead to enhanced thermal stability. In the DTA curve of chitosan at least four thermal events, endothermic and exothermic, were observed up to 630\u00a0\u00b0C. Those thermal events were associated with the dehydration and thermal degradation processes, including depolymerization and decomposition stages of nano chitosan and Chit-Cell@Ni/NiO nano-biocomposite [32]. The exothermic effect on the differential thermal analysis (DTA) curves presented in the inset of Fig.\u00a03b, can be linked to weight loss due to the evaporation of physically adsorbed water. In the case of nano chitosan, that peak presented a maximam at 100\u00a0\u00b0C, while a second peak with a higher mass loss was found at 380\u00a0\u00b0C, caused by the decomposition of chitosan containing hydroxyl and amino groups [33]. Chit-Cell@Ni/NiO nano-biocomposite displayed a broad exothermic peak around 270\u00a0\u00b0C that can be attributed to the decomposition of nano chitosan and also cellulose [34] while the weight loss in the range 450\u2013500\u00a0\u00b0C should be assigned to the further decomposition of nano chitosan residues [33], whereas the band in the range of 600\u2013900\u00a0\u00b0C can be attributed to the reduction of Ni/NiO nanocomposite by reaction with residual carbon.The morphology of Chit-Cell@Ni/NiO nano-biocomposite was investigated by SEM (Fig.\u00a04\n). Fibril shape of attached cellulose to chitosan and randomly shaped shape of Ni/NiO were observed very well. The size of Chit-Cell@Ni/NiO nano-biocomposite was evaluated by TEM [35] (Fig.\u00a05\n). TEM image of Ni/NiO also shows these kind of shape confirming the presence of Ni/NiO in the Chit-Cell@Ni/NiO nano-biocomposite without any change in the morphology of Ni/NiO (Fig.\u00a0S1) The BET isotherms results showed that the Ni/NiO had a pore size of 11.3\u00a0nm with a BET area of 13.8\u00a0m2\u00a0g\u22121. While, Chit-Cell@Ni/NiO nano-biocomposite had a pore size of 16.087\u00a0nm with a BET area of 15.539\u00a0m2\u00a0g\u22121.The photo-absorption behavior of the Chit-Cell@Ni/NiO nano-biocomposite was evaluated by UV\u2013Vis DRS [36]. Fig.\u00a06\na displays the strong absorption edge at 200\u00a0nm\u2013600\u00a0nm. The bandgap (Eg) value of Chit-Cell@Ni/NiO was estimated using Kubelka\u2013Munk equation [36]:\n\n\n\n\u03b1\nh\n\u03c5\n=\nA\n\n\n(\n\nh\n\u03c5\n\u2212\n\nE\ng\n\n\n)\n\nn\n\n\n\n\nwhere, A is the absorption constant, \n\nh\n\u03c5\n\n is the photon energy, \n\n\nE\ng\n\n\n is the energy band gap and n is 1/2 or 2 for direct or indirect optical transition, respectively. As shown in Fig.\u00a06b, direct \n\n\nE\ng\n\n\n of Chit-Cell@Ni/NiO was found at about 3.30\u00a0eV from plotting \n\n\n\n(\n\u03b1\nh\n\u03c5\n)\n\n2\n\n\n vs. \n\nh\n\u03c5\n\n (Tauc plot) [37]. According to the obtained result, Chit-Cell@Ni/NiO could be an excellent photocatalyst for the efficient degradation of drug containments [21].The photocatalytic performance of Ni/NiO and Chit-Cell@Ni/NiO nanocatalysts were studied by degrading ciprofloxacin antibiotics. Before photocatalytic experiments, the adsorption potential of nanocatalysts were tested in darkness for 30\u00a0min. The adsorption efficiency of Ni/NiO and Chit-Cell@Ni/NiO for ciprofloxacin was 4.6% and 7.87%, respectively. The photodegradation of ciprofloxacin in the presence of Ni/NiO and Chit-Cell@Ni/NiO reached 59.05% and 91.88%, respectively, under direct sunlight exposure. Therefore, the effect of adsorption by these nanocatalysts could be ignored. On the other hand, the better photocatalytic performance of Chit-Cell@Ni/NiO than Ni/NiO could be attributed to the unique nanostructure of nanocomposite and synergistic effects of nickel oxides and metallic nickel with nano chitosan and cellulose. The degradation of ciprofloxacin follows pseudo-first-order kinetics (Eq. (1)) [38]. The rate constant for Chit-Cell@Ni/NiO (kapp\u00a0=\u00a00.125542 min\u22121) was 2.8 times Ni/NiO (kapp\u00a0=\u00a00.044641 min\u22121).\n\n(1)\n\n\n\u2212\nln\n\n(\n\nC\n\nC\n0\n\n\n)\n\n=\n\nk\n\na\np\np\n\n\nt\n\n\n\nwhere C is the concentration of ciprofloxacin at reaction time t, C\n0 is the initial concentration of ciprofloxacin, and k is the rate constant (min\u22121).The effects of the initial pH value, the concentration of antibiotic, and dosage of photocatalyst on the degradation of ciprofloxacin in Chit-Cell@Ni/NiO system were studied. At first, the effect of pH variation on the degradation efficiency of ciprofloxacin was examined (Figs 7\na and b). The degradation efficiency of ciprofloxacin was obtained 58.49%, 73.19%, 91.88%, 84.16% and 77.07% in pHs 4, 5, 6, 7 and 8, respectively. In this study, K\napp of ciprofloxacin increased to 0.04396\u00a0min\u22121, 0.06582\u00a0min\u22121 and 0.12554\u00a0min\u22121 as the solution pH increased from 4 to 6 and then decreased to 0.09213\u00a0min\u22121 and 0.07363\u00a0min\u22121 at pHs 7 and 8, respectively. In general, almost neutral pH (pH\u00a0=\u00a06) was more useful for the degradation of\u00a0ciprofloxacin than acidic and alkaline pHs. On the other hand, it was known that pKa value for ciprofloxacin is 5.9 and 8.89 [39]. It means that between these pHs, ciprofloxacin can be found in the form of zwitterion [40]. For further investigation, the isoelectric point of Chit-Cell@Ni/NiO was determined to be about pH 6.2 according to the zeta potential analysis (Inset of Fig.\u00a07a). Because the charge on the surface of the catalyst is positive and negative in highly acidic and alkaline conditions, respectively. The degradation of ciprofloxacin is minimal due to strong repulsion between Chit-Cell@Ni/NiO and ciprofloxacin molecules at acidic pHs [36]. On the other, the presence of more \u2013OH may inactivate the \u2022OH radicals, therefore decreased degradation efficiencies at pH higher than 6 can also be due to high concentration of \u2013OH [41]. It is worth noting that the working pH of Chit-Cell @Ni/NiO is close to neutral pH range, which completely covers the ambient pH of the wastewater [42]. Therefore, Chit-Cell@Ni/NiO can be introduced as an effective photocatalyst for the degradation of ciprofloxacin in wastewater.In the next step, the effect of the initial concentration of ciprofloxacin was examined and the obtained results were displayed in Fig.\u00a07c and d. As seen in Fig.\u00a07c, initially when the concentration of ciprofloxacin was increased the degradation efficiency of Chit-Cell@Ni/NiO was enhanced up to concentration of 10\u00a0mg/L, then decreased. The degradation percentages at the ciprofloxacin concentrations of 5, 10, 25, and 50\u00a0mg/L were 86.65%, 91.88%, 78.88%, and 67.10%, respectively, in a reaction time of 20\u00a0min. Furthermore, the Kapp for photodegradation of ciprofloxacin increases from 0.10068\u00a0min\u22121 at 5\u00a0mg/L to 0.12554\u00a0min\u22121 at 10\u00a0mg/L and then decreases to 0.05558\u00a0min\u22121 at 50\u00a0mg/L. The reason for the initial increase can be short lifetimes of radicals, because they can only react where they were generated. Increasing the quantity of ciprofloxacin molecules per volume unit increased the probability of collision between ciprofloxacin and active species that leads to increased degradation efficiency [43]. In case of decreasing efficiency with further increase of pollutant concentration, one reason may be that high initial concentrations of ciprofloxacin prevent light from reaching the surfaces of Chit-Cell@Ni/NiO photocatalyst and limit the formation of photogenic species responsible for the photocatalytic reaction. Also, more intermediates were produced at high ciprofloxacin concentrations. The generated intermediates compete with ciprofloxacin molecules for contact with active sites of Chit-Cell@Ni/NiO, resulting in a reduced percentage of degradation [44].Also, the Langmuir\u2013Hinshelwood kinetics model (Eq. (2)) can be used to describe this phenomenon [45]:\n\n(2)\n\n\nr\n=\n\u2212\n\n\nd\nC\n\n\nd\nt\n\n\n=\n\n\nk\nK\nC\n\n\n1\n\n+\n\nK\nC\n\n\n\n\n\n\n\n\n(3)\n\n\nln\n\n\nC\n0\n\nC\n\n+\n\nk\n\na\np\np\n\n\n\n(\n\n\nC\n0\n\n\u2212\nC\n\n)\n\n=\nk\nK\nt\n=\n\nk\n\na\np\np\n\n\nt\n\n\n\nwhere \n\nr\n\n is the reaction rate for degradation of ciprofloxacin, \n\n\nC\n0\n\n\n is initial concentration of ciprofloxacin, \n\nC\n\n is the final concentration of ciprofloxacin, \n\nk\n\n is the specific reaction rate constant, and \n\nK\n\n is the equilibrium constant of the reactant. The logarithmic form of the Langmuir-Hinshelwood equation was exhibited by Eq. (3), [46]. As shown in Fig.\u00a07d, the rate constant of 0.12554 min\u22121 was obtained for the degradation of ciprofloxacin using Chit-Cell@Ni/NiO (Catalyst dosage: 0.2\u00a0g/L, ciprofloxacin concentration: 10\u00a0mg/L and pH: 6). The amount of 0.988 for correlation coefficient \n\n(\n\nR\n2\n\n)\n\n confirmed a good fitting of the data by the Langmuir\u2013Hinshelwood kinetics model.The catalytic degradation of ciprofloxacin was evaluated by varying the amount of Chit-Cell@Ni/NiO at pH\u00a0=\u00a06 under direct sunlight. As exhibited in Figs. 7e and f, the degradation of ciprofloxacin increases with an increase in the amounts of catalyst from 0.1 to 0.2\u00a0g/L. In further studies, it was found that when the photocatalyst doses were more than 0.2\u00a0g/L, Chit-Cell@Ni/NiO is not able to disperse well in solution and the degradation of ciprofloxacin changes insignificant. A slight decrease in degradation efficiency is due to that high doses of Chit-Cell@Ni/NiO prevent sunlight from penetrating the solution because of the turbidity of the solution [47]. Also, Kapp of ciprofloxacin increased from 0.05837\u00a0min\u22121 to 0.12554\u00a0min\u22121 as the photocatalyst doses increased from 0.1 to 0.2\u00a0g/L and then decreased to 0.11453\u00a0min\u22121 and 0.10278\u00a0min\u22121 at photocatalyst doses 0.5 and 1\u00a0g/L, respectively. Based on this experiment, the optimal catalyst dose was 0.2\u00a0g/L.For better understanding, effect of Operational variables on the degradation of ciprofloxacin was showed in Table 1\n.The photogenerated h+, \u2022OH, and \n\n\nO\n2\n\n\u2022\n\u2212\n\n\n\n as reactive species were responsible for the degradation of different pollutants and their intermediates [48,49]. To determine the active species in photodegradation of ciprofloxacin, radical trapping studies were carried out using different scavengers [50]. The iso-Propanol (10\u00a0mM, IPA), benzoquinone (10\u00a0mM, BQ), and ammonium oxalate (10\u00a0mM, AO) were applied as scavengers, which indicate quenchers of \u2022OH and \n\n\nO\n2\n\n\u2022\n\u2212\n\n\n\n radicals, and h+, respectively (Fig.\u00a08\n). The maximum degradation of the ciprofloxacin was obtained without any scavenger species. This result emphasizes the photocatalytic degradation of ciprofloxacin in the Chit-Cell@Ni/NiO system. The photodegradation efficiency of ciprofloxacin decreases from 91.88% (without scavenger) to 58.34%, 79.12%, and 38.69 in the presence of IPA, BQ, and AO, respectively. These results show that although \u2022OH and \n\n\nO\n2\n\n\u2022\n\u2212\n\n\n\n were reactive species involved in the degradation process of ciprofloxacin, h+ plays a significant role in this reaction.In order to describe the possible mechanism of ciprofloxacin photodegradation under sunlight, the conduction bond (ECB) and valence bond (EVB) energy levels of Chit-Cell@Ni/NiO were measured using the following equations:\n\n(4)\n\n\n\nE\n\nC\nB\n\n\n=\n\u03c7\n\u2212\n\n\u00a0E\nC\n\n\u2212\n0.5\n\nE\ng\n\n\n\n\n\n\n\n(5)\n\n\n\nE\nVB\n\n\n\n=\nE\n\nCB\n\n+\n\n\u00a0E\ng\n\n\n\n\nwhere, is the energy of free electrons (\u223c4.5\u00a0eV) on the hydrogen scale and \u03c7 is absolute electronegativity of semiconductor and it was calculated by the following equation [51]:\n\n(6)\n\n\n\u03c7\n=\n\n\n[\n\n\n\nx\nA\n\na\n\n\n\n\nx\nB\n\nb\n\n\n\nx\nC\n\nc\n\n\n]\n\n\n1\n\n(\n\na\n+\nb\n+\nc\n\n)\n\n\n\n\n\n\n\n\n\n(7)\n\n\nx\n=\n\n\n\nE\n\nI\nE\n\n\n+\n\n\nE\n\nE\nA\n\n\n\n2\n\n\n\n\nwhere, \n\n\n\u03c7\nA\n\n\n, \n\n\n\u03c7\nB\n\n\n\n and \n\n\n\u03c7\nC\n\n\n are electronegativity of atoms and a, b and c are the number of atoms in the compound. \n\n\n\nE\n\nI\nE\n\n\n\n and \n\n\nE\n\nE\nA\n\n\n\n are the electro affinity and the first ionization energy of atoms, respectively. In this way, \u03c7 for Chit-Cell@Ni/NiO were found 5.26\u00a0eV. The \n\n\nE\ng\n\n\n value of Chit-Cell@Ni/NiO is obtained 3.30\u00a0eV from Tauc plot. Therefore, the ECB and EVB values of Chit-Cell@Ni/NiO were calculated to be \u22120.805\u00a0eV and +2.41\u00a0eV versus normal hydrogen electrode (NHE), respectively. Fig.\u00a09\n presents the photocatalytic mechanism using Chit-Cell@Ni/NiO. When the Chit-Cell@Ni/NiO is exposed to sunlight, an electronic transfer occurs that takes an electron from a lower valence level to a higher conductivity level, creating an electron\u2013hole pair. \n\n\ne\n\u2212\n\n\n\n and \n\n\nh\n+\n\n\n\n represent for an electron in the conduction band (CB) and the deficiency in the valence band (VB), respectively. \n\n\ne\n\u2212\n\n\n\n and \n\n\nh\n+\n\n\n transfer to the photocatalyst surface, where the redox reaction occurs. Since the VB energy level of Chit-Cell@Ni/NiO is more positive than the H2O/\u2022OH standard redox potential (E0\u00a0=\u00a0+2.34\u00a0eV vs. NHE), \u2022OH radicals were formed by the reaction of \n\n\nh\n+\n\n\n with H2O in VB. Also, due to the fact that the energy level of CB of Chit-Cell@Ni/NiO is more negative than the O2/ \n\n\nO\n2\n\n\u2022\n\u2212\n\n\n\n standard redox-potential (E0\u00a0=\u00a0\u22120.33\u00a0eV vs. NHE), \n\n\nO\n2\n\n\u2022\n\u2212\n\n\n\n\n radicals were formed by a combination of \n\n\ne\n\u2212\n\n\n\n with O2 in CB [52]. The generated radicals degrade ciprofloxacin into simple harmless molecules. In addition, immobilizing of Ni/NiO on the chitosan-cellulose surface due to increasing \n\n\ne\n\u2212\n\n\n and \n\n\nh\n+\n\n\n separation and the wide surface area of the photocatalyst can lead to enhancing ciprofloxacin degradation efficiency. As the surface area increases, there will be more active sites for adsorption of pollutant molecules, which can increase the likelihood of interaction between reactive species and ciprofloxacin molecules. Also, the bandgap of the Chit-Cell@Ni/NiO has been reduced compared to Ni/NiO (3.8\u00a0eV) [53], which provides the conditions for the photocatalytic process in the presence of natural sunlight.Based on LC-MS results as shown in Supporting Fig.\u00a0S3, the degradation ways of ciprofloxacin were proposed in the presence of Chit-Cell@Ni/NiO. As shown in Scheme 1\n, the photodegradation of ciprofloxacin generally underwent two pathways, including the degradation of the piperazine ring and the cleavage of the quinolone ring [54]. In pathway 1, at first, the piperazine ring of ciprofloxacin is oxidized to produce intermediate I (m/z 362) through the piperazine breaking, followed by the losing the carbonyl group to form intermediate II (m/z 306). The intermediate II could be converted into intermediate III (m/z 262) via the losing of ethylamine. In the following, the cyclopropyl group is removed from III to achieve intermediate IV (m/z 221) by cleavage. Then, the ring-opening of the quinolone occurs and then diketone VI (m/z 181) is produced through decarboxylation and closing ring. In pathway 2, the \u2022OH radicals can attack ciprofloxacin to generate intermediate VII (m/z 288) by decarboxylation. On the other hand, the piperazine and quinolone rings were destroyed and the intermediate V (m/z 270) is generated. Eventually, the intermediates could be degraded to safe lower mass intermediates.From a practical view, one of the most essential properties of a catalyst is its recyclability and durability in long-term use. The stability studies of Chit-Cell@Ni/NiO were done with fresh ciprofloxacin solution under sunlight in three consecutive cycles. When each cycle was finished, Chit-Cell@Ni/NiO was separated by filter paper, washed with water several times, dried, and reused in the next run. The degradation efficiency of ciprofloxacin for these three repeated applications was exhibited in Fig.\u00a0S4. The obtained results show that the catalytic performance of Chit-Cell@Ni/NiO was only slightly decreased after three consecutive cycles of photodegradation. TEM and FT-IR of the recycled nanocomposite were considered and structure, size, and morphology were not changed after three times of reusability (Fig.\u00a010\n).To show the merit of Chit-Cell@Ni/NiO photocatalytic performance in the comparison with other photocatalysts toward the degradation of ciprofloxacin, some of the previous reports were compared in Table 2\n. The current protocol was compared with the data in the literature based on the ciprofloxacin concentration, catalyst dosage, degradation time and efficiency. As shown Table 2, Chit-Cell@Ni/NiO exhibited relatively higher activity for photodegradation of ciprofloxacin in a shorter time without any auxiliary agent.Biosynthesis fabrication of Ni/NiO and aldehyde cellulose was performed using barely wastes. Nano chitosan was extracted after deproteination, demineralization and deacetylation of shrimp waste. By incorporating the nano chitosan and Ni/NiO nanocomposite to aldehyde cellulose, Chit-Cell@Ni/NiO nano-biocomposite from organisms was fabricated. The catalytic performance of the Chit-Cell@Ni/NiO was evaluated for the photodegradation of ciprofloxacin under direct sunlight. The\u00a0maximum degradation efficiency of 10\u00a0mg/L ciprofloxacin was\u00a0obtained at neutral pH (pH\u00a0=\u00a06) and Chit-Cell@Ni/NiO dosage of 0.2\u00a0g/L. About 92% of ciprofloxacin was degraded within 20\u00a0min, while the degradation efficiency for the Ni/NiO was\u00a059.05% only. Radicals quenching experiments were illustrated that h+ was played a dominant role in degradation process of ciprofloxacin. We believe which this study offers a\u00a0novel perspective for practical photocatalytic degradation to\u00a0effective reduce the number of antibiotics in the water system.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors gratefully acknowledge the Birjand University of Technology and the University of Jiroft for the support of this work.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.jmrt.2022.04.046.", "descript": "\n                  A novel nanocomposite including cellulose, chitosan and Ni/NiO was fabricated from barely wastes, shrimp wastes, and Calotropis procera, respectively. It was characterized by TEM, SEM, TGA, DTA, FT-IR, BET, EDAX, and elemental analysis. 5\u201325\u00a0nm of Ni/NiO were dispersed on chitosan and cellulose. The BET isotherms results showed that the Ni/NiO had a pore size of 11.3\u00a0nm with a BET area of 13.8\u00a0m2\u00a0g\u22121. While, Chit-Cell@Ni/NiO nano-biocomposite had a pore size of 16.087\u00a0nm with a BET area of 15.539\u00a0m2\u00a0g\u22121. Then, Chit-Cell@Ni/NiO bio-nanocomposite was applied to the photodegradation of ciprofloxacin under sunlight. About 92% of ciprofloxacin could be efficiently degraded within 20\u00a0min. Radical quenching experiments confirmed the contribution of active species was in descending order of h+> \u2022OH\u00a0>\u00a0\u2022O2\n                     \u2212 in the Chit-Cell@Ni/NiO system. The possible ciprofloxacin degradation pathway has been proposed according to the intermediates detected by LC-MS. Also, Chit-Cell@Ni/NiO showed high durability and stability after three-cycle ciprofloxacin degradation. In short, this study offers an efficient green methodology to decrease the number of antibiotics in the water system.\n               "}
{"full_text": "Core-shell particles is a class of particles containing a core and a shell, which can be either the same or different materials that can be separately identified (Hayes et\u00a0al., 2014). Their properties can be adjusted by modification of a material with a variety of another materials through surface coating techniques, such as chemical grafting and microemulsion, in order to form layer-by-layer core-shell structures. Chemical grafting is a surface modification method mostly used to adjust the surface properties of the resulting material. This technique involves the use of an initiator that can undergo a copolymerization reaction on the surface of a substrate. Different types of initiators have been employed such as potassium permanganate, ammonium peroxydisulfate, benzoyl peroxide, etc. in order to achieve the desired properties of the materials (Abidi, 2009). Moreover, micro-emulsion is one of the simplest and most effective methods for the preparation of nano-sized particles that can be immobilized on the support materials. A formation of microemulsion basically involves a water, hydrocarbon, and surfactant, which can be divided into two types; that are, the water-in-oil type of micro-emulsion and the oil-in-water type of micro-emulsion (Hanaoka et\u00a0al., 2015).Owing to their adaptable properties, core-shell structured materials have been employed in various potential applications; for instance, the catalytic pyrolysis of lignin to valuable monoaromatics (Xue et\u00a0al., 2020), the Fischer-Tropsch reaction (Chen et\u00a0al., 2020) and the catalytic reduction of NOx (Liu et\u00a0al., 2018). Certainly, like other catalyst types, the core-shell catalysts were found to have catalytic activity varied, depending on several characteristics such as shell/core thickness, exposed active sites, the ratio of components, and most importantly, the order and/or location of active components in the structure. For examples, the effect of the shell thickness of a Co@C@SiO2 core-shell catalyst was studied on the Fischer-Tropsch synthesis (Chen et\u00a0al., 2020). The results indicated that the CO conversion decreased with the increasing SiO2 shell thickness because a larger thickness obstructed CO from entering to the Co active sites. Therefore, the sequence of the materials on a core-shell structure exposed to the reaction mixture is very crucial for the catalytic activity. Morever, the catalytic reduction of NOx was also examined over a Fe/Beta@SBA-15 core-shell catalyst (Liu et\u00a0al., 2018). The SBA-15 shell thickness, modified by varing the ratio of Si/Beta, was found to affect the acidic and redox properties of the catalyst. The optimum thickness (10\u00a0\u200bnm) improved the performance of the catalyst whereas a too thick shell (30\u00a0\u200bnm) decreased the activity. Additionally, the preparation of a core-shell Raney Fe@HZSM-5 catalyst was accomplished using one-pot hydrothermal synthesis method, for gasoline production via Fischer\u2013Tropsch reaction (Sun et\u00a0al., 2010). They found that the core-shell catalyst provided a higher CO conversion and C5\u2013C11 selectivity than the physical mixture of Fe and HZSM-5 catalyst. So, it can be implied that the synergy between the two components of the core-shell catalyst helped promote the performance of the catalyst. In addition, the core-shell form of a catalyst was also constructed in order to prevent the sintering of active metals for high temperature reaction. For example, the core-shell structure of a Ni-yolk@Ni@SiO2 catalyst was developed for the CO2 reforming of methane (Li et\u00a0al. (2016). The TEM analysis suggested that the Ni nanoparticles was trapped inside the silica shell that prevented Ni from sintering; as a result, the catalyst with the core-shell structure gave a higher methane turnover frequency than the SiO2-supported Ni catalyst.1,3-Butadiene is one of the most important chemicals widely used as a building block in the production of polymers such as polybutadiene, styrene-butadiene rubber, and polycholoprene (Fedotov et\u00a0al., 2019). It can be commercially produced by three main processes, including steam cracking of paraffinic hydrocarbons, catalytic dehydrogenation of n-butane and n-butene, and oxidative dehydrogenation of n-butene (White, 2007). However, due to the depletion of the petroleum sources and environmetal issues, the use of a renewable resource becomes more attractive, especially bio-ethanol that is considered as a renewable feedstock for the production of several industrial products such as light hydrocarbons (Chinniyomphanich et\u00a0al., 2016) or heavy hydrocarbons (Choopun and Jitkarnka, 2016) and oxygenate compounds (Sujeerakulkai and Jitkarnka, 2016). Not only is it used as the main feedstock, but bio-ethanol is also used as a hydrogen donor for other feedstocks such as glycerol in order to produce oxygenates, like 1,2-propanediol, acetaldehyde (Kumpradit and Jitkarnka, 2019), co-produced with ethyl lactate (Kuljiraseth and Jitkarnka, 2019). Among ethanol-derived oxigenates, 1,3-butadiene is one of the most important chemicals expected to be produced from the catalytic conversion of bio-ethanol through four principal steps; that are, (1) ethanol dehydrogenation to acetaldehyde, (2) the condensation of acetaldehyde to crotonaldehyde via the dehydration of 3-hydroxyl butanal, (3) the reduction of crotonaldehyde to crotyl alcohol via Meerwein-Ponndorf-Verley-Oppenauer (MPVO) reduction using an alcohol as a hydrogen donor, and (4) butadiene formation via the dehydration of crotyl alcohol (Zhao et\u00a0al., 2020). Therefore, a multi-functional catalyst with an optimum acid and base ratio is needed in order to achieve a high 1,3-butadiene selectivity. Especially, since these are sets of consecutive reactions that need different types of catalysts with multi-functions to drive, a catalyst with a core-shell structure that contains different materials layer by layer were believed to be beneficial to such reactions because they can be driven consecutively on consecutive layers, and selectively geared layer by layer, using various designed materials, until the desired product is formed (Namchot and Jitkarnka, 2016).Based on several research articles, Cu catalysts provided a good efficiency in ethanol dehydrogenation. The effect of Cu content on the ethanol dehydrogenation over Cu/ZrO2 catalyst was examined (Freitas et\u00a0al., 2014). The % content of Cu was varried from 5% to 20%, and the reaction was carried out in a continuous-flow tubular fixed-bed reactor at the temperature range of 473\u2013548\u00a0\u200bK and atmospheric pressure. The results indicated that the highest acetaldehyde formation rate was obtained over the 5Cu/ZrO2 catalyst. According to the results from XPS analysis, 5Cu/ZrO2 provided highly-dispersed Cu+1 species selective for the formation of acetaldehyde, the product from dehydrogenation of ethanol. Moreover, the different morphologies of copper-based catalysts, including urchin-like (CuO-UC), fiber-like (CuO-FB) and nanorod (CuO-NR) catalysts, prepared by a microwave-assisted method, were studied on the dehydrogenation of ethanol (Sato et\u00a0al., 2012). The results indicated that the catalyst with the highest amount of Cu+1 species from the urchin-like morphology exhibited the highest activity of ethanol dehydrogenation to acetaldehyde among all catalysts. Besides the metal content and morphology, thermal treatment condition also influenced the copper species (Cu0/Cu1+) and then the activity on the bio-ethanol dehydrogenation to acetaldehyde over CuMgAl mixed oxide (Campisano et\u00a0al., 2018), and it also had the impacts on the acid-base properties and catalytic activity of ZrO2/Nano-SiO2 catalyst for 1,3-butadiene production from the mixture of ethanol and acetaldehyde (Gao et\u00a0al., 2018).For the acetaldehyde condensation reaction, the roles of acid and base sites on the acetaldehyde condensation were investigated over MgO/SiO2 and ZrO2/SiO2 catalysts (Ordomsky et\u00a0al., 2010). It was found that the selectivity of crotonaldehyde was high on both magnesium oxide and zirconium oxide supported on silica catalyst. In order to investigate the role of active sites, CO2 and pyridine were doped as the acid or base molecular probe in order to poison the corresponding active sites. The conversion of acetaldehyde was found to decrease when pyridine was co-fed with acetaldehyde, suggesting that the lewis acid site played an important role on the condensation of acetaldehyde to crotonaldehyde. In addition, ZrO2 was also studied as a catalyst for MPVO reduction. For example, the effects of various oxides, including Nb2O5, TiO2, and ZrO2, on the MPVO reduction of cyclohexanone to cyclohexanol were investigated with using 2-propanol as a hydrogen donor (Komanoya et\u00a0al., 2015). They found that although ZrO2 had a lower lewis acid density than Nb2O5 and TiO2, ZrO2 was the most effective catalyst for the MPVO reduction of cyclohexanone due to the highest density of base sites provided by the hydroxyl groups on the surface of ZrO2. Such a base site type was very important for the evolution of six-membered ring intermediates on the lewis acid site of the catalyst.Layered double hydroxides (LDHs) belong to a group of anionic clay materials. Their structure consists of positively-charged brucite-like layers of mixed hydroxides and the exchangeable charge compensating anions in the interlayer space. The general formula of an LDH is [M2+\n1-xM3+\nx(OH)2][(An\u2212)\nx/n\u00b7mH2O], where M2+ and M3+ are divalent and trivalent cations, respectively. An\u2212 is mainly an inorganic or organic interlayer anion (P\u0161eni\u010dka et\u00a0al., 2020). It has received attention as a catalyst for the conversion of ethanol due to its high surface area, acid-base properties, and thermal stability (Mishra et\u00a0al., 2018). For instance, the effects of preparation methods on the catalytic condensation of ethanol were determined using Mg\u2013Al mixed oxide catalysts (Leon et\u00a0al., 2011). The reaction was performed at the temperature range of 473\u2013723\u00a0\u200b\u00b0C. They found that the preparation methods strongly affected the acid-base properties of the catalysts. Moreover, the results also indicated that ethylene and acetaldehyde were observed as two primary products from ethanol conversion over Mg\u2013Al catalysts via dehydration and dehydrogenation reactions, respectively. In addition, the effect of adding a tetravalent metal, Tin, into the hydrotalcite structure was evaluated on the MPVO reduction of aldehyde and ketone with 2-propanol as a hydrogen donor (Jim\u00e9nez-Sanchidri\u00e1n and Ruiz, 2014). It was found that the catalyst containing Mg/Sn/Al mixed oxides showed a higher activity and selectivity than Mg/Al and MgO catalysts. The better performance of tin-containing catalysts can be ascribed to the Lewis acid site of Sn+ ion where 2-propanol can be adsorbed more efficiently.Previously, our research group studied some sets of core-shell catalysts composed of a metal oxide shell and a MgAl-LDO-based core promoted with a dot-coated metal oxide promoter (Sricheun, 2018). A variety of core-shell catalysts prepared with various metal oxide shells (MgO, CuO and ZnO), and different MgAl-LDO cores, synthesized by partial substitution of various transition metals (Cu, Zn, Mn and Fe) in the core\u2019s structure promoted with two dot-coated metal oxide promoters (HfO2 and ZrO2), were prepared and subsequently studied for ethanol conversion to 1,3-butadiene. It was found that the catalysts that consisted of FeMgAl-LDO core, promoted by dot-coated ZrO2, and then coated with CuO shell, denoted as \u201cCuO@ZrO2/FeMgAl-LDO\u201d shell@core catalyst, gave the highest 1,3-butadiene yield due to the synergistic effect of all components. Based on the best combination of this shell@core catalyst, further investigation shall be done on preparation techniques that might impact the 1,3-butadiene yield as well. Recently, our preliminary detailed work on grafted ZrO2/FeMgAl-LDO catalyst has confirmed that the presence of grafted ZrO2 on the catalyst core helped promote the condensation reaction of acetaldehyde to crotonaldehyde and Meerwein-Ponndorf-Verley (MPV) reduction of crotonaldehyde to crotyl alcohol, which are the important steps for 1,3-butadiene production (Suwansawat and Jitkarnka, 2020a). As earlier mentioned, since the preparation method can influence the exposure of active components in the core-shell structure and then the catalytic activity, the next step of our catalyst development was therefore to investigate the effect of the shapes (powder and granule) of FeMgAl-LDO core and the method of ZrO2 deposition on the core in order to determine the appropriate exposure and synergy of all active components, which would lead to the highest yield of 1,3-butadiene. Based on this concept, the preparation of CuO shell was fixed for all prepared samples of the core-shell catalyst in order to investigate the catalyst behavior through the alterations of the core and its promoter only. Therefore, in this work, the effects of preparation methods were investigated in two steps; that are, \nStep 1\n: granular formation of the FeMgAl-LDO core with \u03b1-Al2O3 binder by pelletization, and then \nStep 2\n: ZrO2 deposition on the core using either chemical grafting or micro-emulsion. Based on the two methods of two different steps of preparation, the catalysts, prepared and used in this work, were CuO@m-ZrO2/p-FeMgAl-LDO, CuO@gf-ZrO2/p-FeMgAl-LDO, CuO@m-ZrO2/g-FeMgAl-LDO, and CuO@gf-ZrO2/g-FeMgAl-LDO, where p is powder, g is granular, gf is grafting, and m is micro-emulsion.The FeMgAl-LDO core was prepared using the co-precipitation method. A solution of Fe(NO3)3\u22c59H2O, Mg(NO3)2\u22c56H2O and Al(NO3)3\u22c59H2O was mixed into the three-neck flask, containing a 700\u00a0\u200bml of 0.25\u00a0\u200bM sodium carbonate solution with constant stirring. The pH was controlled at 10 using a 5\u00a0\u200bM NaOH solution. After aging for 16\u00a0\u200bh, the precipitate was separated by filtration, and washed with deionized water until pH 7 was reached. The precipitate FeMgAl-LDH was next dried in an oven at 65\u00a0\u200b\u00b0C for 12\u00a0\u200bh, and then calcined at 500\u00a0\u200b\u00b0C for 5\u00a0\u200bh in order to obtain the p-FeMgAl-LDO catalyst, where p stands for \u201cpowder\u201d.The calcined p-FeMgAl-LDO catalyst was mixed with 20\u00a0\u200bwt% Al2O3 binder, and then formed into a granular shape using a hydraulic press machine. The palletized catalyst was then ground, and subsequently sieved to obtain 0.8\u00a0\u200bmm diameter granules using a mesh No.20 stainless steel wire sieve. Next, the sieved catalyst was calcined at 500\u00a0\u200b\u00b0C for 5\u00a0\u200bh to obtain g-FeMgAl-LDO, where g stands for \u201cgranular\u201d.The p-FeMgAl-LDO catalyst obtained from the first step (2.1.1) was promoted with a ZrO2 promoter via the micro-emulsion method. 1.246\u00a0\u200bg of Zr(NO3)2\u22c5xH2O precursor was added into a two-neck flask containing 130\u00a0\u200bml of deionized water, 10.4\u00a0\u200bml of octane, and 0.146\u00a0\u200bg of L-arginine. The mixture was stirred for 4\u00a0\u200bh, followed by the addition of sieved p-FeMgAl-LDO catalyst. After 20\u00a0\u200bh of aging, the solid product was filtered and washed with ethanol and deionized water, respectively. Next, the solid was dried in an oven at 65\u00a0\u200b\u00b0C overnight, followed by calcination at 500\u00a0\u200b\u00b0C for 5\u00a0\u200bh to obtain m-ZrO2/p-FeMgAl-LDO catalyst, where m stands for micro-emulsion.The g-FeMgAl-LDO was also promoted with a ZrO2 promoter via the micro-emulsion method as stated in 2.1.3.The ZrO2 promoter was also deposited on the p-FeMgAl-LDO via the chemical grafting method. First, 3\u00a0\u200bg of p-FeMgAl-LDO was dispersed in 100\u00a0\u200bml of anhydrous toluene at constant stirring, followed by the slow addition of 2.4\u00a0\u200bg of zirconium sec-butoxide. The resulting slurry was stirred at room temperature for 6\u00a0\u200bh. After that unreacted zirconia in the solution was removed by centrifugation, and the resulting solid was then maintained in deionized water for 6\u00a0\u200bh. Next, the solid product was isolated by filtration, followed by drying at 65\u00a0\u200b\u00b0C overnight and calcination at 500\u00a0\u200b\u00b0C for 5\u00a0\u200bh to yield the gf-ZrO2/p-FeMgAl-LDO catalyst, where gf stands for grafting.The g-FeMgAl-LDO was also promoted with ZrO2 using the chemical grafting method as described in 2.1.5.CuO that was encapsulated on ZrO2/FeMgAl-LDO samples was prepared using the sol-gel method. First, 2\u00a0\u200bg of a ZrO2/FeMgAl-LDO was dispersed in a mixed solution of 800\u00a0\u200bml of deionized water, 600\u00a0\u200bml of ethanol and 3\u00a0\u200bg of hexadecyltrimethyl ammonium bromide at constant stirring for 2\u00a0\u200bh. The pH was controlled at 11 using NH4OH. After that, 0.84\u00a0\u200bg of Cu(NO3)2\u22c53H2O precursor was added into the mixed solution at constant stirring, and then aged for 24\u00a0\u200bh. The resulting product was then filtrated and washed with deionized water until pH 7 was reached, followed by washing with 800\u00a0\u200bml of ethanol. Next, the precipitate was dried in an over at 65\u00a0\u200b\u00b0C overnight, and calcined at 500\u00a0\u200b\u00b0C for 5\u00a0\u200bh to obtain CuO@ZrO2/FeMgAl-LDO catalysts.The crystalline structure of the catalysts was investigated using Rikagu SmartLab X-Ray Diffractometer (XRD) equipped with CuK\u03b1 radiation (1.5405). The diffraction patterns were collected in the 2\u03b8 range of 5\u00b0\u201370\u00b0 using the scan speed of 0.02\u00a0\u200b\u00d7\u00a0\u200b(2\u03b8)/0.6\u00a0\u200bs. X-Ray fluorescence spectrometer (XRF) was used to identify the elemental composition of the catalysts using Best Detection-Vacuum method. The temperature programmed desorption of ammonia (NH3-TPD) and carbon dioxide (CO2-TPD) was employed to study the acid-base properties of the catalyst using a Temperature Program Desorption/Reduction/Oxidation analyzer (TPDRO), BELCAT II. To determine the specific surface area, total pore volume, and pore size of catalysts, the Brunauer-Emmett-Teller (BET) technique was employed using Surface Area Analyzer (Quantachrome, Autosorb-1MP) with Multipoint nitrogen adsorption and desorption isotherm plots. A Field Emission Scanning Electron Microscope (FE-SEM) was used to investigate the morphology of the catalyst at high magnification. The catalysts were placed on carbon tape and then coated with platinum by sputtering. The catalysts were then analyzed on Hitachi/S-4800 (accelerating voltage 10.0\u00a0\u200bkV) Transmission electron microscopy (TEM) to investigate the morphology and elemental composition of the catalysts. The TEM measurements were done on Thermo Scientific TALOS F200X equipped with EDS system.The activity testing was carried out in a continuous U-tube fixed bed reactor under atmospheric pressure. 1 g of a catalyst was loaded into the reactor using quartz wool as a bed supporter. After that the ethanol was fed into the reactor at the rate of 3.2\u00a0\u200bml/h. The liquid product was collected in the cooling condensing flask while the gas product was passed from the condensing flask to an online gas chromatograph (GC) equipped with a flame ionization detector with a Plot alumina column. The liquid product from the condensing flask was extracted by CS2, and then the obtained non-aqueous products were analyzed for their composition using a LECO Pegasas 1D-mode Gas Chromatograph equipped with a mass spectrometer of Time-of-Flight Type (GC-TOFMS) using a capillary column, Rxi-PAH (60\u00a0\u200bm\u00a0\u200b\u00d7\u00a0\u200b0.25mmID and 0.10 \u03bcm film thicknesses). The product selectivity and yields were calculated using Equations (1)-(2)\n\n\n(1)\n\n\nS\ne\nl\ne\nc\nt\ni\nv\ni\nt\n\ny\ni\n\n\n\n(\n%\n)\n\n=\n\n\n\nC\no\nn\nc\ne\nn\nt\nr\na\nt\ni\no\n\nn\ni\n\n\n\n\n\n\u2211\n\ni\nn\n\nC\no\nn\nc\ne\nn\nt\nr\na\nt\ni\no\n\nn\ni\n\n\n\n\n\u00d7\n100\n\n\n\n\n\n\n(2)\n\n\ny\ni\ne\nl\nd\n=\n\n\n\ns\ne\nl\ne\nc\nt\ni\nv\ni\nt\ny\n\n\n(\n%\n)\n\n\u00d7\nc\no\nn\nv\ne\nr\ns\ni\no\nn\n\n\n(\n%\n)\n\n\n100\n\n\n\n\n\nThe XRD patterns of CuO@m-ZrO2/p-FeMgAl-LDO, CuO@gf-ZrO2/p-FeMgAl-LDO, CuO@m-ZrO2/g-FeMgAl-LDO, and CuO@gf-ZrO2/g-FeMgAl-LDO catalysts are shown in Fig.\u00a01\n. All the XRD patterns of catalysts show the diffraction peaks at 2\u03b8\u00a0\u200b=\u00a0\u200b25.44\u00b0, 35.02\u00b0, 37.64\u00b0, 43.22\u00b0, 52.42\u00b0, 57.36\u00b0, 61.18\u00b0, 66.38\u00b0, and 68.08\u00b0, which correspond to the Al2O3 phase (Li et\u00a0al., 2017). Moreover, they also show the diffraction peaks of MgO phases at 2\u03b8\u00a0\u200b=\u00a0\u200b43.2\u00b0, and 62.7\u00b0 (Mishra et\u00a0al., 2018). Furthermore, the characteristic diffraction peaks of MgAl2O4 spinel at 2\u03b8\u00a0\u200b=\u00a0\u200b44.86\u00b0, and 65.30\u00b0 (Kuljiraseth et\u00a0al., 2019) are also observed over in all catalysts. It can be concluded that all the layered double oxide-based catalysts were successfully synthesized. No diffraction peak of ZrO2 is observed over all the catalysts, indicating that ZrO2 is very well dispersed or formed as an amorphous phase. For the CuO@m-ZrO2/p-FeMgAl-LDO catalyst, the small diffraction peak of CuO is located at 2\u03b8\u00a0\u200b=\u00a0\u200b38.9\u00b0 (Zhu et\u00a0al., 2018) while the XRD of CuO@gf-ZrO2/p-FeMgAl-LDO catalyst shows the sharp reflection peak of CuO at 35.7\u00b0 and 38.9\u00b0. Additionally, no diffraction peaks of CuO can be detected over CuO@m-ZrO2/g-FeMgAl-LDO and CuO@gf-ZrO2/g-FeMgAl-LDO granular catalysts, possibly because the high peak intensity of crystalline Al2O3 obstructs the amorphous peaks of amorphous CuO from appearance. CuO was also possibly formed as an amorphous phase or very well-dispersed.TEM-EDX analysis was employed to study the morphology of the powder catalysts. The HR-TEM images (Fig.\u00a02\na) suggest that the powder p-FeMgAl-LDO core is composed of a group of the agglomerated stacks of the FeMgAl layered double oxide as can be depicted in Fig.\u00a02 ci. Moreover, the elemental mapping also reveals that the Fe (yellow), Mg (red), and Al (green) particles are homogeneously distributed (Fig.\u00a02b).After the deposition of ZrO2 using two different methods, including micro-emulsion and chemical grafting, the elemental mapping shows that the Zr particles (purple) prepared from both micro-emulsion (Fig.\u00a03\nA b) and chemical grafting (Fig.\u00a03B b) methods are homogeneously dispersed on the surface of the p-FeMgAl-LDO core. In addition, the elemental mapping also shows that the promoted sample prepared by the chemical grafting method provides a higher density of Zr particles than the one prepared by the micro-emulsion method as shown in Fig.\u00a03A d-v and Fig.\u00a03B d-v.After the encapsulation with CuO using the sol-gel method, the HR-TEM images of CuO@m-ZrO2/p-FeMgAl-LDO catalyst show the agglomeration of several core-shell particles. The agglomerated core is encapsulated by CuO as shown in Fig.\u00a04\nA a. Moreover, the elemental mapping also suggests that Cu particles are mainly located at the outer part of the promoted core (Fig.\u00a04A b). For the CuO@gf-ZrO2/p-FeMgAl-LDO catalyst, the result from elemental mapping (Fig.\u00a04B b) reveals that the large particle of cores are surrounded by CuO particles in a crystalline form as observed from the XRD patterns. So, it can be concluded that the core-shell catalysts were successfully synthesized.The SEM-EDX technique was adopted to study the morphology of the granular catalysts. Fig.\u00a05\n illustrates the results obtained from the SEM analysis of the g-FeMgAl-LDO granular core. The elemental mapping (Fig.\u00a05a) shows that the FeMgAl-LDO core is composed of Mg (purple), Al (blue), O (White), and Fe (red) particles that are homogeneously distrubuted. After the deposition of the ZrO2 promoter using micro-emulsion and chemical grafting methods, the elemental mapping images reveal that ZrO2 particles (yellow) are homogeneously dispersed on the surface of the g-FeMgAl-LDO core as shown in Figs 6A b and Fig.\u00a06\nB b. Furthermore, after the encapsulation of CuO, the images from elemental mapping (Fig.\u00a07\nA b and Fig.\u00a07B b) suggest that the Cu particles (green) are mainly populated at the outer part of the CuO@m-ZrO2/g-FeMgAl-LDO and CuO@gf-ZrO2/g-FeMgAl-LDO catalysts.The physical properties of all catalysts are summarized in Table\u00a01\n. For both core-shell catalysts with the powder core, the surface area, pore volume and pore diameter of the CuO@m-ZrO2/p-FeMgAl-LDO catalyst are 117.3\u00a0\u200bm2/g, 0.362\u00a0\u200bcm3/g, and 60.69\u00a0\u200b\u00c5, respectively while the surface area, pore volume and pore diameter of the CuO@gf-ZrO2/p-FeMgAl-LDO are 141.7\u00a0\u200bm2/g, 0.148\u00a0\u200bcm3/g, 35.94\u00a0\u200b\u00c5. In case of the granular shape catalysts, the surface area is significantly decreased with the addition of Al2O3 binder (Suwansawat and Jitkarnka, 2020b). The surface area, pore volume and pore diameter of the CuO@m-ZrO2/g-FeMgAl-LDO catalyst are 107.1\u00a0\u200bm2/g, 0.199\u00a0\u200bcm3/g, and 36.16\u00a0\u200b\u00c5 while the surface area, pore volume and pore diameter of the CuO@gf-ZrO2/g-FeMgAl-LDO catalyst are 106.8\u00a0\u200bm2/g, 0.198\u00a0\u200bcm3/g, and 36.16\u00a0\u200b\u00c5. The elemental composition of the catalysts is also shown in Table\u00a01.The acid and base properties of all catalysts are stated in Table\u00a01. Moreover, the results also indicate that the total acidity and total basicity of the catalysts are suppressed with the presence of Al2O3 binder.The ethanol conversion was performed over CuO@m-ZrO2/p-FeMgAl-LDO and CuO@gf-ZrO2/p-FeMgAl-LDO catalysts in order to study the effect of the ZrO2 deposition method on 1,3-butadiene production. The results suggest that ethanol can almost entirely be converted over both catalysts as shown in Table\u00a02\n. Moreover, Fig.\u00a08\nA indicates that ethylene is formed as a primary product over both catalysts with the yield of 25.1% and 25.8%. Additionally, the yield of 1,3-butadiene is also not significantly different. In addition, the relative yield of dehydration and dehydrogenation products (Fig.\u00a08B) also exhibits that the core-shell catalyst with ZrO2 prepared by micro-mulsion method provides 42.7% yield of dehydrogenation products and 57.3% yield of dehydration products while the one with grafted ZrO2 gives 36.7% yield of dehydorgenation products and 63.5% of dehydration products. According to the results from XRF and TEM analyses, the catalyst with ZrO2 prepared by the chemical grafting method gives a higher Zr content. Based on the results from the previous study, the presence of ZrO2 promoted the dehydrogenation pathway. So, the higher Zr content of CuO@gf-ZrO2/p-FeMgAl-LDO resulted in the higher yield of dehydration pathway. Hence, it can be concluded that, in the case of powder core samples, the ZrO2 deposition method did not significantly affect the formation of 1,3-butadiene. Nevertheless, the catalyst with ZrO2 deposited by the micro-emulsion method is better because it gives higher yield of dehydrogenation products.In the case of the granular core catalyst samples, the ethanol conversion was performed in order to also investigate the effect of the ZrO2 deposition method on the catalytic activity of CuO@m-ZrO2/g-FeMgAl-LDO and CuO@gf-ZrO2/g-FeMgAl-LDO. The results indicate that the conversion of ethanol obtained from both catalysts is the same at 99.9% (Table\u00a03\n) In addition, the relative yield of the products in Fig.\u00a09\nA also suggest that ethylene is produced as a main product over CuO@m-ZrO2/g-FeMgAl-LDO and CuO@gf-ZrO2/g-FeMgAl-LDO catalysts with the yield of 21.6% and 23.8%, respectively. Moreover, the relative yield indicates that the catalyst with ZrO2 prepared by the micro-emulsion method gives 31.6% of dehydrogenation products and 68.4% of dehydration products whereas the chemical grafting one provides 42.3% of dehydrogenation products and 57.7% dehydration products. The higher yield of dehydrogenation products of the chemical grafted ZrO2 catalyst sample results in higher yield of 1,3-butadiene (10.2%). Based on the results from XRF and TEM analyses, the catalyst with ZrO2 prepared by the chemical grafting method provides a higher content of Cu, which helps convert more ethanol to acetaldehyde, resulting in higher yield of 1,3-butadiene.The ethanol conversion was performed over the core-shell catalysts, where the ZrO2 promoter was deposited on the cores using the different methods in order to investigate the impact of the deposition method on the conversion of ethanol to 1,3-butadiene. In the case of the powder catalysts, the method of ZrO2 deposition did not significantly alter the formation of 1,3-butadiene. However, the catalyst with micro-emulsioned ZrO2 exhibited a better activity on ethanol dehydrogenation. In case of the granular catalysts, the highest 1,3-butadiene yield was found over the catalyst that had ZrO2 prepared by the chemical grafting method. Therefore, it can be concluded that the conversion of ethanol to 1,3-butadiene is a set of surface sensitive reactions. Therefore, the specific catalyst preparation method was required to achieve a high 1,3-butadiene yield. Among all catalysts, the granular one with grafted ZrO2 provided the best activity for 1,3-butadiene production, due to its optimum acid and base ratio together with the Cu content that helps promote the formation of acetaldehyde, resulting in the highest yield of 1,3-butadiene.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to acknowledge the financial supports from IRPC Public Company Limited, Center of Excellent on Petrochemical and Materials Technology (PETROMAT), Thailand, and The Petroleum and Petrochemical College, Chulalongkorn University, Thailand.\n\nGC-TOFMS\n\nGas Chromatography-equipped with a Mass Spectrometry of Time-of-Fight Type\n\nLDHs\n\nLayered Double Hydroxides\n\nLDOs\n\nLayered Double Oxides\n\nTPD\n\nTemperature Programmed Desorption\n\nTPDRO\n\nTemperature Programmed Desorption/Reduction/Oxidation.\n\nXRD\n\nX-Ray Diffraction\n\nXRF\n\nX-Ray Fluorescence Spectrometer\n\nSEM\n\nScanning Electron Microscope\n\nTEM\n\nTransmission electron microscopy\n\nBET\n\nBELCAT II. The Brunauer-Emmett-Teller\n\nMPV\n\nMeerwein\u2013Ponndorf\u2013Verley\n\n\nGas Chromatography-equipped with a Mass Spectrometry of Time-of-Fight TypeLayered Double HydroxidesLayered Double OxidesTemperature Programmed DesorptionTemperature Programmed Desorption/Reduction/Oxidation.X-Ray DiffractionX-Ray Fluorescence SpectrometerScanning Electron MicroscopeTransmission electron microscopyBELCAT II. The Brunauer-Emmett-TellerMeerwein\u2013Ponndorf\u2013VerleyThe following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.clet.2021.100088.", "descript": "\n                  Core-shell particles is a type of materials that consist of an inner core structure and an outer shell made from different components. They have been employed as a catalyst due to their unique properties, arising from the combination of the core and shell materials. In addtion, the properties of the core-shell particles can be designed using several suface modification techniques in order to improve the activity and stability of the catalyst. Chemical grafting is a surface modification method that involves the reaction between a metal alkoxide precursor and the surface hydroxyl group of a support. This technique has been reported to be one of the most interesting surface modification techniques that provide a well-dispersed metal oxide on the surface of a support. Micro-emulsion is also considered as one of the simplest and effective methods for the preparation of nano-sized particles with a narrow size distribution, which can be immobilized on the surface of a support. In this work, the impact of preparation methods on the physical and chemical properties of the CuO@ZrO2/FeMgAl-LDO shell@core catalyst and 1,3-butadiene production were investigated. The catalyst samples were characterized using several techniques, including XRD, XRF, BET, NH3-TPD, CO2-TPD, TEM-EDX, and SEM-EDX. The activity of catalyst samples on ethanol conversion was performed in a continuous U-tube fixed-bed reactor at 400\u00a0\u200b\u00b0C and atmospheric pressure. It was found that the conversion of ethanol to 1,3-butadiene was a surface sensitive reaction. Therefore, the specific catalyst preparation method was required in order to achieve a high 1,3-butadiene yield. Particularly, in this case, the catalyst that provided the highest yield of 1,3-butadiene was in the form of ZrO2-grafted granular catalyst (CuO@gf-ZrO2/g-FeMgAl-LDO one). Based on the characterization results, the grafted granular core-shell catalyst was the one that had the highest ratio of total base/acid sites.\n               "}
{"full_text": "In the current energy scenario, the production of heat, power and biofuels from biomass has become of major interest. In this line, gasification is a key technology for the large-scale exploitation of biomass. However, the development of biomass gasification is conditioned by the efficient conversion of the feed and the formation of troublesome by-products, such as tars [1\u20136].The tar is a complex mixture of high molecular weight aromatic hydrocarbons, which cause fouling, corrosion and blocking of downstream equipment, leading to unacceptable level of maintenance for engines and turbines. Apart from the total concentration of the tar, its nature (mainly its dew point) also determines the problems associated with this matter. Nevertheless, the tar contains a significant amount of energy that could be transformed into syngas by acting on the operating conditions, reactor design and gas conditioning systems [7\u201312].The design of conventional conical spouted bed reactors has recently been modified to optimize reactor performance, especially for biomass steam gasification. Conventional spouted bed reactors are characterized by short gas residence times, which is an advantage for pyrolysis processes, but a severe drawback for gasification ones because tar cracking/reforming reactions are avoided [13,14]. Thus, the fountain confined spouted bed reactor has been developed to overcome these problems and improve the overall process efficiency [15\u201317]. Moreover, this novel technology widens the applicability range of conventional spouted beds, as it may handle very fine particles without elutriation from the bed by confining in the fountain the gases produced in the bed, and therefore lengthening their path. Therefore, the gas residence time is increased and the gas-solid contact improved, which is even better under the fountain enhanced regime. The latter regime is characterized by a great expansion of the fountain region, which significantly improves the contact between the gas and the solid, and therefore tar conversion [18\u201321].Catalytic gas cleaning methods for tar removal also entail an additional increase in H2 and gas productions, as they promote tar cracking and steam reforming reactions. These catalysts may be used as primary catalysts directly in the gasifier, or as secondary catalysts in downstream catalytic processes. Thus, in the case of fluidized bed reactors, the use of an active and appropriate in-bed material as primary catalyst is a promising strategy to decrease the tar content in comparison with the use of a more expensive secondary catalytic reactor downstream [22\u201328].A large number of materials with significant activity for tar cracking and reforming have been used as primary catalysts. Natural minerals, such as dolomite and olivine, have attracted most of the attention because, apart from being active for tar cracking and reforming, they are inexpensive and abundant. Although the activity of dolomite is reported to overcome that of olivine, it is very fragile and undergoes severe attrition when used in fluidized beds. Furthermore, olivine has higher mechanical strength, comparable to that of sand [14,29\u201336]. However, the catalytic activity of these primary materials for tar conversion leaves room for improvement by metal phase addition.Ni based catalysts are more effective for converting tar into hydrogen-rich gas, but they undergo a rapid deactivation by coke deposition and are toxic [37\u201344]. Recently, iron based catalysts have gained considerable attention among the catalysts for tar removal. Compared to nickel, the use of iron reduces the catalyst cost and lowers its toxicity [45\u201355]. Apart from the well-known activity of metallic iron for tar reforming and cracking, magnetite (Fe3O4) has also been proven to be active for the WGS reaction [56\u201358]. Therefore, impregnation of natural minerals with iron seems to be an interesting alternative to synthesize primary catalysts.The novelty of this paper is associated with the proposal of a novel and efficient gasification technology. Thus, an original gasification technology based on the fountain confined spouted bed reactor has been developed. This reactor is able to operate under a vigorous fluidization regime (enhanced fountain regime), which greatly improves the gas-solid contact, and therefore the catalyst efficiency. Accordingly, this paper assesses the potential benefits of the fountain confined conical spouted bed for reducing the tar produced during biomass gasification, and the potential to improve the overall process efficiency by using Fe/olivine catalysts. Furthermore, the novelty is also related to the role of active iron species and their behaviour in biomass steam gasification. Thus, this paper analyses the performance and stability of an Fe/catalyst and relate its activity for biomass steam gasification with its physical and metallic properties. A detailed characterization (BET surface area, X-ray fluorescence (XRF), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), temperature programmed reduction (TPR) and oxidation (TPO)) of the fresh and deactivated catalysts was carried out to ascertain the cause of the deactivation.The biomass feedstock used in this study was crushed and sieved forest pinewood sawdust with an average particle size of 1\u20132mm and dried to a moisture content below 10wt%. The ultimate and proximate analyses were conducted in a LECO CHNS-932 elemental analyzer and in a TGA Q5000IR thermogravimetric analyzer, respectively. Moreover, the higher heating value (HHV) was measured in a Parr 1356 isoperibolic bomb calorimeter. The main features of the biomass used are listed in Table 1\n.Olivine supplied by Minerals Sibelco was used in this study as catalyst support and primary catalyst. Based on previous studies [19], olivine in the 90\u2013150\u03bcm particle size range was used in order to operate in the fountain enhanced regime. The performance of the Fe/olivine catalyst was compared with that of olivine, which was calcined in situ at 850\u00b0C.An Fe/olivine catalyst with 5wt% iron was prepared by wet impregnation of the support with an aqueous solution of Fe(NO3)3\u00b79H2O (Panreac AppliChem, 98%), by means of a rotavapor, which allows evaporating the solution under reduced pressure and moderate temperatures. The rotavapor used was a B\u00fcchi rotavapor R-114, which operates under vacuum at 70\u00b0C. A relatively low metallic load (5wt%) was used, as the physical properties of olivine hinder a suitable dispersion of the metal due to its non-porous nature. The iron solution was added to the support and the water excess was evaporated at 70\u00b0C and under vacuum environment. The samples were dried in an oven at 100\u00b0C for a couple of days and calcined in a muffle oven at 1000\u00b0C for 4h.The physical properties of the catalyst (specific surface area, pore volume and average pore size) were determined by N2 adsorption\u2013desorption in a Micromeritics ASAP 2010 apparatus. Before each analysis, the samples were degassed under vacuum at 150\u00b0C overnight. Surface area was calculated based on the BET equation, whereas the pore size distribution was determined by BJH method.X-ray fluorescence (XRF) spectrometry was used to measure the chemical composition (wt%) of both the calcined olivine and the synthesized catalyst. The chemical analysis of the particles was carried out under vacuum atmosphere using a sequential wavelength dispersion X-ray fluorescence (WDXRF) spectrometer (Axios 2005, PANalytical) equipped with a Rh tube, and three detectors (gaseous flow, scintillation and Xe sealing). The calibration lines were determined by means of well-characterized international patterns of rocks and minerals.X-ray powder diffraction (XRD) patterns were obtained in a Bruker D8 Advance using CuK\u03b1 radiation equipped with a Germanium primary monochromator and Sol-X dispersive energy detector in order to analyze the crystalline structure of both the olivine and the reduced catalyst. The spectra were obtained in a 2\u03b8 range of 20\u201390\u00b0. The diffraction spectra were indexed by comparing with JCPDS files (Joint Committee on Powder Diffraction Standards).X-ray photoelectron spectroscopy (XPS) analysis was carried out to record in detail the elements making up the surface, and quantify and analyze their oxidation states. XPS measurements were conducted in a SPECS system equipped with a Phoibos 150 1D-DLD analyzer and a monochromatic Al-K\u03b1 radiation source (1486.7eV). Prior to the analysis, the spectrometer was calibrated with Ag (Ag 3d5/2368.26eV).The reducibility of the materials was determined by Temperature Programmed Reduction (TPR) in an AutoChem II 2920 Micromeritics. The tests were carried out on a 200mg sample, through which a flow of 10vol% hydrogen in argon was circulated. Prior to the reduction experiments, the catalyst was thermally treated under He stream at 200\u00b0C in order to remove water or any impurities. The temperature was increased at a rate of 10\u00b0Cmin\u22121 from room temperature to 900\u00b0C. A thermal conductivity detector (TCD) was used to analyze the hydrogen consumption of the samples and its signal recorded continuously.Carbon deposition on the catalyst was ascertained by temperature programmed oxidation (TPO) in a Thermobalance (TGA Q5000 TA Instruments) coupled in-line to a mass spectrometer (Thermostar Balzers Instrument). This device allows recording the signals at 14, 18, 28 and 44 atomic numbers, corresponding to N2, H2O, CO and CO2, respectively. The coke content was determined based on CO2 signal. Once the signal was stabilized under N2 stream (50mLmin\u22121) at 100\u00b0C, oxidation of the sample with air was carried out by increasing temperature to 800\u00b0C using a ramp of 5\u00b0Cmin\u22121 and keeping the final temperature for 30min to ensure total carbon combustion.The biomass steam gasification experiments were performed in a bench scale unit based on the conical spouted bed technology (Fig. 1\n). This reactor was designed for biomass valorization processes, and specifically fine-tuned and optimized for the pyrolysis and gasification of different solid wastes [59\u201363].The main element of the plant is a fountain confined conical spouted bed, which is also provided with a non-porous draft tube (internal diameter 5.5mm and entrainment zone height 15mm). It enables operating in a wide range of conditions and improving the hydrodynamic performance of the reactor. This reactor may also operate in the conventional spouting regime by using a lid without confiner. The main dimensions of the reactor are as follows: cylindrical section diameter 95mm, height of the conical section 150mm, cone included angle 30\u00b0, length of the fountain confiner 330mm, and total height of the reactor 430mm. The cone base diameter is 20mm, and the internal diameter of the fountain confiner 54mm, with its volume being of around 0.8L. The height from the reactor base to the lower end of the confiner is 105mm. The fountain confiner is a tube welded to the lid of the reactor, whose lower end is close to the bed surface and confines the gases generated in the bed and force them to circulate upwards through the core of the fountain and downwards through its periphery. This device increases the residence time, narrows its distribution and enhances gas-solid contact in the fountain region [17]. Furthermore, the draft tube makes operation feasible in a much wider range of gas flowrates and improves bed stability [64]. More details about the reactor, fountain confinement technology and draft tube have been reported elsewhere [18,19]. A gas preheater is located below the reactor to heat the gases to reaction conditions. A radiant oven made up of two independent sections heats the gasifier (the lower section heats the gas preheater and the upper section the fountain confined spouted bed reactor). The temperature in each section is controlled by two thermocouples, one placed in the bed annulus and the other one at the inlet of the gaseous stream.All the unit elements, i.e., the reactor, the interconnection pipes, the cyclone and the filter, are located inside a forced convection oven, which is 1830\u00d71950\u00d71000mm stainless steel box and features 100mm insulation of quartz wool with fiberglass reinforcement fabric, kept at 300\u00b0C to prevent tar condensation before the condensation system. The high-efficiency cyclone and 5\u03bcm sintered steel filter retain the char and catalyst fine particles entrained from the bed.The biomass was fed by means of a piston dispenser. This system consists of a cylindrical vessel equipped with a vertical shaft connected to a piston placed below the material bed. When the piston rises, the biomass is pushed towards the top of the feeding system and drops into the reactor through a tube cooled with tap water at the same time as the whole system is vibrating by means of an electric engine to prevent biomass agglomeration. Moreover, a very small nitrogen flow is introduced from the top of the feeding vessel in order to ease the solid flow into the reactor and avoid the condensation of steam in the dispenser. A detailed description of the functioning of this device to feed the solid has been reported elsewhere [20]. The water flowrate was measured by an ASI 521 pump and directed to an evaporator to produce steam before entering the reactor. The plant is also provided with three mass flow meters for N2, air and H2, with N2 being used as fluidizing agent during the heating process and H2 for the reduction of Fe catalyst prior to the runs.The volatile condensation system is located after the convection oven and consists of a double-shell tube condenser cooled by tap water, a 1L vessel, a stainless steel filter (60\u03bcm) and a coalescence filter, with the latter ensuring total recovery of the tars.Continuous gasification experiments were performed at 850\u00b0C using steam as fluidizing gas. Thus, a water flow rate of 1.5mLmin\u22121 was employed (1.86 NL min\u22121 of steam) and biomass was fed at a rate of 0.75gmin\u22121, which corresponded to a steam/biomass (S/B) ratio of 2. This high S/B ratio is needed to guarantee vigorous spouting conditions (enhanced fountain regime), as steam not only acts as gasifying agent, but also as fluidizing gas.The bed consisted of 100g either calcined olivine or Fe/olivine catalyst, with their particle size being in the 90\u2013150\u03bcm range. Prior to the reactions, the iron catalyst was subjected to an in situ reduction process at 850\u00b0C for 4h with a stream containing 10vol% of H2 to ensure complete reduction to Fe0 phase.The experiments were performed in continuous regime and the gas chromatography (GC) and micro GC analyses were conducted once several minutes of operation had elapsed to ensure steady state conditions. Moreover, the runs were repeated several times (at least three) under the same conditions in order to guarantee reproducibility of results.Certain limitations in the experimental unit for operation and product monitoring had to be addressed. Thus, the feeder must be refilled after 40min operation and the GC analysis of the tars lasts about 20min. Accordingly, successive 10min reactions were carried out in order to ensure suitable product analysis throughout continuous operation and overcome these limitations.Samples of the volatile stream leaving the reactor were analysed on-line by means of a GC Agilent 7890 outfitted with a HP-Pona column (50m long, 0.2mm in internal diameter and a coating thickness of 0.5mm) a flame ionization detector (FID). The sample was injected into the GC through a line thermostated at 280\u00b0C to avoid the condensation of the tars. The temperature programme used in the GC is as follow: 2min at 40\u00b0C in order to attain a good separation; a sequence of 25\u00b0Cmin\u22121 up to 320\u00b0C and 7min at this temperature to ensure that all products were outside the column. Furthermore, the non-condensable gaseous stream was also analysed online in a Varian 4900 micro GC equipped with three modules (molecular sieve, Porapak (PPQ) and plot alumina) and thermal conductivity detectors (TCD), which allow both identification and quantification of gaseous products previously calibrated. The conditions for the analysis were the same for the modules: column temperature 90\u00b0C, injector temperature 100\u00b0C and pressure 20psi. In this case, the sampling point was located after the condensation and filtering sections. Moreover, the tars retained in the condensation system were identified by Shimadzu UP-2010S GC/MS once they had been dissolved in acetone.In order to assess the gasification performance the following reaction parameters were considered:\n\n\u2022\nGas production (P\n\ngas\n, Nm3\nkg\u22121) based on the mass unit of biomass fed into the gasification process:\n\n\n\n\n\n(1)\n\n\n\nP\n\ng\na\ns\n\n\n=\n\n\n\nQ\n\ng\na\ns\n\n\n\n\n\nm\n0\n\n\n\n\n\n\n\n\n\n\nwhere Q\n\ngas\n is the volumetric flow rate of the gas produced and m\n0 is the mass flow rate of biomass fed into the process.\n\n\n\u2022\nH2 production (\n\n\nP\n\n\nH\n2\n\n\n\n\n, wt%) by mass unit of the biomass fed into the reactor, which is calculated as follows:\n\n\n\n\n\n(2)\n\n\n\nP\n\n\nH\n2\n\n\n\n=\n\n\n\nm\n\n\nH\n2\n\n\n\n\n\n\nm\n0\n\n\n\n\u22c5\n100\n\n\n\n\n\n\n\nwhere \n\n\nm\n\n\nH\n2\n\n\n\n\n and m\n0 are the mass flow rates of the H2 produced and biomass fed into the reactor, respectively.\n\n\n\u2022\nTar concentration determined as the amount of tar (in mass) per m3 of syngas:\n\n\n\n\n\n(3)\n\n\nTar concentration\n:\n\u2003\n\n\n\nm\n\nt\na\nr\n\n\n\n\n\nQ\n\ng\na\ns\n\n\n\n\n\n\n\n\n\n\n\n\u2022Carbon conversion efficiency defined as the ratio between the moles of C in the gaseous product and those entering the reactor.\n\n\n(4)\n\nX\n=\n\n\n\nC\n\ng\na\ns\n\n\n\n\n\nC\n\nb\ni\no\nm\na\ns\ns\n\n\n\n\n\u22c5\n100\n\n\n\n\n\nGas production (P\n\ngas\n, Nm3\nkg\u22121) based on the mass unit of biomass fed into the gasification process:\n\n\n(1)\n\n\n\nP\n\ng\na\ns\n\n\n=\n\n\n\nQ\n\ng\na\ns\n\n\n\n\n\nm\n0\n\n\n\n\n\n\n\nwhere Q\n\ngas\n is the volumetric flow rate of the gas produced and m\n0 is the mass flow rate of biomass fed into the process.H2 production (\n\n\nP\n\n\nH\n2\n\n\n\n\n, wt%) by mass unit of the biomass fed into the reactor, which is calculated as follows:\n\n\n(2)\n\n\n\nP\n\n\nH\n2\n\n\n\n=\n\n\n\nm\n\n\nH\n2\n\n\n\n\n\n\nm\n0\n\n\n\n\u22c5\n100\n\n\n\n\nwhere \n\n\nm\n\n\nH\n2\n\n\n\n\n and m\n0 are the mass flow rates of the H2 produced and biomass fed into the reactor, respectively.Tar concentration determined as the amount of tar (in mass) per m3 of syngas:\n\n\n(3)\n\n\nTar concentration\n:\n\u2003\n\n\n\nm\n\nt\na\nr\n\n\n\n\n\nQ\n\ng\na\ns\n\n\n\n\n\n\n\n\n\u2022Carbon conversion efficiency defined as the ratio between the moles of C in the gaseous product and those entering the reactor.(4)\n\nX\n=\n\n\n\nC\n\ng\na\ns\n\n\n\n\n\nC\n\nb\ni\no\nm\na\ns\ns\n\n\n\n\n\u22c5\n100\n\n\n\nTable 2\n shows the specific surface area, pore volume and average pore size. As observed, the specific surface area of the calcined olivine was as low as 1.92 m2\ng\u22121 and the pore volume 0.0023 cm3\ng\u22121, which are evidences of its non-porous structure. Regarding the synthesized Fe/olivine catalyst, olivine physical properties were improved by iron impregnation. Thus, pore volume and average pore size became larger, which was due to the collapse of the inter-pore structure of olivine. Likewise, the specific area also increased, which may be attributed to the deposition of Fe on the external surface. This trend has also been reported for Ni impregnation on low porosity surfaces [65,66].The chemical compositions of the calcined olivine and the prepared catalyst are summarized in Table 3\n. The content of Fe in the olivine was of around 5.2wt%. After impregnation, Fe content in the catalyst increased significantly, 10.2wt%, which confirmed that the metal content was that corresponding to the impregnation (5wt%) plus that in the original olivine.\nFig. 2\n shows the diffractograms of the calcined olivine and fresh and reduced Fe/olivine. In the case of the calcined olivine, the XRD data revealed the main diffraction lines were characteristic to the olivine structure ((Mg1.81\u00b7Fe0.19)\u00b7(SiO4)). Additional peaks corresponding to secondary crystalline phases may also be observed, such as enstatite (MgSiO3) and quartz (SiO2). According to Michel et al. [67,68] and \u015awierczy\u0144ski et al. [69], numerous phases of iron oxide may appear subsequent to olivine calcination, as are \u03b3-Fe2O3, \u03b1-Fe2O3, Fe3O4 and MgFe2O4. The presence of these iron oxides is explained by the migration of the iron Fe2+ located within the internal structure of the olivine to its surface due to oxidation (Eq. (5)) [69,70]. However, none of these phases were detected in this study. It should be noted that the calcination temperature used for the natural olivine was rather low (850\u00b0C) compared to other studies in the literature, in which they were over 1100\u00b0C. Kuhn et al. [71] performed XRD analysis to olivine calcined at 900\u00b0C during 2h and they neither observed free Fe oxide phases. These oxide phases diffract in the same main lines as the olivine structure, but they were not strong enough to be detected and so inferred their presence. For the fresh and reduced Fe/olivine catalysts, the main crystalline forms were still those corresponding to olivine structure and MgSiO3 enstatite phase, even though the olivine was subjected to iron impregnation, calcination and reduction. However, significant changes in the relative intensity of olivine structure and MgSiO3 enstatite phases were noticed at 2\u03b8\n=21\u00b0, 31\u00b0 and 36\u00b0, which indicated certain modifications in the crystallinity of the samples due to iron impregnation. In fact, the higher intensity of the diffraction lines in the reduced catalyst is evidence of its greater crystallinity compared to the calcined olivine or fresh catalyst, which was due to iron reincorporation into the olivine structure. In addition, hematite (\u03b1-Fe2O3) peak appeared at 2\u03b8\n=24\u00b0 in the fresh catalyst, whereas for the reduced catalyst the presence of an intense peak of the metallic iron phase was observed at 2\u03b8\n=44\u00b0 and a smaller one at 2\u03b8\n=65\u00b0. Iron oxide phases were not detected in the reduced sample, which is evidence of their full reduction. Other authors reported the same main lines for this catalyst [47,48,72]. The SiO2 lines detected in the support disappeared in the catalyst. Michel et al. [67] stated that olivine phase reacts with quartz at 1000\u00b0C to form enstatite phase:\n\n(5)\n\n\n\n\n\n\nM\ng\n,\nF\ne\n\n\n\n2\n\nS\ni\n\nO\n4\n\n+\n\nO\n2\n\n\u2192\nx\nM\n\ng\n2\n\nS\ni\n\nO\n4\n\n+\n\n\n1\n\u2212\nx\n\n\nS\ni\n\nO\n2\n\n+\n\n\n2\nx\n\u2212\n1\n\n\nF\n\ne\n2\n\n\nO\n3\n\n\n\u03b1\n\n+\n2\n\n\n1\n\u2212\nx\n\n\nM\ng\nF\n\ne\n2\n\n\nO\n4\n\n\n\n\n\n\nFig. 3\n shows the XPS spectra for the samples in different binding energy ranges. This analysis revealed the main components on the surface of the samples, which were Si, Mg, Fe and O. No significant changes were observed in Si after iron impregnation and catalyst reduction, whereas more pronounced changes were detected in the peaks corresponding to Mg and Fe. In the case of Fe, its oxidation states are analyzed below in the paper. These variations are also visible in Tables 4 and 5\n\n. Furthermore, peaks of other trace elements, previously detected by XRF, were not observed, which is evidence that they were not located on the surface.\nTable 4 shows the surface composition of the samples. The quantification of each element was carried out by integrating the intensities of Si 2p, Mg 2p, O 1s and Fe 2p using Scofield sensitivity factors. As observed, after iron impregnation, the amount of iron on the catalyst surface increased (from 6.2 to 8%), which suggests that part of the impregnated iron was deposited on the surface of the catalyst, as evidenced by the increase in the BET surface of the catalyst (Table 2). However, the amount of Mg on the surface decreased (from 17.5 to 14.2%) after iron loading. According to Frekdisson et al. [73], after the oxidizing treatments, the surface is enriched in Fe at the expense of Mg. Furthermore, catalyst reduction with H2 led to a decrease in the amount of Fe to 4.4% and an increase in that of Mg to 22.1% on the surface. Under reducing conditions, Fe clustered into large particles and incorporated into the olivine structure [71]. Regarding oxygen concentration, its oscillations on the surface of the catalyst were related to the oxidation state of iron.XPS spectra in the 700\u2013750eV binding energy range of the samples were analyzed to further understand the valence state of the iron in the calcined olivine and fresh and reduced Fe/olivine catalysts (Fig. 3b and c). Accordingly, Fe 2p lines were used instead of Fe 3p because they were stronger. Moreover, Table 5 shows the iron distribution on the surface of the samples. Yamashita and Hayes [74] reported that Fe 2p3/2 peak at 711eV with satellite peak at 719eV and Fe 2p1/2 peak at 725eV with satellite peak at 732eV are characteristic of Fe3+, whereas Fe 2p3/2 peak at 709eV with satellite peak at 714eV and Fe 2p1/2 peak at 723eV with satellite peak at 728eV correspond to Fe2+. In Fig. 3c, the positions of these peaks are marked with dashed lines. Iron in Fe3+ state corresponds to Fe2O3 and MgFe2O4 compounds, whereas Fe2+ state is characteristic of iron in the olivine structure and FeO. In the calcined olivine, most of the Fe was as Fe3+ and doubled the amount of Fe as Fe2+, which is evidence that a higher amount of iron led to free oxides on the surface than those remained within the olivine structure. The presence of free iron oxide phases (Fe3+) stemmed from Fe migration from the olivine structure (Fe2+) during the calcination process [69,73], although none of these compounds were detected by XRD analysis. Regarding iron distribution, the fresh Fe/olivine catalyst followed the same trend as the calcined olivine. However, when the former is compared with the calcined olivine, the amount of Fe2+ in the fresh catalyst increased (from 32. 89 to 35.42%), whereas that of Fe3+ decreased (from 67.11 to 64.54%), although Fe2+/Fe3+ ratio remained approximately constant. These results suggest that, after impregnation, the iron within the olivine structure was preferably in the metallic state rather than forming free oxides. After reduction, a weak peak of metallic Fe appeared at 707eV, which cannot be quantified due to its very small size. It seems that the metallic iron on the catalyst surface was oxidized due to its contact with air, but the iron inside the olivine remained in the metallic form, as was revealed by the XRD analysis (Fig. 2). Moreover, the Fe2+/Fe3+ ratio in the reduced catalyst was higher than that in the fresh one, with the amount of Fe2+ and Fe3+ being almost the same. Thus, the oxidation state of the iron located on the surface changed from a Fe3+ dominating state after oxidation to Fe2+ state after reduction [73]. Meng et al. [72] observed the same trend for the iron distribution on the surface of the catalyst.H2-TPR experiments for the bed materials were carried out prior to their use in the reaction environment. The TPR profile of the catalysts enables determining the temperature needed for their reduction [75]. As well-known, the profile depends not only on the nature of the metallic species, but also on the metal-support interactions. Moreover, as the metallic iron is supposed to be the active phase for hydrocarbon cracking, the reducibility of the catalysts is of great relevance [76].The TPR profiles of the calcined olivine and synthesized catalyst are shown in Fig. 4\n. In the case of the calcined olivine, two small peaks are observed between 350 and 550\u00b0C. A third peak is also observed at a reduction temperature above 600\u00b0C. In the case of the first two peaks, their low reduction temperature is evidence that these species are easy to reduce. Thus, these peaks are attributed to the reduction of iron oxides on the olivine surface [77]. According to the XPS analysis (Table 5), the surface of the calcined olivine is presumably made up of Fe2O3 and/or MgFe2O4, which migrated from the internal olivine structure during the calcination [48,69,71,72]. Thus, the peak at 350\u00b0C is assigned to the reduction of Fe2O3 and the peak at 550\u00b0C to the reduction of Fe3O4, as the reduction of Fe2O3 to metallic Fe occurs in two steps (Fe2O3\n\u2192Fe3O4\n\u2192Fe0) [46,48]. The peak that might appear at higher temperatures is associated with the reduction of iron phases inside the olivine grain, in which reduction is more difficult. The TPR profile of the Fe/olivine catalyst shows a broad reduction zone covering the range from 300 to 700\u00b0C. Three main peaks may be observed, with the first two being associated with the two-step oxidation of Fe2O3 on the olivine surface and the peak above 600\u00b0C to the Fe atoms that migrated into the olivine support to form a very stable MgFe2O4 spinel phase [78]. In the case of the Fe/olivine catalyst, the reduction of iron phases inside the olivine grain is not observed due to the high stability of the olivine structure, i.e., higher temperatures are required for its reduction.The effect of Fe/olivine catalyst on the steam gasification process parameters (H2 and gas productions, gas composition, carbon conversion and tar concentration and composition) was assessed and compared with that of calcined olivine. The performance of these in-bed materials was analyzed based on the following reactions inside the gasifier:\n\n(6)\n\n\nBiomass pyrolysis\n:\n\u2003\nB\ni\no\nm\na\ns\ns\n\u2192\nG\na\ns\ne\ns\n\n\nC\nO\n,\nC\n\nO\n2\n\n,\nC\n\nH\n4\n\n,\n\nH\n2\n\n,\n\u2026\n\n\n+\no\nx\ny\ng\ne\nn\na\nt\ne\ns\n+\nc\nh\na\nr\n\n\n\n\n\n\n(7)\n\n\nSteam gasification of the char\n:\n\u2003\nC\n(\ns\n)\n+\n\nH\n2\n\nO\n\u2192\nC\nO\n+\n\nH\n2\n\n\n\n\n\n\n\n(8)\n\n\n\nCO\n2\n\n gasification of the char\n:\n\u2003\nC\n(\ns\n)\n+\nC\n\nO\n2\n\n\u2192\n2\nC\nO\n\n\n\n\n\n\n(9)\n\n\nTar cracking\n:\n\u2003\nT\na\nr\n\u2192\nC\nO\n+\n\nH\n2\n\n+\nC\n\nO\n2\n\n+\nC\n+\nC\n\nH\n4\n\n+\n\u22ef\n\n\n\n\n\n\n(10)\n\n\nTar steam reforming\n:\n\u2003\nT\na\nr\n+\n\nH\n2\n\nO\n\u2192\nC\nO\n+\n\nH\n2\n\n\n\n\n\n\n\n(11)\n\n\nWater gas-shift\n(\nWGS\n)\n:\n\u2003\nC\nO\n+\n\nH\n2\n\nO\n\u21d4\nC\n\nO\n2\n\n+\n\nH\n2\n\n\n\n\n\n\n\n(12)\n\n\nMethane\n(\nor hydrocarbon\n)\nsteam reforming\n:\n\u2003\nC\n\nH\n4\n\n+\n\nH\n2\n\nO\n\u21d4\nC\nO\n+\n3\n\nH\n2\n\n\n\n\n\nAs observed in Fig. 5\n, all representative gasification parameters were significantly improved on the Fe/olivine catalyst. An increase in gas and hydrogen productions and a decrease in tar concentration was noticeable when 5wt%Fe/olivine was used instead of olivine (Fig. 5a and b). Thus, gas production increased from 1.30 to 1.46Nm3\nkg\u22121 and so did the hydrogen production, from around 5 wt% on the olivine to 6.25wt% on the iron impregnated catalyst. Fig. 6\n illustrates the product gas composition for the runs using 5 wt%Fe/olivine catalyst and calcined olivine. Iron impregnation led to an increase in H2 concentration from 43.2 to 48.2vol% and a reduction in that of CO, which implies that H2/CO ratio increased from 1.41 for olivine to 3.26 for the iron catalyst. Consequently, CO2 concentration increased to 28.2vol%. From these results, it could be deduced that the addition of iron to olivine enhances the WGS reaction (Eq. (10)), as well as light hydrocarbon steam reforming and cracking reactions (Eqs. (8) and (11)). Consequently, tar concentration was reduced approximately to half, from 20.6 to 10.4gNm\u22123, and carbon conversion efficiency accounted for 87.6% (Fig. 5c and d). Likewise, the heating value of the gas increased from 2.44MJm\u22123 with calcined olivine to 8.66MJm\u22123 with 5 wt%Fe/olivine catalyst. According to several authors [73,76,79], the metallic Fe on the reduced catalyst enhances tar decomposition reactions. Moreover, the BET surface area (Table 2) and XPS analyses (Table 4) revealed that Fe was mainly located on the external surface of the catalyst, and was therefore easily accessible to the volatiles and promotes tar cracking and reforming reactions (Eqs. (8) and (9)).Although there are many studies dealing with steam reforming of biomass tar model compounds using a wide variety of supported metal catalyst [80\u201389], those dealing with the effect of metal impregnated in situ catalysts on the biomass steam gasification are scarce, especially those carried out in laboratory pilot plants. Several authors reported the same trend as that obtained in this study for iron impregnated olivine and compared its activity with that of raw olivine using different gasification technologies [45,49,50,90,91]. Thus, Rapagn\u00e0 et al. [49] studied the performance of 10 wt%Fe/olivine catalysts in the biomass steam gasification at 820\u00b0C in a fluidized bed gasifier and obtained slightly higher reaction indices than in this study. They observed that H2 and gas productions increased from 3.5 to 6.6 wt% (the gaseous stream contained 53 vol% of H2) and from 1 to 1.4Nm3\nkg\u22121, respectively when an Fe/olivine catalyst was used instead of raw olivine, whereas tar concentration was reduced by approximately 62%, with the value being 2.25gNm\u22123 with the catalyst. Carbon conversion efficiency reached a value of 80%, which was similar to that obtained with raw olivine. Virginie et al. [50] used the same catalyst as the previous authors, but they used a dual fluidized bed. They reported that tar reduction was more notable in the presence of Fe/olivine in the bed than in the run with raw olivine (5.1 and 2.6gNm\u22123 of tar content for olivine and Fe/olivine at 850\u00b0C). In addition, Barisano et al. [90,91] evaluated the performance of 10 wt%Fe/olivine catalyst in the biomass steam/O2 gasification at 890\u00b0C in an internal circulating bubbling fluidized bed (ICBFB) and they reported 1.2Nm3\nkg\u22121 and 3wt% for the gas and H2 productions, respectively. They also reported a reduction in the total tar content by 38% (from 10.1 to 6.2gNm\u22123), and 98% of carbon conversion efficiency was therefore attained. However, Pan et al. [45] used a lower Fe load in the catalyst (5wt%Fe/olivine) for the steam co-gasification of pine sawdust and bituminous coal in a pyrolysis-reforming-combustion decoupled triple bed system (DTBG) at 850\u00b0C. All the studied reaction indices were improved, but the differences were not as remarkable as those observed for the biomass steam gasification. Thus, they obtained gas and H2 productions of 0.66Nm3\nkg\u22121 and 2.49wt% (10% higher in both cases) and a tar content as low as 4.87gNm\u22123 (17% reduction).Ni loading to olivine also enhances tar reforming activity in the biomass steam gasification, with the performance being even better than that of Fe/olivine catalyst. Thus, Pfeifer et al. [92] studied tar removal activity of Ni/olivine catalyst in a 100kWth dual fluidized bed reactor. After adding 20% of 5wt%Ni/olivine catalyst to a bed of olivine, the tar concentration was reduced by half and gas and H2 productions increased to 1Nm3\nkg\u22121 and 3.93wt%, respectively at 850\u00b0C. Michel et al. [93] used in situ 3.9wt%Ni/olivine catalyst in the biomass steam gasification carried out in fluidized bed at 800\u00b0C and reported a higher efficiency of the catalyst compared to raw olivine. Thus, they obtained H2 and gas productions of 7.56wt% and 1.7Nm3\nkg\u22121, instead of 3.36wt% and 1Nm3\nkg\u22121 with olivine, and less than 1wt% of tar. More recently, Tursun et al. [94] used 5wt%Ni/olivine catalyst in a decoupled triple bed gasification system consisting of a pyrolyzer, reformer and combustor, and reported that the catalyst not only improved tar removal, but also enhanced H2 and gas productions. Their results were slightly better than those obtained by Michel et al. [93], but the Ni loading was also slightly higher. They reported a gas production of 1.59Nm3\nkg\u22121, with H2 concentration being 56.1vol% (H2 production of 7.96wt%) and tar content as low as 0.6gNm\u22123.The tar fraction is a mixture containing polycyclic aromatic compounds larger than benzene, and is commonly studied by dividing into four lumps [7], as are: light aromatics (monoaromatic compounds), heterocycles (aromatic rings with heteroatoms), light polyaromatics (PAHs with up to 3 rings) and heavy polyaromatics (PAHs with more than 3 rings). Apart from the total concentration of the tar, its nature (mainly its dew point) is of high relevance, as it is responsible of problems related to fouling and sooting.\nFig. 7\n shows a significant reduction in the amount of heterocycles and heavy PAHs using Fe/olivine catalyst. In fact, the mass fraction of those lumps was reduced from 10.33 and 10.20 to 7.43 and 5.05wt%, respectively. However, the percentage of light aromatics and PAHs in the total tar amount increased from 14.22 and 62.09 to 19.91 and 65.20wt%. It is noteworthy that the Fe/olivine catalyst managed to reduce significantly the concentration of all tar families, as shown in Table 6\n. Based on the tar formation and PAH growth mechanisms [95], the Fe/olivine catalyst seems to hinder the growth of light PAHs into heavier ones, and the amount of the light PAHs was therefore higher. Furthermore, Diels\u2013Alder reactions involving light alkenes in the permanent gases and phenols may produce light aromatics, and therefore its amount was increased [96\u201398].\nTable 6 provides a detailed composition of the tar obtained with raw olivine and Fe/olivine catalyst. Naphthalene was the most abundant tar molecule for calcined olivine and Fe/olivine catalyst, although its concentration was reduced by 42% approximately with the iron enrich catalyst. Barisano et al. [91] reported a higher naphthalene reduction (of around 58%) in the biomass steam/O2 gasification. Moreover, compounds such as phenol, methyl phenol, 1-methyl naphthalene, dibenzofuran, 1-H phenalene, 2-phenyl naphthalene and pirene were significantly removed, as the catalyst managed to reduce their content beyond 60%. Thus, it is clear that metallic iron is active for CC and CH bond breakdown [76,99]. The results in Table 6 also show the more stable tar compounds, which are those that are more difficult to remove. Using the Fe/olivine catalyst the concentration of toluene, naphthalene and anthracene was reduced, but their amounts were still rather high, as they are refractory to reforming/cracking reactions [49]. Therefore, all the efforts in the development of supported metal catalysts should be directed towards their capacity for removing the most stable tar compounds.The evolution of the gasification performance (Fig. 8\n) and gas and tar compositions (Figs. 9 and 10\n\n) were monitored for Fe/olivine with time on stream. In the case of the calcined olivine, gasification performance remained stable after 140min on stream. The main properties of the Fe/olivine catalyst and their role on the biomass steam gasification explain these results.\nFig. 8 illustrates the evolution of the reaction indexes as a function of time on stream for Fe/olivine catalyst. Even though the performance of the calcined olivine remained stable after 140min on stream, that of Fe/olivine catalyst underwent deactivation and the efficiency of the gasification process decreased with time on stream. Catalyst deactivation was especially evident by tar concentration, which increased by around 90%, from 10.4 to 19.8gNm\u22123, as shown in Fig. 8c. After 140min on stream, the amount of tar produced with the Fe/olivine catalyst reached almost that obtained with the calcined olivine (20.6gNm\u22123). Other reaction indexes also showed the deterioration of the catalyst. Thus, gas and H2 productions declined from 1.46Nm3\nkg\u22121 and 6.25wt% to 1.35Nm3\nkg\u22121 and 5.44wt%, respectively (Fig. 8a and b). However, the gas and H2 productions were still above those obtained with calcined olivine, which suggests that although the catalyst was not able to maintain its original tar elimination capacity, it was still active in the WGS reaction. Likewise, a similar trend was observed in the evolution of the gas composition (Fig. 9). H2 concentration slightly decreased from 48.2 to 45.5vol%, whereas that of CO increased from 14.3 to 20.2vol%. CO2 concentration remained almost stable at 24.9vol%. A comparison of this performance with the stable calcined olivine shows that higher H2 and CO2 concentrations were obtained, whereas the value of CO was lower due to the enhancement of the WGS reaction. Concerning CH4 and C2\u2013C4 light hydrocarbons, they showed a slightly upward trend. In the case of the deactivated catalyst, CH4 concentration was even lower (6.4vol%) than that obtained with the calcined olivine and C2\u2013C4 concentration reached a similar value as that with the calcined olivine (2.7vol%). The latter results are evidence that the Fe/olivine catalyst was still active for steam reforming of CH4 subsequent to 140min on stream.The evolution of tar lumps with time on stream is shown in Fig. 10. As the Fe/olivine catalyst was deactivated, the amount of each tar family was similar to that obtained with the calcined olivine. Thus, the amount of light aromatics and PAHs declined from 19.91 and 64.36 to 15.24 and 57.47wt%, whereas that of heterocycles and heavy PAHs increased from 7.43 and 5.05 to 10.82 and 11.63wt% after 140min on stream. Small differences were observed in the amount of light PAHs between the value with the calcined olivine and that with the deactivated Fe/olivine catalyst, which are related to the amount of unidentified compounds (there were more unidentified compounds with the deactivated catalyst). When the deactivation of the catalyst was not considerable, the Fe/olivine catalyst seemed to hinder the growth of light PAHs into heavier ones, and the amount of the light PAHs was therefore higher than for the calcined olivine. Moreover, Diels\u2013Alder reactions involving light alkenes in the permanent gases and heterocyclic compounds may also have produced light aromatics, which led to an increase in their amount [96\u201398].The prevention and attenuation of catalyst deactivation is a challenging task. Thus, most catalytic processes undergo catalyst deactivation, and therefore understanding the deactivation mechanisms is vital. In the biomass gasification processes, deactivation is mainly caused by sulphur and chlorine poisoning or carbon deposition. However, catalyst physical changes, such as sintering, phase change and attrition may also lead to catalyst deactivation. The deactivated catalyst was characterized in detail in order to understand the main causes of catalyst activity decay.\nTable 7\n shows the values of the physical properties for the fresh and deactivated Fe/olivine catalysts. After 140min on stream, the specific surface area of the Fe/olivine catalyst was significantly lower, with the reduction being even more noticeable in the pore volume and size, which underwent a more severe decrease. Therefore, the pores of the catalyst were partially blocked, which led to a decrease in the total surface area, as well as pore volume and size. The deactivated Fe/olivine catalyst had still a higher surface area and pore volume than the calcined olivine. However, the pore size was higher in the calcined olivine.In order to assess the changes in the metallic structure of the Fe/olivine catalyst after the reaction, Fig. 11\n shows the XRD patterns of the reduced and deactivated catalysts. After the reaction, the main crystalline structures were still the olivine structure and the MgSiO3 enstatite phase, although more diffraction lines corresponding to MgSiO3 phase appeared in the deactivated catalyst. The most significant differences between both spectra are related to the iron phases. In the spectrum of the deactivated catalyst, there is no evidence of the presence of metallic iron, neither in 2\u03b8\n=45\u00b0 nor 2\u03b8\n=65\u00b0 diffraction lines. However, multiple lines of Fe3O4 or MgFe2O4 spinel phase were noticeable, which are evidence of a loss of active phase by oxidation of the metallic iron under reaction conditions. Virginie et al. [50] also reported the presence of intense diffraction lines corresponding to Fe3O4 or MgFe2O4 spinel phase after reaction.XPS analysis of the deactivated catalyst was carried out to determine the components located on the surface of the catalyst after reaction. The XPS spectra of the reduced and deactivated samples in different binding energy ranges are shown in Fig. 12\n. This analysis revealed that, after the reaction, the main components on the surface of the samples were still Si, Mg, Fe and O (Fig. 12b). However, the presence of K and Ca was also observed, although the amount of the latter on the surface could not be quantified because it was very small. Their existence is probably due to the biomass ashes. As shown in Fig. 12c, metallic iron was not detected on the catalyst surface in the deactivated catalyst, which is consistent with the previous XRD results (Fig. 11).The surface composition and iron distribution in the reduced and deactivated catalysts are shown in Table 8\n. As observed, after the reaction there were small differences in the amount of Mg and Fe on the catalyst surface. The amount of iron slightly increased from 4.4 to 5.1% at the expense of Mg, which decreased from 22.1 to 18.4%. However, the iron distribution remained constant (the amount of Fe2+ and Fe3+ compounds was the same), which is an indication that iron migration from the olivine structure into the surface happened. A comparison of this catalyst with the calcined olivine shows that the deactivated catalyst had more iron in the olivine structure and a higher amount of Fe2+ compounds on its surface. Iron migration from the inside to the surface or vice versa occurs in order to reach iron equilibrium in the structure [50,73]. Regarding the amount of K on the deactivated catalyst surface (1.4%), its origin is attributed to biomass ashes. Alvarez et al. [100] reported the chemical analysis of the ashes of the same biomass used in this work and the amount of K2O was 11.3wt%. Moreover, at 850\u00b0C, potassium salts melt and they might have formed deposits on the deactivated catalyst surface.\nFig. 13\n shows the TPR curve of the fresh and spent catalysts. A single peak at 500\u00b0C with a small shoulder at a slightly higher temperature (590\u00b0C) was observed for the deactivated catalyst, which is evidence that the iron in the Fe/olivine catalyst was oxidized during the gasification process. As the XRD revealed, this peak should be attributed to the Fe3O4 or MgFe2O4 spinel phase detected. According to Meng et al. [72], the difficulty for reducing the possible iron oxides is as follows: MgFe2O4\n>FeO>Fe3O4\n>Fe2O3. However, the low reduction temperature suggests that this specie was easy to reduce, i.e., it was probably Fe3O4. Furthermore, the shoulder at 590\u00b0C is attributed to the reduction of a small amount of MgFe2O4 spinel phase. In fact, it seems that most of the MgFe2O4 spinel phase did not undergo oxidization during the reaction, as it is a very stable compound.As carbon deposition may cause catalyst deactivation, temperature programmed oxidation (TPO) was conducted on the spent Fe/olivine catalyst to quantify the amount of carbon settled. The total amount of coke and its composition depend on the operating conditions, mainly temperature and S/B ratio, as carbon deposition is a consequence of a balance between its formation and removal by gasification [101]. The TPO analysis revealed that a negligible amount of coke (0.11wt%) was deposited on the catalyst after the reaction, which is evidence that high temperatures and steam promoted the oxidation of almost all the carbon that may have formed. Fig. 14\n shows the TPO profile of the deactivated catalyst. Two different peaks are observed, which is an indication of the heterogeneous nature of the coke. According to the literature [102\u2013105], the coke combustion temperature on supported metal catalysts is related to its location on the catalyst and composition. Low combustion temperatures are attributed to the coke deposited on the metallic sites (encapsulating coke), which may catalyse coke combustion, whereas higher combustion temperatures indicate that the coke is deposited on the support, which prevents coke combustion by metallic sites. Furthermore, even if the coke is deposited on similar locations, its combustion temperature is higher as the condensation degree is higher, i.e., more organized structures with lower H/C ratios.\nFig. 14 reveals the heterogeneity of the coke. Thus, two different carbon species were detected, with their combustion temperatures being 530 and 606\u00b0C. The peak at 530\u00b0C is attributed to the amorphous coke and the shoulder at 606\u00b0C to a coke with a slightly more condensate structure. It seems that the severe reaction conditions prevent coke formation from the evolving compounds to more condensed ones due to the in situ gasification of the amorphous coke. Virginie et al. [50] observed a similar TPO profile after biomass steam gasification experiments, although their carbon oxidation temperatures were slightly higher than those obtained in this work (585 and 630\u00b0C). As the coke content was very low (0.11wt%), it cannot be stated that coke deposition caused catalyst deactivation.\nTable 9\n shows the chemical composition of the fresh and deactivated Fe/olivine catalysts. XRF analysis revealed that there was no any iron loss due to attrition phenomena, which was also checked by sieving the deactivated catalyst (it had the same size range (90\u2013150\u03bcm) as prior to the runs). Claude et al. [51] and Meng et al. [106] reported that the olivine catalysts synthesized by wet impregnation may undergo attrition, since the metallic species were mainly placed on the surface, and therefore their interaction with the support was rather weak. Some other authors studied this aspect. Thus, Virginie et al. [50] reported an iron loss of 32% after 12h gasification in a dual fluidized bed and Rapagn\u00e0 et al. [49] about 5wt% during 320min operation in a fluidized bed reactor.The gas composition in the gasifier plays a crucial role on the oxidation state of the iron located on the catalyst. During biomass steam gasification, the metallic Fe0 was oxidized as detected by XRD, XPS and TPR analyses. The operating methodology used in this study had some limitations, which may have contributed to the catalyst oxidation, as explained in the experimental section. These limitations caused changes in the reaction environment, as the fluidizing agent had to be changed from N2 to steam and ensure suitable fluidization regime prior to starting biomass feed. Thus, the presence of steam may have induced partial oxidation of the metallic phase at the beginning of the reactions. However, as biomass was fed into the reactor, the reaction environment shifted from oxidizing to reducing due to the high hydrogen concentration, and therefore the iron oxidized under steam atmosphere was reduced again. It should be noted that this problem can be avoided in full scale operation with continuous biomass feed.A similar catalyst deactivation cause was observed in the in-line steam reforming of biomass fast pyrolysis volatiles using 10wt%Co/Al2O3 catalyst by Santamaria et al. [107]. Nordgreen et al. [99] reported that, when the oxygen concentration in the reaction environment is too high, it would oxidize the metallic iron to wustite (FeO) and subsequently to magnetite (Fe3O4), since some locations favour these transformations. Based on the results obtained, when Fe was in the metallic state in the Fe/olivine catalyst, it showed a higher activity for reducing tar than when it was in the oxidized state (Fig. 8c). Changes in tar removal capacity of the Fe/olivine catalyst with time on stream may also be related to the distribution of iron oxides. Nordgreen et al. [99] also stated that the catalyst with metallic iron was capable of reducing the tar concentration above 60%, whereas the catalyst with the oxidized iron only had a capacity of 18%. The catalytic activity of iron oxides species increases with their reduction state (Fe2O3\n<Fe3O4\n<FeO<Fe0) [51]. After 140min on stream, the tar concentration obtained with the deactivated Fe/olivine catalyst and that obtained with the calcined olivine were almost the same, which suggests the presence of iron oxides leads to the same tar removal performance as calcined olivine. The same trend was observed in the evolution of tar lumps. As the Fe/olivine catalyst was deactivated, the amount of each tar family was similar to that obtained with the calcined olivine. Undoubtedly, the oxidation of metallic Fe0 sites led to their decrease, and therefore caused catalyst deactivation, as was revealed by the characterization techniques.Several studies pointed out that different Fe-phases may catalyze different reactions. Thus, Fe2O3 is reported to catalyze shoot and NOx conversion, Fe3O4 to be active in the WGS reaction and metallic Fe to catalyze Boudard reaction and tar removal reactions. Therefore, changes in the oxidation state of iron will drastically influence catalytic properties [73]. However, it seems that, after 140min on stream, the Fe/olivine catalyst was still active for WGS and CH4 steam reforming reactions, as shown in Figures 8a and 8b, which may be attributed to Fe3O4 or MgFe2O4 spinel phase detected on the deactivated catalyst (Fig. 11). However, the TPR analysis of the deactivated catalyst was more conclusive than XRD and XPS analyses, which allows inferring that Fe3O4 is the responsible, as the reduction temperature was rather low. Some authors proved Fe3O4 was active in the WGS reaction [57,58,108\u2013110].The fountain confined conical spouted bed performs well in the biomass steam gasification with primary catalysts. In fact, this reactor allows enhancing the gas-solid contact, and therefore the catalytic activity by avoiding the elutriation of fine catalyst particles.Iron incorporation to olivine proved to be beneficial in the biomass steam gasification at zero time on stream, as it allowed not only reducing tar formation, but also improving syngas production and composition. Thus, gas production was increased from 1.30Nm3\nkg\u22121 with calcined olivine to 1.46Nm3\nkg\u22121 with Fe/olivine catalyst, and similarly did the hydrogen production, with the value being 6.25wt%. Likewise, tar concentration was reduced approximately to half, from 20.6 to 11.4gNm3. This was explained by the positive effect of metallic iron, which greatly favours WGS and light hydrocarbon steam reforming and cracking reactions. At zero time on stream, naphthalene was the most abundant tar compound for both the calcined olivine and Fe/olivine catalyst, although its concentration decreased to 42wt% with the latter.The evolution of the gasification performance and gas and tar compositions with time on stream was also monitored. The stability of the Fe/olivine catalyst was lower than that of calcined olivine, which was still stable after 140min on stream. Catalyst deactivation was especially evident based on the tar concentration, which increased from 10.4 to 19.9gNm\u22123, i.e., it almost reached the value obtained on the calcined olivine (20.6gNm\u22123). Other reaction indexes also showed the deterioration of the catalyst. Thus, gas and H2 productions declined from 1.46Nm3\nkg\u22121 and 6.25wt% to 1.35Nm3\nkg\u22121 and 5.44wt%, respectively, but still remained above those obtained with the calcined olivine, suggesting the activity of the deactivated Fe/olivine catalyst for WGS and steam reforming of CH4.The characterization techniques revealed that the catalyst deactivation was due to the oxidation of the metallic iron into Fe3O4. The presence of steam in the reactor for a few minutes before starting biomass feed may have induced the partial oxidation of the metallic phase at the beginning of the reactions. However, as biomass feed started, the reaction environment shifted from oxidizing to reducing conditions, and the iron that may have oxidized became reduced again. These changes in the iron oxidation state had a great influence on the catalytic properties of the Fe/olivine, and therefore on the evolution of tar removal, as well on WGS and light hydrocarbon reforming reactions.Although the experimental unit used in this study involves certain limitations for the operation with metallic catalysts during the start-up period, the results obtained shed light on the biomass steam gasification using Fe/olivine as primary catalyst in large-scale plants.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors report no declarations of interest.This work was carried out with financial support from the Spain's Ministries of Science, Innovation and Universities (<GN1>RTI2018-098283-J-I00</GN10> (MCIU/AEI/FEDER, UE) and Science and Innovation (PID2019-107357RB-I00 (MCI/AEI/FEDER, UE), the Basque Government (IT1218-19 and KK-2020/00107), and the European Union's Horizon 2020 research and innovation programme under the Marie Sk\u0142odowska-Curie grant agreement No. 823745. The authors also thank the technical and human support provided by SGIker from UPV/EHU and European funding (ERDF and ESF).", "descript": "\n                  The performance of Fe/olivine catalysts was tested in the continuous steam gasification of sawdust in a bench scale plant provided with a fountain confined conical spouted bed reactor at 850\u00b0C. Olivine was used as catalyst support and loaded with 5wt%Fe. The activity and stability of the catalyst was monitored by nitrogen adsorption-desorption, X-ray fluorescence spectroscopy, temperature programmed reduction, X-ray diffraction and X-ray photoelectron spectroscopy techniques, which were conducted before and after the runs. The fountain confined conical spouted bed performs well in the biomass steam gasification with primary catalysts. In fact, this reactor allows enhancing the gas-solid contact, and therefore the catalytic activity by avoiding the elutriation of fine catalyst particles. The uncatalysed efficiency of the gasification process, assessed based on the gas production and composition, H2 production, tar concentration and composition, and carbon conversion efficiency, was consideraby improved on the Fe/olivine catalyst, with tar reduction being especially remarkable (to 10.4gNm\u22123). After 140min on stream, catalyst deactivation was particularly evident, as tar concentration increased to 19.9gNm\u22123 (90% of that without catalyst). However, Fe/olivine catalyst was still active for WGS and CH4 steam reforming reactions, with gas and H2 productions being 1.35Nm3\n                     kg\u22121 and 5.44wt%, respectively. Metal iron oxidation to Fe3O4 caused catalyst deactivation, as the reaction environment shifted from oxidizing to reducing conditions due to operational limitations.\n               "}
{"full_text": "Single-atom catalysts (SACs), created by decorating isolated metal atoms or mononuclear metal complexes on appropriate supports, have received tremendous and continuous attention in chemistry and materials science.\n1\u20139\n In principle, SACs provide 100% metal dispersion on the surface and thus maximize the metal utilization, which is a special and ideal feature for developing efficient and low-cost heterogeneous catalysts, especially significant for the effective utilization of noble metals (e.g., Pt, Pd, Ru, and Ir).\n10\u201312\n Furthermore, the enormous possibilities of agilely pairing the single-atom metal centers with the host materials and the fine control of their local coordination environment give the great potentials to actualize many highly efficient SACs.\n7\n\n,\n\n13\n\n,\n\n14\n As an admitted bridge between homo- and heterogeneous catalysis, SACs render favorable features of homogeneous catalysts, including the uniformity of active sites and the tunable interactions with ligands while inheriting the high durability and excellent recoverability as heterogeneous catalysts.\n15\u201317\n Benefiting from these attributes, SACs have afforded us spectacular opportunities to develop efficient heterogeneous catalysts with competitive activity and selectivity to homogeneous counterparts and, importantly, to understand the fundamental mechanism of heterogeneous catalysis at the atomic and molecular levels.\n5\n\n,\n\n18\n\n,\n\n19\n After a decade of rapid progresses and breakthroughs, SACs have been engaged as promising platforms for chemical transformation and energy-related catalysis, including photocatalytic energy conversion to produce sustainable fuels and chemicals.\n20\u201323\n\nFor the utilization of infinite and freely accessible solar energy, artificial photocatalytic energy conversion provides a promising strategy to overcome the global energy crisis and combat the increasingly erratic climate by reducing greenhouse gas emissions.\n24\n\n,\n\n25\n For example, photocatalytic water splitting is a technologically straightforward and cost-competitive avenue toward sustainable clean H2 fuel production,\n26\u201328\n while photocatalytic CO2 reduction represents a deployable and highly attractive strategy to convert inert CO2 into value-added chemicals and definitively close down the carbon cycle.\n11\n\n,\n\n24\n\n,\n\n29\u201331\n Nevertheless, the performance of the traditional photocatalytic systems, which greatly rely upon the energy band configuration and surface structure of the catalysts, is still far from satisfactory because of the sluggish separation of electron\u2013hole pairs and limited surface-active sites.\n32\n\n,\n\n33\n Despite tremendous efforts, the heterogeneous photocatalysts heretofore still suffer from many deficiencies such as the fast charge carrier recombination and inefficient molecular activation,\n34\u201337\n thus significantly depressing the charge transfer from catalyst surface to reactant species and further transformation of molecules.In such circumstances, SACs have emerged as new options and notions for the design and fabrication of cost-effective and high-efficiency photocatalysts and cocatalysts. Advantageously, the isolated reactive centers in a single-atom photocatalytic system can not only create an increased number of active sites for photocatalytic reactions but also broaden the light-harvesting range and elevate the charge separation/transfer efficiency.\n38\u201340\n The as-constructed single-atom photocatalysts are highly configurable, providing sufficient light traps and fine modification of the surface structure to adsorb and activate molecules.\n33\n\n,\n\n39\n Furthermore, the structural simplicity of single-atom photocatalysts allows us to draw more precise structure-performance correlations, enabling the improved understanding of the fundamental mechanism of photocatalysis and thus promoting the rational design of targeted photocatalysts. Albeit research efforts on SACs applications in photocatalytic systems are still at their nascent stage, these advantages have made them up-and-coming candidates to facilitate the photocatalytic reactions.\n30\n\n,\n\n35\n\n,\n\n41\u201343\n In order to promote such a fledgling yet rapidly evolving field, timely overviews on the recent progress of single-atom photocatalysis are highly desirable to not only reveal the primary working mechanism but also inspire future research directions. As far as we know, although some excellent reviews with the emphasis on the background and synthetic strategies of SACs, as well as their unique advantages for photocatalysis, have emerged in the literature, the fundamental principles that are highly relevant to the energy band engineering and energy transfer route in single-atom photocatalysis have not been covered in the reported review articles.\n33\n\n,\n\n39\u201341\n\nIn this review article, we herein focus on extracting the key principles from literature for the single-atom photocatalysis to understand the working mechanism comprehensively and thus promote the rational design and fabrication of more efficient single-atom photocatalysts. The review begins with a short presentation on the achievements and features of the photocatalytic application of SACs. Then, the synthetic strategies and related structural characterization methods for confirming the successful construction of single-atom photocatalysts have been concisely overviewed. More importantly, as shown in Figure\u00a01\n, we dedicatedly disclose the mechanism for the surface charge separation/transfer accelerated by isolated metal centers and also the adsorption and activation of molecules in single-atom photocatalysis by illustrating some representative examples. The applications of SACs for well known, and emerging photocatalytic reactions are also introduced with the most recent developments. Finally, we present some challenges and perspectives for the future development direction of SACs in photocatalytic energy conversion. It is highly expected that this review would deliver some new insights toward the understanding and engineering of SACs in photocatalysis and further accelerate the development of this important emerging research area.An irrefragable witness of the ever-growing interest in the topic of single-atom photocatalysis could be expressed by the exponentially increased number of publications and related citations from 2014 to 2020 and its sustained growth trend by 2021 (Figure\u00a02\nA). During the 7-year research period, the potential of SACs has been explored in plenty of photocatalytic applications, especially in sustainable energy conversion and small-molecule activation (Figure\u00a02B). Figure\u00a03\n depicts the timeline for the key achievements and features related to single-atom photocatalysis. In 2014, Yang and coworkers have reported the first use of an SAC (0.2 Pt/TiO2) for photocatalytic hydrogen evolution, which exhibits better performance than those of metallic nanoparticles or clusters.\n44\n 2 years later, Zhang and coworkers have demonstrated that the anchoring of single Co atoms in the metal-organic framework (MOF) could remarkably promote the photoexcited carrier separation in the MOF-525-Co photocatalyst and enhance the oriented migration of photoexcited electrons from support to metal centers, which also represents the first report for the use of SACs in the highly efficient photocatalytic CO2 reduction.\n45\n In 2016, Wu and coworkers have directly evidenced that a longer lifetime of photoexcited electrons can be provided from isolated Pt sites decorated g-C3N4 to improve the performance of photocatalytic hydrogen evolution significantly.\n46\n In 2017, Wei and coworkers have further discovered that the generation of mid-gap states in an isolated Co1-P4 site anchored g-C3N4 photocatalyst can increase the light-harvesting ability and provide separation centers to depress charge recombination and elongate carrier lifetime.\n47\n These pioneering achievements have greatly promoted and triggered the booming development of single-atom photocatalysis.Subsequently, researchers have demonstrated the bridging function of single-atom photocatalysts among heterogeneous, homogeneous, and even enzymatic photocatalysis for the photoactivation of molecules, for example, through the well-designed study of Co1-G\n48\n and Cu/TiO2\n\n49\n photocatalysts for the activation of CO2 and H2O, respectively. Meanwhile, a plenty of new strategies have also been discovered to expand the family members and related applications of single-atom photocatalysts, such as single-tungsten-atom-oxide (STAO) for photocatalytic pollutant degradation (2019)\n50\n and plasmonic Cu nanoparticle supported single-atom Ru for photothermal methane dry reforming (2020).\n51\n Furthermore, in 2020, Lou and coworkers have investigated the dynamic changes in chemical valence and coordination environment of isolated metal centers in their multi-edged TiO2 supported single-atom Ru photocatalyst through in situ extended X-ray absorption fine structure (EXAFS) technique.\n52\n This work opens new avenues toward the observation\u00a0and interpretation of the intrinsic working mechanism for single-atom photocatalysis. Based on these great achievements in the past seven years, single-atom photocatalysts have been proved as excellent candidates for various energy conversion processes with high activity and selectivity (Table 1\n), including H2 evolution,\n44\n\n,\n\n71\u201373\n CO2 reduction,\n45\n\n,\n\n62\n\n,\n\n74\u201376\n N2 reduction,\n63\n\n,\n\n77\n\n,\n\n78\n C\u2013O coupling,\n65\n pollutant degradation,\n79\n\n,\n\n80\n CH4 conversion,\n68\n H2O2 production,\n69\n as well as photocatalytic sensing.\n70\n\n,\n\n81\n\n,\n\n82\n\nAs various synthetic strategies for SACs have been well summarized in other excellent reviews,\n7\n\n,\n\n14\n\n,\n\n83\n we here mainly focus on the preparation of semiconductive supports-based SACs for photocatalysis. Due to the high surface energy of the isolated atoms in SACs, the active sites in SACs are rather vulnerable to relocating and aggregating into clusters and nanoparticles, which brings additional difficulties in stabilizing the active sites in high loading contents. Some innovative strategies for fabricating single-atom photocatalysts to combat these challenges have been briefly introduced.To realize the homogeneous dispersion of isolated reactive centers on the semiconductive supports, creating sufficient distance and excellent distribution among every species of the metal precursor is an easy and efficient solution, which prevents the agglomeration of metal sites after post-treatment to form single-atom photocatalysts with a specific geometric environment. Moreover, surface modification with functionalities for the semiconductive supports before integrating with the metal precursor is quite necessary, providing adequate anchoring sites for stabilizing the single atoms. Based on these progresses, we have summarized some detailed approaches for the preparation of single-atom photocatalysts.Wet-chemical synthesis has been recognized as a facile and repeatable strategy for the synthesis of single-atom photocatalysts with well-defined mono-dispersity and deployable scale-up. In a typical process, the metal precursor (e.g., metal salt or mononuclear metal complex) is added to the semiconductive supports, followed by the reduction or activation steps to realize the decoration of isolated metal sites onto the semiconductor. For example, Chen et\u00a0al. have decorated isolated Pt atoms (loading: 0.02 wt %) onto the defect-rich TiO2 support via the wet-deposition approach (Figure\u00a04\nA).\n53\n The Pt precursor (H2PtCl6) has been first added to the suspension containing sodium titanate nanotubes, followed by the calcination of the as-prepared mixture at 400\u00b0C to obtain the Pt single-atom photocatalyst. A plenty of transition-metal-based single photocatalysts (e.g., Pt,\n46\n Pd,\n87\n Ru,\n52\n Cu,\n61\n Ni,\n65\n and Co\n88\n) have also been synthesized via the nearly identical strategy. However, the loading contents for the isolated metal atoms are commonly lower than 2 wt %, which should be attributed to the limited control of the distribution of active sites on the support surface, especially under the effect of capillary forces during the drying process.\n7\n\nThe pyrolysis method is an essential strategy to obtain single-atom photocatalysts via the pyrolysis of the mixed precursors of metal and organic semiconductors (e.g., g-C3N4\n\n89\n; MOFs\n90\n) at relatively high temperatures. In general, a strong coordination bond between the metal sites and the organic precursor is required to prevent isolated metal sites from agglomeration during pyrolysis. For example, Jiang et\u00a0al. have introduced Ga (0.014 \u223c 0.045 wt %) into g-C3N4 by direct pyrolysis of the mixture of GaCl3 and urea at 550\u00b0C (Figure\u00a04B).\n84\n Through pyrolysis of the supramolecular precursor assembled by melamine, cyanuric acid, and silver citrate, Zou et\u00a0al. have synthesized a Ag-N2C2/g-C3N4 photocatalyst with a high content of isolated Ag metal (3.7 wt %).\n56\n\nCoordination confinement has been believed as another effective approach for constructing single-atom photocatalysts in which the encapsulation of suitable mononuclear metal precursors with semiconductive supports can thus be realized. Due to the strong interaction between mononuclear metal and anchor sites from the porous supports, the agglomeration of isolated metal sites is largely avoided. Porous MOFs with suitable anchor sites have been developed as ideal semiconductive supports for the synthesis of single-atom photocatalysts via the coordination confinement strategy. Zhang et\u00a0al. have directly incorporated coordinatively unsaturated Co atoms into a porphyrin-based MOF (MOF-525) to prepare a MOF-525-Co single-atom photocatalyst with the Co loading content of 6.01 wt % (Figure\u00a04C).\n45\n Wang et\u00a0al. have further stabilized a series of isolated metal atoms (i.e., Ir, 1.41 wt %; Pt, 2.74 wt %; Ru, 1.92 wt %; Au, 1.18 wt %; Pd, 3.68 wt %) on zirconium-porphyrinic MOF hollow nanotubes by the coordination between porphyrin units and metal atoms.\n91\n Moreover, Zuo et\u00a0al. have achieved an ultrahigh loading content (12.0 wt %) of isolated Pt atoms on the ultrathin MOF nanosheets by assembling the linker of PtII tetrakis(4-carboxyphenyl)porphyrin (PtTCPP) and the metal nodes of Cu2(COO)4 paddle-wheel clusters (Figure\u00a04D).\n57\n\nApart from the above-mentioned approaches, some other specific avenues have also been developed to synthesize the single-atom photocatalysts. Atomic layer deposition (ALD) could be a controllable approach to synthesize single-atom photocatalysts via the layer-by-layer deposition of isolated metal sites onto the semiconductors. Cao et\u00a0al. have decorated isolated Co metal sites (Co1-N4; loading: 1.0 wt %) onto the g-C3N4 by an ALD method. By using bis(cyclopentadienyl)cobalt as the ALD reagent and further removing the surface cyclopentadienyl ligand through O3 treatment, isolated Co atoms decorated g-C3N4 has been thus developed (Figure\u00a04E).\n85\n By providing sufficient ionized anions and cations with a strong polarizing force for breaking the covalent, ionic, or metallic bonds in the liquid environment of melt salts, the molten-salt method (MSM) could be another powerful strategy for preparing single-atom photocatalysts. Xiao et\u00a0al. have realized the decoration of isolated Ni atoms on the TiO2 photocatalyst through a controllable MSM (Figure\u00a04F).\n86\n The molten salts (a mixture of LiCl and KCl) have provided liquid conditions and space confinement for the isolated dispersion of Ni species and promoted the generation of strong Ni\u2013O bonds on TiO2. Furthermore, chemical etching can also be used to construct single-atom photocatalysts with relatively high metal loading. Zhang et\u00a0al. have further provided an etching method to decorate isolated Ni atoms on the defect-rich zirconia support (Figure\u00a04G). They have prepared Ni-Zr hydroxide from the Ni-Zr sol-gel. Subsequently, the porous Ni-Zr hydroxide has been calcinated and etched with dilute hydrochloric acid to obtain the single-atom photocatalysts with the nickel content varied from 2 to 8 wt %.\n59\n\nIn addition to those main traditional characterizations for elucidating the physical properties (e.g, optical absorption and crystal properties), chemical composition, and band structures for photocatalysts, the dispersion of isolated reactive centers in single-atom photocatalysts requires a whole toolbox of complementary techniques.\n32\n\n,\n\n92\n Here, we mainly focus on the overview of structural characterizations of isolated reactive centers in single-atom photocatalysts by advanced techniques in electron microscopy and spectroscopy.Transmission electron microscopy (TEM) in atomic resolution has been developed as an effective approach for studying the detailed structural information of the isolated reactive centers, as well as their interactions with the supports. In particular, by using aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (AC-HAADF-STEM), we can easily confirm the existence of isolated reactive centers over the supports, provided the metal atom of interest shows a much higher atomic number than the support element. In addition, the attached energy-dispersive X-ray (EDX) detector of STEM could further provide the elemental mapping elucidating the atomic dispersion of the metal atoms on the semiconductive supports.\n49\n\n,\n\n53\n\nAlthough electron microscopy images provide us with effective information for identifying the structural information of the catalyst, we also need to notice that there are some limitations for its applications in structural characterizations. Electron microscopy can only image the structure in the local area and fail to provide overall structural information for the whole catalyst. Furthermore, due to the limited electron penetrability of microscopic techniques, isolated metal atoms decorated in the bulk phase or cavities instead of the surface structure are difficult to be observed.\n28\n Consequently, some additional spectroscopic methods are necessary to provide complementary data and underpin the presence of isolated metal sites in single-atom photocatalysts.X-ray absorption spectroscopy (XAS) is one of the most heavily used and powerful tools to characterize single-atom photocatalysts, including X-ray absorption near edge structure (XANES) spectroscopy and EXAFS. The XANES can deliver the local electronic state information for the probed elements, while the EXAFS provides notable details on the coordination environment and local geometric structure of isolated metal sites in high resolution. Combined with the density functional theory (DFT) simulations, well-defined structures of single-atom photocatalysts could be assigned through XAS observations. Cao et\u00a0al. have forcefully demonstrated the formation of \u201cCo-N4\u201d in their isolated Co decorated g-C3N4 catalyst.\n85\n The relaxed bond lengths determined by the fitted EXAFS spectrum are extraordinarily close to those calculated by DFT from geometry optimization of simulated \u201cCo1-N4\u201d model in a 2\u00a0\u00d7 2\u00a0\u00d7 1 supercell of g-C3N4. As a site-specific characterization technique, Fourier-transform infrared (FTIR) spectroscopy has also been widely applied to evaluate the existence of isolated metal sites according to their significant shift compared with clusters or nanoparticles.\n38\n\n,\n\n93\n X-ray photoelectron spectroscopy (XPS) is another widely used technique to reveal the surface valence structure of single-atom photocatalysts. A distinct shift in binding energy compared with pure metal references may elucidate the oxidation state of isolated metal atoms and exclude the existence of nanoparticles.\n7\n\n,\n\n94\n Furthermore, M\u00f6ssbauer spectroscopy could become another forceful tool to analyze the chemical state and coordination environment of M\u00f6ssbauer-active elements (e.g., Fe and Sn). Noteworthily, Wu et\u00a0al. have validated the existence of Fe(IV) species in a single-atom Fe-modified TiO2-SiO2 photocatalyst by employing the M\u00f6ssbauer spectroscopy.\n77\n\nThe initial step in the photocatalytic process is to harvest the incident photons with a specific frequency (h\u03bd) by the component of the semiconductor for the generation of photoexcited electron (e\u2212) and hole (h+).\n95\n The energy band structures of single-atom photocatalysts, including the bandgap size and positions of the conduction band (CB) and valence band (VB), are the dominant elements to tune the light-harvesting ability and drive the redox reactions.\n32\n The introduction of isolated metal atoms could not only adjust the band structure to enhance the light absorption capability of semiconductive supports but also provide electron pumps to boost the transfer of photoexcited electrons, thus highly accelerating the surface charge separation/transfer in the single-atom photocatalytic systems.To fully utilize solar energy, it is of great importance to narrow down the required bandgap of photocatalysts for the generation of electron\u2013hole pairs under visible-light excitation. Furthermore, the CB minimum (CBM) and VB maximum (VBM) positions of photocatalysts should well match the redox potentials of specific reactions to trigger the proceeding of the photocatalytic process.\n32\n\n,\n\n39\n Experimental and theoretical results have clearly revealed that the introduction of isolated metal sites could significantly modulate the energy band configuration of semiconductive host materials to highlight the light absorption ability and adjust their redox potentials.In principle, when the loading contents of isolated metal sites are in trace amount, limited effect on the inherent band structure of host semiconductors will emerge while some new electronic states are inclined to be generated to boost the light-harvesting ability of the entire photocatalyst (Figure\u00a05\nA). Xiong and coworkers have confirmed that the band structure of the isolated Pt (Pt loading: 0.18 wt %) decorated g-C3N4 photocatalyst is rather similar to that of pristine g-C3N4 by the Mott-Schottky plots measurement and ultraviolet photoelectron spectroscopy investigation (Figure\u00a05B).\n38\n They have further demonstrated the generation of new hybrid states over the surface of the catalyst due to the metal-to-ligand charge transfer (MLCT) between the isolated atoms and the host organic semiconductor. The first-principle simulations further confirm the up-shift of the lowest unoccupied molecular orbital (LUMO) and the highest occupied molecular orbital (HOMO) positions in MLCT-induced hybrid state (Figure\u00a05C). Under the simultaneous contributions of HOMO-to-LUMO transition in local g-C3N4 and MLCT in g-C3N4-Pt units, the single-atom photocatalyst exhibits enhanced broad-spectrum light-harvesting ability. This tentative conclusion has been further verified by other reports combined with the DFT calculations.\n39\n\n,\n\n87\n\n,\n\n96\n\n,\n\n97\n For example, Lou and coworkers have confirmed that the isolated Ru decoration can lead to the generation of new energetic states corresponding to the 4d orbitals of Ru near the Fermi levels of TiO2 (Figure\u00a05D), which significantly improves the light-harvesting capability of the single-atom photocatalyst under visible-light irradiation.\n52\n\nWith the increase of the doped metal contents to a relatively high amount, the band structure configuration of host semiconductor materials can be significantly changed (Figure\u00a05A). For instance, Zhang and coworkers have reported that the inclusion of 2 \u223c 5 wt % of isolated Ni atoms in ZrO2 photocatalyst would directly narrow the bandgap of the as-synthesized Ni-SA-x/ZrO2 photocatalyst by an up-shifted VB level and a down-shifted CB level as demonstrated by both the experimental results and theoretical calculations (Figure\u00a05E).\n59\n As revealed by diffuse reflectance spectroscopy, the Ni-SA-x/ZrO2 photocatalyst herein presents additional light absorption peaks in the range from 400 to 700\u00a0nm compared with the pristine ZrO2 (p-ZrO2) sample (Figure\u00a05F). Furthermore, the modified CB potential of Ni-SA-5/ZrO2 highly satisfies the potential of CO2 reduction into CO reaction, making it an excellent SAC for selective CO2 photoreduction. This phenomenon has also been demonstrated by Wang and coworkers.\n58\n They have validated that the bandgap of zincblende cadmium-zinc sulfide quantum dots (ZCS QDs) could be gradually narrowed by the increase of Ni species in the Ni\nx\nZCS QDs photocatalyst. Both CBM and VBM positions of the single-atom photocatalyst are tailored significantly by the decoration of Ni atoms (Figure\u00a05G), which triggers excellent charge separation efficiency and photocatalytic activity for the hydrogen evolution. The modification of band structures for semiconductors by the decoration of isolated metal sites in high concentration is similar to those of traditional doping avenues with 3d-transition elements or nonmetal elements\n32\n; however, the type of metal has limited influence on the modification, possibly due to the atomically isolated distribution of metals in single-atom photocatalysts.\n98\n\nTo date, the mechanism of isolated metal sites decoration on modulating the band alignments of host semiconductor materials is increasingly evident but still exists contentions. Therefore, further experimental and theoretical investigations are highly demanded to reveal such a role of isolated metal sites, thus providing the guidance of rational design of single-atom photocatalysts for targeted reactions.The photogenerated electrons transfer from VB to unoccupied CB in the photo-excitation process, while the equivalent holes stay in the VB. The following migration of photoexcited carriers from the photocatalyst surface to the reactant molecules is crucial for redox reactions. Nevertheless, a fraction of electron\u2013hole pairs recombines on the surface or in volume with the release of light or heat.\n27\n In this case, decorating the isolated metal atoms on the semiconductor supports could serve as electron pumps to accelerate the photogenerated electron\u2013hole transfer, thus greatly speeding up the whole reaction process.Note that the Schottky barrier in a traditional particulate cocatalyst/semiconductor system can perform as an electron trap for photoreactions and prevent the photoexcited electrons from traveling back to the host semiconductor from the cocatalyst; however, it also blocks the extraction of photoexcited electrons and lowers down the efficiency of electron transfer and injection.\n99\n Distinctively, the covalent coordination between the isolated metal sites and host semiconductor materials makes the single-atom photocatalyst Schottky-barrier free, enabling the electron transfer to be unrestricted and more effective.\n100\n Benefiting from the modern characterization techniques, several electron pump models and local electron trap states induced by isolated metal sites have been found to reveal the charge separation/transfer manner in single-atom photocatalysts.For instance, Wu and coworkers have confirmed that the newly generated electron trap states near the CBM of g-C3N4 induced by MLCT between isolated Pt atoms and g-C3N4 could be the electron transfer and injection channels (Figure\u00a06\nA).\n46\n The isolated Pt sites coordinated with C/N atoms on g-C3N4 can function as electron pumps to elevate the charge separation/transfer efficiency (Figure\u00a06B), resulting in longer-lived photogenerated electrons than those of Pt nanoparticles decorated g-C3N4, as revealed by an ultrafast transient absorption spectroscopy (Figure\u00a06C). The as-synthesized catalysts thus provide more photoexcited electrons to engage in the reduction of H+. This typical electron pump model and unique MLCT-induced electron trap state have been demonstrated as general cases for the single-atom photocatalysts.\n57\n\n,\n\n93\n\n,\n\n96\n\n,\n\n102\u2013104\n Furthermore, Zhang and coworkers have demonstrated that there would be an MLCT-induced electron trap state and an additional vacancy-related state in the single-atom photocatalyst with a defect-rich host semiconductor (Figure\u00a06D).\n100\n These newly generated electron states could collectively and significantly promote the electron trapping capability and extend the lifetime of photogenerated electrons according to the results of photoluminescence emission spectra (Figures 6E and 6F). Moreover, the decoration of isolated metal sites could also form mid-gap states on the host semiconductor to promote charge separation and transfer.\n39\n\n,\n\n105\n\n,\n\n106\n For example, Wei and coworkers have verified the generation of mid-gap states by restricting isolated Co1-P4 sites on g-C3N4 (Figures 6G and 6H).\n47\n The presence of mid-gap states below the CB position improves the light-harvesting capability of the single-atom photocatalyst, which also provides charge separation centers to suppress the electron\u2013hole recombination and prolong the photocarrier lifetimes, as demonstrated by transient open-circuit voltage decay measurements (Figure\u00a06I).Moreover, the strength and efficiency of electron pumps can be further enhanced via the decoration of isolated metal sites onto the dyadic semiconductor composites such as the Z-scheme photocatalyst.\n101\n\n,\n\n107\n Recently, Peng and coworkers have demonstrated the first example of isolated PtN4 sites on a Z-scheme photocatalyst composed of Zn-/Pt-porphyrin conjugated polymer (ZnPtP-CP) and BiVO4 semiconductors (Figures 6J\u20136L).\n101\n The isolated Pt metal sites can perform as the strengthened electron pumps to trap more photogenerated electrons migration from Z-scheme of the ZnPtP-CP/BiVO4 composite and further transfer them to the reactant of H+. Femtosecond time-resolved fluorescence spectra further reveal that the charge separation efficiency is significantly enhanced over this kind of single-atom-metal-Z-scheme composite photocatalyst (Figure\u00a06L).Upon the construction of enhanced electron pumps induced by isolated metal sites, the photogenerated electron\u2013hole could be efficiently separated and transferred, also significantly promoting photocatalytic oxidation reactions (e.g., pollutant degradation and oxidation of organic compounds). For example, Yang et\u00a0al. have implanted isolated Co atoms in the polymeric carbon nitride (pCN) for the efficient oxytetracycline degradation under visible light.\n108\n The introduction of isolated Co sites has significantly extended the light absorption area and electron density for the pCN, thus accelerating the charge separation/transfer and the degradation of oxytetracycline. The generation of reactive species (i.e., h+, 1O2, \u22c5O2\n\u2212, and \u22c5OH) facilitated by isolated Co sites triggers the catalyst to exhibit an apparent rate of 0.038\u00a0min\u22121 for the oxytetracycline degradation, outpacing the pristine pCN by about 3.7 times. Xiao et\u00a0al. have further reported a promoted charge separation/transfer system for photocatalytic hydrogen evolution and benzene oxidation by introducing isolated Cu atoms to the C3N4 layers.\n109\n The isolated Cu sites coordinated with the N atoms in C3N4 have constructed efficient Cu-N\nx\n charge-transfer channels to promote in-plane and interlayer transfers of photoexcited electrons and holes. Consequently, the as-formed single-atom photocatalyst exhibits excellent performance for the hydrogen evolution as well as the oxidation of benzene into phenol with a conversion rate of 92.3% and the selectivity of 99.9% under visible light.It is evident that the decoration of isolated metal sites could improve the light-harvesting ability and accelerate the charge separation/transfer for photocatalytic systems, which offers us a fascinating strategy to elevate the utilization efficiency of photogenerated carriers. Thereafter, adsorption and activation of molecules over the single-atom photocatalyst hold the keys for driving the entire photocatalytic performance. Advantageously, the coordinatively isolated metal atoms have provided adequate unsaturated active sites on the catalyst for photocatalytic reactions.According to the Sabatier principle, the ideal catalytic site should have moderate bonding strength to the key intermediate in the reaction, which is not only strong to activate the intermediate but also facilitates the desorption of the product.\n110\n\n,\n\n111\n This principle is essentially related to the electronic structure of the catalysts.\n112\n In a single-atom photocatalyst, control of unsaturated active sites with diverse coordination environments to modulate the electronic structure supplies numerous opportunities to tune the adsorption energy of molecules or intermediates onto the catalyst surface. Combined with the progress of DFT calculations, researchers could further forecast the catalytic activity of particular active sites and thereby realize the theory-guided synthesis of efficient single-atom photocatalysts.As a typical example, Ma and coworkers have quantitatively assessed the adsorption energy of hydrogen over isolated Pt sites supported on carbon nitride with six coordination environments by following the DFT calculation (Figures 7A and 7B).\n113\n The theoretical calculations directly indicate that the coordination environments of unsaturated Pt atoms can modulate the local electronic configuration of the catalyst and thus influence the adsorption energy of hydrogen on the catalyst surface. DFT simulations further demonstrate that the hydrogen can be absorbed by the unsaturated Pt active sites and neighboring C atoms (Figure\u00a07C). As the Pt-C4 moiety exhibits moderate bonding strength between hydrogen and unsaturated Pt sites, the as-constructed Pt-C4 cocatalyst doped CuS photocatalyst has well-balanced active sites for the proton adsorption and final desorption of H2, thus significantly upgrading the performance of photocatalytic H2 production from\u00a0water (Figure\u00a07D). This work also emphasizes the fact that the modulation of the coordination environment is crucial for stabilizing unsaturated metal sites and tuning the adsorption energy to achieve the desired photocatalytic activity.Moreover, the coordination environment modulation of the unsaturated active sites also significantly affects the electron density of the entire catalyst surface to improve the adsorption capacity of single-atom photocatalysts. Zhou and coworkers have demonstrated that adjusting the coordination environment of isolated Au atoms through changing the vacancy types in CdS could enhance the surface electron density of Au-anchored CdS photocatalyst (Figure\u00a07E).\n114\n Higher electron density can be achieved on the surface of as-constructed Au/Cd1-x\nS photocatalyst because of the electron accumulation at Cd vacancies instead of unsaturated Au active sites. The bonding of CO2 molecules on the catalyst surface thus converts from physical adsorption into chemical adsorption, which significantly promotes the following photocatalytic CO2 reduction.Since the adsorption process of molecules or intermediates onto the catalyst surface is crucial for the single-atom photocatalysis, researchers have begun to evidence the enhancement in adsorption capacity induced by unsaturated active sites through the related adsorption-desorption experiments. Ozin and coworkers have investigated the effect of unsaturated Bi3+ active sites substitution on the adsorption capacity of In2O3 photocatalyst for CO2 molecules via the CO2 temperature-programmed desorption (TPD) analysis and in situ diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS).\n115\n CO2-TPD clearly reveals that the typical desorption peaks for CO2 and other related surface species move to higher temperatures after introducing unsaturated Bi3+ sites (Figure\u00a07F), indicating stronger binding of these surface species over the Bi3+-anchored In2O3 photocatalysts (Bi\nx\nIn\n2\u2013x\nO3). In situ DRIFTS has been further applied to observe the transition evolvement of surface species during the adsorption of CO2 molecules over Bi\nx\nIn\n2\u2013x\nO3 and pure In2O3 (Figure\u00a07G). The enhanced peak intensities for all surface species over Bi\nx\nIn\n2\u2013x\nO3 compared with those over In2O3 imply the empowered adsorption-bonding-activating capacity for CO2 molecules after incorporating the unsaturated Bi active sites. Strikingly, these pieces of experimental evidence further confirm that the unsaturated active sites can serve as efficient promoters to elevate the adsorption capacity for molecules or intermediates over the single-atom photocatalysts.Upon the moderate adsorption of molecules onto the surface of the single-atom photocatalyst, the activation of reactant molecules over the unsaturated active sites by using the photogenerated carriers is the ultimate goal for the entire photocatalytic system. For the traditional heterogeneous photocatalytic system, the molecular activation mechanism is rather complicated due to the diversity of active centers and varied reaction paths.\n99\n Single-atom photocatalysis has provided us with a streamlined model to investigate the photoactivation process for reactant molecules, attributed to the structural simplicity of active centers.\n116\n That is to say, the initial activation process can be traced in the whole single-atom photocatalytic process, including photoexcited carriers separation/transfer, molecular adsorption, intermediates formation, and the final product desorption (Figure\u00a08\nA). On the basis of the above-mentioned advantages of emerging single-atom photocatalysts in promoting the key principles of the photocatalytic process, researchers have drawn several convincing photoactivation cycles, thus evidently proving the high activation capacity of unsaturated active sites for reactant molecules.For instance, Sun and coworkers have proposed a compelling photoactivation process for hydrogen evolution from water over the unsaturated GaN4 active sites supported on carbon nitride (Figure\u00a08B).\n84\n As an efficient electron pump, GaN4 active sites capture and store the photogenerated electrons under light illumination, resulting in the negatively charged GaN4 sites preferring to absorb the positive protons. The absorbed protons further accept electrons from the charged GaN4 site and form the H\u2217, while the GaN4 sites return back to the ground state after releasing the captured electron. H\u2217 tends to desorb from the GaN4 site and generate H2 with other H\u2217 in the system. Upon the generation and release of the H2 product, the free GaN4 active site moves to the next reaction cycle (Figure\u00a08B). For a more complex photocatalytic system like reduction of CO2 into CO, Ye and coworkers have also revealed the photoactivation cycle over unsaturated Co active sites decorated in covalent organic frameworks (sp2c-COFsdpy) by combining the in situ DRIFTS measurements and DFT calculations (Figures 8C\u20138E).\n117\n\nIn situ DRIFTS study confirms the formation of absorbed CO2 (CO2\u2217) and COOH\u2217 as the surface species during the CO2 photoreduction (Figure\u00a08C). DFT calculations directly reveal that the unsaturated Co active sites significantly decrease the Gibbs free energy of the rate-limiting reduction step of CO2\u2217 into COOH\u2217 to activate the CO2 molecules while depressing the process of hydrogen evolution to elevate the selectivity of CO2 photoreduction (Figure\u00a08D). In the photoactivation cycle of CO2 reduction, the unsaturated Co active sites undergo the circulation between initial Co(II) and Co(I) to transport the photogenerated electron flow, thus upholding the selective reduction of CO2 into CO product through a proton-electron coupling path (Figure\u00a08E). These reports demonstrate the structural advantages of unsaturated active sites in delivering the photoexcited carriers for the photoactivation of molecules.It is worth pointing out that the local coordination environment and chemical state of the unsaturated active sites would be dynamically changed in the photoactivation cycle. The related study can gain deep insight for us to access the intrinsic activation mechanism of single-atom photocatalysis. Hyeon and coworkers have evidenced that the valence state of unsaturated Cu active sites on TiO2 is changed after accepting the photoexcited electrons during the photoactivation process, which also triggers the activation of TiO2 support.\n49\n They have herein identified a reversible and cooperative photoactivation cycle at the atomic level for hydrogen evolution from water over the Cu/TiO2 photocatalyst (Figure\u00a08F). That is, under the light excitation, the Cu/TiO2 is transferred from the initial resting state (CT0) to a photoexcited state (CT1) following by the generation of electrons and holes. The flow of photoexcited electrons from the CB of TiO2 to the d orbital of Cu results in the valence change of the Cu active site (CT2). The electron localization at Cu further triggers a polarization field, causing lattice distortion in adjacent TiO2 (CT3). CT3 as an active state manifests enhanced photocatalytic activity for H2 generation. Furthermore, the exposure of CT3 to O2 can revert it to the initial CT0 state and thus complete the photoactivation cycle. Such a systematic work further implies that the activation of water molecules over unsaturated active sites in single-atom photocatalysts could share the fundamental principles resembling enzymes and homogeneous catalysts. Moreover, Reisner and coworkers have recently observed that the reactivity for organic C\u2013O coupling reaction over isolated Ni sites deposited on mesoporous carbon nitride (mpg-CN\nx\n) shows a similar trend to a homogeneous catalytic system (Figure\u00a08G).\n65\n The unsaturated Ni2+ active sites are converted into NiI species after receiving photogenerated electrons from mpg-CN\nx\n, thus providing coupling reaction centers in an activated state for the whole photoactivation cycle.Artificial photocatalytic hydrogen evolution is a feasible way to convert the infinite sunlight into the green and clean energy carrier of H2, providing a sustainable solution to overcome the global energy dilemmas and environmental problems.\n26\n\n,\n\n27\n\n,\n\n52\n As one of the \u201choly grail\u201d reactions, photocatalytic water splitting has great potential to achieve economically practicable and scalable solar H2 production.\n26\n Although plenty of efforts are being dedicated to realizing excellent performance for the H2 evolution from water splitting, the desirable solar-to-hydrogen efficiency for viable application still suffers from the fast recombination of photogenerated electron\u2013hole pairs and the inadequacy of surface-active sites.\n37\n\n,\n\n118\n To overcome these issues, noble-metal materials, like Pt nanoparticles, have been used as cocatalysts on host semiconductor photocatalysts to accelerate the charge separation/transfer and create sufficient active sites for the reaction, thus remarkably improving the photocatalytic performance.\n27\n\n,\n\n95\n However, the low atomic efficiency, high cost, and scarcity still hinder the practical use of noble-metal-particle-based cocatalysts. In this case, the development of single-atom photocatalysts could reduce the utilization of noble metal and offer new possibilities for enhancing the performance of photocatalytic water splitting. Statistical results for the published articles clearly demonstrate that \u223c40% of single-atom photocatalysts have been applied in hydrogen evolution (Figure\u00a02). Among them, host semiconducting materials such as TiO2 and g-C3N4 are most widely used as the supports, showing impressive improvement for the photocatalytic activity of hydrogen evolution.Since the discovery of water splitting on a TiO2 photoelectrode by Fujishima and Honda in 1972, semiconducting TiO2 has become the commonly used photocatalyst and still received continuous attention.\n119\u2013121\n Yang and coworkers have first decorated the isolated Pt atoms onto TiO2 for H2 production from water splitting.\n44\n The isolated Pt atom-based TiO2 exhibits reduced H\u2217 adsorption energy and boosted active sites, thus giving better performance for photocatalytic H2 evolution than the Pt nanoparticle-based TiO2 photocatalyst. Subsequently, Liu and coworkers have further validated that the selective decoration of isolated Pt sites on (101) facet of TiO2 could achieve higher activity and more considerable stability for photocatalytic hydrogen evolution compared with the traditional Pt nanoparticle- or cluster-decorated TiO2 photocatalysts.\n122\n\nThe defects and morphology structure of TiO2 have been finely engineered to enhance the intrinsic photocatalytic activity of TiO2-based single-atom photocatalysts.\n90\n\n,\n\n104\n\n,\n\n123\u2013125\n Li and coworkers have fabricated a high-performance photocatalyst by assembling isolated Pt sites on a defect-rich anatase TiO2 support (Pt1/def-TiO2). A Pt\u2013O\u2013Ti3+ atomic interface is constructed after the decoration of isolated Pt atoms, which significantly promotes the charge separation/transfer and suppresses the recombination of electron\u2013hole pairs.\n53\n The unique Pt1/def-TiO2 delivers record-high photocatalytic H2 evolution activity with an extremely high turnover frequency (TOF) of 51,423 h\u22121, outpacing the traditional particle-based Pt/TiO2 catalyst by 591 times. Schmuki and coworkers have further demonstrated that the high density of Ti3+-Ov defects on TiO2 nanotubes favors the stabilization of the isolated Ir atoms.\n126\n The resulting isolated Ir sites decorated TiO2 photocatalyst exhibits higher activity for H2 evolution than single atoms supported on other nanostructures in their system. Lou and coworkers have further decorated isolated Ru sites over multi-edged TiO2 spheres (ME-TiO2@Ru) for photocatalytic H2 production (Figures 9A\u20139C).\n52\n The sharp edges of TiO2 (Figure\u00a09B), as revealed by STEM observations, have been confirmed to be conducive for the migration of photoexcited electrons from TiO2 to Ru sites. Consequently, the ME-TiO2@Ru shows a highly enhanced hydrogen evolution rate (Figure\u00a09C), exceeding the TiO2 nanoparticles stabilized isolated Ru catalyst (TiO2@Ru) by 2.2 times. This work further highlights the importance of controlling the morphology of host semiconductor supports for improving the performance of single-atom photocatalysts.Moreover, non-noble transition metal atoms have also been introduced to fabricate TiO2-based single-atom photocatalysts,\n86\n\n,\n\n127\u2013129\n which not only reduce the utilization of noble metals but also break the ultimate activity limit of noble-metal atom sites. Li and coworkers have successfully anchored isolated Co and Pt atoms on the surface of TiO2, which contains 13.4% oxygen-coordinated Co-O-Pt dimers (Figure\u00a09D).\n54\n The dual-single-atom photocatalyst exhibits much better performance for H2 evolution than Pt single-atom and cluster-based photocatalysts (Figure\u00a09E). They have confirmed that the Co-O-Pt dimer coupling enables the mutual optimization of electronic structure for Pt and Co centers to weaken the binding of H\u2217, which is the prime factor for breaking the ultimate activity ceiling of the Pt atom.As a representative organic semiconductor, two-dimensional g-C3N4 has also been used as an excellent host support for fabricating SACs, attributing to its abundant surface trapping sites for isolated atoms. Many types of isolated metal sites, including noble metal (e.g., Pt,\n38\n\n,\n\n46\n\n,\n\n94\n\n,\n\n130\u2013134\n Pd,\n55\n\n,\n\n87\n\n,\n\n102\n Rh,\n135\n Au,\n136\n and Ag\n56\n\n,\n\n137\n) and non-noble metal (e.g., Fe,\n138\n Co,\n47\n\n,\n\n85\n\n,\n\n139\n Ni,\n97\n\n,\n\n140\n Cu,\n141\n\n,\n\n142\n and Ga\n84\n), have been decorated on g-C3N4 to synthesize single-atom photocatalysts for hydrogen evolution. Wu and coworkers have decorated isolated Pt atoms on g-C3N4 via a liquid-phase reaction of H2PtCl6 and g-C3N4 followed by the low-temperature annealing. The as-synthesized Pt-CN single-atom photocatalyst presents a longer lifetime of photoexcited electrons by the formation of surface Pt-N/C bonds, which results in boosted photocatalytic performance for H2 production, outpacing Pt\u00a0nanoparticles and pristine g-C3N4 by 8.6 and 50 times, respectively.\n46\n Yu and coworkers have further demonstrated that the single-atom photocatalyst of Pd/g-C3N4 exhibits enhanced charge separation/transfer due to the introduction of isolated Pd sites, thus giving much better H2 evolution activity than the benchmarked Pt/g-C3N4.\n87\n During the rational design of g-C3N4-based single-atom photocatalysts for hydrogen evolution, researchers pay much attention to modulate the coordination configuration between unsaturated metal active sites and C/N atoms to enhance the light response, charge separation, and adsorption/activation processes, thus significantly elevating the photocatalytic performance.\n55\n\n,\n\n56\n\n,\n\n134\n\n,\n\n140\n Zou and coworkers have synthesized a single-atom photocatalyst of isolated Ag sites decorated g-C3N4 with a unique coordination configuration of Ag-N2C2 (Ag-N2C2/CN) through a facile supramolecular approach for H2 evolution (Figure\u00a09F).\n56\n DFT simulations indicate that the C and N co-coordinated Ag sites exhibit better charge distribution than Ag nanoparticles (AgNPs) and isolated Ag-N4 sites (Figure\u00a09G), which renders a faster transfer of photogenerated electrons from C3N4 to Ag and significantly reduces the energy barrier of H2 evolution. The Ag-N2C2/CN thus delivers much better photocatalytic activity than AgNP-, Ag-N4-, and Pt nanoparticle-based C3N4 photocatalysts (Figure\u00a09H).Non-noble metals have also been applied to construct the g-C3N4-based single-atom photocatalysts for hydrogen evolution with impressive performance. Wei and coworkers have grafted isolated Co1-N4 sites on g-C3N4 nanosheets via ALD, and the formed photocatalyst delivers a superior H2 evolution rate of more than 10.8\u00a0\u03bcmol h\u22121, surpassing the pristine g-C3N4 by 11 times.\n85\n Theoretical investigations validate that the coordinated donor N atoms improve the electron density of the unsaturated Co active center and accelerate the formation of the key intermediate of Co hydride, hereby promoting the H\u2013H coupling to speed up the H2 production. They have further fabricated a Co1-phosphide/PCN photocatalyst of isolated Co1-P4 sites decorated g-C3N4 by a simple phosphidation strategy, elevating the H2 production rate over 410.3\u00a0\u03bcmol h\u22121 g\u22121.\n47\n The newly generated Co1-P4 configuration leads to the generation of mid-gap state in Co1-phosphide/PCN, forcefully improving the light-harvesting capacity and inhibiting the recombination of electron\u2013hole pairs. The unsaturated Co active sites coordinated with P atoms are more stable and favorable for the adsorption/activation of water molecules than the N sites in pristine g-C3N4. Moreover, the electric conductivity of g-C3N4 has been enhanced after P doping. These positive factors markedly promote the photocatalytic activity of Co1-phosphide/PCN for the hydrogen evolution from water splitting.Moreover, researchers have also developed various single-atom photocatalysts based on other semiconductor supports (i.e., metal sulfide\n58\n\n,\n\n72\n\n,\n\n96\n\n,\n\n103\n\n,\n\n107\n\n,\n\n113\n\n,\n\n143\u2013147\n and MOFs\n57\n\n,\n\n71\n\n,\n\n91\n\n,\n\n93\n\n,\n\n148\u2013150\n) for hydrogen evolution. Wang and coworkers have decorated isolated Ni atoms onto the zinc-blende Cd\u2013Zn sulfide quantum dots (ZCS QDs) with the finely tuned concentrations of Ni sites for photocatalytic H2 evolution (Figures 9I\u2013K).\n58\n HAADF-STEM observations demonstrate that the shapes of as-synthesized Ni\nx\nZCS QDs (x\u00a0= 0.015625, 0.03125, 0.0625, and 0.125) vary from tetrahedra to truncated tetrahedra with the increase of Ni contents (Figure\u00a09I). Atomically resolved HAADF-STEM investigations further demonstrate that the (111) and (110) facets are the only existed facets for Ni\nx\nZCS QDs in the presence of Ni. The as-constructed surface junctions between the anisotropic (111) and (110) crystal plane in the same phase of Ni\nx\nZCS QDs significantly reinforce the charge carrier separation/transfer and result in enhanced electronic conductivity by the built-in electric field (Figure\u00a09J), leading to optimized surface hydrogen adsorption thermodynamics. The best-in-class Ni0.03125ZCS QDs catalyst herein exhibits an ultrahigh H2 production rate of 18.87\u00a0mmol h\u22121 g\u22121 under sunlight (Figure\u00a09K). Jiang and coworkers have anchored isolated Pt atoms into the Al-based porphyrinic MOF of ((AlOH)2H2TCPP (H2TCPP\u00a0= 4,4\u2032,4\u2033,4\u2032\u2033-(porphyrin-5,10,15,20-tetrayl)tetrabenzoate), simplified as Al-TCPP) for photocatalytic hydrogen evolution.\n93\n The Al-TCPP has affluent coordination sites for the implantation of Pt atoms, thus providing efficient channels for the electron transfer from the MOF photosensitizer to the unsaturated Pt active sites. The as-synthesized Al-TCPP-Pt photocatalyst herein exhibits outstanding activity for H2 evolution with a TOF of 35 h\u22121, outperforming the Pt nanoparticle catalyst stabilized by the same MOF by \u223c30 times.Photocatalytic reduction of CO2 into fuels and value-added chemicals by harvesting solar energy has delivered one of the ideal blueprints for reducing the concentration of atmospheric CO2 greenhouse gas in an ecofriendly manner.\n11\n\n,\n\n24\n Major ongoing research of photocatalytic CO2 reduction is to create efficient catalysts for overcoming the sluggish kinetics and speeding up the conversion rate of the linear CO2 molecule.\n29\n\n,\n\n114\n Due to the distinctive advantages for adsorption and activation of molecules, single-atom photocatalysts have been engaged as promising candidates for photocatalytic CO2 reduction.\n48\n\n,\n\n89\n\n,\n\n151\u2013155\n\nYe and coworkers have first applied SACs for photocatalytic CO2 reduction in 2016.\n45\n By introducing the isolated Co atoms into the porphyrin units of the MOF-525, SACs with unsaturated Co sites (MOF-525-Co) have thus been constructed. The as-synthesized catalyst exhibits a CO production rate of 200.6\u00a0\u03bcmol g\u22121 h\u22121 and a CH4 evolution rate of 36.67\u00a0\u03bcmol g\u22121 h\u22121, exceeding that of the parent MOF by 3.13 and 5.93-fold, respectively. The authors have attributed\u00a0the greatly improved photocatalytic activity of MOF-525-Co to the boosted\u00a0separation efficiency of electron\u2013hole pairs in porphyrin units after the Co introduction. Since then, many research works of photocatalytic CO2 reduction\u00a0into CO or CH4 fuel gas have been reported over various single-atom photocatalysts along\u00a0with the production of H2, where H2O provides the hydrogen resource.\n59\n\n,\n\n74\u201376\n\n,\n\n114\n\n,\n\n155\u2013166\n Li et\u00a0al. have created a unique Fe-N4O site supported on nitrogen-rich carbon for photocatalytic CO2 reduction via a facile top-down strategy (Figures 10A\u201310C).\n167\n The EXAFS observation reveals that the isolated Fe sites exist in high valence due to the construction of a coordinated Fe-N4O configuration (Figure\u00a010B). These optimized unsaturated Fe active sites significantly facilitate the adsorption of CO2 molecules and accelerate the generation of key intermediate COOH\u2217, which promotes the photocatalytic system rendering a stable turnover number of 1,494 within 1\u00a0h and an excellent selectivity of 86.7% for CO generation (Figure\u00a010C).Lou and coworkers have recently decorated isolated Co atoms onto the W18O49 ultrathin nanowires (W18O49@Co) via surface modification engineering (Figure\u00a010D).\n88\n The inclusion of Co active sites has modulated the band structure of the W18O49 support, thus enhancing the redox capability of photogenerated electrons and promoting the charge separation/transfer for the CO2 reduction system. Furthermore, DFT calculations confirm that the anchored Co sites have modified the surface of W18O49@Co to enhance the adsorption of CO2 molecules (Figure\u00a010E). Benefiting from these positive factors, the unsaturated Co sites could work as reaction switches in a tandem system of photocatalytic reduction with [Ru(bpy)3]Cl2\u00b76H2O (bpy\u00a0= 2,2\u2032-bipyridine) serving as the photosensitizer, which results in high activity with a considerable CO production rate of 21.18\u00a0mmol g\u22121 h\u22121 and a hydrogen evolution rate of 6.49\u00a0mmol g\u22121 h\u22121 (Figure\u00a010F).In addition to producing the fuel gases, the photocatalytic CO2 reduction over single-atom photocatalysts can also achieve the generation of value-added chemicals, such as methanol and ethanol.\n60\n\n,\n\n61\n\n,\n\n168\n Yu and coworkers have presented that the isolated Cu atoms-modified g-C3N4 can provide unsaturated C-Cu-N2 active sites for the reduction of CO2 into various C1 products, including CH4, CO, and methanol.\n60\n Zbo\u0159il and coworkers have also constructed a g-C3N4-based single-atom photocatalyst with Ru-N/C active sites for reusable reduction of CO2 into methanol, which can yield 1,500\u00a0\u03bcmol of methanol per gram of catalyst upon six hours of the photocatalytic reaction.\n168\n Wang and coworkers have decorated the isolated Cu sites into the UiO-66-NH2 MOF through a photoinduction strategy (Figure\u00a010G). The as-synthesized catalyst could trigger the solar-driven reduction of CO2 into methanol with a yield rate of 5.33\u00a0\u03bcmol h\u22121 g\u22121 and ethanol with a generation rate of 4.22\u00a0\u03bcmol h\u22121 g\u22121 (Figure\u00a010H).\n61\n DFT calculations directly unravel that the inclusion of isolated Cu atoms on the UiO-66-NH2 significantly reduces the formation barrier of both CHO\u2217 and CO\u2217 during the CO2 reduction (Figure\u00a010I), thus resulting in impressive selectivity for methanol and ethanol. Moreover, photocatalytic conversion of CO2 into C2+ product over single-atom photocatalyst has also been reported recently. By introducing isolated Mo atoms into 2,2\u2032-bipyridine-based COF, Kou et\u00a0al. have constructed a Mo-COF single-atom photocatalyst with a coordination configuration of MoN2, which could reduce CO2 into C2H4 product with a\u00a0yield rate of 3.57\u00a0\u03bcmol h\u22121 g\u22121 and the selectivity of 42.92% under visible light.\n62\n\nIn situ FTIR investigation and theoretical calculation have directly demonstrated that the inclusion of isolated MoN2 sites in COF could enhance the adsorption/activation of CO2 molecules and further promote the hydrogenation process for the generation of C2H4.Furthermore, researchers have also achieved the gas-phase CO2 photoreduction over single-atom photocatalysts, with H2 serving as a hydrogen source. Ozin and coworkers have constructed surface frustrated Lewis pairs (SFLPs) on defective In2O3 through the isomorphic substitution of isolated Bi sites for Lewis acidic In3+ sites in In2O3, enabling enhanced activity for gas-phase CO2 photocatalysis.\n115\n The isolated Bi3+ sites (Bi\u2019), oxygen vacancies ([O]v), coordinately unsaturated In sites (In\u2019) below the CB of In2O3, and the oxygen states (O\u2032) above the VB of In2O3 could serve as mid-gap states (comprising SFLPs) for trapping photoexcited electrons and holes, respectively, greatly inhibiting the recombination of electron\u2013hole pairs and driving the reaction between H2 and CO2 (Figure\u00a010J). Herein, benefiting from the enhanced charge separation/transfer and also the adsorption capacity, the as-constructed Bi\nx\nIn2-x\nO3 photocatalyst shows an excellent CO evolution rate higher than that of pristine In2O3 by three orders of magnitude (Figure\u00a010K), while also renders remarkable reactivity toward photocatalytic methanol production (Figure\u00a010L). This work further exemplifies the distinct advantage of decoration of isolated metal sites for the atom-scale photocatalyst engineering toward the adsorption and activation of inert molecules.The ammonia (NH3) synthesis through the well-known Haber-Bosch process has artificially provided essential nitrogen sources for various N-containing reactions relating to human development, including the production of fertilizer, drugs, and fine chemicals.\n169\n However, this process requires harsh temperature (400\u00b0C\u2013500\u00b0C) and pressure (10\u201330 Mpa) due to the great activation barrier of nonpolar N\u2261N in N2 molecule, which consumes \u223c2% of the world\u2019s energy production and generates \u223c3% of the CO2 emissions worldwide per year.\n170\n It is thus highly demanded for researchers to explore alternative pathways to produce NH3 via the fixation of N2. Recently, artificial photocatalytic N2 reduction under ambient conditions has offered a more sustainable approach beyond the traditional Haber-Bosch process.\n171\n Single-atom photocatalysts have thus become emerging platforms for the photocatalytic N2 reduction reaction.\n63\n\n,\n\n78\n\n,\n\n172\n\n,\n\n173\n\nXie and coworkers have first reported a g-C3N4-based single Cu photocatalyst (Cu\u2013CN) for N2 reduction, which achieves an NH3 yield rate of 186\u00a0\u03bcmol g\u22121 h\u22121 and quantum efficiency of 1.01% under light illumination at 420\u00a0nm.\n63\n The coordination between isolated Cu atoms and edged N atoms in g-C3N4 results in extra active lone-pair electrons on the g-C3N4 support and enhanced adsorption capacity of unsaturated Cu active sites, thus leading to an impressive performance for the activation of N2 molecules. Zhong and coworkers have anchored isolated Pt atoms in the covalent triazine framework (CTF) nanosheets with a coordinated configuration of Pt-N3 for photocatalytic N2 reduction.\n64\n Due to the introduction of isolated Pt sites, the modified CB position of CTF is higher than the electrode potential of N2/NH3, enabling the single-atom photocatalyst more thermodynamically feasible to accelerate the reaction (Figure\u00a011\nA). The as-constructed Pt-SACs/CTF photocatalyst thus exhibits a high NH3 yield rate of 171.4\u00a0\u03bcmol g\u22121 h\u22121 without the addition of sacrificial agents (Figure\u00a011B). The isotopic labeling experiments further confirm the formation of NH3 from photocatalytic N2 reduction (Figure\u00a011C).Apart from organic semiconductor materials, inorganic semiconductors have also been applied to design single-atom photocatalysts for photocatalytic N2 reduction.\n77\n\n,\n\n100\n\n,\n\n171\n\n,\n\n174\n For example, Niu et\u00a0al. have decorated isolated Ru atoms on defect-rich TiO2 nanotubes for N2 fixation (Figure\u00a011D).\n100\n The typical charge-transfer state induced by ligand (TiO2)-to-metal (Ru) has greatly promoted the transfer of photogenerated electrons and turned the Ru active site into a strong electron pump (Figure\u00a011E). As a result, the single-atom photocatalyst exhibits higher efficiency than anchored Ru nanoparticles and pristine TiO2 nanotubes (Figure\u00a011F). Recently, Wu et\u00a0al. have further reported the elevated photocatalytic activity over macro/mesoporous TiO2-SiO2 supported Fe sites for N2 reduction (Figure\u00a011G), which gives an NH3 yield rate of 32\u00a0\u03bcmol g\u22121 h\u22121 in the absence of any sacrificial agents and cocatalysts (Figure\u00a011H).\n77\n M\u00f6ssbauer spectroscopy directly confirms that the generation of photoexcited hole-trapping polarons around the Fe active sites has resulted in the formation of high-valent Fe(IV) species (Figure\u00a011I), which significantly promotes the N2 reduction on the adjacent oxygen vacancy. Moreover, Zhang and coworkers have developed a single-atom photocatalyst of electron-rich Cu\n\u03b4+ sites and oxygen vacancies co-decorated Zn-Al layered double hydroxide nanosheets via a simple coprecipitation method (Figure\u00a011J).\n174\n The unsaturated Cu\n\u03b4+ active sites and oxygen vacancies greatly enhance the charge separation/transfer and the adsorption of N2 molecules (Figure\u00a011K), thus resulting in optimized activity for photocatalytic N2 reduction with a durable NH3 yield rate of 110\u00a0\u03bcmol g\u22121 h\u22121 in pure water under UV-vis excitation (Figure\u00a011L).Solar-driven organic synthesis over homogeneous catalysts (e.g., metal complexes and organic photosensitizers) represents one of the most sustainable and economic strategies to produce valuable organic compounds.\n175\n However, the high cost and nonrecyclability of homogeneous catalysts greatly hamper the viable application of photocatalytic organic reactions. To this end, the construction of efficient heterogeneous catalysts for photocatalytic organic synthesis has received wide attention from\u00a0chemists.\n65\n\n,\n\n176\n Considering the bridging function between homo- and heterogeneous catalysis, single-atom photocatalysts have been regarded as the promising candidates to achieve selective and recyclable photocatalytic organic synthesis.\n16\n\n,\n\n175\n\n,\n\n177\n\nWang and coworkers have confirmed that the isolated Ag sites can serve as effective photocatalytic active centers in their AgF/visible-light system for selective hydrodehalogenation and dehalogenation-arylation of various organic halides without any organic additives (Figure\u00a012\nA).\n178\n The photolysis of AgF results in the in situ generation of both AgNPs and single-atom Ag (SAAg) sites. Under the visible-light excitation, AgNPs can serve as light-harvesting units attributed to the localized surface plasmon resonance, while SAAg provides surface-active sites for the anchoring of activated halides and driving the photocatalytic reactions. Based on the robust synergy between AgNPs and SAAg, the TOF of deiodination-phenylation reaction over isolated Ag center can reach 6,000 h\u22121 under mild conditions. Yang et\u00a0al. have\u00a0loaded isolated Ni atoms onto TiO2 supports for selective sulfonation of enamides into amidosulfones with yields of 99% for 33 examples (Figure\u00a012B).\n176\n The as-synthesized Ni/TiO2 photocatalyst can selectively realize the formation of \u03b1-amidosulfones and \u03b2-propionamidosulfones with considerable recyclability and a high turnover number over 18,963, which also exhibits high tolerance of functional groups and enables easy gram-scale reaction with considerable efficiency.Organic semiconductor supports have also been applied to construct single-atom photocatalysts for organic synthesis.\n109\n\n,\n\n175\n\n,\n\n179\n\n,\n\n182\n Wang et\u00a0al. have anchored isolated Co atoms on carbon quantum dots (CoSAS@CD) with a coordinated Co-N4 structure through straightforward hydrolysis of vitamin B12 in sodium hydroxide solution. The carbon dots serve as both the photosensitizer and the support for isolated Co sites. The elevated visible-light adsorption capacity and charge separation/transfer by the synergy between the Co atoms and carbon quantum dots enable the CoSAS@CD with excellent oxidation ability, exhibiting an oxygen evolution rate over 168\u00a0\u03bcmol h\u22121 g\u22121 for water oxidation and imine synthesis with a high conversion of \u223c90% and selectivity over 99% (Figure\u00a012C).\n179\n Song and coworkers have assembled isolated Ni active sites onto the g-C3N4 with imidazole auxiliary ligand for C\u2013O cross-coupling reactions under visible light, which could efficiently realize the etherification of various aryl bromides with alcohols/water, giving high turnover numbers up to 500 (Figure\u00a012D).\n180\n The dispersion of isolated Ni sites can be well maintained without aggregation after the photocatalytic reaction and be reused with sustained activity, indicating the excellent stability and recyclability of SACs in photocatalytic organic synthesis. Reisner and coworkers have further emphasized that the single-atom photocatalyst of Ni active sites deposited on mesoporous g-C3N4 is a robust, earth-abundant, low-cost, and heterogeneous easily recyclable platform for achieving C\u2013O coupling reactions between aryl halides and aliphatic alcohols under the mild condition without the addition of external ligands, which exhibits a reactivity trend paralleling to that of homogeneous catalysts (Figure\u00a012E).\n65\n\nIn addition, the use of single-atom photocatalysts for biomass reforming to produce value-added organic chemicals has emerged recently.\n66\n\n,\n\n181\n Guo and coworkers have demonstrated that the single-atom Pt-loaded commercial P25-TiO2 (Figure\u00a012F) exhibits superior activity for photocatalytic reforming of acetone into highly value-added 2,5-hexanedione (HDN) and H2 with a selectivity of 93%.\n162\n The HDN production rate can be optimized to 3.87\u00a0mmol g\u22121 h\u22121 over the as-synthesized Pt/P25-TiO2 photocatalyst, outpacing those of other precious isolated metals (Ru, Rh, and Ir) or Pt nanoparticle-loaded photocatalysts by at least 13 times (Figure\u00a012G). Furthermore, they have developed partially reduced Pd-P3 active sites on CdS nanorods (PdPSA-CdS) for solar-driven reforming of bioethanol into hydrogen and a C6 compound of 1,1-diethoxyethane, exhibiting a photocatalytic generation rate of 35.1\u00a0mmol g\u22121 h\u22121 and a selectivity of nearly 100% (Figure\u00a012H).\n66\n By the investigations of in situ attenuated total reflection-infrared spectra combined with the theoretical simulation of reaction pathway, the reinforced electron transfer from ethanol molecule to the unoccupied P 3p and Pd 4d states has been demonstrated as the origin for the highly effective activation of C\u2013H and O\u2013H bonds over the PdPSA-CdS photocatalyst.Organic pollutants, including industrial dyes, agrochemicals, and pharmaceuticals, have undesirable hazards on the water environment and potentially severe danger to the creatures.\n183\u2013185\n It is urgent to explore efficient approaches for removing these contaminants, thus achieving a sustainable planet. The traditional removal of organic pollutants mainly relies on biodegradation and solid-adsorption separation process, which still suffer from kinetic inertness and low processing capacity for trace contaminants.\n186\n Photocatalytic degradation has provided an intriguing avenue to dispose of these pollutants in trace concentration by harvesting solar energy and promoting the generation of radical oxygen species.\n108\n\n,\n\n187\n\n,\n\n188\n Single-atom photocatalysts have also been demonstrated as efficient candidates for photocatalytic pollutant degradation due to the sufficient surface-active sites and accelerated charge separation/transfer.Zhao and coworkers have decorated isolated Ag atoms onto the mesoporous g-C3N4 for the photocatalytic degradation of bisphenol A (BPA) with the addition of peroxymonosulfate (PMS) under visible light (Figures 13A\u201313C).\n189\n The introduction of Ag active sites promotes the capture of photons for g-C3N4 in the visible-light range by narrowing the bandgap of g-C3N4. The suitable match of energy level between Ag and g-C3N4 results in a quick transfer of photoexcited electrons while the addition of PMS accelerates the separation of electron\u2013hole pairs (Figure\u00a013A), leading to a 100% degradation of BPA over 0.1\u00a0g L\u22121 of photocatalysts within 1\u00a0h (Figure\u00a013B). Electron spin resonance spectra (Figure\u00a013C) and quenched experiments for free radicals directly confirm that the construction of SAAg/g-C3N4 photocatalyst could promote the generation of radical oxygen species (i.e., sulfate radicals, superoxide radicals, and photogenerated holes), significantly accelerating the photocatalytic degradation of BPA.Dong et\u00a0al. have developed laminar COF-909(Cu) nanorods for the efficient solar-driven degradation of sulfamethoxazole (Figures 13D and 13E).\n190\n The decorated Cu active sites enhance the light absorption capability of COF and elevate the separation of electron\u2013hole pairs (Figure\u00a013D). The COF-909(Cu) further provides abundant binding sites for adsorbing target molecules in the channels of COF, which delivers state-of-the-art photocatalytic activity for the sulfamethoxazole degradation with a high kinetic constant of 0.133\u00a0min\u22121, exceeding the pristine COF-909 and commercial TiO2-P25 by 27 and 40 times, respectively (Figure\u00a013E).Furthermore, Wang et\u00a0al. have demonstrated the construction of a STAO with W6+ and W5+ sites for the visible-light photocatalytic degradation of various dyes (e.g., dimethyl yellow and methyl red).\n50\n Spherical aberration (Cs) corrected STEM-HAADF observation directly reveals the monodispersed nature of STAO (Figure\u00a013F). The combinations of the XPS, EXAFS, and electron energy loss spectroscopy clearly confirm the distorted octahedron structure of STAO with W6+ and W5+ deriving from the coordinated configuration of W-O6 (Figure\u00a013G). The as-synthesized photocatalyst exhibits an impressive and stable photocatalytic degradation rate of 0.24 s\u22121, outpacing various photocatalysts by two orders of magnitudes (Figure\u00a013H). Based on the systematical experimental results and theoretical calculations, the unsaturated W5+ sites in STAO are demonstrated as the real active centers, which enable the efficient transfer of photogenerated electrons from HOMO to LUMO\u00a0+1 state to drive the entire photocatalytic degradation reaction (Figure\u00a013I). The as-synthesized STAO has also enriched the family members of single-atom photocatalysts.Duo to the highly configurable structure, single-atom photocatalysts have also shown great application potentials in some other photocatalytic systems involving NO\nx\n removal,\n67\n\n,\n\n191\u2013193\n CH4 conversion,\n51\n\n,\n\n68\n\n,\n\n194\n H2O2 production,\n69\n\n,\n\n98\n\n,\n\n195\n\n,\n\n196\n as well as photocatalytic sensing.\n70\n\n,\n\n81\n By sharing the accelerated charge separation/transfer and enhanced capacity for adsorption/activation of reactants, these single-atom photocatalytic reactions also show appealing activity, selectivity, and durability.Liu et\u00a0al. have developed Pd-N3 active sites by the decoration of isolated Pd atoms onto the g-C3N4 support with tunable carbon defects (Figure\u00a014\nA), which exhibit outstanding and stable activity for photocatalytic NO removal, exceeding the pristine g-C3N4 by 4.4 times (Figure\u00a014B).\n67\n Steady-state PL measurements confirm the accelerated separation/transfer of photoexcited carriers after the introduction of isolated Pd metal sites (Figure\u00a014C). The as-constructed Pd-Cv-CN catalyst thus provides sufficient photoelectrons for the formation of \u00b7O2\n\u2212 and OH\u2212, resulting in highly selective conversion of NO gas into NO3\n\u2212.In the case of photocatalytic CH4 conversion, Chen et\u00a0al. have recently loaded isolated Nb atoms into hierarchical macro-mesoporous TiO2-SiO2 for constructing a single-atom photocatalyst (Figure\u00a014D), which exhibits an optimal conversion rate of 3.57\u00a0\u03bcmol g\u22121 h\u22121 with considerable recyclability for non-oxidative coupling of CH4 (Figure\u00a014E).\n68\n The dopant of isolated Nb atoms has facilitated the charge separation and elevated the electron mobility, which is significantly beneficial to the activation of methane and the desorption of ethane product. They have further emphasized that the Nb, Mo, W, Ta dopants on TiO2-SiO2 exhibit much better activity for CH4 conversion than Cu, Ga, Fe dopants, as shown in Figure\u00a014F.The applications of SACs have been further extended for photocatalytic H2O2 production. Ohno and coworkers have developed an efficient single-atom photocatalyst of isolated Sb atoms decorated g-C3N4 (Sb-SAPC) through a wet-chemical method from the precursor of NaSbF6 and melamine.\n69\n The fitted EXAFS spectrum and the DFT simulation reveal that each Sb atom on average coordinates with 3.3N atoms in a typical Sb-SAPC15 catalyst (Figure\u00a014G), directly indicating the isolated dispersion of unsaturated Sb sites over g-C3N4. A solar-to-chemical conversion efficiency up to 0.61% (Figure\u00a014H) can be achieved over an optimized Sb-SAPC15 catalyst for photocatalytic H2O2 production from a two-electron oxygen reduction reaction (ORR). The isolated Sb active sites can deeply trap the photogenerated electrons, serving as the photoreduction centers for two-electron ORR (Figure\u00a014I). Meanwhile, the adjacent N atoms enable the accumulation of holes, which significantly facilitates the kinetics of water oxidation. The collaborative effect between isolated Sb sites and coordinated N atoms thus dramatically accelerates the overall photocatalytic reaction.Guo and coworkers have recently applied the isolated Pd atoms decorated TiO2 catalyst (Pd/TiO2) as a photocatalytic sensing platform for the highly selective detection of the organophosphorus pesticide chlorpyrifos.\n70\n The inhibition effect of chlorpyrifos on the activity of photocatalytic hydrogen evolution over Pd/TiO2 has served as the stable detection sensor for the chlorpyrifos molecules (Figures 14J and 14K), which results in an extremely low detection limit of 0.01\u00a0ng mL\u22121 (Figure\u00a014L). This work opens up new insights for the development of biosensing strategies and extends the application of single-atom photocatalysts.We have summarized the key principles of SACs and their versatile applications for photocatalysis. It has no doubt that single-atom photocatalysts are excellent brand-new candidates to construct efficient photocatalytic systems due to the accelerated charge separation/transfer efficiency and enhanced molecular adsorption/activation capacity. The advantages of single-atom photocatalysts are highly conducive to promoting photocatalytic activities of the photocatalytic reactions and thus covering a greater application range. Apart from the previous outstanding achievements, there are still many challenges in the exploration and practical uses of single-atom photocatalysts, including how to achieve the long-term stability and high loading of isolated reactive centers. Some future research directions and deployable solutions to overcome the ongoing challenges are also proposed as follows:\n\n(1)\nThe relatively low stability of single-atom photocatalysts is one obvious drawback for photocatalysis.\n197\n The poisoning effect derived from the strong bonding of the reaction intermediate or by-product onto the isolated metal sites may deactivate the single-atom photocatalysts. Additionally, the photogenerated electrons induced conversion of isolated reactive centers into zero-valence metal atoms may cause the aggregation of isolated metal sites and the formation of clusters or nanoparticles.\n40\n Enhancing the metal-support interaction and optimizing photocatalytic reaction under ideal conditions may somewhat prevent the aggregation of isolated metal sites during the reaction. Moreover, the preparation of nonmetal-based SACs for photocatalysis with low cost and effective modulation for substrates may provide another intriguing solution to realize the improvement of both reactivity and stability. Although there are no related reports yet, it is highly expected that the synergetic interaction between nonmetal atoms and semiconductive supports may significantly prevent the migration of active sites and thus further explore the intrinsic activity of the single-atom photocatalysts.\n\n\n(2)\nAlthough the energy conversion efficiency over single-atom photocatalysts has shown a distinguished advantage for the average reactive centers, their overall performances are still far from satisfactory because of the low densities of unsaturated active sites. Developing single-atom photocatalysts with high loading of metal atoms favors elevating the densities of both photoexcited electron pumps and photocatalytic active sites. However, due to the high surface energy, isolated metal sites are rather apt to relocate and aggregate into clusters or even nanoparticles during the synthetic process.\n14\n And thus, the surface structure of the host semiconductive supports should be further modified to boost the meta-support interaction for increasing their loading amount. In addition to the creation of cation/anion vacancies in the surface modification process, introducing sufficient anchor sites (e.g., N, P, and S) or grafting specific functional groups (e.g., pyridine and \u2212NH2) could also be applied to provide abundant binding sites for stabilizing the isolated atoms.\n\n\n(3)\nThe precise control of the coordination geometry of the reactive centers and the number of isolated atoms is crucial to tune the activity and selectivity of single-atom photocatalysts but still remains as one great challenge. Accurate control of the interactions between the isolated reactive centers and semiconductive supports during the synthetic process is necessary for constructing the desired configuration, including the invention of a modular strategy to guarantee the geometry of implanted reactive centers. Furthermore, developing synthesis strategies based on the theoretical predictions of the structure-performance relationship may also provide a feasible approach. Furthermore, as the current XAS investigation still exhibits a significant error (\u223c20%) in confirming the coordination number for isolated atoms,\n17\n it is highly desirable to improve the accuracy of the characterization strategies to clarify the coordination configurations of single-atom photocatalysts in practice.\n\n\n(4)\nExploring the synergetic interaction between two neighboring monomers has great potential to manipulate the catalytic properties and deepen the mechanistic understanding of heterogeneous catalysis. The controllable synthesis of photocatalysts with dual-single-atoms or multi-single-atoms should be thus considered. Due to the synergistic effect among adjacent reactive centers in the dual or multisingle-atom photocatalysts, the reaction paths for the reactant may be greatly modified, and it may thus result in the significantly reduced reaction barrier energy and obviously improved catalytic performance. Moreover, the construction of dual- or multi-SACs will further enrich the family members of single-atom photocatalysts.\n\n\n(5)\nPhotocatalytic and enzymic systems in nature have provided a delicate blueprint for solar-to-chemical energy conversion, which generally exhibit prominent activity and selectivity under ambient conditions. From a structural aspect, isolated atom sites in SACs are highly analogous to the simplest active sites in enzymes. By mimicking the configurations of active centers such as enzyme catalytic pockets in natural photosynthetic reactions, the higher-level biomimetic design may be achieved to further release the overall potential of SACs for photocatalytic energy conversion. The integrations of terminal functional ligands or additional active sites around the isolated metal sites could be deployable strategies for developing these bioinspired single-atom photocatalysts. The related investigation will also gain a deep understanding of natural photosynthetic processes.\n\n\n(6)\nLarge numbers of single-atom photocatalysts have been developed based on defect-laden semiconductor materials containing abundant surface binding sites for the coordinated metal atoms, thus achieving the enhanced charge separation/transfer. However, a high concentration of defect may deteriorate the crystallinity of host semiconductor materials and thus increase the massive recombination of electron\u2013hole pairs in the volume or on the surface of single-atom photocatalysts. Plenty of research has focused on the distribution and coordination configuration of single-atom active sites while ignoring the structural control of semiconductor supports. Therefore, the structures (e.g., crystallinity, defects) of host semiconductors in single-atom photocatalysts are suggested to be controlled meticulously due to the above trade-off, their contribution to the whole reaction process should be further identified, thereby realizing the innovation in catalyst design and a fresh round of elevation for the photocatalytic activity.\n\n\n(7)\nThe full interpretation of the charge separation/transfer process in single-atom photocatalysts induced by isolated metal sites is still challenging. Ultrafast transient absorption spectroscopy, particularly combined with electrochemical or microscopy techniques, is a powerful strategy to study the energy transfer and trapping processes in photocatalytic reactions. Developing real-time ultrafast transient absorption techniques to track the dynamics of photoexcited carriers in photocatalysis will further advance the understanding of electron pump models and electron trap states induced by the isolated metal sites, thus elucidating the deeply intrinsic mechanism of charge transfer and energy conversion in the photocatalytic systems.\n\n\n(8)\nTracking the structural evolvements of active sites during the photocatalytic reactions provides not only the deep insight into the single-atom photoactivation process but also the guidelines for the rational design of effective photocatalysts. However, the related experimental pieces of evidence are still highly limited to understand the molecular adsorption/activation mechanism over SACs in photocatalytic systems. In situ or operando investigations combining with various techniques such as Raman spectroscopy, XAS, and XPS could be excellent approaches to detecting the dynamic changes of chemical state and coordinated environment of the isolated metal sites during the reactions. The development and application of these technologies in single-atom photocatalysis may provide more holistic trails for structural evolution, which are highly associated with the adsorption and activation of molecules over the unsaturated metal active sites.\n\n\n(9)\nTheoretical calculation combining with experimental results has become a robust strategy to investigate the electronic structure of catalysts and the molecular adsorption/activation in the photocatalytic process, forcefully revealing the working mechanism of single-atom photocatalysis on the atomic scale. However, the theoretical modeling of dynamic changes of active sites over the photocatalytic reactions is rather limited to achieve a rational cognition of the single-atom photoactivation cycle. The combinations of different simulation methods, such as time-dependent DFT and molecular dynamics, may provide deployable and reasonable approaches for exploring the evolution of single-atom photocatalyst during the reactions and revealing the photocatalytic activation mechanisms.\n\n\n(10)\nAs one of the most forceful parts of artificial intelligence, machine learning based on computer algorithms enables fast and reliable predictions through data mining, which shows great potential in exploring high-efficiency catalysts. The application of SACs has covered a wide range of photocatalytic reactions, while the possibility of undiscovered single-atom photocatalysts is boundless, resulting in infinite combinations of structure\u2013activity relationship and great challenges in the rational design of single-atom photocatalysts. It is thus highly desirable to explore appropriate machine learning methods integrating with theoretical calculation data to predict the catalytic performance of single-atom photocatalysts and figure out the ideal configuration for targeted reactions. The implementation of machine learning will also construct reliant structure-performance relationships for the photocatalytic reactions, thereby enhancing the understanding of single-atom photocatalysis and promoting the rational development of efficient single-atom photocatalysts with highly applicable potential.\n\n\nThe relatively low stability of single-atom photocatalysts is one obvious drawback for photocatalysis.\n197\n The poisoning effect derived from the strong bonding of the reaction intermediate or by-product onto the isolated metal sites may deactivate the single-atom photocatalysts. Additionally, the photogenerated electrons induced conversion of isolated reactive centers into zero-valence metal atoms may cause the aggregation of isolated metal sites and the formation of clusters or nanoparticles.\n40\n Enhancing the metal-support interaction and optimizing photocatalytic reaction under ideal conditions may somewhat prevent the aggregation of isolated metal sites during the reaction. Moreover, the preparation of nonmetal-based SACs for photocatalysis with low cost and effective modulation for substrates may provide another intriguing solution to realize the improvement of both reactivity and stability. Although there are no related reports yet, it is highly expected that the synergetic interaction between nonmetal atoms and semiconductive supports may significantly prevent the migration of active sites and thus further explore the intrinsic activity of the single-atom photocatalysts.Although the energy conversion efficiency over single-atom photocatalysts has shown a distinguished advantage for the average reactive centers, their overall performances are still far from satisfactory because of the low densities of unsaturated active sites. Developing single-atom photocatalysts with high loading of metal atoms favors elevating the densities of both photoexcited electron pumps and photocatalytic active sites. However, due to the high surface energy, isolated metal sites are rather apt to relocate and aggregate into clusters or even nanoparticles during the synthetic process.\n14\n And thus, the surface structure of the host semiconductive supports should be further modified to boost the meta-support interaction for increasing their loading amount. In addition to the creation of cation/anion vacancies in the surface modification process, introducing sufficient anchor sites (e.g., N, P, and S) or grafting specific functional groups (e.g., pyridine and \u2212NH2) could also be applied to provide abundant binding sites for stabilizing the isolated atoms.The precise control of the coordination geometry of the reactive centers and the number of isolated atoms is crucial to tune the activity and selectivity of single-atom photocatalysts but still remains as one great challenge. Accurate control of the interactions between the isolated reactive centers and semiconductive supports during the synthetic process is necessary for constructing the desired configuration, including the invention of a modular strategy to guarantee the geometry of implanted reactive centers. Furthermore, developing synthesis strategies based on the theoretical predictions of the structure-performance relationship may also provide a feasible approach. Furthermore, as the current XAS investigation still exhibits a significant error (\u223c20%) in confirming the coordination number for isolated atoms,\n17\n it is highly desirable to improve the accuracy of the characterization strategies to clarify the coordination configurations of single-atom photocatalysts in practice.Exploring the synergetic interaction between two neighboring monomers has great potential to manipulate the catalytic properties and deepen the mechanistic understanding of heterogeneous catalysis. The controllable synthesis of photocatalysts with dual-single-atoms or multi-single-atoms should be thus considered. Due to the synergistic effect among adjacent reactive centers in the dual or multisingle-atom photocatalysts, the reaction paths for the reactant may be greatly modified, and it may thus result in the significantly reduced reaction barrier energy and obviously improved catalytic performance. Moreover, the construction of dual- or multi-SACs will further enrich the family members of single-atom photocatalysts.Photocatalytic and enzymic systems in nature have provided a delicate blueprint for solar-to-chemical energy conversion, which generally exhibit prominent activity and selectivity under ambient conditions. From a structural aspect, isolated atom sites in SACs are highly analogous to the simplest active sites in enzymes. By mimicking the configurations of active centers such as enzyme catalytic pockets in natural photosynthetic reactions, the higher-level biomimetic design may be achieved to further release the overall potential of SACs for photocatalytic energy conversion. The integrations of terminal functional ligands or additional active sites around the isolated metal sites could be deployable strategies for developing these bioinspired single-atom photocatalysts. The related investigation will also gain a deep understanding of natural photosynthetic processes.Large numbers of single-atom photocatalysts have been developed based on defect-laden semiconductor materials containing abundant surface binding sites for the coordinated metal atoms, thus achieving the enhanced charge separation/transfer. However, a high concentration of defect may deteriorate the crystallinity of host semiconductor materials and thus increase the massive recombination of electron\u2013hole pairs in the volume or on the surface of single-atom photocatalysts. Plenty of research has focused on the distribution and coordination configuration of single-atom active sites while ignoring the structural control of semiconductor supports. Therefore, the structures (e.g., crystallinity, defects) of host semiconductors in single-atom photocatalysts are suggested to be controlled meticulously due to the above trade-off, their contribution to the whole reaction process should be further identified, thereby realizing the innovation in catalyst design and a fresh round of elevation for the photocatalytic activity.The full interpretation of the charge separation/transfer process in single-atom photocatalysts induced by isolated metal sites is still challenging. Ultrafast transient absorption spectroscopy, particularly combined with electrochemical or microscopy techniques, is a powerful strategy to study the energy transfer and trapping processes in photocatalytic reactions. Developing real-time ultrafast transient absorption techniques to track the dynamics of photoexcited carriers in photocatalysis will further advance the understanding of electron pump models and electron trap states induced by the isolated metal sites, thus elucidating the deeply intrinsic mechanism of charge transfer and energy conversion in the photocatalytic systems.Tracking the structural evolvements of active sites during the photocatalytic reactions provides not only the deep insight into the single-atom photoactivation process but also the guidelines for the rational design of effective photocatalysts. However, the related experimental pieces of evidence are still highly limited to understand the molecular adsorption/activation mechanism over SACs in photocatalytic systems. In situ or operando investigations combining with various techniques such as Raman spectroscopy, XAS, and XPS could be excellent approaches to detecting the dynamic changes of chemical state and coordinated environment of the isolated metal sites during the reactions. The development and application of these technologies in single-atom photocatalysis may provide more holistic trails for structural evolution, which are highly associated with the adsorption and activation of molecules over the unsaturated metal active sites.Theoretical calculation combining with experimental results has become a robust strategy to investigate the electronic structure of catalysts and the molecular adsorption/activation in the photocatalytic process, forcefully revealing the working mechanism of single-atom photocatalysis on the atomic scale. However, the theoretical modeling of dynamic changes of active sites over the photocatalytic reactions is rather limited to achieve a rational cognition of the single-atom photoactivation cycle. The combinations of different simulation methods, such as time-dependent DFT and molecular dynamics, may provide deployable and reasonable approaches for exploring the evolution of single-atom photocatalyst during the reactions and revealing the photocatalytic activation mechanisms.As one of the most forceful parts of artificial intelligence, machine learning based on computer algorithms enables fast and reliable predictions through data mining, which shows great potential in exploring high-efficiency catalysts. The application of SACs has covered a wide range of photocatalytic reactions, while the possibility of undiscovered single-atom photocatalysts is boundless, resulting in infinite combinations of structure\u2013activity relationship and great challenges in the rational design of single-atom photocatalysts. It is thus highly desirable to explore appropriate machine learning methods integrating with theoretical calculation data to predict the catalytic performance of single-atom photocatalysts and figure out the ideal configuration for targeted reactions. The implementation of machine learning will also construct reliant structure-performance relationships for the photocatalytic reactions, thereby enhancing the understanding of single-atom photocatalysis and promoting the rational development of efficient single-atom photocatalysts with highly applicable potential.This work received financial support from the King Abdullah University of Science and Technology (KAUST). X.W.L. acknowledges funding support from the Ministry of Education of Singapore via the Academic Research Fund (AcRF) Tier-2 grant (MOE2019-T2-2-049).The authors declare no competing interests.", "descript": "\n                  Artificial photocatalytic energy conversion represents a highly intriguing strategy for solving the energy crisis and environmental problems by directly harvesting solar energy. The development of efficient photocatalysts is the central task for pushing the real-world application of photocatalytic reactions. Due to the maximum atomic utilization efficiency and distinct advantages of outstanding catalytic activity, single-atom catalysts (SACs) have emerged as promising candidates for photocatalysts. In the current review, recent progresses and challenges on SACs for photocatalytic energy conversion systems are presented. Fundamental principles focusing on charge separation/transfer and molecular adsorption/activation for the single-atom photocatalysis are systemically explored. We outline how the isolated reactive sites facilitate the photogenerated electron\u2013hole transfer and promote the construction of efficient photoactivation cycles. The widespread adoption of SACs in diverse photocatalytic reactions is also comprehensively introduced. By presenting these advances and addressing some future challenges with potential solutions related to the integral development of photocatalysis over SACs, we expect to shed some light on the forthcoming research of SACs for photocatalytic energy conversion.\n               "}
{"full_text": "With the growing attachment to the environmental issues, the production of clean gasoline has become a major assignment in petroleum refining industry. To this end, the strict gasoline quality standards have been formulated by countries all over the world, limiting the contents of olefins and aromatics in gasoline. The decrease in the contents of high-octane olefins and aromatics will unavoidably cause the great reduction of gasoline octane number. n-alkane hydroisomerization is a process that converts low-octane straight chain paraffin into their high-octane branched alkanes (Qin et\u00a0al., 2011; Samad et\u00a0al., 2016; Kim et\u00a0al., 2018; Jaroszewska et\u00a0al., 2019). In this process, bifunctional catalysts containing metal sites and acid sites are often adopted in alkane hydroisomerization, and the metal sites provide the (de)hydrogenation active sites and acid sites provide skeletal isomerization active sites (Kim et\u00a0al., 2013; Liu et\u00a0al., 2015; Yu et\u00a0al., 2020). In general, the metal sites are usually supplied by precious metals (such as platinum and palladium) (Lv et\u00a0al., 2018; Oenema et\u00a0al., 2020.), non-precious metals (such as nickel and MoO\nx\n) (Yang et\u00a0al., 2019; Harmel et\u00a0al., 2020) and bimetallic nanoparticles (such as Pt\u2013Ni and Pt\u2013Ni2P) (Eswaramoorthi and Lingappan, 2003; Yao et\u00a0al., 2015), and the acid sites are normally offered by silica-alumina zeolites (such as ZSM-22, Beta, ZSM-5 and Y) (Chi et\u00a0al., 2013; Jin et\u00a0al., 2009; Lee et\u00a0al., 2013; Sazama et\u00a0al., 2018), silicoaluminophosphate (SAPO-n) (Zhao et\u00a0al., 2020; Guo et\u00a0al., 2021; Wen et\u00a0al., 2021), metal oxides (such as SO4\n2\u2212/ZrO2) (Kimura, 2003) and mesoporous materials (such as AlSBA-15 and Al-TUD-1) (Kang et\u00a0al., 2021; Vedachalam et\u00a0al., 2021). Among them, Pt has an excellent hydrogenation/dehydrogenation activity, and SAPO-11 possesses mild acidity and an appropriate pore structure for different processes (S\u00e1nchez-Contador et\u00a0al., 2018). Hence, the Pt/SAPO-11 catalysts are extensively used on alkane hydroisomerization (Zhang et\u00a0al., 2018a). However, mono-branched (Mo) isomers instead of multi-branched (Mu) isomers are the principal products for alkane hydroisomerization over Pt/SAPO-11 due to the small micropore size (0.39\u00a0\u00d7\u00a00.63\u00a0nm) and external surface area (ESA) of conventional SAPO-11. The Mu isomers with higher octane numbers than the Mo isomers are optimal components to improve gasoline octane number (Liu et\u00a0al., 2014). Researches show that the hierarchical SAPO-11 is able to effectively improve the selectivity to Mu isomers (Nandan et\u00a0al., 2014; Zhao et\u00a0al., 2019). Thus, it is of great significance to synthesize SAPO-11 molecular sieve with a hierarchical structure.There are many methods to synthesize hierarchical molecular sieves, such as desilication or dealumination, hard template strategy and soft template strategy (Jacobsen et\u00a0al., 2000; Fan et\u00a0al., 2012; Xi et\u00a0al., 2014; Chen et\u00a0al., 2016). However, in desilication or dealumination, the amount of removed silicon or aluminum cannot be controlled, and the excessive removal of framework silicon or aluminum will affect the crystallinity and stability of molecular sieves (Cartlidge et\u00a0al., 1989; Groen et\u00a0al., 2004). The soft template strategy is limited because the soft templates have a competition with the micropore templates to degrade the crystallization of molecular sieves and cause more environmental pollution in the process of removing soft templates. For the synthesis of hierarchical molecular sieves with a hard template strategy, the process is usually simple and the hard templates used do not interfere with the micropore templates. Thereby, the synthesis of hierarchical molecular sieves using hard templates has aroused extensive attention. Carbon materials, as a kind of hard templates, are commonly applied in the synthesis of hierarchical molecular sieves (Jacobsen et\u00a0al., 2000; Christensen et\u00a0al., 2005). However, the phase separation, which exists between carbon materials and synthesis precursor gel of hierarchical molecular sieves, is easy to occur, and which is attributed to the weak hydrophilicity of carbon materials (Machoke et\u00a0al., 2015). To improve the hydrophilicity of carbon materials. Zhao et\u00a0al. (2019) prepared carbon nanoparticles with abundant -C-O-C- groups by calcining a mixture of urea and polyethylene oxide in N2, and the -C-O-C- groups were converted into the -C-O-H groups in the alkali synthetic solution of hierarchical ZSM-5 molecular sieve. The existence of -C-O-H groups enhances the interaction between the carbon nanoparticles and ZSM-5 synthesis gel, and thus the hierarchical ZSM-5 with large mesopore size was successfully obtained. B\u00e9rtolo et\u00a0al. (2014) adopted a new process to obtain the hierarchical SAPO-11 through the use of commercial Merck carbon and Merck carbon with the treatment of nitric acid as mesoporogen, respectively. Merck carbon treated with nitric acid has more amount of oxygenate surface groups compared to the commercial Merck carbon, but the ESA and mesopore volume of the hierarchical SAPO-11 using Merck carbon with the treatment of nitric acid as the template are almost the same as those of the hierarchical SAPO-11 obtained by employing Merck carbon as mesoporogen.Metal-organic frameworks (MOFs) have aroused considerable attention because of their diversity of pore structures and high specific surface areas (Yang et\u00a0al., 2022; Zhang et\u00a0al., 2022) over recent years. The composites with the uniform mixture of carbon and metal oxides can be prepared through the heat treatment of MOFs (Li et\u00a0al., 2022; Liu et\u00a0al., 2022). Al-MOF-96, a typical porous aluminum based metal organic framework material, has the characteristics of simple synthesis and large-scale production (Deng and Peng, 2019), which can be used to prepare the composites with the uniform mixture of carbon and Al2O3. Additionally, Al2O3 can interact with phosphoric acid, thereby forming the aluminophosphate layers (Zhang et\u00a0al., 2018b). Thus, compared with conventional carbon materials, the Al2O3/C composite has better hydrophilicity in the phosphoric acid solution. However, the hierarchical molecular sieves prepared by using the metal oxide/carbon composite derived from MOFs as the mesoporogen has not been reported.In this work, a novel Al2O3/C composite derived from Al-MOF-96 was used to synthesize hierarchical SAPO-11. The above synthesized hierarchical SAPO-11 has more mesopores, a greater ESA and a higher number of medium Br\u00f8nsted acid centers (MBAC) than the conventional SAPO-11 (S-11) and the hierarchical SAPO-11 obtained employing activated carbon as the mesoporogen (CS-11). As a result, its corresponding catalyst displays enhanced selectivity to branched C10 isomers and low cracking selectivity in the n-decane hydroisomerization.Pseudoboehmite (73.0\u00a0wt% PB) was purchased from Shandong Aluminum Plant. Tetraethoxysilane (99.0\u00a0wt% TEOS), trimesic acid (98\u00a0wt% C6H3(CO2H)3), di-n-propylamine (99.5\u00a0wt% DPA), phosphoric acid (85.0\u00a0wt% H3PO4) and activated carbon (AR 200 mesh) were purchased from Aladdin. n-decane (98\u00a0wt% C10H22), aluminum nitrate (99\u00a0wt% Al(NO3)3\u00b79H2O) and chloroplatinic acid (37.0\u00a0wt% H2PtCl6\u00b76H2O) were provided by Innochem. Deionized water (H2O) was made in the laboratory. All reagents were employed directly as purchased.The synthesis procedures of conventional SAPO-11 were consistent with literature (Wen et\u00a0al., 2021), which is shown as detailed below: firstly, 12.2\u00a0g of H3PO4 and 40.0\u00a0g of H2O were mixed thoroughly, followed by adding 8.1\u00a0g of PB and evenly stirring for 3\u00a0h. Subsequently, 4.7\u00a0g of TEOS and 6.8\u00a0g of DPA were added to the suspension in turn and stirred for 3\u00a0h to afford the mixture with the composition of 40 H2O: 0.95 P2O5: 1.2 DPA: 0.4 TEOS: 1.0 Al2O3. Finally, the above synthesis mixture was poured into a 100\u00a0mL autoclave to crystallize at 200\u00a0\u00b0C for 24\u00a0h. The resultant sample was obtained after washing, drying and calcination at 600\u00a0\u00b0C for 6\u00a0h for the removal of DPA, and it was denoted as S-11.The preparation procedures of hierarchical SAPO-11 with Al2O3/C as mesoporogen were as follows. First, Al-MOF-96 was synthesized through the method described in literature (Liu et\u00a0al., 2015), and then the Al2O3/C composite was obtained after the calcination of Al-MOF-96 at 600\u00a0\u00b0C for 2\u00a0h in N2. Subsequently, 1.5\u00a0g of Al2O3/C composite was added to the mixture with the composition of 12.2\u00a0g of H3PO4 and 40.0\u00a0g of H2O and mixed thoroughly for 4\u00a0h. Afterwards, 7.0\u00a0g of PB and 4.7\u00a0g of TEOS were added in turn and evenly mixed for 5\u00a0h. Whereafter, 6.8\u00a0g of DPA was dropwise added to the solution and mixed for 1\u00a0h. Finally, the resulting mixture was put into a 100\u00a0mL autoclave and heated at 200\u00a0\u00b0C for 24\u00a0h. The as-synthesized sample was washed, dried and calcined at 600\u00a0\u00b0C for 6\u00a0h to remove DPA and mesoporogen, and it was denoted as ACS-11.The preparation procedures of hierarchical SAPO-11 employing activated carbon as mesoporogen were same as that of ACS-11 except that the activated carbon was utilized as mesoporogen rather than Al2O3/C. The obtained hierarchical SAPO-11 was named CS-11.Pt/SAPO-11 catalysts with the platinum mass fraction of 0.5% were obtained via incipient wetness impregnation. First, S-11, CS-11 and ACS-11 were pressed, and then they were sieved to 20\u201340 mesh. Subsequently, these catalysts were prepared by adding a solution of H2PtCl6 to the shaped SAPO-11 samples. Finally, these catalysts were got after drying and calcination at 450\u00a0\u00b0C for 4\u00a0h.A D8 Advance X-ray diffractometer (XRD) produced by Bruker AXS was used to analysis the sample structure (a Cu K\u03b1 radiation, 40\u00a0kV and 40\u00a0mA). A scanning electron microscopy (SEM) was performed to analyze the sizes and morphological properties using a SU8010. An ASAP 2420 physical adsorption equipment was adopted to perform the N2 adsorption-desorption measurement. The micropore volume (V\nmicro) and ESA of the sample were obtained based on the t-plot method (Lippens and Boer, 1965). The specific surface area (S\nBET) and the pore size distribution of the sample were obtained according to the Brunauer-Emmett-Teller method and the Barrett-Joyner-Halenda method individually (Barrett et\u00a0al., 1951).A F20 transmission electron microscopy (TEM) was adopted to obtain the TEM photos and the high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) photos of samples. A NICOLE-750 infrared instrument was adopted to measure the acid properties of the samples. The disk sample with 0.02\u00a0g was heated at 400\u00a0\u00b0C for 1\u00a0h, then the disk absorbed pyridine vapor for 0.25\u00a0h at room temperature. Afterwards, the vacuum desorption of pyridine was carried out at 200\u00a0\u00b0C and then 300\u00a0\u00b0C. Eventually, the pyridine infrared (Py-IR) spectrums at 200 and 300\u00a0\u00b0C were obtained at room temperature. X-ray fluorescence spectroscopy (XRF) was carried out on a ZSX-100e spectrometer to detect the elemental compositions of samples. An Avance III spectrometer with a 4\u00a0mm probe head was employed to analyze the Si coordinated environment of the synthesized samples, and the 29Si spectra were recorded at 79.5\u00a0MHz and a recycle delay of 3\u00a0s. The 29Si shifts are referenced to 3-(trimethylsilyl)-1-propanesulfonic acid sodium salt. H2 chemical adsorption was carried out to analyze the Pt dispersions of the catalysts by the Auto Chem II 2920 equipment.n-decane hydroisomerization was performed in a fixed-bed microreactor. First, the catalysts were reduced at 400\u00a0\u00b0C, 2.0\u00a0MPa and a H2 flow rate of 50\u00a0mL/min for 4\u00a0h. Afterwards, n-decane hydroisomerization was carried out at 340\u00a0\u00b0C, 2.0\u00a0MPa, a H2/n-decane volume ratio (HDVR) of 400 and different weight hourly space velocities (WHSVs) of 2.0\u201325 h\u22121. The products were qualitatively confirmed using a Trace 1310 mass spectrometry. What is more, the n-decane hydroisomerization products were analyzed adopting an SP3420 gas chromatograph, which contains an HP-PONA column (50\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm) and an FID detector.The n-decane conversion, the selectivity to total branched C10 isomers (S\nT), the selectivity to multi-branched C10 isomers (S\nMu), the cracking selectivity (S\nC), the total branched C10 isomers yield (Y\nT) and the multi-branched C10 isomers yield (Y\nMu) were calculated as follows:\n\n(1)\n\nC\no\nn\nv\ne\nr\ns\ni\no\nn\n=\n\n\n\nC\nr\n\n-\n\nC\nP\n\n\n\nC\nr\n\n\n\u00d7\n100\n%\n\n\n\n\n\n(2)\n\n\nS\nT\n\n=\n\n\nC\nT\n\n\n\nC\nr\n\n-\n\nC\nP\n\n\n\n\u00d7\n100\n%\n\n\n\n\n\n(3)\n\n\nS\nMu\n\n=\n\n\nC\nMu\n\n\n\nC\nr\n\n-\n\nC\nP\n\n\n\n\u00d7\n100\n%\n\n\n\n\n\n(4)\n\n\nS\nC\n\n=\n\n\nC\nC\n\n\n\nC\nr\n\n-\n\nC\nP\n\n\n\n\u00d7\n100\n%\n\n\n\n\n\n(5)\n\n\nY\nT\n\n=\n\n\nC\nT\n\n\nC\nr\n\n\n\u00d7\n100\n%\n\n\n\n\n\n(6)\n\n\nY\nMu\n\n=\n\n\nC\nMu\n\n\nC\nr\n\n\n\u00d7\n100\n%\n\n\nwhere C\nr represents the concentration of n-decane in the feedstock; C\nP and C\nT represent the n-decane concentration and total branched C10 isomers concentration in the products individually; C\nMu represents the multi-branched C10 isomers concentrations in products, and C\nC is cracking products concentration in products.n-decane hydroisomerization was considered as a pseudo-first-order reaction (Wen et\u00a0al., 2017). The rate constant (k) was calculated as detailed below:\n\n(7)\n\n\nk\n=\n\nF\nq\n\nln\n\n(\n\n1\n\n1\n\u2212\nc\n\n\n)\n\n\n\n\nwhere F represents the n-decane feed rate in mol s\u22121, q and c represent the quality of catalyst (g) and n-decane conversion individually.The formula adopted to calculate the catalyst turnover frequency (TOF) was as follows (Guo et\u00a0al., 2013):\n\n(8)\n\nT\nO\nF\n=\n\nf\n\nC\nB\n\n\n\n\nwhere f refers to the amount of reacted n-decane per gram of catalyst per second (mol/g/s). C\nB represents the numbers of MSBAC in the per catalyst quality (mol/g).\nFig.\u00a01\n displays the XRD spectra of S-11, CS-11 and ACS-11. All samples appear characteristic peaks at 2\u03b8\u00a0=\u00a08.1\u00b0, 9.5\u00b0, 13.1\u00b0, 15.7\u00b0, 20.4\u00b0, 21.1\u00b0 and 22.2\u201323.3\u00b0, which are ascribed to the AEL topology of SAPO-11 (Phienluphon et\u00a0al., 2015; Wen et\u00a0al., 2017). The XRD results show that the addition of activated carbon or Al2O3/C in the synthesis of SAPO-11 does not change the crystal structure of SAPO-11. Additionally, the XRD patterns of the synthesized SAPO-11 samples appear a weak peak at 2\u03b8\u00a0=\u00a06.6\u00b0 attributed to SAPO-41, which is commonly observed in SAPO-11 (Guo et\u00a0al., 2013; Wen et\u00a0al., 2021). In the synthesis of SAPO-11, the aluminum phosphate (AlPO4) precursors are first produced, and then Si substitution in the AlPO4 framework occurs. During this process, the slow formation rate of AlPO4 precursors and the fast releasing rate of silicon species lead to a change in the composition of partial liquid phases, which results in the formation of SAPO-41 (Ren et\u00a0al., 1991; L\u00f3pez et\u00a0al., 1997).\nFig.\u00a02\n presents the nitrogen adsorption-desorption isotherms of S-11, CS-11 and ACS-11. CS-11 and ACS-11 exhibit obvious hysteresis loops while S-11 presents a small hysteresis loop in the range of relative pressure (P/P\n0)\u00a0=\u00a00.4\u20130.9, showing that there are more mesopores in CS-11 and ACS-11. According to the SEM results (Fig.\u00a0S1), the addition of activated carbon or Al2O3/C could hinder the aggregation between SAPO-11 crystals, so CS-11 and ACS-11 have more intercrystal mesopores in comparison to S-11.The pore diameter distributions of S-11, CS-11 and ACS-11 are given in Fig.\u00a03\n. All samples present a peak at 3.8\u00a0nm, which is caused by the tensile strength effect (Danilina et\u00a0al., 2010). In addition, CS-11 and ACS-11 present high intensity peaks at 6.0 and 6.5\u00a0nm, respectively, and S-11 presents a low intensity peak at 5.5\u00a0nm, indicating that CS-11 and ACS-11 have abundant mesopores while S-11 has a relatively low amount of mesopores. Additionally, the texture properties of S-11, CS-11 and ACS-11 are exhibited in Table\u00a01\n. The mesopore volume and ESA of these samples rise in the order S-11\u00a0<\u00a0CS-11\u00a0<\u00a0ACS-11. In addition, as shown in the TEM images (Fig.\u00a0S2), ACS-11 has much more mesopores than S-11 and CS-11, which is in consistence with the nitrogen adsorption-desorption result.In order to explain the Al2O3/C function in the synthesis of hierarchical ACS-11, the Al2O3/C composite derived from Al-MOF-96 was treated with phosphoric acid according to the following procedures: Al2O3/C (1.5\u00a0g) was added to the solution of H2O (40\u00a0g) and phosphoric acid (12.2\u00a0g) and stirred for 4\u00a0h. Afterwards, the mixture was put into the autoclave and maintained at 200\u00a0\u00b0C for 24\u00a0h. The sample named H3PO4\u2013Al2O3/C was collected by washing and drying. And the H3PO4\u2013Al2O3/C and Al2O3/C were characterized by XRD.\nFig.\u00a04\n displays the XRD spectra of H3PO4\u2013Al2O3/C and Al2O3/C. Al2O3/C shows no diffraction peaks, while H3PO4\u2013Al2O3/C presents several characteristic peaks at 2\u03b8\u00a0=\u00a020\u201350\u00b0, which is in line with the PDF standard card of AlPO4 (PDF # 71\u20131041). The above results indicate that Al2O3 has an interaction with phosphoric acid, forming the AlPO4 structure in the process of treating Al2O3/C with phosphoric acid. Additionally, Fig.\u00a0S3 displays the HADDF-STEM result of Al2O3/C, Al2O3 is uniformly doped in carbon material.SAPO-11 crystals are easy to aggregate with each other due to the absence of mesoporogen in the synthesis of S-11. As a result, SAPO-11 with large crystallites is obtained (Liu et\u00a0al., 2014). For CS-11, activated carbon as mesoporogen is dispersed between SAPO-11 crystals, which reduces the contact of these crystals. As a consequence, CS-11 has smaller crystallites and more mesopores than S-11 (Yu et\u00a0al., 2021). However, activated carbon has weak hydrophilicity, leading to the phase separation between SAPO-11 gel and activated carbon, and thus activated carbon plays an inferior role as mesoporogen. During the preparation of hierarchical ACS-11, Al2O3 doped in carbon material has an interaction with phosphoric acid, forming AlPO4 structure. Thereby, the phase separation between Al2O3/C and SAPO-11 gel is avoided in the synthesis of ACS-11, and Al2O3/C effectively prevents the contact between SAPO-11 crystals. As a consequence, SAPO-11 with small crystallites is obtained. Thereby, ACS-11 has smaller crystallites and more mesopores than S-11 and CS-11.\nFig.\u00a05\n displays the Py-IR spectrums of S-11, CS-11 and ACS-11. The bands at 1455 and 1545\u00a0cm\u22121 correspondingly represent the Lewis acid centers (LAC) and Br\u00f8nsted acid centers (BAC), and the band at 1490\u00a0cm\u22121 can be ascribed to the cooperative action of LAC and BAC (Wen et\u00a0al., 2021). All samples show bands at 1545 and 1455\u00a0cm\u22121, indicating that they all have BAC and LAC.The amount of total B/L acid centers (TB/LAC) and the amount of medium B/L acid centers (MB/LAC) are calculated at 200 and 300\u00a0\u00b0C individually. The calculation formula is as follows (Fan et\u00a0al., 2006):\n\n(9)\n\n\nC\n\nB\n/\nL\n\n\n=\nA\nS\n/\nm\n\u03b5\n\n\nwhere C\nB/L and A are the amount of LAC or BAC per quality of samples (\u03bcmol g\u22121) and the absorbance (cm\u22121), respectively; S and m represent the cross-sectional area (cm2) and the mass (g) of samples, respectively; and \u03b5 is the extinction coefficient (cm \u03bcmol\u22121) (Datka, 1981). As shown in Table\u00a02\n, the amount of TBAC and the amount of MBAC over the SAPO-11 samples follow the order ACS-11\u00a0>\u00a0CS-11\u00a0> S-11. To explain these results, the contents and coordination environment of Si in S-11, CS-11 and ACS-11 are analyzed by XRF and 29Si MAS NMR.According to the results of XRF (Table\u00a0S1), S-11, CS-11 and ACS-11 show similar Si contents. Consequently, the acidities of these prepared SAPO-11 samples are mainly related to the coordination environment of Si atoms (Barthomeuf, 1994). The acidity of SAPO molecular sieves is generated by the substitution of the neutral AlPO4 framework by Si atoms. According to the different formats of Si substitution, it can be divided into SM2 and SM3. In the SM2 substitution method, Si(4Al) is formed through the replacement of one Si for one phosphorus. The silicon islands are formed by the replacement of an adjacent aluminum and phosphorus by two Si atoms in the SM3 substitution method, and Si(nAl,4-nSi) (0\u00a0<\u00a0n\u00a0<\u00a04) is formed at the borders of silicon islands (Barthomeuf, 1994). Small silicon islands are conductive to improve the acidity of SAPO-11, thereby enhancing the activity for alkane hydroisomerization (Yang et\u00a0al., 2017). Fig.\u00a06\n presents the 29Si MAS NMR results of S-11, CS-11 and ACS-11. All SAPO-11 samples exhibit five resonance peaks in the range of\u00a0\u221280 to\u00a0\u2212115\u00a0ppm. The structures of Si(4Al), Si(3Al,1Si), Si(2Al,2Si), Si(1Al,3Si) and Si(4Si) are centered at\u00a0\u221286,\u00a0\u221295,\u00a0\u2212101,\u00a0\u2212106 and\u00a0\u2212112\u00a0ppm, respectively (Fan et\u00a0al., 2012).The proportions of different Si species in SAPO-11 are shown in Table\u00a03\n. The size of silicon islands decreases following the sequence S-11\u00a0>\u00a0CS-11\u00a0>\u00a0ACS-11. Consequently, the amount of MBAC of ACS-11 is the maximum among these samples, which corresponds to the results of Py-IR. According to the results of Py-IR and 29Si MAS-NMR, the addition of activated carbon or Al2O3/C increases the amount of MBAC of SAPO-11. This is because Si atoms are promoted to enter the SAPO-11 framework with the existence of activated carbon, which improves Si distribution in the SAPO-11 framework (Yu et\u00a0al., 2021). As a result, the amount of MBAC of CS-11 is larger than that of S-11. Compared with activated carbon, Al2O3/C is better dispersed between SAPO-11 crystals and thus further improves Si distribution in SAPO-11 framework. Consequently, ACS-11 possesses the largest amount of MBAC among these samples.The n-decane hydroisomerization over Pt/S-11, Pt/CS-11 and Pt/ACS-11 are evaluated, and the evaluation results over these catalysts at 300\u2013360\u00a0\u00b0C, 2.0\u00a0MPa, a weight hourly space velocity (WHSV) of 2.0 h\u22121 and a HDVR of 400 are presented in Fig.\u00a0S4, which shows that 340\u00a0\u00b0C is considered as the optimal reaction temperature for n-decane hydroisomerization over these Pt/SAPO-11 catalysts. Additionally, the n-decane hydroisomerization performances over these catalysts at 340\u00a0\u00b0C, different WHSVs of 2.0\u201325 h\u22121, 2.0\u00a0MPa and a HDVR of 400 are displayed in Fig.\u00a07\n. The n-decane conversion of these catalysts decreases in the wake of the increase of WHSV (Fig.\u00a07(a)). As the n-decane conversion rises, S\nT decreases (Fig.\u00a07(b)) and both S\nMu and S\nC increase for all catalysts (Fig.\u00a07(c) and (d)). The n-decane conversion at the same WHSV follows this order Pt/ACS-11\u00a0>\u00a0Pt/CS-11\u00a0>\u00a0Pt/S-11; S\nT and S\nMu at the same n-decane conversion decrease in the same sequence, and the S\nC increases following the sequence Pt/ACS-11\u00a0<\u00a0Pt/CS-11\u00a0<\u00a0Pt/S-11 in the whole range of n-decane conversion. However, the difference of S\nT and S\nC over these catalysts is small under the low n-decane conversion, while the gap of them gradually rises with the increasing n-decane conversion. According to the typical bifunctional reaction mechanism of n-alkane hydroisomerization, the n-decane isomerization process can be expressed as: n-decane \u2192 mono-branched C10 isomers \u2192 multi-branched C10 isomers, and the isomerization process is accompanied by cracking reactions (Deldari, 2005). In this work, the low n-decane conversion is due to the high WHSV, which results in the short residence time of reactants at the active sites of catalysts (Singh et\u00a0al., 2014). As a result, the cracking selectivity is low and the C10 isomers selectivity is high over the three catalysts. Therefore, there is a little difference in the S\nT and S\nC over the three catalysts under the low n-decane conversion (Yang et\u00a0al., 2019; Wen et\u00a0al., 2020). However, with the decrease in the WHSV, the n-decane conversion rises, and the difference in the S\nT and S\nC over the three catalysts gradually increases. ACS-11 has more mesopores, smaller crystallite size and larger ESA than S-11 and CS-11, which promote the formation and diffusion of isomerized intermediates and suppress cracking reactions. Therefore, there is an obvious difference in S\nT and S\nC between Pt/ACS-11 and the counterparts under the high n-decane conversion.The n-decane hydroisomerization results over the Pt/SAPO-11 catalysts are shown in Table\u00a04\n. The primary n-decane hydroisomerization products on Pt/S-11, Pt/CS-11 and Pt/ACS-11 are 2-methylnonane (2-MC9), 3-ethyloctane (3-EC8), 3-methylnonane (3-MC9), 4-ethyloctane (4-EC8), 4-methylnonane (4-MC9), 5-methylnonane (5-MC9), 2,5-dimethyloctane (2,5-DMC8), 3,5-dimethyloctane (3,5-DMC8), 4,5-dimethyloctane (4,5-DMC8), 2,6-dimethyloctane (2,6-DMC8), 3,6-dimethyloctane (3,6-DMC8), 2,7-dimethyloctane (2,7-DMC8), 4,4-dimethyloctane (4,4-DMC8), 3,3-dimethyloctane (3,3-DMC8), 2-methyl-3-ethylheptane (2-M-3-EC7), 3-methyl-3-ethylheptane (3-M-3-EC7) and 2-methyl-5-ethylheptane (2-M-5-EC7). S\nMu of Pt/ACS-11 is 23.28% at the n-decane conversion of approximately 92%, which is higher than those of Pt/S-11 (18.38%) and Pt/CS-11 (19.53%), and S\nC of Pt/ACS-11 (15.83%) is lower than those of Pt/S-11 (22.95%) and Pt/CS-11 (19.09%). Additionally, Pt/ACS-11 has a superior stability for n-decane hydroisomerization, which is presented in Fig.\u00a0S5. The rate constant (k) and turnover frequency (TOF) values of Pt/ACS-11 obtained at the n-decane conversion of 20% are 15.25\u00a0\u00d7\u00a010\u22126\u00a0mol\u00a0g\u22121 s\u22121 and 25.69\u00a0\u00d7\u00a010\u22121 s\u22121 individually, which is higher than those of Pt/S-11 (7.84\u00a0\u00d7\u00a010\u22126\u00a0mol\u00a0g\u22121 s\u22121 and 20.43\u00a0\u00d7\u00a010\u22121 s\u22121) and Pt/CS-11(10.19\u00a0\u00d7\u00a010\u22126\u00a0mol\u00a0g\u22121 s\u22121 and 22.84\u00a0\u00d7\u00a010\u22121 s\u22121).The n-decane hydroisomerization over Pt/SAPO-11 catalysts follows a typical bifunctional reaction mechanism (Martens et\u00a0al., 1986; Deldari, 2005; Zhang et\u00a0al., 2019), and the details are displayed in Fig.\u00a0S6. Firstly, n-decane is dehydrogenated over platinum to produce corresponding n-decene intermediates; then, these intermediates are quickly transferred to the BAC inside the pore of SAPO-11 and occur a skeletal rearrangement reaction to form mono-branched C10 intermediates. Based on the theory of \u201cpore mouth/key-lock\u201d (Zhang et\u00a0al., 2018a; Liu et\u00a0al., 2020), one side of the mono-branched C10 intermediates can be adsorbed on the pore mouth of SAPO-11, and another side is adsorbed on the BAC of the adjacent pore mouth to undergo skeletal isomerization and generate multi-branched C10 intermediates. These n-decane branched intermediates can also be cracked on the BAC of SAPO-11. Finally, these branched n-decane intermediates are transferred on the Pt metal sites for hydrogenation to produce C10 isomers.The loadings amount of Pt in these Pt/SAPO-11 catalysts are 0.5\u00a0wt%, which meet the hydrogenation-dehydrogenation requirements of alkanes hydroisomerization (Wen et\u00a0al., 2021). Fig.\u00a08\n presents the TEM pictures and Pt particle size distributions of Pt/S-11, Pt/CS-11 and Pt/ACS-11. The average Pt particle size and dispersions over Pt/S-11 and Pt/CS-11 are both 4.3\u00a0nm and 56%, respectively, and which over Pt/ACS-11 are 4.2\u00a0nm and 57% individually. The nearly same particle size and dispersion of Pt suggests that the performance of n-decane hydroisomerization over these catalysts chiefly depends on the physicochemical properties of S-11, CS-11 and ACS-11.\nFig.\u00a09\n exhibits the forming schematics of the of branched n-decane isomers over Pt/SAPO-11 with different crystallites. SAPO-11 with a small amount of MBAC and large crystallites offers few active sites and long diffusion path in n-decane hydroisomerization, which is disadvantageous for the generation and diffusion of isomerized intermediates and products (Noh et\u00a0al., 2018; Oenema et\u00a0al., 2020). S-11 with less MBAC and larger crystallites than CS-11 and ACS-11 provides a smaller number of hydroisomerization sites and a longer residence time for isomerized intermediates and products over Pt/S-11 than Pt/CS-11 and Pt/ACS-11 (Fig.\u00a09(a)). Consequently, Pt/S-11 presents low S\nT and high S\nC. Fig.\u00a09(b) shows that SAPO-11 with small crystallites and a large amount of MBAC provides a great amount of hydroisomerization sites and a short diffusion path in n-decane hydroisomerization, which enhances the generation of branched isomers and reduces the cracking side reactions. Pt/ACS-11 presents the maximum n-decane conversion in n-decane hydroisomerization among these catalysts. This can be attributed to the fact that ACS-11 has the largest amount of MBAC among these SAPO-11 samples, which provides a large number of active sites for the n-decane hydroisomerization. Additionally, Pt/ACS-11 shows the maximum S\nT and the minimum S\nC among these catalysts. This is because ACS-11 has the smallest crystallites and the largest mesopore volume among these samples, which makes the isomerized C10 intermediates easy to diffuse out the pores of ACS-11 and thereby reduces the cracking reactions and enhances the production of C10 isomers. Furthermore, based on the \u201cpore mouth/key-lock\u201d mechanism, the multi-branched alkane isomers are generated at the ESA rather than in the SAPO-11 channels due to their larger diameters than in the pore openings of SAPO-11 (Claude and Martens, 2000), and a high ESA for SAPO-11 favors the formation of multi-branched C10 intermediates. Thus, ACS-11 with higher ESA than S-11 and CS-11 endows the corresponding catalyst with the maximum S\nMu among the three catalysts. Additionally, the maximum Y\nT and Y\nMu of Pt/ACS-11 are 77.85% and 21.44% individually, which is higher than those of Pt/S-11 (73.21% and 16.75%) and Pt/CS-11(73.85% and 17.19%) (Fig.\u00a0S7), and also higher than those of the reported catalysts in literatures (Table\u00a0S2).Hierarchical SAPO-11 molecular sieve was prepared using the Al2O3/C composite derived from Al-based metal-organic framework as mesoporogen. The Al2O3/C reacts with phosphoric acid to generate the AlPO4/C structure during the synthesis of hierarchical SAPO-11, which efficiently disperses the Al2O3/C in the synthesis gel of SAPO-11 and thereby inhibits the aggregation of SAPO-11 crystals. Consequently, the optimal hierarchical SAPO-11 is obtained with smaller crystallites, a bigger mesopore volume (0.13\u00a0cm3\u00a0g\u22121) and a greater amount of MBAC (26.6\u00a0\u03bcmol g\u22121) than the conventional SAPO-11 and the hierarchical SAPO-11 employing activated carbon as mesoporogen. Additionally, its corresponding catalyst displays the maximum selectivity to multi-branched C10 isomers (23.28%), the minimum cracking selectivity (15.83%) and a superior stability in n-decane hydroisomerization among the prepared catalysts. This work provides a new route for preparing hierarchical silicoaluminophosphate molecular sieve-based catalysts with superior alkane hydroisomerization performances.The authors gratefully acknowledge the financial support of Science Foundation of China University of Petroleum, Beijing (Grant No. KYJJ2012-03-03).The following are the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\n\n\nMultimedia component 2\nMultimedia component 2\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.petsci.2022.06.003.", "descript": "\n                  Hierarchical SAPO-11 molecular sieve (ACS-11) was successfully synthesized employing the Al2O3/carbon (Al2O3/C) composite obtained through the pyrolysis of Al-based metal-organic framework (Al-MOF-96) as mesoporogen. Unlike other carbon-based mesoporogens with strong hydrophobicity, the Al2O3/C interacts with phosphoric acid and generates the AlPO4/C structure, which promotes the Al2O3/C dispersion in the synthesis gel of SAPO-11 and avoids the phase separation between them. The Al2O3/C as mesoporogen decreases the crystallite size of SAPO-11 via preventing the aggregation of SAPO-11 crystals. Additionally, the addition of Al2O3/C improves the Si distribution in the ACS-11 framework. Consequently, ACS-11 has smaller crystallites, more mesopores, and a greater amount of medium Br\u00f8nsted acid centers than the conventional microporous SAPO-11 and the SAPO-11 synthesized using activated carbon as mesoporogen. The corresponding Pt/ACS-11 catalyst exhibits the maximal selectivity to multi-branched C10 isomers (23.28%) and the minimal cracking selectivity (15.83%) in n-decane hydroisomerization among these catalysts. This research provides a new approach for preparing hierarchical silicoaluminophosphate molecular sieve-based catalysts to produce high-quality fuels.\n               "}
{"full_text": "Second generation biomass is a renewable feedstock with a lower carbon footprint than fossil crudes and is thus a promising alternative raw material to produce fuels and chemicals [1\u20133]. One of the interesting biomass-derived platform molecules is furfural (Scheme 1\n), which can be produced on an industrial scale by dehydration of hemicellulose from agricultural waste and forest residue [4\u20137] and has been identified as one of the numerous oxygenated compounds in pyrolysis bio-oils [8\u201314].Currently, the most relevant use of furfural as a chemical feedstock is in the production of furfuryl alcohol (FOL) and tetrahydrofurfuryl alcohol (THFOL) from selective hydrogenation, which have significant applications in furan resins industry as solvents or chemical intermediates [5,15\u201321]. Besides, it is also feasible to obtain furan and tetrahydrofuran (THF) via decarbonation and hydrogenation as an environmental benign industrial solvent in the synthesis of plastics, pharmaceuticals and agrochemicals [5,22,23]. Furthermore, the production of biofuels from furfural has received extensive attention during the last decades. Aldol-condensation of furfural with acetone, self-condensation of furfural, or condensation of intermediate products (e.g. furan, 2-methylfuran) with other substrates are promising approaches to produce higher alkanes (C8-C15) for diesel fuel applications [20,24\u201326]. Selective HDO of furfural, on the other hand, yields 2-methylfuran (MF) or tetrahydro-2-methylfuran (THMF) which are useful blending components for gasoline [6,27,28].In the present study, we focus on the production of MF from furfural HDO, because the involved hydrogen consumption is lower, while the carbon yield is higher compared with products like THMF, THFOL, and furan [5,6]. Moreover, MF has favorable properties in fuel applications. Cu-based catalysts have been found being active in the MF production [29\u201332]. However, Cu-Cr catalysts [29] used in the industrial MF production are highly toxic, while Cu supported on SiO2, Al2O3 or ZnO [30] suffers from sintering during long reaction periods. Group 9\u201310 metal catalysts (Ni [5,33], Co [33], Pt [13,34] and Pd [4,5,35]) are also active in the MF production. However, the strong interaction between the furan ring and the transition metal surface leads to undesired ring-hydrogenation or ring-opening reactions; the unstable \u03b72(C, O) adsorbate configuration tends to turn into an \u03b71(C) configuration at high temperature enabling undesired decarbonylation reactions [5].To optimize MF production over pure transition metals (e.g. Ni, Pt), more oxophilic metals (e.g. Fe, Zn) have been added to design alloy catalysts (e.g. PtZn [7] and NiFe [6,27]) for furfural HDO. The addition of more oxophilic elements provides more stable \u03b72(C, O) configurations to the catalyst surface, in which the furan ring is tilted away from the catalyst surface and the carbonyl group is bonded stronger to the catalyst surface by the additional interaction of the carbonyl O to the oxophilic metal [6,7]. Such \u03b72(C, O) configurations suppress furan-ring hydrogenation or ring-opening, and enhance the conversion of the carbonyl group by weakening the C1O1 bond (Scheme 1) [6,7]. By adjusting the Fe loading from 0 to 5\u00a0wt% in FeNi/SiO2 catalysts at a constant Ni loading of 5wt%, the MF yield was improved from 20% for Ni/SiO2 to 80% for 5/5\u00a0wt% NiFe/SiO2\n[6]. Besides alloying metals, ceramic materials like Mo2C [36\u201339] have also been explored for the furfural to MF conversion. Mo2C shows a high MF selectivity of 70% [36\u201338], but suffers from rapid deactivation with the furfural conversion decreasing from 90% to 18% over a period of 3\u00a0h [38]. Metal borides like NiB, CoB and NiPB [14,19,40,41] have also been considered as catalysts in the furfural conversion. The electron-deficient B can attract the O atom of the carbonyl group, which weakens the CO bond and thus promotes its hydrogenation to the alcohol [14,19]. Nevertheless, only FOL, instead of MF, has been obtained as main products on these metal boride catalysts.Metal phosphides have extensively been investigated in hydrodesulfurization (HDS) [42\u201344] and hydrodenitrogenation (HDN) [44\u201348] reactions in recent years. Due to their promising catalytic performance, they are also increasingly applied in hydrodeoxygenation (HDO) reactions [49\u201352]. So far, there was a focus on investigating HDO model compounds[53], such as phenol [50,54\u201356], anisole [2,56], or guaiacol [1,57\u201362], furfural [63\u201365] and some aliphatic model compounds [51,66\u201369]. However, studies of the furfural HDO over metal phosphides were limited to Ni2P-based catalysts. Hence, we decided to explore the catalytic performance of different metal phosphides in the furfural HDO reaction with a special interest in obtaining MF as the most desired product. In addition to changing the metal component, we varied the phosphorus/metal ratios in our samples and explored catalytic properties in response to the changed P/M stoichiometries for both Ni and Co. Previous literature reports the influence of the P/Ni ratio on the performance of nickel phosphides in HDS, HDN and HDO [70\u201373]. However, only the relationship between the active nickel phosphide phase and the catalytic performance has been established, while the presence of P is considered to only determine or maintain the composition of the active phases. In the present work, we provide deeper insight into how the P/Ni ratio influences in each of the active phases the catalytic performance during furfural HDO, by means of furfural-IR and CO-IR spectroscopy.Commercial silica (silica gel Davisil, particle size 90\u2013125 um, SBET\u00a0=\u00a0305\u00a0m2/g) was used as support material. H3PO3 (Aesar, 98%), Ni(NO3)2\u00b76H2O (Aldrich, 97%), (NH4)6Mo7O24\u00b74H2O (VWR, 99%), (NH4)W12O39\u00b7H2O (Aldrich, 99%), Co(NO3)2\u00b76H2O (Aldrich, 99%), Fe(NO3)2\u00b79H2O (Aldrich, 99%), Cu(NO3)2\u00b73H2O (Aldrich, 99%) were used to prepare catalyst samples of Ni2P, MoP, WP, Co2P, Fe2P or Cu3P.Metal and phosphorus species were loaded on silica by impregnation. For NiO-P/SiO2, MoO3-P/SiO2, WO3-P/SiO2, Fe2O3-P/SiO2, Co3O4-P/SiO2, and CuO-P/SiO2 samples, metal precursors were first loaded on silica support by incipient wetness impregnation. After drying overnight at 110 \u00b0C in air and calcination at 550 \u00b0C for 1\u00a0h, the silica supported metal oxides were impregnated with H3PO3 by incipient wetness impregnation and then air-dried at 110 \u00b0C overnight. Before reaction or characterization, the prepared samples were reduced for 2\u00a0h in hydrogen at the desired reduction temperature as determined by temperature programmed reduction measurements (TPR) with a heating rate of 5 \u00b0C/min.The metal loading is 1.5\u00a0mmol/g SiO2; the atomic ratios of P/M (phosphorus to metal) are 2 for Ni2P; 1 for MoP; 1.5 for WP; 0.50 for Fe2P, Co2P and Cu3P, which were empirically found to be needed for the formation of the respective phases. The unreduced precursors are referred to as NiO-P/SiO2, MoO3-P/SiO2, WO3-P/SiO2, Fe2O3-P/SiO2, Co3O4-P/SiO2, and CuO-P/SiO2. A H3PO3/SiO2 sample with a phosphorus loading of 3\u00a0mmol/g was prepared as reference sample. NixMoyP/SiO2 catalysts with atomic Ni/Mo ratios of 1:1 and 2:1 were prepared similarly by co-impregnation with a solution containing both Ni and Mo. The loading of Ni/Mo/P on the NixMoyP/SiO2 samples was 1.5/1.5/3 or 1.5/0.75/3 mmol (for the elements Ni/Mo/P, respectively) per gram of SiO2.NiP(x)/SiO2 and CoP(x)/SiO2 were prepared by incipient wetness impregnation as in our previous work [42]. Catalysts were prepared by a one-step impregnation of metal nitrates and H3PO3 followed by direct reduction of as-prepared precursors. The molar P/Ni ratio was varied from 0 to 2 (0, 0.33, 0.5, 1.0, 2.0) in the precursors. Catalyst with different P/Ni ratio were labeled as Ni/SiO2, NiP(0.33)/SiO2, NiP(0.5)/SiO2, NiP(1)/SiO2, and NiP(2)/SiO2, respectively. Molar ratios of P/Co were varied from 0 to 2 (0, 0.5, 1.0, 1.5, 2.0) in the catalyst precursors. Catalysts are denoted as Co/SiO2, CoP(0.5)/SiO2, CoP(1.0)/SiO2, CoP(1.5)/SiO2, and CoP(2.0)/SiO2, respectively.The metal loading was determined by inductively coupled plasma atomic emission spectrometry (ICP-AES) analysis performed on a Spectro Blue apparatus. Prior to analysis, samples were dissolved in an HF/HNO3/H2O solution with a volumetric ratio of 1:1:1. HF (VWR, 40%) and HNO3 (VWR, 65%) were used to prepare the acidic solvent for ICP measurements.X-ray diffraction (XRD) patterns were acquired on a Bruker D2 Phaser powder diffraction system using Cu K\u03b1 radiation (1.5406\u00a0\u00c5). Scans were taken at a rate of 1\u00b0/min in the range of 10\u00b0 \u2264 2\u03b8\u00a0\u2264\u00a080\u00b0.BET surface area measurements were performed on a Micromeritics ASAP3020 Tristar system and a nitrogen stream was applied for N2 physisorption at \u2212196 \u00b0C. The specific surface area was determined by the Brunauer-Emmett-Teller (BET) equation using the adsorption data.X-ray photoelectron spectroscopy (XPS) measurements were performed on a Thermo Scientific K-Alpha XPS apparatus equipped with a monochromatic Al K\u03b1 X-ray source. Samples were reduced and sealed in a tubular quartz reactor, transferred into a glovebox without exposure to air, then loaded in an airtight transfer vessel and introduced into the XPS analysis chamber for analysis. The background pressure prior to analysis was 2\u00a0\u00d7\u00a010\u22129 mbar. Survey scans were collected at constant pass energy of 200\u00a0eV, and region scans at 50\u00a0eV. The spectra were calibrated to the Si 2p line at 103.3\u00a0eV and fitted with the CasaXPS program.Temperature programmed reduction measurements in H2 (H2-TPR) were carried out between 50 \u00b0C and 700 \u00b0C. Typically, 15\u00a0mg catalyst precursor was heated in a 20\u00a0ml/min H2 flow from 50 \u00b0C to 700 \u00b0C at a heating rate of 5 \u00b0C /min. The effluent gas composition was analyzed by an online mass spectrometer. The mass signals of 2 (H2), 18 (H2O), 34 (PH3), and 30 (NO) were monitored.Temperature programmed desorption of NH3 (NH3-TPD) was conducted between 25 \u00b0C and 400 \u00b0C. Typically, 15\u00a0mg catalyst precursor was reduced at 550 \u00b0C for 1\u00a0h and cooled down to 25 \u00b0C in a 20\u00a0ml/min flow of H2. Afterwards, the reduced samples were first flushed with He for 30\u00a0min to remove physisorbed species, then with NH3/He (10 v/v%) for 30\u00a0min to adsorb NH3, followed by He for 30\u00a0min to remove weakly absorbed NH3. Subsequently, the pretreated samples were heated from 25 \u00b0C to 400 \u00b0C in a 20\u00a0ml/min He flow with a heating rate of 5 \u00b0C /min. The effluent gas flow was monitored by a TCD detector.Infrared (IR) spectroscopy of CO absorbed samples were carried out with a Bruker Vertex V70v Fourier Transform IR spectrometer equipped with a DTGS detector. The catalysts were pressed into a self-supporting wafer with a diameter of 13\u00a0mm and then mounted in an in-situ cell equipped with CaF2 windows. Prior to CO adsorption, the samples were reduced at 500 \u00b0C for 2\u00a0h, applying a heating rate of 10 \u00b0C/min, followed by evacuating the cell to a pressure below 2\u00a0\u00d7\u00a010\u22126 mbar for 0.5\u00a0h at 500 \u00b0C. Afterwards, the reduced sample was cooled down to 25 \u00b0C under evacuation. CO was introduced into the cell with a metering valve until the CO pressure reached 5\u00a0mbar. The catalyst was exposed to CO for 5\u00a0min, subsequently evacuated for 5\u00a0min to 2\u00a0\u00d7\u00a010\u22126 mbar, followed by heating to 425 \u00b0C at a rate of 10 \u00b0C /min under evacuation. The IR spectrum was recorded by accumulating 64 scans at a resolution of 2\u00a0cm\u22121. Spectra of the freshly reduced as well as CO saturated catalysts were recorded during the heating process.In case of furfural-IR, catalyst pretreatment and experimental parameters were the same as those of the CO-IR experiments. Liquid furfural was expanded into the evacuated cell at a pressure of 2\u00a0\u00d7\u00a010\u22123 mbar for 5\u00a0min at 25 \u00b0C and then evaporated until a pressure of 2\u00a0\u00d7\u00a010\u22126 mbar was reached. Afterwards, the catalyst was heated to 300 \u00b0C at a rate of 10 \u00b0C/min at 2\u00a0\u00d7\u00a010\u22126 mbar. Spectra of the furfural-saturated catalysts were recorded during the heating process. Spectra of freshly reduced catalysts were taken to use as background correction reference.The particle size of active Ni-phosphide phase was investigated with a FEI Tecnai 20 transmission electron microscopy (TEM).The catalytic performance of synthesized materials was evaluated in a plug flow fixed-bed reactor with 4\u00a0mm internal diameter. Furfural was fed by a syringe pump (Hewlett Packard 1050) and gasified at 170 \u00b0C using H2 as carrier gas.Prior to reaction, catalysts were reduced in-situ at the desired temperature in a hydrogen flow for 2\u00a0h. After reduction, the reactor was cooled to reaction temperature, then fed with a furfural/H2 flow (H2/furfural molar ratio\u00a0=\u00a074). The products were analyzed by an in-line gas chromatograph using a flame ionization detector (GC-FID) and a DB-1 (30\u00a0m, 0.32\u00a0mm, 1.00\u00a0\u03bcm) column.The reaction was carried out at atmospheric pressure between 120 \u00b0C and 200 \u00b0C. In a typical run, the H2 flow, the feeding rate of furfural, H2 and the loading amount of catalysts were 20\u00a0ml/min, 0.001\u00a0ml/min in liquid phase, and 60\u00a0mg, respectively. This corresponds to a weight hourly space velocity (WHSV) of ca. 3\u00a0h\u22121. All activity measurements were carried out in duplicate to verify reproducibility. The product gas stream was analyzed after 2\u00a0h on stream.Conversion and selectivity were calculated as follows:\n\n\n\nC\no\nn\nv\ne\nr\ns\ni\no\nn\n\n(\n%\n)\n\n=\n\n\nmol\n\no\nf\n\nt\nh\ne\n\np\nr\no\nd\nu\nc\nt\ns\n\n\nmol\n\no\nf\n\nf\nu\nr\nf\nu\nr\na\nl\n\nf\ne\nd\n\n\n\u00d7\n100\n\n\n\n\n\n\n\n\nSelectivity\n\n(\n%\n)\n\n=\n\n\nmol\n\no\nf\n\no\nn\ne\n\np\nr\no\nd\nu\nc\nt\n\n\nmol\n\no\nf\n\na\nl\nl\n\np\nr\no\nd\nu\nc\nt\ns\n\n\n\u00d7\n100\n\n\n\n\nThe metal and phosphorus content of unreduced precursors as determined with ICP and the textural properties are listed in Table S1. Each precursor shows a P/metal molar ratio close to the target value. BET surface areas are between 150\u00a0m2/g and 250\u00a0m2/g, which are all lower than the surface area of bare SiO2 (305\u00a0m2/g) and decrease with increasing P content due to pore blocking [72].Temperature programmed reduction (TPR) was used (Figure S1) to determine the right reduction temperature for each catalyst. As revealed by Figure S1, the NiO-P/SiO2, WO3-P/SiO2 and MoO3-P/SiO2 samples are reduced at 500 \u00b0C, 600 \u00b0C and 550 \u00b0C, respectively. A higher reduction temperature (650 \u00b0C) is required to reduce Co3O4-P/SiO2 and CuO-P/SiO2, while Fe2O3-P/SiO2 reduces at 680 \u00b0C.In case of the NiP(x)/SiO2 catalysts (Figure S1b), the reduction temperature is determined as 450 \u00b0C for Ni/SiO2, and 550 \u00b0C for NiP(x) (x\u00a0=\u00a00.33, 0.5, 1, 2). Although the TPR profiles of NiP(1) and NiP(2) show the most significant signals at 650 \u00b0C as a result of PO4\n3\u2212 reduction, we have chosen 550 \u00b0C as reduction temperature since P species formed by reduction of excessive H3PO3 suffice for nickel phosphide formation.In the case of CoP(x)/SiO2 catalysts (Figure S1c), Co/SiO2 is reduced at 400 \u00b0C, while a high temperature of 650 \u00b0C is required for CoP(x) reduction.X-ray diffraction (XRD) reveals that each catalyst formed crystalline phases after reduction (Fig. 1\n). All catalysts show a broad diffraction peak in the 2\u03b8 range of 15\u201335\u00b0, typical of amorphous silica [60,74,75]. In Fig. 1a, all reduced samples contain pure phases of the corresponding metal phosphides (i.e. Ni2P, MoP, Co2P, Fe2P, and Cu3P) except WP/SiO2, which displays reflections of both WP and unphosphided W0 phases. The sharp diffraction peaks of WP/SiO2, Fe2P/SiO2, Co2P/SiO2 and Cu3P/SiO2 indicate that the metal phosphides consist of relatively large particles due to the rather high reduction temperature. For MoP/SiO2 and Ni2P/SiO2, milder phosphidation conditions were applied and consistently the XRD patterns show broad diffraction peaks, suggesting smaller particles.\nFig. 1b shows XRD patterns of reduced NiP(x)/SiO2 (x\u00a0=\u00a00, 0.33, 0.5, 1.0, and 2.0). Pure metallic Ni0 is formed in Ni/SiO2. Ni3P is obtained on NiP(0.33), Ni12P5 on NiP(0.5), and Ni2P on NiP(1) and NiP(2). FWHM indicates comparable particle sizes.\nFig. 1c shows that reduced CoP(x)/SiO2 (x\u00a0=\u00a00, 0.5, 1.0, 1.5, and 2.0) contains a pure Co0 phase on reduced Co/SiO2. As the P/Co ratio increases, the active phase shifts from Co2P to CoP with Co2P present in CoP(0.5), CoP, CoP(1.5), and CoP(2.0), and both phases in CoP(1).Different from NixPy (Ni2P, Ni3P, Ni12P5) and CoxPy (Co2P, CoP), only one single phosphide phase has been obtained for Mo, Fe, Cu, and W. Hence we explored the effect of metal/P ratio on the catalytic performance only for nickel and cobalt phosphides.Reduced Ni, NiP(0.5), NiP(1) and NiP(2) samples were also characterized by TEM to determine the particle size distributions (Fig. 2\n). Besides some large particles observed in the Ni catalyst, all catalyst show small nanoparticles of a comparable size of\u00a0~\u00a05\u00a0nm, consistent with the XRD data for the NiP(x) catalysts.The XPS results reveal the type and amount of chemical species on the surface of reduced catalysts, which confirm the formation of metal phosphides.\nFig. 3\n and Table S2 show the results of the reduced NiP(x) (x\u00a0=\u00a00, 0.33, 0.5, 1, and 2) catalysts. The contribution at 852.2\u00a0eV in Ni 2p spectra is assigned to Ni0 in metallic nickel, the contribution at 852.7\u00a0eV is assigned to Ni\u03b4+ (0\u00a0<\u00a0\u03b4\u00a0<\u00a01) in nickel phosphide phases (Ni2P, Ni5P2 and Ni3P), and the signal at 856.4\u00a0eV to Ni2+ species [42]. Concerning the P 2p spectra, the three peaks at 134.4\u00a0eV, 133.5\u00a0eV and 129.1\u00a0eV are due to PO4\n3\u2212, PO3\n3\u2212 and P\u03b4\u2212 (0\u00a0<\u00a0\u03b4\u00a0<\u00a01) species in nickel phosphides [42], respectively. PO3\n3\u2212 originates from unreduced H3PO3, while PO4\n3\u2212 forms from the disproportionation or reduction of H3PO3\n[42]. The presence of Ni\u03b4+ (0\u00a0<\u00a0\u03b4\u00a0<\u00a01) and P\u03b4\u2212 (0\u00a0<\u00a0\u03b4\u00a0<\u00a01) in nickel phosphide phases points to an interphase Ni\u00a0\u2192\u00a0P transfer of electron density as a consequence of the electron withdrawing nature of P [76]. The electron density on Ni\u03b4+ is expected to decline at higher P content due to the electron withdrawing nature of P atoms. Note that the presence of Ni0 species in NiP(2), NiP(1), NiP(0.5) and NiP(0.33) catalysts cannot be excluded, since the P/Ni ratios of these samples (Table S2) are all lower than the theoretical P/Ni ratios of Ni2P, Ni5P2 and Ni3P phases (0.5, 0.4, and 0.33, respectively).\nFigure S2 displays the Co 2p and P 2p core-level spectra of reduced CoP(x) catalysts. Peaks at 777.4\u2013777.9\u00a0eV are assigned to the Co0 or Co\u03b4+ (0\u00a0<\u00a0\u03b4\u00a0<\u00a01) sites in catalysts, while peaks at 129.0\u00a0eV are the fingerprint of P\u03b4\u2212 (0\u00a0<\u00a0\u03b4\u00a0<\u00a01) in cobalt phosphides. Similar to NiP(x) catalysts, the presence of Co\u03b4+ (0\u00a0<\u00a0\u03b4\u00a0<\u00a01) and P\u03b4\u2212 (0\u00a0<\u00a0\u03b4\u00a0<\u00a01) in Co2P and CoP phases points to an interphase Co\u00a0\u2192\u00a0P transfer of electron density as a consequence of the electron withdrawing nature of P [76]. The binding energy of Co\u03b4+ is slightly higher than that of Co0 due to the slight positive charge it bears. Different from NiP(x) catalysts, the P/Co molar ratios of the reduced phases are higher than the elemental analysis in the respective phases (i.e. 0.5 in Co2P and 1 in CoP), indicating a P rich surface of Co2P and CoP phases.NH3-TPD is used to characterize acid properties of the catalysts. Fig. 4\n shows the NH3-TPD profiles of reduced Ni/SiO2 and NiP(x)/ SiO2 (x\u00a0=\u00a00.33, 0.5, 1, 2). The TPD profiles consist of two peaks: one at low temperature in the range between 100 \u00b0C and 300 \u00b0C and another at high temperature between 500 \u00b0C and 650 \u00b0C. The low temperature peak in the profile of H3PO3/SiO2 corresponds to the POH groups with weak Br\u00f8nsted acidity, while the high temperature peak in the profile of Ni/SiO2 might be attributed to Lewis acid sites stemming from Ni\u03b4+ species bearing a small positive charge or unreduced Ni2+ species [2]. Both Br\u00f8nsted acid and Lewis acid sites are present on NiP(x)/SiO2 samples due to the POH groups and Ni\u03b4+/Ni2+ species, respectively. The amount of Br\u00f8nsted acid sites evidently increases as the P content increases.To get insight into how P influences the nature of active sites on the catalyst surface, further characterization of the reduced NiP(x) catalysts by CO-IR (Fig. 5\n) and furfural-IR (Fig. 6\n) was performed. The strength of the CO adsorption is directly related to the transfer of electrons from the metal to the \u03c0* orbital of the CO molecule. More back-donation of electrons to the \u03c0* orbital results in a stronger metal-CO bond and a weaker CO bond. The latter is reflected in a lower frequency of the CO stretching vibration. CO-IR can thus be used to probe the CO bond strength after adsorption as well as the electron density of the adsorption site.\nFig. 5 shows the effect of varying P content on the chemical nature of catalysts by CO-IR. The bands at 3745 and 3670\u00a0cm\u22121 are due to the \u03bd(OH) stretching vibrations of SiOH and POH surface groups [58]. The intensity of the \u03bd(POH) vibration (3670\u00a0cm\u22121) increases with the P content, indicating a higher amount of Br\u00f8nsted acid sites on catalyst surface. This is in alignment with the NH3-TPD results shown above.For the Ni/SiO2 catalyst, the band at 2045\u00a0cm\u22121 is assigned to CO molecules bonded to a single Ni atom [77,78]; the band at 1962\u00a0cm\u22121 is attributed to CO molecules bonded to two Ni atoms (bridge-CO species) [77,78]. The band at 2086\u00a0cm\u22121 is assigned to coordination of multiple CO molecules to a single coordinatively unsaturated Ni0 site.The infrared spectra of NiP(x)/SiO2 catalysts show the strong bands between 2096 and 2086\u00a0cm\u22121 (linear-CO bonded on Ni sites [57,78\u201380]), NiP(0.33) and NiP(0.5) exhibit a weak band at 1905\u00a0cm\u22121 (bridge-CO species [77,81]), and NiP(1) and NiP(2) a very weak band at 2202\u00a0cm\u22121 (PCO surface species) [77,80]. The peak broadening at 2096\u00a0cm\u22121 is observed on NiP(0.33) and NiP(0.5) samples likely due to the dipole-dipole interaction of absorbed CO since the XPS results (Table S2) reveal that more neighboring Ni sites (i.e. lower P/Ni ratios) are available on NiP(0.33) and NiP(0.5) surfaces than on NiP(1) and NiP(2) surfaces [82,83].The absorption band at 2086\u00a0cm\u22121 can tentatively be assigned to CO adsorbed on a single Ni in a phosphide matrix. As nickel in NiP(x) is positively charged according to XPS results (see Fig. 3), less back donation of electrons will occur creating a blue shift for the vibrational frequency of linearly adsorbed CO from 2045\u00a0cm\u22121 to 2086\u00a0cm\u22121. If the P content is reduced, more adjacent Ni will be available, increasing the dipole interaction and thus shifting and broadening the peak towards 2096\u00a0cm\u22121. Since both adsorption on positively charged nickel as adsorption of multiple CO on unsaturated nickel atoms lead to less back donation, it is not unexpected that the adsorption band of CO adsorbed on Ni in NiP(x) has a frequency (Fig. 5b colored lines) similar to that of multiple CO adsorbed on unsaturated sites of pure nickel (Fig. 5b, black line). The frequency of bridging-CO bands in NiP(x)/SiO2 (1905\u00a0cm\u22121) is lower than those in the Ni catalyst (1962\u00a0cm\u22121), indicating the presence of stronger bonded bridge-CO species in NiP(x)/SiO2. This allows us to conclude that the addition of P alters the properties of the adsorption site, apparently leading to a stronger bridge-CO adsorption. The intensity of the bridge-CO bands (1962 and 1905\u00a0cm\u22121) decreases significantly with increasing P content, indicating that P lowers the amount of bridge-CO species (Fig. 5b). This is expected since P could block the availability of adjacent Ni atoms, which would act as adsorption sites for bridge-CO [77,78]. Since adjacent Ni sites are required for a planar adsorption of furan, the planar adsorption of the furan ring is also expected to decrease with increasing P content and so do the furan-ring hydrogenation or the ring-opening reaction in the furfural HDO.The strength of CO adsorption on the catalyst surface is also explored by IR spectra obtained during temperature programmed desorption (Figure S3). As temperature increases, the intensity of the linear-CO band declines due to CO desorption. Increasing P content facilitates the desorption of linear-CO (Figure S3f), indicating a weaker CO adsorption on Ni sites at higher P content. This confirms the conclusions drawn from Fig. 5b that the adsorption band of linearly absorbed CO shifts to higher frequencies at higher P content as the adsorption is weaker due to the lower electron density on Ni\u03b4+. Based on this understanding, the increasing P content is also expected to weaken the furan adsorption on the catalyst surface from an electronic point of view, because it contributes to a lower electron density on Ni\u03b4+ sites providing a lower back donation of electrons from to the d-band of Ni to the \u03c0* bond of the furan ring.The adsorption bands (Fig. 5b) as well as the CO desorption behaviors (Figure S3f) are similar for NiP(0.33) and NiP(0.5) catalysts, as well as for NiP(1) and NiP(2), which is consistent with the XPS results (Table S2) showing that NiP(0.33) and NiP(0.5) have a similar Ni reduction degree of 0.67 and 0.68, respectively, and NiP(1) and NiP(2) of 0.74 and 0.75, respectively. This suggests that NiP(0.33) and NiP(0.5), as well as NiP(1) and NiP(2), possess active sites with similar geometrical and electronical properties.In summary, P addition decreases the amount of adjacent Ni sites and the electron density on Ni sites, which is expected to suppress the planar adsorption of furan ring from both geometrical and electronical view.IR spectra of adsorbed furfural on silica supported Ni/SiO2 and NiP(x)/SiO2 are shown in Fig. 6. The bands at 1675 and 1695\u00a0cm\u22121 are assigned to the \u03bd (CO) stretching vibration of the carbonyl group; bands at 1570, 1465, and 1475\u00a0cm\u22121 are due to \u03bd(CC) stretching vibration of the furan-ring [84\u201386]. The broad bands between 1400 and 1300\u00a0cm\u22121 are a result of the background correction.As the P/Ni molar ratio increases, the \u03bd(CO) bands gradually shift to lower frequency from 1675\u00a0cm\u22121 of Ni/SiO2 to 1655\u00a0cm\u22121 for NiP(2)/SiO2, which suggests that the interaction between the carbonyl group and catalyst surface is notably enhanced by P addition. Furthermore, new bands at 1625\u00a0cm\u22121 and 1622\u00a0cm\u22121 are observed on NiP(0.33) and NiP(0.5), indicating the formation of a second adsorption mode of the carbonyl group. In case of NiP(1) and NiP(2), the shoulder at 1635\u00a0cm\u22121 may also result from this new adsorption configuration of carbonyl groups.Based on geometric and electronic considerations, the tendency of planar furan ring adsorption is expected to decrease with increasing P content. Nevertheless, no frequency shift of the \u03bd(CC) band (at 1465 and 1475\u00a0cm\u22121) is observed (Fig. 6). Likely because the interaction between the furan-ring and the catalyst surface is strong and the weakening effect of P on the furan-catalyst interaction insufficient to cause a frequency shift of IR bands. The intensity of the \u03bd(CC) IR bands, however, clearly decrease with increasing P content. We therefore calculated the intensity ratios of \u03bd(CO)/\u03bd(CC) during a temperature programmed heating process to further investigate the adsorption situation of carbonyl groups and furan-rings on the catalyst surface. As temperature increases above 150 \u00b0C, a significant decrease is observed in the \u03bd(CO)/\u03bd(CC) ratios on Ni/SiO2 and NiP(0.5)/SiO2, while only a slight decrease is detected on NiP(1)/SiO2 and NiP(2)/SiO2. This does again confirm that NiP(1)/SiO2 and NiP(2)/SiO2 with higher P content exhibit a stronger carbonyl adsorption than Ni/SiO2 and NiP(0.5)/SiO2 do.The CO-IR investigation of CoP(x) was also conducted (Figure S4). As in the case of P in the NiP(x) samples, the P addition does suppress the presence of adjacent Co sites, i.e. bridging-CO adsorption sites. As argued above, P atoms do likely \u201cdilute\u201d Co atoms at the surface hence suppressing the bridge-CO adsorption [77]. Besides, the intensity of the 2066\u00a0cm\u22121 band assigned to multiple CO adsorbed on unsaturated sites, declines with increasing P content, suggesting saturation of coordinated unsaturated Co sites by P. Furthermore, a decrease in electron density of Co\u03b4+ is expected at higher P content due to the electron withdrawing nature of P, in line with the declining adsorption strength of linear-CO at higher P content (Figure S5f). Since adjacent Co sites are necessary for the planar adsorption of the furan ring and a lowered electron density of Co\u03b4+ contributes to weaker adsorption of furan ring, the P addition exerts a suppressing effect on the planar adsorption of furan rings on CoP(x) surface for reasons of geometric constrains and electronic structure, which is very similar to the situation in NiP(x).The activity of different metal phosphides was tested in furfural HDO at a WHSV of 3\u00a0h\u22121, a pressure of 1\u00a0bar and a temperature of 200 \u00b0C (Fig. 7\na). MoP, Ni2P, Co2P and WP show promising furfural conversions in the order Ni2P \u2248 MoP\u00a0>\u00a0Co2P \u2248 WP with values around 90%, 91%, 53%, and 50%, respectively. Cu3P and Fe2P show almost no activity and are excluded from our study. Subsequently, Ni2P, MoP, Co2P and WP were further tested at different contact times to explore the reaction mechanism of furfural HDO over these catalysts (Fig. 7b,c,d,e).According to previous research employed pure metal catalysts [5,6,21,35], mainly two reaction pathways are proposed for furfural HDO (Scheme 2\n). One is a hydrogenation pathway (pathway 1), in which the carbonyl group is first hydrogenated to an alcoholic hydroxyl group, which is then removed by hydrogenolysis, generating furfuryl alcohol (FOL) and subsequently 2-methylfuran (MF) as products. The other one is decarbonylation (pathway 2) yielding furan as a product. If applied catalysts (e.g., Ni, Co, Pd [5,33]) have a high hydrogenation capacity, furfuryl alcohol (FOL), 2-methylfuran (MF) and furan can be further hydrogenated to tetrahydrofurfuryl alcohol (THFOL), tetrahydro-2-methylfuran (THMF) and tetrahydrofuran (THF), respectively. If catalysts possess high hydrogenolysis capability, the formation of ring-opening products (e.g. butanol, butane, pentanol, pentane, etc.) would be obtained.In our cases, metal phosphides show a promising performance in the furfural HDO with activity orders of Ni2P \u2248 MoP\u00a0>\u00a0Co2P \u2248 WP and a desirable product selectivity towards 2-methylfuran. Product distributions differ across the series of metal phosphides significantly. WP and MoP show highest selectivity towards MF (>90%), whereas Co2P mainly produces the less-desired furan product (>40%). This shows that each metal phosphide sample favors the hydrogenation and decarbonylation routes in a different way.At decreasing furfural conversions (attained by increasing WHSV), all catalysts showed increasing furfural alcohol selectivity at the expense of 2-methylfuran production, revealing the hydrogenation mechanism (pathway 1) of metal phosphide catalysts. The constant furan selectivity revealed a decarbonylation pathway (pathway 2), which is independent from the hydrogenation pathway. Small amounts of THFOL and THMF, and no trace of ring-opening products were detected, indicating that furan-ring hydrogenation and ring-opening reactions are successfully suppressed over these catalysts. Accordingly, these catalysts can contribute to lower hydrogen consumption and lower light-product formation in comparison to traditional transition metal catalysts (i.e. Ni [5,6] or Pd [5]).For MF production, indirect and direct reaction pathways have been reported over metal and metal alloy surfaces (e.g., PtZn, NiFe, and Mo2C) [6,7,20,37]. The indirect reaction pathway consists of hydrogenation of \u03b72(C, O) surface adsorbed species to FOL, followed by conversion into MF via FOL hydrogenolysis [6,20]. The direct reaction pathway involves the conversion of the \u03b72(C, O) species into C4H3O-CH2 or C4H3O-CH intermediates, which are anticipated to directly produce MF in a H2-rich environment [7,20,37]. The presence of FOL at high WHSV over Ni2P, MoP, Co2P, and WP catalysts (Fig. 7) confirms the indirect MF formation mechanism. Yet, the direct MF formation pathways cannot be excluded, especially for MoP and WP, where high MF production can be still obtained at low conversion (Fig. 7).As shown above, Ni2P/SiO2 is practically the most active catalyst among our metal phosphides with a promising product distribution. We therefore decided to study this catalyst in more detail: Nickel phosphide samples were prepared by varying the P/Ni molar ratios NiP(x) (with x\u00a0=\u00a00.33, 0.5, 1, 2) and were tested in the furfural HDO (Fig. 8\n). Since the focus of our work is to optimize selectivity and to develop a mechanistic explanation of the furfural HDO reaction, instead of optimizing reaction rates, TOF investigations at low conversion are not included here. Nevertheless, upper bound estimates of TOFs based on the given data are provided in Table S3.Significant yields of FOL and THFOL are produced over metallic Ni, demonstrating the high hydrogenation ability, i.e. strong interaction between furfural and the Ni metal surface. Higher amounts of MF and THMF are produced at the expense of FOL and THFOL over NiP(0.33) and NiP(0.5), implying an enhanced C1O1 (Scheme 1) hydrogenolysis ability of these catalysts, i.e. a stronger interaction between the carbonyl group and the Ni-P catalyst surface. The decreasing production of THMF over NiP(0.33) and NiP(0.5), as well as the absence of THMF over NiP(1) and NiP(2) reveals a suppressing effect of P on the furan-ring hydrogenation, i.e. a weaker interaction between furan-ring and catalyst surface. High reaction temperatures contribute to high production of undesired decarbonylation and ring-opening reactions, consistent with the effect of reaction temperature on furfural HDO over pure metal catalysts [5,6,33]. The NiP(1) and NiP(2) catalysts show similar catalytic performance, which can be attributed to the fact that these catalysts contain the same active (Ni2P) phase (XRD). In line with this, their surface characteristics as probed for XPS, CO-IR and furfural-IR spectroscopy are also similar. The behaviors of the NiP(0.33) and NiP(0.5) catalysts are clearly different from those of NiP(1) and NiP(2), most likely due to the different active phases of Ni3P and Ni12P5. Reasons of different catalytic behavior of the Ni3P, Ni12P5 and Ni2P active phases are discussed below.The interaction between the furan-ring and the catalyst surface (labeled below as the furan-ring/Ni interaction) is essentially the interaction between the d-band of Ni and the \u03c0* bonding of the furan-ring. As the Ni/SiO2 catalyst accommodates the highest electron density on Ni sites and the largest concentration of bridge-CO sites among all of our catalysts, the furan-ring/Ni interaction is expected to be the strongest here, contributing to ring-hydrogenation and ring-opening reactions in the furfural conversion. Therefore, it is plausible that substantial amounts of THFOL are obtained on a Ni catalyst.As the P content increases, the number of adjacent Ni sites, required to form bridge-CO species, is substantially suppressed (Fig. 5) and the electron density on Ni sites is gradually reduced due to the electron withdrawing nature of phosphorus. From a geometric viewpoint, the furan-ring/Ni interaction is expected to decrease due to the decreasing amount of adjacent Ni sites. From an electronic structural point of view, the interaction between the Ni d-band and the furan \u03c0-system is expected to decline at higher P content due to the lower electron density on Ni\u03b4+ sites. Therefore, the furan-ring/Ni interaction as well as the ring-hydrogenation capacity of the catalysts are expected to decline as the P content increases, which is consistent with our activity results that the product distribution shifts from THFOL for Ni/SiO2 to THMF for NiP(0.33), then to MF for NiP(0.5), NiP(1), and NiP(2).The interaction between the carbonyl group and the catalyst surface, which can be characterized by furfural-IR, is another key factor influencing furfural HDO performance.According to literature, the carbonyl group adopts an \u03b72(C, O) configuration on the Ni surface, which is unstable and tends to rearrange into an \u03b71(C) configuration at higher temperature (Scheme 4c) [5]. As the \u03b71(C) configuration is most likely the precursor for decarbonylation, an increasing furan production is observed as reaction temperature increases [5]. Our results are consistent with those published in literature [5]: the \u03bd(CO)/\u03bd(CC) ratio of the adsorbed species decreases at higher temperature (Fig. 6f) confirming the \u03b72(C, O)\u00a0\u2192\u00a0\u03b71(C) conversion at higher temperature. In accordance with the formation of the \u03b71(C) configuration, more furan is produced at higher temperature (Fig. 8).After P addition, the \u03bd(CO) stretching vibration of adsorbed furfural shifts to lower frequency and the downward shift tends to be more prominent when the P content increases (Fig. 6), indicating a stronger carbonyl/catalyst interaction at higher P content. It is likely that electron-deficient Ni\u03b4+ binds to the lone pairs of the carbonyl O, while electron-rich P\u03b4\u2212 donates electrons to the anti-bonding orbitals of the CO moiety, contributing to a stronger carbonyl/catalyst interaction (i.e., a more stable \u03b72(C, O) configuration, Scheme\u00a04\nd). Such stronger carbonyl/catalyst interaction weakens the C1O1 bond (Scheme 1) in the carbonyl group. Consequently, the carbonyl hydrogenation and subsequent C1O1 (Scheme 1) hydrogenolysis of FOL could be enhanced, which is consistent with the product distribution shift from FOL and THFOL of Ni catalyst to MF and THMF of NiP(0.33) catalyst.Besides, the MF yields are relatively stable for NiP(1) and NiP(2) at low or high furfural conversions. Likely that the Ni2P surface exposes sites with the appropriate geometry and electronic properties that enable a direct MF formation from furfural deoxygenation via C4H3OCH\n[7] or C4H3OCH2O\n[37] intermediates (Scheme 3\n).Another explanation for the changing catalytic performance by P addition is based on the role of Br\u00f8nsted acid sites: In the hydrodeoxygenation of phenolic compounds, a Br\u00f8nsted acid site adjacent to a metal site is considered to facilitate CO hydrogenolysis by protonating the oxygen in OH or OCH3 groups [87,88]. Since the P species in NiP(x) catalysts could also act as Br\u00f8nsted acid sites adjacent to Ni sites, i.e. POH (Figs. 3 and 4), we infer that the increasing P content in the catalyst could also offer a promoting effect on C1O1 (Scheme 1) hydrogenolysis in FOL and THFOL, contributing to enhanced production of MF and THMF.\n\nScheme 5\n illustrates above findings and possible explanation: in the metal phosphides (in contrast to the pure metal) the weakened CC adsorption leads to less furan-ring hydrogenation products, and the strengthened CO adsorption contributes to easier C1O1 (Scheme 1) conversion. Besides, phosphorus could act as Br\u00f8nsted acid which would facilitate the C1O1 (Scheme 1) hydrogenolysis in FOL and thus contribute to a higher production of MF.Since Co2P is the third active catalyst in furfural HDO (Fig. 6) and the active phases of cobalt phosphide are adjustable by varying the molar P/Co ratio, we investigated the effect of the P stoichiometry in CoP(x) phases on the furfural HDO as well (Figure S6). The conversion decreases with increasing P content, indicating the decreasing activity in the order CoP\u00a0<\u00a0Co2P\u00a0<\u00a0Co.Similar to NiP(x), the ring-opening products are suppressed with increasing P content. This is understandable, since as in the case of the nickel phosphides, P suppresses the formation of adjacent Co sites required for the planar adsorption of the furan ring, and lowers the electron density of Co\u03b4+ sites leading to a weaker interaction between furan ring and catalyst surface. In addition, the MF production is enhanced at the expense of the FOL production by increasing the P content as a result of Br\u00f8nsted acid sites facilitating the C1O1 (Scheme 1) bond breaking.A higher furan production is found on CoP(x) than on NiP(x); a possible explanation for this observation is as follows. The \u03b72(C, O) configurations on the catalyst surface are stabilized by the carbonyl O donating electrons to electron-deficient Co\u03b4+ or Ni\u03b4+ sites and the antibonding orbitals of CO accepting electron density from electron-rich P\u03b4\u2212. Since the density of Co\u03b4+ sites on CoP(x) surfaces is lower than that of Ni\u03b4+ on NiP(x) (P/Co\u00a0>\u00a0P/Ni, Table S2), the binding of carbonyl O to the catalyst surface will be lower on CoP(x) surfaces, which contributes to an easier \u03b72(C, O)\u00a0\u2192\u00a0\u03b71(C) transformation and, consequently to a higher furan formation rate on CoP(x) catalysts.Overall, we summarize that increasing the P/Co stoichiometry in CoP(x) samples suppresses the ring-opening reactions but does not suppress the undesired decarbonylation. In comparison with NiP(x), CoP(x) catalysts are not desirable in furfural conversion due to the higher furan production and the lower MF yield (furan has a lower C yield than MF in furfural conversion). Our characterization and activity results of CoP(x) and NiP(x) demonstrate that the type of metal phosphide and the phosphorus/metal ratio plays a crucial role in determining the furfural adsorption on the catalyst surface, as well as the catalytic performance in furfural HDO.A series of transition metal phosphides was evaluated in the furfural HDO reaction, showing the following activity trend: Ni2P \u2248 MoP\u00a0>\u00a0Co2P \u2248 WP\u00a0\u226b\u00a0Cu3P\u00a0>\u00a0Fe2P. In comparison to traditional transition metal catalysts (i.e. Ni or Pd), metal phosphides like Ni2P, MoP, Co2P and WP are promising catalysts for furfural HDO as they produce 2-methylfuran and furan as major products and contribute to lower hydrogen consumption and lower light-product formation.By varying the P/Ni molar ratios in NiP(x) precursors, the effect of P stoichiometry on catalyst properties and performance are investigated in depth. A higher P content weakens the furan-ring/catalyst interaction contributing to lower furan ring-hydrogenation and ring-opening reaction. On the other hand, it provides a stronger carbonyl/catalyst interaction by providing a more stable \u03b72(C, O) configuration with the electron-deficient Ni\u03b4+ binding to the lone pairs of carbonyl O and the electron-rich P\u03b4\u2212 donating electrons to the antibonding orbitals of CO. Such enhanced carbonyl/catalyst interaction weakens the CO bond in the carbonyl group promoting its hydrogenation and further conversion. Moreover, P species could also act as Br\u00f8nsted acid sites that facilitate hydrogenolysis of FOL and THFOL, contributing to a higher production of MF and THMF. A comparable role of P was observed in CoP(x) samples, except that the undesired decarbonylation is suppressed to a lesser extent. This is likely because the density of available Co\u03b4+ sites on CoP(x) surfaces is lower than that of Ni\u03b4+ on NiP(x). This reduces the binding of the carbonyl O to the catalyst surface, thereby contributing to an easier \u03b72(C, O)\u00a0\u2192\u00a0\u03b71(C) transformation and consequently a higher furan formation rate on CoP catalysts.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. We consulted the author's guide when submitting the manuscript. Manuscripts are prepared in accordance with publishing ethics policies described in the Author's Guide.The China Scholarship Council (CSC) is acknowledged for financial support. We thank Tiny Verhoeven for performing XPS measurements and Adelheid Elemans for the ICP measurements.Supplementary data to this article can be found online at https://doi.org/10.1016/j.jcat.2021.01.031.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  The gas-phase hydrodeoxygenation (HDO) of furfural, a model compound for bio-based conversion, was investigated over transition metal phosphide catalysts. The HDO activity decreases in the order Ni2P \u2248 MoP\u00a0>\u00a0Co2P \u2248 WP\u00a0\u226b\u00a0Cu3P\u00a0>\u00a0Fe2P. Nickel phosphide phases (e.g., Ni2P, Ni12P5, Ni3P) are the most promising catalysts in the furfural HDO. Their selectivity to the gasoline additives 2-methylfuran and tetrahydro-2-methylfuran can be adjusted by varying the P/Ni ratio. The effect of P on catalyst properties as well as on the reaction mechanism of furfural HDO were investigated in depth for the first time. An increase of the P stoichiometry weakens the furan-ring/catalyst interaction, which contributes to a lower ring-opening and ring-hydrogenation activity. On the other hand, an increasing P content does lead to a stronger carbonyl/catalyst interaction, i.e., to a stronger \u03b72(C, O) adsorption configuration, which weakens the C1O1 bond (Scheme 1) in the carbonyl group and enhances the carbonyl conversion. Phosphorus species can also act as Br\u00f8nsted acid sites promoting C1O1 (Scheme 1) hydrogenolysis of furfuryl alcohol, hence contributing to higher production of 2-methylfuran.\n               "}
{"full_text": "Data will be made available on request.Supercritical water gasification (SCWG) has drawn considerable attention in recent years as a clean and renewable synthetic natural gas (bio-SNG) production technology. As water acts as the solvent and reactant in these conditions, no drying step is required to convert biomass feeds of high moisture content, leading to higher gas production efficiencies (\u00a0\u2248 70%) compared to conventional conversion technologies [1]. Combustible gases (CH4 and H2) can easily be produced from the catalytic conversion of wet biomass at moderate SCWG temperatures (375\u2013450 \u2218C) at which CH4 formation is favoured. For this however, an active gasification and methanation catalyst is required [2\u20137]. A lot of work has been performed on SCWG catalysts in order to guarantee high activity and long lifetimes, [6,8,9] as well as to understand the main deactivation mechanisms, namely leaching, [10\u201313] sintering, [14,15] poisoning [16\u201318,15] and coking [15,19,20].Many studies already showed the superiority of ruthenium-based catalysts for SCWG, be it in terms of gasification/methanation activity or in terms of stability towards leaching and sintering [8,10,16,21,22]. However, Ru is known to exhibit structure sensitivity in several catalytic reactions such as Fischer-Tropsch synthesis, [23,24] ammonia synthesis [25,26] and ammonia decomposition [27,28]. This is also the case for the methanation reaction, where size sensitivity was observed for Ru/TiO2, [29,30] Ru/C [31] and single crystals [32]. The reason for this size sensitivity is thought to arise from B\n5 sites, introduced by van Hardeveld and van Montfoort, [33] which are a combination of five under-coordinated \u201cstep-edge\u201d atoms creating a three-dimensional adsorption site for reactive species (i.e. CO or N2). For the methanation reaction, CO dissociation seems to be the rate-limiting step and occurs preferentially at under-coordinated sites, as shown over a Ni surface [34]. On Ru, ab initio studies at 400 \u2218C reported lower free energy barriers during CO methanation at stepped Ru surfaces through multiple hydrogen transfer steps leading to an easier C-O bond cleavage and the subsequent formation of H2C. and water [35]. The computational study of Shetty et al.\n[36] showed a lower CO dissociation energy barrier at \u201chollow\u201d Ru sites (allowing high coordination of CO) than on stepped surfaces, indicating that specific sites (such as B\n5 sites) are highly active for CO activation. These very active sites have high probabilities of being found at defined Ru nanoparticle (NP) diameters, as they are purely geometrical features. Based on the work of van Hardeveld and van Montfoort, [33] Jacobsen et al. [25] showed that there was a high concentration of B\n5 sites on Ru NPs of 1\u20133\u00a0nm in diameter. They suggested that the increase in ammonia synthesis activity was due to the disappearance of the smallest NPs (< 1\u00a0nm) due to sintering, which led to larger crystals containing more B\n5 sites. Indeed, Ru NPs smaller than 0.8\u00a0nm exhibit very few B\n5 sites, as there are not enough atoms available to form these special ensembles. Czekaj et al.\u00a0[37] performed DFT calculations of Ru clusters of different sizes supported on graphitic carbon layers. They showed that Ru clusters only stabilised in given geometries on graphite, and that 1.5\u00a0nm Ru NPs contained more B\n5 sites (i.e. 12) than 1.0 nm Ru NPs (i.e. 6), which is in line with the high activity observed with Ru/C catalysts composed of small Ru NPs (1.2\u20131.4 nm) [17].Most fundamental studies on Ru-based SCWG catalysts were performed with activated carbon (AC) as support, as it possesses a high specific surface area and exhibits good mechanical stability in supercritical water (SCW). Unfortunately, using AC as catalyst support is far from optimal for several reasons. On the one hand, the high surface area of AC mainly arises from micropores, which are often too small (\u00a0<\u00a01 nm) to welcome Ru NPs, and are prone to rapid surface area losses and pore blockage by coke deposits, eventually leading to mass transfer limitation and catalyst deactivation [15,38,39]. On the other hand, the heterogeneity, density and intrinsic activity of AC make it difficult to precisely evaluate the different deactivation mechanisms, e.g. coking or sintering.De Vlieger et al. showed that carbon nanofibers (CNF) could be used in continuous SCWG, highlighting the good stability of unsupported and supported CNF [19]. They followed up with another study using Ru/CNF for aqueous-phase reforming of acetic acid, proving the good gasification activity and deactivation resistance of this material in SCW [40]. CNF were also shown to be an ideal support for particle size effect studies due to their high pore volume, mostly open porosity (micropore free), high specific surface area, as well as their purity and inertness [41]. Furthermore, the CNF structure makes it an ideal support for the analysis of supported metal NPs, particularly in transmission electron microscopy (TEM) [42].The geometry of a particle is known to greatly depend on the atmosphere it is exposed to, even at low partial pressures [43]. Hence, the effect of particle size on Ru activity in SCWG may therefore greatly differ from the theory. Despite a few preliminary studies, [44,45] to the best of our knowledge, the effect of the Ru NP size on the gasification activity has never been thoroughly studied in SCWG conditions. In this paper, we elucidated the particle size effect in a model SCWG system, using Ru/CNF catalysts of different Ru NP sizes to gasify aqueous glycerol solutions to CH4.Commercial carbon nanofibers (CNF, NC7000, Nanocyl) were used as catalyst support. They were first sieved to 0.50\u20130.80\u00a0mm, then purified in 1\u00a0M KOH (2\u00a0h, reflux), washed in deionised water (DI H2O) until the filtrate was neutral, dried overnight (110\u00a0\u2218C in\u00a0air), and sieved again to the fraction of interest (0.50\u20130.80 mm). The purified support was then impregnated with RuNO(NO3)3 (31.3% Ru, Alfa Aesar) or RuCl3 \u22c5\u00a0xH2O (38% Ru, Alfa Aesar) dissolved in DI H2O with the incipient wetness method (IWI). The solution concentration was adapted to reach the desired catalyst loading (xRu) based on the support pore volume (Vp = 3.6\u00a0cm3 g\u22121), determined by addition of water to mimic the synthesis method. The impregnated CNF were dried at 110 \u2218C (air, 15\u00a0h) before being reduced in a quartz reactor (i.d. = 45\u00a0mm, L = 600\u00a0mm, with a fritted disc in the middle) for 4\u00a0h at 300 \u2218C (5 \u2218C min\u22121) in H2/N2 (5:95\u00a0vol/vol, 150\u00a0mL\u00a0min\u22121). After, the reactor was cooled down to room temperature and the catalysts were passivated by letting air diffuse through the quartz reactor. This procedure allowed the formation of a RuO2 oxide layer at the surface of Ru(0) particles in a controlled and reproducible way, which will be readily reduced back under reactive conditions [17]. The Ru/CNF catalysts were eventually sieved again to 0.50\u20130.80\u00a0mm before being loaded in the catalytic reactor. At this stage, the catalysts were referred to as \u201cfresh\u201d. High-purity glycerol (\u2264 99.7%, Carl Roth GmbH & Co. KG) was diluted in DI H2O to reach glycerol concentrations ranging from 6\u00a0wt% to 20\u00a0wt% and was used as biomass model feed for the SCWG experiments.The catalytic performance was investigated on a SCWG setup used in a previous study [10] (Konti-I, P&ID shown in Fig. S1). A high-pressure pump (Knauer 80\u00a0P) fed the aqueous glycerol (6\u201320\u00a0wt%) into the system at 28.5\u00a0MPa. A series of three heaters was used to bring the feed up to 405\u2013410 \u2218C at the beginning of the catalyst bed. A 316\u00a0L stainless steel tube (SITEC-Sieber Engineering AG) was used as fixed-bed plug-flow reactor (L = 460\u00a0mm, i.d. = 8\u00a0mm, o.d. = 14.3\u00a0mm), the flow configuration was top to bottom. The Ru/CNF catalyst bed was situated in the middle of the reactor (260\u00a0mm from reactor entry), held in place by three sizes of stainless steel wire mesh \u2013 0.08, 0.25 and 0.50\u00a0mm placed on top of a hollow stainless steel rod. The rest of the catalyst bed was filled with \u03b1-Al2O3 beads (0.8\u00a0mm diameter, 0.03\u00a0cm3 g\u22121 porosity, Alfa Aesar), which was used as inert filling material. A heat exchanger was located at the exit of the catalytic reactor to cool down the effluent. A back pressure regulator (BPR, Tescom), protected by a 15 \u03bcm frit, maintained the system at the desired pressure (28.5\u00a0MPa). After the BPR, the effluent entered a phase separator from where the liquid and gas phases exited the setup. The latter was cooled through a Peltier element (1\u20134 \u2218C) to remove the water before being analysed online with a \u03bcGC (Inficon). An automated sampler located between the BPR and the phase separator was used to collect the liquid effluent at defined times on stream (TOS). Samples (2\u00a0min sampling time, 10\u201315\u00a0mL) were taken every 30\u00a0min to monitor the carbon and ruthenium concentrations. A target WHSVgRu of 4000\u00a0gorg g\n\n\n\nRu\n\n\n\u2212\n1\n\n\n h\u22121 was selected to ensure a final carbon conversion below 50% and thus monitor the catalyst activity (i.e. turnover frequency, TOF) in the kinetic regime.The absence of internal mass transfer limitation was verified through the Weisz-Prater criterion [46], which was in an acceptable range (0.03\u2009\u226a 0.3) for all catalytic tests performed in this study.The gas produced from the catalytic experiments was analysed online with a \u03bcGC 3000 series (Inficon) having two different columns (Molsieve, 10\u2009m x 320 \u03bcm x 30 \u03bcm and PLOTQ, 8\u2009m x 320 \u03bcm x 10 \u03bcm) with TCD detectors. The former analysed H2, O2, N2, CO, and CH4 in He as carrier gas at 120 \u2218C, 25\u2009psi. The latter analysed CO2 and C2,3 in Ar as carrier gas at 70 \u2218C, 20\u2009psi. The carbon content of the unfiltered liquid effluent was analysed on a Dimatoc2000 (DIMATEC) total (TC), total inorganic (TIC) and total organic (TOC) carbon analyser. The instrument determined the TC by oxidising the carbon into CO2 at 850 \u2218C in a quartz reactor containing a Pt/SiO2 catalyst. The TIC was determined by converting the carbonates to CO2 at 160 \u2218C with addition of H3PO4 (42.5%) in a quartz reactor filled with porous silica gel beads. The TOC was eventually determined by subtraction (TOC = TC \u2013 TIC).The carbon gasification efficiency GE\n\nC\n was determined by Equation\u00a01.\n\n(1)\n\n\nG\n\n\nE\n\n\nC\n\n\n\n\n(\n\n%\n\n)\n\n=\n\n\n\n\n\n\nn\n\n\n\u02d9\n\n\n\n\nC\n,\ng\na\ns\n\n\n\n\n\n\n\n\nn\n\n\n\u02d9\n\n\n\n\nC\n,\nf\ne\ne\nd\n\n\n\n\n\u22c5\n100\n\n%\n\n\n\nby knowing the flow of carbon in the produced gas (\n\n\n\n\nn\n\n\n\u02d9\n\n\n\n\nC\n,\ng\na\ns\n\n\n) and the amount of carbon entering the system (\n\n\n\n\nn\n\n\n\u02d9\n\n\n\n\nC\n,\nf\ne\ne\nd\n\n\n) per unit of time.The carbon conversion (X\n\nC\n) was calculated with Equation\u00a02.\n\n(2)\n\n\n\n\nX\n\n\nC\n\n\n\n\n(\n\n%\n\n)\n\n=\n\n\nT\nO\n\n\nC\n\n\nf\ne\ne\nd\n\n\n\u2212\nT\nO\n\n\nC\n\n\no\nu\nt\n\n\n\n\nT\nO\n\n\nC\n\n\nf\ne\ne\nd\n\n\n\n\n\u22c5\n100\n\n%\n\n\n\nwhere TOC\n\nfeed\n and TOC\n\nout\n are the amounts of organic carbon present in the feed and the process waters, respectively.The specific surface area (SSA) and the pore volume (Vp) were measured by N2 physisorption (77\u2009K) on an Autosorb-1 (Quantachrome). The samples were outgassed in dynamic vacuum (10\u22126 bar) for a minimum of 3\u2009h at 300 \u2218C. The SSA was calculated according to the Brunauer-Emmett-Teller (BET) model, the total pore volume was determined at a relative pressure p p\n\n\n\n\n0\n\n\n\u2212\n1\n\n\n\u2265\n0.99\n\n.After synthesis, the catalyst loading was verified by calcination in static air (900 \u2218C, 10 \u2218C min\u22121, 12 h) in a muffle oven (Nabertherm). The ash content of the support was then subtracted and the residual ash content was corrected, as the ruthenium was completely oxidised (RuO2). The loading determination through calcination was in good agreement with the calculated loading from the impregnation step (error \u2264 10%), validating this characterisation method.The Ru dispersion (DTEM) was determined from TEM micrographs acquired with a JEOL JEM 2010 microscope operated at 200\u2009keV and equipped with a LaB6 cathode. Images were recorded by a slow scan CCD camera (4008\u2009\u00d7 2672 pixels, Orius Gatan Inc.). High-resolution TEM images were acquired on a probe-corrected JEOL JEM-ARM200F (NeoARM) microscope equipped with a cold-field emission gun operated at 200\u2009keV and a Gatan OneView camera. The instrument could be operated in TEM or STEM modes. Samples were prepared on lacey carbon grids (Ted Pella Inc.) using ethanol to disperse the ground catalyst. For each catalyst sample, a thorough qualitative analysis was performed and representative micrographs were carefully selected to determine a particle size distribution (PSD). The minimum sample size for each analysed catalyst was 160 particles, except for the fresh 30%Ru/CNF and both spent 1%Ru/CNF catalysts (sample size \u2248 100). The histogram bin size for the PSD was selected by following the guidelines of Alxneit [47] to ensure a statistically representative\u00a0particle size determination. The Ru NP diameters were corrected for the formation of the oxide passivation layer in contact with air, which was reported to reach 0.6\u2009nm [48]. For Ru NPs smaller than 1.2\u2009nm, the size was corrected by the ratio of the Ru(0) and RuO2 bulk densities, as performed in other studies [49,50]. The dispersion was then calculated according to Equation\u00a03,\n\n(3)\n\n\n\n\nD\n\n\nT\nE\nM\n\n\n\n\n(\n\n%\n\n)\n\n=\n\n\n\n\n\n\u2211\n\n\n\ni\n\n\nR\n\n\nu\n\n\ns\nf\nc\n,\ni\n\n\n\n\n\n\n\n\u2211\n\n\n\ni\n\n\nR\n\n\nu\n\n\nt\no\nt\n,\ni\n\n\n\n\n\u22c5\n100\n\n%\n\n\n\nwhere Rusfc,i and Rutot,i are the amount of surface and total Ru atoms in the ith\n NP, calculated from the geometrical equations published by van Hardeveld and Hartog [51], linking the particle diameter to the number of atoms for a truncated bipyramid (see Fig. S2 and Table S1). The detailed calculation steps can be found in the supporting information (SI). The reported Ru NP diameters refer to the main mode obtained from the PSDs (Figs. S3-S6).The turnover frequency (TOF) was used to compare the activity of the different catalysts and was calculated with Equation\u00a04,\n\n(4)\n\n\nT\nO\nF\n\n\n\n\n\nm\no\n\n\nl\n\n\nC\n\n\n\n\nm\no\n\n\nl\n\n\nR\nu\n,\ns\nf\nc\n\n\n\u22c5\nm\ni\nn\n\n\n\n\n=\n\n\n\n\n\n\nn\n\n\n\u02d9\n\n\n\n\nC\n,\nf\ne\nd\n\n\n\u22c5\n\n\nX\n\n\nC\n\n\n\n\n\n\nn\n\n\nR\nu\n\n\n\u22c5\n\n\nD\n\n\nT\nE\nM\n\n\n\n\n\n\n\nwhere \n\n\n\n\nn\n\n\n\u02d9\n\n\n\n\nC\n,\nf\ne\nd\n\n\n is the mole flow rate of carbon into the system, n\n\nRu\n is the moles of Ru in the catalyst bed and D\n\nTEM\n is the Ru dispersion. Thus, n\n\nRu\n\n,\n\nsfc\n =\u2009n\n\nRu\n \u22c5\u00a0D\n\nTEM\n. The initial activity (TOF\n\n\n\n30\nmin\n\n\n) was calculated with the fresh catalyst dispersion, while the steady-state activity (TOF\n\u221e\n) was calculated with the dispersion of the spent catalyst.The reaction rate (r) was used to evaluate the impact of dispersion loss on the activity of the catalyst and was calculated with Equation\u00a05.\n\n(5)\n\n\nr\n\n\n\n\n\nm\no\n\n\nl\n\n\nC\n\n\n\n\nm\no\n\n\nl\n\n\nR\nu\n,\nt\no\nt\n\n\n\u22c5\nm\ni\nn\n\n\n\n\n=\n\n\n\n\n\n\nn\n\n\n\u02d9\n\n\n\n\nC\n,\nf\ne\nd\n\n\n\u22c5\n\n\nX\n\n\nC\n\n\n\n\n\n\nn\n\n\nR\nu\n\n\n\n\n\n\n\n\nThermogravimetric analyses (TGA, Mettler Toledo TGA/SDTA 851e) were performed on the CNF support as well as on selected catalyst samples (fresh and spent). Approximately 10\u2009mg of sample were loaded in an Al2O3 crucible. The samples were first heated up to 110 \u2218C in air (5 \u2218C min\u22121, 30\u2009min hold) to get rid of the moisture. The temperature was then increased to 900 \u2218C (5 \u2218C min\u22121). Analyses were performed with air as reactive gas (10\u2009mL\u2009min\u22121) and Ar (10\u2009mL\u2009min\u22121) as protective gas.The Ru concentration in process waters was analysed using an Agilent 7700x ICP-MS with the following parameters: RF power 1350\u2009W, sampling depth 10\u2009mm, carrier gas (Ar) flow rate 0.93\u2009L\u2009min\u22121. Each sample was acidified to 1% HNO3 using TraceSELECT\u2122 nitric acid before analysis. Two isotopes (99Ru and 101Ru) were monitored at an integration time of 0.20\u2009s in transient analysis mode. External calibration with commercial standards from Inorganic Ventures (ICP Precious Metals Std) was done using five standard points with concentrations of 0, 0.01, 0.1, 1, and 10 \u03bcg L\u22121.The main support and catalyst properties are summarised in \nTable\u00a01. When comparing the as-received support (CNF_AR) and the purified one (CNF), one can see that treatment in KOH did not alter the support properties, as the SSA and pore volume remained similar. The catalyst synthesis did not affect the SSA too much either, with values ranging from 297\u2009\u00b1\u200914\u2009m2g\u22121 (1%Ru/CNF_2) to 226\u2009m2g\u22121 (30%Ru/CNF). However, the pore volume decreased linearly with increasing Ru loading. The decrease in SSA with increasing metal loading was explained by the increase in material density (Fig. S7, right). However, the decrease in pore volume was not fully explained by the density (Fig. S7, left). The volume occupied by the Ru NPs is the most likely explanation for the pore volume decrease, as the 30%Ru/CNF catalyst exhibited the most pronounced loss in Vp.By preparing the catalysts by incipient wetness impregnation, the following hypothesis can be made: 1) 100% of the support pore volume was filled by the Ru precursor solution, 2) the solution was distributed homogeneously throughout the pore volume. Hence, the presence of Ru salts inside the CNF are expected to remain there upon thermal treatment, as observed with other metallic salts in SBA-15 [52,53]. Statistically, 12\u2009\u00b1\u20097% of Ru should be located on the inside of the fibres used in this work. This is in line with results from Winter et al., who proved that Co and Pd NPs were located inside carbon nanotubes (up to 15% and 34%, respectively) after synthesis [54].The catalyst synthesis method applied in this work yielded very small Ru NPs, with DTEM reaching 67\u201371% for the 1%Ru/CNF and 5%Ru/CNF catalysts (dp = 0.9\u20131.1\u2009nm, \nFig. 1). Another 5%Ru/CNF catalyst not shown in this work, synthesised with the same technique, yielded Ru NPs of the same size (dp = 1.0\u2009nm, DTEM = 69%). These figures also highlight the high reproducibility of the synthesis method. The use of a chloride salt instead of nitrosyl nitrate led to a slightly higher dispersion.Very small Ru NPs (\u2264 1.1\u2009nm, DTEM \u2265 67%) were achieved with catalyst loadings up to 5%, while higher Ru contents led to a decrease in DTEM, from 59% for 10%Ru/CNF down to 35% for 30%Ru/CNF. However, even the high loading of the latter catalyst yielded a main Ru NP diameter mode at 2.0\u2009nm, which remains relatively small. The evolution of dp and DTEM as function of Ru loading is shown in \nFig. 2 for the fresh catalysts. These results show that small Ru NPs, i.e. high metal dispersion, can easily be achieved on CNF with a facile synthesis method, even at high loadings.In \nFig. 3, an example of a catalytic SCWG experiment is shown with the 10%Ru/CNF catalyst (see Figs. S8-S14 for all other experiments). The carbon gasification efficiency (GEC) and conversion (XC) overlapped, indicating that all the converted carbon ended up in the gas phase (the discrepancy observed for the first GEC data point is explained by the significant variation in gas flow at the start of an experiment). 10%Ru/CNF exhibited a high initial activity, before stabilising towards 30\u201335% conversion. Because steady state was not reached for all catalysts, an extrapolation was performed in order to have a more accurate estimation of the steady-state values for XC and TOF and ensure a better comparison between the different catalysts. Both parameters were hence fitted with an exponential decay function (optimised through a Levenberg-Marquardt iteration algorithm). An example is shown for the 10%Ru/CNF catalyst in Fig. S15.The carbon conversion was reported for all experiments (see Fig. S16), where two distinct trends were observed. The catalysts exhibiting a lower activity (15%Ru/CNF, 20%Ru/CNF, 30%Ru/CNF) had initial conversions in the range 50\u201365%, stabilising towards \u2248\u200920% at steady state. On the contrary, catalysts of lower loadings and smaller Ru NP diameters (1%Ru/CNF_2, 5%Ru/CNF_Cl, 5%Ru/CNF_1, 10%Ru/CNF) showed initial conversions higher than 65%, stabilising at steady-state values around 40\u201350%. Both 1%Ru/CNF catalysts exhibited high initial conversions, but suffered from drastic deactivation. At the evaluated conditions, full conversion was initially achieved for the 5%Ru/CNF_Cl and 1%Ru/CNF_1 catalysts, preventing accurate estimations of an initial TOF.To understand the implications of the different deactivation mechanisms, the effects of leaching, sintering and coking were assessed for this set of experiments.We showed in a previous work that Ru leaching was negligible from activated carbon supports, with concentrations in the range 0.01\u20130.2\u00a0\u03bcg\u00a0L\u22121, being close to thermodynamic models [10]. Ru NPs may exhibit a decreased metal-support interaction on CNF, as this support is known to be more inert than activated carbon due to its well-defined structure [55,56]. To investigate the effect of Ru loss, time-resolved ICP-MS was used to quantify the Ru loss from the Ru/CNF catalysts. The acquired data (Figs. S17 & S18) shows that the final Ru concentrations in the process waters were in the same range as for Ru/AC. Most catalysts exhibited a similar Ru loss trend i.e. higher amounts at the start before stabilising towards 0.06\u20130.12 \u03bcg L\u22121. The measured concentrations showed that the higher inertness of the support (compared to activated carbon) did not alter the metal-support interaction. Hence, leaching is thought to have a negligible effect on catalyst deactivation.In the case of the 1%Ru/CNF_1 catalyst, it is interesting to note that the Ru concentration in the process waters suddenly increased (8-fold) after 3\u2009h TOS, which coincided with the rapid decrease in XC (Fig. S19). The reason for this Ru loss behaviour remains unclear. Changes in solvent properties around the Ru NPs could be the reason for this increase, as the density and the chemical composition of the medium rapidly changed due to the rapid loss in XC. With a rapid change in XC from 100% to 10%, the gaseous products were replaced mainly by glycerol and its degradation products. However, more experimental data is required to further conclude on the effect of rapid activity loss on the Ru loss increase.As Ru leaching was shown to be negligible, the effect of sintering and coking were investigated together due to the impossibility of disentangling both effects in SCWG conditions. To do so, the SCWG activity (TOF) of the different Ru/CNF catalysts was compared in \nFig. 4, where the top graph regroups the experiments performed at WHSVgRu \u2248\u20094000\u2009gorg g\n\n\n\nRu\n\n\n\u2212\n1\n\n\n h\u22121 and the bottom graph shows the 1%Ru/CNF catalysts tested at different WHSVgRu (3000 and 9000\u2009gorg g\n\n\n\nRu\n\n\n\u2212\n1\n\n\n h\u22121). In the top graph of Fig. 4, all catalysts showed a similar trend in TOF throughout the experiments. The highest activity was recorded for 5%Ru/CNF_Cl, which also exhibited a very high dispersion (69%). Note however that this catalyst reached high initial conversions, close to the thermodynamic value, explaining the plateau observed initially. The other 5%Ru/CNF_1 catalyst was less active at the beginning of the experiment, but eventually stabilised at a steady-state TOF (TOF\n\u221e\n) in the same range as the former catalyst (TOF\n\u221e\n = 55 and 47\u2009min\u22121, respectively). The activity of 10%Ru/CNF was slightly lower than both 5%Ru/CNF catalysts, but still significantly higher than the higher-loading catalysts. For the 15%, 20% and 30%Ru/CNF catalysts, the loss in activity was more pronounced and their TOF\n\u221e\n were consequently lower (26\u201328\u2009min\u22121). Looking at the 1%Ru/CNF catalysts (Fig. 4, bottom), the initial activity was much higher for 1%Ru/CNF_2 than for the other catalysts and reached TOF =\u2009152\u2009min\u22121 at XC =\u200964%), while 5%Ru/CNF_Cl was limited at 105\u2009min\u22121 because of the high conversion (XC \u2248 100%). Due to the relatively high initial conversion, the initial TOF of the 1%Ru/CNF_2 experiment might be underestimated. However, the TOF loss was very rapid in both 1%Ru/CNF experiments, leading to low final TOF values overlapping after TOS =\u20094\u2009h and stabilising at 11\u2009min\u22121. The complete data set discussed here can be found in Table S2.In \nFig. 5, the initial TOF taken at 30\u2009min TOS (TOF30 min), is presented as a function of catalyst dispersion. Note that the initial conversions of the 5%Ru/CNF_1 and 10%Ru/CNF catalysts being above 60% (70% and 76%, respectively), the TOF\n\n\n\n30\nmin\n\n\n may be slightly underestimated (marked with an asterisk in Fig. 5). For catalyst dispersions between 35% and 60%, the TOF\n\n\n\n30\nmin\n\n\n remained relatively stable in the range 80\u2013100\u2009min\u22121. However, very small Ru NPs (DTEM > 65%) presented a higher initial TOF than larger ones. This clearly evidences a particle size effect in SCWG over Ru-based catalysts and highlights the benefit of working with small Ru NPs around 1.0\u2009nm. After the experiments, the spent catalysts were analysed by TEM and N2 physisorption and the results are reported in \nTable\u00a02. All catalysts suffered from dispersion loss, but to different extents as it can be observed in \nFig. 6. Globally, the catalysts exhibiting the most significant dispersion losses were those with the highest initial values. The Ru NP size of all spent catalysts stabilised towards 2\u20133\u2009nm, independent of the initial dispersion and loading. These results could suggest a thermodynamically-stable Ru NP size, as reported by Parker and Campbell [57] for gold NPs supported on TiO2. The pore volume generally increased by 5\u201310%, whereas the surface area decreased by 10\u201320%. Overall however, all catalysts exhibited excellent stability, which demonstrates the compatibility of the Ru/CNF catalytic system with supercritical water conditions.The TOF\n\u221e\n of all catalysts are presented in Fig. 5. As described previously, the dispersion of all tested catalyst decreased to values ranging from 30% to 50% after the gasification experiments. In contrast with the initial TOF values, TOF\n\u221e\n were in disarray with values varying by a 2-to-3-fold factor for catalysts with final dispersions around 40\u201345%. Interestingly, the three high-loading catalysts (15%, 20%, 30%) showed similar TOF\n\u221e\n around 32\u2009min\u22121, both 5% and the 10%Ru/CNF catalysts around 60\u201380\u2009min\u22121, while both 1%Ru/CNF catalysts exhibited very low TOF\n\u221e\n values (\u00a0\u2248 15\u2009min\u22121). These results suggest that the catalyst dispersion alone is not the only parameter influencing the steady-state catalyst activity.In Fig. 6, the rate loss was reported as a function of the loss in DTEM. The dashed line representing \"rate loss =\u2009DTEM loss\" is an indication of the level at which the rate loss would be entirely caused by the loss of Ru dispersion. Since all catalysts lie above that line, another phenomenon than Ru NP sintering contributed to the observed deactivation. The impact of Ru loss by leaching or loss of catalyst debris being negligible, the other cause of catalyst deactivation in this study can only be coking. However, the impact of coking on mass transfer limitation to the inside of the CNF cannot explain the observed difference between the catalysts, assuming a constant fraction of Ru NPs anchored on the inside of the fibers (see Section 3.1). The share of deactivation caused by coking does not appear to be the same for all catalysts. Indeed, the two catalysts with the highest TOF\n\u221e\n in Fig. 4 exhibited the smallest difference between the rate loss and the reference line, indicating that coke had a lower impact on deactivation.One could have expected the catalysts of higher loadings to sinter more significantly, because of the lower inter-particle distance. Indeed, Yin et al.\n[58] showed that a critical particle distance existed, up to which significant NP sintering occurred. In their study, Pt sintering could be avoided up to 900 \u2218C through higher spacing between the NPs, either by using lower metal loadings or high\u00a0surface area supports. However, the lowest DTEM losses occurred for the higher-loading catalysts (15%, 20% and 30%Ru/CNF). The limited sintering data presented in this work suggests that, for all catalysts, the inter-particle distance must have been below this critical particle distance due to the high surface area of CNF.Peng et al.\n[59] reported that a 5%Ru/AC catalysts exhibited an increased stability compared to its 2% analogue during isopropanol conversion (450 \u2218C, 30\u2009MPa), even though the dispersion of the former was lower. He showed that the coke formation rate was lower on the catalyst containing a higher fraction of Ru, although the Ru NPs were larger (5\u2009nm vs. 3\u2009nm). This is in line with the results presented here, showing that the dispersion was not the only factor affecting the stability of the catalyst. One reason for the 3-times higher activity and good stability of both 5%Ru/CNF catalysts with regard to other catalysts of same dispersion could be due to the density of active sites. Indeed, the metal loading can significantly vary between catalysts of similar dispersions.To verify this hypothesis, TOF\n\n\n\n30\nmin\n\n\n and TOF\n\u221e\n were plotted as function of the surface density of active Ru (Rusfc) in the fresh and spent materials (\nFig. 7). As described previously, the TOF\n\n\n\n30\nmin\n\n\n of 1%Ru/CNF_2 clearly stood out due to its very high initial activity at high space velocity. For the other catalysts, TOF\n\n\n\n30\nmin\n\n\n decreased before stabilising at higher Rusfc surface density (\u00a0\u2248 1.5 atomsRu,sfc nm\u22122). The smallest Ru NPs (0.9\u20131.2\u2009nm) were clearly responsible for the high initial activity with Ru surface densities lower than 0.8 atomsRu,sfc nm\u22122. However, the trend changed when looking at the steady-state data (TOF\n\u221e\n), with an optimum Ru surface density appearing in the range 0.4\u20130.7 atomsRu,sfc nm\u22122, corresponding to a catalyst of Ru NP diameter around 2.1\u2009nm with a 4\u20138\u2009wt%Ru loading and 260\u2009m2g\u22121 surface area. At higher Ru surface density, the steady-state trend matched the initial activity one, remaining constant with increasing Rusfc surface density. This is a strong indication that the steady-state catalyst activity can be increased by having the optimal amount of active sites at the catalyst surface. It is important to keep in mind that the mentioned Rusfc surface densities are average densities, and do not necessarily represent the reality when looking at the local surface density, especially for the low-loading catalysts (1%Ru/CNF). A catalyst synthesis method yielding high catalyst homogeneity and narrow particle size distribution as used here (Fig. 1) is likely of great importance.As mentioned previously, coking may be one of the deactivation mechanisms leading to the rapid loss in activity for the 1%Ru/CNF catalysts. The extent of coking investigated by HR-TEM revealed that very little coke deposited on all spent catalysts. Except for the slight increase in Ru NP size, hardly any difference could be observed between the fresh and spent catalysts. However, in the case of the 1%Ru/CNF_2 catalyst treated at high space velocity, very thin carbon deposits could be observed at the surface of Ru NPs, as it can be seen from representative images in \nFig. 8. Note that these images were acquired at very low beam intensity and exposure to limit carbon deposition from the environment. The fact that this observation could only be made on the catalyst that suffered from the most severe deactivation supports coking as the source of deactivation together with sintering. This also suggests that deactivation by coking only occurred by the deposition of a thin layer of C at the surface of the metal.To further investigate coke deposition, TGA was performed on the neat support, the 1%Ru/CNF and 5%Ru/CNF_Cl catalysts (\nFigs. 9 and \n10). For the neat CNF support, a sharp weight loss was observed with a maximum at 590 \u2218C. When Ru was loaded onto the CNF (1 and 5\u2009wt%), the weight loss occurred at lower temperatures and over a longer range, with maximum rates at 570 and 550\u2218C, respectively. The weight loss patterns were different for the fresh and both spent (1%Ru/CNF_1 and 1%Ru/CNF_2) catalysts. The former underwent constant weight loss, with one clear contribution in the DTG curve close to 600 \u2218C. The spent catalysts exhibited broader differential profiles, with the weight-loss offset higher in temperature for 1%Ru/CNF_2 (\u00a0\u2248 510 \u2218C vs. \u2248 460 \u2218C). This indicated an altered/decreased activity of ruthenium for oxidising the support (especially for 1%Ru/CNF_2), which could be ascribed to Ru NP size increase and/or coke deposits. The catalyst treated at lower WHSVgRu exhibited a similar initial oxidation activity to the fresh 1%Ru/CNF catalyst, although it required higher temperatures to fully oxidise the support and the carbon deposits. For 1%Ru/CNF_2 treated at higher WHSVgRu, the initial weight-loss phase was similar to the neat CNF, showing that the Ru NPs had a low activity towards support oxidation. Only at temperatures above 600 \u2218C could the support and coke be completely oxidised. This is in phase with HR-TEM results and supports that Ru was largely covered and blocked by coke.TGA performed on the best-performing catalyst (5%Ru/CNF_Cl, Fig. 10) showed a completely different weight-loss trend. Indeed, the offset was shifted to a lower temperature for the spent catalyst (\u00a0\u2248 300 \u2218C) compared to the fresh one (\u00a0\u2248 380 \u2218C), while the differential profile looked similar. The maximum weight-loss rate of the spent catalyst was shifted to a slightly lower temperature (520 vs. 550 \u2218C). The significant shift in temperature observed for the 5%Ru/CNF_Cl catalyst cannot be explained by deposited coke precursors of lower thermal stability than CNF because of the very small difference in the differential weight-loss curve. The reason for this shift in temperature remains unclear. For both loadings (1% and 5%), the spent catalysts lost slightly more weight than the fresh ones, indicating either a slight decrease in ash content due to the SCW conditions and/or an increase in the carbon fraction i.e. by coke deposition. The observed difference being small, it did not allow us to exclude one of the possibilities and we were not able to conclude on the impact of metal loading on coke deposition.As discussed previously, a higher Ru NP density at the CNF surface seemed beneficial to maintain a high gasification activity (5%Ru/CNF vs. 1%Ru/CNF) and avoid catalyst deactivation by carbon deposits rapidly obstructing the access to the active sites, most probably by covering Ru NPs with nanometric layers of carbon.We showed that Ru/CNF catalysts can successfully be used in continuous SCWG systems and exhibit high gasification activity and stability. Structure sensitivity of Ru was shown for the first time in SCWG conditions, with Ru NP diameters smaller than 1.2\u2009nm leading to the highest initial activity. However, a combination of high initial dispersion and optimal surface ruthenium density was shown to be crucial to maintain a high steady-state activity. The optimal surface Ru density was found to be around 0.4\u20130.7 atomRu,sfc nm\u22122, which is thought to help in delaying/suppressing coke formation by limiting carbon deposition on large Ru-free carbon surfaces. The loss of active phase from Ru/CNF catalysts being negligible, the observed catalyst deactivation was linked to a combination of coking and Ru NP sintering, systematically stabilising to 2\u20133\u2009nm. Coking was found to occur in very small amounts and to be limited to nanometric layers on top of Ru NPs. With an optimal particle size and surface density identified, efforts must now go into improving the stability of Ru nanoparticles towards coking, for instance through doping of the carbon support or ruthenium to limit adsorption of unsaturated compounds.Evaluating deactivation mechanisms in SCWG conditions is not straightforward, nevertheless Ru/CNF proved to be an ideal system to perform in-depth catalytic studies in SCWG conditions. This is mainly due to the well-defined CNF structure and high contrasts generated in TEM, as well as the large CNF pore volume (micropore-free) and surface area allowing high Ru loadings. The good stability of Ru/CNF catalysts gives new opportunities in the field of catalytic SCWG as it allows the preparation of highly dispersed and homogeneously distributed Ru nanoparticles. Its large open (non-microporous) surface allows a much higher loading than activated carbon at optimal Ru surface density, with optimised catalyst stability towards coking. Consequently, the volume of the catalyst fixed bed can be reduced, which should in turn reduce investment costs for SCWG processes, the latter being closely linked with the volume of a vessel at high pressures. The feasibility of using this catalyst on larger scales should however be assessed.\nChristopher Hunston: Investigation, Methodology, Validation, Formal analysis, Writing \u2013 original draft. David Baudouin: Conceptualization, Methodology, Validation, Writing \u2013 review & editing, Supervision. Leo Koning: Investigation, Formal analysis. Ayush Agarwal: Investigation, Validation, Formal analysis. Oliver Kr\u00f6cher: Resources, Writing \u2013 review & editing, Supervision. Fr\u00e9d\u00e9ric Vogel: Methodology, Writing \u2013 review & editing, Funding acquisition, Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This project was financially supported by the Swiss National Science Foundation grants 200021_172624/1 and 200021_184817, as well as by the Swiss Innovation Agency Innosuisse and is part of the Swiss Competence Centre for Energy Research SCCER BIOSWEET. The authors would like to thank Erich De Boni and Pascal Unverricht for designing, optimising and keeping the continuous SCWG test rig running smoothly. We also thank Andrea Testino for helping with TGA measurements and Ivo Alxneit for some HR-TEM images. The Swiss National Science Foundation (R\u2032Equip Project 206021_177020) is kindly acknowledged for the co-funding of the electron microscope.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.apcatb.2022.121956.\n\n\n\nSupplementary material\n\n\n\n.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  Ru/CNF catalysts of different Ru nanoparticle (NP) sizes (0.9\u20132.7\u00a0nm) were assessed for their performance in continuous supercritical water gasification (SCWG) of glycerol. A structure sensitivity of Ru was demonstrated, with high initial turnover frequencies (TOF) for Ru NPs smaller than 1.2\u00a0nm. Deactivation happens essentially through: 1) coking, which occurs more readily at low Ru surface densities and is limited to nanometric layers, and 2) sintering, which led to final NP sizes ranging from 2 to 3\u00a0nm independent of the catalyst loading (1\u201330\u00a0wt% Ru). A correlation between TOF and the density of surface Ru atoms was found, with an optimal surface density (0.4\u20130.7 atomRu,sfc nm\u22122) resulting in high initial and high steady-state TOFs, allowing longer lifetimes through the delay of coke formation. CNF as a support enables high metal loadings with optimal performance, which could help in decreasing the volume of high-pressure vessels and hence the cost of SCWG plants, making this technology even more attractive.\n               "}
{"full_text": "Data will be made available on request.As the only renewable carbon source, the development of biomass energy is conducive to reducing fossil-fuel energy depletion, mitigating greenhouse gas emissions, and promoting sustainable development [1,2]. Gasification is one of the most effective mechanisms for the utilization of biomass [3,4]. However, the presence of tar not only reduces the gasification efficiency but also leads to the blockage and corrosion of downstream pipelines [5\u20137]. Therefore, tar removal is crucial for improving biomass gasification.Thermal catalytic reforming is an effective method for treating biomass tar, which not only removes tar but also increases the gas yield [8\u201310]. Ni-based catalysts are widely used for tar reforming owing to their low cost, easy preparation, and high activity for tar cracking and hydrocarbon conversion [11\u201313]. However, such processes are energy-intensive given that the thermal catalytic reforming of tar is conducted at high temperatures. Moreover, Ni-based catalysts are easily deactivated by coke deposition, sintering, or oxidation [14,15].Photocatalysis can decompose organic pollutants under mild conditions by promoting relevant redox reactions. Among the various photocatalytic materials, TiO2 has been widely studied as a suitable semiconductor material owing to its high catalytic activity, high stability, non-toxicity, and low cost [16\u201318]. However, its applicability is constrained by its limited photocatalytic reaction rate [19].In contrast to the limitations of thermal catalysis and photocatalysis, photothermocatalysis can simultaneously apply light and heat to the catalyst to promote thermochemical reactions at lower temperatures with a higher efficiency [20\u201322]. For example, Li et al. [22] prepared Pt/TiO2 catalysts and conducted the catalytic oxidation of benzene under combined UV irradiation and thermal energy. Their results showed that the photothermocatalytic oxidation of benzene is more effective than photocatalytic and thermal catalytic oxidation individually. Ren et al. [23] found that the photothermocatalytic activity of the TiO2 nanosheets under mercury lamp irradiation at 240\u00a0\u00b0C or 290\u00a0\u00b0C was much higher than its cumulative photocatalytic activity under UV irradiation with the same intensity and thermal catalytic activity at the same reaction temperature. These studies demonstrate the feasibility of photothermocatalysis for the treatment of volatile organic compounds (VOCs), attributing to the synergistic photothermal effect. In addition, Mao et al. [24] prepared a Pt/CeO2 catalyst that was effective for the catalytic oxidation of benzene under UV-, visible-, and infrared-light irradiation. Li et al. [25] conducted photothermocatalytic experiments on typical VOCs and CO2 using a Ag3PO4/Ag/GdCrO3 catalyst. At 363\u00a0K and under visible-light irradiation, the catalyst achieved nearly 100% conversion of toluene. Wei et al. [26] reduced CO2 by constructing In-Em In2O3 nanoflake with metallic In embedded, which exhibited higher intrinsic activity of CO2 reduction than In2O3 under the light condition. Some studies demonstrate that the excellent performance of the synergistic photothermocatalysis results from the introduction of a plasmonic metal through the local surface plasmon resonance (LSPR) effect, which further improves the catalytic performance [24\u201327].Currently, most researches on the photothermal catalysts are focused on noble metals, including Pt [22,24,28], Ag [29,30], Au [31,32]. Experiments have shown that transition metals (Ni, Co) also exhibit a photocatalytic activity. Song et al. [33] investigated the photothermocatalytic activity of Ni2P/TiO2 nanoparticles for hydrogen generation from methanol. The photothermocatalytic activity of Ni2P/TiO2 was approximately 3.6 times the sum of activities associated with the photocatalytic and thermocatalytic reactions, exhibiting a synergistic photothermal effect. Shi et al. [34] prepared Co3O4/TiO2 nanocomposites possessing high catalytic activity and stability. A synergistic effect between the photo-assisted thermal catalysis of Co3O4 and UV photocatalysis of TiO2 could be observed, accelerating the oxidation of benzene on Co3O4. Fiorenza et al. [35] found that CoO-CuO supported on the TiO2-3% CeO2 gave the best results of CO2 conversion. The energy synergism between the thermocatalytic and photocatalytic mechanisms increased the CO2 conversion and favored efficient e-/h+ transfers.Photothermocatalysis is predominantly applied to promote the oxidation of VOCs to CO2 and H2O, or in the CO2 reduction with methane (CRM). Researchers have prepared and explored various catalysts for CRM to produce syngas and reduce greenhouse gas emissions [36,37]. Liu et al. [38] prepared Ni/Al2O3 catalysts and applied them to CO2 reduction by CH4 to produce syngas. As a result of light irradiation, the Ni/Al2O3 activity increased significantly and the syngas yield was doubled. Tan et al. [39] found that Ni/Mg-Al2O3 demonstrated high production rates of H2 and CO for photothermocatalytic CRM under light irradiation, which also showed better photothermocatalytic durability in comparison with Ni/Al2O3. Xie et al. [40] prepared Ni/TiO2 catalysts and applied them to the dry reforming of CH4 at a high temperature. With the increase in lighting intensity, the formation rates of CO/H2 were enhanced correspondingly. In addition, some new photoactive nanomaterials such as Ru/CeO2\n[41], MCM-Ni/Ni-MgO [42], NiCo/Co-Al2O3\n[43], Co/Co-Al2O3\n[44], and plasmonic photocatalyst Cu-Ru [45] were reported for enhancing the high light-to-fuel efficiencies of CRM and the excellent photothermocatalytic durability. These studies demonstrate that the photothermocatalytic effect could significantly increase the yields of H2 and CO through the improved catalytic performance of CO2 reforming. However, photothermocatalysis has rarely been utilized in tar treatment. Chen et al. [46] conducted experiments on the photothermal steam reforming of toluene using Ni/TiO2 catalyst. At 600\u00a0\u00b0C, the toluene conversion was 90.1%, which was 16.3% higher than that achieved by thermal steam reforming alone. Sun et al. [47] applied photocatalysis to the plasma-steam reforming of toluene, which showed that the conversion of toluene could be increased by 4\u20135% through the introduction of UV.This study aims to effectively convert tar (contained in biomass syngas) to syngas by CO2 reforming under mild conditions. Therefore, the photothermal catalysts were prepared using Ni as the active component, TiO2 as the photo-responsive material, and Al2O3 as the carrier. Moreover, the experimental photothermocatalytic CO2 reforming (PTCR) of tar model compounds was carried out to explore the synergistic photothermal catalytic effect.Ni/TiO2-Al2O3 was prepared in this study as a photothermal catalyst using various impregnation methods. Before catalyst preparation, the Al2O3 carrier (Nanjing Aotai Catalyst Carrier Co., Ltd., China) was calcined at 600\u00a0\u00b0C for 4\u00a0h to facilitate the formation of \n\u03b3\n-Al2O3.For the first impregnation method (M1), 10\u2009g of a mixture of TiO2 powder (10\u2009nm, Nanjing Hongde Nanomaterials Co., Ltd., China) and Al2O3 particles were placed in a beaker with a predefined mass ratio, followed by the addition of 20\u2009ml deionized water. Next, the mixture was uniformly blended using an electromagnetic stirrer at 45\u2009\u00b0C for 30\u2009min. Ni(NO3)2\u00b76\u2009H2O (analytical reagent (AR) 98.0%, Sinopharm Ltd., China) was dissolved in deionized water to obtain Ni(NO3)2 aqueous solution, in which the ratio of Ni(NO3)2\u00b76\u2009H2O to water was adjusted according to the desired NiO content in the catalyst. Then, 30\u2009ml of the Ni(NO3)2 aqueous solution was added to the mixture and stirred continuously until the water evaporated completely, yielding the precursor of the Ni/TiO2-Al2O3(M1) catalyst.In contrast to M1, the second method of catalyst preparation (M2) involved the addition of TiO2 after the Ni(NO3)2 aqueous solution was impregnated into Al2O3. The Ni/TiO2-Al2O3(M2) catalyst was prepared according to the following methodology. 10\u2009g of a mixture of TiO2 powder and Al2O3 particles with a predefined mass ratio was prepared. Al2O3 particles were placed in a beaker, following which 50\u2009ml Ni(NO3)2 aqueous solution was impregnated into the Al2O3 consistent with the desired NiO content in the catalyst. TiO2 powder was added following constant stirring of the solution at 45\u2009\u00b0C for 30\u2009min and the stirring was continued until the water evaporated completely to obtain the Ni/TiO2-Al2O3(M2) catalyst precursor.The semi-finished catalysts were dried at 115\u2009\u00b0C for 12\u2009h and calcined in a tube furnace for 6\u2009h to obtain the finished Ni/TiO2-Al2O3 catalysts, which were ground to the sizes of 0.125\u20130.25\u2009mm for the experiments. For the two preparation methods, the catalysts calcined at 450\u2009\u00b0C under a nitrogen atmosphere were denoted as mNi/xTiyAl(M1) and mNi/xTiyAl(M2), where m denotes the mass fraction of NiO in the catalysts; x and y denote the mass ratios of TiO2 and Al2O3, respectively.As shown in \nFig. 1, the experimental equipment consists of a fixed bed reactor, an electric heating furnace, a vapor generator, a gas mixer, a pipe preheater, a light source system, a cooling and collecting system, and a control system.The fixed bed reactor was made of quartz with an internal diameter of 10\u2009mm. First, the catalyst (1.2\u2009ml) was loaded into the fixed bed reactor before the experiment. The reaction zone was heated to the desired temperature (450\u2009\u00b0C) using an electric heating furnace with a light window diameter of 15\u2009mm and the whole system was purged with nitrogen during the preheating process. Following stabilization of the reaction zone temperature, CO2 (36\u2009ml/min) as a reforming agent and N2 (54\u2009ml/min) as carrier gas were injected into the pipeline and preheated to 150\u2009\u00b0C by the pipe preheater. Meanwhile, benzene (analytical reagent (AR) \u2265\u200999.5%, Shanghai Aladdin Biochemical Technology Co., Ltd., China), as the tar model compound (0.6\u2009ml/h), was fed into the system through a syringe pump and heated to 150\u2009\u00b0C by a vapor generator. Next, the gas mixture entered the fixed-bed reactor after the benzene vapor was mixed with CO2 and N2. During the experiment, the reaction zone was heated using an electric heating furnace and illuminated with a xenon lamp (50\u2009W, Beijing Zhongjiaojinyuan Technology Co., Ltd., China). The reaction bed temperature was monitored in real-time by the thermocouple and maintained at 450\u2009\u00b0C. After that, the produced gas was cooled by the condenser and then purified and dried using an absorption device equipped with adsorbent resin and silica gel. Finally, the clean non-condensable gas product was collected in bags with a collection time of 15\u2009min per bag. The reactant gas flow rate and reaction temperature were adjusted using a suitable control system. The benzene reforming reaction lasted for 2\u2009h in this study.The specific surface area and pore volume of the catalyst were measured by the physical absorption instrument (ASAP2020, MAC instrument, USA).The chemical composition of the catalyst was analyzed via an X-ray diffractometer (XRD) (Ultima IV, Rigaku Corporation, Japan). The X-ray tube was a Cu target with a tube voltage of 40\u2009kV and a tube current of 40\u2009mA, and the diffraction angle 2\u03b8 was scanned from 10\u00b0 to 80\u00b0 at a rate of 1\u00b0/min with a step of 0.02\u00b0.The functional group analysis was investigated by Fourier transform infrared (FTIR) spectrometer (Thermo Nicolet iS5). All test measurements were performed according to KBr disc technique.Temperature-programmed reduction (TPR) (AutoChem II 2920, Micromeritics, USA) was performed to investigate the influences of catalyst preparation conditions on the catalytic property. The characterization procedure was as follows. Fresh catalysts were pretreated at 200\u2009\u00b0C for 30\u2009min in 30\u2009ml/min of Ar and then cooled below 50\u2009\u00b0C and purged with Ar at 50\u2009\u00b0C for 60\u2009min. Finally, a constant flow of 10% H2/Ar was injected as the temperature ramped at 10\u2009\u00b0C/min to 800\u2009\u00b0C.The surface topography of catalysts was characterized by scanning electron microscope (SEM) (Ultra Plus, Zeiss, Germany). The accelerating voltage is 15\u2009kV. And transmission electron microscopy (TEM) (TALOS-F200X, Thermo Fisher Scientific, USA) was conducted to investigate the morphology of catalysts. The accelerating voltage was 200\u2009kV.All the gaseous products produced in the experiment were analyzed by GC9800 gas chromatograph (Shanghai Kechuang, China) to obtain the volume fractions.The components of the gas products were CO2, N2, H2, CO and CH4. Because N2 did not participate in the reaction process, the volume flow rate of N2 was constant before and after the reaction. Therefore, the volume flow rate of each component can be calculated by Eq. 1.\n\n(1)\n\n\n\n\nV\n\n\ni\n\n\n=\n\n\n\n\nC\n\n\ni\n\n\n\u00d7\n\n\nV\n\n\n\n\nN\n\n\n2\n\n\n\n\n\n\n\n\nC\n\n\n\n\nN\n\n\n2\n\n\n\n\n\n\n\n\n\nWhere Vi and Ci represent the volume flow rate (ml/min) and volume concentration (%) of gas component i (CO2, N2, H2, CO and CH4), respectively.Since all H atoms in the gas products came from benzene, the tar conversion rate (Xc, %) can be calculated as shown in Eq. 2.\n\n(2)\n\n\n\n\nX\n\n\nc\n\n\n=\n\n\n\n\n\n\n\nV\n\n\n\n\nH\n\n\n2\n\n\n\n\n\u00d7\n2\n+\n\n\nV\n\n\n\n\nCH\n\n\n4\n\n\n\n\n\u00d7\n4\n\n\n\n\u00d7\nA\n\n\n22400\n\u00d7\n\n\nV\n\n\nBenz\n\n\n\u00d7\nn\n\u00d7\n\u03c1\n\n\n\n\n\nWhere A represents the molecular weight of benzene, VBenz represents the inlet volume flow rate of benzene (ml/min), n represents the number of H atoms in benzene, \u03c1 represents the density of benzene (g/ml).The main gas products of the photo-assisted thermal catalytic reforming of tar are H2 and CO. The syngas yield (Yg, mol/kg) can be calculated by Eq. 3. The reaction rate of CO2 (rCO2, mmol min\u22121 g\u22121) and benzene (rBenz, mmol min\u22121 g\u22121) can be calculated by Eq.4 and Eq.5, respectively.\n\n(3)\n\n\n\n\nY\n\n\ng\n\n\n=\n\n\n\n\n\n\n\nV\n\n\n\n\nH\n\n\n2\n\n\n\n\n+\n\n\nV\n\n\nCO\n\n\n\n\n\n\u00d7\n1000\n\n\n22400\n\u00d7\nV\n\u00d7\n\u03c1\n\n\n\n\n\n\n\n\n(4)\n\n\n\n\nr\n\n\n\n\nCO\n\n\n2\n\n\n\n\n=\n\n\n\n\n\n\n\nV\n\n\n\n\nCO\n\n\n2\n\n\n\u2212\nin\n\n\n\u2212\n\n\nV\n\n\n\n\nCO\n\n\n2\n\n\n\u2212\nout\n\n\n\n\n\n\u00d7\n1000\n\n\n22400\n\u00d7\nm\n\n\n\n\n\n\n\n\n(5)\n\n\n\n\nr\n\n\nBenz\n\n\n=\n\n\n\n\nX\n\n\nc\n\n\n\u00d7\n\n\nV\n\n\nBenz\n\n\n\u00d7\n\u03c1\n\u00d7\n1000\n\n\nm\n\u00d7\nA\n\n\n\n\n\nWhere m represents catalyst quality, VCO2-in represents inlet CO2 volume flow rate (ml/min), VCO2-out represents outlet CO2 volume flow rate (ml/min).The Xc and Yg were calculated based on the average values during the stable reaction process.The Brunauer-Emmett-Teller (BET) surface areas, pore volumes, and pore sizes of the Ni/TiO2-Al2O3 catalysts prepared by different methods are listed in \nTable 1. It can be found that the pore sizes of the catalysts synthesized by the two impregnation methods are mainly in the mesoporous range. However, different preparation conditions influence the physical structures of the catalysts. Compared to the 0.2Ni/0.5Ti0.5Al(M1) catalyst calcined in a nitrogen atmosphere, the 0.2Ni/0.5Ti0.5Al(M1)-Air catalyst calcined in air exhibits a decrease in the BET surface area and an increase in the pore size. This change might be caused by the interaction between the catalyst components and oxygen in the air. Moreover, the 0.2Ni/0.5Ti0.5Al(M1)-T550 catalyst calcined at\u00a0550\u2009\u00b0C exhibits a significant decrease in the BET surface area, pore volume, and pore size. It implies that a high calcination temperature is apt to cause the growth and aggregation of crystal grains and the collapse of mesopores, destroying the physical structure of the catalyst.In addition, a comparison between the two different impregnation methods (with the same component content) exhibits a demonstrable difference in the pore structure. In M1, given that the TiO2 is loaded before NiO, TiO2 covers the surface of the Al2O3 carrier to form the stacking pores, hindering the impregnation of NiO into Al2O3. When the Ni/TiO2-Al2O3 catalysts are prepared by M2, as NiO is loaded onto the Al2O3 carrier before TiO2, NiO is apt to impregnate Al2O3, which results in a decrease in the surface area of the catalyst. Compared to the 0.25Ni/0.5Ti0.5Al(M2) catalyst, the 0.25Ni/0.3Ti0.7Al(M2) catalyst (with identical Ni content) possesses an increased BET surface area and pore volume, while the 0.25Ni/0.7Ti0.3Al(M2) catalyst exhibits the opposite result. Moreover, followed by an increase in the NiO content to 30%, the 0.3Ni/0.3Ti0.7Al(M2) catalyst possesses a smaller BET surface area and pore volume than the 0.25Ni/0.3Ti0.7Al(M2) catalyst. Therefore, it could be concluded that a NiO content of more than 25% and a ratio of TiO2 to Al2O3 of more than 50% are disadvantageous to the physical structure of the catalyst.The XRD patterns of the fresh catalysts are shown in \nFig. 2. TiO2 exhibits the anatase phase (PDF#71\u20131166) in all catalysts, and the diffraction peaks of NiO are found at 2\u03b8=\u200937.2\u00b0, 43.3\u00b0, 62.8\u00b0, 75.4\u00b0 and 79.4\u00b0 (PDF#71\u20131179). A diffraction peak of Al2TiO5 is detected near 2\u03b8=\u200926.8\u00b0 for both 0.2Ni/0.5Ti0.5Al(M1)-Air and 0.25Ni/0.5Ti0.5Al(M1)-T550 catalysts. It implies that a new Al2TiO5 phase can be formed by the interaction between TiO2 and Al2O3 when the catalyst is calcined in air or at a high temperature, which may negatively impact the catalytic activity. Compared to the diffraction peaks observed for the catalysts prepared using M1, those for the NiO and TiO2 catalysts prepared using M2 are observed at higher positions, indicating enhanced crystallinity.\n\nFig. 3 shows XRD patterns of the spent catalysts. TiO2 is still in the anatase phase, while the Ni diffraction peaks are observed at 2\u03b8=\u200944.6\u00b0, 51.8\u00b0 and 76.4\u00b0, indicating the transformation of NiO to Ni during the reaction. Moreover, a diffraction peak of graphitic carbon (2\u03b8=26.5\u00b0) is detected for all the spent catalysts, indicating coke deposition.\n\nFig. 4 presents the FTIR spectra of the fresh and spent catalysts with different preparation methods. It can be concluded that the preparation conditions have no effects on the functional groups on the catalyst surface. Both the fresh 0.25Ni/0.5Ti0.5Al(M1) and 0.25Ni/0.3Ti0.7Al(M2) catalysts exhibit the typical FTIR spectra of TiO2-based materials, with absorption vibrational peaks at around 1640\u2009cm\u22121 and 3400\u2009cm\u22121, which were caused by H2O deformation vibrations and -OH stretching vibrations, respectively [35]. These bands may be due to adsorbed water molecules in the air and Ti-OH bonds. The broad band at around 600\u2009cm\u22121 is caused by Ti-O-Ti vibrations [35]. Comparing the fresh and spent catalysts, it can be found that new peaks appear near 2900\u2009cm\u22121 and 1430\u2009cm\u22121, corresponding to -CH stretching vibration, -CH2 stretching vibration and -CH asymmetric vibration. The existence of the foregoing functional groups indicates that there are some hydrocarbon compounds on the catalyst surface produced after benzene cracking. Moreover, the spent catalyst shows a small signal at 1100\u2009cm\u22121, corresponding to the C-OH compound [48], which implies that the alcohol intermediates may be generated during the reaction process. Consequently, a small amount of H2O adsorbed was found on the desiccant.Ni/TiO2-Al2O3 catalysts prepared by various methods were tested by temperature-programmed reduction (TPR) (\nFig. 5). According to previous studies [49], conventional NiO/Al2O3 catalysts calcined at high temperatures (900\u2009\u00b0C) are dominated by the NiAl2O4 reduction peaks, which are observed at approximately 800\u2009\u00b0C. Compared to the NiO/Al2O3 catalysts, the Ni/TiO2-Al2O3 catalysts exhibit reduction peaks at temperatures below 500\u2009\u00b0C that are associated with NiO, having improved low-temperature reducibility. It indicates that the Ni/TiO2-Al2O3 catalysts are reduced during PTCR.Compared to the reduction temperature of the 0.2Ni/0.5Ti0.5Al(M1) catalyst, that of the 0.25Ni/0.5Ti0.5Al(M1) catalyst decreases, while the H2 consumption increases by 24.4%, indicating that an increase in the NiO content can increase the number of the active sites and enhance the catalyst performance. When the calcination temperature is 550\u2009\u00b0C, the temperature of the NiO reduction peak is shifted to approximately 500\u2009\u00b0C. Furthermore, relative to the 0.2Ni/0.5Ti0.5Al(M1) catalyst, the H2 consumption of the 0.2Ni/0.5Ti0.5Al(M1)-T550 catalyst is reduced by 12.3%. It implies that an increase in the calcination temperature weakens the properties of the active metal and reduces the number of active sites. When the catalysts are calcinated at 450\u2009\u00b0C, the NiO reduction peaks are observed at approximately 430\u2009\u00b0C for all catalysts. However, multiple NiO reduction peaks are observed at less than 340\u2009\u00b0C for the catalysts prepared using M2. These low-temperature reduction peaks may be the result of the interactions between NiO and Al2O3\n[50].The TEM images of the two catalysts prepared using the various methods are presented in \nFig. 6. From the particle-size distributions shown in Fig. 6(a) and 6(b), the mean particle sizes of NiO in the 0.25Ni/0.3Ti0.7Al(M1) and 0.25Ni/0.3Ti0.7Al(M2) catalysts are 17.48\u2009nm and 14.13\u2009nm, respectively. These values indicate that the catalyst prepared by M2 can achieve a smaller size and more uniform distribution of the crystalline particles than that prepared with M1.The high-resolution images of the selected regions in Fig. 6(a) and 6(b) are presented in Fig. 6(c) and 6(d), respectively. It can be observed from the images that TiO2 and Al2O3 in the two catalysts possess lattice spacings of 0.352\u2009nm (belonging to the {101} facets) and 0.198\u2009nm (belonging to the {400} facets), respectively. However, the NiO components in the 0.25Ni/0.3Ti0.7Al(M1) and 0.25Ni/0.3Ti0.7Al(M2) catalysts exhibit the different dominated lattice spacings of 0.209\u2009nm (belonging to the {012} facets) and 0.241\u2009nm (belonging to the {111} facets), respectively. It is suggested that the different preparation methods influence the catalytic activity of NiO.The experiments on the CO2 reforming of benzene using 0.25Ni/0.3Ti0.7Al(M2) catalyst were conducted under different reaction conditions, including the thermocatalytic CO2 reforming (TCR) at 450\u2009\u00b0C without light irradiation, the photocatalytic CO2 reforming (PCR) at room temperature, and photothermocatalytic CO2 reforming (PTCR) at 450\u2009\u00b0C under light irradiation.The different conditions have dramatic effects on the catalyst activity (\nFig. 7). The Xc is only 4.36%, while the catalyst is observed to remain nearly unchanged following the TCR experiment. By contrast, Xc is approximately zero under the PCR because H2 and CO cannot be detected in the gas product, while the CO2 concentration remains unchanged, indicating the lack of catalytic activity in the absence of thermal effects. However, under the PTCR, Xc reaches 73.45%, which is approximately 17 times higher than that under TCR. In addition, the catalyst color changes from grey to black over the course of the reaction, indicating the transformation of NiO to Ni or coke deposition, as verified by the XRD, TPR, and SEM characterizations. Therefore, it could be concluded that the 0.25Ni/0.3Ti0.7Al(M2) catalyst exhibits a strong photo-assisted thermal catalytic effect.To illustrate the function of Ni, the CO2 reforming experiments of benzene were conducted under PTCR using Ni/TiO2 and TiO2 as shown in Fig. 7(a). When TiO2 is used as the catalyst, H2 and CO are not detected in the gaseous product, indicating that the reaction fails to occur without Ni. It demonstrates that metallic Ni is the major active site for promoting CO2 reforming of benzene.As shown in Fig. 7(b), the Xc using Ni/TiO2 is about 70% at the beginning of the reaction, but decreases rapidly to about 40% after 1\u2009h. Compared to Ni/TiO2 and TiO2, the addition of Al2O3 can significantly improve the BET surface area of catalysts (based on Section 3.1 BET analysis) and the structural strength of the catalyst (based on 3.9 SEM analysis). Therefore, the performance of the Ni/TiO2-Al2O3 catalyst is enhanced due to the introduction of Al2O3.To verify the role of TiO2, the CO2 reforming experiment of benzene using conventional NiO/Al2O3 (25% NiO content and calcined at 450\u2009\u00b0C) as the catalyst was performed under a PTCR. The results indicate that H2 and CO are not detected in the gaseous product and so the reaction barely progresses. The formation of Ni-TiO2 structure can enhance the separation of e\u2212/h+ pairs under the condition of illumination. Therefore, the introduction of TiO2 plays a significant role in photocatalysis, as demonstrated by the comparative benzene conversion performance using NiO/Al2O3 and 0.25Ni/0.3Ti0.7Al(M2) catalysts.To further investigate the photothermal catalytic performance of the Ni/TiO2-Al2O3 catalysts, the CO2 reforming experiments with benzene under variable photothermal reaction conditions were conducted. Xc achieves approximately 75% at the beginning of the experiment, which is under a PTCR for 45\u2009min (Fig. 7(b)). For the experiment conducted with the PTCR, Xc is maintained at approximately 75% for more than 60\u2009min. However, when the reforming reaction is altered to a TCR (by turning off the light source), Xc decreases rapidly to approximately 25% after 30\u2009min, which is only one-third of that under PTCR. The XRD characterization of the spent catalyst after 45\u2009min of the reaction indicates that the Ni-containing phases detected on the catalyst were all metallic Ni monomers, indicating that NiO on the catalyst is reduced after 45\u2009min. Moreover, the photothermal catalytic activity of the reduced 0.25Ni/0.3Ti0.7Al(M2) catalyst for the CO2 reforming of benzene is significantly higher than only its thermal catalytic activity. However, a small amount of CH4 can be detected in the gas product, indicating that the catalytic cracking of benzene also occurs.The conversion efficiencies and syngas yield of PTCR for benzene based on the catalysts prepared under various calcination atmospheres and temperatures are shown in \nFig. 8. The Xc and Yg values obtained using the 0.2Ni/0.5Ti0.5Al(M1) catalyst are 53.2% and 55.7\u2009mol/kg, respectively, which are 12.5% and 5.5% higher than those obtained using the 0.2Ni/0.5Ti0.5Al(M1)-Air catalyst calcined in air. Moreover, the Xc obtained using the 0.25Ni/0.5Ti0.5Al(M1) catalyst is 59.2%, which is 7.2% higher than that using the 0.25Ni/0.5Ti0.5Al(M1)-T550 catalyst calcined at 550\u2009\u00b0C, while the yields of H2 and CO are 21.4 and 35.8\u2009mol/kg, respectively (increasing by 1.6 and 3.2\u2009mol/kg, respectively). The rCO2 and rBenz of these catalysts are 0.300\u20130.487 and 0.082\u20130.102\u2009mmol\u2009min\u22121 g\u22121, respectively, where 0.25Ni/0.5Ti0.5Al(M1) catalyst achieves the maximum reaction rate. It indicates that the nitrogen atmosphere and the calcination temperature of 450\u2009\u00b0C are beneficial for the catalytic activity, given the formation of Al2TiO5 due to calcination in air or at high temperatures (as verified by XRD characterisation (Fig. 3)). Moreover, the TPR analysis (Fig. 5) indicates that the reducibility of the catalyst calcined at 450\u2009\u00b0C is more favorable than that calcined at higher temperatures.The effect of the impregnation method on Xc and Yg is shown in \nFig. 9. Using 0.25Ni/0.5Ti0.5Al(M2) and 0.25Ni/0.3Ti0.7Al(M2) catalysts, Xc can reach 66.8% and 73.5%, respectively, which are 10.0% and 12.8% higher than those obtained using 0.25Ni/0.5Ti0.5 Al (M1) and 0.25Ni/0.3Ti0.7 Al (M1) catalysts, respectively. Moreover, the yields of H2 and CO are higher for the catalyst prepared by M2 than that prepared by M1. It could be observed that when the 0.25Ni/0.3Ti0.7Al(M2) catalyst is used, the yields of H2 and CO can achieve 25.6 and 50.3\u2009mol/kg, respectively (representing an increase of 4.5 and 4.7\u2009mol/kg, respectively, compared to the yields obtained using 0.25Ni/0.3Ti0.7Al(M1) catalyst). The rCO2 of these catalysts are in the range of 0.487\u20130.545\u2009mmol\u2009min\u22121 g\u22121 and rBenz are 0.102\u20130.129\u2009mmol\u2009min\u22121 g\u22121, where 0.25Ni/0.3Ti0.7Al(M2) catalyst has better performance. BET and TPR characterizations showed that the catalysts prepared using M2 possess a superior structure and more active sites than those prepared using M1.Experiments on the photothermal catalytic feasibility of Ni/TiO2-Al2O3 catalysts show that the introduction of TiO2 is necessary for the NiO/Al2O3 catalysts to possess photocatalytic properties. However, BET analysis indicates that the introduction of TiO2 may cover the Al2O3 surface, which has a negative impact on the physical structure of the catalyst.Therefore, the influence of the TiO2 to Al2O3 ratios on Xc is investigated to identify the optimal value. Xc and Yg for various TiO2/Al2O3 ratios are shown in \nFig. 10. When the TiO2 to Al2O3 ratio is 3:7, Xc attains an optimal value of 73.5%, while the yields of H2 and CO are 25.6 and 50.3\u2009mol/kg, respectively. As the ratio of TiO2 to Al2O3 is further increased, Xc and the Yg decline owing to the reduction in the specific surface area and pore blockage caused by excessive TiO2 coverage on the catalyst surface, which hinders the interaction between the active site and reactants. When the ratio decreases to less than the optimal value, Xc and Yg also decline because of the decrease in TiO2 photocatalysis. The maximum rCO2 and rBenz of this catalyst group is also the 0.25Ni/0.3Ti0.7Al (M2) catalyst (0.545 and 0.129\u2009mmol\u2009min\u22121 g\u22121, respectively).The effect of NiO content on the PTCR is shown in \nFig. 11. When the NiO content is 20%, Xc and the yields of H2 and CO are 57.7%, 23.3 and 40.1\u2009mol/kg, respectively, representing a decrease of 27.3%, 9.8%, and 25.4%, respectively, compared to those obtained by the catalyst with 25% NiO content. The increase in NiO content increases the number of active sites on the catalyst surface, thus enhancing the CO2 reforming of benzene. However, when the NiO content is increased to 30%, compared to the 0.25Ni/0.3Ti0.7Al(M2) catalyst, the Xc decreases by almost 20%, while the yields of H2 and CO are reduced by 6.1 and 16.7\u2009mol/kg, respectively. This may be explained by the fact that excessive NiO leads to pore blockage when the NiO content exceeds 25%, which weakens the physical structure of the catalyst, thereby decreasing photothermal catalytic performance. Moreover, more NiO might cover TiO2 reducing the light absorption, thus leading to the degradation of the catalyst performance. The rCO2 and rBenz of these catalysts are 0.475\u20130.545 and 0.090\u20130.129\u2009mmol\u2009min\u22121 g\u22121, respectively. Also, the 0.25Ni/0.3Ti0.7Al(M2) catalyst achieves the maximum reaction rate.From the comparison of the above catalysts, it can be found that best performance is achieved by the 0.25Ni/0.3Ti0.7Al(M2) catalyst. The space velocity (SV) directly affects the contact time between the reactants and catalysts. Experiments on the effects of SV were conducted by adjusting the total flow rates while keeping the concentrations of benzene and CO2 constant, using the 0.25Ni/0.3Ti0.7Al(M2) catalyst. As shown in \nFig. 12, when the SV is reduced from 4500 to 3000\u2009h\u22121, Xc increases significantly from 73.45% to 88.24%. When the SV is reduced from 4500 to 3750\u2009h\u22121, the H2 yield increases by 2.9\u2009mol/kg, while the CO yield remains essentially unchanged. When the SV is further reduced to 3000\u2009h\u22121, the H2 yield remains nearly unchanged while the CO yield increases by 0.8\u2009mol/kg relative to that associated with the SV of 3750\u2009h\u22121. On the one hand, the molar ratios of H2 to CO are in the range of 0.5\u20130.56, which is higher than the theoretical value for the CO2 reforming of benzene (C6H6\n\n+\n 6CO2\n\n\n\n\u2192\n\n 3\u2009H2\n\n+\n 12CO). As SV decreases, the H2 to CO ratio exhibits an increasing trend. On the other hand, Xc increases dramatically owing to the increase in CH4 yield. This indicates that the decrease in SV is more important for enhancing catalytic benzene cracking than for reforming.To investigate the causes of catalyst deactivation, SEM analysis was performed on the fresh and spent 0.25Ni/0.5Ti0.5Al(M2) and 0.25Ni/0.3Ti0.7Al(M2) catalysts.As shown in \nFig. 13(a) and (d), the morphologies of the two fresh catalysts are similar. Following the photothermal CO2 reforming reaction, the surface morphology of the 0.25Ni/0.5Ti0.5Al(M2) catalyst is severely damaged and many granular substances appear on the surface (Fig. 13(b)-(c)). By contrast, the spent 0.25Ni/0.3Ti0.7Al(M2) catalyst (Fig. 13(d)-(e)) retains its original surface morphology, while granular and flocculent substances also appear. It can be inferred that the proportion of TiO2 has a specific influence on the physical structure of the catalyst, and the catalyst stability is weakened when the proportion of TiO2 exceeds 50% because the light-induced thermal effect causes sintering.Moreover, further magnified images of the sintered and unsintered surfaces of the spent 0.25Ni/0.5Ti0.5Al(M2) catalyst (Fig. 13(g)-(h)) distinctly show clusters and granular substances. Through EDS analysis, the new element on the spent catalyst is identified as carbon, indicating coke deposition occurs. It implies that the main reason for the deactivation of the Ni/TiO2-Al2O3 catalysts is deposited carbon. Therefore, further research is needed to reduce the amount of coke deposition to improve the catalyst stability.In this study, experiments on the photothermocatalytic CO2 reforming of benzene as a tar model compound were conducted using the prepared Ni/TiO2-Al2O3 catalysts. The main conclusions drawn are as follows.\n\n(1)\nThe introduction of TiO2 can convert Ni2+ to Ni during mild photothermal reactions, which enhances the availability of holes for benzene oxidation. The Ni-TiO2 structure brings a better separation of e\u2212/h+ pairs under the light condition, significantly enhancing the photothermocatalytic activity of the Ni/Al2O3 catalyst at low temperatures.\n\n\n(2)\nThe catalyst preparation method involving the loading of NiO before TiO2 is beneficial to the catalyst performance. However, excessive TiO2 content caused a decrease in the specific surface area and pore volume of the catalysts and their structural stability, thereby decreasing the catalyst activity.\n\n\n(3)\nThe experimental results of PTCR show that the appropriate preparation parameter for the Ni/TiO2-Al2O3 catalyst is 25% NiO content with a TiO2/Al2O3 ratio of 3:7 at a 450\u2009\u00b0C calcination temperature in a nitrogen atmosphere. When the PTCR of benzene is conducted using the 0.25Ni/0.3Ti0.7Al(M2) catalyst, a maximum Xc of 88.2% was achieved here.\n\n\n(4)\nThe main cause of catalyst deactivation is coke deposition on the catalyst surface, which is mainly due to the thermal cracking of benzene.\n\n\nThe introduction of TiO2 can convert Ni2+ to Ni during mild photothermal reactions, which enhances the availability of holes for benzene oxidation. The Ni-TiO2 structure brings a better separation of e\u2212/h+ pairs under the light condition, significantly enhancing the photothermocatalytic activity of the Ni/Al2O3 catalyst at low temperatures.The catalyst preparation method involving the loading of NiO before TiO2 is beneficial to the catalyst performance. However, excessive TiO2 content caused a decrease in the specific surface area and pore volume of the catalysts and their structural stability, thereby decreasing the catalyst activity.The experimental results of PTCR show that the appropriate preparation parameter for the Ni/TiO2-Al2O3 catalyst is 25% NiO content with a TiO2/Al2O3 ratio of 3:7 at a 450\u2009\u00b0C calcination temperature in a nitrogen atmosphere. When the PTCR of benzene is conducted using the 0.25Ni/0.3Ti0.7Al(M2) catalyst, a maximum Xc of 88.2% was achieved here.The main cause of catalyst deactivation is coke deposition on the catalyst surface, which is mainly due to the thermal cracking of benzene.\nYutong Shen: Investigation, Methodology, Data analysis, Writing \u2013 original draft and Editing. Jun Xiao: Funding acquisition, Methodology, Writing- Reviewing, Supervision. Qijing Wu: Writing \u2013 review & editing. Jingting Su: Investigation. Li Zhu: Methodology.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Natural Science Foundation of China (51576047).", "descript": "\n                  To improve the conversion efficiency and reduce the CO2 emissions associated with the reforming of biomass tar to syngas, photothermal catalysts were prepared in this study through various impregnation methods. Photothermocatalytic CO2 reforming (PTCR) of benzene as a tar model compound, was conducted using the prepared Ni/TiO2-Al2O3 catalysts. The results show that the introduction of TiO2 facilitates the conversion of NiO to Ni during the photothermal reaction, while the photocatalysis of TiO2 significantly enhances the photothermocatalytic activity of the Ni/Al2O3 catalysts at low temperatures. Moreover, the methodology for catalyst preparation involving the loading of NiO before TiO2 is also beneficial to catalyst performance. However, when the mass ratio of TiO2 to Al2O3 exceeds 1, the specific surface area, pore volume, and structural stability of the catalysts are reduced. The results of PTCR experiment show the following optimal Ni/TiO2-Al2O3 catalyst preparation parameters: 25% NiO content and a TiO2/Al2O3 ratio of 3:7; 450\u00a0\u00b0C calcination temperature and a nitrogen atmosphere. The maximum benzene conversion achieved through PTCR with the 0.25Ni/0.3Ti0.7Al(M2) catalyst is 88.2%. However, a large amount of coke is deposited onto the spent catalyst surface, contributing significantly to catalyst deactivation.\n               "}
{"full_text": "Fuel cells, which directly convert the chemical energy stored in fuels to electrical energy, have been widely investigated and applied in many variations since the first hydrogen fuel cell was invented in 1838. Direct ethanol fuel cells (DEFCs) are promising candidates for portable power applications. Ethanol has an energy density of 8.03\u202fkW\u202fh\u00b7kg\u20131, which is lower than that of hydrogen (32.8\u202fkW\u202fh\u00b7kg\u20131). Regarding the density of the fuel, however, the volumetric energy density of ethanol (6.28\u202fkW\u202fh\u00b7L\u20131) is much higher than that of hydrogen gas compressed at 20\u202fMPa (0.18\u202fkW\u202fh\u00b7L\u20131). Even if we consider realistic operation conditions where a DEFC operates at 0.5\u202fV and a current density of 100\u202fmA\u00b7cm\u20132 with the complete oxidation of ethanol to CO2 via a 12-electron transfer, the thermodynamic efficiency of the DEFC will be approximately 40%, which is comparable to that of a conventional diesel engine (Lamy, Coutanceau, & Leger, 2009). In addition, ethanol has a low toxicity, is easy to store and transport, and can be derived from biomass. Therefore, ethanol is a promising feedstock for low-temperature fuel cells.In a DEFC, the anode reaction is an ethanol oxidation reaction (EOR), which involves a 12-electron transfer for the complete oxidation of ethanol to CO2. However, the kinetics are slow and incomplete oxidation is an issue, thus reducing the current density and hindering the application of DEFCs. The reaction pathways of EOR have been investigated by many techniques coupled with cyclic voltammetry and/or chronoamperometry, including in situ infrared spectroscopy (IR) (Chang, Leung, & Weaver, 1990; Vigier, Coutanceau, Hahn, Belgsir, & Lamy, 2004), in situ Raman spectroscopy (De Souza, Neto et al., 2011; Lai, Kleyn, Rosca, & Koper, 2008), in situ sum-frequency generation (SFG) spectroscopy (Kutz, Braunschweig, Mukherjee, Dlott, & Wieckowski, 2011, 2011b), differential electrochemical mass spectrometry (DEMS) (Wang, Jusys, & Behm, 2004, 2007; Willsau & Heitbaum, 1985), and in situ nuclear magnetic resonance (NMR) spectroscopy (Huang, Sorte, Sun, & Tong, 2015). Recently, Wang et al. thoroughly discussed the reaction mechanism of EOR (Wang & Cai, 2015). This is briefly introduced in this review. Generally, two reaction pathways, namely, the C1-path and C2-path, are recognized, corresponding to the complete oxidation to CO2 with 12-electron transfer and incomplete oxidation to acetaldehyde and acetic acid with 2- and 4-electron transfer, respectively. CO2 and acetic acid are the final products of the reaction, while acetaldehyde can desorb to the bulk electrolyte and can subsequently be re-adsorbed onto the catalysts, followed by being further oxidized to CO2 or acetic acid, as shown in Fig. 1\n(a). However, the acetic acid is difficult to oxidize further and therefore becomes dissolved in the electrolyte. Fig. 1(b) presents a schematic illustration of the pathway to complete oxidation. COad acting as the intermediate followed by CC bond cleavage strongly adsorbs onto the catalyst surface and cannot be removed oxidatively at low potential, thus inhibiting further adsorption of the reactants and consequently suppressing the current density. Till to the high applied potential, COad can be oxidized by the *OH group from water dissociation, producing CO2. To date, low selectivity to CO2 and low current of the EOR in the low potential range remain fundamental issues. Mechanistically, the target catalyst should be able to easily break the CC bond together while also providing sufficient *OH oxidant adjacent to the C1 species to contribute to complete oxidation. In this scenario, the active sites can be renewed and thus adsorb the reactants.Pt was the earliest-reported single-component catalyst and remains the best, making it the most intensively investigated for EOR (Marie, 1929; Marinkovic, Li, & Adzic, 2019; Rizo, P\u00e9rez-Rodr\u00edguez, & Garc\u00eda, 2019; Zhou et al., 2003). The mechanism of EOR has been well studied on a well-defined Pt surface (Colmati et al., 2009a, 2009b; Ferre-Vilaplana, Buso-Rogero, Feliu, & Herrero, 2016; Lai & Koper, 2009, 2010; Tarnowski & Korzeniewski, 1997; Xia, Liess, & Iwasita, 1997), with the finding that a surface with steps facilitates CC bond splitting. Pt nanocrystals with different shapes, such as nanocubes, nanowires, multi-pods, tetra-hexahedrons, and other polyhedrons, have been synthesized and studied for EOR (Han, Song, Lee, Kim, & Park, 2008; Huang, Zhao, Fan, Tan, & Zheng, 2011; Liu et al., 2016; Tian, Zhou, Sun, Ding, & Wang, 2007; Zhou, Huang et al., 2010, 2011). These studies have shown that Pt nanocrystals enclosed by high-index surfaces improve the current density. In addition to Pt metal, Pd (Bianchini & Shen, 2009; Cui, Song, Shen, Kowal, & Bianchini, 2009; Liang, Zhao, Xu, & Zhu, 2009; Tian, Zhou, Yu, Wang, & Sun, 2010; Yang et al., 2014; Zhang, Zhou, & Zhou, 2016; Zhou, Wang, Lin, Tian, & Sun, 2010), Ir (de Tacconi, Lezna, Beden, Hahn, & Lamy, 1994; Du, Wang, Saxner et al., 2011; Zhu et al., 2017), Au (Cao, Li et al., 2020, 2020b; Rodriguez, Kwon, & Koper, 2012; Tremiliosi-Filho et al., 1998), and Rh (Caram & Guti\u00e9rrez, 1992; Li, Fan et al., 2019, 2017; M\u00e9ndez, Rodr\u00edguez, Ar\u00e9valo, & Pastor, 2002; Suo & Hsing, 2011; Zhang et al., 2018; Zhu, Lan, Wei, Wang, & Yang, 2019) single-component catalysts have also been applied to EOR. Although each exhibits a low catalytic activity relative to Pt, the current density of the Pt catalyst can be enhanced together with a second component. Various Pt-based binary catalysts have been developed to improve the current density of EOR by either promoting water dissociation or facilitating CC bond cleavage. The enhancement of the current density in the cases of PtRh (Almeida et al., 2019; Bergamaski, Gonzalez, & Nart, 2008; Le\u00e3o, Giz, Camara, & Maia, 2011; Li, Liu et al., 2019; Rao, Jiang, Zhang, Cai, & Sun, 2014; Zhu, Bu, Shao, & Huang, 2019, 2018), PtIr (Chang et al., 2019), PtPd (Miao et al., 2020), PtAu (Li, Liu, & Adzic, 2012), PtCo (Zhang et al., 2017), PtBi (Zhang, Lai et al., 2019), PtNi (Altarawneh, Brueckner, Chen, & Pickup, 2018; Sulaiman, Zhu, Xing, Chang, & Shao, 2017) core-shell, or alloy nanoparticles has been attributed to the prompted capability of breaking the CC bond due to the electronic effect. The incorporation of metal oxides. such as TiOx (Corchado-Garc\u00eda, Morais, Alonso-Vante, & Cabrera, 2017; Song et al., 2007), CeOx (Bai et al., 2007; Corchado-Garc\u00eda et al., 2015; Men\u00e9ndez et al., 2014; Murphin Kumar et al., 2017; Xu & Shen, 2005; Xu, Zeng, Shen, & Wei, 2005), WOx (Ganesan & Lee, 2006; Wu, Liu, & Wu, 2010; Zhang, Ma et al., 2006), or NiO (Comignani, Sieben, Brigante, & Duarte, 2015), increases the current density by increasing the capability to oxidize the intermediates, owing to the synergistic effect between the oxides and Pt nanoparticles. To date, the most widely investigated binary-system catalysts have been Pt-Ru (Camara, de Lima, & Iwasita, 2004; Dong, Gari, Li, Craig, & Hou, 2010; Fujiwara, Friedrich, & Stimming, 1999; Hu, Zhu, Zhang, & Liu, 2016; Schmidt, Ianniello, Pastor, & Gonz\u00e1lez, 1996; \u015een, \u015een, & G\u00f6ka\u011fa\u00e7, 2011; Spinac\u00e9, Neto, Vasconcelos, & Linardi, 2004; Zhao et al., 2017) and Pt-Sn (Antolini, Colmati, & Gonzalez, 2009; Colmati, Antolini, & Gonzalez, 2006; Du et al., 2014; Godoi, Perez, & Villullas, 2010; Jiang et al., 2005; Lamy, Rousseau, Belgsir, Coutanceau, & L\u00e9ger, 2004; L\u00f3pez-Su\u00e1rez, Carvalho-Filho et al., 2015; Silva et al., 2010, 2011; Wang et al., 2007; Zhang, Liu et al., 2019; Zhu, Bu, Shao, & Huang, 2020), given their apparent ability to improve the current density over pure Pt metal, which can be in the form of a single phase (as a metallic/intermetallic alloy) or a mixed phase (as a core\u2013shell or segregated grains) (\u015een et al., 2011). Moreover, the chemical states of Ru and Sn vary depending on the synthesis method, making the roles of Ru and Sn species in EOR more complicated, as has been reviewed by (Antolini, 2013) and Lamy et al. (2004), respectively. Since the recognition that the current can be enhanced by the introduction of a second component in a binary catalyst, ternary catalysts have attracted the interest of many scientists, who have attempted not only to introduce a second component to supply more oxidants, but also to introduce a third component to facilitate CC bond splitting. Various structural configurations and component combinations for ternary materials have been reported for EOR, which have been partly summarized in recent reviews (Antolini (2013); Bai, Liu, Yang, & Chen, 2019; Marinkovic et al., 2019; Wang & Cai, 2015). However, no review has focused on ternary catalysts for EOR, which offer the promise of overcoming the issues mentioned for single and binary catalysts. In previous reviews, the role of the third species added to a binary system has been discussed in detail. In addition to the type of the third component, the architecture of the nanoparticles (either homogeneous or segregated phases) plays a decisive role in the EOR performance. For example, triphasic PtRhOx\u2013SnO2 catalysts with partially oxidized Pt and Rh cores and SnO2 shells showed a 2.5-fold increase in the CO2 generation rate for EOR, compared with biphasic PtRh-SnO2 catalysts with a metallic PtRh alloy core (Yang, Frenkel, Su, & Teng, 2016). Table 1\n summarizes the structure, components, and composition of existing ternary electrocatalysts. We classified the reported ternary catalysts into single-phase and segregated-phase materials according to their structures. The influences of the components and composition are discussed considering that they had the same structure.Adding a third component to the binary catalysts not only reduces the amount of platinum but also may promote CC bond cleavage and provide more oxidants to completely oxidize intermediates to CO2. To date, many elements have been used as auxiliary components in ternary catalysts to promote the current density for EOR, as shown in Fig. 2\n and summarized in Table 1. Generally, d-block metals (the d-block elements are located in the middle of the periodic table and include elements from groups 3\u201312) were the most frequently used for ternary catalysts, such as Pt-noble metal-nonnoble metal (Pt-Au-Ni (Co and Cu) (Cui et al., 2020; Dutta & Ouyang, 2015; Li, Jilani et al., 2019; Wang et al., 2015), Pt-Ir-Cu (Ni) (Ahmad et al., 2019; Wang et al., 2011), Pt-Pd-Cu(Ni) (Arroyo-G\u00f3mez et al., 2019; Hu et al., 2012; Ma, Wang, Fan, Zhang, & Li, 2015; Ren, Zhang, Liang, Wu, & Shen, 2020; Wang, Xue et al., 2018; Wang, An, & Zhang, 2020; Wang, Zhang et al., 2020), Pt-Rh-Ni(Co, Fe, Cu) (Almeida et al., 2020; Han, Liu, Chen, Jiang, & Chen, 2018; Chen et al., 2019; Erini, Rudi et al., 2015, 2017; Han et al., 2019; Liu et al., 2017; Shen, Zhang, Xiao, & Xi, 2014; Wang, Du, Sriphathoorat, & Shen, 2018; Zhang, Huang et al., 2019; Wang et al., 2017), and Pt-Ru-Co(Ni, Mo, and W) (Garc\u00eda et al., 2012; Li, Kang, & Wang, 2011; Wang, Yin et al., 2007; Oliveira Neto, Franco, Aric\u00f3, Linardi, & Gonzalez, 2003; Ribadeneira & Hoyos, 2008; Tanaka et al., 2005; Wang, Yin, Zhang, Sun, & Shi, 2016, 2019c)); Pt-noble metal-noble metal (Pt-Au-Pd (Chen et al., 2015; Dutta, Mahapatra, & Datta, 2011; Dutta, Ray, Sasmal, Negishi, & Pal, 2016; Zhang, Wu, & Xu, 2012), Pt-Au-Ir (da Silva et al., 2017; Liang et al., 2019), Pt-Ir-Rh (Liu, Chia, Cheng, & Lee, 2011), Pt-Rh-Pd (Huang et al., 2015; Zhu et al., 2015), Pt-Rh-Ru (Lima & Gonzalez, 2008; Nakagawa, Kaneda, Wagatsuma, & Tsujiguchi, 2012; Salazar-Banda, Suffredini, Calegaro, Tanimoto, & Avaca, 2006; Xiong et al., 2020) and Pt-Ru-Re (Choudhary & Pramanik, 2020)); and Pt-nonnoble metal-nonnoble metal (Pt-Ni-Mo (Mao et al., 2017) and Pt-Ni-Cu (Castagna, Sieben, Alvarez, Sanchez, & Duarte, 2020; Hong, Lee, Kim, & Choi, 2019; Huang, Liu et al., 2019; Imanzadeh & Habibi, 2020; Jilani et al., 2020)). In addition, some p-block elements (p-block elements are located on the right side of the standard periodic table and include elements from groups 13\u201318) can be added to form a unique structure to improve the electrocatalytic performance. Examples include Pt-Sn-noble metal (Pt-Sn-Rh (Bach Delpeuch, Maillard, Chatenet, Soudant, & Cremers, 2016; Colmati, Antolini, & Gonzalez, 2008; Dai, Wang et al., 2018;\nde Souza, Giz, Camara, Antolini, & Passos, 2014; Du, Wang, LaScala et al., 2011; Erini, Krause et al., 2015, 2014; Fan et al., 2019; Garc\u00eda-Rodr\u00edguez et al., 2011; Higuchi, Takase, Chiku, & Inoue, 2014; Jiang, Bu, Wang, Guo, & Huang, 2015; Kowal, Gojkovi\u0107, Lee, Olszewski, & Sung, 2009, 2009b; Li, Marinkovic, & Sasaki, 2012; Li, Zhou, Marinkovic, Sasaki, & Adzic, 2013; Li et al., 2010; L\u00f3pez-Su\u00e1rez, Perez-Cadenas et al., 2015; Mai, Chiku, Higuchi, & Inoue, 2015; Mai, Chiku, Higuchi, & Inoue, 2017; Silva, Camara, & Giz, 2019; Silva-Junior et al., 2013; Soares, Morais, Napporn, Kokoh, & Olivi, 2016; Song et al., 2012; Spinac\u00e9, Dias, Brandalise, Linardi, & Neto, 2010; Yang et al., 2016; Yang, Namin, Aaron Deskins, & Teng, 2017), Pt-Sn-Ag (Dai, Huang et al., 2018), Pt-Sn-Ir (Li, Cullen et al., 2013; Ribeiro et al., 2007; Silva et al., 2012, 2013; Tayal, Rawat, & Basu, 2011; Thilaga, Durga, Selvarani, Kiruthika, & Muthukumaran, 2018; Zhao, Mitsushima, Ishihara, Matsuzawa, & Ota, 2011), Pt-Sn-Pd (Lee, Park, & Manthiram, 2010; Wang et al., 2013, 2016), Pt-Sn-Ru (Chang, Liu, Wei, & Wang, 2009; Chu & Shul, 2010; Cunha, Ribeiro, Kokoh, & de Andrade, 2011; Hang, Altarawneh, Brueckner, & Pickup, 2020; Huang, Zheng et al., 2019; Liu, Chang, Wei, & Wang, 2011; Neto, Dias, Tusi, Linardi, & Spinac\u00e9, 2007; Thepkaew, Therdthianwong, Therdthianwong, Kucernak, & Wongyao, 2013; Wu, Swaidan, & Cui, 2007; Xia, Zhang, Zhao, & Li, 2017), and Pt-Sn-Au (Dai et al., 2020; Shakibi Nia et al., 2020)); Pt-Sn-nonnoble metal(Pt-Sn-Mo (Lee, Murthy, & Manthiram, 2011), Pt-Sn-Ni (Beyhan, L\u00e9ger, & Kad\u0131rgan, 2013; HUANG, Lian-Hua ZHAO, & Shi-Gang, 2017; Parreira et al., 2013; Ponmani, Kiruthika, & Muthukumaran, 2015; Spinac\u00e9, Linardi, & Neto, 2005), Pt-Sn-V (Jin, Sun, Huang, & Zhao, 2014; Sun, Zhao, & Yu, 2013), Pt-Sn-W (Anjos, Hahn, L\u00e9ger, Kokoh, & Tremiliosi-Filho, 2008; Ribeiro et al., 2008), Pt-Sn-Cu (Huang, Wu, Wu, & Guan, 2015), Pt-Sn-Ce (De Souza, Parreira et al., 2011; Jacob, Corradini, Antolini, Santos, & Perez, 2015), Pt-Sn-Fe (Almeida, Van Wassen, VanDover, de Andrade, & Abru\u00f1a, 2015), Pt-Sn-Pb (Santos, Almeida, Tremiliosi-Filho, Eguiluz, & Salazar-Banda, 2020), and Pt-Sn-In (Chu et al., 2012; Zhu, Sun, Yan, Li, & Xin, 2009)); and Pt-Bi-Ru(Pb) (Brandalise et al., 2009; Huang, Cai, Liu, & Guo, 2012; Huang, Cai, & Guo, 2013). Theoretical and experimental research has shown that the composition of a hybrid catalyst plays a crucial role in EOR.Generally, oxophilic metals, such as Sn, Ru, and Ni, primarily facilitate water dissociation to form adsorbed oxidant or surface oxides, which oxidize the surface-adsorbed C-containing species. Noble metals, such as Rh, Pd, and Ir, primarily enhance the capability of breaking CC bonds to increase the selectivity of CO2 generation. However, the roles of the components of a ternary catalyst vary dramatically depending on the structure of the catalyst. For example, in PtSn and PtRu binary systems, the degree of alloying, the oxidized state of the Sn/Ru, and the arrangement of the Sn/Ru oxides and Pt nanoparticles plays a decisive role in the EOR performance (Camara et al., 2004; Du et al., 2014; Gupta, Singh, & Datta, 2009; Hu et al., 2016; Liu et al., 2018; Pires, Corradini, Paganin, Antolini, & Perez, 2013; Silva et al., 2011; Zhang, Liu et al., 2019; Zhao et al., 2017; Zhu et al., 2020). In most cases, ternary catalysts have more than one phase, making them considerably more complicated than binary catalysts. Therefore, validating the structure is the first step and critical in investigating the mechanism of EOR. Given that there are numerous ternary catalysts, we classified them into two groups in terms of their structure: those with single-phase and those with segregated-phase structures, as shown in Fig. 3\n. The former includes random alloys with random ratios of components, intermetallic alloys with a particular ratio of components, and metal surface-rich alloys with a concentration gradient of a certain component. The latter includes bi-phase and tri-phase nanoparticles, which are elaborated in the following context. Consequently, it is difficult to attain improved catalytic activity simply by adding one or two factors. In reality, multiple factors interact to produce a beneficial effect. Thus, we will review and discuss the structures of the materials in the following sections.Many researchers have shown that the use of Pt-M binary alloys improves the kinetics of EOR compared with pure Pt. Incorporating a third metal into a binary system has been assumed to be an effective means of further improving the electrocatalytic activity of EOR. Relative to Pt, PtRh alloys are much better able to break the CC bond and attain complete oxidation (Almeida et al., 2019; Rao et al., 2014; Zhu, Bu et al., 2019). However, they have an issue of poisoning resulting from the adsorption of large amounts of CO after CC bond splitting. Introducing a third oxophilic metal, such as Ni, Co, Cu, or Fe, to form a random ternary alloy would alleviate this poisoning (Han et al., 2018, 2019; Han et al., 2019; Wang et al., 2017). Using Cu2O nanocubes as a template, Han et al. synthesized PtRhCu cubic nanoboxes via a galvanic reaction. The X-ray diffraction (XRD) pattern illustrated in Fig. 4\n(a) shows a single face-centered cubic phase, suggesting the formation of PtRhCu alloy (Han et al., 2018). When this is combined with the uniform distribution of elemental Pt, Rh, and Cu, as revealed by energy-dispersive X-ray (EDX) spectroscopy, we can conclude that a ternary random alloy has formed. In a 1\u202fM ethanol/1 KOH solution, Pt54Rh4Cu42 cubic nanoboxes exhibit a remarkably higher EOR current density than Pt58Cu42 cubic nanoboxes. For instance, Pt54Rh4Cu42 cubic nanoboxes have a peak current density of 4090\u202fmA mg\u22121, which is double that of the Pt58Cu42 cubic nanoboxes. They contribute to the improvement of both the facilitated CC cleavage due to the synergistic effect between the elemental Pt, Rh, and Cu, as well as the high oxidation activity of the COads intermediate in the presence of Cu, which favors the formation of OH species. Similarly, PtCuRh-alloyed nanowires were found to have a higher current density than PtCu nanowires because of their outstanding anti\u2212CO-poisoning properties (Chen et al., 2019).In addition, PtNiRh nanowires with an average diameter of 1.63\u202fnm (Fig. 5\n) were synthesized by reducing the metal acetylacetonate with glucose and oleylamine in the presence of the structure-direct agents of tungsten hexacarbonyl, dodecyl trimethyl, and ammonium chloride (Zhang, Huang et al., 2019). High-resolution transmission electron microscopy (HRTEM) images revealed that the thickness of an individual wire is about eight atom layers, as shown in Fig. 5(d). The uniform distribution of Pt, Rh, and Ni in Fig. 5(e) and the single phase, as shown in the XRD pattern in Fig. 5(g), prove the formation of a PtRhNi random alloy. For the alloy, the electron of the Pt tends to transfer to Rh or Ni, leading to less CO being adsorbed onto the Pt. These results reveal the effects of Rh on EOR and show that Pt69Ni16Rh15 nanowires have a considerably lower onset potential than Pt, as well as a mass-normalized current that is 3.26 times higher than that of Pt/C in 0.1\u202fM HClO4/0.5\u202fM ethanol. They contributed to the alleviation of CO poisoning due to the bifunctional mechanism of Rh and the electronic effect among the three metal elements. DFT calculations showed that, with the association of Ni 3d and Rh 4d synergetic d-orbital interplay, the Pt 5d band has been integrally pinned at the higher band center close to the Fermi level EF, thus balancing the reactivity and anti-poisoning and exhibiting long-term stability. The advantages of alloying with Rh and Ni have also been reported for other PtRhNi alloyed nanoparticles (Liu et al., 2017; Shen et al., 2014).PtRhM ternary alloys derived from PtRh binary alloys doped with M transition metal have proven effective for enhancing the current density of EOR (Erini, Rudi et al., 2015; Han et al., 2019; Shen et al., 2014; Wang et al., 2017). To elucidate the role of the third metal, Dai et al. synthesized a series of PtRhM (M\u202f=\u202fFe, Co, Ni, Cu, In, Sn, or Pb) ternary alloys with a fixed ratio of 3:1 between Pt and Rh by using the dimethylformamide solvothermal method (Dai, Wang et al., 2018). With negligible differences in their structures and a similar particle size, as shown in Fig. 6\n(a), they investigated the composition\u2013reactivity relationship for EOR in a 0.5\u202fmol CH3CH2OH/0.1\u202fmol HClO4 solution. At 0.45\u202fV vs. RHE, the specific activity occurred in the order of: Pt3RhSn/C\u202f>\u202fPt3RhIn/C\u202f>\u202fPt3RhGa/C\u202f>\u202fPt3RhPb/C\u202f>\u202fPt3RhCu/C\u202f>\u202fPt3RhCo/C\u202f>\u202fPt3RhNi/C\u202f>\u202fPt3RhFe/C\u202f>\u202fPt3Rh/C\u202f>\u202fPt/C (size of ca. 3.5\u202fnm), as shown in Fig. 6(b), suggesting that the PtRh alloy can be modified by the transition metals. With the help of DFT calculations, they stated that the Rh and M control the adsorption configuration of the key intermediates and thus change the barrier in the rate-determining step. The M further influences the adsorption and dissociation energies of water and tunes the d-band centers of Pt and Rh, thus modulating the catalytic activity.In addition to the improvement in the composition, Pt-based ternary materials can generate more edges with low-coordination sites in the presence of other metals. Candied haws-shaped AuPtNi alloy (Cui et al., 2020), Pt-Mo-Ni nanowires (Mao et al., 2017), ultrathin 2D PdPtCu nano-rings (Wang, Zhang et al., 2020), PtRhTe nanotubes (Jin et al., 2020), and PtPdCu nanodendrites (Wang, An et al., 2020) have numerous unsaturated edge sites that facilitate CC bond scission and CO oxidative removal, thereby enhancing the current density of EOR.Among the ternary alloys, the surfaces of some alloys are enriched by a metal forming a concentration gradient, which has a well-tuned geometric structure and modified electronic structures, resulting in an enhanced current density of EOR. Zhu et al. synthesized (111)-terminated Pt-Pd-Rh nanotruncated-octahedrons (NTOs) with a Rh-rich surface, as determined by extended X-ray absorption fine structure (EXAFS) analysis.(Zhu et al., 2015) Because Rh has the minimum standard reduction potential from RhCl6\n3\u2212, relative to Pd from PdCl4\n2- and Pt from PtCl6\n2-, Rh prefers to segregate on the surface of the particle after the reduction of PdCl4\n2- and PtCl6\n2-. The formed unique surface facilitates CC bond cleavage, and Pd on the surface provides OH for COad oxidation. Erini et al. prepared octahedral PtNiRh FCC alloy nanoparticles with Ni-rich facets and Pt-rich frames in terms of the EDX mapping depicted in Fig. 7\n (Erini et al., 2017) By varying the concentration of Rh, they were able to tune the surface composition, revealing that a material with a greater amount of Rh facilitates EOR. They found that the electronic effect facilitates CC bond cleavage to form C1 products on PtNiRh with a low Rh content. In contrast, C2 was the main product of a PtNiRh alloy with a high Rh content, due to the formation of an Rh shell outside the PtNi alloy core.In most cases, ternary catalysts tend to undergo phase segregation because of the differences in the surface free energy and chemical stability. Bi-phase heterostructures include a core-shell structure and a mixture of an alloyed phase patched by the other phase. Chen et al. synthesized a Ni core-PbPt alloy shell that exhibited significantly superior current density and stability compared with PtRu and Pt catalysts, resulting from the unique core-shell structure with the tuned d-band center of Pt. This weakened the CO binding to Pt and the fast electron transport in the presence of the Ni metal core (Chen et al., 2013). Dog-bone-shaped Au core-PtPd alloy shell nanoparticles were also synthesized to improve the performance of EOR (Dutta et al., 2016). Liang et al. prepared a Au core-PtIr monolayer core-shell catalyst for ethanol electrooxidation, which exhibited a current density that was six orders of magnitude greater than that of the AuPtIr random alloy, as shown in Fig. 8\n (Liang et al., 2019). In the case of the core-shell structure, the monatomic steps within the PtIr layer and the Au-induced tensile strain on the PtIr layer facilitate CC bond cleavage via ethanol dissociative adsorption. Meanwhile, Ir together with Au can dehydrate ethanol at very low potentials. Therefore, at the peak potential, the current for generating CO2 accounts for 57%, which is the highest selectivity towards complete oxidation so far. In addition to pure metal cores, binary alloy cores can also be used. Li et al. developed ternary CoPtAu nanoparticles with an intermetallic PtCo L10-tetragonal structure core-PtAu shell, which exhibited a higher current density and better stability than commercial PtRu and Pt catalysts in a 2 CH3CH2OH/0.1\u202fM HClO4 electrolyte (Li, Jilani et al., 2019). The stabilized Co inside the core induces surface compression on the PtAu shell, as revealed by HR-STEM and EXAFS analyses, thus facilitating the CC bond scission.In addition to the core-shell structure, two segregated phases are frequently formed in ternary systems: a binary alloy with another phase and a ternary alloy with another phase. Kowal et al. developed a cation-adsorption-reduction-galvanic displacement synthetic method to prepare the ternary Pt/Rh/SnO2 electrocatalyst, which consisted of a PtRh quasi-alloy and a SnO2 patch, which produced more CO2 than PtSnO2 (Kowal, Li et al., 2009). Assisted by DFT calculations for RhPt/SnO2(110), they concluded that Rh facilitates CC scission while SnO2 facilitated the dissociation of water to form an OH oxidant for oxidizing the C1 species. A similar phenomenon was observed for PtRh nanowires patched by SnO2 (Fan et al., 2019). Li et al. alloyed Pt with Ir on SnO2 nanoparticles and compared them with PtRh-alloy-patched SnO2 nanoparticles, as shown in Fig. 9\n (Li, Cullen et al., 2013). The PtIr/SnO2/C catalyst, with the highest Ir content (Pt/Ir/Sn\u202f=\u202f1:1:1), exhibited the most negative EOR onset potential together with a greatly improved CC bond-splitting capability. This was attributed to both the ensemble and ligand effects.For a PtRhSn ternary system, some researchers synthesized the structure in the form of a PtRhSn nanoalloy decorated with SnOx. Colmati et al. created a PtSnRh fcc structure with the Pt:Sn:Rh atomic ratio from 1:1:0.3 to 1:1:1 by reduction of the metal precursors with formic acid (Colmati et al., 2008), while Erini et al. obtained a PtRhSn alloy of Niggliite-phase-decorated SnOx with a Pt:Sn:Rh atomic ratio of 41:50:9, as shown in Fig. 10\n(a) and (b). Compared with the Pt/Rh/SnO2 nanoparticle reported by Kowal, Li et al. (2009), PtRhSn nanoparticles with a Niggliite phase decorated by SnOx exhibited current densities that were four times higher, at 0.45\u202fV, relative to RHE. This high current density was attributed to the active surface structure, where the PtRhSn sites interact strongly with the SnO2 moieties nearby, thus promoting CC bond cleavage and further oxidation to CO2. Du et al. synthesized ternary PtSnRh material to obtain a high current density for EOR (Du, Wang, LaScala et al., 2011). An EXAFS analysis revealed the coexistence of a homogeneous Pt/Sn/Rh random alloy and non-alloyed SnO2 throughout the catalyst. This led to the superior electronic effect of the Pt/Sn/Rh random alloy. Similarly, a Pt2SnCu alloy patched with SnO2 was developed to improve the current density owing to the synergistic catalytic effects of the Pt defects and SnO2 (Huang, Wu et al., 2015).Among all the ternary systems, the PtRhSn system has been the subject of the most intensive studies. The architecture of the PtRhSn system varies from that produced by the synthesis approach. Yang et al. prepared a PtRhSn material by reducing Pt and Rh precursors in ethylene glycol using SnO2 nanoparticles as seeds. A scanning electron energy loss spectroscopy (STEM-EELS) line scan revealed the formation of a PtRh core-Sn-rich shell (Yang et al., 2016). They further analyzed the core using EXAFS, as shown in Fig. 11\n(a), where it can be seen that the Pt and Rh are partially oxidized without Pt-Rh coordination. They concluded that the material is a triphasic PtRhOx\u2013SnO2 catalyst with a partially oxidized Pt and Rh core and a SnO2 shell, which realized a 2.5-fold increase in the CO2 generation rate for EOR compared with biphasic PtRhSnO2 catalysts with a metallic PtRh alloy core. This improved selectivity to CO2 can be attributed to the co-existence of metallic and oxidized Pt and Rh on the surface, whereas the metallic phase provides a large and available site for the dissociative adsorption of ethanol by CC splitting, while the oxidized Pt and Rh phases provide mobile O atoms for the oxidation of reaction intermediates, such as CO and CHx. Higuchi et al. synthesized a PtRhSn catalyst with segregated Pt, Rh, and SnO2 nanoparticles, which are partially in contact with each other, as shown in Fig. 11(b) (Higuchi et al., 2014). On the catalyst, SnO2 provides OH species from water dissociation to oxidize the dissociated CO at the Rh sites, while the Pt sites facilitate ethanol dehydrogenation. Thus, both SnO2 and Rh are necessary for an active electrocatalyst. Given its ability to break the CC bond of Re in the reforming process, researchers have attempted to use Re to replace Rh. Generally, a Pt/Re/Sn system has three segregated phases, Pt, Re, and SnO2, that are in physical contact with each other, as shown in Fig. 12\n (Drzyma\u0142a et al., 2018, 2020; Parlinska-Wojtan et al., 2019). This synergistic effect is a result of the assembly, enabling the ethanol molecule to be accessible to the three components, which are in direct contact with each other, with each playing an important role in the oxidation pathway. In addition, the interfaces between the particles are active for EOR.Dutta et al. synthesized NiAuPt nanoparticles on reduced graphene oxide nanosheets (NGs) consisting of tightly coupled nanostructures of Ni, Au, and Pt, which have neither an alloy nor a core-shell structure, as shown in Fig. 13\n (Dutta & Ouyang, 2015). NiAuPt-NGs with this unique architecture produce an 8, 4, and 2 times higher EOR current than is the case with monometallic Pt-NGs, bimetallic NiPt-NGs, and bimetallic AuPt-NGs, respectively, in an alkaline electrolyte. This can be attributed to the synergetic effect of the three nanostructured metals.Practically, if portable DEFCs are to be commercialized, it is essential that ethanol be completely oxidized to CO2 without partial oxidation to acetaldehyde and acetic acid. Although much effort has been expended on the synthesis of Pt-based ternary catalysts, the use of noble metals (Rh, Ir, and Pd) is essential to enhance the ability to break CC bonds with high CO2 selectivity. However, noble metals are a rare resource and, therefore, their cost prevents their widespread implementation. To remove COad or CHx,ad oxidatively, an oxophilic metal is required. However, these metals suffer from severe leaching during electrochemical oxidation over the long term, leading to damage to the structure and dramatically reducing the current density.First, the commercialization of DEFCs will demand that the cost of the EOR catalyst be reduced. To reduce the required amount of noble metals, the adoption of the core\u2013shell structure appears to have great potential. Technically, particles with a core of cheap metal and a shell fabricated from a thin layer of noble metals, configured in a manner that is highly active for EOR, have shown substantial promise. Second, maintaining stability after cycling within a large potential window and over long-term operation at a constant potential is a serious challenge for ternary catalysts. In most cases, the catalysts are degraded by the rearrangement of the bulk and/or surface structure, the agglomeration of particles, and the loss of the active components, specifically in the case of non-noble metals. To mitigate the etching of cheap metal, the metal atoms can be anchored at structural defects through strong interaction with the noble metal within the alloyed shell. In addition, alloying a second non-noble metal with the initial non-noble metal in a binary catalyst is an effective means of preventing the leaching of the non-noble metal. For example, adding Cu atoms to PtNi nanoparticles can stabilize the PtNi nanoparticles significantly, because the Cu atoms have a large orbital overlap with the Ni atoms. Last, increasing the ability to break the CC bond to achieve complete oxidation is the main focus. There are numerous possibilities for the structure of Pt-based ternary catalysts, which should exhibit appropriate selectivity for breaking the CC bond together with fast kinetics for oxidizing the C1 fragments at low potentials. With progress in surface/interface engineering and techniques enabling the precise control of synthesis at the atomic level, we believe that researchers will attain a breakthrough in this field in the near future.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the National Natural Science Foundation of China (grant No. 21373091), Guangdong Basic and Applied Basic Research Foundation (grant No. 2019A1515110035), and the State Key Laboratory of Pulp and Paper Engineering (grant No. 202013).", "descript": "\n                  Direct ethanol fuel cell (DEFC) as a promising device for converting chemical energy to electricity has been paid ever-increasing attention. However, the slow kinetics of ethanol electrooxidation at an anode hinders the application of DEFCs. Although Pt is the best catalyst among all the pure metal catalysts, it still has a relatively poor ability to break the CC bond, is deactivated by the accumulated COad intermediates, and undergoes unwanted desired structure change over long-term operation. In recent years, the addition of other metals to form binary, ternary, and quaternary catalysts have significantly improved electroactivity and stability. Ternary catalysts can have numerous element combinations and complicated architectures and, therefore, have been the subject of considerable research. In this review, most of the reported ternary catalysts will be summarized and categorized according to their structure while discussing the essence of the role of each component.\n               "}
{"full_text": "No data was used for the research described in the article.Friedel\u2013Crafts alkylations are important reactions for attaching alkyl chains to an aromatic ring and hence is of considerable industrial and pharmaceutical significance [1]. It proceeds through the electrophilic attack of an aromatic ring by an alkylating agent, such as olefins, alkyl halides, or alcohols [2\u20134]. Mineral acids like sulphuric acid and Friedel\u2013Crafts catalysts like AlCl3 and BF3 are used in the synthesis of many fine chemicals and pharmaceutical intermediates, though they have inherent drawbacks such as tedious catalyst recovery post reaction and large amounts of acid waste generation [5]. Hence solid acid catalysts such as zeolites, clays, metal oxides, metal organic frameworks have been thoroughly researched as potential catalysts for Friedel-Crafts alkylation reactions [6\u20139].Alkylation of benzene with isopropanol to yield cumene is a vital Friedel Craft\u2019s reaction as the global production of phenol and acetone is largely founded on the cumene process. Cumene also known as isopropyl benzene is a colorless liquid, having high antiknock value. It is also important in the production of cymene and polyalkylated benzenes [10]. Other industrial uses of cumene include the production of phenolic resins, bisphenol A, and caprolactam. During the isopropylation of benzene to cumene, 5\u201310\u00a0wt% diisopropylbenzene (DIPB) isomers are produced always as low-value byproduct. These can be recycled for cumene production, making the production process more economical [11]. A tremendous increase in cumene and phenol production has recently been reported worldwide [12]. In the last decades, several zeolite catalysts like zeolite UZM-8, silica supported beta zeolite, zeolites BEA and MWW catalysts have been reported in cumene synthesis [13\u201315]. Though cumene selectivity increased over beta-zeolite catalyst, benzene conversion dropped during alkylation of benzene with isopropanol. The decrease in benzene conversion was attributed to catalyst deactivation due to the presence of by products like ethylbenzene, p-xylene and 1-ethyl-3-(1-methyl) benzene [16]. In a similar study, Zou et al. achieved cumene production over nano-sized beta zeolite in a submerged ceramic membrane reactor, where the benzene isopropylation and the in-situ catalyst separation could be done [17].Spinel ferrites are important solid acid catalysts for various reactions like nitrogen fixation [18], CO hydrogenation [19], thermolysis of perchlorate [20], waste water treatment [21\u201322], synthesis of pyrazolopyridine derivatives [23] etc owing to their high activity and stability. As they exhibit admirable magnetic properties, these materials have the added advantage that they are magnetically recoverable, post reaction. Among them, cobalt ferrite that exists in a partially inverse spinel structure in which both sites (A and B) contain a fraction of Co2+ and Fe3+ cations are found to be highly active catalysts [24,25]. The catalytic and magnetic properties of cobalt ferrites can be modified by substitution of specific cations such as Cd, Ni, Zn, Sr, Zr and Cr etc as well as non-metals like N and S in the metal sites [26\u201327]. Elements that have not been prior selected were chosen for the present study. Though zirconia is of much technical importance it has not been used in ferrite doping. Also, non metal such as S, N and C have been used as dopants in metal oxides such as titania with excellent results, it has not been done in ferrites.\nIn the present study, we report the liquid-phase isopropylation of benzene using isopropyl alcohol (IPA) as alkylating agent over Zr/N/S doped cobalt nanoferrites. The catalysts were prepared by wet impregnation and coprecipitation techniques. The effect of reaction parameters like catalyst concentration, reaction temperature, and the mole ratio of reactants on reaction rate were analyzed to determine the feasibility of the catalyzed reaction. The structural stability and reusability of these catalysts also were studied in detail.\n\n\n\nZirconium doped cobalt ferrite nanoparticles Co1-xZrxFe2O4 (x\u00a0=\u00a00, 0.25, 0.5, 0.75 and1) and the N/S doped analogues were synthesized by co-precipitation method as reported in detail in our previous publications [28\u201329]. Required amounts of Fe(NO3)3\u00b79H2O, Co(NO3)2\u00b76H2O and ZrO(NO3)2\u00b7H2O were dissolved in deionized water and homogenized. 5\u00a0M NaOH solution was added drop wise with continuous stirring at 80\u00a0\u00b0C. The precipitate was washed repeatedly with deionised water to remove excess NaOH. It was dried for 24\u00a0h in an air oven at 110\u00a0\u00b0C and subsequently calcined for four hours at 700\u00a0\u00b0C. The prepared five ferrite compositions were designated as CoFe2O4, Co0.75Zr0.25Fe2O4, Co0.5Zr0.5Fe2O4, Co0.25Zr0.75Fe2O4 and ZrFe2O4. To assess the effect of non metal doping, N/S co doped CoFe2O4 and Zr/N/S co doped CoFe2O4 were synthesised as detailed before, the only difference being the addition of 0.1\u00a0M thiourea and 0.1\u00a0M urea solutions during homogenisation. In order to study the effect of non metal doping after the formation of spinel structure, samples were prepared by wet impregnation of 0.1\u00a0M thiourea and 0.1\u00a0M urea solutions. The samples were heated at 300\u00a0\u00b0C in a muffle furnace for six hours. The prepared samples were denoted as CF (CoFe2O4), CFTD (S/N co doped CoFe2O4), CFTA (S/N doped after CoFe2O4 formation), CZFTD (N/S/Zr co doped CoFe2O4) and CZFTA (S/N doped after Co/ZrFe2O4 formation) respectively. The N doped samples prepared using urea precursor were denoted as CFUD (N co doped CoFe2O4), CFUA (N doped after CoFe2O4 formation), CZFUD (N/Zr co doped CoFe2O4) and CZFUA (N doped after Co/ZrFe2O4 formation) respectively.\n\nThe crystallite structures of the synthesised particles were evaluated by Rigaku Miniflex 600 X-ray diffractometer (CuK\u03b1 as the radiation source). The particle size and morphology were evaluated using JEM 2100 model High Resolution Transmission Electron Microscope. A Bruker-S4-Pioneer model spectroscope was used to analyse the composition of the samples. Fourier transform infrared spectroscopic analyses were done in Thermo Nicolet Avatar 370 spectrometer in the range 400\u20134000\u00a0cm\u22121 with a resolution of 4\u00a0cm\u22121. Raman spectra were recorded in the spectral range 50\u20134000\u00a0cm\u22121 on a Bruker: RFS 27 Raman spectrometer using a laser source Nd: YAG 1064\u00a0nm. X-ray photoelectron spectra were obtained on Axis Ultra DLD instrument of KRATOS using monochromatic AlK\u03b1 radiation. Magnetic properties of the samples were characterized by a Lakeshore 7410 vibrating sample magnetometer at room temperature.Isopropylation of benzene was conducted in a 500\u00a0mL three-necked round bottom flask with a condenser under constant magnetic stirring at atmospheric pressure and 420\u00a0rpm of stirring speed. The alkylation reaction was studied under different conditions of reaction time, temperature, benzene/IPA ratio, and catalyst loading. The products were characterized by gas chromatography (GC) and GC\u2013MS analysis. GC used was Perkin Elmer Clarus 580 model Gas Chromatograph equipped with an Elite-5 capillary column and Flame Ionisation detector. The injector temperature was 250\u00a0\u00a0\u2103 and detector temperature was 275\u00a0\u2103. The GC\u2013MS used is a Varian 1200 Single Quadrupole mass spectrometer with helium as the carrier gas.Reusability of the prepared catalysts was checked for four consecutive catalytic runs by magnetic separation of the catalysts on the completion of each run. The spent catalysts were washed with acetone, followed by drying in an air oven at 200\u00a0\u2103.The prepared doped nano ferrites were characterised by standard techniques and the results have been detailed in our previous publications [28\u201329]. The obtained X-ray diffraction profile could be indexed to single-phased cubic spinel systems with average sizes in the range of 16\u201326\u00a0nm. Fourier Transform Infrared and Raman spectra confirmed the presence of stable cubic spinel ferrite structure. Nearly spherical particles were seen on TEM images. X-ray photoelectron spectroscopy indicated the substitution of zirconium ions mostly into octahedral sites, suggesting a partially inverse spinel structure. Interestingly, the sulphur and nitrogen were inserted into the lattice structure as cations. Decrease in saturation magnetisation on doping was observed by magnetic measurements at room temperature whereas coercivity increased with reduction in particle size.In this work, the efficiency of the synthesized Zr/N/S doped cobalt ferrite magnetic nanocatalysts for benzene isopropylation is evaluated. Optimization of various reaction parameters such as temperature, molar ratio of substrates, time and catalyst dosage was done by changing the variable while keeping all other parameters constant. Results of the optimization studies of reaction conditions done by taking CZFTD nanocomposite as the reference catalyst are discussed in the subsequent sections.\nFig. 1\n depicts the effect of temperature on isopropylation of benzene at four different temperatures viz 130, 140, 150 and 160\u00a0\u00b0C keeping catalyst amount as 0.5\u00a0g, and benzene to isopropanol mole ratio as 1:3. It can be noted from the figure that the reaction temperature has significant effect on reaction rate. The % conversion of IPA increases from 24.9\u00a0% to 77.2\u00a0% with increase in temperature. The major products found are cumene and di/tri isopropyl benzene. However, the selectivity to cumene, which is high at lower temperatures, decreases gradually with increase in temperature to 160\u00a0\u00b0C. This may be due to subsequent alkylation to di/tri isopropyl benzene, which takes place at high temperatures [30]. Maximum cumene selectivity of 80.8\u00a0% with the good conversion of IPA (70.8\u00a0%) is obtained at 150\u00a0\u00b0C. From the results, considering the reaction rate, selectivity and energy cost, 150\u00a0\u00b0C is chosen as the optimum temperature for further studies.Reactant molar ratio plays an important role in determining reactant conversion and product yield. The effect of mole ratio of reactants on conversion and product selectivity were studied by taking benzene: isopropanol ratio as 1:1, 1:3, 1:5, and 1:7 at 150\u00a0\u00b0C with catalyst weight 0.3\u00a0g for 2\u00a0h. The results are shown in Fig. 2\n. Conversion of isopropanol increases as expected, on increasing the mole ratio of the reactants up to 1:5. Further increase causes reduction in conversion due to the blocking of active surface sites with excess reactants [31]. Selectivity to cumene also increases from 65.7\u00a0% to 83.6\u00a0% with increase in the mole ratio from 1 to 5. Cumene selectivity then decreases to 70.6\u00a0% on increasing the mole ratio to 1:7. The decrease in selectivity of di and tri substituted products at lower molar ratio is due to the transalkylation of these products, attributed to the increased availability of alkyl groups. From this study, it can be deduced that high concentration of IPA is not preferred and hence a minimum ratio 1:5 was maintained for further studies.\nFig. 3\n shows the effect of catalyst concentration on benzene isopropylation. As the catalyst concentration increases from 0.3\u00a0g to 1.5\u00a0g/L, there is a noticeable increase in IPA conversion from 58.9\u00a0% to 90.3\u00a0% and decrease in cumene selectivity from 83.6\u00a0% to 63.7\u00a0%. A possible explanation for these observations is that an increase in the number of active sites accelerates the reaction rate with concomitant enhancement in the rate of disproportionation, which could adversely affect the cumene selectivity [32]. The % conversion of di/tri-substituted benzene also increases slightly with increase in catalyst concentration. Considering the results in Fig. 3, the catalyst concentration of 0.5\u00a0g/L was employed for further studies.The optimized reaction conditions are presented in Table 1\n.The catalytic activities of all the prepared catalysts for isopropylation of benzene were evaluated under the optimized reaction conditions and the results are given in Table 2.\n. Cobalt ferrite shows 60.2\u00a0% conversion of IPA with a cumene selectivity of 63.90\u00a0%. Zr doping increases the catalytic activity. Among the different Zr doped catalysts, Co0.5Zr0.5Fe2O4 shows maximum conversion of 73.5\u00a0% with selectivity of 65.8\u00a0%. The high selectivity of Co0.25Zr0.75Fe2O4 and ZrFe2O4 maybe due to the presence of lattice imperfections that provides catalytically active sites. Further doping with nonmetals such as N and S increases the reaction rate. In Zr/N doped samples, CFUD shows a maximum conversion of 80.5\u00a0%. Cumene is the major product over all the catalytic systems. In Zr/N/S doped series, CZFTD shows a maximum conversion of 82.7\u00a0% with selectivity of 83.6\u00a0%. The catalytic efficiency of doped cobalt ferrite nanocomposites can be attributed to two reasons. One is the increased surface area; Zr/N/S substitution increases the surface area of cobalt ferrite nanocomposites in a noticeable manner. Another reason is the improved surface sites due to the presence of Co3+/Co2+, Fe3+/Fe2+, Zr2+/Zr4+, and non-metal cationic species, which is directly related to the catalytic efficiency towards alkylation reactions [33]. Thus, it can be concluded that the prepared catalysts effectively catalyze the isopropylation of benzene at relatively low temperature and catalyst load.The schematic diagram of the reaction pathway is depicted in Fig. 4\n. Along with cumene, di/tri substituted products also is formed, which then gets converted to the monosubstituted product viz cumene.Friedel-Crafts isopropylation of benzene with isopropanol is a well-known aromatic substitution reaction and it involves a complex reaction network. This reaction is sometimes associated with several side reactions which result in the formation of various products through transalkylation, dispropotionation, dealkylation etc [34]. The formation of n- propylbenzene is always accompanied by the production of cumene, which is formed either by the isomerisation reaction of cumene or from the primary n-propyl cation formation before isopropylation of benzene. From GC\u2013MS analysis, cumene was found to be the major component whereas di/tri isopropyl benzene was the minor product along with minor quantities of propene, n-propyl benzene, and other aliphatic products. Similar observations were also made by other researchers [35]. Based on this a suitable mechanism is suggested for the isopropylation of benzene. Water is a byproduct formed in the isopropylation reaction. Due to the electron deficient dopant cations of Zr, N and S, the catalyst surface becomes highly hydrophilic as evident by the FT-IR analysis. Therefore, the catalysts easily become proton donors by the abstraction of water molecules. The isopropanol is converted into propene by dehydration followed by the proton abstraction from the catalysts sites resulting in the formation of a carbocation. Isopropyl substitution on the aromatic ring is possible by the attack of the carbocation after the removal of H+ back to the catalyst. The main reaction steps can be represented as follows:The reusability of the spent catalyst was studied by separating it from the reaction mixture by magnetic separation on the completion of a catalytic run. The catalyst removed from the reaction mixture was washed thoroughly with acetone to remove organic matter, dried, calcined, and reused for another three catalytic runs under the same reaction conditions. The results of the reaction in 2\u00a0h are shown in Fig. 5\n. No remarkable fall in the activity was observed till two consecutive runs whereupon it slightly reduces in the third and fourth cycles. The selectivity of the catalyst towards cumene remains almost constant till its fourth cycle. The reduction in the activity can be explained by two reasons; the deactivation of the catalyst by structural and surface changes due to its interaction with substrate molecules and the deposition of organic matters, especially the byproduct DIPB [36]\nThe structural stability of the spent catalyst was analyzed by recording the XRD pattern and surface area of CZFTD catalyst. PXRD pattern of the fresh and reused catalyst remains the same, indicating the phase stability of the prepared nanoferrites (Fig. 6\n). Average crystallite size increases due to agglomeration of particles. Surface area decreases significantly due to the incorporation of particles into the surface sites [37] (Table 3\n) (see Fig. 7\n).Zr/N/S doped nanoferrite catalysts efficiently catalyze the isopropylation of benzene offering an alternate process for the selective cumene preparation at low temperature and catalyst loading. Comparatively high activity was observed with nitrogen and nitrogen/zirconium doped samples. The critical role played by the different reaction parameters on the catalytic activity and selectivity is noteworthy. Monoalkylated product is the major product along with negligible formation of polyalkylated products. The isopropylation activity of the catalyst is found to be dependent on the number and strength of active surface sites, which in turn depends on the nature and concentration of dopant. The prepared catalytic systems are found to be reusable and resistant to rapid deactivation. The remarkable feature of this catalytic system is that it can be readily isolated from the reaction mixture by employing an external magnet.All authors contributed to the study conception and design, material preparation, data collection and analysis. Both authors contributed equally towards the data analysis and writing. All authors read and approved the final manuscript.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.", "descript": "\n                  Nitrogen, sulphur and zirconium doped cobalt ferrites prepared by coprecipitation method were evaluated as catalysts for the production of cumene from benzene and isopropyl alcohol. The catalytic efficiency of cobalt ferrite towards the reaction increased due to the doping by zirconium and non-metals. The isopropylation activity is decidedly dependent on the number and strength of active surface sites that in turn depends on the nature and concentration of dopant. Reaction variables like temperature, reactant ratio and catalyst load affected the reaction rate considerably. On successive uses, the catalysts were found to be magnetically recoverable and stable.\n               "}
{"full_text": "In the current transition period towards green energy, the implementation of carbon capture and utilization (CCU) strategies is one of the biggest challenges to achieve the goal of decarbonization and tackle climate change. In this scenario, the technological development of efficient routes for the large scale conversion of CO2 into value-added products is imperative to offset the cost of its capture and storage (Hepburn et al., 2019; Kamkeng et al., 2021; Zhang et al., 2020).The catalytic processes for the conversion of CO2 into fuels and raw materials (as olefins and aromatics) receive great attention for their implementation in the refineries of the future (Garba et al., 2021; Leonzio, 2018; Ye et al., 2019) and are complemented by other initiatives aimed at intensifying the recovery of oil and natural gas (Alabdullah et al., 2020; Palos et al., 2021). As an alternative to the well-developed two-stage hydrocarbon production technologies from CO2 hydrogenation to methanol/dimethyl ether (DME) (Sehested, 2019) and its selective conversion into olefins, gasoline or aromatics (Tian et al., 2015), hydrocarbon synthesis routes in one stage (modified Fischer Tropsch (MFT) or with oxygenates, methanol/DME, as intermediates), through cascade reactions and with tandem catalyst (Ma and Porosoff, 2019; Wei et al., 2021), have the attraction of lower equipment cost and higher CO2 conversion. The thermodynamic equilibrium of methanol synthesis is displaced by the in situ conversion of oxygenates into hydrocarbons, which allows the integrated process to be carried out at higher temperatures and at moderate pressure (15\u201330\u00a0bar) compared with the conventional methanol synthesis. It is remarkable that this facilitates the supply of H2, using commercial PEM electrolyzers.The CO2 hydrogenation route with oxygenates as intermediates is more attractive than the MFT reaction (based on Fe and Co catalysts) for the selective production of hydrocarbons, as it is not conditioned by the Anderson-Schulz-Flory (ASF) distribution. Using appropriate zeolites (SAPO-34, HZSM-5, Hbeta or HY, the most studied), the oxygenate intermediate route can be selectively addressed towards the production of olefins, aromatics or gasoline. The acidity and appropriate shape selectivity are key features of the catalyst for this purpose (Ramirez et al., 2019; Wei et al., 2017). Comparing the deactivation of the catalysts used in oxygenates conversion into olefins, that is, MTO (methanol to olefins) and DTO (DME to olefins) processes, the high partial pressure of H2 in the integrated CO2 to olefins process contributes to minimizing the formation of coke on the acid catalyst (Nieskens et al., 2018), which is a relevant feature that conditions the feasibility of the overall process and the configuration of the reaction equipment (Cordero-Lanzac et al., 2020b; Tian et al., 2015).Although conventional Cu based catalysts (with of Cu/Zn, Cu\u2013ZnO\u2013Al2O3, Cu\u2013ZrO2 and Cu\u2013ZnO\u2013ZrO2 configurations, among others), are very active and selective for the synthesis of methanol from syngas, in the hydrogenation of CO2 (with high H2O concentration in the medium) and especially in the conditions required for the synthesis of hydrocarbons (above 300\u00a0\u00b0C), suffer severe deactivation by sintering and are particularly active for the rWGS reaction (Marcos et al., 2022). The adequate activity of In2O3 for the synthesis of methanol under these conditions, especially from pure CO2 source, is accepted in the literature (Ara\u00fajo et al., 2021a; Martin et al., 2016). Numerous experimental and theoretical studies delve into the reaction mechanism of In2O3, whose activity is attributed to its CO2 adsorption capacity in the superficial oxygen vacancies and H2 dissociation (Frei et al., 2018; Wang et al., 2021; Ye et al., 2013), favoring the advance of the reaction mechanism with formate ions as intermediates (Chen et al., 2019). CO2 is successively hydrogenated: CO2* \u2192 HCOO* (formate) \u2192 H2COO* (dioxymethylene) \u2192 H3CO* (methoxy) \u2192 CH3OH (Chen et al., 2019; Gao et al., 2017; Ye et al., 2013, 2014).However, at the temperature required for the direct synthesis of hydrocarbons as well, In2O3 presents limitations due to: i) Thermodynamics, because secondary endothermic reactions (rWGS and methanation) are favored; ii) partial sintering, favored by the required higher temperature and the presence of H2O and CO (Wang et al., 2021). To increase the hydrogenation activity and the selectivity to methanol of In2O3, and reduce sintering deactivation, different strategies have been used (Wang et al., 2021): i) Improving the dispersion of In2O3 and increasing the oxygen vacancies, ii) promoting the dissociative H2 adsorption and spillover; iii) promoting the activation of CO2; iv) stabilizing key reaction intermediates, and v) generating new types of active sites.In these strategies, the use of supports and promoters has been of great importance. From studies on the synthesis of methanol (Ara\u00fajo et al., 2021a; Zhang et al., 2018) and olefins in one stage (Gao et al., 2018), the role of ZrO2 as carrier in favoring the dispersion of In2O3 and generating new vacancies by the formation of epitaxially-grown In2O3 or solid In2O3\u2013ZrO2 solutions is well established (Nieskens et al., 2018; Ramirez et al., 2019), with the consequent increase in the adsorption capacity of CO2 and to attenuate the sintering of In2O3 (Alabdullah et al., 2020; Garba et al., 2021). Additionally, In2O3 can be combined with other metals, active for hydrogenation reactions, such as Zn (Palos et al., 2021), Ni (Ara\u00fajo et al., 2021b; Jia et al., 2020), Co (Bavykina et al., 2019; Pustovarenko et al., 2020), Au (Rui et al., 2020), Rh (Li et al., 2020), Pt (Han et al., 2021) or Pd (Ara\u00fajo et al., 2021b; Frei et al., 2018; Snider et al., 2019), to increase CO2 conversion and methanol selectivity. Various studies in the literature compare In2O3 based catalysts either for methanol production from syngas (Su et al., 2018) or by CO2 hydrogenation, nonetheless, the literature studying the joint hydrogenation on CO2+CO mixtures is scarce (Ara\u00fajo et al., 2021b). The co-feeding of syngas together with CO2 is interesting from various perspectives: i) It allows the joint valorization (avoiding separation costs) of streams derived from the gasification of biomass or wastes of the consumer society (where CO2 and syngas are present) (Couto et al., 2013; Lopez et al., 2015), ii) considers the real need to recirculate the stream of unreacted gases in the synthesis of methanol and direct synthesis of olefins (Ara\u00fajo et al., 2021b), and; iii) contributes to the necessary supply of H2. Ara\u00fajo et al. (2021a) have verified the capacity of different catalysts (In2O3\u2013ZrO2, Cu\u2013ZnO\u2013Al2O3 and ZnO\u2013ZrO2) for the hydrogenation of mixtures of CO and CO2, under suitable conditions for the synthesis of methanol, verifying the favorable effect of the presence of CO on the controlled formation of surplus oxygen vacancies in In2O3. However, the effect on the yield and selectivity of methanol and on the deactivation of the catalyst by sintering is complex, since CO acts as a reducing agent (which favors the sintering of In2O3 by over-reduction). Another factor influencing the deactivation of the In2O3 catalyst for these feedstocks is the concentration of H2O, which increases as CO2 conversion increases. Several authors (Frei et al., 2018; Ye et al., 2014) have verified in the synthesis of methanol that a limited concentration of H2O favors the formation of methoxy ions, increasing the yield of methanol. However, an excess of H2O leads to annihilate the oxygen vacancies, to the aggregation of In species, and decrease of In0 species, affecting the dissociation capacity of H2, and so, the overall performance of the catalyst for CO2 hydrogenation.The aforementioned results in the literature show the need to progress in the knowledge and improvement of catalysts for the synthesis of methanol from CO2 under the reaction conditions required for the integrated process (CO2 to hydrocarbons), and especially when CO is co-fed given the interplay of CO/H2O. The reaction conditions for maximizing CO2 conversion and hydrocarbon production while limiting catalyst deactivation by sintering must also be optimized, since the CO2 to hydrocarbons process is conducted at a higher temperature than the individual stage of methanol formation and at higher concentration of H2 than oxygenates to hydrocarbons conversion. As to contribute filling this shortage, in this work, the effect of the Zr/In ratio on the In2O3\u2013ZrO2 catalyst has been studied in the synthesis of methanol from CO2/syngas mixtures. Moreover, the selected In2O3\u2013ZrO2 catalyst (based on its good kinetic performance) has been used to conform a In2O3\u2013ZrO2/SAPO-34 tandem catalyst, and its performance has been further studied (in terms activity-selectivity-stability) in the direct synthesis of olefins. SAPO-34 has been selected (Dang et al., 2019) as acid catalyst for this purpose due to its well-known suitable behavior (highly selective) in the conversion of methanol (and/or DME) into olefins. The results are explained according to the properties of the catalysts determined by different analysis techniques.The In2O3\u2013ZrO2 catalysts have been synthesized following a co-precipitation method (S\u00e1nchez-Contador et al., 2018). Metal nitrates solutions, In(NO3)3 (Sigma-Aldrich) and Zr(NO3)4 (Panreac) with the desired Zr/In atomic ratio (0 (In2O3), 1:3, 1:2, 1:1, and ZrO2) and total metal concentration of 1\u00a0M were co-precipitated over 20\u00a0mL of deionized water under stirring, with ammonium carbonate (Panreac, 1M), at 70\u00a0\u00b0C and neutral pH. The mixtures were aged for 2\u00a0h to ensure the complete co-precipitation, and then filtered and cleaned with deionized water until neutral supernate was obtained. Finally, the resulting powders were dried and calcined at 500\u00a0\u00b0C for 1\u00a0h and pelletized, crushed and sieved to the desired particle size (125\u2013250\u00a0\u03bcm).SAPO-34 acid catalyst (ACS Material) was calcined at 550\u00a0\u00b0C for 5\u00a0h, pelletized, crushed and sieved to the desired particle size (300\u2013450\u00a0\u03bcm). The tandem catalyst was composed by physical mixture of pelletized In2O3\u2013ZrO2 metallic catalyst and SAPO-34 acid catalyst in a 2/1 mass ratio. The different particle size of both functions allowed analyzing the spent catalysts independently.The physical properties of the catalysts (BET specific surface area and pore volume) have been determined by N2 temperature programmed adsorption-desorption (N2-TPD) analyses (Micromeritics ASAP 2010) at \u2212196\u00a0\u00b0C. The procedure consists on a previous conditioning stage of the sample, on which degassing is carried out at 150\u00a0\u00b0C under vacuum (10\u22123\u00a0mmHg) for 8\u00a0h to eliminate impurities and remove the H2O adsorbed on the surface of the catalyst sample, facilitating N2 sorption. Subsequently, serial equilibrium stages of N2 adsorption-desorption are carried out until the complete saturation of the sample at cryogenic temperature of liquid N2. The pore volume is calculated with the BJH method using the adsorption branch of the isotherm.The chemical composition has been quantified and qualified by X-Ray fluorescence (PANalyticalAxios) and the structure by means of X-Ray diffraction (PANalyticalXpert PRO) and XRD vs temperature analyses. For determining the metallic properties, H2 temperature programmed reduction (H2-TPR) and CO-TPR analyses were carried out (MicromeriticsAutochem 2920). Briefly, 100\u00a0mg of sample were first swept with He to eliminate impurities and adsorbed H2O. After stabilizing the catalyst in the corresponding mixture (10% H2 or CO, in Ar), the samples were heated up to 800\u00a0\u00b0C at a 2\u00a0\u00b0C min\u22121 rate, and reference and analyzed streams were compared.The same equipment was used for CO2-TPD analyses and for acidity measurements (TPD-NH3). For CO2-TPD the following steps were used: i) 30\u00a0min of He sweeping (160\u00a0mL\u00a0min\u22121) at 550\u00a0\u00b0C, for eliminating possible impurities and adsorbed H2O; ii) stabilization at 150\u00a0\u00b0C with He (20\u00a0mL\u00a0min\u22121); iii) sample saturation by CO2 injection (5\u00a0mL\u00a0min\u22121) at 50\u00a0\u00b0C; iv) He sweeping (20\u00a0mL\u00a0min\u22121) to remove the physisorbed adsorbate; and, v) desorption by heating the sample with a controlled temperature ramp (5\u00a0\u00b0C min\u22121) from 50 to 400\u00a0\u00b0C, the operating reaction temperature.Analogous technique was used for NH3-TPD analyses, using 50\u00a0\u03bcL\u00a0min\u22121 NH3 injections at 150\u00a0\u00b0C for the saturation of the sample and 5\u00a0\u00b0C min\u22121 temperature ramp for the desorption step, up to 550\u00a0\u00b0C.The reaction runs have been carried out in an isothermal PID Eng&Tech fixed bed reactor. The reactor dimensions are: 9\u00a0mm internal diameter and 10\u00a0cm of effective length and is made of 316 stainless steel. The equipment can operate up to 700\u00a0\u00b0C, 100\u00a0bar and with catalyst loadings up to 5\u00a0g. The catalyst was mixed with an inert (SiC) to ensure isothermal conditions and to avoid preferential pathways. The reactor outlet stream was heated up to 110\u00a0\u00b0C to avoid products condensation, and analyzed online in a gas-chromatograph (microGC Varian CP4900). For this analysis three modules were used: i) Molecular sieve (MS-5) to quantify H2, N2, O2 and CO; ii) Porapak Q (PPQ) for CO2, water, C1\u2013C4 hydrocarbons and MeOH/DME; and iii) 5CB column (CPSiL) for higher hydrocarbons. Typically, the reaction runs were carried out at 400\u00a0\u00b0C, 30\u00a0bar, H2/COX ratio of 3 and with 125\u00a0mg of catalyst. These conditions were established in a previous work as suitable for the joint valorization of CO2 and syngas into olefins (Gao et al., 2018; Portillo et al., 2021). H2, CO and CO2 flowrates were adjusted to get a 3.35 gcat h molC\n\u22121 space-time value with the corresponding CO2/COX ratio (between 0, corresponding to 100% CO; and 1, corresponding to 100% of CO2). The reaction system has been described in more detail elsewhere (Portillo et al., 2021).In order to quantify the obtained results, the following reaction indices have been defined. The conversion of CO and CO2:\n\n(1)\n\n\n\nX\n\nC\n\nO\nx\n\n\n\n=\n\n\n\nF\n\nC\n\nO\nx\n\n\n0\n\n\u2212\n\nF\n\nC\n\nO\nx\n\n\n\n\n\n\n\nF\n\nC\n\nO\nx\n\n\n0\n\n\n\n\u00b7\n100\n\n\n\nwhere \n\n\nF\n\nC\n\nO\nx\n\n\n0\n\n\n is inlet molar flowrate in content C atoms, and \n\n\nF\n\nC\n\nO\nx\n\n\n\n\n its analogous at the reactor outlet stream.Similarly, CO2 conversion, \n\n\nX\n\nC\n\nO\n2\n\n\n\n\n has been defined as:\n\n(2)\n\n\n\nX\n\nC\n\nO\n2\n\n\n\n=\n\n\n\nF\n\nC\n\nO\n2\n\n\n0\n\n\u2212\n\nF\n\nC\n\nO\n2\n\n\n\n\n\nF\n\nC\n\nO\n2\n\n\n0\n\n\n\u00b7\n100\n\n\n\nwhere \n\n\nF\n\nC\n\nO\n2\n\n\n0\n\n\n and \n\n\nF\n\nC\n\nO\n2\n\n\n\n\n are the CO2 molar flowrates at the inlet and outlet of the reactor, respectively. Carbonaceous products yields (Y\n\ni\n) and selectivities (S\n\ni\n) (except for CO and CO2) have been defined according to Eqs. (3) and (4), respectively, by grouping the products into the following lumps: methane, C2\u2013C4 olefins, C2\u2013C4 paraffins, oxygenates (MeOH and DME) by the use of the following expressions:\n\n(3)\n\n\n\nY\ni\n\n=\n\n\n\nn\ni\n\n\u00b7\n\nF\ni\n\n\n\nF\n\nC\n\nO\nx\n\n\n0\n\n\n\u00b7\n100\n\n\n\n\n\n\n(4)\n\n\n\nS\ni\n\n=\n\n\n\nn\ni\n\n\u00b7\n\nF\ni\n\n\n\n\n\n\u2211\n\ni\n\n\n(\n\nn\ni\n\n\u00b7\n\nF\ni\n\n)\n\n\n\n\u00b7\n100\n\n\n\n\nbeing ni the number of C atoms in a molecule of component i and F\n\ni\n the molar flowrate of the component i at the reactor outlet stream.The carbon balance in all experiments was closed over 99%.In this section, first, the effect of the Zr loading on the properties of the In2O3\u2013ZrO2 catalyst has been studied. Second, a comparison of the performance of the catalysts for CO2/CO mixtures hydrogenation has been carried out in order to select the most suitable Zr/In ratio to favor methanol production. The reaction runs have been carried out under the reaction conditions required for the direct olefins synthesis process, pursuing to use the selected In2O3\u2013ZrO2 catalyst in a In2O3\u2013ZrO2/SAPO-34 tandem for this reaction. Catalyst screening has been carried out attending to activity, selectivity to methanol and stability criteria. Finally, the performance of the In2O3\u2013ZrO2/SAPO-34 tandem catalyst has been assessed for the direct synthesis of olefins from CO2/CO mixtures.Attending to the physical properties of the catalysts (Table 1\n) determined by N2-TPD analyses, the pore volume of In2O3 is higher (0.25\u00a0cm3\u00a0g\u22121) than that of ZrO2 (0.16\u00a0cm3\u00a0g\u22121). Consequently, the pore volume of the composite In2O3\u2013ZrO2 catalysts decreases with increasing Zr loading in the catalyst. As to the BET specific surface area (SBET) regards, being 53\u00a0m2\u00a0g\u22121 for In2O3 and 96\u00a0m2\u00a0g\u22121 for ZrO2, the SBET of In2O3\u2013ZrO2 catalysts increases upon increasing Zr/In ratio, in agreement with the results reported by Frei et al. (2020). Indeed, higher SBET than expected from the Zr/In ratio has been obtained for the In2O3\u2013ZrO2 catalysts.As to the chemical and metallic properties characterization regards, XRF analyses have been carried out to ascertain the co-precipitation of the metals in the desired Zr/In ratio. In Table 2\n, the nominal and measured metal ratios are listed.As to analyze the morphology of the catalysts, XRD patterns for the different catalysts are depicted in Fig. 1\n. According to these spectra, pure In2O3 and ZrO2 catalysts show the typical peaks for these structures. At 2 \n\n\u03b8\n\n\u00a0=\u00a021.68\u00b0, 30.74\u00b0, 35.61\u00b0, 37.88\u00b0, 40\u00b0, 42.03\u00b0, 43.99\u00b0, 45.86,\u00b0 51.14\u00b0, 52.84\u00b0, 56.14\u00b0, 59.245\u00b0, 60.784\u00b0, 62.33\u00b0 and 63.79\u00b0 for In2O3, and; at 2 \n\n\u03b8\n\n\u00a0=\u00a030.6876\u00b0, 35.5315\u00b0, 51.0336\u00b0, 60.6445\u00b0, 63.0967\u00b0, 74.9399\u00b0 for ZrO2. For the composite catalysts, that is, for the mixture of In2O3 with ZrO2, a combination of those in good agreement with the Zr/In ratio reported in Table 2 is observed. Moreover, the results evidence that ZrO2 coexists in its monoclinic (most favored thermodynamically) (Martin et al., 2016) and tetragonal structure, whereas it changes completely into its tetragonal polymorph with the incorporation of In2O3 in the In2O3\u2013ZrO2 catalysts as determined in the literature (Frei et al., 2019). Moreover, from further XRD vs temperature measurements carried out, the structure of the In2O3\u2013ZrO2 catalysts is expected to remain stable under the reaction temperatures used in the direct CO2/syngas to olefins process. Additionally, the Rietveld calculations carried out evidenced the presence of Zr atoms in the In2O3 structure and of In atoms integrated in the ZrO2 structure, which is consistent with previous findings (Artamonova et al., 2006; Frei et al., 2020; Portillo et al., 2021). For the 1Zr\u20131In catalyst, 42.6% of In2O3 structure and 57.4% of ZrO2 structures were determined. The metal content within the In2O3 structure, is divided into 81.2% of In, and 18.8% of Zr. Likewise, within in the ZrO2 structure, 76.2% stands for Zr and 23.8% for In. These results are consistent with the suggestion in the literature that indium-zirconium composite oxides are not simple mechanical mixtures but generate active composite In1-xZrxOy oxides (Dang et al., 2018).The reducibility of the catalysts has been studied by H2-TPR and CO-TPR analyses (Fig. 2\n, where the TCD signals have been normalized for In2O3 mass). With this approach, gathering information on the H2 splitting activity (H2 desorption) and on the reducibility of the catalyst in the reaction medium is pursued. Prior to both H2 and CO-TPR, the samples were swept with He at 200\u00a0\u00b0C for 1\u00a0h to eliminate impurities and adsorbed water. Later on, the samples were cooled down to 30\u00a0\u00b0C, the inlet gas changed to H2/CO and temperature increased after attaining a stable baseline at 30\u00a0\u00b0C.As expected, ZrO2 is not reduced under the studied TPR conditions and so, it is not expected either under the not so severe reaction temperature used. As observed, the combination of In2O3 and ZrO2 incurs peaks at higher temperatures for In2O3\u2013ZrO2 catalyst compared to In2O3 and ZrO2, for both reducing agents. Comparing H2-TPR (Fig. 2a) and CO-TPR (Fig. 2b), CO presents higher reduction capacity, which is in accordance with the previous results (Chen et al., 2019; Dang et al., 2018; Frei et al., 2018; Martin et al., 2016). The results also indicate a favorable effect of the addition of ZrO2 on the number of In2O3 sites accessible to H2 and CO (greater area under the curve per unit mass of In2O3). It is noteworthy that the presence of ZrO2 favors the reduction of In2O3 with CO at low temperature, as a peak is observed with a maximum between 275 and 300\u00a0\u00b0C for composite In2O3\u2013ZrO2 catalysts. The effect of CO as vacancy generator (Martin et al., 2016; Wang et al., 2021) will contribute to this result, also favoring its adsorption and that of CO2.\nFig. 2 also shows a greater resistance to reduction of In2O3 sites due to the presence of ZrO2. This lower reducibility is consistent with the larger crystal size of In2O3 observed with increasing Zr/In ratio (Table 3\n). In addition, it is well established in the literature that the presence of ZrO2 hinders the sintering of In2O3 (Ara\u00fajo et al., 2021b), which is consistent with the lower reducibility observed in Fig. 2 for moderate ZrO2 contents in the composite catalyst.CO2-TPD analysis have been carried out for all the catalysts to quantify their CO2 adsorption capacity, since this is a key feature for their activity for oxygenates synthesis. The results are plotted in Fig. 3\n. According to these profiles, In2O3\u2013ZrO2 catalysts outperform significantly the CO2 adsorption capacity of the parent In2O3 and ZrO2 catalysts.The responsibility of the oxygen vacancies in In2O3 on the CO2 adsorption capacity and activity for hydrogenation to methanol is well established (Martin et al., 2016; Sun et al., 2015). Consequently, the higher CO2 adsorption capacity of the In2O3\u2013ZrO2 catalyst than of In2O3 observed in Fig. 6\n\n\n is attributable to its higher density of oxygen vacancies. Dang et al. (2018) determine by X-ray photoelectron spectroscopy (XPS), CO2-TPD analysis, and periodic DFT calculations that the incorporation of ZrO2 into In2O3 generates interactions in the electronic structure, the formation of In1-xZrxOy mixed oxide and the formation of additional oxygen vacancies, increasing CO2 conversion.Regarding the characterization of the SAPO-34 acid function: an specific surface area BET of 652 m2g-1, micropore volume of 0.2192\u00a0cm3\u00a0g\u22121, and total pore volume of 0.23\u00a0cm2 g\u22121were determined (Fig. S1). In the NH3-TPD analysis (Fig. S2) 777.6 mmolNH3 gcat\n\u22121 were measured for SAPO-34, and from the profile two types of acid sites were identified, with peaks at 180\u00a0\u00b0C (14%) and 375\u00a0\u00b0C (86%), related to weak and strong acid sites, respectively.The performance of the In2O3\u2013ZrO2 catalysts has been studied in CO2 (Fig. 4) and CO2/CO mixtures hydrogenation (Fig. 5) under the reaction conditions (described in Section 2.3). It is observed in Fig. 4, that for CO2 hydrogenation, parent catalysts (thus, In2O3 and ZrO2) reach slightly higher COX conversion values than combined metal oxides, and even higher oxygenates selectivities. However, all CO2 conversion values are significantly enhanced in the In2O3\u2013ZrO2 catalysts. For the catalysts with a Zr/In ratio of 1/3 and 1/2 similar performance is observed, thus, more than doubling the value of XCO2 with respect to that obtained with In2O3 and upgrading that of ZrO2 over 50\u201360%. However, the selectivity of oxygenates decreases from 60% to 55% for In2O3\u2013ZrO2 catalysts, as a consequence of the increase in the formation of methane and, to a lesser extent, of olefins. It should be noted that the results in Fig. 6 evidence that the weakly acidic sites (Lewis sites) of ZrO2 in the In2O3\u2013ZrO2 catalysts (Dang et al., 2018) are sufficient to activate the dual cycle mechanism, with a reduced conversion of methanol into olefins, justifying the upturning olefins yield when increasing the Zr content. However, ZrO2 itself is not sufficient for the formation of olefins, because the presence of In2O3 is required for an efficient CO2 adsorption and H2 splitting as the first steps for methanol formation.Based on these results, a Zr/In ratio in the range between 1/3 and 1/2 is considered adequate to maximize both pursued targets in the hydrogenation of CO2, thus, CO2 conversion and oxygenate yield. This upgrade in CO2 conversion for the In2O3\u2013ZrO2 catalysts at 400\u00a0\u00b0C is consistent with the characterization results observed in the H2-TPR (Fig. 2a) and CO2-TPD of (Fig. 3) analyses. Both results explain the synergy in the conversion of CO2 by the improvement of the CO2 adsorption capacity by ZrO2 and H2 dissociation on In2O3 sites, due to the proximity of these sites. Frei et al. (2020) already observed this phenomenon in the usual conditions of methanol synthesis from CO2 (300\u00a0\u00b0C) with In2O3\u2013ZrO2 catalysts.Comparing the results in Figs. 4 and 5, a significant effect of the feed composition over the performance of the catalysts is evidenced. In both cases, the high CH4 selectivity is noteworthy, being higher for CO2/CO mixture hydrogenation (Fig. 5). For this feed, higher Zr/In ratio in the tandem catalysts leads to upturn CH4 selectivity, at the expense of olefins and oxygenates yields. This significant CH4 formation, is consequence of the fact that the endothermic methanation reaction is favored at such high reaction temperature of 400\u00a0\u00b0C required in the direct CO2 to olefins process. However, this reaction, which also occurs with methoxy ions as intermediates, as oxygenates formation does (Solis-Garcia et al., 2017), will be suppressed in the direct synthesis of olefins, by the in situ conversion of these ions into olefins, by means of the very fast dual cycle mechanism over the SAPO-34 catalyst (Cordero-Lanzac et al., 2020a) (subsequent section 3.4). The greater CO2 conversion attained with the catalyst with Zr/In ratio of 1/2 (Fig. 4) is interesting for its use in the direct synthesis of olefins by CO2 and CO2/CO hydrogenation.For further studying the effect that adding ZrO2 might have in the deactivation of In2O3\u2013ZrO2 catalysts, the evolution of CO2 conversion with time on stream is compared in Fig. 6 for CO2/CO mixture hydrogenation. It is noticeable that deactivation is only observed for pure In2O3. These trends evidence that ZrO2 addition improves the stability of the parent In2O3 catalyst. In accordance with the results in Figs. 4 and 5, 1Zr\u20132In and 1Zr\u20133In catalysts have also similar performance for the evolution of CO2 conversion with time on stream.This high stability of the catalyst in presence of ZrO2 is also evident in the evolution of products distribution with time on stream for the tested 24\u00a0h. As an example, in Fig. 7\n, the evolution of products yield with time on stream is depicted for the In2O3\u2013ZrO2 catalyst with Zr/In of 1/2 for the hydrogenation of a CO2/CO mixture with a CO2/COX ratio of 0.5. A high CO2 conversion with ZrO2 catalyst is also observed in Fig. 6. However, this result is a consequence of an undesired high CH4 formation (Fig. 5). In addition, Figs. 4\u20136 show that for a Zr/In ratio of 1/1 the conversion of CO2 is remarkably lower, which does not correspond to the CO2 adsorption capacity (Fig. 3). This result is explained because with this Zr/In ratio the amount of In2O3 is not enough for the dissociation of the H2 amount required for the methanol and CH4 formation reactions.All in all, considering CO2 conversion, oxygenates selectivity, catalyst stability and the influence of the CO2/CO composition in the feed, the composite catalyst with Zr/In ratio of 1/2 is considered to give the best balanced results, and so, the best prospects for conforming tandem In2O3\u2013ZrO2/SAPO-34 catalysts for the direct olefins synthesis.As it is sensed by comparing the results in Figs. 4 and 5, the composition of the CO2/COX mixture in the feed has a great effect on products distribution in the synthesis of methanol using In2O3\u2013ZrO2 catalysts. In Fig. 8\n, product yield values after 16\u00a0h TOS corresponding to the catalysts with Zr/In ratios of 1/2 (Figs. 8a) and 1/3 (Fig. 8b) are shown. It is observed that in both cases CH4 formation decays sharply when co-feeding CO2 together with syngas. It is also noteworthy that methanol yield passes through a maximum for feed compositions with equal concentration of CO and CO2. Likewise, the results in Fig. 8b contribute to further standing out the greater production of methane the 1Zr\u20133In catalyst leads to, being it especially relevant for pure syngas feeds (CO2/COX, 0). Probably due to the over-reduction of the catalyst, in good agreement with the highest reducibility under H2 and CO atmospheres reported in section 3.1.Accordingly, these results reaffirm the selection of 1Zr\u20132In as the In2O3\u2013ZrO2 catalyst with best prospects to be used in combination with SAPO-34 in the direct olefins formation from CO2 and syngas mixtures whatever CO2/CO composition.The suitability of the selected In2O3\u2013ZrO2 catalyst (1Zr\u20132In) in the direct synthesis of olefins process has been addressed in this section. Fig. 9\n illustrates the products yields obtained with the tandem In2O3\u2013ZrO2/SAPO-34 catalyst for feeds with different composition. The results evidence the good performance of the tandem catalyst, given almost all the oxygenated compounds formed as intermediates are converted into hydrocarbons, the selectively to olefins, being paraffins the only by-products, and the absence of CH4.Comparing the results of the direct synthesis of olefins (Fig. 9) with those in Fig. 8 of the first stage of the process (oxygenates formation), various features are to be highlighted: i) The upgrade of the overall conversion obtained in the direct synthesis. Thus, COX conversion increases from 0.88% to 4.28% for feeds with CO2/COX ratio of 0.5, and the conversion of the targeted products from 0.51% (oxygenates yield in methanol formation) to 3.11% (olefins yield). ii) The suppression of the undesired methane formation pathway, being it almost undetectable.The mechanism of hydrocarbon formation directly by hydrogenation of CO2 and CO on the In2O3\u2013ZrO2/SAPO-34 tandem catalyst is the cascade combination of the mechanisms of methanol/DME synthesis and the in situ conversion of these oxygenates into hydrocarbons. It is well established in the literature (Frei et al., 2018; Tan et al., 2019; Ye et al., 2012, 2013) that in methanol synthesis over In2O3\u2013ZrO2 catalysts, CO2 is adsorbed on In2O3 oxygen vacancies and on additional oxygen vacancies formed by the presence of ZrO2 and the formation of stable In1-xZrOxOy mixed oxide. The H2 dissociation capacity of the In2O3 in In2O3\u2013ZrO2 facilitates the hydrogenation of adsorbed CO2 to form formate species (HCOO*). The next steps consists of the reaction of these species with H* ions to produce H2COO* species, and the hydrogenation of the latter to methoxy species (H3CO*), which will hydrogenate to form methanol. The presence of these intermediates has been determined by means of experimental and theoretical studies (Dang et al., 2018; Wang et al., 2021). The effect of the presence of CO in this reaction mechanism is controversial. Besides disfavoring the rWGS reaction, it is well established (Martin et al., 2016) that a moderate concentration of CO increases the density of oxygen vacancies, favoring therefore CO2 adsorption and the extent of methanol formation. However, due to its strong reducing character (verified in Fig. 2b), a high concentration of CO can favor the over-reduction of In2O3 and its sintering (Ara\u00fajo et al., 2021b). The results in Fig. 9 are consistent with the commented effect of CO concentration, giving rise to higher olefins yield for the CO2/COx ratio of 0.5 in the feed than for the hydrogenation of CO (CO2/COx of 0), as a consequence of the higher methanol yield.The formation of the C\u2013C bonds of the light olefins from methanol/DME is a consequence of the activity of the acid sites of SAPO-34 (Gayubo et al., 2000; P\u00e9rez-Uriarte et al., 2016). The reaction proceeds through the dual cycle mechanism, with two related routes, with polyalkyl benzenes and olefins as intermediates (Gao et al., 2019). The severe shape selectivity of SAPO-34, with CHA topology (cavities of 10\u00a0\u00d7\u00a06.7\u00a0\u00c5 connected by 3.8\u00a0\u00d7\u00a03.8\u00a0\u00c5 8-ring cages) (Hemelsoet et al., 2013), is suitable for the selective formation of ethylene and propylene when used in the tandem In2O3\u2013ZrO2/SAPO-34 catalyst (Dang et al., 2019; Portillo et al., 2021).The evolution of products yield with time on stream obtained in the direct synthesis of olefins is shown in Fig. 10. These results, corresponding to the most severe deactivation conditions studied, that is, for syngas feeds, evidences that after an initial activity decay taking place in the first 4\u00a0h of reaction, an pseudo-steady state is achieved (in this case and for all the studied feed compositions). From characterization analyses through temperature programmed oxidation (TPO) with air (results not shown), this deactivation is attributed to coke deposition in the acid function (coke content of 4.9\u00a0wt% on the acid catalyst and 0.5\u00a0wt% on the metallic catalyst of the tandem catalyst, for the reaction conditions in Fig. 10). Coke formation takes place fast in the first reaction hours over SAPO-34, reaching almost the maximum reported value in 2\u00a0h TOS (4.6\u00a0wt% and 4.9\u00a0wt% at pseudo-stable conditions). Once at that point, coke formation rate is residual, suppressed by the hydrogenation of the intermediates, and does not lead to further activity decay.The addition of Zr to In2O3 catalysts improves the performance (activity, oxygenates selectivity and stability) for the hydrogenation of CO2/CO mixtures to methanol under the suitable operation conditions for the direct CO2 to olefins process (400\u00a0\u00b0C, 30\u00a0bar). This is a key feature for the configuration of tandem catalysts for the direct conversion of CO2 (and syngas) into olefins via the route with oxygenates as intermediates. The behavior of the In2O3\u2013ZrO2 catalysts is a consequence of its properties. The loading of ZrO2 with a Zr/In ratio between 1/3 and 1/2 increases the yield of oxygenates compared to that obtained with parent In2O3 and ZrO2 catalysts, improves stability, and is suitable for attaining an outstanding conversion co-feeding CO (syngas) together with CO2, albeit with high formation of CH4 (favored with increasing Zr/In ratio).Accordingly, the In2O3\u2013ZrO2 catalyst with Zr/In of 1/2 has been selected as suitable for its use in the In2O3\u2013ZrO2/SAPO-34 tandem catalyst for the direct synthesis of olefins. The results obtained with this catalyst offer a good balance between CO2 and COX conversion, olefins yield and selectivity, and catalyst stability at 400\u00a0\u00b0C and 30\u00a0bar, for different CO2/COX composition feeds. The in situ conversion of the formed oxygenates into olefins, displaces the thermodynamic equilibrium of the methanol formation reactions, and as a result, olefins yield 6 folds compared to the oxygenates yield of the first stage, and methane formation is negligible.The results (obtained at low space time values, to work in demanding conditions for the stability of the tandem catalyst) are encouraging to progress towards the optimization of the operating conditions for the joint valorization of CO2 and syngas, which is an interesting strategy to mitigate climate change.\nA. Portillo: Conceptualization, Methodology, Investigation, Validation, Data curation, Writing \u2013 original draft, Writing \u2013 review & editing. A. Ateka: Conceptualization, Methodology, Investigation, Validation, Data curation, Writing \u2013 original draft, Writing \u2013 review & editing. J. Ere\u00f1a: Data curation, Writing \u2013 original draft. J. Bilbao: Conceptualization, Methodology, Investigation, Data curation, Writing \u2013 original draft, Writing \u2013 review & editing, Supervision. A.T. Aguayo: Project administration, Conceptualization, Supervision, Validation, Data curation.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work has been carried out with the financial support of the Ministry of Science, Innovation and Universities of the Spanish Government (PID2019-108448RB-100); the Basque Government (Project IT1645-22), the European Regional Development Funds (ERDF) and the European Commission (HORIZON H2020-MSCA RISE-2018. Contract No. 823745). A. Portillo is grateful for the grateful for the Ph.D. grant from the Ministry of Science, Innovation and Universities of the Spanish Government (BES2017-081135). The authors thank for technical and human support provided by SGIker (UPV/EHU).The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.jenvman.2022.115329.", "descript": "\n                  The effect of the ZrO2 content on the performance (activity, selectivity, stability) of In2O3\u2013ZrO2 catalyst has been studied on the hydrogenation of CO2/CO mixtures. This effect is a key feature for the viability of using In2O3\u2013ZrO2/SAPO-34 tandem catalysts for the direct conversion of CO2 and syngas into olefins via oxygenates as intermediates. The interest of co-feeding syngas together with CO2 resides in jointly valorizing syngas derived from biomass or wastes (via gasification) and supplying the required H2. The experiments of methanol synthesis and direct synthesis of olefins, with In2O3\u2013ZrO2 and In2O3\u2013ZrO2/SAPO-34 catalysts, respectively, have been carried out under the appropriate conditions for the direct olefins synthesis (400\u00a0\u00b0C, 30\u00a0bar, H2/COX ratio\u00a0=\u00a03) in an isothermal fixed bed reactor at low space time values (kinetic conditions) to evaluate the behavior and deactivation of the catalysts.\n                  The Zr/In ratio of 1/2 favors the conversion of CO2 and COX, attaining good oxygenates selectivity, and prevents the sintering attributable to the over-reduction of the In2O3 (more significant for syngas feeds). The improvement is more remarkable in the direct olefins synthesis, where the thermodynamic equilibrium of methanol formation is displaced, and methanation suppressed (in a greater extent for feeds with high CO content). With the In2O3\u2013ZrO2/SAPO-34 tandem catalysts, the conversion of COx almost 5 folds respect oxygenates synthesis with In2O3\u2013ZrO2 catalyst, meaning the yield of the target products boosts from \u223c0.5% of oxygenates to >3% of olefins (selectivity >70%) for mixtures of CO2/COX of 0.5, where an optimum performance has been obtained.\n               "}
{"full_text": "The huge emission of carbon dioxide (CO2) due to the excessive consumption of fossils has caused severe problems, so the CO2 issue has attracted much attention all over the world [1\u20133]. As a renewable and environmentally friendly C1 resource, the capture and utilization of CO2 has attracted tremendous interest because it is not only an efficient way to alleviate greenhouse effect but also can provide green routes to synthesize chemicals [4\u20137]. By constructing C-X (X\u202f=\u202fN, O, C, H) bonds or direct CO2 hydrogenation [8], CO2 has been applied in the synthesis of urea [9,10], carbonate derivatives [11\u201313], carbonyl derivatives [14,15] or fuels [16\u201318]. However, CO2 is a thermal stable and dynamic sluggish apolar molecule, which makes its utilization be challenging, and its transformation is generally more difficult than other C1 resources (e.g., CO) [19]. In the past decades, the chemical transformation of CO2 has been widely investigated via thermo-, or electro-, or photocatalytic conversion, with focus on exploring new reaction routes and developing catalysts with high efficiency [20\u201323].As an emerging material platform, porous organic polymers (POPs), featuring in tailorable functionalization, large surface areas, adjustable porosity, versatile polymerizations, good physicochemical and thermal stability, have attracted considerable scientific interest [24\u201327] and showed promising applications in adsorption and separation [28\u201330], heterogeneous catalysis [31\u201333], energy storage [31,34] and so on. Especially, the task-specific design of POPs via selecting functional monomers, polymerization protocols, and modification strategies endows them promising candidates for CO2 capture and transformation, and various efficient POPs-derived catalysts have been developed [35\u201337]. In these reported POPs, CO2-philic heteroatoms (e.g., N, F, O, P) [38\u201340] and/or ionic sites [41\u201343] have been incorporated into the skeletons of POPs, which can achieve efficient adsorption and activation of CO2. Moreover, metal species as active catalytic components (e.g., Zn, Cu, Ru, Ag ions or nanoparticles) [44\u201346] can be immobilized onto POPs, achieving metalated POPs catalysts for CO2 transformation. In addition, POPs can be designed with conjugated structures to serve as photocatalysts for CO2 photoreduction to value-added chemicals [47,48].Our group has paid much attention on design and preparation of task-specific POPs for CO2 capture and utilization via introducing CO2-philic groups into the skeleton, constructing conjugated structures, modification with metal active species (i.e., metalated POPs) and so on, and a series of POPs with functional groups, such as, azo, Tr\u00f6ger's base, fluorine, phenolic \u2013OH, N-containing sites, have been prepared with high catalytic performance. In this article, we will introduce our recent work on synthesis of POPs catalysts for CO2 transformation, which include POPs-based catalysts for cycloaddition reactions of epoxides and propargylic alcohols with CO2, reductive conversion of CO2 with H2, photocatalytic/electrocatalytic conversion of CO2. Finally, the challenges on POP-based catalysts for CO2 capture and conversion are discussed.Cycloaddition reactions of epoxides or propargylic alcohols with CO2 are economically favorable and environment-friendly routes to access cyclic carbonates that are widely applied as raw substrates for polycarbonates and polyurethanes, aprotic polar solvents, electrolytes of batteries and so on [49\u201351]. Developing catalysts with high efficiency, good stability and low cost is of significance for this kind of reactions. Besides the ability to activate CO2, the catalysts are also required to be capable of activating epoxides or propargylic alcohols [52\u201354]. In our work, we designed and synthesized POPs-based metal-free and metalated catalysts, which efficiently realize the transformation of CO2 under mild conditions.For the cycloaddition of epoxides with CO2, various catalysts that are capable of activating the epoxides and CO2 simultaneously have been reported, including homogeneous [55] and heterogeneous catalysts [56], metal-based [57,58] and metal-free catalysts [59]. Among the reported catalysts, POPs-based catalysts with or without metal components have been widely investigated, and have shown promising application potentials due to their unique performances [60\u201362]. In our work, we designed metal-free and metalated POPs catalysts for this reaction, and achieved the reaction under mild conditions.Lewis acid metal sites (e.g., Zn2+, Co2+, Ni2+) have been proved to be efficient for the cycloaddition of epoxides with CO2 due to their coordination ability with O atom of epoxides, and various metalated POPs with coordination unites (e.g., porphyrin, azo, salen, carbene, pyridine) have been synthesized over the past decades [63\u201365]. For example, Deng's group [66,67] reported a series of metalated salen-based POPs (M-CMPs, M\u202f=\u202fZn, Co, Al) for this kind of reactions, which were synthesized by polymerization of metal-salen or post metalation. M-CMPs showed good CO2 adsorption capacity, and could achieve cycloaddition of epoxides with CO2 under ambient conditions. Especially, the resultant Zn-salen-POPs showed excellent catalytic performance for the cycloaddition of propylene oxide with CO2, and achieved an ultrahigh turnover frequency (TOF) up to 11,600\u202fh\u22121\u202fat 120\u202f\u00b0C and 3\u202fMPa with good stability [67]. Yang and coworkers [68] synthesized zinc porphyrin-based frameworks (P-POF-Zn) using hydroxyl-containing monomers (e.g., 1,3,5-tris(30-tert-butyl-40-hydroxy-50-formylphenyl)benzene) by two-step metallization, and the as-prepared Zn-P-POF displayed high catalytic performance for the cycloaddition of epoxides. It was demonstrated that the presence of massive \u2013OH groups was beneficial for CO2 adsorption and activation. Though the above reported POPs-based catalysts showed high performance for cycloaddition of epoxides with CO2, the high cost of the monomers and complicated synthetic procedures limit their applications.We developed a simple strategy to synthesize o-hydroxyazobenzene POPs (HAzo-POPs) based on the diazo-coupling reaction of aryl tri/diamines with tri/diphenols in aqueous solution without any template or metal catalysts under ambient condition, as illustrated in Fig.\u00a01\n [69]. In contrast, the feedstocks used in this protocol are cheap and easily available, and the product yield reached up to 90%. The resultant HAzo-POPs displayed mesoporous structures with BET surface area up to \u223c 600\u202fm2\u202fg\u22121. They possess massive phenolic hydroxyl and azo groups in their skeletons, which are favorable to adsorption of CO2 and to coordination of metal ions. HAzo-POPs exhibited high metal adsorption capacity (e.g., 26.24\u202fwt% for Cu2+ by HAzo-POPs-1) and excellent CO2 adsorption capacity (e.g., 7.5\u202fwt% at 273\u202fK and 1.0\u202fbar) and high separation selectivity of CO2/N2 (106). Especially, Zn2+ complexed HAzo-POP-1 (Zn/HAzo-POP-1) exhibited excellent performances for cycloaddition of propylene oxide (PO) and CO2 in the presence of tetrabutylammonium bromide (TBAB), which could realize this reaction under ambient conditions, and showed 10 times higher activity than Zn-CMPs that was reported to be the state of art of the POP catalysts at that time. The good CO2 adsorption ability, mesoporous architecture, and excellent dispersity of metal sites are believed to be responsible for their high performance for catalyzing the cycloaddition of PO with CO2.Subsequently, the aqueous azo-coupling reactions were applied in the synthesis of porous polymers for CO2 capture [70], wastewater treatment [71], sensors [72], organic battery [73] and so on. For example, Chen and coworkers [70] prepared metalloporphyrin-based o-hydroxy azo-hierarchical POPs (ZnTPP/QA-azo-PiPs) by azo-coupling of tetra(4-aminophenyl) porphyrin zinc-derived diazonium salts with multihydroxy benzene in aqueous solution. The resultant ZnTPP/QA-azo-PiP1 not only showed excellent catalytic performance for the cycloaddition of various epoxides under mild conditions, but also it could catalyze the CO2-involved synthesis of oxazolidinones and N-formylated amines under diluted CO2 (15% CO2 in 85% N2, v/v).From the viewpoint of green chemistry, using renewable feedstocks and adopting simple synthetic procedures to synthesize porous polymers with low-cost are valuable. Utilizing renewable feedstocks including chitosan and phytic acid, we prepared a kind of porous metal\u2013organic hybrids (MOH) based on the complexation of metal ions (e.g., Zn2+, Cu2+, Ni2+) with the interaction sites in the organic feedstocks, as illustrated in Fig.\u00a02\n [74]. As an example, the resultant MOH-Zn exhibited mesoporous structures with a BET surface area of 90\u202fm2\u202fg\u22121 and a pore volume of 0.40\u202fcm3\u202fg\u22121. Due to the presence of plenty of \u2013OH, \u2013NH2, and \u2013PO4 groups, MOH-Zn displays excellent activity for CO2 activation, which can make cycloaddition of epoxides with CO2 proceed in the presence of TBAB under ambient conditions, affording a high turnover frequency of 7.8 h\u22121. This work provides a facile protocol for the synthesis of low-cost POP-based catalysts from renewable biomass feedstocks, which may have practical application potential. In a recent work, Zn-modified N-doped carbon (Zn/NC-X) was prepared via carbonization of the chitosan-derived MOHs-Zn, and the obtained Zn/NC-950 exhibited excellent performance for the aerobic oxidative cleavage of C(CO)\u2013C bonds in acetophenone derivates at 100\u202f\u00b0C [75].In most cases, TBAB is required as a co-catalyst for the cycloaddition of epoxides with CO2, which is unfavorable to the purification of product and increases the practical cost. To avoid the use of TBAB, Ding's group [76] developed a Zn2+ and Br\u2212 co-modified porous imidazolium ionic polymer containing P coordination sites for the cycloaddition of PO with CO2, achieving a high TOF of 6022 h\u22121 owing to the high BET surface areas, excellent CO2 uptake and synergistic effect of Zn2+ and Br\u2212.Though great progress has been achieved in the synthesis of efficient metalated POPs for the cycloaddition of epoxides with CO2, the leakage or deactivation of active metal sites is still a big challenge. Therefore, developing metal-free POPs based catalyst with excellent catalytic performance and stability is interesting.Fluorine (F) element has the highest electronegativity and a small radius of 71 pm, which endow it with many special properties and coordination ability [39]. F-containing materials have CO2-philic property, and fluorinated POPs have been intensively studied, which indicated that F modification in POPs can obviously enhance the CO2 adsorption capacity [77\u201381]. For example, Han's group [82] reported that fluorinated-covalent triazine frameworks (CTFs) afforded a much higher CO2 adsorption capacity (5.53\u202fmmol\u202fg\u22121) than that of non-fluorinated-CTFs (3.82\u202fmmol\u202fg\u22121) due to the presence of massive C\u2013F bonds. Besides of F modification, integrating nucleophilic ions (Br\u2212, Cl\u2212) into the skeleton of POPs has been proved to be an effective way to activate the epoxides [83]. For instance, quaternary phosphonium-functionalized POPs prepared via the polycondensation of tetrakis(4-chlorophenyl)phosphonium bromide showed intrinsic catalytic activity for the cycloaddition of epoxides with CO2 under 1\u202fatm [84].Our group synthesized a series of fluorine and nucleophilic ions (Br\u2212 or Cl\u2212) co-modified imidazolium based polymeric ionic liquids (denoted as PILs-X, X\u202f=\u202fCl, Br) by an one-pot ionic polymerization based on the coupling reactions as illustrated in (Fig.\u00a03\n) [85]. It was indicated that all the resultant PILs-X samples were effective for catalyzing cycloaddition of styrene oxide with CO2, and the F content in the catalysts showed positive correlation with their catalytic performance, following the order: PIL-Br\u202f<\u202fF0.5-PIL-Br\u202f<\u202fF-PIL-Br. Notably, F-PIL-Br showed three times higher efficiency for the cycloaddition of styrene oxide than that of the polymers without F modification (PIL-Br) under 1\u202fMPa and 120\u202f\u00b0C, together with broad scope of the reactants, excellent product yields (93%\u201399%), high stability and easy recyclability. In addition, F-PIL-Br showed much higher activity than F-PIL-Cl, which should be caused by the superior leaving ability and nucleophilicity of Br\u2212 over Cl\u2212.F and ion co-modified POP catalysts have been reported to be efficient for other CO2 involved reactions as well. For example, Yan's group [86] synthesized frustrated Lewis pair (FLP) functionalized polymeric ionic liquids (PILs) using 4-styryl-di(pentafluorophenyl)borane. Because of the synergistic effect between two complementary Lewis acidic and basic sites, the as-prepared FLP-polymer exhibited ultrafast response for CO2 adsorption at about 20\u202fs, and it also showed high catalytic performance for N-formylation of amines with CO2 in the presence of PhSiH3, affording a high TON of 14,800\u202fat room temperature.The carboxylative cyclization of propargylic alcohols with CO2 is a green approach to synthesize \u03b1-alkylidene cyclic carbonates that have potential bioactivity with a broad range of applications as intermediates in organic synthesis. Transition metal catalysts (e.g., Ag, Cu, Co) have been reported to be active for this kind of reactions [49,87,88]. To achieve the cycloaddition under mild conditions, we developed several Ag-metalated POPs for this reaction. As illustrated in Fig.\u00a04\n, we integrated the fluorine-containing component and phenanthroline ligand into the backbone of POP via direct Sonogashira-Hagihara cross-coupling of tetrakis(4-ethynylphenyl)methane with perfluorinated aromatic bromides and bromo-substituted phenanthroline, and prepared fluorinated POP with phenanthroline sites (F-MOP) [89]. F-MOPs showed microporous structures with high BET surface area and much higher CO2 adsorption capacity (223\u202fmg\u202fg\u22121) than nonfluorous MOPs (98\u202fmg\u202fg\u22121). Due to the presence of phenanthroline sites, F-MOP could be metalated with Ag(I) to form Ag-metalated F-POP (denoted as F-MOP-Ag). The resultant F-MOP-Ag was capable of catalyzing the cyclization of propargylic alcohol with CO2 at room temperature, and showed much higher catalytic activity than that of MOP-Ag without F modification.Lately, we prepared rose bengal (RB)-functionalized POP (RB-POP) via direct Sonogashira-Hagihara cross-coupling of RB with 1,4-diethynylbenzene (Fig.\u00a05\na) [90]. The resultant RB-POP had a high BET surface area of 562 m2\u202fg\u22121, and could adsorb CO2 with capacity of 72\u202fmg\u202fg\u22121 at 1\u202fbar at 273\u202fK. RB-POP could immobilize Ag nanoparticles with uniform distribution and size around 3.0\u202fnm (Ag@RB-POPs). Importantly, Ag@RB-POPs exhibited extraordinary activity for the carboxylative cyclization of propargyl alcohols with CO2 under mild conditions, achieving a high TOF of 5000 h\u22121, with a wide substrate scope, high stability, and easy recyclability.Recently, He's group [87] developed a Ag(0) nanoparticles modified reduced graphene oxide (Ag-rGO) with massive O sites to serve as catalyst, which can not only catalyze the cyclization of propargylic alcohols with atmospheric CO2 with high efficiency, but also can catalyze synthesis of other value-added chemicals (\u03b2-oxopropylcarbamates and 2-oxazolidinones) from CO2 under ambient conditions.The reductive transformation of CO2 with H2 to chemicals is an important way for CO2 utilization, and can also provide green routes for chemical synthesis [91]. This kind of reactions generally require catalysts that can simultaneously activate CO2 and H2, and transition metal (e.g., Rh, Ru, Pd, Ir, Ni) decorated POPs catalysts can be designed to meet this requirement [92\u201395]. In our work, we integrated the CO2-philic group and ligands that can coordinate with metal species into the skeletons, and designed various POPs-based metal catalysts, which have realized reductive transformation of CO2 with H2 including CO2 hydrogenation to formic acid and N-formylation of amines with CO2/H2. Different from organic molecule ligands with unique structures, the polymers that can coordinate with metal species are only required to have coordination sites, which provide more opportunities to develop new catalysts. To get high-efficiency catalysts for reductive transformation of CO2 with H2, we integrated N-containing sites into the polymer skeletons and prepared various metalated POPs catalysts via post metallization strategy.Tr\u00f6ger's base (TB)-type ligand is one of the most versatile ligands applied in coordination chemistry. We introduced the TB unit into the skeletons of polymers via the reaction of a planar rigid building block (tris(4-aminophenyl)amine) with dimethoxymethane in trifluoroacetic acid at room temperature (Fig.\u00a06\n), and TB-based POP was obtained [96]. The resultant TB-MOP showed microporous structure with a BET surface area up to 802\u202fm2\u202fg\u22121, which could adsorb CO2 with an uptake of 169\u202fmg\u202fg\u22121 at 1\u202fbar and 273\u202fK. As expected, TB-MOP could coordinate with Ru(III) complex via the interaction between the TB ligand and the metal ions. The resultant material (TB-MOP-Ru) showed declined CO2 adsorption capacity (127\u202fmg\u202fg\u22121 at 273\u202fK) compared to TB-MOP, but it could efficiently adsorb H2 with a capacity of 9.5\u202fmg\u202fg\u22121 at 77\u202fK. As a result, serving as an efficient catalyst TB-MOP-Ru realized the CO2 hydrogenation to formic acid in Et3N with a TON of 2254\u202fat 40\u202f\u00b0C.Han and coworkers [46] reported a Ru-metalated N,P-co-functionalized POP for CO2 hydrogenation, which afforded an ultrahigh TON of 25,400. It was discovered that the electron-rich Ru3+ accelerated the H2 dissociation and N-functionalized architecture enhanced the CO2 adsorption. Yoon et\u00a0al. [97] reported Ru modified dipyridyl functionalized POP for CO2 hydrogenation in a fixed bed reactor, which showed substantial catalytic performance with the highest productivity of 669.0 gform. gcat.\n\u22121\u202fd\u22121. To avoid the purification process from formate to pure formic acid, they further developed a Ru metalated pyridine and imidazole co-functionalized POPs, affording a high TON of 2197 for the generation of methyl formate with methanol as solvent in the presence of Et3N [45].The imine-type ligands in the polymer skeleton have good ability to complex with metal species. We presented a novel approach to prepare imine-based POP (Imine-POP) via the reaction of aryl ammonium salt with aromatic aldehyde in water without any catalyst or template, which could metalated with Pd2+ by mixing with Pd(OAc)2 in CH2Cl2 and ethanol solution at room temperature (Fig.\u00a07\na) [98]. It was indicated that Imine-POP possessed micro- and mesoporous structures (Fig.\u00a07b), dominated with mesopores with BET surface area around 200 m2\u202fg\u22121. Since the imine bond can chelate with metal species, the Pd nanoparticles with mean size of 2.84\u202fnm were uniformly distributed onto the surface of Imine-POP (Fig.\u00a07c). The resultant mesoporous Imine-POP@Pd was applied in catalyzing formylation of amines with CO2/H2 at 100\u202f\u00b0C, which was tolerant to primary and secondary aliphatic amines, affording a series of formamides in moderate to excellent yields (48%\u201397%). Mechanism investigation indicated that the N-formylation of amines underwent the Imine-POP@Pd-catalyzed CO2 hydrogenation to HCOOH and the subsequent reaction of HCOOH with amine.In another work for formylation of amines with CO2/H2, we prepared pyridine-functionalized POPs (CarPy-CMP) via radical oxidative coupling polymerization of 2,6-di(9H-carbazol-9-yl)pyridine catalyzed by FeCl3 in chloroform, which was decorated with Ru nanoparticles to form a Ru catalyst (CarPy-CMP@Ru) [99]. The resultant CarPy-CMP exhibited micro- and mesoporous structures with BET surface area \u223c1000\u202fm2\u202fg\u22121, and it showed excellent CO2 uptake capacity (up to 63 and 171\u202fmg\u202fg\u22121 at 0.1\u202fbar and 1\u202fbar at 273\u202fK). CarPy-CMP@Ru showed high catalytic activity for the reaction of secondary amines with CO2/H2, affording a series of formyamides in high yields (89%\u201393%), together with high stability and easy recyclability.The azo-type ligands within the polymer skeletons also show good coordinating ability for metal species. We presented a simple and efficient method for the synthesis of azo connected POPs through oxidative polymerization of aromatic multi-amines catalyzed by \nt\nBuOCl/NaI at room temperature (e.g., 25\u202f\u00b0C, 1\u202fh), and 4 samples (Azo-MOP-N, N\u202f=\u202f1\u223c4) were prepared using the corresponding amine monomers (including tetrakis(4-aminophenyl)methane, A-1; 2,6,14-triaminotriptycene, A-2; 1,3,5-tris(4-aminophenyl)benzene, A-3; and tris(4-aminophenyl)amine, A-4) as illustrated in Fig.\u00a08\n [100]. The resultant Azo-MOPs displayed high thermal stability up to 400\u202f\u00b0C, high BET surface areas up to 706\u202fm2\u202fg\u22121 and high adsorption capacity to CO2 of 134.8\u202fmg\u202fg\u22121 (273\u202fK, 1\u202fbar). Furthermore, the Azo-MOPs could coordinate with Ru(III) complex to form Azo-MOP-N-Ru with homogeneous ruthenium distribution without detectable ruthenium clusters or nanoparticles, and the Ru content in each sample was almost identical around 4.70\u202fwt%. All the Azo-MOP-N-Ru samples could adsorb CO2 and catalyze the methylation of amines with CO2 and phenylsilane with high efficiency under low pressure (0.5\u202fMPa). Especially, Azo-MOP-3-Ru showed the best performance, together with broad scope of amines (including N-methylanilines with both electron-donating and electron-withdrawing groups, dialkylamines), excellent product yields (93%\u201399%), high stability and easy recyclability.Triphenylphosphine (PPh3) is an organic ligand widely applied in organic chemistry, and the PPh3-based POPs can combine the advantages of PPh3 and POP together to generate new functions. In our work, we integrated PPh3 unit and CO2-philic group (e.g., azo group) into the skeletons of polymers, and prepared poly(PPh3) connected with azo bonds (poly(PPh3)-azo) via oxidative polymerization of phosphine-containing aromatic amines, P(m-NH2Ph)3, in the presence of tBuOCl/NaI at 25\u202f\u00b0C (Fig.\u00a09\na) [101]. The resultant poly(PPh3)-azo showed a BET surface area and of 118\u202fm2\u202fg\u22121 dominated with mesopores (Fig.\u00a09b). Treating poly(PPh3)-azo with AgBF4 in tetrahydrofuran under refluxing conditions resulted in the decoration of small amount (0.17\u202fwt%) of Ag nanoparticles with size around 3.0\u202fnm onto the surface of the polymer (Fig.\u00a09c, poly(PPh3)-azo-Ag). Similarly, treating the polymer with RuCl3 in ethanol realized the complexation of Ru(III) with the ligand sites in the polymer as illustrated in Fig.\u00a09d, and no Ru particles were observed in the resultant poly(PPh3)-azo-Ru. Moreover, the Ru content reached a higher value of 3.72\u202fwt%. The resultant poly(PPh3)-azo-Ag could catalyze the carboxylative cyclization of propargylic alcohols with CO2 at room temperature, affording more than 400 times higher site-time-yield (STY) compared with the best heterogeneous catalytic system reported. Poly(PPh3)-azo-Ru exhibited extraordinary activity for the methylation of amines with CO2 under low pressure. The high performances of the catalysts were originated from the cooperative effects between the polymer and the metal species. Moreover, both poly(PPh3)-azo-Ag and poly(PPh3)-azo-Ru showed good stability and easy recyclability, thus demonstrating great potential for practical utilization in catalysis.Recently, Dai's group [38] prepared a phosphabenzene functionalized POPs (Phos-POPs) by substitution of pyrylium-based POPs (Py-POPs) with P(Me3Si)3. After partially fluorinated and subsequently metalated with ruthenium, the resultant Ru/F-Phos-POP-2 exhibited excellent catalytic performance for the N-formylation of amines at 100\u202f\u00b0C, affording an ultrahigh TOF of 204 h\u22121.Because of the high cost of the POPs metalated with noble metals, developing efficient POPs modified with earth-rich metals is more interesting. We developed zinc-metalated and fluoro-functionalized porous N-heterocyclic carbene polymer (F\u2013PNHC\u2013Zn) via ionic polymerization and post metalation [102]. F\u2013PNHC\u2013Zn showed excellent catalytic performance for both formylation and methylation of various N-methylanilines with both electron-withdrawing or electron-donating groups. Remarkably, even under low CO2 pressure (0.05\u202fMPa), F\u2013PNHC\u2013Zn can catalyze the reaction efficiently.Photocatalytic reduction of CO2 (especially with H2O) to value-added chemicals is a promising and ideal way for CO2 transformation, and has attracted tremendous attention, considering its double benefits for CO2 utilization and conversion of solar energy [103\u2013107]. Generally speaking, the photoreduction of CO2 over photocatalysts undergoes three steps: (i) light absorption by the catalyst to generate photoexcited electron\u2013hole pairs; (ii) charge carrier separation and transfer to active sites; (iii) surficial reduction of the absorbed CO2 to chemicals (Fig.\u00a010\n). The photocatalyst is the key to realize the CO2 photoreduction, thus extensive efforts have been dedicated to developing efficient catalysts. To date, various photocatalysts have been reported for photoreduction of CO2, including inorganic semiconductors [22,108] (e.g., TiO2 [109,110], CdS [111,112], perovskite [113\u2013115], g-C3N4 [116\u2013119]), metal/organic hybrids [120] (e.g., heterojunctions [121\u2013123], MOFs [124,125]), and conjugated POPs [126\u2013128]. In particular, conjugated POPs showed promising applications in CO2 photoreduction because of their good CO2 adsorption ability, high surface areas, and tailorable functionalization [47,48].Up to now, great progress has been achieved in CO2 photoreduction over metalated POPs catalysts (e.g., Re [129], Zn [128], Co [130,131]) though they suffer from metal leakage, poor product selectivity, and high cost. Metal-free conjugated POPs have attracted attention due to their excellent CO2 adsorption capacity and tailorable structures [132\u2013134].Integrating CO2-philic elements N, O, and P into the skeleton of polymer, we prepared a N,O,P-containing COP (NOP\u2013COP) via condensation of hexachlorocyclo-triphosphazene with barbituric acid (BA) as illustrated in Fig.\u00a011\n [134]. This sample can effectively capture CO2 with a capacity up to 7.21\u202fwt% under ambient conditions, and exhibits appropriate energy band structure (E\ng\u202f=\u202f2.14\u202feV, CB\u202f=\u202f\u22120.81\u202feV). It was indicated that the incorporation of phosphorus in the skeleton of NOP\u2013COP promoted the visible light absorption, improved the separation efficiency for photoinduced electron-hole pairs, increased the lifetime of photoexcited charge carriers and reinforced the redox ability, compared to those of N,O-containing COP (NO\u2013COP) prepared via condensation of cyanuric chloride with BA. As a result, NOP\u2013COP showed high activity for photoreduction of CO2 in the presence of sacrifice agent TEOA, achieving a selectivity of 90.2% towards CH4 as the sole carbonaceous product with a rate of 22.5\u202f\u03bcmol gcat.\n\u22121\u202fh\u22121 under visible light irradiation. This is the first example of a COP metal-free photocatalyst for CO2 reduction to CH4 under visible light irradiation. However, sacrifice agent TEOA was required.To achieve photocatalytic CO2 reduction with H2O as an proton donor, specific POPs catalysts are required, especially in solid-gas system [135]. We contributed to the synthesis of metal-free POP photocatalysts for photoreduction of CO2 with H2O. Eosin Y is a dye, which can absorb visible light. We introduced Eosin Y unit into the skeleton of polymer, and designed Eosin Y-functionalized COPs (PEosinY-N, N\u202f=\u202f1\u223c3) via direct Sonogashira-Hagihara cross-coupling of Eosin Y with aromatic alkynes (Fig.\u00a012\na, A-1, A-2 and A-3) [136]. The resultant materials possessed porous structures with BET surface areas up to 610 m2\u202fg\u22121, and could adsorb and activate CO2 and H2O simultaneously, confirmed by in situ diffuse reflectance infrared Fourier transform spectroscopy analysis and DFT calculations. PEosinY-N could absorb visible light with suitable band structures as illustrated in Fig.\u00a012c, which make them be capable of catalyzing the photoreduction of CO2 to CO and the H2O oxidation to O2. As expected, CO was obtained as the sole carbonaceous product under the visible-light-drawn catalysis over PEosinY-N, and PEosinY-1 showed the best performance, affording a CO production rate of 33\u202f\u03bcmol\u202fg\u22121\u202fh\u22121 and a selectivity of 92%. However, no O2 was detected in the gas products, while H2O2 was instead detected in the aqueous solution. We found that the generation of H2O2 resulted from the photoreduction of O2 originated from the H2O photooxidation. For this kind of POP photocatalysts, the Eosin Y units and conjugated structure of PEosinY-N were found to be responsible for the light absorption as well as efficient electron/hole separation, which cooperatively realized the photoreduction of CO2 with H2O, generating CO and H2O2.We synthesized a series of pyrene-based POPs using the nickel-catalyzed Yamamoto protocol as shown in Fig.\u00a013\n [137]. To capture CO2 from air and realize its photoreduction, we constructed an ionic liquid (IL)-assisted catalytic system using pyrene-based POPs as the photocatalysts. The task-specific IL ([P4444][p-2-O]) can chemically capture CO2 and absorb H2O via hydrogen bonding from air. In combination with the IL, the pyrene-based CP realized the capture of CO2 from air and further photoreduction to CO under visible light irradiation, affording a CO production rate up to 47.37\u202f\u03bcmol\u202fg\u22121\u202fh\u22121 with a high selectivity of 98.3%. It was demonstrated that the IL enhanced the CO2 photoreduction to CO and suppressed H2 evolution.The natural photosynthesis of CO2 with H2O produces biomass with O2 release and no H2 generation. However, in the artificial photocatalytic process of CO2 with H2O the generation of H2 is hardly avoided due to the presence of competing reaction of proton reduction. Therefore, the rational design of POPs based catalysts is important. In our recent work, we prepared a kind of amide-bridged conjugated POPs via self-condensation of amino nitriles (i.e., diaminomaleonitrile, DAMN; 2,3-diaminobut-2-ene-1,4-dinitrile, DAEN; 3,4-diaminobenzonitrile, 34AB) in combination with subsequent hydrolysis, as illustrated in Fig.\u00a014\na [138]. These COPs are full of CO2-philic groups (CO, C\u2013NH-), thus exhibiting high CO2 adsorption capacities (up to 32.7\u202fmg\u202fg\u22121) in spite of their low BET surface areas. Meanwhile, they could absorb visible light efficiently, displaying suitable energy band structures for reducing CO2 and oxidizing H2O (Fig.\u00a014d). Therefore, they realized the photoreduction of CO2 with H2O without any photosensitizer and sacrifice reagent under visible light irradiation. Interestingly, CO was obtained as the sole carbonaceous product and no H2 was generated. Among the resultant catalysts, Amide-DAMN with energy band structure (E\ng\u202f=\u202f2.19\u202feV, CB\u202f=\u202f\u22120.75\u202feV, VB\u202f=\u202f1.44\u202feV) displayed the highest activity, affording CO with a production rate of 20.6\u202f\u03bcmol\u202fg\u22121\u202fh\u22121, much better than the most reported metal-free catalysts. DFT calculations indicate that the adjacent redox sites of this kind of photocatalysts makes the CO2 photoreduction couple well with H2O photooxidation, inhibiting the generation of H2 (Fig.\u00a014e).Besides amide-bridged conjugated POPs, other POPs with unique chemical structures also exhibited excellent selectivity towards target reduction product in CO2 photoreduction with H2O. For example, 2,5-diphenyl-1,3,4-oxadiazole derived CMPs (OXD-TPA) achieved CO2 photoreduction with H2O to access CO as the sole carbonaceous product in a selectivity nearly 100% under visible light irradiation [139]. In general, the photocatalysts with electron donor and acceptor (D-A) structures can enhance the light-excited charge carrier separation and transmission, thus resulting in better photocatalytic performance. The imine-linked COF with D-A structure (CT-COF) from 9-ethyl-9H-carbazole-2,7-dicarboxaldehyde and tris-(4-aminophenyl)triazine realized photoreduction of CO2 with H2O to CO in a selectivity of 100% with an evolution rate of 102.7\u202f\u03bcmol\u202fg\u22121\u202fh\u22121 under visible light irradiation [140].Compared to conjugated organic polymer semiconductors, inorganic semiconductors possess better photosensitivity, but suffer from high band gap energy and low CO2 adsorption capacity, thus lowering its performances for CO2 photoreduction [110]. The combination of POPs with inorganic semiconductors can overcome this shortcoming, showing enhanced activity [120]. For example, Tan's group [141] decorated porous hyper-crosslinked polymers onto the TiO2 particles (HCP\u2013TiO2-FG), which significantly improved the CO2 adsorption capacity (12.87\u202fwt%) and exhibited high activity for catalyzing photoreduction of CO2 with H2O, generating CH4 with a production rate of 27.62\u202f\u03bcmol\u202fg\u22121\u202fh\u22121. Constructing heterojunction between POPs and inorganic semiconductors is an efficient strategy to prepare efficient catalysts with high performance for CO2 photoreduction. For example, Dai and coworkers [142] developed a series of hybrid catalysts derived from conducting polymers (including polyaniline, polypyrrole, and polythiophene) and Bi2WO6 hierarchical hollow microspheres via in situ deposition oxidative polymerization. The resultant polythiophene/Bi2WO6 exhibited the best CO2 photoreduction activity, affording a methanol production rate of 14.1\u202f\u03bcmol\u202fg\u22121\u202fh\u22121 and an ethanol production rate of 5.1\u202f\u03bcmol\u202fg\u22121\u202fh\u22121. It was found that the improved photocatalytic activity resulted from decrease in the recombination of photogenerated electron\u2013hole pairs caused by the forming of Z-type heterojunction.To construct a Z-scheme photocatalytic system composed of POP and SnS2 nanoparticles, we designed sulfur-bridged CTFs (S-CTFs) nanospheres via nucleophilic substitution coupling of cyanuric chloride and trithiocyanuric acid (Fig.\u00a015\n) [143]. The resultant S-CTFs could adsorb CO2 with capacity of 2.6\u202fwt% and showed energy gap of 2.73\u202feV with CB at \u22120.93\u202feV. Treating the S-CTFs sample in SnCl2 ethanol solution at 180\u202f\u00b0C resulted in the formation of SnS2 nanoparticles on the surface of S-CTFs, forming SnS2/S-CTFs composites, which exhibited energy gap of 2.40\u202feV with CB at \u22120.27\u202feV. It was found that under light irradiation the transfer of the photo-induced electron-hole pairs in SnS2/S-CTFs follows Z-scheme mechanism, thus achieving effective separation of photo-generated carries. As a result, SnS2/S-CTFs displayed high efficiency for catalyzing CO2 photoreduction using TEOA as a sacrifice reagent under visible-light irradiation, yielding CO and CH4 with evolution rates of 123.6\u202f\u03bcmol\u202fg\u22121\u202fh\u22121 and 43.4\u202f\u03bcmol\u202fg\u22121\u202fh\u22121, respectively, much higher than those obtained over pristine S-CFTs and other reported photocatalysts. Notably, the SnS2/S-CTFs hybrid displayed excellent stability for photoreduction of CO2.Along with the substantial reduction in the cost of sustainable electricity, electrocatalytic CO2 reduction reaction (CO2RR) to value-added chemicals has attracted much attention in recent years [144,145], and the studies on POPs-based electrocatalysts (e.g., COFs [146,147] and CTFs [148]) are burgeoning. In 2015, Lin and coworkers [149] applied cobalt porphyrin-derived COFs as the electrocatalysts for CO2RR, affording a high FEco of 90% and ultrahigh turnover numbers of 290,000 with an overpotential of \u22120.55\u202fV. After that, kinds of metal porphyrin and phthalocyanine derived COFs have been developed and applied as electrocatalysts for CO2RR [150\u2013153]. For example, a series of 2D cobalt(II)-phthalocyanine based COFs with high conductivity were developed by Wang and coworkers which exhibited excellent CO faradaic efficiency (FEco) of 87%\u201397% at the potentials of \u22120.6\u223c\u20130.9\u202fV (vs. RHE) [150]. Lan's group [151] developed metallo-porphyrin-tetrathiafulvalene-based COFs with electron-rich tetrathiafulvalene to serve as electrocatalyst, which showed an ultrahigh FEco nearly 100% at the potential \u22120.8\u202fV (vs. RHE). Though great progress has been achieved, developing POP electrocatalysts with high catalytic performance and good long-term stability is still challenging. Just recently, we have developed perfluorinated CTFs (F-CTFs) by polymerization of tetrafluoroterephthalonitrile over Lewis superacids (e.g., Zn(NTf2)2), as shown in Fig.\u00a016\n [154]. Serving as an electrocatalyst in a three-electrode flow-cell system with gas diffusion electrode (GDE), the resultant F-CTF-1-275 showed excellent catalytic performance, affording a high FECO of 95.7% at \u22120.8\u202fV (vs. RHE) with a high current density of 114\u202fmA\u202fm\u22122. It was found that the high hydrophobic of F-CTF-1-275 caused by the high fluorine content (31\u202fwt%) inhibited the HER competing reaction and improved the CO2 adsorption ability. The low-temperature ionothermal strategy great extend the applications of CTFs in the field of electrochemistry.In this review article, we have described our recent work on POPs catalysts for CO2 transformation, with focus on the design strategies for various functional POPs based catalysts as illustrated in Fig.\u00a017\n. For the synthesis of POPs-based catalysts, we introduced CO2-philic groups (such as azo, Tr\u00f6ger's base, fluorine, phenolic \u2013OH) and ligands that can chelate with metal species into the skeletons of the polymers via various coupling reactions, with subsequent immoblization of metal active species onto the surfaces of POPs. A series of POPs-based catalysts for cycloaddition reactions of epoxides or propargylic alcohols with CO2, reductive conversion of CO2 with H2, and photocatalytic/electrocatalytic reduction of CO2 have been prepared, which have shown high efficiency for CO2 transformation. Our research work demonstrates that POPs-based catalysts have promising application potentials in CO2 capture and transformation. To achieve their commercial applications, the following issues are suggested to be considered and paid much attention in the future.Firstly, low-cost and environmental-friendly protocols should be developed. The most used approaches suffer from uses of expensive monomers, massive organic solvents, and noble metals catalysts, together with low product yields, which generally result in high cost and pollution. Biomass-derived feedstocks with multiple reactive sites are easily available and environmentally friendly, which are good candidates of the monomers of POPs. Therefore, biomass-derived POPs catalysts should be explored. Secondly, high-performance POPs catalysts that can achieve CO2 transformation under mild conditions should be explored. Due to its inherent nature, the transformation of CO2 under mild conditions is still challenging. The rational design on POPs that can capture and activate CO2 under mild conditions is highly required, which may achieve efficient conversion of CO2 under mild conditions. Especially, photocatalysis provides an ideal way to transform CO2 to chemicals, thus developing POPs-based photocatalysts with high efficiency is an interesting and important topic, which should be paid much attention. Electrocatalytic CO2 conversion has regard as the most promising approach to achieve CO2 transformation, however, the low conductivity of POPs-based catalsyts limits their applications as electrocatalysts and carbon black is still necessary in most cases. Therefore, developing POPs-based electrocatalysts with high conductivity are highly desirable. In addition, the catalytic stability is a key factor to determine the practical applications of the POPs catalysts. The POPs catalysts should be designed with high stability. Thirdly, the catalytic mechanism should be investigated. Comparing to homogeneous catalysts, the POPs-based catalsyts can combine multiple functions together and play multiple roles in chemical transformation of CO2. Understanding the catalytic mechanism is very important for developing efficient catalysts for CO2 transformation. In all, the study on POPs-based catalysts is still in its infancy, and the related research work is of significance for promoting the utilization of CO2.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors thank the National Natural Science Foundation of China (22121002, 21773266), Chinese Academy of Sciences (121111KYSB20200057) for the financial support.", "descript": "\n                  The transformation of carbon dioxide (CO2) into fuels and chemicals is an interesting topic, which has been paid much attention in recent years. The materials with specific functionalities are highly required for CO2 capture and conversion, which have been widely investigated. As an emerging material platform, porous organic polymers (POPs) have attracted considerable scientific interest due to their distinctive properties such as tailorable functionalization, large surface areas, adjustable porosity, versatile polymerizations, good physicochemical and thermal stability. Our group focuses on designing and synthesizing POPs via introducing CO2-philic groups and organic ligands into the skeletons of the polymers and immobilizing metal active species onto their surface, and a series of POPs with functional groups, such as azo, Tr\u00f6ger's base, fluorine, phenolic \u2013OH, have been prepared for CO2 transformation. In this review article, we mainly introduce our recent work on design of POPs-based catalysts for CO2 transformation, which include POPs-based catalysts for cycloaddition reactions of epoxides and propargylic alcohols with CO2, for reductive transformation of CO2 with H2, for photocatalytic/electrocatalytic reduction of CO2. In addition, the perspectives of the POP-based catalysts for CO2 transformation will be discussed as well.\n               "}
{"full_text": "Azo dyes are identified with one or more azo bond linkage (-N=N-) having the most prominent use of 70% among over 2,000 species of synthetic dyes [1\u20134,70]. Several studies have been carried out on the properties of these dyes which reveal their unique simplicity in synthesis, excellent fastness rating, high solubility, and uptake by the substrate. These properties give it a preferential choice of use among other dyes in textile industries (Oyetade et\u00a0al. [173]). However, these dyes are commonly present in discharged textile wastewater which results in environmental pollution, limiting the quality of life ([9,100]). An estimate of 50,000 tonnes of textile wastewater is annually discharged with 10\u201330% concertation of unfixed dyes from dyeing, printing, pigmentation, and bleeding of the textile substrate [13], (Adetuyi et\u00a0al. [175]; Oyetade et\u00a0al. [174]). The toxic impacts of these dyes in wastewater are based on high resistance to microbial degradation, conventional treatments and chemical transformation of the dye molecules in the effluent to more toxic pollutants (benzidine) [13,27,56]. Although various conventional treatment techniques have been reported. However, limitations of toxic sludge formation, cost and low dye removal efficiency of azo dyes necessitated the use of the improved photocatalytic technique in nanotechnology (10\u22129) for dye remediation [13,100,142,174]. The process of dye degradation/decolorization via the use of photon-active nanomaterials as adsorbents is called photocatalysis [27,56,124]. Studies have shown the effectiveness of these techniques in azo dye degradation among other processes of advanced oxidation processes (AOPS) (Fig.\u00a01\n) (Rauf and Ashraf [176]; Pandey et\u00a0al. [156]). The whole process initiates a redox reaction which spontaneously generates radicals (e.g hydroxides, peroxides, and superoxides) while forming electron-hole pairs. The radicals combine with the organic pollutants (e.g. dyes in wastewater) and mineralize them into compounds such as \n\nC\n\nO\n2\n\n\na\nn\nd\n\n\nH\n2\n\nO\n\n (Shindhal et al. [150] Pandey et\u00a0al. [156]).Although there exists a wide spectrum of nano-photocatalyst materials such as bio-based, metal oxide-based, polymer-based, carbon-based and inorganic-based materials [35,101]. However, the challenges corresponding to their uses as photocatalyst includes photo-corrosion, frequent electron-hole recombination, large band gap, low photon detection and capturing propensity and agglomeration of powdered nano particles (NPs) [6,123,127,130]. These challenges gave rise to the investigation of polymer-based nano-composite from conducting polyaniline (PANI). Conducting polymers such as polyaniline (PANI), polythiophene, polypyrrole, etc., have been known for multiple varieties of applications in electronic devices, sensors, anticorrosive coatings, energy storage devices, catalysis, etc [101].Interestingly, out of these varieties of conducting polymers, PANI stands out with a unique applicable potential in photocatalysis [42]. The polymer commonly synthesized by oxidative polymerization exhibits incredible charge transport dynamics which supports its high photon capturing activity and lower band gap [19,72]. Furthermore, it exhibits other unique potentials such as high adsorptive capacity, ability to form polymeric support for nano powder materials with enhanced morphology, tuneability of bandgap on fabrication with other nano materials as composites and high solubility based on its amphiphilicity [59,112,141]. The multifunctional potentials of this polymer in composite fabrication necessitated the study of improved functional potentials and performance evaluation of the polymeric nano composite catalyst. This present paper evaluates the photodegradation performance of PANI nano composite applied to textile wastewater laden with azo dyes as compared to the conventional catalyst as reported in literature.Azo dyes are the largest class of synthetic dyes which are commonly synthesized by diazotization and coupling reactions via several chemical routes [16]. Although they can equally be synthesized by Gewald reaction (Fig.\u00a02\n) however, the most adopted synthetic pathway for these industrial dyes is by diazotization and coupling reaction to give more brilliant shades, optimum yield, desired particle size and improved dispersibility [14,31,47,114].The process involves the diazotization of a primary aromatic amine, before coupling with electron-rich nucleophiles as described in Fig.\u00a02\nAzo dyes structurally consist of one or more azo (\u2212N=N\u2212) chromophoric group(s) and mostly water-soluble sulfonic (\n\nS\n\n\nO\n\n\n3\n\u2212\n\n\n\n) (Fig.\u00a03\na) group which justifies their affinity for water and excellent fastness rating (Sudha et\u00a0al., [177,90]. These structural features also explain their increasing industrial demands and presence in most textile wastewater). Furthermore, auxochromes linking to phenyl or naphthyl rings such as amine, chloro, hydroxyl, methyl, nitro, sulfonate may be present in their structure which contributes to bathochromic shift thereby enhancing colour intensity (Sudha et\u00a0al. [177]).Generally, dyes are classified into two broad categories concerning their application or chemical structure. Classification based on applications separates dyes into groups such as reactive, acid, basic, sulfur dye, mordant, direct, disperse, pigment, vat, and azo dyes (Popoola [90]). While structurally, they can be categorized as, indamine, diphenylmethane, xanthene sulfur, carotenoid, acridine, quinoline, anthraquinone, nitro, azo, indigoid, amino- and hydroxy ketone, phthalocyanine, inorganic pigment, etc. Although based on their charge they may be either anionic, non-ionic, or cationic dyes . However, based on their chemical reactivity they are classed as either acid, basic, reactive, direct or disperse azo dyes [128]. Furthermore, classification concerning azo constituents could be mono azo (one azo group), diazo (two azo groups), triazo (three azo), tretrakisazo (four azo) or more (poly azo) [13]. However, based on their colour index in Table\u00a01\n, they are classed monozo, diazo, triazo, polyazo and azoic dyes [102]. These varieties of classification, their aromatic constituent(s) and the auxochromic substituent(s) account for their vast application in the textile industries and determine the bonding arrangement with cellulosic, protein, regenerated, and synthetic polymers [90,102]. Popoola [90] and Oyetade et\u00a0al. [174] added that these dyes have strong dye-fiber interaction (Fig.\u00a03b), which accounts for their excellent fastness rating and resistance to bleeding, crocking and generally running off from the substrate applied.Statistically, they have 70% use among other industrial dyes with applications in paint, paper and most especially textile industries [102]. Despite their vast use, azo dyes remain an industrial dye of global threat. Among the assortment of azo dyes, the most widely used reactive azo dyes are associated with the toxic environmental impact on disposal, and difficulty in the treatment of their corresponding effluents (Jagadeesan et\u00a0al., 2021). The toxicity of azo dyes stems from the hydrolysis reaction of loose dyes present in textile wastewater after textile substrate application. On discharge, the dye molecules become persistent in the environment and can bio-transform into aromatic amine products with acute mutagenic and carcinogenic effects on the organism (Bruna et\u00a0al. [178]). Textile effluents laden with azo dyes, especially the commonly used Congo red, methylene blue and green, benzidine may predispose a man to various impaired health effects and hormonal dysfunction such as mutagenesis, quadriplegia, jaundice, cyanosis, tissue necrosis, vomiting chromosomal fractures, carcinogenesis, and respiratory toxicity (Vinothkannan et\u00a0al. [179]). The toxicological impacts of these dyes are connected to the substituent position and the nature of the dyes (Bruna et\u00a0al. [178]). Some of these dyes are less toxic but their presence in wastewater after dyeing, printing, or pigmentation processes may cause a reduction in the azo bond, imparting them with strong mutagenic action (Vinothkannan et\u00a0al. [179]). For instance, direct black 38 (azodisalecylate) gives a breakdown product like benzidine (Fig.\u00a04\n\n) and other derivates such as anilines nitro semis, and dimethyl amines which have carcinogenic inducing effects in humans and animals [100,111]. Furthermore, the presence of these dyes discharged into water bodies constitute a significant reduction of light penetration thereby producing different amine with higher genotoxic and mutagenic effect (Ventura et\u00a0al. [180]; Bruna et\u00a0al. [178]).Literature has reported and recommended the use of various conventional treatment processes for textile wastewater laden with azo dyes and dyeing auxiliaries. These processes are functionally described as physical, chemical, or biological treatment processes (Table\u00a03) ([148] Ventura et\u00a0al. [180]; Bruna et\u00a0al. [178]; Siani, 2017) [111]. Table\u00a02\n describes these conventional treatment approaches and their corresponding merits and limitations. Among the treatment technologies, recent studies have emphasized the use of adsorption and biological treatment process which are classed as secondary treatments given to textile wastewater. However, the problem of disposal of sludge and the recovery of materials used for the treatment process are fundamental challenges of these technologies [85,111]. On the other hand, the use of advanced oxidation processes in recent times has appreciable advantages of faster reaction kinetics and the generation of little or no toxic sludge. However, the sophistication of some instruments for the treatment brought about the study of high dye remediating photocatalysts using which gives provision for the use of various nanomaterials in their pure form and as composites with the possibility of recovery [45,106]. These evolving limitations gave rise to the evolution of the photocatalytic treatment approach for textile wastewater using nano polymeric composites (Jangid et\u00a0al. [181]).The uses of nano technological materials as adsorbents have grain prominence and applicability to environmental science. This thematic field studies the use of nano scale materials (10\u22129) characterized by novel, versatile and multiple functionalized properties for the treatment of toxic dye pollutants [17]. The transition in structural, functional and reactivity of nanomaterials from bulk scale to nano-size offers its desirable usage as a catalyst in photocatalysis [37], Mishra [75]. Although, these materials are generally classed as conductors, semiconductors and insulators. However, this classification is a function of the value of their respective bandgap, which is pivotal for the photocatalytic process [8,130]. Nano adsorbent materials are broadly referred to as either organic or inorganic (Fig.\u00a04) with distinguishing features of photosensitivity, high surface area, high thermal chemical and mechanical stability, appreciable electrostatic features, compressibility, tunability of pore volume and bandgap, high magnetic and adsorptive capacity and enhanced solubility properties due to short intra-particle diffusion [36,124].These distinguishing features account for their application in the photocatalysis of recalcitrant dye molecules in industrial effluents [13]. The use of the photocatalytic technique in nanotechnology is described as a photon-induced molecular transformation that occurs at the surface of exciting photoactive nanomaterial adsorbing organic pollutants (e.g dye molecules) from the wastewater [86]. The mechanism of the degradation process starts with the capturing of photon energy from light by the photocatalyst leading to the excitation of the electrons from the valence band (VB) to the conduction band (CB), forming oxidizing and reducing sites (Fig.\u00a06\n) (Antonio et\u00a0al., 2019) [86]. For electronic excitation to occur, the energy of the photon captured by the material must be equal to or greater than the energy of its bandgap. This excitation leads to the generation of hydroxyl radicals (\u2022OH), superoxide radical anions (\u2022O2\u2013), and hydroperoxyl radicals (\u2022OOH), which are oxidizing species (Xing et\u00a0al., 2018). Dyes adsorbed already by the nano adsorbents from the wastewater combine with the electrons in the conduction band resulting in the formation of dye radical anions and consequently degradation of the dye molecules (Hossain et\u00a0al. [205]; Sioni et\u00a0al., 2020). For instance, in the photodegradation process of azo dyes present in textile wastewater, the energy of photons captured via UV or sunlight generates electron-hole pairs (e\u2212 and h+) which migrate to the surface and site of the adsorbed azo dyes.This migration set-up a redox reaction (Fig.\u00a06), which produces oxidizing radicals that attacks the adsorbed azo dye molecules and degrade them into non-toxic substances such as \n\n\nH\n2\n\nO\n\na\nn\nd\n\nC\n\nO\n2\n\n\n (Comparelli et\u00a0al. [182]; Zhu et\u00a0al. [183] 2013; Soltani and Entezari [184]; Ullah et\u00a0al. [185]). Photocatalyst nano-adsorbents used for this process can be bio-based, metal-organic frameworks (MOFs), carbon-based, polymeric-based or inorganic as described in Fig.\u00a04 However, commonly used nano adsorbents in photocatalysis are the inorganic metal oxides (Table\u00a03\n) [85]. This is due to the promising potentials such as photon detection and capturing, electronic structure, carrier transportation and band gap [30,77,110].The study by Dutta et\u00a0al. (2021) reveals that metal-oxides-based adsorbents exist as magnetic or non-magnetic. An investigative study by Mohamed et\u00a0al. [79] on one of the magnetic classes of the nano-adsorbents (Fe3O4-Nps) has an adsorption capacity of 150\u2013600 mg g\u22121 for rhodamine dye within 30 min, while the fabricated composites such as Fe3O4/CeO2, Fe2O3\u2013Al2O3 have a notable adsorption capacity of six-times higher than the pure metal oxides nanomaterial in dye remediation [73]. This action is due to their surface-to-volume ratio and pore size, which is consequent to bandgap tuneability for improved photocatalytic effect [33,36,112,132]. On the other hand, the non-magnetic metal-based oxides ZnO, TiO2, MgO are often fabricated as a hybridized composite with notable adsorption efficiency [36,85].Examples of these composites with their corresponding adsorption capacity include Co/Cr-co doped ZnO with 1057.9 mg g\u22121 for methyl orange, ZnO\u2013Al2O3 nano adsorbents for Congo red and Ni-MgO having maximum adsorption of 397 mg g\u22121 [62,64,79]. Dutta et\u00a0al. (2021) revealed that apart from their appreciable specific surface area, the charge on the adsorbent surface gives room for electrostatic dye-adsorbent interaction. Also, the polymorphic nature of the composite nano adsorbent creates a more active site for the binding of dye molecules. Although the challenges of toxicity and frequent recombination notably exist for the metal oxides photocatalyst in Table\u00a04\n,which greatly limits their performance in dye degradation [8,42]. Additionally, ceramic nanoparticles are another class of adsorbent chemically existing as oxide, phosphate and carbonates (silica, alumina, titania, zirconia), with high chemical inertness and heat-resistance [36,58]. However, the limitation of the large bandgap of these materials incited the study of carbon-based nanomaterials with unique structural features fit for adsorption and composite fabrication. Carbon-based nanomaterials generally exist as either carbon-nanotubes (single-walled or double-walled) or carbon-fullerenes [103]. One of the unique examples of carbon-nanotubes (CNTs) is graphene sheet which can be rolled in the form of tubes, having characteristic features of thermally-conductive, less toxic photoactive, bandgap tunability and synergistic composite forming potential [38,42,131].Bezerra de Araujo et\u00a0al. [18], Oni and Sanni [88] and Sivakumar et\u00a0al. [121]discussed the high adsorptive performance of GO for toxic dye pollutants such as Direct Red 81 and Indosol SFGL direct blue, crystal violet and methyl orange, and methylene blue respectively at pH 6-7. Sivakumar et\u00a0al. [121] added that the efficiency of GO is due to the interaction of the hydroxyl and carboxylic groups with the auxochormic substituents (functional groups) on the dye molecules. Furthermore, Zheng et\u00a0al. [140] related that the hierarchically unique orientation of GO\u2013NiFe-LDH nano composites enhance its adsorption efficiency for Congo red and methyl orange at 489 and 438 mg/g respectively. Similarly, the composite of GO with metal oxides (GO/MgO) studied by Heidarizad and \u015eeng\u00f6r [49]notable adsorption of 833 mg/l within a contact time of 60 based on the multiple functional sites available for binding. Although GO is thermally, mechanically and chemically stable, however, the need for the use of photosensitizers in place of toxic metal and metal oxide NPs justifies the investigation of p-type conductive polymers especially the novel polyaniline and its composite forming mechanism with other materials with appreciable advantageThe beneficial attributes of high effective surface area, high selectivity and absorptivity, appreciable doping/de-doping technique, effective electrical transport characteristics, well-established binding affinities, and unique textural properties offer polyaniline a leading advantage [71,118]. Mu et\u00a0al. [84] reported a high adsorption capacity of 248.76 mg g\u22121 for Congo red dye graphene/polyaniline and PANI/Fe3O4 respectively. Although in its purest form Smita et\u00a0al. [122] reveal that PANI exhibit an adsorption efficiency of 92% for methyl orange due to the structural versatility and the presence of amine and imine active group. However, higher adsorption and reduced recombination enhance photodegradation when hybridized composites of PANI are formed with other nanomaterials [103]. This also improves the surface area and enhances photocatalytic reaction sites thereby inducing electron-hole separation [92]\nIn photocatalysis, vital parameters such as pH, initial dye concentration, temperature, nature and dosage of nano adsorbent, contact time, irradiation time, and irradiation intensity play a significant role in the adsorption-degradation process as described in Table\u00a03 [15,93,96]. As the initial dye concentration increases the adsorption and degradation increase up to the point where the binding sites on the adsorbent are saturated, beyond this point desorption occurs [35,93,99]. The pH on the other hand plays a very crucial role in the determination of adsorption and consequent degradation of dye molecules adsorbed onto the surface of a nano photocatalyst [95]. Research by Salleh et\u00a0al. [104] reveals that low and high pH enhances the adsorption of anionic and cationic dyes, respectively. This claim was affirmed by Daneshvar et\u00a0al. [29] and Phoemphoonthanyakit et\u00a0al. [89] in their research where higher adsorption of 1093 mg g\u22121 at pH 2 for Acid Blue 25 and 600 mg g\u22121 at pH 7 for Rhodamine 6G at pH of 2 and 7 respectively.Furthermore, photocatalytic decomposition of dye molecules in wastewater has been studied at pH values ranging from 3 (acidic) to 13 (alkaline) for anionic, cationic and neutral dyes in wastewater [69]. The pH value required for the reaction is suggestive of the kind of charge on the surface of the nano adsorbent. This is because at pH <pHzpc, pH>pHzpcor pH=pHzpc it implies that the surface charge is positive, negative and neutral respectively [15]. Hence, cationic dyes are adsorbed more in an alkaline medium to establish electrostatic interaction resulting in increased degradation efficiency [15,53]. Alakhras et\u00a0al., [5] added that at low pH there is a notable reduction in the production of hydroxyl radicals by the positively charged surface which is needful for hydroxyl radical formation. Also, temperature requirement quantitatively determines whether the process is endothermic or exothermic which influences the adsorbent-adsorbate interaction prior to photocatalysis [35].Generally, physical properties such as high selectivity, high absorption capacity, durability, reusability, cost-effectiveness, surface area, optoelectrical properties, crystallite size and distribution, dispersibility, and mechanical and thermal stability are vital properties that determine the choice of nano adsorbents materials in photocatalysis [41,42]. Table\u00a04 describes the physical properties of the commonly used conventional catalyst with metal oxides taking the highest frequency of use [37]. Although the frequent use of these is based on their polymorphic nature. However, electron (e\u2212)- hole (h+) recombination, agglomeration and photo-corrosion of some of the metal oxide nanoparticles limits their performance, recovery and reuse (Meng et\u00a0al. [186]). Another limiting challenge of this conventional catalyst is that their photo capturing propensity is only within the ultraviolent region Beyond this region to the visible, the nanomaterials exhibit low photon detection and capturing potential which lowers the generation of radical species needful for dye degradation [25]. Furthermore, from Table\u00a04, semiconductors such as TiO2,h-2DBN, Nb2O5, ZrO2, and ZnO2 have appreciably high band-gap which limits their photon capturing potentials, especially when irradiated within the visible region [112,130]. Additionally, the use of these semiconductors in their pure form as nano-adsorbent has toxicological impacts and reduced photocatalytic performance [13,136]. Xu et\u00a0al. [187] suggest that the reduction in adsorption capacity during adsorbent-adsorbate contact is due to the saturation of the adsorbent surface and charges present on synthetic dye molecules, especially azo dyes. For instance, cationic azo dyes easily undergo adsorption and degradation when compared to the anionic azo dyes such as Eriochrome Black T due to the lack of electrostatic interaction [74] (Xu et\u00a0al. [187]). This is due to the charge present on the nano-adsorbent, the difference in sorption mechanisms and the value of the pH which is greater than the pHpzc, favouring the preferential adsorption of cationic dyes compared to its anionic counterpart [15,116]. This action justifies the report on the faster adsorption-degradation kinetics of cationic azo dye using \n\nT\ni\n\nO\n2\n\n\n when compared to the anionic quinizarin, having a low adsorption efficiency of 21.8% with the same catalyst (Abuabara et\u00a0al. [188]; Pereira et\u00a0al. [189]).Also, the challenges of recovery and reuse of conventional catalyst after the process of photocatalysis has limited their commercial acceptance in the treatment of wastewater laden with recalcitrant dyes from textile industries (Muhd et\u00a0al. [190]). Hence, tackling the attendant limitations of these nano-adsorbent requires the fabrication of nanocomposite by combining these semiconductors with a photon sensitizing conducting polymer (PANI) characterized by high resistance to corrosion, high adsorptive capacity, chemical and thermal stability and versatile surface area [54,112]. The fabrication of composite mix form using these materials enhances surface modification and bandgap tunability necessary for improved performance (lowering the bandgap- Fig.\u00a07\n) [30,112].There are various types of \u03c0-conjugated conducting polymers with appreciable conductivity and novelty. These emerging conducting polymers include polyaniline (PANI), polythiophene (PTH), polypyrrole (PPY) poly(para-phenylene) (PPP), poly(phenylenevinylene) (PPV) and polyfuran (PF) [17,22,44]. However, the most prevalent among these varieties is polyaniline (PANI). This is based on unique properties such as exceptional electronic and optoelectronic features, cost-effectiveness, ease of synthesis, environmental stability, high electrical conductivity, high thermal power and unique structural ordering [19,65,87,110]. Various approaches such as interfacial polymerization, seeding polymerization, vapor phase self-assembling polymerization, photo-induced polymerization, plasma polymerization, sonochemical synthesis, and electrochemical synthesis can be used in the synthesis of polyaniline [19,109]. However, the solution technique via oxidative polymerization of aniline with ammonium persulphate is commonly used due to its faster polymerization rate and excellent yield (Fig.\u00a07). The resultant products of this polymerization process can be either leucoemeraldine (fully reduced state), emeraldine salt (half oxidized state) or pernigraniline (fully oxidized state) which are described in Fig.\u00a07 [17,19]. Out of these products, emeraldine salt stands out with unique structural characteristics of two benzoid units and one alternating quinoid unit which results in high thermal stability and nanocomposite forming potential with inorganic nanoparticles such as CeO2, TiO2, ZrO2, Fe2O3, Fe3O4, ZnO and TiO2 thereby generating nanostructures such as nanofibers, nano tubes, nano sphere and nano flowers [17,19]These unique structural features and versatility account for its vast application in corrosion protective coatings, energy devices, sensors, water purification and as photocatalyst [100,111]. Current research reveals that PANI exhibits incredible photoelectronic properties and the composite resulting from its incorporation with other materials has appreciable synergic properties higher than the use of polymers [87,109,137]. Furthermore, PANI is amphiphilic based on its conjugated organic part (Quiniond) and\u2013NH\u2022+ formed during protonation, the excellent dispersion and ability to bind with organic and water-soluble dyes in effluent account for its choice among other polymeric adsorbents [19,20,66,141]. Also, the structural ordering of the polymer chain and its high conjugation give it dynamic electrical properties associated with high carrier transport mechanism which is vital for the photocatalytic process [19,109]. The dynamics of electrical transport in polyaniline are described by its intra-chain or inter-chain transport processes [109]. The inter-chain transport process depends on the carrier delocalization of the polymer chain with appreciable conjugation length [63]. On the other hand, the inter-chain charge transport process is dependent on the hopping mechanism, based on the molecular crystalline packing of the polymer matrix [30,109]. The unique molecular orientation of PANI accounts for its improved charge mobility properties which enhances its photosensitivity when existing in pure form or as composites with other semiconductor materials [83]. Apart from the hopping of charge carriers along and between polymer chains, other incredible conductive mechanisms of polyaniline can be tunnelling between high-conductive crystallites embedded into the amorphous matrix or electron-phonon interaction [30].Composite fabricated from PANI is aimed at tackling the inefficiency of conventional catalyst and treatment process for recalcitrant dyes. Various cost-effective and environmentally friendly approaches have been investigated by literature on the fabrication of polymeric nanocomposite catalysts with polyaniline (Liu et\u00a0al. [204]). Studies show that the fabrication of these novel nanocomposites is mostly by ex situ or in situ polymerization (Vargas et\u00a0al. [191]). The ex-situ polymerization involves the mixing under sonication of the semiconductor's particles with already synthesized polyaniline (Cruz et\u00a0al., 2017). However, the commonly used composite fabrication technique (in situ polymerization) involves homogenous dispersion of the nano-grade semiconductors or other nano-size materials of choice during the polymerization process of polyaniline (Fig.\u00a08\n). This method of fabricating nanocomposite is more appreciable than the former because the former exhibits the formation of low\u2010density and non\u2010uniform coverage of nanostructures of polyaniline by the material semi-conductor (Vargas et\u00a0al. [191]). The in-situ polymerization of polyaniline with semiconducting material such as \n\nT\ni\n\nO\n2\n\n,\n\nZ\nn\n\nO\n2\n\n\n, forms coupled nano polymeric composites (Figs.\u00a08 and 9\n) (Tai et\u00a0al. [192]; Liu et\u00a0al. [204])During composite fabrication, PANI acts as a photosensitizer while interacting with the bandgap of the semi-conductors in Fig.\u00a09. This consequently lowers the bandgap of the The presence of the semiconductor provides the composite with appreciable morphological structure for dye adsorption [101]; Jangid [181]). Consequently, composite mix improved the subsequent performance of the photocatalyst by preventing frequent recombination of electron-hole pairs during photocatalysis (Jangid [181]). Furthermore, during photocatalysis, the advantages of high mobility charge carriers, stability, and the strong coupling effect existing between the polymer and the semiconductors, make the nanocomposites function effectively in the degradation of dye molecules within the visible region. (Bingham and Daoud [193]; Riaz et\u00a0al. [99]; Jadoun et\u00a0al. [194]). Ansari et\u00a0al. [195] and Hao et\u00a0al. [196] added that the tuned morphological property of the composite proffer higher performance in photocatalysis compared to the conventional catalyst. Morphologically, the presence of PANI enhances the high hole transporting ability and stability of the material during the remediation process (Vikas et\u00a0al. [197]). Jangid [181] and Shahabuddin et\u00a0al. [112] revealed that the presence of the conducting polymer in the composite mix accounts for the photocatalytic stability and activeness of the material under visible light after five consecutive runs. Studies by Saha et\u00a0al. [101], Shahabuddin et\u00a0al. [112] and Ameen et\u00a0al. [10] reveal that PANI exhibited well-defined tubular morphology, however, incorporation of h-BN and graphene nanosheet correspond to the formation of the granular polymeric network. Consequently, the surface modification improved the adsorption capacity by enhancing the availability of the binding sites of the composite for adsorption [39,101]. Furthermore, in Fig.\u00a09, the adsorption of dye molecules unto the surface of the nano composite of polyaniline is higher as a result of electrostatic interactions and \u03c0\u2013\u03c0* conjugation among the aromatic rings of dye molecules [109,110]. In the fabricated composite, PANI act as a photosensitizer via photon capturing and excitation of valence electrons in HOMO, jumping to LUMO through \u03c0\u2013\u03c0* transitions [19,76,112]. Also, the incorporation of semiconductor material into the PANI polymeric chains by in situ polymerization prevents recombination such that, as the positively charged holes (h+) are returning to HOMO to recombine, the empty conduction band of the semiconductor intercepts the recombination thereby improving it photocatalytic efficiency [8,42,112].The current research scope focuses on the vital need to improve composite fabricated from polyaniline to enhance effective industrial performance, recovery and reuse [85,120,135]. Recent studies discussed the use of immobilization techniques to enhance the recovery and improve the performance of these catalysts for several runs [97,120]. The study suggested the fabrication of nano composite photocatalyst via immobilization of the semiconductors either on the surface or in the PANI polymer matrix [117]. The technology involving the immobilization of the photocatalyst nanoparticle on the surface of the polymer matrix could be via dip-coating, chemical vapour deposition, grafting or plasma treatment followed by UV irradiation (Fig.\u00a010\n) [7,21,66,135].Alhaji et\u00a0al. [7] and Lin et\u00a0al. [66] added that this method enhances the reduction of agglomeration, creation of more active sites, surface modification and its cost-effectiveness. Although the method is disadvantaged by wide particle size distribution. On the other hand, the technology involving the immobilization in the support matrix enhances the reduction in leaching at lower energy and higher catalyst recovery propensity, although may be limited by agglomeration inside the matrix [117,135]. One of the most effective processes in immobilization in the polymer matrix is the in-situ process. This process is characterized by considerable advantages of lower bandgap and higher dye-degrading efficiency as highlighted in Table\u00a05\n. The in-situ process can either be by in situ polymerization (described in Fig.\u00a08: incorporation of TiO2 on PANI nano rods), or the sol-gel method [54].The in-situ polymerization involves a controlled blending of selected semiconductor nano material with the neat monomer before subsequent polymerization [117]. However, the sol-gel approach is a two-stage technique that involves hydrolysis where bond cleavage occurs between the organic matrix and the semiconductor nano powder before condensation which involves bond formation alongside small leaving groups [3,34]. The use of the in-situ process enhances the homogenous dispersion of the semiconductors on the PANI matrix, and reduces possible agglomeration [3,34]. Examples of other processes are mixing, hydrothermal method catalyst deposition, galvanostatic method and plasma-irradiation treatment which have the challenges of nano powder agglomeration and its energy-intensive [46]. Studies carried out by Jangid et\u00a0al. [54] and Shahabuddin et\u00a0al. [112] on the fabrication of catalyst nano composite via in situ polymerization to generate hybridized h-BN/PANI and TiO2/PANI nano rod exhibited excellent photodegradation efficiency of more than 90% and with over four runs due to the immobilization of the nano powder in the polymer matrix which enhances the recovery and reuse. Similarly, Mondal & Sharma [81] and Singh et\u00a0al. [119] also suggested the use of the immobilization method in the polymer matrix to improve the adsorption capacity enhance their photocatalytic water splitting potential and boost the possibility of recovery of titania.Also, Lin et\u00a0al. [66] discussed three possible approaches to improve the functional features of the polyaniline-based nanocomposite. Firstly, the solute-solvent interaction of PANI can be enhanced via the use of Graphene oxides (GO), and CNTs as additives which limit fouling, enhance the rate of flux recovery and improve the hydrophilicity of the conducting polymer. Secondly, the use of secondary amine with molecular width < 4.53 \u00c5 and a pKa>7 to facilitate gelation inhibition. This also provides needful solvent-polyaniline interaction for the enhanced adsorptive potential of the nanocomposite. Thirdly, the use of zwitterions to form zwitterionic polyaniline. The zwitterions are characterized by charge functional groups of both positive and negative. This enhances the Pickering emulsions at the interface between organic solvents and aqueous solution leading to improved solubility of the polymer matrix [30]. Other techniques for the production of high profile multiple functionalized PANI based composites are photo-induced polymerization, interfacial polymerization, electrochemical polymerization, solution polymerization, seeding polymerization, emulsion polymerization, vapor phase self-assembling polymerization, plasma polymerization, and sonochemical synthesis process techniques [109]\n\nTable\u00a06\n shows the polyaniline nanocomposites with their respective degradation time and percentage, while Table\u00a07\n reveals the degradation time and percentage of conventional catalysts used in photocatalytic remediation of industrial azo dyes in wastewater. From Table\u00a06, degradation percentage of 97% and 94.35% degradation of methylene blue at 150 min and 30 min respectively was recorded when the composite fabricated with PANI was used in the presence of UV radiation. Also, for methyl orange 94%,81.3%, and 95% degradation were obtained within 30, 125, and 90 min of photocatalysis. The observed time variation for the reaction is a function of the nature and composition of the composite and dyes, pH, dosage of the catalyst, contact time, and temperature (Reze et\u00a0al., 2017). Furthermore, the composites of PANI used under influence of sunlight energy show a high percentage performance of 92%, 95% and 97% for the degradation of azo in Table\u00a06 at the rate of 45,45 and 180 min respectively. However, in Table\u00a07 when sunlight irradiates the photocatalytic system using conventional catalyst low degradation percentage of 25%, 21.8%, and 56% for the industrial azo dyes at the longer duration of 120, 180 and 180 min was recorded. Ameen et\u00a0al. [213] and Vargas et\u00a0al. [191]) added that the faster photocatalytic reaction of PANI composite to sunlight energy was because of the conduction band of the semi-conductors used and the LUMO level of the PANI are well matched for the charge transfer. This consequently, promotes the electrons p\u2013p* absorption band of the outside PANI film upon irradiation with sunlight light leading to faster reaction kinetics (Sadia et\u00a0al., 2012; Rita et\u00a0al., 2017). Also, when the same dyestuff present in the wastewater was comparatively assessed, at 90 min in Table\u00a06, the degradation efficiency of methylene blue azo dye was 93% and 99.6% while in Table\u00a07, 97%, 92%, and 90% were recorded at longer duration time of 120, 350, 180 min respectively for the same dye. Others include acid blue having percentage degradation of 95% and 73% at 45 min in Tables\u00a06 and 7, respectively and Rose Bengal having 97% in 150 min in Table\u00a02, while it has 81% in 180 min in Table\u00a07. However, the low percentage of degradation of 56% from Table\u00a02 of the same dye may be due to the adopted method of composite fabrication (Vargas et\u00a0al. [191]).Furthermore, the photosensitizing action of PANI when coupled with other semiconductors enhances the photocatalytic performance of the nanocomposites within a shorter duration as compared to its conventional counterpart (Muhd et\u00a0al. [190]; Yang et\u00a0al. [198]). This action is based on the narrow bandgap and enhanced charge mobility of PANI resulting in photon response as the energy level of PANI is incorporated into the semi-conductors [61,133]. Yang et\u00a0al. [198] added that the synergistic effect between PANI and semiconductors such as TiO2, makes it act as sensitizer for its large bandgap (3.2 eV), resulting in the reduction of the bandgap during photon excitation (Yang et\u00a0al. [198]). The surface of the negatively charged TiO2(rod) undergoes electrostatic interaction with positively charged anilium ion from the PANI forming a composite with improved the photocatalytic degradation of phenol in the azo at strong adsorption of light in the visible region (Yaseen and Scholz [214]). Furthermore, the use of PANI as a coating in in-situ polymerization of nanomaterial forestalls the possibility of recombination rate of electron-hole (e\u2013 \u2013 h+) by contributing to the bathochromic shift of the absorption band (Wang [207]; Yang et\u00a0al. [198]; Jiang et\u00a0al. [206]). Although, the use of metals as doping as a nanomaterial for the degradation of dyes comes with high dye degradation efficiency in Table\u00a07 (92.6% and 92%). However, this technique is not cost-effective, not reproducible, and has a problem with catalyst recovery, hence it becomes imperative to use the alternative of nanocomposite developed from polyaniline (Yang et\u00a0al. [199]; Abd El-Rady et\u00a0al. [200]; Sood et\u00a0al. [201]).\nTable\u00a08\n shows various microorganism which serves as bio-catalysts used in the degradation of industrial azo dyes, their degradation time, and percentage. From the report in the Table, the microbial degradation of azo dye molecules from textile wastewater is characterized by a longer degradation time when compared to the photocatalyst used in Tables\u00a06 and 7. The extended time makes the technique inappreciable for industrial commercialization owing to the slow dye decolorization rate and consequent generation of more toxic substances during continuous industrial dyeing and pigmentation processes. (Bruna et\u00a0al. [178]) [55]. Also, from Table\u00a08, only fungal laccase bio-catalyst has the lowest degradation time of 12 h to degrade 99% of methyl orange. This efficiency is based on its high redox potential (400 to 800 mV) resulting in the generation of intermediates via two distinct pathways [134]. However, one major setback of this bio-catalyst is that it is unstable at elevated temperatures and alkaline conditions, with which most effluent emerges after industrial usage (Shahzad and Burhan, [202]). This challenge, however, is an advantage to polymeric composites of polyaniline (Shahzad and Burhan, [202]). Other organisms from the Table exhibiting degradation at a duration lesser than 2 days are fungal\u00a0Aspergillus niger\u00a0having 98%, 94% and 99% for Acid red 151, Orange II, and Congo red respectively, and\u00a0Pseudomonas\u00a0sp\u00a0having 83% dye degradation at 24-33h for reactive black 5. It is also worthy of note to add that some of these microorganisms from the table have low degradation efficiency over an extended time. Examples of these are\u00a0Mutant Bacillus sp. ACT2 with a degradation efficiency of 12\u201330% for Congo red at 37\u201348 h,\u00a0Bacillus cereus\u00a0having a degradation efficiency of 67% for Cibacron black PSG and Cibacron Red P4B during 5days. Although these bacteria cleave \u2013N=N\u2013 bonds reductively and utilize amines as the source of carbon and energy for their growth. However, their stability to high pH, temperature, and the recalcitrant nature of some azo dyes remains a major challenge (Saratale et\u00a0al. [203]; Mohammed and Burhan, 2014). Furthermore, the report by Buitr\u00f3n et\u00a0al [208] reveals that aerobic microbes cannot reduce azo linkages, their ability to destroy dye chromogens is lower when compared to the anaerobic bacterium. This justifies the low degradation of some bacterial microorganisms in Table\u00a08. Other limiting tendencies of this bio-catalyst are problems of early saturation, making the nano-polymeric photocatalyst of greater industrial acceptance than the microorganism (Vikas et\u00a0al., 2013; Shindhal et al. [150]).The use of nano photoactive polymeric composites in photocatalysis incredibly proves to be most effective in performance during adsorption and degradation of toxic azo dyes appreciably present in the effluent. The study was able to establish the stability, flexibility, versatility cost-effectiveness, high performance and the possibility of recovery and reuse of the catalyst composite for treatment of pilot and field-scale wastewater. Additionally, the fabrication of this composite form polyaniline lowers the band gap and creates photon capturing possibility within the visible region, prevents agglomeration of Nps, lowers recombination of electron-hole pair and creates a more active site for binding of the dye molecules via tunability of the morphological properties of the NPs semiconductors used. Although the performance and the integrity of these nanomaterials are a function of the pathways of composite fabrication which is most efficient by the in-situ process. This process involves the incorporation of powered catalyst NPs into an immobilized conducting polymer matrix as stated by this review. However, factors such as irradiation intensity and source, contact time for adsorption-desorption equilibrium before photons are incident on the system, pH, temperature and initial dye concentration also play a vital role in the effective performance of the process. The review also establishes the fact that among the varieties of semi-conductors used carbon-based nanomaterials stands due to their low toxicity, higher adsorption capacity and the excellent synergic effect when coupled with polyaniline to form composites. Hence, continued research should focus on the fabrication of nanocomposite using PANI coupled with carbon-based-nano materials, especially the multi-walled carbon nanotube (MWCNT), having unique morphological features fit for the adsorption and photodegradation process. Additionally, it is imperative to investigate more the possibility of recycling and recuse of the spent catalyst and to quantitatively determine the point of saturation of the fabricated composites during their reuses. It is worthy of note to add that the efficient primary treatment process must be given to textile effluent laden with dyes to remove the extraneous materials to limit the challenges of the photocatalytic process. Furthermore, the challenges of quantitative determination of initial dye concentration before photocatalysis limits the use of this treatment process due to oversaturation of the nano-catalyst at extremely high concentration dye molecules. Hence there is a need to create a robust industrial system to monitor the dye concertation fed into the photocatalytic reactor for effective use. Also, a future outlook on a hyphenated system of biological-photocatalytic techniques should be developed where biological microorganism such as laccase microbial enzymes associated with faster dye decolorizing speed is primarily used for the treatment of effluent before photocatalysis.The authors declare an absence of competing financial interests in personal relationships that could influence the work reported in this paperThis work was supported and funded by Regional Scholarship for Innovation Fund (RSIF) a flagship program of the Partnership for Skills in Applied Sciences, Engineering and Technology (PASET).", "descript": "\n                  Azo dyes in industrial textile and dye effluent (5\u201330%) have become irresistibly recalcitrant and toxic to both treatments and the environment respectively. Global concerns about the persistent nature of these dyes and the limitation of the conventional treatment currently in place have led to this critical analysis and evaluation of the photocatalytic approach using nano-technology. The review of literature has indicated that although this approach is effective, however, the limitation of frequent electron-hole recombination during the process coupled with challenges of agglomeration of nano particle powder, photo-corrosion and photosensitivity of the various nano-materials are still challenges associated with the development of polymeric based nano composite catalyst of polyaniline (PANI). The unique features of incredible charge transport properties, surface morphology and enhanced functional properties gave PANI the choice of use among other conductive polymers for composite fabrication with materials such\n                        \n                           \n                           T\n                           i\n                           \n                              O\n                              2\n                           \n                           ,\n                           \n                           a\n                           n\n                           d\n                           \n                           Z\n                           n\n                           \n                              O\n                              2\n                           \n                           ,\n                           \n                           G\n                           r\n                           a\n                           p\n                           h\n                           e\n                           n\n                           e\n                           \n                           o\n                           x\n                           i\n                           d\n                           e\n                           s\n                           ,\n                           \n                           C\n                           N\n                           T\n                           s\n                        \n                     . Photoactive properties, conductivity mechanical, thermal and chemical stability equally offers the polymer the propensity of bandgap tunability when in composites with other materials. Consequently, effective recovery and reuse of the composite catalyst for more than four runs with efficiency > 90% becomes obtainable. These appreciable advantages offer fabricated nano composite polymeric-based catalysts an effective outlook of use in the remediation of toxic azo dyes industrially as compared to the bio-catalyst and pure nano adsorbent materials. Therefore, the review discusses the treatment process for azo dyes, fabrication and performance evaluation of improved composite catalyst of PANI as an alternative to the conventional catalyst in wastewater and recommends for further investigation in PANI to enhance treatability of azo dyes.\n               "}
{"full_text": "The oxidation of carbon monoxide (CO) is a key reaction in automotive and industrial pollution control (Wu\u00a0et\u00a0al., 2001). Due to its relative simplicity, it has been the subject of numerous fundamental studies into heterogeneous catalytic processes (Szanyi\u00a0et\u00a0al., 1994). The reaction is traditionally carried out over noble metals such as platinum, palladium and rhodium. These materials exhibit high activity and selectivity toward total oxidation of CO as well as other volatile organic compounds (VOCs). However, high cost and limited availability of these materials have discouraged their extensive application in larger scale processes (Agula\u00a0et\u00a0al., 2011). Transition metal oxides function as suitable alternatives to the noble metals for oxidation of CO and are particularly attractive due to their relatively lower cost and good stability under cycled conditions (Yu-Yao\u00a0and Kummer,\u00a01977). Supported oxides of cobalt, iron, copper, manganese and nickel have all shown good performance in CO oxidation (Gravelle\u00a0and Teichner,\u00a01969). Nickel-based catalysts have been employed in numerous experimental studies of supported and single crystal catalytic processes, including mixed oxide catalytic systems (Parravano,\u00a01953; Choi\u00a0and Kim,\u00a01974; Royer\u00a0and Duprez,\u00a02011; Agula\u00a0et\u00a0al., 2011). Having excellent activity also for reforming and catalytic partial oxidation of methane, nickel-based catalysts are considered promising candidates for processes involving a combination of these basic C1 transformations (Raju\u00a0et\u00a0al., 2009).The aim of the present work was to investigate the kinetics of heterogeneous CO oxidation over a commercial alumina supported nickel oxide (NiO/Al2O3) catalyst (S\u00fcd-Chemie) and over a temperature range between 180 and 210\u00a0\u00b0C. The influences of the concentrations of CO and oxygen on the reaction rate were studied to derive a rate expression that can be used over the specific range of operating conditions for large scale reactor design.Consensus has not yet been reached in literature with respect to the mechanism by which the catalytic oxidation of CO takes place. Therefore, prior to kinetic model development, it was deemed necessary to identify potential reactions mechanisms describing the kinetics over the specified operating regime. This section outlines the mechanisms proposed.Research conducted by Yu-Yao\u00a0and Kummer\u00a0(1977) proposed that CO and oxygen are initially chemisorbed on catalytic centres before the surface reaction, after which carbon dioxide is released in two successive steps. These researchers proposed that catalytic oxidation of CO over nickel oxide could be represented by a Langmuir-Hinshelwood (L-H) type reaction mechanism with competitive adsorption of oxygen and CO onto active catalyst surface sites (see Fig.\u00a01\n below).However, earlier work undertaken by Dell\u00a0and Stone\u00a0(1954) to investigate the adsorption of gasses such as CO on thin nickel oxide films concluded that at low temperatures \n\n(\n\n\n20\n\u2218\n\nC\n\n)\n\n, the amount of CO which could be adsorbed was very much less than the amount of oxygen which could be adsorbed at corresponding conditions. Similarly, Gravelle\u00a0and Teichner\u00a0(1969) noted that the adsorption of CO decreases with temperature and Gandhi\u00a0and Shelef\u00a0(1972) found CO adsorption negligible above 140 \u00b0C. As a result, an Eley-Rideal (E-R) type reaction mechanism could also be proposed. In this proposal, CO directly attacks a chemisorbed oxygen complex before carbon dioxide release (see Fig.\u00a02\n).In more recent research conducted by Conner\u00a0and Bennet\u00a0(1976), it was proposed that there are three types of site onto which oxygen can adsorb. Two sites participate in the CO oxidation reaction at 180\n\n\n\n\u2218\n\nC\n\n. The first type of active catalyst surface site \n\n(\n\nS\n1\n\n)\n\n is hypothesised to be a nickel atom onto which removable oxygen atoms are chemisorbed. Carbon dioxide generated during the reaction then exclusively adsorbs onto the second type of site \n\n(\n\nS\n2\n\n)\n\n which strongly bond to oxygen atoms such as found in a lattice, and this oxygen is proposed to be in the O2\u2212 state. In addition, experimental results concluded that no CO adsorbs and that the presence of such would be due only to reaction.Based on these observations, Conner\u00a0and Bennet\u00a0(1976) compiled a probable sequence of reaction steps (see Fig.\u00a03\n) describing the CO oxidation reaction pathway over nickel oxide. This proposal provides a plausible alternative to the reaction schemes listed above.This section will present the formulation of initial reaction rate expressions based on the different reaction mechanisms presented above. Given that isotope or other spectroscopy measurements were not possible within the experimental system, possible rate-controlling steps within each mechanism could not be identified a-priori. Therefore, in the case of the proposed l-H and E-R mechanisms, different rate expressions were developed based on the assumption that each possible forward reaction step can be rate-controlling. Given that the conversion of CO in the reaction system was limited to <10%, it was assumed that the reverse reactions do no occur and thus are not possible rate-controlling steps. In the case of the alternative reaction mechanism put forward by Conner\u00a0and Bennett\u00a0(1976), isotope studies were conducted by the authors to ascertain the possible rate-controlling step. Therefore, the findings of these studies were used to develop a reaction rate expression from the proposed mechanism.There are four steps in this sequence viz. the adsorption steps for oxygen and CO, a reaction step between the adsorbed species and a product desorption step. In this development, focus on initial rate expressions where reactions involving products are considered negligible in their rate.For the case of the adsorption of CO being rate-controlling viz. forward step (1a), the net rate can be described as that shown in Eq.\u00a0(1).\n\n(1)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n=\n\nk\nCO\n\n\nP\nCO\n\n\nC\nV\n\n\n\n\n\u2026where \n\nC\nV\n\n is the total number of vacant active sites on the catalyst surface.Applying the Pseudo-Steady State Approximation (PSSA),\n\n(2)\n\n\n\n\nd\n\n[\n\nCO\n\n.\nS\n\n\n]\n\n\ndt\n\n\u2245\n\nk\nCO\n\n\nP\nCO\n\n\nC\nV\n\n\u2212\n\n\n\n\nk\n\n\u2032\n\n\nCO\n\n\n[\n\nCO\n\n.\nS\n\n\n]\n\n\u2212\n\nk\nS\n\n\n[\n\nC\nO\n.\nS\n\n]\n\n\n[\n\nO\n.\nS\n\n]\n\n\u2245\n0\n\n\n\n\n\n\n\n\n\u2234\n\n[\n\nCO\n\n.\nS\n\n\n]\n\n\u2245\n\n\n\n(\n\n\nk\n\nC\nO\n\n\n\nk\nS\n\n\n)\n\n\nP\n\nC\nO\n\n\n\nC\nv\n\n\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\n\n+\n\n\n\nk\n\nC\nO\n\n\u2032\n\n\nk\nS\n\n\n\n\n\n\n\nAs forward step (1a) is rate-controlling, it can be assumed that: \n\n\nk\nCO\n\n\n<\n<\n\n\nk\nS\n\n\n and \n\n(\n\n\nk\nCO\n\n\nk\nS\n\n\n)\n\u2245\n0\n\n\n\n\n\n\n\u2234\n\n[\n\nC\nO\n.\nS\n\n]\n\n\u2245\n0\n\n\n\nSimilarly,\n\n(3)\n\n\n\n\nd\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\ndt\n\n\u2245\n\nk\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\nC\nV\n2\n\n\u2212\n\n\n\n\nk\n\n\u2032\n\n\n\nO\n2\n\n\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n2\n\n\u2212\n\nk\nS\n\n\n[\n\nC\nO\n.\nS\n\n]\n\n\n[\n\nO\n.\nS\n\n]\n\n\u2245\n0\n\n\n\nBut considering that \n\n[\n\nCO\n\n.\nS\n\n\n]\n\u2245\n0\n\n, it can be assumed that \n\n\nk\nS\n\n\n[\n\nC\nO\n.\nS\n\n]\n\n\n[\n\nO\n.\nS\n\n]\n\n\u2245\n0\n\n\n\n\n\n\n\u2234\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\u2245\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n0\n\n.\n5\n\n\n\n\nC\nV\n\n\n\n\nThe concentration of vacant sites can be calculated as follows:\n\n(4)\n\n\n\nC\nV\n\n=\n\nC\nT\n\n\u2212\n\n[\n\nCO\n\n.\nS\n\n\n]\n\n\u2212\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\n\n\nwhere \n\nC\nT\n\n represents the total number of active sites on the catalyst surface.Then, \n\n\nC\nV\n\n=\n\nC\n\nT\n\n\n\n\u2212\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n0\n\n.\n5\n\n\n\n\nC\nV\n\n\n\n\n\n\n\n\u2234\n\nC\nV\n\n=\n\n\n\n\nC\nT\n\n\n\n\n1\n+\n\n\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n\n0\n\n.\n5\n\n\n\n\n\n\n\n\nFinally, the rate expression could be obtained in Eq.\u00a0(1) as:\n\n(5)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n=\n\n\n\nk\nCO\n\n\nP\nCO\n\n\nC\nT\n\n\n\n\n1\n\n+\n\n\n\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n\n0\n\n.\n5\n\n\n\n\n\n\n\n\nIn a similar way, in the case of the adsorption of oxygen being rate-controlling viz. forward step (2a), the net reaction rate can be expressed as:\n\n(6)\n\n\n\u2212\n\nR\n\nC\nO\n,\n0\n\n\n=\n\nk\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n\n\nC\nV\n\n\n2\n\n\n\n\nIt was shown that in the case where the adsorption of CO is rate-controlling, the concentration of CO adsorbed onto active catalyst sites is negligible \n\n(\n\n[\n\nC\nO\n.\nS\n\n]\n\u2245\n0\n\n)\n\n. This makes sense considering that the adsorption rate of CO onto catalyst sites was proposed to be the slowest reaction step. As a result, in the case where the adsorption of oxygen is the rate-controlling step, it is assumed that the concentration of oxygen adsorbed onto catalyst sites is also negligible \n\n(\n\n[\n\nO\n.\nS\n\n]\n\u2245\n0\n\n)\n\n.Then,\n\n\n\n\n[\n\nCO\n\n.\nS\n\n\n]\n\n\u2245\n\nK\nCO\n\n\nP\nCO\n\n\nC\nV\n\n\n\n\n\n\n\n\n\n\nC\nV\n\n=\n\n\nC\nT\n\n\n\n1\n+\n\n\nK\nCO\n\n\nP\nCO\n\n\n\n\n\n\nAs a result, the reaction rate expression in\tEq.\u00a0(6) can be reduced to:\n\n(7)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n\n=\n\n\n\nk\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n\n\nC\nT\n\n\n2\n\n\n\n\n\n(\n\n\n1\n\n+\n\n\n\nK\nCO\n\n\nP\nCO\n\n\n)\n\n\n2\n\n\n\n\n\n\nIn the case for the reaction between the adsorbed reactant species being rate controlling viz. forward step (3a), then the net reaction rate can be expressed as follows:\n\n(8)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n\n=\n\nk\nS\n\n\n[\n\nC\nO\n.\nS\n\n]\n\n\n[\n\nO\n.\nS\n\n]\n\n\n\n\nHere,\n\n\n\n\n[\n\nCO\n\n.\nS\n\n\n]\n\n\u2245\n\n\n\nK\nCO\n\n\nP\nCO\n\n\nC\nV\n\n\n\n\n\nk\nS\n\n\nk\n\nC\nO\n\n\u2032\n\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\n+\n1\n\n\n\n\n\n\nAs the forward step (3a) is rate-controlling, it can be assumed that \n\n\nk\n\nS\n\n\n\n\n\n<\n<\nk\n\n\nC\nO\n\n\u2032\n\n\n such that \n\n\n\nk\nS\n\n\nk\n\nC\nO\n\n\u2032\n\n\n\u2245\n0\n\n\n\n\n\n\n\u2234\n\n[\n\nCO\n\n.\nS\n\n\n]\n\n\u2245\n\nK\nCO\n\n\nP\nCO\n\n\nC\nV\n\n\n\n\nIn addition,\n\n\n\n\n\nd\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\ndt\n\n\u2245\n\nk\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\nC\nV\n2\n\n\u2212\n\n\n\n\nk\n\n\u2032\n\n\n\nO\n2\n\n\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n2\n\n\u2212\n\nk\nS\n\n\n[\n\nC\nO\n.\nS\n\n]\n\n\n[\n\nO\n.\nS\n\n]\n\n\u2245\n0\n\n\n\n\n\n\n\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\nC\nV\n2\n\n\u2212\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n2\n\n\u2212\n\n\nk\nS\n\n\nk\n\n\nO\n2\n\n\n\u2032\n\n\n\n[\n\nCO\n\n.\nS\n\n\n]\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\u2245\n0\n\n\n\nAs the forward step (3a) is rate-controlling, it can be assumed that \n\n\nk\n\nS\n\n\n\n\n\n<\n<\nk\n\n\n\nO\n2\n\n\n\u2032\n\n\n such that \n\n\n\nk\nS\n\n\nk\n\n\nO\n2\n\n\n\u2032\n\n\n\u2245\n0\n\n\n\n\n\n\n\u2234\n\n[\n\nO\n.\nS\n\n]\n\n\u2245\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n0.5\n\n\n\nC\nV\n\n\n\n\nThen,\n\n\n\n\nC\n\nV\n\n\n\n=\n\n\nC\nT\n\n\n\n1\n+\n\n\nK\nCO\n\n\nP\nCO\n\n+\n\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n\n0\n\n.\n5\n\n\n\n\n\n\n\n\nTherefore, the reaction rate expression in Eq.\u00a0(8) can be reduced to:\n\n(9)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n=\n\n\n\nK\nS\n\n\nK\nCO\n\n\nP\nCO\n\n\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n\n0.5\n\n\n\nC\nT\n2\n\n\n\n\n\n(\n\n\n1\n+\n\n\nK\n\nC\nO\n\n\n\nP\n\nC\nO\n\n\n+\n\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n\n0\n\n.\n5\n\n\n\n\n)\n\n\n2\n\n\n\n\n\n\nIn the case where the adsorption of a reactant species viz. CO or oxygen, is rate-controlling, the driving force for the reaction depends only on the concentration of that reactant species (see Eq.\u00a0(5) and Eq.\u00a0(7) above). In the case where the surface reaction viz. CO oxidation, is rate-controlling, the driving force for the reaction depends on all reactant concentrations (see Eq.\u00a0(9) above).This mechanism differs from reaction mechanism (RM) 1 in that adsorbed oxygen is proposed to only react with CO from the bulk gas to generate carbon dioxide. The modified step is represented by step (2b) above and if this step is assumed to be rate-controlling then:\n\n(10)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n=\n\nk\ns\n\n\nP\nco\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\n\n\nThen\n\n(11)\n\n\n\n\nd\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\ndt\n\n\u2245\n\nk\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\nC\nV\n2\n\n\u2212\n\n\n\n\nk\n\n\u2032\n\n\n\nO\n2\n\n\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n2\n\n\u2212\n\nk\nS\n\n\nP\n\nC\nO\n\n\n\n[\n\nO\n.\nS\n\n]\n\n\u2245\n0\n\n\n\n\n\n\n\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\nC\nV\n2\n\n\u2212\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n2\n\n\u2212\n\n\nk\nS\n\n\nk\n\n\nO\n2\n\n\n\u2032\n\n\n\nP\n\nC\nO\n\n\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\u2245\n0\n\n\n\nAs the forward step (2b) is rate-controlling, it can be assumed that \n\n\nk\n\nS\n\n\n\n\n\n<\n<\nk\n\n\n\nO\n2\n\n\n\u2032\n\n\n such that \n\n\n\nk\nS\n\n\nk\n\n\nO\n2\n\n\n\u2032\n\n\n\u2245\n0\n\n\n\n\n\n\n\u2234\n\n[\n\nO\n.\nS\n\n]\n\n\u2245\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n0.5\n\n\n\nC\nV\n\n\n\n\nHere,\n\n(12)\n\n\n\nC\nV\n\n=\n\nC\nT\n\n\u2212\n\n[\n\nO\n\n.\nS\n\n\n]\n\n\n\n\n\n\n\n\n\n\u2234\n\nC\n\nV\n\n\n\n=\n\n\nC\nT\n\n\n\n1\n+\n\n\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n\n0\n\n.\n5\n\n\n\n\n\n\n\n\nTherefore, reducing the rate expression presented in Eq.\u00a0(10) gives:\n\n(13)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n=\n\n\n\nk\nS\n\n\nP\nCO\n\n\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n\n0.5\n\n\n\nC\nT\n\n\n\n\n1\n+\n\n\n\n\n(\n\n\nK\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\n)\n\n\n\n0\n\n.\n5\n\n\n\n\n\n\n\n\nIn contrast to RM 1, this rate expression describes a linearly dependence on the partial pressure of CO.To obtain the rate expression formulation, the following propositions regarding the probable reaction scheme for CO oxidation put forward by Conner\u00a0and Bennett\u00a0(1976) were adopted:\n\n-\nAt 180 \u00b0C, the predominant surface species on \n\nS\n1\n\n sites would be the \n\nO\n.\n\nS\n1\n\n\n surface intermediate. This was assumed to be the most abundant reaction intermediation on \n\nS\n1\n\n sites.\n\n\n-\nAt 180 \u00b0C, step (3c) is pushed towards the production of \n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n such that the concentration of \n\nC\n\nO\n2\n\nO\n.\n\nS\n1\n\n\n is very low. Therefore, it is assumed that the reverse reaction of step (3c) is unlikely to occur.\n\n\n-\nIsotope studies conducted at 180 \u00b0C suggest that although carbon dioxide can be generated via step (4c), bulk of the carbon dioxide would be generated from subsequent steps (5c) and (6c). Therefore, step (4c) will be neglected in the formulation.\n\n\n-\nThe two-carbon surface intermediate \n\n\nC\n2\n\n\nO\n3\n\nO\n.\n\nS\n2\n\n\n has low concentration such that \n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n is found to be the most abundant surface intermediate on \n\nS\n2\n\n sites. Therefore, the forward reaction of step (5c), which describes the consumption of \n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n species to produce \n\n\nC\n2\n\n\nO\n3\n\nO\n.\n\nS\n2\n\n\n, is assumed to be the rate-controlling step.\n\nApplying these assumptions,\n\n(14)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n=\n\nk\n4\n\n\n[\n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n]\n\n\nP\n\nC\nO\n\n\n\n\n\nFurther applying the PSSA to determine \n\n[\n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n]\n\n:\n\n(15)\n\n\n\n\nd\n\n[\n\nC\n\nO\n2\n\nO\n\n.\nS\n\n\n]\n\n\ndt\n\n\u2245\n\nk\n4\n\n\n[\n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n]\n\n+\n\nk\n2\n\n\n[\n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n]\n\n\n[\n\nO\n.\n\nS\n2\n\n\n]\n\n\u2245\n0\n\n\n\n\n\n\n(16)\n\n\n\u2234\n\n[\n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n]\n\n\u2245\n\n\n\nk\n2\n\n\n[\n\nC\n\nO\n2\n\n.\n\nS\n1\n\n\n]\n\n\n[\n\nO\n.\n\nS\n2\n\n\n]\n\n\n\n\nk\n4\n\n\nP\nCO\n\n\n\n\n\n\n\n\n\n\n(17)\n\n\n\n\nd\n\n[\n\nC\n\nO\n2\n\n.\n\nS\n1\n\n\n]\n\n\ndt\n\n\u2245\n\u2212\n\nk\n2\n\n\n[\n\nC\n\nO\n2\n\n.\n\nS\n1\n\n\n]\n\n\n[\n\nO\n.\n\nS\n2\n\n\n]\n\n+\n\nk\n1\n\n\n[\n\nO\n\n.\n\n\n\nS\n1\n\n\n]\n\n\nP\n\nCO\n\n\n\n\u2245\n0\n\n\n\n\n\n\n(18)\n\n\n\u2234\n\n[\n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n]\n\n\u2245\n\n\n\nk\n1\n\n\n[\n\nO\n.\n\nS\n1\n\n\n]\n\n\nP\n\nC\nO\n\n\n\n\n\nk\n2\n\n\n[\n\nO\n.\n\nS\n2\n\n\n]\n\n\n\n\n\n\n\nSubstituting Eq.\u00a0(18) into Eq.\u00a0(16),\n\n(19)\n\n\n\n[\n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n]\n\n\u2245\n\n\nk\n1\n\n\nk\n4\n\n\n\n[\n\nO\n\n.\n\n\n\nS\n1\n\n\n]\n\n\n\n\n\n\n\n(20)\n\n\n\n\nd\n\n[\n\nO\n.\n\nS\n1\n\n\n]\n\n\ndt\n\n\u2245\n\nk\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\nC\n\n\nV\n\nS\n1\n\n\n\n2\n\n\u2212\n\nk\n1\n\n\n[\n\nO\n\n.\n\n\n\nS\n1\n\n\n]\n\n\nP\n\nCO\n\n\n\n\u2245\n0\n\n\n\nWhere \n\nC\n\nV\n\nS\n1\n\n\n\n represents the concentration of vacant \n\nS\n1\n\n sites on the catalyst surface.\n\n(21)\n\n\n\u2234\n\n[\n\nO\n\n.\n\n\n\nS\n1\n\n\n]\n\n\u2245\n\n\n\nk\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\nC\n\n\nV\n\nS\n1\n\n\n\n2\n\n\n\n\nk\n1\n\n\nP\nCO\n\n\n\n\n\n\nHere,\n\n(22)\n\n\n\nC\n\nV\n\nS\n1\n\n\n\n=\n\nC\n\nT\n\nS\n1\n\n\n\n\u2212\n\n[\n\nO\n\n.\n\n\n\nS\n1\n\n\n]\n\n\n\n\nWhere \n\nC\n\nT\n\nS\n1\n\n\n\n represents the total number of active \n\u03b1\n sites on the catalyst surface.At 180 \u00b0C, the predominant surface species on \n\nS\n1\n\n sites would be the \n\nO\n.\n\nS\n1\n\n\n surface intermediate. This was assumed to be the most abundant reaction intermediation on \n\nS\n1\n\n sites.At 180 \u00b0C, step (3c) is pushed towards the production of \n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n such that the concentration of \n\nC\n\nO\n2\n\nO\n.\n\nS\n1\n\n\n is very low. Therefore, it is assumed that the reverse reaction of step (3c) is unlikely to occur.Isotope studies conducted at 180 \u00b0C suggest that although carbon dioxide can be generated via step (4c), bulk of the carbon dioxide would be generated from subsequent steps (5c) and (6c). Therefore, step (4c) will be neglected in the formulation.The two-carbon surface intermediate \n\n\nC\n2\n\n\nO\n3\n\nO\n.\n\nS\n2\n\n\n has low concentration such that \n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n is found to be the most abundant surface intermediate on \n\nS\n2\n\n sites. Therefore, the forward reaction of step (5c), which describes the consumption of \n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n species to produce \n\n\nC\n2\n\n\nO\n3\n\nO\n.\n\nS\n2\n\n\n, is assumed to be the rate-controlling step.There are three expressions. viz. Eq.\u00a0(19), (21) and (22) and three unknown variables viz. \n\n[\n\nC\n\nO\n2\n\nO\n.\n\nS\n2\n\n\n]\n\n, \n\n[\n\nO\n\n.\n\n\n\nS\n1\n\n\n]\n\n and \n\nC\n\nV\n\nS\n1\n\n\n\n\n. Solving simultaneously gives:\n\n(23)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n=\n\n\n\nk\n\nO\n2\n\n\n\nk\n1\n\n\nP\nCO\n\n\nP\n\nO\n2\n\n\n\nC\n\n\nT\n\nS\n1\n\n\n\n2\n\n\n\n2\n\nk\n\nO\n2\n\n\n\nP\n\nO\n2\n\n\n\nC\n\nT\n\nS\n1\n\n\n\n+\n\nk\n1\n\n\nP\nCO\n\n\n\n\n\n\nThis rate expression is fundamentally different from those obtained from RMs 1 and 2 in that the reaction rate is a function that is linearly dependent on both the partial pressures of oxygen and CO. In addition, this rate expression does not contain adsorption equilibrium constants. A summary of the rate expressions obtained is presented in Table\u00a01\n below.Given that the list of probable kinetic models describing the oxidation of CO over nickel oxide has been outlined (see Table\u00a01), the model parameters can be estimated by fitting these models to experimental data. The experimental procedure used to perform kinetic measurements is outlined in the section below.A commercial G-65 steam reforming catalyst (S\u00fcd-Chemie) was used for all experiments. The catalyst is also a common methanation catalyst. Van\u00a0Herwijnen et\u00a0al.\u00a0(1973) carried out X-ray diffraction measurements on a sample of the catalyst and deduced that the carrier material is \u03b3-alumina. No further information was elucidated. Since the catalyst was used as supplied by the manufacturer, for the purposes of this study only the surface and morphological properties were measured. The textural properties of the catalyst (BET surface area and average pore size) were determined using a Micrometrics ASAP 2020 gas adsorption analyser. Measurements were performed with nitrogen as the adsorbate at \u2212196\u00a0\u00b0C. The samples were degassed under nitrogen at 200\u00a0\u00b0C for 18\u00a0h prior to analysis. The catalyst was found to have a specific surface area of 42.4 m2\ng\u00a0\u2212\u00a01 with an average pore diameter of 100 \n\nA\n\u02d9\n\n. Energy dispersive X-ray (EDX) analysis was performed using a ZEISS Ultra Plus scanning electron microscope (SEM) instrument, and the results were used to determine the amount of active material present on the surface. The catalyst was found to contain 33.6\u00a0wt.% nickel. The commercial pellets were ground and sieved to recover a fraction in the size range of 250\u2013350\u00a0\u00b5m, which were used for all subsequent kinetic tests.Experiments were carried out in a conventional laboratory-scale gas-phase fixed bed reactor (as shown in Fig.\u00a04\n) at atmospheric pressure. The reactor was fabricated from a \u00bd inch nominal sized stainless steel tube (i.d. 9.2\u00a0mm, length 400\u00a0mm) and was placed in a vertical, electrically heated tube furnace. The catalyst was placed on a stainless steel grid attached to a sheathed type K thermocouple mounted through the center of the tube. CO (Afrox, 99.9%), oxygen and nitrogen (Afrox, 99.999%) were supplied in standard cylinders and metered using precision flow control valves. The product stream was sampled and analysed using gas-solid chromatography and the flowrate was measured using a conventional gas flowmeter.In each experiment, a catalyst sample of 1\u00a0g was loaded into the reactor tube. The catalyst was pretreated in situ by heating at 500\u00a0\u00b0C under a stream of nitrogen for two hours, then cooled to reaction temperature. The feed gas was then opened to the reactor and the reaction was allowed to proceed until steady state was achieved (approximately 1\u00a0hour, determined from preliminary tests by monitoring the change in the exit composition of the product gas). Samples of the gas stream were analysed after this period of stabilization. Analyses of the product gas were performed on a Shimadzu 2014 gas chromatograph (GC) equipped with a packed Restek ShinCarbon ST column (80/100 mesh, 2\u00a0m long, 1/8 inch o.d.) and a thermal conductivity detector (TCD) using helium as the carrier gas (25 ml\u2022min\u22121). The temperature program for gas samples was 40\u00a0\u00b0C for 3\u00a0min, 40\u00a0\u00b0C to 220\u00a0\u00b0C (at 8\u00a0\u00b0C min\u22121) and 220\u00a0\u00b0C for 10\u00a0min.Kinetic data was collected at four different temperatures (180, 190, 200 and 210\u00a0\u00b0C) using feed mixtures of CO, oxygen and nitrogen. Each isothermal dataset consisted of 11 experiments. In the first six experiments, the feed concentration of oxygen was varied whilst maintaining the concentration of CO approximately constant. In the last five experiments, the feed concentration of CO was varied whilst maintaining the concentration of oxygen approximately constant. Nitrogen was used as the dilution gas, with the flowrate adjusted to maintain a constant space time in the reactor across all experiments. The total inlet molar flowrate across all experiments ranged between 2.42\u20132.65\u00a0\u00d7\u00a010\u22124\u00a0mol s\u00a0\u2212\u00a01. The partial pressure of O2 ranged between 7 and 20\u00a0kPa and that of CO ranged between 7 and 25\u00a0kPa, corresponding to inlet concentrations of 2\u20135\u00a0mol m\u00a0\u2212\u00a03 and 2\u20137\u00a0mol m\u00a0\u2212\u00a03 for O2 and CO, respectively.The conversion of CO was kept below 10% to ensure differential operation of the reactor in addition to the use of limited quantities of catalyst and high gas flow rates. The modified Reynolds number of the gas flow through the reactor was estimated to be \u223c100 which is close to fully turbulent flow through the packed bed (Chhabra\u00a0and Basavaraj,\u00a02019). Therefore, plug flow is assumed through the reactor and the initial rate could be calculated directly using Eq.\u00a0(24)\n\n\n(24)\n\n\n\u2212\n\nR\n\nCO\n,\n0\n\n\n=\n\n\n\nF\n\nCO\n,\nin\n\n\n\u2212\n\nF\n\nCO\n,\nout\n\n\n\n\nW\ncat\n\n\n\n\n\n\nThe Weisz-Prater criterion (CWP\n) was used to check whether the reaction is internal diffusion limited. This criterion was calculated using the following equation:\n\n(25)\n\n\n\nC\nWP\n\n=\n\n\n\n\n\u2212\n\n\nR\n\nCO\n,\n0\n\n\n\u00b7\n\n\u03c1\ns\n\n\u00b7\n\n\n\nR\np\n\n\n2\n\n\n\n\nD\ne\n\n\u00b7\n\nC\n\nAS\n,\n0\n\n\n\n\n\n\n\nThe effect of internal diffusion on the observed CO oxidation reaction rate is considered negligible when CWP\n is <<1. The parameters used in Eq.\u00a0(25) are presented in the table below.Using the values in Table\u00a02\n, the CWP\n value was estimated using Eq.\u00a0(25) as follows:\n\n\n\n\nC\nWP\n\n=\n\n\n\n\n(\n\n2\n\n.\n95\ne\n\u2212\n06\n\n\u00b7\n6\n\n.\n67\n\n\u00b7\n0\n\n.\n015\n\n\n)\n\n\n2\n\n\n(\n\n1\n\n.\n10\ne\n\u2212\n04\n\n\u00b7\n4\n\n.\n87\ne\n\u2212\n04\n\n\n)\n\n\n\u2245\n0.0021\n<\n<\n1\n\n\n\nTherefore, it can be assumed that internal diffusion within the supported NiO catalyst has negligible effect on the observed CO oxidation reaction rate.The catalyst internal effectiveness factor \n\n(\n\u03b7\n)\n\n also indicates the effects of the internal mass transfer and is used to evaluate the performance of a catalytic reactor. The effectiveness factor \n\n(\n\u03b7\n)\n\n is defined as the ratio of the observed rate of reaction to the hypothetical rate in the absence of mass transfer limitations. Therefore, as the effectiveness factor approaches 1, the effects of internal mass transfer becomes more negligible.The effectiveness factor \n\n(\n\u03b7\n)\n\n was calculated using the initial reaction rate constant (kCO,0\n). Assuming that at the start of the reaction (t=\u00a00) the surface CO concentration is the same as the concentration across the catalyst particle i.e. CAS,\n\n0\u00a0=\u00a0CA,\n\n0, the initial reaction rate constant can be estimated by using the measured initial reaction rate value and initial CO concentration at the surface i.e. -RCO,\n\n0= kCO,\n\n0\nCAS,\n\n0.For a first-order reaction and assuming spherical catalyst particles, the internal effective \n\n(\n\u03b7\n)\n\n can be calculated using the Thiele modulus \n\n(\n\u03d5\n)\n\n, which correlates the catalyst activity \n\n(\n\nk\n\nc\n0\n,\n0\n\n\n)\n\n and radius of the catalyst particle (R) as follows:\n\n(26)\n\n\n\u03d5\n=\nR\n\n\n\nk\n\nc\n0\n,\n0\n\n\n\nD\ne\n\n\n\n\n\n\n\n\n\n(27)\n\n\n\u03b7\n=\n\n3\n\n\n\u03d5\n\n2\n\n\n\n(\n\n\u03d5\ncoth\n\u03d5\n\u2212\n1\n\n)\n\n\n\n\nBased on the Eqns.\u00a0(26) and 27, the Thiele modulus and internal effectiveness factor was estimated as follows:\n\n\n\n\u03d5\n=\n\n(\n\n0.015\n\n)\n\n\n\n\n6.06\ne\n\u2212\n03\n\n\n1.1\ne\n\u2212\n04\n\n\n\n=\n0.0081\n\n\n\n\n\n\n\n\n\u03b7\n=\n\n3\n\n\n\n(\n0\n\n\n.\n0081\n)\n\n\n2\n\n\n\n(\n\n0\n\n.\n0081\n\n\u00b7\ncoth\n\n(\n\n0\n\n.\n0081\n\n\n)\n\n\u2212\n1\n\n)\n\n\u2245\n1\n\n\n\nGiven that the catalyst internal effectiveness factor \n\n\u2245\n1\n\n, the internal mass transfer effects are considered negligible within the catalytic reactor.The Mears criterion is used to check whether the reaction is limited by external diffusion. This criterion is calculated using the following equation:\n\n(28)\n\n\n\nC\nmears\n\n=\n\n\n\n\n\u2212\n\n\nR\n\nCO\n,\n0\n\n\n\n\u00b7\n\n\u03c1\nb\n\n\u00b7\n\n\nR\np\n\n\u00b7\nn\n\n\n\nk\nc\n\n\u00b7\n\nC\n\nAS\n,\n0\n\n\n\n\n\n\n\nThe effect of external diffusion on the observed CO oxidation reaction rate is considered negligible when Cmears\n \u00a0<\u00a00.15.Here, the external mass transfer coefficient (kc\n) for CO in the bulk diffusing through air is calculated by:\n\n(29)\n\n\n\nk\nc\n\n=\n\n\nSh\n\u00b7\n\nD\n\nA\nB\n\n\n\n\nd\np\n\n\n\n\n\nHere, Sh is the Sherwood number and is calculated as:\n\n(30)\n\n\n\nSh\n=\n2\n+\n0\n\n\n.\n6\nR\n\n\n\ne\n\n\n0\n\n.\n5\n\n\n\n\n\n\nS\nc\n\n\n\n1\n/\n3\n\n\n\n\n\nwhere Sc\n and Re are the Schmidt and Reynolds numbers for CO in the bulk diffusing through air, respectively. The Sherwood and Reynolds numbers are found as follows:\n\n(31)\n\n\n\nS\nc\n\n\n\n=\n\n\n\nV\n\nD\nAB\n\n\n\n\n\n\n\n\n(32)\n\n\n\nRe\n\n=\n\n\n\n\u03c1\n\u00b7\n\n\nD\n\n2\n\n\u00b7\nu\n\n\u03bc\n\n\n\n\n\nThe Mears criterion was then calculated as follows:\n\n\n\n\nC\nmears\n\n\n\n=\n\n\n\n\n\n(\n\n2\n\n.\n95\ne\n\u2212\n06\n\n\n)\n\n\u00b7\n\n(\n\n25\n\n.\n1\n\n\n)\n\n\u00b7\n\n(\n\n0\n\n.\n15\n\n\n)\n\n\u00b7\n\n(\n1\n)\n\n\n\n\n(\n\n92\n\n.\n59\n\n\n)\n\n\u00b7\n\n(\n\n4\n\n.\n87\ne\n\u2212\n04\n\n\n)\n\n\n\n\n\u2245\n2.46\ne\n\u2212\n05\n<\n<\n0.15\n\n\n\nTherefore, it can be assumed that external diffusion within the bulk CO-air layer has negligible effect on the observed CO oxidation reaction rate at the catalyst surface.The average relative deviation on TCD calibration for CO and oxygen was 2.3%. Relative uncertainty in the gas flowrate measurements averaged 3%, based on the manufacturers\u2019 specifications for the precision rotameters and the laboratory calibration of the rotameters. The catalyst mass was determined using a digital scale having an accuracy of 0.1\u00a0mg and hence the uncertainty in measurement was considered negligible. Consequently, the average relative uncertainty on observed reaction rates was found to be low (3.7%) and was not accounted for in the rate measurements presented in the results section below.Kinetic measurements were performed to determine the influence of the partial pressure of CO and oxygen on the reaction rate with increase in temperature from 180 to 210 \u00b0C. These kinetic rate measurements allow for the estimation of model parameters contained within the proposed reaction rate expressions. In addition, assessing the influence of these parameters on the reaction rate can allow for preliminary discrimination between the reaction rate expressions prior to the estimation of model parameters. Plots of the relationship between the reaction rate and the partial pressure of oxygen and the partial pressure of CO are shown in Figs. 5 and 6\n\n below, respectively.The experimental data presented in Figs.\u00a05 and 6 both reveal that an increase in the partial pressure of oxygen and CO respectively results in an increase in the reaction rate. As the temperature increases, these observed trends become more pronounced with increase in the reactant partial pressures. As a result, the overall reaction rate must be driven by both the partial pressures of oxygen and CO. This is not represented by reaction rate expressions (1) and (2). As a result, these expressions could be ruled out from the list of possible reaction rate expressions. Given that rate expressions (3), (4) and (5) were still potential kinetic models for the system, further model discrimination was made by measuring the degree of fit between these rate expressions and the experimental data measured at 180 \u00b0C. The experimental data was regressed to the rate expressions using the Levenberg-Marquardt algorithm based on MATLAB software.The graphical representations of the model fits regressed to the experimental data measured at 180 \u00b0C and non-linear regression outputs from MATLAB are presented in Fig.\u00a07\n and Table\u00a04, respectively. Analysis of Fig.\u00a07, as well as, the coefficient of determination (R\n2) values presented in Table\u00a04 reveals that the experimental data measured at 180 \u00b0C best fits rate expression (5).The model variances \n\n(\n\n\u03c3\n2\n\n)\n\n presented in Table\u00a04 measures the magnitude of deviation between the experimental data measured at 180 \u00b0C and the model outputs. These variances reflect on the R\n2 values and can be used to conduct a statistical F-test to assess whether the fit between the experimental data measured at 180 \u00b0C and rate expression (5) differs significantly from the model fits made with rate expressions (3) and (4). The F-test value represents the ratio of the model variances \n\n(\n\n\n\u03c3\n1\n2\n\n\n\u03c3\n2\n2\n\n\n)\n\n where \n\n\n\u03c3\n1\n2\n\n>\n\n\u03c3\n2\n2\n\n\n. For the null hypothesis viz. that there is significant statistical difference between the model fits to be accepted, the F-test value must be greater than the tabulated critical F-test value. At the 95% confidence interval, the tabulated critical F-test values for rate expression (3) and (4) with regards to rate expression (5) are 3.44 and 3.39 respectively. The calculated F-test values for rate expression (3) and (4) with regards to rate expression (5) are 1.74 and 1.25 respectively. Therefore, the alternate hypothesis is accepted in both cases viz. that there is no significant statistical difference between the closer model fit made with rate expression (5) and those made with rate expressions (3) and (4).However, given the better fit to the experimental data measured at 180 \u00b0C, rate expression (5) was selected from the list of probable rate expressions as the most reliable description of the kinetics of CO oxidation within the specified operating region. It is important to note that the lack of agreement between the experimental data measured at 180 \u00b0C and rate expression (5) can be accounted to possible factors such as mass and thermal diffusion resistances, side reactions and deviation from the plug flow reactor (PFR) model.\nFigs.\u00a05 and 6 also both reveal that the degree of influence of the partial pressure of oxygen and CO respectively on the reaction rate increases with temperature. Given that the reaction rate is strongly influenced by temperature, rate expression (5) should also be a function of the reaction temperature. The model parameters contained with rate expression (5) can be related to the reaction temperature through the Arrhenius expression as follows:\n\n(33)\n\n\nk\n=\nA\nexp\n\n(\n\n\u2212\n\n\nE\na\n\n\n\nR\ng\n\nT\n\n\n\n)\n\n\n\n\nIn this form, values for the pre-exponential factor (A) and apparent activation energy (E\na) could be estimated by regressing the non-isothermal experimental data set to rate expression (5) using MATLAB software. Given that these constants can be appreciably sensitive to even small changes in the temperature due to the exponential form of the Arrhenius expression presented in Eq.\u00a0(33), caution was taken when specifying initial guesses for A and E\na required to initiate the MATLAB regression loop.Linearization of Eq.\u00a0(33) gives:\n\n(34)\n\n\nln\n\n\nk\n=\n\n\u2212\n\n\nE\na\n\nRT\n\n\n+\nlnA\n\n\n\n\nPlotting \n\nln\n\nk\n\n against the inverse of the temperature (1/T) gives rough estimation of the Arrhenius constants A and E\na. These plot estimates could then be specified as the initial guesses used in the regression procedure.However, prior to the construction of Arrhenius plots, non-linear regression on MATLAB software was carried out to estimate the model parameters for the remaining isothermal data sets measured at T=\u00a0190 \u00b0C, 200 \u00b0C and 210 \u00b0CGiven that the overall rate of the CO oxidation reaction increases with temperature as shown in Figs.\u00a05 and 6, the model parameter estimates should also increase with temperature. This trend is seen in Table\u00a05 with exception of \n\nk\n2\n\u2032\n\n at T=\u00a0210 \u00b0C. The range of R\n2 values presented in Table\u00a03\n\n\n reveal that with increase in temperature, there is better agreement between the experimental data and reaction rate expression (5). However, given that the range of R\n2 values lies between 0.665 and 0.802, there is still reasonably close agreement between the experimental data and this kinetic model.Arrhenius plots could then be constructed as follows:\nFigs.\u00a08\n(a), (b) and (c) reveal that in all cases, reasonably good linear fits are obtained. As a result, the pre-exponential factors (A) and apparent activation energies (E\na) shown in Table\u00a06\n could be obtained confidently from the y-intercept and slope of the straight-line presented in Figs.\u00a08(a), (b) and (c), respectively.Regression for the Arrhenius constants could then be carried out on MATLAB software:The MATLAB output for the coefficient of determination (R2) when fitting the consolidated non-isothermal data set to rate expression (5) was found to be 0.717. This result represents the best agreement that could be obtained between the experimental data and rate expression (5) over the range of operating temperaturesFigure\u00a09. The apparent activation energies (E\na), and pre-exponential factors (A) shown in Table\u00a07\n for rate expression (5) are found kinetically realizable given that these Arrhenius constants are both positive and finite quantities. Comparison of Figs.\u00a05 and 6 shows that the reaction rate is more sensitive to changes in the partial pressure of CO with increase in temperature. This could be related to the temperature dependence of the constants \n\nk\n2\n\u2032\n\n and \n\nk\n3\n\u2032\n\n in rate expression (5) given the activation energy associated with \n\nk\n2\n\u2032\n\n (the rate coefficient coupled with the partial pressure of CO in rate expression (5)) is higher than the activation energy associated with \n\nk\n3\n\u2032\n\n (the rate coefficient coupled with the partial pressure of oxygen in rate expression (5)).\nThe model most representative of the kinetics associated with the oxidation of CO over NiO/Al2O3 catalyst, among those proposed in this study, is found to be a reaction rate expression based on a mechanism whereby non-competitive adsorption of reactant and product species occur at low temperature. The model based on this mechanism is observed to best fit the experimental data collected at temperatures between 180 \u00b0C and 210\u00b0C despite the lack of significant statistical difference between the model fit results. However, the agreement between the best fitting model and the experimental data appears to deteriorate with increase in temperature. This is possibly due to the occurrence of side reactions and deviation from the proposed reaction mechanism at higher temperatures. Despite this observation, the identified rate expression fits bulk of the experimental data measured within the range of operating temperatures, as indicated by a coefficient of determination (R\n2) value of 0.717. Therefore, the mechanistic model developed and identified in this study provides a useful description of the kinetics of the NiO catalysed CO oxidation system at low temperature.\n\n\n\n\n\n\n\n\n\nA\n\n[mol gcat\u22121 bar\u22121\ns\u00a0\u2212\u00a01]\nArrhenius pre-exponential constant\n\n\n\nC\n\n[mol cm\u22123]\nReaction rate constant\n\n\n\nCAS\n\n\n[mol cm\u22123]\nSurface concentration of species A\n\n\n\nCWP\n\n\n[-]\nWeisz-Prater criterion\n\n\n\nCmears\n\n\n[-]\nMears criterion\n\n\n\nD\n\n[cm/m]\nDiameter\n\n\n\nDAB\n\n\n[cm2/s]\nDiffusion coefficient\n\n\n\nDe\n\n\n[cm2/s]\nEffective diffusivity\n\n\n\nE\na\n\n[kJ mol\u22121]\nApparent activation energy\n\n\n\nF\n\n[mol s-1]\nMolar flow rate\n\n\n\nk\n\n[mol gcat\u22121\ns\u00a0\u2212\u00a01]\nReaction rate constant\n\n\n\nkc\n\n\n[cm/s]\nMass transfer coefficient\n\n\n\nK\n\n[mol gcat\u22121\ns\u00a0\u2212\u00a01]\nEquilibrium adsorption rate constant\n\n\n\nN\n\n[-]\nReaction order\n\n\n\nP\n\n[bar]\nPartial Pressure\n\n\n\nR\n\n[mol s\u00a0\u2212\u00a01 gcat\u22121]/[cm]\nMethane consumption reaction rate/radius\n\n\n\nRe\n\n[-]\nReynolds number\n\n\n\nR\ng\n\n8. 314 [J mol\u22121\nK\u00a0\u2212\u00a01]\nIdeal gas constant\n\n\n\nS\n\n[-]\nActive surface catalyst site\n\n\n\nSc\n\n\n[-]\nSchmidt number\n\n\n\nSh\n\n[-]\nSherwood number\n\n\n\nT\n\n[s]\nReaction time\n\n\n\nT\n\n[ \u00b0C]\nReaction temperature\n\n\n\nU\n\n[m/s]\nVelocity\n\n\n\nV\n\n[cm2/s]\nKinematic viscosity\n\n\n\nW\n\n[gcat]\nMass\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\u03c3\n\n\n[-]\nStandard deviation\n\n\n\n\n\u03d5\n\n\n[-]\nThiele modulus\n\n\n\n\n\u03b7\n\n\n[-]\nCatalyst effectiveness factor\n\n\n\n\n\u03c1\n\n\n[g/cm3]\nDensity\n\n\n\n\n\u03bc\n\n\n[kg s\u00a0\u2212\u00a01\nm\u00a0\u2212\u00a01]\nDynamic viscosity\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nB\n\nBulk\n\n\n\nCat\n\nCatalyst\n\n\n\nCO\n\nCarbon monoxide\n\n\n\nIn\n\nInlet\n\n\n\nO2\n\n\nOxygen\n\n\n\n0\n\nInitial\n\n\n\nOut\n\nOutlet\n\n\n\nP\n\nParticle\n\n\n\nS\n\nSolid catalyst\n\n\n\nS\n\nSurface\n\n\n\nT\n\nTotal\n\n\n\nV\n\nVacant\n\n\n\n\n\n\n\n\n\n\n\n\n\nCO\nCarbon monoxide\n\n\nE-R\nEley-Rideal\n\n\nGC\nGas chromatograph\n\n\nL-H\nLangmuir-Hinshelwood\n\n\nNiO/Al2O3\n\nAlumina supported nickel oxide catalyst\n\n\nPSSA\nPseudo steady state approximation\n\n\nRM\nReaction mechanism\n\n\nTCD\nTemperature controlled decomposition\n\n\nVOCS\nVolatile organic compounds\n\n\n\n\n\nThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors are grateful to the University of Kwa-Zulu Natal for financial support of the study and Ms S. Naicker for laboratory assistance in the experimental work. This work is based on the research supported by the National Research Foundation of South Africa.", "descript": "\n                  The kinetics of carbon monoxide oxidation over alumina supported nickel oxide was studied using a mechanistic approach to model development and identification. Langmuir-Hinshelwood and Eley-Rideal reaction schemes, together with an alternative scheme were investigated, using low temperature, differential rate measurements. Experimental results were found to be most consistent with the alternative scheme, supporting a reaction mechanism based on non-competitive adsorption of carbon monoxide and oxygen onto different catalyst sites without mutual displacement. The resulting kinetic model was observed to have reasonable agreement with the experimental findings and was found representative of the system.\n               "}
{"full_text": "The oxidation of alcohols is considered a benchmark reaction for the development of new catalysts [1]. Besides this, the oxidation of primary and secondary alcohols to their respective aldehydes and ketones is a common laboratory procedure. These reactions traditionally employ toxic oxidants such as chromium VI salts (dichromate, chromic acid, and chromium trioxide), potassium permanganate [2,3], and pyridinium chlorochromate [4] that, although selective to aldehyde and ketones, generally require an excess to ensure better conversions. However, the procedure drawback is the generation of toxic waste [5]. At the same time, these reactions attract attention mainly because aldehydes and ketones are intermediates of many products used in fine chemistry [6]. The oxidation of benzyl alcohol to benzaldehyde is an important example since benzaldehyde is industrially the most important aromatic aldehyde and one of the main aromatic compounds used in the pharmaceutical, cosmetic, food, and perfumery industries [7\u20139].In this context, several works have sought catalysts and nanocatalysts [10,11] that increase the production of intermediate products, i.e. benzaldehyde [9]. Some of these new procedures have applied catalysts with or without an oxidant, highlighting the application of metallic nanoparticles, including, for instance, Au [6,12], Pt [13], Au-Pd [14], Ag [15]; Ag [8,16], Co [17], and Pd [18] in association with O2; and Cu associated with hydrogen peroxide [19] or TEMPO (2,2,6,6-tetramethylpiperidine-N-oxyl) [1,20]. The oxidant choice is also important, being suitable to the catalyst (e.g. Cu and TEMPO), as well as for the process as a whole, relatively stable and, if possible, having a low cost, as molecular oxygen, hydrogen peroxide, and sodium/calcium hypochlorite.The application of sodium and calcium hypochlorite as an oxidant in organic synthesis was extensively reported in the 1980s. Stevens and co-workers applied \u201cSwimming Pool Chlorine\u201d (an inexpensive commercial sodium hypochlorite solution) to oxidize secondary alcohols to ketones [21] and diols/aldehydes to their respective ketones and methyl esters [22]. At the same time, calcium hypochlorite was applied to oxidize secondary methyl ethers into ketones [23], aldehydes to their corresponding carboxylic acids [24], secondary alcohols to ketones, primary alcohols to esters, and ethers to esters [25]. All these applications report hypochlorite as a versatile, effective, safer, and potential substitute for traditional chromium VI salts [3] and pyridine-based [4] oxidants in organic synthesis. Besides this, some alcohol oxidation using sodium hypochlorite has been recognized as environmentally benign and/or a greener process [26\u201330], when compared with that described by Stevens [21].Germanophosphate glasses containing self-supported nickel-based nanoparticles as catalysts were recently described as a new protocol for benzyl alcohol oxidation using sodium hypochlorite as an oxidant [31]. In this study, good conversions were achieved (\u224875\u00a0mol%), with high benzaldehyde selectivity (>99%), employing mild reaction conditions (20\u00a0\u00b0C, acetonitrile as solvent). In this sense, the glass acts as an active substrate for the synthesis of self-supported Ni-based catalysts. The employment of glass materials in catalysis is practically an unexplored topic [32], and the major application of this material is like a sort of support for catalysts [33\u201338].However, Matzkeit et\u00a0al. [32] reported the use of borophosphate glass as a catalyst for bioactive bis(indolyl)methanes molecules (BIMs) synthesis under solvent-free conditions, achieving high yields. Phosphate-based glass materials can be obtained from simple raw chemicals (e.g., KH2PO4, P2O5, NaPO3, and NaH2PO4). Therefore, phosphate-based glasses could represent an unconventional method that moves towards the development of new heterogeneous catalysts.In this sense, considering the relevance of the development of new catalysts and processes that look for chemical compounds of laboratory and industrial interest in an effective and selective way, and the use of less harmful chemical reagents, this work applied a simple and easy to produce borophosphate glass as a catalyst for the oxidation of benzyl alcohol and 1-phenylethanol by sodium hypochlorite 11\u00a0wt%, in an organic-aqueous biphasic system. To optimize the alcohol conversion and ensure good aldehyde selectivity, some experimental parameters of the reaction such as temperature, oxidant amount, catalyst mass and particle size, and reaction media were evaluated and monitored by High-Performance Liquid Chromatography (HPLC) analysis. HPLC analysis of both organic and aqueous phases was also used to propose the reaction mechanism. The glass catalyst was characterized by Raman spectroscopy, X-ray diffraction, density, and Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES).Borophosphate glass was prepared by melting-quenching technique [34,35] with high-purity reagents (Sigma\u2013Aldrich\u00ae) using NaH2PO4, H3BO3, and Al2O3 as precursors. The NaH2PO4/H3BO3 molar proportion was set as 2 and Al2O3 was added in proportions of 0, 3, 5, 7, and 10\u00a0mol%. In a typical synthesis, 5\u00a0g of the aforementioned compounds in the predetermined proportions were weighted and homogenized in an agate mortar for 10\u00a0min, transferred to a covered Pt crucible, and then melted at 1050\u00a0\u00b0C in a resistive preheated oven for 1\u00a0h. The glass sample was obtained by quenching the molten mixture on a graphite mold at room temperature. Posteriorly, the glass was grounded in an agate mortar and sieved through a 150 to <400 mesh range, and stored under a vacuum desiccator until analysis.The glass density (\u03c1) was measured with the bulk glass samples by the Archimedes method using a density module mounted on a Mettler Toledo ME240/M analytical balance and ethanol as immersion solvent.The total content of aluminum, boron, sodium, and phosphorus on borophosphate glasses was determined by ICP-OES using a Thermo Scientific\u00ae iCap 6000 Series Spectrometer. The ICP solutions were prepared in triplicate with 100\u00a0mg of sample 100-time diluted by ultrapure 1% (v/v) HNO3 aqueous solution in an ultrasonic bath. The analytical standard curves (0.1\u201310\u00a0mg\u00a0L\u22121) were prepared using a multi-element standard solution (Fluka\u00ae) diluted by the aforementioned solvent. All samples were prepared in plastic materials to avoid borosilicate glass interference. Emission lines used for quantification in axial view: Al 396.152\u00a0nm; B 249.678\u00a0nm; Na 589.592\u00a0nm, and P\u00a0185.942\u00a0nm [35].Raman spectrum of the glass sample was recorded using a micro-Raman Renishaw InVia\u00ae, laser power 8\u00a0mW, 633\u00a0nm excitation wavelength, and CCD (Charge Coupled Device) detector. The powder glass sample was measured without any additional treatment. Deconvolution analysis of the Raman spectrum was carried out by Voigt functions using the Fityk program (version 1.3.1).The amorphous nature of the borophosphate glass powders was assessed by X-ray powder diffraction (XRD) measurements, using a Rigaku SmartLab SE Diffractometer equipped with the Cu K\u03b1 radiation (\u03bb\u00a0=\u00a01.5418\u00a0\u00c5), and at angles between 15\u00b0 and 80\u00b0 (\u03b8 \u2013 2\u03b8).First of all, the concentration of the NaOCl solution was determined as 11\u00a0wt% by iodometric titrimetric analysis (Supplementary Information, Section 2).Benzyl alcohol (BnOH) and 1-phenylethanol (BnEtOH) oxidations were conducted using NaOCl 11\u00a0wt% as oxidant and acetonitrile (ACN) as a solvent, based on our previous paper [32]. Initially, the reaction parameters oxidant and catalyst amount, temperature, and catalyst mesh were tested to optimize the reaction conditions, as reported in Table 1\n. All reactions were carried out under constant stirring. Thus, the glass catalyst was first dispersed in 10\u00a0mL ACN for 2\u00a0min, followed by the addition of 0.75\u00a0mmol of the aforementioned alcohols. The oxidant, NaOCl, was added in 4 equal portions (total amount tested divided by 4) in the beginning and within 30, 60, and 90\u00a0min of the reaction, being the reaction time adjusted according to the optimization of reaction conditions [28].After the evaluation of oxidant amount, temperature, catalyst mass, and particle size, the standard reaction conditions applied for additional tests (Supplementary Material), the effect of pH, and catalyst reuse were: 75\u00a0mg and 100\u00a0mg glass-catalyst for benzyl alcohol and 1-phenylethanol reactions, respectively, first dispersed in 10\u00a0mL ACN for 2\u00a0min, followed by the addition of 0.75\u00a0mmol of the alcohols. NaOCl was added in 4 equal portions of 1\u00a0mL (total 4\u00a0mL, 6.4\u00a0mmol) in the beginning and for 30, 60, and 90\u00a0min. The reactions were carried out under constant stirring at 50\u00a0\u00b0C for 3 or 5\u00a0h.For catalyst reuse, the reaction media, containing the catalyst, was centrifuged at 3400\u00a0rpm for 10\u00a0min. The sedimented catalyst was separated from the liquid, washed with 5\u00a0mL of absolute ethanol, and the mixture centrifuged at 3400\u00a0rpm for 5\u00a0min. This procedure was repeated 3 times. Then, the freshly washed catalyst was dried at 80\u00a0\u00b0C for 4\u00a0h and reused.The unconverted BnOH, BnEtOH, and the oxidation products benzaldehyde (BnCHO), benzoic acid (BnCOOH), and acetophenone (BnCOCH) were determined using HPLC (Thermo Scientific \u00ae Ultimate 3000). The separation was performed at 30\u00a0\u00b0C using an octadecylsilane C18 column (Ace ltd.\u00ae), a flow rate of 1\u00a0mL\u00a0min\u22121 in gradient elution, with the mobile phase composed of the mixture of acidified water (0.01% v/v phosphoric acid, pH 2.75\u00a0\u00b1\u00a00.05) and acetonitrile (ACN, J.T. Baker\u00ae HPLC grade): initial 30% ACN \u2192 to 60% ACN in 10\u00a0min, keeping this condition until 15\u00a0min. The detection was made by a diode-array detector (DAD) at 210\u00a0nm. Aliquots of the reaction media were collected at the beginning, during, and at the end of the reaction, diluted 10 times with the mobile phase, and filtered in a 0.22\u00a0\u03bcm hydrophilic PVDF syringe filter [31]. Analytical standards (Sigma Aldrich\u00ae, Supelco\u00ae) were\u00a0used as a reference for sample concentration determination.To verify the distribution of compounds (mmol) in the biphasic formed phases, the organic reaction media was centrifuged (3400\u00a0rpm, 6\u00a0min), and a fraction of both organic and aqueous phases were collected for HPLC analysis.The selectivity, linear range, quantification limit, precision, and accuracy of the HPLC method were previously tested to ensure the correct determination of the reaction compounds. BnOH and BnEtOH conversions (C), product yields (Y), and BnCHO selectivity (S) were calculated, respectively, according to Eqs. (1)\u2013(3). n refers to the number of mmol calculated for each compound.\n\n(1)\n\n\nC\n\n\n\na\nl\nc\no\nh\no\nl\n\n\n\n\n(\n\nm\no\nl\n%\n\n)\n\n=\n\n\n\nn\n\na\nl\nc\no\nh\no\nl\n\n\n\ni\nn\ni\nt\ni\na\nl\n\n\n\n\u2212\n\nn\n\na\nl\nc\no\nh\no\n\nl\n\n\nt\ni\nm\ne\n\n\n\n\n\nn\n\na\nl\nc\no\nh\no\nl\n\n\n\ni\nn\ni\nt\ni\na\nl\n\n\n\n\nx\n\n100\n\n\n\n\n\n\n(2)\n\n\nY\n\n\n\np\nr\no\nd\nu\nc\nt\n\n\n\n\n(\n\nm\no\nl\n%\n\n)\n\n=\n\n\nn\n\np\nr\no\nd\nu\nc\n\nt\n\n\nt\ni\nm\ne\n\n\n\n\nn\n\np\nr\no\nd\nu\nc\nt\n\n\n\nt\nh\ne\no\nr\ne\nt\ni\nc\na\nl\n\n\n\n\nx\n\n100\n\n\n\n\n\n\n(3)\n\n\n\nS\n\nB\nn\nC\nH\nO\n\n\n\n(\n%\n)\n\n=\n\n\nn\n\nB\nn\nC\nH\n\nO\n\ne\nn\nd\n\n\n\n\n\u2211\n\nn\n\ne\nn\nd\n\n\nB\nn\nC\nH\nO\n,\n\nB\nn\nC\nO\nO\nH\n\n\nx\n\n100\n\n\n\n\nAll pH measurements were performed by an MS Tecnopon\u00ae digital pHmether previously calibrated with standard solutions (pH 4\u201310). pH adjustment, when needed, was made with HCl 10\u00a0wt% and NaOH 10\u00a0wt% solutions.\nTable 2\n shows the molar composition of P, Na, B, and Al as their respective oxides, determined by ICP-OES [35], for borophosphate glass doped with 10\u00a0mol% Al2O3 used as a catalyst for benzyl alcohol and 1-phenylethanol oxidation by NaOCl.Figure\u00a0S1 shows the x-ray diffraction patterns for a borophosphate glass series with an increase of Al2O3 content from 0 to 10\u00a0mol%. The absence of crystallization peaks and the presence of the broad amorphous regions (halo) indicate the glassy characteristic of the materials. Moisture-resistant borophosphate glasses (molar ratio P/B\u00a0=\u00a02) can be obtained with the addition of Al2O3 (10\u00a0mol%) at a relatively low fusion temperature (e.g., 1050\u00a0\u00b0C) [32,35]. The successive addition of Al3+ ions reduces the moisture absorption of borophosphate glasses and increases the glass transition temperature (Tg). However, concentrations above 12.5\u00a0mol% Al2O3 crystallize when fused at 1050\u00a0\u00b0C [35].The addition of Al2O3 improves the chemical resistance of borophosphate glasses due to the depolymerization of the phosphate network and the formation of P\u2013O\u2013Al bonds [35,39]. The depolymerization of phosphate glass structure can be observed through Raman spectroscopy, Fig.\u00a01\n. The band at \u2248 330\u00a0cm\u22121 is associated with symmetric stretching of the O\u2013P\u2013O bond in metaphosphate structures (Q2, based on Qn terminology, where n represents the number of bridging oxygen that links one tetrahedron to another) [35,39\u201341]. Borophosphate B\u2013O\u2013P band is observed at\u00a0\u2248\u00a0630\u00a0cm\u22121 [41], and the symmetric stretching of P\u2013O\u2013P at\u00a0\u2248\u00a0700\u00a0cm\u22121 [32,35,40,42]. With the addition of 10\u00a0mol% Al2O3, Al\u2013O\u2013Al bending modes can be noticed at\u00a0\u2248\u00a0540\u00a0cm\u22121 [32,35].The band at \u2248 930\u00a0cm\u22121 (inset graph, Fig.\u00a01) is associated with isolated orthophosphate groups (Q0) and asymmetric stretching of P\u2013O\u2013P in Q2 structures [35,40]. The main effect of Al3+ addition to the borophosphate glass network is the depolymerization process, e.g. the reduction of metaphosphate structures (Q2) and the increase of pyrophosphate groups (Q1). This effect can be observed in the 1000\u20131250\u00a0cm\u22121 region of the Raman spectrum: pyrophosphate band \u03bds P\u2013O occurs at \u2248 1030\u00a0cm\u22121 [41], whilst metaphosphate band \u03bds PO2 occurs at\u00a0\u2248\u00a01100\u00a0cm\u22121 [41], and both overlap to form a single band with a maximum at 1070\u00a0cm\u22121, as highlighted in the inset graph of Fig.\u00a01. The asymmetric stretching \u03bdas PO2 occurs at \u2248 1235\u00a0cm\u22121 [35,41].The density of borophosphate glass rises with Al2O3 addition up to 7.5\u00a0mol% Al2O3 and decreases with further addition of Al2O3 (10\u00a0mol%) (Fig.\u00a02\n). The variations in density indicate that aluminum addition changes the O/P ratios determined from the ICP-OES analysis: for glasses with Al2O3 amount below 7.5\u00a0mol%, O/P\u00a0<\u00a03.5, and metaphosphate structures are predominant (Q2). On the other hand, 10\u00a0mol% Al2O3 borophosphate glass has O/P\u00a0>\u00a03.6, and pyrophosphate groups are predominant [43,44]. These variations in O/P ratios also change the predominant Al groups. In glasses with O/P\u00a0<\u00a03.5, the Al2O3 addition tends to replace P\u2013O\u2013P and PO-Na+ bonds by cross-linked POAl(6)-phosphate chains, increasing the glass density, for instance. However, further aluminum additions change O/P\u00a0>\u00a03.5, with the replacement of POAl(6) groups for more open POAl(4) structures, reducing the glass density [43,44].The oxidation of benzyl alcohol (primary alcohol) and 1-phenylethanol (secondary alcohol) to benzaldehyde and acetophenone, respectively, were carried out using acetonitrile as the solvent, NaClO as oxidant, and borophosphate glass (10\u00a0mol% Al2O3) as the catalyst. After the addition of NaOCl to the reaction, a biphasic organic-aqueous system was formed, and its role is properly discussed in section 3.2.1. The reaction conditions were optimized to determine the best alcohol conversion and aldehyde selectivity.Sodium hypochlorite was added to the reaction at 20\u00a0\u00b0C [31] varying its amount between 1.6\u00a0mmol (1\u00a0mL) and 6.4\u00a0mmol (4\u00a0mL). Fig.\u00a03\n shows the effect of oxidant over alcohol oxidation. When added 1.6 and 3.2\u00a0mmol (oxidant: alcohol molar ratio 2.1:1 and 4.2:1, respectively) in both reactions, BnOH (Fig.\u00a03(a)) and BnEtOH (Fig.\u00a03(b)) achieved a constant conversion, indicating that the oxidant is the limiting reactant. Increasing the NaClO amount to 4.8 and 6.4\u00a0mmol (oxidant: alcohol molar ratio 6.4:1 and 8.5:1, respectively), the conversions rise for both reactions. Applying 6.4\u00a0mmol (4\u00a0mL) of NaClO, BnOH conversion after 10\u00a0h of reaction was 58.0\u00a0mol% (Fig.\u00a03(a))) and BnEtOH conversion was 70.4\u00a0mol% (Fig.\u00a03(b)). Thus, 6.4\u00a0mmol of NaClO was set as the standard amount for further reactions.The reaction using NaOCl as oxidant tends to be strongly affected by temperature, and the addition of this compound leads to an exothermic process, requiring, in some cases, the reactions to be carried out in ice baths or under mild temperatures [21,22,45]. For instance, Mombarg et\u00a0al. [46] have reported that 2,3-butanediol oxidation by NaClO catalyzed by NiSO4\u00b76H2O was slow in temperatures above 20\u00a0\u00b0C, whereas an exothermic reaction occurs at 30\u00a0\u00b0C. However, the formation of the organic-aqueous biphasic system allows us to investigate the effect of ambient to 50\u00a0\u00b0C over the alcohol conversion. In this sense, Fig.\u00a04\n shows the effect of temperature on BnOH and BnEtOH oxidation at 20, 30, and 50\u00a0\u00b0C. The increase of the temperature from 20 to 30 and 50\u00a0\u00b0C provides higher conversions. At 50\u00a0\u00b0C, the maximum conversion of BnOH (77.2\u00a0mol%, Fig.\u00a04 (a)) and BnEtOH (75.7\u00a0mol%, Fig.\u00a04 (b)) were achieved after 5\u00a0h. Thus, the temperature rise to 50\u00a0\u00b0C allows a reduction of reaction time by half without loss in alcohol conversion.\nFig.\u00a05\n shows the effect of catalyst mass on the reaction, where the glass catalyst plays a fundamental role mainly in BnOH oxidation. Without the glass catalyst, conversions reached only 12.4\u00a0mol% for BnOH (Fig.\u00a05(a)) and 0.6\u00a0mol% for BnEtOH (Fig.\u00a05(b)). Nevertheless, for BnOH oxidation, the increase of glass mass from 25\u00a0mg to 75\u00a0mg results in conversions between 72.7 and 78.5\u00a0mol%. Additional mass, 100\u00a0mg, does not result in higher conversions. On the other hand, BnEtOH conversion showed to be more affected by glass-catalyst mass (Fig.\u00a05(b)). The increase from 25\u00a0mg to 75\u00a0mg of glass-catalyst led to conversions between 56.2 and 80.3\u00a0mol%. For further mass increases (100\u2013150\u00a0mg), the conversion levels are between 85 and 87\u00a0mol%. These low conversions achieved in uncatalyzed reactions indicate that the glass catalyst and the biphasic organic-aqueous system play an important role in the process, even being NaOCl readily active in the oxidation of primary and secondary alcohols in other reports [5,21,22,45].At least, the glass-catalyst particle size effect in the reaction was evaluated. The previous evaluations for oxidant, temperature, and catalyst mass were carried out with glass particle sizes between 325 and 400 (Tyler) (44\u201337\u00a0\u03bcm). As a general trend, the reduction of particle size of the catalyst or the use of nanocatalysts tends to increase the surface area and, consequently, the conversion [38,47,48]. So, the glass catalyst sieved with a lower Tyler scale (higher glass particle, 150\u2013325 mesh range) results in reduced yields (67\u201372\u00a0mol% for benzyl alcohol and 68\u201370\u00a0mol% for 1-phenylethanol), whereas higher Tyler sieves (325\u2013400 and\u00a0<400 mesh) rise the conversion, Fig.\u00a06\n. However, the lowest particle glass catalyst (<400 mesh) shows an opposite behavior with a reduced conversion (72.3\u00a0mol% for benzyl alcohol and 77.1\u00a0mol% for 1-phenylethanol) in relation to the 400\u2013325 range (higher glass particle size, 78.2\u00a0mol% for benzyl alcohol and 85.6\u00a0mol% for 1-phenylethanol) due to its agglomeration and adhesion to the reaction vessel.The \u201csalting-out\u201d [49] or \u201csugaring-out\u201d [50,51] is a well-established analytical procedure used to extract organic solutes from water [49], allowing its quantitative analysis [52]. The process is characterized by the formation of a biphasic organic-aqueous system when a salt is added to a water-acetonitrile mixture, for instance. Thus, the addition of aqueous NaClO 1.6\u00a0mmol\u00a0mL\u22121 to acetonitrile results in a biphasic organic-aqueous system [31]. The catalyzed reaction proceeds in the aqueous phase and the oxidized products are transferred to the organic phase. At the end of the procedure, the reaction system was centrifuged to separate the organic from the aqueous phase and to determine the concentration of the reactants and products, Table 3\n. Benzyl alcohol and benzaldehyde are present mainly in the organic phase, whereas benzoic acid was only observed in the aqueous phase in a small amount. The maximum percentage of BnOH mmols present in the aqueous phase related to the organic phase was 1.3%, whilst for BnCHO it was 0.5%. The same pattern occurs with BnEtOH and BnCOCH, where the maximum percentage of BnEtOH mmols present in the aqueous phase was 0.5%\u00a0related to the organic phase, whilst for BnCOCH it was 0.2%.Acetophenone is the only expected product of 1-phenylethanol oxidation (secondary alcohol), whilst two products can be obtained by benzyl alcohol oxidation\u2014benzaldehyde and benzoic acid. Benzaldehyde is one of the most important aromatic molecules applied in the cosmetic, perfumery, food, and pharmaceutical industries [8]. Thus, the development of selective routes for aldehyde synthesis using low-cost and greener reagents is highly desirable. The biphasic system enables a high selectivity for benzaldehyde under a wide range of experimental conditions (Table 4\n). The catalytic reaction proceeds in the organic phase, and the benzaldehyde is transferred to the organic phase avoiding further oxidation by sodium hypochlorite, a strong oxidant.Grill and co-workers [5] used commercial bleach (\u22485% aqueous sodium hypochlorite) and a nickel salt to convert benzyl alcohol directly to benzoic acid achieving 89% yield, with or without an organic solvent (dichloromethane). On the other hand, the oxidation of 1-phenylethanol reached only 33% after 4\u00a0h of reaction using dichloromethane as solvent. Mirafzal and Lozeva [53], using ethyl acetate as the solvent, achieved a 93% yield for the aldehyde when tetrabutylammonium bromide was applied as a phase transfer catalyst (PTC) to improve benzaldehyde selectivity. The authors concluded that in the absence of the quaternary ammonium salt, little or no reaction was evident. Okada et\u00a0al. [54] have used NaOCl\u22c55H2O crystals as the oxidant to convert benzyl alcohol to benzaldehyde with a 99% yield. However, the reaction was performed in dichloromethane using 1\u00a0mol% TEMPO and 5\u00a0mol% Bu4NHSO4. Similar results were also obtained by Abramovici et\u00a0al. [55], and Lee et\u00a0al. [56,57] using sodium hypochlorite as oxidant, a water-immiscible solvent, and a phase transfer catalyst (PTC); Vitaku and Christie [26] used bleach as oxidant, ethyl acetate as the solvent, and NaHSO4 as an acid source, and Fukuda et\u00a0al. [45] using NaOCl with an imide compound-nitroxyl radical catalyst system.\nFig.\u00a07\n shows the proposed process for benzyl alcohol and 1-phenylethanol oxidation by NaOCl using the borophosphate glass as the catalyst. Based on the compound distribution in the organic and aqueous phases (Table 3), the pathway for the catalytic process is composed of six steps. First, benzyl alcohol or 1-phenylethanol dissolves in acetonitrile (1). After the oxidant (NaClO) addition and the formation of the biphasic system, a small amount of alcohol is transferred to the aqueous phase (2). During the reaction, the glass catalyst is dispersed in the aqueous phase where BnOH (3) and BnEtOH (5) oxidation occurs. The products are transferred back to the organic phase (6). As a consequence of a small solubility of benzaldehyde in the aqueous phase, benzoic acid is formed only as a by-product of benzaldehyde oxidation by NaOCl (4).To test the proposed mechanism, considering that benzaldehyde is more reactive (it undergoes autoxidation under exposure to air at room temperature) [9] and susceptible to oxidation than benzyl alcohol [1] (that is considered an intermediary on benzyl alcohol oxidation by NaOCl to benzoic acid) [5], the BnCHO susceptibility to oxidation was evaluated applying the standard reaction conditions described in Fig.\u00a08\n. Initially, benzaldehyde is dissolved in acetonitrile (1). After the addition of NaOCl and the formation of the biphasic system, a small amount of BnCHO is transferred to the aqueous phase (2). where its oxidation to benzoic acid took place (3). Even being more reactive than BnOH, BnCHO conversion achieved only 24.4\u00a0mol%. At the end of the reaction, the medium was centrifuged, and compound distribution was determined in both phases (values % described in Fig.\u00a08). Benzaldehyde is present mainly in the organic phase, whereas 0.5% (3.27\u00a0\u00d7\u00a010\u22123\u00a0mmol) is present in the aqueous phase. The concentration of benzoic acid in the organic phase is 17.4% (2.68\u00a0\u00d7\u00a010\u22122\u00a0mmol), whereas 82.6% (1.28\u00a0\u00d7\u00a010\u22121\u00a0mmol) of benzoic acid is present in the aqueous phase.In this sense, whilst the solubility of BnOH in water is 40\u00a0g\u00a0L\u22121 (25\u00a0\u00b0C) [7], the solubility of BnCHO is 4\u00a0g\u00a0L\u22121 (20\u00a0\u00b0C). Thus, the BnOH is transferred to the aqueous phase to react with \u2212OCl, and the BnCHO, once formed, is transferred to the organic phase (ACN), avoiding further oxidation by \u2212OCl, which results in high selectivity. Following the same pattern, the water solubility of 1-phenylethanol and acetophenone at 25\u00a0\u00b0C is 1.95\u00a0g\u00a0L\u22121 and 6.1\u00a0g\u00a0L\u22121, respectively [58]. Therefore, we can infer that even with ketone being more water-soluble than the secondary alcohol, the reaction occurred with satisfactory yields, and acetophenone is transferred to ACN once formed, Table 3.The application of sodium hypochlorite as the oxidant for aldehydes/alcohols/amines conversion using phase transfer catalysis (PTC) is performed with a water-insoluble solvent (usually a chlorinated solvent) [53,55,56,59] and a quaternary ammonium salt, whose function is transporting \u2212OCl to the organic phase [56]. In these cases, the oxidation mostly occurs in the organic phase, whereas in the mechanism of Fig.\u00a07 the reaction takes place in the aqueous phase without quaternary ammonium salt for phase transference. In our reaction condition, no trace of \u2212OCl was determined by iodometric titration (Supplementary Information Section 2) in the corresponding organic phase, supporting this statement.The effectiveness of \u2212OCl in the reaction is pH-dependent, being less effective and slower whereas the pH increases from 8 to 13 [3,46,55,59]. At the same time, pH lower than 7 has the opposite trend due to the fast decomposition of the hypochlorite solution [55]. Borophosphate glass reduces the pH of the solution from \u224813 (the original pH of NaOCl) to\u00a0\u2248\u00a09. The pH decreasing enables the formation of HOCl and, consequently, the oxidation [3,45]. Catalytic reactions were evaluated without the glass catalyst adjusting the pH of NaOCl solution to 9.00\u00a0\u00b1\u00a00.05 with HCl and NaOH. Fig.\u00a09\n shows the conversions obtained by glass-catalyzed reactions and uncatalyzed reactions (without and with pH adjustment). The pH decrease of NaOCl solution from 13 to 9 increases the conversion from 0.6\u00a0mol% to 12.4\u00a0mol% to 50.5\u00a0mol% and 44.7\u00a0mol% for BnEtOH and BnOH, respectively, whereas the glass-catalyzed reactions reach conversions of 87.0\u00a0mol% for BnEtOH and 78.5\u00a0mol% for BnOH.The pH reduction of NaOCl solution in the glass-catalyst reactions occurred due to the controlled release of the catalyst during the reaction. The ICP-OES analysis (Supplementary Information, Table S2) demonstrated the presence of P in the aqueous phase, with values varying between 6 and 15\u00a0wt% depending on the test. The in situ release of phosphate groups reduces the pH and activates the oxidant, i.e. NaOCl. The Al3+ ion released to the aqueous phase was negligible (maximum value 0.14\u00a0wt%). Furthermore, the role of phosphate-based glasses for catalytic purposes [32,35] is demonstrated by the ineffectiveness of four commercial silica-based glasses evaluated (Supplementary Materials Table S3 entry 1\u20134).To evaluate the efficacy of glass-catalyst recycling, the catalyst was recovered by centrifugation, cleaned, dried, and applied again as a catalyst for benzyl alcohol and 1-phenylethanol oxidation by NaOCl. Fig.\u00a010\n shows the glass-based catalyst performance after three cycles for both reactions.The reuse of the glass-based catalyst can be accomplished without a significant decrease in the conversions, mainly between the first and the second cycle of reaction. The conversions reduce just 14.5% and 25.8% for BnOH and BnEtOH, respectively, in the third cycle. In this sense, even with a partial glass release during the process, the catalysis remains with high selectivity.In addition to the glass-catalyst recycling, we have tested acetonitrile recovery by applying two daily common and simple laboratory procedures of solvent recovery\u2014fractioned distillation and vacuum distillation. The recovered acetonitrile was applied as the solvent in a new cycle of BnOH and BnEtOH oxidation, Supplementary Information Table S5. The reaction where ACN recovered by fractioned distillation was used as solvent achieved good conversions after 3\u00a0h of reaction (57.6\u00a0mol% for BnOH and 58.8\u00a0mol% for BnEtOH). In addition, the acetonitrile recovered by fractioned distillation presented only traces (below the HPLC quantification limit) of both alcohols and the reaction products (Supplementary Information Table S4). These results demonstrated that not only the solvent can be easily recovered by a simple method and recycled in a new reaction cycle, but also the reaction products can be easily isolated from the solvent.Besides the high benzaldehyde selectivity, relatively mild reaction conditions, and the possibility of catalyst and solvent recovery, our proposed reaction also allowed the simultaneous oxidation of benzyl alcohol and 1-phenylethanol, as can be seen in Supplementary Information Fig.\u00a0S1. The oxidation of both alcohols was concomitant and allowed good conversions (BnEtOH 79.6\u00a0mol% and BnOH 71.2\u00a0mol%), in addition to exempting the need for isolated reactions if a mixture of aldehyde and ketone is required as the product.Some common organic solvents were combined with NaOCl 11\u00a0wt% to investigate the salting-out formation: acetone, dimethyl sulfoxide (DMSO), and isopropyl alcohol do not form the biphasic system; on the other hand, it was formed with ethyl acetate and dimethyl carbonate (DMC). Based on this, ethyl acetate was tested as a solvent (see Supplementary Information Table S3 entries 5\u20136) instead of acetonitrile, but lower conversions were achieved for both alcohols\u2014BnOH 28.8\u00a0mol% and BnEtOH 17.3\u00a0mol%\u2014compared to reactions that have used ACN as solvent. The same trend was observed for DMC, resulting in similar conversions\u2014BnOH 28.7\u00a0mol% and BnEtOH 13.8\u00a0mol% (see Supplementary Information Table S3 entry 7).The depolymerization of phosphate chains with the addition of 10\u00a0mol% Al2O3 in borophosphate glass (NaH2PO4/H3BO3 ratio\u00a0=\u00a02) is associated with the formation of phosphate-aluminum structures that cross-link to each other, enhancing the glass-network strength, increasing the chemical resistance, and making this glass moisture resistant, which makes this material attractive for applications in various fields. In this sense, borophosphate glass effectively catalyzed benzyl alcohol and 1-phenylethanol oxidation by aqueous 11\u00a0wt% sodium hypochlorite using acetonitrile as a solvent, under mild conditions. The reaction conditions oxidant amount, temperature, mass, and particle size of the catalyst were screened to achieve high alcohol conversions (87.0\u00a0mol% for 1-phenylethanol and 79.4\u00a0mol% for benzyl alcohol) and benzaldehyde selectivity above 95%. The addition of NaOCl to the reaction results in a biphasic organic-aqueous system. HPLC analysis allowed us to infer that a small amount of the alcohol is transferred to the aqueous phase, where it is oxidized. Once formed, benzaldehyde and acetophenone are transferred back to the organic phase. The formation of the biphasic system prevents benzaldehyde oxidation, even employing a strong oxidant, and can be an interesting option for processes that look for intermediary compounds. Furthermore, the proposed biphasic system exempted the use of PTC. The ICP-OES analysis allowed us to infer that the catalytic activity of borophosphate glass occurs due to partial liberation of phosphate-based groups to the aqueous oxidant, reducing the pH of NaOCl from 13 to 9, which enables the formation of HOCl and, consequently, the oxidation. Based on this, the use of phosphate-based glasses as the catalyst is something unexplored and promising: it is easily and quickly produced with low-cost raw material, and its chemical properties can be modified according to the application required, i.e. control the dissolution during the reaction, it can be used as a host for several metal ions, and as an active material for supported nanoparticles applied in catalysis.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Ricardo Schneider would like to acknowledge Conselho Nacional de Desenvolvimento Cient\u00edfico e Tecnol\u00f3gico (CNPq) for funding (grant 422774/2018\u20139). Jorlandio F. Felix acknowledges the CNPq (grant number: 430470/2018\u20135 and 309610/2021-4) and Funda\u00e7\u00e3o de Apoio a Pesquisa do Distrito Federal(FAPDF) (grant number: 193.001.757/2017), for financial support.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.jmrt.2022.05.105.", "descript": "\n                  The oxidation of primary and secondary alcohols to their respective aldehydes/ketones is one of the most important reactions in fine chemistry due to the industrial application of these products. Based on this, a large number of new catalysts and oxidants have been tested using this reaction as a catalytic model, mainly looking for a process that ensures high aldehyde selectivity. In this paper, we have used moisture stable borophosphate glass doped with 10\u00a0mol% Al2O3 as a heterogeneous catalyst in the oxidation of sodium hypochlorite, an effective, greener, and low-cost oxidant, using acetonitrile as solvent under mild conditions. The glass catalyst mass and the particle size were evaluated, as were the reaction temperature and oxidant amount, to determine the ideal reaction conditions where the conversions achieved 87.0\u00a0mol% for 1-phenylethanol to acetophenone and 79.4\u00a0mol% for benzyl alcohol to benzaldehyde, with benzaldehyde selectivity above 95%. Although sodium hypochlorite is a strong oxidant, benzaldehyde was the main product of the oxidation of benzyl alcohol due to the formation of a biphasic organic-aqueous system that protects the aldehyde from oxidation and allows the reaction to occur without the use of a phase transfer catalyst (PTC). HPLC analysis of both phases showed that alcohols, aldehyde, and ketone were mostly present in the organic phase (concentrations above 98.7%). During the reaction, a small amount of alcohol is transferred to the aqueous phase, where the oxidation took place. Once formed, the products are transferred back to the organic phase. ICP-OES analysis indicates that borophosphate glass acts in the reaction by partially releasing phosphate-based groups, reducing the pH of hypochlorite to 9. In this sense, borophosphate glasses prove to be a simple and inexpensive alternative for the development of new catalysts.\n               "}
{"full_text": "catalyst to oil weight ratiofinal boiling pointfluid catalytic crackingheavy cycle oilinitial boiling pointlight cycle oilliquefied petroleum gasesmunicipal solid wasteplastic pyrolysis oilmean squared errorvacuum gasoilcorrelation coefficientmolar concentration (mol cm-3)apparent activation energy (kJ mol-1)certain lumpcertain reactionapparent kinetic parameter (m6 kgcat\n-1 kmol-1 s-1 / m3 kgcat\n-1 s-1)catalyst deactivation kinetic parameter (s-1)reaction ordermolecular weight of HCO lump (g mol-1)mass of catalyst (g)mass of PPO (g)number of experimentsnumber of lumpsnumber of parameterscertain experimentreaction rateideal gas constant (8.314 J mol-1 K-1)contact time (s)reaction temperature (\u00b0C)reference temperature (\u00b0C)volume of the reactor (cm3)weight fractionlevel of significanceactivity termstoichiometric coefficientdegrees of freedomThe development and wellness of the humankind implies an increase of global pollution. One of the consequences is the increasing presence of waste plastics in the municipal solid wastes (MSW), which overflows the management capability of both public and private entities. Consequently, an unacceptable amount of these wastes ends up landfilled, causing the contamination of the soils and the aquifers [1]. In order to solve these problems, it is well established the interest on tertiary recycling by means of thermochemical processes, i.e. pyrolysis and gasification [2].The fast pyrolysis of plastics is performed at low temperature, using high heating rates and short residence time for the volatiles. Moreover, it can be carried out in simple and versatile units equipped with different types of reactors (rotary kilns, screw reactors, fluidized or spouted beds, etc.) entailing a reduced environmental impact and with the possibility of tuning the operating conditions to adapt the production to the type of plastic fed [3,4]. The ideal goal of pyrolysis processes is the monomer recovery, which can be done with high yields in the pyrolysis of polystyrene [5] and polymethyl-methacrylate [6]. On the other hand, for polyolefinic plastics, which constitute two thirds of the plastic fraction found in the MSW [7], is of great interest the production of plastic pyrolysis oil (PPO) because of its possibilities to be used as an alternative fuel [8].Based on its properties, the PPO has been considered as a potential fuel for diesel engines feeding it neatly or blended with commercial diesel [9]. Nevertheless, the PPO does not meet the tough requirements of commercial fuels and requires of physicochemical treatments to adapt its composition [10]. This situation has led to the proposal of integrating the fast pyrolysis of waste plastics with the upgrading of the PPO in refinery units (Waste-Refinery) [11]. The interest of the proposal lays on the capacity of refinery units for valorizing the PPO, either in ad hoc catalytic units or in already existing industrial units. The fluid catalytic cracking (FCC) units are the most appropriate ones in the short term, given their high capacity and versatility to manage unconventional feeds, such as diverse secondary refinery streams [12,13] or bio-oil [14]. Indeed, the chemical composition of the PPO (highly olefinic and free of aromatics) makes it appropriate to be fed to catalytic cracking units with the aim of producing fuels free of sulfur and nitrogen [15]. Furthermore, within the facilities available in the refineries, there are the required fractionation and conditioning units to obtain fuels similar to conventional ones. Among the advantages that the Waste\u2013Refinery strategy offers, the following ones must be highlighted: (i) the recycling of petroleum-derived products with the subsequent savings of raw materials; (ii) the removal of economic barriers that entails the design and construction of new units, which correspond to the high cost of the equipment and of the marketing of non-conventional fuels that would compete against the conventional fuels; and, (iii) the rational organization of the plastics recycling, carrying out the pyrolysis process in a delocalized way in units located nearby of the waste plastics collection and segregation points. The PPO would be afterwards transported from different geographical areas to centralized refineries for its large-scale valorization. Feeding a liquid stream, such as the PPO, into a cracking unit, entails less technical difficulties that the feeding of pure polyolefins, the cracking of which has been also studied [16\u201318]. Nonetheless, these initiatives will require a rigorous control of the feeds, since their composition can be easily contaminated by the presence of different plastics and of additives and pollutants in the waste plastics.In a previous work it has been studied the effect of the properties of different FCC equilibrium catalysts on the production of fuel from PPO, operating at 500\u2013560\u00a0\u00b0C and using a riser simulator reactor [19]. Interestingly, the yields of naphtha (highly olefinic and with a high octane rating) and light olefins were superior to 40 and 12\u00a0wt%, respectively. The proposed initiative is similar to that of cracking wax from Fischer-Tropsch process with the aim of producing high octane gasoline and light olefins [20,21].Both for the simulation and optimization of an ad hoc designed reactor and for the feeding of the PPO to an industrial FCC unit, it is required a kinetic model capable of quantifying the products distribution. Traditionally, the efforts in the kinetic modeling have been focused on the cracking of vacuum gasoil (VGO). Moustafa and Froment [22] were pioneers in taking into account the heterogeneous composition of the VGO and they proposed a kinetic model with a complex reaction scheme that described the individual reactions involved and the formation of coke by means of elementary steps. This type of molecular-level kinetic models has been also applied for the cracking of wax from Fischer-Tropsch process [21]. Nevertheless, most of the works have established lump-based kinetic models that simplify the computing and their posterior use in the design of the reactor [23\u201325]. Apart from the complexity of the reaction scheme, an additional difficulty for obtaining kinetic models is the extremely fast deactivation of the catalyst caused by coke deposition [26,27]. Kinetic models for the catalytic cracking of VGO consider between 3 and 17 lumps and have been collected by different authors [28,29]. These models assume that kinetic parameters are apparent values as a consequence of the diffusional restrictions caused by the components of the VGO (especially the heavier ones) [30].In this work, it has been established a six-lump based kinetic model for the cracking of PPO obtained in the pyrolysis of high-density polyethylene from the experimental data obtained in a previous work [19]. The aim of the work is to provide a tool for quantifying the effects of the operating conditions on the yields of products of interest, such as fuels (gasoline and diesel) and commodities (light olefins). In the modeling, it has been taking into account the catalyst deactivation by coke deposition, which is extremely significant in cracking reactions. Moreover, the analysis of the kinetic parameters obtained for three different FCC equilibrium catalysts with different acidity and porous structure allows for assessing the effect of these properties on the different catalytic steps and on the formation of coke.The plastic pyrolysis oil (PPO) has been obtained at 500\u00a0\u00b0C under fast pyrolysis conditions by feeding virgin high-density polyethylene (HDPE) to a fountain confined conical spouted bed reactor [31].Three different commercial equilibrium FCC catalysts (ECAT-1, ECAT-2 and ECAT-3) have been used in the work. The catalysts have been collected from the catalyst purge stream of industrial FCC units, specifically ECAT-1 from Petronor Refinery (Spain) and the other catalysts from Petrobras Refinery (Brazil). Consequently, they are equilibrium catalysts, since they have been submitted to numerous cycles composed of reaction, stripping and regeneration steps in their corresponding FCC units [32].The catalytic cracking runs have been performed on a laboratory scale micro-riser reactor, specifically designed to mimic the conditions of the riser reactor of industrial FCC units [33]. A schematic representation of the experimental unit together with an explanation of the experimental procedure can be found elsewhere [34]. The operating conditions of riser simulator reactor have been: temperature, 500, 530 and 560\u00a0\u00b0C; catalyst to oil weight ratio (C/O), 3\u20137 gcat gPPO\n-1; and contact time, 1.5\u20136\u00a0s.The catalytic cracking of the PPO has been described by means of a six-lump reaction network. The six lumps are heavy cycle oil (HCO, C20+), light cycle oil (LCO, C13-C20), naphtha (C5-C12), liquefied petroleum gases (LPG, C3-C4), dry gas (C1-C2) and coke (carbonaceous material deposited on the catalyst). The reaction network in Fig. 1\na corresponds to parent reaction network and it accounts for sixteen kinetic parameters, to which it must be added a parameter for catalyst deactivation.The kinetic modeling methodology used is based on the one developed by Toch et al. [35] for catalytic processes with complex pathways and on that proposed by Cordero-Lanzac et al. [36,37], since they included the catalyst deactivation on it. Furthermore, the methodology has been adapted for handling the experimental data obtained in a batch reactor. Likewise, a molar balance to the micro riser reactor has been also required in order to properly describe the behavior of the different lumps [38]. According to the reaction network in Fig. 1a, the reaction rate equations that describe the evolution with contact time of the different lumps are listed below.\n\n(1)\n\n\n\n\n\n\ndy\n\n\nHCO\n\n\n\n\ndt\n\n\n=\n-\n\n\n\u03c6\n\n\n\n\n\n\n\n\nm\n\n\nPPO\n\n\n\n\n\nM\n\n\nHCO\n\n\n\nV\n\n\n\n\n\n\n\n\n\nk\n\n\n1\n\n\n+\n\n\nk\n\n\n2\n\n\n+\n\n\nk\n\n\n3\n\n\n+\n\n\nk\n\n\n4\n\n\n+\n\n\nk\n\n\n5\n\n\n\n\n\n\n\n\ny\n\n\nHCO\n\n\n2\n\n\n\n\n\n\n\n\nm\n\n\ncat\n\n\n\n\nV\n\n\n\n\n\n\n\n\n(2)\n\n\n\n\n\n\ndy\n\n\nLCO\n\n\n\n\ndt\n\n\n=\n\n\n\u03c6\n\n\n\n\n\n\n\n\nm\n\n\nPPO\n\n\n\n\n\nM\n\n\nHCO\n\n\n\nV\n\n\n\n\n\n\n\n\n\nk\n\n\n1\n\n\n\n\n\ny\n\n\nHCO\n\n\n2\n\n\n\n\n\n-\n\n\n\n\n\nk\n\n\n6\n\n\n+\n\n\nk\n\n\n7\n\n\n+\n\n\nk\n\n\n8\n\n\n+\n\n\nk\n\n\n9\n\n\n\n\n\n\n\n\ny\n\n\nLCO\n\n\n\n\n\n\n\n\nm\n\n\ncat\n\n\n\n\nV\n\n\n\n\n\n\n\n\n(3)\n\n\n\n\n\n\ndy\n\n\nNaphtha\n\n\n\n\ndt\n\n\n=\n\n\n\u03c6\n\n\n\n\n\n\n\n\nm\n\n\nPPO\n\n\n\n\n\nM\n\n\nHCO\n\n\n\nV\n\n\n\n\n\n\n\n\n\nk\n\n\n2\n\n\n\n\n\ny\n\n\nHCO\n\n\n2\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n6\n\n\n\n\n\ny\n\n\nLCO\n\n\n\n\n\n-\n\n\n\n\n\nk\n\n\n10\n\n\n+\n\n\nk\n\n\n11\n\n\n+\n\n\nk\n\n\n12\n\n\n\n\n\n\n\n\ny\n\n\nNaphtha\n\n\n\n\n\n\n\n\nm\n\n\ncat\n\n\n\n\nV\n\n\n\n\n\n\n\n\n(4)\n\n\n\n\n\n\ndy\n\n\nLPG\n\n\n\n\ndt\n\n\n=\n\n\n\u03c6\n\n\n\n\n\n\n\n\nm\n\n\nPPO\n\n\n\n\n\nM\n\n\nHCO\n\n\n\nV\n\n\n\n\n\n\n\n\n\nk\n\n\n3\n\n\n\n\n\ny\n\n\nHCO\n\n\n2\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n7\n\n\n\n\n\ny\n\n\nLCO\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n10\n\n\n\n\n\ny\n\n\nNaphtha\n\n\n\n\n\n-\n\n\n\n\n\nk\n\n\n13\n\n\n+\n\n\nk\n\n\n14\n\n\n\n\n\n\n\n\ny\n\n\nLPG\n\n\n\n\n\n\n\n\nm\n\n\ncat\n\n\n\n\nV\n\n\n\n\n\n\n\n\n(5)\n\n\n\n\n\n\ndy\n\n\nDry Gas\n\n\n\n\ndt\n\n\n=\n\n\n\u03c6\n\n\n\n\n\n\n\n\nm\n\n\nPPO\n\n\n\n\n\nM\n\n\nHCO\n\n\n\nV\n\n\n\n\n\n\n\n\n\nk\n\n\n4\n\n\n\n\n\ny\n\n\nHCO\n\n\n2\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n8\n\n\n\n\n\ny\n\n\nLCO\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n11\n\n\n\n\n\ny\n\n\nNaphtha\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n13\n\n\n\n\n\ny\n\n\nLPG\n\n\n\n\n\n-\n\n\n\n\n\nk\n\n\n15\n\n\n\n\n\ny\n\n\nDryGas\n\n\n\n\n\n\n\n\n\n\n\nm\n\n\ncat\n\n\n\n\nV\n\n\n\n\n\n\n\n\n(6)\n\n\n\n\n\n\ndy\n\n\nCoke\n\n\n\n\ndt\n\n\n=\n\n\n\u03c6\n\n\n\n\n\n\n\n\nm\n\n\nPPO\n\n\n\n\n\nM\n\n\nHCO\n\n\n\nV\n\n\n\n\n\n\n\n\n\nk\n\n\n5\n\n\n\n\n\ny\n\n\nHCO\n\n\n2\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n9\n\n\n\n\n\ny\n\n\nLCO\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n12\n\n\n\n\n\ny\n\n\nNaphtha\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n14\n\n\n\n\n\ny\n\n\nLPG\n\n\n\n\n\n+\n\n\n\n\n\nk\n\n\n15\n\n\n\n\n\ny\n\n\nDryGas\n\n\n\n\n\n\n\n\n\n\n\nm\n\n\ncat\n\n\n\n\nV\n\n\n\n\n\n\nbeing yi the weight fraction of lump i, t the contact time between the reactants and the catalyst in the reactor, MHCO the molecular weight of the lump HCO, V the volume of the reactor, kj the apparent rate constant of reaction j, mPPO the mass of PPO fed and mcat the mass of catalyst used.One should observe that the catalytic cracking of the different lumps has been described using irreversible first-order reactions, with the exception of the cracking of HCO lump, which has been considered as an irreversible second-order reaction [28,39]. Additionally, it has been assumed that cracking reactions are non-selectively affected by catalyst deactivation and it has been quantified in Eqs. (1)-(6) by using the same activity term (\u03c6), which has been defined as:\n\n(7)\n\n\n\u03c6\n=\n\n\n\n\n(\n-\n\n\nr\n\n\nj\n\n\n)\n\n\n\n\n\n\n\n\n\n(\n-\n\n\nr\n\n\nj\n\n\n)\n\n\n0\n\n\n\n\n=\nexp\n\n\n\n\n-\n\n\nk\n\n\nd\n\n\n\nt\n\n\n\n\n\n\nwhere (\u2212rj) and (\u2212rj)0 are the reaction rates of each step of the reaction network at t time and zero time, respectively, and kd is the deactivation parameter.The equation proposed for explaining the deactivation kinetics corresponds to a first-order exponential function, which is effective for describing the activity decay in the cracking reactions where a notably deactivation occurs for short contact times (<20\u00a0s) [40].For computing the kinetic parameters, they have been expressed as a function of temperature by means of the reparameterized Arrhenius equation in order to avoid the regression issues derived from the strong correlation between the activation energy and the pre-exponential factor.\n\n(8)\n\n\n\nk\nj\n\n=\n\nk\n\nj\n\n*\n\n\n\nexp\n\n\n\n\n\n\n-\n\n\nE\nj\n\nR\n\n\n\n\n\n\n\n1\nT\n\n-\n\n1\n\n\nT\n\n*\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nbeing kj* the kinetic reaction rate of the j reaction step at the reference temperature T* (500 \u00b0C), Ej the corresponding apparent activation energy and R the universal gas constant.The system of differential equations that describes the catalytic cracking of PPO, Eqs. (1)-(6), has been solved using an in-house written MATLAB code. The code estimated the required kinetic rate constants and activation energies to fit the weight fraction of the different parameters to those experimentally obtained. To find the best fitting values, a loss function in which the mean squared error is minimized has been employed.\n\n(9)\n\n\nLoss Function\n\n=\n\n\n1\n\n\n\n\nn\n\n\ne\n\n\n\n\n\u00b7\n\n\n\u2211\n\n\n1\n\n\n\n\nn\n\n\nl\n\n\n\n\n\n\n\u2211\n\n\n1\n\n\np\n\n\n\n\n\n\n\n\ny\n\n\ni,p\n\n\ncal\n\n\n-\n\n\ny\n\n\ni,p\n\n\nexp\n\n\n\n\n\n\n2\n\n\n\n\n\nwhere yip is the weight fraction of lump i for experiment p, nl is the total number of lumps and ne the total number of experiments. Moreover, superscripts \u201ccal\u201d and \u201cexp\u201d denote the calculated and experimentally determined weight fractions, respectively.The discrimination between the different models proposed has been performed by means of a statistical significance test based on Fisher\u2019s method. The procedure is well explained in the literature [41]. In brief, for two kinetic models with different degrees of freedom (\u03bdA\u00a0\u2260\u00a0\u03bdB) if model B shows a smaller mean squared error than model A (SSEB\u00a0<\u00a0SSEA), the improvement offered by model B with respect to model A will be statistically significant when the following condition is fulfilled:\n\n(10)\n\n\n\n\nF\n\n\nA\n-\nB\n\n\n=\n\n\n\n\n\n\nSSE\n\n\nA\n\n\n-\n\n\nSSE\n\n\nB\n\n\n\n\n\n\nSSE\n\n\nB\n\n\n\n\n\n\n\n\n\n\n\u03bd\n\n\nA\n\n\n-\n\n\n\u03bd\n\n\nB\n\n\n\n\n\n\n\u03bd\n\n\nB\n\n\n\n\n\n\n\n>\n\n\n\nF\n\n\n1\n-\n\u03b1\n\n\n\n\n\n\n\n\u03bd\n\n\nA\n\n\n-\n\n\n\u03bd\n\n\nB\n\n\n,\n\n\n\u03bd\n\n\nB\n\n\n\n\n\n\n\n\n\nbeing F1-\u03b1 the critical value of the Fischer distribution function for a level of significance of 95% (\u03b1\u00a0=\u00a00.05). The degrees of freedom have been computed according to the following equation by taking into account the number of experiments (ne), number of lumps (nl) and number of parameters (np):\n\n(11)\n\n\n\u03bd\n=\n\n\n\n\n\nn\n\n\ne\n\n\n\u00b7\n\n\nn\n\n\nl\n\n\n\n\n\n-\n\n\nn\n\n\np\n\n\n\n\n\n\nThe main properties of the PPO are provided in Table 1\n. It consists of a mixture of hydrocarbons with a broad distillation range that can be divided into 82.0\u00a0wt% of HCO (heavy cycle oil), 12.5\u00a0wt% of LCO (light cycle oil) and 5.5\u00a0wt% of naphtha. These fractions have been defined according to the usual criteria followed by oil refiners: naphtha (C5\u2013C12), LCO (C13\u2013C20) and HCO (C21+) [42]. Additionally, the chemical composition of the PPO obtained by chromatographic means has been already reported in our previous work [43]. Briefly, they are composed of 67.6\u00a0wt% of olefins and 32.4\u00a0wt% of paraffins.Even though a descriptive characterization of the catalysts has been already reported in a previous work [44], their main properties are shown in Table 2\n. In order to compare the properties of these industrial catalysts it must be taken into account its complex configuration, which is composed of an ultrastable Y zeolite (USY) embedded in a meso- and macroporous matrix (consisting of a mixture of clay, silica and alumina) [45]. The highest content of zeolite of ECAT-2 (21\u00a0wt%) is in concordance with its high micropore surface area (139\u00a0m2 g-1). ECAT-3, in turn, has the highest matrix/USY zeolite ratio, which turns into the highest mesopore surface area and mesopore volume (111\u00a0m2 g-1 and 172\u00a0cm3 g-1, respectively). This way, its wide porous structure increases the accessibility of the NH3 to the acid sites, making ECAT-3 the catalyst with the highest total acidity (124 \u03bcmolNH3 g-1), acid strength (130\u00a0kJ molNH3\n-1) and Br\u00f8nsted/Lewis acid sites ratio (1.56). Furthermore, the high mesopore volume of the matrix will reduce diffusional constraints of the long chains of PPO, easing its access to the external crystal surface of zeolites and, consequently, its posterior cracking on the channels of the zeolite.It is also remarkable the presence of rare earths in the catalysts, especially for ECAT-1 (2.50\u00a0wt%), since these elements increase the selectivity to naphtha lump [46]. The presence of P2O5 (with a maximum value of 0.62\u00a0wt% for ECAT-1) aims the formation of light olefins. Metals, such as V, Fe and Ni, are irreversibly deposited on FCC catalysts in the successive reaction-regeneration cycles acting as poisons and causing a reduction in throughput by increasing coke formation [47].In order to validate the model proposed, the results obtained with ECAT-1 for the parent reaction network (Fig. 1a) are shown below. The values of the apparent kinetic rate constants and activation energies that have minimized the mean squared error of the loss function (Eq. (9)) have been collected in Table 3\n. The values of the parameters provide a large amount of information about the relevancy of the different catalytic steps. This way, it can be seen that the steps that govern the catalytic cracking of PPO are those in which the cracking of HCO fraction is involved (steps #1 to #5 in Fig. 1a). The highest crackability of the compounds within the HCO fraction is a common result obtained in the catalytic cracking of hydrocarbon streams and it is coherent with the higher crackability of the high-molecular weight olefins [25,48]. However, the values of some of the kinetic rate constants, in particular those corresponding to steps #8 to #12, are so small that the contribution of these kinetic steps can be considered negligible.Therefore, three alternative reaction networks have been proposed (Fig. 1b, c and d) in which various simplifications have been made based on the results collected in Table 3. This way, in alternative network 2 (Fig. 1b) the naphtha fraction has been considered as a final product, i.e. steps #10, #11 and #12 have been removed from the parent network (Fig. 1a). In alternative network 3 (Fig. 1c), in turn, the steps removed have been those in which LCO fraction is converted into coke and into dry gas fractions (steps #8 and #9). Finally, alternative network 4, which is the simplest one from all the proposed, ignores all the routes removed in alternative networks 2 and 3. Thus, in alternative network 4 (Fig. 1d) steps #8 to #12 have been removed from the parent one (Fig. 1a).Consequently, the fitting of the experimental data obtained for the catalytic cracking of PPO with ECAT-1 has been also performed for the alternative reaction networks. Overall, good fitting results have been obtained for all of them. Hence, in order to perform an appropriate discrimination between the four networks, the statistical significance test based on Fisher\u2019s method described in Section 3.3 has been applied. The results obtained have been tabulated in Table 4\n. Attending to the statistical parameters, it can be seen that the number of experiments and lumps is the same for all the networks, but the number of parameters varies following the trend: np,1\u00a0>\u00a0np,3\u00a0>\u00a0np,2\u00a0>\u00a0np,4. Consequently, the degrees of freedom of the different reaction networks follows just the opposite order. On the other hand, the lowest value for the sum of squared errors has been obtained for scheme 3 (5.844 10-3), whereas the values obtained for the other networks are slightly higher and follow the trend: SSE2\u00a0<\u00a0SSE1\u00a0<\u00a0SSE4. Therefore, since alternative reaction network 4 is the simplest one (less amount of parameters) and the worst fitting has been obtained with it, this one has been taken as reference for the statistical comparison. Thus, it has been assessed if the addition of more catalytic steps is statistically significant. It has been obtained that F4\u20131\u00a0<\u00a0F1\u2212\u03b1 (0.841\u00a0<\u00a02.259), F4\u20132\u00a0<\u00a0F1\u2212\u03b1 (1.569\u00a0<\u00a03.040) and F4\u22123\u00a0<\u00a0F1\u2212\u03b1 (1.569\u00a0<\u00a02.649), meaning that neither parent network nor alternatives 2 and 3 improved in a statistically significant way the fitting of alternative scheme 4.Based on all the previous, the alternative reaction network 4 has been also used for the fitting of the data obtained with both ECAT-2 and ECAT-3. The goodness of fit has been evaluated using parity plots (Fig. S1 in the Supplementary Material) by evaluating the final fit of calculated data for the three catalysts against raw experimental data. As it can be seen, almost a perfect fit between the calculated and the experimental weight fraction has been obtained for all the catalysts, with the exception of the scattering of some points, especially for ECAT-3. Nevertheless, those deviations do not exceed the 5% as they remain inside the region delimited by the dashed lines.In Table 5\n have been collected the values computed for the apparent kinetic parameters and the activation energies of the kinetic steps involved in the catalytic cracking of PPO by the three catalysts. Overall, small differences have been obtained in the values of the kinetic parameters with all the catalysts. These differences lie in the properties of the catalysts, considering the effect of the acidity and porous structure on the activity, selectivity and deactivation of the catalysts [49]. This way, ECAT-2 has the highest value for the deactivation parameter (0.030\u00a0s-1) because of its moderate mesopore surface area (50\u00a0m2 g-1) and mesopore volume (147\u00a0cm3 g-1), which are not enough for easing the diffusion of coke precursors towards the external surface of catalyst particles. Likewise, the confinement of the precursors will block the micropores of the zeolite resulting in the ineffectiveness of its high content of zeolite [50]. It is well-established the role as coke precursors of light olefins in cracking processes, as they undergo oligomerization, aromatization and condensation reactions that are catalyzed by strong acid sites [51,52]. In the same line, do stand out the rate constants of the reactions that form coke from LPG and dry gas fractions (2.8 10-3 and 1.0 10-4 m3 kgcat\n-1 s-1, respectively) using ECAT-2.In contrast, the lower zeolite/matrix ratio of ECAT-3 that entails a higher mesopore surface area (111\u00a0m2 g-1) and a higher mesopore volume (172\u00a0cm3 g-1), will improve the diffusion of coke precursors, attenuating their confinement and, consequently, the blockage of the micropores. Moreover, the enhanced accessibility and diffusion of PPO chains to the active sites in ECAT-3 are in concordance with the high values of the kinetic parameters for the reactions that convert the components within the HCO lump into dry gases (1.16\u00a0m6 kgcat\n-1 kmol-1 s-1) and the components within the LCO lump into naphtha (1.0 10-3 m3 kgcat\n-1 s-1). The similarities among the rest of the kinetic parameters for the different catalysts lie in the synergistic and parallel effects of the porosity and acidity that boost the extent of the cracking reactions.The values computed for the apparent activation energy of the different catalytic steps have been collected in Table 6\n. Unlike the kinetic parameters (Table 5), significant differences are observed between the activation energy required in some of the catalytic steps for the different catalysts. Likewise, the energy barrier that must be overcome for the deactivation step is very different depending on the catalyst used. The lower activation energy of the deactivation stage (79.2\u00a0kJ\u00a0mol-1), added to the high value of the deactivation kinetic parameter obtained (3.0 10-2\u00a0s-1 in Table 5) expose the high tendency of ECAT-2 to be deactivated. Equally, the high amount of acid sites on ECAT-3 explains the low activation energy required for the steps of formation of dry gas from HCO and LPG lumps (17.1\u00a0and 40.0\u00a0kJ\u00a0mol-1, respectively). In addition, the high matrix mesoporosity of ECAT-3, which is the other key feature of the catalysts, reduces the activation energy of the steps limited by the diffusivity of the components. This way, this catalyst reduces the energy involved in the steps that convert the HCO in LCO, LCO in naphtha and LCO in LPG (60.5, 42.5 and 58.3\u00a0kJ\u00a0mol-1, respectively), as well as in the formation of coke from LPG and dry gas (4.4 and 40.7\u00a0kJ\u00a0mol-1, respectively).The evolution of the yields of products with high commercial interest as fuels (LCO and naphtha) and with high content of olefins (LPG) has been obtained for the three catalysts by computing the kinetic model previously described and using the kinetic parameters collected in Tables 5 and 6. One should note that the PPO fed to the reactor has a content of 12.5\u00a0wt% of LCO and a 5.5\u00a0wt% of naphtha, as it has been previously detailed in Section 4.1. Those contents have not been taken into account for depicting the evolution of the yields, in order to assess the formation of these lumps in its real magnitude. On the other hand, in order to fully understand the obtained results, it should be taken into account that conversion has been defined as the ratio of mass of HCO converted to lighter products and to coke to the mass of HCO fed:\n\n(12)\n\n\nConversion\n=\n\n\n\n\n(\n\n\nm\n\n\nHCO\n\n\n)\n\n\nPPO\n\n\n-\n\n\n(\n\n\nm\n\n\nHCO\n\n\n)\n\n\nProducts\n\n\n\n\n\n\n(\n\n\nm\n\n\nHCO\n\n\n)\n\n\nPPO\n\n\n\n\n\n100\n\n\n\n\nThe yield of each i lump of products has been defined as the mass of lump i referred to the total mass of lump HCO fed:\n\n(13)\n\n\n\n\nYield\n\n\ni\n\n\n=\n\n\n\n\nm\n\n\ni\n\n\n\n\n\n\nm\n\n\nHCO\n\n\n\n\n\n100\n\n\n\n\nTherefore, Figs. 2-4\n\n\n compare the evolution with conversion of the yield of LCO, naphtha and LPG, respectively, obtained with the three catalysts at different temperatures. In a previous work [19] it has been detailed the composition of these lumps, stressing out the interest of the naphtha lump (research octane number up to 105) for being added to the stream of gasoline in refinery. It is also remarkable the propylene-rich LPG lump produced.The trend of the curves in Fig. 2 exposes the character of the LCO lump as an intermediate in the reaction network [53], as they go through a maximum at values of conversion of ca. 80\u00a0wt%. Furthermore, it can be seen that high temperatures promote the cracking reactions that convert the molecules within this lump into lighter molecules, resulting in lower yields of LCO. Comparing the results obtained with the three catalysts, similarities are observed between the results obtained. This way, with ECAT-1 and ECAT-2 higher values than with ECAT-3 are obtained, yielding up to 46.5\u00a0wt% with the former catalysts at 500\u00a0\u00b0C.Attending to the evolution of the yields of naphtha and LPG lumps (Figs. 3 and 4, respectively), both are end-products in the reaction network since their yield increases continuously with the extent of conversion. Nonetheless, in spite of the evolution obtained for LPG lump (Fig. 4), the molecules within this lump are cracked to dry gas and condensed to coke as it has been previously obtained in the reaction network (Fig. 1d). With regard to the evolution of the yield of naphtha (Fig. 3), high temperatures promote the production of this lump, since the cracking of molecules within HCO and LCO lumps is boosted. Furthermore, higher yields of naphtha have been obtained with ECAT-3, yielding up to 33.6\u00a0wt% at 560\u00a0\u00b0C. However, the maximum values obtained with ECAT-2 and ECAT-1 have been slightly inferiors (31.1 and 29.6\u00a0wt%, respectively).The evolution of the different yields (Figs. 2-4) strongly depends on the properties of the catalyst used (Table 2) and can be correlated with the values of the apparent kinetic parameters reported on Table 5. This way, ECAT-3 is by far the catalyst with the highest and strongest acidity (124 \u03bcmolNH3 g\u22121 and 130\u00a0kJ molNH3\n-1, respectively), which turns into the catalyst with the highest cracking activity. Moreover, it is the catalyst with the highest mesoporosity that eases the access of the bulky molecules within the HCO and LCO lumps to the acid sites located in the inside of the porous structure of the catalyst. Therefore, the highest yields of naphtha and LPG, together with the lowest yield of LCO should be expected using this catalyst.Even though ECAT-2 possesses a lower amount of acid sites available (81 \u03bcmolNH3 g\u22121), their strength is quite remarkable (126\u00a0kJ molNH3\n-1) making a priori this catalyst a serious candidate for maximizing the yield of naphtha and LPG lumps. However, its microporous nature and its, subsequent, shortness in mesopores are unsuitable for boosting the access of the heavy molecules to inner acid sites. Consequently, the behavior of ECAT-2 is only comparable with ECAT-3 at 560\u00a0\u00b0C as an increase in temperature increases the diffusivity [54]. Nonetheless, ECAT-2 promotes the formation of LPG instead of naphtha, which can be attributed to the overcracking reaction that takes place within the micropores of the zeolite as a consequence of the higher residence time of the reactants. Its low content of rare earths will also presumably contribute to obtain the aforementioned results [46]. Furthermore, the narrower porous structure of ECAT-2 will lead to a faster activity decay of the catalyst.Finally, the configuration and composition of ECAT-1 are the less favorable ones to promote the cracking reactions. Indeed, ECAT-1 has the lowest superficial area (124\u00a0m2 g\u22121), the lowest acidity (40 \u03bcmolNH3 g\u22121) and the weakest acid strength (100\u00a0kJ molNH3\n-1). In addition, the high concentration of impurity metals detected on ECAT-1, especially of vanadium (3335\u00a0ppm), will also contribute to deteriorate the properties of the catalyst. Consequently, slightly lower yields of both naphtha and LPG lumps (Figs. 3 and 4) have been obtained with ECAT-1.The selectivity to each lump i has been defined as the mass of lump i formed respect to that of all the products:\n\n(14)\n\n\n\n\nSelectivity\n\n\ni\n\n\n=\n\n\n\n\nm\n\n\ni\n\n\n\n\n\n\nm\n\n\nLCO\n\n\n+\n\n\nm\n\n\nNaphtha\n\n\n+\n\n\nm\n\n\nLPG\n\n\n+\n\n\nm\n\n\nDryGas\n\n\n+\n\n\nm\n\n\nCoke\n\n\n\n\n\n100\n\n\n\n\nTaking into account that naphtha and LPG lumps are the ones with the highest commercial interest, the evolution with conversion of the selectivity to them has been depicted in Fig. 5\n. Overall, it can be seen how different the selectivity to each lump is. This way, the selectivity to naphtha lump is barely affected by the extent of conversion. The trend followed by the selectivity to naphtha curves depends on the catalyst. This way, it can be seen that with ECAT-1, which is the less active catalysts, the selectivity to naphtha remains almost steady for values of conversion below 80%, to increase exponentially at higher values. For ECAT-3, in turn, the growth can be noticed for values of conversion above 55%. Furthermore, the differences between the performances of the catalysts are more evident at high temperatures and high values of conversion. This way, the selectivity to naphtha has been maximized at 560\u00a0\u00b0C with ECAT-3 reaching a value of 35.8\u00a0wt%, whereas under the same conditions a selectivity of 31.5\u00a0wt% has been obtained with ECAT-1. ECAT-2, in turn, offers an intermediate result and a selectivity to naphtha of 32.8\u00a0wt%.In contrast, selectivity to LPG lump grows exponentially with conversion since the very beginning of the reaction. The effect of the temperature is less marked but the opposite in the case of the LPG. Likewise, an increase from 500 to 560\u00a0\u00b0C for a fixed value of conversion entails just a reduction of the selectivity to LPG of ca. 2.5\u00a0wt%. Focusing on the performance of the catalysts, ECAT-2 offers the highest selectivity at 560\u00a0\u00b0C but also the lowest at 530 and 500\u00a0\u00b0C. This result is characteristic of a partially deactivated catalyst, in which thermal cracking plays a more important role that in the case of ECAT-1 and ECAT-3.Attending to the results collected on Figs. 2-5, operating under the conditions that allow for reaching conversions levels within the range 60\u201380% would be the optimal considering the possibility of varying the reaction temperature between 500 and 560\u00a0\u00b0C. This way, the conversion of HCO would be promoted keeping under control the overcracking reactions that would lead to obtain too much dry gas. Furthermore, ECAT-3 should be the selected one for turning the production to LPG and naphtha lumps, whereas ECAT-1 and ECAT-2 would increase the yield of LCO lump in detriment to the yield of naphtha.Since catalyst deactivation has a notable impact on the results collected in Sections 4.4 and 4.5 about the yields and selectivity, the evolution of the activity term (\u03c6) of the three catalysts with contact time at 500\u00a0\u00b0C has been plotted on Fig. 6\n. One should note that these curves have been obtained by applying the previously proposed deactivation equation (Eq. (7)) and using the corresponding kinetic parameters (Tables 5 and 6). It can be seen that for a contact time of 6\u00a0s, all the catalysts maintain good activity levels as they are above of the 83% of the initial activity. This result is very different to that obtained in the cracking of VGO (benchmark feed in FCC) [55], where the catalyst was totally deactivated. This result exposes the crucial role that the composition of the stream fed to cracking reactor plays in catalyst deactivation and, therefore, in products yield and distribution. This way, the heterogeneity of the VGO, with high contents of aromatics and the presence of polyaromatics, is more prone to the formation of coke than the olefins that predominate in the composition of the PPO [43]. This low deactivation is an interesting result for adopting different cracking strategies for the PPO, such as being co-fed with other refinery streams that deactivate the catalysts in a large extent.Comparing the evolution followed by the catalysts, ECAT-1 and ECAT-3 show almost identical curves of activity vs. time as a difference of<1% for a contact time of 6\u00a0s (ca. 87%) has been obtained. However, ECAT-2 suffers from a higher and more severe activity decay since the very beginning of the reaction. Indeed, the final value for activity obtained for this latter catalyst is of 83.5%. Undoubtedly, the deactivation suffered by ECAT-2 lies in the porous structure of the catalyst, which is by far more microporous (Table 2) than the structure of the other catalysts. Consequently, the coke formed during the reaction will more easily block the channels of the zeolite reducing the accessibility of hydrocarbon species to the catalyst inner micropore network [56].To offer another perspective of the deactivation results, Fig. 7\n depicts the evolution of the activity of the three catalysts with the content of coke deposited. Clearly, the amount of coke deposited on ECAT-2 is higher than that deposited on ECAT-1 or ECAT-3. Consequently, ECAT-2 suffers from a higher activity decay than the other catalysts. In spite of that, attending to the accelerated decrease of the activity obtained for all the catalysts, it can be concluded that the deactivation mechanism is highly affected by the micropore blocking caused by coke deposition. This phenomenon will also restrict the access of the reactants to the acid sites located in the inner crystals of the zeolite. This result is in concordance with the hypothesis of the key role of the matrix mesopores for attenuating the catalyst deactivation by delaying the aforementioned phenomenon.A lumped kinetic modeling method has been applied to the experimental data of the catalytic cracking of plastic pyrolysis oil (PPO) over three commercial FCC equilibrium catalysts. By means of a statistical data analysis, it has been obtained that from the four different reaction networks proposed, the simplest one was the most appropriate for describing the process. From the kinetic parameters obtained in the fitting of the results, it has been obtained that both total acidity and acid strength rule the cracking process, boosting the extent of the different reaction steps and modifying the distribution of the lumps of products. Furthermore, the mesoporous structure of the matrix is a key feature for reducing the diffusional restrictions and, subsequently, for maximizing the formation of the naphtha and LPG lumps. This way, the maximum yield and selectivity to naphtha of 33.6 and 35.8\u00a0wt%, respectively, have been obtained with ECAT-3 for a conversion value of 94%. In contrast, ECAT-1 and ECAT-2 promote the formation of LCO instead of naphtha.The deactivation of the three catalysts in the cracking of the PPO is by far lower than that obtained in the cracking of VGO (benchmark feedstock of FCC unit), because of the absence of aromatics in the PPO. Likewise, for a contact time of 6\u00a0s the catalysts keep a residual activity above the 80%. The lowest deactivation of ECAT-3 (kd\u00a0=\u00a02.3 10-2 s\u22121) has been related to the high mesoporosity of its matrix, which is appropriate for promoting the internal diffusion of coke precursors, attenuating the catalyst deactivation. This way, for this catalyst, the apparent activation energies of the conversion of heavy cycle oil (HCO) into light cycle oil (LCO), LCO into naphtha, and LCO into liquefied petroleum gases (LPG) are 60.5 42.5 and 58.3\u00a0kJ\u00a0mol\u22121, respectively. In addition, those of the formation of coke from HCO, LPG and dry gas are 129.0, 4.4 and 40.7\u00a0kJ\u00a0mol\u22121, respectively.The kinetic model proposed is an interesting tool for facing the manufacturing of reactors designed ad hoc for the catalytic cracking of PPO. Additionally, obtained results could also encourage the future co-feeding of this alternative and waste-derived feedstock to industrial FCC units commonly available in oil refineries. Nevertheless, the kinetic results could be modified by the presence of additives and pollutants in the waste plastics.\nRoberto Palos: Formal analysis, Conceptualization, Writing \u2013 original draft. Elena Rodr\u00edguez: Investigation, Formal analysis. Alazne Guti\u00e9rrez: Supervision, Methodology, Visualization. Javier Bilbao: Conceptualization, Writing \u2013 review & editing, Supervision, Project administration, Funding acquisition. Jos\u00e9 M. Arandes: Software, Resources, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work has been carried out with the financial support of the Ministry of Science, Innovation and Universities (MICIU) of the Spanish Government (grant RTI2018-096981-B-I00), the European Union\u2019s ERDF funds and Horizon 2020 research and innovation programme under the Marie Sk\u0142odowska-Curie Actions (grant No 823745) and the Basque Government (grant IT1218-19). Dr. Roberto Palos thanks the University of the Basque Country UPV/EHU for his postdoctoral grant (UPV/EHU 2019). The authors also acknowledge Petronor Refinery for providing with the catalyst used in this work.Supplementary data to this article can be found online at https://doi.org/10.1016/j.fuel.2022.123341.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  The kinetics of the catalytic cracking of plastic pyrolysis oil (PPO) over three FCC (fluid catalytic cracking) equilibrium commercial catalysts has been modeled. The PPO comes from the fast pyrolysis of high-density polyethylene (HDPE). The cracking runs have been carried out in a laboratory-scale reactor under FCC conditions: 500\u2013560\u00a0\u00b0C; catalyst/oil weight ratio of 5 gcat gPPO\n                     -1; and contact time of 1.5\u20136\u00a0s. Four different reaction schemes composed of six lumps have been compared and it has been obtained by statistical means that the simplest one is the most appropriate for describing the process. The differences in the kinetic parameters have been related to the properties of the catalysts. Among them, total acidity and mesoporous structure have a key role. The former for promoting the cracking reactions and the latter for limiting the diffusional restrictions of both the bulky compounds within the PPO and the formed coke precursors. This way, ECAT-3 that is the most acid and most mesoporous catalyst, maximizes the yields of naphtha (33.6\u00a0wt%) and liquefied petroleum gases (LPG) (18.9\u00a0wt%). In contrast, ECAT-1 and ECAT-2 should be chosen for producing light cycle oil (LCO). For ECAT-3, the apparent activation energies of the conversion of heavy cycle oil (HCO) into light cycle oil (LCO), LCO into naphtha, and LCO into LPG are 60.5 42.5 and 58.3\u00a0kJ\u00a0mol-1, respectively. In addition, those of the formation of coke from HCO, LPG and dry gas are 129.0, 4.4 and 40.7\u00a0kJ\u00a0mol-1, respectively.\n               "}
{"full_text": "As energy crisis and environmental pollution are two major enemies for mankind, developing new energy sources and realizing carbon neutrality have become an urgent task in the 21st century. Hydrogen is considered to be the ultimate form of energy in the future because of its advantages such as high calorific value, pollution-free, and wide sources [1,2], but how to store and use it safely has always troubled everyone.Solid hydrogen storage, which is a promising hydrogen storage technology, owns the advantages of high density of hydrogen storage in weight and volume, moderate operating pressure and high energy efficiency [3,4], compared to high-pressure gas hydrogen storage with low hydrogen storage efficiency, poor safety and to low-temperature liquid hydrogen storage with high cost and large energy consumption [5]. In recent decades, many solid hydrogen storage materials have been discovered and studied, which can be divided into chemical hydrogen storage and physical hydrogen storage according to different hydrogen storage methods. Metal hydrides belong to chemical hydrogen storage, which can store hydrogen in the form of hydrogen atoms in solid hydrogen storage materials, and exhibit excellent safety and cycle performance. Magnesium based hydrogen storage alloys have been extensively studied in the field of hydrogen storage [2,6,7] due to their high hydrogen storage capacity (7.6\u00a0wt.% for MgH2), low cost and abundant resources. Nevertheless, the poor thermodynamic and kinetic barrier of Mg-based hydrogen storage alloys lead to high working temperature and slow kinetics, which limits its practical application in the field of hydrogen energy [8,9]. The modification methods, such as alloying [10], nano-crystallization [11,12] and catalyst addition [13,14], have been used to improve the performance of Mg-based hydrogen storage materials. Noteworthy, alloying can effectively lower the thermodynamic and kinetic barriers. Among various alloying elements, transition metal elements (TM) promote the dissociation of H2 molecules and play a synergistic catalytic role in improving the hydrogen absorption and desorption performance of magnesium- based hydrogen storage alloys [15\u201318]. Rare earth elements (RE) can also promote the hydrogen absorption and desorption properties of the alloys [19\u201322].In recent years, Mg-Ni-Y hydrogen storage alloys have been widely studied. The transition metal element Ni in the alloys can reduce the activation energy of H2 decomposition during the process of hydrogen absorption [23], and Mg2Ni generated by reaction with Mg can be used as a stable catalyst during hydrogen ab- and desorption reaction [24\u201326]. YH2 produced by rare earth element Y and H2 can also catalyze the reaction of hydrogenation [27\u201329]. Remarkably, long period ordered stacking phase (LPSO phase) can be formed in Mg-Ni-Y hydrogen storage alloys [30]. After hydrogen absorption, the LPSO phase decomposes and forms nano-sized Mg2NiHx, MgH2 and YHX, which can reduce the hydrogen diffusion path and promote the activation and the kinetics of hydrogen ab- and desorption in Mg-Ni-Y hydrogen storage alloys [31\u201333]. Whereas, the influence of size, distribution and type of LPSO phase on hydrogen storage performance remains not very clear.In this work, Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys were prepared by melting and ball milling. The microstructure and hydrogen ab- and desorption properties of the alloys were analyzed. The effect of the size, distribution of LPSO phase and ternary eutectic microstructure in the alloys with different contents of Ni and Y on hydrogen uptake and release properties were studied in detail. The hydrogen absorption mechanism of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys with LPSO structure is also discussed in this work. It is found that the catalytic effect is better when the catalytic phases are more evenly dispersed in the magnesium matrix. The enhancement of in-situ decomposition of coarse LPSO phases may be lower than that of fine Mg-Mg2Ni eutectic structure. So the size and distribution of phases and the number of phase boundaries may be more conducive to improve the hydrogen storage performance of the alloy. It is believed that this paper can provide us with an idea that the improving effect of LPSO phases is limited, reducing the size of LPSO phases and eutectic structure can be beneficial to further enhance the hydrogen storage performance of Mg-Ni-Y alloy.The character of raw materials is shown in Table\u00a01\n. The Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 casting ingots are prepared by melting in induction furnace in SF6 atmosphere and cooling in a salt solution. The casting ingots were then broken into powder, and further refined in the ball mill (pulverisette5 planetary ball mill, Feichi company, Germany). The rotation speed of the ball mill is 250\u00a0rpm (positive and negative rotation cycle), the mass ratio of grinding ball to material is 20:1, and the ball grinding time lasts 7\u00a0h.The phase structures of the as-cast, ball milled, hydrogenated and dehydrogenated alloys were determined by an X-ray diffractometer (XRD, RIGAKUD/max250pc) with Cu K\u03b1 radiation, at a diffraction angle (2\u03b8) from 10\u00b0 to 90\u00b0 at 4\u00b0 min\u22121. Scanning electron microscopy and corresponding elemental analysis were performed on a field emission scanning electron microscope (JEOL JSM-7800F) equipped with an energy dispersive X-ray spectrometer (EDS), which perform microanalysis of specific areas. LPSO phases and hydrides were identified via transmission electron microscopy (TEM, Talos F200s, Czech).Non-isothermal dehydrogenation behaviors of hydrides were performed using differential scanning calorimetric (DSC, STA449F3, NETZSCH, Germany) at the heating rate of 5\u00a0\u00b0C/min, 10\u00a0\u00b0C/min and 15\u00a0\u00b0C/min ranging from room temperature to 500\u00a0\u00b0C. The hydrogen absorption kinetics curves and PCT curves at different temperatures were measured by high pressure gas adsorption instrument (Beth company 3h-2000\u00a0pH) at the initial pressure 3\u00a0MPa.\nFigure\u00a01\n shows the XRD patterns of as-cast, ball milled, hydrogenated and dehydrogenated Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys. The as-cast alloys consist of \u03b1-Mg, Mg2Ni, LPSO [27,32,34] according to Fig.\u00a01. Due to the increasing Y content in Mg92.8Ni2.4Y4.8 alloys, Mg24Y5 phase is formed while Mg2Ni disappears compared with Mg91.4Ni7Y1.6 alloys. After ball milling, the diffraction peaks of various phases are broadened, which indicates that the plastic deformation of the alloy results in lattice distortion and grain size refinement. The refinement of grain size is beneficial to enhance the hydrogen absorption and desorption properties. In addition, the types of phases in the alloy do not transform, and the LPSO phases still exist after ball milling. After hydrogen absorption, five kinds of hydrogen absorbing phases can be found, namely MgH2, Mg2NiH4, YH3, YH2 and Mg2NiH0.3. After releasing hydrogen, the XRD analysis shows that the dehydrogenated alloys are mainly composed of Mg, Mg2Ni and YH2. The LPSO phases and Mg24Y5 phases do not form any more, indicating that the decomposition of LPSO phase and Mg24Y5 phases are irreversible.\nFigure\u00a02\n shows the SEM images of as-cast Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys with different magnification. In the Mg91.4Ni7.1Y1.6 alloy, the black elliptical pre-precipitated phase and lamellar eutectic structure can be clearly observed. When the atomic ratio of Ni to Y is greater than 0.5, there are three phases of \u03b1-Mg, LPSO and Mg2Ni in Mg-Ni-Y alloy, and eutectic reaction occurs during solidification:L\u2192\u03b1-Mg+LPSO+Mg2Ni [29]. So occurs the eutectic reaction in Mg91.4Ni7Y1.6 alloys with Ni/Y atomic ratio of 4.4. In addition, based on the XRD data and the SEM microstructure in Fig.\u00a02(a), the Mg91.4Ni7Y1.6 alloys consist of black Mg matrix and a large number of gray eutectic structures, which is composed of fine lamellar LPSO phases, Mg2Ni and \u03b1-Mg alternating layers. The TEM patterns of LPSO phase regions in Mg91.4Ni7Y1.6 alloy is observed in Fig.\u00a03\n. The results of selected area electron diffraction and HRTEM micrograph analysis show that the lamellar LPSO phase in Mg91.4Ni7Y1.6 alloys is 14H-type LPSO structure. The microstructure of the Mg92.8Ni2.4Y4.8 alloy observed by Scanning Electron Microscopy is shown in Fig.\u00a02(b). The Mg92.8Ni2.4Y4.8 alloy is made up of black Mg matrix, a large number of bulk LPSO phases and gray eutectic structure, and the LPSO phases, Mg24Y5 and \u03b1-Mg are distributed alternatively in the small eutectic structure. The TEM patterns of LPSO phase regions in Mg92.8Ni2.4Y4.8 alloy is observed in Fig.\u00a04\n. The results of selected area electron diffraction and HRTEM micrograph analysis confirms the bulk LPSO phase in Mg92.8Ni2.4Y4.8 alloys belong to 18R-type LPSO structure. In Mg92.8Ni2.4Y4.8 there are less eutectic phase and more bulk 18R-LPSO, whose width is more about 10\u00a0\u00b5m or even wider. Whereas the lamellar 14H-LPSO phase is more and finer, its width is about 200\u2013500\u00a0nm, and the content of ternary eutectic areas is more in Mg91.4Ni7Y1.6 alloy.In summary, there will be more grain boundaries and phase boundary where it is easier to diffuse for hydrogen atoms. Therefore, kinetic properties of hydrogen absorption and desorption of Mg91.4Ni7Y1.6 alloys will be more excellent.\nFig.\u00a05\n shows the SEM pictures of the particle size and surface morphology of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys after ball milling under the same conditions. According to Fig.\u00a05, the particles of Mg91.4Ni7Y1.6 alloy are finer and there are more cracks and other defects on the surface compared with the ones of Mg92.8Ni2.4Y4.8 alloy, which can provide surface where H2 can be adhered and increase the diffusion rate of hydrogen atoms.\nFigire\u00a06 and Fig.\u00a07\n\n show the SEM images of Mg91.4Ni7Y1.6 alloys and Mg92.8Ni2.4Y4.8 alloys after hydrogenation and dehydrogenation, respectively. It is found in Fig.\u00a06 and Fig.\u00a07 that the particle size of Mg91.4Ni7Y1.6 alloy is significantly smaller than that of Mg92.8Ni2.4Y4.8 alloy. Moreover, the particle surfaces of Mg92.8Ni2.4Y4.8 alloys after hydrogen absorption is relatively dense, and there are fewer fine particles on the alloy surfaces after hydrogenation and dehydrogenation. However, lots of flocculating particle clusters appear on the surfaces of Mg91.4Ni7Y1.6 alloys after hydrogen adsorption. These nano-sized particles can improve the hydrogen absorption and desorption dynamics of the alloy because it shortens the diffusion distance of H atoms. The TEM, energy spectrum, high resolution electron microscope and selected area electron diffraction pattern of a hydrogenated particle can be observed in Fig.\u00a08 and Fig.\u00a09\n\n. The EDS results show that Mg, Ni and Y elements are slightly evenly distributed in the alloy particles after hydrogen absorption. The calibration results of electron diffraction pattern and high resolution electron microscope data confirm also that MgH2, Mg2NiH4, Mg2NiH0.3 and YH3 phases are uniformly distributed in the alloy particles after hydrogen absorption. After hydrogen desorption, Nano-sized compounds, such as Mg2Ni and YH2, are evenly dispersed on the surface of the alloy particles. According to Nobuko, Hanada et\u00a0al. [35], transition metal and Rare earth metal nano-sized compounds uniformly dispersed on the surface of MgH2 can modify the surface condition of the particle and greatly reduce the activation energy of hydrogen desorption on the alloy surface. Therefore, Nano-sized hydride on the surface of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys can not only absorb hydrogen, but also play the important role of surface modification, leading to better kinetic properties of hydrogenation and dehydrogenation of the alloys.The DSC curves of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 during hydrogen desorption at various heating rates are presented in Fig.\u00a010\n. As one can see, three endothermic peaks of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys can be found. The hydrogen desorption temperature of Mg2NiH4 is lower than that of MgH2 and YH3\n[36\u201338]. In the hydrogen desorption process, Mg2NiH4 first desorbs hydrogen to Mg2NiH0.3 and undergoes a significant volume contraction, causing a contraction strain on MgH2 around it, facilitating hydrogen desorption of MgH2. Hence, the desorption processes of primary MgH2 and Mg2NiH4 are not isolated, the synergistic reaction of them displays the lowest desorption peak temperature: Mg2NiH4 \u2194 Mg2NiH0.3\u00a0+\u00a02H2, MgH2\u2194Mg+H2. The dehydrogenation of Mg2NiH0.3 into Mg2Ni when the decomposition of MgH2 is complete, so the dehydrogenation process of Mg2NiH0.3 will occur at a higher temperature: Mg2NiH0.3\u2194Mg2Ni+H2. The dehydrogenation temperature of YH2 and YH3 is generally around 790\u00a0\u00b0C and 400\u00a0\u00b0C, respectively [36,38]. Therefore, DSC curve shows no heat absorption peak of YH2 and the highest desorption peak temperatures is YH3\u2194YH2+H2. When the heating rate is 5 \u2103/min, the onset and peak decomposition temperatures of Mg91.4Ni7Y1.6 hydride are only 210.3\u00a0\u00b0C and 237.7\u00a0\u00b0C, both of which are lower than that of Mg92.8Ni2.4Y4.8 hydride: 227.9\u00a0\u00b0C and 248.2\u00a0\u00b0C, which indicates that more Ni doped in the alloys would contribute to enhancing thermodynamic property.\nFigure\u00a011\n show the fitted Kissinger curves of the Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys. The DSC curves of the synergistic catalytic de-/hydrogenation reaction between MgH2 and Mg2NiH4 at different heating rates were analyzed. The endothermic peak temperatures of Mg91.4Ni7Y1.6 alloy are 230.7 \u2103, 248.3\u00a0\u00b0C and 262\u00a0\u00b0C, and that of Mg92.8Ni2.4Y4.8 alloy are 248.2\u00a0\u00b0C, 266.6\u00a0\u00b0C and 273.1\u00a0\u00b0C, respectively. The activation energies of hydrogen desorption of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys were calculated by fitting the straight line ln(\u03b2/Tm\n2)vs.1/Tm through Kissinger equation [39]. As shown in Fig.\u00a010, the hydrogen desorption activation energies of Mg91.4Ni7Y1.6 alloy and Mg92.8Ni2.4Y4.8 alloy are 87.7 and 112.4\u00a0kJ/mol H2, respectively, which are lower than that of [40] ball-milled pure MgH2, 250\u00a0kJ/mol H2. Combined with SEM and TEM analysis, the synergistic catalysis of uniformly distributed YH2, YH3, Mg2NiH0.3 and Mg2NiH4, which are generated by hydrogen absorption of cluster-parallel LPSO phases, gives rise to in-situ element catalysis and nano-scale effect [41]. Thus, these kinds of effect would greatly reduce the activation energy of hydrogen release and improve dehydrogenation performance.The experimental results are consistent with the analysis results of microstructure characteristics observed by SEM, that is, the more eutectic phase, the finer and more dispersed LPSO phase, leading to better in-situ catalytic effect after hydrogen absorption.The de-/hydrogenation process of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys includes three chemical reactions:\n\n\n\nMg\n+\n\nH\n2\n\n\u21cc\nMg\n\nH\n2\n\n\n\n\n\n\n\n\n\nM\n\ng\n2\n\nN\ni\n+\n2\n\nH\n2\n\n\u21cc\nM\n\ng\n2\n\nN\ni\n\nH\n4\n\n\n\n\n\n\n\n\n\n2\nY\n\nH\n2\n\n+\n\nH\n2\n\n\u21cc\n2\nY\n\nH\n3\n\n\n\n\nThe respective de-/hydrogenation plateau pressures of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys at different temperatures can be obtained from the PCT curves in Fig.\u00a012\n(a) and (b). The low- plateau pressures corresponds to the transformation of Mg/MgH2 while the high-plateau pressures relates to the transformation of Mg2Ni/Mg2NiHx. Due to the low content of Ni element in Mg92.8Ni2.4Y4.8 alloy, the synergistic catalytic effect is weak, resulting part of the hydrogen is still retained in Mg92.8Ni2.4Y4.8 alloy at 330\u00a0\u00b0C, indicating that its dehydrogenation temperature under 3\u00a0MPa pressure is higher than 330\u00a0\u00b0C. Moreover, there is only one de-/hydrogenation platform, and the higher Mg2Ni platform was not obvious. In addition, we found that the reversible hydrogen storage capacity of the two alloys at 380\u00a0\u00b0C is slightly less than the hydrogen storage capacity at 350\u00a0\u00b0C and 330\u00a0\u00b0C. Combined with the analysis of the hydrogen absorption kinetic mechanism on page 15 of the article, with the increase of absorption temperature, the hydride nucleation rate decreases. There are more hydride grains with large sizes and a thicker hydride layer [42] during the hydrogenation of the Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys at 380 \u2103, which are not conducive to the diffusion of hydrogen atoms, resulting in a lower hydrogen storage capacity. Wenjie Song [41] also reported the similar situation.The corresponding van't Hoff plots for both hydrogen absorption and desorption are shown in Fig.\u00a013\n(a) and (b). According to the fitting result, the formation enthalpy changes (\u0394H) of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 hydride are calculated to be \u221260.6 and \u221268.6\u00a0kJ/mol H2, and the formation entropy changes (\u0394S) are 105.5 and 116.2\u00a0J/K/mol H2, respectively. According to Fig.\u00a012, The Mg92.8Ni2.4Y4.8 alloy release hydrogen incompletely at 330\u00a0\u00b0C, and the complete dehydrogenation PCT curve was not obtained. It is difficult to distinguish the hydrogen release platform. Therefore, it is difficult to calculate the enthalpy change and entropy change of dehydrogenation through van't Hoff plots. In addition, the dehydrogenation capacity of the Mg91.4Ni7Y1.6 alloy is more significant, so the dehydrogenation enthalpy change and entropy change of the Mg92.8Ni2.4Y4.8 alloy can be omitted. The decomposition enthalpy and entropy changes of\u00a0Mg91.4Ni7Y1.6 alloys are 56.9\u00a0kJ/mol H2 and 97.9\u00a0J/K/mol H2, separately.These values are all lower than that of MgH2 theoretical values [43], which indicates that the addition of Y and Ni reduce the thermodynamic stability of MgH2 and show better enhancement effects than just adding Ni or Y. Because of the large atomic radius of Ni and Y elements, the Mg-H bond distance is increased, which reduces the binding energy, thus reducing the stability. In addition, the activation energy and enthalpy of the hydrogen storage alloy will be reduced due to the nanostructure, which will greatly improve the hydrogen storage performance of the alloy [44]. In Mg91.4Ni7Y1.6 alloy, the LPSO phase is decomposed to form the finer and more dispersed nanoscale catalytic phases, and the value of hydrogen absorption enthalpy decreases more than that of Mg92.8Ni2.4Y4.8 alloy. This indicated that the Mg91.4Ni7Y1.6 alloy presents better hydrogen absorption performance and lower de-/hydrogenation temperature.\nFigure\u00a014\n gives the hydrogen absorption kinetics curves of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys. The amount of hydrogen absorbed and the rate of hydrogen absorption reaction increase with the rising temperature. The hydrogen absorption kinetic curve of the alloy is composed of two parts, which are fast hydrogen absorption and stable hydrogen absorption. The hydrogen absorption rate of the alloy rises steeply in the first 5\u00a0min, which is the rapid hydrogen absorption. However, the hydrogen absorption of the alloy shows a slow upward trend in the stable hydrogen absorption stage. The reason is as follows: after the hydride layer is formed on the surface of the particles, further hydrogenation of the alloy requires H atoms to penetrate through the hydride layer and diffuse into the alloy particles. However, the diffusion rate of H atoms in the hydride layer is lower than that in metallic magnesium, which causes the hydrogen absorption rate of the alloy to continuously decrease until it can no longer continue to absorb hydrogen when the size of the hydride layer grows to a certain thickness. Therefore, the hydrogen storage capacity of the alloy is determined by the rapid hydrogen absorption stage, and the high hydrogen storage capacity in the rapid hydrogen absorption stage results in a high total hydrogen storage capacity of the alloy. Representative hydrogen absorption data for Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys are summarized in Table\u00a02\n. The data shows that Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloy can absorb 5.78\u00a0wt.% and 3.84\u00a0wt.% hydrogen respectively under an initial hydrogen pressure of 3\u00a0MPa at 350\u00a0\u00b0C in 5\u00a0min, reaching 90% and 78% of the maximum hydrogen absorption capacity. Apparently, the theoretical hydrogen uptake of Mg91.4Ni7Y1.6 alloy is lower than that of Mg92.8Ni2.4Y4.8 alloy, but Mg91.4Ni7Y1.6 has a higher maximum hydrogen uptake and a faster hydrogen absorption rate.To explain the hydrogen absorption kinetic mechanism of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys, the Johnson-Mehl-Avrami-Kolmogorov (JMAK) model are adopted to analyze the evolution of kinetics [45]:\n\n\n\n\n\n[\n\n\u2212\nl\nn\n\n(\n\n1\n\u2212\n\u03b1\n\n)\n\n\n]\n\n\n1\n/\nn\n\n\n=\nk\nt\n\n\n\n\nEquation transformation:\n\n\n\n\u03b1\n=\n1\n\u2212\ne\nx\np\n\n(\n\n\u2212\nk\n\nt\nn\n\n\n)\n\n\n\n\nwhere \u03b1 is the reaction fraction of the hydrogen storage material converted to hydride corresponding to time, k is the reaction rate constant, n (commonly between 0\u223c3) is the Avrami exponent of reaction order. When the value of 0.5<n<1.5, the nucleation and growth mode of the alloy is three-dimensional growth. After nucleation, the core takes a spherical shape and grows in a three-dimensional manner. When the spherical hydrides grow up and collide with each other, a continuous hydride layer is formed. In contrast to the value of n<0.5, the nucleation and growth method of the alloy is one-dimensional growth, and the core-shell structure will be formed in Mg-MgH2 transition, and the rate of hydrogenation is one to two orders of magnitude slower [31].The hydrogenation curves of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloy at 100\u00a0\u00b0C, 200\u00a0\u00b0C and 300\u00a0\u00b0C under 3\u00a0MPa are analyzed by JMAK model, which is shown in Fig.\u00a015\n (a-c) and Fig.\u00a016\n\n(a-c). The fitted curve is in good agreement with the experimental data (R2>0.99). At 100 \u2103, the hydride nucleation and growth of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys are all three-dimensional diffusion-controlled methods. However, at 200 \u2103, Mg91.4Ni7Y1.6 alloy forms a cored shell structure while Mg92.8Ni2.4Y4.8 alloy still grows at the three-dimensional manner, and the cored shell structure in Mg91.4Ni7Y1.6 alloy is formed at a lower temperature, indicating the nucleation rate of the alloy is very fast at a lower temperature. The rapid nucleation promotes the alloy to absorb a large amount of hydrogen during the rapid hydrogen absorption stage, and the hydride core quickly grows to form a hydride layer.Temperature greatly affects the hydrogen absorption reaction rate. When the temperature is low, the reaction rate is not fast enough to form cored shell structure. After rising to a certain temperature, nucleation and growth rate of alloy is accelerated, the core-shell structure is formed, and this gives rise to reducing the rate of reaction of materials. As grain boundaries and formed compounds act as pathways for hydrogen diffusion [46], interfacial diffusion can be maintained at higher temperatures. Therefore, although the Mg91.4Ni7Y1.6 alloy has a high nucleation rate and is easy to form a core-shell structure, its powder particles are fine, and a large number of fine LPSO phases and phase interfaces are contained in the alloy, which can provide diffusion path for H atoms and reduce the diffusion distance of H atoms, thereby improving the hydrogen storage performance of Mg91.4Ni7Y1.6 alloy.As we all know, hydrogen storage alloys need to overcome a certain energy barrier in the hydrogenation process in order to absorb hydrogen. This energy barrier is called the activation energy, which can be obtained by the Arrhenius equation. According to the Arrhenius equation, the activation energy of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys can be calculated by lnk vs. 1000/RT fitting, their hydrogen absorption activation energy was reduced from 100\u00a0kJ/mol H2 of pure Mg [47] to 25.4\u00a0kJ/mol H2 and 28.6\u00a0kJ/mol H2, respectively, as shown in Fig.\u00a015(d) and Fig.\u00a016(d). Through the analysis of hydrogen absorption kinetics curve and activation energy, Mg91.4Ni7Y1.6 alloy has a better hydrogen absorption kinetics performance. Adding different amount of Ni and Y element, Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys have different structure after smelting. Compared with the Mg92.8Ni2.4Y4.8, the Mg91.4Ni7Y1.6 alloy has finer LPSO phases and more eutectic structures, and the hydrogenated YH2/YH3 and Mg2NiHx phases are more dispersed, contributing to better synergistic catalytic effect and kinetics performance.Based on the analysis to XRD, SEM and TEM, it can be seen that there are a lot of LPSO phases in Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys, as shown in Figs.\u00a01-4. The LPSO phase is a long-period stacking ordered phase containing Ni and Y atoms. Moreover, the LPSO structure with the periodic distribution of Y and Ni has an orderly composition and an orderly stacking in the entire LPSO phase. Combined with Fig.\u00a01 and Fig.\u00a09, the alloys after hydrogen absorption consist of hydrides such as MgH2, Mg2NiH4, YH3, YH2 and Mg2NiH0.3. This means that the hydrogen induced decomposition of the LPSO phase occurs upon initial hydrogen absorption, which can be expressed as:\n\n\n\nLPSO\n+\n\nH\n2\n\n\u2192\nMg\n\nH\n2\n\n+\nY\n\nH\n2\n\n+\nM\n\ng\n2\n\nNi\n\nH\n\n0.3\n\n\n\n\n\nwhere YH2 and Mg2NiH0.3 are further hydrogenated into YH3 and Mg2NiH4, respectively:\n\n\n\nY\n\nH\n2\n\n+\n\n\nH\n2\n\n\u2192\nY\n\nH\n3\n\n\n\n\n\n\n\n\n\nM\n\ng\n2\n\nNi\n\nH\n\n\n0.2\n\n\n\n\n+\n\nH\n2\n\n\u2192\nM\n\ng\n2\n\nNi\n\nH\n4\n\n\n\n\nAfter releasing hydrogen, the XRD analysis shows that the dehydrogenated alloys are mainly composed of Mg, Mg2Ni and YH2. Since the decomposition temperature of YH2 is as high as 790 \u2103 [36], YH2 cannot release hydrogen at the experimental temperature. The dehydrogenation reaction can be described as follow:\n\n\n\nM\n\ng\n2\n\nNi\n\nH\n4\n\n\u2192\nM\n\ng\n2\n\nNi\n\nH\n\n\n0.3\n\n\n\n\n+\n\nH\n2\n\n\u2192\nM\n\ng\n2\n\n\nNi\n+\n\n\nH\n2\n\n\n\n\n\n\n\n\n\nMg\n\nH\n2\n\n\u2192\n\nMg\n+\n\n\nH\n2\n\n\n\n\n\n\n\n\n\nY\n\nH\n3\n\n\u2192\nY\n\nH\n2\n\n+\n\nH\n2\n\n\n\n\nThe LPSO phases do not form any more, indicating that the decomposition of LPSO phase is irreversible.A large number of uniformly distributed nano-hydrides formed during hydrogen absorption have a significant catalytic effect on the hydrogen absorption and desorption performance of the alloy [25,48], which is attributed to the nanosizing and in-situ catalyzing effects. In the process of hydrogen absorption, Ni can promote the decomposition of H2 molecules into H atoms, and the ability to form YHx hydrides is better than that of Mg2NiHx and MgH2 [49,50]. The formation of YHx will causes significant lattice distortion and then promote the formation of Mg2NiHx and MgH2. In the hydrogen desorption process, Mg2NiH4 first desorbs hydrogen to Mg2NiH0.3 and undergoes a significant volume contraction, causing a contraction strain on MgH2 around it, facilitating hydrogen desorption of MgH2. In addition, the multi-phase nanostructure can prevent the growth of Mg grains during the hydrogen absorption and desorption cycle, which can shorten the diffusion distance of H atoms.Although plenty of LPSO phases are contained in Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys, the hydrogen storage performance of Mg91.4Ni7Y1.6 is better than that of Mg92.8Ni2.4Y4.8 alloy. The reason is that the fine eutectic structure in Mg91.4Ni7Y1.6 alloy and many fine particles formed after activation are conducive to increasing the specific surface area, providing many reaction nucleation and diffusion interfaces, and enhancing the diffusion ability of H atoms in the alloy. The schematic diagram of the hydrogen absorption process of Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys is shown in Fig.\u00a017. According to the results of SEM and TEM analysis, there are a large number of bulky 18R-LPSO phases and a small amount of phase interfaces in the Mg92.8Ni2.4Y4.8 alloy, which causes most of the hydrides to nucleate only on the surface of the alloy. The growth of hydride requires H atoms to penetrate through the hydride layer and diffuse into the alloy. The growth of the hydride layer stops when it grows to a certain thickness, which results in part of the alloy structure not participating in the hydrogenation reaction, thereby reducing the maximum hydrogen absorption capacity and slowing down the hydrogen absorption rate of the alloy. On the contrary, Mg91.4Ni7Y1.6 alloy contains a large amount of Mg+Mg2Ni+14H-LPSO ternary eutectic, the size of LPSO phase is small, and there are many phase interfaces in the alloy. The hydride can not only nucleate on the surface of alloy particles, but also nucleate at the phase interface. In addition, the diffusion speed of H atoms in the LPSO phase is faster than that in the Mg phase, and the particles size of the Mg91.4Ni7Y1.6 alloy is smaller than that of the Mg92.8Ni2.4Y4.8 alloy, which makes the diffusion distance of the H atoms shorter. Therefore, the hydrogen absorption rate and total hydrogen absorption of Mg91.4Ni7Y1.6 alloy are higher than that of Mg92.8Ni2.4Y4.8 alloy under the same reaction conditions.In this paper, Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys were prepared by smelting and ball grinding. The microstructure and hydrogen storage properties are studied in detail. The catalytic mechanisms of alloying Ni and Y are elaborated. The results are as follows:\n\n(a)\nThere are three phases in Mg91.4Ni7Y1.6 alloys, including Mg, Mg2Ni and abundant lamellar 14H-LPSO, while the massive bulk 18R-LPSO phases appear in the Mg92.8Ni2.4Y4.8 alloy. With the increase of the relative content of Y and the decrease of Ni, Mg24Y5 are formed in the Mg92.8Ni2.4Y4.8 alloy, and Ni element is completely distributed in the LPSO phases, inducing Mg2Ni phases disappears.\n\n\n(b)\nDuring the hydrogen absorption, LPSO phases decompose, causing hydride Mg2NiHx and YH2/YH3 to generate in situ. These phases distributed more uniformly in the alloy can play a better role of in-situ catalysis. In addition to the better dispersion of these catalysts, Mg91.4Ni7Y1.6 alloys also have more phase boundaries, so hydrogen atoms can diffuse better and improve the dynamic properties of the material.\n\n\n(c)\nA large number of fine particles are contained in Mg91.4Ni7Y1.6 alloys, which exposes more second-phase hydrides to the alloy surface and shortens the diffusion distance of H atoms. Not only can the maximum amount of hydrogen absorption of the material be increased, but also the activation energy of hydrogen absorption can be reduced, and the rate of hydrogen absorption of the material can be improved.\n\n\n(d)\nMg91.4Ni7Y1.6 alloys have better kinetic and thermodynamic properties. Under the conditions of 300 \u2103, 350 \u2103 and 3\u00a0MPa, the hydrogen absorption contents of Mg91.4Ni7Y1.6 alloys reach 4.64\u00a0wt.% and 5.78\u00a0wt.% in 5\u00a0min, respectively. The activation energy of hydrogen absorption was 25.4\u00a0kJ/mol H2, and the enthalpy and entropy of hydrogen absorption were \u221260.6\u00a0kJ/mol H2 and 105.5\u00a0J/K/mol H2, separately. The alloy begins to dehydrogenate at 210 \u2103, with the dehydrogenation activation energy of 87.7\u00a0kJ/mol H2, and the \u0394H and \u0394S of dehydrogenation are 56.9\u00a0kJ/mol H2 and 97.9\u00a0J/K/mol H2, respectively.\n\n\nThere are three phases in Mg91.4Ni7Y1.6 alloys, including Mg, Mg2Ni and abundant lamellar 14H-LPSO, while the massive bulk 18R-LPSO phases appear in the Mg92.8Ni2.4Y4.8 alloy. With the increase of the relative content of Y and the decrease of Ni, Mg24Y5 are formed in the Mg92.8Ni2.4Y4.8 alloy, and Ni element is completely distributed in the LPSO phases, inducing Mg2Ni phases disappears.During the hydrogen absorption, LPSO phases decompose, causing hydride Mg2NiHx and YH2/YH3 to generate in situ. These phases distributed more uniformly in the alloy can play a better role of in-situ catalysis. In addition to the better dispersion of these catalysts, Mg91.4Ni7Y1.6 alloys also have more phase boundaries, so hydrogen atoms can diffuse better and improve the dynamic properties of the material.A large number of fine particles are contained in Mg91.4Ni7Y1.6 alloys, which exposes more second-phase hydrides to the alloy surface and shortens the diffusion distance of H atoms. Not only can the maximum amount of hydrogen absorption of the material be increased, but also the activation energy of hydrogen absorption can be reduced, and the rate of hydrogen absorption of the material can be improved.Mg91.4Ni7Y1.6 alloys have better kinetic and thermodynamic properties. Under the conditions of 300 \u2103, 350 \u2103 and 3\u00a0MPa, the hydrogen absorption contents of Mg91.4Ni7Y1.6 alloys reach 4.64\u00a0wt.% and 5.78\u00a0wt.% in 5\u00a0min, respectively. The activation energy of hydrogen absorption was 25.4\u00a0kJ/mol H2, and the enthalpy and entropy of hydrogen absorption were \u221260.6\u00a0kJ/mol H2 and 105.5\u00a0J/K/mol H2, separately. The alloy begins to dehydrogenate at 210 \u2103, with the dehydrogenation activation energy of 87.7\u00a0kJ/mol H2, and the \u0394H and \u0394S of dehydrogenation are 56.9\u00a0kJ/mol H2 and 97.9\u00a0J/K/mol H2, respectively.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was financially supported by Chongqing Special Key Project of Technology Innovation and Application Development, China (Grant No. cstc2019jscx-dxwtB0029)", "descript": "\n                  Magnesium-based hydrogen storage materials are considered as one of the most promising candidates for solid state hydrogen storage due to their advantages of high hydrogen capacity, excellent reversibility and low cost. In this paper, Mg91.4Ni7Y1.6 and Mg92.8Ni2.4Y4.8 alloys were prepared by melting and ball milling. Their microstructures and phases were characterized by X-ray diffraction, scanning electron microscope and transmission electron microscope, and hydrogen absorbing and desorbing properties were tested by the high pressure gas adsorption apparatus and differential scanning calorimetry (DSC). In order to estimate the activation energy and growth mechanism of alloy hydride, the JMAK, Arrhenius and Kissinger methods were applied for calculation. The hydrogen absorption content of Mg92.8Ni2.4Y4.8 alloy reaches 3.84\u00a0wt.% within 5\u00a0min under 350 \u2103, 3\u00a0MPa, and the maximum hydrogen capacity of the alloy is 4.89\u00a0wt.% in same condition. However, the hydrogen absorption of Mg91.4Ni7Y1.6 alloy reaches 5.78\u00a0wt.% within 5\u00a0min, and the maximum hydrogen absorption of the alloy is 6.44\u00a0wt.% at 350 \u2103 and 3\u00a0MPa. The hydrogenation activation energy of Mg91.4Ni7Y1.6 alloy is 25.4\u00a0kJ/mol H2, and the enthalpy and entropy of hydrogen absorption are -60.6\u00a0kJ/mol H2 and 105.5\u00a0J/K/mol H2, separately. The alloy begins to dehydrogenate at 210 \u2103, with the dehydrogenation activation energy of 87.7\u00a0kJ/mol H2. By altering the addition amount of Ni and Y elements, the 14H-LPSO phase with smaller size and ternary eutectic areas with high volume fraction are obtained, which provides more phase boundaries and catalysts with better dispersion, and there are a lot of fine particles in the alloy, these structures are beneficial to enhance the hydrogen storage performance of the alloys.\n               "}
{"full_text": "Data available on request from the interested.In this era, it is becoming imperative to make a transition towards renewable energies. However, most renewable energy sources have an intermittent nature, precluding their adoption and utilization at their full potential [1]. One energy vector that has the potential to overcome this issue is hydrogen. This carrier can store energy for longer periods of time compared to alternative electrochemical energy storage, e.g. batteries, and can alleviate intermittency issues for daily or even seasonal variations [2,3]. In order to accomplish this, water is electrochemically split into its components, oxygen and hydrogen, both products that can be used in a wide range of processes [4].The study of catalytic surfaces has concentrated mainly on the development of materials that are very catalytically active, but are not very dependent on stability, the majority being analyzed in very short times of less than 20\u00a0h, and in low currents of 10\u00a0mA\u00a0cm\u22122, [5\u20137]. Some methods use cyclic voltammetry of the materials in a very narrow potential window [8,9], preventing study of the corrosion of the materials in process. There are several works in which the cathodic corrosion process is analyzed during the off periods for the formation of reverse current in the cathode and the electrode is subject to degradation. These findings were obtained using electrochemical impedance spectroscopy (EIS) in the open circuit potential (OCP) and during the off processes [10\u201313]. Hence, there is a lack of studies on the stability of electrodes during the oxygen evolution reaction (OER) in more realistic operating conditions, i.e. high density currents and long operation times. Some research works use techniques such as inductively coupled plasma mass spectrometry (ICP-MS) and electrochemical quartz crystal microbalance (EQCM) to measure the loss of material during the electrolysis operation [14]. Acceleration tests have been proposed to study the degradation of electrodes during water splitting. These tests use a series of techniques applied consecutively, such as constant current density, then a constant potential and finally cyclic voltammetry to simulate the inverse current in the off process [15]. Nevertheless, these studies are few in number and, to the best of our knowledge, do not use additional transient techniques to measure or detect the changes on the surface of the electrodes due to the water splitting reactions. Transient and in-situ techniques are powerful tools capable of discerning small changes in the surface of the electrode that allow observation of the degradation of the electrode and the formation of new compounds on the electrode surface.Traditional corrosion studies performed on different stainless steels and nickel alloys immersed in aerated and hot concentrate alkaline media have demonstrated that metals such as Cr, Mo, Fe and Ni dissolve and leave a porous layer of nickel oxides and nickel hydroxides [16,17]. Formation of these nickel oxide compounds depend on the temperature, and it has was found that at 120\u00a0\u00b0C the main compound formed was Ni(OH)2, while at 180\u00a0\u00b0C this was NiO [18]. When nickel was used as catalytic active substrate, a small amount of iron was incorporated, finding that the iron is beneficial because it forms a layer of NiFe2O3 that protects the Ni against corrosion [18]. Studies performed on an AISI-SAE 316 stainless steel (SS) in NaOH at 90\u00a0\u00b0C over 4 months found the formation of a black layer on top of the electrode. The layer was constituted by iron oxyhydroxide, nickel oxide and nickel hydroxide [19]. In the ASI-SAE 304 SS at different concentrations of NaOH and 150\u00a0\u00b0C, stress corrosion cracking in the samples and a depletion of Fe and Cr were found, leaving the Ni on the surface and a under layer of oxides such as NaMO2, where M represents the different transition metals [20].Recently, AISI-SAE 304 SS plates were used in the structure of an electrolyzer, operating with NaOH electrolyte at a pH of 13 and 60\u00a0\u00b0C [21]. The SS plates were subject to electrolysis at low potential for several hours, and subsequently analyzed by XPS and Raman spectroscopy. The electrolyte was analyzed after electrolysis by ICP. Iron dissolution was found during the first hours of electrolysis, as well as the formation of different hydroxides and oxyhydroxides on the surface of electrodes, such as FeOOH, CrOOH, and Ni(OH)2. Depth profiles of composition performed by XPS analysis showed an increase in oxygen and depletion of the different metals, which suggests the formation of oxide compounds on the electrode surface.There are some contradictory studies about the role of phosphorous in the catalytic layer of the electrode for water splitting. Some of these have found that the coating of NiFeP in alkaline media acts as a passive layer that protect the electrode against corrosion [22]. In agreement with that work, Huang et al. found that an increase in phosphorus in the catalytic layer changes the properties of the coating, which become more amorphous, and facilitates the formation of a passive layer [23], creating a coating with fewer phase boundaries and crystalline defects. However, Safizadeh et al., evaluated the behavior of FeMoP catalytic layer in 1\u00a0M KOH and found that the phosphorous in the layer makes it more susceptible to corrosion [24]. This was demonstrated considering the open circuit potential (OCP) measurements after cathodic polarization. The OCP values of the FeMo layer changed to a more noble value with P incorporation; but the FeMoP layer stayed at a corrosion active state during the experiment, while, similarly, the corrosion current was higher with the addition of phosphorus.The current work aims to fill some gaps in the knowledge of the behavior of the NiFeP catalytic layer deposited on stainless-steel electrodes used as anode in the water splitting process, i.e., during the oxygen evolution reaction (OER). Similar conditions to the real operation of an electrolyzer were proposed in the study, considering the application of a constant anodic current density of 400\u00a0mA\u00a0cm\u22122 over2h and the open circuit potential (OCP) monitoring for 2.5\u00a0h, in which time the system stabilizes. At the same time, the condition of intermittence of the electrolysis process that currently occurs with the use of renewable sources is evaluated. Electrochemical transient technique, such as EIS measurements, and in-situ Raman spectroscopy, were also performed in order to evaluate the changes in the electrode surface during the electrolysis. All of these were complemented with other characterization techniques to get a clearer understanding of what happens on the electrode during the OER.The working electrodes were AISI 304 stainless-steel square plates of 0.7\u00a0\u00d7\u00a00.7\u00a0cm2 (approx. exposed area of 0.5\u00a0cm\u22122). The electrodes were polished with sandpaper up to P1500 grit, followed by a final polish with cloth covered with alumina (particle size of 1\u00a0\u00b5m). Non-exposed areas of the electrodes, laterals and back (where the electrodes were connected) sides were isolated, covering with epoxy resin and leaving 0.5\u00a0cm\u22122 of area expose to the electrolyte.In the current work we look for deep information about of the behavior against corrosion of the NiFeP catalytic layer deposited on stainless-steel electrodes used as anode in the water splitting process, i.e., during the oxygen evolution reaction (OER). Accordingly, stainless steel samples with and without anodic treatment and deposited catalytic layer were tested in anodic exigent conditions. The anodic treatment and the nature of catalytic layer and the electrodeposition conditions were selected from the best results of water splitting previously obtained [25] Polished stainless steel substrates were anodized following the procedure reported by Bervian [26]. Briefly, an aqueous solution with 10% v/v sulfuric acid (J.T. Baker\u00ae 95.9%) and 10% v/v glycerol (Pharmaceutical Grade) was prepared as electrolyte using type III water (conductivity \n\u2264\n 2.5 \u00b5S cm\u22121). A current density of 1\u2009A\u2009cm\u22122 was applied for 600\u2009s between the contra (platinum plate) and working (stainless steel) electrodes. The electrodes were separate 2\u2009cm, and a constant potential around 3\u2009V was observed. Finally, the working electrode was washed with abundant type III water.The catalytic coatings of NiFeP were electrodeposited either on polished or on anodically treated stainless steel, following the procedure described by Sridharan et al. [27]. Before coating application, the electrodes were thoroughly cleaned in an ultrasonic bath by immersion in ethanol followed by distillate water. The electrodeposition of NiFeP coatings was performed using a Pt foil as a counter electrode and applying a current density of \u2212\u200930\u2009mA\u2009cm\u22122 for 2400\u2009s. The electrolytic bath composition was: 0.114\u2009M \n\nNiS\n\n\nO\n\n\n4\n\n\n.\n6\n\n\nH\n\n\n2\n\n\nO\n\n (EMSURE\u00ae MERK, \n\u2265\n 99.0%), 0.108\u2009M \n\nFeS\n\n\nO\n\n\n4\n\n\n.\n7\n\n\nH\n\n\n2\n\n\nO\n\n (CARLO ERBA, \n\u2265\n 99.5%), 0.129\u2009M CH3COOK (MERK) and 0.094\u2009M \n\nNa\n\n\nH\n\n\n2\n\n\nP\n\n\nO\n\n\n2\n\n\n.\n\n\nH\n\n\n2\n\n\nO\n\n(ALDRICH, \n\n\u2265\n99\n%\n\n) using type III water. The pH of the bath was adjusted to 2 with sulfuric acid addition. The bath was kept at 40\u2009\u00b0C under vigorous agitation in order to promote good adhesion of the coatings. Finally, the electrodes were cleaned with abundant type III water.Three kinds of samples were studied, namely: stainless-steel with anodic treatment (SS-AT); stainless-steel with anodic treatment and coated with NiFeP (NiFeP/SS-AT); and stainless-steel without anodic treatment and coated with NiFeP (NiFeP/SS).Surface morphology of NiFeP/SS, NiFeP/SS-AT and SS-AT samples were characterized before and after performing the oxygen evolution reaction (OER) process in 1\u2009M NaOH electrolyte. Scanning electron microscopy (SEM) and electron dispersive X-ray spectroscopy (EDS) were performed using a JEOL-JSM 6490LV equipment, with acquisition parameter following the ASTM E1508\u201312 standard, as follow: acceleration voltage 20\u2009kV, spot size 50, working distance 10\u2009mm, dead time 11%, number of counts that arrive to the detector of 1.2kcps. SEM images were obtained using a secondary electron\u2019s detector, after capture the imagen was not modified. EDS elemental analysis was semiquantitative because sense the photons, also the microprobe only carry information of 10\u2009\u00b5m of the sample and the identification percent is 0.1% in weight. The algorithm use was INCAEnergy. Crystal structures were characterized by X-ray diffraction (XRD) Miniflex600 Rigaku with \n\nCu\n\n\nK\n\n\n\u03b1\n\n\n\n (\n\n\u03bb\n=\n1.54059\n\n\u00c5\n\n) as source and operated with source powers of 40\u2009kV and 15\u2009mA, with a step size (\u00b02\u03b8) of 0.02 and dwell time (s) of 1. The scans were made at angles \u00b02\u03b8 of 10\u201380. The data were analyzed using the high score plus program, with an automatic background determination and subtraction, and a smooth using quintic the program and a convolution range of 7. The peaks were indexed using several data bases as COD, PDF2, ICSD and information taken from different papers which concern similar materials. Raman spectra of the surfaces were taken in a Horiba Jobin Yvon (Labram HR) Nikon (BX41) microscope, with a CCD detector (Wright 1024\u2009\u00d7 256 pixels), with a laser of 625\u2009nm. In situ Raman measurements were performed using a homemade cell and the same Raman equipment already mentioned, using 6 repetitions and 50\u2009s of capturing without filter, the range of the capture was from 100 to 3000\u2009cm\u22121 with an objective of x50WD, a slit of 600\u2009\u00b5m, a hole of 800\u2009\u00b5m.Electrochemical characterization for OER was performed using a Zhaner IM6e potentiostat. A typical three-electrode cell was used for all measurements, a Hg=HgO electrode made in-house was used as reference electrode and a Pt mesh was used as counter electrode. The potentials were converted to reversible hydrogen electrode (RHE) according to literature reports [4,28\u201330], as follows:\n\n(1)\n\n\n\n\nE\n\n\nRHE\n\n\n\n\n\nV\n\n\n\n=\n\n\nE\n\n\nHg\n/\nHgO\n\n\n+\n0.098\nV\n+\n0.059\nV\n\u00d7\npH\n\n\n\nWhere, \n\n\nE\n\n\nHg\n/\nHgO\n\n\nis the experimental measured potential, 0.098\u2009V factor is the relation between \n\n\n\nE\n\n\nHg\n/\nHgO\n\n\n\n\nand the \n\n\n\nE\n\n\nNHE\n\n\n\n\n\nV\n\n\n\n\n, the factor of \n\n0.059\nV\n\u00d7\npH\n\n is the correction given for the pH of the solution. All the potentials were corrected by the ohmic drop using \n\n\n\nR\n\n\nu\n\n\n\u00d7\nI\n\n, where \n\n\nR\n\n\nu\n\n\n is the resistance of the electrolyte obtained by electrochemical impedance spectroscopy (EIS), and I is the current. The overpotentials of the oxygen evolution reaction (OER) were calculated using the RHE potentials early calculated and an equilibrium potential of 1.23\u2009V\n\n(2)\n\n\n\u03b7\n=\n\n\nE\n\n\nRHE\n\n\n\n\n\nV\n\n\n\n\u2212\n1.23\n\nV\n\n\n\n\nThe electrochemical performance of the working electrode and its corrosion resistance were evaluated by the following procedure: (i) a period of stabilization in the solution of 1\u2009M NaOH for 2\u2009h, (ii) electrochemical impedance spectroscopy (EIS) performed at the open circuit potential (OCP); next (iii) a linear sweep voltammetry (LSV) at a scan rate of 5\u2009mV\u2009s\u22121 caried out to measure the initial activity of the material to oxygen evolution; then, (iv) cycles of 2\u2009h of electrolysis (OER) at 400\u2009mA\u2009cm\u22122 followed by two and a half hours without pass of current and a final EIS measurement at the OCP. This procedure was repeated 12 times. Finally, a LSV measurement performed at 5\u2009mV\u2009s\u22121 was conducted to evaluate possible changes in the catalytic activity of the samples after electrolysis test. All experiments were conducted at room temperature and using 1\u2009M NaOH as electrolyte, prepare with type III water.The EIS spectra were analyzed using both distribution of relaxation times (DRT) calculations and fit with electrical equivalent circuit, using \n\n\n\nR\n\n\ns\n\n\n(\n\n\nR\n\n\n1\n\n\n\n\nCPE\n\n\n1\n\n\n)\n(\n\n\nR\n\n\n2\n\n\n\n\nCPE\n\n\n2\n\n\n)\n(\n\n\nR\n\n\n3\n\n\n\n\nCPE\n\n\n3\n\n\n)\n\n and \n\n\n\nR\n\n\ns\n\n\n(\n\n\nR\n\n\n1\n\n\n\n\nCPE\n\n\n1\n\n\n)\n(\n\n\nR\n\n\n2\n\n\n\n\nCPE\n\n\n2\n\n\n)\n(\n\n\nR\n\n\n3\n\n\n\u2212\nW\n\n\nCPE\n\n\n3\n\n\n)\n\n circuits for experimental impedance without and with diffusion, respectively (see details of electrical equivalent circuits in Fig. S-7 in the supporting information). In the electrical equivalent circuits Rs is the electrolyte resistance, R\n\n1\n and CPE\n\n1\n are, respectively, the resistance and the constant-phase element associated with the passive layer formed on the surface of the electrode. Similarly, R\n\n2\n and CPE\n\n2\n are associated with a second passive layer formed on the catalytic coating. R\n\n3\n is the charge transfer resistance and CPE\n\n3\n is the constant-phase element associated with the double layer capacitance, and W is the Warburg element of the oxygen diffusion. The electrical equivalent circuits were analyzed with Gamry Chem Analyst and the DRT calculations were performed using the algorithm implemented by Wan et al. [31], using a regularization parameter of 0.005 and a gaussian discretization. Capacitances and resistances were comparable for both methods. The capacitance (C\n\neff\n) values were calculated from Brug\u2019s relationship [32], using the values of CPE and considering the approximation performed by Hirschorn et al. [33], where the Ohmic resistance is included in CPE behavior associated with surface distributions, following the Eq. (3). This approximation has been used in similar conditions [34,35].\n\n(3)\n\n\n\n\nC\n\n\neff\n\n\n=\n\n\nQ\n\n\n\n\n1\n\n\n\u03b1\n\n\n\n\n\n\n\n\n\n\nR\n\n\ne\n\n\n\u2212\n1\n\n\n+\n\n\nR\n\n\nt\n\n\n\u2212\n1\n\n\n\n\n\n\n(\n\u03b1\n\u2212\n1\n)\n/\n\u03b1\n\n\n\n\n\nWhere C\n\neff\n is the effective capacitance, R\n\ne\n is the Ohmic resistance and R\n\nt\n, Q, and \u03b1 are the current properties of the CPE. For the DRT spectra, the peak position was used directly as the effective time constants \u03c4.\n\nFig. 1 shows the scanning electron microscopy (SEM) of the samples before and after different cycles of electrolysis (OER) performed in 1\u2009M NaOH. There is a kind of micro-texture formation on the surface of NiFeP/SS sample after OER (Fig. 1(a) and 1(b)), meaning that a generalized attack occurs on the catalytic layer of NiFeP after OER. The products of the corrosive attack are well adhered to the surface. On the other hand, at the surfaces of anodically treated samples, NiFeP/SS-AT and SS-AT, Fig. 1(c) and (e), some material dissolution was observed during the anodic treatment. The dissolution process occurs mainly around the grain boundaries [36], causing the formation of micro-structured features, like faceted/granular appearance, of varied sizes between 10 and 50\u2009\u00b5m. The dissolution process was controlled in part with the addition of glycerin that diminishes the conductivity of the media, enabling better process control, in which the porous structure is formed. This is desirable, because this causes an increase of the area of the electrodes making it more suitable for water electrolysis. These coatings did not experience appreciable changes after electrolysis, as can be seen in Fig. 1(c) to (f). The surfaces of those samples exhibited similar texture, indicating that there was not severe corrosive attack after anodic polarization during OER. The roughness of the surface achieved by the initial anodizing process performed on the stainless-steel samples was preserved. It was also observed that the catalyst NiFeP coating (NiFeP/SS-AT sample) perfectly copied the texture of the surface initially generated by the anodic pretreatment process (SS-AT sample). Due to the anodic treatment, the NiFeP/SS-AT sample is expected to have a larger surface area than the NiFeP/SS sample. Consequently, lower current density during the NiFeP-catalyst electrodeposition and lower thickness of the layer for the NiFeP/SS-AT sample could be expected, because the catalyst electrodeposition was performed a constant current.The EDS mapping analysis shows surface homogeneity of the elemental composition of the NiFeP coating before and after the OER process (see Figure S-1 and Table S-1 in the supporting information). There was only an appreciable change in the iron composition after electrolysis. The amount of Fe in the catalyst coating diminished during the OER process by around 50%, resulting in a relative increase in Ni content in the surface of the coated samples. In the case of the sample SS-AT, there was only a slight reduction in the iron content. A dissolution process after the electrolysis that only involved the iron in the samples was corroborated, this result is in agreement with that reported by Santarini [20]. In addition, an increasing in the amount of oxygen in the surface of the samples could be expected due to the formation of different oxides [21,37,38], as it will be shown in the next section during the Raman results analysis.\n\nFig. 2 shows the XRD patterns of different samples before and after the electrolysis process. SS-AT and NiFeP/SS-AT samples exhibited peaks located at 2\u03b8 =\u200943.79\u00b0, 50.91\u00b0, and 74.83\u00b0, corresponding to crystalline planes {111}, {200}, and {220} respectively, of the austenite phase of the stainless-steel substrate [39]. This because the thickness of the coatings is very thin, as was suggested by SEM analysis, making evident the diffraction peaks of the substrate. No diffraction peaks associated to the NiFeP catalytic coating were observed, which suggests that the coatings are amorphous in nature. The amorphization of the coating is due to the presence of phosphorus, which causes a distortion in the base structure of the coating, which is also dependent on the amount of phosphorus present in the deposit [40,41]. The XRD patterns obtained after the electrolysis process (\nFig. 3(b)) did not show changes in the diffraction peaks of the samples and the peaks related to stainless steel substrate remain in the patterns, meaning that the amorphous nature of the coating was preserved. Even though there was a dissolution of Fe and a probable enrichment of nickel on the surfaces (EDS analysis) the amorphous nature of the coating and the formed nickel compounds (Raman analysis will confirm further the formation of nickel oxides) is preserved after several cycles of electrolysis. The formation of complete amorphous structures after the electrochemical activation treatment in a basic electrolyte has been confirmed by other researchers [42]. This is convenient in terms of catalytic activity to oxygen evolution, that because electrocatalysts with an amorphous structure exhibit better OER activity than electrocatalysts with a crystalline structure [42,43].Raman spectroscopy is a technique commonly used for surface characterization, mainly for oxides and salt compounds materials. The main ventage of this techniques respect to others are that it allows a rapid characterization of the problem sample material without further preparation, enabling to characterize the actual surface without the interferences or surface changes that can be introduced by sample preparation. In addition, Raman spectroscopy allows to perform in-situ experiments during the electrochemical measurements in aqueous media, allowing to detect changes of the surface sample during the anodic or cathodic processes without water interferences.The ex-situ Raman spectra of the samples before the electrolysis process is shown in Fig. S-2 in the supporting information. There are no appreciable bands in the Raman spectra related with metal-oxide compounds formed on the surface of the samples before the electrolysis process. These results indicate either that the surfaces of the anodically treated samples and the NiFeP catalytic layer preserve their metal character, or that there were not enough oxides to be detectable with this technique. Raman results indicate no metal oxide formation in the samples in the condition \u201cbefore electrolysis\u201d and, as expected, not oxygen was detected in the EDX analysis, see Table S-1 in the supporting information. Conversely, in the condition \u201cafter electrolysis\u201d oxygen content was detected in the samples, which can be related to the formation of NiOOH compound, as will be seen further by in situ Raman spectroscopy measurements.Raman in-situ spectra of the samples were recorded during the anodic polarization at different potential for OER. The Raman in-situ measurements were done in a home-made cell in which the surface of the samples was covered by 1\u2009M NaOH electrolyte. The objective of the microscope was located very close to the surface of the electrolyte to prevent losses in the signal. Raman spectra were measured at different anodic potentials. The polarization potential values were chosen considering the potential reached by the OCP after the on-period of electrolysis at 400\u2009mA\u2009cm\u22122, to identify possible species formed. The polarization potentials for in-situ Raman measurements were 0.81\u2009V, 1.05\u2009V and 1.39\u2009V vs RHE, and finally a current density of 10\u2009mA\u2009cm\u22122 was applied to complete the polarization of the surface at the OER. At this last stage, the number of oxygen bubbles was very large. Following this, a return to the polarization potential was applied, i.e., 1.39\u2009V, 1.05\u2009V, 0.81\u2009V vs RHE. Each polarization potential was applied for 30\u2009min in order to obtain a stationary state and Raman spectrum was then recorded. At least three measurements were performed for each sample, in order to verify reproducibility of the results. In-situ Raman spectra of the samples recorded before and after all the cycles of electrolysis at 400\u2009mA\u2009cm\u22122 are shown in Fig. 3. In the case of the NiFeP/SS sample, Fig. 3(a) and (b), two bands at 475\u2009cm\u22121 and 555\u2009cm\u22121, associated with Ni\u2212O vibration of the NiOOH compound [44\u201346], begin to emerge with the anodic polarization. The two allotropic compounds \u03b3-NiOOH and \u03b2-NiOOH exhibit a pair of bands at these wavenumbers, but with different relative intensities of the bands at 475 and 555\u2009cm \u22121. The ratio of the intensity of the bands I475\u2009cm\u22121 /I555\u2009cm\u22121 of the \u03b2-NiOOH compound is lower than for \u03b3-NiOOH [47,48]. However, it was demonstrated that the presence of Fe in the deposition leads to a decrease in the number of redox electrons, so the formation of \u03b3-NiOOH could be inhibited [49]. Table S-2 shows the ratio of the intensities of the bands at 475\u2009cm\u22121 and 555\u2009cm\u22121 (I475\u2009cm\u22121 /I555\u2009cm\u22121) observed in the in-situ Raman spectra of the samples performed at different polarization potentials. In the NiFeP/SS sample the bands of NiOOH are clearly visible at the anodic potential where a high rate of oxygen bubbles occurs, i.e., 1.5\u2009V RHE (10\u2009mA\u2009cm\u22121), and the presence of the bands remains even after the return of the anodic polarization to low values. With the return of anodic polarization to lower values, the intensities of the bands at 475 and 555\u2009cm \u22121 tend to decrease, indicating a probable reduction process of the compounds. These features occur for both situations, before and after electrolysis at 400\u2009mA\u2009cm\u22122. However, after all cycles of electrolysis it was observed that the ratio of the intensities of the bands at 475 and 555\u2009cm \u22121 is different in some respects to that observed before electrolysis, being larger after electrolysis cycles. This is probably due to the previously detected iron dissolution that occurs during electrolysis. According to the study by Bell [44] the intensity of the 555\u2009cm\u22121 band exhibited by the NiOOH layer increases with Fe content in the Ni-Fe alloys. Conversely, if iron content in the Ni-Fe coating diminishes, the intensity of the 555\u2009cm\u22121 band also decreases and the I475\u2009cm\u22121 /I555\u2009cm\u22121 ratio increases after electrolysis, which is consistent with what occurs in the current study. Due to iron dissolution from the coating, a stable and more catalytic layer of \u03b3-NiOOH can be formed. No more bands were detected in the in-situ Raman spectra, which indicates that no additional Ni or Fe species were formed on the surface of the sample.In the case of the anodic-treated (SS-AT) sample, \nFig. 4(c) and (d), no Raman signals associated with Ni or Fe compounds were observed on the sample before electrolysis, with only a broad band at 1000\u2009cm\u22121 associated with OH deformation being observed before the electrolysis process. However, after electrolysis the characteristic bands at 475 and 555\u2009cm\u22121 of the NiOOH layer appeared at high anodic polarization where a high rate of oxygen bubbles occurs, i.e., 1.5\u2009V (10\u2009mA\u2009cm\u22121). This result indicates that there are changes in the elemental composition of the surface of SS-AT sample, i.e. local dissolution of Fe and increase in Ni due to the corrosion process. This situation induces the formation of the NiOOH layer on anodically treated stainless steel after electrolysis. However, in this sample, the Raman bands of the NiOOH (475 and 555\u2009cm\u22121) are less intense than those of the samples with NiFeP catalytic layer and they disappear with the return of anodic polarization, see Fig. 3(d). This is probably due to the lower thickness of the NiOOH layer formed in the SS-AT sample after electrolysis, which experiences a complete reduction process with the return of anodic polarization to lower values.In-situ Raman spectra of sample NiFeP/SS-AT, Fig. 4(e) and (f), exhibited a similar behavior to the NiFeP/SS sample. As a result of the presence of the catalytic layer of NiFeP, the two bands at 475 and 555\u2009cm \u22121, associated with Ni\u2212O vibration in NiOOH compound, are present both before and after the electrolysis process. Before the electrolysis the bands associated with NiOOH are clearly visible at the anodic potential where high rate of oxygen bubbles occurs, i.e., 1.5\u2009V RHE (10\u2009mA\u2009cm\u22121), and the presence of the bands remains even when the anodic polarization was returned to low values. In addition, after the electrolysis process the NiOOH bands are detected even at potentials lower than that where a high rate of oxygen bubbles occurs (1.5\u2009V vs. RHE).According to the surface characterization and in-situ Raman results shown above, the mechanism that best explains the corrosion and transformation of the surfaces of the samples during OER is that proposed by Jerkiewicz [50], which involves nickel dissolution followed by formation of several nickel species in an alkaline electrolyte, reactions (1) to (3). According to this reaction model, the OH- of the media is adsorbed in the surface, reaction (1), followed by a rearrangement of Ni, in turn creating a 3D lattice, reaction (2). Subsequently, nickel dissolution occurs upon the formation of adsorbed \u03b1-Ni(OH)2, reaction (3).\n\n(1)\n\n\nNi\n+\nO\n\n\nH\n\n\n\u2212\n\n\n\u2192\nNi\n\u2212\nO\n\n\nH\n\n\nad\n\n\n+\n\n\ne\n\n\n\u2212\n\n\n\n\n\nfast\n\n\n\n\n\n\n\n\n\n(2)\n\n\nNi\n\u2212\nO\n\n\nH\n\n\nad\n\n\n\u2192\n\n\n\n\nOH\n\u2212\nNi\n\n\n\n\nquiasi\n\u2212\n3\nD\n\nlattice\n\n\n\n\n\n\nrate\n\ndetermining\n\nstep\n\n\n\n\n\n\n\n\n\n(3)\n\n\n\n\n\n\nOH\n\u2212\nNi\n\n\n\n\nquiasi\n\u2212\n3\nD\n\nlattice\n\n\n+\nO\n\n\nH\n\n\n\u2212\n\n\n\u2192\nNi\n\n\n\n\nOH\n\n\n\n\n2\n,\nad\n\n\n+\n\n\ne\n\n\n\u2212\n\n\n(\nfast\n)\n\n\n\n\nThe \u03b1-Ni(OH)2 can be then converted into a more stable form with the application of an anodic polarization potential, in which it loses water, passing to \u03b2-Ni(OH)2\n[51]. The new compound is better arranged compared with the previous \u03b1-Ni(OH)2, and the reaction is irreversible when a high enough voltage is applied. Upon further anodic polarization the \u03b2-Ni(OH)2 passes to \u03b2-NiOOH, involving the nickel oxidation from 2+ to 3+. If the anodic polarization increases, allotropic transformation of the nickel oxyhydroxide occurs, forming \u03b3-NiOOH, in which nickel has an oxidation state of 3.2.\nFig. 4 shows the results of the electrochemical evaluation of the NiFeP/SS sample after electrolysis process (OER). The procedure was described in the experimental part and consisted of a period of stabilization in the electrolyte, EIS measurement, LSV, on/off cycles, electrolysis process (2\u2009h of electrolysis (OER) at 400\u2009mA\u2009cm\u22122), OCP measurement and EIS, and finally LSV. Fig. 4a) shows the Nyquist diagrams of impedance measurements performed at the different times of electrolysis. EIS measurements were performed during the off period, after the OCP stabilization, of each electrolysis cycle. After the first electrolysis processes (0\u201312\u2009h) two coupled capacitive loops were observed in the Nyquist diagram (for more detail see the Bode diagram of EIS, Fig. 4(c) and Figure S-3 in supporting information). DRT analysis (Figure S-3c in supporting information) clearly indicates that three time-constants should be considered in the impedance response after the first electrolysis cycles. The flattened capacitive loop at high frequency (HF) is related to the previously formed Ni(OH)2 layer, which gradually transforms to NiO oxide [37] and catalytic NiOOH layer; while the low frequency (LF) loop could be related to the parallel combination between double-layer capacitance and the charge-transfer resistance of the coating/electrolyte interface. The formation of hydroxide layer has been pointed out in theoretical studies, which demonstrate that it is not thermodynamically stable materials during the oxygen evolution reaction and propose the dissolution and the formation of a hydrous amorphous layer over the bulk metal, through which the oxygen anion diffuses [52]. Also, the formation of Ni oxide \u2013 hydroxide layer in alkaline media has been evidenced by TEM images, where dense layer of NiO and NiOOH was found on the surface [37,38].After successive electrolysis cycles, both HF and LF capacitive loops decouple and a straight line at 45\u00b0 at the limit of the low frequency region begins to emerge, see Fig. 4(a)-(c) and Figure S-3 in supporting information. Similar feature has been observed in passive layer formed on steel immerse in alkaline media [53]. More detail of the features of EIS results can be observed in Figure S-3, where bode plots and DRT analysis of the impedance measurements performed during the on/off cycles of electrolysis process are presented. The parameter values calculated from DRT analysis are shown in Table S-3, while Figure S-7 shows those values as a function of time, for comparison purpose. From the DRT plots of impedance, it was possible to determined that three processes took place. Some of these were coupled at initial times of electrolysis; however, they decoupled at larger cycle numbers and a diffusion process also begins at the limit of low frequency, this probably due to the formation of the dense layer of Ni oxides and the oxygen diffusion [37]. From DRT analysis an increase of the peak located at the HF region can be seen, indicating the increase of the coverage of the surface by the NiOOH layer during the cycles of electrolysis. On the other hand, the peak at middle frequency region (MF) diminishes, indicating the reduction of the early Ni(OH)2 layer [41]. Likewise, the peak at LF diminishes, indicating that the charge-transfer resistance is reduced over time, due to the formation of NiOOH catalytic layer. The straight line at 45\u00b0, observed at the zero-limit of the LF region, could be related to the diffusion process of oxygen away from the electrode surface. The formation of the NiOOH layer consolidates after successive cycles of electrolysis. Consequently, the impedance diagrams are very similar. The layer formed are expected to become more stable, as has been shown in other studies in which the surface composition stop changing after the initials time of polarization [21].As can be seen in the Tables S-2 and S3, in the supporting information, the values of capacitance, in some cases, are in the order of the mF, which are larger and no typical of films and double layer. The reason for this is that due to the difficulty of accurately calculating the actual area of the electrodes, which is larger than a planar electrode, because the high roughness of the surface, we use the geometric area instead of the real area in order to calculate the capacitances. In a recent published paper [25] we obtained that the geometric area is about two orders of magnitude inferior to the real area of this kind electrodes. Because of that the capacitance values appears one or two orders of magnitude larger than typical values of capacitance for films and double layer. The same situation has been evidenced by other researcher works concerning the capacitance of electrodes for oxygen evolution reaction [34,54], and in tests in alkaline media of steel and iron [55,56]. Furthermore, if we consider that deviation of capacitance values, we estimate that real capacitances of the samples are in the range of 10\u2013100 \u03bcF cm\u22122, and the thickness the oxide films will be in the range of 100 \u2013 1000\u2009nm. Which is typical of this kind of films [37,38].EIS fitting using electrical equivalent circuits was also performed. Figure S-6, in the supporting information, shows the electrical equivalent circuit configurations used to fit EIS experimental results, Figure S-6a for the first electrolysis cycles where no diffusion process was observed and Figure S-6b for the last electrolysis cycles, also considering Warburg impedance to model the diffusion process. The parameter values obtained after EIS fitting with electrical equivalent circuits are presented in Table S-4. The results of fitting using electrical equivalent circuits are in accordance with those obtained by DRT analysis, see also Figure S-7. The reduction in the resistance of catalytic layers during on/off electrolysis process is evident and the straight line of 45\u00b0 observed at the limit of the low frequency of the EIS diagrams performed at the end of electrolysis cycles can be associated with the diffusion of oxygen away from the electrode surface. It is supposed that a large amount of O2 remains on the catalytic layer of the electrode after OER during each cycle of electrolysis, which is then released from the electrode during the off-electrolysis period. As a result, overpotential jumps of the electrode are observed during the on/off cycles of electrolysis, as can be seen in Fig. 4(d).The red vertical lines drawn in Fig. 4(d) indicate the time when the electrolysis process is stopped and the OCP is recorded. A gradual increase in the overpotential during the on period of electrolysis is attributed to accumulation of oxygen bubbles on the catalytic layers of the electrode [57]. Then, during the off period of electrolysis, a gradual decrease in the overpotential is observed due to the release of the oxygen away from the electrode surface. The observed noisy signal was attributed to the excessive production of bubbles that accumulate in the surface, partially blocking the active area of the electrode [14]. These experiments were made at quiescent conditions, so the adsorption of oxygen bubbles on the electrode surface takes on a large importance. The evolution of the overpotential appears to reach a maximum after the fourth on/off cycle of electrolysis, after which a decrease in the overpotential signal is observed, probably due to changes in the nature of the catalytic layer of the electrode or changes on the morphology of the compound formed on the electrode surface, see Fig. 1(b). According to the in-situ Raman and EIS analyses, the catalytic layer NiFeP transforms to a more active NiOOH [37] compound on the electrode surface, and partial iron dissolution occurs during the electrolysis process. This change in the catalytic layer increases the activity of Ni cations for OER [44]. The OCP of the electrode shifts to more positive values after each on/off period of electrolysis (see Fig. 4(e)) as a consequence of the formation of NiOOH layer on the electrode surface. In addition, the formation of NiOOH layer causes a gradual reduction in the size of the OCP jump after each on/off cycle. The rest time in each potentials is longer in each cycle indicating an increase in the thickness of the oxide layer in agreement with the impedance results and other related investigations [37,58]. In addition, the steady state of OCP values are reached in more positive potentials at the end of the cycling process, and an easier release of oxygen bubbles away from the electrode is achieved. Fig. 4(f) shows the results of the LSV performed at 5\u2009mV\u2009s\u22121 before the first and after the last electrolysis cycles. After the electrolysis cycles, reductions of the overpotential at 10\u2009mA\u2009cm\u22122 and 100\u2009mA\u2009cm\u22122 by 13 and 18\u2009mV respectively are observed, indicating improvement of the catalytic activity of the surface after electrolysis.\n\nFig. 5 shows the electrochemical evaluation of the naked anodic-treated stainless-steel (SS-AT) sample during on/off electrolysis (OER) cycling process. It is interesting to highlight that, despite the fact that the SS-AT sample lacks a NiFeP catalytic layer, the main features of its electrochemical response are almost identical to those observed for the NiFeP/SS sample. The in-situ Raman spectra of the SS-AT sample recorded after water electrolysis showed that the OER also induces the formation of the NiOOH layer in this kind of sample. Accordingly, the naked anodic-treated stainless-steel sample can develop a catalytic layer through OER, which could improve its catalytic properties during the on/off electrolysis cycling. This agrees with the activation procedures, which has been used to develop a layer of NiFe over the surface of an austenitic stainless steel. During the activation procedure it was observe the formation of a layer of oxide that began to grow and then pass to a process of densification [37]. The thickness of the layer was 40\u2009nm and the active surface area was around 14 times of a plane electrode. It has been demonstrated by high resolution TEM images that the main oxide in the surface is a Ni-rich Ni-Fe hydroxides and a buffer layer of NiO that reduce the lattice mismatch [37]. This is in accordance with the detected structures of Ni(OH)2, which changes to NiOOH during electrolysis and reduces again. Other studies with successive potentiometric cycles has found similar results with an interlayer of 850\u2009nm of thickness [38].In the SS-AT sample the electrochemical impedance diagrams also show coupled HF and LF capacitive loops and a straight line of 45\u00b0 at the limit of the low frequency region after successive electrolysis cycles, see Fig. 5(a)-(b) and Figure S-4 in the supporting information. The high frequency (HF) loop is related to the NiOOH layer formed after electrolysis. The low frequency (LF) loop is related to the parallel combination between double-layer capacitance and the charge-transfer resistance of metal/electrolyte interface, and the straight line of 45\u00b0 is associated with the diffusion of oxygen away from the electrode surface. The charge-transfer resistance of the sample diminishes considerably after successive electrolysis cycles, see Table S-3 and Figure S-7 in the supporting information. However, the overall resistance values are much higher than those of the samples with NiFeP catalytic layer. According to DRT analysis (Figure S-4c), it is clear that the passive layer was growing after every cycle of electrolysis. The overpotential exhibited by the SS-AT after each cycle of electrolysis rises from 0.7\u2009V, until reaching a stable value of 1.0\u2009V at the 16\u00b0 cycle, see Fig. 5(d). This indicates that oxygen bubbles are accumulated on the surface [57]; but, unlike what was observed for electrodes with NiFeP catalytic layer, the NiOOH formed on the stainless-steel surface does not favor the release of the oxygen bubbles away from the electrode, probably due to incipient formation of the NiOOH compound on the SS substrate. Similarly, the evolution of the OCP values during the on/off electrolysis cycles shows an increase until reaching a stable value, see Fig. 5(e), which also indicates the accumulation of oxygen bubbles on the surface. Similarly, to what was observed for electrodes with NiFeP catalytic layer, the stainless steel undergoes a period of activation due to NiOOH formation, showing a considerable reduction in the overpotential at 10\u2009mA\u2009cm\u22122 and 100\u2009mA\u2009cm\u22122 by 56 and 60\u2009mV, respectively after the electrolysis cycles, see Fig. 5(f). As has been shown by other studies, the products of the corrosion of stainless-steel in alkaline media are majority metal hydroxides [19,21], which is coherent with what was observed by in-situ Raman spectroscopy, where the formation of NiOOH on the SS surface during electrolysis process at the OER potentials was corroborated. This compound is much more catalytically active that the normal oxide of Cr2O3 present in the surface of stainless steels.\n\nFig. 6 shows the results of the electrochemical evaluation of the NiFeP/SS-AT sample after electrolysis process (OER). Due to the presence of the catalytic layer of NiFeP on the anodic-treated stainless-steel, this electrode exhibited similar electrochemical behavior to that observed for the NiFeP/SS electrode during the on/off cycles of electrolysis. Fig. 6(a) shows the Nyquist diagrams of impedance measurements performed at the different times of electrolysis. After the first electrolysis process, two coupled capacitive loops were observed in the Nyquist diagram (see for more detail the Bode diagram of EIS in Fig. 6(c) and Figure S-5). Similarly to what was observed for the NiFeP/SS sample, the flattened capacitive loop at high frequency (HF) of the NiFeP/SS-AT electrode is related to the previously formed Ni(OH)2 layer, which gradually transforms to catalytic NiOOH layer; while the low frequency (LF) loop is related to the parallel combination between double-layer capacitance and the charge-transfer resistance of coating/electrolyte interface. After successive electrolysis cycles, both HF and LF capacitive loops continue to be coupled, but with a low associated impedance, and a straight line of 45\u00b0 at the limit of the low frequency region, associated with the diffusion of oxygen away from the electrode surface, begins to emerge, see Fig. 6(a)-(c). More detail of the features of EIS results can be observed in Figure S-5, where Bode plots and DRT analysis of the impedance measurements performed during the on/off cycles of electrolysis are presented. The reduction in the resistance of catalytic layers during on/off electrolysis process was corroborated, see Figure S-7. This sample exhibited the lowest values of impedance during the on/off cycles of electrolysis, as can be seen in Table S-3, Table S-4 and Figure S-7. In addition, the lowest values of overpotential jumps and more stable OCP values after each on/off cycle of electrolysis were observed in this sample, see Fig. 6(d)-(e). This indicates that the O2 accumulated on the electrode surface after OER during each cycle of electrolysis is more easily released away from the surface than what was observed for the other evaluated electrodes. As was shown in the in-situ Raman analysis performed on the NiFeP/SS-AT sample before and after electrolysis process, the bands at 475 and 555\u2009cm \u22121 associated with the NiOOH were clearly visible even at potentials lower than that where a high rate of oxygen bubbles occurs, i.e., 1.5\u2009V (10\u2009mA\u2009cm\u22121), and the bands continued to be present when the anodic polarization was returned to low values. The best condition for the formation of the active layer of NiOOH was that exhibited by the NiFeP/SS-AT electrode, showing the lowest overpotential values at 10\u2009mA\u2009cm\u22122 and 100\u2009mA\u2009cm\u22122 in the LSV curves performed before and after the electrolysis cycles, Fig. 6(f). Although there was little change in the LSV curves observed for the NiFeP/SS-AT electrode, this demonstrates that the changes undergone by this electrode in catalytic activity after on/off cycles are very low, and that the catalytic layer becomes more stable over time.The catalytic NiFeP layer electro-deposited on stainless-steel and naked anodic-treated stainless-steel experiences selective dissolution of iron after the oxygen evolution reaction during the electrolysis of water. The relative content of nickel in the surface of electrodes increases and a second amorphous catalytic layer of nickel oxyhydroxide (NiOOH) is formed as a corrosion product after the oxygen evolution reaction. The formation of the NiOOH layer after electrolysis, improve the catalytic response of the surface to OER. The formed NiOOH layer becomes more stable over electrolysis time. In addition, the presence of the NiOOH layer causes a gradual diminution in the size of the OCP jump after each on/off cycle, indicating that steady OCP values are reached more quickly at the end of the cycling process, and an easier release of oxygen bubbles from the electrode is achieved.Selective dissolution of a catalyst could be implemented as a key strategy to improve de catalytic behavior to OER and to improve the release of bubbles, which must be considered to future works.\nS. Cartagena: Investigation, Validation, Writing - original draft. J. A. Calder\u00f3n: Conceptualization, Methodology, Funding acquisition, Writing \u2013 review & editing.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors would like to thank Colombian Ministry of Science, Technology and Innovation \u201cMinciencias\u201d for financial support through the Colombia Scientific Program (Contract No FP44842-218-2018).Supplementary data associated with this article can be found in the online version at doi:10.1016/j.corsci.2022.110437.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  Green-hydrogen generation has become a focus for research due to its promising future as an energy vector. In this regard, one topic that has not been explored in depth, is the corrosion of catalytic layers under on/off operation in alkaline media during oxygen evolution reaction (OER) on the anode. Here, we studied the corrosion of, and changes to, the catalytic layer on stainless steel (SS) electrodes under different surface treatments. The results showed that there was formation of a passive and catalytic layer of NiOOH during the anodic polarization concomitantly with iron dissolution.\n               "}
{"full_text": "Steelmaking, one of the world\u2019s key industrial sectors, is highly energy-, and thus, carbon-intensive. The iron and steel sector directly accounts for 2.6 Gt of CO2 emissions annually, corresponding to about 7% of the global CO2 emissions [1]. Modern installations are reaching their limits in terms of efficiency improvement and energy and emission reduction. Further significant CO2 mitigation in the steel sector can only be realized with the introduction of breakthrough technologies, such as carbon capture, utilization and storage (CCUS). According to the International Energy Agency (IEA), CCUS technologies will play a central role to the industry decarbonization portfolio, contributing by about 15% to the required emissions reductions [2].While the removal of CO2 from the blast furnace gases via physical or chemical adsorption/absorption has been extensively studied ([3] and references therein), CCS has not been yet widely deployed in the steel industry, hampered by the high energy consumption of these capture technologies. In that context, CCU, i.e. the valorization of the carbonaceous steel-work off-gases for the production of useful products, holds more promise. The first commercial steel facility integrated with emissions conversion was launched in China in 2018, employing LanzaTech\u2019s low-temperature gas fermentation-to-ethanol technology with a capacity of 60,567\u00a0m3/y [4]. A similar plant is currently being constructed in ArcelorMittal\u2019s steel mill in Ghent, with support from the EU via the project STEELANOL [5]. Carbon Recycling International (CRI), the company that operates the first CO2-to-methanol production facility with a current capacity of 5 million liters of methanol per year, has achieved production of methanol at industrial scale using blast furnace gas from the SSAB steel mill in Lulea, Sweden, in the frame of the EU-funded project FReSMe [6]. CRI\u2019s technology is based on catalytic CO2 hydrogenation over CuO/ZnO/Al2O3-based catalyst under similar operating conditions to syngas-based methanol production, after careful removal of gas impurities [7].Despite the enormous progress, CO2 reduction remains a grand challenge, as commercial CuO/ZnO/Al2O3 catalysts have been reported to have unsatisfactory CO2 hydrogenation performance and low stability due to the negative \u201cflooding\u201d effect of H2O and the use of a hydrophilic alumina promoter component [8]. CO2 conversion becomes even more demanding when the CO2 source is industrial flue gases due to the co-presence of several other gases and impurities (e.g. CO, O2, N2, S- and N-compounds, inorganics, metals, etc.). Commercial Cu-Zn oxide catalysts have been recognized, early on, to be susceptible to sulfur, nitrogen, chlorine, phosphorus and Fe or Ni metal carbonyls [9 \u201311]. Ma et al. showed that Cu/ZnO deactivates sharply in syngas containing 3\u00a0ppm H2S, with CO conversion decreasing from 24.5% to lower than 1% within 7\u00a0h of reaction [12]. Deactivation was attributed mainly to the formation of ZnS and CuS, which destroyed the synergetic effect between Cu and ZnO. The work of Wood et al.\n[13] revealed the importance of the sulfur source, as the rate of deactivation was similar for thiophene and H2S, but negligible for COS.Although there is vast literature on the design and development of heterogeneous catalysts for direct CO2 hydrogenation to methanol for CCU applications (see recent reviews [7,14\u201318], and references therein), little attention has been paid to the effect of flue-gas impurities on catalyst performance and stability. This is especially important for the economic viability, and thus wide-scale deployment of such processes, as low catalyst poison tolerance necessitates the use of advanced and cost-intensive purification methods. In a recent study, Sch\u00fchle et al.\n[19] addressed the performance of In2O3/ZrO2 and commercial CuO/ZnO/Al2O3 for the hydrogenation of CO2 to methanol in the presence of typical impurities of industrial CO2 feed gas streams, i.e. SO2, H2S, NO2, NH3 and hydrocarbons (C1 \u2013 C3). The Cu-based catalyst was more resilient to sulfur, but still methanol productivity decreased by about 36% under severe sulfur poisoning by SO2 and H2S. The negative effect caused by nitrogen (either NO2 or NH3) was much less, in the order of\u00a0\u223c\u00a010%, while the presence of hydrocarbons caused a more than 50% decrease in methanol productivity. The focus of that work was however the In-based system. The poison-induced changes in the rate of deactivation under prolonged reaction times were also not considered.In this work, we investigate in a systematic way the deactivation and stability of a commercial CuO/ZnO/Al2O3 methanol synthesis catalyst, poisoned with typical contaminants present in steel-work off-gases, in the hydrogenation of CO2 to methanol. To accurately control the concentration of impurities on the catalyst surface, we poison the catalyst ex-situ with known amounts of contaminants. Besides the classical nonmetal impurities, i.e. H2S and NH3, we extend our study by considering the effect of Na, Ca and Fe. Fe in oxidic form is the major inorganic component of the dust that is produced in blast furnace and basic oxygen furnace steelmaking, followed by CaO/CaCO3\n[20,21]. Na, although present in smaller amounts [20], is strongly basic and could potentially have adverse effects on the reaction. The poisoned and untreated catalyst samples are characterized and tested under CO2/H2 mixture at industrially relevant reaction conditions as a function of time on stream.The catalyst used in this study was a commercial CuO/ZnO/Al2O3-based catalyst provided by Clariant. The catalyst was supplied in pellet form (6*4 mm). Prior to poisoning, the catalyst was ball milled and sieved to particle size 250\u00a0\u03bcm \u2013 350\u00a0\u03bcm.Poisoning with sulfur and nitrogen was performed ex-situ with H2S and NH3, respectively, to protect the high-pressure testing unit from corrosion. A Linseis STA PT-1750 thermogravimetric analyzer (TGA), offering precise temperature and gas flow control, was used for this purpose. The samples, each comprising 5\u00a0g of catalyst, were placed in a crucible in the TGA. Prior to the poisoning process, initial reduction took place by slowly ramping temperature and H2 content to finally 5% H2/N2 at 250\u00a0\u00b0C and holding this point for 4\u00a0h. The system was then purged with N2 for 1\u00a0h. For the sulfur poisoning, the catalyst was exposed at 250\u00a0\u00b0C for 5\u00a0h to a gas mixture of 0.3\u00a0vol% H2S/He diluted with an appropriate amount of N2 to achieve a final concentration of 400\u00a0ppm H2S in the feed gas. For the nitrogen poisoning, the catalyst was exposed at 250\u00a0\u00b0C for 10\u00a0h to a gas mixture of 500\u00a0ppm NH3/CH4 diluted with N2 to achieve final concentrations of 400\u00a0ppm NH3 in the feed gas. In both cases, the system was then purged again with N2 for 1\u00a0h at 250\u00a0\u00b0C and cooled down in N2. Before removing the sample from the setup, it was carefully passivated at ambient temperature step-wise with increasing concentrations of O2/N2 from 0.1 to 21\u00a0vol%.Poisoning with the Na, Ca and Fe metals was performed via incipient wetness impregnation of the catalyst with aqueous solutions of 1.5\u00a0mg/ml Na2CO3, 2.9\u00a0mg/ml Ca(NO3)2*4H2O and 3.6\u00a0mg/ml Fe(NO3)3*9H2O. After impregnation, the samples were dried in a compartment dryer at 80\u00a0\u00b0C for 18\u00a0h and subsequently calcined at 550\u00a0\u00b0C (ramp 100\u00a0\u00b0C/h) for 12\u00a0h in a muffle furnace.The catalysts are denoted as CuZnAl and CuZnAl_X (where\u00a0X\u00a0\u00a0=\u00a0S, N, Na, Ca and Fe) and refer to the untreated and treated samples, respectively.The treated samples were subjected to physiochemical characterization to determine the actual concentration of the impurities, both pre- and post-reaction. The S content was determined via tube furnace combustion, according to the ASTM D4239 method. The N content was measured with the Kjeldahl method, following a three-step process completed by titration for nitrogen. Finally, the Na, Ca and Fe content was determined by Inductively Coupled Plasma - Atomic Emission Spectroscopy (ICP-AES) on a spectrophotometer Optima 4300 DV (PerkinElmer).X-ray photoelectron spectroscopy (XPS) measurements were performed on a Kratos Analytical AXIS UltraDLD spectrometer equipped with a monochromatic Al K\u03b1 excitation source (\u03bbKa\u00a0=\u00a01486.6\u00a0eV), under ultra-high vacuum conditions (10\u22129 Torr). Binding energy (BE) referencing was employed using the adventitious carbon peak at 284.6\u00a0eV. Survey scans were recorded for surface elemental analysis (pass energy 160\u00a0eV) at an X-ray power of 180\u00a0W, while high resolution spectra were recorded at 20\u00a0eV pass energy with a step of 0.1\u00a0eV and an X-ray power of 225\u00a0W. Deconvolution of the peaks was performed with Kratos Vision software (version 2.2.1) using Shirley background subtraction and mixed Gaussian (70%) \u2013 Lorentzian (30%) functions. Auger electron spectroscopy (AES) was performed both via X-ray excitation under the same conditions as described above and via electron beam accelerated with a voltage of 10\u00a0kV.The catalytic performance of the materials in CO2 hydrogenation was evaluated in a continuous high-pressure dual fixed-bed reactor unit (Microactivity Effi-PID) with an electronic feed control system for both gaseous and liquid feeds. The unit\u2019s feeding system consists of three gas lines, each equipped with high accuracy mass flow controllers, and one liquid feed line, using a high precision pump. The stainless steel reactors (ID: 9.1\u00a0mm) are fitted with porous plates to ensure that the catalyst bed is located in the furnace\u2019s isothermal temperature zone. The reaction temperature is monitored with a thermocouple inserted in the catalytic bed. The exit stream of the reactors is sent to a liquid\u2013gas separator for separation and collection of the liquid and gaseous products.In a typical experiment, the required catalyst amount was diluted with equal amount of SiC (dp: 200\u00a0\u03bcm) and was introduced into the reactor. Prior to the measurements, the catalyst was pre-reduced in-situ by slowly ramping the temperature and hydrogen content to 5% H2/N2 at 250\u00a0\u00b0C and holding for 2\u00a0h. The reaction was conducted at temperature 250\u00a0\u00b0C, pressure 70\u00a0bar, GHSV 6500\u00a0h\u22121 and H2/CO2 feed molar ratio 3. Both liquid and gaseous products were analyzed with a GC 7890 gas chromatograph (Agilent Technologies) equipped with a dual FID/TCD detector. Mass and carbon balances typically closed to within 5%. All the gases used (H2, N2, CO2) were of purity above 99.999%. The molar conversion of the reactants and the carbon-based molar selectivity of the products was calculated as follows:\n\n(1)\n\n\nConversio\n\nn\ni\n\n\n\n\n\nmol\n%\n\n\n\n\n=\n\n\n\nn\n\ni\n,\ni\nn\n\n\n-\n\nn\n\ni\n.\no\nu\nt\n\n\n\n\nn\n\ni\n,\ni\nn\n\n\n\n\u2217\n100\n\n\n\n\nwhere \n\n\nn\n\ni\n,\ni\nn\n\n\n\n and \n\n\nn\n\ni\n,\no\nu\nt\n\n\n\n are the molar flows of reactant i at the inlet and outlet of the reactor.\n\n(2)\n\n\nSelectivit\n\ny\ni\n\n\n\n\n\nC\n-\nm\no\nl\n%\n\n\n\n\n=\n\n\nc\n\nn\n\ni\n.\no\nu\nt\n\n\n\n\nc\n\nn\n\nC\n\nO\nx\n\n,\ni\nn\n\n\n-\nc\n\nn\n\nC\n\nO\nx\n\n,\no\nu\nt\n\n\n\n\n\u2217\n100\n\n\n\n\nwhere \n\nc\n\nn\n\ni\n,\no\nu\nt\n\n\n\n is the C-molar flow of product i and \n\nc\n\nn\n\nC\n\nO\nx\n\n,\ni\nn\n\n\n\n and \n\nc\n\nn\n\nC\n\nO\nx\n\n,\no\nu\nt\n\n\n\n are the C-molar flow of CO2 at the inlet and outlet of the reactor.The ex-situ poisoned CuO/ZnO/Al2O3 samples were characterized to determine the concentration of the impurities on the catalytic surface. Table 1\n shows the nominal (calculated based on the poisoning experimental parameters) and the experimentally-determined concentration of the different impurities before and after reaction. The contamination of the catalyst with sulfur from gas phase H2S leads to high S concentration (\u223c65% of the nominal), confirming the known affinity of the catalyst components to sulfur. Several studies report the formation of sulfide phases, including ZnS, CuS, Cu2S, and CuSO4\n[12,22]. The results are very different for nitrogen, where the actual deposition is more than an order of magnitude lower than the nominal value. This suggests that gaseous nitrogen compounds, such as NH3, do not interact strongly with the catalyst components. Sch\u00fchle et al.\n[19] attained similar results and report the accumulation of only 0.02\u00a0wt% N on a commercial CuO/ZnO/Al2O3 after treatment in 500\u00a0ppm NH3 for 4\u00a0h. In the case of Na, Ca and Fe, the actual content of the metals is in all cases very close - as expected - to the nominal one, due to the impregnation procedure that was followed. It should be noted that the untreated catalyst also contains a small amount of residual Na (0.05\u00a0wt%), probably originating from Na2CO3 typically used as precipitation agent in the industrial preparation of CuO/ZnO/Al2O3 catalysts.The poisoned catalysts were also characterized post-reaction to check whether the contaminants are removed under prolonged exposure to the reaction conditions (see following section on catalyst testing). As shown in Table 1, there is a prominent reduction in the nitrogen content, and a milder, yet significant, decrease in sulfur. This suggests partial reversibility and/or weak binding of the sulfur and nitrogen species on the surface of the CuO/ZnO/Al2O3 catalyst. The contaminants are probably removed from the surface in the form of H2S and NH3 respectively under the high H2 pressure conditions of the reaction. Sch\u00fchle et al.\n[19] also showed substantial decrease in the sulfur levels of H2S- and SO2- poisoned In2O3/ZrO2 catalysts after the CO2 hydrogenation reaction. The metals appear to be more resilient; their concentration varies within experimental error before and after reaction.X-ray photoelectron spectroscopy (XPS) was employed to probe the catalyst surface and obtain insight on the surface elemental states. The Cu 2p, Zn 2p, Al 2\u00a0s and O 1\u00a0s core-level spectra of the ex-situ poisoned and non-treated CuO/ZnO/Al2O3 samples are shown in Fig. 1\n. The Cu 2p3/2 transition (Fig. 1, top left) demonstrates a main peak centered at\u00a0\u223c\u00a0933.3\u00a0eV, with a satellite peak at 941\u2013943\u00a0eV characteristic of Cu2+ species [23]. The Cu 2p3/2 peak is strongly asymmetric in the samples poisoned with nitrogen and sulfur. Via spectral deconvolution, three different Cu species can be identified on the catalysts\u2019 surface. In all samples, the main copper species is CuO at 933.2 \u2013 933.4\u00a0eV [24]. A second well-defined peak appears at\u00a0\u223c\u00a0935\u00a0eV. This high binding energy peak has been attributed to basic copper carbonate, Cu2(OH)2CO3, which forms either due to the exposure of the samples in atmospheric air and/or from the precursor salts used during the preparation of CuO/ZnO/Al2O3\n[25].The presence of metal\u2013oxygen and metal-carbonate bonds is also confirmed by the detection of the respective oxygen species in the O 1\u00a0s core-level spectra (Fig. 1, bottom right). The S- and N-poisoned samples exhibit in addition a third copper peak at 932.2\u00a0eV that corresponds to reduced Cu+/Cu0 species [26]. This is reasonable, considering that both these samples were poisoned in the gas phase in hydrogen atmosphere. Further chemical state differentiation between Cu+ and Cu0 is difficult with XPS alone, as the XP binding energies are practically the same [26], but is possible with Auger spectroscopy. The respective Cu LMM Auger spectra show peaks at 569.4 \u2013 569.6\u00a0eV (calculated Auger parameter 1849.6 \u2013 1849.7\u00a0eV) attributed to Cu+ species and at 568.2 \u2013 568.3\u00a0eV (Auger parameter 1850.7 \u2013 1850.8\u00a0eV) corresponding to metallic Cu. In both S- and N-poisoned samples, the majority of the reduced copper species are in\u00a0+\u00a01 valence state (greater than 80%).The Zn 2p3/2 core-level spectra (Fig. 1, top right) consist of a well-resolved peak at 1021.3\u00a0eV, originating from ZnO or Zn in close interaction with Cu [27]. The Al 2\u00a0s transition (Fig. 1, bottom left) also has one main peak at 118.7\u00a0eV associated with Al2O3\n[28]. The commonly used Al 2p line is not considered, as there is strong overlap with the Cu 3p transition. Overall, the binding energies of the main catalyst components, Cu, Zn and Al, do not show appreciable shifts, suggesting that the poisons (at the investigated concentrations) do not affect in a significant manner the nature of the main species on the CuO/ZnO/Al2O3 catalyst.The XP spectra of the poison elements are presented in Fig. 2\n. Despite the high background noise, which is a result of the low contaminants\u2019 concentration, some useful information can be deduced. The S 2p core-level spectrum (Fig. 2, top left) demonstrates a single peak centered at 168.9\u00a0eV, characteristic of sulfates. The binding energy of sulfur in both copper and zinc sulfate species is within the range of the recorded value [29]. Cu and Zn are known to react with H2S, forming sulfide species. ZnS formation is thermodynamically more favorable and thus, ZnO serves also as a sulfur trap, improving the S-poisoning tolerance of Cu/ZnO catalysts [30]. In fact, sulfur adsorption on ZnO is promoted in the presence of Cu [10]. Still, both ZnS and CuS have been previously detected on sulfur-deactivated catalysts [12,22,31]. Furthermore, CuSO4, ZnSO4, Zn3O(SO4)2 and Cu1.5ZnSO4(OH)3 phases were identified on Cu/Zn/Al2O3 catalysts poisoned under commercial and laboratory conditions [31]. It was suggested that in a commercial reactor, the conversion is so high that the gas phase becomes less reducing near the reactor exit, leading to sulfate formation [9]. Moreover, CuSO4 has been previously observed on Cu/ZnO catalysts, formed as a result of the thermal decomposition of CuS in O2-containing atmosphere during passivation [22,32,33]. Previous reports show that CuS is easily and irreversibly oxidized to CuSO4 by oxygen in humid environments [34]. It can thus be postulated that the poisoning procedure with H2S followed in our study possibly leads to the formation of zinc and copper sulfides on the surface that readily oxidize to sulfates during the passivation process and the subsequent exposure to atmospheric air.With regards to nitrogen, the surface concentration is below the detection limit (Fig. 2, top right) and therefore no information on the nitrogen compounds can be derived. Both Na and Ca exist on the surface in their respective carbonate form (Na2CO3 and CaCO3), as evidenced by the Na 1\u00a0s peak at 1071.6\u00a0eV [29] (Fig. 2, middle left) and the Ca 2p3/2 and Ca 2p1/2 doublet at 346.9\u00a0eV and 351.1\u00a0eV [29] (Fig. 2, middle right). Previtali et al. also detected the formation of CaCO3 by XRD in CuZnAl catalysts doped with 1% Ca [35]. The Fe 2p core-level spectrum (Fig. 2, bottom left) is complex, with multiplet splitting and satellite peaks. Due to the low resolution, unambiguous identification of the iron species is difficult, however the binding energy values and the strong satellites point to the presence of Fe2O3\n[36].Besides information on the oxidation state and the chemical environment of the different elements, XPS also allows the analysis of the elemental composition of the outer catalyst surface layers. Fig. 3\n (left) shows the surface composition of the poisoned and un-treated samples. The main component on the surface of the CuO/ZnO/Al2O3 catalyst is Cu, which is also the active site for the methanol synthesis reaction, with surface coverage of\u00a0greater than\u00a040\u00a0wt%. There is also high Zn surface coverage and substantial amount of carbon and oxygen that originate from the respective metal carbonates, metal oxides and air contamination. The controlled poisoning of the catalyst with the different contaminants causes a clear decrease of the Cu surface exposure, with a concurrent surface segregation of Zn. This possibly occurs due to partial sintering of the CuO/Cu particles during the different heating and calcination steps of the poisoning procedure. It also suggests preferential deposition of the contaminants on the Cu sites. The extent of surface Cu depletion varies between the different poisons, with the smallest effect caused by S and N and the largest by Ca and Fe. Wood et al.\n[13] showed by means of Auger electron spectroscopy and XRD that sulfiding at low levels of H2S (<4 ppm) favors the formation of surface sulfur adspecies on copper, thus decreasing the surface concentration of Cu, in agreement with our results. This trend was however reversed at higher concentrations of H2S, with the surface concentration of copper increasing beyond that of the fresh catalyst, as the ZnO phase started gradually being converted to ZnS. Previous studies on the effect of Na species in Cu/ZnO and Cu/ZnO/Al2O3 catalysts also showed lower Cu surface area due to blocking of the Cu sites by Na+\n[37,38].\nFig. 3 (right) shows on smaller scale the concentration of the contaminants on the surface. The bulk concentration (nominal and experimentally-determined) is also plotted for comparison reasons. It is interesting to note that the non-metallic elements, S and N, have less surface exposure than the bulk, as opposed to the three metallic elements, Na, Ca and Fe, that preferentially segregate on the catalyst\u2019s surface. The most striking difference is observed for Fe that has more than an order of magnitude higher surface concentration compared to the bulk. In multicomponent systems, the surface rearranges thermodynamically to the most stable configuration, i.e., the one with the lowest surface energy. Therefore, the surface becomes enriched in the constituent possessing the lowest surface tension. Experimental surface energy data is scarce and more often, theoretical values, based on DFT calculations, are reported. For CuO, the most stable crystal plane is CuO(111) with a calculated surface energy of 0.76\u00a0J/m2\n[39]. The respective values for the most stable crystal planes of CaCO3 (the calcium compound detected by XPS on the surface of the Ca-poisoned sample) are comparable, in the range of 0.73 \u2013 0.84\u00a0J/m2\n[40]. Unfortunately, no data could be retrieved for Na2CO3 detected in the Na-contaminated catalyst. On the other hand, the surface energy of FeOx phases is reported to be about half compared to CuO, ca. 0.36 \u2013 0.40\u00a0J/m2\n[41]. Therefore, the Fe-enriched surface in the CuZnAl_Fe sample could be possibly attributed to the lower surface energy of FeOx compared to CuO, which provides the thermodynamic driving force for the formation of an iron oxide capping layer on copper. Alternatively, the very high surface concentration of Fe might also originate from gradients of Fe over the macroscopic grains measured in XPS.The catalytic performance of the ex-situ poisoned catalysts was assessed as a function of time-on-stream at constant experimental conditions (T\u00a0=\u00a0250\u00a0\u00b0C, P\u00a0=\u00a070\u00a0bar, GHSV\u00a0=\u00a06500\u00a0h\u22121, H2/CO2 molar ratio\u00a0=\u00a03) to determine the impurities\u2019 effect both on the initial catalytic performance and the deactivation rate. The non-treated catalyst was also tested for comparison reasons. Measurements were taken after\u00a0\u223c\u00a024\u00a0h time-on-stream to ensure that the system reached steady-state. Over all samples, the carbon-containing reaction products consist only of CH3OH, produced via CO2 hydrogenation, and CO, produced via the reverse-water\u2013gas-shift (RWGS) reaction. No other products are detected even in ppm levels, confirming, in accord to the XPS findings, that the poisons do not alter the basic nature of the active site and thus do not catalyze other side reactions (at least at the investigated concentrations and conditions). The normalized conversion of CO2 with time and the activity losses in initial conversion and after 100\u00a0h TOS are shown in Fig. 4\n. The respective results for the normalized CH3OH carbon-based selectivity and the selectivity losses are presented in Fig. 5\n. The conversion and selectivity values were normalized over the initial value (first measured data point) recorded for the non-poisoned reference catalyst.The untreated catalyst exhibits relatively good stability, with about 5\u20136% loss in activity and selectivity after 150\u00a0h time-on-stream. It is widely accepted that the main deactivation process of Cu-based catalysts in syngas methanol synthesis is thermal sintering via a surface migration process and the growth of the Cu crystallites [9,42]. More recent studies show that ZnO plays a major role in the sintering process. Lunkenbein et al.\n[43] claimed that deactivation mainly occurs from changes in the ZnO moiety. The disruption of the Cu/ZnO synergy caused by the formation of Zn-Al mixed phases has also been suggested by Ficht et al.\n[44]. The presence of water, which is generated in abundance in methanol production from CO2, significantly accelerates deactivation. Liang et al.\n[45] investigated in detail the deactivation behavior of a CuZnAl catalyst in CO2 hydrogenation to methanol during 720\u00a0h TOS and showed that the agglomeration of ZnO species and the oxidation of metallic Cu are the main reasons for catalyst deactivation.The presence of impurities clearly degrades the performance of the CuO/ZnO/Al2O3 commercial catalyst. Depending on the contaminant\u2019s nature and concentration, there is decrease in both the initial activity and selectivity and the rate of deactivation as a function of time-on-stream. Nitrogen (which has the lowest concentration on the catalyst) has the smallest impact; it causes an\u00a0\u223c\u00a08% decrease in initial activity compared to the benchmark and negligible effect on the deactivation rate (in the same range as that of the un-treated sample). This agrees very well with the results of Sch\u00fchle et al.\n[19], who report that the treatment of a Cu/ZnO/Al2O3 catalyst for a period of 4\u00a0h with NH3 (leading to a N content of 0.02\u00a0wt%) decreases methanol productivity at 250\u00a0\u00b0C from 0.64 to 0.58 gMeOH/g Cu\n\u22121h\u22121 (9.4% reduction). Furthermore, they report no further deactivation at higher N concentrations.Sulfur, despite its much higher concentration than nitrogen, is moderately worse, with\u00a0\u223c\u00a010% lower initial CO2 conversion and CH3OH selectivity than the untreated sample. It accelerates however deactivation, as evidenced by the more negative slope of the performance curves as a function of time-on-stream in both Figs. 4 and 5. The activity and methanol selectivity drop by about 10% after 100\u00a0h time-on-stream, suggesting that the S-poisoned catalyst exhibits twice the deactivation rate of the non-treated and NH3-treated samples. In the study of Sch\u00fchle et al.\n[19], a 17% drop of methanol productivity at 250\u00a0\u00b0C is reported after 4\u00a0h of H2S-treatment leading to a sulfur content of 0.3\u00a0wt%. This discrepancy with our data could be due to differences in the composition of the employed CuZnAl catalyst. Data on the poisoned catalysts\u2019 stability are unfortunately not reported in the aforementioned study. Older works [31,46] examining the poisoning of Cu-based catalysts in H2S and thiophene-contaminated syngas showed that a Cu/ZnO/Al2O3 catalyst loses about 20% of its methanol synthesis activity with an average sulfur accumulation of 2%. This loss increases to 75% when the sulfur content increase to 12%.The contamination of the catalyst with metals has more pronounced impact on the CO2 hydrogenation performance. The ex-situ poisoning procedure followed for the metals involved the calcination of the catalyst at 550\u00a0\u00b0C for 12\u00a0h after impregnation to decompose the respective precursors. This higher temperature treatment could induce modifications in the structure and active phase dispersion of the catalyst and thus contribute to the observed deactivation. To address this, we calcined the untreated catalyst to these high-temperature conditions and re-measured its catalytic performance under identical experimental conditions. The normalized CO2 conversion and CH3OH selectivity, shown as open black symbols in Figs. 4 and 5 respectively, denote that the catalyst performance does not deteriorate as a result of the exposure to 550\u00a0\u00b0C. Therefore, the pronounced deactivation observed for the metal-poisoned samples is solely attributable to the impact of these elements on the Cu/ZnO/Al2O3 catalyst. Among the investigated metals, Na has the most detrimental effect. Sodium contamination reduces initial conversion and selectivity by\u00a0\u223c\u00a020\u201324% compared to the benchmark. It also accelerates the deactivation rate, with conversion and selectivity dropping additionally by 16% and 23% respectively after 100\u00a0h operation. In the reviews by Kung [9] and Twigg and Spencer [42], alkaline impurities appear to promote the production of higher alcohols and hydrocarbons from syngas; these high molecular weight deposits can eventually block the pores and reduce the activity. No formation of higher alcohols or hydrocarbons is detected here, in agreement with other studies on Na-doped Cu/ZnO and Cu/ZnO/Al2O3 studies in CO2 hydrogenation [37,38], underlining the differences between CO and CO2 conversion. Both these aforementioned studies report a strong decrease in CO2 conversion and methanol selectivity with increasing sodium content.Iron poisoning decreases the initial values of conversion and selectivity by\u00a0\u223c\u00a018%; the respective decrease for calcium is lower (\u223c12%). The order is reverse for the deactivation rate. Calcium demonstrates a higher deactivation rate than iron, so that the Fe- and Ca-poisoned samples present a very similar CO2 hydrogenation performance after 100\u00a0h time-on-stream. Literature on the poisoning effect of alkaline earth metals is scarce. Previtali et al. doped a commercial CuZnAl with 1\u00a0wt% Ca and reported lower CO2 conversion and CH3OH selectivity than the undoped catalyst [35]. More studies exist on the effect of iron, focused however on methanol production from syngas. These studies, summarized in the review of Chinchen et al.\n[46], report a decrease in methanol selectivity and total conversion in iron-doped catalysts, with an increase in the selectivity to methane, higher hydrocarbons and higher alcohols (not detected in this study). In general, the effects of a given level of iron impurity in a catalyst are strongly dependent on its form, oxidation state and distribution [42]. The addition of Fe was reported to enhance stability by preventing the sintering of Cu nanoparticles and inhibiting the oxidation of copper surfaces in CuO-ZnO-ZrO2-Al2O3/HZSM-5 [47] and CuO-Fe2O3-CeO2/HZSM-5 [48] bifunctional catalysts for CO2 hydrogenation to dimethyl ether, and in Fe-Cu/SiO2 catalysts for the high temperature reverse water gas shift reaction [49]. On the other hand, transition metal oxides, such as the Fe2O3 species formed on the surface of the Fe-poisoned catalyst in this study, can inhibit the formation of methanol, probably by surface coverage of the copper crystallites [42].To verify this hypothesis, we plot in Fig. 6\n the normalized initial conversion of CO2 over the poisoned and un-treated CuO/ZnO/Al2O3 samples as a function of the surface copper exposure, determined from the XPS analysis. The data reveal that the catalytic activity is closely related to the exposed Cu surface area. This observation fully supports the XPS findings and confirms that the poisons deactivate the catalyst by blocking part of the Cu active sites, either through association with Cu and formation of new phases (as in the case of sulfur) or mere deposition and coverage of the Cu sites (e.g., iron). Therefore, the poisons do not appear to modify the nature, but rather the number of the active sites that convert CO2. The only catalyst that deviates from this general trend is the Na-poisoned sample, which suffers more severe activity losses than expected based on the measured Cu surface area. We postulate that this stems from the strong basicity of sodium. Sodium contamination probably leads to the generation of basic sites on the catalyst surface that not only block active Cu sites, but also strongly bind the acidic CO2 and thus cause a considerable reduction in activity. Kondrat et al.\n[37] also observed discernable changes in the activities and selectivities of different loading Na+ doped catalysts with however comparable Cu surface areas. They thus concluded that Na+ acts directly as a poison, through increasing surface basicity, by blocking active sites or by inducing phase separation between Cu and ZnO.Besides the reduction in CO2 conversion, all samples, including the un-treated CuZnAl catalyst, demonstrate a progressive decrease in methanol selectivity with time-on-stream. The plot of CH3OH selectivity versus CO2 conversion for all catalysts and different times-on-stream reveals a very interesting trend. It is evident from the data in Fig. 7\n that, despite the differences in the initial CH3OH selectivity induced by the different poisons (see Fig. 5), the further decrease of selectivity is independent of the contaminant type and relies only on the extent of the activity reduction with time due to thermal sintering. As aforementioned, thermal sintering disrupts the Cu/ZnO synergy which is more crucial for methanol synthesis, but less important for the RWGS reaction [50]. This explains why deactivation reduces not only activity, but also methanol selectivity. The above reinforce the finding that the poisons block or render inactive a part of the active sites, but do not modify their intrinsic nature.In this work, we show that typical impurities present in steel-work off-gases, namely S, N, Na, Ca and Fe, can influence the performance and deactivation of a commercial CuO/ZnO/Al2O3 catalyst in the production of methanol from CO2. The exposure of the catalyst to gaseous H2S leads to high sulfur accumulation, confirming the known affinity of the catalyst components to sulfur. The interaction with NH3 is much weaker, resulting in low concentration of weakly-bound nitrogen species on the catalyst surface. The poisoning of the catalyst with Na, Ca and Fe via impregnation from the respective carbonate and nitrate salts leads to almost quantitative deposition of the metals. XPS characterization reveals that the contaminants preferentially block the Cu sites, as they cause a clear decrease of the Cu surface exposure, with concurrent Zn surface segregation. The extent of surface Cu depletion varies between the different poisons, with the smallest effect caused by S and N and the largest by Ca and Fe. However, XPS shows that the contaminants do not affect in a significant manner the oxidation state or the chemical environment of the main Cu, Zn and Al catalyst species.This is fully supported by the results of the catalyst testing in CO2 hydrogenation at 250\u00a0\u00b0C and 70\u00a0bar for more than 120\u00a0h time-on-stream. All impurities decrease the initial CO2 conversion; the extent of the activity loss is directly proportional to the poison-induced decrease in the Cu surface area, confirming that the poisons do not modify the nature, but rather the number of the active sites. The Na-contaminated catalyst deviates from this general trend, due to the strong basicity of sodium that creates basic sites on the catalyst surface that not only block active Cu sites, but also strongly bind the acidic CO2 and thus cause a more severe reduction in activity. Moreover, all catalysts, including the un-treated sample, suffer a progressive decrease in CH3OH selectivity with time, which is independent of the contaminant. We postulate that the preferential formation of CO with time-on-stream is due to thermal sintering that disrupts the Cu/ZnO synergy, which is more crucial for methanol synthesis but less important for the RWGS reaction.Overall, this study demonstrates that industrial CO2 off-gases can be exploited for the production of methanol over classical Cu/Zn/Al methanol synthesis catalysts after appropriate cleaning. The relatively high poison concentration used to accelerate deactivation and mimic the state of the catalyst after several months of continuous operation allows to extend the main findings of the study to real industrial operation, where the impurities would have a permanent presence in the gas feed stream and would accrue on the catalyst surface over time. Moreover, the irreversibility/partial reversibility of the contaminants under reaction conditions suggests that also in real in-situ poisoning, the reactants would compete with the poisons for the catalytic active sites, rendering the results relevant to industrial operation. Regarding flue gas cleaning, purification should primarily target the removal of metal impurities, such as alkali, alkaline earth and transition metals, that cause severe blocking of the active Cu sites. Sulfur and especially nitrogen removal are less critical, due to weaker and partially reversible binding on the catalyst surface.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was funded by the European Union through the Research Fund for Coal and Steel (RFCS), within the project entitled \u201ci3upgrade: Integrated and intelligent upgrade of carbon sources through hydrogen addition for the steel industry\u201d, Grant Agreement No 800659. The authors thank Clariant for providing the commercial methanol synthesis catalyst.", "descript": "\n                  The use of steel-work off-gases as CO2 source for methanol production is a promising strategy for the decarbonization of the steel industry. In this work, we show that typical impurities, namely S, N, Na, Ca and Fe, reduce the activity of a commercial CuO/ZnO/Al2O3 catalyst by blocking part of the Cu active sites. The extent of the activity loss is proportional to the poison-induced decrease in the exposed Cu surface area, suggesting that the poisons do not modify the nature, but rather the number of active sites. The only deviation is observed for Na, as its strong basicity leads to stronger binding of CO2 on the catalyst surface. Moreover, all catalysts suffer from methanol selectivity losses with time, due to thermal sintering that occurs similarly over the untreated and poisoned samples. Industrial CO2 off-gases can thus be exploited for methanol production after appropriate purification targeted primarily at the removal of metal impurities. The presence of sulfur and nitrogen is less critical to the catalyst performance.\n               "}
{"full_text": "The effects of global warming have prompted the search for technological alternatives to mitigate its impact and thus avoid the increase in greenhouse gas emissions, with CO2 and CH4 representing an important part of the total amount emitted into the atmosphere. Among these alternatives, the dry reforming of methane (DRM; CO2\u00a0+\u00a0CH4\n\n\n\u21cc\n\n 2 CO\u00a0+\u00a02H2) has been positioned as a technology that can reduce pollution while also acting as an important energy, source, thus allowing the development of a comprehensive system for capturing greenhouse gases [1].Developing efficient catalysts that allow the application of DRM on an industrial scale is essential for the implementation of this technology and to produce synthesis gas that can be used to obtain synthetic gasoline [2]. As the support plays a key role in the catalytic activity, it must therefore be carefully selected to allow full advantage to be taken of its physical and chemical properties, such as texture, thermal stability, redox properties, storage capacity, oxygen, and surface acidity-basicity [3,4]. This improves the metal-support interaction and increases the dispersion of active metal particles, thus minimizing the effects of C deposition [5,6]. Noble metals such as Ir, Rh, Ru, Pt, and Pd have a higher resistance to coke deposition than non-noble metals. Given that noble metals are more expensive than non-noble metals, an inexpensive way to prevent coke formation involves the use of multi-metal formulations of non-noble metals such as Ni, Co, and Fe with noble metals [7]. These formulations facilitate metal dispersion and generate more active metallic centers [8]. Ni is the only transition metal that exhibits catalytic properties comparable to those of precious metals. However, Ni-based catalysts tend to generate carbon deposits on the catalyst surface and, subsequently, a loss of catalyst activity. The resulting poor stability limits the commercial use of Ni-based catalysts for DRM reactions and, therefore, Ni-based catalysts must be modified\u2014in terms of the nature of the support and the preparation method\u2014to improve their performance and resistance to carbon deposition [4,9,10].Hexaaluminates are an excellent choice as DRM catalytic supports due to their thermal stability [11]\u2014they exhibit high thermal resistance above 1873\u00a0K. The general formula for hexaaluminates is ABxAl12-xO19, where AB represents a large cation such as Ba, La, Na, etc. and a transition (Co, Cu, Fe, Mn, Ni, etc.) or noble metal (Ir, Pd, Rh, Ru, etc.). These materials have been used as catalysts for high temperature applications, superionic conductors and luminescent laser materials, ceramics, and matrices for immobilization of radioactive elements, amongst others [12,13]. In recent years, several synthetic methods for the production of hexaaluminates have been developed, including lyophilization, nitrate decomposition, solid-state reaction, sol-gel, coprecipitation, inverse microemulsion, and hydrothermal synthesis [11]. The most widely used synthetic method is coprecipitation, in which the precursors are homogeneously mixed in the form of ions and precipitated simultaneously. In recent years, this method has been the object of studies aimed at improving the textural properties of the materials synthesized. In this regard, unconventional drying methods, as well as the use of nonconventional sources of raw materials, such as industrial inorganic waste, have been explored.Wastes known as aluminum saline slags are generated during aluminum recycling. These slags contain metallic Al, various oxides, and flux brines as main components, with variations in the percentages thereof depending on the nature of the material to be recycled [14,15]. Due to their limitations for final disposal in controlled landfills, these slags have been used recently to synthesize various materials, including alumina, calcium aluminate, layered double hydroxides, molecular sieves, microporous aluminophosphate, zeolites, pillared clays, and hexaaluminates [15\u201334]. The objective of these works is to synthesize materials with an application and that, therefore, can contribute to the recovery of industrial waste. The objective of the work would be framed within the so-called Circular Economy [38]. Logically, it is necessary to compare these new materials with the materials that are being used in these applications, to reduce and control the new emissions of pollutants generated and, if they can be applied, to analyze economically this route of recovery of inorganic industrial waste.In this work, a hibonite-type Ni/La-hexaaluminate synthesized from an industrial waste is used and compared as catalyst in DRM. The structure, catalytic behavior, and stability during a run time of at least 50\u00a0h of three Ni-catalysts obtained from two commercial supports and two preparation methods were used for comparison.Lanthanum(III) chloride heptahydrate (99.9%, Sigma-Aldrich), nickel(II) nitrate hexahydrate (99% Panreac), polyethylene glycol 400 (Merck), polyethylene glycol monolaurate 400 (PegMn400, Aldrich), and methanol (99.8%, ACS) were used as materials and reagents for the synthesis of the hexaaluminate and supported catalysts. Carbon dioxide (99.996%, Praxair), helium (99.999%, Praxair), hydrogen (99.999%, Praxair), methane (99.5%, Praxair), and nitrogen (99.999%, Praxair) were also used in the characterization and catalytic-performance studies.Aluminum was extracted from the saline slag using a previously reported procedure [33]; briefly, 50\u00a0g of saline slag was added to 750\u00a0mL of an aqueous reagent solution (HCl 2\u00a0mol/dm3) in a reflux system consisting of a 1000\u00a0cm3 Erlenmeyer flask with tube condenser, thus avoiding volume losses. The slurry was heated to 373\u00a0K and kept at that temperature for 2\u00a0h. The solution was then allowed to cool and separated by centrifugation. The most important constituents of the filtered solution were determined by ICP-OES using a VARIAN ICP-OES VISTA MPX with radial vision: Al (9.40\u00a0g/dm3), Ca (1.19\u00a0g/dm3), Fe (1.03\u00a0g/dm3) and Si (0.33\u00a0g/dm3).The synthesis of La-hexaaluminate-support was performed with a La/Al molar ratio of 1:11 using a previously reported and optimized method [33]. The slag solution was concentrated to one third of its initial volume to obtain a yellow liquor. A microelmulsion was then prepared using Methanol/Peg400/PegMn400/Al solution in a volumetric ratio of 1/0.8/0.4/0.6. The lanthanum chloride was mixed with the aluminum solution at 353\u00a0K, with vigorous stirring. After 10\u00a0min, the methanol was added slowly, the mixture stirred for a further 10\u00a0min, then Peg400 and PegMn400 were added and the temperature increased to 373\u00a0K. This mixture was kept under these conditions for 20\u00a0min prior to digestion in the autoclave. The resulting final mixture was heated in a stainless steel autoclave at 493\u00a0K for 16\u00a0h, drying in an oven until the liquid matrix had been eliminated, then calcined at 673\u00a0K for 1\u00a0h and 1473\u00a0K for 2\u00a0h, in both cases using a heating ramp of 10\u00a0K/min (Ni/LHA). Wet impregnation of the La-hexaaluminate support synthesized was carried out using 10\u00a0wt% of NiO, then the catalyst with the impregnated metallic phase was calcined at 673\u00a0K for 2\u00a0h. For comparison, the same method is used to prepare the Ni/LHA catalyst but using aluminum nitrate nonahydrate (\u226598%, sigma-Aldrich) as aluminum source [34]. Three reference catalysts were prepared using two methods, namely wet impregnation (I) and precipitation-deposition (PD), starting from two commercial oxides, \u03b3-Al2O3 (Rh\u00f4ne-Poulenc) and SiO2 (AF125, Kali Chemie), as supports (Ni\u2013I/Al2O3, Ni\u2013I/SiO2 and Ni-PD/SiO2) [39].The structural phases were analyzed using an X-ray diffractometer (model Siemens D5000) equipped with a Ni-filtered CuK\u03b1 radiation source (\u03bb\u00a0=\u00a00.1548\u00a0nm). The main textural properties of the solids were determined by nitrogen adsorption at 77\u00a0K using two Micromeritics ASAP (2010 and 2020 Plus) adsorption analyzers. Prior to the adsorption measurements, 0.3\u00a0g of sample was degassed at 473\u00a0K for 2\u00a0h at pressures lower than 0.133\u00a0Pa. The BET surface area (SBET) was calculated from the adsorption data obtained over the relative pressure range 0.05\u20130.20. The total pore volume (VP) was calculated from the amount of nitrogen adsorbed at a relative pressure of 0.99. Temperature-programmed reduction (TPR) studies were performed using a Micromeritics TPR/TPD 2900 equipment instrument. TPR tests were then performed from room temperature to 1273\u00a0K under a total flow of 30\u00a0mL/min (5% H2 in Ar, Praxair). Finally, the morphological analysis and chemical composition of the samples were carried out using a SEM Phenom XL desktop (Mode: 15\u00a0kV - Map, Detector: BSD Full) and HR-TEM (JEOL JEM 2100F, Accelerating voltage: 200\u00a0kV, Detector: X-Max).DRM was carried out at 973\u00a0K using an automated bench-scale catalytic unit (Microactivity Reference, PID Eng&Tech). The reactor was a tubular, \ufb01xed-bed, down\ufb02ow type, with an internal diameter of 0.9\u00a0cm and a length of 30\u00a0cm. Catalyst samples (25\u00a0mg) were mixed with an inert material (SiC, VWV Chemicals-Prolabo) to dilute the catalyst bed and avoid hot-spot formation. The reaction mixture consisted of CH4 and CO2 with a molar ratio of 1:1 (concentration of 12% in the feed), with helium as equilibrium gas up to a total feed flow of 40\u00a0cm3/min, thus achieving a gas hourly space velocity (GHSV) of 9.6\u00b7104\u00a0cm3/g h. Before the reaction, the catalyst was reduced in situ using 30\u00a0cm3/min of H2 at 973\u00a0K for 2\u00a0h. The reagent and product streams were analyzed using an Agilent 6890 gas chromatography system.The nitrogen adsorption isotherms for the supports and catalysts were of type II and IV in the BDDT classification [40] (see Fig.\u00a01\n). The specific surface area (S\n\nBET\n) and total pore volume (Vp) derived from the experimental adsorption results are summarized in Table 1\n. A decrease in textural parameters for the supported catalysts with respect to their corresponding support can be seen. In the case of the precipitation-deposition preparation method, reduction of the textural properties compared to the properties of the catalyst obtained by impregnation was not so important, and can be related to the formation of a talc-like nickel phyllosilicate structure during synthesis of the catalyst. This structure that is formed has been previously reported by our research group [41]. The incorporation of nickel through the wet impregnation method, and its subsequent drying and calcination to obtain NiO, causes the clogging of the porous structure. Under these conditions, the textural properties of the catalyst, specific surface area and pore volume are reduced, with a decrease from 304 to 225\u00a0m2/g and from 0.840 to 0.615\u00a0cm3/g. In the case of the precipitation-deposition preparation method, the textural properties are practically maintained to those corresponding to those of the support precisely due to the formation of the talc-like nickel phyllosilicate structure.The XRD patterns of the supported nickel catalysts are presented in Fig.\u00a02\n. In the case of the hexaaluminate, a very complex diffractogram was obtained. Based on the synthesis methods used and the presence of La and Ca, the most probable hexaaluminate structure appears to be magnetoplumbite (hibonite-Ca, pattern # 00-007-0785). The hexaaluminate obtained presents crystalline characteristics and different phases, as indicated in Fig.\u00a02. The patterns of the samples prepared by wet impregnation reveal the presence of NiO. In the case of the hexaaluminate, the structure of the support remains perfectly stable, without modification. Logically, supports with a higher specific surface area will favor the dispersion of NiO, which can favor the catalytic behavior of these materials. The pattern of the sample prepared by the precipitation-deposition method corresponds mainly to that of the silica support. The possible formation of talc-like nickel phyllosilicate compounds has been reported previously [39,41]. The crystallite sizes of NiO determined using the Scherrer equation can be found in Table 1.The TPR analysis provides information about the interaction between NiO and the support. Depending on the reduction temperatures, the degree of interaction of the NiO species can be classified into four different types: \n\u03b1\n, \n\u03b21\n, \n\u03b22\n and \n\u03b3\n [42]. The TPR patterns and TCD curves of the supported NiO catalysts are included and compared in Fig.\u00a03\n. A comparison between the TPR data and the main regions of this classification has also been included in Fig.\u00a03 A). The maximum peak temperatures and fraction of the total area represented by each can be found in Table 2\n. Fig.\u00a03 also shows the deconvolution of overlapping peaks using a Gaussian fit for determination. In the case of Ni/LHA, a single peak appears and then undergoes a complete reduction in the region of weak or poor NiO/support intercations, representing 100% of the total peak area. The peak found is not totally symmetric (maximum reduction temperatures at 612 and 652\u00a0K), thus suggesting that two types of NiO particles are present and that they may interact with a different surface that makes their reduction rate slightly different. A reduction behavior similar to the previous one can be seen in the case of Ni\u2013I/SiO2, but in this case it also encompasses the next interaction stage (\n\u03b21\n). This behavior could be related to NiO particles dispersed on the surface of the easily reducible support (651\u00a0K) and other particles occluded in the porous structure, which are more difficult to reduce. The area of the first peak corresponds to 44%, whereas the other peak accounts for 56% of the total. A shift of the TPR peaks to higher temperatures is observed for the samples prepared by the precipitation-deposition method (NiO-PD/SiO2) and considering Al2O3 as support. In these two cases the temperatures of the reduction maxima are shifted to 871 and 981\u20131044\u00a0K, respectively. These catalysts exhibit four peaks with maximum peak area percentages of 71% (Ni-PD/SiO2) and 74% (Ni\u2013I/Al2O3) in the interaction stages \n\u03b21\n and \n\u03b22\n. The main difference between the two catalysts is that the reduction temperatures found for the Ni\u2013I/Al2O3 catalyst are shifted to a higher temperature. Indeed, the highest temperature of one of the reduction peaks in this case is 1044\u00a0K. These results confirm the findings of the X-ray structural analyses, namely that the formation of various nickel compounds makes it more difficult to reduce than the NiO metallic oxide. These results are in according to the degree of interaction/reaction of nickel with the surface of the SiO2 and Al2O3 supports, aspects previously referenced in the literature [41]. The structure of \u03b3-alumina and the size of Ni2+ allow the formation of Ni\u2013Al spinel, especially if it favors temperature. Under these conditions, the reduction temperature of nickel increases considerably when compared to the reduction of NiO particles dispersed on a support. This is the situation observed in Ni\u2013I/LHA and Ni\u2013I/SiO2. When the preparation method is modified so that the degree of nickel interaction with the support surface (Ni-PD/SiO2) increases, the reduction temperature also increases, precisely because of the presence of this strong interaction.The DRM reaction (CO2\u00a0+\u00a0CH4\n\n\n\u21cc\n\n 2 CO\u00a0+\u00a02H2) is affected by several parallel reactions that occur during the catalytic process: methane decomposition (CH4\n\n\n\u21cc\n\n C\u00a0+\u00a02H2), the reverse water-gas shift reaction (RWGS; CO2\u00a0+\u00a0H2\n\n\n\u21cc\n\n CO\u00a0+\u00a0H2O), the Boudouard reaction (2 CO \n\n\u21cc\n\n C\u00a0+\u00a0CO2), CO2 hydrogenation (CO2\u00a0+\u00a02H2\n\n\n\u21cc\n\n C\u00a0+\u00a02H2O), CO hydrogenation (CO\u00a0+\u00a0H2\n\n\n\u21cc\n\n C\u00a0+\u00a0H2O), and steam reforming (CH4\u00a0+\u00a0H2O \n\n\u21cc\n\n CO\u00a0+\u00a03H2). The conversion and ratio of CO2 and CH4, as well as the selectivity with respect to hydrogen (H2/CO), can give an idea of the prevalence of these reactions during DRM.The conversion (CO2\u2013CH4), selectivity (H2/CO), and carbon balance (CB) results obtained for the catalysts during a long 50\u00a0h catalytic test are presented in Fig.\u00a04\n. The carbon dioxide and methane conversions [X]i, selectivity [H2/CO] and carbon balance (CB) were calculated using the following equations:\n\nEquation 1\n\n\n\n\n[\nX\n]\n\n\nC\n\nH\n4\n\n\n\n=\n\n\n\n\n[\n\nC\n\nH\n4\n\n\n]\n\n\ni\nn\n\n\n\u2212\n\n\n[\n\nC\n\nH\n4\n\n\n]\n\n\no\nu\nt\n\n\n\n\n\n[\n\nC\n\nH\n4\n\n\n]\n\n\ni\nn\n\n\n\n\n\n\n\n\n\nEquation 2\n\n\n\n\n[\nX\n]\n\n\nC\n\nO\n2\n\n\n\n=\n\n\n\n\n[\n\nC\n\nO\n2\n\n\n]\n\n\ni\nn\n\n\n\u2212\n\n\n[\n\nC\n\nO\n2\n\n\n]\n\n\no\nu\nt\n\n\n\n\n\n[\n\nC\n\nO\n2\n\n\n]\n\n\ni\nn\n\n\n\n\n\n\n\n\n\nEquation 3\n\n\nS\ne\nl\ne\nc\nt\ni\nv\ni\nt\ny\n\n[\n\n\nH\n2\n\n\nC\nO\n\n\n]\n\n=\n\n\n\n\n[\n\nH\n2\n\n]\n\n\no\nu\nt\n\n\n\n2\n\u2217\n\n\n[\n\nC\n\nH\n4\n\n\n]\n\n\ni\nn\n\n\n\n\n\n\n\n[\n\nC\nO\n\n]\n\n\no\nu\nt\n\n\n\n\n\n[\n\nC\n\nH\n4\n\n\n]\n\n\ni\nn\n\n\n+\n\n\n[\n\nC\n\nO\n2\n\n\n]\n\n\ni\nn\n\n\n\n\n\n\n\n\n\n\n\nEquation 4\n\n\nC\nB\n=\n\n\n\n\n[\n\nC\n\nO\n2\n\n\n]\n\n\no\nu\nt\n\n\n+\n\n\n[\n\nC\nO\n\n]\n\n\no\nu\nt\n\n\n+\n\n\n[\n\nC\n\nH\n4\n\n\n]\n\n\no\nu\nt\n\n\n\n\n\n\n[\n\nC\n\nH\n4\n\n\n]\n\n\ni\nn\n\n\n+\n\n\n[\n\nC\n\nO\n2\n\n\n]\n\n\ni\nn\n\n\n\n\n\n\n\n\nIn the case of Ni\u2013I/SiO2, this catalyst showed a very different thermodynamic behavior from the rest, with the conversion of CH4 and CO2 and selectivity decreasing during the DRM reaction, with slopes of \n\n\n\u0394\n\nX\n\nC\nO\n2\n\n\n\u2248\n\u2212\n\n6\n\n%\n\n, \n\n\u0394\n\nX\n\nC\nH\n4\n\n\n\u2248\n\u2212\n\n14\n\n%\n\n, \n\n\u0394\n\nX\n\nC\nH\n4\n\n\n\u2248\n\u2212\n\n7\n\n%\n\n (see Fig.\u00a04 A) B) C)) and, therefore, an increasing slope for the CB during the first 35\u00a0h (3%), subsequently stabilizing to a mean value of 98%. This behavior can be related to a high deposition of coke, which explains the low catalytic performance as a result of deactivation of the catalyst at a constant rate, practically during the entire test run. The blocking effect of the active sites could be attributed to the morphological transformations that the support and metallic phase undergo during the reduction stage, as an effect of a weak interaction (type \u03b1) in which the nickel nanoparticles are practically free and/or weakly fixed on SiO2. This situation causes a high rate of diffusive migration on the surface of the catalyst, thereby greatly benefiting the sintering and growth of the NiO grains. This effect is induced by the thermal gradient and the differences between the calcination and reduction temperature before the DRM test, as well as the impregnation method used for deposition of the metallic phase. In the case of Ni-PD/SiO2 and Ni\u2013I/Al2O3, the catalytic behavior of these catalysts is much more stable than for Ni\u2013I/SiO2, with a slight increase in the case of Ni-PD/SiO2 with respect to Ni\u2013I/Al2O3. The CO2 conversion presents mean values of between 73% and 75% (see Fig.\u00a04 A)). The CH4 conversion also maintains the same thermodynamic regime for both catalysts with average values of between 82% and 85%. With regard to the H2/CO selectivity (see Fig.\u00a04C)), Ni-PD/SiO2 presents an average value of 98% versus 94% for Ni\u2013I/Al2O3, and in the case of the CB (see Fig.\u00a04 D)), in the first 40\u00a0h they show different behaviors, with Ni\u2013I/Al2O3 presenting an increasing behavior, stopping at 85% and stabilizing at around 95%, whereas Ni-PD/SiO2 remains at around a mean value of 98%. Over the last 10\u00a0h, both catalysts essentially stabilize in the same thermodynamic regime. The precipitation-deposition method (PD) allows a significant improvement in the catalytic performance for Ni-PD/SiO2 compared to Ni\u2013I/SiO2. These results seem to indicate that the strong interaction of nickel with the support favors the stability of the DRM reaction. The hexaaluminate catalyst (Ni/LHA) exhibited the best catalytic performance in terms of yield and stability, with CO2 conversion values of 80% at the beginning of the reaction and 75% at the end (\n\n\u0394\nX\n\n\u00a0=\u00a0\u22122%). CH4 (opposite behavior to that of CO2), in turn, presents an increasing slope of 4% up to a final value of 85%, a behavior which is better than that of the aforementioned catalysts at all times (see Fig.\u00a04 A), B)). With regard to H2/CO selectivity and CB (see Fig.\u00a04C), D)), Ni/LHA continues to show a consistent thermodynamic behavior, with mean H2/CO selectivity values of 99% and a CB of 76%, thus exhibiting the best H2/CO ratio and the lowest rate of deactivation by coke deposition. Similarly to the strong interaction of nickel with the support, the presence of Al in the catalyst also seems to favor the stability of the reaction.SEM, TEM, and TEM-HAADF images for NiO/LHA are shown in Fig.\u00a05\n before the reduction stage and after the catalytic test. The morphology of the catalyst (see Fig.\u00a05 A) and B)) shows spherical agglomerates of NiO/LHA for the fresh catalyst and a rosette-like morphology for the used catalyst. The carbonaceous deposits generated during DRM, which were identified as filamentous carbon and carbon nanotubes (see Fig.\u00a05 B) E)) that displace the Ni0 grains (distribution in the range dp\u00a0\u2248\u00a010\u201350\u00a0nm, see Fig.\u00a05 F)) from the catalyst surface, are also found. Although these forms of coke still block the active metallic sites, they are the least harmful, thus allowing this catalyst to maintain great stability and excellent performance in DRM at 973\u00a0K. Another important aspect to highlight is that the amount of carbon produced is much lower than the critical concentration necessary to completely exhaust the catalyst and cause it to lose its catalytic ability compared to other catalysts, which in turn exhibit better textural properties.A Ni/La-hexaaluminate catalyst has been synthesized using an aluminum saline slag\u2014a hazardous waste generated in aluminum recycling\u2014as aluminum source in the synthesis of the La-hexaaluminate used as catalytic support. The catalysts were synthesized by impregnation. This method generated two types of morphologies, namely rosettes and clusters of amorphous tables, which allow a good distribution of the metallic nanoparticles on the support, as determined by SEM and HR-TEM. The catalyst showed excellent stability after 50\u00a0h of DRM reaction at 973\u00a0K. The conversion of CH4 is higher than CO2 and the H2/CO selectivity is about 99%, thus suggesting the predominance of the Boudouard reaction over the RWGS reaction. The behavior of this catalyst is comparable to that of Ni\u2013I/Al2O3 and Ni-PD/SiO2, which is related to both the Ni-support interaction and the presence of alumina.All the authors conceived, designed, and performed the experiments, analyzed the data, and drafted the manuscript.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The authors are grateful for financial support from the Spanish Ministry of Science and Innovation (MCIN/AdslEI/10.13039/501100011033) through project PID2020-112656RB-C21. JJTH thanks Universidad P\u00fablica de Navarra for a postdoctoral grant. AG also thanks Santander Bank for funding via the Research Intensification Program.", "descript": "\n                  In this work, a hibonite-type Ni/La-hexaaluminate (Ni/LHA) synthesized from an industrial waste is used and compared as catalyst in the dry reforming of methane (DRM) at 973\u00a0K. The structure, catalytic behavior, and stability during a run time of at least 50\u00a0h of three Ni-catalysts obtained from two commercial supports and two preparation methods were used for comparison. An aluminum solution (9.40\u00a0g/L) obtained from an aluminum saline slag waste by acid extraction was used to synthesize the hexaaluminate by mixing with a stoichiometric amount of lanthanum nitrate and methanol/Peg400/PegMn400 under hydrothermal conditions at 493\u00a0K for 16\u00a0h. The Ni/LHA catalyst (10\u00a0wt% NiO) was obtained by impregnation of the synthesized support, calcined previously at 1473\u00a0K for 2\u00a0h. The resulting solids were characterized by several techniques as: X-ray diffraction (XRD), N2 adsorption at 77\u00a0K, temperature-programmed reduction (TPR), scanning electron microscopy (SEM) and transmission electron microscopy (HR-TEM). In order to compare the catalytic behavior and properties of the Ni/LHA catalyst, three Ni catalysts obtained from two commercial supports (\u03b3-Al2O3 and SiO2) and two preparation methods (wet impregnation (I) and precipitation-deposition (PD)) were synthesized. Analysis of the TPR patterns for the catalysts allowed the type of metal support interaction and NiO species to be determined, with a weak interaction with the support being observed in Ni/LHA and Ni\u2013I/SiO2. The NiO species observed, with crystallite sizes between 9.7 and 40.4\u00a0nm, confirm the X-ray structural analyses. The Ni/LHA catalyst was found to be active and very stable in the DRM reaction after 50\u00a0h. The catalytic behavior was evaluated from the CO2 and CH4 conversions, as well as the H2/CO selectivity, with values of 99% over almost all the time range evaluated. The behavior of this catalyst is comparable to that of Ni\u2013I/Al2O3 and Ni-PD/SiO2. The results found indicating that the strong interaction of nickel with the support favors the stability of the catalysts in the DRM reaction.\n               "}
{"full_text": "With the rapid development of the automobile industry, the pollution of the environment by automobile exhaust is becoming increasingly serious [1,2]. SOx generated by incomplete combustion of sulfur compounds in gasoline not only is the main source of acid rain but also can significantly reduce the conversion efficiency of vehicle exhaust to NOx, incomplete combustion hydrocarbons and particulate matter and aggravate environmental pollution [3\u20135]. The sulfur content of gasoline has been strictly regulated worldwide [6,7]. China fully implemented the national V gasoline standard in 2017, which requires the sulfur content to be reduced to 10\u00a0\u03bcg/g [8,9]. The main sulfur compounds in gasoline include mercaptans, sulfides, disulfides and thiophenes [10]. Mercaptans and thioethers are easily removed due to their high reactivity, while thiophene sulfides are the most difficult to remove due to their low reactivity [11\u201316]. The petroleum fraction contains thiophene sulfides. To achieve deep desulfurization, the pressure of the hydrogenation reaction should not be less than 2.0\u00a0MPa, and the temperature should not less than 300\u00a0\u00b0C [17,18]. Moreover, thiophene sulfides account for more than 85% of the total sulfur content in gasoline. Therefore, the conditions for deep desulfurization of gasoline by the hydrogenation method are very strict. Generally, the operating pressure is 3\u20135\u00a0MPa, the temperature is 300\u2013450\u00a0\u00b0C [19,20], and the high-pressure method must reach 10\u00a0MPa, which correspondingly increases the material requirements of the equipment, requires materials resistant to high temperature and high pressure, and greatly increases the operating costs as well as the risk factors related to high-temperature and high-pressure operation [21,22]. As an emerging desulfurization technology, photocatalytic oxidation has attracted increasing attention due to its mild reaction conditions and low cost [22\u201326]. Some materials can be excited only by ultraviolet light, such as Ti, which greatly limits the use of visible light. Therefore, the development of a catalyst with visible light response and its application in the field of photocatalytic oxidative desulfurization has important significance and potential.NiO is a typical p-type semiconductor material and is considered one of the most promising photocatalysts because there are many vacancies in the 3d orbital of Ni2+, which can promote electron migration, resulting in the existence of more holes in the crystal [27,28]. Due to the wide band gap (3.6\u00a0eV) of NiO, the absorption efficiency of visible light is reduced. By combining NiO with an n-type semiconductor, the absorption range of NiO can be changed, and the utilization of visible light can be improved, thus improving the photocatalytic desulfurization performance. Dong et al. [29] successfully prepared NiO/g-C3N4 heterojunction photocatalysts by the ammonia evaporation method and investigated the catalytic performance of the composite photocatalyst under visible light. Compared with that of g-C3N4, the catalytic activity of the NiO/g-C3N4 heterojunction photocatalyst has been greatly improved. The research shows that the main reason for this improvement is that the larger specific surface area and heterojunction structure inhibit the recombination of photogenerated electrons and holes. In recent years, bismuth-based semiconductors have received extensive attention from researchers, and Bi2WO6 has been studied in-depth due to its narrow band gap, large visible light response range, and strong oxidation ability. Zhang et al. [30] investigated the effects of different hydrothermal treatment conditions on the particle size, crystal form and morphology of Bi2WO6 and studied its light absorption characteristics. Research shows that prepared Bi2WO6 has strong ultraviolet-visible absorption characteristics and luminous intensity and has good photocatalytic performance. In this paper, NiO and Bi2WO6 were synthesized to prepare new p-n heterojunction semiconductors by hydrothermal and high-temperature calcination methods. The advantages of NiO and Bi2WO6 are complementary to maximize the performance of photocatalytic oxidation desulfurization and achieve the deep desulfurization of gasoline.To meet the national fuel standards and reduce the air pollution caused by automobile exhaust gas, a photocatalyst with strong light absorption performance and high efficiency of photogenerated electron\u2013hole separation was prepared in this paper, which exhibited high photocatalytic oxidation performance and was applied to gasoline desulfurization. Using Na2WO4\u00b72H2O, Bi(NO3)3\u00b75H2O and Ni(NO3)2\u00b76H2O as raw materials, the NiO-Bi2WO6 composite photocatalyst was prepared by hydrothermal and high-temperature calcination methods and applied to the study of benzothiophene (BT) desulfurization in gasoline. The photocatalyst was characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), ultraviolet-visible diffuse reflectance spectroscopy (UV\u2013Vis-DRS) and other means, providing sufficient evidence for the synthesis of NiO-Bi2WO6 photocatalytic composite materials. In the application of photocatalytic oxidation desulfurization, BT is oxidized into the highly polar benzothiophene sulfone (BTO2) through continuous optimization of the process conditions, which is then extracted by acetonitrile to achieve deep desulfurization.Na2WO4\u00b72H2O, Bi(NO3)3\u00b75H2O, Ni(NO3)2\u00b76H2O, Na2C2O4, C3H8O, C6H4O2, dilute nitric acid, NaOH, thiophene (Th), benzothiophene (BT), dibenzothiophene (DBT), n-octane, acetonitrile, deionized water and absolute ethanol (all the above are analytical pure reagents).The morphology and lattice structure of the catalyst were observed by SEM with a Zeiss Merlin Compact system and by TEM with a Talos F200X system. An Ultimate IV X-ray powder diffraction analyzer (Rigaku Corporation) was used to determine the crystal structure of the catalyst. Cu-Ka rays were used as the laser light source in the experiment. The scanning range was 10\u00b0\u201380\u00b0, and the scanning speed was 5\u00b0/min. The visible light absorption performance of the catalyst was characterized by a U3900H UV\u2013Vis diffuse reflectance absorption spectrometer (Hitachi Company, Japan), and the wavelength measurement range was 200\u2013800\u00a0nm. The elemental composition of the catalyst surface was analyzed by a Thermo Fisher Scientific k-alpha+ XPS system. Electrochemical testing was carried out by an electrochemical workstation (Zahner PP211, Germany). The test conditions were Ag/AgCl as the reference electrode and Pt wire as the counter electrode [31\u201333]. To prepare a working electrode, NiO/Bi2WO6 powder was deposited on conductive glass with a frequency of 0.1\u00a0*\u00a0106\u00a0Hz and a 150\u00a0W cold light source. Organic sulfur in BT was identified by a gas chromatography (GC) system coupled with a flame ionization detector (Agilent 7820A), which was equipped with a capillary column (HP-5, 25\u00a0mm\u00a0\u00d7\u00a00.25\u00a0mm\u00a0\u00d7\u00a00.25\u00a0mm). Heating program: The initial temperature was 120\u00a0\u00b0C, held for 1\u00a0min, increased to 180\u00a0\u00b0C at a rate of 10\u00a0\u00b0C/min, held for 1\u00a0min, and then increased to 250\u00a0\u00b0C at a speed of 20\u00a0\u00b0C/min for 5\u00a0min [34,35].First, 0.01\u00a0mol of Bi(NO3)3\u00b75H2O was added to 50\u00a0m of 0.4\u00a0mol/L nitric acid solution and stirred continuously. Then, 0.005\u00a0mol of Na2WO4\u00b72H2O was added into 50\u00a0ml of deionized water and slowly added to the above solution. After stirring for 1\u00a0h, the mixed solution was transferred into a stainless steel autoclave. The solution temperature was maintained at 150\u00a0\u00b0C for 24\u00a0h. After the reaction, the solution was naturally cooled for 2\u00a0h to room temperature. The precipitates were centrifugally separated, washed with deionized water and absolute ethanol 3 times, and dried at 60\u00a0\u00b0C for 12\u00a0h to obtain the Bi2WO6 sample.First, 0.006\u00a0mol of Ni(NO3)2\u00b76H2O and 0.0003\u00a0mol of NaOH were separately dissolved in 20\u00a0ml of deionized water, and the latter solution was transferred dropwise to the former solution. After adjusting the pH to 8 with nitric acid, the proper amount of Bi2WO6 was added, and the mixture was stirred for 1\u00a0h, ultrasonicated for 15\u00a0min, and then placed into a hydrothermal kettle at 180\u00a0\u00b0C for 16\u00a0h. After the reaction, the precipitate was filtered, washed with water and with alcohol 3 times each, and dried at 180\u00a0\u00b0C for 2\u00a0h. Then, the precipitate was placed in a muffle furnace and heated at 350\u00a0\u00b0C for 2\u00a0h, yielding NiO-Bi2WO6. According to the above method, NiO-Bi2WO6 samples with different NiO loadings of 10%, 20%, 30%, 40%, 50%, and 100% were sequentially prepared.First, 0.2097\u00a0g of BT was dissolved in 250\u00a0ml of n-octane solution and then mixed evenly to prepare a model oil solution with a sulfur content of 200\u00a0mg/L. Then, 40\u00a0ml of model oil solution and an appropriate amount of catalyst were added into a quartz tube and placed in the photochemical reactor. The reaction solution was first reacted in the dark for 30\u00a0min to reach adsorption-desorption equilibrium. Next, the light source was turned on, a 5\u00a0ml sample was taken every 30\u00a0min, and the extract was extracted with acetonitrile. Then, the upper solution was taken, and the sulfur content was determined with a WK-2D microcoulomb analyzer. By comparing the sulfur content results with that of the original solution, the desulfurization rate was obtained, and the desulfurization effect was analyzed. The desulfurization rate is calculated according to the following formula, where \u03b7 is the desulfurization rate, C0 is the sulfur content of the solution before the reaction, and Ct is the sulfur content of the solution after the reaction.\n\n\n\u03b7\n=\n\n\n\nC\n0\n\n\u2212\n\nC\nt\n\n\n\nC\n0\n\n\n\u00d7\n100\n%\n\n\n\n\nFig. 1\n shows the SEM images of the sample, where (a) (b) presents the morphology of pure Bi2WO6, and panels (c) and (d) depict the morphology of the NiO-Bi2WO6 composite photocatalyst. As shown in Fig. 1 (a) (b), the pure Bi2WO6 sample has a spherical shape, which is composed of flakes and has a hydrangea-like structure resembling a bird's nest. Fig. 1 (c) (d) shows that the NiO-Bi2WO6 composite still maintains a complete spherical structure. Fig. 1 (d) shows that there are deposits on the surface of Bi2WO6, and the dispersion is good, with no agglomeration. The supported material was thus preliminarily determined to be NiO.To study the lattice structure of NiO-Bi2WO6 in depth, the samples were evaluated by TEM. As shown in Fig. 2\n (a), the composite material exhibits a spherical structure. Fig. 2 (c) presents a local enlarged view of the red box shown in Fig. 2 (b). In panel (c), the lattice spacing of 2.59\u00a0nm corresponds to the (002) crystal planes of Bi2WO6 (JCPDS No. 39-0256), and 2.09\u00a0nm corresponds to the (200) crystal planes of NiO (JCPDS No. 65-2901), indicating that the prepared material was NiO-Bi2WO6. In addition, the EDS images of the corresponding elements in Fig. 2 (d\u2013h) show that the composite was composed of four elements: Bi, W, O and Ni. The elemental mappings of Bi, W and O were consistent with the shape of the composite, and Ni was evenly distributed on the composite. Fig. 2 (i) presents the content diagram of each element in the composite. The atom number ratios are Bi:W:O\u00a0=\u00a02:1:6 and Ni:O\u00a0=\u00a01:1, confirming the successful synthesis of NiO-Bi2WO6 composites.To study the phase composition and crystal structure of NiO-Bi2WO6, the samples were characterized by XRD, as shown in Fig. 3\n. In Fig. 3 (a), characteristic peaks appeared at 28.4\u00b0, 32.9\u00b0, 47.2\u00b0, 56.0\u00b0, 58.6\u00b0, 75.7\u00b0 and 78.2\u00b0, corresponding to the (131), (002), (202), (133), (262), (391), and (460) crystal planes of Bi2WO6 (JCPDS No. 39-0256) [36], respectively. The diffraction peaks at 37.2\u00b0, 43.3\u00b0, and 62.8\u00b0 corresponded to the (111), (200), and (220) crystal planes of NiO (JCPDS No. 65-2901), respectively. There were no heteropeaks in the figure, and the diffraction peaks were sharp and obvious. This result confirms that we successfully synthesized a NiO-Bi2WO6 catalyst with increased purity and improved crystallinity. Fig. 2 (b) presents the XRD comparison diagram of the photocatalyst before and after the reaction. The figure shows that the catalyst maintains an obvious characteristic peak at the same diffraction angle before and after the reaction. The diffraction peak intensity of the catalyst was only slightly reduced after the reaction, illustrating the excellent stability of the catalyst.The elemental composition and chemical state of the NiO-Bi2WO6 catalyst were further analyzed by XPS. As shown in Fig. 4\n (a), the catalyst was composed of Bi, W, O, and Ni, consistent with the EDS test results. Fig. 4 (b) shows two peaks of Bi 4f at 159.0\u00a0eV and 164.3\u00a0eV after peak fitting, corresponding to Bi3+ in Bi2WO6 [37]. The peaks of W 4f at 35.3\u00a0eV and 37.4\u00a0eV in Fig. 4 (c) were attributed to W6+ in Bi2WO6. The O 1\u00a0s peak fitting results are shown in Fig. 4 (d); three peaks appeared after splitting. Two peaks at 530.2\u00a0eV and 531.5\u00a0eV were attributed to the O atom of BiO in Bi2WO6, while the peak at 529.3\u00a0eV was attributed to the O atom of NiO in NiO34. Fig. 4 (e) shows the characteristic peaks of Ni 2p. The peaks located at 855.6\u00a0eV and 860.6\u00a0eV were assigned to Ni 2p3/2, while those at 872.0\u00a0eV and 879.0\u00a0eV were assigned to Ni 2p1/2 [38,39]. These results indicate that the NiO-Bi2WO6 catalyst was successfully synthesized through hydrothermal and calcination methods.The light absorption performance of the photocatalysts was studied by UV\u2013Vis-DRS under the same scale, as shown in Fig. 5\n. As observed, pure Bi2WO6 has strong visible absorption in the ultraviolet region from 200 to 400\u00a0nm, with the absorption edge at 464.3\u00a0nm. However, the NiO-Bi2WO6 composite materials absorb strongly in the visible light region from 400 to 750\u00a0nm, with an absorption edge of 704.1\u00a0nm, and a redshift occurred. The visible region accounts for approximately 50% of the whole solar radiation energy, while the ultraviolet region only accounts for approximately 7%. Therefore, NiO-Bi2WO6 composite materials broaden the response to light, which is beneficial for improving the photocatalytic oxidation activity. According to the formula of band gap and band edge absorption (Eg\u00a0=\u00a01239.8/\u03bbg, where \u03bbg is the absorption edge), the band gaps of Bi2WO6 and NiO-Bi2WO6 are 2.67\u00a0eV and 1.76\u00a0eV, respectively. In summary, the combination of NiO and Bi2WO6 reduces the band gap, which is conducive to generating photogenerated electrons and holes, thereby further improving the photocatalytic oxidation performance of NiO-Bi2WO6.To evaluate the photocatalytic oxidative desulfurization performance of NiO-Bi2WO6, desulfurization of BT model oil was carried out. As shown in Fig. 6\n (a), initially, the model oil exhibited almost no degradation in the absence of light and catalyst. Then, different photocatalysts (NiO, pure Bi2WO6, NiO-Bi2WO6) were separately added to the model oil. After reacting for 30\u00a0min in the dark, reaching adsorption-desorption equilibrium, desulfurization was carried out under light conditions. Fig. 6 (a) shows that under dark conditions, although different catalysts exhibited a certain desulfurization rate, the rate was relatively low. When the light source was turned on, the desulfurization rate was significantly increased. The results show that the desulfurization rate of the NiO-Bi2WO6 composite was higher than that of pure Bi2WO6 and NiO because light contributes to the generation of photogenerated electron-hole pairs and the combination of NiO and Bi2WO6 reduces the band gap of the catalyst and photogenerated electron-hole pair recombination, thereby improving the photocatalytic oxidation activity.\nFig. 6 (b) further depicts the effect of NiO loading on the desulfurization performance. As the NiO loading increased from 10% to 30%, the desulfurization rate increased from 74.53% to 93.32%. However, when the NiO loading reached 100%, the desulfurization rate unexpectedly decreased to 58.89%. The reason was that with increasing NiO loading, the number of redox-active photogenerated electron-hole pairs increased, increasing the desulfurization rate. As the NiO loading continued to increase, the active sites on the catalyst surface became covered, which unfortunately decreased the desulfurization rate of NiO-Bi2WO6. The experiments show that the ratio of NiO and Bi2WO6 affects the desulfurization effect. When the NiO loading was 30%, the desulfurization effect reached 93.32%.The effect of different catalyst dosages on the desulfurization performance was studied to further improve the desulfurization rate. As shown in Fig. 7\n, when there was no catalyst in the system, the desulfurization rate was almost zero. As the amount of catalyst increased from 0.3\u00a0g/L to 1.2\u00a0g/L, the desulfurization rate increased from 81.36% to 95.37%, reaching the highest value at this time. As the catalyst dosage increased to 1.8\u00a0g/L, the desulfurization rate decreased to 86.84%. When the amount of catalyst added to the reaction system was inadequate, the amount of photogenerated electron holes was insufficient, resulting in poor catalytic effects and a low desulfurization rate. As the amount of NiO-Bi2WO6 increased, the desulfurization rate gradually increased. However, an excessive amount of photocatalyst can decrease the light transmittance of the model oil solution and cause light scattering, which will reduce the light absorption and utilization rate as well as the generation of excited states in the reaction system, leading to a lowered desulfurization rate. The experimental results show that the appropriate amount of catalyst has a positive effect on the photocatalytic oxidation desulfurization performance of NiO-Bi2WO6.To explore whether NiO-Bi2WO6 has wide application prospects for the removal of sulfur-containing compounds from gasoline, the desulfurization performance of different substrates, namely, Th, BT and DBT, which are present at high concentrations and difficult to remove, was studied. Fig. 8\n shows that NiO-Bi2WO6 exhibited the highest removal rate for DBT, followed by BT and then Th; the desulfurization rates were 98.2%, 95.37% and 85.8%, respectively. The efficiency was affected by the electron cloud density of the S atom. The lower the density of the s electron cloud, the more difficult it is to oxidize. The s electron cloud densities of DBT, BT, and Th were 5.758, 5.696 and 5.639, respectively, so the removal efficiency showed a downward trend. The above results show that the NiO-Bi2WO6 photocatalyst has a good degradation effect on various sulfur-containing model compounds and has a wide range of practical applications.To study the mechanism of the NiO-Bi2WO6 photocatalyst for photocatalytic oxidative desulfurization under light conditions, analyze the electronic conduction and transfer capabilities of the material, and determine the separation efficiency of electrons and holes, the transient photocurrent responses and electrochemical impedance spectra were measured. As shown in Fig. 9\n (a), compared with that of Bi2WO6, the transient photocurrent density of the NiO-Bi2WO6 photocatalyst was significantly increased, indicating that as more electrons were excited under light irradiation, the charge transfer performance increased [40]. Thus, the separation efficiency of photogenerated electrons and holes was improved. Fig. 9 (b) shows the electrochemical impedance values of the Bi2WO6 and NiO-Bi2WO6 samples. The resistance value of the NiO-Bi2WO6 samples was much lower than that of the Bi2WO6 samples. A decreased resistance value will accelerate electron transfer, which is consistent with the transient photocurrent response test results. Undoubtedly, the combination of NiO and Bi2WO6 effectively improves the charge transfer capability, reduces the recombination rate of photogenerated electrons and holes, and is beneficial for improving the photocatalytic oxidation performance.To thoroughly study the active species of the NiO-Bi2WO6 photocatalyst during the degradation of BT, an active radical trapping experiment was carried out. In the experiment, isopropanol (IPA), sodium oxalate (Na2C2O4), and p-benzoquinone (1,4-BQ) were used as radical scavengers for h+, \u00b7OH, and \u00b7O2\u2212, respectively. The experimental results are shown in Fig. 10\n. IPA, Na2C2O4 and 1,4-BQ have little effect on the degradation of BT. When NiO-Bi2WO6 and different active radical capture agents were added to the BT model oil, the degradation rate changed. The system containing 1,4-BQ changed the most, and the desulfurization rate decreased from 95.31% to 25.83%, followed by Na2C2O4 and IPA, which decreased the rate to 65.37% and 79.44%, respectively. The experimental results show that when the \u00b7O2\u2212 in the system was captured by 1,4-BQ, the desulfurization rate was significantly reduced; when h+ was captured, the desulfurization rate decreased slightly; and when \u00b7OH was captured, the desulfurization rate decreased slightly. These results indicate that the active species from NiO-Bi2WO6 that played a major role in the photocatalytic oxidation degradation of BT was \u00b7O2\u2212, followed by h+, and \u00b7OH was not the main active species. The existence of \u00b7O2\u2212 was because light irradiation excited NiO-Bi2WO6 to generate photogenerated electrons and holes, and molecular oxygen was reduced by the electrons to generate superoxide radicals. Therefore, the higher the separation efficiency of photogenerated electrons and holes was, the stronger the photocatalytic activity. This result was consistent with the electrochemical impedance spectra.The substance of BT model oil before and after reaction was identified by GC analysis. As shown in Fig. 11\n, before the photocatalytic oxidation of the model oil, the only substance present was BT. After the reaction, a large amount of BTO2 existed in the system, while only minimal BT remained. When the reacted model oil was extracted with acetonitrile, it contained only a small amount of BT. The results showed that the superoxide radicals and holes produced by NiO-Bi2WO6 oxidized BT to the highly polar species BTO2 under light irradiation, which was then extracted by acetonitrile to achieve deep desulfurization.Based on the analysis of the above experiments and characterization results, a schematic of NiO-Bi2WO6 photocatalytic oxidation is proposed, as shown in Fig. 12\n. The valence band (EVB) and conduction band (ECB) positions of NiO and Bi2WO6 were calculated from empirical formulas (E\n\nVB\n\u00a0=\u00a0X\u00a0\u2212\u00a0E\n\ne\n\u00a0+\u00a00.5E\n\ng\n and \n\u03b7\n=\n\n\n\nC\n0\n\n\u2212\n\nC\nt\n\n\n\nC\n0\n\n\n\u00d7\n100\n%\n), and the calculation results are also shown in Fig. 12. Under light excitation, the electrons produced by the NiO/Bi2WO6 photocatalyst were transferred from the valence band to the conduction band, leaving holes in the conduction band and thereby generating photogenerated electrons and holes. Fig. 12 (a) shows that due to the Schottky barrier at the interface, the conduction band position of NiO (\u22120.54\u00a0eV) is more negative than that of Bi2WO6 (+0.55\u00a0eV), and the valence band position of Bi2WO6 (+3.22\u00a0eV) is more positive than that of NiO (+3.06\u00a0eV). Therefore, the photogenerated electrons in the conduction band of NiO easily transition to the conduction band of Bi2WO6, and the holes in the Bi2WO6 valence band easily flow into the NiO valence band, thereby realizing the separation of photogenerated electrons and holes. However, since the standard oxidation-reduction potential of O2/\u00b7O2\u2212 (\u22120.33\u00a0eV) is more negative than that of Bi2WO6 (+0.55\u00a0eV), the conversion cannot be completed [41]. Active radical experiments show that \u00b7O2\u2212 is essential to the degradation process; thus, the schematic shown in Fig. 12 (a) is not the mechanism of NiO-Bi2WO6. The energy band structure presented in Fig. 12 (b) can well explain the mechanism of NiO-Bi2WO6. The electrons in the conduction band of Bi2WO6 first recombine with the holes in the valence band of NiO so that the photogenerated electron-hole pairs on Bi2WO6 and NiO are separated. Then, the electrons on the conduction band of NiO are captured by O2, producing the oxidants \u00b7O2\u2212, \u00b7O2\u2212 and h+, which can oxidize BT to BTO2. Since the position of the valence band (+3.22\u00a0eV) of Bi2WO6 is more positive than that of \u00b7OH/OH\u2212 (+2.40\u00a0eV) [42], a portion of h+ oxidizes H2O to form \u00b7OH and participates in the reaction. According to the above analysis and the active radical capture experiment, three active substances, \u00b7O2\u2212, h+ and \u00b7OH, participate in the photocatalytic oxidation reaction, and \u00b7O2\u2212 plays a major role in the reaction.Catalyst stability and reusability is an important indicator to measure whether a catalyst has application value. Accordingly, the activity of the NiO-Bi2WO6 composite material was studied through a catalyst recycling experiment. Fig. 13\n shows the relationship between the desulfurization rate and time when NiO-Bi2WO6 was used for six cycles. The figure shows that after six cycles with NiO-Bi2WO6, the desulfurization rate was only reduced from 95.37% to 88.96%, and the catalyst still exhibited high catalytic activity, indicating that it has good recycling performance and stability. This result is consistent with the XRD measurement results of NiO-Bi2WO6 before and after the reaction, as shown in panel (b). Based on the cycle experiment and XRD measurement results, the catalyst has excellent recycling performance and great stability.In summary, based on the matching band structure between NiO and Bi2WO6, a heterojunction was formed, and this structure facilitated the generation and separation of photogenerated electron-hole pairs. Therefore, a new flower-like NiO-Bi2WO6 composite photocatalyst was successfully synthesized by a hydrothermal method and a high-temperature calcination method. This result was confirmed though various characterization methods, such as SEM, TEM, EDS, XRD and XPS. Under visible light irradiation, the desulfurization rate of the NiO-Bi2WO6 composite material for BT was much higher than that of pure NiO and Bi2WO6. In particular, when the NiO loading was 30% and the catalyst dosage was 1.2\u00a0g/L, the desulfurization rate was highest (95.31%). According to the ultraviolet-visible diffuse reflectance absorption spectrum and the electrochemical impedance diagram, this optimal result is because the combination of NiO and Bi2WO6 reduces the band gap, causes a redshift, and exhibits an increased electron-hole separation efficiency, thereby improving the photocatalytic performance. NiO-Bi2WO6 showed a good removal effect on many different thiophene-type sulfides. Moreover, the active radical capture experiments and GC results showed that three active species \u00b7O2\u2212, h+ and \u00b7OH participate in the photocatalytic oxidation reaction, oxidizing BT to BTO2 for removal, and that \u00b7O2\u2212 plays a major role in the reaction. Furthermore, the high stability and excellent recycling performance of the NiO-Bi2WO6 photocatalyst indicate that it has good application value.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was financially supported by the National Natural Science Foundation of China (Grant No. 21276156).", "descript": "\n                  In this paper, a hydrangea-like Bi2WO6 material was first synthesized by a hydrothermal method. Then, as a carrier, a NiO-Bi2WO6 composite photocatalyst was successfully synthesized by the hydrothermal method and high-temperature calcination and applied to studying the removal of benzothiophene (BT) from fluid catalytic cracking (FCC) gasoline. The morphology, crystal structure and elemental composition of the catalyst were characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), energy-dispersive X-ray spectroscopy (EDS), X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS); ultraviolet-visible diffuse reflectance spectroscopy (UV\u2013Vis-DRS) and electronic impedance spectroscopy (EIS) were used to characterize the light absorption and charge transfer ability of the catalyst. The characterization confirmed that the NiO-Bi2WO6 composite photocatalyst was successfully prepared. Moreover, the effects of catalyst dosage, NiO loading and different substrates on the desulfurization rate were investigated by photocatalytic oxidation desulfurization experiments. The experimental results showed that under the conditions of a catalyst dosage of 1.2\u00a0g/L and NiO loading of 30%, the highest desulfurization rate of BT was 95.37%. The active species in the reaction process were studied through active radical capture experiments and qualitative analysis of the substances before and after the reaction was carried out via gas chromatography (GC), and the reaction mechanism of the catalyst was explored in depth. The results indicated that the catalyst mainly oxidized BT to benzothiophene sulfone (BTO2) under the oxidation of superoxide radicals to achieve deep desulfurization. Cycling experiments demonstrated that the catalyst still had high stability and catalytic activity after six cycles.\n               "}
{"full_text": "The widespread presence of organic pollutants in wastewater poses a serious threat to ecosystems and public health, many micropollutants are extremely toxic even at low concentrations and require effective removal [1\u20133]. However, many difficult-to-remove organic pollutants, especially antibiotic aromatic compounds, are difficult to remove by conventional water treatment techniques [4,5]. The Electro-Fenton (EF) process, which oxidizes harmful pollutants by strongly oxidizing OH, is particularly suitable for the degradation of such stubborn compounds [6,7]. The generation of active components in traditional EF generally includes two processes: (1) anodic oxidation to generate oxygen and hydroxyl radicals (\u00b7OH yield of the anticathode is much lower than that of the negative electrode) [8]; (2) dissolved oxygen (produced by pumping in air or oxygen) undergoes an oxygen reduction process at the cathode to generate hydrogen peroxide, followed by homogeneous or heterogeneous reactions to generate hydroxyl radicals [9]. Traditional EF anodes are dominated by BDD, Pb, Ti/SnO2, and Au. Although they have higher oxygen evolution overpotential, the expensive cost and high energy consumption limit its application in practical water treatment [8,10\u201312]. If these energy-intensive anodes can be replaced by excellent oxygen evolution catalysts, it may help to reduce energy consumption while achieving self-oxygenation of the system, thus avoiding the need for additional air or oxygen. Recent developments of EF cathode catalysts have been mostly focused on the composites of transition metals and carbon materials [2,3,13\u201316], in which carbon defects and heteroatom dopants are exploited as the active sites for preparation of H2O2 by electrosynthesis [17,18]. In addition, some transition metals (such as Co, Fe, Ni, and Cu) have become promising active components for EF processes due to their cheap cost and good property in activating H2O2 [19\u201323]. However, highly loaded atomically dispersed carbon materials generally require complex preparation process [24], moreover, obvious challenges still remain in the rational regulation of dopants or defects with atomic-level accuracy [25]. Thus, it is of significance to develop bifunctional transition metal catalysts with efficient oxygen evolution reaction (OER) and oxygen reduction reaction (ORR) activity for improving energy efficiency and cathode electrocatalytic activity in conventional water treatment processes.Generally, high efficiency of electron transfer near the catalytic site and suitable free energy of adsorption for OER intermediates at the catalytic site are crucial for the fabrication of highly active OER and ORR catalysts [26,27]. For a long time, electrocatalysts with high oxygen evolution activity are usually noble metal-based materials. However, due to their rarity, more researches currently focus on transition metal-based catalysts, such as Co3O4/CeO2, NiFe LDH, CoSe2, RuO2, etc [28\u201331]. Among them, CoSe2, as a typical transition metal selenide, has attracted a great deal of interest in energy consumption and conversion, including water separation, Li-ion cells and zinc-air cells, due to its availability in abundance on Earth, acceptable low cost, intrinsic excellent electrical properties and remarkable stability\u00a0\u200b[32\u201334]. Dong et\u00a0al. [35] homogeneously immobilised CoSe2 nanoparticles on carbon fibre paper by pyrolysis and selenisation of ZIF-67, resulting in a catalyst (CoSe2/CF) with good long-term stability and an over-potential of 297\u00a0mV in the OER process (10\u00a0mA/cm2). In addition, orthorhombic marcasite-type CoSe2 (o-CoSe2) has been demonstrated to exhibit excellent H2O2 accumulated concentration and selectivity (more than 80%) in 2e\u2212 ORR due to the ability that the adjacent Co active sites to inhibit the breaking of O\u2013O bonds, which comparable to that of noble metals in acidic electrolytes [19]. However, when o-CoSe2 is used as the cathode of the EF process, it is usually necessary to add Fe2+ to the solution to activate H2O2 due to tepid fenton activity of CoSe2, which may lead to the formation of iron sludge [22,23]. Iron (Fe), as a typical catalyst, exhibits outstanding catalytic activity in the degradation process of organic wastewater [7,36], and our previous work has also demonstrated the co-operative interaction of Co and Fe in a Fenton-like process [37], Therefore, combining CoSe2 with Fe-based metals may be an efficacious method to improve the Fenton property of CoSe2. The electronic structure projection of doped Mo can facilitate electron transport and synergistic interactions between various elements, affect the sorption/desorption energy of active substance in electro-catalytic reactions and adjust the catalytic capacity [38\u201340]. Meanwhile, Mo doping also can leads to electronic rearrangements and lattice flaws in CoSe2, which contributes to the electron transfer capacity and catalytic activity of CoSe2 [41]. Therefore, the adjustment of the electronic structure at the CoSe2 surface by the action of Mo and Fe may help to tune the adsorption energy of intermediate \u2217OOH and enhance deprotonation of adsorbed water molecules in the OER and ORR process, thereby enhancing its EF performance.Here, we developed Mo-doped FeCo\u2013Se aerogels to couple OER and EF reactions. Because of the superiority of three-dimensional frame structure and large surface area, the aerogel structure could improve the electrochemical behavior of OER and EF through shortened distance diffusion pathways and accelerated mass transfer. The Mo-doped FeCo\u2013Se anode effectively replaces the common anodes such as Pt, BDD, PbO2, SnO2, RuO2, and TiO2. The doping effect of Mo can significantly increase the carrier density of the anode and lower the reaction free energy of the OER process. Under the current intensity of 10\u00a0mA/cm2, the overpotential is only 235\u00a0mV, which is significantly higher than that of RuO2. The excellent OER performance can provide sufficient oxygen for the oxygen reduction process of the cathode, unlike the general electro-Fenton process, which requires continuous introduction of air or oxygen. Furthermore, on the electro-Fenton side, Mo doping CoSe2 exhibits excellent H2O2 selectivity (ca. 85% at the region of 0.1\u20130.6\u00a0V). Due to synergistic effect of the bimetals, Fe metal effectively activates H2O2 generated on the surface of CoSe2, compared to the CoSe2/Fe2+ system, the production rate of active components is higher, which helps to further obtain the ideal degradation effect, The best Mo0\n.\n3Fe1Co3\u2013Se catalyst can remove 97.7% of sulfamethazine (SMT) within 60\u00a0min (SMT content: 10\u00a0mg/L, current intensity: 10\u00a0mA/cm2). This work provided a novel perspective for the development of transition metal-based electro-catalysts for EF process.Sodium borohydride (NaBH4, 99.5%), selenium powder (Se, AR), ferric chloride anhydrous (FeCl3, AR), sodium molybdate (Na2MoO4, AR), cobalt (II) chloride hexahydrate (CoCl2\u20276H2O, 99%), were bought from Inno-Chem Science & Technology Co., Ltd (Beijing).CoFe metal aerogels (MAs) were prepared by a typical NaBH4-induced gelation process. NaBH4 (0.1\u00a0M) was added in the mixture CoCl2\u20276H2O (0.05\u00a0M) and FeCl3 (0.05\u00a0M), then Na2MoO4 (0.05\u00a0M) was added to aqueous solution. The obtained solution was left for about 8\u00a0h to form a hydrogel. Samples with different metal ratios were achieved by varying the volume ratio of the FeCl3 and CoCl2\u20276H2O solutions. In a typical preparation of FeCo\u2013Se, the selenium powder and CoFe aerogels were positioned upside and downside of the tubular furnace, respectively, and the CoFe\u2013Se aerogel was obtained by selenization at 400\u00a0\u00b0C for 3\u00a0h.Scanning electron microscopes (SEM, Hitachi, SU8010), transmission electron microscope (TEM, F20), and electrochemical tests were performed using a CHI 760D electrochemical workstation (CH Instruments, Shanghai) with incorporates a three-electrode system. Oxygen reducing property-related electrochemical tests collected by rotating disc electrodes (PINE, CPR1 and Wavenow)A catalyst suspension was obtained by mixing 20\u00a0mg of aerogels sample with 200\u00a0\u03bcL of ethanol, 200\u00a0\u03bcL of deionised water, and 50\u00a0\u03bcL of poly tetra fluoroethylene (PTFE) and subjected to ultrasonic treatment for 30\u00a0min. Then, The catalyst suspension was dipcoated in carbon fabric and dried overnight at 60\u00a0\u00b0C to make a working cathode and anode. 10\u00a0mg/L SMT solutions (0.1\u00a0M Na2SO4) was used as a model for the target pollutants. Use the organic filter to extract 1\u00a0mL of the sample solution at set intervals. The micro-pollutants concentration during the degradation was measured by the high performance liquid chromatography (HPLC) (Shimadzu, LC-10ADVP) fitted with a C18 column (particles size: 1.9\u00a0\u03bcm, 2.1\u00a0\u00d7\u00a0150\u00a0mm). The mobile phase solution for the SMT detection was a mixture of 0.1% acetonitrile and formic acid solution in a ratio of 1:4.DFT calculation was conducted by DMol3 as implemented in materials studio (MS). All energy variations were processed by the generalised gradient approximation of the Perdew-Burke-Ernzerhof (GGA-PBE) function. The K point is set to 4\u00a0\u00d7\u00a04\u00a0\u00d7\u00a01 and the size of the vacuum zone has a size of 15\u00a0\u00c5. The Mo doping Co\u2013Fe model was formed by cubic cell (Im-3m 229) consisting of elementary particles with the molar ratio of Co: Fe: Mo\u00a0=\u00a03:1:0.3, where the crystal surface (110) was cleaved as the research object. Meanwhile, the Mo doping CoSe2\u2013FeSe2 model was constructed by a modified CoSe2 cell (Pa-3 (205)), where the Co element particles were replaced by elementary particles with the molar ratio of Co: Fe: Mo\u00a0=\u00a03:1:0.3 and the crystal surface (110) was cleaved as the research object. The surfaces were modified by U\u00a0\u00d7\u00a02 and V\u00a0\u00d7\u00a02 super cell. The adsorption energy of the intermediate state in the OER process is calculated according to Eq. (1):\n\n(1)\n\n\n\u0394\nG\n=\n\u0394\nE\n+\n\u0394\nZ\nP\nE\n\u2212\nT\n\u0394\nS\n\n\n\nWhere \u0394S, \u0394ZPE, and \u0394E were the entropy change, zero-point energy change, and binding energy for the adsorption process, respectively.Synthesis of Mo-doped FeCo\u2013Se MAs including the fabrication of CoFe MAs and the subsequent selenization process (Fig.\u00a01\na). The primary self-assemble aggregates first rapidly bind to form nanowire structures when sodium borohydride is used as a reducing and gelling agent, followed by cross-linking of the nanowires to form hydrogels with a three-dimensional porous network feature in the presence of H2 template. After drying, the CoFe MAs were selenized at 400\u00a0\u00b0C for 3\u00a0h by upstream and downstream methods.It can be seen that Mo0\n.\n3Fe1Co3\u2013Se MAs exhibits an obvious nanowire cross-linked skeleton structure in Fig.\u00a01b\u2013d, and further increase of Mo content may lead to clogging of aerogel pore structure (Fig.\u00a0S1), which is detrimental to the mass transfer of the catalytic process. Compared with the Mo-doped CoFe MAs without selenization (Fig.\u00a0S2), the cross-linked framework structure did not collapse and deform significantly after the selenization treatment at 400\u00a0\u00b0C. This inter-connected framework structure helps accelerate mass transfer and shorten the distance diffusion pathway, which is thought to help enhance electrocatalytic performance. As can be seen in Fig.\u00a01d, the aerogel nanowires have small protrusions on their surface, which may help to further increase the number of active sites available to the catalyst and thus improve its catalytic performance. We performed the corresponding TEM and EDS characterizations to further reveal its morphological features. The nanowires are formed by self-assembly of small primary aggregate units, the diameter of the aerogel nanowires is about 50\u201360\u00a0nm with the core-shell structure features (Fig.\u00a01e and f). We speculate that the internal composition of the nanowires may be Fe (0) according to the different activities of Fe and Co in the reaction process with sodium borohydride, and the lattice diffraction fringes of the surface layer well matched the (111) crystal plane of CoSe2 after the selenization treatment. Fe, Co, Mo, and Se are uniformly dispersed in the aerogel structure. Notably, Se elements are mostly dispersed on surface layers of the aerogel nanowires in the elements mapping (Fig.\u00a01g).The Fe1Co3 MAs, Mo0\n.\n3Fe1Co3 MAs and Mo0\n.\n3Fe1Co3\u2013Se MAs samples are described by XRD (Fig.\u00a02\na), the only peak at 2\u03b8\u00a0=\u00a044.9\u00b0 are indexed to (110) planes of the Co7Fe3 (PDF#50\u20130795). In the XRD of Mo0\n.\n3Fe1Co3 MAs, only the features of Co7Fe3 (2\u03b8\u00a0=\u00a045.1\u00b0) can be observed, suggesting the existing form of Mo is doped or amorphous. After the calcining of Mo0\n.\n3Fe1Co3 MAs in a Se atmosphere, the resulting of Mo0\n.\n3Fe1Co3\u2013Se MAs exhibits a marked difference. The different peaks at 2\u03b8\u00a0=\u00a030.7\u00b0, 34.5\u00b0, 35.9\u00b0and 47.7\u00b0 correspond to the (101), (111), (120), and (211) of CoSe2 (PDF#53\u20130449) [19,41], The additional peak at 2\u03b8\u00a0=\u00a045.2\u00b0 is indexed to (110) planes of the Co7Fe3. Note that a small positive shift (0.1\u00b0) is observed on the characteristic peaks of Mo0\n.\n3Fe1Co3 MAs and Mo0\n.\n3Fe1Co3\u2013Se MAs, further confirming that the incorporation of Mo does not produce new crystalline phases in the high temperature heating process (oxidation) [39,40,42]. The above results fully demonstrate the successful preparation of Mo0\n.\n3Fe1Co3\u2013Se MAs. The specific surface areas of aerogel samples were presented in Fig.\u00a02b and Table\u00a0S1. It can be seen that the specific surface area of Fe1Co3 MAs, Mo0\n.\n3Fe1Co3 MAs and Mo0\n.\n3Fe1Co3\u2013Se MAs are 16.25, 16.70, and 26.13\u00a0m2/g, respectively. The significant increase in surface area of the Mo0\n.\n3Fe1Co3\u2013Se MAs is due to the increased surface roughness of the nanowires as a result of selenisation. Corresponding results can also be seen from the TEM images, with raised CoSe2 crystals on the surface of the aerogel nanowire structure. Besides, Mo0\n.\n3Fe1Co3\u2013Se MAs also exhibits the larger pore volume and pore width (Table\u00a0S1). This can provide a sufficient mass transition pathways and active sites number to improve kinetics of reactions.Next, the XPS spectra of Fe1Co3 MAs, Mo0\n.\n3Fe1Co3 MAs, and Mo0\n.\n3Fe1Co3\u2013Se MAs are studied to investigate their elemental interrelationships (Fig.\u00a0S3). The characteristic peaks at \u223c712 and 780\u00a0eV in the XPS spectra of the different catalysts correspond to Fe 2p and Co 2p [37]. Note that for the Mo0\n.\n3Fe1Co3 MAs and Mo0\n.\n3Fe1Co3\u2013Se MAs, an additional peak attributable to Mo 3d occurs at \u223c230\u00a0eV, implying successful adulteration of Mo elements [39]. In addition, it was noticed that the selenized samples showed more pronounced Se peaks at 55\u00a0eV, indicating the successful selenization of Mo0\n.\n3Fe1Co3 MAs [41]. The hi-resolution spectra of Fe 2p, Co 2p, Mo 3d, and Se 3d are displayed in Fig.\u00a03\n. The spectra of Fe 2p can be decomposed into six peaks. Two peaks at 719.9 and 706.6\u00a0eV are ascribed to Fe (0), and the other four peaks at 724.7, 722.8, 715.9, and 711.8\u00a0eV are ascribed to Fe2+ and Fe3+, respectively [43\u201345], which is caused by partial oxidation of the aerogel when being exposed to solution and air. Six characteristic peaks were obtained by fitting the Co 2p spectra. The two peaks at 778.5 and 797.2\u00a0eV are assigned to Co2+ [40], and another one at 793.6\u00a0eV is assigned to Co3+ [39,40]. The two peaks at about 230.6\u00a0eV and 228.9\u00a0eV on the Mo 3d spectra of Mo0\n.\n3Fe1Co3 MAs and Mo0\n.\n3Fe1Co3\u2013Se MAs are assigned to Mo4+ 3d5/2 and Mo 3d5/2, confirming the existence of Mo4+ in the two catalysts, which is attributed to the reduction of NaBH4 [34]. The two peaks at 55.0\u00a0eV and 54.3\u00a0eV are attributed to Se 3d3/2 and 3d5/2, respectively (Fig.\u00a03c), while another one at 59.4\u00a0eV is assigned to SeO\nx\n [46], which may be due to surface oxide layer and adsorbed oxygen in air. It is noted that two new peaks at 235.5 and 232.4\u00a0eV appears after selenization, which indicates the existence of the Mo6+. The change from Mo4+ to Mo6+ indicates that some of the electrons are captured by the nearby Fe and Co atoms, and this electron transfer phenomenon may be conducive for electrocatalytic behavior [42].The EF catalytic property of the fabricated materials was assessed in a model two-electrode device configured using carbon cloth coated with catalyst as the cathode in the electrolyte. (SMT\u00a0=\u00a010\u00a0mg/L, 0.1\u00a0M Na2SO4). Firstly, different metal ratios were examined, where the material with the molar ratio of Mo: Fe: Co is 0.3:1:3 and annealed at 400\u00a0\u00b0C showed the best catalytic ability (Fig.\u00a04\na). Adsorption removal of SMT is negligible in this process. The removal rate of SMT is close to 100% within 90\u00a0min, and the proposed first-order reaction rate of up to 0.084 min\u22121 (Fig.\u00a0S4a), which is better than the performance of some recently reported EF cathodes materials (Table\u00a0S2). Besides, it is noted that the EF performance of Mo0\n.\n5Fe1Co3\u2013Se MAs becomes worse with higher Mo content. This result may be due to that the increase of sodium molybdate leads to formation of part of molybdate, which is wrapped on the aerogel framework structure (Fig.\u00a0S1b), thereby decreasing the number of available active sites and leading to poor performance. To determine the dominant role of EF cathode in the degradation process, the corresponding comparative experiments are shown in Fig.\u00a0S1c. First, Mo0\n.\n3Fe1Co3\u2013Se MAs/Carbon fabric (CF) is made as the cathode and anode, and the removal of SMT by adsorption is negligible under the condition of no current. CF without catalyst as the cathode and anode exhibits about 16% SMT attenuation performance at the same current density (10\u00a0mA/cm2). CF without catalyst as the cathode, Mo0\n.\n3Fe1Co3\u2013Se MAs/CF as the anode, approximately 34.8% of SMT was removed within 90\u00a0min, and when Mo0\n.\n3Fe1Co3\u2013Se MAs/CF served as the cathode and anode, SMT attenuation efficiency close to 100%. Although the small amount of \u2027OH produced by anode will also contribute to the removal of SMT (in addition to the OER process, the anode can contribute to the formation of adsorbed \u00b7OH under the action of oxidation potential), the main contribution to the SMT removal process comes from the EF cathode. In addition, the same volume of 25\u00a0mg/L methyl blue (MB) was placed in the cathode chamber and the anode chamber, respectively. Take out 1\u00a0mL of the solution at regular intervals for ultravioletand visible (UV) absorbance testing. As shown in Fig.\u00a0S1d, with the increase of time, the absorption peak of cathode chamber solution gradually decreases and approaches 0, and its fading process is shown in Fig.\u00a0S1f. The change of UV absorbance of MB solution in the anode chamber is shown in Fig.\u00a0S1e, and the absorption peak of MB only decreased slightly, and the decoloration of the solution is almost imperceptible (Fig.\u00a0S1f). In addition, the fading process of MB was similar to the above results when MB indicator was added to the anode chamber (Fig.\u00a0S1g) or the cathode chamber (Fig.\u00a0S1h) alone. As the reaction proceeds, the fading of the anode chamber MB solution was almost negligible, while the decolorization of MB in the cathode chamber was almost complete. Given the above comparative experiments, it is not difficult to find that the cathode catalytic process plays a leading role in the degradation process.In general, pH has an impact on the reaction process of EF that involve iron-based catalysts. The percentage degradation decreases with increasing initial pH at any selected electrolysis time (Fig.\u00a04c). The fastest degradation was achieved at a pH of 3.0, achieving complete decay within 75\u00a0min, which can be due to the fact that the dissolution of iron can promote a homogeneous Fenton reaction to produce \u00b7OH. None of the other pH conditions resulted in complete elimination of SMT within 90\u00a0min, reaching removal rates of 99.8%, 69%, and 61.1% at pH 4.0, 7.0, and 9.0, respectively. The evaluation of degradation performance at various cathode currents is shown in Fig.\u00a04d, only 64.5% SMT was removed at 5\u00a0mA/cm2, while the current intensity reaches 10\u00a0mA/cm2, improved removal performance of SMT with increasing current is related to that larger current helps to promote the production of H2O2. Besides, note that the applied current intensity does not take a key factor in heterogeneous EF with further increase in current, since similar removal rate could be obtained. At 90\u00a0min, 99.8% and 100% SMT decay was determined at 10 and 15\u00a0mA/cm2. The removal performance of different SMT concentrations in this EF system is shown in Fig.\u00a0S3b, when the concentration of SMT is 5 and 20\u00a0mg/L, the degradation performance can reach more than 97% and 55% within 90\u00a0min, which indicates that Mo0\n.\n3Fe1Co3\u2013Se MAs still has good catalytic removal performance for SMT below 20\u00a0mg/L concentration. This means that Mo0\n.\n3Fe1Co3\u2013Se MAs is suitable for industrial high-concentration, complex antibiotic removal.To evaluate the catalytic ability of Mo-doped FeCo MAs for EF anode, The OER properties of Mo0\n.\n3Fe1Co3\u2013Se MAs materials were determined by performing LSVs tests in 1\u00a0M KOH with a scan rate of 10\u00a0mV/s. Mo0.3Fe\nx\nCo\ny\n-Se MAs catalysts with different Fe and Co contents were measured to verify the influence of the metal ratio on the OER behavior. Based on the electrochemical data obtained for the different materials (Fig.\u00a05\na and b), the doping of Mo contributes to the OER properties of FeCo aerogels, and the OER properties is further improved after selenization. Moreover, the Mo-doped selenide samples displayed the best catalytic properties, providing the lowest overpotential of 235\u00a0mV at 10\u00a0mA/cm2 current intensity and the lowest Tafel slope of 73\u00a0mV/dec in all samples tested (Fig. 5c), which outperforms many similar materials (Table\u00a0S3). The prepared sample shows higher Tafel slope compared to RuO2, which indicates that the OER performance of RuO2 is slightly better than Mo0\n.\n3Fe1Co3\u2013Se MAs at lower current densities, however, the open circuit potential of RuO2 in the OER process is much higher than that of Mo0\n.\n3Fe1Co3\u2013Se MAs, therefore, Mo0\n.\n3Fe1Co3\u2013Se MAs is more advantageous from the point of actual energy consumption. In addition to this, the Tafel slope of the samples at higher current densities was shown in Fig.\u00a0S5, in the current interval of 1\u20131.65 (log j (mA/cm2)), Mo0\n.\n3Fe1Co3\u2013Se MAs (84\u00a0mV/dec) exhibits a lower Tafel slope than RuO2, which indicates that as the current increases, the Tafel slope as well as the open circuit potential of Mo0\n.\n3Fe1Co3\u2013Se MAs sample are better than RuO2. Considering the possible industrial catalytic processes, the persistence of Mo0\n.\n3Fe1Co3\u2013Se MAs for OER is evaluated by measuring the LSVs profiles of the materials after and before 1000 continuous cycles and the current-time curve in 1\u00a0M KOH with a scan rate of 100\u00a0mV/s (10\u00a0mA/cm2). Mo0.3Fe1Co3\u2013Se MAs displayed almost ignorable decay after 1000 continuous cycles (Fig.\u00a05d), and no obvious current decay was seen and the current maintained 95.3% of the initial current intensity after 12\u00a0h of chrono-voltage measurements (Fig.\u00a07c). Nyquist plot of aerogels samples were shown in Fig.\u00a05e, and the fitted resistance value (Rct) of Mo0\n.\n3Fe1Co3\u2013Se MAs is 24.58\u00a0\u03a9, which is much lower than that of Mo0\n.\n3Fe1Co3 MAs (41.45\u00a0\u03a9) and Fe1Co3 MAs (70.41\u00a0\u03a9), indicating its excellent electron transfer capability.To further illustrate the oxygen reduction activity of Mo0\n.\n3Fe1Co3\u2013Se MAs, CV test of different aerogels in oxygen saturated solution (0.1\u00a0M KOH) is shown in Fig.\u00a05f. Mo0\n.\n3Fe1Co3\u2013Se MAs displays a stronger oxygen reduction characteristic peak (\u22120.17\u00a0V vs. SCE), comparing with Fe1Co3 MAs (\u22120.2\u00a0V) and Mo0\n.\n3Fe1Co3 MAs (\u22120.18\u00a0V), and a larger positive potential and peak intensity indicate that Mo0\n.\n3Fe1Co3\u2013Se MAs has better ORR activity. The peaks at \u22120.61\u00a0V may be attributed to the oxidation process of Co or Fe. The ORR properties of FeCo aerogels were further investigated by testing continuous RDE and RRDE scans in O2-saturated electrolyte (0.1\u00a0M KOH) simultaneously changing the rotation speed between 400 and 1600\u00a0r/min in sequence. RDE tests of Fe1Co3 MAs at different rotational speeds is displayed in Fig.\u00a06\na. The current intensity corresponding to the voltage from 0.15 to 0.3\u00a0V vs. RHE with different speeds was used to calculate the electron transfer number according to K-L equation (Eqs. S1 and S2). The calculated electron transfer number (n) of Fe1Co3 MAs is about 2.9, indicating that it tends to occur 2e\u2212 ORR process. To further illustrate the selectivity of the prepared CoFe aerogels, the corresponding RRDE test results are presented in Fig.\u00a06b and c. Fig.\u00a06b shows that i\nring and i\ndisk intensities increase with increasing rotational speed, and the n value at 1600\u00a0r/min is calculated to be ca. 2.7 according to\u00a0\u200bEq. S3, which is close to the calculated result (2.9) of the K-L equation. LSVs of the prepared materials at 1600\u00a0r/min is shown in Fig.\u00a06d, the positive onset potentials of Fe1Co3 MAs, Mo0\n.\n3Fe1Co3 MAs and Mo0\n.\n3Fe1Co3\u2013Se MAs are identified as 0.67, 0.68 and 0.69\u00a0V vs. RHE, and the potential of Mo0\n.\n3Fe1Co3\u2013Se MAs is close to the thermo-dynamic limitations of the 2e\u2212 oxygen reduction process, which indicates that it has better oxygen reduction activity. Besides, the Tafel slope of Mo0\n.\n3Fe1Co3\u2013Se MAs (Fig.\u00a0S6), is 53.6\u00a0mV/dec in the corresponding voltage range, lower than that of Mo0\n.\n3Fe1Co3 MAs (60.9\u00a0mV/dec), confirming the more effective electron transfer kinetics, which is consistent with its best ORR performance. Among the three samples, Mo0\n.\n3Fe1Co3\u2013Se MAs has the largest i\nring and i\ndisk intensities, especially, this dramatically increased i\nring is an index of low electron transfer number and high H2O2 selectivity. Fig.\u00a06e shows the H2O2 selectivity of the different aerogel catalysts. The selectivity of Mo0\n.\n3Fe1Co3\u2013Se MAs for H2O2 can be up to more than 85% and stable over a wide voltage range, and this performance is significantly better than that of Fe1Co3 MAs (ca. 65%) and Mo0\n.\n3Fe1Co3 MAs (ca. 69%). Correspondingly, the number of electron transfers of Mo0\n.\n3Fe1Co3\u2013Se MAs can be as low as ca. 2.2, while the electron transfer numbers of Fe1Co3 MAs and Mo0\n.\n3Fe1Co3 MAs are around 2.7 and 2.5 (Fig. 6f), respectively. It should be noted that the ORR performance of H2O2 synthesis of Mo0\n.\n3Fe1Co3\u2013Se MAs is encouraging and it exceeds many of other electrocatalysts reported (Table\u00a0S4), which provides a sufficient guarantee for the subsequent activation of H2O2 to generate \u00b7OH.The materials structural reusability of Mo0\n.\n3Fe1Co3\u2013Se MAs was confirmed after OER and stability testing, and the morphological features and XPS spectra of the catalyst after OER are displayed in Fig.\u00a0S7 and Fig.\u00a0S8a. The original nanowire cross-linked skeleton structure morphology of the Mo0\n.\n3Fe1Co3\u2013Se MAs is well preserved without the phenomenon of structural damage. Co 2p spectra (Fig.\u00a07a) shows Co\u2013Se bond is still well preserved, and the Co\u2013Co bond disappears, which demonstrates that the Co\u2013Se bond is fully exposed in the electrochemical catalytic process, at the same time, the internal active site is fully exposed, facilitating the adsorption and reaction of intermediates in the catalytic process. Additionally, for the Se 3d, the strength of Se\u2013O bonds increased obviously after cyclic decay, and the strength of Se\u2013Se has been weakened, which may be related to the oxidation of Se due to the electrochemical reconfiguration process (Fig.\u00a07b). For the spectra of Mo 3d and Fe 2p, there is no obvious variation before and after OER process as displayed in Fig. S8b and c, which further indicates that Mo0\n.\n3Fe1Co3\u2013Se MAs has satisfactory stability in OER process. Fig.\u00a07c shows the i-t of the Mo0\n.\n3Fe1Co3\u2013Se MAs electrode in OER process for about 12\u00a0h. The results show that the current intensity decreased, but the variation is not obvious (maintained at more than 90% of initial current density) compared to the Fe1Co3 MAs (current density attenuation over 40%), indicating the excellent stability of Mo0\n.\n3Fe1Co3\u2013Se MAs catalyst. To demonstrate the stability, Mo0\n.\n3Fe1Co3\u2013Se MAs after stability testing was tested by XRD (Fig.\u00a07d). The XRD pattern still has the typical characteristic peaks of CoSe2 (PDF#53\u20130449). In contrast to the fresh sample, the characteristic peak of Mo0\n.\n3Fe1Co3\u2013Se MAs located at 45.1\u00b0 decreases, which is attributed to the reconstruction of the catalyst during the OER process. Moreover, the occurrence of the remodeling behavior is generally favorable for the improvement of the catalytic performance in OER. After i-t stability test, the Mo0\n.\n3Fe1Co3\u2013Se MAs sample still maintains the 3D network structure of aerogel, and there is no significant change compared with the fresh sample except for a small amount of nanowire agglomeration caused by the Nafion binder (Fig.\u00a0S9).Cycling tests were used to determine the durability of the material during the EF process (Fig.\u00a0S10). After 5 cycles, the degradation rate of SMT was still maintained at about 88%, and the decrease in catalytic performance could be attributed to the shedding of catalyst and SMT adsorption on surface active sites during the cycle. In addition, Fig.\u00a07d also shows the crystal phase structure of the Mo0\n.\n3Fe1Co3\u2013Se MAs cathode after EF, except for the weakened peak intensity, the crystalline features can be better matched with CoSe2 (PDF#53\u20130449), which is basically unchanged compared with the fresh sample. In addition, the v-t and i-t curves in SMT degradation are shown in Fig.\u00a0S11, when the current is set to 5, 10, 15\u00a0mA/cm2, the voltage of the entire SMT degradation process is basically kept stable, indicating that the degradation process is stable (Fig. S11a). At the current density of 15\u00a0mA/cm2, slight fluctuation in the v-t curve may be attributed to the acceleration generation of oxygen at anode, but the overall trend is still stable. In addition, the stability of the degradation process at different pH values was also investigated, as shown in Fig. S11b. At the current density of 10\u00a0mA/cm2, the voltage value shows an increasing trend with the increase of pH, but when the pH is 3.0, 4.0, 7.0, and 9.0, the voltage window is stable in the SMT degradation process, and the voltage is respectively stable at about 2.89, 2.94, 3.17, and 3.57\u00a0V. Based on the above results, electrochemical v-t tests under all conditions indicate the stability of the SMT degradation process. The i-t stability tests were performed using an electrochemical workstation with a three-electrode system (CHI 760E). The initial current density was adjusted to 10\u00a0mA/cm2 by setting a voltage of 2.9\u00a0V in the i-t program, the current density may fluctuate or decay as the reaction proceeds at this set voltage, so the stability of the reaction process is determined by monitoring the fluctuation of current density. As shown in Fig.\u00a0S11c, the overall trend remained stable after 90\u00a0min of degradation, and the final current density is maintained above 98% of the initial current density. The above encouraging results indicate that Mo0\n.\n3Fe1Co3\u2013Se MAs has excellent stability, which can be used as a coupled system to generate oxygen and decay SMT.EPR spectra were conducted to determine the free radical active species using DMPO as the trap agent in the EF-catalytic process. Fig.\u00a07e indicates that a classic 4-fold peak is found in the EPR spectrum of Mo0\n.\n3Fe1Co3\u2013Se MAs cathode and the ratio of peak intensities is about 1:2:2:1, which is a standard feature of DMPO-\u00b7OH adduct. In addition, with the accumulation of time increase, the signal peak intensity is gradually increasing, demonstrating that \u00b7OH is the major active free radical, and the corresponding generation path is shown in Eqs. (2)\u2013(9) [9,20,22], moreover, the metal of the high valence state will gain electrons at the cathode and be reduced to the low valence state, which contributes to the continuous activation of H2O2 [14,17]. In this process, clearly defined DMPO-SO4\n\u00b7- (six hyperfine lines; 1:1:1:1:1:1) also was observed, which may be caused by the activation of electrolyte SO4\n2\u2212 in the solution (Eq. (10)). To investigate the contribution of individual active free radical to SMT degradation in Mo0\n.\n3Fe1Co3\u2013Se MAs system, TBA, MeOH, and BQ (p-benzoquinone) were acted as the quenching agent of \u00b7OH, SO4\n\u00b7- and O2\n\u00b7-, respectively. Fig.\u00a07f shows that in the absence of TBA, the degradation performance of SMT decreased from 99.6% (without scavenger) in the control group to 53.8%. However, degradation performance was only reduced to 71.2% and 80.5% when MeOH and BQ was added into the electrolyte solution, suggesting that \u00b7OH played the key roles and SO4\n\u00b7- and O2\n\u00b7- played secondary roles for SMT decay in this EF catalysis, which is agreement with the results of EPR experiments.\n\n(2)\n\nM\n+\n\nH\n2\n\nO\n\u2192\n\nH\n+\n\n+\n\ne\n\u2212\n\n+\nM\n\n\n\u00b7\nOH\n\n\n\n\n\n\n\n(3)\n\n\nO\n2\n\n+\n\ne\n\u2212\n\n\u2192\n\nO\n2\n\n\u00b7\n\u2212\n\n\n\n\n\n\n\n(4)\n\n\nO\n2\n\n\u00b7\n\u2212\n\n\n+\n\nH\n+\n\n\u2192\nH\n\nO\n2\n\u00b7\n\n\n\n\n\n\n(5)\n\n\nO\n2\n\n+\n2\n\ne\n\u2212\n\n+\n2\n\nH\n+\n\n\u2192\n\nH\n2\n\n\nO\n2\n\n\n\n\n\n\n(6)\n\nF\n\ne\n\n2\n+\n\n\n+\n\nH\n2\n\n\nO\n2\n\n\u2192\nF\n\ne\n\n3\n+\n\n\n+\nO\n\nH\n\u2212\n\n+\n\u00b7\nOH\n\n\n\n\n\n(7)\n\nC\n\no\n\n2\n+\n\n\n+\n\nH\n2\n\n\nO\n2\n\n\u2192\nC\n\no\n\n3\n+\n\n\n+\nO\n\nH\n\u2212\n\n+\n\u00b7\nOH\n\n\n\n\n\n(8)\n\nF\n\ne\n\n3\n+\n\n\n+\n\ne\n\u2212\n\n\u2192\nF\n\ne\n\n2\n+\n\n\n\n\n\n\n\n(9)\n\nC\n\no\n\n3\n+\n\n\n+\n\ne\n\u2212\n\n\u2192\nC\n\no\n\n2\n+\n\n\n\n\n\n\n\n(10)\n\nS\n\nO\n4\n\n2\n\u2212\n\n\n+\n\u00b7\nO\nH\n\u2192\nS\n\nO\n4\n\n\u00b7\n\u2212\n\n\n+\nO\n\nH\n\u2212\n\n\n\n\nThe DFT calculations were utilized to deeply explain the catalytic process of Mo0\n.\n3Fe1Co3\u2013Se MAs towards OER. According to the above results, the electron transfer mechanism of the four-proton coupling is related to the OER process as the\u00a0\u200bEqs.11-14. Fig.\u00a08\n presents the free energy at each reaction stage, which suggested that the transformation of O\u2217 to OOH\u2217 intermediates on the surface of Mo0\n.\n3Fe1Co3\u2013Se MAs is the decisive step in the OER process, while the adsorption of OH\u2212 on the surface of Fe1Co3 MAs and Mo0\n.\n3Fe1Co3 MAs is the rate-control step, because Mo0\n.\n3Fe1Co3\u2013Se MAs exhibits the lowest adsorption energy (1.58\u00a0eV) of OH\u2212, which is much lower than that of Fe1Co3 MAs (1.96\u00a0eV) and Mo0\n.\n3Fe1Co3 MAs (2.11\u00a0eV). The difference in rate-limiting step is attributed to the fact that Fe1Co3 MAs and Mo0\n.\n3Fe1Co3 MAs have properties similar to high-entropy alloys (HEA), in which the low electron density state of the active site of Co near the Fermi energy level leads to its low electronegativity, which is not favorable for OH\u2212 adsorption [47]. After selenization of Mo0\n.\n3Fe1Co3 MAs, the O2 formation on the CoSe2 surface is a peculiar mechanism of adsorbate evolution. OH\u2212 can be readily adsorbed on the surface of the metal Co site and then desorbed to form O\u2217, resulting in the third step becoming limiting. For the entire OER process, the reaction energy barrier was reduced from 2.11\u00a0\u200beV (Fe1Co3 MAs) to 1.96\u00a0eV (Mo0\n.\n3Fe1Co3 MAs) and 1.58\u00a0eV (Mo0\n.\n3Fe1Co3\u2013Se MAs), indicating that the Mo doping and formation of selenide greatly decreases the adsorption of OH\u2212, while promoting the transformation of \u2217OOH, which facilitates the production of O2. Based on the above DFT calculations results and the theory [48\u201350], if the influence of electric potential is not considered, the ORR process can be considered as the reversal process of OER. It can be seen from the step diagram that the energy barrier of OOH\u2217 to O2 step is appropriate for Mo0\n.\n3Fe1Co3\u2013Se MAs (from 1.53\u00a0eV to 1.27\u00a0eV), which implies that 2e\u2212 ORR is more possible because the larger anion increases the separation between adjacent Co active sites in CoSe2 [19].\n\n(11)\nM\u00a0\u200b+\u00a0\u200bOH\u2212\u00a0\u200b=\u00a0\u200bM\u00b7OH\u2217\u00a0\u200b+\u00a0\u200be\u2212\n\n\n\n\n\n(12)\nM\u00b7OH\u2217\u00a0\u200b+\u00a0\u200bOH\u2212\u00a0\u200b=\u00a0\u200bM\u00b7O\u2217\u00a0\u200b+\u00a0\u200bH2O\u00a0\u200b+\u00a0\u200be\u2212\n\n\n\n\n\n(13)\nM\u00b7O\u2217\u00a0\u200b+\u00a0\u200bOH\u2212\u00a0\u200b=\u00a0\u200bM\u00b7OOH\u2217\u00a0\u200b+\u00a0\u200be\u2212\n\n\n\n\n\n(14)\nM\u00b7OOH\u2217\u00a0\u200b+\u00a0\u200bOH\u2212\u00a0\u200b=\u00a0\u200bM\u00a0\u200b+\u00a0\u200bO2\u00a0\u200b+\u00a0\u200bH2O\u00a0\u200b+\u00a0\u200be\u2212\n\n\n\nIn this work, we prepared bimetallic Mo-doped CoFe aerogels by a simple sodium borohydride template method and selenized their surfaces by vapor deposition. Mo-doped FeCo\u2013Se aerogels were used as anode and cathode for electro-Fenton. The experimental results show that the optimal Mo0\n.\n3Fe1Co3\u2013Se MAs catalyst can remove 97.7% of SMT (10\u00a0mg/L) within 60\u00a0min at a current intensity of 10\u00a0mA/cm2, and the overpotential is 235\u00a0mV under the current intensity of 10\u00a0mA/cm2. The superior performance is due to that the unique porous cross-linked structure of aerogel endowed the catalyst with enriched active sites and efficient mass transmission paths. Mo doping can lead to the lattice contraction and metallic character of CoSe2, which is beneficial to accelerate electron transfer. In addition, DFT calculations indicated that the selenization treatment lowered the reaction energy barriers for the OER and ORR processes, thereby optimizing the reaction kinetics. RRDE test indicated that Mo0\n.\n3Fe1Co3\u2013Se has excellent 2e\u2212 ORR activity with H2O2 selectivity up to 88%. Fe active sites can effectively activate H2O2 to generate \u2027OH. The excellent 2e\u2212 ORR and Fenton-like activity ensure its excellent EF performance. This work provided a novel perspective on the exploration of transition metal-based materials for EF process.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Thanks for the support of the National Natural Science Foundation of China (No.21776308) in this work.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.gce.2022.11.003.", "descript": "\n                  Antibiotic pollution in aqueous solutions seriously endangers the natural environment and public health. In this work, Mo-doped transition metal FeCo\u2013Se metal aerogels (MAs) were investigated as bifunctional catalysts for the removal of sulfamethazine (SMT) in solution. The optimal Mo0\n                     .\n                     3Fe1Co3\u2013Se catalyst can remove 97.7% of SMT within 60\u00a0min (SMT content: 10\u00a0mg/L, current intensity: 10\u00a0mA/cm2). The unique porous cross-linked structure of aerogel confered the catalyst sufficient active sites and efficient mass transfer channels. For the anode, Mo0\n                     .\n                     3Fe1Co3\u2013Se MAs exhibits superior oxygen evolution reaction (OER) property, with an overpotential of only 235\u00a0mV (10\u00a0mA/cm2). Compared with Fe1Co3 MAs or Mo0\n                     .\n                     3Fe1Co3 MAs, density functional theory (DFT) demonstrated that the better catalytic capacity of Mo0\n                     .\n                     3Fe1Co3\u2013Se MAs is attributed to the doping of Mo species and selenization lowers the energy barrier for the \u2217OOH to O2 step in the OER process. Excellent OER performance ensures the self-oxygenation in this system, avoiding the addition of air or oxygen in the traditional electro-Fenton process. For the cathode, Mo doping can lead to the lattice contraction and metallic character of CoSe2, which is beneficial to accelerate electron transfer. The adjacent Co active sites effectively adsorb \u2217OOH and inhibit the breakage of the O\u2013O bond. Rotating ring disk electrode (RRDE) test indicated that Mo0\n                     .\n                     3Fe1Co3\u2013Se MAs has excellent 2e\u2212 ORR activity with H2O2 selectivity up to 88%, and the generated H2O2 is activated by the adjacent Fe site through heterogeneous Fenton process to generate \u00b7OH.\n               "}
{"full_text": "Biomass gasification has attracted considerable attention as a technology to reduce environmental pollution and to face the steady increase in the world heat and power consumption. Gasification converts solid biomass into biosyngas, a gaseous mixture of hydrogen, carbon monoxide, methane, carbon dioxide, alongside nitrogen and water vapour, whose presence and concentrations are highly dependent on the gasification agent used. However, biosyngas also contains minor concentrations of species that might damage downstream equipment. These contaminants include nitrogen compounds (NH3 and HCN), sulfur compounds (H2S), halides (HCl), particulate matter (carbon and solid metals), and tar (e.g., toluene and naphthalene). Among biosyngas contaminants, tar requires particular attention.A common definition of tar is that of organic hydrocarbons with a molecular weight (MW) higher than benzene [1]. Otherwise, Milne et al. defined tar as the organic compounds, largely aromatic, released during pyrolysis or gasification of any organic material [2]. The same authors developed a tar classification method based on the maturation of tar from their initial release during the pyrolysis step. The first step starts at 200\u2013500\u00a0\u00b0C, with the conversion of cellulose, hemicellulose and lignin to oxygenated hydrocarbons called primary tar. Above 500\u00a0\u00b0C, primary tar is converted into phenolics and olefins compounds known as secondary tar. When the temperature is increased above 800\u00a0\u00b0C, part of the secondary tar converts to polycyclic aromatic compounds, also called tertiary [2].Being mostly composed of hydrogen and carbon, tar is potentially a fuel. However, tar compounds can cause various problems such as condensation in cold spots, polymerization to form complex structures that condense even at high temperature, and formation of carbonaceous deposits which can deactivate catalysts or plug filter pores, gas lines and heat exchangers. The term carbonaceous deposit is used in this work to indicate a deposit mostly consisting of carbon, that might have originated from thermal cracking, catalytic reactions, or polymerization [3]. To avoid the aforementioned issues, tar is often removed from the gas stream at low or intermediate temperature (below 300\u00a0\u00b0C) via physical methods, or converted into useable gases at high temperature (above 1000\u00a0\u00b0C) via processes such as reforming.Reforming consists in converting tar into CO and H2 using CO2 (dry reforming) or H2O (steam reforming). The process temperature can be lowered by using a catalyst, that can be added inside the gasifier (primary methods) or in a downstream reactor (secondary methods). Catalysts can be classified in several different ways and the most important distinction can be made between metal-based, or synthetic, catalyst (e.g., Ni and Fe based catalysts) and naturally-occurring, or mineral, catalysts (e.g., dolomite and olivine). Generally, mineral catalysts are less catalytically active, but also less expensive. The latter is a significant advantage since the use of catalyst increases the overall system cost [4]. Moreover, naturally occurring catalyst could exhibit higher tolerance to impurities, such as H2S and HCl, or simply be replaced in case of deactivation due to carbon deposition [5]. Therefore, despite the lower activity, mineral catalysts, such as dolomite and olivine, are thoroughly studied [6\u20139].Most of the research on biosyngas tar reforming has focused on secondary and tertiary tar, since the majority of the gasifiers are operated in the temperature interval 700\u20131000\u00a0\u00b0C (downdraft and fluidized bed gasifiers) [10]. Selected representative compounds or real tar have been used in these studies. As an example, Ren et al. used toluene as model compound to study the effect on the reforming activity of different Ni loading methods and oxidation degree of the char support [11]. Recently, Ashok et al. reviewed the progress in the development of catalysts for steam reforming of biomass tar using toluene, benzene and naphthalene as tar model compounds [12]. Various catalysts have been compared by Simell et al. on the ability to reform tar generated in a full-scale updraft gasifier [13,14]. This research found that silicon carbide is an inert material and the activity of the catalysts decreases in the order Ni on Al2O3\u00a0>\u00a0dolomite\u00a0>\u00a0activated alumina\u00a0>\u00a0silica-alumina [14]. Other authors tested different materials, as Fuente-Cano et al., who studied the conversion of tar generated in a steam-blown fluidized-bed gasifier using wood char as catalyst [15]. Furthermore, innovative system configurations have been investigated by Savuto et al. who achieved tar reforming by placing metal-based catalysts inside the ceramic filter candles in the freeboard of a fluidized bed gasifier [16]. Nonetheless, updraft gasifiers present some advantages over downdraft and fluidized bed gasifiers: operability with a broad spectrum of feedstock (in terms of ash, moisture and size), very low particulate matter entrainment (both dust and ash), simplicity of construction, technology maturity and robustness and high efficiency [17]. On the other hand, updraft gasifiers have the disadvantage of producing a larger amount of tar, mostly belonging to the primary tar group.Studies on biosyngas primary tar reforming are limited in literature. Most of primary tar have a molecular weight lower than that of benzene; therefore, they are often not considered as tar. Moreover, problems with tar such as condensation at high temperature and formation of carbonaceous deposits are promoted by aromatic hydrocarbons (secondary and tertiary tar). Conversely, primary tar are volatile compounds (e.g., formaldehyde, furan) or light condensable compounds (e.g., acetic acid, hydroxyacetone) with low condensation temperatures. Nonetheless, also primary tar represent a threat for downstream equipment due to their corrosiveness [18] and, if temperature is increased, their tendency to form carbonaceous deposits and harmful secondary/tertiary tar compounds via cracking or polymerization [19,20]. Furthermore, the wastewater generated by low temperature physical removal methods, such as scrubbing, has to be treated to prevent environmental issues caused by these organic molecules [21,22].Some information on primary tar reforming can be derived from studies on catalytic steam reforming of bio-oil and its representative compounds. In fact, bio-oil is a product of biomass pyrolysis and consists predominantly of a mixture of primary tar as aldehydes, alcohols and acids derived from the carbohydrate fraction of biomass, and phenolics derived from lignin [23]. Detailed reviews on catalytic reforming of bio-oil and its representative compounds can be found in literature [24\u201327]. In general, the activity of transition metal catalyst towards the steam reforming of acetic acid is Ni\u00a0>\u00a0Co\u00a0>\u00a0Fe\u00a0>\u00a0Cu [28]. Ni and Co are much more catalytically active than Fe and Cu, but less active than noble metals [29]. The majority of studies focus on metal-based catalyst, whereas a very limited amount of research has investigated natural catalysts, such as dolomite and olivine [30,31]. As with most tar compounds, one of the main reasons for catalyst deactivation in studies with bio-oil and its representative compounds is the formation of carbonaceous deposits. Studying the catalytic steam reforming of acetic acid on Ni/\u03b3-Al2O3, An et al. concluded that the carbonaceous deposits originated from the catalytic cracking reactions and the CO disproportionation reaction [32]. With Pt/ZrO2 and Pt/CeO2 catalysts, the formation reaction of acetone via acetic acid condensation/dehydration was identified as the carbonaceous deposits formation mechanism [33]. Additionally, since the oxygenates are often thermally unstable, thermal cracking can also lead to carbonaceous deposits formation [34,35].Although the studies on bio-oil and its representatives compounds (e.g., acetic acid) provide some insights, most studies use humidified N2, Ar or He as gas carrier [36\u201338]. However, the presence of CO, CO2 and CH4 in biosyngas can have a significant impact on the catalytic reforming of tar compounds [39]. In fact, CO2 can increase tar conversion via dry reforming and gasify carbonaceous deposits, H2 and CO can shift the reforming reaction equilibrium towards the reactants, and tar and biosyngas components might compete for catalysts active sites [40]. Therefore, this study evaluates the ability of three different catalysts to reform acetic acid, selected as main primary tar compound from updraft gasification, using simulated biosyngas as gas carrier. Particular attention was given to one of the catalysts tested, dolomite, owing to the promising results obtained and the opportunity this natural catalyst offers to decrease system costs. To the best of the authors knowledge, this represent one of the few, if not the first study focusing on acetic acid reforming under updraft gasification representative conditions. The insights presented are expected to help the further development of high temperature gas cleaning, and the commercialisation of energy conversion systems based on updraft biomass gasifiers.In this study, the ability of one metal-based catalyst and two mineral catalysts (i.e., dolomite and olivine) to reform acetic acid was compared using simulated biosyngas as gas carrier. The results are compared with a benchmark test performed using activated alumina. This material was selected since it is often used as support for metal-based catalyst and it shows limited catalytic activity. Biosyngas was used as gas carrier instead of a simple carrier (e.g., humidified N2 and He) as it can have a significant impact on the reforming process. Moreover, the use of biosyngas gives the possibility to evaluate the catalyst activity towards methane reforming and Water-Gas-Shift (WGS) reactions, thus providing additional useful information for the design of tar reformers for updraft biomass gasifiers based systems.Acetic acid was selected as model compound since it is the main component of primary tar, that is the largest group of tar generated in updraft gasifiers [2]. Above 500\u00a0\u00b0C primary tar undergoes the maturation process studied by Milne et al. [2]. Therefore, when the gas is heated up, some of the acetic acid might be converted into different compounds such as higher molecular weight compounds and carbonaceous particles even before reaching the catalyst bed, as observed also by Matas G\u00fcell et al. or by Boot-Handford et al. [33,41]. This conversion process is affected by the heating up conditions (e.g., surfaces available, residence time and temperature). Using acetic acid as initial compound was considered a more interesting option as compared to using only a secondary tar compound, as it provides conditions representative of an updraft gasifier based system, where primary tar compounds, together with higher molecular weight compounds and carbonaceous particles generated when heating up the biosyngas containing primary tar compounds, reach the reformer. Therefore, in the experiments performed, the three catalysts are evaluated on the ability to convert residual acetic acid and higher molecular weight compounds, and to withstand the deposition of carbonaceous particles generated while heating up the gas mixture. All catalysts were tested under the same operating conditions and, in the case of dolomite, additional tests were performed: one at lower temperature (400\u00a0\u00b0C), and one at higher temperature (900\u00a0\u00b0C) and higher steam content (50.1\u00a0vol%). The low-temperature test was performed to investigate the influence of dolomite on the process of heating up biosyngas containing acetic acid. Other catalysts (i.e., Ni based catalysts) are in fact reported to promote the ketonization reaction already at 350\u00a0\u00b0C [42]. The high-temperature and high steam content test was performed to investigate the possibility to completely suppress accumulation of carbonaceous deposits. The results of these additional tests provide a useful direction for future research on using dolomite for reforming tar from biomass updraft gasifiers without the formation of carbonaceous deposits.The tests were carried out in a bench-scale unit consisting of a quartz reactor mounted inside an electric furnace. The Process and Instrumentation Diagram (P&ID) of the unit is illustrated in Fig. 1\n. The electric furnace is an insulated ceramic hollow cylinder with a heating coil wrapped along its length. A thermocouple is placed at the middle of the cylinder height, on the inner cylinder surface, to control the temperature of the heating coil. The quartz reactor has an inner diameter of 2\u00a0cm and the catalyst bed, kept in place thanks to a quartz frit, has a height of 5.5\u00a0cm. A thermocouple is placed next to the top part of the catalyst bed to monitor the temperature during the tests. Before starting the experimental campaign, this thermocouple was used to measure the temperature of the furnace along its length. The furnace showed a temperature gradient and, as a consequence, the catalyst bed was not kept at a uniform temperature: the bottom part was at 680\u00a0\u00b0C, while the top part at 750\u00a0\u00b0C. The top of the catalyst bed was placed at the highest temperature in the furnace. Fig. 2\n shows the furnace and the design of the quartz reactor next to the temperature profile over the setup.A gas flow rate of 380 NmL/min simulated biosyngas composed of 133.1 NmL/min H2, 8.9 NmL/min CO, 74.1 NmL/min CO2, 13.7 NmL/min CH4, 91.1 NmL/min H2 and 59.1 NNmL/min N2 (35.0\u00a0vol% H2O, 2.3\u00a0vol% CO, 19.5\u00a0vol% CO2, 3.6\u00a0vol% CH4, 24.0\u00a0vol% H2 and 15.6\u00a0vol% N2) was used. The volume percentages correspond to the measured values during the first \u201cFlexiFuel-SOFC\u201d project experimental campaign. This EU funded project aimed at the development of a micro scale combined heat and power system composed of an updraft biomass gasifier, a high temperature gas cleaning unit and a solid oxide fuel cell (SOFC) [43,44]. The gas flow rates were regulated using mass flow controllers Bronkhorst EL-FLOW (Bronkhorst, The Netherlands). Steam was added to the fuel gas stream by bubbling the gas mixture (except CO2) in a temperature controlled water bath (humidifier). It was assumed that the gas in the water bath was constantly in equilibrium with the liquid phase. Therefore, the steam content is a function of the liquid temperature, according to Antoine's equation. The piping after the humidifier and the bottom part of the quartz reactor were trace heated and kept at 150\u00a0\u00b0C.The tar concentration at the inlet was 40\u00a0g/Nm3, and was selected based on the results of tar sampling from the \u201cFlexiFuel-SOFC\u201d project updraft gasifier that used wood chips as feedstock and humidified air as gasifying agent. The analysis was performed following the tar protocol (CEN TC BT/TF 143 WI CSC 03002.4) [1]. Liquid acetic acid of 99.7% purity (Sigma Aldrich, USA) was injected at the entrance of the quartz reactor using a peristaltic pump BT100-2\u00a0J (Longer Precision Pump Co., China). The trace heating temperature assured the continuous evaporation of the acetic acid. With the steam and acetic acid flow rates selected, the steam-to-carbon (S/C) ratio results equal to 1.3 and to 1.8 for the tests performed with dolomite at 900\u00a0\u00b0C and different gas composition.At the reactor outlet, a three way valve directed the gas flow either to a series of impinger bottles for tar sampling, or to a microGC to monitor the outlet gas composition. An Agilent 490 microGC with thermal conductivity detector and a CP-COX column for measuring CO, H2, N2, CH4 and CO2 (Agilent, USA) was used. Before reaching the microGC, the gas was passed through a condenser and a desiccator containing silica-gel to remove the moisture contained in the gas. The outlet flow rate was back-calculated from the inlet N2 flow rate and the N2 outlet concentration that was measured with the microGC. This was then used to calculate the flow rates of H2, CO, CO2 and CH4. The outlet tar concentration was measured by bubbling the outlet gas flow in a series of 4 impinger bottles, the first one kept empty and acting as moisture collector, two containing isopropanol at room temperature, and a last one containing isopropanol at 0\u00a0\u00b0C. The isopropanol and water samples were analyzed with a Varian 430\u00a0GC-FID (Agilent, USA) equipped with a Rtx-1 column (Restek, USA).The catalyst used were all in the form of particles with a diameter of 2\u20133\u00a0mm. The metal-based catalyst was a commercially available catalyst called TARGET\u2122 developed specifically for tar reforming by the company Nexceris and consisted of Pt/MNS (MgO, NiO, SiO2). The olivine used is produced by the manufacturer specifically for biomass gasifiers bed, and the material is reported to have a catalytic activity close to that of calcined dolomite [45]. The dolomite was supplied in partially calcined form (CaCO3\u00b7MgO), that is with MgCO3 already converted into MgO and with Ca still in the form of CaCO3. An additional calcination process was performed at 900\u00a0\u00b0C to increase the dolomite catalytic activity. However, the catalyst became too brittle; consequently, the calcination temperature was lowered to 800\u00a0\u00b0C but the dolomite mechanical resistance was still significantly affected. Therefore, the dolomite used in the tests was in the partially calcined form as received from the supplier. The chemical composition of the partially-calcined dolomite and of the olivine is shown in Table 1\n.After having positioned the quartz reactor in the furnace, the temperature was increased to the set value with a ramp of 50\u00a0\u00b0C/h. A nitrogen flow of 100 NmL/min was passed through the catalyst bed during the heating up stage. The gas flow rate was then changed to simulated biosyngas and the gas composition was monitored for 12\u00a0h before adding acetic acid. In the case of the metal-based catalyst, acetic acid was injected after 18\u00a0h. The longer time was used to assure full reduction of the catalyst. The gas composition at the outlet of the reactor was monitored during this time to determine the catalyst activity towards WGS and methane reforming and to be sure that the catalyst was not undergoing any change affecting its catalytic activity.Acetic acid was successively injected and the experiment was kept running for three days. During the day the outlet gas composition was monitored with the microGC while during the night the outlet gas was led through the sampling bottles to measure the residual tar content with the GC. The gas composition results presented are the average values recorded over time. After the three days operation, the setup was cooled down with a ramp rate of 50\u00a0\u00b0C/h and a gas flow rate of 100 NmL/min N2 passing through the catalyst bed. The catalyst was then visually inspected for carbonaceous deposits. An elemental and morphological analysis of the deposit was not performed since the scope of this work was limited to the performance comparison of the different catalysts, and to the identification of a suitable material and conditions to be used in an integrated biomass gasifier SOFC microCHP system. Table 2\n summarises the tests performed and the relevant parameters.Before proceeding with the experimental tests, thermodynamic equilibrium calculations were performed using the software FactSage version 5.4.1 (Thermfact/CRCT, Montreal, Canada and GTT-Technologies, Aachen, Germany) to assure the catalysts were operated outside the theoretical solid carbon formation region [46]. The results of the calculations also provide an indication of the expected gas composition at the outlet of the catalyst bed if thermodynamic equilibrium conditions are reached. The software calculates the concentrations of chemical species when specified elements or compounds react to reach a state of chemical equilibrium. The users specifies the mass of the reactants, process temperature and pressure, and the software solves a Gibbs minimization algorithm based on three constraints: the equilibrium product amounts are positive, the mass balance with respect to the system components is satisfied and correspond to the lowest possible Gibbs energy for the possible products. A detailed explanation on the method followed for the calculation can be found on the specific software webpage [47]. Thermodynamic equilibrium calculations only give an indication of the possibility for solid carbon formation since equilibrium might not be reached during the test. Moreover, the software database used contains only thermodynamic properties of solid carbon in the form of graphite.\nFig. 3\n shows the carbon-hydrogen-oxygen ternary diagram calculated using the software FactSage. A specific mixture containing these elements is represented by a point in the diagram. The operating points without tar (triangle) and with tar (dot) were calculated taking into account all the biosyngas compounds containing hydrogen, oxygen and carbon. In the diagram, two regions can be distinguished, one where solid carbon is formed and one where the mixture completely remains in the gas phase. The two regions are delimitated by a line whose position depends on the temperature. If the operating point falls on the left of the line, then solid carbon is formed at equilibrium. The results indicate that, if equilibrium conditions are reached, no carbonaceous deposits should be present in the whole temperature interval in which the catalyst was operated. Both operating points in fact fall on the right of the continuous and of the dotted line, representing the operating temperatures of 750\u00a0\u00b0C and 680\u00a0\u00b0C, respectively.The software was also used to calculate the equilibrium gas composition of clean biosyngas at 750\u00a0\u00b0C with and without acetic acid, as illustrated in Table 3\n. If equilibrium is reached, methane is almost completely reformed and an increase in the carbon monoxide flow rate is expected because of the reverse water gas shift (RWGS) reaction occurring in the reactor. Despite hydrogen is converted into water by the RWGS reaction, the outlet flow rate is higher than the inlet flow rate due to methane being reformed. Similarly, the flow rate of water remains almost constant since the amount produced via the RWGS is balanced by the amount consumed via methane reforming. When acetic acid is added to biosyngas, its conversion leads to an increase in H2 and CO, and to a minor extent of CO2 flow rates.A pre-test was performed without filling the reactor with any catalyst to determine whether reactions towards equilibrium take place at 750\u00a0\u00b0C even without any catalyst. The gas composition measured at the outlet of the furnace, shown in Fig. 4\n, indicates that the residence time was not sufficient to allow any reaction to a noticeable extent. The minor differences between the set and measured inlet values were probably caused by small inaccuracies in the mass flow controllers\u2019 calibration. During this test, the inlet concentration of acetic acid was also measured by wet sampling and the results show an inlet concentration of 37\u201341\u00a0g/Nm3, corresponding to a gas volume flowrate of 3.4\u20133.8 NmL/min.The reference test with activated alumina beads in the reactor showed that alumina interacts with the gaseous species but it is not significantly catalytically active. Fig. 5\n shows the gas composition at the inlet and at the outlet of the reactor when biosyngas with and without acetic acid was passing through the alumina bed. While the flow rate of CH4 remained unchanged, H2 and CO2 reacted and formed CO via the reverse water gas shift reaction. When acetic acid was added to the biosyngas, there was an increase in the methane outlet flowrate, while the other compounds remained almost unchanged. This shows that part of the acetic acid undergoes cracking to CH4 rather than catalytic reforming leading to CO, H2 and CO2, as observed by Basagiannis et al. [42]. Unfortunately, the microGC was not calibrated for any hydrocarbon other than methane nor for oxygen. Moreover, the water condensed at the outlet of the reactor and the amount and composition of the carbonaceous deposits were not measured. It is therefore not possible to give details on the acetic acid thermal decomposition pathway. The study of acetic acid decomposition mechanism is however considered beyond the scope of this paper.No acetic acid was detected at the reactor outlet by wet sampling. However, an amount of hydroxyacetone between 0.07 and 0.45\u00a0g/Nm3 was measured. In this test, the first bottle of the sampling train was filled with 30\u00a0ml of isopropanol. The mix of isopropanol and condensed water contained in the first bottle turned slightly yellow and had a distinct odour typical of aromatic compounds. This might indicate that part of the acetic acid underwent the maturation process described in Ref. [2]. Nonetheless, the analysis with GC-FID did not show any tar compound. This could have been due to the inability of the column to separate the compounds, or to the high dilution caused by the water condensed in the first sampling bottle. Gravimetric analysis could have been performed on the liquid samples to confirm the presence and quantify the total amount of compounds that were not being detected by GC-FID analysis. To avoid dilution in the successive tests, the first bottle was kept empty and served as moisture collector.At the end of the test, the alumina beads were fully covered with carbonaceous deposits. In literature it is reported that acidic supports as Al2O3 have the tendency to cause a larger amount of carbonaceous deposits as opposed to basic supports which enhance water adsorption [29,42]. The deposit was also present on the reactor walls where the temperature was above 400\u00a0\u00b0C, that is even before the catalyst bed. Therefore, the carbonaceous deposits might have formed directly in the bed, and/or it might have formed during the heating up process, after which the carbonaceous particles accumulated in the bed by a filtering effect. The formation of carbonaceous deposits on the reactor walls might have occurred due to a radial temperature gradient resulting in higher temperature near the surface, or to acetic acid reacting on the reactor surface. From the experiment performed it is not possible to know if the conversion of acetic acid was complete even before reaching the alumina bed. To better understand the thermal cracking of acetic acid and causes for carbon deposits on the reactor wall and catalytic bed, a test with an empty reactor could be performed. Nonetheless, the understanding of acetic acid thermal cracking behaviour was considered outside the scope of this study and the results of the tests with alumina represent the base case for the comparison of the three catalysts tested.\nFig. 6\n presents the gas composition when olivine was used as catalyst. The catalyst showed a very limited catalytic activity towards the reverse water gas shift reaction and did not significantly catalyse methane reforming. Moreover, when acetic acid was added to biosyngas, there was an increase in the methane outlet flowrate. However, also the flow rates of CO, H2 and CO2 slightly increased which might indicate the occurrence of acetic acid catalytic reforming.The wet sampling showed no acetic acid at the reactor outlet and no other compound was detected during the first two days of measurement. Nonetheless, the third day of sampling 0.02\u20130.19\u00a0g/Nm3 of hydroxyacetone were measured, which might indicate that the catalytic activity was reducing over time. For this reason, the test was extended for an extra day and during the next sampling the amount of hydroxyacetone increased to 0.14\u20130.30\u00a0g/Nm3, thus confirming a decreased activity of the catalyst.After having cooled down the furnace, carbonaceous deposits considerably covered the catalyst bed. The acetic acid was therefore completely converted into non-condensable gases and carbonaceous deposits, at least during the first two days of operation. Also in this case, the deposits were found also on the reactor walls. It is not known if the carbonaceous deposits on the catalyst originated from reactions in the catalytic bed and/or might have formed before the catalyst, carried by the gas and then filtered by the catalyst bed.The results in Fig. 7\n show that roughly 10 NmL/min of H2 reacted with CO2 to form CO, thus indicating that dolomite is catalytically active for the reverse water gas shift reaction. The outlet CO flow rate is still far from the value expected at equilibrium, partially due to the lack of catalytic activity of dolomite towards methane reforming, and partially to the reverse water gas shift reaction not reaching equilibrium. When acetic acid was added to the biosyngas, methane outlet flow rate increased while the flow rates of the other compounds remained almost unchanged.Tar sampling indicated that all acetic acid was converted, and no other compounds were found at the reactor outlet. At the end of the test, carbonaceous deposits were present on the reactor walls and on the catalyst. Therefore, acetic acid was converted into non-condensable gases and carbonaceous deposits. Interestingly, the deposits were found only in the first 2.5\u00a0cm of the catalyst bed, while the top part was clean. This might have been due to the conversion of all acetic acid and intermediates taking place before this top section or to the ability of the catalyst to convert acetic acid without carbonaceous deposits formation. Another explanation might be that from 2.5\u00a0cm onwards, the bed temperature was sufficiently high to allow the gasification of the carbonaceous deposits filtered. However, it seems unlikely that the carbonaceous deposits accumulated on the catalyst by filtration were oxidized again under the testing conditions, that is a reducing atmosphere and moderate temperatures. Therefore, it is more likely that the catalyst was active towards acetic acid conversion without carbonaceous deposits formation. Fig. 8\n shows a schematic of the process described.An additional test was performed keeping the top of the catalyst bed at 400\u00a0\u00b0C to investigate the influence of dolomite on the process of heating up biosyngas containing acetic acid. Moreover, considering that dolomite appeared not to suffer any deposition at temperatures above roughly 730\u00a0\u00b0C, an additional test was performed to verify the possibility to suppress carbonaceous deposits accumulation completely. In this second test, a different gas composition was used, with a water flow rate of 190 NmL/min, corresponding to 50% vol, and the top of the bed kept at 900\u00a0\u00b0C. Fig. 9\n compares the measured flow rates with the calculated equilibrium flow rates. At both the temperatures tested, the gas composition was far from equilibrium. At low temperature, that is 400\u00a0\u00b0C, while CO should be absent and the H2 content is supposed to be significantly lower than the measured values, the CH4 flow rate should be higher than the inlet value due to methane formation reaction. However, dolomite is not catalytically active towards methane formation or reforming reactions, as visible also from the outlet methane flowrate measured at 900\u00a0\u00b0C, that is almost equal to the inlet value and higher than the expected equilibrium amount.\nFig. 10\n and Fig. 11\n show the gas composition measured at the inlet and outlet of the reactor in the two tests. The test at low temperature clearly showed that dolomite had no catalytic activity at this temperature. This was confirmed by the sampling at the outlet which resulted in an amount of acetic acid equal to 35.5\u201339.4\u00a0g/Nm3, corresponding to 3.3\u20133.6 NmL/min. At the end of the test, the dolomite at the top of the bed was light grey, indicating a very minor presence of carbonaceous deposits. At this temperature also thermal cracking of acetic acid almost did not take place, in accordance with the observations of An et al. [32]. In the test at high temperature, that is 900\u00a0\u00b0C, the water gas shift reaction occurred while the methane content remained almost unchanged. Interestingly, a slightly more evident increase in H2 and CO flow rates was noticed as compared to the tests at lower temperatures, suggesting the occurring of acetic acid catalytic reforming. No acetic acid or other tar compound was measured at the reactor outlet and no carbonaceous deposits were observed.The results obtained might have been due to the higher temperature or the higher steam content in the gas since both variables are expected to suppress the formation of carbonaceous deposits. The results therefore provide only a preliminary indication of the possibility to suppress carbonaceous deposits accumulation. However, as experienced during the pre-calcination tests at 800\u00a0\u00b0C and 900\u00a0\u00b0C, the catalyst became excessively brittle, with some of the catalyst beads at the top of the bed losing their structure. The loss of mechanical strength of dolomite might be explained by the occurrence of secondary calcination, converting the partially-calcined dolomite (MgO\u2013CaCO3) to full-calcined dolomite (MgO\u2013CaO). Half-calcined dolomite is a rigid and strong material whereas full-calcined dolomite can be very fragile [48]. The secondary calcination reaction depends on process temperature and CO2 partial pressure; upon checking the secondary calcination temperature with thermodynamic equilibrium calculations, it can be noticed that secondary calcination is expected to occur approximately at 750\u2013760\u00a0\u00b0C. Therefore, further tests maintaining the dolomite temperature below 750\u00a0\u00b0C and with high amounts of steam are suggested to identify under which circumstances carbonaceous deposit accumulation can be suppressed without compromising the catalyst attrition resistance.The biosyngas composition changed significantly when passing through the metal-based catalyst bed. The outlet flow rates of CO2 decreased while that of CO increased due to the reverse water gas shift reaction. The H2 outlet flow rate increased due to almost complete reforming of CH4. The gas composition is very close to the expected equilibrium composition presented in Table 3, for both clean biosyngas and acetic acid containing biosyngas. The increase in H2 and CO flow rates when acetic acid was injected indicates the occurring of reforming of acetic acid but also the reforming of the CH4 generated during acetic acid thermal decomposition during the heating up process. Fig. 12\n presents the gas composition measured at the outlet of the reactor with and without acetic acid.Wet sampling showed no traces of any tar compounds, thus indicating the complete conversion of acetic acid into non-condensable gases and a very minor amount of carbonaceous deposits. Carbonaceous deposits were found on the reactor walls before the bed, and only a very minor amount was found on the first 0.5\u00a0cm of the catalyst bed; this deposit might have been filtered or formed in this section of the bed. Irrespectively from the formation mechanism, it can be concluded that at this temperature the metal based catalyst was able to almost completely suppress the accumulation of carbonaceous deposits.\nTable 4\n summarises the results previously discussed to facilitate a comparison between the different catalysts.\nFig. 13\n compares the gas composition measured at the reactor outlet with the different catalysts tested when biosyngas containing acetic acid was passed through the bed. It can be seen that in terms of catalytic activity towards the reverse water gas shift reaction and methane reforming, the metal-based catalyst has the best performances, with the outlet gas composition being almost equal to the expected equilibrium composition. Both dolomite and olivine showed some catalytic activity towards the reverse water gas shift reaction, with dolomite being more active than olivine. Neither dolomite nor olivine showed activity towards methane reforming and only some activity towards acetic acid reforming, with the majority of the primary tar being converted into methane and carbonaceous deposits. The result are in good agreement with literature, where it is often stated that metal-based catalysts outperform naturally-occurring catalysts [14].The metal-based catalyst also appeared as the most capable of suppressing carbonaceous deposits accumulation. The catalysts can be sorted as olivine\u00a0>\u00a0dolomite\u00a0>\u00a0metal-based from the highest to the lowest amount of deposits accumulated during the test. From the experiment performed, it is not possible to know if the carbonaceous deposits found on the catalysts were formed in the bed or before it. However, irrespectively from the formation mechanism, it is clear that olivine similarly to alumina was not able to suppress carbonaceous deposits accumulation, while dolomite above roughly 730\u00a0\u00b0C and the metal-based catalyst already around 680\u00a0\u00b0C did not suffer carbonaceous deposits accumulation. Moreover, while dolomite and metal-based catalyst completely converted acetic acid into non-condensable gases for the whole duration of the test, the performances of olivine appeared to decrease in time and, at the end of the third and the additional fourth testing day, small amounts of hydroxyacetone were found. Hydroxyacetone is a common intermediate product in the conversion of carboxylic acids, such as acetic acid, into ketones via ketonization reaction [35,49].The olivine used in this study is reported to have a catalytic activity close to that of calcined dolomite [45]. However, according to the manufacturer information, the material was sintered in a rotary kiln at 1600\u00a0\u00b0C for 3\u00a0h and, according to Corella et al. the sintering process strongly decreases the catalyst pore structure [50]. This might explain the performance of olivine observed in this study. Moreover, also Quan et al. observed a decrease in the catalyst performance of olivine due to calcination at higher temperatures [31]. However, since there are studies indicating that calcination might improve the performance of raw olivine [51], further investigation of the performances of raw and calcined olivine are suggested as future work. In case of dolomite, the catalyst used had a low iron content and literature indicates the iron promotes catalytic tar reforming and the WGS reaction [52]. Most importantly, dolomite was used in partially calcined form with Mg being in oxide form but Ca in the carbonate form. An additional test was attempted using dolomite that was pre-calcined at 800\u00a0\u00b0C. However, the gas flow rate was shortly stopped due to dolomite becoming powder and clogging the reactor. Moreover, the biosyngas atmosphere in our tests contained a high concentration of CO2 and upon checking with equilibrium calculations, it is possible that CO2 reacted with CaO and formed CaCO3 at temperatures lower than 800\u00a0\u00b0C. This form of Ca is believed not to be significantly catalytically active [53]. Nonetheless, the tests in this study showed that dolomite can almost completely convert the acetic acid and intermediate compounds into non-condensable gases at 680\u2013750\u00a0\u00b0C, with the carbonaceous deposits present only in the bottom part of the bed. Dolomite friability and release of fines might limit its use as tar reforming catalyst. However, this issue could be mitigated by adding a particle filtration system downstream the catalytic bed. Alternatively, more resistant catalysts should be developed, as studied by Miao et al. who prepared catalysts based on dolomite, clay and carboxyl methyl cellulose with increased strength resistance and higher porosity [30].The metal-based catalyst appeared to outperform the naturally-occurring catalysts with regard to reforming and resistance to carbonaceous deposits accumulation. However, it might suffer poisoning from other contaminants usually present in biosyngas, as chlorine and sulfur, and it has the disadvantage of higher cost [4]. Moreover, the lack of catalytic activity of olivine and dolomite towards methane reforming is not necessarily a drawback. This compound can remain in biosyngas and be used directly in downstream processes without causing issues in some applications for heat and electricity production. With some conversion technologies, such as solid oxide fuel cells, the presence of methane in biosyngas even increases the system efficiency owing to the cooling effect of direct internal methane reforming [54].Despite the differences in the operating conditions (temperature, space velocity, gas composition, tar concentration, and steam-to-carbon ratio), the tar abatement efficiency of the catalysts tested in this study can be compared with results found in literature. A short overview is illustrated in Table 5\n, where the catalyst tested and the abatement efficiencies are summarized. The most effective catalyst reported in literature are noble (i.e. Rh-, Pd-, Pt-) and transition (predominantly nickel) metal-based catalyst, followed by calcined dolomite, olivine, biochar, and ash. Ferrous metal oxides, clay, activated alumina, and fluid cracking catalyst are generally less effective. Several studies comparing catalyst with tertiary tar indicate a similar order of the catalyst based on their activity [55,56]. The high tar abatement efficiency of the commercial nickel catalyst (100%) found in this work agrees well with figures commonly found in literature on acetic acid reforming with Ni/Al2O3 [28,42,56\u201358] and several noble metal-based catalyst [59,60]. Furthermore, in this study the complete conversion of acetic acid was observed with dolomite and such abatement efficiencies have been previously reported [30,37]. The olivine in this work initially achieves a complete conversion of acetic acid as reported by Kechagiopoulos et al. [61,62]; however, with time the formation of carbonaceous deposits seemed to reduce the catalytic activity. The alumina catalyst studied in this work displayed a lower catalytic activity than the nickel, dolomite and olivine, which agrees with the lower tar abatement efficiencies of alumina supports reported in literature [29,63]. Although most catalyst in Table 5 obtain lower efficiencies than transition/noble metal based catalyst, these compounds are often used as support materials.The formation of carbonaceous deposits on the catalyst and sulfide poisoning can severely limit the tar abatement efficiency and lifetime of catalyst. Therefore, regeneration techniques such as air oxidation and application H2, H2O and CO2 at elevated temperature are developed to remove carbonaceous deposits on spent catalyst and recover the activity of the catalyst [76]. For example, after oxidizing the carbonaceous deposit on nickel catalyst at 750\u00a0\u00b0C catalytic activity could be completely regained [77]. Regeneration techniques are also very promising for less active catalyst or catalyst more prone to deactivation by carbonaceous deposits as it allows using less expensive catalyst for tar removal [78]. Lind et al. applied air to a catalyst regeneration unit located next to the tar reformer and continuously replaced older catalyst, namely ilmenite (FeTiO3) with regenerated catalyst, thereby continuously removing carbonaceous deposits and maintaining a tar conversion of 35% [78]. Although catalyst regeneration is promising, the cyclic high temperature processing can for example lead to nickel sintering, phase transformations and volatilization [79] and recent studies show that regeneration is still an important topic of research [76].Tar represents one of the biggest bottlenecks in biomass conversion via gasification. Therefore, the goal of this study was to evaluate in a lab-scale reactor the ability of olivine, dolomite and a metal-based catalysts to reform acetic acid and its thermal decomposition products using simulated biosyngas as gas carrier. Upon heating biosyngas containing acetic acid above 400\u00a0\u00b0C, the tar is at least partially converted to higher molecular weight compounds and carbonaceous particles that will reach the reforming catalyst where they can affect the reactor performance. The tests performed assess the reforming ability of the catalysts under conditions representative of the high temperature gas cleaning unit designed by TU Delft to connect a 50\u00a0kW updraft gasifier with a solid oxide fuel cell in the EU funded project \u201cFlexiFuel-SOFC\u201d. The catalyst performance was evaluated assessing the outlet gas composition, the outlet tar concentration, and presence of carbonaceous deposits on the catalyst.Initially olivine appeared able to completely convert acetic acid into non-condensable gases. However, a considerable amount of carbonaceous deposits was found on the catalyst, and the catalytic activity of olivine decreased in time, as indicated by the increasing amounts of hydroxyacetone measured. Dolomite showed promising performances at 680\u2013750\u00a0\u00b0C since acetic acid was completely converted into non-condensable gases and only a minor amount of carbonaceous deposits was found on the low temperature part of the bed. The carbonaceous deposits accumulation was suppressed by increasing the operating temperature and the steam flow rate. However, operating dolomite at 900\u00a0\u00b0C, or pre-treating dolomite by calcination at 800\u00a0\u00b0C, causes the catalyst to become excessively brittle, which leads to powder entrainment in the gas flow, or even reactor clogging. The metal-based catalyst out-performed the naturally-occurring catalysts since it completely reformed acetic acid and suffered almost no carbonaceous deposits accumulation. The metal-based catalyst showed good catalytic activity towards the reverse water gas shift reaction and methane reforming, while both dolomite and olivine showed some minor catalytic activity towards the reverse water gas shift reaction, but neither dolomite nor olivine showed activity towards methane reforming. The results presented are expected to assist in the development of systems based on biomass gasification, such as biomass gasifier SOFC systems.This research was partially supported by the project \u201cFlexiFuel-SOFC\u201d. The project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No. 641229.The following is the Supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.biombioe.2021.105982.", "descript": "\n                  Tar compounds have been defined as Achilles\u2019 heel in biomass gasification. Catalytic reforming solves problems caused by tar by converting them into H2 and CO. Most of the research has focused on secondary and tertiary tar reforming while some information on primary tar can be derived from bio-oil reforming. However, these studies use humidified N2, Ar or He as gas carrier. Therefore, in this work, three catalysts are compared for reforming 40\u00a0g/Nm3 acetic acid as main primary tar compound from biomass updraft gasification using simulated biosyngas as gas carrier. The catalysts were tested over a 72\u00a0h period between 680 and 750\u00a0\u00b0C with a gas composition of 35.0\u00a0vol% H2O, 2.3\u00a0vol% CO, 19.5\u00a0vol% CO2, 3.6\u00a0vol% CH4, 24.0\u00a0vol% H2 and 15.6\u00a0vol% N2. Olivine completely converted acetic acid, but a considerable amount of carbonaceous deposits was found on the catalyst and the catalytic activity decreased over time with 0.2\u00a0g/Nm3 hydroxyacetone measured in the last day of testing. Dolomite showed promising performances by completely converting acetic acid and accumulating carbonaceous deposits only in the low temperature part of the catalyst bed. The carbonaceous deposits could be suppressed increasing the steam content to 50.1\u00a0vol% and the temperature to 900\u00a0\u00b0C. However, the catalyst became excessively brittle. The metal-based catalyst out-performed the naturally-occurring catalysts by completely converting acetic acid with almost no carbonaceous deposits accumulation. These results are expected to help the further development of tar reformers, and the commercialisation of biomass updraft gasifiers based systems.\n               "}
{"full_text": "Over the past few decades, heterogeneous catalysis in industrial processes has received more attention due to the many advantages related to green chemistry principles. The reusability and the ease of separation and saving catalytic process efficiencies are some of the advantages of the improved ecotechnological process in industrial applications; this includes sustainable energy-related processes such as biodiesel production (Ciriminna\u00a0et\u00a0al., 2021). In addition, biodiesel is potentially renewable energy that aligns with some of the requirements regarding green chemistry principles; these requirements include biodegradability, low toxicity, combustion efficiency, availability, and renewability (Ramos\u00a0et\u00a0al., 2019).The use of heterogeneous catalysis in biodiesel conversion is associated with the replacement of basic or acid homogeneous catalysts, particularly those with some basic/solid catalysts, and this brings about greater challenges from the sustainable development perspective (Yusuff\u00a0and Owolabi,\u00a02019). The high level of activity, reusability, recoverability, and low cost are some of the catalyst properties considered for the economical and sustainable process. The physicochemical properties\u2014such as surface area, surface basicity/acidity, pore distributions, chemical stability\u2014are among the influencing characters of the catalyst (Rossa\u00a0et\u00a0al., 2017). Some studies have revealed the effectiveness of Lewis acid catalysts in the transesterification mechanism; some transition metal salts suggest that the strength of the Lewis acid sites is Sn2+ >> Zn2+ > Al3+ (Einloft\u00a0et\u00a0al., 2008).Considering the high cost and the various steps required for the synthesis of some solid supports\u2014such as mesoporous silica (MCM-41) and synthetic zeolite (TUD-1) is of great importance when considering the cost of the entire process. As such, there is a need to develop easily recyclable heterogeneous catalysts with low-cost and sustainable materials. In this regard of adding economic reasonability, low-cost materials such as CaO-based materials, SiO2-based materials, and clay-based materials were widely developed (Fattah\u00a0et\u00a0al., 2020). Biogenic silica is one of the reasonable catalysts for this purpose, and this is due to its ease of production and modifiability (Mazaheri\u00a0et\u00a0al., 2021). Within this scheme, previous works revealed the use of biogenic silica from agricultural waste\u2014such as rice husk ash, bamboo leaf ash, and other cellulose-biomass sources\u2014as the support for ZnO, ZrO2, and SrO. From previous work, salacca leaf ash has been seen to contain high silica content (Triawan\u00a0et\u00a0al., 2021). As opposed to many other agricultural wastes such as bamboo and rice husk ashes, the biogenic silica from salacca leaves is presumably a good source to support the catalyst of biodiesel conversion (Adam\u00a0et\u00a0al., 2012).To the authors\u2019 knowledge, this is the first case study to utilize biogenic silica from the salacca leaf for the synthesis of ZnO/SiO2. The study on the effect of Zn content on the physicochemical characteristics of ZnO/SiO2 was performed using instrumental analysis. The use of salacca leaf ash for supporting biodiesel production will give the enhancement in clean production and sustainable production in term of renewable energy. Additionally, the optimum condition for biodiesel conversion application was statistically determined. Rice bran oil (RBO) was chosen as the biodiesel source due to its low-cost; RBO is a co-product and\u2014in countries such as Indonesia\u2014it is also a byproduct of rice milling (Trirahayu,\u00a02020). Since RBO is also a non-food oil, the use of RBO for biodiesel production is classified into the second generation of biodiesel; this means that its production will not compete with food consumption, and it can potentially be developed with improved economic value and sustainable production (Ju\u00a0and Vali,\u00a02005). The use of non-edible resources\u2014such as biomass feedstock\u2014eliminates the debate of food and energy scarcity, and it much less demanding on land used to provide feedstocks (Ramos\u00a0et\u00a0al., 2019). In addition, the convertible oil content in RBO is quite high, totaling approximately 15\u201325% wt. from rice bran (Einloft\u00a0et\u00a0al., 2008).Salacca leaves were collected from plants grown in a domestic cultivation area of the Sleman District, Yogyakarta Province, Indonesia. The salacca leaf ash (SLA) was used as the main source of silica, and it was obtained by calcining the dried salacca leaves at 600 \u00b0C for 1\u00a0hour in a Memmert muffle furnace (Germany). Rice bran oil was purchased from Oryza Grace (Jakarta, Indonesia). Chemicals consisted of zinc acetate dihydrate, NaOH, methanol, HCl, n-butylamine, and pyridine; these chemicals were obtained from Merck-Millipore (Darmstadt, Germany) and utilized without further purification.The extraction of silica from SLA was performed using a procedure similar to that of the previously published bamboo leaf ash silica extraction procedure (Fatimah\u00a0et\u00a0al., 2019b). The SLA was incorporated into a mixture with 4 M of NaOH solution, and it was subsequently refluxed using a round-bottom flask on an electric heating mantle for 4 hours. The filtrate from the reflux was slowly titrated with HCl 0.5 M until a pH of 8 was reached and the gel formed from the silica. The silica gel was separated from the filtrate by decantation, and the gel was neutralized using distilled water before it was dried in an oven at 60 \u00b0C.The composite of ZnO/SiO2\u00a0was synthesized using the obtained silica gel. The precursor solution of zinc acetate dihydrate was dissolved in water. It was then stirred into the silica gel to form a colloidal solution. The colloid obtained was then kept at 150 \u00b0C in a Teflon-lined autoclave overnight under hydrothermal conditions. The colloid was then dried in an oven at 80\u00b0C before being calcined at 500\u00b0C for 2 hours. To study the effect of zinc on the properties of the composite, the zinc content was varied at 20, 25, and 30\u00a0wt.%, and the obtained samples were encoded as ZnO/SiO2-20, ZnO/SiO2-25, and ZnO/SiO2-30, respectively. Fig.\u00a01\n represents the method for the ZnO/SiO2 synthesis.The ZnO/SiO2\u00a0nanocomposites were characterized by X-ray diffraction (XRD), gas sorption analysis, scanning electron microscopy/energy dispersive X-ray (SEM-EDX), Fourier transform infrared (FTIR) spectroscopy, transmission electron microscopy (TEM), and X-ray photoelectron spectroscopy (XPS). XRD analysis was performed using a Shimadzu X-6000 diffractometer. Measurements were carried out from 2\u03b8\u00a0=\u00a010\u00b0 to 80\u00b0 at a rate of 0.4\u00b0/min using Ni-filtered Cu K\u03b1\u00a0X-rays. For gas sorption analysis, a NOVA 1200 instrument was employed. For each analysis, the samples were degassed at 90\u00b0C for 4\u00a0hours beforehand. The Brunauer\u2013Emmett\u2013Teller (BET) specific surface area, pore volume, and pore radius were calculated from N2\u00a0adsorption/desorption data. The SEM-EDX analysis was performed using a JEM-ARM200F (JEOL) equipped with a 50-mm2\u00a0Si (Li) detector (JED-2300, JEOL) for EDX analysis. Additionally, the FTIR spectroscopy analysis was performed on the Perkin Elmer FTIR instrument (Singapore). TEM analysis was performed using the JEOL instrument. XPS analysis was performed on a V.G. Scientific ESKALAB MKII instrument using monochromatic Al\u00a0K\n\u03b1\u00a0radiation with a photon energy of 1486.6\u00a0\u00b1\u00a00.2\u00a0eV. The samples (0.2\u00a0mg) were gently pressed into a small pellet with a 15\u00a0mm diameter and mounted onto the sample holder. The sample was initially degassed for 4 hours prior to analysis in order to achieve a dynamic vacuum below 10\u22128 Pa.Surface acidity of the catalyst samples was determined using the n-butylamine titration method and the pyridine Fourier transform infrared (FTIR) analysis method. With the titration method, approximately 0.25 g of ZnO/SiO2 powder was mixed with 10 mL of n-butylamine 0.01 M followed by stirring overnight. The mixture was filtered, and the filtrate was titrated with HCl in order to measure the unadsorbed amount of n-butylamine. The acidity was determined by the amount of adsorbed n-butylamine (mmol) per gram of the 0.2 ZnO/SiO2 sample. The surface acidity measurement using the pyridine method was conducted by weighing the ZnO/SiO2 sample after it was dried at 100 \u00b0C for 2 hours; this was followed by placing the sample in a vacuum desiccator. The pyridine vapor fled into the desiccator overnight. After the desiccator was opened at room temperature, the pyridine-adsorbing sample was analyzed by using FTIR spectrophotometry analysis. The spectrum recorded from the analysis contains the peak of pyridine bound to the sample surface by both the Lewis acid-base interaction and the Bronsted acid-base interaction.All reactions were carried out in a batch reactor. For each reaction, 20 mL of RBO was mixed with methanol and 4 g of catalyst. The quantitative analysis of the produced biodiesel sample was performed on a Shimadzu GCMS instrument equipped with a spilt injector and a FID detector. The HP-5MS 5% (phenyl methyl siloxane) capillary column (30 m x 250 um i.d., 0.25 um film thickness) was employed as the separating phase, and helium was used as the gas carrier with an average velocity of 37 cm/sec at 200 \u00b0C. The analyses were set at the temperature of 100 \u00b0C, with a maximum increase of up to 370 \u00b0C at 15 \u00b0C/min, and held for 16 min. The varied reaction condition was set up in reference to the Box-Behnken Design (BBD) of the Response Surface Methodology (RSM). This was done in order to identify the affecting parameters and the optimum conditions of catalytic conversion. The varied parameters selected for optimization were as follows: Zn content in catalyst, catalyst dose, methanol-to-RBO volume ratio time, and time of reaction.The yield (%) was selected as the response for the reaction. The parameters were selected following optimization on a wider range of variables. In addition, temperature was not selected as a crucial factor due to the initial optimization; the results demonstrated that a higher yield was obtained as the temperature increased within the range of 60\u2013110 \u00b0C, and it reached a maximum at 100\u2013110 \u00b0C. Therefore, all experiments were conducted at 100 \u00b0C. The low, middle, and high levels for all of the independent variables are shown in Table\u00a01\n. Each variable was coded using a coding scheme to denote the level of the factor among three potential levels; -1 indicated the lower level, 0 indicated the medium level, and +1 indicated the higher level. Minitab 16 software (version 6.0) was utilized to carry out the regression analysis and to analyze the constructed data obtained from the preliminary experiments. Optimization via the evaluation of interactions among different parameters was thus performed.Reusability of the catalyst was tested by recycling the spent catalyst and utilizing it for further cycles. The recycling procedure was performed by filtering out the solids, washing them in methanol under stirring the mixture for 1 hour, and drying them in an oven at 100 \u00b0C overnight.The XRD pattern of the materials are presented in Fig.\u00a02\n. The SiO2 sample shows a broad reflection ranging between 22\u201325o. This is characteristic of amorphous silica. The pattern is similar with silica extracted from rice husk ash, bamboo leaf ash, and other biogenic sources. Moreover, after modification with ZnO, the ZnO signal in the composite was observed (Aneesh\u00a0et\u00a0al., 2007). The peaks at 2\u03b8\u00a0= 31.64\u00b0, 34.45\u00b0, 36.23\u00b0, 47.61\u00b0\u2014which are indexed as (100), (002), (101), and (102) planes,\u00a0respectively\u2014are associated with JCPDS card no. 29-1487 (Garcia-Sotelo\u00a0et\u00a0al., 2019). In regard to the ZnO phase, the refinement of the pattern was obtained using the Rietica software. As can be seen from the ZnO/SiO2 reflection, the broad peak associated with amorphous SiO2 disappeared along with the formation of the crystalline ZnO. This pattern is similar to that reported in other ZnO/SiO2 synthesis utilizing other silica precursors (Galedari\u00a0et\u00a0al., 2017). From the varied Zn content ranging from 20-30 %wt. and using the Scherrer's equation (Eq.\u00a01):\n\n(1)\n\n\nD\n=\n\n\n0.9\n\n\n\u03b2\nc\no\ns\n\u03b8\n\n\n\n\n\nwhere \u03bb is a wave length of the X-ray (0.154186 nm) and \u03b2 is the measured FWHM, it can be seen that the Zn content contributed insignificantly to the crystallite size, as listed in Table\u00a02\n.The higher Zn content produced a higher crystallite size, and this is similar to the trend found from the synthesis of ZnO/SiO2 using bamboo leaf ash as the silica precursor (Fatimah\u00a0et\u00a0al., 2019). The higher Zn concentration influenced the precipitation rate, which affects the growth of the crystallite size. The higher precursor concentration also accelerates the sol-gel reaction as well as the distribution of ZnO. It is influenced by factors such as the precipitation rate, interaction between the Zn precursor and the support, and the calcination temperature (Liu\u00a0et\u00a0al., 2014). On the basis of the crystallite sizes obtained from the XRD data, the growth of ZnO particles is influenced by the amount of the Zn precursor, which also affects the ZnO distribution in the composite. The formation of ZnO particles also affected the surface characters of the specific surface area, pore volume, and pore radius, as listed in Table\u00a03\n. As seen in Table\u00a03, the formation of ZnO particles lead to an increase in the specific surface area of the material. The SiO2 specific surface area is 45.2 m2/g, and it increased to 80.3; 82.8, and 87.1 m2/g for ZnO/SiO2-20; ZnO/SiO2-25, and ZnO/SiO2-30, respectively. The trend of the increased specific surface area along with the ZnO formation is similar with previous reports of synthesis of composites using bamboo leaf ash as well as the synthesis using tetraethyl ortho silicate (Somoghi\u00a0et\u00a0al., 2021). However, compared to the composites produced by using the bamboo leaf ash, the specific surface area values in this study are lower than the comparable ZnO content. The enhanced specific surface area with the presence of ZnO on the surface also affected the surface acidity, and this is exhibited by the data in Table\u00a03. The surface acidity represents the capability of the surface to interact with the base, n-butylamine, in the case of this analytical method. Moreover, the acid sites of the surface play a role in conducting surface interaction with both the triglyceride and the alcohol within the transesterification mechanism, as schematically represented in Fig.\u00a03\n.The presence of ZnO on the catalyst surface reflects the Lewis acid site contributing to the adsorption site of the fatty acid via the Lewis acid-base interaction; this occurs with \u03c0-bonding of the double bond of the carbonyl functional group (Abedin\u00a0et\u00a0al., 2020). The chemisorption produced surface-bound carbocation that was further attacked by hydroxyl from methanol as a nucleophile. The methyl shift, proton migration, and rearrangement breakdown of bonding between the reactant and the Zn surface was followed by the desorption of methyl ester and alcohol/glycerol as product (Boonyuen\u00a0et\u00a0al., 2018).The presence of surface acidity is also identified by FTIR analysis of the pyridine-adsorbed samples (Fig.\u00a04\n).From previous literature on surface acidity identification using pyridine as a probe molecule, the interaction of pyridine with the surface acid sites is identified by the Bronsted acid and hydrogen-bonded pyridine at around 1540\u20131600 cm\u22121; with Bronsted and the Lewis acid-bound pyridine at 1450\u20131490 cm\u22121, while the Bronsted acid-bound pyridinium cation did so at around 1540 cm\u22121 and 1640 cm\u22121 (Reddy\u00a0et\u00a0al., 2009; Ali-Dahmane\u00a0et\u00a0al., 2019). From the measurement, adsorbed pyridine is represented by peaks at 3067 cm\u22121 and 2925 cm\u22121 corresponding to the presence of C-H stretching from pyridine structure. The pyridine-coordinated through the aromatic \u03c0 electrons at around 1607 cm\u22121, while the peaks at 1465 cm\u22121 corresponds to the Lewis acid and Bronsted acid sites, respectively (Larina\u00a0et\u00a0al., 2019; Pham\u00a0et\u00a0al., 2021). Furthermore, on comparison of the intensity of the absorption, the Lewis-to-Bronsted acid ratio (L/B) was calculated using the following equation (Eq.\u00a02):\n\n(2)\n\n\n\nL\nB\n\n=\n\n\nI\n\n1450\n\u2212\n1459\n\n\n\nI\n\n1540\n\u2212\n1560\n\n\n\n\n\n\nwhere I1450\u20131455 is the intensity of the band at 1450-1455 cm\u20131 and I1540\u20131560 is intensity of the band at 1540-1560 cm\u20131. The L/B values for each sample can be seen in Table\u00a03. The increasing Zn content represented the increasing L/B, which indicated that the Lewis acid-base interaction between the surface and the reactant is predominantly created by the increasing amount of ZnO on the surface (Abedin\u00a0et\u00a0al., 2020).Aside from the peaks representing the surface acidity, some peaks correspond to the presence of ZnO and SiO2 in the composite, such as the absorption spectrum seen at 501 cm\u22121, 803 cm\u22121, 1105 cm\u22121, and 3445 cm\u22121. The band at 501 cm\u22121 comes as a result of the stretching of the Zn-O. Additionally, the intense peaks at 803 cm\u22121 and 1105 cm\u22121 can be attributed to the symmetric stretching and the asymmetric stretching vibrational mode of the Si-O-Si, respectively. The band at 3445 cm\u22121 is due to the O-H bending modes of the hydroxyl groups and the adsorbed water in the powder (Galedari\u00a0et\u00a0al., 2017).The evolution of the composite surface can be seen by the change of specific surface area, and it has also been identified by SEM and TEM analysis. As shown in Fig.\u00a05\n, flower-like morphologies are shown by the ZnO/SiO2 samples, with rod-like forms appearing at the higher magnification. The morphologies are similar to that of ZnO-based composites reported by various methods, and this includes the ZnO/SiO2 prepared using bamboo leaf ash. With varied Zn content, it can be seen that the higher Zn content leads to the formation of larger clods. The pattern suggests the crystal growth in the hydrothermal process as the growth of ZnO to Zn(OH)4\n2\u2212; this occurred as a result of the systematic pattern of ZnO growth influenced by Zn as crystal nuclei (Shi\u00a0et\u00a0al., 2013). The TEM profiles depicted in Fig.\u00a06\n are in line with the theoretical approach of crystal growth; this is influenced by the Zn content, as the larger particle size is represented at the higher Zn content. From the HTEM image in Fig.\u00a06b, it can be seen that the ZnO particle size is approximately 20\u201330 nm; this is in line with the calculated crystallite size from the XRD measurement. Moreover, the HRTEM images exhibit the lattice fringes of the attached ZnO nanoparticles, and the estimated d-spacing value between the two adjacent of the fringes was found to be 2.49 \u00c5; this corresponds to the (101) planes of hexagonal wurtzite ZnO (Fig.\u00a06e).\nTable\u00a04\n presents the average composition of fatty acid methyl ester (FAME) produced from the processes, and one of the GCMS analysis results is presented in Fig.\u00a07\n. It can be seen that the dominant components are the methyl esters from the palmitic acid, oleic acid, linoleic acid, linolenic acid. The stearic acid with the palmitic acid and the stearic acid methyl ester are the minor components. The composition is similar to that found in previous investigations on RBO transesterification with reference to the fatty acid composition (Einloft\u00a0et\u00a0al., 2008; Zaidel\u00a0et\u00a0al., 2019).The optimization of catalytic activity was conducted based on RSM using BBD for selected influential parameters. With reference to the catalytic mechanism, the parameters of the ZnO content, catalyst dosage, methanol-to-oil ratio, and time of reaction were chosen. Table\u00a05\n lists the yield values at the varied conditions. Table\u00a05. The yield values at the varied conditions.The analysis of variance (ANOVA) results of the response surface model are shown in Table\u00a06\n. The statistical parameters demonstrate that the second-order regression model is significant at the 95% confidence level with the predictability of the model.The determination coefficient, or R2, of 0.9891 indicates that the model could be used as a predictor of the response. In addition, a relatively high value of the coefficient of variation demonstrated the high reliability of the experiments. Based on the multiple regression analysis, the following equation is the model fit for prediction of the experimental results:\n\n\n\nYield\n\n(\n%\n)\n\n\n=\n\n\u2212\n133.740\n\n+\n\n4.373\n\nx\n1\n\n+\n\n14.470\n\nx\n2\n\n+\n\n7.320\n\nx\n3\n\n+\n\n0.054\n\n\n\nx\n1\n\n\n2\n\n\u2212\n\u2212\n\n2.089\n\n\n\nx\n2\n\n\n2\n\n+\n\n1.515\n\n\n\nx\n3\n\n\n2\n\n\u2212\n5.435\n\n\n\nx\n4\n\n\n2\n\n+\n\n0.175\n\nx\n1\n\n.\n\nx\n2\n\n\u2212\n1.127\n\nx\n1\n\n.\n\nx\n3\n\n\u2212\n0.986\n\nx\n1\n\n.\n\nx\n4\n\n+\n\n1.627\n\nx\n2\n\n.\n\nx\n3\n\n\u2212\n0.570\n\nx\n2\n\n.\n\nx\n4\n\n+\n\n0.788\n\nx\n3\n\n.\n\nx\n4\n\n\n\n\n\nThe predicted responses were in good agreement with the experimentally obtained response (R2\u00a0=\u00a00.972 and Adj R2\u00a0=\u00a00.936). Furthermore, the model's F-value of 27.94 and the P-value of 0.00 implies that the model is significant. From the P-values of the effect of varied parameters, it can be concluded that\u2014with the exception of the catalyst dose\u2014other parameters of ZnO content, such as the methanol-to-oil ratio and time of reaction, influence the conversion significantly. It can be confirmed that, at the range of catalyst doses, the effect of the surface provided for the surface mechanism is similar. However, the effect ZnO content is significant. Considering that the ZnO content is in line with the provided surface acidity and the catalyst dose is related with the specific surface area of the catalyst, it can be concluded that surface acidity is more dominant in governing the conversion.The plot in Fig.\u00a08\n was created to ensure the adequacy of the proposed regression model; an R2 of 0.964 was obtained, representing the fitness of the model. The contour plot in Fig.\u00a09\na and 9b demonstrate that the higher value of conversion can be achieved by a high catalyst dose and time of reaction. The plots and the model concluded that the optimum condition for producing biodiesel is a catalyst dose of 6 g/L, a time of reaction of 3 h, and methanol-to-oil ratio of 6. Zn content within the range of 20\u201330 % will give a similar yield, as it does not significantly influence the conversion.The contour plots represent that the higher conversion favored an increase of ZnO content and time of reaction; the conversion, however, reached an optimum level at the methanol-to-oil ratio of 6:1 (Fig.\u00a08c). This indicates the excess methanol tends to reduce the surface equilibrium for the transesterification; this subsequently hinders the completion of FFA and triglycerides being protonated at the active sites, leading to a lower FAME yield (Zhang\u00a0et\u00a0al., 2013) . A similar effect of the methanol-to-oil ratio was reported by previous work (Anwar\u00a0et\u00a0al., 2018). However, the Zn content generally has no significant effect on the yield at along the methanol-to-oil ratio. In general, the activity of the catalysts resulting from this research was comparable with other silica-based catalysts used for biodiesel conversion. In previous reported works, the CaO/SiO2 catalyst gave yields of 98.5% and 85.6% (Moradi\u00a0et\u00a0al., 2014). With ZrO2/SiO2, a yield of 96.2% was obtained from the oil-methanol-catalyst ratio of 100:400:2.4 for 3 hours (Faria\u00a0et\u00a0al., 2009). Similarly, the yield ranging from 60\u201380% was reported with the use of alkali earth metal oxides (CaO, MgO, and BaO) supported by SiO2 (Mohadesi\u00a0et\u00a0al., 2014).Catalyst reusability is an important characteristic to consider for application purposes. The catalyst reusability was identified by the change of yield on reaction cycles. The catalyst was recycled by being washed in ethanol and water and dried at 100 \u00b0C. As shown in Fig.\u00a010\n, the catalysts showed good reusability until the third cycle, as the maintained yield at the reduced value was no more than 10% from the fresh catalyst; the yield decreased significantly during further cycles (fourth and fifth cycles).This suggested that the composite could not accommodate the surface reaction in the mechanism due to the loss of catalytic properties. XRD analysis was conducted to evaluate the structural change of the catalyst. The XRD pattern is presented in Fig.\u00a011\n.From the diffraction pattern, it can be seen that there are new phases identified and reflected by the change of ZnO into Zn(OH)2 as well as the presence of silica in the cristobalite and the quartz. The orthorhombic Zn(OH)2 (O) peaks are observed by strong peaks at 32.93\u00b0, 42.8\u00b0, and 49.5\u00b0 corresponding to the (100), (201), and (102), respectively (Ghaedi\u00a0et\u00a0al., 2016; Mousavi-Kamazani\u00a0et\u00a0al., 2020; Wang\u00a0et\u00a0al., 2013). The adsorbed hydrophobic reactant on the catalyst surface also reduced the catalyst porosity. Further investigation on this was performed using catalytic refreshment; this was conducted by calcining the used catalyst at the temperature of 400 \u00b0C for 2 hours. The comparison of specific surface area and the pore distribution of the refreshed catalyst are presented in Fig.\u00a012\n. It can be seen that the refreshment successfully exhibited the remaining catalytic activity, and this was comparable to that of the fresh catalyst. The data suggests that the catalyst is recyclable, and it can be recycled using a simple method. This suggests that the material is potentially a developed, low-cost catalyst from a sustainable resource. The production of ZnO/SiO2 catalyst by using salacca leaf ash for biodiesel production represented the potential usage of waste into valuable material in renewable energy.The results of the present studies show that ZnO/SiO2 was successfully prepared using silica derived from salacca leaf ash. The samples exhibit efficient catalytic activity for biodiesel conversion from rice bran oil. The use of the response surface methodology\u2014which is based on the Box-Behnken design\u2014was applied for the optimization of the following Zn content parameters: catalyst dose, methanol-to-oil ratio, and the time of reaction. With the exception of Zn content, the statistical parameters imply the strong influence of the tested parameters. The catalyst is recyclable, and it can be recycled using the simple procedure of heating at 400 \u00b0C for 2 hours. This suggests that the material is potentially developed for biodiesel conversion.The authors declare that they have no known competing financial interests or personal relationships that cdocould have appeared to influence the work reported in this paper.The authors would like to express appreciation for the support from the Ministry of Ministry of Education, Culture, Research, and Technology through the World Class Professor Program in 2021.", "descript": "\n                  This paper reports on the first case of the recyclable catalysts of ZnO/SiO2 prepared from salacca leaf ash as a source of silica for biodiesel production. The catalysts were prepared using a hydrothermal synthesis method, and the catalyst was utilized for biodiesel conversion from rice bran oil. Physicochemical properties of the catalysts were studied by multiple instrumental analyses consisting of XRD, SEM, TEM, and surface acidity measurements on pyridine adsorption followed by FTIR analysis. The study focuses on the effect of Zn content on the physicochemical character. As such, a varied Zn content of 20, 25, and 30 % wt. was applied. In order to evaluate the influencing parameters for the catalytic process, a response surface methodology based on the Box-Behnken design was applied in optimization. The selected parameters of catalysis included the Zn content, catalyst dose, methanol and oil ratio, and the time of reaction. It was concluded that all tested parameters\u2014with the exception of Zn content\u2014significantly influence the yield of the reaction. The catalyst demonstrated a reusable feature, as there was an insignificant yield value of catalytic activity until the fifth cycle during the simple procedure of recycling. This suggests that the material was potentially developed for biodiesel conversion.\n               "}
{"full_text": "Data will be made available on request.It is widely accepted that hydrogen, a clean and sustainable energy vector, is considered the ideal alternative of fossil fuels. Within the actual framework of the socioeconomical and environmental situation, green hydrogen will play a crucial role for a clean energy transition process [1]. Since many years, photocatalytic water splitting towards hydrogen evolution has attracted numerous attentions because of its green sustainability [2,3]. In spite of the huge efforts to generate hydrogen via powder-based solar water-splitting systems to date have unfortunately fallen short of the efficiency values required for large scale plants [4,5]. The challenge is to develop stable and efficient catalysts that can harvest solar light, using co-catalyst alternative to precious materials that would allow the scale-up and practical applications [6,7]. In this sense, promising studies at pilot plant scale have been recently reported [8\u201311].Within this frame, green H2 production from alcohol photocatalytic reforming reaction appears as a hot topic in the field of photocatalysis [12,13]. From the enormous literature on this field, TiO2 based systems have provided the best performances [14]. However, some factors such as rapid recombination, the occurrence of backward reaction or even the deactivation by the formed intermediates hindered the development of H2 production at large scale [15]. The use of metal co-catalysts, used as charge trapping sites, have demonstrated necessary in order to enhance the efficiency of the photocatalytic reaction by avoiding the electron\u2013hole recombination processes [16\u201318]. It has been found that the addition of these noble metals could have different effects on the photocatalytic activity which is also strongly affected not only by the nature of metal but also by other parameters from sample history and metal features [19]. Alternatively to noble metals, copper-based catalysts have been extensively considered as a cheaper option [13,20\u201323].In addition to this traditional strategy, the combination of heat and light has been recently suggested as a novel approach pursuing the improving efficiency of the photocatalytic process [24\u201326]. Thus, by the combination of classical thermo- and photo-catalytic processes, interesting synergistic effects have been reported that drastically enhances the hydrogen production through photoreforming reaction [25].Due to the mechanistic complexity of this multicatalytic approach, which would involve different simultaneous reactions, the origin of this cooperative effect is still unclear. In principle, the synergistic improvement would be conditioned by the primary importance of single thermo- and photo- contributions on the kinetic behavior and reaction mechanism [27,28].Moreover, most of the reported results do not cover the influence of the co-catalyst loading on such complex mechanism and the catalytic performance [27,28]. In the present paper we present an interesting study on methanol reforming through a multicatalytic approach, by combining heating and photonic excitation. For this scope, we use a Cu/TiO2 system with different Cu loading that could serve to explore how the mechanism is influenced by metal content and therefore condition the overall reaction performance.TiO2 (Evonik P90) was used as photoactive support (Table 1 in Supporting Information). TiO2 Evonik P90 exhibits higher surface area than widely used Evonik P25. Since the aim of this study is to see the effect of the amount of Cu loading on the catalytic activity, we decided to use this high surface area anatase to avoid the effect of metal agglomeration at increasing contents. Metal loading was performed through chemical reduction deposition procedure using copper nitrate as metal source. In this case, 0.5\u00a0g of support was suspended in 100\u00a0mL of water containing the stoichiometric amount of Cu precursor for a nominal loading between 1.0 and 5.0\u00a0wt%. Chemical reduction of metal precursor was achieved by using NaBH4 as reducing agent. After adding the reducing agent at certain excess (1:20 molar ratio with respect to Cu), the suspension was stirred at room temperature for 30\u00a0min. The effectiveness of the deposition was always checked by adding more NaBH4 to the filtrated liquid. We did not appreciate in any case any turbidity or color changes due to the presence of Cu. The obtained systems were filtered, thoroughly washed with distilled water and finally dried at 90 \u00baC. As prepared catalysts were labeled as CunP90, being n the nominal metal loading.BET surface area measurements were carried out by N2 adsorption using a Micromeritics 2000 instrument.X-ray diffraction (XRD) patterns were obtained using a Siemens D-501 diffractometer with Ni filter and graphite monochromator, using the Cu K\u03b1 radiation as X-ray source., We have calculated the anatase fraction as well the mean crystallite size by Rietveld fitting using HighScore-Plus software, using the line broadening of corresponding diffraction peaks.Diffuse reflectance UV\u2013vis spectroscopy was performed using a Cary 300 instrument. Spectra were recorded in the diffuse reflectance mode using Spectralon\u00ae as white standard. Scans range was 240\u2013800\u00a0nm.XPS data were recorded on pellets using a customized system incorporating a hemispherical analyser (SPECS Phoibos 100), a non-monochromatized X-ray source (Al K\u03b1; 1486.6\u00a0eV, Mg K\u03b1, 1253.6\u00a0eV). The analyser was operated at a fixed transmission and 50\u00a0eV pass energy with an energy step of 0.1\u00a0eV. Binding energies were calibrated using C 1\u00a0s (284.6\u00a0eV) as an internal reference. Prior to each analysis, the samples were evacuated to 10\u22129 mbar at room temperature. In a typical experiment, the sample was initially placed in the sample holder and transferred to the spectrometer chamber where XPS spectra were acquired.\nIn-situ FTIR studies were performed in a Harricks praying mantis cell. The spectra were recorded on a Nicolet FT-IR spectrometer at reaction conditions with a resolution of 4\u00a0cm\u22121. Before reaction, catalysts were pretreated in N2 at 50 \u00baC for 30\u00a0minTemperature programmed reduction (H2-TPR) was performed using a Chemstar instrument (Quantachrome). About 30\u00a0mg of the catalysts previously degassed at 50\u00a0\u00b0C for 30\u00a0min under Ar flow. The TPR spectra were collected in 10\u00a0mL/min mixture of 5% H2/Ar from 50 \u00baC to 700\u00a0\u00b0C with a heating rate of 10\u00a0\u00b0C/min.Transmission electron microscopy were performed by using a FEI Tecnai F30 microscope in STEM mode operated at 300\u00a0kV equipped with a Gatan GIF Quantum 963 energy filter. The samples were directly dropped on a nickel grid.Gas phase photocatalytic H2 production tests were performed in a flow cell (\nFig. 1). The powder photocatalysts (50\u00a0mg) were placed in the cell and then degassed with N2 at 50 \u00baC. Methanol steam flow (CH3OH = 20% v/v) at 15\u00a0mL/min was passed through the sample for 60\u00a0min before reaction. After which, the lamp (200\u00a0W lamp housing) was switched on and the sample was illuminated through the quartz top-window of the cell. Effluent gases were analysed to quantify H2 production by gas chromatography (Agilent microGC) using a thermal conductivity detector. Only CO and CO2 were detected as side products. CO selectivity (S\n\nCO\n) were calculated by considering all products detected:\n\n\n\n\n\nS\n\n\nCO\n\n\n\n\n\n%\n\n\n\n=\n\n\n\n\nCO\n\n\n\n\n(\n\n\n\n\n\nH\n\n\n2\n\n\n\n\n\n+\n\n\n\nCO\n\n\n\n+\n\n\n\n\n\nCO\n\n\n2\n\n\n\n\n\n)\n\n\n\u00d7\n100\n\n\n\n\nThermo-photocatalytic runs were performed at 200 \u00baC. As we have evidenced in a previous study, at this temperature the synergistic effect for thermo-photocatalytic process is maximum [29]. As reference, thermo-catalytic runs were performed in the absence of light under similar reaction conditions.The apparent quantum efficiency for the H2 evolution reaction has been determined from the reaction rate and the flux of incoming photons (calculated for the irradiation wavelengths of 365\u2009nm) [20,30].\n\n\n\nAQE\n=\n\n\nrate\n\n\n\n\nmol\n\u00b7\n\n\ns\n\n\n\u2212\n1\n\n\n\n\n\n\n\nrate\n\nof\n\nincident\n\nphotons\n\n\n\n\nmol\n\u00b7\n\n\ns\n\n\n\u2212\n1\n\n\n\n\n\n\n\n\n\n\n\nThe synergetic effect of the thermal contribution on thermo-photocatalytic process was evaluated by considering the difference rate with single process rates [31].\n\n\n\n\n\nrate\n\n\nsyn\n\n\n=\n\n\nrate\n\n\nthermo\n\u2212\nphoto\n\n\n\u2212\n\n\n\n\n\nrate\n\n\nphoto\n\n\n+\n\n\nrate\n\n\nthermo\n\n\n\n\n\n\n\n\n\nFrom this rate\n\nsyn\n value we have calculated the synergetic efficiency (AQE\n\nsyn\n) which specifically accounts the efficiency attributable to the synergy of both thermo- and photo-catalytic processes.Copper deposition over TiO2 by chemical reduction leads to well dispersed Cu nanoparticles over the range of studied systems. As it can be observed from STEM images (\nFig. 2), homogeneous dispersion of Cu nanoparticles of ca. 2\u2009nm is obtained. Moreover, no aggregation of Cu clusters is noticed as metal loading increases.From diffuse reflectance spectra (\nFig. 3), important absorption bands can be found at 350\u2013450\u2009nm and 750\u2009nm\u00a0that can be attributed to Cu clusters at the surface. These new bands progressively increase as copper loading increases, specially bands at 410\u2009nm and 750\u2009nm. Thus, bands within the range 350\u2013450\u2009nm have been ascribed to (Cu\u2013O\u2013Cu)2+ clusters in a highly dispersed state as well as to three-dimensional Cu1+ clusters. [32,33] On the other hand, the band at 750\u2013850\u2009nm has been assigned to \n2\n\nE\n\ng\n \u2192 \n2\n\nT\n\n2\u2009g\n transitions of Cu2+ located in the distorted or perfect octahedral symmetry. [34] From these results, we could expect that Cu species would consist on highly dispersed CuO and surface doped Cu2+/1+. Both bands progress similarly as metal content increases (Fig. 3\n.b).Reduction profile also denotes the evolution of copper clusters as loading increases (\nFig. 4). In all cases a single TPR peak ca. 180 \u00baC is found for all catalysts. This is a rather lower reduction temperature compared with values reported in the literature, and would be attributed to highly dispersed CuO species at the nanoscale. [35\u201337] Moreover, from the deconvolution of TPR peaks it is possible to envisage four Cu species, denoted as \u03b1 to \u03b4. The presence of different reduction peaks in the TPR profile of oxide-supported Cu samples has been extensively reported. [38,39] Thus, it is accepted that large and crystallized CuO particles would be reduced at higher temperature than smaller CuO clusters or highly dispersed Cu2+ species in strong interaction with the oxide support. [40,41] In our case, the reduction temperature around 180 \u00baC clearly points out the absence of bulk CuO, as was stated above from TEM images. For all samples, species reduced at slightly higher temperature (\u03b3\u2009+\u2009\u03b4) are prominent. However, as metal loading increases, low reduction temperature species (\u03b1\u2009+\u2009\u03b2) become more evident (Fig. 4\n.b). Such low reduction temperature species would be associated to finely disperse copper species as support Cu2+ doped and would be hardly detected from TEM analysis. So, only small CuO nanoclusters are noticed. Therefore, for higher Cu content catalyst, Cu species distribution seems to be composed equally by small finely disperse CuO nanoclusters and surface doped Cu2+ species. Quantitative analysis from H2 consumption clearly denotes the efficiency of deposition method. The amount of reduced Cu follows a clear linear evolution being in all cases close to the nominal values (Fig. 4.c).From XPS analysis, the observed broad peaks as well as the presence of a shake-up satellite would denote the coexistence of Cu2+ and Cu+/Cu species (\nFig. 5\n.a). Thus, peaks at 932.0\u2009eV and 951.9\u2009eV would correspond to Cu 2p\n\n3/2\n and Cu 2p\n\n1/2\n spin-orbit split of Cu/Cu2O species. [42] On the other hand, a second contribution at 934.2\u2009eV and 954.1\u2009eV with satellite shakeup at ca. 942\u2009eV, in all the synthesized products should correspond to CuO species. It is also worthy to mention that for lower Cu contents, partially reduced fraction seems to be slightly higher (\nTable 1). This fact could be explaining considering a higher dispersion in these samples. Such low dimensional CuO clusters should be easily reduced under XPS conditions.In agreement with TPR analysis, the surface Cu/Ti ratios calculated from XPS signals also follows a linear progression with nominal values, indicating the homogeneous distribution of Cu species over TiO2 surface even at large metal loadings (Fig. 5.b). The Cu/Ti ratios obtained for samples after thermo-catalytic or thermo-photocatalytic runs at 200 \u00baC do not show a significant variation from fresh sample (Table 1). Therefore, no aggregation should be expected after reaction by effect of temperature.Moreover, as Cu content increases the ratio of oxidized Cu seems to be slightly increased (Table 1), similarly as \u03b1\u2009+\u2009\u03b2 family evolution from TPR experiment. This can be also observed from satellite relative intensity (I\n\nsat\n\n/I\n\nCu2p3/2\n), showing the highest value for Cu4P90 catalyst. Indeed, \u03b1\u2009+\u2009\u03b2 family has been previously associated to small CuO nanoclusters. The higher partial reduced fraction in low Cu content catalysts could therefore be ascribed to highly disperse Cu2+ entities (\u03b3\u2009+\u2009\u03b4 species) that would easily reduce under XPS conditions.\n\nFig. 6 shows the H2 production rates and apparent efficiencies and H2 yields of Cu/TiO2 P90 systems. Thus, the obtained rates for photo-, thermo-, and thermo-photocatalytic runs under UV\u2013vis light illumination for 4\u2009h are summarized in \nTable 2. As it can be observed, the performance for thermo-photocatalytic H2 production are notably higher in all cases with respect to single thermo- and photocatalytic processes (Fig. S1).The apparent quantum efficiencies denote that for photocatalytic performance the maximum value is attained for Cu2P90 catalyst, showing an H2 production of 12\u2009mmol/h\u00b7g after 4\u2009h (Fig. 6\n.a). A higher Cu loading negatively affects to the photoreforming activity. In principle, the higher photoactivity exhibited by Cu2P90 would be associated to the finely dispersion and stronger interaction with support observed from TPR. [43] As metal loading increases, recombination processes would hinder the photocatalytic performance.It can be also stated that thermo-catalytic process also leads to significant production of H2. As we already discussed, H2 formation under these conditions could be explained considering methanol thermal decomposition (Eq. i) [44,45].\n\nEq. i\n\n\nC\n\n\nH\n\n\n3\n\n\n\nOH\n\n\n\n\n\u2192\n\n\n\n\n\n\n2\n\n\n\nH\n\n\n2\n\n\n\n+\n\n\nCO\n\n\u2062\n\n\n\nmethanol\n\n\n\ndecomposition\n\n\n\n\n\nIn this case, AQE values lineally increases in the whole range of Cu content, reaching an H2 yield of ca 30\u2009mmol/h\u00b7g for Cu5P90 catalyst (Fig. 6.a). Therefore, in this case and within the range of metal content studied, thermo-catalytic activity is clearly favoured by the increasing surface Cu over TiO2.Finally, the combined thermo-photoreforming leads to notable values of efficiency, with a maximum yield of ca. 60\u2009mmol/h\u00b7g also for Cu5P90 catalyst. However, in this case the photocatalytic performance of Cu5P90 is close to that exhibited by Cu4P90 denoting certain exhausting effect in the thermo-photocatalytic activity with increasing metal loading. Thus, at this Cu loading the optimum metal content could be considered. (Fig. 6.a).Since calculated rate\n\nsyn\n always show positive values, the combined process cannot be explained as the sum of both single ones, and a certain synergistic effect can be envisaged for all catalysts. Moreover, the calculated AQE\n\nsyn\n clearly denote that the combination of thermo- and photo- processes notably affects to the overall reaction performance and appear clearly affected by metal loading (Fig. 6.b).In this sense, it is also interesting to highlight the evolution of the synergy with Cu loading (Fig. 6.b). For the lowest Cu content, since the photocatalytic performance is rather low, thermal contribution appears important. Minimum AQE\n\nsyn\n is found for Cu2P90, for which the photocatalytic performance is maximum. Thermal-photocatalytic synergy reaches a minimum for Cu2P90 catalyst, for which photocatalytic performance is maximum and thermo-catalytic activity is still low. Catalysts with metal contents higher than 2 showed outstanding thermo-photocatalytic AQE that progressively increases with Cu content. However, though thermal reforming lineally grows with Cu, a maximum in the AQE\n\nsyn\n is found for Cu4P90 catalysts. This fact would point out a process that would hinder the thermo-photocatalytic performance for Cu5P90.In order to understand the evolution showed by AQE and AQE\n\nsyn\n in different studied photoreforming processes, we followed CO and CO2 as interesting intermediates from side reactions that would simultaneously take place (\nFig. 7). Thus, CO can be derived from side reactions that would proceed simultaneously with photoreforming; that is, methanol decomposition as well as formic acid dehydration (Eqs. i and ii) [45].\n\nEq. ii\n\n\n\nHCOOH\n\n\n\n\n\u2192\n\n\n\n\n\n\n\n\nH\n\n\n2\n\n\nO\n\n+\n\n\nCO\n\n\n\nformic\n\n\n\nacid\n\n\n\ndehydration\n\n\n\n\n\nFrom the evolution of CO selectivity, the single photocatalytic process clearly denotes a concomitant formation of CO. The observed CO formation progressively decreases as Cu loading increases. For the thermocatalytic process, low amounts of copper do not lead to CO formation, starting at Cu loading higher than 2\u2009wt%. It is worthy to notice that the combine thermo-photocatalytic process shows a similar trend as photocatalytic one, but with much lower CO selectivity. Thus, for higher Cu loading S\n\nCO\n values are similar to that obtained by thermocatalysis.The marked lower S\n\nCO\n values attained for thermo- and thermo-photocatalytic processes would denote an important CO consumption (probably associated with a higher H2 yield).Thus, two probable processes could be involved to explain the lower CO formation and higher H2 production. On one hand, formic acid dehydrogenation as alternative way to decompose formic acid instead of dehydration; and water gas shift reactions (Eqs. iii and iv). Indeed, the occurrence of water-gas-shift reaction has been previously argued within H2 photoreforming reaction [29,44].\n\nEq. iii\n\n\n\nHCOOH\n\n\n\n\n\u2192\n\n\n\n\n\n\n\n\nH\n\n\n2\n\n\n\n+\n\nC\n\n\nO\n\n\n2\n\n\n\n\ndehydrogenation\n\n\n\n\n\n\n\nEq. iv\n\n\n\nCO\n\n\n+\n\n\n\nH\n\n\n2\n\n\nO\n\n\n\n\u2192\n\n\n\n\n\n\n\n\nH\n\n\n2\n\n\n\n+\n\nC\n\n\nO\n\n\n2\n\n\n\n\nwater\n\n\n\ngas\n\n\n\nshift\n\n\n(\n\nWGS\n\n)\n\n\n\n\nWe may also point out that the presence of CO during photoreforming reaction would probably affect to the photocatalytic activity during the reaction time by CO poisoning of the metal active sites [46,47].In \nFig. 8 we depict the FTIR spectra during in-situ photoreforming reaction. From these spectra, two important regions can be envisaged. In the higher wavenumber region, C\u2013H stretching modes can be found. Thus, C\u2013H contributions at 2920/2820\u2009cm\u22121 have been associated to metoxy species [45,48]. The appearance of these bands is accompanied by a negative broad peak at around 3620\u2009cm\u20131 due to the disappearance of surface hydroxyl groups through the dissociative adsorption of methanol. As described in the literature, the C\u2013O stretching region located in the region 1800\u2013700\u2009cm\u20131 also shows the characteristics bands associated to methanol. Thus, at room temperature it can be notice a sharp band at around 1060 and 1135\u2009cm\u20131 that can be ascribed to C\u2013O stretching modes of methoxy group. The evolution of bands in this region during different experimental conditions is similar for all catalysts, varying the relative intensity of certain bands. Thus, during photocatalytic experiment at 50 \u00baC, new intense bands appear within this region at 1580\u2009cm\u20131 and 1360\u2009cm\u20131. These new bands can be ascribed respectively to the \u03c5as(COO) and \u03c5s(COO) for bidentate formate ions adsorbed on TiO2 surface. The intensity of these bands remains almost similar as Cu loading increases for Cu contents higher than 2\u2009wt% (Fig. S2). It is worthy to mention that these bands do not appear during thermal treatment. In this case, only the band at 1135\u2009cm\u20131 seems to be exalted (Fig. S2).Concerning the thermo-photocatalytic process, as temperature increases, bands associated to formates become notably marked. However, at 200 \u00baC these bands completely disappear. At 120 \u00baC, such bands are noticeable, showing exponential growing intensities as Cu content increases (Fig. S2). Therefore, the formation of formates seems to be clearly favoured by the increasing metal sites. As previously stated, the presence of formates would determine the different evolution of CO.[29] Thus, for photocatalytic runs we have already showed that S\n\nCO\n values were particularly lower with respect to those found for thermo- and thermo-photocatalytic experiments (Fig. 8). These two facts can be correlated by considering formic acid dehydration reaction which leads to the formation of CO and H2O. As above discussed, during thermo-photocatalytic experiment, at moderate temperatures the formation of formates was exalted and increases with Cu content. It is clear that formate formation is favoured by temperature.The absence of formate bands at reaction temperature clearly indicates that they were completely transformed into CO through dehydration. At the same time, we have also stated that S\n\nCO\n drastically decays. Therefore, during thermo-catalytic reforming at 200 \u00baC the outstanding H2 formation can be ascribable to the enhanced transformation of methanol into formate which proceed to CO and then through WGS reaction to H2 formation. From both facts it would be inferred that WGS reaction is playing an important role. In this sense, the higher the Cu load, the more important the WGS reaction becomes in the overall process.Thus, it can be argued that the efficiency of the thermo-photocatalytic reaction over Cu/TiO2 systems would be improved with Cu loading. Thus, the different optimum Cu loading attained for photocatalytic and thermo-photocatalytic reforming was conditioned by the thermal contribution through WGS reaction. From these results, it seems that this later process would be favoured by increasing Cu loadings. Thus, while for photocatalytic reforming Cu2P90 showed the optimum value, for the combined thermo-photocatalytic process Cu5P90 would be the best catalyst.Therefore, we evidenced a clear dependency of the mechanism on the metal content. Moreover, as we stated from TPR and XPS Cu species present in each case is different. Thus, while for Cu2P90 Cu2+ with strong interaction with support are the predominant, for Cu5P90 CuO dispersed nanoclusters would be also present. The influence of cluster size on the photocatalytic performance of H2 photoreforming was studied by Zheng et al. which stated that small clusters usually possess higher charge separation efficiency [49]. In this sense, Li et al. showed that through ligand assisted method, ultrafine size clusters of Cu species (with about 1\u2009nm), dispersed on the surface of TiO2 were obtained [50]. These authors reported that the superior H2 evolution performance of CuO/TiO2 mainly originates from the cluster size and unique interfacial structure formed. On the other hand, CuO clusters modulated in the range 4\u20138\u2009nm by controlling the loading amount of Cu and calcining temperature were obtained for photocatalytic WGS reaction [51]. Therefore, it is clear that different process would require not only a particular metal loading but also an optimum cluster size. Moreover, different copper species distribution could be also correlated with the better performance as temperature increases.Copper supported TiO2 with different metal loadings was used as catalyst for the gas phase methanol thermal photoreforming. From the wide characterization of the catalyst, we have shown that copper metal co-catalyst has been efficiently deposited and well dispersed on the support surface within the studied range of metal loading. We have stated the important synergistic effect in the thermal-photocatalytic reaction. Thus, ca 60\u2009mmol/h\u00b7gcat was obtained after 4\u2009h at 200 \u00baC upon illumination. Different optimum metal loading can be envisaged for different methanol reforming experiments. Moreover, as increasing metal loading, the nature of Cu species seems to be modified, subsequently conditioning the catalytic performance. Thus, while for the photocatalytic reforming Cu2P90 was the best catalyst, the thermo-catalytic reforming performance follows a lineal evolution with Cu content in the range of the studied metal content. For thermo-photocatalytic reforming, the optimum value would be Cu5P90 for which the synergistic improvement started to decay. During thermo-photocatalytic reaction different thermo- and photo-processes are involved, resulting in different optimal catalyst formulation. We have shown that while for photocatalytic reaction low co-catalyst loading leads to the better performance, for thermo-photocatalytic process higher loading is needed. These results clearly highlight the necessity of metal loading optimization due to the occurrence of different catalytic mechanism that simultaneously takes place during the thermo-photocatalytic reaction.\nF. Platero: Methodology, Investigation, Writing \u2013 original draft. A. Caballero: Conceptualization, Methodology, Supervision. G. Col\u00f3n: Conceptualization, Methodology, Writing \u2013 original draft, Writing \u2013 review & editing, Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Authors acknowledge the financial support from the EU FEDER and Junta de Andaluc\u00eda under I+D+i Project P20-00156. We also acknowledge the financial support from PID2020-119946RB-I00 project funded by MCIN/AEI/ 10.13039/501100011033 and, as appropriate, by \u201cERDF A way of making Europe\u201d, by the \u201cEuropean Union\u201d.Supplementary data associated with this article can be found in the online version at doi:10.1016/j.apcata.2022.118804.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  We have optimized the H2 production by methanol thermo-photocatalytic reforming in the gas phase using Cu/TiO2 catalyst by tuning metal loading. Metal co-catalyst has been deposited by means of chemical reduction deposition. We have stated that thermo- and thermo-photocatalytic process leads to a notable H2 production at 200 \u00baC. By in-situ FTIR studies we evidenced that formate formation follows a different evolution depending on the reforming experiment. These surface formate would lead to CO formation through dehydration reaction. At higher Cu content the low CO selectivity denote that water-gas-shift reaction would predominate and exalt H2 yield. Thus, different optimum Cu content is found for each reforming experiment. While for the photocatalytic reforming Cu/TiO2 (2\u00a0wt%) is the best catalyst of the series, we should increase the Cu content to Cu/TiO2 (5\u00a0wt%) to achieve the optimum performance for thermo-photocatalytic reforming of methanol.\n               "}
{"full_text": "In order to deal with the increasing global energy shortage and serious climate issues caused by the usage of fossil fuels, technologies for production of renewable and clean energy on the terawatt scale are urgently required in the near future, with a required minimum production of 10 terawatt (TW) of renewable energy needed by 2050.\n1\u20133\n One of the promising solutions is to use \u201cgreen hydrogen\u201d, hydrogen produced by water electrolysis using renewable electricity.\n3\u20135\n As a consequence, further development of water electrolysis technologies is required to support the terawatt-scale production of H2 fuel.Proton-exchange membrane (PEM) electrolysis, in which proton-conducting polymer thin membranes are used as solid electrolyte instead of liquid electrolyte, was developed earlier in 1950s for hydrogen and oxygen generation under anaerobic environments.\n6\n\n,\n\n7\n Compared to the conventional alkaline electrolysis, the PEM electrolysis shows plenty of superiorities, such as a small footprint, fast response time, capability of reaching high current densities above 2 A/cm2, high Faradaic efficiency, enabling safe differential pressure operation, and requiring only pure water in place of corrosive electrolytes. These advantages avoid most of the shortcomings of alkaline electrolysis and can facilitate the future coupling with solar electricity grid. It is time to rekindle the research on PEM electrolyzers to overcome the long-existing disadvantages of PEM electrolyzer. One of the biggest disadvantages is that only noble metal materials can currently be used as working electrodes, for instance, Pt group metals (PGM) as cathode and Ir-based oxides or metallic Ir as anode.\n6\n The reason for this current limitation is the acidic local environment within the polymer membrane, even though only pure water is used as reactant solution. The high local acidity can lead to fast corrosion of non-PGM inorganic materials under electrolytic conditions, and concurrently the catalytic efficiencies of non-PGM materials are far from matching catalytic efficiencies achieved by PGM-based catalysts.A convictive fact is that most molecular water oxidation catalysts\n8\n\n,\n\n9\n (WOCs) and hydrogen evolution reaction (HER) catalysts\n10\n\n,\n\n11\n were initially developed, evaluated, and deeply studied under acidic conditions. Considering the chemical nature of H+ and OH\u2212, in contrast to non-PGM inorganic materials which are more resistant to alkaline conditions, metal complexes generally are more resistive to acidic conditions, especially in unbuffered pure water systems. In the preliminary investigations by Millet and co-workers, several molecular WOCs and HER catalysts evaluated in PEM electrolyzers displayed demonstrable performances (can reach A/cm2 current density), sufficient stability (1,000\u00a0h of operation without significant degradation for the hydrogen evolution catalyst), and expectable lower sensitivity to metal impurities in the system.\n12\u201315\n For an example, the [Ru(tpy)(bpy)(OH)] (tpy\u00a0= 2,2\u2032:6\u2032,2\u2032'-terpyridine and bpy is 2,2\u2032-bipyridine) tested in PEM electrolyzer achieved 1 mA/cm2 current density at about 1.9\u00a0V applied potential, which is only 100 and 200\u00a0mV higher than when the state-of-the-art catalysts Ir and RuO2 were used under the same test conditions, respectively.\n14\n It has been seen that molecular catalysts may bring new opportunities to replace noble metal catalysts in PEM electrolyzers.\n16\n\n,\n\n17\n\nMoreover, molecular catalysts have many advantages over inorganic materials, as they are atom economical and their discernible active sites and relatively clear catalytic mechanisms allow finely structural design to obtain high intrinsic activities.\n1\n\n,\n\n8\n\n,\n\n9\n\n,\n\n18\u201321\n During recent decades, numerous efficient molecular catalysts with outstanding catalytic performances and well-studied catalytic mechanisms have been developed for water oxidation,\n8\n\n,\n\n9\n\n,\n\n22\n proton reduction,\n11\n and CO2 reduction reactions (CO2RR)\n23\n (several representative series are shown in Figure\u00a01\nA). However, these molecular catalysts are initially designed to be combined with photosensitizers for (photo)electrocatalytic systems, and have been rarely considered in view of their application in PEM electrolyzers.The idea of applying molecular catalysts in PEM eletrolyzers\n6\n and anion exchange membrane (AEM) electrolyzers,\n25\n\n,\n\n26\n which together are known as solid polymer electrolyte membrane (SPE) electrolyzers, deserves wide consideration. The development of SPE electrolysis technologies can potentially be taken in a completely new direction and high efficacies by the application of these ready molecular catalysts,\n16\u201318\n as indicated by the successful precedent reported recently by Berlinguette and co-workers.\n24\n Through implementation of a commercially available cobalt phthalocyanine (CoPc) CO2RR catalyst in a AEM electrolyzer (Figure\u00a01B), production of CO from CO2 reduction can be achieved with >95% selectivity at current densities of 150 mA/cm2, and catalytic current density at 150 mA/cm2 can be maintained for >100 h. Turnover numbers (TON) of 4,000 and a turnover frequency (TOF) of 216 h\u22121 was achieved over 8\u00a0h of electrolysis at 50 mA/cm2 with a Faradaic efficiency for CO production >90%. Although an obvious efficiency drop was observed for operation at high current density, further detailed investigations displayed that the rapid drop is primarily due to the lowered proton inventory inside the membrane-electrodes assembly (MEA) and not due the CoPc decomposition. These fresh and excellent results challenge the widely accepted dogma that there is no chance to employ molecular catalysts in commercially applicable electrolysis techniques.\n24\n Efficient CO2RR molecular catalysts with good selectivity and stability can benefit the production of fuels from CO2 reduction by AEM electrolyzer.With abundant well-developed molecular catalysts in hand, another remaining challenge to efficiently utilize molecular catalysts in PEM electrolysis is the effective preparation of molecular catalysts modified electrodes. Considering the modification stability, conductivity, and homogeneous exposure of catalytic sites, immobilization of molecular catalysts is not a straightforward process. Both adequate conductive supports and well-designed immobilization methods are required to assemble a finally efficient molecular catalyst incorporated MEA. Carbon materials are cost-effective, environmentally friendly, are highly electrically conductive with wide potential windows and display good chemical stability and rich surface chemistry to facilitate their modification with molecular catalysts.\n27\n Moreover, carbon materials are commonly employed as substrates for loading Pt catalysts and as cathode current collectors in PEM electrolyzers, indicating their promising stability as substrates for loading reduction catalysts under acidic and electrochemical conditions.\n6\n The currently available carbon materials also have enough stability to be used as anode substrate for short-term operation at high voltages, a time-frame suitable for electrocatalyst characterization purposes (See details about carbon-material stability in Stability of Carbon Materials).In this perspective review, we highlight advances of molecular catalysts immobilized on carbon materials as a valuable approach to launch the investigations on application of molecular catalysts in PEM electrolyzers. After the introduction, carbon materials, catalyst immobilization, and characterizations of molecular catalyst on carbon materials are described. Then we provide an overview of molecular catalysts immobilized on carbon materials for electrochemical water oxidation, proton reduction, and CO2 reduction reactions, which may be directly employed for the investigations of application of molecular catalysts in PEM and AEM electrolyzers. Concerns about the stability of molecular catalysts are discussed in a separate section. To conclude, we discuss future scientific perspectives and challenges to advance this promising, but currently underdeveloped, technology for solar fuel production through integration of SPE electrolysis with molecular catalysts.Depending on the type and organization of C\u2013C bonds in the structure, carbon materials can comprise different allotropes with different electrochemical properties, e.g., amorphous carbon, diamond, graphite, graphene, carbon nanotubes (CNTs) and fullerenes (Figure\u00a02\n). Amorphous carbon, the simplest structure amongst these, has poor stability under positive potentials. Diamond, containing sp3-hybridized C\u2013C bonds, possesses excellent mechanical properties but low conductivity. Graphite has high electrical conductivity due to the stacking of aromatic planes consisting of sp2-hybridized C\u2013C bonds with delocalized \u03c0 electrons. An important variation of graphite is carbon fibers, which can be used to fabricate conductive carbon papers and carbon cloth. Graphene, CNTs, and fullerenes are advanced carbon materials derived from a single aromatic plane of graphite. Therefore, their intriguing structures lead to unique physical, chemical, and electrochemical properties.\n28\n\n,\n\n29\n A well-known variant based on these structures is non-graphitizing glassy carbon (GC, also known as vitreous carbon), which is widely used as an inert electrode material in electrochemical studies.GC electrodes are commonly used for studying catalytic mechanisms and the kinetics of molecular catalysts because of their high conductivity, extreme resistance to chemicals, and control over the active surface area. Besides GC, CNTs, and graphene have been widely employed to immobilize molecular catalysts. CNTs and graphene are usually processed as thin films, porous sheets, and porous foams, providing super-conductive substrates with large electrochemically active surface areas. Not only do these materials increase the surface loading of the molecular catalysts, the carbon surfaces can also easily be modified with \u2212OH and \u2212COOH functionalities to facilitate reactant and proton transfer.\n28\n\n,\n\n29\n\nCommercial carbon materials, such as carbon fiber cloth, carbon fiber paper, graphite blocks, and GC electrodes (e.g. GC plate and GC foam) can be directly used as electrodes. However, many commercially available carbon materials, including CNTs and graphene, are marketed as powder. Hence, the powdery carbon materials must first be deposited on the current collecting substrate to form an electrode. Figure\u00a03\n shows three preparative procedures widely employed to fabricate carbon electrodes, i.e., drop-casting,\n30\n electrophoresis,\n31\n and vacuum filtration.\n32\n These three procedures have the advantage of simple and mild operating conditions. However, many other approaches can be adopted to prepare carbon electrodes, e.g., chemical vapor deposition (CVD), Langmuir\u2013Blodgett film deposition, layer-by-layer self-assembly and electro-polymerization.\n33\n To exemplify, interesting architectures such as three-dimensional (3D) self-supporting graphene can be produced based on the CVD method.\n34\n Development of more advanced porous carbon electrodes will indisputably facilitate the investigation of artificial photosynthetic (AP) devices based on molecular catalysts. However, the ideal methodology for electrode production for future applicable electrolyzers should be cost-effective and scalable.In the view of future practical application, the stability of carbon materials should be considered. Carbon materials are commonly employed as substrates for loading Pt catalysts and as cathode current collectors in PEM electrolyzers, indicating their promising stability as substrates for loading reduction catalysts under acidic and electrochemical conditions.\n6\n However, the use of carbon materials at the anodic side is limited due to the ability of the material to undergo electrochemical oxidation under the local stringent conditions. Carbon is thermodynamically easy to be oxidized to carbon dioxide or carbon monoxide (C\u00a0+ 2H2O \u2192 CO2\u00a0+ 4H+\u00a0+ 4e\u2212, E0\u00a0= 0.207\u00a0V versus normal hydrogen electrode (NHE); C\u00a0+ H2O \u2192 CO\u00a0+ 2H+\u00a0+ 2e\u2212, E0\u00a0= 0.518\u00a0V versus NHE).\n35\n Under typical conditions applied in PEM electrolyzers, the high oxidative potential, high oxygen concentration, and acidic aqueous solution, corrosion of common carbon materials is rapid. Therefore, they are not suitable for use in the anode of PEM electrolyzer for long-term operation.Athough most carbon materials suffer from this oxidative instability, these carbon materials can still be applied for the purposes of electrocatalyst characterization and performance testing of WOCs.\n6\n\n,\n\n14\n Fortunately, the field of carbon materials is still under continuous development,\n36\u201338\n providing carbon materials of increasing stability such as multi-walled carbon nanotubes (MWCNTs).\n35\n\n,\n\n39\n\n,\n\n40\n MWCNTs, which are widely used for modification of molecular catalysts, already have better stability under the highly oxidative and acidic conditions owing to their unique structure. Yi et\u00a0al. report that after an initial electrochemical oxidation processes, a passivating oxide layer forms on the MWCNTs, effectively protecting the inner graphitic layers from further electrochemical oxidation.\n35\n Moreover, different from the direct application of carbon materials as current collectors, carbon materials used as substrates are covered by molecular catalysts. The coverage of the carbon material with WOCs and the fast catalysis can also limit the corrosion of carbon materials. Considering the abundant advantages of carbon materials, they offer excellent opportunity for initial stages of investigating molecular catalyst based electrolyzers and further development of carbon materials may even boost applicability of these materials in future practical PEM electrolyzers.In the view of future practical application, common carbon materials, such as carbon black, carbon cloth, and carbon paper, cannot resist the high oxidation potentials, oxygen environments, and acidic conditions for long time. More advanced carbon materials, such as MWCNTs, are more stable; however, further research is certainly necessary to understand how MWCNTs and other new carbon materials would perform under realistic conditions. Corrosive resistant oxides, such as TiO2, SnO2, and Ta2O5, can be considered to replace carbon materials for loading molecular catalysts.\n6\n However, these materials generally have lower electron conductivities and smaller surface area than carbon materials. To realize the practical application of molecular catalyst based PEM electrolyzers, development of more advanced conductive substrates with high surface area and long-term stability under acidic and oxidative conditions is essential.A great advantage of molecular catalysts is that each active site is in principle not restricted, and all can participate in catalysis. To retain this fundamental advantage, the approach for loading a molecular catalyst on an electrode must be compatible.\n41\u201343\n Two different approaches for the immobilization of a molecular catalyst are demonstrated in Figure\u00a04\nA. With immobilization via self-assembly, the catalyst is integrated \u201cone-by-one\u201d on the surface of the highly conductive electrode. In comparison, in bulk immobilization catalysts aggregate on the electrode to form a solid-solid contact interface, resulting in the loss of both surface features as well as molecular properties. Immobilization through self-assembly facilitates the electron transfer between the catalyst and the supporting electrode, accessibility of active sites by the reactants, and efficient proton transfer between the catalyst and electrolyte, i.e., it often promotes efficient operation of catalyst molecules. Many simple and conventional methods for the preparation of hybrid electrodes that have emerged do not meet this demand (e.g., drop casting, spin coating, and Nafion conglutination) and do not utilize the intrinsic activity of every catalyst molecule introduced. To achieve one-by-one anchoring of the catalyst on the substrate carbon material, the connection between the molecule and carbon substrate should be established in situ. In other words, the catalyst should be anchored in a molecular fashion via self-assembly from its homogeneous solution.Self-assembly of a molecular catalyst on the surface of carbon materials can be accomplished via various approaches, generally classified based on the type of interaction between the catalyst and the surface of carbon substrate, i.e., covalent or non-covalent immobilizations.\n27\n\n,\n\n42\n To achieve covalent attachment to the substrate surface, the molecular catalyst needs to be modified with a reactive group that can be reacted with the pretreated carbon electrode surface to form a covalent linkage. Various synthetic procedures have been developed for achieving this covalent grafting of molecules on the carbon electrode surface, e.g., amidation, nucleophilic substitution, and 1,3-dipolar cycloaddition reactions.\n27\n The attachment of molecules by covalent linkage provides highly stable connections, however, the pre-modification of molecular catalysts and carbon surfaces is generally complex and time-consuming. Based on the simplicity of molecular modification, mild reaction conditions, and stability of the target covalent bond, the most widely occurring immobilization reactions used for electrode functionalization are the diazonium ion reduction and the copper catalyzed alkyne-azide \u201cclick\u201d reaction (Figure\u00a04B).\n44\n\nIn contrast, non-covalent immobilization of molecular catalysts involves van der Waals interactions, \u03c0\u2013\u03c0 stacking, or electrostatic interactions between the functional groups in the catalyst and the carbon material (Figure\u00a04B).\n45\n Organometallic compounds with aromatic rings in their structure, can physically adsorb onto porous carbon materials, but often molecule immobilization through physical adsorption has relatively low stability and loading amount.\n27\n Generally, decoration with large conjugated \u03c0 systems, as can be found in porphyrines, phthalocyanines, and pyrene, is required for the adsorbed complex to stably adhere to the surface of CNTs and graphene.\n46\u201348\n The molecular catalysts can also be immobilized on the carbon material through electrostatic interactions. Ion-pairing can be achieved by functionalizing the target molecule and the surface of the carbon material with opposite electrical charges.\n49\n Another acceptable method is in situ polymerization of the modified molecular catalyst to form a polymer chain or a thin polymer film of catalysts onto carbon electrodes,\n50\n although the optimization of polymerization conditions and the characterization of the final polymer provides additional challenges associated with this method. Further details and examples of these approaches will be discussed in the following sections.One of the challenges associated with molecular catalyst immobilized on carbon materials is their characterization. Due to the trace loading amount, the sub-nano scale size of the molecules, and the strong UV-visible absorption of carbon materials, some conventional characterization methods, such as UV-visible spectroscopy, X-ray diffraction, and scanning electron microscopy are not capable of analyzing molecular catalyst modified carbon materials. Electrochemical measurements, X-ray photoelectron spectroscopy (XPS), X-ray absorption spectroscopy (XAS), transmission electron microscopy (TEM), resonance Raman spectroscopy, and infrared (IR) spectroscopy have to be used instead for characterization. This section provides a summary of these approaches for the characterization of molecular catalyst modified carbon electrodes.Electrochemical measurements can be used to detail the redox processes of the catalyst. For example, Figure\u00a05\nA shows the cyclic voltammogram (CV) of a cobalt-catalyst-modified CNT electrode for proton reduction (cf.\nFigure\u00a016D for structural details).\n51\n By comparing the obtained CV with that recorded for the bulk solution of the cobalt catalyst, it was observed that the functionalized CNT electrode showed similar electrochemical behavior, displaying a distinct reversible redox character at \u22121.08\u00a0V versus ferrocenium/ferrocene (Fc+/Fc), attributed to the CoII/CoI couple of the cobalt catalyst. Observation of this redox event provides a clear indication that the cobalt complex was successfully modified on the CNT electrode. Since the currents of both anodic and cathodic peaks of the CoII/CoI redox vary linearly with the scan rates (Figures 5B and 5C), the modification of the cobalt catalyst on the surface of the carbon electrode could be confirmed. Furthermore, the loading amount (4.5\u00a0\u00d7 10\u22129 mol\u00b7cm\u22122) of the molecular catalyst can be estimated from the coulomb value associated with the single-electron redox process by integration of the redox wave in Figure\u00a05A.To obtain more detailed information of the electronic structure and coordination environment of the metal center, XPS and XAS can be used. XPS is a very sensitive technique for surface characterization, especially for metal atoms, making this technique a powerful tool for the characterization of immobilized molecules on the surfaces of carbon materials. Revisiting the cobalt-modified CNT example, the XPS spectrum of the Co-CNT sample (blue line in Figures 6A and 6B) shows an increased N 1s peak and two distinct peaks at 781.7 and 796.6 eV for the Co 2p3/2 and Co 2p1/2 levels, respectively.\n51\n The splitting of the peaks with an energy difference of 14.9 eV and absence of any distinct satellite on the Co 2p core-level signals demonstrates the presence of the CoIII ion, i.e., the immobilized cobalt catalyst.XAS is especially useful for the characterization of molecular catalysts modified on carbon materials, as it can provide information of oxidation states and electronic structure from X-ray absorption near edge structure (XANES) spectrum, and local environment from extended X-Ray absorption fine structure (EXAFS) spectrum.\n53\n\n,\n\n54\n Under the fluorescence detection mode, XAS is highly surface sensitive. For example, XAS analysis was achieved for a copper catalyst immobilized graphene with surface loading of 0.050\u00a0nmol/cm2 (cf.\nFigure\u00a010E for structural details).\n52\n The XANES spectrum of the copper catalyst-graphene was found similar as the XANES spectrum of the individual copper molecules in frozen MeCN, indicating the successful modification (Figure\u00a06C). The interactions between the copper molecules and the graphene interfaces were also identified from some minor changes in the XANES spectrum of the immobilized copper catalyst. Samples after bulk electrolysis showed identical XANES features to those characterized before test. Decomposition of the copper catalyst into CuO could be excluded by comparing the obtained XANES and EXAFS spectra (Figure\u00a06D) of the copper catalyst after catalysis and the contrasting CuO. Moreover, it is relatively easy to realize in-situ investigations by XAS measurements, which can benefit a lot for studies of catalytic mechanism.\n53\n\n,\n\n54\n For instance, XAS techniques were employed to successfully track the key 7-coordinate RuV=O intermediate of a Ru WOC modified on electrode surface.\n55\n\nThe third approach to directly observe the molecular catalyst on the surface of carbon materials, involves advanced TEM techniques, e.g., scanning TEM (STEM) and high-resolution TEM (HRTEM). Figure\u00a07\n shows STEM and HRTEM images of the Ru4POM modified CNT (cf.\nFigure\u00a09 for structural details).\n49\n The STEM images illustrate the electron-dense catalysts (bright spots, diameter 1\u20132\u00a0nm), indicating individually separated Ru4POM molecules on the CNT surface (Figures 7A and 7B). The corresponding energy-dispersive X-ray spectroscopy (EDX) analysis showed the distribution of elemental Ru and W in the color-coded elemental map (Figure\u00a07C). The blue and red colors were assigned to the carbon K\u03b1 and tungsten L\u03b1 emissions, respectively. The catalytic ruthenium cores were barely visible around 19 keV because they are embedded in the large tungsten based polyoxymetalate ligands. The HRTEM images show a 2\u00a0nm thick layer of catalyst coated on the surface of the CNT walls (Figure\u00a07D). In a separate study, where the Ru4POM was modified on graphene, the Ru4POM molecules were identified with atomic resolution, as shown in its HRTEM images (Figure\u00a07E).\n56\n These examples illustrate that it is feasible to observe the immobilization of molecular catalysts on the surface of carbon-materials. However, Ru4POM catalysts have numerous heavy metal atoms in the ligands, making their identification easier relative to catalysts with pure organic ligands. For carbon materials modified by molecular catalysts with organic ligands, center metals of molecular catalysts generally can be identified by TEM techniques.\n57\u201359\n By the more advanced single-molecule atomic-resolution real-time (SMART) TEM technique,\n60\u201362\n it is also possible to visualize the shape and movement of single molecules on CNT (Figure\u00a07F), but this technique is not widely available and requires special and strict sample preparation.Resonance Raman spectroscopy is a simple and efficient way to probe the immobilization of molecular catalyst on the surface of carbon materials. For example, the characteristic bands of both Ru4POM (Ru\u2013H2O stretching modes, 250\u2013500\u00a0cm\u22121, and Ru4 core vibrations, 450\u2013500\u00a0cm\u22121) and graphene (typical D, G, and G' bands at 1,350, 1,600, and 2,700\u00a0cm\u22121) were observed in the Raman spectra of the Ru4POM modified graphene (Figure\u00a08\nA), confirming the successful immobilization.\n56\n It should be noted that some vibrational features of the molecular catalyst may not be observable after it is modified on the surface of carbon material, such as the peak at \u223c770\u00a0cm\u22121 of Ru4POM in Figure\u00a08A, because some of the vibrational modes of the molecule could be prohibited due to the electronic interactions between the modified molecule and the surface of carbon material.\n59\n\nRaman characterization not only show the existence of molecules on the carbon surface but can also provide information regarding the interactions between immobilized molecule and carbon surface. The Raman D, G, and G' bands of graphitic materials, originating from the presence of sp2-sp3 hybridization in the hexagonal framework, are sensitive to changes of surface structure and perturbations to the \u03c0-electronic structure.\n63\n For the covalent modification of graphitic materials, sp2 hybridization in the hexagonal framework is decreased due to the formation of anchoring bonds of molecular catalysts, usually leading to the observation of increased D and G band ratios (I\n\nD\n/I\n\nG\n).\n58\n\n,\n\n64\n Taking the Ni pyridine complex (Ni-bpy) modified CNTs as an example (cf.\nFigure\u00a014C for structural details), the I\n\nD\n\n/I\n\nG\n ratio of \u223c1.4 for the pristine CNTs is increased to 1.9 for the bpy-CNTs (Figure\u00a08B).\n58\n Moreover, a shift of the G' band was also observed, indicating charge transfer between CNTs and the modified bpy moiety. In contrast, the noncovalent modification of graphitic materials does not directly change the surface structure of graphitic materials. The influence of graphitic materials to the Raman spectra depended on the strength of noncovalent interaction between molecular layer and the carbon surface. Adsorption of planar molecules, such as phthalocyanines, porphyrins, and tetrapyrroles, on graphitic materials via \u03c0\u2013\u03c0 stacking can induce 2\u20134\u00a0cm\u22121 shifts of the G band and changes of the I\n\nD\n\n/I\n\nG\n ratio, due to the strong charge transfer between the adsorbed molecular layer and the carbon surface.\n47\n\n,\n\n65\n When the phthalocyanine with steric hindrance groups adsorbed on CNTs, the change of I\n\nD\n\n/I\n\nG\n ratio disappeared, indicating less close interaction between the molecular layer and the carbon surface.\n65\n When the distance between the molecular layer and carbon surface is large, Raman spectra of graphitic materials showed no obvious changes, indicating minor interaction between the molecular layer and the carbon surface.\n56\n\n,\n\n57\n\nRaman spectroscopy is generally recognized as a qualitative characterization method. Less known is that Raman spectroscopy, in principle, can also be used for quantitative analysis.\n66\n Nonetheless, it is challenging to make Raman analysis quantitative because the intensity of a Raman spectrum depends not only on the nature of sample, but also on instrumental parameters and operation factors. Advanced Raman spectrometers, smart experimental design, and highly skilled operation are required to realize Raman quantitative analysis. Considering the merits of Raman spectroscopy and the lack of methods for quantifying molecular catalysts modified on carbon materials, it is advantageous to develop Raman quantitative analysis to support future developments in this field.IR spectroscopy, which is complementary to the Raman spectroscopy, is another important and effective method to characterize molecular catalysts modified on carbon materials, especially many HER catalysts and CO2RR catalysts that have CO ligands in their structures.\n67\u201370\n IR can also be used to sensitively detect trace amounts of in situ generated CO products during electrocatalytic CO2 reduction.\n71\n For the covalent immobilization via \u201cclick\u201d reaction, the characteristic stretching vibration near 2,110\u00a0cm\u22121 of the azido group can be used to monitor the grafting process (Figure\u00a08C, cf. Figure\u00a013 for structural details).\n46\n\n,\n\n71\n Importantly, both IR spectroscopy\n69\n\n,\n\n72\n and Raman spectroscopy\n73\n can be combined with electrochemical systems to establish spectroelectrochemistry investigations for tracking catalytic intermediates and deeply studying catalytic mechanisms.In artificial photosynthesis, water oxidation (2H2O \u2192 O2\u00a0+ 4e\u2212\u00a0+ 4H+, E\no\u00a0= 1.23\u00a0V versus NHE at pH\u00a0= 0, NHE\u00a0= normal hydrogen electrode) is considered the ideal way to provide abundant electrons and protons for the reduction reactions (e.g., proton reduction and CO2 reduction). However, this reaction proves to be a bottleneck due to the high potential requirements caused by the high reaction barrier and kinetic complexity of the process. To increase water oxidation efficiencies, molecular WOCs have been investigated during the past 3 decades, providing knowledge and insight into establishing efficient WOCs.\n8\n\n,\n\n9\n\n,\n\n19\n\n,\n\n20\n\n,\n\n74\u201376\n With a wide variety of WOCs established, the essential fabrication of electrolyzers by grafting molecular WOCs on electrode materials has become more attractive. As a promising strategy, molecular WOC modified carbon materials have been studied in electrochemical water oxidation.A pioneering work on noncovalent modification of WOCs on carbon materials was reported by Bonchio and co-workers in 2010. A Ru4-POM (polyoxometalate) complex was assembled on a conductive support of MWCNTs.\n49\n\nFigure\u00a09\n represents catalyst adsorption on the MWCNTs via electrostatic interactions, achieved by decorating the MWCNTs with polyamidoamine ammonium dendrons. The Ru4-POM@MWCNTs hybrid composite was characterized via resonant Raman spectroscopy, small-angle X-ray scattering (SAXS) analysis, STEM, EDX-spectroscopy, and HRTEM to reveal the local distribution of catalysts (Figures 7A\u20137D). Electrochemical water oxidation using this Ru4-POM@MWCNTs composite material was studied by drop-casting the material on an indium tin oxide (ITO) electrode. Using a Pt counter electrode, catalytic water splitting at pH 7 was observed with a TOF in the range 36\u2013306 h\u22121 (\u03b7\u00a0= 0.35\u20130.6 V) for oxygen evolution. An analogous anode fabricated by mixing Ru4-POM with amorphous carbon showed only very poor catalytic performance, indicating the importance of molecular interfaces in controlling and promoting electron-transfer at heterogeneous surfaces.Bonchio\u2019s group further studied the behavior of Ru4-POM on graphene-based electrodes. Graphene is considered more attractive than MWCNTs due to its high stability and conductivity, as well as larger surface area-to-volume ratio. Following the same functionalization strategy, indeed a 2-fold enhancement in electrocatalytic performance compared to that of the MWCNTs analogue could be recorded when graphene was utilized as the support material.\n56\n The resulting electrode displayed catalytic activity for oxygen evolution at pH 7 with an overpotential of 0.3\u00a0V and the activity was maintained after 4\u00a0h of testing. A similar strategy has been reported involving confining the catalyst in a highly porous wet graphene film.\n77\n\nIn 2011, the groups of Li and Sun developed an approach to decorate MWCNTs with pyrene-functionalized WOCs via noncovalent \u03c0\u2013\u03c0 stacking interactions. The Ru(bda)(pic)2 (pic\u00a0= 4-picoline) catalyst was functionalized with two pyrene anchoring groups, allowing the molecule to adsorb on MWCNT modified ITO electrodes (Figure\u00a010\nA).\n48\n The obtained electrodes displayed catalytic oxygen evolution under neutral conditions at a low overpotential of 0.28 V. The molecular nature of the Ru-bda catalyst was retained on the surface of MWCNTs electrode as confirmed by observation of the two distinct peaks at 0.59 and 0.91\u00a0V corresponding to the RuII/RuIII and RuIII/RuIV redox couples, respectively. An average TOF of 0.3 s\u22121 and TON of 11,000 were observed for oxygen evolution in a neutral solution under long-term (> 10 h) electrolysis. After long-term electrolysis, the redox peaks of the Ru-bda catalyst were still clearly discernible indicating the promising stability of this system.Meyer et\u00a0al. found that on metal oxide surfaces an electron transfer mediator, such as [Ru(bpy)3]2+, is required for the oxidation of the Ru-bda catalysts.\n79\n\n,\n\n80\n These results clearly indicate that the highly conductive CNTs are superior for the electrochemical water oxidation by Ru-bda catalysts. The catalytic behavior on the CNT surface was further detailed by Ahlquist and co-workers through simulation of catalyst dynamics using the empirical valence bond method (Figure\u00a010B).\n78\n It was found that not only that the pyrene \u201carm\u201d attached to the axial ligand tightly adsorbs to the surface of CNTs, but also that the \u201cbody\u201d of the Ru-bda catalysts adsorbs as well. This significantly retards O\u2212O bond formation due to the large distortion required for the catalyst to reach the transition state geometry.Recently, Llobet and co-workers reported the Ru(tda)L2 (tda2-\u00a0= [2,2\u2032:6\u2032,2\u2032\u2019-terpyridine]-6,6\u2032\u2019-dicarboxylato) WOC modified MWCNT anode obtained following the noncovalent \u03c0\u2013\u03c0 stacking modification strategy (Figure\u00a010C).\n81\n According to the authors\u2019 calculations based on the foot-of-the-wave analysis, more than a million TONs at pH 7 with an Eapp\u00a0= 1.45\u00a0V versus NHE were achieved by these hybrid solid state materials. The oxidative stability of the working electrode was monitored by electrochemical techniques and XAS spectroscopy before, during, and after catalysis. After electrocatalysis for 1 h, the Ru molecule still can be identified on the electrode, and no obvious evidences indicate the presence of RuO2. With this Ru-tda/MWCNT material, the groups of Llobet and Lewis assembled an n-Si/TiO2/C/CNT/Ru-tda photoanode for photoelectrochemical water oxidation by binding the Ru-tda/MWCNT on TiO2-protected Si photoanode (Figure\u00a011\nA).\n82\n Photocurrent densities of 1 mA cm\u22122 could be achieved with this hybrid device at pH\u00a0= 7 under three Sun illumination. Application of the Ru-tda/MWCNT material to a multilayered hetero-structured WO3/BiVO4 semiconductor photoanode (Figure\u00a011B) further enhanced the overall photoelectrochemical performance.\n83\n\nNon-ruthenium centered catalysts have also been immobilized using the \u03c0\u2013\u03c0 stacking strategy of pyrene. For example, the Cao group reported the immobilization of a bi-functional cobalt corrole catalyst onto MWCNTs (Figure\u00a010D).\n84\n The corrole functionalized carbon MWCNTs were tested for both electrochemical water oxidation and oxygen reduction. For the water oxidation reaction, the fabricated electrodes achieved high electrocatalytic performance and durability under neutral aqueous conditions with an onset overpotential of 0.33 V. To illustrate the benefit of the \u03c0\u2013\u03c0 stacking using pyrene, MWCNTs loaded with pyrene-free cobalt corrole were prepared as control samples. The pyrene-free electrodes displayed much lower performances for water oxidation, clearly demonstrating the importance of strong noncovalent \u03c0\u2013\u03c0 interactions between the pyrene moiety and the MWCNTs in facilitating the immobilization and electron transfer processes.Interestingly, the pyrene has been shown to not only function as anchor for molecular catalysts. The electronic \u03c0-delocalization effect of pyrene on the electrocatalytic performance was revealed recently by Llobet and co-workers.\n52\n They synthesized two copper complexes with the general formula of [(L)CuII]2\u2212 (L is 4-pyrenyl-1,2-phenylenebis(oxamidate) or pyrene free control o-phenylenebis(oxamidate), Figure\u00a010E). In addition to the anchoring function of the pyrene group, it provided electronic perturbation to the system. Consequently, the overpotential for water oxidation was lowered by 150\u00a0mV, in the homogeneous system in the presence of the pyrene functionality, accompanied with a dramatic increase of the TOF from 6 to 128 s\u22121. After assembly of the catalysts on the graphene sheets, the \u03c0-delocalization effect, provided by the support, boosts the catalytic activity of both catalysts. Due to the combined effect of both the pyrene group and the graphene substrate, the pyrene functionalized catalyst displays an electrocatalytic activity with an overpotential of 538\u00a0mV, a TOF of 540 s\u22121 and TONs as high as 5,300.The \u03c0\u2013\u03c0 stacking effect has been widely adopted to modify molecular catalysts on CNTs and graphene.\n48\n\n,\n\n52\n\n,\n\n81\n\n,\n\n82\n\n,\n\n84\u201386\n The success of this strategy highlights the ease and advantages of noncovalent modification. As a result, different noncovalent modification strategies for the fabrication of electrodes were explored. For example, a Ru-pdc (pdc2-\u00a0= 2,6-pyridine dicarboxylate) based WOC was immobilized on MWCNTs using the hydrophobic effect. By pre-functionalization of the catalyst with a long alkyl chain Sun and co-workers (Figure\u00a012\nA) show catalyst aggregation with the MWCNT support.\n87\n The obtained electrodes maintain a catalytic current density of 2.2 mA cm\u22122 at an overpotential of 480\u00a0mV after 1\u00a0h of bulk electrolysis, and a high TOF of 7.6 s\u22121 was observed. This immobilization strategy proved successful for cobalt based WOCs on carbon black as well (Figure\u00a012B).\n88\n The resulting material displayed good catalytic activity for water oxidation under basic conditions, with an onset overpotential of 0.32 V, and current densities of 10 mA cm\u22122 could be achieved at 0.37 V.Good electrocatalytic performance requires not only a well-designed molecular WOC with high intrinsic activity but also a suitable immobilization method. The effects of different non-covalent immobilization methods on catalysis were systematically investigated and highlighted by Cao and co-workers. They described the modification of cobalt corroles, with four different anchoring groups, on CNTs and compared the performance of these functional carbon materials for H2 and O2 evolution (Figure\u00a013\n).\n46\n The CNTs decorated with cobalt corroles with short conjugated linkers displayed the highest electrocatalytic activity for both H2 and O2 evolution reactions in pH 0\u201314 aqueous solutions. The increased performance obtained with shorter conjugated linkers was attributed to (1) the fast electron transfer ability and (2) the high stability of intermediates under strong basic or acidic conditions due to the stronger coordination of the corroles.To prevent catalyst leaching a more stable linkage between the catalyst and the substrate could offer a solution. In 2012, the Sun group reported a strategy to covalently attach a RuII(pdc)(pic)3 (pdc\u00a0= 2,6-pyridinedicarboxylate, pic\u00a0= 4-picoline) catalyst on carbon electrodes, using a diazonium ion electro-reduction protocol to introduce alkyne functionalities to the support, followed by an alkyne-azide \u201cclick\u201d reaction to immobilize the catalyst (Figure\u00a014\nA).\n44\n This study provided a universal method to covalently attach WOCs on conductive carbon surfaces. Lin and co-workers showed that diazonium grafting can also be achieved directly using the catalyst molecules. In their study, covalent immobilization of molecular Ir complexes onto a carbon electrode was achieved through direct C\u2013C bond linkage of the bipyridine.\n89\n The obtained electrode exhibits a TOF of 3.3 s\u22121 and TON of 644 during the first hour of electrolysis (Figure\u00a014B). Compared to chemically driven water oxidation with the corresponding Ir molecular catalysts, electrochemical water oxidation with modified catalysts gave increased rates and stability. The authors suggested that this strategy can be utilized as an alternative way to systematically evaluate catalysts under tunable conditions.Based on these results, the Lin group later studied catalytic performances of various less inert transition metal ions (i.e., Fe3+, Co2+, Ni2+, Cu2+) by a simple electrochemical approach (Figure\u00a014C).\n64\n In the first step, an amino-bpy was bound on the surface of CNT electrodes by diazonium grafting. Various metal ions were coordinated to this functionalized electrode and the resulting electrodes were assessed for water oxidation catalysis. Results showed that the Co-complex immobilized electrode possessed the highest catalytic activity, and is capable of oxidizing water with a TOF of 14 s\u22121 at an overpotential of 0.834 V. Recently, the Ni-bpy modified CNT electrodes were further investigated under basic conditions by Laasonen and co-workers.\n58\n They show that the Ni-bpy modified electrodes are capable of generating a current density of 10 mA cm\u22122 at an overpotential of 0.31 and 0.29\u00a0V in 0.1 and 1\u00a0M NaOH, respectively.The last discussed immobilization strategy of WOCs on carbon materials is through the polymerization of catalyst monomers. An in situ polymerization of catalyst monomers has been carried out by Sun and co-workers to decorate the surface of graphitic carbon (Figure\u00a015\nA).\n90\n In this work, in\u00a0situ polymerization of pyrrole functionalized Ru-bda catalysts was employed to enrich the graphite electrode surface with catalysts. The obtained electrode showed a high initial TOF of 10.47 s\u22121 at an overpotential of 700\u00a0mV, and a TON of 31,600 after 1\u00a0h of electrolysis in a pH 7.2 aqueous solution. The advantages of catalyst monomer polymerization were further illustrated by Du and co-workers. Through the Glaser coupling reaction, multi-layer covalent cobalt porphyrin frameworks were built on the MWCNTs template to form a hybrid material for electrochemical water oxidation (Figure\u00a015B). This hybrid material performed significantly better than its analogue where the cobalt porphyrin monomer analogue was immobilized. Overall, the hybrid materials achieved a catalytic current density of 1.0 mA/cm2 under an overpotential of 0.29\u00a0V in a pH 13.6 solution.\n50\n\nAn additional novel modification procedure has been reported by Heumann and co-workers. They constructed a cobalt-bridged ionic liquid polymer on a CNT for efficient oxygen evolution (Figure\u00a015C).\n57\n In their material, the ionic liquid polymer can act as the counter ion to increase the stability of the Co2+ ion and adjust the electron structure of the atomically dispersed Co creating a favorable environment for an oxygen evolution reaction.Hydrogen has been considered a clean and efficient next generation energy-carrier. Numerous molecular transition metal catalysts have been developed for water oxidation coupled hydrogen production, i.e., overall water splitting.\n10\n\n,\n\n11\n\n,\n\n91\n\n,\n\n92\n Immobilization of HER catalysts on carbon material has been widely investigated for future practical utilization in industry.Artero and co-workers reported nanomaterials decorated with noble metal-free molecular catalysts for hydrogen production and uptake.\n42\n\n,\n\n51\n\n,\n\n68\n\n,\n\n93\u201396\n A nickel bisdiphosphine based mimic ([Ni(P2N2)2]2+) of the catalytic center of hydrogenase enzymes was covalently attached to MWCNTs, as illustrated in Figure\u00a016\nA.\n96\n For this grafting, an activated phthalimide ester functionalized Ni(P2N2)2 complex was prepared and anchored to the amino-modified MWCNTs-electrode material via amide bond formation. The obtained electrodes showed excellent hydrogen evolution activity in aqueous sulfuric acid conditions with a low overpotential of 200\u00a0mV and exceptional stability achieving >105 turnovers. To increase practical applicability, they prepared a membrane-electrode assembly of the Ni(P2N2)2-functionalized MWCNTs. With this assembly, a current density of 4 mA cm\u22122 could be obtained at an overpotential of 300\u00a0mV, and a TON of 105 (\u00b130%) was achieved within 10\u00a0h stability tests. The authors also evaluated capability of the electrode to perform the reverse reaction, i.e., hydrogen oxidation. Under 1 atmosphere pressure of H2, an overpotential of 300\u00a0mV and TON of 3.5\u00a0\u00d7 104 (\u00b130%) was observed for 10-h bulk electrolysis.The synthetic procedure was later improved by Artero and co-workers.\n93\n Similarly, the MWCNTs were firstly functionalized with amine groups by a diazonium reduction. Now, the ligand PCy\n2NR1\n2 was initially covalently immobilized on the surface of the electrode via amide coupling, followed by separate introduction of the Ni active sites (Figure\u00a016B). The obtained hybrid CNTs were further processed into the MEA, and were evaluated for H2 production and H2 oxidation under typically operational conditions in PEM electrolyzers or fuel cells. At 85\u00b0C, the MEA with Ni-PCy\n2NR1\n2 catalysts showed a current density of 38.3 mA/cm2 at \u2212100\u00a0mV versus standard hydrogen electrode (SHE) for hydrogen evolution and 16.8 mA/cm2 at\u00a0+100\u00a0mV for hydrogen oxidation, and the corresponding performances achieved with platinum MEA (commercial Pt/C containing 46% Pt deposit on the gas diffusion layer) are 32.2 and 26.6 mA/cm2, respectively. Under technologically relevant conditions, the performance of platinum-based MEA is likely to be matched by this Ni-PCy\n2NR1\n2 modified MEA. However, it must be noted that although all the electrodes were evaluated under the same conditions in this work, these tests were carried out under half-cell configuration, which is significantly different from the real PEM cells. In addition, the performances of Pt-based electrodes can be influenced by the MEA preparation procedure over several orders of magnitude. Nevertheless, to some extent, these studies and results demonstrate possibilities of applying molecular HER catalysts in the PEM cells.To achieve straightforward and highly convenient attachment of the [Ni(P2N2)2]2+ catalyst on carbon materials, Artero and co-workers further explored non-covalent immobilization of the catalyst on CNT via \u03c0\u2013\u03c0 stacking interactions.\n94\n This was achieved by functionalization of the [Ni(P2N2)2]2+catalyst with pyrene substituents, which was subsequently physisorbed on MWCNTs through the formation of \u03c0\u2013\u03c0 stacking between the pyrene groups and graphene substrates (Figure\u00a016C). After deposition of the active MWCNT material on a gas diffusion layer electrode, an outstanding electrocatalytic hydrogen production activity was observed in 0.5\u00a0M aqueous sulfuric acid electrolyte in the presence of CO. Stability tests performed at \u2212300\u00a0mV versus NHE showed no obvious loss of activity after 6\u00a0h resulting in 8.5\u00a0\u00d7 104 turnovers.In addition to Ni based electrocatalysts, Artero and co-workers also grafted a diimine-dioxime cobalt catalyst on functionalized MWCNTs, as shown in Figure\u00a016D.\n51\n\n,\n\n97\n\n,\n\n98\n Cobalt diimine-dioxime complexes were functionalized with an azido-group, and then attached on cyclooctyne-functionalized MWCNTs by a copper free click reaction. The obtained electrode showed impressive activity towards H2 generation under acidic aqueous conditions, with onset of the reduction at an overpotential of 350\u00a0mV and a TOF of 8,000 h\u22121 per cobalt site. Over the course of 7\u00a0h electrolysis at a potential of \u22120.59\u00a0V versus RHE, 5.5\u00a0\u00d7 104 turnovers were obtained. No change of the anchored complex could be observed after long-term electrolysis, indicating the remarkable stability of the electrode. Afterwards, they investigated the oxygen tolerance of this Co based cathode for hydrogen production using GC instead of a gas-diffusion layer as electrode.\n99\n The electrode retained its excellent activity for H2 evolution using an O2-saturated aqueous acetate buffer as electrolyte, promoting the potential application of such cathodes in water-splitting devices.Direct comparison between an HER catalyst immobilized on a carbon surface and its homogeneously dissolved counterpart has been provided by Roberts and co-workers. To achieve immobilization, the GC surface was pre-functionalized with triazolyllithium groups using azide alkyne chemistry, which were subsequently reacted with an active NHS-ester-functionalized [Ni(P2N2)2]n+ electrocatalyst (Figure\u00a017\n).\n100\n The coupling could be performed for both Ni0 and Ni2+ complexes to achieve surface densities of 1.3\u00a0\u00d7 10\u221210 and 6.7\u00a0\u00d7 10\u221211 mol cm\u22122, respectively. The coupling reaction proved highly successful for the Ni0 species producing a densely packed layer, however, coupling of the Ni2+-species was less clean. Interestingly, the onset potential for hydrogen evolution was unaffected by the immobilization of the catalyst. Loss of activity of their catalyst assembly was rather caused by the decomposition of the surface-confined [Ni(P2N2)2]2+ complex. This decomposition was shown to occur more rapidly under more acidic conditions, suggesting increased lability of the P2N2-ligands through protonation.Non-covalent immobilization has also been used as successful strategy for the immobilization of planar hydrogen evolving complexes, by direct adsorption onto sp2-hybridized carbon materials, e.g., CNTs and graphene. For example, Peters and co-workers reported the preparation of hydrogen evolving electrode by simple adsorption of cobalt tetraimine complexes on electrode surfaces.\n101\n Under controlled potential electrolysis and in the presence of p-toluenesulfonic acid, Co(dmgBF2)2(MeCN)2 could be directly adsorbed on GC electrodes (Figure\u00a018\nB). The resulting electrode showed a catalytic onset overpotential of 100\u00a0mV in acetate buffer electrolyte at pH\u00a0< 4.5. Furthermore, the high electrocatalytic stability of the electrodes was shown by bulk electrolysis for 16\u00a0h at 1 mA cm\u22122, achieving a TON of 5\u00a0\u00d7 106 and Faradaic efficiency of 75\u00a0\u00b1 10%.Lehnert and co-workers also reported a facile adsorption of a cobalt catalyst on\u00a0several carbon materials by a facile soaking process (Figure\u00a018B).\n102\n\n,\n\n103\n Initially, graphene electrodeposited on fluorine doped tin oxide (FTO) plated glass was used as an adsorption substrate for the Co-based molecular catalyst. In a weakly acidic aqueous electrolyte\u00a0(pH > 3), this electrode showed low overpotentials of 0.37\u00a0V versus Pt and TOF > 1,000 s\u22121 for hydrogen production without significant degradation.\n93\n In later works, a variety of carbon materials including bulk highly ordered pyrolytic graphite, graphite, pencil graphite, graphite powder in Nafion films, graphene, and GC electrodes were targeted as substrate for molecular catalyst adsorption. All the electrodes displayed similar activity towards hydrogen production, achieving high activity in moderately acidic aqueous solutions (pH\u00a0< 4), moderate overpotentials of 0.42\u00a0V versus Pt and high initial TOFs of 96 s\u22121. Their work highlights that adsorption is a facile strategy for the immobilization of catalysts. However, it should be noted that interaction of the catalyst with GC and single-layer graphene surfaces was too weak to provide sustainable catalytic currents.\n103\n\nPyrene functionalization, as reported for WOCs, was likewise used for immobilizing hydrogen evolution catalysts, as shown for example by Brunschwig, Gray and co-workers.\n104\n In their report a [Cp\u2217Rh(L)Cl]+ catalyst (L is a pyrene functionalized bipyridine ligand) for hydrogen production is immobilized on a high surface area carbon black support (Figure\u00a019\nA). Appearance of redox peaks in the CVs confirmed loading of the complex on the electrode surface. With p-toluenesulfonic acid as proton source in the electrolyte, the electrode exhibited a hydrogen production activity with steady-state current density about 2 mA cm\u22122, TOF of 0.95 s\u22121 and TON of 206 over 75\u00a0min electrolysis. After 75\u00a0min electrolysis, the current density gradually decreased approximately 30% of activity in every hour, which is indicative of the release of the catalytically active species from the electrode.To investigate if the stability of the immobilized species can be increased, Reisner and co-workers fabricated a poly(cobaloxime)/CNT electrode and compared its H2 production performance with its analogue monomeric catalyst immobilized on the electrode surface (Figure\u00a019B).\n67\n For the monomeric catalyst, the complex was directly anchored on MWCNTs using \u03c0-\u03c0 interactions with a pyrene substituted pyridine as a ligand coordinating to the catalyst. For the polymeric complex, the monomeric catalyst precursor was placed through ligand substitution on the polymer, consisting of pyrene and pyridine substituted monomers. The pyrene groups of the polymer allowed for immobilization of the polymeric catalyst on MWCNTs. The stability of the two electrodes was probed at near-neutral pH, where the polymer-substituted electrode showed TON of 420 towards H2 production, four times higher compared to the 80 turnovers achieved by the monomer-substituted electrode. These results suggested that the polymer structure can enhance the stability of the assembly and the catalysts incorporated therein, leading to enhancement of the electrode lifetime.The importance of the \u03c0\u2013\u03c0 interactions of the catalyst with the electrode surface is made clearly apparent by the work of Cao and co-workers. They adsorbed three different cobalt corrole complexes for hydrogen production on graphene. The complexes bear a pyrene functionality and two axial pyridine ligands (1-py), a pyrene functionality and triphenylphosphine axial ligand (1-PPh3), and a complex that only has the two axial pyridine ligands and lacks the pyrene functionality (2-py) (Figure\u00a019C).\n105\n Overall, 1-PPh3/G showed better activity as compared to its analogues, indicating the significant influence of both axial and equatorial ligands on the \u03c0-\u03c0 interactions between the cobalt corrole and graphene. This catalyst achieved H2 production under the full pH range (pH 0 to 14) with high efficiency and stability. A similar conclusion could also be drawn in the study of molecular catalyst modified carbon electrodes for water oxidation fabricated based on \u03c0-\u03c0 interaction.\n52\n\nThe Lee group systematically studied catalytic activity and charge transfer behavior of (metal-)porphyrin monolayers on inactive graphene (Figure\u00a019D).\n47\n The interfacial charge transfer process was analyzed based on the performance operations of the graphene field effect transistor. They found that the two-dimensional porphyrin-based monolayers induced homogeneous active sites on graphene, showed electrochemical stability and enhanced charge transfer at the electrolyte/graphene interface. In addition, the electronegative pristine porphyrin and Pt-porphyrin monolayers displayed higher HER performance than Ni-, or Zn-porphyrin monolayers, due to their increased interfacial charge transfers. Monolayers exhibiting intermolecular hydrogen bonding also showed better electrocatalytic effects, indicating that surface hydration potentially plays an important role. Overall, this work indicates that the introduction of surface electronegativity, induced by either porphyrin or metal-porphyrin immobilization, plays an important role in the electrocatalytic HER.In addition to the strategy of molecular catalysts engineered carbon materials, which is the main focus of this review, we wish to note that metal organic compounds have also emerged as a promising support for molecular catalysts in recent years.\n106\u2013111\n For example, Marinescu and co-workers successfully integrated the cobalt dithiolene catalyst into a metal-organic surface (MOS) (Figure\u00a020\n). By layering an acetonitrile solution of [Co(MeCN)6][BF4]2 on the top of an aqueous solution of sodium benzenehexathiolate (C6S6Na6), they obtained a film with molecular formula (CoC4S4Na)n.\n112\n They placed the film on GC by simply immersing the GC surface into the reaction solution. The formed electrodes MOS1 and MOS2 achieved a current density of 10 mA cm\u22122 for H2 production at overpotential of 0.34 and 0.53\u00a0V under fully aqueous conditions at pH 1.3, respectively. A faradaic efficiency of 97\u00a0\u00b1 3% for MOS1 was determined for 2\u00a0h of electrolysis.Recently, Downes and Marinescu also reported a cobalt dithiolene polymer MOS-based electrode.\n106\n By the reaction of a cobalt (II) salt and benzene-1,2,4,5-tetrathiol in a basic solution the MOS forms and can be immobilized by simply immersing the GC electrode in the reaction solution. The resulting electrode displays a current density of 10 mA cm\u22122 at \u22120.56V versus RHE for hydrogen production at pH 1.3. A Faradaic efficiency of 97\u00a0\u00b1 3% was determined for 2\u00a0h electrolysis, and the electrodes displayed moderate stability after a 10-h electrolysis at 0.55\u00a0V versus SHE. The decrease in activity was largely due to delamination of the catalysts. In their work, they also immobilize the MOS on p-type Si electrodes for achieving photoelectrochemical hydrogen production. The assembled photocathode displayed a current density of 3.8 mA cm\u22122 at 0\u00a0V versus RHE under simulated AM 1.5\u00a0G sunlight illumination.The catalytic conversion of CO2 to liquid fuel or fuel precursors such as carbon monoxide, formic acid, ethanol, methanol, and methane is viewed as a potential source of renewable energy as well as a strategy to fix atmospheric CO2.\n23\n\n,\n\n113\n The tremendous challenges posed by CO2 fixation are great as CO2 is a thermodynamically and kinetically stable molecule, however, the rewards gained can be enormous. Catalytic mechanisms of inorganic and organometallic CO2 reduction catalysts, offering excellent activity and selectivity under electrochemical or photochemical conditions, have been extensively studied in this field. However, development of active, selective, and robust catalysts still requires huge efforts. As detailed earlier for WOCs and HER catalysts, immobilization on carbon surfaces can offer an attractive approach for increasing catalyst stability, decreasing required catalyst loading, avoiding bimolecular decomposition pathways, and facilitating catalyst lifetime determination. In this part of the review, we describe immobilization of molecular CO2 reductions catalysts on widely explored carbon electrode surfaces. To slightly expand the scope of immobilization techniques, this section will also introduce some immobilization techniques utilizing other useful materials that are doped with carbon nanomaterials.Similar to the previously described catalysts, both covalent and non-covalent techniques have been used for immobilization of CO2 reduction catalysts on carbon supports. For non-covalent immobilization, a similar pyrene substitution strategy is commonly employed as for example shown by Brunschwig, Gray, and co-workers. In their report they immobilized pyrene modified CO2 reduction catalyst Re(P)-(CO)3Cl on graphitic carbon electrode surfaces via \u03c0-\u03c0 stacking interactions (Figure\u00a021\nA).\n104\n The functionalized electrode was prepared by soaking pyrolytic graphite in a CH2Cl2 solution containing the Re catalyst for 12 h. The surface-immobilized Re catalyst reduces CO2 to CO with 70% faradaic efficiency and 58 TON at \u22122.3\u00a0V versus Cp2Fe+/0. This strategy proved useful for different catalysts. For example, Brookhart, Meyer, and co-workers reported immobilization of a pyrene appended iridium pincer catalyst on MWCNTs coated gas diffusion electrode (Figure\u00a021B).\n114\n The prepared electrode was highly efficient, selective, and stable for electrocatalytic reduction of CO2 to formate. For this engineered electrode, the polyethylene glycol coating played a critical role for the overall stability of the electrode. Under optimized electrocatalytic conditions at \u22121.4\u00a0V versus NHE a current density of 15.6 mA/cm2, with TON 54,200 and TOF 15.1 s\u22121 was obtained. The assembly was shown to be stable for over 1 h, converting CO2 to formate with 83% selectivity.Manganese catalyst [Mn(bpy)(CO)3Br] has been reported as highly efficient catalyst for homogenous CO2 reduction to CO, achieving quantitative selectivity for CO over H2 at \u22121.35\u00a0V versus Ag/AgCl.\n115\n The [Mn(bpy)(CO)3Br] catalyst was initially not immobilized on carbon materials, but was casted in a nafion membrane, where addition of MWCNT leads to great current enhancements, providing stable CO:H2 yields at \u22121.4\u00a0V versus Ag/AgCl under pH 7 conditions.\n116\n However, this strategy\u00a0is not always successful as Nafion/MWCNT electrodes functionalized with [Mn(bpy(COOH)2)(CO)3Br] and [Mn(bpy(OH)2)(CO)3Br] catalysts were found to be inactive for CO2 reduction in aqueous conditions.\n117\n Recently, Reisner and co-workers reported that the product selectivity, to CO or formate, can be tuned by altering the catalyst loading of pyrene-anchored Mn catalyst (Mnpyr) on MWCNT (Figure\u00a021C).\n69\n Immersion of the MWCNT electrode in high concentration solutions of Mnpyr (10 and 20\u00a0mM in DMF) gave high surface loading (>30\u00a0nmol cm\u22122), whereas lower concentrations of Mnpyr (0.5 and 1\u00a0mM in DMF) gave low surface loading (<20\u00a0nmol cm\u22122). With higher surface catalyst loadings a dimeric Mn0 species is formed, preferentially leading to CO as major reaction product, whereas lower surface catalyst loading prefer formation of Mn-hydride species leading to enhanced formate production.The dinuclear rhenium(I) complex [ReCl(CO)3(\u03bc-tptzH)Re(CO)3] (tptz-H\u00a0= 2,4,6-tri(pyridine-2-yl)-2H-1,3,5-triazine-1-ide) was examined as an homogeneous and heterogeneous electrocatalyst for CO2 reduction. For both homogenous and heterogeneous systems addition of methanol enhances the catalytic process and lowers the onset potential. For heterogeneous catalysis, the dinuclear rhenium complex casted on carboxylated MWCNTs-PGE shows a higher cathodic current and about 650\u00a0mV lower overpotential compared to homogeneous catalysis.\n118\n Similar results were observed for the mononuclear Re(I) complex, [ReCl(CO)3(phen-dione)] (phen-dione\u00a0= 1,10-Phenanthroline-5,6-dione), investigated under homogeneous and heterogeneous CO2 reduction conditions.\n119\n In these two cases, the Re complexes were adsorbed on the surface of the electrodes via interaction of the functional groups of the ligand coordinated to Re and the carboxyl groups of the activated MWCNTs.Covalent immobilization of CO2 reduction catalysts has also been elaborately explored. As a first example, the functionalization of a GC electrode by electro-grafting of terpyridine ligands and later modification of the electrode by, e.g., cobalt was described by Fontecave and co-workers. The electro-grafting of diazotized terpyridine ligand (tpy-Ph-N2\n+ BF4\n\u2012) on GC electrodes was achieved by performing cyclic voltammetry (+0.80 to \u22120.40 V, at 50\u00a0mV s\u20121 10 cycles). During the cyclic voltammetry experiment the reduction of the diazonium group liberates N2 generating the free radical, which reacts with the GC electrode yielding a functionalized surface. The modified electrode can be metallated by immersion in a solution of CoCl2 in DMF for 3\u00a0h at room temperature to furnish the functionalized working electrode (Figure\u00a022\nA). This cobalt-modified electrode shows excellent stability for proton and CO2 reduction into hydrogen and CO in both organic and aqueous media.\n120\n\nAnother work where the carbon material is not directly utilized for immobilization of the catalyst is provided by Meyer and co-workers. Herein, the properties of CNT are exploited to lower the barrier for electron transfer. The hybrid electrode was realized by anchoring a phosphonate anchoring group containing Ru(II) polypyridial carbene complex on the surface of thin layered TiO2/CNT/TiO2 on FTO glass electrodes (Figure\u00a022B).\n121\n The derived electrode shows short-term activity for CO2 reduction in 0.5\u00a0M NaHCO3 to produce syngas with H2/CO2 ratios varying from 1.5 at \u22120.96\u00a0V to 5.6 at \u22121.16\u00a0V versus NHE with observed maximum TONs of 308 for CO and 597 for H2 for 15\u00a0min of electrolysis. The long-term stability of the electrode is however hampered by the decomposition and partial detachment of the catalyst from the electrode. Detachment of catalysts from electrodes, to achieve longer lifetimes of electrodes, can be prevented by utilizing the stronger adherence of polymers to materials. Recently, Sato, Arai, and co-workers polymerized pyrrole functionalized mononuclear ruthenium catalyst on carbon paper coated MWCNT. The modified electrode selectively produces formate as the product of the CO2 reduction reaction in an aqueous solution, with a current density 0.9 mA cm\u22122 at \u22120.15\u00a0V (versus RHE) for 24 h.\n122\n\nPeptide coupling chemistry can also be used to immobilized catalysts bearing carboxylic acid functionalities on CNT, as detailed by Maurin and Robert. In their work, amine functionalized MWCNTs were suspended overnight with CATCO2H in DMF in the presence of HBTU and DIPEA at room temperature (Figure\u00a022C). The modified MWCNTs were then deposited on GC electrodes by the commonly used drop casting method. The prepared electrode led to efficient electro-reduction of CO2 to CO in water (pH 7.3) with 90% catalytic selectivity and 95% faradaic efficiency at 0.5\u00a0V overpotential.\n123\n An illustrative example that covalent immobilization is not always the best or only solution is provided by the same group, through immobilization of a similar iron porphyrin catalyst on MWCNTs through non-covalent interactions. Iron porphyrins functionalized with a pyrene group (CATpyr) were immobilized on MWCNT via noncovalent interactions and deposited on the GC electrode. Bulk electrolysis performed by using this surface immobilized iron porphyrin on GC electrode shows high catalytic activity (97% faradaic efficiency, 432 TON and 144 h\u22121 TOF), selectivity (CO/H2 is 25) and durability for CO2 reduction to CO in pH 7.3 water.\n124\n Typical iron and cobalt tetraphenyl porphyrin were immobilized on electrodes with MWCNTs. Comparative investigation under similar conditions revealed that the iron porphyrin-MWCNT modified electrode always performed better than the cobalt porphyrin-MWCNT modified electrode.\n125\n\nThe collective works performed on cobalt catalysts for CO2 reduction provide more detail on some of the potentials and pitfalls of catalyst immobilization on carbon materials. Kramer and McCrory reported that the edge-plane graphite (EPG) electrodes modified with cobalt\u2013phthalocyanine immobilized in poly-4-vinylpridine (P4VP) film shows 90% faradaic efficiency for CO2 reduction to CO, with a TOF of 4.8 s\u22121 at \u20140.75\u00a0V versus RHE. In comparison, the modified EPG electrode of cobalt\u2013phthalocyanine without P4VP shows only modest activity and generates mixtures of H2 and CO. The improved activity and selectivity was attributed to the formation of pyridine coordinated cobalt complexes, and to secondary effects associated with uncoordinated pyridine ligands throughout the film.\n126\n Hybrid electrodes with CoPc molecules, anchored on a CNT (Figure\u00a023\nA) were later devised by combining nanoscale and molecular level approaches. Through ultra-sonication and mechanical stirring, the catalyst could be uniformly distributed on the CNT surface, enabling a high degree of active site exposure. The electrode obtained thus showed excellent electrocatalytic activity, improved selectivity, and enhanced durability for CO2 reduction to CO. Bulk electrolysis with the prepared hybrid electrode in near neutral aqueous solution exhibits >95% faradaic efficiency for CO at 0.52\u00a0V overpotential, with 15.0\u00a0mA cm\u22122 current density and a TOF of 4.1 s\u22121.\n59\n Another approach for immobilization of the CoPc catalyst is the formation of organic-inorganic hybrid scaffold electrodes by a polymerization reaction. These electrodes were prepared by microwave irradiation, polymerizing 1,2,4,5-tetracyanobenzene together with Co2+ ions on the surface of pre-oxidized MWCNT. The resulting electrodes exhibit high catalytic activity (90% faradaic efficiency and 4,900 h\u22121 TOF at 0.5 V) and improved physical and chemical robustness compared to molecular phthalocyanine electrodes.\n127\n\nDaasbjerg and co-workers reported a comparative electrocatalytic CO2 reduction evaluation of cobalt meso-tetraphenylporphyrin (CoTPP) under both homogeneous and heterogeneous conditions (Figure\u00a023B).\n129\n In homogenous catalysis, CoTPP performs poorly with a low faradaic efficiency displaying a low product selectivity and requiring high overpotentials. Interestingly, through a straightforward immobilization which includes, sonication, drop casting and drying a relatively efficient and stable electrode could be obtained. Remarkable enhancement in the catalysts CO2 reduction ability is seen with CO (>90%) as major product at low overpotential under aqueous conditions. A CoTPP derivative could also be covalently anchored to a boron-doped, p-type conductive diamond, as shown by Hamers, Berry and co-workers. Through copper catalysed click chemistry, an azide-functionalized diamond surface could be reacted with a CoTPP derivative containing peripheral acetylene groups (Figure\u00a023C). The electrocatalytic CO2 reduction activity of this assembly was promising, displaying good stability and catalytic activity, with an overall 0.8 s\u22121 TOF observed for 16-h controlled potential electrolysis at \u22121.8\u00a0V for CO2 reduction to CO in acetonitrile solutions.\n71\n These works demonstrate how the method of modification of molecular catalysts can significantly influence the performance of the devices.A [Co(qpy)]Cl2 complex appended at the surface of MWCNT was analyzed for CO2 reduction in water at pH 7.3, giving 100% catalytic selectivity and 100% Faradaic efficiency for CO production at reduced overpotential (Figure\u00a023D). The electrodes were prepared by drop-casting GC and carbon paper with the MWCNTs suspension in 1:1 ethylene glycol-ethanol mixtures with addition of the catalyst and a small amount of Nafion (commonly utilized as PEM and stabilizing agent). The prepared electrodes achieved a current density of 9.3 mA/cm2 at only 340\u00a0mV overpotential with outstanding stability (TON of 89,095 in 4.5 h). This hybrid material retained the high selectivity of the homogeneous molecular catalyst whilst introducing the robustness of heterogeneous materials.\n130\n A structurally simple 1,10-phenanthroline-Cu complex (phen-Cu) applied on mesostructured graphene electrode provided efficient and selective CO2 reduction, with a TOF of approximately 45 s\u22121 at \u22121\u00a0V versus RHE. Although the phen-Cu has no anchoring groups in the structure, it could be observed that the Cu complex can be reversibly accumulated close to the graphene surface by controlling the potential. This indicates the significance of interaction between molecular catalysts and carbon materials.\n72\n\nFukuzumi and co-workers explored a Co(II) chlorin complex as an efficient electrocatalytic CO2 reduction catalyst. The Co(II) chlorin complex was adsorbed on MWCNT and through sonication and drop-casting placed on a GC electrode (Figure\u00a023E). The resulting electrode efficiently reduces CO2 to CO in H2O (pH\u00a0= 4.6) at an applied potential of \u20141.1\u00a0V versus NHE with a 89% faradaic efficiency for CO formation, with the remaining electrons consumed for H2 evolution.\n128\n The same technology was efficiently extended to photoelectrocatalytic reduction by using cobalt(II) chlorin complexes adsorbed on MWCNT as CO2 reduction catalyst, and introduction of [RuII(Me2phen)3]2+ as photocatalyst. This photoelectrochemical system yielded CO and H2 with a ratio of 2.4:1 and provided a high TON of 710 in acetonitrile solution containing 5 v/v%water.\n131\n The cathodes with cobalt(II) chlorin on MWCNT have also been applied in devices with the surface-modified FeO(OH)/BiVO4/FTO photoanode.\n132\n The photocathode in this device showed a selective photoelectrocatalytic reduction of CO2 to CO in H2O (pH 4.6) with 83% faradaic efficiency at \u22121.3\u00a0V bias. These works provided a unique strategy for the production of syngas from CO2 and H2O under (photo)electrochemical conditions.The last example discussed for the immobilization of CO2 reduction catalysts is a report that clearly underlines one of the challenges in employing \u03c0\u2013\u03c0 interactions as method of immobilization. Flat catalysts like the CoPc can utilize their extended \u03c0-systems to interact with the carbon support, but they can also interact with each other. For example, the highly soluble and sterically hindered cobalt(II) octaalkoxyphthalocyanine (CoPc-A) shows enhanced catalytic activity, compared to its analogous CoPc, after being successfully immobilized on graphene via \u03c0\u2013\u03c0 stacking (Figure\u00a023F). The alkoxy substitutions on the heterocycle periphery in CoPc-A help to suppress molecular aggregation on the graphene surface, leading to increased accessibility of active sites and a significantly enhanced CO conversion catalytic activity (\u223c5 s\u22121 at 480\u00a0mV overpotential) compared to its analogue CoPc (\u223c2 s\u22121 at 590\u00a0mV overpotential). The introduction of alkyl groups also increased the activity and long-term stability for CO production of the CoPc-A electrode, providing a \u223c6 s\u22121 TOF and 6.7\u00a0\u00d7 105 TON for CO evolution over 30\u00a0h of electrolysis.\n65\n\nThe misunderstanding of the poor stability of molecular catalysts is mainly due to the short lifetime of photocatalysis, photoelectrochemical cells and other lab-scale electrodes currently developed with molecular catalysts. Obviously this is an inappropriate conclusion, because stability of molecular catalysts does not necessarily translate into the stability of devices (Intrinsic stability of molecular catalysts \u2260 Stability of devices).\n1\n The stability of (photo)electrochemical devices has many influencing factors (e.g., detachment of electrode films, decomposition of light absorbers, desorption of molecules, mass transfer problems, and selection of electrolyte solutions), and no systematic in-depth research has proved that catalyst decomposition commonly leads to the performance loss of devices. To the best of our knowledge, there is no specific theory or fundamental that indicates molecular catalysts are less stable than inorganic materials. Actually, many molecular catalysts have been reported with lower overpotential, and much higher TOFs and TONs,\n11\n\n,\n\n21\n\n,\n\n23\n which are scientifically reasonable indicators that represent the true intrinsic activity and stability of the catalysts.\n133\n In fact, the stability of molecular catalysts has been seriously misunderstood and underestimated at present. It is an accepted dogma that molecular catalysts have no chance to be employed in commercially applicable electrolysis techniques due to their low stability.\n24\n\nIt must be also noted that the observed stability of a catalyst depends on the measurement methods, even for inorganic catalysts. For example, large losses in water-oxidation activity of Ir nanoparticles were observed when they were evaluated in rotating disk electrode (RDE) half-cells (It is also common system for lab-scale testing),\n134\n but as known the durability of Ir catalysts is not an issue when they were used in PEM electrolyzer.\n6\n In our opinion the less-developed testing systems and cell-designs in lab-scale studies could be an essential reason that induce the decomposition of the modified molecular catalysts (cf. our recent review paper\n1\n for more in-depth discussions and examples). Indeed, research on developing molecular catalysts and testing molecular catalysts on electrodes are just at beginning stages; much more studies are required before the arrival of an efficient and applicable device.In addition, the current stability comparison between molecular catalysts engineered electrodes and inorganic materials based electrodes is unequal, if one considers the loading amount of catalysts. Under the commonly employed evaluation system, the loading amount of molecular catalyst is in a scale of nmol\u00b7cm\u22122.\n48\n\n,\n\n49\n\n,\n\n81\n\n,\n\n96\n\n,\n\n104\n In contrast, it is usually in a scale of \u03bcmol\u00b7cm\u22122 for inorganic catalyst, which is three orders of magnitude higher.\n135\u2013137\n The far higher catalyst loadings may mask durability losses for a relatively long period, e.g., a few hours.\n133\n\n,\n\n134\n To exemplify, detachment or decomposition of 1\u00a0nmol of catalyst has far greater implications for systems with lower loading than for electrodes with 1\u00a0\u03bcmol of catalyst on the surface where this loss only accounts for 0.1%. Actually, there are already examples of molecular catalysts based electrodes with only several nmol\u00b7cm\u22122 catalyst loading that reach the stability of a few hours, which is common time-frame for evaluating stability of electrodes on laboratory scale.\n48\n\n,\n\n96\n\n,\n\n138\n Therefore, with consideration of loading amounts of catalysts, the observed stability of electrodes may not be a real disadvantage and molecular catalysts have a great chance to be superior. The high variability of molecular catalysts in terms of structural design is also advantageous for regulating and further improving stability.Compared with traditional alkaline water electrolysis, PEM electrolysis has many principle advantages. But as a comparably young technology, PEM technology has many aspects that need further development. Noble metal catalysts, irreplaceable PEM, and corrosion-resistant current collection plates primarily contribute to the high price of PEM electrolyzers. With the increase of market demand and commercialization process, the degree of patent opening, we believe the replacement of proton exchange membranes and other components for the reduction of costs are expectable. However, the most unavoidable and most challenging part is the replacement of rare noble metal catalysts. We discussed in this review that molecular catalysts have a potential in breaking this bottleneck. We summarized the reported molecular catalyst immobilized on carbon materials for catalytic water oxidation, hydrogen evolution and CO2 reduction, as these materials and the involved immobilization strategies as well as characterization techniques can facilitate the advance of a promising but underdeveloped technology, the molecular catalysts integrated PEM/AEM electrolyzers for solar fuel production. From the blueprint to practical implementation of this technology, the following research roadmap and tasks are proposed:First, evaluate state-of-the-art WOCs, HER, and CO2RR catalysts in PEM/AEM electrolyzer setup. The possibilities and advantages of using molecular catalysts in PEM/AEM electrolyzers should be explored and further verified. Molecular catalyst modified carbon materials summarized in this review can be directly tested by replacing noble metal catalysts in the construction of PEM/AEM electrolyzers. The introduced concepts and strategies for modifying molecular catalysts on carbon materials can also be used to design molecular catalyst based PEM/AEM electrolyzers. Based on the preliminary investigations discussed in the introduction, it is expectable to realize the aim in this stage, achieving current density in A/cm2 with comparable or lower overpotential compared to electrolyzers based on noble metal catalysts.Second, increase catalyst loading for higher current density. Most of the reported molecular catalysts modified materials are in the form of single molecule layer. These methods have very limited catalyst loading. More reliable methods are required to integrate molecular catalysts into bulk materials to increase catalyst loading, e.g., molecular catalyst engineered polymers, metal\u2013organic frameworks (MOFs), and covalent organic frameworks (COFs). Exposure of catalysts, charge, and mass transports should be considered in the material design. A proper amount of catalyst loading is important for obtaining higher activity. In addition, the coating of the carbon material with the catalyst assembly may prove advantageous to improve overall stability under the harsh oxidative conditions commonly found at the anodic site in PEM electrolyzers.Third, challenge the long-term stability. Stability of an electrolyzer depends on many factors, including the configuration of the cell, stability of current collectors, stability of the MEA, stability of catalyst loading, and the intrinsic stability of catalysts. These factors may also influence each other. Although the intrinsic stability of molecular catalyst has been demonstrated above, this does not necessarily represent good stability of the constructed electrolyzer. Catalytic behaviors and deactivation pathways of molecular catalyst based MEA should be deeply studied to find strategies for obtaining good stability of molecular catalysts based electrolyzers. For long-term operation of the electrolyzer, the stability of substrates for catalyst loading is also important, requiring special considerations. Carbon based materials discussed in this review are promising for loading reduction catalysts. For anodes, common carbon materials display long term-stability to high oxidation potentials, oxygen environments, and acidic conditions. More advanced conductive substrates showing long-term stability under acidic and oxidative conditions, should be developed for loading WOCs. Sintered Ti particles, which form current collectors for PEM electrolyzer, pre-oxidized MWCNTs, and special conductive metal oxides, e.g., TiO2, SnO2, and Ta2O5, can be potential substrates for supporting molecular catalysts.At last, more intrinsically efficient and low-cost molecular WOCs and HER catalysts operating under acidic conditions should be developed. In addition, for CO2RR catalysts high selectivity and efficiency for production of carbon-based fuels are required. Even though the application of molecular catalysts potentially can reduce the overall high resource requirements due to lower loadings, the terawatt-scale requirement of renewable fuels would still only be reachable with low-cost and earth-abundant-based catalysts as ultimate ideal catalysts for real practical electrolyzers.We acknowledge financial support for this work from the Swedish Research Council (2017-00935), Swedish Energy Agency, Knut and Alice Wallenberg Foundation, and National Basic Research Program of China (973 program, 2014CB239402). L.F., T.L., and Q.M. also thank the China Scholarship Council for a special scholarship award.B.Z. and L.S. proposed the topic of the manuscript. B.Z., L.F., R.B.A., and T.L. wrote the manuscript. B.Z., B.J.J.T., and L.S. revised and edited the original draft. All authors discussed and revised the final manuscript.", "descript": "\n                  Molecular catalysts possess numerous advantages over conventional heterogeneous catalysts in precise structure regulation, in-depth mechanism understanding, and efficient metal utilization. Various molecular catalysts have been reported that efficiently catalyze reactions involved in artificial photosynthesis, however, these catalysts have been rarely considered in view of practical applications. With this review, firstly we demonstrate in the introduction that molecular catalysts can bring new opportunities to proton exchange membrane (PEM) electrolyzers. In the following parts, we provide an overview of molecular catalyst modified carbon materials developed for electrochemical water oxidation, proton reduction, and CO2 reduction reactions. These materials and the involved immobilization strategies as well as characterization techniques may be directly employed in the investigations of application of molecular catalysts in PEM electrolyzers. The future scientific perspectives and challenges to advance this promising, yet underdeveloped technology for solar fuel production, integrating PEM electrolyzer with molecular-level catalysis, are discussed in the conclusions.\n               "}
{"full_text": "Data will be made available on request.Nanomaterials based on oxides (nanostructure and nano dispersed) are a diverse class of materials in terms of electronic structure, physical, chemical, and electromagnetic properties. The application of metal oxide nanomaterials and nano composition based on them is becoming increasingly popular in applied ecology especially; where they can be used as absorbents and photocatalysts as well as a material for the manufacture of environmental monitoring devices. The nano-sized metal oxide-based absorption materials have a large specific affinity for various contaminants. The metal oxide nanomaterials and their nano compositions are based on TiO2, ZnO, SnO2, ZrO2, and Fe3O4 for environmental application [1].Titanium dioxide (TiO2) is considered to be one of the most attractive semiconductor photocatalysts owing to its long-term stability, nontoxicity, and excellent photocatalytic property. On one hand, the wide band gap nature of titania (3.2\u00a0eV for the anatase structure or 3.0\u00a0eV for the rutile structure) makes it absorb only ultraviolet (UV) light, which limits the effective usage of solar light. In spite of this, a low quantum yield of TiO2 is present as a result of its high recombination rate of photogenerated electrons and holes [2]. Doping TiO2 with foreign ions is a promising approach to extend its response to the visible-light region. Furthermore, doping TiO2 with multiple dopants has become an effective and promising approach to improve the photocatalytic performance of TiO2\n[3]. As one of the earliest studied n-type semiconductor photocatalysts, TiO2 has been widely used in environmental purification, self-cleaning, H2 production, photosynthesis, CO2 reduction, organic synthesis, solar cells, etc. Therefore, the number of publications on TiO2 has increased exponentially in recent decades [4].TiO2 has also been used as a gas sensor, which helps to sense the amount of oxygen (or reducing gas) present in the atmosphere. Another application is found in car engines for controlling fuel consumption and environmental pollution [5\u20138].In order to use solar energy more efficiently, most of the investigations have been focused on preparing TiO2 sensitive to visible light during the past several years. Many ionic dopants in different valence states have been investigated, including both metallic (e.g) Ca2+, Sr2+, Ba2+, Al3+, Ga3+, Cr3+, Fe3+, Co3+, Ce3+, Sn4+ and non-metallic ions (e.g) N3+, C4+, S4+, F and particularly TiO2 doped with Sn, has proved to be an effective method and has been widely studied [9,10]. SnO2 is an important n-type wide band gap semiconductor and SnO2-based nanostructures act themselves as one of the most important classes due to their various tunable physicochemical properties [9]. In some of the reports, sn-doped TiO2 was prepared by the vapour transport method of water molecules [11]. It would effectively control the rate of hydrolysis of Ti4+ by adjusting the flow speed of vapour. Sn-doped TiO2 particles with different concentrations and photocatalytic test show all the sn-doped TiO2 have higher photocatalytic activity than that of pure TiO2\n[11].Although the mono-doped non-metal (or) metal atoms can obviously enhance photocatalytic performance of TiO2, they always act as the recombination centre because of the partially occupied impurity bands. The passivity of the impurity band while co-doping two or more foreign atoms have been recognized theoretically and this decreases the formation of recombination centres by increasing the solubility limit of dopants. Furthermore, co-doping can also modulate the charge equilibrium. Consequently, co-doping can effectively enhance photocatalytic activity. Based on the research on the doping effect, two (or) more elements are introduced into the TiO2 lattice to check the chance of electronic structure and band gap energy, including non-metal [12\u201314] and non-metal and metal atoms [15\u201317], and metal and metal atoms [18\u201320]. The modified N-doped TiO2 usually shows a favourable effect on improving the activity in the range of visible light compared to C-doped TiO2\n[21,22], (C, F) co-doping [23], and (C, S) co-doping [24\u201327]. Recently, many researchers found that tri doping in TiO2 crystals is an effective supplemental tool [28\u201332]. This implies that (C, F) co-doping might be a kind of efficient way to improve the photocatalytic activity of TiO2\n[33].Although mono-metal doping can improve the band gap structure, the serious recombination centres also deteriorate carrier transport due to their partially occupied impurity bands. Hence many research want to improve carrier transport through co-doping with two different metal elements such as (Ti, Ni) co-doping [34], (Ti, Fe) co-doping [35], (Ag, W) co-doping(Ag, or) co-doping [36], (Cu, V) co-doping [37] and (Fe, Co) co-doping [38]. Recently S-metal co-doping has also been widely investigated, such as (S, Fe) co-doping [39] and (S, Cu) co-doping [40]. Compared to undoped TiO2, (S, Fe) co-doped TiO2 showed a higher photocatalytic activity under both UV and visible light irradiation, and the optimal methyl blue degradation level was about 96.92\u00a0% [29,41].Azo dyes are the largest class of dyes used in industry. These compounds are characterized by the presence of one (or) more Azo bonds (\u2013NN\u2013) in their molecule that are associated with one (or) more aromatics structures. The presence of Azo dyes in textile effluents makes them particularly harmful to the environment and human health. In fact, their release into aquatic ecosystems may lead to a reduction of sunlight penetration and dissolved oxygen concentration, with deleterious effects on local flora and fauna [42].Herein, we report that the Ni-co-doped SnO2-TiO2 powders were prepared and characterized by means of XRD, SEM-EDX, HRTEM, XPS, XANES and photocatalytic abilities of Ni-deposited SnO2-TiO2 nanocomposite was evaluated. The photocatalytic activity of TiO2, SnO2 and SnO2-TiO2 has been enhanced by Ni complex deposited on SnO2-TiO2.NiCl2\u00b76H2O, EtOH, Imidazole, Titanium isopropoxide (TTIP, 98\u00a0%), Hydrogen peroxide (H2O2, 30\u00a0wt%), Double distilled water, Azo dye, Methyl Orange, Whatman filter paper-41. All AR-grade chemicals were purchased from Qualigens.The IR spectra of the complex and prepared nanocompiste were recorded on a Thermo Nicollet \u22126700 FT-IR instrument (KBr pellet technique). The UV\u2013Visible spectra were undertaken in an aqueous medium at room temperature on Shimadzu model UV-2450 double beam spectrophotometer with a quartz cell. A crystallography study was performed on the prepared nanocomposite by rotating anode diffractometer (XRD-PHILIPSPANALYTICAL, Netherland) using Cu-k\u03b1 irradiation. The HRTEM images were taken on JEOL JEM-2000EX microscope. X-ray photoelectron spectra (XPS) were measured at the beam line-14, Raja Ramanna Centre for Advanced Technology (RRCAT), Indore, India (Ref. IBR/3850/2022\u201305-07/INDUS-2/BL-14 XPS) and the binding energy (B.E.) was calibrated using contaminated carbon as an internal standard (C 1s B.E. 284.6\u00a0eV). Ni K-edge X-ray absorption fine structure (XAFS) spectra (Ref. IBR/3823/2022-04-07/INDUS-2/BL-9 Scanning EXAFS) were taken at the beam line-09, RRCAT, Indore, India. Ni K-edge XAFS of Ni-foil and Ni-SnO2-TiO2 measured by fluorescence mode, while Ni K-edge XAFS of NiO and Ni-precursor complex measured by transmittance mode. Steady state fluorescence emission spectra were recorded on Spex FluoroLog-3 spectrofluorometer (Jobin-Yvon Inc.) using 450\u00a0W xenon lamp and equipped with a Hamamatsu R928 photomultiplier tube. The photocatalytic activity of the samples were evaluated by photodegradation of MO using HEBER Visible Annular Type Photoreactor, model HVAR1234 (Heber Scientific, India), under visible light irradiation using 300\u00a0W Tungsten lamp as a light source.Ni(II)-imidazole complex was prepared by the reported method [43]. Dissolve 6.0\u00a0g of NiCl2\u00b76H2O in 3\u00a0mL. of EtOH. A little warming improves the rate of dissolution. Cool the solution in ice while adding 5.0\u00a0g (5.6\u00a0mL) of imidazole in EtOH. Add the imidazole slowly because the reaction is quite exothermic. Cool and add 15\u00a0mL of cold ethanol to initiate crystallization. Keep cold for 10\u00a0min and collect the product on a Buchner funnel and wash it with two 5\u00a0mL portions of ethanol. Dry in the air.The synthesis of TiO2 nanoparticles was reported and characterization details are given in reference [44]. The reported synthesis method is as follows: titanium isopropoxide (TTIP, 98\u00a0%) and an aqueous solution of hydrogen peroxide (H2O2, 30\u00a0wt%) were used as the starting materials. In a typical experiment, TTIP was added to H2O2 in a 500\u00a0mL beaker with vigorous magnetic stirring at a 70\u00a0\u00b0C water bath for 15\u00a0min in a basic medium (adjust pH using ammonia solution). The solution was made concentrated and the resultant yellow gel was put into a porcelain crucible with a cap and dried at 80\u00a0\u00b0C and then TiO2 nanoparticles were obtained.The preparation process was reported in our group\u2019s previous paper. The same material has been taken for further nickel co-doping process [43].In a typical preparation, the prepared sample of the SnO2-TiO2 nanoparticle (200\u00a0mg) was dispersed in 80\u00a0mL of water in a 250\u00a0mL beaker. The prepared Ni(II)-imidazole complex was dissolved in 20\u00a0mL of distilled water it become (10-3 M) concentration and the basic medium was maintained. The latter solution was added to the SnO2-TiO2 suspension, and the mixture solution was stirred for one hour and put the suspension in a water bath for one hour at 80\u00a0\u00b0C. The resultant suspension was filtered and washed thoroughly using water and ethanol, followed by acetone. The resultant solid sample was collected and dried at 100\u00a0\u00b0C in the oven for two hours.\n\n\n(i)\nTypically Azo dye was added to 1000\u00a0mL of double distilled water and used as a stock solution. About 10\u00a0mg of TiO2, SnO2-TiO2 or Ni-SnO2-TiO2 samples were added to 100\u00a0mL of Azo dye solution. A control was also maintained without the addition of catalyst. Before exposure to irradiation, the reaction suspension was magnetically stirred well for about 30\u00a0min to make up the equilibrium of the working solution. Afterward, the dispersion was put under the sunlight and monitored from morning to evening sunset. At specific time intervals, aliquots of 2\u20133\u00a0mL suspension were filtered and used to evaluate the photocatalytic degradation of dye. The absorbance spectrum of the supernatant was subsequently measured using a UV\u2013Visible spectrophotometer. The concentration of dye during degradation was calculated by the absorbance value at 660\u00a0nm.\n\n\n(ii)\nIn a typical process, 80\u00a0mL of aqueous solution of methyl orange (MO) with initial concentration 2.5\u00a0\u00d7\u00a010-6 M and 100\u00a0mg of catalyst were taken in a cylindrical-shaped glass reactor at room temperature in air and at\u00a0\u223c\u00a07 pH conditions. The flow rate of air was kept at a constant value of 80\u00a0mL\u00a0min\u22121. The mixture solutions were kept in dark for 30\u00a0min before irradiation. Furthermore, prior to irradiation, the mixture solution were continuously aerated by a pump to provide oxygen and for complete mixing. The samples (4\u00a0mL) were taken out every 10\u00a0min, and were analyzed by UV\u2013vis spectrophotometer. The photodegradation efficiencies (PDE) were calculated via the formula PDE = (A0 - At/A0)\u00a0\u00d7\u00a0100\u00a0%, where A0 is the absorbance of initial MO solution and At is the absorbance of MO solution measured at various irradiation time at 465\u00a0nm.\n\n\nTypically Azo dye was added to 1000\u00a0mL of double distilled water and used as a stock solution. About 10\u00a0mg of TiO2, SnO2-TiO2 or Ni-SnO2-TiO2 samples were added to 100\u00a0mL of Azo dye solution. A control was also maintained without the addition of catalyst. Before exposure to irradiation, the reaction suspension was magnetically stirred well for about 30\u00a0min to make up the equilibrium of the working solution. Afterward, the dispersion was put under the sunlight and monitored from morning to evening sunset. At specific time intervals, aliquots of 2\u20133\u00a0mL suspension were filtered and used to evaluate the photocatalytic degradation of dye. The absorbance spectrum of the supernatant was subsequently measured using a UV\u2013Visible spectrophotometer. The concentration of dye during degradation was calculated by the absorbance value at 660\u00a0nm.In a typical process, 80\u00a0mL of aqueous solution of methyl orange (MO) with initial concentration 2.5\u00a0\u00d7\u00a010-6 M and 100\u00a0mg of catalyst were taken in a cylindrical-shaped glass reactor at room temperature in air and at\u00a0\u223c\u00a07 pH conditions. The flow rate of air was kept at a constant value of 80\u00a0mL\u00a0min\u22121. The mixture solutions were kept in dark for 30\u00a0min before irradiation. Furthermore, prior to irradiation, the mixture solution were continuously aerated by a pump to provide oxygen and for complete mixing. The samples (4\u00a0mL) were taken out every 10\u00a0min, and were analyzed by UV\u2013vis spectrophotometer. The photodegradation efficiencies (PDE) were calculated via the formula PDE = (A0 - At/A0)\u00a0\u00d7\u00a0100\u00a0%, where A0 is the absorbance of initial MO solution and At is the absorbance of MO solution measured at various irradiation time at 465\u00a0nm.The theoretical calculation to determine the electronic density of states for dopant induced changes in Ni-Sn-TiO2 based on the Density Functional Theory (DFT) were carried out using the Vienna ab initio Simulation Package (VASP) with projector augmented wave (PAW) pseudopotentials and Generalized Gradient Approximation (GGA) to Perdew, Burke and Ernzerhof (PBE) as exchange correlation energy. The geometrical structure relaxation performed for pristine Sn-, Ni-doped TiO2 using the conjugate-gradient algorithm (CGA). Monkhorst-Pack method was used to obtain the plane wave cutoff and the k-point density.\nFig. 1\n shows XRD patterns of prepared sample. The narrow and intense peaks at 2\u03b8\u00a0=\u00a026.9\u00b0, 38.1\u00b0, 53.8\u00b0, and 55.4\u00b0 are the standard X-ray diffraction peaks of anatase TiO2 (Table 1\n), while the narrow and intense peaks at 2\u03b8\u00a0=\u00a027.4\u00b0, 34.9\u00b0, and 52.7\u00b0 demonstrate that the phase structure is rutile TiO2. The diffraction pattern for pure SnO2 show two main peaks at 2\u03b8\u00a0=\u00a027.2 and 34.4 \u00b0 that referred to SnO2 with rutile (cassiterite tetragonal) structure (ICDD card No. 41\u20131445) of space group P42/mnm. Moreover, there is a diffraction peak corresponding to SnO2 in patterns of the prepared sample. The XRD reflection indicates the presence of anatase phase more predominantly, the peak at 27.4\u00b0 2\u03b8 corresponds to a rutile (110) reflection near that of pure TiO2 (27.4\u00b0 2\u03b8). The rutile (110) reflection appears at \u223c27\u00b0 as 2\u03b8, which is intermediate to that of the pure TiO2 (27.4\u00b0, 2\u03b8) and pure SnO2 (26.6\u00b02\u03b8) positions [44]. The impurity phase of NiO is not observed, it may be because of very limited Ni loaded or Ni substituted in lattice site. XRD result indicates that cassiterite of SnO2 coupled with anatase\u2013rutile of TiO2.\nFig. 2\n depicts FTIR spectrum of Ni-SnO2-TiO2 sample. The entire spectrum display-two characteristics broad band centred at 3408 and 1627\u00a0cm\u22121 which are assigned to the stretching and bending modes of vibrations of physical adsorbed water on titania surface or to hydroxyl groups exist on the surface of the oxides, respectively. A remarkable broad band in the region 680\u2013400\u00a0cm\u22121 is associated with the stretching modes of vibrations of bridged sn-O-Sn, Ti-O-Ti, Ni-O-Ti, and Ti-O-Sn bonds. It should be emphasizing to notice two weak bands at 1389 and 1041\u00a0cm\u22121 which are assigned to the hetero Ti-O-Sn bonds. FTIR spectrum confirmed that SnO2 coupled with TiO2.The UV\u2013vis measurements of TiO2, SnO2-TiO2, and Ni-SnO2-TiO2 are shown in Fig. 3\n. The absorption spectrum of TiO2 consists of a single broad intense absorption around 400\u00a0nm due to the charge transfer from the valence band (mainly formed by 2p orbital of the oxide anions) to the conduction band (mainly formed by 3d t2g orbitals of the Ti4+ cations). The adsorbed TiO2 shows absorbance in the shorter wavelength region while Ni-SnO2-TiO2 and the UV\u2013vis results show a red shift in the absorption onset value in the case of Ni-added tin-titania composite. The adsorption of various transitional metal ions into TiO2 shifts its optical absorption edge from UV into visible light range, but no prominent change in the TiO2 band gap is observed.The SEM images of Ni-SnO2-TiO2 particles are presented in Fig. 4\n. The micrographs show that most of the particles are spherical with various size distributions. The Ni implanted particles are found to be highly agglomerated and some are non-spherical in characteristics. These differences in morphology, particle shape, and size indicate the presence of Ni and SnO2 species in TiO2 nanoparticles. From SEM images, it can be presumed that nickel and SnO2 deposited titania samples consist of very fine free individual grains plus other granular aggregates: very large distributions of shape and dimensions of the particles are observed, with an average diameter of 110\u00a0nm. It is evident that the growths of particles are highly restricted due to Ni deposition, which is of significant importance not only for the design of surface properties but also for tuning the electronic structure of the semiconductor. Ni and Sn were loaded in titania, which were observed from the Energy X-ray Diffraction study. EDX spectra are shown in Fig. 5\n and Table 2\n presents the concentration of Ni and Sn loaded on the solid. Sn is present in a considerable amount of TiO2. Table 2 exhibits the relationship between the density of Ni and Sn in atomic percent (at %) in titania, in which Ni is loaded 0.39 at % and Sn is loaded at 13.11 at% in TiO2. Ti and Sn are present in the nearly same ratio in prepared samples, which means that nearly 13 at% of each Ti and Sn is present in a prepared sample. It indicates that nearly 50\u00a0% of SnO2 and 50\u00a0% of TiO2 are present in the prepared nanocomposite. The EDX confirms the existence of nickel in the SnO2-TiO2 powder.\nFig. 6\n(a) shows the TEM images with a diameter of 5\u201310\u00a0nm were observed in Ni-SnO2-TiO2 sample. The selected area electron diffraction (SAED) on prepared sample (inset Fig. 6(a)) shows the strong Debye-Scherrer rings and additionally complicated bright spots were observed, indicating the coexistence of polycrystalline anatase\u2013rutile or rutile-casseterite crystallites. This is in well agreement with the aforementioned conclusion that the Sn4+ ions doped in TiO2 can promote the formation of new phases. Fig. 6(b) displays the HRTEM images of prepared sample. For the anatase structure of TiO2, the fringe spacing (d) of (101) crystallographic plane is determined to be 3.55\u00a0\u00c5 [45]. Furthermore, a fringe spacing of \u223c3.33\u00a0\u00c5 corresponding to the (110) planes of rutile TiO2 are observed in prepared sample. The crystallographic planes of SnO2 cassiterite (101) fringe spacing observed at 2.64\u00a0\u00c5, which were further confirmed that SnO2 was coupled with TiO2 in prepared sample. No fringe spacing of NiO or Ni-precursor complex observed but amorphous phase can observed surface site of prepared sample.XPS analysis was performed to further study the chemical states of Sn in TiO2. Fig. 7\n shows the survey, Ti 2p, and Sn 3d XPS spectrum of prepared sample. It can be seen that XPS peak positions of Ti 2p3/2 locate at 448.31, 449.36, 454.31, and 453.53\u00a0eV, which indicates that Ti element mainly existed as the chemical states of Ti4+. The doublet peaks observed at 476.38 and 484.70\u00a0eV in the Sn 3d XPS spectrum, it is ascribed to Sn 3d5\n\n/\n\n2 and Sn 3d3\n\n/\n\n2 of the substitutional Sn4+ dopants in the lattice, since the peak position of Sn 3d5\n\n/\n\n2 (484.70\u00a0eV) is located between that of SnO2 (486.6\u00a0eV) and metallic Sn (484.0\u00a0eV) [38]. The XPS analysis proved that the Sn4+ easy to replace Ti4+ in the lattice of TiO2\n[46]. From survey XPS spectrum indicates that very minimum amount of Ni species (0.40 at%) presented on surface of our prepared sample.To further study the coordination structures of Ni over SnO2-TiO2 nanocomposites by XANES study of our sample. The XANES spectra of the pure NiO, Ni-precursor complex and our prepared nanocomposite are shown in Fig. 8\n. Their spectra revealed the well-defined pre-edge peaks at 8346.31\u00a0eV, the main characteristic shoulder features for NiO and Ni-precursor complex at 8392.56\u00a0eV, which are attributed to Ni2+, while little shift was observed at 8398.28\u00a0eV for Ni-SnO2-TiO2 sample. This indicates that Ni is not just deposited in the form of NiO or Ni-precursor complex in Ni-SnO2-TiO2 sample. This result clearly indicated that the ionic state for Ni in Ni-SnO2-TiO2 is\u00a0+\u00a02. Moreover, it could be seen that the spectral shape and features of Ni-SnO2-TiO2 is intermediate to that of NiO and Ni-precursor complex suggesting that small amount of Ni species surrounded with O, N, or C is coupled with SnO2-TiO2 nanocomposite. The structural, chemical and morphological characterizations performed by XRD, XPS, XANES, SEM-EDX, and HRTEM microscopes clearly confirm the octahedral substitution of Sn4+ for Ti4+ and additionally SnO2 and small Ni2+ species are coupled with TiO2 in Ni-SnO2-TiO2 nanocomposite.The aim is to increase in the electron deficiency and a breakdown of the azo molecules. Ni(II) loaded photocatalyst was used to study the photodegradation of Azo dye in an aqueous medium under sunlight irradiation. The effectiveness of the degradation of this substrate by the prepared sample is compared with the photodegradation yields of the substrate by pristine TiO2. From Figs. S1-S2 the UV\u2013vis spectra of azo dye at before and after irradiation finds that a gradual decrease in the absorption bands of an azo dye, when the irradiation time increases, is due to the decomposition of azo molecules. The predominant peak is observed at 583\u00a0nm for azo dye before irradiation (Table S1) but this peak completely disappears after irradiation of dye solution using Ni-SnO2-TiO2 as photocatalyst (Table S2). This indicates an increase in electron deficiency and a breakdown of the azo molecules. The PDE of Azo dye versus irradiation times is shown in Fig. 9\n (Tables 3-5\n\n\n) using pure TiO2, SnO2-TiO2, and Ni-SnO2-TiO2, respectively. TiO2 presents a fair activity for the degradation of azo under sunlight irradiation. The photodegradation rate is been determined for each experiment and the highest values are observed for SnO2-TiO2 and Ni-SnO2-TiO2. Comparatively, the photocatalyst Ni-doped SnO2-TiO2, which is prepared by us, has very highest photocatalytic efficiency. The complete mineralization was confirmed by total organic carbon (TOC) analysis, and COD measurement. This indicates that Ni loaded has improved the photocatalytic behavior of TiO2 as well as SnO2-TiO2 photocatalysts under sunlight irradiation.\nFig. 10\n shows the UV\u2013vis spectra of photodegradation of MO with the Ni-SnO2-TiO2 sample suspended in water at different irradiation times. It can be found that gradually decreases the absorption bands at 268 and 464\u00a0nm, when the irradiation time increases due to oxidation of the MO molecules, which indicates an increase in the electron deficiency and a break-down of the MO molecules. The photocatalytic activities of the pure TiO2, SnO2, SnO2-TiO2, and prepared Ni-SnO2-TiO2 nanocomposites were evaluated by the degradation of MO under visible light irradiation, and the photoefficiencies were shown in Fig. 11\n. Present reported nanocomposites showed the better photocatalytic performances than above reference samples under visible light irradiation. The photocatalytic mechanism of SnO2-TiO2 was reported in our previous works [45,46]. similar mechanism could work in Ni-SnO2-TiO2 nanocomposite, in addition Ni species influence the photoefficiency of SnO2-TiO2 nanocomposite.The photoluminescence (PL) spectra of TiO2 and Ni-SnO2-TiO2 samples with excitations at 330\u00a0nm are presented in Fig. 12\n. Broad emission in the spectral range from 350 to 600\u00a0nm was observed as well as the presence of well-resolved peaks/ shoulders at 468, 483, 494 and 560\u00a0nm. A addition of Sn and Ni in TiO2 was increase the intensity or changes the shape and peak position of the PL spectra compared to spectrum of pure TiO2 sample. It was found that the steady state emission spectra contain a narrower UV emission located near the position of 390\u00a0nm (3.18\u00a0eV) and a widened emission range from 450 to 491\u00a0nm [45\u201348]. The strong green emissions were observed in both TiO2 and Ni-SnO2-TiO2 at 560\u00a0nm (2.22\u00a0eV) in Fig. 12. Therefore, emissions likely originate from surface defects, such as ionizable oxygen vacancies and the recombination of self-trapped excitons (STEs) localized within TiO6 octahedra [44]. The green emission increase in Ni-SnO2-TiO2 sample as compare to TiO2, it indicates that our prepared sample contain more surface defect sites. The defect concentrations and life times of photoexcited species plays an important role in photocatalysis, investigation of photocatalysts through PL spectroscopy is important to obtain critical reasons behind the enhanced photocatalytic activity. Moreover, the contributions of charges from the dopants to the electronic states are clearly observed from the density functional theory calculations demonstrated in Fig. 13\n. These effects are occurring in nanocomposite system as a consequence of charge trapping on surfaces or the inter phases between the two oxide phases [46]. Hence, the PL and DFT studies supported for photocataltic activities enhanced through Ni and Sn presented in TiO2.A first attempt was made to synthesize of Ni-SnO2-TiO2 nanocomposite. [Ni(Im)6]Cl2 is preferably adsorbed on the solid surface forming surface: complex adduct, simultaneously it becomes surface species; SnO2-TiO2:Ni(II). The surface species was identified by SEM-EDX, HRTEM, XPS and XANES measurements. From the DRS spectra, we could conclude that the absorption of composite particles expands to visible region that are able to be excited by visible light. The organic model pollutant of azo dye degradation property of TiO2 photocatalyst under visible light irradiation has been greatly enhanced after the introduction of Ni into the SnO2-TiO2 surface matrix. The present work considered that the low-level Ni-loaded anatase\u2013rutile- cassiterite (SnO2-TiO2) induced photocatalytic behaviours due to the creation of surface synergetic effect and defect sites. Various spectroscopic measurements indicated the presence of structural, optical, and surface composition in the Ni-Sn\u2013Ti system. The photocatalytic properties of Ni-SnO2-TiO2 sample was tested, and the results showed that a small amount of nickel co-doped SnO2-TiO2 which preferably formed anatase\u2013rutile-cassiterite mixed phases had improved photocatalytic activity due to the more efficient separation of photoinduced electrons and holes on its surface. The major important finding in this report: (i) the formation of sn-doped and Ni complex deposited in surface of anatase\u2013rutile-cassiterite mixed phases, no pure NiO was found, (ii) the enhancement of light absorption property in the visible region and shift of absorption edge to the long wavelength side, and (iii) Ni content in SnO2-TiO2 assisted optical character and improve the photocatalytic property. This study could point out a potential way to develop new and more active nickel complex deposited and tin doped titania photocatalysts for wastewater treatments.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Dr. ASG is thankful to National College (Autonomous), Tiruchirappalli, Tamil Nadu for financial support through a college minor research project scheme (No. NCT/SEC/010/2022-2023/19-07-2022). Authors are grateful to Raja Ramanna Centre for Advanced Technology (RRCAT) (Ref.: IBR/3850/2022-05-07/INDUS-2/BL-14 XPS and IBR/3823/2022-04-07/INDUS-2/BL-9 Scanning EXAFS) and thanks to Dr. S. N. Jha, Dr. R. K. Sharma, and Dr. Jaspreet Singh at RRCAT, Indore, India and Dr. D. Bhattacharyya,\u00a0\u00a0Bhabha Atomic Research Centre (BARC), Mumbai, India for supported EXAFS and XPS studies.Supplementary data to this article can be found online at https://doi.org/10.1016/j.rechem.2022.100557.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  A first attempt was made to synthesize a Ni-SnO2-TiO2 nanocomposite by a one-pot simple synthetic method. A [Ni(im)6]Cl2 precursor complex is preferably adsorbed on the solid surface forming nanocomposite: complex adduct, it simultaneously becomes a surface species (Ni-SnO2-TiO2). In this investigation, it has been found that surface interaction of nickel complex ions lead to the formation of surface species that are identified by XRD, FTIR, UV\u2013vis DRS, SEM, EDX, HRTEM, XPS, and XANES analyses. The metal-loaded metal oxide coupled semiconductor solids find their applications as catalysts and in advanced electronics. We demonstrate the dopant induced changes in electronic density of states using DFT VASP calculations. The photocatalytic property of Ni-SnO2-TiO2 sample was tested, and the results showed that a small amount of nickel surface co-doped SnO2-TiO2, formed an anatase\u2013rutile-cassiterite mixed phases with surface defects or oxygen vacancies, and had improved the photocatalytic activity. The organic model pollutant of Azo dye and methyl orange degradation property of TiO2 photocatalyst under visible light irradiation has greatly enhanced after the introduction of Ni into the SnO2-TiO2 surface matrix. This study points out a potential way to develop new and more active tin and nickel co-doped titania photocatalysts for organic pollutant degradation in water systems.\n               "}
{"full_text": "The hydrogenation of aromatic nitro compounds is one of the most important applications of precious metal powder catalysts. Amongst these, one of the most used processes is the hydrogenation of dinitrotoluene (DNT) to toluenediamine (TDA), with more than 1 million tons produced globally on an annual basis. TDA is an intermediate in the synthesis of toluene diisocyanate (TDI), a compound predominantly used in the preparation of flexible polyurethane foams and elastomers [1]. Carbon, silica, and alumina-supported transition metal catalysts such as Pd, Pt, Ru, and Ir have been extensively studied in the catalytic hydrogenation of dinitrotoluene [2\u20139]. However, it is difficult to fully separate powder form catalysts from the reaction medium\u00a0because of the small particle size\u00a0and the degree of polarity of the surface. The separation can be achieved to a certain extent in some cases by filtration [10] or centrifugation [11], and the presence of small residual amounts of the catalyst particles can be ignored. But there are applications when the catalyst loss and product contamination due to incomplete separation is unacceptable [12]. A solution to this problem is the application of catalyst supports with magnetic properties, and thus, the separation can be carried out easily by using a magnetic field. The use of magnetic nanocatalysts, such as spinel ferrites, is becoming increasingly important, especially in heterogeneous catalysis [13\u201316]. The crystal structure of the spinel allows various metal ions to be introduced into the system without significantly altering it, but with this, the magnetic, electrical, and dielectric properties can be influenced [17]. Ferrites can be prepared in a number of ways, including hydrothermal [18,19], sonochemical [20,21], sol-gel [22,23], precipitation [24,25], microemulsion [26,27], and even mechanical alloying [28,29] methods. In our previous study [30], a unique method that combines combustion and sonochemical treatment has been developed and proved to be suitable for the efficient synthesis of magnetic metal oxide nanoparticles. Therefore, this method has been used to prepare magnetic Ni, Co, and Cu ferrite supported Pd catalysts, and their applicability has been compared in the hydrogenation of DNT to TDA.During the synthesis of the magnetic spinel nanoparticles cobalt(II) nitrate hexahydrate, (Co(NO3)2 \u2219 6H2O), copper(II) nitrate dihydrate (Cu(NO3)2 \u2219 2H2O), nickel(II) nitrate hexahydrate (Ni(NO3)2 \u2219 6H2O), and iron(III) nitrate nonahydrate (Fe(NO3)3 \u2219 9H2O) from Reanal Ltd., have been used. Polyethylene glycol (PEG 400, Mw: ~400\u00a0g/mol, Macrogola 400 from Molar Chemicals Ltd.) was applied as a reducing agent and dispersion media. In the final catalyst preparation step, palladium(II) nitrate dihydrate (Pd(NO3)2 \u2219 2H2O, Merck Ltd.) was utilized as a precursor and Patosolv (mixture of aliphatic alcohols: 90\u00a0vol% ethanol, 10\u00a0vol% isopropanol; from Molar Chemicals Ltd.) as reaction media.Hielscher UIP100 hdT tip homogenizer (1,000\u00a0W, 20\u00a0kHz) was used to deposit palladium nanoparticles onto the surface of the ferrite crystals. Bs4d22 ultrasonic block sonotrode (D\u00a0=\u00a022\u00a0mm) was applied to initiate the formation of metal hydroxides in PEG 400 dispersion.All three synthesized spinel magnetic nanoparticle samples were examined by high-resolution transmission electron microscopy (HRTEM, FEI Technai G2 electron microscope, 200\u00a0kV). The morphology of the nanoparticles has been characterized, and particle size distributions have been measured. The samples were prepared by dropping the aqueous suspension of the nanoparticles on 300 mesh copper grids (Ted Pella Inc.). X-ray diffraction (XRD) measurements were used to identify and quantify the crystalline phases. Bruker D8 Advance diffractometer (Cu-K\u03b1 source, 40\u00a0kV and 40\u00a0mA) in parallel beam geometry (G\u00f6bel mirror) with Vantec detector was applied to perform the XRD measurements. The average crystallite size of the domains calculated by the mean column length calibrated method using full width at half maximum (lowest) and the width of the Lorentzian component (highest) of the fitted profiles. The palladium contents of the prepared magnetic catalysts have been measured by using a Varian 720\u00a0ES inductively coupled optical emission spectrometer (ICP-OES). For the ICP-EOS measurements, the samples have been prepared by placing them into aqua regia. The specific surface area of the catalysts was measured by nitrogen adsorption-desorption experiments at 77\u00a0K using Micromeritics ASAP 2020 sorptometer based on the Brauner-Emmett-Teller (BET) method. The carbon content of the spinel ferrite samples was determined by using a Vario Macro CHNS element analyzer with a certified phenanthrene standard (C: 93.538%, H: 5.629%, N: 0.179%, S: 0.453%; from Carlo Erba Inc.) and the samples were loaded into tin foils. The carrier gas was helium (99.9990%), while oxygen (99.995%) was used for oxidation during the measurements.Cobalt, copper, and nickel-containing ferrite nanoparticles were synthesized by using a two-step process that includes sonochemical treatment and combustion (Fig.\u00a01\n). In the first step, iron(III) nitrate nonahydrate and one of the precursors (Table\u00a01\n) were dissolved in 20\u00a0g polyethylene glycol, and then, the solutions were treated by using a Hielscher UIP1000 hdT tip homogenizer for 3\u00a0min (130\u00a0W, 19\u00a0kHz).The color of the dispersions has deepened and changed to brownish red, during which metal hydroxides formed from their nitrate salts.In the second step, the PEG 400 was burned by using a Bunsen burner in air and the metal hydroxide phase remained. After that, the samples were heated for another 30\u00a0min to fully oxidize the carbon content.The palladium nitrate dihydrate precursor (0.25\u00a0g) was dissolved in 50\u00a0ml Patosolv, and 2.00\u00a0g ferrite (Co, Ni, or Cu ferrite) was also added to the solution. The dispersions were sonicated by using the homogenizer (130\u00a0W) for 2\u00a0min. The sonication initiates the continuous formation and collapse of vapor microbubbles in the solution and will lead to intense local heating, high pressure, enormous local heating and cooling rates, and liquid jet streams. The area around the microbubbles is full of energy, and thus, chemical reactions, such as the reduction of palladium ions to Pd metal nanoparticles in the presence of alcohol as reducing agent, could take place easily. During this process, the Pd nanoparticles are deposited onto the surface of the magnetic spinel supports, and the final catalysts are formed. Then, the catalyst samples were removed from the dispersion with an Nd magnet, washed with\u00a0Patosolv, and dried at 105\u00a0\u00b0C overnight. The final palladium content of the magnetic catalysts was determined by ICP-OES measurements.Dinitrotoluene hydrogenation was carried out in a methanolic solution (c\u00a0=\u00a00.05\u00a0mol \u2219 dm3) at four different temperatures (303\u00a0K, 313\u00a0K, 323\u00a0K, and 333\u00a0K) and by applying 20\u00a0bar hydrogen pressure. The reactions were performed in a B\u00fcchi Uster picoclave reactor system (SI Fig.\u00a01) (volume: 200\u00a0cm3) with continuous mixing (1,000\u00a0rpm). The sampling took place after the beginning of the hydrogenation at 0, 5, 10, 15, 40, 20, 30, 40, 60, 80, 120, 180, and 240\u00a0min. The quantitative analysis of the samples was carried by Agilent 7,890\u00a0A gas chromatograph coupled with Agilent 5975C Mass Selective detector. During the measurements, a Restek Rxi-1MS column was used (30\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm\u00a0\u00d7\u00a00.25\u00a0mm). Three analytical standards, 2,4-diaminotoluene, 2,4-dinitrotoluene, and 2-methyl-5-nitroaniline (Sigma Aldrich Ltd.), have been used in the analysis of the samples.The efficiency of the catalyst was determined by calculating the conversion (X%) of DNT based on the following equation (Eqn.1):\n\n\n\nX\n\n%\n=\n\n\nc\no\nn\ns\nu\nm\ne\nd\n\n\nn\n\nD\nN\nT\n\n\n\n\ni\nn\ni\nt\ni\na\nl\n\n\nn\n\nD\nN\nT\n\n\n\n\nx\n100\n\n\n\n\nThe DNT hydrogenation can be seen as a first-order reaction [31\u201333]. Thus, the rate constant of the reaction (k) at different temperatures can be calculated by using non-linear regression as the following equation is considered (Eqn. 2):\n\n\n\n\nc\nk\n\n=\n\nc\n0\n\n\u00b7\n\nexp\n\n(\n\u2212\nk\n\u00b7\nt\n)\n\n\n\n\nwhere c\n0 and c\nk (mol/dm3) are the initial and final dinitrotoluene concentrations, respectively.The activation energy (Ea) was calculated based on the Arrhenius equation (Eqn. 3):\n\n\n\nk\n=\nA\n\u00d7\nexp\n\n[\n\n\u2212\n\n(\n\n\nE\na\n\n\nR\n\u00d7\nT\n\n\n)\n\n\n]\n\n\n\n\nwhere k is the reaction rate coefficient, A is the pre-exponential factor, T is temperature, and R is the universal gas constant.The yield (Y%) of the product (TDA) was also calculated as follows (Eqn. 4):\n\n\n\nY\n\n%\n=\n\n\nn\n\nT\nD\nA\n\n\n\nn\n\nT\nD\nA\n\nt\nh\ne\no\nr\ni\nt\ni\nc\na\nl\n\n\n\n\u00b7\n\n100\n\n\n\nwhere \nn\n\n\nTDA\n is the real amount of the formed TDA, while \nn\n\n\nTDA theoretical\n is the amount, which can be formed theoretically.The PEG was eliminated from the samples by combustion, and thus, carbon is expected to remain in the system. This was also confirmed by FTIR measurements (Fig.\u00a02\n). The absorption bands, which are visible at 1,630/cm, 2,858/cm, and 2,919/cm can be attributed to the stretching of CC and to the antisymmetric and symmetric stretching vibration modes of \u2013CH2, respectively (Fig.\u00a01. A). The ferrite particles are covered with a carbon layer that is visible on the HRTEM images (Fig.\u00a02. B). The carbon content of the samples varies between 0.15\u00a0wt% and 0.64\u00a0wt% and the cobalt ferrite-based sample contains the least, while the nickel ferrite the most. The copper-containing sample\u2019s carbon content (0.17\u00a0wt%) is just slightly higher than in the case of the cobalt ferrite-based one.Two more bands have been located around 580/cm and 3,435/cm on the spectra, which can be associated with the vibration mode of the metal-oxygen (\u03bdM-O) stretching and the surface hydroxyl groups and adsorbed water, respectively. The \u2013OH groups are beneficial for the catalyst preparation\u00a0because they promote the adsorption of Pd2+ ions on the catalyst support. The adsorption mechanism is a complex process, which involves physical adsorption, electrostatic interaction, ion exchange, and surface complexation. Hydroxyl groups can be deprotonated depending on the pH, and thus, the surface of the nanoparticles can be negative, which will further promote the adsorption of the metal ions. The strength of the interaction between the precursor ions and the support (with electrostatic interaction or ion exchange) influences the rate of nucleation and growth of palladium nanoparticles on the surface of catalyst support, and thus, smaller nanoparticles can form.In order to identify the different oxide phases in the magnetic spinel-based catalyst supports, XRD measurements have been carried out. In the case of the cobalt-containing spinel, seven reflexion peaks have been identified at 18.3\u00b0 (101), 30.1\u00b0 (200), 35.5\u00b0 (211), 43.1\u00b0 (220), 53.6\u00b0 (312), 57.2\u00b0 (303), and 62.7\u00b0 (224) two Theta degrees (ICDD card number: 22-1086) each of which corresponds only one metal oxide phase, CoFe2O4, and thus, it is a pure cobalt ferrite sample (Fig.\u00a03\n. A).Three metal oxide phases have been identified in the copper-containing sample (Fig.\u00a02. B). The reflexions at 18.3\u00b0 (111), 30.2\u00b0 (220), 35.7\u00b0 (311), 43.3\u00b0 (400), 54\u00b0 (422), 57.2\u00b0 (511), and 62.7\u00b0 (440) 2\u0398 degrees can be associated with the presence of the CuFe2O4 phase that is the main component (78.9\u00a0wt%, Table\u00a02\n) of the system (ICDD card number: 77-0010). CuO (tenorite) and iron(III) oxide (hematite) phases are also located and can be identified as by-products (11.73\u00a0wt% and 9.97\u00a0wt%, respectively). The low-intensity peaks at 35.5\u00b0 (111), 38.7 (022), 48.8 (202), 58.3 (202), 61.5 (113), 66.1 (022), and 68.1 (220) (ICDD card number: 00-001-1117) corresponds to the CuO phase. The hematite phase in the sample is represented by peaks at 24.1 (012), 33.1 (104), 40.9 (113), 49.6 (024), 54.0 (116), 62.5 (214), and 63.9 (300) (ICDD card number: 33-0664).In the nickel-containing sample besides NiFe2O4 (66.3\u00a0wt%), nickel(II) oxide (NiO, bunsenite, 30.01\u00a0wt%) and FeNi3 (awaruite, 3.75\u00a0wt%) have also been identified (Fig.\u00a03. C, Table\u00a02). On the diffractogram, high-intensity peaks appeared at 18.4\u00b0 (111), 30.2\u00b0 (220), 35.3\u00b0 (311) 37.3 (222); 43.4\u00b0 (400), 53.6\u00b0 (422), 57.4\u00b0 (511), and 63.1\u00b0 (440) 2\u0398 degrees which correspond\u00a0to NiFe2O4 (ICDD card number: 54-0964). The presence of NiO has been confirmed as peaks at 37.3\u00b0 (111), 43.2\u00b0 (200), and 62.9\u00b0 (220), and two theta degrees have been located (ICDD card number: 47-1049). Furthermore, low-intensity reflexions have been identified at 44.1\u00b0 (111) and 51.3\u00b0 (200) 2\u0398 degrees that correspond\u00a0to FeNi3 (ICDD card number: 38-419).The Pd/ferrite catalysts have also been examined (Fig.\u00a03 D-F). In each case, peaks at 40.0\u00b0, 46.6\u00b0, and 68.2\u00b0 two Theta degrees have been identified and associated with Pd(111), Pd (200), and Pd (220) reflexions (ICDD card number 046\u20131043), respectively. Thus, the palladium is in the elemental state in the catalytic systems.Specific surface areas (SSA) of the magnetic palladium catalysts have also been measured, and it was found that the Pd/CuFe2O4 sample has the largest (38.6\u00a0m2/g), and it is more than two times the case of Pd/CoFe2O4 (18.2\u00a0m2/g). The SSA of Pd/NiFe2O4 is 21.1\u00a0m2/g, which is also larger than in the case of Pd/CoFe2O4. The palladium content of the catalyst has also been determined, and the Pd/CuFe2O4 contains the most (4.34\u00a0wt%), and it is followed by Pd/CoFe2O4 (3.98\u00a0wt%) and Pd/NiFe2O4 (3.82\u00a0wt%)The size of the nanoparticles of the supports and the corresponding catalytic systems have also been analyzed (Table\u00a03\n). The average particle size of the main phases (ferrites) remained the same for CoFe2O4 and CuFe2O4 before and after the sonochemical treatment of the samples, which led to the palladium deposition. However, in the case of NiFe2O4, the particle size slightly increased from 21 to 23\u00a0nm. The palladium particles are found in a finely dispersed form, and their average diameters are 6\u00a0\u00b1\u00a01\u00a0nm, 4\u00a0\u00b1\u00a01\u00a0nm, and 5\u00a0\u00b1\u00a02\u00a0nm in the case of Pd/CoFe2O4, Pd/CuFe2O4, and Pd/NiFe2O4, respectively.The electron microscopic analysis of the magnetic catalysts shows that each of them contains highly dispersed nanoparticles (Fig.\u00a04 A-\nC). However, the different phases are inseparable in\u00a0the images.Before the catalyst\u2019s preparation, the palladium-free ferrite supports were tested in DNT hydrogenation. The non-loaded ferrites showed varying degrees of catalytic activity by using the NiFe2O4, 69.89 n/n% DNT conversion, and 18.23% TDA yield was achieved after 4\u00a0h of hydrogenation at 333\u00a0K and 20\u00a0bar pressure (SI Fig.\u00a02 A). In the case of the CoFe2O4, 64.17 n/n% DNT conversion was reached, and 18.23 n/n% TDA yield (SI Fig.\u00a02 B). The copper ferrite sample was the least active, and only 56.12% DNT conversion and 15.61% TDA yield were measured. Although the supports were active, only low TDA yield and DNT conversion were achieved, and thus, the involvement of palladium is essential to reach the desirable high activity. Then, the activity of the three synthesized magnetic Pd catalysts has also been compared in dinitrotoluene hydrogenation. The change of DNT concentration with time was measured, and the experiments were carried out at four different temperatures to achieve TDA. The conversion was excellent in each case. The Pd/CoFe2O4 and Pd/NiFe2O4 were able to achieve full conversion after 40\u00a0min at 333\u00a0K and 20\u00a0bar hydrogen pressure (Fig.\u00a05\n A and C). By using the Pd/CuFe2O4 sample, the reaction was slower, which shows that it is slightly less active than its counterparts, but after two hours of hydrogenation at 333\u00a0K and 20\u00a0bar, it was still able to reach 99.97 n/n% conversion (Fig.\u00a05 B). The lower activity of the Pd/CuFe2O4 catalyst can be explained by the relatively high CuO content (11.74\u00a0wt%) of the sample. CuO is used to improve the selectivity of precious metal-containing catalysts in hydrogenation [34,35]. However, above a certain copper content, the activity of the catalyst decreases (catalyst poison effect), and a similar phenomenon is experienced in the present case.The corresponding reaction rate coefficients (k) were also calculated (Table\u00a04\n). The k values are similar in the case of the Pd/CoFe2O4 and Pd/NiFe2O4 catalysts. The Pd/CuFe2O4 sample was less efficient, and its reaction rate was an order of magnitude lower compared to the other two catalysts. The activation energies (Ea) were determined based on the Arrhenius plots (Fig.\u00a05, Table\u00a04 and SI Fig.\u00a03), and it was found that they are in a range between 32 and 39\u00a0kJ/mol, which is similar to other Pd-containing catalysts [36,37].Next to the main-product (TDA), two semi-hydrogenated intermediates, 4-amino-2-nitrotoluene (4A2NT) and 2-amino-4-nitrotoluene (2A4NT) were identified (Fig.\u00a06\n and Fig.\u00a07\n). Furthermore, three additional larger intermediates (Fig.\u00a06, see species 1, 2, and 3) have also been identified, which could be formed through side reactions. Based on the results of the catalytic tests and previous findings, a possible reaction mechanism, including the detected species, has been envisaged (Fig.\u00a06). As DNT contains two nitro functional groups, two main reaction pathways can be proposed. These channels lead to TDA (Fig.\u00a06, middle section) going\u00a0through six consecutive hydrogenation steps. In the first step, a nitroso-nitrotoluene is formed, which is followed by a hydroxylamino-nitrotoluene formation. The formation of hydroxylamino-nitrotoluene was confirmed earlier, but additional molecules or condensed derivates were not reported [36,37]. Thereafter, one of the above-mentioned semi-hydrogenated compounds, depending on which nitro group is hydrogenated first, 2A4NT or 4A2NT will be formed. Then, if the reaction will go further, the other nitro group will be hydrogenated, and the corresponding nitroso and hydroxylamino species will be produced. In the last step, both pathways will reach the main product, TDA. The detected three larger side-products (1, 2, 3) are the results of condensation reaction steps, which can be explained with the enhanced reactivity of the nitroso groups. Janssen et\u00a0al. detected similar condensed species, which formed from 4-nitroso-2-nitrotoluene and 4-hydroxylamino-2-nitrotoluene in the reaction media [38]. In the studied system, 1 could form by the condensation of 2-hydroxy-4-nitrotoluene and 2-nitroso-4-nitrotoluene (Fig.\u00a06). The formation of 2 can be explained with a possible reaction between 4-methyl-resorcinol (which can be formed from 2-hydroxy-4-nitrotolune in methanol) and 2-nitroso-4-nitrotolune. For explaining the presence of the third condensed side-product (3), the formation and reaction of N-2,5-xylylhydroxylamine and 4-nitrotoluene have been assumed.The partially hydrogenated compounds (2A4NT and 4A2NT) converted to toluenediamine above 323\u00a0K by using the Pd/CoFe2O4 and Pd/NiFe2O4 catalysts (Fig.\u00a07 A,B and E,F). However, regardless of temperature, the less active Pd/CuFe2O4 sample is not able to convert all of the intermediates (Fig.\u00a07C,D). In this case, even after 2\u00a0h, there are still some semi-hydrogenated compounds that remained and were not converted to TDA. Thus, the Pd-copper ferrite nanocomposite was less active in the catalytic hydrogenation of nitro compounds than the other catalysts. This phenomenon can be explained by the structure of Pd/CuO nanocomposite, where the active Pd particles are coated with CuO layers, and thus, prevent them from being involved in the catalytic process [39]. As a relatively high amount of CuO is present in the developed Pd/CuFe2O4 catalytic system, a significant portion of the Pd particles can be covered with copper oxide. On the other hand, due to the inhibitory effect, copper and CuO-containing palladium catalysts are very well suited to semi-hydrogenate alkynes, as the interaction between the intermediate species and the bimetallic surface is weakened because of the encapsulation of the active nanoparticles [40\u201345]. This is the so-called \u2018ensemble effect\u2019 that may have contributed to the decreased activity of Pd/CuFe2O4, and thus, the lasting presence of 2A4NT and 4A2NT (Fig.\u00a07C,D).In the case of the Pd/NiFe2O4 and Pd/CoFe2O4 catalysts, the TDA yield was the highest at 333\u00a0K and 20\u00a0bar H2 pressure and reached 99.8 n/n% and 84.2 n/n%, respectively (Fig.\u00a08\n A). The Pd/CuFe2O4 catalyst underperformed its counterparts and achieved only 54.2 n/n% yield. The Pd/CoFe2O4 sample can be separated easily by a neodymium magnet because its support is a pure magnetic ferrite phase (Fig.\u00a08 B). The magnetic separation is not as efficient in the case of the Pd/CuFe2O4 catalyst. The non-magnetic CuO particles dispersed into the reaction media that is shown by its dark green color. (Fig.\u00a08C). Thus, total catalyst recovery is not possible by using magnetic separation. Despite the high NiO content (30\u00a0wt%), the Pd/NiFe2O4 catalyst was easy to separate by using a magnetic field (Fig.\u00a08 D). Most probably, it is due to the adsorption of the NiO on the surface of the nickel ferrite particles.For comparing the catalytic activity of the different ferrite supported catalysts, the corresponding turnover numbers (TON) have been calculated as follows:\n\n\n\nT\nO\nN\n=\n\n\nn\n\nT\nD\nA\n\n\n\nn\n\nP\nd\n\n\n\n\n\n\nwhere n\nTDA is the amount of TDA formed during a 60\u00a0min reaction.Pd/NiFe2O4 catalyst was the most effective as 110.99\u00a0mol TDA was produced at 333\u00a0K and 20\u00a0bar pressure (Table\u00a05\n). The TON was lower (90.84) when the Pd/CoFe2O4 catalyst was used, while in the case of the Pd/CuFe2O4 sample, only 27.18\u00a0mol TDA formed. The turnover numbers confirmed that the activity of the cobalt ferrite and nickel ferrite supported Pd catalysts are excellent, as shown by the corresponding reaction rates (Table\u00a04).The reusability of the most active Pd/NiFe2O4 catalyst was also tested without regeneration (SI Fig.\u00a04, A and B). DNT conversion continuously decreased after each cycle, and the deviation was 14.3% compared to the first and second cycles. The decrease in TDA yields was even more dramatic, as it decreased from 99.8 n/n% to 38.91 n/n%. Thus, regeneration is inevitable after each cycle to preserve the efficiency and functionality of the catalysts.CoFe2O4, CuFe2O4, and NiFe2O4 spinel nanoparticles have been successfully synthesized by using a combination of sonochemical treatment and combustion. These magnetic nanoparticles have been used as supports to prepare palladium-containing catalysts for hydrogenation reactions. The previously developed fast and easy catalyst preparation method has been applied within which post-treatments such as calcination and reduction in a hydrogen atmosphere at high temperature are not necessary. The formation of palladium nanoparticles was initiated by sonication, and they were deposited onto the surface of the synthesized ferrite supports. The final Pd/ferrite catalysts are in an active form and ready to use from the beginning and can be easily separated from the liquid phase with a magnet as the main component of the supports is magnetic. Despite their relatively low specific surface areas (18.2\u00a0m2/g and 21.1\u00a0m2/g), Pd/CoFe2O4 and Pd/NiFe2O4 were highly active during the DNT hydrogenation reactions. Reaction rate constants (k) were similar at 333\u00a0K and 20\u00a0bar hydrogen pressure (2.6\u221910\u22123\u00a0\u00b1\u00a02.3\u221910\u22124 and 2.3\u221910\u22123\u00a0\u00b1\u00a04.4\u221910\u22125/s). The TON was also calculated (90.84 and 110.99) and confirmed the high catalytic activity of these catalysts. The activity of the Pd/CuFe2O4 catalyst was much lower\u00a0and reached only 54.2 n/n% TDA yield at 333\u00a0K (TON: 27.18\u00a0mol TDA/mol Pd), which can be explained by the inhibitor effect of the high CuO content (CuO: 11.73\u00a0wt%) of the support. Moreover, the copper ferrite catalyst contains a significant amount of hematite (Fe2O3: 9.97\u00a0wt%), which is not magnetic, and thus, the magnetic catalyst separation is inefficient. Possible reaction mechanism, including the detected species, has been envisaged based on the results. The TDA yield was the highest (>99 n/n% at 333\u00a0K and 20\u00a0bar hydrogen pressure) in the case of the Pd/NiFe2O4 catalyst. In addition, despite the presence of non-magnetic phases in the support (NiO: 30.00\u00a0wt%), it was well separable by using magnets. The magnetic separability of the catalyst can be explained by the adsorption interaction between the NiO and the spinel particles. The TDA yield was lower (84.2 n/n%) in the case of the cobalt ferrite-based Pd catalyst. However, it is still very well suitable for the hydrogenation of DNT or other aromatic nitro compounds\u00a0because the cobalt ferrite-based support contains only one pure, magnetic phase, which ensures an easy and complete recovery of the catalyst from the reaction medium.Vikt\u00f3ria Hajdu Methodology, Writing- Original draft preparation.Mikl\u00f3s Varga Methodology, Visualization.G\u00e1bor Mur\u00e1nszky Methodology, Data curation.G\u00e1bor Karacs Methodology.Ferenc Krist\u00e1ly Methodology.B\u00e9la Fiser Writing- Reviewing and Editing.B\u00e9la Viskolcz Validation, Funding acquisition.L\u00e1szl\u00f3 Vanyorek: Conceptualization, Supervision.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The described study was carried out as part of the EFOP-3.6.1-16-2016-00011 \u2018Younger and Renewing University \u2013 Innovative Knowledge City \u2013 institutional development of the University of Miskolc aiming at intelligent specialization\u2019 project implemented in the framework of the Sz\u00e9chenyi 2020 program. The realization of this project is supported by the European Union, cofinanced by the European Social Fund. Further support has been given by the National Talent Program (HU), National Young Talents Scholarship, Ministry of Human Capacities (HU), Human Capacities Grant Management Office (EMET) (contract number: NTP-NFT\u00d6-20-B-0062).The following is the supplementary data to this article:\n\nMultimedia component 1\nMultimedia component 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.mtchem.2021.100470.", "descript": "\n                  Cobalt, copper, and nickel ferrite spinel nanoparticles have been synthesized by using a combination of sonochemical treatment and combustion. The magnetic nanoparticles have been used as supports to prepare ~4\u00a0wt% palladium catalysts. The ferrites were dispersed in an ethanolic solution of Pd(II) nitrate by ultrasonication. The palladium ions were reduced to metallic Pd nanoparticles, which were then attached to the surface of the different metal oxide supports. Thus, three different catalysts (Pd/CoFe2O4, Pd/CuFe2O4, Pd/NiFe2O4) were made and tested in the hydrogenation of 2,4-dinitrotoluene (DNT). A possible reaction mechanism, including the detected species, has been envisaged based on the results. The highest 2,4-diaminotoluene (TDA) yield (99 n/n%) has been achieved by using the Pd/NiFe2O4 catalyst. Furthermore, the TDA yield was also reasonable (84.2 n/n%) when the Pd/CoFe2O4 catalyst was used. In this case, complete and easy recovery of the catalyst from the reaction medium is ensured, as the ferrite support is fully magnetic. Thus, the catalyst is very well suited for applicationy in the hydrogenation of DNT or other aromatic nitro compounds.\n               "}
{"full_text": "As one of the 100 most important chemicals in the world, hydrogen peroxide (H2O2) is a valuable and environmentally friendly oxidizing agent with a wide range of applications, ranging from the provision of clean water to the synthesis of valuable chemicals as well as being a potential energy carrier.\n1\u20133\n The current industrial synthesis of H2O2 involves an energy-intensive anthraquinone oxidation-reduction step, which requires elaborate and large-scale equipment and at the same time generates substantial waste.\n3\n\n,\n\n4\n An attractive alternative route for direct on-site production of H2O2 is through an electrochemical process in a fuel cell setup (anode: H2 \u2192 2e\n\u2212\u00a0+ 2H+; cathode: O2\u00a0+ 2e\n\u2212\u00a0+ 2H+ \u2192 H2O2, E\n0\u00a0= 0.695 V), where the oxygen reduction reaction (ORR) occurs via a two-electron pathway. Substantial efforts devoted in recent years to this fuel cell concept have aimed at efficiently generating electricity with a simultaneous high-yield production of H2O2 in basic media.\n5\u20138\n Indeed, recent results have suggested negligible room for further improvements in the activity and selectivity of carbon-based materials for H2O2 synthesis in basic media.\n3\n\n,\n\n6\n\n,\n\n7\n\n,\n\n9\n\n,\n\n10\n Unfortunately, the production of H2O2 in basic media has several drawbacks: (1) H2O2 is less stable and can self-decompose in bases (especially at pH > 9)\n11\n; (2) there is no commercially competitive anion exchange membrane with comparable stability and conductivity to that of the proton exchange membrane (PEM) for device development\n3\n; and (3) H2O2 is more widely used in acidic media with stronger oxidation ability than in basic media. For example, Fenton's reagent, which is mostly applied in organic synthesis and effluent treatment, has an optimal pH range of 2.5\u20133.5.\n12\n Therefore, there is a great industrial motivation to improve H2O2 catalysis in acidic media, more specifically using PEM-type apparatus.\n3\n\n,\n\n13\u201315\n Previously, mercury-alloyed platinum or palladium\n16\n\n,\n\n17\n nanoparticles supported on carbon, as state-of-the-art catalysts, have been investigated for H2O2 synthesis via ORR in acidic media. However, these catalysts contain precious noble metals and toxic mercury, thus limiting their potential applications in H2O2 production. Although homogeneous molecular catalysts such as cobalt macrocycles are highly selective for H2O2 production via ORR,\n18\n the low activity and poor stability hinder their possible applications. Transition metals such as cobalt particles\n19\n or manganese species\n20\n loaded on nitrogenated carbon materials can also be used to produce H2O2 but lack high activity. Meanwhile, the non-uniform structure in these catalysts hinders their identification of active sites, mechanistic study, and further rational optimization. In short, there is still a lack of cost-effective electrocatalysts with high catalytic performance for H2O2 synthesis in acidic media. In recent years, single-atom catalysts (SACs) with well-defined active centers have drawn great attention for their particularly high activity and selectivity in various chemical reactions.\n21\u201323\n In principle, to increase the selectivity of H2O2 production through ORR, O\u2013O bond breaking needs to be minimized. Benefiting from the desirable features of SACs, in which the active sites are atomically isolated, the adsorption of O2 on SACs is usually of the end-on type, rather than \u03bc-peroxo coordination, which therefore could reduce the possibility of O\u2013O bond splitting.\n18\n\n,\n\n24\n\n,\n\n25\n This implies that SACs would be suitable for H2O2 generation via ORR. Previous studies of metal-nitrogen-carbon materials mainly focus on the electrocatalytic activity toward four-electron ORR to H2O for fuel cells applications,\n26\u201330\n whereas unfavorable two-electron ORR to H2O2 is rarely studied in detail. Although there are few studies of their electrocatalytic activities toward two-electron ORR for H2O2 production,\n19\n\n,\n\n20\n the fundamental aspects such as active center and reaction mechanism as well as practical electrolytic cell device aspects remain poorly understood. Here, by combining theoretical and experimental methods, the relation between the structure of transition metal (Mn, Fe, Co, Ni, and Cu) SACs anchored in nitrogen-doped graphene and the catalytic performance of H2O2 synthesis via ORR was systematically studied. Both theoretically predicted activity-volcano relation and experimental results show that the Co SAC possesses optimal d-band center and can function as a highly active and selective catalyst for H2O2 synthesis via ORR and even slightly outperforms state-of-the-art noble-metal-based electrocatalysts in acidic media.Inspired by previous work,\n16\n we first investigated the ORR process on various transition metal SACs anchored in nitrogen-doped graphene for producing H2O2 or H2O along a 2 e\n\u2212 or 4 e\n\u2212 pathway, respectively, by DFT calculations using a computational hydrogen electrode model (details are given in Experimental Procedures). The two-electron (2 e\n\n\u2212\n) pathway to H2O2 via ORR comprises two proton-coupled electron transfer steps with only one intermediate (*OOH):\n\n(Equation\u00a01)\n\n\n\u2217\n+\n\n\nO\n\n\n2\n\n\n+\n\n\nH\n\n\n+\n\n\n+\n\n\ne\n\n\n\u2212\n\n\n\u2192\n\n*OOH\n\n\n\n\n\n\n(Equation\u00a02)\n\n\u2217\nOOH\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\u2192\n\nH\n2\n\n\nO\n2\n\n+\n\u2217\n\n\nwhere the asterisk (*) denotes the active site of the catalyst. In contrast, for the 4 e\n\n\u2212\n ORR pathway, four proton-coupled electron transfer steps are included, in which O2 is reduced to *OOH, *O, *OH, and H2O in sequence, as displayed in Figure\u00a01\nA. Theoretically, an ideal catalyst for H2O2 synthesis should minimize the kinetic barriers for Equations 1 and 2 to provide high activity. Meanwhile, the catalyst needs to maximize the barrier for *OOH dissociation or reduction to *O and *OH (the intermediates of the 4 e\n\n\u2212\n ORR pathway to H2O) to achieve high selectivity.\n16\n Here, density functional theory (DFT) calculations revealed that *OOH, *O, and *OH were all energetically favored to adsorb on the top site of the metal (M) atom (the most stable configurations are shown in Figure\u00a0S2). Therefore, the activity of ORR is mainly dependent on the electronic interaction of the intermediates with the M atom rather than the geometrical effects. Figure\u00a01B shows that the binding energies of *OOH, *O, and *OH are generally scaled with the number of valence electrons in the M atom from manganese to copper. The larger the number of valence electrons in M, the weaker the binding of these intermediates to the M atom, which is because the d-band center of M atom shifts down in energy relative to the Fermi level from Mn to Cu (Figure\u00a01B).\n31\n In detail, the anti-bonding states derived from the coupling between d-orbitals of M atom and 2p-orbitals of bonded O atom of intermediates are shifted down in energy and thus are more filled, which weakens the M\u2013O bonding from Mn to Cu. Then, to compare the ORR activities of these SACs, we calculated the free energy diagrams (Figure\u00a0S3) and constructed the activity-volcano plots for both the 2 e\n\u2212 and 4 e\n\u2212 pathways by using \u0394G*OH as a descriptor, as shown in Figure\u00a01C. For an ideal 2 e\n\u2212 ORR catalyst, the adsorption of *OOH should be thermoneutral at the equilibrium potential (U\u00a0= 0.7\u00a0V versus reversible hydrogen electrode [RHE]), corresponding to \u0394G*OOH\u00a0= \u223c3.5\u00a0\u00b1 0.2 eV.\n16\n However, in striking contrast to the 2 e\n\u2212 ORR, even for the optimal catalyst, an overpotential of \u223c0.4\u00a0V was required to drive the 4 e\n\u2212 reduction of O2 to H2O, because of the scaling relation between *OOH and *OH (Figure\u00a0S3F), i.e., \u0394G*OOH\u00a0= 0.747 \u0394G*OH\u00a0+ 3.32 eV, and similar results have also been found in other models.\n32\u201334\n From Figure\u00a01C, it can be seen that the ORR on the Ni and Cu SACs prefers the 2 e\n\u2212 pathway but with a large overpotential because of the large *OOH reduction barrier (Figures S3D and S3E) and high O2 activation energy (Figure\u00a01D), implying that these two catalysts would exhibit low activity but high selectivity for H2O2 production.By contrast, the binding of O2 on the Mn and Fe SACs is so strong (Figure\u00a01D) that it becomes more downhill in free energy for *OOH reduction to *O (Figures S3A and S3B). Therefore, the 4 e\n\u2212 pathway dominates over the 2 e\n\u2212 pathway on the Mn and Fe SACs, thus causing much lower selectivity for H2O2. In addition, the Fe SAC should possess the highest ORR activity via the 4 e\n\u2212 pathway among the five SACs because of its optimized adsorption energy of oxygenated intermediates (Figures 1B and 1C). Considering the relatively stronger binding energies of Fe and Mn SACs toward ORR species, it is possible that the backside of these two catalysts are covered by *OH or *O. Supplemental calculation results show that the backsides are possibly covered by *OH under ORR working potentials (Figure\u00a0S4). However, this does not obviously change the 4 e\n\u2212 ORR activity (Figure\u00a0S5), but instead could slightly enhance the 2 e\n\u2212 ORR activity (Figure\u00a0S6) for H2O2 production. Significantly, the Co SAC with optimal d-band center shows \u0394G*OOH\u00a0= 3.54 eV at U\u00a0= 0.7\u00a0V versus RHE (Figure\u00a01D) for the 2 e\n\n\u2212\n pathway, neither too strong nor too weak, being positioned nearly at the vertex of the activity-volcano map (Figure\u00a01C), suggesting that the Co SAC would be highly active for the 2 e\n\u2212 pathway. In addition, the higher barrier for *OOH reduction to the *O intermediate on the Co SAC (Figure\u00a0S3C) compared with those on the Mn and Fe SACs (Figures S3A and S3B) would enhance the selectivity of the Co SAC to H2O2. Combined with the predicted high activity, it can be anticipated that the yield of H2O2 on the Co SAC would be the highest among the five SACs.Subsequently, we synthesized five different transition metal SACs anchored in nitrogen-doped carbon (NC) (Mn\u2013NC, Fe\u2013NC, Co\u2013NC, Ni\u2013NC, and Cu\u2013NC) via pyrolysis of melamine, L-alanine, and the corresponding metal acetate mixture (details are given in Experimental Procedures). NC was also prepared by the same method without adding a metal salt for comparison. This synthesis method can be easily scaled up and Figure\u00a0S8 shows a digital photograph of the SACs obtained by batch synthesis. Figures 2A and S9 show representative scanning electron microscopy (SEM) images of the as-synthesized SACs (Figures 2A and S9 for Co\u2013NC, and Figures S9B\u2013S9F for the rest), which display aggregated two-dimensional (2D) platelets. No obvious Co particles can be observed in the transmission electron microscopy (TEM) images of Co\u2013NC (Figures 2B and S10A), similar to the absence of the corresponding metal particles in the other SACs (Figures S10B\u2013S10F), suggesting that the metal species are highly dispersed in the carbon matrix. The Brunauer-Emmett-Teller (BET) surface areas of the six samples (Mn\u2013NC, Fe\u2013NC, Co\u2013NC, Ni\u2013NC, Cu\u2013NC, and NC) obtained from N2 adsorption isotherms (Figure\u00a0S11A) are in the range of 360\u2013670\u00a0m2/g with total pore volumes of 1.5\u20132.2\u00a0cm3/g (Table S4). All samples have similar pore size distribution. The pore sizes show a wide distribution from several to a few tens of nanometers (Figure\u00a0S11B), and the pores themselves are mainly formed from the folds or holes in the carbon matrix. The atomic-scale dispersion of the metals was confirmed by aberration-corrected high-angle annular dark field scanning TEM (HAADF-STEM). The bright spots with diameter of \u223c0.2\u00a0nm in Figure\u00a02C are atomically dispersed Co species in Co\u2013NC. Very similar HAADF-STEM images of the other transition metal catalysts are displayed in Figure\u00a0S12. The X-ray diffraction (XRD) patterns (Figure\u00a02D) show that all five of the transition metal SACs, together with NC, exhibit a single, similar, broad characteristic diffraction peak of the carbon (002) at 25.8\u00b0, suggesting a low degree of crystallization. No other diffraction peaks of metal, metal nitride, or metal oxide are discernible, agreeing well with the TEM and HAADF-STEM results. The Raman spectra (Figure\u00a02E) of the six samples also show very similar patterns with two vibrational bands: the d-band at 1,350\u00a0cm\u22121 is the characteristic peak of vacancies or defects in graphene\n35\n\n,\n\n36\n and the G band at 1,580\u00a0cm\u22121 is the characteristic peak of graphitic layers, which corresponds to the in-plane vibration of sp2 chains associated with the E2g symmetry.\n35\n\n,\n\n37\n The relative intensities of D to G band for the six catalysts are nearly identical, suggesting that they have similarly disordered or defective carbon structures. To further examine the structure of the catalysts, we measured extended X-ray absorption fine structure (EXAFS) spectra. Figure\u00a02F shows the Fourier transformation (FT) of the EXAFS spectra of five transition metal catalysts, exhibit only one strong peak at an interatomic distance of \u223c1.3\u00a0\u00c5 (without phase correction, the same below), which is typical for metal-N bonds.\n26\n\n,\n\n38\n\n,\n\n39\n In addition, the EXAFS spectra of some commercial metal phthalocyanines (M-PC, M=Fe, Co, Ni, and Cu) were measured and Fourier transformation of the EXAFS spectra are shown in Figure\u00a0S13A for reference. All of them show quite similar peaks at interatomic distances of \u223c1.3\u20131.5\u00a0\u00c5, suggesting that the transition metals are mainly coordinated with nitrogen, as was the case in the synthesized SACs. No strong metal-metal bonds (which have interatomic distances of \u223c2.1\u00a0\u00c5 in metal foils as shown in Figure\u00a0S13B) can be observed in Figure\u00a01F, moreover, the EXAFS data can be fitted well with the model proposed in above DFT calculation section (Figure\u00a0S14 and Table S5), further proving that the metal species are mainly atomically dispersed, consistent with the HAADF-STEM results (Figures 2C and S12). In addition, the valence states of the metals in the catalysts were also examined through measurement of their K-edge X-ray absorption near-edge structure (XANES) spectra and comparison with those of metallic foils and metal phthalocyanines (Figure\u00a0S15). Comparison of the first derivative XANES for M\u2013NC catalysts with references indicates that the metals in M\u2013NC are all in positive oxidation states.\n39\n The composition of the catalysts was analyzed by X-ray photoelectron spectroscopy (XPS) (Figure\u00a0S16). The carbon, nitrogen, and oxygen contents of all the catalysts display very similar spectra and valance states, and the estimated atomic percentages of N in the transition metal SACs are \u223c6\u20137 atom % (Table S6), which are higher than that in metal-free NC (3.5 atom %). Moreover, the much-enhanced relative intensities of N species coordinated with metal (\u223c 398.8 eV) in M\u2013NC compared with bare NC (Figure\u00a0S17), suggesting that the transition metal and N can stabilize each other in carbon materials by formation of M\u2013N bonds. The metal contents in the catalysts determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) are in the range of 1.0\u20131.6 wt % (Table S6). By comparison with the typical binding energies of metallic and oxidic states of the corresponding elements (Figure\u00a0S18 and Table S7), it is deduced that all five metals in the catalysts show positive valance states, agreeing with the XANES spectra (Figure\u00a0S15) and previous reports.\n38\n\n,\n\n39\n\nTo examine the catalytic performance, we conducted electrochemical ORR tests in\u00a00.1\u00a0M HClO4 on a rotating ring disk electrode (RRDE) at room temperature (24\u00a0\u00b1 1\u00b0C). Figure\u00a0S19 shows the cyclic voltammetry (CV) curves of the six catalysts acquired in O2-saturated and N2-saturated 0.1\u00a0M HClO4. All the catalysts exhibit similar curves in N2 atmosphere, suggesting their comparable double layer capacitances. Under O2 atmosphere, the reduction peak of oxygen occurs at 0.1\u20130.2\u00a0V versus RHE for Mn\u2013NC, Ni\u2013NC, Cu\u2013NC, and NC, whereas Fe\u2013NC and Co\u2013NC display reduction peaks of O2 at higher potentials (\u223c0.45\u00a0V versus RHE). Subsequently, we conducted linear sweep voltammetry (LSV) measurements on the RRDE and the results are shown in Figure\u00a03\nA. As expected, Co\u2013NC and Fe\u2013NC show much higher activity for ORR and had onset potentials at around 0.7\u00a0V versus RHE. The ring current of Co\u2013NC, which corresponds to the oxidation of H2O2 also starts at around 0.7\u00a0V and the calculated faraday efficiency for H2O2 (Figure\u00a03B) shows that Co\u2013NC is highly selective for H2O2 production in the entire potential range of 0\u20130.7\u00a0V versus RHE. The kinetic current of H2O2 production over Co\u2013NC reached 1 \n\n\n\nmA\n\n/\n\n\n\ncm\n\n\ndisk\n\n\n2\n\n\n\n\n\n (corresponding to a mass-normalized current density of 40 A/gcatalyst.) at 0.6\u00a0V versus RHE with H2O2 faraday efficiency >90%. In contrast, Fe\u2013NC shows a much lower selectivity for H2O2, consistent with the above DFT predictions and previous reports.\n27\n\n,\n\n28\n\n,\n\n40\n Although the other catalysts, Mn\u2013NC, Ni\u2013NC, Cu\u2013NC, and NC, are also highly selective for H2O2 (Figure\u00a03B), their catalytic activities are much poorer. The electron transfer number is calculated according to reported method\n19\n and shown in Figure\u00a0S20A. The number over Co\u2013NC is close to\u00a02, agreeing with its selectivity for H2O2. The turnover frequency (TOF) values of the catalysts for H2O2 production at different potentials were calculated (Figure\u00a0S20B) and the highest TOF value of 2.5 s\n\u22121 was obtained for Co\u2013NC at a potential of 0.5\u00a0V versus RHE. The above experimental results are consistent with our DFT calculations (Figure\u00a01), which show that Co\u2013NC is at the top of the activity-volcano map for H2O2 production reaction. In addition, we further examined the electrochemical characteristics of the transition metal SACs in N2-saturated 0.1\u00a0M HClO4 containing 0.1\u00a0M H2O2 to monitor their catalytic activity for H2O2 oxidation (H2O2 \u2192 O2\u00a0+ 2H+\u00a0+ 2e\n\n\u2212\n, E0\u00a0= 0.695 V) and reduction (H2O2\u00a0+ 2H+\u00a0+ 2e\n\n\u2212\n\u2192 2H2O, E0\u00a0= 1.763\u00a0V), which are closely related to their selectivities for H2O2. It can be observed from the LSV curves (Figure\u00a0S21) that the oxidation of H2O2 over Co\u2013NC starts at 0.75\u00a0V versus RHE, very close to the onset potential (0.7\u00a0V versus RHE, Figure\u00a03A) of its reverse reaction, namely reduction of O2 to H2O2 (O2\u00a0+ 2H+\u00a0+ 2e\n\u2212 \u2192 H2O2, E0\u00a0= 0.695 V). Therefore, the reduction of O2 to H2O2 over Co\u2013NC is nearly reversible, and this is consistent with the high activity of Co\u2013NC for O2 reduction to H2O2 and our DFT calculations. Figure\u00a0S21 further shows that Fe\u2013NC can efficiently reduce H2O2 to H2O, thus leading to its much lower selectivity for H2O2 production (Figure\u00a03B). Moreover, as deduced from the DFT calculation, Mn\u2013NC should also possess low selectivity for H2O2, because of its strong adsorption of oxygen intermediates. However, Mn\u2013NC in fact shows relatively high selectivity for H2O2, close to that of NC in the ORR (Figure\u00a03B), possibly because the adsorption of oxygenated species is so strong that it blocks the Mn active sites, making Mn\u2013NC behave similarly to bare NC. Meanwhile, Cu\u2013NC and Ni\u2013NC are less active for H2O2 reduction and oxidation because of their too weak adsorption of oxygenated intermediates. Figures 3C and S22 (mass-normalized activity) compares the performances of state-of-the-art catalysts for H2O2 production through ORR. It can be seen that Co\u2013NC is the most effective catalyst for H2O2 synthesis (Table S8), which even slightly outperforms the best previously reported catalyst in acidic media, a Pd-Hg alloy. The effect of the catalyst loading amount on the ORR performance for Co\u2013NC and other SACs was further optimized (Figure\u00a0S23). Reducing the loading of Co\u2013NC slightly improves the selectivity for H2O2. As shown in Figure\u00a0S21, H2O2 as an intermediate can be further reduced to H2O over Co\u2013NC. Thus, decreasing the catalyst loading amount reduces the residence time of H2O2 on the catalyst surface, so that less H2O2 is reduced to H2O, therefore increasing the H2O2 selectivity. The rotating speed of the RRDE was found to have little influence on the selectivity for H2O2 (Figure\u00a0S24). Additionally, NC loaded with Co nanoparticles (CoNPs/NC) was also prepared for comparison (Figure\u00a0S25), but its ORR activity was much lower than Co\u2013NC (Figure\u00a0S26). Considering that cobalt porphyrins and phthalocyanines are known for their high selectivities for H2O2 generation by ORR despite their rapidly decaying activities in acidic conditions,\n18\n\n,\n\n27\n\n,\n\n41\n we prepared tetra-amino-cobalt(II) phthalocyanine (Co-TAPC) loaded on carbon nanotubes (Co-TAPC/commercial carbon nanotubes [CNT]) and tested its ORR performance (Figure\u00a0S27). However, although the H2O2 selectivity of Co-TAPC/CNT was high, the activity was much lower than Co\u2013NC. The catalytic ORR performances of all six catalysts together with CNT were also tested in alkaline conditions (0.1\u00a0M KOH). All of the catalysts exhibited higher activity for ORR in alkaline conditions (Figure\u00a0S28) than in acidic media and had markedly increased current density at the same potential versus RHE. Among the studied catalysts, NC, CNT, and even blank glassy carbon electrode (GCE) exhibited high selectivity for H2O2 production in alkaline media (Figures S28B and S28D). However, these carbon-based materials show very poor ORR performance in acidic conditions, as shown in Figures 3A and S26. This trend is consistent with previous reports,\n5\u20137\n\n,\n\n10\n and might be attributed to the different affinity of protons and hydroxyls toward the various functional groups present on the surface of the catalysts at different pH.\n42\n\nBesides activity, stability is another important consideration for a catalyst in practical use. The stability of Co\u2013NC was studied at 0.5\u00a0V versus RHE in 0.1\u00a0M HClO4 both under static and rotating conditions. The currents of both the ring and the disk electrode remained stable for 10\u00a0h without obvious decay (Figure\u00a03D); the slightly increased current of the ring electrode can be attributed to the gradually accumulated H2O2 in the electrolyte. The selectivity for H2O2 determined by titration method (details are given in Experimental Procedures) remained as high as \u223c88% throughout the entire process. Figure\u00a0S31 compares the LSV curves of Co\u2013NC before and after the stability test. A change can be observed in the kinetic range, which might stem from partial detachment of the catalyst from the electrode or some deactivation occurred. The stability of Co\u2013NC was further confirmed by CV cycling for 5,000 cycles (Figure\u00a0S32). Half-cell experiment at fixed potential of 0.5 and 0.4\u00a0V versus RHE chronoamperometry test was further conducted to imitate real case H2O2 production. An average H2O2 production rate of 80 and 275 \n\n\n\nmmol\n\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\ng\n\n\ncatalyst\u00a0\n\n\n\u2212\n1\n\n\n\n\nh\n\n\n\u2212\n1\n\n\n\n is obtained at 0.5\u00a0V and 0.4\u00a0V versus RHE, respectively (Figure\u00a0S33).The HAADF-STEM images of the used catalysts show that the Co species are still atomically dispersed (Figure\u00a0S34), demonstrating the good catalytic stability of Co\u2013NC in the H2O2 synthesis process.From the above-mentioned DFT calculations, we found that the first step (*\u00a0+ O2\u00a0+ H+\u00a0+ e\u2212 \u2192*OOH) was the thermodynamic potential-determining step for ORR on Co\u2013NC. To shed more light on the reaction mechanism, we performed kinetic analysis to experimentally probe the rate-determining step. First, the kinetic current of ORR on Co\u2013NC was obtained by Koutecky\u2013Levich analysis from the LSV curves acquired at different rotating speeds (Figures S35 and S36). The reaction order of O2 was determined by performing ORR at different O2 partial pressures. Figure\u00a04\nA shows the kinetic current as a function of overpotential at different O2 partial pressures. Then, the logarithm of the kinetic current versus the logarithm of the O2 partial pressure was plotted as shown in Figure\u00a04B, from which, we can deduce that the reaction order of O2 (slope of the line) varied from 0.53 to 0.90 as the overpotential increased from 0 to 250\u00a0mV. Simultaneously, the Tafel slope increased from\u00a0\u223c110\u00a0mV dec\u22121 to \u223c140\u00a0mV dec\u22121, and finally to \u223c240\u00a0mV dec\u22121 as the overpotential increased from 0 to 250\u00a0mV (Figure\u00a04A). Figures 4C and 4D show the effect of pH (H+ concentration) on the activity of ORR over Co\u2013NC. By the same analysis as performed for Figures 4A and 4B, the reaction order of H+ in the rate-determining step is \u22120.05 \u223c \u22120.07, very close to zero, as shown in Figure\u00a04D. Thus, it is suggested that H+ is not involved in the rate-limiting step, namely, the protonation process is fast. By combining Figures 4A\u20134D and Table S9, it is deduced that the rate-determining step of H2O2 synthesis over Co\u2013NC is as follows: *\u00a0+ O2\u00a0+ e\n\u2212 \u2192 *O2\n\u2212, which is covered in the DFT calculation predicted thermodynamic potential-determining step (*\u00a0+ O2\u00a0+ H+\u00a0+ e\n\u2212 \u2192 *OOH).\n43\n In detail, at relatively low overpotential (<50\u00a0mV), it is mainly controlled by the electron transfer step of adsorbed O2. (*O2\u00a0+ e\n\u2212 \u2192 *O2\n\u2212). The electron transfer step becomes faster by increasing the overpotential. Then the overall reaction rate is more limited by the O2 adsorption process, which agrees well with the observed gradually increasing Tafel slope and reaction order of O2. To monitor the change of Co electronic state and coordination environment, operando X-ray absorption spectroscopy (XAS) was conducted to probe the change of the Co K-edge under H2O2 synthesis conditions in 0.1\u00a0M HClO4. The EXAFS spectrum of the Co\u2013NC in air is nearly the same as that of Co\u2013NC immersed in electrolyte saturated with N2, and the Co\u2013N distance is 1.25\u00a0\u00c5 in both cases (Figure\u00a0S37). After switching N2 to O2, a dramatic enhancement of Fourier transformed intensity and slight increase of Co\u2013N distance to 1.35\u00a0\u00c5 are observed (Figure\u00a0S38), indicating the adsorption of O2 onto Co atoms, which pulls the Co atoms out of plane. This result agrees with the DFT calculations, which show that oxygen species are energetically favored to adsorb on the top site of Co atoms. At the potential of 0.6\u00a0V versus RHE, part of the adsorbed O2 is transformed to H2O2 via ORR, and the Co\u2013N distance decreases to 1.32\u00a0\u00c5 (Figure\u00a04E). At lower potentials, the transformation of adsorbed O2 to H2O2 via ORR becomes more rapid, and the surface coverage of O2 becomes much lower because of the limited rate of O2 adsorption, which explains the further decrease of the Co\u2013N distance to 1.29\u00a0\u00c5 at 0.3\u00a0V versus RHE. Additionally, the Co\u2013N bond distance recovers to its initial value (1.35\u00a0\u00c5) after the potential returns to that of open-circuit conditions and the EXAFS spectra in k-space also show similar trend (Figure\u00a0S37). The operando XANES of Co\u2013NC were also collected (Figure\u00a0S37). Comparison of the spectra in N2- and O2- (Figure\u00a0S38B) atmosphere shows an increase in intensity of XANES in O2, suggesting molecular O2 adsorption on the cobalt centers in Co\u2013NC, which agrees with previous report.\n29\n This trend matches very well with the kinetic analysis. Figure\u00a04F displays a schematic illustration of the ORR steps taking place on Co\u2013NC for H2O2 production, in which step 2 is the rate-limiting process at higher potential whereas step 1 becomes rate-limiting at lower potential. After *OOH is formed, further reduction to H2O2 is rapid, and the active sites are vacated by the desorption of H2O2 to complete the catalytic cycle.In summary, DFT calculations predicted that a cobalt-based SACs anchored in a NC matrix would outperform other transition metal SACs for H2O2 production through ORR. Then, those transition metal SACs were successfully synthesized. Just as predicted by the DFT calculations, Co\u2013NC experimentally behaved as a highly active and selective electrocatalyst for H2O2 synthesis via oxygen reduction in acidic media. A kinetic current density of 1 mA/cm2 was reached on Co\u2013NC at a potential of 0.6\u00a0V versus RHE with H2O2 selectivity >90%, and these performance measures could be sustained for 10\u00a0h continuous operation without decay. The optimized adsorption energy of oxygenated intermediates on Co\u2013NC with optimal d-band center as compared with other transition metal (Mn, Fe, Ni, and Cu) SACs is chiefly responsible for its high activity and selectivity toward H2O2 production. A kinetic analysis and operando X-ray absorption study combined with DFT calculations demonstrated that nitrogen-coordinated Co single atom was the active site for H2O2 synthesis through ORR and the reaction was rate-limited in the first proton-coupled electron transfer step. This work combines the advantages of both homogeneous catalysts of cobalt macrocycles (well-defined active sites) and heterogeneous catalysts (high catalytic performance) together, moreover, operando XAS combined with kinetics analysis in this work tracked the dynamic change of nitrogen-coordinated cobalt active center under reaction condition, which together lead to higher catalytic performance with enhanced understanding of the reaction process.Chemicals: melamine (C3H6N6, 99%), L-alanine (C3H7NO2, 98%), cobalt(II) acetate tetrahydrate (Co(CH3CO2)2\u00b74H2O, 99%), manganese(II) acetate (Mn(CH3CO2)2, 98%), iron(II) acetate (Fe(CH3CO2)2, 99.99%), nickel(II) acetate tetrahydrate (Ni(CH3CO2)2\u00b74H2O, 98%), nickel(II) phthalocyanine (C32H16N8Ni, 85%), copper(II) acetate (Cu(CH3CO2)2, 98%), copper(II) phthalocyanine (C32H16N8Cu, sublimed grade, 99 %), hydrochloric acid (HCl, 37%), nitric acid (HNO3, 70%), hydrogen peroxide solution (H2O2, 30 wt % in water), and cerium(IV) sulfate (Ce(SO4)2, 98%) were purchased from Sigma-Aldrich and absolute ethanol was bought from Merck. Cobalt(II) phthalocyanine (C32H16N8Co, 95%, product code: 41496) and Iron(II) phthalocyanine (C32H16N8Fe, 96%, product code: 39262) were purchased from Alfa Aesar. Multi-walled carbon nanotubes (CNT, 10\u201320\u00a0nm diameter, 5\u201315\u00a0\u03bcm length) were bought from TCI chemical company. All chemicals were used directly without further purification. De-ionized water was obtained from Millipore Q water purification system. Transition metal SACs were synthesized according to our previous method\n38\n with slight modification. In a typical synthesis, 12\u00a0g of melamine, 2\u00a0g of L-alanine, and 50\u00a0mg of transition metal acetate were homogeneously mixed by ball milling for 1 h. Then, 15\u00a0mL of ethanol mixed with 3\u00a0mL of hydrochloric acid was added and the slurry was put in a mortar. The mixture was milled in a fume hood until all ethanol was evaporated. The resultant solid was dried in an oven at 60\u00b0C overnight and ball milled again for 1 h. The thus obtained powder was pyrolyzed under flowing N2 atmosphere in a tube furnace with the following ramping program: from room temperature to 600\u00b0C at a ramping rate of 2.5\u00b0C/min, then hold at 600\u00b0C for 120\u00a0min, ramp to 900\u00b0C at 5\u00b0C/min and hold for 90\u00a0min, finally the furnace was naturally cooled down to room temperature. The obtained black solid materials were grinded and then washed by 2\u00a0M HCl aqueous solution at 80\u00b0C for 24\u00a0h under stirring to remove metal particles. For copper-based material, 1\u00a0M HNO3 was used to remove copper metal particles. The acid-washed materials were dried and then annealed again in N2 at 800\u00b0C for 1\u00a0h at a heating rate of 10 \u00b0C/min to recover the crystallinity. The thus obtained SACs were marked as Mn\u2013NC, Fe\u2013NC, Co\u2013NC, Ni\u2013NC, and Cu\u2013NC, respectively, according to the metal acetate used. NC was synthesized by the same method without adding metal acetate and undergoing acid washing. Co nanoparticle supported on NC with Co content of 2 wt % was prepared by impregnation method: 200\u00a0mg of NC was dispersed in 20\u00a0mL mixed solution of water and ethanol (volume ratio 1:1) containing 18\u00a0mg of cobalt(II) acetate tetrahydrate. The mixture was heated to 80\u00b0C in an oil bath under stirring to evaporate the solution. The obtained solid was then calcined at 400\u00b0C in N2 atmosphere for 2 h. The obtained catalyst was marked as CoNPs/NC. In addition, tetra amino cobalt(II) phthalocyanine (Co-TAPC) was synthesized and loaded on CNT according to the method reported in literature,\n44\n\n,\n\n45\n and the catalyst was marked as Co-TAPC/CNT.Powder XRD was performed on a Bruker D2 Phaser using Cu K\u03b1 radiation with a LYNXEYE detector at 30 kV and 10 mA. The morphological information was examined with field-emission SEM (FESEM, JEOL JSM-6700F). Sub angstrom-resolution high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) characterization was conducted on a JEOL JEMARM200F STEM and TEM with a guaranteed resolution of 0.08\u00a0nm. The metal content in the catalysts was quantified\u00a0by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, PerkinElmer). N2 adsorption-desorption was performed on an Autosorb-6 (Quantachrome) at 77 K. Before analysis, the samples were degassed at 200\u00b0C for 5 h. BET surface area was calculated in the P/Po range of 0.05\u20130.2. Pore size destitution was obtained by Barrett-Joyner-Halenda (BJH) method by using the adsorption branch. Pore volume was calculated by the adsorption amount at P/Po\u00a0= 0.985. Raman spectra were recorded on a Renishaw INVIA Reflex Raman spectrometer using 514\u00a0nm laser as the excitation source. XPS measurements were carried out on a Thermofisher ESCALAB 250Xi photoelectron spectrometer (Thermofisher Scientific) using a monochromatic Al K\u03b1 X-ray beam (1,486.6 eV). XAS including both XANES and EXAFS at Mn, Fe, Co, Ni, and Cu K-edge were collected in total-fluorescence-yield mode at ambient air in BL-01C1 at the National Synchrotron Radiation Research Center (NSRRC), Taiwan. The spectra were obtained by subtracting the baseline of pre-edge and normalizing to the post-edge. EXAFS analysis was conducted using Fourier transform on k3-weighted EXAFS oscillations to evaluate the contribution of each bond pair to Fourier transform peak. Operando measurement in a typical three-electrode setup was performed in a specially designed Teflon container with a window sealed by Kapton tape with continuous gas bubbling. X-ray was allowed to transmit through the tape and electrolyte, so that the signal of XAS could be collected in total-fluorescence-yield mode in BL-01C1 at NSRRC, Taiwan.The electrochemical performance of various catalysts was evaluated in a three-electrode configuration with carbon rod as the counter electrode and Ag-AgCl electrode with saturated KCl salt bridge as the reference electrode on a RRDE setup (AFE6R1PT model; disk OD\u00a0= 5.0\u00a0mm; ring OD\u00a0= 7.50\u00a0mm; ring ID\u00a0= 6.50\u00a0mm; Pine Research Instrumentation, USA) and a CHI (760E) potentiostat. 0.1\u00a0M HClO4 was prepared by diluting perchloric acid (70%, 99.999% trace metals basis, Sigma) with Millipore Q water. 0.1\u00a0M KOH was prepared by dissolving KOH pellets (semiconductor grade, 99.99% trace metals basis, Sigma) in Millipore Q water (15 M\u03a9). A RHE was made with two Pt plates as working and counter electrodes to calibrate the Ag-AgCl electrode and H2 was bubbled over the working electrode. Potentials reported here are referenced to the RHE scale as follows: E\nRHE\u00a0= E\nAg/AgCl\u00a0+ 0.197\u00a0V\u00a0+ 0.059\u00a0V \u00d7 PH or standard hydrogen electrode (SHE) scale: E\nSHE\u00a0= E\nAg/AgCl\u00a0+ 0.197 V. To prepare the working electrode, the catalyst ink that was prepared by ultrasonically mixing 5\u00a0mg of the catalyst, 0.98\u00a0mL of Millipore Q H2O (15 M\u03a9), 0.98\u00a0mL of isopropyl alcohol, and 40\u00a0\u03bcL of 5 wt % D520 Nafion dispersion solution was drop-casted on freshly polished RRDE. In a typical measurement, 2\u00a0\u03bcL of catalyst ink was used, which corresponds a catalyst loading amount of 25 \n\n\n\n\u03bcg\n\n/\n\n\n\ncm\n\n\ndisk\n\n\n2\n\n\n\n\n\n. Before collecting the electrochemical data, the electrolyte was bubbled with purified O2 or N2 for 30\u00a0min. Subsequently the working electrode together with Pt ring were cycled for ten cycles between 0 and 1.0\u00a0V versus RHE at a scan rate of 500\u00a0mV s\u22121 to achieve a stable performance. CV curves were recorded in the potential range of 0\u22121.1\u00a0V versus RHE at 500\u00a0mV s\u22121 under static condition with saturated N2 or O2. LSV curves were recorded at a scan rate of 5\u00a0mV s\u22121 with 100% solution ohmic drop correction under 1,600\u00a0rpm or other indicated rotation speed; the potential of the Pt ring in the working electrode was set at 1.2\u00a0V versus RHE. In kinetic analysis, according to the Henry's law, the O2 concentration is proportional to its partial pressure in the gas phase, thus partial pressure is used to control the concentration of O2 in electrolyte. In detail, the partial pressure of O2 was adjusted by diluting O2 flow with Argon at controlled flow rate. For example, O2 partial pressure of 25\u00a0kPa was obtained by mixing gas of O2 (50\u00a0mL/min) and argon (150\u00a0mL/min). For different pH (1.5\u22123.0), KOH tablet was added to 0.1\u00a0M H3PO4 solution to tune the pH and KNO3 was added to adjust the ionic strength.\n46\n Considering that O2 partial pressure and concentration of H+ can influence the equilibrium potential of the reaction (O2\u00a0+ 2e\n\u2212\u00a0+ 2H+ \u2192 H2O2) according to the Nernst equation: \n\n\n\nE\n\n\n\n\n\n\nO\n\n\n2\n\n\n\n/\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\n=\n\n\nE\n\n\n\n\n\n\nO\n\n\n2\n\n\n\n/\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\n0\n\n\n\u2212\n\n\nR\nT\n\n\nz\nF\n\n\nln\n\n\n\n\na\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\na\n\n\n\n\nH\n\n\n+\n\n\n\n\n2\n\n\n\n\n\n\na\n\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\n\n, where \n\n\n\nE\n\n\n\n\n\n\nO\n\n\n2\n\n\n\n/\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\n0\n\n\n\n\u00a0= 0.695\u00a0V (versus standard hydrogen electrode) is the standard potential of the reaction at O2 partial pressure of 1 atm, the activity of H+ (a\nH+) and H2O2 (a\nH2O2) equal 1 mol/L. Here, due to the unknown value of a\nH2O2 in the electrolyte solution, we just assume it as 1 mol/L to get a pseudo-equilibrium potential. Then, the change of O2 partial pressure and activity of H+ would reach a new pseudo-equilibrium potential \n\n\n\nE\n\n\n\n\n\n\nO\n\n\n2\n\n\n\n/\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\n\n. The overpotential is defined as \u03b7\u00a0= E\napplied\n\u2212\n\n\n\n\nE\n\n\n\n\n\n\nO\n\n\n2\n\n\n\n/\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\n\n, E\napplied is the potential (versus SHE) applied to the working electrode, \n\n\n\nE\n\n\n\n\n\n\nO\n\n\n2\n\n\n\n/\n\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n2\n\n\n\n\n\n\n\n is the pseudo-equilibrium potential at designated O2 pressure and H+ activity. It should be noted that the overpotential defined here is not a strict one due to the unknown true value of equilibrium potential, but it still can be used as the driving force of the reaction for kinetic analysis.H2O2 selectivity was calculated on rotating ring disk electrode based on the currents of both disk and ring electrode according to:H2O2 selectivity or Faraday efficiency: H2O2 (%)\u00a0= 200\n\n\n\n\n\n\n\nI\n\n\nR\n\n\n\n/\n\nN\n\n\n\n\n\n\nI\n\n\nD\n\n\n\n+\n\n\n\n\n\nI\n\n\nR\n\n\n\n/\n\nN\n\n\n\n\n\n\nwhere I\nR is the ring current, I\nD is the disk current and N is the collection efficiency of the RRDE (0.25), which is calibrated by the redox of potassium ferricyanide (Figure\u00a0S29).During stability test, H-type electrochemical cell with working and counter electrode separated by Nafion film was used to avoid further oxidation of H2O2 on the anode. The formation rate of H2O2 in half-cell experiment was conducted in O2 saturated 0.1\u00a0M HClO4 in a H-type cell. The working electrode is prepared with carbon paper (1\u00a0\u00d7 1\u00a0cm) by coating catalysts with loading amount of 100\u00a0\u03bcg/cm2.The concentration of H2O2 in electrolyte during stability test was determined by cerium sulfate titration (2Ce4+\u00a0+ H2O2 \u2192 2Ce3+\u00a0+ O2\u00a0+ 2H+) as detailed in literature.\n7\n The concentration of Ce4+ was measured by ultraviolet -visible spectrometer (JINGHUA Instruments, Model: 754PC) at 316\u00a0nm and the calibration curve is shown in Figure\u00a0S30.Spin-polarized DFT calculations were performed using the generalized gradient approximation (GGA) in the form of Perdew\u2013Burke\u2013Ernzerhof (PBE) for the exchange-correlation potentials,\n47\n\n,\n\n48\n the projector augmented wave (PAW) pseudopotential for the core electrons,\n49\n and a 480 eV cutoff energy for the valence electrons as implemented in the Vienna ab initio simulation package (VASP).\n50\n\n,\n\n51\n Transition metals are likely trapped in the vacancies of graphene at various levels of nitrogen doping. The porphyrinic moieties containing 3D transition metal such as Fe and Co are reported to be good ORR catalyst.\n30\n Herein, we investigated ORR of 2 e\n\u2212 and 4 e\n\u2212 pathways on the nitrogen and transition metal atom (TM\u00a0= Mn, Fe, Co, Ni, and Cu) co-doped graphene. The single transition metal atom dispersed catalysts are simulated using a cluster model (formula of C40H16N4M), in which the M atom is bound with four pyrrolic nitrogen atoms. Here, we choose cluster model other than periodic one based on two considerations: first, previous studies reported that the porphyrins containing 3d transition metals show promising activity toward production of H2O2.\n30\n\n,\n\n52\n\n,\n\n53\n If periodic model is constructed using the skeleton of the porphyrin molecule embedded with a metal atom as the basic unit, eight-membered carbon rings inevitably exist, which is significantly different from the honeycomb structure of graphene. Second, previous experimental and DFT studies proposed that the periodic model with MN4 group (M\u00a0= transition metal, N\u00a0= nitrogen) compactly embedded in graphene cannot correctly predict the ORR and CO2 reduction activities of these SACs.\n30\n\n,\n\n54\n Therefore, the cluster model is selected here. The cluster is placed in a box in the size of 30\u00a0\u00d7 30\u00a0\u00d7 13\u00a0\u00c5, with vacuum layers of \u223c13\u00a0\u00c5 along vertical and lateral directions to decouple the interaction between neighboring images. The energies of gas-phase H2 and H2O molecules were calculated in a cubic supercell with length of 20\u00a0\u00c5. All atoms are free to relax until the net force per atom is less than 0.02 eV/\u00c5. We consider the possible spin state of transition metal atom with and without adsorbates and confirm the most stable adsorption configurations of *OOH, *O, and *OH. The corresponding binding energies and spin states are listed in Tables S1 and S2. The stability of SACs is evaluated via calculating the formation energy, which is defined as:\n\n(a)\n\n\n\n\nE\n\n\nf\no\nr\nm\n\n\n=\n\n\nE\n\n\n\n\nT\nM\n\n/\n\ng\nr\na\n\n\n\n\n\u2212\n\n\nE\n\n\ng\nr\na\n\n\n\u2212\n\n\nE\n\n\nT\nM\n\u2212\nb\nu\nl\nk\n\n\n\n\n\nwhere \n\n\n\nE\n\n\n\n\nT\nM\n\n/\n\ng\nr\na\n\n\n\n\n\n, \n\n\n\nE\n\n\ng\nr\na\n\n\n\n, and \n\n\n\nE\n\n\nT\nM\n\u2212\nb\nu\nl\nk\n\n\n\n represent the energies of TM and N co-doped graphene, N-doped graphene, and TM atom in the bulk phase.According to this definition, a more negative formation energy is, the more stable of the SAC will be. The calculated formation energies of these SACs (\n\n\n\nE\n\n\nf\no\nr\nm\n\n\n\n) are all negative and are shown in Figure\u00a0S1, indicating that such single atom dispersed structures are thermodynamically stable. The d-band center of the transition metal atom in SAC is defined as:\n\n(b)\n\n\n\n\nd\n\n\ncenter\n\n\n=\n\n\n\n\n\u222b\n\n\u2212\n\u221e\n\n\n+\n\u221e\n\n\n\n\n(\n\n\u03b5\n\u2212\n\n\nE\n\n\nF\n\n\n\n)\n\nn\n\n(\n\nx\n\n)\n\nd\n\u03b5\n\n\n\n\n\n\n\u222b\n\n\u2212\n\u221e\n\n\n+\n\u221e\n\n\n\nn\n\n(\n\nx\n\n)\n\nd\n\u03b5\n\n\n\n\n\n\n\nwhere \n\nn\n\n(\n\nx\n\n)\n\n\n and E\nF are the projected density of states of the d-orbitals of M atom in C40H16N4M and the corresponding fermi level of C40H16N4M, respectively.Model of pyridinic N coordinated Co SAC was also used for calculation and given in Figure\u00a0S7 for reference.The ORR after 2 e\n\u2212 and 4 e\n\u2212 mechanisms produces H2O and H2O2, respectively. The associative 4 e\n\u2212 reaction is composed of elementary steps (c, d, e, and f):\n\n(c)\n\n\n\u2217\n+\n\n\nO\n\n\n2\n\n(\n\ng\n\n)\n\n\n\n+\n\n\nH\n\n\n+\n\n\n+\n\n\ne\n\n\n\u2212\n\n\n\u2192\nO\nO\n\n\nH\n\n\n\u2217\n\n\n\n\n\n\n\n\n(d)\n\n\nO\nO\n\n\nH\n\n\n\u2217\n\n\n+\n\n\nH\n\n\n+\n\n\n+\n\n\ne\n\n\n\u2212\n\n\n\u2192\n\n\nO\n\n\n\u2217\n\n\n+\n\n\nH\n\n\n2\n\n\n\n\nO\n\n\n\n(\n\nl\n\n)\n\n\n\n\n\n\n\n\n\n(e)\n\n\n\n\nO\n\n\n\u2217\n\n\n+\n\n\nH\n\n\n+\n\n\n+\n\n\ne\n\n\n\u2212\n\n\n\u2192\nO\n\n\nH\n\n\n\u2217\n\n\n\n\n\n\n\n\n(f)\n\nO\n\nH\n\u2217\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\u2192\n\nH\n2\n\n\nO\n\nl\n\n\n+\n\u2217\n\n\n\nThe ORR of 2 e\n\u2212 mechanism comprises of elementary steps (g and h):\n\n(g)\n\n\n\u2217\n\n+\n\n\n\nO\n\n\n2\n\n(\n\ng\n\n)\n\n\n\n+\n\n\nH\n\n\n+\n\n\n+\n\n\ne\n\n\n\u2212\n\n\n\u2192\nO\nO\n\n\nH\n\n\n\u2217\n\n\n\n\n\n\n\n\n(h)\n\nO\nO\n\nH\n\u2217\n\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\u2192\n\nH\n2\n\n\n\nO\n2\n\n\nl\n\n\n+\n\u2217\n\n\n\nThe asterisk (*) denotes the active site of the catalyst.The free energy for each reaction intermediate is defined as:\n\n(i)\n\n\nG\n=\n\n\nE\n\n\nD\nF\nT\n\n\n\n+\n\n\nE\n\n\nZ\nP\nE\n\n\n\u2212\nT\nS\n+\n\n\nE\n\n\ns\no\nl\n\n\n\n\n\n\n\n\n\n\n\nE\n\n\nD\nF\nT\n\n\n\n is the electronic energy calculated by DFT, \n\n\n\nE\n\n\nZ\nP\nE\n\n\n\n denotes the zero point energy estimated within the harmonic approximation, and \n\nT\nS\n\n is the entropy at 298.15 K (T\u00a0=\u00a0298.15 K). The \n\n\u00a0\n\n\nE\n\n\nZ\nP\nE\n\n\n\n and \n\nT\nS\n\n of gas-phase molecules and reaction intermediates are listed in Table S3. For the concerted proton-electron transfer, the free energy of a pair of proton and electron (\n\n\n\nH\n\n\n+\n\n\n+\n\n\ne\n\n\n\u2212\n\n\n\n) was calculated as a function of applied potential relative to RHE (U versus RHE), i.e., \n\n\u03bc\n\n(\n\n\n\nH\n\n\n+\n\n\n\n)\n\n+\n\u03bc\n\n(\n\n\n\ne\n\n\n\u2212\n\n\n\n)\n\n=\n\n\n1\n\n\n2\n\n\n\u03bc\n\n(\n\n\n\nH\n\n\n2\n\n\n\n)\n\n\u2212\neU\n\n, according to the computational hydrogen electrode (CHE) model proposed by N\u00f8rskov.\n55\n In addition, the solvent effect is reported to play an important role in the ORR. In our calculations, the solvent corrections (\n\n\n\nE\n\n\ns\no\nl\n\n\n\n) for *OOH and *OH are 0.45 eV in accordance with previous studies.\n56\n\n,\n\n57\n We used the energies of H2O and H2 molecules calculated by DFT together with experimental formation energy of H2O (4.92 eV) to construct the free energy diagram. The free energies of O2, *OOH, *O, and *OH at a given potential U relative to RHE are defined as:\n\n(j)\n\n\n\u0394\nG\n\n(\n\n\n\nO\n\n\n2\n\n\n\n)\n\n=\n4.92\n\u2212\n4\neU\n\n\n\n\n\n\n(k)\n\n\n\u0394\nG\n\n(\n\nOOH\n\n)\n\n=\nG\n\n(\n\n\n\nOOH\n\n\n*\n\n\n\n)\n\n+\n\n\n3\nG\n\n(\n\n\n\nH\n\n\n2\n\n\n\n)\n\n\n\n2\n\n\n\u2212\nG\n\n(\n\n*\n\n)\n\n\u2212\n2\nG\n\n(\n\n\n\nH\n\n\n2\n\n\nO\n\n)\n\n\u2212\n3\neU\n\n\n\n\n\n\n(l)\n\n\n\u0394\nG\n\n(\n\nO\n\n)\n\n=\nG\n\n(\n\n\n\nO\n\n\n*\n\n\n\n)\n\n+\nG\n\n(\n\n\n\nH\n\n\n2\n\n\n\n)\n\n\u2212\nG\n\n(\n\n*\n\n)\n\n\u2212\nG\n\n(\n\n\n\nH\n\n\n2\n\n\nO\n\n)\n\n\u2212\n2\neU\n\n\n\n\n\n\n(m)\n\n\n\u0394\nG\n\n(\n\nOH\n\n)\n\n=\nG\n\n(\n\n\n\nOH\n\n\n*\n\n\n\n)\n\n+\n\n\nG\n\n(\n\n\n\nH\n\n\n2\n\n\n\n)\n\n\n\n2\n\n\n\u2212\nG\n\n(\n\n*\n\n)\n\n\u2212\nG\n\n(\n\n\n\nH\n\n\n2\n\n\nO\n\n)\n\n\u2212\neU\n\n\n\n\nWe would like to acknowledge funding support from the National Key R&D Program of China (2016YFA0202804), Singapore Ministry of Education Academic Research Fund (AcRF) Tier 1: RG10/16 and RG111/15, Tier 2: MOE2016-T2-2-004, and the financial support from Jiangsu Specially-Appointed Professor program.J.G., H.Y., and B.L. conceived and designed the project. X.H. carried out the theoretical calculations. S.H. and H.M.C. performed the X-ray absorption experiments. W.C., C.J., S.M., X.Y., and Y.H. contributed to the structure characterizations. J.G. and B.L. prepared the manuscript. All authors contributed and reviewed the manuscript.The authors declare no competing interests.Supplemental Information can be found online at https://doi.org/10.1016/j.chempr.2019.12.008.\n\n\nDocument S1. Figures S1\u2013S38, Tables S1\u2013S9, and Supplemental References\n\n\n\n\n\nDocument S2. Article plus Supplemental Information\n\n\n\n", "descript": "\n                  The electrochemical oxygen reduction reaction in acidic media offers an attractive route for direct hydrogen peroxide (H2O2) generation and on-site applications. Unfortunately there is still a lack of cost-effective electrocatalysts with high catalytic performance. Here, we theoretically designed and experimentally demonstrated that a cobalt single-atom catalyst (Co SAC) anchored in nitrogen-doped graphene, with optimized adsorption energy of the *OOH intermediate, exhibited a high H2O2 production rate, which even slightly outperformed the state-of-the-art noble-metal-based electrocatalysts. The kinetic current of H2O2 production over Co SAC could reach 1 \n                        \n                           \n                              \n                                 mA\n                              \n                              /\n                              \n                                 \n                                    \n                                       cm\n                                    \n                                    \n                                       disk\n                                    \n                                    \n                                       2\n                                    \n                                 \n                              \n                           \n                        \n                      at 0.6\u00a0V versus reversible hydrogen electrode in 0.1\u00a0M HClO4 with H2O2 faraday efficiency > 90%, and these performance measures could be sustained for 10\u00a0h without decay. Further kinetic analysis and operando X-ray absorption study combined with density functional theory (DFT) calculation demonstrated that the nitrogen-coordinated single Co atom was the active site and the reaction was rate-limited by the first electron transfer step.\n               "}
{"full_text": "Lignin is a complex three-dimensional natural polymer that is obtained from lignocellulose along with cellulose and hemicellulose [1\u20133] and contains the monomer units of p-coumaryl (H units), coniferyl (G units), and sinapyl alcohols (S units) [4\u20137]. Unlike the formation of aromatic compounds from methane and acetylene [8], the production of phenolics through lignin depolymerization does not involve the complex condensation of small molecules and is relatively straightforward, thus attracting much attention from researchers. In view of the high stability of the lignin structure, the corresponding depolymerization typically requires the use of catalysts and large amounts of suitable solvents at high temperatures/pressures (e.g., supercritical or near-supercritical solvents), which, together with the poor understanding of the related catalytic process, hinder industrial applications [1].Because of the poor solubility of lignin, several solvents including methanol [9,10], ethanol [10\u201312], 2-propanol [9], methylcylohexane [9], aqueous methanol [13\u201315], aqueous ethanol [10,11,13,16,17], tetrahydrofuran [10], amines [18], and others [19] were used for its dissolution. Lignin depolymerization is promoted by these organic solvents [9\u201313,18] and additives [14], and can be enhanced by improving the mobility of solvent-mixed lignin or providing a hydrogen supply through transfer hydrogenation. Although high-dilution conditions favor the formation of lignin oil and lignin-derived monomers, the associated need to handle large solvent quantities complicates the purification process and reduces economic feasibility. Typically, depolymerization is performed at lignin concentrations (typically 0.01\u20130.10\u00a0g/mL in a suitable solvent [9\u201312,15,17,20\u201330]) that are too low for industrial applications and may result in large operation costs associated with purification and solvent removal. Thus, the economic feasibility of lignin depolymerization can be enhanced through the use of higher lignin concentrations, which favor the production of phenolic monomers but lead to char formation and poor processability. Lignin depolymerization is promoted by acids [15,17] and certain metals, as exemplified by the reductive depolymerization of lignin catalyzed by noble and transition metals (Ru, Pd, Ni, Co, Re, Mo, and their mixtures) supported on carbons, zeolites, alumina, and other materials [1,20\u201325]. Among these catalysts, Ru/H-zeolite \u03b2 (H\u03b2) was reported to exhibit high lignin depolymerization activity in our previous studies [14,15,17]. In view of the above, a deep understanding of the roles of metals in lignin depolymerization is essential for the development of optimal depolymerization catalysts.The present study probes the effects of the lignin concentration on the outcome of catalytic lignin depolymerization, elucidates the roles of metal and acid sites in the related catalysts, and provides valuable mechanistic insights, thus paving the way to the efficient industrial depolymerization of lignin. Among the solvents suggested in the literature, aqueous methanol was selected for lignin dissolution based on the results of our previous studies [14,15]. Because of the possible methanol enhancement of the depolymerization reaction [31,32], lignin depolymerization by methanol was also performed without heterogeneous catalysts to determine the catalyst contributions to the lignin depolymerization reaction.All chemicals were used as received without further purification unless otherwise noted. Kraft lignin (KL), tetraamminepalladium(II) nitrate ([Pd(NH3)4](NO3)2, anhydrous pyridine (99.8%), acetic anhydride (99%), fumed silica powder (SiO2, 0.007\u00a0\u00b5m), and titania (TiO2, nanopowder, 21\u00a0nm) were purchased from Sigma-Aldrich (St. Louis, Missouri, USA). Oak-extracted organosolv lignin (OL) was purchased from Sugaren (Yongin, Gyeonggi-do,\u00a0Korea). Sulfuric acid was purchased from J. T. Baker (Phillipsburg, New Jersey, USA). THF for gel permeation chromatography (GPC) was purchased from Honeywell B&J (Morris Plains, New Jersey, USA). Zeolite \u03b2 (ammonium form, Si/Al2 = 25 mol/mol) was purchased from Zeolyst (Conshohocken, Pennsylvania, USA). Zirconia (ZrO2) and alumina (\u03b3-form, Al2O3) were purchased from Alfa Aesar (Havenhill, Massachusetts, USA). 1-Butanol (99%) was purchased from Junsei (Tokyo, Japan). Hydrogen gas (H2, 99.999%), nitrogen gas (N2, 99.9%), H2/Ar (5% v/v), O2/N2 (0.5% v/v), and CO/He (10% v/v) were purchased from Shinyang Medicine (Anseong, Gyeonggi-do, Korea).The content (wt%)\u00a0of acid-insoluble lignin was determined using a two-step hydrolysis process as reported elsewhere [33,34]. A dry sample (0.5\u00a0g) was treated with 72\u00a0wt% aqueous sulfuric acid at 30\u00a0\u00b0C (water bath) for 1\u00a0h upon stirring with a glass rod at 10\u00a0min intervals. The mixture was supplemented with deionized water (84\u00a0mL; 18.2\u00a0M\u03a9\u00b7cm), heated to 121\u00a0\u00b0C for 2\u00a0h, slowly cooled to room temperature, and filtered, and the residue (acid-insoluble lignin) was dried and weighed. All processes were performed using pressure tubes (8648\u201330 ACE glass) and screw-on Teflon caps with O-ring seals.H-zeolite \u03b2 (\u0397\u03b2) was prepared by calcining the ammonium form of zeolite \u03b2 at 550\u00a0\u00b0C for 3\u00a0h. The support powders, including H\u03b2, Al2O3, TiO2, ZrO2, and SiO2, were sieved to obtain particles smaller than 150\u00a0\u00b5m and impregnated with 1\u00a0wt% Pd using a wet method\u00a0[14,15]. The impregnated catalysts were calcined in air at 400\u00a0\u00b0C for 1\u00a0h, reduced in a flow of H2/Ar (5% v/v) at 400\u00a0\u00b0C for 1\u00a0h, passivated in a flow of O2/N2 (0.5% v/v) at room temperature for 30\u00a0min, and stored under ambient conditions before use. For manipulating the metal dispersion, the reduction time and temperature were adjusted as follows: 400\u00a0\u00b0C, 1\u00a0h for Pd/SiO2-A; 400\u00a0\u00b0C, 3\u00a0h for Pd/SiO2-B; 600\u00a0\u00b0C, 1\u00a0h for Pd/SiO2-C; 600\u00a0\u00b0C, 3\u00a0h for Pd/SiO2-D; and 800\u00a0\u00b0C, 1\u00a0h for Pd/SiO2-E.The catalyst acidity was evaluated by the temperature-programmed desorption of ammonia (NH3 TPD) using a BELCAT-B instrument (MicrotracBel, Osaka, Japan) equipped with a thermal conductivity detector (TCD) and interfaced with a BELMass quadrupole mass spectrometer (MicrotracBel, Osaka, Japan). The catalyst powder (approximately 50\u00a0mg) was degassed in a flow of He (50\u00a0mL/min) at 500\u00a0\u00b0C for 60\u00a0min and passivated in NH3/He (5% v/v) at 100\u00a0\u00b0C for 30\u00a0min. The TCD was stabilized at 100\u00a0\u00b0C for 60\u00a0min in a flow of He (30\u00a0mL/min), and the amount of NH3 desorbed upon heating from 100\u00a0\u00b0C to 700\u00a0\u00b0C at 5\u00a0\u00b0C/min in a flow of He (30\u00a0mL/min) was recorded using the TCD and the mass spectrometer (MS). CO chemisorption measurements were performed using a BELCAT-M instrument (MicrotracBel, Osaka, Japan) equipped with a TCD. The catalyst powder (50\u00a0mg) was placed in a quartz reactor and heated to 400\u00a0\u00b0C at 10\u00a0\u00b0C/min in a flow of He (50\u00a0mL/min), oxidized in a flow of O2/He (5% v/v, 50\u00a0mL/min), and reduced in a flow of H2/Ar (5% v/v, 50\u00a0mL/min) at 400\u00a0\u00b0C for 15\u00a0min. He gas was used to remove O2 and H2 before measurements. The treated catalyst powder was cooled to 50\u00a0\u00b0C in a flow of He (50\u00a0mL/min), and pulses of CO/He (10% v/v) were injected to quantify the surface metal sites. The Brunauer\u2013Emmett\u2013Teller (BET) surface areas, pore volumes, and Barrett\u2013Joyner\u2013Halenda (BJH) pore size distributions of the catalysts were measured using N2 physisorption at 77\u00a0K (ASAP 2020, Micromeritics, Norcross, Georgia, USA). High-angle annular dark field (HAADF)\u2013scanning transmission electron microscopy (STEM) images were acquired using a transmission electron microscope (Talos F200X, FEI, Hillsboro, Oregon, USA). Elemental compositions of the catalysts were measured by inductively coupled plasma\u2013optical emission spectrometry (ICP\u2013OES, iCAP 6000 Series, Thermo Fisher Scientific, Waltham, Massachusetts, USA). The crystal structures of the catalysts were probed by X-ray diffraction (XRD) using a Bruker D8 Advance diffractometer (Billerica, Massachusetts, USA) operated at 40\u00a0kV and 40\u00a0mA (Cu K\u03b11 radiation, \u03bb\u00a0=\u00a01.54059\u00a0\u00c5).The catalytic depolymerization of lignin was performed in a 100\u00a0mL stainless steel batch autoclave reactor. In a typical process (\nFig. 1), dry lignin (1, 2, 6, or 12\u00a0g) was supplemented with aqueous methanol (30\u00a0mL, 65% v/v) and catalyst powder (0.2\u00a0g). The reactor was purged three times with H2 at room temperature and pressurized to 50\u00a0bar. The reaction mixture was heated to 280\u00a0\u00b0C for 2\u00a0h upon agitation at 500\u00a0rpm and cooled to room temperature. The liquid product (LP) was collected and analyzed by gas chromatography-mass spectrometry (GC-MS) and gas chromatography with flame ionization detector (GC-FID) using 1-butanol as an internal standard. The method of effective carbon numbers was used to quantify the lignin-derived monomers. The LP was concentrated using a rotary evaporator and further dried at 55\u00a0\u00b0C in a vacuum furnace for 16\u00a0h to obtain the mass of the depolymerized lignin (DL). The product yield (%) was calculated as [mass of product (g)]/[mass of lignin (g)] \u00d7\u00a0100, and the selectivity for component i (%) was calculated as [mass of component i (g)]/[mass of LP (g)] \u00d7\u00a0100. The dispersion remaining after LP isolation was centrifuged to collect the spent catalyst and precipitate solid products, which were further dried at 55\u00a0\u00b0C in a vacuum for 16\u00a0h to obtain the solid residue (SR) mass.The phenolic monomers in the LP were identified using a GC-MS instrument (Agilent 7890\u2009A, 5975\u2009C inert MS XLD) equipped with an HP-5MS capillary column (60\u2009m\u2009\u00d7\u20090.25\u2009mm \u00d7 250\u2009\u00b5m) and quantified using a GC-FID instrument (YoungLin 6500) equipped with an HP-5MS capillary column (60\u2009m\u2009\u00d7\u20090.25\u2009mm \u00d7 250\u2009\u00b5m). Prior to injection into the GC instrument, the LP was filtered using a 0.45\u2009\u00b5m Whatman syringe filter.The DL was dissolved in a mixture of acetic anhydride and pyridine (10\u2009mL, 1:1\u2009v/v), and the solution was stirred for 24\u2009h under ambient conditions, supplemented with ethanol (20\u2009mL), and further stirred for 30\u2009min. The solvent was removed in vacuo at 60\u2009\u00b0C using a rotary evaporator, and the residue was dried in a vacuum oven at 55\u2009\u00b0C for 16\u2009h to afford acetylated lignin. The number-average molecular weight (Mn), weight-average molecular weight (Mw), and polydispersity index (PDI) of the OL, KL, and DL were determined using a GPC device (Agilent 1200 HPLC) equipped with two Shodex LF-804 columns and a UV detector (\u03bb\u2009=\u2009270\u2009nm). The flow rate of the THF eluent equaled 1\u2009mL/min, and the column was calibrated using a polystyrene standard ReadyCal set (Sigma-Aldrich, 250\u20132,000,000\u2009g/mol). Acetylated lignins were dissolved in THF at a concentration of 1\u2009g/L and filtered using a 0.45-\u03bcm Whatman syringe filter prior to injection into the GPC system. The utilized lignin acetylation procedure was described in our previous studies [15,17].The FT-IR spectra of OL and its depolymerization products were recorded on a Nicolet iS10 spectrometer (Thermo Scientific, Waltham, MA, USA) equipped with a Smart Miracle accessory in the transmission mode at UNIST Catalysis Research Center (Ulsan, Korea).Prior to depolymerization, the compositions of KL and OL were measured as reported elsewhere [33,34]. The contents of acid-insoluble lignin in KL and OL were determined as 93.9 wt% and 94.2\u2009wt%, respectively (\nTable 1). The moisture contained in KL (3.6\u2009wt%) and OL (4.5\u2009wt%) was removed by drying at 105\u2009\u00b0C for 2\u2009h before further processing.The reductive depolymerization of lignin was performed over Pd catalysts deposited on several supports, and the effects of the lignin concentration (expressed as the lignin/solvent or lignin/catalyst ratio) on the depolymerization activity were examined. The solvent and catalyst amounts were fixed while the lignin amount was varied. The distillable phenolic compounds in the LP were characterized by GC-MS and GC-FID (\nFig. 2 and Tables S1, S2). OL depolymerization afforded derivatives of G units (selectivity: 27\u201353%) and S units (selectivity: 43\u201367%) as well as in small quantities of heavier products detected at higher retention times. In contrast, KL depolymerization produced G units (selectivity: 63\u201390%), 4-ethylphenol (H units), and heavier products, whereas S units were not detected. The observed differences in product distributions can be attributed to the biomass origin [35]. The lignin extracted from softwood mainly contained G units (90\u201395%), and the lignin extracted from hardwood contained G (25\u201350%) and S (50\u201375%) units. Grass usually contains G, H, and S units. The depolymerization of OL extracted from oak formed both G and S units in this study, confirming that OL was prepared from hardwood oak. Although the origin of KL was not specified by the manufacturer, the depolymerization products of KL (G units) indicate that it was likely obtained from softwood.Compared to the heterogeneous reaction over supported Pd catalysts, the non-catalytic reaction afforded phenolic monomers in smaller yields of 4.8\u201316.9\u2009wt% (OL) and 4.1\u20137.1\u2009wt% (KL) (lignin/solvent = 0.033\u20130.4\u2009g/mL for both lignin reactants), possibly via solvothermal depolymerization [28\u201330,36]. In the case of OL, a larger yield improvement was observed at low lignin/solvent or lignin/catalyst ratios, which was ascribed to the better solubility of OL (extracted using aqueous alcohol) in aqueous methanol. With increasing lignin concentration, the yield of phenolics decreased for both OL and KL. The effects of lignin concentration on catalytic activity are further discussed in Section 3.5.In the case of OL, the combined selectivities for 4-propylguaiacol and 4-propylsyringol equaled 23\u201331% at a lignin/solvent ratio of 0.067\u2009g/mL or a lignin/catalyst ratio of 10 w/w (Table S1). Whereas the detailed distributions of phenolic monomers in the LP did not significantly depend on the catalyst type, high selectivities for guaiacol (7\u201320%) and syringol (7\u201318%) and low selectivities for 4-propylguaiacol (4\u20138%) and 4-propylsyringol (5\u201313%) were observed at all lignin/solvent ratios in the absence of a catalyst (Fig. 2 and Table S1). In addition, under the conditions of high dilution (lignin/solvent = 0.033\u2009g/mL or lignin/catalyst = 5 w/w), the selectivities for 4-propylguaiacol and 4-propylsyringol increased to 10\u201318% and 26\u201333%, respectively. In the case of KL, the selectivity for 4-propylguaiacol decreased from 19\u201329% to 15\u201318% with increasing lignin/catalyst ratio (Table S2). The selectivity for heavy molecules (retention time > 33\u2009min) reached 21\u201336% for all catalytic reactions except for the case of Pd/Al2O3 at lignin/catalyst ratios of 10 and 30 w/w.The compositions of the solid- and liquid-phase products were analyzed further. Liquid-phase products correspond to lignin-derived monomers, dimers, oligomers, and the lignin polymer dissolved in aqueous methanol, while the SR products correspond to insoluble lignin polymer. Solvent removal from the LP afforded DL as a dry powder. The yield of DL equaled 53\u201362% at a lignin/solvent ratio of 0.033\u2009g/mL and decreased to 22\u201330% when this ratio increased to 0.4\u2009g/mL (\nFig. 3). In the case of KL, the DL yield was lower than that obtained for OL, equaling 39\u201352% and 16\u201321% at lignin/solvent ratios of 0.033 and 0.4\u2009g/mL, respectively. The yield of SR increased with increasing lignin concentration (i.e., with increasing lignin/solvent or lignin/catalyst ratio), equaling 58\u201364% and 69\u201375% for OL and KL, respectively. In the case of OL, the DL yield obtained under catalyst-free conditions was lower than those obtained for catalytic reactions. The DL yield was highest for Pd/Al2O3, Pd/ H\u03b2, and Pd/ZrO2 at lignin/solvent ratios of 0.033 and 0.1\u2009g/mL, exceeding that obtained under catalyst-free conditions by 9\u2009wt%. In the case of KL, the DL yield was highest for Pd/SiO2 at a lignin/solvent ratio of 0.067\u2009g/mL, exceeding that obtained under non-catalytic conditions by 15\u2009wt% and those obtained for Pd/Al2O3, Pd/H\u03b2, Pd/TiO2, and Pd/ZrO2 by 7\u201319\u2009wt%.\nFigs. S1, S2, and \nTable 2 present the GPC results of acetylated DL, revealing that depolymerization decreased the Mw of OL and KL by 67\u201383 and 75\u201384%, respectively. The PDI decreased from 3.5 to 1.8\u20132.7 for OL and from 4.4 to 1.7\u20132.2 for KL because of the significant loss of high-molecular-weight compounds and the formation of low-molecular-weight compounds (Figs. S1 and S2). Under non-catalytic conditions, the largest Mw decrease was observed at the highest lignin concentration (lignin/solvnet = 0.4 g/mL) for both OL and KL. As the highest SR yield and the lowest DL yield were also observed at the highest lignin concentration (lignin/solvent = 0.4 g/mL), this behavior was attributed to the significant condensation and precipitation of heavier lignin polymers. For OL, the lowest decrease of Mw and the highest PDI (2.5\u20132.7) were observed at a lignin/catalyst ratio of 10\u201330 w/w (lignin/solvent = 0.067\u20130.2 g/mL), which was indicative of non-catalytic re-condensation during depolymerization. For KL, no such significant re-condensation was observed.The Mw values of DLs obtained under catalytic conditions were not significantly different from those obtained in the absence of catalysts through the lignin concentrations except for the highest lignin reactant concentration (lignin/solvent = 0.4 g/mL) for both OL and KL. These observations indicate that the catalysts suppressed the re-polymerization of heavier lignin polymers to increase the yields of DL and phenolic monomers. The smallest Mw values were observed for Pd/TiO2 in the case of OL and for Pd/ZrO2 and Pd/H\u03b2 in the case of KL, although the effects of the catalysts on the results of GPC analysis were not significant.The analysis of OL, KL, and DL by FT-IR spectroscopy revealed the presence of typical lignin functionalities and confirmed the formation of phenolics (Figs. S3 and S4). The results obtained for KL resembled those obtained for the corresponding DL. As DL is composed of phenolic monomers, these observations indicate that the functionalities of lignin polymers were not significantly different from those of the monomeric products in DL. The bands of C\u2013O bonds in syringyl and guaiacyl rings at 1330 and 1270\u2009cm\u22121, respectively, were observed only for OL and the corresponding DL. Additionally, C\u2013H (3000\u20132800\u2009cm\u22121), CO (1700\u2009cm\u22121), CC (1600\u2009cm\u22121), aromatic C\u2013C (1500\u2009cm\u22121), C\u2013H (1420\u20131440\u2009cm\u22121), and C\u2013C (1200\u2009cm\u22121) stretches were observed for all lignin reactants and products in this study. The C\u2013O peaks of aliphatic ethers at 1150 and 1110\u2009cm\u22121 were more pronounced for OL than for KL.The surface areas of Pd particles and catalyst pore structures were probed by N2 physisorption, CO chemisorption, and NH3 TPD measurements (\nTable 3, Figs. S5\u2013S7). Pd dispersion, expressed as [CO]/[Pd] (mol/mol), decreased in the order of Pd/SiO2 >\u2009Pd/ZrO2 \u2265\u2009Pd/Al2O3 >\u2009Pd/TiO2 >\u2009Pd/H\u03b2, while the quantity of surface acid sites per catalyst mass decreased in the order of Pd/H\u03b2 >\u2009Pd/ZrO2 >\u2009Pd/TiO2 >\u2009Pd/Al2O3 >\u2009Pd/SiO2. Notably, the latter parameter was highest for Pd/H\u03b2, which contained less dispersed Pd, and was lowest for Pd/SiO2, which contained highly dispersed Pd. Interestingly, for a low concentration of OL (0.033\u2009g/mL), which did not contain the catalyst-poisoning sulfur, the smallest monomer yield of 15.8\u2009wt% (Fig. 1 and Table S1) was observed in the case of the Pd/H\u03b2 catalyst with abundant acid sites. This yield was smaller than that obtained under non-catalytic conditions. In the case of OL, high monomer yields were obtained for the Pd/Al2O3, Pd/SiO2, and Pd/TiO2 catalysts with low amounts of acid sites. These observations indicate that the lignin depolymerization activity is determined not only by the quantity of acid sites but also by the properties of the surface metal sites [15,17].The XRD results of all catalysts exhibited no Pd peaks because of the low metal loading (Fig. S8). Moreover, Pd deposition did not markedly affect the XRD results of supports, i.e., the support structure did not change during catalyst preparation. The result of Pd/Al2O3 featured the peaks of boehmite (AlO(OH)) and \u03b3-Al2O3, while that of Pd/SiO2 featured a broad peak of amorphous silica at 2\u03b8\u2009=\u200922\u00b0. Although the catalysts were annealed at 400\u2009\u00b0C, the formation of crystalline silica (e.g., quartz) was not observed. The result of Pd/H\u03b2 exhibited the broad peaks of zeolite \u03b2, indicating the presence of zeolite nanoparticles, while the result of Pd/TiO2 revealed the presence of both rutile and anatase phases. Finally, the result of Pd/ZrO2 featured the broad peaks of monoclinic ZrO2, thus indicating the presence of ZrO2 nanocrystals.The influence of the lignin concentration on the depolymerization activity was probed to shed light on the roles of acid and metal sites. The yield of phenolic monomers decreased with increasing lignin concentration for both OL and KL (\nFig. 4(a, b) and \nTable 4; based on Fig. 2). Under dilute conditions (lignin/solvent = 0.033\u2009g/mL), higher yields were observed for OL than for KL, whereas no marked difference between OL and KL was observed at high lignin concentrations (lignin/solvent = 0.067\u20130.4\u2009g/mL; yield = 4.8\u201312.6% (OL) and 5.6\u201313.4% (KL)). Notably, higher yields of phenolic monomers were observed in the presence of catalysts than in their absence. At high lignin concentrations, the largest increase in the phenolic monomer yield for both OL and KL was observed in the case of Pd/SiO2, while Pd/Al2O3 also achieved the highest yield of phenolic monomers for OL. Lignin concentrations above 0.4\u2009g/mL were not investigated, as the agitation of the corresponding slurries was difficult.The results of depolymerization were described in terms of the phenolic monomer amount produced per catalyst weight (Fig. 4(c, d)), which increased with increasing lignin concentration up to 0.4 g/mL. These observations indicate that high lignin concentrations (up to 0.4\u2009g/mL) did not significantly deactive the reaction on the catalyst surface, although the pore structures and BET surface areas of the catalysts significantly differred depending on the support type (Table 3). The results also indicated that the production of lignin-derived phenolic monomers can be enhanced by increasing the lignin concentration to up to 0.4\u2009g/mL. Notably, reactions at lignin concentrations above 0.4\u2009g/mL were difficult to perform because of the insufficient wetting of lignin by the solvent.The quantity of phenolic monomers produced per acid site was positively correlated with [CO]/[Pd] (Fig. 4(e)), which indicated that high amounts of Pd atoms on the surface facilitate depolymerization, possibly by favoring hydrogen adsorption. At lower [CO]/[Pd] values, the amount of phenolic monomers produced per surface Pd atom significantly decreased, which indicated that lignin depolymerization can be significantly suppressed by the use of lower [CO]/[Pd] values, lower amounts of surface Pd atoms, or acid sites without Pd atoms. These observations indicate that Pd is required for the catalytic depolymerization of lignin, although acid sites are also required, as demonstrated in our previous studies [15,17].The roles of metals were further studied using the correlation between the weight of the phenolic monomers produced per surface Pd site and acid site quantity (Fig. 4(f)). This correlation was positive, which indicated that acid sites facilitate lignin depolymerization. Interestingly, a certain catalytic activity per Pd atom was observed when the quantity of acid sites approached zero, which indicated that Pd atoms could depolymerize lignin even in the absence of acid sites.Based on the above observations, we further probed the effects of Pd on catalytic depolymerization. In view of the fact that acid-induced depolymerization was observed in our previous studies [15,17], Pd/SiO2 without significant acidity was selected to hinder this depolymerization. The properties of the Pd metal dispersions ([CO]/[Pd]) of Pd/SiO2-A, B, C, D, E are described in \nTable 5. While the BET surface areas and BJH pore size distributions of these species were not significantly different (Figs. S9 and S10), their [CO]/[Pd] values varied between 0.129 and 0.602\u2009mol/mol. HAADF-STEM imaging confirmed the results of particle size measurements obtained using CO chemisorption (\nFig. 5).In the case of OL depolymerization promoted by Pd/SiO2, higher yields of phenolic monomers or larger quantities of phenolic monomers per surface Pd were observed at lower [CO]/[Pd] values (\nFig. 6), with the largest yield of 16.4% observed at the lowest [CO]/[Pd] value of 0.129\u2009mol/mol. Notably, catalysts containing well-dispersed Pd ([CO]/[Pd] = 0.602\u2009mol/mol) achieved monomer yields of only 5\u20136\u2009wt%, which were not significantly different from those obtained under non-catalytic conditions and slightly exceeded that obtained using a metal-free SiO2 support (3.4\u2009wt%, data not shown).The analysis of the quantity of produced phenolic monomers per surface Pd also demonstrated that the depolymerization activity increased with decreasing [CO]/[Pd]. Notably, a 14-fold activity difference was observed between catalysts with the highest and lowest [CO]/[Pd] values: Pd/SiO2 with [CO]/[Pd] =\u20090.602\u2009mol/mol achieved a yield of 94\u2009g monomer/g surface Pd, while Pd/SiO2 with [CO]/[Pd] =\u20090.129\u2009mol/mol achieved a yield of 1324\u2009g monomer/g surface Pd.The above observations indicated that catalytic depolymerization on the surface of bulk Pd occurs more easily than on the surface of small Pd particles (\nFig. 7), which was ascribed to the better decomposition of multidentate lignin that securely adsorbed on the multiple sites of the bulk Pd surface. The results of GPC analysis indicated that Pd/SiO2-E, exhibiting the lowest [CO]/[Pd], produced DL with the smallest Mw and thus exhibited the highest depolymerization activity (Fig. S11 and Table S3).Herein, we examined the catalytic depolymerization of concentrated lignin and elucidated the roles of Pd metal in this process. Although acid sites have previously been reported to promote lignin depolymerization, efficient lignin depolymerization required the presence of Pd, as indicated by the correlation between the lignin depolymerization performance and reaction conditions. The Pd atoms exhibited a certain depolymerization activity even at a negligible content of acid sites, while depolymerization on acid sites required a certain content of surface Pd atoms. Regarding the effects of Pd dispersion, bulk Pd atoms featured a 14 times higher catalytic activity than highly dispersed ones, which indicated that efficient depolymerization required the multidentate adsorption of the bulky lignin polymer on the Pd surface and suggested the superiority of bulk Pd atoms as catalysts. Thus, the presented results pave the way to the efficient utilization of lignin for the production of phenolic monomers.\nAliaksandr Karnitski: Investigation, Writing \u2013 original draft. Jae-Wook Choi: Methodology, Visualization. Dong Jin Suh: Conceptualization. Chun-Jae Yoo: Formal analysis. Hyunjoo Lee: Supervision, Validation. Kwang Ho Kim: Validation. Chang Soo Kim: Formal analysis. Kyeongsu Kim: Validation. Jeong-Myeong Ha: Writing \u2013 review & editing, Conceptualization, Methodology.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This research was supported by the Institutional Program of Korea Institute of Science and Technology (KIST) of Republic of Korea (2E31853) and the Technology Development Program to Solve Climate Change of the National Research Foundation (NRF) funded by the Ministry of Science and ICT of Republic of Korea (NRF-2020M1A2A2079798).Supplementary data associated with this article can be found in the online version at doi:10.1016/j.cattod.2022.07.012.\n\n\n\nSupplementary material\n\n\n\n.", "descript": "\n                  Lignin is a natural polymer contained in lignocellulose and is a potential feedstock for the production of phenolics or small aromatic molecules. However, lignin depolymerization is typically performed using low lignin concentrations and insufficiently active catalysts, which hinders industrial applications. To address this issue, we herein investigated the depolymerization of concentrated kraft lignin and oak-extracted organosolv lignin promoted by supported bifunctional metal (palladium)-acid catalysts and elucidated the roles of acid and metal sites. In addition to acid sites, which are known to be critical for lignin degradation, metal sites were found to be required to initiate lignin depolymerization. Palladium atoms promoted depolymerization even in the absence of acid sites, while acid site\u2013promoted depolymerization required the presence of surface palladium atoms. Catalysts with bulk palladium particles were much more active than those with highly dispersed palladium particles, which suggested that efficient depolymerization required the bulky lignin polymer to be adsorbed on the palladium surface in a multidentate fashion. Thus, our work provides valuable insights into the mechanism of catalytic lignin depolymerization and paves the way to the industrial-scale application of this process.\n               "}
{"full_text": "Alternatives to traditional oil-derived compounds, like biorenewable compounds, are necessary due to their environmental benefits and due to limitations on fossil fuel resources [1,2]. Several processes have been developed for biodiesel production because it is a renewable, biodegradable, eco-friendly, and nontoxic fuel [3]. Biodiesel consists of fatty acid methyl esters (FAMEs) derived from the transesterification of vegetable oils or animal fats [2], but biodiesel has some problems, such as a lack of stability, limited use under cold conditions, and a slight increase in NOx emissions [4]. Green fuels, resulting from the hydroprocessing of carboxylic esters or triglycerides, are promising alternatives to conventional fuels [5]. The conversion process is called the hydrotreatment of vegetable oils of fats (HVO). Green fuel can also be produced through Fischer-Tropsch synthesis, and the process is known as gas to liquids (GTL). All these processes produce linear hydrocarbons whose properties have to be improved (particularly, their high cloud points need to be reduced) for them to be valid as hydrocarbon-based fuel [6]. For this reason, the hydroisomerization of linear alkanes has been previously studied. Thus, during hydroisomerization, the properties of the feedstock are improved by transforming normal hydrocarbons to branched hydrocarbons containing an identical number of carbon atoms [7].The isomerization process, known as hydroisomerization, usually takes place in the presence of hydrogen, and it transforms a molecule into an isomer with a different chemical structure. The hydroisomerization of hydrocarbons is usually divided into different fractions, such as C4-7, C7-15, and more than C15\n[8]. Long-chain hydrocarbons have attracted much attention recently because of their increased availability. These long-chain hydrocarbons or n-paraffins (>C15) comprise up to 80% of the material known as wax [7,8]. The hydroisomerization reaction is always accompanied by a hydrocracking reaction that lowers yield. First, isomerization occurs, and then cracking, which is an undesirable and competitive process that favors multibranched alkanes.Bifunctional catalysts containing metallic sites, including Pt, Pd, or Ni, for hydrogenation/dehydrogenation, and acidic sites for skeletal isomerization via carbenium ions are usually involved in isomerization reactions [8]. First, a linear olefin is formed by the dehydrogenation of an n-alkane; then, this compound is isomerized and finally hydrogenated, giving the isoalkane. Metallic sites also activate and dissociate H2, which plays an important role in decreasing cracking.Several studies are published in the literature about hydroisomerization reaction of alkanes using bifunctional catalysts, specially supported on zirconia. It has been established that the selectivity of paraffin isomerization depends primarily on a proper balance between metal and acid functions. WO3 and Pt are commonly used as acidic function and metallic site respectively. Although numerous published studies have reported the physicochemical properties and activity of WO3/ZrO2 catalysts, the catalytic activity of Pt/WOx/ZrO2 and Pt/WOx/Al2O3 has not been carefully studied for the n-dodecane hydroisomerization reaction. No publications comparing two different preparation methods, two different supports and different W loading at the same time were found for this reaction. This work provides insights contributing to the knowledge of catalyst preparation in the sense that no other articles were found carring out this procedure. The reason why the several consecutive impregnation method was carried out is because acidity plays a main role in the hydroisomerization reaction. Acidity favors the stabilization of carbocationic reaction intermediates and thus, the reaction. Since in our catalysts, we introduce the acidity through the incorporation of tungsten oxide, it could make important differences in the catalysts the different way of introducing tungsten oxide on the supports. Once this phase is introduced, it remains as tungsten oxide on the catalysts which would increase or decrease the acidity depending on the WOx particle size and interaction with the supports surface. This produces significantly changes in the conversion results.Considering the results published in the literature for long chain paraffin hydroisomerization and our previous study on the influence of the reduction temperature and nature of the support on hydroisomerization [6], this work aims to study the influence of W loading, preparation methothithi, and nature of the support during the hydroisomerization of n-dodecane in order to improve the quality of linear alkane fuels obtained by HVO or Fischer-Tropsch synthesis, producing a suitable fuel.Thus, in this study, the hydroisomerization of n-dodecane is mainly discussed with respect to the use of bifunctional catalysts with different W loadings (3\u201318\u00a0wt% W). Pt supported (0.3\u00a0wt% Pt) on W-modified alumina or zirconia were used as the catalysts. Correlations between activity and characterization results were studied to establish the properties that predict the best W loading, support type, and preparation method, contributing to the progress in this field.The following salts were used as W and Pt precursors: ammonium metatungstate hydrate ((NH4)6(H2W12O40)\u00b7xH2O) (99%) was purchased from Honeywell, and tetraammineplatinum (II) hydroxide hydrate (Pt 58%) (H14N4O2Pt\u00b7xH2O) was purchased from Alfa Aesar (Thermo Fisher Scientific, Waltham, USA). Concerning the supports, alumina (\u03b3-Al2O3) was purchased from Saint Gobain-NORPRO (Stow, USA) (1.1\u00a0\u00d7\u00a02\u00a0mm trilobes, SA6975), and the zirconium oxide (2x5 mm pellets, Lot Number X28A052) was purchased from Alfa Aesar (Thermo Fisher Scientific, Waltham, USA).The catalysts were prepared by sequentially impregnating (sequential wetness impregnation method) W and Pt on the supports. Before any impregnation, all supports were dried overnight at 120\u00a0\u00b0C to remove excess moisture. The catalysts were prepared by the following two procedures:\n\n1)\nW loading (containing 3, 6, 9, 12, 15, and 18\u00a0wt% W) was incorporated in one step.\n\n\nW loading (containing 3, 6, 9, 12, 15, and 18\u00a0wt% W) was incorporated in one step.The incorporation of W into ZrO2 and \u03b3-Al2O3 pellets by wet impregnation was achieved following this procedure: the dried pellets were placed in a round flask in contact with an aqueous solution (10\u00a0mL for 6\u00a0g of ZrO2 and 20\u00a0mL for 6\u00a0g of Al2O3) of ammonium metatungstate hydrate ((NH4)6(H2W12O40)\u00b7xH2O) under stirring for 1\u00a0h in a rotary evaporator. Then, the solvent was evaporated under vacuum at 65\u00a0\u00b0C in a rotary evaporator for 20\u00a0min, and finally, the recovered solid was calcined under static conditions at 500\u00a0\u00b0C for 2\u00a0h.The incorporation of Pt by wet impregnation was conducted using this protocol: dried pellets with the corresponding W loading were placed in a round flask in contact with an aqueous solution (10\u00a0mL for 6\u00a0g ZrO2 and 20\u00a0mL for 6\u00a0g Al2O3) of tetraammineplatinum (II) hydroxide hydrate (H14N4O2Pt\u00b7xH2O) under stirring for 1\u00a0h in a rotary evaporator. Then, the solvent was evaporated under vacuum in a rotary evaporator at 65\u00a0\u00b0C for 20\u00a0min, and finally, the recovered solid was calcined in air at 450\u00a0\u00b0C for 3\u00a0h [9].\n\n2)\nW loading (containing 6, 9, 12, and 15\u00a0wt% W) was incorporated in several consecutive steps. At every step, 3\u00a0wt% W was incorporated until reaching the desired loading following the same overall procedure previously described for W and Pt incorporation in a single step.\n\n\nW loading (containing 6, 9, 12, and 15\u00a0wt% W) was incorporated in several consecutive steps. At every step, 3\u00a0wt% W was incorporated until reaching the desired loading following the same overall procedure previously described for W and Pt incorporation in a single step.Following these procedures, the catalysts (Pt/WO3/ZrO2 or Pt/WO3/Al2O3) were prepared from their corresponding supports (ZrO2 or Al2O3) and denoted as PtW/Zr-[wt.%] or PtW/Al-[wt.%] where [wt.%] indicates the wt.% of W (mainly as WO3 oxide) on the catalysts. Catalysts prepared by successive impregnations are denoted with SI acronyms at the end. The Pt loading was 0.3\u00a0wt% for all the catalysts prepared using a single step or several consecutive steps. The catalysts activation or reduction process (350\u00a0\u00b0C in hydrogen at atmospheric pressure) was performed previously to the reaction in the reaction system.Textural properties of the catalysts were determined from the adsorption\u2013desorption isotherms of nitrogen, recorded at \u2212196\u00a0\u00b0C with a Micromeritics ASAP 2420 (Norcross, USA). The specific area was calculated by applying the BET method in the relative pressure (P/P0) range of the isotherms between 0.03 and 0.3 and taking a value of 0.162\u00a0nm2 for the cross-section of an adsorbed nitrogen molecule at \u2212196\u00a0\u00b0C. Pore size distributions were computed by applying the BJH model to the desorption curve of the nitrogen isotherms.X-ray diffraction profiles of the samples were recorded with an X\u2019Pert Pro PANalytical diffractometer (Almelo, Netherlands) equipped with a CuK\u03b1 radiation source (\u03bb\u00a0=\u00a00.15418\u00a0nm) and an X\u2019Celerator detector based on RTMS (real-time multiple strip). The samples were ground and placed on a stainless-steel plate. The diffraction patterns were recorded in steps over a range of Bragg angles (2\u03b8) between 4 and 90\u00b0 at a scanning rate of 0.04\u00b0 per step and an accumulation time of 20\u00a0s. Diffractograms were analyzed with a X\u2019Pert HighScore Plus software. The mean domain size was then estimated from X-ray linewidth broadening using the Scherrer equation. Width (t) was taken as the full width at half maximum intensity of the most intense and least overlapped peak.Raman spectra of the samples were recorded in air under ambient conditions with a Renishaw in Via Raman Microscope spectrometer (Gloucester, United Kingdom) equipped with a single monochromator, a laser beam emitting 785\u00a0nm and 300 mW output power, a thermoelectrically cooled CCD detector and a holographic super-Notch filter. The photons scattered by the sample were dispersed by a 1200 lines/mm grating monochromator and simultaneously collected on a CCD camera. The collection optic was set at 20x and 50x objectives. These measurement conditions gave a Raman shift within an accuracy lower than 0.1/cm.Metal dispersions were determined by CO pulse chemisorption. CO uptake was measured using a Micromeritics Autochem II 2920 apparatus (Norcross, USA). A 100\u2013200\u00a0mg portion of the reduced and passivated sample was loaded in the reactor and reduced under a H2 flow (50\u00a0mL\u00b7min\u22121) at 250\u00a0\u00b0C for 1\u00a0h (ramp, 10\u00a0\u00b0C \u00b7min\u22121). Afterward, the sample was cooled to 40\u00a0\u00b0C while it was flushed with a He flow (50\u00a0mL\u00b7min\u22121). When the TCD signal was stable, pulses of CO (75\u00a0\u03bcL) were passed through the samples until the areas of consecutive pulses were constant. The total CO uptake was then calculated.\n\n\nD\ni\ns\np\ne\nr\ns\ni\no\nn\n\n%\n\n=\n\n\nC\nO\n\nu\np\nt\na\nk\ne\n\n\nmmol\ng\n\n\n\n\nM\ne\nt\na\nl\n\nl\no\na\nd\ni\nn\ng\n\n\nmmol\ng\n\n\n\n\n\n\n\nThe acidity of the catalysts was measured by the temperature-programmed desorption of ammonia (NH3-TPD). NH3-TPD was carried out using a Micromeritics Autochem II 2920 apparatus (Norcross, USA). A 110\u2013170\u00a0mg sample was reduced with a H2 flow (50\u00a0mL\u00b7min\u22121) at 350\u00a0\u00b0C and then cooled to room temperature. Next, an NH3 (5%)/He flow (15\u00a0mL\u00b7min\u22121) was passed through the sample for 30\u00a0min at 100\u00a0\u00b0C. In the next step, at 100\u00a0\u00b0C, the sample was swept with a He flow (25\u00a0mL\u00b7min\u22121) for 30\u00a0min to remove the physically adsorbed NH3. Afterward, NH3-TPD was performed in a He flow (25\u00a0mL\u00b7min-1) at a heating rate of 15\u00a0\u00b0C\u00b7min\u22121 from 100 to 700\u00a0\u00b0C. The desorbed NH3 was detected by a TCD.Surface acidity (Br\u00f8nsted and Lewis acid sites) was characterized by in situ FTIR spectroscopy with chemisorbed pyridine in the diffuse reflectance infrared Fourier transform (DRIFT) mode. The spectra were collected with a Nicolet 5700 spectrometer (Waltham, USA) equipped with a Hg\u2013Cd\u2013Te cryodetector of high sensitivity, working in the spectral range of 4000\u2013650\u00a0cm\u22121. A Praying Mantis apparatus (Harrick Scientific Co, Pleasantville, United States) was used as the optical mirror accessory. The samples were placed in a reaction camera equipped with a temperature controller that allows in situ thermal treatments (Harrick Scientific Products, NY). Pyridine (Py) was used as a probe molecule to evaluate surface acidity. The samples were heated from room temperature to 350\u00a0\u00b0C at a heating rate of 10\u00a0\u00b0C \u00b7min\u22121 under a flow of H2 (10 mLN\u00b7min\u22121) and Ar (50 mLN\u00b7min\u22121). The samples were kept at these temperatures for 1\u00a0h to reduce them and clean the surface, and then an infrared spectrum of the solid was recorded. After that, the temperature was cooled to 120\u00a0\u00b0C, the H2 flow was turned off, and Ar was bubbled through the liquid Py for a sufficient time to saturate the sample (10\u00a0min). Subsequently, Ar flow was switched on (bypassing the bubbler) at 120\u00a0\u00b0C for 1\u00a0h to remove the physically adsorbed Py. The DRIFT spectra of chemisorbed molecules over the surface sites were then recorded. In all cases, the spectra were recorded with 128 scan accumulations and a resolution of 4\u00a0cm\u22121. The net infrared spectra were obtained after subtraction of the background spectrum of the solid.XPS measurements were recorded using a SPECS GmbH electron spectroscopy system (Berlin, Germany) with a UHV system (pressure approx. 10-10 mbar) with a PHOIBOS 150 9MCD energy analyzer, a multichannel detector (9 channels), and a monochromatic X-ray source (with double anode Mg). The area of the peaks was estimated by calculating the integral of each peak after smoothing and subtracting the S-shaped background and fitting the experimental curve to a mixture of Gaussian and Lorentzian lines of variable proportions. Binding energies (BEs) were referenced to the Zr 3d signal at 182.2\u00a0eV for the PtW/Zr-3\u201318 catalysts and to the Al 2p signal at 74.5\u00a0eV for the PtW/Al-3\u201318 catalysts [6] to correct for charging effects. Quantification of the atomic fractions on the sample surface was obtained by integration of the peaks with appropriate corrections for sensitivity factors [10].The desired metal loadings were confirmed by elemental analysis via inductively coupled plasma optical emission spectrophotometry (ICP\u2013OES) and X-ray Fluorescence (TXRF). Qualitative and quantitative TXRF analyses were performed with a benchtop S2 PicoFox TXRF spectrometer from Bruker (Billerica, USA) equipped with a Mo X-ray source working at 50\u00a0kV and 600 \u03bcA, a multilayer monochromator with 80% reflectivity at 17.5\u00a0keV (Mo K\u03b1), a Bruker XFlash SDD detector (Billerica, USA) with an effective area of 30\u00a0mm2 and an energy resolution better than 150\u00a0eV for 5.9\u00a0keV (Mn K\u03b1), and the methodology for the experimental procedure, called DSA-TXRF (Direct Solid Analysis), can be found in the following paper [11] (p.79) and was developed in the SIDI (Servicio Interdepartamental de Investigaci\u00f3n) of the UAM (Universidad Aut\u00f3noma de Madrid).ICP\u2013OES was performed as follows. First, digestion of the solids was carried out using a Multiwave 3000 model Anton Paar equipment (Graz, Austria), a high-pressure microwave. This digestion consisted of wet mineralization using an acid solution. Decomposition was carried out in closed Teflon containers to avoid losing the volatile components. In all cases, 20\u00a0mg of the samples were introduced to the containers containing a mixture of acids, 6HNO3:3HF (ml). The microwave operated at 800\u00a0W and 60\u00a0bar. Once the sample was dissolved, it was measured by argon nebulization using the inductively coupled plasma optical emission spectrophotometer PQ9000 of Analytik Jena (Jena, Germany).The catalysts were tested for the hydroisomerization of n-dodecane. The reactor operated in a trickle-bed mode in parallel flow and at high pressure, ensuring that the three phases, gas\u2013liquid-solid, were in close contact. The calcined catalysts in the pellets (1\u00a0g) were diluted in 8\u00a0g of the inert alumina or zirconia supports, placed in the reactor, and reduced at 350\u00a0\u00b0C and atmospheric pressure. Then, the reduction temperature (350\u00a0\u00b0C) was maintained, and the reactor was pressurized. The reaction conditions were: Tr\u00a0=\u00a0350\u00a0\u00b0C, P\u00a0=\u00a02.0\u00a0MPa (20\u00a0bar), liquid flow\u00a0=\u00a00.1\u00a0mL\u00b7min\u22121 and H2 flow\u00a0=\u00a0340 mLN\u00b7min\u22121. The gaseous-phase products were analyzed by an online Inficon 3000 micro-GC (Bad Ragaz, Switzerland) equipped with 4 channels, two 5A molecular sieves, a Poraplot Q, and a Stawilwax. The liquid products were collected and analyzed offline by gas chromatography with an Agilent (Palo Alto, USA) 6850A GC and an FID detector.Under the reaction conditions employed and the trickle-bed operation regimen, the external transport limitations are excluded. In consequence the inter-particle heat and mass transport limitations are minimized. The reaction conditions employed in the experiments reach the laboratory reactor compliance criteria for a trickle-bed reactor described in the bibliography [12]. The reactor is filled with glass beads on the catalysts bed assure a good liquid dispersion in the catalyst bed. L/dp is 200 clearly higher than the required criterium of 100. D/dp is 25 higher than the required criterium of 10. Based on these figures, the reaction conditions used clearly reach the compliance of a laboratory scale trickle-bed reactor.The textural properties of the supports and catalysts were determined by N2 adsorption\u2013desorption. The BET surface area, mean pore volume and mean pore diameter of the alumina and zirconia supports and catalysts are reported in Table 1\n. In general, when the active phase is impregnated onto the support, the surface areas decrease due to the incorporation of tungsten oxide, which partially covers the support surface, partly blocking the support pores and reducing nitrogen accessibility [13]. This decrease is more evident in the alumina catalysts. Notably, the alumina catalysts show a significantly higher area (196\u2013223\u00a0m2/g) than the zirconia catalysts (73\u201390\u00a0m2/g). Alumina support has a high surface area which makes that as the same % W loading is incorporated in the two different procedures, the support surface coverage is similar and BET areas do not significantly differ.To characterize the porosity of these samples, N2 adsorption\u2013desorption isotherms of the supports and catalysts were obtained. All the isotherms belong to type IV (a) according to the IUPAC 2015 classification or type IV according to the BDDT (Brunauer, Deming, Deming, and Teller) classification, given by mesoporous materials [14,15]. According to the IUPAC system, the hysteresis loops obtained in the isotherms (Fig. 1S) can be classified as the H3 type for PtW/Zr and PtW/Al catalysts. In addition, pore types can be deduced from the isotherm hysteresis loop. The zirconia and alumina isotherm hysteresis loops are close to slit-shaped pores (H3 hysteresis type) [15\u201317]. Comparing the support and catalyst isotherms in all cases, the isotherms are visibly very similar, which indicates a high dispersion of the tungsten and platinum active phases on the supports. All the supports and catalysts show pore size distributions (Fig. 2S) in the mesoporous range from 2 to 50\u00a0nm.The crystalline phases present in the samples were studied by X-ray diffraction. In general, the X-ray diffractograms of the catalysts do not show different lines compared with the supports. As an example, X-ray patterns obtained for PtW/Zr-15, PtW/Zr-15-SI, PtW/Al-15, PtW/Al-15-SI, and their supports are shown in Fig. 1\n.For the PtW/Zr-15 catalyst (PDF card 00\u2013003-0515), the most intense diffraction line is located at 28.1, associated with the (111) plane of monoclinic ZrO2, and for the PtW/Zr-15-SI (PDF card 00\u2013013-0307) catalyst, the most intense diffraction line is located at 28.3\u00b0, associated with the (\u2212111) plane of monoclinic ZrO2 (Fig. 1A). The diffraction lines at 2\u03b8\u00a0=\u00a017.5, 24.0, 28.1, 34.1, 34.3, 35.0, 38.6, 40.6, 44.8, 49.2, 50.4, 54.2, 55.4, 60.1, 61.5, 62.7, 65.3, 71.2, and 75.2\u00b0 were attributed to the monoclinic phase of the ZrO2 support. Some minor peaks associated with the tetragonal phase of ZrO2 (PDF card 00-024-1164) are also present. Most WO3 peaks are overlapped with monoclinic ZrO2 peaks suggesting either a homogeneous dispersion of WO3 on the support surface [18] or very small crystallite sizes for the WO3\n[15]. The most intense diffraction line for WO3 is the one at 28.1\u00b0 for PtW/Zr-15 and PtW/Zr-15\u2013SI corresponding to the (200) plane of WO3, but they overlap with the diffraction lines of the support. The absence of platinum diffraction peaks also revealed the high dispersion of this metal on the catalyst, which was expected due to the low Pt loading.For the PtW/Al-15 catalyst (Fig. 1B), the XRD profiles present three main peaks at 36.9, 47.6, and 67.4\u00b0, corresponding to the (111), (006), and (215) planes of \u03b3-Al2O3, respectively (PDF card 00-035-0121). For the alumina support and the PtW/Al-15-SI catalyst (Fig. 1B), the XRD spectra present three main peaks at 37.1, 46.0, and 66.8\u00b0, corresponding to the (110), (111) and (211) planes of \u03b3-Al2O3, respectively (PDF card 00-001-1303). In both catalysts, few obvious diffraction peaks were found for WO3 (PDF card 00-033-1387) or W3O8 (PDF card 01-081-2262), indicating a high dispersion of tungsten oxide in the support. The lower intensity of the diffraction lines for the catalysts is explained by the lower quantity of sample placed in the sample holder.The Scherrer equation was applied to the most intense and nonoverlapping diffraction lines to determine the average crystalline domains. The average crystalline domain sizes of the corresponding zirconia crystallites are similar for each support and catalyst, showing sizes of approximately 10\u00a0nm. In the case of alumina and alumina-supported catalysts, these properties have not been calculated since the alumina phase (gamma) is poorly crystallized (pseudocrystalline state) and pseudocrystals form its structure.The different tungsten oxide structures were studied by Raman spectroscopy (Fig. 2\n). Fig. 2A shows the Raman spectra of the ZrO2 support and the PtW/Zr-3 to 18 catalyst series with profiles corresponding to the monoclinic zirconia phase. In the PtW/Zr catalysts, these bands appear at approximately 180 (with two components), 330, 380, 476, 547 (with two components more visible for the PtW/Zr-18 catalyst) and 635\u00a0cm\u22121\n[19,20].All catalysts except PtW/Zr-3 show bands at \u223c800 and \u223c700\u00a0cm\u22121, which are characteristic of the stretching and bending vibrations of W-O-W, respectively. These bands are also characteristic of crystalline WO3\n[13], but no evidence for WO3 crystals is shown in the XRD diagrams. The band at approximately 270\u00a0cm\u22121 is assigned to the W-O-W deformation mode [19]. In addition, a broad band with two components at 955 and 995\u00a0cm\u22121 is attributed to the asymmetric and symmetric vibrations of W\u00a0=\u00a0O, respectively, corresponding to highly dispersed WOx species [18], in agreement with XRD results. It can be deduced from the spectrum that a 3% W loading is not sufficient for the formation of bands due to W oxide, but these bands are formed at 6\u00a0wt% W loading and above.\nFig. 2B shows the Raman spectra of the ZrO2 support and the PtW/Zr-6-SI to 15-SI catalysts with a profile corresponding to the monoclinic zirconia phase. In the PtW/Zr-SI catalysts, bands appear at approximately 180 (with two components), 335, 382, 476, and 547\u00a0cm\u22121. The band at 630\u00a0cm\u22121 of the monoclinic zirconia phase is not clearly visible in the PtW/Zr-12-SI and PtW/Zr-15-SI catalysts.PtW/Zr-6-SI only shows a weak and broad band at approximately 1000\u00a0cm\u22121, assigned to symmetric W\u00a0=\u00a0O stretching, which is characteristic of tetrahedrally coordinated surface tungsten oxide species [20]. In the PtW/Zr-9-SI catalyst, the band at approximately 1000\u00a0cm\u22121 is more intense, and a broad band at 800\u00a0cm\u22121 due to the stretching vibration of W-O-W appears. For PtW/Zr-12-SI and PtW/Zr-15-SI, W-O-W bands appear at higher W loadings. These catalysts show bands at approximately 800 and 700\u00a0cm\u22121 due to the stretching and bending vibrations of W-O-W, respectively, at approximately 260\u00a0cm\u22121 due to the W-O-W deformation mode, and a band at approximately 1000\u00a0cm\u22121 due to the symmetric vibrations of W\u00a0=\u00a0O [20,21]. These bands are more intense in the PtW/Zr-15-SI catalyst. This means that below monolayer coverage, such as in PtW/Al-6-SI, only isolated tungsten oxide species exist on ZrO2 due to the unique interactions between WO3 and ZrO2. In PtW/Al-9\u201315-SI, the W\u00a0=\u00a0O vibrations are due to polytungstate, meaning two-dimensional oxotungsten species interact with the support [20,21].\nFig. 2C shows the Raman spectra of the \u03b3-Al2O3 support and the PtW/Al-3 to 18 catalysts. The known \u03b3-Al2O3 support has no Raman active modes [21], and the PtW/Al-3 to 18 catalysts only exhibit one broad band at approximately 1000\u00a0cm\u22121 (varying only from 1001 to 1005\u00a0cm\u22121) [22,23], indicating the pseudoamorphous alumina state, in agreement with the XRD profiles obtained for this support and derived catalysts. The band at 1000\u00a0cm\u22121 can be attributed to the W-O stretching mode of the Al2(WO4)3 phase, but it is not very well defined. Considering monolayer coverages of tungsten oxide on alumina-supported catalysts with less than 25\u201330\u00a0wt% WO3, it suggests that this oxide is in a highly dispersed and amorphous state on the alumina surface at low calcination temperatures (500\u2013800\u00a0\u00b0C) [23]. Fig. 2C reveals that very similar Raman spectra are obtained by varying the tungsten oxide loading in single-step prepared catalysts.\nFig. 2D shows the Raman spectra of the \u03b3-Al2O3 support and the PtW/Al-6-SI to 15-SI catalysts. The Raman band at approximately 1000\u00a0cm\u22121, assigned to the symmetric stretching vibration mode of W\u00a0=\u00a0O bonding, tends to increase in intensity with W loading. This weak and broad band centered at 989\u00a0cm\u22121 in PtW/Al-6-SI is due to the stretching mode of mono-oxo W\u00a0=\u00a0O, corresponding to the highly dispersed WOx species. This band is shifted gradually toward higher wavenumbers, from 989 to 1010\u00a0cm\u22121, in PtW/Al-SI. This shift might be explained by the interaction between platinum and surface tungsten oxides, which affects the distortion of the oxo-tungsten species and transfers electrons from metallic Pt to WOx at the Pt-WOx interface, making the W\u00a0=\u00a0O bond stronger [18,21]. The shift in W\u00a0=\u00a0O also suggests that, under ambient conditions, different two-dimensional tungsten oxide species may be present in the PtW/Al-SI catalysts: tetrahedrally coordinated tungsten oxide species at low loading (6\u00a0wt%) and octahedrally coordinated tungsten oxide species at moderate and high loadings (12 and 15\u00a0wt%) [20,21].The dispersion of supported Pt particles in the catalysts with 15% W loading, determined by CO pulse chemisorption, is reported in Table 2\n. Catalysts were previously reduced at 250\u00a0\u00b0C to avoid the partial reduction of tungsten oxide that could result at a temperature of 300\u00a0\u00b0C and higher. In the zirconia catalysts, WO3 is reduced at approximately 400\u00a0\u00b0C, and the dispersion could not be measured at the same reduction temperature used in the reaction (350\u00a0\u00b0C) because the Pt dispersion values would be overestimated since the CO probe molecule can also adsorb to reduced WOx\n[24,25].The obtained results indicate that the highest dispersions are achieved in zirconia-based catalysts. Zirconia catalysts show a higher dispersion than alumina catalysts because zirconia due to its negative charge on the surface favours tetraammineplatinum (II) cation (Pt(NH4)4\n2+) dispersion whereas alumina due to its positive charge on the surface does not favour (Pt(NH4)4\n2+) cation dispersion. Comparing catalysts prepared using a single step method to catalysts prepared using several consecutive steps the platinum dispersion is higher in the catalysts prepared using a single step. The dispersion measurements follow the order PtW/Zr-15\u00a0>\u00a0PtW/Zr-15-SI\u00a0>\u00a0PtW/Al-15\u00a0>\u00a0PtW/Al-15-SI.The acid properties of the catalysts were investigated by NH3-TPD. This technique measures the total acidity and the strength of acid sites present on the catalyst surface [26]. The NH3 desorption profiles for the four series of catalysts are depicted in Fig. 3\n. Fig. 3 shows the NH3-TPD profiles for all the catalysts, with three different desorption peaks that correspond to the following three acid strengths: low acidity sites (desorption peaks\u00a0less than\u00a0250\u00a0\u00b0C), medium acidity sites (desorption peaks between 250 and 400\u00a0\u00b0C), and strongly acidic sites (desorption peaks\u00a0>\u00a0400\u00a0\u00b0C) [27,28]. All the desorption profiles of the catalysts show three main peaks, in agreement with the previous description. For both alumina catalysts, the strong acidic site peaks follow the order PtW/Al-3\u00a0>\u00a0PtW/Al-6\u00a0>\u00a0PtW/Al-9\u00a0>\u00a0PtW/Al-12\u00a0>\u00a0PtW/Al-15\u00a0>\u00a0PtW/Al-18, which is the reverse of the W loading trend. This is due to the formation of tungsten aluminate, which neutralizes the strongly acidic sites. In other words, more W loading implies less strong acid sites in alumina catalysts because they are neutralized by the tungsten aluminate formed. However, in general, the weak acidic site peaks grow in accordance with the W loading: PtW/Al-3\u00a0<\u00a0PtW/Al-6\u00a0<\u00a0PtW/Al-9\u00a0<\u00a0PtW/Al-12\u00a0<\u00a0PtW/Al-15\u00a0<\u00a0PtW/Al-18. These results indicate that with increasing levels of WO3, weak acidic sites increase and strong acidic sites decrease [28].The proportion of strongly acidic sites on the alumina-based catalysts is higher than the same sites on the zirconia-based catalysts due to the higher surface area of the alumina support (Table 1) and the higher dispersion of WOx in these catalysts observed by Raman spectra. As reported in Table 3\n, the total acidity value was measured. The PtW/Zr catalysts show more total acidity than the PtW/Zr-SI catalysts. For the alumina catalysts, the total acidity does not show major differences between the PtW/Al and PtW/Al-SI catalysts. Despite this, the proportion of low strength acid sites in alumina catalysts is higher in the catalyst prepared by sequential impregnation while the proportion of high strength acid sites is higher in the catalyst prepared by the one-step method. This fact is explained by the greater formation of tungsten aluminate in the catalyst prepared by sequential impregnation, since each calcination carried out after each impregnation favours the formation of this aluminate phase, decreasing the proportion of strong acid sites.The adsorption of pyridine as a base on the surface of solid acids is used to evaluate the nature of surface acidity. The use of IR spectroscopy of adsorbed pyridine facilitates the identification of distinct acidic sites [26]. The pyridine FTIR (DRIFT) absorption spectra (Fig. 4\n) of the PtW/Zr-15, PtW/Zr-15-SI (Fig. 4A), PtW/Al-15, and PtW/Al-15-SI (Fig. 4B) catalysts were measured at room temperature and were previously reduced at 350\u00a0\u00b0C. The pyridine FTIR (DRIFT) absorption spectra of the PtW/Zr-3 to 18 catalysts, ZrO2 support, PtW/Al-3 to 18 catalysts, and Al2O3 support are shown in the supporting information (Fig. 3S).The FTIR spectra show the presence of adsorption bands centered at 1607\u20131616, 1575, and 1447\u20131450\u00a0cm\u22121, which are assigned to the vibrational modes of pyridine molecules adsorbed on Lewis (L) acid sites. A broad band at 1540\u00a0cm\u22121 is assigned to the pyridinium ions absorbed on Br\u00f8nsted acid (B) sites. The absorption band at 1488\u20131490\u00a0cm\u22121 is a combined band that originated from both Lewis and Br\u00f8nsted acid sites [18,26,29,30]. The presence of Lewis acid sites in the zirconia-based catalysts can be attributed to the presence of coordinately unsaturated Zr4+ cations, while the Br\u00f8nsted acid sites are likely hydroxyl groups (W\u2013O\u2013W\u2013OH or Zr\u2013O\u2013W\u2013OH) associated with W6+ and W5+ atoms [29]. In alumina, the Lewis acid sites correspond to three possible Al3+ coordination configurations: five-, four-, and three-fold coordinated [31].Surface chemical analyses of the catalysts were carried out by XPS. The XPS spectra for the PtW/Zr-3\u201318 and PtW/Al-3\u201318 catalysts are shown in Fig. 5\n, and the binding energies (eV) (Al 2p, Zr 3d, and W 4f7/2 core levels) and surface atomic ratios are listed in Table 4\n.The evolution of the W 4f and Zr 4p orbitals in the PtW/Zr-3 to 18 (Fig. 5A) catalysts and the W 4f orbitals in the PtW/Al-3 to 18 catalyst (Fig. 5B) peaks is shown in Fig. 5 as a function of W loading. The W 4f signal presents a typical doublet corresponding to spin-orbital splitting. The PtW/Zr-3 to 18 catalysts show a single component with the BE for W4f7/2 at approximately 35.5\u00a0eV characteristic of W(VI) attributed to WO3\n[32] (nanocrystalline) species. The PtW/Al-3 to 18 catalysts present two components, one attributed to the WO3 species at approximately 35\u00a0eV [6,13] and a second attributed to aluminum tungstate, Al2(WO4)3, at approximately 36\u00a0eV [6,33]. WO3 was also identified in the Raman bands observed in the spectra (Fig. 2C).The surface atomic ratio W/(Al or Zr) tends to increase with W loading, as observed in Table 4 and Fig. 5C and D. This increase is linear in agreement with the theoretical density of isolated tungsten oxide species and polytungstated oxide species [32] identified by Raman spectroscopy. The PtW/Al-3 to 18 catalysts show lower values for W/Al, which can be explained by the presence of aluminum tungstate species. Due to the formation of this kind of species, surface tungsten migrates to the interior of the alumina, reducing the W signal observed by XPS [6].In summary, with the characterization techniques used (N2 adsorption\u2013desorption isotherms at \u2212196\u00a0\u00b0C, X-ray diffraction, Raman spectroscopy, FTIR, NH3-TPD, CO pulse chemisorption, and XPS), we can conclude that the alumina catalysts have a higher surface acidity than the zirconia catalysts due to the higher surface area of the alumina support. Catalysts prepared by successive impregnations show more isolated WO3 in the Raman spectra than the single-step prepared catalysts which can produce lower acidity in the SI catalysts. The absence of WO3 diffraction peaks in the XRD analysis revealed the homogeneous dispersion of this oxide on the catalytic supports.The prepared catalysts were studied for the hydroisomerization of n-dodecane. The activation or reduction process (350\u00a0\u00b0C in hydrogen at atmospheric pressure) and reaction conditions (Tr\u00a0=\u00a0350\u00a0\u00b0C, P\u00a0=\u00a02.0\u00a0MPa, liquid flow\u00a0=\u00a00.1\u00a0mL\u00b7min\u22121 and H2 flow\u00a0=\u00a0340 mLN\u00b7min\u22121) in a trickle bed-mode reactor were selected based on previous studies [6].Catalysts showed stability in reactions of 24\u00a0h. We checked the possible formation of coke analyzing the sample PtW/Al-15 after reaction, obtaining a C/Al atomic ratio similar to the ratio obtained for PtW/Al-15 before reaction, which means that a minimal quantity of coke was formed. In general, all carbon-balance is higher than 90%.\nFig. 6\n shows the conversion and selectivity results for all the catalysts studied. Catalysts with a low loading of W showed low conversion, especially on the zirconia catalysts, and this effect was also observed for other similar reactions [9]. The conversion of n-dodecane tends to increase with W loading, which is more obvious in the single-step prepared catalysts. Similar studies of varied W loading have been previously reported [9,34]. The relationship between the tungsten loading and the hydrogen uptake capacity of the tungstated zirconia catalysts has been studied by Iglesia et al. They concluded that hydrogen uptake increases with tungsten loading due to the stabilization of hydrogen by polytungstate and crystalline WO3 species on the surface ZrO2\n[34]. Conversions reach a maximum value at a WO3 concentration of 15\u00a0wt% W, this optimum W percentage occurs because the PtW/Zr-18 and PtW/Al-18 catalysts prepared by a single step showed lower conversions than the catalysts with PtW/Zr-15 and PtW/Al-15. This result is similar to that reported by Iglesia et al. for o-xylene isomerization rates on WOx\u2013ZrO2 samples with various WOx concentrations, reaching a maximum value at a WOx concentration of 12\u00a0wt% W [9]. Similarly, Triwahyono et al. reported that the isomerization of n-butane reached a maximum using WOx\u2013ZrO2 with 13\u00a0wt% W [34]. The maximum reached for the PtW/Zr-15 and PtW/Al-15 catalysts may have been due to their high concentration of protonic acid sites or their hydrogen uptake capacity [34]. The reason why the PtW/Zr-18 and PtW/Al-18 catalysts are less active than the PtW/Zr-15 and PtW/Al-15 catalysts is probably related to the larger size of the WOx particles formed, presumably because the loading of W, 18\u00a0wt%, is too high to adequately disperse over the support and the WOx particles agglomerate. Thus, in the presence of H2, the activities of PtW/Zr, PtW/Zr-SI, PtW/Al, and PtW/Al-SI catalysts for n-dodecane isomerization were strongly determined by the WO3 loading.Comparing both supports, it is clear that the alumina-supported catalysts are more active than the zirconia-supported catalysts. The higher catalytic activity of the alumina-supported catalysts can be related to their higher surface area and acidity, especially in terms of the strongly acidic sites shown by TPD-NH3, which occur at a higher concentration than those on the zirconia-supported catalysts [13]. The FTIR spectra also show that there are more acidic sites in the alumina catalysts than in the zirconia catalysts. In addition, a larger ratio of Lewis acid sites in alumina catalysts is shown, and this ratio is also larger in the most active catalysts prepared by a single step compared with the SI catalysts. The presence of strong Lewis acid sites (monolayer dispersed WO3 particles) facilitates the formation of protonic sites from H2 in the gaseous phase [34], which are the active sites for n-dodecane hydroisomerization. In addition, we checked that the pH of the ((NH4)6(H2W12O40)\u00b7xH2O) aqueous solution was between 6 and 7, the isoelectric point (IEP) of Al2O3 is approximately 8 and the IEP of ZrO2 is approximately 5 [35]. This creates a net positive charge on the alumina surface and a net negative charge on the zirconia surface and favors W incorporation on the alumina surface because W is in the anion of the precursor and has a negative charge. However, the incorporation of W to the alumina supports yields a partial formation of tungsten aluminate, which usually migrates to deeper layers resulting in the lower surface atomic ratios (W/Al) in the PtW/Al-3-18 catalysts determined by XPS (Table 4), this lower surface atomic ratio seems to have a positive effect on their catalytic activity because these catalysts showed higher conversions.By comparing the two preparation methods, we can conclude that the single-step impregnation method works better than several successive impregnations for adding the same wt.% of W due to the higher dispersion of WO3 on the catalyst surface prepared using the single-step process. This produces more accessible acidic sites, improving the progress of the reaction and the conversion results. This fact is explained by the greater loss of surface WO3 in the catalysts prepared by SI, since each calcination, after the successive impregnations, entails a solid-state reaction between this oxide and the support. Another reason is that, as observed in the Raman spectra, more isolated WO3 clusters are produced in the successively impregnated catalysts than in the single-step-prepared catalysts, which results in a lower acidity in the SI catalysts and less activity.Selectivity for isomers (Fig. 7\n) varies considerably with W loading, especially when comparing low loadings with high loadings. In general, branched C12 hydrocarbons are the main products obtained, showing selectivity for C12 (>80%) for catalysts with high W loadings (>9\u00a0wt% W). The selectivity to C11 varies from 7.4 to 13.9% in these catalysts, and low selectivity values for C6-10 were obtained. Catalysts with 3 and 6\u00a0wt% W (independent of the preparation method for 6\u00a0wt%) showed high selectivity for C6-10 hydrocarbons and low selectivity for C12 hydrocarbons, but these catalysts showed a very low catalytic activity for the hydroisomerization of n-dodecane.A comparison of our results with other systems used in the bibliography is compiled in Table 5\n. The results are selected because a similar catalyst is used or because a similar or the same isomerization reaction is studied. The catalytic activity and isomer yield obtained in this work are better than the values previously, especially for the Pt/WOx/Al2O3 series with certain W loading.Four series of catalysts based on Pt-WOx-alumina and Pt-WOx-zirconia have been successfully prepared by a wet impregnation method via a single step or several consecutive steps (sequential impregnation, SI).The characterization of the catalysts indicates that the alumina catalysts achieve higher surface acidity than the zirconia catalysts, which is explained by the higher dispersion of tungsten oxide on alumina due to the larger surface area of alumina and the IEPs of alumina and zirconia, which influence the dispersion of WOx. The single-step preparation method for the catalysts is a better preparation method because fewer isolated WOx species are formed, as it was revealed in the Raman spectra, which improves the activity results. The high dispersion of WO3 on the catalysts agrees across the XRD, Raman, and nitrogen adsorption/desorption isotherm results.In general, the conversion of n-C12 grows with %W loading for both supports and methods, and branched C12 hydrocarbons are the main products obtained (>80%).We can conclude that the PtW/Zr-15 and PtW/Al-15 catalysts are found to be the most active for the hydroisomerization of n-dodecane, meaning that the catalysts prepared via a single step are more active than catalysts prepared via several consecutive steps. The catalytic activity results clearly show that 15\u00a0wt% W is the best W loading and alumina is the best support (67% of n-C12 conversion and 87% of i-C12 selectivity). The different isomerization activities shown by the alumina and zirconia-based catalysts can be related to differences in their physicochemical properties previously mentioned.\nD. Garc\u00eda-P\u00e9rez: Investigation, Formal analysis, Writing \u2013 original draft. G. Blanco-Brieva: Investigation, Formal analysis, Writing \u2013 original draft. M.C. Alvarez-Galvan: Supervision, Writing \u2013 review & editing, Funding acquisition. J.M. Campos-Martin: Conceptualization, Supervision, Writing \u2013 review & editing, Project administration, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.The Agencia Estatal de Investigaci\u00f3n (AEI) (Spain) supported this work through the project ENE2016-74889-C4-3-R. DGP acknowledges her contract (BES-2017-079679) to AEI (Spain). We acknowledge the support of the publication fee by the CSIC Open Access Publication Support Initiative through its Unit of Information Resources for Research (URICI).Supplementary data to this article can be found online at https://doi.org/10.1016/j.fuel.2022.123704.The following are the Supplementary data to this article:\n\nSupplementary Data 1\n\n\n\n", "descript": "\n                  Catalysts based on zirconia and alumina-supported tungsten oxides (3\u201318\u00a0wt% W) with a small loading of platinum (0.3\u00a0wt% Pt) were selected to study the influence of W loading, support type, and preparation methods on the hydroisomerization of n-dodecane. The alumina- and zirconia-supported catalysts show different catalytic properties that play an important role in n-dodecane conversion. Specific surface area and acidity of the alumina-supported catalysts are clearly higher than those of zirconia-supported counterparts. The 15\u00a0wt% W loaded catalysts were found to exhibit the best catalytic performance for the hydroisomerization of n-dodecane for both zirconia and alumina (67% of n-C12 conversion and 87% of i-C12 selectivity) based catalysts. Catalysts prepared by a single-step wet impregnation method were more active in the hydroisomerization of n-dodecane than catalysts prepared via several consecutive steps.\n               "}
{"full_text": "Fluid catalytic cracking (FCC) is an important process for the conversion of crude oil into valuable products including fuels, lubricants, and precursors for making other products. This importance is evidenced by the fact that there are more than 430 FCC units worldwide today [1\u20133]. First used commercially in 1942 in Baton Rouge, Louisiana (USA), the FCC process cracks high molecular weight hydrocarbon chains into lighter hydrocarbons using high temperatures (525\u2212575 \u00b0C) and a heterogeneous catalyst [1,4\u20137]. Until the development of the FCC process, refineries were inefficient at making valuable products such as gasoline and liquefied petroleum gas (LPG); however, the parallel development of the FCC process and a catalyst capable of fluidization and cracking chemistry enabled refineries to upgrade less valuable fractions of crude oil into high-value diesel, gasoline, and LPG products. The FCC catalyst is primarily composed of zeolite-Y (in the form of ultra-stable zeolite Y or USY), which has a high surface area, but also features a matrix used for cracking reactions. The catalytic system can also include additional features such as nickel and/or vanadium passivation technologies and additives to tune product yields or to control emissions.The catalyst facilitates beta scission reactions and is relatively robust \u2013 a necessity for enduring high temperatures and physical stress during operation. Additionally, FCC catalysts are often exposed to metal contaminants, which are typically introduced into the unit via the FCC feed. A common feed contaminant is nickel, which is often introduced with the feed as a nickel (II) porphyrin structure [8,9]. The concentration of nickel in feed varies widely and can be as high as 100 ppm in extreme cases, although values lower than 25 ppm are more typical [10]. A well-known dehydrogenation catalyst, nickel deposits on FCC catalyst in concentrations ranging up to 19,000 ppm. The deposited nickel induces unwanted dehydrogenation reactions, which lead to an increase in hydrogen and coke yields [11\u201314]. Excessive amounts of both hydrogen and coke can be problematic for refiners as they push the FCC unit closer to operating limits. For example, increased hydrogen can constrain the FCC unit's downstream compressor, and increased coke can increase the regenerator temperature towards its maximum limit. However, the nickel contaminant becomes less active as it spends more time in the FCC unit. The oxidative environment in the FCC regenerator oxidizes nickel to nickel oxides. This chemical transformation immobilizes the nickel and greatly reduces its dehydrogenation tendency. Interaction of nickel with alumina phases in the catalyst (such as the low surface area crystalline aluminas used for trapping of contaminant nickel in catalysts designed for processing heavier, residue-containing feedstocks) result in various forms of nickel aluminate [15\u201319]. Once nickel oxides and aluminates are formed, it is important to keep nickel in those states and inhibit its reduction to metallic nickel in the FCC environment.There is precedence in literature that chloride ions can both mobilize and reactivate nickel oxides. Earlier work shows that NiO on activated carbon reacts with hydrochloric acid (HCl) to form NiCl2, a mobile compound, which can then be further reduced by H2 to metallic nickel, a more active dehydrogenation catalyst than NiO [16]. Another study shows the same phenomena using platinum, a metal from the same family as nickel, on zeolite that is exposed to HCl and subsequently reduced with H2. Additionally, a further study observed a redistribution of platinum on the support following the HCl and H2 reactions [20,21]. Another contribution showed that deactivated Ni-erionite catalyst regained its dehydrogenation activity when treated with solutions of HCl or NH4Cl [22]. These examples demonstrate that relatively inert NiO can be reactivated for dehydrogenation chemistry and mobilized by exposure to chloride-containing compounds and set a precedence that this might be possible in the FCC environment. Indeed, industrial reports have noted a correlation between increased chloride content and an increase in unwanted hydrogen production, among other issues. While little literature exists concerning chloride interactions with nickel aluminates, nickel in nickel aluminate exists in a +2 oxidation state (as it does in NiO) and has been shown to be difficult to reduce and less active for dehydrogenation chemistry than metallic nickel [18,19]. Thus, it is hypothesized that chloride could lead to a similar reactivation of nickel aluminate for dehydrogenation chemistry.Interactions of nickel with chloride ions are relevant to the FCC environment as chloride ion sources can enter the FCC both with feed, sometimes a result of insufficient desalting operations, and with fresh catalyst as part of an alumina-based binder used in incorporated catalysts or from the use of chloride-containing precursors/chemicals in catalyst manufacturing [14,23]. The use of the alumina-based binders for incorporated catalysts is needed for particle integrity in order to control the attrition of the final product. Alumina-based binders often contain chloride as a byproduct in its manufacturing. Chloride content in fresh FCC catalyst can be as high as 1.2 wt.%. Chloride sources coming from the feed vary widely. In heavily contaminated feeds, chloride can be as high as 15 ppm. Chloride sources in the feed can react with steam in the FCC to form HCl, while most of the binder-based chloride is released and converted to HCl in the high-temperature, steam partial pressure environment of the FCC regenerator [24\u201326]. While these chloride contaminants are well known to lead to deposits in the downstream fractionator, fouling in equipment, and having a negative impact on metallurgy, their effect on nickel contaminants in an FCC has never been formally investigated [27].The work described herein constitutes the first exploration of the effect of chloride ions on nickel contaminants deposited on actual FCC catalysts using simulated FCC conditions. While circulating in an FCC unit, a fraction of catalyst is continuously added and withdrawn. The continual addition and withdrawal of catalyst introduces an age-distribution of catalyst particles in periodically withdrawn samples that are tested and tracked to monitor performance. This age-distributed catalyst sample is commonly called equilibrium catalyst (Ecat). Ecat samples taken from two different industrial FCC units were selected for this study. These samples were selected due to their differing nickel levels, marked as \"high\" and \"low\". It is important to note that the Ecat samples chosen for this study originate from catalysts manufactured by the \u201cin-situ\u201d manufacturing route. This route differs from conventional catalyst production process in that zeolite is grown in the microsphere after the spray drying step. The zeolite itself acts as the catalyst binder, thus in-situ catalysts do not use chloride-containing binders. As a result, there are no chlorides present in fresh catalyst. In addition, the refineries from which these Ecat samples originate did not report any chlorides coming from the feed. Therefore, this study represents the first time these samples are introduced to chlorides. Table 1\n shows the total surface area (TSA), zeolite surface area (ZSA), matrix surface area (MSA) and average particle size (APS) of the Ecat samples studied. It is noted that there are slight differences in surface areas and rare earth oxide content between the low and high nickel containing Ecat samples. For the purpose of this study, we note that these would not have a significant effect on the expected outcome based on the experimental design. Because the Ecat samples that fit the desired criteria (similar technology, similar manufacturing route, no additive usage, no previous chloride exposure) are limited and are based on refineries operating around the world at the moment of this experimental design, these samples represent the best compromise between using industrial Ecats and laboratory generated (deactivated) samples. Ecat samples were chosen over lab-deactivated catalyst as these samples provide the best representation of nickel age distribution in the unit, since it is known that the introduction of nickel contaminants in a laboratory can lead to a distribution of nickel which does not mimic what is seen in an actual FCC unit [1]. To this point, extensive research is focused on the attempt to develop methods to minimize these testing artifacts [28]. Thus, performing such a study on Ecat samples provides results most relevant to industrial application.The Ecat samples chosen for this study were exposed to chloride ions via introduction of gaseous HCl generated by reaction of aqueous HCl with sulfuric acid [29,30]. This procedure is well established in literature for generating HCl. While literature describes the introduction of HCl via liquid solutions as well, such a method was not included in this study, since FCC catalysts are not normally exposed to such liquid media during FCC operation [22]. As a result, they are not designed to withstand this type of liquid interaction; consequently, FCC structural integrity and catalytic performance can be drastically altered by exposure to liquids. The objective of exposure to HCl is to monitor any conversion of oxidized nickel on Ecat into NiCl2 species. A control experiment was also run exposing Ecat to gaseous N2. Following each introduction of chloride ions or control treatment, each catalyst sample was then exposed to H2 to mimic the reducing environment of an FCC riser and reduce any nickel chloride species formed to metallic nickel. The effect of each treatment was then studied by evaluating the physical, chemical, structural, and catalytic changes of the catalysts using particle size measurement, surface area measurement, X-Ray Fluorescence (XRF), Scanning Electron Microscopy (SEM), Advanced Cracking Evaluation (ACE), and CO Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) analyses. The results are presented and discussed below.A round-bottom flask was charged with catalyst (40 g). HCl(\n\ng\n\n) was produced by dropwise addition of HCl(\n\naq\n\n) (12.1 M, ca. 1 mL/min) into a Schlenk flask containing a stirred solution of H2SO4(\n\naq\n\n) (100 mL, 18.0 M). The generated HCl(\n\ng\n\n) flowed into the catalyst-containing flask via a gas dispersion tube with stirring for 1 h at room temperature and then exhausted into a KOH base trap. The mixture was stirred for 1 h at room temperature. The catalyst was dried using a temperature furnace at 100 \u00b0C overnight [29,30].In a control experiment, the same procedure was repeated with a gentle stream of N2(\n\ng\n\n) (ca. 1 L/min) replacing the generated HCl(\n\ng\n\n). A round-bottom flask was charged with catalyst (40 g). The N2(\n\ng\n\n) flowed into the catalyst-containing flask via a gas dispersion tube with stirring for 1 h at room temperature and then exhausted into the atmosphere. The mixture was stirred for 1 h at room temperature. The catalyst was dried in a 100 \u00b0C furnace overnight.The reduction process was adapted from a literature procedure [16]. A Fisher-Porter bottle was charged with treated catalyst (40 g). The system was filled with H2(\n\ng\n\n) and evacuated five times before being pressurized with 75 psig of H2(\n\ng\n\n). The system was heated to 375 \u00b0C, kept at that temperature for 30 min, and subsequently cooled (total time elapsed ca. 2 h).Particle size is measured according to ASTM D4464-10. Particle size distribution in the range of 2.8\u2013176 micrometers were measured using a Beckman Coulter LS13320 with Universal Liquid Module and Ultrasonic unit. Material is dispersed in water, exposed to a beam of light, and the diffraction pattern of the light is used to determine the distribution of particle size.Surface area was measured using the Brunauer\u2013Emmett\u2013Teller (BET) method on a Micromeritics TriStar II according to ASTM methods D 3663 and D 4365. BET uses adsorption isotherms to determine material surface area. The sample was pulverized, and outgassing was performed at 250 \u00b0C for 4 h. Surface area was measured by N2 adsorption and desorption.X-Ray Fluorescence Spectroscopy analyses were performed using a wavelength-dispersive PANalytical PW2400 spectrometer, calibrated by linear regression to data from standards. All samples were prepared by fusion, using a lithium metaborate/lithium tetraborate flux.The catalyst samples were mounted in epoxy and polished to an ultra-flat surface and carbon coated using a Denton DV-502A Vacuum Evaporation System. The BEI analysis was conducted on a Hitachi 3400S Environmental Microscope at 15\u221225 kV. EDX (Energy Dispersive X-Ray Spectroscopy) results were collected at 25 kV on a Bruker Quantax EDS system with Dual 30 mm2 Silicon Drift Detectors (SDD).ImageJ was used to calculate the circularity of all particles in each sample. A circularity index was calculated per the following equation:\n\n\nC\ni\nr\nc\nu\nl\na\nr\ni\nt\ny\n=\n4\n\u03c0\n*\n\n\nA\nr\ne\na\n\n\n\n\nP\ne\nr\ni\nm\ne\nt\ne\nr\n\n2\n\n\n\n\n\n\nA circularity of 1 equals a perfect circle while a circularity of 0 equals a straight line.Advanced Cracking Evaluation (ACE) is a laboratory-scale FCC testing unit which evaluates the activity and selectivity of FCC catalysts in a fixed-fluidized bed reactor [31,32]. As testing is carried out under fluidized conditions, it is commonly used for evaluating FCC catalysts. Ecat treated by various methods were analyzed on an ACE testing apparatus with the following conditions: Reactor temperature: 532 \u00b0C, injector height, 2.125\", standard vacuum gasoil feed, variable time on stream method, 1.2 g/min feed rate, 9 g catalyst loading, 575 s catalyst strip time, liquid strip multiplier of 12, 110 \u00b0C feed temperature, 116 \u00b0C and 177 \u00b0C temperature of first and second feedline heater (respectively), and catalyst to oil ratios of 9, 7, 5, and 3. Coke on catalyst is obtained at the end of a run on a LECO unit.A recently developed, 3-temperature (3-T) pretreatment CO DRIFTS method was used to characterize the nickel on samples treated with N2 and HCl [33]. This method, as opposed to traditional CO DRIFTS, is needed due to the presence of other impurities, which can lead to ambiguous CO band assignments. The samples were ground into fine powders and pretreated with 2.4 % H2/Ar at 200, 400, and 600 \u00b0C sequentially for 1 h at each temperature. The samples were cooled to 30 \u00b0C then underwent a 30 min exposure to 1% CO/Ar for adsorption and a 30 min desorption in Ar while FTIR data were collected. FTIR characterization was performed on a Thermo Fisher Nicolet iS50 FTIR spectrometer equipped with an MCT detector and a Pike Technology high-temperature environmental chamber with a KBr window. Spectrum collection was performed under diffuse reflection mode. Bands were assigned based on CO interaction with metals of different oxidation states and the change in these band intensities with temperature was recorded, which allows the characterization of nickel reducibility upon different treatments.Ecat samples were treated with HCl or N2, reduced by exposure to H2, and then analyzed by several techniques. XRF, surface area, and particle size distribution of each sample were measured and compared to untreated Ecat as a means of evaluating the effect of each treatment on chemical composition and physical integrity of the catalyst. SEM images were also obtained in order to evaluate changes in catalyst particle shape and nickel distribution across different catalyst particles. An image processing method employing ImageJ was used to quantify differences seen between each SEM image. Changes in the dehydrogenation activity of nickel on Ecat following each treatment method was evaluated using ACE analysis. A standard feed was cracked over a range of catalyst to oil (C/O) ratios with each Ecat. The properties of this standard feed are given in the table below (Table 2\n).Hydrogen and coke yields of the Ecat sample are used as a measure for dehydrogenation activity of contaminant nickel. Finally, the oxidation state of nickel, which is hypothesized to be altered by interaction with chloride ions and subsequent reduction, was evaluated by a CO DRIFTS [33]. The results are described herein.Following the treatments described previously, the resulting surface area and chemical composition of the catalysts were analyzed and compared to untreated samples to evaluate how each treatment method influenced the chemical and physical properties of the catalysts. XRF results are shown in Table 3\n. Aluminum, lanthanum, iron, nickel and vanadium are reported as oxides. These elements (via their respective oxides) are all of interest. Aluminum is present in both the matrix and zeolite phases of the catalyst and plays the key role in cracking in both the zeolitic domain (especially to provide selectivity towards valuable hydrocarbons such as gasoline and LPG) as well as in the matrix domain (especially to crack large molecules in the feed). Alumina phases can also be used for trapping contaminant metals such as nickel. Lanthanum is a rare earth element which stabilizes active cracking sites. Iron is both a contaminant and found in the structural framework of the catalyst. As a contaminant, iron can act as a dehydrogenation catalyst to generate coke and hydrogen but is considered significantly less active than nickel (ca. one-tenth the activity). Vanadium is also a feed contaminant which contributes to coke and hydrogen yields, but it is also considered to be less active in generating coke and hydrogen compared to nickel (ca. one quarter of the dehydrogenation activity). As a result, it is important to track these elements/oxides before and after treatment methods to assess how any change in their amount might affect the reactivity or selectivity of catalyst samples. It is important to note that no other catalyst contaminants known to increase coke and H2 were present on the catalyst in significant quantities (>50 ppm). For simplicity, they are not included.Following treatment by either HCl or N2 and reduction with H2, both Ecat samples contained amounts of Al2O3, La2O3, Fe2O3, NiO, and V2O5 that were within instrumental error of the untreated Ecat samples. This indicates that loss of nickel, vanadium, aluminum, iron or lanthanum does not occur during treatment (as expected) and will not influence coke and hydrogen yields in ACE analyses. It is also worth noting that there are different amounts of iron and vanadium between the high and low nickel samples; however, these differences are small (ca. 1000 pm Fe and 500 ppm V) compared to the difference in nickel (ca. 4000 ppm Ni) between samples. Furthermore, taking into account the much lower dehydrogenation activities of iron (ca. 1/10th of Ni) and vanadium (ca. 1/4th of Ni), these differences can be \u201cnormalized\u201d to much lower levels. When comparing to a 4000 ppm difference in Ni, these small contributions from Fe and V are considered insignificant in this study.\nTable 1 shows the surface area and average particle size (APS) of the Ecat samples before and after each treatment method. These physical parameters are important to monitor, since any change in the structural integrity of the catalyst could influence coke and hydrogen yields, thus clouding any change in nickel reactivity. Neither treatment method resulted in a change in surface area or particle size that was outside of the instrumental error of the original Ecat. This indicates that HCl and N2 treatment methods do not significantly alter the catalyst structure.Scanning electron microscopy was performed on catalyst samples before and after treatment to understand both the change in nickel distribution and the structural integrity of the catalysts before and after exposure to chloride ions. SEM studies focused on high nickel Ecat, since the low nickel Ecat samples did not contain enough nickel for detection in SEM-EDX (Energy Dispersive X-Ray Spectroscopy).\nFig. 1\n shows the SEM back-scattering results for treated and untreated Ecat samples. No fragmentation of particles was observed, and the structural integrity of the catalyst particles was maintained. ImageJ was used to calculate the circularity of each particle. A \u201ccircularity index\u201d of 0 to 1 was calculated with 1 indicating a perfect circle and 0 indicating a line. The values were averaged for each treatment method and the results are shown in Table 4\n. Each catalyst sample had the same circularity index of 0.86, thus confirming that no treatment method was destructive to catalyst integrity and that these treatment methods are an effective way to introduce chloride into the catalyst without influencing the structural integrity of the catalyst particle.The SEM-EDX images of nickel on catalyst particles were also examined before and after treatment. Fig. 2\n shows the SEM images for nickel and aluminum overlaid for high nickel Ecat untreated and treated by N2 and HCl.A redistribution of nickel is not apparent from these images; however, this is not surprising given the design of the experiment. The catalysts were not treated in a fluidized environment nor at the high temperatures experienced in an industrial FCC regenerator. As a result, while nickel chlorides can still form, the temperature and lack of fluidization would not be amenable to nickel mobility. A further study of nickel mobility in the presence of chloride ions at conditions closer to that of an FCC unit will be investigated later.Changes in the catalytic behavior of nickel-contaminated Ecat following exposure to N2 or HCl then reduced by H2 were evaluated using ACE analyses. A standard FCC feed was cracked over a fluidized bed of Ecat at different catalyst-to-oil ratios. Since nickel is a known contaminant that produces hydrogen and coke when present on FCC catalyst, the changes in coke yield and H2/CH4 yield ratios during ACE evaluations were compared as a means of assessing nickel activity following different treatment methods.\nFig. 3\n shows the coke vs. conversion results from an ACE analysis of the Ecat sample containing high amounts of nickel. The results showed treatment with HCl prior to the reduction step gave roughly a 1 wt.% increase in coke yield at a given conversion level. This result highlights that the introduction of chloride ions leads to increased coke yield. Since coke is a known product of dehydrogenation from nickel contamination and it is hypothesized that chloride ions facilitate activation of nickel contaminants on FCC catalyst, a higher coke yield following HCl exposure suggests the reactivation of nickel by exposure to chloride ions.\nFig. 3 also shows H2/CH4 yield ratios as a function of conversion for the high nickel Ecat treated by N2 and HCl. As with the coke yield, the H2/CH4 yield ratio was higher for Ecat exposed to HCl than Ecat exposed to N2 (\u223c0.08 wt.%/wt.%). The increase in H2/CH4 yield ratio with HCl treatment also suggests that there are interactions of chloride ions with nickel which increase the dehydrogenation activity of the nickel contaminants on catalyst.The average yields at 72.5 % conversion are reported in Table 5\n. There is a 0.06 wt.%/wt.% and 1.1 wt.% increase in H2/CH4 ratios and coke yields, respectively, when the catalyst is treated with HCl as opposed to with N2. These experiments confirm an average relative increase of 13 % in H2/CH4 ratios and 18 % in coke yields for samples exposed to HCl. These increases in H2/CH4 and coke are both significant, as the average values of 0.52 and 7.4 for H2/CH4 and coke yields following HCl treatment are not within the standard deviation of the H2/CH4 and coke values of the samples treated with N2. Additionally, increases in 13 % and 18 % in H2/CH4 and coke would be considered significant by industry standards as well. These increases in H2/CH4 and coke seen in Table 5 agree with the trends seen in Fig. 3, thus confirming the increased dehydrogenation which occurs when chloride ions are introduced to the system.The combination of increased coke and hydrogen yields following exposure to HCl indicates that nickel contaminant on the catalyst is more active, and that chloride ions play a role in reactivating nickel on the catalyst.Low nickel Ecat samples exposed to N2 and HCl were also analyzed via ACE. The coke yield vs. conversion plots are shown in Fig. 4\n. There was a \u223c0.5 wt.% increase in coke yield following exposure to HCl. However, this increase in coke is not as large as the increase seen (\u223c1 wt.%) following treatment of the high nickel Ecat sample with HCl. This is expected as there is significantly less nickel present on the Ecat, thus less nickel available for potential reactivation.\nFig. 4 also shows the H2/CH4 yield ratio as a function of conversion for the Ecat samples with lower amounts of nickel. As in the high nickel case, there is an increase in H2/CH4 following exposure of the catalyst to HCl (+0.05 wt.%/wt.%). However, as was seen with coke yields, this increase in H2/CH4 is not as pronounced as seen in the case of catalyst containing high amounts of nickel.Multiple ACE experiments were run with the low nickel Ecat. The coke yields and H2/CH4 yield ratios at constant conversion were averaged and are reported in Table 5. The H2/CH4 yield ratios at constant conversion agreed with the trend seen in Fig. 4. There is a 0.06 wt.%/wt.% increase in H2/CH4 when HCl is introduced.The coke yield at constant conversion agreed with the trend shown in Fig. 4. Exposure to HCl leads to a 0.3 wt.% increase in coke compared to exposure to N2. However, it should be noted, that the increase in coke due to exposure to chlorides is almost within standard deviation of each experimental trial. This is not surprising considering the relatively low amount of nickel present on this catalyst.CO DRIFTS experiments were performed on Ecat samples treated with N2 or HCl then reduced by H2. The goal of DRIFTS experiments is to determine the reducibility of the nickel contaminant. Since the Ecat samples contain 0.8\u20131.0% of iron, the CO adsorption on iron would show overlapped peaks in DRIFTS with the CO adsorbed on Ni. In the literature, CO adsorbed on the bivalent or single valent state of nickel is assigned in the range of 2100\u22122200 cm\u22121, CO adsorbed on top of Ni(0) is assigned in 2000\u22122100 cm\u22121, and CO adsorbed on Ni(0) can also be found at 1813\u22122000 cm\u22121 for single-fold or multi-fold bridged adsorption on larger particles [34\u201337]. CO adsorbed on iron (Fe2+, Fe0) is reported with similar peak positions [38\u201340].In order to differentiate the CO adsorbed on iron and the CO adsorbed on nickel, a sequential CO DRIFTS experiment at 3 temperatures is designed under the pretreatment of hydrogen reduction following the protocol reported in detail elsewhere [33]. In these CO DRIFTS experiments, samples were first treated with H2 at 200 \u00b0C before introducing CO, which allows a partial reduction of iron or nickel to a different degree. The CO was then introduced and adsorbed on samples to reach equilibrium. After CO introduction, the CO was allowed to desorb in argon, and CO DRIFTS data were collected during both CO adsorption and desorption time using FTIR. The FTIR spectra presented in this paper were collected at 30 s of CO desorption, which retain the adsorbed CO on solid and remove all the gas phase CO signals. This process was then repeated at 400 and 600 \u00b0C. At each of these temperatures, the reduction degree of nickel and iron is examined by the adsorbed CO FTIR signals. As two different metal oxide materials, nickel oxide and iron oxide are expected to have different reducibilities [41\u201344]. The 3-temperature trend analysis of reduction allows the separation of nickel and iron when their reducibilities are different. The single-beam FTIR spectrum at 30-second-desorption was processed using the IR background spectrum collected before CO was introduced, which allows the comparison of adsorbed CO bond vibration signals at the different reduction temperatures. From this spectrum, information on the oxidation states of metals on the catalyst were determined based upon CO interaction with these sites [45]. With that, the reducibility of nickel can be isolated from the influence of iron, and the effect of N2 or HCl treatment on the Ecat samples can be clearly examined.The CO absorbance spectra of high nickel Ecat treated by gaseous N2 and HCl are shown in Fig. 5\n. The CO absorbance spectra of low nickel Ecat samples treated with N2 and HCl are shown in Fig. 6\n.The bands between 1940 cm \u20131 and 2060 cm \u20131 result from CO bound to Ni(0) and Fe(0) species. The 2090 cm \u20131 band is a result of CO bound to Ni(0). The 2120 cm \u20131 band is CO bound to Fe(II) species. The 2140 cm \u20131 and 2160 cm \u20131 bands result from Ni(II) species. Bands were deconvoluted and their areas were integrated at different reduction temperatures in order to determine differences in nickel oxidation state between treatments. The 2090 cm \u20131 and 2120 cm \u20131, Ni(0) and Fe(II) bands, respectively, have very small areas and it was difficult to infer meaningful information from them in any sample. Thus, the analysis focused on changes in 2140 cm -1 and 2160 cm \u20131 band areas (Ni(II)) and the areas of the bands in the 1900-2070 cm\u22121 region (Ni(0) and Fe(0)).\nFig. 7\n shows the sum of the integrated areas of the nickel(II) derived 2140 cm \u20131 and 2160 cm \u20131 bands for the high nickel Ecat samples as a function of temperature. The CO adsorption at these two bands is much smaller than the bands in the 1900 \u2013 2070 cm\u22121 region, which supports that the sample contains mostly metallic forms of nickel/iron after reduction. The form of Ni(II) may include NiO or nickel aluminate in the Ecat samples, and possibly NiCl2 in the HCl-treated samples. The contribution to the peaks at 2140 cm \u20131 and 2160 cm \u20131 is believed to come from NiO or nickel aluminate rather than NiCl2 and are indicative of the amount of NiO/nickel aluminate compound (in short, referred to as Ni-O in discussion below) present on the catalyst. These proposed peak assignments can be supported by the observation that the catalyst treated with HCl showed significantly lower band area and, therefore, less Ni-O containing compounds, than the sample treated with N2. This could indicate that during treatment with HCl, chloride ions reacted with Ni-O, forming NiCl2, which could then be reduced to metallic nickel during the reduction step. Additionally, the N2 treated sample showed a higher band area at 200 \u00b0C and a decrease in this band area with increasing reduction temperature, while the chloride ion treated sample was essentially unchanged, indicating Ni-O remaining on Ecat treated with N2 is reduced at higher temperatures, while Ecat sample treated with HCl has much less Ni-O remaining. This result suggests different amounts of Ni-O species remain in the Ecat samples under N2 and HCl treatments, with a higher amount of Ni-O in the N2-treated Ecat.The combined band area between 1900\u20132070 cm \u20131 was also examined. These bands are indicative of both Ni(0) and Fe(0) species; as a result, these combined band areas are discussed for both the high nickel and low nickel catalyst samples to understand whether changes in band area are influenced by changes in Fe(0) or Ni(0) compounds, as the iron levels between the high and low nickel catalysts were similar (within 1200 ppm).\nFig. 8\n shows the combined band area of all bands in the 1900\u20132070 cm \u20131 region for the Ecat samples. It is expected that iron would be reduced before nickel is reduced when exposed to H2. The formation of Ni(0) will become more obvious as the reduction temperature increases, thus the Ni(0) can be separated from Fe(0) in the CO DRIFTS.For the low nickel Ecat, as the reduction temperature increases to 400 \u00b0C, the bands grow for each sample indicating more Ni(0) and Fe(0) are formed. At 600 \u00b0C, important observations can be made. For both low nickel samples, the band area does not increase, which could be an indication that all iron and nickel have been completely reduced to the zero-oxidation state at 400 \u00b0C. This result has implications for the analysis of the high nickel sample. The high and low nickel catalyst samples have comparable levels of iron. Thus, a complete reduction of iron in the low nickel Ecat at 400 \u00b0C indicates that all iron will be reduced to Fe(0) in the high nickel Ecat at 400 \u00b0C as well. Consequently, any changes in band area at 600 \u00b0C for the high nickel samples can be attributed to a change in the amount of Ni(0).The combined area of all bands in the 1900\u20132070 cm \u20131 region for the high nickel catalyst samples can also be seen in Fig. 8. These band areas should reflect the amount of Fe(0) and Ni(0) present. At 200 \u00b0C, the area is similar to the low nickel samples, which is an indication that the primary species being observed here is Fe(0) since the nickel levels are very different between the two catalysts. As the temperature increases to 400 \u00b0C, the band of the HCl treated sample grows much more rapidly than the N2 sample. At 600 \u00b0C, both high nickel samples show an increase in band area, with the HCl treated sample showing a significantly higher increase than the N2 treated sample. Having established that the reduction of iron to Fe(0) is completed by 400 \u00b0C, this would indicate that the increase in band area at 600 \u00b0C is due to a change in Ni(0). This difference in the change in band area would then indicate that the HCl treated sample contains more readily reducible nickel than the N2 treated sample.This large increase in Ni(0) formation vs. temperature for Ecat treated with HCl must be reconciled with the fact that its CO adsorbed on Ni-O band at 2140 and 2160 cm \u20131 does not change with a reduction temperature (Fig. 7). One would expect a large increase in Ni(0) to correspond to a drop in Ni(II). An explanation could be that a nickel species not detected in CO DRIFTS is being reduced to Ni(0) at 600 \u00b0C in the Ecat treated with HCl. Given the HCl treatment applied, it could be that NiCl2 is present on the Ecat treated by HCl and is not distinguishable in the FTIR spectra studied. As this NiCl2 is exposed to H2 at 600 \u00b0C, it is reduced to Ni(0) which is then visible in the analysis. While the exact mechanism is uncertain, the results show clearly that exposure of Ecat to HCl results in significant differences in the reducibility of nickel compared to exposure to N2. This further supports the conclusion that differences in ACE yields are a result of a change in the reducibility of nickel, and that chloride ions are playing a major role in this transformation.This work attempts to demonstrate and characterize the physiochemical and catalytic effects of chloride ions on contaminant nickel in the FCC environment for the first time. Additionally, by performing the study on catalyst samples from actual FCC units, the age-distribution of nickel on the catalyst studied is representative of what can be expected in actual operation. It is acknowledged that uncertainties are introduced by using actual FCC Ecat, but the method development work performed in this study has laid the groundwork to perform future studies in more carefully controlled laboratory conditions. Uncertainties which will be addressed in future work include aspects such as the use of catalyst with the same properties and non-Ni contaminants, examination of the effect of different Ni passivators, the effect of Cl contamination on catalyst activity through in depth studies, and using additional techniques to characterize the state of nickel on Ecat.By studying the change in physicochemical characteristics and catalytic selectivity of FCC catalysts, as well as the reducibility of the nickel on FCC catalyst, clear differences can be seen when catalyst contaminated with nickel is exposed to HCl and then reduced. Catalyst exposed to HCl showed increased coke and H2 yields and contained less Ni-O bonds. These results bridge the gap between existing literature and the FCC environment by showing that chloride ions can interact with nickel contaminant on FCC catalyst. The interaction results in changes in the electronic environment of nickel, which makes it easier to be reduced in the FCC riser. This reduced nickel poses a significant problem to refineries since it is an active dehydrogenation catalyst which produces undesirable coke and H2. This increased coke and H2 brings the FCC unit closer to its operational constraints and inhibits the refinery from reaching the full potential of this important unit operation. The results from this study enable catalyst manufacturers and refiners to further optimize catalyst design and selection as well as operational strategies to limit H2 and coke from nickel contaminants.The authors report no declarations of competing interest.\nCorbett Senter: Conceptualization, Methodology, Data curation, Visualization, Writing - original draft, Writing - review & editing, Project administration. Melissa Clough Mastry: Conceptualization, Methodology, Data curation, Visualization, Writing - review & editing. Claire C. Zhang: Data curation, Visualization, Investigation, Writing - review & editing. William J. Maximuck: Data curation, Visualization, Investigation, Writing - review & editing. John A. Gladysz: Data curation, Visualization, Investigation, Writing - review & editing. Bilge Yilmaz: Conceptualization, Methodology, Supervision, Funding acquisition, Writing - original draft, Writing - review & editing.", "descript": "\n                  Nickel, a common contaminant in crude oil, deposits on Fluid Catalytic Cracking (FCC) catalysts and induces unwanted dehydrogenation reactions. These lead to an increase in hydrogen and coke which inhibits the FCC unit from reaching its optimal operation. Modern catalyst technologies can include nickel passivation strategies to minimize such detrimental effects, and, over time, aging of the nickel on catalyst also diminishes its deleterious activity to some extent; however, reactivation of nickel due to chemical interactions within the FCC unit can retard aging and further penalize the catalytic performance. For the first time, we attempt to demonstrate and characterize the physiochemical and catalytic effects of chloride ions on contaminant nickel in the FCC environment. Equilibrium catalyst (Ecat) samples obtained from industrial FCC units are exposed to chloride ions, and changes in physicochemical characteristics, catalytic selectivity, and the reducibility of nickel are analyzed. These changes indicate the reactivation of nickel and an increase in unwanted dehydrogenation reactions following exposure to chloride ions. Spectroscopic analyses show that the interaction with chloride ions alters the electronic environment of nickel, which makes it easier to be reduced in the FCC riser, and Advanced Cracking Evaluation (ACE) studies show equilibrium catalysts that were exposed to chloride ions gave higher coke and H2 yields. These results bridge the gap between existing literature and the FCC environment by demonstrating that chloride ions can interact and reactivate nickel contaminant on FCC catalysts.\n               "}
{"full_text": "In the last few decades, titanium dioxide (TiO2) has been used as an alternative catalyst for the sterilization process of several pathogenic bacteria. The photo-catalyst performance of TiO2 is proven to be more effective in degrading several types of organic matter contaminants than that of the conventional method of chlorination [1]. Chlorination is an in efficient process as it can cause environmental problems that require further treatment [2]. When organic matter contaminants decompose in water containing TiO2, the photo-catalyst surface becomes much more effective after irradiation with ultra-bandgap light with UV radiation (\u03bb \u2265 385 nm) [3, 4, 5, 6]. TiO2 photo-catalyst can decompose organic materials such as dyes, peptides and microbes through a series of oxidation processes initiated by the formation of holes (h+) in the valence band and hydroxyl radicals (\u00b7OH), while in the conduction band they form radicals (\u00b7O2) in oxidized water.TiO2 photo-catalyst activity is determined by several parameters including crystal structure, surface area, size distribution, porosity and hydroxide density [6, 7]. This performance will affect the electron hole recombination time (e-h+) and the adsorption of organic matter contaminants on the surface of the TiO2 photo-catalyst [8]. TiO2 has three crystal structures namely brookite, rutile, and anatase. The last two crystal structures are thermodynamically more stable [9] with the band gap energy (Eg) of \u00b13.0\u20133.2 eV. When compared to the rutile structure, the anatase TiO2 phase structure has more excellent photo-catalytic properties including electron transfer rate that is 89 times greater, chemically and biologically inert, mechanical toughness, low cost and non-toxic [10, 11, 12, 13, 14]. The photocatalytic process requires photon energy. The anatase phase structure (Eg: 3.2 eV) requires UV light energy with \u03bb \u2265 385 nm, which is the energy required to produce illumination energy in the anatase TiO2 photo-catalyst process. The photo-catalytic process is strongly influenced by the electron-hole recombination time (e-h+), whereas the recombination can be extended if the doping process is carried out using transition metal ions on the TiO2 surface [15].Synthesis of TiO2 has been carried out via different methods such as sol-gel process [16], non-hydrolytic sol-gel route [17], ultrasonic technique [18], chemical vapor deposition [19], microemulsion or reverse micelles and hydrothermal process [20]. High calcination temperatures above 450 \u00b0C are usually required to form its crystal structures. Up to the present time, no method has been reported without calcination to produce anatase TiO2 particles [21]. The nano-sized TiO2 particles doped with metal alloys are of great interest for further development because they can increase the photo-catalytic activity of TiO2 [22, 23]. It is important to note that powder obtained synthetically by sol-gel has several advantages including low temperature, simplicity, microstructure morphology with different phase compositions can be obtained by varying parameters such as temperature, pressure, process duration, chemical species concentration, solution concentration and pH [24, 25, 26].Sterilization is the rate required at cell suspension to inactivate broad-spectrum microbes such as Escherichia coli, Staphylococcus aureus and Bacillus subtilis. The application of TiO2 photo-catalysts to inactivate microbes has been widely reported. Some examples are explained in the following. Using a doped 1% Pd+3 ion on TiO2, it was found that E.\u00a0coli was inhibited by \u00b1 98% after 2 h of UV radiation [27]. Floating TiO2 photo-catalyst has been used for inactivation of E.\u00a0coli [28], S.\u00a0typhimurium [29], and inactivation and inhibition of P.\u00a0aeruginosa virulence factor expression [30]. TiO2 coated with polystyrene foam has been used for inactivation of E.\u00a0coli bacteria [31], and antibacterial against A.\u00a0baumannii [32]. Pt powder doped TiO2 has been used to inactivate L.\u00a0acidophilus, S. cerevisiae and E.\u00a0coli, and it was found that TiO2 can be used to replace conventional disinfectant compounds such as chlorination, ozone, and chloride oxides [33, 34, 35, 36].This paper reports a synthesis of FeCuNi nano-alloy doped TiO2 via the sol-gel method. As has been mentioned previously, dopants from the transition metal group have several advantages because they are the catalysts, low energy levels so that they are easy to capture electrons and hence they can inhibit electron hole recombination [23]. Photo-catalyst activity needs photon energy, one of which from UV irradiation, which has the same energy as the energy gap from TiO2 anatase [37]. Based on this consideration, the FeCuNi doped alloy on TiO2 has antibacterial activity with a higher sterilization rate when synergized with UV irradiation. To study the mechanism in which TiO2 can inactivate microbes is based oxidation process and lipid peroxidation [38, 39], which is usually based on the analysis of refractive index, peroxide value (PV) through the formation of malondialdehyde (MDA) compounds as indicators [40].The synthesis of FeCuNi doped into TiO2 Nanoparticles consists of several phases, initially by preparation of titania sol using titanium isopropoxide (TIP) as the basic element, then mixed in the isopropanol solvent. Diethanolamine (DEA) was used as the additive with a ratio of 1:2 TIP to DEA. TIP addition was done by using nitrogen gas flow. The sol was homogenized for \u00b115 min and then acetate salt was added from Fe, Cu and Ni metals with different composition ratios. The total concentration number was 4 % mol to TIP matrix. Sol solution was then homo-genized for \u00b12 h at room temperature. Then, the sol was oven-heated at 100 \u00b0C\u2013110 \u00b0C for \u00b115 h to allow dry gel formation. To obtain FeCuNi\u2013TiO2 powder, dry gel was burned in the furnace at 400 \u00b0C\u2013600 \u00b0C under nitrogen gas flow of 100 psi for \u00b1 2\u20133 h to prevent oxidation of metal Fe, Cu and Ni. FeCuNi\u2013TiO2 was characterized using XRD (X\u2019 Port PAN Analytical, Rigaku RINT\u20132400), SEM-EDX (JEOL JSM 6360 LA), TEM (Philips CM 12 Analysis Docuversion 3.2 image) and TG-DTA (Quantachrome, Serial 1089111903. Model: AS-68).In this experiment, E.\u00a0coli (Gram \u2212), S.\u00a0aureus and B.\u00a0subtilis (Gram +) were used as models. Nutrient Broth (NB) was used as a media for bacterial culture stock preparation. Pure bacterial culture of dense slant gelatin Nutrient Agar (NA) of 24\u201348 h was inoculated into the NB as sterile liquid medium. Aerobic base was incubated inside a rotary shaker for 24 h, at 37 \u00b0C and 120 rpm speed. Then, cell production was continued in a medium in the same condition. Cells were harvested after 8 h of incubation process and then centrifuged at 8000 rpm, for 15 min. Cell sedimentation was rinsed with sterile aquadest twice repeatedly and then centrifuged again at 8000 rpm for 15 min. The sedimentation was given cell suspension by adding phosphate buffer (pH: 7.0) at 1:10 ratio. Cell suspension preparation for photo-catalyst reaction samples was made by dilution treatment using sterile phosphate buffer up to 103\u2013106 cell/mL cell concentration.To determine the inhibition power of FeCuNi\u2013TiO2 against the growth of bacteria, a diffusion medium which consisted of NA media for bacteria was prepared. \u00b1 15 mL NA was poured into a Petri dish after the medium was frozen, then the surface of the medium was lubricated evenly by the following: bacterial cell suspension with 105 cell/mL cell concentration, at volume of 0.1 mL. A stainless steel cup was used for the addition of 15 mg of FeCuNi\u2013TiO2. Incubation was performed inside a chamber with a vertical lamp radiating UV (\u03bb \u2264 385 nm) at temperature of 37 \u00b0C for 24 h. The intensity of UV radiation was monitored by a detector (Blue Light Safety Detector UV) with intensity set at 3.25 mW/cm2. The results of the inhibition zone diameter were measured in millimeters. The processing was done in an aerobic manner, duplo and aseptic. Controlling was done without addition of FeCuNi\u2013TiO2. Using the diffusion method, inhibition efficiency was determined as the optimum sensitivity boundary of FeCuNi\u2013TiO2 to bacteria cells by adding 0\u20133.5 g/L of FeCuNi\u2013TiO2 powder. Additionally, the inhibition zone was used as a preparation sample to examine the physical injury of microbes with the use of SEM. Sample preparation was done by applying the freeze drying method. Inhibition zones resulting from FeCuNi\u2013TiO2 in bacteria were called the outer inhibition zone and inner inhibition zone, whereas the growing zone is used as a control. These parts were cut into 5 \u00d7 5 mm size, and were steamed with 2 % osmium tetroxide (OsO4). The sample was then dipped into liquid nitrogen steam at \u2212210 \u00b0C and placed into the freeze dryer (Emitech K 750) for \u00b110 h. The sample was then coated in gold plating to the size of 5\u201310 mm and then monitored by SEM.A total of 0.1 mL bacteria cell supernatant was transferred into NA media inside a petri-dish which was then lubricated evenly (spread plate) on the surface of the media, and incubated at 37 \u00b0C for 24 h. After 24 h of incubation, the growth of the colony was examined and counted with a colony counter equipment. The count was then converted using the following Eq. (1) which results in a percentage value reported.\n\n(1)\n\n\n%\n\nInhibition\n\n=\n\n\nthe\n\nnumber\n\nof\n\ncolony\n\ncontrol\n\u2212\nthe\n\nnumber\n\nof\n\nsample\n\ncolony\n\n\nthe\n\nnumber\n\nof\n\ncolony\n\ncontrol\n\n\n\n\u00d7\n\n100\n%\n\n\n\n\nBacteria cell sample from photo-catalyst was used to determine the malondialdehyde (MDA). A 2 mL sample was transferred into a test tube, added with 4 mL of 10% trichloroacetic acid (TCA), then homogenized and centrifuged at a speed of 11,000 rpm for 45 min. 6 mL of 0.67% 2-Thiobarbituric acid (TBA) was added into supernatant, incubated for 30 min in a hot water bath, then cooled down in an iced cup for 30 min. Next, it was re-centrifuged at speed of 11,000 rpm for 45 min. Supernatant was used in absorbance measurement using spectrophotometric at \u03bb \u2264 400 nm.The preparation of photo-catalyst media were transferred into 1 mL a beaker glass, with the initial bacteria cell suspension of 103\u2013106 cell/mL, and into 9 ml of sterile NB media and 15 mg of FeCuNi\u2013TiO2 powder. The variations in conditions were: Varied radiation system of UV \u03bb \u2264 365 nm and without UV radiation, incubation time of 30\u2013210 min, and 0\u20133.5 g/l of FeCuNi\u2013TiO2 concentration. As a control, another experiment was performed without added FeCuNi\u2013TiO2 powder. Then, the mixture of photo-catalyst reaction was stirred in a magnetic stirrer or sonicator (50 kHz ultrasonic wave frequency). Radiation intensity was vertically controlled by putting a beaker glass surface in a 30 cm distance from the radiation source. The intensity was monitored by using a detector (Blue Light Safety Detector UV) of 3.25 mW/cm2. The process was done in an aerobic duplo and aseptic manner. The inhibition percentage of bacteria was quantitatively determined by applying 2 measurement methods namely plate count agar (PCA) that is based on the calculation of the number of colonies and TBARs that is based on the number of MDA product formation as a result of peroxide lipid.The XRD patterns of FeCuNi\u2013TiO2 powder synthesized using varying calcination temperatures (400 \u00b0C, 500 \u00b0C, and 600 \u00b0C) are presented in Figure\u00a01\n. All the XRD patterns are indexed according to an anatase TiO2 standard diffraction pattern with the tetragonal I41/amd space group (ICSD-154604) and rutile TiO2 with the tetragonal P42/mnm space group (ICSD-97277). All three XRD patterns matched well with the standard XRD of the TiO2 phase without any additional peaks, confirming the formation of single-phase products. No peaks corresponding to oxides of each Fe, Cu, and Ni metals we reobserved in the doped TiO2 samples, which thus demonstrates that the substitution of all metals in TiO2 host lattice was successful.At calcination temperatures of 400 \u00b0C and 500 \u00b0C, the observed XRD peaks at 2\u03b8 = 24.8\u00b0, 37.3\u00b0, 47.4\u00b0, 53.6\u00b0, 54.7\u00b0, 62.1\u00b0 corresponding to reflection planes (101), (004), (200), (105), (211), (204) confirmed the formation of single anatase phase of TiO2. When the calcination temperature was raised to 600 \u00b0C, the major peaks characteristic of the rutile phase peaks shown at 2\u03b8: 27.4\u00b0, 35.7\u00b0, and 40.9\u00b0 appeared, as highlighted in Figure\u00a01. This suggests that a phase transition from anatase to rutile initially occurred around 600 \u00b0C, which agrees with previous reports. It was revealed that the calcination process at 600 \u00b0C initially led to the structural transformation from anatase to rutile (A \u2192 R). Higher temperatures can cause all crystal position turns to defect crystal wherein the cutting-off in M-TiO2 atoms occurs. Consequently, the FeCuNi\u2013TiO2 structure experienced restructuring and a transformation occurred on the structure. Also, the structure of anatase becomes unstable thermodynamically at high temperatures which causes the anatase particles to stick together to form larger particles and the interface of the anatase particles will become the rutile phase nucleation, resulting in the transformation from anatase to rutile phase [41, 42]. The formation of both anatase and rutile phases was further confirmed by the refinement analysis discussed below.It was also noticeable that the XRD peaks became sharper as the calcination temperature was increased, indicating an increase in crystallinity. Using the full width at half maximum value (FWHM), the crystallite size of the particles was estimated using the Debye\u2013Scherrer\u2019s equation [43]. The average crystallite size was approximately 13.9 nm, 16.8 nm, and 20.2 nm, which increased with increasing calcination temperature. It is expected that the calcination temperature plays a crucial role to accelerate the crystal growth, leading to an increase in larger crystallite size and rising intensity of the anatase phase.XRD data were then refined using the Le Bail refinement technique using Rietica software [27] to determine the phase formations and crystal structure in detail. The initial refinements considered the structural parameters of anatase TiO2 with a tetragonal I41/amd space group (a = b = 3.7862 \u00c5, c = 9.4951 \u00c5; \u03b1 = \u03b2 = \u03b3 = 90\u00b0) (ICSD-154604). All structural parameters were then automatically refined to obtain the best fits between the refinement patterns and optimize the value of reliability factors (R\n\np\n, R\n\nwp\n, and \u03c7\n2). Figure\u00a02\n shows the Le Bail fits of the XRD patterns of FeCuNi\u2013TiO2 samples. For sample calcined at 400 \u00b0C and 500 \u00b0C, the refinement was done with a single-parameter system, since the XRD peaks only show the presence of a single anatase phase. The profile plots in Figure\u00a02a and b show good fits between experimental and calculated patterns for both samples and all peaks matched well with the Bragg reflection of the anatase phases, indicating the existence of both phases. The refinement results confirm that the synthesized FeCuNi\u2013TiO2 samples at 400 \u00b0C and 500 \u00b0C were a single phase of anatase TiO2 without any formation from rutile or brookite phases, which adopts a tetragonal symmetry with a I41/amd space group.Considering the formation of mixed anatase and rutile phase in the sample with a calcination temperature of 600 \u00b0C (as highlighted in Figure\u00a02), we therefore refined the XRD data using the multiphase refinement system. Refinement was done accordingly using the parameter of anatase TiO2 phase as the major phase and added parameter of rutile TiO2 with a tetragonal P42/mnm space group (a = b = 4.6257 \u00c5, c = 2.9806 \u00c5; \u03b1 = \u03b2 = \u03b3 = 90\u00b0) as the secondary phase. As a result, the profile of refinement plots displayed in Figure\u00a02c shows a good fit of all XRD patterns and provides clear evidence for the formation of mixed-phase of anatase and rutile TiO2, according to the Bragg reflection of each phase. The phase fractions obtained from the refinement data were approximately 82.3% for the anatase phase and 17.7% for the rutile phase.The refined lattice parameters and unit cell volumes are shown in Table\u00a01\n. All lattice parameters essentially increased as the calcining temperature increased, leading to an increase in cell volume. As expected, the increase in crystal volume can be correlated to the increased crystallite size occurring due to varying calcination temperatures. The appropriate value of reliability factors (R\n\np\n, R\n\nwp\n, and \u03c7\n2) justified the accuracy of refinement results. Since the focus of this study was on the stable anatase TiO2 for higher photo-catalytic activity than that of rutile or mixed phases, the sample calcined at a temperature of 500 \u00b0C exhibiting pure anatase phase and higher crystallinity was chosen for subsequent analysis.Thermal analysis of nanoalloys FeCuNi\u2013TiO2 under nitrogen atmosphere was performed to study the effect of mass reduction with the increasing temperature. The mass reduction from the TG analysis and the DTA pattern displayed the effects of temperature on the change of the nano structural phase of TiO2 [44]. The reduction in mass at certain temperatures resulted in a change in the phase structure of FeCuNi\u2013TiO2 into two structural phases, namely the anatase phase and the rutile phase. This can be understood since, thermodynamically, the rutile phase is formed at temperatures above 500 \u00b0C and the rutile structure is more stable than the anatase phase [45].\nFigure\u00a03\n shows the TG-DTA FeCuNi\u2013TiO2 pattern, where there are four exothermic patterns that fluctuate in the temperature range of 200 \u00b0C to 500 \u00b0C. The first pattern in the temperature range of 25 \u00b0C\u2013200 \u00b0C indicates a reduction in mass of FeCuNi\u2013TiO2 due to the release of water or organic solvents from the precursor and additive mixtures used in the synthesis process. The second stage at a temperature of 300 \u00b0C, in which in this condition the mass reduction is greater, indicates degradation of the organic residue. The exothermic pattern at a temperature of 300 \u00b0C\u2013400 \u00b0C shows crystal growth and a transformation of the FeCuNi\u2013TiO2, phase in the anatase phase structure. At temperatures \u2265500 \u00b0C, there is a transformation of the anatase structure pattern to the rutile phase structure with greater weight loss and stability at high temperatures [46].\nThe SEM pattern of FeCuNi\u2013TiO2 resulting from calcinations at 500 \u00b0C is shown in Figure\u00a04\na. Each produced a rough surface like that of a piece of rocky stone where ion dopant particles were distributed evenly and homogeneously on the surface of FeCuNi\u2013TiO2 with different sizes. FeCuNi\u2013TiO2 powder surface shown by SEM indicated similarities. However, using EDX measurements, their different chemical compositions were identified: FeCuNi\u2013TiO2 1:2:1, 97.06 % at 4.5 keV (Figure\u00a04b).The TEM pattern from the FeCuNi\u2013TiO2 powder is shown in Figure\u00a05\na. It is observed that the highest photo-catalytic activity occurs in FeCuNi composition with a 1:2:1 ratio. This FeCuNi\u2013TiO2 nanoparticles form three-dimensional crystals that are regularly structured in a spherical shape. The result of TEM measurement shows that most FeCuNi\u2013TiO2 particle sizes are of 10\u201315.7 nm in size (Figure\u00a05b). The particles are distributed evenly as much as 45%. The results of the TEM measurement on particle size leads to the correlation with the particle size in Debye-Scherrer formula (Eq. (1)).The free radical \u00b7OH attack from the process of electron-hole photo-generation of FeCuNi TiO2 powder at the bacteria cell partition can be indicated as Malondialdehyde compound (MDA) formation. MDA is the final product from the result of an oxide saturated process in the cell membrane. Figure\u00a010 shows the numbers of MDA product which were formed from a series of \u00b7OH radical attacking process in the bacteria cell partition that was determined by applying the TBARs method [47].When TiO2 nano particles exist in a medium containing moisture or water and then receive UV radiation at appropriate wavelength with energy as needed by TiO2 semiconductor, electron-hole photo-generation will occur that produces free hydroxyl radicals \u00b7OH. The \u00b7OH radical is extremely effective as a toxic compound which kills microorganisms. When the \u00b7OH radical interacts with a microbe cell wall, the DNA chromosome of the microbe will develop a thymine dimer that allows knots among thymine base inside the similar DNA strand. This thymine dimer will obstruct the formation of the double helix and disturbs the normal replication of DNA. Cell growth is obstructed and eventually leads to cell death. The inhibition of FeCuNi\u2013TiO2 photo-catalyst at microbe can be explained by the attack from O2 radicals and \u00b7OH of photo-generated electron holes on the catalyst surface. Among the three types of species, \u00b7OH radical is the most reactive because it has very effective oxidation capabilities for various kinds of organic compounds, such as microbe cells [47].The doping process can increase photo-catalytic activity of FeCuNi\u2013TiO2. Doping can stimulate free radical formation with a high hydroxyl series density through a redox reaction on the FeCuNi\u2013TiO2 surface. Doping of transition metal ions which have been multiplied by three can significantly increase the photo-biocatalytic activity in inhibiting microbes. Based on this consideration, the FeCuNi doped alloy on TiO2 has antibacterial activity with a higher sterilization rate when synergized with UV irradiation. The mechanism of TiO2 inactivation against microbes can be studied from oxidation process and lipid peroxidation. Lipid oxidation is usually based on the analysis of refractive index, peroxide value (PV) through the formation of malondialdehyde (MDA) compounds and 2-thiobarbituric acid reactive substances (TBARS) as indicators, as shown in Figure\u00a06\n [38, 39, 40].FeCuNi doped by multiple of three is more effective against bacteria E.\u00a0coli (+++), S.\u00a0aureus (++) and B.\u00a0subtilis (+). E.\u00a0coli, S. aureus, B. subtilis microbes that were chosen as models for the examination of microbe inhibition of FeCuNi\u2013TiO2. These are pathogens in nature when they interact with humans directly or indirectly. E.\u00a0coli is a negative gram bacterium, having a thinner layer of peptidoglycan cell wall when compared to positive gram bacteria such as S.\u00a0aureus, and B.\u00a0Subtilis [46]. Figure\u00a07\nb shows the inhibition power of FeCuNi\u2013TiO2 at E.\u00a0coli greater than S.\u00a0aureus greater than B.\u00a0Subtilis.Microbe cells which were given inhibition treatment with FeCuNi\u2013TiO2 powder that have a wide inhibition zone was used to examine the effect of FeCuNi\u2013TiO2, whereas the zone of microbe growth without the FeCuNi\u2013TiO2 treatment was used as a control (Figure\u00a07b). The reaction effect of photo-biocatalyst was observed based on the interaction between bacteria cells and FeCuNi\u2013TiO2 that were blended continuously and each bacterium was given vertical UV radiation as a function of time. As a result of photo-biocatalytic reaction, a number of bacteria cells died off, thus causing the reduction in the number of initial cells by 104 to 105 cell/mL for each bacterium. This germicide action was examined by Upreti, et al (2018), to inhibit the E.\u00a0coli bacteria inside Luria Bertani culture which was radiated by UV and nanoparticle composite Nd+3 doped TiO2. A longer time of UV radiation caused a reduction in concentration of E.\u00a0coli bacteria [48].The effectiveness of FeCuNi\u2013TiO2 powder as an antimicrobial is determined by the performance of FeCuNi\u2013TiO2 photo-catalyst. In these pictures (Figure\u00a07), it can be seen that the inhibition response of FeCuNi\u2013TiO2 to the three bacteria species as a function of time can reduce the number of bacteria colonies following the extension of inhibition time. Using Eq. (1), inhibition percentage was counted based on the reduction of the number of initial colonies at 0 h and at the end of photo-catalyst reaction time between bacteria cells and FeCuNi\u2013TiO2 powder. Figure\u00a07b indicates the zone inhibition of E.\u00a0coli bacteria cells and the control. Species as the function of time can reduce the number of bacteria colonies following the extension of inhibition time. Using Eq. (1), the inhibition percentage was counted based on the reduction of the number of initial colonies at 0 h and at the end of the photo-catalyst reaction time between bacteria cell and FeCuNi\u2013TiO2 powder.The inhibition percentage of bacteria cells by UV radiation was analyzed based on measurements of cell turbidity during UV radiation at 120 min Figures\u00a08 and 9\n\n\n display physical changes of microbe cells as a result of FeCuNi\u2013TiO2 powder and UV radiation as observed by SEM to ensure the effect of these treatments on bacterial cell damage. Figure\u00a08a, b and c shows the physical appearance pattern of SEM: E.\u00a0coli; S.\u00a0aureus; B.\u00a0subtilis before inhibition by FeCuNi\u2013TiO2 powder, and Figure\u00a08d and e shows the physical appearance pattern of SEM: E.\u00a0coli; S.\u00a0aureus; B.\u00a0subtilis before inhibition by FeCuNi\u2013TiO2 powder. E.\u00a0coli is 61.6% in Figure\u00a09a, S.\u00a0aureus is 52.8%, in Figure\u00a09b and B.\u00a0subtilis is 34.3% in Figure\u00a09c. Without the addition of FeCuNi\u2013TiO2 and UV radiation, the inhibition percentage for cells are as follows: E.\u00a0coli is 12.2 %, S.\u00a0aureus is 8.8 % and B.\u00a0subtilis is 6.5%. Potential synergy was found when FeCuNi\u2013TiO2 was combined with UV radiation in the application of stirring system, where by an increase in photo-biocatalyst activity occurred for each bacterium: E.\u00a0coli 96.4%, S.\u00a0aureus 92.8% and B.\u00a0subtilis 78.3%.When mechanism treatment was applied by means of ultrasonic wave from sonicator at 50 kHz frequency, there occurred an increase in inhibition efficiency by 6% for 120 min application time. The optimum inhibition efficiency of FeCuNi\u2013TiO2 in each bacterium was noted as follows: E.\u00a0coli 1.0 g/L, S.\u00a0aureus 1.5 g/L and B.\u00a0subtilis 1.5 g/L. This inhibition efficiency was determined by applying Diffusion Method (spread plate) based on the calculation of colony.The free radical \u00b7OH attack from the process of electron-hole photo-generation of FeCuNi\u2013TiO2 powder at the bacteria cell partition can be indicated as Malondialdehyde compound (MDA) formation. MDA is the final product from the result of an oxide saturated process in the cell membrane. Figure\u00a010 shows the numbers of MDA product which were formed from a series of \u00b7OH radical attacking process in the bacteria cell partition that was determined by applying the TBARs method [49].In this work, the performance of TiO2 as an antimicrobial agent has been increased via structural modification and particle size using dopant FeCuNi with a ratio of 1:2:1. The performance of FeCuNi\u2013TiO2 is related to the inhibition efficiency improvement against bacteria in the concentration range of E.\u00a0coli: 1.5 g/L, S.\u00a0aureus: 1.5 g/L, B.\u00a0subtilis: 2.0 g/L. The FeCuNi\u2013TiO2 with particle size of 16.8 nm and surface area of 70.98 m2/g provided more effective inhibition activity based on the measurement of the inhibition zone using diffusion method. The inhibition activity from the highest to the lowest is for E.\u00a0coli followed by S. aureus and B. Subtilis. The photo-catalytic activity of FeCuNi\u2013TiO2 powder as the antimicrobial agent was more effective when it was irradiated using UV with \u03bb = 365 nm, which provided an inhibition percentage in the range of 78.2%\u201396.4%. The final product from the series of chemical inhibition processes indicated the formation of MDA product from a series of \u00b7OH free radical attacking process in the bacteria cell partition that was determined by TBARs method.Yetria Rilda: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.Syukri Arief: Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data.Anthoni Agustien: Performed the experimentsEti Yerizel, Nofrijon Sofyan: analyzed and interpreted the data, analysis tools or data.Hilfi Pardi: Performed the experiments; Analyzed and interpreted the data; Wrote the paper.This work was supported by Andalas University (T/9/UN.16.17/PT.01.03/IS-RPBQ/2022, 7 Juli 2022).Data included in article/supplementary material/referenced in article.The authors declare no conflict of interest.No additional information is available for this paper.", "descript": "\n                  This study reports the application of FeCuNi nano-alloy doped TiO2 synthesized via the sol-gel method as an antibacterial with a sterilization rate greater than 95% under ultra-violet (UV) irradiation. The performance was characterized using X-ray diffraction (XRD), thermal analysis (TG-DTA), scanning electron microscope (SEM-EDX), and transmission electron microscope (TEM). The results showed that the sterilization process of FeCuNi\u2013TiO2 in cell suspension of Escherichia coli, Staphylococcus aureus and Bacillus subtilis increased the effectiveness of UV irradiation at wavelength (\u03bb) \u2265 385 nm after 120 min. The optimum growth inhibition of FeCuNi\u2013TiO2 was observed in the concentrations 1.5 g/L of E.\u00a0coli, 1.5 g/L of S. aureus and 2.0 g/L of B.\u00a0subtilis. The highest antimicrobial efficiency of FeCuNi\u2013TiO2 powder was provided by a particle size of 16.8 nm, surface area of 70.98 m2/g. The increased antimicrobial activity in multiplied-three doped ions was related to the increase of illumination energy of UV absorption in the photo-catalyst process. The inhibition mechanism reaction of the three species of bacteria cell affects the lipid peroxidation process at the microbe cell\u2019s wall. This was indicated by the formation of malondialdehyde (MDA). Lipid oxidation was based on the reaction of 2-thiobarbituric acid (TBARS) as an indicator of primary and secondary oxidation.\n               "}
{"full_text": "The hydrogenation of 2,4-dinitrotoluene (DNT) is an important industrial process to produce 2,4-toluelendiamine (TDA). TDA is an intermediate in the production of toluene diisocyanate (TDI), which is one of the main components in the manufacture of polyurethane (PU) [1]. In the catalytic hydrogenation of DNT, carbon-, silica- and alumina-supported transition metals (Pd, Pt, Ni, etc.) or transition metal oxides are the most commonly used catalysts [2\u20138]. Many intermediates can be formed during the process, such as 4-(hydroxyamino)-2-nitrotoluene (4HA2NT), 2-(hydroxyamino)-4-nitrotoluene (2HA4NT), 4-amino-2-nitrotoluene (4A2NT) and 2-amino-4-nitrotoluene (2A4NT) [9\u201311]. The best catalysts have high catalytic activity, it is easy to use them, and their recovery is excellent. However, it is difficult to separate powder catalysts (such as supported-activated carbon) from the reaction medium due to their stable dispersion forming ability. The use of magnetic nanoparticles might be a solution to this problem, because the small, mobile particles can be easily removed and isolated by a magnetic field. This simplifies the catalyst recovery and recyclability [12], and by avoiding conventional filtration and centrifugation methods, the associated catalyst loss is also eliminated. Therefore, the use of magnetic nano-catalysts is a promising alternative, especially in heterogeneous catalysis. One of the most widely used magnetic material is maghemite [13,14]. Various methods have been applied to synthesize maghemite nanoparticles, such as sol-gel synthesis [15,16], microemulsion [17,18], coprecipitation [19,20], hydrothermal [21], flow injection [22], and combustion methods [23]. However, the overall process usually includes several steps, and post-treatment to activate the catalysts is also required. Therefore, attempts to simplify the catalyst production process were made and a new method has been developed in our research group [24,25]. By using this method, an active catalyst has been developed during the impregnation step [24,25]. The essence of the method is the exposure of the liquid medium to intense ultrasonic effects, where the induced sound waves create cycles of high and low pressure. Thus, the vapor pressure of the solvent is decreasing momentarily, which results in the formation of bubbles of a few micrometers in size in the mixture. These bubbles are pulsating and growing until they reach a higher pressure range in the liquid, where they collapse as the pressure increases [26]. At this point (\u201chot spot\u201d), a large amount of energy is released, causing the medium to act as reducing agent in the reaction and to initiate the formation of metal, metal oxide, or metal hydroxide solid particles [27\u201332]. By using our recently developed sonochemical method, in this work palladium and platinum nanoparticles were deposited on the surface of maghemite and tested in the catalytic DNT hydrogenation reaction.Iron(III) citrate hydrate (FeC6H5O7 x H2O, PanReac AppliChem Ltd) and polyethylene glycol with 400\u00a0g/mol molar mass (PEG400, Sigma Aldrich Ltd) were used for the production of maghemite nanopowder. Palladium(II) nitrate dihydrate (Pd(NO3)2 x 2H2O, Merck Ltd), hydrogen hexachloroplatinate (H2PtCl6, Reanal Ltd), hydrazine hydrate (N2H4 x H2O, Sigma Aldrich Ltd) and patosolv (a mixture of 90\u00a0vol% ethanol and 10\u00a0vol% isopropanol, Molar Chem. Ltd) were used to provide the palladium and/or platinum content of the metal supported maghemite catalysts.2,4-dinitrotoluene (DNT, C7H6N2O4) was used as reactant, and 2,4-diaminotolune (C7H10N2), 2-methyl-5-nitroaniline, 2-methyl-3-nitroaniline, 4-methyl-3-nitroaniline, and 4-methyl-2-nitroaniline (C7H8N2O2) were used as standards (Sigma Aldrich Ltd) for the GC\u2013MS measurements during the catalytic tests. Methanol (CH3OH) was used as solvent (Merck Ltd) during these measurements.The maghemite nanoparticles and the palladium decorated maghemite were examined by high-resolution transmission electron microscopy (HRTEM, FEI Technai G2 electron microscope, 200\u00a0kV). During the preparation step, drops of the aqueous suspension of samples were placed onto copper grids (Ted Pella Inc., only carbon, 300 mesh). The size of the nanoparticles was estimated based on the HRTEM images and the original scale bar by using the ImageJ software. Powder X-ray diffraction (XRD) measurements were conducted by using a Rigaku Miniflex II diffractometer with CuK\u03b1 radiation source (30\u00a0kV, 15\u00a0mA). To identify the functional groups on the surface of the maghemite nanopowder, a Bruker Vertex 70 Fourier transform infrared spectrometer (FTIR) was used. The prepared maghemite (2\u00a0mg) was added to 250\u00a0mg spectroscopic potassium bromide, and after homogenization a pellet was formed which was used in the measurements.Specific surface area (SSA, m2/g) of the catalysts was determined by nitrogen adsorption-desorption measurements at 77\u00a0K using a Micromeritics ASAP 2020 sorptometer and based on the Brauner-Emmett-Teller (BET) method. The carbon content of the maghemite was measured by Vario Macro CHNS element analyser equipment, and phenanthrene was used as standard (C: 93.538%, H: 5.629%, N:0.179%, S: 0.453%; Carlo Erba Inc). The carrier gas used was He (99.9990%), while O2 (99.995%) was used for oxidation, and the samples were loaded onto tin foils.A recently developed combined method [24,25] which includes combustion and sonochemical steps was applied to synthesize maghemite nanoparticles (Fig. S1, first two steps). In the first step, 7.0\u00a0g iron(III) citrate hydrate was dispersed in 40.0\u00a0g polyethylene glycol (PEG 400) by using a Hielscher UIP1000Hdt ultrasound tip homogenizer for 5\u00a0min (20\u00a0kHz). The colour of the PEG 400-based dispersion changed to red, which indicated that iron oxyhydroxide (ferrihydrite and goethite) colloid has formed. In the second step, the pegylated iron oxyhydroxide dispersion was heated up and the organic compound was burned. Thus, magnetic nanopowder (mainly maghemite) was formed.Palladium nitrate (0.25\u00a0g) was solved in 50\u00a0mL patosolv, and 2.0\u00a0g maghemite was added to the solution to synthesize 5.0\u00a0wt% Pd/maghemite catalyst. In the case of 5.0\u00a0wt% Pt- containing catalyst preparation, 0.21\u00a0g H2PtCl6 was added to 2.0\u00a0g maghemite, and 1\u00a0mL hydrazine hydrate was also used. The alcoholic dispersion of the precious metal and maghemite was sonicated for 2\u00a0min by using the tip homogenizer (20\u00a0kHz, 78\u00a0W). Pd or Pt was deposited onto the magnetic nanopowder solid. The catalysts were then removed from the cleared and transparent alcoholic phase with a neodymium magnet, washed with patosolv, and dried at 105\u00a0\u00b0C overnight. A bimetallic catalyst (Pd-Pt/maghemite) with 4.5\u00a0wt% Pd and 0.5\u00a0wt% Pt was also prepared as described above.The catalytic hydrogenation of 2,4-dinitrotoluene (DNT) was carried out in a B\u00fcchi Uster Picoclave reactor (200\u00a0mL stainless steel vessel with heating jacket). The pressure of H2 was kept at 20\u00a0bar, and the reaction mixture was kept at 303, 313, 323 and 333\u00a0K. Sampling was carried out after 5, 10, 15, 20, 30, 40, 60, 120, 180, and 240\u00a0min on reaction stream. The initial concentration of DNT was 0.05\u00a0mol\u00a0L\u22121 in methanol, and 150\u00a0mL DNT solution and 0.1\u00a0g catalyst were applied during the tests. The formed by-products and reaction intermediates were identified by using a JMS-T200GC AccuTOF GCx-plus chromatograph and a JEOL JMS-T200GC mass spectrometer. For the GC measurements, ZB-1MS column (30\u00a0m\u00a0\u00d7\u00a00.25\u00a0mm, 0.25\u00a0\u03bcm) was used. The collected data were analysed and detected molecular species were assigned by using \u201cNIST library search\u201d, \u201cMolecular ion search\u201d, \u201cExact Mass Analysis of Molecular Ion\u201d, \u201cIsotopic Pattern Analysis\u201d and \u201cEI Fragment Ion Analysis\u201d. TDA formation was followed by using an Agilent 7890A gas chromatograph coupled with Agilent 5975C Mass Selective detector. To determine the formed products, analytical standards (2,4-dinitrotoluene, 2,4-diaminotolune, DNT, 2-methyl-5-nitroaniline, 2-methyl-3-nitroaniline, 4-methyl-3-nitroaniline, 4-methyl-2-nitroaniline, Sigma Aldrich Ltd.) have been used.The activity and selectivity (towards TDA) of the catalysts were determined by calculating the conversion (X, %) of DNT and the TDA yield (Y, %) based on the following Eqs. (1) and (2), respectively:\n\n(1)\n\nX\n%\n=\n\n\nconsumed\n\n\nn\nDNT\n\n\n\ninitial\n\n\nn\nDNT\n\n\n\n\u2219\n100\n\n\n\n\n\n\n(2)\n\n\nY\n%\n=\n\n\nn\n\nformed\n\nTDA\n\n\n\nn\n\ntheoretical\n\nTDA\n\n\n\n\u2219\n100\n\n\n\nKinetic measuremetns on the studied catalytic reaction in the batch reactor system used were conducted based on initial rates estimation aiming to determine the reaction orders with respect to DNT and H2. Towards this goal, the initial concentration of DNT (cDNT,0) was varied (25, 30 and 40\u00a0mmol/L) at constant PH2 of 20\u00a0bar, while the total pressure of H2 (PH2) was varied (10, 20, 30 and 40\u00a0bar) at constant cDNT,0\u00a0=\u00a050\u00a0mmol/L for the Pd/maghemite catalytic system. The DNT concentration decay after the first 2\u00a0min was used to estimate the initial rate of reaction (v\n0) by the following Eq. (3):\n\n(3)\n\n\n\nv\n0\n\n=\n\u2212\n\n\nd\n\nDNT\n\n\ndt\n\n\n\n\nThe initial rate (mmol/L\u00a0s\u22121) of reaction can be expressed by Eq. (4):\n\n(4)\n\n\n\nv\n0\n\n=\n\nk\neff\n\n\n\nDNT\n\n\u03b1\n\n\n\n\nH\n2\n\n\n\u03b2\n\n\n\n\nThe apparent reaction orders with respect to DNT (\u03b1) and H2 (\u03b2) were determined by the linear fit of the lgc\n\nDNT, 0 vs. lgvo relationship at P\u00a0=\u00a020\u00a0bar and lgc\n\nDNT, 0 vs. lgvo at cDNT,0\u00a0=\u00a050\u00a0mmol/L using the logarithmic form of Eq. (4):\n\n(5)\n\n\nlg\n\nv\n0\n\n=\nlg\n\nk\neff\n\n+\n\u03b1lg\n\nc\n\nDNT\n,\n0\n\n\n+\n\u03b2lg\n\np\n\nH\n2\n\n\n\n\nwhere, keff is the effective rate constant (T\u00a0=\u00a0303\u00a0K).The magnetic catalyst support was examined by HRTEM, and the \u03b3-Fe2O3 nanoparticles are clearly visible (Fig. S2, A). The nanopowder is highly dispersed given the average particle size of 22.0\u00a0\u00b1\u00a06.6\u00a0nm (Fig. S2, B). The FTIR results indicates that the nanopowder contains carbon as well (Fig. S2, C). The presence of carbon was confirmed by the appearance of the symmetric and asymmetric vibrational bands of the CH stretching at 2892 and 2835\u00a0cm\u22121. Another IR band at 1631\u00a0cm\u22121 also shows the presence of carbon as it can be assigned to the stretching of the CC bonds. Carbon remained in the sample as a product of the combustion of polyethylene glycol. The exact carbon content was measured by CHNS elemental analysis, and it was found that the sample contained 2.83\u00a0wt% carbon. The magnetic nanopowder contains hematite as well (6.8\u00a0wt%) next to the main maghemite phase (Fig. S2, D).Powder XRD measurements were carried out to clarify the phase composition of the magnetic support. After the sonochemical treatment, the PEG-based dispersion was filtered and washed with distilled water and dried at room temperature in vacuum overnight. Based on the diffractogram of this sample, goethite (\u03b1-FeO(OH), 10.7\u00a0wt%), ferrihydrite (Fe3+\n10O14(OH)2, 22.1\u00a0wt%), and PEG (67.2\u00a0wt%) were identified (Fig. S3). Based on this, it can be concluded that during the ultrasonication the iron(III) citrate reacted with polyol forming iron oxyhydroxide species which transformed through dehydration/dehydroxylation processes to maghemite and hematite during combustion.BET surface area analysis was carried out for the maghemite-supported catalysts and the support alone. The Pt/maghemite system had the smallest surface area, ca. 10.1\u00a0m2/g, followed by the bimetallic system, ca. 19.2\u00a0m2/g. The Pd/maghemite catalyst had a surface area of 23.1\u00a0m2/g which is more than double compared to the Pt/maghemite catalyst. The specific surface area of the catalyst support alone is 41.7\u00a0m2/g.The morphology of the solid particles of catalysts has been studied by using HRTEM (Fig. S4, A, C and E). The individual maghemite nanoparticles formed aggregates (ca. 100\u2013150\u00a0nm). The Pd and Pt nanoparticles were indistinguishable from the support particles on the HRTEM images. However, the powder XRD results confirmed that reduction of platinum and palladium ions was efficient since the samples contain pure metallic phases. On the X-ray diffractogram of the Pd/maghemite system, Pd(111) and Pd(200) lines can be identified at 40.3\u00b0 and 46.4\u00b0 2theta, respectively (Fig. S4, B and Fig. S5). In the case of Pt catalyst, the Pt(111), Pt(200), and Pt(220) lines were detected, which indicates the presence of elemental platinum (Fig. S4, D and Fig. S6). The presence of the precious metals was detectable in the bimetallic Pd-Pt/maghemite catalyst also (Fig. S4, F and Fig. S7). The average size of the Pd and Pt nanoparticles was 4.6 and 7.4\u00a0nm, respectively, based on the XRD results and using the Scherrer equation.The maghemite support alone showed a DNT conversion as high as 77.6% at 333\u00a0K, while the TDA yield was 30.5% after 240\u00a0min of reaction (Fig. S8). However, due to the low TDA yield, adding a precious metal to the system is essential. The apparent reaction order with respect to each of the reactants and the rate equation of reaction was determined experimentally. The reaction order with respect to DNT (\u03b1) and H2 (\u03b2) were determined according to the linear fitting shown in Fig. 1\n.A fractional reaction order of ~1.4 is obtained, while the reaction order parameter with respect to hydrogen can be considered as unity (0.97\u00a0\u00b1\u00a00.06), meaning a first order kineticsat the studied experimental conditions. The derived effective rate constant (Eq. (4)) of the Pd/maghemite system was found to be in the range of 6.4\u00a0\u00d7\u00a010\u22125 and 3.3\u00a0\u00d7\u00a010\u22125 (mmol-0.4\u00a0L-0.4\u00a0bar\u22121\u00a0s\u22121). The apparent reaction order with respect to H2 (\u03b2) is significantly lower in the case of supported Pt (0.55\u00a0\u00b1\u00a00.05, Fig. S9) than Pd catalyst (0.97\u00a0\u00b1\u00a00.06) which might suggest some strong interactions of H2 with the catalyst surface, thus leading to high chemical adsorption rates [33].In the case of Pd/maghemite catalyst, TDA yield was not changed significantly with the reaction temperature, where 30\u00a0K difference in temperature resulted only in ~8% improvement in the yield. The maximum TDA yield was 82.6% and it was achieved by using the Pd catalyst at 333\u00a0K and 20\u00a0bar hydrogen pressure (Fig. 2\n). The supported platinum catalyst provided only 62.0% yield. However, both catalysts can be easily separated from the reaction medium by magnet (Fig. 2).A higher TDA yield was achieved by the Pd/maghemite catalyst, and thus, this was applied in the reuse tests (Fig. 3\n). The DNT conversion was stable and did not decreased significantly even after four cycles (Fig. 3). The corresponding TDA yields remained above 80% during the tests (Fig. 3).Two intermediates were identified during the reaction, 4-amino-2-nitrotoluene (4A2NT) and 2-amino-4-nitrotoluene (2A4NT), which are the semi\u2011hydrogenated species of DNT. The less active Pt/maghemite sample was not able to convert completely the intermediates to TDA (Fig. S10, A, B), while the Pd containing catalyst was successfully achieved this within a reasonable time (Fig. S10, C, D).Various by-products have been formed during the hydrogenation, and these were identified by using isotopic pattern analysis. Condensed derivatives of DNT and TDA such as p-tolualdehyde-2,4-dinitro-phenylhydrazone (C14H12N4O4) and 2-methyl-1-[(2-methyl-4-nitrophenyl)-NNO-azoxy]-4-nitrobenzene (C14H12N4O5) and others have been formed (Table S1).Although both precious metals deposited on maghemite support were highly active, a bimetallic catalyst was also prepared (4.5\u00a0wt% Pd and 0.5\u00a0wt% Pt) and tested under the same reaction conditions as the monometallic ones. The rates of hydrogenation at 303 and 313\u00a0K were similar (Fig. 4\n A), and the reaction rate constant (k) values estimated were 5.4\u00a0\u00d7\u00a010\u22123\u00a0s\u22121 and 5.7\u00a0\u00d7\u00a010\u22123\u00a0mmol-0.4\u00a0L-0.4\u00a0bar\u22121\u00a0s\u22121, respectively. The reaction rate constants were very similar (1.8\u00a0\u00d7\u00a010\u22123\u00a0s\u22121 and 1.9\u00a0\u00d7\u00a010\u22123\u00a0mmol-0.4\u00a0L-0.4\u00a0bar\u22121\u00a0s\u22121) at 333\u00a0K when the monometallic Pd and Pt catalysts were used, but a significant increase of k value (5.7\u00a0\u00d7\u00a010\u22123\u00a0\u00b1\u00a04.0 \u2219 10\u22125\u00a0mmol-0.4\u00a0L-0.4\u00a0bar\u22121\u00a0s\u22121) was achieved (~ 3 times) with the bimetallic catalyst (Fig. 4 B). After 10\u00a0min of hydrogenation, the DNT conversion was 90.4% at 333\u00a0K compared to 65.0% in the case of Pd/maghemite. The TDA yield reached 86.8% at 333\u00a0K with the Pd-Pt/maghemite catalyst, which is slightly higher compared to the Pd/maghemite system (82.6% at 333\u00a0K).Nanosized maghemite powder was synthesized by using a recently developed combined combustion and sonochemical methods. By ultrasonic treatment, iron oxyhydroxide species formed were transformed through dehydration/dehydroxylation processes to maghemite and hematite phases during the combustion step. The prepared magnetic nanopowder was used as catalyst support. Monometallic palladium and platinum maghemite-supported catalysts and their bimetallic PdPt counterpart were successfully synthesized after using a fast, relatively easy, and efficient catalyst preparation method, which does not include post-treatments. Their catalytic activity for the 2,4-toluenediamine (TDA) synthesis was tested, and in each case full conversion of 2,4-dinitrotoluene (DNT) was achieved after 60\u00a0min. However, the TDA yield was higher when Pd/maghemite (82.6% at 333\u00a0K) catalyst was used compared to the Pt/maghemite case (62.0% at 333\u00a0K). By combining the two precious metals, a more active bimetallic catalyst was developed and 90.4% of DNT conversion was reached after 10\u00a0min. Furthermore, after 30\u00a0min of reaction, full conversion of DNT was obtained over the bimetallic catalyst, while the monometallic catalysts exhibited lower conversion rates at the same reaction time of 30\u00a0min. In the present work, three maghemite-supported magnetic catalysts were successfully produced in an easy and fast synthetic route, and their catalytic activity was remarkable. In addition, these maghemite-based catalysts are easily separable from the reaction medium due to their magnetic behaviour.On behalf of all authors, the corresponding author states that there is no conflict of interest.This research was supported by the European Union and the Hungarian State, co-financed by the European Regional Development Fund in the framework of the GINOP-2.3.4-15-2016-00004 project aimed to promote the cooperation between higher education and the industry. We also thank Angelos M. Efstathiou for his insightful comments and corrections, which significantly improved the quality of the manuscript.\n\n\n\nSupplementary material\n\nImage 1\n\n\n\nSupplementary data to this article can be found online at https://doi.org/10.1016/j.catcom.2021.106342.", "descript": "\n                  Maghemite particles were synthesized by using a combined combustion method and sonochemical step. Maghemite was used as carrier to prepare supported Pt and Pd catalysts after deposition via a sonochemical step. The catalysts were immediately ready to be used (metals were catalytically active) for the 2,4-dinitrotoluene hydrogenation to produce 2,4-toluenediamine. The most active catalysts were the Pd/maghemite and bimetallic Pd-Pt/maghemite. The catalysts were easily separable after reaction due to their magnetic properties.\n               "}
{"full_text": "Methanol is an essential commodity to produce key chemicals and intermediates, and an appealing energy carrier [1\u20133]. Its thermocatalytic production (CO2 + 3H2 \u2192 CH3OH + H2O) from captured CO2 and renewable H2 or CO2-rich feeds obtained via biomass gasification is an attractive approach to effectively combat global warming [1,3\u20137]. In fact, the absolute sustainability of a process for methanol synthesis using captured CO2 was shown by a system engineering analysis based on planetary boundaries [4]. Still, heterogeneous catalysts possessing high methanol productivity and durability and efficiently utilizing costly renewable H2 are needed to meet requirements for industrialization. First uncovered in 2016, indium oxide (In2O3) is regarded as a breakthrough system due to its very high methanol selectivity [8\u201310]. In-depth kinetic and mechanistic studies revealed that suppression of the competitive reverse water-gas shift reaction (RWGS: CO2 + H2 \u2192 CO + H2O) is mostly determined by the heterolytic splitting of H2 on the oxygen vacancies acting as catalytic centers [8,9,11\u201313]. Among many carriers explored, monoclinic zirconia (m-ZrO2) stood out as a unique material, providing activity improvements beyond conventional dispersion effects and ensuring stability for over 1000 h on stream [8,14\u201318]. These observations were attributed to its ability to assist CO2 activation and to induce the formation of epitaxially-grown In2O3 or solid In2O3-ZrO2 solutions, featuring additional and/or superior oxygen vacancies [15,16]. Since H2 splitting is rate-limiting on bulk In2O3, several hydrogenation metals (i.e., palladium [19\u201323], gold [24], rhodium [25], platinum [26], cobalt [27,28], and nickel [19]) were lately evaluated as promoters. While nickel is most suited among low-priced promoters, palladium in the form of low-nuclearity clusters embedded in In2O3 unlocked an unprecedented sustained productivity of ca. 1 gMeOH h\u22121 gcat\n\u22121 aiding H2 activation without displaying the intrinsic RWGS activity of palladium particles [20].In general, catalytic performance in carbon dioxide hydrogenation has so far been determined upon single-pass conversion using purified CO2 and H2, but looking towards practical implementation requires considering that CO2 streams attained from distinct carbon capture technologies or gasification processes commonly contain CO in the range of 0.03\u221220 vol.% (Fig. 1\n) [29\u201334]. Additionally, since recycling of unreacted gases is needed to achieve competitive methanol yields and efficiently utilize H2, also CO produced by the RWGS will be recycled back into the reactors in view of the excessive cost for its separation from CO2 [35\u201338].At present, fundamental understanding of the effect of CO on the performance of In2O3-based systems for CO2 hydrogenation remains elusive, being limited to few studies on the bulk oxide [8,11]. CO co-feeding was shown to significantly boost methanol productivity (CO/CO2 = 4; p\nCO =0.19 MPa) due to the in situ creation of additional oxygen vacancies [11], or deplete it (CO/CO2 = 1; p\nCO =0.09 MPa) due to In2O3 over-reduction [8]. Feeding only CO in H2 led to a total activity loss, with In2O3 being fully transformed into metallic indium [8].Herein, we thoroughly examined the behavior of In2O3 in bulk form, and when supported on m-ZrO2 and other carriers, or promoted by palladium and nickel in CO2-based methanol synthesis using hybrid CO2-CO feeds containing practically-relevant concentrations of CO. By applying cycle experiments, in which CO2 is step-wise replaced by CO and vice versa, CO-induced deactivation was observed for all catalysts except for In2O3/m-ZrO2, which was activated. To access structural and electronic alterations underpinning these effects, in-depth characterization was conducted, identifying a main positive contribution by CO and distinct interconnected (de)activation mechanisms. The more attractive In2O3/m-ZrO2 and Pd-In2O3 catalysts were investigated through further dedicated testing protocols to define operation strategies maximizing their methanol productivity. Overall, this study highlights catalytic assessment with CO-containing feeds as an essential element to broaden fundamental understanding and thus guide upcoming design of industrially-amenable systems for CO2-based methanol synthesis.Bulk In2O3 was prepared through controlled calcination of In(OH)3 precipitated according to a reported method [20]. Briefly, In(NO3)3\u00b77H2O (15.5 g, Alfa Aesar, 99.999 %) and Na2CO3 (20 g, Merck, >99 %) were separately dissolved in deionized water (235 and 200 cm3, respectively). Thereafter, the sodium carbonate solution was added to the indium-containing solution under magnetic stirring at ambient temperature until reaching pH 9.2 (ca. 158 cm3, 9 cm3 min\u22121) and the resulting slurry was aged for 1 h. The precipitate was recovered by high-pressure filtration, washed three times with deionized water (1 L each time), dried in a vacuum oven (2 kPa, 323 K, 12 h), and calcined at 573 K (heating rate =2 K min\u22121) for 3 h in static air to yield In2O3.Supported catalysts with 5 wt.% In2O3 were attained by wet impregnation (WI) using self-prepared tetragonal zirconia (t-ZrO2) [15] and commercial monoclinic zirconia (Saint-Gobain NorPro, 95 %) and alumina (Sigma Aldrich, 99 %). To produce t-ZrO2, 20 g of a ZrO(NO3)2 solution (Sigma-Aldrich, 35 wt.% in diluted HNO3, >99 % trace metals basis) were diluted with deionized water (100 cm3). Ethylenediamine (Sigma-Aldrich, >98 %) was added dropwise (ca. 3 cm3 min\u22121) to this solution until reaching pH 10 and the resulting slurry was stirred at 353 K for 3 h. The precipitate was recovered by high-pressure filtration, washed three times with deionized water (250 cm3 each time), dried in a vacuum oven (2 kPa, 323 K, 12 h) and calcined at 973 K (3 K min\u22121) for 3 h in static air. WI encompassed suspending 2 g of carrier in a mixture of deionized water (54 cm3) and ethanol (70 cm3, Fisher Chemicals, 99.8 %). The resulting slurry was magnetically stirred (500 rpm) for 12 h at room temperature. Thereafter, the solvent was removed using a rotary evaporator (B\u00fcchi Rotavap R-114) at 323 K, keeping the slurry constantly at boiling point by lowering the pressure from 180 to 40 mbar. The solid was then dried in a vacuum oven (2 kPa, 323 K, 12 h) and calcined at 773 K (2 K min\u22121) for 3 h in static air.In2O3 catalysts promoted with 0.75 wt.% Pd or 1 wt.% Ni were produced by coprecipitation (CP, catalysts noted as Pd-In2O3 and Ni-In2O3) and dry impregnation (DI, catalysts coded as Pd/In2O3 and Ni/In2O3) [20]. For CP, In(NO3)3\u00b77H2O (3.5 g) and Pd(NO3)2.5H2O (0.025 g, Sigma-Aldrich) or Ni(NO3)2\u00b76H2O (0.056 g, Sigma-Aldrich, >97 %) were dissolved in deionized water (50 cm3). A solution of Na2CO3 (10 g) in deionized water (100 cm3) was added under magnetic stirring (500 rpm) until pH 9.2. The resulting slurry was aged for 1 h. Thereafter, it was diluted with additional deionized water (50 cm3) and the precipitate was recovered by high-pressure filtration, washed three times with deionized water (1.25 L each time), dried in a vacuum oven (2 kPa, 323 K, 12 h), and calcined at 573 K (2 K min\u22121) for 3 h in static air. Upon DI, Pd(NO3)2.5H2O (0.095 g) or Ni(NO3)2\u00b76H2O (0.21 g) dissolved in deionized water (0.4 cm3) were added to In2O3 (4 g, calcined at 573 or 773 K in the case of palladium or nickel promotion, respectively) upon mechanical stirring. After drying in a vacuum oven (2 kPa, 323 K, 12 h), calcination was conducted at 573 K (2 K min\u22121) for palladium-containing samples and at 623 K for those comprising nickel for 3 h in static air.Inductively coupled plasma optical emission spectrometry (ICP-OES) was conducted using a Horiba Ultra 2 instrument equipped with a photomultiplier tube detector. Samples were dissolved in aqua regia during 12 h prior to the analysis, except for zirconia-containing materials. These were digested with the aid of microwave irradiation using a mixture of HCl (Alfa Aesar, 36 wt.%), H2SO4 (Alfa Aesar, 95 wt.%), and HF (Sigma Aldrich, 48 wt.%) with a volume ratio of 2:1:0.5, followed by neutralization with a saturated solution of boric acid (Fluka, 99 %). Nitrogen sorption at 77 K was carried out using a Micromeritics TriStar II analyzer. Prior to the measurement, samples were degassed at 473 K under vacuum for 12 h. The total surface area (S\nBET) was determined using the Brunauer-Emmet-Teller (BET) method. X-ray diffraction (XRD) was measured in a PANalytical X\u2019Pert PRO-MPD diffractometer operated in the Bragg-Brentano geometry using Ni-filtered Cu K\u03b1 radiation (\u03bb=0.1541 nm). Data was acquired in the 10 \u2212 70\u00b0 2\u03b8 range with an angular step size of 0.025\u00b0 and a counting time of 12 s per step. High-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) coupled to energy-dispersive X-ray spectroscopy (EDX) was measured using a Talos F200X instrument operated at 200 kV and equipped with a FEI SuperX detector. Temperature-programmed reduction with CO (CO-TPR) was conducted using a Micromeritics AutoChem HP II at ambient pressure. Samples were loaded in a stainless-steel tube, dried at 423 K in He for 1 h, and cooled down to 273 K (20 K min\u22121) using dry ice in ethanol. The temperature-programmed reduction was then carried out using 1 mol% CO/He and increasing the temperature to 1073 K (5 K min\u22121). X-ray photoelectron spectroscopy (XPS) was performed in a Physical Electronics (PHI) Quantum 2000 X-ray photoelectron spectrometer using monochromatic Al K\u03b1 radiation generated from an electron beam operated at 15 kV and 32.3 W and a hemispherical capacitor electron-energy analyzer, equipped with a channel plate and a position-sensitive detector. Samples were firmly pressed onto aluminum foil patches, which were then mounted onto a sample platen and introduced into the spectrometer. Analyses were conducted under ultra-high vacuum conditions (residual pressure = 5 \u00d7 10\u22128 Pa) with an electron take-off angle of 45\u00b0, operating the analyzer in the constant pass energy mode.The gas-phase hydrogenation of CO\nx\n (CO\nx\n= CO + CO2) to methanol was performed in a PID Eng&Tech high-pressure continuous-flow setup comprising four parallel fixed-bed reactors (internal diameter =4 mm) (Figure S1 in the Supplementary Material). The undiluted catalyst (mass (w\ncat) =0.1 g, particle size = 0.2\u22120.3 mm) was loaded in each reactor, held in place by quartz wool bed set on a quartz frit, and purged with a He (Pangas, purity 4.6) flow of 40 cm3\nSTP min\u22121 for 30 min at ambient pressure. Under the same flow, the pressure was increased to 5.5 MPa for a leak test. The reaction was carried out by feeding a mixture of H2, CO (Messer, purity 5.0), and CO2 (20 vol.% in H2, Messer, purity 4.5), with a molar H2/CO\nx\n ratio of 4 at 553 K, 5 MPa, and weight hourly space velocity (WHSV) of 24,000 cm3\nSTP h\u22121gcat\n\u22121, unless stated otherwise. The impact of CO on the performance of the catalysts was assessed by means of CO2-CO cycle experiments, which are summarized in Fig. 2\n. During full cycles (FC, Fig. 2a), CO2 in the feed was progressively replaced by CO from R ratio (R = CO/CO\nx\n, with CO\nx\n = CO + CO2) equal to 0 to 0.5 in steps of 0.1 and then replenished by replacing CO in an analogous manner, while maintaining a constant molar H2/CO\nx\n ratio of 4. Half-cycles (HC, Fig. 2b) were conducted starting at R = 0 and increasing the CO concentration to reach R = 0.5, while reverse half-cycles (rHC, Fig. 2c) started at R = 0.5 and ended at R = 0. In both cases, the same step size of 0.1 was kept as in full cycles. The effect of temperature (543\u2212583 K) and pressure (3\u20135.5 MPa) at R = 0.2 was investigated while keeping all other conditions constant.Additional experiments were specifically designed to further assess In2O3/m-ZrO2 and Pd-In2O3. For the former, CO\nx\n hydrogenation was carried out at R = 0, 0.2, or 0.5 after a pre-treatment in syngas (H2/CO = 2) of 4 h at 553 K and 5 MPa (Fig. 2d) and an extended full cycle was conducted, equivalent to the regular full cycles but reaching R = 1. For the latter, threshold limit cycles (Fig. 2e) were applied, alike to the regular full cycles but attaining R = 0.1, 0.2, or 0.3. Long-term cycle experiments (Fig. 2f) were conducted on both systems starting at R = 0, moving to R = 0.2, and then restoring R = 0, keeping each condition for 40 h.Response factors (Fi\n) for each compound i in the effluent stream, respective to the internal standard (20 vol.% C2H6 in He, Messer, purity 3.5), in gas chromatographic analysis were determined by Eq. 1, where Ai\n is the integrated area determined for the peak of compound i and \n\n\nn\ni\nin\n\n\n is the corresponding known molar flow at the reactor inlet.\n\n(1)\n\n\n\nF\ni\n\n=\n\n\n\nA\ni\n\n/\n\nn\ni\nin\n\n\n\n\nA\n\n\n\nC\n2\n\n\nH\n6\n\n\n\n\n/\n\nn\n\n\nC\n2\n\n\nH\n6\n\n\nin\n\n\n\n\n\n\n\nTheir values were calculated using the average of 5 points around the expected concentration of the respective analyte. The unknown effluent molar flow and methanol production rate (r\nMeOH) were determined using Eqs. 2 and 3, respectively.\n\n(2)\n\n\n\nn\ni\nout\n\n=\n\n\n\nA\ni\n\n\u00d7\n\nF\ni\n\n\n\n\nA\n\n\n\nC\n2\n\n\nH\n6\n\n\n\n\n\n\n\u00d7\n\nn\n\n\nC\n2\n\n\nH\n6\n\n\nout\n\n\n\n\n\n\n\n(3)\n\n\n\nr\nMeOH\n\n=\n\n\n\nn\nMeOH\nout\n\n\n\n\nw\ncat\n\n\n\n,\n\u2009\n\u2009\nmo\n\nl\nMeOH\n\n\nh\n\n\u2212\n1\n\n\n\ng\ncat\n\n\u2212\n1\n\n\n\n\n\n\nThe methanol space time yield (STY) was determined as the product of r\nMeOH and the molar weight of methanol. Data reported corresponds to the average of 4 measurements preceding a specific time on stream. The carbon balance (\u03b5) was determined for each experiment according to Eq. 4.\n\n(4)\n\n\n\n\u03b5\nc\n\n=\n\n\n1\n\u2212\n\n\n\nn\n\nC\n\nO\n2\n\n\nout\n\n+\n\nn\nMeOH\nout\n\n+\n\nn\nCO\nout\n\n\n\n\nn\n\nC\n\nO\n2\n\n\nin\n\n+\n\nn\nCO\nin\n\n\n\n\n\n\u00d7\n100\n,\n%\n\n\n\n\nThe equilibrium composition of the carbon-based species in the reaction mixture upon single-pass conversion was estimated in the temperature range of 423\u2212723 K at 5 MPa and H2/CO2 = 4 using the AspenPlus v10 software as described in Fig. S2. In the calculations, the Gibbs energy was minimized by using the RGibbs model reactor, and thermodynamic constants from Eqs. 5\u20137 as well as Soave-Redlich-Kwong equations of state were applied.\n\n(5)\nCO + 2H2 \u2194 CH3OH, \u0394H\n298 K = \u201390.8 kJ mol\u22121\n\n\n\n\n\n(6)\nCO2 + 3H2 \u2194 CH3OH + H2O, \u0394H\n298 K = \u201349.2 kJ mol\u22121\n\n\n\n\n\n(7)\nCO2 + H2 \u2194 CO + H2O, \u0394H\n298 K = 42.1 kJ mol\u22121\n\n\n\nThe per-pass and overall conversion, and R value of the feed with recycling of outlet gases, at 553 K, 5 MPa, 24,000 cm3\nSTP h\u22121 gcat\n\u22121 and keeping H2/CO\nx\n at 4, were calculated using Eqs. 8\u201310 and results from simulations (Table S1) based on a simplified process flowsheet built in AspenPlus v10 and shown in Fig. S2b.\n\n(8)\n\n\nper-pass conversion\n=\n\n\nC\n\nO\n\n2,\n\u2009\nreactor\n\u2009\ninlet\n\n\n\u2212\nC\n\nO\n\n2,\n\u2009\nreactor\n\u2009\noutlet\n\n\n\n\nC\n\nO\n\n2,\n\u2009\nreactor\n\u2009\ninlet\n\n\n\n\n\u00d7\n100\n,\n\u2009\n\u2009\n%\n\n\n\n\n\n\n(9)\n\n\noverall conversion\n=\n\n\nC\n\nO\n\n2,\n\u2009\nfresh\n\u2009\nfeed\n\n\n\u2212\n\n\nC\n\nO\n\n2,\n\u2009\n\u2009\nproduct\n\n\n+\nC\n\nO\n\n2,\n\u2009\n\u2009\npurge\n\n\n\n\n\n\nC\n\nO\n\n2,\n\u2009\nfresh\n\u2009\nfeed\n\n\n\n\n\u00d7\n100\n,\n\u2009\n\u2009\n%\n\n\n\n\n\n\n(10)\n\n\nR\n=\n\nCO\n\nC\n\nO\nx\n\n\n\n=\n\n\nC\n\nO\n\n\u2009\nreactor\n\u2009\ninlet\n\n\n\n\nC\n\nO\n\n\u2009\nreactor\n\u2009\ninlet\n\n\n+\nC\n\nO\n\n2,\n\u2009\nreactor\n\u2009\ninlet\n\n\n\n\n\n\n\n\nThe effect of CO on CO2-based methanol synthesis was explored on a broad set of In2O3 catalysts. Most prominent systems comprised bulk In2O3, In2O3 supported on m-ZrO2, and In2O3 promoted by palladium through co-precipitation and by nickel via dry impregnation. In2O3 carried on tetragonal zirconia and alumina as well as In2O3 promoted by nickel through co-precipitation and palladium via dry impregnation were additionally considered for comparative purposes. These materials were prepared according to reported synthesis methods [15,20]. Compositional analysis of the fresh catalysts (Table S2) revealed that the nominal loadings of indium oxide (5 wt.%) and transition metals (0.75 wt.% for Pd and 1 wt.% for Ni) were closely matched in supported and promoted samples, respectively. The BET surface areas and XRD patterns of all fresh solids were also in line with expectations and previous evidence (Table 1\n and Figure S3a,b) [11,15,20].To assess their sensitivity to CO, the catalysts were tested in a four parallel continuous-flow fixed-bed reactor setup by means of CO2-CO cycles. Initially, in so-called full cycles, the CO2 in the feed was progressively replaced by CO, i.e., from R = CO/CO\nx\n = 0 to 0.5, and then step-wise restored, i.e., from R = 0.5 to 0. Pressure and H2/CO\nx\n ratio were kept at commonly applied values for CO2-to-methanol (5 MPa and 4, respectively), while temperature and WHSV were set at 553 K and 24,000 cm3\nSTP h\u22121gcat\n\u22121, respectively, as a compromise for the distinct expected reactivity of the catalysts. The results are expressed in terms of methanol STY, since CO2 conversion and methanol selectivity cannot be accurately determined without the use of isotopically-labelled compounds. Indeed, CO shall react with lattice oxygen from In2O3 or carriers forming CO2 (vide infra) especially at higher R values, and with the water formed by CO2 hydrogenation to methanol again to CO2. Moreover, CO is produced to a variable extent by the competitive RWGS reaction and may be hydrogenated to methanol over some of the catalysts. These remarks highlight the relevance of isotopic labelling experiments in future research to deepen knowledge of beneficial and detrimental effects of CO co-feeding through mechanistic understanding.The methanol STY exhibits quite distinct trends in the full cycles (Fig. 3\na). Along the forward branches, it barely increased (1%, Fig. 3b) for bulk In2O3, it was boosted by 8% on In2O3/m-ZrO2 while it decreased by 40 and 55 % for In2O3 carried on t-ZrO2 and Al2O3, respectively, and it dropped almost linearly by 20\u201340 % for the metal-promoted systems. In the latter case, the dry impregnated palladium-containing catalyst deactivated in a slightly more pronounced manner than the co-precipitated analog (26 vs.23 %), while the reverse holds for nickel-promoted materials (40 vs. 37 %).The backward branches of the cycles followed quite closely their corresponding forward branches only for In2O3/m-ZrO2 and bulk In2O3, while more pronounced hysteresis behavior in the methanol STY was detected for all other catalysts (Fig. 3b). For In2O3/t-ZrO2 and In2O3/Al2O3, the methanol productivity regained 7 and 13 % from the value reached at R = 0.5. Similar evidence was observed for Pd/In2O3 and Pd-In2O3, recovering 10 and 15 % from the same point in the cycle, respectively. In the case of Ni-promoted systems, the methanol STY further decreased by 1\u20135 %.Overall variations of the methanol productivity upon completion of a full cycle revealed mild (i.e., \u22128 to \u22129%) to strong (\u221215 to \u221240 %) deactivation of the catalysts. It is worth noting that, in spite of a CO negative impact on methanol production, the global loss in methanol STY for Pd-In2O3 across the full cycle is only \u22129 %, which is equivalent to that of In2O3/m-ZrO2 (i.e., \u22128 %). Generally, although methanol production is expected to drop to some extent when lowering the CO2 concentration if the admitted CO is not converted to methanol or does it poorly, the extent of alterations in methanol STY points to concomitant structural reasons. While the beneficial effect of CO is expected to relate to the creation of additional oxygen vacancies in In2O3, the negative impacts appear diverse and of reversible and irreversible nature. Strong CO adsorption on palladium may account for the first type, whereas over-reduction and sintering of the oxide and of the metal promoters would explain the second. Still, sintering shall also be modulated by the variable amount of water dictated by the extent of (R)WGS at different R values.To unravel structural and electronic alterations of the catalysts in detail and detangle individual contributions to the observed changes, in-depth characterization was conducted applying sensible methods. Porous and structural modifications were initially addressed by N2 sorption and XRD (Table 1). The S\nBET and pore volume of bulk In2O3 dropped by more than 4-fold (from 129 to 28 m2 g\u22121) and about 2-fold (from 0.41 to 0.22 cm3 g\u22121) after the forward branch of the full cycle and increased by a very slight extent after the subsequent backward branch. The particle size (d\nXRD) augmented from 7.9\u201318.9 nm going from R = 0 to 0.5 and decreased to 17.3 nm upon returning to R = 0, indicating irreversible sintering of the oxide. These modifications explain the overall loss of methanol productivity at the end of cycle, and hint that a positive change counteracts sintering in the forward branch, which displayed a rather constant trend (Fig. 3a). It is worth mentioning that sintering occurs to a significant extent already under pure CO2 hydrogenation conditions, as reported previously [9], with S\nBET decreasing from 123 to 47 cm3 g\u22121 and d\nXRD increasing from 8 to 18 nm, indicating a major role of water in driving this phenomenon. In2O3/m-ZrO2 underwent a minor loss in surface area and pore volume in the forward branch, which was restored upon returning to CO2 hydrogenation conditions. This indicates that the activation of this catalyst upon introducing CO is mainly related to modifications of its surface properties. In2O3/t-ZrO2 experienced progressive slight sintering along the full cycle in agreement with the irreversible depletion of its methanol STY. Palladium- and nickel-promoted In2O3 systems underwent strong modifications, with their surface areas, pore volumes, and particles sizes levelling to equivalent values to bulk In2O3 in the different stages of the full cycle. This is in line with the drop in their methanol productivity. However, the partial restoration of methanol STY over Pd-In2O3 in the backward branch of the full cycle indicates additional and at least partially reversible deactivation mechanisms acting on this specific material. A close analysis of the XRD patterns revealed a weak reflection at 34\u00b0 2\u03b8 due to metallic indium in bulk and metal-promoted In2O3 after the forward branch and the full cycle and In2O3/m-ZrO2 after the forward branch only (Fig. S3a,b) [39]. Since metallic indium alone was reported to be unable to catalyze CO2 hydrogenation, this might contribute to explaining the decreased methanol productivity of bulk In2O3 and, possibly, of the other systems as well [8,11]. Another weak reflection around 40\u00b0 2\u03b8 characteristic for metallic palladium might be recognized in the patterns of Pd/In2O3 and Pd-In2O3 after the half- and the full cycle (Figure S3b) [20,23].To gain insights into the distribution and structure of In2O3 and metal promoters in supported and promoted catalysts, fresh and used samples were imaged by HAADF-STEM and chemical element maps were acquired by EDX. The results reveal an effective dispersion of indium on both m-ZrO2 and t-ZrO2 (Fig. 4\na,c) and of the metal promoters (Fig. 4b,d) on In2O3 in all fresh samples. The catalyst architecture remains unaltered for In2O3/m-ZrO2 after the half-cycle (Fig. 4a), while defined aggregates of In2O3 were visualized on In2O3/t-ZrO2 (Fig. 4c). As for the metal-promoted catalysts, palladium remained well dispersed on In2O3 after the same test (Fig. 4b), whereas Ni agglomerated to form nanoparticles (Fig. 4d). In2O3 and palladium stay practically unaltered in In2O3/m-ZrO2 and Pd-In2O3, respectively, even after full cycle experiments, highlighting their robustness against sintering in these materials.Since CO is a strong reducing agent, the reducibility of all fresh catalysts was studied by CO-TPR (Fig. 5\na,b). The curve obtained for bulk In2O3 is composed of a small feature with maximum at ca. 370 K, followed by a pronounced multicomponent peak between 480\u2212720 K, and a further steadily increasing signal at higher temperatures. The profiles of the catalysts carried on the two zirconia polymorphs also show three main signals, which based on the results produced by the supports alone, are considered as surface reduction of zirconia and highly dispersed In2O3, followed by reduction of the two bulk phases in two temperature regions (Fig. 5a). In the second signal, In2O3 contributes more than the carrier for In2O3/m-ZrO2 and the third feature is mostly due to the support for In2O3/t-ZrO2. It is worth noting that the second In2O3 reduction peak starts at a temperature comparable to the reaction conditions (ca. 553 K) for In2O3/m-ZrO2, while at much higher temperatures for In2O3/Al2O3 and In2O3/t-ZrO2 (773 and 800 K, respectively) (Fig. 5a). Based on the XRD and microscopy results (Figs. S3a and 4 c, respectively), this could be due to a larger In2O3 particle size and a strong interaction of In2O3 with the carrier for the alumina- and the tetragonal zirconia-based systems. Furthermore, the peak feature at 553 K for In2O3/m-ZrO2 is also associated with a higher CO consumption, indicating that oxygen is extracted more readily from this system, whereby the carrier may also contribute to this phenomenon due to the tensile force exerted. [15] The curves of the metal-promoted catalysts resemble that of the pure oxide (Fig. 5b). For palladium-containing materials, the first peak was observed at a slightly inferior temperature than for In2O3 (i.e., 340\u2013360 vs. 370 K), suggesting that the promoter facilitates reduction by CO. Interestingly, additional sharp peaks between 400\u2212573 K are present in the curves of Ni/In2O3 and Pd/In2O3, some of which could be related to the reduction of segregated NiO and PdO particles or to the formation of NiIn intermetallic phases (Fig. 5b). STEM-EDX of Ni/In2O3 (Fig. 4d) revealed that nickel sinters into nanoparticles upon reaction at R = 0.5, which are likely reduced to the metallic state by the gaseous environment. Since it is well-documented that metallic nickel is a very active methanation phase, the absence of methane product at any stage of the reaction over Ni/In2O3 suggests that the catalytic properties of nickel were modified by In2O3, whereby the formation of a NiIn intermetallic phase is a plausible option. Indeed, alloying of nickel with indium and other metals such as iron and tin has been reported and shown to foster the production of methanol or CO over methane in CO2 hydrogenation [40\u201342]. Regarding the Pd-In2O3 catalyst, the in situ generation of intermetallics is unlikely. Indeed, STEM-EDX of this sample (Fig. 4b) indicates that palladium remains well dispersed after half- and full cycle experiments and that irreversible activity loss is mostly due to indium oxide sintering (see Table 1). Moreover, Pd-In intermetallic compounds have been evidenced to lead to permanent depletion in methanol productivity [23,43,44]. The findings from CO-TPR also confirm that CO can generate CO2 while reducing the catalysts during reaction, with the above-discussed implications on the determination of CO2 conversion, especially considering that the partial pressure of CO is much greater in the reactor than in CO-TPR experiments. Indeed, conducting the CO-TPR analysis at 5 MPa showed a broad signal with unresolved components starting at lower temperature (not shown). It is worth noting that reduction of In2O3, carriers, and promoters by CO generally occurs at lower temperatures than with H2 (ca. 400\u2212500 K), highlighting its stronger reducing power and role in facilitating the formation of oxygen vacancies [8,11,14,15,20]. However, it is difficult to define a clear correlation between CO content and catalyst reduction degree upon hydrogenation of hybrid CO\u2212CO2 feeds because the CO and H2 are present simultaneously and CO takes part in the reaction equilibria, being possibly used as a carbon source to produce methanol, as discussed above.To shed light on the catalyst surface composition and electronic properties, bulk In2O3, In2O3/m-ZrO2, and Pd-In2O3 were examined by XPS in fresh form and after the half- and reverse half-cycles. All samples exclusively contained indium in oxidized form and no significant change in its surface content was detected after the full cycles (Table S3 and Fig. 5c) [16,23,45]. Remarkably, all In 3d core-level spectra of In2O3/m-ZrO2 show a shift of ca. 0.5 eV to higher binding energy compared to those of the other catalysts (ca.444.3 eV), indicating charge transfer from metal to support and a more oxidic character of the indium species carried. This corroborates previous evidence that indium cations are directly bound to the zirconia lattice, either through epitaxial growth or in a solid solution [15,16]. For Pd-In2O3, Pd2+ species in the fresh catalyst were predominantly reduced to metallic palladium upon use (Fig. S4a), with residual Pd2+ species still present being likely due to the short exposure of the samples to air for the ex situ analysis. The surface palladium content almost doubled in all samples retrieved after the full cycles (Table S3). A similar observation was reported for an equivalent material exposed to pure CO2 hydrogenation conditions and was associated with the characteristic removal of lattice oxygen upon catalyst reduction [20,23]. Concerning Ni/In2O3, no reduction of Ni2+ species to metallic nickel was evidenced, which is expected since this metal is more easily oxidized upon exposure to air (Fig. S4b) [46,47]. Interestingly, In2O3 coverage by nickel significantly decreased (i.e., 4.1 to 1.4 at.%) in catalysts retrieved after the full cycle (Table S3), which is in line with STEM-HAADF-EDX results (Fig. 5d) showing major sintering of the nickel phase into large particles.The O 1s core-level spectra (Figs. 5d\u2013f and S4c) were analyzed to access information about the amount of oxygen vacancies, determining the relative contribution of the distinct oxygen species (Table S3). The signals related to oxygen atoms next to a defect (Odefect) were ascribed to the presence of vacancies, as previously reported [8,15,20,48]. While bulk In2O3 (Fig. 5d), Pd-In2O3 (Fig. 5f), and Ni/In2O3 (Fig. S4c) are associated with a slight modification in the content of Odefect, a significant alteration was evidenced for In2O3/m-ZrO2 (from 14 to 22 % and 20 % after the half- and reverse half-cycles, respectively, Fig. 5e). Although the relative contribution of Odefect is quite similar for all catalysts retrieved after cycle experiments, a recent report showed that the increased density of oxygen vacancies for In2O3/m-ZrO2 relates to the formation of additional In-Vo-Zr centers rather than Vo sites typically generated on bulk In2O3, highlighting that the former better activates CO2 and H2 to produce methanol [48]. This claim is supported by the identification of indium cations directly bound to the m-ZrO2 lattice in this study, and the possible superiority of such vacancies was also postulated for In2O3 epitaxially-grown on m-ZrO2 [15,16]. These findings indicate that CO has a moderate beneficial effect on promoted and unpromoted indium oxide upon moving from R = 0 to 0.5, fully obscured by detrimental phenomena, and a greater positive role on In2O3/m-ZrO2, which is not counterbalanced by negative structural changes.The main knowledge gathered by characterization is graphically combined in Fig. 6\n to provide a comprehensive rationalization of the (de)activation mechanisms acting on most relevant In2O3-based catalysts in methanol synthesis from hybrid feeds. Notably, In2O3/m-ZrO2 showcases an increased density of distinctive and superior oxygen vacancies that along with its resistance to sintering explain the improved performance in CO-rich feeds. Unlike activation, deactivation mechanisms are vaster and intricate. Sintering of In2O3 induced by water and/ or over-reduction by CO is a common denominator for all deactivated systems, but can be associated with other factors resulting in greater depletion in methanol productivity. This is clearly evidenced for the metal-promoted systems. Ni/In2O3 irreversibly deactivates also due to sintering of the nickel phase. For Pd-In2O3, CO inhibition of H2 splitting over the promoter due to strong adsorption is put forward as another factor contributing to the drop in methanol STY. Nonetheless, such kinetic effect is reversible and can be modulated by tuning the CO concentration in the feed.It is worth mentioning that state-of-the-art commercial and lab-prepared Cu-ZnO/Al2O3 systems for CO-based methanol synthesis have also been tested in the hydrogenation of hybrid CO2\u2212CO feeds in steady-state under a fixed feed composition and in cycles where the CO/CO2 ratio was varied, providing significantly diverse results. Similar to In2O3-based systems, some catalysts experienced a boost in methanol productivity, while others displayed diminished performance [49\u201351]. These divergent behaviors are traced back to distinct catalyst compositions and preparation method as well as the presence of additives in commercial samples and indicate that, unlike In2O3-based catalysts, there is no consensus on a top-performer applied to CO2-containing feeds.Based on the findings of the full cycles, In2O3, In2O3/m-ZrO2, Pd-In2O3, and Ni/In2O3 were selected for further testing. To gain insights into the catalytic performance at R = 0.5 without prior exposure to any other environment, reverse CO2\u2212CO half-cycles (rHC), i.e., from R = 0.5 to 0, were carried out (Fig. 3c). While similar trends to the backward branches of the full cycles developed for all systems with a decline of methanol STY for In2O3, In2O3/m-ZrO2, and Ni/In2O3 and an improvement of the same for Pd-In2O3 (Fig. 3a), differences in the absolute methanol productivity values were observed at the start of these experiments. For both bulk In2O3 and In2O3/m-ZrO2, methanol STY at R = 0.5 was higher than under the same condition in the full cycles, suggesting that water-induced sintering in the CO2 hydrogenation environment counteracts the positive effect of CO (Fig. 3a). For Ni/In2O3 and, especially, Pd-In2O3, starting the reaction at R = 0.5 was detrimental to the methanol STY. For all catalysts, the methanol productivity at R = 0 was comparable to that determined at the same ratio at the end of the full cycles, except for the palladium-promoted sample, pointing to a plausible greater role of CO than water on structural modifications of the latter. In line with these findings, characterization by N2 sorption, XRD, and XPS of the catalysts after the reverse half-cycle provided equivalent results to those described for the same materials after the whole full cycle (Tables 1 and S3, and Figs. S3a,b, S4a\u2013c, and 5 c\u2013f).To further examine the impact of CO, the effect of temperature and pressure was studied at R = 0.2 since this is the ratio expected in the reactor at equilibrium when recycling CO (vide infra) and corresponds to the upper concentration boundary of CO present as a feed impurity. The methanol productivity increased as a function of pressure within 3\u20135.5 MPa and temperature between 543\u2212583 K for all systems (Fig. S5), which is in good agreement with their performance trends under CO-free CO2 hydrogenation conditions. This hints that these operational parameters do not have a critical influence on either beneficial or detrimental CO effects in the practically-relevant ranges investigated.Since In2O3/m-ZrO2 displayed an improved performance upon the introduction of CO in the feed and Pd-In2O3 still produced a significant amount of methanol even when experiencing deactivation by CO, additional experiments were devised to assess how their methanol productivity could be individually optimized. To predict the most representative R ratio to which the catalysts shall be exposed upon per-pass operation, calculations were conducted using the Aspen Plus software. Firstly, the thermodynamic equilibrium composition (Fig. S2a) of carbon\u2010based species during methanol synthesis was determined under typically applied reaction conditions (T = 533\u2013593 K, P = 5 MPa, H2/CO2 = 4, and WHSV = 24,000 cm3 h\u22121 gcat\n\u22121), which shows that R equals 0.20 at 553 K in a single-pass conversion process. To quantify the benefit of recycling a gas mixture with this composition on methanol productivity at the latter temperature, per-pass and overall conversion levels were estimated based on a simplified flowsheet comprising the methanol synthesis reactor, a flash separator to separate water and methanol from the gases, and a recycle loop for the latter (Fig. S2b) using molar flows of relevant reaction streams as detailed in Table S1. Values of 21 and 67 % were attained for the two parameters, resp6ctively, highlighting that per-pass operation would triplicate conversion in CO2-based methanol synthesis. A similar R value to that obtained for single-pass conversion (R = 0.17 vs. 0.20, Fig. S2a) was determined upon recycling of the outlet gases. Accordingly, 0.2 was applied in the testing of In2O3/m-ZrO2 and Pd-In2O3.Concerning In2O3/m-ZrO2, a pre-treatment in syngas (H2/CO = 2) was carried out for 4 h at 553 K prior to using the catalyst in CO\nx\n hydrogenation at R = 0, 0.2, and 0.5 expecting that such activation protocol would lead to a more defective and thus better performing material. Fig. 7\na shows an enhanced methanol productivity, which is sustained when applying feeds containing CO (R = 0.2 and 0.5) but not using a stream with only CO2 and H2 (R = 0), emphasizing that the catalyst surface dynamically responds to the reducing potential of the feed mixture. Although methanol STY at R = 0.5 is similar to that recorded at R = 0.2, the actual yield with respect to CO2 shall be effectively higher under the former conditions, which shall trace back to the hindering of the RWGS reaction and a better preservation of oxygen vacancies upon having more CO in the feed. No variations in the surface area were recorded (Table S4), indicating that the syngas treatment did not significantly or irreversibly alter the catalyst structure. O 1s core-level XPS spectra indicate that the amount of surface oxygen vacancies present on In2O3/m-ZrO2 and pure m-ZrO2 after activation with syngas does not change during the reactions (Fig. S6). Some depletion would be expected for the exposure to R = 0, but is it likely masked by the large contribution of zirconia to the oxygen signal. The catalytic data in Fig. 7a also reveal that In2O3/m-ZrO2 can hydrogenate CO to methanol in pure syngas conditions, in contrast to bulk In2O3, which fully reduces and melts [8,11]. In line with this, an extended full-cycle experiment spanning from pure CO2 (R = 0) to pure CO (R = 1) hydrogenation (Fig. S7) displayed that CO is converted to methanol over In2O3/m-ZrO2 at H2/CO = 4. This points to a possible utilization of CO as a carbon source for methanol along with CO2 in hybrid feeds. These findings are in line with m-ZrO2 showing high CO adsorption capacity (i.e., 5- to 10-fold higher than for t-ZrO2) and CO activation ability [52\u201355]. The extended cycle evidenced that increasing CO concentration in the feed to R > 0.5 is irrelevant or detrimental to methanol productivity and that the catalyst performance is readily recovered upon restoring CO2 in the feed. Thus, over-reduction of In2O3 to metallic indium is less likely to be the reason for deactivation in this supported system. It is plausible that decreased methanol productivity at R > 0.5 traces back to the lower methanol formation rate upon CO hydrogenation (r = 2.9 mmolMeOH h\u22121 gcat\n\u22121) as compared to CO2 hydrogenation (r = 7.5 mmolMeOH h\u22121 gcat\n\u22121).To evaluate stability under variable conditions, In2O3/m-ZrO2 was subjected to a long-term cycling test comprising three stages and a total time on stream of 120 h (Fig. 7b). During an initial 40 h at R = 0, the system exhibits a stable performance, which is improved with no induction time upon changing feed composition to 0.2 and sustained for an additional 40 h. The latter regime mimics the situation in which the outlet gas stream purified from condensable species is fed back into the reactor upon per-pass operation. The final switch back to pure CO2 hydrogenation occurs sharply as well, and the catalytic behavior is equivalent to the first phase. These findings emphasize that In2O3/m-ZrO2 can effectively undergo reversible cycling and is amenable for practically-relevant operation.Regarding Pd-In2O3, the first set of tests encompassed three full cycles reaching lower R values than in the case described above, to unravel if the methanol productivity could be better recovered when returning to CO2 hydrogenation from hybrid feeds poorer in CO. In such threshold experiments the highest R was set at 0.1, 0.2 or 0.3. As expected, Pd-In2O3 deactivated almost linearly upon introduction of CO in all cases (Fig. 8\na). The recovery in methanol productivity was moderately higher than in the full cycle reaching R = 0.5 (ca. 90 vs. 85 %) and similar regardless of the R value reached in the forward branch of the cycles (vide the backward branches in Fig. 8a).The second experiment comprised the same long-term cycling test as described for In2O3/m-ZrO2. At R = 0 during 40 h, Pd-In2O3 displayed an equilibration phase after which the methanol STY levelled off (Fig. 8b), in line with a previous report [20]. A drop in methanol productivity (from 0.65 to 0.56 gMeOH h\u22121 gcat\n\u22121) was then observed at R = 0.2, which was partially recovered upon returning to the initial condition of R = 0 but with an induction time. Indeed, the same catalyst would be expected to display stable behavior after 93 h on stream under pure CO2 hydrogenation conditions [20]. These findings indicate that the irreversible restructuring of the catalyst described above occurs to a consistent extent also in feeds less rich in CO. Indeed, the sample retrieved after the threshold experiment at R = 0.1 featured a strongly decreased surface area (i.e., 133 to 37 m2 g\u22121, Table S4) indicating substantial In2O3 sintering. However, since the surface area of bulk In2O3 was reported to decrease from 113 to 38 m2 g\u22121 upon reaction with CO2 hydrogenation for a comparable time on stream [8], such structural change is most likely due to the water byproduct rather than only to the CO co-fed. Overall, the results indicate that Pd-In2O3 could not be regenerated in swing operation. However, they hint that this catalyst could show steady, although reduced, performance in a per-pass process for a longer time than probed in this test.The impact of CO on CO2-based methanol synthesis over bulk, supported, and promoted In2O3 catalysts was systematically assessed by a set of purposely devised cycle experiments. This approach revealed diverse catalytic responses despite all catalytic systems share the same active phase. Upon CO addition until reaching an equimolar CO\u2212CO2 ratio, In2O3/m-ZrO2 experienced an increase in methanol productivity, bulk In2O3 remained almost unaffected, and moderate-to-substantial deactivation was determined for nickel- or palladium-promoted In2O3, In2O3/t-ZrO2, and In2O3/Al2O3. Detailed characterization uncovered that a variable interplay between formation of surplus oxygen vacancies, sintering of In2O3 or metal promoters induced by CO and/or water, and strong adsorption of CO on palladium underpins the observed behaviors. Focusing on the more relevant In2O3/m-ZrO2 and Pd-In2O3, operational regimes were probed to identify conditions to maximize their performance in a practical process. The methanol productivity of In2O3/m-ZrO2 in hybrid feeds was boosted by pre-activation of the catalyst in syngas and exposure of Pd-In2O3 to CO-poorer feeds (R < 0.3) enabled a better recovery of methanol STY upon restoring pure CO2 hydrogenation conditions. Moreover, the two catalysts could adapt rapidly to a switch from CO2 hydrogenation to processing of a feed with a CO/CO2 ratio as dictated by thermodynamics upon a per-pass regime and sustain for a longer period their activated and moderately depleted states, respectively. Overall, this study uncovered the positive and negative impacts of hybrid CO2-CO feeds on CO2-based methanol synthesis on a relevant class of catalysts, offering a novel approach to assess performance under practically-relevant conditions, which shall serve as a stepping-stone to accelerate the development of industrially-viable catalytic technologies for this reaction and other CO2-based conversions.\nThaylan P. Ara\u00fajo: Investigation, Data curation, Methodology, Visualization, Writing - original draft. Arjun Shah: Investigation, Software. Cecilia Mondelli: Supervision, Visualization, Writing - review & editing. Joseph A. Stewart: Project administration, Writing - review & editing. Daniel Curulla Ferr\u00e9: Project administration. Javier P\u00e9rez-Ram\u00edrez: Conceptualization, Funding acquisition, Writing - review & editing.The authors declare that they have no competing interests.Dr. S. Mitchell is acknowledged for the electron microscopy measurements, and the Scientific Center for Optical and Electron Microscopy (ScopeM) at the ETH Zurich for the use of their facilities. We are grateful to S. B\u00fcchele for performing the XPS analyses and to J. L\u00fcthi for designing the 3D models of the catalysts.Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.apcatb.2021.119878.The following is Supplementary data to this article:\n\n\n\n\n", "descript": "\n                  Catalysts for CO2-to-methanol are typically evaluated in a single-pass regime using pure CO2streams. In a practical process however, CO shall be present as a feed impurity or as a recycled byproduct. Herein, the sensitivity to CO was evaluated on In2O3 catalysts in bulk, supported, or metal-promoted forms, using cycle experiments with variable CO2 and CO contents at H2/(CO + CO2) = 4. The methanol productivity was decreased (\u221220-\u221240 %) on all catalysts except In2O3/monoclinic-ZrO2, the activity of which was boosted by 10 %. In-depth characterization of the catalysts uncovered controlled formation of oxygen vacancies and resistance to sintering as the main reasons for the activation of the latter and an interplay of CO/H2O-induced sintering and CO inhibition as the origin of performance loss. Focusing on the most representative systems, operation protocols were explored to maximize their methanol yield. We emphasize that assessment with hybrid CO2-CO feeds is key for the design of industrially-viable catalysts for sustainable methanol production.\n               "}
{"full_text": "With the increasing demand for energy supply and the ever-worsening climate change, the development of sustainable technologies, such as the solar-driving catalysis process, to generate clean energy and chemicals has been recognized as the most promising way to address this pressing global issue. Photocatalysis has attracted significant attention due to its sustainable characteristics and limited environmental impacts (Lei et\u00a0al., 2021; Zhu and Zhou, 2019). To date, many semiconductor materials, such as ZnO (Liu et\u00a0al., 2019b; Shekofteh-Gohari et\u00a0al., 2018), TiO2 (Gao et\u00a0al., 2020b; Guo et\u00a0al., 2019), and C3N4 (Patnaik et\u00a0al., 2021; Zhu et\u00a0al., 2021), have been demonstrated with photocatalytic activity. However, as pristine semiconductor materials, their photocatalytic performance is still below the common expectation due to their unsatisfactory photocatalytic processes (i.e., limited light-harvesting capability, severe recombination of photo-induced charge carriers, and poor catalytic selectivity and activity) (Gao et\u00a0al., 2019a; Zhou et\u00a0al., 2018). Therefore, it is of great importance to developing novel photocatalysts with an efficient photocatalytic process to perform high-efficient solar-driven chemicals/fuel production.Recently, single-atom catalysts (SACs), with the active metal species existing as isolated single-atoms (SAs) and stabilized by bonding with the substrate materials or by alloying with another metal (Lai et\u00a0al., 2019; Zang et\u00a0al., 2020), have attracted significant attention in the field of photocatalysis due to the unique advantages of SAs, such as maximal metal dispersion and tuneable local coordination environments, exhibiting excellent reaction activity. For the supported metal nanoparticles, it has been widely recognized that their catalytic performance (product selectivity, reaction activity, and the stability of the catalysts, etc.) is strongly influenced by the metal particle size and the local coordination environment between the metal particle and the supporting materials (Campbell et\u00a0al., 2002; Liu, 2017; Rao et\u00a0al., 2008; Subramanian et\u00a0al., 2004). With the particle size decreasing, the surface atoms ratio to the total atoms will sharply increase and result in more unsaturated coordinated surficial metal atoms (Liu, 2017), which could act as the active sites for the reactant molecules' adsorption. When decreased to the small nano-size (2\u201310\u00a0nm) (Halperin, 1986; Li et\u00a0al., 2021a), the electron energy level of the metal species will split into discrete energy levels (Yang et\u00a0al., 2015; Yang et\u00a0al., 2013), which will directly influence the orbital hybridization and charge transfer at the metal/reactant interface (Wang and Lu, 2020).Generally, SACs possess five main compelling advantages: (1) maximized metal dispersion (Zhang et\u00a0al., 2021b): as a result, with the same metal loading amount, SACs possess more reactive sites; (2) better product selectivity: all the single-atom sites possess the similar composition and coordination structure, resulting in a uniform catalytic environment (Chen et\u00a0al., 2021); (3) bridged homogeneous and heterogeneous catalysts (Chen et\u00a0al., 2018a; Gao et\u00a0al., 2018): SACs have atomically dispersed metal sites on solid supports and consist of well-defined mononuclear active centers, expected to combine homogeneous and heterogeneous catalysts; (4) unsaturated coordination structure or valence state: the unsaturated coordinated structure or valence state will make the SAs act as the reaction active center and directly participate in the reaction (Chen et\u00a0al., 2018b); (5) size effect (Hu et\u00a0al., 2014): the surface free energy of the SAs obviously increases compared with the metal nanoparticles, making them highly active in chemical interactions for reactant molecules (Yang et\u00a0al., 2013).SACs have been at the forefront of catalysis research due to their maximized atom utilization, unique structures, and properties. Among many applications, electrochemical energy conversion is one of the most promising areas, including oxygen evolution (OER), CO2 reduction (CRR), etc (Daiyan et\u00a0al., 2020a). Compared with traditional catalysts, SACs expose more active sites, leading to enhanced electrocatalytic activity. In addition, SACs are also demonstrated with improved selectivity toward the target products. Taking CRR as an example, it was demonstrated that the coordination structure of the SACs was beneficial for reducing the formation energy barrier of the CRR intermediates (e.g., \u2217COOH), thereby improving the product selectivity (Daiyan et\u00a0al., 2020b; Leverett et\u00a0al., 2021; Leverett et\u00a0al., 2022). In addition, SACs have also been applied and studied in the photocatalysis fields. It is found that by constructing single-atom photocatalysts (SAPs) with rationally introduced SAs, the overall photocatalysis process could be significantly impacted (Figure\u00a01\n) (Xia et\u00a0al., 2021). For instance, it is demonstrated that the introduction of SAs can efficiently alter the electronic structure of the supporting semiconductors, resulting in tuneable optical response behavior (Gao et\u00a0al., 2020a), and the enhanced charge transfer kinetics of the SAPs can be ascribed to the unique band bending effect (Su et\u00a0al., 2018). Moreover, due to the tuneable coordination structure between the SAs and the supporting materials, the surface adsorption and activation of the reactant molecules can be boosted (Yang et\u00a0al., 2020). In this regard, SAPs provide a great platform to regulate the overall photocatalytic reaction process. However, it is also worth to be pointed out that the practical application of SAPs is limited due to the following reasons: (1) the agglomeration of SAs; (2) the structure-performance relationship between the SAPs and the surface photocatalytic reaction is unclear, resulting in vague SAPs-based reaction mechanisms. Therefore, it is of great importance to systematically summarize the current research progress of SAPs for the guidance of future studies on SAPs\u2019 synthesis.In this review, the unique phenomenon of SAPs brought to the overall photocatalysis processes (i.e., optical response, the separation, and surface reaction process) will be systematically overviewed. As such, the local coordination environment between SAs and the substrates materials will be deeply discussed, and the sequent impact on products' selectivity will be explained. At last, the current challenge and potential research points of SAPs are pointed out. We believe this review will provide guidance knowledge for the future development of novel SAPs catalysts for photocatalytic reaction.Photocatalysis is regarded as one sustainable method for solar-driven chemicals/fuels generation. Therefore, within a typical photocatalytic reaction process, it contains three steps to obtain the final redox products, including, (1) the photoexcitation of the photocatalysts and the generation of photo-induced carriers; (2) the spatial separation of photo-induced electrons and holes; (3) the redox of surface-adsorbed reactant molecules. As such, the photocatalytic performance of the catalysts is closely related to the aforementioned three steps: (a)\u00a0for some wide bandgap semiconductors, such as TiO2, ZnO, etc., the narrow optical response performance leads to relatively poor light-harvesting capability and limited amounts of photo-induced carriers; (2) during the spatial separation process, severe carrier recombination processes (both surface recombination and volume recombination) are commonly accompanied, resulting in extra carrier recombination loss; (3) the uncontrollable surface adsorption and activation of the reactant molecule lead to poor product selectivity (Figure\u00a02\n). Therefore, rationally regulating the aforementioned three steps is of great importance to obtain high-performance photocatalytic processes.In addition, from the perspective of reaction types, the photocatalysis techniques can be applied to perform various types of solar-driven chemicals/fuels production reactions, including (Wang et\u00a0al., 2021d), CO2 reduction (Tang et\u00a0al., 2022), N2 reduction (Guan et\u00a0al., 2021), NOx reduction (Zhang et\u00a0al., 2021d), biomass oxidation reforming (Wang et\u00a0al., 2021c), and pollutant oxidation degradation (Bui et\u00a0al., 2021). By virtue of photocatalysis, value-added chemicals and green fuels can be generated in a sustainable and environmental-friendly way. As such, it requires sound development of novel catalysts with favorable adsorption and dissociation modes toward the specific molecules. For instance, in the past decades, noble metal/semiconductor catalysts have been reported as one of the most efficient catalysts for the selective oxidation of aromatics (Chen et\u00a0al., 2017a). However, the high cost of the catalysts severely limit the scale-up application of such kind of catalysts. In this regard, it makes the development of SAPs of great practical significance. Due to the insufficient understanding of the influence caused by the introduction of SAs on the photocatalysis process, the study on the application of SAPs for the solar-driven chemicals/fuels production is still in its infancy, calling for urgent investigation.When the particle size is reduced to atom level, the metal single atoms with ultra-high surface energy tend to aggregate and form nanoclusters during the synthesis process (Zhao et\u00a0al., 2020; Zhu et\u00a0al., 2020). Currently, the reports on SAPs are still limited. Fortunately, the widely reported SACs for other catalysis reactions, such as thermal catalysis, electrocatalysis, etc., can bring us some insights. To date, the synthesis methods of SACs are plentiful, such as impregnation, co-precipitation, chemical vapor deposition (CVD), ion exchange, galvanic replacement, thermochemical method (flame spray pyrolysis [FSP], and pyrolysis of organic materials [metal-organic framework, covalent organic framework, etc.]), atomic layer deposition (ALD), atom trapping, and photochemical reduction. Taking the impregnation method as an example, the prepared substrate material is put into a solution containing SAs precursor. The metal ions are adsorbed on the surface of the substrate material, which are then reduced to produce SACs (Wang et\u00a0al., 2018). The mass loading of metal SAs for the SACs prepared by the impregnation method is very low, but the procedure is simple (Xia et\u00a0al., 2022). In addition, the co-precipitation method is also widely used for the preparation of SACs. At least use two cations to form a homogeneous phase in the solution (Qiao et\u00a0al., 2011). For SACs prepared by co-precipitation, the mass loading must be kept below 1% to prevent agglomeration of the metal particles during calcination or reduction (Xia et\u00a0al., 2022).Besides the method mentioned earlier, the following four strategies will be further discussed and compared, including ALD, atom trapping, thermochemical method, and photochemical reduction methods, which are proven to be effective in preventing metal SAs sintering (Figure\u00a03\n).Atomic layer deposition (ALD) is demonstrated to be a precise method that can deposit the SAs on the surface of the supporting materials by alternately exposing the supports to pulsed vapors of various precursors. As the method is based on chemisorption, deposition occurs only in areas with reactive surface sites (Fonseca and Lu, 2021). Generally, the ALD method includes four steps: (1) exposure to the first precursor; (2) purge of the reaction chamber; (3) exposure to the second reactant precursor; and (4) a further purge of the reaction chamber (Cheng et\u00a0al., 2016). The morphology, size, density, and loading of the deposited material on the carrier can be precisely controlled by simply tuning the ALD cycle (Cheng et\u00a0al., 2019). For instance, Cao et\u00a0al. successfully applied ALD technology and synthesized Co-based catalysts (Co1/PCN), which were used in the photocatalytic hydrogen evolution (Figure\u00a04A) (Cao et\u00a0al., 2017). Specifically, cyclopentadienyl cobalt (Co(Cp)2) was used as the cobalt-precursor for the ALD, which was then treated with O3 to remove the cyclopentadienyl ligand, resulting in the Co1-N4 structured SACs. However, nucleation delay and island growth are considered the key issues that need to be optimized (Cheng and Sun, 2017), resulting from the limited functional groups on the surface of supporting materials. Besides, the surface energy difference of metal and support will also bring a challenge to the deposition process. When the surface energy of the support is lower than the free energy of deposited metal, the support cannot be wetted by the deposited metal, which will lead to insufficient adsorption sites and result in an island growth mode finally (Cheng and Sun, 2017).The atom trapping method is another effective strategy to synthesize the SACs with stably anchored SAs (Wang et\u00a0al., 2018). Specifically, to plant SAs in the commonly used semiconductors materials, constructing surface defect sites (i.e., C-defect (Chen et\u00a0al., 2020c; Liu et\u00a0al., 2021a), O-defect (Cai et\u00a0al., 2020; Wang et\u00a0al., 2021e), N-defect (Qin et\u00a0al., 2021; Zhang et\u00a0al., 2022a), S-defect (Wang et\u00a0al., 2020b; Zhang et\u00a0al., 2021e), metal-defect (Wu et\u00a0al., 2021; Zhang et\u00a0al., 2018), etc.) can efficiently capture the SAs. It means that, through defect-engineering strategies, defect sites can be created on the surface of the supporting materials, leading unsaturated coordination environments to the adjacent atoms (Liu et\u00a0al., 2021a), which can be used as the anchors to trap and stabilize SAs. The necessary synthesis condition of atom trapping requires the mobile metal species and trapping sites on the supports (Qu et\u00a0al., 2018). Thus, by adjusting the concentration of the surface defects, the SAs loading amount and the catalytic performance of the SACs can be easily regulated (Wang et\u00a0al., 2019a). For instance, Jones et\u00a0al. demonstrated that the Pt SAs could be captured and anchored in CeO2, forming atomically dispersed Pt1/CeO2 SACs (Figure\u00a04B) (Jones et\u00a0al., 2016). In this process, Pt nanoparticles supported on Al2O3 were aged in the air at 800\u00b0C, and the PtO2 was released and captured by CeO2 in a high-temperature environment, which exhibited quite good thermal stability.Pyrolysis and flame spray pyrolysis (FPS) are two of the current main thermochemical methods to prepare SACs. The major difference between these two strategies lies in the treatment atmosphere, as the general pyrolysis process calls for inert atmosphere annealing, whereas the FPS process can be processed in air condition (Wang et\u00a0al., 2012). For pyrolysis, a common method is to adsorb the metal complexes with N-ligand onto the porous support, followed by a pyrolysis step of the metal-organic framework, covalent organic framework (Liu et\u00a0al., 2019a). For example, it is widely reported that the Co and Fe atoms can coordinate with the N from the support materials to form the Co-N and Fe-N atomic dispersion structure (Wang et\u00a0al., 2019d; Wu et\u00a0al., 2019a; Zhang et\u00a0al., 2021c; Zhu et\u00a0al., 2017). To synthesize such SACs, generally, the nitrogen/metal precursors and the carbon supports will be pyrolyzed and carbonized under inert atmospheres at a quite high temperature. In addition, Wei et\u00a0al. also demonstrated that the nanoparticles of noble metals (e.g., Pd, Pt, Au) can be converted to thermal-stable SAs with an annealing temperature of 900\u00b0C in the inert atmosphere (Figure\u00a04C) (Wei et\u00a0al., 2018). With the continuous study on the pyrolysis method, it is confirmed that controlling the temperature change during the pyrolysis process is crucial for the evolution of metal SAs (Wang et\u00a0al., 2020a).Unlike pyrolysis, the flame spray pyrolysis method allows the SACs\u2019 synthesis to be carried out in the air conditions, making it a flexible way to produce catalysts with controllable morphology and particle size by adjusting the synthesis conditions (Ding et\u00a0al., 2021). During the combustion process, the metal salt solution and the solvent of the supporting materials are sprayed into the high-temperature flame simultaneously. The solvent will be vaporized and the metal salt solution will be burned or hydrolyzed in the high-temperature flame, allowing the SAs to be anchored on the supporting materials in one step (Michalow-Mauke et\u00a0al., 2015; Pongthawornsakun et\u00a0al., 2015; Thybo et\u00a0al., 2004). As shown in Figure\u00a04D, Ding et\u00a0al. reported that, by virtue of the flame spray pyrolysis method, the Pt SAs can be introduced to a series of oxide supports with excellent stability (Ding et\u00a0al., 2021). Compared with other methods, the flame spray pyrolysis strategy shows unique advantages for the catalysts\u2019 preparation: the flame spray pyrolysis method can synthesize the catalyst quickly and is suitable for scaling up; the SACs produced by the flame spray pyrolysis strategy can exhibit a quite small particle size and well-dispersion (Gavrilovi\u0107 et\u00a0al., 2018).Photochemical reduction method is a widely used postpreparation strategy to synthesize SACs. This method generally requires that the support materials and the metal salt precursor should be mixed within a reducing agent solution at first. Then the reducing agent will release free radicals under the irradiation of the UV light, reducing and anchoring the SAs to the supporting materials (Li et\u00a0al., 2018). This method is easy to tailor the SAs loading amount. For instance, by the photochemical strategy, Liu et\u00a0al. successfully planted the Pd SAs into ultra-thin TiO2 nanosheets (Pd1/TiO2) (Liu et\u00a0al., 2016), as shown in Figure\u00a04E. The TiO2 powder was first dispersed in ethylene glycol (EG) by ultrasonic. Then, the H2PdCl4 was added as the metal precursor. Under mild UV conditions, the ethylene glycol free radicals can be formed on the surface of the TiO2 nanosheets, which promoted the Cl\u2212 releasing in the Pd precursor solution and the Pd-O bond formation, resulting in the formation of Pd1/TiO2. The as-synthesized Pd-TiO2 catalysts exhibited excellent catalytic activity for the hydrogenation of C=C bonds and C=O bonds, which was nine times higher than that of commercial Pd catalysts. To further prevent the agglomeration of SAs, Wei et\u00a0al. demonstrated a novel ice-photochemical reduction method to synthesize Pt SAs (Figure\u00a04F), which were confined and dispersed in the crystal lattice of the supporting materials (Wei et\u00a0al., 2017). The Pt SAs were obtained by exposing the frozen H2PtCl6 solution under UV radiation. Then, the molten frozen solution containing Pt atoms was physically mixed with various supports (e.g., TiO2) to synthesize Pt-based SACs. Due to the low-temperature feature of the iced-photochemical reduction strategy, it can further avoid the agglomeration of the SAs, which is unavoidable in the room temperature photochemical reduction processes (Lu et\u00a0al., 2020; Wei et\u00a0al., 2019). Due to the reduced diffusion rate caused by the low-temperature condition, a hindrance to the agglomeration of SAs thus will be caused, leading to better atomic dispersion of SAs (Lu et\u00a0al., 2020).Current research on the SAPs have demonstrated that the introduction of SAs will influence the overall three-step photocatalysis process (i.e., photo-excitation, carrier transfer, and surface reaction). Therefore, the related work will be comprehensively overviewed and compared in this section.The optical response capability of the semiconductor catalyst could directly influence the photocatalytic performance, as narrow light-harvesting behaviors will severely limit the incident light utilization efficiency of the catalysts. Previous work demonstrated that the introduction of SAs is effective to enlarge the optical response performance of the pristine catalysts. Some work claimed that the introduction of SAs will cause impurity energy level, thereby broadening the optical response range of the pristine catalysts (Figures\u00a05A\u20135B) (Yang et\u00a0al., 2016). In addition, some work also claimed that the introduction of SAs could directly reduce the bandgap of the semiconductor catalysts, rather than introducing impurity energy level, leading to enhanced light absorption capability of the SAPs (Figure\u00a05C) (Yang et\u00a0al., 2021).For instance, Jin et\u00a0al. proposed SAPs with Fe SAs implanted into the surface of Bi4O5I2 (Bi4O5I2-Fe30) (Jin et\u00a0al., 2021). As such, Fe SAs were considered as the dopant to replace the Biatoms in Bi4O5I2, which generated impurity energy levels within the bandgap of Bi4O5I2, leading to a broadened light-harvesting range toward the Bi4O5I2-Fe30 SAPs. From the ultraviolet-visible (UV-Vis) diffuse reflectance spectra, a red-shifting light absorption trend was demonstrated when the Fe SAs were introduced to the Bi4O5I2 (Figure\u00a06A). The Tauc plot diagram in Figure\u00a06B also confirmed that the bandgap values reduced from 2.17 eV of Bi4O5I2 to 1.56 eV of Bi4O5I2-Fe30. The aforementioned evidence indicated that the light absorption and electronic structure of pristine semiconductors can be modulated by incorporating SAs. By virtue of both experimental (Figures\u00a06C and 6D) and theoretical calculation (Figures\u00a06E and 6F), the energy band structure of Bi4O5I2 and Bi4O5I2-Fe30 were further determined. It was confirmed that, after the introduction of Fe SAs, impurity energy levels appeared near the conduction band of Bi4O5I2, therefore enlarging the optical response range of Bi4O5I2. Li et\u00a0al. also demonstrated Ru SAs doped monolayered TiO2 nanosheets (Ru1/TiNS) and confirmed that after Ru1 SAs inducing, an isolated impurity energy level was formed, broadening the optical absorption range up of the Ru1/TiNS to 700\u00a0nm (Li et\u00a0al., 2020b).Moreover, it was reported that the introduction of SAs could directly reduce the bandgap of the catalysts, leading to an enlarged light absorption range. For instance, Jiang et\u00a0al. developed a novel Ag SAs incorporated carbon nitride photocatalyst (Ag-N2C2/CN) with the Ag-N2C2 configuration (Jiang et\u00a0al., 2020). As shown in Figure\u00a06G, with the Ag-N2C2 and Ag-N4 coordination structure, the optical response range of the obtained catalysts can be efficiently expanded. The inset in Figure\u00a06G showed a reduced narrowed bandgap after incorporating Ag SAs. The DFT calculation also confirmed the introduction of Ag SAs would lead to the directly reduced bandgap but no impurity energy level (Figures\u00a06H and 6I).The efficient spatial separation and transfer of photogenerated carriers are important for performing high-efficient photocatalytic reactions. But for pristine semiconductor catalysts, severe carrier recombination is unavoidable (Liu et\u00a0al., 2020). This unavoidable energy loss before the surface reaction would result in relatively low photocatalytic activity. The introduction of SAs is also demonstrated to contribute to efficient carrier separation (Dong et\u00a0al., 2021b; Xiao et\u00a0al., 2020). According to the previous studies, generally, when the metal nanoparticle is in contact with the semiconductor, a Schottky barrier is generated on behalf of the energy level matching of the metal and the semiconductor. After the electronic equilibrium is established at the interface, the photogenerated electrons in the semiconductor will pass through the Schottky barrier and transfer to the metal particles (Gao et\u00a0al., 2020a). This charge transfer behavior across the metal-semiconductor interface can be inherited by the SAPs (Gao et\u00a0al., 2020a; Gopalakrishnan et\u00a0al., 2021). It was demonstrated that, when the metal SAs were contacted with the semiconductor substrate, efficient charge transfer behaviors could be expected due to the interfacial barrier (Meng et\u00a0al., 2019; Wang et\u00a0al., 2019a). In addition, it was considered that, by introducing SAs, the charger transfer distance between the light-harvesting units and the photocatalytic sites could be shortened (Gao et\u00a0al., 2020a).For instance, Li et\u00a0al. demonstrated that, by incorporating Pt SAs into ultra-thin covalent triazine framework nanosheets (Pt-SACs/CTF) with unique Pt-N3 structure, the carrier migration in the Pt-SACs/CTF catalysts can be efficiently enhanced (Li et\u00a0al., 2020a). It was demonstrated that the electrons captured by the Pt SAs were subsequently utilized for the nitrogen fixation reaction. In sharp contrast with the ultra-thin covalent triazine framework nanosheets (CTF-PDDA-TPDH), the Pt-SACs/CTF catalysts showed a weaker PL intensity (Figure\u00a07A), indicating the suppressed carrier recombination in Pt-SACs/CTF. The photoresponse of the Pt-SACs/CTF and CTF-PDDA-TPDH catalysts were further studied by the chronoamperometry curves under chopped optical illumination (Figure\u00a07B). The CTF-PDDA-TPDH nanosheet catalysts showed lower photocurrent density, whereas the photocurrent density increased obviously after incorporating Pt SAs. These results elucidated the enhanced carrier separation efficiency of the Pt-SACs/CTF catalyst. The proposed carrier transfer mode was shown in Figure\u00a07C. Under visible light irradiation, the CTF-PDDA-TPDH nanosheets were excited with the electrons on the CB of the CTF-PDDA-TPDH nanosheets transferred to the Pt SAs, bringing about the enhanced carrier separation. The electrons captured by the Pt SAs were then consumed to reduce the N2 molecule into NH3.Similarly, Dong et\u00a0al. demonstrated that the Pt SAs could be anchored on the covalent organic framework (COF) catalysts, linked by \u03b2-ketoenamine (Pt1@TpPa-1) (Dong et\u00a0al., 2021a). As-synthesized photocatalysts showed high activity (99.86\u00a0mmol gpt\n\u22121 h\u22121) and selectivity (100%) for H2 evolution, which were attributed to the successful anchoring of Pt SAs to facilitate the transfer efficiency of photogenerated electrons. The photoluminescence (PL) spectra exhibited two fitting peaks centered at 625 and 710\u00a0nm in Figure\u00a07D. And the PL peak at 625 and 710\u00a0nm was attributed to the bandgap radiative recombination and the \u03c0-\u03c0 interaction between the COF and \u03b2-ketoenamine layers, respectively. Moreover, these two emissions peaks of Pt1@TpPa-1 were quenched significantly compared with TpPa-1, due to the interfacial charge transfer from TpPa-1 to Pt SAs. The charge transfer behavior was further probed through the time-resolved PL (TRPL) decay spectra (Figure\u00a07E). It was shown that the anchoring of Pt SAs (3% Pt1@TpPa-1) led to a shorter lifetime (0.27\u00a0ns) compared with TpPa-1 (0.50\u00a0ns), which was attributed to the addition of Pt SAs in the COF. The possible charge transfer behavior and reaction routes are illustrated in Figure\u00a07F. As a result, the Pt1@TpPa-1 provided more photocarriers for the subsequent surface photocatalysis reactions, thus improving the photocatalytic performance. The protons (H+) produced by the dissociation of H2O were then reduced to the transitional state (H\u2217) and finally evolved into H2. The aforementioned result implied that the Pt SAs could facilitate the efficient migration of photoelectrons, thus improving the efficiency of the photocatalytic reactions.Besides affecting the photocatalysts' light-harvesting and charge transfer capability, it is demonstrated that the introduced SAs could also act as the reaction active sites and directly participate in the reaction (Liu et\u00a0al., 2021a; Wang et\u00a0al., 2021b). Compared with general metal nanoparticle/semiconductor catalysts, the SAPs, maximizing the atomic utilization, exhibit boosted reaction active sites and consequent excellent photocatalytic performance (Figures\u00a08A and 8B) (Jiao et\u00a0al., 2021; Wang et\u00a0al., 2019e; Zhang et\u00a0al., 2022a). It is generally considered that the surface reaction process is strongly influenced by the geometric effect and the electronic structure of the catalysts (Gao et\u00a0al., 2020a). Constructing SAPs provides an effective method to manipulate the local coordination environments of the SAs, offering an easy way to regulate the adsorption/activation mode of the reactant as the surface of the catalysts. In addition, as the SAs offered a large number of unsaturated coordination centers, numerous reaction active sites could be provided for the surface reaction process (Wang et\u00a0al., 2019b).Currently, metal oxides (TiO2, ZnO, etc.) and metal sulfides (CdS, MoS2, etc.) are widely studied as the photocatalysts (Kumar et\u00a0al., 2021; Kusiak-Nejman et\u00a0al., 2021; Premaratne et\u00a0al., 2004). As discussed in section atomic layer deposition method, by creating defect sites (i.e., O-defect, S-defect), the metal-oxides- and metal-sulfides-based SAPs can be easily synthesized by the atom trapping methods. Due to the synergistic effect of the SAs and the adjacent defect sites, it is reported that enormous reaction active sites could be provided.For instance, Xing et\u00a0al. successfully loaded various SAs (i.e., Pt, Pd, Ru, Rh, etc.) onto the TiO2 substrate for photocatalytic hydrogen evolution reactions (HER) (Xing et\u00a0al., 2014). The turnover frequencies (TOFs) of 0.2Pt/TiO2 photocatalytic hydrogen production was about 24 and 6 times higher than that of 2Pt/TiO2 and 1Pt/TiO2 (PD, denoted nanoparticles), respectively (Figure\u00a08C). It was found that the same phenomenon also existed in some other SAPs (Pd, Ru, and Rh), confirming the introduction of SAs could provide extra reaction active sites compared with bulk catalysts. In addition, Wu et\u00a0al. successfully introduced various concentrations of Pt SAs onto the TiO2 nanotubes in different concentrations of HPtCl4 solution (2\u20130.0005\u00a0mM) and applied them for the photocatalytic HER (Wu et\u00a0al., 2021). It was found that compared with Pt nanoparticles, the Pt SAPs also exhibited better HER performance.For metal-sulfides-based SAPs, similar results were evidenced. For instance, Zhao et\u00a0al. reported Co SAs/N-doped graphene-modified CdS (Co-NG/CdS) (Zhao et\u00a0al., 2017), exhibiting efficient photocatalytic HER performance. The 0.25 wt% of Co-NG/CdS showed an H2 evolution rate of 1077\u00a0\u03bcmol h\u22121, which was 1.3 times higher than that of the Pt-NPs/CdS photocatalyst (1382\u00a0\u03bcmol h\u22121), confirming the contribution of SAs on the reaction active sites (Figures\u00a08D\u20138E). Moreover, the turnover numbers (TONs) were calculated to be 58.2 and 474.764 for the CdS and Co-NG/CdS, respectively. For the 0.25 wt% Co-NG/CdS photocatalyst, the TOF was approximately 8.8 s\u22121. These TON and TOF values show that, under the same reaction conditions, the SAPs showed better reaction activity than the metal nanoparticle/semiconductor catalysts.As the coordination environment of the active atoms in SAPs can be flexibly regulated, it affords us an effective effort to regulate the reactant adsorption/activation modes on the catalysts\u2019 surface, thereby altering the reaction pathway and finally achieving high product selectivity in the photocatalytic process (Gao et\u00a0al., 2020a; Wang et\u00a0al., 2021g). As discussed in section principle of photocatalysis, photocatalysis techniques have been widely applied to various solar-driven chemicals/fuels generation, including HER (Alarawi et\u00a0al., 2019; Zhang and Guan, 2020), CRR (Gao et\u00a0al., 2018; Wang et\u00a0al., 2019c), nitrogen reduction reaction (NRR) (Huang et\u00a0al., 2018; Li et\u00a0al., 2020a), etc. Table\u00a01\n overviewed some recent research on SAPs for the inorganic photocatalytic reactions.For photocatalytic CO2 reductions, it was demonstrated that the presence of some SAs can enhance the adsorption of CO2 molecules, stabilize the photocatalytic CO2 reduction intermediates, and accelerate the CO desorption, thereby achieving better CO selectivity in the photocatalytic process (Zhang et\u00a0al., 2020). For instance, Di et\u00a0al. demonstrated that, by replacing Bi3+ with Co SAs, the CoBi3O4Br atomic shell could be negatively charged, which facilitated the adsorption of CO2, as shown in Figure\u00a09A (Di et\u00a0al., 2019). By virtue of in-situ Fourier transform infrared spectroscopy (FTIR), it allowed insight into the reaction intermediates of photocatalytic CO2RR (Figure\u00a09B). The, peaks at 1256, 1337, and 1508\u00a0cm\u22121 were attributed to\u00a0the CO2\n\u2212, symmetrical O-C-O extended bidentate carbonate (b-CO3\n2\u2212) and monodentate carbonate (m-CO3\n2\u2212) groups, respectively. The increasing peak intensity at 1567\u00a0cm\u22121 was attributed to the COOH\u2217 intermediate, which was an important intermediate for the formation of CO. Finally, the CO desorption was also considered to be an important factor in determining the comprehensive photocatalytic efficiency. The temperature-programmed desorption (CO-TPD) curves in Figure\u00a09C demonstrated that Co-Bi3O4Br-1 possessed a lower initial CO desorption temperature, indicating that as-formed CO\u2217 could be easily removed from Co-Bi3O4Br-1 surface and higher CO yield rates (Figure\u00a09D). Therefore, the introduction of Co SAs could promote the adsorption of CO2 molecules and reduce the activation energy barrier of CO2 by stabilizing the COOH\u2217 intermediate and adjusting the rate-limiting step to CO\u2217 desorption (Figures\u00a09E\u20139F), thus exhibiting excellent photocatalytic activity and selectivity.The SAPs have also been applied for photocatalytic NRR. For instance, Li et\u00a0al. successfully anchored Pt SAs to the ultra-thin CTF nanosheet (Pt-SAC/CTF) (Li et\u00a0al., 2020a). The high-angle annular dark-field scanning transmission electron microscopy (HADDF-STEM) images of the obtained catalysts were shown to confirm the appearance and atomic diameter distribution of the Pt-SAC/CTF catalyst (Figures\u00a09G and 9H). The even dispersion of Pt, C, and N atoms corresponding to the EDS mapping image could be observed in Figure\u00a09I. The electronic state of the Pt species in the Pt-SAC/CTF catalyst was explored by the X-ray absorption near-edge structure analysis (XANES) (Figure\u00a09J). The white line intensity of the Pt-SAC/CTF was lower than that of PtO2, but higher than that of Pt foil, indicating that the unoccupied density lay between PtO2 and Pt foil. The extended X-ray absorption fine structure (EXAFS) spectrum confirmed the same coordination structure of the Pt-N3 site in the ultra-thin CTF-PDDA-TPDH nanosheets. The Fourier transform EXAFS (FT-EXAFS) shown in Figure\u00a09K showed a main peak at 2.34\u00a0\u00c5, which corresponded to the metal Pt bond of the standard Pt foil. The sharp peak at 1.57\u00a0\u00c5 for the Pt-SAC/CTF catalyst indicated that the Pt presented as SAs in the Pt-SAC/CTF catalyst. Then under the visible light radiation, the photocatalytic N2 immobilization experiment was carried out. The average NH3 production rate of the Pt-SAC/CTF catalyst was 171.40\u00a0\u03bcmol g\u22121 h\u22121 (Figure\u00a09L), which was 6 times and 1.5 times higher than that of the CTF-PDDA-TPDH and Pt-NPs/CTF catalysts, respectively.Besides the inorganic photocatalytic reaction, SAPs can also be applied in some organic-related photocatalytic reactions, such as biomass reforming, organic synthesis, and pollutant degradation, which were summarized and listed in Table\u00a02\n. For instance, da Silva et\u00a0al. mixed Na-PHI and FeCl3 together to introduce Fe3+ into the poly(heptazine imides) (PHI) matrix, thereby obtaining the target catalyst Fe-PHI (Figure\u00a010A) (da Silva et\u00a0al., 2022). In Figure\u00a010B, the Fe-PHI XRD peaks were mainly significant differences between 25\u00b0 and 30\u00b0. This apparent difference was assigned to the statistical positioning of Fe ions in the mainframe of the crystal. In Figure\u00a010C, Fe SAs could be well distinguished, confirming the successful synthesis of the Fe-PHI catalyst. EXAFS, Fourier transforms (FTs), and wavelet transforms (WTs) spectra were shown in Figures\u00a010D\u201310F. Detailed analysis showed that once Fe3+ was introduced into the PHI structure, Fe3+ would combine with the N atom of the heptazine ring. The DFT calculations showed that Fe3+ ions were located between the PHI layers, with each Fe3+ ion coordinating with four N atoms and two in each PHI layer (Figure\u00a010G). Moreover, earlier, it was claimed that the C-H bond had high dissociation energy and the C-H bonds were easier to be over-oxidized and lead to the side reactions (Liu et\u00a0al., 2017). However, it was found that the geometric structure enabled the Fe-PHI catalyst to promote the selective oxidation of C-H bonds. As a result, the Fe-PHI SAP was applied for the catalytic oxidation of the ethylbenzene. It was demonstrated that the Fe-PHI (0.1%) exhibited the superior oxidation activity (Figure\u00a010H), with the 99.6% ethylbenzene conversion rate and 98.4% acetophenone selectivity.In addition to the aforementioned biomass refining reaction, the researchers put another focus on organic synthesis to promote the C-C coupling reaction (Toe et\u00a0al., 2021). Zhou et\u00a0al. successfully synthesized Pt SAs-loaded TiO2 (PtSA-TiO2) and applied it for the production of 2,5-hexanedione (HDN), an important chemical in biofuels and medicinal chemistry, from low-cost acetone dehydrogenation (Zhou et\u00a0al., 2020a). It was the first application of the in-situ icing-assisted photocatalytic reduction method to anchor Pt SAs on TiO2. The HAADF-STEM image shown in Figure\u00a011A confirmed the presence of Pt SAs on TiO2. The coordination structure of Pt SAs on TiO2 was analyzed by X-ray absorption near edge structure (XANES) spectra in Figure\u00a011B, which indicated that the absorption edge of Pd SAs was higher than that of Pt nanoparticles-loaded TiO2 (PtNP-TiO2) and Pt foil (Figure\u00a011B). The EXAFS spectrum of the PtSA-TiO2 showed a main peak at 1.61\u00a0\u00c5 in the R space and a maximum at 5.61\u00a0\u00c5\u22121 in the k space (Figure\u00a011C), both of which were assigned to the Pt-O bond (Figures\u00a011D and 11E). Subsequently, the photocatalytic acetone conversion was carried out under the irradiation of a 300\u00a0W xenon lamp at 25\u00b0C. The results showed that the HDN production rate of PtSA-TiO2 was 3.87\u00a0mmol g\u22121 h\u22121, which was 6 times higher than that of PtNP-TiO2, confirming the excellent reaction activity achieved by the Pt SA catalysts. The gas chromatography-mass spectrometry (GC-MS) in Figure\u00a011E confirmed that the photocatalytic product contained HDN and H2. In addition, the HDN-production activity of PtSA-TiO2 can maintain four cycles in 16\u00a0h (Figure\u00a011F). To further explore the catalysis mechanism, the attenuated total reflection infrared ((ATR)-IR) spectrum discovered that the two IR peaks at 2921 and 2852\u00a0cm\u22121 were ascribed to the C-H bond in the methyl group of acetones. As for the PtSA-TiO2, these two peaks revealed a sharp decrease (Figure\u00a011G), which implied the activation of methyl and acetone tended to be dehydrogenated at the surface of PtSA-TiO2. In Figure\u00a011H, the electron spin resonance (ESR) spectrum exhibited the CH3COCH2\n\u2022 radical on PtSA-TiO2, which was an important intermediate for the production of HDN by C-C coupling. This result showed that Pt SAs exhibited a significant influence on the acetone dehydrogenation reactions. All of the aforementioned results suggested that the PtSA-TiO2 possessed a relatively low reaction barrier for acetone dehydrogenation reaction, which was also proved in Figure\u00a011I. Similarly, Wang et\u00a0al. successfully synthesized Pt/gC3N4 SAPs using a photo-deposition method, combining the oxidation of benzaldehydes with simultaneous proton reduction (Wang et\u00a0al., 2021c). The benzaldehyde conversion rate of Pt/gC3N4 reached 49.5\u00a0mmol/gPt, and the hydrogen evolution rate of Pt/gC3N4 was 24\u00a0mmol/gPt. Pt/gC3N4 SAPs exhibited nearly 100% efficiency per atom in the production of benzoic acid and clean H2 fuel.SAPs usually consist of two parts, the supporting substrate and the SAs. By decorating the substrate materials with SAs, enhanced photocatalytic performance could always be obtained. Therefore, in the following section, the current research progress of both the substrate and SAs will be comprehensively overviewed.To synthesize SAPs, the applied supporting substrates are usually semiconductor materials. Meanwhile, acting as the substrates of SAPs, the applied semiconductors should be able to anchor the SAs and prevent the aggregation of the SAs. In this section, the supporting substrate catalysts will be categorized as organic, inorganic, and carbon-based materials.Currently, organic materials have been widely used as SAPs substrates, such as metal-organic frameworks (MOFs) and covalent organic frameworks (COFs). For the MOF substrates, the selection criteria are mainly based on the following three aspects (Jiao and Jiang, 2019; Li et\u00a0al., 2016): (1) possessing large specific surface area, which is conducive to the adsorption of reactants; (2) exhibiting optical activity, which can generate photo-induced carriers to participate in the photocatalytic reaction; and (3) providing pore confinement, which can prevent the aggregation of metal with a relatively high metal loading. For MOF itself, its unsaturated coordination sites, defects, and the porous structure can be utilized to anchor metal SAs, making it an ideal substrate for anchoring the SAs (Jiao and Jiang, 2019). For example, the Pt1/SnO2/UiO-66-NH2 catalysts were successfully synthesized by Sui et\u00a0al., applying for the visible-light-driven HER (Sui et\u00a0al., 2021). The obtained Pt1/SnO2/UiO-66-NH2 SAPs showed a superior H2 evolution rate of 2167\u00a0\u03bcmol g\u22121 h\u22121. Further, Li et\u00a0al. synthesized MOF-808-EDTA with implanted Pt SAs (Li et\u00a0al., 2019), which exhibited an excellent photocatalytic H2 evolution rate (68.33\u00a0mmol g\u22121 h\u22121) under visible light irradiation (Figure\u00a012A).As for the COFs, SAs can be confined within the COF through the coordination interaction between the metal atom and the binding groups in COF (Wei et\u00a0al., 2020). Moreover, COFs possess heteroatom-rich pore walls that can facilitate reactant adsorption and charge transfer, resulting in more efficient photocatalytic reactions (Zeng and Xue, 2021). Therefore, the utilization of COFs as the substrates to capture SAs is expected to bring new opportunities for the development of SAPs. For example, Dong et\u00a0al. reported a two-dimensional \u03b2-ketoenamine-linked COF supporting Pt SAs (Pt1@TpPa-1) for photocatalytic HER (Dong et\u00a0al., 2021a). TpPa-1-COF showed special holes and dispersed unsaturated coordinating nitrogen atoms, which made the Pt SAs highly dispersed. The optimal 3% Pt1@TpPa-1 showed the best H2 evolution rate of 99.86\u00a0mmol gPt\n\u22121 h\u22121 (Figure\u00a012B). Besides Pt, the active Mo SAs were also impregnated in the TPBPY-type COF to get Mo-COF, realizing an efficient photocatalytic reduction of CO2 to produce C2H4 (3.57\u00a0\u03bcmol g\u22121 h\u22121) (Kou et\u00a0al., 2021).To date, metal oxides are the most used inorganic substrate for the synthesis of SAPs, as the SAs can be anchored on metal oxides through the metal-oxygen bonds or be stabilized through oxygen vacancies, contributing to the enhanced stability of the SAPs. For instance, Hu et\u00a0al. demonstrated that the Pt SAs could incorporate defective TiO2 nanosheets (Pt SA/Def-s-TiO2) for photocatalytic water splitting (Figure\u00a012C) (Hu et\u00a0al., 2021b). The surface oxygen vacancies could efficiently stabilize the Pt SAs by forming a three-center Ti-Pt-Ti structure, which also contributed to the enhanced charge transfer processes. As a result, greatly enhanced photocatalytic HER was evidenced. Notably, the Pt SA/Def-s-TiO2 SAPs exhibited an enhanced H2 evolution performance (13460.7\u00a0\u03bcmol h\u22121 g\u22121), which was 29.0 times higher than that of TiO2 nanosheets.Similar to metal oxides, the unsaturated coordinated sulfur atoms in metal sulfides could also bond with metal SAs to form SAPs. For instance, Li et\u00a0al. synthesized CdS-Pd SAPs through the photoreduction method (Li et\u00a0al., 2022). It was demonstrated that the CdS-Pd SAPs exhibited considerable structural stability and photocatalytic HER performance due to the synergistic interaction between CdS and Pd, achieving an efficient charge transfer process to the catalysts\u2019 surface. The obtained H2 evolution rate (947.93\u00a0\u03bcmol g\u22121 h\u22121) was about 110 times higher than that of pure CdS NPs (8.64\u00a0\u03bcmol g\u22121 h\u22121).Besides the metal oxides and sulfides, recently, halide perovskites materials are also demonstrated to be a potential substrate material to synthesize the SAPs. Halide perovskites possess fascinating properties such as broad light absorption, long charge carrier migration lengths, etc. (Fu and Draxl, 2019). Currently, the halide perovskites are demonstrated with excellent photocatalytic performance, besides being applied in the photovoltaic fields. In this regard, synthesizing halide perovskites-based SAPs is promising to obtain extraordinary catalytic performance. The presence of SAs is expected to effectively enhance the interaction between the halide perovskite and the reactant molecules (Fu and Draxl, 2019). For instance, Wu et\u00a0al. successfully anchored Pt SAs onto FAPbBr3-xIx (Pt/FAPbBr3-xIx) with high dispersibility and stability (Figure\u00a012E) (Wu et\u00a0al., 2022). The obtained Pt/FAPbBr3-xIx SAPs showed enhanced photocatalytic hydrogen production activity, reaching 682.6\u00a0\u03bcmol h\u22121 (100\u00a0mg). In addition, Hu et\u00a0al. demonstrated that the Pt SAs could be deposited onto the CsPbBr3 NCs (Pt-SA/CsPbBr3) through the formation of Pt-O and Pt-Br bonds (Hu et\u00a0al., 2021a). Compared with pristine CsPbBr3 NCs, the trap levels exhibited in the Pt-SA/CsPbBr3 were ascribed to the deposition of Pt SAs, leading to an enhanced separation capability of the photogenerated carriers. Because of the fast carrier transfer from CsPbBr3 to Pt SAs, the Pt-SA/CsPbBr3 exhibited a superior activity toward the photocatalytic propyne semi-hydrogenation (TOF\u00a0= 122.0 h\u22121).Because of the excellent conductivity of graphene, carbon-based substrates have been widely used to anchor metal SAs for not only electrocatalysis (Su et\u00a0al., 2021a; Su et\u00a0al., 2021b; Tian et\u00a0al., 2021; Wang et\u00a0al., 2021f) but also the photocatalysis field (Zhuo et\u00a0al., 2020). Similar to organic and inorganic substrates, structurally modified graphene can bind with SAs through the coordination interactions with oxygen- or nitrogen-containing functional groups. For instance, Gao et\u00a0al. used oxidized graphene nanosheets as the substrates to immobilize the isolated Co SAs (Co1-G). Under this circumstance, the graphene acted as a bridge to connect the Ru(bpy)3 photosensitizer and the Co SAs, thereby realizing effective charge transfer and CO2 reduction (Gao et\u00a0al., 2018). It was demonstrated that the Co SAs were coordinated with the carbon and residue oxygen on the graphene surface and exhibited outstanding TON (678) and TOF (3.77\u00a0min\u22121) toward photocatalytic CRR. In addition, N-doped carbon substrates are also widely applied in photocatalysis, which provide rich coordination sites for the anchoring of the SAs (Liu et\u00a0al., 2021b). For instance, Zhao et\u00a0al. demonstrated that the Ni SAs-decorated N-graphene/CdS (Ni-NG/CdS) could be efficient SAPs for photocatalytic HER (Zhao et\u00a0al., 2018). In this work, Ni SAs were anchored on the vacancies in nitrogen-doped graphene (Ni-NG). In the obtained catalysts, the Ni-NG acted as the electron storage medium to suppress the carrier recombination and the active site for the reduction reaction. As a result, the Ni-NG/CdS received an outstanding photocatalytic HER performance with a quantum efficiency of 48.2% at 420\u00a0nm.As discussed earlier, to regulate the charge carriers\u2019 generation/transfer and surface reaction process, loading metal nanoparticles to modify the pristine semiconductor catalysts is a generally applied strategy. However, due to the high expense and scarce reserves of noble metal, increasing the utilization efficiency of metal atoms is of great importance, which is also applicable to nonnoble metal species (Li et\u00a0al., 2021b). In the following section, the currently studied SAs species are systematically summarized.Currently, various noble metals, such as Pt, Pd, Ir, Au, Ag, Rh, Ru, etc., have been applied to synthesize SAPs due to their excellent catalytic activities. For example, Liu et\u00a0al. applied g-C3N4 with carbon vacancies (Cv-CN) to anchor Pd SAs (Pd-Cv-CN), applying for the photocatalytic NO reduction reaction (Figure\u00a013A) (Liu et\u00a0al., 2021a). The results showed that the Pd SAs could be successfully anchored to the carbon vacancies and uniformly dispersed on the Cv-CN surface, thereby forming isolated Pd-N3 sites. In the case of photocatalytic NO conversion, Pd-Cv-CN not only exhibited higher conversion efficiency of 56.3% but also higher selectivity and stability toward NO3\n\u2212 generation compared with Cv-CN. Similarly, Li et\u00a0al. prepared Pd/TiO2 SAPs by the liquid-phase reduction method and applied it in the photocatalytic CRR (Li et\u00a0al., 2017). The results showed that the Pd SAs could be uniformly dispersed on the surface of TiO2, leading to improved CRR activity. The enhanced CRR efficiency was attributed to the synergistic effect of Pd SAs and TiO2, as the Pd SAs could act as the electron trap center to capture photogenerated electrons and inhibit the recombination of photo-induced electrons and holes.The nonnoble metal-based SAPs are focused on the transition metals such as Fe, Co, Cu, Ni, etc. The transition metals have vacant orbitals that can accept electrons as electron traps and avoid the recombination of photogenerated electron-hole pairs (Abdullah et\u00a0al., 2017). Ma et\u00a0al. dispersed Co SAs on g-C3N4 nanosheets with ultra-high density of Co-N2C active sites and applied the obtained SAPs for the photocatalytic CRR (Figure\u00a013B) (Ma et\u00a0al., 2022). They demonstrated that the Co-N2C sites served not only as the electron aggregation center but also as the CO2 adsorption/activation sites, which subsequently promoted the photocatalytic methanol generation performance. As a result, the methanol formation rate for 4\u00a0h was 941.9\u00a0\u03bcmol g\u22121 over Co/g-C3N4-0.2, which was 13.4 times of g-C3N4 (17.7\u00a0\u03bcmol g\u22121). Moreover, Zhang et\u00a0al. dispersed Co SAs into MOFs (MOF-525-Co) for the CO2 photoreduction (Zhang et\u00a0al., 2016). They proved that the Co SAs could act as the CO2 adsorption sites. Simultaneously, the photogenerated electrons could transfer from the MOFs to the Co active sites feasibly, thereby improving the photocatalytic efficiency.With empty and occupied orbitals, the atomic structures of some nonmetal elements (e.g., B, Si, etc.) are similar to that of the transition metals (Zhao et\u00a0al., 2022). Compared with metal-based SAPs, metal-free-based SAPs have also been extensively studied due to their low cost and environmental friendliness.\u00a0Although metal-free-based SAPs show weaker catalytic activity compared with metal-based SAPs, they\u00a0yet have good stability and strong resistance to poisoning and deactivation. Ling et\u00a0al. designed a boron-atom-decorated graphitic-carbon nitride (B/g-C3N4) for the photocatalytic NRR (Figure\u00a013C) (Ling et\u00a0al., 2018). By analyzing the extranuclear electronic structure of boron atoms, they found that the sp3-hybridized boron atoms were similar to transition metals with empty and occupied orbitals, which could be used as the reaction active center for the NRR. Furthermore, the modification of B SAs can significantly enhance the visible light absorption capability of g-C3N4, thus promising to realize the visible-light-driven NRR. Lv et\u00a0al. also reported that B SAs could be applied to reduce dinitrogen to ammonia spontaneously (Lv et\u00a0al., 2019).Generally, for SAPs, the metal/nonmetal species are dispersed on the supporting substrates in the form of SAs, acting as the active sites for the photocatalytic reaction. Therefore, it is crucial to clarify the geometric structure, electronic structures, and the spatial distribution of SAs for the deep study of SAPs. Advanced electron microscopy analysis techniques, such as scanning tunneling microscope (STM) and high-angle annular dark-field-scanning transmission electron microscope (HAADF-STEM), can provide advanced understandings of the structure of SAPs, making it possible to identify the SAs at the magnitude at c.a. \u223c 0.1\u00a0nm; spectroscopy techniques, such as X-ray absorption fine structure (XAFS) spectroscopy and infrared (IR) spectroscopy, can also be applied to identify the existence of SAs and clarify the electronic structure and chemical state of the obtained SAPs.The typical electron microscopes techniques, scanning electron microscope (SEM) and transmission electron microscope (TEM), can toughly identify the SAs at the atomic level. In this regard, HAADF-STEM is applied to observe SAs due to the improved spatial resolution of its sub-Angstrom probe (Gao et\u00a0al., 2019b; Peng et\u00a0al., 2004). This technique has been chosen for heavy elements on light substrates for the strong correlation between atomic number and imaged intensity, called Z-contrast (LeBeau et\u00a0al., 2008; Nellist et\u00a0al., 2010). Under dark-field conditions, different atoms have different Z-contrasts, making the atoms distinguishable by observing their brightness in the HAADF-STEM images (Chung et\u00a0al., 2019). In Figure\u00a014A, the spherical-aberration-corrected HAADF-STEM images of O/La-CN SAPs showed the bright dots, which corresponded to the even dispersed La SAs on CN substrate due to the different Z contrasts of La, C, and N atoms (Chen et\u00a0al., 2020a).STM is a characterization technique applied to probe the surfaces and absorb substances at the atomic level with ultra-high resolution of 0.1\u00a0nm laterally and 0.01\u00a0nm in depth. The SAs can be imaged and manipulated with the conductive tips (Gao et\u00a0al., 2020a). For example, Deng et\u00a0al. used STM to reveal the existence of Fe SAs with FeN4 center in the graphene matrix as shown in Figure\u00a014B (Deng et\u00a0al., 2015). The iron center was shown as a bright spot, whereas adjacent atoms (C and N) exhibited a higher apparent height than other C atoms in the graphene matrix. In the simulated STM images, the FeN4 centers embedded in the graphene lattice were consistent with the STM images, which better revealed the iron centers significantly alter the density of states of adjacent atoms (N and C) (Figure\u00a014C).In addition to the aforementioned microscopy techniques, XAFS spectroscopy, including XANES and EXAFS, is another effective way for the characterization of SAPs, which is used to analyze the coordination environment and electronic structure in the material structure. According to the characteristics of peaks and shoulders in the XANES spectrum, the electronic structure and chemical valence state of SAs can be obtained. In the EXAFS spectra, SAs can be identified through morphological imaging characterization and corresponding spectral information sensitive to atomic structure, so as to obtain the coordination number, coordination form, and coordination distance of the planted SAs to the adjacent atoms in the SAPs. For instance, Sui et\u00a0al. performed XAS to determine the coordination environment and chemical state of Pt species in Pt1/SnO2/UiO-66-NH2 SAPs (Sui et\u00a0al., 2022). From the L3-edge image, it could be seen that the peak intensity of Pt1/SnO2/UiO-66-NH2 was closer to that of PtO2, implying the presence of a highly oxidized Pt state (Figure\u00a014D). The Fourier transform expansion X-ray absorption fine structure spectra (FT-EXAFS) of the Pt1/SnO2/UiO-66-NH2 gave only a dominant peak at about 1.63\u00a0\u00c5, which can be attributed to the first shell of the Pt-O bond, rather than the Pt-Cl bond and Pt-Pt bond, suggesting the existence of atomically dispersed Pt sites in Pt1/SnO2/UiO-66-NH2 (Figure\u00a014E).IR spectroscopy can also be utilized to identify the presence of SAs and to quantify the percentage of metal SAs to some extent (Chen et\u00a0al., 2017b; Liu, 2017). The principle is as follows, IR is used to detect the interaction between the substrate and the adsorbed molecule by utilizing probe molecules (e.g., CO, NH3, pyridine, etc.) (Gallenkamp et\u00a0al., 2021). For instance, Ding et\u00a0al. applied the IR spectra to confirm the Pt SAs in the Pt/HZSM-5 catalysts by analyzing the CO adsorption mode (Ding et\u00a0al., 2015). As shown in Figure\u00a014F, the peak at 2115\u00a0cm\u22121 was attributed to CO molecules adsorbed on Pt SAs. Meanwhile, the peak at 2090\u00a0cm\u22121 was ascribed to CO molecules linearly adsorbed on Pt nanoparticles. It could be inferred that Pt existed as SAs on Pt/HZSM-5 with a low Pt loading at 0.5 wt\u00a0% by studying the changes in peak intensity for the four catalysts. And the Pt atoms tended to agglomerate to form Pt nanoparticles after increasing the Pt loading from 0.5 wt% to 2.6 wt%. In addition, diffuse reflectance infrared Fourier\u00a0transform spectroscopy (DRIFTS) offers another technique to gain insight into the local information of the SAPs, as it is usually applied to in-situ collect information of the surface reactive species and intermediates during the reaction (Yang et\u00a0al., 2019). Typically, CO is used as the probe molecule in DRIFTS studies because of its advantages in characterizing the exposed noble metal sites on loaded catalysts (Liang et\u00a0al., 2022). For instance, Fang\u00a0et\u00a0al. collected the in-situ DRIFTS spectra of the CO adsorption behavior over the Al-TCPP-0.1Pt SAPs (Fang et\u00a0al., 2018). After purging with Ar to remove gaseous CO, the peak centered at 2090\u00a0cm\u22121 corresponded to the CO chemisorbed on Pt SAs (Figure\u00a014G). In the ranges of 2080\u20132030\u00a0cm\u22121 and 1920\u20131950\u00a0cm\u22121, no bands that could linearly and bridge CO adsorption on Pt clusters and nanoparticles appeared, implying that all Pt species were atomically dispersed.To date, SAPs have been widely studied in solar-driven chemicals/fuels generation, with various SAPs synthesis strategies being established. The pioneer works demonstrated that the introduction of SAs to the commonly used semiconductor catalysts can directly influence the overall photocatalysis process: (1) by modulating the band structure with impurity level or directly reducing the bandgap, SAPs, therefore, exhibits greatly enlarged optical absorption range; (2) due to the unique band bending effect at the metal SAs/semiconductor interfaces, the spatial separation and transfer of the photogenerated carriers can be significantly promoted; (3) with the tuneable coordination environment of the SAs, SAPs are equipped with boosted reaction active sites and better product selectivity. Moreover, rationally choosing the supporting substrate materials and the loaded SAs species is expected to regulate the surface reaction process efficiently. However, due to the insufficient understanding of the structure-catalytic performance relationship based on SAPs, in the future, the study of SAPs still faces some crucial issues:\n\n(1)\nDue to the strong influence of the local electronic structure of the material, the rational design of SAPs with high loading rates of the SAs is still a significant challenge. The knowledge to prepare stable and efficient SAPs with high SAs loading is still in its infancy. The loading amounts of SAs can reach 23 wt% now for nitrogen-doped carbon and polymeric carbon nitride (Lu et\u00a0al., 2021). However, for other supports, the loading rates are still less satisfactory. Consequently, it will be promising to achieve higher SAs loading rates by ligands protection or through the ambient multistep method to regulate the removal of the ligands from the metal precursors and enhance the associated interactions between SAs and substrates.\n\n\n(2)\nTo date, there is still lack of efficient methods to monitor the reaction dynamics of the catalytic process. Currently, the characterizations of the ligand environment and associated charge transfer processes mainly rely on theoretical calculations. Direct monitoring of the reaction dynamics is still challenging. In the future study, in-situ characterization techniques, such as in-situ electron microscopy or in-situ synchrotron radiation technique, need to be applied to monitor and probe the photocatalytic reaction process. Combined with theoretical calculations, the relationship between their structure and catalytic performance can be better explained.\n\n\n(3)\nThe advanced understanding of the photocatalytic mechanism over the SAPs is insufficient, making this technique challenging to practical application. To date, only a few studies tried to uncover the effect of the SAs on the reaction mechanism. To illustrate the impact of SAs on the reaction pathway, it will be promising to apply homogenized substrates for the SAs loading to investigate the accurate active sites of SAPs. This will benefit the design of some specific SAPs fitted for the targeted reactions on an atomic scale.\n\n\n\n\n\n\n\nAbbreviations\nSACs Single-atom catalysts\nSAs Isolated single-atoms\nSAPs Single-atom photocatalysts\nCVD Chemical vapor deposition\nFSP Flame spray pyrolysis\nALD Atomic layer deposition\nCB Conduction band\nVB Valence band\nCOFs Covalent organic frameworks\nPL Photoluminescence\nTRPL Time-resolved PL\nTOFs Turnover frequencies\nTONs Turnover numbers\nFTIR Fourier transform infrared spectroscopy\nTPD Temperature-programmed desorption\nPHI Poly(heptazine imides)\nEXAFS Extended X-ray absorption fine structures\nFTs Fourier transforms\nWTs Wavelet transforms\nHDN 2,5-hexanedione\nXANES X-ray absorption near edge structure\nGC-MS Gas chromatography-mass spectrometry\nESR Electron spin resonance\nHADDF-STEM High-angle annular dark-field scanning transmission electron microscopy\nFT-EXAFS Fourier transform EXAFS\nHER Hydrogen evolution reaction\nCRR Carbon dioxide reduction reaction\nNRR Nitrogen reduction reaction\nDRIFTS Diffuse reflectance infrared Fourier transform spectroscopy\nSTM scanning tunneling microscope\nXAFs X-ray absorption fine structure\nIR Infrared\n(ATR)-IR Attenuated total reflection infrared\nMOFs Metal-organic frameworks\n\n\n\n\n\nDue to the strong influence of the local electronic structure of the material, the rational design of SAPs with high loading rates of the SAs is still a significant challenge. The knowledge to prepare stable and efficient SAPs with high SAs loading is still in its infancy. The loading amounts of SAs can reach 23 wt% now for nitrogen-doped carbon and polymeric carbon nitride (Lu et\u00a0al., 2021). However, for other supports, the loading rates are still less satisfactory. Consequently, it will be promising to achieve higher SAs loading rates by ligands protection or through the ambient multistep method to regulate the removal of the ligands from the metal precursors and enhance the associated interactions between SAs and substrates.To date, there is still lack of efficient methods to monitor the reaction dynamics of the catalytic process. Currently, the characterizations of the ligand environment and associated charge transfer processes mainly rely on theoretical calculations. Direct monitoring of the reaction dynamics is still challenging. In the future study, in-situ characterization techniques, such as in-situ electron microscopy or in-situ synchrotron radiation technique, need to be applied to monitor and probe the photocatalytic reaction process. Combined with theoretical calculations, the relationship between their structure and catalytic performance can be better explained.The advanced understanding of the photocatalytic mechanism over the SAPs is insufficient, making this technique challenging to practical application. To date, only a few studies tried to uncover the effect of the SAs on the reaction mechanism. To illustrate the impact of SAs on the reaction pathway, it will be promising to apply homogenized substrates for the SAs loading to investigate the accurate active sites of SAPs. This will benefit the design of some specific SAPs fitted for the targeted reactions on an atomic scale.SACs Single-atom catalystsSAs Isolated single-atomsSAPs Single-atom photocatalystsCVD Chemical vapor depositionFSP Flame spray pyrolysisALD Atomic layer depositionCB Conduction bandVB Valence bandCOFs Covalent organic frameworksPL PhotoluminescenceTRPL Time-resolved PLTOFs Turnover frequenciesTONs Turnover numbersFTIR Fourier transform infrared spectroscopyTPD Temperature-programmed desorptionPHI Poly(heptazine imides)EXAFS Extended X-ray absorption fine structuresFTs Fourier transformsWTs Wavelet transformsHDN 2,5-hexanedioneXANES X-ray absorption near edge structureGC-MS Gas chromatography-mass spectrometryESR Electron spin resonanceHADDF-STEM High-angle annular dark-field scanning transmission electron microscopyFT-EXAFS Fourier transform EXAFSHER Hydrogen evolution reactionCRR Carbon dioxide reduction reactionNRR Nitrogen reduction reactionDRIFTS Diffuse reflectance infrared Fourier transform spectroscopySTM scanning tunneling microscopeXAFs X-ray absorption fine structureIR Infrared(ATR)-IR Attenuated total reflection infraredMOFs Metal-organic frameworksThe authors acknowledge the financial support from the Sydney Nano Grand Challenge, at the University of Sydney and Australia Research Council Linkage Project (LP200200615). H.S. is grateful to Lizhuo Wang for his help in the discussions of the characterization section. Dedication: This work is dedicated to Professor Jianzhong Chen on the occasion of his 70th birthday.Conceptualization: J.H.; Writing - Original Draft: H.S.; Writing-Review & Editing: R.T., J.H.; Supervision: J.H.The authors declare no competing interests.", "descript": "\n                  With the ever-increased greenhouse effect and energy crisis, developing novel photocatalysts to realize high-efficient solar-driven chemicals/fuel production is of great scientific and practical significance. Recently, single-atom photocatalysts (SAPs) are promising catalysts with maximized metal dispersion and tuneable coordination environments. SAPs exhibit boosted photocatalytic performance by enhancing optical response, facilitating charge carrier transfer behaviors or directly manipulating surface reaction processes. In this regard, this article systematically reviews the state-of-the-art progress in the development and application of SAPs, especially the mechanism and performance of SAPs on various reaction processes. Some future challenges and potential research directions over SAPs are outlined at the final stage.\n               "}
{"full_text": "Since Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n intermetallic compound (IMC) is utilized in the preparation of Raney-type Ni catalysts, it is considered as an industrially important material and has been highlighted in Intermetallic Materials Processing in Relation to Earth and Space Solidification (IMPRESS) project \u00a0[1\u201313]. Another important application of Al-rich IMCs is in the use of metal surface coatings due to the easier tendency to form Al\n\n\n\n2\n\n\nO\n\n\n\n3\n\n\n thin films\u00a0[14]. Despite the significant application potentials of Al-rich Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC, there have been less research about this IMC phase compared to its Ni-rich counterparts \u00a0[7,15]. Industrial applications such as the preparation procedure of Raney-type Ni catalysts by Al-leaching from Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC, and development of metallic surface coatings need a better understanding of interdiffusion effects and IMC growth at the interface of LIQUID and face centered cubic (FCC) phases\u00a0[16].\nAt temperatures equal to or greater than 1123.15 K, Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n intermetallic compound is formed at the interface between Al-rich LIQUID and Ni-rich FCC phases. Designing experiments at such high temperature is quite complicated. Because of the opaque nature of Al\u2013Ni alloy, it is quite difficult to study the evolution of the interfacial IMC phases during the reactive wetting solely from experiments\u00a0[17]. In this context, multiscale computational models are essential tools to obtain a better understanding of the mechanisms of IMC growth at the interface.\nA major challenge in simulating the growth behavior of Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n grains between the LIQUID and FCC phases is the uncertainty regarding the values of interfacial energies \n\n\n\u03c3\n\n\ni\n/\nj\n\n\n between two adjacent phases i and j. While some information is available on the interface properties for Ni-rich IMC phases\u00a0[18,19], there is little information regarding the experimental or theoretical values of the interfacial energies (\n\n\n\u03c3\n\n\nf\nc\nc\n/\ni\nm\nc\n\n\n, \n\n\n\u03c3\n\n\nf\nc\nc\n/\nl\ni\nq\n\n\n, \n\n\n\u03c3\n\n\ni\nm\nc\n/\ni\nm\nc\n\n\n, \n\n\n\u03c3\n\n\ni\nm\nc\n/\nl\ni\nq\n\n\n) for the interfaces involving Al-rich IMC phases in Al\u2013Ni system. As the Al-rich IMCs are formed during reactive wetting at higher temperatures, it is difficult to determine the interfacial energies using conventional experiments. Alternatively, the current state-of-the-art of quantitative phase field models makes it possible to determine properties such as liquid diffusion coefficient and interface energies, provided the results can be compared with experimental observations directly related to the effects of these material parameters\u00a0[20]. However, when considering IMC growth at material interfaces, too many different types of interfaces are involved to be able to determine the individual properties of the interfaces based on phase-field simulations. Therefore, in this work, interfacial energies are calculated using molecular dynamics (MD). As it is quite challenging to upscale MD simulations to reach the length scales of macroscale experiments, the mesoscale phase field model can be used to scale bridge results from MD computations and the continuum scale. To be specific, the interfacial energies computed from MD simulations are utilized in multi-phase field simulations for the study of the growth of Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC grains at the LIQUID/FCC interface.Numerous studies have been devoted to molecular dynamics (MD) simulations that can measure the interfacial energies between different materials \u00a0[21\u201324]. For instance, Hu et\u00a0al.\u00a0[21] performed MD simulations to measure the solid/liquid interfacial energy of uranium. Critical nucleation method (CNM) and capillary fluctuation method (CFM) were used to measure interfacial energy. Results show good agreement between these methods (CNM and CFM). Benedek et\u00a0al. \u00a0[22] performed atomistic simulations in heterogeneous interfaces with due consideration to the role of misfit. The influence of inter-diffusion and reaction layer was not considered, hence not promising for reactive systems at high temperatures. The interfacial energy also can be represented by calculating the work needed to create two free surfaces, which is the work of separation. This method has been utilized in the work by Yang et\u00a0al.\u00a0[23] and Gong et\u00a0al.\u00a0[24].Bhaskar\u00a0[25] has studied the precipitate coarsening in ternary Ni\u2013Al\u2013Mo alloys using phase field method. The model is based on constant interfacial energy (CIE) formulation i.e.\u00a0a single constant value of interfacial energy (\n\u03c3\n) is employed for the determination of model parameters of a given phase field simulation. For a total of 6 phase field simulations performed with 6 different values of interfacial energy, it has been revealed that the coarsening rate is increased with the increase in the value of \n\u03c3\n. This illustrates that interfaces with different physical structure will evolve differently and cannot be represented by a single value of interfacial energy. Thus, in a multi-phase field system characterized with distinct interfaces, it is very important to account for the effect of the role of heterogeneity caused by the differing magnitude of the interfacial energies. The influence of interfacial energies in the pattern formation during thin film growth has been assessed using phase field method in work of Khanna and Choudhury \u00a0[26]. However, in their work, the interfacial energy has been utilized directly as a model parameter in the gradient term of the phase field equation.The present study is designed on the framework of multiscale simulations. The interfacial energies of the FCC/IMC, FCC/LIQUID, IMC/IMC and IMC/LIQUID interfaces are first computed at \nT\n = 1173.15 K using MD simulations. Subsequently, atomistically informed multi-phase field simulations of the growth of Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC grains at the interface of Al-rich LIQUID and Ni-rich FCC phases are performed at \nT\n = 1173.15 K. The MD computed interfacial energies are supplied into the finite element method (FEM) based phase field simulations through the appropriate model parameter choice. At \nT\n = 1173.15 K, Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n is the only stable IMC phase between LIQUID and FCC. Therefore, from now onwards the term IMC will refer to AL3NI2 phase.Interfacial energy \n\n\n\u03c3\n\n\ni\n/\nj\n\n\n between phases i and j is a physical quantity of interest for materials scientists and engineers. However, we did not find any experimental work measuring the \n\n\n\u03c3\n\n\ni\n/\nj\n\n\n at such an elevated temperature as the temperature of interest for our study, namely around \n\nT\n=\n1173\n\nK\n\n. MD simulation is considered to be the most relevant approach to compute the interfacial properties at the nanoscale, and is therefore chosen in this study for the calculation of the interfacial energies.\n\nThe Embedded Atom Method (EAM) potential developed by Zhou et\u00a0al.\u00a0[27,28] has been employed for all the MD simulations in this work to describe the interactions between the Al and Ni atoms.In order to set up the MD model, slabs of FCC Al, FCC Ni and Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC are defined with an identical geometry (length = 70\u00a0\u00c5, breadth = 70\u00a0\u00c5, and height = 90\u00a0\u00c5) at room temperature (298.15 K). The Ni Slab consists of 42,865 Ni atoms only, the IMC slab is composed of 20,376 Al + 13,090 Ni atoms, whereas the Al slab is defined with 27,562 Al atoms only. Then these slabs were heated to the elevated temperature of 1173.15 K at a linear heating rate of 0.01 K/fs \u00a0[29], and after reaching this temperature each slab is relaxed for 1\u00a0ns. Then this is followed by the procedure of uniting two slabs. When two slabs of different materials are joined as shown in Fig.\u00a01, an interface will be created between the materials. Fig.\u00a01 illustrates the procedure for the construction of a solid/solid Ni/Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n interface. When two slabs of a same material are united to produce a slab, as shown in Fig.\u00a02, a union slab of single phase is created. In Fig.\u00a02, both Ni and Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n are in solid state at \nT\n = 1173.15 K. The illustrated methodology can thus be applied to compute the surface energies of these solids. We ignored possible variations in interface energy with orientation of the interface plan to restrict the computational work. Al is in liquid state at this temperature. For liquid, surface energy and surface tension are equal. Therefore, for Al, instead of preparing the union of two heated Al slabs, it is opted to apply a mechanical approach\u00a0[30] to compute the surface tension based on simulations for a single heated slab.At 1173.15 K, only Ni retains its FCC crystal structure, whereas Al transforms into liquid. Since the properties are computed at this temperature, the subscripts fcc, liq and imc appearing with mathematical quantities nowonwards refer to FCC (pure Ni or Ni-rich), LIQUID (pure Al or Al-rich) and Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC phases. The energy needed to create an interface between two different phases (i and j) at a given temperature is called the interfacial energy (\n\n\n\u03c3\n\n\ni\n/\nj\n\n\n) and is calculated by the following equation \u00a0[31]. \n\n(1)\n\n\n\n\n\u03c3\n\n\ni\n/\nj\n\n\n=\n\n\n\n\nE\n\n\ni\n/\nj\n\n\n\u2212\n\n\nE\n\n\ni\n\n\n\u2212\n\n\nE\n\n\nj\n\n\n\n\n2\n\n\nS\n\n\ni\nn\nt\n\n\n\n\n\n\n\nwhere, S\n\n\n\ni\nn\nt\n\n\n is the interface area in the simulated system. It is equal to 70\u00a0\u00c5 \n\u00d7\n 70\u00a0\u00c5. For FCC/IMC interface, the interfacial energy is denoted as \n\n\n\u03c3\n\n\nf\nc\nc\n/\ni\nm\nc\n\n\n. E\n\n\n\nf\nc\nc\n/\ni\nm\nc\n\n\n refers to the time-averaged total energy obtained for a union material as depicted in Fig.\u00a01, whereas E\n\n\n\nf\nc\nc\n\n\n and E\n\n\n\ni\nm\nc\n\n\n are the time-averaged total energy of the single phase union materials as shown in Fig.\u00a02. For the computation of \n\n\n\u03c3\n\n\nl\ni\nq\n/\ni\nm\nc\n\n\n, the interface energy of IMC/LIQ interface, E\n\n\n\nl\ni\nq\n/\ni\nm\nc\n\n\n and E\n\n\n\ni\nm\nc\n\n\n are obtained in the same way as just discussed for the FCC/IMC interface, while the time-averaged total energy E\n\n\n\nl\ni\nq\n\n\n is computed using the mechanical approach for surface tension determination. While the interfacial energy corresponding to i and j phases is denoted as \n\n\n\u03c3\n\n\ni\n/\nj\n\n\n, the surface energy for solid phase i or surface tension of a liquid phase is denoted as \n\n\n\u03c3\n\n\ni\n\n\n.Following the mechanical approach, the surface tension (\n\n\n\u03c3\n\n\nA\nl\n\n\n) of liquid Al slab can be expressed by the following equation: \n\n(2)\n\n\n\n\n\u03c3\n\n\nA\nl\n\n\n=\n\n\n1\n\n\n2\n\n\n\n\n\u222b\n\n\n\u2212\n\u221e\n\n\n\u221e\n\n\n\n[\n\n\nP\n\n\nN\n\n\n\n(\nz\n)\n\n\u2212\n\n\nP\n\n\nT\n\n\n\n(\nz\n)\n\n]\n\nd\nz\n\n\n\nIn Eq.\u00a0(2), \n\n\n\nP\n\n\nN\n\n\n\n(\nz\n)\n\n\n and \n\n\n\nP\n\n\nT\n\n\n\n(\nz\n)\n\n\n are, respectively, the normal and tangential components of pressure to the interfaces. The \nz\n-axis is defined as the normal to the base area (x\u2013y plane) of the slab. The total height of the slab in this z-dimension is divided into layers, each of width dz. In a 3D cartesian co-ordinate system, the normal pressure component \n\n\n\nP\n\n\nN\n\n\n\n(\nz\n)\n\n\n can be alternatively expressed as \n\n\n\nP\n\n\nz\nz\n\n\n\n(\nz\n)\n\n\n whereas the average tangential pressure component is defined by \n\n\n\n1\n\n\n2\n\n\n\n(\n\n\nP\n\n\nx\nx\n\n\n\n(\nz\n)\n\n+\n\n\nP\n\n\ny\ny\n\n\n\n(\nz\n)\n\n)\n\n\n.The MD simulation is performed in LAMMPS software \u00a0[32]. The Open Visualization Tool (OVITO) \u00a0[33] has been utilized for the visualization of the atoms and their positions. Periodic boundary conditions were applied in \nx\n and \ny\n directions whereas non-periodic boundary conditions were applied in the \nz\n-direction. Canonical ensemble (NVT) was used to control the temperature of the systems. The velocity-Verlet algorithm\u00a0[34] has been implemented to solve Newton\n\n\n\n\u2032\n\n\ns equation of motion at each time step. A time step of 1 fs was taken.To validate the MD calculated values, experimental data for surface tension of liquid Al (\n\n\n\u03c3\n\n\nl\ni\nq\n\n\n or \n\n\n\u03c3\n\n\nA\nl\n\n\n) and surface energy of FCC Ni (\n\n\n\u03c3\n\n\nN\ni\n\n\n) were collected. Pendant drop method\u00a0[35] is one of the most commonly utilized approaches for surface tension measurement. The exact details of the experimental steps required to determine the surface tension of pure Al using this method at different temperatures in ultra-high vacuum condition (7\u00a0\u00d7\u00a010\u22129-9\u00a0\u00d7\u00a010\u22129 Pa) has been outlined in Sun et\u00a0al.\u00a0[36]. For the present work, the average surface tension values obtained at two temperature, namely, 1023.15 K and 1123.15 K are considered. Moreover, experimental values of surface energy of pure Ni reported in the literature [37\u201340] were used for comparison with the calculated data.\n\nThe MD computed values of surface energy of FCC Ni (\n\n\n\u03c3\n\n\nf\nc\nc\n\n\n) for \nT\n = 1023.15 K and 1123.15 K, are presented along with the available experimental values in Fig.\u00a03. The uppermost value of surface energy among the experimental data is 3.7 J/m\n\n\n\n2\n\n\n corresponding to a temperature \nT\n = 298.15 K \u00a0[37]. The lowest measured value is 1.78 J/m\n\n\n\n2\n\n\n corresponding to a temperature \n\nT\n=\n1573\n\n K \u00a0[40]. Our simulation results are within these two extreme values and are thus considered acceptable.\nFig.\u00a04 shows surface tension data against temperature for liquid Al. At \nT\n = 1023.15 K, the surface tension of Al as obtained from experiments is 0.903 N/m\u00a0[36], while the MD simulation yields \n\n\n\u03c3\n\n\nl\ni\nq\n\n\n = 0.909 N/m at \nT\n = 1000 K. The experimental \n\n\n\u03c3\n\n\nl\ni\nq\n\n\n at \nT\n = 1123.15 K is 0.883 N/m\u00a0[36] and the computed \n\n\n\u03c3\n\n\nl\ni\nq\n\n\n at \n\nT\n=\n1200\n\n K is 0.883 N/m. At 1023.15 K, the percentage error of the numerically computed \n\n\n\u03c3\n\n\nl\ni\nq\n\n\n values with respect to the experimental value is obtained as 0.38% whereas at a higher temperature of 1123.15 K, the numerical model deviates from the experimental value by an error of 1.16%. The deviations between the experimental and computational values is very low, indicating that the MD computed data, using the considered interatomic potential and for the considered system, are most probably reliable even from a quantitative point of view.\nThe above benchmark studies thus confirm the strength and acceptability of Zhou interatomic potential for describing the surface and interface properties in Al\u2013Ni materials systems.The computed results of time-averaged total energy (E\n\n\n\ni\n\n\n for single phase or E\n\n\n\ni\n/\nj\n\n\n for two phase slabs) at \nT\n = 1173.15 K are presented in Table\u00a01. The values of interfacial energies \n\n\n\u03c3\n\n\ni\n/\nj\n\n\n obtained from Eq.\u00a0(1) are listed in Table\u00a02.The MD calculation has demonstrated that the various interfaces in the Al\u2013Ni system at \nT\n = 1173.15 K, have different interfacial energies. Hence, for the accurate understanding of the spatio-temporal dynamics of the AL3NI2 IMC grains at the interface of LIQUID (Al-rich) and FCC (Ni-rich) phases, it is essential to develop a mesoscale phase field model that has the capability to address the differences in interfacial energies. Fig.\u00a05 shows a initial condition (IC) of a 2D computational model consisting of LIQUID phase (upper part), FCC phase (lower part) and 16 square AL3NI2 IMC grains at the LIQUID/FCC interface at \nT\n = 1173.15 K. In this study, the width of the computational domain is 1200\u00a0nm and height is 700\u00a0nm. The length of the squares representing the IC of IMC grains is maintained within a range of 60\u201380\u00a0nm. The total area of IMC phase (sum of area of individual grains) at \n\nt\n=\n0\n\n is equal to 7.91E+04 nm\n\n\n\n2\n\n\n. The initial areas of FCC and LIQUID phase are 3.015E+05 nm\n\n\n\n2\n\n\n and 4.59E+05 nm\n\n\n\n2\n\n\n respectively.\nThe phases of the computational domain are represented by a set of non-conserved order parameters (\n\n\n\u03b7\n\n\ni\n\n\n) in such a way that \n\n\n\u03b7\n\n\ni\n\n\n = 1 inside a phase i and 0 outside it. The composition of a single component is sufficient to numerically represent the composition information in a binary Al\u2013Ni system. It is chosen to represent the composition by the mole-fraction of Al, represented by \nc\n, and which is a conserved variable. We use \nc\n to represent the global composition at each position in the computational domain. Besides, c\n\n\n\nf\nc\nc\n\n\n, c\n\n\n\ni\nm\nc\n\n\n and c\n\n\n\nl\ni\nq\n\n\n are introduced to represent the mole-fraction of Al in FCC, IMC and LIQUID phases, respectively.\n\n\n\n\n\nIn a multi-phase system with possibly more than one grain per phase, the material properties at the interface region between two different phases are interpolated in a thermodynamically consistent manner using the following function \u00a0[41]:\n\n\ni\nAL3NI2 (IMC) phase having q grains per phase \n\n(3)\n\n\n\n\nh\n\n\ni\n\n\n=\n\n\n\n\n\u2211\n\n\nj\n=\n1\n\n\nq\n\n\n\n\n\u03b7\n\n\ni\n,\nj\n\n\n2\n\n\n\n\n\n\n\u2211\n\n\nk\n=\n1\n\n\nN\n\n\n\n\n\u03b7\n\n\nk\n\n\n2\n\n\n\n\n\n\n\n\n\n\n\nii\nFCC and LIQUID phases having 1 grain per phase \n\n(4)\n\n\n\n\nh\n\n\ni\n\n\n=\n\n\n\n\n\u03b7\n\n\ni\n\n\n2\n\n\n\n\n\n\n\u2211\n\n\nk\n=\n1\n\n\nN\n\n\n\n\n\u03b7\n\n\nk\n\n\n2\n\n\n\n\n\n\n\n\n\n\n\nAL3NI2 (IMC) phase having q grains per phase \n\n(3)\n\n\n\n\nh\n\n\ni\n\n\n=\n\n\n\n\n\u2211\n\n\nj\n=\n1\n\n\nq\n\n\n\n\n\u03b7\n\n\ni\n,\nj\n\n\n2\n\n\n\n\n\n\n\u2211\n\n\nk\n=\n1\n\n\nN\n\n\n\n\n\u03b7\n\n\nk\n\n\n2\n\n\n\n\n\n\n\n\nFCC and LIQUID phases having 1 grain per phase \n\n(4)\n\n\n\n\nh\n\n\ni\n\n\n=\n\n\n\n\n\u03b7\n\n\ni\n\n\n2\n\n\n\n\n\n\n\u2211\n\n\nk\n=\n1\n\n\nN\n\n\n\n\n\u03b7\n\n\nk\n\n\n2\n\n\n\n\n\n\n\n\nwhere i represents the phase and \n\n\n\u03b7\n\n\ni\n,\nj\n\n\n are the order parameters representing the IMC grains and \n\n\n\u03b7\n\n\ni\n\n\n are the order parameters representing FCC phase or LIQUID phase. In the current model consisting of 3 phases and represented as in Fig.\u00a05, the FCC phase and LIQUID phase are each represented by a single grain. Whereas the AL3NI2 IMC phase is assumed to initially consist of 16 grains and for this phase q = 16. Thus, the total number of order parameters required in the model, denoted by \nN\n in Eq.\u00a0(4), is 18. The computational model is thus for a 3-phase 18-order parameter system.The driving force for microstructural evolution in the computational model is derived from the free energy of the system. The total free energy of the thermodynamic system \n\n\nF\n\n\nt\no\nt\na\nl\n\n\n can be described as the sum of bulk free energy (\n\n\nF\n\n\nb\nu\nl\nk\n\n\n) and interfacial free energy (\n\n\nF\n\n\ni\nn\nt\n\n\n) as following: \n\n(5)\n\n\n\n\nF\n\n\nt\no\nt\na\nl\n\n\n=\n\n\nF\n\n\nb\nu\nl\nk\n\n\n+\n\n\nF\n\n\ni\nn\nt\n\n\n=\n\n\n\u222b\n\n\nV\n\n\n\n(\n\n\nf\n\n\nb\nu\nl\nk\n\n\n+\n\n\nf\n\n\ni\nn\nt\n\n\n)\n\n\n\nd\nV\n\n\n\nThe bulk free energy per unit molar volume, \n\n\nf\n\n\nb\nu\nl\nk\n\n\n, is expressed as following. \n\n(6)\n\n\n\n\nf\n\n\nb\nu\nl\nk\n\n\n=\n\n\n\u2211\n\n\ni\n\n\n\n\nh\n\n\ni\n\n\n\n\nf\n\n\nc\nh\ne\nm\ni\nc\na\nl\n\n\ni\n\n\n=\n\n\n\u2211\n\n\ni\n\n\n\n\nh\n\n\ni\n\n\n\n\n\n\nG\n\n\nm\n\n\ni\n\n\n\n\n\n\nV\n\n\nm\n\n\ni\n\n\n\n\n\n\n\n\nIn Eq.\u00a0(6), V\n\n\n\nm\n\n\ni\n\n\n is the molar volume of phase i\u00a0[42]. In the present study, the molar volume of FCC, IMC and LIQUID phases are taken equal and are assigned the value of 11.4\u00a0cm\n\n\n\n3\n\n\nmol\u22121. The molar Gibbs free energy of each phase, denoted by G\n\n\n\nm\n\n\ni\n\n\n in the equation, is considered to have a parabolic composition dependence and is represented by the following expression\u00a0[43]: \n\n(7)\n\n\n\n\nG\n\n\nm\n\n\ni\n\n\n=\n\n\n\n\nA\n\n\ni\n\n\n\n\n2\n\n\n\n\n\n(\n\n\nc\n\n\ni\n\n\n\u2212\n\n\nc\n\n\ne\nq\n\n\ni\n\n\n)\n\n\n\n2\n\n\n+\n\n\nB\n\n\ni\n\n\n\n(\n\n\nc\n\n\ni\n\n\n\u2212\n\n\nc\n\n\ne\nq\n\n\ni\n\n\n)\n\n+\n\n\nC\n\n\ni\n\n\n.\n\n\n\n\nThe fitted parabolic Gibbs free energy curves of the phases are depicted in Fig.\u00a06(a) for \nT\n = 1173.15 K. Table\u00a03 presents the coefficients A\n\n\n\ni\n\n\n, B\n\n\n\ni\n\n\n and C\n\n\n\ni\n\n\n for the FCC, AL3NI2 (IMC) and LIQUID phases. In Fig.\u00a06(b), a sketch is shown to illustrate how the projected values for the equilibrium compositions of co-existing phases are obtained using the common tangent construction. The composition (\n\n\nc\n\n\nf\nc\nc\n,\ne\nq\n,\ni\nm\nc\n\n\nf\nc\nc\n\n\n) at which FCC phase is in equilibrium with AL3NI2 IMC phase is 0.16245, and the composition (\n\n\nc\n\n\ni\nm\nc\n,\ne\nq\n,\nf\nc\nc\n\n\ni\nm\nc\n\n\n) of the IMC in equilibrium with FCC is 0.58269. Similarly, the equilibrium compositions of IMC and LIQUID when in equilibrium with each other are \n\n\n\nc\n\n\ni\nm\nc\n,\ne\nq\n,\nl\ni\nq\n\n\ni\nm\nc\n\n\n=\n0\n.\n6275\n\n and \n\n\n\nc\n\n\nl\ni\nq\n,\ne\nq\n,\ni\nm\nc\n\n\ni\nm\nc\n\n\n=\n0\n.\n8967\n\n, respectively. The coefficients of the parabolic Gibbs energy functions were chosen so that the equilibrium compositions agree well with those obtained from the phase diagram calculated with Thermo-Calc software using TCNI8 database.The interfacial free energy density (f\n\n\n\ni\nn\nt\n\n\n) of Eq.\u00a0(5) can be mathematically decomposed into the following parts\u00a0[44\u201346]\n\n\n(8)\n\n\n\n\nf\n\n\ni\nn\nt\n\n\n=\n\n\nf\n\n\no\n\n\n+\n\n\nf\n\n\ng\nr\na\nd\n\n\n\n\n\nwhere, \n\n(9)\n\n\n\n\nf\n\n\no\n\n\n=\nm\n\n(\n\n\n\u2211\n\n\ni\n\n\n\n(\n\n\n\n\n\n(\n\n\n\u03b7\n\n\ni\n\n\n)\n\n\n\n4\n\n\n\n\n4\n\n\n\u2212\n\n\n\n\n\n(\n\n\n\u03b7\n\n\ni\n\n\n)\n\n\n\n2\n\n\n\n\n2\n\n\n)\n\n+\n\n\n\u2211\n\n\ni\n\n\n\n\n\u2211\n\n\nj\n\u2260\ni\n\n\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n\n\n\n\n\u03b7\n\n\ni\n\n\n\n\n2\n\n\n\n\n\n\n\u03b7\n\n\nj\n\n\n\n\n2\n\n\n+\n\n\n1\n\n\n4\n\n\n)\n\n\n\n\nand, \n\n(10)\n\n\n\n\nf\n\n\ng\nr\na\nd\n\n\n=\n\n\n\n\n\u03ba\n\n\ni\n,\nj\n\n\n\n\n2\n\n\n\n\n\u2211\n\n\ni\n\n\n\n\n\n(\n\u2207\n\n\n\u03b7\n\n\ni\n\n\n)\n\n\n\n2\n\n\n\n\n\nThus, the quantity f0 contains two model parameters, namely, m and \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n. The gradient free energy density, denoted as f\n\n\n\ng\nr\na\nd\n\n\n in Eq.\u00a0(8) consists of the model parameters \n\n\n\u03ba\n\n\ni\n,\nj\n\n\n.It is important to note that diffuse interface width (\n\u03b4\n) and interfacial energy (\n\n\n\u03c3\n\n\ni\n/\nj\n\n\n) are both related to model parameters \nm\n, \n\n\n\u03ba\n\n\ni\n,\nj\n\n\n and \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n. When the interface energy is the same for all the interfaces, \n\n\n\n\u03ba\n\n\ni\n,\nj\n\n\n=\n\u03ba\n,\n\u2200\ni\n,\nj\n\n and \n\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n=\n\u03b3\n,\n\u2200\ni\n,\nj\n\n, and only three parameters \nm\n, \n\u03ba\n and \n\u03b3\n have to be defined for a given value of \n\u03b4\n and \n\u03c3\n \u00a0[43]. However, as outlined by the atomistic calculation in Section\u00a02, the interfacial energy is different for different interfaces. When considering different interfacial energies, the model parameters need to be calculated following the procedure outlined in Moelans et\u00a0al.\u00a0[44]. More details of the calculation procedure of the parameters \nm\n, \n\n\n\u03ba\n\n\ni\n,\nj\n\n\n and \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n is presented in Section\u00a03.2.The evolution of the conserved variable (c) is expressed by the following equation: \n\n(11)\n\n\n\n\n\u2202\nc\n\n\n\u2202\nt\n\n\n=\n\u2207\n.\n\n\n\n\n\u2211\n\n\ni\n\n\n\n\nh\n\n\ni\n\n\n\n\nM\n\n\ni\n\n\n\u2207\n\n(\n\n\n\n\u2202\n\n\nf\n\n\nb\nu\nl\nk\n\n\n\n\n\u2202\n\n\nc\n\n\ni\n\n\n\n\n\n)\n\n\n\n\n\n\nwhere \ni\n goes over the three phases and wherein the Cahn\u2013Hilliard mobility M\n\n\n\ni\n\n\n is related to phase diffusivity (D\n\n\n\ni\n\n\n), thermodynamic factor (A\n\n\n\n\ni\n\n\n=\n\n\n\n\n\u2202\n\n\n2\n\n\n\n\nf\n\n\nb\nu\nl\nk\n\n\n\n\n\u2202\n\n\n\n(\n\n\nc\n\n\ni\n\n\n)\n\n\n\n2\n\n\n\n\n\n) and grain-boundary mobility (M\n\n\n\ng\nb\n\n\n) as following : \n\n(12)\n\n\n\n\nM\n\n\ni\n\n\n=\n\n\n\n\nD\n\n\ni\n\n\n\n\n\n\nA\n\n\ni\n\n\n\n\n+\n\n\n\u2211\n\n\nj\n\u2260\ni\n\n\n\n\nh\n\n\nj\n\n\n\n\nM\n\n\ng\nb\n\n\n\n\n\nThe grain boundary mobility (M\n\n\n\ng\nb\n\n\n) of Eq.\u00a0(12) is defined as following \n\n(13)\n\n\n\n\nM\n\n\ng\nb\n\n\n=\n\n\n3\n\n\n\u03b4\n\n\ng\nb\n\n\n\n\nD\n\n\ng\nb\n\n\n\n\n\u03b4\n\n(\n\n\nh\n\n\ni\n\n\n\n\nA\n\n\ni\n\n\n+\n\n\nh\n\n\nj\n\n\n\n\nA\n\n\nj\n\n\n)\n\n\n\n\n\n\nwith D\n\n\n\ng\nb\n\n\n being the grain-boundary diffusivity and \n\n\n\u03b4\n\n\ng\nb\n\n\n = 0.5\u00a0nm is the width of the grain boundary channel. For a binary system, there is only one interdiffusion coefficient for each phase, describing the intermixing of both elements.The phase evolution is characterized by the following equation representing the spatio-temporal evolution of non-conserved order parameters : \n\n(14)\n\n\n\n\n\u2202\n\n\n\u03b7\n\n\ni\n\n\n\n\n\u2202\nt\n\n\n=\n\u2212\nL\n\n(\n\n\n\u03b7\n\n\n1\n\n\n,\n\u2026\n,\n\n\n\u03b7\n\n\nN\n\n\n)\n\n\n(\n\n\n\n\u2202\n\n\nf\n\n\nb\nu\nl\nk\n\n\n\n\n\u2202\n\n\n\u03b7\n\n\ni\n\n\n\n\n+\n\n\n\u2202\n\n\nf\n\n\no\n\n\n\n\n\u2202\n\n\n\u03b7\n\n\ni\n\n\n\n\n\u2212\n\u2207\n.\n\n\n\u2202\n\n\nf\n\n\ng\nr\na\nd\n\n\n\n\n\u2202\n\u2207\n\n\n\u03b7\n\n\ni\n\n\n\n\n\n)\n\n\n\n\nIn Eq.\u00a0(14), \n\nL\n\n(\n\n\n\u03b7\n\n\n1\n\n\n,\n\u2026\n,\n\n\n\u03b7\n\n\nN\n\n\n)\n\n\n is mathematically defined by the following expression \n\n(15)\n\n\nL\n=\n\n\n\n\n\u2211\n\n\ni\n=\n1\n\n\nN\n\n\n\n\n\u2211\n\n\nj\n>\n1\n\n\nN\n\n\n\n(\n\n\nL\n\n\ni\n,\nj\n\n\n\n\n\u03b7\n\n\ni\n\n\n2\n\n\n\n\n\u03b7\n\n\nj\n\n\n2\n\n\n)\n\n\n\n\n\n\u2211\n\n\ni\n=\n1\n\n\nN\n\n\n\n\n\u2211\n\n\nj\n>\n1\n\n\nN\n\n\n\n\n\u03b7\n\n\ni\n\n\n2\n\n\n\n\n\u03b7\n\n\nj\n\n\n2\n\n\n\n\n\n\n\nThe kinetic coefficients L\n\n\n\ni\n,\nj\n\n\n of Eq.\u00a0(15) are determined as described in Section\u00a03.2.For a system with varying interfacial energies (VIE) at different interfaces, the following expressions relate the different interfacial energy (\n\n\n\u03c3\n\n\ni\n/\nj\n\n\n) and diffuse interface width (\n\u03b4\n) to the phase field coefficients \n\n\n\u03ba\n\n\ni\n,\nj\n\n\n, \nm\n and \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n. \u00a0[44,45]: \n\n(16)\n\n\n\n\n\u03c3\n\n\ni\n/\nj\n\n\n=\ng\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n\n\n\n\u03ba\n\n\ni\n,\nj\n\n\nm\n\n\n\n\n\n\n\n\n(17)\n\n\n\u03b4\n=\n\n\n\n\n\n\n\u03ba\n\n\ni\n,\nj\n\n\n\n\nm\n\n\nf\n\n\n0\n,\nm\na\nx\n\n\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n\n\n\n\n\n\nThe two functions of \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n namely, \n\ng\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n and \n\n\n\nf\n\n\n0\n,\nm\na\nx\n\n\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n, present in Eqs.\u00a0(16) and (17) are evaluated following the procedure defined in\u00a0[44]. The interfacial mobilities of the different interfaces are defined using the following formulae:\n\n\ni\nFor grain boundaries between two grains of the same phase \n\n(18)\n\n\n\n\nL\n\n\ni\n,\nj\n\n\n=\n\n\n\n\n\u03bc\n\n\ni\n/\nj\n\n\ng\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n\n\u03b4\n\n\nf\n\n\n0\n,\nm\na\nx\n\n\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n\n\n\n\nwith \n\n\n\u03bc\n\n\ni\n/\nj\n\n\n the mobility of the grain boundary between grains \ni\n and \nj\n.\n\n\nii\nFor phase interfaces \n\n(19)\n\n\n\n\nL\n\n\ni\n,\nj\n\n\n=\n\n\n4\nm\n\n\n\n\n\u03ba\n\n\ni\n,\nj\n\n\n\n\n\u03bb\n\n\ni\n,\nj\n\n\n\n\n\n\n\nwhere, \n\n\n\n1\n\n\n\n\n\u03bb\n\n\ni\n,\nj\n\n\n\n\n=\n\n\n\n\nM\n\n\ni\n\n\n+\n\n\nM\n\n\nj\n\n\n\n\n2\n\n(\n\n\nc\n\n\ni\n,\ne\nq\n,\nj\n\n\ni\n\n\n\u2212\n\n\nc\n\n\nj\n,\ne\nq\n,\ni\n\n\nj\n\n\n)\n\n\n(\n\n\nc\n\n\ni\n,\ne\nq\n,\nk\n\n\ni\n\n\n\u2212\n\n\nc\n\n\nk\n,\ne\nq\n,\ni\n\n\nk\n\n\n)\n\n\n\n\n and \ni\n and \nj\n refer to the two phases in the neighboring grains.\n\n\nFor grain boundaries between two grains of the same phase \n\n(18)\n\n\n\n\nL\n\n\ni\n,\nj\n\n\n=\n\n\n\n\n\u03bc\n\n\ni\n/\nj\n\n\ng\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n\n\u03b4\n\n\nf\n\n\n0\n,\nm\na\nx\n\n\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n\n\n\n\nwith \n\n\n\u03bc\n\n\ni\n/\nj\n\n\n the mobility of the grain boundary between grains \ni\n and \nj\n.For phase interfaces \n\n(19)\n\n\n\n\nL\n\n\ni\n,\nj\n\n\n=\n\n\n4\nm\n\n\n\n\n\u03ba\n\n\ni\n,\nj\n\n\n\n\n\u03bb\n\n\ni\n,\nj\n\n\n\n\n\n\n\nwhere, \n\n\n\n1\n\n\n\n\n\u03bb\n\n\ni\n,\nj\n\n\n\n\n=\n\n\n\n\nM\n\n\ni\n\n\n+\n\n\nM\n\n\nj\n\n\n\n\n2\n\n(\n\n\nc\n\n\ni\n,\ne\nq\n,\nj\n\n\ni\n\n\n\u2212\n\n\nc\n\n\nj\n,\ne\nq\n,\ni\n\n\nj\n\n\n)\n\n\n(\n\n\nc\n\n\ni\n,\ne\nq\n,\nk\n\n\ni\n\n\n\u2212\n\n\nc\n\n\nk\n,\ne\nq\n,\ni\n\n\nk\n\n\n)\n\n\n\n\n and \ni\n and \nj\n refer to the two phases in the neighboring grains.The quantity \n\n\n\u03bb\n\n\ni\n,\nj\n\n\n corresponding to the interface of the phase i and phase j is a function of the equilibrium composition and diffusion mobility in phases i and j\u00a0[47]. With \n\u03b4\n = 25.0\u00a0nm, and \nm\n = 2.4\u00a0\u00d7\u00a0108 J/m\n\n\n\n3\n\n\n, the functional values of \n\ng\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n and \n\n\n\nf\n\n\n0\n,\nm\na\nx\n\n\n\n(\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n)\n\n\n are outlined. Then the model parameters \n\n\n\u03ba\n\n\ni\n,\nj\n\n\n, \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n and \n\n\nL\n\n\ni\n,\nj\n\n\n (for both grain boundaries and phase boundaries) are estimated corresponding to the different values of interfacial energies \n\n\n\u03c3\n\n\ni\n/\nj\n\n\n using the procedure outlined in \u00a0[44]. The magnitudes of these model parameters are provided in Table\u00a04. In order to set up the phase field model with a correct assignment of the model parameters \n\n\n\u03ba\n\n\ni\n,\nj\n\n\n, \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n and L\n\n\n\ni\n,\nj\n\n\n at the corresponding interfaces, the following expression is defined in the model: \n\n(20)\n\n\n\n\n\u03c8\n\n\nk\n\n\n=\n\n\n\n\n\u2211\n\n\ni\n=\n1\n\n\nN\n\n\n\n\n\u2211\n\n\nj\n>\n1\n\n\nN\n\n\n\n(\n\n\n\u03c8\n\n\ni\n,\nj\n\n\n\n\n\u03b7\n\n\ni\n\n\n2\n\n\n\n\n\u03b7\n\n\nj\n\n\n2\n\n\n)\n\n\n\n\n\n\u2211\n\n\ni\n=\n1\n\n\nN\n\n\n\n\n\u2211\n\n\nj\n>\n1\n\n\nN\n\n\n\n\n\u03b7\n\n\ni\n\n\n2\n\n\n\n\n\u03b7\n\n\nj\n\n\n2\n\n\n\n\n\n\n\nwhere, \n\n\n\u03c8\n\n\ni\n,\nj\n\n\n represents \n\n\n\u03ba\n\n\ni\n,\nj\n\n\n, \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n or L\n\n\n\ni\n,\nj\n\n\n. The phase dependent interfacial mobility has already been mathematically expressed in Eq.\u00a0(15).The values of interfacial energies at different types of interfaces, and their corresponding model parameters are discussed in the preceding section. In addition to these model parameters, the other parameters and material properties required as input in the phase field simulations are listed in Table\u00a05. The values for phase diffusivities used in the simulations are presented in the table and are obtained from\u00a0[48\u201351]. As mentioned earlier, the interfacial energy (\n\n\n\u03c3\n\n\ni\n/\nj\n\n\n) listed in the table are obtained from the molecular dynamics calculation of the present work. The energies of the boundaries between the IMC grains were not computed with the MD method, as it would have required a huge amount of compute time. Instead, we used the MD computed IMC surface energy as a first estimate for the grain boundary energies of the IMC phase (i.e.\u00a0we took \n\n\n\n\u03c3\n\n\ni\nm\nc\n/\ni\nm\nc\n\n\n=\n\n\n\u03c3\n\n\ni\nm\nc\n\n\n\n) in the phase-field model. The model parameters \n\n\n\u03ba\n\n\ni\n,\nj\n\n\n, \n\n\n\u03b3\n\n\ni\n,\nj\n\n\n, and \n\n\nL\n\n\ni\n,\nj\n\n\n will vary in the simulation domain to represent the different values of interfacial energies, and their values are mentioned in Table\u00a04.The non-conserved order parameters are initially described as a function of co-ordinates (x,y) to define the initial geometrical conditions of the phases represented in Fig.\u00a05. At t = 0, the mole fraction of Al in FCC, AL3NI2 and LIQUID phases are set as 0.15, 0.55 and 0.89, respectively. No flux boundary condition is applied for both c and \n\u03b7\n variables at the top and bottom boundaries of the computation domain. At the right and left sides of the domain, periodic boundary conditions are enforced for both types of variables.The partial differential Eq.\u00a0(11) and \u00a0(14) are solved using finite element method (FEM) in Multiphysics Object Oriented Simulation Environment (MOOSE) Framework \u00a0[52,53]. The GeneratedMesh object of MOOSE has been utilized to construct 2D mesh with QUAD4 elements for the initial geometry described in Fig. \u00a05. In the phase field simulation, the following assumptions have been made in accordance to Kim\u2013Kim\u2013Suzuki (KKS) model\u00a0[54] for the binary Al\u2013Ni system consisting of three phases (FCC, IMC and LIQUID).\n\n\ni\nChemical potential between any two adjacent phases is equal.\n\n\nii\nThe expression \n\nc\n=\n\n\n\u2211\n\n\ni\n\n\n\n\nh\n\n\ni\n\n\n\n\nc\n\n\ni\n\n\n\n relates the global composition of Al with its local phase compositions (\ni\n goes over the three phases).\n\n\nChemical potential between any two adjacent phases is equal.The expression \n\nc\n=\n\n\n\u2211\n\n\ni\n\n\n\n\nh\n\n\ni\n\n\n\n\nc\n\n\ni\n\n\n\n relates the global composition of Al with its local phase compositions (\ni\n goes over the three phases).\n\nWith the information of interfacial energies obtained from molecular dynamics simulation and having outlined the corresponding values of model parameters for the mesoscale phase field models, we can now present the results of a 2D grain growth simulation for Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC grains at Al/Ni interface for the VIE model. Fig.\u00a07 presents the evolution of the grains showing the structure at t = (i) 0.05, (ii) 0.1, (iii) 0.14 and (iv) 0.19 s. The interfacial Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC grains with a composition c \n\u2248\n 0.6, appear in red color according the color scale bar. The LIQUID (Al-rich) and FCC (Ni-rich) regions are depicted respectively by yellow and cyan colors. It is evident from the figure that as the time passes the IMC phase (total sum of IMC grains) area expands at the expense of LIQUID and FCC phases. At t = 0.05 s (Fig.\u00a07(i)), all the IMC grains (regardless of their initial sizes) have nearly equal exposure to the LIQUID phase at phase boundaries. But as the time passes the Ostwald ripening phenomenon influences the relative growth of grains with a different size. While all the grains tend to grow longitudinally at the expense of LIQUID and FCC phase, they are marked by differentiated lateral growths at the LIQ/IMC and FCC/IMC interfaces. At the LIQUID/IMC interface, the locally larger grains tend to expand laterally whereas the locally smaller grains show lateral constriction. As revealed clearly in the image at t = 0.19 s (Fig.\u00a07(iv)), the initially larger grains are facilitated to enlarge their contact with the liquid whereas the initially smaller grains lose their contact with the LIQUID phase. At FCC/IMC interfaces, on the contrary, the lateral dimension of the IMC grains hardly varies, regardless of their initial size. A clearly different morphology is thus observed at these two interfaces.In order to understand whether the choice of interfacial properties affects the growth pattern of phases in the Al/Ni system, it is necessary to compare the result of varying interface energy (VIE) model with those obtained using constant interfacial energy (CIE) model. The simulations in CIE formulation start with the same initial geometry as the VIE model, and use the same bulk material properties. Only the interfacial properties are taken differently. The interfacial properties of the CIE model are taken as given in Table\u00a06. With the interfacial energy for all the interfaces (denoted as \n\n\n\u03c3\n\n\nc\n\n\n) being assigned a value of 1.0 J/m\n\n\n\n2\n\n\n and the interface width taken as \n\u03b4\n = 25\u00a0nm, the model parameter \nm\n = 2.4E+08 J/m\n\n\n\n3\n\n\n in the CIE model, and this is equal to that of VIE model. However, in contrast to VIE model where \n\n{\n\n\n\u03ba\n\n\ni\n,\nj\n\n\n}\n\n and \n\n{\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n}\n\n have different values at various interfaces, the CIE model has a constant \n\u03ba\n and a constant \n\u03b3\n at all of the interfaces. As mentioned in the table, \n\u03ba\n= 1.875E\u221208 J/m and \n\u03b3\n = 1.5 at all of the interfaces of the CIE model. It is to be noted that the kinetic coefficients \n\n\nL\n\n\ni\n,\nj\n\n\n defined by Eqs. \u00a0(18) and \u00a0(19) are influenced by diffusion mobilities of grains/phases besides the interfacial energy values. The ratio of diffusivity (refer Table\u00a05) to thermodynamic factor A\n\n\n\ni\n\n\n(refer Tables 3) for LIQUID phase is 8.35\u00a0\u00d7\u00a010\u221221 m\n\n\n\n5\n\n\n/(J s) whereas this ratio for FCC phase is 6.14\u00a0\u00d7\u00a010\u221224 m\n\n\n\n5\n\n\n/(J s) for FCC phase. In context of model parameters shown in Tables 4 and \u00a06, all elements of \n\n{\n\n\n\u03ba\n\n\ni\n,\nj\n\n\n}\n\n vector are within the range (1.69\u20132.25)\n\u00d7\n10\u22128 J/m and the components of \n\n{\n\n\n\u03b3\n\n\ni\n,\nj\n\n\n}\n\n vector vary just between 1.32 and 1.98. Thus for the Al\u2013Ni system at 1173.15 K it can be inferred that the \n\n\nL\n\n\ni\n,\nj\n\n\n for interface i/j are solely dominated by the magnitude of diffusion mobilities of phase i and phase j, and are generally independent of the variation in interfacial energies.The morphology of a grain (G1) neighboring two smaller grains at its either sides is compared between the two models for t = 0.1375 s in Fig.\u00a08. The image (a) represents the simulation result of CIE model whereas the image (b) is of VIE model. It is interesting to note that the FCC/IMC interface moves relatively faster in the CIE model, and thus G1 grows more towards the bottom in the CIE model. Moreover, the IMC/LIQUID phase boundary length of the G1 is larger for the VIE model. Though not so distinct, the curvatures of the phase boundaries are definitely different for the two models.\nFig.\u00a09 presents a comparison of the evolution of phase areas for the two models. The AL3NI2 IMC grows at the expense of LIQUID and FCC phases in both cases. As shown in Fig.\u00a09(a), the area of FCC phase decreases at a faster rate for the CIE model compared to the VIE model. At t = 0.4 s, the area of FCC in CIE model is reduced to a value of 1.48E+05 nm\n\n\n\n2\n\n\n whereas FCC phase in VIE model attains an area of 1.84E+05 nm\n\n\n\n2\n\n\n. In contrary to this, the phase area of LIQUID phase for VIE model decreases at a faster rate than that of CIE model (Fig.\u00a09(b)). The LIQUID area in CIE model is lowered to a value of 2.64E+05 nm\n\n\n\n2\n\n\n at t = 0.41 s whereas the LIQUID phase in VIE model already shrinks to a value of 2.56E+05 nm\n\n\n\n2\n\n\n within the same time. This dynamics is consistent with the fact that the interfacial energy at FCC/IMC interface of VIE model is larger than \n\n\n\u03c3\n\n\nc\n\n\n (1.0 J/m\n\n\n\n2\n\n\n) whereas that for LIQUID/IMC interface of VIE model is smaller than \n\n\n\u03c3\n\n\nc\n\n\n. A larger interface energy results in a larger curvature driving force which counteracts the growth, since the center of curvature of the FCC/IMC and LIQUID/IMC interfaces is inside the IMC. Finally, as revealed by Fig.\u00a09(c), the total IMC phase area is observed to increase at a faster rate in the CIE model than in the VIE model. At t = 0.4 s, the corresponding IMC phase areas of CIE and VIE models are respectively 4.18E+05 and 3.97E+05 nm\n\n\n\n2\n\n\n. Since the deviation between CIE and VIE model is larger for FCC phase than for LIQUID, the IMC phase in CIE model grows slightly faster than in VIE model. From these observations, it is clear that the selection of interfacial properties influences the growth kinetics of the phases in the simulations.\nHence it is important to determine the interface properties accurately and use a phase-field model that can account for different interfacial energies. This highlights the need to formulate a phase-field model that can take into account different interface energies (VIE) for a realistic description of IMC morphology in multi-phase materials systems. Moreover, it is needed to determine interface properties accurately for all different types of interfaces.\n\nThe following conclusions can be derived from the study:\n\n\ni\nThis work employs the mesoscale multi-phase field method for mathematical description of the spatio-temporal dynamics of AL3NI2 IMC phase at the interface of LIQUID (Al-rich) and FCC (Ni-rich) phases. Sixteen square Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC grains were initially introduced at the interface of LIQUID and FCC phases and the phase field simulation is performed at 1173.15 K to understand the structural evolution of these interfacial IMC grains.\n\n\nii\nThe interfacial energies required to determine the model parameters of the VIE formulation based mesoscale phase field method, have been supplied through computations performed using nanoscale Molecular Dynamics. The MD computation yielded the interfacial energies of 1.2, 1.02 and 0.9 J/m2 at FCC/IMC, FCC/LIQUID and IMC/LIQUID interfaces respectively. For the IMC/IMC grain boundaries, a value of 0.957 J/m2 was used in the phase-field model. This value was estimated, taking it equal to the MD calculated surface energy for IMC, since computing the different grain boundary energies from MD would become highly complex. Additionally, a CIE formulation based phase field model is also designed corresponding to \n\u03b4\n and \n\n\n\u03c3\n\n\nc\n\n\n of 25\u00a0nm and 1.0 J/m\n\n\n\n2\n\n\n.\n\n\niii\nAs compared to the VIE model, phase area reduction for FCC phase is faster in the CIE model whereas the decrease in area of LIQUID phase is slower in the CIE model. Overall there is a net faster growth rate of IMC area in the CIE model.\n\n\niv\nThe choice of interfacial properties can thus ardently influence the size of bulk phases. The development of microstructure models with varying interface energies and calculation of interface energies from MD is thus important to allow for accurate microstructure prediction.\n\n\nv\nThe presentation of methodologies, namely, the utilization of EAM interatomic potential for the computation of surface as well as interface properties of Al\u2013Ni system at nanoscale, upscaling of these interfacial properties and subsequently combining them compatibly with other thermodynamic properties of bulk phases at mesoscale to simulate the structural evolution of Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n grains at the interface of Al-rich LIQUID and Ni-rich FCC phases, in this work, outline the basic cornerstone of virtual experiments of materials with reactive interfaces undergoing high temperature applications.\n\n\nThis work employs the mesoscale multi-phase field method for mathematical description of the spatio-temporal dynamics of AL3NI2 IMC phase at the interface of LIQUID (Al-rich) and FCC (Ni-rich) phases. Sixteen square Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n IMC grains were initially introduced at the interface of LIQUID and FCC phases and the phase field simulation is performed at 1173.15 K to understand the structural evolution of these interfacial IMC grains.The interfacial energies required to determine the model parameters of the VIE formulation based mesoscale phase field method, have been supplied through computations performed using nanoscale Molecular Dynamics. The MD computation yielded the interfacial energies of 1.2, 1.02 and 0.9 J/m2 at FCC/IMC, FCC/LIQUID and IMC/LIQUID interfaces respectively. For the IMC/IMC grain boundaries, a value of 0.957 J/m2 was used in the phase-field model. This value was estimated, taking it equal to the MD calculated surface energy for IMC, since computing the different grain boundary energies from MD would become highly complex. Additionally, a CIE formulation based phase field model is also designed corresponding to \n\u03b4\n and \n\n\n\u03c3\n\n\nc\n\n\n of 25\u00a0nm and 1.0 J/m\n\n\n\n2\n\n\n.As compared to the VIE model, phase area reduction for FCC phase is faster in the CIE model whereas the decrease in area of LIQUID phase is slower in the CIE model. Overall there is a net faster growth rate of IMC area in the CIE model.The choice of interfacial properties can thus ardently influence the size of bulk phases. The development of microstructure models with varying interface energies and calculation of interface energies from MD is thus important to allow for accurate microstructure prediction.The presentation of methodologies, namely, the utilization of EAM interatomic potential for the computation of surface as well as interface properties of Al\u2013Ni system at nanoscale, upscaling of these interfacial properties and subsequently combining them compatibly with other thermodynamic properties of bulk phases at mesoscale to simulate the structural evolution of Al\n\n\n\n3\n\n\nNi\n\n\n\n2\n\n\n grains at the interface of Al-rich LIQUID and Ni-rich FCC phases, in this work, outline the basic cornerstone of virtual experiments of materials with reactive interfaces undergoing high temperature applications.\nAnil Kunwar: Conceptualization, Methodology, Software, Investigation, Visualization, Validation, Formal analysis, Writing \u2013 original draft. Ensieh Yousefi: Software, Methodology, Investigation, Validation, Visualization, Formal analysis, Writing \u2013 review & editing. Xiaojing Zuo: Software, Validation, Writing \u2013 review & editing. Youqing Sun: Investigation, Validation, Writing \u2013 review & editing. David Seveno: Methodology, Validation, Supervision, Writing \u2013 review & editing, Resources, Funding acquisition. Muxing Guo: Validation, Supervision, Writing \u2013 review & editing, Resources, Funding acquisition. Nele Moelans: Conceptualization, Methodology, Validation, Supervision, Writing \u2013 review & editing, Resources, Funding acquisition.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.This work was supported by the KU Leuven Research Fund (C14/17/075). The computational resources and services used in this work were provided by the VSC (Flemish Supercomputer Center), funded by the Research Foundation - Flanders (FWO)\n and the Flemish Government - department EWI. Open access funding was provided by Silesian University of Technology.", "descript": "\n                  Considering its application in developing Raney-type Ni catalysts and in metal surface coatings, the study on the growth behavior of Al\n                        \n                           \n                           \n                              3\n                           \n                        \n                     Ni\n                        \n                           \n                           \n                              2\n                           \n                        \n                      intermetallic compound (IMC) at the Al/Ni material interface is of utmost importance. The present work integrates nanoscale molecular dynamics (MD) calculation with mesoscale phase field model for studying the interfacial phenomena associated with Al\n                        \n                           \n                           \n                              3\n                           \n                        \n                     Ni\n                        \n                           \n                           \n                              2\n                           \n                        \n                      growth in Al/Ni interface at 1173.15 K. The interfacial energies computed from MD are in the range 0.9\u20131.2 J/m\n                        \n                           \n                           \n                              2\n                           \n                        \n                      with FCC/IMC featuring as the interface with the largest value and IMC/LIQUID as the one with the lowest value. Phase field model parameters characterizing a varying interface energy formulation are established to simulate the 2D growth of interfacial IMC grains. With the help of an atomistically informed phase field model, it has been revealed that the phase areas and morphology are obviously sensitive to the interfacial properties. The methodologies and results of these multiscale simulations for IMC interfaced between Al and Ni microstructures offer the complementary and accelerated design route of in-silico studies for materials systems experimented at high temperature.\n               "}
{"full_text": "Heterocyclic compounds have gained great attention in organic synthesis. Five-membered heterocyclic compounds containing one heteroatom such as pyrrole, furan and thiophene are the most prominent scaffolds in organic chemistry having wide-spread biological activities [1]. Especially, pyrroles have an essential role in pharmaceuticals [2], agrochemicals [3], optoelectronic materials [4] and biologically active natural products [5]. Pyrrole shows notable pharmaceutical properties such as anti-oxidative [6], anti-fungal [7], anti-bacterial [8], anti-inflammatory [9] and ionotropic nature [10]. Moreover, pyrroles are the important building blocks for functional materials due to their occurrence, which exhibits inevitable properties [11]. Knorr [12], Paal-Knorr [13], and Hantzsch [14] reactions are the well-established classical synthetic strategies for pyrroles. However, new interesting and attractive methodologies have been introduced to synthesize substituted pyrroles. Over the past decades, plenty of methods have been developed for the construction of N-substituted pyrroles through modified Paal-Knorr or Clauson-Kaas reaction using acid catalysts [15] and metal catalysts such as Mg [16], Fe [17], Cu [18], Zr [19], Bi [20] and Ce [21]. In this context, we focused on the sustainable and environmentally benign protocol for synthetic reactions [22]. We extended our efforts on organic synthesis catalyzed by transition metals such as Mn [23], Fe [24], Co [25], Ni [26], Cu [27], Zn [28], etc. Very recently our group introduced a solvent-free manganese-catalyzed microwave-assisted version of Clauson-Kaas reaction (Scheme 1\n) [29]. To the best of our knowledge, no solvent-free zinc-catalyzed synthesis of N-aryl pyrroles through Clauson-Kaas reaction are reported yet. On comparison with other metals, zinc is the low-cost and less-toxic 3d series transition metal, which has a vital role in synthetic chemistry due to its relatively great abundance in earth and high concentration in ores. In addition, zinc has an attractive biological relevance and is an essential trace element in the human body [30]. Due to the relevance of zinc, in 2017, Tran and co-workers reported an efficient, simple and green protocol for the synthesis of substituted pyrroles through Paal-Knorr reaction utilizing the deep eutectic solvent ([CholineCl][ZnCl2]3) under solvent-free sonication [31]. We continued our efforts towards zinc-catalyzed C-N bond formation and due to the ecofriendly and less-toxic nature of zinc, we herein disclose our findings on the zinc-catalyzed protocol for the synthesis of N-aryl pyrroles through Clauson-Kaas reaction under neat condition.At the beginning of our studies, we have chosen aniline (1a, 1\u00a0mmol) and 2,5-dimethoxytetrahydrofuran (2, 1.2\u00a0mmol) as the model substrates for the reaction, which along with 10\u00a0mol% of Zn(OTf)2 catalyst, and heating at 60\u00a0\u00b0C for 2\u00a0h under neat condition resulted in the formation of N-phenyl pyrrole (3a) in 56% yield (Scheme 2\n).Encouraged by this result, we screened various zinc catalysts such as Zn(OTf)2, Et2Zn, ZnI2, Zn powder, anhydrous ZnCl2, Zn(NO3)2\u00b76H2O, ZnSO4\u00b7H2O, Zn(OAc)2\u00b72H2O and ZnO (Table 1\n, entry 1\u20139). Catalytic screening was started with Zn(OTf)2, which furnished 56% yield of the product within 2\u00a0h. Subsequently, we used Et2Zn, Zn powder and Zn(OAc)2\u00b72H2O, which did not afford the product. When we used ZnI2, anhydrous ZnCl2 or Zn(NO3)2\u00b76H2O, the N-substituted pyrrole product was obtained in 18%, 15% and 31% respectively while ZnSO4\u00b7H2O and ZnO when used as the catalyst, only trace amount of the product could be obtained. From these catalyst screening, it was revealed that Zn(OTf)2 was the best catalyst.After the catalyst screening, we tested the effect of temperature and time. The reaction carried out at room temperature provided only trace amount of the product (Table 2\n, entry 1). On increasing the reaction temperature to 60\u00a0\u00b0C, we obtained the desired product in 56% yield (Table 2, entry 2). When we increased the temperature to 80\u00a0\u00b0C, the reaction did not take place due to the solidification of the reaction mixture within 25\u00a0min (Table 2, entry 3). We presume that the solidification is due to polymerization or other side reaction. The solidified reaction mixture was found to be insoluble in most of the solvents. Then we carried out the reaction at 60\u00a0\u00b0C and increased the reaction time to 4\u00a0h, 6\u00a0h and 8\u00a0h, we observed a gradual increase in the yield of product (Table 2, entry 4\u20136). Then we increased the reaction temperature to 70\u00a0\u00b0C for 8\u00a0h, which provided 94% yield of the product (Table 2, entry 7). No change in the product yield was observed when the reaction was carried out at 70\u00a0\u00b0C by increasing the reaction time to 10\u00a0h (Table 2, entry 8). Later, we increased the reaction time to 12\u00a0h and 24\u00a0h, and noticed a decrease in the yield of the product (Table 2, entry 9\u201310). From these observations, we concluded that the optimum temperature and time for the reaction were 70\u00a0\u00b0C and 8\u00a0h respectively.Finally, we investigated the effect of the amount of catalyst loading in this reaction. We tried the reaction using 10\u00a0mol% catalyst, which afforded the product in 90% yield (Table 3\n, entry 1). Then, we increased the amount of catalyst to 15\u00a0mol% and found a decrease in the product yield to 80% (Table 3, entry 2). Then the reaction was carried out without any catalyst, which did not offer any product (Table 3, entry 3). When we performed the reaction using 5\u00a0mol% catalyst, the product was formed in 94% yield (Table 3, entry 4). The reaction when carried out with 2.5\u00a0mol% catalyst furnished the product in 59% of yield (Table 3, entry 5). Finally, we concluded that the suitable reaction condition for this reaction as 5\u00a0mol% Zn(OTf)2 as the catalyst at 70\u00a0\u00b0C for 8\u00a0h when aniline and 2,5-dimethoxytetrahydrofuran were used as the substrates.With the optimized reaction condition in hand, we conducted substrate scope studies using a wide variety of anilines substituted with either electron-donating or electron-withdrawing groups (Scheme 3\n). From these studies, it is revealed that anilines with electron-donating groups such as \u2013Me, \u2013OMe, etc. provided higher yield than those with electron-withdrawing groups such as \u2013COMe, \u2013NO2, etc. and the anilines bearing substituents at -ortho position with electron-donating groups \u2013Me, \u2013OMe, etc. furnished good to excellent yield than those bearing at -meta and -para positions. 2,6-Dimethyl substituted aniline (3e) provided a relatively higher yield than -ortho, -meta and -para toluidine (3b-3d). Ortho bromoaniline (3i) afforded the product in excellent yield when compared to the para bromoaniline (3j) whereas chloro substitutions at -ortho, -meta and -para furnished the desired product in satisfactory yields (3k-3m). But, iodo substitutions at -ortho, -para position and \u2013OH at -para position did not give the expected excellent yield due to the clogging of the reaction mixture at 70\u00a0\u00b0C (3n-3p). The presence of electron-withdrawing groups such as \u2013NO2, \u2013COMe, etc are resulted the desired products in lower yields (3q-3r). We also tried the reaction with aliphatic amines such as benzylamine and cyclohexylamine, which unfortunately did not undergo this reaction.In summary, we have disclosed a green zinc-catalyzed protocol for the synthesis of N-substituted pyrroles through a modified Clauson-Kaas reaction using 2,5-dimethoxytetrahydrofuran and various anilines without using any solvent, base or ligand. A wide range of anilines reacted with 2,5-dimethoxytetrahydrofuran under the optimized reaction conditions to provide the N-substituted pyrroles in moderate to excellent yields.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.CMA and KRR thank the Council of Scientific and Industrial Research (CSIR, New Delhi) and the University Grants Commission (UGC, New Delhi) for the award of junior research fellowships respectively.Supplementary data to this article can be found online at https://doi.org/10.1016/j.rechem.2022.100350.The following are the Supplementary data to this article:\n\nSupplementary data 1\n\n\n\n", "descript": "\n                  The first zinc-catalyzed simple and convenient protocol for the synthesis of N-substituted pyrroles through a modified Clauson-Kaas reaction without co-catalyst, ligand, base and solvent has been described. N-substituted pyrroles were prepared from various aniline derivatives and 2,5-dimethoxytetrahydrofuran under green condition by utilizing low cost, eco-friendly, non-toxic and easily accessible Zn(OTf)2 as catalyst. A wide variety of N-substituted pyrroles were afforded in moderate to excellent yields from easily available starting materials at 70\u00a0\u00b0C.\n               "}
{"full_text": "With the rapid development of society, the excessive consumption of fossil fuels and the dramatically increasing of energy demand have caused serious environmental pollution and energy crisis. The development of renewable energy and energy conversion technologies is considered to be an effective way to solve these two issues.\n1\u20136\n Especially, electrocatalytic conversion devices composed of electrocatalysts, membrane and electrolytes have attracted extensive attention from researchers because of their high efficiency and cleanliness.\n7\n However, their large-scale application is greatly limited by the slow kinetics of the cathodic reaction. Therefore, the development of highly active and stable electrocatalysts is necessary to improve energy conversion efficiency of emerging devices and propel their commercialization.Traditional heterogeneous catalysts (e.g. Pt-based catalysts and RuO2) usually contain a mixture of metal particles with a wide size distribution. However, only a small fraction of metal particles with appropriate size distribution can be used as catalytically active species, while others are either inert or probable to trigger undesired side reactions, which results in low metal utilization efficiency and poor selectivity of the reaction.\n8\u201310\n Homogeneous catalysts have well-structured active sites and tunable ligand environments, which exhibit excellent activity and high selectivity for the target reaction. However, the catalysts still suffer from limitations such as poor stability and unsatisfactory recoverability.\n11\n Single atom catalysts (SACs) combine the advantages of \u201cisolated sites\u201d from homogeneous catalysts and the structural stability as well as convenience of separation from heterogeneous catalysts, which serve as a bridge between homogeneous and heterogeneous catalysts.\n12\u201314\n\nSince Zhang and coworkers reported single Pt atom on iron oxide for CO oxidation in 2011,\n15\n single atom catalysts have attracted a lot of attention in various aspects, such as photocatalysis,\n16\u201320\n electrocatalysis\n21\u201325\n and thermal catalysis\n26\u201328\n because of their maximal atom utilization (nearly 100%) and high catalytic activity, stability and selectivity. The size reduction of nanoparticles to the sub-nanometer level leads to an increasing in number of low-coordinated metal atoms that can be used as catalytically active sites, which will be maximized at the single-atom level when individual metal atoms are accessible and catalytically active.\n29\n However, these low coordination single atoms have very high surface energy and tend to aggregate during the synthesis process.\n27\n Therefore, it is of great importance to select suitable supports to interact with metal single atoms to prevent their aggregation for the synthesis of single-atom catalysts.\n30\n Common support materials include metal oxides,\n15\n metal sulfides,\n31\n metal nitrides,\n32\n metal surfaces,\n33\n MOFs,\n34\n porous carbon materials,\n35\n\n,\n\n36\n g-C3N4\n\n37\n\n,\n\n38\n and graphene.\n39\n\n,\n\n40\n Among them, nitrogen-doped porous carbon is considered as an ideal single-atom support due to its hierarchical pore structure, high mechanical strength and special defect effect.\n14\n\nMetal-nitrogen-carbon catalysts, in which dispersive metal atoms are coordinated to nitrogen atoms doped in carbon nanomaterials, have been used as effective catalysts for lots of electrochemical reactions, such as carbon dioxide reduction reaction (CO2RR),\n41\u201344\n nitrogen reduction reaction (NRR),\n45\n\n,\n\n46\n oxygen reduction reaction (ORR),\n47\u201349\n oxygen evolution reaction (OER),\n50\n hydrogen oxidation reaction (HOR)\n51\n and hydrogen evolution reaction (HER).\n52\n Precious metal catalysts show excellent activity for these reactions, but their high cost and low natural abundance have forced the search for suitable metal substitutes. In contrast, iron is well-known to be abundant in nature and inexpensive, and also exhibits high durability, tunability in both acidic and alkaline medium, and methanol tolerance, which has led to many iron-based catalysts being explored for numerous reactions, especially in the reduction of C, N and O.In this review, as shown in Fig.\u00a01\n, we firstly conclude several essential methods for the synthesis of Fe\u2013N\u2013C. Then, we present the property of Fe\u2013N\u2013C catalysts, including spin-related interaction of electronic structures as well as the orbital coupling between the electronic structure of Fe\u2013N\nx\n and the electronic orbital of C, N and O. Furthermore, we summarize the electrocatalytic applications about the chemical conversion of C, N and O with Fe\u2013N\u2013C. Lastly, we discuss some challenges in structural characterization, mechanistic investigation and industrial application of Fe\u2013N\u2013C, and propose some possible solutions.Since 2008, MOFs have been widely recognized as ideal sacrificial templates for preparing of highly stable and conductive nanostructured carbon.\n53\n By charring the MOF precursors with the target metals, it was found that the single atoms (SAs) were firmly embedded into the carbon carriers through strong metal heteroatom (S/N/O) coordination bonds, successfully providing single-atom catalysts with excellent catalytic performance.\n54\n The iron source is first anchored into the networks of the metal nodes and organic linkages, as well as in the cavities of the entire MOF crystal to form iron-containing MOF precursors, which are then converted into Fe SACs in one-step pyrolysis treatment. So far, the most commonly used MOF precursor in this method is a zinc imidazole skeleton doped with a small amount of iron because of its good structural tailorability, high nitrogen content and the pore confinement effect. Under solvent or solvent-free conditions, Fe2+/Fe3+ is spatially separated by Zn2+ and 2-methylimidazole ligands, which are atomically diluted throughout the MOF crystals. During pyrolysis at high temperatures, Zn metal readily evaporates out of the system due to its low boiling point, leaving the Fe SAs distributed on the MOF-derived porous carbon framework (Fig.\u00a02\na).\n55\n It is worth mentioning that the nitrogen atoms from the organic linkers will firmly anchor and stabilize these Fe single atoms, thus effectively avoiding their migration and aggregation.To improve the activity of Fe\u2013N\u2013C, Zhang et\u00a0al. replaced ZIF-8 with MIL-101 to act as a precursor to introduce abundant mesopores (Fig.\u00a02b).\n56\n However, the iron-containing MIL-101 has high iron content and insufficient nitrogen content, so amino groups are further introduced in preparing of the precursor to reduce the formation of Fe nanoparticles in the catalyst. Firstly, NH2-MIL-101(Al) was pyrolyzed to obtain nitrogen-doped carbon materials (NC-MIL101-T) at a temperature range of 800\u20131100\u00a0\u00b0C under N2. After acid etching, Fe(phen)3\n2+ and NC-MIL101-T support were thoroughly mixed. FeSAC-MIL101-T catalysts were finally obtained after freeze-drying and the second pyrolysis at 800\u00a0\u00b0C. Later, Yang et\u00a0al. proposed to replace Zn by Cd as a sacrificial metal for the synthesis of Fe\u2013N\u2013C, which could reduce the pyrolysis temperature of the precursor from 1000\u00a0\u00b0C to 750\u00a0\u00b0C and thus help to preserve the individual iron atom active sites. Meanwhile, as a comparison, the ZIF-8/Fe catalyst was prepared by pyrolysis at the same temperature. In contrast to the ZIF-8/Fe catalyst, the Fe\u2013N\u2013C retained a smaller amount of sacrificial metal and formed a higher density of single-atom structures.\n59\n\nIn addition, Jiang and co-workers used a hybrid ligand strategy to prepare high-content (1.76\u00a0wt%) single-atom iron-implanted nitrogen-doped porous carbon (FeSA\u2013N\u2013C) by pyrolysis of porphyrinic MOFs (Fe\u2013PCN-222). The mixed porphyrin ligands, Fe-TCPP and H2-TCPP made the Fe(\u2162) ions form a long spatial distance in the MOF skeleton, which was conducive to the formation of single iron atoms by pyrolysis.\n60\n Later, Jiang et\u00a0al. developed a similar strategy to obtain FeSA\u2013N\u2013C with high Fe loading (3.46%) by the pyrolysis of SiO2@MOF composite (Fig.\u00a02c).\n57\n The pre-synthesized PCN-222(Fe) has a one-dimensional pore structure of \u223c3.2\u00a0nm, which ensures that tetraethylorthosilicate (TEOS) can be fully permeable in the internal space. Under hydrochloric acid vapor treatment, TEOS was hydrolyzed and condensed to silica, forming the SiO2@MOF composite with well-retained MOF crystallinity. Silica interacts with iron atoms to increase the energy barrier for iron atoms migration and thus prevent their aggregation. Moreover, the removal of silica increases the porosity and specific surface area of the material, which facilitates the exposure of active sites and mass transfer. In contrast, Fe particles could be detected from the catalyst obtained by the direct pyrolysis of PCN-222(Fe) without SiO2, which proved that the presence of SiO2 inhibited the migration of Fe atoms under pyrolysis.Moreover, heteroatom doping is an important means to increase the density of active centers and improve the electrocatalytic activity of catalysts. The doping of heteroatoms can change the coordination environment and electronic properties of iron centers as well as change the density of active centers through long-range or short-range interactions. Wang et\u00a0al. designed a Zn/Fe bimetallic mixed-ligand metal-triazolate (MET) as the precursor and 4,5-dichloroimidazole as the source of Cl to obtain FeN4Cl/NC (Fig.\u00a02d). The higher electronegativity of chlorine could change the d-band delocalization and electronic structure of Fe atoms to ensure the presence of FeN4Cl coordination configuration.\n58\n\nIrregular metal-containing complexes and polymers can replace MOFs as precursors to prepare atomically dispersed Fe\u2013N\u2013C catalysts with special structures via pyrolysis. Tang et\u00a0al. designed atomically dispersed iron atoms anchored on N-doped carbon nanosheets (Fe\u2013N\u2013C HNSs) with well-defined FeN4 structures and unique spherical hollow architecture via SiO2-templated strategy. Histidine (His) served as the source of N and C due to high content of heteroatoms and natural abundance. During the process of preparation, the pre-synthesized SiO2 nanospheres could absorb Fe3+ ions through electrostatic attractions after surface modification with negative charges. Subsequently, the His molecules bound with Fe3+ ions to form SiO2@Fe-His nanospheres, which would be pyrolyzed at high temperature and acid leached. The 3D hollow spherical structure prevented aggregation of iron atoms, offered shorten pathway for mass diffusion and exposed more active sites.\n61\n Li et\u00a0al. synthesized Fe\u2013N\u2013C catalysts that possess special atomically dispersed Fe-N\nx\n structure via changing the ratio of acrylic acid (AA) and maleic acid (MA). AA could be polymerized into PAA and chelated with Fe3+ to form a cross-linked hydrogel, while MA could be co-polymerized with AA to increase the carboxylic content of the copolymer (P(AA-MA)). These polymers and cyanamide were treated at high-temperature to obtain PAA\u2013Fe\u2013N and P(AA-MA)\u2013Fe\u2013N. Structural characterization results showed that the introduce of MA elongate the bond strength of Fe\u2013N and create the exclusive Fe\u2013N4/C moiety of P(AA-MA)-Fe-N.\n62\n\nGenerally, high temperature pyrolysis is the most mature method for preparing FeSA\u2013N\u2013C. By adjusting the size, coordination number and composition of the MOF precursor, a more precise tuning of the catalyst can be achieved. In addition, pyrolysis of irregular metal-containing complexes and polymers enable more structurally diverse FeSA\u2013N\u2013C catalysts with more abundant iron sources, carbon sources and nitrogen sources. Remarkably, during the process of high temperature pyrolysis, the introduction of template can further regulate the microstructure and pore size of the FeSA\u2013N\u2013C catalysts, thus improving their catalytic performance and stability.Due to low metal loading in high temperature pyrolysis methods, the chemical vapor deposition (CVD) method was designed to synthesize Fe\u2013N\u2013C catalysts (Fig.\u00a03\na).\n63\u201367\n One of the most important advantages of the CVD method is the ability to select different templates and adjust deposition conditions to precisely control the morphology and pore structure of the resulting catalysts.\n68\n\nWu et\u00a0al. prepared atomically dispersed Fe\u2013N\u2013C catalyst with increased Fe loading via CVD method compared to wet-chemistry synthesis. As shown in Fig.\u00a03b, with the presence of argon gas and heating, gaseous 2-MeIm was deposited onto Fe\u2013ZnO nanosheets and underwent a gas-solid reaction to form Fe\u2013Zn(MeIm)2 intermediate and its crystalline structure gradually evolved from zif towards kat phase. The formation of CVD/Fe-kat increased the number of active sites, which was attributed to the fact that the formed narrower pores could promote the coordination of single Fe sites with N and thus slow down the diffusion and agglomeration of Fe atoms. Finally, CVD/Fe\u2013N\u2013C-kat was obtained under 1000\u00a0\u00b0C heating. According to the analysis of in-situ electrochemical measurements through the nitrite absorption followed by reductive stripping, the FeN4 active site density of CVD/Fe\u2013N\u2013C-kat is about 26\u00a0mmol\u00a0g\u22121, which is higher than most Fe\u2013N\u2013C catalysts synthesized by other methods.\n63\n Similarly, Peng et\u00a0al. chose ferrocene-doped calcium oxide as a template and pyridine as carbon and nitrogen sources to synthesize atomic Fe\u2013N\u2013C catalyst by pyrolysis at 700\u00a0\u00b0C under an argon atmosphere via CVD method (Fig.\u00a03c).\n66\n Later, they used iron(III) acetylacetonate dissolved in pyridine as a precursor, magnesium hydroxide as a substrate for CVD pyrolysis, and finally acid etching to obtain single iron atoms anchored on porous N-doped carbon (Fe-N-PC) (Fig.\u00a03e).\n64\n\nJia and co-workers creatively synthesized Fe\u2013N\u2013C with high active sites by flowing ferric chloride vapor over Zn\u2013N\u2013C at 750\u00a0\u00b0C via CVD method (Fig.\u00a03d). Zn\u2013N\u2013C material was obtained by pyrolysis of ZIF-8 under argon atmosphere at 1050\u00a0\u00b0C, which contained 2.16% Zn and abundant microporous structures. In the presence of FeCl3, the Zn was removed and Fe took the place of Zn under high temperature (>650\u00a0\u00b0C), which contributed to the formation of Fe\u2013N4 sites according to the reaction mechanism: FeCl3(g)\u00a0\u200b+\u00a0\u200bZn\u2013N4\u00a0\u200b+\u00a0\u200bX \u2192 Fe\u2013N4\u00a0\u200b+\u00a0\u200bZnCl2(g)\u00a0\u200b+\u00a0\u200bXCl (X refers to H or Cl). The obtained FeNC-CVD-750 catalyst had an activity site density of 1.92\u00a0\u200b\u00d7\u00a0\u200b1020 sites per gram with 100% site utilization.\n67\n In short, although the CVD method can appropriately increase the loading of iron atoms, its complex synthesis steps and high temperature condition force a preference for simpler methods to synthesize single-atom catalysts.Traditional ball milling method involves mixing metal salts, nitrogen-containing compounds and supports by ball milling and then thermal reducing the mixture to form single atom catalysts. The purpose of ball milling is to improve metal dispersion before pyrolysis. Dai et\u00a0al. reported a universal domino reaction strategy to produce M-SA/NC catalysts including Fe, Co, Ni, Mn, Mo, Pd and arbitrary combinations SA/NC catalysts (Fig.\u00a04\na). Polyaniline (PANI), appropriate metal salt, NaCl and NaNO3 were mixed and ball-milled with the aim of making PANI chains doped with metal ions and wrapped around salt particles. NaNO3 was decomposed and released gases to cause PANI blew up and carbonized with the formation of porous carbon nanosheets by pyrolysis at 1000\u00a0\u00b0C. The role of the gas is to etch the nanosheets for obtaining a microporous structure and also to anchor the metal atoms to the carbon framework and prevent them from aggregating into metal particles. In addition, the structure and pore size of M-SA/NC could be regulated by changing the content of NaNO3.\n69\n\nThe other means of synthesis is to directly high-energy ball milling coordination precursors (such as iron phthalocyanine) and carbon supports to achieve mixing in molecular-level and provide precursors for the subsequent pyrolysis.\n70\n\n,\n\n71\n Deng et\u00a0al. used the ball milling approach to synthesize a highly dispersed single FeN4 center with coordinatively unsaturated iron sites confined in a graphene matrix at a large quantity with iron phthalocyanine (FePc) and graphene nanosheets as precursors (Fig.\u00a04b).\n70\n During the ball milling process, on one side, the outside macrocyclic structure of FePc is disrupted to produce a fragmented structure possessing FeN4. On the other side, graphene produces defective sites and interacts with isolated FeN4 centers.Under the influence of the above two ball milling methods, Baek et\u00a0al. discovered an eco-friendly top-down strategy, namely mechanochemical abrasion method, which means the direct atomization of bulk metals on different supports by means of abrasion, without any solvent and the generation of by-products and waste in the process (Fig.\u00a04c).\n32\n In the experiment, iron balls, N2 gas and graphite were loaded into a steel container and the ball-milling was conducted at a constant rotation speed for 30\u00a0h. Iron balls are not only the source of iron single atoms, but also transfer kinetic energy to drive reactions. N2 gas is dissociated on the surface of the iron balls and enters the graphite framework to anchor the iron atoms. The graphite is served as an active matrix to accommodate the nitrogen and atomized metal. An additional advantage of the abrasion method is that the amount of catalyst can be expanded in equal proportions by increasing the volume of the ball-milling container.In general, Table\u00a01\n lists the advantages and disadvantages of these three preparation methods for Fe\u2013N\u2013C single atom catalysts. Compared to other synthesis methods, ball milling method simplifies the synthesis process and is prone to be applicable for the large-scale preparation of Fe\u2013N\u2013C. However, during the ball milling process, some of iron atoms may be encapsulated inside the catalysts, which in turn results in a low atomic utilization and uneven structures.There are many factors affecting the properties of Fe\u2013N\u2013C electrocatalysts, such as the electronic structure of the central metal, the metal coordination environment, and the metal-support interactions. But little attention has been paid to the role of electron spin of Fe. As the following, we will discuss spin-related properties in Fe\u2013N\u2013C catalysts.In general, FeN\nx\n moieties serve as active sites of most Fe\u2013N\u2013C catalysts, while the number of coordination x usually depends on the synthesis conditions. The coordination environment and the valence state of iron determine the spin state of single Fe atom. Fig.\u00a05\n shows the electronic configuration and spin state of the common FeN4 species. When oxidation state of Fe is +1 or +4, the 3d electron configuration of FeN4 is \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n2\n\nd\n\nx\nz\n\n\n2\n\nd\n\nz\n2\n\n\n1\n\n or \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n1\n\nd\n\nx\nz\n\n\n1\n\n, respectively, belonging to low spin states. When oxidation state of Fe is +2, the 3d electron configuration of Fe(\u2161)N4 can simply be classified into three forms, including \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n2\n\nd\n\nx\nz\n\n\n2\n\n, \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n2\n\nd\n\nx\nz\n\n\n1\n\nd\n\nz\n2\n\n\n1\n\n and \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n1\n\nd\n\nx\nz\n\n\n1\n\nd\n\nz\n2\n\n\n1\n\nd\n\n\nx\n2\n\n\u2212\n\ny\n2\n\n\n\n1\n\n, corresponding to low spin t2g6eg0, medium spin t2g5eg1 and high spin t2g4eg2, respectively. When oxidation state of Fe is +3, the 3d electron configuration of Fe(\u2162)N4 can simply be classified into three forms, including \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n2\n\nd\n\nx\nz\n\n\n1\n\n, \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n1\n\nd\n\nx\nz\n\n\n1\n\nd\n\nz\n2\n\n\n1\n\n and \n\n\nd\n\nx\ny\n\n\n1\n\nd\n\ny\nz\n\n\n1\n\nd\n\nx\nz\n\n\n1\n\nd\n\nz\n2\n\n\n1\n\nd\n\n\nx\n2\n\n\u2212\n\ny\n2\n\n\n\n1\n\n, corresponding to low spin t2g5eg0, medium spin t2g4eg1 and high spin t2g3eg2, respectively. Most of the reports demonstrate that the preparation method and preparation conditions have important effects on the spin of iron in Fe\u2013N\u2013C catalysts, including the type and content of iron, pyrolysis temperature, pyrolysis duration, atmosphere, etc. Additionally, it must be mentioned that heteroatom doping and the modulation of supports could also change the iron electron polarization as well as the electron spin state and complete mutual transformation of different spin states.Commonly used detection tools for measuring the spin state of Fe include electron paramagnetic resonance (EPR), L-edge X-ray absorption near edge structure (XANES) and M\u00f6ssbauer spectroscopy. When there are unpaired electrons in the outer electron orbitals of iron, under an applied magnetic field and an applied electromagnetic wave, the electrons at the low spin energy level absorb energy and transfer to the high spin energy level to produce paramagnetic resonance absorption, which is the reason why EPR can measure the spin state of iron. However, there is a serious limitation of EPR that it can only provide information on iron with half-integer spin multiplicity and cannot measure the valence state of all iron species. Fe L-edge XANES is employed to analysis the structure of Fe, because the valence and spin states of 3d tradition metals significantly affect the L-edge spectra. The L-edge of 3d tradition metals is generated by electronic transitions between the 2p level and the mostly unoccupied 3d electronic states. The L3 edge (706\u2013712\u00a0eV) involves transitions from 2p3/2 to 3d states, while the L2 edge (718\u2013726\u00a0eV) comes from transition 2p1/2 to 3d states. Additionally, it has been found that there is a relationship between the area ratio of L3/L2 and the spin state. A higher ratio indicates that high spins dominate, and vice versa.\n72\u201374\n However, the sensitivity of XANES to the spin state of the fine structure containing iron is not as good as that of the M\u00f6ssbauer spectrum.The 57Fe M\u00f6ssbauer spectroscopy, involving the resonant and recoil-free emission and absorption of \u03b3-rays by atomic nuclei, is used to study the valence, spin polarization and coordination environment of iron in single-atom catalysts based on the values of isomer shift (IS) and quadrupole splitting (QS).\n74\u201376\n Each type of iron with a defined coordination structure and spin state corresponds to a specific value of IS and QS, and the shift of values implies the change of iron species. In addition, with the development of operando spectroscopy, operando M\u00f6ssbauer spectroscopy could be used to characterize the spin state of iron-based materials during reactions. As a result, the M\u00f6ssbauer spectrum has become the most reliable means to characterize the spin state of iron-based materials.Recently, non-resonant X-ray emission spectroscopy (XES) was demonstrated to quantify the average, ex situ spin state of a series of Fe\u2013N\u2013C catalysts. Herranz et\u00a0al. used two-component fitting to analyze the K\u03b2 main lines based on a linear relation between the relative area of the K\u03b2\u2019 spectral peak and the spin state of several reference compounds, and in turn established a potential-induced spin change in the catalysts prepared by pyrolysis of an Fe-porphyrin.\n77\n Excitingly, this method has the potential to be extended to measure the spin state of other transition metal materials. However, it always reflects an average spin state information and cannot help us to determine the influence of the spin state on the reaction mechanism at a deeper level.Numerous experimental evidences show that the spin of Fe affects the occurrence of the reaction as well as the rate and selectivity of the reaction. Next, we will further analyze the effect of spin on the reduction of C, N and O from the perspective of orbital coupling.The activation of CO2 plays an important role in the process of CO2\u2192CO. The HOMO of CO2 localize on the O atom, while the LUMO localize on the C atom and performed as a C\u2013O \u03c3\u2217 orbital. Taking Fe(\u2160) L.S. as an example, the electron-rich iron center conducts nucleophilic attacks on the electrophilic C-center of CO2, which means that occupied \n\n\nd\n\nz\n2\n\n\n\n, \n\n\nd\n\ny\nz\n\n\n\n and \n\n\nd\n\nx\nz\n\n\n\n orbitals of iron could offer electrons to populate these empty \u03c3\u2217 and \u03c0\u2217 orbitals, which is beneficial to the activation of CO2 (Fig.\u00a06\na, d). When CO2 gets an electron, a new splitting of the CO2 orbitals occurs, but the coupling tendency with Fe will not change greatly.\n78\n\nTaking Fe(\u2162) M.S. as an example, the iron centers have empty \n\n\nd\n\n\nx\n2\n\n\u2212\n\ny\n2\n\n\n\n\n orbitals and the half-filled \n\n\nd\n\nz\n2\n\n\n\n orbitals (Fig.\u00a06b). Upon the N2 side-on adsorption, N2-\u03c3 electrons will be transferred to higher-energy empty spin-down d-orbitals (\n\n\nd\n\n\nx\n2\n\n\u2212\n\ny\n2\n\n\n\n\n). The empty spin-down orbitals \n\n(\n\nd\n\nz\n2\n\n\n)\n\n at the higher energy interact with the antibonding orbitals of N2 to weaken the N\u2013N triple bond and lower the N2 adsorption energy (Fig.\u00a06e), which can explain why Fe(\u2162) has excellent NRR performance.\n79\n The low spin Fe(\u2161)N4 lack of unpaired d electrons and the high spin Fe(\u2161)N4 lack of empty d orbitals showed poorly activity of N2 reduction reaction. Similarly, the high spin Fe(\u2162)N4 exhibits worse activity than middle spin Fe(\u2162)N4 and low spin Fe(\u2162)N4 due to the shortage of empty d orbitals.\n80\n\n\nFig.\u00a06c illustrates the major orbital interactions between O2 and high spin Fe(\u2161) during the O2 adsorption process. The antibonding orbital (\u03c0\u2217) of O2 could couple with the half-filled \n\n\nd\n\nz\n2\n\n\n\n, \n\n\nd\n\ny\nz\n\n\n\n and \n\n\nd\n\nx\nz\n\n\n\n orbitals to form four new low-to-high orbitals, corresponding to \n\n\nd\n\nx\nz\n\n\n\u2212\n\nd\n\ny\nz\n\n\n\u2212\n\n\u03c0\n\u2217\n\nBD\n\n, \n\n\nd\n\nz\n2\n\n\n\u2212\n\n\u03c0\n\u2217\n\nBD\n\n, \n\n\nd\n\nx\nz\n\n\n\u2212\n\nd\n\ny\nz\n\n\n\u2212\n\n\u03c0\n\u2217\n\n\nBD\n\u2217\n\n\n and \n\n\nd\n\nz\n2\n\n\n\u2212\n\n\u03c0\n\u2217\n\n\nBD\n\u2217\n\n\n, which increases the d orbital splitting and form a more stable system (LS Fe(\u2161)-O2). From the electronic point of view, it can be roughly considered that the d electrons of \n\n\nd\n\nz\n2\n\n\n\n, \n\n\nd\n\ny\nz\n\n\n\n and \n\n\nd\n\nx\nz\n\n\n\n orbitals leap to the \u03c0\u2217 orbital of O2, resulting in electrons rearrangement that occur to reduce the energy of the system (Fig.\u00a06f). It is also worth mentioning that the interaction between iron species and O2 should not be too strong. Low spin Fe(\u2162) without eg filling has empty \u03c3\u2217 antibonding orbital of FeN4, and leads to a very strong Fe(\u2162)/O2 interaction and a quite stable Fe4+\u2013O2\n2\u2212 bond. Thus, it is difficult for the occurrence of Fe(\u2163)-O2/Fe(\u2162)-OOH transition.\n81\n\n,\n\n82\n\nElectrocatalytic carbon dioxide reduction reaction is a significant strategy to solve the problem of energy shortage and environmental pollution. Carbon dioxide could be reduced to value-added chemical products powered by electricity generated from renewable energy sources.\n83\n Many studies have shown that Fe\u2013N\u2013C exhibits remarkable performance concerning the reduction of CO2 to CO and other simple chemicals (Table\u00a02\n).Hu et\u00a0al. reported Fe\u2013N\u2013C catalyst with atomically dispersed iron sites, which produced CO at the overpotential of \u22120.08\u00a0V in the CO2-saturated 0.5\u00a0M KHCO3 catholyte. When the cathode potential decreased to \u22120.45\u00a0V (vs. RHE), particle current density of CO could reach 94\u00a0mA\u00a0cm\u22122 with FECO higher than 90%. Fe 2p3/2 XPS spectrum and Fe K-edge XANES spectrum results indicated that the iron oxidation state of Fe\u2013N\u2013C was +3. Compared to Fe\u2013N\u2013C, the current density of Zn\u2013N\u2013C synthesized under the same conditions could be neglected at \u22120.1\u00a0\u200bV to \u22120.6\u00a0\u200bV (vs. RHE), indicating that Fe sites was the origin of Fe3+\u2013N\u2013C during CO2RR. Operando XANES displayed that the Fe K-edge of Fe3+\u2013N\u2013C showed no obvious shift at \u22120.1\u00a0\u200bV to \u22120.4\u00a0\u200bV (vs. RHE), indicating the Fe species remained in +3 oxidation state; while Fe K-edge shifted to lower energies at \u22120.4\u00a0\u200bV to \u22120.5\u00a0\u200bV (vs. RHE), the same as the potential of the deactivation of Fe3+\u2013N\u2013C, indicating the reduction of Fe3+ to Fe2+ (Fig.\u00a07\na). Furthermore, it could be observed that Fe3+ was reduced to Fe2+ in the as-prepared Fe3+\u2013N\u2013C at \u22120.1\u00a0V to \u22120.2\u00a0V (vs. RHE) (Fig.\u00a07b). These phenomena proved that Fe3+ sites were more active for generating CO. According to the kinetic and mechanistic analysis (Fig.\u00a07e), CO2 adsorption is the rate-limiting-step for Fe2+\u2013N\u2013C, while the protonation of the adsorbed CO2\n\u2212 to form an adsorbed COOH intermediate is the rate-limiting-step for Fe3+\u2013N\u2013C. Additionally, the CO2RR rate would be also limited by CO desorption for Fe2+\u2013N\u2013C but not limited for Fe3+\u2013N\u2013C. As a result, the superior activity of Fe3+ could be proven to derive from faster CO2 adsorption and weaker CO adsorption compared to Fe2+ sites.\n84\n Liu et\u00a0al. identified low-spin Fe(\u2160)N4 is the reactive center for the conversion of CO2 to CO. Operando 57Fe M\u00f6ssbauer results showed that three doublets were detected in the Fe\u2013NC\u2013S at OCV, corresponding to LS Fe(\u2161)N4, MS Fe(\u2161)N4 and HS Fe(\u2161)N4, respectively. When polarized at \u22120.3\u00a0V (vs. RHE), a new doublet was observed and was assigned to LS Fe(\u2160)N4. When the potential was gradually decreased to \u22120.9\u00a0V (vs. RHE), the relative content of LS Fe(\u2160)N4 increased accompanied by the decreasing of relative content of LS Fe(\u2161)N4, which reflected that LS Fe(\u2161)N4 was reduced to LS Fe(\u2160)N4. Additionally, the new doublet disappeared when removing the potential, further proving that Fe(\u2160)N4 transited from LS Fe(\u2161)N4 was the real active center during CO2RR (Fig.\u00a07c). DFT calculations indicated that the CO2 molecule is activated on the Fe(\u2160) site and then forms the \u2217COOH intermediate after the hydrogenation step. During the process, the singly occupied \n\n\nd\n\nz\n2\n\n\n\n orbital of Fe(\u2160) coupled with the singly occupied \u03c01\u2217 orbital of COOH to generate one fully occupied bonding (\n\n\nd\n\nz\n2\n\n\n\u2212\n\n\u03c0\n1\n\u2217\n\nBD\n\n) orbital and one empty antibonding (\n\n\nd\n\nz\n2\n\n\n\u2212\n\n\u03c0\n1\n\u2217\n\n\nBD\n\u2217\n\n\n) orbital (Fig.\u00a07d). Next, \u2217CO and H2O are generated in the presence of electrons and protons. Finally, the adsorbed CO desorbs from the Fe(\u2160) site to complete the catalytic cycle.\n75\n\nTo further understand the effect of spin state of single-atom FeN4, Chen et\u00a0al. conducted a more detailed analysis on the electroreduction of CO2 to CO/HCOOH. Combined with the Fe2+ radius, energy order and corresponding HOMO\u2013LUMO gap of calculated Fe(II)N4 in the different spin states, the order of catalyzing activity is inferred to Fe(II)N4(MS)\u00a0>\u00a0Fe(II)N4(LS)\u00a0>\u00a0Fe(II)N4(HS). Moreover, the calculation of CO2 absorption energy indicated the adsorption strength decreases with the increase of the spin states for Fe(II)N4. Furthermore, it is clear that the middle spin Fe(II)N4 has the lowest energy barrier for the first-step reduction of CO2 (0.52\u00a0eV) compared to other two spin states. As a result, the middle spin Fe(II)N4 have the highest selectivity and best activity from the perspective of mechanism. The same approach proves that the middle spin Fe(III)N4C favors the conversion process of \u2217CO2 to \u2217COOH as compared with the other two spin states.\n89\n\nThe industrial production of ammonia is mainly dependent on the Haber\u2013Bosch process under harsh conditions, which accounts for 1%\u20132% of the earth's energy supply.\n97\n\n,\n\n98\n Moreover, this process can only obtain relatively low conversions due to the constraints of chemical equilibrium. Therefore, it is of great significance to develop efficient nitrogen fixation routes under mild conditions. Inspired by the fact that bacteria can electrochemically reduce nitrogen in the presence of enzyme nitrogenase, electrochemical N2 reduction reaction via \n\n\nN\n2\n\n\n(\ng\n)\n\n+\n6\n\nH\n+\n\n+\n6\n\ne\n\u2212\n\n\u2192\n2\n\nNH\n3\n\n\n(\ng\n)\n\n\n received a lot of attention.\n99\n Unfortunately, hydrogen evolution reaction (HER) is more likely to occur at similar potentials at most of the metal active sites due to the yield of a large amount of electrons and protons. From the kinetic point of view, the activation of N2 is the rate determining step of nitrogen reduction reaction.\n100\n Fe\u2013N\u2013C is thought to be ideal electrocatalysts for lowering the free barrier of N2, weakening hydrogen absorption, and improving ammonia selectivity due to the dispersion of active sites and the positive charge of the metal (Table\u00a03\n).\n101\n\nThanks to high stability and high carrier mobility, graphene is expected to facilitate charge transfer in catalytic reactions. As early as 2016, Luo et\u00a0al. proposed a new catalyst, FeN3-embedded graphene, for activating N2 and converting it into NH3 at room temperature from first-principles calculations. From the perspective of chemical coordination, the FeN3 center is strongly spin-polarized with a localized magnetic moment, which greatly facilitates the adsorption of N2 and activates the inert NN bond. The synergistic interaction between graphene and FeN3 gives the system novel properties to catalyze the conversion of activated N2 to NH3 via a six-proton and six-electron process at room temperature following three possible reaction paths.\n102\n On this basis, Zheng et\u00a0al. designed and synthesized Fe\u2013N/N\u2013CNTs with built-in Fe\u2013N3 sites by pyrolysis of Fe-doped ZIF\u2013CNTs templates. The NRR performance indicated that Fe\u2013N/N\u2013CNTs possessed the highest NH3 average yield of 34.83\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121\ncat., and corresponding FE of 9.28% at \u22120.2\u00a0V (vs. RHE).\n90\n\nLater, more and more studies have found Fe\u2013N4 structure possess higher intrinsic activity and obvious ability to inhibit hydrogen evolution. Liu et\u00a0al. reported a Fe single-atom catalyst with well-defined Fe\u2013N4 active sites and in neutral media, achieving high Faradaic efficiency (18.6\u00a0\u00b1\u00a00.8%) and NH3 yield rate (62.9\u00a0\u00b1\u00a02.7\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121\ncat.) at \u22120.4\u00a0V (vs. RHE) at room temperature.\n91\n Hu et\u00a0al. developed a method to prepare iron-nitrogen-carbon materials for electrocatalysis N2 reduction reaction by loading iron phthalocyanine (FePc) on nano/microporous carbon at a molecular level. It delivered a high selectivity and activity with a NH3 yield rate of 137.95\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121\nFePc at the potential of \u22120.3\u00a0V (vs. RHE) in 0.1\u00a0M Na2SO4 aqueous solution. On the basis of systematic electrochemical analyses, poisoning experiments and theoretical calculations, it suggested that Fe center in FeN4 was the most active site for NRR rather than N or C sites.\n92\n Yan et\u00a0al. demonstrated a single atomic iron catalyst on nitrogen-doped carbon (FeSA\u2013N\u2013C), which promoted NRR process with a Faradaic efficiency of 56.55% at an onset potential of 0.193\u00a0V and a desirable ammonia yield rate of 7.48\u00a0\u03bcg\u00a0h\u22121\u00a0mg\u22121 at 0\u00a0V in alkaline solution. Molecular dynamics simulation unveiled that N2 molecules tended to accumulate at approximately 0.45\u00a0nm from the Fe site, leading to high local concentrations, which would promote the following adsorption (Fig.\u00a08\na). DFT calculations proved that the energy barrier of N2 absorption is much lower than that of water dissociation (Fig.\u00a08b) and the alternating pathway was prone to achieve NRR (Fig.\u00a08c). These results illustrated that FeSA\u2013N\u2013C was more favorable for nitrogen adsorption than hydrogen adsorption with a small energy barrier.\n46\n\nIn order to understand the relationship between the electronic state of FeN4 and NRR, Zhang et\u00a0al. fabricated Fe and Mo co-coordinated polyphthalocyanine electrocatalyst (FeMoPPc) by a low-temperature melt polyphthalocyanine. The zero-field cooling temperature-dependent magnetic susceptibility measurements and 57Fe M\u00f6ssbauer spectra revealed that Fe(\u2161) changed from high spin to middle spin after adding Mo, which weakened the NN bond and promoted the first hydrogenation of N2.\n80\n In addition, Feng et\u00a0al. found that F surface modification could induce Fe(\u2162) in the high spin state, which facilitated \u03c0-backdonation process, promoted the activation of N2 and reduced the limiting potential of NRR.\n103\n\nThe misuse of nitrogen fertilizers and the consumption of fossil fuels have made nitrate ions one of the most spread water contaminants, posing a serious threat to the ecology and human health.\n104\n\n,\n\n105\n Nitrate reduction reaction converts NO3\n\u2212 to N2 or NH3 under a wild temperature and pressure, playing a vital role in promoting the earth's nitrogen cycle and solving water pollution problems. In recent years, a series of metal catalysts have been used to convert nitrate to nitrogen, including Ru, Rh, Ir and Cu.\n106\n\n,\n\n107\n However, there have been few studies on the reduction of nitrate wastes to value-added ammonia. As a vital competition, NO3\n\u2212 reduction to N2 involves an N\u2013N coupling step, which needs two neighboring active sites. Therefore, selecting suitable single atom catalyst can improve the selectivity towards ammonia. So far, Fe\u2013N\u2013C was reported to perform excellent activity of NitRR, which could help us understand the complex pathways and mechanism of the 8 protons and 8 electrons transfer process (Table\u00a03).\n93\u201395\n\nWang et\u00a0al. synthesized a single atomic Fe\u2013N\u2013C catalyst by a TM-assisted carbonization method with highly mesoporous structures. At a potential of \u22120.66\u00a0V (vs. RHE), Faradaic efficiency of the ammonia increased to a maximal of \u223c75%. A large NH3 partial current density of \u223c100\u00a0mA\u00a0cm\u22122 and a yield rate of \u223c20,000\u00a0\u03bcg\u00a0h\u22121\u00a0mg cat\u22121 (0.46\u00a0mmol\u00a0h\u22121\u00a0cm\u22122) were obtained at the potential of \u22120.85\u00a0V (vs. RHE). The catalyst still exhibited excellent stability with a high NH3 yield rate and FE after 20 consecutive electrolysis cycles under the best NH3 selectivity reaction condition. DFT calculations reveal the minimum energy pathway for NO3\n\u2212 reduction to NH3 on Fe single atom site (Fig.\u00a09\na). Furthermore, Fig.\u00a09b exhibited NO\u2217 is a key intermediate and a limiting potential of U\u00a0=\u00a0\u22120.30\u00a0V is needed to make all steps downhill in free energy. These results proved the high NH3 yield rate and activity of Fe SAC contributed to intrinsic high-efficiency active Fe\u2013N4 centers that exhibiting lower thermodynamic barriers and optimized electrocatalytic conditions.\n93\n\nLater, Yu et\u00a0al. demonstrated a polymer-hydrogel strategy for preparing single atom Fe catalysts anchored on N-doped porous carbon (Fe-PPy SACs). When the cathode potential varied from \u22120.3\u00a0V to \u22120.7\u00a0V (vs. RHE), The catalyst displayed a maximum ammonia yield rate of 2.75 mgNH3 h\u22121\u00a0cm\u22122 with nearly 100% Faradaic efficiency. Besides, The Fe\u2013PPy SACs delivered a twelve times higher turnover frequency than Fe nanoparticles. The NitRR mechanism illustrated that the single active Fe-N\nx\n site experienced a nitrate-preoccupied transition center and efficiently eliminated the competing water adsorption.\n94\n Later, Liu et\u00a0al. found that doping of S could significantly enhance the activity of NitRR.\n95\n The catalyst (Fe\u2013CNS) has many folds and defects according to the electrochemical active surface areas and pore size distributions, which suggested that S-doping could create more defect sites and shift the previously balanced electrons, leading to well electrical conductivity. Compared with Fe\u2013CN, Fe\u2013CNS showed higher NH4\n+\u2013FE at all potentials. Moreover, the potential of maximum NH4\n+\u2013FE (78%) with Fe\u2013CNS was \u22120.57\u00a0V, higher than that of Fe\u2013CN (\u22120.67\u00a0V). DFT calculations reflected that the energy of the basic reaction (\n\n\nNO\n3\n-\n\n\u2192\n\nNO\n2\n\u2217\n\n\n) changed from 2.02\u00a0\u200beV to 2.37\u00a0\u200beV after the doping of sulfur, which was consistent with the better catalytic performance of nitrate removal of Fe\u2013CNS. And the energy changes for the basic reactions from N\u2217 to N2\u2217 and from N\u2217 to NH\u2217 on Fe\u2013CNS are \u22120.14\u00a0\u200beV and 2.78\u00a0\u200beV, respectively, which indicates that the ammonia path is thermodynamically favored. Unfortunately, no study has yet shown that the spin state of the iron center can affect the activity of nitrate reduction.Proton exchange membrane fuel cells have broad application prospects due to their high efficiency and zero emission.\n108\n However, the slow cathodic reaction kinetics severely limits their development. Current electrocatalysts for this reaction are usually expensive, low storage capacity commercial Pt-based catalysts.\n109\n Therefore, it has become a top priority to search for efficient and stable ORR catalysts that could replace Pt. So far, among many single-atom transition metal catalysts, Fe\u2013N\u2013C with atomically dispersed iron sites has shown the best ORR catalytic performance (Table\u00a04\n).\n110\n\n,\n\n111\n Oxygen reduction reactions are normally divided into a two-electron transfer process and a four-electron transfer process. The two-electron process reduces O2 to hydrogen peroxide (H2O2), and the four-electron process directly reduces O2 to H2O under acidic conditions or hydroxide (OH\u2212) under alkaline conditions.\n112\n\n,\n\n113\n Hydrogen peroxide can react with iron sites in the Fenton reaction to produce reactive oxygen species, which in turn continuously leads to catalyst deactivation and degradation, and damage to the proton exchange membrane.\n114\n Therefore, we hope to further find Fe\u2013N\u2013C catalysts with high selectivity towards H2O and well-defined mechanism.Xu et\u00a0al. proposed a defined explain about the influence of local Fe(\u2161) spin configuration on ORR. The higher spin state of iron in FeN4 with bond contraction can create a wider spin-related channel in FeN4, promoting the charge transport during ORR. Moreover, the oxygen molecule can be more easily captured by FeN4 with Fe\u2013N bond contraction because of higher bond order resulted from the spin\u2013orbital interactions between iron and O2, which should be the intrinsic factor dominating the DFT calculated trend of O2 adsorption.\n117\n Liu et\u00a0al. firstly developed Operando 57Fe M\u00f6ssbauer to identify the exact structures and spin state of active atomically dispersed iron moieties during ORR. When polarized at 0.9\u00a0V (vs. RHE), it indicated that O2 adsorbed on the HS Fe(\u2161)N4 sites along with the generation of O2\u2013Fe(\u2161)N5 intermediate and the spin state of Fe2+ transited from HS to LS with the central Fe2+ moving to the N4-plane. When polarized at 0.7\u00a0V (vs. RHE), O2 adsorbed on the LS Fe(\u2161)N4 sites along with the generation of O2\u2013Fe(\u2161)N4 intermediate and the spin state of Fe2+ transited from LS to HS with Fe2+ moving out of the N4-plane (Fig.\u00a010\na and b). As shown in Fig.\u00a010c, the spin crossover of Fe2+ significantly reduces the energy barrier for the dynamic cycle. Quantum chemical studies provide the structural and dynamic evolutions of Fe(\u2161)N4 and spin-crossover-involved mechanism for ORR (Fig.\u00a010d). Due to the exchange stabilization, the interaction between O2 and HS Fe2+ increases the d-orbital splitting, resulting in the conversion of the spin state of iron (Fig.\u00a010e and f).\n74\n\nLater, Zhai and co-workers reported that the incorporation of S in the second sphere of Fe\u2013NC could enhance catalytic activity of oxygen electroreduction reaction via inducing the transition of spin polarization configuration. M\u00f6ssbauer spectroscopy showed that there are three different doublets (D1\u2212D3) existing in three FeNSC catalysts, corresponding to low spin (LS) Fe3+ (D1: X\u2013Fe3+N\u2013Y), high spin (HS) Fe2+ (D2: Fe2+N4) and high spin (HS) Fe2+ (D3: X\u2013Fe2+N\u2013Y), respectively (X and Y refer to S and C). Among them, Fe1-NS1.3C possessed more D1 moiety which had been proven to be active for ORR (Fig.\u00a011\nb). In order to clarify the active sites of Fe1\u2013NS1.3C for ORR, in-situ M\u00f6ssbauer spectroscopy was performed in O2-saturated 0.1\u00a0M KOH at room temperature (Fig.\u00a011a). As shown in Fig.\u00a011c, the D1 content decreased as the D3 content increased at the potential of 0.85\u00a0V (vs. RHE), indicating conversion of spin state from LS Fe3+ to HS Fe2+. At the potential of 0.65\u00a0V (vs. RHE), the D1 content decreased as the D2 content increased, indicating conversion of spin state from LS Fe3+ to HS Fe2+. While at the potential of 0.45\u00a0V (vs. RHE), both the D1 and D3 content decreased as the D2 content increased, implying LS Fe3+ and HS Fe2+ both acted as active sites. Combined with the results that the LS Fe3+ still converted into HS Fe2+ at the same potential without the existence of O2, it is proved that D1 active site is sensitive to O2 molecules and the LS Fe3+ of C\u2013FeN4\u2013S moiety could be the active site for Fe1-NS1.3C catalyst in alkaline ORR. DFT indicated that the doping of S impacted the spin polarization and adjusted the spin state of Fe center, resulting in the decrease of the adsorption free energy of \u2217OH, which is further enhanced the activity of ORR (Fig.\u00a011d\u2013f).\n115\n\nAdditionally, Zhang et\u00a0al. designed dual-metal atomically dispersed Fe,Mn/N\u2013C catalyst and revealed that the reduction of oxygen occurred preferentially on Fe(\u2162) in the intermediate spin state.\n122\n The analysis of M\u00f6ssbauer spectroscopy and DFT calculation proved that the implant of Mn\u2013N moieties led to Fe(\u2162) 3d electron delocalization and caused the spin state of Fe(\u2162) transition from low spin (t2g5eg0) to intermediate spin (t2g4eg1), which easily penetrated the antibonding \u03c0-orbitals of oxygen. Similarly, Zhai et\u00a0al. found that the introduce of Se could also tune charge redistribution and the spin-state of Fe active sites to improve the electrochemical performance for ORR.\n116\n Guo et\u00a0al. introduced Ti3C2T\nx\n as the support of iron phthalocyanine (FePc) and achieved a significant enhancement of ORR activity. Temperature-dependent magnetic susceptibility measurement results unveiled that the introduction of Ti3C2T\nx\n weakened the paramagnetic state of FeN4 moieties and increased the number of unpaired d electron of Fe(\u2161) ions, such that more occupied 3d electrons were easily transferred to antibonding \u03c0-orbital of oxygen. Compared to the pristine FePc, an additional D1 doublet appeared in the M\u00f6ssbauer spectrum of the Ti3C2T\nx\n-supported FePc, belonging to high spin Fe(\u2161). These evidence suggested that van Der Waals forces or hydrogen bonding between FeN4 moieties and Ti3C2T\nx\n induced electron density redistribution and spin-state transition and electron configuration transition from \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n1\n\nd\n\nx\nz\n\n\n1\n\nd\n\nz\n2\n\n\n1\n\nd\n\n\nx\n2\n\n\u2212\n\ny\n2\n\n\n\n1\n\n to \n\n\nd\n\nx\ny\n\n\n2\n\nd\n\ny\nz\n\n\n2\n\nd\n\nx\nz\n\n\n1\n\nd\n\nz\n2\n\n\n1\n\n was thought to be responsible for the enhanced ORR activity through yielding an easier dioxygen adsorption and reduction.\n117\n\nHydrogen peroxide is a versatile chemical, however, its industrial synthesis involves an energy intensive and tedious anthraquinone process.\n123\n Therefore, the synthesis of hydrogen peroxide by ORR offers a simpler and more sustainable approach to tackle this challenge. The production of H2O2 via 2e\u2212 ORR involves two proton-coupled electron transfer steps:\n\n(1)\n\n\n\u2217\u00a0\n+\n\n\u00a0O\n2\n\n+\n\n\u00a0H\n+\n\n+\n\ne\n\u2212\n\n\u2192\n\u00a0\u2217OOH\u00a0\n\n\n\n\n\n\n(2)\n\n\n\u2217OOH\n+\n\nH\n+\n\n+\n\ne\n\u2212\n\n\u2192\n\nH\n2\n\n\nO\n2\n\n+\n\u2217\n\n\n\n\nLiu et\u00a0al. proved that Fe\u2013N\u2013C adsorbed OOH\u2217 so strongly that the selectivity of H2O2 is low.\n124\n Although the Fe\u2013N\u2013C has poor selectivity for hydrogen peroxide, the selectivity can be improved after doping and other treatments due to the modification of metal center. Xue et\u00a0al. prepared a cheap sodium ferric EDTA-derived material (EDTAFeNa-KB-HT1), an iron and nitrogen co-doped Fe\u2013N\u2013C catalyst with micro-mesoporous structure and large surface area via one-step pyrolysis. The catalyst exhibited high selectivity (80%\u2013100%) towards H2O2 with a large current density.\n125\n Choi et\u00a0al. used H2O2 to introduce oxygen functional groups to the carbon surface of Fe\u2013N\u2013C, which changed the reaction path of ORR and led to a 30% increase towards H2O2.\n126\n Yagi et\u00a0al. synthesized Cu-, Fe-, and N-doped carbon nanotubes, (Cu, Fe)\u2013N-CNT as an ORR catalyst. It showed a high selectivity of 99% towards H2O2. Kinetic analysis revealed that the rate constant for the reduction of O2 to H2O2 is two orders of magnitude higher than that for the reduction of O2 to H2O.\n127\n Besides, H2O2 could be trapped in the micropores of powders and further reduced to H2O. Therefore, constructing a freestanding SACs electrode could prevent the further reduction of H2O2 and improve the selectivity.In summary, we concluded common methods for the preparation of Fe\u2013N\u2013C single-atom catalysts, including high-temperature pyrolysis, chemical vapor decomposition (CVD) and ball milling. Then, the relationship between the electronic structure of Fe single atoms and spin configurations is outlined, and based on this, the common methods of spin regulation are briefly mentioned, and the electronic structure of Fe and the orbital coupling of C, N and O are summarized. Finally, the application of Fe\u2013N\u2013C in electrocatalytic C, N and O conversion is introduced, including CO2RR, NRR, NitRR and ORR. Here, we would like to present the challenges and outlook of Fe\u2013N\u2013C, which may provide some opportunities for the future development of SACs.\n\n(1)\nThe loadings of catalysts synthesized with the existing methods are still not high, so it becomes an important task to find more reasonable and effective means to synthesize catalysts with high site densities. It should not be overlooked that the increase in loading can cause inferior mass transfer, thus the structural design of the catalyst is particularly important.\n\n\n(2)\nA reasonable Fe\u2013N bond length can effectively prevent the aggregation of iron atoms and increase the stability of the active sites. However, there is a lack of study in such topic. More efforts shall be devoted to exploring the relationship between catalyst structure and reaction performance from a smaller scale.\n\n\n(3)\nThe structures of SACs do not maintain unchanged during the reactions. Therefore, it is important to develop new techniques to achieve in situ characterization of catalysts to disclose the underlying mechanism and discover better catalysts.\n\n\n(4)\nThe spin regulation of Fe\u2013N\u2013C by external experimental conditions is seldom studied at present. For example, the modification of catalytic properties by applying a magnetic field in the electrocatalytic process has not received enough attention.\n\n\n(5)\nThe spin regulation of Fe\u2013N\u2013C by changing prepared conditions remains to be further developed. Although many reports expounded temperature, heteroatom doping, and other factors can change the spin state of the iron center, the mechanism is not clearly explained. In addition, it is unknown whether some conditions such as the applied magnetic field will exert an effect on the spin state.\n\n\n(6)\nDFT calculations were reasonably used to predict the spin electron transfer trends on different reactions of Fe\u2013N\u2013C prepared under different conditions. Proper utilization of the DFT calculations can lead to suitable thermodynamic data,\n128\n\n,\n\n129\n such as adsorption energy, dissociation energy and splitting energy, which may provide new possibilities for the design of Fe\u2013N\u2013C catalysts and mechanistic explanations.\n\n\nThe loadings of catalysts synthesized with the existing methods are still not high, so it becomes an important task to find more reasonable and effective means to synthesize catalysts with high site densities. It should not be overlooked that the increase in loading can cause inferior mass transfer, thus the structural design of the catalyst is particularly important.A reasonable Fe\u2013N bond length can effectively prevent the aggregation of iron atoms and increase the stability of the active sites. However, there is a lack of study in such topic. More efforts shall be devoted to exploring the relationship between catalyst structure and reaction performance from a smaller scale.The structures of SACs do not maintain unchanged during the reactions. Therefore, it is important to develop new techniques to achieve in situ characterization of catalysts to disclose the underlying mechanism and discover better catalysts.The spin regulation of Fe\u2013N\u2013C by external experimental conditions is seldom studied at present. For example, the modification of catalytic properties by applying a magnetic field in the electrocatalytic process has not received enough attention.The spin regulation of Fe\u2013N\u2013C by changing prepared conditions remains to be further developed. Although many reports expounded temperature, heteroatom doping, and other factors can change the spin state of the iron center, the mechanism is not clearly explained. In addition, it is unknown whether some conditions such as the applied magnetic field will exert an effect on the spin state.DFT calculations were reasonably used to predict the spin electron transfer trends on different reactions of Fe\u2013N\u2013C prepared under different conditions. Proper utilization of the DFT calculations can lead to suitable thermodynamic data,\n128\n\n,\n\n129\n such as adsorption energy, dissociation energy and splitting energy, which may provide new possibilities for the design of Fe\u2013N\u2013C catalysts and mechanistic explanations.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.We are grateful for the financial support from National Natural Science Foundation of China (No. 21974103) and the start-up funds of Wuhan University.", "descript": "\n                  Single atom catalysts (SACs) are constituted by isolated active metal centers, which are heterogenized on inert supports such as graphene, porous carbon, and amorphous carbon. The thermal stability, electronic properties, and catalytic activities of the metal center can be controlled via manipulating the neighboring heteroatoms such as nitrogen, oxygen, and sulfur. Due to the atomical dispersion of the active catalytic centers, the amount of metal required for catalysis can be decreased. Furthermore, new possibilities are offered to easily control the selectivity of a given transformation process as well as to improve turnover frequencies and turnover numbers of target reactions. Among them, Fe\u2013N\u2013C single atom catalysts own special electronic structure, and have been widely used in many fields of electrocatalysis. This review aims to summarize the synthesis of Fe\u2013N\u2013C based on anchoring individual iron atoms on carbon/graphene. The spin-related properties of Fe\u2013N\u2013C catalysts are described, including the relation between spin and electron structure of Fe\u2013N\n                        x\n                      as well as the coupling between electronic structure of Fe\u2013N\n                        x\n                      and electronic (orbit) of CO2, N2 and O2. Next, mechanistic investigations conducted to understand the specific behavior of Fe\u2013N\u2013C catalysts are highlighted, including C, N, O electro-reduction. Finally, some issues related to the future developments of Fe\u2013N\u2013C are put forward and corresponding feasible solutions are offered.\n               "}